TABLE OF CONTENTS

Chapter No.

Al

The Work of the Aerospace Structures Engineer.

..

...
A2 A3 A4 A5 A6 A7

STATICALLY DETERMINATE STRUCTURES

(Loads. Reactions, Stresses. Shears. Bending Moments, Deflections> Equilibrium of Force Systems. Truss Structures. Externally Braced Wings. Landing Gear. Properties of Sections - Centroids. Moments of Inertia, etc. General Loads on Aircraft. Beams - Shear and Moments. Beam - Column Moments. Torsion - Stresses and Deflections. Deflections of Structures. Castigliano's Theorem. Virtual Work. Matrix Methods.

r
j,

;,

THEORY AND METHODS FOR SOLVING STATICALLY INDETERMINATE STRUCTURES

AS A9 A1Q All A12

Statically Indeterminate Structures. Theorem of Least Work. Virtual Work. Matrix Methods. Bending Moments in Frames and Rings by Elastic Center Method. Column Analogy Method. Continuous Structures' Moment Distribution Method. Slope Deflection Method.

BEAM BENDING AND SHEAR STRESSES. MEMBRANE STRESSES. COLUMN AND PLATE INSTABILITY.
A13 A14 A15 A16 A17 A18 Bending Stresses. Bending Shear Stresses - Solid and Open Sections- Shear Center. Shear Flow in Closed Thin-Walled Sections. Membrane Stresses in Pressure Vessels. Bending of Plates. Theory of the Instability of Columns and Thin Sheets.

INTRODUCTION TO PRACTICAL AIRCRAFT STRESS ANAL YSIS

A19 A20 A21 A22 A23 Introduction to Wing Stress Analysis by Modified Beam Theory. Introduction to Fuselage Stress Analysis by Modified Beam Theory. Loads and Stresses on Ribs and Frames. Analysis of Special Wing Problems. Cutouts. Shear Lag. Swept Wing. Analysis by the "Method of Displacements".

THEORY OF ELASTICITY AND THERMOELASTICITY

A24 A25 A26

The 3-Dimensional Equations of Thermoelasticity. The 2-Dimensional Equations of Elasticity and Thermoelasticity. Selected Problems in Elasticity and Thermoelasticity.

;...,.....

l _ ~

....
.j

1-3;

TABLE OF CONTENTS Continued

Chapter No.

FLIGHT VEHICLE MATERIALS AND THEIR PROPERTIES

81

82

Basic Principles and Definitions. Mechanical and Physical Properties of Metallic Materials for Flight Vehicle Structures.

STRENGTH OF STRUCTURAL ELEMENTS AND COMPOSITE STRUCTURES

Cl C2 C3 C4 CS C6 C7 C8 C9 C10
ell

Combined Stresses. Theory of Yield and Ultimate Failure. Strength of Columns with Stable Cross-Sections. Yield and Ultimate Strength in Bending. Strength and Design of Round. Streamline, Oval and Square Tubing in Tension, Compression. Bending,
Torsion and Combined Loadings.

C12 C13

Buckling Strength of Flat Sheet in Compression, Shear, Bending and Under Combined Stress Systems. Local Buckling Stress for Composite Shapes. Crippling Strength of Composite Shapes and Sheet-Stiffener Panels in Compression. Column Strength. Buckling Strength of Monocoque Cylinders. Buckling Strength of Curved Sheet Panels and Spherical Plates. Ultimate Strength of Stiffened Curved Sheet Structures. Design of Metal Beams. Web Shear Resistant (Non-Buckling) Type. Part 1. Flat Sheet Web with Vertical Stiffeners. Part 2. Other Types of Non-Buckling Webs. Diagonal Semi-Tension Field Design. Part 1. Beams with Flat Webs. Part 2. Curved Web Systems. Sandwich Construction and Design. Fatigue.

CONNECTIONS AND DESIGN DETAILS

01 02 03 Fittings and Connections. Bolted and Riveted. Welded Connections. Some Important Details in Structural Design.

Appendix A

Elementary Arithmetical Rules of Matrices.

INDEX
Accelerated MoUon of Rigid Airplane. Aircraft Bolts . Aircraft Nuts Aircrait Wing Sections -

A4. 8
D1.2 D1. 2

Type, A19. I Aircraft Wing Structure Truss Type. A2.14 Air Forces on Wing . A4.4 Allowable Stresses (and Interactions) . Cll.36 Analysis of Frame with Pinned Suppor-ts . A9.l6 Angle Method C7.1 Application of Matrix Methods to Various Structures A7.23 A4.1 Applied Load Axis of Symmetry. A9.4
Beaded Webs Beam Design - Special Cases. Beam Ftxed End Moments by Method of Area Moments Beam Rivet Des Ign Beam Shear and Bending Moment Beams - Forces at a Section Beams Moment Diagrams. Beams with Non-Parallel Flanges Beams - Shear and Moment Diagrams aeams Statically Determinate & Indeterminate . Bending and Compression of Columns Bending Moments - Elastic Center Method. Bending of Rectangular Plates Bending Strength - Basic Approach. Bending Strength - Example Problems Bendtng Strength of Round Tube' Bending Strength - Solid Round Bar. Bending Stresses Bending Stresses - Curved Beams Bending Stresses - Elastic Range Bending Stresses - Nonhcmcgenecus Sections. Bending Stresses About pr-mcipal Axes . Bending of Thin Plates Bolt Bending Strength. Bolt & Lug Strength AnalysiS Methods Eolt Shear-, Tension &
> >

Buckling of Stiffened Flat Sheets under Longitudinal Compression Buckling under Bending Loads Buckling under Shear Loads. Buckling under Transverse Shear Carry Over Factor casngnaao'e Theorem Centroids - Center of Gravity. Cladding Reduction Factors. Column Analogy Method. Column Curves - NonDimens tonal . Column Curves - Solution Column End Restraint. Column Formulas . Column Strength. Column Strength with Known End Restraining Moment Combined Axial and Transverse Loads - General Action COmbined Bending and Compression. Combined Bending and Flexural Shear. . Combined Bending and Tension Combined Bending and Tension or Compression of Thin Plates Combined Bending & Torsion. Combined Stress Equations . Compatability Equations. Complex Bending' Symmetrical Section. Compressive Buckling Stress for Flanged Elements . COnical Shells - Buckling Strength Constant Shear Flow Webs Constant Shear Flow Webs ~ Single Cell - 2 Flange Beam. Constant Shear Flow Webs Single Cell ~ 3 Flange Beam. Continuous Structures Curved Members Continuous Structures Variable Moment of Inertia COre Shear Correction for Cladding. Corrugated Core Sandwich Failure Modes. Cozzone Procedure Creep of Ma.terials Creep Pattern . Crippling Stresses Calculations . Critical Shear Stress Crystallization Theory Cumulative Damage Theory. Curved Beams Curved Sheet Panels Buckling Stress Curved Web Systems Cut-OUts in Webs or Skin Panels . Deflection Limitations in Plate Analyses. . . Deflections by Elaanc Weights

C6.4 CS.6 CS.6 C8.14

All. 4 A7.5 A3.1 C5.5 AlO.l

C2.2 C2.l3 C2.1 C4.2 C7.21 C2.16 A5.21 C4.22 C3.l0 C4.23
A18.l7 C4.23

CIO.16 D3.10 A7.32 ClO. B AS.I A5.7 AS. 6

Deflections by Moment Areas. A7.30 Deflections for Thermal Strains . A7.17 Deflections by Virtual Work A7.9 Delta Wing Example Problem. A23.2 Design for Compression C4.2 Design Conditions and Deaign Weights .. A5.12 Design Flight Requirements for Airplane . A4.6 Design Loads . A4.1 Design for Tension . . C4. 1 Differential Equation of ALB. 12 Deflection Surface. AZD.15 Dtsconttnutttes Distribution of Loads to Sheet Panels . A2l.2 Ductility. . . B1. 5 Dummy Unit Loads A8.6 Dynamic Effect of Air Forces. M.13 Effect of AxIal Load on Moment Distribution. Effective Sheet Widths Elastic Buckling Strength of Flat Sheet in Compression. Elasttc - Inelastic Action. Elastic Latera! SUpPOrt Columns. Elastic Stability of Column Elastic Strain Energy . Elasticity and Thermoelasticity - One-Dimensional Problems ... Elasticity and Thermoelasticity - Two-Dimensional Equations Electric Arc Welding . . End Bay Effects. End Moments for Continuous Frameworks . Equations of Static Equilibrium Equilibrium Equations

All. 22 C7.l0
C5.1

Bt. 5
C2.l7 .'\17.2 ci, 6
A26.l

en, 9
AS.2 A5.1 AlB. 1 A9. 1 Ala. 13 <:3.1 C3.4

ci. 2

A24.7 C3.9 C6.1

CB.22 A14.l0

A25.1 D2.2 Cll.23

All. 10

A2.l

A24.2

cr. 15
C3.1 A13.1
A13. 15

A15.3
A15.5 All. 31 All. 15 C12.26 C7.4 e12.27 C3.2
31. 8

A13. 13 A13.11 A13.2 AlS. 10

D1.9 D1.5

B1. 12

C7.7 Cll.16 C13.1

C13.3 A5. a

Ber:ding Str-engths

D1.3

Boundary Conditions Box Beams Analysis Brazing Buckling Coefficient Buckling of Flat Panels wiui

A24.8
A22.5
D2.4

CS.l
C12.25 CS.6

C9.1 Cll.29
D3.7

ntsetmuer Feces
Buckling of Flat Sheets under Combined Loads. Buckli.ng or Rectangular Plates

A18.20

A17.4 .\7.27

Failure of Columns by Compression. A18.4 Failure Modes in Curved Honeycomb Panels. . . C12.20 FaUure of Structures 81. 1 Fatigue AnalySiS - Statistical Distri.bution ... . C13.4 Fatigue and Fail-Safe Design . C13.8 Fatigue of Materials B1. 14 Fatig-le S-N Curves . C13.13 Fillers. 03.5 Fitting Design . 01. ~ Fixed End Moments All. :3 Fjxed End Moments Due to Support Deflections All. 9 Fixi.ty cceutcterus. C2. 1 Flange Design . CIO.l Flan~e Design Stresses . ClG.2 Flange Discontinuities. ClO.7 Flange Loads Cll. a Flange Strength (Crippling) . CIO.4 Flat Sheet Web with Vertical Stiffeners CIO.l Flexural Shear Flow Distribution AlS.24 Flexural Shear Flow Symmetr-ical Beam Section A14.5 Flexural Shear Stress. A14.1

INDEX Continued
Static Tension StressStrain Diagram . . . Bl.2 Statically Determinate Coplanar Structures and Loadings . A2.7 Statically Determinate and Indeterminate Structures A2.4 Statically Indeterminate Frames - Jomt Rotation Al2.7 Statically Indeterminate Problem . . . . . . AB.l Stepped Column - Strength C2.14 Stiliened Cylindrical Structures - illUmate 3trength . . . . C9.S Stiffness &I Carry-over Factors lor CUrved Members All. 30 Stiffness Factor. . . All. 4 Strain - Displacement Relations . . . . . A24.5 Strain Energy . . . . . A7.l Strain Energy of Plates Due to Edge Compression and Bending A18.19 Strain Energy In Pure Bending of Plates. . . . . . . . AlB. 12 Streamline Tubing - Strength. C4. 12 Strength Checking and Design - Problems . . . C4.5 Stren~_ ".: Round Tubes _ ..nder Combined Loadings . C4.22 sn-ess Analysis Formulas Cll. 15 Stress Analysis of Thin Skin Multiple stringer Cantilever Wing . . . . . . . . . . A19.10 Stress Concentration Factors. C13.10 Stress Distribution & Angle of Twist for 2-Cell Thin Wall Closed Section . . A.6.7 Stress-Strain Curve . . B1.7 Stress-Strain Relations . A24.6 Stresses around Panel Cutout. A22.1 Stresses in Uprights Cll.17 Stringer Systems in Diagonal Tension . . . . . . Cll.32 Structural Design Philosophy. ci, 6 Structural Fittings . . A2.2 Structural Skin Panel Details. D3.12 Structures with Curved Members . . . . . All. 29 Successive Approximation Method for Multiple Cell Beams . . . . . . . AlS.24 Symbols for Reacting
4

ToUlgent Modulus Tangent-Modulus Theory Taxi Loads . . . Tension CHps . . . . . Tension-Field Beam Action. 'renetcn- Field Beam Formulas Theorem of Castiglta.no . . Theorem of Complementary Energy. . . . . . Theorem of Least Work . Theorems of Virtual Work and Minimum Potential Energy Thermal Deflections by MatriX Methods . . Thermal Stresses . . Thermal Stresses . . . Thermoelasttcity - ThreeDimensional Equations. Thin Walled Shells Three Cell - Multiple Flange Beam. - Symmetrical about
One Axis . . . .

Bl.5
A1B.a C13.33 D3.2 en. 1 C11. 2 A7.5

A7.5 A8.2 A7.5 AB.39 A8.14 AB.33 A24.1 A16.5

AIS. 15

Trusses With Multiple Redundancy . . . A8.11 Trusses With Single Redundancy . . . . AS. '7 Tubing Design Facts C4.5 Two-Dimensional Problems. A.26. 5 Two-Cell Multiple Flange Beam. One Axis ot Symmetry . . . . Al5.11 Type of Wing Ribs. . A2l.t tntimate Strength in Combined Bending &: F1exural Shear tnUmate Strength in Combined Compression, Bending, Flexural Shear &: Torsion. Ultimate Strength in Combined Compression, Bending &I Torsion . . . illtimate Strength in Combined Tension. Torsion and Internal Pressure p in psi. Uniform Stress Condition. Unit Analysis lor Fuselage Shears and Moments. . UnsymmetriCal Frame . . Unsymmetrical Frames or Unsymmetrical Frames using PrinCi~ Axes. . . .'Tnsymmeirical Structures " , \t, . 'Jy - Load Factor " .("rC4.25 C4.26

Ct. 24
C4.26 Cl.I
AS. IS" A9.2
AIO.4

Three Flange - Single Cell

Wing

Torsion - Circular Sections. Torsion - Effect of End Restraint . . . . . Torsion - Non-circular Sections . . . Torsion Open sections Torsion of Thin-Wailed Cylinder having Closed Type StiHeners . . . . . . . . . Torsion Thin Walled Sections. Torsional Moments - Beams Torsional Modulus of Rupture. Torsional Shear Flow in Multiple Cell Beams by Method of Successive Corrections . . . Torsional Shear Stresses in Multtple Thin-Wall Closed Section - Distribution Torsional Strength of Round
4Cell

Al9.5 A6.l

A8.16 A6.3
AS. 4

R_

M.tS A9.13

.'

AS. IS

A6.S
A5.9 C4.1'7

.....

A4.7

':-"'A

:''"it

A6.10
AS.7

Tubes Tubes

.....
4

C4.17 A6.S A6.2 Cl.5

Torsional Stresses 1n Muitiple Cell Thin-Walled

.......

Fitting Units . . . .

Symmetrica sections External Shear Loads

A2.3
A14.2

Transmission of Power by Cylindrical Shaft. . . . TriaXial Stresses . . Truss Deflection by Method of Elastic Weights Truss Structures Trusses with Double Redundancy. . .

A7.33

A2.9
AS. 10

Wagner Equations .. Cl1.4 Web Bending &: Shear Stresses CtO.5 Web Design . . . Cll.18 Web Splices . . . . CIO. 10 Web Strength. Stable Webs. CtO.5 Webs with Round Lightening. Holes . . ClO.17 Wing Analysis Problems A19.2 Wtng Arrangements. Al9.1 Wing Effective Sectton A19.12 Wing Internal Stresses A23.14 Wing Shear and Bending AnaLysis . . . A19.14 Wing Shear and Bending Moments . . . . A5.9 Wing - Sbear Lag . A19.25 Wing Shears and Moments AS. 10 Wing Stiffness Matrix.. A23.11 Wing Strength ReqUirements A19.5 Wing Stress Analyl'lis Methods .'1.19.5 Wing - Ultimate Strength . A19.11 Work of Structures Group. Al.2

Y Stiffened Sheet Panels

C7.2C

CHAPTER Al

THE WORK OF THE AEROSPACE STRUCTURES ENGINEER

AI. 1 Introduction.

The first controllable human flight in a heavier than air machine was made by Orville Wright on December 17, 1903, at Kitty Hawk, North Carolina. It covered a distance ot 120 feet and the duration of flight was twenty seconds. Today, this initial flight appears very unimpressive, but it comes into its true perspective of Lmportance when we realize that mankind for centuries has dreamed about dOing or tried to do what the ~rlght Brothers a:campllshed in 1903. The tremendous progress accompliShed in the first 50 years of aviation history, with most of i t occurring in the last 25 years, 1s almost unbelievable, but without doubt, the progress in the second 50 year periOd will still be more ~~believable and fantastic. As this is written in 1964, jet airline transportation at 600 MPH is well established and several types of military aircraft have speeds in the 1200 to 2000 ~ range. ?reliminarJ designs of a supersonic airliner with Mach 3 speed have been ccmpleted ~~d the government is on the verge of sponsoring the development of such a flight vehicle, thus supersonic air transportation should become co~on in the early 1970's. The rapid progress i~ ~!ssile design has ushered in the Space Age. Already many space vehicles have been flown in search of new knowledge which is needed before successful exploration of space such as landings on several planets can take place. Unfortunately. the rapid development of the missile and rocket power has given mankind a flight vehicle when combined with the nuclear bomb, the awesome potential to quickly destroy vast regions of the earth. TMhile no person at ~resent ~~ows where or what space exploration will lead to, relative to benefits to ~nkind, we do know that the next great aviation expanSion besides supersonic airline transportation will be the full develop~ent and use of vertical take-off and landing aircraft. Thus persons who will be living through the second half century of aviation progress will no doubt witness even more fantastic progress than oceurred in the first 50 years of aviation history.
A!. 2 General Organization of an Aircraft Company Engineering Dfvtetcn,

scientific machine and the combined knowledge and experience of hundreds of engineers and scientists working in close cooperation is necessary to insure a successrul product. Thus the engineering division of an aerospace company consists of many groups of specialists whose specialized training covers all fields of engineering education such as PhySics, Chemical and Metallurgical, MeChanical, Electrical and, of course, Aeronautical ~~lneering. It so happens that practically all the aerospace companies publiSh extensive pamphlets or brochures explaining the organization of the engineering division and the duties and responSibilities of the many sections and groups and illustrating the tremendous laboratory and test facilities which the aerospace industry possesses. It is highly recommended that the student read ~~d study these tree publications in order to obtain an early general understanding on how the ~odern flIght vehicle is conceived, deSigned and then prOduced. In general, the engineering department of an aerospace company can be broken down into sax large rather distinct sections, which in turn are further divided into specialized groups, which in turn are further divided into smaller working groups of engineers. To illustrate, the six sections will be listed together with some at the various groups. ThiS is not a complete list. but it should give an idea or the broad engineering set-up that is necessarJ. I. II. Preliminary Design Section. TecrJlical (1) (2) (3) (4) (5) (5;
_~alysis

Section.

Aerodynamics Group Structures Group ~eight and 3alance Control Group Power Plant Analysis Group Materials and Processes Group Centrols AnalYSiS Group

III. Component DeSign Section. (1) (2) Structural DeSign Group (~lng. Body and Control Surfaces) Systems Design Group (All mechanical, hydraulic, electrical and ther.nal installations) Tests Section.

The ~odern commercial airliner, militarJ airplane, missile and space vehicle is a highly
Al.I

IV.

Laborato~J

Al.2

THE WORK OF THE AEROSPACE STRUCTURES

ENGINEER

(1)
(2)

(3)
(4) (5) (6) (7)

Wind Tunnel and Fluid Mechanic5 ~est Labs. Structural Test Labs. Propulsi~n Test Labs. Electronics ~est Labs. Electro-Mechanical Test Labs. Weapons and Controls Test Labs. Ar~log and Digital Computer Labs.

The final results of t.he work of this group are formal reports glv~n~ complete a~plied load design criteria, with ~ny graphs ~nd swumary tables. The final results ~y 61v8 complete shear, moment and no~l forc~s =e~er=~d to a convenient set of :CY2 axes for major airc ra.r t units such as the Wing, rus eIage , e t c . THE WORK OF STRESS ANALYSIS
~\m S~R~~GTH

v.
VI.

GROUP

Flight Test Section. Engineering Field Service Section. Essentially the primary job of :he stress group is to help specify or deter.nine the kind of material to use and the :h:c~~ess, size and cross-sectional shape Jt every struct~l ~eQ ber or unit on the airplane or ~issile, and also to assist in the deSign of all jOints and connections for such ~embers. safety with ~ight weight are the paramount str~ctural jesl~ requirements. ~he stress group ~ust consta~tly work closely with the Structural DeSign Sect:Gn in order to evolve the best structural over-all arrangement. Such factors as ~ower ~lants, bUilt in fuel tanks, landing gear retracting wells, and other large cut-outs can d~ctate the type of wing structure, as for example, a two spar single cell wing, or a multiple spar ~ultiple cell wing. To expedite the initial struct~r~l ~esign studies, the stress group ~ust s~~ply initial structural sizes based on approximate loads. The fi~l results of the work by the stress group are recorded in elaborate reports which show how the stresses were calculated and hew the reqUired member sizes were obtained to carry these stresses efficiently. The r:nal size of a member may be dictated by one or more rae tors such as elastic action, tne Ias t t c action, elevated temperatures, fatigue, etc. To insure the accuracy of theoretical calculations, the stresS group must have the assistance of the structures test laboratory in order to obtain information on which to base allowable design stresses. THE WORK OF THE DYNAMICS
A~LYSIS

SinCe this textbook deals with the subject of structures, it seems appropriate to discuss in some detail the work of the Structures Group. For the detailed discussion of the other grou~s, the student should refer to the various aircraft company publications.
At. 3 The Work of the Structures Group

The structures group, relative to number of engineers, is one of the largest of the ~any groups ot engineers trat make up Section II, the technical analySis section. The structures group is primarily responsible for the structural integrity (safety) ot the airplane. safety may depend on sufficient strength or sufficient rigidity. This structural integrity must be accompanied with lightest pOSSible weight, because any excess weight has detrimental etfect upon the perfo~ce of aircraft. For example, in a large, long range missile, one pound of '~ecessary structural weight may add mora than 200 Ibs. to the overall weight of the missile. The structures group is usually divided 'into sUb-groups as tollows:(1) (2) (3)
(4)

Applied Loads Calculation Group Stress AnalySiS an~ Strength Group Dynamics AnalYSiS Group Special Projects and Research Group

GROUP

THE '"ORK OF THE APPLIED LOADS GROUP

Before any part ot the structure can be finally proportioned relative to strength or rigidity, the true external loads on the aircraft must be determined. Since critical loads came tTom many sources, the Loads Group must analyze loads fram aerOdynamiC forces, as well as those forces from power plants, aircraft inertia; control system actuators; launching, landing and recovery gear; a~ent, etc. The etrects of the aerOdynamic forces are initially calcUlated on the assumption that the airplane structure 1s a rigid bOdy. Atts: the aircraft structure is Obtained, its true rigidity can be used to obtain dynamic effects. Results of wind t~~el model tests are usually necessary in the application of aerodynamic principles to load and pressure analYSiS.

The Dyna~ics AnalysiS Grou) has rapidly expanded in recent years ~elative to number of engineers required because supersonic airplanes, missiles and vertical riSing ai~craft have presented many new and complex problems in the general field of dynamics. In some airc~tt companies the dynamiCS group 1s set up as a separate group outside the Structures Group. 7he engineers in the dynamiCS group are responSible for the investigation ot Vibration and shOCk, aircraft flutter and the establish~ent of desig~ requirements or c~2nges for its control or correction. Aircraft contain dozens of mechanical installations. Vicration of ~~y part of these installations or systems ~y be of such character as to cause faulty operation or danger of failure and therefore the dynamic

ANALYSIS

AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A1.3

groups have a speCial SUb-group which are working on deSign problems that Nill be encountered in the near 1r distant future as aviation proThe major structural units of aircraft such gresses. For example, in the r.t r-uc tur-ea Group, as the wing and fuselage are not rigid bodies. this sub-group might be studying such problems 7hus when a Sharp air gust strikes a fleXible as: (1) how to calculate the thermal stresses wing in high speed flight, we have a dynamic in the wing structure at super-sonic speedS; load situation and the wing Nill vibrate. The (2) how to stress analyze a new type of wing dynamicist must determine whether this vibration structure; (3) what type of body str~cture is 1s serious relative to induced stresses on the best for future space travel and what kind of wir~ structure. The dynamics group is also materials will be needed, etc. responsible for the determination of the stability and performance of miSSile and flight Chart 1 illustrates in general a typical vehicle guidar.ce and control systems. The ~ke-up of the Structures Section of a large dynamics group must work constantly with the aerospace company. Chart 2 lists the many various test laboratories in order to obtain items which the structures engineer must be reliable values of certain factors that are concerned with in insuring the structural necessary in many theoretical calculations. integrity of the flight vehicle. Both Charts 1 and 2 are from Chance-Vought Structures THE ',jaRK OF THE SPEC IAL PROJECTS GROUP DeSign Manual and are reproduced with their permi saton. In general, all the various technical

characteristics must be changed or modified in order to insure reliable and safe operation.

U..f ,. ~EIlOE'-'\STlc:'4

~'IlOE'-'\STIC cU"", i

,..- --l ,

11:'
'
lAKlIlATo'" I
~~u,

I STWCTVI~S
""

I.

'liSI U,,"T "'",ru,,,

, -.c...."I.IC
'NO
: UN"

"Owu

' ..... 1 ""ST

, .....c:..
~~,

'H~

~'''G''

COMPUIATICH GIlOU'

Chart 1. Structures Section Organization Chance- Vought Corp.

-:3

e.13W

ETT?

5ZWF

AI. 4

THE WORK OF THE

AEROSPACE STRUCTURES

ENGINEER

THE LINKS TO STRUCTURAL INTEGRITY

ARE NO BETTER THAN THE WEAKEST LINK

MATERIALS OF CONSTRUCTION
FA.STENERS WELDING BONDING ?LATE~ND BAR FORGINGS CASTINGS XTRUSIONS SHEET METAL SANDWICH ?tASTIC l,OMINAIT BEARINGS

STIFFNESS CRITERIA
FLUTTER
CONTROL SYSTEM STABILITY PANEL fLIJTTER-SKIN CONTOURS CONTROL SYSTEM DEfLECTIONS

STRESS ANALYSIS
SKIN PANElS BEAM ANALYSIS STRAIN COMPATIBILITY STRAIN CONCENTRATION JOINT ANALYSIS BEARI~G ANALYSIS BULKHEAD ANALYSI S fITTING ANALYSIS i1'ERMAL STRESS ,'oIECHAN1CAL COMPONENTS D:PERIMENTAL STRESS ANALYSIS

THERMAL EFFECTS
MtCHANICAL VIBRATIONS

ROLL POWR-O IVERGENCE AEROtWtAMIC CENTER SHIFT DYNAMIC RESPONSE

LOADS AND ENVIROMENT

FLIGHT LOAD CRITERIA
GROUND LOAD eRITER IA FLIGHT LOAD DY~M'CS LAUNCHING DYNAMICS LANDING DYNAMICS DYNAMIC RESPONSE aECOVERY DYNAMICS ,UGHT LOAD DISTRIBUTIONS INERTIAL LOAD DISTRIBUTIONS Fl..D:IBILITY EFFECTS GROUND LOAD DISTRIBUTIONS REPEATEO LOAD SPECTRUMS TEMPERATURE DISTRIBUTIONS LOADS FROM TllERMAL DEFORMATIONS PRESSURES-IMPACT

MATERIALS AND QUALITY CONTROL

DUCTILITY STRESS-STRAIN HOMOGENEOUS MATERIAL

COMPONENT ANALYSIS
UNIT SOLUTIONS

ALLOWABLES
YIHDING FRACTURE ATlGUE WEAR, BRINELLING CREEP DEFliCTIONS Tl'ERMAl O"FECTS STIFFNESS COMBINED LOADINGS 3UCKlING

RESIDUAL STRESS HEAT TREAT CONTROL STRESS CORROSION STABILITY AT TE~PRATURE SPECIFICATION CONFORMANCE BLUE PRINT CONFORMANCE

INOffiRMINATE STRUCTURES WING ANALYSIS

TAIL ,l,NAlYS1S

FUSElJ.GE SHElL A""lYSIS

THERMAL ANALYSIS

DfFlICTION ANALYSIS
STIFFNESS

Chart 2 From Chance- Vought Structures Destgn Manual

CHAPTER A2

EQUILIBRIUM OF FORCE SYSTEMS. TRUSS STRUCTURES

A2.1 Introduction.

The equations of static

equilibrium must constantly be used by the stress analyst and structural designer 1~ obtaining unknown forces and reactions or unkno~n internal stresses. They are necessary whether the structure ..or machine be s tmp.Le or complex. The ability to apply these equations 1s no doubt best developed by solving ~ny problems. This Chapter l~ltlates the application of these important phySical laws to tt.e force and stress ar~lysls of structures. It 1s assumed that a student has completed the usual college course in engineering mecrznlcs called statics.
A2.2 Equations of Static Equilibrium.

force system pass through a cammon point. The reSUltant, if any, must therefore be a force and not a moment and thus only 3 equations are necessary to completely define the condition that the resultant must be zero. The equat10ns of equilibrium available are therefore:-

ZFy = a
l:Fz

l:F x

=a

=0

or

m, = o m, = o m, = o

- - - - -(2.2)

To completely Cefl~e a force, we must know its ~agnltude, direction and ~olnt of ann1ication. These facts regaTding the ~arce are generally refer~ed to as the characteristics of the ~orce. So~etimes the more ~eneral te~ of line Of act~on or lecation is used as a force cr~racteristic in Place of paint of application designation.

A combir~tion of force and moment equations to make a total of not more than 3 can be used. For the moment equations, axes through the point of concurrency cannot be used since all forces of ~he system pass through this point. The moment axes need not be the same direction as the directirns used in the force equations but of course. they could be.
EQUILIBRIUM OF S?AC::; ?J..RALLEL FORCE SYSTlli

A force acting in space is completely defir-ed :: we %now its components in three directions and its ~oments about 3 axes, as for example FX J F~, Fz ~nd ~x, tly and Xz ~or equilibrium o~ a force system there can be no resultant force and thus the equations of equllibri'4n are obtained ~y equating the force and moment cCill~onents to zero. The equations ~f static equilibrium for the various types of force systems 'Nill now be suear-tzeo .
EQ.UILI3RIUi1 SO.UATICNS FOR GSNERAL SPACE: urCN-COPLANAR) ?ORCE SYSTE?:

In a parallel force system the direction of all forces is known, but the magnitUde and location of each is unknown. Thus to determine ~gnitude, one equation Is required and for location twa equations are necessary since the force is not confined to one plane. In general the 3 equations commonly used to make the resultant zero for this type of ~orce system are one force equation and two moment equatiOns. For example, for a space parallel force system acting in the y direction, the equations of equilibrium would be:
ZFy = 0,
SG.UILIBRIUM

ZI1x ::: 0,

n1z:=

- -(2.3)

OF GE:N""'..2.AL

CO-PUu'JAR ~ORCE SYSTEM

l:F x ZF y

= 0 = 0 ZF z = 0

m, = a m, = 0

~ ...

= 0

- - - - -(2.1)

rhus :or a general space ~orce system, there are 6 equations of static eqUilibrium available. T~ree of these and ~o ~or8 can be force equations. It is or t en acr-e convenient to ~ake the moment axes, 1, 2 ~nd 3. as any set ot X, y and z axes. All 6 equat10ns could be no~ent equaticns about 6 ciffere~t axes. ~~e force equations are written for 3 ~utually ~er?endiC~lar ~es ~nd need not be t~e x, y and z axes.
sq,UILI3RI'L11 OF' SPACE ::CNCJEP,S::T
Co~curre~~ 7;e~ns

In this type ot farce system all forces lie in one plane ~d it ta~es only 3 equations to deter.Aine the magnitUde, direction and location or the resultant of such a force system. Either torce or moment equations C3n be ~sed. except that a ~aximum of 2 :orce equa~lons can be used. For example, for a force system acting in the xy plane, the follcwiig co~blnaticn of equilibri~ equations could be used.

ZF y a zr1z .. = a ZF y or L:!":z..= 0 or ZMz ... :: 0 or 1:1z II = 0 z:M: Z ll = 0 Z!"!Zil= 0 z:1z = 0 mz" = a

ZFx

=0 =0

ZFx

=0

2.4

ropes

SYS':'~

that all forces of the

A2.1

(':'he subsc=-lpts 1, 2 and 3 refer to different locations for z axes or moment center-s . )

.AC

A2.2

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

Since all forces lie in the ss~e plane and also pass thr8ugh a c~~,on ~oint, the ~a~~itude and direction of the r~s~ltant of this type of force system is unknown ~ut the location ~s ~own sin~e the ;oint of conc'~iency is on the lin8 of action cf the resultant. ~hus only two equations 0: equilibri~ are necessary to define the resultant and ~ke it zero. T~e combinations available are,
l:F x

: 0 or ZFx = a or ZFy = a or ZMz~ =0 ZI'1z = O ZI'1z a ZFy : a

=0 }
= O

2.5

reacted 'Jy other ext er-ne ; -or-ces , ccmnorLy referred to as reactions which hold the j:,cwn forces on ~he str~cture in equiliJri~ll. Sl~ce the static equa~ions of equil:bri:2n ~vai:~~:e for the various t~?es ot force syste~ are linited, the str~c:ural engineer resorts tc the use of fitting units whic~ establ:sh ths direction 8f an ~n}~i8~TI fc~ce C~ ~~s ~o:nt Jf ap~11c8:~on or both, t~~s decreasins th3 ~~ber of UIL~owns t8 be determi~ed. ~~e ~l~ures which follow illustrate t~e ty;e of ~itt1~g units employed or o:her gene~~l Ieth~ds ~Jr establiShing the ~orcs c~aracterist1cs of dirsction and pOint of application. Ball and Socket Fitting

(The z axis or moment center locations must be other than through the paint of concurrency)
EQUILIBRIUM: OF CO-PLANAR P.4RALLt:L FCRCE
SYS'I'~

Since the direction of all forces in this type or rorcs system 1s known and since the forces all lie 1n the same plane, it only takes 2 equations to define the magnitude and location of the resultant of such a force system. Hence, there are only 2 equations of equil1bri~ available for this type of force system, namely, a force and :noment equation or two moment equations. For example, for forces parallel to y axis and located in the xy plane the equilibrium equations available would be: l:F y : 0 n1 z

For any space or coplanar f8rce suc~ as ? and Q acting on the bar, the line o~ action of such forces must act t..u-cucn the cent.cr- 0': the ball ~! rotation of the bar is prevented. ~hus a ball and socket joint can be used to establ:sh or control the di~ection and line action of a force applied to a struct~re through :~i~ :j~e of fitting. Since the ~oint has ~o rctaticTzl reSistance, no couples in any plane can be applied to it.

=0

or

l:l1z. : a
4I'1z
Ii

=a

-----2.6

(The moment centers 1 and 2 cannot be on the same y axis)

EQUILIBRITJr1 OF COLINEAR FORCE SYSTEM

~'l
-

II

A collnear torce system 1s one where all forces act along the same line or in other words, the direction and location of the forces is known but their nagnt tudes are unknown, thus only magnitude needs to be found to define the resultant of a collnear force system. Thus only one equation ot equilibrium is a'ffiilable, namely

For any force such as P and Q acting i~ the xy plane, the line of action of such a ~~rce
~ust

ZF

=a

or

ZI1 1 = O

----2.7

where moment center 1 is not on the line at action of the force system
A2.3 Structural Fitting Units for Establishing the Force Characteristics of Direction and Point of Application.

pass through the ,in center since t~e fitting unit cannot resist a ~cme~t about a z axlo through the pin center. Therefore, fJr forces acting in the xy plane, th~ direct:cn ~~d line of action are established by th8 pin joint as illustrated in the f:gure. Sir.ce a Single pin fitting can resist :noments atou: axes perpendicular to the ~in axis, the cirectlon and line of action of out of ,lane :orces is ~tere fore not established by s tng l e pfn ~it-:lng umt.s .

-<@o========:j):Jl_
If a bar AB has si~Gle pin f~t~ings at end, then any :orce P lying in the xy plane and anplied to end B ~ust have a direction and line of action co tnc tc t ng ....ith a line jo tn-. i~g the pin centers at end ~it~1ngs A ar.d 3, since the :lttings cannot resist a ~oment about the 3 axis.
eac~

,B

To completely define a force in space reqUires 5 equations and 3 equations if t~e force is limited to one ?lane. In ~eneral a structure is loaded by ~own forces ar.d these !orces are transferred tPIough the struct~re in some ~nner of internal stress distribution and then

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.3

Double Pin - Universal Joint Fittings

z
Q

L.
y

Fh"
liillirn

/-/11

Another general fitting type that is used to establish the direction of a force or reaction is illustrated in the figure at the bottom of the first column. Any reacting force at joint (A) ~ust be horizontal since the support at (A) is so designed to provide no vertical resistance. Cables - Tie Rods

L
Since single pin fitting units can resist applied moments about axes normal to the pin axiS, a double_pin joint as illustrated above is otten used. T~is fitting unit cannot resist ~oments about y or z axes and thus applied forces such as P and Q ~ust have a line of action and direction such as to pass through the center of the fitting unit as illustrated in the figure. The fitting unit can,. however, resist a moment about the x axis or in other woras, a universal type of fitting unit can resist a torsional ~oment. Rollers

Since a cable or tie rod has negligible bending resistance, the reaction at joint B on the crane structure' tram the cable must be colinear with the cable axiS, hence the cable establishes the force characteristics of direction and point of application of the reaction on the truss at point B.
A2.4 Symbols for Reacting Fitting Units as Used in Problem Solution.

~P

-~~t
t
In order to pe~it structures to move at support paints, a fitting unit involving the idea of rollers is often used. For example, the truss in the figure above is supported by a pin :'itting at (A) which is rur-tner- attached to a fitting portion that prevents any horizontal movement of truss at end (A), however, the other end (8) is supported by a nest of rollers which provide no hor1zontal resistance to a horizontal movement of the truss at end (8). The rollers fix the direction of the reaction at (B) as ,erpend~cular to the roller bed. Since t~e fitti~g uni~ is Jol~ed to the truss jOint by a pin, the ,oint of application of the reaction is also known. hence only one :orce characteristic, na~ely magnitude, 1s unknown for a roller-pin type of fitting. ?or the fitting unit at (A), ,oint of application of the reaction to the truss 1s kncvrn because of the pin, but di~ectlon and magnitUde are unknown. Lubricated Slot or Double Roller Unit.
Ty~e

In solVing a str~cture for reactions, member stresses, etc., one must know what foroe characteristics are unknown and it is cammon practice to use simple symbols to indicate. what fitting support or attachment units are to be used or are assumed to be used in the tinal design. The following sketch symbols are com~onlY used for coplanar tares systems.
Pin
ioi

A small circle at the end of a member or on a triangle represents a single pin connection and fixes the point of application of forces acting between this unit and a connecting member or structure.
b..........Kni.fe Edge

Rx

t
tRy

of Fitting

",4",
Lubricated Slot ! ~

b'RI=Jn-'

Double RoUer

~ ',~

The above graphical symbols represent a reaction in which translation of the attachment point (b) is prevented but rotation of the attached structure about (b) can take place. Thus the reaction is lxnknown in direction and magnitude but the point of applIcation is known, namely through point (b). Instead of using direction as an unknown, it is more convenient to replace the resultant reaction by two components at right angles to each other as indicated in the sketches. /1 .

~,

A2.4

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

l~eEdge
!
I

Rollers

~ FA
I

Roller.

minate, and the degree of redundancy dapends on the number of ~~nowns beyond that ~umber which can be found by the equations o~ static eqUilibrium. A structure can be statically indeterminate with respect to exter~al reac~ions alone or to In:ernal stresses alene or to 60th. The additional eq~ations :hat are needed to solve a statically indeternlnate structure are obtained by conSidering the distortion of the structure. This means that the size of all members, the ~~terial ~rom which members are ~de must be known since dist~rtions ~ust be calculated. In a statically cete~lnate structure this information on sizes and ~terial 1s not required but only the configuration of the structure as a whole. Thus deSign analYSiS for statically deter,ninate structure is stra~ght forward whereas a general t~ial and error procedure is required for design analYSiS of statically indete~inate structures.
A2. 6 Examples of Statically Determinate and Statically Indeterminate Structures.

The above fitting units using rollers fix the direction of the reaction as normal to the roller bed since the fitting unit cannot resist a horizontal force through point (b). Hence the direction and point of application of t~e reaction are established and only magnitude is unknown.
fixed

rrrrrt

(bJM.,
'" R X - I I I \ \ \ \

Illy

The graphical symbol above is used to represent a rigid support which Is attached rigidly to a connecting structure. The reaction is campletely unknown since all 3 force characteristics are unknown, namely, magnt 'tude , direction and point of application. It Is convenient to replace the reaction R by two force components referred to some paint (b) plus the unknown moment M which the resultant reaction R caused about point (b) as indicated in the above sketch. This discussion applies to a coplanar structure with all forces in the same plane. For a space structure the reaction would have 3 further unsnowns , name Ly, Rz, I1x
and lly.
A2. 5 Statically DetermLnate and Statically Indeterminate Structures.

The first step in analyzl~g a stTuc~ure 1s to determine whether the str~cture as presented is statically det.ermtnat e . It so, the reactions and internal stresses can Qe found without knowing sizes of members or kind of mater1al. If not statically determinate, the elastic :heory ~ust be applied to obtain additional equa~ion5. The elastic theory is treated in considerable detall in Chapters A7 to Al2 inclusive. To help the student become f~illar with the problem of determining whether a structure is statically determinate, several ex~ple problems will be presented.
~ample

Problem

1.
w '" 10 lb. lin.

A statically determinate structure is one in which all external reactions and internal stresses tor a given load system can be found by use at the equations of static eqUilibrium and a statically indeter.nlnate structure is one in which all reactions and internal stresses cannot be round by using only the equations of equilibrium.. A statically dete~inate structure is one that has just enough external reactions, or just enough internal members to make the structure stable under a load system and if one reaction or member is removed, the structure 1s reduced to a linkage or a mechanism and is therefore not further capable or reSisting the load system. If the st~ucture has more external reactions or internal members than is necessary for stability ot the structure under a given load system it is statically ir.deter-

Fig. A2.1

In the st~~cture shown in Fig. 2.1, the forces or loads are the distri8uted loads of 10 lb. per inch on member ABD. The reactions at points A and C are unknown. The reaction at C has only one ~~own Characteristic, namely, magnitude because the point of application of He is tnrough the cin center a: C and the directicn of Rc must be parallel to line SB because there is a pi~ at the other end B of ~ember CE. At point A the reaction 1s unknown in 11rection and ~gnitude but the point of application ~ust be through the pin center at A. Thus there are 2 unknowns at A and one ~~o~vn at C or a total
~~Qwn

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.5

At reaction points A, E and D, the react.on is known in direction and pOint of application but the magnitude is ~~own as indicated by the vector at each support. At point C, the reaction is unknown in direction because 2 struts enter jOint C. Magnitude is also unknown but paint of application is known since the reaction magnf tudes , must pass through C. Thus we have 5 unknowns, E:x:ample Problem 2. namely, Re, Ra, RO, Vc and He. For a copjanar force system we have 3 equilibrium equations P available and thus the first conclusion might -Pin be that we have a statically indeterminate structure to (5-3) = 2 degrees redundant. However, observation of the structure shows two internal pins at points E and F which means that the bending moment at these two points is zero, thus giVing us 2 more equations to use with the 3 equations ot equilibrium. Thus Fig. A.2.3 drawing tree bodies of the structure to lett at Fig. A2. 2 pin E and to right or pin F and equating moments about each pin to zero we obtain 2 equations ~lg. 2.2 shows a structural frame carrying which do not includ& unknowns other than the 5 a known load system P. Due to the pins at redction points A and B the paint or application l~OwnS listed above. The structure is thereis known tor each reaction, however, the magnitore statically determinate. tude and direction of each is unknown making a total of 4 unknowns With only 3 equations of Example Problem 4. equilibrium available for a coplanar force system. At first we might conclude that the structure is statically indeterminate but we must realize this structure has an internal pin at C which means the bending moment at this point 1s zero since the pin has no resistance to -ceat ton. If the entire structure is in .. c 0 equilibrium, then each part must likewise be Vc in equilibrium and we can cut out any portion as a f~ee bOdy and apply the equilibrium Fig. A2.6 Fig. A2.5 equations. Fig. 2.3 shows a free bOdY of the frame to left of pin at C. Taking moments Fig. 2.5 shows a beam AB which carries a about C and equating to zero gives us a fourth super-structure CED which in turn is SUbjected to the known loads P and Q. The question is equation to use in deter.nining the 4 unknowns, HA, VA' VB and Hg. The moment equation about C whether the structure is statically deter.nir~te. does ~ot include the unknowns Ve and He since The external unknown reactions for the entire they have no ~oment about C because of zero structure are at points A and B. At A due to ams. As in exa~ple problem 1, the reactions the roller type of action, magnitude is the only at A and B have been r-ep Iac ed by H and V comunknown characteristic ot the reaction since ponents ins:ead of ~sing an angle (direction) direction and point ot application are known. At B, ~gnitude and direction are unknown but as an unknown characteristic. The struct~e is statica:ly dete~inate. paint of application 1s known, hence we have 3 unknowns, namely, RA, VB and Hg, and with 3 Exa'Ilple Problem 3. equations ot equilibrium available we can find these reactions and therefore the structure is statically dete~inate with respect to external p p p P reactions. we now investigate to see it the t E I internal stresses can be found by statics arter haVing found the external reactions. Obviously, -, the internal stresses will be affected by the internal reactions at C and 0, so we draw a free ,He fRB .J.VC body of the super-st~cture as illustrated in ?ig. 2.6 and consider the internal forces that Fig. A2.4 eXisted at e and 0 as exterr~l reactions. In the actual structure the members are rigidly ~ig. 2.4 shows a straight ~ember 1-2 carrying a attached together at point C such as a Nelded or known load system ? and supported by 5 struts

of 3. with 3 equations of equilibrium available for a co?lanar force system the structure 1s statically determinate. Instead of uSing an angle as an '~nown at A to find the direction of the reaction, it is usually more convenient to re~lace the reaction by components at right angles to each other as HA and VA in the figure and thus the 3 unknowns for the structure are 3

attached to reaction paints ABeD.

):0

1;A

""llC'fMc "'-Rn

(1)~

~In \~S::

.~

J 1 'L .-

' _

'\{4x5(

A2.6
~ultiple

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

bolt connection. This ~eans tha: all three force or ~eaction charact9ristics, na~ely, magnitude, direction and point of a?pllcation are unknown, or in o~her words, 3 ~owns exist at C. For convenience we will represent these unknowns by three components as sho~m in Fig. 2.6, namely, He, Vc and Me. At jo Int D in Fig. 2.6, the only unknown regarding the reaction is RD a magnitude, since the pin at each end of the nember DE establishes the direction and point of applIcation at the reaction RD. Hence we have 4 unknowns and only 3 equations of equilibrium for the structure in Fig. 2.6, thus the structure is statically indeterninate with respect to all of the internal stresses. The student should observe that internal stresses between paints AC, 80 and FE are statically determinate, and thus the statically indeterminate portion is the structural triangle CEDC. Example Problem 5

Example Problem 5

Fig. A2. 11

Fig. A2. 12

Fig. 2.10 shows a 2 ~a7 ~~~ss supported at points A and B and carrying a known load syst&~ P, Q. All members of the truss are conr.ectec at their ends oy a common ~in at each jCi~t. The reactions at A and a are applied trIough fittings as indicated. The question is wt-ether the structure is statically dete~inate. Relative to external react~or.s at A and B the structure is statically det8~inate ~eca~se the type of support produces only one unkncwn at A and two unknowns at B, name Iy , VA, 'IS and HS as shown in Fig. 2.10 and we have 3 equations of static equilibrium available.
~e now investigate to see ~~ we can ~ind the internal ~ember stresses after :iaving :ow~d the values of the reactions at A and 8. Suppose we cut out joint B as indicated 8Y section 1-1 in Fig. 2.10 and draw a free body as sho'#TI in Fig. 2.11. Since the ~embers of the truss ~ave pins at each end, the loads in theSe ~enbers ~ust be axial, thus direc~ion and 11~e of action is known and only magnitUde is '~~~~own. In rig. 2.11 HE and ~~ are known but AS, CE, ~~d DB are ~~own in magnitUde hence we have 3 unknowns but only 2 equations of equilibrium for a coplanar concurrent force syste~. If we cut through the truss in Fig. 2.10 by the section 2-2 and draw a free body of the lcwer'portion as shown in Fig. 2.12, we ~ave 4 unknowns, namely, the axial loads in CA, DA, C8, DB but only 3 equations of equilibrl~ available for a coplanar ferce sys~em.

Figs. 2.7, 2.8 and 2.9 show the same structure carrying the same known load system P but with different support conditions at points A and B. The question is whether each structure is statically indeterminate and if so, to what degree, that is, what number of unknowns beyond the equations of statics available. Since we have a coplar~r force system, only 3 equations at statics are available for equilibrium of the structure as a whole. In the structure in Fig. 2.7, the ~eaction at A and also at B is unknown in ~gnitude and direction but point of application is ~~own, hence 4 urJcnowns and with only 3 equations of statics available, makes the structure statically indeterminate to the first degree. In Fig. 2.8, the reaction at A is a rigid one, thus all 3 characteristics of magnitude, direction and point of application of the reaction are unknown. At point B, due to pin only 2 unknowns, na~ely, ~gnitude and direction, thus making a total of 5 unknowns with only 3 equations of statics available or the structure is statically indeterminate to the second degree. In the structure of Fig. 2.9, both supports at A and B are rigid thus all 3 force characteristics are unknown at each support or a total of 6 unknowns which makes the structure statically indete~inate to the third degree.

Suppose we were able to find the stresses in CA, DA, CB, DB in some ~nr.er, and we would now ?Toceed to joint D and treat it as a tree body or cut through the upper panel along section 4-4 and use the lower portien as a free body. The same reasoning as used above would show us we have one mor-e unknown than the numberof equilibrium equations available and thus we have the truss statically lncete~1~t8 to the second deg~ee relative to internal member stresses. Physically, the structure has t~o ~ore members than is necessary for the stability of the structure under load, as we cou:d leave out one diagc~l member in each tr~ss panel and

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.7

the strJcture Nould be still stable and all ~ine the axial loads in the members and the reaxial stresses could be found by the actions on the spar. equations of static equilibrium without regard to their size of cross-section or the kind at Solution: The first thing to decide is whether material. Adding the second diagonal member the structure is statically determinate. From in each panel would necessitate knOWing the the figure it is observed that the wing spar is size of all truss members and the kind of supported by five struts. Due to the pins at material used before member stresSes could be each end of all struts, we r~ve five unknowns, found, as the additional equations needed must ~~elYJ the magnitude of the load in each strut. come from a consideration involving distortion DirectIon and location at each strut load is of the truss. Assume for exa~ple, that one known because of the pin at each end at the diagonal in the upper panel was lett out. We struts. We have 3 equations of equilibrium for would then be able to find the stresses in the the wing spar as a single unit support'ed by the ~embers of the_upper ?anel by statics but the 5 struts, thus two ~ore equations are necessary lower panel Would still be statically indeit the 5 unknown strut loads are to be found. terminate to 1 degree because or the double It is noticed that the wing spar includes 2 indiagonal system and thus one additional equation ternal single pin connections at points a and 0'. is necessary and would involve a consideration This establishes the fact that the moment of all of truss distortion. (The solution ot staticforces located to one side of the pin must be ally indeterminate trusses is covered in equal to zero since the single pin fitting canChapter A.B.) not resist a moment. Thus we obtain two additional equations because of the ~NO internal pin A2.7 Example Problem Solutions of Statically Determinate fittings and thus we have 5 equations to :ind 5 Coplanar Structures and Coplanar Loadings. unknowns. Although a student has taken a course in Fig. 2.15 shows a tree body of the wing sta:ics ~efore taking a beginning course in spar to the right of hinge fitting at O. aircraft structures, it is felt that a limited ~eview of )roblems involving the application 1013",(30+ 15)45 of the equations of static equilibrium is quite I 20~ 2 j~stlfied, part~cularly 1: the prOblems are posSibly somewhat more difficult than ~ost or ~he problems in the usual begir~ing course in stat~cs. Since one ~ust use the equations or Fig. A2.15 static equilibrium as ?art ot the necessary equations in solving statically indeterminate st~uctures and since statically indeterminate In order to take moments, the distributed structures are covered in rather complete detail load on the spar bas been replaced by the reLn other c napt ers at :::115 book, only limited sultant load on each spar portion, namely, the s?ace will be given to ~roblems involVing total load on the portion acting through the statics in this chapter. centroid of the distributed load system. The strut .r-eact.t on EA at A tas been shown in phantom ~a~ple Problem 8. as it is ~ore convenient to deal with its components YA and XA' The reaction at 0 15 unFig. A2.14 shows a much SI~plitied wing known tn ~gnltude and direction and for constructure, cor~lsting of a wing spar supported venience we will deal with its components xo 'oy lift and cabane struts wmch tie the wing and YO. The sense assumed is indicated on the spar to the fuselage structure. The distributed figure. air load on the wing spar is unsymmetrical about the center line of the air!r~e. The wing spar The sense of a force Is represented is ~ade in t~ree units, readily disassembled by graphically by an arrow head on ~he end of a vector. The correct sense is obtained from the using ern fittings at points 0 and 0 1 All solution ot the equations of equilibrium since, su~porting wing struts have Single pin fitting a force or ~oment must be given a plus or minus ~l~ts a: eac~ end. The problem is to deterSign in writing the equa~ions. Since the sense of a force or moment is unknown, 1t is assumed, 201ll/in. 401ll/in. l5#/in. if the algebraic solution of the equilibrium and " ' . ~Oi/in. ..I...,....,--r;i t j " Ill' t f ! ,--, equations gives a ~lus value to the magnitude -r i- 45"~ 82" ---iZO'!_60" ---! Hinge then the true sense is as ass~ed, and oppOSite o' , ';0 to that assumed it the solution gives a minus A PIn A 8' B 0" sign. If the unknown forces are axial loads in Lilt Caban , -+Struts truts C~ c 36" ~embers it is camoan practice to call tensile stress plus and compressive stress minus, thus Fuselage -' "30":.......i..Fig. A2.l4 E' E if we ass~.e the sense of an unknown axial ~oad 'i 8ym. 'i ~s tenSion. the solution of the equilibr!um
~ember
i

A2.8

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

equations will give a plus value for the magnitude of the unknown if the true stress is tension and a ~inus sign will indicate the assumed tension stresses should be reversed or compression, thus giving a consistency of signs. To find the unknown YA we take ~cments about point 0 and equate to zero for equilibrium
- 2460 x 41 - 1013 x 102 82YA

To find strut load g'C I point C.

~ake m~ents

about

L~ ~ 1325 x 65 + 2000 x 40 + (58eO - 44JO) 30 - 1500 x 10 - 993 x 35 - 30 (3'8') 30/33.6 0

whence, BtCI = 6000 lb. with sense as

sho~TI.

=0

To find load in member BIC use equation

ZFy '::: 0 =. 1325 + 2000 + 1500 + 993 - 6000 (30/33.6) - 2720 (30/33.6) - B'e (30/54) =0

Hence YA = 204000/82 = 2480 lb.. The plus sign means that the sense as assumed in the figure 1s correct. By geometry XA = 2480 x 117/66 = 4400 lb. and the load in strut ~ equals v4400 3 + 2480 3 = 5050 lb. tension or as assumed in the :lgure.
ZFX

whence, BIC 3535 lb. The ~inus sign means it acts opposite to that sho'Nn in fi~~e or is compression instead of tension. The reactions on the spar can now be deter.nined and shears, ber.ding moments and axial loads on the spar could be round. The numerical results should ~e checked for eqUilibrium or the spar as a whole by taking moments ot all forces about a dif:erent ~oment center to see i t the result "is zero. Example Problem 9.

=-

= 0 = Xc

To find Xo we use the equilibrium equation - 4400 = 0, whence Xc = 4400 lb. To find YO we use,

ZFy = 0 = 2460 + 1013 - 2480 - YO YO = 993 lb.

= 0,

whence

To check our results for eqUilibrium we will take moments at all torces about A to see it they equal zero.
l:11A

= 2460

x 41 - 1013 x 20 - 993 x 82 = 0 check

-r-r-'

On the spar portion OIA' J the reactions are obViOUSly equal to 40/30 times those :ound for portion OA since the external loading is 40 as compared to 30. Hence AIEl. 6750, XOI = 5880, YO' = 1325 Fig. 2.16 shows a tree body ot the center spar portion with'the reactions at 0 and at as found preViously. The unknown loads in the struts have been assumed tens10n as shown by the arrows.

I
24lJ~ \, E'
12

i\~

strut

~ Brace Strut

12

12"
I

---+--

F--

Fig.A2.17
1500",50x30
A

Shock
Strut
300

+
-L
12"

12"

'':In,"
,
"

15

" 30
Flg.A2.16

C"9----~
I I

L __

To find the load in strut Be take moments about s'

!MB' = 1325 x 20 - 2000 x 5 - 1500 x 55

- 993 x 80 60 (BC) 30/33.6 = a

Fig. 2.17 shows a Simplified airplane landing gear unit with all members and loads confined to one plane. The brace struts ~e pinned at each end and the support at C is of the roller type, thus no vertical reaction can be produced by the support fitting at point C. The member at C can rotate on the roller but horizontal movement is ~revented. A knovm load ot 10,000 lb. 1s applied to axle ~nit at A. The problem is to f1nd the load in the brace struts and the reaction at C. Solution: Due to the Single pin fitting of the brace struts, the ~eact1ons
~t ~t

Whence, Be 2720 lb. with sense as assumed.

each end a and D

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.9

are collnear with the strut axis, thus direction the aluminum alloy tubular truss. Trussed type and point of application are known for reaction beams composed of closed and open type sections RB and RD leaving only the magnitude of each as are also frequently used in Wing beam construcunknown. The roller type fitting at C fixes t1on. The stresses or loads in the members ot a truss are commonly referred to as nprimaryW the direction and point of application of the and wsecondarJn stresses. The stresses which reaction Re, leaving magnitude as the only are found under the following assumptions are unknown, Thus there are 3 unknowns Re, He and RO and with 3 equations of static equilibrium referred to as primary stresses. available, the structure is statically deter.ninate with respect to external reactions. The (1) The members of the truss are straight, sense of each of the 3 unknown reactions has weightless and lie in one plane. been assumed as indicated by the vector. (2) The members of a truss meeting at a' point are considered as jOined together by a To find Rn take moments about point B:common frictionless pin and all member axes inZME = - 10000 sin 30 x 36 - 10000 cos 300 x 12 tersect at the pin center. - RO (12/17) 2~ = 0 (3) All external loads are applied to the whence, RO ~ - 16750 lb. Since the result comes out with a minus Sign, the reaction RO has a sense OPPosite to that shown by the vector in Fig. 2.17. Since the reaction RO is colinear with the line DE because of the pin endS, the load in the brace strut DE is 16750 lb. compression. In the above moment equation about B, the reaction RO was resolved into vertical and rorizontal components at point D, and thus only the vertical component which equals (12/17) RD enters into the equation since the horizontal component has a line of action through point B and therefore no moment. He does not enter in equation as it has zerO moment about B.
To find RE take ZFv = 0

truss only at the jOints and in the plane ot the truss. Thus all loads or stresses produced in members are either axial tension or compression without bending or torsion. Those trusses produced in the truss members due to the non-fulfillment or the above assumptiOns are referred to as secondary stresses. Most steel tubular trusses are welded together at their ends and in other truss types, the ~embers are riveted or bolted together. This restraint at the joints may cause secondary stresses in some members greater than the primary stresses. Likewise it is common in actual practical deSign to apply torces to the truss members between their ends by supporting many equipment installations on these truss members. However, regardless of the magnitude at these so-called secondary loadS, it is cammon practice to first find the priwary stresses under the assumption outlined above.
GENERAL CRITERIA FOR DETERl'lINING WHEI'JlER TRUSS STRUCTURES ARE STATICALLY DEI'ERMINATE WITlf RESPECT TO INl'ERNAL STRFBSES.

ZFv

(24,/26.8) = 0

= 10000

x cos 30

+ (- 16750)(12/17) +

HE

Whence, HB = 3540 lb. Since Sign comes out plus, the sense is the same as assumed in the figure. The strut load BF is therefore 3540 lb. tension, since reaction HB is collnear wi th line SF.
To find
ZH

He

take ZH : 0
+ (-

= 10000 sin 300 - 3540 (12/26.8) (12/17) + He 0

16750)

whence, Rc = 8407 lb. Result is plus and therefore assumed sense was correct.
~aments

The simplest truss that can be constructed is the triangle which has three members m and three joints j. A more elaborate truss consists of additional triangular frames, so arranged that each triangle adds one joint and two members. Hence the number of members to insure stability under any loading is:
m = 2j - 3 -----(2.8) A truss haVing fewer members tillL~ required by Eq. (2.8) is in a state at unstable eqUilibrium and will collapse except under certain conditions of loading. The loads in the members of a truss With the number of members shown in equation (2.8) can be :ound With the available equations of statiCS, since the forces in the me~bers acting at a paint intersect at a common point or form a concurrent force system. For thiS type of force system there are two static equl11brium equations available. Thus for j number of joints there are 2j
I .

To check the numerical results take about point A for equilibrium.

ZMA = 84,07 x 36 + 354,0 (24,/26.8) 12 - 3540 (12/26.8) 36 + 16750 (12/17) 12 - 16750 (12/17) 36 = 303000 + 38100 - 57100 + 142000 - 426000 :. 0 (Check)
A2. B Stresses in Coplanar Truss Structures Under Coplanar Loading.

In aircraft construction, the t~~s type of construction 1s quite co~on. The ~cst common is the tUbular steel welded tr~sses :h~t make up the fuselage frame, and less freq~~rr:ly,

/'

,~ .

A2.10

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.
500

However three i~dependent to determine the ex~er nal reactions, thus the number of equations necessary to solve :or all the loads in the members is 2j - 3. Hence if the nw~Jer of truss ~embers is that given by equation (2.8) the truss is statically deter.ninate relative to the primary loads in the truss ~embers and the truss is also stable.
equa~ions a~e neceSsa~J

equations available.

-I
40"

L,

L,

lL"

T
-::,I(' -,
2
1

40

-l

It the truss has more members than indicated by equation (2.8) the trusS is considered redund~~t and statically lnGeterminate since the member loads cannot be found in all the members by the laws of statics. Such redundar.t structures it the members are properly ?laced are stable and will support loads of any arrangement.
ANALYTICAL IlE:rHODS fOR DE:l'SRtrINI'IG

\ ""{'--~---T-' 30 - U U \ U 3 Fig. A2. 18 c 3 f--- 30" -+30" -+- 30" -----.j i i i

1

(1

L~~ ~_:)oo

',1000

Fig. A2. 19

u,~

- ,

PRIMARY STRESSES IN TRUSS STRUCT1JRES

In general there are three rather distinct methods or procedures in applying the equations of static equilibrium to finding the primary stresses in truss type structures. They are otten referred to as the method of joints, moments, and shears.
A2. 9 Method of Joints.

It the truss as a whole is i~ equilibrium then each member or joint in the tr~ss must likewise be in equilibrium. The forces in the members at a truss jOint intersect in a common point, thus the forces on each joint form a concurrent-coplanar force system. The ~ethod of joints consists in cutting out or isolating a joint as a free bOdy and applying the laws ot equilibrium for a Concurrent force system. Since only two independent equations are available for this type of system only two unknowns can exist at any jOint. ThUS the procedure is to start at the jOint Where only two unknowns exist and continue ,rogressively throughout the truss joint by joint. To illustrate the method consider the cantilever truss of Fig. A2.18. From obse~lation there are only two members With internal stresses unknown at jOint L~. Fig. A2.19 shows a free body of jOint L~. The stresses in the members L~ La and L~ U~ have been assumed as tension, as indicated by the arrows pulling away from the jOint L~.

or tension. In equation (0) the load of 1250 in L~L:;l was substituted as a minus value since it was foune to act opposite to ~hat sho~m in Fig. A2.l9. Possibly a tetter p r-ccedur e would be to change the sense of the ar~ow i~ ~he :ree bodY diagram for any solved ~embers Jefore writing further equilibriUW equati~ns. ~e ~ust proceed to joint L:::I instead c r joint U2, as three ~~~own nembers still exist at jOint U2 Whereas only ~NO at jo~nt La. Fig. A2.20 shows f~ee body of jOi~t La cut ou~ by section 2-2 (see ~lg. AS.IS). The sense of the urJL~own member s~ress L:;lU:;l has Jeen assw~ec as COIDpressic~ (pUShing toward jOint) as ~t is obViously act~ng this way to Jalance :he 500 lb. load.
500

L.

~,

~H
250

u.,

tu.
=-

Fig. A2. 20

For equilibrium of jOint La, ZH and ZV 0 ZV 500 + LaUa = 0, whence J LaUa = 500 lb. Since the Sign came out plus, the assumed sense in Fig. A2.20 was correct or compreSSion.
ZH = 250 - LaLl :

0, wr.ence LaLl =250 IJ.

The static equations or equilibrium for the forces acting on joint L~ are ~~ and ZV
!V

Next consider joint Ua as a :ree jody c~t out by section 3-3 in Fig. A2.18 and erawn as Fig. A2.2l. The known member s~resses are showr. with their true sense as ~reviouslY found. The two ~own member stresses U 2L l and UaU l have Jeen assumed as tension.

=o.
Fig. A2. 21

=-

1000 _ L,U. (40/50)

=a

- - - - - - -fe )

whence, L ~U:;l 1250 1'0. Since the Sign came out minus the stress is opPOSite to that aSStooed in Fig. A2.l9 or compression.
Eli

=-

=- 500

- (- 1250)(30/50) - L,L,

=0

- -(b)

ZV

= -500

- 1250 (40/50)

UaLl (40/50) = 0

Whence, LaL~ = 250 lb. Since Sign comes out plus, sense is s~e as assumed in figure

whence, UaLl = 1875 is. (tens 1on as

asaumed ,.)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A2.11

ZH

(-1250) (30/50) - 1875 (30/50) - U.U.

assum~c

=0

whence, U..U a = - 1875 lb. or opposite in

sense to tr2t

ana therefore compression.

The algebraic sign of all unknowns came out positive, thus the assumed direction as shown an Fig. A2.22 was correct. Check results by taking
~B
&~B

Note: The student should continue with succeeding joints. In this example involving a cantilever truss it was not necessary to ~lnd the reactions, as it was ~oSSlble to select joint L 3 as a joint involving only VNO unknowns. In trusses such as illustrated in Fig. A2.22 i t 1s necessary to first find reactions R:l. or R a wni cn then provides a joint at the reaction point involving only two
~cwn

= 0

= 1400 x 150 + 500 x 30 - 500 x 120 - 500 30 - 1000 x 90 - 1000 x 60 = 0 (Check)

forces.

To determine the stress in member Fl , Fa and F" we cut the section 1-1 thru the truss (Fig. A2.22. Fig. A2.23 shows a free bOdy diagram of the portion of the truss to the left of this section.
, 'f~ _

Fig.A2.22
1400
A2. 10 Method of Moments.

a
o , 500 1
Fig. A2. 23 500

1400

~:
1
FIg. A2. 24

~;

For a coplanar-non-concurrent force system there are three equations of statics available. These three equations may 08 taken as moment equations about three different ~oints. Fig. AZ.22 shows a typical truss. Let it be reqUired to find the loads in the members F l , F a J
F", F. J

r , and F o '

Fig.A2.22

ioo""

3 2 'F"

100M 1 500#

10"
HA----"'(----+-\-"''-+-''--.''-'-'-----''!---''>I

50~f

The first step in the solution is to find the reactions at pOints A and B. Due to the roller type of support at B the only unknown element of the reaction force at B 1s ~gnitude. At paint A, magnitude and direction of the reaction are l~own giving a total of three ~~cwns With three equations of statics available. For convenience the wllic~own reaction at A has been re~laced by its unknown H ~~d V components.
Takir~ L~A

The truss as a whole was in equl1ibrium therefore any portion must be in eqUilibrium. In Fig. A2.23 the internal stresses in the members F l , F~ and F~ which existed in the truss as a whole ~ow are considered external forces in holding the portion of the truss to the left of sectien 1-1 in e~Jil1brium in combination with the other loads and reac~ions. Since the ~em bers a and b in Fig. A2.23 have not been cut the loads in these ~embers remain as internal stresses and have no i~fluence on the equilibrium of the portion ot the truss shown. Thus the ~ortion at the truss to lett ot section 1-1 could be considered as a solid block as shown i~ Fig. A2.24 without affecting the values at Fl , F. and F". The ~ethod of ~oments as the name unpIles involves the operation of taking moments about a point to find the load in a particular member. Since there are three unknowns a moment center must be selected such that the moment of each of the two unknown stresses will have zero moment about the selected moment center, thus leaving only one illL~Own ~orce or stress to enter into the equation tor moments. For exa~ple to cetermine load F 3 i~ Fig. A2.24 we take moments about the intersection of forces F l and Fa or paint O.

momer..ts about

~oint

A,
T

Thus
500 x
II

Z~O
"3
...

= 1400

x 30 - 18.97 F,

=0

= 500 x 30 + 100e x 60 + 1000 x 90 30 + 500 x 120 - 150 VB = a Her-ce Vg = 1600 lb. Take ZV = 0
:= '/.

il.ence

= 42000 ")21::: ( or 18.97 = ~ '-' ,,, ~..;. c ompr-eas t cn acting as assumed)

z.v
ZH

-, - 1000-1000-500-500 fore VA. = 1400 lb.

+ 1600

= a there-

momen~

Take ZH = 0

= 500

- nA = 0, therefore HA = 500 lb.

To ~ind the a~ of the force F 3 tram :he center a involves a s~~ll a~o~~t of calClllat!on. thus in general i~ 13 Simpler to resolve the ~~nown force into H and V components at a ~oi~t In ~ts 11ne Ji action such ~hat one of these ~omDonents passes thru the ~oment center and the arm of the other cJmponent can ~sually be dete~1~ed Jy inspecticn. Thus in

A2.12

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

Fig. A2.25 the ferce F 3 is resolved into its component F 3V and F3H at point 0'. 7hen taking

Fig. A2. 2S

the other two ~Jcnowns F~ and ~s lles at i~flni ty. Thus for conditions where two of the 3 c~t members are parallel Ne have a ~ethod of solving ~or the web ~ember of the t~uss ccmr.,only referred to as the ~ethod cf shea~s, or the summation of all the forces nO~ial to the tNO parallel ~{nown chord members ~us~ equal zero. Since the parallel chord ~embers have no component in a direction no~al to their line of action, they do not enter the above equatton of equilibrium.
1000lii

!
Fig. A2. 26 50'"'

1400lii

Fig. A2. 27

500

...."::,,,-----7"- F ,
500 3

/IF,
Fig. A2. 28

moments about point 0 as before:-

11400

Referring to Fig. A2.27 whence, FeH

~

2100 lb. and therefore

(21.6/30)

F.

Viously obtained.

= 2100

= 2215

ZV lb. as pre-

= 1400

- 500 - 1000 - '.I1/v'2):

whence F. = - 141 lb. (tension or OPPOSite to that assumed in the ~igure. To find the stress in illember F? , we cut section 3-3 in Fig. A2.22 and draw ~ free bOdy diagram of the left portion in Fig. AZ.28. Since Flo. and Fs are horizontal, the member F? must carry ~he shear on the truss on this sectior. 3-3, hence the name ~ethod of shears.
ZV

The load Flo. can be found by taking moments about point Ill, the intersection at torces F.
and F. (See Fig. A2.23).

ZMm =

1400 x 60 - 500 x 30 - 500 x 30 -30Flo.=Q

whence, Flo. = 2800 lb. (Tension as assumed) To tinct rcrce F. by using a moment equation, we take moments about point (r) the intersection ot rorces F J, and F 3 (See Fig. A2.26). To eliminate solving for the perpendicular distance trom paint (r) to line at action of F., we resolve Fa into its n and V components at point 0 on its line at action as shown in
Fig. A2.26.

= 1400 -

500 - 1000 F,

=0

l~ence F? = 100 lb. (compression as assumed)

ZMr

= - 1400 x 30

+ 500

x 60

+ 60 F.V

=0

whence, FaV

= 12000/60 = 200 lb.

Therefore F. pression

= 200

=282

lb. com-

A2. 11 Method of Shears

In Fig. A2.22 to tind the stress in member F. we cut the section 2-2 giving the free bOdy for the left portion as shown in Fig. A2.Z7. The mebhod ot moments is not sufficient to solve tor member F. because the intersection ot

Note: The student should solve this example illustrating the methods of moments and shears using as a tree body the portion of the truss to the right of the cut sections instead of the lett portion as used in these illustrative examples. In order to solve for the stresses in the members of a truss most advantageously, one usually makes use of more tha.~ one cf the above three methods, as each has its advantages for certain cases or members. It 1s important to realize that each is a method of sections and in a great many cases, such as trusses with parallel chords, the stresses can ?Tactically be found mentally without writing down equations of equilibrium. The following statements in general are true for parallel chord trusses: (1) The vertical camponent of the stress ~n the panel diagonal members equals the vertical shear (algebraic sum of external forces to one s1de of the panel) on the panel, since the chord

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.13

members are horizontal and thus have zero vertical component.

(2) The truss verticals in general resist the vertical component of the diagonals plus any exter~l loads applied to the end joints ot the vertical.

since no external vertical load exists at joint E. Similarly, by the same reasoning for LH ~ 0, load in DE = O. The load in the diagonal FO equals the value on the diagonal of the panel index triangle or 167 lb. It is tenSion by observation since the shear in the panel to the right is up and the vertical component of the diagonal FD must pull down for eqUilibrium.

(3) The load in the chord members is due to the horizontal components ot the diagonal members ana in general equals the summation of these horizontal components.

ConSidering Joint F. ZH = - FG - FOR = 0, which means that the horizontal component of the load in the diagonal OF equals the load in FG, or is equal to the value of the horizontal side To illustrate the simplicity at determining in the index triangle or - 133 lb. It is negastresses in tITe members of a parallel chord tive because the horizontal component of DF truss, consider the cantilever truss of Fig. pUlls on Joint F and therefore Fa must pUSh A2.29 with supporting reactions at points A and against the joint for equilibrium. J. Considering Joint D:150
+ DO = O. But DFv = 100 (vertical side of index triangle) ", DG = - 100 LH DE + DFR - DC 0, but DE and DFa = 133 (from. index triangle)

ZV = DFv

DC
550

133

36"-+- 40"

Considering Joint G:-

Fig. A2.29
~lrst, compute the length triangles in each panel of the truss as shown by the dashed triangles in each panel. The other triangles in each ~anel are referred to as load ar index triangles and their sides are directly proportional to the length triangles.

= O. But OF 133, and GCR from index triangle in the second panel. Hence OH 433 lb. Proceeding in this manner, we obtain the stress in all the members as shown in Fig. AZ.29. All the eqUilibrium equations can be solved mentally and with the calculations being done on the slide rule, all member loads can be written directly on the truss diagram.
ZH=-GH - GF - GCR

= 300

=-

=-

The shear load in each panel is first written on the vertical side at each index triangle. ~hus, in panel EFGD, considering forces to the right of a vertical section cut thru the panel, the shear is 100 lb., which is recorded on the vertical side of the index triangle.
shea~

For the second panel from the tree end, the is 100 + 150 ~ 250 and for the third panel 100 + 150 + 150 = 400 Ib , , and in like manner 550 ~or fourth panel.

Observation at the results at Fig. AZ.29 show tr~t the loads in the truss verticals equal the values of the vertical sides or the index load triangle, and the loads in the truss diagonals equal the values of the index triangle diagonal side and in general the loads in the top and bottom horizontal trJSs members equal the summation of the values ot the horizontal sides of the index triangles. The reactions at A and J are found when the above general procedure reaches joints A and J. As a c~eck on the work the reactions should be determined treating the truss as a whole. Fig. A2.3C shows the solution for the stresses in the members of a Simply supported Pratt Tr~ss, symmetrically loaded. Since all panels have the same width and height, only one length triangle is drawn as shown. Due to symmetry, the index triangles are drawn for panels to only one side at the truss center line. First, the vertical shear in each panel is written on the vertical side of each index triangle. Due to the symmetry of the truss and

The loads in the diagonals as well as their horizontal components are directly proportional to the lengths of the diagonal and horizontal side at the length triangles. 'Thus the load in diagonal member DF = 100 (50/30) = 167 and ~or member CO = 250 (46.8/30) = 390. The horizontal compor.ent of the load in OF = 100 (40/30) 133 and :or CG = 250 (36/30) 300. These values are shown on the index triangles tor each truss panel as shown in Fig. A2.29. We start our analysiS for the loads in the ~embers 8f the tr~ss by considering joint E tirst.

USing

LV

= 0 gives EF = 0 by

obser~ffition.

.,/;c

,/

A2.14

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

loading, we know that or-e t~l~ of the exter~al loads at jOints U3 and L 3 is supported a~ reaction R:I, and 1/2 at r-eac t.t cn R a , or shear- in

~
33"

5"

30" 50

Length Triangle
50

Fig. A2. 30
50 50

312 + 127.5 = ~SS.S. T~ere~2re, lsad i~ U.~2 = - 499.2. Si:nllarl~' 3.':: jo inc La, :"'aL 3 = 312 .,. 127.5 = 499.5. At ~oint U~, ete ~o~iz:~tal co~~onents of U:l,U~ an~ U~U3 = ~99.S ~ 52.5 = 562 which must be balanced ,:,y - 562 tr; member' U aU 3 '

-499.5 U

-562
.9;> 1'.9 .s
0

lu,

j
i!R,
312
100

;.=

';~~

'1

The reac:icn R~ equals t~e val~e on the vertical side of our i~cex triang:e in the 91.j panel, or 375. This should oe c~ecked usi~g the truss as a whole and ta~lng ~c~ents ~jout
R,

312

~L~ 499.5 100 100 100 6 Panels @ 25" = 150"

p..3

jL.
100

TI:e metal covered carrt t Lever' wing 'N1t;'1 i ts better overall aerodyna~lc e~:iciency and s~f ficien~ torsion~l ri 6 ! : l t y has ~ractlcal1y replaced t~e externally braced wing excep~ for lew speed cornmerc~al or ?rivate )i:ot aircraft as illustrated by the aircra:t 1~ tigs. A2.31 a~d 32. The wing covering is usually fabric and The general procedure fram this paint is to therefore a drag tr~ss inside the wi~g is find the loads in the diagor21s, then in the necessarJ to resist loads in the drag t7USs verticalS, a~d finally in the horizontal chord direction. Figs. A2.33 and 34 shows the genmember's eral structural layout a: such wings. The two The loads in the diagonals are equal to the spars or beams are ~etal or wood. Instead of USing double wires in each d~g truss bay, a va.Iues on the hypotenuse 0 f the index triangles. Single diagonal strut capable ot t~<ing either The sense, whether tension or campression, is tension or compr-ass tve loads could be -rsec , deter.nlned by inspection by cutting ~ental The external brace s t r-ut s are stream j rne tuocs . sections thru the truss and noting the direction ot the external shear load which ~st be balanced by the vertical camponent of the diagonals. The loads in the verticals are deter.nined by the method of joints and the sequence of jOints is so selected that the stress in the vertical member is the only unknown in the equation ZV = 0 for the joint in question. Thus for jOint U3 or U 3L3 = - 50. but
U~13V
,

center panel = (100 + 50) 1/2 = 75. The vertical shear in panel U:I,UgL1L a equals 75 pl~s the external loads at Ua and La or a total of 225 and Similarly for the end panel shear = 225 + 50 + 100 = 375. With these values kTIQwn, the other two sides of the index triangles are directly proportio~41 to the sides of the length triangles tor each panel, and the results are as sho~m in Fig. AZ.30.

If a truss is lcacec unsJ~etr~cally, t~e reactions should be ceter.r.ined ~irst, atter which the index triar.~:es can be dra'HTI, sta:-ting with the end ~anels, since the ;anel shear 1s then readily calculated.
A2.12 Aircraft Wing Structure. Truss Type with Fabric or Plastic Cover

ZV = - 50 - U3L 3 =

a
Fig.A2.31 Piper Tri-Pacer

hence L1U 1 = 100.

For jOint U~, ZV = - 50 - UaL 3 v - U~L~ = 0, 75, the vertical cornpo~ent of Ua L3 rr-cm index t r tang Le. . ", U ~L~ 50 - 75 = - 125. For JOint L~, ZV = - 100 + L:I,U 1 = 0,

=-

Since the horizontal chort me~bers receive their loads at :~e joints due :0 horizontal components of ~he diagonal members of the tr~ss'l we can start a: La and add u9 these horizontal components to obtain the chord stresses. Th~S, I LoLl = 312 (from index triangle). L1L 2 = 312 I from ZH = a for joint L1 At joint U"I.' the i

Fig. A2. 32 Champion Traveler

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A2.l5

11e
Fig. A2. 33
, ,

,;

~ (0. ,
,
I
t:

Wing Tip (End) Bow Plywood Tip Fairing

~

Plywood Leading Edge Fairing --l........;'1

\.,r:=r "----.ll
~'I

~\

Aileron Aileron Spar Compression Rib

in all ~enbers of the lift anc drag tr~sses will be dete~,ine~. A Si~plified air loac:ng Nill be ass~~ed, as the ?ur~ose of this problem is to ;iv8 the st~dent practice in solving sta:ically determinate space t~uss str~ctur8S.
ASSL~D

AIR LCADING:-

Drag wtre
Anti-Drag Wire

D"gwire~ "!
Fitting \
\,i.

-/
:1

(1) A constant spanwise lift load of 45 Ib/in from hin~e to strut point and then tapering to 22.5 Ib/tn at the wing tip. (2) A [erHard uniform distributed drag load of 6 lb/in. The above airloads represent a high angle of attack condition. In this condition a forward load can be placed on t~e drag truss as illustrated in Fig. A2.36. Projecting the air

:: :~' \
:1
:

V
/ I:

,
,

11'1

Leading '1\ " Edge- ':

"

Aileron Rib Aileron Hinge Drag Strut or Compression Tube

~ Forming ~ Trailing

:-+r
I

or Plain Rib Edge

I
I,

Ir--

"

Wing Hinge / ' : Fitting ~

"

"

Butt Rib Beam or Spar

Airstream

~--,...-,-

/i
,

L ift

Fig. A2. 36
Drag

Front Beam

Drag Truss Center of Pressure

Rear Beam

Fig. A2. 34
~a~~le

Problem 10. plane ';ing Structure

~ig.

~~e~nally

Braced Mono-

litt and drag forces on the drag truss direction, the fonvard projection ~ue to the lift 1s greater than the rearNard projection due to the air drag, which dl~ference ~n our exa~ple problem has been assumed as 6 Ib/in. In a low angle of attack the load in the drag truss direction would act rearward.
SOLUTION:

A2.35 shows the structural d~mensional j~agram of an externally braced nonoplane wing. The ~~ing loads on the front and rear The wing is fabric covered between wing beams, bea~s will Je calculated as the first step in ace thus a drag truss composed of struts and the solution. ?or our flight condition, the tie rods is necessary to prov~de strength and center of pressure or t~e air~orces ~ill be assumed as shown in Fig. A2.37. r~gid1ty in the drag direction. ~he axial loads
DRAG TRUSS

=:i"':

;;

-'" 1 3

Wlng Chor-d e 72" (3)

Cr F.Bl
I I
[SRI
1

1------- 72" lO~'+- 36" ---1

e p .

~
I

R.B.
24.21" Fig. A2. 37

--,---

1.1 22.52"
T~e

3 Dimedral
N '

Airplane I

, It ""

?. -i-(5lfi. 6 ,..-Fuselage
Fig. A2.35

-!(6)

--l

running load on the front bea~ will Je 45 x 24.2/36 30.25 Ib/:n., a~d the re~alnder or 45 - 30.25 = 14.74 Ib/in gives the load on the ::-ear beam.

- ' ,1..
I

A2 16

EQUILIBRIUM OF FORCE SYSTEMS

TRUSS STRUCTURES

To solve ~or loads in a t~uss system by a method of jOints, all loads ~ust be transferred to the truss joints. The wing bear,s are sup~orted at one end by the f~sela~e and o~tooard by the two 11ft st~Jts. Thus we calc~late the reactions on each beam at the strJt a~d ~lnge
paints due to the
~~~l~g

'I"A
1/

lift load on each

:;;/.::B.,r
F

/1' x'\.k /!\ I x/1"

R. B.
\.,/
'\.1/

3'---_71

"'-I

/11

Drag Truss
3

R. B.

beam.
Fl"Qot Lift

Front Seam

Truss

R Rear / ; Lift Truss

~ember

lengthS L and :he ~ar.poner.t ra.t ; os fellow by Simple ca:culat~on.

Table .\2.1

Taking moments about

pof nt (2)

lIember Front Be" Rear Beall. Front Strut

gym.:
FB
RB

t,

I V;1. I D;1. I ';1.

0 0 1.9986 1.9986 I

114.5R .. - 114.5 x 30.26 x 114.5/2 - 15.13 x 70.5 x 149.75 - 15.13 x 35.25 x 138 = O.

I ,

5.99
~.99

o o
11

114.34 114.501.05231 114.34 114.501.05231

hence
RJ.

Take

= 3770 lb. ZV = 0 where V direction beam ZV =- 3770 + 30.26 1

R~

Bear Stru.t

1s taken normal to

V ... vertical direction,

'.
+

'F

15 7 9 9

114.34 128.791.45011.08541.8878 i 114.34128.00i.44861 0 1.89301

j 57.49

D drag direction, S
L
~

x 114.5

side direction,
D2 32

(30.26 + 15. 3 ) 70.5 2

=0

~ Vv2+

=0)

hence R~ 1295 lb. (The student should always check results by taking moments about pOint (1) to see it ZM ..

w~th

We stare the solution of joints jy sta~i~g joint (1). Free bOdy sketches cf joint (I)

Rear Beam
(W'"

14. 7#/in.

(4)

1I
R.

Ilftltll;,,/2
114. 5"

ija
R,

70. 5,,--1

The rear beam has the same span dimensions but the loading is 14.74 Ib/in. Hence beam reactions R~ and R3 will be 14.74/30.25 = .4875

are sketchel below. All members are considered two-f or-c e r.ember-s or havtng pf ns at each end, thus magnt .ude is the cn.Iy unknown characteristic of ~~ch member load. The jrag truss ~em bers coming in to jofnt (l) are replaced :,y a single reaction called Jl.' Ar t er- Dl. is r ounc , its influence in causlr.g loads in drag truss ~embers can then be found when the drag :russ as a whole is treated. In :he joint solution, the drag truss has been assumed parallel to drag direction which is nct quire true frc~ Fig. A2.35, but the error on member loads is negligible. JOINT 1 (Equations of ~quilibri~~)

tilnes those for front beam. hence R s = .4875 x 3770 = 1838 lb.
R~

= .4875 x 1295 = 631 lb.

(1)

The next step in the solution is ~he solving tor the axial loads in all the members. We will use the ~ethod of joints and consider the structure made up of trIee truss systems as illustrated at the top of the next col~. namely, a front li!t truss, a rear lift truss and a drag truss. The beams are co~on to ~oth lift and drag trusses.
Table A2.1 gives the V, 0 and S ~rojecticns or the lift truss ~embers as dete~ined from information given in Fig. A2.35. ~he true

SF

~
B

770

D,
~v

4+D

zv = 3770 x .9986 zs = - 3770 x .0523 ZD =

.0523

.;966 F&

.0854 SF + 0, = 0

- .4501 S" = 0- ( i - .36'7'8 SF = 0- ( i ----- - ----! ,

FE
I

SolVing equa t ions 1, 2 and 3, we obtain

F3 SF

0,

== =

8513 lb. (compression)

9333

798 lb. (aft i

l~

. (tensior.)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A2 17

Joint (3) (Equations of equt Lrbr-i ua)

t 1838

RL.-J(3l
~
R

(3)

V-Splane

:5

:V = 1638 x
ZD

1 83 8

(drag truss

reaction on V-Dplane

---D, ptn (3\)

points (2) and (4). In the design of the beam and fittings at this point, the effect of ~he actual conditions of eccentricity should of course be considered. Combined Loads on Drag Truss Adding the two load systems of Figs. A2.38 and A2.39, the total drag truss loading is obtained as shown in Fig. A2.40. The resulting member axial stresses are then solved for by the method of index stresses (Art. AZ.9). The values are indicated on the truss diagram. It is customary to ~ake one of the fittings attaching wing to fuselage incapable of tranSferring drag reaction to fUselage, so that the entire drag reaction from wing panel on fuselage 1s definitely ccnrtnec to one paint. In this example paint (2) has been assumed as point where drag is resisted. Those drag wires which would be in compreSSion are assumed out of action.

RB SR

= -1838
D, +

.9986 - .0523 RB - .4486 SR

x .0523 - .9986 RB -.8930 SR 0 = 0 - - - - - - - - - - - -

= 0 -(4) = 0 -(5)
- - -(6)

Solving equations 4, 5 and 6, we obtain RB = - 4189 lb. (compression) SR = 4579 lb. itension) 03 = 0 Fig. A2.38 shows the reactions of the lift struts on the drag truss at Joints (1) and (3) as found above.

,J><1Xt~X
Fig. A2.38 798

~~3G---f1-36~--i
39.5
118.5
1191 114\

231

37.5 225

58.5

281.5
4189

I
162il -11,933

254

Drag Truss Panel Point Loads Due to Air Drag Load. It was assumed that the air load components in the drag direction were 6 lb./in. of wing acting r orwar-d, The distributed load of 6 lb./in. is replaced by concentrated loads at the panel points as shown in Fig. A2.39. Each panel point takes one halt the distributed lead to the adjacent panel point, except for the two outboard panel pOints which are affected by the overhang tip portion. Thus the outboard panel paint concentration of 254 lb. is Jetermined by taking moments about (3) of the drag load outboard of (3) as follows:
P
To

-13 893

~!~::o il~"'~J ~I~ ij~8' ~ \~~~II 1 ""I~~~

i(2) 13,893 13 -10,313 8513~ -413
1908 '" Drag Reaction 798

Fig. A2.40

Fuselage Reactions As a check on the work as well as to obtain reference loads on ~lselage from Wing structure, the fuselage reactions will be checked against the exterr~llY applied air loads. Table AZ.2 gives the calculations in table form.
Table A,2.2 .Point
lIeliber

Load -13893 - 1908 - 1295 -

D
0 -1908 0 0 0

s
_13870 0

= 70.5

x 6 x 35.25/58.5

= 254

IJ.

III

Dn,
Reactio.. Jt2(React1on)

726 0 129'" 62 630 4205 2055 7520

Si~pllfY the drag truss SOlution, the dra~ strut and drag wires in the inboard drag truss panel have been modif~ed to ~ntersect at hinge

RB R.. (R.eaeUon)

~/in.

,
j

6
Totals

..
F,

1191

631
9333 4579

- 67 1190 - 33
8290 4090

"8
0 -1110

.00

Applied Air Loads:

V

component = (3770 + 1295 + 1838 + 631) .9986 = 7523 lb. (error 3 lb. D component = -185 x 6 1110 lb. (er-r-or e 0) S component = -(3770 + 1295 + 1838 + 631) .0523 = 394 lb. (error 6 lb.

A2.18

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

al~

The wing bea~ ~ue to the distribu~ed l~~t loads acting upon them, are also s~bj~cted to bending loads in add! t t on to the axf.a ; Loads . The wing be~s thus act as bea~-cJll~s. ~he sUbject of beaT.-col~lh~ action is treated in another chapter of this book. skin instead of fabriC, the drag truss can be omitted since the top ~~d batto~ skin act as webs of a beam which has the front and rear be~ as its flange ~embers. T~e wing is then considered as a box beam subjected to combined bending ~~d axial loading.
~lng ~etal Exam~le

Lengths k Directional Components of Cabane Struts Member Front Cabane Diagonal Caba,ne Strl.\t Rear Cabane Strut Sym.

V
30 30

D
10

I s

VIL

DIL
.240 .597 .1485

'IL
.648 .538 .668

CF CD CR

12 7

1 4 1 . 59 1 .721 , .597 .731

30 i27150.17
6

If the

1s covered with

23.5

27 I 40.42

L ;./ vO: +

D' + "

Problem 11.

3-8ectlon Externa:ly Braced

l,.jing. Fig. A2.41 shcws a high wing exterr4lly braced wing structure. The wing outer panel has been made identical to the wing panel of exa~ple prOblem 1. This outer p~~el attached to the cen ter panel by Sir~le pin fittings at points (2) and (4). Placing pins at these points ~ake the structure sta~lcally determinate, whereas if the beams were made continuous trtlo~gh all 3 panels, the reactions of the lift and cabane struts on the wing beams would be statically indeterminate since we would have a 3-span continuous bea~ resting on settling supports due to str~t deformation. The fitting ?in at ~oints (2) and (4) can be made eccentric with the neut~al axis of the beams, hence very little is gained by making be&~ continuous for the purpose of decreasing the lateral beam bending mo~ents. For assembly, stowage and shipping it is convenient to build such a wing in 3 portions. If a ~ultiple bolt fitting is used as ~oints (2) and (4) to obtain a continuous beam, not much is gained because the design requirements of the various governmental agencies speCify ~hat the wing beams must also be analyzed on the assumption that a multiple bolt tltting provides only 50 percent of the full continuity.
1',
(8)

The air loads on the outer ~anel are taken identical to those in exa~ple ;roblem 1. LikewiSe the dihedral and direction ~f the lift struts SF and SR have been made the sa~e as in exauple ~roblem 1. Therefore the analysis :or the loadS in the outer p~~el Gr~g and :i~t tr~s trusses is identical to that in )roblem 1. T~e solution will be cont1~ued assumi~g the r~~,i~g 11ft load on center panel of 4.5 10./:n. and a forNard drag load of 6 Ib./in.
Sol~tion

0:

Cente~ ?~nel

Center Rear Beam

692# (630-;.62)",692#

w'" 14. 7%/in.

(4) ~~~~~~~~~~~r~~(4) ~20

8'

90'

'8

20-1

R a",1650if

R a",1650.f

Fig. A2. 42

Fig. A2.42 shows the late~al loacs on the center rear beam. The loads consist of the d1str:j~teG air load and the vertical component ot t~e ~or ces exerted by outer ?anel on center panel at pin point (4). From Table A2.2 of exarr.ple ;roblem 1, this resultant V reaction eq~ls 630 +
62

692 lb.

"'.= ,"
\4)

The vertical component of the cabane reaction at joint (8) equals one hal: the tetal be~ load due to symmetry of 1cad1~g or 55 x 14.74 + 692 = 1650 lb. Solution of force system at Joint 8
1650

J...
36"

C~
CR
(1)(3)

/L:.. S

D1
C

1650

~D

ZV - 1650 - .7)1 CR = a
whence CR = 1650/.731
(5)
Fig. A2. 41

= 2250

lb.

(tens~on)

(6)

zs = - eRE - 2250 x .565

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A2.19

'whence

eRE

1510 lJ. (compress ion)

ZD = De - 2260 X .1485
whence De = 336 lb. drag Center rront Beam
568#
(Ref. , Table A2. 2)

=0
reaction

(4) are taken fram Tabl~ A2.2 of problem 1. The drag loac of 336 10. at (8) is ~~e to the rear cabane strut, as 1s likewise the bea~ ax1al load of - 1510 at (8). The axfa I beam load of - 2281 lb. at (7) 1s cue to ~eact1on of ~rc~t caba~e truss. The ,a~el ,oint loads are d~e to the given running drag lead of 6 lb./in. acti~6 forward. The reaction which holds all these drag truss loads in equilibrium is supplied by the cabane truss at pOint (7) since the ~ront and ciagonal cabane struts intersect to form a rigid triangle. ~hus the drag reaction R equals one half the total drag loads or 2634 lb. Solving the truss for the loading of Fig. A2.44 we obtain the member axial loads of Fig. A2.45.
1157

tr~ss

(1294 - 726) ::: 568*

w'" 30.26 iii/in.

fo- 20

71

90 Fig. A2.43

R.,; 2535.

rig. A2.43 shows the V loads on the center front beaw and the resulting V component of the cabane reaction at joint (7). Solution of force
syste~

568

568

:;: ,
-17308
(11) Fig. A2.45

at Joint 7
-1730B

CFB
Co

~
CF

2535

-1502

VS Plane

LoadS in Cabane struts Due to Drag Reacticn at Point 7

ZV

= 2535 - .721 CF - .597 CD : 0 zs ' - CFB - .648 CF - .538 Co , 0

.240 CF + .597 CD

263~

ZD

=0
ZD
Zv

kF

"'.

v-n Plane

CD

solVing the three equations, we obtain

eFB = - 2281 (compression)

C:<,

=
=-

2634 - .240 CF' + .597 CD

.721 CF .~97

=0

Co , 1058

= 2635
Tr~ss ~embers

Co

=0

Solving for CF and CD, we obtain CF - - 2740 lb. (compression) Co = 3310 (tenSion) adding these loads to those previously lated for 11tt loads: CF
60
calc~

solution for Loads 1n Drag

Fig. A2.44 shows all the loads applied to the center p~~el drag truss. The Sand D reactions fry,n the outer ~anel at joints (2) and
Panel point Drag Load
for 6-#/in.-")

1157

13m.~/L\J~7 t
I
190a'

f ~5~ i 1~Jl95 ~ -~I-/1\I~

6 1 15 ..

33 6 270

336

CD ,

, ,

2740 + 2635

1058 + 3310 2250 It.

, -

105

4368 lb.

eR

1157
fusela~e

Reactions

As a c~eck on the ~ork the ~~selage reactions will be checked against t:'e applied loadS. ~ajle A2.3 ~i7es the '1, D a~d S com~on2nts of t~e ~~sela~9 rec~t10~S.

n, 26341

122IT

2.rr-

(211

Fig. A2.. 44

'R, 2634#

1190a

I ! ,-

A2.20

EQUILIBRIUM OF FORCE SYSTEMS.

Table A2.3

TRUSS STRUCTURES. 36

Poiut 9

lIeaber Protlt 5tr\,\t

c=,
_105 2260 4368 9333 4579

S
_68 1510 2356

,.
e e
Total.

C.
Rear Strut CR D1a. Strut CD Front
L1~t

-"
1650 2610 4205 2055 10444

-'5 335 _2610 798

sl
72

24

Strut

8290 4090

I,

S.
Rear Lift Strut

'.

o
_1502

F
16178

Applied Air Loads

V

component

= 7523 (oute::- ~anel) = 10448 (check)

+ 65

x 45

Fig. A2.46

component =

--

1110 (outer pane L) 65 x 1500 (er::-or 2 lb.)

The total side load on a vertical plane trIU centerline of airplane should equal the S component of the applied loads. ~he applied side loads = - 39~ lb. (see problem 1). The air load on center panel is ver~lcal and thus has zero S component. From Table A2.3 for fusela~e reactions have ZS = 16178. From Fig. A2.45 the load in the front beam at b of airplane equals - 17308 and 568 for rear beam. The horizontal component at the diagonal drag strut at joints 11 equals 216 x 45/57.6 = 169 lb. Then total S components = 16178 - 17308 568 + 169 = - 393 lb. which checks the side component ot the applied air loads. Example Problem 12. Single SDaT Truss Plus Torsional Truss System. In small wings or control surfaces, fabric is often used as the surface covering. Since the fabric cannot provide reliable torsional reSistance, internal structure must be of suc~ design as to ~rovide torsicnal strength. A single spar plus a special type of truss system is often used to give a satiSfactory structure. Fig. A2.45 illustrates such a type of structure, namely, a trussed Single spar AEFrJ plus a triangular truss system bevNeen the spar and the trailing edge OS. Fig. A2.46 (a, ~, c) shows the three prOjections and dimensions. The air load on the surface covering of the structure is assumed to be 0.5 Ib./in. a i~tensity at spar line and then varying linearly to zero at the trailing edge (See Fig. d). The problem will be to dete~ine the axial loads in all the members of the structure. It will be assumed that all members are 2 force illembers as 1s usually dcne in finding the pr1ma~J loads in t~ssed str~ctures.
~

T
36"

(lVTC I I --

/ 11

S
\

\1
/
T

Fig.46a 1
I

l.s
'\ / "'

'c! \/
D D
G

II;';
*'1

": I ::!l A I~ c,
I

;g

)'"1/

~ 1/I~l2Is:vt;>,,1
N L J ~ 7 panels @ 12" '" 84" Fig.46b - _-----J

ABC ABC

E E

(Q~

,=
~

I 36" r-----.., Fig.46c

c>;

SOLUTION: The total air load on the str~cture equals the average intenSity per squa~e inch ~1mes :r.e surface area or (0.5)(.5)(36 x 94-) = 756 lb. In order to solve a truss system by a method of joints the distributed lead must be replaced by an eqUivalent load system acting a~ the Joints of the structure. Referring to Fig. (d), :he total air load on a strip 1 l~:h wide and 36 inches long is 36(0.5)/2 = 9 lb. and its 0'5. or resultant location is 12 inches from line AE. In i":g. 46e. this z-esu'l tarrt load of 9 ~b./in. is imagined as acting on an ima~inary bea~ located along the lIne 1-1. This rur~lr.g load a~~llec a:ong this line is now replaced jy an equivalent force system acting at jOlr.ts OP~}BED8CA. The results 0: :hls joint distrlbution are shown by the jOint loads in F~g. A2.46. !o illustrate how these joint loads were obtained, t~e calculatior.s for loads at ~oints SSDR will be given.
F~g. A2.48 shows a portien of :he struct~e to be c ons t der-ed. For a run....n tng load 0: 9 Ib ./"!..n., along Ltne 1-1, r-eact t cns -vt Ll. be for

,.g 'JLI!

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.21

my :: (6

66

72

72

36 )36 - 36 Oz :: 0

whence Oz = 252 lb. acting down as assumed. To find Oy take moments about z axis through point (A).

mz = 0
Fig. A2. 48
si~ple beams re~ting at ,oints 2, 3, 4, 5, etc. The distance bet~een 2-3 is 8 inches. The total load on this distance is 8 x 9 = 72 lb. One half or 36 lb. goes to paint (2) and the other half to paint (3). The 36 lb. at (2) is then replaced by an equivalent force system at E and 8 or (36)/3 12 lb. to 8 and (36)(2/3) 24 to E. The distance between points (3) and (4) is 8 inches and the load is 8 x 9 = 72 lb. One half of this or 36 goes to pOint (3) and this added to the previous 36 gives 72 lb. at (3). The load of 72 1s then replaced by an equivalent force system at Sand D, or (72)/3 = 24 lb. to Sand (72)(2/3) ~48 to D. The final load at 8 1s therefore 24 + 12 = 36 lb. as shown in Fig. A 2.46. Due to symmet~J of the triangle CRD, one half of the total load on the distance CD goes to points (4) and (5) or (24 x 9)/2 = 108 lb .. The distribution to D is therefore (108) (2/3) 72 and (108)/3 36 to R. Adding 72 to the previous load of 48 at D gives a total load at D = 120 lb. as shown in Fig. A2.46. The 108 lb. at po tnt (5) also gives (108)/3 = 36 to R or a total of 72 lb. at R. The student should check tee distribution to other joints as shown in Fig. A2.46.

36 Oy

= 0, Oy

To find Ay take moments about x axis 'thr-ough point N. The moment of the air loads was previously calculated as - 31752, hence, ZMx = - 31752 + 9 Ay 3528 lb. To find Ny take l.Fy ZFy

= 0, =a

whence, Ay =

=3528

- Ny

=0,

hence Ny

3528 lb.

To find Nz take l.Fz 0 l.Fz :: - 252 + 756 - N z :: 0, hence Nz 504 lb.

The reactions are all recorded on Fig. A2.46. Solution of Truss Member Loads For simplicity, the load system on the structure will be considered separately as two load systems. One system will include only those loads acting along the line AE and the second load system will be remaining loads which act along line as. Since no bending moment can be resisted at JOint 0, the external load along spar AE will be reacted at A and N entirely or in other words, the spare alone resists the loads on line AE.

To check the equi,mlence of the derived F1g. A2.49 shows a d1a~ of this spar joint load system wl~h ~he original air load with its joint external loading. The axial system, the magnitude and ~oments of each loads oroduced by this loading are written on system must be the same. Addi~g up the total the trUss members. (The S~Jdent should check joint loads as shown in Fig. A2.46 gives a total these ~ember loads.) of 756 lb. which checks the original air load. 144 144 24 72 The moment of the total air load about an x t ~xis at left end of structure equals 756 x 42 :: I -11841 -1184 0 2336 31752 in. lb. The ~oment of the joint load ': I "0 i~ <\"~ :.?o~~ system in Fig. A2.46 equals (66 x 12) + (72 x IN , 36) + (72 x 60) + (56 x 84) + l?4 (24 + 48) + BOO 32 BOO 224 224 (120 x 72) + (24 x 84) = 31752 in. lb. or a check. The moment of the total air load about Fig. A2.49 line AE equals 756 x 12 9072 in. lb. The TRI.ANGUUR TRUSS SYSTE:1 moment of the distributed joint loads equals (6 + 66 + 72 + 72 + 36)36 = 9072 or a check. The load system along t~e trailing edge as causes stresses in both the spar truss ~~d the Calculation of Reactions diagcr~l truss system. ~he support fitting at paint a provides a reaction in the Z direction The structure is supported by Single pin but no reacting ~oment about the x axis. Since fittings at pOi~ts A, Nand 0, with pin axes the loads on the trailing edge 11e on a y axis parallel to x axis. It will be assumed that throu~~ 0, it is obvious that all these loads the fitting at N takes off the spar load in flow to ?oint O. Since the bending strength of Z direction. Fig. A2.46 shows the reactions the tra~llng edge member ~s negligible, the Oy, 0z, Ay, Ny, Nz To find Oz take moments about y axis along spar AEFN.

~i~-"23,,,3,,6"""==-i-'-""7<?"'---;-:='-:n,,",C"'""",,--"--i

A2 22

EQUILIBRIUM OF FORCE SYSTEMS

TRUSS STRUCTURES

load of 36 lb. at Jol~t S in order to be transfer~ed to pOint 0 through the dlago~al truss system must follow the path SDRC~PAO. In like ~nner the load of 72 at R to reach a must take the path RC~~PAO, etc. Calculation of Loads in Dlaeonal
Member
, y x
Tr~ss

reaction on the s;ar force on the spar <~ duces compression in truss and tension ~n Consider Joint R

but does produce ~ couple the v dir~ctlcn ~r.lch ~ro the ~0P chord of t~e s~ar the bottom c~ord.

Members:x/L
.943

z/L
.118

y/L

The load to be transferred to truss ~CJR is equal to the 72 lb. at R ~lus the 36 1J. at S wn.tcb CCII'.es to jot nt R from tr-cs s DRG. Hence load
~n

All Diagonal Truss Members AO, NO

4.5

12 36

38.2

.314

Re =
C~

(~2 +

36)0.5 x (1/.118) and JQ

4.5

36

37.5

.120

.960

457 lb.

Tdhence RJ Consider Joint S truss SEF carL~ot assist in any portion of the 36 lb. lead at S because the reaction of this truss at SF would put torsion on the spar and the spar has no appreciable torsional resistance. The
trla~Jlar transterrlr~

= 457,

= 457

=- 457
= 72
~

Joint Q

Load to be transferred to truss 72 + 36 180 lb.

~BL

Hence load in - 762.

~nence

= (180 x G.5)(1/.118)
BP

Considering Joint S as a free body and writing the equilibrium equations;

l:Fx

QL

=762,
+ 66

= 762,

LP

=- 762
=AO

= -.943
+

DS -.943 GS

=G

X~~:G !Z

Joint P Load

= 180

=246
x O.5)(l!.l18)
~ree

.....hence, DS ::: - GS

Load in PA

= (246

lG40

ZFz = 305

.118 OS - .U8 GS = 0

Whence PN

= 104-0
:ody.
+

Subt. DS = - GS and solving for OS, gives as = 159 lb. (tension), DS = -159 (compression) Consider Joint D

Consider JoInt (A) as a

ZF x = - 1040 x .943
lG22 lb.

.960 AO

= 0,
~J

=
~O

= - 2G22 lb.

In like manner, considering Joint

gi'l8S

Couple Force Reactions on Spar

L.
Let Ty and Tz be reactions ot diagonal truss system on spar truss at Joint D. ZFx

=- 159

x .943 + .943 OR = OJ hence DR = 159 lb.

+

As pointed out ?reviously, t~e diagonal torsion truss produces a couple reaction on the spar in the y direction. The ~gnitude of the torce of this couple equals the y component ot the load in the diagonal truss ~embers ~eeting at a spar jOint. Let Ty equal this reaction load on the spar.
At Joint c:Ty

ZFz = - 159 x .118

159 x .118 - Tz = a

==-

(457 + 457) .314': - 28'7 1::'.

whence Tz 0, which ~eans the diagonal truss produces no Z reaction or shear load on spar truss at D.

Likewise at Joint J, Ty At Joint s:Ty (762

+

= 287
=479

ZFy

=- .314 x
whence Ty

159 - .314 x 159 - Ty

=0

762) .314

=-

100 lb.

, T Likewise at Joint c, y

=479
=326

If joint G is investigated in the same manner, the results will show that Tz a and Ty 100.

At Joint A:m

'y

=-

(1040 x .314)

diagor~l

The results at joint D shows that the rear truss system produces no shear lead

Likewise at Joint N, Ty = 326

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.23

These reactions of the torsion truss upon the spar truss are shown in Fig. A2.50. The loads in the spar truss illembers due to this loading are written adjacent ~o each truss ~ember. Adding these illember loads to the loads in Fig. A2.49, we obtain the final spar truss ~ember loads as shown in Fig. A2.5l.

Level Landing with Inclined Reactions

Level Landing with Nose Wheel Just Clear of Ground

Fig.A2.51

If we add the reactions in Figs. A2.50 and AZ.51 J we obtain 3528 and 504 which check the reactions obtained in Fig. A2.46.
A2. 13 Landing Gear structure

Tail Down Landing Fig. A2.52

The airplane 1s both a landborne and airborne vent c l,e, and thus a means ot operating the airplane on the ground must be prOVided which means wheels and brakas. Furtharmore, provision must be made to control the impact forces involved in landing or in taxiing over rough gro~~d. This reqUirement requires a speCial energy absorption unit in the landing gear beyond that energy absor?tion prOVided by the tires. The landing gear thus includes a so-called shock s~rut cammonly referred to as ~~ oleo strut J which is a member composed ot ~HO telescoping cylinders. 'dhen the strut is compressed J oil inside the air tight cylinders Is forced through an orifice from one cylinder to the other and the energy due to the landing impact Is absorbed by ~he work done in forCing this oil through the critice. The orifice can be so designed as ~o provide practically a uniform resistance over the displacement or travel of the olec strut. An airplane can land safely with the airplane in various attitudes at the instant at ground contact. Fig. A2.52 illustrates ~he three altitudeS of the airplane that are specified by the government aviation agencies for deSign of landing gear. In addition to these symmetrical unbraked loadings, special loadingsJ such as a braked condition, landing on one wheel condition, side load on wheel, etc. are reqUired. In other words, a landir~ gear can be subjected to a cor~iderable number of different loadings ~~der the variouS lar.ding conditions that are encountered in the normal operation of an airplane.

Fig. AZ.53 shows photographs at typical main gear units and Fig. A2.54 tor nose Wheel gear units. The successful deSign of landing gear tor present day aircraft is no doubt one or the most difficult prOblems which is encountered In the structural layout and strength design of aircraft. In general, the gear tor aerodynamic efficiency must be retracted into the interior of the wingJ nacelle or fuselage J thus a reliable J safe retracting and lowering mechanism system is necessary. The wheels must be braked and the nose wheel made steerable. The landing gear is SUbjected to relatively large 10ads J whose magnitudes are several times the gross weight ot the airplane and these large loadS must be carried into the supporting wing or tuselage structure. Since the weIght of landing gear may amount to around 6 percent ot the weight at the airplane it Is evident that high strength/weight ratio is a paramount design requirement ot landing gear, as inefticlen~ structural arrangement and conservative stress ana.Lyst a can add many unnecessary pounds of weIght to the airplane and thus decrease the payor useful load.
A2. 14 Example Problems of Calculating Reactions and Loads on Members oC Landing Gear Units

In its Simplest fonn, a landing gear could consist of a single oleo strut actlr~ as a cantilever beam wIth its fIxed end being the upper end which would be rigidly fastened to the supporting structure. The lower cylinder of the oleo strut carrieS an axle at its lower

A2.24

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

McDonnell Aircraft (Military Airplane)

Douglas DC-7 Air Transport

Douglas DC-8

Jet Airliner

FIg. A2. S3 Main Landing Gear Illustrations (One Side)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.25

Beechcraft Twin Bonanza

Piper-Apache

Piper Tri-Pacer

Navy F4-J

North American Aviation Co. Douglas DC 7 Air Transport

Fig. A2. 54 Nose Wheel Gear Installations

A2.26

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

end tor attaching the wheel and t~re. ~his cantilever beam is subjected to bending in two directions, torsion and also axial loads. Since the gear is usually made retractable, it is difficult to design a single fitting unit at the upper end of the oleo strut that will resist this combination of forces and still per.nlt movement tor a simple retracting mechanism. Furthermore, it would be difficult to provide carry-through supporting wing or fuselage structure tor such large concentrated load systems. Thus to decrease the magnitude of the bending moments and also the bending fleXibility ot the cantilever strut and also to simplify the retracting problem and the carry-through structural problem, it is customary to add one or two braces to the oleo strut. In general, effort is made to make the landing gear structure statically determinate by USing speCially designed fittings at member ends or at support paints in order to establish the torce characteristics of direction and paint at application. TWo example problem solutions will be presented, one dealing with a gear with a Single wheel and the other with a gear involving two wheels. Example Problem 13 Fig. A2.55 shows the projections of the landing gear configuration on the VS and VD planes. Fig. A2.56 is a space dimensional diagram. In landing gear analysis it is cammon to use V, D and S as reference axes instead of the symbols Z, X and Y. This gear l~it 1s assumed as representing one side of the main gear on a tricycle type of landing gear system. The loading assumed corresponds to a condition of nose wheel up or tail down. (See lower sketch or Fig. A2 .52) The design load on the wheel 1s vertical and its magnitUde for this problem is 15000 lb. The gear unit is attached to the supporting structure at points F, H and G. Retraction of the gear is obtained by rotating gear rearward and upward about axis 'thr-ough F' and H. The fittings at F and H are deSigned to resist ~o bending moment hence reactions at F and Hare UIL~own in magnitude and direction. Instead of USing the reaction and an angle as unknowns, the resultant reaction is replaced by its Y and D components as shown in Fig. A2.56. The reaction at G is unknown in magnitude only Since the pin fitting at each end of member GC fixes the direction and line of action of the reaction at G. For convenience in calculations, the reaction G Is replaced by its camponents GV and GD' For a stde load on the landing gear, the reaction In the S direction is taken off at paint F by a special designed unit.

SOUJ'I'I:JN

The supporting ~eactlons upon ~he ~ea~ at paints F, H, and G will oe calculated as a beg~nn1ng step. T~ere are six urJcnowns, na~ely FS, F'V, FO, Hy, HO and G (See Fig. AZ.56). Wlt~ 6 equations of static equilibrium available tor a space ~orce sy~tem, the reactions can be found by statics. Referring to Fig. AZ.56:To find FS take ZS Fs
+

=0

= 0,

hence FS

=a

To find react10n Gv take moments about an S axis through paints F, H. 50 - 24 Gv 0 Whence, GV = 6500 lb. with sense as assumed. (The wheel load of 15000 lb. has been resolved into V and D components as indicated in Fig. A2.55) . With GV known, the reaction G equals (6500) (31.8/24) 8610 lb. and similarly ~he component GD (6500)(21/24) 5690 lb.

l:l1s

=3119 x

1 , 1-.6

=
"

;0

.
II

cJi:~~

J6"

<;

10063

1~3004 !--5690 G
lll09

6500

\'D
Fig. A2. 56

I
a

433

Flg. A2. 57

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A2.1.7

lEy
-r-E I

, , I" u
I
I
I

> ED~'b ...

,,>'

21

24"

1 <>

~ _ C--CGO=31 a e G
"I"'I

31

the structure as a whole by taking moments about D and V axes through point A. = - 10063 x 14 + 6500 x 8 + 11109 x 8 =- 140882 + 52000 + 88882 O(check) = 5690 x 8 - 433 x 8 - 3004 x 14 = 45520 - 3464 - 42056 = 0 (cheek)

26"

L.
0

lel-3u9*

TE=24952

--12"'.JI FS-S :E~s=3920LI~~ "

Fv

By

ro"r
TE

14"--l...f EV=13332 BIV=7840

:;:@-BIS=3920

FS--S 1@)_3920 F.

to

39,20. HO

----~
H

i . E~.2571 I ~. +- 16" ---t

Gv

The next step in the solution will be the calculation of the forces on the oleo strut unit. Fig. A2.58 shows a free body at the oleostrut-axle unit. The brace members EI and CG are two force members due to the pin at each end, and thus magnitude is the only unknown reaction characteristic at points Band C. The fitting at point E between the oleo strut and the top cross member FH is designed in such a manner as to resist torsional moments about the oleo strut axis and to provide 0, V and S force reactions but no moment reactions about 0 and S axes. The unknowns are therefore BI, CG, Es, EV, ED and TE or a total of 6 and therefore statically determinate. The torsional moment TE 1s represented in Fig. A2.S8 by a vector with a double arrow. The vector direction represents the moment axis and the sense of rotation of the moment is given by the right hand rule, namely, with the thumb of the right hand painting in the same direction as the arrows, the curled fingers give the sense of rotation.
T~ fInd the resisting torsional moment take moments about V axis through E. T~

Fig.A2.59

.....

To find Fv, take moments about a D axis

through point H.
l:l'lH(D)
:~ence

ZME(V) = - 3119 x 8

= 16

= 16 x 6500

14672 x 8 - 22 FV = 0 + 14672 x 8 - 22 Fv = 0 FV = 10063 lb. with sense as as+

= 24952
~

TE = 0, hence TE

in. lb.

To find CG take moments about Saxis tnr ougn E.

ZME(S) 3119 x 50 - (24/31.8) CG x 3 24 (21/31.8) CG ~ 0

sumed.
tihr-ough F.

To find Ho, take moments about V axis

ZMF(V)

Tdhence, CG = 8610 lb. This checks the value previously obtained when the reaction at G was found to be 8610. The 0 and V components or co thus equal, CGD = 8610 (21/31.8) ~ 5690 lb. CGV = 8610 (24/31.8) 6500 lb.

=- 6 On =0

22 Hn + 3119 x 14 = 0 - - 6 x 5690 - 22 HD + 3119 x 14

~bence, ~o

HD = 433 lb.

find FD, take ZD = 0

!D

=-

=-

?O

HO

GO - 3119

FD + 433 + 5690 - 3119 = 0

=0

about

Whence, FD
To

=3004 lb.

take ZV = 0 'i.y. = - FV + Gy - Hy + 14672 = 0 = - 10063 + 6500 - HV + 14672 = 0 find

HV

~nence,

Hv

=11109

To find load in brace strut 8I, take moments 0 axis through po irrt z. ~E(D) = - 14672 x 8 + 3 (BI) 22/24.6 + 24 (BI) 11/24.6 0 T~ence, BI = 8775 lb. and thUS, SlV = (8775)(22/24.6) = 7840 lb. BIS (6775)(11/24.6) = 3920 lb.

lb.

To tind Es take ZS = 0
'i.S

Fig. A2.57 summarizes the reactions as found. The results wl~1 be checked for equl11brlum of

Es -

3920

=0,

hence Es

=3920

-~"'-~,~.~

.~'-"

A2.28

EQUILIBRIUM OF FORCE SYSTEMS,

TRUSS STRUCTURES.

To find Eo take ZD
ZD

=0

= 5690 - 3119 - ED

= 0,
+

hence ED

= 2571

To find EV take LV = 0

LV = - EV + 14672 - 7840 Ev = 13332 lb.

6500 : 0, hence

Fig. AZ.59 shows a tree body of the top member F1i. The unknowns are Fy, FO, FS' Hyand HD. The loads or reactions as found tram the analysiS of the oleo strut unit are also recorded on the figure. The equations of equilibrium for this tree bOdy are:1:3

Lower Cylinder
Fig.A2.60

=0 = 22

3920 + 3920 + FS

0, or FS

=0

Torque LinJ<

(l

)~;~O::::iAxle
~

l:l'lF(D)

HV - 3920 x 2

7840 x 20 -

13332x6=0

Whence, HV 11110 lb. This check value obtained previously, and therefore is a check on our work. lJ'lF(V) = 24952 - 2571 x 5 - 22 HD whence, Hn = 433 lb.

=0
=
0

The torque to be transferred in our lem is 24952 in. lb.

~rob

zv - - Fv ...

13332 ... 7840 - 11110

whence, FV = 10063
ZD - - Fn + 2571 ... 433

The reaction R~ between the bvo units 0: the torque link at point (2), see Fig. A2.50, thus equals 24952/9 2773 lb.

whence, FD

=3004

lb.

Thus working through the tree bodies or the oleo strut and the top member FH, we come out with same reactions at F and H as obtained when finding these reactions by equilibrium equation for the entire landing gear. The strength deSign of the oleo strut unit and the top member FH could now be carried out because with all loads and reactions on each member known, axial, bending and torsional stresses could now be round. The loads on the brace struts CO and B1 are axial, namely, 8610 lb. tension ana 8775 lb. compression respectively, and thus need no further calculation to obtain design stresses. TORQUE LINK The oleo strut consists of ~NO telescoping tubes and some means ~ust be provided to transmit torslor.al moment becNeen the two tubes and still permit the lower cylinder to move upward into the upper cylinder. The most cammon way of providing this torque transfer 1s to use a double-cantilever-nut cracker type ot structure. Fig. AZ.60 illustrates ~ow such a torque length could be applied to the oleo strut in our prOblem.

The reactions R 3 at the base of the link at pOint (3) = 2773 x 8.5/2.75 = 8560 lb. ""th these reactions %nown, the strength design of the link units and the connections could be made. Example Problem 14 The landing gear as illustrated in Fig. AZ.61 is representative of a main landing gear which could be attachea to the under 51de of a wing and retract forward and upward about line AB into a space provided by the lower portion of the power plant nacelle structure. The oleo strut OE has a Sliding attachment at E, which prevents any vertical load to be taken by member AB at E. However, the fitting at E does transfer shear and torque reactions between the oleo strut and member AB. The brace struts GD, FU and CD are pinned at each end and will be assumed as 2 force members. An airplane level landing condition with unsymmetrical wheel loadii~ has been assumed as shown in Fig. A2.6l. 80LUTION The gear is attached to supporting structure at points A, Band C. The reactions at these pOints will be calculated :irst, treating the entire gear as a free body. Fig. A2.62

ANALYSIS AND DESIGN OF

14-

FLIGHT VEHICLE STRUCTURES

A2.29

19" ---'- 19" --: r17"----17.. ~!B

,

ZMA(D)

=- 60000 x

9 - 40000 x 29 - 66666 0 whence, BV = 78070 lb.

x 19 + 38 BV
~

28"
I

""' B~ace
-0

4"
Fig. A2. 61

" "
r,
"

"

To find AV' take ZV

0
+

zv

~oDo I tV
36"

=- 78070
=a

60000

40000

66666 - VA

L.
I

whence, Ay = 88596 lb.

5

To find BD take moments about Y axis through po trrt (A).

ZMA(V)

=57142

15000 Whee11 __ 10000 I Wheel 6000

x 19 - 15000 x 9 - 10000 x 29-38BD a whence, So 17386 lb.

40000

To find An take ZD 57142 + 15000 + 10000 + 17386 AD = a whence An 14756 lb.

=-

= = ZD = 0

To check the results take moments about Y and D axes through pOint o.

ZMo(V)
&~O(D)

=5 x 10000 + 14756 x
19 = a (check) 19

19 - 17386 x
+ 78070

= 20000 x 10 - 88596 x 19

=a

(check)

REACTT ONS ON OLEO STRUT OE

64"

Fig. A2.63 shows a tree body of the oleostrut OE. The loads applied to the wheels at
ITE E EO-

r
00
28"

17"'-1- i7"'

-1
9 "

Fig.A2.62

shows a space diagram with loads and reactions. The reactions at A, Band C have been replaced by their V and 0 components. To find reaction Cy take moments about an S axis through paints AB.

571~~
0

~ ~

.:l;G_ - -

4~~

~t
36"

'\!
I

ITE E _ES-J:'

Yi

r
L

28"

005-'

ZXAB

=-

VL

41/4-1

1---" 1-4 1/4

~OF5

0
"0"

(15000

10000) 64
l~. ~ith

Cv = 66666 assumed in Fig. A2.62.

T~ence

+ 24 Cv = 0 sense as

c The reaction at C must rave a line of action along the line ~~, 28G CD since ~ember CD 1s pinned at 24 0 each end, thus the crag component and the load in the strut 2D tollow as a matter of geometry. Hence, Co ~ 66666 (24/25) 57142 lb. CD 66666 (36.93/28) 87900 lb. tension

t 50000"lb.

25000 100,000 lb.

-l

Fig. A2. 63

200,OO@ -, 100,000 lb.

M'o"
t

50,OOO"lb.

To find BV take moments about a drag axis through point (A).

the axle centerllnes have been transferred to point (0). Thus the total V load at (0) equals 60000 + 40000 = 100000 and the total D load equals 15000 + 10000 = 25000. The moment of these forces about V and D axes through (0) are Mo(V) = (15000 - 10000) 10 50000 in.lb. and Mo(D) = (60000 - 40000) 10 200000 in.lb. These ~oments are indicated in Fig. A2.63 by the vectors with double arrows. The sense of the ~cment 1s determined by the right hand thumb

= =

A2.30

EQUILIBRIUM OF FORCE SYSTEMS.

r~le.

TRUSS STRUCTURES.
AD BD

and finger

The fitting at point E is Ges~gned to resist a moment about V axis or a torsional moment on the oleo strut. It also can ~roVlde shear reactions Es and ED but no bend1ng resistance about S or D axes. The unknowns are the forces DG and the moment TE'

19 -+-19 ~ &50000 I

""-A

IV I

32143

EB

Es, ED' OF,

AV~~17
A

_,.
I

t- li~BV
_ B

G.. 4164 47023, .IF

I!

To tind TE take moments about axis OE.

~~E

89215 \

'

77451

Fig. A2. 64

= - 50000

TE

= 0,

whence

T~

=
ZMA(D)

500CO In .u ,

To find ES take through point D.

ZMD(D)

~oments

about D axis whence

=200000 sa

Whence, By = 78070 lb.

28

=a

=- 89215

x 2 - 77451 x 36 -

38 BV

= 7143

Es = 0,

lb.

To r md AV take ZV
ZV

To tind torce DFy take moments about D axis through point G.

lMo(D)

= 89215 + 77451 - 78070 - Ay Whence, AV = 88596 lb.

~oments

=0

=0
t~rougb

=200000
=

- 100000 17 + 34 DFy :: 0

x 17 - 66666 x

To tind BD take
ZMV(D)

Y axis

A.

whe~ce,

DFV

Then DFS and DF

=77451

=77451 lb.
28~ ~:,\l>
17

= 50000 + 32143 x 19 - 38 aD Nhence, SD = 17366 :0.

AD

=0

(17/28) 47023 lb. (32.72/28) lb.

To find AD take ZD = 0
ZD

=77451 = 90503

= 17388

- 32143 14757 lb.

=O.

or AD

To rtnc IXlV take ZV

Zy 100000 - 77451 + 66666 - DGV or lXly = 89215

=0

=0,
lb.

These tour reactions check the reactions obtained originally ~hen gear was treated as a tree body, thus glvln~ a numerical check on the calculations. the forces on each ?art of the gear the parts could be designed for strength and rigidity. The oleo strut would need a torsion link as disc~ssed in example problem 13 and Fig. A2.60.
~lth ~~own,

Then, lXle DG

= 89215

=89215

(17/28)

=54164 lb.
= 104190 = 0,

(32.73/2B)

To r mc ED take moments about Saxis through point D.

l:!1D(S)

=- 25000
ED

=32143

x 35 + 28 ED lb.

A2.15 Problems

The results will be checked tor static equilibrium of str~t. Take moments about D axis through point (0).
l:!1o(D)

(I) For the structures numbered 1 to 10 deter~ne whether structure 15 statically determinate with respect to external reactions and internal stresses.
Pin 'A (1)

= 200000

+ 54164 x 38 - 47023 x 38 - 7143 x 64 200000 + 1949904 - 1892B28 - 457150 a (check)

l:!1o(S)

=32143

x 64 - 57142 x 36

=O(check)
I

~ ],8
:;-,0 10

Pin

~D AB E in"
1- 10
10

Pin

(2)

fIB
A

10

20

(3)

REACTIONS ON TOP MEMBER AS

Fig. AZ.S4 shows a tree body at member AB with the known applied forces as found from the previous reactions on the oleo strut. The unknowns are AD, BD, A~ and By. To tind By take moments about D axIs tr~ough A.

10tr: A (4)

~-10

(5)
B

A
A

'*'

*
I
;

I
B

(6)

10

10

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A2.31

1000*

(2) Find the horizontal and vertical components of the reactionS on the structures illustrated in Figs. 11 to 15.
..J+V

tt4'~0 T Pin
i

52
Slde vtew

~!~
A
(I')

00. 1-8,,1: 10"'"

~10-4 10
10"

+D

Wing
+V

J'

,
Front vtew

+8

21:'
1-16''-'1

(11) Pbi]:

Bi"'!>c

ho"

oj,

~O'
I

l~O
(13)

Flg.20

200f
Cable
(14)

500*

Pulley #~';=f:d~I"
100

k:c---r---,,,-i<'- -,..,..100/1

(6) Fig. 21 shows the wing structure of an externally braced monoplane. Determine the axial loads in all members or the lift and drag trusses for the following loads.

'O~'
A

'YO"

Front beam lift load 30 lb./in. (upward) Rear beam lift load 24 lb./in. (upward) Wing drag load 8 lb./in. acting art

= =

(3)

Find the axial loads in the members of the trJssed structures shown in Figs. 16 to 18 .

PLANV!EW Wing Drag Truss

...J. jl0

15"~' T 30"
10
....J..
( 24

t10

(16)

12 4

124"1
1'000

36"

-+-

36"

-+-

36" ~

45"------1

B'

~'~:::
...L

CF

30 dihedr

_800

10

1-30" -+
(4)

I 2000 I 15"+15'..-.l

(IB)

D,D' l- I 1 30'....
I . I I

.i-

CD CR

:~tL

Fuselage

A.A'
Flg.21

Determine the axial loads in the :nembers of the structure in Fig. 19. The member-a are pinned to supports at A, Band C.

(7 )

I-rl

5 Bays @ 30"

3I
Pin Reactions

~-DragTrUSSLS
t f t

--.j

At t t t l t t t t

t tt t

v t
1100/1

D
60" Fuse-

" '" B "'0. C

~v."'~
fL

at A, B, C
Flg. 19

'-D

.L

lage

::,.... '5 t.-s -E \> Flg. 22

~"'. L
B
:
I

25 lb/in.

' ,

--110'\--

(5 )

Fig. 20 shows a trl-pod trame for hoisting a propeller for asseT.bly on engine. Find the loads in the ~rame tor a load at 1000 lb. on hoist.

Fig. 22 shows a braced ~onoplane wing. For the given air loading, tind axial loads in 11tt and drag trusS members. The drag reaction on drag truss is taken oft at point A.

A2.32

EQUILIBRIUM OF FORCE SYSTEMS.

TRUSS STRUCTURES.

(8 )

(9 )

6"

....,.-_-,A B C

20"

,
I

24"
-13000 lb. Fig. 23 10000 lb.

10000

The r i tt tng at points A and B ~or the Landfng gear structure in ?ig. 23 prOVides resistance to VJ D and S :-8actions and moments about D and V axes. Find the reactions at A and a and the load in member CD for given wheel loading.

L22"----';

In the landing gearof :'1.>;.

member's
2~,

t
B

C F

the br-ace

Wheeo
"113'

wheel

'0"

res.c1:1.ons but only moment resistance about 1J axis. Find reactions at ~ and loads in :::embers SF and BC under given wheel loading.
15000 10000
Fig. 24

wheel

LD

BC and SF are f,'IW r or-ce

.::'itt::-:1g at E:

member's The

. .J500'
333 wheel 2

CD~ A

cr-evices ::-e-

sistance tc
V, 0 ar:.d S

Cessna Aircraft Nose Wheel Installation (Model 182)

Douglas DC-8 Jet Airliner Main Landing Gear Unit

CHAPTER A3

PROPERTIES OF SECTIONS CENTROIDS MOMENTS OF INERTIA ETC.

A3.1 Introduction In engineering calculations,

two terms, center of gravity and moment ot inerIx =J;tJdA, I y =/XQdA, I z =/radA tia, are constantly being used. Thus, a brief review of these terms Is in order. where IX! I y and I z are moments of inertia of the area about the axes xx, yy and zz respectivelY.
A3.2 Centroids, Center of Gravity, The cen-

Then

troid of a line, area, volume, or mass Is that pOint at which the whole line, area, volume, or mass may be conceived to be concentrated and have the same moment with respect to an axis as When distributed in its true or natural way. This general relationship can be expressed by the principle of moments, as tollows:
Lines: -

XL

ZLx, hence

= ZLx = J xdL L --L-

Fig. A3.1 A3.5

Polar Moment of InerUa In Fig. A3.1, the moment of inertia I z =fr~dA about the Z axis is referred to as the polar moment of inertia and vcrueeee- iv = ZVx, hence i = ZVX '=JXdV can be defined as the moment of inertia of an T --Yarea with respect to a point in its surface. Masses:- XI1 = zxm, hence X Zmx: = JXdM Since r~ = X~ + ya (Fig. A3.1) M M Iz <Fir + x ) dA = Ix + I y or; the po Lar- moment or inertia Is equal to the sum of the moments of If a geometrical figure 15 symmetrical with reinertia with respect to any two axes in the plane spect to a line or plane, the centroid of the tigure lIes in the given line or Plane. This is of the area, at right angles to each other and passing thru the point at intersection of the poobvious tram the fact that the moments of the parts of the figure on the opposite sides of the lar axis with the plane. line or plane are numerically equal but of opThe radius at gyration posite sign. It a figure is symmetrical to ~NO A3_5 Radius of Gyration of a solid is the distance tram the inertia axis lines or planes, the centroid of the figure lies to that point in the SOlid at whiCh, it its enat the intersection or the two lines or the two planes, and likewise, if the figure has 3 planes tire mass could be concentrated, its moment of inertia would remain the same. of symmetry, the centroid lies at the intersection of the 3 planes. ThUS,fr a dl1. =faM, where ~ is the radius at gyration A3.3 Moment of Inertia The ten moment at inSince, r"ciM = I, then I =f''' I1 or f =\ IT ertia is applied in mechanics to a number at Vl1 mathematical expressions which represents secBy analogy, in the case ot an area, ond aonents at areas, volumes and masses, such as I =f'A or =Ift fY'dA, r'dY, fr'd1'! etc.

jc-eas r- XA

Sax, hence X

= Zax =fXdA
A A

A3.4 M01Ilent of Inertia of an Area As applied A3,.7 P&rallel Axis Theorem In Fig. A3.2 let I y be the moment of inertia of the area referred to to an area, the term ~ament of inertia has no phYSiCal sig~ificance, but represents a quantity the centroidal axis y-y, and let the moment or entering into a large number of engineering inertia about axis y~y~ be required. y~y~ Is prOblems or calculations. However, it may be parallel to yy. Consider the elementary area dA ~onsldered as a factor Which indicates the inwith distance x + d from y~y~. r luence of tue area i t.se Lf in determining the Than, Iy~ = (d + x ) IlidA total rotating moment of uniformly varying forces applied over an area. =JX'dA + 2dJXdA + d fdA Let Fig. A3.1 be any plane area referred to three coordinate axes, ox, oy and oz; ox and oy being the plane of area. Let dA represent an elementary area, with coordinates x, y, and r as shown.

A3.1

.3 2

CENTROIDS,

CENTER OF GRAVITY,

MOMENTS OF INERTIA

y
I

.c:":
>dA
Fig. A3.2
g.

TABLE 1

Section Properties of Areas

Rectangle

--1- x ....,
I

fe.
I

Area d Y

= bd = bd 3 / 12 = Oct'
3"

I1._1. I _
a
2dA.

The first term,! x represents the acment of inertia of the body about its centroldal
axis y-y and will be given the symbol I. The second term 1s zero because~xdA 1s zero since TJ 1s the centroldal axis of the body. The

f ,-, = .289 d f------------:.-------------l Area = bh

2

the square of the distance between axes yy and

Y1.Y1."

last tern, d2~dA

= Ad

or, area of body tl~es

Thus in general,
I=I+Ad
a

12 This expression states that the amount ot 1 _ inertia of an area with respect to any axis in 2 bh' 3 3 = the plane of the area 1s equal to the moment ot "4 inertia ot the area with respect to a parallel = .236 h centroldal axis plus the product of the area and the square ot the distance bec#sen the two axes.r-----------------~-------------------------_1

3 ,~
~
1

Y = h/3
11._1.
1,;1_a
3 = bh "V"

Triangle

vO

f---- ---i

---r- - - I 'l 2

= bh 3

f ,-,
Area

Parallel Axis Theorem For Masses. Ifinstead of area the mass at the body is considered, the parallel axis can be written: I = I + Md a , where M reters to the mass at
the body.
A3-.7a hss Jloaents at Inertia The product of the mass or a particle and the square at its distance tram a line or plane is referred to as the moment of inertia at the mass at the particle with respect to the line or plane. Hence,

= bh
'2

x = 2/3
11.-1.

h
h~
.3

4,8

Trapezoid

r- bl-j

Area

= d(b

+ b 1.)

'j = d(2b be)

3(1) + bl.)

I = ZMra.

It the summation can be express-

ed by a definite integral, the expression may be wr-L tten I = J r-e dl1 Moments of Inertia of Airplanes. In both ilying and landing conditions the airpl~~e may be subjected to a~~lar accelerations. To determine the magnitude ot the accelerations as well as the distribution and magnitude or the ~ss inertia resisting torces, the moment or inertia ot the airplane about the three coordinate axes is generally required in mEL~ing a streSs ~~lysis of a particular airplane. The mass moments ot inertia or the airplane about the coordinate X, Y and Z axes through the center or graVity at the airplane can be expressed as tollows:

= 0.3

(b 2 + 4bb , +b 2) 36 (b + b, ~

Ix Iy Iz

= Z/fY2 + = Zwx' = Zwx 2

Zwz 2 + Zl.I x
Zwz ll + Zl.I y r,wy2 + Zl.I z

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A3 3

Elliptical

TABLE 1 - Continued

I--------j Area := ted + za)

Ring

~~l:
I

'~f

Ix

:=

y = d/2
Ix
:=

x-l -i~X

4:

n (a 1 3b 1

aa3b:a)

bd 3 - a(d _ 2t)3

12
deb +a)3 _
20

W
6ab:ae

ex = V A}ea
Area ::: .215 a a

zaec Ix - I y

= (dt-t') (b'-bt)
and I:a- II see equagiven for angle.

Tl1
~al1
x

X ::: .223 a

l::

Circular Fillet

u-sectton

r:n.n lIT:[b1

-jtJI-

a~r.o-

Area

y :::

A ::: 2dt 1 + at d a t 1 ~ .5t a a

A

-r~

I'

r-b~=r
1 11

~
Xl

. x

==rr
I ,>,

..L

'I_it-

Parabolic Fillet

x ::: 3 4:

Xl

y = .3 Y.

3
Area = nr'

Circle

2--;-~-2

= nr'

T
4:

y ::: 2 y.

l\~l
Ring

....

-;

"

la_Ii = 5 n r-

'i
TABLE 2

e1-l

= r-

2
(roll - rill)
Solid Cire. Cyl.

Properties of SolidS
Vol.

~ .. '/~
1 ~ ~ -----:..-.l- 1
"
',I

Area ::: n

s
=

n r-'l-L,. (r - radius)

ro '
l

11. -1
.'

:::

n (ro" - rl")

4:

.~
~

Jo.

)5:5' '" ,
'\
.'

M = W (Total wt.)
11
-1

~"
\~

.-

~.r.a

"2
+
-

Area _ n r "

I,~

= M [r a

(L f31])'4
ra~

Serm-crrcfe

Hollow Circ. Cyl.

y:::
tx
v

.424r
.1098 r< = .264 rrt

Vol.
11. -1 1a _ liI

=:::
:::

IT

L{r1 a
a

M(r1

~~~lx

t:x

,
q=outside

I1(r1 a

+ r aa)/2 + r a ll + La/3)/4

For thin hollow eire. cyl.

1
1 _ 1

Area semt-cu-curar Ring

(R

a _

r ")

radius

- Mr a

2
y ::: 4 (R '1+ Rr + ra) appr-ox. y:::2r 3n R+r IT

Vol. ::: abL, I1 ::: W g

11 -

:::

M (a:a

bll)/12

Ix:::.1098(R"-r")-.283Rll!"" (R-r) R+ r

1"1-'01 ::: MLa/12 ll/3 13 -3 :::ML

I x(approx):::.3t(r +R)::S when ~+-----------------------I

i s sm8l1

Lr Sp h ere S o lid
Sphere

Vol. = (4nr'I/3

Iabout axis =
2M r "1/5

Ellipse
I

Area ::: nab

~~ II' j

n a sb Ix = 4xf'x

Thin Hallow Sphere Vol. :::

(r1 :J -ra
r 1 ,. radius of sphere

3
3 )

rt

~~

=a

'2

d1a. 2Mr, '

"""""3""

13.4

CENTROIDS, CENTER OF GRAVITY,

MOMENTS OF INERTIA

TABLE 2 -

Co~t~nued

J;tf'x
y

Ring with Circular Section Y

NAx

- r-e t

~2

sIn 20
4

- (l-cos a

0)

M= w
g

mass per unit volume of body.

1:0: =mnaRaa[R' + (Sa a/4)J 1I a I yy =! ml1 a Raa (4R + 3a )

TABLE 4 Centroids ot ':'rap1zaidal Areas

2
TABLE 3

Lc1JT
b I

J.

ll-y -!-x IRatio a

Section Propert1es of Lines

(t 1s small in comparison to radius)
Circular Arc

Distal1Ce x
0.40992 0.4098-4. 0.40976 0.4988 0.40960 0.40952 0.40944 0.40936 0.4928 0.4920 0.40912 0.40905 0.40898 0.4890 0.4883 0.4877 0.40870 0.4862 0.4835 0.4822 0.40809 0.4795 0.4782 0.4770 0.40758 0.4746 0.40733 o 4721 0.4693 0.4667 0.4641 0.4616 0.40.592 0.4568 0.4545 0.4523 0.4502 0.4482 0.4462 0.4443 0.4409 0.40375 0.4343 0.4312 0.4284 o 4259 0.40233 0.4209 0.4188 0.4168 0.4128 0.4090 0.4060 0.4030 0.4000 0.3975 0.3950 0.3928 0.3848 ~. ~810 o 778 0.3750 0.3725 0.3702 0.3668 0.3636

Distance y
0.5008 0.5016 0.50240 0.5032 0.5040 0.5048 0.5056 0.5064 0.5072 0.5080 0.5088 0.5095 0.5102 0.5ll0 0.5ll7 0.5123 0.5130 0.5138 0.5145 0.5151 0.5165 0.5178 O .5191 0.5205 0.5218 0.5230 0.5242 0.5254 0.5267 0.5279 0.5301 0.5333 0.5359 0.5384 0.5408 0.5432 0.5455 0.5477 0.5498 0.51H8 0.5538 0.55.57 0.5591 0.5625 0.5657 0.5688 0.5716 0.5741 0.5767 0.5791 0.5812 0.5832 0.5872 0.5910
~-

Area = 2 11 rt

11._1. = 11 rat

1.01 1.02 1. 03

1.04

pol ar = 2

11

r"t

E'x = .7(f7r E' polar = rseeu-ctrcte Arc

A = n rt
y

I.::L-__L.LI
Quarter-circular Arc

2~2

= .6366
=
11

1._.
y

-2-

rat

1.05 1 06 1.07 1.08 1.09 1.10 loll 1. 12 1.13 1.140 1.15 1.16 1.17 1.18 1. 22 1. 24 1. 28 1 28 1.30 1.32

t:~~

o:;~~

= .2978 r >t
-2-

Area = n rt

~ r4-.--'-1
t l
2 I

1.3<

= .6366 r

1---;2

1.36 1.38

= 11

rat -4-

1.<0
1.405 1, 50 1 .5.5 1. 80 1.6.5 1. 70 1.75 1.80 1." 1.90 1.95 2 00 2.10 2.20 2.30 2.40

I._iii = .149 r"t

CIRCULAR ARC

;.~~
2.70 2.80 2.90 3.00 3.20 J 40 3..60 J.80 4.00 4.20 4.40 4.60
4.'~

Area

= art
sin a a '

in

Radians

=r

0.5970 0.6000 0.6025 0.6050 0.6072

(a + sin 2(1)

'.00 5.50

~:~:;

6: ~~~~ 0.6152
0.6190 0.6222 0.6250 0.8275 0.6298 0.6332 0.63640

~.~
7.00 7.50 8.00 9.00 10.00

=r

(1-008 c 0)

,l1xx=.W

= rlilt

(j-cos a)

- r"t (0 I xx 2

sin 2(1)
2

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Table"

A3.5

I"
I

'eight IArJaSX
Bo. Bailie fI/I

HorizOIItal lIo.ent
S

.x

Vertical ArJa .. 1 lIo.eut

.Jy

1 2 3 4 5 6 7
8

9
10

Propeller Engine Group Fuselage Group .ing Group Hori. Tail Vert. Tail Tail Wheel Pront Land. Gear Pilot Iladio

180 820 800 SOO 60 40 50 300 200 100

0 in, 46 182 158 296 33.:1 328

o
37120 145600 94800 17760 13400 16400 34500

o o
~18

0
3200 -10800 480 1040 _1000 _ 900

"

115
16.:1 240

8 26 _20 _30

33000
24000 411180

10 5

2000 500
.5480

where Ix, I y, and I z are generally referred to as the rolling, pitching and yawing moments of inertia of the airplane.

I"

Totals

3150

Example Problem 2, Determine the moment ot inertia about the horizontal centroidal axis for the area shawn 1n Fig. A3.4

= weight ot the items in the airplane x, y and z equal the distances from the axes thru the center of gravity ot the airplane and the weights w. The last term in each equation is the summation ot the moments of inertia of the various items about their own X, Y and Z centroidal axes. It w is expressed in pounds and the distances in inches, the moment or inertia is expressed in units of pound-inches squared, which can be converted into slug feet squared by multiplying by 1/32.16 x 144. Example Problem 1. Determine the gross weight center ot gravity of the airplane shown in Fig. A3.3. The airplane weight has been broken down mt c- the 10 items or weight groups, with their individual e.g. locations denoted by the symbol
w

Solution. we rrrst :C1nd t he moment of inertia about a horizontal reference axis. In this solution, this arbitrary axis has been taken ~s axis x'x' thru the baae as shown. Having this moment ot inertia, a transfer to the centraidal axis can be made. Table 5 gives the detailed calculations for the moment ot inertia about axis x'x'. For simpliCity, the cross-section has been divided into the tive parts, namely, A, B,C,D, and E~ lex 1s moment at inertia about centroidal x axis at the particular part being conSidered. Distance tram axis x'x' to centroidal horizontal axis = = :Av = 17.97 = 2.9l w ZA 6.182 BY parallel axis theorem, we transter the moment at inertia from axis x'x' to centroidal +. axis xx, Solution. The airplane center at gravity will be Ixx=Ix'x' -Ajlt=79.47-6.18x2.9P.= located with respect to ~NO rectangular axes. In 27.2 in" this example, a vertical axis t~~ the centerline at the propeller will be selected as a refRadius of Gyration, Pxx = xx =lj27 .2 = 2.1" erence axis tor horizontal distances, and the A 6.18 thrust line as a reference axis for vertical distances. The general expressions to be solved

VI

are:distance to airplane cvg , from ref. axis -S-i5 y = Zwy distance to airplane c vg , from ZW ref. axis X-X Table 4 gives the necessary calculations, whenee x = 417180 = 133.3" aft ott propeller
Zw
3150

= Zwx

Centroidal

Fig. A3.4

Fig. A3.5

5480
3150

-1.74" (below thrust line)

Example Problem #3. Determine the moment at inertia of the stringer cross section shown in Fig. A3.5 about the horizontal centroidal axis. Solution. A horizontal reference axis x~xP is assumed as shown. The moment of inertia is first calculated about this axis and then transferred to the centroidal axis xx. See Table 6.

11)\
.......,F=""-,---- - - 7
Fig. A3.3
Z
I

, Ref. line

.3

, CENTROIDS , CENTER OF GRAVITY , MOMENTS OF

Table 5

INERTIA

Portion 3 + 3' (ref. Tcble 1 )

" "
e

A 0.'"
Al 0.50 1.00

'" a. ,
'.25
2.875
0.375

",

_.ent of inertia of

".
13

portlon about centroldal

lI:

uia.

2.7li 15.13 1/12 x .5

2.75 15,1:3

..

..

, , ! ,
~
0

",

y= .375 +

4 X (.0841 3n

.0725 + .0625 ) = .375 + .54

I
I

.~72

= .547"
:OJ.

appr ox ,

y = .375 + 2

= .375 + .172

=547"

.0< .0<

15.17 15.17
27.52

n
lex= [.1098(.0841+ .0625) + .04x .54.283x .0841 x .0625 x .04] = .002375

,
'0

5.2S 27.50 1/12 x 2 z .53

.ca

2.12

6.10 17,56 1(12 x .5 x 4.25 3 3.19

0.28

20.75 0.29 0.29 0.28

.54
appr-ox ,

0.281 1.00

0.28 bb 3/36 .- .75 4(36.0.01 0.28

D1 0.281 1.00

0.28 0.56

..

lex = .s s .04x .27"= .002365

In~

.. 0.01

i5 \1116 . 182

1.50

0.21 1/12%2x .7:i 3 0.07

17,97

I
Table 6
A,
0 0

179.47111.4

Problem #4. Determine the ::J,oment of inertia of the flywheel In Fig. A3.5a about axis of rotatlc n. The material 15 aluminum alley :::ast1ng (weight = 1 lb per cu. Inch. )

PartiOD I

Area

A,'

AI
lex
Ix 'x' .. lex... Ay2
.00281 .03428 .02078
O~787

Sec. AA

V/// / ~-------,

+ +

l'

.0.
.06~4

o .00281

.706 .04617 .0326 .00168

3'

.0679 _.547 _.0371 .0203 .000475

s~

.1933

.009

y = l'.Av = .009 = .0465" :T933 l'.A

1= = lx~x'

Aj3

= .05787
in~

u~ cb .. \ w~
I
5"

f;i l,.,.,f/ ""'~

1"

r/

2"
I

10"
I
,

I
I

.1933 X
05745 = .55" .1933

IA

Fig. A3.5a

r/////il 1" i . -i--' . :--3'l

-'--

( .0465)' = .05745

Radius or gyration i.ox-x

Solution: ~he spokes ~y be treated as slender round rods and the r-rn and hub as hollow cyl Ind-. ers , (Refer to Table 2) Rim, weight n(R ," = = n(5"'-4.l1)
3_r

Detailed exptanat i on of Table 6:Portion 1 ... l'

lex = 1

a)x3x

.1
)

bd'

12

= (-04

~2' 75')
4 3n

x 2

= .00281 in-

2"'r 2 2 I = .5W(R "' 4 2 ) = .5x8.48 (5 173.7 lb. In'

.3=8.48 lb.

Portion 2

(ref. table 1) y = .375 + 4 Rlil ... Rr -+ r- = .375 +

3n R+r
)

Hub,

'I"
Spokes,

= n(la_ .5 2 ) x 3 x .1= .707# .44 lb. in~ = .5x.707 (l a .... 5:1

r=

(.5

....

5x .54 .... 54 2
+

.375 .. 331 = .706"

( .0

.54)

Length of spoke = 3" '""eight of spoke = 3

By approx. method see Table 1. y = .375 +

g
n

rl, = .375 + z x (.52) = .706"

rt

Icx

= .10ge x

(R a + r-") t (R

r) - .283 Ra + r.llt
R + r

x .25 2 X .1 = .588# I or one spoke = ',.rL:a/12+>"j'r~ .588 x 3'1/12+ .588x2.S a = 4.10 lb. 1n 3 I for 4 spOkes = 4x 4.10 = 16.40

Total I of wheel = 16.40 +

. 1098(.2916+ .25)x .04x1.04-

;44, + 173.7 = 190.54 10 In' I = 190.54/32.16 x 144- = .0411 slug ft~

Problem #5.

.283 x .2916 x .25 x .04

1.04 = .002475 - .000793 = .001682 in ~.

Approx.

lex

= .3trl,"

= .3 x .04 x.52

-.

= .001688 in".

Moment or Inertia of an Airplane. ':'0 calculate the :nO:r:lent of inertia or an airplane about the coordinate 9.xes through the gross weight cen.ter of graVity, a breai< down or the air?1ane weight and its distribution 15
~ample

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A3.7

necessary, which is available in the weight and balance esti=ate of the airplane. Table 6a shows the complete calc~lation of moments of inertia of an airplane. This table is reproduced. 'from :i.A.O.A. 'I'ecnnt ca L no te #575, ~sstimation of ~oments at inertia of airplanes from Design Data.~
t~e

axes, the prOduct of i~ertia of the weight about the reference axes 1s necessary. Column (14) gives the values about the reference axes. To transfer the product of inertia to the c.g. axes of the airplane, we make use of the parallel axis theorem. Thus

Z'NXZ C g -=Z'NXZ(Rer. axes) -Zwxz =48,857,5895325.3xl15.9x77.8=B39,253 lb. in.a Explanation of Table 1 To reduce all values to slug ft.a multiply Fig. A3.5b shows the reference of planes 32:l7 and axes which were selected. After the moments x.J:.... of inertia have been determined relative to 144 these axes the values about parallel axes through the center af gravity of the airplane Hence Ix -= 3061, I y = 6680, Iz -= 9096, Ixz -= 181 are found by use of the parallel axis theorem. Having the inertia properties about the coColumn (1) of Table 6a gives breakdown of airordinate c.g. axes, the moments of inertia bout plane units or items. the principal axes are determined in a ~anner as Column (2) gives the weight of each item. explained for areas. (See A3.13). Columns (3), (4) and (5) give the distance or The angle 0 between the X and Z axes and the the c.g. of the items from the references planes principal axes is given by, or axes. tan 2 0 = 2I xz = 2 x 181 -= .05998 hence >:1 = Columns (6) and (7) give the first moments of Iz- Ix 9096-3061 the item weights about the Y' and X' reference axes. 10 43" columns (8), (9) &id (10) give the moment of inertia of the item weights about the ~eference axes. Columns (11), (12) and (13) give the moments ot inertia of each ltem about its own centroida1 axis parallel to the reference axes. Such items as the fuselage skeleton, wing panels and engine l , ' have relatively large values for their centroid/ ". " ' \ ~ al ~oments of inertia. \ \ The last values in columns (3) and (5) give , ~ \ the distances from the reference planes to the " , , . center of gravity of the airplane. ~';~ Xc g . -=Z'~'X-=617.024 (col. 6) ==115.9 in. l Zn 5325.3 (col. 2)

'. . . ...

. .

ze

'0'

-=

~wz-= _

Zw

414.848 (col. 7) -= 77.8 in. 5325.3

The last values i~ columns (8), (9) and (10) were obtal~ed by ~se of the parallel axis theorem, as follows:Zwxa about e.g. of airplane -= 97,391,595 - E325.3 x 115.9 2 -= 26,691,595. zwa 2 == 33,252,035 - 5325.3 x 77.8 2 -= 992,035 The tmru :'rom the last value in columns (11), (12) and (13) give tr..e .noment.s of inertia or the airplane about the x, y ar.d z axes through the e.g. of airplane. T~e values are obtained as follows: 1y =Zwxa + ZWZli + Z<lI y -= 26,691,595 + 999,035 + 3,120,384 -= 30,304,014 lb. in s Jx = Zwy~ + Z'Nz. 3 + Zl;I x -= 10,287,522 + 992,023 + 2,899,470 -= 14,179,027 lb. In.''' I z -= Zwy a + Z':.,rzl1l + ZIlIz = 3..0,287,522 + 26,691,595 + 5,157,186 = 42,136,303 lb. 1~.2 In order to dete~ine the principal inert:a

:
i

\\
( (
~

,
(

, ',0--_.-- .
\---:...-,'--

\\t-"'; /~
\'~
\
-"'y,

-.---'1-}'

~~

.:

/
L-'-

'\
\

.
~
~

"\' ~
l

_ ~:f ........ ...

>="
Fig. A3.5b

es ,,:

(REF. N.A.C.A. TECH. NOTE No. 575)

A3.8

CENTROIDS, CENTER OF GRAVITY,

MOMENTS OF INERTIA

,,C.ll~e:r

i .

,.
elillt

,.

H.

..
Y
.

COMPU!ATIOHS 0' MOMXiTS OF IBERTIA

S.

,eCl't1oD. no,......C1bly Clln~ll:r oectloQ tc . Center .ecUon rlbl, eee . Flap Outer p&l1el ~ Outer paDfll Outu p&l1.1 rib' Al1nona

.....

106.6 102
204.6 121

.... .. .... ....

".0 ,.0 104.6 105 155.5 120

139

- " - " ee 53 - .5 lsa

lsa '5 ISS 54 lsa sa

" 11,098
24,751 141.462 3,960 10,983 18.672 12,462 5,401 31,966 11,053 55,4&' 141,319

,.

"
6,20a 11,662 4,631 1,166 6,799 10,114 5,747 1,947

. ,,'
.

~'ole

6a

"Y'

1,131,955 2.995,549 1.844,317 71a,8OO 1,153,215 2,240,640 1,735,Oae 928,938

0 0

353,4;;1 664,745 254,705 61,796 441,935 657,4ol0 367,621 120,702

814,~3

..'

10 ..

11.

12
t '

."
361.339 491,24.5 aoZ,164 48,598 164,514 274,478 158,407 55,390 176,378 110.200 69.394 5,184 15,181,

.,,
251,229 4091,245 235,844

lJ

H.
m

532,563 1,411,126 685,388 209,880 713.895 1,213,680 798,861 334,850 3,091,083 1,381,600 4,4075,384 194,400 985,520 61,600 23.800 280,000 l,048,S7<l 10,.72 131,670 152,366
Z8:3, .500

33,680

Ho:rlEOll~al tall Ve:rtlcal tall

87.1 ,.,

31.4 172

2,545,546 3,186,682 2,185,373 764,150

17,601
0

48,596 184,514 374.,4078 175,008 55,390 176,378

31.4 OS.

96.7 125 .1. .0 .0 .0 '0 .0 10.

8,423 11,731, ua 3,925 3,890,586 .35,434 9,726,484 145.800 3.139,026 53.iOO 51,800 1",000 l,417,5l4t 11,858 418,950 373.034 529,200 104,853 53.,900 347,633 49,005 83S.708 1,111.688 145,200 3.073,490

rUI.lac

oklle-

3H.0 1.,. 40.5 ".5

to. Eng1Il.. _ _t
ruztl.~k

( fa.1:rll'11) f1l'n&ll Step.

If . .l.C . .l.

...

'0

11.0 '0 '.0 170

ec
0

490.625 2.060,154 259,200 310,400 70,400 600 .... 000 77S.e56

10,174 1.5713,594 5,184 57,861 1,100

0

.n

1,5'ro,OOO 5,164 56,648 1,100

0

'.""
'70
:l4O

a.aac
0 140

3,4140

c~l1GS

C.. bln &lid. _1lll11.bi ..111 foo-c tI'O~U floor, "'tllg flllOlU 8o'Cto.. eow11Jl4 ~ 0111e fram Al'l'el-C1JWJ ;;1001' 'l'a11 .cl.el parI.. e-Cc. Sld.e doora ae.aase door fabrio &D4 40pe Ta11 cone Co_Unr, atadOI18 1-3
,,~

-c .0 66.5
'.0

18.:1 142 27,0 140 1.3 '.0

'.S 210

," ,. s

SO

, ,.,

..

20 11
0 0

..
"
54

',500 7,18a '.709 '.600

1M 1,995 .,Sa7 3,180
1.073 2,02:1

'.800

.3,200

16,940 2,394

15,470 106,400
0

15,470 108,794 1.976 18,944 24,300

58 '5

I,. 62'

SO
0

'.400 3.267
0 0

9,248 4l,3S2 62,234 151,875 5,160 za.aa4 114,308 7,144 83,200 62,108 108,300 604.472 128,JS5 385,378 142.376 6, '7l3.600 ea4,143 83,538 1,420,300 361,325

'0.

1.Z69 18,944 24,300

I._
2.431

"'.
"'"

sa ase
1.394 11.3 1,040 3 1.140 11,6:12 1.916

100 1,088 720 ".394 120 1.200

0

CIla,e1. (1'_ trected) Retr..cting "ocbWbl., 1'atl _.bee.l

,. fuel .,.,to_
Slectr1aa.l Koht 011c.g
UJU!'L.,t .U .
.qu1~t

l:J:1g1n. accaorl Eng1A. control. PTOp.ller S'ar1:1~ .ye-c_ Ll,lbr1ca'Uq aye

_.e

""

.. ....

332.4 115

1049.0

90.' ,
ae,c

aaa.c

.... .. - " ,.
0

17.0 143 1 .a ,.5 13,0 '.5 12.0 110

,., 0

...

ie

.3 .0 91

5.se8
0
0 0

se

3,302 a,8S8 1,320 26.726 3,146 U.83l

28,288 15,509 0

s
100 720 15,509 120 1.200 23,240
0

raa ,640
199,342 18.111 z64,l60 ::>S2,763 :1...35.400 1,383.0<36 210.7841

23,250

av.acc

Sl.

e77,678 17,875 285.356

0

23,240

no
J3 52

as
54

see.ceo
1,809,171 3.369,600
1.142,~1

91.0 lU ".0 "0

sa

0 0

.0 .3

M,617

'.1.133 a.065 .3,0'73

S._
1.924 83,920

ee
253.658 9,060 177,600 1,374 .32,008 .3,4341 U13,ll84 108,160

ee
253.858
0

718,536 692.640 2.769,3&0 391.030 88,109 185,168 116,120 147,J.08 839,680 356.592 9415,840 1,996.600 l,04.,O!8 70,380

'.71.1. '.520

11.0 1.03 '.3 37.0 sa

10

InetZ'WIIl!'l::l'. 81U'face COAtrol. F\lnl1a.bio.p.

841.0 12. 38.0 10. 81.5 160 160.0

0
0

.5

SO

sa

I._

17.760 3,145 .3.132

.,.

244.9941 118.699 19,201 116,032 U3,798 1.343,488 395.3541 .3.088,400 3,893,160 1,849,869

0
0

sa

.0 '1.

10,496 3,8'16 13.040 .34.960

'.sao 3.498
5,187 13,800 8,157

I,l.OO -

3:13,858

'.060

l'U.834
534,300 331.6341 4010.842 1.0M,OOO 560,593

a.

ea.sce
0

69,300 1,.374

'"

..,

20,008 130,4.QO 400.000

0

SO. 600

35,668 63,233 254,:562 508,1150

lsa

IS
0

SO
SI

100.7 128 '.0 US --3867 106.6 200.0 10' 200.0 205 780.0 132 75.0 Yl 195 '.0 IS' '.0 190 148.7 l7S '0 1.6 '.0 I., 34.0 1.'"

10. 0

IKP'n'

,go 81.' 7i,35O 82.424 -----74.8 41.41,199 .2a9,l38 67,342.573 10,283,443 aa,a63.738 1.8,000 .3.205,000 17,800 8.405,000 66,690 13.590.720 6,375 378.075 148,296 13:1,424 324,900 12.130 4,5.31,307 23.680 I,. 78.854 108,900 3,434 628,864

",""

32,556

ea.

2,890.14 28.800
38,600

...

.3,881.572 33.600 33,800 141,180 3,615

4.980,508 31,090,714

OIl

Pllot Ob.e1"W'.r Fuel Vlry P1eto1 3.:)118 eaDd.l., Flo.. t l1rttta Cbart board., Or1!t 11ght Tlre-C a.1d. Life :raIt U8U1JL LOAO fOULS COflR.u::,.IO.

.....,.

.,.

..

'"

... -, ."
14 14 14
72

U
0

go 21,000 41,000 85.S 1041.990 5,3a5 .5 751.

se "
54

I, 7l.O
25.11

",.

.5

ese

. , . ...
,

10

101.

'55 "0 4,S.34

.. ...

.,

1.7&4
0

t ,

..
J3

'00

1,6ao,OOO 1,584,.300 ;',701,996 541,875 21,938 39.864 46,656 1.031,008 32,693 U .390 12,996 3046,834

172,380
0 0
0

157,560 3,675

'.000 '.000

-,as
0

5.U7

0 0

1457.9 140.5 5335,31 53a5.3 1+5.9

88.7 204.828 1.26.450' 30.549,023

4 ,079 ~88 ,299

5,13'1 see
0

1,890,000 3.649,000 8.603,Oao 452,625 57,038 70,6sa 143,120 3,159,051. .37,824 aQ,837 37,540 461,024

311,898

229,980

1'16,878 17,786.87:1

" .. ,
I

, , ,
,

, ,

J:c.g.

817,0.24 414,848 97,asl,595 10,.387,saa ~S.2,035 2,899,470 3,120,384 5,157,188 48, 857,58a 77.8 r71 200,000 Or~O 000 48, 0 1 33t1 IE.t\9i,5il5 10.Z67.5aa l~a,oj5 a~ ,aS3 1 10.287,5412 as. 691. see 26,691,595 , , , 292 035 1~ 287 52.3 I i , 99. S'~ 36,804,014 4,136,303 , I 1",179, I .3,662 5,784 I 7,912 , , , 51.8 76.1 , , , a

, ,

, ,

, , ,

~--"'~'i"

, ,

, ,

,
,

, , , ,

, , I ,

1'x

J e.g. I a.g.

, :

:_~ :'~---~'3

..,,< ....

'.~y

' ".

~.

"~':;,

> .... ,..,..r"_.~"'-~~"_.

t,

-'):

~",. <:, ,...:<:

. . . >:';.,~~~-~;-"'-::': ...~:~:~i~~ . . .;;.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A3.9

The principal moments of inertia are given by following equation.

I xp = 11: cos"

will be given the symbol Ixy, hence - -(1) Ixy = I xydA - - - - - The unit, like that of moment of inertia, 1s expressed as inches or feet to the 4th power. Since x and y may be either positive or negative, the term Ixy may be zero or either positive or negative. ~oduct of Inertia at a Solid~ The product or inertia or a solid is the sum or the prOducts obtained by multiplying the weight or each small portion in which it may be assumed to be diVided by the product ot its distances trom two or the three coordinate planes through a given point Thus With respect to planes X and Y Ixy = I xy dW
Jxz =
I yz
A3.9

I + I z sin" 0 -

Jxz sin 2

I.

(see

Art. A3.lll
I yp = I y I z = Ix sin" 0 + I z cos" .0 SUbstituting
+

I xz sin 2 .0

Ixp = 3061 x (0.9996)" + 9096 x (0.0300)' - 181 x .0599 = 3056 Izp = 3061 x (0.0300)' + 9096 x (0.9996" + 181 x . 0599 = 9102 Iyp

=6680
Problems

A,3.7b

n
'"i

r Z i ""i

I =I

xz dW
yz dW
S~try.

Product of Inertia for Axes at

1
I
I

'11$
Fig. A3.6

8
Fig. AJ.7

kR

.f0~

,~

It ----IT,

(1) Determine the moment of inertia about the horizontal centrOlda1 axis for the beam section shown in Fig. A,3~6. (2) For the section as shown in Fig. A3.7 calculate the moment or inertia about the centroidal Z and X axes.

It an area is symmetrical about two rectangular axes, the product ot inertia about these axes is zero. This tallows tram the tact that symmetrical axes are centroidal x and y axes. It an area is symmetrical about only one of two rectangular axes, the product or inertia, !xydA, is zero because tor each product xydA tor an element on one side at the axis or sYmmetry, there is an equal product at opposite Sign tor the corresponding element dA on the opPosite side or the axiS, thus making the expression !ydA equal to zero.
A3.l0 Parallel Ax18 Theore.

T,
1

T
A3.S

' 1-'"

1<.'-1"

Fig. A3.9

The theorem states that, "the product at inertia ot an area with respect to any pair at co-planar rectangular axes is equal to the product or inertia at the area with respect to a pair at parallel centrold.al axes plus the product at the area and the distances at the centroid at the total area tram the given pair.ot axes", Or, expressed as an equation,

Ixy F Ixy + AiY- - - - - - - - - -(2) This equation is readily derivable by re(3) Dete~ine the moment or inertia about terring to Fig. A3.l0. yy and XX are centroidal the horizontal centroidal axis tor the section axes for a given area. YY and XX are par-a.lLeL shown in Fig. A3.8. axes passing through point O. (4) In the beam cross-section at Fig. A3.9 assume that the tour corner members are the only The product or inertia about axes YY and XX effective material. Calculate the centroidal is I XY = I(x + x)(y ~ Y) dA moments of inertia about the vertical an~ hori= IxydA + 'is I dA + x I y dA + Y I xdA zontal axes. The last two integralS are each equal to zero, since !ydA and IXdA rerer to centroldal A3..8 Product of Inertia axes. Hence, I xy = IXYdA + i j IdA, which can be In various engineering problema, particu'Nritten in the tarm or equation (2). larly those involving the calculation of the ~oments ot inertia at unsymmetrical sections, the expression I xy dA is used. This expression is referred to as the product of inertia at the area with respect to the rectap~~lar axes x and y. The term, product ot inertia of an area,

A3.10

CENTROIDS, CENTER OF GRAVITY,

MOMENTS OF

INERTIA

1r
_-+-x
y
.l3.1l

By adding equations (3) and (4), ~e obtain I x1.+ I Y1. = Jx + I y- - - - - - - - - - - - - - - - - (.5 ) or the sum of the ~oments o~ inertia of ar. ~rea with respect to all pairs of rectangular axes, thru a common pOint of intersection, ~s c~nstant.
13.12 Location of Axes tor which Product of Inertia is Zero.

In Fig. A3.ll
I X1. Y :l, =!X1Yl dA=!(xcOs~+:;sin,2j)(ycos.z;

xsln0)dA

= (cos s 0 - atnFig. A3.10

~ /

xyG.A'" cos Zl sin 0

(y:a..x a ) d.A

Moments of Inertia with Respect of Inclined

Axes

=Ixy

cos 2

+ ~

(Ix - I y) srn 2 "

Therefore, IX:l,Yl is zero when unsymmetrical beam sections are very comtan 2 0 = ~ - - - - - - - - - - - -(6) mon in aircraft structure, because the airfoil Iy_I X shape is generally unsymmetrical. Thus, the general procedure with such sections is to first find the moment at inertia about some set of A3.13 Princ1pal Axes. rectangular axes and then transfer to other inI~ problems involving ~s~etr~cal bendl~g, clined axes. Thus, in Fig. A3.11 the ~oment at inertia of the area with respect to axis X~X~ is, the moment of an area is frequently used with respect to a certain axis called the principal Ix , = Iy ,, dA = I(y cos 0 - x sin 0)'dA axis. A principal axis of an area 15 an axis about which the ~oment of In~rtia of the a~ea is cosa0! ydA+sln al!x adA-2sin0 cos .0 either greater or less than for any other axis I xydA paSSing thru the centroid of the area. Axes for which the product of inertia 1s Ix cos' 0 + I y sin' 0 - 2 Ixy sin 0 cos 0 (3) zero are principal axes. and likewise in a s1m11ar manner, the following Since the product of inertia is zero about equation can be derived: symmetrical axes, it follows that 3~etrical axes are principal axes. I y1. = Ix sin a 0 + I y cca- + 2; Ixy sin cos The angle between a set of rectan~~lar 0-----(4) centroidal axes and the principal axes 15 given by equation (6).
y

Example Problem 4. Determine the moment of inertia of the angle as shown in Fig. A3.12 about the prinCi?al axes passing through the centroid. Solution: Reference axes X and Yare assumed as shown in Fig. A3.12 and the moment of inertia is first calculated about these axes. Table 8 gives the calculations. The angle is divided into the ~NO portions (1) and (2).

x:_-I,.-

='*"'""'O----''--Ir..-_ x

Fig. A3.11

..
Port
1
A~'

,
'ox
Ax_
,03~1

'0,
,

r
,1" 1.205

A,
,0469

Ax
,281 .06205 .J.oli35

A"
,OO~8

Ax'
,211 ,0078 .2188

'x
,0077 ,9470 ,95.1

Iy

.3705 .500

."
,1205

fi:z

1.5U~3 e

,625 ,6 19

,7800 ,7858

.0781 ,1132

1 1 3 12 :z,,:Z2

0019

1 Jl:1xl 53 070 12 4 '

,281 ,010 .291

,167

~2X2J:.2S3 0026

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A3 11

-lex

of inertia of each portion about their own X and Y centroidar axes. Location at centroidal axes:~nd

ley

~ ~oment

- = l:Ax = .3435 = .392" x

l:T
~ransfer ~oment

= ZAy = .6719 = .767" ~ l:T .875

of inertia and product of inertia from reference X and Y axes to parallel centroidal axes:-

section has been broken down into 16 stringers as listed in co lumn 1. For the top surface, a width of 30 thic~~esses of the .032 skin is assumed to act Nith the stringers and a Nidth of 25 thicknesses or the .04 skin (see Col. 3). On the lower surface, the skin half Nay to adjacent stringers is assumed acting with each stringer, or the entire skin is effective. Col~ 4 gives the combined area of each stringer unit and is considered as concentrated at the centroid of the stringer and effective skin. All distances, x and y, columns 5 and 8, have been scaled from a large draWing.

-, -I y = I y - Ja. ~ .291- .875 x .392

1XY ~
ZAxy - AJ:.y ~ - .150

Ix = Ix-J.jg = .955- .875x .767:1 = .440

2

= .157

.1132 - .B75 x .767 x .392 =

Calculate angle between centroidal X and Y axes and principal axes through centroldal as follows: tan 2 0~2 Ix" ~2(-.150

xp
_X x

:7 -::1 _

F,r.
......2 3

"i

I
I

Yp
.040 Skin 6'

jef.

5,

Ti:!--f~

I y _ Ix

=-.30 =1.06 .157- .440 - .283

~~I
Fig. A3.13

-12 13
.r

! ,
I

.---=-- - ---L... -,-. x

0

XRe f. x

~4

15

I~

2 0

~ 46 0 -

40'

= 23 0

20'

Calculate moments or inertia about centroidal principal axes as follows:-

Yp Y

I.

I
U

1xp=1 x cos" .0+1'1 Sin2121-2Ixy s m c cos a = .44 x .918 2 + .157 X .3965 2 - 2( -.150) x
.3965
-

,
"

a
Area

,
I
<

,
"

Table 9

x .918

504

m'

I Y1l

Ix sin 2 .0 + -I y cos 2 0 + 2I xy sin 0 cos el 2 ~ .44x .3965+ .157x .91a + 2( - .150 ) x .3965 x .918= .092 tn ;"

, " .U

, .. :, .. ., '" I. .... :::' "

"
~f.
"<

,
4.00 6.05 7.00 7.37 .>0 1.50 1.30 6.90 6.50

",

o~

.. . '

"

lx,
~.,

"<

T'
I

q:-:'!-

a .U a .30

I-L -

,"'~i
I ,

~t 2f ,\1' ';""-" , ~-~;

y :>

fL

'+-,( ;,---I

...L

.U -a.nc ." ." ." ." .U ." ." ra .U ." .n .n ." .n .n .n " .n ." ie " a. .30 .30
.J< ata s
-4.90 _5.911 _1.40 _8.13 _8.62 -8.81

a .ra s .J< io U .ur ra .La

." , :t~ r .ra

03 .03 .011

1.14
I

o-t
.0< .0< 0<
.~

.14 .38 17

.,

.U .U .U

0.560 2.240 -J3.1.5 0.8471 .5.124 _29. 28j 2.660 L8.620 _24.8.5 1.253 9.234 -21.18 9.68 I-~~' 60 ~.288 1. 215 9.563 -12.60 1.241 9.0:19 8.60 1.111 8. ceo _ 4.00 1. 885 12.252 - 0.35 _0.561 1. 8.51 _33.2:1 _0.83 4.0112 _29.28 _1.6661 9.913 -24.8.5 _2.2941 16.916 _LS.10 _2. S2~ 20.481 _12.42 _2.612 22.432 _ 6.10 -3.094 21.444 -00.35

...

.0<

-,

_4.6041 153.85 1-18.56 _1. 099 120.02 _24.80 _9.443 234.66 _66. 10 _3.600 18.205 _26.53 _2.8:: _2.142 _1.462 12.51 _10.61 _0.680 2.12 _ 4.69 _0.100 0.04 -00.65 _5.6:13 181.96 111.611 _4.918 1411.16 24.39 _6.9581112.90 41. 40 _1l.191 108.40 42.90 _2.850 4.1.82 31.30 _1. 891 11.54 16. ;10 _0.122 O. il4 1. 08

~:::: :i~: ~~

.,

a. s

Location of centroldal axes with respect to ref. axes,_y='Z.Ay=- 1.465 = - .396" 3.70 l:T
X=,ZAx = - 58.238 - -

Fig. A3.12

Problem 5. Fig. A3.l3 shows a typical distributed flange - 2 cell - wing bean section. The upper' and lower surface is stif:enec by Z and bulb angle sections. _Determine the ~oment of inertia ~f the sect:on aDout the principal axes.
~~~ple

15.74"

l:T

3.70

I
I

Ix = 187.04- 3.70x .396:1

=186.5
2=

in

SoLut t on :

The properties of the cross-sectlon depend upon the effective ~ateria1 Nhich can develop resis~ing axial stresses. The question of effective ~aterial 1s taken up in later c~~pter. Table 9 shows the calculations for the moment of inert~a about the ass~ed rectan~Jlar reference axes XX ~nd YY (see Fig. A3.13). The cross-

1y = 1348.36-3.70x 15.74 .431.7 rn,' I XY = -13.35 - (3.70 x .396 x 15.74) =36.41 In ,

tan 2 0=2

Iy-Ix
20=16-32.5',

I xy

= 2 (-36.41) =-.29696

431.7-186.5
.0=8-16.25'

1xp=Ix cos" .0+ 1y stn" ~-2 1XY sm e cos e = 186.46 x .9896 2 . 4 3 1 . 7 x .1438 8 -2 E36.41X .9896 x (-.1438)J = 181.2 10.-

A3 12

CENTROIDS ,

CENTER

OF

GRAVITY

MOMENTS

OF

INERTIA

I y p = Ix
2

s In e 0+ 'y c os e el+2 I xy s t n 0 cos 0= 186.46 X .:~822 + :;)1.7 X .2695 2 +

e-

A3.14 Section Properties of Typical Aircraft Structural

Sections.
'0 ; Table A3. :0 'thr-ough A3.15 and Char-t ...... si'le the section pr-oper t t es of a few st:-uct.. . ral shapes G8rrnOn to arrcra rt . Use of tr-ees tab Ie s will 08 :TIade in later cha;Jters of th:!.s jook.
v.~

-36.41 x ,9::95 x (-,1"32)

=437

in ...

,
'I =T ,il ---.-.z~1 I R -'-T

,
n"TI
X
,~

f--e-J

x-'

,
I

'i
,::,/-R

AS<T
-T

TaDI ..1.3. 10 PfDlI"rlies of zee Secl'Q/\S

T
X

B I

'I

r+

~T
X

~a--

J ..J,_ __Ai2==:jI
I I

V-R
r

I
I
y

R--) . ) -T ---..'A

Table ..1.3.11 Properties ot Channel ~ectlons

I
SECTION ELEMENTS

"
Z
U

NOMINAL DIMENSIONS

" " "

I
IDcMa

R
125

AREA
Square
Inches

I '" 'nInches
0035 .00<12 00" , 0071 0070 0084 .0109 .0090 Olla .010

Puc

I tOyy
rn '"

Inches 203 .197 255

=t: -+
j+ HI i1

. '00
.625

.eas
.750 .625

050 06J .050 .063

--+

1.000 . 625 .750 .750 1. 000 .S2S

'"

.063 .050

". on ". ", ". ,,, ,,, so

125 125 115
0

" " "

U

"
Z

NOMINAL DIMENSIONS
A

.">REA

B
InChes 375 500

>-

,
Inch 106 153 204 .207 JI5 .327 0995 .146 196 201 245 .247 ~53 362 .319 1~0 .H5 187 193 .229

:>ECTlON ELEMENT:>

:>q..are InChU

'n
0009 0020 003, U045

Iyv

I '"

"0'<

I ~v,'

Incnes' .0060 0075 0090 .010S .0!-l7 J193 00 7

I
I

Inche. .111 195 190

275

0852 0977 1231

00<.
.0062 0122 0146 0069 <lOS3

,I><, 0 o

"" 20'

353 .346 268 336

1-+ I-}

cia

00,

Hl--

r*"
18

~ "

'" .cso
.050

.125 1~5 .094 125

.750 .87S .875 1.000 .625

H-Hl-f--#I~

.078 .050

"', '"

,,, ".
125

a ." ,

..

131 .163 .0977 123 .110

.125

ra

us

, 1~5

".

.125

tl ;::ll::
f--#as

~
~
~

.'w
1.000
.875

--mceo '"
.018

~ .076

".

IT 104 131 111 .147 .177 .153 197 178 .216

.ras
125

'"

.0165 0130 0158 01 1 .0164 0214 0172 ,0245 023 .0276

0148 .0367 .00l>S .0082 .0122 0148 .0367 00"

259 .253
;0

.298

'" '"

.no

ce

.caro

.0122 O14B .0175 0197 ,0241 .0287

.U
139

.CO
.0439 00'

~ J78
~ ~

.,

"

1.00<

sas
,875 1.115
1. 12.5

~ 076

".
.050
0<

.125

., ., .,

.187 .170 .206

..

.0251 .02S4 .0285 .0335 .0363

"

oe

"* "* ... * ... t

;)17 .314
365

304 .312 35J 347 .360 254 .348 .359 .353

.n

259 ~ 3

'" '" m ".

aaa

H: HH~

.750

875 315 .500 .625 .875 .750

I. 000

'"

050 0," 063

'"

".
' 125 .094

.0127 0852 0 17

,n

'"

~ 094
.050 .063

1-+

",
,

.aao
0790 .09\6 115 .131 158 .11. .147

155

.oi rr

296

pt

~ .076

.125
'0 4 125

'"

+64 251 251

Ri: te

i-#.-g.

--*l.or
'~ us

.17

aie

.328 .139

.500

4*

.H

.358 .400

ao

.0147 .0175 .0242 .0287 .0332 .036

0

.396 ~1J9 ~03 .397 .415

'" ." " , ." '" , " ae

.385 .282 .244 306 373

--li'-+.\-

~.
I. 000
.~.

.875
1. 250

s.t ""*" "

.150
1.
2~O

.. . =ll!:

'" ~
~
~ .094

125

".

.226 142 .176 .216

'"

0265 , 0304 .0374 .0441 .0467 .0632 (1421 .04 7

u~32

'"

0083 0197 .0241 .0287

..

sea

.444 237 374 .368 ,364

H\!-+.\aa
as

ae -*

, ~ -1E 094

~
~ ~

.125

155 .137 220 244

". '"

".

m 243 es
294

0159 0009 .0020 0025 ,}047 0055 0065 0079 .0094 0~13 .03ll4 OG26 0029 0049 0058 0083 0096 0113
.0123
~152

,;;
106 .148 144 1ST 2JT
;0

.csro
, 0132 015" 0183 0151 .OIS4 0214 (;276 0376
,n61

,
,

'"" '''' ,,.

~96

290

347
J~9

'" 347 , aaa '" ", ;)58 .14

HI 0

.0211 0252 .0250 0294 0335 .0285 .0335 OJ83

'"

.304 .;H HI 187 231

aaa

2.
131 .158 141 .171

'00
1.12..5

.441 464

.125

.445

LOOO
I.

.125

'W

:jf

* =1t

1. 375

1.000
I. 250

..... ::$.:
~ us
.094 .094

".

." ., " ". ,

.303 .202 .291 .338

.210

".

.0 30 .0637 .0T44 ,0148 .01.

, .72
464 457

." ."
514 501 519 512 543 .562

.503 500 .96

, ,

* :j
--11~ ,..g.

."
.750 .875

~ ~ .078 I : ~94
~ ~ ~ 125
.125

.125

'" ,

." m
134 139 .178

'00

,,.

2n
181 .216 .256 .328 236 279 .359

.0729 0765 0886

0

0511 .0887 1039

'"

558 555

.125

=!t " :::g: -*" =H: =t

I. 500

I. 750

~ I. 2 0 .094 , ,c
000

"

...
,53

210 25~ .103

0662 07 4 .0911 1365 .07 5 .0961 .111 .134 11 4 1377 1925

.12.

aae

,, " . ., s
.0511 1308 0439 0511 .1040

..

.0367

"

.549

.15 .HI

'" ". '"

o.

'sae "

"* " "* Hi-* "* 4-fr

232 .237 .278 .283 .289 238 341 508 SI9

L 125 L 000
1. 375

.,. Pi!:
I .

.125

ase
asa

01,6 041 061J7 02 .0031 805l 0060 01(12 .0117 .0135 0156 .0183 0225 0232 0269

" ,0643
0281 nB5 03:H
,

, ts " '" ,
zas
'159

'"

304 377 394 .02 392 .390

~03

'"

397 .417

087 HO ;86
lH

e9a

"

"

-122

H'

sa
,875 1.250
I. 125
1. 500

~~5. ,.

';0
.53 220 .267 .344 .265 .314

~ .125
~
12
125

=i so
1. 375

,,
csa

125

,500

."

.278 288 377 .383 .560 169 115 .263

~68

.115

265 .114 .284 365 361 408 ;94 J23 385

,,
'" ,
"

L .00 L 'SO L .00 L L 0

".

ras
.125

549 .:; 2 .602 .596 .0\1

"

.525 8'15 1.125

L

~ .094
~

."
,
a

.537 ,407 .403

.0439 .1039 .0511 1040 2346 _0288 0331 0512

.096 707 779

dOl

393 .520

H} ss

~
"

~ 125 ' ~ ~ ;56

125

2.000

'W
m

2,500

1.000

". ", ~ '" ~

"

2194 2621 3852

"

ate
505 .029 ,298

f:i!: ~

\.;00

,.

I2S

00.

.2890 .1383 .3723 .4707

"

sas sae ~ ns avs ~

.,

..
.125

.197 .232 .2.16 279 .359 326 ~22 .513 .244

278 J71 31l'1 .390 :66

."
-a

m
.55 338

1. 125

'" .".

.269 278 380

.0444 oa .2,. 0319 0441 27\ ~52 .0507 :68 .. 5 sea .33 '-1616 .0495 .214 I 4Je .310 .0569 .52 0695 303 m ,66 0661 0757 .0832 I 0930 .53 -123 0071 0493 .0190 .0650 ~57 .0234 'J79 .51 481 355 0698 5n 0334 ,07 .0389 0607 l52 'J484 (>994 .345 -ss 1292 .70 1106 J08 .00 3 .179 0540 171 .00B i)621 517 0169 0705 547 ~68 .0198 543 :66 0818 .0242 S2S .1002 ' 359 .0403 1010 I .352 557 .0504 1246 I .346 SH 059\ .339 .1451 .532 , \710 ,52 2035 .0076 .J76 .177 ;0 I :7 0090 0939 HB ;H .0249 1236 258 .e rz 0289 1485 , 257 0414 350 004 1233

_3 225

,"
, '"
,

H '"

,""

, '"

"

,2,

"00

.955 .941

" "" '" "

.354

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A3 13

11"'7Y

~,~ w , '
I

Table A3. 12 Properties of Extruded. Aluminum Alloy Equal Leg Angles. (Ref. 1938 Alcoa Handbook) Axis XX or YY d I P 0.004 0.004 0.006 0.008 0.012 0.016 0.021 0.029 0.033 0.042 0.059 0.074 0.058 0.074 0.107 0.135 0.096 0.121 0.174 0.223 0.18 0.27 0.34 0.41 0.183 0.220 0.219 0.217 0.311 0.301 0.298 0.293 0.38 0.37 0.37 0.36 0.46 0.45 0.45 0.44 0.55 0.53 0.53 0.52 0.61 0,61 0.60 0.60
,.:

j-i---t Y

----:Ld ~ z Dimensions

'~-Xit a! ---L

Area
R
~.ID.

Axis ZZ P 0.111 O. 142 0.141 0.141 0.199 O. 19:J 0.191 0.192 0.24 0.24 0.24 0.24 0.30 0.29 0.29 0.29 0.35 0.34 0.34 0.34 0.40 0,39 0.39 0.39

3/.

'I'

1 1 1 1

3/' 3/'

3/32 1/16 3/32 1/16 3/32

II'

1/' 1/' 1/'

1/16

II'

II'
II'
3/32 3/16 3/16 3/16 3/16 3/16 3/16 3/16 3/32 3/16 3/16 3/16 1/4

.111 .089 132 .171 .122 178

1/'
3/16 3/32 II' 3/16

, a
3

1~ 1/4 1-1/4 1_1/4 141/4 1_1/2 1-1/2 1-1/2 1-1/2 1_3/4 1-3/4 1_3/4 1-3/4

II'

.23'
339 .230

.30 .43
.56

1/.
3/32 3/16 II' 3/32 1/' 3/16

II'

.28 .36 .53

.59 .32
.42

1/.
3/16 1/4 5/16

.62 .81

II'

II'

1/'
1/4

.,.
1. 16

.49

.n

0.181 .0015 0.199 .0018 0.214 .0026 0.227 .0034 0.211 .0048 0.276 .0066 0.290 .0085 0.314 0124 0.34 .014 0.35 .017 0.37 .025 0.40 .032 0.40 .024 0.41 .031 0.44 .044 0.46 . 057 0.47 .039 0.47 I .050 0.50 .072 0.52 .093 0.53 .08 0.56 .11 0.58 .14 0.61 .17

", ", '" ", ",

'" ", ,,0

, '" '" ",

T
I
~:

,.y
T..ble A3.14 I'rO,",rlies 01 UnequaL Angles

~R

,x

1.-

"
"
0

NOMINAL DIMENSIONS
A

a
Incbe.s 500 625

Y
.050

A~~ R
'------Square Inchel 0535 0818
AREA

SECTION EL.MENTS

i
.126 162 150

625 .751l 875

'" '" H= ", '" ", '" q '" '" ", '"
'45' " 45'

f-t

.oea
.083

750 .625

~
~ ~ 125 ~ 094 ~
~ .125

""

"'''
187 222

lJIChell

.OOIl

.US

at f#
ta ~

1.000

750 875 .750

.US

... ."
.108

aa

.100 .184 .<0

211 I . l40

'" .,

250
272 .320

.178

.n

f# filia

1.250 1.000 .750

1.

.125

.....

fi elf
c#
~

~ ra
~ .125

sec

'.300
1.250
1. 00

-:.m 156

.,.
.228 156

167 .124 .179 .231 .139 .202 .262

."

322
i

.. '"
"I

'"

0027 0029 0049 .00 2

'" "'xx i "''' ., " "''' " ., ,

.H6 01144 0069 0073 .0123

t"yy

.,

.0050 00:'1 008

,DUO .0054 ,0074 .0092 .0124 .0\71 0215 .0077 .0096 . 0182 .0226

231 159 .170 .181 .239 .250

ro,

292

"

.416 .427 373 384

.361
.251 J27 .401l

'"

'~W
I ,
I
I , ,
C z

ti&,

~ ~ 156
.125 .125 ,

.15 .167 .228 239 . 249 .310 .321 .331 .300 .383

es

aa

.534 .472 48~ .433

... ....
.455 .591

1.250 1.750
I. 500

~.m
.156
y
,

3~8 .389 .478

asa

.506 517

'"

.0265 .0347 .0436 .0518 .02 6 0465 .0775 0928

.017 .011)6 .01 2 .0111 .0155 .0197 .0278 .0346 .0217 .0306 0383 .0463 .0584 .0512 .0647 .0712 .0550 .0698 .0834 1078 .1144 1377

, ., ., .,. ,. na
.211

221

"

274 270 .258

.1

217

.,

, .291

2 2 ' 310 257 305 .209 1 .400 .203 ' .393 00 .381
~,

.194 .l91 .465 .282 474 .276 ~68 .271 , csa .312 .365 .462 .360

'"

.389 382

".
'"

"

""

',"

360 ' ~47 ' .542 1 I

". '"

'" s

n
l"Y
T ..ble A3.I3 PrOpertin of

I I

' CJ l-' .:.

,_A~oT
~

-r:7
~

Tee~6eC<lOtl3

-11 X i ~ik-R C ...L

A T
!

~0il~
~!a-j ,

--'-,
'r
X

Table A3.1S I'ro,",rllell of 1.8e<:tionll

6~
,

'----"
Y I

"T

SOMINAL DIMENSIONS

Y I
050 .063

AREA 5q1JaI"e Inchell 094 l27

,
Inch

SECTION ELEMENTS
,~

-+ '" R=j
~1.
r*" H!f.li.-

"

Inches 625 7sa

0

[nches 0028 0055 .0046 .Ol31

",
or

"'V<
173 209 .209 .296 292 .290 379 .192 198 286 371 .371

Ip'IY

Inches 287

, eo
L

,125

.0'
.063

". ,
ias

375

r. 00'
L 2~0

'00

on

bO

!-t: $:1. ,
.

,
;00

,;0

'J94 094 .063

, ne '"

.H

, '"
aas

" "
W 0

NOMINAL DIMENSIONS
A

a
incheS

R
m

SECTION ELEMENTS
AR<A
~

Squate Inches 250 289

1JIches"' 0206 on9 0284 0~2

", 'u '"

Inches . 118 .310 .314
0

0011 .0098 .0129 0158 O1B8 .0 ISS .0169 0244 0200 .0245
'

.228
,

,
,

0<,

240 204

750 ,078 I. 000 t. 000 'J94 t , 250 I. 250 ; .094

I.l~

".
. 125

.lH

.228

'" ua '" no
a
2,5

, ,"

.252

120

" azs
I. l25

.ca tea
.212

I ~L
; 23 24 ,

Ht,

I. 62~

L 125 1. J75 L. 3, L.1 ,

'

, ,

l25 .878 .094 l25

"

'"

.67>

, , ,

750

,
,

I L 125

" 7'"

L L 315 , L .175

ias

", 125 894 I ,

.n,
240 .31l

.0182 .0345 .0052 0080 .225 ,0 60 .232 I J186 0301 Q354 .3H n, .0449 .180 II oJI .0228 .254 314 ,; 39 .047; .3H .0 0, a

,:H2 .327 .141

-" 294 ~ ec '"

321 .328 312
't

ie ia

, ,H=

.750

L '00

'"

~ 078
07

125

'"
,

ats
I. 000

I. 375 L eas L 370

~
094 .094 .H .094

.0245 J2ij 0407 .0263

aea

"
tzra

"

H=,
II
!

aa ,
.341 .388 HI H

-,

0193 .0235 OHI 0375

;

.307 .318 338

'"
-

l. 625
I. ,,0 l. J " l. 02 L B7:;

"

125

'"

,
,

ate
H'

I::tJ ~l.
13

250

,)94

0676 0725 .11 1285 1442 .b 2268

.0375 0626 .078

125

",
458 50S

a cI
H

'" .u '" " '"

HI

516

.431 .321

2. 125

234

o~oa
0314
,)40~

na

,,
,-

".

1,8

,os

,,25 L 625 625

, ,
,

156 ')9. 12:; l56

.45 .310 405

'" H" -,
'"

. 17 .191 252. 333 .3,8

013l .01M 0273 .0412 04 4

234 sa , '" 230 0512 0"; J393 062 0393 0627 'J394 )5l3 G6 9

322 H5 .Hl

.338
40~

" ". Ht, " sa '" ,,""*

rsa

~
~'
"

.oec
.125

.0911 .1023 1543

305~

:' : 1.500 [7 2.000 :!.500 3.000

I. :50 2. COO 2. 500

ia us

'OS
l25 .156 156 125 .188
I

.5H 781 719

"

.,H
0

" 530

as

4~0 531

a ,

2.000 2.000
2.
~OO

Hs~
\ira

,,'
'" 740

3106 3922 .4946 6571

.61; .025 636

H;
.65 626 ~63 464
HI

"

..,44

1183

.789 ~51

an

'"

on:
.0995 1200

4.14

'"

,,"

, ,

.322 .314 I .167 HI , .370 371 .03 .500 J56

""
se
ss

,,"

'--"-' , aa

I. SOO 2. 000 4.000 3.

f--iIT- ,

"

.1104 1546 II 4

'" , ,.
L L L

';0

ua

, a

L 262

, a

coo

'" "

r--#-

3.500 . 000

"" Hm
J 13

l. 52 2.00 2.60 3. ~3

".

I. l88 2, ~24

4. 136 5. 521 368 9. U5

a. 037

L a t , 63d

l5~1 ' ;H1 1184 L l547 i t. se S5d6 I L ;;

,. "

z aa

<0' '0 , iOO

.aaa
550 794 .794

."
I,

, '"

, ss ,,
L

ea "

A3.14

CENTROIDS,

CENTER OF GRAVITY, MOMENTS OF INERTIA

Table A3. 16
SXC'1'IOI'f PROPERTIES OF TYPICAL .URCRAFT RX'I'RUDiD SEC1'IOI'fS

I I
,
"

i xo.
1 2

'Sect.

,
!
,

A
17/32

Dimensions

I
1

I c
I

't

'1

'2

ATOll
!

'rop.,ubM
I

x-x'
e
I

in.

IlI:X ,

e lI:X

AlI:1s

src rU bou~
AlI:l.S

I rY

e rr

e'

13/16 I 1/'
, ,

,040 ,050 1/16 .060 ,050

,050 ,050 1/16 I .060 ,078

1/16 .050 3/32 3/32 3/32 1/16 9/64 I 1/16 3/32 ; 3/32

.068 .113

.00686 ' .317 .01515

,

.319 ,510
,

.00150 ; ,OO15~ ,00162

.148 .117 .125

,112 ,110 ,116 ,130

,

- ...C;..-..-

liZ
lI2

7/32 3/16 3/16 3/16 3/16 9/32

,3665

..l.

~::-[: _e_
'T.

7/'
7/. 7/'

.1025 i .0100 .1031 , .0102

,

,313 ,313

.383 ,383

-,

,
,

9/16 9/16
3/4 3/4

.00215 I .114
!

'-- A----

-*

I ,

.1061

.0120

.329

.345 .435 .574

.00302

.165 ,190 .181 .108 .104 .147 ,206

,133

"

e
,

I
,

1-l/8 1_5/16
3/8 3/'

.0751.075 3/32 1/16 . 3/32 13/32 3/32 3/32 1/16 3/32 I 3/32 3/32 3/16 I 1/8 3/16 1/16 I 1/16 3/32

.1514 I .0239

I
,

.397 ,471 .326 .314

I .00550
,

: .182
.171
, ,
,

-r-F
~

,
II

I .230 I
i

.0507 ,0101 .0136 .0240

.0080

7/8 7/8 1

.095 .1375

,4315: .00111
,

,108 .118 .158 I ,210

I
,

.4375: .00148 .500 .625 .906 1.25 1.125 .133 ,133 ,121 .201 .300 .217 .3::18 .255 .658 .757 .959 .793 ,500 .562
.75
,

. '-'l-'
I 1'0 _L-L., ,

! 10
12 13

I ,

lI2

i .1736
."4

; .37l
I .480

.00376 .00992 : .0555 .0662 I ,

I .0812
,

I
,

1-1/4

,_'t\

I
11

11/16
,

3/32 I ,2320; .0536 3/32

,

l
r

-. 4
C -

1_13/16 11

3/1' 1/.

.2920 .525

I
,

.677

I 2_1/2

14
15

: 2_1/4
1_3/8

1.1/8

,
1/.

illS
' 1/8
!

.581

! .950

I , .338 ,
.362
~

.295

.326 .323 .375

,

l/S
1/16
1/16

.666 ' .524

i
,

.888 .176

,ee ir+-r- ...-/C

I~ri~

~--" ;-~ -;

rr6 .L.
'C

11/16
3/4 3/4

1/16 : .131
!

.00425 .00580 .00589

1 ,

.0178 .0335

,360 .446

I .687 ,

IS

1_5/8
1~5/8

I
1

5/32 5/32
,

1/16

! .168
.306 .359

.186 ,170 .287

.364

17

i 1/16
! 3/32
1/8 3/32 5/32
1/16

5/64 I .206 3/32

I
,

I .0458
i

IS
18 2.

I
,

2 2

1- 1/ 1 6 1_1/4 1-5/32
1-1/2

3/1'

3/16
1/4

! 3/32 I 3/32

i .474 i
I .509
1

I .812
.812 1 0
I
,

: .0252 .0475
!

,0191

I
i

' 3/32
1/'

: ,0821
i

i .479
.590 .326 .302 .281 .282 .514
.348

,1. 0

2
2_5/16

I 1/8
1/16 3/32
1/.

~ .378
.668

,0291 .1075 .0186 .1562 .315

: .279
, , ,

.1095 .2326

i .541 i1. 0
11.156 ,

21

: 5/16
,

I 5/32 !
I

' 5/32

.401 .342 .641

,. Tf"'-f~T 'a I ; as I i.,.e.'J ,

!

j"-:-J:'

22

I
,

1_1/2 1-1/2

i 1-l/8
2
1 1

; 1/16 , ,161
i

; ,0171

23

I
,

l/S
9/32
1/4

I
I
I

1/8

.384

I
,

,0346

I 1_3/4
1-3/4 I ,

2
2_9/16

; 1/8 ; 5(32 5/32

.760 .918 .605 .135 .180 .166 .214

I
,

I .643
I .809

I ,0599
,

.5/32
1/.

.600

.0730 .159 .0163 .0295 .0151 .0281

, ,875 , .815
i 1. 00

I i

.750 ,750
,

I I
i

26

!2
I

i 2_1/2
1
1 1_ 1/

!
l/S
!

5/32 .050

, ,
i

! .385
i

; .800

I' e-T 8--'I..LI

,..A-! i--r~

27

11/16 13/16

.050 ; .050 .050

3/32
3/32

.0218 .0315 .0603 .1382

.14.49

, .404
, .793 ,
i
.795 : :

6625! .78251 .600 ,725 .719 .719 .531 .469 .531 I .525 .655 ,787
1 i

28
28 3. 31

3/16

.060: .060 .050 .050 1/16

1 1/ 1 6

I ,
,

!
i ,

5/S
3/4 3/4 3/4 '/1' 1/2

1~~/2

13/16

1/16

1/16 ; 3/32 3/32

, .595
i .773
I .349
i

.456

.405 .298 .358

.341

I 'r' '.,-J
~
,L

2 2 ,2 : 7/8 7/8

! 1/4
i
1/4

1/16 : 1/16 1/16 : 1/16 3/32

1.00 1.00 1.00 .437 .437 .500

, ,

'I

3/32
3/32 ,1/16 1/16 ' 3/32

.2378 .2855

i
!

; .0282

't-e'..I

132

11/4

! 3/32

.1820 .0120

.oaes
.0056

.356 .238 .197 .234

...
:

: 33 I
'--A ;:I
I 34
35

.050
1/16

.050 i 1/16 .080

.075

.099 .115

.~,

1.,,10-

"

9/16
9/16

I I1
1

1/16

I i
,

.151

38

,1 1/4

.ove
1/18

3/321.170

.0135 ! .341 , , .0239 , , .397 , .475 .0384

: ,

I .00446
I

,00828 .00876 .0194 .0391

L'"
e

a ...-

I I

.625 .625 .750

I I I

I .227
.302 .360 .267 .265

-~

31 38
3. 4.

LII~JI

, ,
,

11/16

1 _ 1/ 4 1.1/2
2 2

.100 .120 I

3/32 : .213 : 05~0 3/32 : .302

1/'
i

I .508

.825
3/4 3/4

.075 .072

.110
.1446 .1826

'

i .012

.102i l 0 2

, ,

: .249 .330

, , .757
.750

.604
i

1.00'

3/32

1.00

I .0180 I .0228

.714

.700

CHAHT A3. 1

PHOPEHTIES OF FOHMED CHANNEL SECTIONS

0.35

.1
~
~

1 I 11 P:llili/lli il1mHW Ii!1~IWilf 'IWi' illi J!:..I.IIIJ I ;~I Imll! LIIIIII :'. Iii J 111m II Ii !II iiiill'11~1.ljjj1!l!!I!lIljjllHmmnn ~
jli

1.0

.. ..
~

w
.1

JIi I'

.,
.,
..

11:1Mo.3O .. ! II
1

,:1: i I 1,

.,
.'_, i:
'+,

II
I .:: I T1

...
~
~

,. ,. rZ
~

i: 'i~

:'1'

, . , I'I
;':. :1

::1.

"" ,i 1

0.25

F.itt0'l!I: i1ii !!II:1 !l~i :II!: ill~~' ;:11 ~ -1" I-!- I~; Iiii1: ill:i : IilliIiII: 1111~1 illIlllllli! ~ '1II: i III Ll!: I!il;:~

,
~

,. I
Z tl
tl
~

'"
~

030

i
0
~

~
0

z " .,
0

Ii
0,15

~
~

Ii w
0
~
~

" .,
r-

.,

'"
'"
~

<:

'" r0
0.10

1";'I'~!I!"!Im!I'!'T"!lIIWI'III'II'I~j0P;IPI'':'IU;;ni'W!UUwnUI'~p!!1 till r._. 1m .. , .m itT. !I. I:-p. Itt! ....... " ", ." .... ','.". "r. " I " . . . " "" . " . "

I~

.,
0.05

'" ., '0c"t
~

1II111l1\l!llHllllllllml!l!!Yf111 IY!!'III1i~:iIIlIJl!lIiHIM1'I!l"J!ml!!! 'III ::1 !I,'I II! II i!!!l1i 111:11 ',IIII! I: !I .,

I~

"

Mom. ollnarUa.:-

1.0

1.0

2.0

2.5

3.0

I
.025

~.

'.

.'-c.

II - IIEIGIiT OF CHANNEL

,
II; HEIGIIT OF CHANNEL

3.5

I;

V~

\1 3

A3.16 A3.15 Problems

CENTROIDS,

CENTER OF GRAVITY,

MOMENTS OF INERTIA
~3.1S,
t~e ;r~~ t~e

(5) ~or the ~ean Beetien in ~i~. calculate the ~oment oi inert~a ~bo~t
ci~al ~xes ~ssu~i~g :~e

fJUY

st~i~;~rs ~s

only effective

~aterial.

-ll-/., l:~;; Ir.. 'i)t.

,.. ,
Fig. A3.14

Ii I-- z" -.J

J
"

/0_

4-... IS I. , .,
l.'

Il"

8 ..;...

10

' I

...+
i

08

Fig. A3.15

~~t

I'.T L I zo'

1'- t

o.?

Od~
15

Fig. A3.19

1"8 i

(1) For the section of Fig. A3.14 determine the moment of inertia about each of the or-merpal axes Xp and Zp. (2) Calculate the moment of inertia of the section in Fig. A3.l5 about the ?rincipal axes.

(6) F~~. A3.19 shows a wing 0e~~ section with a cut-out on the lcwe~ surface. Jeter~ine the ~oments of inertia about the ~r1~:i~~1 axes assuming the eight stringers are the cn:y ef~2c t ive :naterial.

It ~ l : D~
I
5 spaces @5'" 35"

F,g. A3.20

Fig. A3. 17

(3) Fig. A3.l6 illustrates a box type beam section with six longitudirral stringers. Determine the ~oment of inertia of the beam section about the principal axes tor the following assumptions:(a) Assume the beam is bending upward putting the top portion in compression and the lower portion in tension. Therefore, neglect sheet on the top side since it has very little resistance to compressive stresses. The sheet on the bottom side is effective since it is in tension. For simplicity neglect the ver~ical webs in the calculations. (b) Reverse the conditions in (a) thus placing top side in tension and lower side in compression.
(4) For the three stringer single cell boX beam sectien in Fig. A3.17, calculate the moments of inertia about the principal axes. Assume all 'Neb or wall material ineffective.

(7) Fi~. A3.20 shews a 3 cell ~ulti,le flan?e -beam. !he 7 flan~e T.embers on t~e upper face of beams have ~~ area of .3 sq. in. each and those on the bottom skin 0.2 sq. in. each. The bottom skin is .03 inches in thickness. Compute the moments of inertia about the principle axes assuming that the flange :nembers and the bottom skin comprise tr.e effecti~e material.

\_Cutout [or door

Fig. A3.21

..
(

Fig. A3.18

r---

/9 '

(8) Fig. A3.21 shows the cross-section of a small fuselage. T~e dashed line represe~ts a cut-out in the structure due to a Goor. Assume each of the 13 stringers have ~~ area of 0.1 sq. in. Consider fuselage skin ineffec::ve. Calculate the :noment of inertia of the effective section about the principal ~es.

CHAPTER A4

GENERAL LOADS ON AIRCRAFT

engers safely, efficiently and comfortably over various distances between airports. on the other hand the Air Force Fighter type of aircraft has a job at shooting down enemy aircraft or protecting slower friendly aircratt. To do this job efficiently requires a far difterent contiguration as compared to the DC-8 transport. since modern aircraft tly in sub-sonic, transFurthermore the Fighter type airplane must be sonIc and super-sontc speed ranges. furthermaneuvered far mere sharply to do its required more, there Is a wide range ot Wing configurjob as compared to the DC-8 in do Ing its reatIons, such as the straight tapered wing, the qurr-ed job. swept wing and the delta Wing, and many at In general the magnitude of the air forces these wings otten include leading and trailing on an airplane depend on the velocity at the edge devices tor promoting better 11ft or conairplane and the rate at which this velOCity Is trol characteristics. The presence ot power changed in magnitude and direction (acceleration). plant nacelle unt ts J external fuel tanka, etc. The magnitude of the flight acceleration factor are units that etrect the airflow around the may be governed by the capacity at the human body to Withstand these acceleration inertia wing and thus effect the magnitude and distribution at the air torces on the Wing. Likewise, torces without injury which is the situation in the fuselage or airplane body itself influences a tighter type at airplane. on the other hand the airflow over--the wing. The theoretical cal- the maneuvering accelerations for the DC-8 are culation ot the alrloads on the airplane 1s too not dictated by what the human body can withlarge a subject to be covered in a structures stand, but are determined by What 15 necessary book and it is customary in college aeronautical to sately transport passengers trom one airport curricula to provide a separate course for this to another. subject. Designing the airplane structure tor loads In most airplane companies the loads on greater than the airplanes sutters in the perthe airplane are determined by a group at enformance of its required job, obviOUSly Will add gineers aSSigned to the Structures AnalysiS considerable weight to the airplane and decrease Section and this group is otten reterred to as its pertormance or over-all efficiency relative the Aircraft Load Calculation group. While the to the job it is designed to do. work of this group is primarily based on the Te particularly insure safety in the alruse at aerodynamics, it is that phase at aerotransportation, along with uniformity and efdynamics which is conserved with determining ficiency of design, the government aeronautical the magnitude and distribution at the air loads agencies (civil and military) bave definite reon the airplane so that the airplane structure qUirements tor the various types of aircraft can be properly deSigned to support these air relative to the magnitUde of loads to be used in forces safely and efficiently. The engineering the structural deafgn of aircraft. In referring department at an airplane company has a distinct in general to these specified aircraft loads two or separate aerodynamics section, but in general terms are used as follows:their responsibility is the use at the SUbject at aer-odynamrcs, to insure or guarantee the per- Limit or Applied Loads. The terms limit and applied refer to the tormance, stability and control at the airplane. same loads with the civil agencies (C.A.A.) A basic general over-all knowledge at the USing the term limit and the military agenc i es loads on aircratt is desirable in the study at USing the te~ applied. aircraft structural theory, and hence thiS Lim1t loads are the maximum loads anticichapter attempts to give this information. In pated on the airplane during its lifetune of a later Chapter dealing with wing design, this service. Subject will be further expanded. The airplane structure shall be capable or supporting the limit loads withuut suffering A4. t Limit or Applied Loads. Design Loads. detrimental permanent deformations. At all loads Because an airplane Is des Igned to carry up to the limit loads the deformation of the out a definite job, there result many types or structure shall be such as not to intertere with aircraft relative to size, conf1guration and the sate operation of the airplane. performance. For example, a commercial transport like the Douglas DC-8 is designed to do a Ultimate or DeSign Loads. jab of transporting a certain number of passThese 17NO terms are used in general to mean
A4. 1 Introduction.

Betore the structural design or an airplane can be made, the external loads acting on the airplane in flight, landing and take-ott conditions must be known. The complete determination ot the air loads on an aIrplane requires a thorough theoretical knowledge or aerodynamics,

A4.1

>4

A4.2

GENERAL

LOADS

ON

AIRCRAFT

the same thing. Ult~mate or Design Loads are equal to the limit :oads ~ultiylled by a ~actcr of safety (F.S.) or Design Loads = Li~it or Applied Loads times F.S. In general the oV9r-all factor of safety is 1.5. ~he goverr~ent ~eQuirements also specify that these design loads be carried by the structure without failure. Although aircraft are not s~?pOSed to undergo greater loads than the specified 11~lt loads, a certaln amount of reserve strength against complete structural tailure of a unit is necessary in the design of practi8ally any ~ chine or structure. This 1s due to many factors such as:- (1) The approximations involved in aerodynamic theory and also structural stress analySiS theorYj (2) Variation in physical properties of materials; (3) Variation in fabrication and inspection standards. POSSibly the most important reason for the factors of safety for airplanes is due to the fact that practically every airplane is limited to the maxtmum velocity it can be flown and the maximum acceleration it can be SUbjected to in flight or landing. Since these are under the control of the pilot it 1s poss~ble in emergency conditions that the limit loads ~y be slightly exceeded but with a reserve factor at safety against failure this exceeding of ths 11mit load should not prove serious from an airplane safety standpoint, although it might cause permanent structural deformations that might require repair or replacements of small units or portions ot the structure. Loads due to airplane gusts, are arbitrary in that the gust velocity 1s assumed. Although this ~~st velocity is based on years of experience in measuring and recording gust forces in flight allover the world, it is quite POSSible that during the lifetime of an airplane, turbulent conditions near storm areas or over mountains or water areas might produce air gust velocities sli~~tly greater than that speCified in the load requirements, thus the tact or of safety insures safety against failure 1: this situation would arise. The broad general category of external loads on conventional aircraft can be broken down into such classifications as follows:D ue to Airplane Maneuvers. (under the control of the ~ilot). (1) A1r Loads { Due to Air Gusts. (not under control of pilot).

Thrust . (3) ?cNer Plant Loads { Torque.

(4) Take

ott

HOi s t i ng Airplane. , 1 1 I " Towim:: Atr-ul.ane. (5 ) upee a ....oa"'s ~ " 1" I. <.,.. .... ~ " { Beaching of nu_l~y~e ~~~~l~r.~ Fuselage ?reSsurlzi~g. (5)
Wei~~t

and Inert1a Leads.

In resolving external loads :or stress analysis purposes, it is convenient to ~ave a set of reference axes. ~he rs!erer-ce axes XYZ paSSing throug~ the center o~ ~ravity of the air?lane as illustrated in Fig. A4.0 are these no~lly used in stress analYSiS work as well as for aerodynamic calculatior.s. Fer ccnve .. ience the reference axes are often referred to another origin other than the air;lane ~.~.
y

z
x

x
Z
Fig. A4.0

A4.4 Weight and Inertia Forces.

The term weight is that constant force, ,roportional ~o its mass, which tends to draw every phYSical body toward the center of the earth. .~ airplane in steady flight (uniform velocity) is aoted upon by a system of forces in equilibrium, namely. the weight of the airplane, ~he air forces on the complete airplane, and the power plant forces. The pilot can change this balanced steady flight condition ~y cr~ging the engine power or by operating the surface controls to change the direction of the airplane velocity. These unbalanced forces thus cause the airplane to accelerate or de-accelerate. Inertia Forces For ~ot:on of ~Jre Translation cf Rigid 80dles. If the unbalanced forces acting on a rigid body cause only a change in the magnitude of the velocity of the body, but not its direction, the motion is called tranSlation, and from basic Physics, the accelerating ~orce F = ~, where M is the ~ss of the bOdy or ~/g. In Fig. A4.1 the unbalanced force system F causes the rigid body to accelerate to the rl~ht. Fig. A4.2 shews the effect of this unbalanced ferce in ;roduclng

(2 ) Landing.Loads

Landi ng on Land.

(Wheel or

sk1 type).

Landing on ~~ter. Arresting. {Landing on Aircraft Carriers).

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A4.3

F :: unbalanced

external force Motion

Effective Force =1:ma=Ma , /mla/ ma

tw
Fig. A4. 1

- -- --- --- - -- -- - \

Motion
~

lf an acceleration normal to at = ra, and a = n the flight path at A and directed toward the center ot rotation (0). From Newton's Law the effective forces due to these accelerations are:2 F = !".rwlf = Mv /T _
n

rw

Fig. A4.2

where w = angular velOCity at the paint A. a angular acceleration at paint A. r radius at Cllrrnture at :light path at point A. The inertia forces are equal and OPPOSite to these etfective forces as indicated in Fig. A4.3. These inertia forces can then be considered as part at the total force system on the airplane which is in equilibrium. It the velocity or the airplane along the path 15 constant, then at = zero and thus the inertia force F = 0, leaVing only the normal t inertia torce Fn It the angular acceleration Is constant, the following relationships hold.
W -

a force on each mass particle of mLa, m~a, etc., thus the total effective force is zma = Ma. I f these effective forces are reversed they are reterred to as inertia torces. The external forces and tha inertia forces therefore torm a force system in equilibrium. From basic Physics, we have the following relationships tor a motion of pure translation it the acceleration is constant:v -

= at - lf s ve t -+- iat v" _ va If = 2as

Va

(1)
-

(2 )

(3 )

where,
5

= at - Q = wot ... ~a.tlf wa _ W o If = ZaQ

WQ

(6) (7 )
(8)

moved in time t , Vo = initial velocity v = tinal velocity atter time t. Inertia Forces on Rotating Rigid 80dies. A common airplane maneuver is a motion along a curved path in a plane parallel to the XZ plane of the airplane, and generally referred to as the pitching plane. A pull up trom steady flight or a pullout from a dive causes an airplane to follow a curved path. Fig. A4.3 shows an airplane following a :urved path. It at point A the velOCity is increasing along its path, the airplane 1s being subjected to two accelerations, n~elY, at' tangential to the curve at point A and equal in magnitude to
Center of Curvature

= distance

where Q

w = initial angular velocity in rad/sec. w = angular velocity after tune t.

= angle

of rotation in time t.

In Fig. A4.3 the moment To or the inertia forces about the center of rotation (0) equals I'tra(r)= MTaa. The term MT a is the mass moment pf inertia of the airplane about point (0). Since an airplane has considerable pitching moment of inertia about its own center ot gravity axiS, it should be included. Thus by the parallel axis
To = loa
Q

-+-

where I = Mfa and I e.g = moment or inertia of airplane about Y axis through e.g. at airplane. Inertia F0rces For Pitching Rotation of Airplane about Y Axis Through e.g. airplane.

I e.g a

- - - - - - - - - - - - -

(9)

01

r I

Flight Path

an=rwlf=Vlf / ?

=E2~~-MTa

AI

'.g. M? w lf = Mv a / ?

Fig. A4.3

In flight, an air ~~st may strike the horizontal tail prodUCing a tail torc~ which has a moment about the airplane c.g. In some landing conditions the ground or water forces do not pass t~xough the airplane e.g., thus producing a moment about the airplane e.g. These moments cause the airplane to rotate about the Y axis through the e.g. Therefore for this etfect alone the center or rotation in Fig. A4.3 is net at (0) but at

A4.4

GENERAL

LOADS

ON

AIRCRAFT

the e.g. at airplane, or r

o. Thus F and F t n equal zero and thus the only inertia ~~rce tor the pure rotation is I g C, (a couple) and c. . thus the moment of this inertia couple about the

e.g.

= Tc. g = Ic.g.~'

As explained before if the inertia forces are included with all other applied forces on the airplane, then the airplane is in static equilibrium and the problem is handled by the static equations for equilibrium.
A4.5 Air Forces on Wing.

air fOrC(3 on the wing. For example, consider the two air pressure intensity diagr~~ in ?lgs. A4.6 and A4.7. ~hese distributed ~orce systems can be replaced by their resultant (R), which of course must be known in magnitude, direction and location. ~~e location is specified by a term called the center of pressure which is ~he point where the resultant R intersects the a1rfoil chord line. As the angle at attack is Changed the resultant air force Changes in magnitUde, direction and center of pressure location.
R

The wing at an airplane carries the major portion of the air forces. In level steady flight the vertical upward torce of the air on the wing, practically equals the weight of the airplane. The term airfoil is used when reterring to the Shape of the cross-section at a wing. Figs. A4.4 and A4.S illustrate the air pressure intensity diagram due to an airFig. A4.6

Fig. A4.7

Lift and Drag Components of Resultant Air Force. Instead ot dealing with the resultant ferce R, it is convenient for both aerodynamiC and stress analysis considerations to replace the resultant by its two components perpendicular and parallel to the airstream. Fig. A4.8 illustrates this resolution into lift and drag components.
Angle of Attack

12"
Fig. A4.o4

Angle of Attack '" _ 60

Fig. A4.5

stream flowing ar-ound an airfoil shape tor 'both a positive and negative angle of attack. The shape and intensity at this diagram is influenced by many factors, such as the Shape of the airfoil itself, as the thiokness to chord ratio, the camber of the top and bottom surfaces etc. A normal wing is attached to a fuselage and it may support external power plants, wing tip tanks etc. Furthermore the normal wing is usually tapered in planform and thickness and may possess leading and trailing slots and flaps to prOduce high lift or control etfects. The airflow around the wing is affected by such factors as listed above and thus wind tunnel tests are usually necessary to obtain a true picture of the air torces on a Wing relative to their chordwise spanwtse distribution. Resultant Air Force. Center of Pressure. It is convenient when dealing with the balancing or equilibrium of the airplane as a whole, to deal with the reSUltant of the total

~
..
I :

:
'0

1 rAUlODTN..... IC
,J

Cl1:NTEII

o.p

'UGHT DlllltCTIOlI

FLIGHT DllIECTIOIII

'!'?5?:>
Fig. A4.9

Fig. A4.8

Aerodynamic Center (a.c. ). Since an airplane tlies at ~y different angles of attack, it means that the center ot pressure Changes for the many flight design conditions. It so happens, that there is one paint on the airfoil that the moment due to the Lift and Drag torces is constant ~or any angle ot attack. This point 1s called the aerodynamic center (a.c.) and its apprOXimate location 1s at the 2S percent of chord measured from the leading edge. ~hus the resultant R can be replaced by a 11ft and drag force at the aerOdynamic center plus a wing moment Ma.c. as illustrated :n Fig. A4.9.
A4.8 Forces on Airplane in Flight.

Fig. A4.l0 illustrates in general the main forces on the airplane in an accelerated flight condition.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A4.5

r/o--

..,
E

tiplying factor by which the forces on the airplane in steady flight are multiplied to obtain a static system ot forces equivalent to the dynamic force system acting during the acceleration of the airplane. Fig. A4.ll illustrates
Z

LA1 total lift (Wing

Fig. A4.1O

&l Tail)

T_

Fig. A4.11

T
L

=engine thrust.
= total

D = total airplane drag. Ma moment of Land D with reference to wing

a.c. (aerodynamic center) W = weight of airplane. I L = inertia force normal to flight path. In inertia force parallel to flight path. 1 m rotation inertia moment. E tail load normal to flight path. For a hortzontal Constant velocity flight condition, the inertia forces I L, 1D, and 1m would be zero. For an accelerated flight condition involving translation but not angular acceleration about its own e.g. axis, the inertia moment 1 would be zero, but I and 1 L D m would have values. Equations of Equilibrium For Steady Flight. From Fig. A4.10 we can wrlte:l:F

wing lift plus fuselage lift.

forces in steady horizontal flight. L represents the total airplane lift (Wing plus tail). Therefore L = W. Now assume the airplane is accelerated upward along the Z axis. Fig. A4.12 shows the additional inertia torce wag/g acting downward, or opposite to the direction ot accaleration. The total airplane lift L tor the
Z

",L
"...- e.g. 0

w
Fig. A4.12

hg.,

unaccelerated condition tn Fig. A4.ll must be mUltiplied by a load factor n to produce statiC a equilibrium in the ~ direction.

Thus,

x ZF ~.= 0,

0,

D + W sin Q-T cos ~

L

n L - W - ~ e., ::; a g a

=
~ +

ZMy

- '"
+ '"

Since L Hence n
te = a
An

cos

Q+T

sin ~ - E

=W
::; 1
+ a~

0,

-l1a-

La - D b + Tc

cos

Equations of Equilibrium. in Accelerated Flight. D l:F sin Q - T cos ~ I 0,

ZFa

= 0,
0, -

- '"

airplane can at course be accelerated along the X axis as well as the ~ axis. Thus in Fig. A4.l3 the magnitude ot the engine thrust T is greater than the airplane Drag D, which
..L

cos

Q +

T sin

- IL

- E

=
Im

Ma -

La - Db

+ Tc cos ~ +

Ee

=0
~.g.

r Forcee

- Plus ts UD and toward tail.

Signs used: lMoment - C'Iockwf se is positive. Distances tram e.g. to force Plus is up and toward tail.
A4.7 Load Factors.

w
ng. A4.13

~waz g

T is greater than D

The term load factor normally given the Symbol (n) can~dettned as the numerical mu1 The bar through letter Z has no significance. ing without bar. Same mean-

causes the airplane to accelerate forward. It is convenient to express the inertia force in the X direction in terms of the load factor n and the x

-.Zq

A4.6

GENERAL

LOADS

ON

AIRCRAFT
,eloc~ty

weight W of the airplane J hence

n W x
'~

defin~te ~ax~~w~ r:~;~:

:ar

~om

mer-c Ia l airplanes the ve Loc i t y is

imi t.ec to

2.

ax

(See Fig. Al4.l3)

o

reasonable ~lide speed whic~ ~s su ficient to take care of ~easo~a8le flis~t opera~icns.
A4. 9 Gust Load Factors.
~~en a s~a~, edge ;ust strikes :~e ~ir?l2.ne in a di~ection normal to the thrust line (X axis)J a sudden change ta~es place in :~e wing angle of attack with no sudden change in ~ir)la~2 sDeed. The cOrTal force coeffic~ent (C~ ) can

OJ whence T-D-n xw T-O Hence n x W

Therefore the loads on the airplane c~~ be discussed in terms of load factors. The applied or limit load factors are the ~axi~~T. load ~actors that might occur d~ing the ser7ic3 of "the ?articular ai~plane. These loads as jisc~ssed in Art. A4.2 ~us~ ~e taken by t~e ~ir?lane s:r~c ture without appreCiable per.Tanent deformation. The design load factors are equal to the limit load factors multiplied by the factor of safety. and these deSign loads must be carried by the structuta without rupture or collapse J or in other words J ~ornplete failure.
A4.8 Design Flight Requirements for Airplane.

-,
"

be ass~ed to vary linearly wl':h :he a~gle of attack. Thus in rig. (a)J let ?Ol~t (3) represent the nomal air?lane :orce ccef:icient C7
'-')..

necessary to ~inta~n level ~l~sh: velOCity V and point (s) the 'ffilue

Softer

The Civil and Military Aeronautics Authorities issue requirements whic~ specify the design conditions for ~he various class1ficat1on of airplanes. Generally speaking J any air~lane flight altitude can be defined by stating the existing values of load factors (acceleration) and the airspeed (or mere properly the d}~ic pressure) The accelerations en the airplane are produced fram ~NO causes J nanelYJ ~~euvers and air gusts. The accelerations due to maneuvers are subject to the control of the pilot who can manipUlate the controls so as not to exceed a certain acceleration. In hi&~ly maneuverable military airplanes J an accelerometer is included in the cockpit instruments as a ~ide to limit the acceleration factor. ror commercial airplanes the maneuver factors are ~de high enough to safely take care or any maneuvers tr~t would be reqUired in the necessary fl~ght operations of the particular type of airplane. ~hese limiting maneuver factors are based on years of operating experience and have given satisfactory results fram a sarety standpoint without pen-' alizing the airplane from a weight deSign consideration. The accelerations due to the airplane striking an air gust are not under the control of the pilot since it depends on the direction and velocity of the air gust. From much accumulated data obtained by instal:ing accelerometers in commercial and military aircraft and flying them in all types of weather and locations. it has been found that a gust velocity or 30 ft. per second appears sufficient. The speed or velocity of the airplane likewise effects the loads on the ai~plane. The higher the velOCity the higher the aerodyna~ic Wing ~oment. Furthermore the gust accelerations increase with airplane velocitYJ thus it is custcmary to limit the partic~lar ai~plane to

a sharp edged ~ust KU has caused a s~dden change 6a in the a~gle of attack wit~out c~ange in V. The total lr.crease in the airpl&~e load in ~he Z direction can therefore be expressed jy the ratio C7 at 8.
~,

~.

From ?!g. (b) for small angles, 0" KU~l and from Fig. (a) J 6C = m 6.0., wnere cn the ZA Slope of the airplane no~~l force curve (C 7
~A

per r-anian) .

KU~
V Fig. b Fig. a

Tr4e load factor increment jue to the gust KU can then be expressed
KUVS
m 575 ',.,[

(A)

U gust velOCity in tt./sec. K*= gust correction factor depending on wing loading (Cur7es for K are prOVided by Civil Aeronautics Authorities). V indicated air speed in miles per hour. S wing area in sq. ft. W gross weight of airpla~e.
NACA Technical Note 2964 (June 1953), proposes that the alleviation factor K should be replaced by a gust factor, Kg :: 0.88 Mg/(5. 3 + Mg}. In this expression Mg is the airplane mass ratio or mass parameter, 2 W!o.pcgs. In which c is the mean geometric chord in feet andg the acceleration due to gravity.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A4,7

It U is taken as 30 ft./sec. and

c~ange

in C

zA

wit~

~ as the respect to angle of attack

in absolute ~~its per degree, equation (A) reduces to the following

on

~V

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

(E)

Therefore the gust load factor n when airplane is flying in horizontal altitude equals
n ;: 1 ... 3KmV
I,.J/S
1s

(e)

and when airplane

n -

in a vertical altitude
(D)

3KmV

- ',,/s

- - - - - - - - - - - - - - - -

A4.10 illustration of Main Flight. Conditions. Velocity-Load Factor Diagram. As indicated before the main design flight conditions for an airplane can be given by stating the limiting values of the acceleration and speed ~nd in addition the maximum value of the applied gust velocity. As an illustration, the design loading requirements for a certain airplane could be stated as follows: ~The proposed airplane shall be designed !or applied positive and negative accelerations of + 5.0g and -3.5g respectively at all speedS tram that corresponding to CL up ~o 1.4 times the max maximum level flight speed. Further.TIore, the airplane shall Nlthstand any applied leads due to a 30 ~t./sec. gust actl~g in any direction up to the restricted speed of 1.4 times the maximum level flight speed. A design factor of safety of 1.5 shall be used on these applied loads" . In graphical form these design requirements can be represented by plotting load factor and velocity to obtain a diagram which is generally referred to as the Velocity-acceleration diagram. The results of the above specification would be similar to that of Fig. A4.14. Thus, the lines AS and CD represent the restricted pOSitive and negative ~euver load factors which are limited to speeds inside line BD "Nhlch is taken as 1.4 times the maxtmum level fli~~t speed in this illustration. These restricted maneuver lines are ~er.TIinated at points A and C by their intersecticn With the :nax1m.um C values of the a ir-cjcne At speeds L beCNeen A and B, the pilot ~ust be careful not to exceed the ~neuver accelerations, since in general, it would be posSible tor him to ~n ipu1ate the controls to exceed these values. At speeds beloW A and C, there need be no care or the pilot as far as loads on the airplane are concer~ed since a ~neuver prodUCing C~.

would gl ve an acceleration less than the limited values given by l1nes AB and CD. The positive ,and negative gust accelerations due to a 30 ft./sec. gust normal to f:i~ht path are shown on Fig. A4.l4. In this example diagram, a positive !list is not critical within the restrictAd velOCity of the air?lane since the 6Ust lineS intersect the line BD below the line AB. For a negative gust, the zust load tactor becomes critical at velocities between F and D with a maximum acceleration as given by paint E. For airplanes 'Nhich have a relatively low required maneuver factor the IUst accelerations may be critical for both pOSitive and negative accelerations. ~ination of the 3Ust equation indicates that the most lightly loaded condition (smallest gross weight) prOduces the highest gust lead factor, thus inVOlving only partial pay load, fuel, etc. On the dia~am, ~he points A and 3 correspond in general to what is referred to as high angle ot attack (H,A,A,) and low angle of attack (L.A.A.) respectively, and points C and 0 the inverted (H.A.A.) anrt (L.A.A.+ conditions respectively. Generally speaking, if the airplane is deSigned for the air lOBds produced by the velocity and acceleration conditions ~t paints A, 8, E, F, and C, it should be sate from a structural strength standpoint if flown Within the speCified limits regarding velocity and acceleration. Basically, the flight condition reqUirements of the Civil Aeronautics Authority, Army, and Navy are based on consideration ot specified velocities and accelerations and a consideration ot gusts, thus a stUdent understanding the basic discussion above should have no diffiCUlty understanding the deSign reqUirements of these three government agenCies. For stress analySiS ~urposes, all speeds are expressed as indicated air speeds. The "indicated~ air speed 1S detined as the speed which would be indicated by a ~erfect air-speed indicator, that is, one that would indicate true air speed at sea level under standard atmospheric conditions. The relation between the actual air speed Va and the indicated air speed Vi is given by the equation
VI

.~

Va

where Vi Va indicated airspeed actual airspeed

Po = standard air denslty at sea level Pa ;: density of air in which Va is attained

A4.8

GENERAL

LOADS

ON

AIRCRAFT

';''1'''1'''''' ;'1',:: ,.. !'

I, I'

',,;,
~

It;l""

,,""'
1Cl:T

,~~.

I J"

L ;" 1,/
L [,,,
'Qi

'.__

i!~;~~~~;i:;'"lli~~,~;["!
A4.11 Special Flight Design Conditions.

E "'T i.l '1;;, ","" ,f'" t ~:~ ;~"L~ :;.t.r' : ._::: ,:::~::::~:~T:: U~~::

<: ...

',. :~:~~tV:; .:._:~ ':~t$.~

i!i;Wf~ :~iI;; h'::+:; ',:}J:.-

h~RT [,,''I Y1"I'; ", ~

l>i' \,,; i:,=: :E,

'''.l'''l:;;'~

I,; D.""

the loads on the wing should be checked for cases where the engines are attac~ed to the wing and ~e located ~orward of the leading edge. In cases where the land1ng gear is attac~ed to wing or when the :uel and engines are carried in and on the wing, the loads produced on the wing structure ~n a landing condition nay be critical for some portions of the wing structure inboard of landing gear and engine attac~~.ent points.
A4. 12 Example Problems Invcivtng Accelerated Motion of Rigid Airplane.

<L

r;: r': :~;::.:

~1:-"
.::;~:

L i::: k

~ ~rr: ~~~;U ~~ 2; ;lli::" '~ii~l, "'?'. ,;S I , ~~~~~~ ~~imI; ~il:I~~ ~~;t~~~~~~~:~: '~~J:~A:EfSS:! H:' \~~;~ i~~~~ ~~i~ m~;~~~~g~~~;g~ i;~:!~mlli~~lli~tiTIHr~ li::~i :1',::[" ""W i :~I"" :!i:lFi IeJ!iijjJijjjij'iitljijrJlic;'J' Fi~:t4:' I.'

As previously explained, it is general practice to place the airplane under accelerated conditions of motion into a condition of static equilibrium by adding the inertia farces to the applied force system acting on the ai~plane. It is usually assumed that the airplane is a r~gid body. Several example problems will be presented to illustrate this general proceaure. Example Problem 1 Fig. A4.l5 ill~strates an airplane landing on a Navy aircraft carrier and Jeing arrested by a cable pUll T en the airplane arresting tcok. It the airplane-weight is 12,000 Ibs. ar.d the airplane is given a constant acceleration of 3.5g (112.7 ft/sec~), ~lnd the hook pull T, the wheel reaction R, and the distance (d) be~Neen the line of action of the hook pUll and the airplane e.g. If the landing velocity is 60 ~.P.H. what is the stopping distance.
w" 12000 lb.

There are $anY other rli~~t conditions 'which may be critical ror certain portions at the wi~ or fuselage structure. Most airplanes are equipped with flaps, to decrease the landing speed and such flaps are lowered at speeds at least twice that or the m1n1mUm landing speed. Since the flapped airfoil has ditferent values for the magnitude and location of the airro1l Characteristics, the Wing structure must be checked for all possible flap conditions within the specified requirement relative to maximUm speed at which the flaps may be operated. Generally speaking, the flap conditions wtll efrect only the wing portion inboard or the flap and it is usually only critical for the rear beam web or shear wall and for the top and bottom walls or the torsion box. This is due to the fact that the detlection flap moves the center ot pressure considerably aft thus producing more shear load on the rear shear wall as well as torsional moment on the conventional cantilever box metal beam. The a1rplane must likew1se be investigated for aileron conditions. Operation of the ailerons produce a different air lcad on each side or the airplane wing which produces an angular rolling acceleration of the airplane. Furthermore, the deflected ailerons change the magnitude and location of the airfoil characteristics, thus calculations must be carried aut to determine whether the loads in the aileron conditiOnE are more critical than those for the normal flight conditions. For angular acceleration resulting fram pitChing moments due to air gusts on the tail,

Lx
T

24"

Fig. A4. 15

Solution: On contact of the airplan~ with the arresting cable, the airplane is decelerated to the ri&~t relative to ~ig. A4.15. The motion is pure translation horizontally. The inertia ~orce 1s
l1a

=~

= (12~00)

3.5g

= 42000

lb.

The inertia torce acts opposite to the direction at acceleration, hence to the left as shown in Fig. A4.l5. The unknown forces T and R can now be solved tor by using the static equations of equilibrium.

ZFx = -42000
hence,

T cas 10 = 0

T = 42700 lb.

ZFe = -12000 + R - 42700 x sin 10

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A4.9

hence,
~o

= 19420

lb.

hence,
ZF~

R, = 29800 lb. (up)

find the distance (d) take moments about c.g. at airplane, hence,
ill

= 29800

- 9000

R1

=0

c.g.

19420

x 24 - 42700

d = 0

R, = - 20800 lb. (acting down)

hence,

d = 10.9 In.
~

The velocity at end ot catapult track can be found trom. the rof towtng equation
V_Voll~2as

Landing velocity Vo

60 M.P.H. = 88 tt/sec.

V" - 0 ':: 2 x 96.6 x 35

Subt: hence stopping distance s = 34.4 ft. Example Problem 2 An airplane equipped w1th float is catapulted into the air tram a Navy Cruiser as illustrated in F1g, A4.l6. The catapulting torce P gives the airplane a constant horizontal acceleration of 3g{9S.6 tt/sec). The gross weight ot airplane 9000 lb. and the catapult track is 35 ft. long. F1nd the catapu rt ing torce P and the reactions R1 and R. tram the catapult car. The engine thrust 1s 900 lb. :~at is airplane velocity at end at track run?
9000 lb.

or

v=

82 tt/sec. = 56 M.P.H.

Example Problem 3 Assume that the transport aIrplane as illustrated In Fig. A4.l7 has Just touched down in landing and that a brakmg torce at 35000 lb. on the rear wheels is beIng applied to bring the airplane to rest. The landing horizontal vetoo1ty Is 85 M.P'.H. (125 rn/sec) , Neglecting air

forces on the aIrplane and assuming the propeller forces are zero, what are the ground r-eac t i ons R1 and R.. What is the landing run distance With the constant braking forcs?
W ,. 100,000 lb.

Lx

Max.....--c.g.-

T--

-l5",J~,"""===ThrU8tLine _ _

cg:--Ma

-<
A

78"

t,
R,
Fig. A4.16

Solution: The airplane is being decelerated horizontally hence the inertia torce through the airplane e.g. acts toward the tront ot the airplane. Since the braking torce is given we can solVe tor the deceleration factor by the equilibrium equation,
l:Fx

~85"-j

Solution: The torces will be determined just atter the beginning of the catapUlt run, where the car velocity 1s small, and thus the lift on the airplane wing and the airplane drag can be neglected. Horizontal inertia force acting toward the airplane tail equals,

= 35000
II H g

- Ma

henes,

Max = 35000 a

or
whence

X=

35000

/"", = (9000) 3.0g = 27000 lb.

g
From statics: -

35000 (100000) 32.2 = 11.27 tt/sec'

ZFx = -900 - P
To find
2:1:
R~

27000

a, hence P

= 26100

To find landing run (s),

lb.
o -

take

~oments

about point A,

v hence,
s

Vo = 2 a x s

I25' = 2 (-11.27) s

= 9000 x 55 + 27000 x 78 - 900 x 83 - 85R li

=0

= 695

ft.

A4.10

GENERAL

LOADS

ON

AIRCRAFT

To find Re take moments about point (A)

ZM A

Airplane Load Factor

= AirPl~e
7.7

Lift

= 100,000
Re =

x 21

35000 x

9 +

38

Ra

o
111800 - 4100 14000
~ple

47000 lb. (2 wheels)

+ R:l.

ZF

= 47000 - 100,000
R1 = 53000 lb.

= 0

Example Problem 4 The airplane in Fig. A4.lB weighs 14,000 Ib It is flying horizontally at a velocity ot 500 M.P.H. (732 ft/sec) when the pilot pulls it upward into a curved path with a radius ot curvature of 2500 tt. Assume the engine thrust and airplane drag equal, opposite and colinear with each other (not shown on Fig. A4.l8).
Flnd: (a)

(b)

(c)

Acceleration ot airplane in Z direction Wing Llft (L) and Tall (T) forces Airplane Load tactor.

Problem 5 Assume the airplane as used in example problem 4 is in the same attitude as used in that example proo tem. Now the airplane is turther maneuvered by the pilot suddenly pushing the control stick forward so as to give the airplane a pitching acceleration of 4 rad/sec ". (a) Find the inertia forces and the tail load T, assuming the lift force on the wing does not change. (b) Find the torces on the jet engine which weighs 1500 lb. and whose c.g. location is shown ln Flg. A4.19. Assume moment of inertia I y (pitching) of the airplane equals 300,000 lb. sec a in.
93700 lb.

,
- D c :. .

210"
Engine Thrust

J
14000

..

tIll. 800 lb. 210'

I

L.
Solution: -

t"'so'""",

w
Fig. A4.18

Fig. A4.19

Solution: -

V' 732 Acceleration aa = r 2500 = 214.5 tt/sac a or 214.5/32.2 = 6.67g (uj>'IaI"d). The inertia force normal to the flight path and acting down equals
Ma 6 67g = 93700 lb. - (14000) g z -

Fig. A4.19 shows a tree body at the airplane With the lift and inertia forces as found in Problem 4. The additional inertia torce due to the angular accelerat10n a = 4 rad/sec a equals,
Ia=300000x4
y

1,200,000 in. lb.

Placing this torce on the airplane through the e.g. promotes static equilibril~, hence to find tail load T takes moments about wing aerodynamic center (c.p.)
ZMc.p hence - (14000 + 93700) 8 + 210 T = 0 T = 4100 lb. (down)

which acts clockwise or counter to the direction at angular acceleration. The airplane is now in static equilibrium and to tind the tail load T take moments about airplane e.g.
ZMe.g. = 1,200,000 - 111800 x B - 218 T T = 1400 lb.

To find Wlng Llft (L) use

l:F

To find Mas take,

= - 4100

14000 - 93700 + L

o
14000
+

1400 - Mas = 0

L = 111800 lb.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A4.1l

hence,

'I' 992 00) _ ( 14000 g - 7.1 g . . t sec

Find:
(a) The inertia forces on the air-

plane. The e.g. of the engine is 50 inches aft of the airplane e.g. as shown in Fig. A4.19. The force on the engine will be its own weight at 1500 lb., and the inertia forces due to a~ and
~.

It) The resultant load on the pilot

whose weight is 180 lb. and Whose location 1s shown in Fig. A4.20.
IPilot c. g.

Inertia force due to a g equals,

Ma. = (15 7.1g = 10630 lb. g00)

\ 1-372"-1
C .I:....-_7' ~

T::.
Fig. A4.20

Inertia force due to angular acceleration a equals,

Mra

= 3Z~~0~ 12 x 50 x 4 = 778 lb. (dawn)

Solution: -

319000 lb.

'''Y

oopooc!-l-=J::=-,,--,-,_

Then the resultant forca on tha engine equals

1500 + 10630 + 778

= 12908

lb. (down)

Note if the engine had been forward of the airplane e.g., the inertia force of 778 lb. would act upward instead ot downward. In calculating the inertia force on a certain airplane item due to angular acceleration,the equation F = Mra assumes that the particular item had negligible mass moment at inertia about its own centroldal Y axis. In the case of a large item this centroidal mass moment of inertia may be appreciable and should be .included in the I y of airplane. Then to find the inertia torce tor such an item the equation F = Mra. Should be ncct r iec to be
F=(I e.g. ~)/r

The wing l1tt will be neglected in this example problem. The inertia forces on the airplane are rcrces Max and I"'.a~ and the couple I c.g. u , To find I"'A take, x
lF

x = 100,000 Max

- l1ax

=a

or hence,

= 100,000

lb.

ax =
To find
Ma~

100,000 = 11

eOO,OOO) 100,000

19

take,

Where hence

ZFs

= 300,000

- 100,000 - Ma a =

r = distance or ot item.
I

a~

trom airplane e.g. to e.g.

Ma. = 200,000 lb.

moment of inertia at item about e.g. airplane e.g. equals 10 + Mr il where 1 0 is mass moment of inertia Of item about its own eentroidal Yaxis. F = inertia force in lbs. no~l to radius i . Example Problem 6 Fig. A4.20 shows a large transport airplane whose gross weight Is 100,000 Lb, The airplane pitChing mass ~oment of inertia I y = 40,000,000 lb. sec il in. The airplane is making a level landing with nose wheel slightly otf ground. The reaction on the rear wheels is 319,000 lb. inclined at such an angle to give a drag component of leo,Ooo lb. and a vertical component of 300,000 lb.

= mass

a iS

= 200,000 =
11

200,000) _ ( 100,000 g - 2g t

To find the inertia couple I a, take moments about airplane e.g., e.g.
ZM

c.g.

=-

100,000 x 120 - 300,000 x 84

+1

e.g.

0.=0

c,g,

a. = 37,200,000 lb.

~ _37,200,000= hence angular acee 1erat . . on a. - 40,000,000

0.93 ran/sec>.

A4.12

GENERAL LOADS ON AIRCRAFT

Calculations of resultant load on pilot: -

x = 372"
c.g. of pilot
iL _

/'
= 0.93

are for deSign loadS, which in general are 1.5 times the applied loads. It would not be correct to saythat the Wing deflections under the applied loads for these two High Angle of attack condltior~ would be 2/3 the deflections sh0'Nn in the photograph since under the deSign loads a considerable portion of the wing would be stressed beyonu the elastic limit of the mater!al or into c s g , air- the plastic range where the st i rrnass modulus rs plane

Fig. A4.21

Fig. A4.2l shows the airplane e.g. accelerations The torces on the pilot consist of the pilots weight of 180 lb. and the various inertia forces as indicated in the figure.
!lax !las

= e~O)

19

= 180

= e~O) 2.0g = 380

The inertia torce due to the angular acceleration a acts normal to the radius arm between the airplane e.g. and the Pilot. For convenience this normal force will be replaced by its i and x components.
180 Fx = Mza = 32.2 x
F

12 x 40 x 0.93

17 lb.

180 =.Mia = 32.2 x 12 x 372 x 0.93 = 161 lb.

Total torce in x direction on pilot equals

180 - 17

= 163

lb.

Fig. A4.21

Total torce in 8 direction =

380 + 180 - 161

= 379

lb.

Hence Resultant force R, equals

conSiuerably less than the modulus of elasticity, hence the deflections under the applied loads would be somewhat less than 2/3 those shown 1n the photograph. This photograph thus 1ndicates very strikingly that a wing structure is far from being a rigid body. Static loads are loads which are gradually applied and cause no appreciable shock or Vibration of str~cture. On high speed aircraft, air gusts, flight maneuvers and landing reactions are applied quite rapidly and thus can be classed as dynamic loads. Therefore when these d~iC loads strike a ~lexible (nonrigid) airplane cantilever Wing, a rather large Wing deflection 1s produced and the wing tenas to vibrate. This vibration therefore causes additional accelerations of the mass units of the Wing which means additional inertia forces on the wing. Furthermore if the time ~te of application of the external applied forces approaches the natural bending frequencies of the Wing, the Vibration excited can ~roduce large additional wing stresses.

A4. 13 Effeet of Airplane Not Being a algid Body.

The example problems of Art. A4.12 assume that -tne airplane is a rigid body (eur.rer-s no structural der ormat tcn) On the basis of this assumption the applied loads on the airplane to either tlight or landing conditions are placed in equilibrium With the inertia forceS which occur due to the acceleration of the airplane. It is obvious that an airplane structure like any other structure is not a rigid bOdy, particularly a cantilever wing which undergoes rather large bending deflections in both flight and landing conditions. Figure A4.21 shows a composite photograph taken of a test wing for the Boeing 8-47 airplane. The maximum upward and downward deflections shown

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A4.13

Up until World ~ar II practically all airplanes were assumed as rigid bodies for structural design purposes. During the war failure of aircraft occured under load conditions which the conventional design procedure based on rigid body analysis, indicated satisfactory or safe stresses. The failures were no doubt due to dynamic overstress because the airplane 1s not a rigid body. Furthermore, airplane design progress has resulted in thin wings and relatively large wing spans, and in many cases these wings carry concentrated ~sses, such as, power plants, bombs, wing tip fuel tanks etc,. Thus the flexibl1i ty Of_ wings have increased which means the natural bending frequencies have decreased. This fact together with the fact that airplane speedS have greatly increased and thus cause air gust loads to be applied more rapidly, or the loading is becoming more dynamic in character and thus the overall load effect on the wing structure is appreciable and cannot be neglected in the strength design of the wing. General Dynamic Effect of Air Forces on '..ling Loads. The critical alrlaads on an airplane are caused by maneuvering the airplane by the pilot or in striking a transverse air gust. A transport airplane does not have to be designed tor Sharp maneuvers producing high airplane accelerations in its job at transporting passengers, thus the time of applying the maneuver loads is conSiderably more than a fighter type airplane pUlling up sharply from high speeds. Fig. A4.22 shows the result of a pUll-Up ~neuver on the Douglas D.C.3 airplane at 180 M.P.H. relative to load factor versus time of application at load. As indicated the peak load of load factor 3.25 was obtained at the end of one second of time.
4

for applying the load on the wing when striking the air gust. NACA Technical Note 2424 reports the flight test results on a twin-engine Martin transport airplane. Strain gages were placed at various points on the wing structure, and strains were read. for various gust conditions for Which the normal airplane accelerations were also recorded. Then slow pUll-Up maneuvers were run to give similar airplane normal accelerations. The wing had a natural frequency of 3.8 cps and the airplane speed 'HaS 250 M.P.H. Two of the conclUSions given in this report are: - (1) The bending strains per unit normal acceleration under air gusts were approximately 20 percent higher than those of Slow pull-ups for all measuring POSitions and flight conditions of the tests, and (2) The dynamic component ot the wing bending strains appeared to be due primarily to eXcitation of the tundamental wing bending mode. These results thus indicate that air gusts apPly a air load more rapidly to a wing than a maneuver load gl vtng the same airplane normal acceleration for a commercial transport type at airplane, and thus the dynamiC strain effect on the Wing is more pronounced for gust conditions. Figs. A4.23, 24 and 2S show results of dynamic effect at air gusts on a large wing as determined by Bisplinghotr. The results in these tigures show that dynamic effects tend to conSiderably increase wing forces on same portions of the wing and decrease it on other portions.
-000
I

Fig. A4.23

Comparative Shear

t5

:; -400 -200

<

, 1\

,
<,

Dtstribution
Dynamic Analysis
- - - Rigid Airplane

Analysts

o
-30

.2

.4

.6

.a

i'..
1.0 Fig. A4.24

~ 3 <; 2
1

/ ' -,

<,

./
1

-,
1 1.5 2

Fig. A4.22

,
-,

c
~ .;

e -20
-10

Comparative Bending

.5

I'
.2

Moment Distribution

Pull-up of DC-3 Airplane at 180 mph.

I"
<,

The author estimates the natural frequency of the 0.:.3 Wing to be around 10 to 15 cycles per secor.d, thus a loading time of 1 second against a tDne of 1/10 or 1/15 for halt a wing deflection cycle indicates that dynamic overstress should not be appreCiable. In general, it can be said that cynamic over-stress under maneuvering loads en transport airplanes is not as great as from other conditions such as air gusts or landing . Dynamic Effect of Air Gusts. The higher the air gust velocity and the higher the airplane velocity, the less the time
,

o
-3

.6

.6

i
l.0
Fig. A4.25

,
I

g.-1
~

-2

a
2 0

-.

~ c~~~:a~~v;o~~~~i!:Ill

I
I

.. 1

, ['''' I .2 .4 .6 .8 .10 Fraction of Semi-Span

-ii

Abou.. Elaatic Ax1s.

General Data:Wing Span "" 189 ft. GroS8 wt, "" 184000 lb. Airl.a.ne Vel. :z260 mpb.

Report on an Investigation on stresses in Aircraft Struc. tures under Dynamic Loading. M.l. T. Publication.

A4.14

GENERAL

LOADS

ON

AIRCRAFT
W: 15000 lb.

It has also been four.d ~hat landing loads applied through the conventional land:ng gear or by water pressure on a flying boat are applied rapid enough to be classed as dynamic loads and such loads applied to wings of large span produce dynamic stresses which cannot be neglected in the sate design of such structures.
A4.14 General Conclusions on Influence of Dynamic Loa.di.ng on structural Design of -Afrpkane.

-(

e c.g.,

-120'9; , Rl

65"
I

,,
Fig. A4.26

The advent of the turbo-jet and the rocket type engines has opened up a range of possible airplane airspeeds hardly dreamed of only a few years ago, and already trans-sonic and supersonic speed airplanes are a cammon development. From. an aer-odynamtc standpoint such speeds have dictated a thin airfoil section which has thus promoted a high density wtng. Thus tor airplanes With appreciable Wing spans like Military bombers and near future jet commercial transports, which usually carry large concentrated masses on the wing such as engines, fuel tanks etc ,; the assumption that the airplane is a rigid body is not sufficiently accurate enough because the dynamiC stresses are appreciaDle. The calculation of the dynamiC loading on the Wing requires that the mass and st 1tfness distribution or the wing structure be known. Since these factors are not known when the structural design of a wing is started, the general procedure in des tgn would be to first base the design on the assumption that the wing is a rigid body plus correction factors based on past deSign experience or available research information to approx1mately take care of the influence of the elastic wing on the airplane aerodynam1C characteristics and the build up dynwmic inertia forces. ~lth the wing thus initially deSigned by this procedure, it then can be checked by a complete dynamiC analysiS and modif~ed as the results dictate and then recalculated for the modified elastic Wing. This procedure is now practical because at the availability of high speed computors.
A4. 15 PROBLEMS.

90"

---I

zontal deceleration and the stopping distance for the airplane? (3). The flying patrol boat in Fig. A4.27 makes a water landing with the resultant bottom water pressure of 250,000 lb. as shown in the figure. Assume lift and tal~ loads as shown. The pitching moment of inertia at the airplane is 10 million lb. sec.~ in. Determine the airplane pitching acceleration. what is the total load on the crew member who weighS 200 lb. and is located in a seat at the rear end of the r.ull?
Fig. A4. 27 1000 lb.

30"- / '1

20"~ L

25000 lb.

W: 50000

r--'oo" --rr-

__ c. g,

__

'LJ'

0"

g. of Crew Member

(1). The airplane in Fig. A4.26 is being launched from the deck ot an aircraft car~ier by the cable pull T which gives the airplane a forward acceleration of 3.25g. The gross ~eight of the airplane 1s 15,000 lb. (a) Find the tension load T in the launching cabls, and the wheel reactions R1 and R.o (b) If the flying speed is 75 M.P.H., what launching distance is required and the launching time t? (2). Assume the airplane of Pig. A4.26 is landing at 75 M.P.H. on a runway and brakes are applied to the rear wheels equal to 04 of the vertical rear wheel reaction. ~hat is the hori-

(4). The jet-plane in Fig. A4.28 is diving at a speed at 600 X.P.H. when ~ilot starts a 8g pUll-out. Weight of airplane is 16,000 lb. Assume that engine thrust and total airplane drag are equal, opposite and colinear. (a) Find radius of flight path at start of pUll-out. (b) Find inertia force in Z direction. (c) Find 11ft L and tail load T.

T
;,
~

.-l

L t'. "
<i

Fig. A4.28

CHAPTER A;

BEAMS SHEAR AND MOMENTS

AS.l Introduction.

In general, a structural ~ember that suploads perpendicular to its longltudlonal axis is ~6ferred to as a beam. The structure of aircraft provides excellent examples of beam units, such as the wing and fuselage. Very seldom do bending forces act alone on a ~jor aircraft structural unit, but are accompanied by axial and tors1onal forces. However, the bending forces and the resulting beam stresses due to bending of the beam are usually of primary importance in the design ot the beam structure.
~orts

30"
.1

25
Fig. AS. 1 10

.1----15"'

7S lb.

-1

:i
N
~

AS.2 statically Determinate and statically Indeterminate Beams.

Fig. AS. 2

" >1
a
75

g
>-

. I OO . P

Fig. AS. 3

75

A beam can be considered as SUbjected to known applied loads and unknown supporting reactions. If the distribution of the applied known loads to the supporting reactions can be determined trom the conditions of static equilibrium alone, namely, the summation of forces and ~oments equal zero, then the beam is considered as a statically d~terminate beam. However, if the distribution of the known applied loads to the supporting beam reactions is influenced by the behavior of the beam material during the loading, then the supporting reactions cannot be found by the statical equilibrium equations alone, and the be~ is classified as a statically indeterminate beam. To solve such a beam, other conditions of tact based on the beam deta~tions ~ust be used in combination With the static equilibrium equations.
AS. 3 Shear and Bending Moment.

100 ... 1

tclM
:i

::;

H
I.

C .....
d

75

Fig. AS. .o4

~
I.
Flg. AS. S

I OO ' P

75

right side portion as a free body in eqUilibrium as shown in Fig. A5.2. For static equilibrium, ~V, ~H and ZM rouEt equal zero for all forces and moments acting on this beam portione ConsiderIng ZV = 0 In Fig. A5.3: ZV = 75 - 100 = - 25 lb.

- - - - - -

(1)

A given beam is subjected ta a certain applied known loading. The beam reactions to hold the beam in static equilibrium are then calculated by the necessary equations of static equilibrium, namely: ZV = 0, or the algebraic summation af all vertical forces equal zero.
ZH

thus, under the forces shown, the force system is unbalanced in the V direction, and therefore an internal resisting force Vi equal to 25 lb. must have existed on section a-a to produce equilibrium of forces in the V direction. Fig. A5.3 shows the resisting shear force, Vi 25 Ib, which must exist for equilibrium. Considering ZM = in Fig. A5.3, take moments about seme point a on section a-a,

= 0, or the algebraic summation of all horizontal forces equal zero. moments equal zero.

ZM,

=- 75

x 15 + 100 x 5

=- 625In.lb.(2)

ZM = 0, or the algebraic summation af all the

~ith the entire beam in static equilibrium, it follows that every portion at ~he beam ~ust likewise be in static equilibrium. Now consider the beam in Fig. A5.1. The known applied load of P ::ill 100 lb. is held in equt Itbr-tua by the two reactions of 25 and 75 Ibs. as shown and are calculated from simple statics. (Beam weight is neglected in this problem). New consider the beam as cut at section a-a and consider the

or an unbalanced moment of - 625 tends to rotate the portion of the beam about section a-a. A counteracting reststing ~oment M 625 must exist an section a-a to provide equilibrium. Fig. A5.4 shows the free body with the Vi and !'Ii acting. Now ZH must equal zero. The exterr.al forces as well as the internal reSisting shear Vi have no horizontal components. Therefore, the internal farces prodUCing the resisting moment Ml must be such as to have no horizontal

AS.I

A5.2

BEAMS -- SHEAR AND MOMENTS 1000 lb. 300 lb.

6, ,7
6 '1'7

unbalanced force 1 which means that the resisting moment M i in the fo~ of a couple, as shown in Fig. AS.5, or M i = Cd or Td and T must equal C to make ZH O. The tendency of the loads and reactions acting on a beam to shear or move one portion of a beam up or down relative to the adjacent portion ot the beam is called the External Vertical Shear, or commonly referred to as the beam Vertleal Shear and is represented by the te~ V. From equation (1), the Vertical Shear at any section at a beam can be defined as the algebraic sum of all the forces and reactions acting to one side of the section at which the shear is desired. It the portion of the beam to the lett of the section tends to move up relative to the right portion, the Sign of the Vertical Shear is taken as positive shear and negative if the tendency is opposite. Or in other wordS, if the algebraic sum of the forces is up on the lett or down on the right Sids, then the Vertical Shear is positive, and negative tor down on the left and up on the right. From equation (2), the Bending Moment at any section at a beam can be detined as the algebraic sum of the moments of all the forces acting to either side of the section about the section. It this bending moment tends to produce compreSSion (shortening) ot the upper fibers and tension (stretching) ot the lower' fibers' or the beam, the bending moment 1s classed as a positive bending moment, and negative tor the reverse condition.
AS.4 Shear. and Moment Diagrams. In aircrart design, a large proportion

.,[ 5

B,'
B'I .-!

I' 1 1"--4'

-+- 5'~5'
2 '3 RA

-----L.... 3'

= 1110 lb.

RBi", 690 lb.:

Fig. A5.6

Calculations for Shear Diagram: We start at ~he lett end of the beam. ConSidering a section just to the right of the 500 lb. load, or section 1-1, and ~onsldering the portion to the left of the section, t~e Vertical Shear at 1-1 = ZV =_SOO (negative, down on left.)
+

610 lb.

_ 500 lb.

~
_ 500 lb. - 390 lb. - 390 lb.

.,~

610 lb.

~b

3001b,

Fig. A5.7 (Shear Diagram)

~in.lb.

~
ZV - -

~O
-900 in. lb.

-2000 in. lb. Fig. AS. 8 (Bending Moment Diagram)

Next, consider section 2-2, just to left of reaction RA'

sao,

or same as at section 1-1.

ri&~t

ot the beams are tapered in depth and section, and also carry a variable distributed load. ThUS, to design or check the various sections of such beams, it is necessary to have a complete piCture as to the value ot the vertical shear and bending moment at all sections along the beam. It these values are plotted as ordinates trom a base line, the resulting curves are referred to as Shear and Moment diagrams. A tew example Shear and Moment diagrams will be plotted, to retresh the' students kncw.l edge regarding these
diagrams.

Next, consider section 3-3, just to RA' ZV 500 + 1110 left side at section).

of

===-

=610

(pOSitive, up on

Next, consider section 4-4, just to left ot 1000 load.

ZV 500 + 1110 = 610 (same as at section 3-3). Section S-S, to right of 1000 load: ZV 500 + 1110 - 1000 390 (down

on lett).

=-

Example Problem 1. Draw a shear and bending moment diagram tor the beam shown in Fig. AS.6. Neglect the weight at the beam. In general, the tirst step is to determine the reactions. To find RB, take moments about point A. ZMA = - 4 x 500 + 1000 x 5 + 300 x 13 - lORS = a hence RB 690 lb. ZV 500 + RA - 1000 - 300 + 690 = 0 hence RA 1110 lb.

Check this shear at section S-5 by using the portion of the beam to the right of 5-5 as a free body. ZV 300 + 690 = 390, which checks (Sign of shear is minus, because ZV is up on right). Section 6-6, use the portion to right as a free body:

=-

ZV = - 300 + 690 = 390 (minus Shear). Section 7-7: ZV

=-

=- 300

(positive shear, down on right)

ANALYSIS

AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A5.3

Section 8-a: ZV = - 300 (positive shear).

fig. AS.7 shows the plotted values on the shear diagram.

V x = 270 - lOX, and hence, the shear de_ creases at aconstant rate of 10 Ib./in. frem 270 at A to 180 at C. The vertical shear at section D, just to the right of load is, VD = ZVleft = 270 - 10 x 9 - 120 = 60 up, or positive.

Calculation of the Moment Diagram. Start at section 1-1, and consicer the forces to the left only:
ZM

=- 500
= - 500

x 0

=0
A

Since sections 2-2 and 3-3 are only a differential distance apart, assume a section just above RA and consider the forces on the left side only:
ZM
x 4

Ij j

9" -1'20hP

c+n
j

w=10Ib./in.
'3: I I I
j

I l

1 1

i
I

RA=270 270 lb.

= - 2000

in. lb. (Negative

Fig. AS.9
'n-,.,._Vcs180 lb. V =60 lb.

RB=210

moment, because of tension in the top fibers). Consider the section under the 1000 in. lb. load: ZM to left = - 500 x 9 + 1110 x 5 = 1050 in. lb. (positive moment, compressing the top fibers). Check by considering the forces to the right: ZM right = 300 x S - 690 x 5 = - 1050 in. lb. Next, consider a section over RB:
ZM right

Fig. A5.l0

A;~

Fig.A5.~.~ ~
Moment Diagram

~ 15"

1--6"
Shear Diagram -210 lb.

moment, tension in top fibers). At Section 8-8:

ZM right

= 300

x 3

= 900

in. lb. (Negative

The vertical shear be~Neen points D and B, when x is the distance of any section between D and B trom A:

= 300

x 0

=0

VDB

=270

- 120 - lOX - - - - - - - - - -

(1)

Fig. AS.S shows the plotted values. From the above results it may be noticed that when the bending moment is obtained fram the forces that lie to the lett at any section, the bending moment is positive when it is clockwise. If obtained fro~ the torces to the right, it Is positive, when counter-clockwise. The student should sketch in the apprOXimate shape of the deflected structure and determine the signs from whether tension or compression exists in the upper and ~ower fibers. Example Problem 2 Calculate and draw the shear and moment diagrams for the beam and loading as shown in Fig. A5.9. First, dete~ine the reactions, RA and RS:
ZMA

At point BJ x = 36: hence VB = 270 - 120 - 10 x 36 = - 210 lb., which checks the reaction RE' Since the Vertical Shear decreases at a rate ot 10 Ib/in. from D to B, it will be 6" fram D to a point where the shear is zero, since the shear at D 1s 60 lb. This point could also be located by equating equation (1) to zero and solving for x as follows: o ~ 270 - 120 - lOx, or x ~ 1 = 15" frem
00 15
If the shear diagram has passed through zero under the concentrated load, then the method of equating the shear equation to zero and so~ving for x could not be used, thus in general, it is best to draw a shear diagram to find when shear is zero. Fig. AS.IO shows the plctted shear diagram.

A.

= 36 x 10 x 18 hence RS = 210 lb.

1-

120 x
+

9 -

36RB =

O.

ZV = - 120 - 36 x 10 hence RA = 270 lb.

210

1-

RA ; O.

Shear Diagram: The vertical shear just to the right of the reaction at A is equal to 270 up, or positive. This is plotted as line AE in Fig. A5.10. The vertical shear at section C just to the left of the load and considering the forces to the left cf the section = 270 - 9 x 10 = 180 lb. up, or positive. The vertical shear for any section between A and C at a distance x from A is:

Moment Diagram: At section A just to the ri~~t of RA the bending moment, considering the the left, is zero, sinCe the arm of RA The bending moment at any section A and C, at a distance x from the lett RA, is,

reaction torces to is zero.

be~Neen

reaction
(2 )

In equation (2), x can not be greater than

9.

A5.4

BEAMS -- SHEAR AND MOMENTS

The equation for the bending moment between D and B (x greater than 9) 1s

Mx = RAX

- P (x-s ) -

wx
2""

(3 )

= 270 x - 120 (x-g) - lOX' ---2-

= 1080

150 x -

5x~

- - - -

(4 )

At section C, x = 9", substitute in equation (4)

Me = 1080 + 150 x 9 - 5 x 9~ = 2025 in. lb. (positive, compression in top tibers).

At the paint or zero shear, x 15". n = 1080 + 150 x 15 - 5 x 15~ = 2205 Thus, by substituting in equation (2) and (4) the moment diagram as plotted in Fig. A5.l1 is obtained.
AS.5 SectioD. of Muimum Bending Moment.

Thus, the area at the shear diagram between any two points equals ::le Change in ~ending Jloment between these ~NO points. To illustrate this relationship, 80nsider the shear diagram in example problem 2 (Fig. A5.l0). The change in bending moment between the left reaction RA and the load is equal to the area of the shear diagram between these ~NO points, or 270 + 180 2 x 9 = 2025 In. lb. Since the bending moment at the left support is zero, this change therefore equals the true ~oment at a section under the load P. Adding to this the area at the snall triangle between point D and the point of zero shear, or ~o x 6 = 180, we obtain 2205 in. lb. as the max~um ~oment. This can be c~ecked by taking the area of the sh6a~ diagram between the point of zero shear and point 3 = 2~0 x 21 2205 in. lb.

The general expression for the bending moment on the beam at example problem 2 is from equa t Ion (3):

Now, the value of x that will make

ma.x1mum or minimUm is the value that will make

Mx

the first derivative ot M x with respect to x equal to zero, or

(5 )

Example Problem 3. Fig. A5.12 illustrates a landing gear oleo strut ADEO braced jy struts BD and CEo A landing ground load of 15000 lb. 1s applied th:cugh the wheel axle as shown. Let -1 t be required to find the axial load in all ~embers and the shear and bending moment dia~ for the oleo strut.

Therefore, the value or x that will :nake l1x a maximUm or minimum may be found trom the equation

vcr-

15.0

He.....-.e- HA
But, observation or this equation indicates that the term Ra - P - wx is the shear tor the section at a distance x tram the lett reaction. Therefore, where the shear is zero, the bending moment is maximum. Thus, the shear diagram which shows where the shear is zero is a convenient medium tor locating the points of maximum bending moment.
AS.6 Relatian Between Shear and Bending Moment.

-.rrx.o
B
, I";

-j-5.77iVB

HB

T
j

I
D

10"
i

~<f'
Fig. AS. 12

Resistance at Point
:II

~~~.at~~~~~t~~al
~

t
I
I

I
16"

I I

15000 lb.

--I I
16"

Equation (5) can also be written :-

V,

since the right hand portion of equation (5) is equal to the shear. Hence, dM = Vdx - - (6 )
SOLUTION: -

"0"

Which means that the difference dM be~Neen the bending ~oments at ~NO sections that are a distance dx apart, is equal to the area Vdx under the shear curve between the two sections. Thus, for two sections Xl and Xu

To find Vc take moments about point 5, 2MB - - 15000 x 0.5 x 42 + 15000 x 0.866 x 5.77 + 20.77 V c = O. hence, Vc ~ 21550 lb. The axial load in member CE therefore equals

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

AS.'

11550/eos 30 or 13330 lb .'. He = 11550 x 15/26

= 6660
To find HA take moments about point D, ... 10 HA = O. hence, HA
ZMD

Bracket _____

.'
Bracket
1 00 lb.

~[;,""''l--500

E ,:,

II

11550

x 15 - 6660 x 10 - 15000 x 0.5 x 32

= 13340

lb.

To find VB take ZFV

=0,
Fig. AS. 14

hence, VB = 24550 lb. The axial load in member ED therefore equals 24550/cos 30 = 28360 lb. (compression). The reaction H B therefore equals 28360 x sin 30
14180 lb.

ZFv 15000 cos 30 ... 11550 - VB - O.

SOLlJrION: -

Fig. A5.13 shows the oleo strut as a free body with the reactions at A, D and E as calculated. Fig. A5.l3 also shows the axial load, vertical shear and bending moment diagrams. The bending moments due to applied loads without regard to bending'deformation of the beam are usually referred to as the prtmary bending moments. If a member carries axial loads additional bending moments will be produced due to the axial loads ttmes the lateral deflection of the beam, and these oe, :ing moments are usually referred to as seC)ndary bending aomerrts , (Arts. A23-30 covers t.he calculation of secondary moments).
13340

Calculations of reactions at A and B: To find VB take moments about paint A, :MA 500 x 7 - 500 x 6 ... 1000 x 20 ... 1000 x sin 45 x 10 + 1000 cos 45 x 2 - 22 VB = O.
hence, VB 999.3 lb. (up).

=-

zv = 999.3

To r Ind VA take ZV = 0, - 1000 - 1000 sin 45 Q 500 + VA hence, VA = 1207.8 lb. (up). To r inc HB take ZH = 0, ZH = - 500 ... 1000 cos 45 - H B 0, hence

= O.

HB 207.l.

6660

ic= 11550
16

-r.
1
133~O

!L:K:J50

I I

1O '

-1

14180

~ wfl0//@~
-13000 lb.

24550

Axial Load Diagram

e-e.Compz-eaefon

With the exception of the 1000 lb. load at 45, all loads are applied to brackets which in turn are fastened to the beam. Therefore the next step is to find the reaction of the loaded brackets at the beam centerline support points. The load at E and the reaction H g at B will be also referred to beam centerline. Fig. A5.I5 (a,b,c,d) show the cantilever brackets as free bodies. The reactions at the base of these cantilevers will be determined. These reactions reversed will then be the applied loads to the beam at points C, D, F and E.

F7/?/?/'?//hi'T,;==:z:z:J~~~:am ~:84b
-7500 lb.

2'j
50 ,.

~~'Moment Diagram
Fig. AS. 13 -120000 -133440 in. lb.

Bending

~OOO f..l C:~rib=50

Fig. a

!I
8"

,o-rl
8"

j)

L MlY'40ci~P"'500
Fig. b

... 1000 5"

HF"'500 VF"'I000 FIg. c

L~~'O'.1.
~~.,
EFig. d

000

,,<;:)

,,-:.

,'.

F1g. AS.15

Example problem 4. Fig. A5.l4 shows a beam loaded with both transv~rse and longltudional loads. This beam loading is typical of interior beams in the airplane fuselage which support all kinds of fixed equipment. The reactions ~or the beam are at points A and B. Required: - Shear and bending moment diagrams.

For bracket at C, to find He take ZH = 0, or obviously H C ~ 500 lb. In like ~er use Zv = a to tind V c = 5CO lb. To tind Me take moments about point C. :Me 500 x 2 + 500 x 8 - Me = 0, hence Me = 3000 in. lb. The stUdent should check the reactions at the base of the cantilever brackets at D and F (See Fig. h,c).

=-

AS.'

BEAMS -- SHEAR AND MOMENTS

The load of 1000 at 45 and applied at point E"

will be referred to point E the centerline of beam. Fig. d shows the reaction at E due to the load at E'. The reaction at B should also be
referred to the beam centerline. Fig. A5.16

shows the beam with the applied loads at points C 0 E' F and 8'. Figs. AS.17, 18 and 19 show the axial load, vertical shear and bending moment diagrams under the beam loading of Fig.
A5.l6.
500 3000

have rraximum peak moments without the Vertical Shear paSSing through zero. To illustra:e this tact, consider the bean of ?~g. A5.20, namely, a Simple supported beam with an externally applied couple ~o~ent of 10 1n. Ij. magnitude at paint C the center point of the beam. The shear and bending moment dlagr~~ are as ind~cated and a maximum bending ~ome~t occurs at C but the shear diag~am does not pass through zero.
,....-..,MC: 1O " -II'

707.1
4000 -0
1207.8

'R:l
Fig. AS.16

'e '
10"

t,
I Shear Dia.

-a l
Axial Load Dia.
Fig. AS.20

-500 lb. Fig. AS.l7

Fig.A'.lS

-~

----d
707. S,-,lb",,-. _
b l

L_ _

--'-----"-_I__ ...:I-=20=7.~l a
,

-707.1

'"

~"-II'

Bending Moment Dia.

==~O. 7

[I5788.7

Shear Dia.

-999.3----'
Bending

A couple is two equal and opposite forces not in the same straight lir.e. Let it be assumed that the 10 in. lb. couple is made up of torces equal to 100 lb. each and an a~ Jetween them at 0.1 inch as illustrated in Fig. A5.21.

4285
3000

Moment Dia.

Fig. AS. 19

-207.1

'1

1~ ~O 1"
e

lIb.

11
Fig. AS.21

U 99

Shear Dia.

The shear diagram Is determined in the same manner as explained betore. The applied external couples do not enter into the vertical shear calculations. The hending moment diagram can be calculated by taking the algebraic sum ot all couples and moments ot all torces lying to the one side or a particular section. It it is desired to use the area at the shear diagram to obtain the bending moments, it is necessary to add the couple moments to the shear areas to obtain the true bending moment. For example, the bending moment just to the lett ot point E will be equal in magnitude to the area at the shear diagram between C and E plus the sum or all aoplied couple moments between C and E but not inclUding that at E. To illustrate the calculations are: _ (- 500 x 5) + (707.8 x 10) 4578 In. lb. (tram area ot shear diagram). (3000 - 4000) = - 1000 In. lb. (tram sum or couple moments). Thus bending moment at E 4578 - 1000 = 3578 In. lb. lert

The shear diagram is as shown in Fig. A5.21 and now passes through zero 11TIder each of the couple forces. Thus i! we assume the couple ~oment has a dx arm the shear to the right of C is one lb. and then changes to some unknown negative 'falue and then back to one lb. pOSitive as the distance dx is covered in going to the lett. Thus the shear goes to zero twice in the region of paint C.
A5.7 Moment Di2iTams as Made up 01 Parts.

The bending moment at Bright will equal that at Eleft plus the couple moment at E or 3578 + 707

4285 In. lb.

~oments

The student should realize that When couple are applied to a beam it is POSSible to

In calculating the deflection of statically deter.ninate beams (See Chapter A7) and solving statically indeter.ninate structures (See Chapter AS), the area under the bending ~oment curve is required, thus it is often convenient to treat each load and reaction as a separate acting force and draw the moment diagram for each force. The true bending moment at a particular paint will then equal the algebraic summation of the ordinates of all the various moment curves at this particular point or adding the various separate moment diagrams will give the true bending moment diagram. Figs. A5.22 and A5.23 illustrate the drawing of the bending moment diagram in parts. In these examples, we start tram the lett end and proceed to the ri~~t end and draw the moment curve for each force as though the beam was a cantilever with the fixed

ANALYSIS AND DESIGN OF

0\'0'

FLIGHT VEHICLE STRUCTURES

A5.7

~ ~\.Q
I' \.

r"+5"f5"~
r - - - - 15"------l R ..",100 lb. Rll""lOO Due to il, + 11500 in. lb. Due to -====:::=-::-=4000

.;z '" ,0 ~

P.",10*"

w; 10

Ib/ in.
R lI"'4.8

1t

t t

r2''-t--- 10"

* *~

Due to P

R~",62

Due to Rl

p.

Due to Pli

500
........--Fig. A5.-22

Final Moment Dia.

--

Fig. AS. 23

support at the right end. The final bending moment curve for the true given beam then equals the sum of these separate diagrams as illustrated in the figures. STATIC MOMENT CURVES IN SOLVING STATICALLY
INDErERMlNATS STRUCTURES

The usual procedure in solving a statically structure 1s to first make the structure statically determinate by removing the necessary redundant or unknown reactions and then calculating the deflection of this assumed statically determinate structure as one step in the overall solution of the problem (See Chapter AS). In the solution of such structures it is likewise convenient to treat the bending moment diagram as made up of parts. To illustrate, Fig. A5.24 shows a loaded rectangular frame fixed at paints A and B. The reactions at both
Indete~1nate

at B and thus leaVing only 3 unknown elements of the reaction at B. Fig. A5.25 shows the bending moment curves for each load acting separately on this cantilever fr~~e. Fig. A5.26 shows the tr~e bending moment as the summation of the various moment curves of Fig. A5.25. As another solution ot this fixed ended :rarne, one could assume the statically determinate modification as a frame pinned at A and pinned With rollers at B as illustrated in Fig. A5.27. This assumed stricture is statically ceterainet;e because there Pa ::rIO are only 3 unknown elements, namely the magnitude and di!!. =10 rection ot the reaction at A and the magnitude ot the re- P..",10 action at B. For convenience the reaction at A is Pin~HA"'20 B resolved into two magnitudes as H and V components. The ""A"'10 B::r20 reactions VA' H A and VB can Fig. AS. 27 then be found by statics and the results are shown on Fig. A5.27. Fig. A5.28 shows the bending moment diagram on this frame due to each load or reaction acting separately, starting at A and going clockwise to B. Fig. A5.29 shows the true bending moment diagram as the summation at the separate diagrams.
-=::;::::::I-50-Due to P a -60 - 60 o ?'--'"' .....1J.e v t->" Due to V ..J~I00
y

Due to P li Due to P , -601

--=::::J

-50
Fig. A5. 28
'i

..1
-6

'" .I

-60 Fig. AS. 24 -110 .110 -80

j~nu"~
Free
60 - 50 90

\-60

~"r-_,,'O
30 in.

180

lb.

FIg. AS. 25

A
Fig. AS. 29

Final Bending Moment Diagram (Tension on inside of frame is positive moment)

100

...L

5"

A5.8 Forces at a section in Terms of Forces at a Previous Station.

STRAIGHT BEA11S

paints A and B are unknown in magnitude, direction and location, or each reaction has 3 unknown elements or a total at 6 unknowns for the two reactions. With 3 static equilibrium equations available, the structure is statically indeterminate to the third degree. Fig. A5.25 illustrates one ~er in which the structure can be made statically dete~nate, by freeing the end A to make a bent cantilever beam fixed

Aircraft structures present many beams which carry a varying distributed load. tlinlmum structural weight is at paramount importance in aircraft structural design thus it is desirable to have the complete bending moment diagram for the structure so that each portion of the structure can be proportioned efficiently. To decrease the amount of numerical work required in obtaining the complete shear
~.

'( i:t" ,

A5.8

BEAMS -- SHEAR AND MOMENTS

~nd bending ~oment ciagr&~ it usually saves time to ex?ress the shear and ~oment at a given station in terms of the shear and moment at a previous station plus the effect of any loads lying bet~een these two stations. To illustrate, ?i~. AS.3D shows a cantilever beam carrying a considerable ~umber of transverse loads F of different ~~ltudes. Fig. A5.31 shows a free body

H1
Q

~
M1

,/

I
,

r
V,

!- -

--I
d----o-;

MlC

(1)

Lr~v "OM"
(2)

Fig. AS. 32

Fig. AS. 33

Fig. AS.31

at the beam portion 8evNeen stations land 2. The Vertical Shear V1 at stat~on 1 equals the summation of the forces to the lett of station 1 and M 1 the bending moment at station 1 equals the algebraic sum of the moments at all forces lying to lett of station 1 about station 1. Now conSidering station 2: - The Vertical Shear V" = Vl + F 1 _ " , or stated in wordS, the Shear V" equals the Shear at the previous station 1 plus the algebraic sum of all forces F lying between stations 1 and 2. Again considering Fig. A5.31, the bending moment M" at station 2 can be written, MOl = M l + Vld + Fl_"a, or stated in words, the bending moment tl" at station 2 is equal to the bending moment Ml at a previous station 1, plus the Shear V at the pr-evious station. 1 t1.mes the arm c , the distance between stations 1 and 2 plus the moments at all forces lying be~Neen stations 1 and 2 about station 2.
AS.9 Equat.1an.s for Curved Beams.

Then from Fig. A5.33 we can 'Nrite for the resultant torces and moment at point (2) at station 2: -;

Having the resultant forces and moments for a given point on a given station, it is usually necessarJ in :inding beam stresses to resolve the forces ~nto components no~~l and parallel to the beam cross-section and also tr~sfer their location to a point on the neutral axis of the beam cross-section. For examp ;e Fig. A5.34 shews the resultant

Many structural beams carry both longitudional and transverse loads and also the beams may be made of straight elements to r orn a frame or all beam elements may be curved to form a curved frame or ring. For example the airplane tuselage ring is a curved beam SUbjected to torces of varying magnitude and direction along its boundary due to the action or the fuselage skin torces on the frame. Since the complete bending moment diagram 1s usually desirable, it is desirable to minimize the amount of numerical work in obtaining the complete shear and bending moment values. Fig. A5.32 shows a curved beam loaded with a number of different vertical loads P and horizontal loads Q. rig. A5.33 shows the beam portion 1-2 cut out as a treo bOdy. Hl represents the resultant horizontal force at statton I and equals the algebraic summation ot all the Q forces to the left of station 1. Vl represents the resultant vertical torce at station 1 and equals the sum of all F :orces to lett of station 1, and M equals the bending l moment about point (1) on station I due to the moments of all forces lying to the left of point 1.

sP W
v

" "J6

~~
s
Mo.:Ml wNe
Fig. AS. 36

Fig. AS. 34

Fig. AS. 3S

forces and moment at point 1 of a oe~ crosssection. They can be resolved into a normal terce N and a shear for S ~lus a moment M as l shown in Fig. A5.35 where,

N = H cos a + V sin a S = V cos a - H sin a

Later on when the beam section is being des igned it my be round that the neutral axis 11es at point 0 instead of point 1. Fig. A5.36 shows the forces and moments referred to point 0, With M Ne. l o being equal to M

ANALYSIS AND DESIGN OF

AS.10 Tw:sional Moments.

FLIGHT VEHICLE STRUCTURES

A5.9

The loads which cause only bending of a beam are located so that their 11ne at action passes through the flexural axis of the beam. QUite otten, the loading on a beam does not act through the flexural axis of the beam and thus the beam undergoes both bending and twisting. The moments which cause the twisting action are usually referred to as torsional moments. The airplane wing is an excellent example of a beam structure that is subjected to combined bending and torsion. Since the center ot pressure of the airtoil torces changes With angle ot attack, and since there are many flight conditions it is impossible to &liminate torsional moments under all conditions ot flight and landing. For the fuselage, the Vertical tail surfaces is normally located above the fuselage and thus a load on this tail unit causes combined bending and twisting of the fuselage. Fig. A5.34 illustrates a cantilever tube being subjected to a load P acting at point A on a fitting attached to the tube end. The flex-

Fig. AS.34

ural axis coincides with the tube centerline, or axis 1-1. Fig. AS.35 shows the load P being moved to the paint (0) on the tube axis 1-1, however the original force P had a moment about (0) equal to Pr, thus the moment Pr must be added to the load P acting at (0) it the force system at point (0) is to be equivalent to the original torce P at pOint A. The torce P acting through (0) causes bending without twist and the moment Pr causes ~Nisting only. For the resolution of moments into various resultan~ planes of action, the student should refer to any textbook cn statics.
AS.11 Shears-and Moments on Wing.

Arts. A4.S and A4.6 of Chapter A4 discusses the alrloads on the wing and the equilibrium of the airplane as a whole in flight. As expla~n ed, it is customary to replace the distributed air forces on an airtoil by two resultant torces, namely, lift and drag forces acting through the aerodynamic center ot the airfoil plus a wing ~ament. The airflow around a wing is not uniform in the spanwise direction, thus the airfoil force coefficients CL, Co and CM vary spanwise along the wing. Fig. A5.36 shows a typical spanwise variation of the CL and CD torce coefficients in terms of a uniform spanwise variation C L and CO' Any particular type of airplane 1s designed to carry out a certain Job or duty and to do that Job requires a certain ~lmum airplane

velocity With the maneuvering Ilmited to certain maxtmum accelerations. These limiting accelerations are usually specified with reference to the X Y Z axes ot the airplane. Since the directions of the lift and drag forces change with angle of attack it is simpler and convenient in stress analysis to resolve all forces with reference to the X Y Z axes which remain tlxed in direction relative to the airplane. As a time saving element in wing stress analysis, it is customary to make unit load analysis for wing shears and moments. The wing shears and moments for any deSign condition then follows as a matter of simple proportion and addition. For example it is customary: (1) To assume a total arbitrary unit load acting on the wing in the Z direction through the aerodynamiC section of the airfoil section and distributed spanwise according to that of the CL or 11ft coefficient. (2) A similar total load as in (1) but acting in the X direction. (3) To assume a total unit wing load acting in the Z direction through the aerodynamic center and distributed spanwtse a~cording to that of the CD or drag coefficient. (4) Same as (3) but acttng in the X direction. (5) To assume a unit total Wing moment and distributed spanwise according to that of the Cm_ or moment coef~icient. ~.c The above unit load conditions are for conditions ot acceleration in translation of the

A5.I0

BEAMS -- SHEAR AND MOMENTS

airplane as a rigid body. Unit load analyses are also made for angular accelerations of the airplane which can also occur in flight and landing maneuvers. The SUbject of the calculation of loads on the airplane is far too large to cover in a structures book. This subject is usually covered in a separate course in most aeronautical curricula after a student has had initial courses in aerodynamics and structures. To illustrate the type at problem that Is encountered in the calculation of the applied loads on the airplane, simplified problems concerning the wing and fuselage will be given.
AS.12 Ezample Problem 01 Calculating Wing Shears and Moments for One Unit Load. Condition.

Fig. A5.37 shows the half wing plantorm of a cantilever wing. Fig. AS.38 shows a wing section at station O. The reference Y axis has been taken as the 40 percent chord line which happens to he a straight line In this particular wing layout.

aer~mic

center

a.c"

0' ...... Ref.

xL.
I Fig.A5.38

"? Ref.

96"
Axls

The total wing area Is 17760 sq. In. For convenience a total unit distributed load of 17760 1hs. wIll be assumed acting on the halt Wing and acting upward in the Z direction and through the airfoil aerodynamic center. The spanwise distributIon at this load will be according to the (CL) 11ft coefficient spanwise distributIon. For simplicity in this example it will be assumed constant. Table AS.l shows the calculations in table form tor determinIng the (Vz) the wing shear in the Z direction, the bending moment M x or moment about the X axis and My the moment about the Y axis tor a number of stations between the wIng tip statIon 240 and the centerline station O. Column 1 of the table shows the number ot stations selected. Column 2 shows the CLIC L

ratio or the spanwise variation of the lift coeff~cient C L in terms ot a uniform dIstrIbution CL' In Ehis example we have ~aken thIs ratIo as unity sInce we have no wind tunnel or aerOdynamic calculatIons for this wing relative to the spanwise distributIon of the lift force coeffIcient. In an actual problem involvIng an airplane a curve such as that given in FIg. A5.36 would be available and the values to place in Column 3 of Table AS.l would be read from such a curve. Column (2) gives the Wing chord length at each station. Column (4) gIves the wing running load per inch of span at each station Doint. Since a total unit load of 17760 lb. was assumed acting on the half wing and since the wIng area is 17760 sq. in., the running load per inch at any station equals the Wing chord length at that station. In order to find shears and moments at the various station points, the distributed load is now broken down into concentrated loads which are equal to the dIstrIbuted load on a strIp and this concentrated strip load is taken as acting through the center of gravity at this distributed strip load. Columns 5, 6, and 7 show the calculations for determining the (~Pz) strip loads. Column 8 shows the location of the ~Pz load which is at the centroId ot a troplzcidal dIstrIbuted load whose end Yalues are given in Column (4). In determining these centroid locations it is convenient to use Table A3.4 of Chapter A3. The values of the shear V z and the moment M x at each station are calculated by the method explained in Art. A5.S. _Columns 9, 10, 11 and 12 of Table AS.l gIve the calculations. For example, the value of M x = 9884 in Column (12) for statIon 220 Bquals 2436, the M x moment at the preViOUS station in Column (12) plus 4908 in Column (10) which Is the shear at the preVIous statIon (230) tImes the distance 10 Inches plus the moment 2540 in Column (9) due to the strip load be~Neen statIons 230 and 220, which gives a total of 9884 the value in Column (12). The strip loads 6Pz act through the aerodynamic center (a.c.) ot each airfoil strip. Column (13) and (14) give the x arms which is the distance trcm the a.c. to the reference Y axis. (See Fig. A5.38). Column 15 gives the My moment for each strip load and Column 16 the My moment at the various stations which equals the summatIon of the strip moments as one progresses fram station 240 to zero. Fig. A5.39 shows the results at station (0) as taken from Table A5.1.

Fig. A5. 39

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

TABLE A5.1 CALCULA -nos OF WING SHEAR Vz AND WING MOMENTS My AM) My DUE TO TOTAL UNIT DISTRIBUTED HALF WING LOAD OF 17760 LBS., ACTmc; UPWARD m Z DIRECTION AND APPLIED AT AERODYNAMiC CENTERS OF WING SECTIONS (See Figs ~ 37 38 for Wing Layout)
10

A5.11

,
C L/_

11

12

ra

15

15

CL
II

Ratio,
Assumed Unity.

g.

7.2
7.28
605

~;:

Q,,"''S
'40

~g

.. "" . i ... . s-: .. . :;>-:8

>
~

.u >~

~g

~S...;

~.

"

~L

-49.0

1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0

1.0

"

-48.54 49.63

>.0

<2.
48.1 512.7 810.0

o
2.49 242.7 2.49 -490.8 1213 4908 15052 27020 40087 53 68077 83782 99015 115579

o
60s
618

o
1787 1767 3613 7553 14093 21483 29743 38913 49013
50113

49.09 .0 10.00 1 .00 51.27 4.97

7.36 7.53 7.85

230 220
205

50.18 52.36 55.64 58.92 62.20 6 ., 68.78 72.00 75.23

50.18 52.36 54.00 55.6'" 57 27 7.4

2438 9884 309S8 64488 111313 17 1

7."
7.89

1846

1003.5 1813.5

2540 6020 6490 6760

11. 09
8.34 8.85 9.09 3580. 9 9.33 10.06 5545.5 7480 7840 8200 8570 8940 9310 9660 10000 1 00 247683 338705 445920 570060 702949 863174 1042534 1241744 1539 4 10.31 2672.5

'940
65'iO 7390 8260 9170 10100

8 9
15.00 1.00 1 .00 1.00 15.00 15.00 15.00 1.00 1 .00 1007.0 1055.5 1104. 1153.4 1202.7 1252.2 1300.9 1349.7 1 1920.0 908.4

190
175

58.92 60.56 62.20 .84

7 " 7.43

.4J
4538.5

160 14'

65. '" 68.78

100

67.13 70.39

7.43
7. -43

10.55
6601.0 10. 80 11.04 7705.3 11. 28 11.54 11.80 12.04 10061 132880 150915 169700 189210
9

"0

72.00
73. 7. "'3

11100 11480 13270 14480 12.52 15600 13.01 16930 13.49 18200 14.05 26350

75.23 76.89 7.43 78.55 80.18

71593 84863 99343 114943 131873 150073 176423 214011

100

78.55 81. 81 85.10 88.35

91. 62

1.0 1.0
1.0 1.0

as
70

7. i3
12. 28 12.76 13.27 13.71 14.40 7.43 11313

81. 81 83."'8 85.10

ss o

.4J
12614 7.42 9. 139 4

88.35 89.98 15.00 .00 .00 91. 62

9 1

1.0
1.0

.00
98.00

. 0

ie ,

96.00

10.00

1.0

96.00

14.40 17760 318808 19200 1875630 14.40 27848 Sum : 17760 Cheeks Total Limit Load. Assumed all HaJ! Wing.

l~en the time comes to deSign the structural make-up of a cross-section to withstand these applied shears and moments, the structural deSigner may wish to refer the forces to another Y axis as for ex~ple one that passes through the shear center of the given section. This transfer of a force system with reference to another set of axes presents no difficulty.
SHEARS AND ~OMENTS eN AIRPLANE BODY
AS. 13 Introduction.

The body of an airplane acts essentially as a beam and in some conditions of flight or landing as a beam column which may be also subjected to twisting or torsional forces. Thus to deSign an airplane body requires a complete picture of the shearing, bending, twisting and axial ~orces which may ~e encountered in flight or landi~g. Tn the load analysis for Wings, the direct air

!orces are the major forces. For the body load analySiS the direct air pressures are secondary the major forces being of a concentrated nature in the iorm of loads or reactions from units attached to the bOdy. as the power plant, wing, landing gear, tail, etC. In addition, since the body usually serves as the load carrying medium Unportant forces are produced on the body in re sisting the inertia forces of the weight of the interior equipment, installations, pay load etc As in the case of the wing, a large part of the load analysiS can be made without much consideration as to the structural analysis of the bOdy. The load analysis of an airplane bOdy involves a large amount of calculation, and thus the treatment in this chapter must be ~t a slmpli:led nature, and is presented chiefly fer the purpose of Showing the stUdent in general how the problem of load analysis for an airplane bOdy 1s approached.

I _ ..,(

A5.12

BEAMS -- SHEAR AND MOMENTS

A5.14 Design Conditions and Design Weights.

AS.i5 Body Weight and Balance Distribution.

The airplane body ~ust be designed to withstand all loads ~rom specified ~light co~ditions for both maneuver and gust conditions. Since accelerations due to air gusts vary inversely as the airplane weight, it customary to analyze or check the body for a light load condition for flight conditions. :n general, the design weights are specified oy the government agencies. For landing conditions, hcwever, the normal gross Hei~~t is used since it would be more critical than a lightly loaded condition. The general design conditions which are usually investigated in the design of the body are as follows: Flight Conditions: H.A.A. (High angle of attack) L.A.A. (Low angle of attack) I.L.A.A. (Inverted low angle of attack) LH.A.A. (Inverted high angle of attack) The above conditions generally assume only translational acceleration. In addition, it 1s sometimes specified that the forces due to a certain angular acceleration of the airplane about the air~lane e.g. ~ust be considered. The bOdy is usually required to withstand speCial tail loads both symmetrical and unsym~etrical which may be produced by air ~~sts, engine forces, etc. Also, the body should be checked for forces due to unsymmetrical air loads on the wing. Landing Conditions: In general, the body is investigated ~or the following landing conditions. 'The detailed requirements for each condition are given in the government specifications for both ~illtary and commercial airplanes. LandpLanes : Level Level Three Three landing. landing with side load. point landing. point landing with ground loop. Nose over or turn over condition. Arresting. (Usually for only Navy Carrier based airplanes) .

The resisting inertia ferces Que to the dead weight of the body ~nd its contents plays an important part in the load analysis for the airplane body. ~~en the initial aerodyna~ic and general layout and arrangement of t~e airplane is ~de, it is necessary that a complete weight and balance estinate of the air?lane be made. This esti~te is usually ~ade by an engineer from the weight control ~ectlon of ~he engfneer tng cepar tmerrt who has had experience in estimatin~ :he wei~ht and distribution of airplane units. 7:'i8 estimate which is presented in report fo~, gives the wei;hts and (e.g.) locations of all major air;lane units or instal:ations as well as for many of the minor ~~its which nake up these ~jor airplane assemblies or installations. ~~is weight and bal~~ce report forms the basis for the dead weight inertia load a~alysis which forms an important part ~n the load analysis of the airplane body. The use of this weight and balance estimate will be illustrated in the ex~~ple problem to follow later.
AS.16 Load Analysis. Unit Analysis.

Due to the many design conditions such as those listed in Art. A5.14, the 5eneral procedure in the load analysis of an airplane body is to oase it on a series of unit analJ~es. The loads fo~ any part1cular deSign condition then :cllows as a certain combination of the unit results with the proper ~ultlplylng factors. A Simplified example problem follows which illustrates this unit method 8f approach.
AS.17 Example Problem lllustrating the Calculation of Shears and Moments on Fuselage I:e to Unit Load Conditions.

Fig. A5.40 and A5.4l shows a layout of the airplane body to be USed in this example er-eLem, It happens to be the body of an actua... airplane and the wing used in the preViOUS example problem Nas the wing that Nent with the airplane.

($'-1\"1~=r=~~~I-1
hrust Line

II

Seaplanes or Boats: Step landing With and without angular acceleration. Bow landing. Stern landing. ~NO 'Nave landing. BeaChing conditions. Catapulting Conditions (Navy airplanes) . Special Conditions or Forces: Towing of airplane. BOdy superCharging.

~ c
~

_. - -- -

~,.L

I-~==::::===- 2:T72;-,,---=~:.....===:j1

= o

sc.

_ .jBo!_.!..xi.... - - -

G. (groSS wt. )

11'2"

Fig. AS. 40

:1
'

I
I

Front View Fig. AS. 41 - - - - 20'0"'----_

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

TABLE A5.3

AS 13

Table A5.2 gives the Weight and Balance for the total airplane. This table 1s usually fo~ulated by the ~ei~~t and Balance Section of the engineering department and it is necessary to have this information before the airplane load analysis can be made.
esti~te

WEIGHT AND BALANCE OF AIRPLANE LESS WING GROUP AND INSTALLATIONS IN AND ON WING Ref. Axes: 1a up, z. equal &rill frnm thrust line. III distance from z. Ref . .Ax1s 5" !orwarQ at. (x propeller. .. L8 a.ft.

i
Reference

TABLE A5.2
AIRPLANE WEICHT AND BALANCE

..
~

'.m
1

N=.
PawerpiaDt grOlq) Fuselage groull Tail group Surface COlltro18 'Electrical Iylltem T1I.i1 Whee! gro~ Furn1ab.iDgs Wela:ht

w."",
y

Boriz.

"". \

Vert. (Z) arms Ill_red from thrUst line (+ II up) BorU:. (Xj arms meuu:red froID Z ms 5" (+ is aft) !ot'T.l.rd ~ro-p. ~

;
Vert. Mom,
~

1100 '50

19 113.5

,
er

Bori:l:.

Vert.

Vert.

""'m.
~

.um

""'m.
~

30900
39700 31550 10800 7930 10700 25520 22600 1t197oo

o
1

o
3690 -1190

11'

'.m
No.
1

N=.

w. I
wt.
1100 35'
it

Boriz.

.um
19

Bortz.
Moment
~

Vert.

.um
(Z)

" ~
~

.,

e e

130
136 d55

"
"

38'

121
308 118
191

-19
-10

"
10

'"

10 11
12
z

(X) (iD,) 113.5

s
t

e a

10

11

ra

13

Power Plant Fueelaie Gro~ WiD( Grou-p Tail Grou~ SDrface Controls Electrical System Chu;ais Front Tall Wheel Gro~ FuruilItw1ga Radio Weipt emllty PUot Student Fuel Syllt.em Grosl weight 1

20900
311700 72750 31550 10800 7930 16450 10700

n
1

'50 11'

3150

, , ,

" ae 133 " ", '" '" 131 '" '"

38'
73

-18
-l<

-1:ui00

'50

~
0

"'"'"
Plio<

em~ty

'"

,
Z

'30
- 350 1100

GroSl weigllt "

"""~,

200 200

151

33200
19800 19700

es

"Z ". ~

~
SO,
'00

-52

35

11'

4300

'"

20' '00

151
88

..

'50" 22600 I

"8900 '0200 [
37~00

-12200 -10 35' 5 I 1100 10 21380 _20500

"

_ 1190

'04'

Emllty ~

- '"

With \18efUlload. i

""

169700 78.7"

Z.~.2..00"
BG.O" 2.32"

"'"

2~~~

196OO! " 67600 ) _27

." ."
~40280

"50

of C, G. locations: .CalcuJ.ation GroSI wt. x z 376500/4300 z 98,S" a,ft of Re! .Mia z. z -40280/4300 9.4" below line i
thrust

SOLUTION:
WEIGHT At'IITJ BALANCE OF BODY ITEJ"!S.
~~IGHT

DISTRIBUTION.

Table A5.3 gives the weight and balance calculations for all items attached to fuselage or carried in t~e fuselage, except the wing and items attached to the wing as the front landing gear and the fuel. In order to obtain a close approXimation to the true shears and ~oments on the fuselage due to the dead weight inertia loadS, it is necessary t~ distribute the weights of the various items as given in Table A5.3. Fig. A5.42 shows a side view of the airplane with the center of graVity locations of the wei~~t items of Table A5.3 indicated by the (+) Signs. In the various deSign conditions, the direction of the weight inertia forces changes, thus it is convenient and customary to resolve the inertia forCes into X and Z components. ThUS, in Fig. AS.43, the weights as given in Table A5.3 are assumed acting in the Z direction through their (c.g.) locations. The loads as shown would not give a true picture as to the shears and ~oments along the fuselage, thus these loads should je distributed in a ~nner which should si~ulate t~e actual weight distribution. In ~ost weight and balance reports, the weight items are broken down into considerable ~ore detail than that shown in Table A5.3, which ~kes the wel&~t distribution ~ore evident. The ~erson maKi~g the

weight distribution should study the inboard prOfile drawing of the airplane which shows the general arrangement of all the installations and equtpment , Furthermore, he should study the overall structural arrangement as to its possible influence on fuselage weight distribution. The whole process involves considerable common sense if a good apprOXimation to the wei~~t d1s trlbutlon is to be obtained. Fortunately the large dead weight loadS, such as the power plant, tail, etc. are definitely located, thus small errors in the distribution of the minor distributed weights does not change the overall shears and moments an appreciable amount. In order to obtain reasonable accuracy, the fuselage or body is diVided into a series of stations or sections. In Fig. A5.42, the Sections selected are deSigned as stations which represent the distance from the Z reference axis. The general problem 1s to distribute the concentrated loads as shown in Fig. A5.43 into an equivalent system acting at the 7arious fuselage station pOints. Obviously, if a weight item trom Table A5.3, represents a concentrated load such as a pilot, stUdent, radio, etc., the weight can 08 distributed to adjacent station points inversely as the distance of the weight (c.g.) fram these adjacent stations. However, for a weight item such as the fuselage struct~e (Item 2 of Table AS.3) whOSe e.g. location causes it to tall be~Neen stations 80 and !20 of Fig. A5.43, it would obViously be wrong ~o distribute this weight only to the two adjaCent stations since the wel~~t of 350= 1s for ~he entire fuselage. This weight item of 350~ shou1 4 thus Je 11stributed to all station points. The contrOlling require~ent on this distribution is tr~t the moment of the distributed system about the reference axes must equal the moment of the original Height about the s~e ~~es. Fig. A5.44

A5.14

BEAMS _. SHEAR AND MOMENTS WEIGHT DISTRIBUTION TO FUSELAGE STATIONS STATION 170 200

so

80
1

120
I I

220

260

290
I

315

0 11

50

80

120

170

200

230

250

290

315

I I

I I

I
I

110

! I

Fig. AS.42. Location of weight items of Table AS. 3.

Fig. AS. 46.

Weight items from Table AS. 3 acting in X direction.

Z.
011
I I

50

80

120
1 1

170

200

I
I

I
1- ~ I ....

I I
~

I
T
N

220

260
1

290
I

011
I

50

80

120

170
I

200

230

260

290

315

I
I

--1-r1--ig....

'"
I

I
~;
i

'" ~

~M

l~ <:cl

'"
I

,,I

200 220 260 290

I

Fig. AS.43. Weight items of Table AS. 3 acting in Z direction.

I Fig. AS. 47.

Vertical distribution of fuselage dead weight.

011

SO

80

120
1

170
I I

011
I
I

so
I

80

120

170
I

200

230

260

290

315

I I 80< 780 190 --II

I
Fig. AS. 44.

I
I

i
I

I
6511

j -

----i!-- j

30# 100

" rpl
I
I I

I
90 90 .-r-80~ 3:0# 2':34#

I' I

i
I

Results of fuselage weight distribution to stations.

Fig. AS.48.

Fuselage weight referred to X axis plus couples.

SO

80

120

I
I

I
I

I I

170

I I
311*

200

220

260

290
I

215

I
I
7~#

I I
100

I
I

I '

o 11

50

80

120
I

170

200

230

260

290

315

I
11191

I
I

I
I

I'

I
I

I
',9' 'tf: 12e

409*

I
Fig. AS.45.

I
I

11i93
I

8 758 882 1850 '90 10 '274 - 2 -(_288(;-307 -(-'09-C-311G"C- 10 (_21

Final weight distribution to station points.

Fig. AS.49.

Final weight distribution in X direction referred to X axis plus proper couples.

ANALySIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A5 15

shows how the dead weight of 350i: was distributed to the various station points considering the weights to be acting [n the Z direction. Table A5.4 shows the results of this station point weIght distribution rcr the weight item.s of Table A5.3. The values in the horizontal rows opposite each weight item shows the distribution to the various fuselage stations. The summation at the weights [n each ver-t tca.L column at each station point as given in the third horizontal row rz-cn the bottom or the table gives the final station point weight. These weights are shown in Fig. A5.45 for weights acting in the Z direction. The moment at each total_station load about the Z axiS [s given in the second horizontal row tram the bottom of Table A5.4. The summation at the moments in this row aust equal the total WX moments of Table A5.3 or 219700"*. This check [s shown in the last vertical column of Table A5.4. The distributed system must also be distributed [n the Z or vertical direction [n such a

aanner as to have the same resultant e.g. 10cation as the original we Ight system which [s illustrated in Fig. A5.46. Fig. A5.47 illustrates how the fuselage weight distributed system as shown in F[g. A5.44 [s distributed in the vertical direction at the various station points so that the moment at this systern about the X axis [5 equal to that of the original fuselage weight or 350#. For convenience, these distributed fuselage weights can be transferred to the X axis plus a moment as shown in Fig. A5.48. Table AS.4 shows the vertical distribution at the various items at the various station points. The bottom horizontal row gives the moment about the X axes at the loads at each station point, which equals the indiVidual loads times their Z distances. The summation at the values in this horizontal row must equal the total wz aoment ot Table AS.3. This check is shown at the bottom of the last vertical column. Fig. A5.49 shows the results as gtven in Table AS.4 for the weight distribution in the X direction.
~

TABLE

PANEL POINT WEIGHT DISTRlBtn'lON (Monocoque 1';>98 "J FV.lleia&el

No. : di8U.llce " lro... Ref. '" Az1.s

,.~ i
~o.

N=,

I we, I
I

., .
u

so

i Pow. l'\Il.a.nt "fOUP

roUD

urlace contro
Furni

I ElllCtrical s !item
raUl)

io

ra

Radio U i Pilot

", . , ."
..

i
,

1100

8741
,

m"
II

c I no o

-. ,
,
, ,

.
,

. " , , '", ,
I

i
I
,

ire

aoc

i
i

."" ,
, ,
,

"" ,

"" ,

I
I

.'" ,
,
UOO

to

.
1.67

12$ i
,

""'.m ToW

2555 I 893 .02

"", "",

rs rc

..

,-

, ,

,
I
I

Horlo;o"l::I,I MomlMlt :. ""'. 9820 r Ven.1cal Marnellt

3881 \9390

." "'. ""

24560 I

2.47 ,

."

"

te

,
11 uml
' S'l6O

, -"
,

.10 I
,

au
52800
1

5. J

6.45

10 1.0

1181 19.1

491001

isace

w%..

J: z; :

= dlKtaAce

"I

a
0(

758 ,
Z;

682 1

1650 I

reterence U111 _mell III 5" fonra.rd 01 propeller. disbnce aeeve or beloW tllrust l.1D.8 or x wa.

statiol1 lro.ll;l

""

"""I ic

134200 I

'"

''''

- ""

" "'''

.",

"" .W " "" ". 'oo ,

21W100

...

.OO

5920 I

A5.18 Unit Analysis for Fuselage Shears and Moments.

Since there are many flight and landing conditions. considerable tl.:ne can be saved i t a \mit analysis is made fer the fuselage shears. axial and bending forces. The design values in general then folloW as a sunnnation of the values in the \m[t anaLys t s t imes a ~roper multiplicat~on factor. The loads on the fuselage [n general consists ot tan loads. eng:!.ne loads, Wing reactions, landing gear reactions If attached to ruse rage and tner-t i.a fore es due to the airplane acceleration which ~y be due to botn trrans Ia-. tional and an~ular acceleration of the airplane. For s 1.:nplic1t.y, these loads can be resolved into cGIt::onents parallel to the Z and X axes. To illustrate the unit analysis procedure. a unit analYSiS tor our example problem will be carried out for the following unt t conditions:

(1) Unit direction and (2 ) Unit direction and (3 ) Unit down.

acceleration or load factor tn Z acting up. acceleration or load factor [n X acting rorwarc . tail load normal to X axis acting

Unit analyses are also usua.Ll.y carried out for engine thrust and engine tor-que , side load on ~ll and angular acceleration, but to keep the example calculations from becoming too lengthY only the above 3 unit conditions will be carried out in detail. The others will be discussed in detail [n later paragra phs
Solution for Unit Load Factor in Z Direction.

Fig. A5.50 shows the dead weight loads

A:) 16
act~ng

BEAMS

--

SHEAR AND MOMENTS TABLE A5.5

In the Z direction as 'caken :':rorr::'!'3.ble

A5.4 or Fig. A5 . .:i-5. T:'18 'Nin~ ~s attached to t:le rusejage at s--cations 73 ::ond 116 as s::;'~wn en ?~,. AS.50. The fittings at these points are assumed as designed to cause all the :.rag or reaction in the X ::ilrectlon to be taken of f er.t tr-aj.y at the front fitting on station ';'"3. To place the r ....selage In equt I i or-tum, t'::e wing reaction '1'1111 be calculated:
ZFx
lJ"f"S ta t ; or.

I
I

I
1

FUSELAGE SHEARS AND MOMENTS FOR ONE LOAD FACTOR IN Z DiRECTION (Dead Weight Acting Down)

sta, No.
31' :1 290

ZF z RR:

. - 2555

(Note: 21S700 from Table A5.3)

+

= 0, RH + 0 = 0, nonce RH = 0 0 = 219700 - 116 RR - 73 R!i' = 0-iF +

Moment ..:oM Load or V : shear i ex : Dist. i M' Reaction : L w ! between I = VLlx (lbs.) stations (in. lbs.) (in. lbs.) w

i
I

(;, )

0 22 0 118
0

0
22 25

I
,

0
,

RR 0

--

---

-- (3 )

-
-

22 140 140 161

I
30 30

550

550

Solving equations (A)

and (3 ) for RF and

, 260
230

21
0 10 0 76

- 4200 I - 4750 I I
I

RF

==

1780 lb., RR

= 775

161 171
171 247

lb.
200

- 4830
- 5130

9580

30 30

I
-14710
1

Table A5.5 gives the calculations for the fuselage shears and bending moments 3.t the various station points.
sta.O 11 I I
'0
I I I
I

- -

I
I , , , ,

170 120 116 80 73

60

120
I I I

170
I

200

230
I

260
'

290 315
I I

I
I

I I

I
I I

i
311

I
I
76

I ,
I

~
893

: 19

. -

--.--.: .-t--. --t-->: - - t - . +17" 20"

.~.43"~, Ry=1780 R",775

388

307

409

- +- .

21

VN
I
'

- - 31~ - 40: 0 307

247 558
;0

- 7410
I
4 36J
i
,

-22i20

556 967 967 192 192 499 499 -1281

,
-27900 I -50020
I

I
I
I

* .I

22

0 77' 1 I

3668

-53888 I ,
I

.I

i:i'

'--- ~--- --f'-7::.

RW O Fig. A5.50 STA.73 STA.1l6

,
,

I I

- 6912 - 3493
-29463

-60800
1
I
-64~90
I

7
I

0 1780 388 0 893 0

"" I

23 '0

-1281 - 893
39 - 693

-34827

Solution for Unit Load Factor in X Direction.

II -

-348~
I

Fig. AS.51 shows the panel point dead weight distribution for loads acting in the X direction and att, as taken from Table A5.4 or Fig. A5.40. To place the fuselage in equllibrlum the wi::lg reactions at points (A) and (8 ) will be calculated.

(3) Up on left and down on r-ight side of a section IS pcemve shear. (8) Tension in upper fuselage portion is negative bending moment. (I) refers to aft side of station. - refers to forward side of station. M = M at previous station in col. 6 plus 1M in col. 5

lF x

= 2555

- RH

= 0,

hence

RH

= 2555 lb. (forward)

(A)
II 0 I I 50 80 120

Take momentis about point

lI1A ,. 2555 x 17

5920 - 4, RR = 0, hence RR - 1147.8 (up)

+

I
I
19
I

I
I

I
i

I
I I
I

173

200

230

260
I

290 31'
I I I

(5920 equals the sum of the couples from Table A5.4.

~93t ..(-388(-307 (-409

43"

I,'

;",
758

I i

i
I
I

I
r

662

lF z
hence

= 1147.8 - RF = 0, Rr = 1147.8 lb. (down)

I
I
I

1. R':=t~~ ___.B -4' ---STA.73 STA.116

1650 490 10 2249 274 (-311(-760-10 (- 21 (.....],[3.<:-2 2 220 Fig. AS.51

'\

Rp1l47.8

iRR"'1147.8

Table A5.6 gives the calculations for the shears,

moments and axial loads for the loading of Flg. ,.<>..:>.5....

ANALYSIS AND DESIGN OF

TABLE AS.S fUSELAGE SHEARS, MOMENTS" AXIAL LOADS FOR ONE LOAD .-ACTOR IN X DIRECTION lnerua Loada Acting Alt , , ; a I s a
1

FLIGHT VEHICLE STRUCTURES TABLE A5.7 FUSELAGE SHEARS &; MOMENTS FOR UNIT HORIZONTAL TAn.. LOAD IN Z DIRECTION (Load Acting Down) 1 2 Load or Reaction w lbs. 3 4
~X

A5 17

I
,

...
So.

a ,
w" :

mX ,load. ~

'""

:w::iP":.~w" v z w:
i
,

I
s
I i

Dlr. !ZDlr.

'wa.l LGad

SlIear

"'Mt = r c x Couple i D,S!.. Moment Ibetween ;htloD.

I.::.M>lvt:,,:
1

sac - 2029

6 Moment in. tbs. 0 0 0 0

!
t

1 315

asc : I
2l\(J i

" 11~ I
22.!
2li

'1 cI

,<0 ,<0
161

" aa

c
u
I
0
1

o:
220

!
as
30

22411 1

o'
I

-j ,I
c
I
1

V , shear
: Z w
Ibs. 0 0

22g I

sta.

:: di st, between stations

, - ~M ,, V~X
I ,

-,
, ,
, ,

"

,I
,.,

230 : i 200

o:
10 i
01

'"

- 37~
10

I
30 30 30

2029 c 1_ ,- 2303

315 290
272.5

"
u1

c c

1 1
,

c !
!

2303 : _ 2313 1 '.38031 , ,

1
=1

0 0 0 0 0 -100 0 0 0 0 0 0 0 0 0 0 0 475.6 0 0 0 -375.6 0 0 0 0

0 0

25

.! -

rn

" !
,I
311 : 40IIj o

170 :

oI

I
~

120 _
115

'I

...
358

'" '"

1- 49~
1

i
1

o
c

:!
:i

: I
Q

1147.6!

'I
01
,

60 .

3071

I
1

73

2555'-1147.al 3861

" " "

68~ I '" -IH7.6 , '" '" , _1141.8 sat _1147.8 1274

,

l6Sg

L L
1

- I

0 0
12.5

,
! I
,

~_23131

su
4

01.28031 :- 445 3 1 [- 44531 o !_ 3135 1

Q

0 100
17.5

i
,

260 230 200 170 120 116

ae
41320
Q

iSUS ' - 5135

, ,

- I

100 100

,
100 100

30 30 100 100 30

- rse !
'

1214 -1281
-t281 - a93 - 893

-1147. a i

50 ::

u ! ,

. -

o1

, o I

_4343~ ~: I

~:~:; i ,
43462 i 27 !

a035

"

a9~ I

"I

I
1

o, I

I~

"

I I

1 :' 17 x (2555) : 4343:5 refers to alt aide 01 staJ:ion 1 _ refers to forward side of station 1 Col. te pJ,u lor ten!lion in -fllHlage. i ICoL S) M M iU ~rl!Vlou. station In CoL. 9. ll1u .; Ml of col. 6 :'M2 ill 001. a

~I c : "1 "
c
~l...

1750 - 1750 - 1750 4750 - 3000 - 4750

7750 - 7750 - 3000 I -

I
i

100 100 50 100 100 100 -375.5

-37!5.6 -375.6

3000 5000

10750 -10750 -15750 -15750 -16150 -16150

,..

36 7

I-

I !
i

400

80 73

-I -, ,

13521 2629

- 2629 . 2629
a
0 0 0

Solution for Unit Horizontal Tail Load Acting Down.

-375.6 0

The fuselage shears and :noments will be C8mputed for a unit tail load of 100 lb. on the tail acting in the Z direction, with balanc Ing reactions at tna wing attachment pOints. The center of ;Jressure en the hor:!.zontal tail is at station 277.5. Fig. AS.52 shows the fuselage loading. To r ind wing reactions at (A) and (B) : ZMA = 100 x (277.5 - 73) - 43 RR = 0, hence RR = 475.6'" (up)

23 0 0 39

50 11

-I -i

0 0

0 J

(Col. 8) , M= M at previous station in Col. a plus 6M in Col. 5.

ZF z - 100 + '::'75.6 - RF 0, (down) hence R 375.6# r =

Table A5.7 gives the detailed calculations for the shears and moments at the various station :;)oints.
0 11 50 80 120 170
I
1

CALCtru-TION

MOMENTS ANIl

OF APPLED FUSELAGE SHEARS, AXIAL LOADS FOR A SPECIFIC

FLIGilT CONDITION.
,,"0., J

200
I

230

260
127~.
1

290 315

! I ,
I

I
1

I
I

I
!

I ,
,

I
1 1 1

I 1-'100*

+---------------- - - - ; ~ RH-~~4 --2~-_

I

43" --

R F =375. a

-RR=475.6 srA.116

Using the results in Tables AS.5, AS.6, an d the applied shears and ~oments for a g:!.'l8 n flight condition follow as a m.tter of pr'cpor'-. tion and addition. To illustrate, the applied values ~or one flight condition 'Hill be given. It will be assumed that the aerodynamic calculations for this air?lane for the (H,A.A.) hfgh angle of attack condition ;ave the followtng r-esu Lt s , which the student will aave to accept wt -nout knowrenge of how they were cotameo
.- 7

Fig. A5.52

STA.73

A5.I8

BEAMS -- SHEAR AND MOMENTS

Applied load factor in Z direction Applied load factor in X direction Applied tail load
~

- 6.0
down

1.333

~:t.

110 lb. up.

-:

~c.G-.-

Thus with the load factors in the Z and X directions and the tail load :<nown, 7able A5.S can be !111ed in as illustrated. In a similar manner the values for other flight conditions can be found, the only difference being a new set of ~ultiplying factors since the applied loads would be different.
TABLE A.5.a
APPLIED FUSELAGE SHEARS, !>IOMENTS " AX1A1. LOADS FOR niGHT COIQrrlON I. (H.A.A.)

c
Fig. A5. 53

-----

...
,
,

...

,
0 0

c
c
0

c 0 c c
5220 3740 - 3140 - 5Q85

co

c c

29. 3 29.3
186.8 16 .1:1

5Q1O "

"" . -,
,

~60

e
0

200 :
110

'" .i
116 80 :

73 I

S5-10 11620 1_124640 ; 118ZO 1_12898~ I : 173 0 _ 86855 I 6850 17J2.0 286650 1 :290 - 6850 ' 17770 _ 6850 1 17770 I 2910 _313690 I 1290 47250 mo ,_314640 : 1700 H900 0 40 I 1700 '_385704 1_1710

"'"

"

192~

ns as no 'I: 834~ I aac

921:15 '_ 29650 i _ :;~335 52030

,,,

"

",.,

',:;i~:g: ~~:~

'" '"

"

, .,

_208937 ;-1190 -1190

,.,

Col.. 1 Col. 3 Col. 4 Col. 5 Col. B Col. 7 Col. 8 Col. 9 Col.10

_ 6 :l valu In '01""",,, 3 at Talli. A.5.5. 1.333 '" valu'" ill .00umn 5 01 Tabl .. .405.B. _ -1.10 '" valu" ill .00""",n 3 01 Ubi. A.5.7. eol""",,, (1) oolWIIII (2) olumll (3). 8 '" valUH III .01"""," B af 1':lQI. A.5.5. _ 1. 3;13 :l VlI.1ues In .0IW1111 9 of T:lQI.. -'5.6. _ to 10 '" vaiUH ill .01"""," 6 at TaDl. AS.7. _ col""", ... {B) (7) (8). _ 1.333 '" ..uuu In .01 ....... 14) of Tabl. AS.6.

The resistance to these X ~~d Z components of the grcund react~on R is ?rOvided by ~he inertia forces of the airplane in the X ar.d Z directions Tables AS.S and AS.6 show the tuse l.age shears, moments and axial loads for inert~a loads due to one load factor in the Z and X directions respectively. Thus to obtain the fuselage forces for this given landing condition, it is only necessary to mul.t t p.Ly tne va Iues in these two tables by the ?roper factor an~ add Ute results. Thus fuselage forces due to vertical load factor ot 7 would equal 7 t1mes the values in columns (3) of Table AS.S to obtain shear ~.d 7 times column 6 to obtain bendi~g moment. Likewise the farces due to the 2.98 load factor in X direction would equal (-2.98) ti~es the values in col~s (4)7 (5) and (9) of Table AS.S to obtain axial loads, shears and bending moments respectively. The final or ~r~e for~es woult be the algebraic sum of these results. Landing with
.~~gular

Acceleration

AS.19 Example of Fuselage Shears and Moments for Landing Conditions.

Fig. AS.53 illustrates the airplane in a level landing condition. The ground reaction is assumed to pass the center at landing gear wheel and c.g. of airplane. The fuselage Shears, moments and axial loads are r-ecurr-ed when the vertical ultimate load factor is 7. (Gross weight = 430ot).
SOLtlrION:

In a level UL~ding condition. i t 1s sometimes specified that the horizontal component ot the ground reaction must be a certain proportion of the vertical componeht, which causes the line of action of the grO~~d reaction ~ in rig. AS.53 to not pass tr~ough t~g e.g. of the airplane, which creates an external pitching moment on ~he airplane. This ~oment is usually balanced by the inertia forces cue to the angular acceleration produced by the ..in-. balanced moment about the e.g. The shears and moments on the fuselage cue to this externa; moment could be :aund as ex~la1ned in p~t.
AS.20.
AS.20 Inertia Loads Due to Angular Acceleration.

The vertical or Z component of the ground reaction R is specified as 7 load factors which equals 7 x 4300 = 3010oi. One halt ot this is acting on each Wheel. The horizontal or X comDonent of R is 30100 tan 23 425 x 30100 12800# and acting aft. The horizontal load factor on airplane equals 12800/4300 = 2.98.

In some of the flying conditions, it is specified that the a1rplane ~ust be SUbjected to an angular acceleration as well as translational acceleration. ~is angular accele~ation of the airplane produces :nertia forces which must be calculated if the airplane is to be treated as a body in statiC
samet~es

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A5.19

equilibrium. In some cases, a tail load due to a gust on the tail is specified which produces a moment about the airplane c.g. which produces ~~ar acceleratio~ of the airplane. In certain landing conditions, the ground forces do not pass through the airplane e.g. thus producing a moment about the e.g. which for stress analysis purposes is balanced by inertia forces. Moment of Inertia of Airplane The calcUlation of the moment of inertia at an airplane about the center at gravity axes was explained an page A3.5 at Chapter A3. A detailed example solution was given in detail in Table 6A or Chapter A3. The general equations tor t ne moments ot inertia or the airplane about the reference axes are: Iy = Z
WX ...

Fz

='M:y w Xc ly
y

- - - - - - - - - - - - - -

(1 )

=My w zc I

- - - - - - - - - - - - - - -

(2)

From Table A5.9, I y

= 16097600

My was assumed as 100JOOO hence

,z

" =

100000 w -r.; 16097600 ""

= .00621

w Xc

FX

= .00621 w Zc

Z wz

z A Iy

Ix

=Z wy I z =Z wy

...!. wz
...

Z A Ix
+

Z wx

Z A Iz

The last term in each at the above equations represents the moment of inertia of each weight item about its own centroldal axes parallel to the reference axes.
AS.21 Solution for Inertia Loads Due to Unit 100,000 In. Lbs. Pitching Moment.

To illustrate the general procedure ot determining the balanCing inertia loads when the airplane is SUbjected to an unbalanced moment about the c.g., an ar~alYSls will be made for a unit 100,000 in. lb. moment. Table AS.9 gives the necessary calculations. From kinetics: Pitching angular acceleration a
I rac.zsec

Where Zc and Xc are the z and x distances of wei~~t w to the airplane e.g. Columns No. 9 and 10 of Table AS.9 gives the values 9f these inertia components. Fig. A5.54 shows these inertia loads applied to the fuselage. The reactions at Wing attachment pomts Should be computed and then a table of fuselage shears J moments and axial loads should be made up. This unit table could then be used for all conditions involving angular acceleration ot the airplane. It should be realized that the inertia resisting loads in Table A5.9 are only approxi mately, since the moment ot inertia neglects the centroidal moment at inertia of the big items, SUch as the power plant, wing, etc. The example is only for the purpose) or illustrating the general procedure at determining the inertia resisting loads due to angular acceleration. The same general procedure can be tollowed in considering unbalanced external moments about the Z and X axes, commonly referred to as yawing and rolling moments.

:Z)
I

TABL.E A5.i DJo1.AlCING INERTtA F'ORCES FOR maT 100,000 IN. loB. YOWENT ABOUT Y AXlS TIlROUClll: AIRPLANE C.G. (PrI'CKlKG HOhlEHTJ 1 13j 3 1 4
$

where

l'ly = unbalanced external pitching moment

about e.g. of airplane.

!
10

5
i Arm

= pitching moment of inertia of airplane about airplane e.g. = Zwr

a

I ~~. !~!I Arml Arml I ~. I" Z :XI

110 IJlll 200 ,711 130 : 10 : l 1 O ' 21

Arm

z.,

!x:.J. (~.~l"" ". ~;:~I~;:;'I'

1

I..

IFlI:

The tangential inertia force F for a mass

wig due to an angular acceleration a equals J

F = :!!.. g r-c , but a =

1y g

i.1y

'"
Fual.

290

[4.701 IS. il IU.SO 10.40 142.50 2.1.ta 173.$0 501 :W3.50 19. 10 I 28. 1:r.2. <10.0 315 - 0.1 3:17.50

,UBI
i

~.~!170: 5.4'1.00; 1. 00 1 .so i 13.0GI 250:

e:

n,sl

5 i ~.nn.s! l<u.e I
- 14.3 8:1.$ -1I.:l 1011 \ 1S50 20460 i 30:nO I
JlGU 51100

1.~1I

1.

3.1

2180000

~UOOO

I . '"
i
I

204100 535000
3110000
u"",",

I
,

141.3 1_20.11

hence
F

W'l..!lI 1S01_[., 911_ I.G; 11.50 rroutC18IUi..12351_~' 10;11-41.5 i.<11.50

= Ty 'N

My

710 1 - :n I (4l

1M Zl20
313

<17.51

1.50

i~
11091100

133000 4il5ooo

:5.5 1.1

::: l'"~:: I
51.3 U.3
1

~,

wher-e r

1s

the distance from

I
; C<l1 ........ (2J, (3), 6 C<>lumQ {$)

the weight w to the airplane e.g. It 1s convenient to t~eat the inertia force F as resolved into two components Fx and Fz hence,

T\U<8:Q

Colwall

(6)

I C<l1W1111
~ C<l1W1l11

(ill
(10)

Tabla A:i.3. z., , Z - 9.4 1 9H Tabla A:i.3 for x.., X - 51.5 I loeat1lla F . ooeu "lie F ,oo51l .. Sc

fram

,I,. :

e.,.

A5.20

11

50

aD ,
,

120 ,

170
, ,

230

III.

Fig. AS.62 shows t~e ~lan f8~ sf ~ ca~:~ lever wing. Ass~,e a ccnsta~t ~0~al distributed loac on :~e sur~~ce equal to 50 'cb . /sq. ft. ,";rit.e express i ens ~ cr shear and bending moment en '.'ling and r inc 1[a2.':":'25 at 25, 100, lSO and 200 inches ~~JTI ene .

.1

I
O" 10L

r~

Fig. A5.54 A5.22 Problems

\L~~~==-==:-=::::,~to . _ 1:12.0"
Planform

~O"

Fig. AS. 62

1-----

200"

--.....;

(I) Draw the shear, bending moment and ax:al load diagra~s for loaded structures 1~ Figs. 55 to BOc .
100
rIO

jX

-r-

I
90"

."

AS.63

" s o
'5

1"

200 101

5,0

100

~
1

30" 00 1-2,..,'

i 10'~soo
--I10j-

l"L -r-rTl.i A,l.,=a,

in in .

~in.

Plan!orm .: Leading Edge

150"

-.l
1.

a:

(55)

I--- 20'~
400 SO

(56)

}-2H

o~i=:I:::J'=C=IIII::t[Il ]lj'kS3 t 1 .L ,0
I

"""'""1:20"1--

Relative Spanwise Distribution

~1O"f400 (57)

I A
.Cable

60 --j 10"

(56a)

01"

~=;r===i:==~' -L
r't-12"
9"

'\.-.600

a"

Fig. A5.63 shows plan ror-a of a cantilever 'Iring. The total distributed air loa~ nc~al to sur~ace is 10000 lb. The relative spanwise ~istriJut1cr. is shown. Take center of ~ressure at 24 percent o~ chord fro~ leading edge. Divi~e wing i~to 10 inch width strips and calculate VZ 1 M x and My, ane plot curves for same.

AI l' ~2

,
500 lb.

IV. F:g. A5,64 shows an externally Jraced monoplane wt ng , Take an aver-age 'Nir.g 11ft load of 90 l'J./sq.tt. norma I to wing with center of pressure at 27 percent o~ the chord from leading edge of wing and calculate anc draw the tron0 and rear beam prl~ry shears and bencing ~c~ent d1a~rams.

100

Ii

M
I
4

2.0"

60c

,e..L

c
F.B.

(II) Draw bending ~oment diagram for structures and loading in Fig. AS.51, abc
40011
Fig. A5. 64

R.B.

'o'!....M'a.
(0) 16"

100

100f

,j
4~~5
(b)

\ 10';-

I~O / 100
_ _ ::\450
(c)

iii
1--16.. --l

.L

1-12""",

Fig. AS. 61

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A5.21

BENDING MOMENTS - BEAM - COLUMN ACTION

deflections will gradually converge and the member will reach a state or equilibrium. These A beam-column is a member SUbjected to transverse loads or end moments plus axial loads. secondary bending moments could be found by sucThe transverse loading, or end moments, prOduces cessive steps by the various deflection principles given in Chapter A7. However, for pr i smat i c bending moments WhiCh, in turn, produce lateral beams this convergency can be expressed as a bending deflection of the member. The axial ~athematical series and thus save much time over loads produce .~econdary bending moments due to the above successive step method. For ~embers of the axial load times this lateral deflection. variable moment of inertia, the seCOndary moments Compressive axial loadS tend to increase the will usually have to be found by successive steps. pri~ary transverse bending moments, where as If the end loads P are tension, they will tensile axial loads tend to decrease them. tend to decrease the primarY moments; tnua, in Beam-column members are qUite crnmnon in general, the case or axial compression is more airplane structures. For example, the bearnE of important in practical design, since buckling externally braced wing and tail surfaces are and instability enter into the problem. typical examples, the air loads producing transverse beam loads and the struts introducing axial beam loads. In landing gears, one member is A5.25 Equations for a Compressive Axially Loaded Strut with Uniformly Distributed Side Load. usually subjected to large bending and axial loads. In tubular fuselage trusses, ~ateral Fig. A5.66 shows a prismatic beam of length loads due to installations supported on members L subjected to a concentric compressive load P between truss joints produce beam-column action. and a uniformly transverse distributed load W, In general, beam column members in airplane wi th the beam supported laterallY at each end, structures are conrerat ivetv long and slender and with end restraining moments M It 1 and Mg. compared to those in buildings and bridges; is assumed that the general conditions for the thus , the secondary bending moments due to the beam theory hold, namely; that'plane -sec t tons axial loads are frequently of considerable proremain plane after bending; that stress is proportion and need to be considered in the design portional to strain in both tension and compresat the members. sion. This chapter deals briefly on the theorJ At any point a distance x tram the beam end, of single span beam-column aember-e , A summary the moment expression is, or equations and deSign tables is included to2 - M1 ) wLx wX gether with examples of their use. The informa- 11=1'1.+ (118 L xT+ T - P y - - - - - (A5 . 1) tion in this chapter is used frequently in other chapters Nhere practical analysis and design of From appl1ed mechanics, we know that beam-column members is considered. For a comM =EI day pleted and comprehensive treatment of beam-coldx 2 therefore, differentiating equaumn theory and derivation of equations, see tion (AS.I) twice with respect to x gives Niles and Newell-"Alrp~ane structures~.
A5.23 Introduction A5.24 General Action of a Member Subjected to Combined Axial and Transverse Loads.

d P dxa + EI M = w - - - - - - - -

2M

(A5.2)

w
p_

SUb-figure a of Fig. AS.6S shows a member subjected to transverse loads wend axial compressive loads P. The transverse loads W produce a pr~ary bending distribution on the member as shown in Ffg. b. This bending will produce a tr~~sverse deflection c~ve as illustrated in Fig. c. The end loads P now produce an additional second~J bending moment due to the end load P tUnes the deflection 6 , or the bending moment diagram of Fig. d. This first seconda~J moment distribution produces the additional lateral deflection curve ot Fig. e and the end load P will again produce further bending moments due to this defiection. If the axial load is not too large, these successive

"w//#&0z"" "-~t (bl P~I 4 \ n-p , I ws Rua P-r""--u-'i'''''"!-P (el ,,"',.L I - - l

, "?;,.J"j "'''<.t'Ol'!
l.

i //

w f

Fig. A5. 65
_pial

p,;1, .....,..,

Ml

I=:j

I,M2

-r I

rnng Load

0<2<:;'
P
-,0.,0

rt:. . . "JM~e",-:

. Jd l

~2la.stic Curve p_~_~_p

_---_L-='l-p (el
\ I o.ef~tl.,d .... t.,l.t s.."'><'d~1 H."'erot

Fig. A5: 66

For s1mpl1t1catlon, let j =1 !fJ.pI; hence

V-p

P EI

ji"'

which, substituted in (A5.2)J gives daM 1

a:xa

F!1 = w

-,

------~----------'~~~-

AS.22

BEAM -- COLUMNS

solution of this differential equation gives. x x (A5.3) I1=C:l. Sin] + c , cos] + wjll where C1. and ell are constants of integration and sin x and cos x are ~~e limits at an infinite series of variable x. when x
C
1.

y=~

1M Ma -!"!:l. 'NLx 'NX a -:l.+--L- X - T + T

sin

sin j

When x = 0, !1 = !"!:l. and

= L,
-

I1

= Ma ,
J

therefore:
!"h - wj.lll
tan L

(A5.7a) The slope at the elastic curve at any paint is given by the first derivative of equation
(A5. 7al
i =

_ Mil - wja

si-n L

J
(M:l. - wj.lll) cos L

1(Mi I1:l. wL P - l-L - - 2"

+>'HX -

C:\. x c, ~\ Teas J + T 5i:1 x) } - (_ \A;:;.Cj

sin L

AS.26 Formulas for Other Single Span Loadings

J
and Ga

Let D:l. - w.1:ll and Da = Ma substituting in equation (AS.3),

M

= M:l. = I1:l.

Wjll

wj.lll.

Then,

In investigating other transverse loadings ror a single span carrying axial compression, it is round that the expression tor bending moment in the span always takes the torm:
M = C 1. S i n

x J

C:a cos} x + !'I'w) 'I ~ - - - I'~ ~;:;.~

t(w) is a ter::n which does not include the load P or the end moments l11. and M:a. The J expressions tor r(w), C:\. and CII depend on the To find the location of the maximum moment, dit- type of the transverse lead. terent1ate equation (AS.3) and equate to zero. Table AS.I gives the ~alue of these 3 terms tor types or transverse loading on a Single d11 _ c, X Ca sin '!: span which are frequently encountered in airplane dX - 0 = T cos J - J J structureS. The Table also gives equations for whence the pcfrrt of maximum bending moment; and its magnitUde. L Table AS.II 15 a table or slnes, cosines, x C D. - D1 cos tan - = --" = --,--"J IA5.5) and tangents tor L/.1 in radians which is more j c, L convenient to use than the usual type at trigoD1 sin 1 nometric tables. This table is based on values given in Appendix ~ of Air Corps Information The value or x must tall within x = 0 to x = L, Circular #493. The A difference have been added otherwise I11 or Ma is the maximum value. The value of the maximum span moment can be to facilitate rapid use of the tables. For single span beams, the critical value found by substituting the value trom equation of L/j is fi; that is, it the axial compressive (A5.51 in (A5.41, which gf vee load is such that the term L/j = fi, the center = D__ + wJ' - - - - - - - - - (A5.5) region of the beam will tend to deflect until the cos x combined stresses equal the tailing stress of the material. J The moment M at any point x along the span can also be written:

sin L

where

axial

n.ax

I1=D:l. [(tan

J J J where Xm refers to the value at x where the span moment is maximUm, or equation (AS.S). Since it The principle or superposition does not apis customary to locate the point of maximum span ply to a beam-column, because the sum of the bending moment and its value before investigatbending momenta due to the transverse loads and the axial loads acting separately are not the ing other S~l points, the value at tan !m is same as the moments when they act simultaneously. J known from equation (AS.S) and thus is available In combining several transverse load systems to use in equation (A5.7) tor tinding moments at with their accompanying axial loads, the principle other points along the span. ot superpOSition can be said to apply it each If the equation tor the beam detlection is transverse loading is used with the total axial desired, it can be found by SUbstituting the load for the systems which are being combined. value of M trom equation (A5.3) in equation ThUS, in Ta~le A5.I, to tind the :noments tor (AS.l), which gives: several combined loadings, add the values ot C:\., ell and t(w) for the several loadings and use

~ . sin ~rcos ~]+Wjll

(AS.7)

AS. 27 Moments for Combinations of the Various Load Systems as Given in Table AS. I, Margins of Safety. Accuracy of Calculations.

---_:.
ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES AS.23

Table AS. I

Values of Terms ell Cz. and f(w) in Equation

M C 1 sin..!. + Cz
j

cos~+ f(w)

Single Span - Axial Compression - Uniform Section

Loading Equal End Moments. No Ml,. Side Load

C,
.~2 i
sin L

C2

f(w)

Eq. for Point of Max Bending Moment

x=L 2
Tan..!. = M2 - M1 cos~ j j M 1 sin..!::. j

Eq. {or MaX. Span

Bending Moment

P"""'1r-- x _

M max

Ml
2j

Unequal End Moments. No M Side Load M2

P+~~j====I~P I- ~L .
w ;,/ in.

T
wj2 (cos L-l)
sin L

Uniform Side Load. No

End Moments

ttttttttrtTttft p-p C:x~ . l L

Uniform Side Load Plus
End Moments

x: .5 L

T
DZ - D cos L/i em I J
where

M"jf t II t t t f i I tit hM2

Tan..!.: D2 - D1 cos..!:.
wj2
j
J

P+'!=~ o

I""'P

Dr:; Ml- wj2 D2 s M2- wj2

-eD1 sin.,h
j

Mmax'=~twi2 cos x

x <: a, .Wjsinb
sin L

T
- Wj

o
sin.!.
)

o
o

x o a, + Wj sin a

1" tan L
T

Triangular Loading. No
End Moments

p__

-c:rT r 1! t ~JI w~in.

L

..l.

t--X~

(NOTE A) To obtain Max1mum Moment, compute moment at 3 or. " points in span. Draw a smooth curve thru plotted results.

Triangular Loading. No End ~ Momems

w./p'"iJTi

'Fx

U ro-::::r
f L

Couple Loading (Clockwise)

x < a,

- m

cos b sin L

T
)

(See Note A)

X)

a. - m cos..!..
tan L

m cos.!.. J

w or W is positive when upward. M is positive when it tends to cause compression on the upper fibers of the beam at the section being considered. Reference: ACIC '1'493; Niles, Airplane Design; Newell and NUes Airplane Structures For Table of many other loadings, see NACA T. M. 985.

__________

1IIl..,'~'1_

TABLE AS. II

llATUllAL SIlIES, COSIns , AIID TAIlGEllTll OJ' AIlGLEl! III IlADUIlI

L jj

ill Sill

Rad1allS

~jj

t:.
Sill Lj
j

Cos Lj

f::. Cos Lj

Tall Lj

f::. Tall Lj

3.00 3.01
3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 3.10

0.14112 0.00991
0.13121 .00 92

-0.98999 0.00136 -0.99135

-0.14254 0.01019 -0.13235 -0.12219

0.00118 -0.11206 0.00108 0.01011 -0 10195 0.00097 0.01010

O. 0127
-0.99262 -0.99378

O. 2129
0.00993

10 6
0.01013

O. 1136
0.00994 0.10142

-0 99484 0.00996 -0.99581

0.00996

0.09146 0.08150 0.00997 0.07153 0.00998 0.06155 0.00999 0.05156 0.00997 0.04159 0.01000 0.03159 0.00999 0.02160 0.01000

-0.09185 0.00086
0.01008 -0. 817 0.00077 0.01006 -0.0717 0.00066 0.01004

-0.99667
-0.99744 -0.99810

-0.08167

O. 0057
-0.99867 0.00046 -0.99913 0.00037 -0.04162 -0.0318 0.00027 -0.99977 -0.05184

O~
O.
0.01001

003

3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19

3.20

-0.99950 0.00018 9 3 o 00007 0.00003 -0.99997 0.0100e

0.00014

-0.02180

.0 16
0.00160 0.01001

-.
-0.00160

0.01000

-1.00000

0.01001 0.00841

-0.00841 -0.01841 0 -0.02840 -0.03840 0.00999 -0.04839

0.00999 -0.05838 0.00998

O.
00.01841 00.02841

-11.99984 9
-0.99960 -0.99926

3 0.00034 0.03843 0.00043

-0.99883 0.00083 -0.99830 0.05848

0.01000

i
I

0.04845 0.00064

0.01003

0.01004
0.06852

3.21 3.22 3.23 3.24 3.25

-0.06836 0.00997
-0.07833

-0.99766
0.00073

0.01005 0.07857 0.01007 0.08884 0.01009 0.09873 0.01010

0.10883

-0.99693 0.00996 o 00084 -0.99809

0.00996

-0.08829
-0.09825

0.00093
-0.99516 0.00103 -0.99413

0.00995 -0.10820

65

AS.28

BEAM -- COLUMNS

of the bea~, or 4420 x .75 -= 3315"# positive bethese values in the general expression tor M as cause it procuces compression in the top tibers. given at the top of the Table. The moment at (2) due to the cantilever overhang In a beam-column member, the bending momequals (20 + 10) 36 x 16 = 8640"#. Fl;. A5.:58 ents do not vary directly as the load 1s i~crea 2 ed. Thus, the student should realize that margins or safety based on direct ?roportton of mom- shows the beam ?ortion between points (1) and (2) as a free body. ents to loads are incorrect and lie on the unFrom Art. AS.Z5, we have the follOWing presafe side. cise equations for a beam carrying a tr-ansverse It is recommended that four significant uniform distributed load with end compressive figures be used in computations, making use ot loads. the so-called precise equations, since the results in many cases involve small differences x L between large numbers. D~ - Dl cos J - - - - - - - - -(A) tan J
AS. 28 EDmple Problems

D.. sin L
and

Example Problem #1 Fig. A5.S? illustrates a typical upper, outer panel Wing beam of a biplane. Let it be required to determine the maximum negative bending moment between paints (1) and (2), generally referred to as the maximum span moment. To obtain the true bending moments on the beam, the axial beam load as well as the end moments at (1) and (2) are necessary since they influence the deflection of the beam. Solutlon:To obtain the horizontal component Th or the lift strut load, we take moments about the

!'l

max

= ...!2>..cos x
J

w.:;

- - - - - - - - - - -(E)

Evaluating ter:ns tor substitutIon in these equations, we obtain,

Ml
Mil

3315"#
8640"#

= 4420# compression -= 10 in~ given and assumed constant throughout the span.
=1/l.3XIoeXIO = V4420

roo- ----I-- 36''---1

v294I

54.23

P -';"~I~~;b~g;~g~~~~~t'~~B~~.J...
.A.

w= 2Q.IP/in.

lOll/in
Wj

"= 20 x 2941

Wood Sea
~... <;e

D1.
DII

M 1

wja

'\..!~'\.

,0' 'S:'~

lSOQ.IP=Reaction

= M..

- wjll

= 58820 = 3315 58820 8640 58820

55505
50160

from Strut to Lower Wing

J = 54.23 = 1. 844
From Table A5.II sin

100

= .96290 and cos

J -= ....

26981

Fig. AS.67

Substituting In equation (A) L x tan J = 011 D1 cos J

D 1 sin
...... w= 2Of/in.

J
-65156 -53441
1.2192

50180 - (-55505 x - :26981) -5~505 x .96290

tan-' 1.2192 = .88383

Fig. AS. 68

hinge point at the lett end.

l:l'lb 2000 x 50 - 540 x 116

Hence, x = .88383 x 54.23 = 48 ft , which equals the distance trom the lett end at the beam to the point at maximum span moment.

- 1500 hence
Th = 442Qj1

x 100

70.75 Th = 0 Hence

= ..lL:-

cos x

r ,
+

cos ~ = .63419 from Table j

A5.II

J
~
55505 - :63419 58820 28,700"#

The axial compressive load induced by the lift strut at pOint (2) then equals - 4420#. Taking ZH = 0 tor the load system or Fig. AS.S? gives P = - Th -= 4420#. The end moment on the beam at (1) equals the end load timeB the eccentricity or the hinge :ram the neutral axis

To obtain an secondarY bending to the axial load tion, the primary

idea as to the magnitUde at the moment, that is, the ~oment due times the lateral beam deflecbending moment at a point 48~

-,,;; . r

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AS. 29

tram the left end w111 be computed. substituting values of C1 and Cm and f(w) :rom Table A5. I in the above equa.t i ons ; M..II := 3315 + 48x20x 24- 940:<48:= -18765"# I1= (M .. -!11. cos L/j) sin x/j +M cos x/j Thus the secondary bending moment equals sin L/j 1. - 28700 + 18765 = - 9935"# which 1s a large percentage of the primary moment. The transverse 20, OQQ!...lF==={~2='=!=2="=0=6=3=8=,,="=1=T=U=be=J~02~~000t deflection of the beam at the po1nt of max. span . L: 41. 762 ,( moment then equals - 9935 := 2.25 inches upward.
- 4420

Bending Moment at any Point Along Span Fig. A5.70 Let the moment at a pcfrrt 10" from paint (2) be required. In this case, x = 100 - 10 = 90 But, M:a. = 0 in our problem, henes, I1 = Masin x/j M:= Dl.~tan ~ . sin })+ cos + wja (Ref. Eq.

(j)

11

A5.?)

sin L!j

90 x - - = 1. 6596, sin j54.2:3

.99605

=" {Ei: Llj

~ /"'2"'9"'X'l"07''''X-."746' =' ~ = 25 826 VP - V 20000 V66'l

=~=1 25 .826 617
- - - sin L/j = .99892

cos x = - .08867

tan

.s =1.2192 =value
j

tor x at the point of maximum bending moment

x = Ll2 = 20.881 x!j - 25.80 20.881 - .8085 ern x/j = .72327

Hence,

M =-55505 [(1.2192 x .99605) +-.08867] ... 58820 = - 3664'"

suost i tutlng in the above equations tor 11

M = 36000 x .72327 = .99892 26066""

/'

E:x:ample Problem #2 Fig. A5.59 shows a simplified landing gear structure carrying a vertical load of 12000# on the axle. M.ember ABC is continuous thru B and ptnned at C. Let it be recurred to determine the bending moment at the midpoint or member BC and its lateral deflection due to the 12000# vertical design load.

This compares with a prtmarY moment of 36000/2 ~ 18000"#. The deflection at the midpoint or BC := 26066 - 18000 = .403 In.
20000

The maximum moment is gl ven by the equation:

~ 1

~D

26"

"0:"------;;:86:::2"."'"

~ Ie

~ =s~n L and it occurs at x =n2j (See

"--------112 21/2-083.
Round

Tabla A5.I)

A5. 29 Stresses Above Proportional Limit Stress of

Material.

Steel

Tub"
Fig. A5.69

1 I
0

Line of action of DB goes through E

BiE,,(

AxlJ L
A
2000*

The equations as presented in this chapter assume that E is constant or in other words the stresses are Within the elastic range. In aircrart structural destgn the applied or limit loads must be taken without suffering permanent deformation, hence E 1s constant under such loads. However the aircraft structure must take the deSign loads which equal the limit loads times a factor at safety (usually 1.5) without failure. In many cases structural failure will occur under stresses in the plastiC range where the material stiffness Is less and not constant. A good approximation for ~~ effective modulus E' Is obtained as :ollows:(1) Compute Fc = PIA for the given number. (2) With this value of Fc enter the basic column curve diagram for the given material (for end fixity C = 1) and find value of L'/p corresponding to the stress Fc'

"""'"3

Solution:Solving for reactions at C by staticS, we obtaIn :he axial load in Be = -20000. The bending momen~ at 8 due to 3" eccen~ricity of the 'wheel load = 3 x 12000 := 36000"#. Fig. A5.70 shows a r~ee body of portion BC of member ABC. From Table A5.1 x '() M = C~ sin x j +, v~ cos } + ~ w

~~--~-------~~

AS. 30

BEAM

-~

COLUMNS

(3) USing these values of L' /p and FC J compute

"' -""Tti" _Fe (L')' P

to

(4) Determine the bending ~oment at the centerline of the beam-c01~~s shown in Fig. A5.74. Assume ~I 64 J O O O, 000 lb. in. sq.

(4) Then i

>

(E~ry/'

Basic column curves for various ~terials are given in another chapter of this book.
AS.30 Problems

1!
If

800<1
Fig. A5.75

500<

6000*
1

L--l.--

H155"'-!-- 35"
-

50"

---I W= 30<1/in. I sooo " t t t t t .. BOOot .JooOit---4 . - ;-I 1-30''t- 80''---,

300<1

(5) For the beam-column in Fig. A5.75 calculate the bending moment of the centerline of the member. Assume E = 1,300,300 psi. and I = 10 in~

Flg. AS. 71

Fig. AS.72

(1) Fig. A5.71 shows a 1-1/2 - .065 steel tube subjected to both end and lateral loads. Determine the max~um bending moment on the tube. Campare the result with the bending moment due to the side load only. E = 29 X 10 8 ps i . I of tube = .075 in. 40 Compute lateral deflection at point of maximum bending moment. (2) The beam column member in Fig. A5.72 is made of 24ST aluminum. alloy. Calculate and plot a curve of the bending moments on the member. Also plot bending moment due to lateral loads only. E = 10.3 x 10" psi. I:::: 5.0 in ....

500.

50"

----1

;,000# 18000"-IF

Fig. AS. 76

(6) For the beam-column loading in Fig. A5.76, calculate bending moment at center point ot beam. Take S 1,200,000 psi and

10

In~

AS.31 Beam-Columns in Continuous Structures.

(3) Determine the maximum. bending moment for the wood wing beam and loading of Fig. A5.73. I of beam section = 17 in. E 1.3 x
10".

p-r*
V

I t f !

The secondary moments in a partic~lar member due to beam-column~action also effect or influence the deflections in adjacent members of a continuous structure. This rat~er involved problem can be handled quite simply and rapidly by the moment distribution method as explained and illustrated in Arts. All.l2 to 15 of cnap'ter All.

s.

"000..
.

1~

t-<."+-:""--. :. ""-j #rlll tyeoo

IO<I/in.

10001

20007

I I I tIl I

tt

I tl l!

0'"

100001

200" - - - - . - . "
Flg. AS. 74

CHAPTER A6

TORSION.. STRESSES AND DEFLECTIONS

A6.1 Introduction.

Problems involving torsion are common in aircraft structures. The metal covered airplane
wing and fuselage are basically thin-walled

tUbular structures and are subjected to large torsional moments in certain flight and landing conditions. The various mechanical control systems in an airplane often contain units or various cross~sectlonal shapes which are sub1ected to torsional forces under operating conditions, hence a knowledge of torsional stresses and distortions of members Is necessary in aircraft structural design.
A6.2. Torsion of Members with Circular Cross Sections.

sumed that any radial line undergoes angular displacement only, or OB remains stral&~t when moving to OB' The unit shearing strain in a distance L equals,

Let G equal modUlus at rigidity at the and let ~ equal the unit Shearing stress at the extreme fiber on the cross-section.
~terial

Hence,

= E G = r 9 G _ _ _ _ _ _ _ _ _ _ (1)
L

The following conditions are assumed in the derivation or the equations for torsional stresses and distortions: (1) The ~ember is a circular, solid or hollow round cvi incer . (2) Sections remain circular after application of torque. Diameters remain straight atter twisting of section. (4') Material rs homogeneous, isotropic and elastic. (5) The applied loads lie in a plane or planes perpendicular to the axis of the shaft or cylinder. p
(3)
L

In Fig. A6.2 let ~p equal the unit Shearing stress on a circular strip dA at a distance p from O. 'fhen
"0 ,
=-'t"

__ prQG = p 9 G -Lr L

The ~oment of the shearing stress on the circular strip dA about 0 the axis of the bar 1s equal to,
dM .. -r pdA

= '--L-

o'ig}dA

and thus the total in-

ternal torSional resisting moment is,

\nt.
B'

,
\

--L-

ao"QdA

--

J ,,

For equilibrium, the internal resisting moment equals the external t or-ai onaj moment T, and since GQ/L is a constant, we can write,
.T = Mint. =
GQ L Jr
0

Fig. A6.1

p ctA = T

GQJ

(2)

Fig. A6.2

Fig. A6.l shows a straight cylindrical bar subjected to two equal but OPPOSite torsional couples. The bar twists and each section is SUbjected to a Shearing stress. Assuming the left end as stationary relative to the rest of the bar a line AB on the surface will move to AB' under these shearing stresses and this rotation at any section will be proportional to the distance from the fixed support. It 15 asA6.1

Where J = polar moment of inertia of the sr~ft cross section and equals twice the moment of inertia about a diameter. From equation (l) GQ = ~ L r -r J (3 ) Hence, T = r

or
also from equatfon (2), solv:ng for the
Q

(4 )
~Nlst

9,

= GJ TL
(9 1s measured in radians).

(5 )

C-,.' vb

AB.2

TORSION
?;,obl~m

AB.3 Transmission of Power by a Cylindrical Shalt.

2.

done by a tWlst~n~ 8cu~le T 1~ ~ovl~g tn~ough an a~~~lar jis~l~c?~~~t :s ~qual to th3 ~roduct of the ~~gr.it~de :f the cJu~le and the angu lnr c i s cracement in rao Lans . I~ tne angular dt s pLaceasnt t s one r-ev -jut t on , ::18 ',II'::'!:. done equals 2 n T. I: T is ex:ressed in lr.chpounds and N Is the~~ul~r ,elccity in revol~ tions per ~inute, then the horsepower tr3ns~itted by a rot~ting shaft ~y be written,
~ork

The

?i s : AE.4 111ust~ates a~ a:lercn control s~r~~C3, consisting of a circ~lar tor~~e tube

H, P. = 396000

2 n N T

( 0,

. , r----------------" Pressure istributiOn4

Average load on surfac e e 40#/0'
" "'-'

(1-1/4 - .C48 in Size) suppo:,ted on :~ee ~rack3~s ~~d ~lt~ tte c:ntr:l ~od ~:tt:~g attac~~c ~c ~he tor~ue t~be above the center SUDJOrt bracket. F:nd the 8~xi~lli~ torsional shear1ns stress in :he tu~e i t the air lo~d on the ailer:n is as indicated in Fig. A6.4, and also compute the an~le sf twist of tute between horn section anc end of aileron.
hi~ge

.
i

rr
1\ :
I

where 396000 represents inch ~o~~ds of work of one horsepower fer ~ne min~te. Equation (6; may be wr-ttten:
T = H.P. x 396000 2nN
EXAMPLE PRCBLillS.

.:'li '0495ETUbe(.A~~~~ ~ _~_ Tube t . c ~ - - - - -=-=r~e.:r

! I\ ::: ,I I,

63025 H.P.
N

(7)

5"

L
I

24"

24" -15"
Fig. AB.4

Problem l. Fig. A6.3 shows a conventional central stick-torque tube operating unit. For a side load of 150 Ibs. on stick grip, determine the shearing stress on aileron torque tube and the angle of ~Nist between points A and B.

SOLUTION:
airload on the sur~ace ~er.ds to rotate ~round the tC~Gue t~be. ju~ novement is prevented or crea~ed 2y a control rod attached to the torque t~be over the center supporting oracket. The total load 0n a 5t~ip of aileron cne lucre wide = 40(15 x 1/:'44) ~ 4.16 lb. Let w equal intensity of loading per inch of aileron span at tha.,leading edge point cr the aileron surface. (see ~ressure diagram in
~he

the ailer8n

SOLUTION:
Torsional moment on tube AB due to slde stick force of 150# ~ 150 x 26 = 3900 in. lb. The resistance to this torque is provided by the
150*

Fig. A6.41.

Then 3.

(0.5 w)12 hence w

p~

= 4.16
lb.

24 ST At Alloy ,...control Stick l G ::: 3,800,000 psi

= 0.463

ilt----2.' - - - 1
....--.. EI~v. Control Wire

Ai}eron Operating Tube ~

B

I ",.I Ail 0 Ho 356J1J1 II er n rn

forwarG of the centerx 3 = 1.389 lb. and P a the load on aileron portion att of hinge line ~ 0.463 x 0.5 x 12 = 2.778 lb .
line at torque tube

The total load

= 0.463

~ Bearing

.J:

Fig. AB.3

aileron operating system attached to aileron horn and the horn pull equals 3900/11 ~ 356 lb. The polar ~oment of inertia of a It - 0.058 round tube equals 0.1368 in 4

Maximum Shearing stress = ~ - Tr/J = (3900 x 0.75)/0.1368 = 21400 pSi.

The angular twist of the tube between points A and a eq~als
~ - TL _ 3900 x 28 - GJ - 3,800,000 x 0.1368

The torsional ~oment per r~nlng inch or torque tube: = - 1.389 x 1.5 + 2.7~8 x 4 = 9.0 in. lb. Hence, the maximum torque, which occurs at the center of the aileron. equals 9.0 x 29 = 261 in. lb.
_ Tr
1:(:nax.) -

x 0.625 = 2450 J = 251 0.06678

(J

ps t .

= 0.21

radians

0.06678 in 4

. )

or 12 degrees.

Since the tube section Is constant and

~he

torque

'n

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AB.3

varies cirectly as the cistance :rom the end of the ailerons, the angle of twist g can be CQm?uted by using the average torque as acting on entire length of the tube to one side of horn or a distant J.... = 29", :tence 261 x 29
2

twists of non-circular cross-sections is given in Art. A6.5.

A6.5 Elastic Membrane Analogy.

3800000

0.06678

AB.4 Torsion of Members with Non-Circular erose-

sections.

The for-~las derived in Art. A6.2 cannot be for non-c1rcular Shapes since the assumptions made do not hold. In a circular shaft subjected to pure torsion, the shearing stress distribution is as indicated in Fig. A6.5, namely, The maxi~um Shearing stress is located at the most remote fiber from the centerline axis of the bar and is perpendicular to the radians to the stressed paint. At a given distance from the axis of rotation the shear stress
~sed

Fig. AB.5

-E-P=+
Ie
:;;;::-_-~~I
"': max.

Fig. AB. B

Fig. AS.7

The \Shape of a~watD~a cross-section of a non-circular cross-section in torsion is needed in the analysiS by the theory of elasticity, and as a rasult only a few Shapes stich as rectangles, ;eJ:,lJ2.-seS, triangles, etc., have been solved by the theoretical approach. However, a close approximation can be made experimentally for a Iznos't any shape of crosssection by the use of the membrane analogy. It was pointed out by Prandtl that the equation of torsion of a bar and the equation for the deflection of a ~embrane subjected to uniform pressure have the same form. Thus if an elastic membrane is stretched over an opening which has the same shape as the crosssection of the bar being considered and then if the membrane 1s deflected by subjecting it to a slight difference of pressure on the two Sides, the reSulting deflected shape of the membrane provides certain quantities Which can be meaSured experi~entallY and then used in the theoretical equations. However, pOSSibly the main advantage ot the membrane theory is, tr~t it provides a method of visualizing to a considerable degree of accuracy haw the stress conditions vary over a complicated cress-section of a bar in torsion. The membrane ar~alogy provides the follaNing relationShips between the deflected ~em brane and the twisted bar. (1) Lines of equal deflection on the membrane (contour lines) correspond to Shearing stress lines of the twisted bar. (2) The tangent to a contour line at any point on the membrane Surface gives the direction of the resultant Shear stress at the corresponding ~oint on the cross-section of the bar being twisted. .,,,.> (3) The max~ slope of the deflected membrane at any paint, with respect to the edge support plane 1s equal in ~gnitude to the shear stress at the corresponding point on the cross-section of the twisted bar. (4) The applied torsion on the ~Nisted bar is proportional to ~N1ce the volume inclUded be~Neen the deflected membrane and a plane through the supporting edges.
~

is constant in both directions as illustrated in Fig. A6.5, Which means that ends of segments of the bar as it twists remain parallel to each other or in other words the bar Sections do not warp out of their Plane when the bar twists. If the conditions of Fig. A6.S are applied to the rectangular bar of Fig. A6.S, the most stressed fibers will be at the corners and the sGress will be directed as shown. The stress would then have a component no~l to the surface as well as along ~he surface and this is ~ot true. The theorj of elasticity shows that the maximum shear stress occurs at the centerline of the long sides as illustrated in Fig. A6.S and that the stress at the :orners is zero. Thus When a rectangular bar twists, the shear stresses are not constant at the same distances from the axis of rotation and thus the ends of segnents cut trIough the bar would not remain ~arallel to each other when the bar ~Nists or in other words, warping of the section out of ~ts plane takes place. Fig. A6.7 il1ust~ates this action in a twisted rectanb~lar bar. The ends of the bar are warped or suffer distortion normal to the original unstressed plane of the bar ends. Further discussion and a summary of equations for determining the shear stresses and

To illustrate, consider a bar with a rectangular cross-section as indicated in Fig. A6.B. Over an opening of the same sr~pe we stretch a thin membrane and deflect it normal to the cross-section by a small uniform press~re. ~qual deflection contour lines for this deflected ~embrane will take the sr~pe as illustrated in Fig. A6.9. These contour lines Which correspond to direction of shearing stress in the twisted bar are nearly cirCUlar near the center region of the oar, but tend

Sf

A6 4

TORSION

to take the shape of the bar boundary as the boundary 1s approaChed. ~lg. A6.Sa shows a section through the contour lines or the ~erlected membrane along the lines 1-1. 2-2 and 3-3 of Fl~.
AS.9. It is obvious that the slopes of the de-

From Table A6.l it is ~otised that :or large values 0:' bit, tne values 0-:: t he constants 1s 1/3, and t~us for such narrow rectangles, equations (6) and (7) reduce to,
_ _ 3 T
"MAX -

tlected surface along line 1-1 will be greater than along lines 2-2 or 3-3. From this we can conclude that the shear stress at any point on line 1-1 will be greater than the shear stress tor corresponding pOints on lines 2-2 and 3-3. The maxlmu~ slope and therefore the maximum , ,

bt:l

(5 )

_ 3 T g - bt 3 G

(9 )

,
,

,:>
,
,

~.

I __ I
! , ,

._.
\

~
I

,
,

.J

Although equations (8) and (9) have been derived for a narrow rectangular shape, they can be applied to an approx~mate analysis of shapes made up of thin rectangular members such as illustrated in Fig. A6.10. The ~cre generous the fillet or corner radius, the smaller the stress concentration at these junctions and therefore the more accuracy of these approximate formulas. Thus for a section made up of a continuous plate such as illustrated in

Fig. A6.

a
stope

Fig. A6.9

l~l

.LSlope

2~
Fig. A6.9a

3~
Fig. a b can be taken as centerline length for above type of sections

Slope

stress will occur at the ends of line 1-1. The slope of the detlected membrane will be zero at the center ot the membrane and at the four corners, and thus the shear stress ~t these points will be zero.
A6.6 Torsion at. Open Sections Composed of Thin Plates.

.r-:

b " --! ..J::t "

~ET b. ...
fo-bl.~~tl.

Fig. A6.10

Members having cross-sections made up of narrow or thin rectangular elements are sometimes used in aircratt structures to carry torsional loads such as the angle, channel, and Tee Shapes. _;"'~'Ir-For a bar of rectangular cross-section or width b and thickness t a mathematical elastiCity analysis gives the following equations tor maximum shearing stress and the angle of twist per un1 t length.

Fig. (a ) of Fig. A6,.lO, the width b can be taken as the total lteiigthof the cress-section. For sections such as the tee and H section in Fig. A6.l0, the polar moment of inertia J can be taken as Z bt~/ . Thus for the tee section of Fig. A6.10:
9

T =-'=G3 OJ Z bt 3

a b t

(6 )

G (blt~ + bat~)
bl
-

3 T

" - ~ b T '0' t'" G 1n

radl ana - ~

(7 )

Values or a and

are given in Table A6.1.

TABLE AS.l
CONSTANTS (lAND ,
bit 1.00 1.5Q 1.'1'S .2.00 2.50 3.00 4 S a 10 I 0 0 c, O.:W8 0.231 O.~O.2~~~O.~o.m~~~WO.3130.m 0.141 0.19 0.214 O. O. 49 O. 3 O. 1 O. .299 0.307 0.313, O. 333

'toa

For the maximum Shearing stress on leg _Tt l _ 3 T t l 3Tt l 'tb l .fb'tT=blt l S T b:lt a " - and tor the ?late bar _Tt a _ 3Tt a

--J-1

(10)
(

- J - - bltl S + b a t a 3 - - - - - - - -

11)

It t
9
1:'

= t , = t,
3(b

then

G t

3 T
l

+ ba)

- - - - - - - - - - - - (12)
- - - - - - - - - - - (13)

3 T t 1 ( b 1. + 011)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AG.7 Torsion of Solid Non-Circular Shapes and ThickWalled Tubular Shapes.

A6.5

EXAMPL" PROBLEM SHOWING TOPBIONAL STIFFNESS OF CLOSED THIN ',.JALLED TUBE COMPARED TO OPEN OR SLOTTED TuBE. Fig. A6.lla shows a 1 inch diameter tube with .035 wall thickness, and Fig. A6.11b shows the same tube but with a cut in the wall making it an open section. For the round tube J. = 0.02474 in~. For open tube 3 2 = ~ x 3.14 x .035~

Table A6.3 summarizes the for.nulas for torsional deflection and stress for a few Shapes. These for~ulas are ,based on the assumption that the cross-sections are free to warp (no end restraints). Material is homogeneous and stresses are within the elastic range.
A6.8 Torsion of Thin-Walled Closed Sections.

= 0.000045

Fig. AG.lla

Fig. A6.l1b

Let Q1 equal twist of closed tube and Q 2 equal twist of open tube. The VNist, will then be inversely proportional to J since
Q

The structure of aircraft wings, fuselages and control surfaces are essentially thin-walled tubes of one or more cells. Flight and landing loads often produce torsional forces on these major structural units, thus the determination at the torsional stress and deformation of such structures plays an important part in aircraft structural analysis and design. Fig. A6.12 shows a portion of a thlnwalled cylindrical tube which is under a pure torsional moment. There are no end restraints on the tube or in other words the tube ends and tube cross-sections are free to warp out of their plane.

= G~J'

Therefore the closed tube is Jl/J a = 0.02474/ 0.000045 = 550 times as stiff as the open tube. This result shows why open sections are not efficient torsional members relative to torsional deflection.
TABLE Ae.3
FORMULAS FOR TORSIONAL DEFl.ECTION A.ND ,rrRESS

e :: KG :: tw1.1lt

l.n I"ll.dJans per meh of length. T :: Tors1onal Moltlent (In. lb.). G :: Modulus oJ. R~id1ty. K (in4) Froltl Ta.b1e.

~
Fig. A6.12 Fig. All. 13

~qds
q constant

SO
SOLID SQUARE.

SECTION

FORMUloA FOR SHEAR STRESS

SOLID ELLlPITICAL
'tMAX :: ft'iijr (1Lt elida of minOr ana).

'T

Let ~ be the shear force intenSity at point (a) on the cross-section and qb that at point Now consider the segment a a ~ b of the tube wall as shown in Fig. A6.l2 as a free body. The applied shear force intensity along the segment edges parallel to the y axis will be given the values ~y and qby as shown in Fig. A6.l2. For a plate in pure shear the shearing stress at a point in one plane equalS the stress in a plane at right angles to the first plane, hence qa = qa and qb = qb . Y Y Since the tube sections are free to warp there can be no longitudional stresses on the tUbe wall. Considering the eqUilibrium of the segnent -in the Y direction, qb;" 0, hence Qa y qby and therefore Ga = qb or in other words the shear torce intenSity around the tube wall is constant. The shear stress at any point T = qjt. It the wall thickness t Changes the shear stress
~Fy

(b)

D1
SOLID RECTANGLE.

T
K
tO.141a

't MAX :: O.208 aJ (at m1dp::lint of each 'ide).

CJI 2a--.,
~
SOLID TRIANGLE.

ao3[~

3.3e~(1

'tM.\X

= T(3a

- 1. 8bl 8a.2t)%

mtdpo1Dt of 10ll(side.

~]

6~
For an extensive U,t of formula. for many ~, both .soUd and bollo.., refer to book, "Formulas For StreBB and stl"ll.in" by Roark, 19!14 Ed.itiOQ.

= 0 = Qa~

A6.6

TORSION
shea~

changes but the

force q dOes not change,

or
~ata

T~e elas~lc strain energy dU stored in :his element therefore equals,

~btb

= constant.

dU

CS =2 4

-:5

The prOduct ~ is generally refe-rred to as the shear flow and is given the symbol q. The name shear flow possibly came from the fact that the equation ~ = const~~t, resembles the equation of continuity of fluid flow qS = const~~t where q is the flow velocity and S the tube cross-sectional area. We will now take moments of the shear flow q on the tube cross-section about some point (0). In Fig. A6.l3 the force dF on the wall element ds qds , Its arm from the assumed ooment center (0) is h. Thus the moment of dF about (0) is q cs h. However, ds times h 1s twice the area of the shaded triangle in Fig. A6.13. Hence the torsional moment dT of the force on the element ds equals,

nowever the shear strain 6 can be written,

6

=~ G :::;...9:.... Gt'
em : :;

but 4

:=

T" 2A

hence or
U

.- cs , 'ehe mt egraI y f 8Jl:""Gt

cs
~
OIGt

8A.L'IGt

--

cs

is the line integral ar8und the periphery of the tube. From Chapter A7 from Castigliano's theorem.
f"I _

'7

dT qnds 2qdA

aT -

au _ ~

r 4A,.T
L~ 15 J t

- - - - - - - - - - -

(17)

and thus for the total torque for the entire shear flow around the tube wall equals,
T = fA 2qdA and since q is constant
T = qU

since all values except t are constant. equation (17) can be wTitten,
Q

~ 4A,.IlG

- - - - - - - - -

(~8)

and since T

=2

qA, then slso,

(-19)

(14)

Q=~Gfd:at
L

or
q = 2A

Where Q is angle of twist in rcdlans per lli~it length ot one ir.ch cf tube. For a tube length

(15)

Q=i<~f~s--------

----(20)

where A is the enclosed area of the mean periphery of the tube walL The shear stress ~ at any point on the tube 'NaIl is equal to q, the shear force per inch ot wall divided by the area ot this one inch length orlxtor

AG.9 Expressi.on for Torsional Moment in Terms of Internal Shear Flow Systems for Multiple Cell Closed Sections.

(16)
TUBE TWIST

Consider a small element cut from the tube wall and treated as a free body in Fig. AS.14, With ds in the plane at t~e tube cross-section and a unit length parallel to the tube axis. Under the Shearing strains the plate element to- 1 -0\ >0- 1 --<

Fig. A6".l6 shows the internal shear f Low pattern for a 2-cell thln-walled tube. when the t~be is subjected to an external torque. q~, qa and 43 represent the shear load per inch on the three different portions of the cell walls. For equilibrium of shear ~orces at the junction point of the interior web With the outSide wall, we know that
------(21)
m

T , <Is

1. t

'0----1
I

----I
Fig.A6.14

qds~D:
Fig. AS. 15

t ,

t
\
Fig. AS. 16
n

detorma as illustrated in Fig. AS.l5, that is, the tace a-a moves with respect to face 2-2 a distance O. The torce on edge a-a equals q ds and it moves thrOU~~ a distance O.

Choose any ~oment axis such as paint (0). Referring back to Fig. A6.l3, ~e found that the moment of a constant shear force q acting along a wall length ds about a pOint (a} was e4~1 in

_ _ _ _ _ _ _ _......... iII'itMTlllil'i

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A6 7

magnitude to twice the area ot the geometrical Shape formed by radii from the moment center to the ends of the wall element d5 times the shear
flow q.

(0).

Let To moment ot shear flow about point Then from Fig. A6.16,

inch in the web 1-2 and (qe - q~) ~or web 2-3. For eqUilibrium, the torsional moment af the internal shear system must equal the external torque on the tube at this particular section. Thus. from the conclusions of article A6.9, we can write:
(24)

For elastic continuity, the twist of each cell must be equal. or 01. = c, = Q~.

But from equation (21), qa = q~ - qao Substituting the value of qa in (22)

From equation (19), the angular Is

~Nist

of a cell

Q=-q_!~
2AG

or
(25)

Hence, To 2q,A, + 2q,A, - --- - - - - - - (23) where A1 = area ot cell (1) and A3 = area of cell (2). Therefore, the moment of the internal shear system of a ~ultlple cell tube carrying pur~ torsional shear stresses 1s equal to the sum of twice the ~nClosed area of each cell times the shear load per inch Which exists in t~e outside wall or tr~t cell. (Note: The web mn 18 referred to as an inside wall of either cell)
AS.10 Distribution of Torsional Shear Stresses 1n a Multiple-Cell Thin-Walled Closed Section. Angle of Twist.

2GQ=.s.i~ A j t
ThUS, for each cell of a

ture an expression ~s can be written and equated to the constant value 2OQ. Let a l Q represent a line integral 1-0, and ala,

tf

~ultiple

cell struc-

f ~s
a~Q

for cell wall the line in-

tegrals ~s for the other outside wall and interior web portions of the 3-cell tube. Let clockwise direction of wall shear stresses in any cell be pOSitive in sign. Now. substituting in Equation (25), we have: cell (1)

a~o, a2~

and

Cell@
j t a

A~ [q~

au + (qJ. - qa) au]

=209

(26)

=2GQ
Fig. AG.17

- -

- -

- - -

- -

- - -

- -

(27)

Fig. A6.17 shows in general the internal shear flow pattern on a 3-cell tube produced by a pure torque load on the tube. ~he cells are ~umbered (1), (2) and (3), and the area outside the tube 1s designated as cell (0). Thus. to designate the outside wall ot cell (1), we reter to it as lying between cells 1-0; for the outside wall of cell 2, as 2-C; and for the web between cells (1) and (2) as 1-2, etc. q~ = shear load per inch = ~~tl in the outside wall of cell (I), where ~l equals the unit stress and t wall thickness. Likewise, qa ~ata and q3 ~3t3 = shear load per inch in outside walls ot cells (2) and (3) respectively. For equilibrium of shear forces ~t the junction points of interior webs with the outside walls. we have (q. - q~) equal to t~e shear load per

Equations (24. 26, 27 and 28) are sufficient to determine the true values of ql, qa, q~ and Q. ThUS, to 1ete~ine ~rre torsional stress distribution in a multiple cell structure, we write equation (25) fer each cell ant these equations together with the general torque equat:on, s1milar to eq~tion (24), proViQes sufficient conditions for the solution of the shear stresses and the angle of twist.
A6.11 stress Distribution and Angle of Twist for 2-Cell Thin- Wall Closed Section.

For a two cell tube, the equations can be Simplified to give the values of ql' qa and 0 directly. For tubes with ~ore than two cells, the equations become too complicated, and thus the equations should be solved S1~ultaneously. Equations for ~No-cell tube (Fig. A6.18). -

AS.S

TORSION

Cell 2
----(29)

2GQ-367.4 .865 q1. - 5.34 qa

Fig. AS.18

- - - - - -

(34)

Equating (33) and (34)

- 14.195 q1. + 8.505 qa = 0 - - - - - - - - (35)

- - - - (30)

The summation o~ the external and internal reSisting torque must equal zero.
.040" 1 :: 25. 25"

=GTJ

- - - - - - - - - - - - - - - - - - (32)
+

External Torque 83450"* Cell @ 1::25.30"

where A = A1.

Aa

I
I

.050"

A6.12 &ample Problems of. Torsional Stresses in MulUple-Cell-Thin-Walled Tubes. Erample 1 - Torsional Stresses in Un-symmetrical TwoCell-Tube. Fig. A6.19

='
, .032"

Fig. A6.19 shows a typical 2-cell tubular section as formed by a conventional airfoil shape, and having one interior web. An external applied torque T of 83450 in. lb. is assumed acting as shown. The internal shear res ist tng pattern is reqUired. Calculation at Cell Constants Cell areas: - A, = 105.8 sq. In. A. = 387.4 sq. In. A = 493.2 sq. In. Line integrals a = ~s :

83450 - 2 X 105.8 q, - 2 x 387.4 q.

=0

- (36)

Solving equations (35) and (36), q1. = 55.S#/in. and q. = 9Z.St/in. Since results come out ?OSitiv6 1 the assumed direction of coun~er clockwise was correct for q~ and q~ or true signs are q1. ~ - 55.6 and qa - - ~2.5.
q1.~

from Cell 1).

=- 55.6

92.5

= 36.9#/in.

(as Viewed

=26.9 = 1075; 025

aao

au

=~ .04 = 335
032

Fig. A6.20 shows the resulting shear ,attern. The angular twist of the complete cell can be found by SUbstituting values of q1. and ~~ in either equations (33) or (34), since twist ot each cell ~ust be the same and equal to twist of tube as a whole .

= ---:o:;r 25.25

+ ~ + 25.3

05

=''''35 (

Solution by equating angular ~Nist of each cell. General eQuat~on 2GQ ~ ~s. Clockwise tlow at q is pOSitive. Cell 1 Subt. in general equation

tf

2Gg

105.8 q1.

r-

x 1075
qa

SOLUTION BY SUBSTITUTING IN EQUATIONS (29) & (30) - - - - - - - - - - (33)

- 13.33

+ 3.165

~&iwr'

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

A6.9

002456 T
1735 x 105.8 + 335 X 493.2 ] T [ 1735 X 105.a u + 335 x 493.2~ + 1075 X 387.4 3 2

000665 T 000665 x 83450 55.St/ln.

=4

r. 916.7 x 100a La66.7 x 333.3

=
T

+ +

333.3 x: zona + 866.7 x 100 333.3 x 916.7 + 916.7 x 866.7

= 89.76
[ 1735
X

1075 X 387.4 + 335 x 493.2 105.8 3 + 335 X 493.2~ + 1075

] T
X

387.4 3

"2

Q=JT G

G X 89.76

= .01116 ~ (rad.) per unit

length of cell.

= .001107 T = .001107
Example Problem 2.

X 83450

= 9Z.5t/ln.

A6.13 Example 3 - three-Ceil-Tube.

Determine the torsional shear stresses in the symmetrical 2 cell section of Fig. A6.21 when subjected to a torque T. Neglect any resistance of stringers in resisting torsional moment.
SOLUTION:
~

Fig. A6.22 shows a thin-walled tubular section composed of three cells. The internal shear flow pattern will be determined in reSisting the external torque at lOo,OOO~t as shown.

Flg.A6.21

T
q.

Calculation of terms Area of cer.fa: A l = 100 Ali A Al + A3 100 200 Line integrals a =

.03"

,.03"

Iq05To'

o:l<XO'i

.031
.03

CeU@
. 03

Fig. ,A6. 22

SOLUTION:

Calculation at cell constants

Cell areas: Al.

fdtS:a -

1.0

10 + ~ .05 .03

= 866.7

= 39.3
a =

As

100

Line integrals

L~ J t

au

= .~~

= 333.3
a"

10 = 2nxx .025 = 629

au a

.05

10

= 200
333

- 20 altO -:03

+..l2... .04 = 916.7

_ 20 a a e - .03 = 667
a;so

=.l2...= .03

solution of Equations from Article A6.1I: -

.03

ZO

+.lQ.. =
.04

917

Equat1n~ the external torque to the internal resisttng torque: -

1 916.7 x 100 + 333.3 x 200 q.l.=ZL916.7 X 100 2 + ~.3 x 200 2 + 866 . 7

r.

= .00254D

subat i tut Ing r T

78.6 q.l. + 200 q~ + 200 q~ - 100,000 = 0

_ (37)

Writing the expresston for the angular twist of each cell: -

-.1[

866.7 x 100 + 333.3 x 200 ;::; 916.7 x 100 2 + 333.3 X 200 5 + 866.7

Cell (1)
2G1d

= All.

[q;>al.O (q, - q,)

a,.]

A6.10

TORSION
A6.14 Torsional Shear Flow in Multiple Cell Beams by

SUbstituting:
1.3 209 = 39 [629 q, + 200 q, - 200 q,J - - - (38)
Cell (2)

Method of Successive Corrections.

trend in airplane wing struct~al design particularly in high speed airplanes is toward ~ultiple cell arrangement as illustrated in :1g. A6.24, n~.ely a Wing cross-section made up of a relatively large number of cells.
~he

c:=:ITO 0,---,---,---1,--Subat I t.ut tng r

Fig. AB.24

With one unknaNn shear :low q tor each cell,

roe =

l~O

[200 qlil - 200 q , + 667

c,

3:33 q , -

- - - - - - - - - - - - - - - - - (39)
Cell (3)

the solution by the previous eq~tions becomes Quite labor! ous , The method of successive approximations provides a simple, rapid method :or fl~dlng the shear flow in multiple cells under pure torsion.
EXPLANATION OF
SUCG~SSI'lE COFL~ECTION ~HCD.

SUbstituting:
2i}Q

l~

[333 q;) - 333 q.a

917

q~J

Consider a ~HO cell tube as shown in Fig. a. To begin with assume each cell as act1ng independently, and subject cell (1) to such a shear flow QJ. Iceu~ Cell as to make GQJ. = 1. (2) (1) From equation .04 .05 "'J )0" (19) we can 'NTite, I ~s- - - (40)
Gg=..s..~dS
2A
A ..=89.3

'-. . __-...l.'~_"/'U
!- 5" -1
_ Fig. a

Solving equations (37) to (40), we obtain,

q, = 143.4#/in.

Now assume GQ = 1, then

q, = 234.1Vln. q, = 208.e*/ln. q, q, = 90.7*/ln. q. q, = 25:3*/ln.

Since practically all cellular aircraft beams have waLl, and web panels of constant thickness for each part~cular unit, the term for s tmpt tc t tv will be written Z ~, where L equals the length of a wallar web panel and t its thickness. Thus we can write,

p d:

Fig. AB. 23

= -----;L :E cel l t

:2 A

(41)

234. H/in.

Therefore assuming GQ~ = 1 fer cell (1) of Fig. a, we can write ~rom equation (41): _
Based on Paper, "Numerical Transformation Procedures for Shear Flow Calculations" by S. U. Benscoter. Journal, Institute of Aeronautical Sciences, Aug. 1946.

Fig. A6.23 shows the resulting internal shear flow pattern. ~he angle of twist, if desired, can be found by substituting values of shear flows in any of the equations (38) to (40).

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A6.1l

--

2 X 89.3 25.7 10, .04 .05

.-I
I

0.212 lb./in.

(1-2) is SUbjected to a shear flow qJ1l

lin. (counterclockwise with respect to cell (1)

=-

.109#

Fig. b. shows the results.

r-:--. I
,.212

I~

q '" . 109

j
~

(11

I
j

(2)

Fig. b

Fig. c

In a similar manner assume cell jected to a shear flow qa to make GQ 3

2 x 39.3

(2)

= 1.

subThen

and therefore negative), in addition to the shear flow q1. = .212 of cell (1). The negative shear flow qa .109 on web 1-2 decreases the ~Nist of cell (1) as calculated above with the resulting value tor G91. = 0.875 instead of 1.0 as started with. Thus in order to make 091. = 1 again, we will have to add a constant shear flow q'l to cell (1) which will cancel the negative ~Nist due to q~ acting on web (1-2). Since we are considering only cell (1) we can compare cell wall strains instead of cell twist since in equation (41), the term 2A is constant. Thus adding a constant shear flow qi to cancel influences at qJ1l on web l-?, we can ve-tta:

=-

l.Q.
.05

15.7
.03

78.6

723 :. 0.lOS#/1n.

qi

(z

hence

cell (l) - q,

(~) 'Neb 1-2 = 0

- - - - - - - - (42)

Fig. c shows the results. Now assume the two cells are joined together with the ~nterior web (1-2) as a common part at both cells. See Fig. d. The inq w 109 ter i or; web 15 now Subjected to a resultant shear flow at q1. - qJ1l = (.212 .109) .103t/ln. Obviously this change of shear flow on the Fig. d interior web will cause the cell twist to be different tor each cell instead of the same when the cells were considered acting separately. To verify this conclusion the twist measured by the term OQ will be com,uted for each cell.

(If) 'Neb 1-2 ] [ (z i') cell (1)

SUbstituting values in equation (42)

q1. - q~

.- [ .~ +....!2..J_200 - 842 q3 =
~

.04

.05

Cen (1),

GQ. ' -

2A, Z q-t -

L _

7 . + (2' [212 x 25 - 2 x 1 89.3 .C4 ..... 2 - .109)

.~;]

Thus to make GQ1. equal to 1 we must correct the shear flow in cell (1) by adding a constant shear flow equal to .237 tl~es the shear flow qJ1l in cell (2) which equals .237 x .109 = .0258~/in. Since this shear :low is in te~ of the shear flow qJ1l of the adjacent cell it will be referred to as a correction carrv over shear flow, and will consist of a carryover correction tactor times qJ1l. Thus the carryover factor tram cell (2) to cell (1) may be written as

=
Cen (2),

0.375 C.O.F. (2 to

11

Jf)
-(z

'Neb (1-2)

~)

which equals
cell (1)

~S.3

- - (.212 - .109) [.109 x 03

0.4375

15.7

.~;J

.237 as found above in substitution in equation (42) Now consider cell (2) in Fig. d. In

Since 91. must eq~l QJ1I if ~he cress-section is not to c~stort from its original shape, it is eVident ~hat the above shear :lows ~e not the true ones w~en the two cells act together as a unit. Now consider cell (1) in Fig. d. In br-tng-. ing up and attaching cell (2) the common web

bringlr.g up ane attaching cell (1), the common web (2-1) is subjected to a shear flow of Q1. = - O.212*/ln. (co~~terc1~c~vise as viewed from cell 2 and ~~eretore negative). This additior~l shear ~low c~~nges GQ 3 twist of cell (2) to a relative "al~e of 0.4375 instead Jf 1.0 (see preVious GGJ1I calculations). Therefore to ~ke OQ 3 equal to 1.0 again, a cor~ectlve constant shear flow q~ ~ust be added to ~ell (2) to

-----_. - --- -----------------------,-..,..,.,---,=

AS. 12

TOR S ION

cancel the twist effect of q~ = 0.212 on web (2-1). Therefore we can write,

Considering Cell (2)

q:

q.
hence

(z ~) cell

(2) -

q,

(~) web

= q~

(C.O.F.) 1 to 2

= .0256

(2-1) =

.227 .00717#/in.

Fig. f shows the result1ng second set ct corrective constant shear flows for each cell. Since our corrective shear flows are q, (~) web (2-1) q'k '"' .00717 (43) rapidly getting = sffialler, the con(z~) cell (2) {"', . 0139 tinuation of the process dependS on \~ (il (~' SUbstituting in equation (43): the degree of accuracy we wish for ,_ (200) the tinal results. .277 q, .277 x .212 .058711!1n. Fig. f q. - q~ \723 Suppose we a dd one more set of corThus the carryover factor trom cell (1) to cell (2) in terms at q~ to make GQ~ = 1 again can rective constant shear flows q~ and q:, Us tng the carryover correction factors previOusly be written found we obtain,

q.

;:;.-- -

II \
I.

---

jI

C.O.F. (1 to 2)

(z ~) cell

(--) web

(2-1)
(2)

= 200 =
723

.277

qT
q~

.237 x .00717 = .0017~/in.

= .277 x .0139 = .00385#/in.

Fig. e shows the constant shear flow q~ and GQ = 1 for each cell. However these corrective shear flows q'~ e , 0587 were added assuming the cells were again independent ot each other or did not have the cammon web

Fig. g shows the result s . The final or resulting cell shear flows then equal the original shear flows plus all corrective cell shear flOWS, or
:;....-_ _
q'~

qt that were added to make

'" . 00385

0i,~=.OOi71j
\ (I)

t!(21)

(1-2).

ThUS in

Fig. e

--Fig. g
1

bringing the cells together again the Interior web Is SUbjected to be resultant shear flow of q~ - q~. In other wordS it we were to add the shear flows of Fig. e to those of Fig. d, we would not have GQ 1 and GQ~ equal to 1. The resulting values would be closer to 1.0 than were r OUIld tor the shear flow system or Fig. d. Considering Fig. e, we will now add a second to cells set ot correctIve shear flows q~ and (1) and (2) respectively to ~ke GQ, and GQ~ = 1 tor cells acting independently. Considering cell (1), and proceeding with same reasoning as betore~

(r inat )

q. + q'. + q". + qm.

q:

Fig. h shows the final results. To check the final twist at each cell the value GQ will be computed for each cell using the q values in Fig. h.

:;;..-

r . 1787

I~

I'q,=

.2534
(1)

ll(~

fll I

~-Fig. h

cen

(1)

1 [~534 25.7 x "9::; '-' ..... .- ~ x -04

+ .0747

.~~J

'"

.997

Hence
q~

=-,..,.,c'-""--- = .0587

1-2

x .237

= .013911!in.

GQ. =

2 x 139.3 [

.1787

10] x 15.7 .03 - .0747 x .C5 = .997

or
q~

AS.15 Use o! Operations Table to Organize Solution by Successive Corrections.

(C.OS.)

2 to 1 times q~

.237 x .0587 = .0139

Operations Table 1 arranges the calculations so that the steps jealing with the corrective shear flows can be carried out rapidly and with a minimum of thought.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

OPIRA T IOKS TABU 1
Cdl 1

A6.13

Cell 2

C.O. FactoX'
AssUIIled

.277 3

2 J

a a

237

C.O. (1-2) =

(~)

web 1-2

10

CO C.O.
T 2A 0," T

02~8

<15. zo"

00

~
.3""

~~
14 1".

z (t) cell

=
2

.05
..lQ. + ...2Q.

.40

.05

0.1

C.O. (2-1) =

m (~l

web 2-1
cell 1

10

= .~5 = .25

:05

.Explanation of Table 1 Line 1 gives the carryover factors for each cell, computed as explained before. Line 2 gives the necessary constant shear flow q in each cell to give unit rate of twist to each cell acting independently. (GQ = I). Line 3 gives the first set of constant corrective shear flows to add to each cell. The corrective q referred to as the carryover q or C.O.q in the table consists of the q in the adjacent cell times the C.O. factor of that cell. Thus .237 x .109 = .0258 is carryover from cell (2) to cell (1) and .277 x .212 = .0587 is carryover from cell (1) to cell (2). Line 4 gives the second set of corrective carryover shear flows, namely .277 x .0258 = .00717 to cell (2) and .237 x .0587 - .0139 to cell (1). Line 5 repeats the corrective carry over process once more. Line 6 gives the final q values which equal the original q plus all carryover q values.
Example Problem 1 (2 cells) Determine the ~nternal shear flaN system for the ~NO cell tube in Fig. A6.25 when subjected to a torque of 20,000 in. lbs.

Line 2 gives the shear flow q in each cell, when it Is assumed each cell is acting separately and is subjected to a unit rate at twist or GQ - 1. The calculations for the q values are as follows: For cell (1 )

q.
Z

2A.

(~)

cell 1

2 x 100 25 40

.05

For cell (2 )

q.
Z

2A.

(~)

cell 2

2 x 100 4 10 + ~ .05 0.1

A. and A. equal cell area of cells (1) and (2)

respectively. L - length of wall element and t its corresponding thickness. In order not to start out with dec1lllal values of q, the values above will arbitrarily be multiplied by 100 to make q. = 25#/in. and q2 :; 4Oi/in. Since we want only relative values of terms this is permiSSible. These values are shown in line 2 of Table (2). The corrective carryover process proper is started in 11ne 3 of table (2). In cell (1) the amount carried over of q. = 25 to cell (2) equals the C.O. factor times 25 or 0.4 x 25 = 10 Which is written along the vertical line under cell (2). Likewise in cell (2) the amount of q. - 40 that Is carried over to cell (1) equals the C.O. factor t !mes q. or .25 x 40 = 10, .mich is written along the vertical line under cell (1). The second set of corrective carryover constant shear flows are given in line 4 ot Table (2), thus, .4 of the q. :: 10 :; 4.0 is carried over to cell (2) and .25 x value of q. = 10 = 2.5 is carried over to cell (1). Line 5 repeats the process namely 0.4 x 2.5 = I to cell (2) and .25 x 4 = 1.0 to cell (1). This process of carrying over 'lalues of q is continued until the values are small or negligible. Line 8 gives the final q values in each cell as the summation of the assumed q value plus all carryover values of q. Thus in cell (1) q 38.85#/in. and for cell (2) q = 55.5. Line 9 gives the ~orque that these values of cell shear flow can produce. T = 2Aq

r
10"

I-.05

10"

-----.J.-.05

10"

-----I
0.10

.05
Cell 1

0.10
Cell

2
0.10

1
1

.05

Fig. A6.25
OPERlTIOMS TA8U 2

Cell 1

Cell 2 0
O.2~

,
.

CO. l"ac'toJ:"
l!IU11ed

0.' 10
2. S

.0 0

C.O

C.O
0.

10

>n.

0 11

2A 0," T for ~20 000

0 10

,
11100 18870

7770 41

Explanation of solution as given in

~able

(2):

Line 1 gives the values of the carryover factors. The values are calculated as follows: C.O. factor cell (1) to cell (2):

AG.14

TORSION
OPERATIONS TABLE 3 Cell 1 Cell 2 Call 3 .288
o.
I
,

For cell (1) T = 2 x 100 x 38.85 = 7770 in. lb. For cell (2) T = 2 x 100 x 55.5 = 11100 in. lb. Line 10 gives the sum at th9 above two val~as which equals 18870 in. lb. The original requirement of ~he problem was the shear flow system ~or a torque of 20,000"#. Therefore the required q values follow by direct proportion, whence
1

a
3

co
CO 0 C 0 C.O.

Factor

Assumed

, 6
7

23 25.30 3 88

.21

, ,.
10
T

co.

1 " 505 0.244 0.072

18.45 2 3. 394 1 51 0.430 0.13 0.116 O. 0.056

.'88

2.53 0.685

.33
.0

=I88'iO x
q~

20000

38.85

= 41.2#/ln.
= 58.9i/ln.

11

2215 9200 6020 19000 + 6020 + 2215 ~ 1 43S'"1lI Total ~ q _ to;,: T=100,000 183 112.5 '"0

"

20000 = I8870

x 55.5

Explanation of
Cell (1)

sol~tion

as given in Table 3: Cell (3 )

Cell (2 )

These values are shown in line 11 of Table 2. Check on twist of cells under fir~l q values. The relative total strain around each cell 1 boundary is given by the term. A z q tL for the cell. Thus tor cell (1)

"i =

A~

.: 144

1140

= 96 "t= 1040
A.
L

A.
L

= 56. 5

Zt::; 1055

Shear flow q for GQ 1 for each cell acting in dependently: Cell (1)
q

= 2A, =
"~
t

288 = 1140 .253

=
For call (2),

212

Cell (2)

q .: 2A.~

Z L t

192 1040

.1845

L z q~ =

1;0 [(58.9 - 41.2)

=212

.~~

,. 58.9

~03.01)J

Cell (3)

=z 2Ai' = 113 = 1:.. 1055

t

.107

Thus both cells have the same twist. In the above calculations q~ and q~ act clockwise in each cell, hence the shear flow on the !nter1or common Neb 15 the difference of ~~e two q values. Example Problem 2. Three cells The three cell structure in Fig. A6.26 is Subjected to an external torque of - 100,000 in. lb. Determine the internal reSisting shear flow pattern.
I - 12"

To avoid small numbers t~ese values of q are ~ultiplied ~y 100 and er.tered on line 2 of the table. Calculation of carry factors as given in line 1 of Table 3.
Cell (1) to (2)

_(f) web 1-2

C.O. (1-2)

"( ~)
=

cell 2

- 1040 -

240 - .23

Cell

(2) to (1 )

C.O.
I
I

8" .032

---I,

(2-1)

m
(3 )

web 2-1

!(~)cell

240 ;:rw = .21

.04

i
12" .04 Cell 1
.0' Cell 2
I

I
.041 \

Cell (2 ) to

...i...

.04

0032

!y

'Cell 3

I
C. O.

(2-3)

(~)

web 2-3

_ 300

z(~)ce113

- 1055

.234

Fig. A6.26

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

A6.15

cell (3) to (2)

c .0.

_ (i-) web 2-3

(3-2)
- Z

as given in TablA 4. Line 10 shows the correction of q values to develop a reSisting torque of 100,000n#. The ~ultlplylng factor 100,000/80630.

:5

(~)

cell 2

The balance at the solution or procedure in Table 3 is the same as explained for Problem 1 in Table (2). It should be noticed that cell (2) being between two cells receives carryover q ~lues from both adjacent cells and these ~NO values are added together before being distributed or carried over again to adjacent cells. ?or example consider line 3 in Table (3) and cell (2). The q value 5.82 representing .23 x 25.3 is brought over fr~ cell (l) and the q value 3.CB 288 x 10.70 from cell (3). These two values are added together or 5.82 + 3.08 8.90. Pne car-r-y over q to cell (1) is then .21 x 8.90 = 1.87 and to cell (3) is .224 x 8.90 = 2.53. In line 8, the final q in cell (2) equals the original q of 18.45 plus all carryover q values from each adjacent cell. Line 10 in Table (3) shows the total torque developed by t~e resultant internal shear flow is 17435"~. Since the problem was to find the shear flow system for a torque of - 100,000 in. lb., the values of q in line 8 must ~e multiplied by the factor 100,000/17435. Line ~l shows the final q values. Example Problem 3. Four cells. Determine the ir.ternal shear flow system for the four cell structure in Fig. A6.27 when SUbjected to a torsional moment of - 100,000 tn.
lb.
40
(1600)

A6.16 Torsion of Thin-Walled Cylinder HaVing Closed Type Stiffeners. The airplane thin-walled structure usually corrta tns ron-t tuc inat stiffeners spaced around the outer walls as illustrated in Fi~:. A6.28 and A6.29.

Open Type StHfener

Closed Type Stiffener

.025

For the open type stiffener as illustrated in Fig. A6.28, the torsional rigidity of the individual stiffeners as compared t 0 the torsional rigidity of the thin-wal!ed cell 1s so small to be negligible. However a closed type stiffener is essentially a small sized tube and its stiffness 15 much greater than an open section of similar size. Thus a cell with closed type stiffeners attached to its outer walls could be handled as a multiple cell structure, With each st_tfener acting as a cell with a common wall With the outside surrounding cell. Since in general the st11'fness provided by the stiffeners 1s comparatively small compared to the over-all cell, the approximate Simplified procedure as given in NAeA T.N. 542 by Kuhn can be used to usually give sufficient accuracy. In this appraximate method, the thin-walled tube and closed stiffeners are converted or transformed into a Single thin-walled tube by modifying the closed stiffeners by either one of the follOWing procedures: -

16"
(500) Cell 2
(625)
(1600)

14001

(1) Replace each closed stiffener by a doubler

plate having an effective thickness te=tSTs!d, and calculate dslt with these doubler plates in place. The enclosed area of the torsion tube still re~lns (A) or the same. See Fig. A6.30.
d

cen

Cell 3

Cell 4

Fig. A6.27
OPDATIO!(S TABLE
.oj,

Cell 1
1 2 3

Cel]. 2
20. .153 38.8 3 88 , 1< 2 00 188 0.47 022 o '0 12

cell 3
181 .078 28.4 3 as 1. 38 0.36 0.57 0 2 o O. o 02

Cell

.oj,

CO,
0 0 0

s
a
9 10

17' 22.4 1 e 1 69 1 10

",

.16", 20.5 2 22 O. 3 J

Original
~iiiener

o rs

7620 47.1

31560

22550 49.9

8900 29.41

by

Transiormation Procedure (1)

I,-

65.3

----oot , .tsK
I

In Fig. A6.27 the values in the rectangles represent the cell areas. The values in ( ) represent ~he Lit values for the particular wall or web. After studying example problems I and 2 one should r2ve no trouble checking the values

lte:::tsK+tst Fig. AG. 30

,by

Transformation Procedure (2)

Fig. AG.31

t:::tst ... t s K ~

A6.16

TORSION

(2) Replace the skin over each stiffener by a

"liner" 1n the stiffener having a thickness e = tSK dis' (See F'ig. A6.31.) The enclosed area (A) of the cell now equals the original area less that area cross-latched in Fig. A6.31. Procedure (1) slightly overestimates and procedure (2) slightly underestimates the stiffness effect of the stiffeners. The corner members of a stiffened cell are usually open or solid sections and thus their torsional resistance can be simply added to the torsional sti:fness of the thin-walled over-all cell.
t

Lateral shear, H , - mainly ; :-Partiy lateral shear and partl y torsional shear/:-"""",V

fiW

A6.1'l Effect of End RestraJ.nt on Members Carrying Torsion.

Fig. A6.32

The equations derived in the previous part at this Chapter assumed that cross-sections throughout the length of the torsion crembers were free to warp out ot their plane and thus there could be no stresses normal to the crosssections. In actual practical structures restraint against thiS free warping at sections is however Often present. For example, the airplane cantilever wing from its attachment to a rather rigid fuselage structure is restrained against '~rping at the wing-fuselage attachment pOint. Another example of restraint is a heavy wing balkhead such as those carr-ying a landing gear or power plant reaction. The flanges of these heavy balkheads Often possess considerable lateral bending stitfness, hence they tend to prevent warping ot the wing cross-section. Since only torsional forces are being considered here as being applied to the member, the stresses prOduced normal to the cross-section of the member namely, tension and compression must add up zero tor eqUilibrtum. Thus the applied torque is carried by pure torsion action at the member and part by the Iongt tuo i ona.i stresses normal to the member cross-sections. The percentage ot the total torque carried by each action dependS on the dimensions and shape at the cross-section and the length ot the member. Fig. A6.32 illustrates the distortion of an open section, nanely, a channel section subjected to a pure torsional force T at its :ree end and fixed at the other or supporting end. Near the fixed end the applied torque is practically all resisted by the lateral bending at the top and bottom legs of the channel acting as short cantilever beams, thus torming the couple With n forces as illustrated in the Figure. Near the !ree end of the ~emberJ these top and bottom legs are now very long cantilever beams and thus their bending rigidity 1s small and thus the pure torsional rigidity of the Section in this region is greater than the bending rigidity at the channel legs.

AS.IS Example Problem illustrating Effect ot End .. Restraint on a Member in Torsion.

Fig. A6.33 shows ~ I-bea~ s~bjected to a torSional moment T at its free e~d. The prcblem will be to deterT-ine ~rEt proportien of the torque T is taken by ~he flanges in bending and what proportion by pure shear, at two different sections, namely 10 inches ~nd 40 incheS trom the fixed end of the I-beam. .,.u"",""-,+- Fixed
End

r--L 75''-1
F

Fhi .4";#~~ Fig. A6. 33 FIg. A6. 34

'1 3

T'

FIg. A6. 35

Fig. A6.34 s~ows the torque diViding into VNO parts, namely the couple force F-F ~armed by bending of the flanges of ~he I-bea~ and the pure shearln~ stress system on :he crosssection. :<'ig. A6.35 shows the twisting 0:' tne Section through a distance 5. The solution will consist in computing the angle of twist 9 under the two stress conditions and equat1nz them. Let Ta be the proportion of the total torque T carried by the fl2nges in bending forming the couple F-F in Fig. A6.34. From Fig. A6.35, the angle of tNist can be written
2 Tg L"
3

na-IT

.. - ,.
~-

~.

...

~.

ANALYSIS

AND DESIGN OF

FLIGHT VEHICLE

STRUCTURES

A6.17

Note: The deflection of a cantilever ~e~ with a load F at its end equals FL~/3EI, and I the moment of inertia of a rectangle about Its c~n ter axis tb~/12. hence Now let Tt be the ?ortlon ot the tctal torque carried by the ~e~ber in pure torsion. T~e approximate sol~tion for open sections co~posed of rectang~lar ele~ents as given in Art. A6.6, equation (12) will be used.

7he shaft rotates at constant Soeec. The difference in be!t pull on two sides of a p~lley are shown. an tne figure. Calculate the ::2...'(1~~ torsional shearing stress in the 1-3/4 inch shaft bet~een pulleys (1) and (2) and between
(2) and (3).

9t
Equating GS to tit

3 Tr L
Gt'(b + b w)
''iT'
2

(2) rl. 1/2 HP. motor operating at lOCO RR~ rotateS a 3/4-.035 al~~lnum alloy torque tuee 30 inches long which drives t~e gear =ectan~~ tor ~p&ra~ln; a Wing flap. Dete~ine the ~~<: ~~ torslcnal st~ess in the torque sha:t ~cer fUll ~~wer and RPM. Find ~he ang~lar de:lec~ian of shaft in the 30 lr.ch ~ength. ?olar =c~ar.~ ot inertia of tube = .01 in. ~odulus ot r:gidity G = 3300,000 psi.
Tni.n.lb.

we can
3 E h
8L

i te

b3
+

2Ge(2b

b. n)

I I / I

,/

I
I
/

Substituting values when L : 10 inches.

I

'f--- I- "

3 x 10.5

10 4

3.4 x 1.75

io-

X 3.7 X 10 4 X 0.1 11(2 X 1.75 + 3.3}

= 9.65
taken
.T

Fig. AS, 36

'rner-ercre , percent of total t cr-s tona L T :y :e:'.d.ir.g of flanges e~~als,

':'8

-15" :-

.... '1' (100) =,,~:-+"=' ...-(lCC): 90.5 percent. 3 . -t -&

'!'c

Ide t.ertafne t:r.etne crs ona ;

".. (3) I .
~,

__ " "," e ..,.. '!!: "\._ '{:: .;.0 -ee snea r :lo'iJ :::. :-es:~'::::g I tne ext er-na L t orque 0: scccc :n.lb. -,.;,:2: ::.:.:. wall t'u cxnees ar-e gi '''e:: or: tno r:;,;u.:-e. As "'\.::",e the tube 1s 100 In. long and :in~ :~e :c::::~3.1
~

"01'1""'''' __ ..1."" r:: """'

t

.J_

de f t i cn , za cer 15 <:ll:"''"ll.L'1.u::l. (J-= It we ccnsid.er the section 4Q inches :~c~ 32CO , COO ps 1. ) the fixed end, then L = 40 inches. Thus if 40 is orac ec in the above subs t I t at Lcn inst eac of (4) In fig. A6.36 :"2r.:lC':,9 :~e Irrt er-i cr- .035 10 11 the r~s~l:s for TS/T t WQuld be 0.602 and web and cv~pute torsional shear flow and ~e the percent or the torGue carried by ~he rl~~ges rt.ect ions . in bending wOuld be 36 as com~ared to 30.5 perT '" 100,000 in. lb. cent at L = 10 inc~es fro~ support. Thus in general the e::ect of the end restraint decreaSes rapidly with increasing value of L. The ef!9c~ af end restraint on thin walled tUbes with lc~glt~dlcnal stiffeners 1s a more Invc Lved pr ob Lera and cannot be nancj.ec in sucn a si~p1::ied =a~ner. This ?roble~ 15 ccnsi~erec I in other Chapters.
L e c i a t a.I L o y ,

f\

A6.18 Problems.

Fi.g. AS. 37

320*

Fig. A6.35

(1} In Fig. A6.36 ?~lley (1) is ~~e 1rivl~g PUlley and (2) and (3) are the driven pulleys,

(5) In the 3-cell structure of fig. A6.37 detieratne the internal res t st tng shear ilow due to exter~.al torque of 100,000 in. lb. ?or a length of 100 inches calculate twist of cellular structure i: G is assumAd 3,800,000 psi.

A6.18

TORSION
~14'

(6) Remove the .05 interior web of Fig. A6.37 and calculate shear flow and twist. (7) Remove both interior webs of F:g. A6.37 and calculate shear flow and twist. (8) Each of ~he cellular structures in Fig. A6.38 is subjected to a torsional noment of 120,000 In.lbs. USing the ~ethod of successive approximation calculate reststing shear flow pattern.
.0'

14"---i .03
.0'

05 1. .03

.0>" 10" .03

I I

(01

-L

fo-6" --t- 8"

T
10" .04

.03
04 .03

.03

, !
I

10" .03

10" ......l
.03 .03 .03 .0' .03 I (b]

.035
.03

~035

.03

1-==::::::::===

I- 12" '-1-10"-+-8" ........... 8" +8" -+-8"""';

All interior wece e . 05 thickness Top skin .084" thickness Bottom skin .064" thickness Fig. A6.38

I i i I'"

'. ,
,.
The big helicopters of the future will be used in ~ny important industrial and military operat t ons , The helicopter presents many challenging problems for the structures engineer.
(Sketches from United Aircraft Corp. Publication "BEE~HIVE", Jan. 1958. Sikorsky Helicopters)


CHAPTER Ai

DEFLECTIONS OF STRUCTURES
ALFRED F. SCHMITT'

A7. 1 Introduction.
i~portant

Calculations of structural deflections are for tNO reasons: -

(1) A knowledge of the load-defo~ation characteristics of the air?lane 1s of prllnary ireportance in st~dles of the influence of structural flexibility upon airplane performance. (2) Calculations of deflections are necessary in solving for tt.e In~ernal load distributions of complex redundant structures.

Work =-

L: Pi
I~l

6 Oi

in the limit as

60-0
Work

=Lt'J P

dO

The elastic deflection of a structure under load i s the cumuLat;'... ''j result of the strain deformation of the individual ~lements composing the structure. AS SUCh, one method of solution for the t~tal deflection ~ight in,o~ve a vectorial addition of these individual :ontriJutions. ~he involved geometry of most ?ract~cal str~ctures ~kes such an approach prohiJitively ditf~c~lt. For complex structures the more popular tec~~lques ~re analytical rather than vectorial. 7hey::ieal dt r sc t Ly 'Nith quarrt t t Les which are not themselves ceflections but from which deflections ~y be obtained by SUitable operations. ~he nethods employed herein tor deflection calculations are analytical in nature.
A7.2 Work and Strain Energy.

Fig. A7.1

curve ls a straight line whose equation 1s P = k6 and the strain ener~J is readily computed to be p' U equally, U = 2k'
A7.3 Strain Energy EXl)ressions for Various Loadings.

STRAIN SNERGY OF !ENS I ON

A tens 11e load S aniform bar of length A and elastic ~odulus a = SL/AE Hence S

acting at ~he end of a L, cross sectional area E causes a deformation AE a and

L

as defined in ~echanlcs 1s the product of rcr-ce t ines cr stence . It the rorce v~aries over the distance then the work 1s computed by the integral calculus. Thus the work cone by a varying force P in deforming a body an aaount 6 is ',.Jork = Pd.a and is represented by the area ander the load deformation (P-6) curve as shown in Pig. A7.1. If the tefor.ned body 1s per~ectly elastic the energy stored i~ the body nay oe co~pletely r-ec over-ed , the bocy un Loac tng a Lcng the same ?-6 c'Jrve :ollowee for increasing load. T~is energy is ca l Ied the elastic ~ energy q GeforT.aticn (hersa~ter ~he strain energy, for brevi~y) and is denoted by ~~e s~bol U. Thus U = Pdc. Should the cody '_be .Li near Iy
~ork

'Q

Sd5 AE 5'

(1)

Alternately.
U'

S'L
2AE - - -

(2 )

.r

Squations (1) and (2) are equivalent expressions ~Jr the strain energy in a Qni~o~ bar, the farner expressl~g U in terms of the defo~~tior. and the latter expressing U in~e~ of load. 7he secand form of expression is ~ore 80nvenlent for ~eneral usage and succeeding strain energy :ornulas will be ~ut in this form. ?cr a tensile bar haVing non-uni:orm properties (varying AE). or for which the axial load S varies, the strain energy is computed by the calculus. Thus the energy in a dir:erentlal element of length dx is atven by eq , (2) as
dU :: ;;'2AE Where S and AE

elastic (as are ~ost bodles jealt w1~h in structural analYSiS) t~en the load-defo~tion
(1) Design gpec iarist, Convair-Astronautics

s-ex

are average ';alues over the

A7.1

~ength

of element.

A7.2

DEFLeCTIONS OF STHUCTUReS

The total energy in the bar is obtai~ed oy summ:ng over tr.e length of the ba,.
U

U
(3)

=j

dU

=j

"2AE

~.adx

2 ~ '" JL ( n"'AE
c

1-

, 2L nx)' ex . S.Ln

Example ProbleT. 1

01'5" 1.... '0" '-.:; I ) (22'7\(4-0) '" 0.0920 in Ics , 9.87(2)(10 x 2.0 5 ) ' !
STEArN SN~GY C? FLZ:('l)ES
~~ len;th L ~~er the actien of a pure couple ~~jer~oes an angular ro:atlcn proportional :0 the couple. Thus, from elementary strength of Taterials

linea~ly

tapered

al~nin~ll

an axial load of 15Ca Ibs , '1S snovr.

Jar is under .n Fig ..:;'7.2.

Find U.
A,

~ifo~ bea~

TU>AIXI =A,II _A) 40" x 80 L A.,= 2 in" t 1500. Fig. A7.2

Solution:

(E-a~ ~:=..1 f
-15CO#

Fram statics. the internal load S at every section.

S'dx

where the constant S1, the product of elastic modulus by sect~on ~oment of inertia. is called the ~bendlng stif:ness R Since the rotation 0 builds up linearly with M the strain energy stored is
U =

2AE =
In 2 I

= 21-1500)'],' d~ = 4 5 10' 2-1:'

,

h 1b ncr s

( h::tt:/40 intl'odl.leed tal' conven_ ience)

1 2:

MO

or
U =

2 Ef

M'L

14,)

Note that although P was a negative (compreSSive) load the strain energy remained positive. Example Problem 2 Find U in a uniform bar under a load _ nx( q - q , cos 2I Fig .40.7.3).
~inlng

Fo~ a jeam haVing ncn-~~i~orn prcpert~es (varying El) and/or for which M varies along the beam, the strain energy 1s com;uted by the calculus. From sq. (4) the str~in energy in a beam element ot 1ifferentlal length is

q(~ m1

~WT
I.

Hence summing over the complete beam to get the total strain energy one has
1. 40

40 "

.to 2 in

J: _ lO::do 6 pd

'to.

2'

U = j dU = ~ jl1'dx 2 EI
Exa~ple

(5 )

lb/in

Problem 3 the beam of Fig. A7.4 derive the strain energy expression as a function of Mo.
For~ r;.M GEl = constant

Flg~

A7.3
c.

Solution: The equation for

S

(x) was found

by

statics.

1 ! I-- L----.J
Fig. A7.4

Solution:

= q , 2L n sin 2L

mI'

The bending ~oment diagram (Fig. A7-4a) was found and an analytiC expression 'Nrltten ~or v

Substitution into eq. (3) gave

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A7.3

~=GJT

--------------

(6)

x
Fig. A7.4a

where the constant GJ, the product of the elastic modulus in shear by the cross section polar moment of inertia, 15 called the "torsional stiffness" Since the CNist ~ builds up linearly with T the strain ener~J stored 1s

Then

or
(7 )

Example Problem 4 Determine the strain energy of flexure of the beam of Fig. A7.5. Neglect shear strain energy.

varyin~

For a shaft of non-uniform properties and loading one has

HT~7'

--

(8)

GEl

=Constant
z:
L
--01

i-~ - .....Fig. A7.5

2"

In passing it is worth remarking that one often encounters the group symbol "GJ" in use for the torsional stiffness of a non-circular shaft or beam such as an aircraft-wing. In such a case the torsional stiftness has not been computed literally as G x J, but rather as-detined by eq, (5), viz. the ratio ot torque to rate of twist. Example Problem 5 For a certain flight condition the torque on an airplane wing due to aerodynamic loading is given as shown graphically in Fig. A7.6. The variation of torsional sti:tness GJ 1s given in like manner. Find the strain energy stored.
1.
01"":~-

Solution: The bending moment diagram was drawn first (Fig. A7.5a) and analytic expressions were written :cr 11.

__

Fig. A7.5a
Fig. A7.5

Inspection of the diagram revealed that the ener 5J of flexure in the right half of the beam ~ust be identical with that of the left half. Hence
U

o !_ _
Y=O

__

~_~

__

~_~_

Root

y=L Tip

Solution:

=2

STRAIN =G1 OF TORSION

A uniform circular sr~ft car~ying a torque T experiences a total CNist in a length L proportional to the torque. Thus from ele~entary strength ef ~terials

A numerical integration of eq. (8) 'NaS performed using Slmpson's rule. The work is shOwn in tabular ferm in table A7.1. Values of T and GJ ror selected wing stations were taken tram the graphs provided and were entered in the table. For convenience in ha~d11ng the numerical work all the variables were treated in ~on-di~ensional :OrTI, eq. (8) being changed as follows
I U - ~

i~' 0

.:::

GJ

1 LTRaJ'lila dy - -c-r 2 GJR 0 GJ

;,::~ ~.~<~--~
---~'. ':";\':,,,.

~:-~-:.i.?1:

,
A7

DEFLECTIONS

OF

STRUCTURES
~~er~J ~~ ~

where
T -

TlTn , 7=y/L
GJ/GJR

strain

beam under the

act~o~

of

GJ =

transverse shear loading V(l~s). lor this purpose V = ~ x dy x t and t x dy is called A, the ~eaw. cross sectional area. nance
dU
t~erefore t~e

The subscript R denotes Wrcot~ val~e (y=o). The coefficients for 31~pson's ~Jle appearing 1~ colunn (6) were taken from tte express~on for
odd n

= V'dx ZAG

and

=;'r
-

(r 0 + 4f J,. + 2f 2 + 4f 3 + - - - _ +

the enttre
U=

bea~

elastic strain energy of shear fer is ~lven by

_ -

15 f + 3f + ~f ) 4- n-z n-L 4- n

J-2AG

V'dx

- - - - - - - - - - - - - - - - (10,

Sxa:nple Problem 6

TABLE A7.1
STA.
7

Determine the strain energy of shear in the Jeam of Example Problem ~.

Coe!.
. 0667 .267

T
1.0
.98 .9'

(T)~

GJ

(T)'

ill
1.0
1. 26 1. 63 1. 22

m~ Coeff.
.0667 .336 216 305 070 0.0

Sc Iut t on :

1.0

1.0
.78

The shear load diagram was drawn first .

.8
.8

.07
2

."
a

.98 .88

.54
. 37 .24

.133 . 250
' . 200 . 0833

1.0

. 084

.35

.14

Po V (xli I! iii I i_!J~L! i ! ! Ii i ii iii

Ee"

~ 2

2 =-jP o
j

. . . . . . AG CONSTANT

z=
Therefore the strain energy was

994

Then
= LTR' _.- x .994
GJR

2 x

,L/2
0

Poa dx

::: Po aL 2AG
Eq , (9) may be used to comput.e the shear strain ener 5J ~n a thin sheet. The elemen~ dx x dy Is visualized as but one of many in the sheet and the total ener~J is oJtained by summing. Thus U

STRA IN ENERGY OF SHEAR

A rectan~~lar element "dx" by "dy" of thickness "t" into the page under the action of uniform shear stress ~ (psi) is shown in Fig.

A7.6a. From eIementar-y strength ot materials the angle of

shear strain J is proportional to the &hear stress -r as

=~

SS7't gx

dy

- - - - - - - - -

- - -

(11)

~ = ~ where G is the material elastic modulus Fig. A7.6a in shear. For the displacements as shown in the sketch only the down load on the lett hand side does any work. (In general all four sides ~ove, but if the mot~on is referred to axes lying along ~NO adjacent sides of the element, as was done here, the derivation is Simplified). This load is equal to 't' x dy x t and moves an amount a' x ex, Then
dU

Here a double integration 1s required, t~e summation being carried out in both directions over the sheet. In dealing with ~hin sheet the use or the shear flow q Tt 1s often convenient so that eq , (11) rewrites
U

=2 ~

Sf q'dx dy Gt
qas

(l1a)

A very important special case occurs when a homogenous sheet or constant thickness 1s analyzed assuming q 1s constant everywhere. In this case one has
_ q II U - 2Gt

=~

7Et dx dy

=2 G

7'

where sheet.
(9 )

JJ dxdy - 2Gt S = J5 dxdy is

- - - - - - - - - - - (12)

the surface area of the

tdxdy -

Eq. (9) may be used to compute the shear

The sketch is visualized as a side view of an element of length dx taken from a beam of total height dy.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A7.5

Example Problem 7 Find t~e shear strain energy stored in a cant~lever jox bea~ of uniform rectangular cross section under the action of torque T. (Fig. A7.8).

A7.4 The Theorems of Virtual WorK and Minimum Potential Energy.

An important relationsr.ip between load and deformation stems di~ectly from the definitions of work and strain energy. Consider Fig. A7.9(a).
p - --------6 U= PM
~u

j-~7
T" f.-b--l

Limit

flo .. 0,

dU = Pda

"------...JL.j- a
Fig. A7.9a

Fig. A7.8

Solution: The shear flow was assumed given by Sredt's formula (Ref. Chapter A-5)
q=.1.=
2A Zbe

Thus
dU = P (14)

do In words, ItThe ~ 9S chang;e ~ ~ energy ~ respect to deflection is equal t9 the associated load" . ---- Sq. (l4) and the above quotation are state_ ments of the Theorem of Virtual t.or-k. The reader may find this theorem sta~quite differently in the literature on rigid body mechanics but should be able to satisfy himSelf that the expressions are nevertheless compatible. A useful restat~ment of the above theorem is obtained by rewriting eq. (14) as
dU - Pda

constant around the ?erlmeter of any section. Then

U

1: 2Gt
T'L(b+c) 4Gb lie it

q"S

THE TOTAL ELASTIC STRAIN ~GY OF A STRUCTURE The strain energy by its ~efinition is always a positive quantity. I~ also is a scalar quantity (one having TSgllltude but not direction) and hence the total energy of a composite structure, having a variety of elements under various loadings, is readily found as a simple sum.

=0

It is next argued that if the Change in displacement dO 1s sufficiently small the load P re~Alns senSibly constant and hence
dU - d(PO)

=0
- - - - - - - (15)

diU-po)

=0

J'{"d:< Ii q",j.xc.y ZAG 2Gt

+

- - - - - - - - - - - (13)

The integral symbols and comacn use of "XiI an an index of integration should not be taken too 11terally. It is probably best to read these terms as "the sum of so and so over the str~c~ure" ra)her than "the integral of It, for quite often the te~ are formed as simple sums without resort to the 8clculus. ~he 8alculus is only ~sed as an aid in Some ~pp1ica tions. It is seldom that all ~he terms o~ eq. (13) need Je employed :n a calculation. ~any of the loacings, i~ actually ~resent, ~y be of a localized or o~ a secondary nat~e and their energy contribution ~ay be neglected.

The quantity U - PO is called the total potent:'..al or' the system and eq, (15), r-esemb Ltng as it 10es the ~th~~atical condition for the ~inimum value of a function, is said to be a statement of the Theorem of Minimum Potential. From the foregoing i~is clear that the Theorem of Mini~um Potential 1s a ~estatemen~ Of the 1':leorem or 'iirtual .vor-x . In structural analysiS the ~ost Dnportant ~ses of these theorems are T~de in problems concerning buckling instability ~d other nonlinear1ties. No a?plications will be ~de at this poInt .
A7.5 The Theorem of Complementary Energy and casngnaoo'e Theorem.

Again in t~e case of an elastic bOdy, exaoinatior. of the arsa above the load-deformation cu:ve shows that increments in this area (called

(',-

,.-

A7.8

DEFLECTIONS

OF

STRUCTURES

the complementary energy,U), are related to the load and deformation by (see Fig. A7.G(b).

. "P

TI;/' III I/U:, u'

L'

r Ou-

Force {Ibs .") Moment (in. Ibs , ) Torque (in. Ibs.) Pressure (los/in"') Shear Flow (lbs/in)
~force~

Translation (inches) Rotation (radians; Rotation (r~dians)

vojuae (Ln ")

Area (in"')

, , ,

dU ::6dP

iiP

dU'

=6

- - - - - - - - - - - - (16)

This is the Theorem or Complementary STIergy. Now tor the linearly elastic body a very tmportant theorem follows since (Fig. A7.9c)
dU = dU

Any generalizations of t~e ~eanings of and ~deflection~ are possible only so long as the units ~re such that their product yields the units of strain errergy (in. IbS). Once again !or emphasis it is repeated that, while the complementary ~ature ot eqs. (14) and (17) are clearly eVident, the use of eq. (17) (Castigliano's T~eorem) is restricted to linearlY elastic structures. A brief exa~ple will serve tor illustration at the pOSSible pitfalls. The strain energy stored in an initiallY straight uniform column under an axial load P when deflected into a hal! sine wave is

so that
. rt x Y :: Y oSlDT; M::PY
y

, , P~U' ~ a. u
;..,
Fig. A7.9c

Consider Flexural Energy Only

U .fM""" P"Yd L
0 .,

-If(LY' di.)
2

2EI

4EI

dx ; Y3 n: 4L

dU = 0
dP

- - - - - - - - - - - - -

(17)

Fig. A7.10

-~

P~6L~

In worne ,
~The rate of change ot strain ener~J with respect to lOad is equal to the ass0"CTat'8d-detlection".-- - - Eq. (17) and the above quotation are statements of Castigl1ano's Theorem. For a bOdY under the simUltaneous action of several loads the theorem is writte~ so as to apply individually to each load and its associated detlection, thus:

where 0 is the end shortenir.g due to bowing. Because the deflections grow rapidly as P approaches the critical (buckling) load the problem is non-linear. The details of the calculation of U are given with Fig. A7.l0. Now according to the Theorem of ViTtual
~ (eq. 14)

,j3-

dU _ P

but Theretore
nilEI = P

=0

(17a)

P"L"

The partial derl~tive sign in eq. (17a) indicates that the increment in strain energy is due to a small change in the particular load P , all other loads held constant. i Note that by "load~ and "detlection" may be meant: Load ASsOciated Detlection

or
_ n 2 EI P -

r::r

(Euler formula for uniform co nem ) . correct r-esuj.t , Application at Castigliano's Theorem. eq. (17), leads to the erroneous result:
dP

dU = 0
2P6L~

dU dP
The proof of the theorem for the case of multiple ioads is

nAEI

generally formulated more rigorously. appeal to a simple diagram such as Fig. A7. 9c being less effective. se e, for example, "Theory of ElaSticitY" by S. Timoshenko.

See Art. A17.6, Chapter A-I7 for detailed derivation of this equation.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A7.7

(incorrect) Moral: Do not use Castigliano's theorem for non-linear problems. Fortunately the above restriction upon ~he use of Castigliano's theorem is not a very severe one, the majority of avery day structural problems being linear. Castigliano's theorem Is quite useful in performing deflection calculations and a variety of applications will be made in the following section~.
A7.6 Calculations of Structural Deflections by Use of Ca8tigliano's Theorem.

Then

6p

= dll dp = 43.26

x 10- P tt.

Note that in this problem the only use made of the calculus was in the differentiation. A Simple sum was used to form U, as per eq , (2). xample Problem 9 Derive Bredt's to~ula for the rate at twist of a hollow, closed, thin-walled tUbe:
ds

As the examples of Art. A7.3 have illustrated,the strain energy of a structure can be expressed as a function of the external loadings prOVided the internal load or stress distribution is calculable. Having the strain energy so expressed the deflections at points ot load application can be determined with the aid of eq. (17), Castigliano's Theorem. In the examples to follow the deflections ot a variety of statically determinate structures are computed. Methods of handling redundant structures are considered in subsequent articles. Example Problem 8 F1nd the vertical deflection at the point 01' load application of the crane of F1g. A7.ll. Cross sectional areas are given on each member. The stranded cables have effective ~odul1i of 13.5 x 10" pSi. E = 29,000,000 for other members.

Ref. chap. A6.

The strain energy is stored in shear according to 1 'd>:d

U=zJ

rJ Y

This is the only energy stored, secondary etfects neglected. OVer a given cross section q 1s a constant and is given by Bredt's equation
q

= ~ , where A is the enclosed area ot the tube (Ret. Chap. A.5).

Solution: Since the twist per unit length was desired the strain energy per unit length only was written. Thus, assuming y in the axial direction, no integration was made with respect to y. The integration in the remaining direction was to be carried out around the perimeter ot the tube and so the index was changed tram "x" to the ~ore ap propr-tat;e "e ", Hence (compare with Example Pr'ob , 7)

~'1I5'\.l\'~C.
A'

'I
60'

'r 1-i-1.~"~'~:
E

.-$:'
p

~~
0

u
Fig. A7,1l

= J:..

I_

2GJ'

q ads t

=~

~ cis 8AG~ t

~30'+- 60' ~

Th~ symbol ~ means the integration is carried out around the tube perimeter. Therefore

Solution: The strain energy considered here was that due to axial lo~ding in each of the four members. The load distribution was obtained fram statics and the ener 5 J calculation was made in tabular form a~ follows:
TABLE A7.2
MEMBER
I

USE OF FICTITIOUS LOADS

S LaS

AE X

iO"'5
LBS

s.aL x
AE

10"

OA
AS

-1. 50 P

AC
DC

2.50 P 1.58 P -2. 12 P

40.0 50.0
63.0
4,5

136 11.8 11.8 13

0.66 P Ii
26.48 P
13.33 P 2.79 P I;:; - 43.26 P
Ii

In the follOWing example the desired detlec tion is at the free end of the bar where no load is applied. A fictitious load will be added for purposes of the calculation. The rate of change of strain ener~J with respect to this fictitious load will be found atter which the load will be set equal to zero. This technique gives the desi~ed result in as much as the deflection 15 equal to the rate f change at stratn energy with respect to the load and such a rate exists even though the load itself be zero. Example Problem 10 Compute the axial motion at the free end at the tapered bar of Fig. A7.12.

A7.8

DEFLECTIONS

OF

STRUCTURES
eval~ate.

constant J

q -

-flA'"
~

erally are easier to

Ao u- 2L x) q A.o
'" 40.,/in. = 4 in. 4 L = 40 in. E '" 10 x 10e psi

j
, Flo
x

:Q:a'nple ?roblem 11 ~ind the deflection at ,oint 8 of the beam of rig. A7.:3.
p

Fig. A7.12

EI constant
e

Solution: After addition of the ~istitious end load R the axial load frem statics was found to Je
S(x) = p.
-I-

--- -:-L
L : L

A,

3"

Fig. A7.13

Solution: A fictitious load 21. NaS added at point B and the bending moment dia;ram was 1rawn in VNO parts.
0<X<:L/3
Mp

q (L - x)

Hence, since loadings other than tensile are of a secondary nature

U

t
0

S'dx AE

1
2

Po'
2A,E

I~

[Po

IL - x)]'dx XL) AoE (1 - 2

Fig. A7.13a

+ q

0<Y<L/3 0< ;;<L/ 3

p.q JL 1L - x ) dx dx (1 _ -"-)+ A,E 1( _ X-)

2L 0-2L

~
2A,E

IL 'L- x)'dx
( 1 _ -,,-) , 2L

~en

neglecting the energy at shear as being smallU

Betore evaluating the integrals it was observed that the steps to follow, in which U was to be differentiated with respect to P, and the subsequent setting of R = 0 would drop out both the first and last terms. Hence only the second term was evaluated.

1 + 2EI

u =
Then
6

II

X + 2L

A,E

Rq (1 _

L~

r/'
,

2~I
~

L/7

V[(2P +

2)

-I-

.z. 2E!

= dU = 2L'q dP, AoE

11 _ L~ 2)

Differentiation under the integral sign With respect to ?~ gave

r
c

P + P, (L + 3YIJ [gIZL - 3Y) 9

,
dy

,
-I-

3 [IP

2P1.) (L- 30)J do. 9

DIFFERENTIATION UNDER THE INTEGRAL SIGN

0B=~~~

important labor savings may be had in the calculation of deflections "by Castigllano's theorem. In the strain energy integrals arising in this class of problems, the load Pi' with respec to which the deflection is to be round, acts as an independent parameter in the integral. Provided certain requirements for continuity of the functions are met - and they invaria~ly are in these problems - the differentiation with respect to P may be carried out before the integration is made. The resulting integrals genAn

ir J
3yl

L/ 3 [(ZP+
e

P'IL " 3 y)J(L

9

+ ~

3y) dy

- For beams of usual high length - to - depth ratios the shear strain energy is small compared to the energy 01 flexure. Neglecting the shear energy is eqctvarent to neglecting the shear deflection contribution. (see p, A7.14)

. ._

.... 7 _

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

T~e

A7.9

fictitious load P L , havin~served its pur;058, was set equal to zero Jefore complet~ng the wark.

The total loadings

'Nere
Plan View

T=T1.+T:a= TaCOS (1 - cos G)

/3 1 (2L - 3y) (L + 3y) dy + E:I :1),

Q ...

PR

rL

T:!1.us
Fig. A7.14b

U = 2EI
+

_ PR)lI s in" Q Rd g

2;J )n/ 02[T

oCOS

Q +

PR(l - cos

Q)J.:lI

Rd g

Example Problem 12

Fig. A7.14 shows a cantilever round rod of diameter 0 formed in a quarter circle and acted upon by a torque To. Find the vertical movement of the free end. Solution: Fig. A7.14a shows the vector resolution of
applied torque To on

(Note use of ~RdQ~ for the length of a differential beam element instead of "dx"). Differentiating under the integral sign
Fig. A7.14

TI = + -

au

(T, - PR)R"

EI

IT /2 o sin.:ll Q \

dQ

R"

GJ

beam elements. TL (Q) = ToGas Q and the moment nl,{Q) = To sin 'g. APplication of a fictitious vertical load'? (dOwn) at the point of desired 1eflection gave the loadings shown in Fig. A7.14b.

ene

Plan View

Putting P J the fictitious load equal to zero and integrating gave

OVERT = IP

au

= ToR
4

li

4 - IT IT ) ( ~-EI

E Since J = 21 and G ~ 12.6 the deflection was negative (up).

Fig. A7.14a

A7.7 Calculation of Structural Deflections by the Method of Dummy-Unit Loads (Method of Virtual Loads).

The strict application of the calculus to Castlg11ano Js theorem as in Art. A7.6 J leads to a number of clli~bersame techniques ill-suitee to the solution of large complex structures. A more flexiJle approach, readily adapted to improved ~book keeping~ techniques is the ~ 2f ~ Unit Loads developed lr.dependently JY J. C. ~arNell (lB64) and O. Z. Mohr (1674). --- That the eQuations for the Method of Dummy-Unit Loads may be derived in a number of 'Nays is attested to by the great variety of names applied to ~~is nethod in the literature ~. Presented below are t~o deri'lations of the equations stemmir.g ~rom different viewpOints. One derivation obtains the equations by a reinterpretation of the S~bols of Castigliano's theorem - essentially an a?peal to the concepts of s~~ain er.er~J. The other derivation uses the ~rinciples of rigid ~ody mechanics. Based as they are upon a commoe set of consistent assumptions, all the ~ethods must, of course, yield the sa~e result. Derivation of Squations for the Method of Dummy-Unit Loads (Virtual Loads) I - From Cast1glianc's TO_eoram Beginning with the gener-a.I expression for strain energy, aq. (13) II - From the ?r1nciple of Virtual work
~nen a system J:' ~orces whose ~esultant is zero (a system in eQ~1Iibr1um) is displaced a

~ Variously called the Maxwell-Mohr Method, Method of Virtual Velocities, Method of Virtual Work, Method of Auxilliary

Loads, Dummy Loads, Method of Work, etc.

A7.10 I Cont'd.
U

DEFLECTIONS
S "dX + ~ AE 2

= ~

J!'l"u :::1

J T~f
+

OF STRUCTURES IT Cont'd.

- - -

etc.

differentiate under the integral sign with respect to Pi to obtain

6

IS~dX
~

small ~ount withcut dis~urbing the equl:1br1~ balar.ce, the work done is zero - obv1ously, since zero res41tant torse ~ovlng through a distance does zero work. Consider now the set of equilibrium force~ applied to the truss of Fig. A7.15(b). The set may be divided into ~~o parts: the "external system", conSisting of the ~~it load applied at the point whose deflection is ~eslred ane the two reactions fiXing the line of referer.ce. ~d the "internal system" consisting of the axial loads acting on the truss members to produce equilibrium. T~ese latter are denoted by the synbol ~u". T~is set o~ forceS 15 considered small enough so as not to.affect t~e actual behavior of the structure during subsequent application of a real set of ~Jor loads. This unit load set is present solely fer ~thematical reasons and is called a "virtual loading" or "dummy loading". Assume now that t~e struct~e undergoes a deformation due to application of a set of real loadS, the virtual leads ftgoing along for t~ ride n Each ~ember of the structure suffers a deformation denoted by ~ ~ The Virtual loading system, bei~g in equl11b~1~ (zero reSUltant; by hypo'thes t s ,10es work (I'virtual work") equal to zero. Or. conSidering the s~bdivision of the virtual loading system, t~e work done by the external Virtual load must equal that absorbed by the internal vrr-tuaj loads. The work done by the external virtual forces is equal to one pound times the deflection at joint C, the reactions Rl. and Ra not moving. !~at is

J M~dX 1
,1

- - - - - - - etc.

Consider the symbolS

~S

<lP

'"'JP'1'P'""
i i

~M

~T

- - --

- etc.

Each of these 1s the "rate of change of so -and-so with respect to Pi" or "how much soand-so changes when Pi changes a unit a~ount" OR EQUALLY, "the so-and-so loadln~ for a unit load PI' Thus, to compute these partial derivative terms one need only compute the internal loadings due to a unit load (the virtual load) applied at the point of desired deflection. For example, the term elM/elPt, could be computed in either ot the ~HO ways shown in Fig. A7.15a.

RATE METHOD

UNIT METHOD

Fig. A7.15a

Exter~al Virt~l Work = l' x 6 c The internal Virtual Nark is the sum over the l. = x x structure of the products of th~ ~ember vir~ual loads u by the member distortions~. That is, Likewise, d s~ Pc, where Pc is a load (real or fictitious) applied at jOint c ot Fig. A7.15b, Internal Virtual ~ork = ~ u.~ is given by the loadings for the unit load applied as shown. Then equatin~ these works, In practice the use at the WIlt load is 1 x 0c = Z ud . most convenient. Using the notation If the deformations 6 are the result of _ oM elastic strains due to real member loads S then ,m =~, 1 Do = 8L/AE for each member and one has 6 = Z u 8L v a q = aq c AE Pi' - ~ The ar~~ent given above may be extended for the unit loadings, the deflection equation QUickly to include the internal virtual work of becomes virtual bending moments (TIl), t or s i on loads (t), 6 = Sud:< + IlmdX + ~ shear loads (v), and shear flows (q) doing work 1 AE EI GJ during defo~tions due to real moments (M), (T), shear loads {Vr;-ind shear flows + ~ + q qGfdY - - - - - - - - - (IS) torques (q). The general expression becomes
oM

I1 = Pl.X

m = dummy loading

(=~~J

:V

If

-1.0

-1.0

f Sud=< ~ J Ttc.x +J AG +J;q q

6
1

AE

+ ;

EI

...

GJ

VvdX

dXdy Gt

- - - - - (18

Fig'. A7.1Sb

"un loads due to a unit (virtual) load.

~ Note that the deformations are not restricted to those due to ~lastic strains only. They may be the result of elastic

or tnetesuc strains. temperature strains or misalign_ ment corrections.

.
ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

'.?:'"B.'P!

A7 11

In applying eq. (18) the labor of a =efleccalculation divides conveniently into several steps: 1. Calculation of the real (actual) load distribution (S,M,T, etc.)
t~on

11. Calculation of the unit (virtual) load

distribution (u,m,t, etc.) due to a unit (virtual) load applied at the potnt of desired deflection and reacted at the reference polnt(s). 111. Calculation ot flex!bl11tles,
-A~ , "'1

111
"U1,"

loads

"U.'

loads

Fig. A7.16a

Fig. A7.1Bb

'.

etc.

......
AS

TABLJ: AT.3

Lbo

iv. Summation and/or Integration. Here again note the general nature of the terms "load" and "ceriece i on". (See p. A7.6) The following examples apply the method of dummy-unit loads to the determination of absolute and relative deflections, both rotation and translatioo. Example Problem 13 The pin-jointed truss of Fig. A7.l6 is acted upon by the external loading system shown. The ~ember loads are given on the fi~~re. Member properties are given ~n Table A7.3. Find the vertical movement of joint G and the horizontal ~ovement of jOint H.
A 10
BC

CD

,.,
aB

..

BQ

.. " .. .. ..

AIls 10 Ibo
4.1 3.074 3.0T4 5.385 5.385

AI

toslO
8.2'1'0 g.T5' g. '1'1 5.5111 5.1111 1.821 4.1128 14.3.

""Ib

10,500
2,~0

.,
a
c

II.

1.'

.....,. """""
a.. " a a a a
22.01 12.01
_21.~

~s 10 ~JCIO.

2,250
_6,~

... "

-5,2llO .0. T5 1

_21.31

3."
lO.a

_.000
.3, 'liO

a
53."
111.80

50 50

_8, 'I'1lO 1-1.25

3.4'
5.3G 3.0'1'4 3.074 3.0'1'4

00
SF

co
DB

.. .. ..
50

1."

11.320 13.012 13.012 13.012

3,000
_l,OOO

3.000

1000"

1000*

1000'*

a a a

a
a

fj~gl~(?'bgl~~~ l
~"
0

sods

2250

Ie 2250

10

Answer-s :

oG

:::

0.286"

VER

-;o:"g40"
(l
N

'-5250 .5250

\~30" t30"'+ 30"1

2000t
Fig. A7.16

O~OR

::: -.0587" (the negative Sign means

2000t

the joint moves to the LEFT since the unit load was drawn to the RIGHT in Fig. A7.16b). Example Problem 14 For the truss of Fig. A7.16 rlnd the following relative displacements of joints: c) the movement of joint c in the direction of a diagonal line joining c and F. d) the movement of joint G relative to a line joining points F and H.
~elatlve deflections are determined by applying unit (Virtual) loads at the points where the deflections are desired and by support ing such ~nlt load systems at the reference poir.ts of the motion. ThUS, ror solution to ,art (c) a unit load system was applied as shown in Fig. A7.16c and for the solution of

Solution: only the energy of axial loadings in the members was considered. Unit (virtual) loads were applied successively at jOints G and H as shown in Figs. A7.15a and A7.16b. All Sand u loadings were entered in Table A7.3 and the

calculat~on ~or 0

= J Sudx AE
6

was carried out JY

forming the sum ot SuL/A E terms fer the ~embers

, e. of the truss, ..
UL = ~" '" SAE _

A7 12

DEfLECTIONS OF

STRUCTURES

part (d) the system of l~it loads of r1g. A7.l6d .cas used. Table A7.4 completes tne solution, the ~eal loads and member :leXibilities

(A~) being the s~e as for ex~ple problem 13.

o I~,,;> t" iJ; '_. ~Jl f il&

>;)

-.6

roO

-.375

-.375

':c:

, '..,i'

"

O..n

-r

:0

It:I

Rotations, both absolute anQ "elative a~~ unit (virt~al) couples to the member or port~cr. of struc~ure ~hose rotation is desired. The unit ccuple is ~esis~ ed by reactior.s placed en t~e line of reference for the rotation. Thus Figs. A7.16e ~nd A7.l6f show the unit (virtual) loadings fer par t.s (e) and (f) respectively. T~ble A7.5 completes the calculation, the real loads and ~ember ~lexib~llties (L/~) be Ing the same as fer examp.Le
dete~ined by a~ply1ng

6' 0

,. FOG

0" H

Ac

problem 13.

R=U'

.5

1jjo

.5

"us" loads
Fig. A7.16c

"u ..." loads

Fig. A7.16 d

"~~?@?1
R=l!40'll' -.025

1/50*,

/\/0
1/50# R=1/40'"

-. 02~ G'l/50#

TABLE A'l.4

"u e" loads Fig. A7.16e

/G 1/50* "u.!l;" loads

Fig. A7.161

R=1/40#

MEM. AE

...h. x las

In/lb {See Table A7.3} 8.210 9.159 9.7:19

~.591

S,b.
10,500
2, aso

'.
-. ,
0 0 0

'.
- 0

~x1a" Su ...L x 10"

1".

A'r1.llcb.

AS

Be

-.315 .31:1
0 0

-13.11
0 0

-8.23
-8.23 TABLE A7.5 0 0

CD E"
FG

1,%50 5,250 -5,2:la

~xlO.s ~X10S

5.591 S,621 4.926 14.368 .9.320 13. 012 13. on 13.012

-.'
n
0

17.61

..k.
blEM.

:I:

reWIb

GB BE

0
-8,750

c o

c o
71.84
0

AE

(See Table

S,b
10,500 a. 450 2,250 -5,2:10 5,2:10
0

a&lnch-J.

Il....lncb.-l.

AE

A'l.3)

eo
DO

'.000
-3,~0

1.0

.e"
.'20

....
0

AE
Be

6.a70 9. 7S~ 9.
7~9

. 0.5 . D25
.020

1."
.55
.55 . 73 . 73
0 0 0

_. D.5
0 0 0 0 0

.0.55
0 0 0 0

BF CG DB

2.000
1.000

.. ..
0 0

-21.8
13. a 0 13. a

CD EF FG GB BE
BG
DO

.50
0

20.82 .10.41

~.591

.025 .025 i
0 0 0

5.591

2.000

_.50

e.eu
4.926 14. see 9.320 13.012 13. 012 13.012

= 65.87

Z =19.38

-8, 750

0
1. 08 0 0 0 0

'.000
3.750 2,000 -1,000 2,000

.OU
0

".015
0 0 0

.52
0

BF CG

,
I

0 0 0

Therefore the movement of joint c towards joint F was 0 = .06587 inches and the motion of joint G relative to a line between F and H was 6 = -.0194 inches, the negatlve sign indicating an upward movement. Example Problem 15 For the truss or Fig. A7.l6 determine a) the absolute rotation of member DG t) the rotation of member 80 relative to member CO.

, DB

Z = 4.73

.z =0.53

~herefare

DO was 80

DG

= .00473

the absolute rotation of member radla~s and the rotation of

relat~ve

to CG was Q _ CG 3G

= .00053

radians

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.13

a.ample fToblem IS tine the vertical eeflect:on of point C for the cantilever jea~ of ?lg. A7.17 carrying a concentrated load P at its end. Al~o find Slope of elastic ~urve at 8. Solution:

points C, DJ ~, the attachment ~oints of an aileron or a flap. The wing beam deflection bends the aileron or flap structure by applying a lcad at 0 thru aileron s~pporting bracket. To know this force the deflection of the wing beam at D relative to li~e CE ~ust be known. Solution: Origin at 8: Ii

f
with origin at 9
M = - Px

Mmdx

-W-

= 30

= 15

x"

(Fig. A7.17)

For virtual

load1~g
I!l.

(?ig. A7.17a)

= 0, far x cb
-1 (x - b), for x)b
Px(-x + b
Fig. A7.18

Hence

!1mdx

=P

Idx

= (?x'"
=

- Pbx Idx

EI ) b

rL

(x

- bxidx

1.[ x'
EI 3

bX'J L
2 b
~

I'" -----+I ,....-c Fig. A7.l8a

a c
.'

,r ''1
0

Rd'

P [, 3L'b b' 3EI L - -2- + "2

If b
) A

OJ x

<

20

u = -.5 (x-20) =-.5x+10,whenx=20to50 m -.5(x-20)+1(x-50)=.5x-40 J when x = 50 to 80

= zero,
,

then 6

= PL
C

3/3EI

l==
1'
)'

4p
a
B

Fig. A7.17

J
=F:r 1
.... .1.

l1m dx

E1

It
c r.,l

Fig. A7.17a

k
20

50

15 x~ (-.5X + 10)dX + ;;'1 1 " (.5x-40jdx

h
50

80

15 x"

un1 t.

Fig. A7.17b

ooupl_

150x,l50 + l. [7.5 x' _ 600 x' l80 3 J20 S1 4 3 jso

e
For
vi~tua1

c,

=JMmdx

2;1

= ISO; [_ 11.72

+ 6.25 + .300 -

.400 +

loading See Fig. A7.11b

a

76.8 - 102.4 - 11. 72 + 25.0]

= 0,

x cb ,

- I, x>b
?x

17.9 X lOll
EI

Hence !1mdX

(-?xi (-Ii ex

ex
o

J
If b

Mmdx_PfL
SI b

xcx

2EI

= 0,

= 2EI

0' ,

Therefore deflection of point 0 relative to line joining ~E is down because result comes out negative and therefore opposite to directicn of Virtual load. Example Problem 18
1''''

lb/rt
L

Problem 17 Tor tne um r orm.Ly loaded cantilever beam of Fig. A7.18, f~nd the deflection of point 0 relat~ve to the line joining points C and ~ on the elast~c curve of the beam. This is representative of a yr~ctlcal ?rOblem in aBronauttcs, in that A3 :n.1.~ht represent a rear wing beam and
a.~ple

tn
~L-

'; II i i i
2

x~

x-'1
wx e

m diag'l

'2
(b)

",to

'2

'oiL'

j:~~lr,1
R=l-1i'
~ '\ R=l/L unit

1 R=l/L

(a)

( c)

couple

Fig. A7.19

Fig. A7.20

A7.14

DEFLECTIONS

Ol"

STRUCTURES

Find the horizontal deflection ot point C for the frame and loading of Fig. A7.l9. Also angular deflection af C with respect to line CD. Solution: Fig. b shows the static moment curve for the given loading and Fig. c the moment diagram for the virtual loading of a unit horizontal load applied at C and resisted at D.
_JMmdx - - ,. EI
M

dy = J ex

and

~ = A-'

V
~s

J where

cross sectional

area of beam at section and ~s ~cdulus of rigidity: and assuming that the shearing stress
V A is

uniform over the

1 x 0
~s

cr~ss-sectlon.

Therefore

= Vvdx ----A-.

~hen

the total detlec-

tion for the shear slipS of all elements of the beam equals

= wLx
2

wx 2

:n = h

hence

total = L VvdX , AEs

_______ (a)

where V 15 the shear at any section due ~o g~ven loads. v = shear at any sect10n due to unit hypothetical load at 'tae point where tne aer iec-, EI c tion is wanted and acting in the desired direction of the deflection. The reactions to the hypo1 whV' thetical unit load fix the line of reference for 12 EI the deflection. A is the cross-sectional area and Es the To tind angular deflection at C apply a unit modulus of rigidity. Equation (a) is slightly imaginary couple at C with reactions at C and D. in error as the Shear:ng stress is not ~~i~orn F.ig. A7.20 shows the virtual m diagram. over the cross-section. e.g. being ~raballc ~cr a rectan~~lar section. However. ~he average shearing stress gives close results. So ~ 1. WX') ~ dx For a unito~ load of w per unit le~gth, the EI EI 2 2L J;I, , L center deflection on a s~ply supported beam 1s: -

L(WLx WX') 2 - T h d x =r,(WLx'_'/IX') 1!. JL ~ 4 6. EI

=JL

= fL (WLx _
= ,;\
1

= ijL

(W~' - w:: ) dx

[~' - w:~ J

center =

wL;'

2f~~=2 ~ L AE s
,e

= 24 Ef
Linear Deflection of Beams Due to Shear by Virtual Work.

r
J

(wL 2

wx)

1
~

ax

AE s

wL'

SAEs
For bending deflection for a simply supported beam uniformly loaded the center deflection 1s

Generally speakingJ shear deflections in beams are small compared to those due to bending except tor comparatively short beams and theretore are usually neglected in deflection calculations. A close approximation is sometimes made by USing a modUlus of elasticity slightly less than that for bending and using the bending deflection equations. The expression for shear deflection of a beam is derived tram the same reasoning as in previous derivations. The virtual work equation tor the hypothetical unit load system for a shear detrusion dy (Fig. A7.2l) conSidering only dx elastic is 1 x O=vdy where v is shear on section due to unit hypothetical load at paint OJ and dy Is the shear detruslon at the element dx due to any given load system or any other cause.

Henae
wL'

8AE s
5wL"

24

(Er

384EI.

using is = .4:E; I r = radiUS at gyration.

For

beama and cr~els r is apprOXimately ~ d

and tor rectangular sections r = v 12

(1

= depth)

In aircratt structures a ratiO of d 1s

~--

-f-5-' ----ft T,.

Fig. A7.21

"'~dLdy

seldom. greater than

6

12

Sometimes "G", instead.

Seep. A7.4

Thus the Shearing deflection in percent of the bending deflection equals 4.1~ for a ~ ratiO ot for I-beam sections and 1.4 percent for rectangular sections.

iz

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.15

Example Problem 19

t ;
10011"

t '+ lO'4
'O
vA., -

501'

Unit. Load.

IF==='

torsior~l ~oment at any section due to a unit virtual couple acting at section where angle of ~N1st is desired and acting in the plane of the desired deflection. (inch Ibs/inch lb)

Example Problem

Fig. A7.22

Find the vertical deflection of free end A due to shear deformation for beam of Fig. A7.22 assuming shearing stress uniform over crosssection, and AE constant. s
Vvdx AE s V = 100' for x = 0 to 10 10 to 20 V 150 for x v = 1 for x = a to 20

"' _ I

Example Problem 20 Fig. A7.Z3 shows a cantilever landing gear strut-axle unit ABC lying in XY plane. A load of 1000* is applied to axle at point A normal to XY plane. Find the deflection of point A no~l to XY plane. Assume strut and axle are tubular and of constant section. Solution: The loading shown causes both bending and twisting of the strut axle unit. First find bending and torsional moments on axle and strut due to 1000* load.

hence

1 AEs

to ...L 0 1001 dx + AE s

fO
10

1501 dx 20 10 2500 AE s

Y
0 3

10 1 [lOOX] + = IE: s 0

~.[150X]

I /

Method_of Virtual Work Applied to Torsion of Cylindrical Bars.

The angle of twist of a circular shaft due to a torsional moment may be found by similar reasoning as used in previous articles for finding deflection due to bending or shear forces. The resulting expressions are: -

1/ \\ i s: ~~3'
B

36"

I
B
Fig. A7.24

,.

Fig. A7.23

Ttdx
E J

o =

JTEts~

Member AS
- - - - - - - - - - - - (A)

Member BC
(B)

o to 3) M = 1000 x, (for x T = 0 = 3000 sin 20 + 1000 x, (tor x = a to 36)

Tse ; 3000 cos 20 0 constant between 8 and C.

In equation (A), for translation deflections, T twisting moment at any section due to applied vNisting forces. t = torsional ~oment at any Section due to a virtual unit 1 lb. force applied at the point where deflection is wanted and applied in the direction or the desired displacement. (in lbs/Ib) Es Shearing modulus of elasticity for the material. (also nGn) J polar moment of inertia of the circular crcss-eect ron. In equation (B), for rotational deflections, Q = angle in twist at any section due to the applied vNisting moments in planes per?endicular to the shaft axis. Angle in radi~~s.

Now apply a unit 1* rorc~ at A normal to xy plane as shown in F!g. A7.24 and find bending and torsional moments due to this 1* force. Member A8
m

= 1=0 =3

= x,

(r or x

=o

to 3)

Member Be
m

sin 20'

+ 1-

x (for x

=0

to 36)

t 3 cos 20' constant bevNeen B and C.

A7.16

DEFLECTIONS OF STRUCTURES

Subt.
6

I'!mdx -+-

81

JTtdx
E J

The shear :lows shown on ~he (nearly) horizontal edges of the web panels are average values. ?i6' A7.27 ~s an explOded vi~N of the beam shaNing the unit (Virtual) loads.

tr

J:
+

1 1000 X . X dx Er

f:
0

6 K1000 x 1026)

(x

1.026)J dx

-+-

E-J

f36 (2820)

(2.82) dx
Virtual loading. Fig. A7. 27

EI
_1_ (286200)

(16,92500)

Since both axial loads and shear flo'~ were considered, ~he form of deflection eq~tion used was
q q dxdy

Gt
I~tegrations in the ~langes were made assuming linear load variations. Such ~~ integration carried out over a uniform :lange of length L whose real load varies from 8 1 to 5 j and WhOS8 virtual lead varies :rcm ~i ~o Uj yields

EsJ

Note: A practical landing gear strut would involve a tapered or reinforced section involving a variable I and J and the integration would have to be done graphically or numerically.
Example Problem 21

For the trnn-weo aluminum beam of Fig. A7.25 determine the deflection at point Gunder the loading shown. Stringer section areas are given on the figure.
A ... A=.15 B A=.15 C ,..A-.15 D

J
Q

L Sudx =

AE

r I L
20"

t=.0

32

1 l

A=.08

t=.032 1

A=.08 A-.05"

t=.032

--..L

T IS"

The integrations in the trapezoidal sheet panels were made using the shear flows en the (nearly) horizontal sides as average values, assumed constant over the panel. with this
= qAV Gt where S 15 the panel area. The calculation was completad in Table

~ 20" ----..+- -20"

Fig. A7.2S

rF I G

-+-

20"

-4

H+ P.j80"*

Simplification

55

qAV

q,w :ixdy

Solution: It was assumed that the webs did not buckle and carried shear only. Fig. A7.26 is an exploded view of the beam Showing the internal real loads carried as determined by statics:
3. OP_

A7.6.

c==.:===;::::~~~====
: :.:.,. p, ,
'.00 P ,_L,H P, 1.00,_:.011 .un I .,.,., ,

~ID -*~i

2. 184P

3.0P..."..

.04

2:"I84P Fig. A7.26

;~~ .O!~~~~. W I
1.201p

1. 201P

".'"

.'.'M "'-'-:10""'.01'

'u.

".0 P

...

...:,:" ,-',W")
~.r'._.

",,"'''1 i
.,.\<
I

i.,'. ",.i,~......5. ".,

I : 171.0 P."

a.c_r......

<>U'.

""I
: .1

The equations of statics for tapered beam webs are derived in Art. AIS.18, Ch. A-IS.

CD-<l-HI._P'

ANALYSIS AND DESIGN OF

FLIGHT VEH

LE

TR

RE

A7.17

= (271.0
2

13.4) P x 10- 11
-j

.147 (1800) 10-~ in = .265 in.

A7.8 Deflections Due to Thermal strains.

II5'~
L

~hr;l~Tdx
0

As noted in the "virtual work derivation" of the dummy-unit load deflection equations, the real internal strai~s of the structure may be due to any cause including ther.nal effects. Hence, ~rovided the temperature distribution and thermal properties of a structure are known, the dummy-unit load method ?rovides a ready means for computing thermal deflections. Example Problem 22 Find the axial movement at the free end of a uniform bar due to heat application to the fixed end, re~ulting in the steady state temperature distribution shown in Fig. A7.28. Assume material properti9s are not functions of temperature.
T

;-t I-dx
a
Fig. A7.29
b

Solution:
T~e axial deformation of a differential element of the u?per flange (sUbSer}pt u) was ass~~8d given by = a T ax where a was the material ~hermal coefficient of expansion. The lower flange, having received no heating underNent no expanSion.

Au

temperature above ambient temperatlire

an empirical constant depending upon the~l properties and rate of heat addition.

Inasmuch as a thermal expansion is uniform in all directions no shear strain can occur on a material element. Hence no shear strata occurs in the web. The apparent anomaly here - that web elements appear to undergo shear deformations J = a~ (Fig. A7. 29b) - is explained as follows: The temperature varies linearly over the beam depth. The various horizontal beam "fibers" thus undergo axial deformations which vary linearly also in the manner of Fig. A7. 29b giving the apparent shear deformation. No virtual work is done during this web deformation since no axial virtual stresses are carried in the web.
~ith ~he addition of a unit (virtual) load to the free end, the virtual loadings obtained in the flanges were:

Fig. A7.28

Solution: The thermal coefficient of expansion of the rod material was u. Hence a rod element of length ax experienced a thermal defo~tlon a = a . T ax. Application of a unit load at the bar end gave u = 1. T~eretore
T . dx
T~en ~he

L - x
U

(~

- "L)

ceflection equation was

~J~ o n

a T

23 idealized ~No-flange cantilever beam of ?ig. A7.2Sa undergoes rapid ~eatlng of the upper flange to a ~emperature T, uni:orm spanNise, above that of the lower flange. Deterxlne the resulting displaeemer.t of the free end.
2.X8.!l:p19 Proj19m
~he

SXa.m.ple Pr::;ble:n 24 ~he ~irst step in computing the the~l stresses i~ a closed ring (3 times indeterminate) involves cutting the ring to ~ke i~ statically determinate and r1~d1ng the relative movement of the ~NO cut faces. Fig. A7.3Oa s~cws a unifo~ circular ring whose insi~e surface is heat~d to a temperature T above the outside surface. The temperature is constant ar8und the c~rcum!erer.ce and is assumed

A7.l8

DEFLECTIONS

OF

STRUCTURES

to vary linearly over the depth of the cross section. Find the relative movement of the cut surfaces shown in Fig. A7.30b.

Remarks: In the three elementary examples given above no stresse~ were developed inasmuch as the idealizations yielded statically dete~inate structures which, with no loads applied, can have no stresses. Indeterminate str~ctures are treated in Chapter A.a.

Solution:

Flg. A7.30

A7.9 Matrix Methods in Deflection Calculations. Introduction. There is much to recommend the 'lse natr rx methads'~ for the hand Ling of 'the quarrt

of Lty

An element of the beam of length Rd is shawn in Fig. A7.30ce Due to the linear temperature variation an angular change dQ =
~

RaT

d$ occurred in the element. The Change in length of the midline (centroid) ot the section was 4 = R~T d~. Unit redundant loads were applied at the cut surface as shown in Fig. A7.30b giving the following unit loadings around the ring. From. unit redundant couple (X)
~

of data ariSing in the solutions ot stress and deflection calculations ot complex structures: ~he data is presented in a form suitable for use in the routine calculatory procedures of hi~~ speed digital computersj a fleXibility of operation is present which permits the solution of additional related proble~ by a simple expansion of the program; The notation itsel: suggests new and improved methods both of theoretical approach and work division. The ~ethods and notations employed here ar.d later are essentially those presented by Wehle and Lansing@ in adapting the Method of D1..1L1!IlYUnit Loads to matriX notation. Other appropriate references are listed in the bibliography.

=1

(m positive if it tends to open the ring). (= axial load) = a (positive of tensile)

BASIS OF METHOD
Assume the structure to be analyzed has been idealized into a truss-like assembly of ~ods, , bars, tubes and panels (sheets) upon which are acting the external loads applied as concentrated loads Pm or Pn, each with a different numerical

From unit redundant axial (horizontal) load (Y)

my
U

-R (1 - cos

$)
(~)

= cos $

From unit redundant shear (vertical) load

m a
-R sin

$
$
(.)
(b)

ua method is

- sin

The deflection equation by the dummy-unit load

e =
Then

Ju

f ru . dQ

Flg. A7. 31. Idealization into an assembly of bars and panels.

= -

2n R-aT (negative indicating movement to

the right)

Q,)For the reader not famUia.r with the elementary arithmetic rules of matrix operations employed here, a. short appendix has been included. iL. B. Wehle Jr. and Warner LanSing, A Method for Reducing the AnalySis of Complex Redunci:Uit""""StrUCtures to a Rouune-Procedure, Journ. of Aero. Sciences, 19, october

mr.-

4Z

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.l9

subscript. T~us the system of Fig. A7.3la 1s idealized into that of Fig. A7.3l~. With the above idealization an improved scheme may be employed :0 systematize the computation of deflection calculations. The following steps summarize the procedure wh t ch is discussed in detail in succeedi~g sections. I. A set of internal generalized forces, denoted by qt or qj (i, j are different numerical subscripts), is used to describe the internal stress distribution. The q's may represent axial loadS, ~oments, shears, etc. In conjunction 'Nith a set of member flexibility coefficients, Uij, the q's~employed to express the strain energy U. Ulj ~ives,the displacement of point! per unit :orce at point ~ II. EqUilibrium conditions are used to relate the interr~l generalized forces qi, qj to the external applied ~ , Pm or Pn With this relationship the strain energy expreSSion obtained in I, above is then transformed to give U as a function of the P's. III. Gastigliano's Theorem 1s used to compute deflections.

THE STRAIN ENERGY

It is next desired to write the strain energy as a function of the q's. Continuing the illustrative example, write

1 q.j,'X'dx
1.

'-

Observe that each of the integral terms in the above expression is a property of the structural element (variation of El) and of the nature of the associated generalized force (exponent on variable). Introducing the notation

= IL'Y'd Y Ela
c

CHOICE OF GENERALIZED FORCES

Consider for example the problem of writing the strain energy (Of flexure) of the stepped c~~tl1ever beam of Fig. A7.32a, assuming external loads are to be applied as transverse paint loads at A and E. The set of internal generalized forces ot Fig. A7.32b will completely determine the bending ~oment distribUtion in the be~ elements and hence the strain energy. s~t (b) then is a sat1stactory choice of generalized torces. It should be pointed out that sat (b) is not a unique set. Other satiSfactory choiceS (not an exhaustive display) are shown in Figs. A7.32c, d and e. The final selection may be "made for convenience or personal taste. Note that only as ~ny generalized forces are used per element as are required to deter~ine the significant loadings in that element.

=
the strain
ener~J

d y jL.E1.
c

becomes

- - (19)

~
M
=

EI.

A EI,

Equation (19) is an expression for U Which could have been written lmm.ediatelY from physical considerations. Each coefficient aij is the displacement at paint i per unit change in torce at point j. This identity is eaSily seen by applying Castlgliano's Theorem to eq. (19). With this interpretation the first ter.n in eq. (19), representing the strain energy in the outer beam portion, is written by analogy to The remaining three terms, representing the energy stored ~n the lIh~er Jeam segment by qa and q~J are likewise readilY written, With proper account taken tor the cross influence at one :orce upon another (the na a3 qa q~~ term). No'te that
eq , (2) of Art. A7.3 ( U = 1 2: S'L) AE

LaY::-L\~{a)

q,
~q1. q3

{I

I) q.
\
~qg

q,
q1.

(e)

ql)",
q,
(b)

JI

(d)

q1.X

O<x<L1.

qtj

M;Q3+qaY O<y<L.

q. I)

q'f
(e)

=
Fig. A7.32. Some possible choices of generalized forces. -("Relative displacements in the individual member")

= "jl - - - - - - - - (20)

Hence
~

,.

(Maxwell's Reciprocal Theorem)

A7.20

DEFLECTIONS

OF

STRUCTURES

The general form of strain energy expression is (ex,anding sq. (19) by ir.duction) 2U = qlql all
+ qlq~ al~ + - - - - - + ql qN

a(N

, ,
I

.. qaql <l<ill

q~qa

0.a a ... - - - - - - - - -

-I
I , , _I , I , -1

+Q"q1. (l"l + q"q" (l"3 ... - - - _ - _ - - -

+qfoQ1. au

, ,
I
I

, , , ,
,

Fig. A7.32f

qNql aN' +

- - -

- -- - -

+ qN qN (lNN

In matrix notation
q, q.
1

In matrix notation this equation is written (see appendix)

Ql2---a1.~

o.
1

P, P.

1
r
~,

C;3l Cl3 '1------:

, ,

q.
q. (21)

, ,
, , '

,
, ,
qN
, ,

Symbol:cally these relationships are

'NT1~tan

~~-------

,\';'

- - - - - (22)

or, more concisely,

The matrix

[Jim] 1s called the ~1t load

inasmuch as any one
col'~

dlstr1butlon~

at

t h column, gives the values of [GimJ ' say the m

In the matrix [Cl i j ] many (if not most) of

the elements are zero. In the specifiC example, eq. (21) would be 'Nritten

the generalized forces (the Q's) for a unit value of load P~, all other external loads zero.
THE STRAIN ENERGY IN TSR11S OF APPLIED LOADS .

o
o o

q,

q.
q.

If eq. (22) and its transpose are used GO subst rtuts for the q.. "s in eq. (21) one gets

2U

(23)

The problem of computing and tabulating various (lij'S is considered in detail later.
RELATING THE INTERNAL GENERALIZED FORClli TO THE EXTERNAL APPLIED LOADS

For the statically determinate structures considered in this chapter the internal forces are related to the external loads by use of the equations at static equilibrium. A set of linear equations results. ThUS, in the speCific example conSidered, it Pl and Pa are the external loads applied as in Fig. A7.32f, then by statics (reter to Fig. A7.32b)

here i and j are used intaras are ~ and n. Also [~lJ is the transpose of [G1m] I i.e. lntercr~nge of subscripts denotes transposition (see appendiX). It the matrix triple product in eq. (23) is fo~ed and defined as
Chan~eably
notat~on

In the

- - - -

(2?)

then

-' .... ' ._.._m; _\:.

A7.21

2U

i L':rJ1J) 1:1 ,I L J r,
p

- - - - - - - - -

(25)

Sq. (25) expresses the strain energy as a function of the external ~pplied loads. In the ~pecific example being used

1 r La 1L'J 1
0

plication itself yields a summation to complete the calculation. ?rcm this last discussion it is clear that eq , (26a) also may be derived by formulating the Dummy-Unit Load equations (Art. A7.7) in matrix notation. To illustrate the application of the ~trix methods presented thus far, a brief and elementary example 1s worked with those tools already developed. Example Problem 25 Determine the influence coefficient matrix rorxransverae forces to be applied to the uniform cantilever ot Fig. A7.33 at the three points indicated.

DEFLECTtONS BY CASTIGLIANO'S THEORE!1

Application of Castigllano's theorem to

eq , (25),

(25)

t
Solution:

,@

,lID

3"""" 3 -e- 3

I.

b, b
q.

yields
I~ ) I ~p n .f
(26 )

="q q. '

=:J~)q"

c:::::::=::=J.~ q1.

Fig. A7.33

The steps in passing to eq. (26) may be demonstrated readily by writing out sq. (25) for, say, a set of three loads (a, n = 1,2,3), differentiating successively with respect to Pl.. ,'Pa and P" and then re-collecting in -catrr tx

rorn,
The matrix

~mnJ

gives the deflection at

the ext ernal points "a" ror urn t values of the loads? and 1s therefore, by definition, the
n
~trix

T~e choice and numbering at genera11zed forces are shawn on the figure. These forces were placed so that pr~vlously derived expressions for the a's could be used. The following member flexibility coefficients were computed. Note that the only non-zero coefficients ot mixed subscripts (i not equal to j) are those for loads common to an element.

of influence coefficients. ---COt1PARISON .ITH DUMt'lY-UNIT LOADS SQUATIONS

al,l =

It is

instr~ctive

to Nrtte eq.

(26)

out as

This expression was adopted from that developed for a transverse shear force on a cantilever beam segment in the preceding illustrative example.

L/~.dx
Q

L' =

BIEr =

aa::a

= a.....

and compar-e the expression with a typical term from the dumny-unlt load ~et~od equations, say SL 6=zu E: ' :n the ~~x equation (25a) ~he
A

[t~iJ Is the unit

Ing to the symbol

(virt~al) TIU"

matrix correspondThe

This expreSSion is for a couple on the end of a cantilever segment (of length L/3). __
Ua " - au L/ 3

[2. i .:1 J
ing to

are the member' r Lexfb i Lt t Les correspondAi:

in the simple sum.

ydY_L a
EI - 18EI

L...,. T~e

ma tr tx pr'oduct

l'J:l]

:Pn (

J
0

;ives the :nember load distrlb~t~ons due to the real applied loads, hence these are the "B" leads. Finally the operation of matrix :nulti-

This expression is for the cross influence af a couple ~~d a shear load an a cantilever segment. Collecting in matrix form,

A7.22

DEFLECTIONS

OF

STRUCTURES

27
0

0 0 0

0 0 0

as per eq (26). T'he matrix

~~Nell's

G!Il11]

is seen to be svmetc-rc

[a ! ]
l

27
1

L'
3EI

0 0
0

6L
0 0

1 6L 1
0 0

about the main diagonal as it ~ust be: from reciprocal theorsm A .c.' nm (see eq. (20)). mn

27
1

6L
-r-r w
1

A7.10 Member Flexibility Coefficients: Compilation of a Library.

6L

The unit load matrix [G imJ was computed by

Several ~ember ~lexibi:ity coefficients are derived below for various members ~d leadings. A more comprehensive listing 1s available in :he paper by Wehle and Lansing referenced ea:'lier.
BARS

successively applying unit leads at ~oints 1, 2 and 3 and computing the values of the q's by statics.
1
0 0 0 0

The energy in a uniform bar under var-ytng axial force (Fig. A7.34l is

[G1m

1 L/3
1

1 0

1 2L/3 L/3

1
0

Then referring to eq . (20),

Fig. A7.34

Note that the first column of

~ imJ

gives the

values at the q's obtained ror a unit load at point "1" with no other loads applied. The second column gives the q's tor a unit load at point "2" only, and so forth. Finally,

= ...!c.. 3AE
and,
0

1..
'Z1

o
o
o
1

~~=3iIOl
o

a1 j

1 L/3 1 2L/j 0 1.. 27 1.. 6L 0

0

1 L/
1 0

06iL~
o o
0 0 0

a
o
1

~ -u L J~ = ,q ~q = Ai: L
1

0 L/3 0

0 276"

.l. ,];1

6L ~ 3

2L L/3 0

An equally likely choice at generalized forces for the above case is shown ~n Fig. A7.34a. The strain energy is (x measured from .free end)

Multiplying (see appendix),

Fig. A7.3-4a

t~xee

Should the deflections be desired at the points one torms

'U _ 1 JLx , dx - d q/" - AE

" 0

3AE

L'

~'U = AE IJLxdx = oqi)qj

0

= 2AE

L'

.Y!

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.23

In ~he case of tapered ~embers the coefflc1ents are determined by evaluating integrals of the form
o. - 1 JL!(X)dx
o

U =

2~I gql
a11 =...h...
Fig. A7.34e
NOTE: The coen-

and from which

Ij -]"

A"1Xi

Such a quadrature can always be made in these problema. For the linearly tapered bar the results may be obtained as functions of the end area ratios. ThUS, Wehle and Lansing give
ajj

EI

a11

=~.

= 3EI

L'

1"

11
SHEAR PANELS

icients for bars and beams are directly analogous (compare several cases) so that for tapered beams one uses the results {or tapered bars with EI in place of AB.

alj = 6A E !j 1
Fig. A7.34b

For the rectangular shear panel with a uniform shear flow Q on all edges (Fig. A7.34t) i

0jj

3AE . ~jj
1

~
',j\..

3
,

S = surface area
Fig. A7.34f

:
I

The trapezo idal shear panel (Fig. A7.34g) is treated approximately by using the average shear flow on the non-parallel sides as though it were constant throughout the sheet. Thus

S = surface area
Fig. A7. 34c::::::!:d ~e~ v~tes lin~rlY :J: ~

a11 =.....

Gt

= BEAm

Since by statics qj =
Ql -.-, one could use Q
Fig. A7.34g h, as j a. an alternate choice of generalized torce and

The energy in the uniform beam A7.34d is given by

qj

0:

Fig.

ajj -

_(h.)' S. h. lit

TORSION BAR

r1--====<' 1
L

1 q,

A uniform shaft under torque qi has strain

energy Then
0.11

Then

Fig. A7.34d

L = GJ

) a11 =~'U=_L_( ,ql' 3El = Ojj

A7.11 Application of Matrix Methods to Various Structures.

aQl,Qj
An
alterr~te

O'u

=_L_ (= Ojl)

6EI

choice of generalized forces

tor the beam of ?ig. A7.34d is shown in Fig.

A7.34l.

Example ?rob 1em 26 The tubular steel truss ot Fig. A7.35 is to be analyzed for vertical deflections at points ~ and F under several load conditions in which vertical leads are to ~e applied to all joints excepting A and D. The crass sectional areas at tube member~ are given en the figure. Set up the matrix form at expression tor the

7-

A7 24

DEFLECTIONS OF
Con.tant Axial Load; 'l.i-'l.j
-=~

STRUCTURES

'K
1 2 3 0
0 0

2 1.0 0 -1.0 1.25 -1.03 - .75

.848

3
0

1.0 0
0

!ben

"'1.- e(i1'" ~jj'"

OC1j+<:o.j1
.. L!AE

0
C

Fig. A7.35

Fig. A7.35a

deflections at points E and r. Solution: The member flexibility coefficient for a unitorm bar under constant axial load is L/AE. Fig. A7.36a gives the ~umberlng scheme applied to the members and the q's (these being one and the same, since q is constant in a given member). Fig. A7.36b shows the numbering scheme adopted for the external loading points.

~i~

0
0

0
0

1.25

5
6

-1.03
- .75

-1.0 1.13 -.825 .20

0
1.13

7 8
9

.848
-1.55

-.165
0.40

-.825 .20

0.40

Then the matrix triple product

tsl3J3 DJsL
1 2
8

plo

Pa

t=]
rs,
o.
Il,

= Emi] [OiJ]

~Jn]
eq. (26)
257 252 257 252 389 789 252
P,

was formed giVing,

Ip,

-440

pe~

389 927 252

7Sti

Fig. A1.36a

Fig. A7.36b

Member flexibility coefficients were collected in matrix form as

=E

389 257
e..-

p'l P.
J
1nche s

6,

389

789

~
2
3

a a a
1<6

P,

1 92.2

a
92.2

a
0
138

["1J] i
5
6

7
8

a a a a a a a a

a a a a
0

a a a
0 0

a a a a
55.7

a a a
0

a a a a a
185

a
0

a
0 0

a a a a
155

a a

a a a

a a a a a a a
55.7

a a

a a a a a a a a
229

The results here give the deflections of all four points. Since only the deflections of points 3 and 4 were desired the ~lrst two rows
of

tmnJ

may be dropped aut.

The same result

could have been achieved by leaving out the

first two raws of @",i] 0he transpose of

@bl])

In the case of a Pin jointed truss, where only a single generalized force is required to describe the strain energy per member, the matrix or member fleXibility coeffic18nts 15 a diagonal matrix as above. Unit lead distributions were obtained by placing unit loads succeSSively at external loading points 1, 2, 3 and 4 (Fig. A7.36b). The results were collected in matrix form as

The matrix fo~ of equation above ~s useful in organiZing the computation of deflections for a number of different loading conditions. Thus should there be several different sets of external loads P corresponding to various loadn, ing conditions, each set is placed in column fo~ giving the loads as the rectan~~lar matriX

~nk] ,

k different numerical :ubscr'1pts for the

load conditions.

The matrix product

tip'

A7.25

i ."= -" P L-':c-J<....... ~llJlJ . 'nn nk

t~e ~arlous load conditions (k).
~a.T:J.ple

L'
(26b)

au

a"

AB;SEI

LAB~J
= L8~
a ..
GJ

r1SW 21'1e5 t.ne 'deflections at each point (:r,; for

a"

a ..

L;~

3EI

a"

?roblem 27

Deflections at points 1, 2, 3 and 4 of the truss of ?lg. A7.35 are Geslred for the following loading conditions:
Ccndf t t on No.

a,.

BO = a,. = La~ 2EI

= L8~

EI

P,
2500 -1200 1800

P.
2000 -800 1470

1 2
3

eco
-2100
-1200

P,

P, ( see Fig.
450 A7.36bl 01750 -1100
GJ

Collected in matrix form. (noting that

=1.3 EI )
1

2
0 3,' 0 0 0 0 0 0

3 0 0 '.0 0 0 0 0 0

4 0 0 0

5 0 0 0 36,0 0 0 0

,
0 0 0 0 0 36.0 0

8 0 0 0 0 0

Solution:
T~e

'.0 0 0 0 0 0 0 0

0 0 0 0 0 0 40.8 0

matrix product formed per eq. (26b) was

3 4

set up as
440 389 257 389 2500 -1200 2000 - 800

15,500

~'"'J
~ample

1
;;

389 927 252 789

1800
1470

,
7

see
0 0 0

""

sae
0

sae

15.500

257 252 257 252 389 789 252 789

800 -2100 -1200 450 -1750 -1100

The unit load distributions

Pr"oblem 28

~1~ were

obtained
IT~bered

Tor the landing gear unit of SXanple Problem 20, Fig. A7.23 find the matrix of influence soef~lclents relating deflections due to 11ft and drag loads acting at point A and torque about t~e axle A-B.

by applying unit external applied loads, and directed as in Fig. A7.37b.

~
2

2
0 1.0
0

3
0 0

Sol:ltion:
st~ucture was divided into elements set of internal generalized forces ap~lied as shown in ~~g. A?37a. (Torques and noments are shown vectorlally by R.H. rules). Axial stresses were neglected in C-B.

1 1.0 0
0

7he

~nd ~~e

3
4

LO
0 0 1.026
2.311

,342 3 0
0

0 0
-.937

5
6
7

.342
0

0
L..

l.G

At this point the e~g1neer ~ay consider t~e problem as solved, for the re~2inlng computation is a ~outine cperation:

Fig. A7.37a

Fig. A7.37b

~he

following ~e~ber flexibility coef::cients 'Nere oaterainec

Exa."'11ple Problem 29 T~e bea~ at example problem 21 is to be resolved by the T~trlx wethoas ?resented ~ereln.

A7.26

DEFLECTIONS" OF

STRUCTURES

Influence coefficients for points ?, G and H are to be found. Solution: Fig. A7.38 shows the choice and numbering of generalized forces.
q.

The matrix triple pr-cduct

completes the calculation.

Example Problem 30 Deflections of statically indeterminate structures often may be computed successfully by the methods of this chapter proVided that some auxiliary means is employed to obtain an approximation to :he true ~nternal force distribution. The exac~ ~nternal ~orce distribution is not necessarily required in making deflection calculations inasmuch as such a calculation amounts to an integration over the structure - an operation which tends to average Fig. A7.38 out any errors. Thus one may use the er~ineeri~g theory of bending (E.T.B.), experimental data, previous experience, etc. to obtain reasonable No torces were shown applied to the lower tlange esti~tes of the internal force distribution for unit loadings. elements as these were known to ue equal to In the folloWing probl~ ~he ~trix of inthose ot the upper tlange due to symmetry. fluence coefficients is dete~ined tor a sin~le Entries were made for ll in matrix form as beiJ cell, t~Iee-bay box beam (3 times indetermi~2te) low. Entries tor !l33 and !l88 were quadrUpled as by using the ~.T.B. these occur in ~ identical members each on !QQ Fig. A7.39a shows an idealized doubly and bottom. Entries tor C1e1l' au and c were symmetric single cell cantilever box beam having doubled. (See Art. A7.l0 tor coefficient three bays. Determine the ~trix ot influenCe formulae. ) coefficients tor the six point net indicated.

-",==="'_q.. .

/'" (caef!'s a 33 '" a~3 -;. a ~3/ call'td \ ' fr.two ===='~9 ~'?....;, members)

a
1
10

, I ,
e.se
=0'
1

,
10"

t,~

a a
10-

aoes
17.8

....
17 8

....

'-'.
2500

....
8.89

.....

Unit load Yalues were obtained as in Fig. A7.27, considered to be external loading number "two". S1ml1ar diagrams were drawn for unit loads at points "one" (H) and "three" (Ft(llt'. rig.
A.

Stringer Areas Constant", . 33 in a

20" 1:-5"

'-J.-

15.. -J

'

7-2"

1 1 2 3 1.0 .0667 1.201

2 0

Fig. A7.39a

a a
1.0 .0600 11.0: .0496 2.00

~~

a a a
0

Solution: Fig. A7.39b is an exploded view of the beam showing the placement and numbering of the internal generalized forces. Note that only the upper side of the beam was numbered, the lower side being identical by symmetry. Member fleXibility coefficients were computed by the formulas ot Art. A7.l0 and entered in ~trix form as below. Note tr2t all entries tor Which there Nere corresponding leads on the lower surface of the beam were dOUbled. ay this ~eans the total strain ener~J of the beam

a
.0540 2.184

5
6 7 8
9

a a
1.0 .0545 1.00

a
.0447 3.00

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A7 27

flow (q = ~) due to the torque developed in transferring the load to one side. The matriX triple product

completes the solution.

A7.i2 Denecncne and Angular Change ot tbe Elastic Curve of Simple Beams by the "Method of Elastic Weights" (Mobr's Method). Fig. A7.39b

was accounted ror. Note also that entries tor a~~, a 8 8 , a and a 1 0 1 0 were re-doubled as each at these q's act on two (identical) members.

In the calculation of structural deflections there occur many steps involving simple integral properties of elementary functions. The Method of Elastic Weights (and the Area Moment Method to folloW in Art. A7.l4) owes its popularity in large measure to the fact that it enables the analyst to write down many of these integral properties almost by inspectIon, relYing as it does upon the analyst's familiarity with the properties of Simple geometric figures. For finding the deflection of a paint on a simply supported beam relative to a lIne joining the supports, the Method of Elastic Weights states: The deflectIon at point A on the elastic curve f ~ simple beam 1s equal-to the bending moment at A due to the diagram acting as ~ distributed beam load. -- -Spelled out in steps; i-The ~ diagram is drawn just as it occurs due to the applied beam load ii - This dIa~ is VisualIzed as being the loading on a second beam (the conjugate beam) supported at the paints of reference for the deflection desired iii - The bending moment iro this conjugate beam is found at the station where the deflection of the original beam was desired. This bending moment is equal to the desired deflection. To prove the theorem, consider the dummyunit load (Virtual work) equation

ii

Note;

VOID SPACES INDICATE ZEROS.

Unit load distributions were obtained tor successive applications at unit loads to points one through six (Fig. A7.39a). The internal torces Dredicted by the E.T.E, for a load thraugh- the shear center (center of beam, due to symmetry) were superposed on the uniform shear

~
1
2

1 0.3 0.1 0.1

Z Z

0.1

-0.1
0.3 2
2

~L~ =

3
4

5
6 7

.0875 .0375 .0125 2.67 2.67

.0l25 -.0375

0.15
0.05

0.05

_0.05 0.15 1.33 1.33 .0179 0.100 _.0267 .:J713

Z Z

a
9

.0875 2.67 2.67

.0023

0.05 1.33 1.33 .0713 .0267 .0179

10
11

.osea
.020
0023

12

1-.020
'I

13
14

I .0423
3 3

.cess .csss
1 1

-.=
.100 1
1

.oasa

This expression equates the exte~al virtual work done by a unit load, applied at a point de:lectlng an amount 6, to the internal Virtual work on a beam element experiencing an angular change d g = ~~. The sum (integral) of such expressions throughout a beam gives the total

15

3 3

a
2

A7. :.::8

DEFLECTIONS OF

STHUCTURES

deflection at tne ccmt (c r . eq , 103). ..~e now show tr~t the ceflecticr- expression, uSin: :he above equat~on, is the SaT.9 as the bending moment expreSSion fer ~ simple beam loaded by
--"an

In rig. r

co~sider

Mex
81

~3

a load on a

si~~ly

elastic weight 21 In Fig. A7-40, the la~cing at (a) pr ocuces the real moments of (0). Consider the ce r jeotlons ot points 3 and C due to the ar.gular
ch~nge Mdx ~

~ Md.x

supported bea~ a~c dete~ine the benG:n; 200ent at pOints band c G~e ~o ~dx acting at pOl~t c. "IT Mdx L XLdx
4EI

:2
L

BEl

in

a 0 ea~

",1 ~~emen t

a tA" (Fig. A'7-4C':).

1--, -....-L
I L

-1- 4

l.h-Lhj
' 4

I')

values of ~~e te~ bending ~o~e~ts at points 0 and care ldentica: to :~e deflections at band c by the virtual ~ork equa:ions. ~~e noment diagram ~ for a unit ~oad at b a~d c (Figs. d and e) is numerically pr sc t se Ly tne sane as the influence line for illo~e~t ~t ~oi~ts b and c. Therefore deflections of a si~ple bea~ can be determined by cons Ider mg the :1 cur-ve as an
~hese

"

~WEI diagram ~
.., !"odx
(h)

81

imaginary beam loading. ~he bending ~oment at any pOint due to this !1 load.ing equals the 1e-

EI
flection of the beam
ili~der

the gi7en loads.

EIpj
(c)

Mdx

-=

L~4

Likewise it is easily pr)ved :~zt :he ~n~~ lar change at any section of a sl~ply supported beam is equal to the shear at that section due to ~ !'! dt azr-am acting as a beam Lcac . Sol
A1.13 Example Problems

(d) mb, moment diagram for unit load acting downward at point B

r16 L
f ---(el me" unit load at point C
Mdx

Examnle Problem 31. ~ind :he ~2rt~cal deflection and slope of paints?. and b :or be~ and loading sho'Nn in Fig. A7.41. T~e lower F~g. shows the :noment c Iagr-am for load? acting at center at a sl~ple beam.

lEI

r-~1
If)

IP
b

4EI
Fig. A7.40

3 Mdx

4EI

lMdx

f P 2

~-;P
2
L L

EI Constant

For a unit load at point b, Fig. d shows the m diagram. The value of m at the midpOint of dx (point a) = LIS. Hence

PL
16

Fig. A7.11

-6"-

b = ~~ . ~
16

~~

PL'

16

PL'

For deflection ot paint c, ,1raw m diagram ror a unit load at C (see Fig. e). Value or ~ on eleaerrt dx = L Hence
00
.. For srmoncrty the points A, Band C were placed at the onequarter span points. The reader may satisfy himself with the general character of the proof by substituting xA, xB and XC for the point locations and then following through the argument once again.

Deflection at point a 8Guals bending ~cm ent due to M diagram as a load di vtoec by EI. (See lower Fig. at Fig. A7.~1)
6a

PL2 ( 16

4. -

PL2 64

L) 1 11 PL3 12 EI = 768 EI

Db

= (~2

~ i~2 ~ )Elr
=
3 PL2 64 EI

1a ~~3

The angUlar c~nge ot any point equals the shear due to ~JEI diagram as a load.

Ua

= (PL2 _ PL2)2:..
16 64 E1

-,
A7.29

o (Slope is horizontal cr- no change from origi~al direction of beam axis. ) ?rcbl,:::n 32. De t cr-ntne ':;;'a cer Iec t Lcn of a sixple be~~ loaded unifo~y as shown in Fig. A7.42. The bending moment exor ess i on for a unlforn loac M = wlx - ,NX2 or ~arabolic as
~x":.:nJ'"'!

I<i
,

Airplane

1
I
A

50 " - ~:)O"130"

Anercn E
B

rHear Beam

FIA
C

..-- Ercnt Bean-

2"

2"

Plan View I: 2 Wmg

/
25li'iin.
3

shown in F'1;. A7.42a. The deflection at mfd?oint equals the Jending J10ment due to M diagram as a load.

r 0"1 IiffffHHHflH fJ HJ f OHm!! H Hbj"

.,.....Air Load
i- Load on Rear Beam

)
15~/in.

r'
2""
wL

wz in.

lll!

'LPl ! ! If Y+
T
wL

Fig. A7.43

Fig. A7.42

.-?'7A-.<(..M Diagram

Solution:- Due to the beam variable moment ot inertia the beam length between A and C will be divided into 10 equal strips of 10 inches each. The bending moment M at the midpoint at each will be calculated. The elastic weight for each strip will equal Mds, where ds = lOW and I the moment

....!....wL~""Area 24
Fig. A7.42a

24 wL"

of inertia at midpoint of the strip. These elastic loads are then considered at loads on an 1~ aginary beam of length AC and simply supported at A and C. ~he bending moment on this Dnagi~ary J3am at ~o:nt B ~ill equ~l the deflection of 8 with r3spect to line jOining AC. The bending moment at C = 15 x 30 x 15 ?
10

x 15 x 10 = 8250"#

6center

=(2~ NL3 . ~ - d4
c

wL3 .

~ ~)

SlI

The shear load at C = (15 + 25) x 30 =600# -2-

wL4

384

sr

ccent.er

=(2~

14 'NL3 - 2 WL 3) Ell = 0

Slope at supports
SXP~;le P~QbleJ1

= the

reaction

1 wL3 24 EI

33.

Fig. A7.43 shews the plan ~iew of one-half of a cantilever wing. The aileron is supported on brackets at ?oints D, E and F with self-aligning bearings. The brackets are attached to the wing rear beam at ?oints A, B,and C. ~hen the Wing bends under the air load the aileron must likewise bend since it is connected to wing at t~~ee ?oints. In the desi~ of the aileron beam and similarly for cases of wihg flaps this deflection produces critical bending ~oments. Assuming that the running load distriblted to the rear beam as the wing bends as a uni~ is as shown i~ the Fig., tind the deflection of pOint 3 with respect to straight line joining ~oints A and C, which will be the deflection of with respect to line jOining D and F it bracket deflection is neglected. 7he moment or inertia of the rear beam between A and C 'varies as indicated in :he Table A7.6

Bending moment expression be~#een points C and A equals, :M: = 8250 + 60Qx + 12.5x2 , where x = 0 to 100. Table A7.6 gives the detailed calculations tor the strip elastic loads. The I values assumed are typical values tor a aluminum alloy beam carrying the given load. The modulus ot elasticity E = 10 x 106 1s constant and thus can be omitted until the final calculations. The figure below the table shows the elastic loads on the imaginary beam.
Table A7.6 Strip

'0.
I 2 J

d.

'0.

e ,
lIo_nt at mid_ point US63 20063 31063 44580 60580 79050 1000.50 123$50 149550 118150 0

I at JlIidpoint

Elastic load

rt:

...

6 7
0

10

, s
co ro .o
0

IO IO IO IO IO IO

5.S 6.5 7.S

9.5

'.5

10 10 10 10
OJ eeo cOJ

12.0 16.0 20.0 24.0 28.0

21000 30900 41450 52300 63850 6.5940 62530 61770 62320 63600

'"

so

':'J " _0 -o 0'

eo

0
0>
~

co eo eo OJ .c l.D

en

0
0

en
~

0 0
0
~

'"

OJ

A7.30

DEFLECTIONS OF

STRUCTURES

Bending ~oment at paint B due to above elastic loading = 7,100,000 .'. de~lection at B ~elatlve to line AC = 7,100,000 = .71 inch E::IO,OOO,COO Example Problem 34 Fig. A7.43a shows a section of a cantilever wing sea plane. The wing beams are attached to the hull at paints A and B. Due to wing loads the wing will deflect vertically relative to attachment points AB. Thus installations such as piping, controls, etc., ~ust be so located as not to interfere with the wing deflections between A and B. For illustrative purposes a simplified loading r~ been assumed as shown in the figure. EI has been assumed as constant whereas the practical case would involve variable I. For the given loading deter~ine the deflection of point C with respect to the support pOints A and B. Also determine the vertical deflection of the tip ppints D and E.
~r400"
D

A7.14 Deflections of Beams by Moment Area Method

For certain types of beam problems the oethad of moment areas has advantages and this ~eth ad is frequently used in routine analysis. Angular Change PrinCiple. Fig. A7.44 shows a cantilever beam. Let it be required to deter~ine the angular change of the elastic line be~eent~pointsA~8~to~givenlMdl~.

From the equation of virtual work, we have

.,

::.

0 0

0 0

300'~1w

::,
I
EI Constant
B

0 0

'\

0 0

40"

Fig. A7.43a
E

Moment Diagram

o! in.' o 1000
,
01 ., ,

Fig. A7.43b

"" 0,
0,

Fig. A7.43c

!
Fig. A7.43d

Solution:- Fig. A7.43b shows the bending moment diagram for the given wing loading. To tind the deflection of C normal to line joining AS we treat the moment diagram as a load on a 1lllaglnary beam of length AB and simply supported at A and B (See Fig. A7.43c.) The deflection ot C is equal numerically to the bending moment on this tlctic i ous beam, Hence EIO c = 25920 x 40 - 25920 x 20 518000 or 6c EI To tind the tip detlection. we place the elastic Fig. A7.44 loads (area or moment; df agram ) on an imaginary beam Simply supported at the tip D and E (See Fig. A7.43d). The bendlngmoment on this imagiClb = (A ~:n Where en Is the moment at any nary beam. at points A or B will equal numerically )B EI ' sect i cn, distant x trom 3 due the deflection of these points With respect to to unit hypothetical couole the tip points D and E and since points A and B applied at B. But ~ = ~lty actually do not move this deflection will be the at all points between Band A. movement of the tip paints with respect to the beam support points. Therefore as =(A ~ Bending moment at A = 193420 x 700 - 40000 x )B EI 433 - 127500 x 124 = 102200000. .'. Ot1p (Moment area first developed by Prof. C. E. Greene and 102,200,000 published in 1874. )
EI

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.31

Referring to Fig. A7.44, this ex~ressian represents the area ot the M diagram between points
EI

Henceos

= (_

PL2
Z

03

L) ~
EI

= _ PL3
3EI

8 and A. Thus the first principle:- "The :hange 1rr slope of the elastic line of a beam between any ~wo ~o1nts A and 8 is nw~erically equal to the area of the :1 diagram betw"een
EI

these two
Deflect~on

pol~ts.n

Principle In Fig. A7.44 dete~ine the deflection of paint B normal to tangent of elastic curve at A. In Fig. A7.44 this deflection would be vertical since tangent to elastic line at A is horizontal. From virtual work expression Os = I~ ~~ ~, where m 1s the moment at any section A dist~~ce x from 8 due to a unit hypothetical vertical load acting at B. Hence m = l.x = x :or any paint between Band A. Hence Max This -EI x 8 = B eXDression ~epresants the 1st noment of the M uiagracaoout a vertical t.hru B. 'fhus the

Example Problem 36 Fig. A7.46 illustrates the same simplified wing and loadiP~ as used in example prOblem 34. Find the deflection of paint C nor~l to line joining the support points A and B. Also find the deflection of the tip paints D and E relative to support pOints A and S.

g
o
o o

~B

1 T
EI Constant E
Fig. A7. 46

h ~!,ir!'J;
~

= =

Moment
in.~

o Diagram

TijijQ

fA

r-

~67"-'" I ~

r--

~ 21 ifi:!it:I 0 11'0 N a;: 576 " ... 720"

~i ~ ,~o ~4

0 Oi

Solution:Due to symmetry ot loading, the tangent to deflection principle ot the moment area method the deflected elastic l1na at the center line of can be stated as follows:- "The deflection of a airplane is horizontal. Therefore, we will find point A on the elastic line at a beam in bending the deflection of points A or B away tram the normal to the tangent of ~he elastic line at a horizontal tangent ot the deflected beam at pOint point 3 is equal numerically to the statical mO- C ~hich is equivalent to vertical deflection of ment or the !1 area between points "An and "B" C with respect to line AB. . EI Thus to find vertical deflection or A with respect to horizontal tangent at C take moments about point A". of the M diagram as a load between points A and Illustrative Problems EI ExaT.ple Problem 35. Determine the Slope C about point A. and vertical deflection at the rree end B of the cantilever beam shown in Fig. A7.45. 1 1s con- Whence
~I

stant.

(area)

(arm)

1~~0
1--L ------i

(tangent at C)

=E~

(650

~ 646) 40 x ZO

1 EI
Fig. A7.45

(518400)

= deflection of

C normal to AB.

~~am
~

PL2

__

.!L--J
3

2'

To tind the vertical deflection of the tip point D with respect to line AB. tirst rind deflection of D with ~espect to horizontal tangent at C and subtract deflection of a ~ith respect to tangent at C.

Solutlon:- The ~oment diagram ror given load is triangular as shown in rig. A7.45. Since the beam is fixed at A, the elastic line at A is horizontal or slope is zero. Therefore true slope at 8 equals ar.gular c~~ge becNeen A and 8 wm cn equals area of 'nomerrt diagram between A
and B diVided by E1.

~ (respect to tangent at C)

= E; (40000 x 267 +

Hence

"8

= (-PLo

1 LIZ) I

PL Z

1Z7500 576 25920 7Z0) = (10Z700, 000) (See Fig. A7.46 for areas and arms or M/EI diagram). subtracting the deflection of A with ~e spect to C as found above we obtain

ir

- 2EI

l~gO

(respect to Ltne AS) = Ell (102,700,000

The vertical deflection at B is equal to the 1st 518400) = SI (102,180,000) :no~ent of the :noment diag~am about ~oint S diVided by S1, s~nce tangent to elastic curve at A 1s norlzontal aue to ~ixed support.

A7.32 A7.15
Beam Fixed End Moments by

DEFLECTIONS OF
~e~hod

STRUCTURES

of Area

:1A

=-

Pab 2 and MB

lIOllIents

L'"""

=-

?ta 2 / L " l ' o ' h e r e b:: (L-a)

From the rNQ prinCiples of area ~o~ent5 ~s given in Art. A7.14, it is evident that the deflection and 510;e of the elastic curve depend on the amount of bending ~oment area and its location or its center of graVity. Fig. A7.47 shows a beam r1xed at the ends and car~Jlng a single load P as shown. T~e oend ~ng moment shown in (c) can be considered as made up or two parts, namely that for a load P acting on a simply supported beam wmcn gives the triangular diagram Hlth value Pa (L-a)/L tor the moment at the load ~otnt, and secondly a trapezoidal moment diagra~ of negative sign with values of MA and ME and of such magnitude as to make the slope of the beam elastic curle zero or horizontal at the support paints A and B, since the beam is considered fixed at A and B. The end moments MA and ~B are statically indeterminate, however, with the use of the ~NO moment area principles they are easily determined. In Fig. b the slope of elastic curve at. A and a is zero or horizontal, thus the change in slope becNeen A and B is zero. By the 1st

To find the fixed end ~owents ~or a te~~ ""itt variable moment of Iner-t-Ia use the t""iI crucrana in place of :he ~oment clagraos. Problem 37 Fig. A7.48 shows a :1x-ended two concentrated loads. finS the moments MA and roB'
2xa~ple
je~~ carry~ng f~xed-end

M9"......:.-....12"~9" +.~tB
__ I ;

4t=

j 100.

200*

t:',

(a)

Ih)

Fig. A7.48

-l1-lO'~

15MA

~L

(e)

~-----~B

(b]

~lastic Curve
Fig.A7.47
--,

~~T
principle of area moments, this ~ans that the algebraic sum of the moment areas be~Neen A and 8 equal zero. Hence in Fig. c
(-l1A -!'lbl L (Fa (L - a) -2+ L L .2
0 - - - -

~a(t-a)

lEI'' '

(c}

_ -(AI

Solutlon;- Fl~. b shows the static ~o~ent dla~ao ass~i~g the beam simplY supported a~ A and 3. For SImpliCity in finding areas and tak~r.g moments of the moment areas the :noment Qia~~ ~~s been divided into the 4 simple shapes as ShO','iC. "he centroid of each portion is shown together ~ith the area which is shovm as a concent~ated load at the centroids. Fig. C shows the moment diagrams due to .~~ known ~oments MA and Mg. ~he area of these trianglas is shown as a concentrated 10aG at t~e centroids. Since the change in slope of the elastic curve bevNeen A and B 1s zero, the area ot these moment diagrams must equal zer~, hence
5265 + 14040 + 2160 + 6885 + lSMA + 15MB:: 0

In Fi~. b the detlection or B away tram a tangent to elastic curve at A is zero, and also detlectlon ot A away fram tangent to elastIc curve at B is zero. ThUS by moment area prinCiple, the ~oment of the moment diagrams of Fig. C about points A, or B is zero. Taking moments about point A:-

or
15MA + 15MB + 28350 = 0 - - - - - - - - - - (1 ) ~he deflection of point A away :Tom tangent to elastic curve at B is zero, therefore the ~irst moment of the moment diagrams about point A equals zero. Hence, 5265 x 6 + 15 x 14040 + 17 x 2160 + 24 x 6855 + l5CMA + 30G Me = 0 or 150 MA + 300 ME + 444600 ,. 0- - - -(2) solving equations (1) and (2), e obtain MA = - 816 in. lb. ME = - 1074 in. lbs. With the end moments known, the ~etlec~ion or slope of any ~oint on :he elast~c curve between A and B can be found by use of the 2 princ~ples of area :n.oments.

~ L: 0 -

- - - - - - - - - - - - - - (E)

solVIng equations A and B for MA and ME

A7 33

A7.l6

Truss Deflection by Method of Weights

Xla~tic

Ir the deflect~on of several or sll the jOints of a trussed structure are required, the method of elastic ~elghts ~y save considerable ::~e over the method of virtual work used in previous articles of this chapter. !he illethod in general consists of finding the illagnitude and location of the elastic ve i gnt for each member of a truss due to a strain from a given truss loading or condition and applying these elastic weights as concentrated loads on an imaginary beam. The bending aoment on this imaginary beam due to this elastic loading equals numerically the deflection of the given truss structure. Consider the tr~ss of dia;ram (1) of Fig. A7.49. Diagram (2) shows the deflection curve for the truss for a ~L shortening at member be, all ether members considered rigid. This deflection diagram can be determined by the virtual work ex pression 6 = u~L. Thus for deflection of jOint O. apply a uhit vertical load acting down at joint O. The stress m in bar hc due to this unit load = 2 2P = 4P. Therefore

1M c~~-- ,yf1J\
,'l. 0\c
I

90 ~;.-'- . ;Jl
,-

,/1'1\. <ltokih
I

., I

(I) g

r-rr :

6 Panels @ p

Bar be
u6L _

"F" -

---sr--

4P6~L.L Shortening
I.!.. uLb 3r c

~lection Diagram

(2)

.: . Lbc

of Member be Elastic Load on Imaginary Beam t .::.Lbc 3r (3) (4)

~~entCurve
Bar ck
6L(d+6p)

for Elastic Loads

6L
f. rl,

SL, djip t
t

/?\~ (5) ~ection Curve for !J. L Lengthening

a
K

L"'[':'"

1M. d+2p of

Member ck (6)

"'--Ela~tiC Loads on

:3

Rigid Arm

r-

3r

t:. L! d+6W

6pr~ l:.L(d+6p)

.::. Ld Imaginary ~ Beam

The deflection at other = lower chord jOints could 3r be round in a si~ilar ~anner by placing a unit load at these joints. Diagram (2) shows the resu~ting deflection curve. 7hls diagram is ~lalnly the influence line for stress in bar be multiplied by ~Lbc' Diagram (3) shows an imaglnarJ heam loaded ~ith an elastic load 6Lbc acting along a verti4P

~iiiz:
Fig. A7.49 Table A 7.7 Equations for Elastic Weights Elastic- Weight for Chord Members (See Member ab) Lower Chord Upper Chord
w=~L h

cal line thru joint O. the moment center for obtaining the stress in bar be. The beam reactions ~cr this elastic loading are also given. Diagram (4) shows the bea:n bending moment diagram due to the elastic load at point O. I~ is noticed that this moment diagram is identical to the deflection diagram for the truss as shown in
dt agr-am (2).

tab

&1:0""i t
elL w=I

The elastic we Lght of a member' is therefore equal to the =ember defo~ation diVided by the a~ r to its ~oment center. If this elastic load is applied to an i~gi~ary beam corresponding to the truss ~ower chord, the bendi~g ::noment on this Imagfnar'y beam. ,/.111 equal to the true tr~ss deflection. Diagram 5. 6 and 7 of Fig. A7.49 gives a si~ilar study and the resul~s for a ~ lengthenlng of ::nember CK. The stress moment center :or this ciagonal ~ember lies at point O~ which 11es outs1~e the truss. The elastic Neight ~ at rl point 0 1 can be replaced by an equivalent system at points 0 and ~ on the imaginary beam as shown in Diagram (6). These elastic loads prOduce a bending ~oment diagr~ (Diagram 7) identical to the deflection dia~am of diagram (5). Table A7.7 gives a s~~J of the equations for the elastic .vefght.s of truss chord. and 'Neb ~embers together Nith their location and sign.

/y\/\ r
w=r
0

j,L

tab

~\t

i/NJ\I~ .~~' I aL
~

w=

Member Axial Deformation Perpendicular Arm to Moment Center

J.L r

The moment center 0 of a chord member is the intersection of the other t ..... o member-s cut by the section used in determining the load in that member by the method of moments The Signor the elastic weight w for a. chord member is pj.ua if it tends to produce downwardder Iec t t on of its point of application. Thus for a simple truss compression in top chord or tension in botton chord pr-oduces downward or positive elastic weight


A 7 .34

tt

rmn=rry

DEFLECTIONS

OF

STRUCTURES
1000

Table A7. 7 (continued)

/

WEB DIAGONAL MEMBERS (See Member ab}

wO 6L r ...... __ 1,_k:-::.._____

/<'_:'~ - ~/"</\~
,i'''l,e,
~ ~

'\

~( 000

'~~t.:~0$" ~ /3120~ 6870 \ 6870 \/ 3120 ~ --.1...

I

~(.5630

1000 1000 1.::1-7500 if"'

~

~5630;D

1000

W"~:i;P.

bI

'2000 2000 4@30" = 120" Fig. A7.50

CI

d~ '2000

...<iaI>...e

i5000

a
b

Id

FIg. A

p. bw=~=2.L
bJ,Sln 9 r:, a AL "L Q=-a w =~=;'1

I\&i\l\ \Q i
P=Q=T

r.

6L

fQ

~
P Q 6L r

this step are given in the figure, the stresses being w rl tten ad~acent to each member. The next step or steps is to compute the 3ember elastic weights, their location and their sense or dlrection. Tables A7.8 and A7.9 gives these ca.Lculations. Table ..1.7:8
Chord Member
E1as~ic

Weigbts

Kellber Length L

Area A

L~'

For a truas diagonal member the elastic weights P & Q have opposite signs and are assumed to be directed toward each other or away according as the member Is in compressian or tension. In t1g. a, P is greater than Q and P is located at the end at the diagonal nearest the moment center O. Downward elastic weights are plus.
TABLE J,.7. 7

AB

~i, , ,. I
ee
cd

, I ,

~ ~L IA.rJI ~:29z106 I r
1
,

i Elastic
I

l , ~~L ,
.OO~465

Point of allplicatiol1 joint

,.
ac

ao so ao ac ac

.2172 _5630 .2172 _7500 .2172 _5630 .U98 3120 1~~98 .1198 6870 . 1198 3120

".

_.0268 _.0357 -.0256 .0259 .0592 .. 0592 .0269

24,IQ011171 24 .0014871 24 .001117 24 .001117 .002465 24 i .001117

, c ,

I
I

A
C

I" "

(CONTINUED)

Table A7.9

.eb Mesber

Xl.ati~ lfe~ght5.

~
P, I
d

<

Truss Verticals (See Member ab)

Ai~l~
br d

<, " a -

-on
J

PoQ'"

~L

. ,.
Bo
Cd De

. ,I
" } "

'L

w.d

'1

I
I

.il
'1

., " I :1 " I

~!

~I

II

Q -

A'

,c c,

28.25 .2421_5880 ._.0238112.75 .146 47101.0314 .... .1461_~3!>51_.01571 .. " .093 11771 .0123 " " " .09311177 .0123 " .146 _23.55 _.01.57 " " .146. 4710 .0314 " " .242 _.5380 _.0236 " "

_.00185 _.00246 _.00123 _.00095.5 _.00096.5 _.00123 _.00246 _.0018.5

a :12.75 . 0018~ .00246 :1 " .00123 " .000955 " C .00096.5 " .00123 " .00246 " " .00185

, ,
A

e, e,

'""?

p.'" OL

ho

Q Q=8L
d

Fig. A7.51 shows the elastic weights obtained tram Tables 8 and 9 applied to an tmagtna ry beam whose span equals that at the given truss. These elastic weights are the algebraic sum ot the elastic weights acting at each truss jorrrt
co "' 0
~

The elastic weight P acts at toot at vertical and downward it vertical Is in tension. Q. acta opposite to P at tar end at chord member cut by index section 1-1 used in tlnding stress in ab by method at sections.

" 13
0 0

;i
0

c-:
r-0

c~

g g
b

:3
0

g
0

c-

;i
0 0

cOJ

r-.

13 0
0

co
~

co
0 0

Fig. A7.5l

a
.17.17

1 1 1 1
B

Solution of Z%aaple PrOble...

The method at elastic weights as applied to truss detlection can be best explained by the solution at several Simple typical trusses. Example Problem 38 Fig. A7.50 shows a simply supported t~JSs symmetrically loaded. Since the axial detormationa in all the members must be tound, the tirst step is to tind the loads in all the membars due to the given loading. The results ot

Ra

c C d D- e 8 spaces at 15' _ .005412 Re = .005412

J' I

I'

The detlection at any joint equals the bending moment on the imaginary beam of Fig. A7.51. Den. at J,. = (.005412 + .00185)15 = .007252 x 15 = .109" Den. at b = .109 + (.007262 - .000507)15 .109 + .006755 x 15 = .209" Den. at B .209 + (.006755 - .002347)15 = .209 +

.004408 x 15

.275

ANATYSIS AND DESIr.N OF

Fl.Tr.HT

VEHICLE STRUCTURES

A7.35

nen. at c = .275+ (.004408- .0027)15= .301" beam loaded with the elastic weights from Table The slope of the elastic curve at the truss join A7.1l. Table A7.12 gives the calculation for the points equals the vertical shear at these points jOint deflections. for the beam of Fig. A7.5l. Example Problem 39 Find the vertical deflection of the joints of the Pratt truss as shown in Fig. A7.52. The member detormationsAL for each member due to the given loading are written adjacent to each member. Table A7.l0 gives the calculation of member elastic weights. Fig. A7.53 shows the imaginary Fig. A7.54 beam loaded with the elastic weights from Table A7.l0. The detlections are equal numerically to Tabb A7.11 the bending moments on .. t his beam.
0b = .01855

x 25 = .465"
.053 (~in Bar Bb) = .518" x 50 - .00387 x 25 - .833" .031 (~ of ecl = .864 x 75 - .00387 x 50 - .00623 x
!

0B = .465 + 0c = .01855 0c = .833 + On = .01855

b .083
. '"

Ilellber

6 L
.O82~

,
15.0 17.17 20.0

... -.ll ,
.00550 .00320 .00.480 .00322

Joint

....
U

25 = 1.03"
c . 091 d -.091 e .083

..
.n
18
BE

BC

-.056
.096 _.069 06

CD

21.1-6
20.0 17.17 15.00

~ o ~I~~I ~~I 'T r.-l

~
"0>

-.0738 .113
_.015 .108

~i:i~

....."
"

C
J
D

'=,

~o

~e:::,' ~

,~.,,~lI'
.061

BG

.00348 .00565 .00437 .00120

eaber~

r
Q

.061

.061

.066

.066'

.061

EIa.tie Weight ot Web

BCD

lfeiRDer
AD BE

6L

6@ 25" 1S0"

'1

-A!'1

JOin

-ll

Joint
B

"

Fig. A7.52

tc
Table A7.10 Elzstic Weight Chord lteaber
AD
be
Ke~erB

H
061 061 066 -.063 _.091

... =il ,

Joint
b b

.. ....
ID
El

CJ JD

-.0432 9.60 .072 9.60 _.063:1 9.27 .085 10.60 .0496 10.33 10.80 _.0254 9.27 _.0585
.?~!,5

_.00450
_.00748 _.00685 _.00802 .00134 .00480 _.000115 _.00274 _. ?~?28 _.00610

A
C J

.oraa I ~0.33

G7

'.60 9.60

r
G

9.60 .00450 11.20 .00643 10.60 .00600 11.26 .007.55 ~l?~3 _. ~0126 10.93 -.00454 11.26 .00891 10.60 ,00240 .00878 1~.20 9.60 .00610

r
1
I
C J

ac

3
3

CD

CO

3 30 3.

I I

.00203 .00203 .0022 .00277 00304

c C

JOint
b

,
I
lfellber

.lzstic Weigbt ot Web Me.bers

I
I

HI,

'1

p .. 6L

"l

Joint

r2

Q .~

Fig. A7.55

"
b

i
I

.b
be eC CD
ad

Ab

.121119.2 .053 dJ .128 19.2

i 25 .08g 119.2

-.00648 -.00667 _.00124 _.00417

I-'O~~

C
0

19.2 .00648 00 19.2 .00667 25 ' 19.2

Table A7.12

:gg~f~

hllel

PaDel Sher.r

Z Shear
.02 58

.IIloael1t

Shear " de-

.
o

12.5 z Point
A

= = =
~ ~

ql

= 0
q
I C

=
q
I

= 0
0

=
M

q
I

= = 0
~

AD

02158
6 .01350

ileet10D

CJ

~r
N

I
E

qt1;G

Fig. A7.53

IB
YG

-. -. _.02

. -.

~ .
04

.056 2

.07181 .07429

.8. .900
.930 .638

.508 .710

J
D I

Example Problem 40 Find the vertical jOint deflections ror the Example Problem 41 unsymmetrically loaded truss of Fig. A7.54. The Fig. A7.56 shows a simply supported truss ~ deformations ror all ~embers are given on the with cantilever overhang on each ena.' This simFigure. Table A7.1l gives the calculation of plified truss is representative of a cantilever the elastic weights, their signs a nd points of wtng beam the ruseaage attachment points being application. Fig. A7.55 shows the imaginary

error

A7.36

DEFLECTIONS

OF

STRUCTURES

at e and The ilL deformation In each truss memoer- due to the given external loading Is given on the figure. The complete truss elastic loading 'Nill be determined. vt th the elastic loading known the truss deflections from various reference lines are readily determined.

.'.

n"i
;
0

A' .M"".""",

r.---:. .. . .~ ~
~,
d'

:;.. 'if' '" \~"":"''' ",- <at. ~i;~,' ". ~,("p .. ~". ,;~" 0.c:. ~ ~ " tf},(~ . '01n74.~~ \072 '(', . .072 "(,"::,U74\ .U70\"~74\ .:l.."",;
Q

_~~...n'.M...., .o.a E .GSA ;)DS6CD6Q81170A

<

.e'

,,4

c ~~

4 Q 20" , - + - 2 i! 30" - - - 4 @ 20"

g ,

, , ,
0

0_ ~.Oj

a' A' b' 8'

c' C-

D'

~ g 3 ~ g g , ,, ~ " ", , I , I , , , , ,, ,, A a e'

0 0

. . :: . .. .. g s s ~ , , ,
,

Fig. A7.56

lever over-hang portion of the truss r eIat t ve to support points e and e l Since the cantilever :Jor1:iOn is not fixed 2. t e since the r-es tr-a i rrt Is ceter.nined by the truss between e' and e, this fact must be taken into account In loading the cantilever portion. The reactions on the beam of Fig. A7.S8 represent the Slope at e due to the elastic Ioad Ing be twee e and e' . This elastic reaction in acting in the reverse direction Is therefore applied as a ::"oad to the imaginary beam between e and a as shown 1 Fig. A7.59.
N N
~
~,

Elutic W"igj'lt Loadinl

"

"

N
~

Fig. A7.57

~.
I

0'

01

e'

,
,

~.

Table A7 .13

E'

Ilastic Welgbta of Chord embers

IIember

R:.OO:l12

R", .00312 Fig. A7.58

dL
_.070

In finding deflections this overhang elastically loaded portion 15 considered as f1xed at be .074 .0037 B a and tree at e. The bending moment at any po in t I Be -.060 to .0030 0 , on this beam equals the magnitude of the 'ler~ica1 od .070 20 .0035 C , detlection at tnat point. 20 .0028 -.~~6 d ~ .074 20 .0037 0 _.056 IlK Thus to tind the cer l ect i on of the truss er 20 .0028 0 of .012 20 .0036 E (Joint a) we tind the bending moment at pOint a -.0""8 20 .0024 of the imaginary beam or ?lg. A7.59. Hence Ilastic Weights of Web ~ .oera deflection at a ; Z!1a (calling counter:loc'K, dL '1 P ~ I apply '2 Q .<\1, apply wise pos i t i ve ) , ; ( .01922 - .00312 ) 80 + .00248 X rl at '2 70 + , joint .00333 X 60 .. . 00239 X 50 + .00468 X 40 + Jo1ut " .00234 X + .00543 X 20 - .0014-9 X 10 ; 2.13" 30 -.031 8.95 -.003406 I .00346 8.95 .080 8.95 A 8.95 00895 upward. Due to symmetry of 'truss and leading o _.063 8.95 _.00702 I 8.95 .00702 B the truss we know the slope of the elastic curve , Be .075 8.95 ~.00838 B 8.95 .00838 e _.060 8.95 _.00670 c 8.95 ,00670 'C C at the center line ot the truss Is horizontal 0 r 8.95 , o;~ 8.95 -.00781 ! C O~L81 zero. Thus to find the deflection of any potnt dO -.0605 8.905 d 8.95 .00728 0 -. 00 72 8 1 I With .076 8.95 ~. 008.50 D 8.95 .00850 0 reference to Joint f we can make use ot the o. .09' 12.0 12.0 -.00792 .00792 I 0 I !f deflection principle of the moment area method. .02. 12.0 .0020 12.0 I Thus In Fig. A7.60 the vertical ce ri ect i on of a. paint for example joint a, relative to Joint t equals the moment of all elastic loads bet-....een Table A7.l3 gives the calculations tor the nagnltude of the member elastic weights. The signs a - and t: about a. and also the jOint locations for locating the N = = elastic loads are also given. Combining algeN N N 0 0 0 0 0 0 0 braically the elastic weights tor each jOint t:ro~ 0, 0 0 " ~I ~I Table A7.13 the beam elastic loading as shown In , 'I " 9 , o. I, Fixed Freete 0 d C c B b A ar Fig. A7.57 Is obtained. Let it first be required to determtne the R", .00312 (From Span ee') vertical deflection at Joint r relative to the Fig. A7.59 truSS support pOints at e and e ' . To detern,ine the deflections of the truss l:l1a = 2.077" (student should make calculabetween the supports e a d e' It Is only necest i ons ) Previously the deflection of t with r-e sary to consider the elastic weight loading bespact to e was found to be -.0586" . Thus detween these paints. Fig. A7.58 shows the portion flection of ~ with respect to paint e ; 2.077 + of the 1Icaginary beam of Fig. A7.57 be'tNeen these .0586 ; 2.135" which checks value found above. paints. The deflection at f relative to line ee' Let It be required to f'1nd the deflection of Is equal to the bending ~oment at , for the por- joint c relative to a j rne connec t Ing joints b tion of the imaginary beam be~Neen joints e and and d. N = = e' and simply supported at these points. N " N N N N 0 0 0 0 0 0 0 0 0 Hence deflection at t ; - .00312 x 30 + 0 0 0 0 0 0 ~ ~ OJ " " .00232 X 15 ; .0586" (upward since >n.inus Is up) 'I , , i , I , ! ", Find the vertical deflection of the cantil E e d 0 C 0 b B A a
AD

I ,

=&
r
.0040 .0035

.oao

Apply at Joint
A

ac
20 20

..

..... .. ,. ..
""

-,00":

~.0020

, ,

~I

~ ~

~
~

~ ~

~ ~

~
~

~
~

- "',
~ ~
~

~I~ ,

~I

~
~

~ ~

~ ~

~ ~

~ ~ ~

~
~

!<t

Fig. A7.50

________...;,

. '- '.-:~.-:~li--~~.~;.,r:.-~~F.,~:i!!::_

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.37

For this problem we need only t~ consider the elastic loads betN8en pOints band d as

loads on a simple beam supported at band d (See Fig. A7.61) ~he deflection at C with respect to a line bd of the deflected truss
equals the bending moment at point c for the

(7) For the truss in Fig. A7.66 calculate the deflection of jOint C along the direction CEo E = 30.000,000 psi.
C --.-20,000*
S'
B

loaded beam of Fig. A7.61. hence

0c

= .004743
d

x 20 - .00239 x 10 =,07 inches

~ M N

8'
A

0
A

~
2
B

S'

= =
~

~=.004743

~
Fig. A7.61

~l ~l ~1

~ M N

20' Fig. A7.68

'7'

Fig. A7.67

Rb=.004667

A1. 18

Prob 1ems

Fig. A7.62

~
45'
C

A~

O"

(8) Far the truss in Fig. A7 .67, tind the vertical and horizontal displacement of jOints C and D. Take area of all members carrying tension as 2 sq. in. each and those carrying compression as 5 sq. in. each. E = 30,000.000 pSi. (9) For the truss in Fig. A7.68, determine the horizontal displacement of points C and B. E = 28.000,000 psi.

\.

3~'

~ to' 1>\)(): 3~

t- A

J
0

G
H
A

+20"
B

40" . Fig. A7.63

J..

~
IB Ie
2

5000 5000

::i

::'F

,000< 30"

T
I

I I ~-r--r.,--r-.,---,

1000

1000

500*

i10" T

G414~E

5060

5000

Fig. A7.64 Fig. A7.59

e
Fig. A7.70

D 4 @20 = 80"

(1) Find vertical and horizontal deflection of jOint B for the structure in Fig. A7.62. Area ot AS = 0.2 sq. in. and Be = 0.3. E =
10,000,000 psi.

(2) For the truss in Fig. A7.63, calculate the vertical deflection of jOint C. Use AE tor each member equal to 2 x 10 7 (3) For the truss of Fig. A7.64 determine the horizontal deflection of jOint E. Area of each truss member = 1 sq. in., E = 10,000,000 psi. (4) Determine the vertical deflection of joint E of the truss in Fig. A7.64 (5) Determine the deflection at joint D normal to a line jOining joint CE or the truss in Fig. A7.64.
b

(10) For the truss in Fig. A7.69, rind the vertical deflection of jOint D. Depth at tr~ss = 180". l..tIdth ot each panel is 180". The area of each tr~ss member is indIcated by the number on each bar in the figure. E = 30.000,000 pSi. Also calculate the angular rotation at bar DE. (11) For the truss in Fig. A7.70, calculate the vertical a.~d horizontal displacement of joints A and B. Assume the cross-sectional area tor members in tension as 1 sq. in. each and those in compression as 2 sq. In. E = 10,300,000 pat , (12) For the truss in Fig. A7.70 calculate the angular rotation at member AS under the given truss loading.
1000* 5000

"
c
d
~9' ,

12'
I

I
10

' T

10'

i is'1

B iii
Fig. A7.72

Fig. A7.71

e ~a'
I

---.-a'
SO,OOQt-

1.,

Fig. A7.65

(6) Calculate the vertical displacement of joint C tor the truss in Fig. A7.65 due to the load at joint B. tlembers a, b, c and h have areas of 20 sq. in. each. Members d , e , t , g and i have areas ot 2 sq. in. each. E = 30, 000, 000 psi.

(13) For the beam in Fig. A7.71 deter.nlne the deflection at points A and B using method of elastic 'N8ights. Also determine the slopes ot the elastic curve at these points. Take E = 1,000,000 pSi and I = 1296 In.~ (14) For the beam in Fig. A7.72 find the deflection at points A and E. Also the slope at the elastic curve at point C. Assume EI equals to 5.000.000 Ib in. sq.

A7.38 500Jl

DEFLECTIONS

OF

STRUCTURES
100,* 10"---.1
A ,---

t=25"

5"

, 5"

"

-j.<a' t ;r.~1

2"1

50"
Fig. A7.81

~A
lOOOf

'I

Fig. A7.B2

(23) In Fig. A7.81 find the vertical ~ove (15) Fig. A7.73 illustrates the airloads on a tlap beams ABCDE. The tlap beams is supported ment and the angular rotation of point A. Take at B and D and a horn load or 500# is applied at EI = 12,000,000. C. The beam is made tram a 1~-.049 aluminum (24) Determine the vertical deflection or alloy round tube. 1= .01659 in~; E = 10,300,0 point A for the struc~~e in Fig. A7.82. EI = psi. Compute the deflection at pOints C and E 14,000,000. and the slope of the elastic curve at point B. 100#
lOOt 100 "'3" 1"3"

F ...
100

3 .,

F91t -;100

100*

100

9" + 9 " B

j
A

~~1=t=-.~a=58~st=e~el=tU=b~e~~" ri:1 20" Y

A
100ill'

B 3"

Fig. A7.83

Fig. A7.74

Fig. A7.75

(16) For the beam or Fig. A7.74 determine the dert.ectaons at paints C and D in 'terns or EI which is constant. Also determine slopes ot the elastic curve at these same paints. (17) For the cantilever beam ot Fig. A7.75 determine the detlections and slopes or the elastic curve at points A and B. Take EI as constant. ExPress results in terms or EI.

V Fig. A7.B4

L. o r a

Front View

Side View

p." T" "1~

1001 400f

BEMA

Fig. /07.77

(18) For the loaded beam In Fig. A7.76 deterame the value at the t Ixed end moments MA and MB. EI ts constant. Also tind the detlection at points C and D in terms at EI. (19) In Fig. A7.77 determine the magnitude at the tixed end moment I1A and the simple support RB.
p

[ aij] {qj} compute the strain energy in the' truss at Fig. A7.63 (Problem 2). The ~e~ber 11\\ ~ A '11 flexibility coefficient fer a member under Flg.A7.78 Fig.A7.79 Fig.A7.80 uniform axial load is L/AE (see Fig. A7.35a). An". U = 22.4 lb. In. (20) In Fig. A7.78 EI is constant throughout. (28) Using matrix equation (23) compu~e Calculate the vertical deflection and the anguthe strain energy in the beam at Fig. A7.71. lar rotation ot point A. Note: the choice of generalized forces should (21) For the curved beam In Fig. A7.79 rind be made so as to permit computation at the the vertical deflection and the angular rotamember fleXibility coef~iclents by ~he equations tion at pOint A. Take EI as constant. of p . A7.l9. Ans. U =- 3533 lb. in. (22) For the loaded curved beam at Fig. (29) Re-solve the problem of example A7.80, determine the vertical detlection and the problem 25 tor a staowed cantilever beam whose angular rotation at the point A. Take EI as I doubles at point "2 and doubles again at "3". conatant , (HeaViest sect10n at built-in end.)

IF

(25) The cantilever beam of Fig. A7.53 is loaded normal to the plane of the paper by the two loads of lOCI each as shown. Find the deflection at point A no~l to the plane of the paper by the method of virtual work. The rectangular moment or inertia tor the t~be 1s 0.0277 in. E = 29,000,000. (26) The cantilever landing gear strut in Fig. A7.84 is SUbjected to the load or 500# in the drag direction at point A and also a torsional moment of 2000 in. lb. at A as shown. Determine the displacement at point A in the drag direction. The tube size tor portion CB is 2~ .083 and tor portion EA, 2~-.065 round ~~be. Material is steel with E = 29.000,000 psi. (27) Using the matrix equation 2U

tJA

I RI

L qi I

_ " . .; . /' : _:;. i' .:.",

:,.>.:-:~.::~ -:-:-.. :.~,~ :'::::;:<; ~ ..":

<:' ~ -,.

-- -

-<";:. ;','" _~ -::!,~. :',."

~2::'':~'' .....<:\-~. ~-: ';0,\, . ._~r

'3 :./"-:-

"'-~. ~ ,:t ._~ <-.' l: /}",-.f,~~7'f:": N" - ~

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A7.39

(30) For the truss of Fig. A7 85 determine the influence coefficient ~trix relating vertical d.eflections due to loads P1.' P 'U P~ and P~ applied as shown. Member areas are shown on ~he f~gure.
A

1.0

1.0

G : 3.85 x 106 psi

t:.051"

;
1.0

10"
2.0 1.5
10"

r, L
Fig. A7.85

-Jrn

n-l
Fig. A7.86

Answer.

[3,nJ. E
1
~ine

["

44.67 106.0 99.34 38.00

44.67 99.34 99.90

~8.00

44.67 44.67

38.0 33.0

~.~
38.0

(34) Find the influence coefficients relating deflections at points land 2 of the simply supported be~ of Fig. A7 87. Use matrix methods.

33.00

(31) For the truss ot problem (30) ceterwhich of the following two loading conditions produces the greatest deflection of point 4, (All loads in pounds},
)10.

(j)

@+ ( AE:3XI~~lbS
AE:8 x 10 6 lbs
Fig. A7.87

Condition
1

p.
1000 300 500 700

p ,

800
400

400 600

Ans.
(32) Determine the matrix of influence coefficients relating drag load (positive aft), braking torque (positive nose up) and moment in the V-S ~lane (positive right wing down) as applied to the free end of the gear strut assembly ot Problem 26.

7.10~ 12.~
Note to student: It will be highly instructive to re-work problems 33 and 34 using the alternate choice of generalized forces in the stringers from those used in your first solution. See p. A7.22 for alternate generalized forces on a stringer.

Answer.
f2S00
I

-50.7

Lo

-50.7

6.97

References for

Cr~pters

A7 J AB.

T = BOOKS ON STRUCTURAL TOlEORY

the 1eflection of the load applied to the cantilever panel of F~g. A7 86. (Assume the web does not buckle). Use matrix notation . .~s. 0 = 7.94 x lO-~ inches.
(33) F~nd

"Advanced Mechanics of ~~terialsn, F. Seely and J. O. Smith J 2nd Ed., John ~ileYJ New York. "Advanced Strength of Materials", J. P. Den Hartog J McGraW-Hill, N. Y. "Theory of tna st ic rtv", S. T1m.oShenk0 J rrccraw;

arn ,

N. Y.

A7.40

DEFLECTIONS

OF

STRUCTURES

TEXT BOOKS wITH MATRIX

APPLICATIO~~

Langefors, B., ~~alySis of ~las~lc Structur~5 "Introduction to ~he Study of Aircraft Vibration Matrix Transformation with Spec ~al ?,eg-a.r1 ~ and ~lutter", R. Scanlan and R. Rosenbaum, Sere1 Monocoque Structur~ .jour-n , of Aero. Sci. Mac Millan, New York. 19, 1952. Langefors, 3., ~trlx r.ethods ~or Recundant Structures, Journ. of Aero. Sc~ Vol. 2C, :;53. TECHNICAL PAPERS Benscoter, S. U., "The Partitioning of ~atrices !E Structural AnalySTS", .rourn. of Appl , hechs ,
Vol. 15, 1948.
Falke~~einer, H. Systematic )~alysls o~ Redundant 21astic Struct-:.rres 'Jy :-:ea:;s of ~aL:rix Ca:c'.!lus, .Jour-n , or .Aer-o, 3ci:-: 20, ::''.753.---

"Elementary ra cr-ices", R. A. Grazer, v : J. Duncan and A. R. Collar, Cambridge University Press.

L. and Lansing, ~'J A Method for Reducing the Analysis of Ca~p1ex RedUr..~tructUIes tc ~ Routine Procedure, Jo~n. of Aerc. S:~.
~ehle,

Vol. i9, 1952.

Argyris, J. and Kelsey, S., 2nergy Theorems in Structural Ahalysis, Aircraft ~~gineeri~g, cct lS54, e t , seg ,

. .

All .
;..~~ ... ~ .... -.
a';~'~ .

"-\'

Many problems involving calculation of deflections are encountered in the structural destgn of a large modern airplane such as the Douglas DC-a.

CHAPTER As

STATICALLY INDETERJ.\1INATE STRUCTURES

ALFRED F. SCHMITT
AB. 00 Introduction.

A statically indeterminate (redundant) problem Is one in which the equations of static equilibrium are not sufficient to deter.nlne the internal stress distribution. Additional relationships between displacements must be written to permit a solution. The "Theory of Slastlclty" shows that all structures are statically indeterminate when-analyzed in amut e detail. The engineer however, Is otten able to ~ke a number of assumptions and coarse approximations which render the problem determinate. In addition, auxiliary aids are available such as the Engineering
Theory of Bending

tics are individually useful in forming the bases for methods of solution. C[J There are more members in the structure than are required to support the applied roacs , If n members may be removed (cut) while leaving a stable structure the original structure is said to be "n-times redundant".
-eOROLLARY-

(MJ) and

the constant-shear-

flow rules of thumb (q = T/2A) (see Chaps. A-5, A-6 and A-!3 through A-IS). While these latter are certainly not laws of "statics", the engineer employs them often enough so that problems in which they are used to obtain stress distributions are often thought of as being

In an n-times redundant structure the magnitude of the forces in n members may be assigned arbitrarily while establiShing stresses in equilibrium with the applied loads. ThUS, in Fig. AS.I (a singly redundant structure), the internal force distribution of (a) is in equilibrium with the external loads for any and all values of X, the force in member BD.

"cetermnace" . It is frequently the case in aircraft structural arzlysis that, in view of the requirements for efficient design, one cannot obtain a determinate problem without sacrificing necessary accuracy. The Theory of Elasticity assures the existence of a sufficient number of auxiliary conditions to permit a solution in such cases. This chapter employs extensions of the methods of Chapter A-7 to effect the solution of typical redundant problems. SpeCial methods of handling particular structural configurations are shown in later chapters.
AS.O The Principle of SU'Pel"?osition.

_A 20004;'

~.
c-: c-:
D

\.00"*
<a)

-:'Jlc----'!J!' A 2000# '"' 1200"* hoo,,* 200,,*

(b)
Fig. AS.!

~12l
.707X
(e)

Nt-to:

SIngly redundant stress distribution, (a) consisting 01 a stress

in static equilibrium with the awlied loads, (b), with one

zero-resultant stress distribution, (c), superposed.

The general principle of superposition states that the resultant effect of a group of loadings or causes acting Simultaneously is equal to the algebraic sum of the effects acting separately. The principle is restricted to the condition that the resultant effect of the several loadings or causes varies as a linear function. Thus, the principle does not apply Nhen the member material is stressed above the proportional limit or when the member stresses are dependent upon member deflections or deformations, as, for example, the beam-column, a ~ember carrying bending and axial loads at the same time.
AB.! The Statically Indeterminate Problem.

Only the system (b) is actually required to eqUilibrate the external loads (corresponding to X = 0). Note that the system (c) has zero external resultant.

[[j Of all the pOSSible stress (foroe) distributions satiSfying static equilibrium the one correct solution is that one which results rn-kinematically pOSSible strains (displacements), i.e. retains continuity of the structure.
ThUS, for example, there are an infinite number of bending moment distributions satisfying static equilibrium in Fig. AS.l (d) since ~ can assume any value. Of these, only one will result in the zero deflection of the right hand beam tip necessary to maintain structural continuity witn the support at that point.

Several characteristics (and interpretations thereof) of the statically indeterminate prOblem may be pointed out. rhese character is-

AS.!

AB.2

STATICALLY INDETERMINATE STRUCTURES

the Theorem of Least Work. In words; Rthe rate of Change of -strain energy with r9S?ect to a ilxed redundant reaction 1s zero R.

--

M.

-I===:r-<:'.
M

PL/.

AS. 2. 1 Determination of Redundant Reactions by Least Work.

Fig. A8.1d Singly redundant beam with root bending moment

Me

undetermined by statics.

Example PrOblem A By way of illustration, the proble~ posed by Fig. A8.2 'NaS carried to completion. The bending moment was given by (x, YJ ~ measured tram the left ends of the tr~ee beam dlvls~Qns)
M

-eOROLLARY-

bitrarily ~hile establishing equilibrium with the external loads, relative ~ovements of the elements will result, violating continuity at n points. n zero-resultant stress (force) distributions may then be superposed to reduce the relative motions to Zero. The resulting stress distribution Is the correct one.
AB.2 The Theorem of Least Work.

It n member loads have been assigned ar-

(500 + RxJ L + (Rx-500 J Y 2

M Rx (L - g)

O<;;;<L

Then
1

U' 2

f
o

M' dx

EI

'2EI

JLit(500+Rx ) "y ' 1x

0

A theorem extremely uset~l in the solution of redundant problems may be obtained trom Castlgliano's Theorem. Consider first the problem ot redundant reactions such as in a beam over three supports (Fig. A8.2). One ot the reactions cannot be obtained by statics.

+ -1-ffYT500+RxiL + (Rx - 500 Iy]'d Y

2EI

L
+

2~I EX'(L -

g) 'ds

....El = constant

--t--500-2Rx

&

Differentiating under the

p , A'7.8)

lntegra~ 51'S!:,

(see

Fig. AS.2

oR x

au

A singly redundant beam with one reaction given an arbitrary value (Rx)'

It the unknown reaction (say tr~t on the :ar right) is given a symbol, Rx , then the re~zining reactions and the bending moments ~y be determined from statics. The strain energy U ~y then be written as a function of R x , i.e., U = r (Rx). Next form

+ _~ '1

L , fR ., \ 'X (L - 0,
o

'I

-"
'"'

This is the deflection at rtx due to R x o But this ~ust also be zero.-Slnce the support is rigid. Hence

~. a - - - - - - - - - - - - - aR x

(1)

.. (500 +
2

Eq. (I) is true for all redundant reactions occurlng at fixed su;po~ts. 3ecause it corre_ sponds to the T~the~atlcal condition for t~e minimum of a function, sq. (l) is said to state

BE.""...._ _IIliIIiiii.........

..

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.3

Completion of the computation gave Rx

t~e

=-

1500

lbs.

=-

93.8 lbs.,

au ai1=
R

X = [Ll1L+ + 2PL - l1L X] dy 3L 3L

c

negative Sign indicating that Rx was down.

Example Problem E Determine the redundant fixed end moments for the beam of Fig. A8.2(a).

o =l1L J~
Fig. AB.2a A doubly redundant beam with two reactions given
two arbitrary values.

-3~)
+ PL -

dx

~+2PL-'\JL
3L

oX(1 - 3L)dx

Solution: The redundant end moments were designated as ML and MR for the left and right beam ends respectively and were taken positive as shawn. The moment equations for the two beam portions (x trom left end, y from right) were
~+2PL-,\

+!lR

2 Iy

-dY+ 3L

'\

~ J2~.

3L

, 3L
'\

-dy

O'l1L

\-dx+ Je
3L

~ + 2PL 3L

J\.
-r-r

3L
c

dx

11='\+
3L

O<x<L

O<Y<2L

Then Evaluating the integrals and solVing sDnultaneausly gave

4 PL
9

+..l.

2EI

~ - - PL
Differentiati~g ander the integral Sign (see remarks on p. A7-8)

2 9

AB. 2. 2 Redwldant Stresses by Least Work.

The Theorem of Least Work may be applied to the problem of determining redundant member forces within a statically indeterminate structure. ThUS, in-an-fi-tDn8S redundant structure if the redundant member forces are assigned symbols X, Y, ~, - - - etc., the values which these forces ~ust assume for continuity of the structure are such tv~t the displacements associated with these forces (the discontinuities) must be zero. Hence, by an argument parallel to that used for redundant reactions, one writes,

A8.4

STATICALLY INDETE.RMINATE STRUCTURES

au ax
ay =

o
0
__ - - - - - (2)

50
+

au

':;=".l.BC

.~.

j
1

(lOOe
\:

+-2EI~e

etc.
~lth

In words, "the rate of change of strain eneT~Y respect to the redundant forces is zero".

1 +--

60
o (50,000):ld6'

Eqs , (2), like eq, (1) I are statements of the Theorem of Least ~ork. They provide n equations

2EIC D

for the n:tlmes redUndant structure. The simultaneous solution of these equations yields the desired solution of the problem. Example Problem C The can~11ever beam and cable system of Fig. A8.3(a) 1s singly redundant. Find the member loadings by use of the Least ~ork Theorem.

Obviously there was no need to ccnsider the energy in CD as its loading did not depend upon X and hence could not enter the problem. Differentiatln~ under the integral sign

au

58.3 X
~.U3

ax

so v
58.3

(1000 - .30_ x\ '_ _...";,,0,,8.:..,,,,_'-)

EI BC

1x
0

50
a

ex

(b)

-0 )" _0( 50.3

Fig. AS.3 A singly redundant structure with one member force given an arbitrary value (X).

Solution: The tensile load in the cable was treated as the redundant load and was glvzn the symbol X (Fig. A8.3(b)). The strain energies conSidered were those of flexure in portions AC, CD and BC and that of tension in the cab~e AB. ~ergies due to axial forces in the beam ~ort ions were considered negligible. The bending moment in Be (origin at B) Nas

58.3 X ... 11,032 X 6620 X EA trac S; AC AB

<

= 21.44 X 101!!

ELse
Putting
AAB 0.025 in a

!'lac

= (1000 -

5~~3

XJ

x
10.0 in"
gave

InAC, (origlnatA):
MAC

= 58.3 X Y
X

50

613 lts.
#

In eD: l""cD = 50.000

Then

The strain

ener~J

was therefore

~e
M. .',.c

[1000

vv.,,)

30

("3)J 0.1. x = 685x

50

52:3

613

.y

520 y

ANALYSIS AND DESIGN OF

P~0b13~ D A semlci~cular pin-9nded, uniforn ring is support ad and leaded as shown i~ Fig. A8.3(c). As a first ~~prcxi~tlon the ~orlzontal flcor tie is te Je assumed rigl~ aXially. Find the ~ending moment distribution in the ring. Exa~ole

FLIGHT VEHICLE STRUCTURES

AS.5

Dif:erentiatins under the integral sign

au ax
R'P

EI x

R'

J60
0

sin

Q d Q

60'
0

(1 - cos g) sin g d g

""-r
""

R' X f90

sin

II

g d g

6Qo
0

101

Fig. AS.3

.... (d)

- 2Ei'

R'P )90

sin g d g

60'

Solution: The axial lead in the floor was taken to be the redundant (since the floor was assumed rigid, this could have been thou&~t of as a redundant floor reaction from fixed supports). The loading is shown in Fig. A8.3(d). The bending moment distribution 'NaS
11

~ locOS
II

60 0

g sin g d g

Q +

~ sin
] 60'

90 '

II

Q d Q

Evaluating,

X R Sin g - PR (1 - cos g) X R Sin g

PRIZ

0(g<60'

The axial loadings were

+gP

3 (R R') AE-TI
' R) (R rr-TI:

s
S

P ccs Q
X sin Q

X sin

Therefore

The strain energy (for only half the structure) was

U =

zir E sin Q- PR
R 90'
+

2n

3P

' +R) (R EI AE
1. 50

60'

(1 - cos

Q)J'

Rdg

zir

160(X R sin g - PRIZ

Rdg

+ X

sin

Example Problem E The portal frame of Fig. A8.3(e) is three times redundant. Set up the simultaneous equations in the redundant forces. The relative ~ending stiffnesses of the segnents are given on the figure. Solution:

1 000lf

c
2.25
2.25

1
50"

D...1

RdG

Fig. AS.3e

~E

90 '

[x sin

QJlI RdG

J60

The redundant forces selected were the bendlng'moment,-the transverse shear torce and the axial force, all at point A. The four figures A8.3(f) tr~ough A8.3(1) show the bending ~ament diagrams or the structure due to applied loads and due to redundant forces

Zero strain energy in the rigid floor (AE -+ ....).

AS.6

STATICALLY INDETERMWATE. STRUCTURES

acting individually (it being easier to compute the loadings in this fashion). The complete loading ,v,as obtalr.ed Jy superposition.

.1[-50'

;0 Sln g ..

~ .. 50'1(1.

sIn !il) -5crl'll -

-'..:>0

'~~5!ilIJ(,_ . .s.... ~",

"""'" "we...

..

. I"

(l000ll'!"l. '{a' .. SO '{ ~ ~OO ":':(~C - s': 1S'

LZ5

Fig. AB.3f

50,000

Fig. AS.3g

After evaluation of the obtained were

{'1492!l 9.682!l
+ +

~nte~31s

the equations

9.682V

7.459T

1.112 x 10' 288.3 x 10 3

763.1V

484. a!'

-7.45911 - 484.0V + 614.9T

= 111.15

x 10 3

50V

Q.6.
1
s

AB.3 Redundant Problems by The Methods of DummyUnit Loads.

T s'
~

Fig. A8.3h

lT

lOOT

Fl . AB.31

The composite bending moments as funct10ns ot 5, Q and 5' were

I1 AB=Il+Vs

MBc

50,000 sin 9 + 11 + 50 V + 50V sin 9

- SOT (1 - cos g)

MCD = 1000
Then since

s' + Il - Vs' + 50 V-100 T

',.Jhile the Theorem of Least ',.jork may be nace the basis of red~dant problem analYSiS, its direct application by the calculus, as in Art. A8-2.1 and A8-2.2, is often impractical. ~or t~e majority of problems the work is facilitated if carried out by the techniques of the Method of Dummy-Unit Loads. The follOWing derivation is for a doubly redundant tMISS struct~e. The extension te a more general n-times redundant structure, in which ether loadings in addition to axial (flexure, t crst c.i and Shear) are present, is indicated later. Consi~er the doubly redundant truss of Fig. A8.4(a). It may be made statically dete~inate by "cutting" two members such as the diagonals indicated. Application of the external loads to this determinate ("cut") structure gives a load distribution, ~S", computed by satisfying static equilibrium. At this time discontinuities appear at the cuts "xn and "yn due to the strains developed.

l J Il ' dS d U = :2 ~an

aR=;j1i

au

au

au
aT

0,
x

, IP

,P,
y

..

one haS,

au

so

aM =

III

2.50

vsl

ds

S loads

) 0

Fig. AS.4a

loads Fig. AS.4b

ux

Uy loads
Fig. AB.4c

.j .

["-50.000 sin;"!"l" 5OV(1 .. sIn ;1

1.50

,so J .(10009'

.. 11 - 'Is' .. SO '( - 100 ":') d '

Q.=

To compute these and subsequent displacements the Method of Dummy-Unit Loads may be used (Art A7-7). For this purpose Virtual loads are placed alternately at the x and y cuts as in Figs. A8.4(b) and (c). From the dummy-unit load equations

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.7

z S~t}
Z SuyL

(3)

For continuity these net relat~ve displacements must be zero. Equating the above expreSSions each to zero, and rearranging, gives the simultaneous equations
X

AE

llL UX + Y Z

u and u are the unit-redundant stress disx y tributions as indicated in Figs. A8-4b, c. The subscript wow indicates these relative dlsplace~ents occur in the determinate (Wcut W) structure with the Wor i gl na l Wstress distribution. It is now desired to close up the discontinuities by application ot redundant loads X and Y to the x and y cuts, respectively, as In Fig. A8-4(d). Load X causes a stress distribution XUx and, likewise, Y causes a distribution YUy. The relative displacement at cut x due to redundant load X is given Fig. AB.4d by (6 is read wdisplacexx ment at x due to X W ).

AE

uxuyL AE
AE

~ - z "",C AE
AE

u 'L L X Z UyUx + Y Z -y-= _ Z SuyL

--- - -

(4)

ICE

Eqs. (4) are two simultaneous equations in the two unknowns X and Y. Upon solution for X and Y the true stress distribution may be computed as
(5)

For a structure which is only singly redundant, eqs. (4) and (5) are applied by setting Y = a giVing

xZ
or, Simply, SUxL
Z

ux aL = _ Z SuxL AE AE

and at cut.y by (read 0yx as wdisplacement at y due to X W ) .

Ai:
'L
- - - - - - - - - - - - - - - (4a)

Z ...L-

AE

= Z XUx uyL AE
Similarly the load Y causes displacements at the cuts y and x given respectively by
e
YY

and,
- - - - - - - - - - - - (5a)

AB.4 Example Problems Trusses With Single Redundancy.

~ Z YU y ' uyL
AE

=Y

Z uyaL
AE

Example Problem #1

and
Z YU1 llX L = Y Z uyuxL

AE

AE

Now the net relative displacement at each cut under the slmultar.eous action of the three stress systems 5, X-u and Y.u y is
x

-h~l~~ "~,, hoo"" \000'

Fig. AB.5 Fig. AB.6 Fig. AB.7

T~a O'~-g~g

30 -

po.o*

1000*

-.707

+ 6 + 6 = Z SUxL + X Z uxuxL xo xx xy AE AE
+ Y Z u y uxL

AE

and
Z SuyL + X Z uxuytAE

AE

+ Y Z

uyuyL AE

Fig. A8.S shows a single bay pin connected truss. The truss is statically deter.nir~te With respect to exter~l reactions, but staticaily indeterminate with respect to internal member loads, since at any joint there are 3 unknowns with only ~NO equations or statics available for a concurrent force system. The truss is therefore redundant to the first degree. The general procedure for solution is to make the truss statically deter~inate by cutting one or the ~embers;on Fig. A8.6, member bc has been selected as the redundant ~ember, and it 1s cut as shown. The member stresses S for the truss ot

A8 8

STATICALLY INDETERMINATE STRUCTURES

A

Fig. AB.6 are then deter.n1ned, the results being recorded on the r:le::nbers and also entered in Table AS.L In Fig. AS. 7, a unit If tensile dummy load has been applied at the cut section of the r-edundant member be, and the loads in all the m.embers due to this unit load are calculated The results are recorded on the figure and also in Table AS.l under the head of u stresses. The solution for the redundant load X in the redundant aemoer- be is given at the bottom of Table AS.!. The true load in any member equals the S stress plus X times its u stress.
bble U.1

~/ ~/
V'...;

Iiii'

0 1,000

0
Fig. AS. 10 u loads
Ca:c~l~tiJn

Fig. AS.9
S loads

Example Pr'cb Lem s I-A; Deflecti0n a Redundant Truss

1:1

""r

L
30 30 30 30 42.4 42.4

1 1 1 1 2 1.'

s
_1000
0 0 0
14140

u
_.107 707 _.707 _.707 1.0 1.0

';t
0 21210 0 0 0 40000

T
15

u 2L

u.
.d

d.

"

15

True Stress s + Xu

.60'
_559

3"

395 3

21.2

28.3

'55

Calculations of the deflections under load of a redundant structure are made by application of the methods of Chapter A-7. Since, however, there are certain piUalls as regards symbols and also some important special techniques, the following examples. are given at this time. The extension of the method to more complex structures is immediate and no further work on de!lections of redundant structures is given in this chapter (excepting in the case of the matrix methods of Arts. AS. 10 et. eeq.}.

X true X

1~c1

111 redundant _.tier be

."

hi
A. ~

t"r

-aUlO 109.S

_5~9.

Find the horizontal ~ovement of :Joint ftd" of Example Problem 1# 1 under the action ef tne load applied there. Solution: The equation used to find the deflection is Sq. (18) of Chapter A-7. ',iritten for application tv t:-uss deflections it is (see Example Problem 13, p. A7.l1)

Example Pr'oo.Iem t 2 Fig. AS.S shows a Singly redundant 3-member frame. Ftnd the member loadings. Member areas are shown on' the r rgure ,

z~ --AE

- - - - - - - - - - - -(A)

T~OV ~~. -'

100" J.2 .2

.4

0 ho"",

Fig. A8.8

Solution: Member ex: was selected as the redundant and. was cut in figuring the S-loads, as in Fig. AB.9. Fig. AS.IO shows the u-load calculation. The table completes the calcu-

Now for a deflecticn calculation the symbols '8' and "u" ;nust be carefully reinterpreted from their ~eanings in the r-edundant stress catcutat i cn.' For a deflection calculat1Cml the symbols of eq. (A) , above, :nean: "s-Loacs " are the true loads of the redundant structure due to application of the real external leading; "u-Loads " are tae loads due to a dummy-untt (virtual) load applied at the external ;Joint where the deflection is desired and in the directi on of the desired deflection.
Thus , the '8 loads" for use in ;;q. (A) are the true stresses (the sclution] of Example Problem It L

iaercn,
Mem.
AO
BO

S
0

""A
0

SuL

u 2L

""A
1059

True Load
=

s ... X

141.4 0.2

1.224

335.5 625.6 274.1

100.0 0.2 1000 -1. 366 -6. 83x10 5 200.0 0.4

933
500

CO

1. 000

Z Z 8uL
X

-6. 83xl0 5 2492

The "u-loads" represent add1 t i onaj infar:na tion which would, in general, appear to necessl tate another redundant stress calculation. As will be seen, such is ror-cunate Iv not tr.e case. In the present pr'obIem the dummy-urn t load is applied ldentically as is the lOOOJll real load and hence the u-loads ar-e simply equal to the "S-loads" properly scaled .jmvn. The following table completes the calculation.

--

A
ZuliL

=274.1

1':)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8

Mem. ab
bd

S'
3.5 -805 3'5 3.5 -55' 855

u"
.395 -.605 .395 .395 -.559 .855

SuLI A

Mem.
30 30 30 30 42.4 42.4 1 1 1 1
2

SuL
L
A

S'
335.5 625.6 274.1

4,680
AO

A
3. 355x10 5 -3. 128xI0 5

141. 4

.2 .2

10,980

BO
4,680 4,680 6,620 20,660 52.300

100
200

-1

de ca cb ad

co

.4

o
O.227xl0 5

1.5

:. -5

= 22,700
E

=52,300 E

true loads from Exa'Ilple Problem #2 By way of Qemonatration another set of u-loads, called u l , were found for this same problem, this time by cutting aember OA. The corresponding calculations follow:

identical ','lith the "true str-ess" of Table AS.l

si:lply l/lOOOth of the "S-loads" sfnce the dummy-unit load is applied exactly as is the 10001 real load. Example Problem *2-A. Find the horizontal deflection of point 0 of Example Problem 12 under application of the vertical 1000* load shown in Fig. AS.S. Solution; To compute the deflection use

u' 0 .578 -1.155.

rt:
0 1. 808x10 5 -1. 583xlO 5 O.22Sx10 5

Su'L

335.5 625.6 274.1

The results are identical (allowing for round-of! errors). Again the symbols "S" and "u ll are to be reinter?reted for a deflection calculation as explained above in Example Problem II-A. The "Sc-Lcada" are now the "true Loads" coraout ed in Example Problem. #2, above. The lI u_l oadS" are loads due to placir~ a dummy-unit load acting horizontally on the structrre at pOint O. Since this load acts on a redundant structure it would appear that another redundant stress calculation is required. However, this Is not necessary. Theorem; For the u-10ads in a de~lection calculation any-s6t of stresses~loads) i~ static equllT'5ri'i:iinNith the dum.lily-unit load may be used , even :~rom the Simplest of "cut;" S"trueture5:'" - - --- - - -This theorem says that to get the wu_loaasl for this deflection calculation we may " cut " any one of the ~hree ~embers and get a satisfactory set of u-loads by Simple statics I Berore proving the theorem we complete the calculation in tabular forn as shown. The lI u _I oa ds " were obtained by cu~ting ~ember OC and applying a unit load hor~zontallY at O.
PrOOf of Theorem ==

To prove the theorem above we return to the virtual work prinCiple and the argument from which the dummy-unit loads deflection equation, Eq. (18) of Chapter A-7, was derived (refer to p. A7.l0). It will be remembered that the deflection was shown to be equal to the work done by the internal virtual loads (u-loads) moving through the distortions (~) due to the real loads, r . e., 6 Z u e, The internal Virtual loads are those loads due to a unit load acting at the paint of desired deflection.

Now for the statically indeterminate structure these internal 7ir~ual loads (u-loads are, in general, indeterminate since the dummyunit load is applied at an external point ot the structure. However, we recall that, i-any stress df s'tr-fbut f o.t in static equilibrium with the lI a pp l i ed load" (for the ~a::nent now we are thinking of the dummy-unit load as the "applied load") differs from the correct (true) distribution only by a stress distribution having zero external resultant
(~

.'18.1).

ii - a zero-resultant stress distribution moving through a set of displacements does zero work
4

AS.lO

STATICALLY INDETERMINATE STRUCTURES

Mathematically expressed these points 2re;

i - ~RUE

= U STATIC

~=o

where UsTATIC 1s a u-load d1st~ibut1cn cbtained from statics in a simple "cut" structure under the action of the exterr~lly applied d~~y-unit load and )~=o is the zero-resultant u-load system which must be superpozed to give the true u-load distribution
11Z <I x "R=o = a

Fig. AS.12 Ux loads

Fig. AS. 13 1ly loads

It follows, therefore, that any set of u-laads in static equilibrium with the externally applied dummy-unit load will do the same amount of virtual work when the structure undergoes its distortion as would a "true" set of u-loads computed by an indeterminate stress calculation. That is,

We note here the rule by which the degree of redundancy of a planar pmjninted truss can be deter mined. For a truss of m members with p jointa, the truss is n times redundant where n "" m - (2p - 3). For a spatial truss (3 dimensionai truss) n "" m - (3p - 6).

In the present problem n = 11 - (12 - 3) = 2.

~~e calculation is car~ied out in tabular form 1n Table A8.2. The member d~~ensions ~re given in the table.

,
Member t,

"
-

"y

Z <I x "sTATIC + Z <I x "R=O Z <I x "sTATIC

AB
Be

30.5
30.
2~

750 0

.. ..
0

A '";" I -..-- ~
S\I,,1..
0 0 64,000 1250,000 21.6 0 0

..jL
A
0

\I""I..! 1'rue Load A S"X\I".Yu.,

0-27,000
0

0 0 0 0 0

1680

CD

40.5

-'000
U50 0

0 0 0

eo

50 .25

..,
0

i 43.2
I

0 0 0

51.2

I I

.50
255

Q.E.D.
AB. 5 Trusses With Double Redundancy

C<
BE
ED

50 .s
40.1

, ,

o
0

1200 100 1

are
-932 -1015 -190

I
I

30.5 -750 0

Trusses with double redundancy are handled directly by Eqs. (4). By way of illustration, the structure ot Fig. A8.4, from which Eqs. (4) were derived, will be solved tor a loading P:l. 2000,* and P a = 1000#. Choices ot redundants were ~de identical with those at Fig. AS.4a. Figs. AS.ll, AS.l2.and A8.l3 show the S, u and u load calculations r-eepec t i ve Ly, x y

.. ..
AE

B'

.. , .... .... ,W: -..'. -..

0

-300( -.8 -.6,137,000'137,000 36.5 3(;.6 36.6

0 0

27,000 0 0 0 0

0 62.
~

21. 6 1 0 0 0 0 0 0

0375.000 0 77, ZOO 0

~62,

>0,

30.7

-,00<
0

0 0 0

-1550 "'0 -2070 1240

I
!

15.41 105.01

I
!

40.25

:tOO,478, 000 341. 11452_ 6 36.6

SUbstituting from the table into Eqs. (4) gives (cammon fac~or of E diVided out)

I
{
Fig. AB.ll S loads

X Z ~ + Y Z uxUyL
A A

SUxL - - Z! r

X Z Ux~'{L + y Z UrL

= - z S\~t
562,000 478,000

341.1 X + 36.6 Y
+

36.6 X

452.6 Y

So Ivtng ,

X;::

1550*

Y--932H Finally (see Table A8.Z)

True Load

=S

XUx

YUy.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STR

A8.5 Trusses With Double Redundancy. cont'd.

CTURES

A8.11

Example Problem 3 Fig. A8.l4 shows a structure composed at tour co-planar members supporting a 2000* load. With only two equations of statics available for the concurrent force system the structure, relative to loads in the members, is redundant to the second degree. Solution: Fig. AS.lS shows the assumed statically determinate structure; the two members CE and DE were taken as the redundants and were cut at poin~s x and y as shown. The member stresses tor this structure and loading are recorded on the members. Figs. AS.16 and AS.17 give the U and U member stresses due to unit (1*) y x tensile loads applied at the cut faces x and y. Table A8.3 gives the complete calculations for eqs , (4) and (5). The load in the redundant member CE was designated X and that in DE as Y.

Solving the two equations for X and Y, one obtains X = 521* and '[ = 416*. The true load in any member = S + XU + Yu which gave the values y x in the last :olumn of the table.
AS.6 Trusses With Multiple Redundancy.

By induction; eqs , (4) may be extended for application to trusses which are three or more times redundant. Thus for a triple redundancy,

U_ 'L UxuyL uxUsL _ + Y X Z _x z~+ bZ~ AE \ X Z

- Z

SUxL

AE
SUyL AE

uxuyL
AE

+ Y Z

U 'L .:L... + b Z Uyu.,L AE ~ U UsL

- Z

-(6 )

X Z
and

uxuoeL

AE

+YZ

+gZ

uoe aL

AE -

- Z SUsL AE
----(7)

atter solving for X, Y, b, x

+ Yu

~[
E 1200041 Fig. A8.14

r 4' ..,..' ,2' I"-

True Stresses = S + Xu

+ bU

~/y
+200041 Fig. AS.I5

AS.7 Redundant Structures With Members Subjected to Loadlngs in Addition to Axial Forces.

::;qS. (6) are extended readily to cover problems in which flexural, torsional. and shear loadings occur. ThUS, for a three times redundant structure
Xa + Ya +;;a xx xy XS

~ ~,~:./
-~V
ux loads

""."..-::3 "T'

j/
1*

- - - - - - - - -(8) xa sx
+ Ya +:!a .y

Uy loads

Fig. AS. 16

Fig. AS. 17

= - so

X Z -A- + Y Z - A - = - Z -A-

Ux aL

uxuyL

SUxL

where
a

Substituting values from table

2446 X + 2350 Y = 2,253,000.

xx

= Z ux'L +
AE

f :n~dX + J tx'dX E1 OJ
:d Y

X ~ uxuyL + UyaL _ SUyL "A Y Z -A- - - Z -A-

fP;
+

Substituting:
2350 X + 3039 '[

"0'
+

=a

yx

uxUyL =Z --+ AE

f mxmydX "SI

= 2,488,000.

Jtx~
Z _Y_- + AE
U

TABLE A8.3

If qXq~~dY
:ny'dX E1
+

Member

~I

~I

'x

'y

IsuxL x 10 XI0
--.

i_

i7S:! 402.5

~
2150
Q

uxXyL True Load .. S.Xu2:.Yuy

402
1948 0 0

'L

AE BE CE

.286.51
.1 72 .280.5 .393.7

e o
0

.806 1. 1541

o
-2.2S:!

c
2.488

280.5 576.5

2000,.1. 564 -1. 7291

1. 00 0

..,
I
521

.00

- - - - - etc.

c
1. 00

c
0

o
0

D'

0
2446

312.5 3039

ue

iTOtal 1:

-2.25:!

-2.488

2350

and

".~'''~-.c~:~~;;:?":!.,_
,..',..:;)~~~

A8.12

STATICALLY lNDETERML'lATE STRUCTURES

5xO

~ SUXL --+ AE
+

~xOX 1

JTtxox
0"

Solution: The shear flow in the 2heet panels ~as chosen as redundant. Secau2e 0: s~~~et~y :~e problem was only singly red~da~t. ?1g. AS.IS shows the Ux and qx loadings jue to the redandant shear flow X = 1. The real loadir.g in the dete~inate structure consisted of a ccnstant load P in the central stringer alone. The equation solved was (ref. eqs. (3)).

-"-.-

jj

qqxdxdy Gt

Dye

~ --ySu L +
AE

J~ OX -y--+ E1
q<l.r

J OJ

T\ox

ax
Gt

(J

Ux

...... t..

~"ax
ax

If 0-x' ncct '\ =

Y

Gt

,)

dy

+fJ

_ (J SU~E
where

jJ

OXd Y)

Gt

etc. and where 5, M, T, q are the real loads in the determinate structure; U x ' rnx' t x ' ~ are the unit (virt~l) loadings due to a unit load at cut x;

S = P = constant, in central stringer

REAL LOADS

S - o in s ice s'tr-fnger-s
q - 0

Ux - L - x in s1de stringers
VIRTUAL LOADS _ 2(x-L) in

cent ra 1 s 'tr-: r ...nger-s

s,

uy' my, t y ' etc.

are for a unit load at cut

{ ~ - 1.0
When evaluated, (note that the double integrals Simply reduce to a constant times/the panel area)
X

The redundant force(s) need not be an axial force but may be a moment, torque etc. Atter solution for the redundants, True Axial Forces =
S+ X u

C: ~; + 2itL)
.E. ( 2L
1 +

= -

(-P1E')

+ y

+;; u.

x
- - -- - - - - (9)

1 ME )

GtLa

True Bending Moments -

Therefore the true stresses were

1 +

etc. Example Problem 4 The symmetric sheet stringer panel of Fig. AS.IS is to be analyzed for distribution ot load P between stringers. As a tirst approximation, assume constant shear flow in the sheet panels. All stringers have the same area.
////
//

= LX = 2~ --~,"~ln side stri~gers

1 + AEb 1

L'Gt

A
const

1
...
'!-1o-:-b-t:.lLb-~..l;l
p L

x.

l1l

tj

"Clamped" I- L

.j..

-'h

Fig. A8.19

Fig.A8.I8

qx loadings
Fig. A8.19

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AS. 13

Solution: !ha
bendin~

momants at

~cint

C in

~ember

x (L/'Z ... '(AE

621

L')

+ Y \v~ +

(L

(l + \/'2iV\ 3EJ ")

':30 and at point 3 in member- .':'.3 were selected

as radundants yielding (when cut) the pinjOinted deter.ninate struct~e of Fig. AS.20. ~te virtual loadings were as shown in Figs. A8.21 anc. A8.22.
A

=-

PL \" ;'.2 -31

(2L/[ L')
'NaS ass~~ed

For a specific case it

that

~ :: 100
L

g.

giving

'+

'<
B

.3716 X ... 1526 Y 09011 PL

X=1,tcu~-1!L

Y=1" *

{ .1526 X + .8121 y. .3616 PL

X .

~L
Ux> m x loadings Fig. AS. 21

P
Uy, my loadings
Fig. AS. 22

.0645 PL

{ Y .456

PL

Then as usual
~

Note that in Fig. A8.22 the unit redundant loading 'NaS applied as a self-equilibrating set af unit couples. The real and virtual loadings were as follows: (member portion 3D J haVing no virtual loadings J was omitted. It could not enter the calculation.) I uy Member S M Ux !lly "'x
AB
2P$ -2P

True Stresses :: S + X U ... Y U

x
+

rrue Moments
A8.8 Initial Stresses.

= !1

+ Xrn.

Ym y

0 Y2/L

BO

?y

0 -IlL/" 6/Lv'1l" y-L 0 IlL L - YI L

The equations which were solved were (Ref.

eq , (8)).

In a red'.lndant structure In~tlal stresses are developed if J upon assemblYJ certain members must be forced into place because of lack of fit. In some situations intentional misfits are employed to obtain more favorable stress distributlons under load (~prestressing"). "If, in Fig. A8.4(a), the redundant member with the "x cut~ was initially oversize (too long) an amount 0Xi (an overSize J corresponding to a distortion in the positive X direction, is a pOSitive 0Xi)J the modified condition tor continuity at the x cut would be (compare with the equations just preceding eqs. (4)).

Similarly if the Y redundant member were too long

10
After evaluation of the integrals and ~lylng t~~ough by L~ these become
~ulti

6y1

6yx + 6YY = 0

Then using the previous notations J the appropriate equations for the redundant forces ~e
X
_x _ _ ... f' L AE y ~ Ux,\L

rx:. U.
L

-AE :E

= -

~ SUxL _

AE

Xi

1
- -(10)

r'L

=-

fUA~L

- Y1

= _ PL ((2 ... 4v:1"1L _

,-, Ai:

621 )

~\

The ~8 loads" of eq. (10) are present because of applied ext3rnal loads. These mayor may not be zero depending upon the problem.

AS. 14

STATICALLY INDETERMINATE STRUCTURES

Example Problem 6 If in example problem 3 ~e~ber CE was 0.01 inches too short before assembly, determine the stress distribution after assembly and load application. Solution: Data obtained from the ~revious problem was substituted into eqs. (10) along with (negative because
~too short~)

misali~ents

If, as is sometLmes the case, the number of exceeds the number of redundanCies, or it the misalignm~nt does not coincide with the redundant cut chosen but occurs elsewhere, one ~y use the v1rtual work pri~ciple to compute the effect of these misalignments on the redundant cuts proper. ThUS, referring to the ~virtual work~ derivation of the Dummy-Unit load equations, {Chap. A7) one has

where 0 to give 2446 X + 2350 Y { 2350 X + 3039 Y With

E

determinate structure at the X redundancy-cut due to initial imperfections (equivalent to initial strains) 6 throughout the structure.

x"

is the initial misalignment in the

= 2.253

X 10 8 + .0lE X 10 8

= 2.488

29 X 10 8 the redundant forces were

as before, 1& the unit loading due to a x virtual load at cut x.Eq. (11) and similar expressions for the Y, ~, etc. cuts may be inserted in eqs , (10).

X = 985 Ibs.
Y =

57 Lbs ,

Then, as usual,
True stresses
~ple

Example Problem 8 Referring back to example problems numbers 3 and 6, assume that member BE is .025" too long. Determine the initial stresses if the other member-s are of proper length and no external load 1s applied. Solution:

Problem 7 Assume that in the structure of example problem 5 an angular misalignment occurred between members AB and CEO at Joint B such that the end at member AB had to be rotated 2.7. _ clockwise to fit upon assembly. Determine the moments developed Without external loads applied. Solution: The initial imperfection was 0Yl - 2.7/57 3 = - .0471 radians The Sign was determined by noting that the original misalignment was in the negative direction of the redundant couple Y. The equations used from the previous problem were (noting the equations there had been multiplied by L8 x EI/L 3 ~ EI!L). .3716 X + .1526 Y ::: 0 { .1526 X + .8121 Y = .0471 E SolVing, X = - .0258 EI/L. Y = .0630 EI/L True tnitia1 stresses and moments were determined as usual.

To employ the soone equations as those of example prr~lem 3, the initial imperfections occurring at the same x and y cuts used there were computed, in this case due to the initial elongation at BE. Thus
0Xi = Z UxJl i = (-1.564)( .025') = -.0391' 0Yi = Z UyJl j = (-1.729)(.025) = -.0432'

Then, use of those previQusly computed coefficients in eqs. (10) gave, 2446 X + 2350 Y = .0391 E 2350 X
+

3039 Y = .0432 E

With E = 29 x 10' pSi

X = 263 Ibs.

Y = 209 Ibs.

Finally, True Initial Stresses = S +

AB. 9 Thermal Stresses. XU + X

Yu

Stresses induced in redundant structures by thermal strains may be computed by application at methods presented above. The problem may be approached from the point ot view ot computing the

_ _ _ _ _ _ _ _ _ _IIlII.IIIl!II .. 3!!!l1"fiW

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8.1S

relative ~otlons at the cuts of the deter.nlnate Example Problem 9 structure caused by the thermal strains and then The end upright of the truss of ~ig. A8.23 restoring continuity by applying redundant ~em is heated to the temperature distribution shown. ber forces to the cuts. Determine the stresses and reactions developed. Specifically, consider a doubly redundant truss such as that of Fig. A8.4(a). After making cuts ~x" and "y" to render the structure B;:rr determinate, the application ot the temperature distribution IS visualized. Relative displace30" x ments occur at the cuts,dencted by exT and 0yTo --l , These displacements may be computed by the 40" C (permits Y Dummy-Unit Load method as shown in Art. A7.8 of Chap. A? After this calculation Is accompsliding Fig. AS. 23 Fig. A8.24 vertically) lished, the problem proceeds as for initial strains, Art. AS.S. Thus, the continuity condition at the cut gives (compare with the equa- Solution: tions 1mmediately preceding eqs , (4) and assume tor simpliCity that the external loads are The structure was made determinate by cuts x and y as in Fig. A8.24. The unit loadings are absent, making axe = aye = 0) shown in Fig. A8.25.

~ r---l
.6

ET(~;~

0xT + On + 0xy = 0 }
0yT + 0

yx

+ 0

yy

In a truss, the thermal strains produce relative displacements at the cuts given by the ~vlrtual work~ derivation of the Dummy-Unit Load equations (ret. Chap. A7, Arts. A7.7 and A7.8) as
exT = I
0yT =
U

" ,tSJ
vx loading

1#.6

~'~
I,.
Fig. A8.25

~o
Uy loading

a T dx }

- - - - - - - - - - - (12)
U: a Tdx

The thermal coefficient a was assumed constant. The calculation was set up in tabUlar form.

where a is the material thermal coefficient of expansion, T is the temperature above the ambient temperature and U and U are the unit x y load distributions due to virtual loads at the x and y cuts, respectively. The sums in eqs. (12) are written as integrals rather than finite sums to allow for possible variation in a and T, along the members as well as from ~ember to member. Then the final equations for the rial stresses in a doubly redundant truss become
X

, ,.t
lIll

'.11

sol

11.010 I'.UI

,._.Ill"

i UI 1.10.10-"- . 10-'1

.,,,r

".

Substituting into eqs. (13)

5 . 008 X - .90 Y {_ .90 X + 1.50 Y

r 'L rUL .!:....!....

x

+ y

uXuyL
AE

AZ

--=

+ Y

~U'L .!Ai;;

AE

-JU -f
x
uy

=9 a
0

T X lOll

a Tdx
- (13)

So Lvtng ,

a Tdx

= 2.01 a T x 10

II

Equations (13) may, of course, be extended tor application to structures other than ~usses. The expressions appropriate to other loadings have been developed previously (eqs. (8) et seq. in this chapter and other equations in Art. A7.3).

Y = 1.21 a T x 10

1I

True stresses are given in Table AB.4. Problem 10 upper surface of the built-in beam of Fig. A8.26 1s heated to a uniform ta~perature T.
Exa~ple

T~e

-:

AS.16

STATICALLY INDETERMINATE STRUCTURES

Through the depth of the beam the temperature varies linearly to normal (T = 0) at the lower surface. Determine the end moments developed, neglecting axial constraint and influence of axial forces.

Then the equations corresponding to eqs. (13) were written (see also eqs. (8)).

~~
EI constant Fig. AS.26a

r-

X (t:~dS J\:dS) y(J u<t Jmx;{s)

+ + +

Virtual Loading Fig. AS. 2Gb

Solution: The problem was only singly redundant because of symmetry and was made det~inate by cutting the end bending restraints. Application of unit couples (Fig. A8.26b) gave m = 1 = canst. Then (see Example Problem 24, Art. A7.8) the thermal deflection at the ~cut~ was

Evaluated, the equations were The redundant moment; equation was (by analogy to
eq, 13)

....!-x_...B.... + EI .81 Y
-R X EI

'"' O.~ = +

: h

!!lx'dx

--=-0xT EI

1 (1 ,. 2 AE" R') EI Y + 0 " 0

=-

aRT

"Therefore
X
L "IT =-

1 +-'R ' ) ;; OX + Oy + ( -=)~ E1

=0

TaL/h

g
X = - TaEI
-h-

Note that from the last of these equations as it must because of the symmetry of ~~e ring. SolVing the first two equations

= 0,

The redundant moment compresses the upper fibers as was to be expected. Example Problem 11 Complete the problem begun in Example Problem 24 Art. A7.8, viz, that of computing the thermal stresses in a closed ring whose inner surface is uniformly heated to a temperature T above the outSide. Solution:
The ring was made determinate by cutting at the top as in Fig. A7.30(b). The unit loadings ~~d thermal deflections were dete~lned in the referenced example. The results of deflection calculations made preViously were

X = aT:I
y

= 0

A non-zero value of Y would ~roduce a varying bending ~oment which cannot Je because cf symmetry. Hence this result too, is rational.
AB. iO Redundant Problem Stress Calculations by Matrix Methods.

Orr = 2n

RaT

In the following section the Indete~inate structural problem 1s formulated in ~~rix notation. The reader is assumed to be f~iliar with the matriX applications of Art. A7.9 and the elements of matrix notation and arithmetic (see Appendix). The stress distribution of the structure is specified by a set of internal generalized ~orces, Ql' qj'(ref. Art. A7.9). Unlike the cas~
the determinate structure, these qi' qj cannot be
.. In the case of indeterminate structures, wherein some of the

0:rr = -2n R .aaT -h-

support reactions may also be redundant, these reactions also are denoted by qs. (see Example Problem 13a).

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8.17

related ianeaa at etv to the extarna I loads by the equations of stat~cs. Thus a certain subset of the qt' qj are r3dundants and are denoted by
~,

SYMBOLS - continued - the temperature (above nOrTAl) at points i , j. - the ~ember thermal distortions associated with qi J qj'

Q.s' (r , s different numerical suoscr i ots ) .

finally, the redundant forces, ~, _ qs , are computed (by satisfying continuity) the true values of all t~e qi' qj may be found by statics.
~nen,

In the notation of Arts. A8.3 et seq, the final true values of the stresses were expressed as {eq , (5) Art. A8.3)
(5)

8maOLB
G. , qJ t

'1,.'
P:nJ
sr

qs

Pn

- internal generalized forces acting on structural elements and react~ons at support points. - redundant generalized ~orces and redundant reactions - applied external loads terminate stru~turejunder application of a unit external load
P,,=l(Pn=l)

In the notation to be employed hers, this equation is restated as

(14)

- the value of q , (q ) in the de"Im ( =gJn)

Here: [gim] (=[gjn])' is the matrix of unit-load stress distributions in the determinate ("cut!t) structure found by the application of unit (Virtual) ~oads at the external loading points. The product

gtr (=gjs) - the value of qt (qj) in ~he determinate structure due to application of a unit redundant force

'I,.
G

=1

(qs

= 1)

redundant structure due to application at load P = 1 (p = 1) " n G~n (=G ) - the true value of ~ (qs) for a sn unit value of applied load P = 1 (p = 1).

1"

(=0

In

) - the true value of ql (qj) in the

[glm] {Pm} then 6i'TeS

the real loads

1n the determinate structure, corresponding to the "S!t loads of eq, (5).

[g1rJ~[gjsJ) is

the za'tr-tx of unt t-d.oad

ij

- member fleXibility coefficient: deflection at point i for a unit force, q1 = 1 (see Arts. A7.9, 10)'

amn

" coefficient for the de- influence terminate structure: displacement at exterr~l loading point ~ for a unit applied load, P = 1.
n
(a = a ) mn nm

stress distributions found by application ot unit (virtual) forces at the red1.Uldant cuts in the determinate (!tcut Tl ) structure. Hence this 1s the matrix of ux ' u y ' etc. loads. The {~} of course, correspond to X, Y, etc. Note that the [gim] and [gtr] matrices are load distributions computed and arranged in much the sarr,e fashion as was [Gim] of Art. A7.9.
'The

a~n (=u

sm) -

infl~ence coefficient for t~e determinate structure: displacement at redundant cut r (s) fer a unit ac-n -- i ec load, Pn = 1 (P = I),
~

(QrI:t
a~s

= a nr)

small letter It g " 1s used to indicate load distribution in the !tout" structure. By way of illustration, the r ina l result for Example ?roblem 3, Ayt. AS.5 is expressed below. FIRST, in the r cru of eq , (5): True Stresses = S + X

coefficient fer the deer-aune t e (vcut") structure: displacement at red~dant cut r for a ~~it ~edundant force q = 1.
l~fluence

(u,)

+ Y (U )

(a

r-s

- a

sr-

AE o + 521 (.806 ) + 416 (l.154) BE = 2000 + 821 (-l.564 ) + 416 (-l.729)

co
0_

A:nn

influence coefficient for complete Yedundant ctr~ct~e: ~eflect~on at externa: lO~dl~g ~oint n for a unt t applied load, ':l = L

o+

521 (l.00) + 416 (0 )

DE

0 + 521 (0 ) + 416 (l.00)

(.~

:::''1

= Po. :',;::J. )

Note that within each of the sets of subscript symbols (I, j), (r, s}, (m, n) the symbols may be used interchangeably.

AS. IS

STATICALLY INDETERMINATE STRUCTURES

SECOND,

in the form of eq , (14):

of external-point ir.fluence 80efficients in the ~cut~ struct~e: deflection at point n per ~~it load at point n. - the :r..a -n-tx of in~luence coeff:c~ents ~el~ting rela~iye displacements at tne "cut s" :0 external loads: dlsplaceme~t at cut r per unit lead at pOint ~.

Note that in this case [gim] consisted of only one column, inasmuch as there was only a Single external load. In Art. A7.9 the strain energy was written
(15 )

(18)
- the :::a.trix of influence coefficients relating relative displacements at the "cu~s" to ~9dundant :oads at the Wcuts": displace~ent at cut r per unit redundant force at cut s. With the above notation one may Write

where [OiJJ is the matrix of member flexibility coefficients (Art. A7.10). If now eq. (14) and its transpose are used to substitute into (15) the expreSSicn becomes (note the use of (i,J), (r,s), and (m,n) interchangeably)

2U

(LPmJ [~iJ L~J [g,.iJ)X

+

(19)
Now according to the T~eorem of Least ~ork aU/a qr = 0 for continUity. Then, differentiating eq, (19):

x ~iLJ X([!j~{qs}+ [!j~Wn}}

Multiplying out

This last result may be verified by writing eq. (19) out in expanded form, differentiating and then recombining in matrix form. Rearranged, sq. (20) gives

(21)
The reader may satisry himself that the term in the ~ldd1e of the above result is correct by Qbserving that, because of ], the symmetry at
~crass praduct~

[u1J

Eq. (21) is a set of simultaneous equations for the redundant internal forces ~,qs' It :nay be compared With sq. (6) of Art. ,1.8.6, to which it corresponds. Eq. (21) ~y be solved directly from the form there displayed or its solution may be obtained by computing ~rs:JJ the inverse of the matriX of coefficients, giVing

{qs}= - ~rs~ ~rJ{Pn}

ru -,

(22)

The matrix product - fa -~ gives the The various ~atrix triple products occurring L.:rs....J L..:rrrJ above are assigned the following symbOls, each va Iues of the redundant forces ~cr unit values having the interpretation given (compare With of the- external loads. This may be given the eq. (24) of Art. A7.9) sy:nbol ~s~ so that - - - - - -(16) - the matrix

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

If :lOW eq. (22) is substituted into sq. (14) one gets (With exchanse of r ~or s ~d
a

AB.19

L /3

Liz
1

a rcr n)

{qi} = ~i~

{p,,}- ~iJ ~sr:J ~s;::Wm}

-

CiJ

L/2
L

a a
V'/3
Liz

a
0

E:I

a
0

a a

Liz
1

(~~
=

~iJ ~s[] ~sJ ) {Pm}

(~iJ+ ~iJ l:rmJ ){Pm }

The moments q~ and q. were taken as the redundants. With these set equal to zero, the internal force distributions due to application of unit values of P l and P a were determined, giving

The matrix set of~ in parentheses above, gives the internal force distribution per unit value of the external loads. It is given the symbol
(23)

ilL

~~

a a ilL a a

a a

With the applied loads set equal to zero, unit values of the redundants were applied yielding
(r =)
(2 ) (4)

so that

- ilL
(24)

Eqs. (23), (24) constitute the major result, inasmuch as they present the ~eans for computing the internal force distrib~tion in a redundant structure. Example Problem 13 The doubly redundant beam at Fig. A8.27 (a) i5 to be analyzed for the bending ~oment distribution. The beam is loaded by couples over the supports as shown.
EI constant
L
~P2

~!J

1
=

a a
-11L
1

ilL

Note that redundant load qa was applied as a self-equilibrating internal couple, acting on both beam halfs. ----- The following matrix prOducts were formed:

[SJ . [!r!J ~ti] ~J !J

~ f--q,( Ii q,

..!

-Pi
L

----1

., L'" .,
:.1 0 0 -liL

('I
L : 6E:.1

q2( II q,
Fig. AB.27

} (hi

~~
0 0 i _lit. ."

r I" 'J
!..Iz
1

Liz
0 0

0 0

L-/3
L/,

L:,

~L -~L

rrnJ [!r!J rti] ~J;]

So Lut Lbn:

The choice of internal ~eneralized forces is shown in Fig. A2.Z7 (b) .. T~e appropriate ~ember fleXibility coefficients were arranged in ~trix :or~ as (ref. Art. A7.l0 :or coefficient expressions) .

-r.
!
.

J["
1

L/, 1 0 0

0 0

Liz
0

1.-/3

Liz 1

0 L 1 2 .. : mOl

L/2

~ ~

I' .:J
0

... ~

AB.20

STATICALLY INDETERML.,*ATE STRUCTURES

The inverse of Cars] was fOillld (ref. appendix)

~rs ~
II

_J.

_ 5EI - 7[

C -J
2
-1 4

though the ma'tr-tx of member- aexi'Ji2.1 ty ccefflclents was expanded to a 6 x 6, the coefficients for q, and q, were zero. ThUS,
VI/3

L/2
1

a
0

0
0 0

Next, the unit redundant load distribution was found (eq. 22).
1

~lLJ

_ L
- ;;:1

L/2

a
Liz ,

c
0

a
0

a a
0
0

La/3 L/2
0 0

- '7
=

0 0

" r!
0

0 0 0

0
0
0

r-.

286

-.428]

LI43

-.28~

With the redundants q~ and qa set equal to zero, successive applications of llii!t external couples PJ. and Pa gave the stress distribution
0

Finally, the true unit stress distribution was computed.

a
0 0

[!1;]
1
0 0

a
1
0 0

_1/

0
0

= 1 L 0
0

1
0

L
2

1
0
0

286

-.42

.143

-.286

I/L - I/L
0 0

J
-1

and, with PJ. and P a zero, successive applications of unit redundant forces q$ ~id q~ gave
0

0'\
1

1 1.286 - .286L - .429 .143L

.428 -.428L .858 -.286L

The~,

=L

2
3
4-

~lJ

-I
=
2L
1

-I
L

La

(The tabular form at presentation of the matrix Gim, above, is used here only to indicate clearly the functioning of the subscript notational scheme. In general, it should be unnecessary to callout the subscripts in this fashion excepting for the larger matrices, tor the handling of WhiCh, the tabul~r form may prove helpful.)
Example Frob lem 13a

multiplying out per eqs. (17) and (18)

["r;;]
~r~
The inverse was found:

L' 2E1 L'

6Ei

r
-,

J
sl
51
-5]
l~

~5

The redundant beam problem of Fig. A8.27 is to be re-solved using the redundant reactions as the unknowns. Solution: The support reactions under the loads PJ. and P a were given the symbols qe and qe, respectlvely, Positive up. These forces did not enter into the strain energy expression so that, a i-

6EI
7L'

Finally,

rG l L:1m ~

r:. -,
~1~
-

~:CJ ~sr~ ~s~

;:'l

r;; -,

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8.21

1.286

.l25

=L

.':::;CJ..,

-.423L .857 -.256L -.428 -.428

Loading for column {gi2} of the matrix [gim]
Fil{. AB.29

.'+<:.c

.143L -1.256 1.714

(COm'Cl2.r':! 'fIith s ojut i on to Example Problem 13). Problem 14 The oontinuo~s tr~ss of Fig. A8.28 is t.vr cc redundant. :t is desired. to analyze it fa~ stress distributions under a variety of loading conditions consisting of concentrated vertical loads applied at the four external points ~ndic2ted.
~ample

/\t:it:J%f2t\l\/V\
I

.5

1#

.5

~,,~

Fig. AB.30a

{gi3} loading

'R2
8 panels at 20" :: 160"

Fig. A8.28

Solution: The internal generalized forces (qi,G )

j

the axial loads in the various ~eabers. These ~ere numbered from one to thirty-one as ShC~~l on the fl~1re. The member fle:Krbility 80efficients in this case 'Nere of
~ere

employed

the form c , = L/AE (Rei'. Fig. A7.35a).

coeffic1ents arf written as a column matrix below. (They 'NE::ce employed as the diagonal e lements of a square ~~rix in the matrix multiplications, but ars written here as a column to conSer.e space.) Membe~ loads qa and qo were selected as redundants. ~ith qa and Qe set ':!qual to zero (''But''), un t t loads were applied succ ess ive Iv at external loading paints one through four, t~e four stress distributions-thus fo~nd being arranged in four columns giving the matrix
~i~ (below).

The

20 20 20 20 20 20 20 40 40 40 40
40

~
1 2 3 4
0

3 0
.0

5 0
0

6
7

a
9

1.0 .5 0 0 0
0

0 .5 1.0 .5
0 0

10
11

-.25 -.75
-. , 0
o~

a1 =~

40 40 40
22.422.4 22.4-

~1~

22.4
22.4 22.4. 22.422.4

By way of illustration. the used to obtain the second column

loading

fi~ure

of ~i;] 1s shown in Fig. A8.29. Next, unit forces were applied successively at the redundant cuts "three" and "rtve'' as shown in Figs. A8.30a and A8.30b. ~hese loads were arranged ~n two columns to give the ~trlx ~iJ

22.4 22.4
22.4-

m" 22.4
22.~

22.422.4

0 0 1" 0 0 18 I .56 0 19 -.56 0 20 0 561 21 -.56 0 22 -.56 .56 23 .56 -.56 24- -.56 .56 25 .56 -.,:)6 25 v -.56 27 0 .00 29 0 -.00 29 0 .56 30 0 0 31 0 0

12 13 14 10 16

-.25 0 0
0

0 0 -.25 -.75 -.75 -.25 0

_.

AS. 22

STATICALLY INDETERMll'{ATE STRUCTURES

I~
1 2 3 4
0

1
1.0 .50

05 9 [ .0035

.111

.0065 .111

.0035'] .059J

.0065

-.50

The calculation was completed as per eq.

(23) to give [qim] , the values of the member

-.50

7 8
9 10

11 12 13 14 15 16 17 18 19 20 2 22
23 24

.50 - .70 - .25

.50 1.0

forces for unit applied external loads.

.25 .25 .25 .25

Ix
.25 .75 .50 1 2 3 4 5 '6 7
8
I

1 1.0 .529 .059 .031 .003 .001 0 - .50 - .764 - .294.045 - .017 .002 0 0 l.12 -1.12 - .527 .527 .527 .527

2 0 -.445 .111
.059

3 0 .003
.006

,
I

.001
.003 .031 .059 .529

.059
.111 -.445

l.12 -1.12 .56 .06

.56

.56

-.56 .56 .56 -.56

I

25 26 27
~a

9 10 11 12 13 14

29 30 31
NOTE:

-.56 .56 .56 -.56

- .56 .56
-

.56

.56 -1.12 l.12

VOIDS INDICATE ZEROES

16 17 18 19 20 21 22
23

.031

Multiplying out gave

[136

L:a.o

-a. ill
136J

24 25 26 27 28 29 30 31

.031 .031 .031 .001 .001 .001 .001 0 0

.006 .003 0 0 .222 .167 -.085 -.032 -.003 -.001 0 0 0 -.498 .498 -.622 .622 -.056 .058 -.058 .058 -.003
.00:3

0 0 -.001 -.003 -.032

-.oas
0
0

- .017 .04.5
- "'G' " " .... - . , ::
.....

1.0 0 0 .002

.167 .222 0 I .003 -.003 .003 -.003

.058

.':::',,7

.00 -0
0

-.003 .003 0 0

-.058 .058 -.058 -.622 .522 -.498 .498 0 0

.001
.001

.001 .001
.031

.527 - .527 -1.12

1.12

.031 .031 .031 .527

.52""

L~'O

-15.0 0

0 -15.0

-~.J
a1ll
13~

The inverse ot ~r~ was found (ret. appendix)

[136

~.o

Example Problem 14a Fig. AS.31 shows the two bays ot a steel tubular tail fuselage truss Which is loaded by tail air loads to be resolved into t~ree :oncentrated loads applied as shown. The fuselage bulkhead at the attach-paints station (A-E-F-K) is heavy enough so that it may be assumed to be rigid in its own ~lane. Hence, the truss may be analyzed as if cantilevered from A-E-F-K as shown. All members are steel tUbes, their lengths and areas being tabulated below. Solution: The generalized forces were taken to be the member axial loads, these being numbered as in the table below. Member fleXibility coef:icients, L/A, (E set equal to unity for canvenience) were also tabUlated.

found per eq, (22).

Next, the values of the redundant forces, tor unit values of the applied loads, were

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8 23

than apply successive unit loads and forceS at points m = 1, 2, 3 and r = 22, 23, 24 and then carry each loading through the structure, another procedure, often better adapted to large complex structures, was employed. In the method used, the 9quations of static equilibrium were written for eaCh of the seven Joints.* Summation of forces in tr~ee directions on each jOint gave 21 equations in 24 q's (the unknowns) and the tr~ee applied loads P~, P3 and P~. Some typical equations obtained were: 1rom Z F on Joint C.

A1~"~

.f.

.213'. q . . . . 213'. q . . . . 18334 q . ,:334 q. ' -Po

PI
I"~

~ ~
,

II- J--:::-- -'.---G

---

..... B
-,

__

-, ~~, "I ......"

H

"' "

P2

{
::I.......

'" -1.0236 P,

P
3

From Z F on Joint B.
.081759 Cl.1. - 18334 Cl.. - q.
q

-.07617 Cl.1.. '" .5227 q - .91593 Cl.u

, 0

Fig. A8.31

-.1410

q1.'"

'" _.8523 Cl. ..

~ER

N"lJ11BER

L2NGTH

AREA

L/ A
52.8 51.0 .00."" 51.0 62.8 62.3 44.85 30.3 44.85 50.8 62.8 64.1 90.9 55.1429.62 29;52
~

RC

cli

BD CD DE ?G GH HI IJ GJ

1 2 3
4 5 6 7

5
9

J" AG
BG

3H

10 '1 12 13 145 i5
17 ,Q
;

35.47 25.43 9.20 25.43 35.47 35.47 25.34 5.00 25.34 9.20 35.47 40.4 15.0 27.57
11. '70

.565
.49~

.165 .499 .565

.565 .555

And so forth, for the other Joints. Note that in each case the equations were arranged with the applied loads (P ) and the n redundant q's (q33, q33, q.~) grouped on the right hand side at the equal Sign. This arrangement was observed for all 21 equations, atter which the equations were placed in matrix form as

He

Dr
JJ SJ
.' J

Ie

.. ,

11.70 27.57 15.0

40.4-

.165 .565 .165 .565 .630 .165 .500 .395 .395 .500 .160
.630

i,j
n

= 1,2,
= 1,2,3

.. 24

55. "490.9 64.1. 66.1

= 22,23,24

FJ
" ~.
.U

-. "

20 22 23
2.:1

".
I~

sa

37.36 37.36 26.23 17.60 .60

.565 .565 .500

.... ..,0 ; ....0 0

66.1
52."~6

(Note that there were 24 equations ~ere, the adcitiona1 truee equations being the ident::.ties

,-

'Q.

.-

1.06.6 lOE.S

The structure was three ti~es redundant. rra~ework of p jOints, 3p-6 !ndepence~t equations ot statics nay 88 wT~tten (p. A8.10). Here, ~oweverJ stress cetails in the ~l~ne AEKF are t8 be sacrl!iCed, six equations are lost ~hereby si~ce onlY net ~orces in two ~:rec~ions in :his ~lane can Je sUT~ed. 3 x 11- 6 - 6 =21 equations, 24 member' unknown. ~e~bers 22, 23 and 24 ~ere cut. The next steQ Nas to comnute the unit stre~s ~istri~uti~ns [glm] and [~irJ' Rather
a space

On the right hand side of the above matrix equation the ~trices are shown ~partitioned~. T~e ~irst three columnS at [OJ are the coefficLerrt s of P and the last 'thr-ee are the
n

For structures other than trusses the equilibrium equations are written for the various structural elements, equilibrium of joints alone being inappropriate.


A8.24
STATICALLY INDETERMINATE STRUCTURES

coefficients of the q_.

o

The inverse was

~or t~e

rocnc

(see appenc Lx)

The ::latrix equation NaB solved

q.

by finding the

lnver~e

of [C

Thus,

1 j]

" (see appendix).

trs-J
T~en

10

_.

[5555 -.2550

.2580 l

3.C~12

-~ '2:" <::'..... ,-,1.:)

'2~7~

.2775

2.3015

3.023

(referring to eqs. (22) and (22a))

where

IG "I =_Ia LSllJ - LTS

-'J LTllJ ICi "I

= 10

-~

~
2

.18
","

2259 -238

-be."=,

'"'n .". -.<:v1.v

- 31A

121.~

ThUS, the ~~lt stress distributions in the determinate structure were found by a procedure having as its ~ln advantage the reduction of the work to a routine mathematical operation. In the conduct of this work appropriate standardized techniques may be employed. The result in this case was

?lnally. the complete unit stress

1II8.S obta Ined as

distrlbu~~on

1;\
1

1
.5415

a
2~727

, s

, " .. :~
..

.'

_2.72:7 4.0.59

.. . .,.
-c .

5.;'7~

Z2

Zl

-2.:!C59

~::

zrs

.:).5<> .. .

,
,

I
,, ,

~ o
0

X
1
2 3

I ,

.8412 .00017
.01,523 .00017

_7.~

..

,:;,,60

.,

.
,.

,.

.4415

- .0=

" - .aiee .5 ,5 ,
z

."'"
0

.54.

.1 ..l" .. as -3.n4 , . -i .:.009 .eess .01<+5 00 -r.v-a

_2.340

.06<:

.. ..
.&C .22

.."'

.. - -0

.0

4
5

..;..,5
."'~

.6529
-1.4-04 .8443

..

- ...17
.0

2.," .. ,

-1. ..35

c
0 0

r..4J.S
.3523
0 0

6 7 8 9 10
~1

. ,

~
.

, -L'"
.0

.W

. 1.W9
0

-.... 5

.zao

, , ... ,'."

.27E" .2445 ."S06

.:...,27
,

.o se
1.102
.~O'"

1..;,01

_,.~l

-'-.302
0

12 13 14
15

oe o
0

1.0

The member flexibility coefficients L/ A were arranged as the diagonal elements of the matrix [u 1j ] Then, multiplying out according

to eqa , (17) and (18),

-1137.0 - 151. 7 161.6
1150

16 17 18 19 20 21 22 23 24
~ar.mle

.1094
.84.50

.0973 .0494 .6638 .2350 .9185 .5116 .5120 .9193 .1061 .1434 .2023
.2660

~...,.=:: 1.296 2.203 .~38 .31.53 .1190 -2.206 .E;39 -3.008 i -1. 557 1.5247 i .332 .7352 I .761 .095<::'. I .096::. -1.260 -1. -t-50 .0205 .3338 -1.607 I -l.S'~2
.~IV

.3143 - .3896

.7998 445.4457 I .B002 I .3638 .2411

1.6471 .3'796 .543 i

.733
.2641

. , -

.0002
.1364-

.03 4

,-

- .6956
.122.3

.6223 .461 .461 .829 .3205 .8512

.556e

.238 .176

2.269 - .2010

~~
~r;]

[177,9 195.0
-154.4

-2263.0

-6430'~
2273.3

Problem 15

2970 1150 -1148

1221

-1035

-114~
1224-

The doubly symmetric four flange idealized box beam cf Fig. A8.32 is to oe ~n~lyzed for stresses due to load applicat:on ~t the six paints i~QicateQ. Fl~~ge areas taper linearly from root to ~1p while sheet thicknesses are constant in ~ac~ panel. The bea~ :s ~o~ted rl~1dly at t~e ~oot, provid1ns ~u:l restra~nt
These are the "effective areas", being the flange area plus adjacent effective cover sheet area plus one- sixth of the web area. (The factor of one-eixm provides the same moment oC inertia as the distributed web area).

-1035

see references in appendix.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8.25

against war?ing of the root to any torsion loadings.

CT2SS

section

~ue

The ~e~ber :lexlbillty coefficients were collected :n ~trix :crm as shown below. Note t~at entries for u," U e e Q v v and u~o,~o were collected irom ~NO stringers each (as well as being doubled as discussed above). Coefficients for these tapered stringers were computed from the formulas of Ar~. A7.l0.

1)< , !

, IaI , ,

s I
, ,

. i

'

I
,

I
, ,

10

!Z

I IJ I

iaa

! ie , ",

" ace

..
,

..
,
c.
15.2.il

,-.o.~

NOTE:

..

.
,
.~

.
, .,

"

::7 .5

VOIDS INDICATE ZEROES,

=1

Fig. A8.32

Cover sheet shear flows q~, q7 and ql~ were selected as redundants. ~ith these set equal to zero, and with unit loads placed successively at loading points ~one~ trxough ~sixw, the

[gim]

Solution: The generalized forces employed are shown on the exp Ioded View. Fig. A3.33. No forces were shown :or the laHer surface, its ~embers and forces being equal to those of the upper surface because

matrix was obtai~ed. ~~en the cover sheets are ~cut~ the ~NO webs act independently as planeweb beam5. The details of the stress calculation for such beams are similar to those of Ixamp.Le Problem 21, Art. A7.7 and are not shown here.

of

symmetry.

In the

[u 1j ]

I~

1 ,400

matrix this fact was accounted fer by doubling the ~ember fleXibility coefficients for the members of the upper surface.
'\.'1:"
,Qle

3 4 4.

.\00

4,

.200
8 9 5.341 .00
2.571

,200
2.571

~~ \~

10
12 13 14 6.231

5.341

.08_87
,0444S 4.006
5,

.08887
2.003
4.006

,1333 2.003

IS

NOTE:

VOIDS INDICATE ZEROES,

Calculations for [glr] were ~de by suc-

,'.~~,
q,
Fig. A8.33

ceSSively assigning unit values to the reduncants q s, q.., and q u. The calculations are illustrated by the exploded view of the end bay in Fig. A8.34 showing the calculation in that part of the structure for q~ = 1. Note that q:z = 1 was applied as a self-equilibrating pair of shear flows acting one on each side of the "cut;". The r-ibs were considered rigid in their own planes.
?ut
q~

=1
(1: M, Z F)

From equilibrium of end r t o :

A8.26
q ",0

, STATICALLY INDETERMINATE STRUCTURES

T:-.8 f'ollcw1r:g matrix products were ::om.ed:

.~E;--'-<-S\l25
,40",-25.04 .75x~=.5625

I~ .;'" ~q

Per

eo..

(18) : .1769 .2578 .07322 .0520j .07322

.1254-

2.5 j

..

\q

IS"

25.04

~r~

- E

_ lOll

[774 .1789 .0520

lX20",75

15

-IX~~ 5.0
fq,=_1
Fig. AS. 34

-"

~\~
'\

Per eq. (17) :

~;] urn : 10' T

-.=

_.ro:l56

._,
.ceeec

009318 .009318 _.001605 .001605

-.OU56 .OllS6

_.C02Z59

.~2259

_.008216 ,008216 -.005150 .005150 _.002278 .002278

q."

f q,-I
q,..= 1

appendix)

The inverse of Cars] was tOlUld (ref.

=
E 10 IS

q,. -11Cj!.__.

__._...__..':;S q3 " 1

e-:
Then

[3,862 -2.562
- .1137

-2.562

6.137
-2.321

- .113J
-2.521

9.499

q, = -1.0 q, = 1.0
From eq'.l111brlum of 200 rib: q. = -t .5625 q. = .3125 q. = 0

~sJ

=-

~rs -j ~r;;]

.25) = -.3125

[0699 -.0699 .00598 -.00598 .00016 -.00016 = .03593 -.03593 .03409 -.03409 .00403 -.0040 .01568 -.01568 .01872 -.01872 .01578 -.0157

by hypothesiS

So on, into the next bay. In the idealization used here l the ribs have zero stitfness normal to their own plane so that the axial flangeloads are transmitted directly to the flange ends ot the adjacent bay (see q~, q" q .. q"
of Fig. A8.33).

Finally, the true stresses were (per eq.

~i~ I~
1 3
4
0

=
2

~i~
-

~1J ~~
4

1 .330

I~
1 2 3
4

2 - 1.00 1.00 1.00 -25.04 25.04 .3125 .3125 -20.87 20.87 .1387
.1387 -18.79
18.79

12

6
~

5
6

~iJ

7 = 8 9 10 11
12

- 1.00 1.00
l.00

-29.38 29.38 .1955 .1955 -25.44 26.44

0 1 12 3
- LOO
4 15

.0699 .0060 .0060 699 - .0699 .0699 .0060 .330 2. 6 1.75 - .150 .75 2.26 .150 .08 9 .16B .0141 .0341 .0 9 - .0359 .0359 .0322 141 .0 1.w45 2.51 2.83 1.126 4 - .0010 .0776 .0157 - .0157 .0187 - .0010 .0454 .0112 3.47 2.76 2.40 2.76 3.47 1.61

.0060 - .0060

.0050 -.0002 .0002 .0002 -.JCO 2 .:JOC2 -.00'22 .150 -.0040 .00'+ o .150 I .00401-.004.o .0322 -.0040 . 0: 40 C341 .0040 -.' 040 . .168 .0040 -.004-o . 22 1.16 - ....22 1.5<t.5 .122 -.122 .0112 -.0150 .015 o .0187 .0158 -.015 8 .0776 .0150 -.015 o 1.61 .610 "-.610 2.40 -.610 .610

1.00
LOO -31. 71 31.71

13 14 15
NOTE:

VOIDS INDICATE ZEROES.

The reader will observe that the result displays the ~bending stresses due to torsion~ that Is, the bUildUp of axial flange stresses near the root of a beam under torsion when the root is restrained against warp1r.g. The solution for application of a torque is readily

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

AS 27

found by superposing stresses for P~ = 1 and Thus, one :inds that under this condition there 1s a root flange load of

P a = -1.

=;

44

-15.0

ql.

= 3.47

- 2.76

= 0.71

Ibs.

o -15.0 38.0 00 38.0 [ o 0 -15.0

472 .888 .0525

-1~J
144J

for a torque of 15 inch Ibs.

.028

AS.11 Redundant Problem Deflection Calculations by Matrix Methods.

388
~0525

1. 665
.0525

.0975
.888

. 0525
.888 .472

Datlections (and in particular, the matrix of influence coefficients) are readily computed from the results of Art. A8.10. Assume that the redundant forces { have been det.er-atned from eq . (22). The total de~lection of external loading points is then eaSily computed as the sum of, 01~, the deflection due to external loads acting on the

.0975 1.665

.028

qs}

J
.052

Finally, the sum of the above eNO matrices gave

"cut;" structure, [amnJ {P n } (ref. eq , 16) and, TNO, the deflections due to the redundant forces acting on this same cut structure,

~,,;;] ,. ~

~'#
-

-15.9

- 15.9 .052 .028

36.3

.052
.098

.098 .052

36.3
-15.9

-"J
-15.9 144

[a.:nsJ {qs} (ref.

eq , 17).

Thus,

Example Problem 17 Determine the matrix of influence coefficients for the box beam of example problem 15. Solution:
An alternate procedure to that shown by eq. (25) was followed. The influence coefficient matrix was formed as in Chapter A-7, Art. A7.9, by the product.

SUbstituting from eq. (22)

{ 6,. } =

~:nnJ

{n }- ~msJ ~rsJ ~~ {q
P

= (

~mnJ - ~:nsJ ~r(] ~r;] ) {pn}

This product was formed readily, inasmuch as

[G
] im

was evai tame tram example problem 15.

----

The matrix expression set off in parentheses above, giving as it does the deflections for unit values of the applied loadS, is the matrix of influence coefficients. Let

The result was

4363
3500 1522 3500 4363 1243 1522

1243 -

1522 1243 29.7 l243 1522 -29.7 29.7 897.8 594.2 33.7 -33.7 594.2 897.8 -33.7 33.7 33.7 - 33.7 .8 -36.8 9.7 - 33.7 33.7 -36.8 36.8 1

-29.71

so that
- - - - - - - - - - - (26) AS.12 Precision and Accuracy in Redundant Stress Calculations.

~ample

Problem 16 Deter.nlne the matrix of influence coefficients for the redundant truss of example problem 14.

Solution:
~roducts

From the previous work the following wsr9 fo~ed

~trix

Matters of prectsion are dependent upon the number of Significant figures obtained and retained in dealing with the geometry of the structure and in the care with which arithmetic oDerations are perfo~ed. In the discussion to follow it 1s assumed that all due caution 1s exercised with regard to the preCision of the work. ~~t~ers of accuracy'r4ve to 10 with the number of Significant figures :lnally obtained in the answer as influenced by the ~er of formu

A8 28

STATICALLY INDETERMINATE STRUCTURES

lation of the ?roblem. ~~e accuracy ~f the ~e sult may be affected by a n~~ber 0: factors, two of the most important of wh i ch are discussed here. Two factors infl~encin~ accuracy often are considered together under the heading ~choice of redundants n They are:
.~- the number of Significant ~lgures which may be retained in the solution for the inverse of the redundant force :oefficient TIatr-ix , Cars] (eq.21).

inverse ~ar~ot be ro~c (r sr , appendix). condition is terrib12. :::xampl-a:

In the matir-tx
~ents

- the magnitude of the r~dunGant ~orces relative to the size of the dete~,lnate forces. These ~NO factors are concerned respectively With the left hand side and the ri~ht hand side of eq , 21, 'Viz. '--

are measures o~ the str~ctural crossone red~~dant force with a~o:~er. ~he st~ength of t~is cross-coupling is de~end ent upon the choice of red~~dants made ~~ "cu't t tng" 'the s tr-uc tur'e to make it s ta t tca Lj.y determinate. JOR EXAMPLE, the do~bly reduncant bea~ of Fig. A8.35(a) ~y be ~de statically 1eter~in~te by "cutting n ~y two constraints.
cou~l1n; o~

they are discussed :n detail below. - ACCURACY OF INVERSION OF ri

Lr~

5':::::::::::-: ==~ ~ (c)

:l ;

(bl

~
(dl

THE CONDITION OF THE nATRIX.

The characteristic of the mat-~x .. ~ [0 rs ]

Fig. AB.35

which determines the accuracy with which its inFig. AB.35(b) shows the c~oice o~ generalverse can be computed is its condition. The ized torces. Only two (Ql and qa) a~Q required condition of the matrix is an indication of the to describe the strain energy, but the central magnitude of elements otf the main diagonal support react1or~ were also given s~bols as it (upper left to lower r1&~t) relative to those was desired to consider them in the dis~ussion. on. The smaller are the relative sizes of Then elements off the ~in diagonal, the better is cthe condition of the matrix. A weJl conditioned 2/3 1/6 0 0 matrix is more accurately inverted than a poorly conditioned one. 'rNo extreme cases are now 1/6 2/3 0 0 given tor illustration:

a) the diagonal matrix. Its ott-diagcnal elements are zero, so that it is ideally conditioned. Thus. the inverse of

C:I1J =..1. EI

0 ,-0

0 0

0 0

2..

Is eaSily and accurately obtained as

FIRST, suppose the beam was Iade deter.ninate by selecting the support reactions q~ and q~ as redun1ants. The ~cut structure~ in this case may be visualized as the Jeam of Fig. A8.35(c) whose central supports have been removed. Application of unit redundant ~orces q3 = I and q .. = 1 gave

~
1

4
L

- 3

:3
2

L L

b) a matriX all of whose elements are equal in each row. All the elements cff the main diagonal are equal to those on. --rhe dete~in ant of such a ~trix is zero and nence its

~1~

0 <.

;j L
1

:3
0

3
4

ANALYSIS AND DESIGN OF

~ulti~lYi~g

FLIGHT VEHICLE STRUCTURES

A8.29

out,

r'edundant s may be chosen simply.

laEI r,s=3,4

L'~. Lz 71 ~ 8

The condition of this matrix is poor.

pnys i caj jv , a unit load at po trrt "3" causes a Imos't as much cer rec t ron at point "4" as at "3 11

The crosS-coupling is large. SECO~TI, suppose the moments ql and q~ had been chosen as ::-edundants. 'rne "cut structurel! in this case being visualized as in Fig. A8.35(d). Application of unit redundant forces q. = 1 and qa = 1 gave

itself.

I~
1
2

1 1

a
1 1

a
-I:
1

In most structures, however, the additional labor involved in seeking an ortho~onal set of redundants is not warranted. (2) In choosin~ a set of redundants which will yield a well-conditioned redundant matrix, it is best to make such "cuts" as will leave a statically determinate structure retaining as many of the characteristics of the original structure as possible. ThUS, one may consider that the structure of Fig. A8.35(d) retains more of the features of the original continuous-beam structure than 10es that of Fig. A8.35(c). (3) The degree to Which one redundant influences another (extent ot cross-coupling) can be visualized by observing how much their individual unit-load diagrams "overlapH. To elaborate on this last point, refer once again to the above illustrative example. The cross-coupling of q3 with q~ may be expected to be large if their unit-load diagrams are drawn as below in Fig. A8.36(a). This deduction follows eaSily if it is recalled that the dummy-unit load equation for such a cross-coupling term is at the form

I:
2

Jm~.:n~ ex- ,obviouslY

EI

1"

Lar-ge .. or

m~,

L -I:

m..

Strictly, the comparison is with the terms

dx

tIultiply1rlg out,

elements of the ~trix [urs]

J -W-'

:0:

Im:

EIdx

.w h' ' thie on-a.. .,. agcna.J.. , ... ch rorm

r,s=l,2 The condition of this redundant ~~trix is obviously better than that obtained with the first choice of redundants. There is less cross-coupling between the redundant forces.
Thus the analyst, by choice of redundants, determines the condition of the matrix. The choice may be critical in the case of a highly redundant structure, for it may prove impossible to invert a large, ill-conditioned matrix with the l1.mited number of significant figures available from the initiaJ. data. The following statements and rules-of-thumb may be useful in the treatment of highly redundant problems.

m,la
m

z[

~
(b)

Fig. AB.36

St~dy

of

t~e ~it-load j1agr~~ fo~ ~he

redundant choice ql' q~ (Fig. A5.35b) reveals that the cross-coupling sh~uld 88 small her~
....... since an 1rrteetr-a l J ':J" v .9 ~ave a contribution from Hence, the J ::I1.:n'l, ct.:.:: is

r' orm

J:nlffi' l ex EI
ex

can

(1) It is always possible to rind a set of redundants for which tIle cross-co'J.pl1m; is~t a ~inimum-zero, ~n tact. Theoretically t~en, by ~roper choice of redundants, tte ~a~r!x [Q ] rs ~y ~e reduced to a diagonal ~trix (idea~ly conditioned) . (la) The choice of red~~cants which gives zero cr oss-c oupt inz ("orthog:Jnal func-O:ions") 1s no'\: reaai i . I!1 some special ' / found in :::ene p :;' . structures, such as r tngs anc rr-ames, orthogonal

the center

s~an O~ly.

sma11 er h cnan

r-

0'

obvt cus Ly c ons tder-ab Ly

ex 31

or J

r ::I~

~I

Th .. us, a
~han

Visual
q,

~nspect1on

of Fig. A.36 reveals that q1.,

qa is a better choice of redundants

is

c...

.. see eq. (8) Art, A8.7.

AB.30

S!ATICALLY INDETERMINATE STRUCTURES

(4) Combinations of ~ed'~dants ~ay be ~m such a way that the ~rOduct [gir] ] does ro played to yield new '~it-reduncant stress distributions which do not "overlap" as extensively exactly 'Nt-at is jesi~ed, 7iZ., "t~ke Qn~ ti~es as do those of the individual red~dants origin- the q3 coll~~ ~l~us one halt the q~ ccl~~ to ally chosen. give the first column of the new distribution For ex~ple, suppose that in the previous [giPJ ". Then "t-ake cmua one half of the q3 illustrative problem, the choice q~, Q~ for redundants had been made originally, leading to column plus one times ~he q~ column to give the the unit-load dia~acs of Fig. A8.37(a). Inspection of the diagra~s leads one to anticipate second column of the new distribution [giO] a stror~ degree of cross-coupling and hence a new set of redundants is sought. Rather than Multiplying out in the above example return to the s'tz-uc tur-e to choose new "cuts", combinations of the m3 and m~ diagrams are ~ looked tor ~hich will have less "overlap" and 0 - 2 L hence less cross-coupling. It is observed by inspection that two new 1 0 - 2 L stress dlstributions"'which have the desired property may be formed trom the m~, m~diagramB 1 1 by proper combination. Thus, if one halt the 1 - 2 m" diagram is subtracted rr-cm the d'Iagrarnto fOrm one newstress distrIbUtion and one halt 1 1 the ~ diagram ~ subtracted from the m~ dia- 2 gram to form the other new stress distribution, the results are as shown in Fig. A.37b. There Now form the ~trix of redundant coefficients ta. oovtous jy less "overlap" of the diagrams for these new combir.ations. for the ~ew unknowns (subscripts p, 0; [giP] =

[a

--m:

C: eJ

[gj,,] )

L'
24"::' .
Fig. AB.37

New

unit-redundant-force stress distributions (b) obtained by

The condition of this ~trix is ~eatly improved over that obtained :or q~, q~ alone (previously computed), ViZ

combining previous d1st.r1butions (a).

In this way two new unknowns are introducen by linear combination. In matrix notation. the old stress distributions [gir] are transformed to a new set [giP] by forming

~r~ = L~/18EI
(That the

(r , s

3, 4)

[apo] matrix

obtained here happens

to be similar to that obtained for Ql' qa in a previous example, is coincidental.) Once a transformation has been perfo~ed, leading to a new ~it redundant matrix [giP] ,

:3

2 L L

! 3
- :3
0 1
2

L
L

! 3
1 0

~ -:/J
,

1/2

the problem ~y be completed in the "0, 0 system". The appropriate equations ~re abtai~ ad from eqs . (14-), (21), (23) and (25) Simply by replac1ng all "r, 5," by "~, 0". Thus
(29 )
~re

The transformation ~triX [Srp] is written in

where the redundants q (= q ) of P o

the solutions

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A8.31

(30)

and where

auxiliary rules, test information, or even ~n tuitive guess-work which leads to a distribution close to the final true distribution. There 1S no need to set the redundant (c~t) member forces equal to zero in establishin~ the determinate distribution. Instead, reasonable approximate values may be employed fOr them. (The corrections to these values become the unknown----rediiiidants : ) Mathematically, the magnitUdes of the redundant forces are directly dependent upon the magnitUdes of the elements in the matrices [arnJ

The final unit load distribution is

(31)

and the matrix of influence coefficients is given by

(32)

or [apn] on the right-hand side of eqs. (21) or (30). (The right-hand side of such a set at Simultaneous equations is called the non-homogeneous part.) Thus, the relative mertts-orseveral possible determinate stress distributions may be judged by forming the matrix product earn] With each and comparing results. FOR 2XAMPLE, if the doubly redundant structure or Example ?roblem 3, Art. A8.5 were formulated in matrix forn the [gir] and [a t j ] matrices would be (see Fig. A8.38 for numbering scheme)
Fig. A8.38 Generalized force numbering

- THE t".AGNITUDE OF THE REDUNDANT FORCES; SIZE OF ELZME1~S IN [Urn]

Inspection of eqs. (14) or (29) reveals that the redundant forces act in the nature of corrections to the determinate stress dis~ributions origi~~lly assumed (the [gim]). It is

apparent that if these corrections are large, then any inaccuracies in the redundants (arising frcm the difficulties inherent in accurately inverting [ur s] or [apoJ) will have an im-

ecneme in illustrative proclem. q3 and q4 selected. redundants.

3.3

portant effect on the accuracy at the final a 432 0 a stress distribution. ThUS, as a matter of a 720 a a 1 general practice, It is desirable to keep the E aagnttudes of the redundant forces as small as 0 a 402 a possible. a a 312 a It follows immediately that one should use for the deter.ninate stress distribution one requiring a minimum of correction, i.e., Select .806 1.154 as the determinate stress distribution one-which apprOXimates the true stress distribution as -1.564 -1. 729 closely as posslDie. = ~1!J 1.0 a The rule giien in COnnection With the "choice of redundants " (rule 2, above) is an aid l.0 0 in ~king a gOOd selection for the determinate stress distribution. If, as suggested, a deterSeveral Possible deternlnate stress distriminate structure is obtained by making "cuts" butions will now be tried. FIRST, the which leave a system having properties similar to the original, the stress distribution obtained therein by statics should be a fair approxi- stress distribution obtained by statics alone in the "cut" structure (q, = q~ 0) ~tlon to the final true stress distribution. However, it is even ~ore im~ortant to realize that the determinate stress distribution only need be in static equilibrium with the external applied loads and that it may be dete~lned with the aid of any appropriate

[gtmj

AB.32

STATICALLY INDETERMINATE STRUCTURES

(This is certainly a poor approximation to the final stress jistribution). Multiplying out

ExaQple P~oblem 18 It ~s ~9sired to increase t~e ~cc~racy of the calculation in the case ~f the box beam, Example Problem 15. It will Je ass~ned for this ~urpose that the initial data of that proble~ were suffiCiently preCise to warrant an insrease of accuracy. Solution: The first step taken was ~~ examination of the unit-red~~dant stress distributions :~ta~ned with ~he choice pr~vlously zade of q2, ~~ and ~12 as r-sdundant.s . The VITee un i t -r-eduncant stress distributions were represented g~aphically as in Fig. A8.39.

SECOND, a stress distribution in which loads of 0440 Ibs were guessed at for q1 and Q3' (In a t'NO times redundant structure any two forces may be assigned values arbitrarily while satisfying static equilibrium.) q~, Q~ were found by statics. Thus,
0.40 0.258 } 0.40 { 0.0672

Multiplying out one finds

-18.79

1B.79

-26.44

26.44

-31. 71

31.71

= {

-101

9.49}

Note that a reasonable ~ at a stress distribution resulted in non-homogenous terms only one-tenth as great as those obtained by use at the ~cut~ distribution. The magnitudes at redundants are correspondingly reduced. THIRD, as a ~tter ot interest, the true stress distribution, obtained in Example PrODlem 3, was used. The result:

Fig. AS. 39
Unit-redundant-force stress distributions axial flange forces.

Inspection of' the figures showed that the following combinations of 11striJutions should give new distributions likely to have considerably less "overlapll. The non-homogeneous terms are practically zero, as they should be. The redundant forces would be zero also.
(1)-(2)---

lX{gl.}-~g:~~{gl,}+OX{gl'''}

(3)-- The demonstration here suggests a useful check upon the final result of a redundant stress calculation. After obtaining the final true stresses [Gim] (.. [Gjn]), one forms

0X{gl.}+ 1x{g!,}- ;~:~{gl'''} 0X{gl.}+ 0x{gl'}+ 1x{gin.}

form the
transfo~tion

In

~trix

was

o
and compares the result element-by-dement with the matrix previously computed,

-.8330
The "true-matrix" elements ought to be zero, or nearly so, if [Gim] is error-free.

Using [glr] as preViously com;uted, the ~ultl plication gave

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.33

I~
1
2

1
- 1.00 1.00
1.00 -25.04-

3774 10'
E:

.1789 .2678 .07322

.1789

'0520~
.07322 .1254

.0520

3
4

5 6

7
9

25.04 1.023 - .7103 - 1. 023

10 11 12 13

"
.0lD .0lD

. -

14 1 15
NOTE:

- 1.00 1.00 1.00 -29.38 29.38 1.0285 . 833D - 1.0265 .026 .026

- 1.00 1.00 1.00 -31.71 31.71

It remained to select a determinate stress distribution Which would reduce the ~gnitude of the redundants. For this purpose, the engineering theory of bending was employed to comoute stress distributions satisfying equilibrium for each application of a unit external load. The result was (refer to Example Problem 30, Art. A7.11).

~
1 2
4

.30
.10 .10 .0

VOIDS DENOTE ZEROES.

- .10 .30
2.0
2.0

.10

Fig. A8.39a shows plots of the new distributions, these having greatly reduced "overlap" (compare with ~ig. A8.39).

5 2.0 6 .0875 7 .0375

.0125
.0423

.0375 - .0875 2.67 2.67

.0023

.0125 0.15
.05 .05 1.33 1.33

.05 - .15 1.33

.05

9 2.67 1 2.67
11 12 13

.020
.0023

-31.71

31.71

14 3.00 15 3.00

.0713 .0267 .0423 .0179 2.00 3.00 2.00 3.00 .020

1.33 .0179 .0267 .0713

2.00

2.00

.0333 .0333 -.0333 .0333 .lOO 1.0 1.0 1.0 1.0

.100

NOTE:

VOIDS DENOTE ZEROES.

was obtained by pn] multiplying out per eq. (30b)

Fig. AB.39a
New unit redundant stress distributions,

Next the matriX [a

[gip]

~p~

=~

95 -3495 -1841 1841 0 0 ~ 1167 -1167 1554 -1554 -1450 1450 1082 -1082 1445 -1445 1802 -1802

This result compared very favorably With earn] previously obtained, the elements being fram onehalf to one-tenth as large. The solution ~ay be carried to completion in the "p,o system" using the matrices [giP} Capo} [giro] and [Opn] in eqs (31) and (32).
.000017J -.03124 .1254
A8.13 Thermal Stress Calculations by Matrix Methods.

was

The new redundant coefficient matrix [a ] po obtained by :n.ultiplying out per eq , (Soa ) ,

10'
E

2584

-.01132 .2566 -.03124

-.011'32
.0000175

This mat~ix is very well conditioned, being a considerable improvement over that at Cars] round originally in Example
~oblem

The thermal stress problem is conveniently !or.nulated in matrix notation by an extension ot the teChniques presented above. ?irst, consider eq. (21), wTitten
~n

15, ViZ.,

the

form

,-.Ae-f

-<rtf .:{-.J::-i!:
'~-:;:::;:-::;,?
"
~':.)"4

A8.34

STATICALLY INDETERMINATE STRUCTURES

In light ot the physical interpretations given to

may consist of the su~ of several contributio~s should qi act on more than one member. Is desired at each of the reBecause a
rT --

[UrsJ

and [Urn] , the equation is seen to be

dundant cuts

sq. (34) is expanded by wri t~:::g

(35 ) a ~trix-form statement of the condition for { rT } = ~rJ ~iT } continuity at the redundant cuts, ViZ., the displacement at each cut caused bY the redundant torces plus the displacement at each cut caused [~i] is, at course, the transpose of [girJ, by the external loadS, must be equal to zera. the unit redundant force stress distribution in To modify this equation for thermal stresses, the appropriate expressions for thermal disthe dete~inate structure. Substitution of eq. placements at the cuts must be added. Fallowing (35) into eq. (33) gives the argument used in Art. A8.9 one writes

{Orr } + ~r~ {

qs } +

~rnJ { Pn}

=a
Solution of eq. (36) gives the values of the redundants qs' atter which the problem ~y be completed in the usual fashion, viz,

where 0rT is the displacement at the r t h cut due to thermal straining in the determinate structure. The explicit form--ror this te~ will be derived below. Re~Titten,
(33) Eq. (33) is a modHied rom of eq. (21),

It is obvious that the use of combinations of redundants (the ~p,aft system of Art. A8.l2) is possible. One makes a direct substitution or

[giP] for [gir} giving as its solution the redundant forces in pn] for [Urn} etc. into eq . an indeterminate structure under the application (36 ). ot both external loads and a temperature distribution. MEMBER THE:R1lAL DISTORTIONS To der~ve an eXPliCit expression tor 6 , rT It remains to establish the forms for 6 the virtual work concept may be employed to adiT vantage. Thus, following the argument of the Thermal strains on an inrim teet inar eleftVirtual work ft derivation tor deflections (Arts. ment of homogeneous ~terial can cause uniform A7.7, A7.8), the thermal deflection at the rth normal extensions only, so that no shear strains redundant cut must be equal to the total indevelop. Hence only normal (as opposed to ternal virtual work done by the rt..h.-redundant Shear) virtual stresses-need be considered in torce virtual stresses (due to a unit rt..h- r e _ computing the internal virtual work. Note that dundant force) moving through distortions normal stresses associated with flexure ~ust be caused by thermal strains. included. It rollows, that only virtual 'Nark in Since the internal stress distribution is axially loaded bars and in beamB in flexure need expressed in terms of the internal generalized be considered. Hence 6 is zero for all qi iT torces Q1' qj' it 15 convenient to employ these Which are shear flows on panelS or torques on q's in writing the virtual work at straining. shafts. It one lets ~iT be the displacement ot internal BARB generalized torce ql due to thermal straining, The general expreSSion for the virtual 'Nork then the virtual work done by a single generaldone by Virtual axial loads u in a bar under ized torce 1s qi ~iT The quantity ~iT will be varying temperature T is called the member thermal distortion. The total virtual work throughout the structure is obw = u aT dx tained by summing, giving the deflection at the r~ cut as the matrix product Where a is the ~terial thermal coefficient of expansion.

[J.

(34)

Where ~i Is the value at the qi due to a unit (vtr-tua r ) load at cut r , Note that the term 6

*n iT

will be convenient later to designate by qST the solution to eq, (36) when the mechanical loads P n are zero, the

stresses in such a case being purely "thermal".

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.35

Eeveral speCific cases are now treated. A bar under linearly varying load with linearly varying temperature is shown in Fig. A8.40. In this case
u =

8EA11S
~or a bea~ the general form or expression for the virtual work done during thermal straining by a top-to-bottom-surface temperature difference 6T, varying linearly over the beam depth, is (see Art. A7.8, Ex. Prob. 24).

qJ

ql

- qJ
L L

x x

qj~llql
Ti

llll1lUJ.llill T
Fig. AB.40

IYrdx
-h-

T - Tj

Tl - Tj

where m

Then, assuming a constant

= virtual

moment (positive for compression on top fiber)

IYr
h

= TBOTTOM - TTOP = beam depth.

2f 1 + T
6

Ti+ J ql +aL 6
TIay be

zr j

Applied to the case at Fig. A8.42 one gets

qj

This expression

put in the form

aL
wher-e

11

1Yr 1) qi

+ ct, -

(1Yr1

20T,) J q.
'

6.

jT

e a

qj~T"::1
h

= constant

Fig. AB.42

Note that variation in the cross sectional area of the bar does not affect the distortions
'IT' 'jT'

or

The alternate choice of generalized forces for the bar under varying axial load is shown in ,ig. A8.4l. By a derivation similar to that above one finds

w=
where
e ' ; (31Yr 1 6+ IYr

AjT ;;; h

aL (1Yr1 +6 20T~

wner-e

Fig. AB.41

Special forms of the thermal distortion expressions for beams of varying depth may be derived readily as reqUired. Example Problem 19 The upper surface of the beam of Fig. A8.43 is subjected to a temperature 6T above that of the lower surface, varYing linearly ~s shown (i.e.,

The simpler cases of uniform load (qi

=U

constant) and unifo~ T (T = Tj ) follow imi ~ediately by speCialization of the above forms. For example, for a bar under constant load qi = qj' and constant temperature, T i = Tj = T, one has 6. 1,:, = a LT.

( r~=-x ~.~on~ -1
L L

Fig. A8.43
./'

I f', ...

"

AS. 36

STATICALLY INDETERMINATE STRUCTURES

the temperatures are equal at the left end and differ by aTo at the right end). Determine the center reactions assuming a 1s constant. Solution: In the illustrative example of Art. AS.12 this structure was analyzed by emploYing as generalized internal forces the bending moments q, and q .. over the central supports (see Fig. A8.35). The two central reactions were denoted by q~ and q~. The matrix of redundant coefficients, considering q~ and qa as redundants (the better choice, it will be recalled) was (ret. Art. A8.12)
L

2EI

solving

,}= {q s,

EI a OT,{.267} h .933

The final stress distribution was

=0

...

EI a 6T,
h

1 0

0 1 1

6EI

.257

The unit redundant load distribution for and q .. = 1 'NaS

q~

-2 C
1

C .933
2
c

I~
1

1 1 0

0 1 1

EI a. OT 0
h

2
3 4

933 :399/L
-1.60/L

.267

- C
1

C
2

Thus the reactions were

EI a. 6T o Lh

C - C

Member thermal distortions were computed. (Note that OT was negative according to the convention adopted earlier).
aL
2
X

c,

= -1.60

EI a 6To (negative indiLh

cates DO'>,/'N)

~lT

! 3
6

AT ,

11(-)

Example Problem 20 The symmetric sheet-stringer panel of Fig. A8.44(a) is to be ar~lyzed for thermal stresses developed by heating the two outSide stringers to a uniform temperature T above the center stringer. Assume G O.385E.

; -

"" h.

(-)

(2 x ~ OT6 + ~ 0
4

T0"" (""
+ h

Solution: The panel was divided for convenience into three bays. The numbering and placing of generalized forces is shown in Fig. A8.44(b). Transverse members (ribs, not shown on Figure) were considered rigid in their own planes - a satisfactory assumption for symmetric panels. Because of symmetry only one half the panel was handled. All member fleXibility coefficients and thermal distortions were doubled where appropriate. The matrix of fleXibility coefficients was set up as (VOIDS DENOTE ZEROES)

(-l

2 l "" '" - 3: i

0 T~

Then from eq. (36)

..1. 6EI

~ ~ {q'J ~ ~ ~!
14

aL6T,
3h 2aL -T 3h 0 c

q,

=(-)

0 - 1 01

C-

2 L-

o o

P'

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

177 .6
44.4 44.411"7"':'.6
44.~

AB.37

-.'0
3S.

I
< , ., .2 8S.3

T:J.en
22.2 44.4

- .5
31 200

0
.0

(;

2.

0
31,200
31 2 0

0 1.0

0 0 1.0 0 0 - .025 .025

- .5
0

The matrix of was

~ember

thermal distortions
40 40 20

~1~

0 0 - .025 .025 0

0 1.0 0 0 - .025

o o o o o
o

MultiplYing out per eq. 18,

13.8 172.2 13.8

Here, for example AlT = 2 x 2 x 10 = 40 (doubled once because ql acts on two stringer ends and doubled again to account for the other halt or the panel).
", ""'"

~r~
and

= 1

[2,2 13.8 0

86.1

1~'8J
.025 o

-.025

~
",,02

1
(bl

-.005

-."'-'

"1I
"

ao

PCyp
T

1=== 60"

F=

I-

=
'i.
FIg. AB.44

Then the redundant equations, per eq. (36), were

13.8

The structure was three times redundant. Stringer loads q~J qo and qo were selected as redun<1ants. Fig. A8.45 shows the unit redundant load Sketches for q. = 1, qo = land qo = 1.

172.2
13.8

n3lf
~

Solving,
. 1082 } '\= a ~ S .?9944 { } { ."C02

n~

Finally, the complete therTal stress distribution 'NaB

is 16}1 0&<I.1n~ fi.

Fig. AB.45

0~

:::: x 10

-.

- 54.1 - 49.7 - 50.1 -

108.2
99.44

100.2 - 2.70 0.219 - 0.019

A8.38

STATICALLY lNDETERMrnATE STRUCTURES

The result compares favorablY ~lth "exact" solutions made under the same assumptions (ref. NACA TN 2240) as far as the stringer loads are concerned. The shear flow result is not very satisfactory, primarily due to the use of too few "bays ~ in this analysis. Example Problem 21 The uniform fOUT-flange box beam of Fig. A8.46a is to be analyzed for the thermal stresses developed upon heating one flange to a temperature T, unt rcr-a spanwt se , above the other three flanges.

l
L/Z

1
L/ Z

Liz
1

~IJ

1\
).

,. "', ' ~~ ~ ,\~,. "\~''''

~\
\

Liz

ILiZ
Liz

I,

L/2

I
Liz L/2

1
L/2

, ~:,.., \1 --L
(aJ

~)

~':\~, ~'~

~,~~- q\~\~

Mult 1plying out per sq. (16) ,

.500.. B:33 i:"iC7 .
l.k~(b'::)

",~,,- '\~~'

:\~, ~

~'

.625

.375 -83
.375

.1:5 .l25

~.--I

Fig. AS. 46

\~'~ ~ ~ "\~
( 'j

.'25

[iki''' ...

.500

~'!J .

'", t

.s-e
.125

.37'
. 125

~ ...::.::.3
.125

.:25

.JS33

For a spec !fie case let :-::'iLa,

Tr.e

Solution:

i~verse

was computed to je

The beam was df vt ded into four equal bays 1.070 .06587 - .5291 .2337 giving a tour times redundant problem. Four 1.368 .5291 .3613 .1018 self-equilibrating (zero-resultant) independent Gt stress distributions were taken as the unknowns, Grs::J L (0 oj - .2337 - .3613 1.497 - .1934 these being shown In Fig. A8.46(b). Such zero- .06587 .1018 - .1934 1.792 resultant strP.ss distributions are the only oneS possible in a structure having no applied loads. The matrix of member flexibility coefticMember thermal distortions (!Ii'!') were tents was tormed by collecting coefficients from the several members. Un1t redundant stress d1s- computed tor loads Q2. q .. Qa and qa and were tributtons were prepared, taking QJ., q.. Qa and collected from the one heated flange only. Q? as the redundants and setting these equal to unity successively. 0 (Note that if two ad.scc 2 jacent flanges are 0 .6emc" .1667lc" heated equally, one aus't c! .500 set the corresponding ~i a LT 2 ~lU' , (:,o .. e { .1667lc" 0 .667lc" .1667ic" 8 equal to zero; this be2 .eoc cause the virtuai loads 0 .1667lc" .667',(" , .1667lc" being of opposite Sign 1 .500 adjaCent flanges, the .1;;67k~ ( Virtual Ncrk ~ust cane ,"1 ) . where k 2 = Gt / AE (b c) MUltiply1n~ out:

~IT} ~

.""".

Since externaf Ioada are not to be applied it follows from statics that the gener-al.iaed forces for adjacent webs and cover sheets are equal, as are the loads in front and rear spar caps at any given station. Hence an economy of numbering in the generalized force scheme is possible. Much labor is saved in the handling of data when the same symbol can be employed on several members whose loads are known to be equa1.

[]rLl{"IT}=~
Then the solu-::ion to 3q. (36)

a T Sa

f}
i
5

w~s ~Titten

as

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AS. 39

"cut" ~king it determinate, after 'Nhich the temperature distribution is applied producing thermal deflections which, when multiplied out gave
4 ' 079 } 1.938 + c) .8562 { .2413 .

{qST} = - 16

Gt a L T

Ib

Finally the .conptece set at thermal stresses were (still for the case L3k 3 = 1)
4.079 2.039 L 1.938 3.008 L .8562 3.437 L .2413 3.558 L

at the external points (compare with eqs. (35) and (37) ). Simultaneously the redundant cuts will experience relative displacements. SECOND, the redundant cuts are restored to zero displacement by the application at redundant forces (thiS problem was solved in Art. A8.l3) .. The qs are given by eq. (36); they produce additional deflections at points m

The total deflection of point m is then

The result compares favorably with an exact solution (NACA TN 2240) insofar as flange stresses are concerned, an error of less than 1 percent being present at the root. Shear flow values in the sheet are less satisfactory due to the (relatively) crude assumption of constant shear in~each beam quarter.
A8.14 Thermal Deflections by- Matrix Methods.
A. STATICALLY DETElIMINATE STRUCTURES

But

Therefore

The problem of the thermal deflections ot a statically determinate structure was considered earlier in Art. A7.8 in non-matrix fon. It should be apparent from the derivation, that the matrix method presented for the calculation of redundant-cut thermal deflections may be applied equally well to the problem of computing the thermal deflections of external paints of a determinate structure. Thus
(37)

or,

where point

OwT
ffi,

is the and

the~l

deflection at external

[Gmi ]

is the transpose at [Gl~J'

The matrix quantity in parentheses is the total strain (thermal Dlus "mechanical"). ------Fa:r-detinite reasons the equation for thermal deflections has been lett in the form at eq. (38) rather than the more polished forms which might be obtained by substitution from eq. (36), FIRST, the qjT' the thermal stresses, will probably have been solved tor preViously ~~d will be readily available in explicit torm. SECOND, and far more UnDortant from a labor saving standpoint, the unit load distrioution [giml (Whose transpose is used in sq. 38) ~y be any convenient stress distribution satiSfYing atatrcs-rn ~s~t of "cut" structures. One need not even use the distribution (and same

the unit-applied load stress distribution (compare with eq. 35).

8 REDUNDk'lT STRUCTURES

In the case of the redundant structure, additional strains are present due to the the~l stresses set up; the effect of these strains upon the deflection of external paints must be included in the calculation. The appropriate equation is most eaSily derived by VisualiZing the action in two stages. FIRST, the redundant structure is

Bam0l[ilm]

choice of "cuts") as employed in the redundant thermal stress calculation; a more convenient choice of cuts may be employed: In prinCiple, any stress distribution statically equivalent to the LU1it aprlied load(s) may be '..l.sed for [gimJ !!l2.S.. (38). (See p. A8.9). --

A8.40

STATICALLY INDETERMINATE STRUCTURES

_ a L 0 To

Example Problem 22 Co~pute the rotation occurring at the right hand end at the beam ot Fig. A8.43. Solution: To compute the rotation a unit couple was applied (positive counterclockwise) at the right hand end. An additional generalized force, called qa, also was added at that point. Then,
, _ ct,

1.2

!.J
aL:ST o
h

- .289

It is apparent that the values of the redundant ~oments ql, qa could have been chosen arJ1trar1ly in {g1~}' above, with8ut af~ecting

"5T--(-)
h

g 0To (3

20 T 0)_-

-4 a L 15 To
9h

the result.

Here, clearly,

{gi;n} ror any "cut "

and, uSing some results tram Example Prob. 19,

structure visualized w1l1 lead to the same result, qa being equal to ~ity in all cases. Example Problem 23 compute the axial ~ovement of the tree end of the central stringer of the panel of Fig. A8.44. Solution:

L 15 To

9n
2, 5

Note that q~ and q., the intermediate support reactions, were omitted from consideration. They do not enter into any expression for the internal virtual work or the structure; or, equally, they are not used to describe the strain energy of the structure. Hence they are not included in writing the total str~in. (Their 6 are zero.) -iT The member flexibility ccefficient matrix was

An additional generalized ~orce, qlOJ was added aXially to the free end of the central stringer (ret. Fig. A8.44b). T~en
l"n.13

, ..., ... ..., ...,

177 .5

88.8

! , ,
88.8 22.2

,
i

I
I

I
I

I
!2.2.2
I

,22.2 88.8 22.2 22.2 44.41

I
31,2001 !31.20:> !
I

!
i

!
,

i=1,2,5 From Example Problem 19, the true thermal stress distribution was
EI a 0 h

2.2.21

I ,

31,200
44.':'

40

40 20

{ q! } =

To {.267} .933
o
!

1, 2, 5

(qa was zero by inspection) To determine [gi~J' a unit couple qa

=1

o o a o o o o

Taking as the Wcut Wstructure one Where q. = qa = 0 one has Simply

was applied.

Using the qjT as obtained in Example ?roblem 20 (With qlo = 0), the follOWing product was formed:
-11.81

-13.45 - 6.655
11.81

Then

SUbstit~tlng

into eq. (38),

9h {Omr } ts 0!J(~.

{'} ....La 1n, "0 0T ,{.257})

6 40 6EI 14 012 h .933 0

13.45 6.656
-84.24

x aT

6.833
.5 928

2.402

ANALYSIS AND DESIGN OF I'hen , subs t t tut t ng into eq , (38)

J

FLIGHT VEHICLE STRUCTURES

A8.41

a unit load qlo

of

=1

gives (writing the transpose

5",

aT

L&nJ

28.19

{glm})
.50 .50 0 0 0 .025 0 0 1~

26.55 13.34
11.31 13.46
6.656 -84.24

L&nl~ = L 50
Then

O:.r

~ultiplying

= 34.3 a T

out for the thermal deflection: (same answer).

6.833

.5928 2.402 It remained to find the determinate stress distribution for Qlo = 1. As the determinate distribution the

It is apparent ~rom the above result that the Simplest deter~inate ("cut") structure should be used to compute [glmJ; it is completely adequate.
CLC6URE

{gim}

stresses due to a unit load Q~o = 1 were computed in the "cut" structure obtained when q, = qll = q:s = O. (Note that this is a different choice of redundant :uts from that employed in computing the thermal stresses in Exa~ple Problem 20.) rnus

o o
1
1 1

o
o
1

and finally,
o",T = aT

LO 0 0 1

1 1000 1J

28.19

The general the~al stress problem is complicated by the fact that the material properties E, G-and a vary with temperature. The problem created ther-eby is primarily one of bookkeeping - co~puting the member fleXibility coefficients (U and the member thermal distortions i j) (d for a structure whose properties vary from iT) point to pOint with ths temperature. The variations of E, G and a with T Will, of course, have to be known from test data. Two additional complications, not considered here, are the lowering of the yield pOint with heating (and the attendant increased likelihood at developing inelastic strains) and the phenomenon of "creep" (the time-dependent development of inelastic strains under steady loading). Should it prove necess~ry to analyze for the~l stresses under ~ore than one temperature distribution, the me~ber thermal distortion matrix { d may be gener'a Ltzed eaa i Ly into a 1T} rectangular form such as

26.55
13.34

11.81
13.46
6.656 -84.24

where d

6.833 .5928 2.402

34.3 a T
(AJlSWER)

is the member thermal distortion asiR sociated with rorce qi from thermal loading condition R. The matrix [C f j ] of ~ ther.na.l

As a matter of interest, a different unit stress distribution was employed With a different choice of "cuts". If the forces q., q , and qll are set equal to zero ("cut"), app Ltca't t on of

coefficients consists of the constant coef~iclents in the d expressions preViously presented. wr-ile iT T would be the temperature associated with qj jR for condition R. (compare with eq. (26b),
Art. A7.111.

A8.42 A8.lS Problems.

STATICALLY lNDETER1vfiNATE STRUCTURES

Note: Problems (1) through (9) below may ba worked by either the Least Work or DQ~Y Unit Load Methods. The student will be well advised to try some problems both ways tor comparison.
5000* A

(5) For the "King ~ost" truss in Fig. e, calCUlate the load in member SO. Members AS, BC and BD have area of 2 sq. in. each . ':':-le continuous ~ember ADC has an area of 9.25 sq. in. and moment of inertia of 21~ in~. E 1s same for all members.

r"'"""'----7i
100"

A
20"

..........

"I
jA I60

4000*

[-.10'j 60

f-

100" Fig. b

Fig. a

+
Fig. g

-j

C,

(1) Determine the load in all the members

Values in ( ) on members represent the cross-sectional area in sq. in. tor that member. All members at same material. (2) For ~he structure in Fig. (b), determlne the load in each member tor a 700# load at joint B. Areas at members are given by the values in ( ) on each member. All members made at same material.
60"

at the loaded truss shown in Fig. (a).

500.

(6) In Fig. f, AS is a steel wire both 0.50 sq. in. area. The steel angle frame C3D has ~ 4 sq. in. cross section. Determine the load 1n member AB. E = 30,000,000 psi. (7) In Fig. g find the loads in the two tie rods BD and CEo lac = 72 in." ; Abd = 0.05 sq. in. Ace = 0.15 sq. in. E 1s same for all members.

1)

30"

B
I

Fig. c

IE
1000* 1000.

~.r---aD9ct6~" IF:oII
180"
Fig. i

(3) For the loaded truss in Fig. c, determine the axial load in all members, Values in parenthesis adjacent to members represent relative areas. E is constant tor all members.
C

so.ooos
Fig. h

'F-100"-I

1-1

Fig. d

(4) For structure in Fig. d, calculate the axial loads in all the members. Values in parenthesis adjacent to each member represent relative areas. ~ 15 constant or same for all members.
5000*

(8) For the structure in Fig. h, determine the reactions at points A, B. M:embers CE and ED are steel tie rods with areas of 1 sq. in. each. Member AB is a wood beam with 12" x 12" cross section. Esteel = 30,000,000 ps1. Ew-ood = 1,300,000 psi. (9) For the structure in Fig. 1, determine the axial loads, bending moments and shears in the various members. The structure is continuous at joint D. Members AB', 8C are wires. The member areas are AB = 1.2, Be = 0.6; CD = 6.0; SDE = 10.0. The moment of inertia for members CD = 60.0 in."; for SDE = 140 In.'' (10) Re-so1ve Example Problem 10, p. AS.15 using as rsdundants the two restraints at one end (couple and transverse force). Solved in this way the problem is doubly redundant as no advantage 15 made of the symmetry of the st.r-uc'tur-e, (11) Add ~NO additional members) diagonalS FE and ZC (each with areas 1.0 ina) to the truss of Fig. A7.85, Chapter A-7. Find the matriX of

~
Fig.e

I--

l~iI7T~ ~_ t
120"

120"

--j
"'.'~'~'"":"

'.

~"._:_.-,,-: . . . . ' .

,.

T'

~_ "

_..

,._

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A8.43

influence coefficients.

problems of Art. AS.S by 'Nritlng

.ane ,
21.8 27.2 68.5 65.4
26.4-

27.5 85.4 68.4 26.2

18.3 26.4 26.2 20.6

[Amn] = E

27.2 27.5 18.3

where A1 1s the initial imperfection associated with force qi' Refer to the argument leading to eq. (11) or Art. A8.8. (16) USing the equation of problem (15), above, re-solve Example Problem 6, p. AS.l4.
(17) Using the equation or problem (15) above, re-solve Example Problem 7, p. AS.14. (18) USing the matrix methods or Art. A8.13, re-solve Example Problem 9, p. AS.15.

(12) Re-solve the doubly redundant beam of Example Problem B, page A8.3 by matrix ~ethods. The redundant reactions should be given ~q~ symbols. (See Example Problem 13a, page A8.20). (13) Re-solve Example Problem 5, page A8.12 by matrix methods. For Simplicity, make your choice of generalized forces including those designated as X and Y in the example so that Figs. A8.21 and A8.22 can be used to give the gir loadings. (14) By matrix methOds re-solve Example Problem 4, p. A8.12 using 3 equal bay diVisions along the panel (3 times redundant). Use the same structural dimensions as in Example Problem 20, p. A8.36. Campare the results with those obtained from the formulas developed in Example Problem 4.

(19) For the dOUbly symmetriC four flange box beam at Example Problem 15, p, A8.24, determine the redundant stresses qa, q7 and q~. if one flange is heated to a temperature T, unitorm spanwlse, above the remainder of the structure.

nne,
EaT
20Bl}

1599 : : } = - 10' { { 741 qu

REFERENCES (lS) Show that the matrix equation sq. (21) 1s modified to cover the initial stress See references at the end of Chapter A-7.

Douglas Dc~a airplane. Ph.otograph. showing simulated aerodynamic load being applied to main entrance door of fuselage test section.

, C f. ( 'j

A8.44

PROOF TEST OF DOUGLAS AIRCRAFT MODEL DC-6

DOUGLAS DC-8 AIRPLANE.

An outboard engine pylon mounted on a section of wing for static and nutter tests. The steel box represents the weight and moment of inertia of the engine.

CHAPTER A9

BENDING MOMENTS IN FRAMES

BY

Al~D

RINGS

ELASTIC CENTER METHOD

A9.1 Introduction

Assumptions In the derivations Which follow the distortions due to axial and Shear forces are neglected. In general these distortions are s~ll compared to fra~e bending distortions anG thus the error 1s small. In computing distortions plane sections are assumed to remain plane after bending. This is not strictly true ~ecause the curvature of the trame Chan~es this linear distribution of oendtng stresses on a frame cross-section. (correc-, tions for curvature influence are given in Chapter A13. Furthermore it is assuned that stress is proportional to strain. Since the airplane stress analyst TIust calculate the ultimate strength of a frame, this assumotion obviously does not hold with heavy frames ~here the rupturing stresses for the frame are above the proportional limit of the frame material. This Chapter will deal only with the theoretical analysis tor bending moments in frames ar.d ~ings by the elastic center ~ethcd. FYactical questions of bOdy fra~e design are covered in a later c~apter. The followin~ photographs of a ~ortion ot the structural framin~ of the hull of a seaplane illustrate both light and heavy fra~es.

In observing the inside of an airplane fuselage or seaplane hull one sees a large number of structural rings or closed frames. Some appear quite light and are essentially used to ~lntaln the shape of the bOdy metal shell and t~ ~rovlde stabilizing supports for the longitudinal shell stringers. At points where large load concentrations are transferred between body and tail, wing ~ower plant, landln~ gear, etc., relatively heavy fra~9s will be observed. In ~ull construction, the bottom structural framing transfers the water pressure in landln~ to the bottom portion of the hull fraToas which in turn transfers the load to the hull shell. In general the :~ames are of such Shape the load distribution of such character that these fra~es or rings l~~dergo bending forces in tranSferring the applied loads to the ot~er resisting portions of the airplane bOdy. ~hese jending fcrces produce frame stresses in gener~l which are of major importance in the strangth ,roportioning ot the frame, and thus a reasonable close apprOXimation of such bending forces is nec~ssary.
an~

Such fra~es are statically indeterminate relative to internal resisting stress and thus consideration ~ust be given to section and physical properties to obtain a solution of the distribution of the internal resisting forces. Jeneral
~ethods

of Analysis:

~rinciples

There are ~ny ~ethods of applyin~ the of continuity to obtain the solution for the redundant forces in closed rings or frames a~d bents. T~e author prefers the one Nh~c~ is g8ner~lly referred to ~s the ~Z18stiC :::2::t~r~ me thod anc :,"':.s ' . . sec. it for -nany year-s ~n ~Q~ti~8 ~ir~l~"e :~si~n. The 7.et~od was Qri~i ~atec by ~uller-Breslau~. T~e main cifference in this ~et~od as som~ared to Tiost other methods of s~lution is that the redundant forces are assu~~i ac~ing at a s~8cial :oint called th~ elasec...a t Lcns for t ne r edunoant.s Nhic:l a re tnc ~e~d2r.t of eac~ other.
v-

'~MuUer.Breslau, H., Die Neueren Hethden der

FestigkeHslehre und der Statik der Baukonstruktionen. Leipzig, 1886.

A9.1

A9.2

BENDING

MOMENTS

IN

FRAMES

AND

RINGS

Consider a small element ds of the curved beam as shown in Fig. A9.2. Let M s equal the bending moment on this small element due to tee given external load system. The total jending moment on the element ds thus equals, - -(1) (Moments which cause tension on the inside fibers of the frame are regarded as ?ositi ve moments. ) The following deflection equations for point (A) must equal zero:Q = a Ilx -= (, Ily

A9.2 Derivation of Equations. UnsymmetriCal Frame.

= a:

(angular rotation of (A) = zero) (movement of (A) in x df r-ect t on -= 0) (movement of (A) in y direction = 0)

Fig. A9.l shows an unsymmetrical curved beam fixed at ends (A) and (8) and carrying some external loading p~, P a , etc. This structure is statically indeterminate to the third degree because the reactions at (A) and (B) have three unknown elements, namely, sagnt-. tUde, direction and line of action, making a total of six unknowns with only three equations of static equilibrium available.
p.

From Chapter A7, which dealt with deflection theory, we have the following equations for the movement of point (A):-(2)

Ilx=Z~~SOy
:=

-(3) -(4)

Z i1mds
1

P,

\
A

t
Fig. A9.1
ds

P,

I
B

In equation (2) the term m is the bending moment on a element ds due to a unit moment applied at point (A) (See Fig. A9.3). The bending moment is thus equal one or unity A~ for all ds elements C unit moment of frame.
Fig. Ag.3

..J I.-

P
B

Then substituting in equation (2) and using value of M from equation (1) we obtain g := .J"!sds + MAZ l;d _ XAZl.yds + YAZl.XdS = 0 ~ ... 1S EI EI

FIg. A9.2

whence,

In Fig. A9.2 the reaction at (A) has been replaced by its 3 components, namely, the forces X and Y and the moment !1 and the structure A A A is now trea~ed as a cantilever beam fixed at end (B) and carrying the redundant loads at (A) and the known external loading p~, P3' etc. Because jOint (A) is actually fixed it does not sutfer translation or rotation when structure is loaded, thus the movement of end (A) under the loading system of Fig. A9.2 must be zero. Therefore, three equations of tact can be written stating trAt the horizontal, vertical and angular deflection of ,oint (A) must equal zero.

In equation (3) the term m represents the bending moment en a element ds due to a unit load applied at point (A) and acting in the x direction, as illustrated in Fig. A9.4. The applied unit load has a positive sign as it has been assumed acting toward the right. The distance y to the ds element is a plus distance as it is
y

l~~B ,
Y

Fig. A9.4

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A9.3

measured upward from axis x-x through (A). However the bending ~oment on the ds element shown is negative (tension in top fibers), thus the value of m ; - (1) y = -yo The minus Sign is necessary to give the correct bendir.g moment Sign. SUbstituting in Equation (3) and using M from Equation (1):-

place of XA, Y A respectively. A and M The axes x and y through the point (0) are centroidal axes for the values ds/EI of the structure. This fact means that the summationsZ Er=
The

ydS

and

expressions Z xds/EI, Z yds/EI and

Z xydslEI also appear in equations 6 and 7.

In equation (4) the term m represents the bending moment on a element ds due to a unit load at point (A) acting in Y direction as illustrated in Fig. A9.5. Hence, m l(x) x SUbstituting in Y ds equation (4) and USing I'! from equation (1), we lj;obtain,

These terms will be referred to as elastic moments of inertia and product ot inertia ot the trame about y and x axes through the elastic center of the ~e, and for Simplicity will be given the following symbols.
Z

x---~
Fig. A9.5

"""'EI = I y ,

xds

"E'r

yiJds

= Ix

, ""

<;'

xyds = EI Ixy

Equations 5, 6 and 7 will now be rewritten USing the redundant forces at point (0).

hence,

- - -(7)
Equations 5, 6, 7 can now be used to solve for the redundant forces MA' X A and YA. With these values known the true bending moment at any point on structure follows from equation
(1)
- - - - - - - - - - (6)

REFERRING REDUNDANTS TO ELASTIC CENTER

For the purpose of simplifying equations 5, 6, 7, let it be assumed that end A is attached to a inelastic arm terminating at a point (0) as illustrated in Fig. A9.6. The ~oint (0) coincides with the centroid of the ds/EI values for the structure. Reference axes x and y wIll now be taken with point (0) as the origin. The redundant reactions will now be placed at paint (0) the end of the
Y
I

The term ~Z ydslEI is zero since Z yds/EI is zero, thus M o drops out when substituting in Equation (6).

- - - -(10) The term Msds/EI represents the angle change between the end faces ot the ds element when acted upon by a constant static moment Ms. This angle change which actually is equal in value to the area of the Ms/EI diagram on the element ds will be given the S~bol 0s ' that is, 05 : MgdsIEI. ~ith this symbol SUbstitution, equations 8, 9, 10 can now be rewritten as follows:------(ll)
- - -(12)

elastic

Fig. A9.6

inelastic bracket, as shown in Fig. A9.6. Since point A sutfers no movement in the actual structure, then we can say that point (0) must undergo no movement since (0) is connected to pOint (A) by a rigid arm. Thus equations 5, 6, and 7 can be rewritten USing the reduncarrts Xo' Yo' and Ma in

Z 0s x - XoIxy + YoIy

=a

- -

- -(13)

Solving equations (12) and (13) tor the redundant forces Xo and Yo we obtain,

A9 4

BENDING

MOMENTS

IN

FRAMES

AND

RINGS

"'~S

IX (I

\ y -~)

"'~S

MX(~\ ---r:-:-)

Table A9. 1

- - - - - - -(14)

TXIY

y= x= Por- w=ds dist. dist. tion 1 ).05. from x -x y-y

- [Z0s X

60 sY

(!t;-)J
)

AB 3~:10 15
- - - - - -(1,0) Be

Y e

I
y

(I _ I 'x?

wy2= 2250 wx 2=1440 12 -120 +150 i = 30 3 ::750 i - 30 - 0 x y-nz312x3 wy2::10800

TXTY

A9. 3 Equations for Structure with Symmetry About One Axis through Elastic Center.

~12 2

30

360

wx 2 : 0 . - 24 3 ~6 i x ::.-li 3= 0 'v-_,, 12x2 . 12x2 I wx 2:1440

If the structure 1s such that either the x or y axis through the elastic center 48 a axis

CD 30=iO 15 3
Sum 32

wy2'::2250

12

120

150

Ix:

750
16800

iy=~3=0 12x3
3456

of symmetry than the Droduct of inertia Zxyds/EI = I xy = zero~ Thus making the term I xy = 0 in equations 11, 14 and 15 we obtain,

660

=-z s

tds/EI

- - - - - - - - - - - - -(16)

-r;Yo -

_ Zsy

- - - - - - - - - - - - - -(17)
- - - - - - - - - - - - - - (18)

_-Zsx

-r;;y

The terms i x an1 1 are the elastic ~OT.e~t of inertia of each portIon of the fra~e about its centroidal x and y axes. Si~ce I is constant over each pert ian the centroidal T.oment of inertia of each portion ~s identical to that of a rectangle about its centroidal axis. To exp'ln In for member AB:- ~ '_ ll-----O Re re rrv.ng to Fig. a,
ix
1 bh3 1 =12 = 12 I x x3 30~

A9.o1 E:m.mple Problem Solutions. Structures with at least One Axis of Symmetry.
,Exa~ple ~oblem

Fig. A9.7 shows a rectang-

ular frame with

!lxed supports

at ;:JOints A and
D, and carrying

a single load as
Shown. The

B!7:JC I
~'~18"_
10 lb.
30" I

liT
-IX I
!
I

+=t=

= 7:': _,,0,0

<-,I
A

,ds' 30' h

Rf e erring to Fig. b ,

uI '
Y

--1-

Fig. a

ly

= ~bh3 =.l:.x ~o 1 :;:: 12 12'" x3!'

.09 (negligible)

problem Is to determine the bending moment diagram under this loading.

=3

The distance from the two reference axes to the elastic cer.ter can now be calculated:y

8T
!

A X - 'fT11"TT] - - -

IY -

D - - rrrtrrrr-

I
I

ds=30=b

Fig. A9.7 T'ne r irst; step in the solution is to find the location of the elast1c center of the frame and the elastic moments of inertia Ix and I y .

=~ Zw

= J::N

Due to symmetry of the structure about the

Y axis the centroidal Y axis is located midway

between the sides of the frame, and thus the elastic center (0) lies on this axis. Table A9.l shows some of the necessary calculations to determine the location ot the elastic center and the elastic moments of inertia. The reference axes used are x-x~
and

Having the moment of inertia about axis x-x- we can now find its value about the centroidal axis xx ot the frame, by use of the parallel axis theorem. Ix ~ Ix - ~w(y~) = 16800 - 32 x 20.625~= 3188

: 660

32

= 20.625 In. =
0

0 32

-oJ yl ~.!=.!= h
- I ,

U...l.
I

Fig. b

I y : I y - "w(x')

= 3456

- 32(0) : 3456

s-s,

The problem ~ow consists in solving eq~ tions (16), (1:) and (18) for the redundants at the elastic center, namely

..

--~

-~-

..

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A9.5

_ ~ _ Area of static Mil diagram o - ZdS!I - Total elastic weight ot structure M Moment of' static Mil diagram about
X - ZTsY -

along the neutral axis of the frame members. In Fig. A9.10 the area at the Ms/I diagram equals 0 s = 22.5 x 24/2 = 270. The centroid by siillPle calculations of this triangle would fall 10 inches from B. Fig. A9.ll now shows the irame With 1ts MslI or lts 03 load. Os is 0 s = 270 -l' .. positive since I1s is B-...;;:,:-2 C pos rt ive . The next 1375" step is to solve the x---- ..:x equations for the 0: r-edundant at the : 20.625" elastic center. The ; I signs of the distances A 'Y D i x and y from the axes x and yare conven-

o -

x - Elastic moment of inertia about x

ax1s

x ax1s

Yo -Zsx Iy

Moment of static Mil diagram about y axis ==""'-==::'-::T;~="'--;=::'-:; Elastic moment ot inertia about y axis

Thus to solve these three equations we assume a static frame condition consistent with the given irame and loading. In general there are a nlli~ber of static conditions that can be chosen. For exanple in this problem we ~ight select one ot the statically determinate conditions 11l~strated in Fig. A9.8 cases 1 to
~ust

' fl
I

Fig. A9.1l

t i ona L,

,.
I

Thus,
= 32 (from Table A9.1 = -8.437 in. lb.
- (270)

~I ,
-:-:T1

, 10

Xo

= ZsY = 270(9.375)
Ix
3188

=0.7939
=

lb.

Case 1 Fig. A9. S

--ZTsx _ -(270)(-2) Yo y 3456

0.1562 lb.

Fig. A9.12 shows these values of the redill1dants acti~g at the elastic center.
~17.

To illustrate the use at different static conditions, three solutions will be presented with each using a different static condition. Solution No.1 In this solution we will use Case 3 as the static fr&~e concition. The bending moment on the fra~e for this static trame condition Is given in Fig. A9.9. The equations

75

;:8

9.375

20.625"

t -F'9
8.437

):1562

B[rig.
7.5#

~ s diagram 22.5 Ms/l diagram

L06'AJ
Fig. A9.12

~~.81

J
6.06
FIg. A9.13

lo. ,e
A9.9~

Dol ,""'A , ,
'2.j~

Bn

~81

Ms

is posrnve ,

tension

on inside of frame.

The bending moments due to these redundant forces will now Je calculated.
M A

=-

8.~7

- .1562 x 12

Fig. A9.10

.7939 x 20.625 :;: 6.06 in.lb.

Me

for the redundants require 0s the area of the ~s/I ~lagra~. Fig. ~9.10 shows the ~s/I curve which is obtained by divlcing the values in ~lg. A9.9 by the te~ 2 which is the moment of inertia of ~ember Be as given in the problem. Since the equations for Xo and Yo require the moment of the Ms!I diagram as a load about axes trIough the elastic cencer of frame, the area of the Ms/I diagr~n will be concentrated at the centroid of the diagra~ and along t~e centerline of the frame, or more accu~telY

=- 8.437 =- 9.437
=8.437

- .7939 x 9.375 - .1562 x 12 = -17.75 tn Lb.

s

Me
NO

- .7939 x 9.375
+

.1562 x 12 :;:
-14.00 In. lb.
+

.7939 x 20.625

.1562 x 12

9.81 In. lb.

These resulting values are plotted on Fig. A9.12 to give the bending ~oment diagram due to :he redundant forces at the elastic center.

A9.6

BENDING

MOMENTS

IN

FRAMES

AND

RINGS

Adding this bending moment diagram to the static bending diagram of Fig. A9.9 we obtain the final bending moment diagram of Fig. A9.13. The final bending moments could also be obtained by substituting directly in equation (1) ua tng subscript (0) instead of (A). ThUS,
M= M s + M o - Xoy
+

= -Z0s x
ly

= -[-45(-10)-4D5X6-300(-12)-90OX12]
3436

= -(-9180) = 2.656 lb. 3456

The final ~cments at any point can now be found by equation (19). Consider paint 3:lis = -30 from Fig. A9.14 x ~ -12, y ~ 9.375
suot , In (19)

YOX

- - - - - - -(19)

For example, determine bending moment at pOint

a.

For point B, x = -12 and y substituting in (19)

liB

= 9.375,

M s

=0

=0 +
D.

(-8.437) - .7939 x 9.375 + .1562 (-12) = -17.75 as previously round

x

liB

= -17.75 in. lb. which checks

solution.

=-30+51.56-.7939X9.375+2.656(-12)

fi~st

AT

~OINT

= 12,

= -20.625,

M s

=O.

Consider point D:-

liD = 0 + (-8.437) - .7939 (-20.625) + .1562 x 12 9.81 in.lb.

M s = -90, x = 12, Y = -20.625

Subt. In (19)

Solution No.2 In this solution we will use Case 4 (See Fig. A9.8) as the assumed static condition, that is two cantilever beams with halr the exter-ual load or 5 lb. acting on each cantilever. Fig. A9.14 shows the static bending mcment diagram and Fig. A9.15 the Ms/I diagram.

MD

=-90+51.56 = 9.80 in.lb.

7939(-20.625)

2.656 x 12

Solution }Io. 3
In this solution we will use Case 5 (Fig. A9.8) as the assumed static condition, namely a trame with 3 hinges at points A, D and ~, as illustrated in Fig. AS.16.

"5=-45

"~-405

-45

-"62515. 6~5
Fig.A9.14

-1"" ~18"--I

Ill<

~:- 30P
I I

T'::

-'-

r- 12"
t-x
Fig. A9. 16

.08=- 00

30"

Betore the bending moment clagram can be calculated the ~eacticns at A and Dare necessary. To tind VD take moments abou~ point A.

-30

90

!1
-10
Fig. A9.15

-30

Fig. A9.l5 also shows the results or calculating the ~s value for each portion of the Ms/I diagram and its centroid location. Substituting in the equations for the redundants we obtain,

VD=2.5

HD"'l

ZIlA

= lOx6 =0

2~VD

hence, Vo = 2.5

I'1o

=-Z0, Zds I

= - (-45-405-300-900) _ _ ,
3.2

- 01....,6 in .u ,

=0

To tind ~A t~ke ZFy hence VA - 7.0.

= 0 =-10

+ 2.5 + VA

X - Z0 sY - (-45-405)9.375+(-900-300)(-5.625)
o - ~ -

3188

To find HD take moments about hinge at E at all rorces on frame to right side of E and equate to zero. ZME

=0.7939

lb.

= -2.5

x 12

+ 30HD

= 0, hence nD = 1.

Then USing ZFx 0 for entire frame, we obtain ZFx = -1 + HA = 0, hence HA = 1.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A9.7

The ~rame static bending noment Q~a~ram now te calc~lated and dra'NTI as shawn i~ Pig. A9.17.
~an

axes through the fr~e elastic cer.ter are then calculated as -12 ~d -0.625 inches ~espect ively.
; - E0s _ - (-390) ; 12.19 in. lb. Eds/I 32
Xo ; Z0sY Ix Yo

(5)

-30 -30

9.375 \

T
I
I

t\I
I

1[1:1 .- -

,,0s'-150I

I"

x = 12

...J

,y=0.625'

~:i;c:n~er
i
I

"'l50

(6)

-548 3188

= -0 203

Ib

c ,

= -Z~;X

; - t.;~0) = 0.1562

lb.

The final ~oments at any point can now be found by use of equation (19), namely
20

20.625

- - -(19)

Consider point B:y

x
Fig. AS. 17
D
Pin Pin

= -12,

= 9.375, Ms

= -30

SUbstituting Mg

= -17.76
tions)

=-30+12.19-(-0.203)9.375+0.1562(-12)

in.lb.(checks previouS solu-

The Qoment diag~~ is l~beled in 6 parts 1 to 6 as indicated by the values in the small circles on each portion. Most of the calculations frem this ~oint onward can te done conveniently in table form as illustrated in Table A9.2.
Table A9. 2

Consider pOint 0:-

x = 12,
Sub't , in (l9)

= -20.625,

M s

=0

MO = O+l2.l9-(-0.203) (-20.625)+0.1562x12 = 9.82 tn Lb , (checks previous solutions)

v

Mom. A' I Os = Dia- Area for gram of A Beam Por- Mom. Section I tion Dia.

x = dist. Y = dist. from Y from X axis to axis to C.g. of C.g. of

.
.

5,

A,y

The final complete bending moment diagram would of course 08 the s~me as dravm in Fig. A9.13 for the results of Solution No.1. Example
?~oblem

A,
_12.00

Os
_0.625 9.375 9.375 9.375 9.375 -0.625 1800
93

2.

~450

3
2 2 2 2

-150

. 60
15
45

30 _10.67 7.5

320 -281 50 90 70
211

3
4

- 6.67
.
4.00 8.00 i2.00

22.5

5
6

-180 _450

- 90
150
~390

720 -844
93

-1800
I

rectan~~lar closed supported at pOints A and B and carrying the external loads as shown. The reaction at B due to rollers is vertical. The ~rame at po~nt A is continuous through the joint but the react~on is applied through a ?in at the center of 'the j oint . The pr obLem is to d et.ermtne the bending moment; c tagr-an. fra~e

Fig. Ag.1a shows a

Sum

- 540

-658

In order to take mo~ents of the ~5 values in cojumn (4) of the te.oIe , the centroid of each ~ortion at the diagran must be ce~e~ined. For example, the centroid of the ~wo triangular bending morr.ent corticns marked 1 and 6 1s .667 x 30 from the lower end or 20 inches as sho'Nn in ?lg. A9.l7. Thus the 1istanceS x enc y from this 05 location to the y and x

' I n
IB
,P1"2M

w = 10*.1 in.
j

L :: 30
y 4 I' 1

.I !

,L,L I

.2401

12'

I I

1,A
"'"

ii, 18 ' T - - - ~I .-C:- x~ y=9f6;'L=30 -~E. z 1 9.67 cut ,

~~;4
I

1=3
x

L'24I~2f)\
D

c:

\1 B

ii'~""""-----7

(2)

C 3

~240

I:2

Fig. A9.18

'""*'

.1.

Fig. A9.18 Ms diagram for load Pi

A\oo--15~

(5'

240 240D

I-I c
"~~""':'O:i'";;"

-: '~~:.-

A9 8

BENDING

:>10YlENTS

IN

FRAMES

, AND

RINGS

----<~4500
IE I I
1

T:,2 next 5,:e;:;

:s

". 00

C~.'J:Jse c'

5':;.<:1: .::':-:o..:T.e
.' 'If

,a
i

r , [E7.

(7)
,

out
A
D1

I I,

~-4500
Fig. A9, 19 diagram fa, w loading

--:- cut iA

~1
.
9

D 3b ...

360

M,

M,

Fig. A9.20 diagram for P2 load.

stc '::1'.:: \"5 j~:-;;i '.:.; moment ::2.::I""::.;7:. CC i s 5 ':=:L::. Lm ':7'.'::' frq-;:s _0 assur-o-r cut on membe r .'B ~ "':'3~ -,:-::;e -c i:::"': .. :3 10 l:il...ustr?..ted 1" ::.';s. "::' .... :'9 ar;c GiJ ?':r '-' J S1:::c:li:ity ::12 ,"._" me-tent c ur-ve :',2.5 cr-own in "'c:-... r-ee ~ar-:;5, ..' -'~ 83.C:". . ':::=:::5:::::'-2,l!1g only one of tns ... ext.er-nn l - ' t_",,,, i. -:;"'3.::'.5 C , 20 S:~:'N on the s t.r-uc tur e . ?L;s. A.~.::'2, .. ::- 9 ::J,c;;:snt C::'::;V,:;S. these resulti::~ bend tng po!'tions :J1' tnes a benc t ng ::lO::;:.en: ,j 1 '-':.:;-:':o..'~.s numuer-ec '- to 10 2.:-8 5 ['1 own tnc par-errtnes ~2 on each ;Jo:'':icn.

c cnc t t ion en- ce t erntne

-.

'::-,8

-;.~

' : ; - ' 0 " ...

~

\...

~o"',.,

... ~ . .
-~

"",.,..~
,,-~

.~,,-

,.,~o""

0"

, ,_.,

,~

SoL;t~:n:
T~,e first step is to find the j ocat i cn of the elastic center of the rr-ame . Due to sy.mnetry o~ rr-ane about the y ex.;s , 'the elastic center 1.'111 be on a y axis 'thr-ough tt).~ middle o~ the rraae . The ver-t.t ce.I d~s"tance y measured from a ax i s through AD equals,

:''..18 nex"': sts'J '" ;:,'ae s-:lutioD ::':'.3:'3t3 '::l~ r Inc tng the area of the M~/I 2i~;r~~s '"-7'.': 'tr;e first ::lO!1!en'C Gf t ae se j .: 3. ~ :2..i7_S 2.::'0'..1.': tne x 3.:'.::: y axes t nr-ougn t::e 8l3.S~~:: c ent cr-. T;;'858 s i:::~l calC'-llations can jest '.::e cone in tab l e "er-n 3.3 i Ll.uat.ra ted in ~2.ble .'.9.3.
Table A9.3
Portion Area of M, of Ms Diagram Portion ( ) , A
1
2

....

Os ,

dist. y

dist.
X

to Y
Axis

to

Z(dIS) Y y = =
Z

3x24~ 1'" ::> 4 3 x... ,::;X<::,

30+( 24)2.30 4 3, 2

os

372 33.5

= 9.67"

A 480

Axis

sx

I.

OsY
4958 25794 4958 2726

,
-15

1440 3 ~

10. 33 14.33 10.33

7200

i1 -

rhe next step IS to deter::l.ine the elastic moment of inert ta of the rr-azae a::: out x and y axes throug~ the elastic center of the f:-ame.
~oment

3
4 5 6

72004- 1800 1440314403 36002

4801 480

a
15 15
5

01
7200i-

- 5.67
, . 9.67

7206 9000

of inertia about x axis

= Ix
653.9

:-

1800

17406

Members

An
k

and DC,
14.33') 2

45000 4-112501

7.5
15 5 15
5

, Ix =( ~ x '3 X

14.33

7
8

-108000 3 -36000 67500 2-33750 32403 54002

1080 1

2.33 -540000

843751-161212

- 83880 i
,
326352 3064

(1'
Ix

'3

x 9.67') 2

=
= = =

2()0.9

9 10

- 9.67 I - 3.67

-168750
16200 1-

I
!

Member BC
(14.33') = (30/4) (30/2)(9.67")

2700 -77700

9.67

13500 -747225

26109 4645

1540.0
1402.7

Sum

Solving for the red1.U1dants 'It tte i:la;:-t1.c

Ix

3797.5

center,
,-

Moment of inertia about y axis

Me:nbers

=Iy
563

l'!:J

AD and BC.
30'
30 3

- .621<:: = -(-77700) = 20:;.2 =1:dS/T 38.5

-4645 sY -:;'.22 = Z0 Ix = 3797 =

in.la.

Iy

1 x = 12 x '4 1 1 '-2 x 2' x

= =

Xo
Yo

lb.

1125

= -~0'Jx Iy =

-(-74-7225)
~289

= 14.1.28
~.

:"b.

Members AB, CD,

24

acting at

=
'y

3600
5289

Fig. A9.21 shows these redundant r,JrC2S ;;":J" t:. 8 e las c ;c cen:er. . .. ';'9.22 . . j. "'-;-- am cue ",,0 -0 ::>hONS tne bend.ing moment c Iagr cnese
~

r-edundant for-cas. 1';:',8 calculations 'Nith r-er ar-, ence to fig. A9.21 are -

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

A9.9

Total elastic weight of ring

z ~ = It
I

(36) ~ 113.1

IT
1

9 67 .

4124

,'-------'r0
Fig. A9.21

The elastic moment of inertia about x and y axes through center paint of ring are the same for each axis and equal
Ix = I y =
nr~

= n x

18~

= 18300

M A M8
~

= 2018 = 201S
= 2018

14-1. '28 x 15

1.22 x 9.67
X

= -113
= -84 = 4153 = 4124

141.28
+ 141. 28

15 + 1.22

14.33

+ 141.28 X 15 + 1.22 X 14.33

MD

= 2018

+0

,-

1.22 X 9.67

Combining the bending ~oment diagrams of Figs. A9, 18, 19, 20 with Fig. A9.22 would give the true or final bending moment diagram. Example Problem 3. Circular Ring.

The next step in the solution 1s to assume a static ring condition and determine the static (Ms) diagram. In general it is good practice to try and assume a static condition such that the Mg diagram 1s symmetrical about one or 1t posSible about both x and y axes through the elastic center, thus making one or both of the r-edundarrt a Xo and Yo zero and thus r-ecuc Ing considerably the amount of numerical calculation for the solution of the prOblem.

Fig. A9.23 shows a circular ring of constant cross-section sUbjected to a symmetrical loading as shown. The problem is to determine the bending moment diagram.
50 50.

In order to obtain symmetry of the Ms diagram and also the Mall diagram since I is constant, the static condition as shown in Fig. A9.24 is assumed, namely, a pin at (e) and rOllers at (f). The static bending moment at
points (a ) , (b), (c) and (d) are the same magnt tude ann. equal,

Ms =50(18

- 18 cos 45 0)

=265

in.lb.

The sign is pOSitive because the bending moment produces tension on the i~lde of the ring. The next step is to determine the and :2l s Y values.
~s, ~sx

s 1s the area of the Ms/I dtagr-am, however since I is unity it is the area of the ~ diagram. The static Ms diagram of Fig. A9.24 is divided into similar portions labeled (1)
Fig. A9.23 Fig. A9.24

and (2).

Hence

Fig. A9.25 Moments due to redundant forces.

0'

I~ i0~rack:~r~ ~\\ ',

"~MO"_177". -;
,

M=88"*

0s(~) = area of portion (1) where P = 50 lb. and a = 45.

= Pr~(c

-sin

~),

---+-Xo=CX:
I~~

M=-177"# -1~

~
\~

SUbstituting and multfplying by are four portions labeled (1).,

since there

05(.1 ~ 4[50 x IS" {0.785 - 0.707)J ~ 5052

The area of portion labeled (2) equals,

'

.,~

M=88"#

"

-,....-,_.-,c-,-;"o:
Fig. A9.26

Prae(l - cos a).

Since there are two areas (2) we obtain,

Solution. Due to s~etry or the ring structure the elastic center falls at the center of the ring. Since the ring ~as been ass~ed with constant cross-section, a relative value of one will be used for 1.

05 {" )= 2 [ 50 X IS" X; (1-0.707IJ

Hence, total 0s = 15000
+

= 15000

5052 = 20052

Since the centroid of the Xs diagram due to symmetry about ~otn x and y axes coincides

( 11

A9.10

BENDING

MOMENTS
fra~e,

IN

FRAMES

AND

RINGS

with the center or el~stic center of the the te~s Zsx and Zsy will be zero.

Solution: The fi~2t step 15 to 1ete~lfle the elastic center of the f~ame and the elastic ~o~ents of inertia. ~able AG.4 shows the calc~latlons. A reference axis x'-x' has jean selected a~ the ~icpeint of the side AB. Since a static :r&T.e condition has been selected to ~ke the M s diagram sJ:mnetrical ~bout y axes t;Iough the elastic center (see Fig. A9.2S), it is ~ot necessary to determine I y si~ce the redundant Yo will Je zero due to this sJ~T.e~ry.
Table A9. 4

SUbstituting to determine the value of the redundants at the el~stlc center we obtain, Me

= :Zd0S" ~ ~
= Z0,y
Ix

= - (20052) 113.1 = - 177 in. lb.

= 20052(0) = 0
18300

Yo

=-l:0,x =-(20052)(0) Iy 18300

=0
acti~g

at the elastic center and the bending moment diagram prOduced by these forces. Adding the bending moment diagram at Fig. A9.25 which is a constant value over entire trame of -177 to the static moment diagram at Fig. A9.24 gives the tinal bending moment diagram as shown in Fig. A9.26. Example Problem 4. Hull Frame

Fig. A9.2S shows the values

Member
BDB'

Length cis I

w' "f

cis

'NY

IX'X' : tx .. wy2

94.25 1.5 62.8

t!: ..
n

.. 151200: 156600 30 3080 54.00

: 49. i
AB

60.0 60.0 38.4 38.4

1.0 60.0
I.

0 0
42 -42

0 0

18000 ... 0 18000 ... 0

,
,

18000 18000 34772

3477~

A'B'
AC CA'

60.0 19.2 19.2

221. 2

Fig. A9.27 shows a closed frame subjected to the loads as shown. The problem is to deter.nlne the bending moment diag~.
lJ
odl){)"

2. 2.

806 922 .. 33850 806 922 .. 33850

1488

, ,

Totals

262140

In the last col~~ of Table A9.4 the te~ l x is the moment of inertia of a particular member about its own centTQidal x axis. T~~s for member 80B;
D

static moment curve

z.~,

a x ;: .3!a = .3 x 30 /1.5 = 54C0

For members AC and CA ,

T A~ L'38.4

h'"''
Fig. A9.28

-L ~~~~.

~ ~:_jY

_ 1 tx - 12 b L h"

= 12x2"..... 1 1""3' '''04'-920 '-'. "A~

-

Let y = distance from X'X' ret. axis to centToidal elastic axis X-X.

"wy y=-zw=
Ix

221.2

1468

;z

6.64 in.

By parallel axis theorem,

= Ix!

- 6.64"(Zw)

262140 - 6.54"(221.2; = 252400

Fig. A9.29 Fig. A9.30

The next step in the solution is to ccmputs the static moment elastic weights 03 and their centroid locaticns. In Fig. A9.28, the static frame condition assumed 15 a pin at point A and rollers at point A' , which gives the

ANALYSI~

AND DESIGN OF FLIGHT VEHICLE STRUCTURES

general shape of static

in the f Igur-e ,
Cons i der- member- BDB
T~e
P

~oment

curve as shown

1/2
t

x=

10o

38 . 4.
(4690 x 100 x")xdx 78400

sl~ered

t.er-n

~3

bending moment curve will be con~ in two parts, namely (1) and (2). The represents the area of the Ms/I

= 1/2 [469~

x' . 10~ xj
784.00

38.4-

21.77'

c tagr-an.
Thus for pottion (1) and (1' )

o
Vertical distance from line AA' to centroid 21.77 x 24/38.4 13.~". The static moment weight for Ale' 1s same as for AC, thus

~3('l + ~s(") 2b"(d. sin 0] = 2@00~.~ 30"(0.524 - 0.5~ "172800

1he vertical dlsta~ce from the line BS! to centroii of !"'Is curve tor portion (1) and (1') Is,

=
~

~AC + ~A'C'

2 x 784.000

= 1568000

r{l ... cos a - - - - )

y ,

sin au
2.

30(1 0.867 -

a - sin u

0.524 - 0.5

= 10

in.

rig. A9.29 shoWs the trame With the moment weights a located at the centroidS, together WIth the redundant forces Me and Xo at the elastic center. It makeS no dltterence where the frame t s cut to form our residual cantilever, it one or the cnt races 1s attached to elastic center and the other Is considered fixed. With the elastic properties and moment wel~~ts known the redundants can be solved for:
_(,.=.10",Cf7=0:::00~+ .=1,;;72;:;8,.,0;;0...:+...:1:::5",6",80",0'-".) , 221.2
124.30'#
X :: o = ZX'sy I xx

For portion (2) of the M s diagram the area of the ~/I diagram Which equals ~s 1s
,05(.:1)

=PTrlitj

(1 - cos a)= 6000Xl~~"X2.1o._a.967)

= 1,007,000
Consider member AC,
~rom free body clagrarn of bottom portfon of frame (Fig. A9.31~ the equat~on for bending moment 6000 ;4690 6000

=,,10",Cf7=0"0"0-,x,-,48=.-=1,,6...:+~1,,7.;;2:;;8*00,,,,,xC'33=. 3"6::...:.+-=1"5,,6,,60,,,0,-x,,,--,,5"'. 2=4 252400

= -98.5#
a.~d

. . . . . . . . 11
A
I

--

A'
C

I~

Yo 1s zero because of the symmetrical frame loading. The f1nal or true bending moment at any point equals

>0.,"'7

Thus for point 8 on

~ember

AC equals:

MB = 0 + (-12430) - (-98.5 x 23.36) =

-10130'~

Mx :: 4690 x -100 x Q

Area of
aq\..ials,
1/2

~/I c~rle

between A a~d C when I = 2

For point C

Ms at paint C = 4590 x 38.4 - 20OX{38.4)a/2

%
o

38 . 4

= 32700
Hence Me = 32700 - 12430 - (-98.5 x -60.64)
14.300'~

= 1/2

[4690 x" _ 100 X

2 3

0]38.4.

Jo
~/I

784000

fig. A9.30 shows the general shape of the true ~ra~e bending ~cment diagram.

Example Problem 5
Distance to centroid of
;"C iroJ1 A.

curve along line ?ig. A9.31 shows the general details of one

A9.12

BENDING

MOMENTS

IN

FRA:YIES

AND

RINGS

half of a SYTI'-T.8t:'iC3.1 hull rreme that was used in an actual seapl~ne. The ~ai~ external load on such frames is ~he water ~ress~e cn ~he ~ul: bottom pl~ting. The hull bottom s~rtngers transfer the bottom pressure as cancentrations on the ~rame bottom as sho,vn. The resistance to this bottom upward load on the fr~~e is ;rovlded by the hull metal covering Nhich exerts tangential loads on the frame contour. The q~estlon as to the distribution of these reSisting forces is discussed in later chapters. In this problem the reSisting shear flow in the hull sheet has been assumed constant between the chine paint and the upper hea~J langeron. ~cr analysis purposes the frame has been divided into 20 strips. The centroid of these strips located on the neutral axis of the fra~e sections are numbered I to 20 in Fig. A9.2l. The tangential skin reSisting forces are shown as concentrations on frame strips #6 to #16. On the figure these tangential loads have Jeen replaced by their horizontal and vertical components. The sum of the vertical components should equal the vertical component of the bottom water ~ressure. Table A9.5 showS the complete calculations determining the bending moment on the f~ne. Col~s 1 to 7 give the calculations for the elastic properties of the frame, namely the elastic weight of the frame; the elastic center location, and the elastic centroidal moment of inertia about the horizontal ~entroidal elastic axis. A reference horizontal ~is X'X' ~as been selected as shown. All distance recorded in the table have been obtained by scaling fram a large drawing ot the fr~~e.
~or

The moment of inertia of the frame cross sections are given in column 2 of Table /1.9,6 which have been determined from a consideration of the actual dimensions of the frame members.

Bottom member of frame Floor Line

The static condition assumed for computing the M s moments is a double symmetrical cantilever beam as illustrated in Fig. A9.32. The trame is cut at the top to form the free end of the cantilever bea~, and the fixed end has been taken at the centerline bottom frame section. The static bending moment diagram will be symmetrical about the y axis through the elastic center of the frame and thus the redundant Yo at the elastic center will be zero since the term Zsx will be zero. The calculations tor determining the static moments M s in Column 8 of Table A9.5 are not shown. The student should refer to Art. A5.9 of Chapter AS to refresh his thinking relative to bending moment calculatlor$ on curled beams. Columns 9 and 10 give the calculations of the 0s values (area at Ms/I diagram) and the first moment (0 sx values). The summations of columns 3, 9, 10 permits the solution for the red~~dants Me and X as shown below the table. o The final bending ~oment M at any pOint on the frame is by Simple statics equalS,

I'

Fig. A9.31 eutra

Axis

I~

~~

I~

l~

Bending Moment Diagram

Fig. A9.33

Fig. A9.32

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A9.13

Table A9. 5

Calculation of frame Elastic Properties CoL

Calculation of Moment Weights and Solution of Redundants

1
0

3
Elastic Weight
w

4
Arm to Ref. Axis X'X'
w

6
w y'2

8
Static Moment

9
Moment Weight Os ::

10

11
~10

12

13
-Total
Moment

:: - I=~ fiE - " =.

0
~

y'

Arm

"

~,j
ds

~.=
I

= ds
1

'0 Axis
XX
~

es y

-Xoy

MS

:: y'
26. 1 24.4 21. 4 17.3 12.0 5.9 3050 2850 2497 2025 1403 79700 69500 53300 35020 16820 17. 2 15. 5 12. 5

~ 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
-9020 "

1 2 3 4
5

7.6
7.6

6 7 8 9 10 11 12
13

14
15
~.hine

7.6 7.6 7.6 6.4 6.4 6.4 6.4 6.4 6.4 6.4 6,4 6.4 6.4 7.0 7.0 7.0 7,0
7.0

.065 .065 .065 .065 .065

.25
. 25

117.00 117.00 117.00 117.00 117.00 25.60

25.60
25.60 25.60 25.60 21. 30 18. 82 18. 82 18. 82 18.82

.25
. 25

- 6. 3
-12.6 -19. 1 -25. 5 -32.0 -38.3 -44.6 -50.8

151 0 I
-161 -323 -488 -543 -602 -720 _840 -959

892
0

8.4 3. 1 3.0 - 8. 9

.25 .30 .34 .34 .34 .3'

12.0 12.0 18.0 34. a 50.0

1017 4060 9320 13850 19250 27600 57500 48750 1780 1980 1470

f-:

16

18

7, ,

O. 58 O. 58

19 20
Keel Totals

o. 21
O. 14
811.48

O. 39

-55.4 -58.4 -61. 4 -63. 2 -64.8

- 32 - 3' - 24 - 131 - 9
7228

840
588 423237

1220 31. 2 2880 71. 6 -15. 2 6230 159.3 286.5 -21. 5 11260 18470 473.0 -28.0 -34.4 29020 618.0 -40.9 43180 814.0 58980 1110.0 -47.2 77330 1455.0 -53.5 -59.9 99030 1865.0 111700 -63.0 -64. 3 149700 86.7 -67. 3 221400 128. 3 -70. 3 271200 105. 8 -72. 1 I 3086()0 64.9 -73.7 327900 45.9 -75.0 332900 I 7315.0

93.7 - 637 2420 - 6160 - 13220 - 21250 - 33100 - 52400 . 77800 -111300

5570 8650 7440 4680 3385

.. . .. .. .. .. .. . .. .. .. .. .. . .. .. .

..

-9020

16650 15020 12100 8130 3000 2910 - 8610 -14720 -20830 -27100 -33300 -39700 -45600 -51700 -57900 -61100 -62200 -65100 .68100 -70000 -71400 - 72600

890 -6020
10710 -14840 -17510 -18590 -17650 -13300 5540 4360 16610 32110 41600 78500 147300 194100 229600 247500 251280

7630 6000 3080

-348106

Y = 7228/811. 48: 8.9"

IXX:: 423237 - 811.48 x 8. 9 2 :: 358940

= -~Os
'-'w
XX

:: (7315.7 x 1000)
311.;:1

= -9020"#

x, :: Z?s y::. [
"Total.
Moment

348106 x 1000J:: -968# 35894

M at any statton e Ms ...

Mo -

XoY

Columns 11 and 12 record the values of M o and -Xoy for each station poIrrt . For examp.Ie , trie va Iue of -Xoy for station (1) equals - (-368 x 17.2) : 16650 and for stat~on
(20)

2401b

l.-. 6'
,
1

6" ------ooi

= -[-968(-73.7)J = -71400,

Fig. A9.44 shows the shape of the final ~cment carve as the result of the values in c ojunn 13.
A9.5 Unsymmetrical Structures.
Example prootem Solutions.

120~
I I

T I !_l_!..._x
6"
Y::5.242
:E.C.

c , -r-rI

~;1
I

x-

--..

9"

5. 032

Fig. A9.34

'1

10"

.L

Fig. A9.34 shows ~n l~~s~~etrical frame carrying ~he leads as shewn. Dete~ine tte jenQ~ng no~ents at p01~ts A, B, C and D.
Soluti.::m: -

The elas:ic weight of

(15/1)
+

(12/2)

10/1

= 31

~ra~e

= Zds/1 =

.'

I '1' ...,

A9.14

BENDING
j

MOMENTS

IN

FRAMES

AND

RINGS

The distance x f r'cm tne elastic center is,

me

AB to the

x = (15)0

+ 6

31

x 6

10 x 12

= 5.032

in.

Fig. A9.36 shows the centraid loc~tlQns uf the 05 values along the center line of the frame. The ~cment of these 0s values about tte x and y axes will now be calc~lated.
Z0sx

The dist~nce center equals,

y from line Be to

elas~lc

(-2160)(-3.032) + (-21600)(-5.032) + (-4260)(-5.032) = 139700

+

y = 15 x
The

7.5 + ~lX 0 + 10

x 5 = 5.242 in.

Z0sY = (-2160)(5.242)

(-4860)(-6.758)

= 70290

(-21600)(-2.258)

moments of inertia Ix and I y and the product of inertia I xy are ~eGuired.

elas~ic

The 'ffilues of :he recunGants at t~e elastic center can now be cal~~lated USing equations
(ll), (14), (15)
J

nameIy

Ix

= (~)'

+ (15)(2.258)' + (1:)(5.2421" +

(i~)'

M o

= ;;~~7I

= - (-;~620)

= 923

In. lb.

+ (10) (0.242)'
1'0
J

=606.51
2x12 2

= (15)(5.032)'
(10) (6.968)"

+ (12)' +(E)(o 968)' +

= 942.96
(~)(0.968)(5.242)

IxY = (15)(-5.032)(-2.258) +
+

70290 _ 139700 (217.74) 217.74' \ 606 51 (1 - (606.51)(942.96)' .

942.~6

68.46

(10){6.968)(0.242)

= 217.74

The next step 1s to assume some static trame condition and draw the static bending moment diagram. Fig. A9.35 shews that the frame has been assumed cut near pOint C which gives ~NO cantilever beams. The oending moment diagram in three parts for this static condition Is also shown on Fig. A9.35.

= - [Z0sx Yo

Z0sY

I y (1 _..::s:L)

(~)J
4

rxry

= L.:

_ 1139700 _ 70290 ,217.74)J

'606.51 217.74"" ;

-132.36 lb.

942.96 (1 - (606.01)(942.96)

T'i .

(",~I44l:ii0s:S:i( I,.) _ --1,cut

ISl=~2160

t '"

6"

5.242

2)

9.758

M s diagram
~1080

"I

10S3"'-~60
A
.o~

1_f:Jx t U l
OS2'::-

The bending moment at any (19) equals,

11 Ms +
~o

~olnt

from eq.

- XoY

+ Yox,

for exampl e ,

'1-

258

600 6 750

. C.

Consider point A.

x = -5.032, Y = -9.758, M s = -2520

MA = -2520 + 923 - 68.46 (-9.758) + (-132.36)(-5.032) = -263 in. lb. Point 8. x

Fig. A9.35

FIg. A9.36

= -5.032, y = 5.242,
+

~s

= -1440
+

The 05 values which equal the area at the M s diagram divided by the I values of the part1culqr portion will be calCUlated.

Ma = -1440

923 - 58.40 x 5.242

(-5.032) Point C.

=-

(-132.36)

209 in. lb.

0s,
~s.

= 3(-1440)/2 = -2160 = -1440 x 15 = -21600

x = 5.968, Y = 5.242, M s

=0
+

Me

=a

+ 923 - 58.46 x 5.242 (6.968) = -357 In. lb.

(-132.36)

0s, 4.5(-1080) = -4860 :0s

= -2160--

21600

4860

-28620

ThuS, the com~lete be~ding ~cment diagr~ could oe determined by computing several more '/alues such as ~oint D ar.d the ~xternal load points.

ANALYSIS AND DESIGN OF

Exa~ple Problem 2. Fig. A9.37 shows a~ ~n symmetrical closed ::."rame. The bending ~oment diagram will be deter.nlned under the given rr-ame loading.

FLIGHT VEHICLE STRUCTURES

A9.15

Total Ix

=6451.07
= 13.~
30(0.667)' : x IS' +.lx 12

Calculation ot I y Member AB, I y

4500.7

P2

= lOOt
I

_ 1 Member BO, I y - 'j]

x 31

367 x 30' = 791.7

{z x 20(11"
= 1125.0
: 4501.7
30'

- 20 X IS' + Member CD, I y -T

_ 1 Member AD, I y 12

1 x x:r

Totally

= 10919
= -1002.9

Calculation of Ixy ~---30"'--,

11ember AB, I xy = 13.~ (-151(3.343)

Hemoer BO, Jxy

ly'

Fig. A9,37

Solution: (30)(10)

31.67 1 31.67 = -3-(13.343)(0) -12 x - 3 -

The elastic weight of frame equals

~ "I
'<'

= -253.92
= 21 (15) (-1.657) =0

=1.5 30

-+-

31. 67 + 3.Q. 3 1

-+-

30

= 6!::;

~.,

,1:8

Member CD. Ixy

= 497.1

Member AD,

I~J

The location of the elastic axes will be determined with ~eference to assumed axes x'x' ~~d y'yl as shown on rig. A9.37.
=

Tot~l I~J

= -17G3.9

~(0)+()(151+()(301+(4f1(151 =
65.58

15'

(fj (151+() (251+FfI (101+(4f1 (0)

y

The next step in the solution 1s to assume a static fra~e condition and draw the M s diagr~. Fig, AS.38 showS the assumed static condition, namely pi~~ed at paint A and supported as rollers at ?oint D. The bending ~oment dia 6 raID is dra~m in parts as sho.vn.
750"

65.58

=
11.657'"
600 600
\

~uetoP2 I' 121

<Due to Pl

These distances and y locate the x and y e1ast!c axes as shown in Fig. A9.37. The elastic ~oments of inertia and ~he elastic product of inertia will now be calculated. Calculation at Ix Y barAB , -x T-lx-l..-(,e3'''+'1'''-7'')=1'7. . . .-;;:~1 dem 3 1.5 ~ . "'tV ..l. .<)~. 0
. .:::. . . . .

",1111 "
Due to Pl~l,
"

c
Fig. A9.38

=D

Member CD,

Ix=~

i (8.343"+ 11.657") =721.58

+

Member Be. Ix
(10" I

= (31367}(~3.343<i)

l~ (31;367)

{41 ~ ~---=m"""'--"- Due to P 3 and P4 _50 11 -;)UU

=1967.48 Member AD, Ix = (3 (l1.657::1) 20) (0.5)" =2038.30

1 +12x 30 x

The ~ext step is to compute the '~lue of 0s for each ?ortion of the ~oment di~gran. 0 5 is the area of the !'1s / I diagram. For reference the ~ortions of the M s diagram have been labeled 1 to 4.

.t

A9.16

BENDING
~co x -.-.- = ::000 03 = -.~G 1.;J

MOMENTS

IN

FRAMES

AND

RINGS

?ortlon Portion

(1)
(2)

30

03 =-Zx-3e;s
~

600

31.67

~~j ~oir.t e~~als the sta:::.c M s ~lus the ~omen: due to the r-eduncant forces as S:lO'.\TI in !"i.;;;. ).'; ... 0. b~r.ding ~om~~: ~t o~i~:na1

The

~'t:~

o,~

" , Portion \0).

-2-

31.67

T-O ---t::

B
39:~8

Portion (4). - el s z0's

= -5CO(120)

_500(1 ) 2

= -,3000

18.343!
!

I~
:

-165

~l~C-.

L.-i

= 13125

- 5000 :: 8125

~ :4.674

.
:

8.343 "31. 1)71-;

.

-Ai
;""-15 I

iD.L-

11.657

167,,

\
286 240

fi'"1307

The individual 0'8 values are now concentrated at the e.g. of each of the Ms diagrams. Fig. A9.39 shows the location of the ~s values with respect to the x and y axes thro~~ the elastic center.
0s:3167
I

11--"":""'-15''------1 .

167'4;j[~307
Fig. A9.41

Fig, A9.40

Consider
MA

~oint

A: -

M s :: 0
+

= a - 124.8 - 4.674 x :5
M s :: 600

31.07 x 11.657

= 167
Point B.
I
,

in. ::''::1.

I I I

Ds~6000~
;
I

'....5_

I'
I

l8.

15 34 31

13.343
~

Ma = 600 - 124.8 - 4.674 x 15 - 31.07 x

16.3~

'

!
11. 657

,
Point c.
!'Is

= -165 in.lb.

=0
+

r-- 15'-L.........j
Z0sx

. .
i
~

0 s .:5000

---

Me

=0 -

124.8 - 31.07 x 8.343

4.674 x 15

= -314
Ms::
0
4.574 x 15
+

1::1.1'J.

Fig. A9.39

Point D.
MD

= 6000(-15)

+ 3167(-5)

=-105835
= 208659

= 0 - 124.3

31.07 x 11.657

l0sY = 6000(8.343)

+ 3167(15) +

=307
AS.6 Analysis of Frame with Pinned Supports.

in. lb.

3958(13.343) -5000(-11.657)

Fig. A9.41 shows the true bending moment diagram.

The values of the redundants at the elastic center can now be calculated.
MO

=~ ~ Zds/I
= Z0s

-(8125)
65 .58

= -124 8

n,

1b

I y - Z0s x ~)

Fig. A9.42 shows a rectangular frame and loading. This frame is identical to example problem 1 of Art. A9.3, except it is pl~ed at points A and D instead of fixed.

x.,

Ix (1 _

;xy')
x"y

=208659 =- [l0sX

(-105835)(-~)
10919

6451 [1 -

(-1753.9')

6451

The first step will be to dete~ine the elastic weight of the fr~ne, the elastic center location and the elastic moments of inertia about axes through the elastic center.
31.07 lb.

Yo

- l0sY
I

I y (1 -

IXly )
208659C1753.s11 6451 ~_ 4.674 lb. (-1753.9') 10919 x 6451

(-;~

= -[-105835
10919[1 -

J-

The term dslEr of a 'Jeam element of length ds represents the angle change be~Neen its ~#o end :aces when the element 1s acted upon by a unit moment. In this chapter this tern has been called the elastic weight of the element. PhYSically, the elastic weight is the ability of the element to cause rotation when acted upon by a ~it moment. When a unit ~oment is applied to a rigid support, the support suffers no rotation since the support is ~igid, therefore a rigid support has zero elastic weight and therefore does not figure in the fr~e elastic properties.

A9.17

If a su~port is pi!h~ed or hinged it ~~s no resistance to rotation and thus a unit moment acting on a hinge would have infinite angle change or rotation and therefore a hinge or pin possesses infinite elastic weight.

4 46 -14'n

_14~!30'54 ~14.46

"F' 1 I"
L
230

~'

10 lb.

-:
I

~s=270

I ,

'"
Fig. A9.44

Fn -ld(~L
yJE.C. Y Fig. A9. 43

,-:: oo!!J
Fig. A9.45

"
Fig. A9.46

Fig. A9.45 shows the bending moment diagram due to the redundant XO ' Adding this diagram to the original static diagram gives the final bending moment curve in Fig. A9.40.
Ag.7 Analysis of Frame with One Pinned and One Fixed Support.

Fig. A9.42

Due to symmetry of structure about the centerline y axis the elastic center will lie on this axis. Since the two hinges at A ~d B have infinite elastic weight, the centroid or elastic center of the frame will obviously lie midway becNeen A and B. Fig. A9.43 shows the elastic center 0 connected to the point A by a rigid bracket. tor frame is infinite because or the hinges at A and B.
~ds/EI

Fig. A9.47 shows the same frame and loading as in the previous example but point D is fixed instead of hinged.

n4i
I Fig. A'.
y

The elastic moment of inertia about a y axis through elastic center is infinite since the hinge supports have infinite elastic weight. Ix is calculated as follows: For Portion AB For Portion CD For Portion Be

x-Ai - - x

The support D has zero elastic weight and the pin at A has infinite elastic weight, therefore the elastic center of the frame lies at point A. Tne total elastic wei~~t of frame 1s infinite because of pin at A.

=..lx.l.x 3 3
3 3

30'

= =

3000

The elastic moments of inertia will be calculated about x and y axes through A.
Ix

:J:..xl x 30'
= 24

3000

x ~ x 30A 10800

Ix =- 16800
Fig. A9.44 shows the static frame condition assumed to obtain the M s values. The value of 0s for member Be equals the area ot the M s curve divided by I tor 8e, hence s = 4S x 24 x 1 = 270. The centroid of this s value is 10 inches from pOint B. redundant forces at the elastic center can now be solved for
'~e

= 16800 (Same as previous eXaID?le) I y = (3 x24' )+(; x ~ x 24') = 8064 30 x 30 x 12) = 7920 Ixy = (;0 x 15 x 24-) + (2 24

The static frame condition will be assumed the same as in the previous example problem, hence 0s = 270 and acts lOW from B (Fig. A9.48).
/ ' " 5270

2' 2'

r~lO...i
3~'
'I

1S ' n -l -11.92-~

"'
rw

W1l

-270 ir.f:ni ty
y
X = -(270)10 = 0 = -Z5 1y infinity

tl;!- Xo
, 'Yo
Fig. A9. 48

' !MO

"""

~.

5038#

'i 19

28.45

~1

r, I
'a

A-11.92

i. 2582*

~ 6.1'
Fig. A9. 50

Fig. A9.49

_ ~0sY _ 270 x 30 Xo - iX - 16500

=0.482

Since the frame is ~~symmetrical, the x and y axes through the elastic center at A are not prinCipal axes, hence
lb.

CHAPTER AlO

STATICALLY INDETERMINATE STRUCTURES SPECIAL METHOD . THE COLUMN ANALOGY METHOD

AID.l General. The Column Ana Logy- method is a

method tr~t 1S widely uSed by engineers in cetermlnlng the bending moments in a bent or ~lng type structure. The method considers only distortions due to bending of the

structure.
The numerical work in using the column analogy method 1s practically identical to that carried out in applying the elastic center method of Chapter A9.
AID.2 General Explanation of Column Analogy Method.

Now assume we have a frame whose centerl~ne length and Shape is identical to that of the column section in Fig. AIO.I. The width of each portion of this frame will be ?rOportional to lIE! of the member. Fig. AlO.2 shows this assumed ~e. Furthermore, assume tr~t end B of the frame 1s fixed and that a rigid bracket is attached to the end A and terminating at point (0) the elastic center of the frame. The frame is SUbjected to an external loading, W"l.1 w{I, etc.
Y

Ell

I
'-Xo
:f),Mo

Fig. AlO.l shows a short column loaded in compression by a load P located at distances (a) and (b) from the principal axes x and y ot the column cross-section.

w,

x.i,

EX
A

jyo

--r
-ly

f-x
Fig. AlO.2

:11-

Fixed

Fig. AlO.!

To tind the bearing stress between the supporting base and the lower end of the column, it 1s convenient to transfer the load p to the column centroid plus moments about the principal axes. ~hen it we let cr equal the bearing stress intensity at same point a distance x and y rr-ce the yy and xx axes, we can wr t te

This cantilever structure wIll sutfer bending distortion under the exterr~l load system W)., Wv e'tc , , and point (O) will be displaced. Point (0) can be brought back to its original undeflected position by applying a couple and two forces at (0), namely, Mo, X o and Yo as shown in Fig. AlO.2. Since point (0) is attached to frame end A by the rigid bracket these three forces at the elastic center (0) will cause point A to remain stationary or in other words to be fixed. Therefore, for the frame in Fig. AlO.2 fixed at A and 8, the moment and two forces acting at the elastic center cause the statically indeter.ninate moments M1 when reSisting a given external loading causing static ~oments MS. The fir~l true bending moment M at a paint on the frame than equalS M=!'1s-+!"!i'

wnere A is t~e area of the calumn crasssection and pa and Pb the moment of the load P about the xx and yy axes respectively. If we let Pa = M x and Pb = My, the above equation can be written,
- - (1)

From ?ig. AIO.2 we can 'Nrite for a point on the frame such as 8 t~at the lndeterminate bending moment Mi equals,
M. i
~or

=:1

+ Xoy -+ Yox

- - - - - - - - - - (2)

In Crapter A9, Art. A9.3, the equations Mo, Xo and Yo were derived. They are,

-Method of Analysis due to Prof. Hardy Cross. See "The Column Analogy" Univ.lll. Eng. Expt. SU. Bull. 215.
AlD. I

AlO.2

THE COLUMN ANALOGY METHOD

o = Zds/I ' YO = M

- Z0sx
Iy

(4-)

at any

The ter.n z0s represe~ts the area of t~e static Ms/I curve. (E r2s been assumed constant and therefore emitted). Let :~e t&~ Z0s be called the elastic lead and s~ve it a new synbel P. The term Z15/1 equals the elastic weight of the fr~~e and equals the s~ o~ the length of each ~ember tl~es its width which equals 1/1. Let this total ~~~~e elastic weisht be given a new 5J~bol A. In the exoressions ~or Yo a~c Xo the terms Z0s x and Z0sY represent t~e TIcnent of the static Mil curve acting as a load abcut the y and x axes respectively passing t:;rcugh the frame elastic center. Therefore let Z0s x be given a new symbol My and Z0sY a new symbol M x With these new symbJls, equation (2) can now be re'NTltten as follows: M1 - p +

?ig. AIO.3 shews a

~~xed

rectar.g~la~ ~r~~e

with

supports ~t ~oints A anj D. The bend~ng ~o~ents at ~oin:3 A, 5, C and D will Je dete~ined JY the cc~~~n a~alogy ~~t~ed. T~is ~r~~e and load~ns is identical to ex~~?le problem 1 of Art .."-~.? or c-a-ter ;:.g'Nnere the sclution was ~aQe JY the elastic cen~er =e~hoG.

-16"

10 lb.

45"#

18"--1
I : 2 L:. 24

BI;-.L_ _

---:c I
10
~

~
..

L'30,,1
1-3

-A

E;.
-x

+ M{X __

- -

(3)

j
7. 5

Comparing equations (1) and (3) we see they are similar. In other wordS, the indeter.ninate bending ~oments Mi in a trame are analogous to the column bearing pressures 0, hence the name nColumn ~~alogyn for the ~ethod using equaticn (1). ~ith this ~eneral explanation, the method can now be clearly explained by g~ving several example prablem solutions.
AlO.3 Fra.mes with One Axis of Symmetry.

Fig. AlD.3

2.5

Fig. AlO.4

SOLUTICN NO. i
~e first c8r.sider the ~r&~e centerline shape as shewn in Fig. AIO.3 as the crcsssection of a shart column (see rig. AlO.5). The width of each portion of the column section is equal to llEI of the ~ember cross-section. Since E is constar.t, it ~i1l be ~de w~i~y a~d the widths will then eq~al 1/1.

From the previOus discussion we can write, (1) The cross-section ot the ar.alogous column consists ot an area, the shape 0f which 1s the same as that of the given frame and the thickness of any part equals liE! of that part. The loading applied to the top of the analogous col~ is equal to the Ms/EI diagram, where X s 1s the statical moment in any basic dete~inate str~cture derived from the given frame. If Ms causes bendi~g compression on the inside face of the frame it is a positive bending moment and the analogous load P on the column acts downward. The indeter.nlnate bendln~ ~oment M 1 at a given frame ~oint equals the base pressure at this same poi~t on the analogous col~14~. Thus the indete~inate moment at any paint on the frame equals (fromeq.3),

(2 )

i
I

Fig. AlD. 5

Fig. AlO.6

(3)

The first step in ~he ca:culations is to compute the area (A) of the column cross-section in Fig. AlO.S and the moreents of inertia of the column cross-section about x and y centroidal axes.

The calculation of the location o~ ~he centroldal axes and the moments at :nertia Ix

AlO.3

and I, woulc be identical to the ccmnlete calcufations given in Art. A9.4 J and-Table A9.l where this sa~e problem is solved by the elastic ce~ter ~ethod. These cal~ulations Nill not be re?eated here. The results Nere J Ix = 3188 3456. and 1y

Mx 270 x 9.375 2530 (positive because 8ase pressure is compressive on column portion above x axis).
My =- (270x2) =-540 (negative because baSe oressure is tension on column ~ortion to right" of yaxis).

Since the

fra~e

is statically indeterminSubstitution in equation (3). Frame Point A.

I1i

ate the next stea 15 to assume a static frame

condition consiste~t with the given fra~e and loading. Fig. AIO.4 shows the condition assumed for this solution J namelYJ pinned at point A and a pin with rollers at point D. The static M s diagram is therefore as shown in Fig. AIO.4. ~e now load the column crosssection with the Ms/1 diagram as a load as shown in ?ig. AlO.S. The static mocent Sign is positive because the sGatic condition causes tension on the inside face of the frame. In the column analogy method a positive Ms/I loading is a downward or compressive load on top of cOlumn J and therefore a negative Ms/I value would be an upward or tension load on the column.
~x

=-

12"J

=-

20.625"

=1: + MxY + MlX A Ix y

= 8.44-16.38+1.88 =-6.0. tn Lb ,
i

- 270 2530 (-20.625) + (-~5~-12) l'l1-32+ 3188

The true bending moment fram equation (5)

=OJ

see Fig. AIO.5

Equation (3) requires the values of the and My the moment of the ~s/I diagram as a lead about the x and y axes. Equation (3) also requires the total column load P which equals- the ar-ea of the Ms/I diagram. For this problem the value of P from Fig. AIO.S equals J
P

whence, l'lA = 0 - (-6.06) = 6.06 in.lb.

Frame Point B.
l'l
i

=- l2"J

9.325"

= 270 + 2530 x 9.375 (-540) (-12) 32 3188 + 3456

= 8.44+7.44+1.88 =17.77
l'lB = l'!s-l'li = 0-(17.77) =-17.77 in.lb.

=22.5

x 24/2

=270

The centroid of this trianG~lar loading is 2 inches to left of yaxis. Fig. AlO.6 shows the column section with this resultant load P. now USe equation (3) to find the indeterminate moments Ml which are equal to the base pressures on the column. Bquation (3) involves bending moments M x and My and disGances x and YJ allot which must have Signs. The Signs will be determined as follows: When moment ~x produces compression on base on that ~ortion above x axis J then M x is
pos Lt.tve ,
~e

Frame ?olnt C.

= 12

= 9.325
(-540)12 3456

270 2530 x 9.325 - + l'li = 3188 32

= 8.44+7.44-1.88 = 14.0

Me

=l'ls-l'li = 0-(14.0) =-14.0 in.lb.

Frame Point D. Mi = 32
270
+

x = 12 J
3188

=-20.625
3456

2530(-20.625) + (-540)12

= 8.44-16.38-1.88 =-9.82 l'lD = l'ls -l'li = 0 - (-9.82) = 9.82 in.lb.

My produces c~mpress1on on ~ase on that portion to right ot y axis J then My Is pos i t t ve .

~nen ~oment

A distance y is

~easured

posit~ve; ~easured

upward from x ~1s downward 1s negative.

A distance x measured to right fram y axis is positive J to left negative. From Fig. AIO.6: p

= 270

Fig. A10.7 shows the final bending ~ament d1agram, which of course checks the solution by the elastic center ~ethod in Art. A9.4. The student should note that the numerical Nork in the column analogy ~ethod is practically the same as in the elasGic center method.

28.17

Fig. AIO.7

AID. 4

THE COLUMN ANALOGY METHOD

M A

Solution No.2
In this solution a different static ~raoe condition will be assumec as shown in Fig. AIO.8, namely the frame is cut under the load and one-half the 10 lb. load will be assumed as gOing to each cantilever part. Fig. AlO.9 shows the static ~oment diagram and Fig. AIO.lO the static no/I diagram with centroid locations or each ~ort1on or diagram which

= ~'s

-11 1 :: - 30 - (-36.06) = 6.05 in . It: .

which checks solution 1. Frame Point 8.

x

=-

12,

9.375

M - -1650 + 2531x9.325 + (-9180)(-'2) ! - ~ 3138 3456

=-12.23
Me

=M:s-Mi
c.

=-30-(-12.33) =-17.77in.lb.

-30

-90

Frame Point

x :: 12,

= 9.375

_-1650 + 2531x9.375 + (-9180)12 M1 -~ 3188 3'-u.:: 400

=-76.0
-30 Fig. AlD.9 .90

Me

= M s-Ml =-90- (-76.0) =-14.0 In s Ib ,

4'

Fraille Point D.

= 12,
3186

Y = -20.525
(-9160 )12 3456

Ml
(4)

::~ + 2531(-20.525)

32

;' -99.82
MD

(1)

-10

-30

=Ms -M1 =- 90 -

(-99.82)

9.82 1::1. lb.

h2''1
Fig. AIO. 10 Fig. AIO. 11

Thus solution 2 checks sol~tlon 1. The student should solve this problem uSing other static conditions.
AIO.4 Unsymmetrical Frames or Rings.

are numbered (1), (2), (3) and (4). The area of each ot these portions will represent a load P1 , P a , etc. on the column in F1g. AlO.ll. Since the static moment 1s negative on each portion the load on the column section will be upward.
Pl. =-10(30) =-300, p. =-15(6)/2 =-45, P3 =-45x9 =-405 p. =-30x30 =-900

In applying 0he column analogy ~ethod to unsymmetrical frames and r1ngs, the moment of the H/EI dia~ram ~ust be taken ~bout principal axes and the ~aments of inertia ~ith resnect to principal axes. . However, as explained for the elastic center method in Chapter A9, the moments and section properties with regard to centroidal axes can be used if M x, My, Ix and I y are modified to take care of the dis-symmetry of the structure. In Art. A9.2 it was shown that the redundant forces at the elastic center to ~~ s~etrical frame sections was, (see equations 11, 14, 15 of Art. A9.2).
M

p = Zp = - 300 - 45 - 405 - 900 = - 1650

From Fig. AIO.ll

M x

=- (45+405) = 2531
= -9180

9.375+ (300+900) 5.635

My

=300 x 12 + 45 x 10 - 405 x 6 - 900 - 12

=!. +
A

o =~-p( ~ds/EI - T same as , . . or symmetrical

t:-ame)

111

IX
-

MxY + Mt{ y

POINT A.

>

12,
+

y = - 20.625 (-9180) (-12)

+

Ml :: -1650
32

2531(-20.525)
3188

3456

---

(~)

= -36.06

ANALYSIS AND DESIGN

OF FLIGHT VEHICLE STRUCTURES

AlD. 5

As

prev~ously

done in Art. AlO.2 let,

SOLUTION: Fig. AlO.13 shows the cross-section of the analogous column. The length of each ::nember of the column section being the same as in Fig. AlO.l2, and the width of each portion being 1/1. The first step in the solution is to find the column section properties. The total area (A) of column section equals
(15/1) + (10/1) + (12/2)

Fu:thermo!e, let, and (4:!-1 y

and k

!'Ix

= M,;

(2f!--)

= (1

-~I

Ixl y

=31.

we

Then substituting in equations (a) and (b) obtain,

The other properties required are: - Ix,

I y and I xy

From equation (2) we have,

Substituting values of Xo and Yo into this equation, we obtain as the equation for 11i the indeterminate moment in the column analogy ::nethod, the following Ml =~+ (l1x - 11 A :'::Ix

The calculations to determine the location at the centroidal x and y axes and the above properties about these axes would be identical to that given in example problem 1 of Art~ A9.5 at Chapter A9, where this same problem was solved by elastic center ~ethod. The results were Ix 606.51, I y = 942.96 and I xy 217.74. The location ot the axes were as shown in Fig. AIO.13.

y)Y

+ (Mv -

i<I y

i1X )x

- - -

(6

The true moment is the same as for the symmetrical section, namely,

The next step in the solution is to choose a static frame condition and determine the static (~) diagram. Fig. AlO~l4 shows the assumed static condition, namely, two cantilevers because ot the frame cut as shawn~ The figure also shows the static bending moment diagram made up of three portions labeled (I), (2) and (3). The area at each portion of the moment diagram divided by the (I) for that frame portion will give the loads? on the c o Iumn,
240

Thus the solution of an unsymmetrical frame by the column analogy method tollows the same procedure as tor a ~etrical section except that equation (6) 1s used instead of equation (3).
AlO.5 Example Problem - Unsymmetrical Section.

-1440"

(I) \.B , \~cfn:",,","---1 h

cut

Fig. AlO.l2 shows a loaded unsymmetrical frame fixed at pOints A and D. Required, the true bending moments at points A, B, C and D. This problem is identical to example problem 1 of Art. A9.5 where a solution was given by the elastic center ~ethod.

2401b

at
, I JI 120'--r:
9"
1

L 6" -.;... 6" ..,

1=2

I' I

L'T
I
10"
1

-1~40
P 1.
P II
p,
Fig. AlD. 13

Fig. AlO.14 1080 _5"'00"'3"'2 , Y

1-1

I I

1= 1

-'-

-l IA
'::T:7

=- 1440 x 3/2 = - 2160 =- 1440 x 15/1 = -21600

=- 1080 x 4.5/1 ==- 28620
4860

Fig. AlD. 12

zp

AIO.6

THE COLUMN ANALOGY METHOD

Fr~e

These loads act on the centerl~ne of the frame members and t:rough the centro1~ of the geometrical moment d1agram Shapes. These centroid locations are indicated ~y the heavy dots 1n Fig. AI0.l4 and their locations are given With respect to the centroidal x and y axes. The loads Pl, Pa and 23 are now placed en the column in AIO.13, acting upward because they are negative.
we now find the ~crments M x and My wh1ch equal the moments of the loads P about the cer-troldal axes.
i1:x

Po1rot

C.

6.968,

Y = 5.242

M1 =-923+58.37x5.242+132.36x6.S68

= 357.7
~

=Ms - Mi = 0 =

(357.7)

=- 357.7

tn .lb.

FY"ame Point D.

x = 6.968, Y =-4.758

Mi =-923+68.37 (-4.758) +132.36x6.968

-326
-I'1i

=- 2160 x 5.242 + 21600 x 2.258 + 4860

x 5.758 x
= 70290

?'In

= lis

0 - (-326)

= 326

1::.10.

2160 x 3.032'" 21600 x 5.032 + 4860 5.032 = 139700

To solve equation (6) we must have the terms ~' MY and k ,

The above results check the solution of this same problem by the elas~1c centeY" method in Art. AS.S of Chapter A9. The st~Gent should solve this problem by chOOSing other stat~c frame conditions.
AlO.6 Problems

~ = l1x

(*)
y

70290

x 217.74/606.51

= 25234
11' Y

(1) Deter.n1ne the bending moment dlag~am ~or the loaded structures of Figs. AIO.IS to AIO.20.

=11Y
-

(~) I -

= 139700 x 217.74/942.96
= (1
_ 217.74' ) 606.51 x 942.96

~ lSi 15 +151
L_45" I 3 L=30" I =2

400

400

w=lO Ib/in.

= 32258
k

= (1

ix(

x y

11

IL=30"
=2

H~'; I

= .9171

Fig. AlO.l5

""'" Fig. Ala. 16

300 lb.

II

L = 24 =6

L = 24 I =6

SUbstituting values in equation (6) we obtain,

-28620 + (70290 - 32258 )y + 31 0.9171 x 606.51 (139700 - 25234 ) x 0.9171 x 942.96
~12"

200 L 224 , I ; 6 i 100 L=20 L=20i

1;4 1=4:

Ir

+-:0 \Cl I I
29I

'5"~

1<> lCrJ

10

L; 24
1=4

whence,
!'1 i

i-.i.

i 10

=- 923 + 68.37y + 132.36x

x
=-5.032,

1-8
-----(7)

8+i"";

.....

:L==
II

20 5

1L==
I

20 5

i---- 24" -------!

Fig. Ala. 18

For Frame Point A.

100 100 Fig. Ala. 17

Y =-9.758
100

Mi = - 923 + 68.37(-9.758) ... 132.36 (-5.032)

= - 923 - 667.15 - 666.0 = - 2256 I1A = 11. -111 =-1440-1080- (-2256) = -264 i'l.lb.

100 1/,-;--"
, I

400

4~",~~~
2
-..~~I=
I

~ f.. . 1-~- /,10" \ , 1= 2

1

~4'~12'~'''' ,

For Frame Point 8. Mi

=- S.032,

y = 5.242

1.5

200

200

=- 923 + 68.37 x 5.242 + 132.36 (-5.C32)

= - 923 + 358.39 - 666.0 = -1230.6

200 lb. Fig. Ala. 19

Fig. AID. 20

ME

= 11 8 -~

=-1440- (-1230.6)

-209.4 l.n.l'tl.

(2) Solve ?rob:ems (2) and (3) at ~he end of C~~pter A9, Art. A9.9.

CHAPTER All

CONTINUOUS STRUCTURES MOMENT DISTRIBUTION METHOD

All.l Introduction. The moment distribution

method was originated by Professor Hardy Cross.* The method 1s Simple. rapid and particularly adapted to the solution of continuous structures ot a high degree of redundancy, where i t avoids the usual tedious algebraic manipulations of numerous equations. Furthermore, it possesses the merit of giving one a better conception at the true physical action ot the structure in carrying its loads, a tact which Is usually quite obscure in some methods of solution. The method ot procedure in the Cross method is in general the reverse of that useo -tn the usual methods where the continuous structure Is first made statically determinate by removing the continuous feature and the value of the redundant then solved for which will provide the original continuity. In the Cross method each member of the structure is assumed in a definite restrained state. continuity of the structure is thus maintained but the statics pt the structure are unbalanced. The structure is then gradually released from its arbitrary assumed restrained state according to definite laws of continuity and statics ~til every part of the struc~~e rests in its true state of equilibrium. The general principles of the Cross method can best be explained by reference to a speCific structure. Fig. All.l shows a continuous 2 span beam. Let it be reqUited to dete~ne the bending moment diagram. T~e first arbitrarily assume that each span is completely restrained against rotation at its ends. In the example selected ends A and C are already fixed so no restraint must be added to these points. Joint B is not fixed so this jOint is 1magained as locked so it cannot rotate. The bending moments which exist at the ends of eacri member under the assumed condition are then determined. Fig. All.2 shows the moment curves for this condition. (For calculation and formulas for fixed end moments see following articles). Fig. All.3 shows the general shape of the elastic curve under this assumed condition. It is noticed that continuity of the structure at 8 is maintained, hOwever from the ~oment curves or rig. A1I.2 it 1s found that the internal bending moments in the beams over support 8 are not statically balanced, or speci:ically there is an unbalance of 270. The next step is to statically balance this joint, so it is unlocked tram its imaginary locked state and ObViously joint B will rotate (See Fig. All.4) until equilibrium is established, that Is, until resisting moments equal to 270 have been ( Paper - A.S.C.E. Proceedings, May 1930)

set up in the two beams at B. The question is how ~uch of this moment is developed by each beam. The physical condition which establishes the ratio of this distribution to the two beacs at B 15 the fact that the B end of both beams must rotate through the same angle and thus the unbalanced moment at 270 will be distributed between the two beams in proportion to their ability of resisting the rotation at their Bend thru a common angle. This physical characteristic ot a beam is referred to as its stiffness. Thus let it be considered that the stiffness factors of the beam BA and Be are such that 162 is distributed to BC and - 108 to BA as shown in Fig. AI1.4. (The question of stiffness factors is discussed in a following article). Referring to Fig. AII.4 again it 15 evident that when the elastic curve rotates over joint B that it tends to rotate the far ends at the beams at A and C, but since these jOints are fixed, this rotation at A and C is prevented or moments at A and C are produced. These moments produced at A and C due to rotation at B are referred to as carry-over moments. As shown by the obvious curvature at the elastic curves (Fig. All.4), the carry-over moment is of opposite sign to the distributed moment at the rotating end. The ratio of the carry-over moment to the distributed moment, referred to as the carrJ-over factor, depends on the physical properties of the beam and the degree of restraint of its far end. 1Carry-OVer factors are discussed in a following article. For a beam at constant section and fixed at the far end, the carry-over factor is -1/2). In figure AI1.4 a factor at -1/2 has been assumed which gives carryover moments of 54 and -81 to A and C respective* ly. To bbtain the final end moments we add the original fixed end moments, (.the distributed balancing moments and the ca~J-over~aments as shown in Fig. AII.4. With the indeterminate moments thus determined, the question of shear, reactions and span moments follow as a matter of statics.
All.~

Def1n1t10na and Der1Tat10ns 01 Term8

ft

1.

Flxed~end ~oments:

By ~:ixed end moment 1s meant the moment which would ex1st at the ends of a member if these ends were fixed against rota~ion. stiffness Factor: The stiffness :actor of a ~ember is a value pnopor-t tonaj, to the tnagnf tude of a couple that must be applied at one end of a member to cause unit rotation of that end, both ends of the member being assumed to have no movement at 2.

All.l

'."".':_'~':.""'~~-:;;-

'{~1~:

.111.2

THE

MOMENT

DISTRIBUTION METHOD

200

Fixed~
A

VI:

, .I.

10.,/in. . ! I"

i" L1.3~ ----1

r-

0.5 when rar end is fixed. ~he letter ~ will be ~sed to designate carry-over fcc~or.
15''-01 L

1= Fixed

4.

B Fig. All. 1

I
,-480

Distribution Factor: It a ~oment is applied at a joi~t Nr.ere :NO or more members are rigidly connected the distrlbut10n factor for each member 1s the prcport~onal part of the applied moment that ~s resisted by that member. The distribution factor for any ~ember which will be given the symbol D equals KlZK, where K equals stif~ness ~actor of a particular member and ZK equals the sum of K values for all the Joint members. The sum of the D values for any joint must equal unity.
5.

54

Sign Convention: Due to the fact that in ~ny problems where members come into a joint n-om all dn-ect icns as is commonly found in alr~lane structure, the Elastic Curve for Fixed End Spans customary sign convention for moments ~y produce contusion in applying the ~oment distriJutlon method. The follaNing sign convention 1s used in Fig. All.3 this book: a clockwise moment acti~g on the end of a member is positive, a counterclockwise one is negative. It tollows that a moment tending to Elastic Curve due to Rotation of Joint B rotate a jo Int clockwise is negative. It should be understood that when indete~lnate continuity moments are determined by the ~oment distribution Balancing Joint B i method using the above adopted sign convention, ! -1081162 that they should be transferred i~to the conven,./ --=:::::::::... ! Carry Over Moments -811 from B to C & A tial signs before proceeding with the design of ~ ::..::::::..:c.:::...:....:c.::.._ _ the member proper.
.5881.588
I

-8311 Final Moments = Values I in Figs. All. 4 plus All. 2

I

The follOWing sketches illustrate the ed sign convention.

a~pt

Fig. All. 4

Illustrations at Sign Conventions for End Xoments. Fixed end beam with Example 1 lateral loads.
i

.26~~~====~~~~~~~~~~~~~~~~/~1,
rA

I I

831 C Final Moment Curve

Conventional moment Signs

tension in bottom fibers 1s POSitive bending moment.

Adopted sign convention translation. The value ot the stittness :actor will depend in part on the restraint or degree ot tixity of the opposite end ot the member from which the couple is applied. The letter K 1s used as a symbol tor stiffness factor. carry-over Factor: It a beam is simply supported at one end and restrained to same degree at the other, and a moment is applied at th5 simply supported end, a ~oment 1s developed at the restrained end. The carry-over tactor is the ratio of the moment at the restrained end to that at the simply supported end. For a prismatic beam without axial load the carry-over factor 15 3.
-~

-~ '--i-

moments which tend to rotate end ot member clockwise are pOSitive.

ExaI:1ll1e 2

Translation of supports of fixed ended beao.

Conventional I:loment Signs

Adopted s ign::E:::::::::::--. convention ~

-=

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All.3

E:xam:ple 3 External applied ~oment at joint in structure adopted sign Joint

1.L... _

M t~

L~

Pab

S'3M a

Fig. All. 9 Fig. AI1.10 Fig. All.ll

+
AIl..3

moment

Moments which tend to rotate joint counterclockwise are positive.

Fig. AI1.12

Calculation of Fixed End 1I0111euts

Since the fixed end moments are statically facts must be obtained from the laws or continuity i~ order to solve for them. In this book the theorem of area moments will be used to illustrate the calculation of the fixed end moments as well as the other te~ which are used in the moment distribution method. -(Ret. Chapter A7) The following well known principles or theorem of area moments will be used:(1) The deflection at any point WA" on the elastic curve of a beam away from a tangent to the elastic curve at another point WB" is equal to the moment a r the area of the M: diagram beEl tween the points A and 8 about point WAft. (2) The change in slape between two points ~A~ a~d ~Bft on the elastic curve at a beam 1s equal to the area of the M diagram between the El two paints ~A" and "9". The "area moment" theorems will be illustrated by the applications to the solution at a simple problem. Fig. AlI.5 shows a simply supported beam of constant ~oment of inertia and ~odulus of elasticity carrJlng a single concentrated load. Figs. All.6 and All.? show the static moment curve acd the shape of the beam elastic curve. Now assume that the ends are fix ad as shown in Fig. All.S and let the value at the fixed end moments be required. Fig. All.9 shows the shape of the final moment curves wade ~p of the static moment curve and the '~~~O~trl tr-api zo i ca.; moment curve for:ned by the ur ...iknown end moments. Fig. AIl.IO shows the shape of the elastic curve, the slope at the two supports being made zero by fixity at these points.
indete~inateladditional

Since the change in slope of the elastic curve bet'.. . een ends (1) and (2) is zero, theorem (2) as applying to fixed end beams can be restated as follows. The sum of the areas of the moment diagram must equal zero. And fram theorem (1) the static moment of the areas of the moment diagram about any point must equal zero or in equation form: ----(1) ZM = 0 - - - - - - - - - For a beam with variable moment of inertia the conditions for fixity are:jl1dx/El = 0 j!".xdx/EI = 0
==0 ------------(2)

Figs. All.ll and AIl.12 show the static and continuity moment areas, the total area of each portion and its e.g. location. Substituting in Equation (1) M = Pab M1L MaL 0 ------(3)

7+7+2=

and rr on equation (2) M x about lett end = Pab L + a

2'-3-+23+

~"hL

- -(4)
The values of M 1 and t1a tor any value at a or b can now be found by solving equations (3) and (4) Table AlI.l gives a summary at bema fixed end moments for most at the loadings encountered in routine deSign and analYsis.
w

TABLE All. I

'-1
1

I a -t--- b -----:1
EI is constant
L _ _--I

Fig. All. 5

;.-' 12L" 4aL-3a)

w

r;~(

@~
~

Fig. All.6 FIg. Al1. 7 Fig. AI1.S -8-

Fa

iP I

- -

b=1

-..-

WL~_1< + !:J r::WL 8

. 2 ' 2
v ::::'

,
;0 (5u ... 2v)

::::;.-:;::::rITn , 1 I ; I L L'
L ----{

6O(5u ... 3v)

All.4

THE MOMENT

DISTRIBUTION

METHOD

Table All.l - continued

w2~~ 4Jl:Ii:Ii::r::::----=~f w3~

~5~" {lOL <I_ 10aL+ 3a 'i

_W

tangent to the elastic curve at poi~t A relative to a line AS is equal to the shear at A on a si:nply suppor-ted beam AB cue to t ae vr curve be::;1
t~een

<a

A and ':'hus

a acting as a load.
(ML

x 1(3) _
,~ 3EI
~

~~ 60L {5L~ 3a)

<I

2EI
"= 2'E~I ~Jj
~

6EI (Positive Shear Neg. Slope;

ML

x 2(3

wa

a a. 30L<I(lOL -15aL+6a)

p wa' / 20L

( 5L - 4a )

Let 98 Then 1

unity ML or
3EI

Mg

= :lEI

-tj~X(L-XI' ~ t{WX'(L-X)dX ~o
dx

Seam SA of Fig. AII.13. A ~c~ent applied at 3 produces no ~oment at A since end A is freely supported. Thus the carry-over factor for a beam freely supported at its ~ar end is obviously zero. Consider the beam of Fig. All.14. uue to complete fixity at end A, the slope of the elastic curve at A is zero.

r::-

stiff~ess

Factor of'

General case - Any loading

(Ref.N.A.C.A. T. N. #534)

A
or
MA

!1ll!:
2EI

1/3 + ~

2(3

=0

2EI

All.4

Stiffness 7actor; Derivation af:

Carryover ?actar: -

(For definitions of these terms see page

A11.2)
~Ohr's

Consider the beam of Figure All.13. By theorem (see Art. A7.l2), the slope ot a
EI is constant
L

2 Thus the carryover factor for a beam fixed at its tar end Is 1/2. Using the conventional ~om ent signs, the carry over moment is of the opposite sign as shown by the above equation. However, for our adopted sign convention inspection of the shape of the elastic curve as shown in Fig. All.14 tells us that the sign of the carrJover mament is of the same sign as the rotating moment at the ~ear end. That ls, the moment acti~g on each end of the ~ember 1s ir. the same direction, and therefore of the same sign.

= -~

~B
A Elastic Curve

""T

:;1" C'"""=---+---J....l
1.L 3 ML 21
Fig. All. 13

Fig. All. 14

IlJl = !1ll!:
2EI

2/3+~ .1/3=~+~, but

:1A=-

i1E
2

2EI 3EI

3EI 4EI

6EI

ThengB=~-~=~

l2EI

Let 9B

= urn ty,

t.lten~:a

= 4EI
L

sui r rness ract or-

of beam SA ot Fig. All.14. A comparison of the st1f:~ess factor of this beam to that of Fig. All.13 shows that the stif:-

All S

ness factor of a beam freely supported at its far end is 3/4 as great as one r ixec at its far end. j'ur-tnermor-e in one case the carry-over factor is zero and in the other case it 15 1/2. It is t.ner-er or-e obvious that the values or these 1;NO ter:ns depends in part upon t::-te restraint or degree of r ixat i cn of the far and of the beam.
All.S

; 2KF - 2F which is the general exMA ; 3+F K(3 + F) MB pression for carry-over rae tor for a degree of fixation F.

All.5a

Exa.ple ProblellLS

General Express1on~ tor Stiffness and Carryover Factor in Terms ot Fixation Factor (p) at Far End of a Beam

In the beam at Fig. AIL14 (F) fixation factor at A was unt ty since beam had been taken as conpletely fixed at A. It was found that:ME
;

To obtain a definite conception at the true mechanics of the "cross" ;nethoo, the reader is advised to follow thru the detailed solution or the follOwing simple problems. In these pr-oblems, the moment of inertia in any span has been taken as constant and all joints have been assurned to undergo no translation. Problems involving variable I and joint translation will be considered later.
.~./!Il.

4E19B
L
;

Te.ke 9B

Example Problem #1 unity and let EI = K for simplicity.

L

and MA

- 2 EI 9B

~A
~
s K

L
jStUfIl." hclDl' IOllltriblu1OCl Facl<ll' KI.l:K Carry Oo'er Facwr
1'lIed: End Mo",ent.

E2
~1.5 ~
~,
~,

,-W. H/in.

11 11

gil

]} tt tt tl
all

W!!'J
uc nc
.00

51.51

':'hen MB; 4K MA; -2K L~kewise the results for the beam ot F~g. A1L13 give Me = 3K I1A; 0 Figs. All.lS and AII.IS show these results. Fig. All. 17 shows the general case, the r ixation factor at A being F. The difference between F'fga , AII.IS and AII.I? is that the slope at end A has changed but eB the slope at end B remains the same.
F;i~B
tlA=O

iu
.51.5

.,

.515

'-88 7118=.OU3..Lot.:.7118 7ell=.0833Wr..z,,_711a1al13

lal BUanclng Carry 0.-.,.. 1/1

2IId Bll!aJlCUlIj:

1157<E 0 o
c

~-llS!O

57.5 _51.5

c
-710.51710. 5

FI.llaJ. Moll:lefttll
ConvellUO.... 1 MOlQIIDt SIilU'

.S8~

..
8113

." '"

o.rerlla:lg MOIll.nt : .5';51.5)[15.75 . 15x$1.5':17.11: alI3

"".

MA=~2K

MB=4K

Fig. All. 15

F=O

=.5 8B'"1, ~ MAsO MB-3K

Fig. All.

is

The change in :noment at end A when changing beam AIl.lS to that of All. 17 = 2K- 2KF ; 2K(1-F). Since 9B is xept the same value, onenarr of the moment change at end A appears at end 8 but of cppoat te sign, or Mg=4K-l/2 2K(l-F) ;3K-KF=K (3 + F) ;

ZI (3 +F)

F"F~
MA'= -2KF

eA' F(. 5)

MB-K(3+ F)

Fig. All. 17

Thus the ge!leral expression tor the stiffness factor or a beam of constant section equals EI (3 + F') ':'he carry-over factor from 9 to A
;

C-

Example Problem 1 shows a two-span cont1nuous be~ with over-hanging ends. '..le first assume all the joints B. C and D to be locked against rotation or the beams Be and CD are arbitrarily assumed in a fixed-end condt t t on , The !irst line in the solution gives the stiffness factor of each beam. Fr-om Art. AIl.S the general expression for stiffness factor is EI(3 + F)/L. where F ; fixation factor of far end, wni en equals , since all joints have been con" steer-ed rtxec. ThUS, stiffness factor K ; EI (3+ l)/L; 4. EI/L. Since EI/L is the same for each span Be and CD, the Stiffness factor has been wri tten as one since it 15 only the relativa values that are necessary. The cantilever SA has zero s t t r rneas , The second row gives the member distri but t on factor D at each jOint or 0 ; KlZK. ?or eXalIlple at jc ~::.t 3. the distribution factor to 3A = 0/(1 lI(c 0) ; ;- . At joint r !:-o. e 0) = o and to BC factor to CB ; 1/(1 + 1) ; .5 and distribution -c Lfkewt se to v The thll'd row giyes the carry-over factors. From Art. All.5, the carry-over factor C ; 2;;'/(3 + .5 assummg the far end rrxe d F'I = 2 x 1/(3 + 1) or F = 1 for each member. The next step as shown by the 4th horizont'11 row is the calculation of the r ixed end nonants. The signs or tl:e :n.oments are according to our adopted sign convention, t~t is an internal r-es i st.mg moment wni cn tends to rotate the

..

All 6

THE

MOMENT DISTRIBUTION

METHOD

end at a member clock'.... i ee 15 positive. We now begin the solution proper by first unlocking joint 8 from its as suzned fixed state. We rind a moment of -883 on one side and 765 on the other side of JOint or a sta t1 c unba.Iance of -115. Joint 9 therefore will rotate until a resisting moment of 115 is set up iY1 the member-s EA and Be. The resistance of these members to a rotat10n ot Joint B 15 proportional to their stiffness. The distribution factor based on the stiffness factors 15 o for BA and 1 for SC. Thus 1 x 115 = 115 15 distributed to SC at Band Ox 115 = 0 to EA. Joint 8 15 now imagined as again locked against rotation and we proceed to Joint C, which 15 now released from its assumed locked state. Since the jOint 15 already statically balanced, no rotation takes place and the distributing balancing moment to each span 15 zero. Next proceed to joint D, and release it. The unbalanced mament IS, 115 so the JOint 15 balanced by distributing -115 between DE and DC as explained above for joint 8. As painted out in Art. All.5, when we rotate one end of a beam It tends to rotate the far fixed end of the beam by exerting a moment equal to some proportion of the moment causing rotation al: the near end. For beams of constant section and rtxed at their far endS, the carry over factor 15 112 as explained before. Thus the distributing balancing moments 1n line 4 produce the carry-over moments as shown In line 5 ot the table. This completes one cycle at the moment distribut10n method, which is repeated unti~ there Is nothing to balance or carry-over, or in other worca until all artiticial restraints have been removed and the structure rests in Its true state of equt Li br-tun, To continue with the second cycle, go back to joint B and release it again tram its assumed locked state. There is no unbalance since the carry over moment tram point C was zero, thus there is nothing to distribute or carry-over. Proceed to point C, releasing the joint, 'He flnd 1t balanced Wlder the carry-over moments at 57.5 and -57.5. Thus the distributing balancing moments are zero. Joint D Is l1kewise in balance since the carry-over moment tram C 15 zero. All joints can now be released Without any rotation since all joints are in equ1ll brium. To obtain the tinal moments we add the original raxec end moments plus all distributed balancing and carry over momen ts .
All~S

Calculate the carry-over factor C for each end of a:t.l :nembers C = 2F/(3+;-). Thus ror beams f~xed at far end F = 1 and thus C = 112. For pinned at far end F = 0, hence C = O. 5. Calculate the fixed end :noments (t) for given transverse member loadi:Igs or support oer rections, using equations in summary ~able All.l. ~~d ~oments which 'tend to rotate end of :Jearn. cIockwf se are ~ositive moments. (See Art. Al1.2) 6. Considering one joint at a time, unlock it from its assumed fixed state, all other joints remaining locked. It an unbal~~ced moment exists balance It statically by distr~butlr:.g a counter acting moment ot oppOSite sign among 'che connecting members according to their D or distribution factors. 7. These distributed balancing moments produce, carryover moments at far-end of member-s equal to the distributed moment t taes the carry-over raetor C and of the same Sign as the dlst:oibuted moment. Record these carry-over moments far ends for all distributed moments. 8. Repeat the pr-oceedur'e of unlocking each joint, distributing, and carrying over moments until the desired pr-ecf s t on Is obtained, stopping the solution e rt.er- a distr~bution. The final moment at the end of any member equals the algebraic sum at the original r fxed end Jl.Oments and all distributed and carryover moments.
4.
,~

_0

~ample

.SJ/lIl..

*
t,

Problem #2

f pfli j t I , I ! r j it! r 1111 t t! ; t 1, Ij , ''-''''

lJ/lll.. EI l.o Can.w>t

il:~

r:ac'lar

}i.S+---~____t__72 " I'

.!!.~ K

-L
.

12

I'
:

ge------J..s l . ,~
I

F!

q
,

O.OIIl4',*' . 011l41. 013 ...Lz. 013i1 .013il'~',OI39!.011l4

.~

Q1Q~1

D~~:'IOI>~
CarrrOo.r C factor
fl.om E%ld
IOI_C.....

,I.
e .s

.~:l8!. S72
sl.s

"

I
110

I I I
,

.s.s sl.s
I

.sn!.Hs

-~"

_a,

...

'I'
i
I

l&S ..l...L1 ,lllliu:lOl."Lz-'_dzi43:z--.l. .. L1'_43 z lZ ,11 _ 12

l ,e8 "-!.."Ll _ 7M1883 11

Carrr-Qru
:lod8ll\alw:lIl.t:

n Il5 72

1lm 57. slo

::::><:

0 01 961_9&

X~

c _S7.S

u _72

X2~_5!'32_9
_3slo
6.110

~X2.9!24.6~
X;6.5116.5 0136

~><~
_121: 7210 12.31 _:71

car,.,-Oo...

(2.3

3"" 8IIl:uI<:.....
car,.,.OvJr
~tlI

0,IZ.3~lS.ti20.6~~X:.1.1S.?::<11.JIO
7.1 10.l!.10.3 *6,1

8IIlaAc .....

01_1.7~_1.6 ::':'X~X.SZ.6
I-l.] olt.3
.~318U

e.... , ..,oou
Sib SOWle ....

.~ o
I.

~r,:io
I.J!
_l.J10 _8331883

_1.611.8
~

013.9
_1.1 _I. 1 _610. tlalO. 4

7 1.

ll.oal..,..._ ..

,"
_3+413+4

-erc 41610,4

Ie_Ilona! SilPUl
Ge 1:1.8 ra 1 SWIIII&ry of Proceedure

1.

.1.

.1.

.1

'1'

Example Problem 2 15 similar to prOblem 1 but two All computations should be written on, or spans have been added. ',oj'e first assume all joints adjacent to the d1agram. of the structure. locked against rotation. The stiffness factor of factor K tor each 2. Determine the stittness each span 15 :?roportional to EIIL or lIL since EI end at each member. K = (3 + F) EI/L, where F constant. The car:-y-over factor 15 1/2 as in is 15 the degree of fiXity at far end ot member. previous example. Fixed end moments are calculatIt all members are assumed rrxec at far end than ed as shown. Unlock jOint E, the unbalanced mo:nK 15 proportional to IlL assuming E as constant. ent 15 -U5. Balance the JOint by distributing 3. Determine the distribution factor D tor each to BC and zero to EA. Proceed to joint C, 1 x 115 member at each joint of structure. D = K/l](.
1.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All 7

and unlock, all ot~er joints remaining fixed against rotation. The unbalanced moment 1s (-768 + 432) = - 336. Balance by distributing .428 x 336 = 144 to CB and .572 x 336 = 192 to CD. Proceed to joint D, and release. The unbalanced moment is zero which means that joint D is in equilibrium. thus no distribution is necessary. Proceed to joint E and F and balance in a similar ~er. The distributed moments will be the same as the values for joints Band C due to symmetry of structure and loading, however, the signs will be oppOSite under our adopted sign conventions. The next step is the carry-over moments which are equal to 1/2 the distributed balancing moments. This operation is shown clearly in the table. Values of all moments are given only to first decimal place. The first cycle has now been completed. Cycle ~NO is started by again releasing JOint B. We find the joint has been unbalanced by the carryover moment of 72. Balance the joint by distributing - 72 x 1 = - 72 toBC and zero to EA. Proceed to joint C. The unbalanced moment is 57.5. Balance by distributing - 57.5 x .428 = - 24.6 to CB and the remainder - 32.9 to CD. Proceed to jOint D. There is no unbalance at this joint since the carry-over moments are in balance, thus no distribution is necessary. Proceed to joints E and F in a similar manner. The carry over moments equal to 1/2 at the 2nd set of balancing distributed moments are now carried over as shown in the table. The second cycle has now been completed. This operation has been repeated five times in the solution shown, or until the values at the balancing and carryover moments are quite small or negligible The final ~oments equal the algebraic summation or the original fixed end moments plus all distributed and carry over ~oments. One requirement of the final end moments at any joint is that the algebraic sum must equa L zero. The other ~equirement consistant with the cammon slope to all members at any joint is given by equation (5) of Art. All.S. The results at jOint C will be checked using this equation.
.5 t. l",jc = ~ t.l1cb - .5 Hlbc Kcb SUbt. values <ll'Igd -

necessarJ. These modifications usually involve rather long expreSSions :or expressing the stiff ness and carryover factors of a me~ber in terms at the fixation given by adjacent members. It 1s felt that it is best to keep the method in its simplest fOrn which means that very little is to be remembered and then the method can be used in frequently Without refreshing ones mind as to many required formulas or equa't Iona , There are however several quite Simple ~odi f1cations which are easily understood and remembered and which reduce the amount ot arithmet ic required considerably. For example in Problem 12, joints Band F are 1n reality treely supported, thus it is need less arithmetic to continue locking and unlocking a jOint which is definitely tree to rotate. Likewise due to symmetry of structure and loading it is only necessary to solve one half of the structure. Due to symmetry jOint D does not rotate and thus can be considered fixed, which eliminates the repeated locking and unlocking of this Joint. A second solution of Problem #2 is given in gxaap te Problem #3. As before we assume each jOint locked and calculate the fixed end moments Now release or unlock jOint B and balance as ex platned in previous example #2. Before proceed ing to jOint C, carry over to C tram B the carry over moment equal to 115 x 1/2 = 57.5. Joint B is now lett free to rotate or in its natural COndition. Proceed to joint C and unlock. The unbalanced moment = (-768 + 432 + 57.5) =- 278.5 or 278.5 is necessary for equilibrium. This moment is distributed between eNO beams, CD which is fixed at its tar end D and CB ,Yhich is treely supported at the far end B. The stiffness factor is equal to (3 + F) EI/L (See Art. All.5). Hence for CD stiffness tactor = (3 + 1) EI/L = 4 EI/L. For CB stiffness factor = (3 + 0) EI/L = 3 EI/L or in other words the stiffness ot a beam freely supported at its far end is 314 as great as when fixed at its far end. Thus the stiffness factor at CB at C is .75 x .0104 = .0078. The carry-over factor C to B is zero since B is left free to rotate. (See Art. All.S) Example Problem #3. Simplified Solution Of Problem /12
l'IUI.

(610.4-432) -.5 (344- (-432)J _178.6-44 [-610.4 -(-768)J - .5 (883 -768) 107.6-57.3

.~.

1.343

..., r,:-H.

FlU<!,

Ratio of stiffness factors = ~ = ~ .0104 = 134 . Kcb Thus the distribution is according to the K ratios of the adjacent members. Simplifying Modifications - Examule Problem #3 The solution as given in Problem #2 represents the "Cross" method in its flli~damental and most elementary detailed form. Many modifications of the general method have been presented, in the most part for the purpose of eliminating part of the arithmetic or the n~ber of cycles

.
_
~T.S

12~

o . QHl4=ft
Oi",,.,butIOll F1cwr o , !t/!x:

.1$,.. 01Q.p. 001"8 . ou~ ~

lin 2

O\:N IC<:

"
,

Cu..,.""u lK,or (C)

flU<!.t>d mo",onUl
1lI~'iO",'f1

uu c s
.;183 163
'J
ll~_

. le .IM

cu
-4U

c .s
.1e8 432

Ca,..,.ov.r l,2 '0 C B:l!anc. 10"" C Ca,..,.""" to 9. 0

B.t.J..:u>c. ,<lUll 0

Car..,_onr to C
Final mOm.nUl Con.enuon:U &lll'l_ .a831883

..
I
~

c, ..

All.8

THE MOMENT DISTRIBUTION METHOD

The stiffness factor 0: ~he fixed support 1s in- SoL.;t ion #1 finity, that is a rigid support r~s ~n:inita re- zxaco;e Pr cb Lem ""4 sistance to rotation. Joint C 1s now balanced by distributing 278.5 x .36 = 100.1 to CB and 278.5 x .64 = 178.4 to CD. Now carryover to joint 8, a x 100.1 = Stiffness Factor k a and 0.5 x 176.4 = 89.2 to O. Proceed to jOint Oistrtllutioa F:Lctor O-rt o and release it from its assumec locked state. Carry-Over Factor The unbalanced moment is -432 + 89.2 = - 342.8, n"ttd End Moments 1st we balance by distributing 342.8 be~~een DC Carry_Over which has a stif!ness of 0.139 and the support :!nll E which has an infinite stiffness or zero goes Carry_Over 3rd. Ba~ClJ'tg to DC and 342.8 goes to rigid support. ~he carCuryOver ry over moment to C from D is zero since 0.5 x <lUI BalanCing o = O. The t~na1 moments thus equals the sumC.. rryaver 5th Balancinlj: ~tions as shown which of course are equal to thE FlAal l.lomenUi results shown in Problem #2. Cmlventional Signs
a..~CUlg BIil~clng

0.0104' 1/96' .0IO'lI.0139_l/72'.0139!.OI39

0.>

.a831768 01115

112
0172 '-35.3 Oi35.3 11209 OIU.9 ,.6.3

5 .5 _7681<132 1<141192 s7.si108 _70.71_94.8 36. QI.2<I. 0 25.713<1.3

_<1321

"

2l6i21~

17.,111.8
-12.61_16.9 _6. <11_<1. J
~.

-48148 _<17, ~. 23.7,23.7 11.2 .8.618.6 _3. ~i 4. 214. 2 -187.31187.3

'"

016.3 _8831883

616.

.i.

_64<4.~I~H. 3

Example Problem #4Problem 4 is similar to Problem #2 and #3, except that the support at 0 is assumed as having 50 percent fixity. Thus 50 percent of any moment at this point produces rotation of the member DC at D. In continuous wing beams, which fasten together by fittings on a support, it is commonly required that the beam be considered as being tully continuous and also tna't the degree of continuity be taken as 50 percent. Solution 1 of PrOblem 4 1s a detailed solution. The only change that has been made is in the stiffness factor of the support E, Nhich has been taken as equal to the beam DC, thus ~~ unbalanced moment at this poi~t is equally divided between the beam and the support. Solution 2 is a ~odi!ied solution which elimlnates considerable arithmetic. Thus it is unnecessary to lock and unlock joint Band D since ~e know definitely that one is freely sup~orted and the other 50 percent fixed. ~here fore, once we have r'e Leaaee these joints from their assumed fixed end state, we leave them in their natural state. The stiffness and carryover factors for beams CB and CD ~ust then be cetermined for these beams with their ~odified end conditions. By reference to the fundamental equations for stiffness and carryover factors in Art. All.5, it 1s readily seen that the stiffness factor for C8 1s 3/4 as ~uch as when fixed at its far end B and the carryover moment is zero. For beam CD the stiffness ts 7/8 as much as wnen fixed at end D, and the carryover factor is 2F/3 + F = (2 x .5)/3 + .5 = .286. ~ith these modifications the solution Is carried ~hru with a relatively small number of steps. ~hus in solution #2, joint 8 is unlocked. The unbalanced moment of -115 is balanced statically by distributing 115 to BC. The carrJ-over moment of .5 x 115 = 57.5 is carried over to C as shown. Joint B is now left unlocked or free to ~otate. Joint 0 is unlocked next. The unbalanced moment is -432. It is balanced by distributing 216 to DC and 216 to E since support at E is considered to ~iYe 50% fiXity. ~he carry-over moment of

.5 X 216 = 108 is carried over to C. Joint D is left ~~locked or in its true state of ~estraint. Joint C is now unlocked. The unbalanced ~o~ent solution #2
SxampLe

Pr-ob Iem 11'4

SUffn ....s

hotor K

Olst,..bIltlon Fa",o

i1c
liZ

Carry_over Fac,o.
Fixed End Mom..."" 9alanC. Jain' II

-a83j1Sa
ill t 15

Carry.o..... to C
BalanCe JOIRt 0

Cury.o..... to C
Bal:lnce JOin, C

C.. rry_ove.:o 0
Elal:lnc. 0 FIn>,j Mom",,-t.o

equals (-768 + 432 + 57.5

lOB) = - 170.5, or

170.5 is necessary for equilibrium. Joint C 1s balanced by distributing .392 x 170.5 = 66.8 ~o

CB and the remainder of 103.7 to CD. The carryover moment to B is zero and to D it equals 103.7 x .286 = 29.6. The final moments in sol~ tion i W are slightly different than solution #1. If another cycle had been added in solution #1 the descrepancy would be conSiderably smaller.'
All.7
Continuous Beaas with Yielding or Deflected Supports

In Wing, elevator and rudder beams the support points usually deflect cue ~o the defo~2 tion of the supporting str~ts or Nires in the case of a wing, or to the deflection of the stabilizer or fin in the case of elevator and rUdder beams. If these beans are c~ntinuous this deflection of their support points causes ~cdi tional bending ~o~ents in the jeams. The moment distribution y.ethod can of cou:se be ~sed to find the addi~ional ~oments due to this deflection. T~us Exan~le Problem #5, shows a solution illustrating a ~roblem which involves the de:lecting of the supports of a continuous beam. Due to sJ~etry of structure and loading, the Slope at D 1s zero or t~e oeam way be cons1cered fixed at

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

All 9

Joint D. Since the mOment of inertia is conSince the fixed end ~Oments are due to bot~ stant and the spans are cons tant , the relative lateral loads and support def:ect~on, the values stiffness factor of the beam is 1. In the soas listed in the solution table Nill be explainlution shown since beam ~s freely supported at ed in detail. 3 t.nt s joint is lett tree to rotate after reFixed End Moment For Lateral Beam Loading leasing and thus the stiffness factor of beam CB Is 3/4 x 1 = 3/4, when compared to one haVing The distributed airload is trapezoidal in full fixity at B. shape. The fixed-end ~oments for a trapezolcal Since the first step in the solution proper loading from Table All.4 are: 1s to assume the joints fixed against rotation, Ml_2=L.a (5u+2V) (See Fig. (a)) it Is evident that deflecting one support rela60 tive to an adjacent support will produce moments at the ends which are assumed fixed against M2_1==L.a (5u+3v) (See Fig. (a ) I rotation. 65
All.a

Fixed End Moments

Due to Support Deflections

..L_
v

-- --

Ef ,~

M ~

ll.6 ~~
I'l

ML 4EI

i~ ~ -~~
1

1 .

6 Fig. All. 18

J L

Fig. All.IS shows a fixed end beam. The support B -Ls deflected a distance d relative to the support paint C. It the me:nber ls or constant
at the beam mid~oint and ME - Me. By moment area prln-

T
u

-=

Fig. (a]
L

-L
f

CD
For Span Be:
!1bc !'leb

"

~~O~~~~:~~~~~ -~~~lP~~~i
5 L
(3

40'

60
40'

(5x3+1) = 426
(5x3+1.5)

~n.

lb.

cipal deflection a !1L 4El or d == ML a 5El hence

6EIa

L -J1L
(3

60
40'

440 in. lb.

4El

For Span CD:

I'lcct

60 (5 x 3.5 + 1 I
40'

= 494 In. lb.

-V

'tha.magnt tude for the fixed end

Mct.c = 6'Q" (5x3.5+1.5)

= 507

in. lb.

mO:nent due to a transverse support settlement of d. Examole Problem #5. flected supports.
General data; 3 f/in.

?lxed End Moments Due to Sunnort Movement From Art. All.S For span Be: Deflection of B relative to C
!'!be = Me'::l

l1

= ~Eld/L.a = 5/16
inch. Hence,

Continuous beam with de4 '''/In.

049 3'/A1 . U1;J (Alloy tube j , tt.:,1:.03339

It-

o ": ,
-jlOj-- 40

I!

! ; ;
---r-I

=6 x 10,000,000 x

.03339 x .3125/40.a == 390

ABC
t

3/ 16 "

40 - - 40 40 o Defle<:,tion of Supp?rt!l

--110~- -f'11e<I due 10

in. lb. F'Qr Span CD:

?'lcd

DI

---

=t""'dc :: 6 x
1::1.

10, 000, oeo x .03339 X .1875/40

;:: 234

Solution:

A iB
3/4x1~. ~ (t,i

&)

10.

Stiffness factor K Distl'1.bution Factor C.u-ry.over Factor

ZK

011 1/'
50 426

75 1 . i21l .571 o 1/2

n
112

Ftxed Due to End Lateral Load Momentll Due to support deflection

Balance joint B Carry_over 10 C Balance joint C Carryover ~o B. D Balance joint 0 Final Moments Conventional Signs

o 390

390 234

,,.

507

161
50 _50

o 580 580 580

For si;~s of tte ~oments due to tr.ese deflections see Art All.2. Havir~ dete~ined the fixed end ~ooer.ts t~e ~eneral distributing and carrYing over process follows as indicated in the solution table. Thus at joi~t B, the unbalanced moment :: (50-426 -+- 390) :: 14. Balance by df s tr-t out t ng - 14 xl:: - 14 to Be and Zero to EA. carryover.5 x -14 :: - 7 to C. Considering joint C, the unbalanced moment == (440 + 390 -+- 234 - 494 - 7) :: 563. Balance bYdistri but tng - 563 x .571:: - 322 to GD and -563 x .429 :: - 241 to CB. carryover.5 x -322 :: - 161 to D. A~ joint D the unbalanced ~oment = (507 + 234 - 161) = 580. This is balanced by distributing zero to DC and -580 to the fixed support.

All.I0

THE
Check on F1nal Moments

MOMENT DISTRIBUTION METHOD

&1~b .5AMha = ~ _ - -(5) M1ac - .5AI1sa Ka c :: equation (5) is satiSfied the accuracy or the moment distributions is established.

111.8

To satisfy statics, the algebraic sum at the moments in all the oe~bers at a joint must equal zero. This requirement alone will not prove that the final moments are correct, as errors could have been made in the various distributing and carryover moments. continuity tells us that the final slope of the elastic curve at anY joint is common to all ~ambers meeting at that joint, thus for a complete check it is necessary to prove that the final moments are consistent with equal slopes for the various members. The rotation of the joints from their original assumed fixed or locked condition is due entirely to the distributed and carryover moments. The actual rotation will therefore equal the rotation of the end of a simply supported beam when subjected to end moments equal to those produced by the algebraic sum of the distributed and carryover moments. These end moments equal the final moments minus the original fixed end moments and will be referred to as AM moments. From Chapter A7, the slope at any point on a t h simply supported beam equals the shear d ue a ~ e M/EI diagrams as a loading.

All.9

End Moments tor Continuous Frame.orks 'bose )(e=bers are not in a Straight L1ne. Joint Rotat1on Only

Continuous truss frame works are quite common in aircraft construction. Welded steel tubular fuselages are composed of ~embers ~hlch maintain continuity thru the jOints due to the welding. Landi~ gears frequently :~ve ~NO members which are made continuous at their points of connection. The members of such structures usually carry high axial stresses which cause joint translation which in tur~ prOduces bending of the members since the joints are rigid. ~hese moments produce lateral deflection of the members which introduces additional secondary moments due to member axial loads times the lateral member deflections. These influences are treated in later articles. In this article to further familiarize the student with the moment distribution procedure, the effect of JOint translation and secondary moments will be neglected.

The slope QA from Fig.a! equals the beam shear at Ex8!llple Problem #6 i--L/3=i 0 MabL A or equals the reaction Fig. All.l9 illustrates a simplified land1ng ---.--2EI , at A. gear chassis problem. Let it be reqUired to dec.Mab ,. termine the bending moments in the two ~embers Ell 'C.g. due to the vertical load on the axle. The probB lem has been solved using ~hree different degrees ot restrain~ at ends A and B. Joint 0 is 1-1----- L a welded joint and full continuity is assumed thru this joint. The solutions as given in Fig. All.20 give only the moments due to jOint rotation under prumary bending moments. The effect ~c. g. of axial deformation and secondary moments due i to member deflections is omitted in these solu-_!-I-L/3 tions. These factors are treated in later articles . .o.MbaL 2Er" In a practical problem the degree of restraint at points A and B would be dete~ined by 2/3 _ <IIlbaL ~ 2E1 3 the type at tittlng used and also on the r1gidity of the adjacent fuselage or Wing structure. As (&lab + .5 <IIlba) illustrated in later ex~p~es, the moment distr13EKab bution method permits the consideration at the where rigidity of the adjacent structure without adding I any diffiCUlty, While such methods as least work

-AI

Kab

= L or

AB

Since the angle QA must be the same for members meetir~ at A, the general relation bet;.... e en the moment increments at any t'NO members such as AS and AC must be,
&lab + .5 <IIlha M""-ac + .5 &1 c a

1S"

32"

=~
Xac

T .
N

To make this equation consistent with the assumed Sign convention, that is, the carryover moment has the same Sign as the balancing distributed moment, the above equation must be modified as follows:-

Fig.

Ail. 19

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

All.ll

to be freely supported, they will be left in their true state. Thus the carryover moments from end 0 will be zero. Since A and B are both pinned, the relation between the relative stiffness factors of members OA and 08 remain the same as in condition I, thus the same K or stiffness factors that were used in condition I can be used in distributing moments at joint O. Joint 0 is balanced in same manner as condition I but with zero carryover moments to A and B. Solution for Condition III. A is 50% Fixed and B is Pinned. Since each member has a different degree of fixity at its upper end, the stiffness and carry over factors will be considered in detail. In condition I since both members were fixed at their upper ends the relative stiffness factor of each member was proportional to IlL tor the member and this ratio was used. The general expression for stltfness factor is K = EI (3 + F)/L carry over factor 2F. For member OA, F
3+F

equals .5 Since A is 50% fixed, and for member OB, F is zero since 8 is treely supported. Hence for member OA
K=EI(3+ .5) =~=3.5X .1105 E= 0129 E L L 30
K :II Stiffness Factor Values in ( ) are carryover factors
Fig. All. 20

C.O. Factor from 0 to A = 2 x .5 = .286

3 + .5

For member OB
K =

are not practicable because of the large number of equations that must be solved to obtain values tor the many unknowns.

g L

(3+0) =3x .185xE = .0139E 40

Solution torCondit1on I, Fixed at Ends A and B Referring to Fig. All.20, all joints are assumed locked against rotation or fixed. The vertical axle load of 6000# produces a counterclockwise moment of 3 x 6000 = 18000 in. lb. about jOint O. The Sign is positive (See Art. All.2). Release or unlock jOint 0, the unbalanced moment is 18000 or - 18000 is required for static equilibrium of joint O. Joint 0 is balanced by distributing - 18000(.464/.464 + .369) =-10030 to member OB and the remainder or
or -18000(.369/.464 + .369) = -7970 to OA.
~hese

C.O. Factor From 0 to B = 2 x a

3+0

=a

Considering joint a in Fig. All.20 the external moment ot 18000 in.lb. is balanced by distributing - 18000 between the two members in proportion to their stiffness factors. Hence - 18000 (.01291 .0129 + .0139) = -8650 in Ibs. is resisted by OA and the remainder ot -18000 (.0139/.0129 + .0139)= -9350 to OB. The carry over moment rrca 0 to A = .286 x -8650 = -2475 and zero tram 0 to B. (See
Fig. All.20)

Example Problem

#7

distributed balancing moments at 0 produce carry over moments at A and B. Thus carryover to 8 . 5 x - 10030 = - 5015 and carryover to A, .5 x - 7970 = - 3985 Proceeding to joint A which is a fixed joint, the unbalanced moment of -3985 is balanced entirely by the rigid support, or no rotation takes place when joint is released from its imaginary fixed state. Similar action takes place at joint 8. The final end moments are as shown in the Fi~Jre. Solution ror Condition II. End A and B Pinned For this condition the ends A and Bare freely supported. Instead of locking and unlocking these joints which are definitely known to be free

Fig. All.2l shows a structure composed ot 3 members. Member AO is subjected to a transverse load of 120#. Joint A is fixed, B is freely supported C is 25 percent fixed and joint 0 is considered to ~intain continuity be~Neen all members at O. The end moments on the three members due to the transverse loading on ~ember AO will be detennined. Solution #1. Fig. Al1.21 gives a solution USing the "Cross" method in its fundamental unmodified state. The solution is started by assuming all three members as fix-ended. The relative stiffness factor K of each member is therefore proportional to IlL of each member. These K values are listed in ~ig. All.21. The distribution

All. 12

THE

MOMENT

DISTRIBUTION

METHOD

c. O. Factor Member - K. Value factor D for each ~ember at each ~oint which 2F equal K/~ is recorded in c=lon each member' ~(hF) 3- F around each jOint. Thus any balancing ~oment is distributed bet~een the joint members as per 2x1 =1 OA:-l..(3+1J "'.4, 2 these d1stribution factors. The carryover ~ 20 factors for all members is 1/2. The fixed end 2xO OB.i~{3+0) :.3, 3 +0 0 moments due to external loading are computed tor the three members. For member AO, the tix1 2x.25:. ed end moments equal PL/8 = 120 x 20/8 = 300 in. OC=25{3~.25)=~ 3 . 25 154 lb. The other two members having no transverse ZK=.83 loading, the fixed end moments are zero. In this solution the or1er of joint consideration r~s been AOEe and repeat. starting With joint A the joint is released but since the member AO is actually held by a fixed support, no rotation takes place and the balancing moment of 300 is provided entirely by the support and zero by the member AO. The carryover moment C, to 0 is zero. Releasing joint 0, the unbalanced moment of 300 is balanced by distributing - 300 between the three members accordSOL tJrION .2 ing to their D values, thus - 300 x .416 = Fig. All. 22 - 125 to OA; - 300 x .168 = - 50 to OC and - 125 to CB. To prevent contusion it is recomby distri but.fng .75 X 25 =18. '75 to CO and the r-e-. mended that a line be drawn under all distrimafnder- of 6.25 to the support, s tnce the fix!.. ty -r; buted balancing moments, thus any values above the support at C has been assumed as 25 per-cent _ A these lines need not be given tur~her considline is cra~m ~lde~ t~e 18.70 and the ca~ry over eration and only values below the lines need moment of 9.37 is taken over to 0.- One cycle has be considered in later balanCing ot the jOi~tS. now been completed. Retur~ing to ~olnt A, ~e Immediately a:ter dlstribu~ing the moments at find -62.5 below the line. This is balanced by joint 0 the proper carry over moments should be distributing zero to OA and 62.5 to the t!..xed taken over to the far end ot each member, thus support. A line 1s drawn ~~der :he zero distri- 62.5 to A, - 62.5 to Band - 25 to C. Joint buted ~oment to AO and the carryover woment of 8 is next considered. The unbalanced ~oment is zero is placed at O. Considering joint 0 f~r -62.5 and it is balanced by distributing 62.5 the second time the unbalanced mo~ent is 9.37 + to 80 slnce the pin support has zero sti:tness, 31.25 + 0 = 40.62 or the S~~ ot all vaiues below or no resistance to rotation. A line is erawn the col~ horizontal lines. The joi~t is balunder the 62.5 and the carryover moment ot anced by 1istributing - 17 to OA ~~d 08 ~~d 31.25 is placed at O. Joint C is considered -6.62 to OC. Lines are dr~Nn under t~ese balnext. The unbal~lced ~oment o~ -25 is balanced ancing momen~s as shown in Fig. All.21 and t~e carryover mcments are taken over to the far SOL111'ION'IIl eTIQs Jefore proceeding to joint 3. This general process is repeated until ~oint A has been balanced 5 times and the ot~er jOints 4 ti~es each, as indicated ~n the figure the distributl~g values have become quite ~ll and it SWmellB Factor K [ a is evident tha~ a ~igh degrs8 of accuracy has OA=I:- 10 : .10 been obtained. The ~inal end ~oments at each 5;,lO 08= lis joint equal the algebraic sum of the values in . each COIWL~. A double line ~s placed above the DC = 225 =~ r inaa noaerrts as a df s t tnguf sht ng symbol . In ZKJointO=,24 C.O,F",ctor_1/2 the fi~~e the letters band c ~efer to balancing far All Memo. and carr; over ~oments, the sUbscri~ts ~etarring to the member of the balancing or carry over o=eration. Any order of joint consideration can be ~sed in ~eaching the same result.
Sol~tion

#2 of

~oblem

Fig. All. 21

Fig. All.22 gives a second solution. ~ith the end conditions ~own at A, Band C, the ~odi ~ied stitfness factors of the members can be found together with the ~odit1ed carryover factors, thus making it ~ecessary to balance ~oint o only once and car~y over this final far end

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

All.13

of each member. The fib~e gives the calculation of the modified or actual stiff~ess and carryover factors. ~ith these !u~c,vn the solution is started as before by computing the fixed end ooments due to transverse loading on ~e~ber AO. Joints Band C are released and si~ce no fixed end moments exist, no balancing 1s required and the joints are left in their true state 0: restraint instead of locking and unloCki~ as in solution #1. Releasing joint 0 tram its imaginary fixed state the ~~balanced moment is 300 which is balanced by distributing - 300 betNeen the 3 connecting members according to the new distribution factors at joint o. Thus - 300 x .482 = - 145 to OA; - 300 x .361 = - 105.2 to OB and - 300 x .157 = - 47 to ce. ~~e carry over ~oment to A = - 145 x .5 =- 72.5 t.o B = - 108.2 x 0 = 0 and -47 x .154 = - 7.2 to
~oments

assumed to give 50% fiXity to these joints. In Table A11.2 a modified stiffness factor is calculated tor ~embers GI, Fl, and FH using a 50 percent fiXity at their tar ends. The last column of Table All.2 gives the s~tion of the ~ember stiffneSS factors for members intersecting at each jOint.
120* 3200"-' ,.12'",
J.

l-

~:/7\
Engine A

MOWlt 2400\

l~7;''t ~~-~ I~ ,~~ --t1d'rF

/\ /
/
12400".

~~~:'~d
H

'\

Problem #8 figure All.23 shows the forward portion of a fuselage side truss. Due to eccentricity of engine 2o~,t and la~ding gear members, external ~o~e~ts ~e produced on joints A, Band D as shown. Fur~he~ore lateral loads due to eGuip~er.~ installation are shown acting on ~embers BE ~nd CD. Assuming the fuselage welded joints pr ocuce rigi::i continuity of member-s thru ttie joint, t~e problem 1s to find the end moments in all the ~a~bers due to the_eccentric joint ~om en~s ana two lateral loads. The effect of joint tr?-nslation and secondary ooments due to deflect10TIS and axial loads is to be neglected in this ex=..m.ple. So1.:J.tion: ~able All.2 gives the calculation or the stiffness factor for each truss member. The ~uselage truss aft of joints I and H tave been
Exa~ple
_1.98 ( .1.82 b

"'\ -, f
-,
Leuct h
L
2~"

Fig. All. 23

Landing Gear

TABLE AiL 2
1. x
1000

.
1-1

it1ffA
J:. J'aetar

F-''-,-...,-,----j
A IS C

, 1.0240 ,0< , .0240 - ,0<' 34.' .0588 AC - .049 19.25 .0588 2 .049 30.0 .0588 - - ., ,.. _3 .0449 - 34. - 3 .033 - ". 1_1 _1 2 - .049139.5 Be 0178 o 8 1_ 8 _

cr

.049 41.

.,,240

0.582 61

1 _1 2

.S18*
3.06 1 6

S 80 4 81 6 95

D
7

D'

..
:I

,sa
4.42 3.21

6.14

1. 96
2 D.

(lSBB

ell

(I:l~141

o s 0 ,

..

(II: 1. eoutaJIt

fol' Lll _ -

>en)
~OS

.t1ffn fLctor _ 7/8

~ beca~

of

f1xity)

Fig. All. 24

.,84

-"

. -,

(n5

-141 b .178

_187

18 b

1)---i

0-'.

"

c-t

e , ae

b
b

.", "'-, --. '-." ...:l.o

-:....11.
'-.-.i..

.286

-'-' .195

0-'.l15

"' ,

....

"'."' ......,..,....
:::~

"

sr- -s

.
__ .834

:...JJMb

na

--o...ill

~b

70 ',0 b ~

140 ~

aa

. ':""""55'

"

-10 . ~b

.,29 b

_ _ b 140

--1)15

m"

..
n

->;.286

b-rot

-,

~"

.\11.14

THE MOMENT DISTRIBUTION METHOD

Fig. All.24 gives :he solution of the problem. The procedure in this solution was as follows: The stiffness factor K for each member as computed in Table All.2 15 recorded in the circles adjacent to each truss member. The carry over factors for all ~embers 15 1/2 except fcr modified members GI, FI and FH for which the carryover factor to the 50% fixed ends is .286. The distribution factor for each member at each jOint is recorded at the end ot each member, and equals K/LX. The next step in the solution is to cocpute the flXed end moments due to the transverse loads on members. For member BE MBE=Paba/L li =120x29.25x12.11/41.2S ll =-2J8"#

8Y distributing 183 to AC and 15~ to AS with carry over mo~ents of r~l: these values to ends C and B respectively. The student should now be able to c~eck the rest of the solution as given on Fig. All.24. The solution could be ~de with any order of joint consideration. If any particular jotnt appears to be nearly balanced, it 1s best to skip it tor the time being and consider those joints which are considerably unbal~~ced. 7he ~inal moments at the end 0: each me~ber are given below the double :i~es.

Example Problem #9 Fig. All.25 represents a cross section of a welded tubUlar steel tuselage. The top and bottom members which are web members in the top ~~d bottom fUselage trusses are subjected to the l1E:B =120 x 29 .25' x 12/41.25' =725"# equipment installation transverse loads as shown. l1cD = 100 x 20' x 10/30' = - 445 Let it be required to deter.nine the end bending moments in the rectangular trams due to these l10c = 100 x 10' x 20/30' = 222 transverse loads assuming full cont:nuity thru These moments are placed at the ends of the joints. members on Fig. All.23 together with the eccenSolution: tric joint moments. The process of ur~ocklng the joints, distributing and carrying over momFig. All.26 shows the solution. The dlstr1ents can now be started. In the solution as bution factors based on the member sti!fness facgiven the order of joint consideration is tors are shown inD at ends of each member. The ABCDEFG and repeat, and each joint has been bal- first step 1s to compute the fixed end moments anced three t1.mes. due to transverse loads, on members AS and CD usConsider joint A:ing equations from 7able All.l. The ~gnltudes Unbalanced moment = 2400. Balance by disare 1890"# for AS and 2025~# for CD. tributing - 2400 as tollows:Joint B is now released f~om its assumed To AC = - 2400 x .527 =-1258. Carry over fixed state. The unbalanced moment of 1890 is to C =-634 balanced by distributing - 1890 x .247 = - 467 ~o To AB 2400 x .473 = -1132. carry over SA and the remainder of -1423 to BD. T~e carry to B =-566 over moment to A = - 46S-x .5 = - 233. Due to symmetry ot struc~~re and loading only one t~lt Proceed to Joint B:of trame need be considered and hence these carUnbalanced moment = (-566 + 3200 - 298) = ry over moments to A are not recorded. However, 2336. Joint 1s balanced by distributing - 2336 in balanCing joint A 1t will throw over to 8 the to connecting members as tollows: same aagrrrtuee of carry over moments as thrown To BA =- 2336 x .569 =- 1330. carry over to over to A from B but of opposite sign since the A = - 665 original fixed end moment at B is minus. Thus To Be = - 2336 x .310 = - 724. Carry over to 233 comes to 8 tram first balance of A as shown C - 362 in the figure. The distributing moment to a of To BE =- 2336 x .121 = - 282. Carry over to -1423 produces a carry Qver moment of -1423 x .5= E = - 141 -712 at D. ~27" I 300# 300# The convenient device of drawing a line 7" 10--t-r-10"-j under all balancing moments is used to prevent contus1on in later balances of the joint. 1"-.035 Tube Proceed to Joint c:K= 0~~37::. 000458 Unbalanced moment = (-634 - 362 - 44S) =-1441 The joint 1s balanced by distributing 1441 as tollows:0:1 0 0:1 To CA:= 1441 x .44 = 635. Carry over to A = 318 36" ~ 0 ~ ::: . ,. 0 CB = 1441 x .214 = 309. carry over to B = 155 CE = 1441 x .064 = 92. Carryover to E = 46 '~I~ 0 <;gOM :<;g 11 CD =1441 x .282 =406. carryover to D = 203 This process is continued tor the ~emainder or 600# the trJSS jo rrrts, ar tar all joints has been I. : balanced once, on ret~1r~ to joint A we rind -Llr13.S'~13.5"-W below trie lines an unbalanced moment; of (318665) =-347. The jotnt is balanced a second time 1~118 - .035 Tube

I~ ~

III~_~.J

K=L:01782= 00066 L 27 .

Fig. All. 25

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

All. 15

A~.247
~~ X 1/2 -27

~:ill X 1/2 233

~1

1890

tors will not be considered here but will be treated in a later article. (A) Calculation of Fixed End Moments for a Given Beam Load Fig.All.27 shows a fixed end beam with a variable moment of inertia and carrying a Single concentrated load of 100#. The beam ~oment diagram for this load is considered in three parts. F1g.All.28 shows the static moment curve assuming the beam simply supported at A and 8 and carrying the load of 100#. Flg.A1l.29 shows the other two parts, namely the triangular moment diagram due to the unknown moments MA and MS which produce fixity at the two ends. In this figure MA and MS have arbitrarily been taken as 100, instead of unity. Fig.A11.30 shows the MIl diagram for the 3 moment diagrams of Flg.All.29. The beam is divided into ten equal strips, and the MII curves are obtained by dividing the moment values at the end of each strip or portion by the corresponding I value from Fig. A11.27. From the cond1tions of fiXity at the beam ends J we know that the slope of the beam elastic curve 1s zero at each end. Likewise the deflection of one end of the beam away from a tangent at the other end is zero. Stating these facts in terms of the moment area principles, we obtain

~-'L -15
b.~ 1710 0

C -2025f
X

660 1/2
f

.322
2025

~-421 ~-93
~
X

1/2

210

1/2

-......:c.;:
b.

b3

~30

47

Fig. All. 26

15 -10 1743

Joint D is considered next: The unbalanced mom ent of (2025 - 712) = 1313 1s balanced by d1stribut1~g - 1313 to the connecting members. ~o member DB 1313 t .678 = - 892 and the remainder of -421 to DC. The carry-over moment to B is - 446. The carry-over trom C to D i& Dne-half the balancing moment S = - 421 but of oPPosite Sign or 2l0 J due to symmetry as ex?lained before tor member AB. Returning to joint B, the unbalanced as recorded below the single lines is (233 - 446) = - 213. ~o balance 160 is distributed to SD and 53 to SA. Carry-over 80 to D and bring over from A to S .5 x-53 =- 27. Continue this ~rocess until joints A and D have been balanced 4 times or 4 cycles have been completed. The tinal moments are shown below the double lines. ?ig.All.25a shows the resulting moment diagr~ on frame.

=-

~o 11~
L

- -(5) (Area of ll/I diagrams

equal zero)

171

r
~

:rwo~1710
Final

- (6) (Moment of the WI diagram as a load about o either end equals zero) (Note: Since E is usually considered constant it has been omitted trom denominator or the above equations. )

L~ -O I -

Moment ~ Diagram

J-

~17~a~1743
4050
Fig. All. 26a
All. 10 Continuous Structures with Members of Variable Moment of Inertia

S\

_. ......, a a , I I
" r ,,.

TABLE All. 3
Static all
C1U"'N

Trial -

In Arts.Al1.3, 4 and 5 consideration 'HaS given to the deri~~tion of expressions for fixed end ~oments, stiffness and carry-over ractors for beams of ~iform crass-section. Many cases occur in routine deSign where members have a var~able cross section. This article will illustrate ~he calculation of the fixed end ~oments, stiffness ~~d carry-over factor for a beam with var~able ~omen~ of inertia. ~he effect ofax:al load on these ~ac-

A'r::' ,. "',. ..... ,. .... . , 1-;' I ,.. ,... "" '" " , ... '" ". ". ... r. . '" ,... .. .. " ,. . I ,... ... i '" , " ., :ig , ".
a::i Ord. lett left . 1 etl'1.z e.d.

.. _.
t

<:Ur"N

Trial

~C1U"'N
r

h.
("

.~

h.

A~

left

I
I

1120 1760 615

....

2.67 U. r 2. 67 1 6.:n :l1ll0 U3 I 6.07 ,::~ ::.9 ,~~~~ U. 10 11l 14.10 111600 1130 13.i" 1810. 70.0 1O 18.07 26000 1110 17.96 li761 90.0 18.06 U30 22 06 ,38900 21 92 1970 110 0 :z:.:I cs 211.88 i41600 67.3 2ll.1l8 117<lO!~~~.1l 26.04 2lJ. 77 aeoec 40.7 29.76 1 121.5!117.0 211.94 33.1 "00 33.llJ i4180 17.11 33.8. ~~I

i -eza ..

,,

!1!!!!!

.. 21.1.5"

....

179Ctl1l
i

M.5.7

743 8 :2 1 111122

"

1812:1:

,,11922 ..14.1" 845.7

i . ~ 2....."

, ,
I

143.8

Actual area u; 4 u .... t o rnults For ca1c.l.at1_ ot U;lt .... "d. _ l i t e sZc.o. facto", 0Il.1,. ...lati_ Talus are

a.II._.

_''IT

All. IS
A

THE

MOMENT P } 10"*

DISTRIBUTION

METHOD

Fixed & 24"

f Fixed

16" 2,,0 in.

Fig. All. 27
1.0 ( 1)

.75
( 2)

1.2/'
55

.5

.5
(4)

.5
( 5)

.5
(6) , (7)

. 55

.75

I Curve
19)

(3)

18)

(101

identical to F1g. All.30e, ?or the conditions of su~por~ ~ss~~eG, the deflection 0: A away rr ce 3tangent to t~e elastic c~rve ~: 3 1s zero. Th~s by the de~1=c:1Q~ principle of area ~o~ents, t~e ~oment ot the Mil dlagr~s of Figs. All.30 (a and b) about end A eq~als zero. ThUS,
845.7X14.1+l'5O
74,3.8
~Bx24.4=0

aoo
Fig. All. 28 169 320
I

960"*

480
I

or
curve
240

Fig. All. 29 Trial MA : 100

:1B= -66"# Since MA was ass~~ed 100, then carryover :actor tram A to 3 =
-66

=-

.66.

100
To

r
00
(a)

curve

find carryover factor from 3 to A, take moments about B ~~d equate to zer8. Hence 845.7 M ....~,.... 743.8x 15.5.;.l'CO ~.Ax .:::V.~ = 0

(b)

whence MA = - 53.2"#
~heretore

A
Fig. All. 30
13

=-

53.2
100

carry over = - .532

:~ctor

3 to

(e)

,13.4 36

i.

I
37.86

ver.t~on

60 180

2.67 6.2

10:1'6 14j 18.06 24.4

?or t~e no~ent sign ccnused in this book c~~~y over factor Nould be ?lus.) (C) Calculation of Beam St~~f :;.ess Factors

(~ota:

'Jhen a ~eam :s freely supported at one end A and rixed at 743.8 the rar- end 8, the st i rrness raetor at the A end :s measured by I Table A.l1.3 shows the calculations tor computing I the ~oment necessary at 'A to ~roduce ~it rotathe areas and the centroids ot the three Mil tion of the elastic c~ve at A. diagrams of Fig. A11.3D. Substituting values In Art. Al1.4 it was nroved for be~s of tram Table in equations (5) and (6). uniform section tr2t EgA; MAL/4I or MA = ~EIgAI L. In a continuous structure at any joint all 845.7 M 743.8 8504 0 -(7) 100 A + 100 MS + ::: -members have the sa=e g, thus 4E9 is constant and the st~f1"ness K of any ~rismatic beam is 11922 18122 proportional to IlL. For beams of variable - -(8) 100 MA';' 100 i1S + 179809 ::: 0section the stiffness iactor K may Je 'Nritten:The value 100 in the denominators is due to the K=cIo/L - - - - - - - - - - - - - - - -(9) fact that trial values of 100 were assumed for ~here Ie is the ~o~ent of inertia at a particuMA and Ms. Solving equations (7) and (8) we lar referer.ce section and c 1s a constant to ~e obtain; found for each non-uniform aember. T~us for MA = - 315~# and ME = - 785"#. non-uniform ~embers (B) Calculation of carry Over ~actors EgA =MAL/4cIo - - - - - - - - - - - - (10) To dete~ine the carryover factor trom A to By the ~oment ~rea slope ~rl~c1ple, t~e B consider end A as freely supported and B as I slope at A when 3 is rIxec equals the ar-ea of fixed. A moment MA = 100 Is applied at A ~hich I the MII df agr-an oet.veen A and B. prod~ces the ~JI curve of Fig. All.30b. Due to Tak~ng XA = lOa, ~s Nas ~o~~d ~J equal the derlec~~on or the be~u ~~der this loading a -06"#. Thus EgA = 4(2<;'5.7- .6t3x'7~3.3) =l~::'d. restra~ning ~oment at a to cause tar~ent of ';,'he va Iue 4- is cue :0 t.ne etr ic i ..1d.t:. since alast1c c~rve to remain ~or:zontal w~ll exist. true area is Nan:ed ~nd on:y avera;e ordinate ';,':a5 unknown fixed moment at 3 has ar-ot trar i iv was used in Tab~e All.3. been taken as 100, uhus pr cduc i ng a :-;;'I cur-ve Equa t rng ::,15 r-esu.;'; ':0 gcua t t cn (:0)

1------

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All. 17

' 19 = loa 40 1'+ ac x x l ' whence c

.7 0 7

10 was taken as the value at the A end or 1. (See Fig. All.27). Therefore K A B = .707 IA/L Similarly ~or end 8:100 x 40
;:gB

= 4(743.8-.532x845.7) =

'!exl

whence c = .85 and KBA = .85 IA/L

All. 11 Frames with Unknown Joint Deflections Due to Sidesway

In the example problems so far treated the of the particalar structure were assumed to r8tate without tra~slation or with a definite amount of translatory ~ovement. Translation of the joints may however, be produced by shortening and lengthening of the members due to axial loads and by lateral sway due to lack of diagonal shear ~e~bers. The problem relative to the effect of jOint translation due to axial stresses is treated in a later article. In this article only the effect of sldesway of rectangular frames on the frame bending moments ';;111 be considered. :1g. All.31 illustrates cases of tr~es where only rotat1on of joints takes place (neglecting axial deformation) whereas Flg.All.32 ll1~strates conditions in which sidesway takes place and the joints suffer translation as well as rotation.
~oi~ts

bent car:ying a distributed side load on one leg as shown. The K or IlL values for each member are g~ven on the figure. Under the given loading it is obvious that the frame will sHay to the right or in other words, jOints 8 and C will undergo considerable horizontal movement. The moment distribution method assumes that only joint rotation takes place. To make this assumption true for thiS structure we will add an imaginary support at joint C which will prevent sidesway of the frame as illustrated in Fig.All.34. The end moments in the frame will then be found by the moment distribution process. Fig. All.34 shows the results of this process. To explain, the Solution begins with computing the fixed end ~oment on member AB !,.[L"'/12 (300x25 "')/12 15.63 thousands at foot Ibs. This value with the proper sign is written at the head of a column of figures on member AS as shown in Flg.All.34. Now considering Joint B, the unbalanced joint moment of 15.63 is distributed as follows:-

To Member SA To Member 8C

over to Joint A = - 3.27/2

=- 15.63

(40/190)

=- 3.27. =- 1.63
=-

Carry Car-

ry over to Joint C = - 12.36/2 = - 6.18

= - 15.63

(150/190)

12.36.

4 6' I
B I

10'
1

'r-

1= 10, 500 K=L'= 150

cl
0

~
~

g "
D

" -1-'
rnr
A

~~!J '"

T
25'

Fig. All. 33

Symmetry of Structure & Loading Fig. A11.31

Frame prevented from sidesway by external restraint

Symmetry of Str-uctur-e, Loading Unsymetrical Fig. A11 ....3.2

Symmetry of Loading but Structure Unsymmetrical

~ ~II~I'~I~I~ Ll'> <:"lOOIO'I_

... , ,,'

.... .c.c..::::t.o

- . ,I;[ i
II'"

(O.OO)f

(0.00)

-12.36 b 1 . -

+ 2.44

-6.18
4.88

_ _ bl.

'" " '"

.1.93b", -0.96 + 0.38 . . - b 2+O.76 O. 30 b~__ -0.1:>

_b,.,+0.12

There are several ~ethods of dete~~ning the bending mcments Gue to sidesway. Only one ~ethod will be presented here and it can Jest be explained by the solution of example problems. ~ample P:oblem 1. Fig . .4.12.33 shews " single bay r-ec tangu.lar-

.O.04b ... -11.75

_0.02 b..... 0.02 -1. 53

Fig. All. 34

All. 18

THE

MOMENT DISTRIBUTION METHOD

horizontal shear reaction at the upper end of the vertical legs of the frame equal to 3606 lb. Bending Moments Due ta Sidesway. ASsume the structure sways sideways as illustrated in Fig. A1I.36, but With no rotation of the ~pper jOints. It 'NaS proven in Art. AII.8 that the end moments produced by the lateral movement of one end at a beam whose ends are fixed against rotation are equal to 6Eld/La where d is the lateral ~ovement. In Fig. All.36 the load? causes both vertical members to suffer the same horizontal dis-idl-jd l .~---~-'placement at their : IB :C 00 upper endS, hence their end moments due to this : displacement are proFig. All. 36 ~i portional to EI/L a.

Fig. All. 35

Now consider JoInt C:The unbalanced moment Is -6.18. To balance we distribute to CB = 6.18 x 150/190 :;; 4.88. Carryover to B 2.44; to CD 6.18 x 1.30. Carryover to D = 0.65 40/190 The balancing and carryover procedure is now repeated for jOints Band C, until the unbal~ced moments become at negligible magnitude. l' ig.AILM shows that 4 cycles have been carried through. By keeping the ratio 150/190 set on the slide rule the unbalanced moments at Joints B and Care distributed'and chased back and torth as rapidly as one can WT'ite them down. Since joints A and D are assumed fixed they absorb moments but do not give out any. J

:/

Fig. AlI.35 shows a free body diagram at each portion at the rrame , The end moments are taken tram the results in Fig. AI1.34. Consider member AS as a tree body. To find the horizontal reaction HB we take moments about point A.
Z!lA

= 300X25 x 12.5 + 11750 -17570 He = 87930/25

25HB = 0

hence

=3517
=

lb.

Now consider member CD as a tree body. To rmc He take moments about D.

Since EI and L are the same for each vertical Sidesway Without member in our s tr-acture , Rotation of Joints the fixed end moments will be the same magnitude for each member. Therefore, tor convenience we will assume fixed end moments of 10,000 ft.1bs. are produced by the atdeswav, T..le now use the moment distribution process in permitting the upper jOints to rotate as illustrated In Fig. All.37. The procedure is similar to that in Fig. All.34. For example the solution is started by considering joint B. The unbalanced moment rs -10. This is balanced by distributing 10 x 150/190 = 7.69 to Be and the remainder at 2.11 to BA. The carry over moments are 7.89/2 = 3.95 to C and 2.11/2 = 1.06 to A. Due to symmetry of loading and structure, the distributing and carryover moments at joint C will be same as at jOint B, hence it is needless work to show calculations at these joints. The carryover moments tram C to B will be identical to those tram B to C. FIg. A11.37 shows the 5 cycles have been carried out to 0Etain the final end moments as shown below the double lines.
.. 7.89 b .3.95 - 1 / 2 - 3.12 b 1.56 - 1 / 2 +1.23b

lJ1D

= 1530 + 770 - 25Hc

0, hence

He

=89 1b.

We now place these horizontal forces on the top member Be as a tree body as shown in the upper portion or Fig. A11.35. The unknown imaginary reaction He at potrrt C that; was added in Fig. Al1.34 to prevent sidesway is also shown. To find He take ZFh = a

_" 0.62 -0.49 b - 0.25 ~b

t

8.47

Due to symmetry of loading and structure it is not necessary to consider right hand portion as calculations will be identical as shown for lett side.

1:Fh

=3517 + 89 - Be

= 0, hence Be = 3606 Lb

Since the external reaction of 3606 does not exist, we must eliminate it and rind the bending moments due to the sidesway or the frame. In other words, the frame will sway Sideways until bending ot the frame develops a

Il~
O~N~=t C:~~"":~OON
I .. , .. , I

I
Fig. All. 37

~",ooooci

o
o o

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

All. 19

Fig. All.38 shows a free body diagram of the frame vertical members with the end moments as found in ?ig. AIl.37. The shear reactions HB and He are each equal to (8470 + 9230)/25 = 708 lb. These shear reactions are then placed on the free body of member BC in Fig. All.38. For equilibrium the summation or the horizontal forces must equal zero. Thus P = 708 + 708 = 1416 lb. and acting to the right as shown in Fig. AII.38. Since the reaction Rc = -3606 in Fig. AII.35 must be liquidated and since the liquidating force P produced by the moments in Fig. AIl.37 is 1416 lb. it is obvious the values in Fig. AIl.37 must be multiplied by a factor equal to 3606/1416 = 2.545. Therefore, the final bending moment values equal those or Fig. All.34 plus 2.545 times ~hose in Fig. All.37. Fig. All.39 shows the results and Fig. All.40 the final moment diagram.

100

:--8" ......,.....8" ....,

T
12" 100 lb.

1:=16 1=10 15"

1=24

_J
Pin

12"

1SOLUTION:

A Pin

Fig. All. 41

Relative Stiffness Factors; KBA = 3I = 3 X 24 = 3 24 ' L

KBe

= R = 4x 16
L 16

=4

-c
708 +' ,..----~_=,.;tc- P' 1416 lb. ~B 708 8470 "FHB::lI708 _ 8470
HC 1ll 708

11"

-n

=3I L

= 3

x 10 = 2 15

T
25" .

The distribution factors for each member at Joints B and C, which equal K/ZK are recorded in the small 0 on Fig. AlL41a. Fixed End Moments: For Member AS = PL/8 = 100 x 24/8
For Member Be

1 -6
9230

Fig. All. 38
D

= PL/8 = 100 x 16/8 =200

= 300

Ins Lb,

9230 - 11.75 ~
1. 53 ~ + 20.03

,
IB

9.81

Fig. All. 39

~~,
~~ ~
~,

~ :lJ~
_-10

+1,

~r
~ ~~
~ ~

mil\'( ~

'7'~
"
II
~

..

d..;j .
" 1I II

9.81

_"='~"b7T

J
4 1.07

22.73

Fig. All. 40

1
.l>.o~E-

.... ~~ ~. N g:

~I e ...:...,
M

~~~ ...

As explained in example problem 1, the moment distribution 's carried out in two steps, one for jOint rotation only and the ~ther for effect of sidesway or horizontal translation of joints B and C. Fig. Al1.41a shows the moment distribution for no sidesway by placing an imaginary reaction Rc at joint C. The process is started at joint A and the order of jOint balancing 1s ABCBC. As soon as a joint 1s balanced the carry over moments are 1mmedlately carried over before proceeding to the next jOint. When a joint 15 balanced a horizontal line 15 drawn.
F.E.M. 200
200 F.E.M. Bl. 143 - - - 71.5 - 4 2 . 8 - - 85.7 a , B. 24.4 -u:2 - 4.1-- -8.1 B~

"
0

B,-W--U

<I~~
.IN ~ r;::1~

0 - 0 _

.~

~<r;::

~I< ;:

el ,-.

t:

::i
"

-363.2

B lI

<D

"''''''' 0 . " . OO~I~I='I~

o.a <'") _ _ co ....,
,_
<'")

..

47.3

'cD
M

"''''''' C01_1"'IIM
... I I

"

..
I " ,

'"1"'1 0 1.;

Example Problem. 2. Bent with Unequal Length Legs and Pinned Ends. Fig. All.41 shows a loaded unsymmetrical trame. The final bending noments at Band C will be determined.
A

0011 0

j jI

... I

Fig. All. 41a

,':);EifF"'c
,-'-..' -.-':).:::';>it..

All. 20

THE MOMENT
~oments ~own

DISTRIBUTION METHOD
B),119 - 5 9 . 5 54.4-108.8 B), s , -31. 0--15. 5 "5":2- 10.3 B 2 B" _ _ 1.5
~

Rr can be calculated by a consideration of the

free body of each member as shown in Fig.
Al1.41b.
61..,. 363. 2 _
B

With the end

the reaction

-----Rc'

61.85

'r
~

H
B"

3.15 C 65 HC=3.15 c-;

::i
"'l ~
t-

144.6
....
;:Q
I'l

---l......Q. 53

162.6

-" .

"T"
! D1_
3 . 15

;~<:Di~

<:0

~~I.''''I= oiIMI~"'; IB i

C:U::l

cl

10JI

35

A
Fig. All. -lIb

0 ~ ;1 0110 QA
;l;'"

'i
.

l'
, ,

F~

01

;ic5'
oJ

Fig. All. 41c

~
p=6. 02" 10.85

To find Ha take moments about A, 100Xl2+362.3-24Ha = 0, Ha "" 65 lb.

To find

.;

oJ
6.0L B 10.85 C I, _ =16.87 lb.

144.6

He take moments about

He

D,

..r:-:~ 02
~.

'0.85_")'62.6

-47.3 + 15He = 0,

= 3.15

Placing these shear reactions on member BC and writing equilibrium of horizontal forces we find that Rc = 65 - 3.15 = 61.85 lb. ~e now must liquidate this reaction RC and per,nit :rame to sldesway. We will assume that the frame 15 :ixed ended and that an unknown horizontal torce ? at C will deflect the trame Sideways. The fixed end moments for equal horizontal deflection of a and C will be proportional to EI/L a ot the vertical member.
I . 24 _ F ormemer bAB 'P-24: a F or memb er DC ,
'

-lA

ol

Fig. All. 41d

To liquidate the reaction Re = 61.85 it will therefore require 61.B5/16.37 : 3.67 tioes the moments in Fig. All.4lc. Therefore the !lnal bending moments equal those in ?~g. All.41a plus 3.67 times those in Fig. All.4lc. Fig. All.4le shows the results. T~e bendl~g moment and shear diagram and the ~xial loads now follow as a s iapje aat t ar of jsta t i cs .
-363.2

0417

"""'lS6:8

47.3 ..ll2..Jl.
643. J

I - 15 10 " r::a

0445

For convenience we will assume 417 in.lb. as the fixed end ~oments on AB and 445 on DC. The assumed fixed ends will now be eliminated by the moment distribution process as shown in Fig. Al1.41c. The order of joint balance
was ABDCBCBC.

ti gil; ",
Mlt:Il_

Ci: ~1:;

li~ ~I~

I
I
\1

DOl O

IA

00110

Fig. All. He

Fig. All.4ld shows the shear reaction on the vertical members at Band C. These forces reversed on the top ~ember show an unbalanced force of 6.02 + 10.85 16.87 lb. Therefore a force P = 16.87 was necessary to ~roduce the bending moments that resulted on the frame due to sidesway.

SXample P:oblem 3. Involving One Panel Member.

3enc~~g ~Oilients

~n ~~uss

~ithou~ 01a~Gnal

Stear

Frequently, in aircraft struct~r~s ~ :rU2S is used i~ which a diagonal member ~us~ os l~f~ out at one or ~are truss bays. T~e sy.:er~al shear load on such bays ~ust je carried by :he truss chord members in bending. Fig. All.42 shews a 3 jay truss witt no 11ag~na: ~e~ber in the center bay. The jending ~cments on tte truss members will be dete~1ned rJ~ the truss

______________________..,RIlIR.'.t!liiilll
ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES
All. 21

loading as
a

shown In the figure.

10

Then the distribution factors are: 10

~d

.,

10

'c

Gi

.:zJ

ld

To Member cb

=4/7

- .57
.14

-r
I

To Member

cr = 1/7 =

CD
3.5
g

CD
@

CD

CD
@

60"

To Member cd = 2/7 = .29

e -'-

1_ 60" --+- 60"

--+- 60" -1

Fig. All. 42

SOLUTICN: The relative moment of inertia values for each member are given in the circles on the truss in Fig. All.42. The distribution factor to each member
at each JOint IS then computed and equalS KIZK. The values are recorded in t ..8 0

In this problem there are no loads applied to the members between joints. The external shear load on the truss to the right of the center truss panel which equals 10 + 10 = 20 must be carried through the center panel by the truss chord ~embers in bending. This bending causes the center truss to sNay or deflect until a resisting shear :orce equal to 20 is developed. We will assume the truss center panel is fixed at joints b, c, f, and g. The right end of this assumed fixed ended truss will be given an upward deflection. This deflection will cause fixed end moments in members bc and gf which are proportional to I/L 8 for each ~ember. Since L is the same for each member, the !ixed end moments will be proportional to I of the member.

on each member in Fig. All.43. Th; stiffness factor K Is proportional to IlL frr the member. For example for joint c; 1 =~7 + 60 SO
~

100 F.E.M.

F.E.M.

100

-13.6 9.8
~4.

-27.2

-28.5 -27.2

~-57

--=T:"4
LO
4:

o:s
-0.3

4.9 -2.8

----0:5
~

4.0 -2.8

-13.6 8.0 1.4 0.9 36.9

-29 5.9

4.1

~~~-14.5 11.8
.r.s
.:.2.1.
-2.6
0.2 2.0

-0.9 Q.:.i

-19. 4

-24.9
b

45. 7

~~ -7.1
-3.4
g!1 -3.6 -6.7 -0.4 -0.9
-0.1

de
3,0 0.5 -0. 5

-0.8 -0.4
-0 1 -11. 8

"""0:0
~ 3.0

1.2 0.3

ha ~

o
I

<s!

1.5

'.3
" 16
I
I

-TI.":'T i

'~.

3 I !' 1

-0.3 Q.Jl 0.0

Q..2.

~ 2.0

-12.2 10.3 -2.6 1.1 ~

-24.4
5. 1

-U
0.6 -22.3

---l!...i
Fig. All. 43

All. 22

THE MOMENT DISTRIBUTION METHOD

Therefore the fixed end mcments on ~ember gt will be 75 percent of those on member bc. Assume 100 as the fixed end moments on be. Then a 75 x 10 : 75 1s the accompanying fixed end moment on gr.
w~ now remove the imaginary fixed supports on the center truss and let them rotate to equilibrium by the moment distribution process as shown in detail in Fig. All.43.

The first step is to record the assumed fixed end moments of 100 at each end ot ~ember be and 75 at each end at member gt with due regard for Sign. The moment distribution process will be started at jOint C. The unbalanced moment is 100 or -100 is necessary :or static balance. Using the distribution factors on joint c, we find -57 goes to Cb, -29 to cd and -14 to cf. Short lines are drawn below each at these numbers. Fifty percent of these balancing moments are carried over to the tar end of each of these members. This process is repeated at each jOint ~ntil the remaining balancing moments are negli~ible. In Fig. All.43 the order of jOint balance was ctdebghaCfdebghacfdebg. If the student will tollow this order he should be able to check the figures in Fig. All.43.
~ig. All.44 shows free bodies of the truss members bc and gf with end moments from Fig. AIl.43. The shear reaction at ends c and r by statics equals

to axial loads ~rcduces :r~~sl~tiQn of the joints of the structure. Since the ar~les between the members at a joint re~aln the same due to continuity, this translation of joinGs bends the members between joints. (2) ~~e bending ~omen:s In ~he ~e~o~rs d~e to exte~nal joint or lateral loads or those ~ue to ~oin~ rotation produce lateral deflection of the members Jetween joints. The member axial loads times these lateral deflections produce seconda~J ~oments. These secondary ~ament3 can be r~ndled by the general method of mcment distribution however the stiffness and :arry over factors and the fixed end ~o~ents are ~ct constant but become nL~ctions of the axial loads. Fig. All.45 shows a prismatic beam Sl~ply supported at A and fixed at B, Nith a moment M A applied at A and carrying an axial compressive load? Sub. Figure a, band c show the 3 parts which make up the ~cment jiagr~ on the beam. Without the axial load P the pertien (c) would be omitted.

28.7 Hf~f
60"
-L

He + Ht

:II

36.9 + 45.1

45:7

.3

34
MOMENT DIAGRAMS

60
+ 28.7+34:=

Fig. A1L44

60

Fig. A'll. 45

1.377

1.045 = 2.422 lb.

The external truss shear at line cf 20 lb. Therefore it will take 20/2.422 = 8.25 times the final bending moments as found in Fig. All.43 to develop a bending shear reaction of 20 lb. Thus the tinal moments are 8.25 times those in Fig. All.43.

By definition the stiffness factor of beam AB is measured by tne necessary moment MA to pr-oduce unit rotation (9=1) tor elastic Lme at point A. By the moment area pr~nciples, the slope at A equals the shear at A d~e :0 the ~oment diagrams acting as a load en a slnply supported beam at A and B. ror a given unit slope "at A tt is obvious that the r-eq. i r-ec MA The solution as given neglects the effect is less when the load P is acting as it proat axial loads upon the value of the stiffness duces a 20ffient load i~ (Fig. c) of the same sign and carryover factors. Art. All.l2 explains as the-~oment load due to MA' Thus the stitthow to include these effects. ness at AB is less when carrying an axial compreSSive load compared to that wltho~t it. For All.12 Effect of Axial Load on Moment Distribution a tension axial load diagram C becomes a ~ega tive moment diagram and thus MA must become In the preViOUS articles the effect at larger to give ~it slope at A. Or in other axial loads upon the member end moments of a words a ~ember carrying a tensile axial load continuous structure was neglected. The axial has a greater stiff~ess tr~n one without axial loads produce bending in the members of a continuous structure in two ways. (1) The short- I loads. It i s evident that the moment diagram. ening or lengthening of the various members due I due to P will also effect the ~gnitude of any

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All. 23

rixed end tor.

All.l3

~oments

and also the carry over fac,

, ,
,

Fixed End Yoments. Stiffness and Carry over Factors for Beam Columns of Constant Section

r,

In deriving expressions for fixed end moments, stl::ness and carry over iactors the beam coltwh~ to~ulas ot chapter AIO, must be used in finding slopes, deflection and moments. Mr. B. W. J~es in an excellent thesis at Stanford University and later published by N.A. C.A. as Technical Note #534 has derived expressions tor these factors and has provided graphs of these factors for use in routine design. Figures All.~6 to All.56 inclusive are taken :rom this thesis. The use of these Figures can best be illustrated by the solution of several pro b.Lems
All.l4 Illustrative Problems

! !
i

i i

i
,

,I
!

i
,

, , ,

,
i , !

i
,

! !

, I ,., i I

, e_ :rI
A
1

!
-

I I
I

I
,

I
!

!
!
,

.. ..

Example Problem #12 F~g. AII.57 shows the same continuous beam as us ed in Example Problem #4 (and solution #2). For this example it has been assumed tr~t such axial compression loads exist in spans Be and CD as ~o make the term L/j =2.5 and 2.0 for these spans respectively, where j = v~I/P. Due to the axial loads, new values at stittness and carry over moments as well as fixed end moments ~ust be determined as follows. Span BC
K

/ I I /1 i I I .:.---! i I j I i r- ---l I , , >t__ , :......+-..' , i I ! I i~ ., , , , I I I I i i I I I ! ! ,

, ,
,

, i , I
,

Ii I , , I/ I , V
,
,

I
1

i ! , , , , , : I i
1

, ,pb,

,
I

i /,
,

I'"
,

A<" I ~~

,
i

I
i
!
j

I
,

I , ,
i
, ,

i ,

, ,

carry-over factor.
Fig. All. 46

~l

I i I I

,
r,

= t = 9~ = .0104

, , I I
I
,

, ,
,

'---'---'-1 ' , I
i
,

From Fig. All.47 when L/j = 2.5, correction factor for stiffness factor = .775 when tar end is fixed and .36 when far end is pinned. Hence,

, , r.
r.

I
, , ,

i Vi
/I ~/ '/
,

I I I I
, ,

I
r_I..o.'.
,
,

I
, ,
1

tu' _
, ,

'tl_.,

, , ,/

KBC = .0104 X .775 = .00B06 KcB = .0104 x .36 = .00374

Frem Fig. All.46, when L/J = 2.5 the carry over taetors are Cac = .73 and CCE = a it B is considered freely su~ported. From Fig. All.48 when L/j = 2.5 fixed end moments ~or unito~ load = wL~/lO.67 = 1 x 96al 10.67 = 865 in. lb. Span CD

,
\

i, ,I

i
,

r, 0

,
i
,

, ! ~ Vt_lo:... ! , , V, VI , I ...--1 Vi I
i
I

i
,

I I

)/

I
fU _

,
1

, , , I
!

III"':"

i
!
1

!
j
,
1

, I I
I
,
,

, ,

",-<:: 'A''''

NY
"C
!

, ,
1

..

\'
,

= IlL = 1/72 = .0139, L/J = 2.0 ?rom Fig. All.47 correction factor = .86, hence KeD = KDC = .86 x .0139 = .01198. carryover
K

'i
:

\i-e_....

l_. fU . . . IU"

factor from Fig. All.46 = .62. Fixed end moment~ from F'ig. All.48 = WL"'/l1.2 = 1 x 72"'/11.2 = 462 I The dlst~ibutlon factor at joint C equals (.00374/.00374 + .01198) = .238 to CB and the ~eoainder .762 to CD. The balancing and carT:! over process is I s imi Laz- to tnat in zxamrne Problem #4.
II

tlJ'Stiffness factor coefficient. K = C I

Fig. All. 47

!
I 'r ~',
I ",

All. 24

THE

MOMENT DISTRIBUTION METHOD

",
w

, !",-I
!
I

_,1"II./I.a.

,
, ,
I

, ,

, ,

'Io~
1

,fJ
iA
i
, , , ,
, ,

!
,
... ..

,t
!
I
1

V I,
,

, ,

i
i

"

I , I
,

y-t- r "
, , ,
, ,

r : rtr
1Ira

,;I

R
'\!
!

, ,

"
u

Y
I
I

!
!

1
1

,
,

I
, ,

A V ,
1

I
1

,
!

! !
,

,
,
,

I ,

I u

'"
,

,
,

, ,

,
,

!
,

~I!

r1

! ! :
,

'\ 1\

\i \

, ,

,,' f . . . ..1...

!
I

:
I ,

'\J

!
, , , ,

I ! ,

i\ i : '\ I
,
I
,

!
I

.u

I
, ,

!\

,
1

1 ! ;-," I -'I-'-1--'HH-i---'+--+--+_--+--' , \, I "

"

"

1 I
!

'\

\i
\ I

Fixed-end moment coefficient. uni rora Load.

Fig. All. 48

",

Fixed-end ~oment cceffic1ent. UnifOrmly varying load. Axial .compression.

Fig. All. 49

'"

~ ! ' ,
,

..
,

~ .... _~ i

,
,

"

'. (" I
: "
,

~Ii

~/2

\") ..,
T_ _
:
,
, ,
! "

, :
1

,
,

-"to ~

..
'. '. I ,
~

"

I ,
I

I
i

! !
1

i
,

!
;~
,

I
,

i I ,

I I
I
I

I,

/1
,

"
,
,
1

,I 'Y
i

/
,

,/ .
J /1
!
, ,

',
,

, ,

V V,
,

I
,

!
I
1

!
!
, , ,

1
I
1

I
,

i I

!
I
I

! 0. 1/1

v,
I

/
1

,
1

!
i

i I : Y,
,

>,,
I
~;
i
,

, ,
I

i :~I
, , ,

~ll I
I
!
, ,

! !
!
I
1

,
,

, ,

:
,
1

!
,

j--{' !

!
I

i
I
,

i
~

, ,
, i
i
I
,

I
!
,.llJ

~.~--

i
,

"

!
I

Y
'

!
,

I
,

..J...--+1

I i

I ..:-1

....--1
I

, ,

I
,

\, , \
,

I
1

I
,

I
, ,

I
!

i
,

'{

,!

Fixed-end Uniformly vary'ing load. Axial tension.

Fig, All. 50

! I , I I , '" coefficient. moment

, ,

!\
,

, , ,

! \ , 'I' coefficient. ?lxed-end ~oment

, ,

Concentrated load at
Fig. All. Sl

~id-span.

. --.-.-.-------

ze= IIIIIII!!l!!!!~

ANALYSIS AND DESIGN OF

2.0

FLIGHT VEHICLE STRUCTURES

All. 25

f-I.l.l~
r.e I-

\~

1"0 ~~
D..L)
'b 2 "
Col

-, I 1-+-l-+7'f'=--f-+--l--t---+-t---+-t---j-+--1
v
I

I-

!.fA, .. WI;

4.0

I
J.5

i
.,
2.0~

1/ j .. 1 ~ a compreu1on--..
LI.i '" 1. 0 tendon ---'.

- 2.0
. JiO
4.0
!

!
.1
Fig. All. 52

6.0 ....... I .J

t- I

..

'.0
.5

.,

alt or bit

.9

Flxed-end moment coefficIent. concentrated load.

I I
I

I
I

l.~(~~ I

1:.-I
!

! '

'I

'N.

. I'
,
,

, .

1/'f'_ _

'.
!

.
,
I

, . . . -. . . . i I iI Ii!
~

(/'1
":

'''01_' ,
": ":

1/1 iii
I
,

1/1// i !If/"
Y YI!
I

i
,

I
,

I'

I :

- - -"' ~ ~.
,~

. I
I
,

I , ,

, .1 ,
,.,:

I
I

-e

L~

'

_..
...

I \
, ,

; ... '. ""

\l-~i ,
\

, '. ,
:

..... 1

Ii

i
i

\
, ,

';

Iii Ii , I : I ii I, :I II i II II II , , II I, I I , ,I , , , ! I i Ii i ,i I ',.1 'II., , I I , , , \ v \i '\ 'I~ I I

, ,

Ii

;1

II !II

i ,

~~Irl-'--; ",,,1_,,';'-'f; p,.-,,,,--:.C;,,Tr ..

I' /'

=, ." " , : : , , "

l.2L._

.1,
:

r -ri i

'!
ij
!
I

I
,

I ill ,
;1
,

iI

i~L-.~

."

.,

"

/ II

Fig. All. 53

"'1

, II

All 26

THE

MOMENT

DISTRIBUTION METHOD

1.a

ri-f
I
I

~
,

i
'

I:
,

1
, ,

1
I

i
I

..
a,
r,

I , I
I

,
!

I I
,
i
I

I ,
I !

III, I
!

I
i

<;.:......eO"

1 1
I
I

!
I
i

i ,

I
,

I w.3~1 w:32 ~16 i::l::O 1 ttl tIl t 1 CIT' 1 r~~ ~ 7~ 1..@j.@;2JJll.. .:L..iQ...15.J.Q:6Jg:2 ~ T FB
2

IXrnr&d~IFB
j ~~.
B' '~....
.. ,

B'

lO~"

A'

84"

lOOt~60"

1
,

C'

A' '....cabana /A
',strut~

i I
! i
1

I I
, ,

,.q_lIl_

, "

,
I

II
I

I
1

r.

i
./

,%
Y I ./'! I

i
I

I I
I
I
I

' I

1=17.21

IT' , : . , . , I
..
. . ,__

Fig. AI1.58 Spruce

............

B ,.. '
"

,,,'B

uael. /:.~~\.

" c:,....

v....

T 6"

6"

Boam Loads
B

[;U75BJ -~ ~
A

1=22

A'

,.

i : ]
I
i
I
I

SpanAB, A'B'

Span4AA'

i
I

Fig. All. 59

r-c,

NI
,

I I

r-,
I

'I
I I
1

I'
I:
~

1,,1

I
- a.

!
o

I I::>

'~/J~

Culum..'"l dtstrlbutlon factor.

Fig. All. 56
"&1*11n,\ L/j.2.5

2~

Fig. All.59 shows the total beam axial loads in the various portions of the beam by adding the values shown in Fig. All.S8. The outer span AB due to the axial loads 1nduced by drag wires ot drag truss 1s subjected to a 'mrJing load. As is customary in design of such wing beams the ax1a1 load in span AB will be taken ~s equal to the average load or (-9571-8565-7559)/3=-8565#. The be~ bending ~omer.ts at the support pOints will be determined by the ~oment distribution method. Calculation ot Factors;Span AB

lis Cllnata..llt, L/J=2.0

I = 17.21,
B

L = 100,

c
"00374 .01198 aa 76 .865
146 0

D E

!'1eAll.57
StiU. Fact. II:
D b<ttiOll:le

E ' 1300,000,
j

IlL p, 8565
8565

= .172

~51.5"-!---"-----+

__ 72"_
.01198 _

c.O. F1I.eter F1ud End. Mom.

c.o. rc c

e Joint B

e o .883 c

,~,1jl300,000 x 17.21 , 51.1

P

.00808 .73 lI6S 18

c .52

."
a

-4<'

52

L/j , 100/51.1 ' 1.96

Ba.1. Jo1Dt C
Car -ceerBa1aJ:lee D CAl' -cverFInU Mom.

13.2

From Fig. All.47 when L/j = 1.96, correction factor = .86 for fixed far.end and .52 tor pinn ed tar end. Theretore
KAB ' .52 X .172 ' .0895

_88

_7

KBA ' .86 X .172 '

.148

Comparing these results with those of Problem #3 From Fig. A11.46, when L/j = 1.96, the carryover we find moment at C is increased 15.7 percent fa.ctors are:and that at D is decreased 8.5 percent. For C.O' BA = .62, and zero from A to B since 5 larger values ot Llj the difference would be is considered in its true state or freely supgreater. ported. From Fig. All.48 when L/j = 1.96 equaExample Problem #13 tion tor fixed end moment for uniform load = WL 2 / 11 . 25 = 32 x 100:1/11.25 = 28420"#. Fig. All.58 shows the upper wing ot a b1The overhand moment Mac = 60 x 26.6 x plane. The wing beams are continuous over 3 (32+16)/2' 38300"#. spans. The distributed air loads on the tront Center Span A.A' :beam are shown in the figure, also the axial loads on tront beam induced by the lift and drag I ' 22, IlL ' 22/84 ' .262, P , - 11758 truss. The bay sections or ~he spruce are shown ~,300,OOO x 22 = 49.4 J" ='V" 11758 in Fig. All.58. The moment of inertia in each span will be assumed constant, neglec~lng the L/j , 84/49.4 ' 1.70 influence or tapered filler blocks at str~t points. .262 x .895 .234 KAA I =KA'A C.O. Factor .585

&

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All 27

Fixed end moment

= wL a/11.35 = 31
19300"#

x 4a/l1.35

For member Be
p

The moment distribution process is given j in Fig. All.60. It the axial loads were neglected, the bending ~oment, at support A and A' would be 19480, thus the axial influence inFrom Fig. All.47, stiffness factor creases the ~oment at A approximately 7.5 per= .00681 cent. From Fig. Al1.46 C.O. Factor = .66
FlJ.

= - 27150, IlL = .0083, I = .332 = e9,00~7~~~X .330//2 =18.82, L/j =184.82 = 2.12

= .32

x .0083

an.eo

...~ Ie
C' B'

.... n

.~31

._16

For Member AC
p

IA'
1

= 10400#,
2.42

IlL

= .0029,

= .137
L 47.20

c .148 uu

.234

.caas
.52 0
2 420 _2B420 383 ~880 0

.137 j ;;V29,OOO,000x 10400

= 19.58, 1 =19.58

a a,

Fig. All.53 shows the moment distribution solution which includes the effect of axial loads on joint rotation. Comparing the results in Figs. o no _l85 Al1.62 and 63, the moment MeA of 24050 is 29 per 38300 38300 20909 2090S1 .38300 38300 cent larger than that in Fig. 52, and the moment at B is 18 percent larger. The effect on the ax Exa~ple P:cble~ #14 ial loads or these new final moments will be ?lg. All.51 shews a triangular t~lSS comquite small, and thus further revision is unposed of two ~embers fixed at A and B and rigid- necessary. ly joined at C to the axle bar. Let it be required to determine the end moments on the two ~embers considering the eftect or axial loads on joint rotation and translation.

,'0

,.e125 :lle1 .828 1265 _918 -341

St1tt'ness factor = .0029 x l.18 = .00342 C .0. factor = .39

saa

.3901148

Et'fect on End Moments Due to Translation of Joint C Due to Axial Loads. The ~ovement of joint C normal to each member will be calculated by virtual work. (Reference Chapter A7). Fig. AII.54 shows the solution:virtual loading of 1# normal to each ~ember at C The ~gnitude of the axial loads in the The Table shows the calculation of the normal de members is influenced by the unknown restraining tlections at C. ~oments at A and B. To obtain a close approximation or the axial loads, the end moments in Table the ~NO members will be deter.nined Without con_ I 3~bL. I I sideration of axial loads. ~hus the external a 1 L/A.! ", I !,z.L -n- I All: i joint ~ament of 4 x 18000 = 72000 In. lb. at C , ,C .00000"9 1.89 ! ,0962 I 1.5 .0816 is distributed be~Neen the two ~embers as shown ''''00 , I _1.60 i' .202 1060 -1.89 .1250 I sc I, .000002"" I in Fig. All.52. With the ~ember end ~oments .207 known the axial loads and shear reactions at A and B can be found by statics. ~he resulting Thus the deflection of joint C normal to axial loads are I BC equals .202" in the d ir-ect i on assumed tor the unit load, and the deflection of C normal to PAC = 10400# and PSC = - 27150# and the AC = .20'7" shear reactions, SA = 590# and SB = 2000#.

1-.

""

_~'71~

~ith the axial loads known :he beam factors can be deternined.

~odifteG

I"", " "J ~::..'

All. 28

THE

MOMENT DISTRIBUTION METHOD

All. 15 Secondary Bending .Moments in Trusses with Rigid Joints.
A

Fig. All. 64

c
The fixed end moments due to support deflection equals M = 6Eld/L a as derived in Art. All.? However, the secondarY moments due to axial-loads times lateral deflection modity thIs equation. ~James~ has shown that the modified equation is,
d i t K'L! and R' L !'lt1xed end = L" 6Eld (2~-a)'

Often in alrpla.~e structural design tr~sses with rigid joints are used. Rigid joi~ts ~re produced because of welding the ~embers together at the truss joints or by the use of gusset plates bolted or riveted to each member at a truss jOint. when a truss bends l1nder loading the truss joints undergo 11ffer9nt amounts of movement or deflection. SinCE, the truss member-s are held rigidly at the jctnt, any joint displacement will tend to bend :~e truss ~ember8. The bending moments prOduced in the truss members due to the truss joint (e!lectlons are generally referred to as secondary moments and the stresses produced by these moments as secondarJ stresses. Since fatigue strength 15 becoming ~ore important In aircraft structural design, the question or secondary stresses beca.nes of more Unportance than in the part. The moment distribution method prOVides a Simple and rapid methoc for determining these secondary moments in truss members due to truss deflection. T~e general procedure would be as follows: -

' than

!'l ' 6Kffi 2{l-a ,

Fig. All.56 shows a plot ot For Member scrR' .202/40' .00505

2~--<t ' 1. 08

2~-a

against L/j.

K' IlL' .0083,

ror LI J ' 2.12. Hence, tixed end moments due to translation or JOint C equals!'lBC , !'loB ' 6x .0083x .~~X29,OOO,OOO,-6750"#
For Member AC
R' .207/47,2' ,00438

(1) Find the horiZontal and vertical displacements of each truss joint due to the criti~al des~gn condition. (See Chapter A7 for methods of !inding truss deflections).
(2) Fram these displacements, the transverse

when hence

K= ,0029, 2~--<t' .92

deflection of one end of a member With respect to the other can be found fer each truss member.
(3) Compute the fixed end moments on each member

LlJ = 2.42 and member is in tension.

!'lAC '1'lo.l,6X .00438 x .0029x29,000,OOO 92

due to these transverse displacements .

(4) Calculate stiftness and carry-over !actors

-2400"#
The signs are minus because inspection shows that the moments tend to rotate ends at members counter clockwise. (Ret. Art. All.2) Fig. All.65 shows the moment distribution tor these moments. The magnitude ot the moment at B is 6.7 percent ot that in Fig. AlI.53 and 10.4 percent at Joint C, however, it is relieving in this example.

for each truss member.

(5) Calculate distribution factors ~or each

truss joint. carry out the ~oment distribution process to find the secondary moments at the enes at each member.
(7) Calculate the S~resses due to these second-

ary moments and cambine with the pr1~ry axial stresses in the tr~s members ~ue to truss action with pinned truss joi~ts.
All. 16 Structures with Curved Members.

Fig. All. 65

T:18 moment d.i s tr-i but i on me thod can 88 applied to continuous str~ctures which :'..~cl~de cur-ved as -,','ell as scr-a rmt nenoer-s . :':::e equations :or fl~jlng the st~:~ness iac~8rs, carryover factors and fixe~ ~nd ~e~2nts ior stTa~~ht ~2~bers canr.ot je ~sed ior 2UTVed ne~ jers. ~he ~lastic center ~ethod as ~resented

we em

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

All. 29

in Chapter A9 provides a rapid and Simple method for deternining these values for curved members. The use or the elastic center method in deter.nining the value of stiffness and carryover factors will now be explained.
All. 17 Structures with Curved Members.

diagram. Thus to find the distance fram E to this centroid we take moments of the MlEI diagram about B and diVide by ~ the total area at the MlEI utagran, Thus
X : -2( .667L) + l( .333L)

- 2

:..:...1..=
- 1

Before considering a curved member a member or constant EI will be conSidered. Fig. All.66 shows a beam treely
strai&~t

Thus the centroid of the total ~oment weight ~ which equals one lies at paint A or a distance L to left of B.

Fig. All. 66

#
~

n MA 5 ..

-~

_ ,I- .d,= 1 Had.

~ A

EI COnstant '

.-. -,~ deflected beam M

L
,~ I
B

M diagram

Fig. All. 68

M/EI Diagram

-1:"

In USing the elastic center method to determine the stiftness and carry-aver factor tor a straight beam, we aSSume that the bending moment curve due to a moment applied at A is of such magnitude as to turn the end A through an angle of 1 radian. As shown above, the moment weight ~ for this loading is unity or 1 and its centroid location is at A. Then by the elastic center method we tind the moment required to turn end A back to zero rotation. The value of this mOment at A will then equal the stiffness factor of the beam AB. In order to simplify the equations for the redundant forces the elastic center method refers them to the elastic center. From Chapter A9 the equations for the redundant forces at the elastic center for a structure symmetrical about one axis are; 11 0 ::; -1: 0s

Z ds/EI

- - - - - - - - - -

(1)

supported at end A and fixed at end B. A MA 1s applied at end A of such ~agni tude as to turn end A through a unit ar~le ot one radian as illustrated tn F~g. AlI.66. By Qefinition, the necessary moment MA to cause thIs unit rotation at A is referred to as the stiffness of the beam AB. In Art. AIl.4 it was shown that thi s required moment was equal in magnitude to -4EI/L. It was also proved that this ~oment at A produced a noment at the fixed end a of 2EI/L or a moment of one hal~ the magnitUde and of opposite sign to t~at at A. Fig. All.6? shows the bending ~cment d~agrs~ Nhich causes ene radian rota:~on of end A. ?ig. All.58 shows the M/EI c tagr'am, wn..Len equals 'the moment diagram jivi~ed by SI ~hich has been assw~ed constant.
~oment

- - - - - - - - - - - - - (2)

- - - - - - - - - - - - -

(3)

Fig. AIl.59 shows beam of Fig. All.56 reby a beam with the reaction at end A replaced by a rigid bracket terminating at point (0) the elastic center of the beam, which due to symmetry of the beam lies at the midpoint of the beam. The elastic mcment loading is ~s ::; 1 and its location is at A as shown in
~laced

Fig. All-59.
~

<,'

1-0---- -

L
2

------l
'Mo

The total moment we Ight as exp Ia Ined in Chapter A9 equals the ~rea of the MIS I diagram. Thus 0 for the MIS I dlagr~~ in f~g. All.58 equa:s,

! Y L---:
o
1

A~
Fig. All. 69

Or
B

Solving fer redundants at (0) by equations (1),

(2) and (3).

In other words the total elastic moment ~elght equalS one or unity.

'L
"'"0

Z dS/;r

09

LiEr -

-1

- _ SIlL -

o will

The location of this total moment weight coincide with the centroid of the M/SI
i ", , , ,
v'./

All. 30

THE

MOMENT DISTRIBUTION METHOD

Xo

Z ~S Y = 0, because
Ix

y the vertical

Is removed and end A is cor~ected by a rilid bracket te~inating at the elastiC center of the structare. ~e will new find the required forces at (0) to cancel the unit rotation ~t A which 'NaS assumed in Fig. All.71.
,Y ,

distance of ~s load to x axis through elastic center (0) is zero. Fig. All.70 shows these forces acting at the elastic center.

,0'= 1

XD1~\Xt
-: Yo . ' J \ 'B ..J...!: I
A

yo -d. O.6366r
-L-

5 "....

~rl6E~I/-L-2---{' "
Fig. AII.70

EI/L

'y
Fig. All. 73

The bending moment at end A equals

By definition the stiffness factor is the moment at A which is yeQUlred to turn end A through an angle of I raddan , Thus 42I/L Is the stiffness factor and thIs result checks the value as previously derived In Art. Al1.4.

The total area of the M/sr curve ~or the curve in Fig. All.72 it calc~lated ~rould equal one or unity as explained in detail for a straight ~ember. The centroid of this M/EI diagram would if calculated fall at 901nt A. Thus in Fig. All.73 we 2.?ply a unit ~s load at A and find the redundant force at O. Due to symmetry of structure about a vertical or y axis the elastic center lies on ~h~5 3ymnet~ical axis. The vertical distance ITom :~se ~:ne AB to elastic center equals Y= .6366~. (See page A3.4 ot Chapter A3). The elastic moments ot inert~a Ix and 1y can be calculated or taken ~r0m reference sources such as the table on ~age A3.~. :"thence, Ix = .2978r"t J but t Hence Ix
.2978r" SI

The bending moment at B in Fig. All.70

equals,
- -

_ M

5EI (.b.) _ 2EI L + La 2 - L

= SII

Hence the carry-over factor from A to B Is .5 and the carry-over moment is ot opposite Sign to that of the moment at A.
All.18 Stiffness and Carry-Over Factors For Curved Members.

Fig. All.71 shows a curved member, namely, a halt circular arc of constant EI crosssection. The end B Is fixed and the end A is freely supported. A moment MA Is applied at A of such magnitude as to cause a rotation at A at 1 radian as illUBtrated in the Fig. Fig. All.72 shows the general shape of the bending moment curve which IS statically indeterminate. In FIg. All.73 the support at A

Solving the equations for the redundanta at (O)J remembering ~s = 1 and located at pOint A.

Me - _ -

= .=.....ill.:::
nr

- E: I
nr

ZI
V

Zsx _ - (l)(-r) _ 2EI nr.3 - nr" Iy -

ZSI

Xo

= Zsy =
Ix

(l)(-.6366r)

O.2g7e~ 3 SI

= at A Jearn near

Fig. All. 71

Fig. All. 72

Fig. All.74 shows these forces acting the elastic center. T~e Jendi~g ~oment at equals the stiffness factor ~or the curved :1xed at rar end 3 and :reely supported at

ANALYSIS AND DESIGN

OF FLIGHT VEHICLE STRUCTURES

All. 31

end A. The ratio ofpoint B to that at point A 'Nil:' cive the carryJver fac:or.
Bend.tng

t~e

:ending

~o~er.t

at

The method of hew to handle this re~aining redundant force can best be explained by presenting some example problem solutions.
All. 19 Example Problems. Continuous Structures Involving Curved Members.
Ex~~ple

:!loment at A,
MA

Problem 1
B

Fig. All. 74

-SI;'I = nr - 2.14 (0.6366r) r

Send~ng

;'3

(281)r nr "

= -0.318EI

1.:;81 _ 0.536 SrI:

2.314- :::1

moment at 3, 1.36EI + O.636EI

r r

:13 ::: _-,,0-,-.3,,',,""8",E",I,

T~er9fore

1" 042 E1 r

circ~lar

the stiffness fac:or for a halfarch of constant EI is 2.314 EliI'.

Mg to ~A or (-1.042

The carry-over factor equals the ratio of El/r)/(-2.314- EliI') = C.~2. It sr.ould be noticed that the carryover mo~ent has the same sign as the applied ~cment at A as compared to the opposite Sign ~or straight members. In other wor1s, there are t~o points of inflection in the elastic curve for the curved arch as compared to one for the straight member.

The fixed end moments on a curved member for any external loading can be dete~ined q~ite rapidly jy the elastic center ~ethod as 11:~strateG ir. Cha~ter A9 and thus the ex?lanaticn will not be repeated here. The student should realize or understand that ~hen the end moments on a straight member in a ~cntinuous structure are found tr~m the ~c~ent cistrijution ~rocess. the remaining end f~r:es are statically dete~inate, whereas for a c~~~ed ~emjer in a continuous structure, Lr.cwing the end ~c~ents does not ~ke the curvet ~e~ber statically ~eter.n~r.ate, since ',':8 have six unknowns at the two supports as il:~strated in Fig. All.75 and only 3 equations ~~ static equilibrl~. ~ven when
tne end mcment s

Fig. All.76 shows I =1 a frame consisting of both straight and curved members. Although simplified L::r60" relative to Shape, I 2 this frame is somewhat representative of a L::80" L=80" fuselage frame with twa 1=1 I =1 cross members, one between A and C to support installations above cabin ceiling and the other between F and 0 w: SOl/in. to support the cabin I I floor loads. The frame L=60" supporting forces are 1" 6 assumed provided by the fuselage skin as I :: 2 shown by the a~ows E on the side members. Eccentricity of these skin supporting forces Fig. All. 76 relative to neutral axis of frame member is neglected in this stmpl!tied example problem, since the main ~urpose of this example problem is to illustrate the application of the moment distribution ~ethod to solVing continuous structures involving curved members.
SOLUTION

Due to symmetry of structure and loading, no translation of the frame jOints takes place due to frame sidesway. The ~rame cross members AC and FD prevent horizontal movement of joints A, C, F, and 0 due to bending of the two arches. .~y horizontal movement of these joints due to axial deformation is usually of minor importance relative to caUSing bending of frame members. Therefore it can be assumed that the frame joints suffer rotation only and therefore the ~oment distribution method 1s directly appllC~ble. Calculation of stiffness (K) values for each member of ~~a~e: Upper curved memJer ABC: mh' K ,il,Be::: K CEA::: 2.314 E_I. ,_~, ..e va 1ue was

are de t erninec t: om the ncment pr-oc ess there still r-enatna

one unknown,

Fig. All. 7S

the horizcntaa reaction at one of the

n~ely

beam encs ,

der:ved in the

~revlous

Art. AII.IS.

All. 32

THE MOMENT

DISTRIBUTION METHOD
dis~ributa1 nome~t.

SUbstituting: ..."d.... X - = .0770 (E is con30 s~ant and therefore omitted since only ~elatlve values are needed for the K vaLues K,
ABC

,;: K,

:BA

;:

"2~'

'The curved :r.enbe:-s :::r' the .:'ra:J.e have ::c applied external lcadin~s ~ence the .:'~xed end caments on t~e 2urved :r.emjers is zero.
For ~embe~ AC, :lxeC end ~cment equals WL 2/12 = 10 x 60 2/12 = 3000 in. lb. and tor member F'D = 50 x 60 11/12 ;: 1.5000 in .lb Since the supporting skl~ .:'arces on the side members have jean ass~~ed ac~1::g along centerline of .:'~2.Ee membe:-s, the ~1xed end moments on ~e~bers AF and CD are zero. Moment DistriJution Process: Fig. All.7? shoWS the calc~lations in carrying out the successive cycles of the moment distribution process. Due to symmetry

~he relative moment o! cross-section ot eac~ fr~e on Fig. All. 76.

i~ertia ~ember

of the is given

=2.314x2=.1540 30 For all straight members, the sti~fness factor equals 4EI/L. Hence,
K,..,
,'~D

=X,., wEii'

K AC

= KCA = 4x 2/60
4 x 6/60

0.1333 0.4000 0.0500

= ~D = KaF = 4 x 1/80 = K ,A AF = K..KCD

= Kac = a x 1/80 =0.0500

~ACTORS

uISTRIBUTION JOINT A.
ZK

AT 2ACH JOINT: (

car ry over from C ' -.452 X CI-880\

= 0.1333 + .0770 + .0500

= 0.2603

Let D equal symbol for distribution factor.

D illC

~ "'I~ . ".I~ " "I~ i f " " "I ~

;':<:0<0
-:'

' O' F. . O' 452

01'" t-'''' -=I0t-,M"l"IN-I,

<:;0..".

'I'

'I' 'I

"0>

,'0

'"'

'I

'-

.5121
AC

I C

D~C
D AF JOINT F. ZK

= .0770/0.2503 = 0.296 = .1333/0.2603 = 0.512

=
.05CO/0.2603

CA

0.192

-3000 3000 1536 -1536 -768"--- - -122

I
I

Ut-I
"

"d
':1

----sr--82

----:u-=
.0500+0.400+0.1540 .0500/.6040 .4000/.6040 .1540/.6040

0.6040

--r2--

-25

D FA D FD D'ED

= =

= =

.082 .663

6
F
1

-d
-2353

T r-CillJ------l
.

= .255

Carry-Over FactorS: For the straight members the carr-J-over factor is 0.5 and the moment sign is the same as the distributed mament when the Sign convention adopted in this chapter is ~sed. For a half circular arch at constant I, the carrJ-over factor was derived in the previous article and '~s found to be 0.452. T~e Sign or this car-r-y-over moment; was the s ame Sign as the distributed ~cment at the ~ther end of the beam. However, ~sing the sl~n convention as adopted ~or the ~oment ~lstr: bution in this book, the ca~ry-over faC~Gr would be ~inus or of or~osi~e Sl~, :0 the

(0 DF 9750/Y
-15COO
--1875

15000

2122
-106lJll'E
469

-234/ 105 -52/ 23 -8753

Fig. AIL 77

of ~r~~e ~~d l~aGi~g, the ~r~ce5s ~eec only Je ca"riec c~t .:'cr one-half 0: fra~e, t~us in ?i 6~re Al:.~~, ~n:y j01~ts A ~~d ? are cor.s1de"ed since the mzme r t ca I r asu.l t.s a: D anc C 'No~11 be the sane as at )'. and F' respec c ivatv , In the :l~~re t~e c12trlj~tl:n .:'actors are shewn in

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All. 33

the 0 at each joint. The process is started by placing the fixed end moments wi~h due regard to s~gn at the ends of members AC and FD, n~nely, -3000 at AC, 3000 at CA, -15000 at FD and 15000 at OF. ~e now unlock Joint A and find an unbalanced moment of -3000 which means a plus 3COO is needed for static balance. Joint A 1s therefore ba:anced by distributing .512 x 3000 = 1536 to AC, .296 x 3000 880 to ABC, and .192 x 3000 = 576 to AF. Short horizontal lines are then drawn under each of these distributed values to indicate that these are balancing moments. Carry-over moments are immediately taken care of by carrying over to joint F', .5 x 576 = 288. From A to C the carry-over moment would be .5 x 1536 768 and therefore the carry-over :rom c to A would be .5(-1536) = -768 which is recorded at A as shown. For the arch member ABC, the carry-over moment from A to C would be -0.452 x 880 = -401 (not shown) and therefore fram C to A = -0.452 x (-880) 401 as shown at joint A in the figure for arch member CBA. Joint C in the figure has been balanced once for the purpose of helping the student understand the sign of the carry-over moments which flow to the left side trom the right side of the trame.

inertia about x and y axes through the elastic center.

Bly I.. 1 - 1 - ---l d:; 4.6r:-T
i

---.-~
I

19. 1"

k
Fig. All. 78

y =

19.1 x n (30/1) : 14.4-9" 60 'IT x 30 +1 2

(NOTE:- 19.1 equals distance from line AC to centroid of arch member ABC.) Calculation of moment Of inertia I : x

r.:ember ABC Ix
rtrd a = -1- + 1.3r~

= .3 X 13D;!
rlember AC = ..Q.
2

+ --1n x 30
X

x 4.61 a

= 10100

After balancing joint A and taking care of the carry-over moments, we Lrnaglne A as fixed again and proceed to joint F where we find an unbalanced ~oment of -15000 + 288 = -14712, thus plus 14712 is necessary tor balancing. The balancing distribution is .255 x 14712 = 3750 to FED, .663 x 14712 = 9750 to FD and .082 x 14712 = 1212 to FA. The carryover ~oments are .5 x 1212 = 606 to A, .452 x 3750 = 1695 from D to F by way of the arch ~ember and -.5 x 9750 = -4875 from D on ~ember FD. We now go back to joint A which rzs been unbalanced by the car-ry-over- moments and repeat the balancing and carrying-over cycle. In the complete solution as given in Figure A11.77 each joi~t A and F was balanced five times. The :lr~l bending ~oments at the ends of the members at each joint are shown beloW the double short lines. The arch member ABC has 3 unknown forces at each end A and C or a total of 6 u~~owns. ~~th 3 equations ot static equilibrium available ?lus the ~own values of the end moments at A and C as round from the accent df s'tr-t bu-. tion process. :he arch nember is still statically indeterninate to one degree. ~hus the horizontal reaction at A or C as provided by the ~xlal load in oember AC ~ust be found before the bending moments on arch ABC can be calculated. The :irst step in this problem is to find the elastic center of the frame portion com;JQSed of members ABC and AC, as snown 1:1. Fig. Al~.78, and then :in~ the elastic moments of

l4.49 a

6310

Ix total

= 16410

Total elastic weight of structure equals Z dsll = (n x 30/1) + 60/2 = 124.24. The next step 1n the solution is to draw the bending mOme~t curves on this frame portion due to the given load on member AC and the end moments as found by the moment distribution process in Fig. All.77. It 1s composed of three parts labeled (1) to (3) in Fig. All.79. Portions (1) and (2) are due to the end mOments and ?ortion (3) due to the distributed lateral load on member AC.

1094 ~.UCilllli~""':liCcc..:d'-l Fig. All. 79 2353

The next step is to f~nd the 03 (area of Mlr curve) for each ~ortlon and its centroid location.

All. 34
~Sl

THE MOMENT

DISTRIBUTION METHOD

= 1094 x n x 30/1 =
60

= 103000

0.,
~S~

x 2353/2

=
l:0 s

70590

(-.667x60x4500)/2 -90000

63590

Fig. All.30 shows the ~s values concentrated at their centroid loca~ions and referred to the x and y axes :hrough the elastic center.
y

46 lb. The member AC also suf:e~s an ~x~al l~~~ due to the shear ~eactions at the tJ~ o~ =e~:~"2 FA and DC. ?ig. All.Bl shsws ~re8 te118S :: ~he s1de member ?A and DC wi:h the en~ =~,e~t~ as found in ~ig.All.77. The shear reaction at A and C can be found by taking ~oments abo~t lower 80" ends. Thus for FD,

l
!

0$
0

4.61"......"

.,.II

= 103000 'x--~--~,---x 0 I

A'

It

Gl, .70590.

c-LFig. All. 80

14.49

~:: ~90000

iF 38.1 lb. Likewise ~-.:... ,-' He = 38.1 lb. These 1791 1791 ~eactions react on Fig. All. 81 cross member AC in t~e oppOSite directions thus Giving a compression load of 35.1 lb. i~ member AC, which ~ust be added to the tension load of 46 l~. :rom the arch reaction to cb~ain the fir~l load in the cross-member.
~

RA

(1259 + 1791)/30

jo

Bending Moment in Lower The frame has been imagined cut at A and the arch end at A has been connected by a ~igid bracket to the elastic center. The redur,dants at the elastic center which will cancel any relative movement of the cut faces at A can now be computed.
= - ~ 05 -03590 Z d8/1 124.24
be

Arc~

Member

?~~

The horizontal reaction on arch ?SC must found before :he STue bendin~ moments on ~he arch can be found. The proc?cure is the sase as ~or the upper arch. 2ig5. AIl.SI, 22, 63 show the results for the lower fra~e ~ortlon.

=-

673 Lns Ib ,
Fy D

Xo = Z/8 Y = [102000 (4.61) + 70590

T \ _ IE.c. ...i....X_\_ __
E

(-14.49) _ 90000 (-14.49)J / 16410

=46

-"';: I -::/
Iy

3.34" 19.1 t "~';1 1 . / ,...., x..:.:.L:.. f I 22500 . 3 3 E

F~ID/

'I

8753 \.\/592

:::J

lb.
x

Fig.AIl.SI

Fig. AIl.82

Yo =

-z1sx
y

~ 0 because

~ zero.
Y

= 19.1 x

n(30/2)
+ 60

n x 30

The bending moment at any point on ABC or AC equals that due to Mo and Xc plus the moments in Fig. All.79. For example,

26
15nx3.34'1 = 4578
(60/6)(15.75)'

z~ = 57.1 I

Ix =

.3X + Z30.3

Ix
2rom P:g. All.82, -8753 x 60/6

= 7056

= 2478

At point A on member ABC,

MA = 1094 - 670 + 46 x 14.49 = 1091 (should be 1094 since moment as found in Fig. All.77 is correct one. Small error due to slide r~le accuracy. At paint B on
MB
~ember

= - 37530

.667 x 22500 x 60/6 = 1500GO 05 3 = - 5962n x (30/2)

ABC: :::II

= -326000
0s = -265500

= 1094-670-45xl5.51

-289 in. lb. :;'::'g. ).11.33)

The horizontal reaction at A will not ~roduce any bending on nember AC, thus t~e values in Fig. All.79 are the true moments. The axial load in ~ember AC by statics from F:g. AIl.BO equals Xo or a ~er.sio~ load of

So Iv tng for Redundant; "or-ces i"LJ anc Xo (See

M - - Z 2iS '0 - Z dS/I

= - (-265500) =
57.1

..',.650

22

su

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All. 35

x .Z0s Y Ox15 . 76- 328000 (- 3 . 34 ) Example Problem 2. External Loads on Curved : - 8753OX15 . 75+15CCO :-lembers. . o -x 7056 X o = 294. lb. Fig. AIl.85 shows a frame which has exy ternal loads applied to the eurved member as F I D well as ~o the straight members. The fra~e supporting torees r~ve been assumed as acting c~ 1 ' ),Mo'll4650./ l~. 76 uniformly along the side members AD and CEo x _ - - ;Xo=294J-X--+The problem will be to determine the bending i 14.24 moments at frame points ABCDE.
E

Fig. All. 83

1000

1000
I

Bending moment at F on member TED,

MF = -6962+4650-29'1,x15.76
ME = -6962 + 4650 + 294

=-6962

\ .,.. \.15

in.lb.
30
A
I

x 15.76 =

2340 tnv Lb ,

'(. /1 s 2

--

,.)\
,
C
1=1 L=60

The axial load in member FD = 294 Lb , compression plus 38.1 1bs. tension due to sfiear reaction from side members at pOints F and D or a resultant load of 255.9 lb. compression. Fig. All.84 shows the final bending moment ~iagram on each member of the frame.
B

w s 101b/in.

Fig. All. 85

I =1 L-60

1000 lb
00 v
,-

30'

259

Member ABC

I : 3,

L a 60

SOLUTION:\

1094 A~ 2147

Calculation of stiffness factors K. Member ABC.

Member AC

KABC = 2.314 I/r = 2.31'1,xl.5/30 = .1158

=4x 2/60 = .1333

KAC KAD

= '1,IIL

='1,I/L
ZK

= 4 x 1/60 = .0667

2353 1259qA

2353

KDE = '1,I1L

=4

x 3/60 = .2000

CS'259

f;7
\ MemberAF Member CD

Calculation of Distribution Factors D.

JaINI' A.

.1156 + .1333 + .0667

.3158

DABC :: .1158/.3158 DAC DAD

=
= =

.366
.422 .212

.1333/.3158

0667/.3158
ZK
~

1791

13747

JOINT D.
Member FD

.0667 + .2000

= .2667

DDA DDE

= .0667/.2667 = .25 = 2000/.2667 = .75

8753

5952

The carry-over factor for ABC is -0.452 as previously derived for a half circle arc, and 0.5 tor the straight members.
962

Calculation of
2340

~ixed

~d

Moments
1
12 wL:a

Member FED
E

For member AC, riA = ME

= 12 x 10 x 60

_ I I __

3000 ?or DE, !'IE = !'IE

Figs. All. 84.

Final Bending Moment Curves on Each Member of Frame.

= ?L/8 = 1000 x 60/8 =7500

All. 36

THE MOMENT DISTRIBUTION METHOD

Curved member ABC.

~ember

total elastic weight of ~emjer and its elastic mcments o! iner~~a abo~t ~es t~rJugh :~e
e Ias t Lc cent.er

The ~lxed end moments on this curvec due to the ex:ernal loads will be determined by the elastic canter method. ~he assumed static frame condition will be an arch pinned at A and su?ported on rollers at 8. (See Fig. All.B6)
1000

TOTAL

~~STIC

~~IGHT

nr ': Z ds/I =--

= n x ~O
l.v

62.83

Fig. AIl.S7 shews the general Shape of the static moment.cu~re. :or the fr~~e portion between the reactions and the load paints, the bending moment equation is M=P(r-r cos a). For the beam portion between the two 1000 lb. loadS, the bending moment 1s constant and equals, M=P(r - r cos 30 0 )

I 00:" 30~_ .- .AA, -r-~--

I'

!~\I
/

=P

1000 : P

Distance y from line AC to elastic center of arch ABC equals 0.6366r = .6366 x 30 ': 19.1 i~. Ix 0.2978r 3 / I 2978 x 30 3/1.5 = 5350

_---,.<~ Os~ 167300

Fig. All. B6
(2)

-/-.---'~ (1'

~,0Mo
I

"0-

i
19 1

0'" -Os' =28800 "'l 1 1

(1)
Fig. All. 87 Static Moment

\
-- -

'lIT....-.%-.

Curve.

Fig. All. 88

Calculation of tar.

redunda.~t

forces at elastic cer--

Calculation of 0s Values.
C~les.

110 - Xo

Z 09
Z

osvt

= -(28800 +' 67300) = 62.83 in.lb.

The g values equal the area of t~e X/I The moment curve in Fig. AII.S? has been broken down into three Darts labeled (1)
(1') and (2). -

-~-

_ Z 0sY _ 167300x5.7+28800(-9.l)
;:)350

129.3 lb.

0S 1 + .elsa = 2ti a(a - sin

a~

Yo =

=0

(because x

= 0)

= 2~0~~5X 30 '(.524 _ .5 ~ = 28800

The vertical distance y trom line AC to centroid of S1 and .elSa values is,
y

The fixed end ~oments at ends A and C will equal the moment due to the redundant forCes ~~ and Yo since the static ~cment assumed was zero at A and C.

= r(l- cos
c
c

-~)

= 30(1- .867 e
= 120 0,

0.5/2)

MA = Me = M o
Xament

+ Xoy

= - 3122

129.3 (- :9.:;

s in c

.524-.50

= - 652 in. lb.

ro

in.

~istribut1on

Process

):/sa

no:

=- Prae (1 _ I cos a ) .

30 0

., = 1000~30'X2.1
.0

(1- .867)

167300

Vertical distance y from line AC to centroid of

.elSa equals,

2r sin 8/2 = 2x30xO.S67

2.1

= 24.8 in.

Fig. All.SS shows the 0s va1ues and their location with respect to x and y ~es ~hrough the arch elastic center. The elastic 8entar
~ethod

reqUires the

Having determined the fixed end ~cmen"'Cs, distribution and carry-over tacters, ~~e =oment distribution can now be carried o~t. Fig. AIl.89 shows the solution. The first cycle wi:: be explained. Starting at joint (A) the unbalanced moment is -3000 - 652 = - 3652. The joint is balanced by dis~ributlng .~2 x 3652 = 1543 to ACj .364 x 1543 ': 1333 to AbC and .212 x 3652 = 776 to AD. The carry-over moment from C to A = .5(-1543) : -772; from C on member eBA to A = - .452(-l332) ': 601; and trom A to D = .5 x 776 = 388. Now ~roceedl~g to joint 0 the unbalanced TIoment is -7500 + 388 = -7112. ~~e jOint is balanced by distributIng .75 x 7112 = 5340 to DE and the re~ainder 25 per :ent = 1772 to DA.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

All. 37

__ over from C _.-lS2. (_6S2.\ = car,

u '" ...,'" ""ao ....'''' >''''' -,<:>,.,..... 1::1"'''''$'''''-''''''''''''''<''''<"i'~: ";'1'7'71' 'I' ",

(/ r : ~

Static Mom. Curve

366

"I'

,.212.
I

~ :;ii/
302./

.422.1-",;;=------:"::0'"01 C
-3000 i.t.M. I

\~
I

(1)

( 1)

2513 Static Mom Curve

Fig. A1l.90

151

-l:~/

.~~/

-r/a

wLi!. = 4500

.~~/
~
-2513

-0&.
7500

I I

0s

= 2513

x 60/2 = 75390

o
Fig. AU.89

l..(illl--------' "
.7500 ~340/ _2670

);0 s

= 218290

Ja

Calculation of elastic center location and Distance y from line AC to elastic center "
19.1 x
1\

-~~g/
602/ .401

-.ill/ -156

---!ll/ -60
TOTAL
Ix

nx30+60

x 30/1.5 = ~ 92.83

= 12.97"-

1.5
EL~STIC

="
T~e ca~y-over

WEIGHT

=92.83.
2

.Z978x30:'l +..1!..E2.9. (6.13)2 +.Qx 12.97 11

moments are: -

1.5

1.5

Fr-om S to D

= 12713
.5 x (-5340)
2

-2670

To A

~rom

D = .5 (1772) = 886

Fig. All.91 shows the elastic center location and the s values together with their centroid locations.

The :irst cycle is now completed. Five more cycles are carried out in Fig. AII.89 in order to obtain reasonable accuracy of results. The fl~al end moments are listed below the double short lines. On arch member ABC the end moments of 584 are correct. However before the ene moments at any other potnt on the arch can be found the horizontal reaction on the arch at A must be determined. This reaction will be dete~ir.ed by the elastic center method.
?ig. All.90 shows the bencing moment

---:7f::::::::::.f/ = Il"': . ,:=:.,

11.63-,-

IB 0., 167300
0
-

-f---~co-'~'--'~' x,
'T-'------::;,;t' ~ I

I 6.13 -.

--T-"I- x
\

Ps, '2BBOO

Y =, 12.97
C
14610

'<,

for ~2mbers ABC and AC as 5 parts labeled: to 5.

c~ves

~de

up of

tis
Fig. All. 91

= -90000 '" 75390 = -

Solving for redundant forces at elastic center. Calculation of 0s val~es Nhic~ equal area of each ~oment curve ji,1ded by r of member.
2~Sl

= 25800
=
167300
rt

l: 0s M o = z ds/I = -218290 92.83

=-2356
12713

0 '. 0s , 0s ,

=..usL=
Ix

167300 x 11.83+ 36800 x 6.13 +

= 58~ x =-

x 30/1.5

= 36800 =
-90000

28800 (-2.971 - 14C:O (-12.97)= 181 12713 lb.

( .567 x 4.SCO x 60/2)

AIL 38

THE

MOMENT

DISTRIBUTION

METHOD

The bending ~oment at ars pOin~ en T.ember AC or ABC equals the values in Fig. A~l.~O ~lus those due to X o and Xo in ?ig. All.Sl. Thus at
~oint

we ao tb/tn.
T

20"
Fig. 7

A on member ABC,

, ,

400~

1 ' L~h'/ ~~O

I

I =6 1~400 ]:::4 1=41 10

-~

MABC = 584-2356+181x12.97

554 in. lb.

which is the same as found by the moment distribution process. At point B on arch: ME tram Fig. AIL90 Mg from Fig. All.9l
100

Fig. 8

400 600 :-12 ~

= 584 + 4000 = 4584

-2356 -181 x (30 - 12.37)

r-I O
I

400

~_100

Pin

': 4"

1- 5" ..1

14

I '" 6

1= 3
11=2

AI=l

:"'-8"----'
I

= -5476
Hence MS = 4584-5476 ,. -392 tnv Lb,
STRUCTG~

lPin J.. r- 24"-f- 20"----i 5" Ii""""

Fig. 9

[=2

;I=3

I--- 16"

Fig. 10

'ITH JOINT

DISPLAC~S

For unsymmetrical frames or for frames loaded unsymmetrically, the assumption that only jOint rotation occurs may give resulting moments considerably in error. Joint displacement can be handled in a manner as previously explained and illustrated for frames composed of straight ~embers.
The distribution of the skin supporting

(3 )
~12"~

12 lb/in.
Itt

rt'
f-20',-+---40" , .

---!--32" --!--Fig. 11

40"_20'~ ,

forces on the frame boundary are usually taken as following the shear flow distribution for the shell in bending as explained in a later Chapter.
All. 20 Problems.

(1) Deter.nlne the bending moment diagram for the loaded structures sho,vn in Figs. 1 to 6.

In the loaded bea~ of Fig. 11, the supports a and E of the elevator beam deflect 0.1 inch ~ore than su,ports C and D. Com?ute the resultant bendinz mOwents at the suppo~ts and fine reactions. EI = 320,000 Ib.i~.sq.
(4) Fig. 12 illustrates a continuous 3 span Wing bea~, carrying a uniform air load of 20 lb./ln. Deter.nlne the beam bending moments at a strut points A and 8. Take
IAB=171n., IAA I = 20 in., and E = 1.3x:l.O e ln~

1 i f-8"-+-16"-----l
Fig. 1

1100

t:.t= ' ~
1100
Fig. 2

w= 101b/in.
I

"I
In.

24"----1

I-- 16'.--1---- 12"-~

Fig. 3 ;
100 .... 12~ ,

I::2

1=1

'fo w=201b/. 200 y" . . .~ fll'=P h+ito16 ----{

Fig. 4

Neglect e~fect of support deflection due to strut axial defor.rnat:on.

~50"~
f ,

110,-"---100'+ 110'~50'1
I ,

.A

f- 14"
,

hI

'!

1~ 12 ......
t

.<W=.20.rb~in,'

1_2

36"--~ 16"
Fig.S

, ""

-j
Fig. 12
w = 121b/in.

100 ,

20C

I I

~6" +-12"-+12"-.L'- 16"-+ 12"-j

Fig. 6

I
I 1.5

I"

I - 2

Ii

: I

*i

12

(5) Figs. 13 to 15 are leaded structures :hat suffer jOint translation. Jeter.nlne be~ding moment dfagz-am ,
500 l~
24" iI = 2 ,;

(2) Find the bending moments at all ~oints and sup~or~ ~oints Jf the ~oaced s~ruct~res in Figs. 7 to 10.

-1..""'''''''T

111 ...... ~.:.. 1-6"

'N

....." 12

-<5f~O~~

100

200

1 -:;

=10
'I

,
200 __I

-_ ..-c ~~
_

I=4

'

,300
1_5
N II

:1= 4

10" -~

f- 20"-;
Fig. 13

Fig. 14

-20"

-r-- 24"--":
Fig. 15

1:.:
I

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

CHAPTER Al2

SPECIAL METHODS . SLOPE DEFLECTION METHOD

A12.1 General.

The slope-deflection method is

Fig. A12. 1a

another widely used special ~ethod for analyzing all types of statically lndeter.nlnate beams and rigid frames. In this method as in the moment distribution method all jOints are considered rigid from the standpoint that all angles between ~embers meeting at a joint are assumed not to change in magnitude as loads are ~pplled to the structure. In the slope deflection TIlethod the rotations of the joints are treated as the unknowns. For a nember bounded by two end joints J the end moments can be expressed in terms of the end rotations. Furthernore for static equilibrium the sum of the end moments on the members meeting at a joint ~ust be equal to zero. These equations of static equilibrium provide the necessary conditions to bzndle the ~own joint rotations and when these unknown jOint rotations are found, the end moments can be computed frum the slope-deflection equations. The adv~~tages of the slope deflections method will be discussed at the end of the chapter after the method has been explained and applied to problem solution.
A12.2 Derivation of Slope Deflection Equation.

I
I

,.ill f-I------,=.,...--:;c=-:-----I1(2) EI Constant "I

1..- C
Fig. b
d1 MJ,

-Y- __

-L

-.< - ~ - '-'_Q, - - -

->V-

'il'

d.

i) M.

Figs. Al2. 1

The problem 15 to dete~lne the relationships between the displacements Of the end supports or a beam and the resulting end moments on the beam. Fig. A12.1a shows a beam restrained at ends 1 and 2. It is assw~ed unloaded and at constant cross-section or moment of inertia. This bea~ is now displaced as shown in Fig. b, namely, that the ends are rotated through the angles eJ, and e a plus a vertical displacement dJ, and d lil of its ends rr-cm its original pos i , tion, which ?rOduCes a relative deflection 6 of its two ends. The angle 0 representir~ the swing of the member equals 6/L. The problem is to derive equations for MJ, and M a in te~s of the end slopes 6J, and 6 3 , the length of the jearn and EI. Figs. c, d and e lllustrate hew the total beam deflect10 in ?ig. b is broken dcvm into three separate cer iect ions , 1:1 F':g. c , the end (2) is considered fixed and a moment Mi is applied to rotate end (1) through an angle e 1 In Fig.a, end (1) is considered fixed and a moment M: is applied at (2) :0 rotate end (2) through an angle 6 3 In Fig. e, both ends are considered fixed and end (2) 1s displaced downward a
A12.1

distance !J. which causes end moments M~ and M":., . In deriving the slope deflection equations each of the three beam deflections will be considered separately and the reSUlts are then added to give the final equations. Fig. A12.2 shows Fig. c repeated.
Fig. A12. 2

M'a b 2

~M,Ml
~

'--2-

The applied moment Mi, is pOSitive (tension in bottom fibers). Since end (2) is assumed fixed an WL~own ~oment M~ is produced at (2). Fig. Al2.3 shows the bending moment diagram ~de up of two triangles. Mia being unknown '...il1 be

A12.2

SP ECIAL

M E l' HOD S -

DEFLECTION METHOD S L 0 P'-"E--="'-'-"c:..::...:.:.:::...:.:.....:::.-"-''-''-''-''---

assumed also ~osltlve, as the algebraic solution will deternlne the true sign of r.~.
~NO ~oment area theore~s will be used 1~ deriving the slope-deflection equations, namely (I) that the change in slope of the beam elastic curve between two ~oints on the beam is equal in magnitude to the area of the M/EI diagram between the two points, and (II) the deflection of a paint (A) on the elastic curve away from a tangent to the elastic curve at (B) is equal in magnitude to the first ~oment about A of the MjEI diagram between (A) and (B) acti~g as a load.

Fig. A12~6 shows ~he ~oment diagr~ ~0r ~~ applied bending noment ~~ which rotates en~ (2) thro~gh an angle 6 2 Nhen other ~nd (1) is ~!xed (See ?1g. A12.5). In a similar ~nner as
desc~ibed befo~e,

Sa
~ak1ng

(-M'lL/2) + (-~~L/2) E1 ~I

= :-~(~l1~"~.;+~,,;.~~~)"_L
2E1

_ (5)

moments of M/EI diagram about (2),

+

(l1~L)!:
\2E1 3

(l1~L) 2L = 0
2E1::3
- - - - - - - - - (6)

In Fig. A12.2 since Sa 0, then 6~ will equal the area of the M curves divided by Sl.

or,
whence,
(1)

e =
1.

(11\L/2) + EI

(11~L/2

) = (11'~+i1~)L
2E1

E1

e. = M~L 21

an

d Mil - 2ElS"
~

(7\.

The deflection at end (1) away from a tangent at (2) equals zero, thus we take moments or the moment diagram about (1) and equate to zero.

Then from equation (6) M " -

4E16 a L

- - (S)

h (!ill,,) 2L (lille) 2E1 3 2E1 3

+

whence -tl'llI
diagram.

=!1'J2
tr~e

--

(2)

Fig. A12.8 shows the third ~art of the beam deflection, namely, support (2) is jef:ec~sd downward a dista~ce 6 assuming both ends fixed.

Fig. A12.4 shows the

shape of moment

SUbstituting ~esults in equation (2) in equation (1) we obtain,

or

I1' 4E1el. ,L

- - - - - - -

(3)
Fig. A12.9

,
M'''.,h , 2

Then !rom equation (2)

11'. =

2EI6j
L

- - - - - - - - - - - - -

(4)

M'{'

c:==:====-==

.............-- .
, '" M

M'a~t

Fig. A12. 10

Fig.AI2.5

"r (, ~_"'-=- __
I---.i; L ----r-Fig. AI2. 6

The change in slope of the elastic curve between the enes equal zero, thus Jy the moment area theorem the area of M/EI diagram equals zero. Hence, Whence
M:~

+ M~ )L

2El

o
- - - - - (9)

=-:1':-

Frem the deflection theorem,

Mi,': ""
Fig. A12. 7

e:t====="

M'~

(10)
Subt. (9) in (10)
= - 6EL\.
~-

(11 )

ANALYSIS AND DESIGN OF

+ 61:.6-

FLIGHT VEHICLE STRUCTURES

A12.3

= --'i-'T', 41

(12)

(lea)
(lSb)

The ccmo tned effect of deflection Or , o, an~ ~ can now be obtained by adding the separ3.te results.

A12. 3 Hinged End. Slope Deflection Equation.

Let K ~ ~C and 0 ~ ~/L

Consider that end (2) of beam 1-2 is freely supported or in other wards hinged. This means that Ma_l. is zero. Thus equation (18) can be equated to zero.

(13)
ana
;1a_ ..

whence
2K~ ~ -K81. + 3K0' -

= -2K (28,,+9 .. -30)

(l4)

.SMFa

These end ~oments due to the distortion of the supports must be added to the moments due to ~ny applied loads on the beam when considered ~s :lxed ended. Let these fixed ecd moments be l~beled MF l and MF". Then equations (13) and (14) can be written including these mcments.

Substituting this value in equation (17)

A12.4 Example Problems.

Problem 1 Fig. Al2.11 shows a ~NO span continuous beam fixed at ends (1) and (3) and carrying lateral loads as shown. The bending ~oments at points lJ 2 and 3 will be dete~ined. Relative values of I are shown for each member.
100 lb.

(15) (16)
MCCI?I~D

SIGN

CO~~ION

Squatlons (15) and (16) have been jeveloped uSing the conventional signs for bendi~g ~cments. In general there are adV2~ntages of uSing a statical Sign convention 2.3 '.'las ueed in the moment d t s t r tbut ton method in :hapter All. Therefore, the following sign :onvention will be used in this chapter: (1) The rotation of a jOint at member 1s posi-

~ 12" ~

-.:21Ji!i

w = 24 lb/In,
L 24" 1=48

L ; 24"
I~24

Fig. A12.11

SOLUTION: Calculation of fixed end mcments due to

app l I ed loads -

tive 1: it

tu~s

in a

cloc~Nise

direction.

(2) An end mcment is considered Positive if it tends to rotate the end of the member ~loc~N1se or t~e jOint counterclockwise.
This ~dopted s~gn convention is the sa~e as adopted :or the ~o~ent distribution method (See Art. Al:.2).
~Gen
~or :h~s

MF .._ a

:::

-PL/8

100 x 24/8

-300 in. lb.

Mp,_, = PL/8 = 300

equations (15) and (IS) are revised new Sign ccnvention they become: -

Ml._ a 11 41_
..

2K (26",+6 2 - 3 0 ) +MS- ..
-

(17)

~ 2K (29 2 + 8 l
(~7)

391) +?'!S-"

(18)

The S1gns as shown are deter.nlned from the tact that the end moment at (1) tends to rotate the end of the T.ember counterclockNise which is a ?ositive noment according to our adopted sign conventicn. gy similar reasoning the fixed in moment a: (2) is positive because the moment tends to rotate end of ~ember clockNlse. for span 2-3: MF a-3

~q~a~ions

and (18) are referred to as the sl~pe-ceflectlcn equat~0ns ~here K = SIlL and o = oiL.

=wV /12 = 24 x 24 /12 = -1152

a
~

Inc Lb ,

?or no Yielding or ~ransverse movement of sU~yorts, ~ ~ 0 a~d equa~1o~s (17) and (lB)

K vaIues

will

~e

K = SIlL. s ince considered '~ity.

~

is constant it
1

FOR BEAM 1-2, K = IlL ~ 24/24 FOR SEA!'! 2-3, K = 48/24 ~ 2

A12.4

SPECIAL

METHODS

SLOPE DEFLECTION METHOD

Ex~mp1e P~cble~

SUbstituting in slope-def1ec~ion equations lea and 18b. (~= a since supports do not translate. )
Beam 1-2.
Mlo.

2
100 100

100*

81

0 because of fixity at (1).

1 1

(1)

--12'--00.;
t

L8"+8"~a"~ IIIlli\l\
Fig. A12.12

= 2K (28 + e.l +i'IF I1 = 2xl (0+8.)-300 Ml._,. = Z8,. - 300 - - - - - - - - - - M.- = 2K (28 + 8:1.) +i'IF '1 M '1 _ .. = 2xl (26. + 0) + 300 l'l~, = 48'1 + 300 ---- - ---- -1 - ,. 1

L ' 24 I = 24 K =1

(a)

Fig. A12.12 shows a loaded 3 span beam fixed at points (1) and (4). ~hls ~~xity at (1) and (4) causes 81 and 8~ to be zero. Fixed end moments:

It)

MFl,,_g
!iF -'1-1

=-

PL/8

= 100 x 24/3

= - 300,

Beam 2-3.

6.

= 0, because ot fixity at

(3) .

= 300
= (100x8xlS Il / 24 )
=-534
+

At

= 2K (28 a + 6,,) +i'IF M._ a = 2x2 (26. + 0) -1152 M._ a = 86 - - - - - - - - - - - (e) l'l_. = 2.'{ (28" + 8 +MF" 11 = 2x2 (0+9,.) + 1152 M,,_a: = 49 a + 1152 - - - - - - - - - - - (d) joint (2) l:l'l = 0 for static equt r rbr-tum,
11 11"

MF-'1_::I

(lOOx16xa 2 / 2 4- 2

a-1152

Ml;"

'''-a

= 534

a )

_ .

l'lF._. = 400
SLOPE DEFLECTID)J ::QUATIONS (0
!11._11

0)

= 2 X l( 0 + 8:a) - 300 = 28 2 =

300

(a)

Hence

Mg-l,,:: 2xl(28 g+0)+3CO

48 11+300
2

{b )

MII _

"

= 2 x 2 (Z8,. + 6" ) - 534 = 86

+

+ -i-8::1

SUbstituting trom (b) and (e)

4Ba: + 300 + S9 a -1152

- 534 - - - - - - - - - - - - - - (e)

534 - - - - - - - -400 :: 4e 3 - 400-

I co )
(e)

whence 8 11

::

71

!13_.
M~_"

=2 x 1(28" + 0)
eQul1ibril~

With 8 11 known the final end ~oments can be round by substituting in equations (a), (b), (0) and (d) as follows: _
!11_.

= 2x1(0+8,,) +400 = 28"

+400

Joint-Moment
JOINT (2)

equations: -

=2x71-300
x 71
+ 300

=-158 in.lb.

M.._.. :: 4

= 584

whence
48 g + 3 0 0 + 8 8 a+48 3 - 5 3 4

!1._3 ::
M"_II

a x 71

- 1152 = - 584

= 4 x 71 + 1152

= 1436

or
- - (6)

Changing the resulting ~ament signs to the conventional sign practice, gives

JOINT (3)
80:]1'"48 a+534+48.,-400

=0

or
- - - (n )

I'1,,_g

= -1436

128:] + 4.6:a

134 = \) - - - --

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A12.5

Solving equations (g) and (h) for Sa and gives,

8 01

8~

or free to rotate we use equation (19) in writing equations for M a _ 1. and M 3_ 4

Ma_1.
M:~-1.

= 26.15,

8~

=-

19.9

= 3K8

l1

+MF a_1. -0.5 MF1.-a

(0

ls zero)

Substituting in equations (a) to (f) inclusive to !ind final end ~oments.

i'!1._a:: Z8 a -300

2x25.15-300

M"_4
!1~_4

-247.7 in.1b. Ma _ 1.
M2_~

= 3 x 1 x e a + 250 - 0.5 (-250) = za, + 375 - - - - -- = 3K63 + MF~-4 - 0.5 MF4-~

= s x i x s,
-500-0.5 (500) :: 39 3
- - - - - - 750 - - -

(a)

=46a + 300 =4x 26.15 + 300 =404.6

= S8 a
+48~

(b)

-534

= 8x26.15+4
=

(-19.9)

- 534
M~_a

=- 404.6
8 (-19.9) +4x26.l5

Using equations lBa and leb and substituting,

Ma-~

= 88~
=

+48 01 +534

+ 534

479.4

Ma-

M3 _

48 -400

= 4-

(-19.9) -400 =

M~_a

- 479.4
!1"'_ 3
~oment
::

!1~_a

= 2K{28 a + 8 ... M'l,;" , , = 2xO.5(Z8 + S +0 = 26 + 6 - =2K(Z8 +. ea ) +MIi" . = 2 x 0.5 (2e~ + e a ) + 0 = 28 + 6 3 )

l1 3 )

(c)

~-a

11-

(d)

28 3 + 400

=2

(-19.9) + 400

= 360.2

For statical equilibrium af joints: JOIN'~

(Note: - Student should convert to conventional signs and draw complete bending ~oment diagram) .
~ample

(2)

Ma _

1.

+ 11'a_3

=0
=0
- - - - - - - - (e)

Sutrt , 36 a + 325 ... 29,. + 8 3

Problem 3.

Bea~

with

Si~Dle

Supports.

or

sea
(3)

+ 8~ + 375

=0
750

Fig. A12.13 shows a leaded 3 span continuous beam with hinged su?ports at points (1) and (4), which ~eans that ~1.-a and !14 - 3 O. The ~oments at supports (2) ane (3) wi:l be dete:r:nlned.

Jourr
or

28 3 + 58 3

ea ... 38

3 -

=0
- - - - - (f')

... ell -

750

=a

Solving (e) and (r) for ell and 8 3 gives.

(21
L = 20 I ,. 10 K = 0.5

131
L I

ea
K=l

109.4,

6a

=172
(-109.4). + 375

= 20 = 20

T~e

end ~oments at (2) and (3) can now be found from equation (a) to (d) inclusive.
MlI _ 1.
M:a_ll

Fig. A12.13

= 36" + 375 =3
= Z8 a
+ 63
",

= 46.8
= -45.S

in.Ib.

SOLUTION:

=2

(-109.4):;' 172

Fixed end
~F1._a

~oments:

!'!3-OI

= 26

Sa

= 2 X 172 + (-l09.~) = 234

= 3 X 172 750 - - 234

i1... 11~

,~

. 3-4

=- PL/8 = 100 x 20/8 = - 250, !1F = 250 =!'!F ~_a = 0 (No load on span = FL/8 = 200 x 20/8 = - 500,
a_
1.

M~_4

=383 - 750

A12.5 Loaded Continuous Beam with Yielding Supports.

2-3)
~ovable contr81 surfac~s of an air~lar.e, the elevator, rUdder ~G aileron aTe attached at several ro ints to the s tact i icer ,
na~ely

The

!14_~

500

Slope-def1ection

equa~ions:

{~; ~re ~i~~ed

Since suppor:s at (1) ane

fin and wine r-esr-ec t tve Ly. T:"_ese suppor t t ng surfaces are usua:ly ca~tl1ever strUctU:-8S and ~hus t~e suppcrting ~olnts ~o" ~~= 7.cvab~e ccn~rJl sur~aces suffer a ~lsp1ace~ent :hus ~"o ~o~:~~ a cJ~tin'..:.ous ~e~~ cr. Jlelc~~s cr de~:ec:ed

-"."',

A12.6

SPECIAL

METHODS

SLOPE

DEFLECTION METHOD

supports. The slope deflection equations can take care of this support displacement as illustrated :n the following exa~ple ~roblem. E:<:ample Problem.
3 Ib/l.n

MF"'-3

40" (3 .0 =W 0 x

_.0

; -)

= -07
0U

The be~ has a CQnstant K for all s~ans equals,

~ 4 Ibl in

sec:~an,

hence,

14!1IilllilTIVIJll!,j!!
~10'~40"-

*--1.
" I

t3

40" -

40"--+-- 40"
'

I
I

~10't\'

'r

K = EI/L

(10,000,000 x 0.03339) 40

8347

Calculation of
wi~h

~ val~es.

Fig. A12. 14

tube

~/16'"
deflected sUPPJrts.

Span 2-3. The settlement of support (2) respect to SUDDor"': (3) :: (0.5 _ .':'875) = .3125 inches = ~

---------.

=- 6/L =-

,3125/40

=- 0,007812

cad,
~espect

Fig. A12.15

(Since (2) noves co~terclockN1se with to (3) the Sign of 0" is nega t tve , Span 3-4.

Fig. A12.14 is representative of an elevator beam attached to detlec~ing stabilizer str~cture at the five reaction points as indicated in the figure. The elevator beam 15 a round aluminum tube 1-1/4 - .049 in size. The moment or inertia (I) of this tube equals 0.03339 and the modulus of elasticity (S) or the material equals 10 million psi. The air load on the elevator beam is variable as indicated in Fig. Al2.l4. Fig. A12.15 shows the Shape of the deflecting supporting structure, which means that supports (2) and (3) move downward 'thr-ough the distances indicated on the figure. The problem will be to determine the bending moment at the supports under the combined action or transverse loading and settling of supports.
80LlITION: -

.1875 inches
cad,

=-/l/L =- .1875/40 =- .004688

Since the be~, external loadly.g and supper: settlement 1s sJillmetrical about support (4), the slope of the beam elastic cur-ve at (.;) is nor t , zontal or zero and theref0re 9.. = O. T~us cnly one-half of the structure need be considered in the solution. Due to the fact :hat (2) is a Simple support with a cantilever overhang, the moment M~_l is statically dete~l~ate and equa~s 5 x 3 x 3.33 :: 50 m.rc, Then for stat ;c equilibrium of joint (2), I1 11- 3 must equal - 50. Substitution in slope deflection equations (17) and (18): M.a_3

=2K(28
11

11

+ 8 3 - 30") + I1F 11-3

- 50 or

= 2X8347

[Z9.a +93 -3(,007812)J -426

Calculation of fixed end moments: For a trapeqoidal loading as shown in Fig. a, the fixed end ~omen~s are, M:i._ - 60 (5u + zv) Ma-:i. - 60 (5u

= 333886
~_a

+1669483 +15

=0

- - - - - - -

(1)

= 2K(29 3 + 911 = 333888

3

- 30') + :11'\_11

= 2 x 3347
(2)
Uj

_ L" _ L"

[293 +9'i1 -3(-.007812)J +440 Whence

~_II

+ 166948::1 ... 831 - - - (ao ,'"'... -

3v)

'"\'

SUbstituting in these equations tor the loading values as shown in Fig.

A12.4,
M'F.a_3 ::

Fig. a

Whence !13_.... = 333888" - 259

M..._"

-----(3)

= 2K(28...

- 8" - 30')

40"

(5x3+ 1) =-426

in.1J.

~_3
whence

MF3 _

11

=:gil (5x3+1.5)

= 440

=2 x 8347 [0 + 9" - 3(-.004688)J + 507 !1 ..._" = 166940 + 742 - - _ (4)

3
1I+I1 3 _ ...

For static eqUilibrium of joint (3), M3 _

= 0,

hence

3338893 + lE6g4e~ + 331 ... 33386~ - 259

=0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A12.7

whence , 567766;,+166948 11+572 = 0 - - - - - (5)

9;,

MFl._.lII

= =

(100xSx12-)! ZOa + (100x12xS)!

ZOIl =-480

Solving equations (1) ~nd {5J for 8 g and we obtain 6,.. = 0.00439 and 6" =- 0.00967.

I1F._, I1F,_,

480

SUbstituting i~ equations (2), (3) and (4) gives the moments at these points.

=100 X 24/8
= 10 x 20'/12

= 300.

11,_.

=33388
+ 831

(-.00967) + 16694 x .00439

=334.

334

582

in. lb.
259
+

11,_. = 33883
M4 ;,

(-.00967)

= -582
= 580

6 10

Due to symmetry of loading we know that =- 6 g and 8" 8., which fact will shorten the solution.

=-

= 16694 (-.00967)

742

Slope deflection equations: I11o_.lII M1o_.lII I110_.

converting these signs to the convent1onal signs would give Mil =-50, M" =--582 and
11 4
::::

=2K
=2

(28:1, +

e.)
-

+ MF:1,-.
) -

-580.

xI
10 -

(26 1
480

91

480

A12.6 Statically Indeterminate Frames. Joint Rotation Only.

= Z6 = 2K

--- - - ---- - ,
(26 1 + 8 a )

(a)

The slope deflection equations can be used in solving all types of framed structures. In fr~~es as compared to straight continuous be&~s, the axial leads in the members are usually more important, however the influence en Joint dls~lacement due to axial deformation of ~embers is relatively small and is therefore usually neglected in most simple framed structures. To illustrate the slope-detlect1on method as applied to fr~~es, the structure shown in Fig. A12.16 will be solved. SXample Problem 1.

11.-,

(2610+8~) + MFl.

M 1o
M1o _

= 2 x 1.67 = 6.678 1 = 2K

300

"

+ 3.338" + 300

-- --- 300 - -

(b)

M~-1o 11~,

(Z8~+e1o) + I1F"-10

M,,_l. I1"_4

= 2 x 1.67 (28,,+81.) = 6.678" + 3.336 300

1 -

----

(c)

= 2K
=

(Z6,,+8.) + I1F,,_.,

L,.. L4.. -L,. .J I

,

100

100

M"_4 M':l-4

= 2xl

(28,-8,) + 334
- - - - - - - - (d)

Z8,,+3.'34

L = 20 I = 20

Static Joint Equations: JOINT (1).

K"
L: 24 1 0 0 - I = 40 K = 1. 67 L: 24 I =40 lot-- 100 K = 1. 67

which gives,
281. - 480 + 6.676 1 +

3.338"

+ 300

=0
- (e)

whence 8.678:\, + 3.338" -180 = 0

JOINT (3

i.
which gf.vee ,

Fig. A12.16

6.676;, + 3.3361. - 300 whence

61.

26"

334 = a
(f)

?lg. A:Z.16 shows a closed rectangular

subjected to loadings on its four s:des wh:c~ hold the frame in equil~brium. The bending ~oments at the frame joints will be deuerrmned ,
fr~e

8.678" +3.336:\, +14

0 - - - - - - -

Solving equations (e) and (f) for 8" and gives,

81.

SOLUTION: -

= Z6.3,

9;,

=-13.94

A12.8

SPECIAL

METHODS -

SLOPE DEFLECTION METHOD

Subt. in equation (a)

Ml.-e

Due to the flXity

(5), 81.-a

=2x25.3-?80

=-428 in.D.

unknowns are e a -~, e.. _~, g e _

9:5-"

= 9&_., = O.

a~ j~i~ts
e

(1), (3)

~nd

"het-e r or-e , the

:::. nd 6.

Subt. in equation (d)

M:5-"

= 2(-13.94) +334 = 306

ShryNS

There are t;~ee statical ~?int equatior.s of equilibrium available, nameLy,

tnv Ib .

Fig. Al2.17 moment diagram.

the resulting bending Another equation is necessary beca~se there are tour unknowns. This additional equation is obtained by applying the eq~atic~s of statics to the free bodies of the Jartical ~embers. Fig. Al2.l9 shews the free bccies.
M
i

372

Fig.A12.17

233

f~

e-!l

T
I

Li,J
....1,.

194 A12.7 Frames with Joint Displacements.

'-'
In the previous example problems. the conditions were such that only joint ~otations took place, or it any transverse joint dis~lacement took place as in the ex~ple problem of Art. Al2.5, these displacements were known, or in other words the value of in the slopedeflection was known. In ~~ practical frames however, the joints suffer unknown displacements as for example in the frame ot Fig. A12.l8. The term ~ ~/L in the slope-deflection equation was derived on the basis of transverse dlsplacement of the member ends when both ends were fixed. Thus in Fig. A12.18 under the action of the external loadS, will sway to the right as magnified by the dashed lines. Neglecting any joint displacement due to axial deformation. the upper end of each member will move through the same displacement~. Then we can write,

I'_a, .

,J.,--H~

15

Me_t'l Fig. A12.19

M~_...

Treating each ~ember sepa~tely. we ta~e moments about the upper end ~.d equate to zero and then solve for the horizontal reaction at the bottom end. The results are,

With these horizontal reactions known the static equation ZH = 0 can now be applied to the frame as a whale, which gives. (20) Equation (20) is generallY referred ~o as the bent or shear equation and is an equation that supplies the necessary extra conc I t j on to take care of the additional unknown ~.
A12.8 Example Problems of Frames with Unknown Joint Displacement.
Exa~ple

Problem 1

with

Fig. A12.20 shows an unsymmetrlcal ~Ta~e ~~symmetrlcal loading. The bending ~oment c~ve will be determined.
SOLUTEN: -

Fig. A12.18

The relat:ve

~8rnent o~

:~ertia

of each

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

~embeT ~s

A12.9

given on -;:[';,e r rgur-e . The IlL r-eIa t tve K values are also i~dicated aGjacent to eacn member".

96.

M _ liI
6

..

= 2 x 1.667(28 .. + 8 .. -0)

-144

1-

-1

M .._.. = 6.6678 .. +3.3338.. -144I1.._ a

-------

(c)

= 2 x 1.667(26, + 6 a - 0)

+ 144
(d)

I1.._ a = 6.6678.. + 3.3338 1i1 + 144 -

Relative values of .0 due to sldesway of frame: -

11,_
M,,_ ..

The angles .0 are proportional to the ~/L for each member where ~ is same for member-s (1-2) and (3-4) and zero for member (2-4). (See Fig. A12.10).
Hence,

= 2xl(O+6,-3xO.l,0)+0 = Z8, - 0.60 - - - - - - = 48, - 0.60 -

---

(e)

Fig. A12.20

11_, = Zxl(Z8, +0-3xO.10) +0

11,~

----------

(r)

Equilibrium equations: JOIN]' (Z).

Substituting: 2.6678.. - 0.26670 + 103.68 + 6.6676 11 + 3.3336...

Fig. A12.10

- 144

" .. = 10 1 )(.1,,_

0.100

whence,

=0

9.33361i1-0.2667~+3.3338, -40.32

- - - - - I1.._ a + I1,_;s

- - - - - - -

- (g)

.0:a_-4.= 12 = 0
FL"'CZD 2ND :-!OMEN'I'S: -

JOIN]' (4).

=0
=0

6.6676, + 3.3338 11 + 144 + 48", - 0.60 whence, 10.6676, + 3.3338.. - 0.60 + 144

tlember 1-2.
I1F

.-.

=0

- (h)

=-(48xo"x9)/lS" =-69.12 tnv Lb

Writing the bent equation (See Eq. 20).

- (1)

'M:ember 2-4.
I1F._

In order to get H~ and H s in terms of end moments see Fig. A12.21.

=-

(96 x 12)/8 =-144

MF..-a = 144
~ember

3-4 has no lateral loads and thus end moments are zero.

~ixed

The slope-def1ection equations are: (17)

(18)
Writir~ above equat10ns for each ~ember and noting that e~ and 6 3 are zero jecause frame is fixed at support points, ~lves,

Fig. A12.21

For free body of member 1-2 take moments about (2) and equate to zero.
MII_~

+ I1~_a - 6 x 48 + 15H:\.

=0

111.-a
M:\._a

= 2xO.667(0+e"l-3x .0667Z1) -69.12 = 1.3336.. -0.2667Z1-69.12- - - - (a)

ne

h nee H - 288 - (I11i1 _ :\. +M~ a) ,.~Lc

1

Subt. 'ralues of M lI _
H~

and !11._.. in the equation, + .03555.0 + 16.88 - - - - (j)

:1"-1

=2xO.667(Ze"l +O-3x.0667.0) +103.68

= 2.6678.. 0.2667.0 + 103.68 - - - (b)

=- 0.26676

M a_

:\.

?or free body of member 3-4 take moments about (4) and equate to zero.

Al2.10

SPECIAL METHODS -

SLOPE DEFLECTION METHOD

I1e_ +!"',",-_., + lOR.,

=0

Ing the symme'tr t ca.L cf s tr-Lbut.ed load as shcwn .

w: 12 lb/in.

Substituting values of equations, gives,

H,

M~_~

and

M~_~

in above

I
10"

..

=- o.se.

0.120 - - - - - - - - - (k)
H
1.

SUbstituting values of (1), gives,

and

H~

in equation
- (1)

-O.2667ell-.6e~+.15550-3L12 =0-

t
20"
---L

: 10

' 20

Fig. A12.23

I : 10

L = 20":

Solving equations (g), (h), (1) for the unknowns e., e~ and 0 gives,

15
1'iiTTT1

0=

196.9, ell ;;: 12.17,

e~

=-

6.22

40"

The final end ~aments can now be :ound by SUbstituting these values in equations (a) to (t) inclusive.
M1._.

The bending moment diagram under the given loading will be dete~lned.
SOLUTION: -

=1.333x12.17- .2667x196.9
- 69.12

= -105.4 In. lb.

!1._1
11_~

=2.667x12.17- .2667x196.9
+ 103.68 =-83.6

Relative Stiffness (K values) Members 1-2 and 4-5, K

= :0/20 = 0.5

= =

6.667x12.l7+3.333 (-6.22)

- 144
M .._.

= 83.S

Members 2-3 and 4-3, K = 20/22.4 = 0.692 Fixed end

I1F~. MFa_
~cments:

6.667 (-5.22) +3.333x 12.17 + 144 142.9

(-6.22) - 0.6 x 196.9

= wLV12
=

= (12 x 20' )/12 =-?DO m Lb

s

M..~
M3_

= ,; =2

= -142.9

400, MF .._.,
~:

= 400,

MF;,_..

= -

4-00

(-6.22) -0.6x19S.9 =-130.4.5

Relative values of

Fig. A12.22 shows the bendIng moment diagram. The stUdent should also draw the shear diagram, fInd axial load in members since all these loads are needed when the
174.7
83.6 b
80.46

83, 6",,==0..;';;;

Due to the sloping ~embers, the relative transverse deflections of each member are 40t as obvious as when the vertical members are connected to horizontal members as in the previous example. In this example, joints (1) and (5) are fixed. Because of symmetry of frame and loading, joints (2) and (4) will ~ove ~ut~ard the same distance ~ as indicated in ~~g. Al2.24. Furthermore, due to s~etry, jOint (3) will undergo vertical movement only. In Fig. Al2.24, draw a line from ~oint (2') parallel to 2-3 and equal tn length to 2-3 ~o locate the point (3"). Erect a perpendicular to 2 ' -3" at 3" and where it intersects a ver-t i ca I through (3) locates the point (3 ') the final location of joint (3). T~e length at 2'-3' ' , the s ice equals 2-3. In the triangle 3-3' _3 1 3 =~. The relative values of ~ which are measured by 6/L for each member can now ~e calculated.
1_3"

130.4

105.4
Fig. A12.22

strength of the fra~e members are computed anQ compared against the stresses caused by the frame loading. Example Problem 2. Fig. Al2.23 shows a gable
~rame,

carry-

.I1h-ll = - 20 =- .05.0

ANALYSIS AN

ESI

N 0

v
I1._", = 2xO.S(-29. +O-3x .050) +0

A12.11

l ' /t/i E
3"1 3

\C/.
2'

0"-0i

=-

28. - 0.150 - - - - - + 8. - 30)

(g)

= 2K(2S",

+:1r ,-.

3'--_

= 2xO.5(O-8. -3x .050') +0

M",_

=- e:a

- 0.150 equation.

- - -

(h)

JOINT (2).

Equl1i~rl~~

I1 a_1 + M a~
5

= 0,

subs t i 'tu't tng -

28. + O.15 + 3.5688. - O.5352 - 4CO

FIg. A12.24
~~_

whence, 5.56S8 a -0.32520-400

=0

- - - - (i)

=-

0.1 SlaDe Deflection

~auations:

Substl:~tlQn i~

We ~cow that 8 L 8", 0, due to fiXity at jOi~ts (1) and (5). Also 9~ = 0 due to sjmmetry or only vertical movement of joint (3). Further8ere due to symmetry e. = - Sa

= =

The jOint equilibrium equations at (3) and (4) will not provide independent equations becaUse in the ~revicus substit~tions in the Slope-deflection equations e~ was made equal to zero and 8. was equal to - e:a

Shear 2quaticns: Due to symmetry ~~e ~orizontal reactions at points (1; ~~d (5) are equal and opposite and therefore in static balance. Si~ce we ~ave ~NO unknowns 6:a and 0', we need another eq~tion to use with equatien (i). This equation can be obtained by stati~g that the horizontal reaction Ha- l on ~ember 2-1 at end (2) must be equal and opPosite to Ha_~, the horizontal reaction at end (2) of member 2-3.
Fi~. Al2.25 and 26 show free body sketches of members 1-2 and 2-3.

M.._;II = 2xO.5(O+S;II+ .05x3) +0 Ml.-a

= Sa + 0.150
a

- - - - - - - - - - (a)

Ma _

I1._ l

=2xO.5(Z8 + .05x30) +0 =28.+0.150- - - - - - -

- - - (b)

12*/in. 1 l I I I I 1 11 \

11._" = 2xO.3S2(29. +0- .3xO.10) -400 M.-J = 3.5688. - 0.53520 - 400 - - - - (c) MJ _
a

T
20"

H _

______ _

1~ ~_a ~----r-.
3

10" .--L-

H._.

,2
240
,,-,- H 1

'
20,,-----f

=2K(28
= 2x
=

+ S. -30) +MF

,-.
(d)

M~_a

.892(0+8. -3xO.10+400

Fig. A12.26

\-.11

3-.

1. '7249 a -

0.53520 + 400 -

Fig. A12.25

In Fig . .1..12.25 take

~cnents

aJout (:) ane

;;:J_.
M J - ..

= z x .392(0-8;11 +3xO.10) -400

= -1.7848.
+

equat;e to cer-c ,
M l-a + r. a- .. - 20H a_1

0.53520 - 400

=0
~oments

- - -

(e)

. wnence,

H .\I-l

= M.

20 .<li!

+M

:1._ J I1 .._
J

=2x .892(-28. +O+3xO.:0) +400

=-3.5888" +0.53520+400

ane in Fig. A12.26, taking

about (3),

- - - (f)

!'! a-" + 11 3-a + 240 x 20 - 20 x 12 x 10 -lOH .\1-3

=0

wr.ence, :f

a-~

"+M,, 2+24:00

10

A12.12

SPECIAL

METHODS -

SLOPE DEFLECTION METHOD

?:~. Al2.27 ShC~2 t~2 ~~2~1~:r:~ :e~~:~; cement c iacr-em, ~::'rs: ::'.::"'2.','.7', tn oar-as i.'cr 2S.,:::j member of t~e fr~~e ant ther: addec :0 fCrT, :~e cO~Dosi:e f::"'2.~e. d:.a~~~. ~lJt:ec. :ir~c:ly

Shear equa t 1or- , H .\1_1.

= H 2_"

whence,

:r:

:~e

Substituting values for the er:d

Sa + 0.150 + Z8!i + 0.150 :::; 3.56882
-

Iome~ts:

stears an::'. ~~~2.l l::acs ~cllcw as ~atter of statics. ~lth t~e 2cment ~i2.~r~ Y~ovm the fr~s det1sc~sd ~~~pe ~5 ~ead~~y c3.1c:11a:ed.
:~e

2.

0.53520

A12.9 Comments on Stope-Deflection Method.

20
- 400 + 1.7848 2
-

10
0.53520 + 4,00 + 2~OO

10

Whence,
0.38528a - 0.1220 + 240

=0

The ex~~pls problems solut~ons i4c::a~e that the met.hoc c: s Iope-def Lect t cn nae t.vo advantnges , nameLy, (1) i t !'SCc:.C8S t::s number- 0:' equatior.s to )8 solved s:~~1:2~e:uslY a~G (2) it ,resents equations trzt are ea5:11 a~c rap:cly ~or.nulatec..
- - (j)

Solving equations i and

[2

for S.\I and [2,?;i vee

=2805,
= =

ea =

265.8

Substit~ting

these values in equations (a) to (h), the end moments are obtained as follows: M1.-a

Thus for structures ~lth a hi~~ ~e~ree o~ the slo~e-deflectio~ !1ethod should be considered as 80s51j1y the jest !1e~hod o~ sclution. The solvin~ of the equations in :~is method are readily programmed ~or solution jy high speed computing ~achlnery.
red~ncancYJ

A12. 10 Problems.

6.. +0.150 = 265.8+0.l5x:2505 = 666.5 Ins Ib ,

Z9 a + 0.150

MlII-1.

= 531.6 + 420.7 = 952

(1) Determ,ne the bend~"g Go~ents at suoport points A, a, C, 0, for t~e con~lnuous te~~ shown in Fi~. A12.28.
100 100

=3.5689.. -0.5352.0-400 = 3.568 x26S.S - 0.5352 x2S05 -400 =- 952 !'b-a =1.'7848" - 0.53520 + 400 =1.784 x 265.3 - 0.5352 x 2605 + 400
M lII3

~7'!..l.6"...L7" +-12"

j~~~,,20

150 lb.
I

jB

12~..a_6_.,.;C_'_; r;;,',6d

w,., 101b/in,
r ;

~D

628
+ 0.53520 -400:::; - 628

Fig. A12, 28

M::I_...

= -1.7846.11

M._.
M ..._
6

= 952
= - 952
=-686.5

(2) Same as problem (1) but consi::'.er sup~crt A as ~reelY supported instead of fixed.
(3 ) Determine the bendinz
~oment d~agram ~or

the variOUS loaded structures :n r:g.Al2.2S.

w,., 12 lb/in.
200 lb.

M:~_ ...

~8"'" L:Il 16
A I =32

C~D

I :

.:

..

I
I

36 L=30 160

~~ 12" 1 Pin,e,toJ

300 lb.

100~=30

15 i

.60

...1.,TTm

, I

m'\'l..L.-

A
1000
,

, M
' IS
I

100 lb.

1000 lb.

952
I

w,., 20 Ib/in.

~5"~5"
D

"
686.S

1 _ 2 L,., 16

1 ,., 20
Fig. A12.27
A

20"

1=3
A
~

,
-L

Fig. A12.29

,--20"---......J

CHAPTER A13

BENDING STRESSES

Al3.0 Introduction.

The bar AB in ~lg. a Is subjected to an axial compressive load P. If the compresSlve stresses are such that no buckling of the bar takes place, then bar sections such as 1-1 and 2-2 move parallel to each other as the bar shortens under the compreSSive stress.
I, ,2 I' , 2
Fig. a

Pj

""' ,''"
,

,
,

IP

Sbear~

nJL L]
---tx:+
Fig. b

applied moment the beam takes the exaggerated deflected shape as shown. The applied bending moment vector acts parallel to the Z axiS, or in other words the applied bending moment acts in a plane perpendicular to the Z axis. Consider a beam segment at length L. Fig. d shows the distortion of this segment when plane sections remain plane atter bending of the beam. It will be assumed that the beam section is homogeneous, that is, made of the same material, and that the beam stresses are below the proportional limit stress of the material or in other words that Hook's Law holds. From the geometry of similar triangles,
e _ Be or

/' .,/ ....:.\1oment Dla.

Fig. c

e = :i.. ec
c

(A)

..............

We have from Youngs Modulus E, that

E

In Fig. b the same bar Is used as a simply supported beam with two applied loads P as shown. The shear and bending moment diagrams ~or the given beam loading are also shown. The portion of the beam between sections 1-1 and 2-2 under the given loading are subjected to pure bending since the shear Is zero in this region. Experimental evidence for a beam segment ~ taken in this beam region under pure bending shows that plane sections remain plane atter bending but that the plane sections rotate with respect to each other as illustrated in Fig. c, where the dashed line represents the unstressed beam segment and the hea'sy section the shape atter pure bending takes place. Thus the top fibers are shortened and SUbjected to compressive stresses and the lower fibers are elongated and subjected to tensile stresses. Therefore at some plane n-n on the cross-sectlon, the fibers suffer no deformation and th~ have zero stress. This location of zero stress under pure bending is referred to as the neutral axis.
A13.1 Location of Neutral Axis.

stress = = unit unit strain

where cry is the bending stress under a deformation e and since it is compression it will be given a minus sign. SolVing tor cry.
cry:_E ec y

cL

The most remote fiber is at a distance y = c. Hence

= _ E ee
L

whence
cry - crc y c - - - - - - - - - - - - - - (B)

Fig. d s~ows a cantilever beam SUbjected to a pure moment at its free end, and under this

For equilibrium of the bending stress per?endicular to the beam cross-section or in the X direction, we can write Z Fx = 0, or
Z F'x
'= -

'JM Z
y

~ z-I-t.:z
Beam Section

crc c

yda

= o.

x
Fig. d

LJ

however in this expression. the term Oc 1s not c zero, hence the term yda must equal zero and this can only be true i ! the neutral axis coincides with the centroidal axis of the beam cress-section.

A13.1

A13.2

BEAM BENDING STRESSES

The neutral axis does not pass through the beam. secticn centroid when the beam is nonhomogeneous that is, the modulus of elasticity is not constant over the beam. section ~~d also when Hook's Law does not apply or where the stressstrain relationship 1s non-linear. These beam conditions are described later in this Cr~pter.
A13.2 Equations for Bending stress, Homogeneous Beams, Stresses Below Proportiooal Limit stress.

Thus, Y~ (y - x ta.~ 0) cos 0 yeas 0 - x sin 0 - - - - - Then Sq , (1) bec ames

o :::; k (y cos

(2 )

0 -

s tn 0) -

In the following derivations, it will be assumed that the plane of the external loads contain the flexural axis of the beam and hence, the beam is not subjected to torsional forces whiCh, it present, would produce bending stresses it tree warping of the beam sect~ons was restrained, as occurs at points of support. The questions of flexural axes and torsional etfects are taken up in later chapters.
m

P
Plane

of loads

This equation contains three unknowns, v, For solution, two additional equations are furnished by conditions of equilibrium namely, that the sum ot the ~oments of the external forces that lie on one side of the section ABCD about each of the rectangular axes X-X and y-y ~ust be equal and opposite, respectively, tc the sum of the moments of the internal stresses on the section about the same axes. Let 11 represent the bending ~oment in the plane of the loads; then the moment about axis X-X and Y-Y is Mx : :; M cos g and My M sin Q. The moment of the stresses on the beam section about axis X-X is / a da y. Hence, taking moments about axis x-x, we obtain for equilibrium,

k, and 0.

J1 cos

= / c da y

- / k (cos 0 y"cta - sin 0 xyda)

x-----'b-tt-.::::.:y---'- X
m

= k cos 0 / y'da - k sin 0 / xyda

In similar manner, taking moments about the Y-1 axis

(4)

Fig. A13.1

Fig. A13. 2

Fig. A13.l represents a cross-section of a straight cantilever beam with a constant crosssection, subjected to external loads which lie in a plane making an angle Q with axis 1_1 through the centroid O. To simplify the figure, the flexural axis has been assumed to coincide with the centroidal axiS, which in general is not true. Let NN represent the neutral axis under the given loading, and let 0 be the angle between the neutral axis and the axis X-X. The problem is to tind the direction of the neutral axis and the bending stress 0 at any point on the section. In the fundamental beam theory, it is assumed that the unit stress varies directly as the distance tram the neutral axiS, within the proportional limit of the ~terlal. Thus, Fig. Al3.2 illustrates how the stress varies along a line such as ~ perpendicular to the neutral axis }I-N. Let c represent unit bending stress at any point a distance Yn from the neutral axis. Then the stress 0 on da is
0=

whence

11 sin

= / 0 de x

M sin Q = - k sin

x~da +

k cos / xyda(4a)

Al3.3 Method 1. Stresses for Moments About the Principal Axes.

In equation (4), the term. / y~da is the moment of inertia of the cross-sectional area about axis X-X, which we will denote by Ix, and the term / xyda represents the product at inertia about axes X-X and Y-Y. ~e know, however, that the product or inertia with respect to the principal axes 1s zero. Therefore, if we select XX and IT in such a way as to make them coincide with the principal axes, we can write equation (4):

M cos

= k cos 0 Ix

- - - -

(5)

In like manner, from equation (4a)

I1 sin g
k

sm 0

Iy

(6 )

kYn

- - - - -

(1)

where k Is a constant. Since the position of the neutral axis Is unknown, Yn will be expressed tor convenience in terms of rectangular coordinates with respect to the axes X-X and Y-I.

To find the unit stress 0 at any point on the cross-section, we solve equation (5) for cos 0 and equation (6) for sin 0, and then substituting these values in (3), we obtain the following expressing, giVing 0 the subscript b to represent bending stress: _

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A13.3

moment of inertia about the neutral axt s, hence

b -

_ [11 cos (g-~)JYn=

(9 )

Let the resolved bending moment M cos g and M sin Q about the principal axes be given the symbols Mx p and M,f'l..' Then we can write ou
(7 )

In

In can be determined from the relationship expressed in Chapter A3, namely,

The minus Signs have been placed before each term in order to give a negative value for crb when we have a positive bending moment, or M xp is the moment of a couple acting about Ap-X p' positive when it produces compression in the upper right hand quadrant. M yp is the moment about the Yp-Y p axiS, and is also positive when it produces compression in the upper right hand quadrant.
BENDING STRESS EQUATION FOR SYMMETRICAL BEAM SECTIONS

A13.5 Method 3. Stresses from Moments, Section Properties and Distances Referred to any Pair of Rectangular Axes through the Centroid of the section.

The fiber stresseS can be found without resort to pr tnc i pa.L axes or to the neutral axis. Equation (4) can be written: M X = k cos Ix - k sin I X Y
M cos 9.

----(11)

where Ix :; / ya da and Ixy :; / xyda , and M x =

In like manner,

Since symmetrical axes are principal axes / xyda ~ 0), the bending stress equation tor bending about the symmetrical XX and YY axes 1s obviously,
(tei~

11y : - k sin ~ ly + k cos ~ I XY

- - - - - (12 I

cr
b

= _ MxY _ Myx
Ix
Iy

- - - - - - - - - - - - - (7a)

SolVing equations (11) and (12) for sin and cos 0 and substituting their values in equation (3), we obtain the following expreSSion for the fiber stress crb: (11yl x - 11Xl xy) x _ (11xl y - 11yl xy l Y _ (13) IXl y Ixl y - I~y

A13.4 Method 2. Stresses by use of Neutral Axis for Given Plane of Loading.

li:r

The direction of the neutral axis NN, measured from the XX p principal axis is given by dividing equation (61 by (51. Tan 0" = (8)

For Simplification, let

K. K. K,
: :
~

lxy/(Ixly - I\y) ly/(IXl y - I "\y) Ix/(IXl y - I\y)

SUbstituting these values in Equation (13) : crb = - (K,11y - K~Mx) x - (KaMx - Kl~Y)Y

- (141

The negative Sign arises from the :act that Q is meaSUT3d from one principal axis and ~ is measured in the same directicn :rom the other principal axis. Since equation (8) gives us the location of the neutral axis for a particular plane of loadIng,the stress at any point can be found by reSOlving the external moment into a plane perpendicular to the neut~al axis ~-N and using the

In Method 2, equation (8) was used to find the positi0n cf the r.eutral axis for a given plane of loading. ~he loca~ion o~ the neutral axis can also be r ound r e Ia t Ive to any pair of rectangular centroidal axes X and Y as follows: Since the stress at any point on the neutral axis must be zero, we can write frcm equaticn
(141 that: (K Sl '1y - K1Mx)X = - (KJ1x - KaI":ylY for all points located on the neutral ax:s. ?rom ?lg.

A13.1 tan

~ ~

y x

Thus tan 0 = _ (KJ1y - KlMX) - - - - - - - (15)

(K,I'lx - K. 11yl

A13.4

BEAM BENDING STRESSES

Al3.8 Illustrative Problems. Example Problem l.

It frequently happens that t~e plane of t~e bending moment coincides with e~ther the X-X or the Y-Y axiS, thus making either M x or ~y equal to zero. In this case, equation (15) can be simplified. For example, it My = 0
tan

o = Ixy
I,.

------(16)

and It

l1x = 0
Ix ____ - - - - - - - - - - - ._I xy
(17)

tan 0

A13.6 Advantages and Disadvantages of the Three Methods.

Method 2 (bending about the neutral axis tor a given plane of loading) no doubt gives a better picture of the true action of the beam relative to its bending as a Whole. The point at maximum fiber stress is easily determined by placing a scale perpendicular to the neutral axis and moving it along the neutral axis to tind the point on the beam section farthese away tram the neutral axis. In airplane design, there are many design conditions, which cr~ge the direction of the plane of loading, thus, several neutral axes must be computed for each beam eectlon, which is a disadvantage as compared to the other two methods. In determining the shears and moments on airplane structures, it is common practice to resolve air and landing forces parallel to the airplane XYZ axes and these results can be used directly in method 3, whereas method 1 rsquires a further resolution with respect to the principal axes. Methods 1 and 3 are more Widely used than method 2. Since bending moments about one prinCipal axis produces no bending about the other principal axiS, the principal axes are convenient axes to use when calculating internal shear flow distribution.
A13.7 De!lec:tions.

Fig. A13.3 shows a unsvnme t r i ca I one P y " 6000# J~~~ f, cell box beam wi th i ;1 1 . four corner r lange / ,/ IPx:1600~i member's a, o , c and d. I ,1/ ' Let it be required to ~' I I 1/ I det ermme the bending S? 'I ii 'I I, I - - - - - - - o - t i ! axial stress in the N ,1/ 1/ four corner members / II II // " due to the loads ?x 1/ II and Py acting 50" /.10",' ,;/ , " II from the section a .5" ," ! I abed. ,yf'l In this exa.. n ple bI IT : JI I 12" solution the s nee t I' at! ccnnec t.i ng the corner d ( -LX' members w::.ll be con 80' ,#1/.'-------,-16"'''' --"-:I:, .40" atder-ed inarrect ive in' ' bending. The stresses Fig. Al3. 3 will be determined by each ot the three methods as presented in this chapter.
I /

J'

il/'

1/1

xtL

SOLUTION

The first step common to all three meth~ds is the calculation of the moments of inertia about the centroidal X and Y axes. Table A13.1 gives the detailed calculations. The properties are first calculated about the reference axes x'x r and ylyt and then transferred to the parallel centroidal axes.
TABLE A13.l
Area nA"
1.0 0.5 0.8 0.4

Mem.
a
b

y'
12 8 0 0

x'
-16
0

Ay' 12
4 0 0

Ax' -16
00 -12.8 0 .28.8

Ay,2
144.0 32.0 0 0 176.0

Ax,2
256.0 0 204.8 0 ! 460.8

I Ax'y' I
-192 0 0 0 -192

c
d

-16
0

Totallll I 2.7

16

Calculation of Centroid of Section: x =

ZAx' lA
28.8 = -2:7
~

The deflection can be :ound by using the beam section properties about the neutral axis tor the given plane of loading and the bending moment resolved in a plane no~l to the neutral axis. The deflection can also be :ound by resolving the bending ~oment into the two prinCipal planes and then using the properties about the principal axes. The resultant deflection is the vector sum at the deflections in the direction ot the principal planes.

- 10.567"
= 81.18
= 153.58
- 21.33

Y=

lAy' 12..- = 5.926'" lA = 2,7

Ix ::: 176

2.7 x 5.926~

I y = 460.8 - 2.7 x
I XY

10.667~

= - 192 - (2.7 x - 10.667 x 5.926) =

\,,:~'.-->,~.

_;_-,

-/::.--~_~)_:34 ~~~::"-~-""-"_!'><:';~~::i.~_':',,~" ~"~::",";:-"1~',;-,; :'.,~,: ,~.: c~. :.._I;,~.,,;=.--c: ~"'":: :)r"'~~ .:':~.~-J'~.~T;.-:"/:~~::~'1:.'~'i:~i,.'!._

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A13.5

Solution by Method 1 (bending about principal axes)

The angle 0 between the x axis and the principal axes is given by the equation,
2 I xy 2 (-21.331 - 42.66 tan 2 0 = ,,------'''i_ = 153.58 - 81.18 72.40 I y - Ix

right band quadrant I.

(Positive moments produce compression in upper

Calculation at Stresses

=
Stringer a

"D - - - . - - "lCP

Mxp Yp

Myp

ryp
= 4.45

xp

= - .589
2 .0 = 30 - 30' ,

0_15=15'

Xp

=-

6.74,

Yp

I x p = Ix

cos~

Iy

5in~

.0 - 2 I xy sin
(-

cos 0

hence
b = -

= 81.18 x .9648 a + 153.58 x .2636 a - .2636

21.33

268700 x 4.4.5 75.38 15900 - 6600

- 156200 (- 6.74) 159.34 22500i/ln

x .9648)2
in.~

= 75.56

10.67 - 10.85 = 75.38

Stringer b
Xp . 9 . 75 W , yp.4.82

I y p = Ix sin a .0 + I y cos a 0 + 2 I xy sin 0 cos .0

81.18 x .2636 a + 153.58 x .9647 a + 10.85 -

268700 x 4.82 75.38

; - 17180 + 9570

- 156200 x 9.75

159.34 7610i/ln

159.34 In. ~

I'
a
I

IYp
Fig. AI3.3.
9.75,,--..j

Xp

! f-6.7l_~ -..i.,. __
"10'

I 1
I

Stringer c
Xp O'b
'II 'II -

3.58,

Yp

'II

7.12 159.34

/
b

- 268700 x - 7.12
75.38

- 156200 (- 3.58)

X - .",... -

::

c
3

.58

Ypt-

-".1,-= ~ \- - - - X II ''iEl5'-t '.82 f ! 200 " 5.926 I ~ d .;--1"" I I 11.80 ~

y
I

L'D.667"

25400 - 3520 21880

and similarly for stringer d, 0b

= 21900.

y'

Solution by MethOd 2 (Neutral Axis Method)

Let g' angle between yy axis and plane

Fig. A13.3a shows the location of the principal axes, as well as the perpendluclar distances from each corner member to each at the principal axes. These distances are readily determined from elementary trigonometry; The external bending moment about the X and Y axes equal J
~~

at loading.
tan hence
Q~
Q'

=80000 = 300000

.2667
+

= 14

- 56' and Q = 14 - 56

15 - 15' ,.

= 6000 x 50 :

300,000w~ 80,OOOw~

My

= - 1600 x 50 = -

Let a = angle the X p axlS~ tan a IXp tan Q

~'--,~-

be~Neen

neutral axis NN and

These ~oments will be resolved into bending moments about the x p and yp principal axes.

Mx P 300,000 x
i1y

- 75.38 (- 0.5816) - .275 159.34

cos 15 - IS' - 80000 x hence a = 15 - 24' (see Fig. A13.3b). Since the angle be~Neen the X axis and the neutral axis is only 9' , we can say IN = Ix = 81.18. ResolVing the external bending moments normal to neutral axis, we obtain

sin ISO - IS r = 289700 - 21000 = 268700 w

p

=-

300000 x sin 15 - 15'

80000

cos ISO

15' = - 79000 - 77200 = _ 156200 wi

A13.6
~

BEAM BENDffiG STRESSES

= 300000 X .9999 + 80000 x sin Oe - g' = 300200"#. The bending stress ~t any point is given by,

K. = Iy/(Ixl y - I' xy) 153.58/12016 a K~ - Ix/(IxI y - I X Y) - 81.18/12016

= ,012794
::z

.00674

Stringer a
Y n

Stringer a

= 6.074

5.33 x .0025

= 6.087"
5I

x=-5.333",

y
+

= 6.074"

300 200 x 6.087 81.15

_ 22500i/ln.

0b - - [.00674 (- 80000)

.00177 x 300000J x -

Stringer b
Y n = 2.055"
b
~

[.012794 x 300000 - .00177 (- 80000)J y hence 0b = (8) x - (3697)y;

:II _

- 300 200 x 2.05 81.18

7570

hence crb = 8 x - 5.33 - 3697 x 6.074 = - 22450 Stringer b

Stringer c
Yn = - 5.92

x = 10.667,

Y = 2.074

Ob

=-

300 200 (- 5.92) = 21850 81.18

0b = 8 x 10.657 - (3697) 2.074 = 85 -7660 = -7575

Stringer c
x .. - 5.333,

Stringer d
Y = - 5.95
b =

y=-S.926

5.333 - (3697 x - 5.926)

42 + 21900 - 21862

300 200 (- 5.95) = 22000 81.18

Stringer d
y Q' : 14581 ~::z 15015' Q:30011' plane of .I:0ading 9' 52' N -1515'

10.667,

=-

5.926

0b = 8 x 10.667 - (3697 X - 5.926) =

~

85 + 21900 = 21985

"p

Y p

NOTE: The stresses Ob by the three methods were calculated by 10" slide rUle. ~ence the small discrepancy betNeen the results for tte three methods.

Fig. A13.3b

Solution by Method 3 (Method Using Properties About X and Y Axes)

and My.

Error in Stresses Due to Assumption that Section Bends About X and Y Centroidal Axes Due to M x
(Equation 7a)

Mx

=300000,

My = - 80000
I y = 153.58,

Ix .. 81.18,

Stringer a y

IXY = - 21.33

= 6.074,

x '; - 5.33
- 80000 x - 5.33 l53.S8 = - 25190

= 81.18 x 153.58 _ 21.33 ll = b = - 21.33 12016 = - .00177

- 21.33

300000 x 6.074
81.18

Stringer b
y
z

2.07,

= 10.567

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A13.7

Stringer b - contJd.
300000 x 2.07 8l.18 - 80000 x 10.667 153.58

Mx 14260

- 2110

x 50

713.000'~

I'ly = - 760 x 50 - 38000'#

In like

mar~er

for stringers c and d and O'b

These moments are resolved into bending moments about the prinCipal axes, as tallows: -

O'b

= 19130

= 27440

M x = 713000 x cos 0 - 38000 x sin

P p

;1 =

700,OOO~~

Comparing these results with the previous results it is noticed that considerable error exists. Under these eroneous stresses the internal resisting ~oment does not equal the external bending moments about the X and Y axis.
Example Pr.ablem 2.

I'lY

=-

71300 x sin

0 - 38000 x cos 0 '-140000'#

From equation (7), the general formula for 0b is:

Fig. A13.4 2-cell stressed ample, the beam and the section

shows a portion ot a cantilever skin wing box beam. In this exsection is considered constant, is identical to that used in
14260*
I I I 1 , I I I I
1

Stress on Stringer 1: yp = 1.85 11 and Xp = - 17.86 11 (scaled from fUll size drawing).
(700000 x l. 85) - 140000

I:

11

I
I I
I!

II
I:

11761l1]1
II

I I I 1
I I I I

II 'I 11 II II

' I

I!
I;

i I'
I:

II
II

I I I I
I

I I I
I I

I I I

I '"\ II 11 ,1 [I

,I 11 " 11

I
50"

x (- 17.86)
437 in.

181.17
7150 - 5700

12850~/sq.

II
II

I'
II

II II
II

',J..I II III r- 4-, ! III , I1 II 111 III: II I'I[ II

I i 11 II II I I ~ I 1I II II -+-r -J. 1- Ir+- --+ i-I -r-t r- T1 II I I : I l i t I ,I II I I 3 4

II! I I1 I

1r..-l-

"1

Stress on Stringer 9:
Yp = 9.04",

xp

14.24"
140000 x 14.24 437

"b

--

(700000 x ,).04)

181.1/
+

;11

= - 34,900

4560 ::. - 30340t/sq. in.

X-.-_I
B

Stress on Stringer 12:

Fig. A13.4

700000 x (- 6.80) 181.17

- 140000 x (- 8.22; = 437

26200 + 2630 - + 23570i/sq. in.

?lg. A3.13 of Chapter A3 tor which the section
SQlytion by Method. 2 (Neutral Axis Method)

properties were computed and are as follows: Ix = 186.6 In.

;1 =

I y = 431. 7 tn ."
Ix
p

= 36.41 in ."
437

In fig. Al3.5 let Q' be the angle between the Y-Y axis and the Plane of the resultant bending moment. Resultant bending mement, M v713000 3
tan g' - 38000 713000

eO - 16.25"

18l.2

The resultant air load on the wing outboard of section ABC is 14260~ acting up in the Y direction and 760i acting tcrNard in the X direction, and the location of these resultant loads is 50 from section ABC (Fig. A13.4). ;he bending stress intensity at the centroids or stringers number (1), (9) end (12) will be calculated uSing all t~llee methods.
Solution by Method 1 (Bending about Principal Axes)

.-

38000 3

714600" lb.

.0533, hence 9'

=-

3 _ 3'

Let Q equal the angle between the plane of the resultant moment and the Yp axis. Then a = 9' + 0 = 3 3'
+

8 - 16'

11<;1 - 19'

From equation (8), the angle be~Neen the Xp axis and the neutral axis ~ Q, and -tan a I xp tan 9
I yp

The bending moment at section ABC about the X and Y axes is,

181.17 x (- .200) 437

Al3.S

BEAM BENDING STRESSES

Plane of

~',. f Resultant gl-!

I "
\ \
\ I '

Y YO

- 36.41 I;;JXY) = 186.46 x 431.7 - 36.41;;J

Moment-

= - .00046
36.41 ; J

431.7 793000

.C0545

= 186.46
Stress on Stringer 1:
Fig. Al3.5

793000

.0'02355

Y1 = 4.39",
~

Xl

= - 1""'.41"

.083.

Hence, a :;; 4 G

45'

Fig. Al3.5 shows the relative positions of the neutral axiS, principal axes, and plane of loading. The component of the external resultant ~~ment about the neutral axis N equals: -

crb, = [.002355

X (- 38000) -

(- .00C46 x (13000)J

(-17.41 )

MN =

714060 x sin 83 G
cos~

26' = 709350 in. lb.

a
01

- [.00545 x 713000 - (- .00046 x - 3EOOO)J 4.39

(238.5)(- 17.41) - (3868) 4.39

IN = I x p

a +

I yp sin a
+ 437

= 191.17 x .9966 a
Yn~

x .0878

183.37 tn ."

4150 - 17000 = - 12850#/in.;;J

Stress on Stringer 1:

Stress on
+

Str1~ger

9:

= (Yp

x p tan a) cos a = (1.85

.083) .9966 = 3.32'

17.86 x

Yg = 6.S9", Xg = l5.39"

0b~ = - [238.5J 15.39 = [3868J 6.39 :;;-30320;/in .Ii

In.

crb

,=

n Yn = M - --In

709350 x 3.32
183.37

=_ 12850*/sq.

Stress on Stringer 12:

Y1a

Stress on Stringer 9:
Y n

= - 5.55",
(- 9.11) - [3868J (- 5.55)

= (9.04 - 14.24 x .083) .9966 = 7.84'

crb ..

=- [238.5J

23620i/ln.

Stress on Stringer 12: Yn1.ll = (6.80

0b 1a
+

NOTE:

8.22 x .083) .9966 = 6.10

= - 709~~.~7- 610 = 23600t/sq. in.

Solution by Method 3

In the three solutions, the distance from the axis in question to the stringers 1, 9, and 12 have been taken to the centroid of each stringer unit. Thus, the stresses obtained are average axial stresses on the stringers. If the ~imum stress is desired, the arm should refer to the most remote part of the stringer or the skin surface.
JWproximate ...Method.

Mx
Ix = 186.46

= 713000"1,

My =

38000'~

I y = 431. 7
K~,

Ixy = - 36.41

The constants K1 J tatnec -

and Ka are first deter-

It is sometimes erroneously assumed that the external bending moments M x and ~y produce bending about the X and Y axes as though they were neutral axes. To show the error of this assumption, the stresses will be computed for stringers 1 and 9.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

ab

A13.9

Stringer 1:
Yl = 4 + .39 = 4.39
Xl

r-TB

-TAr
2.64

.I
3"

= - 33.15

15.74 - - 17.41

-----, -L
2" ...<--cB2;64 CA'

--t
l

a b = - MxY

1 3"

Ix

ab

=-

713000 x 4.39 186.46

1- 38000 x - 17.41) =
431.7
ah

Fig. A13.8

Fig. A13.9

stress intensity at bottom edge of portion A is

- 18330#:!i n :l .

12 . 25 / 3 ) = 0.75
TA

Ob.

Stringer 9:

= CAr = Average

stress times area

Y.
ah

6.89,

Xg

= 15.39
431. 7

= - 713000 x 6.89
186.46

(- 38000 x 15.39)

- 25000#! l n :l .
Example Problem 3.

The distance from the neutral axis to the centroid of the trapizoidal stress loading on portion A is 2.64 inches. In like manner,
TB

The previous example problems were solved by substituting in the bending stress ~quatlons. The student should solve bending strs's problems by equating the internal r-es ts t Lig moment at a beam section to the external bel.ding moment at the same section. To 111ustrlte Fig. A13.6 shows a simply supported 10adfd beam. The shear and bending moment diagram for the given beam loading is also shown. Fig. A13.7 shews the beam section which is constant along the span.
500#

CB'

= (0 +2 Ob)

x 0.25 = 0.375Ob

The arm to TB is 0.667 x 3 = 2 inches. For equilibrium of the beam free body, take moments about point (0) and equate to zero.

ZMo

= - 500 x 48

2.54 x TA
2

2.64 CA'

2 TB

eB '

1000

.61
A
690

300

1'-2" -1

SUbstituting values as found for TA, T3 etc,

48" +-1110

1:1'

120"--r36'~

T
8"

- 500 x 48
+

2.64 x 1.15 O'b + 2.64 x 1.15 O'b

2 x 0.375 crb + 2 x 0.375 crb = a O'b

5~t---300

610';-_--,

1
Fig. A13.7

hence

= 24000 7.58 = 3170

pe f ,

Solution using Bending Stress Formula.

b

e -I= 11

where 11
e
I

- 24000 = 3" = Moment inertia = Neutral of axis.

about

Calculat10n of r , For Portions A and A'

:or 'lie b Portions b, h",

1 I - 12 x 0.25 x 6' =

4.50 in"

ITotal = 22.76 in" .

A13 10

BEAM BENDING STRESSES

henc e

which checks first

Example Problem 4.

salut~on.

Fig. A13.10 shows a loaded beam and Fig. Al3.1l shows the cross-section of the bearn at section a-a'. aet ormme the magnitude of the maxl~um bending stress at secti1n a-a' ~~der the given beam load~~g. The beam section is unsym~etrical about the ~orizontal centroidal axis. Simple calculations locate the neutral axis as shown in Fig. A13.ll.

ternal force action from this nethod of solution. In solv:ng nonhomogeneous beams a~d beams stressed above the elastic l~mit stress, this ~ethod of solution often proves necessary or advantageous because no simple beam bending stress for.nula can be derived. In Fig. Al3.1Z, let oQ be the intehsity on the most remote fiber, or 3.09" above the neutral axis. Table A Shows the calculations of the total stress on each of tte portions of the cross-section and their noment about the neutra: aXiS, all in te~ of the ~own stress, cr~.
TABLE A
I

2
Area
A

3
Average Stress
in

12*/in.

Portion

' 11' I I II I i~l~

100"---1. 30"-;:-'00"

Total Load

5#/in.

See
Fig.

y : arm to N. A.

Resisting moment
(4) x (5)

(2)

I---

Fig. A13. 10

RA

He

-th-2"

-4-

---I

-11-rI

1/2 _-<= __

Jl r -li I
I

A13.13
I

terms of c b
1.045 1. 00 1.00 1. 09 0.56 -.337crb -.757 -.837 .350 .578 . 821crb

x (3)

-.353crb

1. 39
2.35 2.62

- .490 O'b
~1.

2
3

-.757
.837 .387 .323 1. 2370b
O. OOOOb

780

I
1-27.2 in. 4 3t

'I

N- -I-f- N

3~'6"
1

- 3'
4

-2.190 -0.557

-1. 44
.1. 93
-2.57

- 5'
6

I -0.622
-3.180 Cb -8.82 O"b

i I ,
I

Fi. AI3.11

11 44 t=-3

-.--

(~t~I 2.9;..

1. 50
6.19

Totals

- , - I-- 2" -1
?ig. Al3.l2 shows a tree body of the portion af the beam to the ri~~t of section a-a. The bending stress intensity diagram is shown by the horizontal arrows acting on the beam face at section a-a. Fig. A13.l3 shows the cross-section of the beam at sectien a-a. The general procedure will be to determine the total bending stress load on each portion, 1 to 6, of the cross-section and then the moment of each of these loads about the neutral axiS, the summation of which must equal the external bending moment.

Explanation of Table A. Column 3 gives the average stress en each of the 6 portions. For example, the stress on portion (1) varies from zero at t he neutral 2.09 axis to 3.09 crb = .574 crb at the upper edge of (1). ThUS, tne average stress _ .674 ab + 0 = .337 abo On part 3 or 3' J the stress on the lower edge .674 crb and a b on the upper edge. Then, the average stress = (1 + 2. 674 ) crb = .337 abo Col~ 5 1s the distance from the neutral axis to the centroid of the load on each of the blockS. For portion (1) the stress pattern 1s triangular, and y = 2/3 x 2.09 = 1.39". On portions (2), (3), (5) and (6) the bending st~ess dlstr1b~tion 1s trapezoidal, and the arm 1s to the centroid of this trapeZOid. The total internal res~sting ~oment of - 8.82 at ~rom Table A equals the external bending moment of 36670"1. ThUS, - 36670 Cib = 8.82 = - 4l60#/in. a. For equil18rium, the total co~pressive stresses on the cross-section of the beam ~ust be equal to the total t ens i Le stresses, or l.H must equal zero. Col~ 4 of Table A gives the
2

\
_N~
2 91"
_

Fig.AI3.13

I ,

a
Fig. A13.12

This method of solution involves ~ore calculations than that required in SUbstituting in the bending stress fo~ula, however, the student should obtain a better understanding of the in-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A13.11

total load on each portion of the cross-section and the total of this column is zero. The uending stress on the lower fiber of the cross-section 1s directly proportional to the distance from the neutral axis or, ab1 ower =

R ,:. r
T~
-- 21"T

--..j

2"1-

4.1
3

Maple

3.33"
I

SOLUTION USING BENDING STRESS FOR:-mA.

I NA

= 27.2 in."

hence
- 36670 x 3.09 crb (upper fiber) = 27.2 = - 4160 psi.

1
Fig. Al3.14

! ]" I

L I1j

rr I
1"

11.,l-NIA.
2.73"

I ~teel
-

16
22.3 x 1.5 = 33.4"
Fig. Al3. 16 Transformed Section into Equivalent Spruce

Fig. A13. 15

which checks the above solution.

A13.9 Bending stresses in Beams with Non-Homogeneous Sections, Stresses within the Elastic Ranges.

ant materials is:

Espruce = 1,300,000 psi.

E"ap1e

In general, beams are usually made of one material, but special- cases often arise Where two different materials are used to form a beam section. For example, a commerical airplane in its lifetime often undergoes a number of changes or modifications such as the addition of additional installations, fixed equipment, larger engines, etc. Th1s increase in airplane weight increases the structural stresses and it often becomes necessary to strengthen various structural ~embers at the critical stress points. Since space limitations are usually critical, it often necessitates that the reinforcing ~terlal be stt:fer than ~he original material. Years ago, when spruce wood was a common ~terial for wing beams, it was normal practice to use reln~orcements at critical stress points of stifter wood ~terial such as maple in order to cut dawn the size of the reinforcements. In aluminum alloy beams, the use of heat-treated alloy steel reinforcements are otten used because steel is 29000,000/10500,000 or 2.76 times stiffer than alu:ninum alloy.
SOLUTION BY
~

= 1,600,000 psi. = 29,000,000 psi.

Estee1
SOLUTION:

The first step in the solution is to transform the reinforced beam section of Fig. A13.l5 into an equivalent beam sectIon composed of the same material throughout. This 1s possible, because the modulus of elasticity of each material gives us the measure of stiffness for that material. In this solution the reinforced beam will be transformed into a spruce beam section as illustrated in Fig. Al3.l6. Emp1e Iepruc e

= 1:... . 6~0~0,:..~0~0~0 = 1. 23
0

1,300,000

Esteel = 29,000,000 = 22.3 1,300,000 Espruce Thus to transform the maple reinforcing strips in Fig. A13.l5 into spruce we increase the width of each strip to 1.23 x 0.5 = 0.615 inches as shown in Fig. A13.16. Likewise; to trans:O~ ehe steel :ein!orcing strip into spruce, we ~ke the Width equal to 22.3 x 1.5 = 33.4 inches as shown in Fig. A13.16. This transforned equivalent section 1s now handled like any homogeneous beam section which is stressed within the elastic :lm1t of the ~terials. ~he usual calculation Nould locate the neutral axis as shown ar.d the ~oment of inertia of the transformed section about ~he neutral axis would give a value of 51.30 in.~. Bending stress at upper edge of beam section: - 60000 x 3.33 51.30

OF TRANSFOFJ1ED

SEAM SECTION.

To illustrate how the stresses in a composite beam can be ~ete~ined, ~NO example problems will be presented.
Example Problem 5.

Fig. Al3.1~ shows the original beam section which is entirely spruce wood. ?ig. A13.l5 shows the reinforced or ~odlfied beam section. ~NO ears of maple wood have been added to :he upper part as shown and a steel strap h~? been added to the bottom face of the beam. The problem is to find the stress at the top and bottom points on the beam section when the ~eam section is subjected to an external Jend~ng ~oment of 60000 in. lbs. about a horizontal axis which causes compression in the upper beam portion. The modulus of elast:city (E) for the 3 dif~er-

=-

3900 psi.

Referring to Fig. A13.15, this stresS would be

A13.I2

BEAM BENDING STRESSES

the stress in the spruce sect:on. Since the ~e intorcing strips are maple, the stress at the top edge of these maple strips would be 1.23 times (- 3900) = - 4880 pSi. The bending stress at the lower edge of the transformed beam section of Fig. A13.l6 would be:
_ bt

alloy. ~he original area of 0.1 sq. in. 9ach have been ~ultiplied by these st1f~ness ratio values.

60000 x (- 2.73) -_ 3200 i 51.30 ' ps.

4.65

The stress in the steel reinforcing stra? thus equals 22.3 times 3200 = 71500 psi. Since all these stresses are below the elastic limit stress for the 3 materials the beam bending stress formula as used is
app.l Icabj e ,
Example Problem 6.

+-

y = 5.35"

Fig. A13. 18 Transformed Beam Section into Magnesium Alloy.

Fig. A13.17 shows an unsymmetrical beam section composed of tour stringers, a,b,c and d ot equal area each and connected by a thin web. The web will be neglected in this example problem. Each ot the stringers is made from different material as indicated on Fig. A13.l7. The beam section is subjected to the bending moment Mx and My as indicated. Let it be required to determine the stress and total load on each stringer in reSisting these applied external bending moments. I<---- 6" - - j

The solution for the beam section of rig. A13.18 1s the same as for any other unsymmetrical homogeneous beam section. The first step is to locate the centroid of this section and determine the moments of inertia of this section about centroidal X and Y axes.
y

Z Ay ZA =

.446 x 10 + 0.1615 x 10 = 5.35" 1.1365

r
10"

ZAX = ZA = 0.1615

x 6 + .431 x 4 1.1385
+ 0.531 x +

= 2.365" = 28.27
in.

steel 10

.10

ra

"",':rues.
Steel

,1
My'" 5000".

Ix 0.6075 x 4.65 I y 0.546 x 2.365

5:35

0.165 x 3.635 + 0.431 x

1
Emg.
Esteel

x--L.
Fig. A13.17

1.635
I xy

M:x:'" 10000"*

= 6.34

in.
- - 4.90 = 2.73 = 1.27 - - 3.77
<6

= 0.446 (- 2.365) (4.65)

0.1615 x 3.635 x 4.65 0.10 (- 2.365)(- 5.35) 0.431 x 1.635 (- 5.35)

MagneSium Area of Each stringer :II) sq. in. c .1 Alloy .10 Alum.Alloy 4 " ---j

TOTAL - - 4.67 in.

sowrrON;

The bending stresses will be calculated by USing method 3 at Art. A13.S.

Since the 4 stringers are made of different materials we will tranSform all the materials into an equivalent beam sect ron with all 4 stringers being magnesium alloy.

From Equation (14) Art. A13.5 --

= 6,500,000 =29,000,000
= 10,500,000

Esteel

= 29 = 4.46
6.5 25 = 6.5

- 4.67 28.27 x 6.34 - 4.67* = -1-~' 6 = - 0.0296 0' 47

Iy

Esta1n.steel= 28,000,000 Salum.allOY

~.s.

Emg.
E",ag.

= 4.31
Ixl y -

~1\..ill1.

= 10.5 = 1.615 6.5

Using the ratio of stiffness values as indicated above gives the transformed beam section of F1g. A13.18 where all ~ter1al 1s now magnesi'~

Ix Ixl y 10,000

= 28.27 = 0 1797 157.4

in. lb.

My

5000 in. lb.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A13.l3

Consider Stringer (c): -

to see if they equal Zero. Y

- 5.35

2.365,

p := 6

1167

162

1010

= 1 check.

Substituting in ab equation
Gc

=- [0.1797

x 5000

- (- 0.0296)(- 100 00 )Jx

- [(0.0403)(- 100,00 ) - (- .0296 (898

a c - 602

x 5000

)J

Forthermore the moments of the stringer loads about the X and Y axes Toust equal the applied external bending moments. Take moments about a Y axis through stringers (a) and (cl.
6

x 162

- 4 x 1010

... 5000

296 ) x - [(- 403 ) + 148 ] Y

x + 255" y
x

=(e) and (0),

12 In. lb.

- - (18)

Take moments about X axis through stringers

For stringer (c) hence

Gc

=- 2.365,
+

=- 5.3S

Z Mx = 1162 x 10 - 162 x 10 - 10000

=Oln.lb. (The calculations in this example being done on a slide rule can not provide exact checks)
A13.10 Bending Stresses of Homogeneous Beams stressed above the Elastic Limit Stress Range.

= - 602 = 1423

(- 2.365) 1360

255

(- 5.35)
ps t .

= 63

The total toad Pc in stringer (c) thus equals 0c (Area) = 63 x 0.1 = 6 Ibs. tension. Stringir (a):

In structural airplane design, the applied loads on the airplane must be taken by the structure without suffering permanent strain x = - 2.365, y = 4.65 which means the stresses should fall within the elastic range. The airplane structural deSign tram equation (18) loads which in general equal the applied loads times a factor of safety of 1.5 must be taken 0a 602 (- 2.365) + 255 x 4.65 = 2613 pSi. by the structure without collapse or rupture with no restriction on permanent strain. Many Pa = 2613 x 0.446 - 1167. lb. tension. airplane structural beams will not fail until the stresses are considerably above the elastiC Since the true area of this stringer is 0.1 stress range for the beam material. square the stress in this steel str~nger equals Since the stres~-strain relationship in 1167 /0.1 = l1670 psi. tension. the inelastic range 1s not linear and also since the stress-strain curve for a material in Stringer (o ) : the inelastic range is not the same under tensIle and compressive stresses (See Fig. A13.19), x = 3.635, y = 4.65 the beam bending stress for~ulas as previously derived do not apply since they were based on a ab - 602 x 3.635 + 255 x 4.65 = - 1006 PSi. linear variation of stresS to strain. Experi~ental tests hcwever. have shown that even when Pb = - 1006 x 0.1615 = - 162 1bs. stressed in the inelastic r~~ge, that plane sections before bending remain plain after True stress in stainless steel stringer after bending, thus strain deformation is still = - 162 /0.1 = - 1628 pSi. linear which fact simplifies the ~roblem si~ce the stress corresponding to a given strain can Stringer (d): be found from a st~ess-strai~ curve for the beam material. x = 1.635, Y = - 5.35 A general method of approach to solving beams trat are stressed above the elastic range OJ 602 x 1.635 + 255 t- 5.35) = - 2345 psi. can best be explained by the solution of a problem. x 0.431 1010 Ibs.

=-"

Example Problem 7.

- 1010 True stresS = 0.1

= - 10100 PSi.
str:~ger lo~ds

To check the results, check total

?ortlon (al of Fig. A13.20 shows a solid round bar made rr-on, 24ST aluminum alloy material. Fig. A13.l9 shows a stress-strain curve for this mater~al. Let it be assumed

A13.14

BEAM BENDrNG STRESSES

that the ~imu~ failing ccmpressive stress occurs at a s~raln of 0.01 in. per inch. The problem 1s to deterwlne the ultimate reslstln~ moment developed by this round bar and then compare the result with that obtained by uSin~ the beam bending stress for~ula based en linear variation o~ stress to strain. ,--

!.l .r ~11_~.,; . _.. _.~.! -J-..-"'--.i,.-,- ;l.30:~ "'i''j -., . :'...:. ig

'

c,~ t' ',ri.'it '..,.'... ! I

IrLT:~ i',,
.. ! I
IX wIO ,,,

i .1 ---;._/ . . ~i;!!l.:..,..~._ _.T---;---

'i /
I I

1
I

A13.19 shows that :or a given strai~ ~~ tr.e ~n elastic range the resulting tensile stress 1s ~1gher than the ~esultlng compressive stress, and :urthe~ore from internal eqUilibrium the total tensile stress on the cross-section must equal the total compressive stress. The problem as stated assumed that a compressive unit strain of 0.01 caused failure. Fig. b thus shows the strain picture on the beam just before failure since plane sections remain plane after bending in the ~nelastic range. Table A13.2 gives the detailed calculations for determining the internal reslstl~g moment developed under the given strain condition.
TABLE A13.2

1
Strip
No.

2
Strip Area "A"

3
y

,
a

'! 1/.: ...

~ :;

~_ . ,.

., ...._F'iu-. Al.tn

e
00867 .00773 .00685 .00591 .00494 .00398 .00302 . 00205 .00108 .00012 -.00084 -.00181 -.00276 -.00374 -.00470 -.00566 -.00663 00759 -.00846 -.00937

Unit stress IF =

oA

Res. Moment 1M = Fr 2760 4300 5040 4820 4310 3200 1920

1 2
3

s
6

.058 .102 .135 .153 .165 .180

(Q)
TCHSIO ...

Fig. A13.20
Te1oI$IL

5Tltasa

..
(b \C.,,,.
IJ~JIf""

.195 .197 .200 11 .200 12 .197 13 .195 14 .185 15 .180 16 .165 17 .153 .135 16 19 .102 20 .058 Total 3.140 Col. Col. Col. Col. Col. Col.

7 8 9 10

.185

0.935 0.840 0.75 0.65 0.55 0.45 0.35 0.25 0.i5 0.05 -0.05 -0.15 -0.25 -0.35

-0.45
-0.55 -0.65 -0.75 -0.84 -0.935

53000 52500 52100 51500 51000 43000 33200 22800 12500 3200 - 7250 _17800 -29500 -35500 -40000 -43000 -44800 -46000 -47200 -48000

3075 5350 7025 7870 8410 7740 6140 4450 2460

640
-1450 _3510 -5750 -6560 _7200

940 280 10 130 660

1650 2540 3510 4170 4710 4880 4210 2690 56735

.:~~~g
I
I
-6210 -4810 -2780

740

~.

(e)

5 Unit stress for e: strain from Fig. A13.19 6 Total stress on strip. 7 Moment about neutral axis. r=(y-.0375J

1 Rod diVided. into 20 strips . 1" thick:. 3 Y = distance from centerline to strip c. g. 4 e = strain at midpoint", (y - .0375)/103.75

co......,.,,
Sf.au

SOLUTION: Since the stress-atrain diagram in tension is different from that in compression (See Fig. A13.19) the neutral axis will not coincide with the centroidal axis ot the round bar rsgardless of the fact that it 1s a symmetrical shape. Thus. the method of solution is a trial and error one since the location of the neutral axis can not be solved for directly. In Fig. b of Fig. A13.20 the neutral axis has been assumed 0.0375 inches above the centerline axis of the bar. It Nas assumed toward the tension side because observation of the stress-strain curve of Fig.

The summation of column (6) should be Since ~he discrepancy is 740 lbs.~ it ~eans that the assumed POSition for the neutral axis is a little too high, however the discrepancy is ~egliglble. The total internal r es ts t tng mcment is 56735 in. lbs. (Col. 7). It we take a maximum unit c cmpress tve strain of 0.01 we find the corresponding st~ess from F'ig.A13.l9 to be 48500 ps t I f this stress is used as the failing stress in the bea~ ~ormula M - cr I we obtain, zero.
c
M

= 48500 x 0.785 = 38000 In. lbs.

(0.785
I =c of round bar =
nT~/4)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A13.15

Thus using the ~ame failing stress at :he far extreme fiber the beam for~ula based :n linear stress-strain relct~onship gives an ultimate bending strength of 38000 as compared to a true strength of 56735 or only 67 percent as ::J.llch. Fig. c of Pig. Al3.l0 shows the true stress :istribution on the cross-section, which explains why the reSisting ~oment 1s higher than Nhen a triangular distribution is used. The problem ot the ultimata bending strength of struct~al shapes is discussed in ietail in Volk~e II. A13.11
Curved Range.

given in most books on advanced engineering mechanics and will not be repeated ~ere. It is convenient however to express the influence of the beam curvature in the for.n of a correction factor K by which the stresses obtained by the beam formula for straight beams can be multiplied to obtain the tr~e stresses for the curved beaIT~. Thus for a curved beam the maximum stress can be calculated from the equation
cr = K M c
I

- - - -

(19)

Beams. Stresses

Within

the Elastic

The equations derived in the preVious articles of this cr~pter Wer9 fer beams that were straight. Thus in ?ig. ~13.2l, the element or length (L) used in the derivation was constant over the depth of the beam. The strain (6L/L) was therefore directly ~roportior~l to 6L which had a linear variation. In a curved beam, the assumption that plane sections remain plane after bending still applies, however the beam se~ent of a curved beam cannot have squal width over the depth of the beam because of the curvature as illustrated in Fig. A13.22, or in other words the length of the segment is greater on the outside edge (L ) than on the inside edge (Ll). Thus in calcuI4L--! f-=-i

~-I--"ivl
~ I
I,

Fig.A13. 21

Table A13.3 gives the value of K for various beam section shapes and be&~ radius of curvatures. The table shows that far only rather sharp curvatures is the correction appreciable. In general for airplane fuselage rings on frames the curvature influence can be neglected. However there are often fittings and mechanical structural units in airplane construction whose parts involve enough curvature to make the influence on the stress of primary importance. ~he concentration of stress on t~0 inst1e Bd~e of a curved unit in bending may In r Luence the rat tgue strength of unit considerably, thus a consideration of the possible influence of curvature should be a regular ;art of design procedure. In the inelastic or plastiC stress range, the tn rjuence of beam curvature should be consider~bly less since the stiffness ot a material in the inelastic range is much less than in the elastic stress range and changes rather slowly as the stress increases. A13. i2 Problems. (1) Fig. A13.23 shows the cross-section of a single cell beam with 12 stringers. Assume the walls and webs are ineffective in bend1ng. Calculate load in each stringer by use of beam formula. Also calculate stringer loads by equating internal resisting ~ament to external bending ~oment. Bach stringer area is same and equals 0.1 sq. in. applied bending moment M 100,000 in. lb. x

Straight Beam

FIg. Al3. 22 Curved

Stress Distribution

Beam

=-

lating the strain distri~ution over the beam depth the change in length ~L at a point must be divided by the segment Width at that point. Thus even though plane sections remain plane the strain distr1bution over the section will not be linear. '!'he width of the segment at any point is directly proportional to the radius of curvature of the segment and thus the strain at a point on the segment is inversely to the radius ot curvature. This gives a hyperbolic type of strain distribution as illustrated in :ig. A13.22, and if the strains are within the elastic limit the stress distribution will be similar. The development of a beam :ormula based on a hyperbolic stress distribution is

L5"-r5"--.l....5"-S"__' 5.. -1..5.. -'

, I I _' " I _ I _ I'

bee

(2) Same as problem (1) but change external bending moment to My = 200,000 in.lb. (3) Fig. A13.24 shows a beam section with 4 stringers. Assume web and walls inef!ectlve in bending. Stringer areas

A13.15 TABLE A13.3

BEAM BENDlliG STRESSES

,

TABLE A13.3 (continued)

VALUES OF K FOR USE ill THE BEAM FORMULA o : ~ I

VALUES OF K FOR USE IN THE BEAM FORMULA a : Mc

SECTION

R C

IN"",E
3.41 2.40 1. 96 1. 75 1. 62 1. 33 I. 22 1.14 1; 10 1. 08 2.89 2.13 1. 79 1. 63 1. 52 1. 30 1. 20 1.12 1.09 1.07 3.01 2.18 1.87 1. 69 1.58 1.33 1. 23 1.13 1.10 1. 08 3.09 2.2:5 1. 91 1. 73 1.61 1. 37 1.26 1.17 1.13

FACTOR K vUTSID,t, FmER FmER

0.54 0.50 0.65 0.68 0.71 0.79 0.84 0.89 0.91 0.93 0.57 0.63 0.67 0.70 0.73 0.81 0.85 0.90 0.92 0.94 0.04 0.60 0.65 0.58 0.71 0.80 0.84 0.88 0.91 0.91 0.58 0.62 0.66 0.70 0.73 0.81 0.86 0.91 0.94 0.95 0.5. 0.54 0.62 0.65 0.68 0.76 0.82 0.87 0.91 0.93 0.58 0.63 0.67 0.70 0.72 0.79 0.83 0.88 0.91 0.92

.'
o,

SECTION

c
1.2 1.4 1.6 1.8 2.0 3.0 4.0 6.0 8.0 10.0 1.2 1.4 1.6

FACTOR K lliSIDE OUTSIDE FIBER FffiER 3.55 2.48 2.07 1. 83 1. 69 1. 38 1.26 1. 15 1. 10 1. 08 2.52 1. 90 1. 63 1. 50 1.41 1. 23 1. 16 1.10 1. 07 1. 05
0.67 0.72 0.76 0.78 0.80 0.409

Pi ~I
dJl I
I

'-R

1.2 1.4 1.6

I.a

I-R-j

TtIJT 1 1i
II

ib--.j

~R--I

t=:JTI i
1-3b b
1

-oj

~e::Jl , 1- !
I

r
~

--401---1 R

r-

5b

--I

t;-I
,

'=ill
f.-

i<it.<!1- I T
......JCf-

., 1 I,
R~

2.0 2.0 4.0 6.0 6.0 10. a 1.2 1.4 1.6 1.8 2.0 3.0 4.0 6.0 6.0 10.0 1.2 1.4 1.6 1.8 2.0 3.0 4.0 6.0 8.0 10.0 1.2 1.4 1.6 1.8 2.0 3.0 4.0 6.0 8.0 10. a 1.2 1.4 1.8 I. a 2.0 3.0 4.0 6.0 8.0 10.0 1.2 1.4 1.6 1.8 2.0 3.0 4.0 6.0 8.0 10.0

0.224 0.151 0.108 0.084 o, 069 0.030 0.016 0.0070 0.0039 0.0025
u

0.292
0.224 0.178 0.144 0.867 0.038 0.018

0.86 0.89
0.92

0.94
0.95 0.67 0.71

~~~~lO
0.408 0.285 0.208 0.160

0.204 0.149 0.112 0.090 0.041 0.021 0.0093 0.0052 0.0033 0.168 0.128 0.102 0.046 0.024 O.all 0.0060 0.0039 0.336 0.229 0.168 0.128 0.102 0.046 0.024 0.001 0.0060 0.0039 O.,5. 0.243 0.1'19 0.138 0.110 0.050 0.028 0.012 0.0050 0.0039 0.41. 0.299 0.229 0.183 0.149 0.069 0.040 0.018 0.010 0.0055

~tI--41

l' f-

I ,

3T~ I .L 'L_: r...:' , R ----oj

I

I.a

0.75 0.77
0.79 0.86 0.89 0.92 0.94 0.95 o. 73 0.77 0.79 0.81 0.83

~:~~

2.0 3.0 4.0 6.0 8.0 10.0 1.2 1.4 1.6 1.8 2.0 3.0 4.0 8.0 8.0 10.0 1.2 1.4 1.6 1.8 2.0 3.0 4.0 8.0 8.0

0.127
0.058 0.030 0.013

g.g~:~
0.453
0.319 0.236

2.37
1. 79 1. 56 1.44 1. 36 1.19 1.13 1.08 1. 06 1. 05 3.28 2.31 1.89 1. 70 1. 57 1. 31 1.21 1. 13 1. 10 1. 07 3.55

O. 183 0.147
0.067 0.036 0.016 0.269 0.182 0.134

0.88
0.91 0.94 0.95 0.96 0.58 0.64

g. ~~~~

0.68
0.71 0.73 0.81 0.85 0.90

0.104
0.083 0.038 0.020 0.0087

0.92
0.93 0.67 0.72 0.76 0.78 0.80

1.11 3.14 2.29 1.93 1.'14 1. 51 1. 34 1. 24 1.15 1.12 1.12 3.63 2.54 2.14 1.89 1. 73 1.41 1. 29 1. 18 1.13 1.10

10.0

g. g~~;

4~
I

Trent I
2 ,
I
t,

rf

41

-!
t

I
I

t/2

1.1

I- R-1

j... C

I'

1.2 1.4 1.6 1. 8, 2.0 3.0 4.0 6.0 8.0

0.409
0.292 0.224 0.178 0.144 0.067 0.038 0.018

2.48 2.07 1. 83
1. 69 1. 38 1. 26 1. 15 1. 10 1. 08

0.86 0.89 0.92 0.94

0.95

10.0

0.010 0.0065

e equals distance from centroidal axis to neutral axis. References: Wilson and Quereau. "A Simple Method ot Determining stresses in Curved Flexural Members" "Advanced Strength of Materials", by Seely.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

s~own

A13. 17

in () on

fi~Jre.

ExterTIal applied

Je~ding ~cments

are;

M x= -500,000 in. lb. a~d My = 200,000 in. lb. Find stress on all four stringers by all three ~ethods which were explained in this chapter.
a
1.2)

(1.0)

(0.4)

"i -,-5"
'-..L
Fig. Al3. 24

(0.5)

(7) Fig. A13.27 shows 3 dif~erent Je~ sections. They are wade af aluminun alloy whose stress-strain G~agram is the sa~e as t~at plotted in Fig. AI3.19. Determine the ultimate internal resisting moment it the naximum compressive strain is limited to 0.01 in./in. Consider that upper portion is in compression. Compare the results obtained with formula M = 0bIlc, where crb = compreSSive stress when unit strain is
0.01.

r"

+-C

dI 12"---,

(4) The Zee section shown in Fig. A13.25 is sUbjected to bending moments of M x = 500 in.lb. and My 2000 in. lb. Find bending stresses at pOints a,b,c,d.

ED
500*

1-1"-1

1-1/21

" ---1 1-- 1

01
I--

1"--1

lJ
8

:{y4
!

1 1/4"

...

Lx

314

'~ "L 1-;;'

.L
15
c
d

-n'-

Fig. Al3. 27

fu4

Fig. A13. 25

--..L

50011

2" - - - 1

-Jl/4j-

(5 ) Fig. A13.26 shows a beam section composed

of three different materials. Find the stress at top Magnesium and bottom ~1/8 paints on Alum. I 1/8 beam due to a Alloy bending moment
MX= 80,000 In. lb. Et"'".ag.= 6.5x10 s

Steel

Esteel = 29 x lOS
EAlum.= 10.5x10<!l

-=r "T"
2'...,

' l
....

*"

5"

-l

~.15"

--I 1-0.4'4

0.1"

0.1"
~

1/1/8
1/16

.4;~~O.I"
-L~
Sec. 1
Sec. 2.

Fig. A1J. 26

~
'0.1"

0.6"

Sec. 3

(6 ) Fig. Ai3.26 shows

(8) Fig. A13.28 shows a curved beam, carrying

a cross-section of a wood beam composed at 3 kinds of wood labeled A, 3 and CJ glued to~ther to for.n a composite beam. If the beam is subjected to a bending moment M 75000 in.lb., x find intensity of bending stress at Fig.AI3.26 tap edge of beam. Also find total end load on ?ortions 3 and C.

two 500 lb. loads. Find bending stresses at points C and C", when beam cross-section is made 3 different ways is indicated by sections 1, 2 and 3. Use Table A13.3.

A = Spruce. 3 = Maple. C ::: Fir.

E = 1,300,000 ?si E = 1,500,000 psi E = 1,600,000 psi

A13.18

BEAM BENDING STRESSES

DOUGLAS DC-8 AIRPLANE.

Over-all view of the test wing section representing center wing section of DC-8.

-_.I ___ a".

A ctoae-cp view of wing test section showing details of wing ribs, stringers, etc.

CHAPTER A 14

BENDING SHEAR STRESSES - SOLID AND OPEN SECTIONS SHEAR CENTER

Al4. 1 Introduction.

In Chapter A6, the shear stresses in a member sUbjected to pure torsional forces were considered in detail. In Chapter A13, the sub-

Fig. A14.2

Load

-idxiD' F'

ject of bending stresses in a beam subjected to

pure bending was considered in considerable detail. In practical structures however, i t seldom r~ppens that pure ~endlng forces (couples) are the loading forces on the bea~. The usual case 1s that bending ~oments on a beam are due to a transfer of external shear forces. Thus bending of a beam usually involves both bending (longitudinal tension and compression stresses) and shear stresses. The same assumptions that were made in Chapter A13 in deriving the bending stress equations are l1kp.wlse used in deriving the equations for flexural shear stresses. With flexural shear stresses existing, the assumption that plane sections remain plane atter bending 1s not completely true. since the shearing strains cause the beam sections to slightly warp out at their plane when the beam bends. This warping action is usually referred to by the term nshear lag~. However, except in cases of beams with wide thin flanges, the error introduced by neglecting shearing strains is quite s~all and therefore neglected in deriVing the basic flexural shear stress formula. The problem of shear lag influence is considered in other chapters.
AB.2 Shear Center.
T~en a beam bends without tWisting, due to some external load system, shearing stresses are set up on the cross sections of the beam. The centroid of this internal shear force system is often referred to as the shear center for the particular section. The reSUltant external shear load at this section must pass through the shear center of the section if twist of the section is to be prevented. ThUS, if the shear center is known, it 15 POSSible to represent the external lead influences by eNO systems, one that causes flexure and the other which causes only t'Nist.

_ _ _ ryo __ N. ""----'
dA ---"-'-

Beam

N'A'~'
crt

seC~ion cr c -jdxt"- cr '

~l_'_,o

, ,
, ,

-X

Ct

"

, ,

" C o 0 l;

t!====:!==,:===j o F
Fig. Al4.!

AU,

C1tf

at

'~lL ~J,j~ / 0 F\.

crt crt
Fig. A14.4

JA~-:r::r

c. .

Fig. Al4.3
cr~

is as indicated. The stress crt is greater than because the bending moment due to the given beam loading is greater at beam section DO' than at FF'. Now consider that this beam portion dx is further cut as indicated by the notch DeEF in Fig. Al4.l, and this segment is shown in Fig. Al4.4 as a free body with the torees as indicated. Let crt = ~lmum tensile stress at a distance c from the neutral axrs , Then the stress at a distance y tram neutral axis is cry = at y/c. The total load on an element of area dA of the beam cross-section (see Fig. Al4.2) thus
equals C1ct / y dA

Now, referrin~ to Fig. Al4.4, the total tensile load on each face of this segment will be calculated. Total load on face CD = ctJC Y dA - - - - Yo

(1)

Total load en face FE = crt

c

.r

y dA

(2 )

Yo

AB.3 Derivation of Formula for Flexural Shear Stress.

2ig. A14.1 shows a loaded Simply supported beam. :dhen the beam bends downward due to the given loading. the beam portion above the neutral axis is placed in compreSSion and that below the neutral axis in tension. Consider a short portion dX of the beam at ~oints OF on the beam and treat it as a free body as shown in Fig. Al4.3. The variation of tensile and compressive stress on each face of the beam portion

From Chapter A13, the equation for flexural stresS cr was derived, namely crt = Me/I. Let M equal the bending moment at beam section DO' and l M tr~t at beam section FF' and let I and I' the moment of the inertia of the cross-sectional area about the neutral axis at these same beam sections respectively. Then substituting value of crt in equations (1) and (2) we can 'NTite, Total load on face CD = ~JCy dA
Yo

- - - - -

(3)

Al4.l

A14.2

BENDING SHEAR STRESSES

SOUND AND OPEN SECTIONS

SHEAR CENTER

Total load on race FE = l'

M'JC
Yo

ydA

(4 )

Now let ~ b dx equal the shearing force on race CE at the segment in Fig. Al4.4 where ~ equals the shearing stress and b dx the Shearing area. For equilibrium of the segment, the total forces parallel to x-x axis must be zero. It the beam is at unitorm cross-section, which is the case in our problem, then I = If and c and Yo are the same in both equations (3) and (4). Then the resultant horizontal force on the sides at the segment equals the ditference of the values in (3) and (4). or Resultant horizontal load

Fig. A14.5

MI M' fC ydA.
Yo

neutral axis are given on the tigure.

SOLUTION: -

For equilibrium of segment in x-x direction,

Z Fx

Shear stress at neutral axis x-x

MM' =- 1- JC ydA
y,

+ ~ b dx

"t"x_x
- - - - - - - - - (5)

"0'
V

J3.09

oY d.A

hence ~ I ; dx However M

M M' JC dA
y,
M'

Table A shows the calculation of' the term

3 . 09 [ , Y dA.
TABLE A. REFER TO FIG. a.
PORTION 1 2
Area d A

---ax

dll = a:x=V=the external shear on the beam section ..

v hence "t" ::: IE

JC ydA
y,

(6 )

Y
1. 045 2.34

ydA

It Is Important to nots that equation (6) applies only to beams ot unitorm section (constant moment at inertia). In airplane wing structures the common case is f'or beams to vary in cross-section or moment at inertia, and it this variation is considerable, equation (6) should not be used and resort should be made to equations (3) and (4). This tact Is Illustrated in example problem 2. This matter at variable cross-sections is discussed later in this chapter.
A14.4 Example Problems. Symmetrical Sections. External Shear Loads Act Thru Shear Center.

2.09

0.5

3 x 0.5 0.5 0.5

x 0.5

3 3'

, , ,

1. 045 1. 50

1. 096 3.510
0.710 0.710

0.25

x 0.5 :; 0.25

2.84 2.84 SUM

6.026

Fig. a

x-c-.

hence

~
3'
1

~}"h~ ,
,

Fig. b

-lyx

x----.:

., JY
!

I I ,

Example Problem 1. Fig. A!4.5 shows the cross section at a beam symmetrical about the 1-1 axis. Assume that a beam with this cross-section is Subjected ~o a loading which produces a shear load in the Y direction = to vy = 850 lb. Its location is through the shear center of' the section which lies on the centroidal y axis ot the beam section due to the symmetry ot the section about this axis. Let it be reqUired to determine the shearing stress at the neutral axis x-x and at pOints 1-1 and 2-2 ot the cross-section as shown in Fig. Al4.5. The neutral axis location and moment at inertia of' the section about the

"t" ~ 850 x 6.026 ; 377 psi. 27.2 x 0.5

Calculation ot shear stress at point 1-1, : 1:'1_1

V f3.09y dA = IxYb
2.09

Fig. b shows the eftective areas in this intregration. thus in Table AJ we leave out portion 1, hence I y dA = 4.93. Substituting,
~l-l

= 27.2 x 0.5 = 308 psi.

850 x 4.93

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A14.3

This Shearing stress is at a sect~on just adjacent to portion (2). If i t was taken just adjacent to portion (1), then the width b in the eq~tion would be 3 inches instead of 0.5 inch and the shearing stress would be 0.5/3 times that shown above, or in other words the shearing stress changes abruptly when the shear area cr~ges abruptly. Shearing stress at point 2-2 en cross-section: Fig. c shows effective area, thus in Table A, the portions (1) and (2) are omitted and

SOLUTION:

Bending Moments: 11132 = -600 x 132 = - 79,200"i 11175 = -600 x 175 = -l05,OOO"i 11218 = -600 x 218 = -130,500"* Table A14.1 shows the results of calculating the bending stresses at 3 points on each side of the neutral axis for the 3 stations. For example for station 132

3 . 09

y d.A equals 1.42.

850 x 1.42 = 27.2 x 1.0 = 44.4

2.59
~2-2

"MAX.
psi.

=~

(I1 Ie ) =, -

(79200 x 4)
38.67

='_

8180 pSi.

(tension at top edge and compression at lower

edges) .

Substituting,

(The width b = 0.5 + 0.5 for the two portions 3,3' i. The Shearing stresses as calculated act in the plane of the beam cross-section in the Y direction and also with the same intensity paralle to the Z axis which is normal to the beam section.
Example Problem 2. VARIABLE 110l1Th'l' OF INERTIA. Fig. A14.6 shows a cantilever beam loaded

For a paint 1 inch from either edge of the beam

=+ (79200 x 3) = 6135 38.67

TAJ:ILE AB.1

psi.

z
On Top tnt lI\t I or 1" From 'I 2" From 1 Bottom Top or Top or i p~,rt~~n ~,rt:~D i ~rt~~11 fiber Bottom Bottom' A B r 'C y-4",y-3" 1,"2,,1
,

With a single load of 600 lb. at the end and acting through the centroid of the beam crosssection. The beam section is constant between stations a and 132, then it tapers uniformly to the sections shown for stations 175 and 218. The shear stress distribution on the beam crosssection at stations 175 and 218 will be deter:nined.

BenQ1ng Moment M

132 79,200"f =.8180 175 1 I 105, OOO"fl .!oU80 Z1. 130,800'" =.8180

:t6 U 5
+8135 +6135

+4090 +4090

7157i1 i

t4090

;~:~ : ~~~: i

H12' I 1023* 1023 , 1023 i

~ ~i-----;;c:-~O;::+-;-;;;;-;;;:r-;o.,-------;;2
Sta. '218

!!-__;-__'-_-,
Sta:175

Fixed

'-43"~43"'1I -

+--- 1 3 2 " - - - -

f"0.

sci.. 132

Stl.O

~.::1
B

From the results in Table Al4.l, it should be noticed that the change in moment of inertia between the three stations is directly porportlonal to the change in bending moment, hence the same value for the bending stresses for all three stations. Columns 6, 7 and 8 give the total bending stress load on portions A, B and C of the three cross-sections (see Fig. A14.6). These values equal the average stress on the portions times the area of the portion; for example, for station 132, the load on portion A:
load portion C:

= 6180

6135

x 1 x 1

7157 Jil, and for

sts, 218 I : 64.0 in. 4

sta. 175 I ::; 51. 33 in. 4

Sta. 132 I: 38.67 in. 4

load

= 40902 +

a x 2 x .25 - 1023*

Fig. A14.7 shows the tension and compressive stresses acting on a portion of the beam between

._lS4

A14 4

BENDING SHEAR STRESSES

SOUND AND OPEN SECTIONS

SHEAR CENTER

Fig. A14.8 shows the resUIting horizontal shear stress pattern resulttng trom the loads in Fig. A14.7. For example, it we take a section along the beam 1" from the tap or bottom edge ot the beam and treat this portion as a free body as shown in Fig. A14.12 applying ZH = 0,
stations 175 and 218.
ZH

shear stresses between stations l32 and 175 would be the same, since the change in bending moment and moment of inertia have been ~de the same as between stations 175 and 218. Figs. A14.9 and A14.10 show the shear
stress patterns i f the formula -r = V

~~

dA be

=-

7157+7157+'t'x43xl.O=

0, hence

't'"

=0

~ 613

7157. -~.~ 09 ~'---B 1023*

X - 7157 IB - -- _10224*

--- --- --18180

~a
Iv

IB

1- 1023
I--1023*

used for each station. The discrepancy is conslderable as the equation does nat apply to beams of varying section. To illustrate the calculation by the shear stress formula. the shear stress will be calculated at the neutral axis for the beam sect~on at station 175.
't'=Tb V

.!.Qill.
B ._AI
~

---10224* Ik:!--7157f

;;

.;

Fig. Al4.7

~ ~

where

f>

oYdA

cIA

=1

x 1 x 3.5 -+ 2 x 1
0.25

x 2.5

+ 2

1 9.0

Fig. A14. 8 Fig. Al4.10 Fig.A14.11

,--J
477*/0"

('4'2

~
II

hencs. -r = 600 x 9.0 = 420 ps t , as compared 51.33 x 0.21; to the true shear stress of 477 in FIg. A14.8.
TABLE A14.2 MAXIMUM SHEAR STRESS FOR SIMPLE SECTIONS
C ross section Max. Shear Stress
1

409 20'

~ 49:6
!.--420

1r431

to.~
to 59.5
I

LOCatiOD of

Max. Shear Stress

True
Shear

~ ~
7157 10224

Pattern

Sta 218 Sta 175 Shear Stresses By Equation (6) 43"

N-B-N N-ON
~7
N-~N

3V l' = 2A

I
4V 't' = 3A

-il-1/4"

-- -F sx43x.25 Fig. A14.13 43"

====~-F sx43xl

7157 15336

I I

I I
,
I
Dd)

"'1"1-

D7~

I 4V( I -r = 3A 1'"

na .. d.

Fig. A14.12

Stmllarly, treating the portion between the edge or the beam and a point 2" tram the edge as I a tree body diagram as shown in Fig. A14.l3,
ZH

=- 7157
25

+ 7157 ... 15336 + 10224 + " x 43 = 0, therefore -r 477 psi

tN-E\i-N eL -' 0
.t"-.~-~

-r = 2A

3V

e='6
!

l'

9V = 8A

e=8"
I

Obviously, the shear stress on portion C is constant, since the end load on this portion at both stations Is the same, or 1023*. Fig. Al4.8 shows the general shape at the shear stress d1stribution on the beam section at any paint between stations 175 and 218. The

N ---- N

(j~e

4V 't = 3A

I ,
I

e 0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.5 Maximum Shear Stresses for Simple Cross-sections.

A14.5

Table Al4.2 gives the value of the maximum shear stress on a few simple sections and where it occurs on the cross section, (e is distance from neutral axis to point of maximum shear stress). V equals the shear load normal to the neutral axis and it acts along the centerline axiS, thus no twisting on the section. A is the total cross-sectional area. The maximum shear stress is given in terms of the average shear stress which equals VIA'
AB. 6 Derivation of Flex\J.ral Shear Flow EquaUon. Symmetrical Beam section.

qYb dy "'" (C
qy

y - Cy)

= 0,

hence

=-

( Cy' - CY)

But (C

y-

Cy)

=!1X I~ !1x

J:

dy

z dA (See Art. A14.3)

bz Ja dA

hence

qYb = - :

To emphasize further the fundamental relationships, a second derivation of the equation for shear stress distribution will be presented. Fig. A14.14 shows a portion of a cantilever beam carrying a load P at the tree end as shown. _ This load is so located as to cause the beam to bend in the XZ plane without twist about a Y axis. The problem is to derive relationShips which will give the magnitUde and sense of the shear flow distribution on the cross-section at the beam. Fig. A14.15 shows a free body of a small portion of the beam cut out from the upper flange of the beam at points (a b) in Fig. A14.14. Under the given external load P it is obvious that the upper half' of the beam is subjected to compressive stresses. In Fig. Al4.15, Cy is larger than Cy since the cantilever bending moment is greater at station Y'

however;' = Vz , the external shear in the direction.

b hence qy = _ v z z dA b Ix Ja

- (7)

Equation (7) gives the Change in shear flow force qy between points (a) and (b) and since in figure A14.15 the value of qy at (al 15 zero because of a tree surface, the value of qYb in equation (7) is the true shear flow force in lbs. per inch at point ("b). The stUdent should realize that equation '(7) gives the shear flow q in the Y direction. The minus sign in equation (7) means that the pOSitive sense as assumed by the arrowhead on qYb in Fig. A14~15 1s incorrect or should be reversed. The initial problem was to determine the shear flow force system in the plane of the beam cross-section or the ZX plane. From elementary engineering mechanics, we know that if a shearing stress occurs on one plane at a point in a body, a Shearing stress ot the same intensity exists on planes at right angles to the first plane, or in general at a point,
(8 )

Su. y
/ &===~~~;E!J~~ St2.y Sta.y .L
dy

r,

;-I""'_X-;--;

'~b \Y

Sla.f' 1iJ1}--~d

xAI
Fig. A14.14 Fig. A14.16

C' Y
Fig.A14.15

The free edge of beam flange forms the right side face of the element in Fig. A14.15 and thus the shear flow force on this face is zero as indicated. The shear flow forces on the 1nternal or cut faces 1n Ibs. per inch are qy and qx as indicated. Since the sense of these shear torces is unknown, they will be assumed as acting in the positive direction. (See Fig. A14.l6 for positive sense of forces acting parallel to each of the coordinate axes XYZ. Now consider the equilibrium of forces in the Y direction for the element in Fig. A14.15 ZFy=o,or

Before the shear flow in other planes is completely defined its sense (pOSitive or negative) must be :<nown. Equation (7) gives the magnitude and sense for qy at any desired paint on the cross-section. The question of the sense at the associated qx and qz is easily determined trom an observation involving equilibrium at moments. This tact will be explained by referring to a number of tree body diagrams. Fig. A14.17 shows a free body of a small element cut from the beam in Fig. A14.14 at point (a) on the cross-section. The forces on this free body are the compressive forces C~ and C on the front and rear faces and the shear forces on the various faces as indicated. The right side face of the element Is a tree surface and thus q on this face is zero. The shear flow qy on the left side face is calculated from equation (7) namely

A14.6

bENDmG SHEAR STRESSES.

SOUND AND OPEN SECTIONS.

SHEAR CENTEH.

r-7C~
-@
CIt':';'

I;>.,,I.-.!:...... ;
dy'l:r
I

q_# a

" x I 1>:>'" I, .....t..

'-e-~=iI~

c'
Flg.A14.21
C

"
b

,~di- r-:""~
C'

C'

Fig.A14.l8

Flg.A14.l7

t.

""'~=01

'. t@;r
~

back side and through a point en the line of action of c' and C, the shear flow qz on the front face must act downward in order to balance the moment due to qv . Fig. Al4.22 shows the results tor a free body of the lower flange plus a small web portion. qy is POSitive from equation and thus qz must be downward for mament equilibrium. From the results obtained for these 6 different locations, a simple rule can be stated relative to sense of shear flows in the plane of the cross-section, namely: If the calculated shear flow is directed toward the boundary line between the two intersecting planes of the particular free body, then the shear flow on the other plane is also directed toward the common boundary line, and conversely directed away if the calculated shear flow is directed away.

I,

00(>:"1

d-qx

--e
qx

" ~

Fig.A14.22

Fig.Al4.19

T'

Fig.A14.20

Fig. A14.14 shows the sense of the shear flow pattern on the beam section as determined for the given external loading.
A14.7 Shear Stresses and Shear Center for Beam Sections with One Axis of Symmetry.

For the given beam loading in Fig. Al4.l4 the load P is up, theretore Vz has a positive sign. For ~~y portion at area (A) ot the cross-section the distance z in the above equation is therefore POSitive. Therefore in substituting in the above equation qy cames out negative for any point above neutral axis, and likewise for any point on beam section below the neutral axis the distance z would have a negative Sign and qy would came out positive. Therefore the sense of the calculated shear flow qy on the 13ft side of the element in Fig. Al4.l7 is negative or as indicated by the arrow on the force vector. Now to find the sense of the shear flow qx on the front face take moments about a Z axis acting in the plane of the rear race and through pOint (0) Which is on the line of action of the forces C' and C. Only two forces have moments, namely the side shear force qy dy and the front tace shear force (qx!2)dx. It is obvious by observation that qx on front tace must act to the lett as shown it the moment is to equal zero. The total shear force on tront face is (qx/2)dx because shear flow at right edge 1s zero and it varies linearly to qx at left edge at front face. Figs. Al4.l8, 19 and 20 show free bodies ot elements taken at the other three corners of the beam section whiCh are labeled 0, d, e in FIg. Al4.l4. For all tree surfaces q is zero. The sense of qy as before 15 given by the equation as explained betore, hence qy is negative in Fig. A14.18 and POSitive in tl~~es 19 and 20. A Simple consideration of moment eqUilibrium as explained tor Fig. A14.l7 gives the sense of the shear flow Qx as shown in the 3 figures. Fig. A14.21 shows a free body of the entire upper beam flange and a short portion of the beam web. qy from the equation is negative and this acts as shown ~n the figure. Now if we take moments about a X axis in the plane of the

Example Problem. CHANNEL SECTION. Fig. Al4.23 shows a cantilever beam with channel shaped cress-section carrying a 100 lb. downward load as shown. The problem is to determine the lateral position of this load so that the beam will bend without ~Nist. This posit:on will coincide With the lateral pOSition ot the centroid of the shear flow system on the beam cross-section-which holds the external load in eqUilibrium Without twisting of the beam section. The cantilever beam has been cut at a section abc d (Fig. A14.23) which is tar enough rr-on the fixed end or the beam (not Shown) so that the etfects of beam end restraint against section warping can be neglected. In Fig. A14.23, the internal forces holding the beam in eqUilibrium are sketched in. They consist of a longitudinal stress system of tension and compreSSion and variable shear flow system in the plane of the cross-section. In this problem we are only considered with the internal reSisting shear flow system. For solution of thiS problem the moment of inertia Ix must be known. If calculated it would be Ix 0.2667 in.
SOLUTION: -

From equation (7)

A (8)

.-

z Z z V Ix

we know that the shear stress Is zero at a free edge, thus the solution at equation (8) is started at either paints (a) or (d). The shear Vz is - 100 lb. Thus equation (8) for our problem reduces to,
qy

=-

- 100 __ 375 Z 0.2667 Z z A

A - - - - - - -

(9)

s.u

j."::'--';::;;;'_"""i.;H~_

""= .""

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

100 lb.
~b-,

AB.7

Free end
q: 563

Shear Flow Diagram c Fig. A14. 25

"-,.---,..
Fig. A14.23 Fig. A14. 24

C.

Qy

We will start at point (a) in solving equation (9) and proceed around the section.
Point (a) Point (b)
q

= a (free surface) Y a

Z = - 1, A = area between (a) and (b).

qy =qy b a

+375Z~ZA

Yb

= a + 375 (- 1)(1 x 0.1) - 37.5 lb/in.

Point (0) on X axis.

qy

- 37.5 + 375

Zb

z A

boundary line as shewn in Fig. Al4.24. Cammon sense tells us that the resisting shear flow qz on the channel web must be directed upward because it is the only force system that can balance the 100 lb. load as far as ZFz = 0 is concerned. In general the shear floW is continuous around the section and only reverses when it passes through zero which only happens in closed tubular sections. In general, it is possible in most cases by observation only, to determine the sense of the shear flow at some one point on the beam cross-section. The shear flow being like a flow of liqUid will continue in the same general direction along the center line at the parts that make up the beam section. The small arrows on the beam section at Fig. AI4.23 show the sense of the shear flow pattern over the beam section. In Fig. Al4.25, the shear tlow values as calculated at the various paints are plotted to fonn a shear flow diagram for the beam section. Between points a and b or d and c, the arm z in equation (9) is constant and thus q varies linearly as plotted. Between b and 0 or 0 and c the arm z changes and the area is also a function of z, thus q varies as za or parabol1c as plotted. The initial problem was to locate the centroid of this final shear flow system which is generally referred to as the shear center. In Fig. A14.25, Qab' Qbc and Qed represent the resultant of the shear flow force system on these three portions of the beam cross-section. Each torce 15 equal in magm tude to the area of the shear flow diagram tor the particular beam section portion. Hence,
~b

= 1 x 37 .5/2

= - 37.5 + 375 (- 0.5)(1 x 0.1) = - 56.3 lb./in. Point (c). qy

= 18.75

'"be 2 x 37.5 + (56.3 - 37.5) x 2 x 2/ 3 = 100 lb. Cd = 1 x 37.5/2 = 18.75 lb.

lb.

= -56.3+375 (0.5)(1 x 0.1) =-37.51b./in.

Point (d). qy
d
~e

The resultant R at these three shear forces will now be determined. ZFz=lOOlb., Hence

=-37.5+375 (1) (1 x 0.1) = 0 (tree surface)

Z Fx

= 18.75 - 18.75 = a

know that. the intensity of shear flow in the

ZX plane at any point equals that in the Y di-

R ::

(z

F z " + Z Fx 2) J. / a = (100" + 0)J./2 = 100 lb.

rection at the same paint. Thus qx or qz equal the qy values above. The sense of the qx and qz shear flows must be known before they are completely defined or known. Fig. Al4.24 shows a tree bOdy at a small element at point (a) on the end of the lower beam flange. For any point below the X centroidal axis equation (9) will give a minus sign ror Gy. Thus in Fig. Al4.24 qy acts as shown, or directed toward the boundarJ line be~Neen the side face and the front face. Then by the simple rule as given in the previous article qx is also directed toward this cammon

The ~oment of the resultant about any point such as (b) in Fig. Al4.17 must equal the moment of the shear flow forc~ system about point (b). Let e be distance from point b to line of act10n at the resultant R. Hence R e :: Z 11 b (of shear now system) or
100 e

= 18.75

x 2

or e

0.375 inches.

Thus the centroid of the internal shear reSisting force system lies on a vertical line

A14 8

BENDING SHEAR STRESSES

SOUND AND OPEN SECTIONS

SHEAR CENTER.
~ethod.

0.375 inches to left of point b as shown in Fig. AI4.17. For bending about the centroidal Z axis without twist the resultant of the internal shear flow system would obviouSly, due to symmetry of section about X centroidal axiS, lie on the X axis, hence shear center for the given channel section is at point 0' in Fig. Al4.l7. The external load of 100 lb. would have to be located 0.375 inches to the left of the centerline of the channel web It bending of the channel without twist 15 to occur.
AB.8 Shear Stresses for Unsymmetrical Beam Sections.

Method 3.

The k

(14

In chapter Al3, which dealt with bending stresses in beams, three methods were presented tor determining the bending stresses in beams with unsymmetrical beam sections. The bending stress equations for these three methods will be repeated here: Method 1. The Principal Axis rretnoo .

a _

(9)

Method 2.

The Neutral Axis Method.

- - - - - - - - - - - - - - - (10)
The Method USing section properties about centroidal Z and X axes. For brevity this method will be called the k method.

,Method 3.

EXPJ1PLE DROBLE1 USING THE T:ffiEE DIFFERENl' =HODS. Fig. A14.26 shows a Zee Section subjected to a 10,000 lb. shear load acting through the shear center of the section and in the direction as shown. The problem will be to calculate the shear flow qy at two pOints on the beam section namely paints b and c as indicated on the figure. The shear flow at these ~NO places will be calculated by all 3 methods. Since all 3 methods require the use at beam section properties and since the direction ot either the principal axes or the neutral axis are unknown the z tirst step in the I i solution regardless ot which method is used .1" Is to calculate the section properties 2 about centroidal X and Z axes. Table A14.3 x-- -c r-r-x 3" gives the calculations. The section has been divided into 4 portions labeled 1, 2, 3 and 4.
V = 10000 lb. . 4'J:
..L-

"4 ~dl ,tl 1b~

A~I"l
k
:1"
Fig.AB.26

a = - (k.
where,

Hz -

k,

!'lx)

x - (k.

!'lx -

k,

Hz)

z (11)
Portion M" IA)
1

TAi:lLE A14.3

Mm

Arm

1
I

Ax>

",'

Ax"

! "

I
I.

'x

a
4

= Ix

I z - I' xz

Iz

0.10 1. 4'5 0. 4"J . 06525 0 : 0 10.14 0.70 0 '0.141 0.701 0 0.10 .1.451 0.4:)1.06525 Z .1305

.21025 I .02025 i .00838 000017 . 000117. 02287 .068,601 0 1000117.02~87 .06860 0 .21025 .020251.008381.000017 .55770 .040:)0 .01690 .04577

, ,
I

IX'

= Z A

In referring back to the derivations ot equations (5), (6) and (7) the above equations (9), (10) and (11) can be written in terms ot beam external shears instead ot external bending moments as tallows: Method 1. The Principal Axis Method.
VXp VZp q y : _ - l : zp A - l: x p A I xp I zp

Iz

+ Z i x = .6035 in", = Z A Xlii + Z i z = .0574

Zl1

I xz = Z A xz = - .1305

(Note:

In Table A14.3 t x and i z are the moments ot inertia of each portion about its awn centroidal axis).

- - - - - -

(12)

SOLUTICN BY PRINCIPAL AXES METHOD. (Method 1) Let ~ be angle between Principal axes and the X and Z axes. From chapter A13, tan 2 ~ = 2 I xz

The Neutral Axis Method.

l:

Zn

I z - Ix

A - - - - - - - - - - - - - (13)

2 (- 0.1305) _ 0.0574 _ 0.6035 - 0.477S

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Al4.9

hence 2
sin

= 2S 32.2' = 0.2210 and

0 -

or
cos

$ = 12

_ 46.1'
.

Now SUbstituting in eqUation (15)

qb = -

= 0.97527

The moments of inertia about the principal axes can now be calculated.

0.~16

(1 x 0.1 x 1.5136)

Ix

= Ix cos

+
,

I z sin ~ - 2 I xz sin

cos

6792 - 0.02782 (1 x 0.1 ) ( - 0.1184) = - 1756 + 2890

= 0.6035 x .97527 + .0574 x .2210 - 2 (-0.1305) x .97527 x .2210 = .63316 in'. I z := Ix sin' + I z cos' ~ + 2 I xz sin ~ cos~
p

--'

= 1134 Ib/in.
Calculation of shear flow at point (c). For portion (2) area A = 1.4 x 0.1 zp .70 x .97527 0.6825 in.

= 0.6035 x 0.2210 + .0574 x ~ + 2 (-0.1305) x .97527 x 0.2210 = 0.02782 in~.

The equation for shear flow q is,
- - (15)

0.14

xp = .70 x .2210 = 0.1547

The shear flow at paint (c) equals the shear flow at point (b) plus the effect of the portion (2) between points (b) and (c)., hence
7348 ~ = 1134 - 0.63316 (0.14

x 0.6825) -

In Fig. A14.26 the external shear load is 10000 lbs. acting in a direction as shown. ResolVing this shear load lnto z and x components, we obtain, V z 10000 x cos 30 8667 lb. V x = 10000 x sin 30
0

0.~2~~2

(O.lA x 0.1547)

= 1134 - 1109 - 5285 =

- 5260 Ib./in.

5000 lb.

The shear stresses at these two points (b) and (c) would equal q/t = 1134/0.1 and - 5260/0.1 or 11340 psi and - 52600 psi respectively,

Resolving these z and x components further into components along the principal axes we obtain,

SOLUTION BY NEUTRAL AXIS METHOD.

(Method 2)

= 8667 x

.97527 - 50000 x .2210 .2210 + 5000 x

=8667 x

=7348 lb. .97527 = 6792 lb.

Calculation at shear flow at point (b). (See Fig. AI4.26). Fig. Al4.27 shows the pasi~ion of the principal axes as calculated. The shear flow at the free edge of the upper portion (1) is zero. For the shear flow at point (b), the area to be used in the summations Z zoA and Z xpA Is the area of element (1). The arms zp and x p can be calculated by Simple trigonometry. Fig. A14.27 shows the value of these distances, namely xp = - 0.1184 and zp = 1.5136.

In thiS solution n~ it is necessary to find \ i the neutral axis for the given external load: .O~~6" iag. In Fig. A14.28, the angle Q is the angle zn: .186 between the plane of loading and the zp pr tn~ """."\."' Cipal axt e , and this .g angle Ii) equals :30 e + 12 0 61 0 . 4. - 46" = 42 - 46". 15Q.,2S' ' 'pn Let a equal angle between xp principal axis and neutral axis n-n. From chapter A13, we find, Ix tan Q tan a = - -"-,p,---

\1

.l12

Fig. A14.28

= _ 0.63316 x 0.9245 = - 21.052 0.02782

Whence. a = - 87" - 17'< (See Fig. Al4.28 for location of neutral axls. sin a = 0.9989 , In = Ix
Fig. A14.27

cos a = 0.04742 sin 2 a, substituting,

cos 2 a + I z

'.) r

("""~

A14.l0

BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS.

SHEAR CENTER.

In:: 0.63'316 x .04742

+ .02782 x ~

= .02919

q = - [34.25 x 5000 - (- 7.406 x 8667)J Z xA -

The component or the given external shear load normal to the neutral axis n-n equals
V n 10000 x sin 45 - 29' 7130 lb.
From equation (13) q : - Vn Z In

[3.257 x 8667 - (- 7.406 x 5000)J L z A, whence

q - - 235438 Z x A - 65258 Z z A Shear !low at point (b)

Zn

For portion (1), x = - 0.45 in., z : :; 1.45 in.

A
A :::; 1

x 0.1 :::; .1

SUbstituting, Shear tlow at point (b): The distance from the neutral axis to the
centroid or portion (1) equal

Zn -

qb

=-

235438 (- 0.45) 0.1 - 65258 x 1.45 x 0.1

0.0466 in. Point (c)

hence
7130 qb = - :029I (1 x 0.1) (- 0.0466) = 1135 1b/in. Shear r10w at point (c): -

= 10597

- 9462

= 1135

lb./in.

For portion (2), x:::; 0, Z :::; 0.7, A :::; 1.4 x .1

Zn = 0.1868

In.

.14

Clc = 1135 - 235438 (0) .14 - 65258 x 0.7 x 0.1 Clc

=1135
=

.~~9

(1.4 x 0.1) .1868 5255 1b./in.

= 1135 - 0 - 6395 = - 5260 lb./in.

1135 - 6390

=-

SOLl1l'ION BY l'lE:l'HOD 3 - (The k Method).

In this method, anl) the section properties about the centroldal X and Z axes are needed. These properties as previously calculated in Table A14.3 are,
Ix

=0.5035,

Iz

:: 0.0574,

I",

=-

0.1305

The shear tlow equation tram sq. (14) IS, q = - (k. Vx - k, Vz ) Z xA - (k. Vz - k, Vx )
Z

General Comments. The author prefers solution method number 3 slnce it avoids the calculation of additional angles and section properties as required 1n methods 1 and 2. Furthermore, in calculating the shears and moments on the airplane Wing, fuselage and other major sturctural units it Is convenient to refer these shears and moments to the conventional X Y Z axes, and thus these values can be used in method 3 without further resolution. From an investigation of many airplane stress analysts reports, it appears that the engineers at most airplane companies preter to use method 3.
A14.9 Beams with Constant Shear Flow Webs.

- -

- - --- - - - - -

(16)

F1g. A14.29 shows a

0.1305 - - 7.406 =- 01762

1<0 1<0

= Ix
= Ix

Iz

Iz -

Irl =

0.0574 .01762

=3.257

Ix I z - I xz

0.6035 =0.01762 = 34.25

Resolving the external shear load ot 10,000 into x and z components, we obtain, Vx = 10000 sin 30' = 5000 lb.
Vz = 10000 cos 30 0
equation (16) we obtain ::

8667 lb.

Substituting values of Vx ' Vz and k values in

beam composed ot heavy tlange members and a curved thin web. For bending about the X-X axiS, the web on the compreSSive, x--:-~ side ot the beam absorbs ~ i I very little compressive ' I q stress, since buckling ot .......Flange the web will take place under low stresses, parZ Fig. A14.29 tlcularly when the curvature of the web Is small. on the tension Side, the web will be more eftective, but if the flange areas are relatively large, the proportion of the total bending tensile stress carried by the web is small as compared to that carried by the tension flange. Thus for beams composed of indiVidual flange members connected by thin webs it is orten assumed that the flanges develop the entire longitudinal bending resistance w~ic~ therefore means that the shear flow is constant

~ITe
.
,,~

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A14.ll

over a particular web. In other words in the z I z d A, if the area shear flow equation q : V of the web is neglected then q is constant between flange members.
RESULTANr OF CONSTANT SHEAR FLOW FORCE SYSTEm

EXA!1PLE PROBLEM - RESULTANT OF A CONSTANr FLOW FORCE SYSTEM. Fig. A14.30 shows a constant flow force system thru points ABC o E with q 10 lb. per inch. The resultant or this force system is required.

Ix

Fig. A14.Z9 shows a beam assumed to be car-rying a downward shear load (not shown) and to cause, bending about axis x-x without twist. Assuming the two flanges develop the entire bending reSistance, the shear flow q is constant on the web and acts upward along the web to balance the assumed external downward load. The resultant of this reSisting shear flow force system will give the lateral position of the shear center for this beam sectlon. The problem then is to find the reSUltant ot the shear flow system. Let q = load per inch along web (constant). Let R = resultant of the q torce system. From elementary mechancis, -+- Z qz ,where qx and qyare the x R =.jz and y components of the q forces along the web. SiLce q is constant, Z qx is zero, hence,
(17)

SOLUTION: -

Draw clOSing line between the beginning and end points ot ror-ce system. (line AE). The length h ot thiS closing line is ZO incheS.
From eq, (17) R

qx

The direction of the resultant is parallel to line AE or horizontal In this problem. To tlnd the location of R take moments about any point such as (0). D~ lines trom point (0) to points A and E. The enclosed area (A) equals

= q h = 10

x 20

= 200

lb.

s x ic- s x io- .5nxS -+-lOx

From eq, (19)

50 = 189.3 sq. in.

Equation (17) states that the magnitude ot the resultant of a constant flow torce system is equal to the shear flow q times the straight line distance between the two endS of the shear flow system. Since Z qx is zero, the direction of the resultant 15 parallel to the straight line JOining the ends of the web. The location of the resultant torce is found by using the principle at moments, namely, that the moment or the resultant about any pOint must equal the moment or the original force system about the same point. In Fig. Al4.29 assume point (0) as a moment center. ThenRe=qLr
butR~qh

e 2 hA

:a

Z x zlg9.3 18.93 In.

Fig. A14.30 showS the resultant at ZOO lb. acting at a distance e from (0) and parallel to line AE.
A14.10 Example Problems for Beams with Constant Shear

Flows Between Flange Members.

EXAMPLE PROBLEM 1.

Beam Section Symmetrical About One Axis. Fig. A14.31 showS an open beam section composed ot 8 flange members connected by thin

hence e _ q L r _ L r

-(f1l-T

(18)

In equation (18) the term L r is equal to the area (A), where area (A) Is the enclosed area tormed by drawing straight lines trom moment center (0) to the ends of the shear flow force system. Thus
~Nice

e:"11 - - - -

2 A

(19)

The shear center thus 11es at a distance e to the left of point (0), and the external shear load would have to act through this point if ~Nistlng were to be eliminated.

Fig. AB.31

A14.12

BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTER.

sheet to form the webs and walls. The flange members are numbered a to h and the areas of each are given on the figure. It will be assumed that the webs and walls develop no bending resistance and thus the shear flow between adjacent flange members will be constant. The problem is to determine the shear center for the beam section.
SOLUTION: -

in order to indicate at what point the shear flow ~ is being calculated. Qab = - 2.5 Za z A = - 2.5 x 5 x 0.1 =

= 1.25 Ib./in.
The first letter of the subscript refers to the flange member where the shear flow q is being calculated and the second letter indicates 8n which adjacent side of the particular flange member. Hence qab means the shear flow at flange (a) but on the side toward (b). Qab = qba = - 1.25 (Since no additional flange area is added, and thus shear flow is constant on sheet abo
'lbc = qba - 2.5 "b Z A - - 1.25 - 2.~ x 5 x 0.4 - 6.25 lb./ln.

Since the beam section is symmetrical about the X axis, the centroidal X and Z axes are also principal axes, since the product of inertia I xz is zero. The vertical pOSition of the beam section centroid due to symmetry is midway between the upper and lower flanges. To find the horizontal pOSition of the centroid; take moments of the flange areas about the left end or line be:

x=

lo A x = 0.4 x 15 + 0.2 x 10

ZA

1.6

0.2 x 5

= - 6.25
Sco - 2.5 !c z A

= 5.625 In.

The moments of inertia tor the section about the centroidal x and y axes are: -

= - 6.25 - 2.5 x = 'led =- 1.25 'ldc 2.5"d z

qed = 'lde = 0 qer = 0 - 2.5 "e z qre = qer .. 1.25 qrg %r qgh qhg qha

(-5xO.4)= - 1.25

1.25 - 2.5 (- 5 x 0.1)

=
A

I z = 0.8 x 5.625 4.375

0.2 x 0.625

0.2 x

=o -

2.5 (- 5 x 0.1) = 1.25

0.4 x 9.375- = 64.4

In~.

= 1.2:
t:

HORIZONTAL PCSITION OF SHEAR CENTER: -

- 2.5 "r z A = 1.25 - 2.5 5 x 0.2) = 3.75

The horizontal pOSition of the shear center will coincide with the centroid ot the shear flow system due to bending about axis xx without ~Nist. For simplicity, to elimir~te large decimal values tor shear flow values an external shear load V 100 lb. will be assumed z and the internal reSisting shear flow system will be calculated for this external loading. From equation (8)

= qrg = 3.75 = 3.75 - 2.5 = qgh = 1.25 = 1.25 - 2.5

"g z A

= 3.75-2.5x5

x 0.2 1.25

= =

loh z A = 1.25 (checks free edge at h). 0

2.5 x 5 x 0.1

SUbstituting values at V z and Ix

100 " z A
40

2.5 " z A

We could start the solution at either at two points (a) or h since these points are free edges and thus qy is zero. In this solution, we will start at the free edge at paint (a) and go counterclockwise around the beam section. The area or each flange member has been concentrated at a point coinciding with the centroid of each flange area. In solving for the q values the subscript y will be omitted, and subscripts USing ~~e flange letters will be used

The Sign or sense of each shear flow is for the shear flow in the y direction as explained in the derivations of the shear flow equations. The procedure now is to determine the sense of the shear flow in the plane of the cross-section or in the xz plane. It is only necessary to determine this sense at the beginning point, that is in sheet panel abo The surest way to determine this sense 1s to draw a simple tree body sketch at flange member (a) as illustrated in Fig. Al4.3l. The shear flow on the cut face is qy(ab) = - 1.25 and this value 1s shown ~n the free body. By s1mnle rule given at the end of Art. Al4.6, the shear flow in the plane of the crossFig. A14.31

"'" : '

.~. <: ~ -v- ;,. ::.. '~ .~_:~'-_ ~ ~f,;~~-;'.. ":<h:~2.:-._;~ '".:: ~ :.

,>:. ~:, :.L" ~~: ~ '..<~~' ~ :', ; -: ," .~,,;)~":':' ~.~ ;: ~.,~

~ ~ '~

:,:::,.

----

,~.;:.;. '~~.r . ~~? 5J-:~~. .

t';-

ANALYSIS A D DESIGN OF

FLIGHT VEHICLE STRUCTURES

A14.13

section is also directed toward the cammon boundary line and thus qx(ab) has a sense as shown in Fig. Al4.3l. The sense of the shear flow on the cross-section will now continue in this direction until the sign changes in the origianal calculation, which means therefore the shear flow sense will reverse. Fig. A14.32 shows a plot of the shear flow pattern with the sense indicated by the arrow heads.

We will again start at the free edge adjacent to flange (a) where qy - O.
~b

=- 1.55

Za x A 0.0971 Ib./in.

=- 1.55

(- 0.625 x 0.1)

qba = 'cab = 0.0971 qbc = 0.0971 - 1.55 Zb x A = 0.0971 - 1.55 (- 5.625 x 0.4) = 3.592 'leb = qbc = 3.592 'led = 3.592 - 1.55 Zc x A = 3.592 - 1.55 x 5.625 x 0.4) = 7.087 'ldc = 7.087 'lde

=7.087

1.55 (- 0.625 x 0.1)

= 7.184

qed = 7.184 qet = 7.184 - 1.55 x 4.375 x 0.1 = 6.504

Fig. A14.32

The results will be checked to see it static equilibrium exists relative to Z Fx and
Z Fz O.

qte qrg

= 6.504 -

= 6.504

1.55 x 9.375 x 0.2

=3.589

qgt = 3.589 'lgh 3.589 - 1.55 x 9.375 x 0.2 - 0.674

Z F z = 100 (ext. load) Z Fx = - 5 x 1.25

1.25
+

x 6.25 - 10 x 3.75 + 1.25 x 0.5 x 4 - 1.25 x 0.5 x 4 = 0 (check).

~~

5 x 1.25 - 5 x 1.25 + 5 x

=0

(check).

The shear flow force system in Fig. A14.32 causes the section to bend about axis xx without twist. The resultant of this system is 100 lb. acting down in the Z direction. The position at this resultant will thus locate the lateral pOSition of the shear center. Equating the moments of the shear flow system. about some point such as (c) to the moment of the resultant about the same point we obtain: 100 e 1.25

Fig. Al4.33 shows the plotted shear flow results. The signs of the calculated shear flows are for shear flows in the Y direction. Simple Consideration at a free body of flange member (a) will ~ive the sign or sense of the shear flow in the plane of the beam section. Thus in Fig. Al4.34 qx must act as shown when qy is positive.

= 10

x 3.75 x 15 - 1.25 x 0.5 x 2 x 5 -

x 0.5 x 2 x 10 + 1.25 x 0.5 x 2 x 15 hence e = 562.5/100 = 5.625 inches.

i ~~4
b

----

.0971

.674

3.589 11

e=6.43"

Thus the shear center lies on a vertical line 5.625 inches to right of line be. CALCULATION OF VERTICAL POSITION OF SHEAR
CENTER.

Fig. AU.33

=--=,,---'L
I
I

-----For convenience as betore, we will assume a shear load V x = 100 lb. and compute the resisting shear flow system to reSist this load in bending about axis ZZ Without twist. The resultant at this shear flow system will give the vertical location ot the shear center. The shear flow equation 1s,
q

~
Ox
a
Fig. A14.34

I I
I

1".,...l
_5.625

-I I

S. C.
Fig. A14.35

t 6.43

= _ Vx Z x
Iz

::I'

~Z 64.4

X A

1.55 Z

A14.14

BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTER.

Checking to see it Z FZ = 0 and Z Fx = 0: Z Fz = 3.592 x 10 - 3.589 x 10 .5 x .0971 - .5

x .0971

Calculation of Ix, I z and I xz --

Ix Z A z~

= .5
8

(5.333 a
8

2.667 8

5 x .674 - .5 x .674 ... .5 x .0971 - .5 x .674 - .5 x 6.504 + .5 x 7.087 = 0 (check)

1"

Z Fx

= 100

+ 5 x .0971 + 5 x .674 - 5 (7.087 +

7.184 + 6.504) = 0 (check) The resultant R of the internal shear flow system is a horizontal force of 100 lb. acting toward the left. To find the location of the resultant take moments about a point 0.5 inch below paint (c).
Re=ZIl

(5.333 ... 6.667 ) 90.667 8 I z = Z A x~ = 1 (10.667 ) + 2 (5.333 8 ) = 170.667 I xz = Z A xz = 1 x 6.667 (- 5.333) + 1 (- 5.333) (- 5.333) + 0.5 x 10.667 x 2.067 + 0.5 x 10.667 (- 5.333) The constants
I xz Iz k~,

=-

21.333

k 8 and k 3 are now deter.nlned -- 21.333 90.667 x 170.667 - 21.333 a

100 e 5 (.0971 + .674) 11 + 10 x 3.589 x 15 + .5 x .0971 x 5 - (.5 x .674 x 10) + .5 x .674 x

15 + .5 x 6.504 x 15 + .5 x .674 x 10 - (.5 x

- 21.333 15019
Iz Ix I z -

.00142 170.667 = 0.01136 15019 90.667 15019 0.006037

.0971 x 5) 100 e = 643

Iz

Ix -

I xz '

= 643/100

6.43 inches

Fig. A14.35 shows the resulting shear center

Resolving the given shear load of 141.14 into z and x components, we obtain, Vz

location for the given beam section.

EXAMPLE PROBLEM 2.

141.14

x sin

45
~

100 lb. 100 lb.

unsvnmetrr ica). Beam Section.

Fig. Al4.3S shows a tour 1'lange beam seCt ron, The areas at each flange are shown adjacent; to tlange. The external shear load V

V x 141.14 x cos 45 From equation (14) qy

equals 141.14 lb. and acts in a direction as shown. The problem is to tind the line ot action ot V so that section will bend Without twisting.

= - (k 3 V x

- k 1 Vz ) Z x A - (k. Vz - k 1 Vx ) Z z A

SUbstituting

I
a 1.0 I Z

qy = - [.006037 x 100) - (- .00142 x 100J

Load: 141.14

It-450
d

[.01136 x 100 - (- .00142 x 100)J Z z A whence

Fig. A14.36

-!,,:__
b

to

HI!

+=- '!,:5 --r

I

a" 0.5....L
c

H,a"-j

qy = - 0.7457

A - 1.278 Z

SOLUTION: -

We will start at flange member (a) where qy 15 zero on the tree edge side of the member.
~b

To solve this problem, method (3) will be used. To locate centroida1 x and z axes: =
1 x 12 + 0.5 x 8

=- 0.7457

1 (- 5.333) - 1.278 x 1 x 6.667

- - 4.544

=5.333

in.

= ~b = - 4.544 = - 4.544 - 0.7457 x 1 (- 5.333) - 1.27S x

1 (- 5.333) = 6.249 1b./in.

= (0.5 + 0.5) 16 = 5.333 in. 3

= %c = 6.249

E51

v
A14. 11 Shear Center Location By Using Neutral Axis
Method.

A14.15

qed = 6.249 - 0.7457 x 0.5 x 10.667 - 1.278 x 0.5 (- 5.333) = 5.680

~c
~a

= qed = 5.680
= 5.680 - 0.7457 x .5 x 10.667 - 1.278 x 0.5 x 2.667 = 5.680 - 3.977 - 1.704

In a beam subjected to bending there is a definite neutral axis position tor each different external plane of loading on the beam. The shear flow equation with respect to the neutral axis is, qy

o (checks free edge at d where qy

must be zero.)

=_ Vn

Zy,A

(20)

In

where, V Fig. A14.37 shows the resulting shear flaw ren = Shear resolved normal to neutral axis sisting pattern. The sense of the shear flow in In Moment ot inertia about neutral axis

V= 141.14
d

liR'
b

U
~

0"",
Ftg.A14.37a Ox a

Zn

= Distance

to neutral axis

141.14

q =5.680

qe 6.249

Fig. A14.37

the plane or the cross-section 1s determined in web at flange member (a) by the simple tree body diagram or stringer (a) in Fig. A14.37a Check Z Fx and Z Fz to see if each equals
100. 6.249 x 16 " Fz

In finding the shear center location ot an unsymmetrical section, it is convenient to assume that the Z and X axes are neutral axis and find the shear flow system for bending about each axis by equation (20). The resultant ot each of these shear flow force systems will pass through the shear center, thus the intersection ot these two reSUltant forces will locate the shear center. Example Problem The same beam section as used in the previous article (see Fig. A14.36) will be used to lllustrate the neutral axis method. Fig. A14.38 shows the section With the cerrtr-cfda.L axis drawn in, The X axis will now

99.99 (CheckS V x = 100)

=-

12

4.544 - 8 x 5.68

=-

99.94 (checks
Vz=lOO).

The resultant of the internal resisting shear flow system equals 141.14 lb. To locate this resultant we use the prinCiple ot moments. Taking pOint (b) as a moment center,
141.14 e = 8 x 5.68 x 16 - 4.544 x 30 n 212

Y 100i

+ lOO~

--l. _
5.333

1*1ll66 ~
a
I

_ _

-x

~! ~~.47
b

c
Fig. A14.39

Fig. A14.38

be assumed as the neutral axis for an external plane of loading as yet unknown. T~e 1'1111 further hence e = 141.14 = 1.50 inch. assume that when this unknown external loading is resolved normal to the X neutral aXis, that Therefore external load must act at a distance it 1'1111 give a value of 100 lb., or V z = 100. e = 1.50" tram (b) as shown in Fig. A14.37. The From the previous article Ix = 90.667. load so located will pass thrti shear center of section. To obtain the shear center location, Since the X axis has been assumed as the another loading on the beam can be assumed, and neutral axis, equat10n (20) can be written where the line ot action of the resultant of the resisting shear flow system intersects the reqy _;!:z. Z ZA, hence, lx sultant as fo~~d above would locate the shear center as a Single point. It the shear center Qab 9~~~67 x 6.667 x 1.0 -7.35 lb./in. location is desired it 1s convenient to assume a unit Vz and V acting separately and find the x 'Ibc 7.35 - 9~~~67 x (-5.333)1 = -1.47 horizontal and vertical locations of the shear center from the 2 separate shear flow force systems. qed = -1.47 - 9~~~67 (-5.333) 0.5 1.47

==-

Fig. A14.39 shows the resulting shear flow values.

A14.16

BENDmG SHEAR STRESSES.

SOUND AND OPEN SECTIONS.

SHEAR CENTER.

ZFx
ZFz

=1.47 x

16

=23.52
li

lb. 99.96 lb. (check)

tan

= 7.35x12+8xL47 =R =..; 100 + 23.52 = 103 Q = 23.52/100 = ,2352

Q

Take moments about (b) and let (e) equal distance to "esultant R.
100.8e = 3.125x1'tx5 a+3.13xSx16

lb.

e = 753/100.8

= 7.47

In.

hence

-= 13 - 16'

l0o-tS R
23.52

= 103

Let e distance fram resultant R to point b.

Fig. AI4.40 shows the position of this resultant torce. T~ere it intersects the ?revlous resultant force gives the shear center location.
A14.12 Problems

Equating moments ot resultant about (b) to that of shear flow system about (b),
lO3e = -7.35 X 6ft a + 1.47 x 8 x 16

=- 644 103 = -6.25

In.

Fig. A14.40 shows the location of the resultant. We know the shear center 11es on the line ot action of this resultant. Thus we

~.

"~~ 6.25 3.13

Fig. AH.41

(1) Fig, Al4.42 shows the crosssectlo~ ot a wood beam glued together on lines a-a and a-b. The beam Is subjected to a vertical shear V : 2400 lb. Determine shearing stress on sections a-a and a-b. Find maximum shearing stress on beam section. (2) Fig. A14.43 shows a Zee section loaded by a
1000 lb. load act1ng through

~1+1+1"1
b b

It
-.-

4"

r-f""'4c----,

r i---.L

-L
Fig. A14.43

Fig. A14.42

,:;;'7>::~-Ra1D0.8 Shear

,~.'Lr---a"';~~---'
T

1-1"1

Center
Fig. A14.40

must obtain another resultant force which passes through the shear center before we can det1nitely locate the shear center. Theretore we will now assume that the Z centro1dal axis is a neutral axis and that a resolution at the external load system gives a shear V x = 100 lb.
qy

the shear center as shown. Find the shear stress at sections a-a
and b-b by three

1000 lh. '

different methods.

L-J'l / ~1-1"-1 .

b-~-b

1r

l / 16

is. c.

-l2'I'

I I-,-

=- ~Z XA,
* - 17~~~67x

Iz

= 170.667

Qab

(-5.333)1

=3.).25
=

lb./ln.
I"

_ 100 qbc - 3.).25 -170.66 (-5.333)1 Qed

6.25 lb./1

= 6.25-1~g~66

(0.5)(10.667)

=3.13
Fig. A14.47
a =6

Fig. A14.41 shows the shear tlow results.

ZFx ZFz
R

=-6.25 x 16 =-100 lb. =-8x3.13+12x3.125 = 12.5

100.8 R=100.8

1.0

.4 d

-6
.4c Fig. A14.46

lb.
bLO

=v' looa + 10.Zli ;: tan Q = 112 = .125 0Z

12.5~

r---1O" -----j

(4) Determine the shear center location tor the beam sections in Figs. A14.47 and A14.48. As-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AB.l7

sume flange members develop entire bending stress resistance.

l~o,~/1
Fig.A14.51

q=lOi1/in.
b

Ll50"~
(5) Determine the shear center for the beam section at Fig. A14.49. Assume only the 8 stringers as being effective in bending. Area of stringers (a) and (b) 2 sq. in. each. All other stringers 1 sq. in. each.

~ r- "-i -.r
10"
I

,...

1-+ I

10"

10" _+_

20

FIg.A14.52

(7) In Fig. Al4.5l, the shell structure is subjected to a torsional moment M 50,000 in. lb. The shell skin shown dashed 15 cut out, thus the torsional moment IS resisted by the constant shear tlow on the two curved sheet elements ac, and bd. Determine the value at the shear tlow. (8) Determine the moment at the constant flow torce system in Fig. A14.52 about paint {Ole Also tind the reSUltant at this torce system.

r--l2-J-" 8 t10""
1 sq. in.

c
.5 sq. in.
I

1.5"
1~'

1. 0 0.8 1.

R1I7.5 I

liS

1.2
Fig. AB. 53

o. T>"

0.8

1.0

.8 sq. in.

20'

.1

!-7+-15"-+8-1 Fig. Ai4. 54

2a'

1----19"
Fig.Al4.50

I .5 sq. in.

(6) Determine the shear center tor the unsymmetrical beam section of Fig. A14.50. Assume sheet connecting the four stringers as ineffective. Areas of stringers shown on Fig.

{g} In Fig. AI4.53, the tour stringers a, b, c and d have the same area. Assume the webs ineffective in resisting bending stresses. Determine the distance (e) to product bending about the horizontal axis without twist. (10) For the wing cell beam section In Fig. A14.54, determine the location ot the shear center. Assume webs and 'Malls ineffective in bending.

A14.18

BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTER.

STRUCTURAL TESTING IS AN IMPORTANT PHASE OF STRUCTURAL DESIGN.

CHAPTER A 15

SHEAR FLOW IN CLOSED THIN - WALLED SECTIONS

Al5.! Introduction.

The wing, fuselage and em-

pennage structure ot modern aircraft Is essentially a single or multiple cellular beam With thin webs and walls. The design ot such structures involves the consideration ot the distribution ot the internal resisting shear stresses. This chapter introduces the student to the general problems of shear flow distribution. Chapter A14 should be covered betore taking up this chapter.
AlS.2 Single Cell Beam. Symmetrical About One Axis. All Material Effective in Resisting Bending

open sections we could start the summation at a tree surface Where q would be zero, thus the summation to any other point would give the true shear flow qy. In a closed cell there Is no tree end, therefore the value of qy Is unknown
for any point.

Stresses.

Fig. Al5.l shows a single cell rectangular beam carrying the load of 100 lb. as shown. The problem Is to find the internal resisting shear tlow pattern at section abed.

10""
I

The general procedure Is to a~ a value or the shear flow qy at same point ~d then find the shear tlow pattern tor bending without twist under the given external load. The centroid of this internal shear flow system will be the location where the external shear load. should act for bending without twist. Since the given external shear would have a moment about this centroid, this unbalanced moment must be made zero by adding a constant shear flaw system to the cell.

tor bending about the X axis lit thout twist.

gquatn on (1) g1ves the shear d1str1but10n

To illustrate we will assume qy to be zero at point 0 on the web ad.

I
I I L

qo = O.
_

The term

~=
x

100/62.5 = 1.6

/
/

'lao = -1.6 Z~zA = -1.6 x2.5 x 5 x 0 .1 = -2 1b/1n.

"

.05

qba
.05

= -2-1.6 Z~zA = _2_1.6x5x20x .05 = -10 = -10-1.6 L:~I zA=-10-1.6x2.5x5x .05

-11

-~-

.05

1---'20" --~
Fig. A15-1

x-f
z
l1 ]

qo'b
Cleo' 'led

=-11- 1.6 z~,

3

zA

=-11-1.6 x (-2.5)5 x

.05 =r -10

-10 -1.6

z~

zA = -10 -1.6(-5)20 x .05

Solution 1
qod

= -2

Due to symmetry ot material the X centroldal axis lIes at the mid-height of the beam. The shear flow equation requires the value of I X I the moment ot inertia of the section about the X axis.
Ix

= -z-1.6

z~ zA =-2-1.6(-2.5)5xO.1=0

{zX.15XIO;'+Z[30X.05X5

=62.5

Fig. A15.2 shows a plot at the shear flow reSUlts. on the vertical web the increase in shear is parabolic since the area varies directly with distance z. The intensity at qx and qz In the plane at the cross-section Is equal to the values of q found above which are In the y direction. The sense of qx and qz 15 determined as explained In detail in Art. A14.6 or Chapter A14.

in ....

From Chapter A14, the equation tor shear

flow Is 1
qy

-~ Z x

z.A - - - - -

(1)
In

This equation gives the change in shear

flow between the limits of the sunnaatn on,
Al5.!

./.

16~'

A15.2

SHEAR

FLOW

IN

CLOSED

THIN4WALLED

SECTIONS,

SHEAR

CENTER.

2 H/in.

10 JI/in.

Adding this constant shear flow to t~at of Fig. A15.2 we obtain the tinal shear tlow pattern of A15.3.

IIh~'-".~ q

=11 i1hn.
-8.34 i/in.

(11
'--!44~-

.66 If/in.

2 >/In.
10 H/in.

Flg. A15-2

1. 66 Jii/in.

-6.34 ft/in.

Fig. A15.2 also shows the resultant shear flow force on each of the four walls of the cell. The resultant shear force cn each equals the area of the shear flow diagram on each position. For example,

Fig. A15-3

It this constant shear flow at -8.34 was

~b

=( 2 ;
= 10

10 )20

= 120

lb.

not added then the external load of 100 lb. would have to be re-Iocated so that it passed through the centroid of the shear flow pattern in A15.2. To find this centroid location, we can equation the moment at the internal shear flow pattern about some point to the moment at the resultant at the system about the same point. gesuj.tant R

'<da = ~ x 2 x 10 = 6.67 lb.

Qbc x 10 + 0.667 x 1 x 10

=106.7

V ZFx '"+ Z'i'y i:l: V 100" = 100 acting down.

~

The internal shear flow force system as given in Fig. Al5.2 will now be checked forw equilibrium with given external shear loading at 100 lb. as shown in Fig. A15.l.
ZFz = 100 (external) + 6.67 - 106.67 0 (check)

Take moments about paint d. distance to resultant. \

Let x equal

RiC
100

= l:!1.:l

(internal system)

ZFx

= 120

- 120

=0

(check)

Equilibrium at moments must also be satisfied. Take moments ot all forces about poirrt d. The external load has no moment about d.
~

load would have to be moved 33.33 inches to right it the pattern at A15.2 would hold it in eqUilibrium. Since we assume q = zero at point 0, this means an open cell with the tree end at 0 would bend without twist it the external load was moved the distance

x=120 x 10 + 106.67 x 20 hence x = 33.33 in. Thus the external

x.

=120 x 10 + 106.67 x 20
; 3344 in. lb. clockwise

The stUdent should work this same problem by assuming q at some other point is zero instead ot point 0 as assumed in the above solution. Solution No.2. Shear Center Method.

Thus we have an unbalanced moment which must be made zero it we are to have equilibrium. The unbalanced shear tlow of 3334 in .Tb , can be balanced by adding a constant shear in a qounter-clockwise direction around the call. The value ot this balancing shear flow would equal,
q

In this solution, we determine the centroid at the internal shear flow system for bending of the closed section about axis X without twist. This point is called the shear center. The external shear load can then be resolved into a shear forcs acting through the shear center plus a torsional moment about the shear center. We start the solution, exactly as in solution 1, by assuming the shear tlow q a at paint O. In a sense we are cutting the cell at o and making it an open section. The resulting shear flow is as given in A15.2. This open

=- ~ = 2~0 = -8.34
= 10

lb/in.

(A equals area at call

x 20)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.3

section will bend without twist if the external shear load acts through the shear center of the open section. The closed section will be assumed to bend without twist, and the resulting shear flow pattern will be determined. The equation for angular twist Q (See Chapter A6) per inch length of beam is,
Q

6.36 If/in.

,. Q '" 23.8*

~~4~:
4.36 */in.

s. C,. - - - ..,i+III'rl;1'

.64 Him.

8-Q=23 ..

=_1_ L: qL
2AG

t'

where L equals the length

of a web or wall.
or 2AQ

=~

~L.

The right hand s ide or

Fig. A15-4

this equation represents the total shearing strain around the cell which must be zero for no twist of cell. Since G is constant, we can assume it as unity as only relative values of strain are needed in the solution. Thus the total shearing strain 6 around cell is proportional to,

o =Z

i:' - - - - - - - - - - - - - -

The location of the resultant of the shear tlow force system of Fig. Al5.4 will locate the horizontal pOSition of the shear center. Due to symmetry of the section about the X axiS, the vertical pOSition at the shear center will be on the X axis, because for bending about the Z axiS, the shear tlow would be symmetrical and thus, the resultant would coincide with the X axis.

(2)

Using the values of q from Fig. A15.2 and SUbstituting in (2),

A
'..I

Fig. A15.4 also shows the resultant shear load Q on each portion at the cell wall which equals the area at the shear flow diagram for various portions as shown.
ZFZ = - 56.9 - 43.1 = - 100, which balances the external load of 100 lb.

= ..s!o =[2 x 5 t ~ x .1
'"

10 x 5 (10 2) 20 2 .05 + .05

+ (n - 10) (.667 x 5)1 2 .05

:J

=7000

ZFx a by observation. Take moments about point 0, the intersection at axis XX and side ad.
- _~ _ 43.1x20+ (-23.8+16.58)10_ x - ZFz 100 7.90 in.

Since the section and shear flow pattern is symmetrical about the X axiS, the substitution above is written for one-halt at the cell and the results multiplied by 2. 'dhen the shear flow q varies over a portion the average shear flow is used in the above substitution. If the cell is not to twist the relative VNlst of 7000 must be cancelled by adding a constant shear flow q around the cell to give a total shear strain at -7000. The shearing strain for a constant q equals,
0= Z qL t

Hence the shear center lies on the X axis, 7.9 ft from side ad. The moment ot the external load of 100 lb. about shear center equals 100 x 7.9 = 790 in. lb. clockwise. The moment or the internal shear flow ot Fig. A15.4 is zero thus we have an unbalanced moment ot 790. Therefore tor eqUilibrium. of moments we must add a constant shear flow q aro~~d cell to develop -790 in.lb. or

=qLS [9.1

+....2. +
.05

.JLl z =-7000 .O~

=-

1.98 lb/in.

_ hence q -

7000_ n06 - -0.36 1b/in.

Fig. AlS.4 shows the resultant shear flow pattern if the constant shear tlow ot -6.36 is added to the shear flow pattern ot A15.2.

Adding this constant shear flow to that of Fig. A15.4, we obtain the tinal shear flow pattern which will be identical to that in Fig. A15.3 or the results at solution 1.
AlS.3 Single Cell - 2 Flange Beam. Constant Shear Flow Webs.

As discussed in Art. A14.9, the common aircraft cellular beam 1s made up of thin sheet

AlS.4

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR

CENTER.

wallS and webs which are stiffened by members usually referred to as flange members. A simp11fying common assumption is to assume that the flange members alone develop the resistance to the bending moment. This assumption therefore means that the shear flow ts constant between flange members. Fig. A15.5 shows a Single cell beam With two flanges at A and 8. Find the internal shear flow force system when the beam carries the external load of 100 lb. as shown.

hence q~ 15.21 Ib/in. (Sign cones out positive, hence assumed sense of q~ 15 correct). Fig. A15.7 shows a plot of the internal shear tlow resisting system. A single cell beam having only two flanges can resist only external loads which are parallel to the line AB, and thus a two flange box beam. is not used very often in aircraft structure. When the bending moment in a plane at right angles to line AB 15 small, the resistance of the curved panel to compressive bending stresse~ may be sufficient to resist such external bending mcments and thus be satisfactory.
AIS.4 Shear CeRter of Single Cell - Two Flange Beam.

;----',"'1 ,
.~

f--IO,J

V-lOO#

100

6.21

I I

.--r....

.J:: 1/2"
z

,",b A

i IO"~
Fig. Al5-6

Let it be required to tind the shear center or the beam as given in Fig. AlS.S. In other wor-ds where would the external load have to be placed so that the beam would bend without tw1st.

~/2"
q

2.05

7.95

Fig. A15-6

It is assumed that the two flanges develop the entire bending stress reSistance. This means that shear flow is constant on each web. Let q1 and q. be the constant shear flow as shown in Fig. Al5.5. The sense or direction at these shear tlows will be assumed as indicated by the arrow heads in the figure. The torce system. in the plane at the craBs-section has 2 unknowns, namely q~ and q. and thus we can solve for q~ and q. by Simple statics. To find q. take moments about point A.
~A

q constant q

Fig. A15-8

Fig. Al5-9

Fig. AlS-lO

100 x 10 - q. (2 x 80.52)

=0

hence q IOCO/Ib 1.04 = 6.21 lb/in. The Sign of q. cames out plUS, thus the assumed sense at q. as shown in Fig. Al5.5 is correct. In the above equation of moments about paint A, the moment at the constant flow system. q. about A, equals Q.. times double the enclosed area tormed by lines running tram A to theends at the web which carries the shear rlow q.. In this case the area is the area or the cell, or 80.52 sq. tn. To tind the r-enatntng unknown q1. we use the eqUilibrium equation,
ZFz

Fig. A15.8 shows the resulting shear flow system in reSisting the 100 lb. external load acting as shown in Fig. A1S.5. This shear flow system will cause the cell to twist. Therefore we add a constant shear flow q to the cell to prOduce zero twist (Fig. AlS.9). The centroid of the combined shear flow system will then locate the lateral location of the shear center. To tind q we must write an expression which measures the twist when SUbjected to the shear flows ot Figs. A15.8 and 9 and equate the result to zero, and then solve tor the one unknown q.

A=.QLu T-

o.

(Clockw1se

1s pos1t1ve)

Substituting, using the shear flows of Figs. A15.8 and A15.9,

= 0 =- 100

- 10 x 5.21 10q.

=0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AIS. S

o = _ 16.21 x 10 .04
+ 24.28q = a

6.21 x 24.28 + 1Qi .025 .04

.025

hence q : 8.26 lb. in. Adding this constant shear flow to that or Fig. Al5.8, we obtain the shear flow at
Fig. Al5.10.

of the cell, the web and wall thickness does influence the amount of twist for a given torsional load. In the shear center solution, it 1s known what portion of the shear flOw 1s due to torque or pure tw1st, and also that due to bending Without twist, which fact 1s sometimes of importance.
TORSIONAL DEFLECTION OF CELL

The lateral position of the shear center is given by the location at the resultant at the shear flow system at Fig. A15.10. The resultant R ::. ./ IF

The angular twist as gl ven by the f1nal shear flow pattern of Fig. Al5.8 equals 2QAG =- IqL/t, whence
2QAG = -6.21x24.28_16.21Xl0= -10082 __ (3) .025 .04

' x ,I

+ IF

ZFx = a
ZFz

= 10

x 7.95 + 2.05 + 10 = 100 lb.

Theretore R = 100 lb. Equate moments or R about point A to moment at shear tlow system. about; A.
Re

After finding the shear center location, we found that the external load had a moment of 1330 in. lb. about shear center, which was resisted by a constant shear flow at - 8.26 lb/In. The angular twist under this pure torque shear tlow should therefore give the same result as equation (3) above.
2QAG = - 8.26 x 24.28 _ 8.26 x 10. _ I0082 .025 .04

= l:MA

which checks the result or equation (3).

(2 x 80.52)
A1S.5 Slngle Cell-Three Flange Beam. Constant Shear Flow Webs.

100e

=2.05
= 3.30

in.
(See Fig. A15 .10) .
100 lb.

Thus shear center lies 3.30 inches to

lef't or line AB.

Thus it the given external load at 100 lb. acts through the shear center, it will produce the shear flow system of Fig. Al5.l0. However it acts 13.3 to right at shear center, hence it produces a clockwise moment at 100

-Is"
Fig. Ai5-12

1--25 lb.

13.3

equilibrium, this moment must be balanced by a constant resisting shear rlow around cell which wt i i produce a moment or -1330. The required q = 2 x 80.52 = - 8.26 lb. in. which if added to the shear rlow system or A15.10 will give the true shear floW system
or A15.8. - 1330
c'+_+~

=1330

in. lb. on the cell.

For

Area Cell 128. 54 sq. in.

Fig. Ai5.

. Fig. A15.11

ia

Thus having the shear center location, the external load system can be broken down into a load through the shear center plus a moment about the shear center. The shear flow due to each is then added to give the true resisting shear flow. It should be noticed that the web or skin thickness dOes not influence the magnitude or the shear flow system in a Single cell beam. A change in thickness, however. effects the unit shearing stress and therefore the shearing strain and thus tn computing angular twist

Fig. A154ll shows a sIngle cell beam with three flange members, A, B and C, carrying the external load as shown. A three flange box i t the flanges are not located in a straight lfne can take bending in any direction and therefore Is otten used 1n deSign because at its simplicity. For such a structure, there are six unknowns, namely, the axial load in each stringer and the shear flow q in each of the three sheet panels that make up the cell. For a space structure, we have six static equations ot equilibrium, thus a three f Lange single cell

Al5.6

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS,

SHEAR

CENT ER.

beam can be solved by statics if we assume that the three flange members develop all the bending stress resistance. thus producing constant shear flow webs. Fig. A15.12 shows the crosS section ABC. The tr~ee unknown resisting shear flows have been assumed with a positive sign. (Clockwise tlow is positive shear flow). These three unknown shear tlows can be determined by statics. To find ~a take moments about point B and equate to zero.
ms 100 x 5 - 25 x 10 + C1ca (128.54 x 2)

100

I 'lac ( I "be 'lab,

C

.- Qac

,...- -.::r A

.j.5"

1_
+
B
Fig. A15-15 Fig. A15-16

-'

Fig. AlS-14

The three unknown resisting shear flows will be assumed with the sense as i~dicated by the arrow heads. To find Qac take
~aments

~ hence 'lea - - 257.08

=- 0.972
=
0
10~b

about 9

1b!ln. 2MB

To tlnd qab take ZPz ZFz

~b

=100 x 5

- ~c (128.54 x 2)

= 100

- 10 x 0.972 1b!ln.

0 ZPx

= 1.945

lo/in.

= 9.13

To tlnd qbc take EFx ZPx

- - 15 x

0.972 - 25 - 15qbc 1b!ln.

=0

ZFz

=- 15 x 1.945 + 15Qbc = 0 qbc = 1.945 =100 - 10 x 1.945 - lO~b =0

~b

hence qbc

=- 2.639

8.055 1b!ln.

The signs of Qca and qbc came out negative,

hence the sense of the shear flow on these cell wall portions is opposite to that assumed in Fig. A15.l2. The resulting shear flow pattern is plotted in Fig. Al5.13. The student should realize the thicknesS

The algebraiC signs ot the unknown Q value all come out POSitive. thus the assumed direction of shear flows in Fig. A1S.l4 is correct. To make the cell VNlst zero. we must add a constant shear flew q to the cell (see Fig. A15.15). The relative twist under the shear flow of Figs. 14 and 15 will be equated to zero.
Z qLt

or the wall elements does not influence the

shear flow distribution if we assume the three flanges develop the entire resistance to the bending moment.
AlS.6 Shear Center 01 Single Cell-Three Flange Beam. Constant Shear Flow Webs.

=_ 1.945 x 20.71
.03

1.945x15 + 8.055x10 .025 .04

+ 20.71q + lOq + l5q .03 .04 .025

=a

Let it be required to determine the shear center location for the beam in Fig. AlS.ll. The shear center is a point on the beam crosssection through which the resultant external shear must act it the cell is to bend without twist. The shear center location ~ill be determined in two steps, first its horizontal location and then its verti~al location.
C~lculation

T~ence, q = 0.322 Ib/in. with sense as assumed in Fig. A1S.lS. Adding this constant shear flow to that of Fig. A15.l4 we obtain the shear flow system ot Fig. A15.16. The resultant R of this shear flow system is obviously - 100 lb., since the external load was 100 lb. The location of this resultant R will therefore locate the horizontal pOSition of the shear center. Equate moment of resultant R about point 8 to the moment of the shear flow system about B, Whence.

100e

of horizontal location:or e

=1.623

(128.54 x 2)

We will assume any vertical shear load, as the example, the same vertical shear as used in the prOblem of Art. Al5.4, namely, a 100 lb. load acting five inches from A, as illustrated in the following Fig. A15.l4.

= 417/100

= 4.17 in. from line AB.

(Fig. A15.16)

Calculation of Vertical Position of Shear Center A convenient

~orizontal

shear load will be

~ '~ ~~~ ;~~~'~ ",~'

;:

:',~,,/ :/~ _~'< :", \ ",:~:,;)',:,:;:c~:1 :~~: <-~~-~;.

"'., ._:~~ ':'s- ~>:~: ,:L<~;.; <~_~ "'-:~," ~-

I:,~~ :.,~.<'~ ~ ~-;':~-/,,",:r;~;'6:"~~--'~"~:

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Al5.7

assumed acting on the cell. Since we used a 25 lb. load in the example problem of Art. 15.5, we will assume the same load in this solution. Fig. AlS.l7 shows the loading and the assumed directions at the three unknown shear ~lows.

A1S.7 Single Cell-Multiple Flange-One Axis of Symmetry.

Fig. A15.20 shows a Single cell beam With ~embersJ carrying a 100 lb. shear load. The resisting shear flow system will be calculated.
8 flange

~-t:J
'lea

2S

.1!l4\~
_

Sab

qc

R:l2.1. T e
I

...
~

1.01
Fig. A15-19

Fig. A15-17

Fig. A15-18

-T-"'IIb----r ---il"
10"

S"

.40"

.10"

.10"
g

.20"

Solving for the three unknown shear flows in Fig. A15.17.

OMB = - 25 x 10 + qca (12S.54 x 2) 0 qca

__L

'l-d

=i:.

!_
Fig. A15-20

= 0.972

1b/in.

;: f.-S"+S"+S"~
The beam section which is symmetrical about the X axis 1s identical to the beam section relative to flange material which was used in example problem 1 of Art. AI4.10.
SOLUTION:-

ZF x

ZFz

=- 25 + .972 x 15 + 15qcb = 0 ~b = 0.695 = 10 x 0.972 - 10~b =0

~b

= 0.972
Assuming the 8 flanges develop all the bending stress reSistance, the shear flow will therefore be constant between flanges. Since the beam section is a closed one the value of the shear flow q at any point i s unknown, Thus we will imagine the top cover cut between flange members a and h J thus ~klng q zero in this panel due to the free end at the cut. ',,/e now find the internal reSisting shear flow system for bending at this open section about axis x-x under a external shear load V 100 lb. z

A constant shear flow q is n0W added to cell to make twi5t zero (Fig. Al5.l8). Writing ZqL/t for both loadings and equating to zero:Z L/" - 0.972x20.71

... -

.03

0.695x15 + 0.972 x 10 .025 .04

+ 20.71q + lOq + l5q

:= 0

.03

.04

.025

Whence, q = - 0.324 Ib/in. Adding this constant shear flow to that of Fig. AlS.l7 we obtain the values in Fig. A15.19. R (the resultant)

= 25

The calculations would be exactly like those in example problem 1 of Art. A14.10 and will not be repeated here. Flg~ AlS.21 shows the plotted results as recopied from Fi~. A14.32.

lb.

Equating moment of resultant about 8 to ~oment ot shear flow system about B,

Re

7 S '

100 lb.

"1
cut

=OMS

,1.25

--

25. = 0.648 (128.54 x 2)

Therefore e = 6.65 inches. Thus shear center lies 6.55 inches above B, and 4.17 inches to left of B as previously found.
Fig. A15-21

A15.8

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR

CENTER.

It ZFx and ZFz are considered tor equilibrium or external and internal loads, they will be found to equal zero. To check ZMy zero take moments about same point such as C. ZMe : - 100 x 7.5
+

10 x 3.75 x 15 - 1.25 x 5

1.25 x 10 + 1.25 x 15 = - 187.5 In.lb. Therefore to make ZMc 0, a constant shear flew q equal to M/2A : (187.5/2 x 11 x 15) = Oe57 Ib./in. is required. Adding this constant shear flow to that in Fig. A15.21, we obtain the tinal shear flow pattern at Fig. A15.22. This r fna L pattern is not much di!rerentrram that or Fig. A15.21, the reason being that the location at the imaginary cut to make q equal zero ,was not tar from the true tact, since the final q in thiS panel was only 0.57. It we had started the solution by assuming the web be cut or qbc = 0, then the correction constant flow that would be needed to satisfy ZMy 0 would have come out q = -5488, since this ts the tinal q in web be. Since the shear flow which is a load on the cell wall intluences the required thickness ot sheet required, it is good practice to try to place the imaginary cut at a point where the shear is near zero, so that preliminary estimates in routine deSign relative to shell thickness required will be based on shear flow values that are near the tinal values.

Fig. A15-23

Solution 1.

Using Section Properties and External Shears with Reference to Centroidal Axes (The K Method).

In Art. A13.8 the calculations for this beam sectton gave Ix

= 81.18,

Iz

=153.58,
z
1.0

Ixz

=- 21.33

Fig. Al5.Z3a shows the location of x and z centrotdal axes.

I
_--=:b+-_ 2.074
d 928

J:.' I I .--1
.80-1

1.82

!
5.333 I

10.667

"1 .1 .
0

.'.0

_I

Fig. AlS-23a

In Art. A14.8 ot Chapter A14, the method ot solution was referred to as the K method. The shear flow equation (see Eq. 14 of Art.
Al4.8) Is,
.68 1.82 Fig. Al5-22
A1S.8 SIngle Cell - Unsymmetrical - Multiple Flange.

1.82

qy

= -(k, Vx-k, VzI

J:xA -(k, Vz-Ie, Vx ) ZzA

-21.33 81.18 x 153.58 - 21.33' =

-21.~3

Example Problem 1 Fig.U5.23 shows a 4 flange unsymmetrical Single cell beam carrying two external loads as shown. This beam is identical to the one used in example problem 1 ot Art. A13.8 which dealt wtth bending stresses.
k ..

12016

=- . 001775

= Iz/IxI z

Ixz'

=153.58/12016 = .01279

k , : Ix/IxI z

Ixz' = 81.18/12016 : .00674

For the given beam loading the external shear loads at sectton abed are, V z

= 6000

lb.,

Vx

=-

1600 lb.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.9

Substituting in the equation tor qy as given above,

qy = - [.00674 (-1500) - (-.00177 x 6000UZxA - ['01279 x 6000 - (-.00177)(-1500UZzA

Adding this volume of q to those in Fig. A15.24 we obtain the final shear flow reSisting pattern in Fig. Al5.25. Solution 2. PrinCipal Axes Method

whence
qy

=0.16

ZxA - 73.91 ZzA - - -

In using equation (4) to compute the shear flow pattern we will imagine top panel ab cut, thus making the shear flow qy = zero in this panel. Subt. in (4) - - - ~c

The shear flow system can of course be found by referring section properties and external shear loads to the prinCipal axes of the beam section. The equation for shear flow 15 (see Eq. 16 of Chapter A14), V V qy = - ~ Z zpA - ~ xpA - - - - - (6) xp zp (The subscript p refers to principal axes.) The section properties about the principal axes were computed tor this same beam section on page A13.5 at Chapter A13. The values are:Ixp

=0 +

.16 x 1 (-6.333) - 73.91 x 1 x 6.074 449.78 Ib/in.

=-

Qed

=- 449.78
= - 100.05

+ .16 x .8 (-6.333) - 73.91

=76.38,

I zp

=169.34

x .8 x (-6.926) = - 100.05
+

.16 x .4 x 10.667 - 73.91

x .4 (-6.926) = 76.80 qba

Fig. AIS.26 which was alSo taken from page A13.5 shows the location ot the principal axes and the distances tram the tour flange members to the principal axes.

= 76.80

.16 x .6 x 10.667 - 73.91

x .5 x 2,074

=0

Fig. A15.24 shows the plotted shear flow results. This pattern satisfies ~Fz 0 and ZFx O. To check eqUilibrium of moments

~~8'j
I

6000f =V z

cut
b

Iq=449 .78 I L.J,~o~-J

Fig. A15-24

- Vx

1600=

q=75.8 Fig. Al5-26 Fig. A15-25

about a y axis. paint d.

~

Assume a y axis going through

Before substitution in equation (5) can be made, the given shear loads V z = 6000, and V x = -1600 must be resolved normal to the principal axes.
= 6000 cos 15-15' -1600 sin 150_15'

= 6000 x 8 - 1600 x 8 - 449.78 x 12

': 5367.9 lb. V x = -6000 sin 15 0-15'-1600 cos 15 0_15' , ~ -3121.8 lb.
P B nenc s -1xp

x 16 = - 61160 in. lb.

Thus for equilibrium a moment of plus 51160 in.lb. is reqUired. This is prOduced by adding a constant shear flow q around the cell 'NallS, where
q

V z

71.21
J

-19.69

=J1. 2A = 261160 x 160 = 169 . 88 (A = area of cell = 160)

Ib . in

Subt . in equation (5)

qy

= -71.21

Z ZpA + 19.59 Z xpA - - - -(6)

! t f-

Al5.10

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR

CENTER.

Assume qy
q,c

=0 in

top panel abo Solution. be used.

k

=-71.21

The K method of solution will

x 4.15 x 1 + 19.59 x (-6.74)

x 1 = -448.92

=
:0.

Qed
'ldb

=-448.92
=-

I xz IxIz-Ixz"

36.41 lS6.5x431.7-36.4l'"

- 71.21 (-7.12).8 + 19.59

(-3.58).8

=- 99.41

= 36.41 79185
k

'" 0004598

71.21 (-2.90).4 + 19.59 99.41 x 11.80 x .4 75.05 It/ln.

Iz

IxIz-Ixz II

= 431.7
79185

:=

.005452

qba

=75.65

- 71.21 (4.82).5

19.59

x 9.75 x .5 = 0 (check)

ka

Ix = 186.5 ~ .002355 IxIz-Ixz a ~

These shear flows are practically the same as obtained in solution no. I as recorded in Fig. A15.24. Discrepancies are due to slide rule accuracy. For equilibrium of moments, take moments
about (b).
Z!1b '" 6000 x 8 - 1600 x 8 - 448.92 x 12

qy = (k~VX-k.V2)ZXA - (k.Vz-k 1VX ) ZzA

Substituting in above equation,

qy ~ -( .002355 x 400 - .0004598 x 1000)l:xA

-( .005452 x 1000 - .0004598 x 400)ZzA

qy -0.4822 ;:xA -5.268 ZzA - - - - -

(7)

x 16 = -50993 In.lb.
A constant shear flow q around cell must be added to produce 50993 in. lb. tor equilibrium. This balancL~g shear flow is,

Which Is the same as in solution no. 1. Example Problem 2 Fig. Al5.27 illustrates a typical single cell wing beam With multiple flange members. The external shear load on this beam section is Vz 1000 and V 400 located as shown. x The internal Shear flow resisting pattern will be calculated.

Since the value of the shear flow is unknown at any poi~t on the cell walls, it will be assumed that the cell wall is cut between flange members 1 and 10, thus making q zero on the sheet panel numbered (1-10). Then using equation (7) the shear flaw is calculated in going clockwise around the panel. Columns 1 to 10 of Table AlS.l show the calculations in solVing equation (7). FOF explanation on how to determine sense of shear flows qy in Columns 9, 10, and 11, review Art. A14.6 of Chapter A14. The shear flow values in column 11 Would be the results if the external loads as given were so located as to act through the centroid of this shear flow force system. Since they do not we will solve for the unbalanced moment on the beam section about paint (0) the centroid of the beam cross-section. The moment of the shear flow force on a sheet panel be~Neen any two adjacent flange members 1s equal in magnitude to q times double the enclosed area formed by drawing lines from the moment center (0) and the ends of the particular sheet panel. Fig. A15.28 illustrates this explanation. The value m in column 12 of the table lIsts the double areas of these various triangular areas. Taking moments of all forces both external and internal about point (0),

ThiS beam section is the same as that used in example problem 5 of Chapter A3, where the calculations of the section properties were made. The results were:

Ix

= 186.5,
V7.:1000~

I z = 431.7 J
7.

Ixz = 36.41
8 9

_
Vx:4OOf

2"
5

p-

LX

Z!1 o '" 1000 x 2 + 400 x

s +17123 = 20323

in. lb.

10

(17123 equals summation of column 13)

1._
z

15

161

Fig. AlS-27

Thus for eqUilibrium a negative moment of -20323 is needed. This moment is prOVided by adding a

-e ,

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Table A15 1
1 2 3 Arm Z 4 ZA 5 Arm
X

AlS.11

6 XA

7 ZZA

8 ZXA
qy

10
qy

11

12 m

13

14
Final q' qxz -20.6 -20.6 -18. 53 ~ 14.70 I. 4.94 11. 99 19.31 26.79 34.29 46.98 33.16 21. 98 9.84 - 2.01 -11.43 -1 .57 -20.

Member

Area A

= ~O. 4822
(ZXAI 0 1. 175 2.089 3.758 4.205 4.275 4.017 3.432 2.470 0.318 -2.279 -3.720 -4.216 -3.7 4 -2.544 -I. 434

q =q = -5.268 .~co19 (ZZAI

col 10)

sq. in. qxzm

1 2 3 4 5 6 7 1 15 14 13 12

11
1 1

O. 14 0.14 0.38 0.17 0.17 0.17 0.17 0.17 0.29 0.3 0.31 0.31 0.31 0.28 0.17 0.17

4.396 .446 7.396 7.766 7.946 7.896 7. 96 7.296 6.896 -8.411 .8.224 -7.734 -7.004 -5.554 -4.504 -2.904

0.615 0.902 2.810 1. 320 1.351 1. 342 1.308 1.240 2. 000 -2.964 -2.587 -2.395 -2.171 -1. 554 -0.766 -0.493

0 0 -17.41 _2.437 0.615 -2.437 -13.54 _1.896 1. 517 -4.333 9.11 ~3. 462 4.327 -7.795 5.44 _0.925 5.647 _8.720 - 0.86 -0.146 6.988 -8.866 0.534 3.14 8.340 -8.332 7.14 1. 214 9.648 -7.118 11.74 1.996 10.888 -5.722 15.39 4.463 12.88 -0. 59 15.39 5.386 9.924 4.727 9.64 2.988 7.327 7.715 1. 629 3.32 4.982 8.744 2.96 -0.917 2. 13 7. 27 - 9.11 -2.551 .276 1. 25 -13.54 -2.302 0.493 2.974 -17.51 _2.976 0 0

- 3.240 7.991 -22.794

-29.748 -36.865 -43.935 ~50. 825 -57.358 - 7.894 -52.280 -38.862 -26.245 -14. 19 - 6.632 - 2.597

0 2.065 5.902 19.036 25.543 32.590 39.908 47.393 54.888 67.576 54.559 42.582 30.44 18.59 9.176 4.031 0

55.2 44.2 32.0 38.2 33.0 33.2 40.2 32.6 251. 8 46 48.2 46.4 47.6 36.6 42.6

0 114 261

OS

976 1075 1325 1905 1789 1702 2509 2052 1413 885 336 172 0 Z 17123

constant negative shear flow around cell where magnitude equals

_1'l_20323_ q - 2A - 2 X 493 - -20,6 lb./1n.

AlS.9 Two Cell-Multiple Flange Beam. Symmetrical

About One Axis.

(493

=area

of cell)

Adding this constant shear flow to that in column 11, we obtain the final shear flow in column 14. Fig. A15.29 shows true shear flow pattern.

Fig. A15.30 shows a two cell cantilever beam with 10 flange stringers. The cross-section is constant. Let it be required to determine the internal shear flow in reststing the 1000 lb. load acting as shown. For simplification, the top and bottom sheet covering and the three vertical webs will be considered ineffective in taking bending flexural loads. Since the beam section is symmetrical about the X axis, the beam will bend about this axis in reststing the given external load. The moment of inertia of the section about the X axis equals 250 in."

.11

.12

,13

,14

.15

Fig. A1S-28

-.

1199 19.

34.
7
8

Values from col. 14 Table A1S.1

Fig. A1S-30

Solution 1 (Without use of shear center)

1.98 33.16

Fig. A1S-29

The internal shear flow is statically indeterminate to the second degree, stnce the

_~It-,~

AIS.12

SHEAR

FLOW

IN

CLOSED

THIN4WALLED

SECTIONS.

SHEAR

CENTER.

shear flow at ~y pOint in each cell is ~~known. Therefore, to make the flexural shear flow statically determinate, ~ value for the shear flow q in each cell will be assumed at some paint, and the flexural shear flow for each cell will then be calculated, consistent with the assumed conditions. These resulti~g static shear flow systems Will, in general, produce a different total shearing st~ln around the perimeter of each cell, or in other words, produce a different cell twist. Since full continuity exists between cells, this condition cannot eXist, and therefore an unknown constant shear flow of q:l. in cell (1) and c, in cell (2) must be added to make the ~Nist of both cellS identical. This fact gives us the basis for one equation and the other equation necessary tor the Solution of the two unknowns q:l. and qa is given by the requirement of eqUilibrium, namely, that the moment of the external and the internal shear forces about ~~y point in the plane of the cross section must equal zero. In Fig. Al5.31 the flexural shear flow has been assumed as zero just to the left of stringer c in cell (1) and just right of stringer c in cell (2). The balance of the flexural shear system consistent With this assumption is calculated as follows:
lOf/in.

zerQ.
cc ' Is na:nely

Thus 'the shear flow in the ver-t ica t web dete~lned by the stringer c alone,

c' 'lcc ' =-4Z c zA=-4xlX5 =-201b.lin. We can now continue around cell (2) starting with stringer en where we were ,revlously stopped. _
~Idr

= qb' c '

+ 'lee' -4Z,zA c = 0 2 0 - 4 x (-5) 1 = 0

01'

x (-5) 0.5 = 10 lb./in.

qe ' e = 10

4 x (-5) 1

=30

lb./in.

Qdc 10 - 4- x 5 x 0.5 ~ 0, which checks the assumed value of q = 0 in panel cd.

The shear flows in cell (2) could of course been found by starting in ~anel cd where the shear has been ~ssl1med zero and proceeding clockwise around cell as for exa~ple

Clde
Qee'

=a

- 4- x 5 x .5 = -10 lb./in.
-30 Ib./in.

= -10 -4xSxl::a

c
50>/iD. --l'0:"

qe'd'
-30./1n.

= -30

4x (-5) 1

=-10

lb./tn.

'1i' c' .. -10 - 4x (-5) 0.5 :: 0 lb./ln.

The magnitude of the results are the same as previously calculated but the Signs are opposite. As emphasized previously the shear tlow calculated together with its Sign Is in the y direction or qy. The direction of the shear flow along the cell walls 1n the xz plane can be determined by drawing Simple free body diagrams as illustrated in Cr~pter A14 but it Is simpler to use the automatic rule of Art. A14.11. To illustrate, the solution started in panel cb and proceeded counterclockwise around cell (1). By the rule this means that the shear flows in ~he xz plane will have the same sign as qy. Since the signs of qy were negative, the d1r9ction of ~he shear tlows in the xz plane .vjLl, be counter-cjoctcvtsa around cell (1). To obtain the shear flow in the vertical web eel it was necessary to start at c and go toward c!. T~iS direction Is counter-cloc~~lse with respect to cell (2) or clockN!Se with respect to cell (1). qy for this panel from the equation was found to be -20. If we consider web cc n as part of cell (1) then the direction in solVing the s~atlon ZZA was :loskwise wi:h respect to cell (1) and by our ~Jle the shear flow in the Z direction will have the 0P9bs1te sign to qy or plus 20, which ~eRns clockwise

Fig. A15-31 The general shear flow equation is,

qy
V =---". ZzA = -

1000 ZzA = -4ZzA 250

Ix

Cell (1). Starting in panel cb where the shear has been assumed zero and proceding counter-clockwise around cell qCb
%a Saa'
~rb'

=zero
=

(assumed)
lb./in.

= -4ZzA = 0-4x5xO.S = -10

-10-4xSx2

=-so

lb./in.

= -50-4x (-5) x2;:: -10 lb./in.

qb'c' -lO-4x (-5) xO.5 = 0 lb./in. We cannot proceed beyond stringer o' because there are two connecting webs with unknown shear flows. ~e can get around this difficulty by going back to stringer c, where the shear flow on each side of c was ass~~ed

" -~,.~. ,s.'~:c .:':_: __ .,~.~:_" .. ~ . . . . . '~:.,,;

:<,:,',/: ,~,:~~._. -~~:. '_.

':~,

'.,=;,_-.

-v-:

''': __-'.'''\.'

':;"-.:.;-,,,...-',- -",".;:.:-~~i;:;

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Fig. (Al Fig. (Bl

A15.13

Since there is continuity between cells, 0:1, = Q. Also since area at each cell is the same, A, = A,. Equating (1) end (2), l533q:1, - l583q - 10840 = 0 - - - (3)

One other equation is necessary to solve tor q~ and q. , and it is given by the moments at the external and internal shear forces about any point in the plane at the cross section, which must equal zero tor equilibrium.

unknowns

Take moments. about point (b) at the shear tlow system ot Figs. A15.3l and A15.32 and also the external shear load or 1000 lb. J which in this case has no moment about our assumed moment center.

ZI1tJ:=I
Fig. A15.32 illustrates the unknown constant shear flow systems q:1, and q. which must act on cells (1) and (2) respectively to produce the same cell twist when added to the snear flow system at Fig. Al5.31. The sense at q:1, and q. has been assumed clockwise or positive in each ce~l.

-50xlOx5+ZQxlOx5+l0x30x15+Z00q 1.

+ ZOOq. = 0

hence, 200q1. +200q +3000:: a - - - - - - - (4) SolVing equations (3) and (4) tor q:1, and q., we obtain
q,

=-4.07

lb./in.

q,

=-10.80
=

lb./ln.

The tinal or true internal shear flow systam then equals r.hat at Fig. A15.3l plus that at Flg. A15.32 when q, -4.07 end q, -10.80 lb./ in., which gives the shear flow diagram or Fig.

Al5.33.
14.07
Fig. Al5-32

10.8

.8
Q

The equation tor the angular twist unit length at beam is,
L 2AGQ Zq:;:

per

-. 8

using the shear tlow values in Figs. Al5.3l and 32, the angular twist or each cell will be calculated by SUbstituting in the above equation.
For cell (1)
2.A. GQ :-10x 5 _ 50 x 10 _lOX 5 +20xlO +~+

(Values ln lb./in.)
Fig. Ai5-33

Solution Z

(By use of shear center)

1.:1,

.03

.05

-.-m-

.03

.os:

3 x lag,

.w

_l~,

. .
- - - - - - - - - - - - - (1)

- 6670 For cell (2)

In this solution, we find the flexural shear flow for bending about axis X-X Without twist. The centroid of this internal shear system locates the shear center. The moment of the external shear load about the shear center produces pure torsion on the 2 cell beam. Thus, adding the shear due to this pure torsion to that of pure bending, we obtaIn the final reSisting internal shear tlow. In bending about axis X-X without twist, the shearing strain tor each cell as given by equations (1) and (2) must equal zero. Hence;
1200 q, -333q, - 6670

=a =a
-

(5 )
(6)

4170 - - - - - - - - - - - - - - - - - - -

(2)

-333q :1, + 1250q + 4170

A15.14

SHEAR

FLOW

IN

CLOSED

THIN WALLED

SECTIONS.

SHEAR

CENTER"

Solving equations (5) and (6) for q s and qa' we obtain qa = - 2.0 lb./in. q~ : 5.00 lb./in. Therefore, taking these values of ql and qa in Fig. A15.32 and adding the results to that of Fig. Al5.31, we obtain the shear flow pattern of Fig. Al5.34 which is the shear flow system for bending without twist about X axis. The centroid of this shear system locates the shear cent er ,
In Fig. Al5.34,
ZV

the reSisting internal torque must therefore equal -3600. Therefore,

qt(l)

= .00254(-36001 = -9.17
z[

lb./:n.

SolVing for qt(2)

_ lr-ac"A, + qt (2) 31190000

auA]

lb., which checks the external shear of 1000 lb. ZH = 0 by observation of Fig. A15.34.

= 0 = -10 x 45 - 10 x 28 - 10 x 27 =- 1000
v- 1000,*
-5
8

=2 1

19.. Q
(.03

+ 10 + -lQ) 100 + .!9. x 20;1 .03 .05) ,03 ~Im= 00245T 31190000 .oj' .

hence:

qt(2)

.00245x3600 -8.85 lb/in.

Therefore, it we add to the shear flow system ot Fig. A15.34, a constant shear flow of
-9.17 lb./in. to cell (1) and -8.85 lb./in. to 28

-4

cell (2), we will obtain the true internal reSisting shear flow of Fig. A15.35, which checks solution 1, any discrepancy being due to slide rule accuracy.

-5

5
Fig. A15-34

(Values in lb./in.)

To find the horizontal position or the centroid ot the shear flow in Fig. Al5.34 take ~oments about point"a:

l:Ma = 10x27xlO +10x2Sx20 + 5 x 8 x 10 - 5 x 2 x 10 = 8600 in.lb. 8600 =. 8 6' t 0 th e rlgn-... . h ence , x = lOO6" of web aa ' .
Torsional Detlections The external shear load ot 1000 lb. acts 5 ft to the right of aa' , and therefore causes a moment about the shear center equal to (8.6 - 5.0) 1000 = 3600 in. lb. To resist this torsional moment, a constant torsional shear flow qt(l)
and qt(2) must act on cells (1) and (2)

Fig. A15-35

The angular ~Nlst of each cell is the same. The value of the angUlar ~Nlst Q per anit length ot the beam can be founo USing the shear flow pattern of Fig. A15.35 which is the true ,resultant shear flow, or the pure torsional shear flows
or qt(l) -9.17 lb./in. and qt(2)

respectively. The values of qt(l) and qt(2) can be found by USing equations (16) and (17} at Art. AS.ll of Chapter AS. Thus

-8.S5 lb./in.

may be used it desired.

'

The results Will, of course, be the same. For example: For cell (1) due to qt(l) = -9.17 Ib./fr..
and qt(2) 2GGA
1

-8.S5 lb./in.

-'t- ( = BOOO

- ZqL _ (9.17xlO) 3+9.17X10 8.S5xlO .03) .05 .03

Cell (2)
2GGA ZqL
, t

(8.85X10)3 8.85xlO 9.17xlO .03 I + .04 - .03

= 8000
Since the external torque equals 3600 in.lb.,

~ .""':c~}~:i~:,~;.':~::~,;;~'~~~~~i{fS~l\~\~-(:.~~;~,,~~~;~~k~:~!;-',}?,o:~:0~~~;~~F;-~:~;~~:';~?,}~ :'.. :~:~:~::~::~: ~~/~ :~~~:~.':,~~~:-f ::~~1:;~:.~~;': ~:~~~?{E~:~

--

.-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AlS.15

Cell (1) Final stresses - Fig. A15.35

200A _ZgL_(14.17x5)2. (4.17X5)2.54.17X10 t - ( .03 ) ( .03) .05 26.68x10 = 8000 .03

cell the same, since it any ~Nisting takes place, all cells must sutfer the same amount. Furthermore, for eqUilibrium, the moment of the internal shear flow system plus the moment of the external shear load must equal zero. For bending about axis XX, the flexural shear flow will be assumed as zero at a point just to the left of stringer a in cell (3) and just to the left and right of-stringer c in cells (1) and (2) respectively. one might consider the cells as cut at these three points Fig. A15.37 shows the flexural shear flow under these assumptions. Since the leading edge cell (3) has no stringers and the covering Is considered ineffective in bending, the shear flow will be zero on the leading edge portion since the shear flow was assumed zero just to the left of stringer a. The resulting flexural shear flow tor the 3 cell section will therefore be identical to Fig. Al5.3l and the calculations for the flexural shear flow will be identical to those in Art. Al5.7.

Cell (2) Final stresses

200A _ZgL_(1O.85x5)2. (0.85x5)2.26.68 x 10 t ( .03 ) ( .03) .03 -19.15 x 10
04

= 8000

A15.10 Three Cell - Multiple Flange Beam. Symmetrical About One Axis.

Fig. A15.36 shows a 3-cell box beam subjected to an external shear load of 1000 Ibs. as shown. The section is symmetrical about axis XX. The area of each stringer is shown in parenthesiS at each stringer point. The internal shear flow system which resists the external load of 1000 lbs. will be calculated assuming that the webs and walls take no bending loads, or, the stringers are the only effective material in bending. The moment of inertia about the XX axis of effec.lve ~terial equals 250 in ~ (Note: this beam section is identical to the two cell"beam ot Fig. A15.30 plus the leading edge cell (3).

I
I
I I

l~
lOi/in.

~ 50.JI/in. -,
laflin.

I' I I I I
l:\

---I
j~
20t/in.

10 t/in.

30./in.j

I~

10tt/in.

025 .05

.03
(II

Fig. A15-37

.04

+"--- _._-------(3)

x
10"
(2)

3'

b'

.03

c'
1(1.0)

d'

.03

e'

1(2.01 1(5) I 5"---1-5"----1- 5" -

1(.5) I(La) 5"(-1-5"--1

Fig. A15.38 shows- the unknown constant shear flows q~, q~, and qa which must be added to the flexural shear flow of Fig. A15.37 to make the twist Q of each cell the same. The sense of each has been assumed pOSitive in each cell.

Fig. A15-36

- ---r r If "I If
<1 <3

Solution No.1

(Without use of shear center)

(21

I I I
I
I

The system is statically indeter.ninate, to the third degree, since the value of the shear flow q at any paint in each cell is unknown. The value of the shear tlow will be assumed at a point in each cell and the flexural shear tlow tor bending about the XX axis will be determined consistent with this assumption. A constant unknown shear flow q J,' q a' and q , for cells (1) and (2) and (3) respectively will be added to the static flexural shear flow so as to ~ke the angular twist Q of each
Fig. A15-38

The angular twist Q for each cell equals Q =

I 2AG Z q Lit

Using the values of q in Figs. Al:.37 and Al5.38, the 'ffilue 6G will be computed for each cell.

AIS.16

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR CENTER.

Cell (3)

2QA G = ZqL e t

-12.12/F/in. -2.12*/in. -7

2.91 ill/in.

ZQ x39.4G = 10 x 50 ... 15.71 q ~ ... 10~ _ 1031. .0S ~ -;-os- .00

or Q1

I'
,

22.91 I/in. ~';

( 1) (2)

=10.5q.-2.55q.+127 = ZqL
t

- - - - - - - - (1)

-37.620/ln.

-24.97 if/in.

Cell (1)

2QA G

2Qx10OG

-50x10_2(10X5) +10x20+3(10Q.) .05 .03 .03 .03

Q:}

-12.12*/in. -2.12li"/ in. _7.0911/in. Fig. A15-39 A15.11 Shear Flow in Beam with Multiple Cells. Successive Approximations.

in.

... 10q1. _ 10q~_10qll hence -:05 --:ob:-cr;- 33.34 Cell (2)

= 6q1.-q3-1. 67 Q a
(2)

Method of

- - - - - - - - -

2QA G

=ZqL
t

2Q x 100 x G

=_ 20.03 x 10 + 2 (10 x 5) + 10 x 30 + .03 .040

.04.03

3x 10 Xqa ... 10q,_lOq; hence QJ

.63

= 6.25Q. _
- - (3)

1.67q1. ... 20.83 - -

Taking moments at the internal shear flow systems at Fig. Al5.37 and A15.38 and the external load of 1000 lbs. about stringer a and equating to zero:-

l:Ma

= 10 x20 x 10 + lOx 30x 20 - 5 x 1000 + 78.6Qe'" 200q1.'" 200Q. = 0

= 3000+7S.5Q.+200Q. +20Oq,=0 - - (4,)

The general trend in airplane structural design appears to be to the use of a relatively large number of cells. There are various reasons for this trend some of which are: (1) USing multiple interior webS, the detrimental effect at shear defo~tion on bending stress distribution is decreased; (2) the fail sate characteristic at the wing is increased because the wing is-made statically indeterminate to a high degree and thus failure of individual units due to fatigue or shell fire can take place without greatly decr-eas tng the over-all ultlmate strength of the wing; (3) the ultimate compressive strength of wing tlange units is usually increased because column action 1s prevented by the multiple webs which attach to flange members. In Chapter AS, Art. A6.l3, the method at successive approximatton was presented by determining the t: eetsutng shear flow system when a multiple cell beam was subjected to a pure torsional moment. This method at approach has now been extended to determine the resisting shear r Low when the beam is s ub jec t to flexural bending without ~Nist. USing these ~NO methods the shear flow in a beam with a relatively large number of cells can be determined rather rapidly as compared to the usual method of solving a number of equations.
PHYSICAL EXPLANATION OF THE =HOD

Solving equations (1) (2) (3) and (4,) tor the

unknown q1.' qll' qe and 00 we obtain:

q.
Q. Q.

= - 2.12 = - 7.09
-14,.5

lb./in. 1b./in.

1b./in. = Q1 = - 19.9

Adding these constant shear tlows to the flexural shear tlow of Fig. A15.37, we obtain the true internal resisting shear flow as shown in
Fig. A15.39.

Fig. AlS.40 shows a 3-cell beam carrylng and external shear load V acting through the shear center of the beam section but as yet unknown in location. In other wordS, the beam bends about the symmetrical axis X-X without twist. The problem is to determine the internal reSisting shear flow system for bending without
''The Analysis of Shear Distribution lor Multi-Cell Beams in Flexure by Means of Successive Numerical Approx:i~ mations." By D. R. SAMSON. Journal of the Royal Aeronautical Society, Feb. 1954.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.17

twist. In this example, it is assumed that the bending moment is resisted entirely by the flange members as represented by the small circles on the figure, which means that the shear flow will be constant between the flange members. The first step in the solution is to make the structure statically deter.ninate relative to shear flow stresses for bending Without twist. In Fig. A15.41 imagine each cell cut at points a, band c as shown. For the given shear load V, the static shear flow qs can be calculated, assuming the modified section bends about axis X with nO twist. Fig. A15.41 shows the general shape of this static shear flow pattern.

-x

~ll (1)- -

Cell
(2) -

Cell
-(3) -

-X

Fig.A15-40

% -q;a
e-,
b

Now consider each cell as a separate cell. The static shear flow qs acting on each cell will cause each cell to ~Nist. Since zero ~Nist is necessary a constant shear :lo~ q~ to cell (I), q~ to cell (2). and q~ to cell (3) must be added as shown in Fig. A15.42. and the magnitudes of such value as to make the ~Nist of each cell zero. However. the cells are actually not separate but have a common web between adjacent cellS, thus the shear flow q~ acts on web 2-1 which is part of cell (1). and thus causes cell (1) to twist. Likewise cell (3) is twisted by q~ and cell (2) by both q~ and q~. Therefore to cancel this additional cell twist. we must add additional constant shear flows q:, q; and q~ as shown in Fig. A15.43, and considering each cell separate again. However, since the cells are not separate these additional shear flows effect the twist of adjacent cellS through the common web. As before this disturbance in cell VHist is again cancelled or made zero by adding further closing shear flows q~. as shown in Fig. AI5.44. This procedure is repeated until the clOSing shear flows become negligible. In general the converging of this system is qUite rapid and only a few cycles are necessary to give the desired accuracy of results.

q:, q:

Fig. A15-41

--'Is

The total closing shear flows q:.. q" and q, are then equal to -

". ~--t
I

hi \

:<I'

'>':!

1- -I 1"2 (21 I ___t

~

--'Is 'Is

q. q. q.

= q~ + q~ + q~ + = q~ + q: + q: +
= q~
= Qs
+ q; +

q: +
pane I

".

The t 1nal shear flow on any equals, (See Fig. A15.45)

q

then

+ q:. + q" + qs - - - - - - - - - (1)

The centr01d of this final shear flow system locates the shear center of the section, relative to bending about the X axis.
DERIVATION OF EQUATIONS FOR USE IN SUCCESSIVE APPROXIMATION METHOD

Fig. AI5.46 shows cell (2) or the 3-oe11 beam shown in Fig. A15.45. qs is the static shear tlow and ql' ~ and ~ are the redundant or unknown shear flows. Since "2 (1) r ~ cell (2) does not twist (2) under these shear flows 11 we can write in general,

"s --II

l~ = 0

II

- - - - - -

(1 )

Substituting the var-i-, Fig. A15-46 ous shear flows on cell (2) in Fig. A15.46 into eq, (1).

II --".

A1S.18

ANALYSIS

AND

DESIGN

OF

AIRPLANE

STRUCTURES

Z q - - q Z -... qJ..z -"'q3 Z-- 0 - - - - (2) " s t a" t :1._ t 3_t

L _

and thus equation (7) becomes,

The SUbscript (a) on the summation symbol tmplies summation completely around cell (2) whereas the subscripts :1.-8 and 3-a implies summation only along webs 1-2 or 3-2 respectively. L is the length of a sheet panel and t its thickness.

Thus by repeating the above procedure, a power series of the carryover influence factor is obtained. In general the convergency is rapid and only a relatively few cycles or operations are needed for sufficient accuracy for final shear flows. A solution at a problem wll1 now be given to show how the necessary operations form a very sim?le routine.
A1S. 12 Example Problem Solution. p r-cblem No. 1.

- - -

(3)

The first term in equation (3) represents the proportion or the static shear flow qs which must act as a constant shear flow around cell (2) to cancel the twist due to qs' The resulting value of this first term will be given the term q~. The second and third terms in (3) represent the constant closing shear flows required in cell (2) to cancel the twist of cell (2) due to the influence of qJ. and q~ in the adjacent cells acting on the c~~on webs between the cells. The ratio in equation (3) before q. will be referred to as the carryover influence factor fram cell (1) on cell (2) and will be given the symbol C1.-'" and the ratio before q~ in equation (3), the carryover influence factor from cell (3) to cell (2) and it 1"1111 be given the symbol C:s_:a' Thus equation (3) can now be written as,
(4 )

Fig. A15.47 shows a cellular beam with five cells. The flange areas and the. web and wall thicknesses are labeled on the figure. The prOblem will be to determine the internal shear flow pattern when resisting an external shear load of V z = 1000 Ibs. Without twist of the beam. Having determined this shear flow system the shear center location tallows as a Simple matter. Fig. A15.48 shows the ass~ed static condition tor determining the shear flow system in carrying a Vz load of 1000 lb. without ~Nist. The static condition is that all webs except the right end web have been imagined cut as indicated thus making the shear flow qs at these points zero. In this example problem it will be assumed that the flange members develop all the bending stress resistance, which assumption makes the shear flow constant between adjacent flange members. The total top flange area equals 5.5 in. II , and also the total bottom flange area. Due to symmetry the centroidal X axis lies at the mid-depth point. Hence, Ix
0_

As explained above, q~ is the value of the necessary closing shear flow tor zero twist when the adjacent cell shear flows are zero. Hence first approximations to the final shear flows in each cell can be taken as neglecting the effect of adjacent cells, or in other words each cell is considered separate. Hence the first approximations are,
q, q.

= (5.5x5~)2 =275

in."
l:zA

= q. =
qll

= -1;;" Vz ZzA -

1000 ZzA = -3.636 - 275

q~ q~

(5)

q~

Starting at the lower left hand corner, the static shear flow Qs will be computed going counter-clockwise around beam.
Qab qed

By SUbstituting (5) in (4) a second approximation for q~ is obtained, namely,

=- 3.636(-5)2 = 36.36

lb./ln.

qbc = 36.36 -3.636(-5)1

q~ + C._ q~ ... C:!-11 q~

q~ + q: - - - -

- - - - - - (6)
(7 )

=54.55 lb./ln. =54.55-3.636(-5)0.5 =63.64

where q: is the correction added to the first approxi~tion. In a similar ~nner corrections q~ and q; are made to the approximations for q:l. and q3 T~erefore as a third estL~ate for q3' these ~urther corrections should be added

Continuing in like manner around the beam, the values of qs as shown in Fig. Al5.48 would be obtained. The solution from this point onward is made in table form as shown in Table A15.2 which should be located oelow a drawing ot the

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.19

cellular beam as illustrated, and the numbers in the Table should be lined up with respect to the cells as indicated. The solution as presented in Table A15.2 Is carried out in 17 simple steps. The first step as given in row 1 of the Table is to compute for each cell the value for L: ca J. where sheet panel taken
t qs is the static shear flow on each

C a-l.

=(~L =.9. = .2336 L: L 856

.1:

panel of a cell; L the length of the and t its thickness. Values for qs are from Fig. A15.48. For example, tor CellI
Z

We are now ready to start the solution proper by successive apprOXimations. In row 5 of the Table, the first apprOXimation is to assume a value q' added to each cell which will cancel the twist due to the static shear flow when the cell webs are not cut, but each cell is considered separate or independent of the other. This constant closing shear flow ql equals,
ql

qg ~

= 2(36.36

x 10)

.~

= 18180

zqgt"
Z.b.
t

The minus Sign is necessary

The sign is positive because Qs Is positive. (Clockwise shear flow on a cell is positive.) Row 1 in the Table shows the values as calculated tor the 5 cells. The second step as indicated in row 2 of the Table 1s to calculate the value of the expression L: Lit for each cell. For example, for CellI,

because the twist under the static shear flow must be canceled. The values for q' are recorded in row 5 of the Table. For example for cell (1),
,18180 q--856_ 21.238

For cell (2),

qI
3

_---gsa27275 _ _ 28 . 71

z...h.=...lQ....
1.

.064

(10)2. 10
.04

.05

= 856
Steps 6 to 13 as recorded 1n rows 6 to 13

For cell 2,

at the Table are identical in operation, namely, the carryover influence tram one cell to the

z...h.=
t

-:05

10.2(10)
~

.12.. = 950
.04

The third step as indicated in row 3 is to calculate the value for the Lit of the common web be~Neen two adjacent cells. For example, for web bb' between cells
(1) and (2),

adjacent cell because of the cammon web between the cells. As a closing shear flow 1s added to each cell to make the cell twist zero when they are considered separate, this result is continually disturbed because at the crnnmon web. Gradually these corrective shear flows become smaller and smaller until the cells reach their true state and possess zero twist. In the Table, arrows have been used for two cycles to help clarity the operations. For example in row 6, the carryover shear flow from cell (1) to cell (2) is,
- 6.700 x .2105
1s

f\t

L)

lo_

=..l.. =200 .05

The fourth step is to determine the carryover factor from one cell to the adjacent ~ell. The results are recorded in row 4 of the T2.ble.
expla~ation.

=- 1.414 =- 3.000 =3.215

From cell (2) to cell (1), the carryover value

(-4.480 - 8.330) 0.2336

Referring to equation (3) for general the carryover factor from cell (1) to cell (2) is,
200

From cell (2) to cell (3),

(-4.480 - 8.330) 0.250

3

Cl._<;J ..

950

.2105

and the carryover factor from cell (2) to cell (1) equals,

A15.20

SHEAR

FLOW

IN

CLOSED

THIN~WALLED

SECTIONS.

SHEAR

CENT ER.

,
Fig. A15-47 Flange and Web Data

b'

d'

.04
10" X .064 Cell (I)

.04 .05 (2) .04 .04

.04 (3) .04

.03 (4) .0'

.5d

"

,
.03 (5) .03 .03

.03

.04

.04 .5 c 5 Cells at 10" == 50"

.5 e
-------------' 72.73
ut

Fig. A15~48 Assumed Static Condition ~or Shear Flow qs

t,
0

~'

36.36

Cell
(I)

36.36

1
I

54.55

c'
ut
0

63.64

d'

ut

(2)

(3)

( 4)

r
To
e

81. 82

f '

,
,

ut

(5)

roo]

54.55

63.64

72 73

81.82

Table AIS. 2

Row
I. 2. 3. 4. 5. 6. 7. 8. 9. 10.

OPERATION

Z q" L!t for each cell

Z L!t for each cell L! t of cell web Carn OVer Factor (C) 1st Approx, q' == -EQs Lit!!. Lit q" == Cq' (Carry over) q" : Cq" (Carry over) ETC (Carry over) Carry over Carry over Carry over Carry over Carry over q.q'~q".q" ---) 1st Reiteration q' from row 5 q 2nd Reiteration q' from row 5 q

18180 856

27275 950

31820 1000

I
I

48487 1250

I
I

87880 1333

11.
12. 13.

200 250 333 250 .250 I . 2633 . 200 I .250 .2501.266 .2105 I .2336 ~31. 82 .... -38.79_ ~65. 910 ~21. 238...... _-29.7" - 6.700~-4.480 -8.38~7.170 -9.70~6.364 -17.560~- 9.700 - 3. 000- ...... -1. 414 -4.430'" ....3. 216 ~5. 975 ......- 3. 372 -2. 580~ ..... - 5.975 -0.631 -2.518 -1. 460 -1. 489 -1. 837 -1. 592 1. 489 - 1. 362 0.736 -0.287 ~O. 777 _0.787 _0.872 -0.590 ~O. 396 0.872 -0.155 -0.436 _0.266 _0.246 -0.332 -0.232 0.246 - 0.248 0.141 - O. 138 -0.052 -0.135 -0.148 ~0.141 ~0.lO2 -0. 065 0.044 -0.029 -O~076 _0.047 -0.042 ~0.O58 ~O. 037 - 0.045 0.003 -0. 009 -0.003 ~O. 003 ~O.OO3 -0.002 ~O. 012 0.003

14.
15. 16. 17. 18. 19.

-33.47_ -12.27 -21. 238 -33.51..... 12.23'-21~ 238

20.

-33.'7

~63. 39 .... -73.92_ -84.38 -52.52 .... -. -7.05 I -16.690><:13.13 1-18.4a>-<:12.68 1-22."'><--'-16.46 ~65. 91 -28.71 -31. 82 -38.79 ~73. 91 __ -52~ 45 ....... -63.43 .......... .......~84. 39 ---7.05 1-16.70>-<:13.10 1-18.'&"-12.69 I -22.'4~-16.'6 ~28. 71 _38.79 -31. 82 ~65.91 -52.46 _73.92 ~84. 39 -63.40

Fig. A15-49 Closing Shear Flows to Make Twist of Each Cell Equal Zero

1---1 r~ --t~----tj----tfl--I

f I

q'33.51

II

It

q=52.45

II II

q=63.43

II II

q=73.91

II It

q.64.39

----~~-- - tl--- tj-2.65

If

--2.57 10.48 15.61

l
Fig. A15-50 Final Shear Flows. (Fig. A15.48 plus Fig. A15.49)

t
x
.;. 84.391, hence leOoX

f""

2.10

0.27

1. 18

1
I

16 94 . 2.10 ,
0.27

f 1
t

10 46 .
1. 18

2.85

2.57

CALCUL'.TION OF SHEAR C&\'rER LOCATION

In Fig. A15.47 let = distance tram lett end ot bea~ to shear center. Taking moments about upper left corner of the shear flow forces in Figs. A15.48 and A15.49 and equating to 1000 X.
lOOoX = 10(36.36+54.55+53.64+72.73+81.82 )10 ... 100 x 10 x 50 - 2 x 100(33.51 + 52.45 .;. 63~43 + 73.91 = 19472 or 19.47 tnchee ,

x=

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.21

In row 13 of the Table, the carryover values are 50 small that the process Is terminated. The final constant shear flow that must be added to each cell to cancel the twist due to the static shear flow equals the algebraic sum of the values from the beginning row 5 to row 14. The results are shown in row 14 of the Table. The results in row 14 are obtained after a considerable number of multiplications and additions of numbers, thus it is easy to make a numerical mistake. To check whether any appreciable mistakes have been made, we take the values in row 14 and consider these values of constant shear flow in each cell as that causing zero twist if cells ~~e separate. Then bringing the cells together, through the common webs causes a disturbance in tHist and this is made zero by the carryover values. This step in the Table is referred to as a reiteration and is indicated in row 15. Then adding the values in row 15 to the initial approximation ql in row 5, which value is repeated in row 16, we obtain the final value of q in row 17. The values in row 17 are practically the same magnitude as in row 14, thus no appreCiable mistakes have been made. If the difference was appreCiable, then a second, and if needed, even a greater number of reiterations should be carried out. In the Table a second reiteration is shown in rows 18, 19, 20 and the results in row 20 are practically the same as in row 17. It will be assumed that the solution was stopped after first iteration, and thus the valueS in row 17 are the constant shear flows that must be added to the static shear flows to produce bending without twist. Fig. AlS.49 shows these final closing constant shear flows. Adding these values to those in Fig. A15.48 we obtain the final shear tlows in Fig. AI5.49. The lateral location of the shear center for this given 5 cell beam coincides with the centroid of the shear flow force system in Fig. Al5.50. The calculations for locating the shear center are given below Fig. Al5.50. Solution 2 of Problem 1 In solution 1, the assumed static condition involved cutting all vertical webs except the right end web. Thus the static beam section became an open channel section and the resulting static shear flows must obviously be far different t~ the final true shear flow values, since the webs always carry the greater shear flows in bending without twist. This fact is indicated by the relatively large number of steps required in Table A15.2 to reach a state where successive corrections were small enough to give a desired accuracy of final result. Thus it is logical to assume a static

condition where the static shear flows in the webs should be much closer to the final values and thus hasten the convergency in the successive approximation procedure. Thus in Fig. A15.51, we have assumed the top panel in each cell as cut to give the static condition. The static shear flow is now confined to the vertical webs and zero values for top and bottom sheet. Table A15.3 shows the calculations tor carrying out the successive approximations and needs no further explanation. It should be noticed that atter the first approximation was made in row 5, only three carry over cycles were needed in rows 6, 7 and 8 to obtain the same degree at accuracy as required in 8 cycles in Table A15.2 for solution 1. Fig. A15.52 shows the final shear tlows which equal the constant shear flows in each cell fram row 9 of Table added to the static shear flows in Fig. A15.51. These values check the results of solution 1 as given in Fig. A15.S0, within slide rule accuracy. In Table A15.3 no reiteration steps were given. The student should make it a practice to use such checks.
A15.13 Example Problem 2. All Material EUective in Bending Resistance.

The general trend in supersonic wing structural design is toward. a large number ot cells and relatively thick skins, thus in general, all cross-sectional material or the wing is effective in resisting bending stresses and thUS the shear flow va!ies in intenSity along the walls and webs at the beam cells. Fig. Al5.53 shows a ten cell beam with web and wall thicknesses as shown. It will be assumed that all beam material is effective in bending. The shear flow resisting system for bending about the horizontal axis without ~Hist will be determined. The centroid of this system will then locate the shear center.
has

Fig. A15.54 shows the static condition that been assumed. namely, that the upper sheet panel in each cell has been imagined cut at its midpoint, thus making the static shear flow zero at these points. The static shear flow values Qs are shown on Fig. A15.54. To explain how they were calculated, a sample calculation will be given. The moment of inertia at the entire crosssection about the horizontal centroidal axis lS, Ix for top and bottom skins, .. 78.0 = (50 x 0.125 x 2.5')2 Ix of all webs = (0.912x5~)/12 = 9.5 Total Ix ::; 87.5 in" For convenience an external shear load V ::; 8750 lb. will be assumed ac~ing on this z beam section.
Hence, q

=-~ZzA =_875 9 ZzA =-100 Ix 87.~

~zA
I
~ .
,. .......1

.)

AIS.22

SHEAR

FLOW

IN

CLOSED

THIN_WALLED

SECTIONS.

SHEAR

CENTER.

Now consider Fig. A15.55 which shows a sketch or cell (1) plUS halt or cell (2). As previously explained the upper cell panels were assumed cut at their midpoints (a) and (ml. Solution II out
I

Starting at point (al in cell (1) where the shear rlow is zero and going counter-clockwise around the cell. the static shear flows are as follows: out out out out

Fig. Al5-S1 Assumed Static Condition foE" Shear Flow eu and ReIlU1t1ng lis Values.

0 0 I 0 '~ 0 0 0
(1)

36.36

18.18

(2)

9 09

(3)

9.09

(4)

10.09

01
(5) 18.18

1
I
2633
0 0 1000

I
I
1

Table A1S. 3

Row

OPERATION -2045

1. E qs Lit for each cell 2. Z t for each cell 3. L t of cell web 4. Car over factor C 5. 1st approx. ,= qg L/t/Z Lit 6. "_C I (carry over) 7. ", =C .. (carry over) 8. ETC car over

"8

-1362 950

758
1250

I
I

3031 1333

-z

9.

..". .... --"

200 250 .2105 \ .2336 .250 I. 1.434 2.389 0.503 0 0.334 0.118 0.071 .055 O. 029 0.025 _.007 2.870 2. 081

250 .200\ .250

O. 358 _.152
0.126 _.152 0.031 0 0.211 0.211

2.87 Fig. AlS-52 F1na.l Shear Flows (Row 9 Plus Ftg. A15-S1) Compllu'e Results with Solution I (See Flg.AlS-SO)

2.081

333 .250 .266 _0.606 0 -.605 0.041 -.041 -0.005 -.040 -1.256

-2.274 -0.152 _0.152 0 _2.578 2.5 S

1.206

33.49

18.96

10.98

10.55

f
1.256

10.41

15.60

2.87

2.0S1

0.211

2.578

Example Problem 2 Ten Cell Beam - All Material Effective in Bending.

Top Skin -

Fig. Al5-53

.
g
78
_7 8

(1)

.
~

(2)

..
~

(3)

.=
~

-.

125 Inches Thick

(4)

.
g

(')

..

i!:

(')

= ~

(7)

;;;

(S)

" ~
~

(9)

" 0

(10)

8T 5"

Bottom Skin - 125 Inches Thick

Fig. A15-54

_9 _7

-100 - - -1" ".. " -0- -0-0a

-_78

-rse

78 7So 78 7S o 78 78 078

7S o

-m

'-.156f18S.4f -7878

7878 0 78 78 0 7878 0 7S 7S o .,..156 181.3

_

-1" - -15"
178.5
_0 _

I--- 5"--1

185.

185.4

IS5.4

181.3
0 _

- ;7~5~f
_0

7878 a 78

0_

_0 _ _ 0_

_0

-7878

_n78

-~m

.78m
1350 195 \

-7878

.78m

Table AIS. 4
2;0,

Lit

ZL/t cell Lit web

C q' =-EQsL/t/l: Lit

C.O. C.O. C.O. C.O. C.O.

q

Reiterate

I,

U90 0 1 -4760 211.3 219 227. '7 53.2 53.2 53.2 53.2 53.2 61. 8 61. 8 69.5 69.5 .285 252 .285 .28S .285 .285 .285 .28S .285 .285 .304.316 .292 .304 .317 .329 .306 .317 -10.42 0 _5.63 -6.92 0 20.91 0 0 0 0 ..2.980 _1.7 .63 0 0 0 1.97 0 -2.1 1.71 00 0 -4).860 ~.75 0 ..0.57 0 0 0 -1.11 1.60 1.48 ".5 6-1.20 ..(l.21 .25 .(1.53 0.15 0 00 0 0.1 0.47 0 ".2 0 ".1 ..0.50 .. Q 11 .00 -0.1 3..o.07..oJ.0 ..0.19..0.070 0 0 ~ll 0.20 0.19 ".1 90.l! ..0.00 ..0.03 ..0.08 ..0.0 ..{l.O7.{1.0 ..0.02 ..0.07 ..0.08 -'J.O a ..0.09 .03 00 0.0 0.06 0 _3. 53 _11.28 -1. 29 -1.14 -2.8S -9.00 -5. as 5.56 22.58 -<l.2S ..0.89 -3.21..{).36 -1.01-0.32 -Q.38..o.81 ..0.3 -2.56..0.8 -1.34 -2.74 1.53 -1.241.S3 -1.6 7.15 1.70 -10.42 0 0 0 _5.63 -6.92 0 0 20.91 0 _3.57 -11.31 -1. 33 _2.8S -1.17 -9.07 -4.27 -5.04 5.55 22.61 2200 211.4
0 186.4 0 IS6.ol 0 186:4 0 186.4 0 203.6

".1

89 31

ee.es
17'1'.66

.JH.57 74.H .79.33 76.67 7".17

".
187.11 78.33

15.12_87.07 ~8.93 ~2.Z7 13.73_~3.04 7~9

- .,-

83.55 _55.39

[00,61

109.3 t

183.16

185.24

191.59

176.SU

182.07

167.9\

181.H

120.51

Fig. Al5-S7.

Flnal Shear Flow Values. (Note: Shear Flow at Ends of Webs Equal Sum of Shear Flows in Adjacent Sk1n Panels. )

- ."

ee.es 81.'

--

14.4' .79.33

76.87.79.17

- -

.so.se

rs.ra .a7 ,~7 5R.93 aZ.27 .3.13 _d3.04 ;2.96.72.45

- -

--

U.S5 _55.39 100.61

- -

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A1S.23

cut

Procaading to (g)
h

11
Qa

T XT
5"

qg = qhg+100Z~ = 156+100 (1.25x2.5x

h

.094) = 156 +29.4

c - - (1) - - - - g J - - ( 2 ) - -

= 185.4

1b./in.

rZA qrg = 185.4+10OZ = 185.4+100 (-1.25) g (2.5 x .094) = 156 1b./in. qrk = qrg - qra -156 + 78 = -78 qkr =-78-100Zl<zA =-78-100 (-2.5)(2.5x

Fig.Al5-55

=0
a

qb =_100Z

bZA

=-100x2.5x2.5xO.125

0.125) =-78+78 = 0

= -78 1b./in. qc = qb-100ZcZA =-78-100 (1.25x2.5

b

Fig. A15.56 shows a plot at these calculated values. The arrows give the sense ot the shear flows.

x 0.064) =-78-20 =-98 1b./in. qd

8 ~7~8~)IDjjjJiIi;;;~~~~~~~iiiijjiii;;=

=-98-zdu
c

=-98-100 (-1.25)(2.5x

0.064) = -98 + 20 = -78 qa =_78_Z aZA =-78-100 (-2.5)(2.5 x

d

I
98

78
(2)

0.125) = -78+78 = 0 qra O_ZrZA = 0-100 (-2.5)(2.5 x 0.125)

Fig. A15-56

78

=78

At point (f) there are two other connecting sheet panels so we cannot proceed past this joint in calcUlating the shear flow in the two connecting sheets at (f). Thus we go back to point (a) and go clockHlse,
Qa

Fig. A15.54 shows the calculated static shear flow values for the entire 10 cells. The "falues are recorded at the ends of each sheet panel and at the midpoints of each sheet panel. Clockwise shear flow in a cell is positive shear flow. Since an interior web is part of two adjacent cells, the Sign of the shear flow on vertical webs is referred to the left hand cell in order to determine whether sense is positive or negative. Having determined the static shear flows which will be referred to as qs' we can now start the operations Table Al5.4. The first horizontal row gives the calculations of the twist at each c ell under the static shear flows, which 1s relatively measured by the term ZQs ~ for each cell. With all mater1al effective in bending the shear flow varies along each sheet. Fig. A15.56 shows this variation on the sheet panels of cell (1). The term Zqs Lit is nothing more than the area of the shear flow dia~am on each sheet divided by the sheet thickness. To illustrate,
consider cell (1) in Fig. A15.56.

qha = 0+100zhzA = .0+100 (2.5x2.5xO.125)

= 78

1b ./in.

With two other webs intersecting at joint (f) the shear flow summation cannot continue past (r), hence we go to paint (m) in cell (2) where shear flow is zero due to the assumed
cut at point 1m).
~

hZA qhm = 0_100Z =-100(2.5x2.5xO.125)

= -78

zn

Now at joint (h) we have the shear flow at 78 magnitude on each top panel, thus the shear flow in the vertical web at (h) equals the sum at these two shear flows or 156.

Upper sheet panel: Zq

...b. =
s t

-(0 + 78) .....&... +(0 + 78) 2.5 = 0 2 0.125 2 0.125

Al5.24

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR

CENTER.

For lower sheet panel: ZQs ~

listed in last row in Table A15.4: M = +(-11.31-3.S7 -1.33 -1.17 - 2.88 - 9.07 -4.27-S.04+S.SS+22.51) 2x25 =-S20

(Same figure as for upper sheet)

For lett hand vertical web: Treating the shear flow diagram as a rectangle with height 78 and a parabola with height 98 -78 = 20,
l:<ls.k=-78XS _(20xSXO.667) =-7142 t 0.064 0.064 For right hand web or cell (1),

Total Moment 175000 + 22520 - 520 = 197000 hence X: = 197000/8750 = 22.5 in , , which equals the distance from left end of beam to shear center location. Referring to the final shear flow values in Table A15.57, it will be noticed that the final results are not much di:ferent from the assumed static shear flows with the possible exception of the two end webs. If we had assumed all the webs cut except one to form the static condition, then Table AlS.4 would have required several times as many carr-y-over- cycles to obtain the same accuracy of final results.
Al5. 14 Use of Successive Approximation Method for Multiple Cell Beams when SUbjected to Combined Bending and Torsional Loads.

The shear flow diagram is likewise made up at a rectangle and a parabola.

zn_J..= lS6XS+ (186 S-lS6)S 0.667 -9342 '" t 0.094 x 0.094 -

Therefore for entire cell (1)

Z% ;

= -7142 + 9342

2200, whi ch is the

value in row (1) or Table A1S.4 under cell (1).

The results for the other nine cells as calculated in a similar manner are recorded in row I ot the Table. The procedure as followed in the remaining rows or Table 4 is the same as explained in detail for example problem 1. In Table Al5.4 only one reiteration is carried through as the values in the bottom or last row are practically the same as arrived at atter the firth carryover cycle. Adding the constant shear tlow values in the last row in the Table to the static shear tlow values in Fig. liS.54 we obtain the tinal shear flow values of Fig. Al5.57. The resultant of this shear flow pattern is a force at 8750 acting dawn in the Z direction. Its location would be through the centroid at the shear flow force syatem.. Let X equal distance from upper lett hand corner of beam to line of action at shear flow resultant~force.

The internal shear flow reSisting force system for a beam subjected to bending and twisting loads at the same time is carried out in VNO distinct steps and the ~esults are added to given the true final shear flow system. First, the shear flow resisting system 1s found for being without vnist as was explained in this chapter. The results of this first step locates the shear center. The external load system- is then transferred to the sheat center, which normally would produce a torsional moment about the shear center. The internal resisting shear flow system to balance this torsional moment 1s then handled by the successive approximation method as explained and illustrated in detail in Art. A6.13 or Chapter A6.
A15.15 Shear Flow in Cellular Beams with Variable Moment 01 Inertia.

The previous part of this chapter dealt entirely with beams of constant moment of inertia along the flange direction. In airplane Taking moments of shear flow torce system wing and fuselage structures, the common case of Fig. Al5.54 plus the constant shear closing is a beam at non-uniform section in the flange values in each cell as given in the last row direction. In cases where the change of the or Table Al5.4 and equating to 8750 i, we crosB-sectional areas is fairly well distributed obtain; between the various flange members which make up the beam cross section, the shear flow results Due to uniform static shear flow on each as given by the solution for beams ot constant Web: moment of inertia are not much in error. For beams where this is not the case, the shear flow M = (lS6 xS) (S + 10 + lS + 20+ 25 + 30+ 35 results may be considerably different from the +40+45) +78x5x50 = 175000 actual shear flows. This fact will be illustrated later by the solution of a few example Due to parabolic static shear flow in each web:problems
M = (29.4xSxO.667)(S+10+1S+20+2S)
+ 25.3x5xO.667(30+35) +22.5x5xO.667
AlS. 16 The Determination of the Fle:mral Shear Flow Distribution by Considering the Changes in Flange Loads. (The 6,P Method. )

(40+45) +20xSx 0.S67xSO = 22520

Due to constant clOSing shear flows as

Fig. A1S.58 shows a Single cell distributed

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.25

flange beam. Consider the beam acts as a cantilever be~~ with the bending moment existing at section (A) being greater than that existing at secti~n (B) and that the bending moment produces compression on the upper surface. By the use of the flexural stress equations, the bending stress on each stringer can be found, which if multiplied by the stringer area gives the stringer axial load. Thus at beam section (B), let p~, P~, P3 , etc. represent the axial loads due to a bending moment M. The external bending

flange members, see Fig. A15.60. whence

_
~P3 + ~P~ +

qyd

Again ZFy

=0,

orqy--

(6P,. 6P.)

Therefore starting at any place where the value of qy is known, the change in the average shear flow to some other section equals
- - (1)

It the summation 1s started where qy is zero then equation (1) will give the true average shear flow qy. Fig. A15.6l shows sheet panel (3,3 1 ,4,4! ) isolated as a tree body, Taking moments about corner 4' and equating to zero for equilibrium,
l:M
~l

=d(6P,)b_ bd=O d qx

Whence, qx

= ~P3/d

- - - - - - - - - -(2)

Thus for rectangular sheet panels between flange members the shear tlow qx or qz equals the average shear qy.
Flg.A15-58

The same rules as previously presented to determine the sense ot qx or qz atter haVing qy can be used and will not be repeated here. To show that equation (1) reduces to the shear flow equation preViously derived and used, consider a beam with constant cross-section and take a beam length d = I inch. Then,
I\I'l
ilP

= Vzd = Vz(l) = Vz
=~ZA =..:!.z...ZA (Where A = area of
Ix Ix stringer) .

Fig.A15-59

Fig.A15-60

Fig.A15-61

moment at section (A) is M + ~, hence the stringer axial loads at section (A) will equal P~ + ~p~, P a + ~Pa' P 3 + ~P3' tc. These stringer axial loads are shown on Fig. A15.58. Imagine the upper sheet panel 2, 2 1 , 3, 3' is cut along line (a-a). Furthermore consider stringer number (3) cut out and shown as a free bOdy in Fig. A15.59. Let qy be the average shear flow per inch over the distance d on the sheet edge bb. It has been assumed pOSitive relative to sense along yaxis. For eqUilibrium of this free body, ZF y = 0, hence whence qy = ~P3

From equation (1) qy

=
~P

Z6P
found above,

SUbstituting value of
qy

=_:!.z.. ZZA Ix

- - - - - (3) which Is the

shear flow equation preViously derived for beams with constant moment of inertia.
AlS.17 Example Problem to Compare Results in Using Equations (1) and (3).

qyd = 0

~P3/d
inclucir~

For a free body

two str:ngers or

Fig. Al5.52 shows a square single cell beam with six flange stringers. Between points Band C, the beam has a constant flange section which is shown in Section 8-B. ~he numerals beSide each stringer represent the area at the stringer. Between points B and D, the flange material tapers unifon4ly with the flange material at Doint A as indicated in Section A-A. It should be noticed that the increase in flange area 1s

Al5.26

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS

SHEAR

CENTER

B t!A

~A

'i

10--ro"
. k" +."~
~-l
SECTION A-A I x=SxS 2x2=250in4

~ 20"

-+B

lID
30"

with the shear flow system of F1g. A15.63, which therefore 1s the tinal shear flow system for this method of solution.

----I
Fig. AlS-83 Fig. A15-84
b'

Ole ~ bl"J"
c'
I

...1..

Solution No.2.

11-5"+5"..,1
SECTIO;," B-B TO "'-C Ix=3x5 x2=150in

Considering AP Loads in Flange Stringers. (guation 1)

in.lb.

Bending moment at section AA = 1000 x 50 ::: 50000 Bending moment at section BB = 1000 x 30 ::: 30000 in.lb. Considering Section 8-B: Bending stress intensity at midpOint at stringers by the tlexural formula:
- M.:, _ 30000 x 5 - 1000 pst. "b-I 150 x-

Fig. Al5-62

obtained by increasing only corner stringers b and c. The shear flow on section A-A will be computed using equation (3) which applies only to beams with constant section and also by equation (1) which applies to beams with varying moment ot inertia. Solution 1. USing Shear Flow Equation for Beams at Constant Cross-8ection.
(Equation 3)

Axial load in each at the stringers a, b, and

c

= 1000 x
crb

= 1000

lb.

Since q at any pOint on the cell 15 unknown, it will be assumed that the upper surtace on Section A-A 15 cut through the midpoint at r Lange stringer (a), thus making the shear flow qy equal to zero on this free surface. One-halt at stringer (a) thus acts with each side at the top surface. In this solution the webs and walls will be. assumed ineffective in resisting bending stresses, thus the shear flow is constant between adjacent stringers. Starting at midpoint of stringer (a) and going counter-clockwise around cell,
~

ConSidering Section A-A:

_ liz = 50000 x 5 -IX 250

- 1000 pai , -

Axial load in stringer (a)

= 1000xl = 1000

lb.

Axlalload in stringer (b) or (c) ::: 1000x2:::

2000 lb.

These resulting axial loads are shown act1ng on the portion between points A and 8 in Fig. A15.65, which equals the results as shown in Fig.
A15.66.

=0

(assumed cut)
I

=-~ l:ZA =- f~oO x 5 x 0.5 = -10 lb/in 'leA = -10

'lac
1000 'tcc ' -10-25QX5x2=- 50 Ib/in.

~ .... ~
/ /

,., # '"
e
b

l ~V '~~
~
/ .!'I;;~I
Fig. Al5-8!i

d'~h,4 "'~ /Iw' / J'" I";"

.. ~
;Or ":lIb'
Fig. Al5-66

Proceeding around the cell the balance ot the shear flow ccufd be ca.Lcu.Ia'ced , but due to symmetry enough values have been found for the shear flow to draw the complete shear flow piCture tor bending about the X axis when it is assume~ that one-halt ot the area of stringer (a) acts With each adjacent web. Fig. A15.63 shows the resulting shear flow diagram. The resultant at this shear tlow pattern is a 1000 lb. torce in the Z direction and its location through the midpoint at the box since the tlow 1s symmetrical. The external load at 1000 lb. also acts through the midpoint or the cell hence the external load 1s in eqUilibrium

Having tound the AP flange loads over a length (d) of 20 w , the shear flow can be computed by equation (1).

It will be assumed that one-halt at the 4P load in stringer (a) will tlow to each adjacent wet. However, there 1s no AP load in stringer (a) hence Qab Qac O. Then from equation (1),

qbb' =

O-Z~= O-~=

- 50 lb./in.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AlS.27

qb I a I

== -50 + Z -d-

t.Pb'

::

-50 -

-1000 20

== 0

Due to symmetry the lett side at cell would give the same results. The results are plotted in Fig. AlS.64. Since the increase in section moment ot inertia between beam points 8 and A is increasing at the same rate as the external bending moment, the average shear of 50 Ib./in. is constant between the ~NO beam section A and 8. Comparing Figs. 63 and 64, we tind the first method gives a shear flow ot 10 lb./in. in the top and bottom webS, whereas actually it is zero. This seems reasonable since the entire increase of flange area was placed in stringers band c. Example Problem No.2

Fig. A15-70

The same beam as in Problem 1 will be used except that the cross-sections at beam points B and A are as shown in Fig. Al5.67. The increase in flange area between beam points 8 and A has been placed entirely in stringer (a) which Changes from 1 sq. in. at 8 to 3 sq. in. at A.
3

Fig.A15-71

whence q:l, = Vb a

b:

But V = qgb a , whence

1
b

Section Ix = A-A 250in c' a' b'

a Ix'

q, =(~:)"q. - - - - - - - - - - - - Section Fig. AlS-57

(4)

150 in4
1

c' .' b'

B-B
1

From Fig. A15.71

'la=.:!...=~' b. 0.
q,

hence q. = b %

The results at using equations (3) and (1) relative to the shear flow pattern are given in Figs. A1S.68 and A15.69 respectively. (The stUdent should check these results.) It should be noticed that the true shear flow is greater in the top and bottom skin than that given by equation (1) which applies only for beams of constant cross-section.

Substituting value of q

in (4}

=0:) % - - -

-- - -- -

(5)

Thus having the shear flow on the stringer edge of the sheet panel, the shear flow on the large end of the tapered panel can be found by equation (5).
A15.19 Example Problem of Shear Flow in Tapered Multiple Flange Single Cell Beam..

"
Fig. A15-58 AlS.18 Shear Flow in Tapered Sheet Panel. Fig. A15-59

Fig. A15.72 shows a tapered single cell beam with 6 spanwise stringers or flange members. The beam 15 loaded by a 1000 lb. load located as shown. Assuming the webs ineffective in bending the internal reSisting shear flow pattern will be de'termtned ,

Major aircraft structural units such as the Wing, fuselage, etc., are tapered in both plan form. and depth and therefore the sheet panels between flange members usually are tapered in width. Fig. A15.70 shows a cantilever beam tapered in depth and carrying a load V at its end. The flange reactions at the left end have been found by statics. A free body diagram of tr.e web is shown in Fig. AlS.7l. Take moments about paint (0) and equate to zero.
ZMo

In this solution the shear flow at Station 120 will be determined by considering the ~p flange loads over a length of 30 n or between Stations 90 and 120.
Consider section at Station 120:Bending stress O'b
Me =-1== 1000 x 120 x 5 450

1333.33 psi.

=(~~) b.

- q,b,d

The horizontal component of the axial load in a stringer equals O'oA (where A = area of the stringer)

Al5.28

SHEAR

FLOW

IN

CLOSED

THIN~WALLED

SECTIONS.

SHEAR

CENTER.

Fig. A15~72

T3.:,;
8.5"

f-ll"-1 25 I
a
Il" c 1.75 Ix=271 Sec. SU. 90
b'

Since panel ab is tapered in width from 6" at station 120 to 5.5" at station so, the shear flow ~(ab) at station 120 can be found from
equation (5). qX(ab) = qY(bc) qX(bC) 24.87x5.5/6

L , a'
3.25
1000f

c
75

=- 23.2

1b./in.

2.5

= 24.87-471/30 = - 40.57 =- 40.57x5.5/6 = - 37.2

40.57-196.6/30 = - 47.12

qZ(cc')

47.12x8.5/10

4O.01b./in.

Pa

=4 =2

X 1333.33 x 1333.33

=5333.33 = 2666.7

Since the 6P loadS are the same on the lower stringers but tension the shear flow calculations if continued would give the same values as found on the top surface. Fig. A15.74 shows the shear flow pattern on station 120.
lb.

Pt
Pc

= 3 x 1333.33 = 4000 lb.

lb.

IN PLANE FORCES PRODUCED BY INCLINATION OF FLANGE MEMBERS

Consider beam section at Station 90:,,~ -

- 1000x90x4.25 - 1411.5 pSI. 271 -

Since the box tapers in depth and width, the flange stringers are not normal to section 120, thus X and Z force components are produced on section by the stringer loadS. These in plane force components are:For stringers a and a' ,
Px Pz

hence stringer loads are,

=3.25 x 1411.5 = 4587.4 Pt = 2.50x 1411,5 = 3528.7 Pc = 1. 75 x 1411.5 = 2470.1

Pa
The change in axial load IU' in the stringers between stations 90 and 120 equals the difference between the above loads, whence

=5333.33 x 2/120 = 88.9 lb. = 5333.33 x 3/120 = 133.3 lb.

,

For stringers b and b l Px Pz

<\Pa

=748,

<1Ft

=471,

<\Pc

= 196.6

FIg. Al5.73 shows these ap loads acting as wIng portion. Since shear flow Is unknown at any point, assume q equals zero in web aa' . The average shear flow in each sheet panel over a length d 30 inches.can now be calculated by USing equation (1).

= 4OGOx 0/120 = 0 = 400G x 3/12G = 100

lb.

For stringers c and c' , Px Pz

= 2666.7 x 2/120 = 44.4 = 2666.7 x 3/120 = 66.7

10",,*

lb. lb.

qy

= 'la' a - -Z<\P d-

Qar a 0 (assumed)
- 24.87 Ib ./in.

Fig. A1S.74 shows these in plane force components due to the flange axial loads.
133.3 889

.-

1,3"1 I a b 23.2

100

66.7 37.2
- -<; -44.4

f t
j

J
746

10"

40.0~

Fig. A15-74

'23.2 b' 37.2

8" 133.3

;
-

88.9

1 lOa

8"

j -

44.4

66.7

471

Fig. A15-73

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A15.29

The forces in Fig. AIS.74 will be checked for equilibrium.

ZFz = 1000 - 266.6 -200 -133.4 -10 x 40 = 0

ZFx = 0 by observation. Take moments about stringer (a),

l:Ma

= -3 x 1000 + 200 x 6 + 133.4x 12 + 10 x 40

Taking moments about point a of the forces in Fig. Al5.74 but replaCing the shear flows in the top and bottom panels by the values found above, we would obtain an unbalanced moment at 7970 in. lb. instead of 8670 preViously found. The correcting shear flow would then be g = -7970/240 -33.2 instead of -36.1 as previously rcuno , The tinal shear flow pattern would be as shown in Fig. AIS.76, which values should be compared to those in Fig. A15.75.

x 12+(+88.9 - 44.4)10 + 23.2 x 6 x 10

+37.2x6xlO = 8670 in.lb. i'.4
33.2
15.4 12.9 1.1

2.1
Fig. AI5-76

For equilibrium a moment of -8670 is needed, which is provided by a constant shear flow q M/2A

2.1

-8670/240 = -36.1 lb./in.

Adding this constant shear flow to that of

Flg. A15.74 gives the

36.1
12.9

3.9
1.1

A15. 20 Problems

1000
Fig. Al5--75

final shear tlow as

shown in FIg. A15.75.

!-,,,tp
..L

Solution No. 2 This same beam and loading will now be solved using the shear flow equation derived for beams of constant cross-section. Since the stringers are not perpendicular to the beam cross-section they have a z force component 'Nhich thus assists in carrying the external shear load in the z direction. These Pz components at station 120 have been calculated in the other. solution.
ZPz =- (2xl33.3+2xlOO+2x66.7)=-600 lb Total Vz (ext.mal) = 1000 lb.

Trr===;.0;;'3:=:'=;i .u04 cf4 10''' ~,05 _ ...072 1,2''" - .072 _ -.04 [.03 L I .2U''-----1 \03
Fig. AI5-77 Fig. A15-78

r-

r"..j 2000 lb.

12

j'-t-- 12;'-I

(1) Determine the resisting shear flow pattern for the loaded Single cell beam as shown in Fig. AIS.77. Assume load P = zero in this problem. Assume all material eftect1ve in bending. Make two solutions, one or them involving the use at the shear center.
(2) 8ame as problem (1) but add load P = 1000 lb. (3) FIg. A15.78 shows an unsymmetrical Single

Let VZ(net) be shear load to be taken by cell walls,

VZ(net)

= Vz - ZPz = 1000 - 600

a 400 lb.

Calculation ot static shear flow assuming q in sheet panel aa" is zero.

ZZA

=~x5X4=17.8
450 lb.in.

-.L

' TD =
5

cell beam loaded as shown. Assume all material effective in bending. Determine resisting shear flow diagram..
L."

1.

400 lb.

I 10"

.04

.03
Fig.AI5-79

Fig. A15-80

B (.4 sq.in.

This corresoondS to value ot 23.2 in previous solution (see Fig. A15.74).

qY(bC) =

17.8+~gX5X3

= 31.1 lb./in.

As canpared to 37.2 in Fig. AlS.74.

qy(CC') =

(4) Flgs. Al5.79 and AlS.eO show two loaded single cell - 2 flange beams. Assume the flanges develop all the bending resistance. Determine the shear flow reSisting pattern by two solutions, namely, without and With use or shear center. (5) Fig. AIB.Sl shows a Single cell - 3 flange beam SUbjected to loads as shown. Assume the 3
r

31.1+~gx5x2

= 40.0 lb.ln.

which is the same as in Fig. Al5.74.

-j

TC

A15.30

SHEAR

FLOW

IN

CLOSED

THIN-WALLED

SECTIONS.

SHEAR

CENTER.

flanges develop the entire bending resistance. Determine the internal shear flow reststing system.
300 lb. 100 lb. 1,1

l-_~ .8....., I"

r
81< "'"1
A15~81

r~.02
Fig.

45

f
10"

60""

A Q '.02 .0'

(11) Fig. A15.84 shows a ~ultiple flange-ctrcular beam section. Find the reststing shear flow pattern when carrying the external shear load of 5000 lb. as located in figure. Assume cell skin ineffective in reSisting beam bending stresses

'\_1" bl-l -F10'Y

I
B .03
Fig. Al5-82

t1000 lb.
( . ' \ 500 lb

tV= 3000 lb.

0.5
.1

(6) Find shear center location for beam in Fig. AlS.81 if the 3 flanges provide the entire bending resistance. (7) Find the internal resisting shear flow pattern tor the 3 flange-single cell beam of Fig. AlS.82. Assume webs or walls ineffective in bending. (8) Determine shear center location for beam at Fig. A15.82. Webs and walls are ineffective in bending.
.3

rU
[-'0'1'0''"1
Fig. A15-85

.04 I .03
.3
.1

.1

.1

4 panels a 5"

= 20"

Fig. A15-86

(12) Determine the shear flow resisting system for the beam section of Fig. A15.85. The 6 flanges r~ve areas at 0.2 sq. in. each. Skin is .032 thickness. Assume skin ineffective in bending. (13) Find the shear flow resisting system for the unsymmetrical beam section in Fig. A15.8B. Flange areae and skin thicknesses are given on figure. Assume skin ineffective in bending . (14) neterrune shear center location for beam section in Fig. A15.SB.

1 00*
Top Skin .03
Bottom Skin .03

2 0

.4

.04[
10" ,4

Fig. A16-a3

.035

.3

f---

T
T 8" ...L
.05

8 panels at 5"

=30"

-1

.06"
.05

.03

[:303 .04 .03

$02
I
FIg. A15-88

f 500 lb.

(9) Fig. AlS.83 shows a mUltiple flange-single cell beam section. Find resisting shear flow system ir webs and skin are ineffective in resisting bending stresses. All skin flange members have area of 0.1 sq. in. each. (10) Find shear center location for beam section in Fig. A15.83, it webs and skin are ineffective in bending.

1--12"--+-8" -I
Fig. A15-87

(15) Fig. A15.S7 shows a 2 cell beam section. Consider all material effective in reSisting bending stresses. For the given beam loading determine the internal reSisting shear flew system.

In Fig. A15.87,

(16) Find shear center location for beam section All material effective in bendIng.

R 320"

x-.---

5" = 0.1 sq. in. ;---t 5" each. Skin -- - - ;---t 5" thickness .035
:II

All flange areas

(17) For the 2 flange-2 cell beam in Fig. A15.88, deter.nine the reSisting shear flow pattern when beam section is loaded as shown. Webs are ineffective in bending. (18) Fig. A15.89 shows a 4 cell beam section with 6 flange members. Assume walls and webs ineffective in bending. (a) As a first problem assume trat the left and right curved sheet panels are removed,

~---t

7"

FIg. A15-84

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AlS.31

8"
lat!

1000

lb.
.5C1"

04 Cell Cell Cell Cell 03 12" . (I) .05 (2) .04 (3) .04 (4) .
Fig. A1S-92 .03 .03

I .L

.SCI"

;-12.5''1

1 - 1 2 " + 8"--j
FIg. A15-89

flow equaticn for beams of const~~t section. Campare the results. All six flange members have 0.2 sq. in. area each at section A-A and taken uniformly to 0.1 sq. in. at end c-c,
'S
1000f
,

leaving a 2 cell beam section. Find the resisting shear tlow system under the gl ven loading. (b) Now add the lett curved sheet to form cell (1) thus giving a 3 cell beam. Find shear flow system.
(0) Now add the right curved sheet to

5",5"1 et r r 1,
,

~iH -r-r-

-=---120'~'- - - - '

form cell (4), thus giving a 4 cell beam section. Find the shear flow system.

(22) Add two interior webs to the beam of Fig. A15.9Z, connecting flange members a-a~ and b_b n , thus making it a 3 cell beam. Find the shear flow resisting pattern at section A-A by the liP method.

(19) Fig. A15.90 shows a 10 cell multiple flange beam section. Area ot each ot the 22 flange members equal 0.3 sq. in. Assume webs and skin ineffective in resisting bending stresses. Find internal resisting shear flow system for 2000 lb. shear load acting through shear center of beam section. Find location at shear center. For solution use method ot successive approximation.

Fig. Al5-93

-x-

S;,-.I!R1S!3}0G f1 ~
\

(23) Fig. A15.93 shows a ciycular Single cell beam with 8 flange members. The area of each flange member 1s 0.1 sq. in. throughout the beam length. For the given 400 lb. externaj loading determine the resist1ng shear flow pattern at section A-A ~s1~g the ~p method over a distance of 25 inches between sections A-A and B-B. Assume cell wall ineffective in resisting bending stresses.

0 .1 1

-x-

0.1 .08 -~c:-=-::-;-;::;""---I 7 cells a 6" : 42"

Fig. A15-91

(20) Fir.a the shear center for the 7 cell beam section of Fig. AIS.91 for bending about XX axis without ~Nist. All beam material effective in ~ending. For solution use method of successive apprOximation. (21) Fig. A15.9Z shows a Single ce11-6 flar.ge tapered beam carrying a 1000 lb. load as shown. Calculate resisting shear flow pattern at section A-A by two methods. (1) By ~p method over a distance of 20 inches between sections A-A and 8-3, and (2) By USing general shear

-'

i:t:L 'I

AlS.32

SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS.

SHEAR CENTER

.:1,f!;1

~!W.~~~::'~'~ ~ ;;r:;:~s.!:.~, "':~~' -.~. ,~:.,,~;~-,;i

~

;~'i

.; . .t. " ~

:.,. :~r~;::;-l~J;i;;;.iMf:;:;S!:~

DC-a Fuselage Main Cabin Framing Structure Also Window Belt Treatment

DC-a Fuselage Lower Cargo Compt. Basic Structure DOUGLAS DC-a JET AIRLINER. BASIC FUSELAGE CONSTRUCTION

&

CHAPTER A16

MEMBRANE STRESSES IN PRESSURE VESSELS

ALFRED F. SCHMITT
A16. 1 Introduction.

The structural designer IS orten called upon to develop a vessel which Is to contain a fluid under pressure. Occasionally the design of such a vessel 1s not critical trom either a weight or shape standpoint and almost any suitably strong sealed vessel will suffice. More often, the strength, weight and torm ot such a unit are closely prescribed and rigidly controlled. Thus, the pressurized cabin or a modern aircraft Is a sealed pressure vessel containing an atnosphere at near sea level pressures and whose functional requirements include: 1 - the transmIttal ct heavy loads from
the tail surfaces and trom internal dead loads,

at adjacent meridian curves and a pair of adJacent parallels.

11 - the necessity for nonstructural cutouts for doors and windows,

111 - an efticient shape tram both the aero-

Fig. A16. 1

dynamiC and space utilization points at view, The radii R m and Rt shown on the figure tv - a minimum or weight. are found by erecting local normals to the surstructurally, the most efficient torm at tace at the element at its corners. Rm is the pressure vessel is one in which the lateral radius of curvature at the meridian curve: it pressures are supported by tensile stresses may be either positive (inward pointing), negaalone in the curved walls at the vessel. Extive (outward pointing as in Fig. AI6-l), or inamples at such Shapes are those assumed by finIte (at inflection points or straight-line pressurized rubber balloons 8-~d canvas tire meridian segments). Rt is the radius at curvahoses and by the tree surface at a drop at ture of the section normal to the meridian 'Hater (in which the surface tension forces procurve. For Simple forms of pressure vessel Rt Vide the support). The walls of these vessels is always positive; all radii Rt point inward have zero bending stlftnesses and hence have the and intersect the axis 0-0, although not genprooerties ot a membrane. The stresses developerally normal to 0-0 (see radius Rt erected trom ed, lying wholly in tangential directions at pOint C or Fig. A16-1). each point, are called membrane stresses. Fig. Al6-2 is a detail at the surface eleIn shells of technical tmportance, the ment. The forces per unit length in the meridwalls do, of course. have same bending stiffness ional and tangential directions are denoted by and hence may carry some transverse loadings by and Nt, respectively. Shear stresses are flexural stresses. Indeed, the boundary conurt rone imposed on the, shell may be SUC~l as to necessitate some localized bending near edges and seams. An efficient pressure vessel design R is one in which the configuration minimizes these nejar-tures trom a tpue membrane stress system, d i.e., nimniaas the degree of local bending stresses induced.

Nm

I"

',-

[J
I"

"

-',

=. -

"

\ ".

AlS.2 Membrane Equations of Equilibrium: Shells of Revolution Under Rotationally Symmetric Pressure Loadings.

Consider the equilibrium of a differential element cut from the srEll of revolution of Fig. A16-1. (The figure is drawn to resemble a familiar folding paper Christmas bell, since such an Object may aid in Visualization.) The element is cut out by the intersection at a pair
Al6.!

',-

fftUtl1t - - It

Fig. Al6.2

Hereafter referred to as "stresses" although their units are pounds per inch rather than psi.

i ../r-,
i -.;- '/

Al6.2

MEMBRANE STRESS:c.S IN PRESSURE VESSELS

absent due to the symmetry of the problem. The included angles between the pairs of nerldlo~~l and tangential rcr-ces are :is!ll/Hm and dSt/:Rt, respectively. Summing forces normal to the dlf:erential element, one has

its support, Is neglected.)

In the lower hem-

or,
p - - - - - - - - - - - - - - - Fig. Ale.4a

Fig. A16.4b

Here p is the Internal pressure, positive outward. Note that the shell wall thickness does not enter eq , (1). The pressure p may vary in the meridional direction but is constant in the tangential direction by hypothesis (rotational symmetry assumed). Eq. (1) is one equation containing two unknowns. Another equation may be obtained by the condition of equilibrium ot a portion of the shell above or below a parallel Circle. ThUS in Fig. A16-3, the pressures acting dmvnward on the lower portion at the shell are equilibrated by the upward vertical components at the meridional stresses, Nm.

ispherical portion meridional strssses are required as shown in Fig. A16-4b. Hence, in this class of problems it is best to derive the necessary Second equilibrium equation (corresponding to eq. 2) by considering the individual characteristics of the struct~e.
Al6.3 Applications to Simple Pressure Vessels.

Example Problem 1 Determine the membrane stresses 1~ a cylindrical pressure vessel of circular cross section (radiUS Ro), haVing hemispherical ends, i! the internal gas pressure is p. Also find the greatest combined normal strese.

Fig. Al6.5 Fig. Al6.3

SOLUTION:

Summing torces vertically p

rt

(~ sin ~)' = 2

rt

(~ sin ~)

I'm

sin ~

In the hemispherical ends Nm = Nt by symmetry and, at course J Rm = Rt = R o Hence eq. (1) is SUfficient to determine the stresses in this portion ot the structure. One
has

Solving,

Nm =

P Rt

- - - - - - - - - - -

(2)

Eqs. (1) and (2) determine completely the membrane stress state in the rotationally symI'letric shell problem: the problem is thus seen to be statically determinate. We note that eq. (2) should not be used in cases at hydrostatic pressure loadings. The basic cqncept is that of shell eqUilibrium, and consequently tor this class of ~roblema the manner of shell support must be considered. Thus, in the tank ot Fig. A16-4a, the upper cylindrical portion requires no meridional stresses sL~ce the load is reacted at the supporting ring O~. (In these analyses the structural weight, wnich always requires some stresses for

Stress = t

Nm

P R.

= 2t

Rm

In the cylindrical portion the radii are = 00 (the curve at Fig. A16-5 is the meridian curve and this is a straight line fer the cylindrical portion).

Eq. (1) becomes

I'm.
00

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A16.3

Nt

P RQ

Stress

=Nt
t

=p

R,
t

N b Therefore
Nb cos ~

=
=

2.7 Ho P
2

The meridional stress in the cylindrical portion

Is found from sq. (2);
p

2.7:, p (.92.)

R,
2
o

= 1.25 Ro P

Stress

=P R
2t

Finally, the compressive ring load, F, is (ret.

Fig. A16-6c)

Since shear stresses are absent everywhere in the meridional and tangential directions) these are the directions of principle stresses.

2F = r:1.25 R,

p sin

9 R, d 9 = 2.5

R~

Hence the greatest combined normal stress Is

identically equal to the greatest meridional or tangential stress as just computed. It Is seen to be

In computing reinforcing ring stresses

tram this result it is necessary to include

some effective skin frmn the adjacent shell walls when the ring cross sectional area is
figured.

"max = p R,;t
~ple

.
Example Problem 3. Another form of bulkhead used to close a circular pressure cylinder is elliptical in section as shown in Fig. Al6-7. Such a bulkhead shape provides tangential meridional forces at the seam (requiring no reintorcing ring as in the last example) and yet is reasonably s!flcient as regards space utilization. Problem: determine the membrane stresses in such a bulkhead.

Problem 2. .~ important problem in pressurized cabin design concerns the Shape ot the end bulkheads. ;fflhl1e hemispherical bulkheads (such as used !n Example Problem 1) are highly desirable from a stress standpoint, such fQ~ are uneconomical as regards space utl11zat~on. on the other extreme, a flat bulkhead, while ?roviding far more useful volume, cannot resist the pressure loading by membrane stresses and hence 15 structurally inefficient. A compromise configuration is that shown in Fig. Al6-6, in which the bulkhead is a spherical surface of low curvature, (~dished head~) supporting the pressure loading by membrane stresses. A reinforcing ring, placed at the seam, resists the radial component of these s-tr-eeaes . Problem: find the compressive load acting in the reinforcing ring.

Fig. A16.7

1..

Solution: In cartesian coordinates the equation of the bulkhead meridian is

From the calculus, the radius of curvature of this ~eridlan curve is

Solution; as shewn in the explOded view, Fig. A16-6b, the radial components of the Julkhead stresses are resisted by the reinforcing ring, the cylinder wall bei~g presumed to after no resistance to concentrated transverse :orces. For this case

9 ~ arc

sin

2:7 ;

21.7.

The radius Rt 1s found ~ost readily by observing that it is normal to a tangent to the meridian curve (see figure). After finding the J""'
~

,~

,1 ,

A16.4

MEMBRANE STRESSES ill PRESSURE VESSELS

slope of the tangent (angle $), one computes Rt = X/Sin~. The results are

SOLUTION:

From sq. (1) the "hoop" loadings in the upper and lower cylindrical lobes are

The meridional stress ts found from eq, (2).

Summing forces horizontally at the floor joint:

Nt

Thus,
Nm--2--~-

P (RU cos al, + RL cos a.~)

_ P Rt _ p ("a:..'..;.Y,-'-,--:,--,,--,-,--

Assume all stresses are equal and are given by

s.
Substituting the expreSSions for Rm, Rt and into eq, (1) one r inds

Nm

Su = s = P Ru/tu SL = S = P RLitL
Sf = S =

f: t

(RU cos al, + RL cos a, )

or particular interest are the stresses at the seam. Here y = a and x = a. One tinds

Letting the weight denSity of the material the cabin is (w times the developed length of walls and floor).
be wJ the weight per unit length (axially) along

Nm=~

Nt=pa(l-~)
This last result is import~~ since it indicates that compressive tangential stresses are possible it a> b VZ. As will be seen below, such a situation is undersirable because at the large resultant difference in radial expansions between the cylinder and bulkhead (the bulkhead actually contracts radially it Nt is negative) producing high seconcary bendIng stresses in the vessel walls. Example Problem 4. Determine the weight per unit length or the double-cyllnder tuselage cross section as a function ot the internal pressure J allowable stress and the geometric parameters ot Fig. AlS-8. For structural efficiency it is desired to maintain equal membrane stresses in the skin and raccr-,

w = ',V

12 tuRu
+ 2

(n - a,) +

2tLRL (n a:l. }

a.)

tt RU sin

Solving for the various thicknesses tram above and substituting, one finds (to obtain a result symmetric in appearance use was made of the tact that RU sin a:l. = RL sin aa)
W

=Z; 1RU

[n -

(1:l. +

sin 2
2

a:l.) +

R~ (n - a,

~ sin

a,) f

Since one may show that the cross sectional ar-ea 01 this fuselage Is
A
z

RU\'1t - a1 +

,r.

1 '2 sin

2 U:l. t RL

),1 \1t

1. ) aa + '2 Inn 2 a

an important consequence 01 this calculation is that the ratio of shell weight to shell volume 1s W. 2wp

A --.and Is therefore independent of the combinations of Re, RL, U:l. and a. used. The designer 1s thus free to choose these

shape parameters 80 as to satisfy other requirements.

Fig. A16.8

Example Problem 5. Determine the membrane stresses in a conical vessel of height h and halt apex angle a. The cone forms the bottom of a large vessel f1lled With a liquid of speCif1c weight w and havtng a head of l1quid H above the cone. The compt.ete..unt t is supported from above.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AlG.S

__J-r _-~-, H

Fig. Al~. 9

SOLUTION:

these secondary stress peakS, ar~~ing that local yielding of the material will level them out. However, such stress peakS may prove to be of great importance in cases of ~epeated loadings wherein fatigue failure g; considered likel? To point up the major weakness in the membrane analYSiS one need only compute the radial displacements In the two different elements that make up the pressure vessel of Fig. Al6-5, viz., the cylinder and the hemispherical bulkhead. By Hooke's law, the tangential strain in the cylinder is (IJ. Poisson's ratio, = .3 for

The meridian cur-ve or the cone has a radius of curvature Rm cc ano , at any point a distance y down from the top at the cone, the radius R t is

alum1num)
P R,
Et

.S5 =~(l-~)= Et 2

R=
t

(n - y) tan a

By integration the radial displacement of any

cos a

point on the cylinder is seen to be (ref. FIg.

A1B-10)

Then from eq. 1, at any level y

N ~ - w (H + y) (h " y) tan a t - p cos a.

To tind the meridional stresses Nm in the cone, the equilibrium of a segment of height h - Y is considered (Fig. Al6-9b). Summing torces vertically,

Fig. Al6.10

.85 - Et

P R,

Nm

2 n (h - y) tan a cos a

AdjaCent to tne seam, the tangential strain in the hemt spner tca), bulkhead is
P R, EtSLKllD = 2Et (1 - ,.)

n (h _ y)& tan a a w (H + y) +
+ w ~ (h - y)' tan' a {h - y)

= .35

Hence, by integration

Solving
w tana " =" "m G cos a.

= .35
1 (h - y) ] (h - y) [ H + y + 3

"'"'Et""

P R,

Al6.4 Displacements, Boundary Conditions and Local Bending in Thin Walled Shells.

It is appraoriate at this point to examine some of the raresaing illustrative cases to determine Whether or not the membrane stresses computed gave ss.t i srac tcr i tv accurate measures at the shell stresses. Anticipating the answer, we state trat, while the membrane analysis will give the primary streSS system in a shell-like pressure vessel, a careful (and often lengthy) analYSis of induced bendi~ caused by boundary effects will reveal localized ,secondary stress peakS. In static strength analyses of properly designed- vessels it is the practice to neglect
- various codes and standards give proportions ol common vessels which will correctly limit secondary stresses. See for example reference (1).

Thus the cylinder tends to expand more than the bulkhead - a situation prevented by the seam bevHeen these elements. It follows then trzt the seam experiences a transverse shearing action as indicated in Fig. A16-ll. These shear forces in turn produce bending moments in the sr.ell wall as shown on the figure.

6'olkhd.

~oTl

1'ill-ijjj-,=.....=..;:...::.1
Fig. Al6.l!

1_,

1("

A16.6

MEMBRANE STRESSES ill PRESSURE VESSELS

While it Is not our purpose here to take up shell bending in detail, some indication of the character and magnitude of these bending stresses should be available to place them in proper perspective. The most striking thing about t.hese wall moments Is that they are quickly damped out, becoming negligibly sma.!!Jdown to l';b of their maximum value) at a distance of about 4~ from the seam. Thus. for an instance. In a circular cylindrical shell of 40" radius and .065" will thickness, these moments are so damped at 6.5" from the seam. The next important consideration is an appreciation of the magnitude of these secondary bending stresses. For the case of the pressure vessel of Fig. A16-5, the meridional stresses are increased about 3~ at the point of maximum moment, while the tangential stresses are increased only about 3t. Fo rtunateiy, in this class of vessel, the tangential stresses are the ones designed by (they are twice as great as the meridional stresses) and hence the secondary stresses have little importance for this case (see Chap. 11, pp. 389~~22 of reference 2). In other configurations one 15 not aipys so tortunate, and detailed analysis may be required. (see references 3, 4, 5 and 6).

13% abav o

t~e

membrane stress (as against only

3Sb above for the nemtspber tcz,; bu.lkhead ) (reference 2, p , 410). For this type of bulkhead,

boiler codes sometimes pemit a ratio of alb as high as 2.6, however.

A16.4 Special Problems in Pressurized Cabin Stress

Analysis.

Because at functional req~irements over and above those of a sL~ple pressure vessel, the pressurized cabin shell of an airplane ~~s a number of stress ar~lysis problems peculiar to its confi~uration. Several of the more general of these will be considered ~ere.
DISTRIBUTION OF STRESSES BEr"EEN Sr:ELL AND STRINGERS.

The situation at the seam of the above vessel is typical at many seams or boundaries where elements are jOined which would experience ditterent expansions it loaded separately. .~ong such seams and boundaries are those: i-where the meridional cur\~ture changes abruptly. It changes from Rm = Ro to Rm = 00 at the seam in Fig. A16-5. i1 - Where a sudden change in d1rection of the meridian curve occurs. In Example Problem 5, above, considerable shell wall bending would be induced near the seam. In fact, a reinforcing ring would probably have to be added at the seam as was done in Example Problem 2, above.* l1i - at which structural members of different stitfnesses and different loadings join. In Example Problem 2, the cylinder tends to ex?and the most, the bulkhead quite a bit less and the reinforcing ring, being loaded in compression, tends to contract. Other seams and/or boundaries ot this type are those where an abrupt change in shell wall thickness occurs (addition at a doubler) or where a shell is fastened into a foundation. Good deSign tends to minimize the magnitude of the secondary bending stresses by avoiding combinations at elements which would have highly incompatible distortions. ThUS, the analysiS ot ~ple Problem 3 shews that if one closes a circular pressure cylinder w1th an elliptical bulkhead in which a = 2 b, compressive tangential stresses would develop in the buliL~ead. In such a case the bulkhead would tend to contract radially While the ~in cylinder would tend to expand as always. Thus, the shear and induced moment at the seam would be aggravated, producing (as it happens) a tangential max~um stress
.. certain details of the design of such reinforcing rings are given in the codes and standards.

To stabilize the shell wall in transmitting hea'ry tail loads tr~ough the fuselage, longitudinal stringers ars added. T~ese same stri~gers will also help to carry the meridional pressure loads. The skin and stringers must, of course, have equal st~ains in the longituditudinal. directions but, beca~se the skin 1s in a ~No-dimensional state of stress, they cannot rave equal longitudinal stresses: hence the follOWing analysis. Let the ~eridional (longitudinal) stresses in the skin and stringers be ~ and sL, respectively. St will be the tangential (hoop) stress in the skin. From eq. (1) we again have
p R, St = - t -

If N is the total number at stringers, each of cross sectional area AL, then eqUilibrium longitudinally requires

The condition of equal longitudinal strain in the skin and stringers yields

where

~ is Poisson's ratio (= .3 ~or al~1num). Solving these three equations one !inds

p R, -t-

pR. (1 + 2 ~ a) p s, ":n = 2t (I + al = 2t
sL = 2t
p

(1 + .6 a) (1 + al
.~

Ro (1-2u)_ p R, (1 + a 1 - 2t (l

a)

Where a = NAL/2n Rot is the ratio of total stringer area to skin area. A little s'tudy will show that t(l + a) is a sort or ft et f ect i v6 shell wall thickness": i t is the result of taking all the cross sectional area (skin plus st~ingers)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A16.7

and

it uniformly around the periOn this basis, the results are a little disappointing: the stringers are carrying only 40% of the stress one ~ight expect if the net longitudinal load (p n R~) were distributed evenly over the entire cross sectional area (2 n Rot (1 + al). The meridional skin stresses are reduced by the factor (1 + .6 a)/(l + a) from what they would be Without the stringers. For structures of usual proportions this decrease may amount to 20 to 30% but clearly can never exceed 40%. Inasmuch as the bending stresses due to tail loads will be superposed on these pressure membrane stresses, the reduction is certainly beneficial.
distrib~ting
~eter.

ations of these effects in a speCific case are to be found in references 2 (pp. 395-406), 6 and 8. One proposed solution to the ring-shell interaction problem is the floating skin. BaSically, the idea is to reduce the radial stiffness of the connection be~Neen the shell and the rings 50 as to allow the shell to expand freely under the pressure loading. The connection between ring and skin must still retain its shear stiffness so that ring loads may be transferred to the shell wall by tangential shear flows. Fig. A16-13 shows the basic idea of the radially flexible connection. Many

INTERACT I ON BET.'EEN RINGS AND SHELL.

Because of the neceSSity for transmitting various concentrated loads fram within the cabin and from the wings and tail to the main she.ll a.~d because it is also necessary to provide some lateral restraint which will stabilize the stringers and skin against an overall instabilit failure, the pressurized fuselage ot an airplane contains a considerable number of rings and frames distributed along the length of the shell. These rings are seldom, if ever, spaced closely enough such that they can be considered effective in carrying a part of the hoop stresses (in the way the stringers were effective in carrying part of the meridional stress). Rather, they act more like widely spaced restraining bands having the effect shown exaggerated in Fig.
A16-12.

Fig. A16.13 Ring skin cross section showing the action of a radially flexible connection.

variations of this type or "mount~ suggests themselves, some of which may nave mer:!.t for other reasons. For instance, the transmittal of wind and other Vibration noise into the cabin of a high speed transport is a problem which might be treated simultaneously by the proper choice of connection be~Neen the ring and the shell.
DOORS AND I,JINDOWS.

1
Fig. A16.12 Restraining rings along a pressurized tank. The action is representative of a fuselage with widely spaced rings inside.

The various cutouts in the shell of a pressurized cabin require special ccnsideration if an excessive weight penalty is to be avoided. Consider the panel removed from the pressurized cylinder of Fig. A16-14a. Following a

It is obvious trat the rings in this case will produce secondary bending stresses in the skin and hence may have a detrimental effect on the Simple membrane stress system. Equally ~ul are the tensile loadings developed in the rivets joining the skin and rings. Detailed analyses which will permit quantitative evalu If one looks at the problem from the point of view of a

------1d__-.,::-_-_-_-:.-_-_1r
_____
. _ ~ _ _

-~,""

pA

(a)

Ii

11 1

~
I

Fig. A16.14 stiffened shell, loaded priIDarily by bending and shear loads from the tail, an which the pressure membrane loadings are to be superposed, an interesting effect appears. Because the internal pressure tends to stabilize the curved skin panels on the compression side, the effective width of skin acting with the stringers is increased. The section properties of the cross section may then change in such a way as to produce little or no variation in the maximum tiber stresses. Indeed, the maximum tensile stresses may actually be reduced by the addition of the internal pressure loadingl (see reference 7).

(b)

common practice in dealing with c~tou~s, we determine what forces the panel-to-be-removect applies to the main structure around the border of the cutout, and then superpose a set of equal but opposite, self-equilibrating stresses to cancel these. The c~tout border is then unstressed and the panel may be removed without disturbing the new stress s}~tem in the ~in structure.
I <:"l
j .'

.I
j

Al6.8

MEMBRANE STRESSES IN PRESSURE VESSELS

In examining the figure to dete~ine wr~t sort of canceling stress system must be supplied, we see that the tangential hoop stresses bordering the cutout cannot be canceled by a selfeqUilibrating set since they have a radial component. However, the radial component of these stresses will actually be supplied by the door or window pressing oU~Nard against its fra~e. Hence, it is only the component of the hoop stresses along a chord which need to be canceled
(Fig. A16-14b).

The llmmediate problem bec~mes one of deSigning a structure to effectively support a set of uniformly distributed self-equilibrating stresses acting in the plane of the chord connecting the upper and lower edges of the opening
(Fig. A16-15a).

Fig. A15-16 shews the typical cutout structural arrangement. ~hile analytical approaches have been tried, it is probably safe to say trAt the ~rue elastic stress distribution in such a configuration cannot be computed. The necessity for avoiding high intensity stress concentrations (With their attencant fatigue likelihood) makes empirical inforTation mast useful in such cases. On the other hand, a simple rational analYSiS, based on principles outlined above, will very likely suffice tor a static strength check and tar most design purposes. (,:"150 see reference
8, pp. 16-23).

The above discussion has concentrated attention on the problems of carrying the hoop stresses around a cutout. The longitUdinal pressure stresses, while being s~Aller themselves, are intensified by bending stresses fram the tail loads. Hence, the longitUdinal stresses across the cutout ~y make this condition (or the combination) most severe.

Fig. Al6.15

All that appears necessary to support the stress system is to provide horizontal headers at the top and bottam of the cutout, Which, as beams, will carry the loads across to the Sides of the frames where the loads cancel (Fig. Al6l5b). For cutouts of usual sizes in pressurized fuselages, the stress system to be supported in this manner is quite large and it proves uneconomical to design a single horizontal frame member of sufficient bending stiffness to resist them. Instead, the shell wall itself is employed to help carry these loads across. The skin is used to form a beam of considerable depth, the skin being the web ot this beam, with the horizontal frame member and one or more longitudinals forming the beam flanges (Fig. Al615c)

Fig. Al6.16 Structural arrangement around a cutout. Most or all of the shaded skin area would probably be doubled.

LARGE DEFLECTIONS OF ?I...AJ."JE PoA...r.~LS; "Q,UILTING".

Because of the heavy shear flows and direct stresses developed, the skin is usually doubled in this region. Additional stringers may also be added to relieve the stresses. The rings bordering the cutout (and forming part of the frame) are extended same distance above and below the cutout proper (unless they coincide with a reguXar ring location, in which case they carryall the way around).
.. Clearly one of the design requirements will be to make the frame sufficiently stiff in bending against radia.l forces so that the door or window can bear up evenly against the

frame.

The use ot flat skin panels in a pressurized fuselage cannot always be avoided. Since the thin skin has little bending stiffness, it cannot support the lateral pressure as a be~~ ("plate", more correctly) end hence must deflect to develop some tensile membrane stresses which will then carry the loading. The resultant bulges of the rectangular skin panels between their bordering stiffeners give a "qUilted" appearance to the surface. Even in the case of curved skin panels quilting will occur: if the internal stiffening framework (transverse rings and frames and longitudinal stringers) is relatively rigid and is everywhere tightly fastened to the skin, then each skin panel is restrained along its

,.

'.

.:

'"0

..

...

.-:"

"-_' ...

~.F

'

.,'~.

~ ..

'

"'

-,_

~~,

"

~,

-~.,

_.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A16.9

four sides (borders) against the radial expansion normally associated with the shell membrane stresses. The result is a sort of three-dimensional-case of the behavior depicted in Fig. Al6-1Z* From a structural viewpoint, the unfortunate aspects at qUilting lie in the high concentration of stresses occurring near the panel edges and in the tensile loadings on the rivets which join shell to stiffeners. The aerodynamic characteristics of a qUilted surface are highly undesirable in a high performance airplane; "hence again quilting is to be avoided. Computations of stresses in quilted panels, inasmuch as they involve large, nonlinear deflections, are difficult. An additional (and quite necessary) complication is that of having to introduce the stiffness properties of the bordering members. The reader is referred to Chapter A.l7 for a further discussion ot the problem. A Simplified approach, indicative ot trends, is given there along with further references to the literature.
Al6.5 Shells of RevolutiOn Under Unsymmetrical Loadings.

REFERENCES (1 ) (2 )

API-ASME*

Unified Pressure Vessel Code 1951 Edition, e t , seq. - -

Timoshenko, S. "Tl1eory of Plates and Shells" McGraw-Hill, N.Y., 1940 Watts, G. and Lang, H., Stresses in a Pressure Vessel With a Conical Head,Trans. ASME, vaL 74,-1952, pp. 315-324. - - - - - - - - - - , Stresses in a Pressure Vessel With a Flat Head Closure, Trans. ASME, vol:-74,-l952, pp. 1083-1090. - - - - - - - - - - - , Stresses in a Pressure Vessel With a Hemispherical-Head, Trans. AS'ME, voL 75,-1953, pp. 83-89-.Roark, R. J. wFormulas tar Stress and Strain lt , McGraw-Hill, N. Y. &:d Edition, 1954. Howland W. and Beed, C. Tests Of Pressurized Cabin Structures, Journ. Aero. Sci. vo1:-a:-NOV. 1940.

(4)

(5 )

(6 )

(7)

Problems in which the shell af revolution experiences unsymmetrical loadings are not unccrmcn in a ircraft structural analys is The nose of a fuselage, the external fuel tank and the protruding radome are shells of revolution which may be loaded unsymmetrically by external aer-ocynentc pressures. Again, the same external fuel tank shell receives an unsymmetric internal hydrostatic pressure load tram th~ weight at tuel directed normal to the shell axis. Because of the unsymmetry at the problem, membrane shear stresses are now present and so the analystm"ust solve not two, but three equations in three unknowns (Nm, Nt and Ns). Moreover, these become differential rather than algebraic equations. Because the derivation at the differential equations of equilibrium. is rather Leng'thy , and because their general solution cannot be written (rather, only specitic solutions for certain cases may be be found), no details are reproduced here. The reader is referred to pp. 373379 of reference 2 tor the derivation of the equations ~~d for an example problem.

Designs by Krafft Ehricke of Convair

for

Space Travel

One deSign which reauces quilting in the curved skin. fastens rings and frames to the inner surface of "hat" section stringers only. Thus the ring is not directly fastened to the skin which is therefore not continuously restrained around each ring circumference. The result re a modified floating skin.

Outer Space Vehicles will Present Many New Problems to the Aeronautical Structures Engineer.

American Petroleum Institute - American Sodety ol Mechanical Engineers.

A16.10

MEMBRANE STRESSES IN PRESSURE VESSELS

Douglas

OC~8

Under Ccaatructton and Assembly.

Fuselage is Pressurized to Permit High Altitude Flight.

~:,

"

Section of

DC~8

Fuselage being Lowered into Hydrostatic Test Tank.

CHAPTER A-17

BENDING OF PLATES
ALFRED F. SCHMITT
At7.1 Introduction.

It was seen in the last. chapter that thin curved shells can resist lateral loadings by neans of tensile-compressive membrane stresses. As w1ll ~e seen later, thin flat Sheets, by deflecting enough to provide both the necessary curvature and stretch, may also develop mem-

brane stresses to support lateral loads.

ar~lysls

In the

of these situations no bending strength 1s presumed in the sheet (membrane theory). In contrast to the membrane, the plate 1s a bvo-dlmenslonal counter?art of the Jearn, in which transverse loads are resisted by flexural and shear stresses, with no direct stresses in its middle plane (neutral surface). The skin may also be classified as either a plate or a ~e~brane depending upon the magnitude of transverse deflections under loads. Transverse deflections of plates are small in comparison with the plates' thicknesses - on the order ot a tenth of the thickness. On the other hand, the transverse deflections of a membrane will be on the order of ten tL~es its thickness.
Unfortunately for the engineers' attempt at an orderly cataloging of croclems, most aircraft skins fall between the above two extremes and hence behave as plates having some membrane stresses.

with another SUCh, we write the equations down by a direct appeal to past experience and intuition. Fig. A17-l shows the differential element of a thin, initially flat plate, acted upon by bending moments (per unit length) Xx and My about axes parallel to the y and x directions respectively. Sets of tNisting couples Mxy (= - Myx) also act on the element.

yx

Fig. A17.l

Plate bending investigations ~Ave for a longttme been important in aircraft st~uctural analyses in their relation to sheet buckling problems. Recently they have assumed new importance With the introduction of thick skinned construction and still more recently With the use at very thin low aspect ratio wings and control surfaces which behave much like large plates, or even are plates in same cases. It is the purpose of this crapter to present briefly the classic plate formulas and same applicattons. Appropriate references are cited in lieu of an exhaustive treatise, which could hardly be presented in one chapter (or even one volume) as witness the voluminous literature on the SUbject.
A17 .2 Plate Bending Equations.

As in the case of a beam. the cu~rature in the x, z plane, a a.N/aX.:l, is proportional to the moment M x applied. The constant of proportionality is l/EI, the reCiprocal of the bending stiffness. For a unit Width of beam I = t 3 / 12 . In the case Of a plate, due to the Poisson erfect, the moment My also produces a (negative) curvature in the x, z plane. ThUS, altogether, With both moments acting, one has

Where
um)

~ Is poisson's ratio (about .3 for aluminLikewise, the curvature in the y, z plane

Is

These two equations are usually rearranged to give the moments in terms of the curvature. They are written
i" ay" -

Technical literature in this field abounds with many excellent and elegant derivations of the plate bending equations (references 1 and 2, tor inStance). Rather than labor the subject
As will be seen later, the presence or absence of membrane stresses is not wholly depenctent upon the magnitude of deflections, but is also determined by the form of ceflection surface assumed by the sheet (in turn dependent upon the shape of boundary and loading).

aX

w)

(1 )

My

aX w =D ( a y.:l +

" OX
e-

a x.ll

w) __
~X)

(2 )

Where

D ,t'/12 (1 -

The twist of the element, aaw/axay

(= a:aw/ayax) is the change in x-directlon-

.. the assumptions implicit in the following analysis are spelled out in detail in Art. AI7.5, below. A17.1

Slope per unit distance in the y-dlrectlon (and

A17.2

BENDING OF PLATES

(and visa versa). It ~s proportional to the couple M xv ' A careful analysis (see references 1 and 2) gives ~he relation as
~Nistlng

In a similar manner, a the x axis yields

Q

~oment

summation about

l1x y

= D (1 - ~) axdy

a'

(3 )

=al1y+~

ay

a x

(6)

/,'.. :?-"'" 1 i
~ ~/

Equations (1), (2) and (3) relate the applied bending and twisting couples to the distortion of the plate in ~uch the same way as does M = EI d~Y/dx~ :or a beam. While a few highly instructive problems may be solved with these equations (see reference 1, pp. 45-49 and reference 2, pp. 111-113), they are of little technical importance. Hence we move on to consider bending due to lateral loads. Fig. A17-2 shows the same plate element as in Fig. A17-l, but with the addition of internal shear forces Q x and Qy (corresponding to the ~V" of beam theory) and a distributed transverse pressure load q (pSi). With the presence of these shears, the bending and twisting moments now vary along the plate as indicated in Fig. A17-2a. (For clarity, the several systems of forces on the plate element were separated into the two figures of Fig. Al7-2. They do, of course, all act stmultaneously on the single element)

(Equations (4) and (5) correspond to V = dM/dx in beam theory). One t1nal equation is obtal~ed by summing forces in the z direction on the element:
q

aa ~x

aa ;y _ _ _ _ _ _ _ _ _ _ _ _ _

(6)

Equations (4), (5) and (6) prOVide tr~ee additional equations in the three additional quantities Qx, Qy and q. The plate problem is thus completely defined. To summarize, we tabulate below the quantities and equations obtained above. For comparison, the corresponding items from the engineering theory at beams are also listed.
PLATS T!:Eoo.Y BEAl1

CLABS

IT'"

THEORY
x
y

~4Z
tIC.,

:2

'~~41C~.,

' ..-&P::'I:I":;!';~'
,)
(b)

Geomstry

Coordinates De!lections Distortions axa ,

K~.4IC~

V---..#
~L

::;w
Loa<Ungo

a'"'w

o-W

dxdy

aa.,Of

~ ax
EI
M

,,(

'

.....

Structural Bending Characteristic Stiffness ccceree

E:t~

0- 12

11 - \-'0 ")

oar,

"p

(a)

t Latera
l1oment-

x' ,

Fig. An. 2. The differenti3ls are increments which should be written more precisely as, for instance,

"Hooke's La.'

4Q r (aQr/eY)4r.

Distortion Relation

!1x = D ('" ax-").l. ,W) ay:r

My

= 0 ( ii? .....

a aw

The next relations are moments in turn about the x ample, we visualize the two A17-2 acting s11nUltaneously ment, and sum moments about

obtained by summing and y axes. For exloading sets of Fig. on the single elethe y axis.

Mxy

=0

(1 - ;..) axdy

,,,

ax"

oa.,Of)

11"::.I

dX'"

day

Equilibrium

Moments

v = d i'! cx

IIx ely + (llyx + d lIyx) dx + (Q.x + d Q.x) dx dy =

Forces

(11x

+ d l1x) ely + l1yx dx

DiViding by dX dy and discarding the term at higher order gives

or,
rv.: -sx

Finally, one verJ l.m.portant equation is obtained by eliminating all internal forces (Xx, My, Mxy, Qx, Qy) between the above six equations. The result (whfcn the student shoUld obtain by himself as an exerCise) is a relation between the lateral loading q and the deflections w:
(4)

= a l'lx
ax

a l1x y
a y

(7

If

W', the deflection function, is a continuous function of x and y (as it must be, at course, I.n any technically important plate problem) then at each point aaw!axdy : a"vlaydx, as is proven in the calculus.

the corresponding equation for a simple beam is q/EI ; d"Y!dx".

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AI7.3

The plate bending problem is thus reduced to an integration of eq , (7). For a gfven lateral loading q (x, y), a deflection function w (x , y) 15 sought which satisfies both eq, (7) and the soecifled boundary conditions. Once found, w (x, y) can be entered into eqs , (1) to (5) to determine the interr~l forces and stresses.
Al7.3 An illustrative Plate Bending Analysis.

A. 'lmn ,lID = D nt.

(:n' n')' a
~

-+- ~

Hence the reqUired deflection surface (~~d the solution to the prob1em-j has the equation
w

= ~(m,l n')' sin

n n
~-+-~

~ a

sin nb ny

- (10)

Assume a lateral loading applied to a rectangular plate having all edges Simply supported (hinged). The coordinates are chosen as in Fig. A17-3. With foreknowledge of the general usefulness of the result, we assume a sinusoidal loading of the form
q _ Qmn Sin

The maximum deflection is seen to occur where the trigonometric functions have values at unity and q is also a maximum. If eq, (iO) is substituted into eqs . (1), (2) and (3) one obtains
m nx sin """ b' r.') s i n -+ " b' (a' -oa' ~)' (~' a
;lL +

m ~ x sin n ~ y
m, n

(8)

= 1, 2, 3,

Mx =My ;:

r: ;,: y

- n' (", ? +~)'

cmn

o .n")
m' 7'""

b'

Sin~ sin n a y

--,---

l1 XY

em (1 - :..L) mn =n' ("'+n')' ss ~ ~

cos

mn x

it y cos - b -

Fig. Ai7. 3 Sinusoidal loading on a rectangular plate. Sections through the loading shown for m=3, n=2.

In a Similar manner the transverse Shears

may be found from eqs , (4) and (5).

To find the resulting deflected shape of the plate we try a solution of the form
w =
-tmn
.!._____

With such results as these the plates' stresses may be determined as desired. For example, the maximum direct bending stresses are seen to occur where the shear stresses (due to Mxy) are zero. Thus
(9 )

in

In

n: x sin n n: y a b

sx (_
and hence

MC) = ::1xt
1 21

6 i'!x

t'

where ~ is the unknown deflection amplitude. This trial deflection function is known to satisfy the boundary conditions on the plate since at x = a J a and at y.:: 0, b we have w= 0 (zero deflection at the supported edges) (zero moment at the hinged edges; see eqs. land 2)

- - - - - - - - (11)

a' w ax"

It remains only to find the value of A mn which will satisfy eq, (7). Substituting (8) and (9) into (7) one obtains

The reader having a familiarity with Fourier ~ methods will recognize immediately that the above analysis provides the key to the solution of the problem of any general loading q (x, y) on the same plate. SUch an applicatiOn is made by determining the proper combination of sinUSOidal pressure terms (each of the form of eq. oj such that their sum will closely represent the desired loading. The sum of the corresponding deflection functions (each of the form of eq, 10) gives the desired solution. Details of this type of analysts are to be found in reference 1 On PI'. 113-176 and 199-256.

In common with all problems which are rorennet ec in terms of a partial differential

= 'lmn/D
or
the common factor sin m ~ x sin ~ has been dlvided out.

~-

the uniqueness of solutions to the differential equation of the form of eq, (7) is a classical proof appearing in numerous advanced texts on mathematics and matbemaucal pbystca. Since the equation is known to have a unique solution, then any solution fOWldlor it is the one ana only correct solution.

Al7.4

HENDING OF STRESSES
TABLE A17.1
Stress and Deflection Coefficients for a. UniIormly Loaded Rectangular Plate Having Various Edge Conditions
I

equation. the solution of t~e plate Je~jin~ problem depends strengly up0n the Joundary conditions (both the shape of the boundary and :~e :J?es of support provided t~ere). ~~e above exanple ~y be said to rave been deceptively easy jecause of both the simple shape of the bCw4dary and th~ type of support. Plate ,racIsms whersin the plate planfor.n is not a si~ple geometric f~gure must be solved by numerical ~9ans. 's to tte type of support, a f~ll discussion of boundary conditions fer ~lates is to be found in reference 1. ;p. 59-95.
A17.4 Compilations of Results for Plate Bending Problems.

I Long
All Sides Pinned

I
b/.

Sides Pinned, Short Sides Clampea

! Short

I
!

Sides Pinned, Long Sldes Clamped.

I ,
I
I
,

All Sides
C'lamp ed
J

a
i
I

1.0 1.0443 . 2874 ! . 0209 1.2 .0616 I .3756 I .0340 1.4 .0770 .-1518, .0502

.420 I I .0209 ! .420;.0138 522! .0243 .462 .0188 .600 i .0262 . .;86 .0226

!
I

.3078 .3834 I .4356

Fortunately for the practicing engineer, it 1s not necessary to perform analytiC computations as discussed above for the great Tzjority of practical plate problems. Problems of the type illustrated above, plus the ~iad variations pOSSible, became very fashionable exercises amongst mathematicians following the discovery by LaGrange of eq, (7) in the year 1811. The results of mBJlY researchers' labors have Jesn compiled in various torms for handy reference. A common and Impor-tant case is that of a uniformly loaded rectangular ~late (Fig. A17-4). The major engineering results are the values of the maximum deflections ar.d the max~um stresses developed. These may be put in the form (a is the length of the short side);
~ = a ~ E t ' - - - - - - - - - - - - - - - (12)

1.5172 ! .06581 .654 ! .0273 I .500 .0251 I .4680 .0906 .1017 r 5688 ! . 0799T. 690 .0279 .502 i .0267 .4872 2.0 .11061.6102 I .0987 i .714 1.0284 1.504! .0277 I .4974 3.01.13361.7134: .128!.750 I I 4.0 I .1400 I . '7-110 ! I 5.01.1416 .7476 I I <XI l . 1422 ! . 7500 i . 1422 i . 750 .0284 .498 .0284 .498
1.6
1.'

- , - -

Rectangular Plates Under Various Loadings

S. r ixcenenko , IITheory of Plates and Shells II , pp. 113-176, 199-256.

J. P. Den Hartog, IIAdvanced Strength of Materi2ls~, pp. 132-134.

R. J. Roark, ~Formulas for Stress and Strain", pp. 202-207. Circular Plates Under Various (same t.ta-ee references, in
Loadin~s

crd~r)

"!lAX = jl to - - - - - - - - - - - - - - where the coefficients a and ~ are given in Table Al7.l tor the tour most cammon edge conditions.

q a'

(13)

55-84, pp. -p. pp. 194-201.

257-287. 209-211.

128-132.

A17.5 Deflection Limitations in Plate Analyses.

Fig. A17.4

Similar presentations may be made for many dozens of other cases. With the ready availability of comprehenSive catalaglngs of these problems in references devoted to the purpose, there appears to be little virtue in duplicatlon here. Hence the follOWing list of selected references is presented. Additional references are to be r ounc in turn within these works. 'Ae note that, because of the linearity of the plate bending problem, superpositton of solutions is POSSible to extend even further the usefuL"1ess of these extensive listings.

In the intrOductory remarks of this chapter it was stated that a plate may be distinguished from a membrane by the sma.Ll, order of' its deflections (on the order of a few tenths of its thickness). We will re-examine this statement here to show that this is ~ot so much a definition as it is a~ acc~acy limitation imposed by one of the assumptions ~de in the plate analYSiS. There are several fa~iliar assuaptlons from beam theory which, of course. carryover here, inasmuch as the plate analysis resembles the beam analysis rather closely. These "beam theory assumptions ll are: - elastic stresses only are presumed, ii - small slopes (so that a~w/ox~ and a~/ay~ aregood approximations to the curvatures), iil - at least one transverSe dimension (length or width) be large compared to the thickness so trAt shear deflections may be neglected.

&

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A17.5

However, the beam theory assumptions do not place a very severe restriction on the ma-gTIitude at detleclfrons permitted. Deflections of several times the plate thickness would be permissible i: these were the only restricting assumptions. In deriving the plate bending equ~tions it was assumed t~~t GO stresses acted in the middle (neutral) plane of the plate (no membrane stresses). ThUS, in s~ing forceS to derive sq. (6), no membrane stresses were present to help support the lateral load. NoW in the solutions to the great ~jority of all plate bending problems (solved as in Art. AI7.3), the deflection surface solution found is a non-developable surface, i.e., a surface which cannot be formed fram a flat sheet without same streching of the sheets' middle surface-. But, if appreciable middle surface strains must occur, then Large middle surface stresses will reSUlt, invalidating the assumption upon which eq. (6) was derived. Thus, practically all loaded plates deform into surfaces which induce some middie surface stresses. It is the necessity for holdins down the magnitUde of these very powerful ~iddle surface stretching forces that results in the more severe rule-or-thumb restriction that elate bending formulae apply accurately only to problems in which deflections are a few tenths of the plates' thickness.
A17.6 Membrane Action in Very Thin Plates.

<-----t
(0)

Fig. Al7-5b shows the desired one-dimensional problem which now resembles a loaded cable. The differential equation of equilibrium is obtained by summing vertical forces on the element of Fig. A17-5c (draw with all quantities; loads, deflections, slopes and curvatures shown positive). One obtains

or

- - - - - - - - - - - - - - - (14)
where s is the membrane stress in psi. Eq. (14) is the differential equation of a parabola. Its solution is

There is still another source of middle surface strains in plates: this is the restraint against in-plane movements offered by the edge supports. ~bile not important in problems wherein deflections are limited in accordance with the restriction of the last article, such rest~aint does assume great linportance in the case of large deflections of very thin plates which support a major share of the load by membrane action. It is, in fact, useful to consider the lliniting case of the flat membrane which cannot support any of the lateral load by bending stresses and hence has to deflect and stretch to develop both the necessary curvatures and membrane stresses. The two-dDnensional membrane problem is a nonlinear one whose solution has proven to be very difficult. Rather than attempt to treat the complete problem, we can study a simplified version whose solution retains the desired general features. The one-dimensional analYSiS of a narrow (unit ',vidth) strip will be treated. This strip is cut tram an originally flat membrane whose extent in the y-direction 1s very
great (Fig. A17-5a).

W=2~Xt

(a-x) - - - - - - - - - - - - - (15)

The (as yet) unknown stress in eq. (15) can be found by computing the change in length or the strip as it deflects. This ~stretchn 1s given by the difference be~Neen the curled arc length and the original strai&~t length (a). Thus:

J:JdW' dx'

- a

[/1. (:/

dx - a

Since the slope ctw/dX is small compared with unity, we use the binomial theorem to 'Nrite

dw ) ' l' /' _1 + ( 1+ ( dX)

1(dW)' 2 dx

- The cone and cylinder are examples of developable surfaces, the sphere is a nondevelopable one. It is a familiar experience that the skin of an orange cannot be developed into a flat sheet without tearing.

.. here "ds" is the differentia! arc length of the calculus and has no kinship with the s which denotes the membrane stress throughout the remainder of the analysis.

A17 6

BENDING OF PLATES

Hence
6 =

J} ~ (:YJ
+

dx - a

n:r

dx

Substituting through the use of eq. (15) and integrating we find

gJi a 3

at the middle of the long side of ~he panel. We note that the limiting case, alb = 0, corresponds to the one-dimensional case analyzed earlier .. Unfor'tunaue Ly, an extraporat i cn of these ~No-dimensional results to that limit does not show agreement with the one-1imensianal result. Presumably the jiScrepancy way be traced to the excessive influence of inaccuracies in the assumed deflection Shape Jf the ~embrane as used in the approxDnate two-dL~en sional solutions. Experimental results reported in reference 4 show good agreement with the theory ~or square panels in the elastic range.
TABLE A17.2

O=Z4s't'
Now by elementary considerations
Membrane Stress and Deflection Coefficients

I
i
i

o ; s a

alb
n, n2

1.0 .318 .356

1.5 .228 .37

2.0 .1. .336

2.5 .125 .304

3.0 .10

Equating these last and solVing we find

.272

I I I

4.0

5.0

.068! .23

.052[

! .205!

s = .347 [ E

T (q

a) 'J'/' - - - -

- - - - - - - - (16)
Al7.7 Large Deflections in Plates-.

It eq. (16) is sucet trut ec into eq,

one gets for the maximum detlection

(x :

(15) ~ )

"'MAx

q a)'/' .360 a ( Et

- - - - - - - - - - (17)

Equations (16) and (17) display the essential nonlinearity of the problem, the stress and the detlection both varying as fractional exponents ot the lateral pressure q, Solutions of the complete two-dimensional nonlinear membrane problem have been carried out, the results being expressed in forms identical with those obtained above for the onedimensional problem, ViZ.,

- - - - - - - - - (18)

'I' - - - - - - - - - - (19)
Here ftaft is the length at the long side of the rectangular membrane and n~ and n e are given in Table A17.2 as functions of the panel aspect ratio alb. The maximum ~embrane stress (sMAX) occurs
The work of Henky and F01'pl is summarized 10 reference 3, PI'. 258~290 and in reference 4. The partial differential equation solved is given in reference 1 on p. 344 (eq. 202) and. the approximate method 01 solution usually employed is sketched out on pp. 345, 346 of this same reference. The reader who would compare presentations amongst these references should note the differences in the definitions 01 the plate dimensioning symbols "a" and ''b''.

In the preVious articles of this chapter the results of analyses were outlined for the ~NO extreme cases of sheet panels under lateral loads. At one extreme, sheets Whose bending stiffness is great relative to the loads applied (and which therefore deflect only slightly) may be analyzed satisfactorily by the plate bending solutions. At the other extreme, very thin Sheets, under lateral loads great enough to cause large deflections, ~y be treated as ~em branes whose bending stiffness is ignored. As it happens, the most efficient plating designs generally fall between these two extremes. on the one hand, if the designer is to take advantage of the presence of the interior stiffening structure (rings, bulkheads, stringers, etc .}, which is' usually present tor other reasons anyway, then it is not necessary to make the skin so heavy as to behave like a ~pure~ plate. On the other ~~d, if the skin is made so thin as to necessitate supporting all pressure loads by stretching and developing membrane stresses, then permanent deformation reSUlts, producing ~quilting~ or ~'Nashboardingn. The exact analysis of the two-dimensional plate which undergoes large detlections and thereby supports the lateral loading partly by its bending resistance and partly by membrane action is very involved. A one-dimensional
.. The discussion to follow will be concerned primarily with problems dealing with the support 01 a uniform pressure load on a fiat skin panel. It may, therefore, help the reader to fix his ideas if he visualizes the discussion as applied to the problems of analysis of a single rectangular sKin panel taken between the stringers and bulkheads 01 a seaplane hull bottom. EqUally useful is the picture of the very nearly fiat panel between rings and stringers in the slightly curved side of a large pressurized fuselage.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A17.7

analysis, parallel to that at Art. Al7.6, is to be found in reference 1, pp. 4-10. A more elaborate two-dimenSional analysis is shown on pp. 347-350 of this same reference. An approximate solution of the large deflection plate problem can be obtained by adding together the flat plate and membrane solutions in the following way: Solve sq. (12), the plate bending relation, for q; call it q' ,
_Et' q' _w. niv.

The approximate large-detlection method outlined above ~~s serious Shortcomings insofar as the prediction of stresses is concerned. For simply supported edges the maximum combined stresses are known to occur at the panel midpoint. Fig. A17-7 shows plots ot these stresses for a square panel as predicted by the approximate method (substituting q' and q" into eqs. (13) and (14) respectively and cross plotting With the aid at Fig. A17-6-). Also shown are the maxunum stresses computed by the exact large-detlection theory (reterence 5).
l<>
.X&o~

a a6

Now solve eq. (18), the membrane relation, tor q; call it q",

approx.

--:.-e--,,.-'

..
o

'0

'

- '00

10

./'

The sum of these two pressures gives the total lateral pressure, called Simply, q.
q = q' + q"
q

o o - - - (20)

/'

100

=a

Et 3 1 Et - - b - '''l1AX +,--,-..".. a40 (_)6 n, a

Ww.x

:5

Fig. A17. 7 Largedeflection theories midpanel stresses; simply supported square panel.

-:.4

1'0

a Eq. (20), we see,. is based upon summing the individual stlttnesses at t~e two extreme behavior mechanisms by which a flat sheet can support a lateral load. No interaction between stress systems is assumed and, since the system 1s nonlinear, the result can be an approXimation only. Eq. (20) is best rewritten as
q "

Because ot the obVious desirability at using the results of the more exact theory, s~e or these are presented in Table A17.3. The treatment at additional cases (other types ot edge support) may be found in reference 6, pp.
221, 222.
TAaLI Al1'.'
Lup (lIILIIGnD

E t 40

a \~

U"'MAX)(~)'
b

1 ~

(WMAX\' _- - - - ( 2 1 ) t 7
i I
I

...
I

I or Slr_
_It ,,~/1\~

Fig. A17-6 shows eq. (21) plotted for a square plata USing values at a. and TIl as taken from Tables A17.1 and Al7.2. Also plotted are the results of an exact analysis (reterence 5). As may be seen, sq. (21) is somewhat conservative inasmuch as it gives a deflection which is too large for a given pressure.
"0
'00

I
I

'L'
,430

1'810
$.10
1.10

~/nal 0.10
'ab / gJ i 4." y 11E11 1 1. 2I ' .. 1L
_It 1

3.110

11 ta.-

.IM

,. " ... .. L. .. .. I ,.. _11.24

('ll

D<Ia.-. P....... l..oui

~, """'..... ~I

p_

C-mcl-.

P-..C.,.rtb:1

I.U

,. , ,.n
14.:10

.. t/D4

no
,.~

,n

1.1'0

lD.ao

,u

...

1$.10

11.011

1.10

, .ee

'-"

1.44

II

_.'"

.,

'M' ..""'..II

.xac~

&pprox._

---

"0
100

o o

~0.' vlt

VI
-

II

J.

lOt&l . n l l .. .

--

1.11 1.40

jlO.)O
4.1'

,U

14 1.'1

'.11

l'.~
..

1.1'11 I.M

'.M
1'.4 10.10

11 0 1.10 II.U

- .... ... .... . .... ..

I.U

'.n

II. III

10. :10

"

10.110 L1.

.n.... 0lIl_01. ..........

_ _ 01.

'v.

A17.8 Considerations in the Applicationa of Large-Deneeuce Plate and Membrane Analyses.

-'

.~~i.~:!..~
l.~

Before concluding this chapter it 1s pertinent to note several serious omissions in the developments outlined above with regard to their application to flat pressure-panel analyses within a ship hull or fuselage. The

1.0

2.0

Fig. A17.6 Deflections at the midpotnt 01 a simply supported square panel by two large.deflectian theories.

but 1181ng n2 . 260 in eq, (19), This value gives the stresaes at the center of a square panel whereas n2

.356 in Table A17. 2 is for stresses at the panel edge.

,O~

,c<

A17.8

BENDwa OF PLATES

large-deflecti~n plate and the ~embr~ne analysGs were dev al cped for applications where the plate bending analysis appeared inadequate. However, these analyses the~elves presumed condit10ns seldom encountered in practice. FIRST, the analyses assume unyielding supports on the boundaries of the sheet panel. In practice, the skin is stretched acrOBS an elastic framework ot stringers and bulkheads. It follows, therefore, that the heavy membrane tensile forces developed during large deflections will cause the supports to deflect towards each other thereby increasing the plate deflection and relieving some of the stresses. A simple one-dimensional analysis for a membrane strip having elastic edge supports (parallel to the analySiS of Art. A17.6), shows errors an the order of 25 per cent are likely if the framework elasticity is neglected (referenc~ 7). At this writing no two-dimensional treatment of this problem is known to the writer. SECOND, it is seldam that the analyst has to cheek a panel for lateral pressure loads alone. Most Often, the entire ~fleld~ of panels on the framework of stringers and bulkheads must Simultaneously transmit in-plane loadings from the tail load bending stresses and the cabin pressurization stresses. Inasmuch as the large-deflection plate and membrane analyses are nonlinear, it tollows that correct stresses cannot be found by a straight superposition. The magnitude of the error introduced by such a procedure is difficult to estimate in the absence of an exact analysiS. A one-dlmensional analysiS, parallel to that of Art. A17.6, but With elastic supports and axial load, is given in reference 7. These results, which indicate the effect of the axial load to be quite important, may be used as a guide in lieu more complete two-dimensional studies. The interested reader is referred to the origInal work for details.

RE...J;lERENCES
1.

2.

Timoshenko, S. "Theory of Plates and Shells", McGraw-Hill, N. Y., 1940. Den Hartog, J. P. ~Adv2.nCed Strength of Materials~, McGraw-Hill, N. Y., 1952. Sechler, E. and Dunn. L. "Airplane Str~c tural Analysis and Design", John Wiley.
N. Y., 1942.

3.

4.

5.

Heubert, M. and Sommer, A., Rectangular Shell Plating Under Uniformly Distributed r.yd~Q static Pressure. NACA TM 965. (selected large-deflection plate references). a) Moness, E. Flat Plates Under Pressure, Journ. Aero ..Sc1., ., Sept. 1936.
b) Ramberg, W., McPherson, A. and Levy, S.,

Nonnal Pressure Tests of Rectangular Plates, NACA TR 748, 1942. c) Levy, S. Square Plate With Clamped Zd~es Under Normal Pressure ?rOducin~ Lar~ 5'e"f'Iec~ NACA TR 740, 1942. --------

d) Levy, S. Bending of Rectangular Plates With Large Deflections, NACA TR 737, 1942.
e) Chl-Teh Wang, Nonlinear Large Deflectlor. Boundary-value Problems or Rectangular Plates, --- NACA TN 1425. . . 1948. f) Chi-Teh ~ang. Bending at Rectangul~r Plates With Large ner'iecurons , NACA "i'N
1462, 1948.

6.

Roark, R. J. ~Formulas for Stress and McGraw-Hill, N. Y. 3~ Edition, 1954.

Cab Ina , NACA TN 2612, 1952.-

Strain~

7.

Flugge, W. Stress Problems in Pressurized

of

CHAPTER AlB

THEORY OF THE INSTABIUTY OF COLUMNS AND THIN SHEETS

(BY DR. GEORGE UANIS)

PART 1

ELASTIC AND INELASTIC INSTABIUTY OF COLUMNS

Ala. 1 Introduction,

Thus the total bending moment at section z

will be:-

Part 1 of this chapter will be confined to the theoretical treatment of the instability of a perfect elastic column and an imperfect elastic column. The column Is the Simplest of the various types ot structural elements that are. subject to the phenomenon of lr~tabl11ty. The theory as developed tor 001umns forms the basts tor the study of the instability of thin plates, which subject Is treated in Part 2.
AlS.2 Combined Bending and compreeeton of Columns.

M=

Pu +

Q~Z on upper portion - - - - (d)

Q(l-a) (l-z) on lower portion (e)
1

11. Pu

From mechanfc s of simple bending, we have

the deflection equation,

d"u
--a

dz

= -

11
Ely

- - - - - (1)

Consider a column with one end simply supported and the other end hinged (Fig. AlS.l) under the simultaneous action of a compreSSive load p and a transverse load Q. Without the load P the bending moment due to Q would be:p

Thus the deflection u(Z) ot the column

IS,

Qaz
1

, (0

1 - a)

d"u

(1-a) (l-z)
1

El dZ" = - Pu - Q
I
a
I

(1 -

1)

- - -

(2)

.ll:....-Jf+-Q I
I
I \ \

If we introduce the notation,

(3)

The general solution of eq, (2) is:

U ...

--\u
A p

C:L cos Kz +. C. sin Kz -

Qaz
PI

'
(4a)

(0

z 1 - a)
+

- - - - - - - - c, sin Kz 't

~!-

u = C;5 cos Kz

o (I-a) (l-z)

'

(l-a::z::l)

- - - -

(4b)

Fig. AlB.l

On the lower portion or the column _ caz 11, - -1-- ----

Where C~J Ca, C3 J and C40 ~re constants of integration to be deternined from bo~~dary conci tions.

----

-- -

(a)

On the upper portion

M~ :: ~ (l-z )
(b)

= 1,

For eqe . (4), s tnce u = 0 tor z = a and it follows that:

C~

= 0 and C3 = - c. tan K1 - - - - - (!)

Due to the deflection u(z), the axial load P contributes to the bending moment by the amount:M~

At z = (1 - a) the two portions of the deflection curve given by (4a) and (4b) respectively must have the s~e deflection and slope. From these two conditions we determine Cll and C.
Al .1

= Pu

- - - - - - - - - - - - - -

(c)

AlS.2

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

c,=

~ sin Ka C =_q sin K(I-a) PK sin Kl' to PK tan K1

(g)

Substituting (r l and (g) into eq, (2),

we obtain:
u

Q. sin = PK sin

Ka ca K.l sin Kz - PI

(05 Z S 1 - a) - - - - u

- - - - - -rsa )
q(l-a) (l-z)

=~

sin K(I-a) s1 K(I-z) PK sin Kl n

PI

the bOdy would cause only infinitesimal Changes in the displacements and the body recovers 1f the added loads are removed. When the displacements are continuously increased With little or no further increment at loads, the system is unstable. If the body will remain in the displaced position after the removal ot the disturbance, the body is said to be in neutral eqUilibrium. Havir~ these definitions, we will not investigate the behavior of the column betore and atter the critical load !s reached.
p

(l - a S z s 1) - - - - -

- -(5b)

It load Q is applied at the middle of the column the maximum deflection is:q tan 19,. 2
2PK

GA(~ s "-...:;;J __ ~

_/

Fig. AlS.2
Ass~e, as shown in Fig. A18.2, that a Simply sunported colwmn loaded by an axial load P is bent by a small disturbance. If the deflection IS u, the bending noment due to P is Pli. From basic mechanics, we know that,

-----(6)

It is obvious that for ~ = ~ , tan ~I _ _ . Thus the maximum

~

ceriecta on of
n and from

column becomes infinite for K1 eq, (3).

1 R =
EI

-M , whence

Per = - I'

n'E1

- - - - - - - - - - - - (7)

If = -Pu

- - - - - - - - - - - - -(sa)

Equation (7) is an Important equation derived first by Euler. It gives the critical compressive load which causes infinite detlection in a column and it specifies the ultimate strength or a column in compresSion. It is obvious from eq, (7) that Euler's critical load is independent of the magnitude ot the transverse load Q. It seems, therefore, that even in the absence of the transverse load Q, the maximum deflection becomes infinite under the action ot only a compress iva load as
.g1ven by eq , (7).
AlS. 3 Elastic Stability of a Column.

The exact expreSSion for the curvatuve of the neutral axis i5:1 d9 R = dS ' where s is the arc length at the

deformed axiS, and Q the angle between the tangent to the curve and the z axis. ThUS, EI.@+Pu=O
ds

- -(8b)

Differentiating (8b) with respect to s and since ~~ = sin Q, we obtain:

E1

~~9

+ P

stn 9

= a - - - - - - - - (Bc )

The above conclusion as to the critical load was based on purely mathematical reasoning. We have found a critical value of a compressive load which causes infinite deflection.

Multiplying (ec) by dQ and noting that:d'9 d9 d5T -_ lis d9 d (d9) t1 ng dB' an d 1rrcegr-a
QdQ

Far more important, however, is an investigation at the stability of a column which should ~e based on phySical arguments. The question arises as to what hanpens before the load P reaches its critical value as given by eq. (7) and also how the column behaves if this critical value is exceeded. An elastic system is called stable under given loads When infinitesimal loads added to

EIj ~ d (dQ) ... P sin

ds d5

= C,

or

2" cs

EI (d9)' _ P cos 9

=c

-(8d)

Since at end A, g =- a and M = E1 d9 ds we find that C = - Peas Q.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A1B.3

Now let k

~~

- - - - - - - - -

(8e)

- - - - - - - - - - - - (8n)
2P

Then sq. (ad) becomes, 1 d9

Q-

n~
cos a
1

------------ .

(80)

V2 ds = -k V cos

or

J~

Let us now write the bending moment M ~ PO at the middle paint in non-dlmentional form:
m=..L.~=~ Per 1 1t
(8p)

The total length ot the column in the deflected shape 1s given by:= _Ll
d9

Since p = sin
(8h)

-a or 1 = fa -a

12 k

..;

cos

- - Q _ COS <l

elliptic integral K and the ratios P~r and calculated tram equations (sa) and (80). Thus

% is a

function at a so is the

d9

2k/ stn"

~ -sln"9/2

----(81)

P~r

is a function or

calculated by

Denoting sin ~ by P and introducing a new variable 0 :sin

means at tables giVing elliptic integrals. Thus m can be plotted against 0/1 as shown in
FIg. A18.3.
011

~=

p sin
O~/l

Equation (81) then becomes,

l=:f:-iz
1<:

a / I-p"' sina 0

d9

=2K

- - - - (8J)

where

Is called the

Fig. A1S.3

complete elliptic integral of the first kind, and i t can be found in tables. It a and therefore p Is very small, 'then p3 sin It III can be neglected in equation (8j) and then
1

" ' - - - - - - - - - - - - - - - - - - . ; ,.. m

1a 2

=~ =n
l'

Ei
(8k)

Let us now examine the stability at various eqUilibrium configurations. Assume that a load P~ is acting on the column and the column has a certain maximum deflection where P~ does not correspond to O. The non-dimensional maximum bending moment is:p' 6 :n' _ _ (_I Pcr i- - - - - - - - - - -

whence P ;: n a:Er

_ _ ____ _ ___ __

The deflection at the midpoint of column 15:- 9 = 0, du = ds sin El, and trom Eq . (Sf)

(Bg)

u(z= .1)=0'.1... La
2 2k
0

sIn Qd9

(81)
2

lalna.!. _ atn" 9
2

The m~ versus Oil curves are straight lines. The column is in equilibrium if m = m~, or in other wordS, it the m~(O/l) curve intersects

the m(O/l) curve.

We see from fig. Al8.3 that these curves intersect at the origin only i f p lo <. Per. The column, therefore, has only one pOSSible equilibrium form, tor example, that for which all - 0, which is the straight form. When p~ ~ Per, there are two points of intersection,

or in terms of .0:- (8m)

From equation (8j) and equatton (8k), we obtain

AIS.4

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

one when 0/1 = 0 and the other (point A) tor which 6/1 ~ O. The column thus bAS two possible equilibrium forms, one straight and

(b)

one bent.
Let us now assume that at 6 = 0, the column Is displaced by a small disturbance and acquires a deflection O~. For pl C Per' we see trom Fig. 3 that m s- m1.. Thus p1. Is not sutficient to maintain the column in eqUilibrium in the bent form and i t will spring back to its straight rorm . Thus for p.l. ... Per' the straight form Is stable. (c)

At PIPer = 1, or at point (A), a bifurcation at equilibr!um occurs and the column starts to acquire two POSSible neighbor pOSitions of equilibrium, the straight and the bent. Above PIPcr = 1, the column has two pOSSible equilibrium positions 0/1 0 and ~ /1 .,. o.

It p1.

>

Per' then m1.

>

m. Thus m1. will

bend the column still further. This means that if p1. - PeT' the straight form of equilibrium Is unstable. The column will continue to bend until m1. becomes equal to m (point A in Fig. 3). It the column Is displaced further tram A, the deflection becomes larger than O. and m _ mJ. at the new position. The column will spring back to pOint A. Point A is therefore stable. At P .: Per, the mJ. versus 0/1 Une is tangent to the m curve at the origin. Therefore, ror an infinitesimal disturbance, the column will remain in equllibrium at the dIsplaced POSition since for such small disturbances mJ. remains equal to m. The column is therefore in neutral equilibrium.
AlB. 4 The Failure of Columns by CompreSSion.

Thus as far as initiation of instability is concerned, the Euler load as given by Eq. 7 can be considered as the critical load. The question arises Whether this load has a practical use for design purposes. A logical design criterion is obviously the maximum load which a column can sustain. We observe from FIg. A18.4 that the load P increases tor increaSing displacement 0.' This behavior is due to the development of large deflections due to bending. However J over a considerable range of deflections oJ the P -+ 6 curve is practically horizontal (for instance, between points A and B the ratio 0/1 varies trom zero to - 0.4). For such large deflections for which the column load does not change practically, i t is obvious that the column ceases to fUnction properly. Therefore, trom this point of view J the Euler load can be considered that which characterizes the maximum strength ot the column.

In diSCUSSing the ~tability ot a column in the previous section, it was shown that below the critical Euler load (Eq. 7), the straight form is stable, above the Per the bent torm is stable and at Pcr the eqUilibrium is neutral. By plotting the curve PIPer versus 0/1 as shown in Fig. AIS.4, we observe the following behaVior.
(a)

The riSing part ot the curve BD holds as long as the material behaves elastically. At some paint OJ however, inside the almost flat portion ot the c~rvr. the inner fibers of the column acquire maximum stress equal to the yield stress. If we carry out an elastlcplastiC analYSiS of the subsequent behavior J we observe that the curve drops almost 1mm.ediately. Again this maximum load PB is very near the Euler load. For deSign purposes, therefore, the Euler load J which is a buckling load, is a very good approximation to the Ultimate load which the column can sustain. Another argument will contirm the above conclusion. In discussing the DuCkling of columns in the pr-evf ous paragraphs , we have assumed that the column 15 initIally straight, centrally loaded and made of homo~eneous material. Actual columns, however J are imperfect due to initial crookedness (for instance, due to unaVOidable tolerances in their manutacture), due to slight load eccentricities and due to lack of complete hamogenity. Therefore, a certain amount of bending Is always present even for small loads.

Below PIPer = 1, there Is only one eqUilibrium position, 0/1 O. PIPer

P;a./Per
1.0

.~

o. I

0.2

0.3

Fig. Fig. AlS.4

AlS.5

Let us now examine the behavior of such Initially imperfect columns by assuming a certain initial deflection uo ot the column axis (see Fig. Ala.S). For small deflections,

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AlB.S

the change of curvature due to subsequent bending (after loading) is:-

approaches the horizontal line PIPer = 1 asymptotically. ThiS curve, however, is valid for small deflections for which the approxlmation:is valld. By a treatment similar to that in the previous paragraph, we will find that for large deflections the load deflection curve raises after the paint I (curve FIH). Due to the onset of plasticity, the actual curve drops at the point I~ (curve FII~Hl). The failing load at I~ can be either greater or smaller than Per' but it is usually very near to it. In the above discussions we have shown that for all practical purposes the Euler buckling load can be considered as the ultimata load which a real or practical column can sustain. Besides its closeness to the actual ult~te load, the critical load can be eaSily calculated from equation (7) without the necessity of carrying out a lengthy calculation which will include the initial imperfections and plastiC effects. It should be noted, however, that the buckling load ~iven by equation (7) is valid when the uniform stress due to a compreSSive load (0 = PIA, where A is cross-sectional area) is below y1eld stress. It a is above the yield stress, the theory of plasticity predicts another value for the buckling load. Referring now to equations (11) we tind:Pn =
n~Pcr

In the differential equation of deflection one can prove that llR is th8 change of curvature which for an initial straight column coincides with the curvature itself. Thus in the present case, where the bending moment is Pu, the equation of deflection becomes:-

.-

k'u

- - - - - - - -

(9)

Let us express uo in Fournier series:~ nuz uo .. , Z 3 sin-n 1

- - - -(10)

Substituting (10) in (9), we find the solution which satisfies the boundary conditions (u = 0 for z .. 0, Z = 1) is:u ::.1:
~

on sin T

nz

- - - - - - - (lla)

n=l
where

on ,,-_:;-,,:-

- - - - - - - - - (llb)

The deflection of the column at the center

15:-

(Per tram equation 7)

- -(12)
If we plot the deflection versus the load we obtain the curve (Fig. AlB.B), which
piPer

~n

Thus as P approaches Per' we see that

c
H

etc.

1.0 A

I' -

~-

G
<,

, 'I _..,;6'i';-,,;_ ' x-v ":Il8. - PIPer

'H~

-;

In a buckling test we measure middle point. Thus:-

a:

Omax

- 6, where 6 is the initial deflection at the

o ;;: 6:nax -

6. ""O'L-_
Per
P
1

and,

h-::!-----------"1:"/I
Fig. AIS.6

Per p - 6 = 6:l,

- - - - - - - - - - (13)

113

AlB.8

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

If in a buckling test we plot 6/P versus

0, we can obtain the critical load experi~entallY

1 1

0 0

1
0

without kn~Nlng the initial deflection 6:1,' It is simply the inverse of the slope of this curve.
COnditi0118.

a
sin kl - sin kl or sin kl whence

cos kl - cos kl

1
0

AlB. 5 Buckling Loads of Columns with vartcas End

=0

or kl = nn (n

= 1,

2, 3

- -I
(lea)

From the conclusions reached in the previous discussion, we can consider the buckling problem as an instability prOblem of an initially straight column. Thus we assume a certain deflected position near the straight configuration as another possible equilibrium form and seek the loads under which the nonstraight form is possible. Furthermore, only a small deflection analysts is necessa~J. The general differential equation of bending-buckling 18:-

Thus for Pn equal to the right hand side of equation (16a), we have a possible rorra at equilibrium of the bent form. The smallest value of Pn occurs at n = 1, and this 1s the buckling load:
Per- =

nllEI -r -

- -

(16b)

The buckling load for other end conditions can be derived in Similar manner.

- - - - - - (141
and the general solution is:-

INELASTIC COLUMN
+

STR~IGTH

U3

c~

sin kz + C. cos kz

C3 z

c. - - (15)

AlB.6 Inelastic Buckling.

Introduction.

-The coefficients C~, C., Ca and C. depend on the conditions at the end supports. The various end conditions are:Free end: Pin end: -

d'u dz lll =
u =0
~

dZF =
d'u dz'" = du
dz 0

d'u

Euler's theory of buckling 1s valid as long as the stress in the column nowhere exceeds the elastic limit of the column material. We have seen that the analysis tor perfect and imperfect elastic columns leads to the same result, namely, equation (7). The case of the inelastic buckling, that Is, instability under axial load exceeding the elastic limit stress, presents some difficulties. As we will see, the perfec~ column analysis leads to a different expression !ar the critical load than for the perfect column. This is due to the tact that 1n the plastic stress range, the material behaves differently under loading and unloading, as illustrated in Fig. AIS.? Let us now examine the two cases:
--:::::: Loading

Fixed end:- u

Thus we have 4 end conditions. These give systems at four linear homogenous equations. A trivial solution of these is the zero solution. For the buckling state, however, C1 , Ca , C3 , C. are not all zero. The condition at non-zero solution of the above system is that the determinant at the coefficients of C1 , C., C3 and C. is equal to zero. From this equation, we calcUlate the buckling load. For example, in a simply suoported beam, d'u (u = 0, dZllI = 0, at both ends), the end conditions furnish give

;;

o
/

/
/

/~nJoading

Deformation E:
Fig. AlB.7

c,

+ C.

= 0

C.

=0
AIB.7 Perfect Column.

Reduced Modulus Theory.

For buckling we

~ust r~ve:-

Let us assume tr~t the perfect column 1s first compressed uniformly up to the stress 0. To study the critical value, ocr, ot 0 for wh1ch

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AlS.7
-

the column becomes unstable we assume:(1 )

15 8 11 = E 15

Ell

(19)

That the displacements are small so that the relation between the radius of curvature R and the deflection u of the elastic exrs is, - - - - - - - - - - - (17a)

The distribution of the compressive (-) stress and the tenSile (+) stress due to bending is shown in Fig. Ala.sa. The stress becomes zero on line (a a1.), which is at a distance e from the centroid c. For eqUilibrium of stresses on the cross-section we have,
_

(2)

Plane sections remain plane, therefore the change of straIn due to bending at a distance h on the plane of bending IS,
0<

~~

OS1d.A

+ ~all OS.dA

=0

- - - - (20)

and for eqUilibrium tor moments,

= if =

h dz'

dllu

- - - - - - - - - - (17b)

(3)

The stress-strain relation tallows the Simple tension curve for the material. The plane of bending is a plane ot symmetry of the cross-section.

Due to the linear distribution at stress, we have:-

(4)

os,

00,
=-h

a,

>

Assume now a column with the cross-section as shown in Fig. Ala.sa be compressed in the

os.

Ocr.
=-h

- - - - - - - - - - -

(22)

a.

IntrOducIng now (17b), (18), (19) and

(22) in equation (20), we obtain,

-Et
where,
Q~

~, +

EQ. = 0

- - - - - - - - -

(23)

= ~a1 h1. dA

, Q,1I

= ~a. h.dA

(24)

are the moments of the cross-sectional areas to the right and left at lIne a a.

'., ,
!I
Flg. AlS. Sa

Fig. AlS. Sb

From eq, (21) we Obtain,

- - - - - (25)

where, plastic stress range and that the compressive stress prior to instabilitybe-a. To consider the condition of buckling, let the column be slightly deflected transversely. The stress on one side of the column will then increase due to the bending following the stress-strain curve, while on the other side the stress will decrease and will therefore follow the unloading elastic line shown in Fig. Ala.ab. For small changes of the stress, on the first side, the variation of stress Is related to the variation of strain by:6 81

E=

- - - - - - - - - (26)

E is the so-called reduced modulus, and I1. and III being the moment of inertia of the two sides.
We observe that the position of the neutral axIs in terms of the axial stress is gtven by eq, (24), While the buckj.tng eq. (25) is Similar to the e Laat tc buckling eq. (14). However, the value of K here is not given by eq , (3), but by,
K =

= Et

(a)

Oe1.

- - - -

- - - (18)

where Et (aJ Is the slope of the stress-straIn curve at stress cr. On the second side the Changes will follow the elastic relation, that 1S,

- - - - - - - - - - - - - (27)

Therefore all the results of the preVious analysis will be valid for the case of Inelastic buckling. For instance, ror a SImply

AlS.8

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

supported column according to eq, (7), we will have,

(28)

1/2

Since E is a function of Ocr given by eq. (28) and the value of Et at the unknown Ocr, the calculation of the critical stress requires a trial and error sireultaneous solution of equations (23), (26) and (28).
AlS. S IInperfect Column. TangentModulus Theory.

1/2

The tangent-modulus theory, originally proposed by Engesser (Ref. 1), is based on the assumption that at the critical state, no stress reversal takes place, and the critical stress, therefore, is determined only by the tangent modulus Et. This theory was abandoned early since according to the preViOUS discussion with the classical definition of instability (perfect column, bifUrcation of equilibrium) strain reversal does take place. In recent years, hOwever, this tangent modulus theory has been proved useful. Under the assumption at ho strain reversal both sides of the cross-section in Fig. AlB.Ba, will be characterized by the same linear stress distribution, corresponding to the tangentmodulus Et. Thus the buckling equation will be, EtI

Fig. AlS.9

ing the effect of a load P above the critical load Pt corresponding to tangent-modUlUS, Shanley proves the relation,
P

= Pt

where

1 +

~+
2d

1+-r) - - - - - - (32) 1--r

~:~

.;. Pu

=0 -

- - - - - - - - - (29)

It must be emphaSized that the buckled configuration Is a stable one similar to that considered in ~he refined Euler's theo~J. Shanley has recognized the fact. tnat such a stable configuration ~y exist after exceeding the tangent modulus load.

and the critical stress for sunply supported end conditions becomes,
crt

= n'E;;!
AI'

__ -

(30)

It R = PIPt , Shanley found that the relation between the variation of stress due to bending and the compreSSive strain 6t corresponding to at is:Concave side: et
- - - (33)

Since I .. + I a from (26)

I and

J:" ~

Et, it tallows
(31)

l: ~ Et

and err ~ crt - - - - - - - - -

2 (R-l)

The critical stress crt, therefore predicted by the tangent-modulus theory 1s smaller than or from the reduced modulus theory. Although tor perfect columns, the assumption or no strain reversal is in contradiction to the material behaVior in the plastic range, most expertments have given results more closely to the results by the tangent-modulus theory. To resolve this controversy, Shanley (Ref. 2), proposed the following explanation. For Simplicity, let a ~No-flange buckled column be formed by two rigid legs (see Fig. AlB.9) jOined in the middle by a plastic hinge. Assume that this column starts to buckle as soon as crt is reached. By consider-

Convex side:

In Fig. A18.10, OE3,/Et and OEa/Et are plotted against R tor ~ = 0.75. We observe that while the strain on the concave side increases very rapidly and reaches an infinite value at the reduced-modulus load, the strain on the convex side decreases initially verJ slowly. Due to this picture we can conceive that in a real column, which ras initial imperfections, the compressive strain will increase more rapidly. Furthermore, the rapid increase of 6E~ will cause a fast reduction of Et. The column, therefore, loses its usefulness after the tangent-modulus has been slightly increased. Thus the tangent-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A18.9

,
~

.
II

;,
~

f-+-+-l'::::=f7~~k-+-l-+--+-I-+-

~;"";----'~j.-+--h,.4-+-+-l----l-i--+

I
\
".~
.

-I-i--+--+-+-,.4--+~-l----l-i--+-

"
;.

, ,

r-,

A.n nl .... /(11 L_

f'oI._"""". _

m _ ~ su_ RAno (1..+)

I.

,~

HS.T..tOi _ _ .... (0.- i-. oti_l.

......

t'l~
Modulu

FIg. AlB. 11

".a

GfU,T .. TUTS

..,
FIg. AlB.lO

. . ,.. t18 1'UI1

mOdulus, even though it does not actually def1ne an unstable configuration, it represents the lower limit ot a spectrum ot possible buckled contigurations, the upper limit ot which is the reduced modulus load which corresponds to intinite deflections.
Thus to summarize, sut't1cient experi-

~ r-,

I~
<~

~ .,...
!'I..

~. l~~

f':::: ~
r-,

.......
~

t--.
~

mental results are available to show that the tailing stress at a column in the inelastic range can be round by replacing E by the tangent modulus Et in Euler's equation, or, Ocr ..
(34)

t,

S i. J.

/ ..1 _

...... _ IrSoT."'p IIHi'j.

sU:l8UlIU1l1ATlO,(l.Jf) bf IHii.,

FIg. AlB.l2

Figs. AlB.ll and 12 show how experimental results check the strength as given by the Euler equation USing the tangentwmodulus Et.

References: Ret. 1. Engesser F., Schweezeriche Bauer Zeltung. Vol. 26, p. 24, 1895. Ret. 2. Shanley, F.R., Inelastic Column Theory, Jour. Aeronautical SCiences,
1947, p. 261.

..... _.

A18.10

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

PART 2

THEORY OF THE ELASTIC INSTABILITY OF THIN SHEETS

AIS.9 Introduction.

Thin sheets represent a very common and llnportant structural element in aerospace structures since the major units ot such structures are covered with thin sheet panels. Since compressive stresses cannot be eliminated in aerospace structures, it is unportant to know what stress intensities will cause thin sheet panels to buckle. Equations tor the buckling at thin sheet panels under various load systems and boundary c ondt t ions ha. va been dert ved many years ago and are readily available to design engineers. Part C ot this book takes up the use at the many buckling equations in the practical design ot thin sheet structural elements. The purpose at this chapter is to introduce the student to the theory at thin plate instability or how these buckling equations so widely in use by design engineers were derived. For a broad comprehensive treatment at the subject at instability at structural elements, the student should reter to some at the reterences as listed at the end at this chapter.
Al8. 10 Pure Bending of Thin Plates.

curvature at the middle surface can be found aw oW approximately by omitting powers at -a ,-.-, x oy as compared to unity, as it has been done for the curvature at beams. Thus tne curvature at the deflected middle surface in planes parallel to xz and yz planes respectively are:
1
Rx

=_

aa: ax

- - - - - - - (la)

Another quantity used in the problem of plates Is the so-called twist of the middle surface given by:

- - - - - - - - - - - - - - - (lb)
The strains can now be expressed by means of curvatures and twist at the middle surface. In the case at pure bending of prismatic bar a rigorous solution was obtained by assuming that cross-sections ot the bars remain plane atter bending and rotate so as to remain perpendicular to th' deflected neutral axis. Combination at such bending in two perpendicular directions brings us to pure bending ot plates.

To dart ve the theory ct Instab1l1 ty or thin plates, we must" t1rst derive the theory or the pure bending at thin plates.

In the tolloWing the analysis will be contined to small detormations. Let x, y be the- middle plane or the plate betore bending occurs and Z be tpe axis normal to that plane. ,-X Points at the x, y plane undergo small displacements, w In the z-directlon, Which will be reterred as the y detlection at the plate. The z slope at the middle-Burrace in the x-. and y-directions x atter bending are

Fig. 2a

r------

-d{jX-. ---,
,

_ aw
ix -

ax'

iy = ay

aw

. ,, , f+-l.. ,.
i '

dr

z
Fig. 2b

Fig. 1

For small deflections, the

Let Fig. (2a) represent a thin rectangular Plate loaded by unifOrmly distributed bending

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

AIS.11

M x , My per unit length at its edges. These moments are considered pOSitive when they are directed as shown in the fIgure, I.e. when they produce compression in the upper surface at the plate and tension In the lower. Let also (Fig. 2b) be a rectangular element cut out at the plate with sides dx , dy, t , The thickness t Is cons fder-ed very small compared with the other dimensions. ObViOUSly the stress conditions at the edges at all such elements will be identical to that ot Fig. 23. Assume now that the lateral sides at the element remain plane during bending and rotate about the axes so as to remain nonmal to the detlected middle surtace. ' Due to symmetry the middle surface does not undergo any extension and It is therefore the neutral surface. hom the geometry at the above described torm ot deformation the displacements in the x, y, Z directions can be found as tollows:
A point B on the middle sur-race has been

exy

= as axay = Rxy

a.llw

2z

- - - - - - - - - - - (3)

Since we treat the problem ot plates aS,a plane stress problem. we find by means at Hook's law
"x

Ez I V (I_Va) (RX + Ry)

(l-i7i (ax'

Ez

a.llw
+ V

a"'w

ay')

- -

- - - - -

(4a)

"z = (I-V') (Ry

Ez

+ R

11

x) =

These normal stresses are linearly distributed over the plate. thickneSS. Their resUltants must be equal to Mx and My respectively:

J h/2
-h/2

"x z dydz

=MxdY =MydX

displaced to B~ by W in the z-direction. element at surface dzdy has rotated by an angle equal to the slope at the deflected middle surtace in the direction so as to remain normal to the middle surface. See
Fig. 3.
B

An

J h/2
-h/2

cry z dxdz

Substituting tram (4) we rtnd:

M

-D (a.llw + V a.ll,;) ax.ll oY

(Sa)

My==-D

(o"'w ay.ll

7J ow) ax.:! - - - - - - ( 5b )
tlexural rigidity or

where D
1

EW"3 = 12(I-V.) = the

z
FIg. 3

This angle tor small displacements is obviously equal to ~;. Thus the horizontal displacement Ux in the x-direction at a pOint at distance Z trom. the middle surface is; ux ~ - Z ax (The Sign - indicates negative displacement tor positive z ) , In a similar manner we tind the displacement in the y-df.r-ec t fon , The complete displacement system is;

the plate. It besides the tlexural moments :M:x , My, there are uniformly distributed twisting moments Mxy and Myx along the sides of the plate of Fig. 2a, these must be equal to the resultant at distributed shear torces crxy, 0xz along the sides at the element at F1g. 2b.
From eq, (3) we obtain: "xy = cryx. Rxy - 2Gz axay ~ - - - - (4<1)
MxydX
2Gz _

aw

aw

J h/2 "xyz dXdz , MyzdY -h/2

=M yx-

J h/2
-h/2
D(l-lI)

oyxz dydz, M xy

Ux -zax' Uy. -Zay' Uz =-w(x,y)

The corresponding strains are:
Ey

aw

aw

- - (2)

aaw -nay

(5c)

Equations (5) give the moments per unit length tor pure bending and twisting at a plate.
a.llN
R

=-

ayr ==

Z
y

Ala. 12

THEORY OF THE

rNSTABrLITY OF COLUMNS AND THIN SHEETS

A1S. 11 The Differential Equation of the Deflection Surface.

- - dxdy - - - cxdy + Qy cxdy

To develop thg theory at small deflections ot thin plates we ~ake one more assumption. At the boundary at the plate we assume that its edges are free to move in the plane ot the plate.

a!'!xy ax

aMy ay

=0

or

a;- - ay +

a!1xy

a!1y

Qy = 0

- - -

(6b)

From the moment equilibrium in the y-direction we find:

al1xy
--+ ---

aMx

ay

ax

Q x

- -

(6c)

By elImIr4ting Qx, Qy :rom 6 a,b,c, we tind the equilibrium relation between the moments:
y -+ax:il ay:il-

a.llMx

a:ilM

a"!1 xy

axay =

- q

(7 I

Fig. 4

Thus the reactive forces at the edges due to

the supports are normal to the plate. With these assumptions we can neglect ~~y strain in the middle plane during bending. Let us consider. Fig.4, an element dxdy at the middle plane. Along its edge the moments M XJ My, Mxy are distributed. These are the resultants of the bending and twisting stresses distributed linearly along the thickness of the plate (see eqs. 4 and 5). If the plate is loaded by external forceS normal to the middle plane in addition to the above moments there are ve~ical shearing torces QXJ Qy, acting en the sides of the element ot Fig. 4.
O'xzdr J Qy

To represent this equation in terms of the deflections w of the plate, we make the assumption that the expreSSion (5) derived for pure bending holds approxtmately also in the case of laterally loaded plates. This assumption is equivalent to neglecting the eftect on bending of the Shearing ~orces and the compressive stress cr~. This is an exter~icn or the engineerir~ theory of bending of bea~. As in the case of beams it gives good approximation for bending or plates under transverse loads. Introducing equation
--+ --+
(4)

into (7') we find:

(8)

aflw

ax'

a"w ayfl

=f

. h/2
-h/2

"yzdz -

(6)

The problem of bending of plates is thus reduced to integrating eq. (8) tor w. The corresponding shearing forces in terms of th= displacements are found from eqs 4- and 6b and c: a a.llw e-w aMxy a!1 x =--+--= - D (o-x.ll + 3y :il ) ---(9al ay ax

ax

Let Q be the transverse load per unit area acting normally to the upper tace of the plate. ConSidering the force equilibri~ in the z-direction ot the element of Fig. 4 we find:

_ aMy
Qy

- ay

~ax -

a ae w
- D-

a.llw
+

ay (ax:il

3ya)

(9b)

aQx

aQy
ay or

- - dxdy + - - dydx + qdxdy ::; 0

ax

=
(6a)

The above analYSiS is sufficIent to seek solutions ot specific problems. The general procedure is to find approximate solution ot the fourth order differential equation (8) whiCh satisties the given boundary displacement and force conditions.
Ala.12 Strain Energy in Pure Bending of Plates.

-- + -

aQx
ax

aQy
ay + q

=0

- - - - -

Taking the equilibrium ot the moments acting in the X-direction we obtain:

In evaluating the strain energy of a thin plate we shall ignore the contribution ot the shear strains which are generally small for small detlections.

ANALYSIS AND DESIGN OF

K

FLIGHT VEHICLE STRUCTURES

A1S.13

dVy + dVxy + dVyx

<lxdy
dx

ax+a x.a
z
Fig. 5a

sw a2 w

Substituting !1x ' My, M~ in terms ot the displacement tram eqs. (5) we find:

I I I
Mxy

aw
y

i;!+~dx
oy a xay

- - - - -(11)

\ \

!oJ
Mxy

\
Fig. 5b

This expreSSion will be modified later when we will consider the superposition of compressive loads in the plane of the plate which are related to the problem of plate buckling.
AlB. 13 Bending of Rectangular Plates.

The strain energy stored in a plate element is obtained by calculating the work done by the moments !1x dy , and M~ on the element during bending. Since ~he sides ot the element remain plane during bending the work done by Mxdy is obtained by taking halt the product a: :1xdY and the relative angle of rotation of the two sides at the element. Since the curvat~re in the x-direction is a.a w - ax a , the relative angle of rotation of the sides 1 and 2 of distance dx will be - - , - dx. ax Thus the work due to Mxdy is: dV x = -"2 Mx axlil dxdy Similarly the work due to MydX is: dVy - -"2 My axa dxdy
I

a'w

The general ditterential equation tor bending plates was given in section C2.3 (eq. 8). Two very usetul methods at solution have been widely used, namely, the Fourier Series MethOd and the Energy Method. Both methods will be developed in the tollowing for rectangular plates and various edgesupporting conditions. The edge support conditions are classified as tallows: a) Built-in edge or Fixed: The deflection along the built-in side is zero and the tangent plane to the deflected middle surface is horizontal. Thus if for instance the x-axis coincides with the bUilt-in edge these conditions are:
(wl y
0

aaw

----Cal

aa w

(b)

- - - -(l2a)

The twisting moment 1'1xydy also does work against rotation of the element about the x-axiS. The relative angle or rotation of the two secti ons 1, 2 is oovtoustv axay dx. the work done by MxydY is:
dVycy

a'w

Thus

b) Stmply supported edge: The deflection along the simply-supported side is zero and the bending moment parallel to this side IS also zero. Thus if the plate is sUnply supported along the x-axt s we have;

= 2' Mycy

axay dxdy

aa w

- - - - - - -

(c)

and the work due to Myxdx = MxydX is:

- - (d)

c) Free edge: The bending moment, twisting moment and shear torce along the tree side is zero. Thus if the free side coincides With the straight line x = a L we have:
(<<x)

(It is noted that the twist does not affect the work prOduced by the bending moments, neither the bending affect the work produced by the torsional moments). Thus the total strain energy per unit volume at the plate is;

x=a

=0

However, as was proved by Ki rcnorr two boundary conditions are only necessarJ to find a unique solution of the bending problem. He has shown tr~t the two last equations at the above conditions can be replaced by one condition.

Al8 14

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

These boundary conditions are satisfied take: Expressing the condition (MX)x=a = 0 in terms ot w we find the final tonn of the boundary conditions along the tree edge:

w= z

.. ..
l:

~f

we

("
'"1IlIl

:n=l n=l

sin ~ sin nny a b

---(14a)

By substituting in sq. (18) with q given by eq . (13) we find:

- - - - - - - - (12c)
In the following solutions tor various edge conditions will be developed.
1.

m=l n=l

z Cmn
~

I: amn stn-a:- sin-b n=l

mrtx

nny

- - -

- -

(14b)

Simply supported rectangular plates This relation is identity it:

Lat a plate with sides a and b and axes x, y, as shown in Fig. 6, be simply supported around the whole periphery and loaded by a distributed load q = f (x,y). Two methods ot solutions will be developed: a) Navier solution by means ot double Fourier Series:
We can always express t (x,y) in the form

and thus:
W

= n"'D

m=l n=l

.. Z

sin mnx sin ;lny a b

- - - - (140)

of a double trigonometric (Fourier) series:

~

q = !(x,y) =

l:

l:

"mIl ern

nrty a- ern "F" -(13a)

In the case of a load qo uniformly distributed over the whole surtace we have:
f

m=l m=l where: amn

4 =ab fc,a .{,b t(x,y)

(x,y) = qo = const.

mrtx nny sinTstnodXdY

"__

~WJ.

=4q, ab

fa f b s tn ~ ern nTty dxdy 0 0 a b

169

- - - - - - - - -(13b)

~----- (15a)

H-'I

-"

t::

I'

where m, n are both odd integers. If either or both are even amn = 0 and substituting in
(140 ) :

F
<\

I--

169 0

00

0.:>

W=n"D

Z Z m=l,3,S n=l,3,S

s ina- sin --0-mn {m aa

a

anx

nrty

+)' b~

ffiII1Jtq
The

- - - - - (15b)
with maximum oernecnton at the center,
Fig.

boundary conditions are:

OZn"D

W= 0 , M x ; a W = 0 , My ; 0
(1) W

at x

= a,

=a
or:

m=l,3,S n=l,3,S
-----(15c)

(-1) 2

~-l

at y = 0, y = b,

=0

(2)

~; = 0

at x

= 0,

=a

(3)11=0

This 1s a rapidly converging series and a satisfactory approximation is obtained by taking only the first term. For a square

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AlB.15

plate this approximation becomes; 0.0454 qoa

,
(for V : 0.3)

.T b/2
b/2

1'--'

~
~
Fig. 7

Ell'

f----t...-.~x

which is by 2-1/2% is error with the exact solution. The expressions for bending and twisting moments are not so quickly convergent. To tmprove the solution another series solution can be developed as follows:
b)

1
y

where for symmetry m = 1,3,5 This equation can be satisfied for all values of x it:
+ - - Ym :::

Levy alternate single series solution:

The method will be developed for uniform load qo const. Levy suggested a solution in the rorm:
W:

m-1

m"j(" a'

o - - - - -

-(1gb)

Ym (y) sin

"a

TtX

(16)

The general solution or (1gb) is:

Ym (y) =
+

where Ym is a function of y only. Each term. ot the series satisfies the boundary conditions
w

qa" [",- cos D'~

('1_

~ a + 8m. ~ a sin
ri..

~ a

a t x = a. It remains to determine Ym so as to satisfy the remaining aaw = 0 at y = b. two oouncar-y .cono i t i one '1'1 = 0, ay"
0

a'w = 0, ax'" :::

'"'!U

sin

h mny +

'"'Ina

~cos h

y] mn a

-(20)
Since the deflectIon is symmetrIc wIth respect to the x-axis it tallows that em =

A turther Simplification can be made it we take the solution in the fonn

W=WJ.+W a

Om : O. Thus:

- - - - - - - - - - - -

(17a)

where
- (17b)

(L cos h rnn:y + lL mny sm h mny) sin ~ "'m a '-ma a a

or

is the deflection of a very long strip with the long side in the x-direction loaded by a uniform load qo supported at the short sides x = 0, x = a, and free at the two long Sides. Since (17b) satisfies the differential equation and the boundary conditions at x ::: 0, X = a, the prOblem is solved it we tind the solution of:

w: q~"

(-b + Am cos h ~y

Em ~

y) sin h n: sin ~ where m 1,3,5 . Substituting this expression into the boundary condItions:
w : 0,
Am

--.
a"W a
ax'
Fig. 7).

.--, 0

a"w a
ay'

(18)

ay": 0

aw

for g : -

+ b

"2

we tind:

With w a in the form of (16) and satisfying together with wJ. of eq. (17b) the boundary a'w b conditions w = 0, ayr::: 0 at y = - 2' (see Substituting (16) into (18) we obtain:
' - _ Ym) .srn - _ : 0

cos h am+am Em sin h Cm,+lleme = 0

(An + 2Bm) cos h Om +

am

8m sin h am = 0

- -(21a)

where

am

= mrtb/2a

mton"

ID1tX

From these equations we find;

a'

- -(lea)

AlB. 16

THEORY OF THE INS:rABILITY OF COLUMNS AND THIN SHEETS

Am

2(~

tanhOm

2)
(V

nllm" cos

ham
- - - - - - - - (21b )

ux}mox

'

qa::l + (l-V)qa'n' 8

8m =
Thus:

2
n"m ll cos

Z m=1,3,S

ham

'(

Am -l-V

2V

8m )
0<:1

2:... [1- ('m

m"
cos h 20mY
b
+

tanham2)
2 cos h

(Mylmax

rJ:n

na Va- (l_V)qa'in::l
II

n.=1,3,5
- (2l!)

am
2 cosham

. 2 y sin h 2a mY ]
b

b
- - -(210)

sin mrnc

c) Solution by means of the principle of Virtual work From the discussion of the method (a) we can represent w in double Fourier Series:

The maximum deflection occurs at the middle

X

_ a 2"'

y - O

- .
H)
m=1,3,5

m-l
2

Wmax

__ 4qa'
~'D

~_am

w=
tan ham. 2)

mal

Z Cmn sin - . - sin -bn=l

nrtx

nrty

---(22a)

mil

2 cos ham

- - - - - - - -(21d)

The coefficients C mn may be considered as the co-ordinate defining the deflection surface. A virtual displacement will have the fo~:
'-'....

The summation of the first series of terms corresponds to the solution of the middle of a uniformly loaded strip, sq. (17b). Thus; OQ m-l 5 qa" 4qa' Z (-1)-2Wmax :I 384 0 - 1"(-3D m=1,3,5 fu"

~..

'-'''mIl'

'0....

1n

= -a.

sl n nny "F" -

- - (22b)

The strain energy Vi can be found by substitutlng (22a) Into eq, (6). Atter a few algebraic manipulations we find:

am tan h am' 2
2 cos ham - - - - - - - -(21e)

Z
ne L

C!nIl1l

This series converges very rapidly.

Taking a square plate, alb = 1, we find from (21a): e.t.c.

(23)

- 384""ll - - - (0.68562 - 0.00025 .... )

~'D
40

qa 40

4qa. 40

Let us now examine the deflection or the plate at Fig. 6 with a concentrated load P at the point with co-ordinates x Y = n, The increment of strain energy due to the increment ot the deflection by:
:II { ,

= 0.00406

qa

Ow

= ...

mroc nny ,-,Cmn sin -a- sin "F"

- - - - - -

(24a)

We observe that only the !1rst term of the series in (2le) need to be taken into consideration. The bending moments are found by substituting (210) Into eqs , (5). The msxlmum

is found trom (23)

n 40 ab m' n"::1 OV I =-4-Cmn (. . . . b") 0 Ctnn

- - - - (24b)

bending moments at

x =

t 'y =0

The increment of the work of the lead Pis:

ow = P0Cmn sin m! sin

are:

n~n - - - - - _ (240)

From OVi - OW = 0 we obtain:

_1i

4 11 --.1I&1II 111ill. . . . .!!! Lil

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

.

Al8.l7

,rr'ab " [--4-mn

P sin

~ .

sin

n~~ ]

Let us consider the case of a uniform load q. we write the deflection in the form:

6Cmn = a

- - - - - - - -

~(26a)

Since 6C~ is arbitrary its coefficient must be zero. Thus:

Cmn

4P

sin a

3..

sin
ll

E!!!l b
),

Where w~ is the deflection of a simply supported strip of length, a, which for the system of axes of Fig. 8 can be written (see Levy's method in previous section):
sln, a-

n"abD (m.a ,. n

artJT

mnx

TrJ

_~

4P

and w is represented by the series: 2

ee

mol

w2

L.

m-=1,3,5

Y - - - -(260) m (y) sin ~ a

sin ---a-

mrtx

sin -b-

nrty

- - - - - -(24d)

where w. being a solution of

This serieS converges rapidly. plate (alb = 1):

Z
Z 1

For a square

a"w2 ax'

.--.
a"Wa

aY'

m=l n=l By
takir~

(m2 + n 2)2

- - - -(240)

is found as in the previous section:

Y m = qaDt. (Am cos h~ +

only the first four terms we find

Em mrry sin h mny

a a
h mny)

Fa' Wmax:! 0.01121D"Nhich is 3-1/2% less than the correct value. (2) Rectangular plate with ~NO opposite edges Simply supported, the third edge tree and the fourth edge built-in or simply supported.
,--,..-----r.... x

em

sin h

~,. n __

-m

2!!!!. a cos

_ -(26d)

It is obvious that the two first boundary conditions are identically satisfied by w = w~ ,. w a The coefficients Am, Bm, Gm, Om must be determined so as to satiSfy the last tour boundary conditions. USing the conditions (25b) we obtain:
L _

+.1
y

4,
1"(lImli ,

.nm - -

em=-Dm

By the conditions (25c) we find:

.. 0_'_.
1\''''.

!3"1I1(1-1I) co.sIl%+211coShB.,.-!l(l-!l)~ !ln~a.,..(l-/l) :3+11)(1-111 cOShs..:" (3+111 (l-ll)stnh

nl/)a., .. (."1/).

;, 0_'_.
Fig.

___ _ (25s)

a., COShs,.V(l"!I)stn~B.t-II(l-lI)a"

CO!h~-(l-II)'3:l

n"",'

(3+JlI(l_lIlcos~'3,,+fl_I/)' s,.(1+11)

Assume that the edges x ~ a and x = a are simply supDorted, the edge y = b free and the edge y = 0 bUilt-in (Fig. 8). In such a case the boundary conditions are:
w -= 0,

Substituting Am, Em, em and Dm in eq, (2ed) we find the deflection. The ~imum deflection occ~r~ at the middle of the free edge.
AlS. 14 Combined Bending and Tension or Compression of Thin Plates.

a'w = aX"""

0, for x = 0, x = a (a) tor

aw w = O'ay=o,
a w (ax"".
2

(b)

a'w

--

(25 )

ay.a) - 0,

ta'W ay' .2(1-V)

1.:!!...] Clray = for y =

b (0 )

In developing the differential equations of equilibrium in previous pages, it was assumed that the plate is bent by transverse loads normal to the plate and the deflections were so small that the stretching ot the middle plane can be neglected. It we consider now the :ase where only edge loads are active coplanar

AlS. 18

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS r-------;.. x

with the middle surface (Fi~. 9) we h~ve a p lane stress problem. I f we assume that the stresses are ~niformly distributed over the thickness a~d denote by Nx' Ny, N~J' Nyx the resultant force ot these stresses per ~~it length of linear element in the x and Y-directions (~ig. 9) it 15 obvious thRt: N x .: hox ' Ny .: hoy
Nxy :;:: Nyx .: hoxy - - - - - - - (27)

Nx

dx

ax
N + oW"dx

ow

ax

y
Fig. lOa

hoyx

.-.".-_=-_--'<__.... x

Consider an element of the ~iddle surface dx. dY (Fig. loa). The conditions af force eqUilibrium in the x, y-direc~1ons are ~~ven by eq, (28), Consider now the pr-ojec t i on ot the stresses Nx, Ny, NXY in the z-directlan: (a) Projection of N x: From FIg. loa it follows t~at the resultant :Jrojectlon is:
-

I X

aw ax +

(N

x +awx ax

dx)(~+a2w rjxl

ax

a;{"l

Fig. 9

and neglecting terms of higher order:

The equations of equilibrium in the absence at body forces can be written now in terms of these generalized stresses N x , Ny, N~j by substituting frem eq. (27 ) to the equations of equilibrium.

(N

'xaxa
2

aaw +

aNx

ax

'ax
aw)

aWl

----tal

(b) Projection of Ny: By similar argument we find that this ~rojection 1S equal to:

--.
aNx

aNxy
ay

ax

aNy - 0 , ay

--.

aNxy

(N
(28)

Yay'

a w + aNy

ay ay

-------(b)

ax - o - - - - -

the other hand if the plate is loaded by transverse loads the stresses give rise to pure bending and tWisting moments only, The equations of eqUilibrium for the latter have been given in before (see eqs. 6, 7, 8). If both transverse loads and coplanar edge loadS are acting simultaneously, then tor small vertical deflections the state of stress is the superposition ot the stresses due to Nx, Ny, Nxy and MX J My, I1xy. For large vertical deflection af the plate, however, there Is interaction of the coplanar stresses a.~d the deflections. These stresses give rise to additional bending moments due to the nonzero lever arm of the edge loads fram the deflected middle surface. as 1n the case of beams. When the edge loads are ca~pressive this additional moments might cause instability and failure of the plate due to excessive vertl~al deflections,
On

,0:;,

dy

1
(c)

Nxy

-= a Zttd
y
I

Fig. lOb

Projection of Nxy and Nyx: we find:

From Fig. lOb

of

pla~es

In thiS chapter the problem of instability will be examined.

and neglecting terms of higher order:

'N aw + \ xy axay
2

the edge loads are compressive and give rise to additional bending moments eq. (8) must be mOdified.
'~en

aWl dxd ax ay ....y

- -

(c)

SimilarlY we find the projection of NyX = Nxy:

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

AIS.19

- - - - - - - (d)
Thus in eq . (6a) the terms given by (a), (b), (e) and (d) should be added (divided of course
by dxdy}:

ViN

1 II (NXEx+NyEy+Nxyd"xy) =2
0 0 0

dxdy

=
(a)

2~1 I

(N; +N; -2!INxNy + 2(1+V)Nxy') '- -

aut due to the equation of equilibrium (28) the two terms inside the parentheses in (e) are zero. Thus: ao.x dX
+

0iY y
aaw

+ q

+Nxax a + axay ::

aaw

During bending due to transverse loads or/and due to buckling we assume that the edge loads and consequently N x, Ny, Nxy, remain constant. Its variation is thus zero and we do not consider it in the following. Let us apply now the transverse load that produces bending. (~e can also consider bending due to other transverse distUrbance, which is the case of bUCkling). It u, v, are the displacements at the middle surface due to the coplanar loads (which are assumed constant across the thickness) and w the bending deflection of the plate it can be shown that the strains are:

Nyayr +

2Nxy

aaw

------(29)

ex
Ey

= au +

ax

1.. (a w ) .
2

ax

Eq , (29) replaces eq , (ea ) when edge loads are

present. Eqs. (6b) and (6c) are, however, still valid since they ex~ress moment eqUilibrium of the element dxdy ln wnlen ~ne con~rl0U~~0n ot N x, Ny, Nxy is zero. Thus eliminating Qx , Qy between (6b), (6e) and (29) we find:

=~+ 1:. ~). ay 2 ay

O'J:Y -

_ au
ay
+

av

ax +

ax . ay

aw

oW

- - - - - -

(b)

Let us apply now bending with constant coplanar stresses. Due tn stretching of the middle surface the energy Is;

(q+)Jx

ax 3 +Ny ay" +2Nxy axay)

aa w

aa'N

aaw

II
- - - -(30)

(NxEX+ NyEy+Nxy oxy) dxdy

- - - -

(0)

Eq. (30) replaces eq , (128) when edge loads are present and the deflections are large so that instability ~ight occur. The distribution of the coplanar stresses N x, Ny, Nxy can be found trom eqs , (28) by solving the plane stress problem. In the following the above theory will be applied to rectangular pla~es.
AIS. 15 Strain Energy of Plates Due to Edge Compression and Bending.

Introducing (b) into (c) and adding the strain energy due to bending, eq , (11), we find the total change of strain energy due to bending which Is:

The energy expression for pure bending, eq. (II), must be complemented to include the contribution of the edge coplanar loads. Assume that first the edge loads are applied. Obviously the strains due to the stresses N xy are: X1 N Y1 N

Here u, v are the additional coplanar displacements after bending has started. It can be shown by integrating by parts that the first integral is the work done during bending by the edge loads. For instance taking a rectangular plate this integral becomes; au oV au av] fa afa b [ Nx ax + Ny ay+ NX"J (ay +ax) dxdy =

- - - - - - - - - - -(31)

fa b (I Nxul~ + INxyvl~

) 1y + faa

(I Nyvj ~ + 1 NXYU! ~)dxdxdy - - (e)

The strain energy is:

rarblU(aNx+~) +V(aNY+~)l -o t> ax oy ay ax

I.J C

AlB.20

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

Obviously the first two integrals represent the work done by the edge loads while the second integral 1s zero due to the equilibrium equations (28). Thus the work of the edge loads is:

_ II [ N ai{+Ny au ,av au + av] IItI ay +Nxy (ay ax) dxdy- - (r) x

We assume now that for small deflec~ions the stretching at the middle surface of the plate is negligible. (This 15 the so-called 1nextensional theory at plates). In this case by zeroing the strains 1n eq, (b) and substituting in (f) we find;

buckled conf1guratio~ (ceflections different from zero) is possible. It was found in ~he case of column that this latter solution (Suler's solution) approaches asynptotlcally the first at the l1~it where the deflections become extreillely large but for even small deflections the edge load acquires a value very near the Euler's critical value. The latter teChnique is mathematically more convenient and it gives ~or plates also a ~ery good estimate of their compresSive strength. In the following we shall use this latter approach by assuming a plate with edge loads and no transverse load. Sq. (30) beco~es in this case; a"w +2 a"w ." a'w - 1 (N a'w +N a"w +21N ;;.'l.w u , ax" ax 3ay2' ay" - D ! X I y if xy axay i

" \'IN

1 = -2

II

[ Nx

(a w) .

ax

(-)' - dxd y +N j, Y ay +2Nxy ax , ay

(32a)

aw

ew awl

axr

(33)

In the strain energy expression, eq. (a) the first two terms cancel each other and the strain energy is due only to bending:
Vi

= ID!I
2

lax:!

f("lI W+

a W' lII ayliJ

li

a'" ' w a'w 2 (l-lI) [ a ax' ay" - (axay)'

]1

dxdy - - (32b)

In the absence ot transverse loads the work of external torces is Simply due to the edge loads:
- - - - - - - - (320)

ExpreSSions (32b) and (c) will be used in solVing the buckling problem by means ot the principle ot Virtual work.
AlB.16 Buckling of Rectangular Plates with Vartous Edge Loads and Support Conditions.

General discussion In calculating critical values of edge loads for which the tlat torm ot eqUilibrium becomes unstable and the plate begins to buckle, the same methods and corresponding reasonings as tor compressed bars will be employed. The critical values can be obtained by assuming that the plate has a slight initial curvat~re or a small transverse load. These values ot the edge loads for which the lateral detlection w becomes infinite are the critical values (see Part 1 for similar treatment in columns) . Another way of investigating such instability is to assume trat the plate buckles due,to a certain external disturbance and then to calculate the edge loads for which such a

By solving eq. (33) 'Ne will find trat the assumed buckling made is possible (w ~ 0) for certain defi~ite values of the edge loads, the smallest of which deter.nines the critical lead. The energy method can also be used in investigating buckling prOblems. In this ~ethod we assume that the plate is i~itially under the plane stress conditions due to the edge loads and the stress distribution is assumed as known. We then consider the buckled state as a possible configuration of equilibrium. T~e change of the work is given by eq. (32a). We inter,ret here w as a Virtual displacement though we do not use the variation symbol O. Thus the increment of work ow is ~iven by (32a) and the increment of strain energy 6Vl is given by eq. (32b). If 6~Vl for every possible shape of buckling the flat equilibrium 1s stable. If 6W~Vi for a certain shape at buckling then the flat :onfiguration is unstable and the plate will buckle under any load above the critl:al load. It OW = OVi' the eqUilibrium is neutral and fro~ this equation we find the critical load. The critical load therefore 1s found :rom the equation:
1 II [ N (aw. - '2 x ax) sw + Ny (aw ay) + 2Nxy

ax ay

dxdy

zlll(ax"

! a'w +ay')' a'w [aw -2(1-11) ax'

a'w a'w 'ay" - (axa/j,dxdY

(34)

Jl

- - - - - - - -

Here w is a certain assumed deflection which satisfies the boundary conditions (Virtual deflection) .
Ala.17 Buckling of Simply Supported Rectangular Plates Uniformly Compressed in One Direction.

Let a plate of sides a and b (Fig. 11) Simply supported around its periphery be compressed by load Nx uniformly distributed along the sides x = a and x = a. From the

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A1S.2l

;.......
~

.... I:::::::i

a+c+e+ b + d + f +

- (bl

Nx

Fig. 11

obvious solution of the corresponding plane stress problem we find that the state of stress is everywhere a simple compression equal to N x (the load at the periphery). The deflection surface of a simply supported plate when bending takes place have been found preViOusly (see eq. 14a). Its general expreSSion can be written in a double series rora:
~

ThiS fraction has some intermediate value between the maximum and the minimum of the tractions (a). It follows that i f we wish to make the fraction (b), which is Similar to the fraction of eq. 35d, a minimum, we ~ust take only one term in the numerator and the corresponding term in the denominator. Thus to make the fraction of eq. (35d) minimum, we must put all the parameters Cu , Cu , Cn except one, zero. This is eqUivalent to assuming that the buckling configuration is of simple sinusoidal form in both directions, i.e.

Wmn = Cmn sin a-

sin 0-. The minimum expreSSion tor Cmn obtained by dropping all the terms except Cmn becomes:
- - (350)

mrtx

nrty

W=

m=l n=l

Cmn s in '"':'

atn

- - - (35a)

The increment of strain energy found by sUbstituting (35a) in the right-hand side of
(34) Is:
:: fl

ab D
8

c;'
m=l

Z
n=l

It is obvious that the smallest value of N x is obtained by taking n = 1. This means that the plate buckles always in such a way that there can be several half-waves in the direction or compression but only are half-wave in the perpendicular direction. Thus ror- n = 1, eq , (35e) becomes:

The increment of work done by the external forceS is found by SUbstituting (35a) into the left-hand side of eq . (34) and Nx .,. const , Nx :: Nxy ::::r O. Thus:
6'.

(~Wtd.."'{dY = =~NX [arb o ." oX = = n'b Z m' Col..,' ea Nx Z m=l n=l

~

The value of m (in other words the number of half-waves) which makes this critical value the smallest pOSSible depenrts on the ratio alb and can be found as follows: Let us express (36a) in the form:

-- ----

(350 ) (N )

x cr

= k -n'D ba

- - - - - - - - - -

(36b I

From the equality OW obtain: naaaD =

Z
ce

OVi' solving for N x ' we

N x =

m=l n=l
Z

= a ? " + ~a)a L: Cmn (

ee

a, where k is a numerical factor depending on (bl. From (368 I and (36b) we have:
k

m=l n-l

Z m' C""

---

(35d)

b = -aa

1 aa (m + _)a - - - - - - - - - (36c) :II. b a

It we plot k against ~ for various values of the

Here Cmn is arbitrary. ~e are interested, however, to find that values of Cmn which make N x minimum. To that effect we use the following mathematical rea~oning: Imagine a series of fractions:

integer ~ = 1,2,3, ... we obta1n the curves of Fig. (12). From these curves the critical load factor k and the corresponding number of half-waves can readily be deter.nined. It is only necessarJ to take the corresponcing point (;~) as the axis of abc1ssas and to choose that curve which ~ives the smallest k. In Fig. 12 the portion of the various curves which give the critical values or k are shown by full lines. The transition from m to (m + 1) halfwaves occurs at the intersection of the two corresponding lines. Fromeq. (36c) we find:

15" 'd

.. ,

_ - -

(al

If we add the numerators and the denominators we obtain the fraction:

Ala.22

THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS

7 6
5 4 3

\\
\

'm'"
"\

n=3\
,\

~I:

"mO<'

'

A:

2
I

, I,

, , , , ,
,

, , ,

\ x," ~ /
~,,"

".-

; , ,
,
3/12 4
~

Fig. 13

'16

alb
_ (3'79.)
a~d

Fig. 12

(m.+l)b znb a -+-= ---+ a JIb a

or:
---(36d)

by introducing
00

t~e pa~~et~r:

~=vm(m+l)
b

nOD =-.a h

- - - - - -

- - - - - -

(37'))

',.Je obtain:
0e (n
a ...

Thus the transition from one to two half-waves occur for:

- n~)

'18.

- -

(37.:::)

~
rr-om

=j

1 (1+1)

=f i
= 16
and so on.

Taking any integer :n

t'NO

to three for
2 (2+1)

deflection by:

s~rface

a~d n the corres~oncing of the buckled )late i8 given

~ =j

mnx nrty \-.'mn :: (' -tnn 8 I n -a- s~ ..n -b-

(38)

Tne number or half-waves increases with the ratio alb ann for very long plates m Is veFy large.

and the cOY7espondlng Ox, cry are g1ven by (37c) which 1s a straight line in the diagram cry, Ox (Fig. 14). By plotting such lines for various pairs of m and n we !inc the region of stability and the critical combination of ox, 0y which is on the periphery of the polygon for.ned by the full lines of Fi~. 14.

Thl S means that a very long plate buck.Lee in

half-waves the lengths of which approach the width ot the plate. The buckled plate Is subdivided into squares. The critical value or the compression stress Is:
I

'
I

/
...
...............

'-0
I

\~

,-

= (Nx 1cr =
h

--=:.:.:..:=-:-,.
12 (l-v"1

t"
b"

---(36el
I

/
I

(t : thickness)
AIB.18 Buckling 01 Simply Supported Rectangular Plate Compressed in Two Perpendicular Directions.

Fig. 14 A18. 19 Buckling of Simply Supported Rectangular Plate Under Combined Bending and Compression.

Lat (Fig. 13) N X 1 Ny the uniformly distributed edge compressions. Using the same as before expression for the deflections (eq. 35a) and ap~lYlng the energy equation (33) with N x, Ny - constants (whiCh is the sol~tion of the corresponding plane stress problem) we find:

Let us consider a Simply supported rectangular plate (Fig. 13). Along the sides x :: 0, X = a there are linearly distributed edge loads ~lven by the equation:
Nx
:=

No (l -

A{) - - - - - - - - -

(39a;

ANALYSIS AND DESIGN

OF

FLIGHT VEHICLE

STRUCTURES

A1S.23

which is a combination of pure bending and ~ure compression. Let us take the deflection a 6 a i n in the foron:
w =

rle examine for each value of m the solutions of the system (38f). Starting from m 1 and denoting:

"" .z
m=l

oo

Cmn

~ TJtX 1 nny sm -,-. !5 n -b-

- - -

(39b)

Ocr =

(No)cr
h

- - - - - - - - - "

(39g)

No
I

we obtain from (39f):

I
b

C'ln

[(1 + n:l~) - Ocr ::~ (1-})]allih =co CUni al-crcrn"D Z (n:ol_jlll)a-------- (39h)
1

y
Fig. 15

eq (34) we

Substituting in the right-hand side of find the variation of strain energy:

while the increment of work 1s:

;w=.lrarbN (l_,L)(aw). dxdY= 2 Jo Jo 0 A b ax
~

:'I 4

OCIaa

.z

:1

TIlI1

ar-n=l

1 OC m:1:1 n _~ Z

m=l

:1:1

1'[

These are homogeneous equations in a~~, a~a .. etc. The system possesses a non-zero solution (which indicates the possibility of buckling of the plate) if the determinant of eq. (39h) is zero. From this condition an equation is obtained for Ocr' Obviously the system (39h) is of infinite number of equations (n = 1, 2, 3 . ). A sufficient approximation is obtained by taking a large but finite number or terms and find the solution of the rrm te determinant (USing for example digital computers). Thus a curve of ecr versus aID is obtained for m 1 like that or Fig. 12. Repeating the same calculation for m = 2, 3 etc., we find similar curves of two, three, etc. halt-wave lengths. The regions of the curves with minimum ordinates define the region or stabU1 ty as In F1g. (12).
AlS. 20 Inelastic Buckling of Thin Sheets

[4

"= . z
!1=1

Cmn:l- - - Z
1'[:01

Bb' ""

Equating (39c) and (39d) and solvir.g for No

we find:

- - - -

(39.)

The problem of the inelastic buckling of thin sheets has been extensively studied by various authors. The main difticulty in such studies is in reference to the stress-strain relations at plasticity under complex states at stress. Many controversial discussions have appeared in literature without resolving the theoretical difficulties. For this reason we will not develop the theory of inelastic buckling in this chapter. Some ot the better references on this subject are listed below.
Cnapt er- C4 presents the plasticity correction factors to use in calculating the inelastic buckling strength of thin sheets.
AlB.21 References.

"!he coefficients now Cmu are so adjusted that (Nol er becomes minimum. 8y taking the derivative of expression (38e) with respect to each coef~icient emn and equatir~ these to zero we f:"nd:

(1)
(2) (3)
(4)

Bleich,?: Buckling Strength of Metal Structures. Book by McGraw-Hill. Stowell, E.Z., A Unl~ied TheoFJ of Plastic Buckling of Columns and Plates. NACA Report S9S, 1948. Gerard and Becker: Handbook Of Structural Scabi11ty. nACA T.N. 3781, 1957. Gerard: Introduction to Structural Stability Theory. Book by ~cGraw-Rll1 Co., 1962.
........ '
'r)

"-~~~f.
"

:;:,:;0""

':-

A18.24

Courtesy of The Boeing Company, Seattle. Washington

This multiplEl exposure photograph of a Boeing supersonic transport model shows the variable-sweep wing in three configurations: forward for takeoff and landing, swept part way back for transonic flight. and swept completely back as an arrow wing for 1800-mtie-an supersonic cruise.
whour

SPECIFICATIONS (Basic Design)

Gross Weight Payload. lWge Takeoff Distance (Max. gross weight) Landing Distance Cruising Speed Takeoff Speed Approach Speed Wing Span (forward position) (aft poSition) / Length
Height

430, 000 pounds 150 passengers 4, 030 miles 6, 000 feet 5, 800 feet I, 800 miles an hour (Mach 2.7) 175 miles an hour 136 miles an hour 173 feet, 4 inches 86 feet, 4 inches 203 feet, 10 inches 48 feet, 4 inches

CHAPTER A19

INTRODUCTION TO WING STRESS AL'\[ALYSIS

A19,l TYPical Wing Structural Arrangement

For aerodyr.a~lc reasons, the wins crosssection must have a streamlined shape commonly refeTT9d ~o as an airfoil section. The aerodJ~~~lc ~orces in flight chan~e in ~gn1tude, ~irection and location. Likewise in the various landing o8erations t~e l~ads ctange in ~agnl tUde, direction and lecation, th~s the required 3tructure must be one that can efficiently resist loads causing combined tension, compression, bending and torsion. To provide torsional resistance, a portlon of the airfoil surface can be covered with a metal skin and then adding one or ~ore internal metal webs to produce a single closed cell or a ~ultiple cell wing cross-sect~on. The external skin surface which 15 re Lat tve Ly t ntn for subsonic a i rcrarr is ef:icient for resisting torsior~l shear s~resses and tension, but quite inefficient in ::-8sisting ccmpressive st~8sses due to 8end1ng of wing. To provide strength efficiency, spanwise st~f:ening ~~its commonly referr~d to as :lange stri~gers are attached to the inside of the surface skin. ~o hold the skin surface to airfoil Shape and to provide a ~edium for transferring surface air pressures to the cellular Jearn st~cture, chordwlse fo~ers and ribs are added. To transfer large concentrated loads into the cellular beam structure, heavy ri~s, cOIT~lonly referred to as bulkheadS, are used. Figs. A19.l and AlS.2 illustrate typical structural ar-rangement s of wtng cross-sections ~or sU8sonic airc~ft. The surface Skin is F'ig. a

relatively thin. In general the wins st~uc~ural flange arrange~ent can be classified ~nto two types; (1) the concentrated :lange type where flange ~aterial is connected directly to inter~al webs and (2) the distributed flan~e ~J?e where str1ng~rs are a~tached to sin between internal webs. A9.3 shows several structural a~an~e for wing cross-sections for superscr.lc aircraft. SupersoniC airfoil shapes are relativeLY thin compared to subsonic aircraft.
F'~g.
~ents

Fig. d

l
/--I ~ I "
0

-------

Fig. e

...

----

Fig. t

i _____ ] ,L

Fig. g

r--------~
"'---------'!'"
, ,

Fig. b

)"
Fig. c

-l ----

F 1::;. h

~ '---.. ,

T 1
L

T! .,-

I
Fig. A19.1

F'1,~

\::]

Concentrated ?ls.ncc8 ":ne of ,';in;:: 8ea:n. Deshed line represents sec cndar-y s t.ruc tur-e . aany cases this por t t on ~E :ab::-ic c over-ed,

::1
:'YP8S of

,
'.~1 ng

I
Seem Flange

r' eJ
j:.1.7

In

Fig. A19.2

C:)rr:Y:lon

Ar~ange:ner:t.

A19.1

--.

Al9,2
~o

ANALYSIS OF WING STRUCTURES

withstand the h1~h surf~ce preSsures a~d :0 obtain sufficient strength ~uch ~hicker ~i~g skins are usually necessary. Modern ~il:ing machineS pe~it ta~ering of skin ~h~C:~1es2es. To obtain ~ore rlanse ~ater~al intesral flange ~~its are ~achined on the thick s~in as 111~s trated in Fi~. k.
?lg. j
< ,

merrt (Fig. ,:"19.2), a cor-t t on c< t:-,e::::::,',',::":::g on

][ I[

j[ I[

Fig. k
<c

the c0mpression si~e is ~a~e effec:172 since :: 1s attaChed to closely spac8c stringers cr SQrT~ ga t t ons . However , the r Iange s tr tnger s cc t.veen cell nebs are Supported July at rib or j~lkr.ead points ~~d thus suf:er CJl~~ action no~,al to the cell coveri::g. ThiS factor reduces the flan~e allcwable c8mpressive stresses since it is not practical to space wing rl~s less ~han 12 to 18 times the flange stringe~ :e~th. Th~sJ i~ ~here were no other ~ontr2l11n; =~ctcrs, or.e could easily ~2ke calc~lations to dste~lr.e which of the above would prov~ li~htest. In general, if the tor-s i cna L ror-c es cr; t.ne .v tng are sraaLl , thus recu ir-ma cn::'y a m;n c ove r tng , the concentrated flange ty;e of Nln3 st~~cture should prOve the li~htest.

Fig. 1

Light Weight Core

In general, the flange ~aterial should be placed to sive the largest moment 0: inertia in the Z direction, which ~eans ~n general that the flan~e ~ater1al should se placed bet~een the 15 and 50 per cent 01' ':Iing chord r r'om the leading edge.
The secondary or ~istrij~ting structure aft of the st~~ctural box bea~ should be ~de as light as pOSSible and thus in ;eneral the farther forMara the rear closing web at the box be~, the lighter the wing structure as a whole. In the Layout of the main spanwise flange members bends Or changes in direction should be avoided as added weigh: is required in spl~cing or in transverse stiff~ners whl~h are ~ecessary ~o change the direction of ~he load in ~he r:ange members. If flan~e members must be s01iced, care should be ~ken not to splice them in the region of a max imum cross-section. Pur-thermcr'e , in general, the smaller ~he ~umber of fit~ings, the lighter ~he structure.
(2) Wing-Fuselage Attachment:

Fig.A19.3 Wing Sections - super-soni c Aircraft In a cantilever wing, the wing bending moments decrease rapidly sparmise frem the maximum values at the fuselage support points. Thus thick skin construction must be rapidly tapered to thin skin for wei~~t effiCiency, but thinner skin decreases allowable compreSsive stresses. To promote better effiCiency sandwich construction can be used in outer vortion 01' wing (Fig. 1). A light weight sandwich core is glued to thin skin and thus the thin skin is capable of reSisting high ccmpressive stresses since the core prevents Sheet from buckling. Al9.2 Some Factors Which Influence Wing Structural Arrangements
(1) Light Weight: ~

The structural deSigner al'H.aYs strives for the minimum weight which is practical 1'ram.a production and cost standpoint. The higher the ultimate allowable stresses, tee lighter the str~ctures. The concentrated flange ~ype of wing structures as illustrated Fi~. (a, b and c) at Fig. A19.l per.nits high allowable compressive flange stresses si~ce the fl~ge members are stabilized by both web and covering sheet, thus eliminating col~~ action, which permits design stresses aporoaching the crippl~ng str~ss of the flange ~emb~rs. Since the flang~ nembers are few in number, the size or thickness required is relatively large, thus giving a high cri~pling stress. On the other hand, this type of desi~n does not develop the effectiveneSS of the ~etal covering on the compressive si~e, ~hlch WllSt be balanced against the saving in the weight of thel r range tnember-s , I In the distributed type of
flan~e

If the airplane 1s 0= the low ~nng or the high wing type, the entire ~ing structure can co~tinue in the way of the 2i:-plane body. However, in t~e ~id-wing tj~e or se~i-low wing type, limitations may prevent extending the entire wing through the :uselage, and some ot the shear webs as well as the wing covering nust be terminated at the side of the ~Qselage. If a distributed flange type of cell structure were used, the axial load in the ~lange stringers would ~ave to be transferred to the T.embers extending through the r use Lage . '1'0 cr-evice fer this transfer of large loads requires struct~l weight &~d thus a concentrated fl~nge t:~e of tox str~cture X:~ht prove the DRst type of structur-e.
(3) Cut-ours in Wing Surface:

arrange-

The ideal arranse~ent where continUity of structure is aa i nta mec oyer t ne en t i r-e sur-race

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Al9.3

of the structural box 1s sel~om obtainec in acair?lane cesigr. due to cut-outs in the Ning su~~ace ~or such items as retractable landing gears, ~il compartments and bomb and gun bays. r: the distributed ~~ange type of box beam is used, they are interrupted at each cut-out, which requires that means nust be ?rovided for drifting th~ flange loads around the opening, an arrangement wntcti adds weight because conservative overlapping assumptions are usually made in the stress analysis. The additional str~cture and riveting to provide for the transfer of flange load around large O?enings adds c8nslcerably to the production cost.
t~al

the lower side of the Wi~g. They are usually fastened to two spanwise stringers With screws and the re~ovable ~anels are ef:ective i~ res Let Ing bending and shear load. (See Fig.A19.5)

Fig. Al9. 5

Removable panel for assembly and inspection purposes.

?or landing gears as well as many other installations, the ing cut-outs are confined to the Lower- surface, thus a structural arrangement as illustrated in Fig. A19.4 is quite common. The upper surface is of the distributed flange type whereas the lower flange ~terial is concentrated at the two lower corners of the box. In the normal flying conditions, the lower surface is in tension and thus cell sheet covering bet~een the cut-outs is equally effective in bending if shear lag infl~ence is discounted. For negative accelerated flying conditions, the lower surface is in compression thus sheet covering between corner flanges would be ineffective in bending. However, since the load factors in" these flight conditions are approximately one half the normal flight load factors, this tneffectiveness of the lower sheet in bending is usually not c~itical. Cut-outs 11~ewise destroy the continuity of intermediate interior shear webs of such sections as illustrated in Figs. (c and i)3 and the shear load in these interrupted webs must be transferred around the opening by special ~ulkh3ads on eac~ side of the cut-out, which means extra weight.

Cutouts in the wing structural box destroys the continuity of the torsional resistance of the cell and thus special consideration must be given to carrying torsional forces around the cut-out. This special problem is discussed later.
(4)

Folding-Wings:

For certain airplanes, particularly Carrier based Naval airplanes 3 it is necessary that proviSion be made to fold the outer wing panels upward. T~is dictates definite hinge pOints be~Neen the outer and center wing panels. If a distributed flange ty~e of structure is used, the flange forces must be gathered and transferred to the fitting points, thus a compromise solution consisting of a ~ll number of spanwise members is cammon practice.
(5) Wing Flutter Prevention:

Fig. A19.4

With the high speeds now obtained by modern airplanes, carefUl attention to wing flutter prevention must be given in the structural layout and design of the Wing. In general, the critical flutter speed depends to a great extent on the torsional rigidity of the wing. 'Ahen the mass center of gravity moves aft of the 25 per cent of chord paint, the critical flutter speed decreases, thus it is important to keep weight of the wing fon'ard. At high speeds where ncompressibility" effects became ~portant, the torsional forces on :he wing are increased, which necessitates extra skin thickness or a larger cell. DeSigning :or flutter prevention is a highly specialized problem.
(6) Ease and Cost of Production:

cases, c~t-outs in the leading edge are necessary due to ~ower ?lant installaticns, landing gearxells, etc. Furthe~ore, in many airplanes, it is desirable to ~ake the leading edge ~crtion remova~le for inspection of the many Ins ta Ha t Ions cvnf ch occupy thiS space in the oor-t.tcn of the wtng near the fuselage. If such is t~e case, ~hen an interior Neb should be lecated near t~e ~ront Jf the wing section.
I~ ~ny

Ins~ectio~ 2QCrS ~8r the central oor~~cn of box beam s t ruc t ure ar-e usua Ll.y Ioc a t ec on

The airplane industry is a mass production industry and therefore the structural layout of the wing must take into account production nethods. The general tendency at this tL~e is to deSign the wing and bOdy structure, so that s~b-assenblles of the various parts can be made, which are finally brou~~t together to form the ~iQal assemblY af the wing pa~el. To make this ~rocess ef~!cient requ1res car9t~1 consideration 1~ the ieta:ls and layout ot the win~ structure. ?hotogra?h A1S.6 illustrates the sUb-asse~bly

------

-,

<_C'--

A19.4

ANALYSIS

OF

WING

STRUCTURES

.
'

.- .

,
::1')1v
.,\, /

\
\

Designing To Facilitate Production.

Photographs by courtesy of North American Aviation Co.

ANALYSIS AND DESIGN OF

break~down

FLIGHT VEHICLE STRUCTURES

A19.5

of ~he struct~ral parts of an airplane of a leading airplane company. Fabrication and asse~bly 0: these ur.its ~ermits the installation of ~uch eGui~ent jefcre assembly of the units to tr.e final assembly.

A19.3 Wing Strength Requirements

~NO major strength require~ents must be satisfied in the structural design of a wing. They are: - (1) Ur.der the applied or limit loads, no part of the structure mus~ ~ stressed beyond the yield stress of the material. In general, the yield stress is that stress which causes a permanent strain of 0.002 inches per inch. The terms applied or l1mit refer to the same loadS, which are the maximum loads that the airplane should encounter during i~s lifetime of operations. (2) The structure sr~ll also carry Design Loads without rupture or collapse or in other words failure. The magnitude of the Design Loads equals the Applied Loads tines a factor of safety (F.S.). In general, the factor of safety for aircraft is 1.5, thus the structure must withstand 1.5 times the applied loads without failure. In miSSiles, since no human passengers are i~volved, the factor of safety is less and appears at this ti~e to range between 1.15 to 1.25.

is necessary. A~othe~ l~portant distinction is that a statically dete~inate wing structure nas just e ncugh member-s to produce stability and if ene ~ember is removed or ~ails, the ent1re st::-ucture will usually tail, whereas a statically indeterminate structure has one or more additional ~embers t~~n are necessary ~or static stability and thus some members could ~ail Without causing the entire strJcture to collapse. In other words, the structure has a fail safe characteristic in that a redistribution of internal stress can take place it some members are over loaded. In general, statically indeterminate structures can be deSigned lighter ar.d with smaller overall deflections.
?"!:S'I'HODS OF STRESS

ANALYSIS :;'OR STATICALLY INDE"I'E:R..:."1INATS ":HNG STP.l'CTITRES

TWo general methods are commonly used, namely,

(1)
(2)

FleT~al beam theory with assumptions.

sL~pli~yir~

SolVing for red~"1dant forces and stresses by applying the principles of the elastic theo~J by various methods such as virtual work, strain ener~J, etc.

Aircraft factors of safety are rather low compared to other ~ields ot structural design, chiefly because weight saving is so important i~ obtaining a useful transportation vehicle relative to useful load and performance. Since safety is the ,aramount design requirement, the correctness of the theoret1cal design XllSt be checked by extensive static and dynamic tests co verify whether the structure will carry the design loads without failure.
A19.4 Wing Stress Analysis Methods

In ~any of the previous chapters of this book, internal forces were calculated for bo~h statically determinate and statically indeter.ninate structures. The internal loads in a statically deter.ninate str~cture can be found by ~~e use of the static equilibrium eq~tions alone. The over-all structural arr~ngement of members is necessary, but the size or shape of no individual ~ember is required. In other wordS, design consists of finding Inter~~l loads and then supplying a member to carry this load safely and efficiently. In a statically indeterminate structure, addltlopzl equations beyond :he static equilibrium equations are necesscTJ' ~o find all the 1r.terr.al stresses. The additional equations are supplied tr7.n a consideration of structural distortions, which ~eans that the size a~d shape and ~i~d of ~aterlal for members of ~he structure ~ust 8e known barer-a internal s tr-easea can be deter~ined. ~hls fact means a trial and er~0r method

The second method is no doubt more accurate since less assumptions are necessary. A wing structure ccmposed of several cells and many spanwlse stringers is a many degree redundant structure. Before the development of high speed computing ~chinery, so-called rigorous ~ethods were not usable because the computing requirements were ~Pcssible or entirely ilnpract1cal. However, present day computer facilities have changed the situation and rigorous methods are now being ~ore and more used !n aircra~t structural analysis. Art. A7.9 and AS.IO in Chapters A7 and A8 present ~trix methods for finding deflections and stresses to oe used with computer faCilities.
A19.5 Example Problem 1. 3-Flange Wing. Single Cell

Fig. AIS.7 shows a portion of a cant~lever Wing. ~o prOvide torsional strength a single closed cell (4) is fomed by the interior web AB and ~he ~etal skin cover for~ard of this web. Thin sheet 1s relatively weak in resistir~ compressive stresses thus 3 flange stri~gers A, B and C are added to develop efficient ber-ding resistance. The structure to the rear of suar AB is referred to as secondary str~cture and 20nsists of thin metal or tabric covering attached to chordwise wing ribs. The air load on this Dortion ~s therefore carr~ed [orNard by the r~bs to the single cell beam. A wing 1s subjected to many flight condi-

A19.6

ANALYSIS

OF

WING STRUCTURES

tions. The engineers who cal:ulate the ~ppl:ed loads on the wing usually refer the resulting shears and moments to a set o~ conver-ient x, Y and z axes. Fig. A19.7 shows the location of these reference axes. The job o~ the stress engineer 1s to ~rovide structure to resist these loads safely and efficiently. ~he ge~eral procedure is to find the stresses or loads 1~ all parts of the cell cross-section at several stations along the spanwise direction and tram these loads or stresses proportion the required areas, thicknesses and shapes. In this example, the internal leads will be calculated for only one section, n&~ely, that at Station 240. It will be ~ssumed that the design critical loads from the critical flight condition are as follows.

1-7"-=22" -1/2

~60.500

VZ"1/ IOO~MY"
'5

Fig. A19. 8
SOL~ION.

~10"""""

Vr700

Mx"'l,lOO,OOO

Mz",aoooo

ASSUI1?T'IONS: - It will be assumed tnat the 3 flange stringers A, Band C develop the 3r.tire resistance to the bending ~oments about ~he Z and X axes. FJr skin ~~cer compression t~~s assumption 1s nearly correct since the skin will buckle under relatively low stress. Since sheet can take tensile stresses, this assumption is conservative. However, since the thin sheet cover must resist the shear stresses we will ~ake this conservative a3s~~.~tion. ~he main advantage of this approximate assumption is that it ~kes the structure statical:y determinate. Fig. Alg.a shows ~~e wing c~t at Station 240. The wlliL~own forces are the three axial loads in the stringers A, Band C and :he three shear flows Qab' ~c and qbc on the three sheet panels, ~king a total of 6 unknowns and since there are 6 static equilibrium equations available for a space structure, the struc~ure is statically determlr~te. Since ~he size and shape of flange ~e~bers A, B and C are unknown, we guess their centroid locations as indicated by the dots in Fig.A19.8. The axial load in each at the 3 stringers has been replaced by its x, y and z components as shown on the figure. The external applied loads -a:re given at the reference origin (0) as shown
in Fig. A19.S.
~e now apply the equations of to tind the 6 '~owns . 8Guil~bri~~

Mx l1z My

= 1,100,000 in. lb. = 80,000 in. lb. = 460,500 in. I'D.

V z Vx

= 11,500 lb. = 700 lb.

Fig. Al9.8 shows these shears and ~oments referred to the reference axes with origin at point (0). Moments are represented by vectors with double arrow head. The sense at the moment tallows the right hand rule.

t.J ii' ,2. 33 ,t r-..L~" -=::::' :::J) I T

3.
STA.O
32"-"""t

7.27"

,.. "'
I

;<

PC!.:J.ts A

To find Cy take and 8,

~coents

about Z axis

t~~ough

l:l1z(ab) - - 22 Cy + 60000

Cy = 3636 lb.

= 0,

whence

STA.240

r=
Fig. A19.7

The result CJmes out with a .plus sign thu~ indicating that the assumed sense of tension was correct. To find By take moments about
'thr-u (A).
~

X ax:s

lMX(A) = - 11 By + 1,lCO,OOO + 3636 x

0.125

=0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A19 7

whence, By To find
ZFy
Ay

= 100,043
take ZFy

lb. tension as

assumed.

-9065 whence qbc = """'6'65:: - 13.66 Ib./in. and having the sense as aSSQ~ed.

=0
+

=-

100,043 - 3636

whence, Ay 103679 lb. and compression as assumed. Since the d~rection of the 3 stringers is known, we can find the X and Z components of the stringer loads by simple geometry. The r, x, and z length components of the three stringers from the dimensions given in Fig. A19.7 are found to be,
Member
A

Ay

= 0,

In the above moment equation the moment of the shear flow qbc about point (a ') equals qbc times twice the area of the cellar 665.
To tind Qac take ZFx
ZFx

=0

=- 2879 +

2584 + 316 - 22 Qac - 22 X 13.66 - 700 = 0

whence Qac = 35.45 lb./in. with sense as assumed.

To tiM 'lab take ZFz
ZFz

=0
11500 - 11 'lab

=-

1296 - 12.'10 + 45 + 35.45 x 0.5 +

s
240 240 240

x
6.2 6.2
20.93

z
3 3 3

11.5 x 13.66

=0

whence, Qab = 805 lb./in. The loads on the stringers and sheet panels have now been determined. The axial load in the stringers is practically equal to their y component since axial load equals their y torce component div!ded by the cosine of a small angle. The stress engineer would tind s1m1lar stresses at a number ot stations along the spa.~. These 6 stresses are generally reterred to as primary stresses. Usually in most structures there are secondary stress effects which must be considered betore tinal member sizes can be determined. For example, internal webs at a box tYJ)e beam are designed usually as semitenSion tield beams. Tension tield beam theory shows that the flange members are Subjected to additIonal stresses besides the primary stresses as tound above. The subject at secondary stresses and the strength deSign ot members and their connections to carrJ given stress loads is taken up in detail in Volume II.
A19.6 E::o::unple Problem 2. Metal Covered Wing With Single External Brace Strut.

The force components are therefore: Az Ax Bz

= 103679

x 3/240

= 103679 x = 100043 x 3/240 = 1250 lb. Ex = 100043 x 6.2/240 = 2584 lb. Cz = 3636 x 3/240 = 45 lb.
Cx 3636 x 20.93/240

= 1296 lb. 6.2/240 = 2579 lb.

=316

lb.

Fig. A19.9 shows these forces applied in the plane of the cross-section at Station 240, together with the unknown shear flows and the external forces acting in the plane of the cross-sectlon.
1
1296 " ---22'~' ....".0:-'111fo!

2Jil

3181

267,-+ A

'lac

2161 C 6
45 /

%c

Fig. A19 9

To find qbc take moments about point (a')

:Mar =

+ 2679 X .5 - 2584 X 11.5 - 316 x 0.375 - 45 x 22 + 700 x 12

- 11500 X 39 + 460,500 -

(665)

=0

coo

In Chapter AZ, the stress analYSiS ot an externally braced fabric covered monopl~~e wing was considered. To provide sufficient torsional strength and rigidity, two external brace struts were necessary. However, 1t a wing 15 metal covered, a single external brace strut can be used, since ~he closed cellar cells for,ned by the ~etal sheet covering and the internal webs provide the torsional resistance and the wing can be des fgned as a S1.mply supported beam with cantilever overhang. An excellent example of this type ot wing structure is the Cessna aircraft Model 180 as shown in the photograph. An excellent airplane relative to pertormance, ease of ~nutacture and maintenance. To introduce the student to the general approach Of stress ar41YZing such a wing struc-

A19. B

ANALYSIS OF

WING STRUCTURES

~ions

ture , a Liznl t ed -nscuss i c- -"i':h a few ca LcuLa-. ~111 be ~rese~ted.

4fii\

--

~-;.;~

.'~>;~'-~

Fig. A19.1D and AIG.Il shows the wl~g d1mensions ane genera: str~c:~~3.1 :ayout of a monoplane wing wl~h o~e external brace s:rlt. The wing panel is attached ':0 ~~sela~e by Single ;in fl':tlngs at points A and 8 with ~1n axes parallel to X axis. The ~ating 1~~8 ot the :lttlngs at paint A are ~ace snug :1t but those at B with sane gap, thus ~ag reaction of wing loads on fuselage is resisted entirely at fitting A. Since the fittings at A and B cannot resist moments about x axiS, it 1s necessary to add &~ external brace strut DC to make structure stable. The p&~el str~c:ure consists of a main spar ACE anG a rear spar 3F. Tte entire ?anel is covered with metal skin forNard of the rear spar.

Cessna Aircrart Model 180 Metal Covered Wing with One External Strut

-t-tr-

....L

27"

ala- Front Spar

15"

-'C-90"

A SL~pllfled air load has been assumed as shown in Fig. A19.12, ~~mely, a unifc~ load w = 30.27 Ib./in. of span acti~g at the 30 ~er cent of chord point. ~ben this resultant load is resolved into z ~nd x :o~~or-ents the results are Wz = 30 lb./in. and Wx = ~ lo./in. as Shown in Fig. A19.l2.
C

l---

BO"

-.---Strut

C. y

t' L.

Fig. A19.l0

The general physical ac:ion of the wing structure in carry~ng these al~ loads can Je considered as 3 rather distinct actter-s, namely, (1) The front spar ACE resists the bendl~g moments and shea~s due to load w . , z (2) The skin and webs of the two-cell tube resists all mrr~ents about y axis or broad:y speaking tors~onal moments, (3) the entire ~anel cross-sect ton resists the bending moment and flexural sher due to drag load wx, with the top and bottJm skin acting as webs and the t~o spars as th~ flanges of this box beam. General Calculations; -

The unknown external reactions (see Fig. AI9.ll) are Ay, Az, Ax, By, Bz and DC., or a total of 6. Since 5 static equations of equilibrium are available, the reactions are statically deter.r.lnate. Reaction DC is also the load in brace strut CC.

~_J

"/m4t~m.
-tilFig. A19.l2

30.27

To :lnd reaction DC take moments about x axis through pOints A, 3

ZMx(AB)

= (-

=0

170 x 30 x 170/2) + DC(60/99.4)60

whence, DC = 8979 lb. (T::e Sign cernes out plus so the sense assumed in Fig. A:9.11 was correct.) The load in the strut 1s therefore 8979 lb. ~ensicn.
To find B., take moments about a z. exts thrcugh point rAJ.
~~(A)

=-

(4

x 170 x 170/2)

27 By

=0

Fig. A19.11

whence, By
aesumed ,

= 2141

lb. ac t Ing wt tb sense

ANALYSIS AND

DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A19.9

To find A y take ZFy

ZFy

=0

=2141

- 8979 (80/99.4) + Ay

=0

points A and B. This couple force equals the moment divided by distance between spars or 15300/27 = 566.7 lb. Front spar (ACF) loads due to Wz

whence Ay To tind
through (A).

= 5085
take

lb. about y axis

=30 lb/in:

Bz

~oments

lliy(A)

=+

(30 x 170 x 3) - 27 Bz

=0

t t t t t lltttt ===c y-,-"...c:::= __

7226/
~~.

w'" 30 lb./ln.

whence 8 z

= 567

lb. acting as assumed.

To tind Az take ZFz = 0

ZFz

Cz.:t5329 59
Flg. A19.14

80

=170
+ Az

x 30 - 567 - (8979 x 59/99.4)

=0

80"

whence Az = 796 lb.

actlr~

as assumed. ACE.
(A)

Fig. A19.l4 shows free body of front spar To rind strut load DC take moments about

+-

90"

--i

The tittings at A and 8 should be designed to take the reactions at these pOints as found above. The external strut DC and its end fittings ~t carry the tension load ot 8979 lb. The next step is to find the stresses and loads on the structural parts of the wtng panel.
~oments

lMA

=(- 30 x
=
0

170 x 170/2) + 60 (DC x 80/99.4)

Whence DC 8979 lb. value previously found.

Tension which checkS

Since the spar ACE ~~ resist the bending about x axis the airloads in Fig. A19.12 are ~oved to the spar centerline as shown in Fig. A19.13.

The y and z components of the strut reaction at C will then be,

<:-Il!"/ln~
Flg. A19.13

.,....,=:::-J~3.30.90".

30#/in.

= 8919 Cz = 8979
Cy

(80/99.4) (59/99.4)

= 7226 = 5329

lb. lb.

These values are indicated on Fig. Al9.l4. To find A y take ZFy ZFy : - 7226 + Ay

=0
hence Ay

The torsional moment or 90 in. lb. per inch ot span is resisted by the cellular tube made up of two cells (1) and (2). In many designs the leading edge cell Is neglected in resistir~ the torsional moments due to many cutouts, etc., thus cell one could be assumed to provide the entire torsional shear resistance and the shear flow tor this case would be q = Mv/2A where My equals the torsional moment at a ~iven panel section and A the enclosed area of cell (1). It both cells were considered ef~ective then the sheet thicknesS is necessary before solution for shear flow can be computed. Refer to Chapter A6 for computing torsional shear flow in ~ultiple cell tubes. The maxfmum torsional moment would be at the fuselage end of the wing panel and its ~gnitude would be 170 x 90 = 15300 in. lb. Since the top and bottom skin is not attached to fuselage, this torsional moment must be thrown off on a rib at the end ot the panel and this rib in turn transfers this moment in terms or a couple reaction on the spars at

=0,

=7226 lb.

In finding the reaction Ay previouslY, the value was 5085 lb. The difference is due to the drag bending moment which tends to put a tension load on front spar and compression on the rear spar.
wx"'4*/in. lffftlfttffff BY-B Rear Spar

Front S
C

Fig. A19. 15

Fig. Al9.15 shows the air drag load of 4 lb./in. The bending moment on panel at a distance y tram the wing tip equals W x (y) (y/2) = 4y /2 2y. The axial load Py in e I trier spar at any distance y from tip thus equals bending moment divided by arm ot 27" or Py = Zy'iJ/ 27 = .07 4y a. The axial load at points A and B thus equal .074 x 170 = 2141 lb. (tens ton in front spar and compreSSion in rear spar). Thus each spar is subjected to an axial load

A19.l0

ANALYSIS OF WING STRUCTURES

increasing from zero at tip to 2141 lb. at fuselage attachment ,oints and varying as y~. At point C on from spar the axial tension load would be .074 (90') = 600 lb. The design of the tront spar between points C and E would be nothir~ more tillL~ a cantilever beam subjected to a bendir~ torces plus an axial tensile load plus a torsional shear flow. The design of the spar between points C and A is far more complicated since we have appreciable secondary bending moments to determine, which must be added to the pri~ry bending moments. Fig. Al9.lsa shows a free body or spar portion AC.
w::; 30

would ~3ve t8 be replaced by two fltti~gs, one on the upper flange ane the other an :he lower flange in order to be able to resist a ~x moment. Fig. A19.l6 shaNs this modification. The fitting at B could remain as before, a single pin fitting.

fltltt~ ~Mc::;30x90x45
5085-<:- -. -

--

'

Fig. Al9.16

Fig. Al9.15a

The lateral load of 30 lb./in. ~ends the beam upward, thus the axial loads at A and C will have a moment ar.n due to beam deflection which moments a&e referred to as secondary moments. To find deflections the beam moment of inertia must be known, thus the design of this beam portion would fall in the trial and error procedure. Articles A5.23 to 28 of Chapter AS explain and illustrate solution of problems involving beam-column action and such a procedure Would have to be used in actually deSigning this beam portion. The rear spar SF receives two load systems, namely a varying axial load of zero at F to 2141 lb. at B and the web of this spar receives a shear load rrcr.m the torsional moment. The rear spar is not SUbjected to bending momen~s. In Fig. A9.l0 the secondary structure consisting or chordwise ribs and spanwise light stringers riveted to skin are not shown. This secondary structure is necessary to hold wing contour Shape and transfer air pressures to the box structure. This secondary structure is discussed in Chapter A2l. The broad SUbject at designing a member- or structure to withstand stresses sarely and effiCiently 1s considered in detail 1n later chapters.
A19.7 Single Spar. Cantilever Wing Metal Covered

The stress analYSiS of this wing would consist of the spar AE reslstl~g all the Mx mcments and the V z shears and acting as a cantilever beam. The torsional moment about a y axis coinciclng With spar AE would be resisted by shear stresses in the cellula~ tubes tanned by the skin ane the spar webs. The d~g bending and shear fo~ces would be resisted by the bean whcse flar~es are the tront and rear spars ant the web being the top and bottcrm skin.
A19.8 Stress Analysis of Thin Skin - Multiple Stringer Cantilever Wing. Introduction and Assumptions.

The most common ~ype of wir~ con5t~uctlon is the ~ultiple stringer type as illustrated by the six illustrative cross-sections in Fig. A19.2. A structure with ~any stringers and sheet panels is statically indeter.ninate to many degrees with respect to inte~al stresses. Fortunately, structural tests of complete wing structures show that the Simple beam theory gives stresses which check fairly well with measured stresses it the wing span is several times the wing chord, that sweep back is minor and wing is free of ~jor cutouts and discontinUities. Thus it is common procedure to analyze and design a wing overall by the beam theorJ and then investigate those portions of the wing where the beam theory ~y be in err~r by USing more rigorous analysiS methods such as these explained and illustrated in Art. AB.IO of Chapter AB.
ASSUMPTIO~S

- BSAM

~~ORY

A Single spar cantilever wing with metal covering is often used particularly in light commercial or private pilot aircraft. Suppose in the single spar externally braced wing of Fig. A19.Il, that the external brace strut DC was removed. Obviously the w1~g would be unstable as it would rotate about hinge fittings at ~oints A and S. To make the structure stable the Single pin fitting at (A)

In this crapter the wing bending and shear stresses will ~e calculated USing :he unsymmetrical beam theory. The two ~in assumptions in this theory are: (1) Transverse sections of the 8eam originally plane before ben=ing remal~ plane a~ter be~di~g of beam, This assumption means that Longftudf -. nal strain varies cirectly as the jlstance from

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A19.11

the neutral axis or strain yariation is linear. (2) The longitudinal stress distribution is directly proportional to strain and therefore trom assumption (1) is also linear. This assumption actually ~eans that each longitudinal element acts as it it were separate tram every other element ~~d that Hook's law holdS, namely, that the stress-strain curve 15 linear. Assumption (1) neglects strain due to shear stresses in skin, which influence is commonly referred to as shear lag effect. Shear lag effects are usually not ilnportant except near major cut-outs or other major discontinuities and also locations where large concentrated external forces are applied. Assumption (2) is usually not correct if elastic and inelastic bUCkling of skin and stringers occur before failure of win~. In applying the beam theory to practical wings, the error at this assumption is corrected by use ot a so-called effective section which is discussed later.
A19.9 Physical Action of Wing Section in Resisting External Bending Forces from Zero to Failing Load.

corner flange members (a) and (b) are stabilized in two directions by the skin and webs and usually fail by local crippling. Now continuing the loading of the wing after the skin has buckled, the stringers and corner members will continue to take additional compreSSive stress. Since the ultimate strength of the stringers is less than that ot the corner members, the stringers (c) will start to bend elastically or inelastically and will take practically no ~Jrther stress as additional strain takes place. The corner members still have considerable additional strength and thus additional external loading can be applied until finally the ultimate strength ot the corner members is reached and then complete failure of the top portion of the beam section takes place. Therefore, the true stress - strain relationship does not follow Hook's Law when such a structure is loaded to failure. In the above discussion the stringers (c) were considered to hold their ultimate buckling load during considerable additional axial strain. This can be verified experimentally by testing practical columns. Practical columns are not perfect relative to straightness, uniformity of material, etc. Fig. Al9.l8 shows the load versus lateral deflection ot column midpoint as a column is loaded to failure and fails by elastic bending. Fig. Al9.l9 shows similar results when the failure Is inelastic bending.
Fig. A19.18 Fig. A19.19

Fig. Al9.l7 shows a common type of wing cross-section structural arrangement generally referred to as the distributed flange type.
skin

a
web

'-~vr--~

b web

c(stringers)

La

Fig. A19.17

The corner members (a) and (b) are conSiderably larger in area than the stringers (c). The skin is relatively thin. Now assume the wing is subjected to gradually increasing bending forces which place the upper portion of this Wing section in compression and the bottom portion in tension. Under small loading the compressive stresses in the top surface will be small and the stress will be directly ?roportional to s~rain and the beam formula 0c = Mxz/l x will apply and the moment of inertia Ix will include all of the crosssection material. As the external load is increased the compressive stresses on the thin sheets starts to buckle the sheet panels and further resistance decreases rapidly as further strain continues, or in other words, stress is not directly proportional to strain when sheet buckles. Buckling of the skin panels however does not cause the beam to fail as the stringers and corner members are lowly stressed compared to their failing stress. ~he stringers (c) are only supported transversely at wing rib points and thus the stringers ac~ as columus and fail by elastic or inelastic bending. The

B '" Central Deflection

6 '" Central Deflection

The test results show that a compreSSion member which fails in bending, normally continues to carry approximately the ~aximum load under considerable additional axial defo~tion. Thus in the beam section of Fig. Al9.l? when the stringers (c) reach their ulti~ate load, failure of the beam does not tallow since corner ~embers (a) and (b) still have remaining strength.
A19.10 Ultimate Strength Design Requirement

The strength design reqUirements are: (1 )

Under the applied or limit loads, no structural member shaIIibe stressed above the material yIeld point, or in other words, there must be no permanent de:crmation or deflection of any ?art of the structure.

A19.12

ANALYSIS OF

WING STRUCTURES

(2 )

Under the ~ loads which equal the applied loads ti~es a factor of safety, no failure of the structure shall occur. The usual factor ot safety for conventior41 aircraft is 1.5, or the structure must carry loads 50 percent greater than the applied loads wi~hout failure. The higher the stress at failure ot a member the less material required and therefore the less structural weight. The stress engineer thus tries to deSign members which fail in the inelastic zone.

-l. f" .- -!W, i';' .,- r,L

I

, . , , I ,

I ,

, I
-ow,

I I

I 1 .

itFig. A19.21

kf-. ~ I';'

'\

The bending stress beam ~ormula crc = MclI does not take care of this non-linear stressstrain action and thus some modification of the moment or inertia of the beam cross-section is necessary it the ultimate strength of a wing section is to be computed fairly accurately. The stress engineer usually solves this problem by using a modified cross-section, usually referred to as the effective cross-section.
A19.11 Effective Section at Failing Load.

stress equal to the stringer stress. l~ Fig. A19.2l ~N is the et:ective sheet width ~o go with each stri~er. The total stress on the effective Widths carrying a uniform stress eq~al to the stringer stress equals the total load on the sheet panels carrying the actual varying stress distribution. The equation for effective width is usually written in the ~or.n,
2w = kt

(E!crstl.
= 1.9,
whence
t

a widely used value for K

In order to use the beam formula which assumes linear stress-strain relation, corrections to take care of skin ~uckling and stringer buckling must be fOr.TIulated. The effectiveness of the skin panels will be considered first.
When a compreSSive load 1s applied to a sheet stringer cambir~tion, the thin sheet buckles at rather low stress. For further loading the compressive stress various aver the panel width as illustrated in Fig. Al9.20. The stress in the sheet at the stringer at~ach ment line is the same as in the stringer since it cannot buckle and therefore suffers the same strain as the stringer. Be~Neen the stringers the sheet stress decreases as shown by the dashed line in Fig. Al9.20. This variable stress condition is difficult to handle so the stress engineer makes a convenient substitution by replacing the act~l sheet with its variable stress by a width ot sheet carrying a uniform
Stress Distribution on Stiffener

2W = 1.90
0st

- - - - - - - - - (1) (E!crst)in stringer

=stress

Therefore if we know the stress in the stringer we can find the width of sheet to use With the s~ringer to obtain an effective section to take care ot the sheet buckling influence. Effective Factor for Buckler
Stringe~s.

Consider the beam section in Fig. Al9.l7.

If we take a stringer (c) and attach a piece of

'r '

1 /

IK /:1' ,' I , I ',

I \
,
,

tn,

7"'-. "'>J
'I

f ;;,>,
III
I

stress Distribution on Shee_~t _

1 1\

H j,! -I-H
'T

, 1\, - 'Ii \-r' / 1 - ,

,

, I I' ' ' !

I',

'I

'I"

I; , : ',

VI"
,

II

I -,__i.Ost

-'lac

.L

~~
~

I~ 1--,
,

l-r

, -

Ii Ii I
I~

%'-1

1(

sheet to it equal to ZW, the effective width and test it in compresSion and brace it in a plane parallel to the Sheet, the resulting test stress versus strain Shortening curve (c) of Fig. Al9.22 will result. ~he length of the test specimen would equal the rib spaci~g in the wing. The corner members (a) or (b) in Fig. A19.l7 being stabilized in two directions will tail by local crippling, thus it a short Piece of this member is tested to failure 1n compreSSion the test curve (A) in Fig. Al9.22 is obtained. C~le (C) shows that the strir.ger holds apprQx~tely its maximum load ror a considerable strain periOd. C~ve (A) shows that for the same unit strain member (a) can take considerable higher stress. It we take a unit strain ot .006, the poir.t ?t which the ~imum stress of 47000 is obtained in member (a) (see paint (3) on Curve A) then the stress at the same strain for ~ember (C) will be 38000 (see pomt (2) on Curve C) .

Fig. A19.20
., The buckling of sheets 14 taken up In detail in

Volume II.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Al9.13

Fig. Al9.22

sheet to use with each rivet line depends on the stress in the stringer to which it is attached. The failing stress for the stringer will be used, which means that the failing stress of the stringers and corner flanges must be- ~~own betore the effective width can be found. For this example the zee stringers have been selected of such a size as to give an ultimate column failing stress of 38000 psi, and the corner flange members have been ~de of such size and shape as to give a failing stress of 47000 psi. These failing stresses can be computed by theory and methods as given in Volume II. The sizes have purposely been selected to given strengths represented by the test curves of Fig. A19.22. The effective width with each rivet line from equation (1) would be,

For stress analysis procedure using the beam formula, we assume a linear stress variation fram zero to 47000 psi. Since stringers (c) can only take 38000 pSi at the same strain, or in other words, stringer (c) is less effective than members (a) and (b). The effectiveness factor for stringer (c) ~quals its ultimate strength Givided by the ultimate strength of member (a) or 38000/47000 808.
Al9.12 Example PrOblem

For zee stringer 2W 1.90 x .04 (10.5 x , 108/38000)a = 1.25 inches Thus the area of effective skin: 1.25 x .04 = .05 in. a The area of the zee stringer is 0.135 sq. in. which added to the effective skin area gives 0.185 sq. in. which value is entered in Column (2) Of Table I opposite zee stringers numbered 2, 3 and 4 in Fig. A19.23. The same procedure is done for the c orner- members land 5 With the resulting effective areas as given in Table I. On the bottom side which is in tension all material is effeotive. The skin width equal to one-half the distance to adjaCent stringers is assumed to act with each stringer. Taking the area of the angle section as 0.11 and adding skin area equal to 6 x .035 = .21 or a total of 0.32 sq. in. which value is shown in Column 2 of Table I opposite stringers 7, 8 and 9. The areas of the lower corner members plus bottom skin and web skin would come out as recorded in the Table. The next step is to correct tor stringer effectiveness in compression. The failing stress fer the zee stringer was given as 38000 and fer the corner member-s 47000 psi. The effectiveness factor for the zee stringer thus equals 37000/47000 = .808. This factor is recorded in Col~ 3 of Table I. for the corner members 1 and 5 and all the ~ension ~embers the factor is of course unity. The balance of Table I gives the calculation or the effective moment of inertia ZAz~ about the x r.eutral axis.
T~e compreSSive stress intensity at the centroid of the zee stringers ~hus equals,

The wing section in Fig. AI9.23 is subjected to a design bending moment about the x axis of 500,000 in.lb., acting in a direction to put the upper portion in compression. The problem is to determine the margin of safety for this design bending moment. The material is 2424 aluminum alloy.

f-

6" ___ 6" - - ' - - 6"

I L04 I

6" -----l ~

10" - 10--.05
\
....L.

(1)

(10)

L .035
Fig. A19.23

SOLUTION

The beam !o~ula :cr bending stress at any point 15 crb : Mxz/I x. To solve this equation we must have the effective moment of i~ertia of the beam cross-section. The bottom surface being in tension under the given design ~er.ding moment is entirely effective, however the top surface has a variable effectiveness since :he skin, stringers and corner tlange ~embers ha7e different ult~te :ailing stresses. From equation (1) the effeotive width of

crb = Mxz/I x : 500,000 x 5.57/59.30 = 46600 pSi. The true stress equals the effectiveness factor times this stress: .808 x 46600 = 37400 psi. The failing stress equals 38000 hence margin at safety = (38000/37400) - 1 .01 or one percent.

=
~

:: n
"

c:

A 19 14
TABLE Al9.1

ANALYSIS OF

WING STRUCTURES

I 6

Stringer Area Slringe Plus Number Effective SIdn

Stringer jEllectEUect<Ve iveness Area Factor IA)

z'

Az'

z::;:z'-l. Az'

1 2
3 4

0.37 O. 185 0.185 0.lS5 0.370 0.417 0.320 O. 320 0.320 0.417

1.0
0.808 0.80S" 0.808

0.370 0.149 0.149 O. 149 0.370 0.417 0.320 0.320

5
6

7
6

10

1.0 1.0 1.0 1.0 1.0 1.0

Z

4.50 1.664 5.47 11.05 4.60 0.685 5.57 I 4.62 4.60 0.685 5.57 ! 4.62 4.601 0.685 5.57 4.62 4.50 1.664 5.47 11.05 -4.60 -1.920 -3.63 5.50 -4.63 -1.480 ~3. 66 4.28 -4.63 -1.480 -3. 66 4.28

The stress analysis of this wing would show the resulting bend:;'ng and shear stresses for a number 0: spanwt se stations for the cr-t t i cal c.esign load condi t i ons , In this example sorut i cn the bending longit"J.d.1nal stresses 'Nill be deter:r.lned on cross-sections at two stat i ons namely, stations 20 and 47.5, and the shear stresses w111 be deter:nlned for station 20. In this example problem, the lead!.ng edge cell wil 1 be cons leered tne r r ec t Ive as well as any atr-ucture to rear of rear beam, hence s-cruct;.u-e Is a one cell beam with ~ultiple stringers. A second s o.Lutt on including t he leading edge col., to :0r-a a two cell beam w111 also be presented.
A.i'{:'.LYSIS FOR

BEND IN3- Lm;-G I ':'!J:)1:;).L STPLSSES

0.320 -4.63 -1.480 -3.66 4.28 0.417 -4.63 -1.920 -3.63 I 5.50 2.981 -2.897 Ix::;: 59.80

LongI tuc inaj stresses (tension or coropression) are produced by ext.erna; rcrces nor:r:.a 1 to the cross-section and by bending ~oments about x and z axes. The stress equations are:

z' ::;: distance from It x axis to ceatrotd of stringer area

"n

::;::Z Az' /:ZA::;: -2.897/2.981", .97 in.

= PIA - - - - - - - - - - - - - - = longitudinal stress

external = treid of

(2 )

where "n
P A

on corner members (1 ) and (5 ) the campress ive stress 500,000 x 5.47/59.80 = 45000 ps t , With the failing stress being 47000 the margin of safety Is 2 percent.

load acting ":hraugh ceneffective w~ng cross-sectle n

= effective area of cross-section.

Now suppose we would have omitted consideration of the stringer effectiveness factor and omitted column (3 ) of Table J. Carrying through the calculations at Table I, the value of Z would be -0.76" and Ix would be 63.08. The stress intensity on the zee stringers would then be 500000 x 5.36/63.08 42500 ps t , Since the failing stress for stringer is 38000 the aarg m of safety would be (38000/42500) 1 a negative 10.5 percent. The previous result was a plus 1 percent, thus a difference of 11.5 percent in the results.

For any given flange ~ember 'Ni~h ar-ea (a) the load Pn on the member would be.

Pn = ana

----

- - - - - - - - - - - (3 )

T~e stresses due to bending acments are rr-cn Chapter A13. Art. A13.5: -

"t
where sion

= -(K Mz-Kl. Mx)x - (K. T1x -x, M z) z- (4 ) "b = bending stress with plus being ten3

The purpose of this Simple example problem was to emphasize to the student that failure of real aircraft stiffened skin structures occurs u.~der non-linear stress-strain conditions and the elastic t heory must be modified d fairly accurate ea t imatea of the failing strength of a compOSite s'tz-uc tur-e is to be obtained.

K,

= I xz/( Ix = I x/( Ix

Iz

..

K. = Iz/(I x I z
J(,

- I xz ,..) - Jxz 2)
Ixz'" )

Iz

A19.13 Bending and Shear stress Analysis of Tapered _ Multiple Stringer Cantilever Wing. Unsymmetrical Beam Method.

In general cantilever Nings are taperec. in both depth and planform. Fig. A19.24 i uuetrates a typical structural layout of the outer wing panel for a sma.Ll. airplane. The structure consists of a front and rear beam (spar) with
spanwt sa
o' C'':'''''c! ~ t...r .l.ng~~ ~

" 3efcre equations 2, 3 and 4 can je solved. Tayering in cross-sectiorzl material is obtal~ed the effective cross-sect~cn area ~us~ be known by decreasing size of ~embers by cuttir~ off as well as :he ~oments of ine::-tia and ~roduct portions of the spanw1se stringers and corr.er of ~nertla about x and z c9~troidal axes. flanges and decreasing the skin and web thickness.

bet .....e'-' the ... no h ceams ,

o. .

.'

The normal component or ~he axial load in a rLange member equals "b a) where (a) is the area of the flange member. Since the angle between the normal to the beam section and the centr-oiuaj axis of a stringer is generally qUit e small. the c t rrcr-ence between the cosine o~ a small angle and unity is negllg~ble and thus the nonal component can be considered as the axial load in the stringer.
I

ftJ

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A19.15

WING BODY SECTION PLAN VIEW LOOKING INBOARD FROM TIP

I
2.

--r/4 "ii2.
"

/3

/2

"

10

UPPER SURFACE STRINGER AND SKIN ARRANGEMENT

'00 120 '.0 'PO :ITA FIO'-----"F----'T'---'T'-----''t'---'"IL--q.''!.-2.1
<475" 70

f-I

::lJ1
II
)-."4

.03.2,.

,0z.5

-\1 Skin

thickness between beams

UPPER SURFACE STRINGER AND SKIN ARRANGEMENT

'0

..

uo

Fig. A19. 24 Structural Layout of Tapered Cantilever Wing

A19.16

ANALYSIS OF

WING STRUCTURES

-'

1/ -\
~
I

rOt:

'""-'-I

-to-

1.

"---i

T Upper Front Flange

area of angles", . 508 area of skin "'. 090 area of web "'.050

,
.050

.b4

1:

r--- 2

1/4----..-;
Upper Rear Flange area of angles", . 508 area of skin "'.090 area of web ",.032 . 630

.648

Upper Stringer. No. 2 & 6 (t "'.064) area stringer'" . 257 area skin '" . 096

T
1

-:353

.04 upper skin

No.3,4 & 5 (t '" . 050) area stringer '" . 205 area skin '" . 096 . 301

Z'

- - - - - .. ~ .c

~--L_ ~I----------""='T'-'-,r--,j <f"'----,:;-;;:;:;

o

X Centroidal Axis

N ~ 14 C!

-~

'"
8

.032 lower skin

,7/8

Lower Surface Stringer area of bulb angke e 11 area skin '" . 16

I 5"
2
Lower Rear Flange area of angles '" . 508 area of skin "'.120 area of web "'. 064 . 692

-:27

1
Ir
3.75 Fig. A19. 25

Lower Front Flange area of angles '" . 508 area of skin "'. 160 area of web "'. 100

r
2

,
WING SECTION AT STATION 20.

:768

EFFECTIVE AREAS.

Z'

. 032 upper skin

.
o

0\ ~e'"

,\\.o-c.

7 reference axis

~:----=====-~r======-=,,========...::;~===..::"::=======-=..:::j::=i

.-

["".0.. _r

, X'

X Centroidai Axis

14 13

12
Z

10_j

Z'

"'- . 025 lower skin

Fig. A19.26 wrsc SECTION STATION 47. 5

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A19.17

Fig. A19.25 shows the cross section at Station 20 divided into 14 longitudir~l units numbered 1 to 14. Since the external load condition to be used places the top surface in compression, the skin will buckle and thus we use the effective width procedure to obtain the skin ?ortlon to act with each stringer. Fig. Al9.25 shows the effective skin which 1s used with each flange member to give the total area of numbers (1) to (7). The skin on the bottom surface being in tension is all effective and Fig. A19.25 shows the skin area used with each bottom flange member. The next factor to decide is the stringer effectiveness as discussed and explained in the previous example problem. For the structure at Fig. Al9.25 we will assume that the compressive failing stress of the stringers is the same as that for the corner members, thus we will have no correction factor to take care of the situation of flange members having different ultimate strengths. Table A19.2, columns 1 to 11, and the calculations below the table give the calculations for determining the section properties at Station 20, namely A, Ix, I z and I xz Table Al9.3 gives the same ~or wing section at Station 47.5. The areas 1n column (2) are less sir-ce sizes have Changed between Stations 20 and 47.5. Calculation of longitudinal stiess due to M x and My bendi~g moments: The deSign bending and are as follows: Station 20
t'lx
~oments

Substituting in equation (4)

crb

=- [.00098 x
= 3.3

- 285000 - (-.0002125 x

1300000lJ x - [.004378 x 1,300,000 (-.0002125 x - 285000lJ

whence, crb

x - 5639 z-

-----(5)

Column 12 of Table Al9.2 gives the results of this equation for values of x and z in columns 10 and 11. Mu1ti~lying these bending stresses by the stringer areas, the stringer loads are given in column 13. The sum of the loads 1n this column should be zero since total tension must equal total compression on a section in bending. Stresses at station 47.5: Ix

=157.4,

U9.3 )

Iz

= 700,

Ixz

=-35.4
=

(Table

K,
K,

= 35.4/(157.4 x
-35.4/108950

=-.000324

700 - 35.4')

= 700/108950 =-

= .00643

K,
crb

= 157.4/108750 = .001447
[.001447 x -215000 - (-.000324 x 1,000,000)J

x - [(.00643 x 1,000,000

(-.00324 x -2150001J z

whence
crb

will be assumed

= -14.5

x -6360 z

= 1,300,000

in .n.,

t'lz
Station 47.5
"x
!'!z
~

=-

285,000 in.1b.

Column 12 of Table A19.3 gives the results of this equation and column 13, the total stringer loads at station 47.5. The stresses 1n column 12 of each table would be canpared to the failing stress of the flar~e members to obtaIn the margin of safety.
ANALYSIS FOR SHEAR STRESSES IN ',4EBS A.."ID SKIN

= 1,OCO,000 = - 215,000

in.lb. tnv Ib ,

The moments about the y axes are not needed in the bending stress analysis but are ~eeded in the shear analysts which will be made later.
Ka

To solve equation (4) the constants K., and ~ must be known.

For Station 20 from Table A19.2, Ix = 230.3, I z = 1030 and Ixz - - 50, whence Kl. = Ixz/(I x I z - Ixz'i)

shear flow distribution will be calby using the change in axial load in the stringers between stations 20 and 47.5, a method c~~only referred to as the AP ~ethod. For explanation of thiS ~ethcd, refer to Art. A15.l6 of Chapter Al5.
~he c~lated

-50/(230.3 x 10.30

The shear flow in the y direction at a point n of the cell wall equals,
Za d

- 50') K,

= -50/235500

= -.0002125

n oP

= I z/235500 = 1030/235500 = .004378

- - - - - - - - - - -

(6)

K3

= I x/235500 = 230.3/235500 =

.00098

where qo 1s a known value of shear flow at some point 0 and the second term is the change in shear flow between paints 0 and n.

A19.18

ANALYSIS OF

WING STRUCTURES

TABLE A19. 2
SEC1'10N PROPERTIES ABOUT CENTROIDo\L X AND Z AlC4S
.in~

Section at

St~t1on

20

(Co~press10n

On 8

~pper

~urface)

,
.6<18 .353 .300 .300

3
Z'

I Flange Are" , Mo. A

z I 3 i o
6 7 8 8 10
'i.i.

AZ'
,

5 AZ2 19.61 1':::.30 10.28 9.25 7.65 6.33 7.95 48.90

,

I I

8 AX';': 0 11.30 37.60 85.00

10
Z
~

II
~

12

!
"

13

X'
-0.10 5.65 1l.:W 16.85 22.40 28.00 35.72 35.72 27.80 22.60 16.70 10.90

AX'

AX'Z' _0.36 11.78 19. 6 7 1 28.10 34, 00 43.50 80.00 -208.00

Z'

- z! X
,

I
I

5.50 5.90 5.85 5.55

3.56 2.08 1, 76 1.66 1.52 1, 55 2. ;.!4 -5.81

_0.06 1. 99 3.36 5.05 6.72 9.89 23.15 24.70 1.50 6.10 2.67 2.94 L51 .08

6.68 7.08 7.03 6.73 6.23 5.58 4.73 -7.22 _7.32 -7. J2 -7.12 -G.82 _6.32 -5.32

I I

-17.4 -11.65 - 6.10 0.45

.' -

"b

= "'b

_37660 _39940 _39620 -37950 -35115 ! -31450

i
! !
,

_24390 _14070 -11890 _11370 _10510 -11090 _16700 28230 11180 11170 64)0 10390 9600 23020 0.00

! ;
I

.300 5.05 .3511 4.40 .630 3.55 .692 _8.40 .2701-8.50 .270 -8.50 .160 -8.30

150.50 277. 00 826.00 881.00 208.00 138. 00 44.60 32.00 8.45 0 2700

I
I

I
!

::;.10 iO.70 18.42 13.42

10. SO 5.30 0.60

_26560 ! .;0750 41280 41250 40100 I 38400 35570 ! 29950 I

i

i
I

_2.30 _2.30 _1.33 _2.16 _2.02 _5.00 _6.56

19.55 19.55 11, 05 17.30 15.20 32.50 238.0

I

12 13 1<

."

.27

.768

-8.00 -7.50 _6.50

(LID

5.60

!
i

64. 00 52.00 I - 22.20 - 23.50 - 11.35 0.50

i

!
I

,
I

1:. 5 8 4

96.44

_163.8

I ,

6.40 _11.7 _17.40

_6.56 _1.176 5.584

96~44 ;.

11.3
230.3 in.'; 1030 in.'; x 17.3
~50

238 _ 5.~84 ~ 1.176~ 2700 ~ 5.584 ~ 17.3 2

I~z

Ge ne r a l Nates:" 5" l' ig. A19.25 for Section at Station 20. shown. Rete e enc e axes X'X' and 2'2' are assumed Properties are calculated with respect to th"se axes and transferred to the c e nt r ci da I X c nd

"

Z axes.

-163.8

5.584

x-l.17~

"'b *

~.3

X _ 5639 Z

TABLE A19. 3
SEC1'IOK PBOPERTIES ABOUT CIDfTROlnAL X AND Z AXES llinf:" Section a t 1 Station 47.5 (Col:lpression on upper 7
AX'
S~rtace)

!
II

3
Z'

AZ'
I

5
AZ'~

G
X

s
AX 2 0 6.16 26.60 64.75 1::0.00 226. :;0 478.00 :iS5.00 165.00 107.00

9 AX''Z'

1"lnn;.;o Area )(0. A

I 3 5 I , 6

,
1

.;7G .318 .2136 .266 .266 .318 .476

.35'::

5.(0
~.83
~.75

I
I
r

2.66 1. 86 1, 53 1, 44 1, 30 1, 37 1.71 -3.93 _1.74 1. 74 -0.90 _1.62 _1.4 _3.27 _2.80 _0.51"

14.95 10.80 8.80 7.81 6.40 5.88 6.17 27.90 1:::.90 12.90 6.48 11.20 9.20 17.70
1~9.

0.10 4,40 10.00 15.60 ~1. :::0 :':6.70 31.70 31.70 :i!5.50 21. 30 15.10 9.60 4.50 0.10

0.05 1. 40 2.56 4.15 5.64 8.50

0.27 8.15 15.30 22.40 27.60 35.60

,. - . "
Z' X

10

12

13

_ 1r

"b
_3927:; -40790 -4025
_3~075

"'"b .J,
~18700

6.21 6.44 6.36

5.'Z 4.90 4.30

i
I

7 8 8 10
II 1 c~

I _7.10 3.60

6.03 5.51 4.91

, , i
! ,

-15.60 _11, 10 5.50 0.10 5.70 11.20 16.20 16.20 11.00 5.80

_12990 _10710 _10200 9320 _10000

! , ,
1

-38400 I -31360 !

i
!

! ,

.23:; .':;35

i -7.40 -7.40 I
!
I

i
-

I ,
I

15.10 17.51 6.23 5.00 1.89 :':.::6 1. 06 0.06

54.40 -124.50 ';6.10 37.18 13.60 15.GO 6.61 0.33

4.21 -6.';9 _6.79 _6.79 -5.59 -6.29 _5.84 _4.79

I ,

13

"1

.1:.:5 I -7.20 ;:3<: ;_e.. 90 . 23:; ,i _6. 2S , 50~ -::'.40

..
r
Ix I, I"

4.61
-~'.

I
aO/4. 61

. -. -.9 IS!!. 1
1304

71.3/4.61

. .

l8.50 :0:1.70 4.7S 0

1804

71. 3

I - 79.0
b

- 0.40 _55.90
_11, 00 _15.50

I
i

-27030 41050 43090 43180

I
I

_12880 22770 10140 10160 5240 9410 8470 18600 0

, ,

.;1900 I I 40085 I i 36060 I 30725

,
!
I

15.50"

4.61 :t.61 2 4.61

15.5':

. .

157.4 700

General Notes: See Fig. A19.26 for section at Station 4 . 5 Reference axes X'X' .nd 2'Z' are assumed

-14.5 X

6360

"

shown.

4.61 x -.61

15.5

- 35.4

ANALYSIS AND DESIGN OF

~p equals the cbAnge in stringer over a distance d in the y direction. ~~1al

FLIGHT VEHICLE STRUCTURES

A19 19

load

Since the cell in our problem is closed the value qo at any point is unknown. we assume it zero on web 1-14 by imagining that the web 1s cut as shown in Fig. A19.27. Equation (6) thus reduces to,
- - - - - (7)

areas are referred as ~ values (See Fig. A19.27) Column 6 of Table A19.4 records these double areas which were obtained by use of a planimeter Column (7) gives the moment of each shear flow about the e.g. and the total of this col~ gives the moment about the e.g. of the complete shear flow system of Fig. Al9.27 or a val~e of 256060 in.lb. The double areas (m) can be found approximately as follows: The moment of the shear flow q on the web (2-3) about point 0 equals q times twice the area (A 1. + A a ) . In most cases, the area Aa can be neglected.
By Simple

Columns 1 to 5 of Table A19.4 show the solution of equation (7). The shear flow values of qy in column 5 are plotted on the cell wall in Fig. A19.27, remembering that qy qx qz. For rules giving direction of qx and qz refer to Chapter Al4, Art. A14.6.

1--- x,

----1
2

x,

= =

206.'
I - -;.

245.9

288.8
4

", ,

374.6

5, - z

-6--"

414.2 553.2

-rj

'L_~3,-
I'O.T

T~rea A\;.
.~; - - _ . - Fig. Al9.27
TABLE A19.4

11

10. - .Qx_L_

'

-W area
I

qz~

A1.

:1
I

= 1/2 rx, Za - X",Z;s)' The moment of the shear flow q on web (2-3) thus equals q (X;, Z. - X a Z;s)' Since values of X and Z for all flange pOints with reference to section (e.g.) are given in the Table A19.2, it is unnecessary to use the planimeter except for regions of sharp curvature.
!10l1ENr OF EXTERNAL LOADS ABOUT e.g. OF STATION 20

geometry, the area

F ~
,

_~~ -, .

3T
z,

A.

Calculation of Flexural Shear Flow Assuming q 2 'Zero in Web Between Flange Members (1) and. (14). (Shear values are average between stations 20 and. 47. 5)

2,p

e
m
sq. in.

p p Flange at Station at Station No. au 47.5

q.

q' mq

LP -27.5 Z 2f.5

"

q .. ql

As stated before the engineers in the applied loads calculation group supply the shears and moments at various spanwrae stations. ve will assume that these loads are: V z = 12000 lb., Vx +2700 lb., My 390,000 Lm Ib . The location or the reference y axis used by the loads group will be assumed as located at point o in FIg. A19.28 relative to cross-section at Station 20.

=-

1
3

~24390JII

e 7 a s
10 11 11 13

-14070 -11890 -11370 -10510

~11090

14

-16700 28230 11180 11170 5430 10390 9600 23020

-18700 -12990 -10110 -10200 - 9320 -10000 12880 22770 10140 10150 5240 S410 8470 18600

206.8 39.1 42.9 42.5 43.3 39.6 139.0 ~198. 7 _ 37.8 _ 36. 7 ~ 43.2 35.15 - 41. 1 -160.3

206.8 48.4 10000 -328 -121. 2 245.9 43.8 10780 -328 - E12. I 288.8 42.0 12120 -328 - 39.2 331.3 42.6 14100 -328 3.3 374.6 43.6 16340 -328 46.6 414.2 42.0 17500 3213 86.2 SSJ.2 197.0 108900 -328 225.2 354.5 47.4 161300 -328 26.5 315.7 37.2 11800 -328 - 11.3 280.0 40.8 11420 -328 - 48.0 236.8 42.2 10000 -328 - 91. 2 201. 2 38.0 7650 -328 -126.8 L60.1 54.0 8650 -328 -167.9 204.0 0 0 -328 -328

VZ'12r~q:
v x- 27001 11.8 ~ .01--L-

:
-

!~.~.

My=39~33.3
Fig. A19.28

Total mq

Therefore moment ot external loads about

c vg, is,

256060

For equilibrium af all the forces ln the plane of the c~oss-section ~y ~ust equal zero. ror convenience we will select a morr.ent y axis through the c.g. of the c~oss-section. The ~oment of the shear flow q on any sheet element equal q times double the area of the triangle formed by joining the e.g. with lines going to each end of the sheet element. These double

ZMc.g.

=12000

x 33.3 - 2700 x 11.8 - 390000

In.lb.

= 41800

Moments Produced by Inclination 01 Flange Loads With Beam Section.

Since the flange members in general are not normal to the beam secttons, the flange loads

A 19 20

ANALYSIS OF

WING STRUCTURES

have components In the Z and X direct1ons. Columns (4 ) and (7) of tree Table A19.5 g1ve the values of these In ;Jlane c:::mponents. The slopes dx/dy between stations 20 and 47.5 are f'olU1d by scaling from Fig. A19.24. F1g. Al9.29 shows these 1nduced In plane forces as fOW1d in Table A19.5.
TABL A19. 5
[n

Due to external Loads

= 4.1800

in. lb.

T;;'e~ the total unbalancec 256060 + 4l8CO 302795 Lnc Ib ,

:::o:n.e~t

J.C"'~ ...-.A)

For equt Lfbr tum, this ::!lust be baIancec oy a constant shear- flow q,. hence

Plane Moments About Section e. g. Produced by Components o{ Flange Loads Station 20

In

Plane

!1 q, = 2A

--

302796 2 X 461.5

--

328 lb ./in.

1
Flange No.
I

a
dx

s
d, dy
.025 .022 .022 .022 .022 .025 .025 .018 . 017 .015 .015 ,017 ,018 .018

7
Fz

(Note:
8

461.5

= total area of cell )

Px = p dx

dY
-.046
0

dY
0

I M e. g I = Px z
7480 0 0 0 0 0 -1660 -4300 0 0
Q Q Q

"

Me. g, =

The shear stresses q, are listed In Column

(8 ) of Table A19.4.

p.!!!..
dy
610 -310 -262 250 -231 -278 -417 -508 190 -167

(P z Xl -10600 3610 1600

112

2 3

, ,
5
7 B

10 11 12 12 14

-24390 -14070 11890 -11370 -10510 -11090 16700 28230 11180 11170 6430 10390 9600 23020

1120
0 0 0 0 - 350 594
0
Q

The final or resultant shear flow qr at an y pOint therefore equals qr = q

+ q,

0 0 0 0 ,021 ,021 0
0

1180 2970 7770 9350 2000

885 57

0 0 0 -.046

0 0 0 1060

. -177

"

The resulting values are given 1n Column 9 of Table A19.4. r1g A19.30 illustrates tr..e :-esults graphioally.

56-10

173 -416

1130 2020 7250 2224

7160

NOTES: --Column (2); P Ircm Table Al9. 2 Column (5) and (8): Values of Z and X are found iII Colwnns 10, 1 I of Table Al9. 2

~~ II 7.~
0' '/

-_.- _.
/
/

-5-

,--,
/

+,.g.
13 12
11

--'(L//
lli

9 10 '-_.---8

810

106

r"
14

417

Fig. A19.30

310

282

250

I'

231

5 1

278

16
9

+,.g.
lil
1;7

418
Fig. A19.29

~3

1,1
18 9

10 d7

1~3

t
8 594

Final shear flow diagram. For values see Column 9 of Table Al9.4.

1~0

508

Having deter:nined the shear flows, the shear stress on any panel would be qjt. I:l checking the sheet for strength In shear and combined shear and tension, interaction relatianShips are necessary. The strength deSign of sheet panels under combined stresses Is covered in considerable detail i::1 Voli.J:l8 II.
A19,14 Bending and Shear Stress Analysis of 2-Cell Multiple stringer Tapered Cantilever Wing.

In plane forces prOduced by flange axial loads.

The moments of these in plane components about the section e.g. are given in Columns (5 ) and (8) of Table Al9.5. In general, these moments are not large. Total I'!oments of All Forces About Section .:.; . at atat i cn 20: Due to flanges = 7160 - 2224 4936 in.lb. (Ref. Table Al9.5) Due to assumed static shea:- :low 256060 1n.lb. (Ref. Table A19.4)

A two-cell beam IS also quite common In wing structural design. A 'DNa-cell structure In bending and torsion Is statically indeterminate to the second degree since the shear flow at any one pOint In each cell Is unknown, However, due to continuity between cells the angular twist of each cell must be the same, which gives the additional equation necessarJ for solving a ~No-cell beam as compared to the single cell analysis.

Example Prcb Ieta To avoid re'Oet1t1on of similar t

calcu-

SKU

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A19.21

lations as was used in the previous single cell The first 7 columns of this table are the same problem, the bending and shear stresses will be as in Table A19.4, since no stringers have been dete~lined for the same structure as in the added to cell (1), and the shear q is assuoed previous example except that the leading edge zero In cell (1). cell is considered effec:lve, th~s ~king a 2-cell structure. Since there are no spanwise ~o ~e the twist of each cell the same and st~ingers in the leading edge, very little skin also to make the summation of all torsional on the compreSSive side will be effective. On forces zero w111 require two unknown constant the tension side, the leading edge s~in would shear flOWS, q~ in cell (1) and qa in cell (2). be effective in resisting bending axial loads Thus VNQ equations will be written, namely: and thus the moment of inertia would be slightly different from that found in example problem 1. (8 ) Since this pr0blem is only for the pur,ose of illustrating the use of the equations, the (9 ) Zl1c .g , = 0 leading edge s~in Nill be neglected in computing the bending flexural stresses. ~lth this The VNist Q per unit length of a cell assumption, the bending stresses and flange equals loads at stations 20 and 47.5 are the same as _ 1 for the previous prob Lem. (See values in ccj.cnn Q - 2AG 1:qL/t 12 and 13 of Tables A19.2 and A19.3.)

uo:

Shear Flow Calculations: To compute the static shear flow, each cell is assuned cut at one point as shown in Fig. A1S.31, and thus the shear flow is zero at points (a) and (b).
Fig. A19.31

The modulus of rigidity G w1l1 be assumed constant and thus will be omitted.
C01UITL~S

Consider cell (1): 10 ~~d 11) Area of cell (1)

(Refer to Table A13.6, sq. in.

=23.5

~r
~83.5
14
1
!

Area

:z

Cell (2) 461. 5 sq. in.

1.3

12

'.'

10

[
TABLE A19. 6

Q,

_1

2A,

q Lit. Z q, Lit. 1: q.

L/~
whence

Q,
Q1

x1 83 5
q~ -

+ 1278

q~

- 230

qJ,

=7.65

1.378 q.- - - - - - - - (a)

For cell (2): Table A19.6 (Column 5) giyes the va.Iue of the static shear r icw under- these assumptions.
Ar-ea

of cell (2) :: 461.3 sq. in.

Flant:e i

<0.
1 2 3

P P i at St;... i :J.t_Stal_ '; p. , .1.0> ! cO , 4 5

, ,

I .l I."
,

SHEAR FLOW VULTIPLE FUNCJi TAPERED BEAM 2 CELl. (Aver:li:C Sh.. nr Flo.. Between Stations 20 and 47.5)

(i;. )

10
LI<

11
q (Lit)

12
Ce~~
(2)

13

14

),,~P_ 1-7.;,

sl

"'

I ~(~)

~(~t c'(gl
1<4 137 137 137 137 137 382

'1 0 0 0 0 0 0 0 0 0 0 0 0 0 _62.8 -62.8

'2 _317 -317 31 7 _317 _317 -317 _311 -317 _317 -311 _311 _311 -317 -311

'r ,
Cl-KIl-KI2
I
1

1- 24390
i -11890
I -14070 .11310 _10510 -11090 -16700 ::::8:<30 11180 11170 6430 10390 9600
~3020

,
5 6

- eaac
_10000 _128S0 22170 10140 10160 5240
9410
,

_18100 _12990 _10710 _10200

206.8 39.1 4:.\.9 42.5

~3.3

s
10 11 12

13
14

i
:

8470 18600

1 Lead
cd!>:e
,

1 ,

39.6 139.0 -t9a.7 _ 31.8 _ 36,7 _.43.2 - 35. 6 ... 41.1 160.3

206. 48.4 245.9 43.S 288.8 4:;:.8 331.3 42.6 374.6 ~3.5 4.14.2 42.0 553.2 197.0 35~. 51 ~7.~ 316.71 37.2 2S0.01 40.8 236.81 42.2 201. 2 J8.0 160.11 .54.0 o 204
!

10000 5.15 10780 5.50 12120 5.50 14100 5 .50 5.50 1634lJ 5.50 17500 108900 12.25 16S00 I '-' 5.0 11800 11420 5 10000 5.' 5.3 '""0 8650 '-' 11.15 0 33.5 256060

.04 .04 .04 .04 .04 .04 .032 .032 .032 .032 .032 ! .032 ! .032 , .051 ! .032

I r vz
I
0 0

156 181

1S1
2" 166 172 230

29800 33700 39600 45400 .51400 56700 211500 61000 49400 50650 42900 33400 21500 0

11.1 28.2 - H.3 57.6 97.2 236.2 31 .5 0

-1l0.2

I
I : I

I :

80.2 _115.8 _1.56.9 _2.54.2 -62.'

1

J7

eee

,
,

i !

J7l

i 1048 i
,
1218
~

I ,
133050

S~

I
~

I
t

2469

MOTES:

length of web web thickness.

she~t

betweea flange s.mbers.

Cal.

.. (Cal. 2 + Col. 3)/27.5

A19.22

ANALYSIS OF WING

STRUCTURES

2 X

~1.5 ~733050
~ust

- 230

q~
o, O~,

A19.15 Bending Strength of Thick Skin - Wing Section

+

2469
It)

G. = 794 equatlr~

.2495 q, + 2.678

For continuity O~ (al and (b): 7.899

q~

equal

hence
(c)

- 4.056 qa - 794 = 0

For equilibrium the summaticn of all moments in the ,lane of the cross-section abo~t the section (e.g.) ~ust be equal to zero, or
Z

ric.g. = 0

The ~~ent of the external loads about the section e.g. is the same as in previOus problem.
!1 external forces

?igs. i and k o! Fig. AIG.3 ill~s~rate srzpes for air~oils c~ supersonic aircraft. Such e u-ro i ;s nave r eiat ive iv low thlc~~ess ratios and since supersonic ~l:itary aircraft have comparatively h:gh wl~g loaclngs, it is necessary to go to thick skin in order to resist the wing bending ~cments efficiently. The ultinate compressive stress of s~ch structures can be made rather uni rcrn and o c curr-t ng at stresses considerably above t~e yield ~oint of the material. Since struc:ures wust ca~ry the design loads without failure, it is necessary to be able to calCUlate the ultimate bending resistance of such 3. wtng section i t t::e ~argin of safety is :0 be given ~or various load conditions.
approxi~ate

= 41800 in.lb.

The induced moment due to the ~n plane components of the flange axial loads is likewise the same as in previous problem (see Table
A19.5).

The question of the ultimate bending resistance of beam sectlor~ that ~al1 at stresses beyond the elastic stress Yar~e is treated in Article A13.l0 and example problem 7 of Chapter A13 and should be studied again betore ~ro ceeding with the following example ,~oblem.
A19.16 Example Problem
~a illustrate the ;roced~re of Art. A13.l0, a portiO~ o~ a thick =kin Ni~g section as 1l1L.:.strated In F'fg , A19.34 'N111 be cons Lder'ed ,

M due to flange loads

=4936

in.lb.

The torsional moment due to ~he static shear flow rr ce Column (7) or Table Al9.6 equals 256060 in. lb. The torque due to the unknown constant shear flows at q1 and q~ 1s equal to twice the enclosed area ot each cell times the shear flow in that cell, whence
l'l(

6"

6'"
T
1 '" 22

11.-3/ S'"
+

due to q::,. and Gil) = 2x83.5 q~ -+-2 x 461.6 qll = 167 q:l, -+- 923 qll
q~ +

l~ 3" ~l2

.t..
L

3 ~ 1","" 8 2

3 -l13/8"
---x

Therefore ZMc.g. = 167

302796

=0

923 qll . .
- - (d)

3 l-iS

3 16

I--fa,..-,

Solving equations (c) and (d), we obtain, q1 = -62.8 lb./in., qll

"

C'--J

'=='

= -317

Ib./in.

Fig. A19.34

These values are listed in columns (12) and (13) ot Table A19.6. The final or resultant shear flow qr on any sheet panel equals the sum at q .+ q~ + qll' The results are shown in cojunn (14) of Table A19.6. Fig. Al9.32 shows the potted shear flow pattern. Comparing this figure With Fig. A19.30 shows the et!ect of adding the leading edge cell to 'tne Single cell or the previous problem.

For s~pliclty the section r~s been symmetrical about the x-x axis. The mater~al is aluminum alloy. In this problem the ~aterial stressstrain curves will be assumed the s~e ih bot:: te!'lsion and compression. The ?rOblem 1s to determine the ~gin of safety for ~~is beam section when subjected to a j eaf gn bending IDoment M x = 1,850,000 in.lb. SOLUTION: Since it is desirable to use the beam ab = ~xZ/lx, it is necessary to obtain a mod::ied beam section to correct ~or the !'lonlinear stress-strain relationship since the give structure will tail ~nder stresses in the inelastic zone. The =axilnUffi compressive stress at sur~ace of beam will be ass~ed at 500CO pSi. ~ls value could be calculated from a conslderto~ula

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A19.23

ation of crippling and column strength at the sti:fened Skin, a subject treated later. Curve (A) of Fig. A19.35a is a portion of the compressive stress-strain diagram of the aluminum alloy material from zero to 50,000 PSi. Due to symmetry about the x-x axiS, we need only to consider one halt or the beam section. We divide the upper half of the beam section into a horizontal strips, each 3/8 inch thick. Each beam portion along these horizontal strips can be placed together to form the areas labeled (1) to (8) in Fig. b or Fig. A19.35. Since plane sections remain plane after bending in both slastic and inelastic stress zones, Fig. c shows the beam section strain picture.
Fig. A19.35

The values in colljffiTI (3) of Table A1S.7 represents the true compressive stress at the ~idpoint ot a strip area when the beam 1s resisting its maximum or failing bending'nament. The values in column (4) represent the compreSSive stress at the midpoint ot the strip areas it the bending stress is linear and varying trom zero at neutral axis to 50000 psi at edge of beam section (Curve B of Fig. A19.35a). To illustrate, consider strip area number (2) in Fig. b ot A19.35. ~oject a horizontal dashed line from midpOint of this strip until it intersects curves A and B at paints (a) and (b) respectively. From these intersection paints project downward to read values of 48000 and 40600 pSi respectively. In using the linear beam formula, the stress intensity on strip (2) would be 40600 but actually it is 48000. The ratio between the two 1s given the symbol K. Thus to modify the linear stress to make it equal to the nonlinear stress we increase the true strip areas by the factor K, giving the results ot
co Lumn (6).

The modified moment at inertia (column 8) equals Ix = 115.52. The design bending moment was 1,850,000 in.lb. Consider point at Table A19.1 Shows the calculations for obtaining the ~odifled moment of i~ertia of the cross-section to use with the linear beam formula.
1 Portion 1
~idpoint

of strip (1)

2.8125 Inches

0b

=Mx

Z/I x

= (1,850,000

x 2.8125)/115.52

= 45100

psi.

,
Area
A

TABLE Al9. 7 3
4 5
K=

7 Arm

This stress is based on the modified strip areas. The true stress Ob(t) on strip 1 thus equals Kab = 1.049 x 45100 47400. The allowable stress at failure equals 49200 from column (3) or Table A19.7. Hence margin ot safety (49200/47400) - 1 .04 or 4 percent.

Oc

Oc'

ac/ac' Ae-KA

AeZ'

a
3
4 5 6

4.500 49200 46900 1.049 2.437 48000 40600 1. 182 0.211 46000 34400 1. 340 0.211 43200 28100 0.211 39200 21840
1. 539 1. 790

7
8

0.211 33200 15620 2.121 0.211 20500 9360 2.190 0.211 7000 3120 2.240

4.71 2.8125 37.30 2.88 2.4375 17.14 0.282 2.0625 1. 20 0.324 1. 6875 0.92 0.378 1. 3125 0.65 0.449 10.9375 O. 38 0.461 10. 5625 O. 15 0.473 0.1875 0.02 Z 57.76 Ix

The margin of safety for points other on the beam section will likewise be 4 ~ercent. For example at mtcpo trrt of strip (3), Z ::: 2.0625. K:::l.34 whence O-:t ::: [(1,850,000 x 2.0625)/115.52J1.34= 44400. whence
~rg1n

ot safety

(46000/44400 )

Col. (2) Col. (3) Col. (4) Col. (5) Col. (6) Col. (7) Col. (8)

= Area of Strip

= 115. 52

--'

- 1 = .04

Stress at midpOint of strip as read from Curve (A) Stress at midpctnt of strip as read from Curve (B) Nonlinear Correction Factor K = crcia c' Modi1ied Area = Ae = KA Arm rrom Neutral Axis to Midpoint of Strip Modified section moment of inertia

The moment of inertia Without modifying the strip areas would come out to be Ix = 104.42 hence the stress at midpoint at strip (1) would calculate to be crt ::: (1,850,000 x 2.8125)/104.42 = 49900. The allowable str~ss ror linear stress variation would ~e 46900 from column (4) of Table. Hence margin of safety would be

A19.24

ANALYSIS OF

WING STR lie T URES

A19.l8 Shear Lag Influences
I~ the beam ~heory, the aSSk~~tiQ~ ~s ~ade plane sections ~errzi~ plane after be~cin~. In a beam ~nvolving sheet and stringer ~anels, this asswm~ticn ~eans that the sheet panels have inf:n1te shearing rig:d~ty, whlct o~ course is not tr~e as shearing stresses produce shear~ng strains. The effect of sheet ~anel shear strair.s is to cause Some str1~gers to ~esist less axial load than those calculated oy Jeam theory. T~is cecreased eifect:veness of stringers is referred to as "sneer ::'2.g~ effect, since some stringers te~d to lag bac~ f~8m the pos~tion they would take if ~lane sec!ions r8~in plane after bencing.

(46900/49900) - 1 = -.06. Tte elas:ic theory thus gives a ~a~gin of safety 10 percen~ less than the strength given when t~ue st~ess-strair. or non-linear relationship is used.

tha~

If the s&~e comparison was ~de for bendl~g about Z axis of this same beam section the difference would be consi1erably more than 10 percent as more beam area is acting in the region of greater cescr-epancy between cur-ves A and B.
Al9.l7 Application to Practical Wing Section

A practical wing sec~lon involves these facts: - (1) The section is ur~ynmetrical; (2) external load planes change their direction under different flight conditions; (3) the material stress-strain curves are different in tension and compression in ~he inelastic range. Since the stress analyst 3ust deter.nine critical margins of safety for ~y conditions, it would be convenient to have an interaction curve involving Mx and Mz bending moments which would cause failure of the wlr~ section. This interaction curle could be obtained as follows:(1) (2) Choose a neutral axis direction and its location. Assuming that plane sections remain plane, and taking the ::lax11ll.um sera in as tnat causir~ tailure ot the compressive flar.ge, use the stress-strain cu~re to dete~ine the longitudinal stress and then the internal load on each element of the c~oss section. A check on the locat:on of the assumed neutral axis is that the total compression on cross-section must equal total tenslon. Since the location was assumed or guessed, the neutral axls must be moved parallel to itselt to another location and repeated until the above check ls obtained. Find the internal resisting :noment about the neutral axis and an axis normal to the neutral axes. Resolve these moments into moments about x and z axe'S or Mx and M z These resulting values of M are x and bendlr~ moments which acting together will cause failure of the wing in bending.

In general, the shear lag effect ~n sheetstringer structures is not appreciable exce,t for the following sltuat~ons: -

(1)
(2)

Cutouts which cause one or more stringers to be discontinued. Large abrupt changes in applications . Abrupt changes
ex~ernal

load

(3)

~n str~nger

areas.

(3)

In Chapters A7 and AS, strains due to Shearing stresses were considered ~n solving for d~stortions and stresses in str~ctures involVing sheet-st~1nger construction. Even ~n these so-called rigorous methods, s~plify1ng assw~ptions nust be made as for example, shear stress is constant over a pa~icular sheet panel and estimates of the modulus of rigidity for sheet panels under a varying state of buckling must be made. The number of stringers and sheet panels in a normal wing is la~e, thus the structure 1s statically indeteroinate to ~ny degrees and solutions necess~tate the use of high speed computors. Before such ~~lyses can be made, the size and thickness of each structural part must be kno.vn, thus rapid approx1wate ~ethods of stress analysis are desirable in obtaining accurate prellminar; sizes to use in the ~ore r~gorous elastic analysis. To illustrate the shear lag problem in its Simplest state, consider the three stringersheet panel unit or Fig. A19.36. The three stringers are supported rigidly at B &id equal loads P are applied to the two edge str1ngers labeled (1) at point (A). The center strir~er (2) has zero ax1al load at (A), but as end B is approached, the sheet panels transfer some of the load P to the center stringer by shear st~esses in the sheet. At the support paints B the transfer of load trom side stringers to center stringer is such as to make the load in all three str1ngers approx~ately equal or equal to 2P/3.

Mz

(4)

Repeat ste~s 1, 2 and 3 for several other directions for a ~eutral axls which results will glve additional combinat10ns of and Mz moments to cause wing failure. Thus an interaction curve involving values of Mi and Mz which cause failure of wing in bending is obtained and thus the margin of safety tor ~~y design condition ls readily

MX

obta Inab Le ,

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

( 1) A

A19.25

2P/3 Sheet
2P!3
B B
'(2)

J
x*b

Sheet

2P/3 . . -

(1)

-L- p
A

equals 3b. At beam section 2-2 stringers (5) and (7) have become effective since they intersect triangle at points (a) on section 2-2. At section 3-3 point c, stringer (6) becomes fully effective. To handle shear lag etfect in a practical wing problem another column Would be inserted in Table A19.l between columns (3) and (4) to take care of the shear lag effect. The shear lag effectiveness factor which we will call R would equal the effectiveness obtained trom a triangle such as illustrated in Fig. A19.3S. For example, the shear lag factor R at beam section 1-1 in Fig. A19.38 would be zero for stringers (5), (6) and (7) and one for all other stringers. At beam section 2-2 stringers (5) and (7) have a factor R = 1.0 since they are tully effective at paints (a). Stringer (6) is only 50 percent effective since section 2-2 is halfway from section 1-1 to paint (c), thus R = 0.5 for stringer (6). At beam section 3-3, stringer (6) becomes fully effective and thus R = 1.0 for all stringers. The tinal modified stringer area (A) in column (4) of Table A19.l would then equal the true stringer area plus its effective skin times the factors KR. The procedure tram this point would be the same as discussed before. Thus shear lag approximations can be handled quite easily by ~odifylng the strir.ger areas. USing these modified stringer areas. the true total loads in the stringers are obtained. The true stresses equal these loads divided by the true stringer area, not the mOdified area.
A19. 20 Approximate Shear Lag Effect in Beam Regions where Large Concentrated Loads are Applied.

Fig. A19.36

The theoretical load in center stringer can be calculated by methods of Chapter A7 and AS, and the results would ~ive the solid curve of Fig. A19.37. To simplify the solution, it is common practice to ass~~e the load distribution in the center stringe~ to varJ according to the dashed curve in Fig. A19.37 which indicates that in a distance 3b, the load 2P 1s equalized between the three stringers.

2P/31-- - - -- - -"
Load ill Stringer
(2)

0';;B

~
I'
'<,

' "

--

---.

Theory Approximate

--'-

----3b

~A

Fig. Al9.37 A19.l9 Application of Shear Lag Approximation to Wing with Cut-Out.

?ig. A19.38 shows the top ot a multiple stringer wing wh1ch includes a cut-out in the surface. The str1~gers (5), (6) and (7) must be discont1nued through the cut-out region.
3

,--'
-,
-,

'r--! 'H, h
I

:: :~

I
I

I
I
a1.-/

/Cut-out

I I

'

1.-/
c ----...

~I
b l

<, <, .-/

~

Wing and fuselage structures are otten required to resist large concentrated torces as for example power plant reactions, landing gear reactions, etc. To illustrate, Fig. A19.39 represents a landing condition, with vertical load. The wing is a box beam with 7 stringers
Top Surface
A
<,

"I
I

<,

:--,

1-3b_ I

f---3b

-.I
,
'-'
b -;1

Effectiveness Tri-

an 1e

1
3

I
Fig. Al9.38

II

,
~
~

It 1s assumed that the ef:ectiveness of these 3 inter~~pted stringers 1s given by the triangles in the figure. At beam section 1-1 these strir~ers have zero end load. The stringer load is then assumed to increase linearly to full effectiveness when it intersects the 51des at this triangle whose height

~ Dead wt. inertia forces

IV~

Bending Mom.

Fig. A19.39

A19.26

ANALYSIS OF WING STRUCTURES I

T

and flange nemoera . Fig. (a) shows the bending moment diagram due to ~he landing gear react~cn alone. The internal resistance to :his ~ending moment cannot be uni:orm on a beam section adjacent to section A-A because of the shear strain in the sheet panels or Nhat is called shear lag effect. To epproximate this stringer effectiveness, a shear lag trian~le of length 3b 1s assumed, and the same procedure as discussed in the ~revlous article on cut-outs is used in finding the longitudinal stresses. It should be understood that the bending ~oments due to the distributed forces on the wing such as air loads and dead weight inertia loads are not included in the shear lag considerations, only the forces tr~t are applied at concentrated points on the structure and must be distributed into the beam. A side load on the gear or power plant would produce a localized couple plUS an axial force besides a shear force as in Fig. AI9.39. The resistance to this couple and axial force would likewise be based on the effectiveness triangle in Fig. AI9.39.
A19.21 Approximation of Shear Lag Effect for Sudden Change in Stringer- Area

6"

cr-----......
C B
20"~

Tn '; l !
ISO"

I I

I I
I I I

-I 9"

!- 21"
Cell

60"

-r/

~-.a:.t,,-II_..EO..
Fig. A19.41

-=====-

Stringers of one size are otten spliced to stringers of smaller size thus creating ~ discontinuity because of the SUdden change in stringer area. Fig. A19.40 shows the stringer arrangement in a typical Sheet-stringer wing. Stringer B 1s spliced at paint indicated. The stringer area A. is decreased SUddenly by s~licing into a stringer with less area A1
Top Surface of Wing

intensity acting upward in the z direct10n and 0.25 lb./in.a average Intiens i ty acting rearward in the x direction. The center of pressure for z forces is on the 25 percent ot chord line measured f~om the leading edge edge and at mid-height of spar AB tor the x air forces. Assume the 3 stringers A, 8, C develop the en~ire resistance to external bending moments. Find axial ~oads 1n stringers A, 8, C and the shear :low in the 3 sheet panels of cell (1) at wing stations located 50~, 100~ and 150" :ram wing tip. Consider structure to rear of cell (1) as only carrying a1rloads tcnvard to cell (1) and not resisting'wing torsion or bendi:Ig.
Fig. Al9.42

Single
Pin FIttings

-3b

3b"'"

Fig. A19.40

To approximate the shear lag effect, assume the area of stri~er 8 at splice paint to be the average area of the two sides or (A 1 + A.)/2. This average area is then assumed to taper to As and A1 at a distance 3b tram the splice point. "The shear lag effectiveness factor R will therefore be greater than 1.0 on the side toward the smaller stringer A, and less than one on the side toward the stringer with tfie greater area As, since the average area was used tor the splice point.
A19.22 Problems

'c
I
I

S I

(2)

(1)

Fig. A19.41 shows a cantilever, 3 stringer, single cell wing. It is subjected to a distributed airload or 2 lb./in.~ average

Fig. A19.42 shows a monoplane wing w1th one external brace strut. ~he wing is fastened to !uselage by Single pins at points (a) and (b). ':'he r i t t tng at (b) 1s designed to take of: drag reaction. The alrloads are Wz = 40 Ib./in. of wing span, with center of pressure at 30 percent of chord tram leading edge and acting upward,

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A19.27

and Wx = 5 lb./in., acting to rear and located at mid-depth of wir~. Find reactions at points (a), (b) and (d). Find axial loads on front and rear spars. Find primary bending moments on front spar. Find shear flow on webs and walls. Neilect structure forward of front spar and rearward of rear spar. (3) Fig. A19.43 shows a portion of a single cell - multiple stringer cantilever wing. T~e external air loads are: (4)

Table A gives the str~nger areas at stations 0 and 150. Assume stringers have linear variation in area be~Neen these two stations. Use 3Dt as effective skin with compression stringers. Find axial loads in stringers at stations 150 and 130 and detennlne shear flow system at station 150. Same as problem (3) but add an internal web of .04 thickness connecting stringers (3) and (8). Same as problem. (4) but add a leading edge cell with radius equal to one-half the front spar depth. Take skin thickness as .04 inches.

Wz = 100 lb./in. acting upward and whose

center at pressure is along a y axis coinciding with stringer (3). W x = 6 lb./in. aC4ing to rear and located at mid-depth of wing. (5)

'D'
J...
6

TABLE A

5 ..i

f-3 3 +-" . 3j -12

12

1
4

<1--

-+4

i
I
I I

I I
,

I
I
I

I
I
I

I
I

1 ~-.050
12"

,.032

T
.032~~

10"

~5""

52

C
5

025

+5j

~-J
150"

STA.150

STA.O

Fig. A19.43

...

. -. . ' .

. . .

,-

.., ,',.

.'

-'.0"-

.-.

" ' , . . ..

'~"'.::1

~.t.l.

,'t,,,

A19.28

ANALYSIS OF WING STRUCTURES

North American Aviation FJ3 "Fury Jet". View Shows Bottom Wing Skin. Note Integral Construction of Skin and Stringers.

North American Aviation F-1OO "Super sabre" Wing. View Shows End of Outer Wing Panel. Note Thick Skin.

. ...

Douglas

DC-a Wing.

View ShOWing Lower Surface of Outer Wing Panel Between Center and Rear Spar.

CHAPTER A20

INTRODUCTION TO FUSElAGE STRESS ANALYSIS

A20.1 General.

In general the pur-pose of an air-

plane 1s to trans?ort a commercial payload or a military useful load. The commercial payload of a ~odern airliner may be 100 or ~ore passengers and their baggage. These passengers must be ~ransported safely and comfortably. For example, an airliner flies at high altitudes where temperatures may be far below zero and where the air denSity 1s such as not to sustain human life. These facts mean that the body which car~les the passengers must be heated, ventilated and pressurized to prOVide the necessary safety. Air travel ~ust be acceptaJle to the passengers," thus the airplane body must shield the passengers from excessive noise and vibration. and further.nore efficient, restful and attractive furnishings must be provided to ~ke travel enroute comfortable and enjoyable. The portion of the airplane which houses the passengers on payload is referred to as the fuselage. Fuselages vary greatly in size and confi~uration. For example, the ~uselage of a supersonic military airplane ~y house only one passenger, the pilot, the remainder of the fuselage interior Spdce being used to house the ,ower plant, to prOVide retracting space for landing gear, and to house the many mechanical and electronic installations which are ~ecessary to :ly the airplane and carry out the various operations for whlcn the airplane was designed to acccmplish. Many groups of engineers with various backgrounds of training and experience are therefore concerned with the deSign of the fuselage. The structures engineer plays a very L~portant part because he is responsible :or the strength, rigidity and light weight of the fuselage structure.
A20. 2 Loads. Basic Structure.

to Chapter A5, where ex&~ple calculations of fuselage shears and moments are presentee. The basic fuselage structure 1s essentially a single cell thin walled tube with many transverse fr&~es or rin~s and longitudinal st~ingers to prOVide a combined structure which can absorb and transmit the many concentrated and distributed applied forces safely and efficiently. The fuselage is essentially a beam structure subjected to bending, torsional and axial forces. The ideal fuselage structure would be one rree of cut-outs or dt soont.tnut ties, however a practical fuselage must have ~y cutouts. Fig. (a) shows the basic interior fuselage structure of a small airplane with skin removed. It consists of transverse frames and longitudinal stringers. Photographs 1, 2 and 3 illustrate fuselage construction of late model large aircraft.

Fig. (a)

The wing, being the lifting bOdy is subjected to large distributed surface air forces, whereas the fuselage 1s subjected to relatively small surface air forces. The fuSelage is subjected to large concentrated forces such as the wing reactions, landing gear reactions, empennage reactions, etc. In addition the fuselage houses ~y items of various sizes and weights which therefore SUbject the fusela~e to large mer-t ra forces. In addition, because of high altitUde flight, the fuselage ~ust withstand inter~al ~ressures. and to ~andle these i~ternal pressures efficiently requires a circular crosssection or a combination of circular elements. The student should refer to Chapter A4 for further ciscussion of loads on aircraft and also
A20.1

PHaro. NO.1

Fuselage Construction of Fairchild F-27 Transport

A20.2

FUSELAGE

STRESS

ANALYSIS

PHOTO. NO.2 View LOOking Inside of Rear Portion of Fuse lage of Beechcraft TWin- Bonanza Airplane.

PHOTO. NO.3 Fuselage Construction of Boeing 707 Jet Airliner.

(FOR GENERAL DETAILS OF DOUGLAS DC-a FUSELAGE CONSTRUCTION sEE PAGE A15. 32)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A20.3

A20.3 Stress AnalySiS Methods. Effective CrossSection.

be~

(1)

It is common practice to use the Si~plified theory 1~ calculating the stresses in the skin and stringers of a ~uselage structure. If the fuselage is pressurized, the stresses in the skin due to this lnterr~l pressure ~ust be added to the stresses which resist the flight loads. In wings the skin in the ~iddle region of the (2) The remainder of the curved sheet between airfoil 1s relatively flat and thus the skin is stringers, namely, b-(w l + w.) carries a usually considered as made up of flat sheet maxtmum compressive stress Ocr = .3 E t/r. panels. In fuselages, however, the skin is This value for ocr is conservative. E is curved and curved sheet panels have a higher the modulUS of elasticity of the skin critical compressive buckling stress tr~n flat material, t the skin thiCkness, and r the panels of the same size and thickness. In radius of curvature of the skin. These small airplanes, the radius of curvature of the curved sheet elements are shown by the fuselage skin is relatively small and thus the hatched skin lengths in Fig. A20.l. additional buckling strength due to this curvature may be appreciable. A simple procedure ot Since the thin curved skin between the approximately including the effect of sheet stringers normally buckles under a compreSSive curvature will now be explained. stress far below the buckling strength of the stringers, the curved sheet is treated as an Fig. A20.l illustrates a distributed element with varying effective thickness which stringer type of fuselage section. Assume that dependS on the ratio of the curved sheet buckexternal loads are applied which produce bending ling stress ocr to the bending stress crt existof the be~ about the Y axis with compression on ing at that point for bending of the fuselage the ~pper portion of the cell. section. Hence the effective sheet thickness for the c~led sheet panels can be written,

A small width of sheet w l on each side of the attachment line of skin to stringer ~s considered as carrying the same compressive stress as the stringer, as was discussed in Chapter A19. These effective sheet widths ""l are shown as the blackened elements adjacent to the stringers on the compressive side in Fig. A20.1.

',"

- - - - (1)
or an effective area can be written
----(2)

y
_ - J L N. A.

where b l 15 the width at curved sheet between the effective sheet widths w1 , w., etc. (See F1g. A20.1). To illustrate this approach in obtaining the effective cross-section of a fuselage section, an ex~ple problam will be presented. The example problem will be broadened to some extent for the purpose of intrOducing the student to deSign procedure.
A20.4 Example Problem.

Fig. A20.1

UP to the pOint of buckling of the curved sheet between the skin stringers, all the material in the beam section can be considered fully affective and the bending stresses can be computed by the general flexure formula CO ~ Mz/I y , where I y is the centroldal ~oment of inertia of the entire section. Wnen a bending compression stress is reached which causes the c~red sheet between stringers to buckle, are-arrangement takes place in the stress distr~bution on the section as a whole. Theory as well as experimental results indicate that the ultlmat) compressive strength of a curved sheet with edge stringers can be approximated by the following c~o assumptiOns.

Let it be required to determine the stringer arrangement for the approximate elliptical Shaped :uselage section shown in Fig. A20.2. The following data will be assumed: Design bending moment about y axis = 160,000 in. lb. (pr-oduc tng compression on upper portion). Zee stringers, one inch deep and with an area equal to 0.12 sq. in. shall be used. The ultimate compressive strength of the zee stri~ger plus its effective skin and a ler~th equal to fuselage frw~e spacing is assumed to be 32000 psi. T~e skin thickness is .032 and all material is (2024) al~inum alloy with E = lO,3CO ,OCO psi. The ~uselage stringers are to be symmetrical about section center lines.

A20.4

FUSELAGE

STRESS

ANALYSIS
'II

= 1.7x .032'/10,300,000/32000 = .975

~:1.

...;'

x'"
4

~.75
3 2 1

Which eqlials a
t~icknesses.

-+,0
" 9

;Z " "t:
8 11

7
Height", 50" Width '" 30"

wld~h of .975/.032 = 30.5 sheet Since the bending stress decreases to zero as the neutral axis is lpproached, ~~d since the curved sheet Jetween the Z stringers can carry loads up to its b~~kling stre~g~h, a preliminary value of e r r ec t ive width 'Ii 4:0t will be ass~ed acting with each stringer. Thus total area of st~inger plus effective skin equals 0.12 ... 40 x .032:01 = 0.16 sq. 1::1. The number of st~ingers required is theref)re 2.00/ 0.16 = 12.

e-

y' -'-......-a--lo-=~...J.---+------+-

y'

;::=F=-::-::...:=--=::j:=N~.~A.Trial

2 N. A. Trial 1

Z", .135"

15 16
1

Fig. A20.2

18

19 20
21 2223

Symmetrical about the Z axis Location of stringers below y' - Y' same as above Y' - Y'.

Fig. A20.2 shows how the stringers were placed to give 12 stringers on the top hal:. Since the skin on the lower half is in tension and therefore fully effective, the neutral ~xis wtll fall below the center lir.e and thus the two stringers on the center line will be considered as part of the requi~ed 12 stringers. A t~selage cross-section has ~OW been obtai~ed. The desired :inal result is that the naximum compressive stress will be near but nat Over 32000 pS1. The procedure f~om this ;oint is still a trial and error ~rocess since the effective sheet on the compressive siGe depends on the ma~itude of the compressive bending stress which in turn 1s influenced by the ~ount or effective sheet ~~d the buck11ng load carried by the curved sheet.
Usin~

the preliminary stringer

~rrangement

Solution: The first thing to do Is to determine approximately how many Z stringers will be required so that a section can be obtained to work with. Since the internal resisting ~om ent must equal the external bending moment, one can guess at the internal resisting couple in terms ot total compressive flange stress and an effective internal couple arm. For elliptical and circular sections with distributed flange ~terial, the approxi~te effective resistir~ ar.n of the internal couple can be taken as 0.75 times the height ~, and the average tension or compressive stress as 2/3 the maximum stress. ~hus equating the external bending ~ament to the internal resisting moment an approximate total area A c for the compreSSive side of the fuselage section can be obtained.
=:

ot Fig. A20.2, Tables A20.1 and A20.2 give the

calculation of the e~fect~ve ~ornent of inertia of the section about the horizontal neutral axiS, Table A20.1 deals with the stringers and the effective sheet elements and Table A20.2 deals with the curved buckled sheet elements.
In the trial No. 1, tions are made:
~he

following assump-

(1) A width of 30 thicknesses of skin act with each stringer en the upper or ccmpress1ve side.
(2) The area of the curved sheet betwee~ the effective sheet widths as fOW1d in (1) is modified to give ~~ eftective area by ~ulti plying by a K ~actor of 0cr/ab. where a cr 1s the buckling compressive stress and 00 is the bending stress at the center of the curved sheet element assuming 32000 at the extreme upper fiber of the beam section and zero at the horizontal center line, With linear variation in between these points.

Ac (.67 0b)(.75 h), Whence,

160000 2.0 sq In .75x50x .67 x 32000 ...

Part of this total area is pr~vided by the effective skin area. The ef~ect~ve width to use with each rivet line equals w = Ct ,;E/crst. We will take C = 1.7 which is a commonly used
va Lue ,

NOTE: Since the entire skin on ~he lower half t s e r rec t tve , a mor-e logical aasumpt t on '.... cu.Ld be to guess at the l~cation of the neutral axis and use a variation of ab between the neutral axis and the extreme fiber. ~hiS approaCh will not be used in this example.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

TABLE A20.1

A20.5

TRIAL NO. 1

, -,

Area a aq.i:n. .~

.-.

;1
I
A~

."

!
5

,
'N

TRIAL

'0.
10

7
N

az 2

z'
24..21 18.2 22'1 13.3 6.9 I -3.2 _6.9 -10.1 -13.3 -15.8 -18.2 3.631 3.30 2.73 1, 99 1. 04 -0.72 -0.83 _2.28 _1.60 _2.53 _2.19 -3.24 _2.64 _3.80 .. 2.90 _2. 19 1

, , I "

. :1 e
010 .:: 12 ~ 13

.15
lS

87.' 72.6 I 25.38

, ::: 0 :;~ ,0 ., , ~, , . ," b " 27.58 _30020 1.0

.. - ., -.. .. --
1.03 1.12 1.28 1.66 2.88

=~.o

-.

,. " I "' >.

11

12

E:<planatory Notes for Table Ala.l

TRIAL NO. 1 Col. 1 For DUlllbering of sUffiiiei='S'iDd sh.ee't elellMlDts, .ee Fig. xao.a. Col. 2 Stiffener area .12 30 032 x .032 .18 re. s'ti!feners 10 Below 2. the centerline each s'titfener co~idel"ed aC'ting sepal"ate1y. The entire skin be'tween conllider_ s'tifteners ed as uui't. CoL J All z' m llIeaaur~ e' to horizontal cea-, terlille aJtis.

.,
4..19 3.88 3.36 2.69 1.78 0.72

,z2

::: 15

-16 .,7 .1.

." I
':21
::: 22 .23 ; 19 '>120

.15 .15 .15 .12 .224 .1201 .224. .12 .160 .12 .16 .12

.20.31 _22.0 .16 _23.7 .12 -24.2 .088 24.91

_27700 4.9.71 21.58 -23550 26.4 16.68 _18200 7.2 10.28 -11200 3680 3.38 2.3 -0.18 5.7 _3.52 c 23.0 _6.72 21. 3 -9.92 40.0 ! 12.42 39.81-14.82 65.81 16.92 ~ 58.0i-18.62 90.0 20 32 70.2 1_20.82 ~ 54.5 -21. 52

.032 .033 .036 .04.1 .053 092

.152 .153 .156 .161 .173 .212

"

.226

"

.120 .226 .12 .160 .12 .16 .12 .18 .12 .088
2.627

.04 _0.42
-1, 52

_L19 _L99 _1.78 _2.70 _2.24 _3.25 -2.50 _1.89 _2.86

115.4. 98.5 72.4. 44.7 18.3 2.' 0. T 1.5 10.2 11.8 24.7 26.4 45.6 41.6 66.0 52.1 4.0.7

'

.. '....'. '. .
'"
,
,~

"

TRIAL NO. 2 ~

Col. 7 Col. 8

aUK. 2.4.9

-12.23 714..3j

.73

z s dis'tallce 'cc neutral aXis found fro. l"l!suits of Trial No. 1. 1600000 z/1470 ~h ~ effective widtb based .n S'treS8 1111 Col. 7

,.

TABLE A20.2
TRIAL NO

c
j

.. ", .
~

"
2.27 4.G4 ".04 4.04 6.04 6.52

,
,
.032 .032 .032 .032 .032 .032

OM" o

" ~

1 3 5 11

11 11

. . ... ," - ". , . -.. 5 6 7

<,

TRIAL NO

10

11

12

_b

"'

<.

"

z'

aa ' az,2

<

- .- - .. " ... , , .- .. . ... -t ...

, .N
N N

13

Ofl

;l

0
0

"~
<,

15

16

17

18

'b

"

oz

.,2

>" ON

>< ,

: i

" ae "
ae

-9000 -9000 _4510 _2600 _2600 -2600

-31900 -30300 -26000 _20200 _12900 _ 4100

.282 .296 .173 .128 .201 .63

.020 .038 .022 .016 .039 .131 .266

50 12.4 ".91. 23.70.90 21. a 20.30.45 9.1 23.68 15.8 0.25 3.9 19.18 10.1 0.40 0 3.2 0.42 1.3 1 i1~:::
2.92 52.0

_29500 l;~:;: -25800

_30800

_20900 _14700 7200

.292 .305 .174 .124 .177 .362

12.,. 3.99 3.93

13.80 5.53
4..73

.021 .sa 16.7 .039 1.06 28.7 .022 0.52 12.3 .01510.29 5.5 .031 0.42 5.7 .0155 0.36 2.'

TOfALS

TOTALS

.183

3.24 71.3

Exp taoato rv Notes to r Table

Azu.z

Explanatory Notes for Table

Azu.z

Trial

1 Col. 1, 2, 3. Col. 5 Col. 6

Col. 9

xo.

Cb val"ieS as

(see Fig. Ala.l for mean1ng of terms) .E 10,300,000 for alW11inwa alloys :II. s'tl"a1ght line fl"Olll 32000 top of cell to Z8l"O at centel"11n8 distance frOlll c81l'tro1d of element to z' centerline ax1s of cell

.,

Trial

xe, 2
Col. 12 Col. 13 Col. 14 z - distance to ~eu'tl"ll.l ax1. as found ill results of Trial 110 1 Ob ; 1600000 z/1470 based on s'tress (1) of Col. 13

Results of Trial !l'o.

:!.

Results of Trial !fo. 2 Total effective al"ea (2.627 + .183)2 5.62

Considering resuLts of both Tables and multiplying by l since only one half of cell was considered: ia total effec'tive al"ea (2.4.9 + .266)2

~az. (_0I.86 + 3.24.)2.0.76

Z ... 76/5.62 135" above N.A. of trial 1, rNA .. 2(673 + 71.3) - 5.62 Jt .135' ~ 14.89

_ 5.51 sq. in.

,az, (_12.23 + 2.92)2. _18.82

~Zaz'/Za

INA 2(714.3 ... 52) _ 5.51 x 3.38

.. _18.62/5.51 .. -3.38" 2

14.70 in.

A20.6

FUSELAGE

STRESS

ANALYSIS

Due to the s~~etry of the section, Tables A20.l and A20.2 ~~'le c~lculatior.s for o~ly onehaIr of the material, ttus the results are ~ultiplied by two. General explanatory ~otes are given below each table. The results of trial No. I give a neutral axis 3.38 n below the center line and a moment of inertia of 1470 in.'. In Trial No.2, the effective sheet widths are based on the ~oment of inertia of 1470. The results of trial No.2 give a moment of inertia of 1489 in.' with a neutral axis .135 n above the first location. If a third trial were used, making use of the 14S9 moment of inertia, the change would ce quite small since the effect of a small change in stress on the effective sheet width is negligible. The compressive stress on stringer No.2 using the resulting moment of inertia and neutral axis location, therefore beccmes
"b
T~e

q~f
.

, .., ....

= 0+ i~&
shear stress
~

19.57

= .0132 Vz It./~n.
.0132 Vz / . 032 = .413

= q/t

V z

The average shear stress on ~~e secticn would be 'Tav' = Vz/ 2ht = Vz/2x30x .032 = .312 Vz. Thus for this sr2pe 0: crass-sectien and stringer a~angement the ~axi~um shear stress is .413/.312 times the average shear s:ress or apprOXimately 4/3 times as large. The procedure as given above is quite conservative relative to the true or actual margin of safety, because a linear variation at stress with strain has been assumed and failure of the section is assumed to occur when the most remote stringer reaches its ultimate compressive stress. Actually in a static test of a fuselage to destruction, the f~selage section as a whole will not collapse when one stri~ger ~uckles, but will continue to take increasing ~oad ~til other stringers have reached their ~ltixate strength, Furthermore, in a typ i ca I ruse Iage struc~ure, stringers of various sizes, sha~es a~d Lherefore different compressive strengths are used. ~~d thus to obtain a better ~easure of the ultimate strength of a ~uselag8 section, modifications in stress procedures are ~ace to measure stringer effectiveness. T~is subject waS discussed in some ceta11 in Arts. :1 and 12 of Chapt er- A19. ':.'0 illustrate s'tz-tnger- effectiveness in fuselage bending stress analYSiS, a simple example problem will be ,resented.
A20.5 illtimate Bending Strength of Fuselage Section. Example Calculation.

= l'lyZ/I Y =

160000 X 27 45/1459

29500
~si

The allowable stress was 32000, hence the margin ot safety is (32000/29500) - 1 = .OS or eight percent. If a s~aller ~rgin of safety was desired some ~aterlal would be eli~inated and the calculations of Tables A20.l and A20.2 would be repeated. Calculation of Shear stress in Skin at Neutral Axis The equation for the shear flow q point on the skin is,
q. qo - Vz zaz - -- - - - - - Iy
~t

same

(3)

Due to symmetry of cross-section about the Z axis the shear flow qo 1s zero at a point on the center line Z axis. The summation of the term az between a point on the Z axis and the neutral axis is given in Table A20.3. The values of areas (a) and arms (z) are taken trom Tables A20.l and A20.2.
TABLE A20. 3
B1e_llt xo, (l)
(J) (4) (5)

Fig. A20.3 shows the cross-section of a circular fuselage. The Z stringers are arranged s~etricallY with respect to the center line Z and X axes. Three sizes at Z stringers are ~sed as in Fig. A20.4 and are labeled S., Sa and 8 3 , These symbols are usee. on Fig. A20.3 to indicate where each type of stringer is used. The strlr~ers en each side cf the section are numbered 1 to 13 as shown on Fig. A20.3. Fig. A20.5 shows a plot of the stressstrain curve ~or the three stringer t~?es loaded in compression and with a col~, ler~th equal to the fuselage frame spacing. Fig. A20.5 also shoWS a tension stress-strain clagram for the material which 1s aluminum alloy (2024). The ultimate bending strength will be calculated ~or bending which places the upper portion in compression.
~llustrated

..
_ _ _ _ _ _ _ _" _ .13) .13) .13) .U)

'2l

(&,

(la)
(U)

'8l
(9)

(7)

(12)

.021(28.28 .152(21.58 .039(27.08 .153(25.38 .022(23.68 .156(::1.58 .015(19.18 .161(16.68 .031(13. <18 .173(l0.28 .055(6.58 .212(3.38

.13)
.13)

.13)
.13)

.13)
.13)

.13) - .13) Total

..

0.59 <1.17 1.05 3.86 0.52

3.3<
0.28 2.66 0.<11 1.75 0.35 0.69 19 67

Substituting in equation (3)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A20.7

2 Skin =- .032 4
5

S, S, S.

~'

.006

S,

T
-,
7" S,
10

CL
X

AxiS""

Since the location of the neutral axis is unknown, a location will be assumed, namely, 7 inches below the center line axis as shown in Fig. AZO.3. The entire calculations for deter~lning the effective moment of inertia cab best be done in Table form, as shown in Table A20.4. Due to Sj~etry about the Z axis only one-half of the structure need be considered since the results can be multiplied by two. Column (1) lists the stringer numbers relative to location and Column (2) according to types Sl' Se and 8 3 , Column (3) gives the stringer area. On the tension or lower slde of the section, the skin 15 all effective and that area of skin halfway to each adjacent stringer is assumed to act with stringers numbered 9 to 13 and this skin area 15 recorded in Column (5) an the compressive side the skin is only partially effective. The effective width w tor each stringer rivet line depends on the stringer stress. We will take the effective width w 1.9t The effective area Ae will then equal wt. These effective skin areas are recorded in Column (5). In solVing this equation the stringer stress crST has been taken as -36500 psi on stringer number (1), and then varying linearly to zero at the neutral axis as indicated in Column (4). This assumption Is not true but accurate enough to obtain effective skin areas. To illustrate, consider stringer number (1). The effective area As equals wt- = 1.9 x ,032"(10,300,000/36500)i- = ,032 sq. In, Column (6) gives the sum ot the stringer and effective skin areas or As + Ae.

~medY
Neutral Axis

,,( '8' 11

S,
~

IQ": Y

12

13

- - - - 0036
Fig. a Strain Diagram

Fig. A20.3

l'~~

r:~I"" -'Lno?l~
B
Stringer 81. Area =.135 Stringer SII Area = .18
Fig. A20.4

. .!!

13 r- 16 ..,

vE7OST.

C~ [t 1:

Stringer S 3 Area = .08

In this example problem, the effectiveness of the curved sheet panels be~~een the sheet effective widths will be neglected since its influence is small. It could be included as illustrated in the preViOUS example problem. Column (7) lists the distances tram the assumed neutral axis to the centroid ot each stringerskin unit. We now assume that plane sections remain plane or a linear strain variation. Referring to Fig. A20.S. it is noticed that when a unit strain of .006 15 obtained in stringer S1. type the compressive stress 1s 36500, which represents its ultimate stress. Stringer (1) is of Sl type and Is located tartherest from the neutral axis. SUb. Fig. (a) of Fig. A20.3 shows the strain diagram with .006 at stringer (1) and varying as a straight line to zero at the neutral ~~is. Column (8) of Table A20.4 records the unit strain at each stringer centroid. The true stress at each stringer pOint due to :hese strain values is read from the curves on Fig. A20.5 and recorded in column (9) of the Table. It should be noted that stringer (3) although closer to the neutral axis than stringer (1) carries a higher stress than stringer (1). This is pOSSible because when stringer (1) reaches Its ~aximum-stress, it bends but continues to hold the same stress with

A20.8

FUSELAGE

STRESS

ANALYSTS

TABLE A20. 4

3
Stringer
Area

4
Linear Stress

5
Effecttve Skin
Ar-ea

7
Arm

8
Unit Strain
E

9
Stress ATE.

10

11

12

13

stringer
No.

Type

A,
0.135 0.135 0.180 0.135 0.080 0.080 0.080 0.080

a' b

Total Stringer Area

I
K=o/cr1.
1. 00 1. 05

Ae
.032 .034 .036 .039

A
.167 .169 .216 .174 .124 .134 .154 . 080 .296 .296 .296 .296 .296

Z' (in)
35. 7

KA

KAZ'

KAZ' (Z = Z'I

1 2 3 4 5
6

S, S, S. S, S, S, S, S, S, S, S, S, S,

-36500 -34700 -31000 -26600 -20500 -13400

-.00600 -.0057 -.0051 -.0048 .0034 -.0023 -.0012

-36500 -36500

. 167 .178 .272

5.96

213
I

33.8
30.3 26.0 20. 1 13.7

6.00 8.25 6. 10 3.85

203 250 159

I
I I
I

I -39100
-36000 -31500 -24000 -12500

1. 26

1. 35 1. 54 1. 79 1. 75

.235 . 191 .240 .270 .080

.044
.054 .074

77 45 13

I
i

3.30

7 8 9 10
11

-7150 0
6130 12280 16800 20400 21700

7. a
10.0
- 6.0

I 1. 89

a
.216 .216 .216 .216 .216

a
.0010 .0020 .0028 .0034 .0036

a
10000 20500 30000 35000 38000

1.0
1. 63 1. 67
1. 78

a
,

a
17

O. 080
0.080 0.080

.4831 -2.88

-12.0 -16. 5 -20.0

-21. 2

I .494
.526

i
I

- 5. 91
-8.69

I
I

71 144 202
232 1626

I
I
I

12 13

0.080
0.080

1.71 1. 75
Z

.506 .517 4. 159

-10. 10

1- 10. 93
-3. 16

increasing strain, but stringer (3) which has not reached its maxi~um strength of 39000 continues to take increasing load. Since we wish to use the beam formula CO = MyZ/I x in computing stresses, we must modify the stringer areas to give a linear stress variation since the formula is based on a linear stress variation. The stringer modification factor K equals the ratio of the true stress in column (9 ) of Table to linear stress value in column (4) or K = a/a 1.. The results are recorded in column (10) The modified stringer areas are then equal to KA and are recorded in column (ll) Column (12) gives the first moment of the modified areas about the assumed neutral axiS, giving a total value ot -3.16. The distance Z fram the assumed neutral axis to the true neutral axis is thus,

to Z. The effective moment of inertia is there tore twi ce the sum of Column (13) or 3252. Calculation of Ultimate Resisting
M~ent.

The maximum stress at the most remote ' is 36500. From stringer which is number (1 -i the beam formula,

M x

="b1x/Z = (36500 x 3252)/30.7 + = 3,260,000 in. lb.

0.7

This bending strength when compared to any des1 gn bending moment about the X axis Nould give the margin of safety relative to benci~g strength. If the moment of inertia had been computed without regard to non-linear stress variation, or in other wordS, using K equal 1 for all stringers the neutral axis would have come out 4.9 inches below the centerline axis and the moment of inertia would have calculated to be 2382 in. The reSisting moment developed would then be (36500 x 2382 )/33. 6 2,600,000 in.lbs. Thus the true strength is 25 percent greater than the strength for linear stress variation. This result explains why such structures test overstrength if deSigned on linear stress variation basis. Atter stringer stresses are obtained

Z=Z

KAZ' (ZKA

- 4.159 - - 0.76" TIle true N.A. would fall about .70 inches below assumed pOSition. The effect on total sum at Column (13) would be negligible, thus Table A20.4 will not be revised. Column (13) gives the calculaticn of the er.rect r ve moment of inertia with Z 1. being equal

_ -3.16 _

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

A20.9

uSing the modified areas of Table A20.4, the true stringer areas ~ust be used ~o tind the true stringer loads, which must be used in the shear flow analysis.
A20.5 Shear Flow Analysis for Fuselage Structures

I y = (15:1 X .Lx z ) + (13.86 0 1 1 + 1 0 . 6 1 : 1 + 5 . 7 4 011).1

180 in.4.

The shear flow analysis can be made once the effective cross-sections of the fuselage are obtained. The procedure is the same as was illustrated tor Wing structures in Chapter A19. To illustrate, two example problems will be presented. Example Problem 1. Symmetrical Tapered Section.

Table A20.5 gives the necessary calculations for determining the flange bending stresses and the net total shear load to be taken by the cell skin. Since the cell 1s tapered, the stringers have a z component, thus the stringer axial loads help resist the external shear load. The summation of column (8) ot Table AZO.5 gives -333.4 lb. for a summation for halt the fuselage section. Hence, net web shear at station 0 equals:
V W eb == Vext + Vtlange

Fig. A20.6 shows a portion of a tapered circular shaped fuselage structure that might be representative of the rear portion of a fuselage for a small airplane. Since this example is only for the purpose of illustrating shear flow analysis, it will be assumed that the 16 stringers are the only effective material. In an actual stress analysiS, the effective cross-section would have to be used as illustrated in previOUS articles A20.3 to A20.5. The problem will be to deternine the stringer stresses and the skin shear flow stress system at Station (0) under a given load system at Station (150) as shown in Fig. A20.S.
Solution No. 1 - Solution by Considering Beam Properties at Only One Section.

= 2000 +

(2 x -333.4)

-= 1333.2 lb.

The results in this particular problem show that at station 0 the flange stringer system resists one third of the external shear load. At station 150 the web system will resist the entire external shear load at 2000 lb. since the load in the stringers is zero. In actual deSign the net Neb shear should be used since in many cases it will decrease the sheet thiCkness reqUired one or more gauges.
Calculation of Flexural Shear Flow.

If the Change in longitUdinal stringer or flange material is fairly unitor.n this method can be used nith little error i~ the resulting shear tlow stresses. Moment of inertia of section at station (0) about centroidal Y axis:

VZ(web) Z a Z 1333.2 = qo - 180 Z az ly (A) -= qo - 7.40 Z az

= qo -

--

Due to symmetry ot the sectIon about the Z axiS, the flexural shear tlow in the web at the center line is zero. Therefore, qo will be taken as zero and the summation in equation (A) will

Area of all Stringers

16

:E

1" Frame spacing '" 15"

2000t

15 14 _stiffeners 13

12 11 10
"

S
150" Length not to scale Fig. A20.5

Section at Station 0 STA.O

.\
STA.150

Section at Station 150

A20.10 TABLE A20. 5

2

FUSELAGE

STRESS

ANALYSIS
q3~

= -7.06 = -7.06

23.66

16.60 etc.

a rea O'b '" Px=b -1667z (3)(') a (paf] lba.

-25000 -23100 -17700 - 9550 0 9550 17700 23100 25000 -1250 -2310 -1770 - 955 0 955 1770 2310 1250

6
dz

q~8

27.91 = 20.85

sueener No.

Arm

dX
-.0333 0 -.0308 .0127 -.0236-.023 -.0127-.030 o -.033 .0127-.030 .0236 -.023 .0308 -.012 .0333 0 -41. 6 o -71.3 -29.4 -41. 7 -41. 7 -12.1 -29.5 0 o -12.1 29.5 -41. 7 41. 7 -71. 3 29.4 -41. 6 o -333.4 0

5
5 7

2 3

8
9

.05 15 13.86 .10 10.61 .10 5.7 .10 o . 10 - 5.7 . 10 ~10.61 .10 -13.8 .10 -15 ".05

Fig. A20.7 shews the results i~ graphical tor,n, on the left side o~ the section the shears are of the same Sign and t~erefore aGd together.

Shear taken by stringers",

1/2 of stringer one is assumed acting with each hail of cell.

NOTES: Col. 4 ab'" -Mz/l y '" -2000 x 150 x z/lBO = -1667z

Col. 5 Total x component of load in stringer member. For practical purposes, it equals axial load in stringers since cosine of a small angle is practically one. Col. 6 The slope of the stringers in the z and y directions can be calculated from the dimensions at the two end sections and the length of the cell. (see Fig. A20.6) Col. 8 The in plane components of the stringer axial &: 9 loads at stancn O. Fig. A20.7 Solution No.2. Shear Flow by Change in Stringer Loads Between Adjacent Stations. t.ll' Method.

start with stringer (1)

The shear ~low will be calculated by C8nsiderlng the change in the ~x~a: load in the lor.gitudlnal str1~gers between fuselage sections at stations (0) and (30). Fig. A20.8 shews the bea~ section at station 30, the st~inger areas jeing the Saffi8 as at station 0, but the sec t t cn as a wno Le is smaller due to the taper of t~e cell.
area of each stringer"" . 10
1 2

qu
qUi
qa~

= 0- 7.40 x
= -

.05 x 15

= - 5.55 lb./in.

5.55 - 7.40 x .1 x 13.86 - 15.80 x 10.61

= - 15.80 - 7.40 x .1

=- 23.66
=27.91
~

qu

=- 23.66

- 7.40 x .1 x 5.74
~oment

The torsional

T about the centroid

ot the section at station (0) equals 5 x 2000

10000 in. lb. (clockwise Nhen looking toward station 150). Due to the symmetry of t~e sectton at station 0, the in-plane components ot the stringer loads produce zero ~oment about the section centroid. For eqUilibrium a const~~t inte~~l shear flow q 1. is necessary to make Z Mx :: 0
q ,. -10000 =-1 ""'-Ls ="7 = -7.06 2A =-=--=""' a x n "x

Table A20.6 gives the calculations tor the shear system. ~he procedure ~s the same as illustrated :or wing struct~res in Chapter A19.
tle~Jral

lb ./~n.

Adding the torsional shear- f Low q l.. ':0 t.ne flexural shear flow Q, the ~ollowlng results are obtained:
qu

q."

= -7.06

= -7.06

+ 5.55

.,.. 15.80

=-1.51 = 8.74

Ib./1:1.

Comparing the r-esu.Lt s ot column 13 wi th the flexural shear flow as ~ound by sol~tion No.1, we find the second sol~ticn gives a ~axi~w~ shear flow of 28.95 lb./~n. aga~~st ~ value cf 27.21 for the first soluticn. T~e first ~ethod deaLs wt th the pr-cper-t i es at only ov:e section and this cannot include the etfect~ of change in ~anent of inertia on the shear ~low. The

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES TABLE A20. 6

1

AZO.11

10

11

12

13 Flexural Shear Flow

Sta. 0 S13. 30 Strtnzer Load Panel ~P Sta. 0 Sta. 30 Sta, 0 sta. JU Taper 30 Area b-;; Area Sta. a Sta. 30 crb'" Stringer -16672 -16222 P x '" a a P x ' - (Col. 8 - Col. 9) Corr. Arm Arm No. (psi. ) (pet. ) "b a Factor 30 sq. in. sq. in. z "b a

ao
5. 76

dPK

q=Z~ 30 lb. lin.

S. 76

.05 .10 .10 .10 .10 .10 .10 .10 .05

. os
.10 .10 .10 .10 .10 .10 .10 .05

15. 00

14.00 25000 -21300 -1250

-1085 -1970

6.17
11. 33 8.83
4.65

.935
.935
.935 .935 0

2
3

13.86
10.61

12.93 -23100 -19700 -2310

9.90

10.60
16.36

-17700 -15050 1770

-1505

8.25

4
5 5
7

5.74
0

5.36
0

- 9550 - 8155 - 955 - 815

o
9550 17700
23100

24.61

4.34
28.95
0 28.95

0 955
1770

0 815

- 5.74 - 5.36
-10.61

8155
15050 19700
21300

- 9.90

1505

- 4.65 - 8.83
1l.33
- 6. 17

.935
.935

- 4.34 - 8.25
-10.60

24.61
16.38

8 9
NOTES:

-13.86 -12.93

2310
1250

1970
1065

.935
.935

5.76

-15.00 -14.0

25000

- 5.76

Col. 6 O'b = -2000 x 150 z/180 '" -1067 z

Col. 7 Cib ~ -2000 x 120 z/1S7. 2 = -1522 z

Col. 10 Change in axial load in each stringer between stations 0 and 30 divided by distance between Stations. This result represents the average shear flow induced by the loading up of each stringer between stations 0 and 30. Col. 11 The width of a skin panel at Station 0 is 5.88 inches and 5. 5 inches at Station 30. The shear flow on the edge of the panels at Station 0 equals (5.5/5. a8) AP/30. (See Art. AlS. 18 of Chapter A15 for explanation). This refinement is usually neglected and the average values as given in Col. 10 are used which are conservative. Col. 13 Due to symmetry of structure, the shear flow is zero on z axis. progressive summation of the shear flow values in Col. 12. Thus shear flow at any station equals the

second method is recommended tor practical analysis procedure.

will assume the stringers are the only effective material. In actual design practice the effectiveness of the skin and each stringer would have Since the section is symmetrical, there are to be considered as explained in Articles A20.4 no moments i~Guced by the in-plane components of and 5. the stringer forces at station O. The problem will be to determine the stringer stresses and the skin shear flow valThe torsional sh~ar flow forces are the ues at station (0) due to the given external same as in solution ~ethod No. 1 and these are added to the values ot column 13 of Table A20.6 loads ot ?z = 4000 lb ?y = 1000 lb. and Px = 1500 acting at station (150) as .snown in Ff.g , and give a pattern similar to ~ig. A20.7. A20.9.
SOLUTION:

A20.7 Example Problem. Tapered Circular Fuselage with Unsymmetrical Stringer Areas.

Fuselage cross-sections are seldom all symSince Ne choose to use the ~ method in metlical relative to stringer and skin areas finding the shear flow system at station (0). because the practical fuselage has cut-outs such we will find the stringer loads at two stations, as dcors, etC. To illustrate the unsymmetrical namely, station (0) and station (30). The first case a simplified case will be presented. steo is to find the moment of inertia of each fuselage section about centroidal Z and y axes Fig. A20.9 shows a portion of a ~apered and the product of inertia about these axes. ~~se1ag9. The stringer a~eas are such as to Table A20.7 (Col~~s 1 to 11) gives the calcUmake the cross-sections unsymmetrical ~elatlve lations of ~he section properties for station to oending material. Again tor Simplicity, we
. ~.>~
(l

I (J
-~

A20.12

1-

FUSELAGE

STRESS

ANALYSIS Fig. A20.9

8 .4 8 --I,
.1
(.1)

.1)1 12"

'i

,c

8.48

(.6)

1
Cep-erUpe

d
(.2) (.2)

.2)

150" Sta. 0

150

L- x

,I

Section at Sta. 150

The skin stringers are located symmetrical:y with res~ect ~8 :he Section at Station 0 centerline axes, however' the stringer areas as g tven 1::1.( ) or. the figure are not symmetr~cal With these axes. It is assumed in this ;roblem that the stringers taper uniformly between the values as given for station Gane 150. The cell would of course have interior transverse ~ra~es which are not shown on the figure.

Z'

TABLE A20. 7 Section Properties at sta. 0

1 2 3
Arm

Total Stringer Loads at Sta. 0

a
ay,2

10

11
yo

12

13

14
P s ",a(Ob-od
(2) (12 13)

Stringer Area No. a

"
10.5 18.98 22.50 18.98 10.5 -10. 5 -18.98 -22.50 -18.98 -10.50

Arm y'
-12.00 - 8.48 0 8.48 12.00 12.00 8.48 0 - 8.48 -12.00

,,'
6.30 1. 90 2.25 1. 90 8.40 -8.40 -3.80 -4.50 -3.80 -3. 15

az,2

.y'

az'y'

z' -

O'c",F/Za

y' -

"b

'" -1500/3.40 -441 -441 -441 -441 -441 -441 -441 -4'11 -441 -441

a b c
d

e
f

h
i
j

.80 .10 .10 .10 .80 .80 .20 .20 .20 .30

66.20 36.00 50.80 36.00 88.10 88.10 72.00 1~1. ,0 72.00 33. 10

-7.20 86.50 75.50 11.36 -0.85 7.20 16.10 19.84 -0 0 0 23.36 0.85 7.20 16.10 19.84 9.60 115.10 100.80 11. 36 9.60 115. 10 -100.80 - 9.65 1. 70 14.40 32.30 -18.13 u 0 -21.60 0 -1. 70 14.40 32.30 -18.13 -3.60 43.25 37.80 - 9. 65

-14.46 -10.94 2.46 6.02 9.54

-15080 -21960 -22600 -16742 - 7692 9. 54 11958 6.02 18798 - 2.4' -10.94 -14.46 4592

i;;~~

-9312 -2239 -2304 -1719 -6506 9216 3673 3,0, 2634 1246 -1500

Sum

3.40

-2.90 643.9

8.40 403.2

- 37.70
General Notes: Col. 12 0b'" 307. Or - 936. Lz

Reference Axes Z' and y' are taken as the centerline axes. (see FIg. A20.9) Location of centroid and transfer of properties to centroidal axes. Z '" -2.90/3.40 '" -. 855"

y '" 8.40/3.40

Col. 14 . Since the total tensile stresses equal to total compressive stresses in bending, the sum of Col. 14 should equal the external applied normal load.

'" 2.46"

Iy '" 643.9 - 3.40 x 855 2 '" 641. 4

Iz '" 403. 2 - 3.40 x 2.46 2", 382.6 I%y'" -37.7 - 3.40 x 2. 46 x -.855", -30.55

_.~

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES TABLE A20.B Section Properties at Sta. 30

A20.13

Total Stringer Loads at Sta. 30 8 9

3
Arm

10

11

12

13

14

Stringer Area No. a

Arm y'

az'
4.90 1. 77

az,2

ay'

ay,2 62.8 6.3 0 6.3 83.0 83.0 12.6 0 12.6

32.6 299.2

z:

y: y' -

az'y'
a' - Z

c"'F;Za

.Ps

=-1500/2.98
-503
-503 -503 -503

a(Ob + Gel

a
b

c
d e
f

.50 .10 .10 . 10

9.80
17.72
21. 00

17.72
9.8

.66 .66
.20 .20 .20 .26
2.98

- 9.8 -17.72
-21. 00

-11.2 - 7.92 0 7.92 11.20 11. 20 7.92 0

- 7.92 -11.20

2.10
1. 77 6.47 -6.47

i
j

-17.72 - 9.a

-3.54 -4.20 -3.54 -2 55

-3.29

48.1 -5.60 31. 4 -0.79 44.1 0 31. 4 0.79 63.2 7.40 63.2 7.40 62.8 1. 58 ,".2 0 62.8 -1. 58 25.0 - 2. 92 520.2 6.28

-54.9

-14. a 0

10.90 18.82
22.10 18.82

-13.31
-10.03 - 2.11 5.81 9.09

-14800
-21090

-7651
-2159 -2195 -1615 -5011 7100

14.0

-21447 -15647
- 7088

72.5
-72.5 -28.0
0 28.0 28.6

10.90 - 8.7 -18.60 -19.90

-16.60 - a.70

503
-503 -503 -503 -503 -503

9.09 5.81
- '.11 -10.02 -13.31

-11282
-17583
17::143

3416

12135 3570

,."" 2327
797

Sum

-26.3 N Z'

-1500

~
Reference Z' and Y' axes are taken as the centerline axes.
b

Z = -3.29/2.98 = -1.10"

I:~ 1. 10
.10 7.9 2'[
Z
i

~.:ic.

.1~.?
2' .66 e Section at Station 30

Y=
ly

6.28/2.98 = 2.11

.sot9.a

= 520.2 - 2.98

x 1. 10 2 ::516. 6

Iz = 299.2 - 2.98 x 2. 112 :: 286.0

I zy = -26.3 - 2.98 x 2.11 x -1.10:: -19.4

If.~l=-)
j

,~l.
.26 .20

;1.-

tr y Z2.11
.20

z'" ~ l:;.:,{
y
I

ll.~~

Fig.A20.10

(0) and the similar columns of ~ab1e A20.8 gives the calculations for station (30).

Before the bending and shear stresses can be calcalated, the external bending moments, shears and normal forces at stations (0) and (30) Calculation of Bencting Stresses. must be known. Station (0): At station (0): ~he bending moment about y neutral axis at station (0) equals,

My M z Pn

= 4000 x 120 + 1500 X 8.10 = 492150 in. lb. = -1000x120+1500x2.11 = 116820 tn Lb , = -1500 lb., Vz = 4000 lb., Vy = -1000
s

lb.

where
K1.

My

=Pz
=

(150) + Px (7.85) 4-000x150+1500x7.85

= 511800 tn.J,b ,

K"
K~

= Iyz/(IyI Z - Iyz:il) = Iz/(Iy!z I y z .)

= Iy/{Iyl z - I y Z Ji)

~z

Py (150) - Px (2.46)

= -1000 x
4000 Ib . and Vy

150 + 1500 x 2.4-6

= -14-6310

in .Tb ,

The shears at station (0) are Vz = P z ?y = -1000 lb.

SUbstituting values from Tables A20.7 and A20.S:

K1.

= -30.55/(641.4 x 382.5 -30.55.:1) =

to

cen~oid

The normal load Pn at atat i'on (0) referred. of section eq~ls ZPx = -1500 lb. In a similar
~a~~er,

= -30.55/244670
K~

= -.0001248
.00156

K. = 382.6/244670 =

the values at station

(30) are.

(see Fig. A20.10)

= 641.4/244670 = .00262

A20 14

FUSELAGE

STRESS

ANALYSIS

Substituting K values in equation for ab: r.00262 x-1453l0 - (-.0001248 x

61.1800)J y - [00156 x 5E200 (-.0001248 x -145310)J z

~cments

all ~n~e~nal anc exbe zero. Col~~ (7) of Table A20.9 ~iJes :~e ~oment of ~he flex~al shear about tn; s potnt , (See notes and 3"ig. be Low ':'able ""'''''.~ '" expjanat i on. )
)l~ne, Q~

in :he

ternal

~orces ~us~

TABLE A20. 9 SHEAR FLOW CALCULATIONS

whence crb

= 307.0

y -936.1 z

(pl~s

crb is tension)

4
q

5
e

7
mq

Station (30):

K
K,

=-19.4/(516.6 x 286 - 19.4')

=
-19.4/147620

Stringer Px at P x at No. Sta. 0 su. 30

30

.6.Px Z~
,JU

sqc m,

ql

= -.0001315
.0035

a
b

= 286/147620 = .001936
= 516.6/147620 =

c
d

K" ab

e I
h
i j

=- [0035 c -116800 (-.0001315

(-.0001315 x 492150] y - [001936 x 492150-

-9312 -2239 -23 4 -1719 -6506 9216 I 3673 3809 2634 1246

x 116820I] z

a
NOTES;

55.37 2.67 - 1 347 -1615 -5011 49'83 ...7100 -70' 53 _ 8' -7 3416 3489 -10'~" 2327 -lO'2~ 797 -14'97 .

-7651 -2159

55.37150.04 83001-52 3.37 58.04202.46118201-52 6.04 61.67202.4612500 -52 9.67 65.14 150.04 9780 -52 13.14 114.97 252.0 29000 -52 62.97 44.44 150.04 6660 -52 - 7.56 35.87202.46 7270 -52 -16.13 25.20202.461 5100 -52 -25.8 14.97 150.04 2240 -52 -37.03 0 252.0 0 -521-52.0 Sum 92670 I

whence
"b

= 344.3y-

937.7z

Column (12) in Tables A20.7 and A20.8 gives 'th..e results of solving the equations for ab' Since an external load of 1500 lb. is acting nor.nal to the sections and through the section centroids, an axial compressive stress crc is pr-oduced on the sections. (See CnLtzane 13). The total load Ps in each strlr~er equals the area of the stringer times the ccmbined bending and axial stresses. (See column 14 of each table).
Calculation 01 Flexural Shear Flow q,

Col. (2) and (3) from tables A20.7 and A20. 8 Col. (4) .6.Px :: [PXSt.a. 0

P~ta. 3~

CoL (6) m :: double areas (see Fig. a). Col. (7) mq '" moment of shear flow q on each web element about 0' (Fig. a)

Moments Due to In Plane Components of Stringer Loads.

Table A20.9 gives the necessary calculations to determine the shear flow at station (0) based on the change in stringer loads between stations (0) and (30). The correction of the average shear due to the taper ln ~he skin panelS as was done in example problem (1), Table A20.6, column (11), is omitted in this solution since it tends toward the conser~;ative side. Since the ef:ective cross sec~ion is unsymmetrical, the value of the flexural shear flow qat any point is unknown thus a value rcr q at some point 1s assumed. In Table A20.9 the shear flo''i q in the web a j is assumed zero. Column (5) gives the results at other points under this assumption. Moment of Shear Flow about Intersection of Centerline Axes For equilibrium in the pl~~e of the cross section at station (0), the summation of the

Since the stringers are not normal to the section at station (0), the stringers have ~n plane 'components which :nay produce a eoment about the lntersection of the s~etrical axes which has been selected as a ~oment center. Table A20.l0 gives the calculations for the inplane components and their moments about paint
0' .
Moment of External Load System About Point (O').

The 1000 lb. load at station (l50) acting in the Y direction ~as a noment arm of 7 n about the pOint Or of station (0). Hence external moment

= 1000

x 7

= 7000

in. lb.

Therefore the total mcment about the assumed moment center 0' =

ANALYSIS AND DESIGN OF TABLE A20. 10 MOMENTS DUE TO IN PLANE COMPONENTS OF STRINGER LOADS

FLIGHT VEHICLE STRUCTURES

~.

A20.15

2
Stringer
No,

3
dv

4
Py = p2Y. xdx

6
dz dx

Px (lbs. ) -9312 -2239 -2304 -1719 -6506 9216 3673

Mo'
P y z'

OX
.0026 .0188

Pz M o '" Px..!!I. dx P z y'

a
b

c
d

0
-.0188 -.0266 -.0266 -.0188

e
f
h

248 42 0 32 173 245 69

0

2600 .0233 798 -.0421

0 -,050 608 -.0421

-1820 -.0233 -2570 .0233 -1310 .0421 0 .050 930 .0421 346 .0233 -1634

38 9
2634 1246

0
.0188 .0266

49 33
Sum

217 94 115 72 152 214 154 190 111 29

-2600 796 0 610

1820 2560 1307

y.

0 941 348
1612

NOTES: ---Col. (2) from Table A20. 7 Col. (3) equals the slope of stringers inyandz directions. (see Fig. A20. 9) Col. (5) Values of z' and (8) and y' from Table A20. 7. Fig. b shows the P y and Pz components from Cols. (4) and (7). Total moment about O' '" -1634 + 1612 '" -22".

Fig. A20.11 Shear flow distribution.

where large concentrated loads are applied can be determined by the procedure given in Articles 18 to 20 of Chapter Al9. A more rigorous a~~lysiS can be made by the application of the basic theory as ~iven in Chapter AS. The problem of shell stresses due to internal pressures 1s presented in Chapter A16. The strength deSign or the fuselage skin involves a question of combined stresses. The broad Droblem of the strength design of structural elements and their connections under all types of stress conditions 15 covered in Volume

92670 -22 7000

due to shear flow q due to in plane components or stringers. due to the external loads. in. lb.

II.
A20.9 Problems.

Total= 99648

(1 )
~'S'
I

Therefore :or equilibrium a moment of -99646 is required which can be provided by a constant shear flow q:l. around the cell, hence
q, ..

~.

T 2A

=-:-",9,,96;;,48;;_ i- = -52 2 x 957

lb./tn. (957 enclosed area or cell)

value or q Is entered in column 8 or Table A20.9. The ~esult1p-g shear flow 1~ any web ~ortiQn qr equals ~he algebraic sum of q and ql.' (See Col. 9, Table A20.9). F'~g. A20.11 shows the results in graphical form.
~is

+
,

20"

1/
i ,

Skin .035

Fig. A20. 12
A20.8 Discontinuities - Shear Lag - Pressurization Stresses - Combined Stresses.

A practical :uselage has ~ny cut-outs. The approximate ef~ect of these discontinuities as well as the shear lag er!ect at sections

Fig. A20.12 shows the ~ross-sectlon of a circular fuselage. All stringers have same a~ea. naTlely 0.12 sq. in. Skin thickness is .035 inches. Stringers are 1 inch in depth.

A20.16

FUSELAGE

STRESS

ANALYSIS
<-.-.....--'---'-= -~-=-

All ~terial is aluminum alloy. E = 10,500,000 psi. The ultimate compressive str~ngth of stringer plus its effective skin is 35000 pai' l For e r.rec t tve sheet width use w = l.9t (E/OST )"i". For buckling strength of curved panels use ocr = .3 Et/r. Determine the Ulttmate bending moment that the fuselage section will develop for bending about horizontal neutral axis. Use linear stress distribution. Follow procedure as given in example problem in Art. A20.4.
(2 )

""
/032

.~
J'

S,
.Ok

S,

s,
S:..
s~

s,

Fig.A20.13

Fig. A20.13 shows the cross-section of a rectangular fuselage. The dots represent stringer locations. Three types of stringers are used, namelY, S l' S.lI and 5:5' Fig . AZO.14 shows the ult~te compressive stress-strain curve for each ot the three stringer types and also the tension stress-strain curve ot the material.
(4) Sa~e as Problem (3) but change area of stringer no. (2) to 0.3 sq. in., thus makfng an l~symmetrical section.
A20. 10 Secondary Stresses in Fuseiage stringers and Rings.

Determine the Ultimate bending resistance ot the P~elage section about the horizontal neutral axis it the maximum unit compressive strain is limited to .008. Reter to Art. A20.5 for method ot solution.

ADDITIONAL DATA.
S.lI :: .25

pat

Area s'tz-tnger- S l :: .12 sq .fn, ; sq.in.; Sa :: .08 sq.i~. E = 10,500,000

(3) Fig. A20.1S shows a tapered circular fuselage with 8 stringers. The area of each stringer is 0.1 sq. in. A$sume stringers develop entire bending rBsistance. Find the axial load in stringers at station (110) due to Pz and Px loads at station (0). Also find shear flow system at station 110 USing ~P method. Use properties at station (90) IN OBTAINING AVERAGE
SHEAR FLOwE.

7
.110 Fig. A20.15

The stresses that are found in ~he stringers or longerons at a typical fuselage by use of the modified beam theory or by the more rigorous theory of Chapter AS, are referred to as primary stresses. Because of the necessity ot weight saving, most fuselage structures are designed to permit skin buoxj.tng , which means ttat shear loads in the skin are carried by diagonal semi-tension field action. This diagonal tension in the skin panels produces additional stresses in the stringers and also in the fuselage rings. These resulting stresses are referred to as secondary stresses and must be properly added to the primary stresses in the strength desi~ of the indiVidual stringer or ring. Chapter CII covers the subject at these secondary stresses due to diagonal semitension field action in skin panels. It is suggested to the stUdent that atter st~dying Chapters Al9 and A20, that Chapters ClO and Cll be referred to in order to obtain a complete stress picture for skin covered structures.

CHAPTER A21

LOADS AND STRESSES ON RIBS AND FRAMES

A21. 1 Introduction.

wing

For eer-ocvnamt c reasons the in the chord direction must be ~~lntaln~d ~lt~c~t appr~clable distortion. Unless t~e ~ing sk1n 1s qUite thick, spanwlse str~nge~s ~ust be ~ttached to the skin in order to Inc~ease :he bencins efficiency of the wing. Therefore to hold the Skin-stringer wing surface to c~ntour sha~e and also to 11~it the length of strln~ers to an efficient column compressive strength, 1~t2~~~1 support or brace units are reqUired. ~hese str~ctural units are referred to as wtnz ribs. The ribs also have anot nermajor pur~cse, na~elYl to act as a transfer or distribution unit. All the loads applied to tne '.'ling are reacted at the wtng sucportlng pOints, thus these a~pll~d loads must be transrarre~ into the Nlng cellular stru:ture ~om, posee o~ skin, strin~ers, spars, etc., and then reacted at the Wing support ?o!nts. The aJvlled loads nay be only the distributed surface ai~ leads which require relat~vely light internal ribs to ~ravide th~s carry throush or tr~nsfer ~equlr9ment, to ~ther rug~ed or heavy ribs which ~ust absorb and transmit large concentrated applied loads such as those from landing gear reactions, power ?lant reac~lons anc. ~~se Iage reactions. In between these two extreaes of applied load ~a~nit~des are such loads as reactions at sup~orting points for ailerons, flaps, leading edge high li:t uni~s and the many internal ceaa .vet ght leads such as fuel and military anmament and other installations. Th~s ribs can 'ffiry from a very light structure which serves ~rlmarlly as a former to a hea'TY str~cture which must receive and transfer loads involving thous~~ds of poundS.
c~ntou~

loads into the fusela~e shell such as those from landin~ gear reactions, wing reactions, tail reactions, power slant reactions; etc. The dead wei~ht of all the payload and fixed equipment inside the fuselaroe must be carried to frames by other structure such as the fusela~e floor system and then transmitted to the fuselage shell structure. Since the dead weight must be multiplied by the design acceleration factors, these internal loads become qUite large in ~gnitude. Another important purpose or action of ribs and frames is to redistribute the shear at discontinuities and practical wings and fuselages contain many cut-outs and openings and thus discontinuities in the basic structural layout.
A21. 2 Types of Wing Rib Construction.

Figs. A2l.1 to 6 illustrate the co~non types of win~ construction. Fig. 1 illustrates a sheet metal channel fer a leading edge 3
Skin

flange

.
Beam
Web

OiO

Since the airplane control surfaces (vertical and horizontal stabilizer, etc.) are nothing more than small size wings, internal ribs are li~ewise needed in these structures. The skirr-stringer CJnstruction which fOrTIS shell of the I~sela~e likewise needs internal forming units to hold ~he ~uselage cross-section to contour shape, to limit the col~un length of the stringers and to act as Fig. A21. 2 transfer agents ot internal and externally applied loads. Since a fusela7e ~us~ usually r~ve clear internal space to house the ?ayload such as passengers in a commercial trans2ort, theSe internal fuselage units which are usua:ly referred to as frames are of the aDen or ring type. Fuselage frames vary in size and strength from very light for.ner type to rugged heavy types which nust transfer large concentrated
~he

Fig. A21. 3

A21. 1

A21. 2

LOADS

AND STRESSES ON RIBS AND

FRAMES

.,( Wing Skin / Shear Tie

,-

- ....

Ribs in 3 Spar Wing.

Fig. A21. 4 Fig. A21. 5

o o Removable
Leading Edge Portion.

0 0

o 0 000 O~~~~~~~
Fig. A21. 6

stringer, single spar~ single cell wing structure. The rib 1s riveted, or soot-welded, or glued to the skin along it bo~~dary. Fig. 2 shows the same leading edge cell but with spanwise corrugations on the top skin and stringers on the bottom. On the top the rib flange rests below the corrugations, whereas the stringers on the bottom pass through cutouts in the rib. Fig. 3 illustrates the general type at sheet metal rib that can be qUiCkly made by use at large presses and rubber dies. Figs. 4 and 5 illustrate rib types for middle portion ot wing section. The rib flanges may rest below stringers or be notched ~or allowing stringers to pass through. Ribs that are SUbjected to considerable torsional torces in the ?lane of the rib should have some shear ties to the skin. For ribs that rest belaw~trlngers thiS shear tie can be made by a rew sheet metal angle clips as illustrated in Fig. 5. Fig. A21.7 shows an artist's drawing at the wing structure or the Beechcraft Bonanza commercial airplane. It should be noticed that various types and shapes of ribs and formers are required in airplane design. Photographs A2l.l to 3 illustrate t~~ical rib construction in various type aircraft, both large and smaLl , Since ribs compose an appreciable part at the wing structural weight, it is important that they be made as light as safety permits and also be'efficient relative to cost of fabrication and assembly. Rib development and deSign involves considerable static testing to verify and assist the theoretical analysts and deSign.
A21. 3 Distribution of Concentrated Loads to Thin Sheet Panels.

were used to transmit exter~al loads into the wing cellular beam structure. Concentrated external loads must be distributed to the rib before the rib can tra~sf~r the load to ~he wing beam st~~cture. In other wordS, a concentrated load applied cirectly to the edg~ 0: a thin sheet would cause sheet to buckle or cripple under the localized stress. Thus a str~ctural element ~sually called a web stiffener or a 'Neb r lange Ls fastened to the 'Neb and the concentrated load goes into t~e stl:fe~er which in turn transfers the load to the web. To ~et the load into the stiffener usually requires an end fitting. In general the c isur i , buted air loads on the wing surface are usually of such ~gnitude that the loads can be distributed to rib web by direct bearing of flange normal to edge of rib web without causing local buckling, thus st1f~eners are usually not needed to transfer air pressures to wing ribs.
EXAMPLE PROBLEM ILLUSTRA.TING ':'RJ.NSFER OF CON-

CENTRATSD LOAD TO SHSST PANEL.

Fig. AZl.8 shows a cantilever beam composed at 2 flanges and a web. A concentrated load of 1000 lb. is applied at paint (A) in the direction shown. Another concentrated load of 1000 lb. is applied at paint (E) as shown. To distribute the load ot 1000 lb. at (A)J a horizontal stiffener (AB) and a vertical stiffener (CAD) are added as sho'ND. A t1ttlng would be reqUired at (A) which would be attached to both stiffeners. The horizontal component of the 1000 lb. load which equals 300 Ib is taken by the sti~fener (AB) anc the ver~ cal component which equals 600 lb. is taken y the

In Art. AZl.l it was brought out that ribs

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A21. 3

10G
~lange

L-------'"'

Fig. A21. 7 General Structural Details of Wing for Beechcraft "Bonanza" Commercial Airplane.

vertical stl:fener CD. The vertical load at E would be transferred to stiffener SF through fitting at E. The problem is to find the shear tlows in the web panels, the stiffener loads and the beam flange loads.

Figs. A2l.9 and A2l.l0 show free bodies of that portion including web panels (1) and (2) and stiffeners CAD and AS and ~he external load at (A). In rig. A21.9 the shear flows q J, and q ,

10"-+-'0"--f-- 10"-r'O"i
F
D

9" q t

t~q,
(1)

s,
D

9"
Web

(I)

:/'ti!fener
/'

(3)

II
-, flan g e
Fig. A21. 8

- !3" fE

(2)

/H

r 1000 lb.

l~ 1000 lb. 6001 b.

~800

lb.

TW3" q

'f

(21

~800 'e IY
I

q':~'1I
qll~
B
Fig. A21. 10

--- D

s,

,"-x

(2) 1,600 lb. E---C q.

10"

---l

Fig. A21. 9

SOLUTION;
~langes

It Nill be ass~ned that the bea~ develop th8 en~~re resls~ance to bea~

on the top and bottom edges respective~y have been as.sumed 'filth the sense as ShC'NTl. Ta:-clr.g mcments about ~olnt E,

bending ~oments, thus shear flow is constant on a 'Neb pane L,

The shear flews on .veb pane Ls (I} and (2)
-1'/111 be

mE = 800 x 3 - 12 x lOq
20 lb./ln.
Zrx

J,

= 0,

whence q ~ whence qll

=
=

computed t.r-eat tng each c omccnerrt Gf the 1000 lb. load as acting separately and the results added to give the final shear :low.

= 800 - 20 x 10 -lOqa = 0,
50 Ib./in.

()1_~

:v"'0 "C

A21. "

AND STRESSES ON

RIBS AND

FRAMES

PHOTO. A21.1

Type of Wing Ribs Used in Cessna 180 Model Airplane, a " Place Commercial Airplane .

,
..........
~

..

..

'''#o>..-e~

'~'---

.~~

,. 'j ,
v

>
J.

PHOTO. A21. 2

Rib Type Used in Outer Panel-Fuel Tank Section- of Douglas DC-8 Commercial Jet Airliner.

,
, .. ,
;

.,.,. ,.
PHOTO. A21. 3 Rib Construction and Arrangement in High Speed, Swept Wing, Fighter Type of Aircraft. North American Aviation - Navy Fury - Jet Airplane.

"

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A21. 5

Reterring to Fig. A21.10, ZME

ZFx

Solving equations (1) and (2) gives,

q. = 70 lb./in., q , = 10 lb./in., which checks first solution.

= 600 x 10 -

12 x 10q .. = 0, whence q .. =
qa

50 lb. in.

= -50 x 10 ... lOqa = 0, whence

50 lb./in.

The shear flow qa in web panel (3) is obtained by considering stiftener EBF as a free body, see Fig. A21.16.
ZFY ::

Combining the two shear flows for the two loads,

q~

= 20

+ 50

=70
= 10

12q;s'" 10 x 3 - 9 x 70

Ib./in. Ib./in.

-1000

50

T
9"

Fig. A21.11 shows the results. Fig. A21.12 shows stiffener AB as a tree body, and Fig. A2l.l3 the axial load diagram on stiffener AB, ~hich comes directly from Fig. A2l.11 by starting at one end and adding the shear flows.

1
3"

q,

II t Iq
B

whence, qa :: 133.33 lb./in.

, ",70
,\0

Cl~ E

t:.

1000

The shear flow q~ could also be found by treating entire beam to right of section through panel (3). For this free body,

--------=1
q.. =70

BOO#( tension) Fig. A21. 13

Fig. A21.16

ZFy

=-600 - 1000 + 12q, =0

q~

whence,

= 133.33

= qa
= 10

-=::"A

Fig. A21. 12

Fig. A2l.17 shows diagram of axial load in stiffener EF as determined from Fig. A21.16 by starting at one end and adding up the forces to any section. After the web shear tlows have been detennlned i g' A2 1. 1 7 the axial loads in the beam flanges follow as B _ 570 lb. the algebraic sum of the shear flows. Fig. AZI.18 E 1000 lb. shows the shear flows (tension) along each beam flange as previously found. The upper and lower beam flange loads are indicated by the diagrams adjacent to each flange
F

Fig. A21.l4 shows a free body of the vertical stiffener CAD. and Fig. A21.l5 the axial load diagram for the stiffener.

Tq ,
9"
3"
I

'10 +_A qa 600

I

='0,
=10.

...L..

A~
C
~lows

~
;
/

630> (tension)

Fig. A21. 14

Fig. A21. 15 4700 Tension 700 lb.

q. and qa could of course be dete~ined USing both components of forces at (A) acting Simultaneously. For exa~ple, consider tree body in Fig. A21.ISa.

'rhe shear

133.33

--133.33 133.3=L

13333

70

J- -r -;
t-10"-1
q

~
Fig. A21. lSa

133.33

133.33

10

'1000 100* tension

_800 600

L CIa t'IB
x
E- -

&800<1
lb.

~ 600
C

-+ 3"
.....L...

3900* compr essron

q.

~ig'A2L18

<'

ZFx :: 800 - 10q. ZFy

(1 )
(2)

= -600

9q.

In this example problem the apolled external load at point (A) was acting in the plane at the beam web, thus eNO stiffeners were sufficient to take care of its two components. Often

A21. 6

LOADS AND STRESSES ON .RIBS AND

FRAMES

loads are applied Nhich have t~~ee recta~gular comDonents. In this case, the structure should be arranged so that line of action of applied force acts at inter5ect~on of two webs as illustrated in Fig. A2l.l9 where a load P is applied at point (0) and its co~ponents Pz , Py

Fig. A21. 20

tLx

s,
Fig. A21. 19

and Px are distributed to the web panels by us ing three stiffeners S J.' S a and S,3 intersecting at (0). In cases where a load must be applied normal to the Neb panel, the stiffener must be designed strong enough or tranSfer the load in bending to adjacent webs. In this chapter, the webs are assumed to resist pure shear along their bouncarles. In most practical thin web structures, the webs will buckle ~~der the compreSSive stresses due to shear stresses and thus produce tensile field stresses in addition to the shear stresses. The subject or tension field beams is discussed in detail in Volume II. In general the additional stresses due to tension fIeld action can be superimposed on those found tor the non-buckling case as explained in this chapter.

with q known the shear and bending ~cment at variouS sections along the rib can be dete~ ~ined. For example, consider the section at 9-3 in Fig. A21.20. Fig. A21.2l shows a free body of the portion rorNard of this section.

The bending
MB

~oment

at section 8-B equals:

shaded portion. Let Fx equal the horizontal component of the flange load at this section.

=2qA

1.

where A 1. is the area a f the

Fig. A21. 21

A21. 4 Rib for Single Cell 2 Flange Beam.

Fig. A2l.20 illustrates a rib in a 2flange single cell leading edge type of beam. Assume that the air-load on the trailing edge portion (not shown in the figure) produces a couple reaction p and a shear reaction R as shown. These loads are distributed to the cell walls by the rib which is fastened continuously to the cell walls. Let q = shear flow per inch on rib pertmeter which is necessary to hold rib in eqUilibrium under the given loads
P and R.

The true upper flange load Fu = Fx/cOS Q1. and the lower flange load equals FL Fx/cos Q~.

The vertical shear on the rib web at 3-8 equals the vertical component of the shear flow q minus the vertical components of the flange loadS. Hence

Taking ~aments about some paint such as (1) at all ~orces in the plane at the rib:
ZM1.

Illustrative Problem

= -Ph

ZAq

=a

hence
q

The rib in the leading edge port:on of the wing as illustrated in Fig. A2l.22 ~ill be analyzed. A distributed external load dS shown will be assumed

=Ph/2A.

(A = enclosed area or celli

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A21. 7

Solution: The total air load aft of beam = 8 x 4-0/2 = 160 lb. The a~. to its e.g. location fr8~ the fl b=a~ equals 40/3 = l3.33 Hence the react~ons at the beam flange paints due to the loads on the trailing eGge ~ortion equals:
P

w = 8/F/in.

's I

-rf30

= 160 x 13.33/10 = 213.2

lb.

(See Fig. A21.23 )

Iv:e:_l_
7.67"

Area of Shaded

Portion = 15.4 sq.tn.

Fig.A21. 24

Shear reaction Vr
w =

= 160

lb.

I Is

30
F

8 Ib./in.

A21. 5 Stresses in Rib for 3 Stringer Single Cell Beam.

-----10"

~===dll.. -""1--- 10" I

w '" 8 lb./in.

40"
Fig. A21. 22

Fig. A21.25 shryNS a rib that fits into a single cell beam with 3 stringers labeled (a), (b) and (c). An external load is applied at point (a) whose components are 5000 and 3000 Ibs. as shown. Additional reactions from a trailing ed~e rib are shown at points (b) and (c). A vertical stiffener ad is necessary to distribute the load of 5000 lb. at (a). The following values will be determined; (1)

Rib web shear loads on each side of stiffener ad.

Rib flange load at section ad.

1~

~! It!
A '" 139.3

t ttl f

;~ I

(2 )
p= 213. 2 lb.

~\11

'"' ~

Area

sq.
_

in.

~
"'=:: _ _

j I

(=L
r
10" 160

(3 )

Rib flange and web load at section just to lett of line be.
Fig.A21.25

> -20~ ~/i

(l)

P =213. 2 lb.
Fig. A21. 23

/' t./ d:------f1

I
Area:=160"'A2

I
5O I

9.3"

t 1~2 ~qcb

1"

,/.r

Let q be the cor.stant flew reaction of the cell skin on the rib perimeter which is necessary to hold the rib in equilibrium under the applied air loads. Take moments about some point such as the lower flange (1).
Z!1

web

~"=,=:_~-lJ...=;_",,,,;e="'_="'~b

at

3000 lb.

qba

':1.- 500

J/-

5000

I---

15'

--1

=0

= -213.2xIO+8x15x7.5+2x139.3

whence, q = 1232/278.6 = 4.42 lb./in.

~ith the applied forces on the rib known, the shears and bending moments at various sections as desired can be calculated. For example, consider a section B-B, 2.5" from the leading edge. Fig. A21.24.

SOLUTION. It will be assumed that the 3 stringers develop the entire Wing beam bendlr~ resistance, thus the wing shear flow is constant between the stringers. The wing rib Is riveted to the wing skin and thus the edge forces on the rib boundary will be assumed to be the same as the shear flow distribution. In othar wordS, the three shear flows Qadc' qba and qcb hold the external loads in equilibrium. The sense of these 3 unknown shear flows will be assumed as shown in Fig. A21.25.

Bending moment at section B-B

8x2.5x1.25+4.42x2xI5.4

=
161 Inv Ib ,

To find Gacc' take moments about point (b)

A21. 8

LOADS AND STRESSES ON RIBS AND FRAMES

ZMe = -2(A1.
x 11.5

A.) Qadc .;. 5000 x 15 - 500

whence, assumed.

=-2(60 + 160) QadC QadC = 157.3 Ib./in.

= 5000
<lob
+

=0

75000 - 5750

=0

215B-Ib-t
d T'

~d

To find web shear take ZFz = 0

= 5000 - 9.3 x 74.6

- 9.3 q~d

74.611,

with sense as

I I"ad

9.3"

2034.....J i.-

To tind qcb take ZFz = 0 ZFz 200 - 157.3 x 11.5 - 11.5

'C 5000

c,l 3000

=0
1b./ in.

whence, %d

=463

=0

To find r Lange load C' take ZFx 0, conSidering joint (a) as a tree body,

Fig.A21. 27

whence, Qeb = 295 lb./in.

ZFx

=2034+3000-C! = 0,

whence, C!

= 5034

lb.

To tlnd qba take zrx

=0

ZFx = -500 + 3000 + 500 - 157.3 x 15 -15 qba 0

At joint (d) T' obviously equals 2158 lb. The sti:!ener ad carries a compressive load of 5000 lb. at its (a) end and decreases unifOrmly by the amount equal to the two shear flows or 463 + 74.6 = 537.6 lb./ln.

whence, qba

=42.7

lb./in.

With these supporting skin torces on the rib boundary, the rib is now in equilibrium and thus the web shears and tlange loads can be determined. Consider as a tree body that portion or the rib Just to the lett at the stiffener ad centerline as shown in Fig. A2l.26.
T

The results obtained by conSidering Fig. AZl.27 could also be obtained by treating the entire rib portion to the lett ot a section just to right ot stiffener ad, as showr. in Fig. A21.28.
To find rib flange load T' take moments about point (a).

ZMa
whence,

= (16/17) T' x9.3-157.3x60x2 T' = 2158 lb.

=0

I~T'
I'
To tind tlange load T take moments about point (a),

~""';=="'t3000

l~

9.3'"
C'

5000 Fig. A21. 28

ZMa

= (16/17)T

x 9.3 - 157.3 x 60 x 2 = 0

whence, T

=2158

lb.

To tind flange load C' take ZFx = 0

ZFx

To tind flange load C take ZFx = 0 2158 (16/17) - C 0, whence C 2034 lb.

To tind wab shear Qad take ZFz = 0 ZFz = 2158 (6/17) - 157.3 x 9.3 + 9.3 Qad = 0 whence, Qad 74.6.1b./in.

=-e' +3000 + 2158(16/17) whence, C' = 5034 lb. To rind %d take ZFz = 0
ZFx ZFz whence,

=0

= 2158(6/17)
%d
~d

- 157.3 x 9.3 + 5000 - 9.3

=0
Ib./in.

T6 tind the shear in the web just to right of stiffener ad, consider the tree body tormed by cutting through the rib on each side at the stiftener attachment line as shown in Fig. AZl.27. The torces as found above are shown on this tree body.

=463

The above values are the same as previously obtained. The rib flange loads and web shear will be calculated for a section just to lett of line

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A21 9

cb. Fig. AZl.29 shows the tree body for the rib to left ot this section.
8500
t

6000

Fig. A21.30
skin

6'1'50
7500

~t ]- Fg~~
Web
30"

t- ~:.~.
b

L.
Fig.A21. 29

iI

CQ

skin

8.40" I .....i...-4500#

5000

2~0

SOL1JrION:
To tind tlange load C take moments about
pOint (b).
l:IIb

=-157.3x2 (160+60) + 5000 x 15 -11.5 C = 0

= 500
lb.

-9000 lb. and Vx --500 lb.

The total shear load on the wing in the Z direction equals Vz -6000 - 5000 + 2000 =

= =-8500 + 7500 - 4000 + 4500 =

whence, C

To tind tlange load T take ZFx =

ZFx

The boundary torces on the rib will be equal to the shear flow torce system on the cell walls due to the given external torce system. From Chapter Al4, page Al4.8, equation (14) the expreSSion tor shear flow is,
qy

= 3000-157.3x15-15x42.7-500+T = 0
=
0
q~b

whence, T = 500 lb.

To find ~b take ZFz IT z

-(K15Vx-K1.Vz) .r.xA- (K.Vz -KJ, 11x)

ZZA - - - - - - - - - - - - - - - (1)

=5000 -157.3 x 11.5 -11.5

= 278
lb./in.

=0

whence, qbb

The ~bove results could have been obtained with less numerical work by considering the forces to right of section cb in Fig. AZI.29.
A21. 6 Stress Analysts of Rib for Single Cell Multiple Stringer Wing.

The constants K depend on the section properties at the wing cross-section. Table A2l.l gives the calculation of the moment ot inertia and product ot inertia about centraidal Z and X. axes. In this example the 4 stringers a, b, c and d have been considered as the entire effective material in resisting wing bending stresses.
1

When there are more than three spanwise stringers in a wing, there are rour or more panels in the cell walls, thus the reactions at the cell walls upon the rib boundary cannot be round by statics as was possible in the 3 stringer case of the previous example problem. Fig. AZl.30 illustrates a wing section consisting at tour spanwise flange members. The concentrated loads acting at the four corners of the box might be representative or reactions tram the engine mount or nacelle structure and the reactions tram a rib which supports the wing rlap. These loads must be distributed into the walls or the wing box beam which necessitates a rib. Serare the rib can be designed, the bending and shear forces on the rib must be determined. The calculations which follow illustrate a method at procedure.

,
Am A

3 Z'

, , ,
X'
AZO AZO'

TABLE A.21.1

10

11

F1an<e
No.

AX' AX"

a
b

'.00
I. '"

5.50

U.OO 4.43

60.5

AX' " 0 0
132.9

Z.ZO- Z XaX:-X

6.38 4. oil

-U.8 18. :z 18. :z U.8

3.55 30

15.72 37.50 U25

e
d

1.15 -8.40 30 - 9.68 1. 70 8.50 8.10

82.20 34. SO 1035 289.8 7.54 71. 90 0 0 0 5.64

o -U.04

- 5.2'7 230.3

72.00 2160 -1511.9

=ZAZ'/ZA =-5.27/6.10 = _.865 X =ZAX' /ZA = 72.0/6.10 = 11.8'

z
Ix
Iz Ixz

ft

Centraidal X and z moments of inertia:

=230.3
= 2160

- 6.10 x .865- 6.10 x 11.8'

= 225.8

= -156.9

= 1310

- 610 x -.865 x 11.8

=-94.7

A21. 10

LOADS

AND STRESSES ON RIBS

~no'Nn,

AND

FRAMES

~lth the Ning section ~ro~ert1es the constp-nts K can be calc~lated.

Z1'!r. ,:>. _
~

= -11000 x 11.8 -

3500 x 6 .36 - 7500 x

~x =-6~84CO

5.65 - ';'000 x 4.41- 4500 x ? .54 - 2000 x 18.2 2x90-74Bx2xlC8.7-461x2xG9

KJ. = Ixz/(Ixl z - Ix(2) = -94.7/(225.8x1310-94.7') =

1:1.10.

-94.7/286700

-.00033

Ke = Iz/286700 =
K,

1310/256~00

= .00456

Fer equ11ibrlum Z~.g. must equal zero, therefore a constant ~low shear qJ. acting around the rib perimeter t s necessary wn i ch wtLl. er-educe a moment of 648400 in.lb.
q, =
~ 2; ....

= l x/286700

= 225.8/286700

= .000786

SUbstituting in equation (1),

qy = - G000786 (-500) - (-.00033) (-90008

25~"'_E",.~,,_ 0",0,= =~ = 2x368.5

880 Ib./:n.

(Note:

368.5 = total area of cell)

ZxA - E00456 (-9000) - (-.00033)(-500) ZzA

whence,
qy

=3.363

ZxA

+ 41.205

ZzA - - - - - (2)

Adding this shear flow ~o that of Fig. A21.31, ~~e resulting force sj~tem at F:g. A21.32 1s obtained. The reactions of the beam cell walls on the rib have now been cete~ined and the bending mo~ents and shears on the rib can new be calcul~ted.

Since the shear tlow at any point on the cell walls is unknown, it will be assumed zero on web ad, or imagine the web is cut as shown in Fig. A21.31. The static shear flows can now be round.
'lab

8509..-

Ia
il t =
7500

6000 __

q-=-.i3~
Fig.A21. 32

= 3.363
x 2

(-11.8)(2.0) + 41.205 x5.36

=444

lb./in.

qbc

~d

= 444 + 3.363 x 18.2 x 1.25 + 41.205 x 4.41 x 1.Z5 = 748 Ib ./in. =748+3.363xI8.ZxL15+41.205 (-7.54)(1.15) =461 Ib./ln.

I '" C -\I=:'-'-=;-=q"'."'.'"',"'9=_=,"",=-_=," i'- _ 4500

Id
5000 Fig. A21. 32 2000

~1

"l,

---i9 00

t:

To

illustrate, consider the rib section

B-B which passes through the e.g. or the beam

section. ~lg. A21.33 shows a ~ree body of the bulkhead portion to the left ot section 8-B.

These shear tlows are plotted on Fig. A21.31. Refer to Chapter A14 regarding sense of shear flows.

rooo
8500' a
6.36

Fig. A21. 31

-9..=444

5.~4 7500 d

'= - ro. -F -p-5000

'"""~r

Area=900..

"~

~I

- -

_b
.-

"00

~~ ea~.~

Ax --".0,

, ' ,

,~ ~ea= lOB. 7

I~ 4.4,
t; r

~~~e~ .. 99 ', '~ t~ -I~q ;;"4BUb. /...- - tC . .00

11. B
I

'/'//)"/;-...,..
lB.2

: ~ 754
2000

I--

---1

Fig. A21. 33

Fig. A21. 34

Total Cell
Area", 368. 5"

Moments at section B-B will be referred to

the point (0);

The moments ot the forces in the plane of the rib will now be calculated: Taking moments about the e.g. at the beam cross section (See Fig. AZl.31):

ZMo = -11000 x

U.8 - 3500 x 5.36 - 7500 x 5 .64 + 436

x 2 x 36 + 880 x 2 x 70.8 + 419 x 2 x 38.3

-38200 1n.lb.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A21. 11

The resultant external shear force along the section B-B equals the summation of the z components of all the forces.
'f

= ZFz = -11000 + 12 x 3eO - 436 x 0.36 + 419

x 0.96 = -195 lb.

The resultant load normal to the section 8-8 equals the summation of the force components in the x direct1on.
H

the boundary forces on a rib located adjacent to a c~t-out is to find the applied shear flows in the wing on two sections, one on each side of the rib. Then the algebraic sum of these two shear flows will give the rib boundary forces. Aith the boundary forces known the rib web and flange stresses can be found as previOUSly illustrated. The procedure can best be illustrated by example problems.
A21. 8 Example Problem. Wing with Cut-Out Subjected to Torsion.

ZFx = -8500 + 7500 + (436 - 419) 11.8 = -800 lb.

Fig. A21.34 shows these resultant forces referred to Doint (0) at the cross-section. If we assume that the rib flanges develoD the entire resistance to normal sta-esses., we can find flange ~oads by simple statics. To find upper flange load Fu take moments about lower flange point.
zrI

Fig. A21.35 shows a rectangular Single cell wing be~ with tour stringers or flanges located at the four corners. The upper surface skin is discontinued in the center bay (2). The wing is subjected to a torsional moment ot

'i: - T 4Q
c,
~

STA.O
'I

'I I IRib,lF'
,

_l~a

,,
AI

"000,,

= 12.6
~u

Fu-38200-800x6.6

=0

whence,

= 3443 lb. tension

To find FL ase ZF x - 0
Z~X

STA. 50..1

/ ::
STA.

21)0/

1'7/
J,

= 3443 - 800 - FL = 0, whence FL =

2643 lb. compression.

The shear flew on web equals V/l2.6 = 195/12.6 = 15.5 lb./in. This result neglects effect of fl~~ges not being no~l to section 8-8, which inclination is negligible in this case. the entire cross-section of rib is ef!ective in bending, then the web thickness and flange sizes of the rib would be needed to obtain the section ~oment of inertia which Is necessary in the beam equation for bending stresses. The forces at (0) would then be referred to neutral axis of section before bending and shear 3tresses on the rib section could be calculated.
I~

,q ST~O

;::.~ ~

"

4.~o

Rib

jJ'

~~~;:c:~moved in

-----.1-:1--4000
---1./,1 . I

lb. Bay (2)

r Yr--' ;,
(Rib

' II,I! I !

Rib (A)

::;1

~o in. lb.
40" ----i

Fig.A21.35

I\

80000 in.lb. at Station (70) and a couple force at Station (50) as shown in Fig. A2l.35. The problem will be to determine the applied torces
on rib (A).

SOLUTION: The applied shear flow on the cell walls will be found for two cross-sections of the wing, one on each side at r tb (A). In bay (1) the torsional moment M is 80000 in.lb. The applied shear flew en a crosssection of t~e wing in bay (1) thus equals,
q

To obtain a complete ~icture of the web and flange forces, seve~a1 sections along the ~ib span should be analyzed as illustrated for section B-B.
A21. 7 Rib Loads Due to Discontinuities in Wing Skin Covering.

= ul'! = 2 x30000 = 100 10 x 4,0

lb .ym

As referred to before, rib~ i~ acdition to transmitting external loads to wing cell structure are also a means of re-distrlbuting the shear forces at a ~iscontinuity, the ~ost comnon discontinuity being a cut-out in or-e ar acr-e of tne Nebs or 'Nalls of the .vt ng beam cross section. ~he usual ~rocedure in f1nding

This shear flow system is shown on Fig .. ..;21.36 wntc a is a rree body of rib (A). In bay (Z), since the top skin is removed. the tcrsional moment must be taken by the tront and rear vertical webs, since any shear flow in the bottom skin could ~ot be balanced. The torsional ~oment in bay (2) is,

c, . -"..:-"<.. C

A21. 12

LOADS AND STRESSES ON RIBS AND

FRAMES
Bay (2)

q:: 100 '1 t I

l ,~~-_ ..9. ':....lQ..o _ -"'1

q:: 10

t~;;:::;;;:::;~~~;;::;;;1lt-4000#
t
q':300
lUGl + J -4000Jil

!,1.q'",300

..."..Rib A

h .110"

Bay II)

---rl----- L. 30"-----!--_ Bay (3) Fig.a P-.i~;::='=,=,=,=,=,=,===~ p Vl3000

~:t=;:~~===:=]lrv:iooo
I
_p

Fig. A21. 36

VL

M:: T

I ~ --------===::J

e------

Fig. b

Beam Bending Moment Diagram

_

11

80000 + 4000 X 10

120000 tn.lb.

V.3000!.......

The total shear load an each vertical web thus equals 120000/40 = 3COO lb., which ~ives a shear flow q' = 3000/10 = 300 Ib./in. on each web. ThiS applied shear is shown on the tree body of rib (A) in Fig. A21.36. On the left end of the rib a shear flow of 100 is acting
Fig. A21. 37
q:: 100

Beam Shear Diagram

Fig. c

q:: 300

--- -

- - -:1
Fig.d

flange loads at the bea~ ends are P = 45000/10 :. 4500 lb. (See Fig. a). The deflection of the rear bea~ would be the reverse of ~ig. a. and thus all forces would also be r~versed.

up and on the other side a shear flow of 300 Is acting down, thus the rib web must take the difference or 200 acting down. On the right end of the rib the load on the rib web is 200 Ib./ in. up. The loads on the top and bottom flanges at the rib Is obViously 100 Ib./in. Fig. A2l.37 shows the loads applied to the rib boundary when the torsion in bay (1) and the external couple force is transferred to the cross-section at bay (2).
ADDITIONAL EFFECTS DUE TO DIFFERENTIAL BENDING OF BEAMS IN BAY (2).

Fig. A21.39 shows bay (1) 0:: the wf ng as a free bedy acted upon by the :lange loads due to bend:ng of the beams in bay (2). These interr.al flan~e forces from bay (2) must be held in equilibrium oy ~he internal stresses ~n the adjacent wl~g structure of bay (1).

I:!

~
;-Rib (Ai

". / :i

..

'-'lr@

The torsion In bay (1) and the external couple force are thrown off as couple torce on the front and rear beams of middle bay (2), with the total shear load on each beam being Fig. A21. 39 3000 lb. as previously calculated. These beam shear loads must be transmitted to bay (3) and thus cause bending of the bea~s in bay (2). According to the well known ~rinciple of Since each beam is attached to relative rigid mechanics formulated by Saint Ve~~at, the box structures at each end, namely bays (1) stresses resulting from such an internal force and (3), the beams tend to bend with no rotation system will be negligible at a 1istance from of their ends. If we neglect the deflections the ror-ces . ~hls distance in case of a cut-out of these end box structures, we can assume that is usually assumed as approximately equal to the beams bend with no rotation of their ends ~he width at the cut-out, or in g~neral ~o ~he or each beam is fixed ended. Fig. A2l.38a width of the adjacent wing bay. Thus in Fig. t LLus'tr-a'tee the deflection of the front beam in A2l.39 the flange loads 0: 4500 lbs. each are bay (2) under the assumption of no end rotation. assumed to be dissipated at a unifJr.TI rata for The beam elastic curve has a point of inflection a distance of 20 inches. Thus the shear flow at the span midpoint. Figs. 38b, c show the created by each stringer load which equals the beams bending moment and shear diagrams. change in axial load per inch in the strip~er In bay (II equals 4500/20 = 225 lb. The end moments are M = VL/2 = 3000 x 30/2 = 45000 in.lb. Assuming the b~a~ flanges Fig. A21.40 ShONS a segment 1 inch Nide develop the entire bending resistance the beam ~ut trom wing bay (1) with the ~p load in each

'-.",,~,'''''

-,~.

<-

-=
A2L 13-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

:lange member. ~o !ind the shear ~lcw on ~~e cross-section the frent web is tirst assumed cut, and thus the static shear floW qs = Z6P from cut face where q~ is zero. Fig. A21.40 shows th1s static shear flow.

bay (1) as shown. The areas of corner stringers a, 0, c and dare shown in ( ) adjacent to each stringer.
B800 lb. No Skin on
Bottom of Bay (2).

r-- 15,,--t
Rib

Fig. A21. 40

~O

T
For equilibrium of the cross-section, the moment of the forces in the ;lane of the crosssection must equal zero. Taking moments about lower left hand corner of the qs force ~ystem,
i1 = 225x40xlO = 90000 tnv Ib , For equilibrium a moment of -90000 1s necessary .. Therefore a constant shear flow system q must be added to develop a moment of -90000. Thus

1"

M!2A

(-90000/2 x 10 x 40) -112.5 lb./ln.

12"
Rib

Adding this shear flow to that for qs in Fig. A21.40 gives the final values in Fig. A21.41. This shear flow system represents the stress
112.5
='" - -

Fig. A21. 43

I
Fig. A21. 41

30"

The mi"dle bay (2) has no skin on the

112. 5

system caused on c~oss-sectlon of bay (1) due to the differential bending of the beams in bay (2). This shear flow system must therefore be resisted by rib (A) as it ~ust terminate at end ot bay (l). Therefore the shear flows in Fig. A21.41 are applied bounda~y loads to rib (A) and these T,Ust be added to the rib loads in Fig. A21.37 to give the final rib loads of F~g. A21.42. ~lth the final rib loads
212. 5 312.5 Fig. A21. 42

surface, or in other words, the middle bay has a channel cross-section, which fact often happens in practical ~ing design as for exa~ple a space or well tor a retractable landing gear. The problem will be to find the shear flow in bays (1) and (2) and the boundary Loads on rib (A) between bays (1) and (2).
botto~

Solution No.1 This method of solution will make use of the shear center location for bay (2) in order to obtain the true torsional moment on bay (2). ~ith this torsional moment known, the ~rocedure is similar to the previous example involVing win~ torsion only.
'.-'e wf Ll. first c~lc~late the shear flow in Wing bay (1). ~ig. A21.44 shows the c~osssection.

tl

- - -I'

-4000*

____ 212.5

t312.5

'_4000*

known, the rib fl~nge and web stres5es can be found as previously explained.
AZl. 9 Example Problem. Wing with Cut-Out Subjected to Bending and Torsional Loads.
~ . . g. A a pertl on 0 f -. 21 .~ " s ..ows "

L.
I

ho:-,-'5" 1
~

-,6"

:-t

aBOO/!'

(0.5)
b 4,,1

3.

single cell cantilever beam composed of 3 bays ~o~ed by the four ribs. The loads on the structure consist of loads applied to end of

2400 4 str 1nger-1

3" d * ...L:.(0.5 -r- 6" TO l!--~------""ro:sr _ I

~10"
~x-~

4'"

30"

----i

Fig. A2l. 44

'--"--I

o ,' . . /

A21. 14

LOADS

AND

STRESSES ON
~2eded

RIBS

AND

FRAMES

The section moments of in calculating shear flows.

Ix == (j x s e x z )
+

i~=rtia

are

we ::~s~~~r jay (2) in l~s ~:~e bot:om skin ~~movec. r1;. A21.?7 shows t~~ cross-section of bay (2).
c8~d~tlon w1t~

~herefore

(O.5x";''i x2) = ae 1:1:

tn ,
In~

X = 'i:ki:/ZA

= (1 x30)/3 = 10

I- 15"---1

8800

Iz::l (2xIO Il ) + (lx20 1l ) = 600

Vz
::I

~-.._~"'=-:,~
c
Fig. A2l. 47

2400 12"H 'lad

8800 lb.,

Vx == 2400 lb.

3::r-ll ~ "d

qy:ll -~

ZzA -*ZxA , substituting

(a)

qy = - 100 l:zA - 4 l:xA

':'he 'thr-ee shear- :'l'J',':sC3.r. be cc t cr-atncc by statics. !F x ::: 2400 - 30 Qab = 0, whence
z:-''(J.

Since the shear flow is unkno~n at any paint on cell, we will assume frent web (ad) as cut or carrying zero shear.
~c

ceo

= SO

2400 x 3 - 3600 x 15 - SO x 30 x 12 + (8x30)

qbc

= -100 (-6)(1) -4 (-10)(1) 640 lb./in

0, whence Gbc 640 12

Qcb qba

640 -100 (-4) (0.5) - 4 (20) 0.5 (4)(0.5) -4 (20) 0.5

600

ZFz

= 8300 - 9 x 640 + 2 x 50 whence Qad

:=

Sac

== C

= 800-100
Ia-

= 560
(2)

320

Fig. A21.45 shows these static shear flows.

15'---,
.~8800

Fig. A21.48 shows the ~esults. ~~1s shear :low system 1s t~e final or tr~e shear on bay

J.Q ~i'::.-

-toJ.l:: !~._

nlr
d

bl'll
c

640

----640Fig. A21. 45

Fig. A2l. 48

To this shear flow, a constant shear flow must be added to make Ll1 = O. Take moments about point (d).

ZMa

= -8800 x 15

+ 2400 x 3 + 560 x 30 x 12 + 800

Since we have a. channel or ooen 'Ni::g C:"05Ssection in bay (2), any torsional moment on this bay :nust be transmitted by ji::erent~al bendin~ of the front and rear beams. ~o obtain the torsional :noment an bay (2), the shear center location must; be xncwn ,
~orizontal location of shear cent3r: Assume the section bends about centroidal X axis without twist under a Vz load of 8800 lb.

x a x 30 = 268800 in. lb., or -266800 is reqUired for eqUilibrium, hence the reqUired constant shear flow q = -M/2A = -268800/2 x 300 = -448. Adding this shear flow to that of Fig. A21.45, we obtain the shear flow of Fig. A2l.46.

= - 1; Vz

ZzA , or q

=_

100 ZzA

Qcb Qta 'lad

192
Fig. A2l. 46

=- 100
= 200

(-4)(0.5)

= 200
=0

100 (4)(0.5)

=0

- 100 (6)(1)

= -EOO

This shear flow 5yste~ would J8 the shear flow system for all 3 bays if the bottom skin in bay (2) was not re~oved. 2e~ovlng the bottcm skin in bay (2) ',vill :nodif':l tneee shear flows of
:<'i -. ,1.21.+6.

Fig. A21.49 shows the shear flow results !or bending about x-x wit~out twist. 7he line of action of the resultar.t of :his shear flow force syste~ locates the horizontal ~csitlcn of t~e shear center.

X ..

(200x8x30)/8800

= 5.45 in.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A21. 15

600

l~~.~
+d R::8800 Fig. A21. 49

ential bending at beams in bay (2), thus equal the beam shear times halt the span at bay (2) diVided by the beam depth.
For tront beam P

= 3841 x 12.5/12 = 4000

lb.

For rear beam P = 3841 x 12.518 Vertical position of shear center: Assume section bends about centroidal z axis without twist under a load of Vx = 2400 lb.
q

= 6000

Lb ,

Fig. A21.52 shows these.flange loads applied to bay (1). These loads are dissipated

= _.'!x. ZXA =_ 4 I
z

ZXA

'<cb. -4x20xO.5

=-

7:--1 _

40 lb./ln.

qba:: -40-4x20xO.5 = - 80
<lad

..L
~

= -80 - 4

(-10) 1

=-40

Fig. A21.50 shows the shear flow results.

1/ "

/
Rib A

IIC!L--~

1'1.

Bay (1)
Fig. All. 52

c
Fig. A21. 50

unifOrmly 1n bay (1) over a distance of 30 inches, or the shear flow per inch prOduced by these flange loads equals tJ' :: P/30, whence
<lPa

The vertical distance Z trom point (a) to the line of action of the resultant which locates tne vertical location ot shear center IS,

= <lPd = 4000/30 = 133.3 and = 6000/30 = 200 lb.

<lPo

= <lPc

Z = l:!1a/2400 = (40x8x30)/2400 = 4 In.

Fig. AZl.5l shows the shear center location and the external loads. The moment about the shear center which equals the torsion on the wing bay (2) equals,
Ms c

Fig. A21.53 shows an element or bay (1) ana inch wide with these dP loads. The shear flow q assuming the front web cut equals Zt.P. The resulting static shear flows which equals
Z<lP Is shown In Fig. A21.53.

8800 x 9.55 - 2400 x 13 = -115240 In. lb.

e.c.
8800

Fig. A21. 53

2400

.-::r-l i,.9..,,~ ~{Hk'"

lb
c

Fig.A21. 51

Thi~ torsional moment must be resisted by front and rear beams. Hence shear load on

The moment of this shear flow system about = 133.33x30x12-66.7x8x30 = 31980. For ZM = 0, we need a constant shear floW q - 31980/2 x 300 = -53.3 lb./ln. Adding this constant shear flow to that or Fig. AZl.53 gives the shear flow system at FIg. A21.54. These results represent the effect on bay (1)
point (d)

each beam

= 115240/30 =3841

lb.

As in the previous example problem involving torSion, the beams in bay (2) will be assumed to bend without rotation ot their ends, or in other words the bending moment at midpoint of bay 1s zero. The flange loads at paints a, b, c ~~d d on bay (1) from the d1fter-

I
53.3.

--~

80

Fig. A21.54

'0

f 120

'01.3 j

Fig.A21.55

-- , ..- 272192

232

A21.16

LOADS

AND STRESSES ON RIBS AND

FRAMES

of removing the botto~ skin tn bay (2). Adding the shear flo~s of Fi~. A21.5~ to ~h1se c:' 5"1 ~. A21'~J ~e obtain :~= final 3~ea~ ~lcws in ':lay (1) as sho~n in fiS. A2l.55.
aOmJD.;;RY LOi..OS 0;' 813 (A)
T~e boundary leads on rib (A) ~ill equal the ~1f:erence b0tween the shear flows in ~ays (1) and (2). FiS. A21.56 shows ~ free Jody af rib (A) with the shears :lcws obtai~ed f~am figs. A21.55 and ;"21.48.

192 with sense ~s shown. Since this s~lrr is nt satng e reverse this shear' r lc.v and rmc tne reSisting shear flows )n t~e ~ther ~~:se siees of tr:e bay crOss-sectiJn. Fi~. ~21.5S shows the sec t i on, with the 3 unknown shear ~1:J'I'./s Qab, qbc and Qa1'

1\_ '1a.2-" 1;': _

To r Ind qat use L:F x = 0, 192 x 30 - 30 c-a::,

=0

whence,

~b

= 192
-30 x 192 x 12 + 8 Qbc x 30 = 0

llt:t =
whence, qbc

= 288
x 192 - 8 x 288 + 12 'lao = 128

3F z = 4

=0

whence, The resulting applied boundary forces to the rib equal the alGebraic sum of ~he shear flows on each side of the rib which ~ives the values in Fig. A2l.57.

Gad

AddIng the shear flows of Fig. A21.58 to those of Fig. A21.46 sives the final shear flows in bay (2i as shown in Fig. A21.59. ~r.ese re-

181.3r~~~
272

272

I: 408

-80
320 ;LI_ _- - ; ; - - - ] : 640

Fig. A21. 59

Fig. A21. 57

,
With the rib boundarJ loads known, the stresses in the rib can be found as previously illustrated in this chapter. Solution No.2 This method of solution first finds the shear flow in all bays assuming bottcm skin is not removed in center bay (2). This gives a shear flow in the bottom skin. However, the skin in bay (8) is actually removed so a corrective set at shear flows on bay ,(2) along th~ boundary lines at the bottom skin ~ust be applied to eliminate the shear flows found in the bottom skin. The problem then consIsts at finding the influence at these corrective shear flows upon the shear flows as found for bays (1) and (2) when bcttom skin in bay (2) 'NaS not reJloved. The f~rst step is to find the shear flows in all bays assumi~~ bottom skin in bay (2) is not removed. The calc~lations would be exactly like those in solution (1) and the shear ~low in all bays would be those in ?Ig. A2l.46. The bottom skin in Fig. A2l.46 has a shear flow of

sults check the results in Fig. A21.48 obtained in solution method (1).

I;

l'
192

Fig. A21. 60

-------

?1~. A21.60 shows the corrective shear flows of Fig. A21.58 applied :c bay (2). On ~~e bottom sAin the correc~ive shear :lcw is sho~TI on the bouncary of the cut-ou~. These shear

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A21. 17

flows cause di~~erential bending of the front and rear beams in bay (2). If we make the ass~~~tion that the beam end sutfer no rotation, the be~ding ~oment is zero at niQPOint of the bay and thus th~ fl~nge loads at ~oints a, b, c and d of bay (1) equal the algebraic s~ of the shear ~lows on each side of a flange times half the span of bay (2) or 12.5 inches. Thus :rQ~ FiS. A21.60,

Pa

= (192 + 128)12.5 Pb = (288 + 192)12.5 Pc = (288 + 192)12.5 Pd = (128 + 192)12.5

= 4000 = 6000 = 6000 = 4000

lb. compression lb. tenRion

Frame in Center Portion of Fuselage for "Vought" FaU Airplane

PHOTO.A21.4

compression tension

Referring to Fig. A21.52, we find that the P values above are the same as the P values obtained by solution (1). Thus the remainder of solution (2) Nould be identical to that in solution u i, and therefore the Calculations wi:l not be repeated here.
A21. 10 Fuselage Frames

Frames in a fuselage serve the same purpose as ribs in Wing structures. Ribs are usually of beam or truss construction and can be stress analyzed fairly accurately by statics. Fuselage !ra~es however, are at the closed ring type of structure and are therefore static ~lly indeterminate relative to internal stresses. Once the applied loadS on a frame are known the internal stresses can be found by the application of the elastic theorJ as covered in Chapters A8, A9, AlO and All. The loads on fuselage frames due to discontinuities in the fuselage structure, 3uch as those due to windows and doors, can be approximately determined by the procedures preViously presented for wing rros . The photographs on page 32 of Chapter Al5 show some of the frame construction of the Douglas DC-8 airliner. Other pictures of fuselage construction are given ln Chapter A20. Photographs A21.4 and 5 illustrate typical fra~e construction and arrangement.
A21. 11 Supporting Boundary Forces on

PHOTO. A21. 5 Fuselage-Wing Portion of "Martin" 404 Transport A21. 12 Calculation of Frame Boundary Supporting Forces.
Ex~ple

Problem 1

a circular fuselage.

Fuselage Frames.

'When ext$rnal concentrated loads are applied to a ~~sela~e fra~e through a suitable fitting or connection, the frame is held in equilibrium by reacting fuselage skin forces Nhich are usually transferred to the frame boundary by rivets which fasten fusela~e skin to fr~~e. Since the fuselage shell is usually stress analyzed by the beam theory, it is therefore consistent to deter.nine the distribution of the supporting skin forces by the same theory.

Fig. A21.61 illustrates a cross-section of Two concentrated loads of 2000 lb. each are applied to the fuselage frame at the polnts lndicated. The problem is to determine the reactlng shear flow forces in the fUselage skin whlch will balance the two externally apolled loads. This tusela~e section ~1ght be considered as ~he aft portion of a ~edium size fUsela~e and the loads are due to air loads on ~he horizontal tail surfaces. To make the numerical calculations short the fuselage stringer arrangement has been assumed
symmetrr t ca L,

Solution: In this solution the fuselage skin resisting forces will be assumed to vary according to the general beam theory. The general flexural shear flow equation ~or bending about the Y axis is,

A21. 18

LOADS AND STRESSES ON

RIBS AND FRAMES

=-?Y

= -31.82 ZzA ,

6 .275 x .30 x ::"3.5 = -57.22

where Vz = 4000 lb.

:=

-57.22 -5.275 x .loS x 10

= -86.62

The moment of inertia Iy of the fuselage cross-section is required. In this 5 tmoj i r i ed illustration, the area ot each stringer plus its effective skin will be taken as .15 sq. In. The student should of course realize after stUdying Chapters Ale and A20 that the true effective area should be used on the compr'esa tve side and that the skin on the tension side of the fuselage 1s entirely effective. These facts would tend to make the effective c~oss-sectlon unsym~etrica1 about the Y axis. Since the only purpose ot this illustrative solution is to show how the trame loads are oalanced, the section being assumed as ~etrica1 which will greatly decrease the amount of calculations required.

.-.

= -66.62-6.2?5x.l.Sx.5 = -71.32

Due to symmetry of effective c~css-5ectlon, the shear flow is s~~etrical about the Y axis. As a check on the above Nor~, the s:~a tian of the z corn~onents of the shear flew on each skin Danel between the stringers shaule equal the external load of 4000 lb.

flow q on that panel, or

ZF of skin qhear flow equals the vertical prOject~d length of each ~anel times the shear
ZFz:=li G-.4x16.S7+2.7x31.82+3.5X57.22
+ 5xSS.62+3x71.32] 4

Double angles at

3,3',9,9'.

200~O

11.5

,e-i
1

2000 11. 5 ' J

4000 lb.
(check)

I'

bulb angle stringers

Fig. A2l.62 shows the frame with its balanced load system. The internal stresses can' now be found ~y the methods of Chapters AS to All.
2000 2000

t
6' Y

Fig. A2l. 61

11

11'

Moment of inertia of fuselage section about Y axis which is the neutral axis under our slmplltled assumptions.
Iy

.15 (17 .S- + 1S.2- + 13.5- + 13.5- + 10- +

5-)4': 637 in,'
Fig, A2l. 62

Due to symmetry ot effective section and external loading, the shear flow in the fuselage skin on the z axis or between stringers 1 and 1 and lIar 11 wtll be zero. Thus starting with stringer (1) the shear flow in the skIn reststing the external loads of 4000 lb. can be written around the circumference at the section.
q

Example Problem 2.

Uns~etr1cal

Vertical Loading

.-. =

Vz ZzA Iy

=_ 4000 ZzA 637

=_

6.275 ZzA

-S.275x .rs x iv.s = -lS.S7 Ib./in.

q .-, = -16.S7-6.275x .lSxl6.2:= -31.82

In certain conditions in flying and landing, uns)m~etrical concentrated loads are a~?lied to the fuselage or r.ull structure. For ex~~ple, Fig. A21.63 shows the same section an~ fr&ne as was used in Problem 1. Due to an uns~metrical load on the horizontal ~ail, the reactions from the tail on the fuselage are as illustrated in the ~lgure. The total load ~n the z cirect10n 1s still ?COO lb. but the loads a~e not sym~etrlcal about the z axis. For analYs~s purposes, consider tha loads as ~ransierred to the

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A21. 19

e.g. of the section as indicated in Fi 0 . A21.64. The ~o~ent of the two ioaGs about the c.g. = 1500 x 11.5 - 2500 x 11.5 = -U500 in. lb. The sh3ar load Vz = 4000 produces the same shear flo~ pattern as Fig. A21.62. To balance ~he moment of -11500, a constant shear r Low q :I. around the fra~e is necessary. - M11500 a = 5 .6lb /1 n , Q-2i-2xnxlS 0

I-..
/
/
/

8"

-4h

8"

-+f

8"

~f"0

(3)

(1)

e
(2)

I'~
b'

500

;;.
/

/ /

web

(4)

(A

= area

of fuselage cross-sectton) (2)

+500

600

I"

500

I"

t' 1000

Fig. A21. 66

Adding this constant force system to that ot Fig. A2l.62 J gives the final boundary supporting forces on the frame as illustrated in Fig. A2l.65. The elastic stress analysis of the t~~e can now proceed.
1500

2000

l -~.I.- ~ I --I

2500

Rib

Fig. A21. 67

! i I

I. \

fX-1
Fig. A21. 63 Fig.A21. 64

"
10"

2000

\,

Fig. A2l.67 shows a wing rib inserted in a 3 flange single cell wing.beam J which is subjected to the external loads as shown.

(1\
\2500
(2)

Flnd rlb flange loads ,t (e) and (d). Find rib web shear flow on each side at stiffener cd , Find rib flange and web loads at section 5" to left of line abo

(3)
(3 )

t
d web

l OOa

(2)

I
A21. 13 Problems.

7 1000 I

web

600

to

500

~
b

Fig. A21. 68

-'

(1) Fig. A2l.66 shows a cantilever Jearn loaded as shown. Find the shear flow ~n each of the 4 web panels. Draw axial load dia~ram for each ot the vertical web sti~:eners and also the horizontal stif:ener be. Plot axial load diagram tor beam :la~e ~embers as obtained ~r~m web shear flews.

?ig. A21.68 shows a 3 stringer single cell wing beam. A rib is inserted to distribute the concentrated loads as ShONTI. (1)
(2)

?lnj shear ~lcws in rib web (2) and (3).

~anel

(1)

Find rib flange loads at sections de and ab ,

A21. 20

LOADS AND STRESSES ON RIBS AND

1000

FRAMES

(4)

Fig. A21.69

shows a 2 stringer, 2 cell wing beam. A ~ib is inserted to transfer 1000 lb. load to beam structure. Find shear flow in rib web in each cell aojacsnt to line abo Also rib flange loads adjacent to pOints
(a) and (b).

(8) Sa~e as preblem (7) but with top skin removed instead of lower ski~. (9) S&~e as (5) but with read spar web removed instead of bottom skin.
(10) Same as problem (7) but with rear spar 'Neb removed instead of bottom skin.
(11) In Fig.

Fig. A21. 59

r---10"-j

-/1_1_

4' '"r,,'_JL.- -:P I

IjRibA

~ ~

iOi': ....
0(;'

~ :-/(----=-I :1 RibB

4' ~
"

~71:

1
I

.J-S:=;-'(~ P,
I
1

AZl.7l the external bulkhead leads P and P2 equal 4000 lb. each and ?~ equals zero. The fuselage stringer ~terial consists of tour omega. sections with an area of .25 sq, in. each. Determine the skin reSisting forces or. the bulkhead in ba lanc tng the above loads. Neglect any effective skin in this problem.

T
is-

_ JL'_ ,

\.~"

Fig. A21. 71

'+ " r-I ta:LJ

oS' __
".f(
a

(12) Same as' problem (11) but and Pa = 6000. (13)

~ke p~

~OOO

-'-P,
Corner stringer Areas:{a} I sq. in. (h) . 8 sq. in. (c) "'.5 sq. in.
(d) '" 0.4 sq. in.

Same as problem (12) 'but add

p~

.- 30CO lb.

-l

:: j r\

(14)

In a water

r-

'--,'-r-- P2

landing condition the hull frame

of FIg. A2L 72

L I

s " -1

Fig. A21. 70

(5) Fig. A21.70 shows 3 bays of a cantilever single cell, 4 stringer wing beam. The bottom skin in bay (2) is removed. Find the shear flows in all bays and boundary loads on ribs (A) and (B) when the external wing loads are as follows: T = 56000 in.lb., P 1 = 0, P a = 0, p ~ = 2000 lb., p .. = 2000 Lb , , Po = O. (6) Same as problem (5) but upper skin in bay

(2) is removed instead of the lower skin.

(7) Same as problem (5) but with the following external loads.
T

in.lb., P,=50001b., (15) Same as problem (l4) but consider that Pa = 2000 lb., p~ P, 0 and P, = 1000 lbj ths ~ater pressure is only acting on one side at the bottom of the frame.

= 56000

is SUbjected to a nor.nal bottom pressure at 200 I lb. per in. The area ot the 40" e bUlb angle stringers is "" I .11 sq. in. I each and they d are 7/8 in. deep. I 8.5" The area 0r the I e:.-.-. 24" _ Z stringers is , .18 sq. in. each Fig. A21. 72 and the depth 1.5 in. The area of the stringers a, b, c , d and e ~s .20 sq. in. each. Neglecting any effective skin determine the skin .ras i st mg torces on the frame in balanCing the bottom water pressures.

"',
~,

4-

= =

CHAPTER A22

ANALYSIS OF SPECIAL WING PROBLEMS

ALFRED F. SCHMIIT

A22. 1 Introduction.

In a previous chapter (A19) analyses of wing beams were carried out using the engineering theory of bending and rational ~odlflcatlons thereot. As discussed there, wing configurations which depart radically fram the usual conception ot a "beam" present the engineer with the choice of making apprOXimate and/or empirical corrections to bea~ theory, or of following a complete ar~lytlc treatnent of the structure. This chapter illustrates the latter approach to several syeclal problems associated with aircraft wing structures, viz., Art. Art. Art. Art. A22.2 A22.3 A22.4 A22.5 stresses around a panel cutout shear lag problem

the one finally analyzed, are treated only lightly; additional references are cited where appropriate. The analyses shown are strictly applicable to (reasonably) thin-skinned Wings only, wherein the "constant shear flaw" assumptions are valid, viz. i-the sheet carries shear stresses only 1i - normal (direct) stresses are carried 1n the flanges (spar caps and stringers) with effective areas of skin lumped in. In all cases handled here the "skin was assumed tully effective (stresses below skin buckling stress see Art. AI9.Il, Chapter A19). To enhance the usefulness of these problems, all the structures chosen for analYSiS were taken from referenced NACA (National Advisory Committee for Aeronautics) publications Wherein the reader may find detailed discussions of the problems, other methods of analysis and data obtained from tests upon the specimens. Where available, these data nave been-used herein for comparison.
A22. 2 Stresses Around a Panel Cutout

cutout in a box beam swept wing box be~

Aside from presenting one analytic treatment of these problems, a discussion is given of the physical nature of each phenomenon. An understandi~g of the nature of the problem 1s of prtne importance, since no one ar~lytic technique can be all-powerful in the solution of stress problems. The analyst must exercise jUdgment and ingenuity in approaching each new situation.

In this chapter all analyses are made using the matrix formulation of the Method of Dummy Unit Loads (Chapters A7, AS), a familiarity with which is assumed.
Such problems as those listed above are too unwieldy to be studied here in great detail; hence no attempt at exhaustive analyses has been made. To bring into relief the main features of each problem, the structure selected tor analysis is one Rhich is Si~ple in construction and so loaded as to exhibit clearly the phenomenon under study. Many practical detailS, such as the effects of sheet wrinkling, rivet and fitting "give", stress concentrations, etc., have been Side-stepped so as not to becloud the objective. Further, the problems of idealization of the original structure, into
One other important special problem - the so-called "bending stresses due to torsion" - is not treated here specifically. As indicated in Chapter AS, the general box beam analysis presented there encompasses this problem (Example Prob. 15, p, p. AS. 24 through AS. 27).

"Cutouts in wings and fuselages constitute one of the most troublesome problems confronting the aircraft deSigner. Because the stress concentrations caused by cutouts are localized, a number of valuable partial solutions at the problem can be obtained by analyzing the behavior, under load, of simple skin-stringer pane IS" (l).* Thus, in the case of a wing beam with a panel cutout of the upper surface (Fig. A22.l), it would be feasible to analyze the section !mmediately around the cutout as a flat sheetstringer panel under the action of axial stringer loads and edge shears (coming from the spar webs). The axial stringer forces could be computed with sufficient accuracy by the engineering theo~J of bending (E.T.B.) since these are removed SUfficiently far fram the cutout pr-ope,- , The edge shear nows are readily computed by those elementary conSiderations which give the spar-web shear flows

.. Numbers in parentheses refer to the bibliography at the end of the chapter. A22.1

A22.2

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

*2 / w3
Ft!C. A22.1 Fig. A22.2 Idealization of the half panel by use of substitute stringers

The sheet-stringer panel may, in general, contain a large number of longitudinal elements (stringers). The labor involved in treating this multi-element structure in detail is prOhibitive, and thus an appropriate idealization must be made. First, it is likely that the panel may be considered to be symmetric about a longitudinal axiS, so that only the halt-panel need be handled. Second, the complex, multi-stringer structure is replaced by one ~ving but three stringers. As indicated in Fig. A22.2, these stringers are: *1, a substitute stringer having for its area all the effective area of the fully continuous members to one side of the "combing stringer" (the stringer bordering the cutout) and placed at the centroid of the area of material for which it substitutes. The stress which this stringer develops is then the average stress tor the ~terial it replaces. '2, the combing stringer, being simply the main continuous stringer bordering the cutout. '3, another substitute stringer, this one replacing allot the effective material made discontinuous by the cutout. It is located at the centroid of the material it replaces, and its stress is the average stress for this same material." The sheet thicknesses used are the saQe as those of the actual structure .....
.. An alternate id.ealization, in which stringers iH and #3 are located along the lines AB and CD, respectively (Fig. A22.2), was used in Reference (2) for a box beam loaded in torsion.

Fig. A22.3 gives the geometry of the idealized panel.

.-

A:,'"

.703
.212

in in

,-.r~--n-----------I

+
+ U
-'-~

,=

A,= 1.045

in
in

2
.2

F=H==;,-----, __
__lL__
~ ~ ~

AR'" 0.25
t:,

'" 0.0331 in
'"

:-b. -ba ...

o~ _0 f---, A:l., Alii A 3 - t:, - t a

:L

t2

0.0331 in
in
In In

b:, ,.5.96
ba

=7.56

Fig. A22.3

L:15.0

SOLUTION:

Fig. A22.4 is an exploded view of the halfpanel Showing the placement and numbering of the internal generalized torces (Art. A7.9, Chapter A7) and the external loading. Note that the applied axial stresses were assumed to be constant chordwise, giving stringer leads ~ro portional to the stringer areas; their sum is P JiI , one ot external loads. The applied edge shear flOWS, coming fr~ the spar web, were assumed constant spanwise, as from a constant shear load. Other load distributions may be tandled by allowing these applied shears to vary trom ~anel to panel. For very extreme load variations additional transverse members could be inserted to create more spanwise panels allowing a better fit to the spar shear variation. The applied shear flows were considered as the other external load and deSignated P~. ?anels on the centerline have zero shear cue to symmetrJ (Fig. A22.3).

.... When the longitudJ.na1 members themselves contribute to the shear stillness of the cover (as is the case for "hat" section stringers riveted to the skin so as to form small closed cross sections), an effective thickness must be used. This point is discussed in Reference (3). In this source, however, the increase in shear stiffness is accounted for, not by increasing skin thickness, but by decreasing the panel width an equivalent procedure.

"-'--'--

--_._-_ -- ..

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A22.3

ql, -r"" 359 (P 2+3PJ.L)

qu=108 (Pa +3PJ.L)

ql.lI

"',533 (P a +3Pl. L)

The hal: ,anel was twice indete~inate. Member loads qll and qa were selected as redundants. With these set equal to zero, successive apOlications of loads PJ. and Pa were made. Also, suCCeSS1?8 applications of unit values of qll and qa were made. The results:

~~
1 2 3 4 5 6
Fig. A22.4

1
4.755

2 .316 .316 -1.316 -1.316

6 -1.68

8 - .840 .840 1.68

9.010
-19.76

.840 -1.68 1.68 .840 1.00 1.00

-39.51

a a
-12.09

.268

a a
1.00 1.00 .056
.056

7 8 9

a
- 1.317 .711

.268

a a
.0355

a a
.1l2

- 1.317 - .0450

Experience has shown that for symmetric panels Symmetrically loaded it is satisfactory to consider transverse members to be rigid in their own planes (4). Thus, in this problem, member flexibllities for forces qll' qll' q, and qa may be taken to be zero. In the actual NACA test specimen with which results are to be compared, those transverse members bordering the cutout appeared to have been heavily reinforced (to an extent unknown to the writer). Hence it is logical to take their stittnesses as great. Member fleXibility coefficients were collected in matrix form as below using the for.nulas of' Chapter A7. Note that the coefflvlents for subscripts 5, 6, 7 and 8 were set equal to zero (rigid transverse members).
=C ij

10

.056 .056

II
12 13 14 15 16 17 18 19

.0450 -

a
- 1.599

- .112
a a a a a a a a

a a a
16.16 4.86 23.98

- .035q
.359 .108 .533 .359 .108 .359

a a a
0

a a a a

MEMBER FLEXIBILITY COEFFICIENTS (E

;a

1)

(Voidl denote zeros)

10

11

12

13

l'
3.56

15

16

17 3.56

16

I.

5
6 7 8 10 11 12 13 14 15 16 17 18

1 14.22 3.56 2 3.56 14.22 47.17 11.79 3

11.79 47,17

11. 7

11.

3.56 11. 7 3.56 11. 79

7 024 7 024
7 024

8 0
'1.11 23.58 4.785

The "off dtagonal" values have negative sign because the sense of those internal generalized forces haVing subscripts (14). (15). (17) and (18) was taken OPPOsite to that used in the derivation in Art. A7.10. A change in sense requires a. change in sign In off-diagonal coefficients only.)

I.

7.11

23.58 '.7

A22 4.

ANALYSIS

0 F

SPECIAL

WING

, PROBLEMS
,

The following matrices Nere rorned: Per eq. (17) of Chapter A8:

[-ern]

= C"5206 -6633 = [392.1 293.5

or

-262.'0 -262.8

Per sq. (18) of Chapt.e r AS:

rc

r s]

293U 392.1
thIs last was found;

The inverse

5; '5~15/'5~5 l
II
i

"'1,( 11/
Fig. A22.5

d('" ~", J' I 1111'c?1 Y 7

.93 ' . ." ~.
I

.~

, .

>i'

J(IJfl
.5
MEASURED

CALCULATED

~rs -~=10-[5.802 -4.343

-4.343J 5.802

Comparison between calculated and measured stresses (psi) on the hall-panel.

Finally, per eq. (23) of Chapter AS, the

unit load stress distribution was,

~
1 2
..

inserted in theOCij matrix (in place of the zeroes used above) were

1 -10.93 -18.34 - 4.08 -11.66 1.40 1.40 3.79 15.88 .270 .510

-.65 -.65 -.351 -.351

~
~i~
=
5
5

5 18.03
5.04

6 5.04 10.08

3 4 5 6 7 8

7 8

.115 .383 .115 .383 .019 0 -.019 .0355 -.0355 .359 .108 .533 .359 .108 .359
~

18.03 5.04

5.04 10.08

The problem. was solVed retracing the same steps as before, but us Ing rthe mOd1fIedf=i~ matrix, to yield the stresses:

[G1m]

9 10 11
12
13

X
1 2 3 4 5
6

1 -11.11 -16.83 - 3.90 -13.17 2.14 2.1'4

2.52

2 -.621 -.621 -.379 -.379 .104

.372

- 1.15 0 - 1.60 0 0 0 16.16 4.86 23.98

14 15 16 17 18 19

.104 .372 .017 0 -.017 .036 -.036 .359 .108

.533

~ImJ FLEXIBLE
RIBS

8 9 10 11 12 13 14 15 16 17 18

14.61

.26 .62 0

The above analytical results are compa.ed with NACA test data (1) in Fig. A22.5 for the loading P II = 1, P:l. = O. Agreement is seen to be good.

- 1.05

EFFECT OF RIB FLEXIBILITY:

To investigate the influence of rib fleXibility, the problem was reworked assuming aluminum rib caps, of constant area A s .25 in., as the transverse members bordering the cutout. The appropriate member flexibility coefficients

- 1.60 0 0 0 16.11 4.86

.359 .108

19

23.98

.3<i9

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRU.CTURES

A22.

Comparison of this result with the previous one for the rigid ribs reveals that the most important effect of rib flexibility was to increase the concentration of stresses in the combing stringer bordering the cutout. It should be noted, however, that for this symmetric panel, the use of a very fleXible rib as compared With a rigid rib led to stress increases of the order of only 10% in the combing stringer. Thus, the "rule of thumb" that transverse tlexibilities may be neglected in summetric panelS is re-affirmed.
A22.3 Shear Lag Analysis of Box Beams

shear. The action may be comprehended readily by visualizing an extreme case: a large degree of ~lagff would occur if the load transferring skin were made of a highly flexible material such as a plastic sheet or even rubber. In such a case the inside stringers would be out of action almost entirely! With the inside stringer stresses lagging, the outSide stringers and spar caps must carry an over-stress to maintain equilibrium (~actual~ in the figure). Fig. A22.7 shows the beam analyzed herein.

"The bending stresses in box beams do not always conform very Closely to the predictions of the engineering theory of bending. The deviations fram the theory are caused chiefly by the shear deformations in the cover of the box that constitutes the flange of the beam. The problem of analyzing these deviations from the engineering theory of bending has become known as the shear lag problem, a term that is convenient thou~~ not ,very descriptive." (3) Fig. A22.6 illustrates the basic problem. The beam cover sheet is loaded along the edges by shear flows fram the spar webs. These shear flows are resisted by axial forces developed in the longitudinal members (spar caps and stringers). According to elementarJ considerations, the stringer stresses should be uniform chordwise at any given beam station ("elementary theoryff in the figure). Actually, the central stringers tend to "lag behind" the others in picking up the load because the
r, , '

-j3@8"1-

-L
A ...3497

-+

12"

6 '5"

Flg. A22.7

The beam is an idealization of one tested by the NACA and reported in reference (3). Note that the beam has no lower cover sheet and that it is symmetric about a vertical axis. Transverse bulkheads are located at stations 12", 24~, 36 ~ and 48 ~ from the root. The actual beam spectmen had three more stringers than shown in the idealized structure, these being located one each midway between the pairs or longitudlnals shown on the beam cover of Fig. A22.7. In the idealizing process, these extra stringer areas were divided equally be~Neen adjacent longitudinals. The stringer areas shown are the effective areas, with those in the top cover tapering linearly from root to tip. All skin was considered efrective in carrying direct stresses. Same detailed discussions of the techniques of idealization of practical beams are given in references (5) and (sa).

:""

1--"
--Elementary Theory - - -Actual stress
Fig. A22.6

SOLUTION:
intermediate sheet, which transfers the loads in from the edges, is not perfectly rigid in To permit the handling at the calculation in a limited space, it was elected to analyze

'0.J ..)

9 ,..--,

A22.6

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

the structure for a single transverse (vertical) tip load symmetrically placed. In that case, because or symmetry, it was necessary to treat only one-half of the structure. In addition, no shear flows could appear in the middle panels. Further, it is known that t~e influence of rib fleXibility on shear lag is slight for symmetric systems, so that the ribs were considered rigid in their own planes; hence no generalized forces were needed on the ribs to describe their strain energies. Fig. A22.8 shows the placement and numbering ot the generalized forces on the halt-be~~.

to zero, the stress distribution due ~o a onehalf pound load at the tip ( a unit load divided equally between beam halves) was readily computed.
.07692 0 -.9230 .9220 0

.07692
0 -1.8461 1.8461

~lmJ

.07692
0 -2.769 2.769 0

.07692
0

-3.692
3.692
0

fleXibility coefficients were computed with the fonnulas of Chapter A7 and arranged in a matrix.
Memb~r

\.

The shear flows q., q, q:La and q u were selected as redundants. Setting these equal

Next, the unit redundant stress distribution was computed. Fig. A22.9 illustrates a typical calculation, showing the stresses in the t t p bay for qll = 1, q"l' = q1.ll = qu = O.

OC\ , MEMBER FLEXIBILITY COEFFICIENTS (E -= 1)

I};
1 2 3 4 5 6 7
9

1 2'71

10

11

12

13

14

15

16

17

18

19

20

16 6"10 1 .74

2.684

0.0 17.99
3271
16 640

4. 724 4.202

2.684
4.724
4.202

8.939
17.84 15.78

1. 785

8 1
11 12

4.255
3.708

32 1
16 640
1. 785 4.255 7.142
I. 785

1 14 1 16 17 16 19 20

16.084
3.708 1.785

3.840

13.89
3271

3.324

16 540
3.571

3.840
3.324

7.518 6.439

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A22.7

rrj =

175 -1848 -1260 - 460.7}

16,810 32,160 9,536 3,041 9,536 9,536 25,030 3,041

rFig. A22.9 Application of a sell-equilibrating unit redwl.dant stress qll.

[,210 16,810 9,536 3,041

3,041 3,041 18,650

3'~~

The inverse at the latter was formed:

r:". -~ = 10-' -1.505 L.:: s.J - ,656

- .183

and, finally, per eq. (23) of Chapter AS, The complete redundant stresses were:
.07692 .03453 - .9230 .5086 .4144 .07692 .03136 -1.846 1.055 .7907 .07692 .02401 -2.7691 1.690 1.079 .07692 .01005 -3.692 2.493 1.199
Fig. A22.10 shows the above computed stresses and those reported by the MACA as obtained by test. Agreement is seen to be qUite good. 4.589,.1.S85
~ 1.

' 285

-L505

4.219 -1.000 - .279

- ,656 -1.000 4.687 - .495

- ,1m
- .279 - .495 5.519

X
1 2 3
4,

12

17

a
1

a
-12 12

5
6

a a a
-12 12

a a a a a a
1

a
-12 12

9 10 11 12 13 14 15 16 17 18 19 20

a a a
-12 12

a a a
-12 12

a a a a a a a a a a a
1

a a a
-12 12

a a a
-12 12

-12 12

a a a
-12 12

a a a a a a a a a a a a a a a a
1

SSS

a
-12 12

CALCULATED 3.416 2.364 1.278

5, 1

The following matrix products were formed (per eqa , (17), (18) of Chapter AS):

This is an obvious place in which to use combinations of redundants to decrease the structural coupling (reference Chapter AS, pp. AS.29, 30). (Recommended as an exercise
lor the student.)

MEASURED
Fig. A22. 10 Comparison between calculated and measured stresses (psi) in a box beam

A22.8

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

A22.4 Stress Analysis of a Box Beam With a Cutout.

In Article A22.2, one technique was employed for computing the stresses around a cutout. In that analysis the effect of the cutout was presumed to have been localized about the cutout region; consequently, the problem was treated by isolating the affected panel. QUite often, when the cutout is placed well inboard on the wing, its influence on the root stresses is appreciable. Therefore, it is desirable to be able to consider the overall ;roblem of the box beam with a cutout for such cases. nThe most convenient and the most rapid method of analyzing structures with cutouts is the indirect, or inverse, method. The analysis by the indirect method is made in two steps. First, the structure is analyzed for the basic condition that exists before the cutout is made. The results of this basic analysis are used to calculate the internal forces that exist along the boundary of the proposed cutout. External forces equal and OPPosite to these internal forces are then introduced; these external forces reduce the stresses to zero along the boundary of the proposed cutout, and consequently the cutout can new be made without disturbing the stresses. n (3)It is desired to ~odi!y the calculation of the preViOUS article (A22.3, nShear Lag Analysis") to allow for the presence at two cutouts symmetrically placed. Panel nq.~n was removed while the Single tip load remained. Since the unit remained symmetric, the data tram the previous analysis, in which the transverse rib stitfnesses were taken to be infinitely great, should still yield satistactory results.

T~~ calculat10n :or ste~ (2) above, will new be carried out oy conslder1ng a unit shear flew, q12 = 1, applied ~hile all other loads are zero. Under this loading con~iticn the relative displacements at rec~~dant c~ts 2, 7 and 17 are equal to =<2,1'1' 0<::", 1aand OO:::l",,~a respectively, where these coef:!cle~ts have been computed pr'evtoi.rs Ly 1.:1 Ar t i c Ie ";22.3.

To restore continuity (reduce the relative displacements to zero) th~ee red~~dant ~0r:es are applied, one eaCh, at the cuts 2, 7 ~nQ 17. The appropriate equations spec:~y1ng cont~nu1ty are

Note that the~e equations say si~?ly that the deflections at the reduncant cuts due to the (unkncwn ) r'ecundant forces must be equal and oppoat t e to the deflections due to q::.:a = 1. All coefficients in the above equation were computed in Art. A22.3. SpeCifically,
' ,210

16,810
32,160

16,810 [: 3,041

3,04;]
3,041 18,550

{q,}
c,
q;:,-;o

.::

3.041

9 '036} 9,536 {3,041

The a~Jve matrix is the [ocrsJ of Art. A22.3, with the nq u TOWn and nq u column" removed.

SOLUTION:
The calculation was accomplished in three

The equations were salved to give the values of the redundants for a unit applied load, q.2 = 1, as
.1399
.2134

steps:

1) The stress distribution was found in the "basic structural! (no cutout). This work was carried out in Article A22.3 where it was found that q .. = .024011bs./inch.
2) A stress distribution was found for an napplied load" ot q u = l. Such a loading has zero ~xternal resultant.

1.00

- .1055

The q12 force 1s included in the above for later convenience. Then the complete stress distribution, due to applying a unit shear flow
q

1, was

3) .02401 times this last stress distribution was subtracted from the ~irst

The procedure described here is quite generally useful for studies of the effect of remoVing one or more members; such might be required for an analySiS of the effects at structural damage.

See Appendix for a method of "extracting" the inverse of this matrix from that previously found for the complete

~rJ

matrix.

_,_

. _ - - , . . ' _

C-=~,

~._-

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A22.9

where the row matriX ~12,mJ is simply the w12 r ow" or@im]BASIC

o
- .1399

COMPARISON WITH TEST DATA:

Reference 3 reports test data for the case analyzed, the stringer stresses being plotted ~elow in Fig, A22.ll. In the actual test specilnen a stringer passing through the cutout was severed, it having zero stress at stations 12 and 24, therefore. However, during the idealization process discussed in Art. A22.3 (tor the beam without cutout) the area for this stringer was placed partly with the combing stringer and partly with the spar cap. In the same 'NaY, some effective sheet trom the midportlon at the panel, now made discontinuous by the cutout, was added also to the spar cap and combing stringer. It follows then that the full idealized areas at the combing stringer and spar cap ShOUld not be used in figuring the stresses at stations 12 and 24 (produced by forces q,J q~o, q14' q1e)' ~ith these areas reduced by the appropriate SUbtractions, the stresses were computed and are plotted in Fig. A22.l1. Agreement with test data is seen to be qUite satisfactory.

1.579 -1.579

- .2134

4.240 -4.240

1.00

-7.760 7.760

- .1055

-6.494 6.494

( [glrJ taken from Art. A22.3)

wi~h

stresses were SUbtracted from the wbasic stress distribution of Art. A22,3, as previously eX~lained.

To obtain the stresses in the loaded bea~ the cutout, 0.02401 ~.nes the above w

.07692 .03789 - .9230 .4683

.4547

CALCULATED

l = ks l -.02401{G l: { GImjcUTOUT( L'nfBAS IC in.:;J

.07692 .03648 -1.8461 .9536 .8925 .07692

-2.7691 1.8763 .8927 .07692 .01258 -3.692 2.649 1,037

MEASURED
Fig. A22. 11 Comparison between calculated and measured stresses (psi) in a box beam with cutouts.

Note that q1~ is now zero and the cutout panel may be "lifted out w In the case of a structure under a variety of external loadings (m = 1, 2, 3 ---), the ~ore general equation, carresponding to the above, i s

A22.5 Analysis of a Swept Box Beam.

WExperi~ental investigations of swept box beams have shown that the stresses and distortions in a swept wing can be appreciably dit~er ent in charac~er from those that would exist if the ~oot were ~o~l to the wing axis. The principle ef~ect of sweepback on the stresses occurs under bending loads and consists in a concent~atlon ot bending stress and vertical

IG l = IG l _fG} L G ,,,,,J L imJcU'l'OUT li;nJBASIC 'liJl.ll

A22.10

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

shear in the rear spar near the fuselage. With regard to distortions, the effect of sweep is to produce some twist under loads that would produce only bending at an unswept WIng and some bending under loads that would produce only twist of an unswept wir~.~ (6) In the following example a swept box beam is analyzed by the matrix methods of Chapter AS and, in particular, by the specific techniques of reference (7). The method accounts for the interaction between the swept cover panels and the longitudinal members. It is this action that is responsible for the distinctive structural craracteristics of the swept box
been ,

found in reference (8). One method of ~n~lyz ing such wings is given in Chapter AZ3.
THE STRUCTURE:

The structure shown is F:g. A22.12 is a~ idealization at the NACA test beam of references (6) and (9), in which a Single substitute stringer has been employed along the cover sheet ta allow for the ant:tl~ated shear lag effect. The figure shows only one-half or tte complete unit, which was built symnetrically about the axis corresponding to the langitudln~: axis at the airplane.

Only tip loads were to be applied (at points A and B). The outer section of the be~ Again, we emphaSize that the method used was assumed to carry stresses which could be here 15 strictly applicable to thln-sklnne~ calculated reasonably well by the engineeri~g theory at bending (E.T.B.). For this pur,ose Wings at beam-like proportions only. Considering the wide variety at structural layouts it was jUdged satisfactory to consider the which may be employed in swept w1ng configuraouter 66~ at the Je~ as a Single bay (A-B-D-CJ. tions, a comprehensive treatment carillot be given If loads were to r4ve been a~pl:ed inboard of here. An excellent review at methods better the tip, it would :iave been necessary to conadapted to thick-skinned construction and to steer additional bay dtvf s i ons between .;'-3 and "plate-like" (very thin, Wide) wings, may be C-D (that is, insert additional ribs at statlcGs

/"-.

52"

.050

.050

--i.e

.050

.050 66"

Section C-C (Steel) :;--t=. 125

,..A-.539

t,,-

.078
A A

t,

t l
A 1)
(2)

+ J
.078

B_1
,

-f

A=l. 966 Section B-B

A=I.430

"

:> .L

r-'~,
Fig. A22.13

~..d5"

.[.-15"

Flg. A22.12

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A22.11

of load application). Rib C-D was located at one of the actual rib locations in the NACA test specimen and was assumed rigid in its own ;>lane. The choice of bay C-D-F-E as a single bay was somewhat arbitrary. For improved accuracy, additional ribs inboard of C-D could have been used. Note that any ribs placed inboard or point F will produce triangular skin panels in the cover sheets. Examples of treatments for such panels may be found in references {9},
(10) and (11).

affects total deflections only and can be omitted in a stress analyses where ceriect icns ar-e not sought. Since only symmetric loadings were considered in this analYSis no shear was transmitted by the carry-through bay and hence no shear flows were shown in that portion. Sets at additional axial forces (q18 through q1e:) were applied to the ends or the flanges and stringers adjacent to the Obliquely cut ends of the cover sheet panels in bay C-D-F-E. These forces are necessary to account for the interaction between the swept covers and the longitudlnals. As shown in Fig. A22.14, the pure shear flow on the oblique edge is obtained by superposing onto the panel a zeroresultant system consisting of a uniform tensile stress of intenSity 2q plus a pair of concentrated balancing loadS. The balancing loads must be contributed by the bordering longitudlnals and hence react on these as tensile loads (Fig. A22.l4C). The balancing loads applied to the stringers are shown dashed since they are internal forces within the bay and are not to be entered into the equilibrium equations for the structure.
P=qL

Rib E-F was considered flexible in its own plane, it being known that the flexibility of a rib is important at a locati0~ where a structure changes direction.* Note that this rib was made at steel in the test speCimen. Effective areas of longitudlnals as shown in Fig. A22.12 were computed by considering all of the skin to be effective. The spar cap areas are equal to the sum of the areas of the angle member at the cap location, plus onehalf of the effective area of material be~Neen the cap and the substitute stringer (this area i~cludes several stringers as well as Skin) plus one-sixth of the attached spar web area!* The substitute stringer area was collected in like ma~~er from the half-panelS to either side. The method used in calculating the effective areas of the rib caps (E-F) is given in detail in ~efersnce (9), from which the va lue used here was taken. The "carry-through bay" cover sheet thickness is equal to tnat used on the specimen (.050") plus a weighted increase to allow for the presence of splice plates along the plane of symmetry (see reference 9).
INTERNAL GENERALIZED FORCES:

2J
(a)
(b)
(c)

P"'qL Fig. A22.14

Showing how the uniform. shear stress on an oblique panel end (b) is created by superposition of a uniform tensile stress plus two balancing forces (a). The balanCing forces react as tensile loads on the bordering longitudin:aJ.s (c).

Fig. A22.13 shows the choice and numbering of the generalized forces.

From an energy Viewpoint, these dashed forces account for the additional strain energy stored by the axial components or shear flows in the non-rectangular panelS. This energy is The beam was rigidly supported at points stored in the cover panels themselves (and is E, F and at the two corresponding points on the accounted for in this manner since the longlother beam half. These might correspond to the tudinals contain the effective area irom the :uselage ring attach points in an airplane. The vertical end caps on rib E-F were considered cover sheet** ) and in the longitudinals Which rigid ~xially, so that no fleXibility coreact against these components. efficients were associated with the reactions "Dashed loads 11 are applied to the longiqal5 and qae:' Flexibility of these members tUdlnals adjalsnt to any obliquely cut panel end. Similar dashed loads would be applied to
The effect of neglecting this rib's flexibility is demonstrated later in this example.
-** This much of the energy could be accounted for in another fashion by modifying the member flexib1llty coeUlcient for the sheet panel. See Reference 12. where this was done. However. that referenc e incorrectly neglects the additional energy stored in the longitud1nals, as was demonstrated in Reference 13.

** The factor of 1/6 is used so that the effective area contribution of the web results in a structure having the same moment of inertia about a horizontal ax:I.s as the original. Some of the problems of idealization are discussed in reference (2), p. 16.

.~i~;~';~.~~:~~2~;~~~'~~;~~1k:;(~~~~~:{;>i~~\;~~~;:~;~~~;~f);:~,~~:;,~;.: ~.';~: . ~. '. {yZJ!;::~7 ~.~{~~

,_.

---

A22.12

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

the outboard ends of the panels in bay C-D-F-E if they too were cut obliquely. Such ~anel configurations arise often in swept wing construction having ribs parallel to the airstream. Formulae for ~ore general quadrilateral panelS are given in Reference '!'HE STRESS DISTR!BUT I O,:S

Stresses in inner bays: According to the discussi~n of Art. A8.12, Chapter AS, the deter.ninate stress d~stributicn,
may be any stress distribution in

tmJ,

8qU~:'

For the symmet.r-t c rcadmcs considered here the structur e was indeterminate only two times since the auter bay was assumed to be determinate by the S.T.B. Stresses in outer bay by S.T.B.: Flange stres ses at rib C-D (for both

ibrlum With the applied loads, and preferably one close to the final true stress distribution. The magnitUde of the redundant fcrces is reduced by use ~f ~ satisfactory estiillate of the true stresses. The stresses in ~he two ~nner bays were determined for both gim and glr si~ultaneously. Since this inner portion of the structure is two times indeterminate we can estimate two loads. For this purpose the two flange loads q 111 and q lo4 were written as
qu

P1 = 1 and Pg = 1 )
II I

.. 66"*

88.57 i n . , C = 3.5"

= q~11
=

-+ q~11

IlC to -1'" 2 .608 psi

qh

q~ .... q~.

Therefore,
qa = q,

= 2 .608 x 1.121

= 2.924 lbs.

where the (Single) pri~ed values are approxi~te values determined by the .T.B. and the double primed values are the unknown corrections (the redundants). Using My/I at stations 66" and lIS" from the tip gave "

q.

2.608 x 1.373 = 3.581 Ibs.

For a unt t transverse load at the shear center (midpoint , because of symmetry)
Ibs Qa =~= 2x66 02713 Trl

= 5.228 qi. = 3.899

q~11

p~ -+
P~

5.228 PI1

-+ 3.899 Pl1

The eq~illbrium equations for the elements of the structure were written next by summing forces and moments. Joint F
q., + q

E'%

qb=

.O2713+~

2.924

... 07143 Ibs

Til

q .3 + qlla = 1.414 q16 = 5.513 P~ + 5.513 PlI + 1.414 q~"

The unt t load was shifted 15" to either side by application a r a torque, T = 15 in.lbs.

Joint E
q.. + q = 0 q.. = 1.414 q.. q = 7.392 P, + 7.392 p. + 1.414

The unt rcra shear flow superposed was

15 q=~=2X 210 = 03571 Ibs in

q~2

Finally, su perposition gave the stresses tor the outer ba yas

about E-F
q..

I1Z
1 2

qu + q

= 11. 918

P, + 8.888 p.

1 .1071 .06284 .00858

.03572

2 .03572

3 4 5
6

.00858 .06284 .1071

... This is a rather crude way to estimate these loads and is used here only for simplicity. The analyst is generally better advised to exercise a little more ingenuity in making these estimates, even to the extent of being guided by other swept wing solutions.

2.924 3.581 2.924

2.924 3.581 2.924

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A22.13

0 -.7115 0 0 0 -21,2 0

0 0 21.2 1,682 0 0 0

0 0 0
1 1

-1 0 21.2 0
1

0 0 0 0 0
1 -1

0 0
1

~j
C

"

q. q.
q""

l~
H

"

"'j ~
.

q.

01'
F ,;
~

0 0 0 0 0

iN = = .

0 0 0
1

q" q .. q" q ..

q"

0 0

-21.2 0

Shear Flows around the nonrectangular panels. (Check by summing moments about E, G and F.)

- .1403 .04431 2.756

- .2111

o
.01923

.04431
2.756

o o
.707 .04545

Cap EC
qe - .7115 qa = .04431 P:l,

o
o

+ .04431 Pa +

.04432
.2831 -3.696 -5.466

.04432
.3540 -3.696 -2.436

.01923 q~a

o
-.707 .707

Rib Vertical at E
q. - q ..

o o
.707

=-

.1403

Ii -

.211 p.

Cap

EX}

Atter inverting the matrix ot coefficients on the lett hand side at this last equation, and multiplying through thereby, the stresses were obtained as

Joint G
.707 q:l,::)
q

c,
q1,,)

.1006

.0295

.00676 -.00676

.707

q ..

=0 + q .. =

q.
0

.0791 - .0208 -.01753 _.00950 .0013 - .0411 .00401 .0230 .00675

q"

Cap GF

q"
q..
qll1,

.0422 .2409

.1135 -.00675 .2409

.00676 -.00676

3.087
-2.379

.964
-1.473

-.9355
-.2284

-.3444
.3625

Rib Vertical at F

q..

Cap DF
ql.l + 1.662 q1,Q

= .04432

(p1, + P II

The complete determinate and unltredundant stress distributions, USing the results up to this paint, were therefore:
)

+ .04545 q~40

The fi~st fIve of the above equations were readily solved by substitution, y~elding:

q"
-qllo

= s., = 2.756 P, = q.. = 3.696 P, q.. = 5.466 p "

equll1brl~

,.. 2.756 P, + .707 q~40 + 3.696 P, + .707 q~ " .707 q" + 2.436 P, u

- .707

q~ ..

Also, from

of joint G:

The remaining seven equations were arranged tor solution thus:

~!

-437~'"'"

A22 14

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

1"'(
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
24

1 .1071 .06284 .00858 .03572

2.924

2 .03572

12"
0 0 0 0 0 0 0 .00676
-

14" 0 0 0

q..

= 2 = = -.989 P
~Pn~
lo

X '25 - x .0

52.3

1-

:::I

X (- ) q , = -12.5 q, .2191 q~a

X

.260 Fa + .360
r,

.1168

q~

.00858 .06284 .1071

q.. =

~Pclt

73.5 '5 2 X 125.8 X ~

I-I qe = -17.5 q,

a
0

'"' -1.384 Plo

.3068 q;a .1662 q"

q~a q~a

2.924

Similarly, for the panel HGFD; qlo7 -.0146 Plo

q
l.8

3.581 2.924 .1006 .0791 .0013

.0422

3.581 2.924 .0295

0 0

-.0244 Pl,

.460 P a -.0449 .773 ?a -.0754

.2576 .4324

q~ q~

- .0208 - .0411
.1135 5.228 3.444 3.899 .260 .360 .460
.773

- .00676 .01753 - .00950

.00401 .0230 .00675 0 -1.0 1.0
.1188

The following matrix

17.70 [<><r,J = [ -173.7

produc~s

were rornec:

(c r eqs . 17, 18, 23 of Chapter AS)

250.2J - 50.43 42';IJ 83.04 0.01483

- .00675
1.0 -1.0 0 .2191 .3068

1m:gl~

5.228 7.729 3.899 - .989 1.384 - .0146 .0244 .2409 3.696 3.087 2.379 2.756 3.696 5.466 2.756

[-=rJ = [114.5 42.96 C=rsj =[ as,

0.01082 .005563

.00556~
was found
2 .03572

Finally, the true stress

dis~ributlon

.1662
- .2576

- .0449
- .0754 .00676 - .707

X
1 2 3 4 5 6 7 8 9 10 11

.4324 .OC676 0
.3444

.1071
.06284 .00858 .03572

.2409 -3.696 .964 -1.473

.00858

- .06284

- .9355 - .2284
0 .707 0

.1071
2.924 3.581 2.924

.3625 .707 0 .707

2.924
3.581 2.924 .0747 .0740 .0581 .0681 4.070
5.213

2.756
3.696 2.436 2.756

25

- .707 - .707

.0051 .0112 .0039 .1481

26

DASHED LOAD CALCULATIONS:

In the above matrix, loads qu' ql, 5' q.. and ql,a were obtained tram q. and ql,O following formulas given in reference 7 for general quadrilateral panels. The equations applicable to a parallel-sided panel are:

~1mJ

12 13
14

2.241 4.292
6.038

6.573

15 16 17 18 19 20 21
22

.925

.P~T r L pdq
l),.;...

v: 0)'

/',

-1.295

Pc "'c:;a qL x 2 P D", c
+

.140 .201 .0431 .073 .2060

.6514
.2150

-1.093 -2.877 3.249 -1.145 4.646

d qL x 2

--..L

Balancmg Load (= D~shed. Load)

-1.584 3.022 - .015 4.268

1.:::184

ForJP1i1ae from Ref. 7

c t d = 125.8", L

For the panel CEnH; c = 52.3", c ~ l5~. Hence

= 73.5" ,

23 24 25 26

2. 877
4.394 4.652

3.036 4.268

..iiiiiiiiiiiiiiiiiiiiiiiiii

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

MATRlX OF MEMBER FLEXIBIUTY COEFflCIENT-' (E = 1) {FormulaB from Chapter A7:lnd Reerenee 7)

A22.15

1 15.400

.,

10

11

"
1 .46

ia

14

15

"

17

"

19

ac

21

! 22 "

,.

" "

1 2 60 1 40 70.170 50.0 52.33 112 130 15.46 8.984 6.542 8.984 8.984 6.542

400 46000
5
0
,

15.46 .984

30.92 17.97

30.9Z 17.97 17.97

.42
15.46 8.984 8.984 6.542 30.92 17.97

.08
30.92

13.08

."

17 97 17.97 17.97 7.97 13.08 13.08 2130 9.046 4. 2

4 '"

04 4.523 4.523 9.046

s.
4

19.58

CO!lPARISON WITH TEST DATA:

5.2 3.8

~.6
/

2~
/

Root

2.6

118)......

2.6

-,
(66)

2.8

f\

~.6
2.' 2.'

3.2

requirement. More details concerning the testing teChniques and method at data presentation would probably resolve this conflict. Both test and calculated values clearly exhibit the characteristic build-up at stresses in the rear spar of a swept Wing.
RIB E-G-F RIGID:

2.'

Leading ...........

Edge

't'
/

As a matter at interest, it was decided to investigate the effect upon stresses when rib E-G-F is taken to be rigid. Such a calculation is readily achieved by putting the member flexibility coefficients for the rib equal to zero .
MEASURED

CALCULATED

Station (0) ...-

Fig. A22.15 Comparison Between Calculated and Measured stresses (psi) in a. Swept Box Beam.

Thus, in the matrix those coetticients with Subscripts 19, 20, 21, 22 and 23 were set equal to zero and the complete calculation was repeated. The results, tor spar cap loads at the wing root (a most sensitive point), were:

E: IjJ

Fig. A22.15 shows a comparison between the calculated stringer stresses and those measured by the NACA as reported in Reference 6. The stresses showa are for a unit tip load, centrally placed (p. = p. = 1/21b.). Considering the limItations on the analysiS the agreement Is generally satistactorj. Thus, the discrepancy in the MY/I stresses at station 65 may be attributed to a lack ot precise knowledge of the test parameters. The calculated stresses in the leading edge spar between stations 65 and 118 cannot duplicate the experimental variation since only a single bay was employed in this region in the idealization. The fact tr~t the calculated root stresses run consistently above the test values is 1ifficult to explain. Inasmuch as the calculated stresses satisty equilibrium, the test values, all being lower, would seem to defy this runcamerrta;

~l~RIGID RIB

~
=
12 14

1 3.824 6.993

2 1.792 6.610

A comparison with the fleXible rib ca1culations follows: SPAR CAP FLEX. RIB RIGID RIB % DIFF. LOADING FORCE Simple 6.311 5.616. 11 q" Bending; 12.61 13.60 q,. 8 Flo ::P~=l Torsion;
Plo=-P~=l

q" q,.

1.829 0.535

2.032 0.383

11
28

,.
,x:: /)
)--. t'?

A22.16

ANALYSIS

OF

SPECIAL

WING

PROBLEMS

ConSidering that rib SGF was relatively rigid to begin with - being nade of heavy gage st3s1 - it may be seen that neglect of the flexibility of a corresponding all-al;~inum rib could lead to serious arrors.
REFERENCES

(7)

weme ,

L. B., and Lans tng , 'A A Metccd

~Gr

Reducing the AnalYSiS ~! Com~lex Retun1~nt Str~ctures to a Routine Procec~re, J8urnal Q: Aero. SCi., 19, Oct. 1952
(S) 'williams, M. L., A Review of Certain

Analysis Meth~ds for Swe~t Wing Structures, Journ. of Aero. SCi., 19, p. 61~, lS52

(1) Kuhn, P., Duoer-g , J. E. , and Diskin, J. H.,

Stresses Arolli,d Rectan~ular Cut-Outs in Skin-8tringer Panels Under ft~lal Load - II, NACA 1rJR w6B (ARq 3J02), Oct. 1943
(2)

(9)

~e1denfels, R., Zender, G., and Llbove, C., Stress ~d Distortion Ar~lysis of a Swe~t Box Beam Having Bulkheads Perpe~c~cular ~o the Spars, NACA TN 2232 .

Rosecrans, R., A Method for Calculating Stresses in Torsion-Box Covers with Cutouts, NACA TN 2290, Feb. 1951 in Box Beams: Methods of Analysis and Experimental Investigations, NACA TR 739,
1942

(10) Bispl1nghoff, R., and Lang, A., An In-

(3) Kuhn, P., and Chiarito, P. T., Shear L3.g

vestigation of Defo~tlons and Stresses in Sweptback and Tapered ~ings with Discontinuities, Mass. Inst. of Tech. Rept., July, 1949 (See also, Journ. of Aero. ScL. ..la, p , 705,
1951)

(4) Haft, N. J., and Libby, P. A"

Recommendations tor Numerical Solutions of Reinforced Panel and Fuselage-Ring Problems, NACA TR
934, 1949

(11) Denke, :., The Matric Solution of Certain Non-Linear Problems in Structural Analysis, .jour-n, of Aero. se i . ~' 1956
(12) Levy, S., Compirtat Ion of Influence Coefficients z'or Aircratt Struc-:ures witt'..

(5) Kuhn, P., Approximate Stress Analysis of Mu1ti-3tringer Beams with Shear Defor.natlan of the Flanges, NACA TR 636, 1938
(sa) Kuhn, P., Deformation Analysis of ~lng Structures, NACA TN 1361, July 1947

Discontinuities and Sweepback , Journ. of Aero. SC1., 14, Oct. 1947

(13) Is Lfnger , J. S., Stress Analysis and

(6) Zender, G., and Llbove, C., Stress and Distortion Measurements in a 45 Sweot Box Beam Subjected to Bending and Torsion,
NACA TN 1525 ..

Stress Measurements for a Swept Back Wing Having ~ibs P~ra11el to the Airstream, McDonnell Aircraft Corp. Report 1127, April
1949.

CHAPTER A23

STRUCTURAL ANALYSIS OF A DELTA WING ANALYSIS BY THE "METHOD OF DISPLACEMENTS"

ALFRED F. SCHMlTI'

A23. 1 Introduction

The pur~ose at this chapter Is two-fold: first, to present a discussion and an example of techniques for handling the structural analysts of highly redundant low-aspect ratio wings (typified by the delta wing) and second, to illustrate the nDlsplacement Method" of structural analysis (l)~J a methOd fundamentally different from those of Least Work or DummyUn! t Loads.

The very low aspect ratio, thin wing IS a structural configuration of relatively recent origin. It Is a bUilt-up structure of ribs, spars and a cover sheet but yet 1s so thin and so highly redundant that its structural characteristics are actually closer to those of a tapered flat plate than ~o a beam. In a beam the prlmarJ stresses are longitUdinal; indeed, one of the basic assumptions of beam theory is that transverse stresses are negligible and that cross sections remain undetormed. For the very low aspect ratio wing however, chordwise stresses and deformations are of great importance. The degree of redundancy of these low aspect ratio wings is very high because of the multiple pathS by which a load ~y be carried over the gridwork at spars and ribs. In the wing beam problem examples of Chapters AS and A22 redundant internal loads were determined by use of the Method of DummyUnit Loads - a variant of the Least Work theorem. ~nl1e the ~ethod is perfectly general, it becomes increasingly difficult to obtain satisfactory accuracy as the degree of redundancy of the system increases. To some extent, accuracy may be retained by skillful choice of redundants and through the use of carefully chosen determi~te stress distributions (see Art. A8.12, Chapter AS). A high degree of engineering jUd~ent nust be exercised. how~ver. SVen so, it has been found difficult to apply successfully these Least ~ork ~ethods to the low aspect ratio ~ulti-5par wing. Several authors (2), (3), (4) have presented methods of aralysis for the highly redundarrt wing based upon use of the "Displacement MethOd" of analySiS. These references differ primarily in the techniques advocated
Numbers in parentneeee refer to bibliograpny at end of chapter.

for the breakdown and idealization of the structure into primary elements. In the following example, a method following that proposed by Levy (2), is applied to an idealized delta wing structure. The choice of this method for presentation is primarily for pedagogical reasons - it being the least detailed and consequently the easiest to grasp conceptually. The reader who 1s interested in actual application is recommended to Reference (4) for techniques which are probably better able to handle practical problems because of their greater generality. fleXibility and growth potential.
A23.2 Basis of the "Method 01 Displacements"

The Method of Displacements draws its name from the fact that displacements rather than forces are dealt with as the independent variables. In earlier chapters the relation between structural detlections and applied loads 'NaB written through the use of the fleXibility influence coefficients as (see Eqs. 24, 26,
Chapter A7)
(1)

where Om 1s the detlection at paint m and Pn is the external load at point n. If one forms the inverse of ~mnJ this equation may be rewritten as

'

{P

n }=Gmn-~{~}= ~ronJ {~} - - - (2)

is called the st1!tness

where

[KmnJ =. [Amn-:LJ

matrix.
~nile the fleXibility influence coefficients give the deflections per unit load, the stiffness TIatr~x lves the loads ger ~nit deflection. Thus, anyone column (say the :n"') of K nm gives the values of the forces (reactions) developed at all numbered points ot the structure when ~he corresponding point (paint m) is deflected a unit ~~ount, all other points held fixed.

ConSicer, for illustration, the grid-Ilke structure of Fig. A23.1 into ~hich a delta wing structure has been idealized. Note that the torSional stiffness of the Shear-carrying Cover sheet has been accounted for by the presence of
A23.1

A23.2

STRUCTURAL

ANALYSIS

OF

DELTA

WING

quadrilateral "torque boxes" connecting ~olnts ot intersection of the spars and ribs. This

/ Rib

are written for ~he individual elements of the st~ucture, they ~re Simply ~dQed to g~ve the stiffness ~trix of the composite. Finally, the resulting stif~r.ess ~trix ~zy be inv8:ted to yield the matrix of ~lexibility influence coefficients (Eq. 2 "in r-everse"}. We note here, ~s an aSide, ~hat t~e Monent Distribution Method of Chapter All, the Slope Detlec~ion Method of Chapter A12 and the~od of Successive Corrections of Cr~pters A6 and A15 are all ex~~ples ot the "Displacement Method" o~ structural analysis. In each of these methods the displacements are taken as the independent variables and these are adjusted to achieve equilibrium of the loaded str~cture. The nadjUBtment n may appear as a systematic relaxation or artificial constraints (~oment Distribution and Successive Corrections) or it may be dane mathematically !~ one stroke by the solution or a set at simultaneous equations (Slope Deflection Method). The latter approach solut10n of a set of simultaneous equations - 1s essentially that :'ol1owed herein, the "so Iut t on" being effected by ~trix inversion. It 15 to be said of the r.ethcd ot Displacements that it 1s complementary to the Least ~ork method in that it is better suited to the handling of the highly redundant pr-ob.Iem , While Least Work is better suited to problems of tew redundancies.
A23.3 A Delta Wing Example Problem

Fig. A23.1

Idealized Delta Wing Structure

idealization is discussed in more detail in the example. Assume that the points of intersection have been numbered as in Fig. A23.2.

The idealized structure of Fig. A23.l will be analyzed to determine

Fig. A23.2 Reactions Developed at a Net of Points when a Single Point is Displaced.

(a) the influence coefficients (b) the internal stresses as a function at the applied loadS. The grid points are numbered as in Fig. A23.2. Note that the numbers increase to the rear and outboard.

always symmetric.

It one point, such as point 6, is displaced a unit amount, the other points remaining fixed, reactions are developed at these various points as shown. These reactions may inclUde moments as well as forces, should rotations have been taken as cer-ement displacements in addition to the translations (see points 5 and 7 of Fig. AZ3.2). The values of reactions so developed would for.n the 6th column at [Knmland the 6th row, since [KnmJ is

IDEALIZATION:
In the structure un~er consideration member 1-2-3-4 lies on the airplane centerline. The bending stiffness or this member includes that or the half-fuselage plus one-half of the "carr/-through structure." More detailed representations or the structure in this region are generally deSirable, the overSimplified model used here being employed to limit the amount at data to be handled. Some techniques which may be applied in idealizing the structure in this region are given in Reference (5). Idealization, particularly with regard to effective skin areas, has been discussed in some detail by Levy (2). The cc~plexities of

~ :,_~.

_~ .

~'._" <'~.'-~~'.'k"". >_~,>,: _ .~._':

-~-'.. :'--;e/'--~':~''>--'

,c . .

,.{::.1..

'T'... ~,_.<.;_.":~:_,_:...,.,,::..:;-.:::

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A23.3

this phase of the problem are too great to permit an expansive treatnent here. Briefly, Levy's recommendations are; (a) all skin may be considered effective between spanwlse spars when computing the cap areas of such spars. This assumption is subject to modification, of course, if spanwise stresses are anticipated which will buckle the skin. (b) for streamwise ribs an effective width of O.362L, where L is the rib length, may be taken as acting with each rib cap (Fig. A23.3a).

Beam Element Prooerties (Moments of inertia, in (inches"), are assumed to varJ linearly between numbered
points) li2 Fuselage Beam (1-2-3-4)
3@48"
1= 2000

Spar 2-5
-48"--

==

"II

--

2700

3400

2200

(e) for the leading edge spar an effective width of skin of .181 of the span between spanwise spars is taken as acting with the cap of each such spar segment (Fig. A23.3b)

Rib 5-6-7
48" - - 48"-

8par 3-6-8
_48"_48"_
'

1
I = 9.24

r
20.79

!
3.59
47.23

!
40.82

12.33

Rib 8-9

Spar 4-7-9-10
_ __3@48"

'V'"
I

48"-

3.82

(b)

[
1= 6.23

!
2.87

[
5.09

4.78

!
0.28

Fig. A23.3

Effective Widths of Cover Sheet for Rib and Oblique Spar Caps (After Levy, Ref. 2)

Leading Edge Spar (1-5-8-10) The properties of the structure (Fig. A23.1) after idealization, are summarized as fol1 ....S: Net Point
1 2 -----3@87.8~'l_-----

f
I '" 16.02 12.29

,
8.30
O. 26

Beam Depth
(inches)

(see Fig. A23.2)

- - - -

5.56 9.12
7.80

SIGN CONVENTION Al'Jll NGrATION:

The Sign convention and notation adopted in conjunction with the grid numbering scheme of Fig. A23.2 is as follows:

3
4
5

- - - -

3.16
4.88

Forces: Pm - transverse torce at jOint m, ~ UP Mm - moment at joint ~ acting about a pitching axis, + NOSE UP

6
7 -

7.26 3.06 4.02 2.74 0.72

8 9 -

Nm - moment

10 -

at joint m acting about a rolling axis, + ~IGHT ~ING UP

All cover skin t = .051"

Qm - moment at joint m acting in the plane

of an obl1que beam member (a :nember neither parallel nor normal to the st'reamwtse direction), + CLOCKWISE

A rational, systematic means of treating cover sheet panels is given in Reference (4).

',/HEN THE MEMBER IS VIE'';ED ',-lITH ITS JOINT NUMBERS INCREASINJ LEFl' TO RIGHT

A23.4

STRUCTURAL

ANALYSIS

OF

DELTA

WING

Displacements:

6m - transverse
+UP
~

displace~ent

of jOint n,

- rotation of JOi~t m about a pitching axis, + NOSE UP axiS,

+

~ - rotation of joint m about a rolling

RIGHT

~ING

UP

Tm - rotation of JOint m in the

~lane of an oblique member, + IN DIRECTION OF + Qm (see page .'123.31

For a special purpose it will be convenient later to introduce another set of displacements to be called "relative displacements." ~nere any confusion can exist, the displacements defined above, which may be visualized as being measured with respect to a set or reference axes fixed in space, will be referred to as "absolute displacements. " the above sign convention, any beam element (spar. rib or oblique member) which is viewed so that its joint numbers increase from left to right, will have positive forces and displacements taken in the same sense as every other member. This paint is illustrated in
Fig .'123.4.
~ith

~herev~r addi:l)~al knowl~d~e about the effects of various loadin;s are required. C8~es po~dlng forces or ccup~es should be ~dded. T~us, the influence of additlop~l forces at paints lnte~ediate to the grid intersections, may be accounted fer by the adjitlon of approprla:e extra forces along beam spans. Couples may be applied at any pOints where deflection Slopes are required, e.g., the streamise slope at the trailing edge ~lght be needed in an aero-elas~ic analysis, in which case c oupLee :-': .., ?1.,., M II and I"l:::., 0 (c r Fig. ;'23.2) wouj,c oe emp.l.oyed , The effect of couples from a~x~llary aeradyPZOic surfaces and/or actuators ~y be desired, in which case appropriate additions may be made.

In the present exa~~le, ~ecause of space limitations, only the ~inimum num~er of torces and corresponding displacs~ents are conslderec, viz:
P~ N~

through Pl through
N~

forces on net paints spars at the line

alr~l~a

- rolling couples on through-

center-

M:l.' N 8

Me' N e, M!I - pf t.chtng and rolling

couples at points at Intarsect~on of three be~~ elements

A23.4 Calculation of Element Sillfness Matrices

(a)

1
.p.

Rib 5-6-7

,
ip..,
"

(b)

Spar 4-7-9-10

Q.

'Po
Fig. A23. 4 illustrating the Homogeniety of the Sign Convention

The task of computing the stiffness matrix for any one element at a configuration is a relatively straight-forward structural problem. This problem may be either a statically determinate or a redundant one depending upon the geometry at the element anq upon the nunber of pertinent forces and displacements associated with it. If i t is a redundant problem. it 1sa small one by comparison with the overall structure of which it is an element. One might say that a feature 0:' the Displacement Method of AnalYSis 1s that it "~kes (many) little ones out of bIg ones"! For instance, looking at Spar 2-5 of the present problem (Fig. A23.5a), we see t~~t there are four forces (Teactions) aC~lng on it
The implication here is that beam torsional stiffnesses are not considered in such interactions. This assumption is probably quite satisfactory in general, bending stiffnesses being much greater than torsional stiffnesses 'for most beam elements. On occasion, a beam will have considerable torsional stiffness and it will then be necessary to account for it. The leading edge spar might be such a beam if it replaces the "D" nose section of the Wing in the tdealization.

CHOICE OF PERTINENT FORCES AND DISPLACEMENTS:

Before beginning the analysis proper. the analyst must decide upon which torces and displacements are to be considered pertinent. One will, at necessity, have to consider trar~verse forces (and translatorJ displacements) at all net points (ten paints in the example). In addition, rolling moments, N, (and rotations ~) must b~ considered along the airplane centerline on all spars carrying across the centerline. T~erever three or more beam-like members intersect (e.g. paints 5 and 8 of Fig. A23.2), their bending stiffnesses will react against each other and hence couples in two planes (M, N) must be considered at these points.

.. Couples Q:a" Q 0 and Q!I' in the plane of the oblique Leading Edge spar, are included hereby, since any such couples may be resolved into M:a,. N::.,; Mo. N e and Me. N!I components. See "Transformation of Matrices for Oblique Beams". below.

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

A23.5

la)
Fig. A23.5

(b)

element stiffness matrices involves first f1ndir~ their flexibility influence coefficients and t~en invArting these ~atrice5. By way of illustration, the calculation of the influence coefficients is carried out in detail below tor spar 4-7-9-10. (wnile the calculation is ef:ected here by the Least Work matrix method, any of numerous other me thods (Moment Area, Elastic Weights, etc.) may be employed at this stage, Should the analyst see fit.) Fig. A23.4b shows the forces applied to spar 4-7-9-10, corresponding to those considered pertinent for the wing (see above). For the umuediate calculation of influence coefficients, forces p. and p~o will be considered as fixed end reactions: deflections thus computed will be measured relative to the beam element ends. To distinguish the deflections thus obtained they will be referred to as relative deflections and will be denoted by the lower case Greek letters 6 and ~. Fig. A23.7 shows the internal generalized forces for this spar.

to be related to the tour pertinent displacements at its ends. That is, we seek the ralatl0

between the loads (reactions) Nand P and the displacements and~. The first column of the above matrix relates the four loads to the rotation at the left end with ~l! = s , = I l l ! = O. The situation is shown in Fig. A23.5b where it 1s seen to be a fa~11iar single redundant problem. ThUS, to obtain the complete stiffness ~trlx above requires the solution ot four Simple redundant problems, one for each column at the matrix.

ia

For Rib 8-9, on the other hand, we seek the relation

(.'9 -",
Fig. for

q,

==~" s ,
10

A23.7 Internal Generalized Forces Used

in Influence CoeUicient Calculation

Spar 4-7-9-10

as may be seen from Fig. A23.5, which depicts the calculation of the first column of this matrix, th9se are statically determinate calculations.

Member fleXibility coeff1cients for the tapered beam segnents are computed next using the formulas of Art. A7.l0 of C~~pter A7. Collected in matriX form these became (E = 1)
3.213 1.623 6.841 1.864
0 0 0 0

[ail] =
Fig. A23.6

1.623
0
0

1.864

8.55
4.923

4.923

21.54

Obviously, one can proceed in the above fashion to work through all the elements of the s~ructure solVing for the stiffness matrix of each by application of these Simple arguments. 3ecause a large number of beam stiffness calculations of the above type appear in this prOblem, an alternate (but equivalent) procedure was developed for systematically treating these members. This procedure 1s detailed in the follOWing section. Beam Element Influence 80efficients The ~ethod em~layed to ~btain the beam

Next, unit values of the loads N., p~ and Pa are applied succeSSively to obtain the unit stress distribut1on:

I~
1

N,
1

P,
0

p.
0

~imJ =

2
3

.567
.333
0

-16 -8
0

-8 -16
0

Finally, the following matriX triple product is formed to give the i~:luence coefficients (ct. Eq. 24, Chapter A7).

A23.6

STRUCTURAL

ANALYSIS

OF

DELTA

WING
Stlf~ness

Co~plete 3ea~

Eleffient

Matrices

[1~~::0
-119.9

-141.6

1 .J.O:I.::'

2566

27715

. . ;::.;::. 2"
3104

0 __

'J
?~ -(3)

o~J

written out as an equation; For Spar 4-7-9-10

The stiffnsss matr~x given above r91~tes the 1eflections relati,e to the beam element ends to the loads a?plied to the bean element, end reactions exclUded. It 1s desired next to obtain the "compjete" beam element stiffness matrix in which ~bsol~te beam deflect~cns are related to all the loads applied to the bean element incl~ding the end reactions. Consider deflections ~irst. It is desired to transform from the relative deflections of Fig. A23.3a to the absol~te deflecticns of ?ig.
A23.8b.

5~

~ .}

llo.20 -141.6 -119.9

-141.6 _1l9'J~{N.} 2776

2566

6.

2566

3104

P;

Note again that for this bea~ element the flexibility influence coefficients just computed are for relative deflections (note symbols used). The end transverse deflections (A~, A~o) and the end transverse reactions (p. J P~o) do not appear i~ these results. Beam Element Stiffness Matrices (Relative Deflections) The next step involves finding the beam element stiffness matrices tor relative deflections by invert~ng the above influence coefficient matrix. The results after inverting will be of the form

---

Lm-r

-==L

6 m+ r

(al

(bJ

DATUM

Fig. A23. S Geometry of la) Relative and (b) Absolute Displacements

--

From the geometry

+

c~

the figure, one

r2S

'c'!ll+s - ~ L

6m+r = -

L - kn+r
L

where ~kJ may be called the Wrelative stiffness matrix. Specifically, for Spar 4-7-9-10

<Im+.

_ iL + iI,n+. - iI,n - p+s L

{P"

N .) [,3554 r , = .02303
- .005306

.02303
.003013 -.001600

~:~~~:~~{::) 00143~
6.

For comparison with the matrix form to follow, these equations are rewritten as

-T
o",+r

+ + D.:n+r

T
I...m+r
L

~+s

We note that the inlluence coeUicient matrix to be inverted does not appear to be "well conditioned" in the sense of Art. AS. 12 (see p. AB.28). The situation is more apparent than real, however, and arises because of a difference in units among the coefficients of Eq. (3). Thus, the appearance is readily altered by "scaling" (factoring out appropriate constants). For instance, we may write

I...m+r - L
L

~+s

<Im+. =

- L

+ +

",,+.

:: (~o'P_} ( IuJ. ~:::: -::::: -::::~ 31.0~

6./,0
l1.96 25.66

~+.

lOP.

Scaling in this fashion does not actually enhance the condition of the matrix (which is basically related to the size of the determinant of the matrix) but it does permit one to assay the problem better.

In matrix form these equations are written so as to prOVide a transformation fram absolute to relative displacements. In matrix f0~ the index (subscript) is understood to inc~ease monotonically down the column, and inasmuch as the grid numbering scheme gives this same property to the joint numbers on an element one has:

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

A23.7

- - - - - - - - - -

(5)

where
1
0

~
where

!'1m

Nm~
Pi:!

=
WITH END REACTIONS

[s]

~
I

-p~

!Nm 'Im1
t
0 I I
I I
I

-TA1THOUT END REACTIONS

(6)

r ------1-----7-----I

-L"

r J

l2:----~

1 :T
I
r

[TJ =

r r r r

Q
r

I
I I

-L1

!.m-L

I
I

I
I
r

[I]
~---~

I I

I
I

-!.m L

I I I I

~lm - L)/.!J J -----------,-I

t:
o
J

! ~------:J
J I
I
J

I
I
J J

0 0

,
0
r

I
I I

J I
I

---- -- ----1-- -----T

I I I

I I

[sJ =

[I]

r
I I

I
I

I
I

a __
~

r t

1.
I

I 0

Here [1J denotes a unit matrix.- The matrices are shown ~partitioned~, those portions which transform the end rotations being shown separated frcm those which transform the transverse deflections. The transfor.nation matrix C:TJ in Eq. (5) is not square, haVing two more columns than rows to allow for the added two deflections, one at each beam end. Consider next the loads. The applied loads may be transformed, using the equations of statics to yield the end reactions.
~m+r
M --Lm+r-

1 T

I 1

o :[2;-~:

lm/L

I 1 I T 1

The above transformation matrix 1s not square, having two more rows than columns, these extra rows yielding the end reactions.

[sJ

PrOVided the Sign convention is observed as originally adopted, and provided the grid numbering scheme 15 such that joint numbers increase to the rear and outboard, then the above transformation matrices will apply equally well to ribs, spars and oblique beams: replace N by M or Q, ~ by Q or ~and ~ by 9 or T.

IIJ-

I .;;==~~~~~~d

M m+ s

- L

----Fig. A23.9

P m+ s

For the case of beam elements haVing couples applied along the beam (e.g. the L.E. spar, Fig. A23~ simple modit1cations of these transformation matrices are necessary. The relative stiffness equation, Eq. (4),

From Fig. A23.9 one has

}'-. 1 -T-T

m~r Pmer- (L - Lm.+r) - "L+ s

Z a-r P!l1+r Lm+r + N."L+ s

'L

Pm +s =

-?-- t

is the relation be~Neen loads and relative deflections at a beam element. Substituting from eqs , (5) and (6), into (4), one' obtains

In matrix form, the general expression which introduces the end reactions in te~s ot the other applied loads is

~i

[sJ [kJ[TJ

tll ----ABSOLUTl:
DEFLECT IONS

(7)

WITH END
We have used and { } to denote row and column matrices, respectively (see Appendix A). ThUS, is a row of zeroes.

L .J l..P - - - - - 2.J

REAC~IONS

on 't is introduced here as a symbol for the relative rotation of a joint in the plane of an oblique mem~

A23 8

STRUCTURAL

ANALYSIS

OF

DELTA

WING

Eq, (7) is the relation between absolute deflections and the co~plete loading on the element. ve let

For Spar 2-5

be the "complete" Note that it must be if

ele~ent

stiffness

~trix.

[SJ is
[KJ

the transpose of

G'J,
[TJ

n[=
Pa PI -

-.1218

,1218

-.1218 .1218

.004208
-.004208 -.08054

-.004208

.K"

1.555

.004208 .0605?

-.08054 .02054 2.317

L~'~n
2:.",
~~

~"

as For Spar 3-6-8

.'

Is to be s~~etric.

To continue the illustration, the and [sJ matrices for Spar 4-7-9-10 are now written. For this member the rotation at station 10 Is not considered and the couple at that end is

. .

p.
p.

P,
:;

[sJ

zero, hence the last row and column of [TJ and are omitted in writing them from the general expressions. Equation (5) written for Spar 4-7-9-10:

I
"

~~
c-

-.02423
.00275';'

.:265 _.00.:;.1:2

_.02228
.J013~3

-.03423
.1055 _.02::..29

_.::;0-'1.1:2
.:J01:X:S .01725

,,-<> . '''''' J,.....

_ ,'O:J3CC'6
.OOl~C8

_. ::;03066 _.05.03...

':;':;: t::
.,,3,,0~ ~.

OJ
'~1"''''''

-.2534

. :)3908

1.33

1.

For 1/2 Fuselage Beam

n
0,. -

[_oo,~ 0 _.667
0

a
1.0

6.

-.333

-.333 1.0 -.667

a a

OOOO~

~,
il, il,
/,

M, P, P,

158.5 5.560
.2652

-4.312 -.2385

-4.312 .1557

.,
D

-1.533 .09993
~

5.580 -1.533 .2652 -.2385 .09993 -.017l8 .4343 -.2701 .07429 ............... -.1027 -.2701 .0:::;
I c.~
~

8. 6
-,

-,

P,

" ........ .04560 -.01713 . 07429 -.J:v':::'!

,,,
t

/'"

For Rib 5-6-7

Equation (6) , written for Spar 4-7-9-10:

M.

r-

La

a-

P, -.006945 - .667 - .333 P, = a 1.0 a a P, a 1.0 P"

The matrix

{::)
"~l~
.0.:65
.C~2238

rJ 0
3 '. ?
?

-.G253

.0275

'"

-.02253
.0276 :'

.0,536:0
_ .OC12C4 .::J,3570

_.00:2:>. .CClS3Z

-.005066"

_.c,c:s

.v,"'" v _. '~::~rl o :o:;;~ :: J

I
c'
I '

.006945

For Rib 8-9

.333

- .667

[KJ

is now formed per eq . (a)

For Spar 4-7-9-10:

-.01605

M ') = 63322 (
P!l

-.006910

-.006910
.006910

?g

,0230'3
-.001637

_.C:::::306 .0,6252 _.001600 .OOH38 _ .::J,'i-B3Z

rJ IT'
p. ?

i. 1

For Leading Sdge Spar ......

I
.3>178 -.01846

-.01606

.0~,;g52

.02303

-.001637 .0,:5252

.ooaoia
-.00160::J

p.

-.006306
-.001622

p,.

.0.:65

-.0~46~
.::J,2229

:: j
:: 0

0,.:22:3:0

.C1:;'~

I .-.l~O

I'~

P, p.

-.01946

.C,~';'90 -.O,d901

::J -.0,394:1

-.':;1756) .0030<34 -.01.0.5"

l.~';'

In the manner of the above illustrative case one goes through the calculations for the remaining beam elements to obtain:

,01".;..s -.0,.54,,,0

.c.saes
.OC3~24

c I o!
,0
i

P.
.';'1&0 ::J -.01766 .21...07 JO~ 1001
I
I

.3C16
--q,

- .01259

.3016

. cc:;:;sz I

The notation . ON means there are N zeroes following the decimal point and preceeding the first significant figure,

e. g 0.. 276 '" . 0000276.

.. Because of its Size, this matrix relation (and some to follow) is written in a condensed tabular form. The correct interpretation should be obvious.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

~ransfo~~tiJn

A23.9

of Matrices for Oblique

8ea~s:

T~e complete stiffness matrix of any beam element not either ~arallel to or nor.nal to the strea~ise direction, such as the leading edge spar in the present example, requires an addiI1 8 = sin J\.. Q.. .707 Q.. tional modification to yield expressions relatand ing couples and rotations about pitching and rolling axes, rather than in the plane of the Na = cos A Q,s = .707 Q,s member. Such a modification is readily made. Similar relations for the other couples along In the present example, a pitching rotation this member lead to the matrix transformation of joint 5 (9 s) results in a rotation of jOint 5 for this element's loads: in the plane of the leading edge spar given by 9, 707 9" ~~ = sin A .707 0 0 0 0 0 0 11, where A- is the sweep angle of the spar. Like.707 a N, 0 0 0 0 0 wise, a rolling rotation of j~::nt 5 (~,) has a Q, P, 0 1 0 0 0 0 0 rotatlor~l component in the plane of the leading P, edge spar given by P, 0 0 0 1 0 0 0 P, = cos .'- 9, = .707 ~, P, 0 0 0 1 0 0 0 P, Tten, when joint 5 experiences both pitching and 11, = 0 .707 0 0 0 0 0 rolling rotations, the total rotation in the Q, N, 0 0 .707 0 0 0 0 plane of the leading edge spar is Q, rn 0 0 0 .707 0 0 11, 0

Next, consider resolution or the in-plane element cQuyles. Counles Q~, Q.8 and Qa have components in the pitching and rolling directions. ~or exa~ple, at joint 5
bea~

T:

J. II

.7079 11 + .707 ~,

s~ilar ones for joints 1 and 8, when put in matrix fOIiJ, yield the ~atrix transformation for the displacements.

This last equation, and

N,
P"

0 0

0 0

0 0

0 0

0 0

.707 0

0 1

T"

.707 .707

0 0 0

0 0 0

0
Q

i 9,

r:: 1
,,
, .i,o)

o
0

c
0

o
0 C 0 0

c
1
Q

0
0

, -,
e,

~l ".

0
0

0 0

.7C17 .707

0 0

o
0

707

.tsn
0

o l-

I.

'.

",

" I'"::'0

The above two matrix transformations - one for displacements and one for loads - (and note that they are the transpose of each other) are now applied to the complete stiffness matrix equation for the leading edge spar as given in tabular form in the last sectIon. The operat10n involves premultiplying by one transformation matrix and postmultiplying by the other. The result is:

Leading .
9,
M~

Edg~

SDar cv namDlete
Il, Il,

Stl~Pness ~atr1x

_.,

(~rans~ormed)

~,
.4487
.4487

",
-.01376 -.01376 .03;:)4-90
0 -.01248

9,
.2094 .2094 -.01248 .002534

~,

9,
0 0 0 .006469 .1507 .1507 .3889 .3889 .002391

~,
0 0 0 -.00890 .1507 .1507 .3889 .3889 .002391

.4487

.01376
.01376
- . 3 5490

0
J
0

.2094.2094

Il" 0

N, P,
p.

.4487
.01376 0 .2094 .2094 0 0 0

0 0 0 0 0 .002391 .002391

-.01376 -.01376 0

-.01248
.009947

.01376 -.0 30490 .2094.2094 0 0 0

.0 39428 -.0 33941 -.0 33941 .002534.002534-.00890 -.00890 0

.0~4422

.002534 -.00890

P, M, N,
11,

.009947
.7168

.006469 -.04,4998

.009947 .009947
.006469 .006469

.7168
.7168

-.01248
0 0 0

.7168
.1507

.1507 .1507
0

N,
P"

.1507
0

- .0, 4998

.0,4998

A23.10

STRUCTURAL

ANALYSIS

OF

A
~

DELTA

WING

Torque Box Stiffness Matrices In the case of the torque boxes, the stiffness matrices may be computed directly (they are dete~inate). The following approximate analysIS is suggested as being satisfactory for most torque boxes. A ~ore detailed analysis may be In order for boxes with extr~me geometries. An alternate method is presented in Reference (2) and an appropriate discussion may be found in Reference (6). A completely different treatment at the cover skin is given in Reference (4). Consider the quadrilateral box of FIg.
A23.10

= P~~brn
u,;

Lnbm

+
Ln

Lmbn

Since it is only the deflection of one corner (say, point m) which 1s to be related to the corner reactions, the box is now rotated about the effective root axis to reduce on to zero. The resulting total detlection of point Jl is

.,

:.wTOTAL -

_,,-+Lm 5
4Il

1:;-

n
+ "-'-'TIl

= P:nLm (l.nbm. + Lmb n) It GJ'""" \" Ln n

Re-solvlng this, we write

T
I

(r~

Ln -

"'T

L / 2 - - L / 2 ' --,. _ _ .Ii ___ . l - +bn

'

- :

GJ

Lo

'<n)

Pm = Lm(bn + bml (t-nom + Lmbnl Cn>rCTAL

This last expression relates the load at point m to the detlection at that point, with the ather three corners undetlected. 7he reactions developed at these Jther corners are now found in terns of Pm USing the equations of statics (s~ation ot forces and summation ot moments- about two axes). The r-esuj t , expressed in matriX fo~ is

~s,
L......

t~Axia \-----.lbm : (m)

\ .......
I

,........,
Equivalent , Root

w"

'

Representative Section

Fig. A.23. 10

For purposes or this immediate analYSiS, the box 15 considered cantilevered from one end, such as r-s. An elastic axis (e.a.) is assumed to exist midway between the long sides and an "etfective root" 15 employed. The torque on the box is

where

A, = Lnbm
An (assumed) average rate of twist, Q, Is computed approximately by using the GJ at a representative section halt way along the box.*

+ ~bn +

Ln(b r
+

bs (Lm - Lnl os)

A.

"

-(Lnbm

Lmbnl + br (lm Ln (or + b s )

- Lnl

Q = O~ = Pmbm O~ Pnbn
Then the deflection
On

On

and ~
I.

are given by

It now the corner pOints have absolute displacements Om, dn, dr and tis, one can find from the geometry of the unit that the total relative displacement of point m may be writ~en

= QLnbn = l'mbm OJ - Pnbn -nbn

Cn>rarAL =

A,

Om
By

=QLmbm = Pmbm dJ - Pnbn

~"'1Il

h ..

~ing moments about the effective root, Pn 15 found in terms at Pm. and then eliminated from the above equations to give

On

= PmLnbn
OJ

Lnbm

Lmbn

Ln

When these last three equations are combined to relate the four corner reactions to the absolute corner displacements, the torque box stiffness ~trix 1s seen to be t~e square, symmetric matrix,

* Calculation of. this

GJ is discussed later. The symbol "s'' used. in these few equationa bears no relation to that appearing in the main body of the method.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A23.11

= (Lnbm+l<nbn) GJ Ln/Lm (bn+bm)

1 -L,n/Ln
A,

~ ~ A, ~
----(9)

For box 6-7-8-;) GJE=l = 34.3

Finally, the stiffness matrices for the three boxes at the example become: (c r .eq, 9) For box 2-3-5-6
1 [KJ = 0.000898 -1 -1 1 -1 1 1 -1 -1 1 1 -1 1 -1 -1 1

A, In the above derivation the torsional stiffness OJ at a representative section is used. The stiffness is obtained fram Bredt1s equation tor the twist of a single cell thinwalled tube (Eq. 18, Chapter A6)
g _

-~

Ie.:
[KJ

For box 3-4-6-7

1 -1 1 1 -1 -1 1 1 -1 1 -1 -1 1

fram which, by definition

GJ

=0.000482

-1 _1 1

Here A is the area enclosed within the tube cross section by the median line of the tube wall and the integration is carried out around the tube perimeter (index s gives distance along the perimeter). For the torque boxes encountered in delta wings it is probably satisfactory to neglect the ds/t contribution from the vertical webs, it being small compared with the corresponding contribution from the cover sheets. In the example wing three boxes (2-3-5-6, 3-4-6-7 and 6-7-8-9) are to be used. These boxes are each 48 inches square in plan and have average depthS (assumed here to oe the unlfor.n depthS ot the representative sections) at 7.26~, 5.32" and 4.27~, respectively. Fig. A23.ll shows the assumed representative cross section of box 2-3-5-6 and its OJ calculation.

For box 6-7-8-9

1 -1 [KJ = 0.000310 -1 1 -1 1 1 -1
-1.

1 -1 -1 1

1 1 -1

A23.5 Complete Wing Stiffness Matrix

ie ,
~

l_ t

= . 051 ,

The stiffness matrix tor the composite wing now may be obtained by forming the sum or the complete stiffness matrices tor the beam elements and the torque boxes. For this purpose a large matrix table Is laid out and entries fram the individual stiffness matrices are transferred into the appropriate locations. Wherever multiple entries occur in a box these are summed. Betore the large wing mata-tx is laid out it is necessary to observe first that the matriX which would be obtained as just indicated would be singular, i.e., its determinant would be zero, and hence it could not be inverted to yield the flexibility influence coefficients (see Appendix A). This condition arises due to the tact that the equations represented (19 in number in the example problem) are not independent; three of these equations can be obtained as linear combinations ot the others. That there are three such interdependencies may be seen from the existence ot the three equations of statics which may be applied to the Wing (summation at no~l forces and summation at moments in two vertical planes): hence three ot the reactions expressed by the structural equations in the matrix may be found

48"

A: 7.26 x 48
~

=348 in

..
~

f t:= 2 x. 051 " 1. 88 x 10

Note: G

:= 2:6

4(348)8. 1 . .. GJE=I" 1. 88 x 10~x 2.6" 99.3 lb.fn.

(ds/t contribution of vertical webs neglected)

Fig. A23. 11 Calculation of GJ for the Representative Section of Box 2-3-58

In similar fashion one finds: For box 3-4-6-7

A23.12

STRUCTURAL

ANALYSIS

OF

DELTA

WING

tram the others by the equations of statics. To remove the "Singularity" rr-cm the stiffness matrix it is only necessary to droD out three equations - achieved by removing :hree rows and corresponding columns (so as to retain a symmetric stiffness matrix). The act ot removing the three equations selected is also equivalent to assuming the corresponding deflections to be zero. In this way a reference base for the deflections 1s also established. The choice of reference base is somewhat arbitrary, but, following a suggestion of Williams (7), a triangular base will be employed as shown in Fig. A23.l2.

singular. However, if this ste) ~s carr~ed au: one fines that :he resultar.t inf:~e~ce coefficients are those for- onjv t.ne "':e.l.::'-wl:-,: acting alone and supported by ccr.s:r~in:s as ass~ed above. To account for the ~resence o~ the other ::al! of the vnng , i t is necessar-y to speCify additional geo~etr~c con1:tlons along the airplane centerline. This ste? 15 acconplished by assumir~ the following deflect lens zero (ellmir~ting their corres~ondlng rows and columns from the matrix): for symmetric loadings (zero lateral slope or rolling rotation along the alr,lane centerline) for antisymmetric loadings (zero transverse deflections and pitching rotations along the airplane centerline) It will be seen that in both cases an addItional 3 equations are eliminated from the 16 x l6 matrIces reducing theo to 13 x 13'5. (In.general, there is no reason why the matrIces :or the s~etric and antis~etric cases have to be the sarne size. A rotation i~, for instance, would be zero in one case
only. )
~rltten below 1s the 13 x 13 wlr~ stiffness matrix for ~he ant1syremetric case (~~ = ~~ = ~~ = ~~ = ~& = 9~ = 0). As explained earlier, each entry therein is the sum o~ all corresponding stlffnesses for all elements ~eeting at the point. A typical multiple entry occurs for joint 0 - (row P /I' column 6/1)' These comprise:

5'

.~

'5

Fig. A23.12 Deflection Reference Base

Here the deflections at paints land 5 are zero, fixing the reference triangle, since the point corresponding to 5 in the other halt at the wing (say, 5'), will have zero deflection also due to symmetry. The third conditi In is applied to point 2: pOint 2 will be assumed to have zero rolling rotation (~~ = 0) for symmetric wing loadings and to have zero transverse deflection (6 8 = 0) for antisymmetric loadings.Hence the following equations (rows and columns) are to be omitted trom the wing stIffness matrix:
Pu Po'

N.. ~,
P"

P ., P"

} ",) ", ",

", ",

0.007178 0.001832 0.000310 0.000898 0.000482 0.01070

from spar 3-6-8 from from from from

':'OTAL

tor symmetric loadings

for antisymmetric loadings

rib 5-6-7 box 6-7-8-9 box 2-3-5-6 box 3-4-6-7

The 16 x 16 wing st t rfness matrices thus

obtained can now be inverted as they are non-

- The antisymmetric loading pattern is one wherein the wing is loaded equally, but in opposite sense, on corresponding points o( the two wing halves. Any general loading may be resolved into the sum or one symmetric and one anti. symmetric loading.

.L ;~~,\"~<~;~;i'/::,;;9~~;~~; g~r:~/~~:;:;;t%;i?r:.~~L{~~~;~~i:::~~~;~~<::;~~~V~} /',l~?T:~JL~;~~~j~

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A23.l3

[:.::nnJ ' . . ing

P,
p.
p. p.

Stiffness Matrix, Ar,tisymmetrlc Case

(Note:

Voids denote zeroes)

",
.01070 .001420 .003376 .0,310 .02760

'.
.004066
.G~310

",

'.
.O~310

-.GOl<;ZO -.003376
.0~310

'"
-.0~4632

" " .02760 !

I

I,

I,

j,
.1065

I.
.02303

I.
.009947

I.
-.05634.045.:::>5
,

_.00191 .001892

.0~2238 -.0~066i

.002604 -.0,4537 .009947

-.0~441

-.0~4537 -.0~4998

.009947!-.o,441 .006910 .002391 1.555 .1507 .7211 .2094.1507 .2094-

-.02228

-.00191 .0,2238 - .OOS06E

-.005306 -.001662

."
0

-.0~~998 -.0~4632 -.0~2729

M, M.

.7168
.1507 .2094 4.302 1.5s.:; 3.256

.002391 .1507 .3889

.006910

.002391

N.
N,

.4487

N,
N.

.1065 .02303

-.02223 _.005306 -.001662 .009947 .7168 .002391 .1507 .1507 .3889 .2094 1.555

-.2334 .3554 3.0338 .1507 1.719

N, N,

.05634

.04555

-.2834

.1507

m:::

'.WG FLEXIBILITY 11ATRIX:

The Wing flexibility influence coefficients are now found by forming the inverse of the above matrix (see Sq. 2). An automatic digital computing machine is the essential tool for this step. The details of the procedure and techniques employed in forming such inverses are not of concern here and only the result 1s presented. It is assumed that this phase at the work can be handled with dispatch by experienced computer personnel.
s~ifrness

~AmnJ

' Wing In!luence Coefficients,

0

Antlsymmetrlc Case

..

P,

p.

P,

P ~o
2"'19 537<5 7692 13,610

n.
-17.14 -16.32 -36.09
-~a.09

M,
-14.S9 -38.23 -25.10 -32.32 -131.0 .09551 .2701
.02~25

N,
7.750 a.610 17.57
1 18 01

N,
_.113...0

N,
-15.77 -19.96

N.
-2.3.52
_70.~

N.
.5092 -.1342 -1.555 -.5052 -.7S<06 -.201.2 _.07823

N.
-14.22 -24.2.2 -66.22 -67.49 -123.3 .4227 -.2044 _.2198 -.01744

.>,

" '. '. , "

I, t

73';'.3 988.7 1386

15G4 Z"'lS

se8.7
2'~03

1386
2118 4001 4"'61 7692

:,';644030

.04SS1
.5620

2118
~30

4751 5-558

-23.72
-34.28
-~7 .C4

-30.20

.1::.26 .2836 .07'273

.~73a

-6:5.70
-26.56 .3438 .6358 -.1592 .'::614-

-~Q

53,6

13,510 2.5,670 -;:>a.69

27.75
_.~77g

-17.14 -18.32 -36.09 -33.09 -58.69 1.<25

-IL5!? -33.231-25.10 -32.32 -1.'31.0 .09.551

.0.'303

-.00806':;' .02523

.2976
-.1527

7.'7"60 8.510

17 .57

13.01

2"'.75

-.47-;;) _.0-2'73

-.00a06... 2.,513

-.131:
-.17.;.;3

-. ::'~40 .1825 .2836 _1."3.7"7 -19.96 -2.3.72 -34.28 4 7.04 .;503

.--~O;2

.0';"':':>1 .,')620

.04-38
-.1.':27

.2S54 .00834'::'
-.17.;.3

.80834-7 .0)9086
.~,

.2[;76

.;:---'"

.02309 I .1::;28

I ~.

-2.3.;:;2 _70.:'0 -30.':00 _65.70 -25.56 .-34:38

-.1342
-l.~55

.6338
-.072.23
-.2Q44

-.1592

'OJ sees .-~61'"

'3.271

1-. ':]0251-+

.)1...20
.V"...:>.::....

I, I,

- .50.'52 -.-846 -.2012

- .1311
-.2198

- .-::'<:309 -.0025H . ...a21

".

~~

-H-.22 -24.22 _'36.22 -'57.49 -123.3

..;.2"-

-.017;;.4 . 1~2S

-.014,,0 (.04,82...

2.079

" In spite of the exercise of great care and much ingenuity, problems of such great size and such a nature will arise occasion-

ally as to defy satisfactory inversion. For these problems such clever physical concepts as "block" solutions of portions of the structure are available. The interested reader is referred to a discussion of Ref. 4 and to techniques of block and group relaxations in the literature on Relaxation Methods (e. g. Ref. 8).

A23.14 A23. 6 Wing Internal Stresses

STRUCTURAL

ANALYSIS

OF

DELTA

WING

following t~e calcul~tion of the Ning influence coefficients, a relationShip is available between the a~plied wing loads (assumed given) a~d the wing qefo~tlJns. In a sJ1nbolic fashion,

Equation (10) provides t~e cesired relation between ~xternal loads applied tc the wing and t~e forces acting on a speCi~lc eleme~t. These ~crces on the element might be locked upon as those necessary to hold the element in the deflected srape conforming to that of the wing. Of course, with these element forces known the finer details of the stress distribution within the element are readily found by standard techniques. Note that the ~atrix [KJEL~\~ in Sq. (10) will have to be "blown up" to an accr-opr-tata size be Io re it may be pre:nultiplied:mto [;I,.lln:G' This enlargement is accomp11shed by the inserticn of columns of zeroes at each column location corresponding to a wing defo~atlon which does not affect the speci~lc element under consideration. For illustratlcu, the indicated steps are now carried through for spar 4-7-d-lO. Referring to previous work we find for tt.e element matriX:

Earlier in structural the forces (complete)

the analysts, defor.nations of a el3ment (beam or box) were related to acting on that element by the element s~lffness ~trix:

~~}
---

EL=rr

Now since the deformations of an element must conform to those of the wing,

b
~

.8m lm }
EL=rr

={.1m} 8m
~
W!m

---

~.
II,

.::.,
-.016013
.O~~952

o.

.3554
-.0160(

.0230:3 -.005306 -. ;:01562

-.001537
.0~52.52

".

"

Hence, the previous equations may be combined immediately to y1eld

[KJ: ,

p. p.

.0,155 .0.2238

.023C .3 -.001637

.003013 -.J01600

P, - .oor .306
.

.0.5252 -.C016CO
':""Q

.001438 -.:J.4632
-.0.';'632
.

P.,

-.u....J,662

a . --

0 2-->8 ;,;: ...

J~2229

Expanding (and rearranging to agree in numerical order with the wing matrix),

A, N.
0 0 0 0

AT
.0230:3 -.001637 .003013 -.001600

A,
0 0 0 0

A.

A"

e,
0

9,
C

~.
0 0 0 0 0

~.
0 0

~,
0 0 0 0
0

~.
.3554 -.01606 .0230:3 - .005306 -.001662

~,
0

~.
0

-.005306 -.001662

p.

rKJ
J

LL'"1'!ENT

- E p.
p.

.36252
-.001600

.0,,165

a
0 0
0

0 0 0 0

a I
0 0 0

0 0

.001438

P"

'3 2238 a -'3 4632

.3 -'3 4632 '3 2229

223 8

a
0 0

a
0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

A23.1!i

Then multiplying out with [~~WING per Eq. (10l, we get the five forces en spar 4-7-9-10 (see Fig. A23.4-b) in ter,ns of the 13 wing applied loads as:
p.

P,
P,

p.

N.}
p.

-.C~367

-.0017'7

.o, 794
1212

.0,304

.002694 .0.1736 .0.3834

.0.2706

0 .0.24-9

-.0.8n
.00145
-.00299

1.00
-.02696

-.0.555

p ..

-.0304
.06294

-.1945
.3735

.2190

.4985 -.2915 _.9174-

-.0,320 -.001257 .0,180

.0,5408 .001850

p.
p.
p.

,2098 -.'0191

-.0,330 -.0.625

-.00375
.002905

-.1650

.05466 .3787

-.0,107 .0.975
-.0,111 -.0,682

.0,1133 .0.486
.0.348

M. M, .001451 _.00281 -.0,135 -.001107 N,

.03148 .0,1705 .007071

-.0.687 -.004:34

-.01364 .03347 -.07,,65 .7092

-.0.1202 -.0.5815 -.001665 .0.3325 -.001618

, N,
N.
N.

s:1. :::u:

N,
N,

WHIG

Each column ot torces in the above matrix must satisty the equations of statics on the element. A variety of checks on the accuracy of such a result are thus available.
The student will find it instructive to study carefully this last result to observe in what manner a load applied to the wing appears on this spar element. For instance, one sees that of the load p~ ~ 1 applied to the wing, .3735 goes onto spar 4-7-9-10. Examination ot Fig. A23.l reveals that the remainder, 1.00 .3735 = .6265 must be taken up by rib 5-6-7 and the torque boxes 3-4-7-6 and 6-7-8-9.
REFERENC!l3

4) Turner, l1. J., Clough, R. 1,./., Martin, H. C., and Topp, L. J., Stiffness and Deflection AnalysiS of Complex Structures, Journal or the Aeronautical Structures, 23, September, 1956. 5) Kroll, W." Effect ot Rib Flexl'bili ty on the Vibration Modes or a Delta Wing Aircraft, Institute or Aeronautical Sciences, ?reprint No. 585,
1956,

6) Wooley, Ruth, Check at Method ror Computing Influence Coefficients of Delta and Other Wings, National Bureau of Standards Report 3655, 1954.
(Avallab1e as ASTIA No. AD40866).

1) Argyris, J. H. and Kelsey, S., Energy Theorems and Struc~ural Analysis, Aircraft Sr.gineering, Oct. 1954, et. seq. 2) Levy, S., Structural Analysis and Influence Coef:icients for Delta ~ings, Journal ot the Aeronautical SCiences, ~, 1953. 3) Schuerch, H., Delta Wing DeSign Analysis, Society or Automotive Engineers National Aeronautic Meeting Preprint No. 141, September IG53.

7) Williams, D., Recent Develonments in the structural Approach to Aeroelastlc Problems, Journal or the Royal Aeronautical Society, 58, 1954 (see also, Aircraft Engineering, 25, 1954) 8) Southwell, R.V., "Relaxation MethodS in Science", OXford 1940.

Englneerlr.~

A23.16

STRUCTURAL

ANALYSIS

OF

DELTA

WING

AIRCRAFT WITH "DELTA" WING SHAPES

CONVAIR SUPERSONIC F-I02 INTERCEPTOR

CONVAIR SUPERSONIC B-58 BOMBER

",,-"

~;::~~~~~,]t4?~f~"'",.
~".~

~~

COMMERCIAL TRANSPORT FOR THE 1970s? -- This is an artist's conception of a passenger transport of the future -- cruising at three to five times the speed of sound at 60,000 feet or higher. This is one of hundreds of configurations considered by Convair Division of General Dynamics Corporation at San Diego, Calif., in its studies of supersonic airliners.

.-~c

..-_,.

,-.,'~'

PART B

FLIGHT VEHICLE MATERIALS AND THEIR PROPERTIES

CHAPTER BI

BASIC PRINCIPLES AND DEFINITIONS

Bl.1 Introduction.

The flight vehicle structures engineer

faces a major design requirement ot a high

degree at structural integrity against tai lure , but with as light structural weight as possible. Structural faIlure in flight vehicles can otten prove serious relative to loss ot lIte

and the vehicle.

However, experience has

shown that i t a flight vehicle, whether

military or commercial in type, Is to be

satisfactory tram a payload and performance standpoint, major ettort must be made to save structural weight, that Is, to el1minate all structural weight not required to insure against failure. Since a flight vehicle Is sut~ected to various types ot loading such as static, dynamiC and repeated, which may act under a wide range or temperature conditions, it is necessary that the structures engineer have a broad knOWledge of the behaVior of materials under loading it sate and etticient structures are to be obtained. This Part B provides information for the structures design engineer relative to the behavior at the most common tlight vehicle materials under load, and the . various other conditions encountered in tlight environment, such as elevated temperatures.
Bl.2 FaUure of Structures.

Failure in a ditferent manner can resUlt from a structural unit being too rfext tn e , and this fleXibility might influence aerodynamic torces SUfticiently as to prOduce unsatisfactory vehicle flying characteristics. In some cases this fleXibility may not be serious relative to the loss of the vehicle, but it is still a degree at failure because changes must be made in the structure to provide a satisfactory operating vehicle. In some cases, excessive distortion such as the torsional tw1st of the wing can be very serious as this excessive deflection can lead to a build-up of aerodynamic-dynamic torces to cause flutter or violent vibration which can cause failure involVing the loss of the airplane. To illustrate another desgree of failure of a structural unit, consider a wing bUilt in ruel tank. The stiftened sheet units which make up the tank are also a part or the wing structure. In general these sheet units are des tgned not to wrinkle or buckle under airplane operating conditions in order to insure against leakage at tuel around riveted connections. Therefore if portions of the tank walls do wrinkle in operation resulting in tuel leakage, which in turn require repair or modification at the structure, we can say failure has occurred since the tank tailed to do its Job satisfactory, and involves the item ot extra expense to make satistactory. To illustrate rurther, the flight vehicle is equipped With many installations, such as the control systems for the control surfaces, the power plant control system etic , , which involves many structural units. In many cases excessive elastic or inelastic deformation at a unit can cause unsatisfactory operation ot the system, hence the unit can be considered as having tailed although it may not be a serious failure relatiVe to causing the loss of the vehicle. Thus the aero-astro structures engineer is concerned with and responsible for preventing many degrees ot failure at structural units which make up the flight vehicle and its installations and obviously the greater his knowledge regarding the behavior of materials, the greater his chances of avoiding troubles trom the many degrees of structural failure.
Bl.3 General Types of Loading.

A flight vehicle like any other machine, is deSigned to do a certain job satiSfactorily. It any structural unit of the vehicle sutters effects which in turn effects in some manner the satiSfactory performance of the vehicle, the unit can be considered as haVing tailed. Failure at a structural unit is theretore a rather broad term. For example, tailure at a structural unit may be due to too high a stress or load causing a complete fracture at the unit while the vehicle is in flight. It this unit should happen to be one such as a wing beam or a major wing fitting, the failure is a serious one as it usually involves loss of the airplane ~nd loss of lIfe. Likewise the COllapse of a trut in landing gear structure during a land1&' operation at 'the airplane can be a very erious failure. Failure of a structural unit :,ay be due to fatigue and since fatigue rat iure ,LS ot the fracture type Without warning indl,,' cations of impending failure, it can also prove , to be a very serious type or failure.

Failure of a structural member is influenced by the manner in which the load is

BL 1

81.2

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

applied. Relative to the length or time in applying the load to a member, two broad classitications appear logical, namely, (1) Static loading and (2) Dynamic loading. For purposes or explanation and general discussion, these two broad classitications will be further broken down as tollows:Cont i nuous Loading. Gradually or slowly applied Static Loading loading. Repeated gradually applied { loading.
Dynamic Loading

thus repeated loads are important In design of structures. Dynamic or Impact Loadlng. A dynamlc or impact loading when applied to a member produces appreclable shock or Vibration. To produce such action, the load must be applied tar more rapid than in a static loadlng. This rapid application at the load causes the stresses In the member to be momentarily greater than it the same magnltude or load was applied statically, that 1s slowly applied. For example, it a welght ot magnitUde W is gradually placed on the end of a cantilever beam, the beam will bend and gradually reach a maximum end deflection. However i: thls same weight of magnitUde W 1s dropped on the end at the beam from even such a small helght as one toot, the maximum end detlection will be several times that under the same statlc load W. The beam will vibrate and finally come to rest with the same end detlection as under the static load W. In bringing the dynamic load to rest, the beam must absorb energy equal to the change in potentlal energy or the falling load W, and thus dynamic loads are otten reterred to as energy loads. From the basic laws ot PhySlcs, torce equalS mass times acceleration (F = Ma) and acceleration equals time rate at change at velocity. Thus it the velocity or a bOdy such as an airplane or missile Is changed In magnitUde, or the directlon of the velocity of the vehicle 15 changed, the vehicle 1s accelerated which means torces are applied to the vehicle. In severe flight alrplane maneuvers like pulling out ot a dive rr-on high speedS or In striking a severe transverse air gust when flying at high speed; or in landing the airplane on ground or water, the torces acting externally on the airplane are applled rather rapldly and are classed as dynamlc loads. Chapter A4 discusses the subject ot airplane loads relative to whether they can be classed as static or dynamic and how they are treated relative to deslgn at aircraft structures.
B1. 4 The Static Tension Stress-Strain Diagram.

nnpact or rapidly applied loading. { Repeated impact loading.

Statio Continuous Loading. A continuous load ts a load that reaatne on the member ror a long period or time. The most common example is the dead weight or the member or the structure 1tselt. When an airplane becomes airborne, the weight ot the wing and its contents 1s a continuous load on the Wing. A tank subjected to an internal pressure tor a considerable period of time is a continuous load. Since a continuous load is applied tor a long time, 1t 1s a type ot loading that provides tavorable conditions tor creep, a term to be explained later. Far airplanes, continuous loadings are usually associated With other loads acting simultaneously. Static Gradually or Slowly Applied Loads. A static gradually applied load Is one that slowly buildS up or increases to its maxunum value without causing appreciable shock or vibration.. The time ot loading may be a matter at aeccncs or even hours. The stresses in the member increases as the load is increased and remains constant when the load becomes constant. As an example, an airplane which 1s climbing with a pressurized tuselage, the internal pressure loading on the fuselage structure 1s gradually increasing as the difference in air pressure between the inside and outside ot the tuselage gradually increases as the airplane climbs to higher altitudes. Static Repeated Gradually Applied Loads. It a gradually applied load Is applied a large number of times to a member It 15 referred to as a repeated Load, The load may be ot such nature as to repeat a cycle causing the stress in the member to go to a maximum value and then back to zero stress, or tram a maximum tenslle stress to a max~ compreSSive stress, etc. The situation envolving repeated loading Is important because it can cause tallure under a stress in a member which would be perfectly sate, it the load was applied only once or a small number or times. Repeated loads usually cause tailure by fracturing without warning,

The intormation tor plotting a tenslon stress-strain diagram ot a material is obtalned by loading a test specimen In axial tension and measuring the load with corresponding elongation over a given length, as the speCimen 1s loaded statically (gradually applied) tram zero to the falling load. To standardize results standard size test specimens are specified by the (ASTM) Amerlcan Society For Testing Materials. The speed of the testing machine cross-head should not exceed 1/16 Inch per Inch ot gage length per minute up to the yield point at the material

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Bl.3

and it should not exceed 1/2 inch per inch of gage length per minute from the yield point to the rupturing pOint of the material. The instrument for measuring the elongation must be calibrated to read 0.0002 inches or less. The information given by the tension stressstrain diagram is needed by the engineer since it is needed in strength design, rigidity design, energy absorption, quality control and many other uses. Fig. 81.1 shows typical tensile stressstrain diagrams of materials that fall in three broad classifications. In the study of such diagrama- various facts and relationships have been noted relative to behavior of materials and standard terms and symbols have been provided for this basic important information. These terms will be explained briefly. Modulus of Elasticity (El. The mechanical property that defines resistance of a material in the elastic range is called stiffness and tor ductile materials is measured by the value termed modulus of Elasticity, and, designated by the capital letter E. Referring to Fig. Bl.l, it is noticed that the first part of all three diagrams is a straight line, which indicates a constant ratio between stress and strain over this range. The numerical value or this ratio is referred to as the modulus ot El.eas't t c.t ty (E). E is.stherefore the slope of the im tial "st'ra fght portion --or--the stressstrain diagram and its numerical value is obtained by dividing stress in poundS per square inch by a strain which is non-dimensiona or E :: tiE, and thus E has the same units as. stress, namely poundS per square inch. The clad aluminun alloys rave two E values as indicated in the lower diagram of Fig. Bl.l. The initial modulus is the same as for other aluminum alloys, but holds only up to the proportional limit stress ot the sott pure aluminum coating ~terial. Immediately above this paint there is a short transition stage and the material then exhibits a secondary modulus of Elasticity up to the proportional limit stress of the stronger core material. This second modulus is the slope of the second straight line in ~he diagram. Both modulus values are based on a stress using the gross area wntch includes both c are and covering material. Tensile Proportional Limit StresS. (Ftb)' The proportional 11mit stress is that stress which exists when the stress strain curve departs tram the initial straight line portion by a unit strain of 0.0001. In general the proportional limit stress gives a practical dividing line between the elastic and lnelastic range of the ~terial. Th~ modulus of elastiCity 1s considered constant up to the

Ultimate Tensile Stress

_--7--

Proportional Limit (a) Material Having a Definite Yield Point (such as some Steels) Strain
~

Inches Per Inch

______-:::::-:=z.--.
Ultimate Tensile stress Proportional Limit
(b) Matertals not Having a Definite Yield Point (such as Aluminum AllOYS, MagneSiUm, and Some steels)

.002

Strain - Inches Per Inch

(c) Clad Aluminum Alloys strain - Inches Per rece

Ultimate Tensile Stress

Fig. S1. 1

proportional limit stress. TenSile Yield StreSS (Fty)' In referring to the upper diagram in Fig. 81.1, we find that some materials show a erarp break at a stress considerably below the ultimate stress and that the material elongates considerably with little or no increase in load. The stress at which this takes place is called the yield point or yield stress. However many materials and most flight vehicle materials do not show this sharp break, but yield more gradually as illustrated in the middle diagram of Fig. Bl.l, and thus there is no definite yield point as described above. Since permanent deformations ot any appreciable amount are undesirable in most s~ructures or machines, it is normal practice to adopt an arbitrary amount of permanent strain that is considered admissible for deSign purposes. Test authorities have established this value of permanent strain or set as 0.002 and the stress which e xi st.ec to cause this permanent strain when released from the material is called the yield stress. Fig. Bl.I shows how

B1.4

BEHAVIOR OF MATERIALS AND THEIR PROPE RTIES

it is determined graphically by drawing a line fram the 0.002 point parallel to the straight portion of the stress-strain curve, and where this line intersects the stress-strain curve represents the yield strength or y1eld stress. Ult1mate Tensile Stress (Ftu)' The ultimate tensile stress is that stress under the maximum load carried by the test specimen. It should be realized that the stresses are based on the original cross-sect1onal area or the test speclmen without regard to the lateral contraction of the specimen during the test, thus the actual or true stresses are greater than those plotted in the conventional stressstrain curve. Fig. B1.2 shows the general relationship between actual and the apparent stress as plotted in stress-strain curves. The dit!erence is not appreciable until the higher regions at the plastic range are reached.

Figs. Bl.3 and B1.4 compare the shapes of the tension stress-strain curves for some cammon aircratt materials.
Bl.5 The Static Compression Stress-Strain Diagram.

.,....-------...,
~ ACtualStre~/
Appar... Stress

it~ ..

~ j
Fig. Bl.2

~ '---::-,..-,::--,----

Because satety and light structural weight are so tmportant in flight vehicle structural design, the engineer must consider the entire stress-strain picture through both the tension and compressive stress range. This is due to the fact that buckling, oath primary and local, is a common type at failure in rll6ht vehicle structures and failure may occur under stresses in either the elastic or plastic range. In general the shape or the stress-strain curve as it departs away fram the initial straight line portion, is ditterent under compressive stresses than when under tensile stresses. Furthermore, the various tlight vehicle materials have dirferent Shapes far the region of the stress-strain curve adjacent to the straight portion. Since light structural weight is so tmpartant, considerable etfort is made in design to develop high allowable compressive stresses, and in many flight vehicle structural units, these allowable ulttmate design campressive stresses tall in the inelastic or plastic zone. Fig. Bl.5 shows a comparison or the stress-strain curves In tension and compression tor four widely used aluminum alloys. Below the proportional limit stress the modulus at elastiCity is the same under both tension and campresslve stresses. The Yield stress in compreSSion Is determined in the same manner as explained for tension. Compressive Ulttmate Stress (Fcu)' Under a static tension stress, the ultimate tensile stresS of a member made trom a given material is not influenced appreciably by the shape of the cross-section or the length of the member, however under a compreSSive stress the Ultimate compreSsive strength at a member Is greatly intluenced by both cross-sectional, shape and length at the member. Any member, unress very short and ccecect , tends to buckle laterally as a whole or to buckle laterally or cripple locally when under compreSSive stress. It a member Is qUite short or restrained against lateral buckling, then tailure for some materials such as stone, wood and a rew metals will be by detinite tracture, thus giVing a definite value tor the ult1mate compreSSive stress. Most aircraft materials are so-ductile that no fracture 1s encountered in compress1on, but the materIal yieldS and swells out so that the increaSing cross-sectional area tends to carry increaSing load. It Is therefore practically impOSSible to .select a value at the UltImate cccpresetve stress or ductile materials Without having

"

:;:

."
53

180 120 80

" 53

,;

40

Fig. Bl.4

ANALYSIS AND DESIGN OF

:!.4S-T3 SH[ET
TllleM"!!:$$ ..

FLIGHT VEHICLE STRUCTURES

Bl 5

'0 .0

24S-T4 [XT"USIONS
TIUCIl.Jrl(SS" 0.250-11'1.

.0

Q,:2~O

-IN

.0

, ~-T!NSIO!'!:_. -~~ ;_1 I

.L X-CO"'''~l!~OH\ ""oTiNSION

/ ,/
f, - - - - -

1'[1'1'10"",,\

'0 '0 '0 '0

W'COt,lPIlSSION

'0
'0
'0

-::s

"- ......... I~sslo"

/

--

I'-.

-/1. _.
I

, 'I , ,

.7', ,

"

, , , ,,

0'0
.0,-

/
e

I! ;( 101 IN/IN

'11
'0 00 .0

/
a
10 1 IN/IN

Fig. Bl.6

E X I~ PSI

'0

E X 10- PSI

.'N~ ~~ '0f,of-....
.0
<0
ICOMPIIl(SSlOIll w-ttNIIOH

COMPllU:U01. 'l'ENSIO'" I I

.. -t("'IO

m
.0
'0 '0 '0 '0

r-- t-- ..J po

/

<

30

I V '0f'0

I ~~

V II/

;;

"-

The slope of this tangent gives the local rate ot change of stress with strain. The secant modulus Es Is deter.mlned by drawing a secant (atr-at ght line) tram the origin to the point in question. This modulus measures the ratio between stress and actual strain. Curves which show how the tangent modulus varies wt th stress are referred to as tangent modulus curves. Fig. 81.5 illustrates such curves tor tour different aluminum alloys. It should be noted that the tangent modulus Is the same as the modulus or elasticity in the elastic range and gets smaller in magnitude as the stress gets higher in the plastic r.ange.
B1.7 Elastic
~

Jnela.stic Aettoll.

f.'LCt. ...a 7SS-U SHEET

.....D "LATt
TH'CII:7SS 0jIS-O'jS.'l"

a x 10'

I"""

x_(',.._

'0/
0 0

/
755-T. !XTI'USIONS nUCIlNIE:SS.. 3-tN

'0 E x 10-- PSI

1I to. INn"

e x 10'"

'0
PSI

"

Fig. Bl. IS

sane arbitrary measure or crlteron. For wrought materials It is normally assumed that Fcu equals Ftu. For brittle materialS, that are relatively weak In tension, an Fcu higher than Ftu can be obtained by compressIve tests or short compact speCimens and this ulttrnate compressive stress is generally reterred to as the block compressive stress.
S1.6 Tangent Modulus. Secant MOdulus.

It a member is subjected to a certain stress, the member undergoes a certain strain. It this strain v.anlshes upon the removal or the stress, the action is called elastic. Generally_speaking, tor practical purposes, a material Is considered elastic under stresses Up to the proportional l1mi t stress as preViously detined. Fig. 81.7 illustrates elastic action. However, it when the stress is removed, a residual strain remains, the action Is generally reteITed to as inelastic or plastic. Fig. BI.S illustrates inelastic action.

In

"

Modern structural theory tor calculating the compressive strength ot structural members as covered in detaIl in other chapters of this book, makes use of two additional terms or values which measure the stiffneSS ot a member when the compressive stresses in the member tall in the inelastic range. These terms are tangent modulus or elasticity (Et) and secant modulus ot elastic! ty (Es). These two modifications of the modulus ot elastiCity (E) apply in the plastic range and are illustrated in Fig. 81.6. The tangent modulUS Et Is determined by drawing a tangent to the stressstrain diagram at the pOint under consideration.

t::- i~
Elastic Action
Q)

Permanent

Sir....

"t

j-

straiD straiD

Fig. Bl. 7

Fig. Bl.8

Bl.6 Ductility.

The term ductility tram an engineering standpoint indicates a large capacity or a material for inelastic (plastic) deformation in tension or shear without rupture, as contrasted with the term brittleness which indicates little capacity for plastiC deformation without failure. From. a phys t ca.L standpOint, ductility is a term which measures

Bl.6

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

the ability of a material to be dra'Nn into a wire or tube or to be forged or die cast. Ductility is usually ~easured by the percentage elongation of a tensile test specimen arter failure, tor a specified gage length, and 1s usually an accurate enough value to compare materials. Percent elongation = (

stress-strain
Curve"\.
"

;; "
:>

fp

La L
Q

L ,) 100

= measure

of ductlli tv.

where LQ original gage length and La = gage length after fracture. In referring to ductility in terms of percent alangation, it is Unportant that the gage length be stated, since the percent elongation will vary Nith gage length, because a large part of the total strain occurs in the necked down portion of the gage length just before fracture.
SI.9 Capaci.ty to Absorb Energy. Resilience. Toughness.

C H I E Unit Strain (e)

'" ~

~I
K

dE

Fig. Bl.9

high and the strain at this stress is high, or from equation (2), when the proportional limit stress is high and modulus of elastici~J is low. In Fig. 81.9 if the stress is released trom point D in the plastic range, the recovery diagram will be approximately a straight line D& parallel to AO, and the area CDE represents the energy released, and often referred to as hyper-elastic resilience. Toughness. Tougr~ess of a material can be defined as its ability to absorb energy when stressed in the plastic range. Sir-ce the term energy is involved, another definition would be the capacity of a material for reSisting fracture under a dynamic load. Toughness is usually measured by the term mOdulus of Toughness which is the amount of strain energy absorbed per unit volume when stressed to the ultimate strength value. In Fig. 81.9, let t equal the ~verage stress over the unit strain distance d~ from F to G. Then work done per unit volume in stressing F to G 1s fd~ which is represented by the area FGRI. The total work done in stressing to the ultimate stress f u would

Resilience. The capacity of a material to absorb energy in the elastic range is referred to as its resilience. For measure of resilience we have the te~ modulus of reSilience, which is defined as the maximum amount of energy per unit volume which can be stored in the material by stressing it and then c~pletely recovered when the stress is removed. The maximum stress for elastic action for computing the modulus of resilience is usually taken as the proportional limit stress. Therefore for a unit volume of material (1 cu. in.) the work done in stressing a ~terial up to its proportional limit stress would equal the average stress f' p / 2 times the elongation (eo) in one inch. If we let U represent mOdUlus at reSilience, then

Gp) Ep ---------- ----hence

(1)

But e p = tplE, U

=(t;) (t,:) =!p"/2E

(2)

Under a condition of axial loading, the modulus at resilience can be found as the area under the stress-strain curve up to the proportional limit streSS. Thus in Fig. 81.9, the area OAB represents the energy absorbed in stressing the material tram zero to the proportional limit stress. High reSilience is desired in members SUbjected to shOCk, such as springs. From equation (1), a high value ot resilience is obtained when the proportional 11mit stress is

then equal I ok fde, which 1s the area under the entire stress-strain curve up to the Ultimate stress point, or the area 0 A J K 0 in Fig. 81.9 and the units are in. lb. per cu. inch. Strictly speaklng it should not include the elastic resilience or the ener~J absorbed in the elastic ~~ge, but since this area is small compared to the area under the curve in the plastic range it 1s usually included in tougr~ess ~easurements. It should be noted trat the capacity at a ~ember for resisting an axially aDplied dynamiC load is increased by increasing the length at a member, because the volume is increased directly with length. However, the ultimate strength remair.s the same since it is a function of cross-sectional area and not of volume of the material.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

81.7

Toughness is desirable characteristic when designing to resist impact or dynamic loads as it gives a reserve strength or factor of sarety against failure by fracture when over-loading in actual use should happen to cause the member to be stressed rairly high into the plastiC zone. The property at ductility helps to produce toughness, but does not alone control toughness as illustrated in Fig. Bl.lO, which shows the

BI. 11 Construction of a Stress-strain Curve Through a Given Yield Stress by USing a Known Test stressStrain Curve.

s
Unit strain

Materials in general are produced tosatisfy certain guaranteed minimum strength properties such as yield stress, ultimata tensile stress, etc. Thus in the deSign ot important members in a composite structure, the minimum guaranteed properties must be used to provide the required degree at safety. In general, most materials will give properties slightly above that guaranteed values, thus the test stress-strain curve or purchased material cannot be used for design purposes. The stress-strain curve passing through a given Yield point stress can be readily obtained for a test stress-strain curve as tallows ; In Fig. Bl.ll, the heavy curve 0 8 C represents a known typical stress-atrain curve tor the given material. Let the minimum guaranteed yield stress be the value as shown at point A, using the 0.2 percent method. Then proceed as tallows:
(1) Draw a straight llne through potnt 0 and

Fig.

sr.ro

stress-strain curves for three ditferent materials. Material (A) is strong but brittle, Whereas material (C) is weak and ductile, and material (B) represents average strength and ductility. However, all three materials have the same moduius or toughness since the areas under all three curves is the same.
B1. 10 Poisson's Ratto.

paint A which will intersect the typical curve at point B. Point B may be above or below the typical curve.
(2)

When a material is stressed, it will derorm in the direction at the stress and also at right angles to it. For axial loading and for stress below the proportional ltmlt stress, the ratio at the unit strains at right angles to the stress, to the unit strain in the direction of the stress is called poisson's ratio. It is determined by direct measurement in a tensile or compressive test at a speCimen, and is approximately equal to 0.3 tor steel and 0.33 tor non-terrous materials. In many structures there are members which are Subjected to stresses in more than one direction, say along all three coordinate axes. Poisson's ratio is used to determine the resultant stress and detormation in the various directiOns. pOisson's Ratio in Plastic Range. Information is somewhat limited as to Poisson's ratio in the plastic range and particularly during the transition range tram elastic to plastiC action. For the assumption ot a plastically incompressible isotropic SOlid, poisson's ratio assumes a value or 0.5. oeram and Weldrorm. tound tram their research, that the transition or poisson's ratio rrmn the elastic value or around 1/3 to 1/2 in the plastic ra.'1ge is gradual and is most pronounced in the yield pOint region at the stress-atrain curve.

Locate any other point on the typical curve such as point C, and draw line
rr-ca 0 through C. Locate point D on line 0 C by the follow-

(3)

ing rat.t o r-,

OD=~XOC
(4) Repeat step 3 to obtain a number or pOints as shawn by dots on Fig. BI.II, and draw smooth curve through these points to obtain desireu stress-strain curve.

t ( . . - . . I. . . .

Fig. Bt.11

B1.12 Non-Dimensional stress-strain Curves.

The structural designer is constantly cOnfronted with the design or structural units which fall by inelastic instability. The

BI.8

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

solution ot such problems requires information given by the compressive stress-strain curve. Since flight vehicles make use of many different materials, and each material usually has ma.ny different states of manufacture which give different mechanical properties, the question at time required to obtain certain design intormation from stress-strain curves becomes important. For example, in the aluminum alloys alone there are about 100 different alloys, and when elevated temperatures at various time exposures are added, the number at stress-strain curves required is turther greatly increased. Fortunately, this time consuming work was greatly lessened when Ramsberg and Osgood (Ref. 1) proposed an equation to describe the stress-strain curve in the yield range. Their proposed equation specifies the stress-strain curve by the use of three parameters, the aonnua at elasticity E, the secant yield stress F O. '1" which is taken as the Una at slope 0.7E drawn trom origin (see Fig. Bl.12), and a parameter n which describes the Shape of the stress-atrain curve in the yield region. In order to evaluate the term. n, another stress Yo is needed. which is the intersection of th9 curve by a line or slope of O.SSE through

on the stress-strain curve. Table 81.1 gives the values of F o . ,., to.ee. n , etc . for many flight vehicle materials. Notice that the shape parameter varies widely tor materials, being as low as 4 and as high as 90.
HI. 13 Influence of Temperature on Mater1al Properties.

Before the advent of ~~e supersonic airplane or the long range miSSile. the aeronautical structures engineer could design the airframe ot aircraft using the normal static mechanical propertles ot materials, sInce the temperature rise encountered by such aircratt had practically no effect on the material strength properties. The development of the turbine Jet and rocket Jet power plants provided the means ot opening up the whole new fIeld of supersonic and space flight. The flight environmental conditions were now greatly expanded. the major change being that aerodynamic heating caused by high speeds in the atmosphere caused surface temperatures on the airframe which would greatly etfect the normal static material strength properties and thus temperature and time became important in the structural deSign ot certaIn types of tlight vehicles.
Bl. i4 Creep of Materials.

tho origin (s Fig. Bl.12).

E/l

O.8SE O.7E

Stress-Strain

c.....

_ o-cJ .....Jl...
)..........
~

FIg. 81.12

The Ramaberg and Osgood proposed three parameter representation of stress-strain relations in the inelastic range ta r-,

..

Fo.'I'=FO'.~~Fo.?

Ee

3(

)n.

It Is obvious tbat temperature can weaken a material because If the temperature is high enough the material will melt or tlow and thus have no load carrying capec t tv as a structural member. When a stressed member Is subjected to temperature, it undergoes a change at Shape in addition to that of the well known thermal expanSion. The tenn creep is used to describe thiS general intluence at temperature and time on a stressed material. Creep is detined in general as the progressive, relatively slow change in shape under stress when subjected to an elevated temperature. A simple illustration ot creep Is a person standing on a bituminous road surface on a very hot summer day. The longer he stands on the same spot the deeper the shoe soles settle into the road surface, whereas on a cold winter day the same time ot standing on one spot would produce no noticeable penetration of the road surface.
(3)

---

- - - - -

The sQuation for n Is,

n = 1 =. log. (17/7)/log.(F IF )

(4)

Fig. Bl.13 Is a plot of .quation (4). The quantities EeIF-o . ? are non-dimensional and may be used in determining the non-dtnenslonal curves of Fig. Bl.14. E. n, and F 0.'1' must be mown to use these curves 'n obtaining values
...
~t..-

High temperature. whe..l used in reference to creep, has ditferent temperature values for different materials for the same amount of creep. For example, mercury, which melts at -38OF. may creep a certain amount at -75 0F, whereas tungsten, whf en me 1 ts at 6170OF, :nay not creep as much at 200QoF as the mercury

uncer -75OF. All materials crp underconditions Of temperature, stress and time ot stress application. The simplest manner in which to obtain the etrects at creep Is to study its ettect on the static stress-strain diagram tor the material

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Bl.9

Fig. B1. 13
50

.0
30
25
20

""'-

r-,
-,

-,
~

15

-,

toe. (17fT)
11-1+

10

~. (FO.7"'0'.86)

<,

I'....
<,

r--,
t"-

1. 5 1. 0 1.01 1.02 1.03 1.D41.051.08 1.08 1.1 1.15 1.2 1.3 L'" 1.51.8 L8 %.0

"0. T/FO as

Fig. Bl.14

1.'
1.'
1.0

, I ,

<
~
/"

0.'
f

FQ.";

0.'
0 .

0.'
o

1/ 1/0.'

~~
o

~~

f:?
-a

,
"

1
0.'
1.0

0.'

1.'

l.S

'.0
C / -::..- t'

c'"
~,

,- .,:
Bl.10
BEHAVIOR OF MATERIALS AND
Table 81. 1 Values of Ftu> 1""",. Eo:. 1"0.7. F 0.85, 'to or Temp. Temp. MATERIAL E"". OF

THEIR PROPERTIES
Unci

v arlOUS !Iota . rer ia1s

H,.

i
I ,

I
I

e,

., Roo m &. El evat odTempera ures (F rom R

I"tu".

koSi
125 125

-.

~.

STAINLESS STEEL AISI 301 114 Hard Sheet Tr:uuJ'ferse Compresllion L.olJiitudirlal Compresllion A1SI JOI 1/2 Hard Sheet TJ"Innene Compression Loq:ltlldlnal Compression

1/, 1/2 1/,

1/2

RT RT RT 400 '00
1000

zs
as
15

'r I" :Y:' ., i , , '"

ao
43 118
106.5 107.5

E,
106pl!l i 27. a 26.0 27.0 23.2 20.9 16.2 26.0 22.4

..
t
kosi

6)
,

1"0.7

kosi
73

1"0.as

I !
!

150

1/' 1/' 1/, 1/,

1/2

na no na no

AlSI 301 3/4 Hard Sheet Trana'l'eMie Compresaton

Lofli1,tu4lna1 Compresllion

1/, 1/, 1/, 1/' ,/2 1/, 1/' 1/, 11'

RT 400 '00
1000

15

150

"'

ee

ee

28.2 116.5 106.5 106.5 94.5 45.5 44 40 163.5

I ,

sa aa

5. ,

s. 2
U

'""
as
J7

96.5 83.5

S. a 6.0

a.e

RT <00 600
1000

12

<7,

ae

53.3 52.8 45.2

ac. 1
15.6 27.0 24.1 22.4 16.9 26.0 23.3 21. 5 18.2 27.0 25.1 23. a

"

180

I
12

138
11J 175

11'

'"
'" "a

RT 400
60Q

'" ua
"a "a

153 127 65
65.5

138
78 71 70.3 59.3

'"to
ss
<83

I
I

31 30.5 151. 5 142.5

4.4 4.7 3. s 4.3

13.2 13.2 11.2 19.2

140
1Jl 61.5

AISJ 301 Full &rei Sbeet Tranuerse COl'nprepwll.

Loutudln.a1 Comprnllwn

1000 RT

l'l"'" PH B&r II. Fol"l1nP

17.7 PH (THtOSO) Sheet, Strtp II. Plate. t '" .010 tD .1251D.. 177 PH (RlI9S01 Sheet:, Strip II. Plate, t ... 010 tD 125 In. 19.9DL (AMS 5526) 6 19-9DX (AMS 55381. Sheet, StZ'tp &: Plate 19-QDL (AMS 5527) C. 19_QDX (AMS 5539) Shnt, Strip &: Pl.a.te PRl!l~7Vo (THl0501 Sheet &: Strip. t ... 020to.187ir1. PHU-7Mc (RH950) Sheet &: Strip, t '" . 020 tD .187 in. L.OW CARBON &: AL.LOY STEELS AJSI 1023 1I 1025 Tube. Sheet" Bar, Cold Flniab.ed AJSl4130 NQrmaJUed, t ~ .188 In.

1/, 1/, 11' 1/, 1/, 11' 1/' 1/, 1/, 1/' 1/' 1/, 1/, 11' 11' 1/'

.00 600
1000

165 168 15'

131 185 Ii(

na

se
46

56.5
172

I I
,

7.5
'.6

e. a
5. S

179 130 65. 60. a

79.9 66.3

17. <7,
141. 5 77.5 74 74

18.

ie
18
21.5 5.2

<8

RT
400

a e

21.6
26.0 24.2 22.9 20.6 27.5 25.3 23.1 21.2 29.0 27.11 24.9 20.3 29.0 29.0 29.0

1000

'00

RT .00 700 1000

RT

'" 146 '" as

131

"a

'"ea

135.5

180,

165
105.5 62.6

ea

sa
<80 129 145

59.5

42.5

s
3.
J4

168

'"

"7

180

.00 700 "00

RT RT RT

I
JO

'" H. ea
JlO

18'

'"60

er sa
47

I
I I

<8 7. 1 7.4

H.

ue

'146

61.5

'" n7

sa

'" ".
196
32
74

I ,

1/, 1/,
1/2 11'

ss
190

""45
90 170 '00 I

I 8.4 e

s. a

""

16.4

36.5

7.5 7.2
22.5

RT

RT

I "e 4 I I
aa
J3

'"

65
171

28.0 28.0

134 189

aae

are
32. '7 61. 5

'-'
J4

i ,

RT

1/2 1/,

AISl4130, 4HO, 4340 Heat Treated

AISI 4130, 4140. 4340 Heat Treated

1/, 1/' 1/' 11' 1/, 1/'

1/' 1/' 11' 11' 1/2 1/, 11' 1/, 1/, 11' 11'
1/2

11'

RT '00 900 1000

RT

ee

ae
70

aa
18.5

11'
AISI 4130. 4140, 4340 Heat Truted.

500 .60 1000 RT '00 550

1000 RT

125 113

..
a,
58

90

SO

I , I

34
150 135

61. s 48.2 30.8 113 98.3 58.9 49. '7

14'

105

'"88.5
63.8

AIB14130. 4140. 4340 Heat Treated

500 650 100'

RT

,
I

76
15

180

500 '50 1000

I
r

13.5

200 180 140 104 140

115

'" '"sa
'"sa

17'
15' 109.3

A-2M (AMS m!5Aj

."...,

Y Sheet, ?Iate

I ,

" 17' , '" 87.1

19'
95 88.4 81. '7 50.3 JO

29.0 29.0 27.3 23. a 40.8 29.0 27.3 23.2 20.8 29.0 27.3 23.2 20.6 29.0 27.3 23.2 20.6 29.0 27.3 23.2 20.6 29.0 24.4 19.8 14.2 Z6.0 26.0 31.0 28.9 28.4 23.2

ss
40

aa III se
66.5 45.5

aa
102

31. 5 53 48 32.5

e.

7.3 5.' 4.7

,

SO 61. 5 41

rae

145
58

140

I s.a I " as
as

10.9 10.9

17'
15' 109.4

"

'"
57

63.5

17'
58

153 10'

.
eo
.46

18. 5 10.9

19'

" 172..5
93

aa s.e
JO

19'

121.5 57
57

16'
117

83
57 ai

as
19 14

I I

K-MONEL Sbeet, Age Hardened MONEL Sheet:. Cold Rolled '" Annealed tNCONEL-X

1/, 1/, 1/2 1/, 1/2 11' 1/,

11'

I I

1<:00

RT 600 1000 1400 RT RT RT .00 '00

15

15 35

12' 70
155

20

ae
10'

162 141 10.

20 10.

er 50 ee

ts

95.6

90. a

aa

94 a8.6

17 100 89

ea

47

13.5 12 .5 15. 3 13. 5 23.5

17

I
I ,

"

78.8

'.4

ar

18. 5

I
I ,

ANALYSIS AND DESIGN OF

1'lOble al 1 Valu

FLIGHT

VEHICLE STRUCTURES
()fL.

m. n
6j (C0 nUnu ed) F O8S

"

of Fw

I''Y' Ee, FO 7

FO 85

n, [or V;u1.01lB MaterwlI Under Roo m &: Ele v aled l'om", r:;ltl1rell (FrQ m Ret

~L

Temp.

MATERIAL ALUMINUM ALLOYS

2014~T8 ~Naiona

""' '"'..
2 2 2 2 1/2 1/2 2 2 2 2 1/2 1/2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Temp.

'F

o.

F""

FC Y'

RT
300 420 600 300 420
RT

t ::!' O. 499 in.

"" 60 " 28 to
Sl

""

" I 106pSl
10.7 10.2

ea
8.0

-. .
Fo 7. Ul'"

'"
50.3

I ,
18.5

53
4.1.5 20.5

42.5 21 4:1.5

3.2 7.'
10.2

2014-1'6 Forpngs t ::!' 4 In.

300 600 300 ..0

ea sa

31

sa
41

28

9.2
10.7 10.2

ae

S 44.0

"

19.5

,. S

" 25 s.,
25 29 20
19 12. e

42.5 25.2

'50

29 10

22 7
43 25.5

9.2
7. ,
10.2

52. :1 40.5 21.5 42.5

53
32
12
65.'~Q

..

SO
38.5

20 3.0

3.2
15.8

I
I

2024-1'3 Sheet & Plate, Heat Treated. t :: . 250 In.. 2024-1'4 Sheet &. Plate. Heat Treated, t ~ O. SO In.

-:

RT
300.

9.2 8. , 8. , 1.7
10.3 10.7 10.3

SOO 700
RT

12

300 '00 700

RT

"
60

34
24 7

*"
36

.a35.7
2U 6.2
22 60

280 c

" 21. 36.

33.5

28 7.

22.8

"6 11.' ,

31!.7

2024-1'3 Clad Sheet &. Plate, Heat Treated, t = .020 to .062 In. 2024-1'6 Clad Sheet &: Plate, Heat Treate<l, t ? O. 063 In. 2024-1'8 Clad Sheet &. Plate, Heat Treated, t < O. 063 In. 2024-1'81 Clad Sheet, Heat Treated, tcO.064 In. 6061-1'6 Sheet, Heat Treated &: Aced, t<O.25In. 7075-1'6 Bare Sheet &. Plate,
t~0.501n.

12

8.' 6.'
10.7

"""
35.7-

34. s.,
30.5 21

s.s

15 ... 10.9 j 8.2 '

ia e

RT

300 SOO 700 8 300 SOO 700

ea
60

. ..
"

27 34 24.5

B.S

,0. 3 B., 6. ,
10.7

25
20

RT
300 SOO 700

22 6

47 43.2

10.7 10.3

10.3 8.' 8.'

8.' 6.'

.. .
21.7

25

30.3

'.8

"

14. e 10.2 18.5 12

11 7.9
l}5

44.3
47

40.7

31.5

7.0 42.3 29.s '.00

28 6.0
43

11 8.3 6.8
10.6 10.8

RT
300

31 6

"
42

" 50.s

10.7 10.3 10.t

ee ae

51. :I

51.8 46.5 34

2L7 28 '.0

7.8
U

11.2

io

7075-~

Extrusions,

t~0.251n.

7075-1'8 Die ForginjJ:', t ~2 ill.

7Q75- 1'6 HaDd Forginp. Area :s 16 sq. In.

7075- re Clad Sheet &. Plate, t ~O. 50 In.

2 1/2 2 2 2 2 1/2 2 2 2 2 1/2 2 2 2 2 1/2 2 2

a a

1/2 1/2 1/2 1/2 2

RT
300

'50
BOO

"
7

RT

76

RT

...
300 .00 "0

300 425 '00

.....
54

,11.5
7.5

29.5

10.5

25.5

8 30 7
73 70

10.5

'50

RT

300

600 "0
RT

54 22 8 25
71 47.S 18.5

.. , .. ,
8.1 '.2 B.l 7.8 '.3 7.B 8 7.B '.3 7.8

3 B 7.0

"

28
17.7

31 28
10.9 15. :I

19.3

to. .
55.8
25.4-

6.6

6.2
B3 52.5 22 '.2 32. e

34 .s 72 58.5 21.3

7.2

9.2 15.8 12.1 3.7 12.8 11' 7.2 3.2 B.8

15.2

re

" 54.'

"

10.5

,..s
53

47.8 17.3

"
te

IL. '.3

.Ll

1e.e

12

300

7.0 22
72
B3 51.6

10.5

'50
RT

.5O
70

600

2'" 7.6 24

300

.
59

7.6

'.3

~1.2

"

'.0

3.7 22
61.5

19
10.9

63.8 6.0

25
21.5 13.7

60

20.3

7
10.5

60

7079_ TlS HaPd Fof"iinP. t ~ 6.0 In.

MACNES M .... LLO AZ61A Extrusions,
t~0.249il1.

2 </2 </2 1/2 1/2 </2

'50
RT 300

:l0.5

600 "0

,
,

7.7 23 67

7.8 '.3 7 7.8 '.3 6.3

....
54
19.7

se.e
7.7
27.2 59.5 4S.5

"'.0
25.3

er.e

19.5 19.5

<6

to. 5

'600 W

67
21

7.0
14

20

e.,

.'"
I
I

51.7 17.5 S.s 25.3 57.5

20 3.6 23

,.
is

12.4

3.s
12. J

12

'.0

RT
</2
1/2

12

ae

12.9

HK31A-0 Sheet t 0.016 to O. 250 in. KK31A-1i24 Sheet, t-0.250ill.

RT
300 .00 '00 300 .00 .00

</2 </2 1/2 </2 1/2 </2 1000

RT

10 10

I I

30 20
34

12
11.1

6. S
6.16 4.94 3.77

10

is

a.s

10 22
17

'.3 '.9 19
17.7 14. a

11 120 130 10. 67

70
~

7.a

6.s L2

7.s 3.3

e.s
'.6
I . 14.6 12.6 10.5

8. ,

.s a

17.3

IL 6
13. 1

I
I

2.2 6.2

3.8

'.7
119.5

S.2 102
124.5

e.,

'.9

THMn ~eaJ.~
Plate &. Strip
&. Sheet, t
~.

Sheet, Bar 167 ill.

RT RT
.00 600 800
1000

n-ao-w Annealed

i/a 1/2 </2 </2 </2

123

'"

15.5 16.0 14.1 t3.0 11.8

sa
t

se

... s SO .s

127 67
85.5 80.5

83

". ,

aa

'.7

er

-n

."
"

..

13.7

22
21. 5

59.5

0, ......

.....--

Bl.12

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

Fig. Bl.lS (curves A and B) show the stress-strain curve for a material. Curve (A) is tor a low elevated temperature condition and curve (B) that ot a high elevated temperature condition. The results were obtained by a normal testing ~chine procedure requiring a short time test period, hence the results can be considered as independent ot time.
~;":'t

B1.15. The General Creep Pattern. A typical manner of plotting creep-rupture test data is illustrated in Fig. Bl.1S. For metals tested at high value of stress or temperature, three stages in the creep-time relation can be observed as shown in Fig. Bl.lS. The initial stage, otten called the stage ot primary creep, includes the elastic deformation and that region where the rate ot creep detormation decreases rather rapidly with time, which no doubt Indicates an influence ot strain hardening. The second stage, otten referred to as the secondary creep stage, represents a stage where the rate ot strain has decreased to a constant value (except tor high stress) tor a conSiderable time period, and this stage represents the periOd ot minimum creep rate. The third stage, otten called the tertiary creep stage, represents the period where the reduction in cross-sectional area leads to a higher stress, a greater creep rate and tinally rupture Transition Point. The inflection ~oint between the constant creep rate ot the second stage and the 1ncreasing rate ot the third stage is reterred to as the transition paint. Failure generally occurs in a relatively short time attar the tranSition point. Transition points may not occur at very low stresses and ~y also not be definable at very high stresses. Minimum creep rate is that indicated in the second stage, where the creep rate is practically constant.
Initial

~!:'!-:;,

p',,-

'.rr

pc

I:';",:
,,~,

"r:~":::'
'';::~~~,
-'0'

::p'c_
0

i
~r. I

' J..- i ,~, ...~l..._-L_

;

I'

W,
.~~

:~;i .- -7'
~:

//c ., .~

~~~~i~ """~

... :
':L

:.::.;,z-_~_

.... ... HMr'-.... -... -__..... --f .

j
; ;
DIQlIt1'-'tWE TIJeOlIl

...

-.

.-,.
"
~~.

-!
!

I.

?f.
.~ ,

..
,
Fig.

.
~.?
"

,i'.~,

e" ... ~, , / r, ,."., .. f.'.! L". .' II I' :" FO-" !&~ . , "y '," ..... ... i;.'!c ,. r.. ?:".i' , r.S J """"'~~
-. ,- I
~-::~-.

t"

- ~\
~;"

,<"
' .

I l . .Ii

H'
0

i---

i,

'. t, f i,:

. .
,

....

Bl. 15 (Ref.

2) - Effect of temperature and time on the strength cba.raetertst1cs of metals.

--

The tigure shows that the higher temperature (curve B) reduces the ultimate strength, yield strength 'and modulus ot elasticity or the material as compared to curve (A) which is a test at a lower temperature.
EFFEX:T OF TIllE

Second Stage
(Approx. Constant)

Third

stage :
I

Stage

Rupture
Point

It temperature and stress are ot such combination as to prOduce appreciable creep, then t1me becomes an important errect , For example, in Fig. Bl.lS, it the materIal at low temperature (curve A) Is stressed to a value P within the elastic limit ot the material, and this stress Is maintained tor a considerable time periOd, very little creep, it any, will be detected, and when the stresS Is removed, the material will practically resume its original dimensions. However, it the materIal Is stressed to the same value P but under a high temperature condition, creep will occur Just as long as the stress P remains applied and. the stress-strain curve will take a Shape like curve (e) In Fig. Bl.15. This time dependent strain will follow a characteristic pattern. The material will never return to its original Shape atter creep has taken place regardless it the stress is removed. It both stress and high temperature continue, rupture produced by creep will tinally occur.

, , , ,
I

, , , ,
I

., Transition Point

~ Slope"" Creep Rate

Creep Intercept
Time - firs.

Fig.

Bl.16

Typical creep-rupture curve.

Fig. 81.17 shows a plot ot cree-time curves tor a material at constant temperature and several stress levels. It is noticed that increasing stress changes the creep-time relationship considerably.
ai.

16 Stress-Time Design Charts.

In many structural deSign problems i~ volV1ng elevated temperatures such as power

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

B1. 13

i
o

_.
Fig. Bl.17 (Ref. 2) Creep curves for a material at constant temperature and various stress levels showing the characteristic stages of creep behavior.

Fig. St. 19 (Ref. 3) Effect of exposure at elevated temperatures on the room-temperature tensile yield strength (Fty) of 7079-T6 aluminum alloy (band forgings).

plants, the critical design tact or is not strength but the permiSSible amount at creep that can be allowed to still permit the structure or machine to tunction or operate satiSfactorily. Extensive tests are usually necessary to provide reliable creep design informatIon and such test intormation IS often recorded in the form as illustrated in Fig. Bl.1B.

i I
a
Fig. SI.20 (Ref. 3) - Effect of temperature on the tenstle yield strength (Fty) of 5 Cr-Mo-V aircraft steels.

-'

I ,

:m
'II
I

,
"
, ,

II

d',

:'

, II!
,

,
I
I

"
i

,
,

i:

, I'

II11

II

IIII
,

II II
'1111

II II , III

--- -

i II I i

Fig. BI. 18 (Ref. 2) - Stress-time design chart at a single constant temperature for selecting limiting' stress values.

strasses at a given temperature when heated tram zero up to 2000F per second under constant load conditions tran when loaded attar the material has been 1/2 hour or longer at constant temperature. Increasing the temperatura rate tram 2000 F to 20000F per second or more results in only a small increase in strength. Figs. Bl.2l and Bl.22 (Ret. 4) show the etfect ot temperatura rates up to lOoOF per second upon the yield and Ultimate

SI. 17 Effect of Time of Exposure.

In general materials can be roughly classified into those which time ot exposure to elevated temperatures has great influence on the mechanical properties and those where such exposure produces relatively small etfect. This general tact IS illustrated in Figs. 81.19 and 81.20. The yield strength ot the aluminum alloy In Fig. 81.19 IS tar more influenced by time or exposure to elevated temperature than the steel alloy as shown in Fig. Bl.20.
S1. 18 Effect 01 Rapid Rate of Heating

",~-l--+--+-+--+-+--+----1

. .--

"'f---+-+--+-+-~~+--+-

i
Fig. S1.21 - Tensile yield stress of 2024-T3 aluminum alloy for temperature rates from O.20F to lOOop per second and of stress-strain tests for l/2-hour exposure.

SupersoniC aircraft and missiles are subjected to rapid aerodynamic heating. The results of tests indIcate. that in general metals can withstand substantially higher

Bl.14

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

~'C-I- , - - - , - - - - , - - , - , - - - - ,

Fig. Bl.22 - TensUe rupture stress of 2024-T3 aluminum alloy for temperature rates from O. ~F to l00 0F per second and ultimate tensile stress of stressstrain tests for 1/2-

Fig. 1.23 (Ref. 3) - Effect of Temperature on tntimate strength (Ftu) of 20l4-T 6 Aluminum Alloy.

hour exposure.

strength respectively of aluminum alloy as compared to values when loaded after the material has been exposed 1/2 hour at constant temperature.
B1. 19 General Effect of Low Temperatures Upon Matertal Properties.

The development or the missile and the space vehicle brought another factor into the ever increasing number ot environmental conditions that etfect structural design, namely, extremely low tetnperatures. For example, in space the shady side of the night vehicle is subjected to very low temperatures. Missiles carry rueta and oxidizers such as liqUid hydrogen and oxygen which boil at -423 and _297oF respectively. In general, low temperatures increase the strength and stiffness at materials. This etfect tends to decrease the ductility or the material or, in other WOrdS, produce brittleness, a property that is not desirable in structures because of the possibility ot a catastrophiC failure. In general, the hexagonal closely packed crystalline structures are best suIted for giVing the best service under low temperatures. The most important at such materials are aluminum, titanium, and nickel-base alloys. Fig. 81.23 shows the effect of both elevated and low temperatures 0n the ultimate tensile strength ot 2014-T6 aluminum alloy under various exposure times.
B1.20 fatigue of Materials.

vehicles. Fatigue failure is failure due to being stressed a number of times. For example, a beam may be designed to sately and efficiently carry a deSign static load and it will carry this static load indefinitely without failure. However, if this load is repeated a large enough number of times, it will fail under thiS static deSign load. The higher the beam stress under ~he static deSign lead, the less the number of repeated loadings to cause failure. To date no adequate theory has been developed to clearly explain the fatigue failure of materials. Fatigue failure appears to begin with a crack starting at a point of weakness in the material and progreSSing along crystal boundaries. A microscopiC examination of metals indicates there are many small crackS scattered throughout a material. Under the action ot repeated stress these small cracks open and close during the stress cycle. The cracks cause higher stress to exist at the base ot the crack as compared to the stress it there were no crack. Under this repeated concentration of stress, the cracks will gradually extend across the section or the member and finally causing complete failure of the member. Fatigue testing consists of 3 types:(1) the testing of material crystals, (2) the testing of small structural test specimens, and (3) the testing or complete compOSite structures. A tremendous amount or test information is available for the second type of tes~1ng. More and ~ore attention is be1~~ given to the third type of testing. For example, a complete airplane wing or fuselage is often subjected to elaborate fatigue testing in order to insure the safe deSign lite of the airplane.

DeSigning structures to provide satety against what is called fatigue failure is one of the most important and dlfttcult problema rac ing the structural designer at flight

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

B1. 15

The presence or cracks initially is not necessary to start a fatigue or progressive failure as irregularities such as slag inclusions, surface scratches, pitting, etc., can cause corrosive action to start, thus supplying the condition tor the promotion at cracks and the resultant progressive failure. The strength of ferrous metals under repeated stresses is otten reterred to as the endurance or fatigue limit. The endurance limit stress 1s the stress trat can be repeated an infinite number of times without causing fracture of the material. Non-ferrous materials such as the aluminum alloys do not have an endurance limit as defined above but continue to weaken as the stress cycles are increased. Due to this fact and also since the required service life of structures and machines vary greatly, it is customary to refer to the strength under repeated stresses as endurance or fatigue strength instead at endurance limit. Thus the fatigue strength is the maximum stress that can be repeated tor a specified number of cycles without producing failure at the structural unit. The results of testing a speCimen under repeated stresses such as tension, compression, bending, etc., is often plotted in a torm which is referred to as the S-N (stress versus cycles) diagram, as illustrated in Fig. 81.24.

magnitUde at stresses in aircraft structures Is discussed brietly in Crapter A4. The limited discussion which tollows will deal only with the effect or impact loads on the behavior of materials.

IMPACT TSSTING METHODS

There are in general two types of tests to determine the behavior of materials under impact loads. The usual tmpact test which has been conducted tor many years Is referred to as the notched bar test and consists of subjecting notched specimens to axial, bending and torsional loads by the well known Charpy or Izod impact testing machines. In both of these machines an tapact load is applied to the specimen by SWinging a weight W trom a certain vertical height (h) to strike and rupture the notched specimen and then stopping at a vertical height (ti I). The energy expended in rupturing the specimen is then equal approximately to (Wh-Wh I). This type or test 1s primarily used tor studying the influence ot metallurgical variables. The other type ot impact testing is made on unnotched speCimens and the general purpose is to obtain the stress-strain dia~ at materials under impact load or the loaddistortion diagram of a structural member or composite structure as the unit is completely fractured under an impact load.
Bl.22 Examples of Some Results of. Impact Testing of Mater1als.

!
'Jl

.-. 11::t 10" 10 X 10"

9 x 10 8 x 10

-,

g
'"

'"r--...
10" 10" 10 'r Number of Cycles for Failure, logH-

Figs. 81.25, 26 and 27 show the resUlts at impact tests upon the stress-strain curve as compared to the static stress-strain diagram (Re t . 5).

7 x 10 10

looj : :
I '

!
i
, ,

j
i ,

Fig. B1.2.

The problem of fatigue design ot aircraft airframes is covered in Chapter Cl3 at this book.
Bl.21 Effect of Impact Loading on Material Properties.

A '.t.' I~ ?isol u ~0'

' ~ 50
1 :

I :r

' "

I I

04 f.'jsec
i

i
'

i~i
,

: ::rstar,c
I ,

_ 40' , ~ :1
III ,

I I

I:
I
,

~o:?O

An impact load when applied to a structure produces appreciable shock or Vibration. To prOduce such action, the load must be applied rapidly, that is, in a short interval at time. The efrect at linpact loads differs trom that at static loads in that impact loads appreciably ettect the magnitude at the stresses produced in a member and also the resistance properties or behavior of the material under load. The importance and etfect at dynamiC loads on the

12
Unit

.~4 ~on9aIiQn.lnlin.

.s

'

20

28

.J2

Fig. Bl.25 S'ress-strain curves, 24ST aluminum alloy.

Bl 16

BEHAVIOR OF MATERIALS AND THEIR PROPERTIES

,

so

I I

I I I Slotic
,

I '" /5C f:'sC'cl

, ,

'I'at:e B1.2 snows additional impact testing results as compared to the static test results.
Table B1. 2 Comparison of Strengths, Ductility and Energy Absorption Under Impact and Static Loads. xxx Ratio of Impact to Static Value For, Material Yield Point 1.328 1. 996 1.224 1.29 1. 387 1. 522 1. Maximum
Load

~30

~40 c-

Ih

]V '" co rl/s9Ci

,
I

i
'I

,
I

'.20
~ 101

,
I
IJ2

, I ,
04

, , I , , , , ,
I

, ,
I

, ,

I I

i ,

I
,

I I

I I I

,
/4

,0 '2 C6 :J8 I;IlI elongotion, in./in.

16

/3

Elongation 2.205 1. 418 0.992 0.682 1. 1. 163 1. 628 1. 27

Energy 2.865 1. 376 0.946 0.600 1. 203 1.6

Fig. Bl. 26 - Stress-strain curves, Dow metal X.

SAE No. 6l4()X SAE No. 101 SAE No. 101 18-8 Alia 17ST Alum. Al.
Brass

Aluminum Copper
240

I
I

I I
'/
~

I
i

~I
~

" s

,
"

i!OO

I I I

!
80

frl$f1C

, ,
I

--<.C:-t-- :

, , ,
i
I

x Oil quenched from 162QOF. XX Cold-rolled. :c:a From "Stress-Strain Relations Under Tension and Impact Loading" by D. S. Clark & G. Datwyler, Proe. A.S. T. M. 1938. Vol. 38.

"

1. 020 1. 285 1. 397 1. 212 94 1. 142 1. 323 1. 390

I
I

I
I

:;/~1I,1 r\ "-.' I I I I I i \, ,
!

REFERENCES

-, . ,

,
I

I'/ ~

JSO fI/s.cr"-:- I

Ret. 1.
Ret. 2.

I I I
,
,

I
:

'"
0

,
ce

I I i I
I

,
,

I
I /2

, ,

I I I
.,0

I I

i
(.Inil

I
06

04

ef0l'l90hen. '''',lln.

co

I i
.14

Ret. 3. Ref. 4. Ret. 5. Ref. 6.

Fig. Bl.27 - Stress-strain curves, SAE 6140. drawn lO20oP.

Ramberg & Osgood. Description of Curves of 3 Parameters. NACA, Tech. Note 902. T1me and Temperature Greml1ns of Dest.llctlon. By L. A. 'rerxorvicn, Correll Aero. Lab. Research Trends. Se"'; 1956. Military Handbook (~IL-HDBK-5) Aug. 1962. NACA Technical Note 3462. NACA Technical Note 868. From Structures Manual, Convair Astronautics.
Stres~-Strain

CHAPTER B2

MECHANICAL AND PHYSICAL PROPERTIES OF METALUC MATERIALS FOR FUGHT VEHICLE STRUCTURES
General Explanation. It would require several hundred pages to list the properties or the many

materials used in flight vehicle structur~l design. The metallic materials presented in this chapter are those most widely used and should be SUfficient for the use of the student in his structural analysis and design problems. All Tables and Charts in this chapter are taken tram the government publication '~111tary Handbook, MIL-HDBK-5, August, 1962. Metallic Materials and Elements for Flight Vehicle Structures~. This publication 1s for sale by the Supt. or Public Doc~~entsJ Washington 25, D.C. The properties given in the various tables are for a static loading condition under room temperature. The effect of temperature upon the machan1cal properties 1s given in the various graphs.
AISI ALLOY STEELS

Table 82. 1 (AIBI) Alloy Steels Alloy ....

AISI 4130. 8630, and 8735

AISI 4130. 4140. 4340. 8630,8735, and 8740

I
Form ....... ........... . Sheet, strip, plate. tubing
I N

4140 4340 8740

AISI 4340

All wrought forms

Condition .. .. .

I
>0.187

Heat treated (quenched and tempered) to obtain F I .. indicated

Thickness or diameter, in..

'j

, :<0.187

Basis .. ....... . ..... . .... . . ... ... .

..... .. . .. . .. ... ........ . .. .. ...

90 70 70 55
.. ......

........I
200 176 198 119 272 355

(0)

Mechanical properties: F h ksi.. ... F ,,,' ksi. .. .......... ... . .. F.", ksi ..
'

F ,.., kai . . . . ...... F h ... ksi:

(eID~1.5)

...

(ejD=2.0)

.. . .... .. .

I I
.!

96 75 75 55

125 103 113 82 194 251

150 132 145 95 219 287 189 218

I ,

180 163 179 109 250 326

260 217 242 149 347 440 312 346

~

140
, ,

140

Ph". ksi (e(D=1.5) . (eID~2.0) . e, percent.

. ...
.. .. ...

.,,I
"
,

I ,

151 180

230 256

255 280

i
,

See table
2.3.1.l(b)

E, 103 psi. . E., 103 psi. G. 10e psi .....

i---'---------'-----i
-!

See table 2.3.1.l(c)

29.0 29.0 11.0

L 10

T 3

--------1------------------Physical properties: oM, Ibiin.3. C, Btu, (lb) IF) .. K, Btul[(hr) (W) (F),'ftj. ""' lO-~ in ..in./F.
1

0.283. 0.114 (at 32 F). 22.0 (at 32 F). 9.3 (0 to 200 F).

82.1

: :t ..
-"'.. -

'....J

B2.2

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES

A:SI
~140 ,

~~LOY S~2SLS (Cc~t.)

'.,~~~=._._._.

E
o

e
~

Slrenq1h Of temperature
E~posure

'

--'

up to

1/2 hr.

~120

cr
-0 \00
-e
,

o o

,.,
,

4QO

600

ecc

1000

'200

1400

,600

'"
o

"

r."'(l.rof~"',F

E.

Fig. 82.2. Effect of temperature on the compressive yield strength (Fey) of heat-treated AISI alloy steels.

'"

80
,

e o 60

FiT
,

Ft. I
,
,

.. I
e

roo

~
0

40

!
, , , ,

. .. .'

eo
600

"
, i ,
a
.cc

ecc

1000

IlOO

I-lOO

600

Temperal"llI, F

20
,

Fig. 82.3. Effect of temperature on the ultimate shear strength (Fsu) of heat-treated AISI alloy steels.
,

o -400

400
Temperature
1

800

1200

F Fig. 82. 1. Effect of temperature on Ftuo Fty. and E of AISI alloy steel

~
j

l
$00

soc

1000

;lOO

1400

1600

T_ral...... ,F

Fig. 82.4. Effect of temperature on the ultimate bearing strength (FOnJ of heat-treated AISI alloy steels.

i.
,
J
0

.
ec

~==SE=~~~l~~~~~

--=-=-..o.

._;-;;:.:.~=

" '0

"

400

600

ecc

'000

'200

1400

1600

r''''p''QfUI'e, F

Fig. B2. 5. Effect of temperature on the bearing yield strength (F'bry) of neat- treated AISI alloy steels.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

B2.3

5 Cr-!'lo-V AIRCRA?!' STEEL

Table B2. 2 Design Mechanical and Physical Properties of 5Cr- Mo- V Aircraft Steel

Alloy ......

...

. ......

Condition ...... ... Section size .. .. Basis ....... , ..

Form ........... . ....

...

.. .....

5Cr-Mo-V aircraft steel. All wrought forms. Heat treated to obtain the F II. indicated. Up to 12 in. diam, or equivalent.
(0) (0) (0)

... . . ... ...

Fig. 92.8 Effect of temperature on the tensile and. compressive modulus (E and Eel of 5 Cr- Mo- V aircraft steels.

I
260 220 240 280 240 280 170
....

Mechanical properties: F h ksi ...... ........

240 200 FIll. 220 F. II ! ksi ..... . . . . . . . . I , 145 F .... ksi ...... . . .... F~ .... ksi: ... .... .. .... . (e/D-1.5) .. 400 (eID~2.0) . . . ... . F II , ksi: . ......... ... . (e/D-1.5) .. 315 (e/D-2.0) . . . e, percent: 9 Bar, in 4D ... .... . 5 Sheet, in 2 in. Cb} . 8 Sheet. in 1 in ......

I I ksi .... ..... ......I

155

.. . ...... . ... .
435

. . ....
465

.. . . ...... ... .
340 8 5 7

..

. . .....

365 7 4 5

Fig. 32. 9 Effect 01 temperature on the compressive yield strength (Fey) 01 5 Cr-Mo- V aircraft steels.

-,

i.
1

E. 10' psi .. .. . ...... E 10 8 psi .. ' G, 10' psi. . . ....

Physical properties:
I.

. . Ibjin. 3 . . . C, Btul(lb)(F) . K, Btul[(hr) (It') (F)lltl. a, 10-~ in./in./F .....

30.0 30.0 11.0

.;

0.281. 0.11(') (32" Fl. 16.6 (400 to 1,100" F). 7.1 (80 0 to 800 0 F); 7.4 (80 0 - 1 . 2 0 0 0 F).

Fig. B2. 10 Effect of temperature on the ultimate shear strength (Fsa) of. 5 Cr-Mo-V aircraft steels.

Minimum properties e:q:le1:ted when heat created as recceimended in section 2..U .0.
~ For sheet thickne~ greater than 0.0li0ineb. Calcn.lated value.

I I
.1

!
l

Ftg. H2. 11 Effect of. temperature on the ultimate bearing strength (Fbrn) 01 5 Cr- Mo- V aircraft steels.

-,

Fig. 92. 6 Effect of temperature on the ultimate tensile strength (Ftu) of cr- Mo- V aircraft steels.

-,

Fig. 92.7 Effect 01 temperature on the tensile yield strength {Fty} of. 5 Cr- Mo- V aircraft steels.

-,

.'

Fig. H2. 12 EUeet of temperature on the bearing yield strength lFbry} of 5 ce- Mo- V aircraft steels.

82.4

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERL\LS FOR FLIGHT VEHICLE STRUCTURES

17-7 PH

STAI~1.~S

STSEL

Table B2. 3 Design Mechanical and Physical Properties of 17-1 PH Stainless Steel

Alloy.

17-7 PH
Sheet, strip, and plate(")

Form .. Condition.

Bars and Forgings(b)

TH 1050
0.005 to 0.500 (--1

RH 950
0.005 to 0.1874
0.1875 to 0.500

TH 1050

RH 950

Thickness or diameter, in ..

6 and under

Basis

s
180 150 158 117 2'97 360 210 190 200 136 346 420 200 180 189 130 380 400 270 297

S (,)

Mechanical propertiesFt ksi . F, ksi . F ksi . F , ksi. . F~ ksi: (eID-1.5) . (e/D~2.0) . F~.." ksi: (eID-1.5) . (elD~2.0) . 8, percent ....

170 i 140 ! 147 ' 111 280 340

185 150 158 120 305 370 225 247 5

225 I 285 247 I 313 See table 2.7.2.1(b)

210 231 6

E, 10' psi. E 100 psi G, 10 psi.

.
.

29.0

30.0
11.0

Physical properties:
tal.

lb/in.~

C, Btu/Clh) (F) . K, Btullihr) (ft')(FJ/flj. a, 10-4 in.lin./F.

0.276. 0.11. (~) 9.75 (at 300 0 F). 6.3 (70 0 to 600 e F) for TH 1050. 6.8 (70 0 to 600 0 F) for RH 950 .

Test d.1reetJon longitudina.l for widths less than 9 tn.: transverse for widths 9 in. and over. Test dlneUon longitudinal; these properties ncr eppucabre to the

short transverse (tbtcknesal direction. Vendors guaranteed minimums for F,., Fro, and e. Calculated value.

j
~

Fig. B2. 13. Effect of temperature on the ultimate tensile strength (Fm) of 17-7 PH (THI050) stainless steel.

Fig. B2. 14. Effect of temperature on the tensile yield strength (Ftyl of 17-7 PH (TH1050) stainless sreet.

ANALYSIS AND DESIGN OF FLIGHT

VEHICLE STRUCTURES

82.5

17-7 PH STAINLESS

STE~L

(Cant.)
roc

''''

! ..
l
Fig. 82. 15. Effect 01 temperature on the compresl!Iive yield strength (Fey) 01 17-7 PH (THI050) stainless steel.

.~

.....

Fig. 82. 18. Effect 01 temperature on the ultimate bearing strength {FbruJ of 17-7 PH (THI050) stainless steel.

I J
.:

Fig. B2. 16. Effect of temperature on the ultimate shear strength (Fsn) of 17-7 PH (THI050) stainless steel.

Fig. B2. 19.

EHed of temperature on the bearing yield atrength (i'bry) 0117-7 PH (THI050) stainless steel.

I J
,J
!
l

1
e
&CO

100

IOOCI

1200

l..co

I.cG

-.'

Fig. 82. 20. EUect of temperature on the tensile and compressive modulus (E and Ed of 17-7 PH (THI0SO) stainless steel.


82.5
MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEIITCLE STRUCTURES

17-4 PH

STAINL~S

S?S2L

Table 82.4 Design Mechanical and Prryaical Properties of 17-4 PH Stainless Steel

,
Alloy, .

....
Plate
H 900

17--4 PH
Bars and forf;ings(b)

i
l

Form .. Condition .. ... Thickness or diameter, in....

Basis .............. ......

H 900 8 and under

Fig. 82.23 Effect of temperature on the compressive yield strength (Fey) of 17-4 PH (H900) stainless steer.

S ('J

!
I
i
l

Meehanical properties: F, ksi ... ... . .. . .... F ,... ksi .......... .... .... F~ ksi ........ F ksi ........... ... F~n. ksi:

190 170 178 123 313 380

190 170 178 123 313 380 255 280

e'/D-1.5). .
e'/D-2.0) ........ e'/D-1.5) ...... e'/D=2.0) ...... e, percent:
F~'''J kei,

Fig. 82.24 Effect of temperature on the ultimate shear strength (F au) of 17-4 PH (HSOO) stainless steel.

255
280 10
...

In 2 in............ 104.D ..........

E, 10 psi .. E 10' psi G, 10' psi
8

I I
,
)

10

l
.
.

29.0 30.0 11.0

Fig. 82.25 Effect of temperature on the ultimate bearing strength (FOru) of 17-4 PH (H900) stainless steel.

PhysioJ properties:
w,

lb/in.3 ...

0.282.

0.11 (32" to 212" F). C, Btu/(1b)(F) K, Btu/[(hr)(ft') (FJlrtl. 10.3 (at 300" F); 11.2 (at 500" F); 13.1 (at 900" F). a, 10- 1 in. lin. IF. 6.0 (70" to 200" F); 6.1 (70 0 to 400 0 F); 6.5 (70 0 to 900 0 F).
Vendon guaranteed minimums (or F, F, .. and (. Test direetion longitud.in.a.l; these properties not apptlceble to the shon transverse (thiclme5S) direction.

i I
,
;
l
Fig. 82.26 Effect of temperature on the bearing yield strength (Fbry) of 17-4 PH (H900) stainless steel.
"

i I
l
l

"

Fig. B2.21 Effect of temperature on the ultimate tensile strength (Ftu) of 17-4 PH (H900) stainless steel.

Fig. 82.22 Effect of temperature on the tensile yield strength (Fty) of 17-4 PH (WOO) stainless steel.

Fig. 82. 27 Effect of temperature on the tensile and compressive modulus (E and E c > of 17-4 PH (H900) stainless steel.

A::fALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

82.7

~~-360 S7AINL~S

STSSL

Table 82.5 Design Mechanical and Physical Properties of AM-350 Stainless Steel

Alloy, Form .. Condition .. Thickness, in ... Basis.

AM-350
Sheet and strip" DA Sheet and strip"

seT
0.187 and under

0.187 and under

s
165
135

s
185 150 158 120
305
Fig. 82.30 Effect of temperature on the compressive yield strength. (F cyl of AM~350 stainless steel (double-aged).

Mechanical properties: F,~, ksi . F!I/, ksi . F CI/' ksi .

F.~,

ksi. . F"r~. ksi: (eiD = 1.5) . (eiD=2.0) ...

142 107
272 330
202 223 10

370
225 247

j
I
-s I

F"rl/' ksi: (e/D=1.5) (eiD=2.0) e; percent.

. ,

10 29.0 30.0
11.0

!
l
Fig. 82.31 Effeet of temperature on the tensile and compressive modulus (E and Ed of AM~350 stainless steel (double-aged).

E, 10 psi.
6

E" 10 5 psi., G, 106 psi.

Physical properties: w, lb/in. 3 . C, Btu!(lb) (F) . K, BtUl[(hrj(W) (FHtj .. a, 1O-~ in.iin./F .

0.282. 0.12 (32 0 to 212 0 F). 8.4 (at 100 0 F); 11.7 (at 800 0 F). 6.3 (70 0 to 212" F); 7.2 (70" to 932 0 F).

!
.'

Test direction longitudinal for widths less than 9 m.; transverse for widths 9 in. and over.

!
l

e'tg , 82.32 Effect of temperature on the ultimate tensile strength (Ftu) of AM-350 stainless steel (SCT).

j
!
l

Fig. 82.28 Effect of temperature on the ultimate tensile strength (Ftu) of AM-350 stainless steel (double-aged).

Fig. 82.29 Effect of temperature on the tensile yield strength (Fty) of A11-350 stainless steel (double-aged).

~ 19. 82.33 Effect of temperature on the tensile yield strength (Fty) of AM-350 stainless steel (SeT).
'.~ ~-,1

:'. . :- t.

c-:

B2.8

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES

AM-350 STAINLZSS S!EEL (COilt.)

J
J .r
!
l

~ '" J '"
;

"" ~

"' ~ l

'"
e

sec

ecc "'"TtrroerQrun, """ F

"00

cc

,"00

Fig. B2.34 Effect of temperature on the tensile and compressive modulus (E and Ed of AM-350 etatnless steel (SeT).

Fig. B2.36 Effect of temperature on the compressive yield strength (F cy) of AM-350 stainless steel (SCT).

! J
;

j
J
;

J
~ l
FIg. B2. 37 Effect of temperature on the ultimate bearing strength (Fbru) of AM-350 stainless steel (SCT).

Fig. B2.35 Effect of temperature on the bearing yield strength (Fbry) of AM350 stainless steel (SCT).

FIg. B2.38 Effect of temperature on the ultimate shear strength (F su) of AM-350 stainless steel (SCT). .

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE

STRUCTURES

B2.9

AISI 301 STAINLESS STEEL

Table 82.6 Design Mechanical and Physical Properties of AISI 301 Stainless Steel

Alloy ....... .... . . .. . .... Form ........... Condition ..... Basis ........

AISI 301

.. ...
Annealed
S

Plate

~,

sheet, and strip

,
h -hard
S

.. . . ......

H hard
S

hard
S

Full hard
S

..... . . ....

Mechanical properties: F I ,.. ksi: L .... ....... .. . T ... . . ...... . ... Fill' ksi: L ...... . ...... .. T ... ..... ....... F.", ksi: L ........... ... . T .. . . .... . . ....

75 75 30 30 35 35 40 150

125 125 75 75
43 80 67.5

150 150 110 110 58 118 80 300

175 175 135 135 76 160 95

185 185 140 140 85 179 100

F .., kai . ..

.. . .....

F br . , ksi: ('/D-1.5) ...... (,/D=2.0) . . ' . ..

F~.".

........ .. . ....... ......... .. . . ...... . . . .......

250 350 240
(.)

370
270
( .)

ksi:

(,/D=1.5) .. . . (,/D-2.0) . .. . e, percent ... ...... . . E, lOG psi: L .... ... . . .... T ..... . .. E., lOG psi: L .... .. . ......... T .. ....... . ..... G. 10' psi. .. ...... ...

... ..
50
( .)

. .........
140
(.)

...... ....... . .. . . . . . .. .. .
200
( .)

29.0 29.0 28.0 28.0 12.5

27.0
28.0 26.0

26.0

26.0

28.0
26.0 27.0
11.5

28.0
26.0

26.0 28.0 26.0

27.0
12.0

27.0
11. 0

27.0
11.0

Physical properties: Cot, Ib/in.3.... C. Btu/(lb) (F) . .. . K, Btu/[(h.) (ft')(F)/ft). a, 10- 6 in./in./F ...

0.286. 0.108 (at 32' F) . 7.74 (at 32' F). 9.2 (700 to 2000 F).

_ Properties rcr annealed condition allIo applicable to anneated AlSI 302,303.304.321. and :W-, Only annealed condition applicable to plate. ,see table 2.8.1.1 (b). :-!OTE.-Yield strength, partiCUlarly in compn!5IIion. and modulus of el.aaticity in the longitudinal direction llIay be raised appreciably by thermal stres;s-relieving treatment In the range 500- to 800" F.

"

-'

-.

' - -'

'-.

' '.

_ -

", , -;,

-~,

r~f;1 -~&J

82.10

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERrALS FOR FLIGHT VEHICLE STRUCTURES

j
i
J ,
Fig. 82.39 Effect of temperature on the ultimate tensile strength (Ftu) of AISI 301 (half-hard) stainless steel.

AISI 301

S7AIN~SSS

STEEL (Cout.;

i i
.'
l

J
Fig. 82.44 Effect of temperature on the bearing yield strength (Fbry) of AISI 301 (half-hard) stainless steel.

Fig. 82.48 Effect of temperature on the compressive yield strength (Fey) of AISI 301 (full-hard) stainless steel.

! i
.' J

! I

,
Fig. 82.40 Effect of temperature on the tensile yield strength (Fty) of AISI 301 (half-hard) stainless steel.

j
.
J

Fig. 82.49 Effect of temperature on the ultimate shear strength (F su) of AISI 301 (full-hard) stainless steel.

J
1
Fig. 82.45 Effect of temperature on the tensile and compressive modulus (E and Ed of AISI 301 (half-hard) stainless steel. Fig. 82.41 Effect of temperature on the compressive yield strength (Fey) of AmI 301 (half-hard) stainless steel.

J
J ,
J

i i
J ,
Fig. 82.42 Effect of temperature on the ultimate shear strength (Fsu) of AISI 301 (half-hard) stainless steel.

i i
~

Fig. 82. 50 Effect of temperature on the ultimate bearing strength (Fbru) of AISI 301 (full-hard) stainless steel.

J ,
Fig. 82.46 Effect of temperature on the ultimate tensile strength (F tu ) of AISI 301 (full-hard) stainless steel.

I i
s
! ,
Fig. 82. 51 Effect of temperature on the bearing yield strength (F'bry) of AISI 301 (full-hard) statntess steel.

i i
J J

,
Fig. 82.43 Effect of temperature on the ultimate bearing. strength (Fbru) of AISI 301 (half-hard) stainless steel. Fig. 82.47 Effect of temperature on the tensile yield strength (Fty) of AISI 301 (full-hard) stainless steel. Fig. 82. 52 Effect of temperature on the tensile and compressive modulus (E and Ed of AISI 301 (full-hard) stainless steel.

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

82. 11

NOTE:

2014 ALUMINUM ALLOYS (SHEST & PLAT~J SXTRUSIONS, fORGINGS) Values in (A) colwmns are nin:~un guaranteed values. Values in (B) col~ will be met or exceede~ by 90 percent of material supplied.

Table 82.7 Design Mechanical and Physical Properties of 2014 Aluminum Alloy (Sheet and Plate)

Alloy.

2014
Sheet and plate

Form.

Condition.

-T6l1
1.5012.000

2.001-

3.001-/1

Thickness, in.

0.02<HJ.039
B

0.040-0.499

0.500-1.000

1.001-1.500

3.000
A

4.000
A

Basis.

I - --- - - - --- - - - - - Mechanical properties: Flit, ksi L.

T ST
FIll, ksi

65 64

67 66 60 59

68

67
60

70 69 62 61 62
63

68

67
.

70 69 62 61 62
63

67 67 59 59 61 61
41

68 6B

65 65

63 63
58

59 59
54
55 55

L.

58

60

T. ST
Feu, ksi

57

59 60 61
41

59
60

62 62
64 64 41

59 59 61 61
40

.'57 57

53 61 41 102 129
84 96

51 57
57 57

D.

T. ST
F, ..,ksi. Fbr.., ksi c (../])=1.5).

42 105 133 87
99

42 105 133 87 99
..

59 59 59 39 89 129
80

37
88

(e/D=2.0)
Fbrr/' f.:lliC (r/D=1.5) (el D =2.0) e, percent L. 7'

102 129
84

101 127 83
94

102 129 87
99

98
124

112
ii

83
94

96 8 8

91 4 2 1

88 3
1.5

ST
E, 10 psi. E c, 10' psi. G, 10. psi

6 6

6 ,
10.5 10.7
-.l.()

" 3

Physical properties: 1oI,lb/in. l C, Btu!Ob)(FI K, Btu/[lh,)(ft')(FI!ftj. <J:, 10-- in./in./F

0.101 0.23 (at 212F) !)O (at i7F) 12.5 (68 to 212F)

B2.12

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VElllCLE STRUCTURES

2014- ALt.n'1I.\'tl1 ALLOY:::; (SHEET & PL.;\TE. ~TRUSIOIlS. ?ORGINGS)

(cont.)

Table 82.8 Design Mechanical and Physical Properties of Clad 2014 Aluminum Alloy (Clad Sheet and Plate)

Alloy.

Clad 2014
Sheet and plate

Form ...

Condition.

...

.. ...
<0.039 0.040-0.499

_T6 cJ
0.500-1.000
A

Thickness, in.

.. . .... . ........
. .......

1.001-1.500

Bui8.. .............

Mechanical prcperdee:

PtMo

Jt;.n

F trh

L .. ..... .. .. ......... T ....... ............. ST. . . ... .. . ...... ... /en L .. ... .. . ... ...... . ... T .. .... . .. ........... ST .. .. . .. . .. . ........ L .. ........... ...... . T .. .. .. ........ .. . . .. ST .... ... .... .. . . ...
k.ri

64

64

65
64

67
66

65
64

63
..

63

..
56 55 58

67 66
60

64 M
..

66
65

56 55

60

58

57
..

59
..

57
..

59
60

57 57 59 59 39 96 122
80

60 60

.'

p~

56

57

Ft . .. len .. .. .
P"".,/eno

...... . ...

..

56 57 39 96 122 78
90

58 59 39 98 124 81 93 8 8

60

58

61 40 101 127
84
96

59 39
98
124

61
40

62 62 40 98 124
84
96

39
96

(./D-1.5) .. (./D-2.0). ....

p.........

122
78

101 127
84
96

(./D-l.5). ... .. ... (./D-2.0). ......... ...

e, percent L .... . .. . ............. T ....... .. .......... . E, 10' pn.... .. . ... . ..... . Ee, 10' pai. ... ....... .. .... G, 10' "... .... . ..... . .. . ...

90

81 93 6 6
10.5 10.7

91 6
4

7 7

4.0

ANALYSIS AND DESIGN OF

201~

FLIGHT VEHICLE STRUCTURES

B2.13

ALUMINUM ALLOYS (SHZET & PLATE, EX1RUSIONS, FORGINGS) (Cant.)

Table B2. 9 Design Mechanical and Phystcal Properties of 2014 Aluminum Alloy (Extrusions)

Alloy.

2014
Extruded rod, bar, and shapes

Form.

Heat u-ted andapd by"'"

Condition.

-T6e <25
....... 0.125-0.499 0.500-0.749 0.750-1.499

-T82
>25, <32
Up to 32

Ceose-secncnel area, in.s

Thickness, in. 1I

1.500--2.999 3.llOO--l.499
A
B

;;::0.750

0.125"'-'99
A

Basis .. .........

Mechanical properties: F tvo kri L. 60 T ................ eo F'I1I, kn L. ...... 53 T .. ... ... . .. . 53 PCfI1lui L. . ........... ...... 55 T. 53 F N. kn.. ......... 35 FM, hi (,ID-1.5). 90 (,ID-2.0). 114

61 61 57 57 59
56

64 64

68

67 62 59
64

68 63

73
66

68

73
63
6.5
55

61

68 58

73 61
65 53

se
58

68

eo se
53

58 55

eo
54

6.5

eo
52 62 57 39

eo
49
62

58 67 61 42

47

47
53

eo
58'

35

37

62 39

62 57 39

67 61
42

57
39

67 61 42

.. ..
39

48
35

95 95 88 95 88 92 96 102 88 116 122 129 109 117 109 117 109 117

.. .. ..
6 1

..

P/wv.Jcn
(,ID-L5). ....... (,ID-2,O) .. .... . percent

74 85

80

91

81 93 7 5

87 99

78
84

85

78
84

85

78
84

85

91
..

91

91

I,

L ..

Td ..

..

... .. ...

.... . ... .

7 5

..

7 2

7 10.7 4.0

..

E, 10- ]Mi.. E e, 10- pn . . ..... G, 10- pn..

2 .. 10.5

7 1

..

-,.,..-,

"

., '-I
..;..

,~

B2.14

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VElfiCLE STRUCTURES

2014- AUNINUM ALLOYS (SHE:S7 & PU.TS, EXTRtTSIOl-JS J FC?Gn;oS)

(Cont )

Table 82.10 Design Mechanical and Physical Properties of 2014 Aluminum Alloy (Forgings)

Alloy

2014
Die forgings Hand forgings Length <:3 times width

Korm.

I I
-T6
<6 inches

Hand forg:ngs Length> 3 times width

Condition.

-T4

I
<:16 >16, <:36 >36, <:144

Thickness, ill ....

....

<4 inches

Ceoee-eeeuonal a.rea., in. l

>144, <:256

<16

>16, <:36

>36, <144

> 144, <:256

ilasis..

........

--Mechanical properties: Ftv, k&i L. ..... .... . . ..

T . .. ... ..... ST

55

.... .....

52 30
28

65 62
55

65 63 60 55 55
55 55 55

65 6.
60

62 59 56 53 52 52 53 52
38

60 58 55
"0 0-

65 63 60 55 55 55 55 55 40 91
117

65 63 60 53 53 53
53

62 59
56

60 58
55

FIV,bi. L. .... .. . ....... . . .. T. ....... S7' .. . . .. .. .... F CVo in L. ...... .. T. .. ..... ....

52
55

53 53 53
53 53

50 50
.')2

53 52 52
.j3

52 50 50 52

ST
[l ft>

m ..

...

30 28 34

....

..

52 39

50 37
84

53
4Q

40 91 117
77
88

40 91
117

F",.,.,k.ri

(t:/D-1.5). .... (,ID-2.0). F'In-v,1csi. (,ID-1.5). (t:/D-2.0). e, percent L. .... . .......

87 112
74

91
117

108 73 83 5 2
1

87 112

84 108
73

74
85

77
88

85
7

74 85 9 3
2

74

ss
7

8.

11

T. ............ ST. E, 10' ]Mi. E e 10' psi. G, 10' psi ..

7 3

10 6 3

9 5 2

10
4

3 1
10.5 10.7

2.5 1

.,
1

4.0

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

B2.15

EFFECT OF

TE:i'!PER.~TURE

ON 2014 ALUMINUM ALLOYS

"'J
"'0
,ec

~--

n
TC"~(.UII.[

STUtC.'t

(J.'osvu

"

r e '0 000

..
-

.---

-~

-- --

-.

-0:=

--~

o
'''00

'0 '00 '000 '0 000

'/2

.. -=-==.. .. ., ..
...-

--

---

--

--

-lOO

~=

-'00 0 '00 200

--

-- JOO

-'00

'" - -,==:== --500

-lO

- - --)00

T04"E:FlATtRE. .".

Fig. B2.53 Effect of temperature on the ultimate strength (Ftu) of 2014T6 aluminum alloy (bare and clad sheet 0.020-0.039 in. thick; bare and clad plate 1. 501-4. 000 in, thick; rolled bar, rod and shapes; hand and die forgings; extruded bar, rod and shapes O. 125-0.749 in, thick with cross-sectional area ~ 25 sq. in.).

,'"
''''
'00

-, -.

'':L

-~

: -'-'-t+.....':1. __
~~

'.......--..

,,:~-

- ."-"._;" ~_~_tE:'

-+-l-~.;

'-- .

'J

,.,

:-7 AI ,,.... El'OSIII[ v~

Snuu"

H ..

~[tUvU

10 10 000 u

80

z w
V

-..;::: " ,
--~.,

40

, - --'/2 .. '0 '00 - tc '000 000

-

"

.--..., I

.. .. .. ..

,.-,---'

-"-200

:.::2E?~-

:~ :-- ~-'-.,~~: rr;

-lOO

-'00

200

JOO

'00

---

500

c=

t=;:-=
lO

700

Fig. 82.54 Effect of temperature on the tensile yield strength (Fty) of 2014-T6 aluminum alloy (bare and clad plate 3.001-4. 000 in. thick; rolled bar, rod and shapes; hand and die forgings; extruded bar, rod and shapes 0.125-0.499 in. thick with cross-sectional area a 25 sq. in.).

B2. HI

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES

EFFECT OF TEMPERATtTRE ON 2014 ALFMINlR1 ALLOYS (Cont.)

j
I
! ,
Fig. 82.55 Effect of temperature on the compressive yield strength (Fey) of 2014-T6 aluminum alloy (all products except thick extrusions).

.'

i I
J

Fig. 82. 58 Effect of temperature on the bearing yield strength (F'bry) of 2014-T6 aluminum alloy (all products except thick e.. x trastcns).

!
I

I l
!
I
Fig. 82. 56 Effect of temperature on the ultimate ebeer strength (Fsu ) of 2014-T6 aluminum alloy (all products except tbick extrusions.

r" ,
J
! ,
Fig. B2. 59 Effect of temperature on
the tensile and compressive modulus

---_ -,.... _""'-

.. Fig. 8.2.57 Effect of temperature on the ultimate bearing strength (Fbru) of 2014T6 aluminum alloy (all products except thick extrusions).

(E and alloys.

Ed

of 2014 and 2017 aluminum

2024 ALUMlNUl1. ALLOY (BARE SHEE:I' & PLATE, EXTRUSIONS, BAR, ROD & '"lIRE)

Table B2. 11 Deatgn Mechanical and PhySical Properties of 2024 Aluminum Alloy (Sheet and Plate)
Alloy
Fol"ID .

..... .......

2024

Sheet end pl.te

Heat
t~t.ld

Coiled. !heet HelOttree.ted

Condition.

ThieJmes (in.}.

sw...
Meo:h&nica..t properUe8:
lI,.,

Heat treated by uaerHeat treated..no rolled -T42 -1'3 -T36 T< 0 : = 0.501 1.001- 2.(lOI- 0.0100.2500.5011.001-1 2.001O.OZQ-O.500 <0.250 0.500 1.000 2.000 3.000 0.249 0.500 1.000 2.000 3.000 A A A A A """AT ,.\ B A B B_~_+.~- _A_I_8_

------12...
es

_T<
0.012-Q.128
A 8 ------

.en:
T

L ..
lily> .q,.' L.

62 62
4() 4() 4()
.

" "
38 38

62 62

60 60

.18
36 38 38 J7 03

I"C'It

.en
L.

36 36 38 36 36
00 II.
53 61

.18
36 38 96

T.

4()

P_Us:.
p.... /t;JA.c
(eID_I.5) (eID_2.0)
P~bi

J7

03

118
56

122
53 61

118
53

196
53

f,

(eID_l.5). (eID_Z.O) pell:ent

" .. .. . ... " ' 1' .. " , '" " " .. " " " '"
56

se ee

68

" "

"
66

36 38

sr
42
43

4()

43

" eo " ... ... " ...

63 62
67

61

63
46

56

ss

70 69
60 52

ra
71

62 62
4() 4() 4() 4()

66 66

..

ea
54

41
'I

4()

4()

4()

36 38

4()

46

36 42

.. 38 41

38 43 36

42

42

.7

42

'I

136

36

36

..,

49 56

...

se

51

41 41
4()

J7

102 I 92 96 102 101 129 12. 121 120 129 118 122

105
133

10'

,..

Eo, 10' pn'. G, 10' pM.

" '"

81
8

er

sa n
19

62

14

"
12

69
78

12

'"

'"

;: I:;
8

ea
(d)

eo ee

,_:

"

96

108 137
88

03

118

128

se

100

10.5 10.7

I
,

i',

i"

"

st ee

'.0

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

B2.17

2024 ALU11IN1JM ALLOY (BARE SHEEr & PLATE, E:XTRUSI ONS, BAR, ROD & '.IRE) (Cant.)
Table B2. 12 Design Mechanical and Physical Properties of 2024 Aluminum Alloy (Extrusions)

Alloy.................... .
Form...........

2024
Ertruded bars, rods, and shapes Heat treated by user Heat treated cold worked and aged

...........

Heat treated

Condition .....

.. ...........

-T4
0:0500.249 0.251}0.499
:<25

-T42 0.7501.499 1.5003.000:;; All thick-

-T81
:<0.250

Thickness, in ...

............
....

0.501}0.749

2.999

ness

Cross-sectional area, in. l

I
B
A

:;;::2:5, :;;::32
A

=<32

Basis ........ ........

... . ..

B I-

MechanicaJ. properties: Fl. k&i L ... .... .. . ...... 57 57 T ... F~. k&i 42 L .. ........... .. 42 T. ...... F 38 L. .. . ....... 38 T ....... .. . 30 F. kJli ... ...... . F m. kai

61 61 47 46 41 41 32

60

60

62 62

60 60
44

62 62 47 45 42 42 33

65 ;8 46 41
44

70 61 54
44

70 ;4

74 57 ;4 41 ;2
44

70 ;0 52 36

74 ;3
64

57
SO

..........
........ ..

64

44. 47 43 46 39 39 32 42 42 33

;2
38
SO

42 39 39 32

39

38 36 38 38 30

56

k"

42 34

;2 48 38

42 38

SO' ;2 42 44

40

38

40

..... . ... .......... .. . . . . . . .

('ID~U)

. ..

(e(D=2.0l.

. ... .. 85 91 85 91 85 91 85 91 85 91 85 91 108 114 108 114 108 114 108 114 108 114 108 114 .. .
;9 67 12
..

85 .......... 108 .. ........

F..... lcai (,/D=U). .. . (,/D~2.0) . ..... e, percent

66 75

60 69 12 6

66 75

60 69

66 75

61 71

66 75

62
73

66 75

62 73

66 75

53 .......... 61 .. ........ 12 ..... ..

L .. ....... ... To .... E, 10- pai ... . . .. ..... E e. 10' pai.... G, lOt psi.

... .

12 ... . 10 6 .. 5 .. .

10 . .. 10 2 . ... 2 10.5
1O.i

... . .. ......

. .. . . .. . . . . .

4.0

82.18

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES
p 2024 ALUI1INl.J11 ALLOY (3 ..e.... .E SHE2I' "'= ?L{'~S, ~T:\USIC:NS,

SAH, ROD & '~'::\.:::) (Corrt . )

Table B2. 13 Design Mechanical and Physical Properties of 2024 Aluminum Alloy (Bar, Rod, and Wire); Rolled, Drawn or Cold Finished; Rolled Tubing

Alley.
Form ..

.. 1 Bar. rod and wire; rolled. drawn or cold- finished

I--------~----
Tubing
Heat treated by user'
1

mu

, - - - - - - - - - - - - - - - - -----,.-----i

Condition ..

I
1

-T4 or -T351

Heat treated

I
1

I
Cress-sectional area, in. Thickness. in ..

~36
I
I I
1 ' I ' :

Heat treated cold worked

-T3

.
I

-T42

i,nd aged -T81

Bas;,, .. ..

l-A-I~-A-!-A-!-A-i-A-I-A-' A

I Up to 11.001- 12.001-~ 3.001-~ 4.001-: 5.001-! 6.001-:0.018 to 0.5001 0.01811.000 ! 2.000 '3.000 14.000 5.000 4 1 6 . 0 0 0 4 1 6.500 a , 0.500

i--- ---

-1'-'--'1'-:-1
I
!

Mechanical properties: r.; ksi

1 I'
: -

- 1 'I

'
1

'1

r.; ksi

L. T ...

62
61

82
59

62
57

62
55

I
i

62!
54

Li T. '1 F~II' ksi i

L .. .. .. ..

40 40 32 37

~
39 32 37 93 118 56 64 10 8

~
38 32 37 93 118 56 64 10 6

~
37 32 37 93 118 56 64 10 4

~
37 32 37 93 118 56 64 10 2

521 ~I ~
36 . 32

62

62

64
1

70 46 46 42
105

I
1
1

a'
42

j. 1
1

62 40 ' 40
39

68 60

"I
I

F br ksi

F .... ksi.. ksi. (e/D-1.5) " . " (erD-2.0) "

T...

.i
.. . '1
L

'1' . . . . .I.
93 118 56 , 64 I

37
118

r,..

37

39 93
118

93 56

,
1

133
, ,

96 122 58 64

(eID=1.5).

(e(D=2.0)...... e, percent L. " " T. " ".

'

64 !.

59 67

64 74

10' 1 10 ,
1

10
2

101
1

1'1
!

E, lOS psi ...

E<, IDS psi ..

10.5 10.7

G. IDs psi. .

4.0

Physical properties:
J

Ib(in. Btu/ilb) IF). K. Btul{(hr) (ft 2HF)/ftj a, 10-s in.iin./F.

w, C,

'.1'.

0.100 0.23 (at 212" F) 70 (at 77 F) 12.6 (68 0 to 212 0 F)

Table 82.14 Design Mechanical and Physical Properties of Clad 2024 Aluminum Alloy (Sheet, Plate, and Coiled Sheet)
,

Alloy "orlll

.... ",

'"

Clad lI):lt

-_ ..... -~-

Slwct llnd Illate

Coiled shed lIellt treated b)" user' lIellt rreared -T< lIellt treated IIl:ed

811eet and Illato lIellt treated, cold worked,llIloJ aged

lleat trented end fvlled

IIeat trellted

"d

----Ooudntcn. 'fl<ickness,lo._
Width, III. Oasis ..
10'.. ,1.'11

-1'3

-'fJ6

-T<

-1'42

-TO

--

-Tltt

-T'"

...... .... .......

- ... - ...

.....

----.-

0,1163- o 250- 0, !Jon - I Olll- a 001- 0.010O.:UII 0.400 1.000 ' 2,000- a.on- 0,062 ----- - - - - - - - -, ' - - - - - - - - - ..... -- ------ - - - - - - - A A A A A A A A A A A A A A A -- A -- -- -- ---- --- -'- - - -

o.Hlo-j
0,11fi2

1,250OOOJ- O/lW'/I;,()llJ-j O.2dl 0002 0400 0.&10- 0,400

~500-

IOlJl- 2.001- O,flJIt1.000' 2.000- 3.000' 0.062

"003- 0.010-1.00:1- <0,003 0.0,-IO063-IO,2ro-l 0.128 0.062 0.2411 ~0,0li3 0.002 02411 0.49Q O. SOlI-

>Z >-

<"

M,eh"llkallrol~"II"s

" " " "

l;~

" " -" "

I~I

" " "

Jij,
IJ~'

--

--~

t. r
10', '"i

t.
T f', . hj
J.

r
,. ,hj " ,bi 4

:6

(e/Dj_U
(fiD)-2d

f'" . .\.Ii" (f!D)_U

(,'DJ~2.U f.
1~'rl"'JlI

'/'

E, lUlllli

"Ii
K,.,
,-:" lilli'll

., " " s se " m " """ " " ." " " as ae ,,- '"' ". ,,' " " ' " ., " , '" " " " " m ' , " '" " " " '" " " " " " " ae '" " " " a, '" ae '" " '" " ae , " " " .. '" " ae n '" " " " '" " I, " " " '" " " "' " a:, " " " " " " " " " '" " " "" ,. " '" '" " "' "' " " " "' '" '" '" Ii, ., ", " " "' "' ,:I " '" " " " " ,II 6. ~:I '" " " " " "' . " " " , ~:I ", ,,, ,::1 ' , " "' , , " -, '" " '" " "
f,1
.~'t

Ii:! Ii.'
li~

Ill, 1\, Ii'l

li.~ f~\

fl,
Illi
.~~

f,~

li.~

iii

.,~

~l;

S.~

:1"

I~

.,

S2

.
~'l

1\' li2
~li

.,

("

5~ I

ea

62

41;

IJ~

ali

a, J:, a,
:I~

u;

ta

Jij,

ali

:l~

4~

3,

!IJ 11,\

!'.~

U~I

1I~ l~"

1211<'11

12,~

1111 WI lUI 101 12, 1:t1 12, lJl

ll~

12ll12~

116121 112 JIll

III

Iii

,1\

l"
11I.0
11I.2

. .
11

a;

a,

lOll 110 112

Il~

.,

., ., .. " . ". ,-" " " " " '" . " " . " '" .. " '" . " '" " "
36
ati a,

-- ---

<0'
--,

" "
A

-- --

" " -" -A A A

N 0 N

r "' " '"

~

r-

,0

(t). ,

T'

" '" "" "" " " "' '" '" "' '" 0, eu ee es 36 a, st ea " '" sc "" " es ae 0' '" " es '" " " 31; as sa " " " ,,' r,; '" '" '" '" " " as " '" " " Ii, " '" a, " :I. '" " 3. " '" " " '" " '" " " " " " " " IIf, " I 12U " "' " Ul> Ill) 12, " lin " Hl:l 111I '" " " '" " " '" " '" lUI lU " " '" '" " " "" " 1:/111 "TU; '" '" " " " " 1O~r:~; , , ,, a , "J
62
6~

rs rn rs ;n u

Ii'l 6:1

"
H

..

'" >Z

fll; fll; III

a,

~ " '" '" c ~ Z C "- O C~

"
."

a,

J.~

a~

3\1

~
Ul

ll~

iii

JlXI 100 1Il610'J Ul,) ItIl, 11I3 l~ 11M In I:? 131 IJ~ la~1 LJa la~ lal 133

511

..

ll~

.,

." t<
~

l'i tJ
OJ

102 1Il4
4 ...

... '"
~

Cl

IO~

(oJ

"

lJ.~

I '" I " I
I
1II,2

I I
os

IU.U

I I
os
!I.'

lUll

j"
11...

I'rl

III" 11..

"'''H~I " I
0.100 0.23 (.. t :mQ F) 86 (at 77" F) 12.6 (611" to 212"

r '" ., "j '" n

n
H

<:

t<

\Ilfl

f:;
111.2

K""
fI, 1I1',,,j
l'h)"~kolIIlIOI'''rlk~'

,-

I I
II.

I "' I
Il.UWI
1f,2;l ('11 2120

IIU!

1 1 1 1'",I, ,1;:1 "' I

1Il,2

"

Ul

l'i tJ

101,11>11,,' C,ll1u/(161(IO

t')

'" n " ... " ''" "

'" '" ...

K, 11lu:IChr)(ft '/(F)I
III 10-' III.flll./t'..
'"

711 (at
12."

ijO I")
(.,~o

10 2120 F)

".J

'"

hlilsllLud, liS II ruunt! h.st 11..... ml'n l~ n'lulr." lor lt'stil'lt', The Yllhh'S j!:IYl'lI hHe lor thickm'ljS 0..)11/) 'nei, ,uul Im'al"f hay" ~"n DlljUbll,1 10 r"l'n....nt II,, U~"ruj!:U "fOl,,'Hh's across 111, ...liull' l.-'ctiou, hll'luolinj!: 11l.I,ItIlI,I'nll. ~ 5," Tuhl" a.I.I.I.I . ',1'10,-"IolIl:IIlIon vuhlt's 1m'IIlI'lit- .... IUn"'~ Ilrt' r",uu'ln Til!>'" 3.2,3.1l (t'l

For the ~lH'~S n'lloved {(lnl_r-TaM lI'I'lI<' Ollt}), nil \'oIlut'$ .tiI" Ill. -1'4 l"'IlI"'f al.pl)" IO'ltl, Ih.' elrel'lloli ul ~' wldeh lIIay tN.. som"I\'h,,1 IO\\"'r ." Ileul Irrull',1 hy Ullef" rd,'rs 10 lilt n',l!("riILJ sohlllllll Ill'ut 1""all',1 Ii)" Ih., lllit'f r"jl:ar'!ll'ss of Iho "flor telll''''f of Ihe rtlJtlerlal. fllledft''llUon mllllnmms lor rllI,l nll,l"rlal U.Jij.llllt'lIlhkk lind h"al'I"r lin' lor lIll' j-ur., IlIat.'rllll,

.82.20

MECHANICAL AND PHYSICAL PROPERTmS OF METALLIC MATERIALS FOR FLIGHT VEmCLE STRUCTURES

EFFECT OF TE!'PERATURE ON 2024 ALtmINUM ALLOYS

!
I

.'

! I

, I
I

Fl.g .82.60 Elfect of temperature on the ultimate tensile strength (Ftu) of 2024-T3 and 2024-T4 aluminum alloy (all products except extrusions).

Fig. 82. 64 Effect of exposure at elevated temperatures on the roomtemperature ultimate tensile strength (Ftu) of 2024T3 and 2024-T4 aluminum a.1loy (all products except thick extrusions).

Fig. 82.68 Effect of temperature on the ultimate bearing strength (FOro) of clad 2024-T3 and clad 2024-T4 aluminum alloy (sheet).

I I
!
I FIg. 82.61 EIfect ot temperature on the ultimate tensile strength. (Ftu) of 2024-T3 and 2024T4 aluminum alloy (extrusiOns). Fig. 82.65 Effect of temperature on the compressive yield strength (F cy) of clad 2024-T3 and clad 2024-T4 aluminum alloy (sheet). Fig. 82.69 Effect of temperature on the bearing yield strength (Fbry) of clad 2024-T3 and etas 2024-T4 aluminum alloy (sheet).

! I
__

,
!

i I

'-....- - 1

100 __ f _ ",,,oODQO'"

'-'
Fig. 82. 62 Effect of temperature on the tensile yield strength (Fty) of 2024T3 and 2024-T4 aluminum alloy (all products except extrusions). Fig. 82.66 Effect of exposure at elevated temperatures on the roomtemperature tensile yield strength (Fty) of 2024-T3 and 2024-T4 aluminum alloy (all products except thick extrusions). Fig. 82.70 Effect of temperature on the tensile and compressive modulus (E and Ed of 2024 aluminum alloy.

j
!
I

FIg. 82. 63 Effect of temperature on the tensile yield strength (Fty) of 2024-T3 and 20Z4-T4 aluminum alloy (extrusions).

Fig. 82. 67 Effect of temperature on the ultimate shear strength (F su) of clad 2024-T3 and clad 2024-T4: aluminum alloy (sheet).

Fi.g. 82.71 Effect of temperature on the elongation of 202C-T3 and 2024-T4 alumtnum alloy (all products except thick extrusions).

;or-_

Table 82. 15 De.sign Mechallical and Physical proper-t ins of 7075 Aluminum Alloy (Sheet and Plate)

Alloy. ....... . ........... Form. Condition Thickul-'I:I8, in. ... Basis. ....... Mechanical properties: /o'jll.' klJi L. T

7075 Sllcet end plate

T6"
0.015-0.039
A
~-

0.040-0.249
A

0.250--0.500
A

0.501-1.000
A

1.001-2.000

2.001-2,500

2.501-3.000

3.001 3.500
A

3.501-4.000
A

-- ---78 78 69 68 70 73 47 117 148 67 110

17 17

B B A --- -- - - - - - - - -A- -B B
17 17

-70 70
64

:'z :r~

76 76 66 65 67 70 46 114 144 62 106 7 7

76 76 70 66
7l H

ST
/0"/1' hi 1>

..

76 76 60 68
7l

76
17

82
80

78
17

80 70
7l

73 73 67 62 62 61 65 65
64

75 75 69 65 65 63 67 67
66

72 72
66

68 68
64

66
66

62 56
56 64

'3 en
>t:
H

., "' '"
:Z

t'l
~

67
66

7'
3T It'e/l, klSi I.. T

67 66 60 69 '6

66 66 66 66 47 110 142
00

72 66 72 72 40I 115 147

64

68
66

66
7l

60 60

62 62
60

58 58
56

56 63 63 62

68
7l

71 47 110 142
00

68 68
..

71 47 112 144 62 106

65 65
63

37'.
r, ksi r-; bi"
(,jD ~ 1.5).
(,/D~2.0)

46 116 146
64

47 119 150 68 H2

46 109 140
88

43 102 13\ 81 63 5 3
\

45 105 135 8' 67

..

'\

43 10\
130

62 62 62 49 65
\22

60
60 60

b Sl
[>J

>-

" ~

C>

" "

30 6' 110 73 8' 2

. 108
139 87 100 8 8

.. ..
0

r-

68 126 78
00

P,,'1I' kISt"
(e/ D

= 1.5)

'" " '"

(,/D~2.0)

107 8 8

104

104 6 6

108

102 5 4

81 03

75 87 5 3 1

5: tJ
~
rn

'"

e, percent L. .... T

..

.. ..

.. ..

37' . E, 10' pili. e; 10- psi. ........

...

..

..

..

.. ..

..
10.3 10.5

5 3 1

..

"U

5: -3
to

0, 10' llsi .... _.. Physical properties: w,lb/in. '. C. B'u/(lb)W). K. B'u/I,,)(f")(F)/f,1 II, 10-' in./in 'I".
For tIuo .u-. &aaper -1'''1, aD til. p~pLlon ot , . .A.ppUcable F ... a lu r.u"Jnu,. (ff&.) Di..",,~ of 0.15"'1.000 L

3.9
0.101 0.23 (at 212F)

76 (at 77"F) 12.0 (68" to 212"F)

nII_"

uUte

tor Ua. -Tt to80_:

e-pu

.pplF "Ua

''''

1.001-1.500 1.101 -a 000

L L

(A - ' _ , )

..

"8M T..L.. '.1.1.1.1.

..

..

I I I
,

Rl
Table B2. 16 Des ign Mechanka.l and Physical Properties of Clad 7075 Aluminum Alloy (Sheet and Plate)
w w

Alloy. Form Coudiuou

.-

.....
...

Clad 7075 Sheet anti plate

0.015 0.039
A
0.040--

'I'hioknesa, in. . . ..... .

Beeia.

'

...

0.063-

0.062

O.IM7

u
73 73
64

A
~-

Mechanical properties:
Pill, k8i~L

-- c-- -- - - -- -- - - - A

0.1880.249 A

0.2500.499 A n

T6' 0.6001.000'

n ,.
r1.001-

'o " ~

3:

2.001
2.500~

2.000'
A

2.501 3.000b
A

3.001

3.500'
A

n
79 77 68
66

n -78 77

-73 73 69 62 62 63
64

n -70 70 66 60 60 60 62
62

--A

3.501 4.000"

T. SOl'
PIli' bi-L.

70 70

72

74

73

..

72 63 62
64

74 65
64

73
64

75 75 66 65 67 69 45 112
142

75 75 65
64 66

77 77 67
66

75 75 65
64 66

77
77

77 75
66 64

-o

-o

~ '"
o ,. r~

76 75 66
64

sr
/I'CII. bi-L.
1" ..., ksi c . i'fn", ksi--(el D

T.

61 60 62
64

..

71 71

67
60

63 65 67 44 110 139
90 102
..

67
66

63

..

69 67 69 69 46 109 140
90

sr

66 43

66 68 44
III \41

65 67
44

68 45 112 142 91 104 8 8

68 70 46 116 146 94 107

66

= 1.5)

(e/D=2.0).
1<'/)'1/, k:siC-(e/ D = 1.5). (elf) =2.0). e. pcrcent-c-L

7' .
81'

42 105 133 85 08 7 7

..

60 60 46 108 130 87 100

66
66

69 69 47 \11 142 88 102

66 66 45 \06 137 86 99 5 4

lOB
137 88 101 8 8

110 139
90

91 104
..

102 8 8

92 106

45 105 135 84 98 8 8

46 108
139

86 90 6 6

60 61 62 62 64 42 09 128 78
90

64 66 43 102
131

68 68 64 58 58 59 61 61 62 40
95

66

64

66 64
56

64 62 54
54

104

81 93

E, 10' psi -Pri.

5 3 1
H1.:l 10.0 I tJ.5

122 75 87 5 3 1

63 41 08 \26 78 90

56 56 60 60 62 30 02 119
73

"

58 58 60
38

84 5 3 1

90 115 70 81 2

~ '" ,.. '" r;'j rOJ '" "' 0 rn ,. tn '" c-, 3:

0

>r' ~ H

ill

-o

f:>

" ~
u

... '" ,.
rr-

n
~ ... OJ ~ r-

s;

Sec.

111.:1
U.5 10.5 !1.7

10' lJ8i-l'ri.
::il~C ..

10.3 !I.8
10.5
IIU)

a, 10" }lsi

Physical properties: w,lb/ill. 1 C, Btu/(1L)(F) K,lItu/{(hr)(ft(l")/ft] Ie, 10-' in./ill./l<'

..
~-J
L';
0

'"
L-'

en

It).:!
-~

'" ~: 0 ." '" n r'" "

-u

..

0.101 0.23 (at 212"1<') 7li (al 77 F) 12.U (tiB" ttl 212"1")
II Tb....e valuell except In lhe ST direction have been adju.ted to include the IntlueD"", of the 1 1/2% per .Ide nomInal claddlDlJ thickneae. "See TaLIe 8.1.1.1.1.

"

q "
<:
OJ

a 10'01' the ..tr"", reli"ved !.emper -T66J, aU ...alu~ lor the -Tll "I;""ption 01 ,.'~ . Alllllic.aLle ,..'~. valu"" ar.. q 'ollow.: Thoi(/,;lI"" (i-n.J l>iu"li..... 01 ted "'ell (A tl4l"e., 0.260- 0.499 L ea 0.600-2.000 I. 2.001-2.600 I. eo 2.601-8.000 L

"'
r-

.
"

" " " '"

n
ci
OJ

'" ... '"

-s

Table B2. 17 Des igu Mechanical ~U1d phystcat pr-oper-ties of 7075 Aluminum Alloy (Extrusions)

----~---------- - - - - - - - - - - - - - ; : ; = - Alloy

7075 - - - - - - ---------._---------

Furm .
Condltioll
Cro:;:;-bl~d,i(Jllulllrcu, iu. 2

Extrusions (rod, harH, /Iud shnpes}

.
in.
~~-

-_'1~"c:;.

==----- I
A ---

ThidlJu~HS ~,

Habb. --- ... - -

-_._-------78
76

<20 !J 0,250- O.4ml 0.5000.749 II A A II ----._.- _._-_.--- - -

._O.750-1.4!l9 - -"'

1.500-2 .H!)!)
A

---- -

II

II

--

<20 >2t1, <32 3.000-4AWJ

<32
4,5UO~5.0

II .. A II A II ---_ ---- ---- - - - - - -

--0 0 --0

en

Meehani!",ll pruplJrtie::;:

p ...,

k,;j

L l'
F,~,

82 78
7 -l.

81
77

85 70
77

81 73
72 U3
73 73

85 75 7U
66 77 77

81
72

85
74

81

85

81
62

eo
73

es
7U 58 75 72 47
102 136

84 U4
74

78
62

81 tH 73 5U 72
H

78 UO

81 U2 71
55

r-

".

e...
00 00
~

>Z >-

kl>i

r.; ksi
L.

L 1'.

70 U4 71 71 43 101
125

73

7 U
75 75 45

UU
74 74

U8 78 78 47 102
t3li

72 U2
72 73 45

7U U4 7U 75 47
102
l3l1

71
54

5U
72

5U 75 U8 4U 101
12ti

70 54 U8 43 04 117 77 01 U I

us
53

".

>Z

1'. /t'n, ksi

F~, ..,

45 07
130

k,;i

"
U7 130 70 101 7 4

47
102

60 45
07 130 70

72 U5 45

"'
;J
c: en H
0 Z

r:: 0

" " '"

00

o
0

!:1
en
~I

43
lH

(1:/= 1.5).
(1:/=2.0) .
F ~'"'
J,;,;j

JO 7
13I

!J7
130

07
122

13U 84 IOU 8

07 125 80 U5

07
J 25

'" '" ro
~

117 75 88

0:

(t'/J)= 1.5). (el J)= 2.~) ,

01 U8
7 5

oU
104 8

80
102

85 108 8

70
101

M 106

JOI
7 I

84 IOU 8

78 02 7 I

81 96 8

78 02

&J 0
H
~,

e,

pl"J'(~I"IIL

L 1'. - - - - - - '-'-'-'-'-'-'-'--'-' B, loa -psi ..

H~,

----

7 5

7 3

8
10.3

U I _._,,--

8
~
0

"' ''" "

0:

r-

o
00

I (Ja IMi, .

10.5
---3.9 0,101 0,23 (lit 212]<') 76 (lit 77 [") 12.9 (68 to ZI2F) ~~~~~~~'--n1d11~n

'"

--(}!. _!~~~~--'-'-''-'-''-'-':.'-'.'- ~- - - - - - - - - Phyaieul

w,
Ilr()II.~rlil"::;:

"
H

Ih/iu,3 ..

0, HLlI/(Ib)(J<'). 1\., HI u/I{hr} (fLl) (F)/fLl

~.~.~~~~_~0)_;'_'-I_F

"l

'_'_''_'_''_'_'". .~"--Au.. (,q. h.)

I"ur tI ... stress rcll"""l! 1"rHl'n~ -'j'li.'t1O ..Ill! -'1'6:i1l. ull vuhir~ Jar the -'1'6 temper llflVly, wllh llo~ """'1'1["10 of V.... I. A1JI,lkllbh~ I.... \'llllll'~ erettstco uetow:

Ii".)

Aria ('f, in.l

Thitt,u.t (f".)

mUd/ono/lnl

}.,. (A ""lUI,)

<u.2lU
O.25IlU.HU 07WI,tUll

<20
:<20 <20

I. J.
I.

10

l.lillll---2.llllU 3.00lH.tll1l t.WO-5.j)OO

<'"
<32 :<32

l'irldlon 0/ 1'1" I. I. I.

F..CA

oalu~l)

"' 'c: " o s: c: '" "' ''" "

00

on

10

"
11

I .'or eltmslon~ with olllsllllhHIII: leila, Llle load clirrylnlllil>lIlty such ICl:stilllill he dt'II'OIl!l,C<\ 011 the bll.!ils ot the 'lro,wrlle:l ot ure llj'I'Wl'rllllo eclurun (:()(resl>ou<lhl~ 10 Ihe Irg Ihlckllt,s.:J

0'

ee 0'

B2.24

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES

7075 ALUMINUM ALLOY (EXTRUSIONS, FORGINGS, BAR, ROD, WIRE) (Cont.)

Table B2. 18 Design Mechanical and Physical Properties of 7075 Aluminum Alloy (Hand Forgings and Die Forgings)

Alloy.

.. ... . . . . , . .
. .........
H&.ad-fo~ stock,

7075
length <3 times width Hand-forged stock, length > 3 times width

Form...

Di.
forgings

Condition. ............... Thickness, in .......

_T64

-T5
<3

Croee-eectioD1l &rea,

m.

<16

>16, <36

>36, <144

<16

>16,

<36

>36, <144

Iluis ...

............. .. .

Mechanical properties: Fto .bi L .. . . . . . . .. .. . .

r,., .en:

T ........ ....... ST. , ... .. . .... .

L ..

75 75
72
64 63 63 64 63

73
71

71

69
66
60 58 58 60 58

68

75 73 70 63 61 61 63 61 45 97 135 88 94

73 71 68 60 59 59 60 59
..

71 69 66 59
57

75 71

T ....... . . ..... ST. ... . .. .. .....

F~Ui

....... .....

61
60 60

es
62
6.i

57 59 57 43 85 114
77

L ..
F... Ui.

T ... . . .. ...... ST. . . ... .

'

.. .

.........

61
60
44

........ . ...

..

..

58 45

45 97 13.5
90

43

44 95
124

" .... Iui (.ID-l.5). ... (.ID-2.0) . ......

95 124 79 91
7 .,
2

ss
114 78
90

" .... hi (.ID-I.S). .... . (e/D-2.01. ..... t, percent L. ... .. ...... E, 10- pIi .. ....... Eo 10- pri . . . .. . . . .. G, 10- pn. .......
...... .. . T. ST .. .. . ........

96

78 90 7 3
2

88

9
4

4 2 1

, " .,
10.3
10.5

,
2
I

3.9

ANALYSIS AND

DESIGN OF

FLIGHT

VEHICLE STRUCTURES

B2.25

7075 ALUMINUM ALLOY

(SXTRL~IONS,

~ORGINGS,

BAR, ROD,

~IRE)

(Cant.)

Table B2. 19 Design Mechanical and Phys icnl Properties of 7075 Aluminum Alloy (Bar, Rod, Wire and Shapes; Rolled, Drawn or Cold-Finished)

Alloy ........... ...... Form. . . ........... Condition ..... ...... Thickness. in. ... Up to LOOO d
A
I

7075

Bar, rod, wire and shapes, rolled. drawn or cold-finished

-T6 or -T651
I

I
I

L001 2.000 d
A

2.001 3.000 d
A

3.001 4.000 d
A

Basis ...

.. . .........

Mechanical properties: F l .. , ksi:

r.; ksi:

L ...... . . .... LT .. . . ... L ... .. . ..... LT ..........

77 77 66 66

77 75 66 66 64

77 72 66 63

77 69 66 60

F, ksi:

L ............ 64 LT . . . . . . . . . . . ...... ...... ....... F.... :':~i. .... ... 46

F hn .. ksi: (,/D= 1.5) ....

..... ...... ......

46 100 123 86 92 7 3 10.3 10.5 3.9 46 100 123 86 92 7 2

64

. ... . ..

64

46 100 123 86 92 7 1

(,/D=2.0) .... F~ ..., ksi:

('ID=I.S) ....

100 123 86 92
'7

(,/D=2.0) .... e. percent: L .. .. .. .. .. . LT . .. .. . .. . .

E, 10e psi. .. ... E., 10 pet . G. loe psi .

82.26

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MA.TERIALS FOR FLIGHT VEmCLE STRUCTURES

j
I
! ,
Fig. B2.72 Effect of temperature on the ultimate tensile strength (Fm) of 7075-T6 aluminum alloy (all products)

Fig. 82.75 Effect of exposure at elevated temperatures on the r-oomtemperature tensne yield strength (Fty) of 7075-T6 aluminum alloy (all prccucts}.

! I
! ,
Fig. 82.79 Effect of temperature on the bearing yield strength (Fbryl of 7075-T6 aluminum alloy (all products).

.'

! ,
Fig. 82.76 Effect of temperature on the compressive yield strength (Fey) of 7075-T6 aluminum alloy (all products).

.. : .0

,
a

,
,~

,-",--.'
"'"

0.00

FIg. B2.73 Effect of temperature on the tensile yield strength (Fty) of 7075~T6 aluminum alloy (all products).

f ,
I . '
!

Fig. 82. 80 Effect of temperature on the tensile and compressive modulus (E and Ee) of 7075-T6 aluminum alloy.

Fig. 82.77 Effect of temperature on the ultimate shear strength (F su) of 7075-T6 aluminum alloy (all products).

1 !

Fig. B2.74 Effect of exposure at elevated temperatures on the roomtemperature ultimate tensile strength (Ftu) of 7075-T6 aluminum alloy (all products).

Ftg. 82. 81 Effect of exposure at elevated temperatures on the elongation of 7075-T6 aluminum alloy (all products except thick extruatons}.

-,

Fig. 82.78 Effect of temperature on the ultimate bearing strength (Fbru) of 7075-T6 aluminum alloy (all products).

Table B2.20 Design Mechanical and Physical Properties of AZ31B Magues tum Alloy (Sheel and Plate)

Alloy

AZ31B Sheet und plate

Form .... Condition .... Thickness (in.)

.

-0

~H26

0.061- 0.250- 0.501- 0.016- 0.250-- 0.375- 0.501- 1.001- 0.250- 0.375- 0.439- 0.501- 0.751- 1.001- 1.5010.0160.374 0.600 1.000 2.000 0.374 0.438 0.600 0.760 1.000 1.600 2.000 0.249 0.500 2.000 0.249 0.060 I-~-I-~-I-- --I--,---f---I-- - - - - - - - - - - - - - - - A B A

,.
Z

BnsiB

A B ------- -- ---------------- -------------- --B

'"
''" "
~
~

Mecha.nical properties: 1",.., kai. L T

. .

32

36

32

36

32

30 16

39

40

...... 40

38 39 26
''29

37 38

36 37 22 26
13
.

34 36 20
23

39
40

38 39 26 29
21

38 39
26 29

37 38 26 28 17

37 38
23 26

35

35

36

36

(.~:~;/':!i;
'{,': 18
12

19 13
23 62 61

16 12 17 60 60 29 29
12

19
10

~
24
18
68

30 26 26 60 70 44
44

r.; kai.
L

24 27
16
18
64

27 30
22

22 25
16

21 24
14

10

20 18 66
66

10

18 18
56

16

TF F_

.
.. . . 17

kai .. klli.
(,/D~

23 62 61 31 31

17
60

18

18

1.6)...

60
60

(,/D~2.0)...........

60
27

68
43

63 ...
34 .

58 68
40 40 8 10 6 8

66 65
39

66
36 36
6

1"6"" kai.
(,/D~1.6).....
(,/D~2.0)

29 29
12

30 30
18

27

43 10

38 '38
8
10

34
8

39
6 8

.
6 6 6

e, percent L T

. ............

19

12

6 8 8 ....

10

8 10 6.~

6 8

o c:

c;

. .'"
~

6.6 2.4
Physlcel properties: w. Ib/in.' C. BTU/(Ib)(F) K. BTU/(I,,)(lL')(FJ/fl
. .
.

''" "

0.0639 0.25 (a.t 78 0 F)

66 1212 to 672" F)

_. 10' in./in./F
Ihthll.tL.....

14 (6f; to 212 0 F)

Trll.llsverllll F tdlowtlbles ere 6fJUlli Io or cruUt than tb, 1lJbIttud.lna1 POI .UO....blu.

(C')
".; iJ\,

B2.28

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEmCLE STRUCTURES

AZ31B

~AG}ITSIUM

ALLOY

(SH~~

& PLATE) (Cont.)

j
1 ..

.I
!

.,

Fig. 82.86 Effect of temperature on the ultimate shear strength (F s u) of AZ3IB-H24 magnesium alloy.

r:=I~o''''

Fig. B2.82 Typlcal stress-strain and tangent. modulus eurves for AZ31B-O magnesium alloy at room temperature (longitudinal).

I .I
.' !
l
Fig. 82. 87 Effect of temperature on the ultimate bearing strength (Fbrn) of AZ31B~H24 magnesium alloy.

j
.I
!
l
Ftg. B2.83 EHeet of temperature on the ultimate tensile strength (F tu ) of AZ31B-H24 magnesium alloy.

j
. . !
l
Fig. 62.84 Effect of temperature on the tensile yield strength (FtyJ of AZ31B-H24 magnesium alloy.

.I

Fig. 82.8a Effect of temperature on the bearing yield strength (Fbry) of AZ318-H24 magnesium alloy.

I
.I

l
Fig. B2. 85 Effect of temperature on the compressive yield strength (Fey) of AZ31B-H24 magnesium alloy. Fig. 82.89 Typical stress-strain and tangent-modulus curves for AZ31B~H24 magnesium alloy at room temperature.

Table 82.21 Design Mechanical and Physical Properties of HK31A Magnesium Alloy (Sheet, Plate and Sand Castings)

Alloy . .
Form . .

. ...
~~--

HK31A
...

. ...
-0
0.01ti0.261-

Sheet and plate

Sand castings

..
Z
OJ

Condition. ....

..

. ...

-1124 0.5011.000
1.0010.016-0.126 0.1260.261-

-TO 0.5011.000
1.0012.001-

~
~

r-

Thick ness, in. . . . .

... . ..... .

0.250
A

0.1iOO

3.000

0.250

0.1iOO

2.000

3.000

... . . .....
(.)

Buala . . . . .

............

" ~

.
W

"'
w
~

t:l t:l
W

30 18 12 22 43 51 24
24

,.
.. . . .....
27 13 13

'"
o
0
~

Mechentce! properties: PI., kai . ..... ... . ..... F,u. kai ..... 1'~u, k"i . ... .. . . . ...... ksi. .. . ... . ... . . ... 1'..... ksi

r.;

32 19 13 22 40
54

30
l'

30 16
10

29
14 10

10 22

22 . .......

34 20 20 23 49 67
34

30 28 21 23 52 00 35 35 0

34

23 19 23 49 57 33 33 4 6.5 . 5 2. 4

36 24 22 23 51 58 30 30 8

34

21 18 23 49 . .... 57 . .... 31 31 4

34 33 18 16 11 13 23 .. . . ....

33 15

...
to

f:

......

11

("/D~I.5).

("/D~2.0)

. .. . . . .. .. .

. . .... . ..
'

43 . ....... 51 ...... . . 21 21 12

..... ... . .......

.. . ....... ..... . . .. ... . . .. . ........ ..... . .. . .... .....

tl

Bi

Fu, ksi
("/D~

1.5) .... ("JD~2.0) ..

.. . .. . .. . . .
.

25 25 20

...... ........ ........ .. . .. ...

12 12

34 4

........ . ...... ........ . .. . ..... ... . . ... . . . . . , . , ... . . .. . ." .......

10

e, percent. . ... " . ........ . . E, 10 pai . . . . EI, 10' pai. . . ..... G, to' p.ti. .. . ... . ... . ... Physical properuea: .... . . ..... w, lb./in.'.

12

10

''" " '" ee '" " "" 'rr

'0

'" r-

'"
c

<:
~

co

1.;
H

'" ''" " '" Z 0 g "

0

0,0647

C, BTlJ/Ob.llF) ... , ... K, BTU/(h,.)(ft.')(F) ft.

:

0.25 (320 to 2120 F)

60.0 (at 68 Ji') 15 (68 0 to 392 0 F)
Mechanl(llli proponlea are ba.sod UJ!'llD tho ruaranteod te.D.1lI1o propertlea from IIllplmllely-(:ll4t tlll& ban. 1'he IIWIchanlcal propenle, Ilf ban cut from eNUnl' U1aY be .. klw as 16 porocnl of the tabulated values .

'" " '"

'"

to-I In./io./I<'. .....

. .....

f'rop"rU~s tor 'I.heet lIud pllliu IlJ'(l lako:!D plUlIllello lile direction of rollin.. Tl'anIVCraoproPOrtIall are equal to or Irellter than tho Irmll;IUldlulll propertloa. RekreliOO III10uld be lIllOdll to tllll q>eclll(l roqulremt:lIl1 of tho procur1P, or oarllt\cllllu, qeDey with rellUd to tho USil of lhe abovo value.lll tho delll,u eaatlDp.

ct'

82.30

MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGIIT VEmCLE STRUCTURES HK3lA l1AGNESI11!1 ALLOY (SHEET, ?LATE & SAND CASTI)iGS) (Corrt . )

"'"

f '" '"
.'
0

t '" e ,
so
~

'"
~

,
0

, '"

s-

,(X)

zoo

XlO

-00 sec T _ '..... F

600

roo

ItOO

Fig. 82.90 Effect of temperature on the ultimate tensile strength (Fta) of HK31A~H24 magneeuim alloy.

Fig. B2.93 Effect of exposure at elevated temperatures on the room-temperature tensile yield strength {Ftyl of HK31A~H24 magnes iurn alloy.

[00

zoo

300

-00

ecc

600

roo

ItOO

T~t_.F'

Fig. B2.91 Effect of temperature on the tensile yield strength (Fty) of HK31A-a24 magnesium alloy.

Fig. B2.94 Effect of temperature on the ultimate tensile strength (Fta) of HK31A~T6 magnesium alloy (sand casting).

Fig. B2.92 Effect of exposure at elevated temperatures on the room-temperature ultimate tensile strength (F tu ) of HK31A- H24 magnesium alloy.

Fig. B2.95 Effect of temperature on the tensile yield strength (Ftyl of HK31A-T6 magnes ium alloy (sand casting).

Ll

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE

STRUCTURES

82.31

AZ61A, AZ63A, AZ80A M.;GNESIUM (EXTRUSIONS, FORGINGS, CASTINGS)

Table 82.22 Design Mechanical and Physical Properties of AZ61Aa Magnesium Alloy (Extz-ustcns and Forgings)

Alloy.

AZ6IA

Form.

Extruded bar, rod, and solid shapes

Extruded Hollow shapes

Extruded Tubes

ForginC

Condition.

........ ........

'

-F

Thickness, in ....

<:0.249

0.250-2.499

0.02lHl.750

Besis ...

... . . ....

Mechanical properties: FhI , kn L.

38

39
24

36

36

r.; hi
L ..

T. T

..
16
..

38

.. ..

21
..... . ..

16
11
.

22

F11. k.ri L.

T. . .
F,," m. FIIrv, k$i

... .......

.....

I'

I'
19
45 55 28 32
9

11

I' ..
19
50 60 28 32

19
.........

(,/D -1.5).
(,jD-2.0). F.... kM (,j D -1.S). (,jD-2.0).

'5 55
28

. .. .

32
8
.

e, percent .. E, 10- psi. Ee, 10' p"':. G, 10' p.ri.

6.3 6.3 2.4

Physical properties: w,lb/in. t . C, Btuj(lb)(F). K, Btuj[Cb,)(fto)(F)/ftj. ~, 10-' ir-../in./F

<J PToperti~s
tll.ll:~n oaralI~1

0.1)647 0.25 (at is''F) Ir 46 (212" to 5i2"F) 14 (65" to 212"F)

during fabrication, ~ Estimated.

'-or extruded b&l'S, rods. shapes, tubes. and fol'ltings are to the dlrecttcn of ~xtru.sion or maximum ffil'tAI new

B2.32

MECHANICAL AND PHYSICAL PROPERTmS OF METALLIC MATERIALS FOR FLIGHT VEffiCLE STRUCTURES

AZ61.A , AZ63A, AZ80A MAGNESIU11 (:::XTRVSIONS, FORGINGS, CASTINGS) (Cont , )

Table B2.23 Destgn Mechanical and Physical Properties of AZ80Aa Magnesium Alley (Extrusions and Forgings)

Alley.

.U80.'
Extruded bars, rods, and solid shapes Forgings

Form.

I
i

Condition.

-F
....
,.,

-T5

-F
1.5002.499

-T5

Thickness, in ....

...

<'0.249

0.2.101.499

1.5~

<'0.249

0.2501.499

2.499

Huia... ........

.. ............

Meehaniea1 properties: F ta kn L .. . ........ ... ....... T. ............... . . ... Fly, k.ri L .. ............ . .. .... T ... . .. . .
' 4

43

43
28

43

..

47
30

48
33

48
33
..

42 2" 18
20

42
28

28
..

28

JI~

kn L .. . . ... .. . . .. . . .. . . . T ....... ....... . .. .. p... hi. ......... .......

'

.
17 19 48
56

..

17 19 '8
56

..
20

28

27

19 48
56

20

..

25 ,0
50

20

PIw.,kn (.ID-1.5). ... ...... (.ID-2.0). ........ F.... kri (.ID-1.5) .. ......... (.ID-2.0) .. ........ e, percent...... ..... ........ E, 10' pIi......... .... B" 10' pn. ..... ... . ..... G, 10 1 p:ft... ........... .. .
,

..

70 42

36
40

36
40

36
40

4 6.5 6.5 2.4

..

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

B2.33

AZ6lA, AZ63A, AZSOA MAGNESIUM (E:CTRUSIONS, FORGINGS, CASTINGS) (Cant.) Table B2. 24 Material Specifications for AZ63A Magnesium Alloy

'''.. _~
A_r,n,il. 1I

e.com_ _

-,train

Il

-ttrQIn

C. ren ..l.

lan~l' ~ ..l_

o'C_'~lft04IIl",

Specification

Type of product

QQ-M-56 ..
QQ-M-OO ....

Sand' castings Permanent-mold cutU1gs

TABU: 4.2.3.0<"). Denp MccA4nical aM PhyftctU P-ropmiu of AZS3A Mag7Ulnum- Allov (elUting_)
Alloy.
. ....

AZ63A

s.
'.

Form.

Sand and permanentmold castings

5M*, QOOl in.A\. Torw,.nl MadIll.... tf psI

Condition. ...

---I - - --- --..

-F

-T4

-T5

-T6

Fig. 82.96 Typical stress-strain and. tangent-modulus curves for AZ63A-T4 map,estum alloy (sand casting) at room temperature.

Thickness (in.) ..
Basi8~

A.r.....II. ttnn-,traln

'-'"

.. ...... . ......

e.c~" C.Ttnlil, ~

...-~n
lr'OClul...

Do~"",~

Mechanical properties:
F tur lui .... . . .. . . ... Phi. k..n.. .. . - ........ Ft:fI, Jeri .. ... . .. . ." . F,.." kai ... F/wv., kn (,/D-L5). (,/D-2.0) .. Fl1rT/.kn (,/D-L5). (,/D-2.0) . ..... t, percent .. E, 10- pn.. Ee 10- pn... G, 10 pori.

24
10 10

34

24

34

10
10

11 11

16
36

17 36 50 32 36 7 6.5 6.5 2.4

16 16 19
50

50
28 30

..

65

36 45 3

'.

$traWl,QOOI inA\ TGftqenf Mocl"~ 10 ""

Fig. B2.97 Typical stress-strain a.nd tangent-modulus curves for AZ63A-F magnesium alloy (sand casting) ar room temperature.

Table B2.25 Design Mechanical and Physical Properties of 8Mn TItanium Alloy

Alloy

8Mn

J
J
!
1

j
J
1
1000 1200 1400 1600

<xl ~ w

..
>t-

~ o

Form

Sheet, plate, llud strip

~ n

Coudiuon

AlIllealed
Fig. 82,98 Effect of temperature Oil the ultimate tensile strength (Ftu) of BMn annealed tttantum alloy.

Thickllt'::lll, ill

Fig. 82,101 Effect of temperature on the ultimate shear strength (Fau) of 8 Mil annealed titanium allOy.

~ '~ "
00

ljasis

~
o-J
H

o >r'o

A
'00

Mccllll.uieal properties: 1",.., k~i L

I~O

''" " ... ' " i'l s ,o-J

'"

~ H

120

rr-

1"'11' bi

I-

flO 110
110
200 400 600 lIOO lCOO 1200 1400 1600 200 400 600

'"
tn

1"~II' kei

L T p...,
k~i

~ ~

'" 0 '" ;: ... '>"

rt-

eoc

ICOO

1200

1400

1600

110

T'~f"'_,f

1.,,,p.''''u<t. f
Fig. 2.102 Effect of temperatur-e all uie ultimate
hear-lug st rength (to'bru) of 8 Mil annealed titanium

8'
110

n.; k.~i
e,

Ftm.. /.:~i (e/l) = 1.5) (el D =2.0)

Fig, 82.99 Effect of temperatur-e Oil the tensile yield strength (Fty) of 8 Mn annealed titanium alloy.

o-J

r
H

'U

n ~ ... M
~

to
"3

alloy.

(el/) = 1.5) (el D = 2.0)

..

130

!
10 15.5

Iwr"l~nt

E, 10 0 pei Re , 10" JI$i G, HI" JI$i

Physical properties:
w,lll/ill.' C, Btu/(Ib)(r').

ro.u

j
.:
I
1000 1200 1400 1600

,00

eo eo
eo

I,
I

r"' co '" '" 0 '" '"

'0

<

,Ijjijjjilf
,
J

' ,,:I ~ IJ '

!ii::~

Cl
M

r-

'"

;'j

m
M

0.171
0.118 (ilL 68"1<')

S ,

" eo
0 0

.,. \I' Iii .nun

,~"

.@!JW
,:!,!I!:;

n r-

'"
'" '"

lillirii

..'

... '"
"00 "00

1\,IHu/{(hr)(ft')(I<')/ftI a, 10- 0 ill./iu./F.

e.a
4.8 (ilL 2IXjOJ")

'00

,00

"00

o " ... ct

Fig. 82.100 Effect of temperature 011 the compressive yield strength (l<'cyl of 8 Mu annealed trtanium alloy.

Fig. 2.103 Effect of temperature Oil the tensile and compressive modulus (~ and Ed of 8Mn tltaniutu alloy.

'"
,/

ANALYSIS AND DESIGN

OF

FLIGHT VEHICLE STRUCTURES

B2.35

6Al-4V TlTANIUl1 ALLOY (BAR & SHEEr)

Table 82.26 Des tgn Mechanical and Physical Properties of 6Al-4V Titanium Alloy

Alloy. ...... . . .. .... . . ....

6AHV
Bar

Form.... ......... . Condition..

Thieknese , in.

.....

I
Annealed.

Rheet.

...

. ....

....

<1.5

<:0.181

l -"I(lQ

000.:10

tWO

r_ t

IlaoUo ........

...

. ....

Fig. B2.104 Effect of temperature on Ftuo Fty and E of 6Al-4V annealed titanium alloy (sheet and bar)

~Ieebanical

pecperties:

... . .... . ... . ....

fi't.. hi L. T. .. .... ... F'll, hi L .. .. .... T. ......

130
..

130 130 120 120 126 126 7. 191 244 163 198
10
Fig. 82. 105 Effect of temperature on the compressive yield strength (Fey) of 6Al-4V annealed titanium alloy (sheet and bar)

120

r.; kJti

..

L.. .... T. ..... . ... . .....

1"... ksi.
p"",, k.si
....... . ....

126
80

L .. .... . .. ... T. . . ..... .. ..... E(, 10- pn L .... .. . ... . .. T . . ..... . .... G, 10- pai.. . . ..... ....
Physical properties: lIf, Ib/in.'. ....... C, Btu/llb)(F) ... K, Btu/[(h,)(lt')(FI/ltj. II, 10-' in./in./F..

(,!D-2.0) ..... .. p.... kn (,I D -1.51. ... (,ID-2.0) .. .... t, percent .. ........... E, 10- pn

(,!D-1.5I.

196 248 174

205 10

16.0
..

15.4
16.4

16.4
6.2

16.0
16.9

..

0.160
0.13S (at. 68F) 3.8 (at 63F) 4.6 (at ZOOOF)

Fig. B2. 106 Effect of temperature on the ultimate bearing strength (Fbru) of 6Al~4V annealed titanium alloy (sheet and bar).

-'

J j

Fig. B2. 107 Effect of temperature on the Ultimate shear strength (Fsu.l of 6Al-4V annealed titanium alloy (sheet and bar).
~1'::t
~"---

:><...

B-2. 36

MECHANlCAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES

INCONEL X NICKEL ALLOY (SHEE'r')

Table 82.27 Design Mechanical and Physical Properties of Inconel X Nickel Alloy

Alloy.

Inconel X

Form.

Sheet

Condition.

. ......

Precipitation heat-treated

Thickne8!, in .......... . .
Basis.. .... . .... .... . ...

Fig. 82. 108 Effect of temperature on the ultimate tensile strength (Ftu) of precipitation heat treated Inconel X nickel alloy.

~Ieehanical

properties:

L. ...... . ..... T. . . . .. . . ..... FtJIokri L ... ... . . . .. . .. T. ....... . fi'~/IJ hi L ... ..... T. ...

P,... hi

155 155
100 100

Fig. 82. 109 Effect of temperature an the tensile yield strength (Fty) of precipitation heat treated Inconel X nickel alloy

105
105

F,.. ksi.....
Fbm. kin
(t>ID-1.5). (,fD-2.0). .. F1)r", kri (t'/D-1.5).

lOS
286

(,/D-2.0) e, peeeent. E, 10- pri.. ... . ....... E~ 10-pn. .... G, 10- pn.. ........ .

186
20

31.0 31.0

Fig. 82. 110 Effect of temperature an the compressive yield strength (Fcyl of precipitation heat treated Inconel X nickel allay.

Phy&ical properties: w, lb/in.e. C, Blu/ObHF). K, Blu/[(lu-)(fl')(F)/ft] z, 10.- in./in./F ...

0.304 0.109 8.7 (80 to 212F) 6.4 (100 to 200F)

'_f
Fig. 82. III Effect of temperature an the ultimate shear strength (F su) of precipitation heat treated Inconel X nickel alloy.

Fig. 82. 112 Effect of temperature on the tensile modulus (E) of Inconel X nickel alloy.

'-'

Fig. 82. 113 Effect of temperature on the ultimate bearing strength (FOru) of precipitation heat treated Inconel X nickel alloy.

-,

,-,
Fig. 82. 114 Effect of temperature an the bearing yield strength (FbryJ of precipitation heat treated Inconel X nickel alloy.

PART C

PRACTICAL STRENGTH ANALYSIS & DESIGN OF STRUCTURAL COMPONENTS

CHAPTER C1

COMBINED STRESSES.
Ct. 1 Uniform Stress Condition

THEORY OF YIELD AND ULTIMATE FAILURE.

~l ~l

Aircraft structures are subjected to many types ot external loadings. These loads otten cause axial, bending and shearing stresses acting simultaneously. It structures are to be deSigned satisfactorily, combined stress relationships must be known. Although in practical structures uniform stress distribution 1s not common, still sufficient accuracy tor design practice 1s provided by using the stress relationships based on uniform stress assumptions. In deriving these stress relationships, the Greek letter sigma (a) will represent a stress intenSity normal to the surface and thus a tensile or compressive stress and the Greek letter tau (~) will represent a stress intenSity parallel to the surface and thus a shearing stress.
Cl.2 Shearing Stresses on Planes at Right Angles.

~llD[J
~l

\4~O

..+

g;:.~~
0

,+

l~l

Fig. Ct. 3

o'

Fig. Ct.4

stresses. Fig. Cl.4 shows a tree body atter the block has been cut along a diagonal section. For eqUilibrium the sum of the forces along the x-x axis equals zero.

0(1) + 2 (~l cos 45 0) cos 45

hence,
0

=0
(2 )

=2

('t"

1 cos 45 0) cos 450 _ i - -r

2 1
B

't"zlbr't"z
2 1
Fig. Cl. 1

Therefore when a point in a body is subjected to pure shear stresses or intensity 't", normal stresses or the same intensity as the shear stresses are produced on a plane at 45 0 wit~ the shearing planas.
Cl.4 Principal Stresses

Fig. C1.2

Fi3. Cl.l shows a circular solid shatt subjected to a torsional moment. The portion (A) at the shatt exerts a shearing stress 't"z on section (1-1) and portion (B) exerts a resisting shearing stress 't"z on section (2-2). Fig. Cl.2 illustrates a differential CUbe cut from shatt between sections (1-1) and (2-2). For equilIbrIum a reSIsting couple must exist on top and bottom face at cube. Taking moments about lower left edge of CUbe:
~x dxdy (dz ) ~z dzdy (ox )

For a body Subjected to any combination ot stresses 3 mutually perpendicular planes can be found on which the shear stresses are zero. The normal stresses on these planes of zero shear stress are referred to as principal stresses.
Cl. 5 Shearing Stresses Resulting From Principal Stresses.

=0
- - - (1)

hence, 't"x

='t"z

- - - - - - -

Thus If a shearing unit stress occurs on one plane at a point in a bOdy, a shearir~ unit stress ~f same intenSity exists on planes at right angles to the first plane.
C1. 3 Simple Shear Produces Tensile and Compressive Stresses.

In Fig. Cl.5 the differential block is subjected to tensile principal stresses ax and Oz and zero principal stress a. The block is cut along a di~gona~ section glvlng the free body of F1~. C1.6. The stresses on the diagonal section have been resolved into stress components parallel and normal to the section as Shown. For eqUilibrium the summation of the stresses along the axes (1-1) and (2-2) must equal zero.
ZF ~-1.

=0
Q

an dudy - ax dzdy cos

- Oz dydx sin
Q + az dyctx

=0,
Q

Fig. Cl.3 shows an el~mentary bloc~ of unit dimensions SUbjected to pure shearing
e1.l

Whence an = ax dzdy COS dudy

sin dudY

C1.2

COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.

0" z dydx

ct. 6 Combined Stress Equations

Fig. Cl.7 shows a different1~1 block subjected to normal stresses on two planes at r~g~t angles to each other and with shearing forces on the same planes. The maximum no~~l and shearing unit stresses will be determined. Fig. Cl.8 shows a free bodY dia3ram of a portion cut by a diagonal plane at angle Q as shown.
z dydx

/
/

2,
I
, I
'I

I <Judy

: . "'o&~
/

g
/

"
/

---"-,--

, , / ,,/

,,
2
Fig, CL 7 I

a z dydx
Ftg.

cr.s
an dudy

dzdy _ dydx _ But dudy - cos Q and dudy -

e in

Q,

whence

an = Ox cosa'1 + Oz s1n ag - - - - - - - - - (2a) The normal stress an at a po1nt 1s always less than the ~tmum princ1pal stress Ox or o"z at the po irrt ,

,,
Q/ / / '!// /
/

-,
I

~;;;"--

--

ZF._. = 0
ax
But dzdy dUdy

Fig. Cl.B

dzdy sin 9
Q

+ o"z dydx

cos

=0

= cos

and dydx

dudy

= sin

For eqUilibrium the sum ot the forces in the z and x directions must equal zero.
ZFx

=Oz sin Q cos Q - ax cos ' " or, - ox) sin Q cos Q '" = (Oz or, (1/2) (az - ax) sin 2 Q, = ' " maximum principal stress and ax Is
hence, principal stress. Since sin 2 Q is maximum

Q sin Q

0
Q -+ '!

where Oz Is :nin!::!l.lm

an dudy cos =0
ZFz

dudy sin Q - ax dzdy - 't'xzdxdy

--------(4)

----0

an dudy sin
when Q :: 45 0 ,
(3 )

Q - ~ dudy cos Q a z dydx ~xzdzdY =0---- - - - - - - -(5)

'" = (amax

- amin )/2 - - - - - -

By dividing each equation by du and noting that

Stated in WOrdS, the maximum value of the shearing unit stress at a paint in a stressed body 1s one-halt the algebra1c differences of the maximum and minimum principal unit stresses.

dzdy cos dudy -

and dydx dudy

= sin

we obtain:
Q

(an-Cx) cos g + ('" - "'xz) sin

- - - -(6)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

Cl. 7 Mohr's Circle for Determination of PrinCipal Stresses.

Cl.3

(an - azl sin g - (T - TxZ) cos g

- - -(7)

The maximum normal stress an will be maximum when Q equals such angle Q' as to make ~ = zero. Thus if ~ = a and Q Q' in equations (6) and (7), we obtain,

(an

ax l cos g' azl sin g'

'txz sin g'

(an

TxZ cos g'

=a - - - =a - - - -

-(8) -.(9)

In equations (8) and (9) an represents- he principal stress. Dividing one equation by another to eliminate Q' , an-ax:

"xa

"t'xz on - Oz

whence,

On

=ax

+ Oz + ~ '(ax 2 --V

ZOZ )

+'txz- - (10)

In equation (10), tensile normal stress is plus and compression minus. For maximum an use plus Sign before radical and minus sign tor minimum. on. To tind the plane or the principal stresses, the value of Q' may be solved for trom equations (8) and (9), which gives: tan 2
Q'

It is sometimes convenient to solve graphically tor the principal stresses and the maximum shear stress. Mohr's circle furnishes a graphical solution. (Fig. C1.9a). In the Mohr method, two rectangular axes x and are chosen to represent the normal and Shearing stresses respectively. Taking point a as the origin layoff to scale the normal stresses ax and Oz equal to OB and OA r-espeot t vely. It tension, they are laid ott to right of point a and to the lett if compression. From B the shear stress ~xz is laid ott parallel to Oz and with the sense of the shear stress on the face DC of Fig. Cl.9b, thus locating pOint C. ~ith point E the midpoint of AS as the center and with radius EC describe a circle cutting OB at F and G. AD will equal Be and will represent the shear on face AS of Fib. b. It can be proven that OF and OG are the principal stresses Omax. and Omin respectively and EC 15 the maximum shear stress "t'max.' The prinCipal stresses occur on planes that are parallel to CF and CG. (See Figs. c and d). The maximum shear stress occurs on two sections parallel to CH and CI where HEI is perpendicular to OB. Ifox should equal zero then a would coincide with A.

n(max)

2 ~xz ax - Oz

----(11)

Q' 1s measured :rom the plane at the largest normal stress ax or oz. The direction of rotation of Of trom this plane is best dea. termined by inspection. Thus if only the shearing stresses 'txz were acting, the maximum. 0 LTc principal stress would be one or the 45 0 planes, the particular 45 0 plane being eaSily deter(e) mined by inspection or the sense ot the shear o~,..,o1') stresses. Further.nore it only the largest normal stress were acting it would be the maximum principal stress and Q' would equal zero. (d) T Thus 1t both a and 't act, the plane at the principal stress will be be~Neen the plane on ZT Tc which acts and the 45 0 plane. As stated a. betore cr refers to either ax or Oz whichever is the largest. Z

I
I
I

J1~

I
D

,
/

J;foX:

'/ R

.-Lax-az ,--2

(zj

Fig. Cl.9

MaXimum Value or Shearing Stress. The maximum value or equals, "max,

1:'

(~max.)

tram equation (3)

---(12)

Cl. 8 Components of Stress From PrinCipal Stresses by Mohr's Circle.

(an (max. ) - an("ln.) )/2

Substituting the maximum and minDnum values at an tram (10) in (12), we obtain maxDnUD. shearing stress as tallows:

In certain problems the prinCipal stresses may be known as in Fig. Cl.9 and it 1s desired to tind the stress components on other planes deSignated by angle Q. In Fig. Cl.lI the axes x and z represent the normal and shear stresses respectively. The principal stresses are laid orf to scale on ox giving points D and E respectively. Construct a circle with A the midpoInt at DE and with diameter ED. Draw angle CAB equal

C1.4

COMBINED STRESSES.

THEORY OF YIELD AND ULTIMATE FAILURE.

to 2Q. It can be proven that 08 represents the normal stress on the plane defc of Fig. CI.IO, and CB represents the shear stress 't 'on this plane.

't xz= 5000

Fig. Cl.12

0x-!--

"g

,J~
,

! /

---

-r-o-x
g

z
I

,'-

e
Fig. C1.10

0-.

crz. ax are principal

stresses.

On on plane (defc)

---j
I 0n(min) :;.
: -2070
A

.___.. I
Ox 10000
D B

+ - an(m a x)=120 7 Z

-jLJ
+

Fig. C1. 13

I I
I

Substituting values,

an

= 1000~

+ 0

V(1000~

O)~

. . 5000" = 5000

---+1OX20,~
Fig. Cl.l1 Cl. 9 EDmple Problems.

... 7070 hence, 0n(max.) : 5000 ... 7070 : 12070 pSi

On(mln.) = 5000 ""'X. 7070 = - 2070 pSI

(1/2) (On (max. )

On(mln.)) (ReLEq.l2)

= (1/2)(12070
Example Problem 1. The maximum normal and shear stresses will be determined tor the block loaded as Shawn in
Fig. Cl.12.
The graphical solution making use of MOhr's circle is shown in Fig. CI.13. From reference axes x and z thru paint 0, the given no~al stresses Ox = 10000 is laid ott to scale on ox and toward the right giving point B. From B the shear stress 'txz = 5000 is laid oft parallel to oz to locate point C. With E the midpoint or OB as the center at the circle and with radius EC a circle is drawn which cuts the Ox axis at F and G. The maximum and minimum. principal stresses are then equal to of and oG which equals.12070 and -2070 respectively. The maximum shear stress equals EC or 7070.

- (-2070)) = 7070 psi

'tmax.can also be computed by equation (13), whence,

""'X. =

vieOOO~ .

0)" 5000' - - 7070 psi

tan 21;11'1 hence,

=Ox 2 'tXZ = 2X5000=1 10000 - 0 - Oz Q~ =22.5.

Example Problem 2.
T~e maximum normal and shear stresses will be determined for the block loaded as shown in

Fig. Cl.14.

~
II

O"z

= 20000
I

'

't'xz '" 12000

Algebraic Solution:

From eo , (10),
Ox' 10000

~lDl~
~ 't'xz
I

Ox = 10000

= 12000

Fig. C1. 14

O'z .. 20000

--

--------------_.._--.. I.
ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES
0
3 -

_-~
ci. 5

Fig. Cl.15 shows the graphical solution using MOhr's circle. From ,otnt 0, Ox = 10000 and O'z = -20000 are lald off equal to 08 and OA
respect 1 ve Ly ..

(ox
2

Oy

oz)a~ + (OXOy + oyaz + 0xoz

-r YZ

_Yx:y2)a -

(OXOyOz

2YyzYxzYXY - 0XYyz-;J

?arallel to OZ at A locating C. rlith ~ the mtdpo tn t of AS as the center of a circle of radius EC a circle 1s drawn which cuts the ox axis at F and D. The maximum no~al and shear stresses are indicated on the rlgura.

TXZ

equal to 12000 is lald off

- OyYxz - OzYXya

=0

- - - - - - - - - ~ -(14)

z
Fig.

cr.is
y

Fig. Cl. 16

Q''''180~15'''' _

--.i..

I I
I

Fig. Cl.l? shows the principal stress system which replaces the system of Fig. Cl.16. It can be shown that the maximum shear stress ~. is one of the following values.
-r:nax.=::t(OL all)
- 0'3)

ron(max)' -2420

~ ~ ;:::to",,",] "n(mini1420~
"
+ Oz

or r: max.
cr
t

+ 1 CO. = -"2 + 1 = -"2

- - - - - -(15)

max,

(0, - OL)

Algebraic Solution

=ax
= 10000
2

- az)" + -rxz ..

- 20000 :!: .,j(1000o-~-20000)) ', 12000'

The planes an which these shear stresses act are indicated by the dashed lines in Fig. Cl.IS J namelYJ adhp., bdge and dcef. The largest or the shear stresses in equations (15) depends on the magnl~ude and Signs or the principal stresses J rememberin~ that tension is plus and compression is ~inus when making the SUbstitution in equations (15).

5000 t 19200

hence, On(max. ) = -5000

n(mln. )

-5000 +

19200 = -24200 pet 19200 14200

.:~
q, '"
Fig. Cl. 17

't"max.
Q'

= :t

19200,
_ 50'

tan 2Q'

x 12000

= ISO

1000 - (-20000)

.8

C1. 11 Principal Strains

Cl. 10 Triaxial or Three Dimensional Stresses

For bodies which are stressed in three directions, the state or stress can be defined completely by the six stress components as illustrated in FIg. Cl.16. USing the same procedure as was carried out for a two-di~en sional stress system, it can be shown that there are three prinCipal stresses O'LJ a. and 0., whose values are the three roots of 0 in the following cubic equation.

The strains under combined stresses O3,':'"e usually expressed as strains in the direction or the prIncipal stresses. Consider a c?se or Simple tension as illustrated in Fig. Cl.l9. The stress 0 1 causes a lengthening unit strain e in the direction or the stress OLJ and a Shortening unit strain ~' in a direction at right angles to the stress OL. The ratio of e' to e is called Poisson's ratio and is usually given the symbol~. thuS,
IJ.

= ~I /e

Cl. 6
E

COMBINED STRESSES.

THEORY OF YIELD AND ULTIMATE FAILURE.

2.-1.--

tcr,

: I, '
,

--~

'
cr,
E'
/

, ,

)/

-i--cr,
Fig. Cl. 20

The strain energy can be expressed in terms of stress by substituting values ot E in terms of G from equations (17) into equation (19), which gives,
U

E'

2"

:",-r--" 01. Fig. Cl. 19

Since e

:-,

2"

cr,

lcr,

=..l.. ~ 2E

0:+ 0: - 2iJ.(01.0.... G,GIS

+ 0,,01.)

- - - - - - - - - - - - - - - - - -(20)
O"~ ~ 0

= 04/E,

we obtain,

For a two dimensional stress system, and equation (20) becomes

- - -

- -

~(16)

Now consider the cubical element in Fig. CI.20 sUbjected to the three principal stresses cr~, cr. and G e , all being tension. The total unit strain e 4 in the direction or stress cr~ will be expressed. ObViously, 01. tends to stretch the element in the direction or o~ whereas stresses a. and a~ tend to shorten the element in the direction or a~, hence,
f:l.

C1. 13 Structural Design Philosophy. Limit and Ultimate Loads. Factors of Safety. Margin of Safety.

e,

'=-

-Et - ~ - ~ , whence -T - pa. + oe)

e~,
'Z

and Similarly for e. and

CSII

cr. - E'O'e u. E ...

0'1.

-----(17)

For a two-dimensional stress system, that is, etr-essee acting in one plane, as = 0 and the principal strains become,
1 E, = - (0 E '
E. = E.

- flO. )
-----(18)

t (cr. - flO,)
1
(a~ +

:; E

The basic philosophy governing the structural design of a flight vehicle is to develop an adequate 11ght weight structure that will permit the vehicle to accomplish the operations or missions that were established as deSign requirements. The jOb of a cammercial airliner is to carry passengers and cargo fram place to place at the lowest cost. To carry out this job a certain amount ot flight and ground maneuvering is required and the loads due to these maneuvers must be carried safely and efficiently by the structure. A military fighter airplane must be maneuvered in flight far more severely to accomplish its desired job as compared to the commercial airliner, thus the flight acceleration factors for the military tighter airplane will be considerably higher than that of the airliner. In other wordS, every type of flight vehicle will undergo a different loan enVironment, which may be repeated frequently or infrequently during the lite of the vehicle. The load environment may involve many factors such as flight maneuvering loadS, air gust loads, take off and landing loadS, repeated loads, high and low temperature conditions, etc. Limit LoadS. Limit loads are the calculated maximum loads which may be SUbjected to the flight vehicle in carrying out the job it is designed to accomplish during its lite time of use. The term limit was no doubt chosen because avery flight vehicle is limited relative to the extent of its operatiOns. A flight vehicle could easily be deSigned for loads greater than the limit loadS, but such extra strength which is not necessary ror safety would only increase the weight at the structure and decrease the cammercial or military payload or in general be detr1~ental to the design. Factor or Safety, Factor Of safety can be defined as the ratio considered in structural design of the strength of the structure to the maximum calculated operational loadS, that is, the limit loads.

cr. )

Equations 17 and 18 give the strains when all the prinCipal stresses are tensile stresses. For crnnpressive prinCipal stresses use a minus sign when substituting the principal stresses in the equations.

cr, 12

Elasttc strain Energy

The strain energy in the elast1c range for the unit CUbe in Fig. Cl.20 when subjected to combined stresses is equal to the work done by the three gradually applied principal stresses Ol" cr." and O"IS. These stresses produce strains equal to el.' e. and Ee and thus the work done per unit volume equals the strain energy. Thus if U equalS the strain energy, we obtain,
U

0,,2';) - - - - - -

-(19)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cl. 7

Yield Factor at Safety. This term is detined as the ratio of the yield strength of the structure to the limit load. Ultimate Factor of Safety. This term is defined as the ratio of the ultimate strength at the structure to the limit load. Yield Load. This term Is defined as the limit load multiplied by the yield factor of satety. Ultimate Load. This term can be detined as the limit load multiplied by the ultimate factor of safety. This resulting load is often referred to by engineers as the design load, Nhich is misleading because the flight vehicle structure must be designed to satisty both yield and ultimate failure and either one may be critical. Yield Margin or Safety. This term usually expressed in percent represents the additional yield strength at the structure over that strength required to carry the limit loadS. Yield Margin or Safety = Yield Strength _ 1 L1m1t Load Ultimate Margin or Sarety. This term usually expressed in percent represents the additional ultimate strength of the structure over that strength required to carry the ultimate loads. _ Ultimate Strength Ultimate Margin at Satety - Ultimate Load
Cl. 14 Required Strength of Flight Structures.

incapable of performing its required function. Failure may be due to rupture or cOllapse or due to excessive deflection or distortion.
Cl. 15 Determination of the Ultimate Strength of a Structural Member Under a Combined Load System. Stress Rat1olnteraction Curve Method.

Since the structural designer or flight vehicles must insure that the ultimate loads can be carried by the structure without failure, it is necessary that reliable methods be USed to determine the ultimate strength of a structure. Structural theory as developed to date is in general SUfficiently developed to accurately determine the ultimate strength ot a structural member under a single type at loading, such as axial tension or ccmpression, pure bending or pure torsion. However, many of the members which compose the structure at a flight vehicle are SUbjected Simultaneously to various combinations at axial, bending and torsional load systems and thus a method must be available to determine the ultimate strength ot a structure under combined load systems. A strictly theoretical approach appears too difficult for solution since failure may be due to overall elastic or inelastic buckling, or the local elastic or inelastic Instability. The most satisfactory method developed to date is the so-called stress ratio, interaction curve methOd, originally developed and presented by Shanley. In this method the stress conditions on the structure are represented by stress ratios, which can be considered as non-dimentional coefficients denoting the fraction of the allowable stress or strength for the member which can be developed under the given conditions ot CQmbined loading. For a single Simple stress, ratio can be expressed as, R = stress ratio
tt~

Under Limit Loads:The flight vehicle structure shall be ,designed to have surrtc ient strength to carry simultaneously the limit loads and other accompanying enVironmental phenomena tor each design condition without undergoing excessive elastic or plastic deformation. Since most materials have no definite yield stress, it is crnmnon practice to use the unit stress where a .002 inches per inch permanent set eXists as the yield strength of the material, and in general this yield strength stress can be used as the maximum stress under the limit loads unless definitely otherNise specified. Under Ultimate Loads:The flight vehicle structure shall be designed to withstand simultaneously the ultimate loads and other accompanyIng environmental phenomena without failure. In general no factor of safety is applied to the environmental phenomena but only to the limit loads. Failure of a Structure:This term in general refers to a state or condition of the structure which renders it

stress

-(22)

where r is the applied stress and F the allowable stress. The margin of satety in terms or the stress ratio R can be written,
1 M,S:il-l. O

- - - - - - - - - -(23)

Load ratios can be used instead of stress ratios and is often more convenient. For example ror axial loading, R P/P a , where P = applIed axial load and Fa the allowable load.

;, o

~-;.

",,-...1 ...;

ci. 8

COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.

For pure bending, R = MlMa , where ~ = applied bending moment and M a the allowable bending ~oment. For pure torsion, R TITa, where T is applied torsional moment and Ta the allowable torsional moment. For combined loadings the general conditions for failure are expressed by Shanley as follows:R~ + R; + R~ + -----

1. 2. 3. 4. 5. 6.

~3X1mum Principal Stress Theo~y Maximum Shearing Stress TheorJ Maximum Strain Theory Total Strain Energy Theory Strain Energy of Distortion Theory Octahedral Shear Stress Theory

The reader may review the explanation and derivation ot these 6 theories by referring to such books as listed at the end of this chapter. Test results indicate that the yield strength at a point in a stressed structure is ~ore accurately defined by theories 5 and 6 followed in turn by theory 2. Since theories 5 and 6 give the same result, they mi~ht be considered as the same general theory. In this chapter we will only give the resulting equations as derived by theory 6, since theories 5 and 6 appear to be the theories used in tlight vehIcle structural deSign.
Cl. 17 The Octa..hedral Shear Stress Theory.

= 1.0

- -(24)

In this above expression, R1 , R. and R3 could reter to compression, bending and shear and the exponents x , s, and z give the relationship tor combIned stresses. The equation states that the ra i ture of a structural member under a combined loading will result Only when the sum ot the stress ratios Is equal to or greater than 1.0. For some ot the stmpler combined load systems, the exponents of the stress ratios in equation (24) can be determined by the various well known theories of yield and ra tjur-e that have been developed. However, in many cases or combined loading and tor particular types or structures the exponents In equation (24) must be determined by making actual failure tests of combined load systems. SInce the stress ratio method 'NaS presented by Shanley many years ago, much testing has been done and as a result reliable interactiop equations with known exponents have been obtained for many types or structural members under the various combined load systems. In a number ot the tollowing chapters, the interaction equations which apply will be used in determining the ulttmate strength deSign of structural members.
C1. 16 Determination of Yield Strength of a. Structural

Since this theory gives the same results as the well known energy of distortion method it is often referred to as the Equivalent Stress Theory. The octahedral shear stress theory may be stated as follows:- In elastic action at any point in a bOdy under combined stress action begins only when the octahedral shearing stress becomes equal to 0.47 fe' where f e is the tensile elastiC strength of the material as determined from a standard tension test. Since the elastic tensile strength is somewhat indefInite, it Is common practice to use the engineering yield strength Ft y In this theory it is assumed that the tensile and compressive yield strengths are the same. Figs. Cl.21 and Cl.22 illustrate the conditions ot equilibrium involving the octahedral shear stress. In Fig. Cl.21, the cube Is subjected to the 3 principal stresses as shown. A tetrahedron is cut irom the cube and shown in Fig. Cl.22. Three ot the sides of this tetrahedron are parallel to the
3
3

Member Under a Combined. Load System.

As explained in Art. Cl.14, the flight vehicle structure must carry the ltmit loads Without yielding1 which in general means the Yield strength ot the material cannot be exceeded when the structure is subjected to the limit loads. In some parts of a flight vehicle structure tnvoLvtng compact unf t or pressure vessels, biaxial or triaxial stress condit~ons are often prOduced and It is necessary to determine whether any yielding will occur lUlder such combined stress action when carrying the limit loads. For cases Where no elastic instability occurs, the follOWing well known theories of failure have been developed.

I
I, -

I -+--1,
2
I,

,-

, ,
,

==--2

Fig. Cl. 21

Fig. ci. 22

-n~-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STH

CTURES

C1.9

prinCipal axes, while the n~rmal to the fourth side makes equal angles with ~he principal axes. The octahedral shear and normal stresses are the resulting stresses on the fourth side. The equation for the value of the normal octahedral stress 15, foct-3'(f1.+fa+f,,)
_ 1
t s yx
------(25)

For a triaxial stress system J

For a biaxial stress system,

=
j'

0
ll = ,; f~ + f~ - fxt z + 3f sxz

f y

fS yz J

- - -(32)

The equation for the octahedral shear stress 15,

C1. 18 Example Problem 1.

Now the octahedral shear stress is 0.47 of the normal stress. Let r be the effectlve axial stress In unlaxial tension or compression which results in the given octahedral shear stress.
_
f

A cylindrical stiffened thin sheet fuselage is fabricated from 2024 aluminum alloy sheet which has a tensile yield stress Fty = 40000. Find the yield margin of safety under the following limit load conditions.
(1) A limit bending moment produces a bending stress ot 37000 pSi (tension) at top point at fuselage section. The tlexural shear stress is zero at this point.

= t s oct / 0 . 47

3 =v'2 f Soct

( - - - - - 27)

(2) Same as condition (1) but pressurization at fuselage produces a circumferential

tension stress of 8600 psi and a longitudinal tension stress ot 4300 psi.

3/~ we obtain for a condition of principal

Therefore multiplying Eq. (26) by

triaxial stresses,

(3)

- 1 I a a a r=/2 v(t,-f,) +(t,-t,) +(f,-t.) - - - -(28)

Same as condition (2) but a yaWing maneuver of airplane prOduces a limit torsional shearing stress of 8000 pSi in fuselage sk~n.
Condition (1)

Let F equal the allowable tensile or compressive stress. If the yield strength is being dete~ined then I1a.rgin of safety M.S.=

SOLUTION:

This is a uniaxial stress condition for point being considered.

Fty 40000 Yield M.S. = r;;--1= 37000 - 1

-=- t

-(29)

.08

For a biaxial stress system taking f:s = OJ we obtain,

f = v/ f' 1. + f' II
-

SOLUTION:

Condition (2)

t II .1,],

- - - - - - -(30)

It is often more convenient to use the x J y and z component of stresses instead of the prinCipal stresses. Fig. Cl.23 illustrates the various component stresses.
f,

There are no flexural shear stresses at the fuselage point being considered. Since no torsion 1s being applied to fuselage no torSional shear stresses exist. The stress system at the paint be1r~ considered 1s thus a biaxial stress system and flo and f 21 are principal stresses. f1. 37000 + 4300 : 41300 ps i

is

t II = 8600 pat
~-+-_ix
is xy

From equation (30),

Fig. Cl. 23

J 41300

21

8600a

41300 x 8600

Cl.IO

COMBINED STRESSES.

THEORY OF YIELD AND ULTIMATE FAIT..URE.

whence

= 37700 psi

The vessel wallIs to be stressed to the yield stress at 42000, thus t = 42000.
vhence

SOLUTION:

Condition (3)
f

(42000)

Since a torsional shear stress has now been added, the new stresS is still two dimentional, however the given tension stresses are nat principal stresses due to the addition at the torsional shear stress.
!x=41300 psI.

tx-ottx
,z
! ......... --f s_=8000 f

SoIvtng , p
PROBLEMS

= 970

psi.

f~

(1)

-xa

=:8000

The combined stress loading at a point in a structure is as to11ows:- f z = -1000, t x -2500, t s 2000. Determine the magnitUde and direction ot the principal stresses. Determine the maximum shearing stress. Solve both analytically and graphically.

!z=8600 ps f ,

!s=8000 ps i ,

tinding the principal stresses and USing Eq. (30), we will use the t x and t z stresses and use Eq. (32)

Instead

at

(2) Same as Problem 1, but change t z to 4000 and t x to -3000 and t s to 2500.
(3)
A solid circular shatt is subjected to a limit bending moment at 122000 inch poundS and a torsional moment of 250,000 inch pounds. It di~eter is 4 inches and the yield tensile stress is 42,000, what is yield Margin or Satety. A thin walled cylinder ot diameter 6 inches is subjected to an axial tensile load ot 15,000 pounds, and a torsional moment or 12,000 inch pounds , l..lhat should be the wall thickness it the permissible yield stress is 30,000 psi. A closed end cylindrical vessel is 15 inches in diameter and a wall thickness ot 0.25 inches. The vessel 1S SUbjected to an internal pressure ot 10,000 pSi, and a tensile load at 22,000 pounds. If the yield tensile stress or the material 1s 75,000 p 1, what torsional moment can be added without caUSing yield.

="
?

41300 1I+ 8600- 41300 x 8600 + 3 x 8COO

pst

= 40200

x.a,

=~ - 1 =-.01 40200

(4)

Thus yield is indicated since M.S. Is

negative.
Example Problem 2.

A cylindrical pressure vessel is 100 Inches in diameter and 1 inch thick. The vessel Is made ot steel with Fty ; 42000 pSi. Determine the internal pressure that will produce yielding. SOLUTION: This applied stress system 15 biaxial with no rteccuraj, or torsional shear.
Let:

(5)

p equal internal pressure t = wall thickness = 1 in. d = diameter = 100w stress due to pressure p
and

t.

=ct reuererenetar

-Q-t,
I

t,

REFERENCES: -

Nadat , Theory or r'tcw and Fracture or Solids. Ttmoshenko, Strength ot Materials. FreUdenthal, The Inelastic BehaVior at Engineering Materials and Structures Marin, J., Engineering Materials.

t =pd 2t

t :I.-it -pd

From Eq. 30

r=/t:+t:-

Seely & Smith, Advanced Mechanics of Materials.

CHAPTER C2

STRENGTH OF COLUMNS WITH STABLE CROSSSECTIONS

C2. 1 Methods of Column Fatlure.

Column Equations.

In Chapter Ale, the theory ot the elastic and inelastic instability of the column was presented. The equations tram Chapter Ala tor a pin end support condition are:For elastic primary failure, Fe =

The range AS in Fig. C2.1 Is tor a range or Lip values of below 20 to 25, and represents a range where failure Is due to plastic crusntng of the column. In other words, the

column Is too short to buckle or bow under

end load but crushes under the high stresses. This column range or stresses 1s usually referred to as the block compression strength.

--"--=-(Lip) ,

nil::;

- - - - - - - - - (1)

A column, however, may fail by local

bUCkling or crippling due to distortion of the column cross-section in its own plane. The horizontal dashed line in Fig. C2.1 represents the condition where the prtmary column strength is limited by the local weakness. This line moves up or down aCCOrding to the value or the local weakness. The deten:nination ot the column strength when failure 1s due to local weakness is covered in another chapter.
C2.2 Column End Restraint. Fixity Coefficients. Column Effective Length.

For inelastiC primary failure,

Fe

- - - - - - - - - - - - (2)
st;~ss

Where Fe = compressive unit failure = PIA stress.

E

at

= Young's modulus Et = tangent modulus L = column length p = radius or gyration

section

of cross-

Fig. C2.1 shows a typical plot or Fe versus Lip. It the column dimensions are such as to cause it to fail in range CD in Fig. C2.1, the prtmary failure Is due to elastic instability and equation (1) holds. This range of Lip values is ofteL referred to by engineers as the long col~~ range.
A

The column strength is influenced by the end support restraint against rotation and by any lateral supports between the column ends. The letter C is commonly used to indicate the end fixity coeffiCient, and C = 1.0 for zero end restraint against rotation, which can be prOduced mechanically by a pin or ball and socket end support !1ttlng. Thus including the end restraint effect equations (1) and (2) can be written, FC = , Fe =
(Lti> )"
CTt.lEt

c I
Local Crippling Limit

- - - - - (3)

(Lip)"

Fe

Let L' = effective length of the column which equals the length between inflection points or the deflected column under load. Then L' =

Live
Fe -

(4 )

Thus equation (3) can be written as,

L'/~

Fig. C2.1

Fe

- - -

(5)

(L'/p)'

(L'/p)'

The range Be represents the range of L~ values where failure 1s due to inelastic instability of the column as a whole and equation (2) applies. This range Be 1s often re~erred to as the short column range.
C2.1

If we let P = tailing or critical load, equation (5) can be written as equatton (6) by realizing that p FeA and p ~.

naSI
p

naEtI
p

=(L' )'

=(L' )'

- - - - - -

(6)

C2 2

STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS TABLE C2.1 17-7 PH(THI050) Stainless Steel Sheet Ftu "" 180, 000 E:: 29 X 105 Fey:: 160,000

Fig. C2.2 shows the deflected colu~~ curve under the load P for various end and lateral support conditions. The effective lengths L1 and the end f~xlty coefficients are also listed.
20,000
30,000 40,000 50,000 60, 000 70,000 80,000 89, 600 96,000 102,000 107,400 112,200 117,000 121, 500 125, 800 130,000 134, 000 137,700 147,000 158,100 167, 600 173, aoo

Fig. C2.2
C2.3 Design Column Curves for Various Materials.

For routine design purposes it 1s convenient to have column curves ot allowable tailing column stress Fe versus the effective slenderness ratio Il/p. In equation (5) we will assume values ot Fe' then find the tangent modulus Et corresponding to this stress and then solve for the term L' /p. Tabla C2.1 shows the calculations for 17.7 PH (TRIOSO) stainless steel sheet at roam temperature. The reSUlts are then plotted in F1g. C2.7 to give the column strength curve. Similar data was calCUlated tor the material under certain exposure time to different elevated temperatures and the results are also plotted in Fig. C2.7. Figs. C2.3 to C2.l5 gi ve column curves tor other materials under various temperature conditions. Use ot these curves will be made in example problems later in this chapter. The horizontal dashed line is the compressive yIeld stress. Values above these cut-ott lines should be substantiated by tests.
C2.4 Tangent Modulus Et from Ramberg-Osgood Equation.

29 x 10 5 29 x 10 5 29 xlO 5 29 x 10 " 29 x 10 e 29 xl0'" 29 x 10 " 29 x 10 os 27.55xlO" 26. 10 x 10 " 24.65 x 10 5 23.20 x 10 e 21. 75 x 10 " 20.30 x 10 e 18.85 x 10 e 17.40x10" 15.95xl0 5 14.50 x 10 e 11.60 x 10 " 8.70 x 10 e 5.80 x 10 " 2.90x10"

119. 56 97.62 84.54 75.62

69.03
63.91 59.78 56.49 53. 19 50.22 47.57 45.15 42.81 40.58 38.43 36.32 34.25

32.22
27.89 23.29 18.47 12.82

This equation is plotted in Fig. C2.16. For a given material, n,.F o . ? and S ~ust be known. Then assuming values of F, we can find corresponding values of Et/E from Fig. C2.16. For values of Z, Fa.? and n refer to Table Bl.1 in Chapter Bl.
C2.5 Non-Dimensional Con... .m Curves.

QUite useful non-dimentional column curves have been derived by Cozzone and Melcon (See Ret. 3).
The E:uler column equation is F = n at / (L1 / p ) a , which can be written,

(L'/p)'
F

n'

tor

Et

The baste Ramberg-Osgood relationship

is given as tallows:
1

(See Ref. 1)
- - - - - (7)

The problem there~ore resolves ~tself into obtaining and expreSSion fOT Et/F from the non-di~ensional relationship. To do this multioly both sides of equation (~) by Fa ?/F and equate to B a.

(;~~ ~. 9

=-;:;---;;---;:;----:= B' -L.- + ~ n (-E....-)n

Fa.? 7 Fa.?

- -

(8)

Et = tangent

modulus ot elastiCity

E = modulus ot elastiCity
For definition ot other terms see Article B1.12 ot Chapter 51.

Fig. C2.l7 shows a pla~ of this equation as taken trom Ref. 3, and shows FIFo.? versus B for various values ot n. The shape of the knee of the stressstrain curve is given by the Shape parameter n and the abscissa B incorporates the

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

C2.3

120
R. T.

RoT.
Flg. C.4

Fig. C1 .1 _Sl<!eI, AlSt 4130, H40; 43olO t"'" O. las Temp. E:::qlo....re up to 11: IIR. -F ...QOOO<HRo~. 1---.

--~ 110 S!JOGF'

roe

.!lee1 AJS[ 4130,' ~l4O. 4340 -1t.U-Tl'eamlnFiia-~ t:zsOOO'"

T_lIIp. Blip'-up
-- .-------,

to 1/2

II&.

-"!'":':""..'1~---

."

P';7 .. -,oo (It. T:)

~ ~-_.--~-'

-.-"-", ..""-,-,,,--,-

- I 0I)lI0'P -:-

l()ll(l9F

"
10
-~--,--~-----~._------------.-

10

"~20

30

40

50

80

70

80

90

100

L'IP

Figure C2. 4

Figure C2. 3

180

"---=

---R;~ -.-:----:- - - . -

I~O-----R;-'l';

rie, C%.~ Steel Als:l 4130. 4140. 4340 --.--- H:iU'1'.i'nted ToFi;u;- "-15'OO
1:10'-

.500S'E,-__ c__ '

-'-'-.!-~;?i4~~'-~-

130":"--~

120--'----

I:O~---'---~ -~~""85Q01f

110 -------.--100 ,_ -.C._-._ _ _ _ '-_._"

liD .: ....-.-.-;~

aecsn
so

50'

--_._~_.

20

30

80

90

100

10-20

30

40'- ~o-- eo
t'! P

r:/p

Figure C2. 5

Figure C2. 5

:, ,-f " "",,,~'--:>K~

C2.4

STRENGTH OF COLUM!'iS

WITH STABLE CROSS-SECTIONS

1B0;.,.....,.....~-

RoT.

._~,---

170-----

'"
Fig. C2. l'

Ro-T.--'
F~; C2.8 11......Pll stainl.....a steel (Bars'" Forg1D.gsl .. -Temp.' E:qI. ~p to 1/2 fiR.

11_1 P!f(TH1050\5ta1ll.IeS& 9:.... 1

.~

'so

Strip&! P!:Ite"
-

150 ---140

L 010 _ 0; 125 In. :r"mp., Exp', up to 1/2: HR.

F
tu

"iaooco

.-

'"
'"

Pm'" ISOOOO
Fey-'" 1ll5QOO

_Fer -. 182000

130 '"

'"
110 ,. 100 .-

~
80!--'---

,"

sc

"
70

''::~~'''':',.i-.:_~

:;,

40 ;....:.:

. !-

"...---.:'-',---o,,-,.,..
...

~,----,----.-

...,.........

.
eo
90

".- ......

20

:~_~:::~~;"""--'----:~-_.----,.--.-'-,.~.~i-';'- ~---,. "-~'';;;::'''~''-i'--''~'-i''''''-'' '.

10203040

..- '.---' 50- 50

L>/f'

;--,-,-,,--~o~

ec 'go" 'ioo

Figure C2. 7

" ., L'/f'

'"
R. T;"

Figure C2. 8

---~---

Bue,~ .,Pllle,

Flg. C2.10 .. ---01U'r!i;,_TlJ A!wn; Alloy t ~ O.W Ia.. _._ P'tq-- 16000.

Fey-'" 6'lllQO;
'-----:---:-T~

Ezp;'.-lBB:

'c'

;,-,...

:.~:_:. ~-"':_:::.....c.:i._

50

...:..-i------'-.c.:::;..:::.':.:'::.:.:,:....:,-,.c..:.:.:.,-;-.:..,_
,~

t>. :,:;, .' ""

.

. ", ... "

".' .. __

' . " "

,., " . . ."",.

._..,._.:,._ .

"'-

.,

._-----"----'----,-.._---"--.- .,. ,".-,'

,

. . . . . .,

"-'--,

~--'i'7---.--.-----'~----.,..;-,--'

'._~-

---10
~

30

40

50

70

80

90

100

L'IP

ac - - - - - - - _ . -_._---'--

Figure C2. 10

10--rD""30~--~60--70

ae

90

100

Vip

Figure C2. 9

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C2.5

,
R. T.
-

eo

'00'"

f':Ilt. C2,l1 70'l5Te Alum. Alloy BJltl''l8iOll. t ~ O.2Ull. Temp. EJql. Z HR. Fro'" 75000_ FC'U .. 7oo

..
eo

FIi", C'.12 __ 'lO'r.!.Te_Alum. Allay Clad ~ I:! a.50!D. 1'e~. EXp a 1m: Fta--'lOOOQ--------- -

r<:i.. 64QOO-

30

20'"
20

-'-""45Q:offi ilal,:'-""-
-.--.--

-----_ .._--_._."

iu ........ =:-:::::;:::~

ro

L'IP
10 W
~ ~

50

ro

-~6

100

L'IP

Figure C2. 12

Figure C2. 11

Fl(; C%.IS 'l0'r5-:t'll Ahzm. All>:lT {Ole Fcc;mgat . "-'-... .-.TflIIP---'-.'Z BR.-__,)'__,"_71O!lO ....,. __.._
---r-~Z'!li:"----~-'

"<-40:_

--- Fcr--saooo

30

\0

20

ro

ro

SO

L'II"

Figure C2. 13

C2.6

STRENGTH OF COLUMNS WITH STABLE

CROSS-SECTION

-7l119_~A1W:IIo-

Fig. C2.1. A1lo7'iI:ul<t Ftinf I .:t&.O lIL.Temp. exp 1/2 HR. ------"Iii &"67000

130-

R. T,
Fig.

ca. 15
T1tan1um Alloy
Annealed

__ ":~1 '::"~

n,

'"

Tl-&AJ~4V

au '" Sheet
Ft:T~

t ~ O. 187

In.

TtmlVc Exp;; sl!2lJR:. F hl13oooo lUOOO

'"' "

40

--- --------

-------

--------

20 -------- ----------- .----.---

'"
L'IP

---_._-------~~--'--------

~3040501lO108090

,-.-,-"20--30---40----50-60-70-80-L'!P

90

Figure C2. 14

Figure C2. 15

-:::;1. 00

..; .80

s
~

z o

i'-.

-c
1'-

\'r .90

r-, <,
I'--..
<,

l--.'

t--.

-,...

t---- r-,
l--.

f----

t--. <, ~" .:'-(

"" 0 I"::i-,

r-,
.,0

, I"\n- 0

1. 00

\ .'0
I

1'-

1"<,

I
I

.e 0
E,
E .5 0

r-, 7':> l\ l\ \ 't-. \ ~\ l\ ~\\ 1'--- r-, .o-f'- ~ k 3 R: ,\ ,\ : I 1'.5p:;; 1:::-\- i\\ ~\\ I ! t"'" ~ \'\:
-r-,

h"

1".\.9

..n"-I
,

i
1
I,
,

1
1 i ,

I
I

1
"

.90

I , I I
,
, , ,
"

i
1 1 1

,
.80

, ,

, ,

, ,

, ,
I
1
1
,

I
I
1
, ,

, I ,

, ,

!"":ile.\\

i ,

.4 0

, , I
1

1
I
I

,
!
I

W!

, ,

!~ I'-

t-t--..

r--:
,

I , 1
,

I I I I I i
I
, , , , ,

.70

.0

I
i

n ~ 2.

.I .3 0
.a 0 ,
I

i E,

I
+7

i
1

I
, ,
,

I" ,Ii:'\'-- ",N

I~
1

~I
,

I-----:....J.

,
,

1'--.'

i 'i---J.
I N
.1
,

i-----. '-1
,

9 I
I

.50

I
1

.40

--j
.30

E =1 .! n (..,..L..tn- ~
F o ...

I i
, I
1

[:\ - ,

,"-. 1

'i-..... 1

~:
1
,

I
. 10

I I

,
i

1\\ ~. <, ]"'--...,

Fig. C'. I. Dimensi.onless tangent modulus stress curves. I

!
i

I :
F/F" ...

~ ~""'-. 'i----: I, I~ , ~~ 1--..---"""" -s~ I , ,0

.90

'C

1
,

.10.20.30.40

.so

I
.60 .70

I
.80

\\ 'R~~~ ,---"l -!~ ~Il ,<f ~ 1.001.10 1.201.30 1.40 1.50

(1::

Fig. C2.16 (Ref. NACA T. N.902)

sis].

1 .0.i i
0.9
0 .
0.7 Fe
-F- 0.8 ,"

,. ,. r''" " '" ,.

Z Z

tl tl
l'l

\d'pl~

ilt
I , .

~IJr
I'.
;" -,.

'" c
Z
0

>oj

0.7 0.5

.'"
o
l'l
i

'" '" r-

.70
i

'" r-

i ,

l'l

l'!
0.1

>oj '"

.'. !
0.2

" c:
l 'l

'c":
>oj

0.4
Fig. C2.17

3.2

3.4

'"

'"

Ee

il

II

C2.8

STRENGTH OF COLUMNS WITH STABLE CROSS SECTIONS

~ropertles

particular

of the material EO.7/F.

Tt~t/(L'

Inser-t Ing va Iue of F equation (3), Et

/p)a in

Thus in Fig. C2.l8, ~he deflection of paint (0) away from tangent at nidpoi~t c equals unity in our assumed conditions and i~ also equals the first ~o~ent of :he area of the M/E! diagra~ between (0) and (C) about (0).
(Fig. C2.I9).

n"

Et/( L '/.0 ) a)

~ ( , , . ,'\ B'

-E-J=
\ P

or B

=.!c
n

j ,',., ()
E

The value of the ordinate for M/EI at any point x rr-cn 0 is :1 sin 1"'! LX
----(9)

diagr~n

The total area under area = ,I .!

P .L

~he

curve is,

The use or the curves in Fig. C2.17 will be illustrated later 1n the example problem sotut i ons .
C2.5 Strength of Columns With Variable Cross-Section or Moment of Inertia.

sin
"

rXdx II [-

cos

= EI

-I[~

(- IIJ[- *(IIJl ir
"
~alf

(*+~)
ru"

To save weight 1n a built up column or forged column, the member 1s tapered or Is made ~lth a non-uniform cross-section. To find the ulttmate strength or such columns,

_2PL 'o_PL hence area tEL and haL ar- "a - --:::;-;-

The center of gravity or the

area is

it Is usually necessary to use a trial and

error method. The general method of solution involving a consideration of column deflection Nill be illustrated for a case of a long column with unlfor.n cross-section. Fig. C2.18 shows a pin ended column in a deflected neutral equilibrium position when carrj1ng the ultimate or critical load P. Assume that the shape of the deflected column follows a sine curve relationship with the deflection at midpoint eqU?~l to unity (see
Fig. C2.IBI.

AX=/xda
PL(~Ol) x
"oW

= ':'.1. ~. /
L/2

oL/2

x sin

rt L x ex ,

L -x =/ n

x sin

LX "

dx

Integrating this Simple expression and solving for x we obta~n:

x

The

curve is y =\T . ; It ? is the end load, the ~-.--I bending moment at any point = M = Py = P '1 sin az,
" L

-CfL"'n x '\

~t'Qn

of the deflected column

= Lin

Taking moments about point (0) of the M/E1 diagram between 0 and C about 0: hence,

L'
which is the Euler equation, and ~hus the assumed sine curve was :he proper one ~ur deflected elastic curve of the column.

By the well known "mament area" principle (see Chapter A7; Art. A7.l4), the deflection ot a point (A) on the elastic curve away from a tangent to elastic curle (8) equals the first moment of th~ M/EI diagram between (A) and (B) about (A).

th~

L
P

' r
p

Suppose that the elastic curve of deflected column had been assumed as a parabola with unit deflection at ancpomt . Ff g . C2.20 shows the M/EI diag~~. The area or one-half the
diagra~

SIll.

C.UItVe:

;
I I

[-~'I
C
~

2' ('3

_ PL

LX,I) - 3EI

L.

I-

'~

Taking moments about 0 ot the area between 0 and C;

Fig. C2.18 Fig. C2.19

'U

-I=..!:.. Y o 3EI

16 L = 48EI

5PL"

bence ,

ANALYSIS AND DESIGN OF

FLIGHT

VEHICLE STRUCTURES

C2.9

P -=

9.6 EI 48EI 5L a = L'

which co~pares with P = 9. 9EI/L of the Euler equation or an error of 3 percent.

(9)

Solve for load P by writing and expression tor the cerj.ec t t on at the center point which equals unity. This is done by using the moment area principle as was done in the previous example problem involving a column with uniform section.

Fig. C2.20

In the above outlined procedure, E has been assumed constant or, in other words, the column failure is elastic or failing stresses are below the proportia~.al limit stress of the material. The practical problem usually involves a slenderness ratio where failure is due to inelastic bending and thus E Is not constant. For this case, a trial and error method of solution in necessary using the tangent modulus of elasticity which varies with stress in the inelastic stress range.
C2.6 Design Column CUrves for Columns with NonUniform Cross-Section.

We can now apply the same procedure to a colunn with non-unifonn cross-section. The steps in this procedure for a column symmetr'ical about the center point are as follows:(1)

Assume a sine curve tor the deflected column with unit deflection a center paint. Plot a moment of inertia (I) curve for column cross-section. Find the bending moment curve due to end load P times the lateral deflection. Divide these moment 'falues by the EI values to obtain MlEI curve. The modulus of elasticity E is considered constant. Find the deflected column curve due to this M/EI loading. Campare the shape of the derived column deflection curve with that originally assumed as a sine curve. This can be done by multiplying the computed deflections by a factor that makes the center deflection equal to unity. Since the assumed sine curve is not the true column deflection curve, the computed deflection will di~fer somewhat fram the sine curve. With the computed deflection curve, modified to give unity at center point, repeat steps 3, 4, 5 and 6. The results this time will show derived deflection curve still closer to the assumed deflection curve. To obtain th~ desired accuracy, the procedure in step (7) will usually have to be repeated again.

Figs. C2.2l and C2.22 give curves for rapid solutIon of two types of stepped columns. Figs. C2.23 and C2.24 gives curves tor the rapid solution at two forma of tapered columns. Use at these curves will be illustrated later in this chapter.
C2. 7 Column Fixity Coefficients c for Use with Columna with Elastic Side Restraints and Known End Bending Restraint.

(2)

(3)
(4)

Figs. C2.25 and C2.26 give curves for finding fiXity coefficient c for columns with one and two elastic lateral restraints and Fig. C2.27 gives curves for finding c when restraining moments at column ends are known. Use at these various curves wili be illustrated later.
C2.8 Selection of Mater1al8 for Elevated Temperature Conditions.

(5) (6)

(7)

Light weight is an Unportant requirement in aerospace structural design. For columns that fail in the inelastic range of stresses, a comparison of the FCY/W ratio or materials gives a fairly good picture of the efficiency of compreSSion members when SUbjected to elevated temperature conditions. In this ratio Fey is the yield stress at the particular temperature and w is the weight per CU. inch of the material. Fig. C2.28 shows a plot of Fc /w for temperature ranges up to 600 0 F. With 1/2 hour time exposure tor several important aerospace materials.
C2. 9 Example Problems.

PROBLEM 1.

(8)

Fig. C2.29 shows a forged (I) section member 30 inches long, which is to be used as

i
I
CRITICAL W"lllH<lOtHJNlfORM COLUWNll

CIUTICAL WAOI:I-N()NUHI~R"COLUMNS

Fig. C2.21

bJubl.lit!lPl'!'1 . PIA IlDded

'." .(j'
.. I'

.....

Ca"

f:-.
i.

I' I H -.I.-',---I.-.-j
:11'1'".";;1;;':1"1
, '"'
i "

fUl,

(1:1:1,

"
, , I:il :ri " I n , ! III I, I
, ,

I ,

~~llJ

" :;

~ _.

l'

..j .'

I"

, .' I~ t'

,i
't

>

..
..
.T

'-1-,

,I
COI"tonl
T~,c1",e

'j:

!iii
, .. II.

,I'!

, .,.I , ,
,

, , ,

"'>,j....
CIUTICAL 1.o"OO-NON-lItllf'ORM
~

.. . .. .. ..

'":J ..

". ;;1

CRITICAL LOADS-NON-UNIFORM COLUMNli Solid Priam. Tap"rloK in W\4UI """ ~

- "".pu 1" PloQ-rorn,

Solid eyliliden Taperlll( to Coo...

..... _

a(~)!

Fig. C2.24

... '" >III

r-

"'
I

~
i..

I I
I

'

..

.1

'" '" '" o "' ...

z

'o"

r.lIo

'"

I
Fig. C2.23

'I '

. . ..
Reference - N. American Avaiation Structures Manual

'.1

I ,
(QlJlnla

.. 1: .. ..
!
I

II

II

.. ...
!

SIMPLY SUPPORTED COLUMNS WITH ONE ELASTIC RESTRAINT

4.0.

3.8
3.6 :

140 120

3.. 3.2

K'

'"

4.0

3.0 2.8 2.6

C 2.4

3.8 "

100
80
~....,

Iq

2.2
2,0

' 80

1.8 1.8 1.'

1.2 . . . . .

40
20

!
0.1 0.2

1.0

Fig. C2.25

x/L

O.S

0.6
C 2. 8 1

2.8 '

2)
8

2.2,ill
I:
i'

, I'
I

i'if
i

,.J
""j ,

"' '"
o
Z
~

,. " e
Z

o ..,

I ';,l,'

7
6

2.0 .' :

1.8 1.6

2;

I'

i
1I

I
0.1 0.2

i
1.'

1.0

,
0.8 0.9 1.0

0.3

0.4

0.5

0.6 0.7

10

20

Fig. C2.26

xlL

Ref. Convair Structure Manu.al

.l.!o
EI

Fig. C2.2'l

(j

C2.12

STRENGTH OF COLUMNS WITH STABLE

Fig. C2.28

CROSS~SECTIONS

Solution: Since the column may fail by bending about either the X or Y axes, the col~~ strength for bending about each of these axes Nill be cal~u lated. Since the column strength is a fUnction of the radius of gyration of the cross-sectien, the first step in the solution will be the calcujat.Lon of Ix and I y, from which o x and Py can be found. Calculating Ix: In Fig. C2.30 the section will be first considered a solid rectangle 205 x 2.75 and then the properties of portions (1) a.~d (2) will be subtracted.

(1) AISI Steel. F tu= 180, 000 (2) 17-7PH. stainless steel, F tu = 2l 0, 000 (3) 7075-76 Alum. Alloy
(4) AZ31 B Magnesium Alloy

(5) 6AL-4V Titanium Alloy

Ix (rectangle) ::: l~ x 205 X 2075 3

Portions (1) and (2)
----.;....;.:.........;.-_._._-. ._-

:::

4.32

Ix

:I

-l~ x 1.5 x 1.25 3

4( .625 x .25 x 1.292>l) :::

-1.29
(1 0 or (2) about its x centraidal axis is negl1gi ble )

Ix = 4032 - 1.29 ::: 3.03 -in ;"

100 200 300 400 500 600
TEMP. of

Px

=15. , Area A ::: 205x2.75-2x 075xl.25A

a compression member. Find the ultimate strength of the member i t made fram the following materials and subjected to the given temperature and time conditions. Case 1. Case 2. Case 3. Case 4.
~J1terial

4x 025 x .625) = 4.375 sq. in.

Px =

V3.03/4.375
1 = 12 x

= .83 in.

Calculation at I y: I y (solid) 2.75 x 2.5'

7079-T6 Alum. Alloy hand forging and roam temperature.

= 3.58

Same as Case 1, but Subjected 1/2 hour to a temperature of 3000F. same as Case 2, but tor 600oF. Material 17-4 PH stainless steel, hand forging at room temperature.
Fig. C2.29

Portion (1) = -(1.25 x .75x .875')2 -(1.25 x .75'/12)2 = - 1.52 Portion (2) = -(.25x .625x .833")4-4(.25x 1.25'/36) = - .488
Iy
Py

= 3.58

- 1.52 - .488
.60'

= 1.58

in.'

= V 1.5814.375 =

,Y

Column strength is conSiderably influenced by the end restraint conditions. For failure by bending about the x-x axis, the end restraint against rotation is zero as the single fitting bolt has an axis parallel to the x-x axis and thus c the fiXity coefficient is 1. Forfailure by bending about the y-y axis we have end restraint which will depend on the rigidity at the bolt and the adjacent fittir~ and structure. For this example problem, this restraint will be such as to make the end fixity coefficient c = loS.

Fig. C2.30

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C2.13

?or failure about x-x axiS,

L' : L/VC: 30/VT: 30, L'/Px: 30/.83 : 36

Thus we make use ot the curves in Fig.

C2.17.

For failure about y-y axis,

L' : 30N'"T.5: 24.6, L' /p y : 24.6/ .60

: 41

Case 1. Material 7079-T6 Alum. Alloy forging. Table 81.1 of Chapter 81 s~rlzes certain material properties. The properties needed to use Fig. C2.l7 are the shape factor n, the moduls Ec and ~he stress FQ.~o Referring to Table Bl.l, we find that n = 26, Ec : 10,500,000 and Fo.~ = 59,500. The horizontal scale in Fig. C2.17 involves the parameter,
(L' /p)
Subs t t tut tng i-.

Therefore failure is critical for bending about y-y axiS, with trip = 41. Case 1. The material is 7079-T6 Alum. Alloy hand forging. Fig. C2.l4 gives the failing stress Fe for this material plotted against the t'/.o ratio. Thus uefng L r /.0 = 41 and the room temperature curve, we read Fe = 50500 psi. Thus the falling load if P = FcA : 50500 x 4.375 = 220,000 lbs. Case 2. USing the 300 0F curve in Fig. C2.14 for the same L'/.o value, we read Fc = 40,400, and thus P 40,400 x 4.375 177,000.

Case 3.

USing

~he

and thus P : 6100 x 4.375 = 26700 lbs.

600 0r curve, Fc reads 6100

Using Fig. C2.17 with 1.01 on bottom scale and projecting vertically upward to n 26 curve and then horizontal to scale at left side of chart we read FcIFo.'l'_ = .842.

Thus

subjecting this member to a temperature of 600 0F for 1/2 hour reduces its strength tram 220,000 to 26,700 Ibs., which means that Alum. Alloy is a poor material for carrying loads under such temperatures since the reduction in strength is quite large. :ase 4. ~~terial 17-4 PH stainless steel forging. Fig. C2.8 gives the column curves for this material. For L' /.0 41 and using the roam temperature curve we read Fc = 135,200 and thus P = 135,200 x 4.375 = 591,000 rbs

Then Fc : 59,500 x .842 = 50,100, as compared to 50,500 in the previous solution

USing Fig. C2.14.

Case 2. From Table B1.1 for this material subjected to a temperature of 60QoF for 1/2 hour, we tind n = 29, Fc 9,400,000 and Fo.'T 46,500.

Then B :

~ j 9,~O:goOO

(41) :

.917

C2.10 Solution Without Using Column Curves.

From Fig. C2.17 tor B = .917 and n = 29, we read FcIFoo'T = .88, thus Fc = 46,500 x .88 = 40,900 as compared to 40,400 in the previous solution.

vnen primary bending rat iure occurs at stresses above the proportional limit stress, the failing stress is given by equation (5) which is,
Fe = n'Et/(L'/p)' Since Et 15 the tangent modulus of elastiCity, it varies with Fc ' and thus the relation at Et to Fc ~ust be known betore the equation can be solved. To plot column curves for all materials in their ~y manufactured forms plus the various temperature conditions would require several hundred individual column charts. The use or such curves can be avoided if we know several values or parameters regarding the material as presented by Ramsburg and Osgood and expanded by Cozzone and Melcon (see Arts. C2.4 and C2.5) for use in colunh~ design.

EXAMPLE PROBLEM 2.

Fig. C2.3l shows an extrUded (I) section. A member composed of this section is 32 inches long. The member is braced laterally in the I_lt~l x direction, thus failure will occur by bending about x-x axis. 1'2 x _ The ~ember 1s ?in ended - - -1t , and thus c = 1. The , 1 material is 7075-T6 , Extrusion. The problem is to tind the failing " stress Fc under room Figo C2.31 temperature conditions.

it

-1--

This (I) section correspondS to Section 15 in Table A3.15 in Chapter A3. Reference

C2.14

STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS

to this table gives, A = .594 sq. in.

L'

Px = .618
L'/px

L, stn c

= 1,

= 32/.618 = 51.7

is due to the fact that the stress existing under a L'/p value of 51.7 ls near the proportional limit stress or Et is not much dlfferent than Ec the elastic modulus. To illustrate a situatlon where the 7075 material becomes more efficlent in co~parlson to the 2014 alloy, let us assume that our ~ember has a rigid connection at its end whict wlll develop an end restraint equivalent to a fiXity coefficient c 2.

Fig. C2.1l gives the column curves for this material. For L'/p = 51.7 and room temperature we read Fc = 38,500 pSi. Solution by using Fig. C2.17,
n == 16.6,

From Table Bl.l tor this material we tind Be :. 10,500,000 and F 7 = 72,000.
Q

Then L' = 3ziV2""= 22.6 and L' /p = 22.6/.618 36.7.

Then, B

=~ / 10,~~~~go

(51.7)

1.36

For the 7075 material fram Fig. C2.1l, Fc = 58,300 For the 2014 materlal, we use Fig. C2.l7
_ 1

= 1.36 and n = 16.6, we read FC/F . 7 = .537, hence Fc :: .537 x 72,000 = 38,600.
From Fig. C2.17 tor 8

Consider the member is subjected to a temperature at 4500F for 1/2 hour.

8 -

10,700,000 (36.7) - .823.

53,000

From Fig4 C2.11, Fc

Us1ng F1g. 02.17:-

=21400

pSi

From Fig. C2.17 tor B = .823 and n = 18.5, we read Fc /F 0 . 7 = .87, when Fc = .87 x 53,000 = 46,100 as compared to 58,300 for the 7075 material, thus 7075 materlal would permit lighter weight of required structural materlal.
The stude~t should realize that if the stress range Is such as to make Et Ec' then the bending failure is elastic lnstead of inelastic and equation (5), USing Young's modulus of elasticlty Ec' can be solved directly Without resort to column curves or a consideration of Et, since Et is equal to Ec

From Table 81.1, n :: 8.8, and Fe/F 0 . 7 = 29,000.

B

Ec = 7,800,000

=1. j
n

7,800,000

2!l,000

(51.7)

= 1.00
7

From Fig. C2.17 we find FC/F 0

.74

Then Fc :: .74 x 29,000 :: 21,450 psi A very cammon aluminum alloy ln aircraft construction is 2014-T6 extrusions. Let it be required to determine the allowable stress Fe for our member when ~de ot this material. Since we have not presented column curves for this material, we will use Flg. C2.l7. From Table Bl.1, tor our material, we find n ~ 18.5, Ec = 10,700,000 and Fa 7 = 53,000 .
Th n B

The student should reallze tr~t equation (5) ls for strength under primary column failure due to bending as a whole and not due to local bUCkling or crippling of the member or by twisting tailure4 The SUbjec~ of column design when local failure is involved is covered in a later chapter. In example problem 2, we have assumed that local cripplir.g ls not critical, which calculation will show is true as explained and covered in a later chapter.
C2.11 strength of Stepped Column.

=.1 j n

53,000 10,700,000 (51. 7) = 1.16"

From Fig. Cl.17 tor B 1.16 and n we read Fc/F o 7 = .71, hence Fc = .71 x 53,000 37,600.

= 18.5,

The result shows that the 2014-T6 materlal gave a failing stress of 37,600 as compared to 38,900 tor the 7075-T6 material which has a Fcy of 70,000 as compared to Fcy = 53,000 for the 2014-T6 material. The reaSon for the 7075 material not showlng much higher column failing stress Fc over that for the 2014 alloy

The use of curves in Fig. C2.22 will be illustrated by the solution for the strength of two stepped columns in order to lllustrate both elastic and inelastic tallure at such columns. Case 1. Elastic failure.

Flg. C2.32 shows a double stepped pin ended column. The member is machined from a 1 inch dlameter extruded rod made from

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C2.15

7075-T6 material. The problem is to find the maximum compreSSive load this member will carry.
Portion 2 Portion 1

i ~3/4" Dia.

+1"Dia.

Portion 2

J.

This is a relatively short column so the failing stress should fall in the inelastio range where E is not constant, therefore the solution is a trial and error procedure. we will base our first ~~ess or trial on an averagb Lip value.
p for portion 1 is 0.25 inches
p for portion 2 is 0.1875

Fig. C2.32

Average p = (6 x .25 + 6 x 0.1875)/12

0.22

PORTION 1

PORTION 2
All

Then Lip = 12/0.22

= .7854 in. II .0491 in .... = 10,500,000

= 54.5,

use 55.

= .4418 I. = .0155 Ec = 10,500,000

This is

L' the Euler equation for failure under elastic bending. It the ratio aiL equals 1 or a uniform section, B becomes n~ or 10 as shown in Fig. C2.22. The curves in Fig. C2.22 apply only to elastic failure. Since the member in Fig. C2.32 is rather slender we will ass~e the failure is elastic and then check this assumption.
Ell

From Fig. C2.22) Pcr = B(EI,) .

Fig. C2.11 is a column curve for 7075-T6 Alum. Alloy extruded material. With LIP = 55, we read allowable stress Fc = 33,500 psi. Therefore
P
f
l

= Fc A : 33,500 x 0.7854 = 26,300 lb.

:

33,500

and

til

= 26,3001.4418 = 59,500.

The stress t in portion 2 is above the proportional llmit stress so a plasticity correction must be made in USing the curves in Fig. 02.22. Referring to Table 81.1 in Chapter Bl, we find the following values for 7076-T6 extrusions:- n = 16.6, F o T = 72,000, Eo = 10,500,000. The tangent modulus the stresses f 1 and f ll

aIL = 30/60 El a = 10,500,000 x .0155 = 3.17, = 0.5

10,500,000 x .0491

=3.17 we

From Fig. C2.22 for aiL = 0.5 and EIl/EI. read B 7.0

'...' hence, Per

=7 x 10,500,000 x .0491
60

Et will
?

be found for

For Portion 1, fl/F o .

1000 lb.

= 33,500/72,000

.466

The stresses in each portion are,

t,

= 1000/0.7854 = 1280

psi

Referring to Fig. C2.16 and using 0.465 and n = 16.6, we read Et/E = 1.0, thus Et = E and thus no plastiCity correction for Portion L For Portion 2, f./F o .
?

t. = 1000/.4418 = 2270 psi

These compressive stresses are beloN the proportional limit stress of the material so Ec 15 constant and our solution is correct. Case 2. Inelastic Failure.

=59,500/72,000
= .826

From Fig. C2.l6, we obtain Et/E = .675 whence, ~t = .675 x 10,500,000 = 7,090,000.

EL
El a
8

The column has been shortened to the dimensions as shown in Fig. C2.33. The diameters and material remain the same as in Case 1.
Portion 1
J

10,500,000 x .0491 7,090,000 x .0155

= 4.7 = .5,
we obtain

= 5.6.

From Fig. C2.22 tor aiL

It.

r-""3" -+--- .:6" _b.3'1

L=12" Fig.

Then Pcr

= BEll.
L'

5.6 X 10,500,000 x .0491 = ----'------'---144

------>1

: 20,000 lb.

C2.33

C2.16

STRENGTH OF COLUMNS

WITH STABLE CROSS-SECTIONS

OUr guessed strength was 25,300 lb. Our guessed strength and calculated strength must be the same so we must try again.

1-1/4 -.058

A 1-l/4. 058

B 1-1/4 . 058

<

Trial 2. 23500 lb.

Assume a critical load P

29900
~

t, = 235001.7854

*
16.6, we read

E F I+-- 30" - _......)1;

t.

23500/.4418

53100

Fig. C2.34

Portion 1.
!:tIE = 1.0.

f /F o T
~or

= 29900/72000 = .415
n

From Fig. C2.16 Portion 2.

fa/F o T

= 53100/72000 =

The member AS is welded to th~ee adjacent tubes at joints (A) and (B). Since ~hese tubes are the s~~e at joints (A) anj (a), ~he fixity at the ends (A) and (B) of member AS is the sane. Referring to Ff g , C2.27, the tern ~ is detined as the bending restraint coe!:icien~ spring constant expressed as inch pO~dS ~er radian. In Fig. A, the moment M reE1 (constant) quf red to rotate 8 end (A) through 1 radian when tar end is fixed is 4EI/L. Fig. A For derivation of this value refer to Art. All.4 of Cr~pter All. It the far end (3) is pinned in (Fig. A), a moment M = 3EI/L will rotate end (A) through one radian. To je slightly conservative, we will assume the far ends of members coming into joints (A) and (B) as pinned. Thus ~ = 3EI/L. The s~~ of ~ ~ 3EI/L will be camputed for the 3 members which form the support ot member- AB at end
(A)

.738

From Fig. C2.16, EtlE = .90, whence Et = .90 x 10,500,000 = 9,450,000.

EI, EI a

10,500,000 x .0491
9,450,000

x .0155

= 6.62.

From Fig. C2.22 tor ail = .5, we read

6.62

The Per = BEll.

x 10,500,000 x .0491 =
144

L'

23,650 lb.

This practically checks the assumed value, the answer is between 23,500 and 23,650 and it further accuracy 1s desired another trial should be carried through.
thus

The other types ot columns With nonuniform cross-sections as shown in Figs. C2.21, C2.Z3 and C2.24 are solved in a Similar manner. These charts are to be used only with pin ended columns. The end fiXity coefficient c for tapered columns is not the S~.e as for uniform section columns.
C2.12 Column Strength With Known End Restraining
Moment.

M:ember AC:- I = .03867, IlL = .001289 Member AE:- I .02775, IlL .00071 Member AF:- I .02402, IlL .000962

= =

= =

= Z 3EI/L
=
3( .001289 + .00071 + .000962) 29,000,000

I..l. :::

258,000

.Fig. C2.27 shows curves tor ~lndlng the end fiXity coefficient c for two conditions of known end bending restraint. To illustrate the use or these curres, a simple problem will be solved. Fig. C2.34 shows a 3-bay welded steel tubular truss. The problem 1s to deter.nine the allowable cQmpressive streSS for member AB. This strength is influenced by the fixity eXisting at ends A and E. The diameter and wall thickness of each tube in the truss is shown on the figure. The material is AlSI steel, Ft u : 90,000, Fty = 70,000, E = 29,000,000.

In Fig. C2.27 we need term I..l. L/~I. The L/El refers to ~ember AB. Thus ~ LIEl = (258,000 x 30)/29,000,000 x .0367 = 7.28.
~e use the upper curve in Fig. C2.27 since restraint at both ends of member AB is the same. Ttus tor I..l. LIEl = 7.28, we read end fiXity coefficient c = 2.58.

Then L'

= Live = 30;';

2.58

18.5.

o for member

.422 inches.

L'lp = 18.6/.422 = 44.0

From column curve in F1g. C2.3, we read

ANALYSIS AND DESIGN

OF

FLIGHT VEHICLE STRUCTURES

C2.17

alloNable failing st~ess to be Fe = 55,200 psi. If the far ends of the connecting members were assumed r txed instead of Pinned, then ~ = 4EI/L, or we ca.~ multiply previous value of 7.28 by 4/3, which gives 9.7 whiCh, used in rig. C2.27, gives c = 2.80. L! /p = 30/v'2':'8 x .422 = 42.5. Then from Fig. C2.3, Fe = 56,600 psi. Since the far ends are less than fixed, the assumption that far ends are pinned gives fairly accurate results. In a truss structure all members are carrying axial loads and axial loads effect the ability of ~embers to resist rotation of their ends. Art. A11.l2 of Chapter All explains how to take account of the effect of axial load upon the stiffness of a member as required in calculating the end restraint coefficient ~.
C2. 13 Columns With Elastic Lateral Supports.

C2. 14 Problems.

(1)

6061-T6 Al~lnum Alloy sheet, heat-treated and aged has the following ~r~perties: (a) Under room temperature:- F O 7 = 35,000 pst, Ee = 10,100,000 pst, and n = 31 For 1/2 exposure at 300 oF:- F O 7 = 29,000, Eo : 9,500,000 and n = 26.

(b)

For the above two cases (a) and (b), determine Et (tangent modulus values) n-om Fig. C2.l6 and then calculate and plot column curves for these 2 material conditions. (2) Fig. C2.36 shows the cross-section of a compression member. Calculate the failing compressive load under the following cases:-

,"t

Figs. C2.25 ~~d C2.26 provide curves for finding the end fixity coefficient c to take care of elastic lateral supports at pOints ~idway between the column ends. To illustrate the use of these charts, a round bar 0.5 inches in diameter and 24 inches long is braced laterally as shown in Fig. C2.35. The bar is made of AISI Steel, heat treated to Ftu = 125,000. The spring constant for the lateral support is 775 lbs. per inch. ~ L=24 Moment of Inertia of 1/2 :rod = .003068, Radius of Gyration .125 inches.

Case 1. L = 25 inches. Material AISI Steel 4140, Ft u = 180,000. Take end fiXity coefficient c = 1 for bending about x-x axis and 1.5 about axis y-y.
(3)

I-z-l
)'1

Fig. C2.36

Same as Problem (2) but member is sub_ jected to a temperature of 850F tor 1/2 hour. extruded channel sect tons identical to Section No. 50 in Table A3.1l in Chapter A3, are riveted back to back and used as a column member. If member is 26 inches long and end fixity is C = 1 and material is 7075-T6 extrusion, what is the tailing compressive load. If member is fastened rigidly to adjacent structure which ?rovides a fiXity c 2, what will be the failing load.
~NO

(4)

1
Fig. C2.35

=
120

775 X 24;5 29,000,000 x .003068

?rom Fig. C2.25 for x/L = 10/24 and q 120, we find c 2.92.

(5) .416

~de

Consider the column in proble~ (4) is from 2014-T6 Aluminum Alloy extrusion. Find tailing load.

Then L' : L/ve: 24/1"2."92: 14.08

L' /0 : 14.08/0.125

= 113

Fe

E = (L'n /p)3

:: n x 29,000,000
(113)3

= 22,500
t:>-psi

The pin anced single stepped column as shown in Fig. C2.37 is made Qf AISI-4130 no~allze1 steel, Ftu = 90,000, Fcy = 70,000. Determine the ~tmum compressive load member will carry.
1" Sq.
~r

If the stress is above the proportional stress for the naterial, then the trial and er~or approach ~us~ be used as i1lust~ated in the problem dealing with a tapered column.
l~lt

f.----

a = 21" '_1'_9_" =1 L " 30" --.I Fig. C2.37

C2.18

STRENGTH OF COLUMNS WITH STABLE

CROSS-SECTIONS

(7) (8)

Same as Problem (5) but member Is exposed 1/2 hour to a temperature of SOOoF. Same as Problem (5) but change dimension (a) to 10 inches, and L to 14.28 inches. Find the failing compressive load for the doubly stepped column in Fig. C2.38 if member is made from 7079-T6 hand
forging.

(11) The cylindrical tapered member in Fig. C2.39 15 used as a compression member. If member Is made :Tom AISI Steel 4140, Ftu = 125,000, what Is the failing load.

-+- ae 12 -----oj<- b9
30"
Fig. C2.39

:I

(12) Same as Problem (7) but change dimensions to a = 6", b ::; 4.0", L = 14 inches. References:
FIg. C2.38

(10) Same as Problem (7) but Change dimensions a. 6 ft , b 4", L = 14 ft

(1) (2) (3)

NACA Technical Note 902. Non-dimensional Burkllng Curves, by Cozzone Z Melcon, Jr. of Aero. Sciences, October, 1946. Chart from Lockheed Aircraft Structures
Manual.

CHAPTER C3

YIELD AND ULTIMATE STRENGTH IN BENDING

C3. 1 Introduction.

Members subjected to bending alone or in combination with axial and torsional loads are quite co~on in flight vehicle structures. The lImit design loads on a structural member must be carried without permanent distortion and the ultimate deSign loads must be carried without rupture or failure. The well known bending stress e~~tlon tb = Me/I, assumes a linear variation ot stress with strain or, in other words, the equation holds tor stresses below the proportional lImit stress or, in general, the elastic range. Failure ot a ~e~ber in bending, unless there Is local weakness, does not occur at stresses in the elastic range but occurs at stresses in the inelastic range. Since the Ultimate strength of a member is needed to compare against the ultimate deSign load to be carried, a theory or procedure Is necessary which will accurately determine the ultimate and yield strength of a ~ember in bending.
C3.2 Basic Approach to Finding the Bending strength of Members.

compreSSion. In this example solution, we will find the internal resisting moment when we limit the unit strain at the extreme edge on the compreSSive side of the beam section to 0.010. Now plane sections remain plane after bending in both elastiC and inelastic stress conditions when member is in pure bending. We will guess the neutral axis as located 0.0375 inches above the centroidal axis as shown in Fig. C3.1b. Having assumed the max1mum unit strain as .010, we can draw the strain diagram at Fig. C3.lb. We now divide the cross-section in Fig. C3.1a into 20 narrow horizontal strips. Having the strain curve in Fig. C3.lb, we can tind the unit

The problem is to determine the internal reSisting moment of a beam section when subjected to stresses which fall in the inelastic range of stresses. This stress can be taken as the ultimate tensile or compreSSive stress ot the ~terlal or limited to some stress or deformation in the inelastic range. To obtain the true internal resisting moment, we ~ust know how the normal tension and compreSSive stress varies over the cross-section. The stress-strain c~-ve far the material provides the sa~rce for ob~aining the true stress pic~ure. I: a nate rial has a different shape in the tensile and compressive inelastic zones, the neu~ral axes does not coincide with the centroid~l axis, thus adding some difficult to an analySis 4ethod. The analysis procedure for deter.Jining the true internal resisting ~oment is best explained by an exa~ple solution.
C3. 3 Bendlng Strength of a Solid Round Bar.
~ig. C3.1a shows the cross-section of a round solid bar ~ade of alQninum alloy. The stress-strain curve u~ to a unit strain of .010 in. per inch is given in Fig. C3.2. Note that the shape of the curve in the inelastic zone 15 not the saue for both tension and

.01

.008 .006 .004 .002

.002

.004 .006 .008

(al
TeNSION

Fig. C3. 1

(e)

UHIT

C3.1

"'AI"
~<~~~~~~:~~~:~:~~;t~~.?~f- :;,~~;~.
-

----.---7'r:.~ ~. ~,: .~;_:' ';. ;" ,/;~. ;.~:,~~-,> ~~~,~:,;~' . ~. :</:/~<:.~~~ :,~/'.:,'~ ~.;.~~:::-;.- _":. :1~.:i: ~= ';-"1,(" ~. ~,~" ~ ,:~>~ -'
~

; .. -,

C3 2

YIELD

AND

ULTIMATE STRENGTH IN

BENDING

strain at the -ncoo i nt of each str-t. .c t th the strain known on each strip, the stress existing can be !ound by use of the stress-strain curve in Fig. C3.2. The total loae on each str1;J then equals the stress times the area of the strip. The internal resisting ~o=ent then

equals the s~tlon of the load on each strip times the distance from the strip to the neutral axis.
Table C3.1 shows the detail ~alculatlons. If the neutral axis has been selected in the correct position, the values in col~~ (6) ot the table should add up to zero since total tension must equal the t~tal compression on the beam cross-section. The small discrepancy ot 740 poundS in the summation of column (6) 1s not enough to change the location of the neutral axis or the total internal resisting moment appreciably. Column (7) gives the total internal resisting moment as 56735 in. Ibs. when the strain is limited to the .010 strain as previously discussed. The stress at this strain tro~ Fig. C3.2 is 49000 psi. USing this stress in the well known beam fonnula M rr/e , we obtain M: = 49000 x 0.785 = 38450, which is much less than the true

given bending ~oment, and al~J ~J~ teter~lnir.g t~e true internal resisting ~:~e;.t Jf a bea~ section, structural deSign 2~gi~eers ~ake ~se o~ a fictitious failing ben~ir.s stress ?o' wtn cn is re re n-ec to as a cccujus 0:' ructu-e stress in pUTS bendin~, T~3n the ultimate bending ~om3nt that can be developed jy a given beam cross-secticn ~~d a ~lven ~aterial is M = FbI/C. DeSign cu:ves fer finding ro' the modulus ot rupture, are given later in this chapter.

Since there are ~ny flight vehicle materials and all kinds ot SD2Des used in structural members, the basic approach for solution as illustrated in Table C3.1 becomes very time consuming. DeSign engineers always search for simplified methods which give sufficient accuracy. Thus Cozzone (Ref. 1) has developed a Simplified procedure for finding the modulus ot yield or rupturing bending stress Fb' The method is Nicely used in t~e aerospace industry in structural deSign.
C3.4 The Cozzone Simplified Procedure.

value or 56735.
TABLE C3.l

1
Slrlp No.

2
Strlp
Area

3
y

4
E

5
Unit

5
F :crA

Res.
Moment M= Fr

..

Stress
CO

"A"

1 2 3 4 5 6 7 8 9 10
11

12 13 14 15 16 17 18 19 20
Total Col. Col. Col. Col. Col. Col.

.058 0.935 .102 0.840 .135 0.75 0.65 .153 .165 0.55 0.45 .180 .185 0.35 .195 0.25 .197 0.15 .200 0.05 .200 -0.05 .197 -0.15 .195 -0.25 .185 -0.35 .180 -0.45 .165 ~0.55 .153 -0.65 .135 -0.75 .102 -0.84 .058 -0.935
3.140

.00867 .00773 .00685 .00591 .00494 .00398 .00302 .00205 .00108 .00012 -.00084 -.00181 -.00276 -.00374 -.00470 -', 00566 -.00663 -.00759 -.00846 -.00937

53000 52500 52100 51500 51000 43000 33200 22800 12500 3200 - 7250 ~ 17800 -29500 -3'5500 -40000 -43000 -44800 -46000 -47200 -48000

3075 5350 7025 7870 8410 7740 6140 4450 2460 640 -1450 -3510 -5750 -6560 -7200
-7100 -6850 -6210 -4810 -2780

2760 4300 5040 4820 4310 3200 1920 945 280 10 130 680 1650 2540 3510 4170 4710 4880 4210 2690
56735

The Cozzone method in its si~ple5t fern assunes a symmetrical rectangUlar bea~ section and the same shape of the stress-strain curle in both tension and compression. ?1;. C3.3 represents the true bending stress '~riat1cn over the bea~ cross-section when failure occurs. Cozzone now replaces this tru~ curve by a trapizoidal stress variation as sh0wn :n Fig. C3.4. The stress f o is a fictitious stress which is assumed to exist at the neutral axis or at zero strain.
Fig. C3.3 Fig. C3.4

fMAX

True Stress

[2 -- --- NA.lO L
IMAX=f fMAX

[tMAX=fml~-oI~fb'"
f

bT ..[]I
r

1_ Im _!

N. A.
Assumed Stress Fig. C3.6
~

""-

Assumed Stress

f MAX[

---;--

740

OS

1 3 4 5 6 7

Rod divided into 20 strips. 1" thick. y ,. distance from centerline to strip c. g. E ,. strain at midpotnt '" (y ~ .0375)/103.75. Unit stress for E strain from Fig. C3. 2. Total stress on strip. Moment about neutral axis. r = (y - .0375).

I --- /
?

.....-//

I
I

~Assumed Stress- I
Strain Curve

i
True Stress-Strain Curve Strain e
Fig. C3. 5
II

Since it is desirable to use the bean !'ormula In !1nding bending stresses due to a

_" .:~~).r,[.:~':'-~:'-"":{-~~;,~~;~'-~'.~1,:~~;'~(~~~~~~::":~:f~~,L~'::+;~'~~;''''''5~';'~:i;t''':~;';~~~I.:--.~",,:~-

--

-,. ';"~+ -~;/~~,',

r-"-:"...~.,

'."~,'.

~,~v

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C3.3

The value of f O is determined by ~king the r~quirement that the internal ~o~ent of the t~~e stress system must equal the moment of the assumed tra~9zoidal stress system which results from the assumed stress-strain curve as stown in rig. C3.5. Fig. C3.6 shows the trapez~ldal stress pattern drawn to a larger scale and showing only one half of the sy~~trial ba~~ section. Thp, trapeZOidal stress pattern has been divided into a rectangle (r) and a triangle (~) as shown in tne r tgur-e , Let, Mb = total lnternal resisting moment. internal my,nent developed by mr
portion (r).

Fig. C3.7

-I-

k Factor for Some Typical Shapes

,
./'
r.s

-K~~~-

I.

, .

-0-0-

,. ,.,
u

Z~

-0- ..,

mb = internal

mom~nt

develooed by

portion (b).

.._.
','rrrr-rrrrrrrrrrrrrrrrl
h',1

--_ ._.-1...... ,,,,._.._........ _-_.. .. _-

'

Sin:e fb varies linearly from zero to tb' the stress is elastic and thus the beam equation holdS, or mb : fbIlc for entire bea~ section. The stress variation on porticn (r; is constant or rectangular, thus

.
"

.'J

.. . r-,

..

00

..

F i g . Section Fa.ctor K for r, Cl, and C Sections (Ref. 2)

T~en m~

ro2~

for entire

bea~

section

C3. 5 Deetgn Curves for Finding Modulus of

Rupture (>D).

But fb Thus,

= f m - f o (from Fig. 2f oQ, lib = (fm-f a) 1.. c

I f m f o (I/e - 1)
2(1

C3.6 )

or

Moe

----

---

(1)

Let k -VC - 2~
k is a beam section shape factor.

----(2)

Let Fb = Mbe/I, then from equation (1)

Fb
mQ~ulus

= f!D.

+ fa

(k - 1)

- - - -

(3 )

Fb is a fictitious Mc/I stress or the

sec~lon

of rJpture for a particular crossat a ~iven ~im~~ stress level.

The modulus or rupture Fb may be a yield modulus, that is, in equation (3) the value at f m is equal to the ~~eld stress of the material. It ~y also be the ultimate modulus of rupture, in which case the value at t m in equation (3) equals the ultimate strength or the material. The modulus at rupture may be limited to a stress between the yield and ultimate stress of the material because at local cripoling or by excessive distortion. Regardless of what value is used for f m in equation (3), the corresponding value at fa must be known before the value or Fb can be determined. Figs. C3.9 to C3.23 give strain curves tor various material and the corresponding fa curve. The use or these two curves per.nit the determination at Fb if the k shape factor for the particular be~ section being considered is knO\VTI. In deriving the valueS of fa, the following assum,tions are made.

The values at k vary between 1 and 2.e. If calculatea value of ~ is greater than 2 use 2.0. Fig. C3.7 shows the valu~ of the sh~De factor k for several typical shapes. Ff g . C3.8 snows curves for the r-atn d detenoination of the k factor for 3 c~on beam sections.

(1) The stress-strain curve is assumed the same in tension and compression. (2) The neutral axis is assumed to coincide with the centroidal axis. (3) plane.
Duri~g

bending plane sections remain

(4) The cross-secticn is not subject to local or torsiQnal instability.

C3.4

YIELD AND ULTIMATE STRENGTH IN BENDING

(5) Beam-column, curvature and shear lag effects are considered negligible.
C3. 6 General Accuracy of Method.

to give the value of fa. The value of this chart o~eration ~ives f o 2 29000. from equation (3) FbYield = 65000
+

~o

fro:n. T!':er.
p~i.

29000 (1.5 - 1) = 79500

(1) It 1s exact for a rectangular section under pure bending with moment vector parallel to a principal axis.
(2) For double s~etric sections under pure bending and moment vector parallel to a prinCipal axis, the accuracy should be within 5 percent ..
(3) Single symmetric sections will vary tram practically exact to definitely unconservative (moment vector normal to axis ot

Then yield bending moment = M yp

= Fb1/c

Thus M yp = 79500 x .0938 = 7460 In. lb.

For finding the ultimate resisting bending, we use Ftu which is 75000 as the value of f m in equation (3). Ag~in going to Fig. C3.17 to stress of 75000 on stress-st~ain curve and the vertically dOem to to curve, we obtain f o = 70500.
Then F b (Ul t ) 75000
+

symetry) .
(4) For sections subject to combined bending and axial load, the results will vary tram practically exact to conservative. (5) For unsym:netr1cal bending, with and without axial load, the results will vary from practically exact to conserlative.
C3. 7 Example Problems 1n Finding Bending strength.

70500(1.5-1)= 110250 pSi

Then MUlt ': Fbllc = 110250x.0938 = 10370 in.lb. Let us assume that 1s 1s desired to limit the strain in the extreme tiber ~o .03 lnc~es per inch. What would be the bending ~oment developed under this limitation. From Fig. C3.17 for a unit strain of .03 the corresponding stress from the stress strain curve is 74700 and the f o stress 1s 61200.
Then Fb = 74700 Then M = FbIle
+

EXAMPLE PROBLEM 1.

A rectangular bean section is 0.25 inches wide and 1.5 inches deep. What yield and ultimate bending moment will the section develop when made trem 7075-T6 extruded alum1nu:n alloy.

61200 (1.5-1) = 105300 .0938

= 105300 x

= 9900

In. lb.

EXAMPLE PROBLEM 2.

Solution: The modulus ot bending stress is given by equation (3),

Fb = t m + to (k - 1) - - - - -

- (3)

k _ 2Q. - 2 x 0.. 75 x .25 x .375 x

= .1408
.0938

- DC k

(.l.) .25 x 1.5'(1/.75) 12

The symmetrical I beam section in Fig. (a) 1s subjected to an ultimate design pure bending moment M': 14000 in. lb. ~h~t Is the margin of safety if the beam is made' of magnes i urn forging AZ6lA 1.375 125 which has Ft u = 38000
and Fty

= 22000.

= 1.50
From

Ix

= 1~ x 1.375 x 2 3

2!JEl 1
Fig. (a)

. . ..fI*.
.125

The value of k could also be found in Fig. C3.7. Material is 7075-T6 aluminum alloy. Fig. C3.17, Ftu = 75000, Fty = 65000.

12 x

1.25 x 1. 75 3

= .916 - .558 = .358.

Ix/c = .358/1 = 0.358

Q = 1.375 x .125 x .9375 .4375 = 0.209
k
+

To tind the yield bending strength, the value at t m in equation (3), the maximum stress.permitted on the most remote fiber is 65000, the yield stress of the material. To find to, we go to Fig. C3.17 and find the pof rrt on the stress-strain curve that corresponds to a stress of 65000. This paint 1s proJected vertically downward to intersect the curve f o This paint is then proJected to the streSS scale at the edge of the chart

.875 X .125 X

=~ =
Ix/e

2 X

.209 0.358

1.17

k could have been ~rom Fig. C3.8 when using m = .125/1.375 = .091, and n = .125/ 1.75 = .0715.

ANALYSIS AND DESIGN OF FLIGHT VEHICt.E STRUCTURES

C3.5

From Fig. C3.l9, for t m = Ftu = 38000 we tind in projecting vertically downward to to curve gives r 0 = 23700. Then subt. In equatlon (3) Fb = 38000 + 23700 (1.17 - 1) = 40770
Mult = FbI/c = 40770 x .358 : 15000 in. lb.

llar.ln of Safety = M"M1t _ 1 = 15000 _ 1 14000

= -'(J7

bendIng strength was developed. From Fig. C3.l9, the unit strain when stress Is 38000 15 0.35. Then sInce plane sectIons remaIn plane atter bending, the unit straIn at point .50 lnch from neutral axls 15 (.5/1) (.035) = .0175. From Fig. C3.l9, the stress existing at thIs strain is f = 31000 psi. A linear variation ot stress as used in the flexural equation would give halt the maxImum stress or 38000/2 = 19000 psi as agaInst the true stress at 31000.

Assume we desire the stress inte~sity at a point 0.5 inches trom neutral axis it full

.
50

..

10

0.112 0.02 0.0" 0.011 0.08 0.10 0.12

0.0"

""lA

0.01

,."

""'lA
..
~-=~@i:j~i4io:~"" '-';:'::'':-o:;::~

,::-:2:~t;-':~!;~:",,,~:f~~i;~'I~:;::t

E~;_:,~.-'i:;g~":;if~?::;~:~:". ,~::.:'.::: :~'f:..-

-==....:=::-::. ~-:-:d~'::i::2:J.~l:~""

-,:":':;:"-'=if::~~,,.='000::'::..:

';

s "

~ ... .:l

s
- .,:"-"01 .:..~.::-t:~~;E~E:~ii~E';ii,--'=-'::E~-:"o;f~':'~~"=r_~"'::;,":''':-;-':

,."

'.N

, ""lA

0.011

0.08

0.10 0.11

;~:" ,0-::::;i~:Ol;';:.~::f0~g'~;:::~'-:.,~q~,i~~E"5'fE2f~ 0.01 0.02 0.03 0.04

~~f.#~",~'o-' 0.05

e 10/. .

r=:.::~8:~'f++~~~i~::;~~;::r-~:tn::':J=~:':~:~ttt~ "~,T:;:j~~t:r ~tF1

_._._-----.!~~~~j:_~'E:i~".,.-

._....:~,~_.__ .i_ic~L~~!_~:,Al.1ar~~,PIaIe'~

..:..r:so-~

-c-=---c,---,.--.-:---:,:-~;~,.,.~==~~=
-----------_._.~-

. . -'-,=Cjii.f~
0.12

: ""::"""'"'-

0.02

0.0 1

0.08

""lA

0.08

0.10

0.0'
Fb

0.0"

0.01

, ""lA

0.08

0.10

0.12

Curves tor finding Fb'

= tm +

f o (k - 1).

C3.6
;::~'.~i:~~"''';:,:b;~,".t<'':\~.

YIELD AND ULTIMATE STRENGTH IN BENDING

,:.,: ;, j

10fE:o:"'~-' -':,~::~::::.~~~.:.:.=~~.~~_.

s
....

'""~.-

-.-.

o .~.;~~.c~t~-=~;~
0.01

0.01

0.02

0.02

0.04

0.05 0.055

u.oa

0.03

0.04

0.05

e WID

W"

~~-,-oo;:~...,..:.;,:.:~'----'---,---o

-::~~~~-~,::::.:~=~t~;;~=~ -,,--.
-

-~--...,.~--,-,---c-;;-

. :';.;." ..::.:,. :':;.u.. Sl$'lD'-;'-"'; -"

~;U!imi",,' Hljnd'''''''''hlliri

.,

~ ~I

'

,,'~'" .""";",-,.~;-~''"~'~'-'~---.

0.01

O.OS

0.03

0."

0.0$ 0.01 0.02

, WID

rJ~~;~L;'7:,':--;'2:~;':";---.

....
,<0

,'+riL.~~J,',_
.

..,
-"-"~'?:"-~--?"'=--"-----"-".--" --.....---.----~." ~

t1Iia.-~k;'. r~-

" .J

';

.
<0
0.02 0.03 0.008 O.012 0.018 0.020 0.024 0.028 0.032 0.035

, W ..

I':. \lDcbel/lnl:/ll

Curves tor finding Foo

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C3.7

---- __--_~-:::--~:.:;::.:_;:;:;__-C==~===
.. --_.--"...:..:
.....::...-'=~

portion on each side ot the neutral axis ot the entire section. Label the beam portion below the neutral
axis as (1) and that above by (2) Portion 1. Fig. (d) shows
~

._..

_-----_._~",-,-.:....:.

__ . __ _ 'C--,_-.,

-~._,---=--

;,

! .'

how the lower portion (1) is made a symmetrical section about neutral axes
by adding the dashed

1-0.1"

portion. The internal bending resistance will be tound tor this entire
section in Fig. (d). one

-~r-u''-{--N.A. 1.39
_ J:. \
Fig. (d)

T , ", T 278,,1,1.39

halt ot this amount will then be the true moment developed by portion (1).
E:CAMPLE PROBLEM 3.

Unsymmetrical Section.

1 =12 bh

_ 1 a - 12 X 0.1 X 2.78

= 0.178

Fig. (b) shows a tee beam section, symmetrical about the vertical axis. It the material is 17-4 PH stainless steel, what ultimate bending moment will be developed it bottom portion is the tension tlange.

1,/c,

= 0.178/1.39
.193

Q, = 1.39 X 0.1 X

= .128 .695 = .0965

=
1 5

2Q.,

--<..1_1-1/2 ---+I

i Lo.
0"

IN.A. --

I"

----1-----iO.610
y=1.39

----

Ell"'. 0154

0.193 0.128

From Fig. 03.22, Ftu

= 180000 which equals

____ ___l 1
o. 1"~
I.... Fig. (b)

.~
E1. =.03 Fig. (e)

to trom curve
Then Fb,

= 156000

The neutral axis will tirst be determined.

=t m + to (k - 1) = 180000 + 156000 (1.5 - 1) =258000 (Fb, 1Jc, It =258000 X .12B X

0.5

16550 in. lb.

Fig. (c) shows the unit strain picture. The lower edge at the be~~ section is strained to the maximum value at .035 as shown on the stress-strain curve in Fig. C3.31. Since plane sections remain plane the unit strain E2 at the upper edge ot section is E. = .035 x
.61/1.39

The factor 1/2 is due to the tact that portion (1) is only one halt the beam section
in Fig. (dl.

Portion 2. Fig. (e) shows the developed symmetrical section tor the upper portion (2) at the beam section.

= .0154.

, ~l.s--->I

e =.0154

Solution 1 Equation (3) was'derived tor a symmetrical section about the neutral axis. The equation involves finding the resisting ~oment developed by one halt the beam section and ~ultiplying by 2. This is permissible since the unit strain at both top and bottom edges is the same. In this solution we will continue to use equation (3). To do this it Is necessary to ~ke symmetrical sections for the

o0.tT'lC -1 +
o. -61
.3-

N.A.~, - - - - 1.22

~S[:~~

C::==_::-:::::l _
Fig. (e)

I I

--M-I"5"4Fig. (I)
-

The unit strain picture Is shown in Fig.(f)

III

= l~

x 1.5 x 1.22 3

l~ x 1.4 x 1.02a = .104

3Cf.-

C3.8

YIELD AND ULTIMATE STRENGTH IN

BENDING

I./e, = .104/.61 = .1704

Qa ::
2'~
iii

Laic a = .0514/.61 = .0842,

QII ::

.61 X .1 X .305

1.4 X 0.1 X 0.56 ;. .097

= 0.194
k.

0.61 x 0.1 x .305 = .0971

+ 1.4 x .1 x .561

k,

= 0.194/.1704 = 1.14

.0971/.0842 = 1.15

The stress for a unit strain of .0154 fram the stress strain curve in Fig. C3.22 is 172000, and to :: 129400. Then Fb,
M,

As explained in solution (1), for e =

.0154, f m :: 172000 and to :: 129400.

= 172000

129400 (1.14-1)

= 190100

Fbll = 172000
M,
Mtota1

129400 (1.15-1)

191400

191400 x .0842

= 16100
+

=~

(Fb,,1,/e,)

= 190100 x
?

.1704

= 16200
?

=M

M3 = 16620 32720 In. lb.

1

16100 =

Total resisting moment:: Ml. 16200 32750 In. Ibs. Solution 2.

Mil ::

16550

EXA11PLE

PROBL~

sym~etrical

4. Fig. C3.24 shows an unI beam section. The material is

Instead or making each portion a section as was done in solution (1) and dividing the results by two, we will tind the internal bending resistance at each portion as is when bending about the neutral axis ot the entire section. Equation (3) now becomes tor each portion at beam section,
s~etrical

7079-T6 aluminum alloy die forging. The upper portion 15 in bending compreSSion. It will be assumed that the compressive crippling stress for the outstanding upper legs of the section is ,65000 psi. (The theory and method ot calculating crippling compressive strength is given in another chapter.) ihe ultimate deSign bendIng moment is 16500 in. lb. Find M.S.
1---1.5-01
E=.009

Fb

=t m + to

(k - 1) - - - 3'
,

where kl.

= I~/' , e,

The section modulus at each Dortion refers to neutral axis otentire beam section. Fig. (g) shows lower portion (1).

-N.A.-

= .0895
I,/e,

1] ~:~~
=
.0645
b_! 1+-0.1
FIg. (g)

e ~::. 0129 Fig. C3.25

.OB95/1.39

Solution:

QJ,:: 1.39 x .1 x .695 = .0965 kl.

= Q./IJ,/cl.

= .0965/.0645 :: 1.50
,

From equation (3 1 ) (1.50 - 1) 258000

Fb = 180000

156000

MJ, = Fbl. X Il./cJ, Fig. (h)ShoNs uDper portion (2)

r, :: ~ X 1.5 x .61:5_

=258000 x
11 -

.0645 :: 16620

The mqximum compressive stress permitted is 65000 ps i . From Fig. C3.23, USing the stress-strain curve, we obtain a unit strain ot .008 for this stress. The unit strain at the bottom edge of sectIon is .009 (1.1781 0.822) :: .0129 as shown in Fig. C3.25. From Fig. C3.Z3 this strain causes a stress ot 67500 pet . Upper portion:
Fig. I)

x 1.4 x .51:5

t~ 161=c2 N.A_
Fig. (hi

15 --+1

(See

Tr-;f~i
0.822

_1 x 1 .~ I NA -'3

1 . 822' - '3

:.i"

','

0.1

= .0514

xL4 X 0.722 3

-LN.A.-- Fig. (0

--N.A.

.102

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C3.9

I/e = .102/.822 = .124

Q = 0.822 x 0.1 x 0.411

::: .1418

1.4 x 0.1 x 0.772

section will develop in bending about x axes and M x the design bending.
(2) Using M Y 1 carry out the sane procedure for bending about y axes and tind,

= .1418/.124 = 1.142

r~ax

= 65000. From
+

Fig. C3.23, fa
~

= 30000
69260

Then fb = 65000 m.

30000 (1.142-1)

Then .the moment ratio Rb tor combined bending is.

Rb

= fbI/e = 69260 x .124 " 8580 in. lb.

= Rbx

+ Rby

Lower Portion.
INA =

~x

0.75

x 1.178"
3

1 -3 x

0.65 X 1.078

I!_N.A.
1.178

Then margin

or

safety M.S.

=R~

- 1.

C3. 9 Section With One Axis of Symmetry. with Moment Vector not Parallel to Either Axis.

INA

.137, I/c = .137/ 1.178 " .1164

"0.75-; To. i
Fig. 0)

Since the symmetrical axis Is a principal axis, the procedure in this case Is the same as for the double aymnetrlc case.

k = .1425/.1164 " 1.222

f max

= Rb x
II.S. Rb

Rb y

: 67500

".l.._1

From Fig. C3.23, f o ::: 44500

tb :: 67500 + 44500 (1.222-1) :: 77:!70 psi
!D. ~

C3. 10 Unsymmetrical Section with No Axis 01 Symmetry.

.1164

x 77370 = 9000 in. lb.

Fig. C3.27 shows an unsym-

Total internal allowable resisting

mccent > m1. + m,

metrical section subjected to the applied moment

or

Ma

= 8580
II

vector I1.
+ 9000

= 17580 in. lb.

For this case the procedure 15 as follows :
Fig. C3.27

Rb " 11a " I7580 = 0.94 (Load ratio) II.S.

16500

= R~ - 1 = 0.

194

- 1

= .06

(1)

C3.8 Complex Bending. Symmetrical Section. Moment Vector Not Parallel to Princi.pal Axis.

Determine principal axes location by equation,

2Ixy

Fig. C3.26 shows a

tan

2Q '=

double

sym~etrlc

section.

Iy-I x

The x and y axes are

therefore principal axes. The following procedure can be followed which IS qUite conssrvat I ve . (1) (2)

--.
I
I
Y

where x and yare centroldal axes, Ix and I y are mo~ents of inertia about these axes and Ixy the ?roduct of inertia.
(2)

Fig. C3. 26

Resolve the given moment M into components M x p and M yp ' Follow the
sa~e

(3) Resolve the given ~oment into cry,nponents about x and y axes, or M x and My. USing M x follow the procedure as given in the exanple problems and find Rbx = MxlMax' where Max is the internal moment (4)

procedure as betore.

The stress rat10 Rb

= Rbxp

Rbyp

C3.10

YIELD AND ULTIMATE STRENGTH IN BENDING

C3. 11 Alternate More Exact Method for Complex

Bending.

a shear V = 600 Ibs. The problem is to find the margin of safety under this combined loading. For pure bending only the stress ratio is
- M - 15000 14000 = ' 933 Rb-I1 (the va.iue 1 t u '5000 i s cr a obtained from example problem 2).

A beam sectIon when resl~ting a p~e external bending moment bends about an axis that is called the neutral axiS. No matter what the shape of the beam cross-section for any given external moment, there is an axis about which bending takes place. The general case involves an unsymmetrical beam crasssection and material which has different stress-strain curves in compression and tension in the inelastic range. The neutral axis therefore does not pass through the centroid of the cross-section and thus the method at solution is a trial and error approach. The solution procedure is outlined in Chapter A19, Article Al9.l7, and there tore will not be repeated here. Also the chapter dealing with the design at beams with nonbuckling webs explains and illustrates .how the ultimate bending resistance at an entire beam section is determined.
C3.12 StrengtbUnder Combined Bending and Flexural Shear.

The stress ratiO in shear is R s = fs/F au , where f s is the flexural shear stress and Fsu the ultimate shear stress of the material. The problem therefore is to find the value of ts' The equivalent trapeZOidal and tr iangiuarbending stress distribution will be determined tor the deSign bending moment of 14000 in. lbs.

Fb

=Mell = 1400010.358 =39150

For a triangular stress variation,

pSi.

From example problem (2) the shape factor

k was 1.17.
k =

The previous part at this chapter has dealt with the determination of the strength at a beam section in pure bending. The usual beam deSign problem involves tlexural shear with bending. In tlnding the true internal resisting moment, the Cozzone simplified method derives a trapezoidal bending stress distribution which will produce the same internal resisting moment as the true internal bending stress system. A triangular stress system is then derived which will also give the true bending moment. Now the equation for flexural shear stress tor a triangular bending stresS distribution

On Fig. C3.l9, the curve for 1.17 has been plotted. Starting with the Fb stress of 39150 at the left scale, ~un horizontal to an intersection with the k = 1.17 curve, the prOjecting vertically downward to intersections with the stress- strain curve and the fa curve to give 35800 and 19700 for fa' The stress results are shown graphically in Fig. ~~.28a and Fig. C3.28b.

t 11-- 'i--.{

19700--+1

1 .... 16100

/+39150..

J _:""==Tt~](
-I
~8

....[

I.-

Is
- - (A)

1/2 Beam Section

970b Fig. C3. 28a Case 1

Fig. C3. 28b Case 2

Thus to use equation (A) for a trapizaidal bending stress, a correction factor (C) must be applied or equation (A) becomes ts - S. It
- (B)

The flexural shear stress is a function at the rate of change at the bending stress. Thus we can obtain a shear correction factor C by comparing the bending stresses in the two ~tress distribution diagram. The shear stress 1s maximum at the neutral axis in this particular problBffi. The total normal force on the cross-section of beam above the neutral axis equals the stress times the area. For simplification, the beaL, section will be divided into the two portions labeled (a)
and (b).

To illustrate how the correction factor (C) can be determined, the I beam section used in example prable~ (2) will be used. We will assume the ulti~ate design moment at 14000 in. lbs. is produced by a load of 600 Ibs. acting on a cantilever beam at a point 23.30 inches from the fixed end of the bea~. Thus the be~ section at the support is subjected to bending moment of 14000 and

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C3.11

For Case 1:Load on portion (a) is,

19700 16100

the cross-section will be determined. The correction factor tor this paint is found by comparing the loads on portIon (a) tor Case 1 and Case 2 stress patterns.

x 1.375 x 0.125
+

= 3380 lb.
Case 1. Load on (a) : 3380
Load on (a) = 6300.
+

2600 = 5980

14100 "" 1.375 x 0.125 = 2600 lb. 2

Case 2.

Portion (b)
19700 x 0.875 x 0.125 14100 x 0.5 x 0.875 x 0.125

Hence, C = 5980/6300 "" .95

= 2155 = 771
8906 Ibs ,

C3. 13 Strength Under Combtned Bending Flexural Shear and Axial Compresston.

Total Force For Case 2:Portion (a)

The SUbject of the ultimate strength design under combined loads is treated in detail in a later chapter.

A conservative interactIon equation tor combined bending, shear and axial load is,

39150 ; 34250 x 1.375 x 0.125 = 6300

Portion (b)
34250 x 0.5 x 0.875 x 0.125 1874 = 8174

or M.S. = .; (Ra + Rb)'

res,

Rb inclUdes effect of secondary bending moment due to axial load ttmes deflection.

Thus the correction factor Is,

C 8906/8174 = 1.09
C3. 14 Further Values of fm and

fa.

Then f s at neutral axis Is,

s ""

It -

Cvq - 1.09 x 600 x 0.209 = 3060 1 0.358 x 0.125 ps

Table C3.2 gives the yield and ultimate values at f m and to tor a number of other materials common to the aerospace field of structures. The yield and ultDnate modulus in bending Fb Is found by substituting in the equation.

The ultimate shear stress for this particular magnesium material Is 19000 psi. (See chapter on material properties.) Stress Ratio R8
:::

1'8 3060 F ::: 19000 ::: 0.161 sw

TABLE C3. 2 Values at (m <U1d fa for FilId.tnr Fb

Fb'"flll+to(k-l)
M:U:ertal Yield Ultl.....

S...
t s 4 in. to! 0.02 e ~ 2.0 t s 6. a t:50.25 .016-0.25 ""0.188 r e O. las

fCl"Fty

Co tm'"Ftu
18.8 17.5 25.5 33.5

The interaction equation for combined bending and flexural shear Is

Co
57.5 40.5 70.0 67.0 15.7 22.6 30.8 83.1 86. a !l8.0 146. a 172.0 192. a 161. a 160. a 116. a

Rb

+ R~

1 1

Whence, Margin at Safety =

1

h;

R~

-"

or, M.S.

= /.9'&:/'+

.161" - 1

= .05

ZOI4-T8 AJ.Ua:l. AJ.. 01. Forgtnp 6061-TS Alum. AI. Sheet 7075-T8 Alum. Ai. me Forgings 7079-T6 Alum. AI. Hand Forgln(s(L) A. Z81A MlIcnesluCl AI.:i::rtru.. (Long) lDCIA_O M.ap.elli\l.ln AI. Sheet ZK60A Magnesium AI. Forg1ngs (Lolli) AISt Alloy Steel (Normal1zedj AISI Alloy Steel (Normalized) AISI Alloy seeet (Heat Tre"ted) AlSI Alloy Steel (Heat Treated) AISt Alloy Steel (Heat Treated) AISI Al10y Steel (Heat T~ated) tT-7PH Stainless Steel PHIS-7 MQ {RH9501 Stainless Steel TI-8MN Titanium Alloy

52

" es ea
ae
21 16
TO

ea 52 ts
71 38 30
52

. c
T.5

rca
16'
lT5

'"
ISO

"

200

n,

43.2 47.1 ISL!l 66.0 53.0 65.0 62.0 5$.0 36.0

so

'" 16' aas

16' 16'

" '"

Thus the effect of the flexural shear stress was to reduce the margin or safety of .07 in pure bending to .05 in the combined stress action. AS further calculation at the shear correction factor C, its valu8 for the shear stress at the upper edge of portion (b) of

'"

PROBLEMS (1) A round tube 1-1/2 inches in diameter has a wall thickness of .095 inches.

It is made of aluminum alloy whose stress-strain CU~le Is shown in Fig. C3.2. If the maximum unit strain in

C3.12

YIELD AND ULTIMATE STRENGTH IN BENDING

compression is limited to .008, what bending resisting moment will the section develop. (Note, since stress-strain curve has different shape in tension and compression, neutral axis does not coincide with center line axiS, thus use trial and error method.

Fig. 1

--*.I+--

--+1

j"'t:.:

O. 1

(2) Same as Problem (1) but use a square tube

with 1-1/4 inch outside dimension and .081 inch wall thickness. Use the Cozzona method for solving the tollowing problems.
(3)

t
Fig. 4

...... 0.1

t~l -'+-16
8 I

.1<4- 14"-l

-Jol

1 I+2

Find the ultimate bending moment that each ot the following beam sections will develop when bending about the principal axis and made trom each of the following materials.
(a)

Fig. 5

Fig. 6

(4)

7075-T6 Alum.1num. Alloy Extrusion. Ftu = 75000, Ft y = 65000.

(b)

Ti-6Al-4V Titanium Alloy. Ftu = 130,000, Fty Z 120,000. Ftu

A simply supported beam has a span of 24 inches. Depth 0: bea~ limited to 2 inches. It must carry an ultimate load of 4000 Ibs. located at midpoint of beam. ~aterial is 7075-T6 aluminum alloy extrusion. Design an r shaped section to carry this load. Neglect areas ot corner fillets that would be used in extruded Shapes.

(c)

ArBI Alloy Steel, heat treated. 150,000, Fty = 132,000.

REFERENCES

Ret. 1.

Bending Strength in Plastic Range. By F. P. Cozzone, J. Aeronautical Sci.,

l1a.y, 1943.

Ret. 2.

Vought Structures Ma.'1.ual.

.O.T,

.,_~

~.~

CHAPTER C4

STRENGTH AND DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBING IN TENSION, COMPRESSION, BENDING, TORSION AND COMBINED LOADINGS.
C4. 1 Introduction.

Berare the advent or the stressed-skin structure tor ai rc ra.rt , or during the period when tuselage and wing were rabric covered, round, oval and square tUbing were used in designing the major structure or the fuselage and wing. It the wing and tail were externally braced, streamline tubing was used. The development or the metal covered structure el1minated the use or tubing in fuselage and wing design, however, tubing continued to be used for landing gear structure, engine mounts, control systems, fixed equit:m,ent such as passenger seats, etc. With the opening or the space age, tubing as a structural unit in space vehicles is again being widely used because drag in space is not an important factor. Round tub mg is the best shape tor transmitting torsional torces and thus Widely used in control systems. Round and square 'tub Ing permit Simple connection or end fitting design. The metals industry has made available a large number at diameter and wall thicknesses and thus the structural designer has a large number ot sizes to select rroa.
C4.2 Design for Tension.

steel and 2024 aluminum alloy. Observation or Fig. C4.l shows that for temperatures below 3500F, aluminum alloy is lighter unless the steel is heat-treated to Ftu 180,000 or above. Above 350 oF, the ult1mate strength of aluminum alloy falls Off rapidly, but steel continues its rather uniform decrease in tensile strength. The graph explains why aluminum alloy cannot be used entirely for the surrace of supersonic airplanes flying at speedS around 2000 miles per hour, as aerOdynamic heating would produce surface temperatures in the region where the strength Of aluminum alloy decreases rapidly.
to

:''':':;'::-,i.=:,:c:,:::,~:~::''-:; ,L:::::: ,:h .:~~:.: :;c::;,ltt;~~,~'~::r-:':;:~~3:-:_'-~:. =j

700

:~:~~~~;;~~->~C~;2~fu~;;i~
j;~~~~~~-"~'"~;;~=:~~~~~~~~i
~~"c:"_j':::.::::3':=:-=-:::

g
2
~

600

~ : f~:~g~~~~;~:'~~2-~~~;:_;~:~;~rf~~~~
f"'"-'''
300

In general the strength deSign requirements are that the l1m1t loads must be carried without exceeding the tensile yield stress (Fty) of the-material and the ultimate deSign load which is equal to the limit load times a factor of safety must be carried without failure, which means the tensile stress cannot exceed the (Ftu) Of the material. In general for aircraft, the factor of safety is 1.5. For unmanned missiles and space vehicles the factor of safety ~y be as low as 1.2. Since the ratio of Ftu to Fty for ~terials varies Widely, sometimes the yield under limit loads is more critical than failure under the ultimate deSign loadS, thus the stUdent should always be sure he has the critical situation. Since elevated temperature and time of exposure effect the yield and ultimate strength of materials, the problem of material selection relative to light weight becomes an important design factor. Fig. C4.l shows a plot of the Ftu/w ratio versus elevated temperatures up to 800 degrees, where w 1s the density of the material. The tUbe materials are AISI alloy
C4.1

: : ~~~~tiii~~-f~i~iiti
100 200 300 400 500 600 700 Temp. OF. 1/2 Hr~ Exposure FIlJ. C4.1 600

~~~~~~~~i~~ii:~.~iE~~;~~~z_ ;~=t=~~=-:_

,",",.."",=,

.. -. . '

'-'"""E--- --. :-:-"'':;:';:;'1

-.~.''''''',~,.,,~'''''

i_

._~=:--o:

Practically every structural tubular member in a flight vehicle structure must be fastened or attached to another adjaCent member. The connection can be made by USing same sort at end fitting which is fastened to the tube by rivets or bolts or by welding. It rivets or bolts are used, holes are cut in the truce walls which means the tube is weakened since tube area is cut away, thus net area on any tUbe cross-section must be used in calculating the tube tensile strength. If welding is used in the connection, the welding heat causes grain growth in the tube material adjacent to the weld area, which decreases the tUbe strength, thus a welding correction tact or must be used in calculating the ultimate tensile strength. DeSigning for tension will be illustrated later in this chapter.

Ct.2

STRENGTH &: DESIGN OF ROUND, STREAMLmE, OVAL AND SQUARE TUBING IN TENSION, COMPRESSION, BENDING, TORSION AND COMBINED LOADlliGS.

C4.3 Design for Compression.

The strength of members with stable crosssections when acting as columns can be calculated by Euler's equation it the bending failure Is elastic, or E 1s constant (eq. C4.1) and tor inelastic bending faIlure, Euler's equation ~lth the tangent modulus Et replacing E (eq. C4.2) checks test results. (The student should refer to Chapters AlB and C2 for theory on column strength.)
(L'/p)" Fe' (L'/p)"
L/~

(C4.1) - - - - - - - - (C4.2)

any tendency of the test curve to rise rapidly tor very short lengths where failure is by block compression. Table C4.l shows the resulting short and long column equations after values of Fco and E have been substituted in equations C4.3 and C4.4 and E in the Euler equation. The column headed transitional L'/p, represents the value at L' /p where failure change trom inelastic to elastic failure or, in other wordS, it is the dividing paint bevneen the so-called long and short column range. Thus if the equations are used, the Lr/p value must be known in order to select the proper equation.
C4.5 Column Formulas for Aluminum Alloy Tubes.

U Is the effective length and equals

where c is the column end fiXity coerr icrent
Long -and Short Columns For many years the prOblem or SUbject or inelastic column strength or !allure was

Fram (Ret. 1), the basic short column equations for aluminum alloys are:Fe Fe o [1 - 0.385 (L' /p)/nV E/F e o] Fe (C4.5 )
(C4.6)

= Fco [1

0.333 (L'/p)/nV E/Feo]

treated almost entirely trom a consideration ot test results. That is, sufficient tests were made to establish the shape of the failing stress curve in the region where the failure was at stresses above the proportional limit stress of the material. MathematIcal curves were then derived to tit the test results. Engineers referred to the .columns which tailed by inelastic bending as short columns, and thus reterred to the equations that tit the test data as short column equations. The columns that ~ed by elastic bending were then referred to as long columns. The test CUI"le tor long columns would follow the Suler column equation (C4.l) and thus tests were not necessary to establiSh allowable falling stresses in the so-called long column range. Thus over the years short column equations based on test results have been presented by official government agenCies for use in structural design. The official publication tor the aerospace tield is the Military
Handbook IUL-HDBK (Rer. 1).
C4. 4 Column Formulas for Round Steel Tubes.

Fe Fc o [1 - 0.272 (L'/p)/nV E/F e o] For long columns:Fe

(C4.7)

= nOE/(L' /p)"
The
equatior~

------

- (C4.8) The table

given in Table 04.2 (from Ref. 1).

ma.ter-ta.Ls ,

for determining Fco are

also indicates which ot the three short column equations to use for the various aluminum alloy To illustrate the use of Table C4.2, the column formula for 2024-T3 aluminum alloy tUbing will be derlved:From Chapter 82, we r ind the following strength properties for 2024-T3 tubing, Ftu
a

54000,

Fey. 4200C

From Table C4.2, the equation tor Fco is, Fco ~ Fey (1 + ,; Fey/lOCO
Fe o 42000 (1 +

SUbstituting,
42000 (1 + .205)

V42/1000

From (Ret. 1) the basic short column equations are:Fe' Fcc [l-F e o (L'/p)"/4n'E] - - - - - (C4.3) Fe

=
From Table C4.2, we note that short column equation C4.5 applies. SUbstituting in this equation,

50600

Fcc { 1- 0.3C27 [(L' /p )/n

V EIF e o]

r,oj

---

(C4.4)

Fe = 50600 [1-0.385(L'/p/nV 10,5oo,000/50600J or Fe = 50600 - 431 L'/p - - - - - - - - (04.9)

C4.6 Column Formulas for MagneSium Alloys.

Where Fc o is the colunn yield stress (upper limit at column stress for primary failure). It can be determined tram test results by extending the short column curve to a point corresponding to zero length, ignoring

From (Ref. 1) the following short column equations for various ~gnesium alloy materials

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES Table C4. I Column Formulas for Round Steel Tubes Short columns (a) Material Ftu. its! Fty. Itsi Fcc, Itsi Column formula (b) 1025 .. , .......... 4130 , ... , ..... , .. , Heat-treated e alloy steel. , , ... ' ... Heat-treated alloy steel, ..... '" . Heat-treated alloy steel ......... 55 95 125 150 180 36 75(0) 103 132 183 36 79.5 113 145 179 36,000-1. 172(L'/p)3 ", 79,500-51. 9(L'/p) J,. 8,. 113,000-11. 15(L'/p)' . 145,000-18. 36(L'/p)" .. 179,000-27. 95(L'/p) a Basic
equat~on

C4.3

Transitiona! (C) Long columns (b)

L'/a
122 91 73 83 58 266 x lO'/(L'/p)' 286 x 10o!l/(L'/p).11 288 x 10'/(L'/p)' 286 x lOo!I/(L'/p).II 286 x 10o!l/(L'/p).11

Local failure

C4.3 C4.o4 Co4.3 C4.3 C4.3

... (d) ... (d) ... (d) ... r')

...

(d)

a Equatfun C4. 1 may be used in the short column range if E is replaced by Et obtained from the combined stressstrain curves for the material. b L'/p '" L/py'C: L'/p shall not exceed 150 without specific authority from the procuring or certificating agency. c Transitional L/p is that above which columns are 'tong" and below which they are "short." These are approxtmate values.

d Not necessary to investigate for local lnBtabUity when O/t 50. e This value is applicable when the material is furnished in condition N (MILT-6736) but the yield strength is reduced when normalized subsequent to welding to 60 Itsi.

are given in Tables C4.3 and C4.4.

Table C4. 3 Column Formulas for MagnesiumAlloy Extruded Open Shapesa GENERAL FORMULA K(Fey) Fe' (L',jJ)m (Stress values are in lesi) Alloy
, MIA . , . ' AZ3IB, AZ61A, AZ80A. AZ80A-T6, ZK60A-T5 .

a
1/2 1/4 1/4

Max. Fe 0.90 Fey

or inelastic instability of the column as a whole. As the slenderness ration L'/p gets smaller, the Fc stress increases. Now it the diameter of the tube is relatively large and the wall thickness relatively small or, in other WOrdS, it the diameter/thIckness (D/t) ratio 15 large, tailure will result by local crippling or crushing ot the tube wall and this local tailing stress is usually represented by the symbol Fcc. The values or Fcc in general have been determined by tests (see design charts for Fcc versus D/t ratio).
C4.9 Design Column Cha.rtl!I

.. .

. .

180 2,900 3,300

1.0 1.5 1.5

O. 6 Fey

Fc~

a Formulas given above are for members that do not fail by local buckling. Table C4. 4 Column Formula for AZ31B-H24 Magnesium-Alloy Sheet (1.05 Fcy)'(L',p ) , Fc'" 1.05 Fcy4 n"E
Max. Fc .. F cy

In design, column strength charts are a great time-saver as compared to substituting In the various column equations, thus a number at column charts are presented in this chapter to facilitate the strength check at columns and the strength deSign of columns. Fig. C4.2 is a chart of L'/p versus Fc for heat treated round alloy steel tUbIng. Fig. C4.3 Is a similar type or chart for aluminum alloy round tubing. Fig, C4.4 gives column charts for magneSium alloy materials. All three charts are taken tram (Ret. 1). Figs. C4.5 and C4.6 represent a turther S1mplication ror the design at steel and aluminum round tubing.
C4.10 Section Properties of Round Tubing.

C4.7 Short Column Equations for Other Materials.

For other metals tor Which short column equations are not available, the use of Euler's equation, USing the tangent modulus St can be used (eq , C4.2). Refer to Chapter C2 for intornatlon on how to construct column strength curves using this equation.
C4.8 Column Failure Due to Local Failure.

The equations as presented give the allowable stress due to failure by bending of the column as a whole and the action is elastic

Table C4.3 gives the section properties of round tubing. A tube is deSignated by giving its outside wall diameter (0) and its wall thickness (t). Thus a 2-1/4 - .058 means a tube with 2-1/4 inch outside diameter and a wall thickness in inches ot .058. Since a tube is symmetrIcal about any axis, the polar moment or inertia, which Is needed in torsion problems, equal twice the rectangular moment of inertia as given in Table C4.3. For weIght comparison, the weight at steel and aluminum tubing is

C4.4

STRENGTH & DESIGN OF ROUND, STREAMLlliE. OVAL AND SQUARE TUBrnG IN TENSION. COMPRESSION, BENDING. TORSION AND COMBINED LOADrnGS.
Table C4. 2 Column Formulas for Aluminum Alloys

Alloy and Temper

Product

Short Columns

Transitional L'!p

Long Columna

2014-T'J, T", T451 202....T3. T351, T36, T4, T42

Sheet and Plate.; Rolled Rod, Bar and Shapes; Drawn Tube

5052-All Tempera 6083--All TempeJ'l!l

~AUTempera

54M-All Tempera MM-All Tempe" 6Ofl1-T4, T"Sl, T4510, T4511

All

AU Products

F (1

+..,; F,./lOOO)

Equation

1.732."/ ElF..

Equation

C4.5

C4.8

Cut Tempen

Alloys

and

Sand and Permanent Mold CaatinlJ8 Enru.sions

2OU-T3310, T3S11, T4, T42

2014-T6, T651
~T8,T81,T86,~1

7075-T6, T6S1 7173-T6, T6S1

6081-T8, Teal 201.....T8, T651, T6510, T6SlI, T6S2 2024-T8, TOI, TaStO, 2024-T6, TOI, T85tO, T8611, T862 7075-T8, T8610, T862 70'1l1-T6, TastO, T6SlI, T6S2 7178-T6, T6S10, T6S11

Sheet and Plate*j Rolled Rod, Bar and Sha.pes; Drawn Tube

F H

+../ F .,/1333)

Equation

1.346 ../E / F..

Equation

C4.6

C4.8

Sheet and P1.te

nm,

Eztruaiona, Forgings

6Ofl1-T6, T6S11

T65l,

T8StO,

Rolled Rod, Bar and Shapes; Drawn Tube; Extrusions Sheet and Plate.

F ,.(1+ ,; F .,/2000)

Equation

1.22. ."/E/F..

Equation

C4.7

C4.8

X202O-T6, T651

'"IDdudes eJad u weUu tNIre sheet and plaC.l. TtmsttioDal L'I_ Is that above _hleh the columns .,. "1onC'" llnd beknr _b1ch t.be1 ..... ''short''.

Equation C4. 8 may be used. In tbe 'hort uolumn rllZI.P If E' Is l'l!plaeed by , obtained from the eompreulva StrlSSoStraincurve for the

matllr1&l.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C4.5

given in tfie last two columns of the

C4.11 Some General Facts in Tubing Design.

ta~le.

Table C4.4 Tension Allowables Near Welds in Steel Tubing (X-4130) Type of Weld *Tapered Welds of 3()O or Less
All others

1.

For a given area, the larger the tube diameter, the greater the column strength if failure due to local crippling is not critical. The higher the D/t ratio of tube the lower the crippling or local failure strength. If columns fall within the long column category, the use of higher strength alloy steel or aluminum alloy will not increase strength at column since E is practically constant tor all Chrome-moly steel alloys and l1kewtse for all aluminum alloys. Failure is due to elastic buckling of the column as a whole and is therefore a function only at I, L' and E. The col~ end restraint eftects the necessary tube size. Consult the design requirements of the Army, NaVYJ and C.A.A. in this matter. In general with welded steel tubular trusses a coefficient of C = 2 is permissible except tor engine mount and Nacelle structures. For trusses with riveted joints a value of not over 1.5 is generally permissible. The student should realize that practical limitations such as clearance requirements may determine the diameter ot the tube instead of strength-weight considerations. Thus design can consist of checking the tubes available under the given restrictions.

Normalized Welded after HT HT after Tube Welded or Norm. after Weld Welding .947 Ftu 90,000 psi 80,000 psi .90 Ftu

2.

.841 Ftu

.80 Ftu

* Note: Gussets or plate inserts considered 00 "taper" with . . ** For

3.

(X-4i30) Special, comparable, values to the FtuJ equal to 90, 000 and 50, 000, are stresses 94, 500 and
84, 100 pst, respectively.

Ret. Anc-5

C4.

13 IDustrative Problema in Strength t:hecktng

and Design of Round Steel Tubes as Columns and Tension Members.

4.

PROBLEM 1

Tube sIze: 1-1/2 - .058, Length L : 30 In End tix1ty coefficient C = 1. Material:- Alloy steel J Ftu 95000. The tube is welded at ends. Ult~te design loads are:- P -14,500 lbs. compression, and P ~ 18500 lbs. tension. Required the Margin ot Satety (M.S.)

5.

C4. 12 Effect of Welding of steel Tubes Upon the Tension and Column Strength.
Si~ce welding etfects the grain structure ot the tube material adjacent to the weld, tests show the strength of the material adjacent to the weld is decreased as compared to the unwelded material. If a tapered weld 1s used J the etfect of the weld is decreased. Table C4.4 shows the allowable stresses in tension to use when tension loads are carried.

Solution: The ccnpreeatve (M.S.) will be determined first. As the Simplest solution, we can use the column curves in Fig. C4.5. For a length ot 30 and C = 1 J trom the upper right chart we project upward to the intersection with the 1-1/2 diameter tube and then horizontally to the lett hand scale to read the column strength or 14800 Ibs. which we will call the allowable tailing Pa Pa 14800 M.S. = P - 1 I4500 - 1 = .02 The tube strength could also be tound by using Fig. C4.2 as tallows: L' : L/V-C- : 30/v-r- : 30 L'/p : 301.5102 : 58.7. p 15 round rrom Table C4.3 as well as the tube area 0.2628 sq. in. Using 58.7 tor L'/p on lower scale and projecting upward to the Ftu = 95000 curve, which is the lower curve, and then horizontally to left hand scale we read Fe 56500 psr ,

In short cOlumns, the prunary column failing stress may be greater than the local crippling strength ot the tube adjacent to the weld at the end of the tube. This local tail1ng stress Jue to welding 1s reterred to as the weld cut-otf stress and the column compressive stress F c should not exceed this value. This cut-off weld stress is shown by the horizontal Itnes in Fig. C4.2 and C4.5.

14850

Whence, Pa = FcA lb.

= 56500 x

.2628

The solution obViously could be made by SUbstituting in the short column equation for steel haVing Ftu = 95OO0 J or

C4.6
"0

STRENGTH & DESIGN OF ROUND, STREAMLINE, OVAL Ai"ID SQUARE TUBING IN TENSION, COMPRESSION, BENDING, TORSION AND COMBINED LOADlliGS.

... \
II

-I
1\ FlU' ISO k$1
rcce at Weld Cutoffs

""
160
150

Heot Treated After Weldin Heat Treated Cutoff,ksi Reference 10 Flu,ksi

i \.,
\

180 150 125

Normalized Cutoff Meattreofed ofter welding
Ftu~ISOksi

144 120

REF. 1

95
76.6 67.5

rI-

100

""
1

1\

95

110

-.... ...

\~
1\

.0
____ Unwelded COlulTV'l ollowobles
tu~125 ksi

'0

;;;
"
e E

eo

\1
I" - \ r e l d cutatf treot 4130

a 0

~u

"

JO
20

cne-

heat

,,\ {
'\

, Ftys7S kSl

~,

20

.,

Ftu=95 kSl

Fig. C4.3 2024-1'3 and 6081-T6 round aluminum alloy tubing.

'\
0

-,
<,

....
I60 I

20

60

80

I00

I20

4 10

eo

'00

220

L'Ip Fig. C4.2 Allowable column stress for heat-treated. alloy-steel round tubing.

50
~

.5 40

in
o

~ 30 _ 0 96 Fey

c E .2 20
U'

VV ~V
.
60

Euler curve, E = 6.5 .lIOe psi I---AZ80AT5 and ZK60A-T5 e.lfruded shapes (except hollow) AZ80AF e.ltruded sncpes (except hollow I

I
I

I I
I'
I

I
I

+;

.I' ~~Round tubes,

Fey = 10 ksi

u.."
10

Fey

I
20
40

"""'~ I I
80

i
140

!
I

100

120

180

LIP
Fig. C4.4 Allowable column stresses for magnesium-alloy columns.

.
o

10

~
;;i

'" 1;

."'''""
~
(.J

z 4
2

" 6
(:=2
C=l

"' ~ is 20

.'"
0

20

40

6{)

80

~ 16

COLUMN LENGTH - INCHES

[g

g 12
e z

. .'" ."' ''""

z
,.J

38

~ 32

8
0

..

'" 28

" 0
(.J

:Ii

C=2

C=1 0

20

40

80

80

COLUMN LENGTH - INCHES

0 50 " '"

i'l z
0

.
"' Z
Q

.'" .'" '" .'" "'

0
,.J
(.J

[g

~ 24
20

o 18 z

40

"' " 30 g

<

" 0

:Ii

12

8 4

. ."'
(.J

.". 'c z 20 ''" "

z 10
0 0
40 80 120

C=2 C=l

20

40

60

80

100

COLUMN LENGTH - INCHES

" 6
C=2
C=l

:Ii

Fig. C4.5
STRENGTH OF CHROME MOLY ROUND STEEL TUBES

160

FTY '" 75, 000 PSI

FTU 95, 000 PSI

COLUMN LENGTH - mCHES

C4.8

STRENGTH & DESIGN OF ROUND, STREAMLmE, OVAL AND SQUARE rusmo IN TENSION, COMPRESSION, BENDINCi, TORSION AND COMBlliED LOADINGS.

7000

.r .'--~-I'

'Ftt;;;q4;6--:-

~
I

S 6000 ....

I:r.l

5000 ..:::.t'

' !'f ' ~~CiO~'~'?~.~'Ci"'~'irEi'~~!-:-' ;iil;~ii~~~~~Ft .:

t':: ...

.... --. 20~TJ ~.T!JBn-:::

--l:;~;~jJLI~~:~L
'-'::-::1- -

:;--.r-ru:'" a

~:.= ~"! Uop& PIlI

>--

:':':';':~.~--"';~:lOi.:ii.hti~;:.-

'" 4000 < '" ~ 3000 '"

0:4 .J

~_.

. _.=-~~:l::::~=::.~r~~-:~~ .-':~~:~_"';-:_~~~~f=;!~ IlIiz;; .

~-':::""'l.":- -:-...:,o:.-:~_--:

':':_:i:::;:~':;:::.t.r:~;

*'::::::::; ';::

;:::~?t':;-:~l::~:-4~J

~1000

~: ~;:~ ~:i~~-~~~-~5 :.::.-:::::'=.~f~~ ~+-:-=-"':-~::t;~:- e,:=...-; -=-t:=:J

'"

-=::::~I
60

-_~~ __ .~. ::-:::":'---:L:=~:~!=~:f~NW-~*~

120304050

o
35000

20

40

60

80

100

120

COLUMN LENGTH U'lCHES (C .. 1)

COLUMN LENGTH - INCHES (C '" Xl

14000

S 12000

..
Q

1 ..
,
Q

30000

..
.J .J

~'OOOO
8000

~ .J
< '"

25000

~ 20000

~ '" < '"

'"

i...
'" '"

8000

..
.J

~ 15000
~
'0000 5000

<

4000

;:
.J

2000
0 35000
~

20 40 60 80 100 120 COLUMN LENGTH - INCHES (C'" 1)

~t:.~:~

.''""

30000

~ ~ 20000 ~
; 15000 '

25000

.;-wi::~-F~*~
~. ~~~~g~_::

t:J;=IiG'E~" .:.'

..
Eo<

'" <

I
10000

:l

5000

lii" - c:
: :.1__

20

40

60

80

100

120

COLUMN LENGTH INCHES {C .. 1}

,--.-.

,~""'--._;

--

---~ :-::--._---:--~

- ..:......_-

~ -~~ ~~--:::-:-~

-= ---;-=:- -=--"-------

Wlltght LIt/iOO i_. Dla.

lla

.P
.0.10 .0781 .1231 .120' .116. .167:1 .11141 .1618 .2113 .2090
.204"

I
.000103
.OOOU2

I/Y

D/'
11.3. '.33 13.3. 10.12 1.85 11.0 14.2' 10.20 22.30 11.85 12.77 10.11 28.80 21 ....a 16.30
12'"

StOilI

1/' 3/'

.022 .028

.OU1. 01953 .03063 .03731 .08018 .04.162 .0:1.113 .0""3 .05252

.oooan
.000878 .003...6 .003tU ,003836 .004 ...1 .006659

1/'

.... ....
.02' .035 0.0

... ...
Dural

Wilight Lb/IOO iu. Dia.

/'

Go
.035 .0.1 .058

.P

I
OU.5 .11120 .1367'1

I/Y
.0113 .12500 .14511 .16168 .200110 .14288 .18896 .1851'" .2301 .2588 .1824 2134 .236 .29:13 ,33211 .2287 .2655 ,2960 .3888 ....16 .215 .3233 .3698 .4501 .50" .3868 ....305 15398 .810" .7ISII .41580 .11077 .837. .7211 .1808 .6114 .7t36 .8ua 1.0"11 14

D/.
53 0 38.25 32.30 2'.80 22.'0 0&0.80 30&.0&6 30.75 2.... 10 21.011 0&5.80 31,'0 34.80 27,15 23.70 61.00 43.10 311 30.10 a8.30 68.10 47.40 42.30 33.111 :18.85 n.70 4'.20 36.16 31.61 25.00 66.io 60.00 38.16 34. aD 27.10

etl 5.73 7.811 1.38 10.47 13.25 10.03* 11.11 14.18 18.11 '.5' 1l.SO 12.84 1'.01 18.22 10.88 12.60 14.0._ 11.86 aD. 3.. 11.71 13 0 lIS.IO. 19.70. 2:1.48
18.18

Dural

1 1/8

.000442 .000548 .000882 .001180 .001390 .00178' .002346 .002833 .003704, .004185 .OOCl ..5 .00503. .008161 .001601 .001278 .006S81 .008181 .010882 .01248" .013853 .012388 018594 .019111 .020870 .01782 .0:u.02 .02176 .03062 .oa481 .03339 .03887 .G4280 .03309 .0449a .011213 .06113 .04324 .058111 .08841 .076:18 .08305 .05528 .015..0 .08118 .0.707 .119811 .08938 .09...78 .11048 .12230 111138

'81.08 1.43 1.17 1.0&61.'6 1.48 1 2.51 21 3

.eas

. DOn.....
.007603 .009065 .011852 .0134.211 .011053 .OU.21 .01718a .02:031 .oa201 .016289 .018853 .024&1 .02853 .03UI .02"n .0331 .03822 .OC193 .03188 .04270 .04833 .01421 .03.... .063...2 .06187 .08818 .04814 .08534 .011183 .08341 .06181 .07141 .011:11 .1007. .124.07 .08803 .0.2:19 .lQ801 .11948 .147151 .079a7 .10832 .1:162'" .13.71 .11.:11'

.... .7'
.M
1.05

.31 .3. .'1

.eee

.083 .0.1 .06a .0611 .0811

.20:13 .2811 .3311 3616 .4873 .3003 .353' .3961 .....18 .68611

."61 .'~1
.&342 .8800 .8a61 .614.5 783 .7'1.3 .17153 172a .7887 .762 .8887 .8636 .8813 .U60 .8&0' .8661 .06211
....34

501

.....

.nu

.18787

2 1/4

.eaa

'j' 'j'
7/.

.n.....

.14288 .18816 .U5a .2300 .2588 .2052 .2.01 .2685 .3322 .3741 .2834 .3318 .3681 .4801 6191 3783 .4446

....
'.00 5.06
5.'18

.... ....

b
.'"
eZ
;> Z ;>

3.311 3.7" . 73 3.5S-

.02' .035 .04' .058 ,028 ,0311 .OU .0118 .065 .038 .035 .0". .058 .066 ,036 ,0... ,068 .085 .035 .049 .058 .085 .036 ,049 .058 .066 ,035 .04.9 .058 ,065 .035 .04. .0118 .0811 .0.3 .0311 .0.9
,058

.08.81
.08867 .10331 .08351 .07882 .107.1 .1260. .13961 .07"61 .0.238 .12716 14881 .18Ml .10811 .U640 .17164 .190.3 .11.85 .16564 .184.2 .21860 .13360 .1848' .2172 . 2420 .1473 .2041 .24.00 .2875 .1'11 .2234 2628 .2.30 .36.5

.2018 .25U .2631 .2485 aUII .a433 .2988 .2913 .a925 .28' .28615 .34l" .3387 .3337 .3314 .3858 .3808 .3780 .31156 .4287 .250 .42lf .411' .4738 4691 .4681 .4838 .6181 .5132 .1510a .11071 .6018 .5622 .111175 .1164'" .51120

11.63 31.:13 26.00 11.86 16.10 IS." 38.68 20.40 11.20

1,80 2.23 3.06 1.0' ' l.a. 3.51 1.42 3.'6 2.11 2.82 4 3.60 1.2' 4.2a 1.11 88 1."

... ....
.7'

."

.e8

a 1/3

....
.0"1

.048 .051 .065 .0.3 .095

. s....
.4... 21 .5851 ....32

.3388

3."3 ".011. 4.52: 5.72 8.61 3.83 4.50 11,03* 8.38 1.27 4.20 4 8 6.55-

>oj

.oal

.3773 .4450

.4.7:1
.6302 .7178 .4168 4901S 6"'83 .8964 .7124 .6361 .19'3 .7808 .8870 1.0U7 .5818 .8504 .la51 .'418 1.1100 .1014 .tilt 1.0182 l.a742
.16:15

.083 Ot5

a 3/4

16.S8

....
1.41

3.01. 1.01. 1.48 4.15

I.U

1.93

3 1/...

1 l/B

32.10 a2 1S 11.40 11.30 3!i.10 211.60 19.2a 31.26 28.05 23.70 al.15 4a.1O SO.80 alS.86 13.01 18.08 46.40 33.16 a8.00 :16.00 181 60.00 35.70 30.20 28.80 111.10

3.40 1.2:1 4.8' 1.88 6.11 1.31 a.lIo '.14

.... ....
.095 .0.$ .120

.04' .058 .0811

.......
393

.'18 .8.01
.11802 .6451 .8091 .'1118 1.1278 .1410 .8261 1.0381 1.1721 1 ....472 1.0348 1.3012 1.473. 1.8220 1.a171 1.8080 1.8228 2.21185 I.U67 1.8517 2.2228 a.711U 3.873a 1.0282 1.0a13

..au

7."

" .. ., " '" '., " " ..

w

;> '" n Z
g

,..
0

8~O3

.058 .0811

1.0403 1.0380 1.031a 1.021. 1.0181 1.1:187 1.1283 1.laOI 1.1180 1.1074 1.2141 1.20U I,aO3 1,1868 1.3031 1.2.88 l.2.n 1. a841 1.3116 1.3852 1.3810 1.312i1l 1.4551 1.63'. 1.'143

18.6 21.66. 24.66 30.18 18 7 18.40 23.3" 28.18 33.43 18,81:1 26,20 :18.70 3.,00 21.30 :17. 30.84. 3'.70 22,75 ai, 16 33.n 41 ...0_ 48.10 '0.40* 18.25_

5.42 6.08 1.708.78 11.00

Ii. 8.

1 1/4

n.n

....

3.7' 1.35 6.23 17 6.111 a.20 2.46 31/2 1.41 2.01 a.43 2.70

1 3/'

17 6.78* 8.80 1.68 4.118 '.32 7.ta 30 10.47

....
.083 .095 .120 .086 .083 .095 .120

.058 .066 .083 .095 .120

8.11 1.35* 8.62 11 5

63.80
42.20 38.85

7." 9.01
10. as. la.89

21.15

1 1/2

....
2 ' 3.14

3 3/4

1."

2. ae-

.168:1 1.0908 1.3885 .8035 1.0214 1.1615 1 ....21 1.1321 a.1281 a

U

.151.
.8722 1.2036 .1779 .119' 1.1114 1.377' 1.1408 2.2347 2 583

67 0 46.20 38.60 31.2i6 81.60 ...a.20 42.10 33.33 31.16 28.80 26.a5

o.

11. Q.tIt 13.1.2 8,12 10.3a 11.78 14 Olt 17.6S1t

'.81

7.80

- ... '" '" ., "" " '" " ., '" " '"
0 '.", ...
>oj

'" 0 ...
0
Z

r>oj

>oj

r-

'" z >oj
d
>oj

1 5/8

.'"
1 3/4

.065

.1741 .3-&26 .a8511 .3188 ...021

.11185 .2618 .3083 .3441 .43..7

.un
.606D

4.96 1.77 ' 7 a.48 .ot 2.81 3.:13 '.06 11 ...0 08 1.n 6.32 7.42* 2.611 8.73* 3.12 3.48 '.n 1:1.33 4.40

1/... 4 1/2

.065 .083 .081S .120

.'",
'"

. 1"

.031 .049 .018 .0611 .083

.15'
.1"

n.lla27.201t

.8017 .6986 .11982 .1801

... 3/..

IIUIIJ)lIlD ",81.0

C4. 10

STRENGTH &: DESIGN OF ROUND, STREAMLINE, OVAL A..~ SQUARE TUBING lli TENSION COMPRESSION BENDllia TORSION AND COMBINED LOADINGS

Fe

=79500

- 51.9 (L' /p)."

L' = L/ve = 40/'1'2 = 28.4

= 79500 - 51.9 (58.7)"

= 56500 psi

The short column equation applies since, as shown in Table C4.1, the transitional L'/p is 91 and the value for our tube is 58.7. TenSile Strength

In Fig. C4.5 we can use L ::: 40 and c = 2 scale at bottom of chart or use c 1 scale and L' = 28.4. Reading the chart we obtain Pa ::: 9200 lbs. Thus the c = 2 fixity increased the strength of the tube trom 6000 to 9200.

Case 4. Since the tube is welded, the tube material adjacent to the weld is weakened. The weld correction values are given in Table C4.4. We will assume a weld other ttan tapered. Let Fa : allowable or failing tensile strength of tUbe.

Same as Case 3 but heat treated to Ftu ::: 150,000 after welding.
L'/p = 28.4/.425

= 66.8

From Fig. C4.2 USing 150,000 curve, we read Fc = 63000, whence Pa = FcA

Pa

Ftu (weld factor) (area at tube) lbs. 1 =

= 53000

x .1849

= 11650 lb.

0.13, thus compression 1s critical.

PROBLErl 2

= 95000 x .841 x .2628 = 21000 !I.S. = (Pa / p )- l = (21000/18500) -

Case 1.

Tube size 1-1/4 - .049, L = 40 in. c = 1 Material: Alloy steel, Ftu = 95000 Find ulttmate compressive load it will caITy From Fig. C4.5, Pa = 6000 lbs.
It tube was heat treated to Ft u = 150,000, what compressive load would it carry.

In this case heat treating produced additional strength, whereas in Case 2 it did not. The reason for this Is that failure occurs in the inelastic stress range and heat treating raises the material properties in the inelastic range. The end fiXity changed the column tram a so-called long column to a short column.
The strength could be found also by substituting in the short column equal for 150,000 steel as given in Table C4.1, Fc = 145000 - 18.36 (L'/p)

Solution: Case 2.

= 145OCO
PROSLEM 3 Case 1.

- 18.36 (66.8)' = 63000 psi.

Solution: Fig. C4.5 cannot be used since Ft u 2 150,000, thus we will use Fig. C4.2. L' = Live = 40NT = 4. From Table C4.3, p = .425 and area (A) .1849.

Tube size 2 - .065, L = 24, c = 1.5 Material Ftu = 95000. Welded at ends. Ultimate deSign load' 25000 lbs. What is M.S. L' = L/ve =
24/~

L'/p 40/.425 94. From Fig. C4.2, using the 150,000 curve, we find Fe = 32500. Then Pa FeA 32500 x .1849 = 6000 lb. Thus heat treating the tube rrcm 95000 to 150,000 for Ftu did not increase the column strength. For a L' Jp =a 94, it is a long column and failure is elastic and E is constant.

Solution:

= 19.7

The strength could also be calculated by Euler's equation from Table C4.1. Fe

From Fig. C4.5 for L = 19.7 on c = 1 scale, we project upward to the 2 inch tube and note that it intersects the horizontal weld cut-off line which gives an allowable column load at lett scale of Pa = 26700 lb. Failure in this case is local cripplIng adjacent to welds at the tube ends.
M.S. Case 2.

=286,000,000/(L'/p)'
= 286 ,000,000/(94) a
-

= Pa/p = 26700/25000

- 1

= .[17

32500 pe i , the

same as preViously calculated.-----Case 3. Same as Case 1, but assume tube Is welded to several other tUbes at its end and that the end fixity developed is c 2.

Assume tUbe 1s heat treated to Ftu = 125000 atter welding. wbat Is tube strength. L'/p = 19.7/.6845 = 28.8

USing Fig. C4.2 with L'/p 28.8 and projecting up to 125000 curve, we again note that horizontal weld cut-oft line is intersected

<,

--

,~,>,. ',~-'.

t ~:: -,'~~'.'~'>.~"_\~":~~~~~~'~'~:.>~.~~'~~'_'>.~~"

;,:.':;,~.,,~ ~:">~::';'~_:~~,~~,:"'f ':-~~_. :~ ~ .. _,:~~::~,,~f~:~/J:~~(:

------- ..

~ -

. -~--

--------OF FLIGHT VEHICLE STRUCTURES C4.1l

ANALYSIS AND DESIGN

giving Fe = 95000, whence Pa = 95000 x .3951 = 37500 Iba ,

If the tube had not been welded at ends the dashed part ot the column curve could have been used, thus giving additional strength.
PR08LEM 4

The results show that 1-1/4 - .035 is the lightest. Since there is danger in welding .035 thickness to the other heavier tube gauges particularly the engine ~ount ring which is usually relatively heavy tor this size engine, a ~ln~um tUbe thickness or .049 will be used, hence the 1-1/8 - .049 tube will be selected. Consider Member (2)

Fig. C4.7 shows a steel tubular engine mount structure for a 1050 H.P. radial engine. The ultimate design tension and compressive load in each member as determined tram a stress analysts for the various flying and landing conditions are shown in ( ) adjacent to each member. The true length L ot each member Is also shown. Using chrome-moly steel tUbes, Ftu = 95000, select tube sizes for the given loads. It Is cammon practice to assume the column end fixity c 1 tor engine mount members, since the mount Is subjected to considerable vibration and the true rigidity given by the engine mount ring is difficult to accurately determine.

Design loads 11650 tension and 4250 compression. Since the tension load appears critical, the tUbe will be designed for the tension load and then checked for the compressive load. The Ftu of the material equals 95000 psi. Since the engine mount in a welded structure, the strength at the tube adjacent to the end welds must be reduced to .841 x 95000 ~ 80000 psi (see Table C4.4). Hence tUbe area required = 11650/80000 0.146 aq, in. From. Table C4.3, which gives the section properties of round tUbes, we select the following sizes: 1- .049, Area ... .145, 11.S (.146/.146) -1 1- .058, Area
1-li8 1-3/8
035,
.049, Area

=
=0

= .172, M.S. = (.172/.146) -1 ....19 = .166, M.S 14 Area = .147, M.S = 0

To obtain a reasonable margin at safety, the 1-1/8 - .049 will be selected.

Front View

Fig. C4.7

Consider member (3). Ultimate deSign load = - 9250. Referring to the column charts or Fig. C4.5, we find for C :: 1 and L :: 31.4 the fOllowing tUbe sizes for a strength near - 9250 Iba .
1-1/2- .035, Pa

Many structural deSigners prefer to have large margins of safety on engine nount members as considerable trouble has been encountered in the fanura or cracking at engine mount members. The strength or the 1-1/8 - .049 tube as a column for length = 31.4 and c := 1 equaLs -6700 lbs. tram Fig. C4.5 which gives a margin of sarety or (6700/4250) - 1 = .58 on the maximum compressive load. The student should select a tUbe size for member (l).
C4.14 illustrative Problems USing Aluminum Alloy and Magnesium Round Tubes as Columns and
Tension Members.

= 8850

(weak). (Note-Pa = allowable load).

1-3/8 - .049, Pa - - 10350,weight::a5.78 lb./lOO", M.S. = (10350/9250) - 1 = .12 1-1/4- .058, Pa = -10000, weight = 6.15, M.S. = 10000/9250) - 1 = .08 Thus use 1-3/8 - .049 since it is the lightest as well as the strongest. Consider member (4), Load c 1

=- 5470,

= 30,

Fr'om Fig. C4.5: 1-1/8- .049, Pa = -7100, 'Nt.= 4.68, M.S.: .30 1-1/4- .035, Fa 1"
'=

In general alloy steel round tUbes must be heat treated to around 180,000 to 200,000 ulttmate tension strength before they can campare favorably with aluminum round tubes on a material weight basis. However, aluminum alloy as used for tubes cannot be welded satisfactorily and ti:us in a truss structure the end connections involving riveted and bolted connections add weight and design difficulties as compared to welded connections in steel trusses.

-6500, 'Nt.= 3.78, M.S.= .19

- .058, Fa:: -6000, wt . 4.86, M.S.= .10

C4.12

STRENGTH &: DESIGN OF ROUND, STREAMLmE, OVAL A..lIffi SQUARE TUBING IN TENSION, COMPRESSION, BENDlliG, TORSION AND COMBINED LOADlliGS. C4.15 Strength of Streamline Tubing.

PROBLEII 1 Case 1. Tube size 1 - .049 round. L = 24, c = 1. Material 2024-T3 Find failing compressive load.

solution: The column curves in Fig. C4.6 are slightly conservative because the equation used was slightly different tram the equation now epec i ri ec In (Ref. 1). Use L 24, we read for 1-049 tUbe a tailing load ot 2800 lb. C4..3. As a second solution, we will use Fig. L' = L/ve = 24./v'"""L 24.. L' /P = 241.3367 = 71.3. From Fig. C4.3, we read Fc 20000. Then Pa FcA 20000 x .1464. = 2930 lb.

If a round tUbe is exposed to the airstream, the air drag 15 about 15 times greater than if it were given a streamlined shape, thus streamline tubes are used when the member is exposed to the airstream. Streamline tubes are drawn from ro~~d tubes. In designating a streamline tUbe, the round tube from which it was made is used and then the fineness ratio is also given. The fineness ratio is the ratio LID, which dimensions are shown in Fig. C4.8. The ~ost common fineness ratio used is 2.5 to 1. Table C4.4 shows the section properties of streamline tUbing having a fineness ratio of 2.5 to 1. Figs. C4.9 and C4.l0 give curves for finding the column failing stress F c and the local crushing stress Fcc.
Fig. C4.8 B '" . 944L R '" 19D L :: 2. 5D D '" . ~714d d =- Equiv. Round Dia.

The answer could be obtained by substitUting in equation C4.9, 4.31 L'/p Fc = 50600 = 50600 4.31 (71.3) = 19900 Pa 19900 x 1464. = ~ lb. A column may also tail by local crushing or crippling ot the tube wall, thus the crushing stress Fcc should be determined to see it it is less than the pr~ry bending tailing stress tor the column. For our tube the diameter over thickness ratio D/t 1.0/.04.9 20.4.0. Values or D/t are given in Table C4.3.

PROBLEM 1.

Case 1.

A streamline tube made tram a basic round tube of 2-1/2 - .065 size has a fineness ratio at 2.5 to 1. The length L is 30 in. T~~e c 1. ~~terial is alloy steel Fty = 75000. Find the ultimate compressive load the member will carry. Solution: From Table C4.4 tor 2-1/2 - .065 size we find the following section propertles:Area (A) .4972, P (major axiS) .5137 In. Then L' = L/V-C- 30/y-r- 30, and L'/p 30/.5137 = S8.5 D/t value for tube:: 2.5/.065 :: 38.5

Referring to the small chart in the upper right hand corner ot Fig. C4.3, we tind for a D/t ot 20.4. that Fcc = 4.7500 pSi. Since this stress is greater than the bending failing column stress at 20,000, it is not critical. Case 2. same as Case 1 but use c = 1.5 and change material to 6061-T6 aluminum alloy.

From Fig. C4.9 tor L'/p = 58.5 and D/t = 38.5, we read Fe = 46500 psi. For D/t = 38.5 and reading from small chart in upper right hand corner at Fig. C4.9, we read Fcc = 66500. L' = Llve= 24./v'"T.'S = 19.7, L' /p = 19.7/.3367 Thus Fc Is critical and Pa = 46500 x .4972 = 23000 lb .: 58.5 Case 2. Same as Case 1 but change material From Fig. C4..3, Fc = 22S00 to 2024-T6 aluminum alloy. Whence Pa = 22500 x .14$4 :: 3300 lb. For this material we use Fig. C4.l0. For L'/p 58.5, we read Fc = 26000 psi. Fcc tor D/t = 20.4. from Fig. C4..3 38500 (not For D/t = 38.5, we read Fcc = 37500 (not crttical). critical). Thus Pa = 26000 x .4972 = 12900 lb. Case 3. Same as Case 2 but change material to _ C4. 16 strength of Oval and Square Shaped Tubes magnesium alloy, Fcy :: 10,000. in Compression.

For L'/p = 58.5 and using lower curve on Fig. C4.4, we read Fe = 7600. Then Pa = FcA -= 7600 x .1464. = 1110 lb.

Tables C4.6 and C4.7 give the section properties for square and oval Shaped tubes respectively. For the design of these Shaped

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C4.13

Streamline TABLE
1~

C4.4 SECTION PROPERTIES OF STREAMLINE TUBING (Fineness Ratio 2. 5 to 1)

Mat<><

,
Minot

Minot
J~

~"i

Wail
Dec'! Ca,.

A...

RO""G

0.0

Mo;ot

Mil\Ot

"0

.O\S 049

20 la 10 18 11

1.0114

4286

07116
.1079

167}
}668

oem
0021 0018 OO}1 0041

14116

OIOl

.14lS

1>

on

i.isco
I.H8S

osa

049

sooc

"''' 1272
.1489 .1Oti! 1116

.3140 4113

, .cse
~, ~,

ms

1Yi .O}S 20

."
20

ta

1.1170

". " "

lJ.(

.",
714l

0}5
~, ~,

20

"

2165

... "" '''' " ..

ien

.5061

0111 .11S; 0148 1703 0168 .1671

0151 .10211

"'"
0118 0181

con

'50' "" ,W,

'm!
01)9 0278

0195

""

Hll
nIS

"'.
~, ~,

IS

s.rm

\.4185

H9S

.~Bl5

.7]~9

"'" .,,, ""

.m
.."
618~

00" 00",
~,

0100 0118 0196 0253 02n 0320

1912 1919 1291 2140 1105 1181

.con
OOM OIOl

coer

.01S?

.OJl9

0341
~,

usn

Hi

Hi

Hi

" ". " " 015 ". " " on " .ota " "" " .oe " " 0)5
20
~, ~, ~,

.18049

1111 .14ll)
U'\04}

'''' "" "

78S7

,~,

aacc
.167S

20

1.0m

8S71

1611 .1617

U"

19H

50"
1.191)

1.1106

". " " " .m "" "

~,

.oe

20

""

.1416

Hi

.on
~,

20

"

" ...,,, ,.,

1MB

1.]600 1.<:'000

Iii

"" ."" ". .m ""

.oss
~,

rr ". " "" " "" " "" " ms " '" "" "
10
~,

.4341 49}9 1.S18S 1.0714

. ". , .., ""

1B~~

''''

1.08}
1.361
1.~51

1.170
1.478 1.619

"" "" .m '''" "" "" "" ",. "" " "'" on8 ~" .", "50 ",U "" ''''' "' "" 50
7S93
0239

. """ .. " .""

8158

00.. 0116
OIl)

0246 ollS
Ol11
0406 0400
.~I

2559 1509
14TI 241\

.o;SS
.0618

.US

0145
.0118

,." " ''' wn "" ''"' ,." .w"

.0714

"'"

.,.. ocr .,,, .,,, "'" .'" "'''

0151 0145

" " "" " "" u a

". " "
"" "
~,

,m ."',
"50
"51
7178

..,
"

MalO'" M-o.,ot

,
Minot

Miaot

.10t8 1611

.nu
.1815 1.414 1364

a ..

.oco

1J(

'''' 049

18

3.706S 1.)714

.490'

"'''
,~,

4S73

.oe n . .nc a
.083

.~,

.1914
4.04~~

.... "" ,L"" ""

}.l11 1.0}7

.m

.J'"

4'"5

.'51'\04

0300 .UlS

.01P .0181 .019$

.1m
1745

.0'50'1

..719 i096 .)014

.cns
.OSSS
.0611

.048} .304S

.1989

.0316

814S

..

.on7 .29n
})6S 3314 0448 3181 0118 m6 088} .3191 0976 .3149 .0498 0611
)Ii}}

0116 051}

.0410

"" "'" "''' om "'" "'" ''''' ''''' "''' .nu "", om """ ''''' use "'" ",n m' ".,
~n

~" oeec "" ,.,. "'" '''' cts "" .S581

.08IS
6161 .61)1 .6108

'" " " .no a "" " ", ,,,

,

11

1.1141

"'" "''' 1.08S1

\.}959 4.3818 1.8S71 .5816
S1~8

24n "" ,.... "M "" '.m "" "''' "" ,m ''''
1.~86

1'9) 1'765

2311

.1l14 124S
1732

.,,, ..... "'" ."

.1\6~9

"'"

.nc
~,

os, " "" " m

Hi .065

'" "" " " a .nc "" ". '" .oa

16

" a "

,.

"'" ,,,
1.1800

3.691 4.146

''''

.)119 .3925 .4814 .2961

Lm 2.111 50m
},Wl 4,011 H6} 2.38'
3-019

1.5186

.61~1

"" "50 "" 50"

16

4.noo

2.lXlOO

.7014
89LO 1,0162. 1.1141

1.64l4 '.0570 1.1418

m. ,."" """
USB
l.I,1
}.708 4.6S1

).4S'
4.})1

.~1}4

.m, "

6)74

.",

.cs

.110

'm
OHI

0511

.08106 1042
.11'\04

ens

}583 .3551 lSlS 3461

HI8

Ion
1!11

rt

.HI1

,..e

O}S

2)4

.osa .oe .uc u

~,

049

'" "

" " " " " " " " "

10 18

.-1613 5111 1.6970 1.142.8

1161

"'"
J-O}4} I.l.8S7

"

..., .,,'"
.l388

,." ""

lH9 j9H

J""

.6431

"" '''" "' " ~'" L"" "'" urn "" ~" r.sn ''"' "'" 'm "''' "'"
.0411
1.116
1.1~7

6818

0515 .}902 .0719 }SSI 0900 lBI5

0993

\.jBiI

.1114 .l119

.J~

.1HS .l6S7

7]4S

0)16 OSlO

1.203
I.Hl 1.699

.10}4 1141 119B .1SS6

.1144

"''' "'"

.Hl'

.}7+4

nl1 , "'" ."''' ,.,,, "" ua "''' ,.,,, '''' ""' .mll "''' "" "" "" 50" "" aeea "" nc ""
1~91

"''' "'50 Ill}

'M'

.176S
1191

.14U

..

116} .1116

,"

n " a

L""" 1.168S
1.7641
1.0114

!
16 14

"" "" ,m

.os}

O6S

"'''

.116} .1101

.nc

.095
.188

I}
11

1.16n

.7061

.156

.168S
.2186 .2719

.1296

.lSlS

.TIll
.7690

7614

.1076

8113

,~

.1731

.I!m

!2111 8117 .SIS6

.8115

1.l~1 \.]~S

.1l19
.2020 2411

.14SO

".,

.," .'M .cs '"' '''' .," , us

"" ."" .uc

.06'

i:
n
a ,. ,.
H

H.

L..,., L...,

1.14S1

'.1.~1l LUllS

L"'" 1.1<\00

.'"

,."'" 1.J930
1.;141
2.1111
1.'~403

1."10

M683 1.5113 USH

1.6~ll

.<S"

.1111 H70 .42"

LSI7 1.911

1161

"''''

11116 1.7}0

.1198 .15)0

"" ,,"

.l8l'

,." .,."

50" "" "" "" """

.9ll)

,os. sm
9Jt5

. ,.'"
51

" ,.

1.9411

30

I
r-,
20 40 so 80 00

60
50

40

~50"

I I i" ,OJ' I

30
OtBoSlC rour'lCl tube dlornefe~ II! I
L'/p

020~60ao

00

c
. ," Fig. C4.'9

=~

Fig. C4.10 Streamline 2024-T3 Tubing

10 _

o~
o
20

ALLOW ABLE COLUMN AND CRUSHING STRESSES . CHROME MOLYBDENUM STREAMLINE TUBING. F ty = 75000 PSI

~,

.
40 60 80 Lip
100 120

14<l

STRENGTH & DESIGN OF ROUND, STREAMLiNE, OVAL AND SQUARE TUBING C4.14
iN TENSION, COMPRESSION, BENDING, TORSION AND COMBffiED LOADINGS.

Squares

TABLE C4.6 SECTION PROPERTIES OF STREAMLINE TUBING

, -

~.

...

'" '" '"

~,

.on
,~
~,

'" "

,~

..

" " " " w " " " " zc

.. . ". "
u " zo

re

~" ~u

"

~~

= om coos

oon
oo~
~,

," '"

.035

oou

~w

~"
~,

"''' "'''' "'''

.~

.,"" "'"
.1071
.12~1

,."
~

oon "'''' oon

m",

.,,..
.crJ"61

'H

. " ." " '" " " ." " " ." "
n

" " " "

.,,'"
.,6ft
.UIl

""
"'" " "
,,~ ,,~

,,..
nos
'17'

00"
con =, om
OWl

"'''
~,

c.. .ce,
.081

".

cots
al41

",. "'"
,,~

.c-s

.""
.. "
.izo

O'l,

ra '" " "

1}

6549

.
1}

t.cu

9O'l,

oes

OW,

0881

.1051

1.195

Ute
1.i17 2.136 1.065 1.25' 1.402 1.774
2.019

1436 .1608

"OJ

.1203 .10101

,,.,

.2586

"'" "'.

1915

.oss

0119 01"

"" O'l'
""

'" " is
n is u

.
" ."
u ...
3592 .4236
.47)2

~"

"'" =.
.OU8
0129 .01:14
.0101 0125
~

''''' "", m"

.0161 011] 0)41
~~

2.484

.1451 .1691 .1877 .23U .2618 .3175

.1658 .1932

.1146
.2663

.2992

.3618
2557 .2838

.oe
.0~8

1,221

.5711

''''' ''''

.1IOS

aa

.1021 .1271
.~5

. . ""'".. .. "'" . "" ..""

.,Ifl

.7184

'"'' "
.m

"'"
~

.~

.0191
.02~'

"m

n
n

.. . " """
" .lJ36"
.)716
.14H
~.
,~,

.,m

"'"
un

noo

UI1

..3479 .7819 6717

,,~

"" "" ""

.~

'H

'" rr " " u " u '" " "

.1'119

. '"
\.l46 \.414

1.114 1.26]

""" "" "'" .,,'" ~" ~" "". """ m'"

m~

.0591

0124

.~~

.0271

"'"
~u

'''''
,m
m ..

... .,,' .ce,

.1~6

.nc

.oss

.ce,

.6827

.8~)O

"" ""

'''''

n'"
2~57

1.609 2.038 2321

2838

"'"

.i7W i769

.7744
7679

3H6 .3981
.48'9

.H16 .)931

7636
.7~48

1.0912
..-:l'1

i.sco 3.710

.6010

.6010
.2738

""
",n

.7422

'" ,"

. , ." " ... .., " .,," ..

.rac
.156
.058

is

1.377
l.624 1.816
2.)02 2623 3.281

.n41
.7716 .9652 1.2)72

.4778

eoco
6918

.O'l'

1}

,.., "'"
-nr
7981

"'''' '''''

3556 6149

,,'"

4.206

.5321

is

s ..
eeo

... " 2.022

1.$66

.,., w,

.8765

."

.8598
.~18

",. ,,.,

.7549

"" ~" m,

.nc
.H6 .iee
.0~8

.O'l,

..... .sez
.6896

'''' A(l35

...
9541

"" eses

.9757
.9678

"',
ssu

1}

.l014 11)8

"'.
~

.U,I

1415
.75'16

.
u

. "

.- ...
.~

"" "'" "" .,,n, "" .'"

.19Oll
~161

.... ....

.0178

~"

~"

~"

.4161

LIn

,m. "'"

.1438
1718
.lOO)
~

"n nse
,~,

"""

.1761

ml

"" "" '''' ,,.. ,

8116 1_178

"''' m,

" .... """"

.., .. " ,... "" '''' .," ... ."

ie
.6517

u I ... ...

1.0775 1.)831 1.0479

.O'l'

1}

..8)26

2.925 l-663 4:ro2 5.602 1.)191 2.229

U~

l.2255 1.4257
.5792
.i~59

'""

1.31S}

"'" "'"
.7806 .5497

.941l
.9301 1.0;63 1.(7)7
1.0671 1.0628

.g.t81

. '"
""

.,~

'h

1.1898 1.5291
l.32~S

3.227 4,CoI5 5.198

2.436

l.a724 1.3213
1.6~9

.m.

e.ace

1.1121

1.93)4

l.J2iS4

1.0541 1.0410 1.0296

.os, .os,

.7166

sesc
1.239<4
1.4032
Li325 2.179')
U~3

.1166

,=

,-

.n",

.l771

""

.,'" "" .rss

1}

.zrs

'"

... ...
ll.

.9102 1.0183

1.)011 1.67'\0

r.ssst

,....
4.426
6.798 7.847
~

''''' ).530

.m,
.9351

.6573

1.1550
1.4529
LiOO2

~:i~~ I 1.1544
1.1406 1.1381 1.1184

1.1730

I I

2.:lOlll

2.8873

1.9249

'j S~~ i~ ~~ ~.~ ~!! ~~~~~~~ ~~i ~~~~ ~i~~~ !~~= ~~~~i ~~s~ ~~~3~S ~~~!~5 S~~~~5

'" ..
0.

'"

N~ i~~~iiiili~~~~~~;~;~~a!5~~~~~~E~~a~;iii~~~i~~~z~E!~~~~!~~~5~

iJ

lIl"' o 0-"'::<'

.,

--

:;:<

"l-"'O.,

_o

-z

-::\l.

"''''':;1:---

" 'o "

o.z
0'"
0",

"0

z-

-" ....
"'<
;:;i~:3~= i;~:::~i "'> i-i ~~ ii ~~ ~~~ U~ ~E~i ~~~ h~ ~!~; s~S;!~ ~~;! ~:~ ii!~~ 5~i~!~ -. -.-. .. ~0I--=t---------,,.--,..---=---''-=--'''==='='':============='''-'=='-'4
V

.,;

00

- -- --- ----- --- -- -

'" '" '" < ..

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C4.1S

tubes the primary column strength can be found by uSing the curves in Figs. C4.9 and C4.l0. The crushing stress Fcc for oval shaped tubes can conser'ffitively be taken as that for stre~line tubes as given in FigS. C4.9 and C4.l0. For square tubes the local crushing stress Fcc can be taken as the crippling stress ot a flat plate. For this stress refer to the chapter which covers the buckling and crippling stress of flat plates with various Widths, thickness and boundary edge conditions.

C4.18 Problems Involvtng Bending Strength of Tubes.

PR08LEM 1 A 1-1/4 - .058 round tube is used as a Simply supported beam with the supports at the ends. The span or length or the beam is 24 inches. It carries a uniform distributed load w in pounds per inch. Find the value of w to cause the tube to tail in bending it the tube is made fram the following materials:- alloy steel Ftu = 95000, 606l-T6 aluminum alloy, and 6AL-4V titanium. Solution: The maximum bending moment occurs at the midpoint of the span and equals wL z/8 = w x 24 a/8 =-72 w in. lb. The section properties for the tube are D/t 21.55, and I/y .06187 obtained rrca Table C4.3.

ULTIMATE BENDING STRENGTH OF ROUND TUBES

C4. 17 Charts for Finding Modulus of Rupture Stress.

Chapter C3 was concerned with the theory and methods at determining the Ultimate yield and failing stress of a section in pure bending. It was concerned with finding a fictitiOUS stress Fb which, when SUbstituted in the well known beam formula M = FbI/c, would give the value of the bending moment which would cause failure. The same procedure as was used in Chapter C3 can be used to find the modulus of rupture stress (Fb) for round tUbes. However, since round tubes have been a standard. and available structural member for many years, much testing has been done on tubes and as a result rather complete deSign curves are available for finding the modulus of rupture in bending (Fb) for round tubes. Figs. C4.11 to C4.l4 inclusive give curves for tinding modulus or rupture F'b round tubes when fabricated from alloy steels, aluminum alloys, magnesIum alloys and. titanium alloy.
270
~
,
,

at rupture Fb is M:
w

The beam equation involving the modulus = FbI/y. Substituting 7~ tor M, we obtain:

=-72y

FbI

Consider the alloy steel, Ftu = 95000. From Fig. C4.l1 tor D/t = 21.55, we read Fb = 117000. Then w (117000 x .06187)/72 100.2 lb. per Incha Consider tube material as 6061-T6 which has a Ftu = 42000 pat ,

From Fig. C4.12 for D/t = 21.55, we read that Fb/Ftu 1.06. Thus Fb 4.:2000 x 1.06 = 44500.

260

,
i

,
!:
c.c,

, Flu 160000 PSI

,
,

Then w = (44500 x .06187)/72 Tube material 6AL-4V titanium.

=38.1.
-----

" Ii:
\

..; 250 a:
240

-,'
,
!

,
I
!'

a: 230 Ti, , "o

"

, ,

From Flg. C4.l4 for O/t = 21.55, we read Fb = 191600. Then w = (191600 x .06187)/72 = 118.7 Ibs. per in. PR08LEr! 2 A beam sunply supported at its endS has a span of 30 inches. The ultimate deSign load consists at two equal loads of 2000 lbs. each. The beam is symmetrically loaded with each load located 12 inches tram the ends.
I

3 2'0 ,
" g

220

,
, ,

!<:
i:'<
,'I>
;"-

~,~:

,I
I
,

,
,

;;",=::j'
i:;"

,
,

'!,,'f

,
!
, , ,

'~ "

200

i
, ,

190 i5 z w 180

,
,

, ,
I

r
~

!
,

"I

,
,
!

, ,

, ,
I
,I

'" u.~

170
160

""

,~

I , , .. ,-

Select the lightest round tube when made tram the following material: (1) alloy steel Ftu = 220000, (2) 7075-T6 aluminum alloy. compare the resulting tube weights. Solution: The maximum bending moment is constant and occurs between the load points. M

10 15 20 25 30 35 40 45 50

01'
Fig. C4. 14 Bending modulus of rupture for round 6A14V tubing. (Titanium)

= 2000 x

12

= 24000

in. lb.

,IL ,1\\
4

'~,\

u,

I. ',
2

,, ::,..
' ~

.: I

'-

i'-. -.
<,

r.....

o9
o8 o7

'.

---

--

-~

-- ..,--606l-T6
~14-T6ond

7075T6

400

2024-T3,T4 ......

"- ... 70

00

I
10
2D

'0

'"

011

'"

so

80

'50

Fig. C4.12 Bending modulus of rupture for aluminum-alloy round tublng.

'00

'a

200

Flu. 260 ksi Ft u, 240 hi

F,u, 220 hi
Flu. 200 k$i

2
'fA,

200

60

\\

.
~

Flu.180 hI

~
'50

f,u. 150 k,i

f,u ,125 ksi

1\\ ~ ~ 1\ <,

~ <,

'0

i"'-

-60

~
70

~ ./

'ey 300,r ./ 25k5!_ 00,1 15 0,1 100,1 r5 0,1

F tu 90 k$i

F. u,95 ksi

- '1'
50

+t+
'0

'0

20

30

40

00

011

Fig. C4. 11 Bending modulus of rupture for round aHoy-steel lulling.

'\.
80
90

'0

20

Fig. C4.13 Bending modulus of rupture of magnesium-alloy round tubing.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C4 17

Since the allowable or fal1lng bendlng stress Is a function of O/t, and slnce we do not have a tube slze, the design or solution procedure Is by trial and error.

Therefore we will select a tube tram Table C4.3 that has a I/y value in the region of .30 to .37.

Try 2-3/4 - 058. !/y .3233, O/t 47.4. Observation of the modulus at rupture From FIg. C4.12, Fb/Ftu = 0.90. Then Fb = .90 curves show that as Olt increases Fb decreases. x 77000 = 69300 psi. Then lie. = Fb I/y = 6930C ThiS is due to the tact that failure In bending x .3233 = 22400. This is less than the design Is a tocai ra i.iure and the thlnner the wall and bending moment of 24000 so this tube is weak. larger the diameter, the lower the buckllng or crushing stress. However, the larger the D/t Try 3-058. I/y .3868, n/t 51.7 value the greater the section modulus I/y of Fb/Ftu = .885 , Fb = .885 x 77000 = 68000 the tUbe, which means increasing bending lie. Fb I/y 68000 x .3868 = 2630C reetetance. Thus we have two influences which !l.S. (263OC/240OC)-1 .09. act OPPOSitely relative to effecting the bending strength. A study of other tubes in Table C4.3 shows that no other tUbe would be lighter In weight. There are many ways of guessing a tube size tor checking purposes. In this example problem we will aSSume two values for D/t and Tube weight: 30 x .5361 x .101 = 1.70 see what I/Y would calculate to be. The two lbs., as against 2.22 lbs. tor the alloy steel values ot D/t will be 45 and 25. heat treated to 220,000. ThUS aluminum alloy tubes from a weight e'tandpo mt usually yield results better than most materials. This Consider the material alloy steel Ftu=220000:conclusion applies to only low temperatures, For O/t = 45 trom FIg.C4.11, Fb=232000 be low 2500 F , as aluminum alloys lose strength rapidly for temperatures above 2500 to 3000F . Then I!y = M/Fb = 24000/232000 = .103

For O/t = 25, Fb = 266000 Tnen I/y 24000/266OCO .089

Therefore we will refer to Table C4.3 and select tubes that have an I/y value near the 089 to .103 range and then find their true bending strength. Table (A) shows the selection and the necessary calCulations, using Fig. C4.11.
Table A

The student should calculate the lightest titanium tube and the lightest magnesium tube uslng Figs. C4.l4 and C4.t3 respectively and compare the weight results with the steel and aluminum as found above

ULTIMATE TORSIONAL STRENGTH OF ROUND TUBES

C4. 19 Torsional Modulus of Rupture.

Tube Size

1/1

Area

D/l

Fb

M=FbJl1
20800 26400 23200

<::Jill:1
-0.13

1-7/8. -. 035 .09136 .2023 53.6 227000 -3/4 -.049 .1083 .2618 35.7 248000 1-5/8 -.049 .0928 .2426 33.15 250000

+0.10
-0.04

The lightest available tube with a pOSitive margin of satety Is l~ - .049 and Its weight for a 30 inch length Is 30 x .2618 x 0.283 = 2.22 lbse Consider the tube made tram 7075-T6 aluminum alloy material which has a Ftu = 77000.
Pr-om

In Article A6.2 of Chapter AS, the torsion tormula for circular sections, f s : I Tr/J, was derived. This equation assumes the maximum shear stress on the cross-section of a round bar or tube does not exceed the proportional limit of the material, or the stress variation is linear as shown in Fig. C4.l5 and this situation exists under the flight vehicle limit loads. Betore a round bar made of ductile

FIg. C4.12;-

T f. 1

For Dlt = 60, Fb/Ftu = .84, thus Fb = .84 x 77000 = 64700

I/y : M/Fb = 24000/64700

= .37

Fig. C4.15

Fig. C4.16

For D/t = 30, Fb/Ftu = 1.045, thus Fb = 1.045 x 77000 = 80500 I/y = 24000/80500 - .30

material fails in torsion, the shear stresses tall in the inelastic or plastiC range and

C4 18

STRENGTH & DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBrnG IN TENSION COMPRESSION BENDING TORSION AND COMBINED LOADrnGS

thus the internal shear stress distribution is similar to that indicated 1n Fig. C4.l6. The shear stress-strain curve is similar in shape to the tension stress-strain curve.and is equal to approximately 0.6 of the ordinates. Thus to tind the ultimate internal torsional resisting moment, we can divide the crOsssection into a number of circular elements. From the stress curve in Fig. C4.16, the stress at the midpoints of the circular elements can be found. Multiplying the area of the element by this stress times the distance of the element from the center axis gives the moment developed, and adding up the moment of the shear force on all the circular elements gives the total allowable or ultDmate internal resisting moment Ta. torsion equation, we must replace the true stress curve in Fig. C4.16 by a triangular stresS curve with maximum value Fst which will produce the same internal resisting moment as the true stress curve. Thus Ta
It this moment Ta is to be used in the

,"
"" ,
j
~
1(1

~Q

<,

S~lic~/iQ"

WW-T- e2'
,

1-""":--,

""!II' 36 _si

1
I I
i

'"

I
0

"

ec
~.

"

"

eo

Fig. C4.31 Torsional modulus of rupture for magnesium-alloy round tubing.

will it develop if made of the following materials: (1) alloY steel Ftu 95000, (2) aluminum alloy 2024-T3.

= Fst

J/r, or Fst

=Ta

Solution: nit =25. J/r 2(I/y) = 2 x 0.11948 0.2390. LID = 16.25/1.525 10. For alloy steel, Ftu = 95000, we use Fig. C4.l7 where for Dlt ; 25 and L/D = 10, we read Fst/l000 = 47.3. Thus Fst = 47300.

r/J

- - - - (C4.10)

Fat Is called the torsional modulus at rupture stress. It is a fictitious stress, being higher than the ultImate shear stress or the material. Tubes SUbjected to torsion usually fail by inelastic or plastic instability or by elastic instability if the D/t ratio is large. To obtain reliable values ot Fatl resort is usually made to tests, and since the round tube or rod Is the most efficient and most available shape, much testing has been done over the years thus design curves are readily available for round tubes at the most cammon flight vehicle materials.
C4.20 Torsional Modulus of Rupture Curves.

Ta = Fat J/r = 47300 x 0.2390

= 11300

in.lb.

For 2024-T3 aluminum alloy, we refer to Fig. C4.26 where we read Fst 29000.

.@Q in. lb.

Then Ta

Fst J/r = 29000 x 0.2390 =

Problem 2. A round tUbe 10 inches long Is to carry an ultimate torsional moment of 7000 in. lb. Select the lightest tube it made tram aluminum alloy 2024-T4 and alloy steel Ftu 200,000 and compare the resulting weight of each. Solution: Since the modulus of rupture depends on D/t and LID and since the tUbe size is unknown, we will use the trial and check approach. The design calculations can canveniently be made in table form as follows;
Trial Tube Size
Area
Oil

Figs. C4.17 to C4.24 inclusive present curves for finding the modulus of rupture stress Fst tor steel alloys tram Ftu 95000 to 260,000 pSi. It should be noted that the torsional strength 15 influenced by the D/t and the LID values of the tUbe. Figs. C4.25 to C4.30 inclUSiVe give curves far the various aluminum alloys and Fig. C4.31 gives intormation tor one magnesium alloy. All of these curves were taken from (Ret. 1) .

J/r

LID Fst

M.S. T FstJ/r ~-1 T

Alum. Alloy 2024-T4 (Fig. C4.27) 1-1/2 -.065 .2930 23.05 .2016 6.7 1-3/4 -.058 .3083 30.20 .2525 5.7 1-7/8 -.049 .2811 38.25 .2500 5.3 2.0-.049 .3003 40.30 .2860 5.0 30000 6050 28000 7060 26000 6500 26000 7440
-.14

The curves are based on a theoretical investigation by Lee and Ades (see Ref. 2), and have been tound to be in good agreement With experimental results.
C4.21 Problems mustrating Use of Torsional Modulus of Rupture Curves.

.01

-.07
.06

Alloy Steel Ftu ::: 200,000 (Fig. C4.21) 1-.035 1- .049 .1061 28.56 .0485 .1461 20.40 .0664
10
8

98000

4850

-.31 .01 .06

Pr'ob lem 1.

inches long.

r~t

1-5/8 - .065 round tUbe 15 16.25 ultimate torsional moment

10 106000 7050

1-1/4 -.035 .1336 35.70 .0790

94000 7430

, ,

..
.
,"
o
",0

lID
'30

o
'12

"0

);

..
"

110

1/2

.
"

. .

.
0

so
30
0

"

"

..
F tu-240 ksi

"

.."

10

20

'0

.
DIo

20

'"

"
fiu 260 ksi
-

80

Fig. C4.2t Torsional modulus of rupturealloy eteete heat treated to Ftu .. 200 kat.

.
0

Fig. C4I.22 Torsional modulus of rupture alloy steels heat treated La FLu'" 220 kat.

-- -

Llo

'"

"
LlO

- -

'''' '"

''''
'30

,/2

'.

-- -

'"
--

'"
'0

So

110

..
'0

- ---

--- -

"
so ao c

'0

'0 o

20

10
20

so
70 80

10

'0

30

'0
01,

ec

70

eo

'0

en
01,

50

Fig. C4.23 Torsional modulus of rupture alloy steers heat treated to FLu = 240 kst.

Fig. C4.24 Torsional modulus of rupture alloy steels heat treated to Flu = 260 ks i.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C4.21

211 .... 'OIGING

F""'6~ ~~

"
,~

m"'l "AI"<:;
Ftu64k$i

"

,"
on

,
...... JO

'n

"
"'011,
~

"
II <;O-.lIlD_ 'll'. _OOUHO' .... LQIII_ ~ lIJ 01' 1t0000ID 11lllNGtM ....ONG TIOI nH<l ..........
IA"'~ 1\.11)....

"
"

C\,II.., "'"WOO"......... ,U .....

"

"

"

"

tit

...

"

"

"

"

"

"
Ftc. C4.28

"

" ,,,

"

"

Fig. C4.25 Torstonal modulus of fupture 2014-T8 alumiDum alloy lorrtni'.

Torstonal modu.Iu of rupture 20244T3 ulUDtDum alloy tubing'.

m"" "-",....
Ftu'6Z klJ

II:t
, ,

, ,

, , , ,

""

,~

::
"
, , , ,

, ,

'" ,n , ,

"
"
"

"

, " "

,
"

"

,,,

"

"

"

"

'"

"

FII. C4.28 TontoDlll modulus of rupture 80el-T! aluminum alloy tubtDg.

Ftg. C4.27 Torstonal!XlOdUlua of nqlture 2D24-T4 aluminum alloy tubin(.

1GF,... OO1.UD 1lQ~' Ftunk~

"

,~

u.

-r-r-

."
"
"

,, "

u.

'"

'"

" " "

HOlt, 1>4 C\,II\1It!'lll!!t.. n.........11...... " ~ ....... (VD~" Co-,,'UllOl . . . . . _ . _OOUHO.....LQllllOOI

.
Tlli ""Slltl" .... OIItOLlCllO '1IlI1OG"" ALO>'G TIOI'''''NO ... _ .

"
" "

"

"

"

"

..

"

"

"

"

"

Fig. C4.29 Torstonal modulua of rupture 7D'75-T8 alumlnum alloy rolled rod.

Ftg. C4.30 Torsional modulua of rupture 707~-T8 aluminum alloy (orglnc.

C4.22

STRENGTH & DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBING IN TENSION, COMPRESSION, BENDING, TORSION AND COMBfNED LOADINGS.

The lightest alumin~~ tube .with a positive margin of safety is 2 - .049. The weight of a 10 In. length = .3003 x .101 x 10 = .303 Ibs. The lightest steel tube is 1-1/4 - .035 and its weight is 0.1336 x .283 x 10::: .378 lbs. Although the steel is heat treated to Ftu ::: 200,000, it still is heavier than the aluminum alloy 'tube ,

equation C4.14 is not used.

1,1

1.0

.,
.8

Fig. C4.32 Combined Bending & Compression

Rc+Rb':1

STRENGTH OF ROUND TUBES UNDER COMBINED LOADINGS

Reference should be made to Artiale CI.lS of Chapter Cl for general explanation of stress-ratios and interaction equations as used in determining the ultimate strength of structural members under combined loadings.
C4. 22 Combined Bending & Compression.

.7
.6

.5
.4 - - - - - - - "

In tubes sUbjected to combined bending

and compression, the stress due to compression

. 1 .;

is uniform over the crass-section whereas the bending stress is not uniform over the crosssection. The following stress ratio equatton Is pOSSibly somewhat conservative but is recommended by (Ref. 1).
Rc+Rb'=l

o
C4.23 illustrative Problem Involving Combined. Bending and Compression.

- - - - - - - (04.11)

whiCh can be 'Nritten

~+!b' =1.0

---

(04.12)

Fig. C4.33 shows a round 2024-T3 tube acting as a beam-column. It Is supported at pin A and by the inclined strut BC at point C. Let it be reqUired to determine the margin of safety when carrying a uniform lateral load ot
10 Ib./ln.

Fc

Fb

Where f b ' equals the maximum bending stress including the secondary moments due to axial load times beam deflection

To find the axial load in tube AC, take moments about paint (B).
2MB

=-

40

Fa ::: Bending modulus of rupture stress

to ::: Compressive column aiial stress Fe ::: Allowable column streSS
Equation C4.12 could also be written,

RAH

x 10 x 20 + 20 RAH = 0, hence = 400 lb.

Thua axial load in tube is 400 lb. compression. The column strength a! a 2024-T3 1-1/4 .049 round tube 40 inches long with end fiXity

P M -+-=1.0 Pa Ila

- - - - - - - - - - - - (04.13)

C ::: 1 15 obtained tram FIg. C4.6 and equals - 2100 Lb , , hence stress ratio
R Column load _ -400 o - Column strength - -=2lOO -

Where P = column load Pa :Ii allowable column load at failure 11 :: bending moment

191

Ma.

=allowable

ult 1mB.te bending moment

Fig. C4.32 shows a plot of equation C4.11. The margin of safety equation is,
M.S. =

The bending moment will be maxtmum at the center Of span because of symmetry of loading and the value ot the moment is obtained from the following equation which includes the secondary moments due to deflection. (Reference table AI041 ot chapter A10)4
"LX 'lila

Rc + Rb' - 1

- - - (04.14)

= w," ,

(1

1 ) sec 2T

- - - - - - - - - -

(A)

Fig. C4.32 shows margin of satety curies tor estimating closely the I1.S. values ir

,= / IEI ! = , T p

10,300'400000 x .0334

=.~ v 860 = 29.3

ANALYSIS AND DESIGN OF

40"
w :: 10#/in.

FLIGHT VEHICLE STRUCTURES

C4.23

bending and tension on round tUbes, which is widely used, is

- - - - - - - - (C4.15) (Ref.2)

1-1/4 - .049 Rd. Tube

Fig. C4.33

This equation Is plotted in Fig. C4.34. The stress ratio Rt : ft/Ftu. The figure is based on D/t : 10 and in general is conservative for other Olt ratios.
1.1 1.0

wj" : 10

860 : 8600
sec

.9

,j

L j = ~ 58.6 = 683 2

...h.. = 2j

1 ---=-.,-L
COS2j

.8
.7

. ,
. __ :-i.::. -_..... _.
~--

-1

>

From Table AIO.2 or Chapter AIO cos or .683 :

.7757.

""

6 6

.4
3 2
.1

.
o
.1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1

From Fig. C3.2, the mOdulus at rupture Fb for a 2024-T3 round tube which has D/t value of 25.5 equals 64000 psi. Therefore the bending
strength M

= FbI/y = 64000 x .0534

3420 InvIbs ,

Rt
C4. 24 Combined Bending and Torsion.

Stress ratio in bending Fb : ~g

member- Is not weak.

: .728

Re + Rb = .191 + .728 : .917 which shows that

1

The interaction equation for combined bending and torsion tram (Ret. 1) is, Rb+Rst=l
-(C4.16)

Margin ot safety M.S.:

= ( .191+.728 1 ) M.S.
f

This equation is plotted in Fig. C4.35 tor ~ O. Curves showing various M.S. values

The margin of safety could be read directly tram curves in Fig. C4.32. The student should notice that the maximum bending moment ot 2490 in. lb. is 24 percent greater than the primary moment which equals wL"/8 = 10 x 40"/8 ~ 2000. The lateral deflection at the midpoint of the tube thus equals
490/400

11

~:~:-

+ . .:

<:: ~:f _.

-t : :

.T

J~:'~~,~:::-~~,~ --=-di$.~=~1

1.0 "
.9~ . .......- . : :

_ . -~6in8d=-'_." T:-~~=i~.:&~_~~~~1

.7---

8 ==-';'~---'~_--"-><--''---''-<i-----=--= : _ J -__ t _ - :~ _--. ....

- ,.;.

, .

---!

:L

1.22 Inch.

The secondary moments due to lateral deflection do not vary linearly, so if design loads were increased the calculation of the maximum bending should be repeated instead of assuning that the moment would increase directly as the applied load to the beam.
C4.23 Combined Bending and Tension.

.3
i

I
1

The interaction equation tor combined

.1

.2

.3

.4

.5

.8

.7

.8

.9

1.0

1.1

llst

C4.24

STRENGTH &: DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBlliG m TENSION, COMPRESSION, BENDING, TORSION AND CQMBrnED LOADINGS.

tor various values cr Rb and Rst are also shown

on the r tgur-e .
The expression for Margin of Safety Is,
- - - (C4.17) ILLUSTRATIVE PROBLEM 1

T8

=M,J 1 + (T!l1).lI

- - - - - - -

- - (C4.21)

Having the value of Te J select tube sizes will develop this torsional moment Te as was done in Problem Z of Art. C4.2l. These sizes are then checked for combined bending and torsion as illustrated above in the example problem.
tr~t

_.C4.24 tnttmate strength in Combined Compression, Bending and Torsion.

A 1-1/2 - 058 round steel tube (Ftu :: 125000) 1s 30 inches long. It 1s subjected to an ultimate design bending moment ot 10,000 in. los. and a torsional moment or 6000 in. lbs. Find the Margin or Safety.
sctutaon: D/t

The interaction equation for combined compression, bending and torsion fram (Ref. 1)
1s,

= 25.85,
=

I/y

= .0912, LID

= 20
Flgs. C4.36 and 37 show a plot of this equation. The expression for the Margin of Safety is,

tor D/t = 25.85 and F'tu :::I 125000 we read Fb a 148000. Then Ila FbI/y 148000 x .0912 = 13500 In. lb.

To tind FbI we reter to Fig. C4.11 where

Then Rb

= Me. ::

10000 13500 ... .74

Rc

V Rb'

- - - - - (C4.23)
+

R;t

To fInd Fst, we reter to Fig. C4.18 where

[or D/t a 25.85 and L/D = 20 we read Fat/l000 to be 61, whence Fat'" 61000. Rst- T - Ta

To illustrate the use of the interaction curves, let us assume the following values for the three stress ratios: -

= 11120 6000 = 54
1 1

Rc

= .333,

Rb = .20,

Rs t = .20

Then Rst/R c

= .20/.333 = .60.

" Rb' + Rst"

=r====.:=-:::1"=--:;:-:;; _1 =_1__ 1 = .09

,j .74.11 "'" .54.917

thus the ultimate strength ot the tube has 9 percent Margin ot Safety under the combined loading. In a design problem which involves a trial and error procedure, USing an equivalent torsional moment Te which will produce the same torsional stress as produced by the combined bending and torsional loads is quite useful in shortening the trial and error procedure.
Let [s(max)

In Fig. C4.36 locate pOint (a) at the Intersection of Rc .333 and Rb .20. Since the intersection pOint (a) 11es inside the RstIR c = .60 curve, we know that a pOSitive margin at safety eXists. A line 15 now drawn through (0) and (a) and extended to an intersection with the RstlRc = .60 curve at point (b). PrOjecting vertically downward to Rc scale, we read Rc = .538. Then M.S. = (.538/

.333) - 1

= .62.

= Ter/2I

- - - - - -

(C3.18) (C4.19)

fs(max) also equals .j r; '" (tb/Z)

Also [s

. .L.
1 I ..--.
~

=Tr/2I,

and [b

=Mr/I

SUbstituting these values in C4.l8,

[s(max)

c.:'."'__ ; . '"

-I

= z'; [Mv' 1 + (TIM)'J

- - - - - - (C4.20)
1.0 .9
.4

.3

.2

.1

Equating C4.20 and C4.18 and solving for Te,

-------/.- - - - . ::::-..

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C4.25

pr-oject tng horizontally rr-cm pornt (tr) to Rb scale, we read Rb = .323. Then M.S. = (.323/.200) - 1 = .62; which checks value previously found. If the ratio of RstlRc is greater than one, we use the curves in Fig. C4.37 and use the ratio Rc/Rst. Substituting in
1

2~ = 2 x ;~.43
sec 2
Lj

586 : 2:. : 1.005

: cos lL/ 2 j :

.53~1

: 1.87 (See Table AI0.2 )

hence,
M

.333 +

as given

max.

ro x 154.7

(1-1.87)

= -1347
Rc :

in. lb.

.20" + .20.:11

from use of curves.

1.0

The column strength tor a 25 inch length and end fixity c 1 can be read tram Fig.
-t~---_---t----

_c'cR,;t..iQ;':1b~F ~_;~=--='-=i=-=::~~~-';;~~_,,::-i

------

C4.6 and equals 3700 lb. 1600/3700 : .432.

Then

P/Pa :

To tind the ultimate bending strength, we reter to Fig. C4.l2, where tor D/t ~ 1.125/.049 22.95, we read FblFtu = 1.04. Then Fb : 1. 04 x 62000 : 64500. Thus Ma FbI/y 64500 x .0427 : 2760 in. lb.

R' b

Therefore R6 = 2760= .487. To tind the ultimate torsional strength we reter to Fig. C4.26 where tor D/t = 22.95
and estimating location of LID : 25/1.125

1347

22.2 line, we read Fat

.1

=27500.
2350
650

CC_,

- ,- --_,

; -----.+---

-:._::_:.--t-:-~- ~.
, ~ ~

-::.:::- :':'.: --=:--i -=-.:..-.:-r

_~

Then Ta : FstJ/r
Whence Rst

= 27500 x

.0427 x 2

2350

; -.; ----_--T~-:=._~~_ ~ ..--- .-:.r -----~-- - . ---,..

-~----

__

,---

,.----

=....l.. =
Ta

~ 276

.1

.a

.3

-'-

.5

.6

.7

.8

.9

1.0

Rot
+

ILLUSTRATIVE PROBLEM
1 A 1-1/8 - .049 round tube made from

" Rat

2024-T3 aluminum alloy carries the ultimate design loads as shown in Fig. C4.38. Find the margin of satety under the combined loading.
"t-s.o~
1600

.432 +

.487'" + .276 8

-1 : .01
Ct. 25 Ultimate strength in Combined Bending and Flexural Shear.

Gl' . , , , , !'
w

= lO*/in.

l.--- .,,,...l-lI8 -.049 .o-n R<l. Tub (!.,'~'~

-Fig. Ct. 38

,,

+ , , , +

1600*

The interaction curve trom this type ot combined, loading rran (Ref. 2) is,

Rb

R = 1
:

- - -1

(C4.24) (C4.25 )

Solution:
M.S.

The maximum bending moment due to symmetry will occur at midpoint ot tube. For a beam column carrying a unitorm side load with no end moments, the maximum. moment is given by the following expression. (See Chapter AlO, Table
AlO.l). l'inax. : wj" (1 - sec j :

h b+

R;

- - - -- - - -

The allowable flexural or transverse shear stress is taken as 1.2 times the allowable torsional stress ot the tube (Fat).

t.J)
.; 154.7 : 12.43

ILLUSTRATIVE PROBLEII
A 1-1/4 - .058 round aluminum alloy 2024-T3 tube is used as a Simple beam and carries an ultimate design load ot 600 lb. as shown in Fig. C4.39.

Jf:i : jlo,3001~gg x .0240 :

----- -~,". -: ,:.; ;:'~.:::.~;:'":~;< .::.< .: ':7',~:~,,;<,

',?,;:r

~' ~.

<:: ~,'. ,'-.'.

-------,-~: >~':::'~';:" ~;~:,f~:~}:~

32. 2~

..

C4.26

STRENGTH & DESIGN OF ROtrnD, STREAMLINE, OVAL AND SQUARE TUBING IN TENSION COMPRESSION BENDING TORSION AND COMBINED LOADINGS.

1---12.5--'1' 25"
300
//////

600

, ,
...>750
/ / /'7-;-

Shear Diagram

Interaction curves for this equation are given 1n Fig. C4.40, where Rs the stress ratio tor flexural shear can be found as explained in the prev10us article
=> =>
","

BendingMoment
Diagram

Fig. C4.39

The critical section is adjacent to the midpoint at be~~ where shear and bending forces are largest. The ultbnate bending strength can be found from Fig. C4.26 where tor D/t a 21.60 we read

Fb/Ftu
4100. --

and therefore

= 1.07.

Ma

Then Fe

.915. ----' The allowable shear stress Fs can be taken as 1.2 Fst (Ref. 2).

Then Hb

= MIMe.

= FbI/y = 66300 x .06187 =

3750/4100

= 62000

x 1.07

= 66300

To tind Fat, we reter to Fig. C4.26 where tor D/t = 21. 6 and LID = 20, we read Fst ~
28400. Then Fs

1.2 x 28400

= 34000.

=>

The maximum. shear stress in a round tube which occurs at the centerline axis is given by the equation,

...,

'.".,
-:
~

where V = vertical beam shear load A : tube area D = outsIde diameter d inside diameter

'"
----------"-*",--"---- =

An approximate formula is t s = 2 A' tor D/t -10 error less than 1%.
USing the approximate equation,
f

= 2 .2172 x 300 = 2760

= ~=
Fs

pe

Rs

2760 34000
1

"--.J

0811

C4.27 Ultimate Strength in Combined Tension and Torsion.

The interaction equation tor this type of loading as presented in Ret. 3 is,

= .09
The ettect ot the shear stress is less than 1 percent. In us1ng Fa 1.2 Fst, it the result 1s greater than Fau tor the material, use Fau tor Fs '

Rt+Rst=l
'~'here

- - - - - - - - - - - (C4.27)

Rt
- - - - - - - - (C4.28)

C4.26 Ultimate strength in Combined Compression, Bending, Flexural Shear and Torsion.

C4.28 Ultimate Strength 1n Combined Tension, Torsion and Internal Pressure p in psi.

The interaction equation for this combined loading as presented 1n Ret. 2 is,

Ref. 3 gives the following interaction equation.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C4.27

(C4.29)
Where Rp = pip . max where p
max

(3)

= 2tFtu d
(4)

where t

= wall

thiCkness and d = tUbe diameter.

If the truss of Fig. 2 is heat treated to 180000 psi atter welding, how much weight could be saved over the results obtained in problem (2). Same as problem (2) but change material to 2024-T3 aluminum alloy round tubes. Members to be fastened together by rivets and gusset plates. For deSign of tension members assume that rivet or bolt holes cutout 10~ or 'tube area. Fig. 3 shows a typical tubular engine mount structure. The engine is supported at paints A and B. For design purposes assume that engine torque is reacted 60% at A and 40% at E. Tube material is steel Ftu = 95000 psi. Use C = 1 tor all members. Determine tube sizes for the tOllowing deSign conditions: Condition I Vertical Load factor = 10 Thrust Load factor
(down)

The expression for Margin of Safety IS,

1

- - - (C4.30)

PROBLmS
(5-)

(1)

Fig. 1 shows a portion ot a steel tubular fuselage of a small airplane. The critical tension and compression load IS shown adjaCent to each truss member. Assuming and end fixity coefficient c = 2 tor all members, select tube sizes tor all

members or the truss.

The minimum size to

be used Is 3/4 - .035. The top and bottom longerons should be spliced at least once using telescoping sizes. The material Is alloy steel Ftu 95000 and truss 1s welded.

=
I
,

(rorwarn ) Engine Torque Load factor = 2

Condition II {Same as I except vertical load factor Is 5 up.

=2

U
I

I-"f+ eroo
.:\040 U

TI I
35~
I

..'"

.3ns ,.., .~ ...'\.

-7400

.1

C.

, f- 28+ --:-- 2216

c.
'

-uoo

"''''

.
,
]8

(!b '0
",

-,
c

,, , /' .j '"

. ..

"" ".

General Data:

~el~ht

at power plant instal:ation = 440 lb.

"

Maximum engine thrust = 400 lb. Engine H.P. = 120 at 2000 R.P.M.

ft -t- "

....J-H.l-i , ,

T
.1.

;"""

-=-= --=-=-7-/ T
C

C'

Fig. 1

.ll._
A

(2 )

For the cantilever welded steel tubular truss of Fig. 2, select the lightest members for the truss loading as shown.
40001

T
30"

t
20

T'

17t 16

1
30"
r II I

JI

-~:-:_-

30"

---I
2000#'

(6)

10001

Fig. 2

The loads shown in rig. 4 are to be transmitted to the support at the lett, or in other words. a cantilever structure. The problem is to design the lightest truss configuration ~sin~ round tubes of alloy steel Ftu = 95000 and Nelded together at the tru~s JOints. Use C = 1.5 for end fixity of all members. There are

The top 'and bottom longerons should be continuous members. Minimum size 3/4 _ .035. Use C = 2 for web members and C 1.5 tor langerans. ~ater1al (chrome-moly steel Ftu 95000 PSi).

~ ~.~:: -~:.~:~: ~~!- ~>~; '/,~:: c;~f. ~~. :.:::" _:~:o<!. !~ -~ ;;'"~"'~'1Z~'" ,~~. ~1 ~:',.~ .~.r;. '., '.~ ... :'"-" >-ij~~~~"::,~ .. ~ '. '~:., ~ :,='" .'~,~'~, '." "...' ,-_- ." ~~/1~' ~~.~..r"
>

_......----

3-23
.,-~ ~ ~

.....

C4.28

STRENGTH & DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBING IN TENSION COMPRESSION BENDffiG TORSION AND COMBINED LOADmGS.

no restrictions on tJ~e arrangement of members, is the lightest truss. of weight of any gusset joints.

of truss or however, the goal Omit consideration plates at truss

1200

c--h -1--C',D'
I!.', S'

'Thrun Llne

h F
(7 ) (8 )

1500

l'------..:k,-----,l
28"

1400

2'"

-+-

-+-

28"

-+-

CO
2'"

a
til

lb.

-I

Fig. 6

Same as problem (6), but instead of a cantilever truss use a Simply supported truss with supports at points (A) and (B). Fig. 5 shows a tront beam and tront 11ft strut in an externally braced monoplane. The wing beam and lift strut are in the same vertical plane. The ultimate design loads on the beam tor the critical conditions are w a 50 Ib./in. and w -30 Ibs. per inch. Minus means load is acting

down.
(a) (b) Design a streamline tube to act as the litt strut. Material 1s 2024-T3 aluminum alloy. same as (a) but made trom alloy steel Fty = 75000. Compare the weights of the two designs,
112" - - - - I . . - - 40" wi/ln. 1Pin

---l

Fig. 5

PIn I

Fig: 6 illustrates the strut and wire bracing tor attaching tloat to fuselage ot a seaplane. Determdne the necessary sizes tor the streamline struts AC and ED for the tollowing load conditions. V = -32000 Ibs., H:II _ 8000 Ibe , Condition 2. V = - 8000 Ibs., H -28000 Iba , Material 2024-T3 aluminum alloy. Use C-l. (10) Tube size 2 - .065 round. L = 44 in., C'= 1.5. Material alloy steel Ftu = 95000, welded at ends. Design Ultimate loads equal 22000 lb. compreSSion and 28000 Ibs. tension. rind margin of safety. (11 ) same as Problem 10 but heat treated to Ftu :II 150000 atter welding. Condition 1.

(12) The ultimate design load it 20,000 lbs. compression. L = 30 In. Use C = 1. Design the lightest round tube from the following materials and compare their weights. (a) Aluminum alloy 2024-T3 (b) Alloy steel Ftu = 180,000 (c) Magnesium alloy Fcy = 10,000 (13) Same design load as in problem (10) but design the lightest streamline tUbe from 2024-T3 aluminum alloy material. (14) A round tube Is to carry an ultimate pure bending moment of 14000 in. lbs. Select the lightest tube sizB from the following materials and compare their weights. (a) Alloy steel Ftu 240,000, (b) 2024-T3 aluminum alloy, (c) MagneSium alloy Fcy = 30000, (d) Titanium 6AL-4V alloy. (15) A round tube 20 inches long is to carry an ultimate torsional moment of 15000 in. lb. Select the lightest tube size from the following materials and compare their weight~ , (a) AllOY steel Ftu : 180,000, (b) Aluminum alloy 2024-T3, (c) Magnesium alloy Ftu = 36000. (16) Determine the lightest 2024-T3 aluminum alloy round tUbe 10 inches long to carry a combined bending and torsional design load ot 4500 and 3000 in.lbs. respectively. (17) Same as Problem 16, but change material to alloy steel Ftu = 95000. (18) A 1-1/2 - .065 2024-T3 round tube 50 inches long is used as a beam-column. The distributed load on beam Is 12 lb. per inch and the axial load is 700 Ibs. What is the M.S. under these loads. (19) It the tube in problem (19) was also subjected to a torsional moment at 1400 in.lb., what would be the M.S.

Reterences (1)
(2) (3)

Military Handbook. MIL-HDBK-5. March, 1961Lockheed Report 2072. M. A. SadOWSky, nA PrinCiple or MaXimum Plastic Resistance," Journal of Applied Mechanics, June, 1943

_-"------

-CHAPTER C5

BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR, BENDING AND UNDER COMBINED STRESS SYSTEMS
CS. 1 Introduction.

Chapter Al8, Part 2, introduced the student to the theoretical approach to the problem of determining the buckling equation tor flat sheet in compression with various edge or boundary conditions. A similar theoretical approach has been made for other load systems, such as shear and bending, thus the buckling equations for flat sheet have been available for many years. This chapter will summarize these equations and provide deSign charts for practical use in designing sheet and plate structures. Most et the

COLUMN

\r

material in this chapter Is taken tram (Ref. 1), NACA Technical Note 3781-Part I, "Buckling ot
Flat Plates" by Gerard and Becker. This report Is a compr-enens t va study and summary ot practically all important theoretical and experimental work published betore 1957. The report is especially usetul to structural design engineers.
CS. 2 Equation for Elastic Bucldlng strength 01 Flat Sheet in Compression.

Frcm Chapter A18, the equation tor the elastic instability at flat sheet in compression
1S,
(~)

Fig. C5. 1 (Ref. 1) Transition from column to plate as . supports are added along unloaded edges. Note chanps in

buckle configurations.

- - - - - - (C5.1)

Fig. C5.2 gives curves for finding the buckling coefficient kc for varlous boundarY or edge condi tiona and alb ratio of the sheet. The letter C on edge means clamped or fixed against rotation. The letter F means a free edge and S5 means Simply supported or hinged. Fig. C5.3 shows curves for kc for various degrees of restraint (e) along the sides of the sheet panel, where e 1s the ratio of rotational rigidity Of the plate edge stiffener to the rotational rigid1ty of the plate. Fig. C5.4 shows curves for kc for a flange that has one edge free and the other with various degrees at edge restraint. Fig. C5.5 illustrates where the compressive stress varies linearly over the length at the sheet, a typical case being the sheet panels on the upper side of a cantilever wing under normal flight condition. Fig. C5.6 gives the kc factor tor a long sheet panel with two extremes of edge stiffener, namely a zee stiffener which is a torsionally weak stiffener and a hat section
CS.l

Where kc = bUCkling coefficient which depends on edge boundary conditions and Sheet aspect ratio (alb) E = modulus of elasticity V e = elastic Poisson's ratio b = short dimension of plate or loaded edge t

= sheet thickness

CS. 3 Buclding Coefficient kc

Fig. C5.1 shows the change in buckled Shape as the boundary condf tiona are changed on the unloaded edges tram free to restrained. In Fig. (a) the sides are tree, thus sheet acts as a column. In Fig. (b) one side is restrained and the other side free, and such a restrained sheet is referred to as a flange. In Fig. (c) both sides are restrained and this restrained element 15 referred to as a plate.

<.:,-~,~",":.:.~,- -1

':-~P-1

CS.2
16

BUCKLING STRENGTH OF FLAT SHEET rn COMPRESSION, SHEAR, BENDING AND UNDER COMBINED STRESS SYSTEMS

~f---!

-->
I
I

,
l0

, I , I I' 1
I

I,

,
A

I-

I"
I,

I
I
e
H

\1" \~
,I ,I
\

\
\

3< ~'WE
','.'."

"I _.-

,
0
T

r----I
I
-~

II

I'uu~:~
'~

<

'"
s

~"

..'''',.'''.-.
.
~

K,
8 6

~.~'A

'\

,,

- - - - U/IJU1I!/1 SNWS CL_D

LDA/1U /HIS SfWU

'~
I
! '

,-=
I ,
,

-~

,
, \

-.:,

......

:::-------

f-

,'-I,
I

-';-,!

l,'

\. l/'....
i\~'
i \',

...,.... , I - - -...

(JA~

JA-

<:

'-- ~

50

20
Io

-----...L____
,

~~-:::

2 I

~
.

..

ts

, \' \ <, I

.. -

'k ' I . ---.,---_.J.__

',i:---_~__

I
0

. .

II

'

%.
Fig~

"

CS.2 (Ref. 1) compreeetve-cucsnng coefficients for flat reetangular plates.

..
ts
iT 16

Fig. CS.3 (Ref. 1) Compressive-buckling-stress coefficient of plates as a function of alb for various amounts of edge rotational restraint.
7

OJ

=1
I

-~tfj!==
-: : a--b

,.
a

14

-l

..!.. ~:- __ I==:

I

I ,

I-

D1

", ,
.~

lZ

so
20

K,_
5

~~
I~

U
ta
9

'D

\J

---~
'

'V<>i
I

.to:'. .

AT

a/b'(O

.;",

2
8

500
" 2.41')

-5

-.
~

2DO
1.60 r- I.D._

7
6

--'-~-

,
E-

/
Z.O

o
JO .J5

<

a>

25

a/b

Fig. C5.4 (Ref. 1) Compressive-buckling-stress coefficient of flanges as a function of alb for various amounts of edge rotational restraint.

Fig. C5.5 (Ref. 1) Average compressive-buckling-stress coefficient for rectangular flat plate of constant thickness with linearly varying axial load. kca v ,,":2 E I' ) ". ~ av =- - ' ' ' - - - - b 2 12 (1 - Zle )

i.~

..

....

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CS.3

which is a closed section and. therefore, a relatively torsionally strong stiffener. Fig. C5.6a gives the compression bUCkling coefficients ke for isosceles triangular plates.

Substituting in Eq. C5.l, O'er

n'x4.0xl0.700,OOO (.04)'. 2480 ps i . 12(1-.3') 5

...--,-----,..---,..---.,---,

This stress is below the proportional limit stress for the material, thus equation C5.l applies and needs no plasticity correction.
C5.4 Equation for Inelastic Buckling Strength of Flat Sheet in Compreeetcn.

f---";""--""'."

'" '1-f:-,,4---+----+--+--t---;

-TOI'SIONAU.J'

1f:f:o"EDfl/~w:lJ---+----+---j----j 'k---l!--t I---'---' LSTTF~NDt Z

, I

If the buckling or instability occurs at a stress in the inelastic or plastic stress range, then E and V are not the same as for elastic buckling, thus a plastiCity correction factor is required and equation C5.l is written,
12(1-Ve '

- - - - - - - - - (C5.2)

Where ~ is the plasticity reduction factor and equals Ocr plastic/ocr elastic.
b/t

Fig. CS.6 (Ref. l) CompreSSlv6 coefficient for long rectangulaJ.' stiffened panels as a function of bit and stiffener
4buckling

torsional rigidity.

The values of Icc and Ve are always the elastic values since the coefficient contains all changes in those terms resulting from inelastic behavior.

,
I
,

"

'" .-~ .... '."

~
'

.... ,,' .... /

/

..

L .....,' ...........,...... "J I ! ! j ! !

-t-

. _.

"',-

k,

"

,
00

\.~':'""~ \-S/MMr SU~IfTE:,

'.0

'0

"

Fig. C5. 6a (Ref. 1) Uniform Compression.

Illustrative Problem. Find the comoress1ve buCkling stress for a sheet panel with (a) 10 and b = 5 inches, thickness t = .04 and all edges are stmply supported. Material is 2024-T3 aluminum alloy. Solution: E = 10,700,000. Ve = 0.3, alb = lOIS 2. The boundary or edge condition corresponds to Case (c) in Fig. C5.2. Thus USing curve (c) for alb = 2, we read k c = 4.0.

A tremendous amount of theoretical and experimental work has been done relative to the value of the so-called plasticity correction factor. POSSibly the first values used by design engineers were ry = Et/E or ~ ~ = Esec/E. Whatever the expression for ~ it must involve a ~easure of the stiffness ot the material in the inelastic stress range and since the stress-strain rel~tion in the plastic range is non-linear, a resort must be made to the stress-strain curve to obtain a plasticity correction factor. This complication is greatly simplified by USing the Ramberg and OsgOOd equations for the stress-strain curve which involves 3 Simple parameters. (The reader should refer to Chapter 81 for information on the Ramberg-Osgood equations.) Thus using the Ramber-Osgood parameters (Ref.l) presents Figs. C5.7 and C5.8 for finding the compressive buckling stress for flat sheet panels with various boundary conditions for both elastic and inelastic buckling or instability. '
C5.5 Simple Problems to nlustrate Use of Curves in Figs. C5.7 and C5. 8. crcr

The sketch shows a 3 x 9 inch sheet panel. The sides are simply supported. The material is aluminum alloy 2024-T3. The thickness is .094". E = 10,700,000.

55

9":a

'.,

. '.

/. .

.;:i!~ >7i~t:\'4t~

~3b"

:~

cs.

BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR, BENDING AND UNDER COMBlliED STRESS SYSTEMS

8S 0;,,, S 1~., <:.: ,4 0;;,.:. ie, ,>},,~ Z5'

TM.v.

hb

....

0;,

Ii.,.,,:. n, oS

121--+--+--I--+--+-......:--t--+-:79-L'----::::l.
o

rl.'.... t.vUi'
pLi\~U~~ G"',..~

.. }~b
t
8

,401.4""

, p"I",.,.

........ 11,'
q;,.,. ..
r",tof-:.

42. 't.

CTer,.o

I)

or

f---t--t--+--W7-7.:l'?'--,--t----j--+--1
f---f---f---,~'7""-+_--.;_--..,..---

1',0.,> a

'{OJ

1 ,., -1'
'D~b

FCc/F I). '1 B

rt..::~

Z"'",;L;
,Jrttl,c'

-f--t_---z~C-t_-t_-+--+_-

S$~

,
,
%

II i .2lo----~(L-+--+---+---+-__+-+-...,-+____i
(
.2!#.6 4
iOt2 14/.6 1.8Z0

~
' I

kcnaE til 12 (1 - vea)O e , : . Cf;'>

Fig. C5.7 Chart of Nondimensional Compressive Buckling Stress for Long Hinged Flange ~

=(EslE)(1

- Ve'l/(l - V'l.

n,v.ir 1'.i81.10 ,

Ocr lOa ."

FcclF,.: s r--+--!--~~.p.;"""-+--+----i
-1---+-, ,

or

-<;.

UJ

lZ

kcn 3 E t a 12 (1. Vell)O'o .,

tb')

'.

20

FIg. cs. a Chart of Nondimensional Compressive Buckling stress for Long Clamped Flanges and. for Supported Plates with Edge Rotational Restraint.
....: / .~':~_

,-'~

~.:.;-."::":':':::-J

..

(Esl2E){ 1 + 0.5 [1 + (3Et/E.l]

'I'} (1 - 1/0')/(1 -1/').

. . . . . . . . . . . .

I: . I:~I-:*- - ~~~~~ii~~~::::~
__ Z= -

.-:-c---.-.:::c ..__._'._-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C5.5

Va :: 0.3.

Find the buckling stress ocr.

Solution: We use Fig. CS.8 since i t covers the boundary conditions- of our problem. The parameter for bottom scale IS,
- - - - - - - (A)

Thus a further correction must be made for alclad sheets because of the lower strength clad covering material. Thus the buckling stress tor alclad sheets can be written: - - - - - - (C5.3)

I
For alb = 9/3 :: 3, we find kc trom curve (o) ot Fig. C5.2 equals 4;0.

n/a

The use of Fig. C5.S involves the use of a 0'. r and n the Ramberg-OsgoOd parameters. Referring to Table Bl.1 or Chapter 81, We find for 2024-T3 aluminum alloy that a o. '7 -= 3QOOO and the Shape factor n 2 11.5. ~ SUbstituting in (A):4.0 n a x 10,700,000
12 '(1 - .3") 39,.000

~ 'trom Table C5.1.

Cladding

Fig. C5.9

and n = 11.5 curve, we read on left hand scale that ocr/a e , or 0,84., /~.
';I

From Flg. C5.8 using .98 on bottom scale

Reference 1 gives Simplified cladding reduction factors as summarized in Table C5.l. Thus the buckling streSS tor alclad sheets 1s determined tor the pr~ry strength properties as normally listed tor such materials as illustrated in the two previous example problems. The resulting ocr is then reduced by use of equation C5.3, using valueS ot

Then ocr

= 39000

x .84

= 32800

-------psi. "

It we neglected any plasticity effect, tpen we would use equation CS.2 with ry = 1.0, or, ocr = n"x4.0xIO,700,OOO (,O~)" =~,PS)l
12 (1 - .3")

Whereas the actual buckling &tress was 32800, or in this case the plasticity correction tactor is 328/384 = .854. The sheet thickness used in this example ot .094 is relatively large. If we change the sheet thickness to .051 inches the results would be practicallY no correction within the accuracy of reading the curves, and the buckling stress ocr would calculate to be 11200 psi, which 1s below the proportional limit stress and thus no Plasticity correction.
C5.8 Cladding Reduction Factors.

o
o c1

e
Fig. C5. 10 (Re!. 1) Stressstrain Curves for Cladding. Core, and AIclad Combinations. a/o cor e S 1 - f +fl f; fl-aClb core-

Table C5. 1 (Ref. 1) Summary of SimplUied Cladding Reduction Factors

Loading

Aluminum allay sheet is available With a thin covering at practically pure aluminum and is widely used in aircraft structures. Such ~aterial is referred to as alclad or clad aluminum alloy. The mechanical strength properties of this cla( material is considerably lower than the core material. Since the clad is located at the extreme fibers of the alclad Sheet, it is located where the strains attain their highest value when buckling takes place. Fig. C5.9 shows make up of an alclad sheet and Fig. C5.10 shows the stress-strain curves for cladding, core and alclad combinations.

a cl< o cr ooc crpi

1 + <:;1'/4)
1 + 3f

o cr:' apl
1 + 3f

Short plate columns

Long plate columns

CompreSSion and shear panels

1 + 3f

1 + 3f

1 + :;1f
1 + 3f
1 + 3f

" I ,J\,:

C5,6

BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR, BENDING AND UNDER COMBINED STRESS SYSTEMS

BUCKLING UNDER SHEAR LOADS

C5.7 Buckling of Flat Rectangular Plates Under Shear Loads.

BUCKLING mmER 9ENDUJG LOADS

C5.8 Buckling of Flat Plates Under Bending Loads.

The critical elastic shear buckling stress

tor rlat plates with various boundary

conditions 1s given by the following equation:

n' k s

- - - - - - - (C5,4)

The equation for bending instability of flat plates in bending 1s the same as for compression and shear except the buckling coefficient kb is different from kc or ks' 'dhen a plate in bending buckles, it involves relatively short wave length buckles equal to 2/3 b for long plates with simply supported edges (see Fig. C5.l4). Thus the smaller buckle patterns cause the buckling coefficient kb to be larger than kc or ks.

It buckling occurs at a stress above the proportional lbn!t stress, a plasticity correction must be included and equation C5.4
becomes

t
b

f 1 """';"'~-i-:-=-:-r-=-""T":7''"7T-::-t:=
b

o 0 } =---'---'-----'-

o .L....J:=

rtr.cr = 7(. "n'

"

Fig. C5. 14 (Ref. 7) Bending Buckle Patterns

ks E

12(l-Ve

- - - - - - - (C5.5)
IS.

For bending elastic buckling the equation

Test results campare favorably with the results ot equation C5-.5 i t 77 5 'II Gs/G" where G 1s the- shear mOdulus and Gs the shear secant modulus as obtained tram a shear stress-strain diagram tor the material.

12 (l-Ve ' )

- - - - - --- - (C5.5)

For bending inelastic bUCkling,

A long rectangular ~late SUbjected to pure shear produces internal compress1 ve stresses on planes at 45 degrees with the plate edges and thus these compressive stresses cause the long panel to buckle in patterns at an angle to the plate edges as illustrated in F1g. C5.12, and the buckle patterns have a half

Ocr
12 (1 - Vel

- - - - - - - (C5.7)

wave length or 1.25b.

Where kb is the buckling coefficient and is obtained from Fig. C5.l5 for various alb ratios and edge restraint E against rotation. In the alb ratio the loaded edge is (b). The plast1city reduction factor can be obtained tram Fig. C5.8 USing simply supported edges.
BUCKLING OF FLAT SHEETS UNDER COl'lBINED LQADS

Fig. C5.12 (Ret. 7)

F1g. C5.13 is a chart of non-dimensional shear buckling stress for panels with various edge rotational restraint. This chart fs SDmilar to the chart in Figs. C5.7 and C5.8 in that the values cro.~ and n must be known for the material before the chart can be used to flnd the shear buckling stress.

The practical deSign case i~volving the use or thin sheets usually involves a combined load system, thus the calculation or the bUCkling strength at flat sheets under combined stress systems is necessarJ. The approach used involves the use ot inter-action equations or curves (see Chapter Cl, Art. CI.15 tor explanation at inter-action equa t1 ons )
C5.9 Combined Bending and Longitudinal Compression.

The interaction equation that has been widely used tor combined bending and longi-

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

CS.7

.2

7{
= (Es/E) (l ~

1.0

Fig. C5. 13 (Ref. 1) Chart of Nondimensional Shear Buckling Stress for Panels With Edge. Rotational Restraint.

21e 1I)/(l

~ 1/~.

1$

60

13

52 -

tm:,/Ut
,

"

1/

SYMMCTRIC "-

k.
9

'7i
I

'0
6

. ~~~
1/\
;
I

/)~

)///)

' --.J
, ,
,

tAMPED rDGES

,\'1 A ! .
Y ,\..
~'

,......,.

~r

100

'"
10

i
I

, ,
I
I

60

I
i
I ,

i
,

F l=F

\
,I
1

AN77SYIIMErR/C

Ij

MOD; I

l,lMETRIc i MODel
ZJ

~ ~~
I
i
I
:

':-;--0
,

,.
,

,
I

""""""----1... /~/NGED
2'

s I o

l:

Eo/ES/

l
O/b

I
IS

,
17

I
19

,
,

II

21

Fig. C5.11 (Ref. 1) shear- Buckling-Stress Coefficient of Plates as a Function of a/b for Clamped and Hinged Edges.

Fig. CS.15 BendingSuckling Coefficient of Plates as a Function of alb for Various Amounts of Edge Rotational Restraint.

C5.8

BUCKLING STRENGTH OF FLAT SHEET rn COMPRESSION, SHEAR, BENDING AND UNDER COMBINED STRESS SYSTEMS

tUdinal compression is,

Rb 1
."

l.l

+ Rc

1.0

- - - - - - - - (C5.S)

1.0

Fig. C5.16 Combined Bending & Shear.

This equation was originally presented in Ref. 2 and the interaction curve tram plotting this equation Is found in many of the structures manuals of aerospace companies. Fig. C5.l5 Is a plot or eq. C5.S. It also shows curves for various margin of safety values.
,Q

.9

R;"'R;-l

.8

.7

.J'
"10'

'"

.6

.5

1.0

9
.8

Fig. C5.15 Combined Bending & Long. Compression Rt,1.70+ ftc '" 1

.3

.2
.1

5H
a

.7m!
.6

.1

.2

.3

.4

.5

.6

.7.8.91.01.1

:Yo

B.s.

(Hs + RsU. Rst

.3

Flg.C5.17 is a plot of equation C5.1l It the direct stress is tenSion, it is included on the figure as negative compression USing the compression allowable
CS. 12 Combined .compresSion, Bending &: Shear.

2
.1

.1

.2

.3

.4

.5

.6

.7

.8

.9

1.0 1.1

R"
C5.10 Combined Bending & Shear.

The interaction equation tor this combined loading (Rer. 1 & 2) IS, (C5.9 ) '!'he expression tor margin at satety IS,
!l.S.

From Ret. 5, the conditions for buckling are represented by the interaction curves ot Fig. C5.l8. This figure tells whether the sheet will buckle or not but will not give the margin at safety. Given the ratios Rc, Rs and Rb:- it the value at the Rc curve detined by the given value at Rb and Rs Is greater numerically than the given value ot Rc, then the panel will buckle.
.9
.8

=-;:::~l==", _

v'Ri+R;

(C5.l0)

"0.0
.7

.1 .2,

Fig. C5.16 is a plot at equation C5.9. Curves Showing various M.S. values are also shown. Rs Is the stress ratIo due to torsional shear stresS and Rat is the stress ratio tor transverse or flexural shear stress.
CS. 11 Combined Shear and Longitudinal Direct Stress. (TenSion or Compression.)

.8
Rb .5

..
.3

.5
.6 .7

.3

.8
.9

.2

The interaction equation is (Ret. 3,4) ,

.1

.95

!1. + R3

= l.0
- 1

(cs.rr:
(C5.l2)

o .975
o

.1

.2

.3

.4

.5

.6

.7

Fig. C5. 18 (Ref. 5)

Rs

.8

.9

1.0

..

~.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C5.9

1.8 1.6 1.4

1.2

'" "
0

;; 1.0
.8 .6 .4 .2

","

2.0 1.8 1.6 1.4

RL Longitudinal Tension

.2

.4

.6

.8

1.2

Rt. Longitudinal Compression

11

.,

..

..",

..

",

..

.J

.J

1.0

Flg. C5. 19 (Ref. 5) Combined Compression, Bending and Shear.

C5.10

BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR, BENDING AND UNDER COMBINED STRESS SYSTEMS

The margin of safety of elastically bUCKlsd flat panels may be determined from Fig. C5.19. The dashed lines indicate a typical application where R~ a .161, Rs = .23, and Rb = .38. Point 1 is first determined for the specific value of Rs and Rb. The dashed diagonal line fram the origin 0 thro~gh paint 1, intersecting the related Rc/Rs curve at point 2, yields the allowable shear and bending stresSeS for the desired margin at safety calculations. (Note when Rc Is less than Rs use the right hal! of the figure; in other casas use the left halt).
C5. 13 illustrative Problems.

centroid of the cross-section, a ~ending moment is prOduced about the x-x axis equal to 1400 x 3.7 = 5170 in. lb. = M X I where 3.7 is distance from load P1 to x-x axis.
P

P,

Area of Zee Stringer'" . 18 Area of Corner Member '" 0.25 sq. in.

In general a structural component composed or stiffened sheet panels will not fail when buckling ot the sheet panels occurs since the stiffening units can usually continue to carry more loading before they tail. However, there are many design situations which require that initial buckling of sheet panels satisfy certain design speCifications. For example, the top skin on a low wing passenger airplane should not buckle under accelerations due to air gusts which occur in normal every day flying thus preventing passengers fram observing wing skin buckling in normal flying conditions. Another example would be that no buckling of fuselage skin panels should occur while airplane Is on ground with tull load aboard in order to prevent publiC from observing buckling at fuselage skin. In many airplanes, tuel tanks are built integral with the Wing or tuselage, thus to eliminate the chances of leakage developing, it is best to design that no buckling of sheet panels that bound the fuel tanks occur in flying and Landing conditions. In some cases aerodynamic or rigidity requirements may dictate no buckling Of sheet panels. To insure that bUCkling will not occur under certain load requirements, it is gOOd practice to be conservative in selecting or calculating the boundary restraints of the sheet panels. Problem 1. Fig. CS.20 shows a portion of a cantilever wing composed of sheet, stiffeners and ribs. The problem is to determine whether skin panels marked (Al, (8) and (Cl will buckle under the various given load cases. The sheet material 1s aluminum alloy ~ . Load Case 1.

Fig. C5.20

The sheet thicknesses, stiffener areas and all necessary dimensions are shown on Fig. C5.20. The total cross-sectional area of beam section including all skin and stringers is 3.73 sq. in. The moment of inertia about x-x centrolda1 axis calculates to be 49.30 in.. Since the beam section is symmetrical, the top panels A, B and C are subjected to the same stress under the P1 load system. CompreSSive stress due to transferring loads P1 to centroid of beam cross-section is,

CompreSSive stress due to constant moment of 5170 in. lbs. is,

bendip~

P1

=700

lb.,

P~

= 0,

p~

=0

!c = MxZ/l x = 5170 x 4.233/49.30 = 444 pSi

Total t

With only loads P1 acting, the one cell stiffened cantilever beam is subjected to a compressive axial load of 2 x 700 = 1400 lb. Since the P1 loads are not acting through the

: 375

444 : 819 psi.

The skin panels are subjected to compression as shown in Fig. a. The boundary edge conditions given by the longitudinal st1~~'~~~a

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CS.l1

and the rib flanges will be conservatively assumed as Simply supported. (Fcer is same as O'er)
Feer =
na k E (~)' e 12 (l-V ') b

The Rc + R~ = .431 + .309:a = .526. Since the result is less than 1.0, no buckling occurs.
2

(See Eq. C5.1)

a/b of skin panel

= 15/5

=3
FIg. (a)

.431 +

.431 a + 4 x .309 a

- 1

= .69

Load Case 3.
Flo

From Fig. C5.2 for Case C, we read k c- = 4..0

F

=700,

Fa = 500, P3 :;: 100 lb.

eer-

n'

X 4.0 X 10,700,000 (.035 ) '

12 (1 - 0.3')

= 1900

psi

Since Feer the buckling stress is less than the applied stress f c , the panels will not buckle.
M.S.

= (Fe er/f e ) -1 . = (1900/819) -1 = 1.32

Plo :;: 700 lb., Fa

The two loads P3 prOduce bending and flexural shear on the beam. The bending moment prOduces a ditterent end compressive stress on the three sheet panelS since the bending moment is not constant over the panel moment. To simplify we will take average bending moment on the panel.
Mx(av)

Load Case 2.

= 200

52.5

= 10500

in. lb. X

=500,

P3 :;:

f e due to thiS bending 4.233/49.3 = 903 pat ,

=MxZ/Ix =10500
psi.

The two loads Fa acting in opposite directions produce a couple or a torsional moment of 500 x 16.5 = 8250 in. lb. on the beam structure, which means we have added a pure shear stress system to the compressive stress system of Case 1 loading. -") The shear stress in the top panels A, B and Cis,
fS

Total t c

=903 + 819 = 1722

Re

= fe/F eer =1722/1900 = .906

The two loads P3 produce a traverse shear load V = 200 lb. The flexural shear stress must be added to the torsional shear stress as !ound in Case 2 loading. Due to symmetry of beam section and P3 loading the shear flow q at midpoint of sheet panel (8) is zero. We will thus start at this pOint and go clockwise around cell. The shear flow equation (see Chapter A15) is,
.20 T/ 7.20
10.94 1.50/,

=T/2At = 8250/2

X 138 X .035

= 854

(Where A 1s the cell inclosed area) The shear buckling stress 1s

Fs

pai ,

n.:a k

er 12

(l-V e ' )

- (See Eq. C5.4)

i
q
q

(1 -

a/b = 15/5 = 3. From Fig. C5.11, for hinged or Simply supported edges, we read k s = 5.8.

!~5~5~

3.69 4.23 ; ; ; ;
12.34

f 3)1///
Fig. (b)

= ..!!- ZZA Ix

200 - 4.05 ZZA - 4.93 -

The sheet panels are now loaded in combined compression and shear so the interaction equation must be used. From Art. C5.12 the interaction equation is R c + R; = 1.
Re

=0

(Refer to Fig. b)

= - 4.05 x 2.5 x .035 x 4..23 = - 1.50

- - 1.50 - 4.05 x .18 x 3.69 = - 4.20

= felF eer = 819/1900 = .431

= - 4.20 - 4.05 x 5 x .035 x 4.23 = 7.20

- - 7.20-4.05x .25x3.69'" 10.94

Rs : fs/F s cr = 854/2760 : .309

= - 10.94-'1.05x.051x3.69x3.S9/2 = -12.34
(See Fig. b tor plo t at shear flow)

CS.12

BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR, BENDING AND UND.'eR COMBINED STRESS SYSTEMS

The shear flow q on panel (A) varies from 4.20 to 7.20 or the average q = (4.2 + 7.2)/2 = 5.7. Thus the average shear stress is 5.7/.035 = 163 pSi. It is in the sa~e dIrection as the torsional shear flow and thus is additive. Total f s

(4)

A sheet panel 5" x 12.5" x .051 has all edges s1~ply supported. The panel 15 subjected to c8mbined compreSSion and shear loads which produce the following stresses:-

= 163

+ 854

= 917

psi

t c = 2400 PSi, applied normal to 5 11 side.

fa ps i . 'will the sheet buckle the given load system if made of aluminum alloy 2024-T3 mater tat . '.-ihat is the margin of safety.
l~der

Rs = fs/F scr

= 917/2760 = .332
11

= 2800

Rc

Subt.:- .906 + .332 :: l.016, since the result Is greater than 1.0, initial b~ckling has started. The margln of safety is slightly negative and equals,
+

R; = 1,

(5)

- .01 (6)

If the material in problem (4) is changed to alloy steel Ft u = 95000 psl, what would be the margin of sarety. It sheet was heat treated to Ftu = 180,000, what would be the M.S.

In this example problem, the panels were assumed simply supported, which 15 conservative. Reference to Fig. CS.6 shows that kc could be assumed higher as the panel Is riveted to a Zee Shaped stringer whlch has same torsional resistance and thUS panel is not free to rotate at its boundaries. Panel (C) is less critical because the flexural shear is acting OPPOSite to the torsional shear stress, thus f s total = 845 163 = 682 psi. Ths Rs = 682/2760 = .246. Re +- R .9OS + .246 .9S6. S1nce the resUlt 1s less than 1.0, panel w1ll not buckle. Panel (B) carr1es a small shear flow, being zero at center of panel and increasing uniformly to 1.5 lb. per inch at the edges, and tlowing in opposite d1rections trom the centerline. Thus transverse shear will have nsgligible errect. Thus Rs 854/2760 .309.

A 3" x 12" x .040" sheet panel is subJected to the following combined stresses. t c = 3000, fb = 10000, f s = 8000. The f c and fb stresses are normal to the 3" side. If sides are simply supported, will panel buckle if made ot 707S-T6 aluminum alloy. What 15 M.5. What will be the M.S. if material is Changed to Titanium Ti-eMn.

References:(1) (2 )
(3 )
(4)

s=

11

NACA Tech. Note 3781, 1957. ANC-5 Amendment 2. NACA AHILNo. L6A05 NACA ARR.No. 3K13 ANC-S Revision or 1942 General References on Theory Aug. 1946.

(5 )

(B) is on the verge or buckling under the

assumptions made in the solution.
PROBLEMS

Rc

+-

R = .9OS

.309 11 = 1.00, or panel (6) (7)

NACA Tech. Note 3781. "Introduction to Structural Stability Theory". By Gerard. Book publiShed by McGraw-Hill. "Theory at Elastic Stability", by Timoshenko, McGraw-Hill Co.
~A Unified Theory Plastic Buckling of Columns and Plates", MACA Report 898. By Stowell.
~Plastic Buckling at SLmply Supported Compressed Plates~, NACA T.N. 1817. By Prlde & Heimerl.

(1)

A sheet panel 15 3" X 9~ x .051" in size. The 3" side is simply supported and the 9" side is tree. Determine the buckling load it the compressive load 1s applied normal to the 3" sides. Do so for 3 different materials, (1) aluminum alloy 7075-T6, (2) magnesium llK3lA, (3) Titanium Ti-8Ml:t. In Problem (1) it all edges were Simply supported, what would be the buckling load. In Problem (1) if the 9" sides were clamped or r txec and the 3" sides simply supported, what would be the buckling load.

(8) (9)

of

(2)

(10)

(3)

(11)

"Buckling of Metal Structures" by Slelgh. McGraw-Hill Co.

,,.

,~

-,

'

','

~"'

-.

"

-'< ':. -,

".~-",.,,:

CHAPTER C6

LOCAL BUCKLING STRESS FOR COMPOSITE SHAPES

C6.1 Introduction.

Thin flat sheet 1s inefficient for carrying compressive loads because the buckling stresses are relatively low. However, this weakness or fault can be greatly tmproved by forming the flat sheet into composite Shapes such as From Fig. C5.2 of Chapter CS, kc .43, angles, cnanne.Ls, zees , etc. Most of the many then composite sr~pes can also be made by the extruding process. Formed or extruded members are Widely used in Flight Vehicle Structures, Ocr - - - C6.1 thus methods of calculating the compressive strength of such members is necessary. If the buckling streSSes are above the proportional limit stress, use Fig. C5.7 in .ce.2 Compressive Buckling Stress for Equal Chapter C5 to take care of the plasticity effec~ Flanged Elements.

From Equation C5.1 or Chapter C5, the bUCkling compressive stress for a long flange IS,

'"'I

The stmplest equal flanged member that can be formed is the angle shape. Other Shapes With equal flanges are the T section and the cruciform section as shown in Fig. C6.l.

For formed angles, the flange w1dth b extends to centerline of adjacent leg, but for extruded angles, the width 'b extends to insIde edge of the adjacent flange or leg.

~e~~~
Angle

ca.3

CompresSive Buckling Stress for Simple Flange-Web Elements.

T-Section

c~~~~~m

Fig. C6.1

These sections can be considered as a group of long flanges, as illustrated, for the angle section in Fig. C6.2. Since the flanges which make up the section are equal in size, each flange will buckle at the same stress. Therefore each flange cannot restrain the other and thus it can be assumed that each flange is Simply supported along the flange junction as illustrated in Fig. C6.2.

The most common flange-web structural shapes are channels, zees , and hat sections. A flange has one unloaded edge free, whereas a web has no free unloaded edge and thus has an unknown restraint on the 'boundary between the web and the flange. Fig. CG.3 shows the break down of a Z sectien into two flange and one web plate elements.

III ~I!"
"'I

II :

~.L J.,I,

SS

~
ra.

~ Fl".....e

~
:::
Q

,,,
58
Web

]I

--

" '"
-a;

".
58

-'l Flang

Ocr

1 I 1
55 55

..

~
55
f

..
I-b-l
1 1

r"'t1.t"'..,

58 10" \, 58 58 f f , T T T T T Breakdown of Z Into Flange and Web Elements.

Fig. CG.3

r;;

Fig. C8.2

The buckling strength of the web and flange elements dependS on the boundary restraint between the two elements. If this restraint which is unknown could be found in terms of a known rotatIonal restraint E as presented In Chapter C5, the buckling coefficIents could be found from charts in Chapter CS. Having the
C6.1

C6.2

LOCAL

BUCKLING STRESS FOR COMPOSITE SHAPES

buckling stress for each element, the critical buckling stress will be the smaller or the ~NO. The buckling load based on the buckling stress IS not the failing load as more load can be taken by the material in the corner.reglons before local failure or crippling takes place. The SUbject ot local crippling of formed and extruded Shapes IS covered in Chapter C7. USing the moment distribution method or a step by step analysiS procedure, several research studies have determined the restraint factors between web and flange elements tor simple Shapes lIke channels, Z, H, square tubes and formulated design charts for such Shapes. (Rets. 1 to 5 inclusive.)
C8.4 Design Charts for Local Buckling Stresses of

For the plasticity correction of Shapes covered in Fig. C6.4 (Ret. 4), the plasticity correction tor a flange free on one edge can be used with accuracy. Thus we can use chart in Fig. CS.7 of Chapter CS.to ccrrect for plasticity eftects. The parameter for bottom scale of Fig. CS.7 is equation (Aj divided by cro.~, or 56100/39000 = 1.44. Using this value and the n = 11.5 curve, we read from Fig. CS.7 that ocr/a 0." = 1.02. Therefore the local buckling stress is
crcr

= 39000

x 1.02

= 39800

psi.

PROBLEM 2.

Some Composite Web- Flange Shapes.

F1gs. C6.4 to C6.7 1nclus1ve g1ve charts tor determining the local buckling stress ot channel, Z, H, square tube and hat shaped sections. For tormed sect10ns, the width b extends to centerl1ne ot adjacent element and tor extrUded sections the width b extends to inside edge at adjacent element.
CO. l5 Problems illustrating Use of Charts.

If the member in Problem 1 is SUbjected to a 300F temperature for 2 hours duration, what would be the local buckling stress. From Table Bl.1 tor this temperature condition, 00 .
"

= 35700,

Ec = 10,300,000,

15

= 2.9 n' x
1.51.

10,300,000 12(1-.3') (35700)

!'ROBrm 1.

The Z section in Fig. (a)

1s tor.med trom aluminum alloy 2024-T3 sheet. What ccnpreeslve stress will start local buckling ot an element ot the member. Solution:-

USing this value on bottom scale of Fig. CS.7 and n curve = 15, we read ocr/O 0 ~ :: 1. 03.
Thus crcr
Fig. (a)

= 1. 03

x 35700

=~

pe t ,

PROBrm 3.

llw

br = 0.75 - .032 = 0.718 brlbw =0.718/1.436 = .50 tw/tr = .064/.064 =1.0

From Fig. C6.4, we read kw = 2.9
- - - - - (AI

= 1.5 -

.064

=1.4:36

Same as Problem 1, but change material to

Titanium TI-SMn Sheet.

Ec

= 15,500,000,
= 2.9

From Table BI.l we obtain tor this material:cro.~ = 119500, n = 13.7.

n' x 15,500,000 ( ,064)' 12(1-.3') 1.436

crcr

81200 i _ _ ps .

Th1s stress is near the proportional limit stress so plasticity correction should be small
it

any.

Ocr/a 0 ~ This stress is above the proportional limit stress at the material, thus a plasticity correction must be made. The buckling occurs on the tlange. From Table Bl.l at Chapter Bl, we abtain tor 2024-T3 alum1num alloy:- oo.~ = 39000 and n = 11.5.

= 81200/119500 = .68

From Fig. CS.7, USing n = 13.7 curve, we read ocr/ao.~ = .68. Then ocr. 119500 x 0.68 = 81300, thus no plastic1ty correctlon_
PROBLE!'! 4.

The rectangular tube has the dimensions as shown in Fig. (b). It is extruded from aluminum alloy 2014-T6. Determine the local compressive buckling stress.

r-----,=------,-=----;------,---,.---,
7

...

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

ce.s

h'---'-.-=""+--\-{----\t-.nAN6 8UCJa.D
I

I'IIfST

I
k.
3

I i

K",

v t,

f--+-'l:ttiA''M-+'c---'<-+>,--I-----i

{}
WEB."

,03

..
7

L--,
0 0

ia 'J

12
.2 .4

.2

)Z016j
6
~

10

'w
twa
lil

Fig. ca.4 (Ref. 2) Channel- and z-eecttcn stiHeners.

kwn:a E Ocr'"

Fig.

ce, 5 (Ref.

2) H-sectton stiffeners.
kwnllE

12 (1- Vea, bw

tw:.
)

o cr '" --'-_-ll
12 (1 - V e

bw'i

,------,---r_-r_-----,-,---,----,

IC""
3
,

s 1--i--+'r''ri'~~:_~-+-+--1

t.g
I I
6

to

'f----------\-+--'<--'-'>,,--",,-"'*'o--f----1

..
'0
0

o0 "

'"

to
--

B~
','

~'#f

.
i

,
,.

't/b

'"

<Z

,.

Fig. CG.6 (Re!. 2) Rectangular-tube-section stiffeners.

cr cr '"

KhTt2E

(t
-

Fig. CG.7 (Ref. 5) Buckling stress for hat-section stiffeners.

kTnllE
t '" tf '" t w '" tT; Ocr=' t

h) '

12(1- Ve 41)

12 (1 - l.Ie ll )

- . (Data of Ref. 12.) bTa

~;-.1 : ~~ ...~.>.~-~- ,c'.~ -~~~:, ~;.,.>'~.i::--~'(.,.:\:~'.~;~:::~;, . .,.":'-~ .. ~ ....,:;z~~

--

r:

';":. )' . -': .: -:-:.. _~~,~.~'~',

---~

:.~.;..;~

.--

C8.4

LOCAL

BUCKLING STRESS FOR COMPOSITE SHAPES

Solutlon:b

.08 = .92

~fT
I'
2.0

b/h tb/th

= 2 .08 = 1.92 = .92/1.92 = .479

= 1.0

~-':IO

'I

In the deSign at rectangular tUbes J the designer should select the tube thicknesses for both long and short sides so that buckling occurs on both Sides, thus giVing the lightest section tor buckling strength. As pOinted out previously in thiS chapter, the load on the member which causes local buckling is not the failing or maximum load for a short length of the member. This local failing or crippling stress is treated in the next chapter. Since the buckling stress may fall in the inelastic stress range, the buckles will not entirely disappear when load is removed. Since 11mit or applied loads must be Carried without permanent distortion, it is thus important to know when local buckling starts. For those missile and space vehicles that carry no humans J the factor at safety on limit loads is conSiderably less than for aircratt J thus the spread between local buckling and local failing strength becomes Important in design.

Fig. (b)

From Fig. Cl.S, we read kh = 5.2.

"cr

5.2 nil x 10,700,000 (...:.2.)' 12(1-.3') 1.92

= 21900 ps t

As shown in Fig. Cl.S, buckling occurs on the h side of the tube. The camputed buckling stress IS below the proportional l1mit stress, thus no plasticity correction.
PROBLEM 5.

same as Problem 4 but change the thickness of the h side to .072, but leave the b side .04 in thickness. Solutlon:b
h

ca.5

Buckling of Stiffened Flat Sheets Under Longitudtnal Compresston.

=1 =

.144

= .856
=
.555

bill

2 - .08 = 1.92 = .856/1.92 = .446

tb/%

.04/.072

From Fig. C6.6, kh

O"cr=

= 4.3
= 58600
ps t .

In supersonic aircraft, it is important that the surface skin J partiCUlarly that on the wing, not buckle under flight conditions since a buckled surface could etfect the aerodynamic characteristics of the airflow around the wing, thus it i s important to know when the skin or its stiffening units initially buckle in order to deafgn so 'tna't SUCu Juckl1ng will not occur under flight conditions. Gallaher and Boughan (Ref. 6) and Boughan and Baab (Ref. 7) determined the local bUCkling coefficients for idealized web J Z and T stiffened plates. The results of their stUdies are shown in Figures C6. 8 to C6.12 and were taken from (Ret. 4). The initial local "buckling stress for plate or stiffener is given by the equation:-

4.3 n' x 10,700,000 (~)' 12(1-.3 II} 1.92

This stress 15 above the proportional limit stress thus a plasticity correction Is necessary (Ret. 4) gives no value tor a plasticity correction but recommends the correction for a
clamped long flange which Is a slightly con-

.. ;

servative correction. This plasticity correction should also be used tor the hat shape as shown in Fig. C6.7
Thus Fig. C5. 8 ot Chapter C5 can be used to correct tor etfect of plastiCity. From Table Bl.l, for our material a 53000 and n = 18.5.
Cl T

- - - - - - - - - - (a)
It the buckling stress is above the pro- "'. portional limit stress of the material, correct! tor plastiCity effect by USing Fig. C5.8 or j Chapter C5.
I

Problem Illustrating Use of Charts. The value of the parameter ro .. ~ bottom scale of Fig. C5.8 15 ocr/at,. T = 58600/53000
1.10.

Fig. C shows a plate with idealized Z section stiffeners. The material 1s 2024-T3
tf=.0625
tw

From FIg. C5.8, using n = 18.5 curve J we read ac~/a Cl. T = .9l. Therefore ocr = 53000 x
.91

...

/b"

e ,

48100.

=.0825
ts

Thus by Changing the long side of tube trom .04 to .072 J the buckling stress was increased tram 21900 to 48100 pe t ,

= 125

.... 4.0 Fig. (c)

.1

==-- -

--~- ~--

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C6.5

.
t

-.
"s
4

k,
3

,.

,.

'~7.,
'.
4

00

.. L..1..:.f
.2

.-

"

!!

,~

j
6
b,//)~

a to

"

Fig. C6. 9 (Ref. 6) Z-section stiffeners.

twits" O. 50 and 0.79.

10

"

"
,

Web stiffeners. 0.5 < twits < 2. 0

Fig. C6.8 (Ref. 7) Compressive-local-buckling coefficients for infinitely wide idealized stiffened flat plates.

Ocr""

_.naE

12 (1'" tie a) b

(t,,~' -

" s f----'--.....l---'-+-'r'f'r-l '

e

..
~

"

"

Fig. C6. 10 (Ref. 6) Z-section stiffeners. twits ,. O.63 and 1. O.

r.

,
2

4,/b,
3

ie

:1:U.
~
\ t.

l ',~ ,

'. '

"

Fig. C6.11 (Ref. 7) T-section sttffeners.

tw/tt

'" 10; _ Ow, . 0.25. 1.0; _::tf bs

Flg.

ca. 12 (Ref. 7) T-section stiffeners. twltf '" 0.7; bf/tr ,. 10; bw/bs :> O.25.

-'.

~,~>. ~ ~ ~:~
:

: : ,,'".

'.

::.~ ~

'. :

~> .:::: ,"~ ~: \::.:;.::.~.:~ ::-.' ,:

.: -

-~~~ ~-:_~.

. .: , <:

=-.c'-~~7-~ r,~_ ':cO", ~=

ca. a

LOCAL

BUCKLING STRESS

FOR COMPOSITE SHAPES

Re.r~rer.ces

aluminum alloy. Determine the initial buckling stress under longitudinal compreSSion.
(1 )

OW tw

br _ 0.5

r.s

7'7 .~,

Ow _ 1.5 - 375 , -----. bs 4.0

(2)

LundqUist, Stowell and Schuette: Princi~les of Moment Distribution Applied ts Stability of Structures composed of Bars ar.d Plates.
NACA WRL-326 , 1943.

15"

0.125

.0625 =

50

USing the above three values and referring

to Fig. CS.9, we read the buckling coefficient kg to be 4.2. Substituting in equation (B)

Kroll, Fisher and Heimerl: Charts for Calculation of the Critical Stress for Local Instability of Columns with I, Z, channel and Rectangular Tube Sections. NACA
WRL-429, 1943.

= 4.2 n x 10,700,000 (0125)3 = 39800. 1

12(1-.3') 4.0
ps

(3)
.

Kroll: Tables of Stiffness and Carry-over Factor for Flat Rectangular Plates Under Compression: NACA WRL-39B~ 1943. Becker: Handbook of Structural Stability. Part II. Buckling of compOSite Elements. NACA TN.3782, July 1957. Van Der Maas: Charts for the Calculation of the Crt tical Compress! ve Stress for Local Instability of Columns with Hat Sections. Jour. Aero. Sci. Vol. 21, June
1954.

This stress 1s no doubt above the proportional limit stress so a check for plastiCity etfect will be made. For this effect we use Fig. C5.8 of Chapter C5.
For our 2024-T3 rnat.er-taj , we r inc from Table Bl.1 ot Chapter Bl, that ocr/croo7 = 39000 and the shape parameter n = 11.5.

(4)

(5)

The bottom scale parameter on Fig. C5.S 1s equation (B) divided by 00 7J thus it equals 3980D/39000 = 1.02. Using this value and the n = 11.5 curve on Fig. C5.8, we read on lett side scale that Gcr/Go . ? = .86. Therefore the buckling stress ocr - .86 x 39000 33600 pSi.

(6)

Gallaher and Boughan: A Method for Calculating the compressive Strength ot Z Stiffened Panels that Develop Local Instability. NACA TN.l482, 1947. Boughan and Baab: Charts for calculating the Critical Compressive Stress for Local Instability of Idealized T Stiffened Panels. NACA WRL-204, 1944.

(7)

--_.-~-"

--~--.;:::-

CHAPTER C7

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PAL'IELS IN COMPRESSION SHEET EFFECTIVE WIDTHS.
C7. 1 Introduction.

COLUMN STRENGTH.
presented in (Ref. 1). In this me thod the member section is divided into equal or unequal angles as illustrated in Fig. C7.2. The strength of these angle elements can be established by theory or tests. The ultimate strength or failing strength can then be found by adding up the strengths of the angle elements that make up the composite section.

Chapter C6 was concerned with the local buckling stress of composite sections when loaded in compresSion. Tests of short lengthS of sections composed of flange-plate elements often show that after the section has buckled locally, the untt still has the ability to carry a greater load before failure occurs. In. other wordS, the local buckling and local Needham made a large number at tests on angle and channel sections. From a atudy of failure loads are not the same. For cases these test resu~s well as other published where local buckling occurs at low stress, the test data on ChannelS, square and rectangular crippling or failing stress will be higher. When local buckling occurs at high stress such tubes, etc., he arrived at the following equation for the crippling or failing stress as .7 to .8 Fey, buckling and crippling stress are practically the same. Fig. C7.1 illustrates ot angle sections. the stress distribution on the cross-section after local buckling has occurred but prior to Angle. I Angle 4 such b .1- ~ unit local crippling or failure~ units

~
---L

Basic Angle Unit. Two edges free.

FIg.

f:
f

FIg. a

T,

~ Fi~.

rr=F4

-ttijf
Basic 1Angle Unit. No edge free.

Basic Angle Unit. One edge free. Fig. C7.2

C7.i

l(b')".111 () IF " Fes_eyo,=Crt - - - - - - - C7.1

where, As the load on the section is increased, the buckles on the flat Dortions get larger but most of the increasing load 1s transferred to the much stifter corner regions until the stress intensity reaches a high enough value to cause sufficient deformation to cause failure.
Fcs
F cy

c,.:....> '-

= crippling stresS (psi) = compression yield stress (pst) = Young's mOdulus or elasticity in

compression (psi) b'/t = equivalent bit at section =

Ce

(a b)/2t = coefficient that dependS on the

degree of edge support along the A theoretical solution for the local edges of contiguous angle units. crippling stress for all types of sr~pes r~s Specifically they are:not been developed as the boundar/ restraint Ce ~ 0.316 (two edges tree) between flange and plate elements is unknown C~ = 0.342 (one edge tree) and also the manner in which the stress builds C; = 0.366 (no edge tree) UD in the corner regions is not well lli~derstaod. ConsequentlY, the methods ot solution are semi~he crippling stress for angles, c~~~~els, empirical in character, and the results af such ze8S and rectangular tubes can be determined methods have been sufficiently ?roven by tests. directly from use of equation C7.1. The ~o methods of calculating crippling stresses crippling loac on ~~ angle unit is then, will be presented in this chapter. C7.2 METHOD 1. THE ANGLE METHOD, or
NeedhaIn Method.
the

(C7.2)

where A is the area of the angle .. This method which will be refe~ed to as the angle method or the Needham method was C7.i The crippling stress of other formed

C7.2

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COM RESSION

(a)
(b)

structural shapes can be determined by dividing the Shape into a series of angle units and computing the crippling loads for these individual angle units by use of equations C7.1 and C7.2. The weighted crippling stress for the entire section is obtained from the following equation:F = Z (crippling loads of angles)_ (C7.3) cs Z (area of angles)
C7. 3 Design Curves.

Fig. C7.3 gives curves tor determining the crippling stress of angle units as per equation C7.l, and Fig. C7.4 gives curves tor determining the crippling loads tor angle units. USing these curves and equation C7.3, the crippling stress at compOSite shapes other than angles, channels, zees and rectangular tubes can readily be calculated. Illustrative problen4~, uSing this method, will be g1ven later and the results compared With method 2. Crippling stresses for compos1te sections should be limited to the values g1ven in Table C7.l unless substantiated by test results.
C7.4 METHOD 2. For Crippling stress Calculation. (The Gerard WelJlod).

Fig. C7. 5 stresses and displacements of flat plates after buckling under conditions of uniform end shortening (reference 4). (a) straight unloaded edges, (b) stress free unloaded edges free to warp in the plane of the plate. C7.6 The Gerard Equations for Crippling Stress.

The following equations are taken from

(Ref. 3).

For sections with distorted unloaded edges as angles, tUbes, V groove plates, multi-corner sections and stiffened panelS, the following crippling stress equation applies within: 10 percent limits:Fcs/FCY = 0.56\
[ Igt'/A)

IntrOduction: Reterences 2 and 3 gives the results ot a very comprehensive study by Gerard on the subject of crippling stresses. From a thorough stUdy at published theoretical studies and most available test or experimental results, Gerard has developed and presented a more generalized or broader, semi-empirical method of determining crippling stresses. In one sense it is generalization or broader application at the Needham method which was presented as method 1. The student and practicing structures engineer should refer to the above references tor a complete discussion of how the resulting crippling stress equations were obtained and how these check the extensive test results. In this short chapter we can only present the resulting equations, design curves for same ~~d example problems in the use at the information, in the determination of crippling stresses.
C7. 5 Stre8ses and Displacements of nat Plates After Buckling Under Conditions of Uniform End Shortening.

(ElF ey)' ,

/ ] e

- - (C7.4)

For sections with straight unloaded edges such as plates, tee, c ruc t form and H sections, the follOwing equation tor crippling stress applies within 5 percent limits.

Fcs/FC Y = 0.. 67

f. ~A)(ElFey)'/ , ]"" - - (C7.S) L(gt

For 2 corner sections, Z, J, and channel sections, the fOllowing equation applies within ~IO percent limits.
Fes/Fey = 3.2 [It

~A)(ElFey)';,r'" -

- (C7.6)

Fig. C7.5 shows a picture at the resulting stress distribution on flat plates after buckling under conditions of uniform end Shortening as determined by Coan in (Ref. 4). The Gerard method recognizes the effect of distortion at the tree unloaded edges upon the tailing strength of the member section.

Fcs Fcy t A E g

= criPpling stress for section (pSi) = compressive yield stress (pSi) = element thickness (inches) = section area (in. a ) = Young's modulus of elasticity = number of tlanges which compose the compOSite section, plus the number at cuts necessary to divide the section into a series of flanges. See Fig. C7.6 tor method of cutting ComDOSite sections to determine value of g.

The cut-oft or maximum criPpling stress F~s tor a composite section should be limited to the

. - ~ - - -.21 -

-'

2.0
/

18

0.09
--

16

--<

!.L.

~2P\ ,
i":::
- r-

ONE EDGE

F\~TWO EDGES

fREE _ FREE _

0.08

r-

-l
I

l'

- r- .

'..L
-,L

\
Pce

jz /j7' '
r--

NO [DuE

fREEiL

60
0.01

14

I 40

t0.06

~~tb~
I

>
Z

o
H

12

V"E ~-'l/l(.Y'-I.o
o.

t--

- r- t-r--

120

ee
Z

r=

I -I I00

0.05

\\1
1\

.~

/'
~-

004

o
6o

t\\ 1'0
,~

o.o~

o.6

o4
o2

'5 b-J
a

~~

, ,nn,

.. .
o
.fJl

"
r-

'"

"
T

R;;3t

0.02

,
\ \
I

'<~ ""I
-.

r:::

.,
<
o

.'" ..

40
1

'"
H

r-

20

'"
50 60
70

10

<0

30

40

80

10

20

.30

40

Ji_ .<'ctl> < 2t

30

60

70

80

(Nole: Fcc same as Fcel

Fig. C7.3 Dtmenalonleea Crippling Stress VB. b'/t (Ref. 1)

(Note: P cc same as PCB) Fig. C'l.4 utmenejonteee Crippling Load VB.

by/t

(Ref. 1)

C'l' 4

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET STIFFENER PANELS

Distorted Unloaded Edges

rn COMPRESSION

C7.7 Correction for Cladding.

Angle

Plate

Tube

L
Basic Section

Cut
1 Cut

, , i ,

-f++,
4 Cuts 8 Flanges

g.2

2 Flanges

Since many formed sections are made from alclad sheets, the clad covering acts to reC:.l.:::e the value of buckling stress and thus a correction factor ~ must be used to take care of this reduction in strength. This correction from (Ref. 3) 1S,

3,.

12

=g

~ =

[1+ 3

(ocl/ocr If ]/(1

3n - - -

(C7 7)

where,
Straight Unloaded Edges
T~Section

Cruciform

T
Baste Section

+
o Cuts
4,.

H-Sectton

rut
1 Cut

O'cl ... cladding yield stress Ocr = bUCkling stress t = ratio of total cladding thickness to total thickness. t = 0.10 for 3.1c1ac 2024-T3 and .08 ror- a Lc Iad 7075-T3.
--The cut-ott or maximum crippling stress fo ra composite section should be limited to the values in Table C7.l unless test results are obtained to substantiate the use of higher crippling stresses.

.. 3

4 Flanges

70.

6 Flanges

C7.8 Maximum Values for Crippling Stresses.

Fig. C7.6 Method of cutting simple elements to determine g.

Table C7.1

values in Table C7.1 unless higher values can be Slbstantlated by test results. The cut-ott values given in Table C7.l are no doubt slightly conservative. DeSign curves tor equation C7.4 J 5 and 6 are given in Figs. e7.?, C7.8 and C7.9.

Type of Section Angles

Max. F c s .7 F ev
F ev
.8Fcv

v Groove Plates
Multi-Corner Sections, Including Tubes stiffened Panels Tee, Cruciform and H Sections

F ev

. a Fey
.9 F cy

2 Corner Sections.

Zee, J. Channels

Fig. C7.7 Curve for Crippling Stress F es for Angles, Tubes, V -Grcove Plates, and Multi-Corner Sections.

Fes '" O. 56[(gt <I; A)(E!Fcy)

Fey

1./<1] 0

A.,

See Table C7. 1 for Cut-off Values.

r,;
.2
.2

Fe.

'lEIlI.I1IIIBlIi
.15 .2.25.30 .4 .5.6.7.8.91.0
1. 5

2.5 3

6 7 8 9 10

(...iL) (Fey) ,/. gt B Ee

--_..... ~.

---

--

."

"

.'

,".:".

',

'.,<,;

~.~.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES 1.5

C7.5

FIg. C7.8 Curve for Plates, Tees, erect Form ~ and H Sections. -Fcs : O.S7[(gtli/AHE/Fcy) l./li]O'400 Fey

1.0 .9 .8 .7 .6 .5

~""""~

See Table C7. 1 for Cut-off Values

Fe. Fey

.4
.3 .25 .2

.15
.1
.1

.15

.2 .25.3

.4

.5.6.7.8.91.0

1.5

2 2.5 3

5 6 7 8 910

(;.)(~)'I'

Fig. C7.9 Crippling stress Fcs for 2 Corner Sections, Z, J and Channel Sections.

~C8~", 3. 2[<eIAHEdFcy) ],/~] o.n

cy

See Table C7. 1 for Cut-off Values

1:=
Fey

.2

1.5

2 2.5

'5
t

6 7 8 9 10

(~) li ~)'I'
~

C7.6

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS

rn COMPRESSION

C7.9 Restratnt Produced by Lips and Bulbs.

quite otten in formed sections, the :lange element which has a free edge is rather small in width as illustrated in Fig. a. Also for extruded sections, a bulb is often used as illustrated in Fl~s. b. The question then arises, IS the lIP or bulb SUfficIently large enough to provide a simple support to the adjacent plate element. Since the compressive buckling coefficient for a plate element is

From Chapter C6, the compressive buckling coefficient USing V~ = .3 is 0.388 for a flange element and 3.617 for a plate ele~ent. Therefore, {C7.l0; From equations C7.9 and C7.l0, the following relationship is obtained, - - - - - - - (C7.11)

Lip

Lip

Bulb

Fig.

B~aub
b

Fig. C7.1l shows the results as a curve.

UP PROVIDES AT LEAST Sll<lPlE SUPPORT~OR FLANGE

~.
4.0 and 0.43 for a flange element, the use of a

I I i,,1 P
I

small lIP or bulb can increase the value at the coefficient considerably above 0.43 and thus produce a more efficient load carrying element. The problem of determining the dimensions at a lip or bUlb to give at least a simply supported edge condition to the adjacent plate element has been investigated theoretically by Wlndenburg (Ret. 5). The results at his studies gives the tollowing design criterion.
2.73 btt' - btt

~.

7 l'IUCKLIN." OF UP ~

t,

I I~tt" It I
I

.'.,1>,-",",,',;>";

',>\1.~~ i

:
r

11

'I

71

I
b.

T
Fig. C7. 11 Minimum lip dimensions required for nanga to buckle as simply supported plate (Ref. 5).

--!l....

~~

=5

- - - - - - - (C7.8)

Where IL and AL are the moment or inertia and area ot the 11p or bulb respectively. (See
FIg. C7.10).

In extruded sections, a circula~ bulb IS otten used to stiffen a free edge as illustrated. in Fig. C7.l0b. The moment at inertia of the bulb area about the centerline of the ~late element is,
I nD =- . -no 64 4
4

,ll>- - -;,1
Lip (aj

li-J>L--i

Bulb

D t (--=-.). 2

,bt
I

tr

ss

(1))

As tor the case at the lip, the buckling stress of the bulb must be greater or equal to the buck11ng stress of the adjacent ~late element, which gives,
D), (D, D). (t -1.6 t) - 0.374(t

= 7.44

br/t

- (C7.12)

FIg. C7.10

From Fig. C7.lDa tor the lip, AL = bLt, and IL = tbL 3 /3. In substituting these values in equation (C7.8), the dimensions of the lip are expressed as
0.910 (btL). _ btL = 5 b: (C7.9)

Fig. C7.12 shows deSign curve representing the above equat10n.

(~Mim
30 10 20 IIf 30 40 50

To determine bL and t, an additional reqUirement is specif1ed, namely, that the buckling stress of the lip must be greater or equal to the buckling stress at the adjacent plate element.

Fig. C7. 12 Minimum bulb dimensions required for buckle as simply supported plate (Ref. 5).

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C7.7

C7. 10 ntustrative Problems in Calculating Crippling Stresses.

Solution Method 1. Fcs

.Problem 1. Find the crippling stress for the equal leg angle shown in Fig. a. The material is aluminum. alloy 2024-T3. .S olution by Method 1. Material properties are:1"---lrT ~: .:.105

= Fcc = .033oVI10,000 x 15,500,000

= 43000 psi

Solution Method 2.
( .093 ) ( 110,000 ),/- = 1 565 2 x .05"" 15,500,000

Fey: 40000, Eo

= 10,700,000

1"

16

Ffg.a

Fcs

= 0.38 x

From Fig. C7.7, Fcs/FCY = 0.38, whence 110,000 = 41,800.

Arep.093
Fig. a

Problem 4. Find the crippling stress for the channel section shown in Fig. b if the material is aluminum alloy 2024-T3. Fcy = 40,000, Ee = 10,700,000. Solution Method 1.

For this method, we use Fig. C7.3. The parameter for bottom scale ot Fig. C7.3 Is (a + b)/2t, where (a) and (b) are leg lengths measured to centerline ot adjacent leg ot angle. For our case a = b ; 1 - .025 : 0.975. Thus (a + b)2t = 1.95/0.1 = 19.5. From Fig. C7.3 using 19.5 on lower scale, and the curve tor two edges rvFe, we read ~n the left hand scale that Feel cy~ = .0330. Since we have only one angl~, the crippling stress Fes,- Fcc = .0330xY40,000x10,7QO,OOO
:II

fo--

1.50 ~

pe i ,

Solution by Method 3 (Gerard Method) For angle sections we use equation C7.4.
A plot of this equation is given in Fig. C7.7.

As shown in Fig. c , the channel is composed of 2 equal angle units (1) and (2). Since they are the same size, we need only calculate the failing stress for one angle.

16
Area = . 137
Fig..

FT .75
1..
-oj fo-.05

1- ......

The parameter for lower scale Is, A Fcy l./~ (gta) ( E) , where A ; section area and g equals the number of flanges plus cuts, or g = 2 tor an angle section. SubstitutIng,
.093 )( 40,000 )'/- = 1 1'8 (2 x .05~ 10,700,000 '-'

2t

aeb

= .725+

.725

0.1

= 14.5

From Fig. C7.3 tor bl/t = 14.5, we read Fcc/vFcyE

= .045

(for one ed~e free) : .045 x

~

Then Fcs

= Fce

V40,OOO x 10,700,000 = SOlution Method 2.

psi

Using this value in FIg. C7.7, we read Fcs/FCY = 0.50. Therefore Fcs = 0.50 x 40,000 "" 20,000 PSi. .Problem 2. Same as Problem 1, but change material to aluminum alloy 7075-T6. Fcy = 67,000, .Solution Method 1. Fcs Ec = 10,500,000

For a channel section we use equation C7.6 which 1s plotted on Fig. C7.9. The pa~eter for bottom scale of Fig. C7.9 ls,

= Fcc

= .0330./67000 x 10,500,000

= 28,000

From Fig. C7.9 Fcs/FCY = .65, whence, Fcs = .65 x 40,000 = 26,000 pSi. psi Problem 5. Same as Problem 4 but change material to aluminum alloy 7075-T6. Fey = 67,000, Ec = 10,500,000. SOl'ltion by Method 1. Fcs = .045 x J67,OOO x 10,500,000 = 38200 psi SOlution by Method 2.
(.137)(
~

Solution by Method 2.
( .093 )( 67,000 ),/- = 1.485 2 x .05 4 10,500,000

From Fig. C7.7, Fcs/F cy; 0.40, whence Fcs = .40 x 67,000 ~ psi. Problem 3. Same as Problem 1 but change material to Titanium Ti-~J1. Fcy = 110,000, Ec = 15,500,000.

67,000 ),/. = 10.17 10,500,000

. \::~,.::~~.\.:;' :.: :~ ~i.~,:<~~', ::r'1~'~~;I::::;l:'~~;"; ~,:: ::>:.;/i) ~: ":07 .~: '/':'~~~' ;'~';L;E~~2~~.Ii1

C7.8

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION

Fes = .57 x 67,000

From Fig. C7.9, FcS/F C Y = .57, whence = 38,200

.8 x 11,100 :: 8900 psi. Thus unless tests substantiate higher values, t~e crippling stress should be taken as 8900 pst ,

Problem 6. Find the crippling stress tor the square tuba as shown in F1g. d. Material 1s 2024-T3 aluminum alloy. Fcy = 40,000, Ec = 10,700,000. Solution by Method 1. The square tube 1s considered as made up of 4 equal angles with no edge free.
b ' /t = (a + bl/2t

Problem 8. Same as Problem 6, but change material to stainless steel l7-7PH(THl050), Fey = 162,000, Ec = 29,000,000 Method 1.
F es

.039Z,)162,OOOxZ9,000,Ooo = 85,000 psi

Method 2.
.373
Fig. d

162.000

1/~

_
-

(12 x .05') (29,000,000)

.9~

= 1.95/0.1 = 19.5

Fro.m Fig. C7.7, Fcs/FCY:: .595. 162,000 x .595 = 96,200 pSi. Problem 9.

Fes

From Fig. C7.3, uSing upper curve, we obtain

Fee/v'FeyE = .0:392. wllenee,

Pes .. Fcc = .0392 xv'40,OOO x 10,700,000 =

25,600 psi Solution by Method 2. Area A = .373. g = number of cuts plus tlanges or 4 + 8 = 12. For rectangular tubes we use equation C7.4 or Fig. C7.7.
(Fey),/a. (.373 a 40.000 gta E 12x .05 ) (10,700,000) A
1. 5

CD2
.375

1 -'-~I.

=t==.o=:::::j=1
Fig. e

-r: -

0.758

From Fig. C7.7, Fcs/F cy

= .70

Determine the crippling stress for the tormed section shown. in Fig. e it material is aluminum alloy 2024-T3. F cy = 40,000, Ec = 10,700,000. Solution by Method 1 (Needham) The section is diVided into 6 angle units by the dashed lines in Fig. e. They are numbered (1) to (3) since we have symmetry. The procedure will be to find the tailing load for each angle and add up the total for the 6 angle units. The crippling stress will then equal this total load divided by the section area.
Angle un! t (l I (One edge
~ee)

Therefore Fcs = 40,000 x .70 = 28,000 psi Problem 7. Same as Problem 6, but change material to magnesium HK3lA-O Sheet, subjected to a temperature ot 300F tor 1/2 hour. Solution by Method 1. Fey

From Table Bl.l ot Chapter Bl, 11,100, Be = 6,160,000.

Fcs = .039ZVl1,100 x 6,160,000 = 10,250 pSi Solution by Method 2.

.528

b'/t = (a+bl/2t = [(.375- .02) + (.5- .02)]/.08

= 10.44 )From Ftg. C7.3, Fec/vFcYE = .06

whence, Fcc

Fram Fig. C7.7, Fcs/F cy = .96

Fcs

= .96

x 11,100

= 10,600

pSi

.06 x y'40,ooO X10,700,000 :: 39,300 psi

From Table C7.1, the cut-ott or maximum crippling for rectangular tubes is .8 Fey or

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.9

Area of angle (1)

P cc

= .0309 =A
=
.0459 (no edge free)

as Section NO.2. 0.113 sq. In.

The area tram that table is

.0309 x 39,300 :. 1215 lb.

Solution by Method 2 (Gerard)

10,700,000.

Angle unit (2) b'/t

Area

= (a + b)/2t = (.48+ .73)/.08 = 15.1

Fee/vFc;:E = .0475

For this material Fcy = 53,000; Ec

From Fig. C7.3

Fcc 'Y40,OOO x 10,700,000 x .0475 = 31,100 psi Pee = 31,100 x .0459 Angle unit (3) b'/t

= 1428

The first question that arises is the bulb size sufficient to give an end stittness to the (a) leg so that the bulb may be equivalent to the normal corner.
In Fig. t, bt

lb.

Area = .0509 (no edge free)

=0.78,

hence br/t = 15

= (.73
=

+ .605)/.08 = 16.7

F'fom Fig. 07.3, Fee/vFc;:E = .046

Fcc .046 x /40,000 x 10,700,000 ;; 30,200 psi

Pcc = .0509 x 30,200

= 1540

lb.

Fes ;; ZPcc/area

= (2x1215+2x1428+2x1540)/0.255 =
32,800 psi Solution by Method 2 (Gerard)

Referring to Fig. C7.lZ, we observe that tor a br/t value or 15 we need a D/t ratio or at least 3.8. The D/t'value for ~ur bulb angle is (7/32)/.05 4.4, thus bulb has sufticient stiffness to develop a corner. The next question that arises should the bUlb angle still be classed as an angle section ror which equation C7.4 applies or be classed as a channel or 2 corner section with the bulb acting as a short thick leg ot the channel. For this case, equation C7.6 would apply. The crippling stress will be calculated by both equations.
By equation C7.4 or Fig. C7.7:-

gtr

(!z.) ,/. Eo

cuts = 12

g = number ot flanges plus number of

+ 5

It bulb is considered as a tull corner then g = 4 flanges plus 1 cut 3 5.

gt8

= 17.

(Fe y),/. = (
Ec

5 X .OS'*

.113

)(

53,000 ,/"_ 10,700,000) -.636

SUbstituting in the above term,

From Fig. C7.7,

) ( 40,000 ),/._ ( 17 .255 x .041 10,700,000 - .573
FeB

Fcs/Fcy

:=

.82, hence Fee

.82 x 53000 43500

= 40,000 x .895 ;; 35,600 psi.

From Fig. C7.7, Fcs/FCY;; .895, whence

By equation C7.6 or Fig. C7.9,

From Table C7.1, it Is recommended for eutm-oomer sections that Fes max1mum be limited to .8 Fey unless tests can prove higher values.
FcSmax

From Fig. C7.9, Fcs/FCY = .7, hence Fes =

.7 x 53,000

=37,100.

= .8 x 40,000 = 32,000 pst. Since

this Is less than the above calculated values, it should be used.

Problem 10. Find the

PosSibly the best estimate at the crippling stress Would be the average at the two above results or 40,300 psi. In Table C7.l, the so-called cut-off stress for angles 1s .7 Fcy and channels .9 FCY' If we use the average value or .8 FC Y' it gives Fes 8 x 53,000 = 42,400 as maximum permissible because at 11mited test reSUlts on bulb angles. For the case where the bulb or lip does "\ not develop the stiffness necessary to assume a rull corner, then the bulb is only

crippling stress tor the extruded bulb angle shown in Fig. f it material is 2014-T6 extruded aluminum alloy. This particular bulb section is taken trom Table .\3.16 of Chapter A3

Fig. f

C7.10

CRIPPLlliG STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION

Fig. b Sheet stress distribution before buckling

considered as an additional flange and the g count would be tour instead of 5, thus reducing the crippling stress. EFFECTIVE SHEET
C7. 11 Introduction.
~IDTHS

((,I(ll(llOIUIIJ]
Sheet stress distribution after buckling

The previous discussion in this chapter has dealt with the crippling stress of formed or extruded sections when acting alone, that is not fastened to any other structure along its length. However, the major structure at aerospace vehicles, such as the wing or body, involves a sheet cover1ng which 1s strengthened by attached or integral fabricated stiffeners such as angles, zees, tees, etc. Since the sheet and stiffeners must deform together, the sheet will therefore carry compressive load and to neglect this load carrying capacity of the sheet would be too conservative in aerospace structural deSign where weight saving is very Important.
C7. 12 Sheet Effective Widths.

Fig. d

I i I . I I i

! '

Fig. C7.l3a illustrates a continuous flat thin plate fastened to stiffener and the entire unit is SUbjected to a uniform compressive load. Up to the buckling strength of the sheet the compressive stress distribution is uniform over both stitteners and sheet as in (Fig. b) assuming same material for sheet and stiffeners. As the load is increased the sheet buckles between the stiffeners and does not carry a greater stress than the buckling etrese , However as the stiffeners are approached, the skin being stabilized by the stiffeners to which it is attached can take a higher stress and 1mnediately over the stiffeners the sheet can take the same stress as the ultimate strength of the stiffener, assuming that the sheet has a continuous connection to the stiffener. Fig. c shows the general stress distribution atter the sheet has buckled. This distribution dependS on the degree or restraint provided by the stiffeners and the panel dimension. Various theoretical stUdies (Ref. 6) have been made to determine this stress distribution atter buckling. In general they lead to long and complicated equations. To proVide a Simple basis for deSign purposes, an attempt has been made to tind an effective width of sheet w which would be considered as taking a unifor.m stress (Fig. d) which would give the same total sheet strength as the sheet under the true non-unl~orm stress distribution of Fig. c. The question of sheet effective sheet has been consIdered by many individuals. The names of VonKarman, Sechler, Timoshenko, Newell, g'ranzj.anc , Margurre, Fischel, Gerard, and many more are closely associated with the present knowledge on effective sheet width.

Fig.

a
I I I

n
....
'I(

lil/lf{ltfl

Sheet- stiffener panel

Fig. C7.13

From Chapter CS, the buckling compress ive stress at a sheet panel is,
(C7.13)

.If we assume that the stiffener to which the sheet is attached provides a boundary restraint equal to a sunple support, then kc = 4.0, and if Poisson's ratio V e 1s taken as 0.3, then equation C7.l3 reduces to,
Fcr

= 3.60E(t/O)'

(C7.14)

The Von-Karman-8echler method as first proposed consisted of solving equation C7.l4 for a width (w) in place of (b), when Fer was equal to the yield stress of the material since experiments had shown that the ulttmate strength of a sheet simply supported at the edges was independent of the width of the sheet. Thus equation C7.l4 changes to, Fcy =
w
3.60E(t/w)~

whence,
- - - (C7.15)

= 1.90t .;E/Fcy - - - - -

Since the crippling or local failing stress of a stiffener can exceed the Yield . strength of the material, equation C7.15 was later changed by replaCing Fcy by the stress in the stringer FST' thus giVing,

i .

~.

--------_...
w = 1.90t VE/FST

-'

{ '; '.....

- '-. 1'1'

- '.<

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.11

- (C7.16)

Same early experiments by Newell indicated the constant 1.90 was too high and for light stringers a value of 1.7 was more realistic, thus 1.7 has been widely used in industry. If we assume the stiffness of the stiffener and its attachment to the sheet as developing a fixed or clamped edge condition tor the sheet, then Fer = 6.35E(t/b)" or w = 2.52t vE/FST For general deSign purposes, it Is felt that 1.9 or equation C7.15 Is appropriate tor determining the effective width w. It stiffener Is relatively light, use 1.7. Fig. C7.14 illustrates the effective width tor sheet-stiffener units which are fastened together by a single attachment line tor each flange of the stiffener.
2 '"' Rivet Lines
One

The effective sheet width as calculated by equation C7.l6 assumes that no inter-rivet buckling of the sheet occurs or J in other words J the rivet or spot welds are close enough together to prevent local buckling of the sheet between rivets when the sheet is carrying the stiffener crippling stress. The subject of inter-rivet sheet buckling is discussed later in this chapter. Fig. C7.l6 1llustrates a procedure to follow tor determining the effective width w when sheet and stiffener are integral in manufacture.
Case 2
tf ~ 2t s

Tee Section

'.
I-VI

'-I

no!

Fig. C7.14

Rivet Line

+4

Fig. C7.18

The crippling stress is determined for the stiffener alone. This stress is then used in equation C7.15 to determine the effective widthS w. The total area then equals the stiffener area plus the area of the effective sheet width w. The radius of gyration should include the effect of the effective skin area. Fig. C7.l5 illustrates the case where stiffeners are fastened to sheet by two rows of rivets on each stiffener flange. In this caseJ the rivet lines are so close together that the effective width w for each rivet line would overlap considerably. A cammon practice in industry for such cases is to use the effective width for one rivet line attachment as per equation C7.l6 to represent sheet width to go with each stiffener flange. However J in calculating the crippling stress of the stiffener alone J the stiffener flange which ts attached to sheet is considered as haVing a thickness equal to 3/4 the sum of the flange thickness plus the sheet thickness.

For Case 1, find the crippling stress for the tee section alone J assuming the vertical stam ot the tee has both ends Simply supported. For value at t in equation C7.15, use (t s ... tr}/2. The effective stiffener area equalS the area of the tee plus the area ot the sheet at width w. For Case 2, determine the crippling stress for the I section acting alone. calculate w/2 tram equation C7.lS to include as effective sheet area. The column properties should include I section plus effective sheet.
C7.12 Effective Width Wl. for Sheet with One Edge Free.

In normal sheet-stiffener construction, the sheet usually ends on a stiffener and thus we have a free edge condition tor the sheet as 1llustrated in Fig. C7.l6a. The sheet ends at

Fig. C7.16a

L===;
I-~-I

Staggered Rivet Rows

Fig. C7.15

a distance b' from the rivet line. For a sheet free on one edge, the buckling coefficient in equation C7.l3 is 0.43, thus equation C7.l3 reduces tO J Fcr = .387E(t/b,)a J and replacing b' by w J we obtain,
wl. = .62t JE/FST

I
I

I
I w--1

- - - - - - - - - - (C7.17)

C1.12

CRIPPLlliG STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION

Then the total effective sheet width for this end stif:ener would thus equal wl + w/2.
C7.13 Effective Width When Sheet and Stiffener Have Different Material Properties.

In practical sheet-stiffener construction it is co~on to use extruded stiffeners which have different material strength propertles in the inelastic stress range as compared to the sheet to which the stiffener is attached. For example, in Flg. C7.l7 the stiffener material could have the stress-strain curve represented by curve (1) and the sheet to which it is attached by the curve (2). Now when the stiffener is stressed to point (B), the sheet directly adjacent to the stlffener attachment line must undergo the same strain as the stiffener and thus the stress in the sheet will De that given at pOint (AI in Fig. C7.l7. This difference in stress will influence the effective width w. Correction for this condition can be made in equation C7.l7 by multiplying it DY FSHIFsT, which gives w = 1.9ot(FSHIFsT)

before the crippling stress of the stiffener 1s reached, which means the sheet is Isss effective in hel~ing the stiffener c~rry a suo jec t ed compress 1ve load. Thus, in genera L, to save structural Nelght, str~~tur~: ~2~:~~~r~ select r-ivet scac irxs that 'N111 or-event Int.er>rivet buckj t ng of the sheet. I:1 general, the rivet spacing a~ong the stif~eners in th~ u?per surface ot the wlng will ~e :loser together than on the botto~ surface of ~he wing since the design com~ressive loads on the top surface are considerably larger than those on the bottom surface. The following method is widely ~sed by engineers concerned wit~ aerospace structures relative to calculating inter-rivet buckling stresses. It 1s aSS1~ed that the sheet between adjacent rlvets acts as a col~~ With fixed ends. The general coIumn equation from crapter C2 for stable cross-sections 15,

(~F~T

Fc = Cn'Et/(L/p)'
) - - (C7.l8)

- (C7.19)

Where tST is the stiffener stress and fSH is the sheet stresS existing at the same strain as eXisting for the stiffener and obtained from a stress-strain curve at the sheet material.
70
~

Where C Is the end fixity coefficlent and varies from a val~e of 1 for a pin end support to 4 for a fixed end support. The effective column length L' thus equatlon C7.19 can be written
~

Live,

'"
50

,,

Curve 2 Curve 1

-I I

- - (C7.20) Let p the rivet spaclng be considered the column length L. Assume a unlt of sheet linch wide and t its thicY~ess. Then moment of inertia of cross section = 1 x t~/12, ~~d area A = 1 x t ~ t. Then radius of gyration p ~ O.29t. Then substituting in equation C7.20 to obtain the Inter-rivet buckling stress Fir, Fir = n:llEt

:I:

"l

-c

'"'
30
20

;;
~

~ ~

I!! 10
~

(kiO .29t)'

- - - - - - - - - - (C7.21)

.002.004 .006 .008 .010 .012

STRAIN - IN./IN. Fig.

For clamped ends C :: 4, thus Fir' (p/G.58t)'

- - (C7.22)

C7.i7

For a rather complete study and comparison or the various effective widths theories as compared, see article by Gerard (Ref. 7). Equation C7.l5 Is In general :onservative for higher bit ratios.
C'7.14 Inter-Rivet Buckling Stress.

To olot this equation, the tang8nt modulus Et tor the material must be known. However, we can use the various column curves f~ Chaoter C2 which show a plot of Fe versus L' /0 and in equation C7.22 the ter.n P/O.58t correspondS to L'/p. The fiXity coefficient C ~ 4 can be used for flat head rivets. For spot welds It should be decreased to 3.5. For the BraZier rivet type use C :: 3 and for counter-sunk or dimpled ~ivets use C ~ 1.

The effectlve sheet area Is considered to act monolithically with the stiffener. However, it the rivets or spot welds trat tasten the sheat to the stiffener are spaced too far apart, the sheet will buckle between the rlvets

[--

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C7.13

Figs. C7.18 and C7.19 shaw a plot ot equation C7.22 for aluminum alloy materials.
It the inter-rivet buckling stress calculates to be more than the crippling stress of the stiffener, then the effective sheet area can be added to the stiffener area to obtain the total effective area. This total effective area times the stiffener crippling stress will give the crippling load for the total sheetstiffener unit.
l~en the sheet between rivets buckles before the crippling stress of the stiffener is reached, the sheet in the buckled state has the ability to approximately hold this stress as the stiffener continues to take load until it reaches the stiffener crippling stress. This buckling sheet strength can be taken advantage ot by reducing the effective sheet area. Thus effective sheet width equals,

-r 5/8
Fig. C7. 21 Detail of Point Ul .064

J_
.I.
1:.072
web

Fig. C7.20

corrected = W(F i r IFcs )

- - - - - (C7.23)

The area is then added The crippling stress of the

of the corrected effective sheet to the area of the stiffener. load then equals the crippling stiffener ttmes the total area.

problem 15 to determine the crippling stress of the stiffener, the effective skin area and the total compressive load that the unit can carry at the failure point. Since the stiffener is braced laterally by the web and the skin, column bending action is prevented and thus the criPpling strength is the true resulting strength of this corner member under longitudinal compression. (Additional stresses are produced on these corner members i t web buckles under shear stresses and web diagonal tension forces are acting. This SUbject is treated in the chapter on semi-tension field beams.) Solution:

The use of sheet effective widths in fi~ding the moment of inertia of a wing or fuselage croSs-section is a widely used procedure in the analysis for bending stresses in c~nventional wing and fuselage construction. Reference should be made to article Al9.l3 of Crapter Al9 and article A20.3 of Chapter AZO for practical illustrations in the use of effective Widths. C7.i5 illustrative Problem Involving
Effective Sheet.

Area of stiffener"" 0.24 sq. in. For 7075-T6 extruSion, Fcy "" 70,000, Eo ""
10,500,000.

The Gerard method will be used in calculating the crippling stress. Equation C7.4, or the design curve of Fig. C7.7, applies to this multlcorner shape. The lower scale parameter for Fig. C7.7 is,
gt

Conv8ntional airplane wing construction is illustrated in Fig. C7.20. The wing is covered with sheet, generally referred to as skin, and this skin is stitfened by attaching tormed or extruded Shapes referred to as skin stiffeners or skin stringers. A typical wing section involves one or several interior straight webs and to tie these webs to the Skin, a stringer often referred to as the web flar~e member, is required to facilitate this connection. Fig. C7.21 is a detail of the flange member and the connection at point (1)
in Fig. C7.20.
~he stiffener or flange member is an extrusion ~f 7075-T6 aluminum alloy. The skin and 'Neb sheets are 7075-T6 aluminum alloy. The skin is f~stened to stiffener by two rows of 1/8 inch diameter rivets of the Brazier head type, spaced 7/8 inch apart. The web is attached to stiffener by one row of 3/16 diameter rivets spaced 1 inch apart. The

~ (F ey),;_ = 0.24 ( 70,000 )'1- = 0.63 Ec 6 x 0.72 a 10,500,000

using this value, we read fram Fig. C7.7 that Fcs/F ey 0.82. Thus FeB = .82 x 70,000 = 57,400 psi. The g value ot 6 was determined as shown in Fig. (a). Effective Sheet Widths: Equation C7.16 will be used to determine the sheet effective widths. For the skin t = .05. Material is 7075-T6 aluminum alloy. Ec = 10,500,000.

-t
f f
f

cu t

Flanges (f) : 5 No. of cuts: 1

g

:'6

Fig. a.

W= 1.9t vE/Fcs "" 1.9 x 0.5 vlO,500,OOO/57,400

= 1.28

in.

Thus a piece of sheet 1.28/2 = .64 wide acts to each side of the rivet centerline. Observation
-z, -; , ...-/._---()

ao

70

.1 1 ;

.. ..
.II
~

I ,
,

~ .. .~
~
~

eo

Fig. C7.18 Materlal Alum. Alloy Bare Curve Designation Thlckneea Basis -(BARE) -2024-T4 <.260 Z024-T3 2024-T36 2024-T6 2024-T81 2024-T86 7075-T6 7075-T6

70

Fig. C7.19 Material Alum. Alloy Clad Curve Designation 'rtuckneea Bas Is

.II

9 .
~

..
eo
'0

--

(CLAD)

--

2024-T4 2024-T3 2024-T36 2024-T6

2024~T81

<.

064

I
I

M
~

'0

~ ..

2024-T86 7075-T6 7075-T6

..
20 10

a 1'1
~

ri

ao

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.15

of the di~ensions in Fig. C7.2l shows the effective Nidth with each skin rivet line would overlap slightly thus we use only the width between rivet li~e (see Fig. b). For the web t = .064. Sinca the web has a free edge, the effective width calculation will be in two steps.

~ 1 29 x .064 Jlo,.500,000/57,400 : 0.82

from equation C7.1S 62t E/F es = .62 x 064 VIO,500,OOO/57,400 ~ 0.55 inch. Fig. (b) shows the effective sheet width as calculated. Total effective sheet area = 1.28 x .05 -+- .875 x .05 + (0.55 + 0.S2) .064 = .195. Total area = .195 + 0.24 ...435.

usually of the flUSh surface type, either countersunk or dimpled. If we mak~ the skin rivets of the countersunk type, the end fixity coefficient ~ust be reduced to 1 to be safe. Then the corrected pit ratio to use with Fig. C7.1S would be (V4 I vl)p/t. 2 x .S75/.05 = 35. From Fig. C7.l8 and curve 8, Fir = 29,000 whi:h is far below the cal:ulated crippling stress, thus the rivet spacing would have to be reduced. Use 9/16 inch spac Ing corrected pit = 2 x .5625/.05 = 22. From Fig. C7.1S, Fir 57,400 pSi, which happens to be the crippling stress and therefore satiSfactory
C7.16 Falling strength of Short Sheetstiffener Panels in Compression.

~!~+."5

+ .. ~
\
Fig. b

wi'

-r ... =.55
1

w 2'",.82

+-

The total compressive load that entire unit cc~ carry before f~11ure is then equal to AF c ; = .435 x 57,400 ~ 25,000 Ibs. This result assumes that no inter-rivet buckling occurs under the stress 0f 57,400 psi in the sheet between rivets. The skin rivets are Brazier head type spaced 7/8 inch apart or p ~ 7/8. As discussed under inter-rivet buckling, the end coefficient c for th13 type of rivet should be less than 4 or assumed as 3. Fig. C7.18 gives the inter-rivet buckling stress versus the pit ratio. This chart is based on a clamped end condition or C = 4. Since C = 3 will be used for the Brazier type rivet, we correct the pit ratio by the ratio

Gerard (Refs. 2, 3) from a :omprehensive study of test results on short sheet-stiffener panels i~ compression, has shown that his equation C7.4, or Fig. C7.7, can be used to give the local monolithic crippling stress for sheet panels stiffened by Z, Y <L~d H at shaped stiffeners. The method of cal~ul~ting the value of the g f~ctor is illustrated in Fig. C7.22. Fig. C7.23 Is a photograph showing the crippling type of failure for a short p~~el invol7ing the Z shape stiftener.
C7.17 Failure by InterRivet Buckling.

HowlRnd (~ef. 8) assumed that the sheet acts as a wide column which is clamped at its ends and whose length is equal to the rivet spacing. The int~r-rivet buckling stress equation is then,
cnll'n t s II' Fir = 12 (1 _ Ve')<-p-l - - - - - - - - (C7.24)

4 for

The end fixity coefficient C is taken as fl~t head rivets and reduced for other types as previously explai~ed for equation

C'7.Z1.

14 1/.31.16.

17 Is the plasticity correction factor 17 is the clad correction factor

Ve is Poisson's ratio (use 0.'30)

The corrected pit .. 1.16 x .875/.05 .. ZO.3 From Fig. C7.18 USing curve (8) which is our materl~l, we read Fir = 50,000 pSi, thus skin will not buckle between rivets as crippling stress is 57 J 4 0 0 pSi. The web rivets are of the flat head type and C = 4 can be used. Spacing is linch. Hence pit = 1/.064 = 15.6 and from Fig. C7.1B, curve 8, we read Fir - 64,600, which is considerably more than the unit crippling stress. In wing construction, the skin rivets are

t s = sheet thickness, inches. p = rivet spacing or pitch in inches. For non-clad materials the curves of Fig. C7.24 can be used. ThiS figure is the same as Fig. C5.8 of Chapter C5. For the clad correction see Table C5.l of Chapter C5.
C7. 18 Failure of Short Panels by Sheet Wrinkling.

In a riveted sneee-eet rrnner panel, 1! the rivet spacing 1s relatively large, the sheet will buckle between rivets, such as illustrated in the photograph of Fig. C7.25. This inter-rivet buckling stress was discussed in the previous

C7.16

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION

Average g = 16. 83
(c)

Hat-stiffened panel.

g=8

g=8

g=8

g=8

2 Cuts 8 Flanges Ssg FF

Average g = 7. 83
(b) Zstiffened panel.

A ~ i~1 ,~! ,I'I 1""9+

c.
A;

FF;FF c
19 = g

I C

Ig;rl

19:19

,1
';18

5 Cuts

Average g "" 18.83

(.) Y-stiIfened panel.

14 Flanges

Flg. C7.22 Method of cutting stiffened panels to determine g.

Fig. C7.23 A 24S-T aluminum-alloy Y-stiliened panel (on the left) and its 75S-T counterpart after failure.

ANALYSIS AND DESIGN OF

a~ti~le.

FLIGHT VEHICLE STRUCTURES

C7.17

This sheet bUCKll~g does not deform the flange of the stiffener to Nh~l-;h the shee t is attached. However, if the rivet or spot weld spacing is such as to prevent inter-rivet buckling of the sheet, then failure often occurs by a larger wrinkling of the sheet as illustrated in F16. C7.26. The larger wrinkle snape SUbjects to flange of the stiffener to which the sheet is attached to lateral forces and thus the stiffener fl~nge often deforms with the sheet wrinkle shape. This deforming of the stiffener flange produc~s stresses on the stiffener web, thus Nrinkling failure is a ~Qmbination of sheet and stiffener failure. The action at the wrinkling sheet to defor.n the stiffener flange places tension loads on the rivets, thu~ rivet design enters into the failir~ strength of sheet-stiffener panels under compression.

aJproxirnate criterion for rivet strength from Ref. 2 i5,

- - - - - - ("C7 .'2.7 j

The tensile strength of the rivet Sr is defined in terms of the shank area and it may be associated with either shank faIlure or pulli~ of the countersunk he~d of the rivet through the sheet. For aluminum alloy 2117-T4 rivets whose ens i Le s'tz-eng'th is s ... 57 kst , the cr-t ter-i.a are:t

= 57 kSi., de/t:!.,! -;

1. 67
(C7.28)

where t av . is the average ot sheet and stiffener thickness in inches. The effective diameter de is the diameter for a rivet made from 2l17-T4 material. The effective diameter of a rivet of another material is,
d e /d = (8./8)

Fig. C7.25 Inter- Rivet Buckling

Fig. C7.26 Wrinkling Failure

'I,

- - (C7.29)

Several persons have stUdied this wrinkling 'Nhere Sr is the tensile strength of a rivet or forced crippling of riveted panels (RefS. 9 defined as maximum tensile load divided by shank area in ksi units. and 10). A rather recent study was carried out by Semonian and Peterson (Ref. 11), which is reviewed and sl~pllfi9d somewhat by Gerard in C7.21 Problem 1. lllustrattng Calculation of (Ret. 2). The results as given in (Refs. 11 Short Panel Failing strength. ~~d 2) used to calculate the wrinkling stress. Fig. C7.29 shows a sheet-stiffener panel composed of formed Z stIffeners. The material C7.19 Equation for Wrinkling Failing Stress F w is aluminum alloy 2024-T3. Fcy = 40,000. From Ref. 2 we obtain, Fo.~ a 39,000, n = 14.5, Eo = 10,700,000. The problem is to determine the compressive tailIng. k" ,,' 1) ry E ~, strength of a short length of this panel unit. - (C7.25) '" = 12 (1-/),') (b s )

kw, the wrinkling coefficient is obtained from Fig. C7.2ry. This coefficient 1s a function of the effective rivet offset f Which is obtai~ed from Fig. C7.28. Having determined ~ from Fig. C7.27, equation C7.25 can be solved by use at Fig. C7.24.
C7.20 Rivet Criterion for Wrinkling Failure.

/
.:

A crlteriJn for the rivet pitch round from test da ta vn i cn results in a wr~nkl1n'?; mode rat lure 15, Ganeral
(C7.26)
~.nel Dat~:

Fig. C7.29

The 19.teral force required to make the stringer at.tachaent flange conform to the wr tnk Ied sheet, loads the r1 vet in tension.

tw ts
An
bs

= .064
.064

bw
bf

= :2.4.'37
= 0.905

bA '" 0.593

bo

= 0.343

2.00
~" ,,,-,,I
,

: >~ ,'c' -

::i(: _.:-~;~~L-:,:/ ,::::;': :'; t

,.','..:.: j,:;. ,,~.:'-~; :, : ;~. ,::::. ::::~'.: ,:~~.

-~--

-------

..>'

-~:~':.~< :i.;1't ~?i~

-: ;

C7 18

, PANELS rn COMPRESSION CRIPPLmG STRENGTH OF COMPOSITE SHAPES AND SHEET STIFFENER

r-'

1.2

I
,

rJC

IF w .8 Fo. "

I-

s
Fir

Fo . "

..

:V
a

l-

~
,

~ I~ V
,
s
c n2
E

~~

I~~ ~2-

.:

~VJ;~
~
Q

~:

I I I
Fig. C7.24 Curves for finding Fir as per eq, C7. 24 and F w as per eq, C7.25.

1.2

.
~naE

.
."

20

12 (1 - Ve:il)F"."
0

12 {l - 1I e)F 0

s (-) bs

II

I ,\ '. $I,.-f: G

,,\' \ l\' '., "\ 1\\1\ I\..

i:\:
<,

\ I\ I \1 \' \ \ \

i ,! 1
I

I
,

i , I ,
I

!
,
1

i
\
.2
,

i,

..i ,
,

i
,

,
I
,

I I I

1 I
I
! !

'I

"

I
I
I

, ,
I ,

.2

~ i ,., " " ,.. "

,
,

~ f-..

I'...: r-, r-,

i'i--I :::::t-+-l--I

Fig. C7.27 Experimentally determined coefficients for failure in wrinkling mode. (Ref. 2)

"

,
i
I
"

I
,

, I
1

!
,

i
,

~..ilf
I

'-/1'

,
\

I,.

+.

~ .i-r ' 7'_, ,--r

-;V'/.vi' ,
._ s

:.-+-

I
,

~ ...................
I
1
,

...... /
1

;
I

i
j

,
1

I
1

zz

- iI
I

,
i

.
,-

i
i

I I

"

I
!

,,","d'

'--

I ,

I
r

. ..." I--

!i
0

, ,

Fig. C7. 28 (Ref. 2) Experimentall y determined values of effective

rivet offset.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.19

whence,

bw/tw = 38,

0.372 5.36

Fir = 39,000 x 1.1 = 43,000. This value is far about the stiffener or panel crippling stress as previously calculated so inter-rivet buckling 1s not at all critical. Failure by Sheet or Face Wrinkling. (Fw)

The rivets are 3/32 diameter Brazier head type AN456, 2117-T3 material spaced at 0.75 inches. Area of Z stiffener = 0.852 in. Area of skin tor bs = 2 inches = 2 x .064
a

0.128 Total area of sheet and stiffener = 0.380 in. a

The wrinkling failure stress by equation C7.25 is,

Fw

Crippling Stress of Stiffener Acting Along.

lo" n'n E 12(1-Ve') To determine value of a~d C7.27.

p/d

Fcs(ST)'
Since we have a Z type of stiffener, equation C7.6 and Fig. C7.9 applies. The lower scale parameter in Fig. C7.9 is,
\

Figs. C7.28

kw,

we use curves in

= .75/.0937 = 8,

bo/tw

= 5.35

(0.~o2)( i

.064

40,000 )'/' 10,700,000

= 9.56

From Fig. C7.9, Fcs/Fc~ = 0.6. Hence, Fcs = .6 x 40,000 = 24,000 P i = Fcs(ST)'
Crippling Stress of Panel Considered as a Monolithic Limit. FCS(M)" Equation C7.4 or Fig. C7.7 applies for monoli thic failure of sheet-stiffener panels. The lower scale parameter on Fig. C7.7 is,
( 0.380 )( 40,000 )'/' 7.83x .064 3 10,700,000

From Fig. C7.28, we read f/tw = 6.5, whence r = .054 x 5.5 = 0.415 r/bw = .415/2.437 = .17 (bw/tw)/os/t s 38/(2.0/.064) 1.21

From Fig. C7.27, we read kw = 4.4. To solve equation for Fw we use Fig. C7.24. The lower scale parameter is,

kw n3 E ts .3 12 (1 - Ve ')(F,.,) (b s) =
4.4 n x 10,700.000 (.064)a = 1.15 12 (1 - .3')(39,000) 2.0 For n = 11.5, we read from Fig. C7.24 that Fw/Fo :-;=- .9"; whence Fw .9 x 39,000 = 35,100 psi. Thus wrinkling failure 15 not critical as Fw is larger than Fcs(M} and Fcs(ST}o
3

= 0.73
g = 7.83 (see Fig. C7.22b)

From Fig. C7.7, we read Fes/F ey = .725, hence Fes = 40,000 x .725 = 29,000 psi = FcS(M)' Inter-Rivet Buckling Stress (Fir) The rivet type is Brazier head and the spacing p is 3/4 inch. Equation C7.24 ~ppli~S and Fig. C7.24 is used to solve the equation. The lower scale parameter in Fig. C7.24 is,
C n.3 E 12 (1 -l)e 3 )
suos t t tut tngi-,

The results show that the crippling stress for the stiffener alone of 24,000 pSi is the smallest value, or the stiffener is unstable as it fails first. The entire panel unit will not reach its failing strength when stiffener stress is 24,000 because the skin wrinkling stress f w is higher. An approximation suggested in (Ret. 2) is to assume stiffeners carry the same stress as the skin up to Fcs(ST) and beyond this the stiffener carries no additional load. Thus the panel failing stress F(F) can be calculated trom the follOWing equation.

F'O.7

For Brazier head rivet C

= 3. u =

.3.

The Shape parameter n for our material is 11.5. Reference to Fig. C7.24 for a value of 5.45 on bottom scale which is off the scale, we estimate the Fir/To"~ as above 1.1. Thus

The total load carried by each stiffener plus its sheet is 27,800 x .380 10.600 lbs.

The falling strength of the riveted panel cannot exceed the monolithic panel failing stress FcS(M), which was 29,000 psi :Oar our panel or

C7.20

CRIPPLING StRENGTH OF COMPOSITE SHAPES AND SHEETSTIFFENER PANELS IN COMPRESSION

f~:lln~

greater than the calculated 2",aOO pSi.

stress of

Check of Rivet

Str~n~th.

From

ex~ression

C7.26

(5) p/D ~ 8 - - - - - - - - - - Face wrinkling (5) Tensile strength of rivet or spot weld attachmen~ ~er inch should be ~ 0.05 ~~yts in order to prevent failure in wrinkling. (7) As a rou,h guida do not use bent up stringers it bs/t s < 30 in or~er to prevent face wTin~ling wea~ess.
C7. 23 Y Stiffened Sheet Panels.

0z':g:=

.:375

-<

1.0:27/(4.4)

:../~

::. .603(sa-:isfactory)

The criterion for required rivet strength to make the stiffener flange tollow the wrinkled sheet is from 07.27

s
S

r>EsTdd
(

0.7

be

(F ) ,

r->

0.7 '( 2 )(0.75)(35 100)' 10,700,000' 3/32 3/32 '

A Y shape cannot be fonned ~rom sheet, thus it must be extruded. To make the Y Shape efficient, the various parts usually have a different thickness. Furthermore, the extruded material has different mechanical properties in the inelastiC stress range as compared to rolled sheet that is used for the panel, thus these effects must be considered in calculating the crippling stress of the stiffener and the complete panel unit. The effective thickness ~N of the stiffener 1s determined by the following equation (Ref. 3):

Sr::OO 13.7 pe t . From expression C7.28 s = 57 xs i for de/t av.

~

- - - - - - (C7.29) 1.57 where bi and ti refer to the lengtn and thickness, respectively, of the cross-section elements. Wnen the yield stress F cy of stiffener and sheet are different and effective Fcy can be estimated as follows (Ref. 3):

For given panel de/ta v := 3/32/.064 = 1.46, thus rivets have plenty of tensile strength to produce a wr-tnkkfng failure.
C7. 22 General Destgn Limitatio113 to Prevent Secondary Failure in Sheet-Stiffener Panels.

Sheet stiffener units can be designed as if the secondary forms of failure such as inter-rivet buckling and face wrinkling are avoided. The following design rules referring to Fig. (A) will usually avoi1 these secondary weaknesses.
col~

Fey = {FCY(SH) Fcy(ST) [(twits) -l]}tw/ts

--------(C7 .30)

The monolithic crippling stress for the sheet stiffener panel can be Galculated fram equation C7.4 or by the curve in Fig. C7.7. Equation C7.4 is

- 0.56 Fes(M) I Fey -

[(gt '/)( A EIF C Y ).1']""

The constant 0.56 in this equation applies for Y and hat-stiffened panels when the ratio twits = 1.0. For other ratios the correction of this constant which is referred to as J?g in Gerard's basic equation can be obtained fram curve in Fig. C7.30.

bs

"'r--r--,--,---------,
,

Fig. A

..... , s",,.rNCD , S'''''F1WE#

II

_,I ,
~ANELS

(1)

tf~/ts ~

0.5 - - Promotes overall crippling

<

(2) 0.4 - bra/bN

0.5 -- Rolling versus local buckling

04

0,6

i-rl

(3) Make bo as small as ?OSSiblp- -- Face wrinkling (4) p/b s -< 0.5 - - - - - - Inter-rivet buckling

Fig. C7.30 fig correction for 'twIts effect on Y- or hat. stiffened panels.

,.;:.;;.._-.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C7.21

C7. 24 Example Problem Y Stiffened Panel.

The compressive monolithic failing stress of a Y stiffened panel, as illustrated in Fig. C7.22a, will be calculated by the Gerard method. Fig. C7.3l shows details of the panel unit. The stiffener is extruded from 20l4-T6 aluminum alloy for which Fey = 53000 and Ec ~ 10,700,000. The skin or panel sheet is aluminum alloy 7075-T6 for which Fcy ~ 67,000 and Ec = 10,500,000.

unit. Ec will be taken as 10,600,000 whi~h is the average E for stif~ener and sheet. SUbstituting in the above parameter:.808 ) ( 64,700 ) ./. = 686 18.83x .0763 x .064 10,600,000 .

From Fig. C7.7 we read

Fes/F ey

= .76

fo---,------4v- 4. 21" ----j ~ 1-1.111 ~ =C=::l

1361 -

.068

"I
~

Thus the monolithic crippling a failing stress FesM = Fey x .76 = 64,700 x .75 = 48 ,200 pat', The curve as plotted in Fig. C7.7 is for a tits ratio of 1.0. The ratio for our panel is .0763/.064 = 1.19. The correction factor from Fig. C7.30 is 1.03. Therefore FCS(M) = 1.03 x 48200 = 49,600 psi. Load carried by one stiffener-sheet unit
: 0.808

::--=i

I'L .c, g. axis

x 49600 = ~ lb.

COLUMN STRENGTH
Fig. C7.31
C7.25 Column Curve for Members With Unstable Cross-Sections.

Stiffener area = 0.538 in. a Sheet area = 4.21 x .064; .270 in. a Total area (A) per stiffener unit = 0.808 in. a Radius. of gyration of stiffener alone = 1.123 in.

Chapter C2 dealt With the column strength at members with stable cross-sections. For example, if we took a round tube with relatively heavy wall thickness and tested various lengths in compreSSion to obtain the falling stress and Solution: then plotted these stress values Fc against the slenderness ratio L'/p of the member, the test Since the elements of the Y stiffener have results would closely follow the curves ABFC in different thiCknesses, the effective tw by equation C7.29 is needed. Fig. C7.32. The type of failure would be elastiC overall bending instability at stresses Z biti between points 8-e and inelastic bending = instability at stresses for most of the range Zbi EA. Euler's equation as sha'Nn in Fig. C7.32 1.111 x .136 + 1.432 x .068+ 3.2x .064 + 1.188 x .064 can be used to determine the f3.lling stress tor 1.111 + 1.432 + 3.2 + 1.188 both elastic and inelastic bending instability with the tangent modulus Et being used in the inelastic stress range. or tw = .0763 inches Since stiffener and sheet have different material properties, an effective Fcy from equation C7.30 will be calculated. Now suppose we test various lengthS or a member composed of the same material as used tor obtaining curve ABC, but use a member with an Open cross-section, such as a channel, hat section, etc., with relatively small material Fey = J 67,000 53000 (.:..Q2.)-1! /~ = 64700 thiCkness. The test results for such members \ .064 .064 would often follow a curve similar to DEFC in FiS;. C7.32. Thus it is obvious that Euler's The curve in Fig. C7.7 will be used to soj.ve the equat i on for Fcs(M). The lower scale equation cannot be used in the range OEF, as the true failing stresses are far less than that parameter for Fig. C7.7 modified by twand FCY given by the Culer column equation. A short is, length of the member With L'/p less than 20 will A fail at practically the same stress, thus the g t ts failing stress for lengths up to L '/0 = 20 will be practically the same and this stress has been The value of g from Fig. C7.22a is 16.83 given the name of crippling stress (Fc s) and the as the average value for a 6 stiffener panel previous portion of this chanter has been con-

l'

C7.22

CRIPPLrnG STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS rn COMPRESSION

FC""(L:L/p)~

n'E t

\
\

\
\
\
E

o
.\

The equation gives a parabolic curve trom the crippling stress a~ LI/o = 0, and becomes tangent to the Euler curve at a stress value equal to one half the crippling stress. Fig. C7.33 showS a plot or equation C7.31 for aluminum allOY material for various rdlues of the crippling stress which is the Fe stress at L'/p = o.
startlr~

Euler

Curve

Crippling Stress Region

I
I

1'[ .2 E Fe = (L:L/ p )~

ThiS method is quite Simple to use as the only additional carcutat i on reow.rec is the cripplIng stress of the column s~ction which is obtained by methods previously explained In this chapter. The other method~ usually involve the buckling stress as well as the crippling stress. Since the criPpling stress is normally constant below Ll/p = 20, the assumption that Fcs is zero at L'/p = 0 is slightly conservative.
=OD2.

Fer

1- Transition Region
L:L/p
F",. C1.32

_I

The following method or slight variations or It appears in the structural design manuals of a number ot aerospacd com~anies. The method or procedure for determining the column failing stresses in the so-called transition range involves the use of the basic column curve for stable cross-sections. The procedure can best be explained by reference to FIg. C7.34.
A
+--

cerned with methods of calculating this local crippling or failing st~ess. At point (F) the elastic bucklin~ of some part of the crosssection begins. Between stress points F to E the action ror ths member involves both overall elastic bending instability plus loc~l buckling which bec~es more extensive as the stress increases. The portion EF of the column strength curve is otten referred to as the tran5iti~n range. At present no relIable theoretical theory has been developed for determining the railing stresses in this transition range, thus resort Is made to semiempirical methods which have been checked against test reSUlts and round to give reasonably close results.
C7.26 Methods Used for Determining the Column Failing Stress in the Transition Region.

Fey

Column Curve 1'[2 Et F C "" (L:L / o ) .2

Fcs---....J{

FpL

-,
"Fer

ME.'!'HOD 1.

JOHNSON-FULER EQUATION.

.9Fcr - . -

POSSibly the first method used In calculating the column taIlIng stress Fc In the transition range EF In Fig. C7.32 was the well known Johnson-Euler equation which involves the cripplIng stress. The equatIon is,
F
Fc z F cs a

c
L:"/p

Fig. C7.34

4~~E {L'/p).2

where, Fe = col~ failing stress (pSi) Fcs = crippling stress, assumed to occur at L'/p" 0, where L' - Live.

The curve ABC is the Euler column curve for a column with a stable cross-section and for a given material. It involves USing the tangent modulus Et in the inelastic stress range. The following steps are taken to determine the column curve for the so-called transition range:-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.23

(1)

Locate point (J) on the basic column curve by drawing a horizontal line through an Fe value equal to FCY J the yield stress ot the particular material being used. Draw a horizontal line starting at paint D. Point D is at an Fe value equal to the (Fes) crippling stress tor the column section being considered. Point (E) on thiS line is determined by projecting vertically downward tram point (0). Locate point (F) at a value at Fe = .9 Fer Where Fer Is the buckling stress tor the cross-section. Draw a horizontal line through point (F) to intersect the column curve at pOint (G). Connect points E and G with a straight line. The line EO then represents the col~ taIling stress Fe for values at Lt/p between points E and G.

METHOD 3.

Another method that is widely used also uses a parabolic curve to represent the column strength in the transition range (see Ret. 12) The parabolic approximation has the tollowing torm:-

(2)

_c_ = 1 _ (1
F

- - - - - - - (C7.32)

cs

(3)

where Fc is the column tailing stress. Fcs Is the crippling stress. Fer Is buckling stress for the column cross-section. FE 15 the Euler column stress tor the particular column being considered as tound tram equation FE = naE/(L'/p) 8 The equation applies tor Fc > Fer. For cases Where Fer> FpL where FpL is the proportional limit stress tor the material use FpL instead at Fer in equation C7.32.

(4)

This method requIres the use ot the column curve for sta~le sections and the determination of the buckling stress and thus requires more calculations than Method 1. It also requires ~ gra~hical construction.

....,

1'1 ,"I 'J:,

.... U. v ' '. ,

C7.24

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET~STIFFENER PANELS IN COMPRESSION

C7.27 Example Problems Involving the Finding of the Column Strength of Columns With Unstable Cross-Sections.

C7.33 for a value of L'/p = 60 Fc = Fes = 38,200 at L'/o = o. Column Strength

~r.d

a value cf

oy

nethod 2

PROBLOJ'! 1.

A rectangular tube 21 inches long has the cross-section as shown in Fig. (a). The materi~l 15 alumInum alloy sheet 7075-T6, Fig. a with Fcy ~ 67,000 and Ec = 10,500,000. If th= member bas a pinned end condition, what is the column falling stress.

This ~ethod requires a graphical construction which involves th~ calUDL~ surve for the given material and for' a stable er-osesection. Fig. C7.35 shows the colum~ c~rve for ou: material for a column Nith staJle crJSs-sec0ic~. It is Identi~al to Fig. ~2.l0 of C'napter 82 for the roam temperature condition, except the curve in Fig. C7.35 has been dra~m to a smaller vertical scale.

Solution. The area at cross-section = .234 in. 3 The least radius ot gyration 15 0.42 Inches. Betore the column failing stress Fe for the 21 inch length can be found, the crippling stress Fes and the buckling stress Fcr must be determined.

The crippling stress will be calculated by the Gerard Method. The parameter tor use with -Fig. C7.7 is,
~ (&:).1"= li
gt

Ec

12 x .04 3 lO,506,00C

.234

67,000

)'1"

"'.

097'
b

From Fig. C7.7, Fcs/FCY .57, hence Fcs .57 x 67,000 = 38,200 psi. The initial buckling stress Fcr will be determined by the theory ot Chapter CS.

20
10

10

20

30

40

50

60

70

80

L'/p

= 1 - .08 = .92, h = 2 - .08 = 1.96 b/h = .92/1.96 = .479, tb/th = 1.0

b

The graphical construction to obtain the column curve in the region be~Neen the cr1~pling stress and the bUCkling stress is as follows;

From Fig. CS.S of Chapter C6, K h = 5.2

Fcr

12 (1 _ V 3)
3

IS,

n"E c e

(b"

h")

Fer

=5.2n 12

x 10,500,000 04 3 11- .3') (1.92) = 21,500 ps i .

Column Strength by MethOd 1 (Johnson-Euler eq.) L = 21 inches

L'

= Live' 21/vl = 21

L'/p = 21/142 = 50
The Johnson-Euler equation is,

At a column stress Fc = F c = 67,000, draw a horizontal line to intersect "he bas tc c,Jl"WDIl curve at point (0)."' Draw a horizontal line tram point (D) which equals the crippling stress of 38,200 pSi. Locate pOint (~) by prOjecting vertically downward from point (0). Locate point (0) at an Fe stress equal to .9 fer = .9 x 21,500 = 19,350 psi. Connect points (E) and (G) by a straight line. This line represents the column failing curve for the member cf our problem far L'/p values between ooints (E) and (~). Taking L'/p = 50, we ~roject upWard to EG line and then horizontally to Scale at 19ft ta read Fc = 27,700 pSi. Column Strength by Method 3 In this method equation C7.32 1s used.
1 (1

Fc = Fcs
Fc

substituting, 38,200 3 - 4 n3 10,500,000 (50)a ~ 29,400 psi

= 38,200

-----(A!
Fes
~

This value could be read directly from Fig.

Fcr = 21,500,

38,200

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.25

FE is the Euler 80lumn stress for column with a stable cross-section. To find FE, we project upward from L'/p = 50 to the curve AOC in Fig. C7.33 and then horizontally to read Fe = FE ; {1,700 psi. Substituting in eq . (A),
F 38200 ,

SubstitutIng in Johnson-8uler equation,

Fc = 30,600 -

4 n:1I

x 10,500,000 .'-'.

30,600

lJ

("-"

8,,2_ 25 ' 0 ,v6 ,

I ,

psi

Column Strength by Method 2 FIg. C7.35 paint S' i5 located at ~ 30,600. Point a r 1s located at .9 x 20,900 = 18,800 ps t . The line E 'G r is the column curve. For L'/p = ~S.3, we project upward to line E'G' and then horizontally to left scale to read Fe = 25,000 psi.
On

= 1 - (1

21,500/38,200)(21,500/41,700)

Fc

Fes

F c ; 38,200 (1 - 0.825) ; 29,500

psi

Example Problem 2.
A formed Z section as shown in Fig. (b) is used

j::':;4i

Column Strength by Method 3

as a column. The length L ; 30'1. The member Is braced in the x-x directi0r thus column bending failure must occur about x-x axis. Material is aluminum alloy, Foy = 57,0~O, Ec = 10,500,000. Assume the end fixity coefficient c = 1.5.

.04~I ...T

T
1.5

x--- 1-* bw

To find FE, use Fig. C7.33 with L'/p = 45.8 and basic ~~ler carve, gives a v~~~lue of FE = 47,000 psi.

_I
FIg. b

1-.75-1

Substituting in equation (A):30,600

Fc
1 (1

20,900/30,600)(20,900/47,000) (1 - .141)

=30,600

=26,300

psi

Find the column failing stress. Solution:

Area of Section; 0.117 tn ," Px = .535 in.

C7.28 Column Strength of Stiffener With Effective Sheet.

The initial buckling stress is needed in finding the column strength.

bw = 1.5 - .04

Column members are often attached to thin sheet by rivets or spot welding. If the rivet spacing-is such as to prevent inter-rivet buckling, then the sheet will assist the stiffener to carry a compressive load and to neglect the sheet would be too conservative. If th~ attached sheet 15 relatively thin, that Is, less than the stiffener thickness, the method of using the effective sheet width as a part of the column area Is widely used by structural deSigners and ~111 be used in this example solution. Example Problems. For an example problem, we will assume that the Z stiffener in Problem 2 is one of several stiffeners riveted to a sheet ot .025 thickness and of the same material as the stiffener. Solution.

= 1.46,
= 0.5,

bf

0.75 - .02

= 0.73

bF/bw
YoW

= .73/1.46

tw/to

= 1.0

= 2.9.

From Fig. C6.4 of.Chapter C6, we read

Fer = 2.9 11'"

12

t 10,50~iOOO
1 - .3

(1.c::,) '" = 20,900 ps i

~o

The crippling stress will be calculated by the Gerard Method USing Fig. C7.9, which applies for Z shapes. The lower scale parameter Is,

1. ~=T
Fig. c

From Fig. C7.9, we read, Fes/F ey = .455, hence, Fcs = 67,000 x .455 = 30,600 psi. Column Strength by Method 1
L' = L/ve = 30 -11.5
L'/p

= 24.5

= 24.5/.535 = 45.8

Th~ rivet or spot weld spacing is ~de such as to prevent inter-rivet buckling. Thus the column area will be as shown in Fig. 0, namely, the stiffener area Plus the area ot the sheet for the effective width w. Since the effective shaet width W 1s a !Unction of the stiffener stress and since the stiffener

C7 26

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND

SHEET~STIFFENER

PANELS IN COMPRESSION

stress 16 a fUnction of the radius of gyration, the design procedure Is of the trl~l and error category
.F1rst Tr13.1.

Where, Po = radius of gyration of stiffener alone. P = radius of gyration of sheet

~~d sti~fener .

Assume the etfective sheet width Is based on column strength at Z stiffener acting alone. The average column falling stress by the 3 methods in the Problem 2 solution was (25,860 + 25,000 + 26,300)/3 = 25,700 pst. The effective width equation to be used Is,

S = distance from centerline of sheet to neutral axis at stiffener.

t

= shget thickness.

w = sheet effective width

Equation C7.33 has been put In curve form as shown in Fig. C7.36, which will now be used to compute the radiUS at gyration Po for the stiffener plus effective sheet.

w = 1.9 t vE/FST

= 1.9x

s
= 0.965 x
= .117
.025 ;: .0242

= .75
=

.0125

= .7625
= 1.425

.025x/lo,SOO,OOO/2S,700. ,965 In.

Po tor stiffener alone = .535

S/po .7625/.535

nrrect ive sheet area

.trea of stiffener

= Ao

W';I .965

Total area = .1412

wt/Ao
Adding the effective sheet to the stiffener will change the radius of gyration. Mr. R. J. White (Ret. 13) has developed equation C7.33 which gives the variation in the radius of gyration in terms of kn~Nn variables for any stiffener cross-section. Since the failing stress at a column is directly proportional to the radius" or gyration squared, equation C7.33 can be equated to the ratio of the column stresses.

,965 x ,025/,117: .2063

~

From Fig. C7136 for the above values of S/po and wt/Ao ' we obtain .775, whence p~ = .535~ x .775, which gives p = .471 in.
(p/po)~ ~
-

Then L'/p = 24.5/.471 = 52.

Use Method 1 for column strength;-

= +

1 [1

(S.~ wt = !....

po-'I Ao

FST

- - - - - (C7.33)

Fig. C7. 36 (Ref. 13) Curves for the Determination of

O \ "_ ,,&, (-Po) - FST

2.0

:--0

"[!; .... ,

"I
e,

~I.O
t-3C.~Ff
STIFFENER ALONE AO - Area of Stiffener Alone Po Radius of Gyration of StUfener Alone

"-I ~

.
,

"I" ,..
00

"~
I I

1.0

2,0

3.0

4.0

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

C7.27

The revised effective width based on this stiffener stress is,

(4) Same as (2) but subjected to 300 0F for 1/2 hour duration.

w = 1.9 x .025 VIO,500,000/24,500 = .98 in.

This value is only .02 inch more than the value of .96 previously, thus the effect on the radius of gyration p will be negligible. The column failing stress for the sheet stiffener combination is therefore 24,500 pSi, and the compressive failing load would be 24,500 x .1412 3460 Lbe,

18~
(4 )

I-:- 1-1(2-1

r[I

-1

Fig. f

l05

"C7. 29 SheetStUfener Panels With Relatively Heavy

Find the crippling stress for sections at Figs. g and h when formed from 2024-T3 and 7075-T6 aluminum alloy material.

Sheet Thickness.

For Z, Y and hat Shaped stiffener-sheet panels as previously discussed relative to local failing stress, use entire sheet and stiffener area in computing radius of gyration or, in other wordS, due not use effective sheet widths but use entire sheet as effective.
PROBLEl1E
(5)

:a T
Find the sections as given Material

11-3(4:::'l

.05

r
1
I.!. 4

Fig. h 1/8

F7(8

-l

(1)

Find the crippling stress for the angle sections of Figs. (a) and (b) when formed from the following materials. Use both Needham and Gerard Methods. (L) 2024-T3 aluminum alloy. (2) 7075-T6 aluminum alloy. (3) 17-7PH(THI050) stainless steel. (4) Same as (3) but sUbjected to an elevated temperature of 700 0 F for 1/2 hour duration.

crippling stress for the extruded numbers 4, 8, 16, 22, 28 and 34 in Table A3.16 of Chapter A3. 2014-T3 aluminum alloy.

(6 )

!~Ig.a
7(8

fLL:,g.b
7/S".l..

1.

I- 7(8 ...., ~064

16

1.

.072

Fig. i shows a corner member, in a stiffened wing section. The skin Is fastened to the stiffener by one row of 1/8 diameter rivets at 3/4 inch spacing. The web Is fastened to stiffener by 2 staggered rows of rivets, with rivet spacing in each of 1-1/8 inch, and rivets are of the flat head type. Material Is 2024-13 aluminum. alloy. Also use 7075-T6 aluminum material. Find crippling stress for stiffener. Will inter-rivet buckling occur. Find effective sheet widths and total failing load unit will carry.

16

1.

~ '-1/2

--IT

(2)

Tf
7/8

Find the crippling stress for the two channel Shapes of Figs. c and d when formed of following materials. (1) AISI 4130 steel, heat-treated to Ftu = 180,000. (2) Same as (1) but SUbjected to an elevated temperature of SOOoF for 1/2 duration. (3) TI-8l1n Titanium. (4) 7075-T6 aluminum alloy.

1-3(4
Fig. c

1.

16

Tr!lr
3(4

1--'-1/4~

16

.04

1...

Fig. d

...."".064
(3 )

(7 )

Find the crippling stress of the rectangular tubes in Figs. e and r when formed trom folloWing materials. (1) 2024-T3 aluminum alloy. (2) HK3lA-O magnesium. (3) Ti-8Mn Tltansium.

Determine the compreSsive failing strength of a short Z stiffened sheet panel unit such as that illustrated in Fig. C7.29. The various dimensions are as follows. Refer to Fig. C7.29 for meaning of symbols.

tw = .072
ts

bw

= 2.50

= .072

bF bA co

= 0.625 = 0.375

= 1.0

bs

= 2.50

C7.28

CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS

rn COMPRESSION

Stiffener and sheet material is 2024-T3 a Lumtnum alloy. Rivets are lie diameter Brazier Head type 2117-T3 ~~terlal and spaced 7/B inch. (8) (9)

{5J Windenberg, D. F.:pp 54-en, 1939.

Proe. 5th International Congress f0r Applied Mechanics.

(6)

as (7) but change material to 7075-T6 aluminum alloy.

Sa~e

Experimental Study of Deformation and Effective Widths in .~ially Loaded Sheet Stringer Panels. NACA TN 684, 1939. Gerard, F.:- Effective Width of Elastically supported Flat Plates. Aero. Sci., Vol. 13, Oct. 1946. Jour.

same as (7) but change material to

Tltanlun.

Tl-~n

(7)

(10) In the Example Problem on a Y stiffened panel as given in Art. C7.25, change sheet thickness from .064 to .081 and calculate the resulting panel strength. (11) The hat stiffener in Fig. g or Problem (4) 1s one of several stiffeners riveted to skin or thickness .032 and of the same material as the stiffener. It the length of the panel 1s 20 inches, what will be the column failing stress if end fixity c = 1.5. Also for c = 2.0. Use method involving effective sheet widths. (12) Same as Problem 11 but use Z stiffener ot Fig. h of Problem 4. References
(1) Needham, R. A.;The Ultimate Strength or

(8)

Howland, W.L.:- Effect of Rivet SpaCing on Stiffened Thin Sheet in Compression. Jour. Aero. Sci., Vol. 3, Oct. 1936. Argyris, J. H. and Dunne, P.C.:Structural prinCiples and Data. Handbook of Aer-o, , 4th Edition, 1952.

(9)

(10) Bizlaard, P. D. and Johnston, G. S.:CompreSSion Buckling of Plates Due to Forced Crippling of Stiffeners. FTepr~nt No. 408. 8.M.F. Fund Paper. lnst. of Aero Sci. Jan. 1953.

Aluminum Alloy Formed Structural Shapes in Ccr,npression. Jour. Aero~ Sci. Vol. 21 April, 1954.
(2)

(11) Semcniau. J. W. and Peterson, J. ?:An AnalYSiS of the Stability and UltL~~te Compressive Strength of Short SheetStrin~er Panels with SpeCial Reference to the Influence of Riveted Connection Between Sheet and Stringer. ~4CA TN 3431, March, 1955.
(J.2) Gerard, G.:-

Gerard, G.:- Handbook of Structural Stability. Part IV. Failure of Plates and ca~pasite Elements. NACA TN 3784. August, 1957. Gerard, G.:- The Crippling Strength of Compression Elements. Jour. Aero. Sci., Vol. 25, Jan. 1958. Coan, J. M.:- Large Deflection Theory for Plates with Small Initial Curlature Loaded in CompresSion. Jr. Applied Mech., Vol. ia, J'..Ll'le 1951.

Handbook of Structural Stability, Part V. Compressive Strength of Flat Stiffened Panels. NACA TN 3785. Aug. 1957.

(3)

(13) Thesis by R. J. ~hite:- The Ultimate compress t ve Strength at Thin Sheet t'!etal Panels. Appendix A. C.I.T. 1935.

(4)

CHAPTER CB

BUCKLING STRENGTH OF MONOCOQUE CYLINDERS

This :hapter presents information on the buc~llng st~ength of ~lr~ular cylinders under ccapress tve , bending and torsional loads acting separately and in combination, without and with internal pressure. Sqme information on the buckling strength of conical cylinders 1s presented.
ca. 1 Introduction.

Before the advent of the high speed aircraft and particularly the missile and space vehicle, the use of the unstlttened cylinder or monocoqu8 type of structure was quite limited. However, the arrival of the space age has caused the thin walled cylinder to become important in the design of missiles and space structures. The classical small deflection theory which has proved adequate for determining the buckling strength of flat sheet structures as ~overed in Chapter Ale, has not proved adequate for determining the buckling strength of thin walled cylinders or curved sheet panels since the theory gives results much too high when compared with the experimental or test results. A more recent large deflection theory has shown closer agreement with experimental results, however, the theory requires a prior knowledge of cylinder imperfections. As a result ot the theoretical limitations to date, th? strength design of such structures is based primarily on best fit curves for experimental or test results, using theoretical parameters and eurves when they appear applicable.

for the strength design of monocoque structures indicated that the design curves given in (Ref. L) were widely used. It thus seems appropriate to give some deSign material from
(Ret. 1).

Cylinders are usually given four classifications relative to length:-

(1)

Short Cylinders. Short cylinders tend to behave as flat plate columns {cylinders with infinite radius}. They develop buckles in a sinusiodal wave form similar to nat plate buckling. The end fixity has considerable influence. Intermediate or Transition Length of Cylinders. Buckling involves a mixed pattern combining slnusiodal_and diamond shapes. The effect of end fiXity and cylinder length are only of nominal importance. Long Cylinders. Such cylinders buckle in a diamond shaped pattern, and the length and ftnd conditions are not of much impor-tance ,

(2)

(3)

(4)

A missile structure in handling operations and flight maneuvers 15 subjected to tensile, compressive, bending and torsional load systems. The cylinders may be pressurized or unpressurized. The structural design of such structures thus requires a knowledge of the bUCkling strength under the various load systems, acting separately and in co~bination. This Chapter will be confined to giving test data and design curves for the buckling strength of cylinders under the various types of s~ress systems.

Very Long Cylinders. Such cylinders buckle by over-all column instability or act as a Euler type column.

Fig. CB.l shows a photograph of the b~ckle failure of an intermediate length cylinder under axial compression. The name of Donnell is prominent in the development of a theory for the buckling strength of cylinders. From (Ref. 2), Donnell's eightorder differential equation can be used to provide a snall-deflection theoretical solution to the behaVior or cylinders 1n the short to long length range. A solution of Donnell's equation by Batdorf (Ref. 3) gives the critical stress in terms of the buckling coefficient Kc, which for sim,ly supported cylinder ends 1s defined by the equat1on:-

ca.2

Buckling of Monocoque Circular Cylinders Under Axial Compression.

The term monocoque cylinder means a thin walled cylinder without longitudinal skin stiffeners or transverse intermediate frames attach~d to cylinder skin. Investigation by the author of the procedure used by a n~~ber of aerospace companies

Kc = [(m" +p")"/m"] + [12Z"m"/nO(m" +p")]

- C8.1

where,
!IT.

= number of half waves in

longitudir~l

cs.r

direction

,:: ~ .~ :-.:~~;~;'t~~::'i-;~~~: ~: ~,-:,:,,- .," " 2~:" ,; '~:~'.', ,.<:.:,',' :~",.::)~;'-,:;~ . "::" _ :~:.'~\::' . ~.~,' >;;'L ~~':~ ;"~'
"

'-

--

----..-,.---.---ca.2

----_.~-

BUCKLI:-lC STRE:-iGTH OF

MONOCOQUE

CYLINDERS

Fig. ca. 1 Type of Failure Axial Compression and No Internal Pressure. (Ref. 1.)

Fig.

ca. la

Type of Failure Under Axial Load With Internal Pressure. (Ref. L)

Fig. ca. tc Type of Buckling Failure Under Pure Bendtng Load. No Internal Pressure. (Ref. 9.)

Fig. CS.Id Type of Buckli.ng Fiilure Under Pure Bendina With High Internal Pressure. (ReL 9. l. -

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

cs.s

fJ = L/,\
A = half wave length of buckles in
circumferential direction. The compressive buckling stress is given by,
...~ -: t Kc "-(" Fcr/T] = 12(1 - V') TI

Observation shows that the theoretical results are far above the test values. Fig. C8.7 (from Ref. 1) is a set of def:~n curves of ~ ver-sus Z for various rlt va Lces and for 90 percent probability.
ca.3 Additional Convenient Design Charts for Determining
Compressive Buckling Stress.

- - - -

- - - C8.2

Where, t = wall thiCkness, L is the cylinder length and V is Poisson's ratio. The term ~ is the plasticity correction factor and equals 1.0 for elastic buckling. Minimization ot Eq. Ca.l with respect to the parameter (TIl" + fl") "1m ", gives the critical buc~ling coefficient for long cylinders in the following form:Kc

Figs. C8.8a and C8.8b are more convenient design charts as the buckling stress FCcr can be read from the chart, thus avoiding the calculations involved in using curves in Fig. Ca.7. Fig. C8.8a is for 99 percent probability and 95 percent confidence level and Fig. C8.ab for 90 percent probability and 95 percent confidence level. The curves are based on tests of steel and aluminum alloy cylinders only. The accuracy of these curves when used for other materials has not been substantiated by tests. The 99 percent probability curves are recommended as design allowables for structures whose failure would be highly critical. The 90 percent curves can be used for less critical structures.
ca.4 Plasticity Correction.

(4v~/n: ~}Z = O.702Z

- Ca.3
- C8.4

where Z = (L '/rt I,ll-v'

Substitution of Eq. C8.3 into Eq. C3.2 reduces to the well known classical equation Fer = Cs(t/r) where, C
2

- - - C8.5
11 )

1/'1'3 (1- .3

= 0.1'305 for lJ = .30

The bUCkling coefficient for Simply end conditions in the transition range can be dete~ined by substituting the 11~lting values Ofj3 = 0 and ~ = I in Eq. Ca.l. A similar solution for cylinders with fixed edges deviates from the solution for Simply supported cylinders only in the flat plate and transition ranges.
sup~crted

The plasticity correction for cylinders in the long range (L 2 / r t ~ 100) is given by the curves in Fig. e8.9 taken from (Ref. 4). In general most practical cylinders in aerospace structures will fall in the long cylinder range.

.f---------l----,-----'---:-----::J
;11

Figs. C9.2 to C8.S inclusive taken rro~ Reference 1, show the theoretical curve which shons the buckling coef~lcient as a function of the geamet~ical par~~eter Z. Theoretically, th8 short cylinder range would occur at Z = O. Tte values of Kc of 1.12 and 4.12 for Z = 1 correspond quite closely to the buckling coefficients of Simply supported anG ~ixed ended plate col~~s. The long cylinder behavior is represented by a 45 degree sloping straight line portion of the curve. The curve con~ecting the s~ort and long cylinder r~nges is re~erred to as the transition range. ?igs. :8.2 to ~8.5 show tte plot of extensive test data and a GO per cent proba~il:!.ty curve der-tvec by t he author of (Ref. 1) by ~ statistical ~pprcach. ~1g. CS.6 from (Ref. 1) shews a plot of ~uch test data on a Ioga r ; 'thmt c char-t -0:' :-/-: ver-sus tee :or- Z z; 10,000. A best :it curve and 90 and 99 ~er cent ~robabllity derived cur79S are also seen as Nell as the tr.eor9~ic9.1 curve :or C : .605.

{-,r-----El
" D

II

K" 1:'
Fii. ca.
~

12(1-V,,")F,.
(Ret.
~) ~ond\m..n.. lona1

(fl'

Bw::1<.Hng Chart for ,u'a.lly Compr......ed Long Circular

Cylu><len.

7/.

(E"iE) [(Ec/El\1 lI" ")/(1 -V')] - ' . .'l.ppllcaCle

Wilen L "rt 100.

ca.5 Buckling of Monocoque Circular Cylinders Under Axial

Load and Internal Pressure.

Exnerlments conducted many years azo deflnit:ly showed th~t -:~e compressive ~uckling strength af monoccque cylinders was inc~eased if internal pressure was added to the closed cylinder. Since ~el~ht savi~g is very important :n ~isslle deSign, the ~S9 of pressurized

C8.4

BUCKLING STRENGTH

OF

MONOCO

UE CYLINDERS

. ..........,

lA~.~
'/lIUION _

"

'0'

! '

90' PR06A8ILJTY DESIGN CURVE vlto:3001

"

- - " " " ' " ....".,.., EDOES -..oMP<D EDOES

"

"

..
,""""

10

.. .. ..
;
,

0'

o'

10'

z
Fig. ca. 3
-1lDT \1oIAA1 _a..- HllU

Fig. ca.2
&tU.JlSn:DT
_aICK'!' INA'" _ Q..UlIl 1I ItOU
~

rlt:: 100 to 500.

1I NOJl_

r/t:: 500 to 1, 000.

,U. _

,,,,,,,,"a l'\,M !ICMUlI

KMEMJT~

.""""

.eo"

iW' 1 I - . t

,
"

",,2 E[Ir 'f' ';,{o-p.' "

:~C1t'.l(
z
Z.~~

,
,

, "i"

: I!:'

"

, , Tl-IEORE11CAl. CURVE ,
, , ,

:!!!!
,
,

)',

I ,"

,j/

, ! l:~>
"

'i

,
',0

..

ego
"

.,,;y ,

, ,

'II,
,

I I

, , , , ,

'I

, 'i

OESK>N CURVE

"""""'UN

(r/hl:5001
'i
"

V!
,

' SIMPLY SlJPPQRTEO EDGES : : : - --et.AMPED EDGES

.fi"' ,
ic

.. ... .. .. , ..
,

,,

'l;.y

!::!

0'

i , : ilill a ,
z

u:

, '

0'

11,1 i : !1 ,

II !I!!

'"

'"

Fig. ca.4 r/t = 1, 000 to 2, 000. Fig. ca.5 rlt over 2, 000. Fig. ca. 2-ca. 5 (Ref. 1) Compressive Buckling Coefficients for unpressurtaed Circular Cylinders.
oUL>..DISTGT a >'lAGIO(lII
,~

.'
l , ,

ollUCH"!' IUo41 oQ..Alll<lI..a.1"

,,.,

''':'----''''''---'E''''''

.1.O,~a~

''''''-:IOfII.~~

.~

'Q-li

"

"

.
.:

":i.l.

..
.~

..

--~._-,._~

..
,

~--.m'CIJII"o'(
i!I'-.m'~

.,....

--......., "'""""'" ""'"

---0,.,AI0lPED EOGS
I

i!!

l !!ill

i! !1i!!
i

Fig. ca.6 (Ref. 1) Compressive Buckling Coefficients Function of rit.

J.$

a
80Z Z Z 6
IQ:J
.. 6

!O:J

Fig. car (Ref. 1) Compressive Buckling Stress Coefficients

for Unpressurized Circular Cyl inders, (90 Percent Probabtl tty)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

cs.s

-E-

Feet

I
,

1xlO".'

.,", :' - :<Jr$i:~ , .

S

t~:+
1 I t

, 'J.:.!

Ii

it:

,]1;

Fig. C8. 8a

BUCKLING

STRENGTH

OF

MONOCOQUE
~;-

CYLINDERS

"",.

FIg. CS. Sb UNPRESSURIZm. UNSTIFFENED, CIRCULAR CYLINDERS n'I AXIAL COMPRESSiON (C1:lmped Eoo.) (Taken by PentllUlon fnlm General J)yna.mici Fort Worth
Structures 1ofa=;Ll)

lxl-'L, '

01111111111111111111111111111111111
,I>

Fig. C8.8b

ANALYSIS AND DESIGN OF

FLrGHT VEHI

LE STRUCT

car

cylinders in missile structures has become a common type of missile structural design. The famous Atlas missile was one of the first to USe a pressurized monocoque type of structure. The major reason for the large discrepancy between the actual test strength and the theoretical strength by the linear small deflection theory that is generally accepted is that the discrepancy is due to geometrical imperfections and the associated stress concentrations. Now large internal pressures should smooth out such imperfections and approaCh a perfect cylinder and thus the resulting buckling strength should approach that given by the linear small deflection theory. However, much of the available test data for pressurized cylinders gives values below that given by the linear small deflection theorJ. (See Fig. C8.la for buckling action of pressurized cylinder.) One of the first experimental and theo~etical investigations of the effect of internal pressure upon the buckling compressive stress of monocoQue cylinders was by Lo, Crate and Schwartz (Ref. 5). They analyzed the problem of long 9ressurized cylinders using an extension of the large deflection theory of Von Karman and 'I'sLen (Ret. 6). Plotting their results in ter.ns of the non-dimensior;a1 parameters (p/E) (r/t)" and (Fcr/E)(r/t), they found that the bUCkling coefficient C increased from the Tsien value (Ref. 7) at 0.375 at zero 1.nternal pressure to the maximum classical value of 0.605 at (p/E)(r/t)a: 0.169. Fig. CS .10 taken from (Ref. l) shows the large deflection theoretical curve. Also shown are the experimental values obtained in (Ref. 5) as well as those obtained by the investigators in (Ref. 1) and by other investigators in (Re~. 8). FrOm Fig. C8.10 it is apparent that large discrepancies exist between the theoretical predicted values and the experimental values. La, Crate and Schwartz suggested that better correlation with test results could be obtained if the increment in the buckling stress parameter (6Fcr/E)(r/t) were plotted as a function of the pressure parameter as shown in Fig. ca.ll. The increase in the critical stress 6Fcr due to the inte~l pressure directly represents the beneficial effect of the lnterr~l preSSure. The total critical stress is thus obtained by adding the critical stress for the unpressurized cylinder to the increase in the critical stress jue to the internal pressure. In order to plot the test points in Fig. Ce.ll. it was necessary to determine the unpressurized critical buckling stress for each test. The 90 percent pr8bability values from Fig. CS.? were ~sed. As Shown in Fig. Ca.ll, the general trend of the test data agrees fairly Nell with the theoretical curve. T~e investigations in Ref. 1

have shown a best fit ourve and a 90 percent probability curve obtained by a statistical approach and this 90 percent curve in Fig. CS.ll is recommended as a design curve for taking into account the effect of internal pressure.

IO:.~~
,

,,
I
i

".' ,

1111

...

10

Fig.

ca.l0

Compressive BUCkling Stress for Pressurized Circular Cylinders. (Ref. 1)

.01 .0<

Fig.

ca. 11

Increase in Compressive Buckling Stress Due

to Internal Pressure.

BUCKLING OF

~ONOCO~UE

CIRCULAR

CYLINDERS UNDER PURE SENDING

ca.5

Introduction.

~llght vehicles are subjected to forces in flight and in ground operations that cause bending action on the structure, thus lt is necessary to know the bending strength of cylinders. ~NO rather extremes have been used in past deSign practice. One deSign assumption takes the value of the bending bUCkling stress as 1.3 times the buckling stress under axial compression. The other assumption is to ass~e the bending buckling

ca.s

BUCKLING STRENGTH OF MONOCOQUE

CYLINDERS
.~.

stress is equal to the axial compressive buckling stress. The ~irst assumption is cons tder-ed by some designers as somewhat unconservatlve while the second assumption is no doubt somewhat conservative. It is relatively recent (Ref. 10) that a small deflection approach has been completely solved for a cylinder in bending. Tests of cylinders in bending show that the theoretical result is lower than the test results but higher than the buckling stress in axial compression. No large deflection analysis which involves a consideration of initial imperfections of the cylinder has been formulated to date for the buckling strength in bending. Since the stress in bending varies from zero at the neutral axis to a zaaxtmum at the most remote element, the lower probability of imperfections occurring within the smaller highest stressed region would lead one to conclude that higher buckling stresses in bending J as compared to the buckling stress in axial compreSSion, should be expected.
ca.7 Available Design Curves for Bending Based on Experimental Results.

00N<ll

........
41,'

.~

,
~o CI liIL,_
0

(3'~~=SWf:_

'l 't ~ !

c:

01

,~~~

- ~~i
,

imo.-:
I
.~_.

.- _.

~'II!

.,

'--:"'1-1 """

Fig. ca. 12 tjnpressurtzed Bending Buckling Stress Coefficients as a Function of r/t.

The same investigatcTS (Suer J Harris, Skene and Benjamin) that carried out tests on cyll~der in axial compression (Ref. 1), have also carried on an extensive investigation of the buckling strength of ~onocoque cylindrical cylinders in pure bending (see Ref. g). As originally developed by ~lugge (Ret.ll) :or long cylinders J the buckling stress in bending is expressed as:Fb

percent probability and confidence level of 95 percent. These curves are from the structures manual of the General Dynamics Corp. (Fort Worth.) Their manual states t~at most of the test data upon which the curves are based fall Within the range or cylinder dtmensions .25 < L/r < 5, and 300 < r/t < 1500, and the curves are based on tests of steel J aluminum and brass cylinders only.

ca.a

Buc!ding Strength of Circular Cylinders in Bending With Internal Pressure.

The Published information on the buckling strength of circular cylinders in bending with internal preSsure is very limited and the status of theoretical studies to date leave much unknown regarding this SUbject.

Reference 9 gives the reSUlts or a series or tests of circular cylinders in bending With internal pressure. Fig. C8.l4 is taken fro~ The theoretical vlue of the bending bUCkling coefficient as found by Flugge was that published report. In ~ig. CS.14 the exabout 30 percept higher than the corresponding perimental data are Plotted in ter,ns of the classical bUCkling coefficient of 0.605 in increment ACbp to the buckling coefficient Cb' The increase i~ the bUCkling stress coefficient axial compression. ACbp represents the beneficial effect of Fig. CS.l2 (f~om Ref. g) gives a plot of internal pressure. The tctal value of the considerable test data and a plot of Cb versus buckling coefficient is obtained by adding the (r/t). A best fit curve J a 90 percent probabuckling Coefficient ror unpressurlzed cyli~ders ~illty curve and a 89 percent prObability curve J to the increase in the buckling coefficient cue are shown. T~e dashed curve is a plot of the to the internal pressure. In order to plot the 90 percent probability curve as previously data, it was first necessary to determine the given in Fig. e8.S for buckling in axial comunpressurized buckling coefficient for each specimen. The 90 percent probability deSign p~essionJ thus ~lving a comparison between curve of Fig. CS.12 was used ror this purpose. bending ann compressive buckling strengths. As The direct benefit ot lateral internal pressure indicated by the ~lotted test points J the test to the stability ot cylinders in bending is data above an rlt value of 1500 is quite indicated by those speCimens with no net axial li~itedJ thus the "accuracy of the curves is stress (the balanced speCimens) represented by somewhat unknown. the circular symbols. At large ';alues of the Figs. C8.l3 and CS.l3a give convenient ~ressure yarameter, the additional benefit of deSign curves for finding the bending buckling ~he axial pretention is clearly demonstrated by stress based on 99 percent prObability and 90 the large increase in 6Cbp of the pretensioned
- - - - - C8.6

cr = CbE(t/r)

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

e8.9

. ,I'

,:t~ 'i

i, ,

1xl0

Fig.

cars

-----:3.10

----._.",.~.

---STRE~GTH

BUCKLING

OF

MONOCOQUE

CYLINDERS

1xlLJ

-. illill

1~,1111

' '"'0--

--,

1 I

_ ~:~#:~
Z

I ! t

lJC.10 1

I lX,WJ
Fig.

ca. 13a

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C8.11

speCimens, represented by the triangular symbols, over trat for balanced specimens. The limiting value of the increase in the buckling stress coefficient for pretensioned pressurized cylinders at very high values of the pressure parameter is given by the line o~r ~ pr/2t. The analysis of the pressurized cylinder data was achieved by selecting a best ~it curve tor those specimens in which the axial pretension was balanced. This curve (shown in Fig. CS.14) was selected by the investigators as best indicating the general trend of the experimental data. At large values of the pressure parameter, the curve is drawn to approach an asymptote agreement bevNeen the best fit curve and experimental data is apparent in Fig. C8.l4. A statistical analysis of the test data was performed for the speci~ens with no axial pretension to establish the 90 percent probability design curve shown in Fig. C8.l4. Because data were available only fram the tests made by the investigators, they indicate the sample may not be representative and a lower probability curve should perhaps be used for design purposes. The data was not considered sufficient to permit a statIstical analysiS of the pretensioned test data, and therefore they suggest a lower bound curve be used in ~he design of pretensioned cylinder. Additional tests are needed too for unpressurlzed cylinders with r/t ratios greater than 1500 to verify the sr~pe of the design curve.
"
"

The bUCkling compressive stress for this type of loading is given by the following equation (Ref. 12):k o 11 2 E t a = 12 (1 - Ve 'J (Ll -

cs.c

Values of the buckling coefficient k p are given in Fig. C8.1S. Equation C8.7 is for bUCkling stresses below the proportional limit stress of the material.
C8. 10 External Radial Pressure.

Under an inward acting radial pressure only the circumferential compreSSive stress produced is f c ~ pr/t where p is the pressure. The buckling stress under this type of loading from (Ref. 12) is,

= 12 (1 -lie "J(Ll

ky

rt

2E

C8.S

Values at the buckling coefficient k y are given in Fig. Ca.l6. Equation C8.a Is for buckling stresses within the proportional lImit stress of the material.
C8. 11 Buckling of Monocoque Circular Cylinders Under Pure Torsion.

I 0 .IJQ4t.

A,llJ,lij. ~ ~

ST1'lES$ \11"/21 EW..AN:EO $\' CCNFI'ISSVE I.CAO \11"/,1 I '

l.C'foal !'nINO 0JlM:

:!

Iii

Fig. Ca.l? (from Ref. 12) shows the results of tests of thin walled circular cylinders under pure torsion. The theoretical curve in FIg CS.l? is due to the work ot Batdorf, Stein and Schildcrout (Ref. 15). Their theoretical Investigation utilIzed a modified for.n of the Single eqUilibrium form of Donnell (Ref. 2) and by use of Galerkins Method obtaIned the curve shown in Fig. C8.17. This theoretical curve falls above the test results and thus for safety a lowered curve should be used for design purposes. Fig. Ca.lS shows a desi~n curve which appears in the structural design manuals of a number of aerospace companies.

"0

The torsional buckling stress is given by the equation:- - Ca.9 Fig. Ca.lS gives the value of the torsion bUCkling coefficient kt and applies for bUCkling below the proportional limit stress. To correct for plasticity effect when buckling stress 15 above the proportional limit stress, the non-dimensional chart of Fig. C8.19 can be used. Figs. CS.20 and C8.21 give other convenient deSign c~rves involVing buckling stresses for 90 and 99 percent probability.

Fig.

ca. 14

Increase in Bending BUckling Stress Coefficients Due to Internal Pressure.

SUCKLING OF

MONOCO~UE

CIRCULAR

CYLINDERS UNDER

~TE.'qNAL

PRESSURE

ca.9

External Hydrostatic Pressure.

Under this type of loading, the cylinder shell is placed in circumferential compresSive stress equal to twice the longitudlnal compressive stress.

---~.L

--'.-

carz

BUCKLING STRENGTH OF MONOCOQUE CYLINDERS

10J.----------~----------....,...---"/_,

r,/frll-Ji'/ ~
/O' __-----c---------------<>-:""-"""'-----i

/0

TEST DATA

o
c

WfNDENBERG AND TRILLING

SnJRM

I I

10
ZL

10'
L' r-.,.,...= ::-J rt 1 - lI,'

"

Fig.

ca. 15

Buckling Under External Hydrostatic Pressure.

IOJ

r=-- - - - .,..-- - - - . .,. - - - - - - - - - - .,..-- - - -r-..,

rr/frY 1-1'/

10'r------1'------'------------~(9-"'------1

/0

_OF IlArOORF, --~,....."'------_T----------1

DA~

OF STURM

I I

"

10

10'

Fig. Ca.16 Buckling Under External Radial Pressure.

,-

-~
.,~,.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

ca.rs

c,-----

-,

k.

0 0 0

_~9

- -" -- ~

.......
-,
".

-r:~.

z,
Ftg. C8. 17 Comparison of Test Data and Theory for Simply Supported Circula.r Cylinders in Torsion.

10 3

\l

"

'1'_.

10:01

10

-,c---=;U--"~H

For Transverse

..:~_:c_:------_
- -- - - -'-~-- .:<

Shear-

l.0

10

10'
ZL

10'

iO:---------------l0

L' 1 =rt/

Ve'l

Fig.

ca. 18

Buckling of Simply Supported Circular Cylinders in Torsion or Transverse Shear

---.,.---- - - ca.l<!
BUCKLING STRENGTH OF MONOCOQUE CYLINDERS

7,----------------....,---------,",
,
f-------------------~7/

"
F Scr

-~""~~~~~~;

f-----'---'--ff7:5~r____:_--------_l
1---,------,~7""-------

F 0.'" J

.2 f-----#;~---------

.2
2

kt: n E t 12 (1 -ilea) (F o . ,. ) (1:'")

:I

Fig.

ca. 19

Nondimensional-Buckling-Slress Curves for Long Cylinders in Torsion.

ca.12 Buckling Under Transverse Shear.

tit test data. was:Rst~ +

Shear stresses are also produced under bending due to transverse loads. These shear stresses are maximum at the neutral axis and zero at the most remote portion or the cylinder wall, whereas the torsional shear stress Is uniform over the entire cylinder wall. Limited tests indicate a higher buckling shear stress ~der a transverse shear loading as compared to the torsional buckling stress. A general procedure in industry 1s to increase the shear buckling stress under torsion by uSing 1.25 times kt. Thus in Fig. CS.lS tind k t tor buCkling under torsion ~~d then mUltiply it by 1.25 in using Equation e8.S to tind buckling stress under transverse shear.
ca.13 BUCkling of Circular Cylinders Under Pure Torsion With Internal Pressure.

The derived interaction equation - - - - - - - - - - - - - CB.IO

Rp

=1

where,

st

ratio of

applied torsion shear stress allowable torsion shear stress

applied internal pressure Rp~tiO of external hydrostatic buckling pressure Note that Rp has a negative Sign. The value at the exterr~l hydrostatic buckling pressure can be dete~lned by use of Fig. Ca.15. Fig. C8.22 shows a plot from E~uatlan Ce.IO and its comparison with test data.

Internal pressure places the cylinder walls in tenSion, thus the torsional buckling stress is increased as torsional buckling is due to the compressive stresses that are prOduced under shear torces. Hopkins and Brown (Ref. 13), uSing Donnell's equation. calculated the effect of internal pressure on the bUCkling streSS of cirsular cYlinders in torsion and the results were in fair agreement with test results. Crate. Batdorf' and Baab (Ref. 14) utilized an empirical interaction equation to
J

ca. 14

Buckling of Circular Cylinders Under Transverse Shear and Internal Pressure.

For this load system, the derived interaction equation is similar to Eq. Co.IO.
C8.11

where, applied transverse shear stress allowable transverse shear stress Rp is same as eXPlained in Article C8.13.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

CS.15

.F"

lxl~.1111

lxl-'f'1111
.6

'I'
Fig. C8.20

3. 16

BUCKLING

STRENGTH OF

MONOCOQUE

CYLINDERS

lxlO_J -0

'1111

lxl_O

J 1111 -3'

.....
P"
-4 lxl!l.-

.1

;1111

lxl....

-5

;Iil

6,,_, ','0 ! 1xl01

1, ",,11111111 .
, ,
Fig. ca.21

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

Table ca. L

ca. 17

'0

Summary of tlIteraetion Eq\lattons for Bucklini of PressuriZe<! and Url~re5s\lrized Circular Cylinders

Combined
~

Loading
fflOII IN1'E1fACrtON

F ster
~5i

..,"'0

ConditiOn

tnteracnoe Equation

Eq\lll.ti.on For Marn of Billety

'"

~.t'M',.

L.ongitlldlnal ..) . Compression andl He" Rb ~ 1 Pure Benatng Longitlldinal R.::~Rst~"l COmprtllaiOn and. (See Ftg. ca. 23\ Torsion t.ongibl.dtllal

!. 1 1M-So "He... Rj,L

M. S. = He ... ,j~ . m:t

I
.1

I
i
1 1

S
t

'teestcn and.
Toraion Pure Bending and 'Torsion

I(See R:t - (RtJ" L Note II

RbI.~Rst~"L

I
I
1

(See Fig.

ca. 20\

Pure Bendinl <U\d Tl"UIsverse lRb~-Rs3,," , (See Fig. ca. 23) Shear Longitudlnal Compression, Ptu'e Bending and Transverse
Sh~'

He",~ R8 .. ~ .. 1

'0

f'tnt. P en. ern. ' psi.

Fi!i. C8.22 Effeet of Internal Pressure on Torsional-Bucldmg Stress of Long Cylinders. Test Results are from Ref. (14I.

Longitudinal Compression, _ Pure Bending and. 'Torsion.

Lo~bl.dlna1

Rc"'Rb"'~t .. 1

,
M.S. "R.:: ...~ .. j{Rc" ~ ~ .. 4RSt - 1

i
~
1

Compfeallion. Pure Bending.

T~sverse

. a:t..';Rj ..<\6" L

I
i

ca.15 Buckling of Circular Cylinders Under Combined

Shea.r 1l.lId 'terewn

NOTE 1.

Load Systems.
R

It is very seldo~ that a circular cylinder used in an aerospace struct~e such as a missile is sUbjected to a load system causing only one internal etress system such as axial stress, bending stress, and shearing stress. Therefore it is necessary to be able to safelY deter.nine the buckling strength under the vractical cases of combined stress systems. ~hiS problem is handled by the use of interaction equations. Table C8.l s~rizes the interaction equations that appear in the structures deSign ~nuals of several aerospace companies. These equations no doUbt have been proven reliable by Checking against test data. Fig. C8.23 gives a ?lot of t~ee interaction curves. T~ese curves are ~seful in quickly ObserVing Whether cylinder ~s weak and to determine margins of safety.
ca. 16 illustrative Problems for Finding the Buckling

applied tensile stress t .. compressIOn bucltiing alloWiJlIe

Rt...O. a

I
1

:,i' -.

... ,-:_:--::.. _'.:,.::,~,:::"".,+- -.-

'0-::,1::' :",~.

.7

.4

Strength of Circular Monocoque Cylinders.

.3

Problem 1.

.txial Compressive Strength.

.a
o/
/

----;>-"-:-----*,)~~_2C.:~;p.;j~:~O~:.C~\L.
.._ _ ~. ,_ " : :

r.. ------ ,~-:,,'~~.:;.,i '::;L;~

A cirCUlar cylinder has a radiUS r = 50 inches, a length L = 75 inChes, and a wall thickness t of .05 inches. T~g material is aluminum alloy ZOZ4-T3, for which Ec 10,700,000 psi. '~at compressive load will it

. 1 -------;;.......~-- -~---F'tg..---cBi-2S<

Inte-raeUoo-ClaV-es,-",-;- -.,'-- I--;~_L:o:-.-_ -;'---,-

I -1<;--,-,,--;,,;;-

.1

.2

.3

.4

"",. ""

.5

.6

.7

.e

.9

LO

----~--

CB. 18

BUCKLI:-fG STRENGTH OF MONOCO

~sing

UE

CYLINDERS

carry

1eslgn curves basad on 90 percent

pr-cbab t Li ty ,

The bending moment developed at this buckling stress is.

11="' ... r' nr'il t . ocr I/r = F'.vc

. , t :: 50,' .OS ,. 1000, L/r

= 75/50

:: 1.5

:: 1710 n x 50'il X .05

= 670,000

In. lb

(See EQ. C8.2~

If we use Fig. Ce.l3a based on 90 percent probability and 95 percent confidence, ~e read for rlt : 1000 and Llr = 1.5 that Fbcr/E is .000160. Thus Fb~r : 10,700,000 x .000160 = 1710 pSi. Since it 1s difficult to read Fig. Ca.12, it is recommended that Fig. C8.l3a be used in design.

To find the buckling coefficient kc, we use F'lg. ''::3.7.

- :: ~..j 1
rt

V", a
-

z = 50 ,,:) X

... ,...:a

.05

/ 1 - . 3 2=2140
a

If we require 99 percent probability, we use Fig. C8.13 and obtain Fbcr/E = .00092, which 2140 and r.ft ::
1210

From Fig. ce.? uSing Z

....

lOCO, we read k c :: 280, then

gives Fb : 10,700,000 x .00092 = 985 pSi as cr against 1710 for gO percent probability. Problem 3. Torsional Strength.

"cr- -

_ 250 n:li x 10,700,000 (.05,a

12 (1 _ .sa)

75'

= __ PSi

The buckling axial co~presslve .oad p :: 2nrt tcr .. 2n x 50 x .05 x 1205 = 18950 lb.
Ii' we use the des rgn curves of Fig. CS.8b

Same cylinder as in Problem 1. what torsional moment will this cylinder develop. Solution. Using design curve in Fig. C8.18
75 :;j

ret:/S ... 000121. Thus F'ccr = For this value of Z. we read the torsional 10,700,000 x .000121 = 1295 psI. It we ~ultlply buckling coefficient Kt from Fig. C8.!8 to be this value by the ,95 confidence value, we
1.5, t:v.t!""
~s

based en 90 percent probability ~1 95 percent conr icsnce , we r eec for r/t = loce and L/r =

=-.J 1-)). = o -0 X. 05/1- .3" = 2140 r-c

L:;j

r--r-rr:

obtain 1230 pSi ~hich is practically the 0btained above using Fig. eg.7.

sa~e

170.

(See Eq.

cs.s:

Pa :: 1295 x 2n x .05 x 50 :: 20350 Ibs.

95
~/t
per~ent

If we require 99 percen~ probability and confidence. we use Fig. CS.8a. For = 1000 and Llr :: 1.5, we read FScrlE =

.000082 or Fs~ 2 10,700.000 X .000082 :: 877 ~Si. which woule give an ~ial b~ckli~g load of 13,~00 Its. T~as re~u~ri~g ~ d9 percent ~r?bability dec~eases the buckll~~ load

The torsional ~o~ent developed by this buckling stress is,

cone i ceracrv .
Problem 2.
The be
~sed

3er.dl~;

Strength

= Fe tcr- .2rr.r'ilt
"'35 X 2;'( x 50 a X .03

~o~ent ~ercent

s~e cyli~der as used i~ ~?blem 1 will in this exampl~ problem. ~hat bending Nill buckle this cylinder ~nder 90

... =

= 5780ClO in .1:05.

?robabil:~y.

Sol~ticn.

Th~ curve in Fig. C3.12 will be For r/t = 1000 and Llr = 1.0, we read ~r0m ?i~. :8.12 that the bending buckling ccer r i c t cnt ':b = .1.5.

The reeu Lt wl Ll, a Ie c ::3 ca LcuIa t ed us mg and CS.2l o~sed O~ GG ;ercent ~nd 90 percent pY'oba:l11ty respectively ar.c 95 percent con~icence. USing Fig.:3.21:~l~S. ~8.20

used.

For- Y'/t :: 1000 and L/r = l.5, ':;e read fst/S = 000082. 7hen ?st : .000082 X lO,~OO,cOO = 37'5 PSi. : 876 x 2rr. x so ' X .05 = 688000 ~n.lbs .
US1rr~ F!~.

::'SxlO,700,oaox .05/50:: 1710 os i

:e.20

Nh:c~

:s f?r 89 Jercent

~roba-

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

ca. 19

bility, we read ?st/E = .000060 which gives Fat = 642 ps t wht ch , in turn, gives allowable t:::rque"T = 504000 in. los. Problem 4. Combined CompreSSion and Bending.

:=

2 .527+ . 42l+y I (.527+ .421) 2+4x.13--1=035 -'--

Problem 6. The cylinder of Problem 1 is subjected to an axial compressive load P of 10,700 lbs. and a bending moment M of 282,000 in. 1bs. rAhat 1s the margin of safety under this combined loading for 90 percent probability and 95 percent confidence. Solution. The interaction equation for this type at combined loading from Table Ca.l is,
Rc + Rb = 1

Combined Compression, Bending, Torsicn and !ransverse Shear.

The same cylinder as in Problem 1 is subjected to the following 10ads:-

P = 10700 lbs. compression, M = 282,000 in.lbs., T = 715000 in.lb., V:= 1790 lb., (transverse external shear). Will the cylinder carry this combined loading under 90 percent probability.
Solution: From Table ca.l the interaction equation for this combined loading is,

Rc = P/Pe.' (Pa tram Problem 1 solution is 20350 is.: RC

= 1 - - - - - - - (8)
Re = 10700/20350 = .527

= 10700/20350 = .527
M'a =
570,000 in.lbs.)

Rb = MIMa (From Problem 2 results:-

Rb

= 282000/670000 = .421

Rb

= 252,000/670,000 = .421
= .527
tr~n

Rst = 71500/688000 = .104

The equation tor shear stress due to transverse shear load V is,

Rc + Rb is less safety.

.421 = .948. Since the result 1.0, we have a small margin at

+

I for c yl inder- section

1tI' 3t:

Preblem 5.

Com~ined

Compression, Bending and

/ydA nr-t x .6366r = 2 r't

fs
f

Torsien. Same cylinder as in Problem 1. The loading is the same as Problem 4 plus a torsional ~cment T of 89300 in. lb. Fing the M.S. under this combined leading tor 90<:t; probabill tv . Solution. interaction equation ~or this combined load system from Table Ca.l is,
~he

= nr"t
:=

(2 rllt)

:=

1'U"t
:=

1790/n x 50 x .05

228 psi.

From Article C8.12 we use buckling coefficient for transverse shear to be 1.25 times that tor torsional buckling. From Problem 3 the torsional buckling stress calculated to be Fst = 876 psi. Therefore the transferse shear buckling stress is 1.25 x 876 := 1095 pSi, = F s

- -

(A)

From Problem 4, Rc = .527, Rb = .421, Rst = T/Ta.' (From Problem 3 results: Ta is 688,000 in. lb.) Rst = 893001688,000 = .1Z
Subt , in :?,q. (A):-

Rs = fs/F s 228/1095 = .208

Substituting in Equation (B), .527 + .104.01 ,. V.208" .... 421" = .975. This result is less than 1.0 so a small margin of safety ext sts . Problem .., Combined Compression, Sending and Transverse Shear.

.527 ,. .421 ,. .13:1 = .965, (Since this result is less than 1.0, we have a small margin of safety. From Table Ca.l,

n.s.

Same cylinder and loads as in Problem 5, but without the torsional moment load. Solution: The interaction equation 15,

ca.ao

BUCKLING STRENGTH OF

MONOCO

UE

CYLINDERS

- - - - - - - -

(8)

Rc

= .527,

Rb = .42l

~nd

Prcb Lem :3)

Rs

= .~08

ca. 17 Problems Involving Internal Pressure with External Loadings.

(from The same cylinder and ma~erial as waS ~sed in Problems 1 to 9 will be ~Sed in t~e folloWi~g e~ple nroblp.ms, namely, r = 50, t = .05, L = 75, E = 10.7 X 10 's. Problem 10. Axial Compressi8n with Internal Pressure.

suct .

in Equation (C),

.527 + ~.208~ + .421 3 = .964 (less than 1.0 thus cylinder will not buckle.) Problem 8. Combined Bending and Torsion.

Using the sa~e cylinder as in ~revl:Jus problems, the combined loading 1s:-

n = 178,000 in.lbs., T = 358,000 in. lb.

~ill

The cylinder is subjected ~o an axial compressive load ot 50000 Ibs. and an internal pressure of 5 Ibs , per sq. in. What is the margin of safety under thiS combined lead system wlt~ 90 percent probability. solution. From Problem 1, the bUCkling compreSSive stress Fecr with no interr~l pressure was 1295 psi. To obtain the increase in the compreSSive bUCkling due to the inter~l pressure, we use Fig. C8.ll. The horizontal scale parameter is,
p (r _ 5 E t l - 10, 700,000

the cylinder buckle under this loading.

Solution: Rb = 470,000/670,000 = .701 Rst = 358,000/688,000 : .520 The interaction equation trom Table ca.l is,
Rb ... e + Rs t ' = 1

(.3 00 5),

".0"7

From Fig. Ca.ll, we read on vertical scale,

.7011.5 + .520' = .88 less than 1.0, therefore cylinder will not buckle. The ~rgln oi safety can be determined graphically by us~ng Fig. C8.2l as tollows:Find paint (a) as indicated on Fig. Ca.2l using Rb = .701 and Rst = .520. Draw line :rom origin a through paint (a) and extend it to intersect interaction curve at point (b). Projecting downward and horizontally from paint (b), we read Rb = .78 and Rs t = .58. The factor of utilization U = .70/.78 = .52/ .58 = .895. Therefore l1a.rgin ot Sa re't y e (l/U) - 1 (1/.895) - 1 .12. AFCcr
:=

= .21, whence
(",,?x .21 x lO,7CO,OOOx

.05)/50 = 22507'l ps t

Since the stress is below the ?roport~onal limit stress the plastiCity correction is zero or n = 1.0. Thus Arc cr = 2250 'I - pSi. Then the total buckling stress flith the internal pressure is 1295 + 2250 = 3545 pSi. Now the internal pressure prOduces a longitudinal tensile stress in the cylinder wall. This tensile stress is, Ft
~

prOblem 9.

Combined Compression and Torsion.

pr/2t

= 5 x 50/2 x .05 = 2500 pSi

tensile stress
~ust

Sane cylinder as in prev~ous probl~ms. The :oacs are P = 12900 Ibs. T = 358000 in.lbs.
~,;1".at

is t"!.S.

This walls can due to an allowable

be

cancelled before The

be SUbj:cted to a compressive stress

SoL.tion: The interaction equation from Table C8.1 is, R-: + Rst' = 1
Ro

external compressive axial load. failing load Pa thus equalS,

+ ~Fccr +

Pa = 2nrt (F ccr

Ft)
+

= 12900/20350 = .534

= 2n

x 50 x .05 (3545

2500)

=~

IJ.

Rat

=058000/588000
= .634

= .520

:-l.S. = (Pa/P) -1 = (95000/60000) - 1 ' .90

~is result shows the tremendous effect of internal pressure in i~creasing the ccmpress1ve strength of ~ thin walled ~~stl:fened circ~lar cylinder. ~lth no internal pressure, the tailing cQmpressive load as calculated In

.534 + .520 2 = .904 (less than 1.0, thererors no buck.l Ing , )

n.s.

2
+

(.534<1 + 4 X .520')

l./a -

= .08

ANALYSIS

AND DESIGN OF

FLIGHT VEHICLE

STRUCTURES

cs.ar

?rob:3m 1 was :nly 203~O lbs. as ~om~ared to 95000 Lbs -.1i t h .'3 os t internal pressure.
PrC:Jl~m

Solution:
Intera~tiQn

equation

i~, ~~

Rb = 1

11.

Bandin,," -.vi t:l Intern;?l Press:.lre. From Problem la, the allowable load Pa for compresSion plus 5 psi, i~teraal pressure is 9.5000 rcs ,
R~

If the sa~e cylinder as used in the preceding exa~ple problems is subjected to a b~ndlng moment o~ 3,000,000 in. Ibs. when internal pressure is .5 pSi, what is the M.S. use 90 percent pr'obab i Lt ty.

::: P/P a

51000/95000

= .536

S01ution.

For bending Without internal ~ressure, the bending buckling stress Fb was cr ~ounj in Problem 2 to be 1710 pSi. ~ne in:re~se in this benrting buckling stress due to ~nternal pressure is obtained from the curves in Fi~. ~8.l~. The horizontal scale parameter is,
t
I

From Problem 11, the allowable bending moment ~~ with an internal pressur8 is 3,760,000 in. Lbs ,
Rb

= MIMa

1475000/3760000

.092
tha~

Rc + Rb = .536 -+- .392 :: .928 (less will not buckle)

1, thus

2::

P (L\:;J =

10,700,000

(50 ,a .05'

4.67

M.S. = (1/.928) - 1 = .08 Problem 13. Torsion with Internal Pressure

~rom

Fig. C8.14, we read

(f) := .51

rcr gO percent prooabt Lt t y ,

(.51 x 10,700,000 x .05)/50

If the same cylinder is Subjected to an internal pressure of 5 DS1, wha~ pure torsional moment T~ will it carry without buckling.

wnence ,

~FtL:r::

Solution: With no internal pressure, the torsional Shearing buckling stress as cal:ulated in Problem 3 was 876 psi. for 90 percent probabill ty. The increase in shear bucklin~ stress due to the internal pressure Hill be determined by use 0: ~nteraction curve in Fi~. CS.22. To use this curv~, the external hydrostatic pressure (Pext.) to cause buckling ot the cylinder must be deter.ni~ed. This buckling stress FCcr is determined by equation CS.?, which is, k1Jne t:a F'or = 12 (l - 1.Je 'J (L")
The buckll~g coeffi~i3nt use o:(Fi s . ::;::.V;.
Z=h.:.(l rt
/)e ~)-./~
k~

5350 oai . As calculsted in ?ro~lem 11, the internal Qf 5 ~si ,"educes a lor.gitudinal te~si:e stre3S in the cyll~dp,r Nall equal to rt ~ 2500 ?S1. Tr.e allowable bendi~~ ~oment that cy l Inde r C3.n deve Lcp is,
~ressu~e

= ?~(tota1,)I/r -

= F-ctotat

x nr:;Jt

=
!1a..
::a

1710 .... 5350

-+-

2500 = 8.'560 ps 1 .

3560 x n x 50 2 X .05 ;;;: 3760000 in. ibs.

1s

:o~~d

by

!1.S.

(?1s./'?1) - 1 = (3':"'30000/3008000) - 1 = .25

note the ~reme~dous ef~ect of ;ressure o~ the bending strength. ~itt no internal pressure, the f~i1ing bendl~g mo~ent was 570000 as com~ared to 37S0000 with
A~~lrr

= 2140 (See

Pr-cb Iem 1)

inter~al

F:"om Fi 5

C3.l~,

we read k p :: 60
::: 2.58 psi

~si

Inter~Bl ~ressure.

= SOn. ~ x 10 700. COO

I

~obl'Om

12.

:'::mb'!.::J.ed Cmr::;:ression, 3endim:: and

Inte~a1 ?~ess~,e.

12(1_.3 2

T~e

exte"nal

~ressure

to

~roduce

this

sane cyl r ader .vt Ll. be sub j sc t ed tc 8.:1 ~xlal lead P ::a 51000 Ibs., a bendin~ ::lo~ent ~ = l~?~OOO in. Ibs., and an interne1 ~ress~re of .5 ps t . 'J:::8 90 per-cent cr'obab i Lt t y , rih8..t is tte ~.rgin of safety under these :J~d
':':-~e
cJn:i~tl:ns.

c Lrcumrer-ent ta I s tr-es s in tne cy l mder' wall

','mu,l::

be

i~ ~he

:,~

-c-

t/r = 253 x .06/50 = .2:)S

)81

Fig. :3.22 is now used to ~ind ::l8 i~cre~se :orsicn~l shear j~c~ll~g stress due to

---==----'--

C8.22

BUCKLr:'{G STRE:-IGTH OF

MONOCOQUE

CYLI:'{DERS

7:1:5 is less than l.0, 'tnus cyI inde r '."ill not buck Ie under tl:'_~ ::riT~n. ::C;::lbir.,::c Loac i ng,
C8.17 Buckling Strength of Thin-Walled (Monocoque) Conical Shells.
~ultl-stage ~lssile, the ~9per s~ages he.ve smaller di~eters than :he l~Ner stages, thus a conical s~ell pr8vid~s a type of str~cture to permit che.nging the ~lssile jia~eter bet~een the various s~ages. nOr.T.~lly

In a

= 5280 (2n x

SO~

x .OS =

~,lSO,OOO

in. lb.

Re rer-ence ~,J tne tnt erac t t cn cur-ve in F'ig. :3.22 ShO',iS nest of 'the t es t data ra Ll tng be Lo the 2'.1:''19. A ?Q:Jr 99 percent prcbabf Lt tj 'Nc'cll:l give a curve cons i der-ab Iy be Low the
sh:;\'ITI. A correction rae tor of around .8 seem t'J be in order or .8 x 5230 = --.;220 ;JS1.
:::J.T'r~ yo~ld

~s~

G~

The buck11ng stress ca~ Je salculated by ths 1~~eract1cn equation:

As for the case of th~ cylindrical monocoque shell, the theoretical analys~s ~~r th~ b~skllng strength of conical shells by either the small or large deflection theories gives results cons~derably above those given by tests, thus deSign of conical shells at present is based ~r1marl1y on reSUlts obtained by tests. The ~~ter1~1 presenCac in th~s Chapter will thus be limited to ~reser.t~~~ a f9W desi~n buckling c~rV3S.
ca.18 Allowable Compressive Buckling Stress for Thin-Walled Conical Shells.

~/

= 5.0/.225

= 12.4 (use

~1nus

Sign)
~Si.

.:..L 373 a

"

=1

19.4, 0!" f ST -~

= 3940

'I'!:",,'lS to cause t or-s Lona L ':luckllng With 0

rig. CS.2~ s~~ws desi 5n curve~ fer t~e compressive tucklln~ stress o~ a csnleal shell as derived from a statistical study of test data by Hausrath and Dittoe (Ref. 16). ~he expr-ess i on ~or the buck Ling c ouor-ess ive St::'-8SS along ~ ge~erator at ~:d-~elght cf cone ts,

;JSi

t nter-na L

:s':'c~-o..r

pr-easur-e requ Lr-e s a torsional stress 0-;: ~ ps t .

The~ ths torsian~l :n')men~ developed 3G40 (2n x 50 2 X .OS) ~ 3,100,000 ir..lb.

=
wn~re,

Per 2 n p teos'" P::r is the tcac on cone .

tu~kllng

axial

canpressl~e

?r)clem 14.

Combined Ccm~ress10~, ae~ding, 7Jrs1on an~ Internal ?ress~rs.

3 = rad:us af ne t zht .
~

curvat~re 'Jf

cone at mld-

-o-.ot nec Lcac i ng

; =

~OOJOOO

sr.:::

J:

=~C""50,

M =

1,::~C,COO. ~s1.

= cone wall

th1c~ness.

~n:~r~~l

Jressure 0: 5

Data

:'~CJ: ::'~'J

:25t3 OJ' vs z-Lcus 1:'.'18.'3':i-

..;:.

3,--':O,COC

caeors wer-e 5t<;.::5:1::'O.l'..:.; ~';2.l,:,~.:~-: "or- expec t ea :::~::, ;0 per-cent So!:::: 08 pe r-cent ;::':':3.:1111::,.' s t.r-enzt.a levels. :Jls:'2:-si:::n ::' ca t a Na~ rcun; to be slL:::-:tlY le~s than ':::::.to:' nonocoque ::ylin:ers, N0~-11~e~r e~~e:ts )~ ~dl~3 :: t ru cknes s r-at tc c r s t r-eng'th vie t er-tc ra t l cn ',.,,1:1"', ::'~;l.;;t',-: ~J rae Las rar rc '::9r9 ::0: c i s t sr-n b l e .

, "

.~'--::,JCO!3,lC~,JOC

= .22":

~e!':::2!".:~:' :::e:::-:~i:-::
:>~:'~~~:: ~:r.:'~_-:::?~C2j

'::23:

--:;::'_.i.O:.':::'~.vit
::'.~

.:

C:-'.:;:

:;'_"!..':

.s ,

~::".::::'..8~.-~9

is

_~:":''':::'~;;:::'_T2

:;":~2:'.C:::~S

?:':~"':-:~~
',~r!

:-

-::>:~~'Oc ::~'"

c : ':.ll :-:::D.?':; _":-~".:~ "-.:,," :!.::'.'~:" c r-ve . :';:'2

.~31

.... -+-.;-; ...

ANALYSTS

AND

DESIGN

OF

FLIGHT- VEHICLE STRUCTURES

3uc~ling

CB.23

5.0 " " " - - - - - - - - - - - - - ,

1.0 :-

Under Combined LQsQing

0.5 ~
0.1 c50 100

MEAN

from Boeing R~port (02-3617) th~ er~~c~s of si~ultaneous co~press1on. cer.ding, shear ~4d overpressure can be considered ~sing the interaction eq~ation:(R

"...,,/<
T

BLEVEL ALEVEL

c + Rb + Rv)
R~

1 2

Rq 1

1I

::

1 Rv

where.

= fclF c,

Rb

= fb/Fb,

= 1.4

~ st

'0..

500 1,000 5,000 10,000 P

ca.20 Example Problem.

Fig. ca.24 (Ref. 16)

~::e !fA" and "a" curves are recommended for ;ractical je 19r.. The "An level curve 1s racJ~ended or use for those structures where ~ si~gle rai ~re NQuld result in catastrophiC loss or lnj~ry to personnel. T~e stress level at :~e small en1 of the cone should be checked to preclude the poSSibility of an early ~al1ure precipitated by inelastic stresses.

A cenical shell w1th 'NaIl thickness t = .05 and other dimensions shown in Fig. (a) is fabricated from aluminum alloy (S = 10,700,000). Determine the ~ollowlng 20 values :-

(1) Allowable comprese ive load Pc.(2)

Allowable bending moment t"'..-::r.

1.

..

" TlJ
1- I "-1
0

I -20" I _I 13 52 -1 . 1 Fig. a.

~-l

ca. 19 Additional Design Buckling Curves for ThinWalled Conical Shells.

(3) Allowable torsional moment Ter . (4) Sxternal buckling ?ressure Ger'

Figs. C8.25, 26, 27 gives c~rles for ~ete~lni~g the allchable b~ckling stress for ~hln-walled conical shells in compression, :e~dl~6 and torsion resp~ctlvely. These curves are reproduced by pe~lss1on of th9 Boeing ~cmpany fr:m their report D2-3617. ~ig. Ca.28 ~ives a deSign ~~rve ~or the allowable external buckling pressure for 'ttn n-waLl.ed cJnical Shells, also ~rom sa~e Boeing Re~ort. 7hese curves were used ~or nlnute~~n interstage an~ skirt jesign ~~d the so~ical s~ells had ring s~if~eners 1eslgn8G by nethod ~iven in

Solation:
p

L = 20/c08 10 0 :: 20.32 In. Lip = 20.32/10.16 = 2.0

To f1nd Pc:-

= 10/cos 100 = 10.16 In. pit = 10.16/.05 = 203

From Fig. C8.25 with pit = 203 and L'o = 2.0. we read Fc x 10 3 / E equals 1.20, ~henee F'e 1.20 x 10,700,000/1000 12,520;lS 1 . Then p~ F' c 2 'IT. r-t 12820 x 2n x 10 x .03 ~O,300 Ibs.

= =
L

= =

so~pressive Juckllng stress (psi) allowable bending buckling stress (~sf) allowable torsional Jucklin~ s~ress (~s~; coc;ressive stress bendir:g stress tr~~sv8rse s~ear stress sl~~t je:~ht of conical sjell 'Nhll ttu cknesa one ~al~ ~one apex angle P]iSSsn's r"?.':io ::lin~:r..r;n r-actus of cur-vet n-e averag-e -ac tus cr cur'va t ut-e ex:ernal ~reSSU~8 (~si) allowable external pressur~ allc\~ble

Fr-om F'1g.2S.26 .vt t a oft = 203 and Lzo = 2.2. we read :='J x l03/E = lAc. vnence f'J = 1.~S x le.-CO,eCa/IOCO = 15820 ~si. Then
i":i~r ~

= 15220

?c ~ :rOlt
~

x lOll X .OS = 24e,5Ca in. lb.

2.J

T8 T

.51 x :O,-JO,OOO/lOOO = 6520 pst. Then :: 1 t 2 'IT. rOlt s = e520 x 2n x 10 2 x .05 = 200.000 in.

Fr-om ri;,.::S.2'7, 'N1th 8/t :: 203 and :"",'0 -= we read 1 s t x l03/E = .61. "Jnenc8. r st
l~.

,0 9.0 8.0 '.0 8.0 5.0 4.0

tn ci

i'
,

4.0 3.0

I
!
I

3.0

'.

t ,
'! 2.0:
,

" 'r"z

\
,

I
F

, , ,
i ,

!
!
I

2.0

, X IO

1. 5 :

!.
,

1.0

i ,

: ~/R -4
I

; , I

,
,

1.5

.90 '
.80 :
.70

-.!!x lOll
E

o. p

.60,
.50 i
,-40,

1.0 .90 .80 . '0 .60 .50 .40 .30 .25

ii::

o c:

o o

,
\1:
.1

.30

"' o
rZ

.20

.20

!ii

' i.
.15

, ,

.'I,

. 15

" '"
til

'"
>-<l

"'

a, 6"

II

'r

1000

Lip

pi'

"

_. ,----,.---_.-_.-

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C8.25

g
-

, ,- -

- -

'.

-. -J

.-.-.

:..:'~ c--"'-c-.c:c~='~~~'----,-_~~_"",--,-

__

-.-

__'U>""~a.-,",_=,--~

'-~-:--,-

...

,_.

10.::'========
~-.'--'-

..

.,~-_.--=.,:;o_

'-

;-~.

;;.;::

-.-.-----..:.-------..::.----T-htD-WaHcd efni:cii;l:Sbi::U";
--'--~:'--~--------_.-

. Bui:klinf~~s'friri:::~;:

f~~_!~

(--~~;~~7~,7""h

::: ::..:-.:::::,:
.... _--;--~---c

- - - - - - - - - - - - - - - - -------'---'-;'

. ::.-i

".----------.;=::::::='::::.t=:.=
1. IY-

LO

.,

" '10

----------

,--;--;--;--:-7'.1000

C8.26

BUCKLING STRENGTH OF MONaCO

UE CYLINDERS

To find external bUCkling pressure qcr:Use 15 made of Fig. C8.28.

Pavg

(4)

;: 10

13.52

2 cos 10

11.92 In.

Pavg/t = 11.92/.05 = 238

A conical shell has a small end diameter of 30 inches, a length of 30 inches, a semi-vertex angle of 8 degrees and a wall thickness of .057. Material is aluminum alloy with E = 10,500,000. Determine the ultimate compressive load, the ultimate bending moment, and the ui t icaee torsional moment the cone will sustain when each load 15 acting separately. Deter.m1ne the external pressure that will buckle the cone.
REFERENCES

The lower scale parameter of Fig.

C8.28

Is,

11~~2~.05 .; 1

- .3"

660

(1)

Ky

From FIg. C8.28 for Z = 660, we read The equation for buckling pressure from Fig. C8.28 Is,

HarriS, Seurer, Skeene and Benjamin. The StabUity of Thin-Walled Unstiffened Circular Cylinders Under Axial Compression. Jour. Aero. SCiences. Vol. 24. August,
19~7.

=27.5.

Ky Etoln:l

L" 12 (1 - lis")

27.5 x 10,700,000 x .05$ x n:l

11.32 x 20.32 x 12 (1 _ .3 ')

=~

pSl

PROBLEMS

(1)

A monocoque circular cylinder is fabricated from aluminum allay fbr which E is lO,500~OOO. Wall thIckness Is .064. Cylinder diameter is 80 inches. Length is 60 inches. Using a probability factor of 90 percent, determine the fallowing values:(a) (b) (c) Buckling compressive load for cylInder. BUCkling bending moment tor cylinder. Buckffng torsional moment for cylInder.

(2)

The cy1lnder ln Problsm (1) Is subjected to the following deSign loads. Will the cylInder fail under these various deSign conditions:Condition 1 P = 30000 lbs. compression M = 50000 in. Ibs. Condition 2 P = 25000 lbs. compression M ~ 45000 in. lbs. r = 50000 in. lbs.

(3)

the cyf tnder- in Problem 1 is pressurized to 6 'PS!. what ulttmate compressive load will it carry. What Ultimate bending moment will it carry. What Ultimate torsional moment will it carry .
If

(2) Donnell, L. H. Stability of ThinWalled Tubes Under Torsion. NACA Report 479. (3) Batdorf, s. B. A SimplUied Method of ElaBttc Stability Analysis for Thin Cylindrical Shells. NACA Report 874 (1947), (4) Weingarten, Morgan and Siede. Final Report of DeSign Criteria for Elastic Stabillty of Thin Shell Structures. Contract AF04(647)619 Air Force Ballistic Missile DiVision). Space Tech. Labs. Dec., 1960. (5) La. Crate and 8ehwartz. Buckling of Thin Walled Cylinders Under Axial Compression and Internal Preasure. NACA T. N. 2021, Jan. 1950. (6) VonKarman and Taien. The Buckling of Thin Cylindrical Shells Under Axial Compression. Jour. of Aero. Sciences. Vol. 8, June 1941. (7) Taein. A Theory for Buckling of Thin Shells. Jour. of Aero. Sciences. Vol. 9, August i942. (8) Fung and Sechler. Buckling of Thin Walled Circular Cylinders Under Axial Compression and Internal Pressure. Jour. of Aero. Sciences. Vol. 24. May 1957. (9) Harris. Sew-er. Skeene and Benjamin. The Bending Stability of Thin Walled UnatWened Circular CyUnders Including Effects of Internal PreSl5Ure. Jour. of Aero. SCiences. Vol. 25, May 1958. (10) Slede and Weingarten. On the Bending of Circular CyUndrical Shells Under .Pure Bending. Jour. of Applied Mechanics. Vol. 28, March 1961. (11) F1ugge. Dte Slabititat der Kruszyllnderscbole. Ingenieur Archiv. Vol. 3. 1932. (12) Gerard &t Becker. Handbook of Structural Stability. Part In. Buckling of Curved Plates and Shells. NACA T. N. 3783. Aug. 1957. (13) Hopkins and Brown. The Effect of Internal Pressure on the Initial Buckling of Thin Walled Circular Cylinders Under Torsion. R &t M. No. 2423. British A. R. C. 1946. (14) Crate, Batdorf and Bub. The Effect of Internal Pressure on the Buckling Stress of Thin Walled Circular Cylinders Under Torsion. NACA WRL67. 1940. (is) Batdorf, Stein and SchUdcrout. Critical Stress of Thin Walled Cylinders in Torsion. NACA T. N. 1344 (1947). (16) Hausrath & Dtttoe. Development of Destgn Strength Levels for the Elastic Stability of Monocoque Cones Under Axial Compression. NASA T.N. D-1510 (1962). (17) Hess and Garber. Stability of Ring Stiffened Conical Shells Under SLmultaneous Lateral Pressure and Axial Compression. General Electric Report No. R58SD226. April 19:8.

I
f

~.

CHAPTER C9

BUCKLING STRENGTH OF CURVED SHEET PANELS. ULTIMATE STRENGTH OF STIFFENED CURVED SHEET STRUCTURES. PART 1. BUCKLING STRENGTH OF CURVED SHEET PANELS
e9.1 Introduction.

Curved sheet panels represent a common external part of flight vehicle structures. Ex~ples are the skin of the fuselage and the missile. If the curved sheet has no longitUdinal stiffeners, failure will occur when buckling occurs. If the curved Sheet has stiffening elements attached then the composite unit will not fail on the development of sheet buckling, that Is, the composite unit has an ultimate strength much greater than the load which caused initial buckling ot the curved sheet panels between the stiffening units commonly referred to as sheet stiffeners. In some flight vehicle deSigns i t may be specified that under a certain percentage of the limit loads that no buckling of the curved sheet panelS shall occur. Thus it is necessary to be able to determine what stresses will cause curved sheet panels to buckle and also to determine what external loads will cause a stiffened sheet panel to fail.
C9.2 State of the Theory.

Fig. C9.l gives curves for determining the buckling coefficient Kc. The theoretical derived curve is shown along with the recommen~ed design curves which were dictated by test results. These curves apply tor buckling stresses in the elastic range.
C9.4 Shear Buckling Stress of Curved Sheet Panels.

The equation for the buckling shear stress is,

- C9.2

Figs. C9.2 to C9.S gives curves for finding the buckling stress coefficient Ks to use in Equation C9.2. These curves apply when buckling stresses are in the elastic ~ange.
e9.5 Buckll..cg Strength of CUrved Sheet Panels Under Combined Axial Compreeston and Shear.

In Chapter Ala, the small deflection linear theory was used to determine the compressive bUCkling stress of flat sheet panels. The theoretical results compare favorably with expertmental test results. However, when the same theory is aoplied to unstiffened curled sheet panels or thin-walled cylinders, the theoretical results are considerably above the test results on such units. A large deflection theory gives closer comparison to test reSUlts, but this theory involves a consideration of initial imperfections in the sheet and thus an un~nawn quantity. For the application of both small and large deflection theories to the buckling of curved sheet and cylinders, the reader is referred to the references listed at the end of Chapter C9, oar-t icurar-ry those i-epor-cs which gi ve the results of such investigators as Do~~ell, Batdorf and Gerard.
C9.3 Compressive Buclrling Stress of Curved Sheet Panels.

The studies by Schildcrout and Stein (Ref. 1) gave the follOWing interaction equation for combined longitudinal compreSSion and shear on curved sheet panels:- - - - - - C9.3

where,

RL = -F-

c ccr

If the longitudinal stress is tension instead at compression, then RL is conSidered as a negative compression uSing the compression allowable. The equation fo~ margin of safety 1s:- - - - - - C9."
e9.6 Compressive Buckling Stress of Curved Panels with Inter-nal Pressure.

l~~der

The expression for the buckling stress axial compression is of the same form as for flat sheets, the 'mlue of the buckling coefficient Kc having a higher value.
- - - - - - - C9.1

As for the case of monocoque cylinders, internal outward pressure increases the axial buckling compressive stress of the curved sheet panel. Rafel and Sandlin (Ref. 3) and Ratel (Ref. 4) performed tests on curved panelS under axial compreSSion and internal outward pressure. The results correlate with the interaction equation:- - - - - - - - - - - - - - Cg.S
e9. 1

10'0
0

7
6 0

10" 0
B

7
0 0

Kc

10 0
B

e
0

a
2

10

10'

10'

10'

..,.

,,'

.,

>

'" z
c

" " '"

o ..,

'n" " " " ''"

>;

>; '"

'"

Fig. C9.2 Shear Buckling Coefficients for Long Clamped Curved Plates

-I

'I,

'-(~

"',i:~l

.')

Fig. C9.3 Shear Buckling Coefficients for Wide Clamped Curved Plates.

ANALYStS AND DEStGN OF FLIGHT VEHICLE STRUCTURES

e9.S

JO'r--------,------~----, Fig. C9.4 Shear Buckling Coefficient for Long Simply Supported Curved Plates.

Ksn E ta F Scr = 12 (1 - Veal

(tJ

Zb

= ~;(1 - Ve ") ,/,

(From Ref. 2)

40

30

'0

zo
10

of

~~

.............

I I

I '
10

Dl
1
I

I ,-,,

I
,,'

f
I

10'

Zb

/0'

Fig. e9.5 Shear Buckling Coefficient for Wide Simply Supported Curved Plates.

KsnaE

t,

FS cr = 12 (1 - lie") (Ii)

Zb=~(l-lIe")'/'
::.

(From Ref. 2)

Ks

30

ZO
10

to :-;
cnJNDER
/5 2.0

'-../
I
2

II

!,I,I,
10

I" ,

, I r~-~jl I

-, !

:;.--

!
io'

, 1.1.1,1 Id

Zb

~t'2-'

C9.S

BUCKLING STRENGTH OF CURVED SHEET PANELS. ULTIMATE STRENGTH OF STIFFENED CURVED SHEET STRUCTURES.
Z
1)

where Rc = tclF ccr Rp = ratio ot applied internal pressure external inward pressure that would buckle the cylinder ot which the curved panel is a section. Found by use ot Fig. C8.16 In Chapter C8.

2j (3)

In using Equation C9.5, the value ot Ro is negative as inward pressure is opPosite to the inward acting outward pressure.
C9.7 Shear Buckling stress of Curved Sheet Panels with Internal Pressure.

,~>J'''$-k
~

:!; l:

I1
19.30

(4)

Panel A Panel
5)

I
I
I
Z
Fig. e9.6

2"'r

As in the case at monocoque cylinders, internal pressure increases the shear buckling stress ot the curved sheet panel. Brown and Hopkins (Ret. 5) solved the classical equilibrium equations to determine the ettect at radially outward pressure upon the shear buckling stress at curved panels and obtained tair agreement with test data by Ratel and sandlin (Ret. 3). The test data also correlates With the interaction curve used tor the etfect at internal pressure upon cylinders in torsion (see Chapter C8). The interaction equation is; RS.:ll + Rp = 1 where Rs Rp
~

The problem Is to determine whether skin panels (A) and (B) will bUCkle under the given combined loading on the fuselage section. Solution. To ~lnd the bending stresses, the moment ot inertia of the cross section about axis y-y is necessary, which axis Is also the neutral axis Since all material is eftective. The moment ct Inertia wlll equal 4 times that due to ~tertal in one quadrant.
I y due to stringers is,

- - - - - - - - C9.6

ts/F Scr

= ratio

at

applied internal pressure external inward pressure that would buckle the cylinder at which the curved sheet panel is a section. Found by use af Fig. C8.l6 of Chapter C8.

I y == 4 [.075x20.:ll .... 15 (19.30.:ll"'17.34 1l1"'14.l4 1l1

... 10"+5.l8.:l1)] = 720 in." In using Equation Cg.s, Rp is given a minus sign.
C9. 8 Example Problems.

Due to skin:Iy

=n

rJt

= nx20 J x.04 = 100S

In. ".

PROBLEM 1.

Total I y = 720 ... 100S = 1725 in .", Consider Skin Panel (A): r/t = 201.04 = 500,
alb

Fig. Cg.s illustrates a circular fuselage section with longitUdinal stringers represented by the small circles. The area ot each stringer is .15 sq. in. The skin thickness is .04 inches. All material is aluminum alloy with Ec ~ 10,700,000. The fuselage frame spacing (a) is 15.75 inches. The ruaeLage section is SUbjected to the following load system:My
~

= 15.75/5.25

= 3.0

To deternune the compressive buckling stress, use will be made of Fig. C9.l.ta tind the buckling coefficient Kc.

z =.2:(1_V .:lI)l/.:lI rt e
5 25" 1 __ 20 ~ .04 (1 - .3")'" 32.9

600,000 in. lb. (caUSing compression on

t-op halt)

V z ~ 5175 lbs. (acting up) T = Torsional moment = 210,000 in.lbs. (acting counterclockN!se)

From Fig. C9.l,

Kc = 14

=="

------~--

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

e9.7

ICc nil E ....t... a F Per- = 12 (1 - Ve ') ( b 1

[S(total) = 269
2359/11200

+ 2090

= 2359 psi. =
cr

= 14 nil x 10.700.000 ( .04)11 = 78S0 psi

12 (1 - .3') 5.25

Then shear stress ratio Rs = fslF's

.21.

To find the shear buckling stress, we use Fig. C9.2. Z is the same as calculated above or 32.9. ThUS from Fig. C9.2 we read for alb = 3, that K s = 20.
_ 20 n.ll x 10,700,000 04 FScr 12 (1 _ .3') (525 10 = 11,200 psi

The interaction equation for combined compression and shear is, RC+RaIl=l
.874 + .21 11 = .918. This is less than 1.0 so panel will not buckle.
M.S.

The bending stress at midpoint of Panel (A) will be calculated:

.874 + V.874a + 4 x .2111 - 1 09

t c = l1Z/I y = (600,000 x 19.71/1725 = 6850 psi

Thus stress ratio Rc = fclF'ccr = 6850/7850
a

Consider Skin Panel (8).

Arm

Z to midpoint at panel = 15.88 in.

.874

tc = Mz/I

= 600,000 x

15.88/1725

5520 psi

Shear stress on Panel (A) due to torsion is,

fS

Rc

=5520/7850 = .704

= T/2At, where A Is inclosed area ot

ruse jage cei.; ,

all panels or f s

The torsional shear stress Is the same on = 2090 psi as previously calculated. Shear tlow q due to transverse shear load:
q
s

fs

210,000/2 x n x 20 11 x .04

= 2090 psi.

The panel Is also subjected to shear stress due to transverse shear at Vz = 5175 1bs

fZZA = 3l:ZA.
Calculation at ZZA at upper edge at panel:

The shear flow equation is,

For stringers = .075 x 20 + .15 (19.3 + 17.34)

= 7.0
The shear flow will be zero on Z axis. The shear flow at top edge of Panel (A) will be dUe to effect of one-halt the area ot stringer (1).
q._o

For skin:

Area = 2 x S.25 x .04 = .42.

=3

x .075 x 20

= 4.5

=r

Vertical distance Z to centroid of Ski~ portion sin a/a. a = 30 0 The result is Z = 19.1
In. The

lb./in.

ZA =

Area at skin between stringers (1) and (2)

Is 5.25 x .04

= .21 =A.

8.03 + 7.0 15.03. = 45.09 Ib./ln.

19.1 x .42

Then q = 3 ZZA

8.03.

Total l:ZA

=3

x 15.03

Distance from centroid of rallel (A) fram neutral axis is Z = r sin a/a. a = ISO. which gives Z = 19.8 In.
Thus qa_:l. = 4.S+3x .21,;{19.8 = 17 Ib./in.

A s~ilar calCUlation tor shear flow at lower edge at Panel (B) would give q equal 55.0. Thus average shear flow on panel Is (S5 +
45.09)/2

50.04.

1251 psi. The total shear stress f s on panel then equals 1251 + 2090 = 3341 pet ,
RS

Then t s

= q/t =

50.14/.04

Then average shear flow on panel Is

(17 + 4.5)/2 = 10.75.

= ts/F s cr
= .704

= 3341/11200

= .299

Then shear stress = qlt 269 psi.

= 10.75/.04 =

Rail

Substituting in interactlon equation He + + .299- = .793. The result Is less than 1.0, thus panel will not buckle. The student should check other panels for bucklIng and compare their margins at satety.

This shear stress has the same sense or direction as the torsional shear stress so we add the two to obtain the, true shear or:

C9.8

BUCKLING STRENGTH OF CURVED SHEET PANELS. ULTIMATE STRENGTH OF STIFFENED CURVED SHEET STRUCTURES.
~-Rp=l

General Comment:In general the compressive stress is the do~inant factor in causing the panel buckling. Thus to increase the buckling stress of the panels and also to give a more effective stringer arrangement to carry the bending moment, the stringers should be spaced closer in the top and bottom regions of the crosssection and with increased spacing as the neutral axis is approached.

rca

ccr

f,' 7850a - 2.06 :: 1, or

t c = 13750 pat

Thus the internal outward pressure .pt 5 psi increases the axial compressive buckling stress from 7850 to 13750 psi. Shear Buckling Stress Under 5 pSi Internal Radial Pressure. From Art. G9.7, the interaction equation is, RSiJ+Rp::l From Problem 1, Fscr or our panel was 11200 psi. The value of Rp is determined as above or Rp = 2.06. SUbt. in interaction equation:
~

PHOBLEM 2.
The fuselage section in Problem 1 is subjected to an internal outward pressure of 5 PSi. What would be the compressive buckling stress of a panel and also the shear buckling stress with this internal pressure existing. Solution. From Art. C9.6, the interaction equation is, RCa+Rp=l, From Problem 1, the compressive buckling stress FCcT = 7850 psi. To evaluate Rp' the external inward. radial acting pressure that would cause buckling at a circular cylinder having a radius equal to that ot the curved sheet panel must be determined. Use is made of Fig. G8.1S at Chapter C8 to determine the backing stress under such a loading. The lower scale parameter for Fig. C8.16 is,

scr

- 2.06 ::. 1

The internal pressure of 5 psi thus increases the shear buckling stress from 11200 to 19500 psi.

PART 2. ULTIMATE STRENGTH OF STIFFENED CYLINDRICAL STRUCTURES.

cs. 9
Introduction.

15 .75' v/'-:1:-----::3'"' -. .::: 296 = 20 x .04

From Fig. G8.16, we read K y = 19

=
=

Ky n
a

12 ( l - l I . ) (Tl

iJ

19 n x 10.700)000 (---!...9!...)a 12 (1 - .3' 15.75

= 1215 psi.

The external radial pressure to prOduce this buckling stress is,

p
Fc

or tlr

= 1215

x .04/20 = 2.43 psi

A cylindrical structure composed at a thin skin covering and stiffened by longitudinal stringers and transverse frames or rings is a common type ot structure tor airplane fuselages, missiles and various types of space vehicles, and such structures are often referred to as the semi~onocoque type of structure. The design at a semi-monccoque structure involves the solu~lon ot two major problems, namely, the stress distribution in the structure under various external loadings and the check of the structure to see if these stresses can be safely and efficiently carried by individual components of the structure as well as the structure acting as a whole.
C9.10 Types of Instability Failure of Semi-Monocoque Structures.

Rp

= ::~ = 2.06

Subt. in interaction equation with a minus Sign for Rp,

(1)

Skin Instability.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

e9.9

In general, thin curved sheet panels buckle under relatively low compressive stress and thus it design requirements specified no buckling of the sheet under lunit or design loads, the sheet would have to be relatively thick or the stringers placed very close together and the fuselage or body structure would be unsatisfactory from a strength weight standpoint. In missile structures, internal pressurization increases the buckling stress greatly, thus the buckling weakness of thin sheet is improved, but keeping a structure pressurized under all operating conditions has its difficulties. In a semi-monocoque body. the longit~dinal stringers provide efficient resistance to compressive stresses and buckled sheet panels can transfer shear loads by diagonal tension field action, thus the bUCkling of the sheet panels is not an impor-tant factor in l1miting the ultimate strength of the over-all struc~ural unit. ~nen buckling of the skin panels takes pf.ace , a stress redistribution over the entire structure takes place, thus it is important to know when skin buckling begins. Furthermore, design requirements may often specify that no skin buckling should take place underv a certain percent of the limit or deSign loads. The equations and design curves in Part 1 of this chapter can be used to determine the buckling stress Of curved sheet panels under various stress systems.

panels transfer fUrther shear forces by semidiagonal tension field action which produces additional axial loads in stringers and also bending which must be considered in arriving at the panel failing strength. This subject is treated in detail in Part 2 of Chapter CII. (3) General Instability.

In general instability, failure is not confined to the region between two adjacent frames or rings but may extend over a distance of several frame spacings as illustrated in Fig. C9.7b for a stiffened cylindrical shell in bending. In panel instability, the transverse stiffeners provided by the frames on rings is sufficient to enforce nodes in the stringers at the frame support points as illustrated in Fig. C9.7a. Any additional stiffeners in excess of this amount does not contribute to additional buckling strength. General instabllity may thus occur when the stiffeners of the supporting frames is less than this minimum value.
C9. 11 The Determination ot the Stresses in a Stifiened Cylindrical Structure Under
External Loads.

The stresses in a stiffened cylindrical structure such as used in typical fuselage or missile deSign can be fairly accuratelY determined by the modified beam theory as presented in Chapter A20. A more rigorous approach (2) Panel Instability. is given in Chapter AS involVing matrix formulation but this approach requires the use of a The internal rings or frames in a semilarge electronic computor to handle the reqUired monocoque str~cture such as a fuselage divide calculations. For details at applying the the longitudinal stringers and their attached modified beam theory, the reader should refer skin into lengthS called panels. If these to Articles AZO.3 and A20.4 of Chapter A20. In the example problem solution as given in frames are SUfficiently rigid, a semi-monocoque structure i f suo.ject ed .co bending will fall on Article AZO.4, the effective area to use tor the compression side as illustrated in Fig. the curved sheet was based on the ratio of the C9.7a. The stringers act as columns with an buckling stress of the curved panel to the bending compressive stress on the panel due to effective length equal to the panel length which is the ring spacing. Initial failure thus bending of the entire effective cross-section occurs in a Single panel and thus is referred at the fuselage under the deSign loadS. In the to as a panel i~stability failure. In general, example problem as given in Table A20.2, a this type of failure occurs in most practical conservative buckling compressive stress equal to .3 Et/r was used for the curved panel and no aircraft and aerospace semi-monocoque structures because the rings are suffiCiently consideration of the effect of shear stress on stiff to promote this type of failure. Since the compressive buckling stress was considered. the inside of a fuselage carries various loadS, A more accurate procedure would be to calsuch as passengers, cargo, etc., the rings must culate the effective area at the curved panels act as structural units to transfer such loads taking into account the influence or combined to the shell skin, thus requiring rings of compression and shear on the buckling strength considerable strength and stiffness. Even of the panel. Thus in Table A20.2 on page AZO.5 lightly loaded frames must be several inches deep to provide condUits reqUired in various of Cha,ter AZO, the shear stress on each curved panel should also be calculated and then the installations to pass through the web of the ring cross-section, thus prOViding a relatively buckling stress of the panel under the combined compreSSion and shear calCUlated. stiff ring for supporting the stiffeners in their col~ action. ~hen the skin buckles under shear and ccmpressive stresses, the skin The buckling stress under pure compression

ANALYSIS AND DESIGN OF FLI

T VE

ICLE

TRUCTURES

C9.1l

and shear should be calculated uSing Equations C9.l and C9.2. The buckling stress under combined compression and shear is given by the interaction equation:-

C9.13 Calcula.tion of General InstabUtty.

~ great deal of theoretical and experimental work has been done on the SUbject of general instability ot stIffened shells. The general goal in the design of such structures is to insure the trames have sutt1cient stiffness so as to prevent the type at failure illustrated in Fig. Cg.7b or, in other words, to insure the type of failure illustrated in Fig. C9.7a which is panel instabilIty.

Re

fc/Fccr" Rs fsh scr' The expression for margin of safety is,

a a

Rs "

1, where R e

11.3.

R e

2
+ y'R c a .,. 4Rs a - I

Let t c oe the compressive stress that will buckle the curved sheet panel when subjected to combined compression ~nd shear when the ratio at the applied compressive and shear stresses in a constant. Then,
2

Shauley (Ref. 6) has derived an expreSsion for the reqUired tTame stiffness to prevent general instability failure of a stiffened shell in pure bending.
(EI)f

= Cr

!1D"/L

Rca.,. 4RS a)

- C9.6

In a study of available test data. Cr was

found to ba 1/16000. (El) r Thus,
----C9.7

These t c values should then replace the values in column 5 at Table AZO.2. The author has noted that one aerospace company in their missile design uses only 90 percent ot the theoretical buckling stress in computing the effective area of the bUCkled curved panels. This correction assumes that the curved sheet fails to hold the buckling stress as the fuselage section as a whole Is further loaded and the curved sheet surrers more buckling distortion.
C9. 12 Panel InstabUity Strength.
I Panel instability means failure of the stringer and its effective skin between two adjacent frames. The bending of the stiffened shell as a whole prOduces a compressive load or stress on the stringer. The semi-tension field action of the skin atter bUCklIng produces an additional compreSsive load on the stringer and also a bending moment.

!1D '/16oo0L

where, E : modulus of elasticity I = moment of inertia ot trame section o ~ diameter ot stittened shell
L "" frame spec iog

M "" bending moment on shell

Becker (Ref. 7) in a comprehensive study ot must published theoretical and expertnental material relative to the general instability at stiffened shells, summarizes the results of his stUdies as given in Table C9.l. Bending. For the case at bendfng , the constant or
4.80 in the equation gIven in Table eg.l is for

The compressive stress due to bendIng or the stiffened shell as a whole is tound by the methods discussed in Ar~icle C9.ll. The addItIonal stringer loadS due to semI-tension field action are determined by the theory and procedure given in Part 2 ot Chapter CIl. These calculated stringer loads are then ~ompared to the stringer strength to determine whether a pOSitive margin ot safety exists. The local crippling and column strength of a strInger plus its effective skin can be found by the theory and analysis methods given in Chapter C7. The bending strength at the stringer cross-section can be found by the theory and analysis method given in Chapter C3. The strength at the strtnger in combined compression and bending is tound by use of the proper interaction equation.

the condition where the frames are attached to the skin between the stringers. For frames not attached to the skin between stringers, the constant should be 3.25. The eftectIve sheet width tor use with the str tnger-s may be r ound tram the aqua t i on,
C9.S

where, We = ettective width of skIn per s tde or

str Inger (In.)

stringer SpaCing (In. ) FCcr = compr-eaatve bucklIng stress for curved skin panel Fe compressive stress at bending general instabilIty (PSi)
b

C9.12

BUCKLING STRENGTH OF CURVED SHEET PANELS. ULTIMATE STRENGTH OF STIFFENED CURVED SHEET STRUCTURES. Table C9. 1 (Ref. 7) THEORETICAL GENERA.L INSTABILITY STRESSES OF ORTHOTFOPIC CIRCULAR CYLINDERS

I I
Loading Bending

(Results are based on the assumption that spacings of longitudinal stiffeners and circumferential frames are uniform and small enough to permit assumption that cylinder acts as orthctroptc shell) Moderate-length cylinders F c ::gE (Irt)J./a/Rts Long cylinders

g =4.80[(b/d)(Os/OI)('s/tf)"(Ps/b)"j

-I, CPf/R)"

I , I
I
f

External radial or hydrostatic pressure

t 0 F = 5 51E (...!)'C-4l- I"C R) Y tf Ri:

F y = 3E

I
I

Torsion

FsT = 3' 46

C~)'I' (2!,J,I'(-!!l-I" Rli L t

FT' I . 754 s

('sj-I'(..!!.,J'I' t R2

Fe
Fy

:: Compressive stress at bending general instability (psi) :: Circumferential normal stress under external pressure at general instability (pSi)

FsT:: Shearing stress at torsional general instability (pSi) b d

R

:: Stringer spacing (in. ) :: Frame spacing (in. ) :: Cylinder radius {in.} :: Skin thickness (in. ) :: Stringer area (in. 2) :: Frame section area (in. 2) :: Distributed stringer area :: Distributed frame a rea
>

A, Af
ts
tl

As/b

Af/d

if II
0,

:: Bending moment of inertia of frame section (in. ~) :: Distributed bending moment of inertia of rrame :: Str-tnger- section radius of gyration (in.) :: Frame section radius of gyration (in. ) :: Length of cylinder (in.) :: Modulus of Elasticity.
e

ff/d

Pf

L
E

---~.......,.g"'~i41-'~.J.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C9.13

For the frames the effective skin width should be taken as the total trame spacing (d). For Inelastic stresses, the use of the secant modules appears to be applicable on the basis ot limited test data. External Radial or Hydrostatic Pressure. The effective skin width to be used in computing the stringer and tr.ame section properties may be determined tram the following equation.

where

= applied moment

Mo = moment caua tng bending general instability acting alone T = applied torsional moment To = torSional mament causing torsional general instability acting alone General Instability in Combined Transverse Shear and Bending. (Ref. 7) concludes there is no interaction for this combination of transverse shear and bending loads. General instability occurs only tor either type ot loading acting alone and thus both loadings may be examined separately.

The subscript s refers to stringer and f refers to trame. The term d is the rrame spacing. Torsion. The effective skin to be used can be determined fram the following equation:-

e9. 14

Buckling of Spherical Plates Under Uniform

External Pressure.

Classical Theory USing 0.3 for POisson's ratio gives the following buckling stress for a perfect spherical shell subjected to a uniform external pressure:Fer = 0.606 Et/r - - - - - - - - - - - - C9.13

where Fscr is the shear buckling stress tor the curved skin panel. Fat' the torsional shear stress at torsional general instability (pSi). The equations for torSion as given In Table C9.l would not apply to shells in which there is a strong tension field that could Introduce appreciable secondary stresses in the trame. The reader should refer to Part 2 at Chapter Cll for a treatment of this subject involving the effect of semi-tension field action In the skin panelS. Transverse Shear General Instability. From Ref. 7, it 15 stated that a conservative shear general instability shearing stress may be made by utilizing the relation. - - - C9.11

Available test data on practical shells show this theoretical buckling stress to be much too high. Thus to satisty experimental results, reduced values must be used. The buckling equation which is similar to that tor curved plates, under external pressure (tram
Ref. 2) IS,

- - - - - - - C9.14 Fig. C9.S shows curves tor determining the buckling coefficient K p and shows how test data falls considerably below the theoretical bUCkling curve. Equation C9.l4 is for buckling stresses below proportional limit stress ot material. Report AS-D-568 or the xeta-oneutacs Division of General Dynamics corp. trom a statistical s'tudy of test data gives the fOllowing equat i ons for the bUCkling stress ot spherical shells under uniform external pressure for use In preliminary design. For Mean expected value:0.1561 Et Fer = r-Iern a)1,/3 - - - - - C9.15

where Fs 15 the transverse shear stress under transverse shear general instahility. General Instability In Combined Torsion and Bending. From (Ret. 7; the following interaction relation may be used to compute the permissible combinations of applied torsion and bending moments to a stiffened cylinder.

For probability = 90 percent and .95 confidence factor: F _ 0.1138 Et cr - r(sin a)~/3 - - - - - - - - - - - C9.16

MIMo'

(TITo) , = 1

- - - - - - - - - -

C9.12

For 99 percent prObability and .95 eont1dence ractor :

C9.14

BUCKLING STRENGTH OF CURVED SHE~T PANELS. ULTIMATE STRENGTH OF STIFFENED CURVED SHEET STRUCTURES.

10""""-------------------

PROBLE:MS

".

(1)
I./NC,AR / 1'NCORY~/

~
,

./0
/

The fuselage cross-section as given in Fig. C9.6 of example problem 1 is changed by increasing the skin thickness to .05 inches. The des tgn loads are increased to the following values: My = 700,000 in. lb.,
T Vz
~

--'1--

6040 lbs.

= 245,000 in. lb.

..
10 ~-------'-__7_---'-_;&~-....,_----___,
(2)

Will any of the skin panels buckle under this combined loading. The ruselage section as given in problem 1 above is subjected to an internal outward pressure of 6 pSi. what would be the compressive and shear buckling stress for the skin pa~els under this internal pressure.
REFERENCES

, ,

10

~o

70

IrJI

'100

10'

(1 )

Flg. C9.8 (Ref. 2) Test Data for Spherical Plates under External pressure Compared wtth Empirical Theory. k p ~'l E t 'l Fer = 12 (1 lie a) fd)

Schildcrout & Stein. C~itical Combinations of Shear and Di~ect Axial Stress for Curved Rectangular Panels. NACA T.~. 1928. Gerard & Becker. Handbook or Structural Stability. Part III. Buckling of Curled Plates and Shells. NACA T.N. 3883. Rafel & Sandl~n. Zffect af Normal Pressure on the Critical CompreSSion and Shear Stress ot Curled Sheet. NACA ~nL-57. Rafel. Effect of Normal Pressure on the Critical Compressive Stress of ~~rved Sheet. NACA 'tj'RL-258. Brown & Hopkins. The Initial 8~ckling of a Long and Slightly Curved Panel under Combined Shear ~nd Normal Pressure. R&~ No. 2766, BRITISH ARC (1949).
Sr~nley F.R. Simplified ~nalysis of General Instability of Stiffened Shells in Pure Bendi~g. Jo~r. af Aero. Sciences, Vol. 16, Oct. 1949.

(2 )

Fer = r-fs tn u ) "'/3

ar~le

0.0908 Et

-----09.17

(3)

For explanation of a see Fig. A.

'when c. = 90 or a hemisphere (sin a.)~/3 1s one. The test range covered was
7 0::;

~L:ll 1t
Q

~. r:.
FIg.A

(4 )

(5 )

a :; 90 0
(6)

175 :; rlt :;;:. 2000

The equations are for buckling stresses

below the material.
pr~portional

limit stress of the

(c)

Becker H. Hand~ook or St~~ctural Stab11ity. Part VI. Strength of Stl rrenec curved P'la'tes and Shells. NACA 1'.)I. 3'786.

---------~

~-. .. ,

CHAPTER ClO

DESIGN OF METAL BEAMS. WEB SHEAR (NON-BUCKLING) TYPE.

RESISTAL~T

PART 1. FLAT SHEET WEB WITH VERTICAL STIFFENERS.

gyration ot the beam section as large as possible, and at the same tUne ~lntain a flange section The analysis and deslg~ of a ~etal beam which will have a high local crippling or crushcomposed of flange ~embers riveted or spoting stress. Furthermore, the flange sections for large cantilever be~s which are frequently weIded to 'Neb members is a frequent problem in used in wing deSign should be of such shape as a i r-cane structural dee tgn , In this chapter, the general theory :01" beams with non-buckling to permit efficient tapering Or redUCing of the webs is considered. In Chapter ell, the more section as the beam extends outboard. This tapering of section should also be considered common case where the beam web wrinkles and goes over into a semi-tension field condition trom a fabrication or machining standpoint. ~he Is considered. The advantages and disadvantages most efficient flange from a strength/weight standpoint might be ~ery costly or entirely or the non-buckling and the buckling or semltension field web are discussed in Chapter Cll. L~practlcal from a fabrication and assembly The general beam theory as given in this chapter standpoint. Is bae tc to that given in Chapter ell, thus the student should study this Chapter bef~re Cll.

ctu.i Introduction.

F~_
,
,

QCQQQO_

Jf_S;in F-~--' /=1'1":~ l~in

Web

OQQQQO

(a)

t
-

'

Web

Web

Web

(b)

(0)

(d)

1
Stiffener Stiffener
Web Web

I'

....
Fig.

'. '0

(g)

Fig. cro.z

cio, 1 (From ACI Cfr-cuLar No. 622)

Fig. f Wing been of Boeing ~81ip per~ flying boat. aas ic r Lange sec~ion 1s square tube.

Fig. e

7r~ss

Fig. 210.1 shews a few typical ~etal ~eam sections ~O!" exte!"~2.11y braced wings. Such C10.2 Flange Destgn. bea~s ~~st carry l~rge axial ;cmpres=~ve leads Fer s'tr-eng'th/wef ght ~r:ic:1.':!::Cj, t ce beam 2.5 Nell as benatng loads. F1~. '::10.2 shows ~langa Sfiould :e ~esfgned tJ ~aka the radius ot typical be&~ ~12r,~~ sect:1.8TIS for ca~t!lever Cia. 1

type of be~~ t:ar.ge :ne:nber.

wi~h

ctznnel 3ectior.

' . ...1 1 .

"'II

CIO.2

DESIGN OF METAL BEAMS.

WEB SHEAR RESISTANT NON-BUCKLING) TYPE.

bea~. ror the approximate deSign of beam flanges, the ~ethod given in thiS chapter can be used.

metal covered wing cons t ruc t Lon, Sections (a), (~), (c) and (d) are typical beam flange sections for Nide box beams where ~ddltlonal stringers or skin stiffenerS are also used to :rovlde jendl~g resistance. These flange sections are generally ot the extruded type alttough such sections as (b) and (e) are frequently ~tie :rom sheet stock. Tr~se flange sections are aL~o5t always used With a beam web composed of flat sheet, which is stiffened by vertical stiffeners riveted or spot-welded to the Neb ar stlftened by beads or flanged lightening holes.
F'tga . (e) and (t) illustrate two types of flange sections used in truss beams which land themselves readily to connection with truss web 4embers. Be~ flange sections (g) and (h) are typical sections for wing construction in which ~o additional spanwlse stringers are used. In section (g) tapering of sectional area is provided by first omitting the reinforCing plates, and then gradually decreasing the extruded Shape by machining until only a small angle remains. In section (tt) a gradual decrease in section area ls ~roduced by milling out the center ?ortlon to forn an H section and then cut t tng this section down ri.; Ll y to a simple angle.

The beam flange sections (a), (b), (c) and (d) of F1g. CIO.2 are stabi11zed by the sheet covering and also by the beam web; thus compressive tests of short lengths to obtain crippling stress, and a test of a length equal to the wing rib spaCing should give sufficient information on which to base design allowables. Several lengthS of the flange section tor the truss type of beam should be tested in compression to obtain the column curve since the distance between panel points of the truss Nill vary.

Sefore deSigning any test beams, the structural designer would like to know approximately what his proposed beam flange sections will carry from a stress standnoint, since it is desirable to make test speCimens closely approximate to the sections to be finally used in the completed structure. For most of the sections of Fig. ClO.2, the ulttmata stresses can be calculated approximately by the methods of Chapter C7. For heavy sect tons Similar to (g) and (h) of Fig. ClO.2, where tte ultimate stresses fall tar above the yield strength of the material, and wher-e sene parts a f the section buckle before other parts, ~nd also where two different kinds of material are used C10.3 Allowable Flange Design stresses. in the same flange section, a logical procedure The calculating at the stresses in the beam in trying to calculate Ultimate strength of the flanges 15 in general not a di~ficult procedure section would be to make use a t the stressit the ~sual assumptions are made in the flexure stra1n diagrams of the ~terials. theory. T:te question as to wnat flange stresses will cause failure is the dif~lcult one from a C10.4 Use of Stress-Strain Diagrams in Computing Beam theoyetic~l star.dpoint. The only sure way to Flange Bending Allowable Design Stresses. determine the design allowable is to ~e sufri~ient static tests of specially deSigned In the be~ type ot wing construction, test beams. where the flange material 15 concentrated over the web ~cmbers instead of distributed over the For beam sections as illustrated in Fig. surface in the form of stringers, the allowable ClO.I, the following tests are usually necessary ultimate compressive stress which ca.~ be tor tO~ing the basis of deSign allowable developed is considerably ~bove the yield stresses. strength ot the material since the flange is composed of a section wit~ thick elements which (1) A test s~bJecting the beam to pure promotes a high crippling stress and since the bendi~g. beam flange is stabilized in both vertical and horizontal directions JY t~e r~b and Skin cover(2) A test ar a short length at be~ in ing respectively, the ~nfluence of cclwnn action is neg Ltg tb Le . Fig. (a) illustrates this type bendlr~ so t~t ~allure will occur in Neb instead of ~lange. of beam. The general flexure for.nula ~ssumes that stress is proportional to strain 'Nhich is (3) A test 0: 9. short Lengt n of beam in cor-r-ec t for stresses beLew the proportional pure ccmpressicn to obtain local cripplir~ limit of the beam material, but the ultimate strength. resisting stresses for the flange 0 f a beam such as in Fig. (a) is tar above the proportional (4) Several tests at be~s in combined l~~it, thus the actual stress distribution is bendi~ and compression, uS1ng different ;atios mor-e like the dashed line in i"1:;o;. (b) instead 0: bendi~g to compreSSive loads. ot the triangular distribution as assumed in the co~uon be~u theory. Thus to obtain beam ~iber Enough da~a f~om the above tests can stresses above the ?roportional limit, i t is usually be obtained to give rather complete necessary to consider strain and the stress whfch accompanies such a s tr-a i n , wr"iG'~ re Iat t onallowable stress cu~:es for the design of the

~i~t;~1'~1d*i~fE&~~~~A~~:7~::~i~,~iN~;~:k:YS; .:;';':-;,', ;.~:';.~:":~':~:~:~if:2~~; :":'\ :"~~~~7:;~~;i:~H~ ~~ 91;~~~~! ---- _.- . . . ----- ... _-, - -~~-

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTUFIES

CIG.3

Assumed stress / distribution in " ~ beam formula.

Actual stress

Fig. b

distributi.on above yi.eld point.

usually fail at a strain between .008 to .01 inch per inch if column failure is prevented. In Fig. CIO.5a, the compressive unit strain at the upper beam fiber has been taken as .008" /" . The neutral axis at the section has been assumed at 1.25~ above center line of beam. Taking zero strain at this point the beam section strain :'~"1_:" _ .l.::i as shown in Fig. C10.5a. For equilibrium the total compressive bending stresses above the neutral axis ~ust equal the total tensile bending stresses below the neutral axis. Tables ClO.l and ClO.2 give the detail calculations for calculating the resisting moment. The bending unit stress in column (5) is obtained from Fig. C!0.4 USing the unit strain in column (4). It the summation of column (6) in each Table is the same, the assumed location of the neutral axis is correct. The total Ultimate resisting moment for this section equals 1032000 + 1360110 = 2393000 in. lb. Using the ordinary beam formula With properties about the geometriC neutral axis as given in Fig. ClO.3 and taking extreme fiber stresses of 46000 and 52000 psi which correspond to stresses as per strain diagram or Fig. CIO.Sa, the internal reSisting moment would equal, M = t l/e 46000 x 436/969 = 2070000 in. lb. as compared to 2392000 In. lb. by the more logical solution, which is a difference of 16 percent. The discrepancy would still be larger it the outstanding leg (a) did not buckle, since the more exact solution only allowed 35000 psi on this element whereas the general beam formula stresses lt to 46000 pSi. Trial and Error Approach

ship can be obtained trom the stress-strain diagram of the material. A straight line distribution for strain, that Is, plane sections remains plane after bending 1s a reasonable one, ~~d verified by tests. In a beam in bending, one s1de is in tension and the other in compresSion. The tensile and compressive stress-strain diagrams for materials like aluminum alloy are different above the proportional limit, and the same unit strain will cause different stresses on the two sides of the beam. In frequent cases of large ~eam design, the be~ flange may be composed of two kinds at material and certain portions for attaching to skin or web ~y buckle before the ultimate strength of the section as a whole is obtained. The solution for the Ultimate internal resist1ng moment for such a beam ~ quires that consideration be given to the stress-strain diagram of the various materials and units making up the beam section. This general ~ethod of approach in studying the ultimate internal reSisting moment of a beam section can best be explained by an example problem. Example Problem rig. CIO.3 shows the cross-section at a in a metal covered wing. The main flange members are composed of heavy 245T extruded srapes. The extrusions are reinfOrced by 24SRT sheet strips. The beam web is made from 24ST alclad material. The problem is to calculate the ultimate internal resisting moment for this beam section.
be~

The location of the neutral axis is unknoWn, thus the calculations in Tables ClO.l and 2 are for the flnal assumption which is the true location of the neutral axis. The general procedure would be as 2nd Trial N. A. Fig. ClO.4 shows the stress-strain diagram follows: Assume neutral Above ll. ror these various materials. The 1/8 inch thick axis as center line axis r----";) outstanding legs of the extrusions act as a of beam, and flnd total plate stiffened on three sides and free on the axial load on each side fourth. T~ese legs will buckle at a stress or of axis. The results wll1 -e usually show excess load 35,000 pSi in compreSSion as determined by the on one flange. For the ... -1,--.,....-,-methods given in Chapter C7. The stress-strain x '" distance I diagram of the eNQ outstanding legs will be next trial move neutral ~ from tE. to o I horizontal at 35,000 psi as sho'Nn in Fig. 810.4. axis so as to give excess ~ true N. A. I AlthOugh the legs buckle, they will tend to load on other flange. ~ 1st Trial N. A. Plot the results as indihold the buckling stress under farther flange on 'l. strain. In Fig. CIO.S, each beam section has cated in Fig. C10.5b to I Dist. from" to N. A. been broken down into narrow horizontal strips obtain true location or des tgnat sd ~r01Il (a) to (w). Only that portion neutral ax1s and then make Fig. CIO.5b of the web in way of flange has been considered ~lnal calculations as 1llustrat~d in Tables in this example. Fig. ClO.sa shows the strain clstrlbutlon assumed for the trial solution. A CIO.l and 2. hea~J aluminum alloy section in compreSSion will

CIO.4

DESIGN OF METAL BEAMS.

WEB SHEAR RESISTANT

(NON~BUCKLING) TYPE.

"

l/4~

;~~4SR: 2_
\24S-T Extrusion

4-l/8xl-3/4 Plates

1/8

-60 24SRT -50 -40 -30 -20

0

-*-

.!-L

4-

1/4 J rJ

1~21
..L.

T;T
I

=9

24ST Ex.
COMPRESSION

,051 - 245- T Alclad Sheet ,. UPPER FLANGE

r;::e::-68'~ I--*N. A. at Stresses

1.25"

Fig. ClO.5

g-lO
~

;; a c,

---,L

_._--

..

_"--._-~-~-_

OlD _ 008 ..006-

_~Q04.

._002.:......:....

UNIT STRAIN .-,-" ,_ . . . . _ ._------,~~;-

below Prop. Limit. (I::z 436 in. "")

i2
ll:

'"

10
..

18" .OOS"/"

~ ~

20 30 40 50

24ST Extrusion

:;:C_;~::-'~~:':"'~:-:-:=~~ ',:, -::; -;d~- /--'''''''-.- --~:-:F'lg;--clO;'t~ ci .. -",

-:~====r:f""":"------'-~--,,--~";,~f~~~~~'~~~:E::.:.;;-: ,.
-----~'-='[IT~--,-,--~'--:~-

"--:":':"':':"'--stre~stram d i~.'c"''---'

._..-

1/8

EFig. CIO.3

10.:i~~~1.25 -

60
.000876y 70 -' .._._-

---_._----

---~-_._-"-

..-

Fig. CIQ. Sa

. ._- -, --, . ... --.. _._-------

...

_,-,--=

"" _..:.:.:::.:J.~.----,:.~,::..l-;'~.:

..

Table ClO. I

Table CIO. 2

1te.:I..'t1Dg Wo.ellt o:t

Strip ;2 area , , . u.a 11ll1t strain t e.UIl1t Load ou xc.. about _ .00876y !ltnt S1::rip .. Poz'1:. ot !ltr1p to .triP If.A. _ !~ A 1 f. , troa If.A. I I

I"

CCl~.

Seading Stroll abou't If.ut:ra1 A::II::I..

!
I

fl'. I
,I
) ~ t
~
~I!l

I:

.312~

.1230
.12~O

9.0625 1.875
8.62~

.1250 .12-'0 ,6250

.082~

8.375
8.12~

re

I .0825 I "1 1 .0625 I .01275 01275 I , = t . .0121:,; .01275

.062~

.0625 062:5

1=
0

, . ~: i~ .;~875
.01275 .2187.5 .21875 .. :t 'u 1875
I

7.875 1.62:5 1.375 7.12:5 6.875 6.625 6 .375 7.315 1.125 6.875 6.62:'; 8.375 8.3125 8.1875 8.0625 7 6875

.00795 .00778 .00756 .00734. .00712 .00691 .00868 .00646 .00625 .00603 .00581 .005:';9 00646 .0082.5 .00603 .00581

.3~000'"

10940

....800

-4~400

_451.50
-4~OOO

e...,
27900

"

.005:59
.00728 .00718

,'TOTALS
Buckling litre

:~~~

_44700 ....300 -44000 ...3300 -43100 -42700 -42000 -40700 -40100 _39700 _.38000 _38600 -5.5500 _55200 .55100
-S

.. " ..
m.
2770 2750 2710 2695 2870

99200 50800 '''00 47200

-...
"

".lIItiDg ""Ilt o:t

!!!!:!..lli

Bead. StrllSll. . about Heut::n1 Ax1.

p A .. U'lta r ~ 0.1 t IItra1.D It_un:l.t Lo.d 011 11:0 abou't , of lI1:r1p to .'trip 000876" .. 1:nt Strip .. If.oL.. f, .. !1: A 1

t .

220000
21150 20300 19300 18610 17700 16720 383' 3""

4~700

...
51

,0<
A97

ti~~~ ,

12230 12<> ec

. " .. ... ....

-w~

: t " " ij
1
0

; ,

,
e

.31250

.0937:5
.09375 .09315 .09375 .5937~

11 .s62~ 11.37~ 11.12~

10.875 10.625 10.375 10.12:5 9.875
9,6~

.010U .00987

.00976 ,009"
.00931 .00910 ,00887 .00886 .00844 .00822 ,00800 .00778 .00868 .00844 .00822 .00800 .001"78 .00948 .00936 .00926

51950

.046'
,0489 .0469 .0469 .0469 .0488 .0127$ .01215 ,01275 .012.75 .01275 .21875 .21875 .21875

ereeo
.51500 51400 51300 51200

511

3130

101800 98700 97100 91450

~q " r

c:: a ,

9.375 9.125 8,875 9.875 9.625 9.375 9.125 8.875 10.8125 10.6875 10.5625

,..... .... ..... .... .... .... 22_ ..., ..... ,...
521J50

162. .
4870

'2000

51800 .5l7M 511ee

moo

860

55450 "'00

51800

'''''00
2415 2410 2405 2400

318000

..sse

23900 232S0

21950

21300

....00 49300 411200 49100 67500 67300 6nOO

14770 14720

."

.2 .2. .27

633

<00'

5720

ssoc

14700

1511600 157200 156300

"""U
t

C10.5 Flange Strength (Crippllng).

In many cases of beam design, the r Lange ls braced laterally because It Is part at a cell construction ane. the sheet covering which ls fastened to the beam flange provides a cont1nuous lateral bracing to the flange or prevents lateral bend1ng-col'~ action for the beam flange. The beam web prevents column bending

of the flange in the web direction, thus the flange fa1ls by local crippling action and the crippling stress ls detenn1ned by the methods 0 t Chapter C7. In many cases the beam loads are relatively small and thus the area requ1red for the flange may be relat1vely small, which means a nange shape with elements of small thicknesB thus the fa1ling local strength :nay be In the elastic stress range or re Iat 1vely low.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CIO.5

In many cases, such as a frame in a fuselage, the inner flange of the f~e crosssection is free from a lateral brace, thus provisions must be taken to brace the flange laterally or the flange must be designed for lateral column action. This sUbject is discussed further in Chapter 03.
ClO.6 Web Strength. Stable Webs.

Equation C5.7 of Chapter C5 gives the bending buckling stress. It is,

b kb E - 12(1- V. ')

_11 n'

(1.)' b

- - - (Cl0.2)

In solving this equation, Figs. C4.l5 and C5.8 at Chapter C5 are used. Buckling at Web Panels Under Combined Shear and Bending From Art. C5.1l ot Chapter CS, the interaction equation tor a flat sheet panel under combined bending and shear is,
- - - - - - - - - - (Cl0.3)

A stable web beam is one that carries its deSign load without buckling of the web, or in other words it remains in its initially flat condition. The design shear stress is not greater than the buckling shear stress for the individual web panelS and the web stiffeners have sufficient stiffness to keep the web from buckling as a whole. In general, a thin web beam With web stiffeners deSigned for non-buckling is not used widely in flight vehicle structures as its strength/weight ratio is relatively poor. In bUilt-in or integral fuel tanks, it is often desirable to have the beam webs and the skin undergo no bUCkling or wrinkling under the deSign loads in order to give better insurance against leaking along riveted web panel bOWIdaries. The student should realize that the buckling web stress is not a tailing stress as the web will take more before collapse of the web takes place, thus in ~eneral the web is not loaded to its full capacity for taking load and the web stiffeners are only designed tor sufficient stiffness to prevent web buckling and not for the full failing strength ot the web. Equation for Web Buc:<1ing Shear Stress Equation C5.5 at Chapter C5 gives the bUCkling stress of a flat sheet panel WIder shear loads. The equation is
F ecr-

The expression for margin at safety is,

1

- - - - (C10.4)
CIO.7 Web Bending and Shear Stresses.

Since the web is deSigned not to buckle under the deSign loads, the web will be effective in taking bending stresses and the follOWing well known equation applies. tb = My/Ix - - - - - - - - - - - - - - (C10.5)

Where Ix is the moment or inertia at the beam section and the web is included.
known flexural

In Art. Al4.3 of Chapter Al4, the well shear stress equation was derived, namely,

ts

=.:L I Ib

ydA

- - - - - - - (C10.5)

_11s n'

kS E

- 12(1 - V. ')

no.r:

The curves of Fig. C5.l3 and C5.ll of Chapter C5 are used to solve this equation and the reader should reter to Art. C5.8 of Chapter C5 for a review of sheet buCklli.:g under shear loads. Equation for Web Buckling Bending Stress If the web does not buckle it will be subjected to longitUdinal bending stresses at compression and tension with zero stress at the beam neutral axis. Thus in general. the beam 1s subjected to a combined shear and bending stress system.

Since the 'cera I ydA is maximum for a section at the neutral axiS, the shearing stress in a beam will be maximum at the neutral axis. In general, the webs of aircraft beams are relatively thin; thus the term I ydA for the web is quite small so that the shearing stress intensity over the web is approxDmately uniform. Thus a simple formula t s = V/hb has been widely used for calculating the maximum web shearing stress. In this equation h is a distance which will make the shear stress f s check the maximum value of the shear stress as given by equation ClO.5. A Simple conSideration at the internal stresses on a small element cut from a beam in bending and shear will indicate what value at h to use in the Simplified shear equation f s = Fig. ClO.6a shows a beam element at length cut from a beam which 1s SUbjected to a

V/hb.
dx

..

__ __
..

.--~--

ClG.6

DESIGN OF METAL BEAMS.

WEB SHEAR RESISTANT NON-BUCKLmG) TYPE.

C '" Cli!

c,
dx

(e)

the bending stress centroids. For a rectangular section the effective arm 1s obViOUSly equal to 2/3 the beam he1ght, but for the common beam sections as illustrated in Figs. CIO.Il or CIO.12, the distance between bending stress centroids 1s not so obvious particularly it the web 1s considered effective in bending. A close approximation of the effective arm (h) and a procedure which is common practice 1s to assume (h) equal to the distance between the centroidS ot the web-flange rivets. The student should take several example beam sections and check th1s assumpt10n tor (h) US1ng Eq. (C10.6) against the exact values by Eq. (CIO.S).

Same structural deSigners make assumptiOns as to the proportion of the total vertical beam shear which is carried by the beam web. For fsbdx example, it is sometimes assumed that the Web (d) takes the entire beam. shear, or it may take only 90 percent. The percentage of the shear load carried by the web dependS of course upon the size or the flange sections and the torm of the bending moment which prOduces compression on the upper portion. The bending moment on web section. For example, in Fig. CIO.7, the section (AB) is M and on section (CD) M + ~, same flange is connected by a web which is thus the vectors representing the stress on attached to the flanges in two ways as illusface (CD) are drawn longer than on face (AB). trated 1n F1gs. (a) and (b). In F1g. (a), the shear flow on port1ons (AB) and (CD) help res1st The beam element is also subjected to a total shear force V on each face. C, and Cl.. the external Shear, whereas the shear !low on re~resent the resultant or the total compressive these portions in Fig. (b) act in the same torces on each ~ace and T a and TJ.. the resultant direction as the external shear load; thus at the t.ens t Le stresses. Fig. (b) shows the caUSing the shear load on the web to be greater same free body but with the tensile and than the external shear load. (See Chapter Al4 and Al5 for general discussion of shear flow in compressive stresSeS on each race replaced by open and closed sections). a simple torce C and T which tendS to move the block with the same results as the system ot Fig. (a). In Fig. (c) t.ne beam element ot Fig.', (b) has been cut along the neutral axts , and a force applied to the cut taces equal to f s bax, Where t s equals the horizontal shear stress B intensity.
Fig. ClO.6
W,b

Writing the equilibrium equation, that the sum ot the horizontal fOrces on the upper portion must equal zero, we obtain C - fa bax = 0, hence C = f s bdX
(.)

eb
(b) FIg. CIO. 7

and likewise for the lower

hence T = f s
bdx ,

f s bdX = 0,

Ie

ID

Fig. (d) shows the free bOdy ot Fig. (b), but With C and T replaced by their above eqUivalent values. Taking moments about point (0)

--

CIO.8 Shear Resistance Provided by Sloping Flanges.

1'"
fs

=fs

bdx . h - Vdx = 0, hencs

= bh

(elO.6)

Thus to obtain a value f s equal to the maximum value given by equation (a) use an effective arm (h) equal to the distance be~Neen

A large majority or tbe beams in airplane wing and tail surfaces have sloping flanges because of the taper of the structure in both plantorm and thickness. This Sloping of the beam flanges relieves the beam web ot conSidering shear load and should not be neglected. Fig. CIO.8 shows a beam (abed) carrying a load system Pl' Pli!' etc. The top flange 15 slOPing as shown. It both ilanges were extended, they would intersect at point (0). Let M = bending moment at section (ac)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES ctO.93. Flange Fig. CtO.8

CtO.7

Discontinuities.

From a weight saving standpoint, it is necessary to taper flange sections in order to approach a beam at constant strength relative to the applied loads.

---~'I
Then C = Th = M/h (h = distance between flange stress centroids) The vertical component Tv or the load T in the upper tlange equals Ttl hiLa, but Th = I1/h, hence
Tv

Fig. a illustrates how such tapering ot the tlange section may produce local eccentric flange loads. Between sections (1) and (2) the upper tlange tapers in side view as shown which

h La La Let VF shear load carried by beam flanges

- - - - - - - - - - (C10.?)

= nJ!.. = JL

1J[]lE ;]If]
F, (3)
Fig.

(2)

(i)

(4)

~~

Fig. b

Then VF = l1/Lo

a.

Thus the shear component carried by the axial loads in the sloping flange members equals the bending moment at the section being considered divided by the distance tram the section to the point ot intersection (0) or the flanges. The above derivation was based on the assumption that the entire resisting moment ~ was developed by the flanges. ~ith the web eftective in bending, the moment developed by the web should be SUbtracted fram the total bending moment ~. Let I = IF1.
+

IWeb
IWeb

The moment developed by web = M 1where

M = total bending moment on section

In airplane construction the centilever beam with sloping flange members is the c~on case, and the shear resistance by the flange axial loads is an important factor which should not be neglected it an efficient structure is desired tram a streDgth-weight standpoint.
ClO.9 Effect of Variable Moment of Inertia on Flexural Shear stress Dlstributton.

displaces the flange neutral axis as shown. Assuming there 1s no change in bending moment over the beam portion as shown, the rcrce F must be greater than F since resisting arm d is less than d. For eqUilibrium this moment due to F1. and F not being col1near must be balanced in some manner. It the flanges are rigidly connected to web and stiffeners, this moment can be balanced by an additional shear stress on the web panel between points (1, 2, 3 and 4) as illustrated in Fig. b. Thus in cases ot rather abrupt Changes in flange sections which produce the eccentricity as illustrated the web and stifteners should be checked for the additional shear flow load on the web. It such displacements in the flange neutral axis occur in the plan view, the skin covering should be investigated tor the additional Shearing stresses.

cto. 10 Stiffener Size to Use with Non- Buckling

Web.

The runcaaeneat shear stress equation ClO.5 as derived in Chapter Al4 applies strictly to beams at constant moment at inertia. For airPlane beams the cammon case is one with variable V hy,""/:s - 2 . 29d ~ament at inertia; thus the stresses obtained by IV - - - - - - - - - (C10.B) - - t - 33E equation CIO.5 are incorrect, although the discrepancy in most cases is not large. The where student should realize this fact in studying Iv = moment of Inertia ot stiftener the shear flow picture in tapered wing structures. See Art. A15.l5 ot Chapter A1S. d = center line distance between stitfeners hw = depth or web plata

A web stitfener is used to decrease the size at web panel; thus when bUCkling ot the web searts, the stittener tends to keep buckles trom extending across the stirtener or causes the sheet to buckle in two panels instead or one. Mr. H. Wagner in a paper presented betore a meeting ot the A.S.M.E. in 1930 oftered the tollowing expression as the recurred moment or inertia ot a stitfener to be used with a shear resistant web.

4-'-

' . - _

__

,-':;.~""-';'

....;_~7....,-~:

10.8

DS!GN OF M~TAL SEAMS. WEB SHEAR RESISTANT (NON-BUClCLINGl TYPE.

,. -'" ::---------"'""':::------------~------=------.::=--=""'=====" s ::,-:;--::.:,_=-.:.::=::-=

--

---1.- .
':=-'0

I:;)~-'"'-'--

:--.-,.---'~,

... -

;-~

----.....;.-'-

---..--:...-_-:::-:::~:...~::-:_.- :~. r

":":::~-

-'~

-----"~-...---~-'1

:.::-==-_:~
r~-

_ --', ... ---I - ,-,"'---'--;:-----, --':---------, .. - ... .. ----'_ ... _--,---_._-~'.....,..~

~ ------'~>_.,.---

'~-'

-.

--'

'--"

v
t

= vertical shear at section

:

web thiCkness

= ~oculas af elasticIty
A more recent criteria [or stl::ener stiffness (Iv) for both rlat and c~rred webs 15 given by the curve in Fig. C10.9. When the stl!tener 1s used pu~ely aB such and not as a means to transfer a concentrated external lead to the beam ~eb~ the question arises as to what 15 the mln~um number ~f :asteners reQUired in at~achlng the stltfener :0 the web. For nonbuckling Nebs, two criteria ~re suggested:(1) The S~i!tener should be attached to the flange at each end.
~he

consists of separate web and flange members, tastenec together by r1vets~ bolts, spot welding or continuous welding. In the design ot such heams it 15 thus necessary to know what loads the rivets, bOlts~ etc., are Subjected to in order to provide the proper connect1ve strength. It 1s quite easy to substitute in Simple {onnulas to tind the loads on beam flange rtvets. however, the student should be sure that he understands the ~~ental beam act10n behind these

formulas.
FIg. CIO.IO illustrates a beam portion equal in length to the flange rivet plt~h p. The beam section at (A.A.') 1s subjected to a bendf ng moment M and 11 + &"1. at section (CG' ). The bending stress Qistr~bution on the beam faces is indicated by the stress triangles. In this exampl~ section it will be assumed tr4t the w~b takes ~o oending stresses. Let PF equal the total Dull on the flange angles due to bending stresseS at section AS due to bending moment~. ~hen total pUll on flange angles at section 20 due to a moment M ? An on beam section C~' eQualS ?r ~ ~Pr. Under the action of these two forces the tlange angles would move to the lett, but this ~ovement 1s prevented by the rivet WhiCh ties

(Z)

r1vet PttCh (spaCing) should

accor~1ng

to (Ref. 1) be at the mcs't equal to 1/4-

times the stiffener spaCing, or 1/4 the web height i t this is smaller, in order to justify tne assumption ot SimPle support at the edges or ~he Neb panel. Normal practice useS ~ore rlvets.
CIO. 11 Notes on Beam Rivet Design.

;Xcept for very small beam sections Nhtch. $aY oe extruded as one piece, the usual beam

--------_...-

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CIO.9

I-.:PF + 6.PF
\

P
'\it.

~-PF
I
I
M

\
\

M+o
11' I I

\ V

N.A.

" " " " I'

"

equation (CIO.9) would be wrong because the moment of the bending stresses on the web about our moment center is not included. Thus to make the stmplified equation (CIO.Il), check the exact result as given by equation (CIO.12), the distance (h) would have to be greater than the distance between the flange stress centroids. In fact, taking (h) equal to beam depth would not be far off from the results o( equation
(CIO.12)
CI0. 12 Loads on Rivets Attaching ReinforCing Plates to
Flange Member.

\
D'

B' \
\~i.1

Pp + 6.PF

J--PF

C'

A'
Fig. CIO.10

" "
"

\I

the flange angles to the web. Thus load on rivet equals aPr. The same reasoning applies to the rivet holding the bottom flange angles to the web. The beam portion of Fig. CIO.IO as a whole is in equilibrium under the bending stresses and the shear load V on each face. Taking moments about the intersection of the lower flange bending load and the beam face

The beam flange is commonly composed of a main unit plus several reinforcing plates or parts which are held to the main unit by rivets or spot welds. Fig. CIO.II illustrates typical beam flanges. The basic section of the upper flange is reinforced by the plates (a) and (b), and the lower flange by the plates (c) and" (d). The purpose of the rivets is to keep the reinforCing plates trom sliding along the tlange tee section due to the bending ot the beam; thus making the plates effective in bending. This horizontal force tending to slide reinforcing plates and which is prevented by the rivets in shear is given by the fundamental shear flow equation (ClO.12), namely
q -

cc' ,

[Pr - (PF
hence,

OFF)] h' Vp =

a - - - -

(CIO.9)

i;

ydA, whfch

(CIO.IO)

Equation CIO.IO says that the change in tlange axial load in a distance p equals the vertical shear on beam times rivet pitch p divided by the distance between flange bending stress centroids. 6PF is also the horizontal shear flow produced, by flange angles over a length of p. The horizontal shear (q) per inch due to nange would be
q
111

V/h '

(CIO.ll)

The general expression for shear tlow (see chapter A14) is,
q =
te~

t; ydA

- - - - - - - - - -

(CIO.12)

I and I ydA of Eq. (ClO.12) are not required, and the distance h can be estimated closely without calculation, and will be greater than the distance between the centroids of the r Iange areas.

;q. (CIO.II) 1s easier to use since the

equals the horizontal b b shear per inch along the beam. To tind the load or rivet section 1-1 of the upper flange of Fig. CIO.ll, the term / ydA equals the area of the plate (a) times the distance tram its centroid to the neutral axis. For shear load on rivet at section (2-2), the term I ydA equals the area ot plate (b) times distance to neutral axis. Fig. CIO.11 On the lower flange rivet section (3-3) is critical since "both reinforcing plates are on same side, and the entire shear flow produced by plates (c) and (d) must be resisted by rivet section (3-3). The term. I ydA would thus equal area or both plates ttmes the distance to the neutral axis of the beam. A Simplified method which yields good results is based on the relative areas of the units or the total flange. To deSign connection of flange to beam web, the total horizontal shear q prOduced by entire flange is always necessa~J and involves the use of the entire flange area in the shear flow calculation

Equation (CIO.II) was derived on the assumption that the beam web took no bending stresses. In general this is not true or only partially true. With the web taking bending,

q=t/YdA.

ClO. 10

DESIGN OF METAL BEAMS.

WEB SHEAR RESI5TANT (NON-BUCKLING) TYPE.

The shear flow produced by a reinforcing plate 1s then taken as proportional to the area of the plate over the total flange area times the total tlange shear tlow. USing Simplified equation (CIO.II), we can 'Nrlte v ap

Ry

= 1.. n

- - - - - - - - - - - - - - - (C10.14)

where v = net web shear

n = number of rivets in splice.
It butt splice, n, equals the number of rivets on one side of splice.

h'

AF

- - - - - - - - - - - - - - (C10.13)

where

= area of plates under consideration

total area of flange = load per inch on rivet and h' same as explained before For rivet loads in beams with sloping flanges the shear V is the net shear as explained before in discussing web shear stresses.
ClO.13 Web Spllces.

In bending the splice rivets must transfer the bending moment due to the bending moment M developed by the web. The largest rivet load on a rivet due to bending will be on the most remote rivet, e.g., rivet (a) at distance r a trom center ot rotation ot the rivet group. Then load on this rivet due to web moment equals
R_ -'"'8.

= M;"ra Zra

- - - - - - - - -

- - (CIO.15)

where Usually in designing a sheet girder beam, it is necessary frcm a weight saving standpoint ~o use several web sheet thicknesses, which means web splices. Fig. CIO.l2 illustrates typical web splices. Fig. (a) is typical tor a
_ T

Zr a = moment at rivet group which equals the sum ot the squares at the distances of the rivets tram the center or rotation" at bolt group. The resultant load on the critical rivet will equal the vector sum of the values or equations (C10.14) and (15). Since in most cases only two rows of rivets are used in a web splice, a close approximation tor the moment load on critical rivet can be written by USing the vertical distances (y) to the rivets instead or the radial distances (r), the resulting torce acting in the horizontal direction. Hence

-I-

--r
ra

++ tot
+

-r-i+_~_ -- .:!:. HI
HI

+I,!,+

+tl++
++I1"t ... +1++ !++ H
Splice Plates

to+l

++:tt
I

+ +-l-."La ++1 S Ice .-1-, '

++- ttH1+t
.j. ...

+l-

.... ..j..~~

+1....1.+ ...

:+1. __ +71= +1... ;- ..

(b)

++,

H, H,

Rma i1,yYa Zy"

- - - - - - - - - - - - - - (CIO.16)

The resultant combined load on critical rivet is

(C10.17)

(al Fig. C10.12

ccrmparatively heavy web which prevents joggling at web as in the case or Fig. (b). In the case or Fig. (b), the lap is usually made under a web stiffener which provides a support for the web in driving the rivets through the thin web sheets. Loads on Web Splice Rivets The web is SUbjected to shear loads and for stable webs, the web undergoes bending stresses. For rivet design it is usually assumed that ~he web shear stress 1s constant over the depth. .hus the vertical component of load on each web splice rivet is the same or

The student should review Chauter Dl tor more detailed information on rivet'loads due to moment loads on riveted connections. When the web at one beam Is joined to that at another beam using shear nclips,n a special problem may sometimes arise regarding the adequacy or the clips. This deSign problem is discussed in Chapter 03.
CIO. 14 Example Rivet Problem.

Fig. ClO.13 shows the cross-section ot a riveted beam. Ir the deSign vertical shear on the section is 3000 lb., check the strength of the riveted connection.

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

,

CtO 11

1-'-, -I
i

1-'-. -Iri6
r

ydA t1rst moment

N.A.
~

t tlange area about

Deta.U of

Flange Angles

3/4

,*-*'T ~U'99
"
Area = .089 IN.A. =.004

I-t-l.L

+- -+,'." 'Ji' 10, i--l t i

All Rivets 1/8 Dia.. 2117-T3

...l

1.'7"

4-.051

angles = 2 x .089 x 2.731 :: .486 Reintorcing plate

Total
~ 1~75

j
.57

x .0625 x

2.961

= .324

r
Web

.810

.-

I--i-~
5

-4- 9'

Fig. Cl0.13

Depth

SUbstituting In squatlon (A)

q 1.125

= 6-7/8"

3000 x 6.I7x

.810

= 443 lb.
V =h'

Tension Flange

Shear rlow by S1mpli!1ed equation q hi = beam depth" hence

q :: 1.125 x
7.062~

solution: The loads on the rivets will be calculated by the ~exactn and also by the sunplified approximate equations. The exact shear flow equation i5

(see Art. ClD.lD)

73.~~ =

478 lb. which is con-

servative compared to 443 by exact expreSSion.

The web is attached to angles by 1/8

q.:J.../ I

y dA

The first step will be the calculation of the moment Of inertia of the beam section about the neutral axis. The bending load3 are such as to put the upper flange in compression. The moment of inertia will be calculated about the centerline axis of the beam section and then transrerrec to the neutral axis. Table ClO.3 gives the detailed calculations.
Table CtO. 3
..",
1JllpoIlr

diameter 2117-T3 aluminum alloy rivets and are acting in double shear.
From Chapter D1:

DOUble shear strength or 1/8 - 2117-13 rivet = 2 x 388 776 lb.

Bearing strength on 2024-T3 - .051 web plate:: 630 Ib~, thus bearing is critical and Margin or safety .. 443 - 1 :: .42 and w1 th the apprOXimate method the margin ot safety
would equal (630/478) - .1
630

\"1'_ . lA,
!
.1780
.1~3

,
3.301 3 .531

aAC1. .

Ull~J' reULtazoe-1 ml' plat.

I I .3,510

" ,,' .s .. '.M .,.. , ,

1.31 .067 -

.31.

.001

1.9"1

Check ot Cover Plate Rivets:J::

l/n J:: 305 1 6.87S = 1.38

il

Rivet spacing 1.5" with two rows or ri vets.

By
6.43

.00'

I I

I_.o~ ~3.012

.206l
I

exact equation, load on two rivets

w

Still

.1143

.4"

1.51/ ydA 3000 6.l7 x .1093 x 2.961 - 236 lb.

~

-; lli
U.

j""7j43

.H"- er-

1.5 x

tw .-'-... 6.4.3 _ .7"'3 J:: .:17 3 .. 8.17

La."

Load per unit:: 236/2

8y Slmpllrled formula:

118 lb.

1~

to .. _ l i t of I.aert1a altOll:t ita 0WlI celltlrOidal uie.

Rivet load on upper flange rivets which attach angles to web: Rivet pitch
~

(:J... h'

a 5 plate) 1 5 = ( 3000 X .1093) a 7.062 .2873 1. flange

1-1/8 inch

243 lb. = 122 lb./rlvet.

Horizontal shear load per 1-1/8 inch 'cj.stance equals

q 1. 125

The rivets are in Single shear which is crit1cal and equals 388 lb. as given above.
(A)

Y/

ydA - - - - - - - - - - - - -

Hence margf.n or safety

388 ill - 1 = 2.28

CIO.12

DESIGN OF METAL BEAMS.

WEB SHEAR RESISTANT (NON-BUCKLING) TYPE.

T
8"

I-

240q lb.

1 ~I~

mllU~. . r
...

25"

1 /,
I I

Flange Rivets 1/8 Dia. At. 1" Spactng 2700 lb. I 25"

Web Stiffeners 1/2 x 1/2 x . 04 Angles. 2400 lb. 25" 25"

."

...

-I

..

... ...

Bav 1 web e . 057

Bay 2 - Web = . 072

2
Bay 3 - Web'" . 072

""

.
I
I

.
Bay 4 - Web"" . 057

Fig. CIO. 14. Sketch of Beam (See Fig. C10. 15 for Cross-section of Beam)

C10.15 Example Problem. Strength Check of Beam.

Fig. ClO.14 shows a bUilt up (I) section beam, Simply supported and carrying its three concentrated loads as shown as the design load :or the beam. Check on Bending Strength of Beam Since the beam crosS-Se~~ton is constant, the critical section IS at the midpoint of the beam where the bending moment is:
11"ax 3750 x 50 - 2400 x 25 = 127500 in.1b.

. i - - . - i- ~

L
I

N.A.

T 1
3/4

I-~

3/4

--I .L
.214

1-1- - - I I . A . !
Upper Flange Angle Area :.132 IN.A. .00 Material 2014..T5 ExtrtlStion

31"""

As indicated in Fig. CIO.14, the beam is rivetad to the sheet covering. On the upper flange which is in compreSSion under the given loading, a certain effective sheet width will a~t with the beam flange. This effective sheet Width depends upon the beam flange stress which is unknown as yet. As a preliminary value, the assumption that a width or skin sheet equal to 30 t as acting with each rivet line is a reasonable one. on the tension or bottom side, the entire skin sheet is effective or 2 inches to each side of the beam which Is the distance half way to the first skin stiffener on each side. Fig. CIO.15 shows the details of the effective cross-section at midspan of the beam. Three rivet holes are assumed in the tension flange. Table CIO.4 shows the calculations for the section properties, first about centerline reference axis, then tra~~terred to the neutral axis. Bending stress at ~idpoint of horizontal leg or upper anglas:- IN.A. ~ 10.44.
!b =
~/I

~J Rivet Holes 1. l ~.020 -t::::~4f"~=~.~ L.:::=:;==:::::> .199 4" I.1 r--- 3/4 ---j T
; 7-7/8" .7/18

w._-.,t~
I
Part
Upper angles Upper skin
Web
Area A

I~

tt - --N.Al

,\ :::7:::-

ill

Fig. CIO.I5

Table CIO. 4
y
Ay:l Ay

.254

3.786 1. 000

3.786.012

3.798 0.604

.0375 4.012 O. 1504 0.604 .4490 0 0 1/12 x .072 x

2.310

7.875
.. 2.31 Lower angles Lower skin .1780 3.801 -0.676 .080 -4.01 -0.321

I 2.570
1. 290

.008

2.578

1. 290

Skinrivetholes,-.0206 3.959 0.082 Web flange rivet holes

.SUM

-0.323

-0.323
-0.288 9.980

127500 x 3.614/10.44 = 44200 pSi

-.0227
1-3.5621

0.081 0.316

-0.288

The assumption ot 30t ~ 0.75 inch as the skin effective width w will be checked. From Chapter C7J
'N

.9249

y'" .316/.9249 =.34

~. A. '" 9. 98 . 925 x .34
1 '"

= L9t .j E/FST

9. 87 in.

Section with. 072 web:IN. A. .. 10,44 (N. A. Loca.tion Same. )

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

CIO. 13

FST

flange compressive stress which 44200 psi. Then,

Check of Web Buckling Stress The maximum Shearing stress occurs at the support point and equals 3750 lbs. The web thickness at this point is .057 and the web stiffener spaCing is 3.57 inches. The maximum Shearing stress on the web by the Si~litled equation is,
fs

w ~ 1.90 x .025 J 10,700,000/44200 =

.74 in.

Since 30t or 0.75 inches was assumed tor w, the erTor is quite s~ll and no revision is necessary. Calculation of Allowable Failing Compressive Stress for Flange. Since the flange is braced in two directions at right angles to each other by the skin and the web, column action is prevented and the allowable stress will be the crippling stress of the flange unit. Chapter C7 gives methods at calculating the crippling stress. In this example problem the Needham method will be used and the crippling stress for a single angle will be calCUlated.
b: = (.75 _ 3/64)/3/32 = 7.55

= V/ht = 3750/7.125 x .057 = 9230 pSi

By

ts

!~

the more exact equation, 37x 50.057 / ydA ~ 9.87 (.264 x 3.446 +

.0375 x 3.673 + .057 x 3.6 x 1.8)

= 9450

psi

Since the bending moment in the first web panel adjacent to beam end support is practically zero, the web can be considered as subjected to shear stress only.
C10.lo

The buckling stress is given by equation

n kg E
(1- ')
b ::: 3.57 inches

From Fig. C7.5 of Chapter C7, we obtain Fcs/v FcyEc :: .068

For 2014-T6 Ext., Fey = 57000, Ec = 10,700,000. Whence Fcs = .068 x ,; 57000 x 10,700,000 :: 53,000 PSi Since the angle legs which attach to web are riveted together, the crippling stress for these legs would be higher than for the angle legs connecting to the skin, thus the value of 53,000 is somewhat conservative. The margin or satety tor the upper flange is (53,000/46,700) - 1 = .14. Check of Bottom Flange Tensile Strength;fb

a = distance between rivet lines::: 7.125 alb 7.125/3.57 :: 2.0. Assuming simply supported edge conditions, we obtain kg = 6.4 trom Fig. CS.ll at Chapter C5.

F = ,,X6.4x10,700,OOO (.057). = 15,800 psi ecr 1.2 (1 - .3 ') 3.57

This stress is below the proportional limit stress or the material so no plastiCity correction is necessary or 71 s ::: i ,o, The margin of safety against buckling is thererore (15,800/9450) - 1 = .67. This value is somewhat conservative as boundary condition tor web panel is no doubt larger than tor Simply supported, which condition was assumed. Check ot .057 Web at End at Bay 1. The web will be more critical at this point because the beam is still Subjected to an external shear load at 3750 lbs., plus a bending moment of 3750 x 23.22 :: 87000 in.lb. at the midpOint ot the end panel. Since the web is clamped be~Neen the flange angles, buckling or the web will occur adjacent to the lower end ot the upper flange angles or at a distance at 2.91 inches tram neutral axis.
f

=~~/! = 127,500 x

4.35/10.44

53,000 pSi The material for lower skin = 2024-T3 sluminum alloy, which has an ultimate tensile strength of Ftu :: 65,000 PSi; and the extruded ar~les have a Ftu = 60,000. Thus M.S. (60000/53000) - 1 = .13. As a practice problem for the stUdent, the bending resisting moment should be calculated by the method of Article CIO.4 and compared with the above margins af safety. The result would shew a greater margin or safety.
~ltimate

=MY/! = 87000

x 2.91/9.87

=~

The bending buckling equation is,

;l-';t>L:-':';;~~~~::~f:::;t<t~~'~,,~l'~:,* I''::',: ~"'}~~ .,.(.. ~;- j-;:Y~,"~: ~~~t:."<. ::-~i:/ .,'-~ ":~: "~>~'~

---

;.:; r: ;>~ -:.;,.:

Y{""

~~, f :<:~" .~ ,., zc~_...>!~"~~y

--_.-- .... . .
-~.....,.

CIO.14

DESIGN OF METAL BEAMS. WEB SHEAR RESISTANT (NON-BUCKLING) TYPE.

_ n 3X6.4x10,700,OOO (.072). = FScr . 25300 ps t 12 (1 - .3') 3.-;7

= 5.5 in. taken as distance between of flange angles since web Is clamped in between flange angles. a = stiffener spacing = 3.57. The alb: .55. From Fig. C5.lS of
b
~dges

RS
Rb

= 2710/25300
1

.107
.94

= 34200/36300 =

Chapter C5:-

M.S. : ';.1073 + .943 - 1 = .01 Check of Stiffness of the Web Stifteners.

kb tor Simply supported edges = 36 kb tor clamped edges = 50 We Will assume an average value of kb = 43.
b er
Ii'

= n a X 43 X 10,700,000 (.057). = 32200 2

12 (1- .3
)

5.5

ps

The web is stiffened by 1/2 x 1/2 x .04 angles on one side of the web. From Equation (C10.8), the moment ot inertia required ot web stiffener to prevent web bUCkling is,
_ 2.29<1 (V hw) ,I. I ST - - t - 33 E

15800.

The shear stress at this section will be same as at support since the external shear load Is the same. Thus f s = 9450 and Fs =

cr

2.29x3.57 (3750X 7.875 )'/' .057 33 x 10,700,000

The interaction equation for combined bending and shear IS,

Rb::ll .,. Rs::ll

= .00050

tb/Fbcr = 20600/32200 tS/F s cr = 9450/15800

M.S. = -;==;=",1=;;= J Rb'i .,. RS'i - 1
11.8. = .01

The moment of inertia ot stiffener crosssection is .00094, thus stiffness is satisfactory.

.795

= .598

=J

The required moment of inertia will also be checked using curves in Fig. CIO.9. The ,lower scale parameter in Fig. ClO.9 is d/h Vk; 1 in buckling equation must be .795'1 + .598'1 - 1 where k s as used multiplied by 1'1 3 / 1 - .3 3 = .9OS. ThUS,
d/h ~ = 3.57/7.125 V 6.4x .905 = .208

Check of .072 web at Centerline ot Beam The bending moment at the midpoint at the web panel adjacent to the beam centerline is 123,200 in. lb. The shear load on the panel
Is 1350 lb.

Fram Fig. ClC.9 we read 1.1 for Iv/dt~. Whence Iv = 1.1 x3.57 x .057~ = .0007'35. (at i r renerO.K.)

Check of Rivets Attaching Web to Flange Angles.

tb = r1Y/1
F

= 123200 x 2.91/10.44 = 34200

x 43 x I O, 700 , O OO (.07~)3 12 (1 - .3") 6."

psi

End Bay.
PSi

V = 3750 lb., t = :057

'cr

=n

= 51500

The horizontal shear load q per inch by approximate formula CIO.lI is q = V/h' = 3750/8 = 468 lb. per inch By the most exact ClO.12 equation,
q
q

This value is in the inelastic stress range so correction must be made for plasticity. ~~e curves of Fig. CS.8 of Chapter C5 will be used. For 2024-T3 Sheet, Fa.? = 39000 and n = 11.5. The value at the lower scale parameter in Fig. C5.8Nill be Fbcr/Fo.? : 51,500/39000 = 1.32. From Fig. C5.S for n 11.5 we read FDc/F o . ? = .93. T~erefore, Fbcr = 39000 x .93

if

ydA

= f;~

(.264x3.446+ .0375x3.673)

=398

= 36300

PS1.

The shear stress is,

fs

The Single shear strength of' 1/8 diameter rivets made tram 2117-T3 aluminum alloy is 388 Ibs. Since rivets are in double shear and 1 inch spacing, rivet shear strength 1s 2 x 388 = 776 Ibs. versus the deSign load ot
398 1bs.

- V

IT

ydA - 10.44x .072,1.51

1350

= E1Q

ps t

The bearing strength of a 1/8 diameter rivet on .057 web is 790 Ibs. as against load

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CI0.15

a~rt

of 398 lb. Thus rivets could be spaced further as margin of safaty is rather large.

The rivets attaching the skin to the upper should be spaced to prevent inter-rivet buckling of the skin, since the skin was assumed effective in computing the beam moment of inertia. The rivets attaching the lower skin to the lower angles would be checked for the shear flow load and not for inter-rivet buckling since skin is in tension. The general subject of rivet design for structures 1s treated in Chapter Dl.
fl~ge

beads or a series of flanged lightening holes which stabilize web against buckling and provide low fabrication and assembly cost. ThiS type of web deSign is treated in Part 2 of this chapter.
CiO. 15a Use of Longitudinal Stiffener to Increase Bending BUCkling Stress of Web Sheet.

The web stiffeners at the external load points and at the beam support paints mus~ be designed to transfer the concentrated load to the beam web. Refer to Chapter AZl, which treats of loads in such stiffeners. General Comments The reader should understand that the margin of safety for the beam web in the preceding beam check was based on the design reqUirement of no initial buckling under the design load. The buckling stress as calculated is not the stress that wculd caUSe web to fail or collapse, as the web could take conSiderably more load in the buckled state. The subject of beam design with bUCkling webs is treated in chapter Cll. In general, this type of beam web design is not Widely used in flight vehicle structures becaUSe it is heavy construction since the web thickness must be relatively large to prevent buckling and the cost of fabrication and assembly is relatively high because of many parts and much riveting. The aerospace structures engineer decreases these disadvantages by using a web sheet With closely spaced

The strength check of the beam in the example problem. showed that the compressive stress on the web due to beam bending was the factor that had great effect in producing the web buckling. This web weakness can be tmproved by adding a single longitudinal stiffener on the compression side as illustrated in Fig. CIO.I6. Theoretical and experunental information on the effect of such a stiffener is quite limited (see Ref. 3). Such a stiffener can raise the buckling coefficient kb to around 100 or more, thus the Web can be made somewhat thinner if bending stresses are critical. Adding this longitUdinal stiffener means another structural part and more assembly cost and therefore such construction is not widely used although it is a structural arrangement that will save structural weight under certain conditions ot beam depth, span and external loading.

,.
~
" II

Vert.

" II !l1 ...... Vertl~ y stlff,-~

It II

stiff.

I-- Web

Fig. CI0.16

C10.16

DESIGN OF METAL BEAMS. WEB SHEAR RESISTANT (NON-BUCKLING) TYPE.

PART 2. OTHER TYPES OF NONBUCKLING BEAt"r WEBS.

(BY W. F. McCOMBS)

C10. 16 Other Types of Web Design.

Table C10. 5
t

At this point it should be noted that the web design discussed in Part 1 required a large number of parts (the stiffeners) to achieve lightness. To keep manufacturing costs down, the number of parts ~ust be kept to a ~inlmum. A balance. or economic compromise must be found between manufacturing expense and weight penalty. This can be arrived at since in every a t t-p.Iane design some "dol Lar-" penalty can be assigned to every extra pound of weight. The exact figure will depend upon the type of airplane being designed. Thus, eliminating pa~s may incur some weight penalty but if the savings in manufacturing cost is greater, then the reduction in parts is economical. Three types of shear resistant, nonbuckling webs are frequently used in aircra!t design to save the expense of stiffeners. Actually, the web in most cases is as light, or lighter, than a web with separate stiffeners. There is a general limitation, however, in that a stiffener must be provided Wherever a significant load is introduced into the beam. ~he web types are:
a)
b)

B
,1.65

.020 .025 .032,,.040

.95
1. 05

.32
.40

.072 .081

1. 15 1. 30
1. 45
1. 63

1. 73
1. 81
1. 92

1. 16,,1. 27 1. 42
1. 55

.52/

.091
.102 .125

.64 .81
1. 02

.051
.064

2.12

2.00

The allowable shear flow, q in pounds per inch at collapse, 1s given in Fig. CIO.IS, by the solid lines. The dotted lines indicate initial buckling setting in but this is not failure (whiCh is given by the solid lines). This 1s the strongest of the web systems not having separate stiffeners. Failure occurs
Fig. CIQ. 18

BEADED SHEAR PANELS

DESIGN CURVES FOR CLAD 2024-T4 AND 7075-T6 SHEAR PANELS STIFFENED BY MALE BEADS AT MINIMUM SPACING

web with formed vertical beads at a minim~ spacing. web witn round lightening holes haVing 45 fOrmed flanges at various spacing. web with round lightening holes having formed beaded flanges and vertical formed beads be~Neen holes.
;

c)

T'!1e webs with holes. (b) and (c), also provide lots of built-in access space for the many hydraulic and eleccrical lines and control linkages in airplanes.

"~ " " ~lt ... " $! ~

;

.
0

cro, 17

Beaded Webs.

Figure CIO.l? shows a web having ~malen beads for.ned into i~ at the ~inimum spacing f?~ing will allow. The cross-section of the ~ead is described in Table CIO.5.

i~i , . s: I g , l "

i~;

(0'000000":0 DO O~ODD "

. . . . . . . . . ..
Loading / Stiffener /" Fig. CIO. 17

~~%~~~~~~~. R~R
. , View A-A

. ! Flange

~ Load

25

! , ! i

~ 's ~

See Table CIQ. 5

i l:

REFERENCE: STRUCTURAL DESIGN MANUAL SECTION 4.230

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

C10.17

when the beads collapse. An example will [ollow haVing this type of hole. ReferrIng to Fig. later in ClO.19. A sUitable stiffener must be CIO.l9, the allowable shear flow is, from used (replacing a bead) wherever a load Is Ret. (2). Introduced to prevent earlIer collapse of the bead. The allowables are for pure shear only, qALL.= k t (1 ') + t sc no normal loads (see Art. D3.S).

~Sh

(~)

AJ C~

ctO. 18 Webs With Round Lightening Holes Having Formed 450 Flanges.

- - - -(C10.18)

Where
k

This is a Simple easily formed web. Fig. CIO.19 shows such a beam with a cross-section through the web at the holes. The geometry of the hole and its flanges is given in Table ClO.6 for typical forming.
Flange

.85 - .0006

Collapsing shear stress of a long Plate Of width h and thickness t; obtained from Fig. CIO.ZO.

+ Load
'L-

= Collapsing shear stress of a long

Plate of width c and thickness t, obtained from Fig. CIO.20.
LOading

'+-+-+-+-0
b h C'
C

= hole

centerline spacing.

Stiffener

= height of web, between flange to

web rivets. C-2B where B is given in Table CIO.6 for a typical hole.

Fig. CtO.19

b - D

The N.A.C.A. has developed, from an extensive test program, an empirical formula that gives the allowable shear flow (COllapse) for webs
40,000 38, 000 36.000 34,000 32,000
I

For this type at web deSign, in general, those webs designed to take Ultimate load will probably not show any permanent set at yield
20, 000 18, 000 16, 000 14,000 12,000 rJi

I
I I I I I

"

...
en
l
0

,
I

...

0:
~ 30, 00 0

II I I
I I
,

1\ I
I \I \
I \

0: I 10, 000 ~

.:!'

.;.'

28,000 26,000 24,000 22,000

I

1\ ,
I

II
I
I

(in.) I-

-"
8,000 6,000 4, 000 2,000
0
~

. 015

I
I , ,

_f'

I
I

I
I
I

I I
I

.020 ~ .025 I-+-L032 , .040 I- Ofrt l-

.oee
I I

1\
60

20, 000

! I 20 40

I I
80 100 120 140 160 180

NAbI A R. "- DEC I

200 220 240 260

9'
280 300

320

h/t or c/t
COLLAPSING SHEAR STRESSES, fSh OR fs c FOR SOLID WEBS OF 24S-T ALUM. ALLOY Fig. CtO.20

C10 18

DESIGN OF METAL BEAMS Table C10. 6

WEB SHEAR RESISTA."fT (NON-BUCK! ruG) TYPE

D
1. 31 1. 56 1. 81 2.08 2.33 2.58

D
3.61 3.86 4.11 4.36 4.61 4.86

D
6. 16 6.66 7. 16 7.71 8.21 8.71 9.27

13 .13 .13 .16 16 16 .16 16 19

. 13 013 . 13 .16 .16 .16 .19

.19 .19 . 19 .19 .19 .19

o

19 .19 19 019 19 .19

.25 .25 .25 .32 .32 .32 .37 .37 .37

.25 .25

In addition to the above (co Lrapse ) , the web should be checked for net shear stresses through the holes to be sure that sNET <: " ~ su ' at ultimate load, q x h. For a more complete discussion of this typ e of web design and the test data, the reader should review Refs. (2) and (1) . Design charts can be prepared from formula (18), Fig. CIO.lS, and Table CIO.6 for use in designing without having to resort to :or:nula (18) Figure Cl0.2 1 shows such a chart taken from Ref. (2) for the cases of Dlh = .80 and D/h = .50.
ClO.19 Webs With Round Beaded Flange Lightening Holes and Intermediate Vertical Male Beads.
A third type of web has round holes 'Nith beaded flanges and vertical "maLe n beads between the holes. Such a beam is shown in Fig. ClO.22. The vertical bead is as described in Table ClC.5. The beaded flange shown is described in Table ClO.7. For the particular .6 and the hole spacing is case where ..Q..:.Q...:.h the allowable shear flows shown in aqual to h, Fig. ClO.23 apply.
.~

I .25
.25 .25 .25 .25 .25 .25

2.83
3.08

.16i15.11

19

5.36

.19 .19

3.36

.19115.61

I . 19 I .19 I .19

9.77
10.27

Load if the shear stress at the yield load h over the net sections, C' x t and h-D x t, IS

about equal to the yield stress of the web cat er-tat , FSy' In general, it Nill be found, from formula that the larger holes (about 0 ':::::< .an) with wider spac mgs , b, will give the lightest web in designing for a given shear flow q , but the stiffness will be less, of course.
(18) ,

Again, the allowables ~er tor.nula (18) are for pure shear only, no nor.nal loads on the web Thus a stiffener ~ust be prOVided wherever a load t s introduced into the beam and also in areas where the beam may ha~e significant curvature as in round and elliptical bulkheads, see Chapter D3.8. An example is gl"lsn in Art. CIO.19.
:i 900

............

~Load .. . . . .

....

:~
...

. . ..

'.

. ..........

: 00000
'-- Loading Stiffener Bead

. h ' . :1 .1.
tr-

lilT

Fig. ClO.22

~
Ccjl
vtew
A-A

~1'f1 i R fHolel~6
I

O.D.-..I

Hole Geometry

~ ...
'" '"
~
I

800 I 700

: ,

. - -- -I- f- f, , I
;
I

D/jh",

5Q

1
,

I I I I
I

,
I I ,
I 1 ,
b-2125O

1
,

i
I

,!it -l4"
I ! lI I I I
!
!

I I I I i I , I I I
1 1 1

1
!

,
! h~"
,

, i 't
,

I ' I
I

, I

D~h.' 8

I
,

pP. QD I

I
i

h=4" I , ill of I , I I I I I I ! { i '7" /i ,.:7" iJ I I I I ! I I i z , ; I II ~ I I I I I I i I ! I ~ 400 '" J4" Ii , i{ I I ;'lhJ4" I !b =It. 5D I i ! i, I V I, I j , I , 10'~ 1 '/1 'I ,I": 7,,1 I ! I Y I' I " ! I 1 S 300 I I , I I ; j /~1", , V.f i~ I . .J' 'lO'r I i "I . , 101' I i I! ,,( ,j "'i I J/ A' 1 V" '10'r <'" ~ 200 " V1 ',Y ,/'0' , I I I IXJ.. /1 ~v 101' I 'I ,.{ ,I A o ,..( i I 1 III ,%'0 ( : I , , >~ ... i=-::::f"" , / y I , , .t ...{.. .Y-" .? I , I yr- I.;:A" I ~ 100 I i ~A",,~ :::-+--": i i Y1~.----r , I I I i I ':"l"AlCAJ A: R.iR.' DE '" I i I r '" . 01 .02 .03 .04 . 05 .06 .02 .03 . 04 . as .06 .02 .03 .04 .05 .061942
I

~ 600 ~ {ii 500 (3 !

I I

I I i I
,

I I
, ,

I I ,'I

I i
,

1f!1"'fI"l VI I I

,I

I I

v,

"

'"

, ,

'"

Web Thickness

In.

ALLOWABLE (COLLAPSING) SHEAR Fl.OWS FOR 24S-T WEBS WITH ROUND LIGHTENING HOLES HAVING FORMED 45 Fl.ANGES Fig. CIO.21

ANALYSIS

AND DESIGN OF FLIGHT VEHICLE

STRUCTURES

CIa. 19

Table CIa. 7
0.0.

L D.

0.0. 3.88 4.43 4.94 5.43 5.94 6.44 6.94 7.44 7.94 8.44
8.94

LD.
2.94 3.31 3.81 4.31 4.81 5.31 5.81 6.31 6.81 7.31 7. 81 8.31

1. 69
1. 81 1. 94

.81
.94 1. 06 1. 19 1. 31 1.44 1. 69
I

[9

.2'
.38 .38 .38

[9

.19
[9 [9 [9

reduced, the allowablef ~ill :8 ~educe1 jut no data is included for other ~eometries. ~s:n other cases, Arts. CIO.l~ ~nd C10.l~. t~e ~2:a 1s for pure 5h?ar o~ly. no nQ~al lo~js. ~2~ce loading st t rreners must be used (r ep Iac i ng the vertical bead locally) and. Ar-t i c Le D.3.a shou.Ir; be reviewed. Of course, in all three types ot f0~,~d webS, a Small &~ount of normal loadir~ can be tolerated, but no li~it is def1ned. Typical cases are wing ribs and fuselage frames where only relatively light no~l airloads are distributed by the rib Without loading stiffeners . In such cases some extra margin of safety above the shear load is usually used, depending UDon the designers jUdgment ~~d/or sUbstanti~ting element tests. ~here large distributed loads are involved, as in a beam supporting a ~uel cell, or Where the beam curvature is significant (See D3.8), stiffeners should be used for the lightest design. As a final note, whenever USing we8S with formed beads as in CIO.16 and CIO.lS, it is important trat the beads be formed with a length long enough to extend as close to the beam flanges as assembly will alloW. Short beads, ending well a'NaY from the flanges, will not develop the strength indicated by the al Lowab Les at ven in tne figures. Rivets attaching the 'Neb to the flange above ~ hole also need be nore closely spaced to take the higher "ne't" shearlocally .
ClO.20 Example Problems.

2.06 2.19 2.31 2.63 2.75 2.38 3.00 3. 13 3.38 3.63

.38 .38 .38 .38 . 38 .38 .38 .38

.25 , 1. 81 I .25 .25 1. 94

2.06 2.19 2.44 2.69

.2' .25 .25 .25

9.45

The solid li~es of F!g. CIO.23 give the ultimate strength, q, of the web as a function of web height, h. This Tepresents the total collapsi~~ strength of the web. The dotted lines indicate the shear flow, q. at wh1ch initial ouckllng begins. If the 0.0. is greater than .0 x ~, or if the spacing of holes is
Fig. CtO.23
CHANCE VOUGHT AIRCIAFT, INCORPORATED

REFERENCE: STRUCTUFlAL DESIGN MANUAL SECTION 4.230

~ N01'ZS: i 1. vuI

u. lor

. _ t .... Plnt....
0 D

2.

oll1Y.

qIlIIY_+::/'i- .G;
'i,'

BEAD SPAC!NC...:: I AND

+\? :J'+

1-- '.----1
0

For the beam shown 1n Fig. CIO.24. determine the web gages required for a) a beaded web, as in CIO.16. b) a 45 0 flanged hole web as in CIO.l:.
.~ ...J......o;...

"

c) a beaded {lange hole NeD with 1nter~ediate beads as in CIO.IS. sho'Nn in Fig. CIO.24 has the same 1imensions and design load 5ystem ~~ the bea~ shown in ~lg. CIO.14 and ~sed ~Qr the examp.Le problem in Part 1 of this chapter. i n ~~ese eX~TIple ~roblems. we are replac!r-g :~e web des~gn by t~e ~hree variations (al. (~) and (c). S:nce the webs cannot be cor.sid~red ful:y
be~~
exter~l

,
.051

The

..... i --,
. -_ _ I

,-~-r

~_+_
"

--- -1-_._
i

"

----:=t,: _ ...... ~~-r-"'-r-

2400#

2700-*1

u
h _ IIEIGHT

or

1'>UlL

~CIIES

T.
7-1/8": BayA."
i

DESIGN CURVES fOR CLAD 2024-T4 AND 7075-T6 SHEAR PANELS WITH CVC-3030 LIGHTENING HOlES AND STIFFENED BY MALE BEADS

..l.:..

BEADED SHEAR PANELS WITH LIGHTENING HOLES

1;750. 25"

+
,

Bay "8"

Bay"B"

:,
25" - - \ - - 25"

+
;
A

'l.i.

Bay "A"

::!
25"

37~

Fig. ClO.24

" ., 1~":'~~ ;'.<:: .~., , ;- ~:-

"'f~ : ~ .t;~.

:5 ~,;"\ :-;-:' '/' ":>" -.:" ~<./ .. :...~: .

----'1' :',-;

,~:-' ~.

-;,' _ '

"1:_. :,<,'.--{

~ :;:

;,,.

-~~~;.~

~"~ .~.~ j.,,~~, .... 1;-,.!'~ ,"~ -1f"...

....
CIO.20
DESIGN OF METAL BEAMS. WEB SHEAR RESISTANT NON-BUCKLING TYPE.

effective in resisting Jearn bending stresses, the flanges would be stressed slightly higher than found in the exanple problem of Part 1. USing the s~~e beam dimensions and beam design loading will provide a comparison of the web weights ror the various web cesigns. USing the simplified formula, the shear flows in bays A and 8 are
q~

f Se :: 13,800

Substituting values in equation (13)

qALL.= .79(.081) [13,800 [1-

(7S.~).] +

13,800

S.61 7.125

]~ 12.5

=5....::2 h

3750 7.125

= <::'26
v

'b /i n. _.

.064 [13,200 (.38) + 13,800 (.887)] (.521)

'-iB -

_ 3750 - 2400
7.125

189 lb./in.

:: 573 Ib./in.

Thus, for lightest weight, 2 web gages should be used, the lighter one for bays "8", ~d these could be spliced at the loading stiffeners introduCing the 2400 lb. loads. a) Beaded
~eb

Also, the net shear stresses an any through the holes must be less than Checking a vertical section through hole we get
rsNEl'

section Fsu' the

Gage Requirements.

~~terir~ Fig. CIO.IS with q = 526 Ib./in. and h = 7.125" we find that the minimum acceptable gage is .051". Thus the web of Bay A should be .051", giving ~llaw.= 770 lb./in., from Fig. CIO.lS.

(h-D)t

526(7.125) = 3750 (7.125-5.61)(.081) .123

=30,500
Thus, M.S. CO LL =
M

~~ -

=~
" 1

M.S.

770 526 - 1

=~

I '

'~lT:"T'
.n",,~

Repea~ing tor the web of Bay 8, enter with q = 189 lb./in. ~nd h = 7.125" and find the minimum acceptable gage is .032, which has qALL :; 310 Ib./in. Thus for Bay S,

-r-'0 ,0 -00 sNEr - ....

Fsu

_ 37,000

.21

Repeating with the above geometry and using t = .051 for Bay "E" where Q = IS9 Ib./in., we get
qALL.= .77(.OSl) [7900(.38) +8300(.887)] (.521)

310 M.S. = 189 - 1 = .64

oj
A 45 0 Fl~nge

Round Lightening Role.

.0393 [10,630](.521) = ~ lb./in.

Since the larger diameter (and more widely spaced) holes are more efficient weight"NiSe, try a hole navtng D ,... Bh or 0 = 5.51" from Table CIO.S. Soacing two holes tn the 25" bays 'Nould inc icare a spacing J = 12.0". Then, tram C10.19
C = b-D = 12.5-5.5 = 6.9

M S 217 1 2 ". OOLL. 189 -

.15 =

and
M.S., . ~ .""..

= 37,000
17,400

_ 1

1.12 =

a=
~l=

.19,
C -23

~r~m

Table CIO.6
(.19)

= 5.9-2

= '3.52

Now determine the a)lowable value of q for, say, t = .081" for Bay A.
:I t

The reader can repeat the above for smaller sizes of holes and thi~~~ess to see it a lighter arrangement (NET web weight) can be obtained. It can be seen that the example ga~es are contlr.ned by the data of Fig. ClO.2l extrapolating in the case of the .081 web.
0)

7.125 :: 38 .081
.0006 (68)

A Beaded Flange Hole with V-::rtical Beads.

Inte~ediate

= .8.5 n

= .80
=
38 and

Fr-om Fig. CI0 .20, ror .~

The geometry ~or t~is is as previously discussed (O.D. - .6h and b = h).

85

fa

:3,200

First determine gage for Say A. Entering Fig. CIO.23 with h = 7.125" and q = 526 lb./in., the closest acceptable gage is t = .064 (solld line), ~lving

ANALYSIS

!\NO DESIGN OF

FLIGHT VEHICLE STRUCTURES

CIO 21

qALL. = 700 lb./ln.

I1.S.

= 526 - 1 = .33

700

Repeating for Bay B with q ~ 189 and h = 7.125 ff the closest accdptab1e gage is t = .032.
qALL. = 208 lb./ln.
I1.S.

.1
3

I-~

r'- -

f-I6
24"

: 1a9 - 1 '" ;,,1

208

Actually, a hole with 0.0. = 4.43 (from Table ClO.?) could be used even though 0.0. = 4.43 = 62 is a little larger
h 7.125

~
3

~ 1-.0&4

than .6, since the M.S. above is well above zero. A comparison of the total weights of each of the 3 types of webs is given in Table ClO.S, using a hole diameter of 5.6l ff for webs (b) and 3.31" (from Table CIO.?) for webs (c) and a density 0: .10 lb./in. for aluminum.
Table CIO. 8
Type of Web (a)

'1' to,-

T
I.

3"

--ll,L

Fig. CIO.25

(2)
Total Web wi,

1-----,--'---,------1
Bays "A" (2)
1.82

Weight in Pounds

Bays "B" (2)

1.14

Design a butt web splice for the beam section of ClO.l3 for a design shear load V 3000 lb. and a bending moment of 50,000 in. IbS.

2.96 3.36
3.09

(b)
(e)

2.06

1.30
1. 03

2.06

Web of Fig. CIO.14 (with extra

Stiffeners)

5.82

Thus for a "non_distributed W load, web (a), a beaded web, is lightest for Bays WAn and (c) is lightest for Bays nSw. If naccessw :or lines is required then a web ot type (c) or (b) should be used for Bays "A",

Fig. CIO.26 shows a Simply supported beam carrYing a 2000 lb. design load located as shown. The cross-section ot the beam is shown in Fig. CIO.27. The design requrr-ement for the web is no Int t fa.L buCkling under the deSign external load. Check the given deSign for strength and modify it too weak or too strong, or in other words, ilnprove the design. Assume beam flanges are braced against lateral column failure.

The weight of the web and vertical ~II~----stiffeners designed for no initial buckling under combined bendir~ and shear gave a value ot 5.82 Ibs., or ~uch heaVier than the other designs.
PROBL;"'!'.S

\ Fig. CIO. 26

Web stiffeners 1/2 x 1/2 x 035 angle on one side of web.

(1)

Fig. ClO.25 shows the cross-section of a wing beam. Calculate the ultimate resisting moment for the beam section using the stress-strain curve for the extruded 248T material given in Fig. ClO.4. Use .008 unit strain at the extreme fiber of the upper flange. Compare the results With the reSisting moment given by the general beam for.nula M = fbI/yo

(4)

Re-destgn the beam of Problem (3) to USe the 3 types of stiffened web as presented in Part 2 ot this Chapter. Compare the web weights with the web weight required in Problem 3. Fig. ClO.28 shows the external dimensions of a tapered cantilever beam carrying the distributed deSign load as shown. The web is not to buckle under the design load.

(5)

CI0.22

DESIGN OF METAL BEAMS.

WEB SHEAR RESISTANT (NON-BUCKLlliG) TYPE.

r "Ti'
1--1-1/2"
6" "'" For Section .051 Properties of Tee, See
Section No. 22 of Table A3. 16 of Chapter A3.

--I g6

Make a complete structural design of beam showing size of all parts. For flange use 7075-T6 extrUsion material and 7075-T6 clad material for ~eb and web stiffeners. Show rivet design.
(5)

Same as Problem (4) but use web with vertical beads.

REFERENCES

(1)

Kuhn, Paul:- The Strength and Stiffness

at Shear Webs With and Without Lightening
Holes. NACA. ARR. June 1942.

Same as Upper Flange

Fig. CIa. 27

(2)

Kuhn, Paul:- The Strength and stittness of Shear Webs with Round Lightening Holes Having 45 0 Flanges.. NACA. ARR. L-323

Dec. 1942.
(3 )

BleIch, F.:- Buckling Strength at Metal Structures - Book - PubliSher, MCGraw-Hill

pp. 418-24.

~k====~-====ll. : 8 '0
-'t,Beam-

2 5"

96"

--

---I -I 2.

1.

5"

Fig. Cla.28

CHAPTER Cll

DIAGONAL SEMI-TENSION FIELD DESIGN PART 1. BEAMS WITH FLAT WEBS. PART 2. CURVED WEB SYSTEMS. PART 1
Cll.l Introduction.
T~e aerospace structures engineer Is constantly searching for types ot structures a~d ~ethods of structural analysis and design which will save structural weight and stIll ~rovlde a structure which Is satisfactory from a fabrication and :conomlc standpoint. The development of a structure in which buckling of the webs is permitted with the shear loads being carried by diagonal tension stresses in the web Is a striking example or the departure of the design of aerospace structures from the standard structural design methods in other fields of structures, such as beam deSign for bridges and buildings. The first study and research on this ~ew ~ype of structural design l~volving diagonal semi-tension field action in beam webs was by Wagner (Ret. 1), and because ot this tact, this type of beam design is ot~en referred to as a Wagner beam.

Cll.2 Elementary Approximate Explanation of Tens tonField Beam Action.

Fig. Cll.l shows a Single bay truss with double diagonal members (A) and (B) and carrying an external load P. The load P will cause a compressive load in ~ember (A) and a tensile load in member (B). It member (A) is quite flexible it will buckle as a long column as shown in Fig. ClI.Ib, when load P is relatively small, however, panel will not collapse. As the load P 1s increased, the member (A) cannot take any more load but it will practically hold its column buckling load as the bending deflection or bOWing gets larger. However, member (B) being in tension can take further load until it reaches its ultimate tensile strength. Thus any increase of the shear in the panel due to an increase of load P after diagonal (A) has buckled can be carried by a further increase of tension load in member
(B)

In :ha,ter CIO, ?art 1, the SUbject of beam deSign with shear resistant (non-buckling) flat webs was covered. This type of web design leads to a comparatively heavy weight, which fact prevents its wide use in aerospace structures. Part 2 of Chapter CIO dealt with webs stiffened by closely spaced beadS, flanged ligh~enlng holes, etc., a design which shows an ~provement relative to we~ weight over the flat sheet web with vertical stiffeners. However, a large proportion of sheet panels used in aerospace struc~ures is part of the external surface ~~d holes and deep beads in the surface skin cannot be ~er.nitted, thus continuous sheet is required and to save structural weight, semi-tension field action in the webs and surface sheet panels ~ust be permitted, which means a wrinkling type of structure. Since the original work by Wagner, much further study ~id testing of structures involvir~ semi-tension field deSign has been carried out by both industry and government agencies, hence a fairly accurate procedure for the design of such structures has been developed and this Chapter is concerned with a limited presentation of the principles involved and the design procedures that have been
deve Ioped ,

(.1

Fig.

en, i

(hI

Flange
W.eO

; ;

ft~fC
fc ft

;f'p:
~

Flange

Fig.

(.1

en. 2
(0)

tp

Fig. Cll.2 shows the same panel but with the ~NO diagonals replaoed by a flat sheet web. Under a small load (P), the web will not buckle and the stress picture on a small web element is shown in Fig. CII.2a. and fc = ft ~ f s where f s is the shear stress in the web at this particular point on the web. Now thin flat sheet 15 relatively weak under compression, thus when panel load P is increased, the compressive stress f c reaches the buCkling stress of the panel in the diagonal direction and the web buckles, however, the panel does not collapse as further inc~ease in load P can be handled by

c n.u

Cll. 2
~UTther

DIAGONAL SEMI-TENSION FIELD DESIGN

increase

i~

ft or diagonal tension in

the web sheet. The ~eb h~s an ability to hold the f c stress ttat c~us3d buckling but cannot increase i t . ?:!.g. Cll.2b shows the 'Neb stress
~i::ure after the loac P has been increased ccnSideraJly, thus increasing the d18gonal web t ens t Le stress as shown by tno length o r the vector fer ft. Since the shear load on the panel is transferred by diagonal tension in the web a~d since flat sheet is efficient in tenSion, this ~ethod of carrying the shear load oertat t s t.ne use of relatively thin 'Nebs because of the hi~h allowable design stresses in tension. Fig. Cll.3 shows a photograph ot a thin web bearn under load. Since the diagonal wrinkling appears severe, the external load being carried is no doubt approaching the failing pOint of ~he web. The student should realize that such a degree of 'Neb wr tnk'l mg dces not occur under nor.nal flying accelerations since the loads carried in ~or.nal flying conditions are cnly a f~action of the design loadS, and thus the bUCKling ane wrinkling is barely noticeable ~nder accelerations cf 1/2 gravity, Nhich nay be encountered otten in flying in gusty weather conditions.

rig. :~1.4 shows a cantilever beam with parallel chords a~d vertical sti~feners SUbjected to a Single shear load V at the free end. The dashed diagonal lines indicate the direction of the wrinkles as the thin ~heet buckles ~der the load 11. Assuming that the flange anglss develop all the bending resistance, the vertical and horizontal shearing stress is constant over the .veb and equals

- - ... - - ,- - ... - - ... - - - (1)

where
t is web thickness h is taken as the distance between centroids of flange rivets. V = vertical shear load. %

...

~':2'

ell. 3 Elementary Derivation of Approximate TensionField Beam Formulas.

~~ :\ :\ ~~" , " ~,,~ 0-"~

~~
~,,~
......... ...
-,
,

-, ,"-:: I~'" ~:: ~:::: 0"-'" ' ;;:~ !~,~

- -

(See Fig. Cll.5lJ-t:::' ...... ... ...... ......

~ ~

,
--~-~

\~"./x

--...... -... !~,,~

!
I

In order to give ~he student a general p i c tur'e of t.he tnr Luence of Neb tension field action upon the beam component ~arts, an ~lementary apprOXimation of the beam equations :;:ill be given.

-"" l-d+d+d-j-d+ d~ ~

<, ,~ ,~~' "-~ "" -, < -,-;~ " , .,a;~~ ....... ..

T -,0;" ,,'0 ~,,~! ~~.

~ \;"

~':

~~~-,j

Fig.

cu.s

Fig. Cll.3 Loaded Cantilever Beam Showing Severe Diagonal Web Wrinkling (Ref. 3)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE

STRUCTURES

en. 3

Fig. ell.S shows the free body diagram of a ssall tri~ngula" segnent of the web cut from the upper ~ortion of the beam.

Rivet Loads The sheet element is held to the flange angles by the rivets along line (AS). The rivets are subjected to two loads, one parallel to AB and one normal to AB; and each force equalS f s tdx. The reSUltant rivet load therefore equals Vl2 f s tdx; and if dX is taken as one inch, the resultant load will be ylf f s t. But f s = V/ht, hence Rivet load/inch web Stiffener Load
~

Fig.

cn.s

1.41 V;fi

_____ (6)

From elementary mec~nics the horizontal and vertical shearing stresses produce compressive and tensile stresses on planes at 45 wt th the shearing p lanes , The web, being very thin, can carry very little compression stress on the surface (AC) of Fig. Cll.5 before buckling; thus this small stresS which produces buckling on AC will be neglected or fc = O. The edge Be is subjected to tensile stresses which the ~heet can carry effectively. The forces acting on the sheet element are shown in Fig. Cll.5. For equilibrium of the element: ZFx = 0 or -f s tdX
whence
__________ - - (2)
-+-

has

The tendency of the web is Fig. ClI.4 which broken down into the tension bandS, is to pull the flanges together; this action is prevented by the vertical stiffeners which keep the flanges apart. Thus if a pure tension field is assumed, the axial compressive load P s in the stiffeners from Fig. ell.S equals the vertical component of the web tensile stresses over a distance (d), the stiffener spacing. whence
Ps = P sin 9

tdx ft - - .

v2

-= 0 .,f2

But
P
Fig. Cl1.6

hence
1 Pa =_1_ - ! t dt .,f2 .f2

= 't

dt/2

or the web tensile stress equals twice the web shearing stress:
From (1)
fs

but !t whence

= v/nt
______ - - (3)

=2

f s and f s

= V/ht
(7 )

whence ft ,. 2V/ht Lik9wise for eqUilibrium: ZFy = 0 or f v tdx - ft wnence ft :-tence

Stiffener load Ps = Vdln (compression) Flange

~~lal

Loads

v:2

tdx

.,f2

Fig. Cll.7 shows a tree body or the port ions of the beam to the right of a section a distance x from the end of the beam.
(4)

=2

fv

_________ - - - -

r, = f s

_________ - - - - -

(5)

Let Mx = external bending moment at Section AB. For eqUilibrium the internal ~esisting moment on Section AS must equal external bending moment Mx ' Taking moments about point B; h' ZI1B = i'1x- Ft h ' _ ft cos 45 hi t . cos 45 '2 = 0

______ (8)

err. 4

DIAGONAL SEMI-TENSION FIELD DESIGN

,1

M V Ft =V-Zcot a
_ M V

- - - - - - - - - (11)

Axial load in the compression flange: FC-h'!-""2cota

- - - - - - - - (12)

Axial force in stiffeners:

FSt if f.

=-

% tan a

- - - - - - - - (13)

In the above equations:

Fig. CIl.7

v = applied shear load h = distance between cent~oids of flangeweb rivets h ' = distance between centroids of flange sections t web thickness d = vertical web stiffener spacing a. = angle of web buckle (see Fig. Cll.7)

where h' = distance between f'lange centroids.

But
:LC

= 2

s and !'S = F/h't, hence ft

ft

V =2 h't

Buos t

Ltut tng this value of

"x ""

in Sq. (8)

l't n

,.,."

Vb' 2

o
- - - -(3a)

whence

In deriving equations (10) to (13) Wagner assumed that the beam flanges were infinitely stiff in bendIng. Actually the flanges due to the lateral pull of the web tension field will act somewhat as a continuous beam over the web stiffeners as supports, as illustrated in Fig. CII.8a. The deflections of the beam flanges relieves the web stress tn the midportion of the panels and concentrates the web stress near the stiffeners where deflectton of the flange Is prevented (see FIg. CII.Bb).

Then

~o ~ke

ZH = 0 on Section AB:

~c=hT+2

M x

(9 )

Thus the compressive flange axial load due

to bending is increased by a value equal to V/2

Ibs. ar.d the tension flange load due to bending is dec~eased by Vl2 due to the horizontal cam~onents of the Neb tension field.
Cil. 4 General Wagner Equations for Tension Field Beams.

Fig.

en. Sa

Fig.

cu.ss

Wagner (Ref. 2, Part III) has developed a correction factor R to take care of this web stress concentration due to flange deflection. This stress ratio factor is obtained from Fig. Cll.9 and the following equations:

T~e appr8xinate elementary derivation g~ven in the previous article was for the purpose of' wd = 1.25 d sin u (I +\!...)h - - - (14) giving the student a general idea of the u influence of a complete tension field beam on d = stiffener spacing the various beam stresses. The angle a is in general not 450 jut depends on ~lange areas, I u and I L = moment of inertia of upper and lower beam flanges about their own neutral ~xls. beam height, stiffener spacing, etc.

The general equations derived by (Ref. 1) are as follows:

~agner

wagner also has derived the :ullowing equations for deter.nining the web buckle angle a as follows:
s in" a = / a JI
+

For beams Nith infinitely Tigid and parallel !'langes with vertical web stiffeners: Diagonal tensile stress in web:
ft =

a _ a - - - - - - - - (15)

ht sin 2 a.

2V

- - - - - - -(10)

+ AL dt ht As - Au + At

Au

ht

- - - - - - - - - - (16)

Axial load in tension flange:

ANALYSIS AND DESIGN OF FLIGHT VEHICLE

STRUCTURES

cu.s

" ""'"X"_:_::\,,,,<_: :-::--,- --

--

_.__..

--_.

----

-',,,~::- -~-

:,0:"- ",:::_-"

:"::s: "'\;- :- ,_C- "-'C-"_: :"::- ~"

(3) The remainder of the beam shear after SUbtracting th~t carried by (1) and (2) is considered carried by'the web in a buckled state in the form or a diagonal tension field. The above assumptions apply to beams with parallel flange members. If the beam has sloping flanges, part at the beam shear will be carried by the shear component of the flange axial loadS, and thus the assumptions should be applied to the net beam shear.
CIL 6 Shear Carried by the Beam Flanges.

The general flexural shear flow equation is

q

= V \YdA = v;<

(it !ydA 15 gt ven symbol q)

4 wd radians

Au and AL = area of upper and lower flanges including reinforcing plates but neglecting web material. AS = area of stiffener METHOD 1
Cil. 5 Modified Wagner Equations for Use in Design.

For beam flange and web arrangements commonly used in aircraft structures the shear stresses are apprOXimately constant over the web, that is, between the centroids of the flange-web rivets. Using this assumption, the value Q in the above equation equals the static moment of the flange area about the neutral axis of the beam, and I equalS the moment ot inertia of the flange area about neutral axis. (web area is neglected). The shear load resisted by the web alone therefore equals

_ vqn V Vw - -1- = I/Wl - - - - - - - - - - where

(17)

The Wagner equations are too conservative for design, particularly in the web stresses and the loads in the vertical stiffeners. The shear stress carried by the web before it buckles is in many cases an appreciable part of the total resistance, as the buckled sheet normal to the di~gonal compresSive stresses has the ability to hold this buckling shear lead after 'Nrlnkling. The load in the vertical stiffener is too conservative, since the stiffener is riveted to the web and the web is riveted to the flange; thus the web a:ts as a large gusset plate instead of a pin end condition as assumed in the ~agner equation. The following general ~ethod of analysis of Wagner bea~s has been used by ~any airplane companies. Instead of assuming that the entire beam shear is taken by the web in diagonal tenSion, the following assumptions are made relative to the resistance for carrying the be~ ver~ical shear.

h = effective web depth = distance between centroids of flange-web connection rivets. The total beam shear V which equals the resistance of both web and flange, equals
V

w Qil -- --

(18)

The difference between (17) and (18) gives the shear carried by ilanges.
Cll.7 Shear Load Carried by Web.

Up to the buckling stress of the web plate, the shear flow is assumed to be of constant intensi~Y over the effective web depth. when the web bUCkles, it is assumed that the web maintains the diagonal critical compressive stress that produced the bUCkling of the plate. For further increase of shear load on the web, the entire resistance is ~rovlded by the increase in diagonal tensile stresses with no increase in (1) The shear strength or ths beam flanges the diagonal compressive stresses. In other words, for loads above the web bUCkling point, is not neglected. the web acts as a pure tension field beam. Fig. (2) The shear carried by the web before it ell.lO illustrates these assumptions. buckles, th~t 15, as a shear resistant nember is Shear Load Car~ied by ~eb at ~eb BUCkling Stress cansidered as ~n effective resistanc~ and not Fig. (CII.lOa) shows the shear resistance neglected.

CIL 6

DIAGONAL SEMI-TENSION FIELD DESIGN

(c)

stress of equation (21) must be multiplied by a rivet correction factor lIK r to obtain the
true fSax.

Hence for riveted connections:

!t(max.)
Vcr, Load
up to web buckling
Vt taken by web after buckHng

= (E!.. s

Fs

- - - - - cr )....L Kr

(23)

Vw '" Total

whence
-----(24)

load on web '" Vcr + Vt

Fig. Cll.IO

where

on the web face (aa) when the web is shear resistant. The vertical shear stresses on (aa) have been replaced by the diagonal compressive and tensile stresses. The intensity of these diagonal stresses equals the intensity of the critical shearing stress Fs ' . cr Hence the shear load' carried by the web at the web buckling strass equals Vcr = F Scr h t
- - - - (19)

K = rivet spacing - rivet diameter r rivet spaCing

Squation (22) would apply for webs spat welded to flange members. The vertical components at the total ft an the web effective depth (h) would thus equal the shear load Vt developed by the web after buckling.
st~esses

The critical shear buckling stress is given by the following equation from C5.4 of Chapter C5.
(20)

For s~ot welded connections:

fl~nge,

web and

stif~enArs

Vt = (ttmax - F'Scr) Rht s tn a cas 11 - (25)

where k s is a function of the aspect ratio alb of the shear panel and of the edge c ondi t i ons Fig. C5.11 of Chapter C5 gives the value of k s for edges simply restrained. Taking a vajue of ks tor this edge condition is no doubt slightly conservative. Shear Load Carried by Web atter Buckling Fig. CIl.lOb shows the web stress distribution that is assumed to be subjected to the web when the web shear load is increased above that which ca~sed the web to break down into a tenet en r ie ic . The diagonal tensile stT8SS ft tends to pull the beam flanges together, and thus to bend the fl~nges. The diagonal tensile stress ft ~cr the shear resistant~eb does not produce such action. To obtain the ~imum combined tensile stress in the web, the stress tt must be multiplied by a concentration factor llR trom Fig. CIl.9. Hence ft(max.) =
So.Lvtng rcr f
t

For riveted connections: Fs c ....

-

K " ) Kr

Rht

sm a cos a
- - (26)

It t r _ -nax , is taken as the tensile yield

paint stress Fty or the ultimate tensile stress Ftu equations (25) and (26) Will give the shear load carried by the web above the buckling load when the web i s stressed to the yield and ultl~te tensile strengths respectively. Thus for yield strength:
Vty

= (""' SeT) rty-~

Kr- R ht at ' n a cos a.

(riveted connect.)
Vt y

- (27)

(Fty-FScr) Rht s In c cos c (23) (Spot we Ided connect. )

(E.! rscr ) R

- - - - - - - (21)

For ultimate strength: Fs cr Vtu = (Ft,J.- K ) K.... R ht sin a cos 11 r " (riveted connec t . ) - (29) (30) Ytu = (Ftu-Fscr) Rht s m c cos c (s~ot welded connect.) The total yield shear resistance or the web equals

- - - - - -

(22)

If the Neb is ~iveted to the ~langes and web stiffeners, part of the web material 1s cut away due to the rivet holes thus the tensile

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUC

RE

Cll.7

V W ty Total

= (Vcr

Vty) web
shea~

(31) resistance:

ulti~te

v,Ntu' (Vcr

Vtu)

- - - - - - - - - (32)
Fig. A

Total beam shear resistance equals equations (31) or (32) multiplied by

~ (Ref. Equation 13)

Hence
1

err. 9 Rivet Loads. The loads on the rivets connecting the flanges to the web consists of two parts, (1) that due to the web acting. as a shear resistant member subjected to stresses which cause web buckling~ and (2) that due to the web tension field for web stresses above the web buckling streSs. Rivet Load at Web Buckling Stress
_V",cr,-;-IF:...

= WI (Vcr ~

(33)
(34)

= Q.~
30)
rOT

(V,r +

Substituting values from equat18ns (27 to for Vty and Vtu, spot welded connections:

h I

- - - - - - - -

(39)

where
It r, "QLFscr
+ (F t u -F scr

11

) R sin a cos

a]
(36 )

= load on rivet parallel to flange in

of inertia of flanges about beam neutral axis = moment of inertia of total beam section = distance between flange rivet centroids : buckling shear strength of web, lbs.

= moment

lb./in.

For riveted connections:

Vy i l d

It[ ,< F

Scr

(Fty -

Scr FK ) Kr R sin a cos a

r

----(37)

ult.

K R sin acos
-

aJ
(38)

Rivet Loads tor Tension Field Action In Fig. Cll.ll the web element (abc) is attached to the flange along line (ab). A vertical depth (bc) of 1 inch has been taken. The shearing stress on this len~n due to Vtu equals Vtu/h. This load represents the vertical component of the tension field, hence
- - - - - - - (40)

CII. 8 Beams with Parallel Flanges but with Oblique Web Stiffeners.

wagner has developed equations for the condition where stiffeners are placed at an ang19 with the parallel flange me~bers, (Fig.A). In this case equations (27) to (30) should be ~ultiplted by a correction factor equal to (1 - tan a cotjJ) where fJ is the upright angle. For oblique stiffeners the wrinkle field angle a is equal to jJI2 and the concentration factor R c~n be taken as unity. Hence the equations for Vty and Vtu become as follows: l-cos/l - -(36a) Vty = 1/2 (Fty - K~ ) Kr h t SinjJ Vtu = 1/2 (Ftu -;t-) K,. rFS.. . r
h t

Resolving the tension load Pt into x and y components on rivet line (ab) and dividing by the length ab to get rivet loads per inch, we obtain
Py

F'S . . r

=
=

"tu sin

0. sin a hsinacosa

=""h t an
= Vtu
h

Ytu

--(41)

1 - cos

/l

Px

Vtu cos a sin a h sin a cos a

(42)

sinft

- -(36b)

(For spot welded JOints, omit term KrJ

Combining the three component rivet forces as given in equations 39, 42 and 41 to obtain resultant load R on rivet:-

err.a

DIAGONAL SEMI TENSION FIELD DESIGN

r~n
a
Fig. ClI. 11

equation which neglects the resistance of the web in bending before bUCkling.

l"=-r
h
I
tan a ).] ./.
- - - (43)

- + My fb - - IF

- - - - - - - - - - - - -

(45)

-L--~C~
Pr-

The total flange loads can be calculated by equating the internal reSisting cou~le to the external reSisting couple to the external bending moment, or
M Fc = Ft + - hi
- - - - - - - - - - - (46)

where Fc and Ft = total compreSSive and tensile flange load respectively

[(V';:' IIF + V~U),+ (V~u

h'

= distance between flange

centroids

Cll.10 Flange Loads.

pri~ry

The axial flange loads are due to two causes, namely

Equation (46) neglects the resistance of the web in bending. Flange Axial Stresses Due to Web Tension Field Due to the horizontal components of the diagonal tens ton fIeld each flange Is subjected to a compressive load Jqual to
Fe : Ft

(1) Stresses due to primary bending of the beam by the usual flexural theory. (2) Additional stresses produced by the web tension field.
In addition to these two primary effects, secondary bending stresses due to the bending of the flange because of the tension field are prOduced as illustrated ~n Fig. ell.B. Stresses for Primary Bending:
fb

= - 2"" eot

Vt

- - - - - - - -

(47)

(Reference see Equation (12) and general derivation when a = 450 see Equation (8) and (9).
In Equation (47) Vt = shear load carried by tension field action.
seconda~J

+ Mer Y + (l'! - Mer)Y - -1-IF

(44)

Bending Stresses

Where
I : moment of inertia of total section

IF

web about neutral axis = moment of inertia of section without web about neutral axis
includir~

For estimating the secondary bending moments on flanges due to lateral pUll of web tension field, the flange can be treated as a continuous beam with spans equal to the stiffener spaCing. The component or the web diagonal tensile stresses normal to flange. Vt = 11 tan a (pounds per inch) - - where V t = shear carried by web in diagonal tension For a continuous beam ot equal spans, the moment over the supports = 1/12 Wv ct a , where d equals the stiffener spacing. The detlection ot the :langes tends to relieve the pull in the center portion between stiffeners which then decreases the continuity moment over the support. ~agner (Ref. 1) gives a relieving factor C (See
(48)

Mer = bending moment for load which causes web buckling

= total bending moment on section

The first term in the above equation gives the bending stresses at the paint where the web breaks down into a tension field. The web is thus effective in computing the moment of inertia I. The second term in the equation gives the bending stresses when the beam web acts as a tension field web, or in other words the buckled web is ass~~ed ineffective in bending. To be Slightly conservative the bending stresses can be computed by the following

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

en, 9

Fig. CIl.9) for use ','lith the bending moment expression. Therefore, the secondary bending moment on flanges is
!'Isec

stiffener design, the NACA method be used. The method above, proposed originally by Wagner, can be used for quick prelimina~J design at the web stiffener.
Cl1.12 Beams with Non-Parallel Flanges.

1/12 C

Vt

tan u - - - - - - (49)

Cll.ll Loads in Web Vertical Stiffeners.

The following method of Checking the strength of vertical web stiffeners was originally proposed by Wagner (RefS. 1, 2) and can be considered conservative. The axial column load in the web sti rrenerequals

In many aircraft beams the flanges are nat parallel but have a slight taper. For this case equations (46) and (47) give only the horizontal components of the flange loads. The total flange farces and their vertical components can be computed from the horizontal components and the slope at the rtenges , Thus from Fig. CILl2a SUbtracting the vert1cal components of the flange forces we obtain the net web shear load V wn ,
- (51)

= -

- -

(50)

where net web shear for beam with nonparallel flanges

V,N

This is the same as equation (13) except V is replaced by Vt, the shear carried by the Neb in diagonal tension field action. Under this load Wagner considers the stiffeners as columns with elastic supports as the Neb tension restrains the struts trom buckling out of the plane of the web. ~agner's calculations yield a reduction factor Cg which the actual length of the stiffener is multiplied by to obtain a reduced lengt h L I , or

= web shear for parallel flanges

L =
1

C~L

- - - - - - - - - - - - - -(5Qa)

With this reduced length L 1 , and the actual cross section of the strut, the column strength can be calculated by the methods of Cr~pters 81 and 85. Fig. Cll.12 by wagner shows a plot of the reduction factor C which is a function of the parameter as shown. Since the Nark of Wagner, the NACA has conducted an extensive research program on the strength and deSign of sem1-tension field beams and the strength design of the web stiffeners has been based on a ~are rational basis. The NACA nethod is given later in thiS c~pter and it is recommended t~t for final check or web

F----

Fe. to" Be

Fig.

en, 12a

Cll.13 Example Problem Using Method 1.

Fig. Cll.13 shows a cant1lever beam or constant cross-section carrying a 13500 lb. lead at the free end. The beam will be strength checked tor the ~iven load. The mater1al properties are:
Web:

2024-T3 alum. sheet. Ftu = 54000, Fty = 42000, E 10,500,000.

Flanges:
Ftu

= 10,300,000.

= 78000, Fey =

7075-T6 alum. alloy extruded.

70000,

Investigation of Web Strength Web panel size be~Neen web stiffeners and centroid of flange rivets = 10 by 28.56. Aspect ratio or panel 2.855.
Fig. Cll.12

= alb

= 28.56/10

The critical buck11ng shear stress is given by Eq. C5.4 tram Chapter CS;-

Cll.10

DIAGONAL SEMI-TENSION

FIELD DESIGN

(A)

The angle of the diagonal tension r t eId with horizontal is

The ratio of oJtained gives k s

F

2.856 and simply supported edges !s

buckling coeff!cient k s for an alb

(see Eq. 15)

where
1 + Au

from Fig. Co.ll of Chapter C5 which = 5.8. SUbstituting in (Al,

s-:::r

= n x5.8xIO,500,000 (.025)2 =:342 P5i

12 (1 _ .3 2
)

ht
+ AL

10

"AS- Au
Therefore shear load resistance developed by web u~ to buckling stress equals
Vcr=: FScr h t

dt

ht
+ AL

(see Sq. 16)

Substituting
1 + 28.56
lOx .025

= 342x28.56 x

.025 = 244 lb.

.675 + .378
ZB.56

x .023

The shear strength of the web acting as a tension field after buckling is given by the following equations, when stressed to the yield paint in tension

.23

x .025 .675 + .378

hance sin 2 a .::

FScr (Fty - - - ) Kr Rht sin a cos a.

J 4.0Sa
a

+ 4.05 - 4.05

= .4.7
1"

K r

sin a = .685,
i~

= 43,

cos

0 ~

when stressed to ultimate stress

F s ........ Ytu :: (Ftu -~) Kr r

tension

Rh t

s tn CL cos a (See Eq. 29)

To determine the web stress concentration factor R, we solve for term wd and use ~urve or Fig. CIL9.

wd = 1.25 d sin a j/(Iu +tIL)h

wd

(Ref.Eq.l4.)

I. I
o

50" 2ll7T3 Rivets 2 Rows - 5/32"

13500#

1.25x10x .685

3/4"

= 2.4.3

.025 :;..,(-.;cl0:;;7"5,....,.=;~"'-..,;-;~., + .0291) 28.56

~

From curve Fig. Cll.9, R

(d)

.86

Since the web is riveted to the flange, correction must also be made for net area of web.

15/2

Jc

1/ 8

-.

K .:: rivet spacing-rivet di~meter r rivet spacing

7"'

.75-.156 .75

=
For 2024-T3 material, Fty = 42000 and ttu : 64000 ps t ,
Vt y ':: (42000 NA

.79

Lower Flange
Fig. Web Negleered
~=;"

cu.

7075-T6 Extrusion 13 Web Effective

10.94

'

--h-1- -'- --NA

,~~~--",
' !

---r' .17.42

J.:T
.l.

:~) .79 x .86 x28.56 x .025

19.06 Total Area > 1. 053 in. INA'" 210 in.~ Q '" 7.15

J..Jb

'i-'l.

x.685x.73

Vty = (41565) .242.:: 10100 lb.

Vtu

= (64.000

:::;;1

.242

= 154.00

lb.

Total Area e 1. 79 in. a INA'" 270. 5 tn.'

Tae total shear load lncl~ding strength ot flanges is given by equations (33) and (34),
nameIy

Betore solving these equations, the terms 0, R, and K r must be determined:

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

CIL 11

shear load carried by the tensile field under the given shear load of 13,500 lb. From Fig. Cll.13
13500 = 210 (244 + Vt) 7.15 x 28.56

IN ,. A. of section without web

Q of flange aJout N.A.=7.15

= 210

in,'

13500

whence
210 (244+ 10,100) 7.15 x 28.56

Yt = 12,850

10,670 lb.

equatIon (B) tor Pr

Substitute in Equation(B)

rs., which represents Vt'l in

-r .ut

210
.

7.15 x 28.56
ultl~te

The design hence

~.S.

shear load

= 13,500

P,.= [ ( 244 ~+ 12850). + 12850 x .933']'/' 28.56 270.5 28.56 28.56

lb.
Pr

[(6+4.50)'1 +420

a ]J./a

=.

624 lb./per

inch.

= (16,070/13,500)

- 1

= .19

Load per rivet pitch of .75 inch

624 = 468 lb.

.75 x

The applied shear load uSing a factor of safety of 1.3 is 13.500/1.5 = 9000 lb. The yield strength is 10,670 lb. Thus for margin of safety of yield strength over applied load, we have (10.670/9000) - 1 = .18. The results show tr~t the beam shear resistance is proportioned as follows for a total ultimate strength of 16,070 lb., (244/16,070 x 100 = 1.5% carried by web in pure Shear, and 15,400/16,070) 100 = 96.0% carried by web as a diagonal tension field. or 2.5% is carried by the shear strength of the flanges. This value is relatively low for this particular beam, as in general the flange may provide considerably more of the resistance. Due to the thin web thickness and the large web panels, the web buckles at a relatively low stress, thus the ~ercent of the external shear load carried by the web at the bu:~ling stress is qUite small.
T~e re~inder

r i vet = 596 x .86 = 512 lb.

Single shear strength or 5/32 - 2117-T3 (See Chapter D1). Bearing strength on 2024-T3 sheet.: 1.24 x

392 486 lb. (See Chapter D1).

Bearing is critical and the rivet strength per rivet pt tcn is 2 x 486 = 972 lb. (2 rows or rivets) Margin of Safety = (972/468) - 1 = 1.08
Due to large X.S., the r1 vet diameter could be made 1/8 inch, or 5/32 rivets could be s paced farther apart.
CHECK OF FLANGE STREr-GTH

Check

of

Rivet Attachment -

',veb to

Fhnge

The .025 web is attached to the flange member by 2 rows of 5/32 diameter 2117-T3 rivets at 3/4 inch spacing.
7he resultant load on the rivets per inch

The end bay of the cantilever beam involves speCial deSign considerations such as the two beam fittings and the special end stiffener. Therefore, the basic flange areas will be checked at Section A-A, which is 50 inches from th~ load ooint. which is where t~e end fitting woula possibly begin to take load out ot the flange :nembers. DeSign bending moment at beam Section A-A
15,

of

~lange

is given by equation (43),

+

na~ely

Pr -

_[VcrIF' (h I

50 x 13,500 = 675,000 in. lb.

Ytu _,_,a +
h)

'h

f_"_'"

v-;

tan
-

a)

,/.
--(8)

The web is not stressed to its ultimate tensile stress since we ~ave 4% margin of s~fety. ~e Nill therefore solve equation (AJ for Vtu using the given external shear load 1[:; 13,500 lb., and call this value Vt. th"i!

Since Y . . ~ = 244 Lb , , which ts the shear load the webv~arries without buckling, the bending moment cue to a shear ot 244 lb. will be resisted by the entire cross-section including the web. Above this value the ~emaining moment is resisted by tne section with the web neglected in computing the moment of inertia.

Cll.12

DIAGONAL SEMI-TENSION FIELD DESIGN

Thus bending moment at web buckling load = 244 x 50 : 12200 in. lb. Bending moment for tension field beam = 675,000 - 12200 = 56280b in. lb. Bending Stresses Upper Flange Bending Stresses:

Average stress:tC(av.)

= -33900 -10170 = -44070

F1~ge

PS1.

Lower

- Extreme Fiber

tt = 60980

18180

42800 pst.

Average ft : 60500 - 18180 = 42320 pSi. Since the tension field actIon tendS to decrease the tension load in the lower flange, it is good practice not to include the entire relieving etfect as tension tield action 1S not an exact theory. Flange Secondary Bending Stresses Since the web Is in a diagonal tension condition, it pulls on the flange members, or in other words, each flange acts as a continuous beam with the web stitteners as the support points and the transverse load on the flange is equal to the vertical component ot the web diagonal tenSile stresses. This secondary bending mament is approximated by equation (49). Msec.
Vt = 1/12 C 11 d

to = -

12200 x 12.58 270.5

662,800 x 10.94 210

(See Eq. 44)

= - 567 - 34600 = - 35167

ps t

If the entire bending moment were assumed resisted by the flanges alone, then,

rb

x 10.94 =- 675000 210

::I

35200 p sl.

Lower Flange Bending Stresses

f

o-

_ - 12200 x - 17.42 + - 662800 x 19.06 270.5 210

783 + 60200 = 60980 pSi.

tan a

(Note:

Section properties were computed without taking out one 5/32 rivet hole in vertical leg of flange tee, thus stresses are slightly ~iconservative. To be on the safe side, the net section properties should be used in figuring stresses)

From Fig. Cl1.9, C = .925 when wd whence

Msec

= 2.43
3230 In.lb.

. 12850 /2 x.9 25 X"2"8':56xlO = 11 x .933 -

Secondary Bending Stresses on Upper Flange Average Axial Flange Loads Due to Bending

IN.A, upper flange = .1075 (See Flg. C11.13)

Total Flange Load =

h~ = 6;95.~0 = 22900 lb.

Upper Flange t c (aver-, ) = -22900/.675

-33900 pe i .

I/y lower flber = .1075/1.16 = .0925 I/y upper flber = .1075/.338 = .318 tb(lower tlber)

= 3230/.0925

=34900

psl

Lower Flange ft(aver.) = 22900/.378 =

50500 PS1.

CompreSSion as flange bends over the support.

fb(

Flange Axial Loads Due to Tension Field Action The axial load in each flange due to diagonal tension in web equals
Ft

upp

er 'lb
J.

er

= 3230/.318 = 10140 (tension)

psl.

The combined stress on the lower tiber then would be -34900 - 44070 = -78970 pSi. The combined stress on the upper tiber
would be 10140 ~5337

= FC a -.5 'it cot a (See Eq. 47) = -.5 x 12850 x 1.07 ~ -6860

= - 35197 pSl.

Average stress on upper flange f c = - 5860/.675 = -10,170 pSl. Lower flange, ft = -6860/.378 = -18180 psi. Combined Flange Axial Stresses

Secondary Bending Stresses on Lower Flange I/Ylower tlb _ er = .0291/.217 = .134 (from Fig. 011.13) I/Yupper tiber: .0291/.939 = .031 f O(upper fiber)

= 3230/ .031 = lQ4(JOO

pSl

t c .:: -35157 - 10770 =

Upper Flange - Extreme Fiber -4.5337 ps t .

(Compression) tb (lower fiber) = 3230/.134 : 24100 pSi tension

ASALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

Cll.13

The combined stresses on lower flange are:f(upper fiber) = 42320 -104000

f (lowe~ ft ber ) == 42800 + 24100

=-c1880
= 56900

pet

pat

Calc~l~tion of Failing Compressive Stress for upper Flange.

however, the 0.5 inch increase in depth at the Tee would raise the ~oment of inertia at the Tee considerably. (2) The strength of the vertical leg of the Tee can be improved by adding a lip. This addition will likewise increase the flange moment of inertia, which Is needed to decrease the stresses due to the secondary bending action. Fig. (A) shows a re-design or the flange section with the same area as betore, thus no weight is added. The Tee has been made 0.5 inch more in height and an 0.5 inch wide lip has been added to bottom of Tee. The thickness and bulb size has been reduced to cancel the area added. The new section properties are given in Fig. A.
1

It will be asauned that lateral column actlJn is prevented by lateral bracing fram adjaCent structure, thus failure of flange wtll be by local crippling. The crippling stress for-the flange as a~ whole wtll be calculated by the Gerard method as given in Chapter C7. The bulbs on the Tee Section (see Fig. Cll.l3) will be assumed as only partially effective in producing a sim,?le support to the adjacent plate. The g number for the Tee Section will be taken as 4 which is n~ doubt slightly conservative. The design curve of Fig. C7.7 of Chapter C7 will be used.

gtr

(!J:::)'I' =
Ec

0.675 ( 70,000 )'1' = -72 4x .l.562 a 10,300,000 .0

Fig. A

From Fig. C7.7 we read Fcs/F cy = .90, hence Fes .9 x 70000 == ~ psi. compression.

"--,1.

I--t--l fila
Area:lll .675 INA'" O. 35

Without the effect of secondary bending of the average axial stress on upper flange was previously calculated to be -44070 psi. Thus margin of safety would be (63000/44070) - 1 = .43. However, the upper f!~~ge must also carry the stresses due to the flange acting as a continuous beam with the web vertical stifreners as the support points. This bending moment over the support points was previously calculated to be 3230 inch Ibs. The bending moment midway bet'....een the support paints would be about 50 percent or this value since lateral deflection of flange produces a relieving effect. The previous calculations gave the combIned stress on upper tiber as -35197 p~i, and on the lower fiber as -78970 psi. This compressive stress is above the Fey value of 70000 for 7075-T6 material and thus the vertical leg of the Tee is obviOUSly weak. The permissible average stress on the vertical leg can be calculated by considering it part or an equal angle section and computing the crippling, stress by the Needham method or Chapter C7. The result would be -51500 psi, thus with a stress of -78970 existing weakness 1s indicated and re-design is necessary.
fl~~ges,

1/y upper Uber '" . 584 I/y lower fiber = . 25

The secondary bending stresses w1l1 now be recalculated.

_ 3230

fb(upper) -

.25

=-

12900

b(lol'ler)

= 3230 = 5530
.584

The combined stress on the lower fiber is -44070 - 12900 = -56970 psi as against the previous value of -78970 before Tee was modified. Adding the lip on the bottom of the Tee would raise the allowable stress above the -51500 previously calculated, thus stress of -56970 can now be carried. This example problem has brought out the fact t:1at the secondary bending stresses can be a major stress factor unless proper attention is given to designing the flange to reduce the secondary bending stresses. Another approach to calculating or checking the strength of the flange would be to use the inter-action equation,

The question now arises, what changes can be made without adding any ....eight to the present flange. Two obvious faults exist in the flange design. (1) For a beam 30 inches deep, increasing the depth of the flange by a small amount such as 0.5 inch would not effect appreciably the over-all beam bending strength~

He

+ Rb

= 1

Where Rc is equal to the flange load due to beam

CH.l4

DIAGONAL

SEMI~TENSION

FIELD DESIGN

bending and web diagonal tension effect, divided by the crippling stress Fcs of the flange section. is the ultimate bending moment that the flange can develop as per the method presented in Chapter C3, divided by the design secondary bending moment.
Rb

From Fig. C7.S of Chapter C7, Fcs/J FcyE ::: .07. Fcs = .07'/53000xIO,700,OOO = 52500 ps i . using Johnson-Euler equation of
07,
Fe
Ch~pter

= Fcs - 4~~E (L'/p).Ol

52.500 ( 52500 - 4nJ 10,700,000 40.5) = 41800 psi.

Strength Check Lower

Flar~e

(~ension

Flange)

As previously calculated, the stress on bottom fiber 'NaS 66900 pSi tension and -61880 psi compression on the upper tiber at the lower flange. The tension is O.K. since the Ftu of the material is 78000 psi which should give enough margin of safety to tak8 care of rivet holes. However, the compressive stress of -61880 is too high and re-design is necessary. The difficulty is due to the large stress from secondary bending. Thus making similar changes for the lower flange Tee as was done for upper ~'~,ange Tee would solve the pr-ob Iem without adding arorec tame weight. Check of Web Vertical Stiffener Strength.

Stiffener strength = P .253 = -10600 lb. The load

bein~

= FcA =41800

only 4200 lb., the stiffener

is far overstrength and should be re-designed to

save structural weight.

IlETHOD 2

Cll.14 NACA Method of Strength Analysis for Semi-Tension Field Beams with Flat Webs.

sq.

Column load in web stiffener is given in

(50).
F
f" stir'. . = -Vt h- tan a.

Method 1, or the modified wagner equations

and procedure used in example problem, are
,1

=
web

-12850

281.~6 x

.9::3

-4200 it.

The r-educed co Iumn length factor C of the stiffener is obtained from Fig. Cll.12 for a parame-ter- of
d h

(cot a. cot jJ)

= 28.56

10

(cot 40)

= . 304

somewhat conservative relative to web and web stiffener design. To eliminate this conservatism and plac~ the web and stiffener deSign on a more rational or truer basis, the NACA carried on a comprehensive study and testing program to develop a better understanding of semi-tension field beam action and to present a design procedure for use by the aeronautical structures engineer. The results of this program are summarized in (RefS. 3 and 4). The material which follows is taken tram those reports .
NACA SY"!BOLS

~nd ~ives C~
~ence

= .38

L'

.33 x 23.56

= 10.8"

Ass~~lng an effective width of web sheet equal to 30 t ~s actln~ Nith the 1 x 1 x 1/8 an~le stiffener the radius of gyration equals .267 and the area equals .253.

The list of symbols which follows 1s the same as USed in Refs. 3 and 4 except cr, ~ and cro have been replaced by f n, f s and Fu ~espectively, in order to be consistent wt tn the symbols used in t.he "Lrs't par-t of this chapter.

A S
G P Q

cross-sectional
YOw~g's ~odulus,

a~6a,

square inches.

KSI.

shear- modulus, KSI. force in KIPS.

S
::1

static ~oment about neutral axis of parts of cross-section as specified by subscripts. in. coefficient of edge restraint. transverse shear force, ~ips.
spac ing of uor-i errts , Inc hes

e
~

.937S/.l25

= 7.5

h k

distance from ~edi~n DIane of web to centroid of (Single) upright, inches. depth of beam, inches. diagonal-tanslon factor.

ANALYSIS AND DESIGN OF t

FLIGHT VEHICLE STRUCTURES

Cll.15

thickness, inches (used without subscript) signifies thickness of Neb). angle between neutral axis cf beam and direction of diagonal tension, de~ees. centroldal radius of gyration of crosssection of upright about axis parallel to web, inches (nc sheet should be included). normal stress k51. shear stress k51.

creasing beyond this critical buckling value, the buckle pattern will approach a pure tension field as indicated in (b) of Fig. Cll.14b.

r
h

Oi
.

Subscripts
DT
F S
u
TrJ

diagonal tension
flange shear
upright
web

(a) Nonbuckled ("shear-resistant") web.

cr ult e

crt tical ultimate effect! ve

SpeCial Combinations

internal force in upright, kips. shear force on rivets per- inch run kips. per inch. Rtot total shear strength (in single shear) of all rivets in one upright, kips. de upright spacing measured as shown in Fig.
~u

R"

(b)

Pure diagonal-tension web.

Fig. Cll. 14 - State of Stress ina Beam Web.

Cll.16a.

he

depth of web as measured as shown in Fig. Cll.l6a. he depth of beam measured be~Neen centroids of flanges, inches. hr depth of beam measured between centroids of web to flange rivet patterns, inches. hu length of upright measured between centroids of upright to flange rivet patterns, inches. ksg theoretical buckling coefficient for plates Nith simply supported edges. F'u "bas i c" allowable stress for forced crippling of uprights. 4 r---,:-wd flange fleXibility factor (.7d ,)1(1 ;1 ) he where Ie and IT are c T moments of inertia of compression and tension flanges respectively.
Cll. 15 Engineering Theory of Incomplete Diagonal Tension.

In the usual practical thin web beam in aircraft construction, the state of stress in the web is intermediate between pure shear and pure diagonal tension. The engineering theory as developed by the NACA considers tPAt this intermedIate state of incomplete diagpnal tension may be based on the assumption that the total shear force in the web can be divided Into two parts, a part Ss carried by pure shear and a part SOT carried by pure diagonal tension. Thus under this assumption one can write,
S

=ss

+ SOT,

whIch can be written in the form

(52)

SOT

= l<S

55 = (1-k)5

In a beam with a thin flat sh~et as a web, if the external shear load is less than the buckling load for the web, then the web is in a st~te of Dure shear at the neutral axis as indicated" in Fig. Cll.l4 (Fig. a). It we neglect the normal stresses due to bending over the depth of the web, this shear stress arrange~ent can be assumed constant over the full depth of the web. If a web is ~~in. it will buckle under a certain critical load and if the load 1s in-

where k Is the "diagonal-tension factor" which expresses the degree to which the diagonal tension is developed by a given load. Thus the state of pure shear is ~easured by k = 0 and the state of pure diagonal tension by k = 1. Fig. CII.IS illustrates the stress condition for the limi ting cases of k = 0 and k = I and for the intermediate case. The letters PS, DT and PDT as labeled on Fig. CIl.IS mean pure shear, diagonal tension and pure diagonal tension respectively relative to web stress conditions.
Cll. 16 Formulas for Stress Analysis.

Lilnltations of Formulas: The NACA believes the formulas whIch follow will give reasonable strength predictiOns i~

Cll.16

DIAGONAL SEMI-TENSION FIELD DESIGN 2 f, f --n-)'Sin2a.

fSI~~~ ~~N L~
f.;')
n s
f n '" (1 - k)f s

~~I' Y'<.JXX:'~ '-~

r' \ In '" SIii'2"Ci' 2kf,
0< 1 -< 1

fn=-f s

f n=-(I-k)fs

ported edges are given in Fig. Cll.17. To the right of the dashed line the curves in Fig. Cll.17 are plots of the theoretical equation,
F ecr F. = k ss(d )'
t

- - - - - - - - - (54)

k I

Fig. Cll.15 Resolution of Web Stresses at DiUerent Stages of Diagonal Tension.

normal design practices are used. The following limitations should be observed.

and may be used tor most aluminum. alloys. To the left of the dashed line, the curves represent straight line tangents to the theoretical curves in a nonlogarithmic plot and are valid only for 2024 aluminum alloy. If the critical shear buckling stress is above the proportional 11mit stress for the material, a plasticity correction must be made. Fig. Cll.18 presents curves for correcting the calculated elastic values for this plasticity ettect. The plasticity correction for other materials can be obtained as explained in Art. C5.8 and Fig. C5.13 ot Chapter C5.
Cll.18 Loading Ratio.

(1)

upri~hts

be too thin.

T ~ 0.0

on web stiffeners should not

(2) The upright or web stiffener spacing should not be too much outside the range

0.2

-<

d 11

<l.0

The tests by NACA did not cover beams with very thin or very thick webs, hence some pOSSibility of inconservative predictions may
(3)

ex1st i t ~ :- 1500 or less than 200.

CIL 17 Critical Shear Stress.

The loading ratio is the ratio tafF s cr where fa is the depth-w1se average or nominal shear atr-ess , wben the depth of the fl~~ges is small compared w1th the depth of the beam and the flanges are angle sections, the stress fs may be computed by the formula
te

In the elastic range, the critical shear stress between two web uprigrxs is calculated by the rornuia i -

Sw = he t

- - -

- - - - - - - - - -

(55)

where,

In beams with other cross-sections, the average nominal shear stress shOUld be computed by the formula

=theoretical

bUCkling coefficient (g1ven In Fig. Cll.16a tor panel length at he and width dc with simply supported edges.

re

Sw Q.F =-r-t-

2 Qw (1 3 Q.F) - - - - - - - - (56)

= width at sheet be~Neen uprights measured as shown in Fig. ell.loa. = depth at web measured as shown in Fig. Cll.1Sa.
Rh = restraint coefficient for edges of sheet along upright (See Fig. Cll.16b)

Where QF is the static moment about the neutral axis of the flange material and Qw is the static moment about the neutral axis of the effective web material above the neutral axis. For the computation of I and Q, the effectiveness of the web must be estimated in the first approximation. As second and final approximation, the effectiveness of the web may be taken as equal to (1 - k), where k is the diagonal-Gens ion tactor determined in the next step. Thus in computing I and Q the eftective web thickness is (1 - k) t.
Cll.19 Diagonal-Tension Factor k,

Rct = restraint coefficient for edges at

sheet along flanges. Cll.16b).
(It dc
~

(See Fig.

he, substitute hc for dc; dc for h c; Rct for Rh; and Rh for Rd)'
Curves of the critical shear stresses for plates at 2024 aluminum alloy with SimplY sup-

Having determined the loading ratio fsIF s cr' the diagonal tension factor k can be read from Fig. Cll.19.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.17

Cll.20 Average and Maximum Stress Web Stiffener.

in

Upright or

and the axis of the beam can be determined by the use of Fig. Cl1.22. ell.22 Allowable Stresses
in

The average stress over the length of the upright for a double upright (stiffener on each side of web) can be calculated by the formula.
fu

Uprtghtlll.

Four types of failure are conce Ivab.Ie . (1) (2) (3) (4) Column Failure. Forced Crippltng Failure. Natural Crl~pllng Failure. General elastic instability failure of web and stiffeners.

__ a_ =__k_f-"s_t_an
r;

~~

- - - - - -

(57)

+ .05 (1 - k )

This equation can be evaluated with the use of Fig. ClI.20. The stress f u is uniformly distributed over the cross-section at the upright until buckling of the upright begins. Eq. 57 assumes the values fs. k, and a to be the same in the panels on each side of the stiffener. If they are not, then average values should be used, or a conservative check should be made uSing the largest shear value. Single Upright (or Web Stiffener on one side at
'Aeb)

Column Failure Column failures in the usual meani4g of the word (fallure due to instability, without previous bowing) are pOSSible only in double uprights. When colunn bowing begins, the uprights will force the web out of its original plane. The web tensils forces will then develop components normal to the plane of the web which tend to force the uprights back. This bracing action is takJn into a:count by USing a reduced ~effectlve~ column len~h L e of the upright. which is given by the following empirical formula,
hu k a (3 - 2

The stress t u for a single upright is obtained in the same manner, except that the ratio Auldt is replaced by Aue/dt, where Au

.'!.) h

- - - - - - -

(60)

For the single upright fu Is still an average over the length of the upright, but it applies only to the median plane at the web along the The problem of "c cLumn" failures in a single line of rivets connecting the upright to the upright has not been investigated to any extent web. In any given cross-section of the upright. and test results are greatly at variance with the compressive stress decreases with increasing theoretical results. Two criterions are distance fram the web, because the upright is a suggested for strength deSign, na~ely:column loaded eccentrically by the web tension. -r, Thus formulas for local crippling based on (a) The stresses t u should be no greater uniform distribution at stress over the crossthan the column yield stress for the upright sections do not apply. ~terial. This accounts for the upright acting as an eccentrically loaded compression member. Maximum Stress in Upright (J) The stress at the centroid of the upThe stress f u in the upright lTaries from a right (~hich is the average stress over the maximum at (or near) the neutral axis of the cross-section) should be no greater than the beam to a minimum at the ends of the upright allowable column stress for the slenderness (~gusset effect~). The maximum value 1s given ratio hu/2p. This is an attempt to take in by tre following equation:account a two-wave tY"Pe of buckling failure that ~~s been observed in very sl=nder uprights, t r~. = f u ( ~x.) _ - - (59) Forced Cripoling Failure where. (f~ax./fu) is the value at the ratio when the web has just buckled, Fig. Cll.2l gives the lTalue at this ratio. Cll.21 Angle of DiagOnal Tension. Having determined k and fults, the angle a between the direction ot the diagonal tension The shear buckles in the web will force buckling of the upright in a leg attached to the web, partiCUlarly it the upright leg is thinner than the web. These buckles give a lever a~ to the compressive force acting in the leg and therefore produce a severe stress condition. The buckles in the attached leg will in turn induce buckling of the outstanding legs.

The stress f u at which ~olumn failure takes place can be found using the standard column curve with the slenderness ratio Le/~ as shown.

Cll.IS

DIAGONAL SEMI-TENSION

FIELD DESIGN

In s tng La upr ~~ht;.-; the out s tandt ng Legs are

Cll.23 Web DeSign.

to a cons t de-ab Le extent by virtue of ~he Lac~ t~~t the compresslv~ stre2s jecr:3~Sa8 "Ii ':h distance rr-on tile 'Neb; t he allowable st~ssses f'0r sinsl: ~prt~ht5 ~r~ t~er~for9 sone.vha t h Igher than those Cor double uprights. Be~8.US9 the for:ed crl~pli~g is of local ~:ur:3, it is aS3~~ed to depend on the peak va Iue ru of the upright stress ra t her- than
!':311~lI3d

For design pur-ocses , the peak va Lue nomi:lnl web shea~ stress Nithi~ a bay is as,
fs

0: the
tai<~n

max

= f s (l+-:<:::l.)(l+-l.;:C a )

- (52)

aax

on the average

val~e.

T11e upr-Ight stress at wnIch final eo Llapse occurs is obt~ined by the following empirical method;Compute the a l Lowab Le va Iue of F Umax for a perfectly, elastic upright ~:erial by the for:nula:(1)

Refer to Figs. Cll.23 and Cll.24 to determine values of factors c, and Ca' The c, tern constitut~s a corr~cti0n factor to allow for the angle a of the ji~gonal tension field diff~ring from 45 degrees. C~ makes allowance for the stress concentration due to thp. flexibility of the bea~ flangp,s. Allowable
Shea~

Stress Fs .
shea~ str~ss

The allowab12

For 2024-13

Al'~inum

Alloy:-

by tests and depends on the va Lue of the diagonal-tension factor k as ~ell as on the

Fs is

det~ruined

det8.ils of the ~eb to flange and web to upright fastenings. Fl~. CI1.85 gives empirical allowab Le cur-ves for t'NO a Iumt num alloys. It should Fu=Z6000k a / : 5 (tu/t)l./::5 (For Single uprights)-(51b) be not~d that these curves contain an allowa~ce for the rivet factor; i~clusion of this factor For 7075-T6 Aluminum Alloy:in these curves is pOSsiJle bec~use tests have shown that the ultimate shear stress based on Fu=26000ka/::5(tu/t)~/::5(Fordouble uprlghts)-(61~~ the gross section (that is, without reduction F u=32:500k a /" (tu/'.: ) l./" (For single upr-Lghtisr- (;31cl.) ot rivet hol~s) is al~ost constant within the normal rang~ o! rivet f~ctJr (CR> 0.5). (2) If Fu exceeds the proportional li~it for the upright material, use as ~llo~able For the allowable streSS for other val~e th~ stress corres~o~dln~ to t~g ccmpresma t er-Ia Ls re rer- to Fig;. ':::11.42. s ive strain Fu/E. Permanent Buckling of Web. Fu can ~lgo be obtained for various ~terialg from Fig. Cll.38. A check for the development of permanent shear buckles can be ~de USing Fig. Cll.46. Na~ural Crlppli~g Failure In this fi~re Fs P . B. is the allawable web gross shear stress for no permanent buckles. The tern "natural c r-tpp.l tng failure" i s The Air Force usually s,ecifies no per~anent use-r to denot.e a cr-t opl tng failure resut t mg from a compr~s31ve stress unifo~ly di~trlbut~d buckles at limit 108d. Over the ~ross-sactl~n of the upright. By ~his definition it ~a.~ occur only in ~ouble unri~hts. ell. 24 Rivet Design. ~o avoid r2tural crippling failure, the peak ~eb to Flqnge R:vets stres3 in the upright ~~~. in ~he upr~ght snoul.c 'De less than the c z-t poLing stress of the The load per inch run acting on ~eb ~o section fer Lip _ O. It appears that crippling flange rivets is taken as, re.i iure does not appear- to be a cont ro Ll tng s~ t~ctJr in ~ctual deslsns. R" =-h (l +- 0.414 k ) - - - (53) .g General El~sti~ Inst~bility of ~eb and Upri~hts ',.Jith coub Ie uprights the 'Neb to upright Test experi~nce so far ~~S not indicated rivets ~ust orovide suffi~ler.t longitudinal that general ~la5ti8 ln~t~biitty need be ~on shear at r-eng'th to aac- the two upr Lgtrts act as s icer-ec in des Lgn . Appar-ent Ly the web system an integral unit until column C~ilure occurs. is safe aga t ns t general eIas t t c instability i f The total shear strength (single shear strength t.ae upr-Ight.s are de s Igned to rail by column 0: all ~ivets) reqUired in an upright 1S , action ar by farc8c crl)pli~g at a shear load 2 F' co "'.: not ~uch l~ss than th~ shear stre~gth of the - (64) R = total b Le .veb
Fu=ZlJOOk~/::5(tu/t)l./::5(Fordoub La UDY'U,LltS\-(51~)
"

-.

A:fALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.19

'.vher-e

flange, especially when the (See Chapter 03.) ::'-';0

~

uprl~ht

is joggled.

co luan yield s t'reng t h of upright .nate rfa I . (If Fcc is exp res sed in
kS!, Rtot~l will be in ki)S.)

Cll.25 Secondary Bending Moments in Flanges.

...

stat ic moment of cr-oes-eect.t on of one ~pright ~hout an axis in the medLan plane of the web in.~

The secondary ~ament in a flange, caused by the vertical com)onent of the di~gonal tension may be taken as,
lkft M = 12 s d'C,
(69 )

width of

outstandin~

lag of upright.

:1U' L e :: rat i o cb ta Inab Le from ror-nuje (60).

The rivets aus t also have surr tc lent tens lIB strength to prevent the buckled sheet from lift Lng off the stiffener. The necessary strength is ;::.ven by the criterion. Tensile at reng't h per tncn of rivets :.0.15 x t Ftu (65)

where C~ is a factor given in Fig. Cll.24. The moment given by this roruuja Is the maximum. moment and exists at the ends of the bay over the uprights. If c~ and k are near unIty, the moment In the middle at the bay is halt as large as that given by formula (69) and of opposite sign.
Cll.26 Shear stiffness of Web.

wne r-e Ft.u

1;'3

the tensile strength of the web. The theoretical effective shear modulus at a web 0e 1n partial diagonal tension 1s given in Figs. Cll.26a and Cll.26b. In Fig. Cll.26a, GIDT is valid only in the elastic range. The correction factor for plasticity is given for 2024-T3 alumInum allay in Fig. Cll.26b. The effective modulus should be used in deflection calculations. - - - - - - - - - (70)
Ge , for example, should be used in place of G

?or Web to Upright Rivets on Single Uprights. The r2quired tensile strength is given the tentative c!'iteriJn, Tensile strength pe!' inch of rivets'
.22 x t Ftu - - - - - - - - - (56)
~y

(The tensile sur-eng-ch of a rivet is defined as the tensile load that causes any faIlure; it the sheet ls thin failure will consist in the ;)ulltne; Q: the r-t vet through 'the sneet.. )
No c r-tt.er-ion for shear strength (If the !'ivets on Single ~prigh:s has been established; the criterion for tensile strength is probably adequate to l~sure a satisfactory design.

in the flexibility coefficients for shear panels in Art. A7.10 of Chapter A7, if buckling skins or webs are present.

The pitch of the rivets on sinBle uprights should be small enough to prevent inter-rivet buckling 0: the web (or the upright leg if thinner than the web), at a compressive stress equal to f~ax. The pitch should also be less tnan 1/4, in order to justify the aasumet Lon on ~dge support used in the dete~ination of
~ SCI"

Cll.27 Example Problem. USing NACA Method. (Method 2).

The be~m used in example problem ot Art. Cll.l3 will be checked by the NACA Method.
withi~

First check to see if given be&~ falls the limitations Of the NACA formulas.

Prom Art. Cll.IE:-

Upr1ght to Flange Rivets These rivets :nust carry the load existing between upright and beam flane;~. These loads are, Pu :: f u A-..: (for double uprights) Pu :: f u AUa (for single uprights) (67)
(68 )

-t- shoul.d be grea'ter- than 0.6

For our beam t u the leg thickness of the is .125~ and the web thickness 1s .025~, henc~ .125/.025 = 5, which 1s greater than .6.
uprl~ht

t ll

"

Also from Art. ClI.l6, the d/h value for the beam shoul1 fall between .2 and 1.0. The d/h value for our beam 1s 10/30 :: .333 which falls within the ~iven range.

~hese formul~s neglect the gusset effect (decrease of f u towards the ends of the upright) in order to be conservative. Two fasteners shoulrt be ~sed to attach the upright to the

ell. 20

DIAGONAL SEMI-TENSION FIELD DESIGN

Calculation or the Critical Shear Stress (?scr) Use will

be mae or Fig. crr.r-,

T -

de _

10 ~ 400 ::; Stiff~ner sDacing (See Fig. .02.5 Web thickness Cll.l6)

Hence, f u = l&~50 x .92::; 16900 pSi, which represents the average stress over the length of the upright but applies only along the median plane at the web or along the line at rivets connecting upright to web. Calculation ot Maximum Stress Upright

h c =~::;? 79 See Fig. C11.16 tor de 10 .....,

he.
Value of

h~

'=

281.~0 ~

.352

crr.r-

Usi~g

the above values, we tind from Fig. where, d ::; upright spaCing h u ; distance between centroids at upright-flange rivet connections. Using .352 as value at d/h u and k = .69, we tind tram Fig. Cll.21 that tUmaxltu = 1.17. Hence,
f

that

Fs

er = 370 psi.

Calculation ot the Loading Ratio fs/F scr From equation 55 Sw 13500 f s ::; - - = het 29.45 x .025 ::; 183.50 pet . where,

Umax . = -16900 x

1.17 -19800 pSi.

Calculation at Diagonal Tension Angle a From Fig. Cll.22 for values ot k = .69 and tu/t s = 16900/18350 = .92, we obtain tan a ::; .815. calculation of Allowable StresseS for Upright Check allowable stress for failure as a column:Since the design is of the single upright type, we check the following two criteria:(1) Stress f u shOUld be no greater tr~ the column Yield stress for the upright material. In the previous solution at tto.is beam by Method 1, the cripPling stress of the web upright was calculated to be -S25oo pSi., which corresponds to the col~~ yield stress Fc o Since f u (average) was -16900 p5i., the upright Is tar overstrength tor this partIcular strength check.

Sw
he

= distance

external shear load on web. between flange centroids.

Equation 55 was used because the flanges will take very littl~ of the shear load. Therefore the loading ratiO,
F

t _s_ = 18350 = 49 6 370 Scr

Calculation of Diagonal-Tension Factor k With fs/F Scr ::; 49.$, we use Fig. Cll.l9 to find value ot k ::; .69, for zero curve since sheet is flat or R = O. Calculation of Average Stress In Upright
ext~~ded

Upright consists of one I x 1 x 1/8, 2024 angle section. obtain effective area, we use eq. 58

To

Au. where,

Au
1

(~)'

=1

.234

(.3025) .. .29

= .112

(2) The streSS at the centroid of the upright should be no greater than the allowable column stress for a slenderness ratio at hu = 20 = 28.50/2 x .29 ::; 49
The
Johnson-Eul~r col~

equation 15,

Au ::; area ot web uprl~ht or stiffener e ::; distance from median plane at web to centroid of Single upright. p = centroldal radius of gyration of cross section of upright about axis parallel to web.
112

::'')500 ' v~ (49)' = c= 52500 . -4n.oJxlO,700,OOO -38000 pat

~.8.

= (38000/16900) - 1 = 1.14

lOx .025 fults =

Allowable St~e~S for Unright for ?orced Cri ppl1r::~ For a Single member upright equation (6Ib) applied

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.21

Fu ' 26000 k'/' (tu/t)

'I'
34600 pat

(23700/19600) - 1 . 21.

Fu

::I

26000 X .69 a/ 3 (.125/.025) 1./3 =

Check of Rivets Loads per inch of run web to flange rivets is given by the following equation.
R'

The Fcy for 2014-T6 Ext 53000 Correcting tram Fcy = 42000 tor 2024 material by directly as the V at the yield stresses, which 1s approximately true, we obtain,
(57000) '/' Fu ' (42000)'1- x 3460038800 pSi

~ r

(1 + 0.414 k)
+

28.50 131500 (1

0 . 414 x .69) = 510 lb.

M.S (38800/19800) - 1 0.96 The stress is below the proportional limit stress or the material so no correction is necessary . The stress Fu could also be found by use or curves in Fig. Cll.38. The web stiffener or upright is too much overstrength and should be re-designed. Check Web DeSign The peak value of the nominal web shear stress withi~ a beam bay is taken as, r Smx .
::I

= 456

Load per rivet PitCh of 3/4 inch: .75 x 610

lb.

These values are practically the same as by the first methoa of solution in Art. Cl1.13. Single shear strength of 5/32 2117-T3 rivet 512 lb. Bearing strength on .025 web = 486
lb. (critical).

Strength per rivet pitch = 2 x 486 = 972 lb. M.S (.72/456) - 1 = 1.12 Web to Upright Rivets The required tensne strength is given by the criterion, that the tensile strength or rivets per inch of stiffener should be greater than 0.22 to Ft '4 which equals 0.22 x .02.'5 x 62000 = 340 Ib./inch. In our rivet problems we have specified no web-stiffener rivets. Thus rivets should be specified that will develop 340 lb. tension strength per inch of stiffener. See Fig. Cll.37a tor tension strength for some fastenerskin combinations. Refer to further discussion in latter part of Art. Cll.32. Upright to Flange Rivets

(l ... k

C1.) (1 ... k Cal

From Fig. Cll.23 using tan a = .815, we find C1. .022.

USing Fig. Cll.24, we must tind value of term wd = 0.7 d j

4'r-----:,-_
t
(Ic + IT) he

before the value of

ell can be found.

wd = 0.7 x 10

;'---=---= 1. 96 .025
(.1<175
+

.0291) 29.45

The web uprights are fastened to the flange both upper and lower by 1/4 dia. AN steel bolt. The load on the upright at its ends is, Pu t u A U a 16900 x .112 = -1895 lb.

hence from Fig. Cll.24, Cil = .075

Suoe't t tuting:rsma.~.

= 18350 (1

+ .69

x .022)(1

+ .69

x .075)

Shear strength of 1/4 bolt = 3681.

= 19600 pSi.
Allowabl~

= 1.215 x

Bearing on vertical 3/32 leg at lower flange 2340 2930 lb. (critical).

Web Shear Stress

M.S. = (2930/1895) - 1 = .54

Since a l/~~ dia. rivet Al7ST, FSU a 38 kai has a shear strength of 1970 lb. it could be used instead at the 1/4~ bolt and the M.S. would be .04. Check of Flange Strength Section
50~

From Fig. Cl1.25a for k = .69, the allowable F sa l l . tor 248T (Ftu = 52000 psi.) equals 21000 pSi for web without rivet 'Hasher and 23700 pSi !or web with rivet washer. The margin of safety for web with rivets Without washer 1s (21000/19600) - I = .07. For rivets with washer the margin ot safety would be

trom end.

DeSign external

ell. 22

DIAGONAL SEMI-TENSION FIELD DESIGN

!1l0m~nt ~

bending

Ji~g0nal tens~on f~ctor

50 :< 13500 = 675000" lb. = .69

Hence FF ::; - .69 X21~~00 x 1.228 = - 5710 lb. Total Flange Average Axial Loads Upper flange Fe = -21350 - 5710 ::; -27060 lb. Lower flange Ft ::; 21350 - 5710 = 15640 lb. Average streSS upper flange::;
-27060
.675

Thus bending !1loment developed as a shear resistant bea~ = (1 - .69) 675000 = 209000" lb.
The re~ainder of the bending moment is deve Ioped as a :JUTe tension fi91'_1 beam, or a moment of 466000" 10. Bending
5tr~sses

= -40000 ps t

fi:>er) f o

Upper flange bending stresses, (extreme 209000 :< 12.~8 466000 x 10.94

Average s'tr-esa lower nange = .378 ::;

40300 psi.

15640

=-

270.0 9700 - 24300

210 34000 psi.

Lo'.ver Fl'lnge Bending Stresses (extreme fiber) fb

209000 x -17.42
:= -

Comparing these values Nith the values obtained by the first method of solut10n we find -44000 and 424.50 respectively. Thus the NACA method decreases the load on the flanges. Check of Flange Stresses Due to Secondary Flange Bending Moments From Formula 65

27 0.-') -

-466000 x -19.06 2'0 ...

= 13~SO
Avera~e

42400 = 55850 psi.

Axial

?l~nge

Loads Due to Bending From Fig. Cll.24, C3 Hence,

M:

As a shear resistant beam average axial flange load equals

id,
)

::;

.97 when wd = 1.96

::;.....L x
12

.69 x 18350 x .025 x lOa X .97

where M (1 - k) = 675000 (1 - .69) = 209000" lb. a: t i-ae of 'Neb buckling. he = distance bet~een flange centroids I 3 moment of inertia of entire beam section about ne~tr~l axis I w = mament of inertia ot web about beam N.A. Substi r.ut ing Flange load Average
~ial

2560 In.1b.

This compares with 3210 in.lb. by the first method of so Iutnon . The flange stresses could. be found as in solution method 1.
Cll. 28 General Conclusion.

209000 (270.5-60.5):1 __10 b 29.45 270.,:> 00.J.

be~ i~ ai~gonal

flange load due to bending with tension state,

In general the NACA method gives higher margins of safety. The NACA method is recommended for actual design of semi-tension field beams. The MACA method has been extended to cover curved webs and this subject 1s presented in Part 2 of this chapter. In the example ~robl~m the web ;ViaS relatively thin and the diagonal tension factor k of .69 means that wrinkling is quite severe as 69 percent of the shear load is carried by diagonal tension. In heaVily loaded and racher srallow depth beams, such 13.3 wing spars or wing bulkheRds sUbjectad to larg~ external 10~ds, the webs are ~uch thicker and the k factor much less.

= 456000
29.~5

=t

1-840 lb
c
Axi~1

Total Flange Average Upper r Lange Fe Lower r Lange Ft

~

Loads Due to Sending:

-15840 - 5510 - -21350 It. Fe ~ 21350 lb.

Flange Axial Loads Due to Tension Field Action L'F - --2-co" a. SOT shear load carried by di~gonal tenslcn field action::; ks, where x = .69
_ SDT

over that The 5reat saving in web weight required for a non-buckling web deSign easily exceedS the weight increase in thA flanges and the web uprights that diagonal tens ian field ac~ion produces. The rivet design in semitension field design presents more detailed design problems since the z-i vet loadS are larger

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

Cll.23

and more complex than for a beam with nonbuckling web.

Cll.29 End Bay Effects.

w = k q cot a

= .69

(18,350x .02.5) x 1.228

= 388

lb./ln.

The previous discussion has been concerned with the ~interior" bays of a beam. The vertical stiffeners in these areas are subject, prDDarily, to onlY axial compression loads, as discussed. The outer or "end bay" is a special case. Since the diagonal tension effect results in an inward pullan the end stiffener, it produces bending in it, as well as the usual compression axial load. This action can be clearly seen in Fig. CII.? Obviously, the end stiffener must be considerably heavier than the others, or at least supported by additional members to reduce the stresses due to bending. Actually an end bay etfect exists wherever a buckled panel ends and structural members along the edge of the panel must carry the bencing due to the diagonal tension loads. Typical examples, in addition to the end stiffener discussed, would be the edge members bordering a cut-out or a non-structural door in a fl~t beam, or a curved fuselage or wing panel. Ill~strdtions are shown in Fig. ClI.27.
Flange
UYCut Out

This is a severe loading tor the end stiffener to carry in bending in addition to its other loads. A heavy member would be required. There are in general 3 ways of dealing with the edge member subjected to bending, the object being to keep the weight down.

(1)

Sunply ~beef-up" or strengthen the edge member so it can carry all of its loads (thiS is inarr ic i ent for long unsupported lengths). Increase the thickness 0 t the end bay panel to either make it non-buckling or to r-educe k and thereby the running load producing bending in the edge member (thiS is usually inefficient for large panels).

(2)

Provide additional member (stiffeners) to support the edge member and thereby reduce its bending moment due to 'N. (thiS requires additional parts). Actually a combination Of these methOds might be best.
(3)

Stiff.

Members
Provided to Frame
Cut~OUt

Tens.
Fteld~

Pull

Consider method (3) above. This will sometimes increase the local shear in the bay and should be considered. Assume that 3 additional stiffeners are added (equally spaced in this case) to support the end stiffener against bending and analyze the end bay internal loads resulting for the beam Of Fig. ClI.13 and the analysis used for it in Art, Cll.27 (NACA
l1ethod)

(.1

(b)

Fig. ell. 27

Fig. Cll.28a shows the end bay and the loads applied to it.
10,280t

The component of the running load per inch that produces bending in such edge members is given by the formulas w=kqtana for edge members parallel to the neutral axis (stringers) and
w
w

=38St/in

-.[~
y;

I"'- 10..--t 13, 500*

7. 5" 7. 5" 7.36"

13, 500-1

70>
45

13, 500>j 29.45"

q -

457lit/inl

V'" 13,500

k q

cot a

1, 14.Q.!L...;o

7.~5"
(a)

for members no~l to the neutral axis (stiffeners or rings). The longer the unsupported length of the edge member subjected to w, the greater will be the bending ~oment it :nust carry. For example, consider the end stiffener of the beam of Fig. C1I.13 in Art. ClI.13. Using the data from Art. C1I.27,

45 70>

..-

<bJ

5,7~

38St/in.

l~ f",4'1

a fA.

:: 1"A"?''1
v"143~

80> 10.~

5,71.QJt.......

m.@ r 500 13, 500# J! I

388>/

t88~inj

13, 500#

\"'v"42~
~

0>1, 1~

(c)

-~
(dl

3~

I I

2. j

Fig. Cll.28

Cll.24

DIAGONAL SEMI TENSION FIELD DESIGN

Note that in Fig. (a) there are two sets one is the basic applied shear load of 13,500 lb. and the other is the component of the tension field load, w, calculated as 388 lb./in. earlier in this article.

at applied loads.

Fig. (b) shows the shear flows in the end bay panels, taken as constant in all the panels, due to the 13,500 lb. load applied. The shear tlows are shown as they act on the edge members. Fig. (c) shows the shear flows in the panels due to the applied load w. Average shea tlows (in the centsr of each panel) are shown. Actually the shear flow in the end bay will
vary tram a maximum of q ::I 388 x 22;~o = 571 Ib./tn. to 0 lb./tn. at the center. The values of this variati~n at the center of each bay are shown for analysts purposes.

or q = 29.45 = 457 rs., in. existing in the other bays at the beam. This means that the diagonal tension effects in this area will be greater (tor the same web thickness) and must be considered locally here tn checking the upper flange, the end stittener, the added support stitfeners, the rivets, the web, etc. Having determined the basic internal loads, the members involving the end bay can be checked for strength using the methods and data of Art. Cll.14 to Cll.26. When, as in the case at the upper added support stiffener, the shear flows are difterent in the adjacent bays, average values or q and k should be used in cnecktng the stiffener for strength. Formula 57 assumes equal shears in adjacent bays.

13 500

In general, there is no stmple analytical way at calculating exact tension tield load variations when shear flows vary trom panel to In Fig. (d) the load systems of Co) and (c) panel in a structural network. The procedure out11ned above is but an elementary ~approx1 are added to obtain the final (preliminary) matlon~ that can be used tor deSign purposes. loads. It can be seen here that the shear flow in the upper two panels at the end bay ts If all margins are near zero, sUbstantiattng Significantly increased over the nominal value element tests are in order.

____________________1

ANALYSIS

AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CII 25

'0

I
-l

SrI-8
Fig. Cll.16 Graphs for Calculating Buckling Stress of Webs.
"''U

he - Ld)- t - -i-

r-li~

L1'.8
1.6
L"

I
f ,

+' '1-- +-: r- r, L

d.-J- ~

r-l'l.JF
'e
=I,

,,_.W
p-

-5 ,...,

12
LO

,
4

"

r-,

? .. ~ ~.9
.6
.4 /

,I

""T

. -.:::
de,
2

.2

'_;'~"'S~~7:

"
coefficients $uPQorted

(0) The cr-etic c! with simQly

fer- etctes
edges.

,':l)

!mpirico!

I"astl"oll'lt

ccet fk::a:'1ts.

40

1\ !
20
'6
,

1\ i
I

I~ i\ ~ j\
,
,\\ \

20

, , ,
h,
de

,.0 t- t-

_ 12
~

, 1

6
4

Fig. Cll.17 Buckling Stresses F Ser for Plates with Simply

Supported Edges. E = 10, 600 kst. (To left ot dashed line, curves apply only to 24ST aluminum alloy. )

1
,

1
i

\~

i
1\'\
.'.'

/'

,
1

./
~

1.5 2

3 X"'

f-h e
4

1-'
,

de II' \
i 1.5f

I
I
200
I
I.

n:p~\ I ?"

i.. . ~ i" ~~
~

-.

. .. .."
0

" ',,, 1"-1

600

40

Alclad

400

30

F ecr ' kar

10

Ip

j:;7 -: V

~ \--

1/

7~S-T6

24S-T3

I-

Fig. Cll. 18

oV
o

1/
-:::::.~/
10

20

~O

40

'0

60

Fscr, elastic' kst

Plasticity correction.

C1l. 26

DIAGONAL SEMI TENSION FIELD DESIGN

1.0

. ltd
I

~~
l

s
.6
2

V V

.I

---\
h

k
.4
{

, II

//.

3 4

.2

V
6

.~
eo 100
~

I.J.
oz:~"

10

20

30 40
fslF s c r

60

200

400

600

1000

Fig.

cu. 19

Diagonal-tension factor k,

(J! h ,. d, rePlace: by

; if

%- (or -7}) >

2, use 2.)

For Flat Sheet Use Zero Curve.

1.0

,
1

I
I

9f"
a
.1
6

II
15

f -a,OS

.10
i

I
,

Au,
dl

5
4
I

3
2
I

0 0

:/ .I ;1 / /0, y: /. / { , : / / I I I ,I I Vi ./ / j:/ Y}9.1/// ; / I I I , . . IV ..< I / ~ : / I' 1;1 /1 /. :/.: VI. )~~6' / / / /1./ I 11/ / 'V /. ~ ~ /. 2 0 / I '/ / /. /. V / . / . /. '.L' / , 11/ Y /. V V, V ' / / /14".::~ / / / / /: /1 ~ /:/ , J /I.~o / / ,I " y /. V ' ./ /1 / . / . / . / 200 I {i I I .; / / , / ' V 'v X/" /. . / . / . / 250,'~. I / 1'1 ,( /. V / / . / /' ./ . / . / I I /, ' , / ' ./ . / . / /~0071 "~I /VVV/Y/,/,.//.// .....5 .0dC-J /;1//1 /VV/vV././/// / / " . ,
1/
j

II II
:

~
{

20

, 'I ,i 1 Y , 1/ 1/ I / 1 / / / 25 30135 .1. V V '/ '/ , , ( ! , TI-r; .4550 . /, ! / ~

Ij

, ,(
'

!/
/1
!

;/
/'

I
I.

,I

,I

Ii

55,&,

,A

i /,'

A
I.

'? /., ,90,-",//

/"0
9

1.0

~./

Fig. Cll.20 Diagonal-tension analysis chart.

ANALYSIS AND

DESIGN OF

FLIGHT VEHICLE STRUCTURES

Cll. 27

18

r-,
<,
'<,

~
.2
"<,

"

-'-u12

fumax

r-- "-

t---....

t---....

.........
"<,

r--

-- r-- r--. ~ t-. -:-...

!--.
4
~

I---- r::::::: -.........

~~

.6

'u

Fig. Cll.2l Ratio of Maximum Stress to Average Stress in Web Stiffener. NOTE for use on curved webs:

For rings, read abscissa as

f; for stringers, read abscissa as %-.

J2 JO .oS
I

1.0

I'>h. f::: l0 3
I

!
I

I"

I'- 1'-

.02 .0
i
,

tan a

.8

1'-- r1"

,
7 .8 ton a .9

l.Q

Fig. Cll. 23 Angle Fa.ctor Ct.

7

6
0

I
.5
10
15

1.2
20

'u r;Fig. Cll.22 rncomplete Diagonal Tension.

1.0

o"
-e

.8
.6

< 0
N

.4

2 00
Fig.

2 wd
(wd .o.7d

cu. 24

Stress-Concentration Factors C2 and C3'

VUe ,'IT) '" )

ell. 28
40

DIAGONAL SEMITENSrON FIELD DESIGN

40

,---r-..,--,--.,--..,--,--,-..,--,-,

,5
<,

I
Heavy

25

"--

<,

1---_

WC5~~

-'y
I

,-

i ,,

... '"
.~

20

I
I

l-

T
I I

... '"
.~

20 f-+---if-+-i--+---t---'--+-+-~

~ 15 f-+--,f-+--t--+--+--+--+-~-1 I r..
I : I

'a

I
I

-rT TTed
'
I

Ion
k

I~U" i62("
.6

"--

72 kSl'_

.8

'0

.4

.6

1.0

(a) 2024 Aluminum Alloy.

(b) Alclad 7075 Aluminum Alloy.

Fig. Cll.25 AlloW"3.ble Values of Nominal Web Shear Stress.

1.0 .8 .6

""

~
,
1.0 4

"'

II~

~
G
4

,-

.2

o,

I I
234610

,.
204060
fs

.. 1.

00

200

600 "'

F Scr

(a) Modulus Ratio for Elastic Web.

1.0

.8

I
G 6

~T 4
2

I I I I ! I I I .'I I \ I I I I I i 1 , I I I I 1\ , i I 1 I I I ! I i I ; i I , I I I I , , I ,I t t , I I I i 1 I I I 1 I I I I I I
I,

I I I I I

"-

'6
fs ' kai

(b) PlastiCity Correction for 24S-T3 Aluminum Alloy.

Fig. Cll.26 Effective Shear Modulus of Diagonal-Tension Webs.

ANALySIS AND DESIGN

OF FLIGHT VEHICLE STRUCTURES

cu. 29

PART 2. CURVED WEB SYSTEMS.

(BY

w.

F. Mct:OMBS)

Cll.30 Dtagonal Tension in Curved Web Systems Introduction.

Th13 type of structure has an important in the design ot light metal structures. ~he structur~l designer should ~~ve as good an understanding of it as he ~us~ have for the some.vhat; s tmp.Ie r o.Iane 'Neb system. Actually, most airframe shear web syste~s are curved r~the~ than tl~t, the fuselage of a mode~ aircraft being me oustanding example. To make a fusel~ga sKtn entirely non-bucKling would require a very thi~k skin and/or a closely s9ac~d substructure supporting it. This would involve a considerable wei~ht penalty compared to the buckling skin arrang~ment. The typical metal skin in a ~odern fighter or transport air~lane thus carri~s its limit and ultimate loads Nith a considerable degree ot skin bU8kli~g. In view of thiS, the need tor an ~derstanding 0: di~gonal tension effects in curved web systems is obvious.
pl~ce

are present. Comparing this structure to a plane web beam, the stringers correspond to the flanges, the rings correspond to the uprights and the skin corresponds to the web. ThUS, the stringers carry (or resist) axial loads. The rings support the stringers and, if not at the "tloating" type, also divide the skin panels into shorter lengths. The skin carries (or resists) shear loads. Now, assume the structure to be subjected to a pure torSion, T, as shown in Fig. Cll.29. Betore the skin buckles, thIs torsion produces a shear in the skin panels given by the wellknown formula

only the skin is loaded. As In the case or the upri,shts ot a pl~ne web beam, the rings are not loaded. There is no load in the stringers.

Cll.31 General Discussion.

Sefore getting into the details of design, a ~eneral discussi)TI of what happens in a curved web system as the web buckles is helpful. Consider a se~i-~onoc~que structure with a circul~" (o~ elli~tlcal) cross-sectional Shape as shown in F~~. Cll.89.

AS the torsion is increased, however, the skin shear stress eventually becomes larger than the critical buckling stress and the panels buckle. Any further increase in torsion must now be carried as diagonal tension. Five main things then occur as the torsion is increased above the buckl1~ value I as illustrated in Fig. Cl1.30.
1.

Ring Attached to Skin

~Ring
~in

The skir. panel~ buckle and flatten out between' rings fastened to the Skin, from their original curved Sha?6. This gives a polygonal cross-section (away from a ring). The &~gle ot diagonal tension is less than that for a pl,~~e web bea~, however, in the range of 20 0 _ 30 0 The stringers now feel an axial load, due to the pulling on tne ends, (r~1l.30b), of the structure by the buckled skin, Just as in the case at the pl~ne ~eb beam. The stringers also feel a nQ~1 loading tnat tends ':'J bend, or "bow", t nem inward bet-.veen supporting rings, as shown in
Fl~.

~Ring
..Floating" Ring, Attached Only to Stringers
Fig. ell. 29

The structure v8nsists of a

n~~oer

ot axial

3.

member-s (called "s'tr Inge ra'' i f they are numerous and "Longe rons" if they are few in number, say

3 to 8) ~hich are supported by fr~es or rings and covered with a skin. The r~ngs ~ay be attached to the skin and "notched" to let the stri~gers pass through, as in Fig. (b), or they ~y J8 located entirely ~~de~ the stringers and not, therefore, attached to the skin. In this latter case they are called "floating" rings, as in Fig. (c). Sometimes both types of rfnga .

Cl1.300.

4.

The su~nortlng rings teel an inward loading wtitcn put.s them in "hoop" compression. For rings attached to the skin this loading is applied by the stringers and the skin, and is thus "spread out". For' floating r-Ings t.at s load.ing Is ap plied only by the

Cll.30

DIAGONAL SEMI TENSION FIELD DESIGN

compressive axial load, P, is being applied, sl~ultaneously, as shown in Fig. Cll.31.

IIL:-'\\.- --

'....-,L_ -

-- -:::...:: ~~
--)~
Fig. Cll.31

(b)

Generation of Axial Loads in the Stringers

The following Rill now rappen due to the presence of P.

1.

Net Loads Applied by the Skin to the Sub-Structure (Stringers and Rings)
~"-PST

The stringers Will, Of course, have to-, carry the compressive load, P, Rhich will be divided among them. There will be some "effective" skin to help. Less obvious, but very important, is the fact that the diagonal tension loads due to the torSion, T, will be considerablY affected by the presence ot the axial load, P. The larger P is, with respect to T, the greater will be its effect upon the diagonal tension etrects. This is as follows. a) The skin panelS will nON buckle at a lower amount of applied torsion since they are now also strained axially in compresSion. Actually there 1s a IIcom_ blned~ bUCkling consisting of compression a~d shear buckling. This can be obtained tram an interaction equation, discussed later. b) Since the critical shear buckling stress is now lower the diagonal tension factor, k , is larger. c) All of the diagonal tension effects dependent ~pon k are increased. These inclUde the axial loads induced tn the stringers., the normal loads bending the stringers inward, the loads induced in the rings and the loads felt by the fasteners.
d)

Pull of Skin

2.

~G
Id)
(c)

Loads on Stringer and Support by Rings (PRG)

End View of Stringer Showing Loads & Supports

Loads on Ring Due to Supporting Stringers. (Additional Loads also Present Due to Skin)
Fig. Cll.30

stringgrs, coming entirely from (3) above, and is thus concentrated at points. This concentration produces not only compression but also internal bending moments in the floating rings. This is shown in Fig.
:U.30e.

6.

Any t~stene~s splici~g skins together or ras t ent ng the skins to t ne end rings feel ~ct ouly a shear panel ~ype of loading Jut also 3. norma L loading, 'is 1:1 the case of a pl~n~ Neb ~eam. Also, the "folds" in the skin due to the diagonal buckles pry on the rivets as they "attempt" to extend 'leross the rlv~t lines at the stri~ers an': r-tngs .
~~e

The ang Ie of c iagcna L 'tens t on dill be la:-gar-, closer to 4.S:J.

Thus we see that the etfect 0f compression is to increase the loads due to diagonal tension. Now assume that instead of being compression, the axial load, P, ~s tension. In this case, the ef~ects of diagonal tension, due to T, are reduced.
1.

InportQnt thing to realize here is although only a pure ~orslon has been app If ed , 2;JGS.. fe rab Le axfa l Loads have been g~nerated in the strir~ers and rings. And even some bending noment s have been t ccucec in the at r t nger-s and in the rings of the "floating"
tr~~,

type.

~~rs1~n

NON, assum8 th3.t at che same t1~e the load 1s being applied an 1nr.reasing

The stringers and skin feel tension jue to P, ~hich opooses compression due to diagonal tension.

ANALYSIS AND DESIGN OF FLIGHT VEH CLE 8TR

CURES

11. 31

2.

The skl~ panels can carry a larger shear stress be fare bUCkll~ in shear. In fact, a relatively small amount ofaxl~l tension can prevent them from buckling at all, as will be discussed later.

landings, catapulting requirements, etc. Obviously all of this complicates the calculations and experience and jUdgment are of great help; but the method of gOing about it is fundamental and ~11l now be discussed. The N.A.C.A. has conducted an extensive program over the years with the object at determining a system tor the design or structures having curved webs in dfagonaL tension. The theory for this sys~em, as in the case for the plane web beam, 'NaS given by Wagner and others and modified as necessary fram the results ot many tests. The design method 1s tully d1scussed in Ret. (3) and the substantiating test progr~ in Ref. (4). Tne reader Is encouraged to consult these reterences for a fuller presentation at the theoretical devel~pment and test results. As mentioned earli~r, curved web systems are of twa general types. one of these has an arrangement whioh results in the web, or Skin, panels being longer in the axial direction, d, than in the circumferential direction, h. This Is typical ot the stringer system in a fuselage. as shown in Fig. Cll.33.

3. 4.

The diagonal tension t~ctor, k, will be smaller (or zero if no buckling occurs). All diagonal tensl~n effects dependent upon k are reduced and the dl~gonal tenato angle Is srraller.

Finally, suppose that instead ot an axial load, P, a bending moment Is being applied sl~ultaneously with the torsion, T, as in Fig.
Cl1.32.

In this case, also, as the torsion Is being increased from a small mount, M Is also being
l~creased,

(the ratio of M to T being constant,

Skin

as was the case for axial Load , p). Then, from the usual bending theory, t ~.;;: !1z , the
I

Panel
d

following will occur. 1. The stringers (and skin) above the neutral axis will teel compression loads, the further a~vay the greater the load. The ~pper skl~ panels will thus buckle earlier in combl~ed compreSSion and shear and produce the largest diagonal tension loads on the stringers and rings. skins Jelow the ~e~tr~l axis will (or ~ot at all) due to the tension str~ln ?roduced by X. Thus the diasor~l tension e:fects ~ill be 5~al1e~ (or ~:rc-exist~nt) in the rings and stri~ge~s in this region.
T~e

::::?1
J! h >
1 0

Fig. Cll.33

That is, the geometry at the stringer spacing, h, and the ring spaCing, d, is such that

2.

b~ckle l~ter

"Floating" rings, not being attached to the skin, do not determine the spacing, d. A second type of curved web structure may be referred to as the langeron system. Its main characteristic is that the skin panelS are long in the circumferential direction. This tyue or structure . . . ould, typically, in a fuselage, consist at a tew axial members (a minimum of 3 but more usually 4 to 8 tor a "tail sate" deSign) and a large number of closely spaced frames. The fr~mes WOUld, in this system, be at about a 4" to 6" spacing as compared to a 15"-20" spacing for a strir.ger system. This gives

3.

The skins near the neutral axis Nill feel litt19 or no strains due to bending and will buckle about as in the case for the ~ure torSion, T, prOducing equivalent ~ftects.

In an actual airolane structure, a fuselage example, the applied leading is ~ore complex. I~s~ead of Si~ply an applied torsion, there is also, usually, a vertical and perhaps a sideward set of loads ~hich produce shears that ~ry from panel to panel. And there may be not only a Jendi~ mo~ent, whi~h changes, alJng the r~selage, but also axial loads due to
~or

$-<1 n
as indicated in Fig. Cll.34

Cll.32

DIAGONAL SEMI-TENSION FIELD DESIGN t (t)

Longeron

Eo h

< 1 0

where 5imple support is usually conservatively assumed in deter.nlntng ks'

l~en axial loads (and strains) are present, as in practical structures subjected to bending as well as shear, the situation is more complicated. As discussed in Chapter C9, the presence of compresSive stresses together with shear stresses causes the panel to buckle at a lower ~lue of shear than it no compression were present. The presence of tension stresses, along With shear stresses, enables the panel to stand larger shear stresses before buckling occurs. In tact, a relatively small amount at tension stress may prevent the panel tram ever bUCkling in shear.

Fig. ell. 34
I~ the long8ron system the frames are attached to the ski1 ~nd longerons, there are no "rjoat ing" rings.

Most of the NACA data and design method is for the stringer system. The design method for the langeron syStem, also presented herein, has evolved from Ref. (3) and also from the results of an i~vestigatlon and test program at Chance-Vought Aircraft corp. These latter checks are in so~e pl~ces different from the ones in the stringer system. The main ,roble~ is the determi~~~ion of the angle of diagonal tension, for either system. Once this 15 known all of the di~gonal tension stresses are known. The stringer system will be discussed tirst and the longs ron system will be discussed secondLy. The basic approach is si~llar to th~t for plane web beams.
Cl1.32 Analysis of Stringer Systems in Diagonal Tension.

It is tmportant to determine the actual shear bUCkling stress when axial stresses, partIcularly compression, are present. The reason 1s that this afrects the diagonal tension factor, k, and all of the ensuing stresses affected by k First consider a shear panel 8ubject~d to a shear stress f s and a compressiJn stress f c' It has been de~onstrated experL~entally at the NACA, Ref. (5) tha~-a curved panel, thusly loaded, buckles aCcordlng-to-the interaction formula

Before the 1Iagonal tension effects can ' = 1 .0 ( ~) - - - - - - - (71) be calculated, the pr1~~ry internal loads in F ser the structure, due to the applied loads, must be determined. This can be done as discussed in Chapter A20. The engineers theory of bending where FCcr and F Scr are the crt tical panel can usually be used to deternine the axial buckli~ stresses for pure compressiJn and oure loads in the stringer~, as in Art. A20.2-A20.S. shear respectively. From Chapter C9, the -'The sheaT :lows in the various skin panels can buckling stress tor a curved sheet panel, 1n end cam?ression 15 given by the equation, be determined as in A20.~-A20.9. In the case of s'tr-rnger- cons'tr'uc t ton a more accurate n.arykc E det8~i~tionJ~ the skin shear flows may, if FCcr = 12 (1 _ lJ 2) de~i~ed, ~e obtain~d if the skin panels are e constder-ed r ra ; , r-a t her- than curved, between stringars. This 1s because the panels, atter Now for any partiCular panel, o:.lckli:lg, actually flatten out, g ivtng a ~~lyg~nal cross-5ectl~nal Shape, except FCcr 1~ediatelY adjacent to non-floating ri~gs, as A , (a constant) - - - - - - - (72) ser shown 1n Fig. Cll.30a.

r---:=

The di~gonal tension effects can now be evaluated. 1.

DT~ffi1INATION

and

OF

3KI~

OF CRITICAL BUCKLING STRESS PANEl..B.

The buckling strength ot curved panels under pure snsar and compressive stresses 1s cover-eo in Chapter cs . The equa t t on for buckling shear stress F~ 1s, "'cr

Now, also for any gi.en ap91ied loading condltiJn being check~d we can calCUlate --f c (a at the center ot the panel (before buckling) and also f s (preViously done). These stresses will bear a constant ratiO to each other unttl buck11ng occurs, atter wh:ch the

l"'f)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.33

compression stress no longer increases. C3.n VITi te

Thus

- - - - - - - - - - - - - - (73)
or

fo:: = 8 f s

Since RT 1s always greater than 1.0 when tenSion stresses are present, the actual shear buckling stress will always be greater than Fs , the cr buckling stress tor pure shear loads only. Inspection at the term in parentheSis in (78) making up RT shows that as the tension stress becomes several times larger than Fe the

Suos t t tut lng these rormutas for Fcer and t e bao::k tnto the interaction tormul3. (71) we obtain,

cr

B fs + 1_f_ '_)'" -c' A FScr . Scr

:::

value at t Scr will became several times larger than Fs cr' ThUS, it tt is large enough no buckling will occur even when large shear stresses are present.
2. DIAGONAL TENSION FACTOR, k.

LO
The next step is the determination ot the diagonal tension tactor, k. This is a functIon fs or -f-where, as discussed above,

SolVing this quadratic for f s we get

= F
[-

"cr-

~ j 2(tl '
+

+4]

- - - -

(74)

scr

where f s is the actual shear stress at which the panel buckles due to the oresence at compression stresses. Calling this ~~ress fs and calling the expression in tre brackets cr 3c
f

or

- F ecr- ser

0_ UT

depending upon whether compression or tension stresses are present. For a curved panel the formula tor k, as determined tram much test data, Ret. (3) is the empirical relationShip,
k

Scr

=:

l"scr Rc

(75 )

where Rc is always less than 1.0 when compression stresses are present.
~ Next, consider a shear panel sUbj~cted to a ~ stress, f s, and a tension stress, ft. For this case it has been experimentally d&.nonstrated, Ref. (6), that the interactiQn relationship defining buCklinf, is

= tanh

[(.5+300

~)

f - -] - (80) log" - f

s scr

R ::: radius of curvature with the auxiliary rules that

a)
it

..,fS _..!..""ft = LO2 rc r scr er and Fee r are as where FS cr

~ ,. 2 use only 2
:>

------'(75)

b)
~revlously

ir h

discussed.

d replace h by d ("longeron"
h

The shear stress, f s, at which the panel will buckle when the tension stress present t s ft can be solved f0r directly tram equation (76),
f
t

system) and, in this case, if 'd> 2 use only 2. Rather than calculate k from the formula, it can more eaSily be obtained from Fig. Cll.l9. 3. STRINGER LOADS, STRESSES, AND STRAliIl'S

2Fccr

or,
c
SCI'

(78)

As in the case of the plane web system, the total stringer load will consist of the primary axial loadS, Pp , due to applied bending moments and/or axial loads, plus the diagonal tension induced loads, PO.T.
PSTR

Salling the expression in parentheSiS in :78) :l.T we get,

f Ser ~ FScr RT

Pp + Pu,T, - - - - - - - - - -

(811

- - - - - - - - - - - (78)

Pp 1s ceter.uinp-d as in Chapter A20.2-A20.5. PO.T. is determined, Simil~rly to the c~se for the flanges ot a plane web beam, fro~ the "pulling" of the buckled skIn on the and trames.

Cll.34

DIAGONAL

SEMI~TENSION

FIELD DESIGN

Re,T depending upon whether axial co~pression or tension is ~resent. Again, k and R~ or RT are also average values for the p~nels~on eac~ side of the stringer. Thus we ~an write
Pp
+

k q h -ot u

AeSKIN

A + .5ht (l-l<)Rc,T STR

k

or
Fig. Cll.35

f s cot a

tit
As shown in ii"ir:s. Cll.35, PO.T. is the
diagon~l

ASTR + .5(l-k)R'T ,,
- (85)

tension load in a stringer bounded by pane L "a" on one side and panel "b" on the oth~r. Let the width of panel (a) be ha and the Nidt':1 at' panel (b) be hbo (For equally spaced stringers, of course, ha = h)). Let the shear- flow in pane I (a) be % and that in panel (b) be qb' Then we can write
PDT

where all terms are known on the right side or the equation except a. The total stringer strain is then
"STR
stre~s,

h~nd

=~

f'STR

- - - - - - (86)

(82 )

T,:: kee p the ca Lcu.La't Ions Sl!l.)ler (as will be apcr-ec t.sted later) we can accept; Some inacc~acy and use average values for the Ina + ho Qa + qb , respecti ve ter.ns \ 2 2 ' e'tc , ) and get

If fSTR is larger than the proportional limit or in the neighborhood of the yield stress, ~ is not a constant and a stress strain diagr~~ should be used to read E directly, USing f STR. Again, tSTH. and hence ESTR. cannot be determined until a is known (l~ter).
There is also a secondary loadir~ on the stringer whieh tends to bend or ~bow~ it imvard. This is caused by the fact that the taut sk~ns are pUlled flat on each si1e ot the stringer. Thus as in Fig. ell.30 there is an inward component of the skin diagonal tension loading that pulls the stringer inward. This loading is not a Simply distributed one; it is largest in the middle ot the stringer and becomes smaller at the supports. Ref. (13) recommends that the eJfect of this loading be ~onsidered as ~roducing secondary bending moments in the stringer, taken as fS ht d 3 k ta:1 a

PJ.T. = k

q ~

cot a

- - - - - -

- -

(83)

r-emember-Ing that this is an "average II Load , which, for closely spaced stringers, 15 =ufficient, especially tor preliminary design. The total stri:lger load is then, from 50 and 82

PSTR

= Pp

+ k q h

cot a except a.

wner e all ter-ns are

T~e

knOI'Jl1

stringer stress is obtained by dividing the terns on the right side by their respective e r r ec t Ive areas fSTH.
~) Ae~

------+ A A ?
e

_ Ps7R _ Pp

k q h

cot a AeO. T.

is the tJtal effective area, stringer a~d skin, used in deta~1n1ng the ~rlmary loads ~s in :ha~ter ~20. If Pp is tension the area is equal tc ht, ~anel is fully effective.

of curv. T:~is represents a ~~eak~ moment at the middle and at the ring supports. It 'Nill produce tension on the inRl1e of the stri~ger at the middle and compression on the inside of the st~lnger at the supports. The recammended va Iue of r"-STR is the result ,:>f many test measurements, therefore it is of a semiempirical nature.
4.
STRESS~

MsTR

24R

= radius

- -

(87)

AND STRUNS IN RINGS

A.,

'D.T.

= ASTR

.0

ht (1 - k) R, see ',"
~

?ef. (3)

- -

--

------

(54)

There are two types o~ rings, those attached to the skin and those nat attached to the Skin, called "floating rings," which suppor-t , and are therefore loaded only by, the stringers. Rings attached to the skin are usually "notched" to let the stringers pass through.

Ref. (3) suggests a more accurate calculation.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.3S

'I'he st.r Inger-s are attached to thp.se rings locally by some shear clip arranga~ent. Tae rings feel an inward acting loading which puts them in "hoop compression". These loads come fram the stringers and from the skin. The stringer being pUll~d inward by the skin, as describ6d above and in Fig. Cll.30, in turn pushes inv~rd on the supporting rings. The skin, not being flat at the ring, also pullS inward on the ring. The result of all of this 15, essentially, according to Ref. (3) a hoop co~~resSion s~ress i~ the ring for the case or a cylinder under pure torsion. This is because the loading is approximately equivalent to ~~ evenly distributed inward acting radial loading. For this case the radial loading can be taken as, per inch along the ring, f s dtk tan a
=~-;;----

rs
dt

tan u

ARG

- - - - - - - - - (91)

where all terms on the right hand side are known ex:.;ept a. The axial strain is given by
----(92)

or can, of course, be gotten from a stress strain diagram for the material. The maximum bending moment present in a floating ring is given by
k fa

t h 3 d tan a
12R
- - - - - - - (93)

The axial (hoOp) compression load in the ring

wt Ll then be

R :. radius of curv. This occurs at the Junction with the stringer. There is a secondary moment, half as large, midway between stringers in the ring.
5. STRAINS IN THE SKIN PANELS

~G=

PRG : pR

~ fs

dt k tan a

- - - - - - (88) The strain in the skin panelS is given in

(3)

The axial compression streSs in the ring will be k -f" tan a s

dt +
Nhe~e

ARG

(89)

ser.

as

.5 (l-k)
a~e

e =

r; [Si~\U sin
+

2u (1-1,)(1+

U)]
(94)

all terns on the right The

axi~l

known eXc3Pt a.

- - - - - - -

strain in the ring will be

- (90)

where u poisson's Ratio = .32 tor aluminum, every term on the right hand side or .the equation ~s kno,vn except a. Fig. Cll.36 is ot help in calculating e, giVing the value of the bracketed term.
6. DETER!1INATION OF a"

where tRG and hence 8RG are unknown until a is de t erai ned. wnen loads other than pure torsion are a~plied f s and % will ~~ry from panel to panel and the hOOp compression stress will not be ~onstant around th~ ~ing. There will also be soma varying seccndary shear stresses in the panel due to unequal "puj Ls" on each side of the panel, at the stringer, by tne buck Led skin.
n ~11en "tloating rings are used concentrated inward acting radfa l loads <ire ap~'lied by the str~ngers. This prOduces hoop compression and, since all the loads are concentrated, also some bending moments. There is, of course, ~o effective skin acting with these rings. The axial compression load is

For the stringer system (d > h) Ref. (3) shows th~t a is related to the stringer strain, the ring strain and the web, or skin pa~el, strain by the formula
E - EST 8 - 8RG
- - (95) (..) 3 R R =. rae ius of cur-v . strain (+) and 8RG and
+

...l.. 24

where 8 is a tension eSTR are entered as negative members for compression, thus adding to e.

? D.T. = t a

a is determined by suc:essive approxi~t10n using the three prior formulas for 8, 8ST and ERG and then checking with the formula (95) above. That is,
... For a faster estimate or for preliminary design, step 6 can be skipped for panels in compression and shear and a simply assumed to be 450

d~ v-

tan a

The axial com?resSion stress is then,

Cll.36

DIAGONAL SEMI-TENSION FIELD DESIGN

1) Assume a value for a.

2)
3)

At a splice (or opening) along a ring (a vertical splice) the loading is

Load/inch; f s t

Determine E, ERG and ESTR using this assumed value of a. Substitute E, ERG and ESTR into the formula for tan~ a and get a "new" value for a. Repeat steDs (1) to (3) ~s many t~es (say 3) as necessary to get a to a "converged" value.

[1+

k (_1_ -1)] - -

sin

Cl

(97)

4)

It rings attached to the skin are being used then ERG in step (2) above is obtained from (90), using (89) tor fRG' If "floating" rings are used then eRG is obtained from (92), using (911 for fRG.
~~d

Note that in either case, if the panels involved are made non-buckling (on each side of the splice) k =0 and the load per inch is the same as for a non-buckled web. It is only around a "c~t-out~ or opening that the panelS are made either non-buckling or made to bUckle to a lesser extent, and this is done to "relieve" the loading on the edge member rather tPRn on the rivets (see Art. Cll.29). The second type of rivet loads, the prying loadS, are not determinable by any formllla. Ret. (3) recommends that an arbitrary criteria be used as follows. The tenSile strength of the rivet-skin combination, Pf, should be such that it is as large as the number given by

Once a is determined, the stringer stresses ring stresses are kno~vn, having been used in getting the strains eSTR and ERG for the tinal check of a in (95). The bending moments in the stringers can then be calculated fram (87) and the bending moments in floating rings, it used, from (93).
LQ'DS ON THE R!"IEl'S

Pr per inch

= .22

t FTU

(98)

Where FTU is the ultimate tensile strength of the skin or web material being used.

The only remaining internal loads to be calculated are those acting on the rivets. These a~e at two types. a) There are the primary loads, in the plane of the Skins which try to cause shear or bearing failures at the riveted jOints, as in any spliced skin. There are alSO, "prying" forces on the rivets which try to "pop Oft" the rivet heads a r to pull the ski n up and around the rivet heads. This latter may occur, particUlarly, when countersunk or dimpled skins and flush head rivets are used and the rivet diameter is too small or the rivet spaCing is too large. These are ~secondary~ loads.

PT is usually most critical for flUsh attachments. As an aid in getting this, Figs. Cll.37s and Cll.37b give information on the tensile strength of various r-ivet types and sizes.
Cll.33 Allowable Stresses (and Interactions).

1.

STRINGERS

b)

The primary rivet loads OCCUr. whenever the skin is spliced, which is usually, but not necessarily, along a stringer or along a rir~. These loads would also be present it the skin panel ended at stringer or ring, as at an opening or "crc-cut " and the panel had not been re-enforcad by a doubler to prevent buckling.
At a splice parallel to a stringer the load per inch along the rivet line is due to the same effects as discussed for the plane web beam. It is
Load/inch; fst

Just as there are ~NO types at basic loads (and stresses) in the stringers and rings (the primary ones and the ones due to diagonal tenSion etfects) there are also two types of allowabl3 stresses for local f~ilure. An inter2ction fo~ula is ~used to predi~t adequate strength. This 1s,ror the stringer t -E-+ ; 1.0 - - - - - - - - - (99) F'cc
wher-e

fp

= stringer

stress aue to the app Lfed loads, this is the first term on the right hand side of equation (80). stringer, obtained as in Chapter C7.
fSTMAX
f

Fcc = the allowable cripPling stress for the

fSTtlAX = fST x

ST
rl~ht

[1 + k (_1_ -111 _ cos a

J

(96)

fST is the second ter.n an the hand side Of equattJn (85).

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

ell. 37

fSTMA.;{
t'sT

is a ratio

obt~ined

from Fig. the

pl~ne

Cll.Zl. It is explained web b~am discussion.

occurs in the ring midway between stitfeners. Fcc is the ~normal~ allowable criPpling stress tor the ring. It FRo ~ Fcc then use Fcc for

i~

IRG'
Floating rings, not being subject to forced criPpling by the skin, are checked in the usual manner tor the stresses due to hoop compression loads and the accompanying bending moment,
equations (17) and (19). The interaction

Fs'r= the "forced crippling" (diagonal

tensiJn effect) allowable stress tor the stringer is obtained from Fig. Cll.38, as in the case for the uprights in the plane web beam, or calculated tram the tormul~s (also zcr FRG and tRG)' FST- 2G,OOO k a/ 3 (tSTR)~/~ for 248T material.
tW"...B

equation Is,
f + f RG I1 --""~..2.;._ 1 ;: 0 - - - - - - - - - -

Fcc

(101)

FS~=
-

32,500 kala (tt STR ) 1/ 3 tor 758T material with the same restrictions noted in Cl1.22. (The total axial stringer stress is t p + tST at the area near the supporting rings, and f p + f3~MAX in the middle between supports. )
WEB

fRO is a constant stress between stringer junctions; it does not have a maximum peak as do the rings attached to skin.
3. GENERAL INSTABILITY

A general instability check tor the stringers and skin can be made tram the empirical criteria presented in Fig. Cll.39. This is obtained fram test data and recommendations of Ref. (7) and Ref. (3). The allowab1es

The stress in the stringer due to the bending moment, MsTR' at equation (87) can be calculated and added to t p in the proper manner.

are based upon pure torsion tests. The adequacy at the structure is checked by
f

Ref. (3) also suggests the stringer be checked as a column, fully tixed at the supporti~ rings and carrying the stress f STR. Some allowance shoufd be made for beam co hzan action, due to the bending moment present in the nrccte of the stringer. This can be done using so~e affective loading prOducing the moments discussed or by carrYing some extra margin of safety over and above the column buckling stress.
2. RINGS

FSI NST

- - - - - - - - - - (102)

It is suggested that a respectable margin of safety be held (M.S. ~ .15). The radl1 of gyration in Fig. Cll.39 should be made assuming the full width of sheet to act with the stringer or ring respectively and that the sheet is flat because the criterion was obtained under these
assumpt i ons
4. ALLOWABLE STRESSES IN THE SKIN (OR WEB)

The rings have allowable stresses similar in nature to the stringers. The interaction equation is,

Ret. (3) recommends that the allowable stress in a web or skin be taken as, for nonflush attachments,
f SALL

---2.. + Fcc
ror the
c~se

= f sALL

(.6S + A)

- -

(103)

1.0

- - - - - - - - (100)

of rings attached to the skin.

f p is the stress due to the ring carrying loads other than the tension field ones (as a bulkhead analysis would Show). FRG is the allowable forced cripoling stresS for the ring, obtained in the same ~nner as that for the stiffener, (rIg. Cl1.38), obtained in the same RG man:ler as discussed for the stiffener. It

= f RG x

fRGj<.AX

where 6. = .3 tanh ARG + .1 tanh AsT . t. can, dt ht ' more easily, be read tram Fig. Cll.40. t;ALL is obtained from Fig. Cll.4lb or Cll.4lc after obtaining apDT from Fig. Cll.4la. Data trom tests by the Chance-Vought Aircraft Corp. indicate that the allowable web stress can Si~ply be read from Fig. Cll.42. In either case, the allowables apply to t s ' the ~ stress in the panel. The net shear stress between rivet holes can be carried up to the ultimate shear stress Fsu at the material.

Cll.38 Cll.34 Example Problem.

DIAGONAL

SEMI~TENSION

FIELD DESIGN

The foregoing explanation can be better explained or clarified through the presentation at an example problem. Assume that we have a ruaaIage with a structural arrangement as in the example proble or A20.5 and Fig. A20.3 and A20.4 ot Chapter A20. Also, assume the moment ot inertia and neutral axis for a linear bending stress distribution to apply. These values are given on page AZO.8*. Also let it be assumed that the supporting rings are attachsd to the skin and spaced at 15". The rings are 1" x 3" X 1" 24ST ~ sections, .040" in thickness (see Flg.C1l.43).
An analysis wlll be made of stringer 1t3, skin panels 2-3 and 3-4, and the ring. It is assumed also that an axial skin splice occurs along stringer i3 and a vertical splice along the ring to show a check or the fasteners.

634,000 29,g x-

2-x

61.6 + 113 = 175 lb .Ztn.

_ 11,700[.167(38.3) + .169(36.4)+.216(32.9)] %-4 2382 + 113

96 + 113

= 209 Ib./ln.

The diagonal tension effects will now be calculated assuming average q's, bending stresses, etc., for the panels 2-3 and 3-4.
QAV2-4

=175 ;

I t92

209 = 192 Ib./in. .1 = .QQQ. ps 1. 2 0392

sAV2_4

#3*2

l.-- e-". -".~'! \.-15~ 0" 1< . .~i~; _ "0 ..... '{\.~.\.. ...
.')1

~ ~",

Ring "b" Ring "a" M=1,387,OOO"jt

The average critical shear buckling stress will now be calculated USing equation (75). First, it no skin buckling occurred the average compression stress in the ~HO ?aTIels would be, apprOXimately, the stress in the stringer between them, thus f'Cpanels

61;,0
v

0"#

!14
11,700*,

= t pAV. = 20,400

PSi.

d 15 h""7.2S",,2.07

then the constant, B, would be

B - f c _ 20,400 : 3.40 6,000

Fig. Cll.43

First the internal loads in the stringer and skin panels are deter.nined. The "average" loads at the middle of the bay will be used in this example to get the diagonal tension effects. At the center of the bay, M = 1,475,000 in. lb. T = 634,000 In. lb. 'l = 11,700 lb.

-r;-

The critical pure shear buckling stress equation trom Chapter C9 IS,
n~

F SOT

12

(1

- V s ')

(.!<.),
b

From Chapter C9, kg = 16.9 ..

O(

= n lll x 16.9 x 10,600,000

12 (1 - .3")

c.e ':.

(.032)1lI =

7.25

Max. stringer stress occurs at Ring b:_:'!Z r P!".AX -r-"I

3140 pe i .

1,563,000 (32.9) = 21,600 I 2382 ps .

For pure compression the critical buckling stress equation from Chapter C9 1s,
n lll k
FCc r

E
kc = 15.9 (rr-cm
ChaP. tar C9). a
r-:r""",~ C:.J.~
A

= 1,475,000 (32.9)
2382

= 20,400

psi.

= 12 II : Ve') I;)'

Shear flows in skin pane1s:-

=nlllx 15.9xIO,600,OOO {.032):lI= 2980 s i 12 (1- .3') 7.25 --p-

'lz-3

"T" + 2A ' where A consists of area

or 26 triangles
2382
+ .169(36.4)1.

VQ

Thus, A

= FCcr/FScr = 2980/3140

= .950.

= 11,700[.167(38.3)

A non-linear bending stress distribution can be used. but it also 15 affected by the diagonal tension compressive stresses and involves considerable iteration. Linear ones are often
used,

Since buckling stresses are below the proportional li~it Stress or the material, no Plasticity correction was necessary as Is usually the case in thin walled structures. Thus stress ratio (A) could be obtained directly
as kc/ks'

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.39

Next: A . + 4 !2(~) _ -

ERG

8,300 (1.0)

10.6 X 10'

= 782

10-'

Now solve for "new" a USing equation (95)

(1036 + 2450)10-' + 782)10-' l(7.25). + 24 30

_ 3.40 + j(~)' + 4 .950 .950 2

= 0.822

tan a

= .907
= 42.2 0,
which 15 less than the assumed
45.

and, from equation (75)

rScr2_4 = FS c r x

a
= 3140 (.26)

He

= 815

psi

which is quite small, due to the presence of compression. Next,

7.36

As a second trial assume a = 42.40.

~= r

1.84 (rr'om Fig. Cl1.36) 6000/10.6

= 1.84 x

10'

= 1041 x
x 10'
X
ll

10-' = 2505

0STR

= 20400 + 5570

(cot <l)/10.6

x 10-'

now

can be determined:
300

f:

RG

= 8,300xtan a/10.6xlO = 752

_

la-II
.84

= 300

~0032)

(2.0)'= .64

tan

a - (1041 + 752)10-' + .00243

(1041 + 2505 )lo-'

From Fig. Gll.19, k = .75 The expressions for stringer and ring stresses can now be written in terms ot a: SUbstituting into equation (85) and uSing f PAV for the first term on the right Side,
20400+ , .75 (6000) cot <l 180

whence a. = 42.6 0 as against 42.4 0 assumed. The accuracy of the theory does not ~ t a closer check , tnus CI. will be taken as 42.40 The corresponding stringer and ring stresses are as follows:
r

STR

= 20,400 + 5570 (1.092)

(7.25)(.032)+5(1-.75)~6)

= 20,400 ... 6090

TheSe are not added tor local strength checkS. They are added for a col~abilitY check. The term on the right is the average stress, forced crippling, due to the diagonal tension effects. . .

= 20,400
this 1s the

5570 cot a stress in the stringer.

~average~

Substituting into equation (89) tor rings, and us tng ARG = (1~+3~+1")(.040~) ~ .20 m..", 75 (6000) tan a

(15i~~032)

R O Also, trom equation (87),

.75)
M__

8,300 (.917) = 7,610 psi .

+ .5 (1 -

= 8,300 tan a.
a will now be determined, by successive apprOXimation using equations (94), (86), (92) and the above expressions for stresses in them, and also Fig. C1l.36 for equation (94). Since there 15 so much compression involved, assume a = 45, to start With. From Fig. Cll.36
eE _ 1 83 f s -.
E

6000(7.25)(.082)(15)'(.75)(.917) .13TR24 (30) 299 lb./in.

and rb = MBTR c = 299 (.532) I .0306

5,200 ps i .

STR

= 1.83 (6000) = 1036 x 10- 11

10.6
X

10"

The rivet load/in. along the axial splice is, tram sq. (96),
Load/ln.= 6000(.032)[1'

_ 20,400 + 5570 (1.0) ESTR 10.6 x 10

= 2450 X lO-e

.75(.~4) -1)] =
243 lb ./in .

.. Since *:;:00 2.0, use 2.0 (See page Cll. 33)

ell. 40

DIAGONAL SEMI-TENSION FIELD DESIGN

and, along the circumferential splice, tram (97)

Load/in.- 6000 ( .032l[1+ .75(

thus
20,400 +6090 39,500 RINGS .67 (surr i c rent )

.6~5) -1)] =
262 1b ./in.

From the above stresses, fMAX/! ratios, and the allowables, checks tor adequate strength can be made as !ollows:STRINGERS:-

The rings attached to the skin are Subject only to the forced crippling stresses, fRG. Midway between stringer junctions thiS action is a maximum and for adequate local strength,
!RGr<.AX

,,

At the junction with ring (b),

!STR

=
c

(21,600 + !b) + 6090 psi.

1.0

oo

For adequate local strength,

!

+!
F

b ...

r
ST.: 1.0 F

USing the data in Fig. C11.21, !RGnAx !RG

ST

ST

= 26,OOOka/3(t~T)J./3

7,010

x 1.0 =

7,610 psi.

i'

= 26,000(.75)'/' (.~)'/' = 24,900 ps i

(Also determinable trom Fig. C1L38). 23,100 . 33 (su!!iclent)
7,610

Thus,

21,600 ... 5200 6090 39,500 ... 24,900

=.

92 (surrtc ienn)

At the center at the stiffener,

!8TMAx

!STrlA;< !ST ( ! ) ST
tSTMAX

trom Fig. C4.21,! Therefore,

!ST~JlX

-1.02.

The same compression stress (aDd load, P = ARG) exists at the stringer junction. It the ring is notched to let the stringer pass through, as is usually done, the net section at the ring must be made capable at carry1ng this load, which is located at the centroid ot the un-notched ring cross-section. Usually this means some "beef_upn, locally, around the notched section; sometimes Incorporated into the clip attaching the stringer to the ring.
tRG

ST

= 6090 (1.02) - 6210 psi.

Actually the rings, like the stringers, are subject to the average ot the shear stresses in the panels fore and. art ot them. These have been assumed equal in this example.
SKIN

then we get, using the Interaction formula,

20.400 + 5200 +

39,500

6210 - 90 (surr rc i errt ) 24,900 - .

The allowable shear stress taken fram Fig.

Cll.42 15
F s = 21,800

Checking as a fixed end COlumn,

p

!l = ./ro06 = 412 = VA .18

Using the method at Ret. (3)

FSALL = 20,SOO (!rom Fig. C11.41a ar.d b)
~

L/Zp

= 2(.412) 15 = 18.2
Thus

.38 (!rom Fig. C11.40)

F ~ 39,500 (very short column range) CO L.

F s = 20,800 (.65 .... 38) = 21,400 psi.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

CII.41

Either case shows there to be a large margin of safety for the actual Shear stress of f s = 6000 pSi, (average value shown). (Panels nearer the neutral axis will, ot course, feal a larger shear stress.)
FASTENERS
Any tasteners used in this area must, at course, be able to tr~ster the load/inch at the splices as previously calculated. This criterion might not design the spacing however. T::Iey must also ha va a tens ion allowable, when installed in .032 skin, at, equation (98),

and the margin at safety against collapse is,

- 1-

3G,900 _ 6,000 - large

other panels, ot course, might have a larger shear stress and a smaller allowable FSI NST and thus be more crttical, but there is toc much stiffness available, as seen, to expect any instability troubles.
LONGERON TYPE SYSTFl'l
Cll.35

Tens.All.coad/lnch' .22 (.032)(62,000)

=435

lb.

From Fig. Cll.37 it can be seen that 1/8~ fluSh head aluminum rivets in a dimpled .032 skin at .50~ spacing should be adequate, or l/8 w brazier head rivets at a spacing at .~5w it a ~fluShw joint is not required. The tension requirement 1s most severe in this case.

GENERAL

I~TABILITY

The longeron type of structural system. 15 somewhat simpler tran a total anajyata standpoint. This is, at course, primarily because there are tewer members carrying the axial loads and not as many shear panels with varying shear loads. This type at structure may, or may not, be the most opt1.mum. arrangement trom a weight and manutacturing cost consideration tor a particular airplane. But this is not the SUbject ot this discussion since it depends upon an opttmization study. The methods at analysiS presented here would, however, have a place in the calculations behind such a study.
Sane typical types at longeron structural systems cross-sections tor a fuselage are shown in Fig. Cll.44.

A check against overall instability ot the at rings and stringers can be made USing Fig. Cll.39. This Is analagous to the column stability check tor the uprights at a plane web beam. In the case at curved webs, however, no Whelp" 1s gIven by the taut skin in preventing instability.
Wnetwark~

For this example problem the values used

in calculating the shear stress in the sk1n at

which wcallapse" due to instability would occur are: PST 414 (Stringer S. at Fig. A20.4 in Chapter A20)
PRO

CLOO
(.)

(b)

{c}

FIc. Cll."
Fig. (a) shows the minimum arrangement, as to number of longerons, since at least 3 axial load carrying members are necessary tor equilibrium when bending moments in more than one plane are involved. This arrangement, however, has a disadvantage in that it is not a "fail safe" deSign. This means that the fat lure at anyone member will not leave a structure capable of stHl carrying some arbttrary percentage (usually SC% to 6/'%) of the deafgn ult1lllate loads. The system shown In Fig. Cll.44b Is capable at dOing this and is, theretore, the min~ type acceptable tram the "rail sare~ standpoint (4 longerons). More longerons may be used, as in Fig. (e), occasionally, depending upon other factors or design and manutacturtng. The longeron system, however J requires more closely spaced rings than does the stringer

= .819

(Both of these radii of gyration are gotten by including a full width at skin acting with the stringer, we = h , and with the ring, we = d , per Ret. (3). Then
(PST PRG)'I'/s xlO'"
(dh):l.ja R.3/'"

(.4l4x .819)'1'/8 xlO'" (15 x 7.25)'/' (3G)'/'

x 10 = 29 133

= .386
From Fig. Cll.39
F SINST

Ec

3.0; hence FSINST

3 x 10.3 x1Q

1I ;>

30,900 pSi

Cll.42

DIAGONAL SEMI-TENSION FIELD DESIGN

~entioned

system for strength-weight efficiency, as before. The optimum spacing of the supporting rings, .or- frames, is determined after many trial calculations have been made involving different gauges and sizes and spacings of rings and thicknesses of skins.

As in the previous discussion of stringer construction, the rings support the Skin, diViding it into smaller panelS, lengthwise. They also support the longerons, similarlY to the uprights (stiffeners) in a plane web beam, against bending due to unequal tension tield "pulls" by the skin on each side of the longsron. There are no "tloating" rings in a Longer-on structural system..

and (A) is determined for a curved panel of length "d" between rings and height (h) between longerons measured along the circumference as in Fig. Cl1.42. (B) is the ratio of the compreSSion stress to the shear stress, (fc/f s), for the particular loading condition being investigated. The compreSSion stress Should be calculated as if the panel being calculated had not yet buckled. Then, as in equation (75)

gives the reduced shear buckling stress, f scr' When tension strains, rather than compressive ones, are present with the shear we have, as in equation (78) RT::l.O+ and then f Scr = RT F Scr as in equation (79).
3.

After buckling, the skin in this system has tension diagonals Which, rather than being "flat" between closely spaced stringers, instead lie on a hyperboloid of revolution. That is, they flatten diagonally between the closely spaced rings. This action is discussed in Ref. (8) and the reader should consult it for the basic theory as presented by Wagner originally. In this system the ring spacing "d" is such that d c h; see Fig. CIl.34.
Cll.38

tt
2Fccr

Next, the loading ratio, fs/!scr can be calculated using fs as determined in (2) above. cr 4. Following this, the diagonal tension factor, k, can be obtained from Fig. CII.19.
5.

The engineering procedure for calculating stresses and allowables tor the longeron system is somewhat similar to that used tor the stringer system. The reader will note the differences.
1.

The total axial stress in the longsron can now be written as

First, at any bay being checked, determine the prtmary internal load distributions in the longerons and shear panels due to the applied loads. This can be done as in Art. Cll.34 USing the engineers theory of bending in most cases. In other cases where jUd~ent and experience and the nature of the structure indicate it, the ~ethod of Chapter A8 may be used in determining the primary load distribution due to the applied loads.
2.

tL = t p r D.T. k, t s, cot ul. = tp 2AL .25 (l-kl)Rc hl.t1. cot Q. lI k. t s. 2AL .25 (l-Ka)Re hat ll

--.
--.

- - - (104)

Next determine the critical shear buckling stresses in the skin panels. Since compression stresses are nearly always also present in practical situations, pure shear buckling does not cccur. Thus, as discussed in the case for stringer design, some rational interaction must be used to obtain a "reduced" shear buckling stress. This can be done, for example, by USing some "average" compression stress in the panel, weighted toward the high side for conservatism. Thereby the interaction ~ethod of Article Cll.32 can be used where
4

whe~e

f p = prlmarJ longsron stress from Step (1) (+) if tension and (-) if compression AL = Longsron Area
and

k, f s, cot u, Rc (or RT), h, and t are as previously defined. One set, subscript (I), 1s for the panels above the longeron and the other, (2) for the panel below the longeron.

, as in (74)

The average stress in the supporting ring (or frame) due to di~gonal tension effects is given by the following formula (and note that this 1s different from the strtn~er case)
6.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.43

k f
~

-'-HG

--.
ARG e
Qt
~

s tan
.5 (1

t:o.

- - - - - -

(1051

- k)

The angle, u, ~ust satisfY the equation, trom Ret. (3), and this is different from the casa for stringers,

here, as in equation (58) ,

ARC e

tan' u

~G
1

e - eL ----..,-Oi----tan a. e eRG

1 (d),

"8 If

(106)

(~)

where, as be:Core, from equation (94)

:This rorauja rs s im j ar- to the one for the 'a:factive n area at a single uprlght in a plane vee beam system, (Art. G11.20), where

e =

f: [Si~k2U
TL
t ~ -..B&1.
8 f (f

,
1
i

srn z c (l-k)(l-VI]

I
,I

or use Fig. Cll.3S for bracketed expression.

eL ~ L

e : p

d1~tance

from c.g. of

~ing

to the skin

from equation (104)

radius of gyration of the ring (crosssectIon) about an axis parallel to the sk1n.

eRG

(f

RG

from equation (105) )

Equation (105) assumes that k, t s ' d, t, etc., are the same tor the panels on each slde of the rI~g. If not, then some average-value should be used, or else fRG must be written as the zum of 2 diagonal tension effects (:or each side p~~elf-as was dene for the longeron 1n (5) above. The nmaxl~um~ stress, tRG1t~ , in the ring wl11 be, as before, eq. (59)
fROm.x
(f

Equation (lOS) must be solved by successive approximation as was done in the stringer system discussion.

8. Strength cnecks can then be made as follows, USing stress ratios.

a/ Langerons;

r- ~ ----r:- cc
L

r P -'- f O.... ~

1. 0

I I
I

I I

,
r
e

RG

where is obtained trom Fig. Cll.21.

7. T:~e next problem is the determination at the ar~18 of diagonal tension, Q. There are several ways in which a can be evaluated. Four are as tollOws:-

where r ec 1s the ~natural~ crippling strength and F 15 obtained trom F1g. L C11.38. Frequently, in the case ot thicker longerons, the following is used as a strength check:

t p '"

f D T

cc

1.0

a)

When there is a si~lficant ~ount or compressive strain, s1m~ly assume a. '= 450. there 1s no, or very little, ax1al strain present (only shea~) use Fig. Cll.45 to obtain a.
~ben

That is, the primary and diagonal tens10n effects are simply added and Checked ag~inst r ec .
0)
~

b)

~RG ~
'RG

1.0

c)

I: sufficient tension stress (or strain) is ~resent to prevent ouckling, rr-om eq . (78). then, of course, 0. "" 00.
Calculate a. by the ~ethod of successive approXimation as follows. T~lS ts the most tedious ~ethod and (a) and (b) above will suffice in most cases, especl~llY tor prell~l~ary design.

d)

It the rings have stresses in them due to loads other than diagonal tension (I.e. bulkhead type loadS, see Chapter A21), then an interaction equation is used.

~
?cc

... t?to FRC

1.0

Cll. 44

DIAGONAL SEMI TENSION FIELD DESIGN

where other
CI

o is the stress due to loads th~~ dl~sor.al :ension effects.

Shear Clips

Skin

,I...

1. 7 -oj

-oj .50f-

Ring

Skins: Strength check: r ~~ 1.0 ,s where F's is obtained from Fig. Cl1.42. Pennanent Buckling: Usually no perm. buckling is allowed at Lrnt t load.
fSL'~I"
F
-.~
c

~.O-.j Ring Cross Section Splice Longeron A '" , 0787 ~n.: Plate ~~",,~ Cross-Section I "', ~252. m, Detail "A" A",,60in.a .0 "'.:J65 In. I "',197 in.1o Clad 7075-T6 Sheet Mtl. p .580 in. 7075~T5 Extruded MU. F cy=70, 000 psi Detail "A" E e "'10. 5x 10 11I

~5 JE:'15,I:JL'032(Typ) J.1, it",. )~o 1,=

=

= 1.0
b

Skin. ' 025" 7075-T6

s p. b.

where Fs
d)
G~neral

is obtained trom Fig.

.

Cross-Section and Details of Members Fig. Cll.47

Cl1.46. ?

T Instability:
1. 0

~ 600,000 tn. lb. (reversible) V = 18,000 lb. (produces camp. in upper

,,---"--- =
F S1NST

!s

skIn)
<

These are the loads at the middle of the bay being checked.

where FSINST Is obtained fram Fig. Cl1.39. It Is recommended that a margin of safety ot .15 to .20 be malnt~ined here. Note in Fig. Cll.39 tr~t no effective skin is used and tuat PLON(} is at most .34. e) The loads on rivets at splices, etc. are the same as for the stringer system, eq , (96) and (97).

52.3" 23.1"

( ,

{'

I 1.. -t-

"'\'r "
I

33. 0"

I Neutral

I 9. 9"

J...
4.8,0
Geometry Cross-Section Area 2830 in." Fig. Cll.48

;s;;r~-

13.2" ....L

Effective section in
Bending. Skin Shown

The procedure outlined in (1) to (8) above can best be illustrated t:1rough the use of an example problem.
Fig. Cll.37 Example Problem.

'Dashed" is not Effective IN. A. :a 1895 in. 10

Consider a longs ron type fuselage structure having a cross-sect i cn as shown in Fig. Cll.47. The details and section pro~ertles of the longerons and rings are included in the fIgure. The properties of the structural crosssection are given in Fig. Cll.48. The properties shown are for the case where the bending moment causes the upper skin to be in compression and the lower in tension. The upper skin thus buckles out early, as indicated. For this example problem assume applied loads are,
tr~t

1. Using the geometry and properties of Fig. Cll.48 and the engineering theory of bending, we rind the primary internal stresses to be,

long.

=]g.= -2,135,000(33) =
I 1895

-37,200 pSI (comp.)

:eStop skin

= 2At = 2(2830)(.025) - 4240 pSI

+

600,000_

fSside Skins"" - 2At

VQ.
+ ZIt

= 4240

the

18,OCO(2x .60x33) 2 (1895)(.025) = 11,780 psi

M = 2,135,000 tn.lb. (comoresSion in

upper Skin) flower long.;::

~ =

-2,135,000(-13.2) = 1895
+14,800 (tens.)

ANALYSIS AND DESIGN OF

FLIGHT VE

ICLE STRUCTURES

en.as

28..

Check top skin panel for t s cr and k. Use properties ot Fig. Cll.47 and the

300 t h = 300 ~0025) (2)'= .5

R d k

applied loads to get R,.., eq. (74). The skin coopresslan stresses will be "fictitiOUS" since

= .81

the skin buckles out early, but will give the proper constant, B, for interaction. Assume fCskln at a point 2/3 up from upper langsron to top surface as governing compressive buckling.
fCsktn

Next, repeat steps 2a-4a for the side panels. This involves an arbitrary interaction to serve as a criterion for bUCkling undercombined shear and compression. 2b. The stresses at the top (upper longeron) and bottom (lower longeron) of the side panels are -37 , 200 (comp.) and 14,800 (tension) respectively. This gives an average "fictitious" stress of -11,200 axial and 26,000 psi bending stresS (eqUivalent). The actual reduced buckling stress for the side panels could be obtained from an interaction for these axial and bending stresses. Another way is to use an axial s tress only, "weighted" on the high side of the average, arbitrarily, to account for the effect of the bending stress. For this problem a compreSSion stress value half-way between the average, -ll,200, and the maximum, -37,200, is used. Thus t c

=T
_

MZ

-2,135,000(27.7) = -31 200

isss
600,000

PSi (c;mp.)

s - 2At - 2(2830)(.02:5)

= 4240 psi

Then, rr-ca eq , (73),

ror this u'.:,p~.:- ';J2_~31 .~~" ::::r.g (c l , 41.8" wide (h) and R =< 3(1". -:ii>? cr i t t ca l pure shear and

compression tuc~llr~ stresses can be found from buckling equations in Chapter ce.

= -11,200 ~ 37,200

=_ 24,200

pSi

kc

::=

From Chapter ce , for Ql.IT panel djnenai ons , 10 and k s = 12.83, thus
:s

Feer
FS

n x lOx 10.300,000 (~)'= 2310 i 12 (1 _ .3 2 ) 5 _ _ ps

12 (1 - .3')

then, since t s = 1l,780,

c B = _ff_.

= -,2"4",;;20,,,0:;. =
11,780

2. OS

er

= n.a x 12.83 xlO,300,OOO = ;:~: = ~~~g =

(.Os25)a;: ~ psi

and, as before,
A

hence, from eq. (72)

A .774

.774

Hence
+

Then, from eq. (74)

_11+
A

+ 4

---+

7.35

.774

/(7.3.5)' + 4

2
and
f

.335

.774

= .105
and

Scr =

Rc

FScr = .335 (2980) ;: 998 pSi

3b.

The loading ratio is then

11,780 = 11.8
998

f Scr ;:
3a.

~~

FScr = .105 (2980) ;: 313 psi

, is
4b.

The loading ratio,

"a f Scr

From Fig. Cll.l9, using 11.8 trom (3b)

k = .79

= 3W =

4240

13.'54

5. The upper longer on stress can then be gotten rr ca eq. (104); * 5inCe:a- =- 2. 0 use
2.0 (See page ell. 33).

4a.

k is then, from (3) and Fig. Cll.19, for

Cll.46

DIAGONAL SEMI-TENSION FIELD DESIGN

-27,900 (.967)

10.3 x lOa then, .81 (4240) cot

2 (.6)
a~

= -.00262

(41.8)(.025) + .5 (1-.81)(.105) .79 (11,780)cot a,

ten ~

a~
E

e ~ - eL
RG

+ - (-) 8 R

d. tan3. a.

52.~((~~25) +
= -37,200 2970 cot
a.~

.5 (1-.79)(.335)

.00215-(-.00262)+~ (3~),(.705)

-'."'00"'2,,1"'5'--_(:--...0.,,0"'0.5"'0,,7-')___ = 1 .00

- 9810 cot c ,

Sa. The stresS tn the rings supporting the upper skin is obtained tram eq. (105). First,
AP.G

ThUS, all; = 450 which is close to the 440 assumed. Let a. = 45. This bears out the statement th-~t any signif1cant com?reSs10n stress (or strain) forces u towards 45 0 Had there been no significant average axial strain (or stress) present a could be gotten from Fig. Cl1.45 which is based on pure shear (no axial loads present). Using this for the above side panels, for example, we would proceed as follows:

~G : --=':-1 + (~)' p

.0787 = -='-"';",...-: .0285

1 + (.?5), .08

Then, .81 (4240) tan

a~

= lO.3xIO
11,780

.0285 - (1 81) 5(.025) +.? -.

(..)' -- 23. 5 30

As discussed earlier, this assumes the upper pa~els on each side of the frame to have the same shear stress.
6b. The stress in the rings supporting the side panels is, similarly,

Then from Fig. meters,

a
Th~9

Cll.~~,

using the above para-

36 0 (tor no compreSSion)

5( .025)

.79(11780) tan .0285 +

a~

.5 (1- .79)

the etfect of compreSSion, as previously calCUlated, is Signif1cant in forcing u towards 45, as frequently assumed for Simplicity.

a~ and a.~ in the above equations refer to the upper panels and side panels respectively.

7. large

Angle of diagonal tension. The compression in the upper panel is so th~t a.~ can be reasonably taken as 45. same is probably but this will be successive (106), (105), (104) FrOm

8. Now that u has been established as 450 tor the upper skin panels and for the upper portion of the side skin panels, the stresses are available (from (5) and (6) ) and strength checks can be made. '

a.

Longe rcns :
fp

For the side panel the true, in the upper portion, checked USing the method of a~proximati8n and equations
and (94).

=-37,200;

D T.=-2970(1.0)

-9810(1.0)

= - 12,780
Fcc = 55,000 Using Fig. Cl1.38, and tne r ormuja for "elf in the figure,
F RG

Ass~e a. = 440 (and u ~ 450). Fig. Cl1.36 and equations above,

=Nx c

= 43,500 (1.28) = 55,600

= _c = -.00361- .00028(1.19) _ .00095 (1.19) = -.00507

-.

then USing the interaction equation sug>!;estad, 37,200 55,000

+

12,71:30 55,600

.801

<

1.0 (adequate)

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.47

b.

e. The ring under the side panel skin will be most critical (fs 1s larger);
fRO: 27,900 (1.0) : 27,900
fR~

General Instability Using Fig. Cll.39, with PLO NO :: .554 and PRG ::: .565 (Using no effective Skin, per the figure)

: 27,900 (1.15) : 30,500 (Using

Fig. Cl1.2l)

.39 ( PRO ) ' /

(dh ) ~/Ji R3

/ to

X 10"

=11.35
30,900 ps I

Then, USing Fig. Cll.38,

From Fig. CII.39

FSINST.
Thus

ror R :: 30" N = 19,200

F NxC RG:: M.S.

=3.QxExlOFSI NS T
fs

::

19,20Qx1.1S"':: 22,800

22,800 _ 1 : - .25 30,500

M.S.

: 30,900 _ 1 : 1.62
11,780 (large)

Thus the ring 1s not adequate. Either more area (thickness) 1s required or a closer ring spacing (or both) are needed to lower t RG Skins:

. ,.

Rivets
The rivet requirements can be checked in

the same man.~er as was done for the stringer type structure along any skin splices and tor the tension field npryin~" forceS. The load transferred between the longer on and the ring can be calculated as PRG XARG ' from (Bb) and (6a) RG e ::: 27,900 x .0285 : 795 lb.
f

The side panel r~s th~ highest stress and will be, therefore, more critical from an ultimate shear strer4th than the top panel.
fa

= 11,780

PSi

From Fig. Cll.42, for k = .79 and F'TU 72,000, Fs ALL 25,000 psi. M.S. =

.=

(25,000/11,780) - 1 : 1.13.
d.

Cneck

f~r

Permanent Buckling

Usually th~re 1s the requirement that no permanent s~in buckles shall occur at limit load. This 1s checked as follows:
f5 =--:: ll11it 1.5
USin~

Tnis load is actually carried by the rivets attaching the outer ring flange to the Longer-on flange (next to the Skin) and also by the gusset action of the skin at the ring-longeron junction. Two fasteners, staggered if necessary, should be used to attach the outer ring flange to the longeron r lange . (See discussion ot joggles in Chapter 03.).
Cll.38 Summary.

is

11,780 :::7,840p51 1.5

rig. Cl1.46
4 (2,980) x 10e ::: 1.87 10,3 X ic- x 62,000

4 F'S:r x 10" E:.,x Fey

From these examples inVOlVing both stringer and longeron type construction it can be seen that the effect of axial compression stresses (or strains) alon~wlth the shear stresses in the panels 1s tWo-fold:
a)

FSP,B _
- - - - 2.8, F SP

i"Scr

: 2.8(2980) : 8,340 psi

M.S P. B

.=

_ l

-7'40

= 8.340 _ 1 : 06-

"

.0

It brings about, through interaction, earlier buckling of the panels than would result from shear stresses ~lone. The result is, at course, a higher value of k and, hence, of all the ensuing loads and stresses tpat are a function of k.

Cll.48

DIAGONAL SEMI-TENSION FIELD DESIGN

b)

The angle of diagonal tension is forced to approach 450, more than would result from shear stresses only.

CU.39 Problema for Part 2.

1.

The effect of tension stresses is just the opposite and can thus be conservatively ignored, or evaluated if desired.
2.

In the example problem ot Art. Cll.34, what M.S. would exist it the ring were made of .032 2024 (instead of .040). (ASSume a. :: 450). In the example problem of Art. Cll.34, how wide could the ring spaCing, d, be made and still show a pOSitive M.S. tor the ring. (Assume u :: 450.) In the example problem of Art. Cll.34, how much additional torSion, T, could be applied before the ring would show a M.S. :: O. (Assume a. = 450.) Repeat example problem Art. ClL34 USing applied loads
11. :: 2,600,000 in./lb. T =0 V = 40,000 lb.

From a time-saving standpOint, in preliminary design, an arbitrarily large value of k and an angle of diagonal tension of 45 0 can be assumed where significant axial compression strains are present.
TIle exact magnitude 0 f the various dfagona tension effects throughout a network Of Skin panels defies simple evaluation from an analytical standpoint. This is particularly true when both shear stresses and axial stresses change from panel to panel as in most Dractical structures and loadings. No simple analytic expressions are available. But some rational approach is necessary to complete the design, or specimens tor test progr~~, and the approaches given in this chapter represent one such procedure. It margins are extremely small, element tests tor sUbstantion are in order. The reader is encouraged to consult the references for a ~ore thorough understandln of the basic theory and its limitations, particUlarly with regards to areas where sub_ stantiating test data is relatively meager.

3.

4.

AsSume the general section properties (I neutral axis location) remain the same as in Cll.34.
and

5.

In the example problem of Art. Cll.37

a) What (standard) gauge or 7075-T6 sheet

aluminum would the- ring have to be to show a minimum positive M.S. b) What ring spacing, d , would be r-equtr-ec for the .032" ring to show a mintmum positive M.S. (M.S. - 0)

The stringer system is usually found, for example, in fuselage structures where there are relatively few large Wcut-outs" to disrupt the stringer continuity. This is more typical of t~nsport, bomber and other cargo carrying aircraft. The longeron type structure is more efficient and suitable where a large number wquick-access" panels and doors and other "cut-outs" are necessary to service various systems rapidly. These would "ChOp-up" a stringer system rather severely, making it qUite expenSive trom both a weight and manufacturing standpOint. Therefore, longeron systems are more usually found in fighter and attack type aircraft, and in others With unusual features. There are alSO, of course, other factors intluencing the choice of structural arrangement. Some further notes concerning this general subject are inclUded in Chapter C3. This includes "beef-upn of panels and axial members bordering cut-outs or non-structural doors, which is also related to "end-bay" effects discussed in Cll.24.

6.

Repeat the examPle problem using as applied loads.

11. T

or Art.

ClL37

= 2,400,000 in. lIb.

V :: 28,000 lb.

=0

(Assume section properties are not Changed.)

7.
!~t is required to eliminate any negative margins of safety in Problem (6).

CU.40 Problems for Part 1.

(1) The beam. as shown in Fig. (A) is SUbjected to a shear load at 8000 lb. as shown. Determine the margin of safety tor the given loading for the ro Lkowtng units: (l) web, (2) web flange rivets, (3) web stiffeners. Use NACA method.

(2) Same as prOblem (1) but with web upright spacing = 4".
* It
is not conservative to ignore the reducing effect of tension stresses on k if the compression stresses due to diagOnal tension are being relied. upon to reduce any primary load tension stresses in stringers or longerons.

(3) Design a semi-te~~ion field beam for the beams and loading of Fig. (B). Take web as 2024-T3 with minimum thickness being .020.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

3 3 x 16 1 S11 'Ifener iX'4 one Sid
'f'~JJ~++ ...

Cll.49

I21 . 4 ,
1

~eb

.. ... +++

(>

Ii___

8"~
+- +
-t'
+-~

11
12" 11

'" '~X1X3/32
\.. 3

,1/8

Dia. @ 3/4"

uprights and flange members to be 20l4-T6 extrusions. Rivets to be 2117-T3. Show calculations for at least four sections along beam length. Assume the beam braced laterally by .025 skin on top and bottom.

16

16

)--

..

. 04 Web 2024- T3
~++ ++ .....
i -,,

All Rtvets 17ST.

Fig. (A)

80,000#

1 ,
"
Fig. (B)

1xlx3/32

(4) Fig. (c) shows the cross-section of a single spar wing beam. Strength check the following units of the beam section. (1) web, (2) web uprights. The upright spacing is 8 inches and the design shear load on section is 35000 lb. web material is 7075 alclad.
REFERENCEtl

(1)

H. ~agner, "Structures of Thin Sheet Metal, Their Design and Constructlon~, N.A.C.A. Memo. 490.
H. Wagner, nFlat Sheet Girder Metal Web. n Part I - N.A.C.A. Techn1cal Part II - N.A.C.A. Techn1cal Part III - N.A.C.A. Technical

(2)

With Thin Memo. 604 Memo. 605 Memo. 606

70'lIDJTIlin. Design Loading

.

....,...T-,:;4~0#lin. i OTt f If J

(3 )

Kuhn, Peterson, Levin;- A Summary ot DIagonal Tension. Part 1. Methods or Analysts. N.A.C.A. T.N, 2661. Kuhn, Peterson, Lev1n:- A Summary of Diagonal Tension. Part 2. Experimental Evidence. N.A.C.A. T.N. 2662. N.A.C.A. T.N. 1347.

'.

180"----~

(4)

'05~~!~ 2 -*J_~ 3/16

(5 )

r
1...
5/8

(6)
(7)

Ebner, H., The Strength of Shell Bodies. N.A.C,A. T.N. 838,

N.A.C.A.
N.A.C.A.

T.N. 1197,
T.N. 774.

lrl
16
Web

1_1/2

(8)

Stiffener 26"

tz~3/8
T ..... V "'-,..",

.064 Web 7075 Clad

-i....I
5/8

1-

--1

Fig. (e)

l.

50
3.4 3.2 3.0 2.8 2. 2.4 2.2

DIAGONAL SEMI-TENSION FIELD DESIGN

I I

eE
f

"'stn2a + sin 20. (l-k)(l+i-")

~

2k

- 0.32

,
10
.8

2.0 18

..
.4

I.
1.4 1.2 1.0 .8 20

.2

o
Rivet

o
8
25
30
I
I I

r-Rtuld h

eene-

head (AN4551

-_301

~raf heed 'I"' (AN442l

35
CI, dec;

40

45

.......

i'eQd (AN4561

Fig. Cll.36 Graph for Calculating Web Strain (Ref. 3)

2
0

"'1@7

.2

.6

~" !hidu1ess Net diameter

Fig. Cll.37a Tensile Strength of Four Types of A17ST3 Aluminum-Alloy Rivet in 24ST3 Aluminum-Alloy Sheet. {Ref.3}

Approx.imate i~o.28

3.
I

50

,
I

'I',

:7
I

I
I
I
:

, ,.)/1
I

, ,

I I

,
,

I ,

8 0

2
Rivet

seeet

diameter

""""'"

.8

I .2

QEY
4
.0

,
8

Fig. Cf L 37b Tensile Strength of NACA Machine-Countersunk Flush Rivets of"A17S-T3 Aluminum-Alloy in 24S-T3 Aluminum A.lloy Sheet. (Ref. 3).

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.51

10

1.2
....._-_ ..
--.
-

.. _._.. _--.-

----.

~--~~---

:.:.~::.:..-:

"

-----_.~-~

_._~--

.-. -.

.403

en, 52

DIAGONAL SEMI-TENSION FIELD DESIGN

, ,
f i
I I
, , ,

;;;-

An

: :

.' .'
,

.,
I

I
,

,
,
,

I ,

.i

-,

<II ,

1.5
.4

6 6

,
I
,

,
I

I ,
,

:
!

,
I

I ,
i

t::.. Q.3tenll~+ OJI~II~

,,
I

,
o o
'0

"-.HAt:.V

",",

Fig. Clt. 40 Correction for Allowable Ultimate Shear Stress in Curved Webs.

"

"

ao

,.R ,.

35

,
,

11~RR

~ if

"
'

'0
5

~
ksi

,..-1;\

30
25

~
i

, , ,
,

I ,
, ,

I
,

,
, ,

,
,

I
,

,
,

, ,

,
,
:

, ,
,

, , ,

20

, ,
,

,
.8

~,k

20

25

30

35
"POT

40

50

.2

.4

.6

to
a ult
'" 72 ksi

RR....EL; Rs A
RG

ht

Fig. Cll.41c Alclad 7075-T6 Aluminum Alloy.

'" AST

Fig.

cu. 4la
2 20
,

30
,

,
,

,
,

,
, ,

25
, , , ,

~~
QpOT,deg
11+RR

"

fsal 1 , kat

20
15
,

,
, ,

, ,
,

,
, ,

,~ es
, ,

'0 .

45 40

'0

,
25
30

,
,

, ,

,
,

es
20

, ,
10

, ,

, ,

ss
"POT

40

,
6
.8

.4

La

Fig. Cll. 4ld Angle of Pure Diagonal Tens ion.

Fig. Cll.4lb 2024-T3 Aluminum Alloy. O"ult '" 62 ks1. Dashed Line is Allowable Yield Stress.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Cll.53

Cll.54

DIAGONAL SEMI-TENSION FIELD DESIGN

i i Ii WJ __ LLJifJUL-_l_LJjjJUL-_-LLLJ inn
APPROXIMATE VALUES OF THEANGLKOEDIAGONALTENSION

a...--..

FOR: PaNELS WITH' b",. d (LONGERON SYSTEMS)

-Cut-()ff- . --~---,.- .... - ----.--,-.-'-.--.

NOTES: . . . . . '

l)Thi& curve-ta:for panela Ioaded in.'_s~~:_. .' only, no,signiIicant axial, strains

.. "

...... _

2} The,cune. is an,~.~~~~~.~_~!~l~_~t~~_,_~ __ .

tau

II

ri'-""

. ',' -----.' . - . K+ eoaaca (l-k)(l+i!). s1IL-lI.a, +

----.-.--~

a:

.. -~ ,--

c.. 1 .At.Lbt~z:(t;'k} -._-.~

_ 3}

~"'. '. - -. :"'. :...~~. . ' . represemedon,-thia c~rveis,from;~z. to-l~.5~. --:-:-. ..-. ,..-;----~-:; Values, of: thiS parameter outside- this range ~for values: at f , <i.l_lI,: _ , 2,0) . . ..,.. _.-:-~~ s .... must beapplted:to' the :equaUoDabove (Note 2):: to; determmri!'il. -:.-::~~~~~--:-.=.--:::::-.:::--:::-~~.,::::::::_::;.::.~:_:_i Chance V01Jb!lt :a1rcrafta::-Ine.:::, _....._

K.~ siJtt_ 0. [~J.~ ..It)(l:+'!-J;} eC)s~_g. ..t. . . ' _. ".1.,...- __ '+T(iia!an .: ;..u ARG.,!d'T\l,kl ...

The'1:angeo!'1iLv~s-

-, t At;

.__ lJ_ . . .:

1.

10.0

100.0
--..,---...,....~.,...~

..-..- - . - - . ----..!L .fs-- R,-..

4-l'

_.__... _2

.. '-._---

-.

-- .. _----_._-~~~

234567891

3 4 5 67891

34567891

3 4 567891

4 5 67891

NOTES:
This curve good for fiat or curved panels. Conventtonal panel stiffening members 'must be attached to edge me mbers 2) Curve does not apply to panels with stringers. 3) Curve may be used for any material at any temperature provided proper values of FSCR ' E c & F cy are used in the parameters.

II

30
20

I
3
2

N.
I
,

1. 0 9 6 5 4 3
2

I
FSCR' Panel Shear Buckling Stress Ee FeY

I
,
'" Modulus of Elasticity in Compresston '" Compression Yield Stress

~,

,
I I

II I III

I III
.10 ,5

I
1.0

I
5 10

1 . 01

III IIII1I

I III 11III1 I I IIII

Ec x FcY

50

100

500

1000

4xFSCR. X lOll
Fig. Cll.46

CHAPTER Dl

FITTINGS AND CONNECTIONS. BOLTED AND RIVETED.

01. 1 Introduction.

The ideal flight vehicle str~cture would be the single complete unit of the same material l~volvlng one ~anuf~cturlng operation. Unfortunately the present day types of materials D1. 3 Fitting Design Loads. Minimum Margtns of Safety. and theIr illdthod J! working dictates a composite structure. Furthermore gener~l requirements ot As discussed in other chapters, limit repair, malntena~c~ and stowage dictate a loads are the maximum loads, which a flight structure of several main units held to other vehicle may be subjected to during its lifeunits by main or prua~ry fittings or cont~e, when carrying out the required ground nections, with each unit incorporating many and flight operations. The limit loads must be primary and secondary connections involving carried by the structure Without exceeding the fittings, bolts, rivets, welding, etc. No yield stress at the material used in the doubt main or primary fittings involve more structure. The ultimate or design loads are weight and cost per unIt volume than any other the limit loads multiPlied by a factor of part of the aeros~ace structure, and therefore safety, usually 1.5 tor aircraft and less tor fitting and jOint design plays an important missiles. The structure must have sutticiant part in aerospace structural design. strength to carry the ultlmate or design loads without failure.
D1. 2 Economy in Fitting Design.

cost at inspection and rejections ot costly fittings because of faulty workmanship or materials are qUite frequent, thus adding greatly to the unit cost.

A blanket factor of safety for all types of fittings or load conditions is not logical. The manner in which a load is applied to a jOint often involves a dynamic or shock load, as for example joints or fittings in landing gear. Single pin connections often undergo rotation or movement between adjacent parts, thus pro?or economy o~ fabrication, the str~ctur~l ducing faster wearing away of mat~rial in oper~tion. Repeated loads often present a fatigue designer should have a good knowledge of shop problem. In ~n airplane or missile there are processeS and operations. T~e cost of fitting fabr'Lca t Lon and assembly var-ies greatly with the certain main fittings whiCh, if they failed, would definitely causa the loss of the vehicle. type of fitting, shape and the required tolerances. Poor l~yout of major fitting arrangement Thus the design fittin~ requirements or the military and ci7il aViation agenCies involve may require very expensive tools and jigs for many special or higher factors of safety. This shop fabrication and assembly. is so particularly in deSign involving castings. FittIngs likewise add conSiderably to the
01.1

FOT structur~l econamy the structural designer in the initial layout of the flight vehicle should striv8 to usa a min~um number of fittings parti~ul~rly those fittings connecting units which carry large loads. Thus in a wl~g structure splicing the mai~ beam tlan69s or intrOducing fittings near the centerli~e of the airplane are far ~ore costly than splices or fittings pl~ced f~rther outboard Nhere member sizes and loads are considerably ~~ller. Avoi1 Chang3S in direction of heavy members such as wing beams and ruselage longerons as these involve heavy fittings. If joints are necessary in continuous beams place them near points of inflection in order that the bending moments to be tr~nsferred through the joi~t be k~pt of small magnitude. In column design with end fittings avoid introducing eccentr ic i t res on the beam and on the other hand make use of the fitting to increase col~ end fiKity thus compensating some of the weight increase due to fitting weight by s~ving in the Neight of the beam.

The normal flight vehicle structure inVolves many parts whIch are joined together by various types ot connections. In general, an additional blanket factor of safety is required for design of these connections. ThiR blanket ~actor of safety is norma11y 1.15 to 1.20. The stress analysis of most connections or fittings is more complicated than for the prtmary structural members due to such factors as combi~ed stresses, stress concentration, bolt-hole tightness, etc., th~s an additional factor at safety is necessary to give a similar degree at strength reliability tor connections as provided in the strength design of the members being connected.
01. 4 Special or Higher FactorS of Safety.

4Su
,

._'~~..--:

. " c

---~---

'

01.2 01. 5 Aircraft Bolts.

FITTINGS AND CONNECTIONS.

BOLTED AND RIVETED.

inch diameter should be the smallest size to be used in any fitting. Bolts connecting parts having relative motion or stress reversal should have close tolerances to decrease shOCk loads. For bolts connecting members having relative motion a lubricator should be incorporated in the surrounding parts of the fitting; the fitting should not be drilled to prOvide lubrication. Bolts should be used in double or multiple shear i t possible in order to increase strength efficiency in bolt shear and to deCrease bending tendency on bolt.
D1. 6 Aircraft Nuts.

The aircraft bolt is used primarily to transfer relatively large Shear or tension loads from one structural member to another. ?ig. 01.1 shows three standard aircraft bolts in common USe. There are other types but they ~ill not be presented in this limited chapter on connections. Hexagon Head

crevis

Internal wrenching
Fig. D1. 1

Fig. 01.2 illustrates four standard steel aircraft nuts. Nut material is more ductile than bolt materIal, thus when the nut 15 tightened the threads will deflect to seat on the bolt threads. The aircraft strength in it so Castle n~t IS probably the nut. It develops the full of the bolts. The nut has that nuts can be prevented most cammon rated slots milled from turning.

The hexagon head bolt is an Army-Na'ry standard bolt made from SAE 2330-3.5 percent nickel steel, heat treated to an ultimate tensile strength of 125,000 pSl. The bolt head is 0: sufficient size as to develop the full tensile strength of the bolt. The internal-wrenching bolt Is a high strength steel bolt usually heat treated tram 160,000 to 180,000 psi. It 1s especially suitable for main splice fittings because of its high strength and the relatively small space required for the bolt head. The clevis bolt is referred to as a shear bolt because its head is not designed to develop the full tensile strength of the bolt. The clevis bolt is usually USed when a group or cl~ter of bolts is required to tranSfer a load by shear loads on the bolts. The smaller bolt heads thus save weight and also prOVide greater bolt head clearances. The clevis bolt will develop about One half the tensile strengt of the standard AN hexagon head bolt. The three bolts are also ~ade from al~~inum alloy for diameters over 1/4 inch. In many fitting designs weight can be saved by uSing aluminum alloy bolts. General
R~l~s

The Shear nut Is only one half as thick as the Castle nut, and thus has only threads enough to develop one half of bolt tensile strength. It is used With a clevis bolt which has a screw driver slot to limit the torsion in tightening the nut. The nut is also castellated tor cotter pin lock.
COTTER

SHEAR NUT AN - 320

PLAIN NUT SELF LOCKING AN - 315 AC 365

4

Fig. D1. 2

The Plain nut which is very seldom used in present deSign is used for permanent locations and is loCked by "peening" or riveting the end of the bolt over the nut, an operation which destroys the bolt protective coating or finish. Aluminum Alloy Nuts. Aluminum alloy nuts are not used on bolts designed for tension. Sometilnes aluminum alloy nuts are used on steel bolts on land planes to save weight provided the bolts are cadmium plated. If the bolt is used in places Where the nut is repeatedly removed, neither bolt or nut snoujd be aluminum alloy beCause 0: the danger of injuring the relatively so t't threads.

in Using Bolts.

Bolt threads should not be placed in bearing or shear. The length of the bolt shank should be such that not more than one thread axtends below fitti~ surface, which can be done by the use of washers. Bolts less than 3/8 diameter should not be used in ~jor fittings. For steel bolts, 3/16

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

013

Salf Locking Nuts. Self locking nuts are widely us~d in aircraft industry and there are a number of types on the market. ~lg. 01.2 shows one type of self locking nut involving a fiJer ins~rt. ~~en the bolt reaches the fiber collar it tends to push the fiber up because the hole in the collar is smaller than the bolt, and is not threaded. The fiber ring thus sets up a heavy do.vnward pressure on the bolt automatically throwing the load carrying sides of the nut and bolt threads into positive contact. Thus all play in the threads is eli~inated and a friction is set up between every bolt and nut thread in constant. This constant pressure which is maintained on threadS, provides the friction which prevents nut from moving under vibration. The use of the selt locking nut reduces the assembly costs as it eliminates the bothersome cotter pin which takes extra operatiOns at the mechanic and is very difficult to install on the nut in the ~ny JOints and corners at an airplane. In another type at selt lOCking nut, the lOCking torce is provided by the spring action at the upper part of a specially designed nut.
D1. 7 Bolt Shear, Tension &: Bending Strengths.

Table 01. 2 Ultimate Shear and Tensile Strengths of Steel Internal Wrenching Bolts (Etu 160,000)

UlL

Size
Dia.

Tensile Strength lbo. 190 9,820 15,200 20,600 27,400 34,800

Double Shear Strength lbs. 9,300 14,600 21,000 28,600 37,300 47,200

tnr,
Size
Dia.

Tensile Strength
Ibe,

Double Shear Strength Ibe,

1/4
5/16

a.

5/8

3/4

7/16 1/2 9/16

3/8

7/8
\.0

1-1/8 1-1/4

43, 600 63,200 86, 100 114, 000 144,000 180,000

300 83,900 114,200 149,200 188,900 233,200

sa.

Table 01. 2a Ultimate Shear, TensUe and Bending Strengths of 2024 Aluminum Alloy Bolts
SHEAR F su - 35, 000

TENSION Ftu-62, 000

1,059 1,975

BENDmG Fb-72,OOO

99'
1,717 2,684 3,868 5,261 6,871 6,697 10,738 15,463 21,046 27,489 01. 8 Bolts in

46
110

3, i89
4,937 6,663 9, 104 11,563 14,719 21,573 29, 520 39,759

216
373 59. 884 1,260 1,730 2,980 4,740 7,070

Table 01.1 gives the section properties and the ultimate Shear, tension and bending strengths ror AN Standard Steel bolts at room tem?erature. Fig. D1.3 shows the correction to be applied to the strength values in Table Dl.l when bolts are subjected to elevated temperatures. Table 01.2 gives the tensile and double shear strengths ot Steel Internal Wrenching bolts. Table Dl.2a gives ultimate shear~ tension and bending strengths tor aluminum alloy bolts. The value of Fb' the modulus of ruptur~, was determined by the method given in Chapter C3.
Table D1. 1 Ultimate Shear, TensUe &/; Bending Strengthll of AN Steel Bolta (Ftu .. 12S,OOO, F su " 7S. 000, Fb 180,000)

Combined Shear and

Tension.

When bolts are SUbjected to both shear and tension loadS, the reSulting strength is given by the following interaction equation:- (Ref.
i, 2)

x3 s" 'iT + bi" = 1 - - - - - - - - - - - - - - (D1.l)

where, x y a b

ISlze 01 pin
I

Area of .solid section.

in.'

or bolt

ULtimate stngle InUmate Moment of sllear strength Ulttma.te tellllUe Ben<l1!1g inel'tla 01 at lull strength Moment soUd., in.' diameter, lb. (in thread), lb. in. ills.

= shear load

= tension load = shear allowable load tram Table D1.l = tension allowable load tram Table 01.1

Figs. 01.4 and Dl.5 is a plot of equation for the various AN steel bolt sizes. The curves are not applicable where shear nuts are uBed. ~he curves are based on the results of combined loau tests with nuts tingertight.
D1. Sa BushingS.

"5 80

-~

---~-

100

.-----"-------,- ~.200 300 400 500 600

It is customary to provide bushings in the lugs of single bolt or pin fittings SUbjected to reversal at stress or to Slight rotation. Thus i t wear and tear takes place a new bushing can be inserted in the lug fitting. Steel bushings are commonly used in aluminum alloy

45.3

01.4

FITTINGS AND CONNECTIONS.

P Sa l
x

BOLTED AND RIVETED,

30

M.S. '"

'ow - tor Ptallow

.- - ,

Interaction Formula x3 a." .. b'" = 1 Where: x .. Shear Load y .. Tension Load a. '" ANC-S Shear Allowable b '" A.1lfC-5 Tension Allowable

z:

~""

"0

single bolt ritting lugs to increase the allowable bearing stress on the lug eince the bushing increases the bearing diameter l/S inch since bushings are usually 1/16 inch in wall thickness. It bushings are not useo on Single bolt connections sufficient edge distance should be provided to ream hole tor next size bolt in case or excessive wear Or the unbushed hole. If considerable rotation occurs a lubricator should be provided for a plain bushing or an oil-impregnated blli;lhing should be used.
01. 9 Slngle Bolt Fitting.

~""-
ptallow
~>

~"""

15

bh:,
a

Possibly the simplest method of jOining two members together is the use of a 6ingle bolt or pin connection. Such a jOint can transmit relatively large loads and yet the jOint 1s easily and qUickly disconnected. Fig. Dl.6 illustrates the four general methods of connecting two members by a single bolt. First the connection is made symmetrical about the centerline of the load on the joint. Thus is Fig. DI.6a the load P on the male part of the fitting divides equally and symmetrically to the two female plates or units which make up part of the fitting unit. It the male and female parts 0 t the connection are to be tied together by a Single bolt it is evident that the connecting plates will be weakened due to the bolt hole unless extra material is added at the bolt hole section. Fig. DI.Sa shows a fitting consisting at three rectangular plates or uniform section throU~hout fastened together by a single bolt. ObviOUSly the weak section for the plates in tension would be a Section through the centerline of the hole. If this section is strong enough to carry the given Loads, then the remaining part of the !1tting members are conSiderably over strength. To avoid this over-strength which means extra weight of fitting units, a Single bolt fitting unit is often made like one at the examples indicated in Fig. b, c and d of Fig. DI.6. In Fig. b, the plates are made constant thickness but increased in width in the Vicinity of the hole section. In Fig. c, the width of the plates is kept constant but the thickness of the plates are increased in the Vicinity of th~ hole section, and in Fig. d, both width and thickness ot plates are changed.
)

.~,

"

1St

' ,;

0 0

10

20 PS

allow

25

SHEAR LOAD IN THOUSANDS OF POUNDS

:Fig. Dl. 4 Combined Shear and Tension on AN Steel Bolts (Ftu = 125, 000, Fsa .. 75, 000)

Where: 130
120 110
xooShea.r Load y..Tension Load a..ANC-5 Shear Allowables b..ANC-S Tension Allow-

"',

Interaction Formula

..

I , , ,
,,

? +b"1'"

x"

y a

20

'0

01. 10 Methods of Failure 01 Single Bolt Fitting and the Allowable Failing Loads.
10 W 30 40
~

60

ro

90

100

SHEAR LOAD IN THOUSANrn:: OF POUNDS

Fig. D1.5 Combined Shear and Tension on AN Steel Bolts

As the load on a fitting is transferred fram one side or the ritting to the other, internal stresses are produced which tend to cause the fitting to fail in several different ways.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

01.5

}I '~
i "

E
Fig. 01.7

~Flg.b
Flg. a Fig. b

eft

c:::@"'!=~;;:' cmo
: ==
Fig. c
Fig. Dl.6

~
Fig. d

Pu~ u:~
t

Flg. 01.8

Failure by Bolt Shear. In Fig. 01.7 the bushing Is not continuous between the plates, but each of the three plates have separate bushings. As the pUll P is placed on the fitting it tends to shear the bolt at sections (1-1) and (2-2), (Fig. Dl.7a). Fig. Dl.7b illustrates the forces or pressures on the bolt and the failure which can take place it the stresses are sufficient. Let Pu represent the maximum or ultimate load on the fitting. This force Pu must be resisted by t~e shear strength of the bolt at the two sections (I-I) and (2-2). Hence, PU(bOlt shear) - FsuA2 - (1)

R,

'8~el~ 1 QP l' 0)
Flg. a Fig. b Flg. c
1

Fig. 01.10

Fig.

Fig. b

MSTHOD 1 OF BOLT & LUG STRENGTH ANALYSIS Failure in Tension. Fig. 01.8 indicates how a fitting plate can pull apart due to tension stresses on a section through the centerline of the bolt hole. Both the male and remaLe parts of the t1tting must transfer the load past the centerline of the hole, thus both parts must be considered in the deSign of a fitting. Equating the allowable load Pu to the ultilnate reSisting tensile stresses at points (a) and (b) Fig. 01.8, we obtain,
PU(tensl0n)

where Fsu = ultimate shearing stress for bolt material. A = cross-sectional area of bblt. Failure by Salt Bending. The main concern regarding bolt bending is that under the ~ loads any bending deflection of the bolt be not permanent as such deformation would make removal of the bolt di!f1cult. The sUbj~ct of bolt bending stresses is discussed in a later article. Failure of
L~g

= Ptu

(2R - n)t

- - - - - - - -

(2)

where rtu = ultimate tensile strength of plate material. Equation (2) assumes that the tensile stress on the cross-section Is unitorm. This Is not true as the flow of stress around the hole causes a stress concentration. To take care ot this stress concentration requires a margin of safety of 25 percent. Failure by Shear Tear Out. Fig. Dl.4 illustrates the manner in whIch failure can occur by the shearing tear out of a plate sector in front at the bolt. In Fig. a,

portion or Fitting.

The lug ~ortion or the fitting refers to that portion ot the fitting that involves the hole tor the si~gle bolt that connects the ~le and female parts of the fitting unit. The Simplified assumptions regarding tailing action and the reSulting equations which tallow have been Widely used tor quiCk ap?roximate check of the lug strength. Tr.e procedure which follows will be referred to a3 Method 1.

D1.6

FITTI:IGS AXD CONNECTIONS.

BOLTED AND RIVETED.

= 0"1.'.5&:3 t ae bol t :':10':: s ncvn ) to cress the plate ~ro~~~ tC3 jJlt t~l~ ~d~e. 5tressss.:>.re cr-ocucec whfch tend 'Co cause t.ae
':::-.e Loac
~~

as be core , P;;,
~

por-t t cn (a) Ln
~q~~:ing

?1~.

:<'br D

the

resista~ce

c:

10~d

(:;) ?u tc

",::,'J

t.ee r 'Jut S2 shown.

- - - - - - - - - -

(5)

t~e ~l:i~~te
'NP

'Ch3 ~a~eri~l ti~e3

the tvo shear- cut ar82..'3,

shsar the area or can ':1":-1 t e ,

------(3)

A ~r~in of safety cf 50 percent shoul~ be malntain8c. If the :"itting is Subjected to tnr requent rotat rc- under' lead cut with Load
Invo Ivfng no shock or v tbra t Lon , r-equf re a

whet-s

su
As

= ult1~lte

plate ~terl3.1. she-ar- out area.

shearln~ strengt~

of the

It is ve:-y C0~on practice to tak~ th~ stear out area As equr.l to the edga distance at the centerline o~ the hole times the th:;Kness t of the plate ti~es t~o since there are two shear are~s. ~hi5 is slightly conser~ctive because the actual shear area is larger. Ref. 1 pernits one to use the area alJug line 1-1 whic~ is li~itad by the 40 degree Ltne as sho.vn in Fig. D'l , sc .
?~il~re

margi:1 o~ safety of 100 percent. I~ shock or Vibration with infrequent rctati1n is ?resent, require a ~~rgi~ of safety of 150 per~ent. S~ock 1s cansidered to occur in s~ch structures as l~~ding gears, g~~ m~unts, halsting, moortn~ lnd towing conn~ct10ns.
GenAr~l

Comment en

Mar~ins

of Safety

~or

LuSs.

by

3ear~ng

of BUShing on Plata.

I~ FiS. Dl.lOa, the pUll P causes the bolt (not sho~n) to press against bushln~ ~~ll which i~ turn presses against the plate wall. If the ~ressure ts ~i~h BTIough the ~late ~~ter1al adjacent to the hole will start to crush and flow thus allowing the bolt and bushing to ~ove ~hich results in the elongated hole as illustrated in ~lg. b. Squati~g the load Pu to the ulti~ate bearing strength on t~e bearing surfaces '.V8 can wr-t te , -

In gene~al it is geod desi;n pra~tice to deSign lugs cJnservatively as the ~elght of lugs is small relative to their ~Dortancs in insuring the se.retv c t t:19 fl1~r,t vehicle. Inaccuracies in ~~nufacture are di:"ficalt to control. It is ~)~d 1esign ?~ctlce to Drcvide sufficient ~terial to pe~lt drilling for a b~Ehing if bUShing is not USBd ~n orig1rzl desi~. If castings are used as fittings, ~uch ~i~her factors of safety on the li~it loads are specified because of the low ductlbility at the ~teri~l in ~as~lngs.
01. 11 Method 2. Loading. Lug Strength Analysis Under Axial

---(4-)

where Fbr = allowab18 bearin~ stress D ~ diameter of bu~r.lng t ~ plate thickness It is good of safety of 50 ?ailure by
a~d

~ractice per:~nt.

to require a
3~Shi~~.

~2~gin

Due to a sQ~prehensive st~dY and test program by Cozzone, ~elcon and Habelt (Refs. 3 and 4), the proeec.ure as ~lven in Method 1 1s somewhat modified. ~he l~)Ortant 1if:erence is that curves derived frsm test results give the stress co~~ent~~tion factor to use for tension on the net secticn and the shear out failur~ ~s 2SS1~,~d i~ neth~d 1 has Jesn repl~ced by a combined shear-out bearing failure. Fig. D1.11 ShCNS ~~e lug-pin ~o~~ir~tior.s and t~e~ of f~ilure ~s ta~en trom R~~. 3.
Bushing

3ea~lng

o! 3Git

~ :ushlng i3 ;ressed lr.~o the )lat2 hole thus it is 2onS~~erec as a ti~ht tit. ~ r t t t t ng bolt 1s ";5:lii.':"ly cons t ae r-ec 8S removac Le :herefore a cer-te t n t oIer-ance ce tween ':h," :'J1t and busht ag 1."',:31.'::'e diamet.er-s is -ieceeearv in or-der- to inser-t anc r-emove bolt. 1:' :'itti::,q: is SUbjected ':0 reversible leaes :~~ ST~ll slo~ in ':~e :itting tends to cause sheck on t~~ :ittlr:~. A151 th2 fittin~ ~ay be suc~ that =ll~ht rotation t:i.KSS ::12.oe on tile bOlt,.'Ihl-:r, t.ends :>J cause wear- between :01:: 'i::~ ous 'u nz . :t is ther-e ror-e cus'tomar-y tc check th.' b..,.a.r~:1z

ffi ~
a.

-p

fB\. t@)
b.

.-p

-, -

~~~~~~:eo~e~~:e~u~~~n~C;;~~~1,:~~~~~~~~n~nS~:l:e
explained
i~ t~~ ~revi~us ~r':i~le ie~lj;..;;;;:~~--:>:

[81-:,nsrL~
c.
- e ~ Shear Bearing Failure - ' d.

ma.~~er

Hoop Tenston F:111ure_

mg wi tn bear-i ng c:"

- ,. -

~~

Fig. 01. 11

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES The meth0ds ~f f~1luTe ~~d the methods of
Shea~ O~t-Bear~~~

D1. 7

lug s tr-ength ana.Iysf s ar-e as follows:

~ens1on A~ross

Strength.

the Net Section.

8acause of stress ~oncentration, the stress on the net cross-section cannot be taken as ~lfo~. The ulti~te allowable tension load Pu for lug equals,
- - - - - - - - - (6)

Failure due to shear out and bearing are clossly rel~ted ~nd are Covered oy a Single calculation based on empirical curves. The ultimate or failing load in thi~ shear-bearing type of failure ~s:-

where Kt is the stress concentration factor as found from Fig. 01.12 and Table Dl.3. Ftu = ulti~ate tensile strength of the material and At = net tension area.

Dl.13.

The values of Kbru. the shear-b~aring efficiency factor. 1s given by curves in Fig.

For stear-bearing yield strength the equation 1s,

~0'I'1:!,

p,.-"'t'"
I.
1.

M' , v.,

A.'"'''
,

1.. . - , _

m",.loc<

.' . 11

_ r.....

_.

"".

"r('~*
Fig. Dl.12 Lug Design Data Tension EffiCiency Factors fOI Axially Loaded Lugs (Ref. 3, 4) Table m. 3 Curve Nomenclature for Ax1a.1 Loading for Fig. D1. 12 L, L T and ST Indicate Grain in Direction F in Sketch L LOngitudinal L T - Long Transverse ST Short Transverse (Normal) MATERIALS Curve 1 - 2014TB and 7075TB Die Forging (L) 4130 and 8630 Steel 2014.T6 and 7075T6 Plate ~ O. 5 (L, LT) 7075-T6 Bar and Extrusion (L) 2014-T6 Hand Forged Btilet :;, 144 in. a (L) Curve 2 - 2014-T6 and 7075-T6 Plate"" 0.5 in. ~ )., 0 in.
(L, LT)
.... , no
,~.

:=;:::::;:--=="'-=-=_:::_= :.. .....-t.

, ..__ c.... .'" " ".'.'"

...e ...

."
..,

- ,.. _" ,,,--_.'" "_.n _

...... _, .....
'.1

. ""'..00_ , '.1._
"' ' . " - ". "

....- " .." .. _ ..... "'....... _

.... --._.-_ . , .....,,,---'--11 00 '.. _ .. _,

... ' - ' ._ - ._ , ,....... ,. . "'lot.. _ . I

_0.. ...... _' , "'1 _

"-

_ '.. L"'.-- .... ,

c'

__-_ ..,-,. ... ..

t.. , _ , , , _ , , ,

, .... __ ,.. ,.. _,

~.':..": "'"

_,""_ ,-. ..
_ _Il

"

_ n

_,.

'.0

I.'

.,

...

Fig. Dl.13 Lug Destgn Data Shear-Bearing Efficiency Factors for Axially Loaded Lugs (Ref. 3, 4)

Fig. Dl.l4 gives curves ~O~ ~tnd1n~ Kbry' Bus-hing Yield. Take Acr as the smal:er of the beari~g areas 01' bus~.1.::,;: en pt n 0:- Jus:;'1:J.;;" on lug. (':':-:e latter ~y be smaller as a :-es~lt Jf ~X~er~al c~~fer of the bUShing, oil gTocves, ~tc.; r~e allJwacle y1eld bearin~ load Qn bUShing 1s then,

Curve 3 Curve 4 Curve 5 Curve 6

Curve 7

7075-T6 Extrusion (LT, ST) 2014-T6 Hand Forged Billet:>o 144 in .:II (Ll 2014-T6 Hand Forged Billet ~ 36 in .:II (LT) 2014-T6 and 7075T6 Die Forgings (LTl 2024-T4, 2024-T2 Extrusion (L, LT, ST) 2014-T6 and 7075T6 Piate,.. I in. (L, LT) 2024-T4 ~ (L, LT) 2024T3, 2024T4 Plate (L, LTl 2014-T6 Hand Forged Billet ~ 36 in.~ (LT) Aluminum Alloy Plate, Bar, Hand Forged Billet and Dte Forging (STl. NOTE: For Die Forgings ST Direction Exists Only at parting Plane. 7075T6 Bar (T) - AZ91C-T6 Mag. Alloy Sand Casting 356T6 Aluminum Alloy Casting

Where Fey is c0m~"ess1ve Y~el: s~ren~tj of

busn1~~ $~ter~~l.

Bolt :,r Pin She.?!" Strer:.sth. the

Ss~e ~n:.er ~s

~he

bolt shear

;f,en in ~~thOd 1.

s~ren~th

is calculated in

D1.8

FITTINGS AND CONNECTIONS.

BOLTED AND RIVETED.

Pu" Kt,ry Abr Fty FOR ALL MATERIALS

2.0 0.6. atIci lUgher tiD valuu

1.8 1.6

'ID

#.
1.5

::;:::====

"2
0.4 0.3
~,

~ O'2 .15
.12
10

1.2

I .... n_"'~Ii1.". . . I M _ ...... 1i1,. I. l40t tID 1M. __ Olio!' , _

- - : . _ - - - - .08

1.0

J. " '

... _ _ _ ..'c_.
T.
P*.IO~1

_ .. u.s.

. . -...

1.0

-------."

.8 6

"'........

V~&r<If

__..
~

c...-

IMp~~

....

... lI11lOollooctor

M.S.

~o.u

,.

..

. __._ ..:....a.:.~ __

.5

.4
-13=":- -----;-:~-.~

.2
1.0 1.5 2.0 2.5 3.0 3.5 4.0

. -"--- ---"-14

0.5

./D Fig. D1. 14 (Ref. 5)

Values of Shear-Bearing Factors of Lugs

.2

.6

.8

1.0

1.2

1.4

Bolt or Pin Banding. subject of bolt bending strength is treated in Art. 01.14.
~he

Aav
Abr
Fig. D1.15 Lug Design Data Tension Efficiency Factors for Transversely Leaded Lugs (Ref. 3, 4) (See Table D1. 4 for Curve Nomenclature) Table 01.4 (To be Used With Fig. D1. 15) Curve Nomenclature for Transverse Loading Curve Curve Curve C).1rve -Curve Curve Curve Curve Curve Curve 14130 and 8630 Steel thru 125 KSI H. T. 2 - 4130 and 8630 Steel 150 KSI H. T. 3 Ktv for All Aluminum and Steel Alloys 4 - 41'30 and 8630 Steel 180 KSI H. T. 5 - 356-T6 and AZ91C-T6 sand Casttngs 6 - 2024-T3 and 2024-T4 Plate ~ 0.5 in. 7 - 220-T40 Sand Casting 8 - 2014T6 and '7075-T6 Plate =!:. O. 5 in. 9 - 2024-T3 and 2024-T4 Ptate > 0.5 in. also 2024-T4 Bar 10 - Approximate Cantilever Strength for All Aluminum and Steel Alloys. If Ktu is Below this Curve a Separate Calculation as a Cantilever Beam is Warranted. 11- 2014-T6 and 7075-T6 Plate> 0.5 in. ~ 1. 0 in. 7075-T6 Extrusions 20l4-T6 Hand Forged Billet ~ 36 in." 2014-T6 and 7075-T6 Dte Forgings 12 2024-T6 Plate, 2024T4 & 2024-T42 Extrusions 13 - 2014-T6 and 7075-T6 Plate > 1 in. 14 - 2014- T6 Hand Forged BUlet :> 36 in.<1I

" f

D1. 12 Lug Strength Analysis Under Transverse Loading.

Cases arise where the lug ot a fitting unit 1s subjected to only a transverse load. Melcon and Hobbit in (Ref. 4) express the ~ltl~ate transverse or failing load by a single equation:- -(10) Similarly the y1eld strength of ?ty = Kty Abr Fty
e:!:c~8nts

lu~

is,
- -( 11)

The efficiency failing and Yield coKtu ~nd Kty are given by the curves in ~ig. Dl.15. The curve nomenclature for the c~rves ~n ~i6' 01.15 is given in TaJle Dl.4. ~n ~sing Fig. 01.15, a value called Aav is needed, the value of which 1s shown in the equation shown on Fig. Dl.15
D1. 13 Lug Strength Analysis Under Oblique Loads.

Curve

Curve Curve Curve

Fitting lugs are often subjected to Jbllque loads. Ref. 4 gives the following a??r~ac~ to this loadi~ case. Resolve the applied load into axial and transverse components. Tr-en use the following inte~actlon equation;__ - - _ - - - - -(12)

or margin of safety is,

\ _ _ _ _ ,.;.0)
~

,;

where,

R~

= axial component 0: applied ultimate

load divided by the smaller of the

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

D1.9

values obtained for equatiQns (6) or (7). Rtr = transverse component of applied ultimate load divided by the values of Ptu in equation (10).
01. 14 Bolt Bending strength.

shown. On the vertical scale we locate point (y) at tension load of 5800 Ibs., and draw horizontal dashed line to give point (a). A straight line through paints (a) and (a) is drawn and extended to intersect the curle for a 5/8 AN bolt at point (b). Projecting downward t~om point (b) to lower scale, we obtaIn PS(allow) = 22400 lb. and projecting horizontally we obtain Pt(a1Iow)

!n general static tests of single bolt fittings will not show a failure due to bolt bending failure. However, it is tmportant that sufficient bending strength be provided to prevent permanent bending deformation of the fitting bolt under the lUnit loads so that bolts can be readily removed in maintenance operations. Furthermore, bolt bending weakness can cause peaking up a non-uniform bearing pressure on the fitting lugs thus influenCing the lug tension and shear strength. The unknown factor In bolt bending is the true value at the bendIng moment on the bolt because the moment arm to the resultant bearing torces Is unknown. An appr'oxamat e method (Ret. 4) for determining the arm (b). to use in calculating the bending moment on bolt Is given in , Fig. 01.16, which gives b : .5t~ + .25t~ + g where g is clearance or gap between lugs. ThereSulting bendIng moment Is considered to be conservative. (See Ret. 4 for ather refinements relative to determining moment arm b.)

= 7700

lb.

Then M.S. = (?S(allow)/X) - 1= (22400/ 16700) - 1 = .34 or M.S. = (Pt(allowl/Y) - 1 = (7700/5800) - 1 34. The specified required tl.S. was .25, thus bolt strength 1s satisfactory. PROBLEM 2. Fig. Dl.17 shows a single pin fitting. The lug material is AlB! Steel, heat treated to Ftu ~ 125000 pSi. The bolt is AN steel, Ftu = 125000. The bushing is steel with Ftu : 125000. The titting Is subjected to an ultimate tension load at 15650 lb. The fitting will be strength checked for the design load. The check will be made by both Methods 1 and 2.

!.>
2

Load on Pin

1/2" Dia. AN Steel Bolt 1/16" Thickness Steel Busbtng 1-3/16

P... t. 15x 15650 ... 18000* I.

_J_
p

Lug A

R=3'2

2"

f
i.,

" L_"
Flg. Dl.17

"

FIg. 01.16

80LUTION BY METHOD 1.
D1. 15 TIlustrative Problems.

PROBLEM 1. An engine mount fitting is attached to the nacelle fitting by a 5/8 diameter AN steel bolt. The ulti~te or design loads on the bolt are tension = 16700 lbs. and shear: 5800 Ibs. Due to Vibration a 25 percent margin of satety will be required. Solution:- This Is a problem ot combined tension and Shear on a bolt. Use is made of curves in Fig. D1.4 Which are interaction curves for combined tension and shear. On Fig. Dl.4 at a value an lower scale equal to the shear load of 16700 lb., labeled point (x) on the figure, we erect a vertical dashed line as

A fitting factor of safety of 1.15 will be used which is standard practice for ~ilitary airplanes.

DeSign Fitting Load: 1.15 x 15650 = 18000 lb.

~neck

of Bolt Shear Strength.

Bolt is in double shear. ?rom Table 01.1, Single shear strength of 1/2 inch diameter AN steel bolt is 14700 lb. whence Pu : 2 x 14700 : 29400 lb. M.S. = (29400/18000) - 1 =.62

4Sb

Dl. 10
:~e~k

FITTrNGS AND

CON~ECTIONS.

BOLTED

A~D

RIVETED.

Bending of 3clt. to
:~g.

SO:"'L":'I~X

3Y :'Tl'::CD 2.

?=fe"rir~
-'.~~

Cl.lo, the

~c~ent ~~

ca Icu Ie.t tng benc rng ncnent on belt is,

1/6~

St~en;~~ ~re

:: = 0.5
fj ::

(.818;+0.25 (.3~5)".J156:: .2:;'3 incase Benc mg aomant on belt:: .52'::: = .5 x 18000 x .218

:: :: J.St l + O . 2 5 t a + g; (take ;; as

tnch

Bolt Shear Strength and Bolt Bending calculated in th~ same manner as in ~e~hJd 1 and thus the calculations w1~1 not bs re ceat ec ,

Tension Net Section

:: 1965 Lnc Ib , I"'.r/I :: 0..965 x .2.5)/.003069:: 150000 ;5':.

(S~~ ~able

01.1 for bolt moment of inertia) 01.1, Fb for AN Steel bolts is Kt is the tension efficiency factor to take of stress concentration due to the hole and is determined tr~ Fig. 01.12. Table 01.3 says to use curve number 1 for all steels. To use Fi~. Dl.12 requires the ratio WID :: (1.1875/.625) :: 1.9. Then from Fig. 01.12 we read Kt = .98, whence,
~are

2r~m ~able

1:;0. OCO.

Pher e rure M.S. ::: (180000/160000) -1 ::: .12

T~1s lug is more critical trzn lug B 3ir.ce t~ickn~ss of lug a is more than one-half J~ ';'ug A.

Pu

= .98

x 125000 x (1.1875 - .625).375 28500 lb.

M.S. = (28500/18000) - 1 = .58.

C~ec~

of

T8~si0n

Through Bolt

~ole.

F:..;.

125000

x (1.187.5 - .625) .375 :: 26400 lb.

Fig. 01.12 says that a M.S. of .15 is appropriate over tr~t of all reqUired fitting factors ot safety. thus our ~.S. of .58 provides more tr~n this additional M.S. of .15.
Shear Bearing Strength.

~.s.

(26400/18000) - 1 ; .46

~ethcd

T0 take cars or stress concentration, 1 says ~int~in a ~.S. of .25, thus lug

t~nsl1e

strength is O.K.

:~eck

of Shear Out Stren5th at Lug.

Kbru is the shear-bearing efficiency factor and is obtained tram Fig. Dl.13.
D/t e/D

= .5937/.525 ; .95

= .625/.375 = 1.55
Kbr~

Pu:: 82000 x {.S937 - .3125}2 x .375 ::

lb.

From Fig. 01.13, we read

= .80

Tiis 'Jalue is under ~he deSign fitting load or :eooo lb. It ~s ~~~isSlble to take a Slig~tlY 5~eatar shear OU~ di9ta~ce tr~n the edge ~~sta~ce as ~sed (see ?ig. DI.ge). The e.cci t ioaar c tstance to ens 40 degree :"1n8 would ~:d enough shear out area to give a ?osltive ~rg1n of ~atety. ~~1s ~eth0d o~ calculating she~~ out ~trengt~ ~= conservati7e as Nll1 be er-e-rant out :y the result in i"!etho1 2 so Lut t on ,

?bru = .80 x 125000 x .625 x .375

~.S.

= 23400

lb.

(23400/18000) - 1

= .30

The reader should note shear out strength by Method 2 is considerably larger tr~n by ~ethod 1. Fig. 01.13 says a .15 i"!.S. is appropriate, thus our 0.30 is satisfactory. BUShlnp,: Yield.
F'~.,

~~.~

bushln~

s:~en~~h
~:"

ccl t

and lug have the same ultL~ate (?t~), thus bearing will be critical In ous atng since bea rfng area is Leas .

., = E3C'OO

tor steel

?~ry

= 1.36 x

113000 x .500 x .375

= 390CO

l~.

?.
~n ~xt,~

~.s.

= (39000/13000)
Conclusion.

50 per-cent margin of safety 15

~y

G9ner~l

r equ t z-ec cr' :::8 a l Lowab Ie bear-Ing stress Fbrs~]uld

Je

i:,:~e1

1.5.
- 1 :: .35

."
~.s.

':';?OOO/l,.SO) .50 x .375 :: 24300 lb.

= (2~OO/l800C:

S~noe ~11 margins of safety are ~os~~ive, the strengt~ ot f~ttir.g ~~lt is sat1sfact8ry. It could be redesi~ned to save weigh:. Mov1ng

ANALYSIS AND DESIGN

t~~

OF

FLIGHT

VEHICLE STRUCTURES

01.11

hole s11~~tly back of the center ~f the r-ac ius woul.d ne ;c shear out s tr-engt n. ~hls c~2nge woul1 pe~lt decreasing the thickness o~ l~g sli~htly. Decreasl~~ ~he lug ~hlckness ~c~lc cecrease t~e jolt bending moment and ;Jssijly pe~it ~se of ~/15 claneter bolt. The
-'--6

stadent

sh~uld retesl~n

the iu;.

PRCBLZM 3.
A IU~ identical to lug A in Problem. 2 is suo jec t ec to a transverse loat, Pt"J.1 '''!r.le:, is the ra i Ling load. tind the load P~u.
SCL:.:rrlON:

,
~~

I--'-"':32
16

17

;
J,

w~l.1.. 0-

D
.L

SOLU':'ION:

:
I

from eq~tion (10) the faillr~ load lS,

Fig. D1. 18

DeSign fitting load;: .;. 33000 tension and -25000 Ibs. compression. These loads include the fitting factor of safety of 1.15.
T~e fail:!.r~ coef:lcient K!u is determined by use of c~ves in Fig. 01.15. The lower scale parameter for Fig. 01.15 is Aav/Abr, where fittl~~

Let Pu equal ~ltimate strength ot the unit in the various failing ~odes.
Strengt~

Check Shear
6

of Hinge Pin.

s;

+ 1:...+
A2

1
A"

.;..l...
A ...

For the meaning of these (A) areas, refer to sketch of lug ~n Fig. 01.15. SlmDle calculations give the following values :01'" these areas.

The hinge pin diameter is 5/6 and ~oerial is AlSl Steel, heat treated to Ftu "" 150000. Fsu = 95000. Pin is in double s near-,
Pu
z

F su As

= 95000

x .3068 x 2 ;: 58300 lb.

A.

M.S . (58300/33000) - 1 77.

~~eck

A.

= .14 , =

A,

= .',

= 105
=
.126

BendIng Strength of Hinge Pin .

a~

whence , A av

6 1 1 1 3 + + + .,14 lOS .14 .10S

The bendIn~ moment rig. 01.16 ie

(J) on

~in :ro~

Abr = .375 x .625 = .234, "'-ar/A-br = .126/.234

=0

b = .5t l + .25t a +~. taken as 1/64 inct. b

T~~ clea~anca

g will Je
.329 in.

.54

= .5 x .3125 .;. .25 x .625

~

1/64

steel
:r~m

Tacle D1.? incicates curve 1 is used tor ~tar1al N~th ftu u~ ~o 125000 ~Si. ~hen Fig. Dl.15, we rsad K t u = .74. . 74 x .234 x 125000 ;: 21500 lb.

Pin ~ending moment ;: 5410 in.lb.

.5P~

;: .5 x 33000 x .323

::hen, ?tu
?RCBL::!1 ~.

fa =
be

~/I

= 5410 x .3125/.00749 = 226000 PSi .

of
the
r~pture

T~e ~cdulus

in Janding Fb Will
of
~hapLer

calc~la~ed ~y

~ethod

C3.

Fig. D1.15 shows a steel forkac fitting to a dou~le crsnr.2L section ~de ~:rom "'C75-T6 a.Iumimm alloy. The hinge ptn 1s 5/3 ~l~eter AlSI Steel, heat treated to ~tu ;: 150000. 5:8el t~ttir.g is also 150000 steel. A strength check of the fitting Rill be carried out.
~olted

~rom

we

f~r.c

Table S3.2 fa = 146000.

~cc

st=el w1:h

.' tu ;: 1500:0,

For a solid r0und bar K ;: 1.7 {sae :'lg. 83.7 of Chapter C3) .

D1. 12

FITTINGS AND CONNECTIONS.

BOLTED AND RIVETED.

'0 150000 + 146000 (1.7 - 1) 258000 psi

~.S.
C~eck

= (258000/226000)

- 1

= .14

A stress concentration factor Kt fram Fig. Dl.12 will be used to be conservative. WID = (1.125/.3125) 3.6. Fig. 01.12 gives Kt = .91.

Shear-Bearing Tear Out of Lug.

Pu = .91

x 150000 x .305 = 41700 lb.

- 1 26.

Pbru = Kbru Ftu Abr D/t

M.S.

= (41700/33000)

= .75/.625
~

= 1.2, elD

= .6875/.75 = .92
~

Check Section of Steel Fitting at Section 3-3. Load tram Fig. 01.19 = 16500 lb. Net Section (1.125 - .3125).25 .204 sq. in. Use same Kt as before.

From Fig. Dl.13, Kbru = .72 Pbru .72 x 150000 x .75 x .625
51500 lb.

M.S. (51500/33000)-1 at least 0.15),

= .56

(margin desirable

Pu = .91 x 150000 x .204 27900 lb.

M.S.

= (27900/16500)

- 1

= .69

Check Tensile Strength of Lug Section Through Pin Hole.

Check Shear Strength or 4 - 5/16 Dia. Bolts. Bolts in double shear.

Pu

Pu = Kt Ftu At
Kt obtained from Fig. 01.12 using curve 1. WID = 1.125/.75 = 1.5

x 2 x 5750 = 46000 lb.

- 1

M.S.

= (46000/33000)

= .39.

From Fig. 01.12, Kt

= .99 = .06.

Check Bearing of Bolts on Steel Fitting. Section 3-3 or 4-4 1s critical as bearing area is less. Allowable bearing stress for 150000 steel is 21S000. Pu

Pu = .99 x 150000 x (1.125 - .75).625 = 34800

M.S (34800/33000) - 1 .15.
C:~eck ~ensile

It would be desirable to have a M.S. of Strength of Section 2-2 of Steel

= Fbr Aor

= 218000

x .3125 x .25 = 17000 lb.

Fitting. From Fig. 01.18, four 5/16 diameter AN Steel bolts are used to attach fitting to channel members. Since bolts are the same size, it will be assumed that each or the 4 bolts transfers 1/4 of the total fitting load. Fig. 01.19 shows the load in the steel fitting and the channel members. The total load passes by section 2-2.

M.S.

= (17000/8250)

- 1

= 1.06.

Bearing of BOlts on 7075-T6 Channels. For for 7075 aluminum alloy = 106000.
Pu

= 106000 x

.3125 x .25

= 6300.

M.S. = (8300/8250) - 1 = .01. Check Tensile Strength of Channels at Section

4-4.

9 o

0 z,
a

Load on Section Net area

= 33000.
= 78000

= .70 - .25 x .3125 = .622.

Ftu for 7075 material

~

Pu

= 78000 x .622 = 48500.

(48500/33000)-1 = .47 (which will take care of any stress concentration.
B~.

47sa ~ /h500 .' la;~ ;;'!>T~I&tJTlOIII

=.
t

3"0,,"

:1.S.

o
Fig, 01. 19

Check Shear Out of Channels Behind Bolt Shear out distance .625 .156 469.

Net tension area

= (1.125 - .3125).375 = .305.

Shear out area = 2 x .469 x .25 = .234.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Dl.13

Pu

= Fs u

AS: 43000 x .234: 10100 lb.

Therefore we can write,

P

M.S.

= (10100/8250)

- 1 = .23 (figured conservatively by Xethod 1)

B~.

Pn = P (2!!..) Z Ps
Example Problem of Bolt Load Distribution.

(A)

Shear Out of Steel Fitting Behind Bolt

F'su :

95000

Shear out area As

Pu

= (4375

- .1565)2 x .25 = .141

= .141

x 95000

= 13400.
- 1

Fig. Cl.20 shows a multiple bolt fitting un1t subjected to a concentric load of 100000 Los, Determine the load transferred by each bolt.

M.S.

= (13400/8250)
= 1.85
Fey Abr' steel.

= .62.
: 113000 for 125000

Bushing Yield Strength.

?b~y

F cy

Pbry

= 1.85 x 113000 x

.625 x .625 Load = 33000, not critical.

= 81700

lb.
Boll
Sym.
A

Fig. Dl.20 Boll

Dla.

D1. 16 Bolt Loads for Multiple Bolt Fitting. Bolts Sizes Different. Concentric Loading.

Double Shear Strength P, 23010 x 2 ~ 18637 x 2. 11272 x 2. 5751 x 2 ZP 46020 37274 22544 11502 117340

58
9 18
7 16 5 16

In designing or strength Checking a ~ultiple bolt fitting, the question arises as to what proportion of the total fitting load does each bolt transfer. This distribution could be affected by many things such as bolt tit or bolt tightness in the ho Ie ; bearing deformation or elongation of the bolt hole; shear deformation of the bolt or pin; tension or compressive axial deformation of the fitting members and the member being connected, and a number of other ~inor influences. Since aircraft materials have a considerable degree of ductility, it the fitting is properly designed, the loads on the bolts when the loads on the fitting approach their maximum value, will tend to be ln proportion to the shear strength t tne bo Lt ; That is , if the combined shear strength of the bolts is the crltical strength, the yielding of the fitting material in bearing, shear and tension will tend to equalize the load on the bolts in proportion to their shear strengths. For stresses below the elastic limit of the fitting plates the bolt loac distribution no doubt is ~ore closely proportional to the bearing area at eac~ bolt. Since the primary interest is tailing strength, the bolt load distribution in proportion to the bolt shear strengtr~ is usually assumed. Let:P = design load on fitting. Ps : allowable shear strength of any bolt. Pn = load on bolt n. Fs n ~ allowable shear strength of bolt n.

= =

Distribution ot loads to each bolt:Pa

=P

(P sa ) Z Ps

= 100000 x

46020 117340

= 39200

lb

Pb = 100,000 x 37274/117340 = 31800 lb.

Pc : 100,000 x 22544/117340 = 19200 lb. Pd = 100,000 x 11502/117340 : 9800 lb. Total load on the four bolts adds up to 100,000 lb.
D1.17" Multiple Riveted or Bolted to Eccentric Loads.

Joints Subjected

Fig. 01.21 shows a plate fitting attached to another member by means of four bolts or rivets. (The circles represent the bolts). The fitting plate 1S sUb~ected to the loads ~d and ~l acting as shown. Let it be required to find the resultant loads on the bolt group due to the given loads. The centroid of resistance Por the bolt group will then correspond to the centroid of the bolt areas. Fig. 01.22 sh~NS the fitting unit with the :orce ~rl and ~{ replaced by an equivalent forca system at the bolt group centroid point (O). This equivalent force system will be:-

Pv' -200

lb.,

PH

= 1000 lb.

D1. 14

FITTINGS AND CONNECTIONS.

BOLTED AND

RIVETED.
~roup

::--! ----- I +~

~~;-+.'"

a.S

b:;
,q"",,,

rc

For the bolt rd ~ 1.25".

of

?i~.

Dl.22 r a = rb

-tr-

-+;;-,
i

Hence. I ; Zr" = 4 x 1.25 a = 6.25 in. Therefore -Nill equa L

F
mo~ent

load Fm

o~ eac~ bol~

c--6-IFig. D1. 21

- ~I = -600~X ;.25 = 120 m--rs.as

lb.

Fig. C4.3l shows the resulting H, V and moment loads applied to eaoh bolt. The resultant load can be found graphically by drawlP~ the force polygons as sho'Nn in Fig. Cl.23. The resultant bolt loads can likewise be deter.nlned analytically. For example consider bolt (e) ZH = 250 anc M'J = -200 x 1.75 - 1000 x .25
Slnc~

120 x

1.~5

= 346

lb.
lb.

= -600

in.lb.

xv

= -50 - 120

o = -122

?V and ~~ act truough t~e Jolt centr-c ic t.aey :/ill be reacted equally by eac h )01:, hence V load on each bolt due to ~{ = -200/4 ~ -50 lb.; H load on each jolt due to ~{ = 1000/4 ~ 250 lb.
~omerrt

Case ~here Bolts or Rivets are of Different Diameters.

~ben the joint bolts or rivets are not all the same size, the moment load on each bolt is proportional to the bolt area ttmes its distance to the bolt group centrold. Thus the bolt areas must enter equation (15).

The load ~roduced on eac~ bolt due to the load ~~ = -6CO in.lb. will vary directly ~s the distance ot the bolt from the center of r es i s tance ',IIhic.': co mc tdes with the bolt group centrc td ,

Let r a equal the distance from the bolt group centroid to bolt (a). Then the resisting ~c~ent developed jy Jolt (a) equals Fara, where :a equalS the load on bolt (a). Since the bolt leads are proportional to their distance from 0, the resisting moment l"!:b deve Loped jy bolt (b) 'Nill equal

hence,
jJr.A

a ra
cee
A r "+n A r a

n A r a+:e .; r

aaa

"bbb

~+n

ddd

- - - -(16) n

= number of bolts of each size.

:1c

" and for bolt = Fa -'ra

The total
~oment resis~ance o~

-.

and

si~llarly

rar bolt (0)

Since theory of loads on a ~ultiple bolt group ~s only apprOXimate, reasonable ~rgins of safety should be ~aintained.
RIVETE~

CONNEC7ICKS

the

bol~

;rJUp

t~ere~8re
~+'"'r

equa:5.
a+::-.,..

D1. 18 Types of Rivets .

... .,..

t a' a

tab

.,,-.;

"+"'r

ad

- - - -(14)

From an aercspace str~~tara1 s~anGpOlnt rivets may ~e )laced into t~o ~eneral classifications, namely:(1) The ?rotruding Head type of rivet.

----(15)

(2) The Flush

t~e

rivet.
protr~ding

.vner-e I = Z r

a
t~e

Therefore for

loads

o~ t~~

other bolts,

rig. D1.24 illustrates the tj?e of rivet. Fig. 01.25 l~~~str~tes a number of mod1~lcatlons of the protruding head type of rivet ~hat have been used in the past.

head

Fig. D1. 24

:- ,_~. :~:\~.~ "'<:;'~~-~~ ~'_~,~~~t!:~':':',-:~:~j:'~.~'4 ~.-. ;.~:~ ~: .: <~~" ~'\:':"~1\' ~;~~ ~ ~.~ ~~ ~ ,:'~~_ .~~. ~.:;:~.~~~~
---' -

,r. ::

~ -~~:,.~. '~~~~ >{. :\~'::~,;:~~,'('~~j:t

-

-....,~.............,,-:~

...~-'~,

A::-iALYSIS

A.ND DES!GN OF

FLIGHT VEHICLE STRUCTURES

D1 15

-6'

I '~'; 8
v ,I
I

- 8 - -B---J, L a - v - 8 - u ~ I~ ~u/ .....,u

CD 1 I cD! LJ
~A~
, 1

~A-

U
!

01 ; @I
LJ
-A-

''" I T ' "IT

I I
I

-rr-

?ig. Dl.27 illustrates the aporOxlmate sheet thickness 11~ltations tor the :hree ~ethods of flush ri~etlng. AP9rox. S~set Li~itations ?or Machine Countersunk R1'lets (;'1:-426)

~Ai
"

U LJ
I
8

i
L' "'''.

,-8~

~A-

v.

R-+
1 :

Fig. D1. 25

l~G~

'''':.1 ..

1/8 Dia. Rivet.

!NO.
, 1

Rivet Types
Round Head

@I I

U -';A-

I 2 , 3

,
5
6

12A . 75A Mushroom Head 12A .625A Brazier 12. SA . 50A Modified Brazier 2A .25 to. 33A Flat Head 2 A . 4A Ctsk. Head l.alA . SA

L"'"
f

.....':1

5/32 Ole.. Rivel:

For ~y years the round head rivet was tar all interi~r work and before the era of high speeds it was used as a surface ri7et as well. ~~en wind tunnel experi~ents showed t~at such rivets ;ave appreciable drag, designers turned to rivets with less head ~rotrus10n, thus the develop~ent of the 3razier ana modified Brazier type of rivet heac. Then as the age of relatively high airplane speeds a~rived a flush surface was needed, ?artlcularl on certain sensitive ~or~ions ot the airplane surface, thus various ~od:fications of the countersunk head involving press and machine counterSinking of the sheets were developed.
~sed

L'''''

","''''''1 =-l

..

3/16 Dla. Rivet.

Approx. Limitations For Press Counters~~ or Double Dimpled R~vets (~~-425)

1/8 uia. Rivet.

?lg. 01.26 illustrates the flush type of rivet. As illustrated in Fig. Dl.26, this :lush :ype rivet can be used in several different ways, thus the method ShC'NO in Fig. Dl.26a is referred to as the machine countersunk type; that in Fig. Dl.26b as the preSS countersunk double dimpled type; and that in Fig. Dl.26c as the combined press ~~d ~chine ccuntersur_~ type or the dimpled ~chine co~~tersunk type.

... ... ",.

AO" I'H"
~.

5/32 Dja , Rtvet;

....... '0'
.N'IHIII

:L

~t

3/16 Dia. :uvet

Machine
Fig. a

Counters~nk T}~e

ApDrox. Limitations ?or ?~ess-~achine Countersunk S1vets (AN :26) ?rsss

Ccunters~~

- Double

D iaotec Type.
Fig. b

1/8 Dia. R17et

Scmblned Press anc ~chine :oun:e~s~~; or Ji~pled

~nchine :0~~t2rsu~~ ~ype.

r .......
T..J.

~J

5/32 nia , Stve t

Fig. c

Fig. D1. 26

3/16 Dla. Rivet

Fig. D1. 27

m. 16
Dl.19 Rivet Material.

FITTINGS AND CONNECTIONS.

BOLTED AND RIVETED.

Pa = FSll A n where, Pa = FS ll
A
~

----(17)
ulti~te

Since aluminum alloy is by far the most widely used ~terial in the aircraft i~dllstry, it follows that aluminum alloy is the material ~ost Widely used Cor rivets. Table 01.5 (column 1) lists the 5 aluminum alloys used tor rivets and the ultimate shearing stress FSll tor each material. Rivets made from 2017-T3 (FSll = 34000) and 2024-T31 (Fau = 41000) are rivets that must be driven soon after heat treatment or betore age hardening takes place. The aging or hardening is slowed ~y keeping rivets in refrigeration atter heat treatment. The other rivet material is less hard or less brittle in the aged state and thus can be stored in air ar.d driven anytime. These so-called sotter rivets have less shear strength, but since a great deal of aircraft construction involves thin sheet, bearing is often critical and thus rivet shear is not critical. Most surface or skin riveting involves the softer rivet, usually 21l7-T3 (F Sll = 30,000).
Dl. 20 Strength of Rivets. Protruding Head Type.

shear strength of rivets

(Ds.)

ultimate allowable shear stress for rivet (pSi) = area of rivet cross-section = 'It 0)'4, where D equals the nominal rivet hole diameter. = number of shear areas per rivet.

Reference (17) shows that the sheat strength of protruding head aluminum alloy rivets is affected by increasing Dlt (Diameter of rivet over sheet thic:~ess) ratios. The conclusionS in Reference (17) are as follows:Rivets in Single Shear:For values of D/tup to 3:Single shear strength = basic allowable single shear strength. For values of Dlt greater than 3:Single shea~ strength = b~sic allowable single shear strength times
[1-0.04 (D/t-3)]

Rivets are widely used in airplane st~Jctures to fasten or tie together two or more structural units. Standard methOds of stress analyses of riveted jOints consider ~NO ~rimary types of failure, namely, the shear ot the shank of the rivet and the bearing or compreSSive failure of the metal at the point where tne rivet bears against the connecting sheet or p.Iat.e. Fig. Dl.28 illustrates the main forces cn a ~lvet in transferring a load from one plate to another. The load is transferred to the rivet from the plate by bearing of the plate on the rivet. The load is then transferred along the rivet and resisted by bearing action on the other plate. Slnce the plate bearing forces on the rivet are not in the same line, the forces tend to shear and bend the rivet. 3ending of the rivet is usually negl~cted if the~e are no intermediate filler plates. In Fig. (a) of D1.28, the rivet is in Single shear, whereas in Fig. (b) the rivet is in double shear.
( b)

For Rivets in Double Shear:For values of nit up to 1.5:Double shear strength = basic allowable double shear strength. For values of nit greater than 1.5:Doub13 shear strength = basic a11owabl~ double shear strength t~es
[1-0.13 (D/t-1.5J]

Table 01.5 (from Ret. 2) gives the shear strength of protrUding and flush head aluminum alloy rivets and the corrections to take care of the D/t influence on the ~ivet shear strength. Table Dl.8 gives the all~wable bearing strengthS between the protruding head rivet and the various alw~inum allOY sheet and plate materi~l. The bearing values are given ~or t~o elD ratioS, namely 1.5 and Z.O, where e is the edge distance measured rrom ~he center of the hola to the edge of the Dlate. Any reduction in edge distance may cause bUlging of the edge of the sheet due t~ driVing dnergy. Edge distance shou11 not be less than aiD = 1.5.
Dl.21 Strength of Rivets. Flush Type.

Fig. D1. 28
~lven

The ultimate shear strength or a rivet 1s by the following equatlon:-

Since flush rivets have no protruding head on the flUsh end of the .ivet and also since flush riveting involves ~chine counterSinking or press counterSinking or both, the strength of the tlush type rivet is dlrrerent till4~ the common protruding head type.

p'

ANALYSIS AND OESIGN OF

FL GHT VEHICLE STRUCTURES

01.17

Fig. 01.29 illustrates a machine counterrivet. Due to the ?ull P on the two sheets whi~h are held together by the rivet and ind~ced force Pi is produced on the sloping side of the head of the rivet. Th1s incticed force tends to shear and bend the portion 1-1 or the rl~8t head. The sharp edge of the countersunk sheet at pOint (a) tends to cut into the rivet. These combined i~fluences tend to cause excessive deflections and finally failure as roughly illustrated in Fig. 01.30.
s~~

Inserted Installed FIg. 01. 32 Jo Bolt

Fig. 01.33 Cherry Type

Fig. 01.30

In the press countersunk or d~pled type of flush rivet connection, see Fig. 01.26b and C, because of the interlocking of the sheets due to the dimple, the joint could transmit a load without a rivet it the sheets were held together. Since there is no clearly defined bearing or shear surface in this type of joint, the manner in which the loads are transferred is quite complex. As a result resort must be made to tests to establish design allowables. ~ables 01.?, 01.11 and 01.12 glV~ the ultimate and yleli strength of flush type rivets
(Rat. 2).
01. 22 Blind Rivets.

Fig. 01. 34 Deutch Type

The name WBlinct W rivet is given to that type ot rivet whl~h can be completely installed trom one side or the Joint, and is therefore almost exclusively used where it is impossibla or impractical to drive the normal rivet, which requires access to both sides of the JOint. There are two general types of blind rivets, namely where the i~slde or blind head is formed ~echanically or where it is formed by an eX~loSive force. Fig. Dl.~l illustrates the Dupont explosive type and Figs. 32 to 34a inclusive t Llus'tr-ate the mec nant ce.Ll.y formed head type.
;

FIg. 01.34a Huck Lock Bolt

01. 23 Riveted Sheet Splice Information.

In splicing or connecting two sheets together by means of rivets or bolts, the joint or co~~ection may fail in the various ways as explained in detail for single and mu:tiple bolt fitting units. Thus one must check the Shear strength of the rivets; bearing ot rivets on the Sheets; tear out of the sheet edges and tension on sections through the rivet holes. Types of Sheet Splices or Connections.

~!i"
Inserted

m:~
Installed

Fig. 01.35 illustrates the various types of sheet splices. In the offset lap splice between two sheets or different gauges, the offset should be in the heaVier material. For a Single shear butt splice, the butt splice olate should be equal to the thinnest of the two sheets being spliced and likewise in the

Fig. 01. 31 Dupont Explosive

Dl 18
:'::..::;~13 2:'~2.:~::J

FITTI~GS

AND

CONNECTIONS

BOLTED AND RIVETED

51~e

A::' -:

'::'3

~.,..

cut t scc ic a ::,B 5;J11:9 -Lates the S2...":l.':: tji,::KY:.RSS anc shou.Ic
thir~~st :l~:~ be1~g 5~lice1.

rivet should not

~e

l~~s t~~r.

that glv9n

be

in Tao135

A ~n~ B.
Table A

~~~~1

-~

t~~

I
I I

PROTRUDING HEAD RIVET NORMAL MINIMUM SPACING

Rivet Diameter

1/8 1/2
Table B

5/32 9/16

3/16
11/18

1/4

I Normal Minimum Spacing [

7/8

Offset Lap Splice

NORMAL MINIMUM SPACING PRESS AND MACHINE COUNTERSillH( FLUSH RIVETS

Rivet Diameter Fig. D1. 35

Normal Minimum

1/8

5/32 27/32

3/16
1-1/32

1/4
1-114

Spacing

11/16

-...------------ ----,
'fo- ---E>
~

"~;;'--..,o)

.a.

e- iD---

gener~l the ~l~i~~~ rivet row spaCing be such as to ~~~ the distance bet~een any two r i vets in tne two r-o.vs not less tnac the ~inimun ~lv~t spaci~s fJr :he rivet size be1r~ used.

In

st~u:j

3l.::<:t-8i::.;lB Splice

s t rap

Splice Sheet Tension

Sffi~l?ncy.

Proper Ri'/9t Size to Use.

N'J exact rules can be gfven z-e Iat tve to the opt rmra z-t ve t far a gtven splice, because a nunber ~: ;J~cti:al considerations usually dn~~:- into the jes1ln J~ a sheet splice. Small ~lvets, r~ely 1/16 and 3/32 di~eters are r~rd to jancle and :3.:-8 seldQm used as structurql ri~ets. The :O~tiQ~ 5i3e of rivets are liB, ;5/32, 3/16 and 1/:<. tnc n diameters. 'l'he Lar-ger s taes snouj.d not be usee in sheet 5,;11:.:es unless th~~e is a Jp.ckl~~ up structure as the sheet ~y buckl: ur:.der the dr1vi3~ of the la:-ge rivets. ~r8m a str~ctu~l 1eSi;n s:~~d ~o~~t, the 8p:i~um j0int ~s one i~ Nhi:h ~he sllowable rivet shear strength ~nd je~ring strength o~ the ~lven sheets are p~acti~p.lly ene same ~or tlle l~r~est prac~lcal SiZ8 ri~et.
~i7et

;aen a sheet is spliced by ~eans of ~ivets or ;olts it means the sheet is Neakened since the rivet holes cut ~Nay a part of the sheet ma:erlal. The ratio or the tension strength of the spliced sheet to the unspliced sheet is called the sheet tension effiCiency of the jOint. If the ~inimum rivet spacing 1s used an.I only one row of rivets the sheet err ic rencv will be 8.!"oun-j 70 to 75 ;:ercent. The deSigner should strive for a hl~her ef:lclency.

/
,
I

0;,
'""';
r-rt: ~5
R'VETS

/
/ .

Fig. D1. 36

Spacing.

Shest Sags Distance.

7he allowable rivet-sheet Jsaring loajs as giJsn in (Re~. 1) are Jasen on an ed~e d1st~nce of tvo df ame t er-s . 7herefore in genera), no ecce distance in a jOint should be less than 2 :-ivet ctamet.er-s !or pr-ot.ruc tng head r-t ve t s , and 2-1/2 d1a~eters for press and ~c~1ne countersur~ :lush rivets. The
li~8

\' M
(

/-:i-J.- ii: RIVers W</ i' COlliR ,...d::*h~ L~~~, ;' !o11_~!
I

//J;I ,: .06,3 COLLA I<! ' ,

""0

OJ

'\~:::41 '\
\,

-,

~r I2tJDDE~ l DlK:..J ;}-.0419 '

TOOQUE'

~j
i

S
"

I
I

24ST TUOf

nlr.im~T.
:-ivet~

:!

)l:ch (distance betNee!1 cente!' in the same row) for a -iven

R'VDOfR iUl'?QUf\ -""""'" TU8E HORci\

~
G

'

\!
t

I
400
,

-~~

CAeLr;: PULL

lj

A~ALYSIS

AND DESIGN

OF

FLIGHT

VEHICLE

STRUCTURES

01. 19

':,.1.1)e roz- a

~"ll

a i rc lene ,

"he cor-n is

the tube :,:," t',VC cc Ll.ar-s rfve t ed to t~3 j~rn and l~ke~lse :0 the tube. ~he design load ~~~le ~ull is 400 lb. w~ich in:ludes a f~':ting fa:tor of safety a! 1.15. ~he rivet ~~terial 1s 211~-T3 al'~inum alloy. The horn and tube is 2024-T3 aluminum alloy. The ~~rgln ~f safety for the riveted connection ~ill be
':8

rae t enec

x .c

= ~34

Lcec on each r-t 'jet due "':0 tor-cue 200'-:)/10

lb.
t.";;

T~~ hO:":Z0~t~1 ~orc8

Jf

~~o

1:.

C2.!l be taken

d i rec t sS<'l,ri'1:S between co Ltar-s

and. tubB. The ~ivets ~re in s1:~la sheqr. The Sl~~l~ shea strength Jf a 5/32 riv~t ~rom Tab12 DI.S ... 596 x .996 ;:: 594 LJ. (T':1e va.Lue of .996 is the correctiJn from ~1jdle tabl~ of Tao18 D1.0.
M.S.:::o (594/3;34)-::'

ca l cuja t ed ,

50l,),t10n:The horlz~ntal cabl,; pull of 400 lb. can be replaced by an equl1~l3nt force system at the c ent er-Lt-ie of the tube , cons Lst tng of a torsional ~omer.t of 400 x 5 = 2000 in.lb., and a hor-izonte L 400 10. force.

.~9

,':;onsUer
shear .

att:'ich!n~nt

of co11a:--5 to horn:-

6 - 1 ':: :'1.--:'::.8t;:, r t vst.s E..Te ve ec tr, couo l e

3earing strsngth ~n .0~~ tub~ wall from Table 01.9 for .050 thtckne ss '0....'1.1 Fbr = 100,000 is 795 lb. corr~ct1~g to .04S = (.049/.050) 79S = 780. For 2024-T3 tubs ~a~er~2l and e/D ~ 2.0 we abt~in a ~ater1al factor of 1.24. Ther8roce bea r tng s t r sngth of one r t ve t IJn tub,'! wall is 1.84 x 7eO ~ 967 lb.
M.S. ; (967/334) ::. 1.9

Load ;Jer rivet due to torque = 33.+ L"). (arve t '3..:'.:1 is linch).

2000/5 x 1
PHQ8La1 2.
Fig. 01.37 ShO'NS a al.at e fitting at tached to a double cl:annel sect1Jn by 5 - 1/4 d1a~etar r17sts. The 1esign fitting loads ~re sho~m in the figure. 7ne r-i ve ted connection 'Nill be checked for s t reng'th under- tne given ces i gn f1 t't Ing loads.

= 100/E
~3~

Load per- r i ve t cue ::,J hcr tz onteI ro rce = 57 lb.

~ost

Resultant load on 57 = 401 lb.

critical rivet 1s

?rom Table Dl.3 the sing1~ shear strength 1/8 jiametar r~vet of2117-T3 material is 388 lb. Since the r-t vet.s ar-e in doub le shearthe shear st:-engt~ o~ one rivet Noulrl be 2 x ~a8 = 770 10. Referring to the table at the bottom c: Table 01.5, we find. a rivet factor 8: .9~) to apply for a I/o r~vet on .063 sheet thickness. tner-arore r-ivet strength is .93.:; x
Jt
~

Solutlon:The ;siven force system will be rsp1aced by forcB system acting at the center of gravity of the rivet group. ~hls force system. 'N111 consist of:
~n equlv~lent

756

718 Lb ,
~n

Hc.~.

= 9000 lb . V c. g = 3000 lb.

8000 x 0 ... 3000 x 3 :;; 9000 1::1. lb.

~.s.

rivet shear = (718/4Ol)

= .79

Me .~. ;::

::;o.c:, r-ivet bears on t,IJC co Har-s each Gf 'Nhich

~s .063 tn :hl~kn~~s. ~he be~r1ng strength of a 1/3 z-tve t on .063 2024-!3 a Iumt num alloy ~p.ter1al is Jbt'ii~ed from Tables Dl.9 and J:.:.

~71.z0zq.n .:u,o
Fig. D1. 37
.071

From Table Dl.9. bearing str~ngth of 1/8 rivet on .063 sheet basad on 'in ~llowajle bearir~ ~tress ~br = 100,000 13 B10 l~. Then referri~g to '!'aj13 Dl.S :Jr 2024-T3 ~~~:rial ~nj ~/D of 1.5, we find ~ correct1~n "ac t ar t .98. Th:.4..S oear tng s tr-engt.h is .S8 x 2~C ~ ~94 IJ. S1nce ~~c~ r~vet :~P.::rs on ~wo :cll~~s, ':he bearin~ str~~~:h for one ~ivet = 2 ~ ~9~ : lS88 lb.
ji~~eter

~.s.

; l3S3/40l

2.95

01.20

FITTINGS AND CONNECTIONS.

BOLTED AND RIVETED.

Table D1. 5 Shear Strengths of Protruding and Flush~Head Aluminum-Alloy Rivets

Diameter of rivet, in. Shear strength, lb: 5056, F... = 28 ksi . ... 2117-T3, F...=30 ksi , 2017-T31", F...=34 bi. .. .. ~01i-T3, F =38 kei , 2024-T31", F ... =41 bi....

li',
99 106 120
13S 145

l;i
203 217 24,
273

Ji
363

;li
556 596

!1i
802

U
1,450 1,550 1,760

!1i
2,290
2,460 2,790 3,110 3,360

Ji
3,280 3,510
3,970 4,450

388
442

862

675

on

296

'94
531

753
815

1,090
1,180

1,970
2,120

4,BOO

Single-shear rivet streDgth fACtors Sheet thickness, m.: 0.016 0.018. 0.020 0.025. 0.032. 0.036.

0.964 .984 .996

1.000

0.972

1.000

0.040.
0.045 0.050.

0.964 .980 .996

1.000

0.964 .980 .996

0.972

0.063.
0.071

1.000

1.000

0.080.
0.090. 0.100. 0.12.5 0.160. 0.190 0.230

0.964 .980 .996

0.964
.974

1.000

.984 .996

0.972

1.000

1.000

Double-shear rivet strength factors Sheet thickness, in.: 0.016. 0.018. 0.020. 0.023

0.688 .753 .792

.870

0.714
.818 .857

0.032
0.036 0.040 0.045 0.050. 0.063 0.011 0.080.

.935
.9i4 .987

0.688
.740 .792 .831 .870 .935 .974 0.688 .i40 .792

.896
.922

1.000

.961

1.000

.883
.919 .948 .974

1.000

0.090
0.100 0.125

0.714 .818 .857 .896 .922

0.688
.740 .792 .831 .870

0.688
.753 .792

1.000

.961

0.n4

1.000

.935
.987

.883
.935 .974

JUB
.883

0.160 0.190 O.:2:7JO

1.000

.935

1.000

1.000

Shear .u'...... in tabl. 8.1.1.1.10 eOlTlSpondlnl' to 90 NOT.: V.lulS ot Ibeu Itrl!netb .bould b. multiplied by p"reent probability d.ta an t1Hd _b.revu uaiiabJI. the tacton riven butin _b ....ever ttt. Dlt ratio i. I.. r;-. Inou.b to require IUeh .. eorrec:tion. Shwt tbiekn..& i. that of tb. tbinn.t Ibe<l1: in .inl'lI.b.-r joint& aDd .l:b" m.iddl" Ib~ in dou.b~tr...r jointll. Sb.ar Ir&lu.. an b..ed on .r.... eorrellpond.inl' to tb. IlOmiu&I 110101 diam.un IPftill.ecI. illl table '.1.1.1.1 (4l, not . The -T31 designation refers to n,et5 that !lava heen heat-treated aD<I then m::linmin~<l in the beet-rrreted C'Ondition until dn.-inf.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

Dl. 21

Table Dl. 6 Standard Rivet-Hole Drill Sizes and Nominal Hole Diameters

Rivet

~ize,

in..

\(,

)i

W,
11
0.191

V.
F
0.257

W,
P 0.323

11
W

Drill Xo.. Xominai hole diameter, (in.)

51 0.067

30
0.1285

21

0.096

0.159

0.386

Table D1. 7 tntimate and Yield Strengths of Solid 1000 Machine-Countersunk Rivets

Strength, Ib
Rivet material. Clad "beet material.
2111-1'3

I
I
11I

2017-1'3

I I
h

2024-1'31

2024-1'3, 2024-T4, 2024-T6, 2024-1'81, 2024-T86, end 7075-T6

Rivet diameter, in ....

.........

l~

)i

!is

11I

w,

~-:i

t:Itimate strength Sheet thickness, in.: & 0.020. 0.025. 0.032 .. 0.040 .. 0.050.

.... .

132 156 178 193

206

163
22\

272
309

'250 '348
~418

... "25

.. . .. . C416

....... .

..
. .......

. .......

'324
'"55.1)

......
'726 '859 '917 '969
1,015

0.063..
0.011 0.080.

216
..

340 363
313

'479 523 542

560

'll28 705 739 769

795

'580

... .....
'"1,200 cl,338
cI,452 cl,552 el,640

'657 600
720

756 886 942

992

'"975

'"1,290
'"1,424

0.000.
0.100. 0.125 ... 0.160. 0.190.

..

575

746

1,00.;
1,073 1,131

..
....

818
853
..... 755

I,OM 1,000

'"1,543 el,641 el,738

1,817

i.rra
1,891 1,910 1,970

2,000

......
1,180

Shear

...

217

388

596

862

1,000

2,084 2,120

Yield strength

0.020.
0.025. 0.032. 0.040 0.050

0.063
0.071. 0.060

91 113 132 163 188 213

98 150 198 231 261 32\ 348

....

110
200 265
204

273

o.oso
0.100 0.125. 0.160. 0.190
..

..

32\ 402 463 498 537

389
471 538

616
685

270 345 401 481 562

633

419 5\'
.'i57

filO

706
i88

362 ;38 614 669

i61

811
902

'94

623
746 854 1.018

982
1,053 1,115 1,357 1,694 1,925

861
1,017 1,313 1,574 1,753

745 836

842 913
1,021

I
be computed.

NOTII:: Th. v..h... te tail table .re bued 011 "cood" m.llu r."turinl' pn.ctic:e, ..ad aay d.-vi.tioa from Uailo wiD produee lil'11ifl.untly re<luc:e<l valu...

G ShMt ..... ia th.t of tbe collllWnllnc ,hftt. III _ . wbere th. lower Ibeet ia tbiaaer thaa Ua. Utipel'. Ua. aa-ubn.rinw allowable for tlw 10_ ItMoK-ri.. ~ combill.tian oJIouJd

~ InereMed .ttelltion Ibauid be paid to dataU d.il'11 ia ~ ...ban D/t>4.0 becallH of ~ibJy lr'rfttar illcidenee of dlflleulty in ""ice,

e Yla!<l .. ahl_ 01 the IbMt-rh'et COIIlbtn.dolll .... leu Ulan 2/3 of the io<licated ultimate ...Iu-.

01. 22

FITTINGS AND

CONNECTIONS.

BOLTED AND

RIVETED.

Table nr. 8 Aluminum-Alloy She.et and Plate Bea F a t i rtng- actors (K - ra 10 a actual bear-ing strength to 100 ksi)

)'!aterial

Thickness, in.

'"

A values

K (Cltlmatel

--------I~------!:-','CI D':::::2::,.~OI_,''CiD~:,I'.:~'lllc''C! D..:::2~'ol e/ D = 1..')1 eiD = 2.1)!-:iD =

2024-T42 (heat ( <0.250 treated by user) . , ( O. 250-0. 500 . 501-1. 000 2024-T3. <.250 r .250-.500 ~024-r4 .. \ . SOH. 000 2024-T36 ... -;:.500 2024- T4 (coiled) ... <0.063 <.063 Clad 2024- T42 .063- .249 (heat treated by 25~ .499 user) . . ... 50D-1. 000 .010- .062 Clad 2024-T3 . .063- .249 25~ .499 Clad 2024-T4. 50~1. 000 .019- . 062 Clad 2024-T36 .063- .500 Clad2024-T4 r .012- .062 (coiled) . . .063 <.063 r Clad 2024-T6. . \ 5. 063 <.063 Clad 2024-T8l . 5.063 <.063 I Clad 2024- T86. 5.063 016- .029 040- .249 7075-T6. .063- .488 .500-1. 000 .015- .039 .040- .062 .063- .187 Clad 707.5-T6. .188- .249 .250- .499 .500-1. 000 .015- .044 .045- .249 717B-T6 .. .250- .499 .500-1. 000 I .015- . 044 i .045- .249 Clad i1 rs- T6 .250- .499 .500-1. 000 ! <.039 ,I Clad 2014-T6. 040-1. 000 , I 2014-T6. . I 040-1. 000 I 5052-H32 O~H) 5052-H34 r;~H) . .... .

.
K O'jeld)

\-aJlJes

1\ Cntimare)

K ('Yield)

I ,.'j e!D= !.O! el

1

D =

1..5

I ( I

1. 18 1. 22 1. 18 1. 24 1. 24 1. 20 1. 33 1. 18 L 06 1. 12 1. 18 1. 14 1.14 1. 20 1. 20

0.93/ 0.641 0.56/"1 .96 .61 '~I' .93.61.53 1, .~ .ro .69 1.29 1.02 ,,98 . 74 . 64 1. 27 t. 01 .% .ro .62 l.~ 1.~ 1.05 .96 . 8 4 1 . 3 7 1 . 08 1
.~

.
0.82 i8
0.71

.77
1. 00

.M

.56

1.26

.~

1.16
1. 20 1. 27 1. 10 1. 16 L 14 1. 18 L 22 1. 27 1. 33 1. 35 1. 44 1. 49 1. 39 1. 42 1. 33 1. 37 1. 39 1. 42 1. 35 1. 39 1. 58 1. 60 1. 51 1. 51 1. 44 1. 48 1. 40 1. 40 L 22 1. 24 1. 29 . 65

Ii

I{

l{
i I

I I

I{

5052-H36 (H). ..J 5052-H38 6061-T4_ . 6061-T6.

O~H

.71
. 78 .62 .63 . 88

'

I..

.1

.84 .54 .481.10 .89 .58 .50 1.16 .93.61.53 .90 .58 .50 . .90 .73 .64 1. 18 .95.74.64 1. 24 .95 .74 .64 1. 24 .92.67.59 1. 24 .95.88.77 1. 25 1.01.93.81 1. 31 .87.59.52 1. 16 .92.61.53 1. 20 .90.75.66 .93 .78 .69 ... .96.90.78 1.00 .94 .83 1.051.04 .91 1.061.09 .95 1.141.06 .92 1. 48 1.161.07 .94 1. 50 1.081.00 .87 1. 42 1. 10 1. 04 . 90 1. 47 1. 05 . 98 . 85 1. 39 1.081.01 .88 1. 41 1.101.02 .90 1. 42 L 46 1.121.04 .91 1.05.98.84 1. 39 1 . 0 8 . 9 9 . 8 6 1 1. 42 1. 61 1.24 1.17 1.021 1.261.181.04 1. 63 1.181.11 .96 1. 1.181.12 .97 1. 55 1.141.07 .94 1. 481 1.17 1.10 .96 1. 52 1. 09 1. 04 . 90 1. 44 1. 44 1. 09 1. 05 . 91 .96.90 .78 1. 22 1. 27 . 98 . 93 . 81 1.02.96,84 1. 33 .50 .34 .29 . . . 5 4 . 3 8 . 34 1 . . 59 . 46 . 41 .. . . . , . , . 62 .531.461. ,..I .481.26 , .67 . 58 . 50 . .. 1 1 1 I

.87 .92

.....I' .
.76 .78 .78 .74 .93 .96 61 .64

.66 .56 61

.69 .67 .88 .57 .49 .53

.93 .98 .98 .96 .99 L 04 .92 .95

.6i .69 . '9 .64 .81 .84 .53 .56

1. 1. 1. 1. 1.

551

17 19 10 15 10 1.11 1. 12 1. 16 1. 08 1.11 L 27 1. 29 1. 20 1. 20 1. 17
I. 20 1. 12 'I 1. 12 . 96 ,

1. 10 1. 12 1. 04 1. 08

1.02 1. 04 1.06 1. 07 1. 00 1. 02 1. 20 1. 22
1.14 1. 15 1. 10

1. 14 L 06 1.08 90
99

1. 01 \

.96 I

I
! i

1.

"

.,

O~.I ....I.
... ,

.97 .98 .90 .94 .90 .91 .92 .94 .87 .88 1. 05 1. 06 .99 LaO .96 .99 .92 .93 . is .64 .67

i
,
!
I
I

.22'/'

Far tiD >;>a.Jtles between I.S and 2.0 bea.riug factars may be obtained by liDea.r interpolation.

(t-cdge L!istanc~. D-hale -ncmevcr.'

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

D1. 23

Table D1. 9 Unit Bearing Strength of Sheet on Rivets, Fbr '" 100 kst

'Unit bearing strength for rivet diameter indicated, IbQ

Sheet thickness in.
~& in.

~in.

H in.

~in.

~e

in.

H in.

~in.

~ in.

0.012. 0.016. 0.018 .. 0.020 .. 0.025. 0.032. 0.036 0.040. 0.045. 0.050 .. 0.063. 0.071. 0.080. 0.090 .. 0.100. 0.125 .. 0.160 .. 0.200 .. 0.250.

.....

...

80 107 121 134 168 214 241 268

302
335 422 H6 536 603 670

173 192 24<> 307 346 384 432

..

...
321

411
463 514
578

....

.....

838
1072 1340 1670

480 605 682 768 864 960 1200 1536 1920

2400

643 810 912 1028 1157 1283 1606 2056 2570 3210

509 5i2 636 il6 795 1002 1129 1272 1431 1500 1988 2544 3180 3970

688 764 860 955 1203 1356 1528 lil9 1910 2388 3056 3820 4770

1285 1619 1825 2056 2313 2570 3213 4112 5140 6420

2035 2293 2584 2907 3230 4<>38 5168 6460 8070

2741 3088 3474 3860 4825 6176 7720

Table D1. 10 Unit Bearing Strengths for Pin Size Indicated; lb. a

Sheet thickness, in.

Bearing strength of sheet for rivet size indicated, Ib

,
He in.

rb in.

H in.

%, in. He in. X in. Yte in. Ys

1 - - - ------

in.

in.

0.025. 0.032. 0.036. 0.040. 0.045. 0.050. 0.063. 0.071. 0.080. 0.090 0.100. 0.125. 0.160. 0.200 0.250.

156 200 225 250 I 281 1 313 394 1 4441 5001 5631 6251 781 1 ! , 1, 0001 1. 250 I I 1 '631,
,

,,:)

234 313 . . .... . ..... .. 300 400 500. ..... . 338 450 563 675 . ... 500 625 750. 3751 I .. 422 563 704 845. I 1 .. 469 625 781 940 I, 250 . 590 i881 985 1, 180 1, ~!~I 1,969[ .... .. 665 8881 1,110 1,330 1, 1/,)1 2,2191 2,663 . 7501 1. 0001 1,250 1,500 2, 0001 2.500 3,000. I ... 8451 1. 1251 1, 4071 1,690, 2,250! 2,813 3.3751 4, 500 5.0001. 938 1,250 1,562) 1,8751 2,5001 3. 125 3, I 1, 1iO 1, 563 1.953 2. 3401 3. 1251 3. 9061 4, 6881 6.2501 7,8121. 1. 500,1 2,000 2,5001 3. 0001 4, 0001 5, 0001 6. 0001 S.OOO! 10.0001 12.000 1. 8i5 2. 500 3, 125 3. 7501 5. 0001 6. 2501 7. 5001 10. oooi 12. 500 15. 000 7. 5001 ~O. 000 I I '68816 ..oJ', '01 813'i . 9 3" 18. 7501!' ~l. 875i 25.000 M, 3 441 3, 15' - I3 , 916 "t. / ,)1. 12.500! 1'.625 ! i I ! I I

..

% in.

d
IS

10.

1_ _ m_._

J I
i

iSOI

'

I,

Rear:lnlt ealues are based Oilareas computed uslnll:the nominal Pill

dlamet~~

tndtccted.

D1. 24
Si~ce ~ivets ~re

FITTINGS AND

CONNECTIONS.

BOLTED AND RIVETED.

~rotruding

ass~~3d

same size, all rivets are to share equally in resisti~g H and V

2117-T3 material and of the type.

head

Loads

on each rivet due to Hc.g. = 8000/6 = 1333 acting in H direction and to the right. Load on each rivet due to Vc g 3~)O/5 = 500 lb. ac t tng down.
Lo~d

The ultimate desi~n tensi8n load in the sheet Inc Iudfng a 1.15 fitting factor of safety is 1000 lb./inch. The limit fitting load i~ 2/3 x 1000 = 567 Ib./in. The margin ot satety of will be deter.nined.
th~

sheet splice

From equ~:ion (15), the load on a rivet due to ~.g. on rivet group equals F ~ Mr/I. 1= !r:l ::;: 1.625:1 x4+J.625:1 x z
~onsider

= 11.4

rivet

mark~d

c;
lOOO*/ln.

1.625 = arm to e.g. or bolt group.

Fe = l'lr/I = ("000 x 1.625)11.4 = 1280 lb.

Sinee rivets b, d and e are the same dt s tance as rive": c from the e.g., t he moment load on these bolts Nill also equal 1280. Fig. Dl.38 shows the H, V, and M loads on the rivets b, c, d and e. Since the ar.n r to the rivets f a~d g is only 0.~25, the load due to moment will be cons tderab Ly sraa.lLer' and thus these rivets will not be critical. Observation of ?i 6 . ~l.~a shows rivet c is the rivet with the l~rgest resultant load.

O F ' -- ,:3
~

-I 1-16
I

00 00

, 0 0 - .L
00

0 0 - .;-

I"

lOOO*/in.

o o

-I 1- 5/ 16

fO'.....

Fig. Dl.39

ScLtti:m:As an ~nalysls ~~l~J a width at sheet equal to the riv~t pitch ot 1 inch will be used. ~us load on 1 inch unit = 1000 lb. Check Tension in Sheet at Sect10n Through Holes.

~FH

VZ3+1280xl ..S/1.625:;: 2513 lb.

LFv :;: -500 - 1280 x 0.625/1.625

= - 992

Pt(~llOW)

= A ~tu

A
Hence , rt :;: /251:3:1
+

99Z:I 32700 D.
doubl~

= net area = (1- .159).04" .0336

Tabl~

(.159 = drill diameter tor 5/32 rivet, The riv8ts ~re 1~ ~at8rial is Zl17-T3l. shear. Rivet
D1.6)

Ftu for 2024-T3 clad = 60,000 psi. From Tahle 01.5, Single shear value = 1760 lb. or do~ble shear strength = 3520 lb. Bearing strength or 1/4 rlv8t on the.071 clad char~el section from TaJles 01.9 ~nc Ol.~ is 1325 x 1.20 = 2190. Since rivet bears on two ~hannA1s, bearing str~ngth of Jne rivet ~ 2 x 2190 = +380 lb. ~ivet shear
202~-T3

Pt(allOw) = .0336 x SOOCO ,. 2016 lb.

n.s ,

= (2015/1000) -1 = 1.01

Check Shear of Ri 'Tets.

i3

:riti~"l.l.

M.S.
~ ~~j

= (3520/2700)-1

= .'30

As a ~ro~l~m for the reader, c~nge rivets g to 3/15 dl~~et8r and cete~lne whether ~~J8t attachment stIll shows a poSitt78 ~rgin o~ s~fety (use equation 16).
PF..J3L:::M '3.

Rivets are in slngl~ shear and two rivets act in the 1 inch w~lt which was assumed. From Table D1.5, single shear strength ror 5/32, 2117-'1.'3 rivet ~s 596 lb. The strength factor ~lddle table ~r Table 01.5 ~or .04 ~heet thlckness i3 .964. Thus for ~NO rivets the shear strength is 2 x .964 x 596 = 1150 lb.
~.s.

= (1150/1000) -l ""
Jf Rivets
~n

Check

Be~ri~g

.04 Sheet.

::g. J1.39 shows a lap joint tnvo Lvtng rows of riv~ts as shown. Sheet ~aterial is 202~-~3 ~l~d, ~nd r~v~t3 ar~ 5/32 di~~eter and
~N0

From ~a~l~ Dl.9, :h~ ~ltl~ate bearing jased en ~br of 100,000 psi for 5/32 rlve~ on .04 sheet = 536 lj. Then ~eferring to
stren~th

ANALYSIS AND DESIGN OF

r~bla

FLIGHT

VEHICLE STRUCTURES

Dl. 25

Dl.8 for

202~-T3 c~~d

mc.terial and an ,I I.;

6"../
~

e-D r-at t o o~ 2.0, ',lie find correction factor K ; 1.14. Tr.ersf0re rivet bearing strength is 1.14 x 636 x 2 ; 1~50 lb.
~.s.

Ij
I

1/

;.;i'
8(./

(1450/1000) - 1

:::r.

.45

275

ill

25

Check

Riv~t

Shear Out.

Since edg~ distance is 5/16 in. or aID 2.0, shear out strength is satisfactorJ.
PRGSLEr'! 4.
ASSU,rn.8

Fig. Dl. 40

~/,~,., s
r i vets are changed to the so11d

""!
150

d1.:::npled type. ..Jhat 'NOUl1 be the M.S. for the rivets. Referring to Fig. D1.27, we find the sheet thicknesses are such as to prevent doubl~ di~pling. From Table 01.11 and Dl.12, we obtain the ult1mate :311d yield strength of a 5/32 rivet on .04 sheet as 635 and 506 Ibs. r-espec t I 'fe ly .
vnence , U1tLnaC8 M.S.

100 0

- .

425 Point 3

~~" i25~~
150 Point 7 Sketch b - - -

Sketch a

(2 x 60.5/1000) -1 = .27

Y:i.=1(i :1.8. ; (2 x 506/567) - 1 = .49

Ri'.fet line (3). The skin is cont Inuoue OVer stringer at point (3). Sketch (a) shows a free body of the skin and stringer at pOint (3). Since the s~ati~n of the fore as parallel to the stri~ger must equal zero. it is observed that the load tran~ferred to the strlng3r is
150 Ibs .vm.

NCTS:

In checking tensile strer~t~ of sheet through hol~ section, the drill size for dimple1 riveta is sligh~ly larger than for protrudiQ~ ~ead type.

Rivet lines (4) and (5). Since the sheets end over the str-inger. the load in rivet lines (4) and (5) are ~~ and 275 Ibs./in. respecti1ely. Rivet line (6). 150 los./in.
Rtve t

PR.08u:r1 5.

load

= 275 -12.5 =

This is a typical problem involving the rivet loads in a Sheet-stringer type ot cons tr-uc t Ion as t Ljus tr-at ed in Fig. 01.40. Ber ar-e the t-t vet size and SpaC1:J.g at the cernes (1) to (10) can be det8rmined, the rivet loads at these pOints must be known. The shear flow in dlrecti~n and ma~itude on the webs and skin are shown on the fi~ure and are in 1bs. per inch. These values represent the results in one of the flight conditions. The structural designer must look at all the shear flows in the ~arious Clight and landi~~ conditions in order to obtain ~he critical rivet loads. It is assumed the shear flows as shown include a~y diagonal tensi:)n ef~ect in the varicus sheet pane 15.
!he ri1et loads
i~

Rivet line (7). The skin is lap spli~ed ever the stringer at paint 7. Sketch (b) shows a free body. The load produced on the stri~er 1s 1.'10 rr-on equt Ltor tum, 'Phus the worst Shear load on the :rivet is 150 Ibs./in. which 1:=; greater tr~n the shea, on another cross-section of the rivet ~hich equals l~~ lbs./in. as the shear flow in panel 6-7. Rivet Load at (3) ~ 175 - 25 ; 150 Ibs./in. Rivet Load at (9) = 17~ Ihs./ln. Rivet Load at (10) = 576 L:s./i:1..
01. 25 Rivets in Tension.

Ibs./in. at rivet

li~es

1 to 10 will be as roj.jowe r-

Rivet li~e (1). Since .05 vertical Neb ends at ~oint (1). the shear flow of 1075 lbs./in. in the ver~~cal web ~ust obViously be reected by the rivets in ~17et line (1), thus load on rivet line
~l)

is 1075 Ibs./in.

Rivet line (2). By the same reasoning since s~i~ ~nds at )oi~t (2). the load on rivet line (2) equal~ she~r :l~w in panel 2-3 or 575
Los ./in.

Great: jud3;llent Should be used in :lSing rivets in tension. There ~s a general saying, nNever uSP. a r~vet in tsnsion. n If this r~ qut reaent HaS s cr fc t Iy rc i icwed, it vou.Id be difficult to deSign a c~nventional airpl~ne. For example, t~e s~!r. ~n the upper surfa~e of the wi~.g.jue t o tne upward suc t t on a t r rorces pl~ees the rivets that holj t~e skl~ to t~e stringers ~nd ~ibs i~ tenSion, however thAse tension loads in ~ost cases are rel~tlvely small. The followin~ general c"ite,i~ a;ply relative to rivets in tension.

Table 01. 11 tnumate gt reugth of Solid 100 Dimpled Rivets

lJll.imllu.~ :>tfl~lIglh, III

._-------------Rivet matunut
2117 '1':1

..

~--------_.-

'.!U17 -'fa

'.!(I'l-I Tal
._-~-

._---------

t
Hivd ,lililuetu (in.) KI...,'I II,i"kIW";'-;, in.
tlUlti

""~ 1 T:l, 202-1 '1'4, '.!tl'.!-I TH,

alLll

'1.1I:l1
1'l:Ili

Ta. d '1'-1

,.
Z
t;;)

'.!U:H '1'6, 2(l:!1 TMI, '.!1J'.!ITHli, 71175 Tt)

"lO'.!-I-l':~,

'.!tt.!-I-- 'I'!'ili
lUlll

:.!U:lI,T-I, '.!1l:.!1 -Tti,

7075-'I't)

'.!O'.!-I-'I':I nntl '.!u'.!-I-T-I

'.!tt~-I

'ru,

:lUU '1'1:11,

nnd
2trl1 -Tlil
.-----~

:.!ll:.!l THti, uu,1 7075--'I'ti

-------

:lH'.!-1 Ta, nud '.!O:.!I n

'.!U'.!I Tti,'.!O'.!1 '1'1:11, '.!ttl-1 TXli,

aUII

707ri 'I'li

:ltl21--'I'HI

o z z

,.
~.

,.

..

- - - -.'>"

.._--

--

,.
I

." -

~:12
-

------ ------

" ."

---_.

" ."
--

H
-----_.--

kl
---_.--

'1_-'

."

V.
-

,1(;
--

'" o
Jj

----

.-

'o"
H

00

177
:"H\I

. .

. .. . ..
7-14 7Hli !IH'.! I,I!'I'.! 1,'.!77 1,:1:1'1. 1,:IXn I,I:.H -IIi'.! !',!l1I liIl;1 771:1
l:lItl

III

O_H'.!tI lHI'.!;-' otl:l'.!

'.!:\:.
:l:,7

uutu
II 0;10

'.!TI

:l\l\l :liitl lI:l -I;'i I

all'.! :11:1:1 171

r-

4:d ;-,liX

n'1.
HaU
\110

n!'l

:,tl;-l lj:I:1
;,11'1 li\l:l T\li
-cc

X!H

-I HI titMI 7'.!H

san

mH
!)l15 1,0!17 I,:l-U1 1,:ltII 1,:157 1,-1ll:1

ti7:.! 77f!

<xl

Il.tll':l

tUl71 (Joxl tl,tl'.ltl

II. teo

,.101

,01'1. ,HI!'I ,071 ,tl!IH

Htii

I,U:Ui Kill 1,1-12 II'.!:.! 1,I!ltl !I!'JH t,'l.:\U 1,2H7

am
\l:m
\157

H'.!'.! 1,lltlll I, 15:t I ,'.!t\7 I,:ll!i I,:I!'IX I,:I\IH

8"
I,a:l'.! l,ti!l5 I,HS:I 1,9!l5 2,11f! 2,:l:lU

I,IOH 1,5l1H I,Hm I,u:m 'l,H-clA '.!,1-15 2,:l:l:l

11,11

1,110 I,:.!:m I ,'.!!) I 1,:1-10 I,:tl:l'.!

H7!1 1,:lW l,n7 I,HH:I '.!,n'.!!i '.!,15U '.!,'.!5:1

1,:lOll 1,711ri '.!,om 2,1!'IO 2,'.!litl 2,:m:1 2,lfi!'1

''" "
Z
H

t;;)

,.

t;;)

Non:; The v..lue. In lhl" tllLle are bll:>t:d on 'a"""" any devi,li.. n 'rom thl. will I.roduce IIllfnll\cantl, red""ed II Theile ,llowlbl"" aplll, to double dhnpkd .heel.ll and into a ma"hine-cuunter"unk lower abeO:!t. Sheet Kille i"

manul,ct"rlnll prldlce ".01 valu"". to the upper IIheet dl"'Ilell that .d tbe thln"""t .hOlt

,,,~ double "lml.l"d jointa and or ."nk j .. lntlo. Tlte lhleknea8 0' tlhulated Kkllol thicker tban the by utrapolaliu" 'ur .kln lIall""

th" "pper dimple Ihe<l!t 'or dimpled, mlchine-counlerthe mad,inecounte""nk 8h....t mud be at l..ut 1 upper .b....t. IJI JlO ea.e IhlU allowa"leB be "blained ..tiler tban tboae .b..wn.

., " '" ''" "

t;;)

Table Dl. 12 Yidd Strength of Solid 1000 Dimpled Rivets

Yidd lMeugth, III

Itivd lIlall~l'ial

2117-Ta

2017--'I'a

:.m:H-'I':11

('Ia,1

~1.1'1'1

11l;lll'riitl

2 11~ 1 'I':~

:!OJITI
:m.~1

Tli ,

:!021 '1':1, JO:!-I '1'4, :W:,H Tli,

JII21 TSI,

2tt~1

-'1':1,

and

2024 -'1'4, ~tl2'1- 'I'll, lIud

:!O~l--'I'SI

:.'.O:.'.4'!'Mti uml 7075'1'1)

:.'.0:.'.4 '1':1, ~O~ .. '1'.1, :m:H-'1'(1, ulill

:!():.'.')

;!m!t -THti

20;!-t- '1':1,
IUlll 'lIt!-1 '1'1

~I~H-Tti

:.!O:H 'I'Xli
,,1111

ami
71175 Ttl

und
:.!O:H-Txl

TSI

7U'if, Tii

" o
'"
~

'"
o
Z

\ :!O~H 'I'H I

aUlI
~1I:!-I-'I1iti
----

o .., ..,
r0: ...

Ilivd ,Jiallu'I"1

uu.t

.,.:t:!
-

-_...

----

'.
'.

:1,
-

." '"

,.

-- -----

," " ----_.

,.

--- ..

'.

%;
- -

~:i
----

---- -

'"

''

,.

---- -

'.

,
y.
----~-

'" ---_.-

a>

h
-

-_..

_.---

'"

7-l

,
-

'IIi

H
_.----

-- -

- --

"';III"I'lllli"klll''''',

ill,:"

umu
1l0~O O,IJ~:,

IX I :11 1\1
\I

""

. .

..
.

... .
711fl Hli7 I ,1M)? I, II t
..

257

.
110
fll:.! :i25 lill liUti 7r-.7 ii77 KII j:!!1 7.'i2
.. ..

:Uf,
WI Hti

:~2,1

;I:ili tHO IX:I 7S2 [IH\l UOr-. IISI H!I.'\ 7IH 1,0:1-1 liX 1,U70 I, ron flHi

lI_Il:U

am -t:m
f,Oti .'i71 lilH till

omu
011.-,0

IH

7:iO
1,IMIIl 1,ll'iti I,IO:! I,I-I:!

lUlti:1 U.o'l
1I.IJI'iO
1l,O~)O

'"*'

"'" !lHt
uau

1I.IlMI

1W 5XI Hili ti75 1,IH7 tnn I,llli Slti I,libti KI2 I,H711 2,U57 2,:l'lU

0:

"' '"
o

flH~

r.tsn
I,I!IH 1,:!:11

tl7M 1,fIOB I,HO:.l I,!l:m

2,OH
:!,I-Ifi

tiliti 7:IH H:l5 1,045 1,152 1,2-lli

H7~)

tim l:Uti

7XII !Hi2 1,:lOH 1,5ti1 1,711 I,!I:!H

r-

usa
1,1r-.:! 1,:l77 1,:1:12

:l,2:l2

I,:UIH 1,[lli-1 1,711 1,!tlB 2,121 2,25fJ

ntH
l,lItig 1,115 1,177 l,a2'"

I,:nm
I,H-I

2,121
2,2liH

!l7S 1,fl.1;1 1,!HiS 2,1-10 :l,:!tiO 2,;!tia :!,-15!'1

'"

NUTlo:: 'fbe v.luea In U,ill Lable .re a..sed on "lJooo1" any uHi.. tion frum thia will produee allolnifieanlly reduee<t CI 1'he~e .. Uow.blea apply Lo dt,,,b1e dlml,led aheela .nd Into m .,blneeounl"rllunk low".. ab""t. Sheet ""1101 la

mllnuf.d"rh.&, .. r.etie" and val"ea. to tbe "I'per abed dimpled tbat of tbe thinn"t IIbeet

for duul>le dimpled join'" and of aunk jointa. Tb" thiekn_ ot t.. bulat"d lIalle thicker tblan tbe b)' extrapolation tar akin &,allea

tbe llpper dimple abeet t"r dimpled, m.ebln"-"ount,,rthe ma"blne-e"unteraunk ah"et mud be at least 1 upper ..bed. In Du elUle IIhall .llowabl.... be obtained other tban tboae IIb"wn.

01 28

, , FITTINGS AND CONNECTIONS

, BOLTED AND

RIVETED

(1 ) 'l'ens t on on rivets shall be restrict~d to condt t i ons in ~hich tens10n load Is inc t 1ent8.1 to the :najor shear' ca.rryi:lg pUT';;ose of the r ive t , 'Ailen It Is :1!flcCllt to ce te rmt-ie If t.ne tension cQmponent Is inc i-iente i or -ea jcr , a bolt shall be :.lsed.
(2)

Table A PROTRUDrnG HEAD RIVETS (AN4'70, AN442) ULTIMATE TENSILE STRENGTH

E[]3
ut6
.020 .025

E@]:3
223 31l 420 561 736 758 758

The

r-t vet s

8.re examples of jOints where are considered to be satiSfactory tBnstun c~rrylng mediums.
[a ) Skirt atrtachment to ribs and frames
Co) Atrtacnmerrt of sheet panels to beam

ro Ll.owtng

Use this table for 24ST Alclad sheet and harder Allowable Rivet cad s. Por rver Sheet Gaulle l/4 5/16 3/8 3/32 1/8 5/32 3/16

.u32
.040 .051 .064 .0'72 081

159 214 277 277

l~~

142 197 269 353 471

flanges and stringers, where interrivet buck.Lfng or diagonal sheet wr ink'l Ing produce tension loads on r ; vets.
(c)

495

354 47. 549 8 981

1094 1094

568 799
lu" 1245

929
126" 1482 1'721 1669 1952 2622 3353 4336 44'70 4470
22~~

1440
1651 1882 1982 1982

.091
.102 125 156 .188 .250

i~~:
3130

2890

5:<.i.:1 attach!:lent on a pressurized nac eLl.e or body.

,13u

(3 )

Do not uso r t vets to raat en control brackets to a supporting structure.

It' there Is no load reversal on the asse:::lbly, the tans ton 8.11owablss ~i'lJ3n In the folloWi:1g tables can be used.

Table B lOCO FLUSH HEAD RIVET (AN426) MACHrnE COUNTERSUNK JOINT ULTIMATE TENSILE STRENGTH

(4)

~\ L-tmin

--.l

(51

It there is load :-eversal on the assembly, the tension load on the r-i'let enoujc not ex:::eed 25 percent of the values In the
t'i~l~.

(6 )

Rtvet.s loaded in both shear and tension

~
.051 .064

Use this table for 24ST Alclad sheet and harder Allow Ie Rivet oad L a. Per Rivet Sheet l/4 5/16 3/8 3/32 l/8 5/32 3/16

cnecked tor combl::1ed stresses, :.lSi:lg the tnter-ac t f on equation,

shou.Ld be

191 249 2.9

.072
.081 .091 .102 125 156 188 .250

319 438 446 4.6

Rt lll
IC)

Rs

=1

501 592 683 583

653 773 912

'85

985

sufficisnt numb8r or rivets shall be l..;'s~d to insure tnat failure of any one rivet due to improper installation, cracked nead , etc. , shall not result 1n the ta t Iur-e of the 3tructure that Is being :-,911 t.ocetner- by the rivets.

12'75 1698 1783 1783 . 102

1941 2660 2817 2817 . 125

2950 3827 4023 . 156

tmtn

.040

.051

.064

.072

D1. 26 Rivet Tension Strengths.

Table C 1000 FLUSH HEAD RIVET (AN426) DOUBLE DIMPLE ULTIMATE TENSILE STRENGTH

aererence to the strllctures des Ign manuaLs 0: vaz-tous at rc r-art companies shows ':hat rivet t-::':.slle s tr-engt hs ar-e not the same or, In other wcrcs , r'.:Jt suanca rd ized as 1:1 the case of shear ':.nd bear-tng rarc se ,t.., B & C ;;'~'''''J been taken ::Y',~:::, :1"16 ;:=\ef. 0 ':'he va.lues g iven ar-e " '~:J:1:3-'Jr'.'2. t ~ ve re Lvt t ve ,0 val'..l.es r ounc In other ~om7a~y manuals.

tmm--=r~t

.l-

r- t min

Use this table for 24ST Alclad sheet and harder Allowable Rivet Loa Lbs. Per Rivet I Sheet I 3/32 GaUlle 1/8 5/32 3/16 I .020 103 .025 137 168 .032 185 233 271 2., .040 305 40s 382 . 051 408 552 465 .064 446 635 737 .072 850 .081 970 .020 .040 .025 .032 tmtn

1/4

694 I 931 ,
1077 1242 .051

ANALYSIS

A~D

DESIGN OF FLIGHT VEHICLE STRUCTURES

D1. 29

PROBLEMS
(1) The si~le bolt ~ittl~g ~nit as shovm i~ Fig. 1 is suo jec ted to a design r i t't t ng load of 12000 Ibs. in axial tanslon. The pin 1S an _~~ Steel bolt 3/8 inch diameter. Bushl~ is 1/16 wall and steel Ftu ~ 125000. Lug materi~l is 2014-T6 bar, Ftu ~ 58000. The fitting is not subjected to shock or vibr~ti~n. Strength cbeck the bolt and lug (A) and give all margins of safety. Same fitting as in Problem 1 but design fitti~g load is a transverse load of 1O,!JOO lb. s tr enctn check and stve all aa rgfus of safety. Same fitting as in Problem 1 but l~g (A) is eubj ected to a design fltting load acting at ~~o with a value of 11000 lb. Strength check for this loading.

/7(JQQ

!0/~o

(2)

__

,, , ,, , I , ,' ,, , , , , , ,, , , , I I ,, .--.1
,

-"

--

:'
T

"

" "

"

Fig. 2

(3)

~ srlt'l. ?tAT~$

/I.r. ..t8Q1JIO

'.

t
+-p

,--.----'---!,T---,:-- - p.... 1/2";::: L~ B

LugA,_,,_, _ _ : :: :'

~l/4
_P/2

'0-;-'<---,----";.-' -+-"LnaA',_ 1 ','1 "0 - "~

"-J_

P/2

!~;;"'j
Fig. 3

Fig. 1
( 4)

JJ
(8 )

;11

A 1/2 inch diameter AN st8el bolt is SUbjected to a combined shear and tension load. The shear load on the bolt is 10,000 lbs. and the tension load Is 12,000 Ibs. Find margin Of safety under this combined lO~ding. Design a hi~ge pin USing a standard AN steel bolt and a male lug to carrying an axial "tensile load of 25,000 Ibs. Use fitting factor of 1.15. Use steel bushing. No shock or vibration. Assume lugs of female part ot fit"ti~ 1/2 as thick as mala lug. Design the ~le lug from two materials. (1) 2024-':'4 alumt num alloy and (2) A1S1 steel, Ft u = 180000. Fig. 2 illustrates an end fitting for a streamline strut. The tube is flattened Slightly at the end to fit a Simple bl:ck fitting. Loads ShOiVTI ~~e design strut loads. Using a. fitting factor of 1.20, check the strength of the ent~re fitting unit. Assum~ no shock or Vibration. Fig. 3 illustrates a :l~tlng unIt on the end at an extruded (I) section. The web on the (I) section extends out to form

part of the fittL'1g lug. which is re tnror-ced by steel fitting plates. Strep~th check the fitting tor a deSign load of 45000 lb. Use fitting factor or 1.15. Fig. 4 shows a typical beam end-single pin fitting unit. The fitting plate is att~ched to beam secti~n by the rivet pattern as shown. The loads shown on f1 6ure are applied loads. USing a factor of sRfety on applied loads of 1.5 and a flttlr~ factor of 1.2, determine the ~argln of safety of the rivet attachment of fittl~g plate to be~~ section.

(5)

;t (202'FI'J /IN(;.I.O")
'0

..f0IA.2117-TJ

=91-<I?T= :,

0- -:

I;; rf0

(6 )

\..;;,-'

~ Ol-i, Zfl7-1J
" RiVETS

.~

r
I
,

\ '-I-- ""
\ ' -

(lOl"-TJ~

,>+-;
,

.JL,-Fig .- ,-

~ G
~8
- s:

'- ""

(7)

Dl. 30
,;:1

FITTINGS AND
::L~

CONNECTIONS.

BOLTED AND RIVETED.

.3 Sh0'NS

':3.

::-1,,\:; hinge

::itt~:1~

a t tached

(12) In

s~?)orting st~uctur~

by 4 -

l/~n

cia.

A:'; s t ec I bo l t.s . :'::.~ belts e r-e 1:1 -iouo l s s aear-. Check t.ne .aos t c r-t t t ca L bolt in

shear
ac t tng

~nd

bearing
shown.

fJf

12ac of 3500 lbs.

2.5

Fitting matez-ta L is

8 find the 5iz9 )~ 201ry-r3 r~v~ts to carry the ~lt~~~~s fi:tlng deSign lo~d 0~ 3000 lbs. as ShO;ffi. The r i t't tng p la te is st.ee l 1/B thtck , Ftu = 85000, and the 2024-T3 channel fr~e is .081 lncn. ':hlck..
nes~ssarJ
,;~FKAA;;

Fi~.

201';"'-1'3.
CJ.J/UJ. . .
!

c:

J:'- j rt
i: "i~
~

(<'t: Of~. gr;~T~

...- 9-

3500

,
+---H----r--~

~,rtJii.

"I
, i
l...L4~
I

..J
Fig. 5

"

- --8

,-+

I.:: che eccent r tca Ll.y Loaded mul t tp.Le bolt !ltti~~ ~f Fi~. 5, ~~t-3~ine t~e resultant 1,::8.'.1 en ea-n cr <:;:-.':3 :'173.A.J.\I s tee I bolts in
resisti~~ t~e 2~2CC lb~.

Fig. 8

(13)

load acting as

the lig~test Qverlap sheet s~lice for .061 cl~d 2024-T6 ar.d protruding head tYPe rivet. Design tension lead on sheet = 2350 lb./ln. Give all details of jot~t.
Des~gn

(14) Rework design in PrJo19m 13 ':0 USl douc Le

,jimpled r-tvets.

\
(1)
Fig. 6

ANC-5.
!'"'..2.rC[I.

Streng~1

or

Metal

Aircrat~

Slements.

19.55.
~IL-HDBK-5. Aug~st,

(2) (11) 1:1 is atteched to a f~s~:age fT~ne by ::~~r l/~ di!'l.. _~\I ste~l bo1:s. ~~e ~eSi;:l fittiu 6 lJ!'l.ds are as s~~~n. Jete~l:l~ ::he ~~r~in or safety en ::::'e ~ost criti:;2.1 bolt. Use Fig. Dl.4
Fi~. :~e fitti~6

Military Handbook

Met~llic ~ateris.ls

and

~lements

lS62. for Flight

veni c te Struc turee ,

(3) Cozzone, Mel~on, and Hoblit: Analysis of Lugs and Shear Ptns Made of Aluminum and Steel Alloys. F~oduct ~g1neer1ng Vol. 21,

tc obtain equa t I cn f)'!' c omctnec tens Ion and she3,r stresses.

:-raJ,

1~50.

--...-<

l.J
-

-t

2~1:
I

1t//,-rT

1J-t
,I

tte Lcon and Hoblit: Sngineerlng.

(=: \ J~n~,

"DeveIopment.s in the
Pr-ocuc't
1~53.

AnaIyees of Lti;;S ann Prns."

-3 ---,--,------{3-! --:::---1. , --= ei

7800 ~T

s,
_
4

':.'--' ,'.-78
.

,"',
, . I

i,~-r

(.::'

Fig. 7

CHAPTER D2

WELDED CONNECTIONS

D2. 1 Introduction.

(3)
~verall

Since the

structure of

~n alrpL~eJ

missile or space v~hlcle cannot be fabricated as a si~gle continuous unit, such structures InvolvB many struc~ural parts which ~ust be
f~staned together. For ~e~t~ln materials and tY?8S o~ structural units, welding plays an tnportant role in jolnl~g or connec~lng s't ruc tur-a.L un; ts . Research 15 constantly going an to develop better welding machines

In general avoid weIds in t ens ton since they produce a weakening effect. In some connections it is dif:icult to avoid all tension loads on welds, thus weld stresses shoul~ be kept low and if possible incorporate a fishmouth joint or tinger patch to put part at weld in shear.
A weld should not encircle a tube in a

(4)

and

weldln~ technl~ues

and

al~o

to dev310p new
pr-oduc tng

materials t ha t can be we Lded

'N1t!lCUt

a detrlment~l strength influenCe on the base or '~welded ma~erlal. A tair Size book could ~5 written on the sUbject of welding and design ~orNel,jlng. This brief chapter C8.TI only be a tr:ef Introductl~n to the sUbjec~.
02.2 Gas Welding.

plane perpendicular to the tube length. Standard splices or joints for overlapping tubes, and end socket fittings in tubes have been developed, which require no strength check. These are the diagonal weld and the fisnmouth weldS as illustrated in Fig. D2.1.
(5)

There are two types ot gas weldl~g. namely. oxyacetylene and oXYhYdrogen. Practically all gas welding in aircratt work is oxyacetylene. Some welders pref9r the oxydrogen flame in welding a'lumt num a l Lcys because the flame is not so hot. The major aircraft structural units in which gas welding plays an important part are welded steel tubular fusela~es, engine mounts, and landing gears and the attachment ot ~late and machined fittings to such sta-uce.c-e .
(6)
02.3 General Notes on the Practical Deatgn of Welded .rctnta.

Tapered gusset plates should be incorporated in all important welded joints to insure gradual cr~nge in stress intensity in ~embers. These gussets lessen the danger of fatigue failure by reducing stress trrtens t t y.

BUTT WELD

FISHMOUTH WELD

FIg. D2.1

A weld over a weld should not be made.

(7)

To prevent burning ot sheet welds should not be ~de on both sides at a thin sheet. If two weldS are placed close together la:k of Shrinkage Space may cause craCking. Cracks usually on bends.
devel~p

The deSigner of welded structures in steel can greatly help ~he we~cer obtain good joints or co~~ect10ns by adherring to the following gener-a L rules.
'i)

(8)

(3)

if

weldi~g

is dcne

It is

~ucj easi~r

to

obt~in

a good weld

when the parts bei~g wel~e~ together are of equal thickness. It is general design ~ra~tlce to try and keep thi~kness ratio between the two welded ;arts lASS than 3 to 1. Seme deSigners try to keep within a 2 to I ~atio in order to elim:nate posSibil1tles 0: welders bur~i~g the thinner sheet.
(2)

(lO)

~hen tub~s are spliced by welding locate splice near one end at the tube, to avoid effecting column pr0perti~s. In g~neral it is not goOd practice to weld brackets to the middle of column members. Clamps are preferable.
inta~~l

Ill) In welding members together local

DeSigners usually
~inim~~ th1~kness
cons~der

.030 as

~he

to be welded in gener~l practical st~uctu:al joints as there is considerable danger that :he aver~ge welder might burn a thinner gauge.
02.1

stresses are set up. On most welj assemblies it is therefore customary to Rnor.nalize R the assemblies after welding. This heating permits' the equalization at the interr~l localized stresses thus preventing cracking in service.

02.2

WELDED CONNECTIONS

(12)

aircraft bolts should not be in ~lace since they are made of nl~kel steel and therefore cannot be satiSfactorily welded. Since standard ai~craft nuts are made of 1023 steel they can be welded in place if desired.
wel~ed

St~ndard

02.4 General Types of Welded Steel Fitting Units.

Fig. 02.2 taken from aircraft t~bing handbook of the Summerill Tubing Company summarizes the common t~es of tube termin~ls and disc~sses their structural merit. Fig. 02.3 illustrates the conventional concentric butt welded fuselage joint which tests Show is satisfactory where vibration is not present. Tests have shovm chat the fatigue strength of a Nelced jOint as illustrated in Fig. 02.3 -vhen the members are subjected to r-ever-se bending is reduced conSiderably, thus it is cc~~on practice to add additional joint rei~f?rcement such as finger ?lates or insert gussets as illustrated in Figs. D2.4 and D2.5 to jo Irrta subjected to vibration. to 02.8 illustrate methods of SPlicing a longeron at a truss jOint. The vertical and diagonal members strengthen the butt weld on the spliced member.
Fl~S, 02.~

nJl.
Fig. D2.15

Fig. D2.16

Figs. 02.9 and D2.10 illustrate fitting plate attachments to tubes. Except for light fitting loadS, the fItting plate should extend through to both sides of tUbe or to the adjOining members. The fitting type illustrated in Fig. D2.ll Is only used tor secondary conditions where loads or plate are relatively light. Since eccentricity of~ember forces on ~ jOint ?rod~ce bending in the connecting ~embers which ~y l~wer the fatigue strength of the jOi~t, such cases of jOint eccentricity as illustrated in Figs. 02.12 to D2.14 should be eli~inated in joint desl~.
02.5 Electric Arc Welding.

Angle "A" to be not less than 30

Fig. D2.17

3QO or more

of welding is based upon the generated in an electric arc. Arc welding tJ a lImited extont has been used for many years in aircraft fabrication. No doubt the fleXibility and general all around good reSults Jbtained wlt~ gas welding retarded its extensive \J$e, ~owever 1n recent years its use is inc~easlng rapidly as its eC8nomies and ad';antages ~ecome apparent to the deSigner, In arc welding :he apPlied heat is more concent rat ed and qufcke r welding results With less eXpanSiJn and warning as compared to gas welding. In ~he design Jf tubular joints, care shoul1 be taken to ~ke all welds as accessi~le as )OSs1ble. To sec~re proper s~ress distribution in arc wp.lded JOints the cesi~ner should folloW the ~ecomnendations a5 illustr~ted ~n Figs. D2.15 tJ D2.18.
~is ~ethod ~e~t

Fig. 02. 18

The fact that Leas expanat on and war-r-Ing takes place since the heat is concentrated makes it posSiblp. to hold to closer tolerances on parts requiring :nac:l~nt'1g after we Io tng an allowance of 1/16 ~r.ch t~ generall sufft~t~nt on most aSsemblies. Slectric welding Jerm1ts weldin~ of thin sheets as low as ,016 inch thickness.
D2.6 Effect of Welding on Base Metal.

Tests snow that cja in cat-bon ano c hr-cme-. mOlybdenum steels surfs:- very little in L~ss ~r tensile strength due Co 'Nelni~g. rGr CJld rolled sheet or tuo t ng t ne r-e r t nemenc ~:t cra m

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES Fig. D2.2

TYPICAl. WELDED STEEL TUBE TERMINALS

D2.3

CO'~.ECT I~.I
<

.-=--~
~

~
-~--.,.':"'"

.-

Fig. D2.9

jr::~""::'/%<~']?:~,:;:,,';. '""" _ at'."'W "-H

",~,<It".

"-",,"- _'" ~ '" ~ """" 'S

",,"y.~

I""

w",,"

"""Jft 01 ...

-:\tJoo
I

:-----------------1

;",",.I"""""_~ ",,,'I'''~ '",!Jr"'at'~,.,"",w""~,,,_

~.<..

...,....

' '.D .' n

; !:.

U'

may """l' , ......._ " " . .

Fig. D2.10

~~-.--",,-,, .""'''1>' ......, ,.-- <'>II:t a' f!W _

-~,...,-"" iit"'9u'.-I_tr _ _ 111'

i.~f N ~

~""""afa>i<l-""'9" ~

-11<I, '''''>f><J
IJ-.,

_ltltf"".11/'

.~ <""'pn1~. ;" ~

Fig. D2.11

Reproduced by permission from Summerill "Aircraft Tubing Data"

Fig. D2.12

Fig. D2.3

Fig. D2.6

Bending on tube tends to out lower weld in tension I~~

.--: I,;

"

Fig. D2. 13
Fig. D2.7

Fig. D2.4

---=.....-,-. -. 7":'"

Fig. D2.5

Fig. D2.8

Fig. D2.14

02.4
due

WELDED CONNECTIONS
the respectiye :r.in1":lllin tensile ul.t tmat e test 7alues.

:0 cc La vorkt r.g is lost in the material adj:;.cent to the weU wnl.ch Lower-s strength to a s:!J.3.11 degree. '(j' e Ldtng , however, dOes produce a mor: brittle aat.e rfa.L Nhlc':l has lower resisto.nc=; 'Co shock, vt crat i on and reYersal or stress, t aus it is customary to assume an efficiency or ;veld JOin.ts less than 100 percent.
T3.ble D2.l gives 'the allowable ultimate tensile stress ror alloy steels for material ac jacent to weLd when structure is welded after heat treatment. Table 02. 1 (Ref. 1) Allowable Ultimate Tensile Stresses Near Fusion Welds in 4130, 4140, 4340, or 8630 Steelsa (Section tmcknees 1/4 inch or less) Ultimate tensile Type of joint stress. ksi b 0 , 90 Tapered joints of 30 or less All others. 60

Table D2. 2 (Ref. 1) Strength of Welded Joints Matertal Heat treatment Fs u Ftu. subsequent to welding k,s1 ksi 32 I 51 51 32

Carbon and alloy steels. None.

I Alloy steels
Alloy steels Alloy steels Steels 4130. 4140. 4.340.

None, Stress renevec Stress relieved Quench and temper 125 Itsi 150 Itsi 180 ksi

43 50 60

72

85

100

63

75 90

105 125 150

a Welded after heat treatment or normalized after v-std. b Gussets or plate inserts considered 00 taper with center line.

02.7 Allowable Load for Welded Seams. The allowable lO8.d on a welded seam can be calculated by the following 8Quatlon:-

For we11ing member-s sub j ec ced to bending , the a Ll.owab Le zrocuj.us of rupture for alloy steels when welded after heat-treatment suould not exceed the following as specified In

Pa

F s u or FST (Lt )

- - - - - - - - - - - D2.1

(Ref.

1)

where,

For tapered joints of 300 or less, '-.lse modulus of rupture Fb equtva.Lent to that for steel ;..zav tng .< tu 20000 psi.

Pa

For all other types of welds, use Fb equal to .8 of that for steel ha'ling Ftu = 90000 psi. rnap ter C4 gives chart for dete:r:nlning the modu.lus of ruptur-e Fb for alloy steel tubes and Chapt.er C3 gives a procedure of detem iru ng Fb for other shapes Subjected to bending. Strength or 9ase zerer-te i -rnen Structure is Heat-Treated Art er- 'N'e11:tn~.

allowable load in Ibs . r s u 6,::1d FST from Table D2.'2 Iengun or welded seams in tnc nee L t t hi cknesa of tnmneec sa cer-re j joirl.ed by the weld In the case of lap welds between two steel plates or between plates and tUbeS. (1nches) t = av.:;rage t ruckneas In tncnes of the weld metal In the case of tube assemblies, but not to be greater 'than 1.25 ti:nes the thickness of the welded stock. 02.8 Brazing.

= = =

-eat-ereet.ec a rt er

Reference {li says that for

:tater~als

we Ld i ng , the allolhac Ls

stresses in cne parent nat e.r ta), near a welded .J J tnt may equal the a l Lcwabj,e stress ror the neet-ct'r-ea ted :::laterial. io other wor-ds , no yeductlon :Oor welding. However, it Is good ces ~~n practice to be conee rvat tve on welded j ot nt.s , tnus a reducti:Jn of 10 to 20 ?erc~nt ,1:' the heat-treated pr:Jper'ties is c reen used in ca Icuja t mg the tens 1l e or oeno i ng st.r-eng't o In ':he member adjacer..t to tne .veIc
D2.6 Weld-Meta.! Allowable Stress,

as app.l t sd :J a i rc r-ar t .'Iork is a steel oar-t s by -aeans or' a cop;er-z~nc atx-are , ,'11"-.1 G:-: is app.l t ed jy .ne11:1:: '....i '::-J. ao air-g::;.s :'2.':''"ne 0" ~" d1 p~i.n,S' ~n-:o tne .. o :' :r.e j-J l:::t 'no Lt en arxun-e . The ",r ~.~. ~""'~" de'Jencis on 'CM suriace areas of :~~t:'..c:: ana the
.' crocesa ' ,
~

Braat ng

um t mg

"",

",,.,,..,.~1,-

c Iear-ance oe tween

t.ae

car-t s to be

o t cec .

Al t nough t ce cr'az Ing .mxt.ur-e X?_y cevelop a ' shear-Ing s-crenc-;c 0::' 4,OOCQ 0 0 , a gener-a I ~llow ':l. __ ..:0 ~"" 105 <1052::1 1n a 1 rc :''1:t Nor k va Iue 0( leOOQ
>'~
~

,.~

~",<

because
t
j'J~r,:.

:ll.. any rac t ors , ~.,- :-: ~ ~ _~J.'"~:'lj- :;:~ s;';'~ll ne vor-sraan , e r rects the 5'::-eI:6':1", Jr a cr-azed

0'

Table D2.2 (from :te:". 1) gi Ires the allowable wejo-aeea; st.r-engt ns ~)r the var-t ous 5:ee1s. These design al1'Jwable stresses :cr the /'Ield mat er-ta I e re based on 35 per-cent ')f

':'he z-equt rement-s ~ :' cne or-ocur tng or govsr-rrnenc 3.gencie~ shOu:': be noted. ce ror-e us n braz i ng :n a tr-c ra rt vc r,c.

--

~~"" :::;><:~~~ ;~;-( 'J,; ,~:,"') '. ~ .;: ~' .~j~~,<,v; ~~. ~,"..: ' ,~ -:~:.'"
< _ ,"

.<_ :. ".~ <:

_

.:~"

;--.. '", '\ ~ _: :;_0" .',,,', ':::"~<~"~':-~::\{ .. ~: :~;~.:,~':; .;:.~" ~:{-f

>

_,~

"\

~.

++

~ . . . ."';>.o

.r:.,.t."

ANALYSIS D2.9 Welding of Aluminum Alloys.

AND

DESIGN OF

FLIGHT

VEHICLE STRUCTURES

D2.5

The
r~farred

~e~t-treatable ~lloys,

commonly

as 201,:;', 2024, 70';'.'3, cannot be we Ideo tlit::-t the ~rJacetylene torch without de8troyi.~ ~heir ~e~ha~l:al properti3s, whl:h ~re not restored Bien i~ heat-treated after weliing. These all~ys are g~nerally cl~ssed as unweldable. 20nstant research is ~oine on to develop aluminum alloys that have relatively hi~h s~cength Nhich cqn be welded without appreciable dec~ease af the strength proparties. A recent ceva l cpment ":ly the Al.umt num Company ot America is a new alloy deat gna ted X7005, wni cn develops h1 0 h strength after weldi~g.
~s st~en~th ~lloys, suc~

to

the

This 1s conservative since weld does not extend across the ~late, th~s any ~ecrea3e of t ens t Ie strength or-oper-t i es sncuj.d be less than that assumed al0ve.
P!~"OBL2M

2.

______

~/;:'--f1 L

1'! @,,[.ii!'Pt -- ~::..::."<@

lJ L
1

Ii

!ljI''iIIMfT ONLY)

The strain-har1ened allOYS, namely, 1100 ani 3003, are readily joined by gas wel1ing. E::l tner aa oxyacetylene, or an oxyhydrogen r Iame is used and shee~ thic~~esses as low as .020 a~e succes3fully we1rted. It is common ~ractice to use these ~te~ials in welded r~el or oil t~nk for aircraft.
D2. 10 nIustrative Problems Involving Welding.

"2JOo'(ifTJ~"'OD)

: ,

WELD qNE S/Df QNLY

TUM THJcx}Y,s~ .049

FIg. D2.20

Fig. 02.19 shows two plates welded together to ~orm a lap joi~t. The materfal 15 alloy steel Ttu = 95000 psi. Find the ~argln o~ safety of the weljed seams under the load of 5200 lbs. acting as shown.
PIWBL,;:j1 1.

Fi3. 02.20 shows a gusset pl~te inserted between the end~ of two tubes of a truss. The gusset is used as a fitting to taxa th~ pUll from a 3/16 di~netar steel tie rod. Determine the margin of safety of can:.ection of gusset to tubes. All material steel, Ftu = 95000. Resolving the wire pull into components parallel to the tubes, we obtai~ P = 2100 x sin ~SO = 1480 x 1.2 (fitting factor) = 1780 lb. The allowa~le weld load is governed by thinnest materi~l or .049 of the vertical tube.
P3.

\
~

-r,

<7"0'

(
;
"-.O7l.

L
" 1 -f'I

= Fs u Lt

= 43000 x 1.125 x .049 = 2370 lb.

~.s.
?~OBLE!1

= (2370/1780) - 1 = .33

3.

Fig. D2.19
rro~ eq~a:10n

D2.1,

Pa

= Fsu Lt
2x1

Fsu f~~m Table D2.2 = ~OOO L = total weld seam length t = .065 Pa = 43000 x 2 x .065 = 5580 M.S. = (5850/5000) - 1 = .11

= 2 in.

In general welded fittings involVing Jlates and tubes present conditions for whlc~ it 1s diffisult to dete~ine the actu31 stress flow through the joint, thus the general procedure i5 to ~ake conservative assum~tlons regarding the stress flow distribution and chec~ the ~lttlng units for these conservative assum)tions. ~he following exam~19 ll1~strates this a)prOximate pr-ocedur-e 0 ~ strength checking a ve Ided fitting j otnt ,

In

Fi~.

D2.21 the fitting plate

whic~

1s

Tensile strensth o~ .065 plate ~sin~ a reduced allowa2le stress due to welcing c~ 30000 as ?er Tabl~ 02.1, gives
Pa

weIced to t ae three s t.ee I t ubes ie suo jec t ed to

a tension load of 14000 lb. as s~c~n. The fitting ~111 je investigatec :o~ ~os3iole Neakness.

= 1 x .065 x

SCOOO

= 5200

lb.

)2.6

WELDED CONNECTIONS.

This value will be decreased to 50000 because of welding effect on material properties. M.S. = (50000(49700) - 1 =: .01 This margin of safety is conservat1ve since shear out area is conservative. Tension Stress on Section Through Bolt Hole:
I'-O"~ ....... --MTU8';.'.lr.' fB'"

Area of Section

Tr~ough

Hole:

- TV&f-i4,"

r
!

fI):

1-,

')-./i,
,

l .'.

I - .

"

'

Main olate = (1.5 - .625) .1875 = .164 ~ashers = (1.25 - .675) .1875 = .118 Total net area = -:282 sq .m.

tt

= 140001.282 = 49700

psI.

Ftu from Table 02.1, allowing full correction for welding effect, equals 80000 pSi. M.S. = (50000/.9700) - 1 = .61
Fig. D2.21

Solution Shear
Strer~th
~

Tension stress on fitting plate at Section 1-1 (See FIg. D2.21). Net area = (2.5 - 1.25) .1875 .235 sq.ln.

of Clevis Pin:

Load on pin

14000 lb.

The entire load of 14000 will be assumed to pass this section, Which ~s no dOUbt conservative. ft
~

- 1 = 1.1

Double shear stre~~th of 1/2 di&~eter jU~ clevis bolt 2 x 14722 = 29444. M.S. =(294444/14000) Bearing or
~levis

= 14000(.235 = 59500, Ft

= 80000

psi. o.k.

Check of Connection Between Fitting Plate and Bolt on gUshing: It will be assumed that the horizontal component of the wire pull will be transferred to tube (A) by the weld between the tube and the fitting ~lace. This is a conservative assumption. Horizontal load component

_l4000~ : 59500 pSi Searing stress ib : .., x .3(5

Ultimate bearing stress: 175000 psi. Thus a large ~rgin of safety is a~ailable to take care of Near due to Slight rotatiQn or shock. Bearing StreSS Bushing on Lug:

= 14000 x

2/~=

12500 lb.

.. = .525 14000 ~b x .375 = 59500 psi

Allowable beari~g stresS FbI' Chapter B2). 140,000 pSi (See

The weld length between tube (A) and the !itting plate is 1.5 inches on the u?per t~be surface and 2 ;'ncheS on the lower surface. To be conser'latlve, a total weld length of 2 x 1.5 = 3 inches Will be assumed a~tir.g. The fitting plate is welded to the tube on both sldes and since t~lce .065, the tube thickness 15 less than the plate thiCkness, a total weld length based on tube strength is 2 x 3 = 5 inches. ?a

The result shows that bearing on lug is not at all critical. Shear Out Strength of Fitting Plate: Shear area main plate = (.75 - .3125) .1875 x 2 = .164 'ashers (.625 - .3125) .1875 x 2 .118 Total shear out area =: .882 sq. in. f s = 14000/0.282 = 49700 psi

= Fs u

Lt : 50000 x 5 x .065 = 19500 lb.

- 1

M.S.

= (19500/125001

= .56

rsu ~or steel when rtu = 95000 is 55000. See 8hapter B2.

It will be assumed that the vertical component of the wire pull will be taken into tube (B) by the weld along each side of the tube. Load

= 14000 x l/vS = 6250 lb.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

02 7

The weld length on one side of tube is 0.625 inches long and I inch on the other. A total weld length of only 2 x 0.625 = 1.25 inches will be assumed which is conservative.
Pa

(4)

Attachment of hinges, brackets and fittings to supporting structure. At jOints in trussed structures. At juncture points of stringers with ribs unless a stop rivet is used. At ends of stiffeners or stringers unless a stop rivet 1s used.
On each side of a joggle, or wherever there

(5)
(6)

50000 x 1.25 x 2 x .058

7250.

Thus even under the assumed conservative assumption, the weld attachment for transferrin! verti~al component to tube (B) is more than adequate.

(7)

(8)

spar w'ELDING
D2. 11 Spot Welding. After many years of research and testing, spot welding of aluminum alloys, magnesium alloys and corrosion resistant steels has become a reliable established practice of joining many parts or units of flight vehicle structures. The spot welding procesS is accomplished by clamping two or more sheets of metal between cop~er or copper alloy electrodes, under comparatively high pressure and causing an electric current of low voltags to flow beONeen the electrodes for a predetermined interval. The current creates an intense heat at potnt ~A" (See Fig. a) which melts the metal locally due to the resistance set up by the shBets. As soon as the metal is molten to the extent shown at "B", the predetermined tline of current tlow Is completed and the sheets are forged together by the pressure on the electrodes np". This pressure depends on the thickness of the sheet.
(9)

is a pOSSibility of tension load component unless stop rivets are used. In general most aluminum and aluminum alloy material combinations can be spot welded. Table D2.3 gives information on this subject.

Table 02. 3 Acceptable Material Combinations for Spot Welding

0
N N N

S e-

"' "' "'

0
~
~

MatertaJ.
5052. 5052. 6061. 3003. Clad 2024 1100 . 1100 . Clad 7075 c Bare 7075 b
Bare

~ ~

~ ~

0 0

N N

0 0

-e

<3

:; :;
~

-e

<3

1Il " " 1Il

N 0 N

0 N

0 N

a O'"l

-e

2024 b

Clad 2014 . Bare 20gb

a The various aluminum and aluminum-alloy materials referred to in this table may be spot welded in any combinations except the combinations indicated by the asterisk (") in the table. The combinations indicated by the asterisk (") may be spot welded only with the specific approval of the procuring or certificating agency. D2. i2 SpotWelding of Aluminum Alloys. In general. aluminum alloy spot welded joints should not be used in primarJ or critical st~uctures without the specific approval of the military or civil aeronautic authorities. The follOWing are a few types of structural connections in aerospace structures where spot welding should not be used. (l) (2) (3) Attachment of tlanges to shear webs in stiffened cellular construction in wings. Attachment or shear web t'langes to wing sheet covering. Attachcent of wing ribs to beam shear webs.
D2.13

b This table applies to construction of land- and carrierbased aircraft only. The welding of bare, l1igh-strength alloys in construction of seaplanes and amphibians is prohibited unless specificallyauthorized by the procuring or certificating agency. c Clad heat-treated and aged 7075 material in thicknesses ieee than 0.020 inch shall not be welded without specific approval of the procuring or certificating agency.

Spot Strengths.

Design shear strength allowables for spot welds in aluminum alloys are ~iven in Table

D2 8

WELDED CONNECTIONS

D2.4, for magnesium alloys in Table D2.5, and for steels in Tat:le 02.5. The ~lnlmum edge -:11stances from s;:ot weIds 1s also given in the tables.
F1g. D2.22g1 ves the maxl~um statl~ strength of spot weIced jOints having the same pitch in all rows in aluminum alloys together

Table D2.6 Spot-Weld Maximum Design Shear Strengths for Uncoated Steelsa and Nlckel Alloys (Ref. 1)

I Nominal Thickness
of Thinner sheet, in.

Material Intimat!! Tensile Strength. lb.

150 kai and above

90 Its! to 150 sst

57 85

Below 90 kat

with the maximum plt~hes with which these values can be obtained. For joints having larger pitches, ase Table D2.4.
Fig. 02.23 gl ves the tensile strength single spot weIds in 7075-T6 clad ma ter-ta.l .
D2. 14 Reduction of Tensile Strength of Parent Metal Due to Spot Welding.

0.006 ..........
O. 008 ......... 0.010 ........ 0.012 ..........

Spot welding decreases tne ultimate tensile strength of the sheet material being spot welded. ?1g. 02.24 gives the efficiency in tension for spot welding of alumlw~ alloy sheets.
Table D2.4 (Ref. 1) Shear strengths and Minimum Edge Distances for Bare and Clad Aluminum Alloys Materials & Tensile Strength, ks! Nominal I Minimum Thickness 19.5 and Edge of Thinner Above 56 28 to 56 20 to 27. 5 below Distance, Sheet, inches Inches Shear strength of Sheet, Pound. 0.012 0.016 0.020 0.025 0.032 0.040 0.051 0.064 0.072 0.081 0.091 0.102 0.114 0.125 0.156

0.014 0.016 0.018 0.020 0.025 0.030 0.032 0.040 0.042 0.050 0.056 0.080 0.063 0.071 0.080 0.090 0.095 0.100 O. 112 0.125

........ ........ ......... ....... .... ........ ...... ..... ......... ....... ....... ............ .... ..... .......... ....... ......... ........... ...... ......

70 120 165 220 270 320 390 425 580 750 835
I, 168 1,275 1,700 2,039 2,265 2,479 3,012 3,540 4,100 4,336 4,575 5,088 5,665

.....

127 155 198 235 270 310 425 565 623 850 920
1,205 1, 358 1,558 I, 685 2,024 2,405 2,810 3,012 3,200 3, 633 4,052

70 92 120 142 170 198 225 320 403 452 650 712 955

I, 166 1,310 1,405 1,656 1,960 2,290 2,476 2,645 3,026 3,440

60 86
112

148 208 276 384 552 678 642

1020 1230 1465 1698 2400

52 76 106 140 186 248 354 500 589 691 810 960

,
I

1085 1300

2' 56 80 116 168 240 329 451 524 620 703 760 803 640

I
I
I

16 40 62 88 132 180 240 320 364 424 484 548 591 629

3/18
3/16

a Refers to plain carbon steels containing nat more than O. 20 percent carbon and to austenitic steels. The reduction in strength of spotwelds due to the cumulative effects of timetemperature-stress factors is not greater than the reduction in strength of the parent metal.

3/16 7/32 1/4 9/32 5/16

11/32

3/8
13/32

._- ecer -TS

_ . - 2024 -

101!5 - TS CLAD

CLAD

7/16
7/16

,
I
2

781

MIN.

7/16 9/16 5/8

Table D2. 5 (Ref. 1) Shear Strengths for Magnesium Alloys

I
I
!

Magnesium Alloy ~ominal Thlckness i QQ.NI-::l4 : QQ-M-44 of Thinner Sheet,: Shear Strength Inches of Sheet, Lbs. 0.020 0.025 0.032 0.040 0.051 0.064 0.072 0.081 0.091 O. 102 0.114 O. 125
, ,

I Minimum Edge
I
Distance, Inches

I I

, ~~-r -l
...

~:: ~ ....... ~ ~ (;
~

/ / ,, , ",' , ~ ,

~ // ~ ,

1<1}/

...- ;::::--

~J'/' L~ .......
I
liNCH)
I

;;

,
2

--

-~I

- --1 --."'-l.Tl!'L

8 8
Q Q

"

o.

,.
I

! I
~Q.'"

TH!Cl<NESS OF THINNE":'T

I I

...0'"

sHlO"FT
....

! I

, I , ,

, "
0

,~

-"00

'00

H'

I ,
!

51
71

102 137 198 242 279 320 368 433 488 544

69 97 129 185 251 328 378 434 498

586

3/16 7/32

''''
112

~--

,
"' .

1/4
, ,

9/32 5/16 11/32

-,

,,.

, ,

-.
-

v. ~s
~

Iii/Gte

112

3/8
13/32 7/15 7/16 7/16 9/16

, Note I~I:m )OIll~ ~.:..,,-: :::_,:;,ltI 'IlIQII 0l:l0Ve

3/42
~
x

<

658 735

Fig. 02.22 (Ref. 1) Maximum Static Strength of Spot Welded Joints in Aluminum Alloys and Corresponding Maximum Spot Weld Pitch

~~~LYSIS

A'S"D DESIG:-i
---,

OF

FLIGHT

VEHICLE
Sa~9

STRUCTURES

D2 9

'5

V.'.TERIAL

-..,....

--

"'07'5- -:'"6 CLAO ~

as ?rQal~m _ bu: t~b~ anG fl::1~g is heat-treated a~ter Nelding to ?tu =

1.50,000.

" ~
~
i
l e

i :
/

,I

/
I

zj-r, ----..

I 0~-------......,-r.,--+-,,"'--~,
----------,--+__1-+.,4----

1\
.;

.1

-t-:'~ '~'~'~'~ '~'~' ~'~' ~-..~

- ------ ~.-;~l
c
-It'L-----~

3
~

Y i : I Ii V I I I c IV o ~_+_l,4--'--'-!_+-+-+--__1 , i /f i I I I : I I I
" <
I

;::5~-......,-__,_---r-r_17f-+-__t-t_--_I

, Vi

i/ i !
I

'-

>

71 -

020 .032 ,040 050 063

Q71

0800')0 ,1000

'25

~ .. .... :'/
../
Fig. A

!"~

f'l, 'UASJliFli5
7l

,-

SHEET THICKNESS (INCH)

Fig. D2.23 (Ref. 1) Static Strength of Typical Single Spot Welds In Tension Using Star Coupons

(3)

A -.051, 202~-T3 aluw.ln~ sheet carries an ultimate tension load of ~OO los. per inch. It is spliGed by a lap joint involving one row of spot welds spaced at 0.5 inch. Is the snot weld stren~th satisf~ctQry.
alumin~ shaet ~arries an tensile stress of 30000 pSi. The sheet 1s to be solieed. Design a soot welded joint :or a lao joint.
~ltimate

(4)

A 7075-T6

(5)

Fig. 02.24 (Ref. 1) Efficiency in Tension for Spot Welding Aluminum Alloys

In a wing sect ton involving skin and stringers, the shear [low in ~djacent panels to a ,~rtlcular str~nger 1s 400 and 600 Ibs. per inch, and ~c:ing :G the sa~e direction. Assuming no di~[~nal :e~sion action due to skin ~rlnkling, Nnat s~ot spacing t s required to fasten the s tr-t ncerto the skin i: ths skin is .04 thick ~~G 2CZ4-TS aluminum ~llcy ~~te;l~l.

?ROBLZI'S (2..;

ng.
~.lt

(.1)

ill'J.st::ates a welded pIa t e

to a
ro~~d

r i t.t ins
30th

(3..)

:1ilitary Hancbook - XIL-HJSK-5,

1962.

_~.ll~ us t ,

fitting plate ana tu~e = 95000. ;nat 1s the T~xi~An desl~n 10a1 ? Nhlch the tlttir.g C~~ je 3~bjected to 1~ a ~itC1~g factor ~f 1.2 is used. ,1ttin~ 1s not sUbject~d to vibratlJn or ~0tation 0.' a i nge -nn .

~astened

steel tube. a,e ste~l ftu

~~~

-~.--:'"L--":'

... ---"~.

_ ~ -

~...

...-'-.... ...... .".'

~~

~~-.~~

",

....

~-~_.~---

:
--

';',''''

~~'~

--.- ---

.~{,

-----

CHAPTER D3

SOME IMPORTANT DETAILS IN STRUCTURAL DESIGN

BY

WILLIAM F. McCOMBS IDESIGN SPECIAIlST CHANCE VOUGHT CORP.)

D3. 1 Introduction.

In th~ design and fabrication of an airplane the major caIDPcnents receive a thorough ~evl=w and evaluation. ~~ny of the smaller ~arts, however, are designed at the last minute
and, not recelvl~g so ~uch attention, somet~es have faulty details. It Is these which

"b"

frequently lead to trouble in service and in tests. This cr~pter represents an attempt to ?olnt Jut Some of the more cammon details that seem, somehofl, to be overlooked from time to ti~e. ~hls should be of help to those involved in designing or dealing in other ways wIth the str~ctural co~ponents of airplanes or ot sl~11ar types of structures. With regard to specifiC details, many aircraft companies have standard methods of design. The reader should al~ays consult his company's data on these, if available. In the event such are not available, the following suggested practices should be of practical help.
D3. 2 Shear Clips.

Fig. D3.2

is cantilevered oft of a hea'TY piece of structure, (b). The load, P, passes through the shear clip as a shear load from web (a) to (b). There is also the oending moment, P x 1, to be transferred. Additional splice plates, "8", are provided for this purpose. They transfer the moment in the form ot axial loads trom the flanges of beam (a) to ~ember
(b)

Shear Clips are usually seen in two

roras ~
(a)
bent up sheet metal or extruded angles (the angle oeing anywhere from 00 to 180 0 between the legs). extruded "tees".

There are hundreds of these ~n a typical airplane. They are used ~n joining together both pri~ry structural components and secondary structural parts such as equipment mounting brackets, etc. The function of the shear Clip is to transfer a shear load from one part to another. It is not intended to transfer axial load or bending moment or twist, only stear.
m~lltary

(h)

These Shear clips are shown in their "mintmum acceptable" tonn in Fig. D3.3. The

A typical example ~s shown in ~ig. D3.1. Here bracket, or beam, (a) is supported by beams (b) and (0). The load P is thus "beamed out" to (b) and (c), passing as a shear load tr~ough the clips into the webs of (b) and (c) '3.S illustrated.

Fig. 00.3

Fig. D3. 1

When a Significant axial load or bending moment ~ust also be transferred, additional m~bers must be provided or the shear clip must oe replaced oy a heavier fitting. This is illustrated in Fig. D3.2. Here a be~, (a),

minimum requirement is that eac~ leg of an angle type clip must have at least 2 fasteners through it. The "top" of the tee t}~e clip and its leg must also have at least 2 fasteners each, as shown ~n (e) and (f). The load "balance" of an angle type shear clip is illustrated in Fig. D3.4. The corner edge of the clip should be assumed to carry

D3.1

03.2

SOME

IMPORTANT DETAILS

r~

STRUCTURAL

DESIGN

only the shear being tranSferred, taken as :000 lb. in this :ig~re. The net loads on ~he fasteners are then as illustrated. Once the loads are known, the clip and fasteners can be checked for strength USing standard ~ethods.

thicker part of the member being attached. If the joggle is a Significant one, say to the

C-

Final

-lt~i'":::J~' l.T'J'~ ~ Sal:} ~500' "':'OOOX~

Q

'r
, , , ,
I
~-

-,
,
',/;
~

:~
--:.,

(-,

:",<,\

,~

", ," ," , "I< ::"" "

'"
~

,<

~ .:._ .... ~ ,

\
~.'-

, ,,

,
0
(hI
,

,~

c '

(al

Exampl.,

"

'I

"

/'
(

"

'-i

'~

.-

'" '"
iH

cr'

"

"

Obtained by Adding Above Together

O~$,

Let a"'. 40", b=l.O"

Fig. 03.6

500

Then Q; 1000X-:"""":'ll,.4 =400""

1. U

Resultant Rivet Load:

R ",.j500 a .. 'l00 a ,. 640*

Fig. 03.4

Use 2-5/32" Alum. Rivets Then clip thickness required is t 032". 75ST Ale.

=.

Fig. 03.5 illustrates '.. .hat will happen if only one fastener is provided in a leg of an angle clip. The single fastener cannot (ignoring friction) balance any shear at the corner. In other words, it can receive only a shear from the Neb to which it is fastened. ThiS, in turn, puts a ~Nist, P x a, into the other leg of the cliP and, hence, into the other web. This is unacceptable, of course, sinCe a much thicker leg would be needed to carry the tOLSion, and an undue twist would be present in the other web being joined. Of course, several fasteners, rather than just two, may be used when space allows. Clips of type (a), (c) and (8) in D3.3 are more efficient than are types (d) and (f). The latter are used When all that space limitations will allow. cases the dimension nan should be kept as practical installation will allow. Fig. (b), this ~s In all as small

order of the clip's thickness or greater, it can considerably reduce the clIP's rigidity and cause it to function as 2 none rivet clips" with the adverse twist effects mentioned preViously. In this case at least 2 fasteners should be prOVided on one side of the joggle in the joggled leg, as illustrated in Fig. D3.6{b), to maintain rigidity and proper functioning. The load should be aSSumed to be carried by the 2 fasteners above the joggle, similar to case (b) or (d) in Fig. D3.3. Joggles are discussed further in Art. 03.4.
03.3 Tension Clips

These are also quite numerous in ~il1tary airplanes, being used to splice relatively light tension loads from one member to another. The tension clip is a very inefficient tj~e of splice. It has a relatively poor fatigue life J particularly, and should be used only when the load is ~ll and other design factors prevent the Use of the more efficient lap shear splice. It is usually -resorted to when so~e structural member such as a bulkhead Neb ar flange or fitting cannot be efficiently nopened upn to let an axially loaded member pass through. It is also ~requently used to attach cantilevered brackets to bulkheads Or ribs or other structure. Consider Fig. D3.7. ~ember (a) 1S, say, on one side of a bul~~ead and is to be spliced to member (b) on the other side. ~he~e 1s an axial tension load to be trans~erred and since the bulkhead cannot bs cut, a tension ~llP arran~ement must be used as shO'Nn. Angle cli;S in this case are 111~strated.

For loads on longer leg in figure, let a '" . 4'",

b "" 1. 0".

1000*

An "unacceptable"
Type Shear Clip

lTwist=1000il'x .4'"=400". This is required to balance the 1000* which is out of the plane of the longer leg. This is unacceptable.
Fig. D3. 5

"b"

Another type of deficiency someti~es ariSeS wnen a minimum type shear eli? IS being used. This IS 111ust~ated in Fl~. D3.6 where it has been necessary to njoggle n one leg of the ~~gle :lip, say to fit over some locally

P..___

~
\,

.. ,

: ", ..... ". , .

~.'" -----'P ,.
""

a.

ens ion Clips

Fig. 03.7

- v"

.;,:;".;-,-;l,-'_""<~~ ~":~_ ,:.......~.,_":.~

. -

,".y,~:.'~;
.

""_'.'''':-',
.

' . r~~:,'~'.;,,;~ ...

;

-.'; _~_:':"-:;._,~:e::'t.i'~~;""'?":"""'&""""'~'"
_.,,_.

;'v;-';:'f<,:?"
-'. -'f;

----.

---..-.;-''''''

ANALYSIS
Si~p18

AND
~sec

DESIGN OF
~0~

FLIGHT

VEHICLE

STRUCTURES
p

D3.3

tension

~~i~s ~c~

~e~~~i,3Iy ~i~~t ':.li~s

l:ads, lsr~= ::sts re~~lri:1g a zac ai ned r t t t t ng of a "ba tr.tub" t ype . 'l'hese

ar-e usue l Iy 3eer::. 1:1 .3 ror-as .

single
~etal a~gles

- either (back to
extr~ded

j~nt

up sheet

ar

ext~usicns.

,. ,
\ Q.,

double

an~:es

~ack)

(0 )
~ype

clips :ut from (c) is the

t~ickness.

tee sections stif!est

~or

stron~est ~nd

(01

~iven

To obta1n maximum s~rength and stiffness, Jolts should be uSed fer attach~ent purposes. Allowable load data is given in Fig. D3.8 for the Single ang.Ie clip arrangement illustrated. T~e ~ethod of cbtaining the allowable load is alsG illustrated by the dashed lines in the
:~~'.lre.

Fig. 03.9
~oments

more on each other. F~cm statics, taking abDut the center of ~ressure on the a+b taes, .~ = p x (--.--), Thus ~ ~s gr3ater than Obviously a small ~nough jolt will yield or ~ail in tension before a thick clip will yield or fail in bending near the waSr~r (~li~ = ? x e). There is also a ~rYi~~ ac~lon in the tee type clip, as illustrated.

P.

YIELO LOAO FOR

gaLT
SPACl~G

~'1NGLE

A."GLES

BOLT HEAD CLEARANCE {NCHES

.30
'~' 25

JO

."
100
THIClOlESS OF ANGLE INCHES YIELD LOAD PER BOLT' Las.
1000

~his prying action 1s the reason why the designer should be cautious in using rivets even ~or light te~sion loadS, as is sometl~es done. When rivets are used, as in ~ounting equipment brackets, it 15 best to use steel types and car9:ully check the prying load maintaining an a~ple ~argin of satety. In any event, riveted clips are inferior and no design data for them is given here.

:-fOTES;
I. In these te$tll tile anglu procrndea

aOLT HEAD

C:'.EARASCE -----, 2. critic3.L _ ...~." BOLT... 3 V:lJues 3.re lor l"OOm tempenture ....:;960 /.J-2024-T ANGLE' use only.

t .H6..;:::m,
I
I

LOAD

.u e,ther end beyond Ute It of the

bolt <l d\st;Lnce of 1/2 the bolt spacLng. For thi~l< angles the bolt may be

Another Doint in ~Sing tension Clips is frequently overlooked. The structure to which the cliP is attached must be capable of taking the loads appll~d to 1t. These loads Gonstst of the tension load frOm the bolt and the load :ram the toe action. Several examples are sho'Nn 1n Fig. D3.10. In these examples the te~ nunac~eptablen ~eans that the allowable loads of Fig. 03.8 are not applicable.

'~~~~
Fig, D3.8
(Ret, Vought Structures Manual)

It 1s noted ~t~t ~J~ lar~3r Clip :~ick ~ssses the belt ~ay b2~c~e c~itical and Jegin :0 yield (the sa~e would ap?ly ~o smaller bolts t~an ~hose spec~f~ed ~D~ :he t~icKr.esses S~cNn). T~ls is bec~use of the ~rYing action 0:1 t~e bolt. Beca~se J: t~e ~rY1n~ ~ctL8n, the load in the bolt is always greater ::zn the a;~lled load. Co~s1der Fi~. D3.9(a) illustrating an ar.gle :j~e eli? a~ange~ent. As "cne app.Ltec Ioans , P, c es t n to "open-up" t~e :11;S t~e ~olt feels an increasing tension loa:.,~, and the "toes" ot: t.r;e ell; bear:o"lffi

~~l
101'
Heavy BaCK-Up Structure Acceptable

!P

(b]
Light Back-Up Structure
Unacc epr-

1
j
Back to Back Clips Accentable

~J'~1 f ic;Jr
Eccentr-ic Clip Load Path
Ijnacc eptable
CrissCross Clips Unacceptable

t"

tP

able

Fig. D3.10

:~ses (b), (c) and te) reQu1re a rearTa~ge~ent

of, 8r

~jd1tional st~ucture

in, the baCK-Up

03.4

SOME

IMPORTANT

DETAILS

IN

STRUCTURAL DESIGN

structure whict is ~ecelving the load from the ~liPS in order to achieve the full allowables ~f ?lg. D3.8. The resultant load on the back~p str~:t~e is, of course, the applied vector, P. aside from having a low and stiffness, also exhibit a r~latively poor fatigu8 life. If the load is of such a nature as to occur many times, say due to symmetrical flight conditions such as p~ll-ups and gusts, there should be a large =arg1n of sa~ety at Itnlt load levels in the neighborhood of 200% or more. If the load 1s ~~e only to some non-recurring type of loading, such as a crash condition or nj~edn system load, ~he large ~rgin of safety would not be :1ecessary.
~11PS,
st~t1c strer~th

Tension

Other suggested practices involVing tension

c Lt ps are:

Some airc~ft camoanies have specific strength data and prac~1ces !or the design of joggled members. This should, of course, be consulted by the designer, if available. Some companies use a 6:1 joggle length to depth ratio, others use a 3:1 ratio, or both ~4Y be used. Strength or sti:~ness data for one ratio should not be used blindly fo~ another. Some of this data indicates that When, in the case of angle members, the depth of the joggle is to the order of the thickness of the joggled leg or more, the loss in strength is about equ1valen~ to the loss ot the leg outboard or the bend radius. A "rule-ot-thumb ll deSign practice 1s therefore suggested as follows. Assume that the net etfect at the joggle, tram a strength and stiffness standpoint, is equivalent to a slot cut 1nto the joggled leg that extends iITMard to the bend radius tangent point. This 15 illustrated 1n Fig. 03.12.

1.

Keep the bolt head as close to the bend radius or fillet radius as is ;,ossibie. Avoid their usage, if possible, When repeated loadings are dominant.
View A-A

2. 3.

If part of the structure is continuous ac~ass a jOint and part is interrupted, do not use tension clips to jOin the 1nte"ru~ted structure - instead a heavier, stiffer, ffiachined f~ttlr~ 1s r equt red.

Joggled Member

Efiective Section Through Joggle

Fig. D3.l2

D3.4 Joggled Members.

A "joggfe" is an offset formed in a member. It usually Invo lvee one 'or more flanges of a ~ember of the nopenn cross-sect10n type. JJggles are quite common in typical metal ~~:r~lane structures. They are used most often when it 1s desired to fasten together two ~~:ersecting ~embers without using an extra ?ar: at the joint. The Jogg~e 1s a compromise. It 3aves an extra part but the price paid 1s a IJSS ~n strength and stiffness or the joggled ~e:nte". I~ the load in the ~ember at the ~olnt ~s small eno~~h, the saving 1s justif1ed. If ~~t. an extra part, ir~tead of or in addition :J the joggle, must be used. A typical joggled Insta Llat ton is shown 1.:: Fig. 03.11 wher-e one leg Jf a 4 angle ~ember has ~een joggled over and fastened to another ~ember lying in the same pjane .

With this assumption, the flat portion at the joggled leg will carry no axial load across the joggled area but will prOvide support for the curved element. The effective net section, Fig. 03.12(c), can then be checked USing standard methods of analysis for whatever forces are acting on it. It is obvious that the net section shown will have little strength for carrying bending moment normal to the re~ining leg. ThUS, care should be taken to insure that any axial loadS are introduced as near the corner as possible - ~hich in turn ~eans that at ~east two fasteners should be used on eac~side of~e joggle.
~he above approach, cor-sidering the ~cggle eqUivalent to a slot, will ~i7e t~e ~esigner a ~uch better nfeel~ of what he 1s really doing when he s;ecifies a joggled member. T~e basic reason ~or the loss of stren;th and stlf:ness can be seen 1n ?lg. D3.13.

Angle

2
II

Skin or Floor is
~ U,uajly

3 3 3
Fig. D3. 11

Presen)

:--- 0~~: J ~l~~::led

I

__ -Ii
~

Req'd. Balancing Loads at Breaks

c:

=== =::...
'

""

Fig. 03. 13

ANALYSIS AND DESIGN OF

FLIGHT VEHICLE STRUCTURES

D3. 5

The axial loads in t~e joggled leg being inclined to each other require a balancing load. Such a balancing load is not available except ~3 shear in the thin leg and ttls resultS in the less of stiffness and strength. If the symmetrical leg of a tee member were joggled there would be more stiffness than in the case of an angle, but the sa~e approach, though more conservative, is recommended in the absence of test data.
AS an example of the foregoing discussion assume that an angle member is supporting another member locally which is loaded by the forces Q as shovm in Fig. D3.14. If a skin is present, as sho,vn, part of the load can be carried across the joggle by the gusset effect of the skin. This can be approximated by USing the ~ethods of calculating inter-rivet buckling of skins discussed in another chapter. The rest of the load must be carried across by the net effective section of the angle in the joggled area.

section to carry, then an additional memD~r should be provided locally. Two ways of doing this are sho.vn in Fig. D3.15 Sometimes local requirements are such as to necessitate both legs of an angle member being joggled. In such cases it should be assumed that the angle has no significant load carrying ability at the joggle. ThUS, the eXistence Of any Significant load at the joint would require an additional member and the angle should be ended just short of the jOint rather tran joggled up onto it.
Add Member

Cut from Tee Extrusion

Fig. D3.15

N.~\~:lo_~ 1 j 1~
_,

\kin

Vtew AA "b"

Fig. D3.14

Thus if the tctal load at the joint were 2Q and the load carrying ability of the skin were R, then the net section or the angle would be Subjected to a load P = 2Q - R and a bending ~oment M = 2~ x a - R x b. The stress at the lower curved edge would be the sum of the compressive stresses,
Ie =
T"!C

The suggested effective net sections of members haVing other types of cross-sections are shown in Fig. 03.16 where the legs indicated by dotted lines are joggled. In general it the joggle is slight, conSiderably less than the thickness of the joggled leg, its effect can be ignored, but ?roper fasteners should still be provided as discussed. The smaller the length to depth ratio used for joggling, the greater the ef:ect of the joggle. Joggled members lose stiffness and strength when subjected to tension loads as well as when under compreSSion (but any skin present is, of course, ~uch ~ore eftective as a gusset than when in compreSSion).

.E
Net Section

~et Section
Fig. D3.16

~or ultimate strength, f c could be carried up to Fey, conservatively, in the typical case.

D3. 5 Fillers.

In order to realize the ~imum s~rength ana stiffness, the load in the net section must be app.lLed in the "corner". T~is:5 to pr-event stresses due to oending out of the plane of the remainir~ leg. This requires that a minimum of 2 fasteners be prOVided to receive the load at the jOint. The reasoning here is the same as 1iscussed in Art. D3.2 concerning minimum type shear clips, and the fastener loads can be calculated in ~he same manner as discussed there. If the load is too large for the net

As the name implies, tillers are used to fill up a 70id. It is ~hen they become a ?art of the structural load path that they ~eed particular attention. Fillers also represent an item that is ~uite common in typical large or complicated metal airplane structures. AS an examole con~ider rig. ~3.l7. Here two tees, "a" and "b" carrying an axial lead, P, are seen to be spliced together by a pair of an~les "c". Since the lower leg of "0" is thicker than that of "a", a filler is needed. This ~111er is part of the str~ctural load

03. 6

SOME IMPORTANT

DETAILS

IN STRUCTURAL DESIGN

Case I.

I'
. 06"

"a"

P :;; 3000 Ibs.

1.
Extended Filler

Filler Fig. D3.17

Load

View A-A

Rivets required to "art to comb Ina t i cn No. Rivets

=:
",':::;;.1

=;li~e 0:' "~1!

3000 lbs. from crus filler:

31~~~ =

1.65, c r- 2

2.

Load carriec by
"b ") :
D

~iller

(to be spliced to

path, rr-cm "c" to "'3.". In tni s case , to realize ~ull strength ~f th~ tasteners, the ~iller mus t ce "extended" and additional :asteners provided ~o tie the ex~ended portion rntc :TI.e:I:~&r "'3.". If this is not done, it is ee ic t-c be a "rtoe t ing" :'lller. As explained later, a floating filler, i~ thick enough, will ~ause a loss in fastener strength. In the above example let the total load ? be aooo Ibs. ~~1 '3.ss~~e ~~~t 2000 lbs. of tnrs must; be t.rans rer-r-ed n-ora "c" to "a" by :he t~o fasteners in t~e :iller area. Th~ ~art of the load put tnte ~he filler ~y :earing pressure can je taken as . Filler "

Filler
1500 los.

3.

Fasteners required to transfer P to "b" F No. Fasteners

= 1630

1500 _ - .92 or

l
+

Total Fasteners required = 2

=:

3.

Splice is adequate since 3 rivets are present. Had no filler been present, 2 fasteners would have sufficed. Case II. P = 5000 Its.

= 2000 x .04 .04 = .... C6

8GO lb.
1.

Repeating the same steps as 1.n Case I: No. Rivets required

=:

Sut:icient :asteners should be ~ut in the extended part of tta tiller to transfer this 500 Ibs , into member "a ". a thick ~111er is inserted being spliced together in shear, tr.e ~iller should be assumed to be a part o~ on~ o~ the~. The ?ar~ of the total splice load it Nill carry can be ~alculated as illustrated ~bove. ~he fasteners :an then be ~:~sldered ~s bei~g in ~NO sets. One set ~uSt ~ve :~e strength to splice t~e total shear load frcm the stngle member to the combination !f:'1l1er plus memcer-". 1'1".e ot her set o-: ~asteners zust r~ve the strength (zeaning in 5~~ar a~G ir. :ear~~g) to trenafer t::8 calculated ~:ller load t~ the ~e:nber it is ass~~ed to be ~ ~~rt o~. ~n~~her exa~~le uses ?1g. D3.1a.
T~~s, ~henever
:et~een ~NO ~embers

-000 ~620

= 2.75 or

2.

-0-0 .072 = Load in ?iner = o U x .072 + .072

2500 Its.

3.

2500 No. Rivets required = 1630

=:

1.54 or 2

Thus the 3 rivets are required to trans~er load, P, from 1!a~ ~nd 2 a:dit1onal rivets ~re needed to unload the fUler into :ne:nt:er "b". The filler s hould be extended over "b ' and 2 e.ddlt~onal fasteners added as Sh01ffi by the dotted lines ir. rig. 03.18.
a~pr0xl~te,

Gt 2, . a:
~

Hi Shear- (Steel,) Rivets ,.081

Strengths: Rivet Shear" 1820"

the above ~r0cec~re ~s it ~rovldes a quick Nay of evalua~in~ :je ef~ect or :he ~1.11er. ~r.en t.ae r i j je r ttnckneas is less tran about 15-:1; 0! the faster.er dia~eter, its ~resence ~an 'ce ignored..
~~~lttedly,
b~t ~he ef:ect of the filler is ~o reduce tne allowable strength of the fastener. The reason ~cr this can je seen :r~m Fig. D3.19 wr.ere t~e ~resence of the filler causes ~reater ~ry1.n~ loads ~nd iance :nore tension in the raat aner-, a l onz wt t n the shear load.

7075- T6 Slum. Alloy Sheet Mtl , (All Memnez-a}

8earmg In .072 ,. 1630Bearing In .081 ,. 1840.

Fig. D3. 18

Asaune filler to be cencer-, "'J".

t nt.egr-aL

Nith t::e jcwer

riny filler ~n the structural load ~ath course, be T~de :rc~ a ~aterial compat~ble in sti:fness with that of the
s~ould, o~

1"

ANALYSIS

ANO DESIGN OF

FLIGHT

VEHICLE STRUCTURES

D3.7

P,
prytng Force
~

"Nor-mal" Eccentricity and Prying Forces, with No Fillers.

g Z

Larger Prying Forces Due to Eccentricity of Floating Filler

Smaller Prying Forces with Structural (Extended) Filler Fig. 03.19

j -"
I
,

12" ----+01 j: -- - - - -l:1 =,!;a='1: :-Cut au,.....: 7" !, ....c~ - - - t3~ = = :::1 J'= .. - ...
1--12" 12"

P,

t5~ :~:r lr"o

....

jli

P,

Fig. D3.20

Before the cut-out was made t~e ~emters sho,vn by solid lines (flanges, stit~er.ers, webs) are present. The members "a" and "b" are added to frame the cut-out. as sho'Nil by the bro~en lines. There are 2 ways to determine the loads in the area framed around the cut-out. The first is to assume a shear !low equal and opposite to that present with no cut-out (qc in the figure above) and determine the corresponding balancing loads in the framed area. Adding this lOad system to the original one will give the firzl leads and. of course, q = 0 in the cut-out panel. rhe other method is to use standard procedures assu~ing the shear to be carried in reascnable pr~por~i0ns on each side of the cut-out. The tirst method will be illus~rated here. All shear flows are sho'Nn as they act on the edge members (on the flan~es and stiffeners) in this discussion.
It there were no cut-out there would be a constant shear flow. qQ, in all of the panels. as shown in Fig. D3.21a. Next a shear flow equal and opposite to that in the center panel of ~ig. (a) is applied to the center panel ot

str~ct~e around it. That is. one should not use a soft aluminum filler between high heattreated steel parts or a ~henollc or fiberglas filler bet~een aluminum parts. The need for fillers arises not only f~om design considerations but frequently from ~nufacturing problems. In these latter cases "ars-eercn" between parts sometl~es occurs in assembly. Tc prevent expenSive rS-Nork. stFJctural fillers must be used to make the spliced area adequate. In these cases detailed attention is necessary. In the occasional instances when floating ~111ers cannot be avoided, the fasteners should r~ve quoted allowables well in excess of the shear being transferred locally. if the filler is ot Significant thickness. It is common practlce also to use a bonding agent (glue) in addition to the fasteners in installing fillers.

03.6 Cut-outs in Webs or Skin Panels.

The aircraft structure is continually faced with requirements for opening up webs and oanels to ~rovide access or to let other ~embers such as con~rol rods. hyaraullc lines, electrical wire bundles. etc., pass through. ~he designer or liaison engineer should be ~ami:iar with some of the various methods of ?rovlding structurally sound cut-outs. outs. are: There are several ways o~ providing cutT~~ee ~ill ~e mentioned here. These (a)
(b)
F~~,ldin~ sUl~able ~ramlng ~emoers
a~ound

P,P,--

"oj ~ "0.
~

qo

It ~ III q;;:: II r lit ~ III ""-- ~! qo ,Ir

Before Cut-OUt (al

P,

- p ..

@qolll 7/8 go i 17116 go 11

I li2
the
C~t-out.

'!o \7/1oqGlo- i

III

~7/g-,g.g.

-qo

II
i

li2 qo

II
1

,7/l6~

Pr-cvt oIng a cicub Ler or "bent" wher-e

~r~lr~

as in (a) cannot be done.

Self-Balanc mg Internal Loads (Due to Application of Equal and Opposite qo Assumed in the Cut-OUt Panel) (bl

(oi

Prcvldlr~ star.aard round tlan~ec holes which have publiShed allowables as discussed in chapter on ~ea~ c.eslgn.

(a)

?r~~lng ~~t-OUtS

in

~ebs

P,

Fi nat Shear Flow Distribution. (al'" (b)

(01

As an example assume ~hat a beam web requires a cut-out as shewn in Fi~. C3.20.

Fig. 03.21

03.8

SOME IMPORTANT DETAILS IN

STRUCTURAL DESIGN

Flg. (b). Stnce th1.s represents a selfbalanc1.ng load system, no external reactions outside at t~e fr~~ing areas are required. This is an important concept and the reader should think about it. The loads in the tr~med areas due to qo 1.n (b) are next determined. To eliminate redundancies, it is usually ass~~ed that the s~~e shear flow exists in the ~anelS above and below the cut-out. It is also assumed that shear flows are the s~~e in the panels to the left and right of the cut-out. a) the shear flow in the panels above and beloN the center panel must statically balance the force due to qo, or Since ZFt :. 0, qo

These
~pon

~ill add or SUbtract, denend:ng their directions, to any loads oresant before the cut-out ~as made (as in the case of the be~ flanges). The axial loads due to the cut-out can be gotten from Fig. (b). Or the total axial loads in all of the members can be gotten from (c). These are illustrated in Fig. 03.22.

P=Pa-1Z"xteQO

P=Pl. +lZ"Xft %

!
i 9/16 tlo

gh6 qo

I 1_7/8 ~

Axial Load Distributi.on in Upper Flange from Fig. D3. ate

(a)

xr

1-7/8 ~

9/16 <to

x (5+3) q 7/8 qo
q

1-1~

\ 1-1/2'10
I6Qo

Paxial""lZ'X

P axtal:12'~1-l/2qofiqol Compression

j)

the shear flows in the panels to the left and right of the center panel ~ust also statically balance the ~orce due to qo' Since ZFy = 0, qoxl2:= qx(l2+12) q=1/2Qo

15

Axial Load Distri.bution in Framing Member Above Cut-OUt Obtained from D3.21c (Same Result Could be Gotten from Fig. D3.21b). (bl

9/1.'10

c)

the shear flows in the corner panels must also balance the force due to the shear flow in the (any) panel between them. ConSidering the panels in the right hand bay

1-1/2 <to~

jF,'_7/.'lo"al""S"x(1-7/8Qo-9/16Qo)
p

9/1.

'IoJ

'r-- axial=3"(21/16Qo) Compression

p

axl

""S(21/16 qo) Tension

'1-7/8Qo

ZF z = 0; 1/2
q d)

Go x 7

q x (5+3)

1 qo x7 7 =2 8 =16"0

Axtal Load Distribution in Stiffener Bordering Cut-OUt on Left Side, from Fig. D3.21c (Same Result Obtained from D3. 21b) Fig. D3.22

tr~ final shear flows are ~otten by adding the values in (a) and (b) together, algebraically. Note that:

Once the internal loads are known, the members can be checked far strength using standard methods of stress analysis. The cut-out could have been framed without extending the framing members into the bay on the right of the cut-out. This caSe is illustrated in Fig. D3.23. Had the 7 n deep ~ut-out Jeen required at the Jottom of bay, the franiTIg could have ~een dcne with only one member (as could the pr9cedir~ cases also) as illustrated in Fig. D3.24. This represents the ~inimum of adequate fr~ing for any cut-out. That is, :here must be a mini~um of one redistribution bay on one side of the cut-out and at least ~NO redistribution bays on the other side, anc there must be the traming members 1eflni~g the bays. These rra~lng ~embers will ~l,vays be loaded ax1.ally. Note trAt tn the prev1.ous examples in CaSe (b) the sum of the loads on all edge members (framing members) 1.5 zero. No external loads are needed ~or eqUilibrium. ~his 1s

1.

the shear flow in the center panel (the cut-out) is q :. qo - qo = 0, as it should be. the shear flows above and below and to the left and right of the cut-out add, giving a nunber greater than the original qo. the shear flows in the ~orner ~nels are smaller than the original value of qo'

2.

3.

This is the way the changes al~ays occur in the area f~-med about a cutout.
a)

Finally, and importantly, there are axial loads developed in all of the framing ~embers due to t~~ c~t-out.

ANALYSIS AND DESIGN OF

P'I-r

FLIGHT

VEHICLE STRUCTURES

D3.9

qo
P,

'It I .

I i~ 'la-.

'!Q...... _

ill 0;;,~

W _qo
I(a)

. 'lo_

qo_

III I
II

in Fig. D3.25.
-

qo

II'

il!r %

P,

~ .E!!5o

~,"/8

t
,

7/~

'!L.

11 1 ~ 7/1l..90.- 1

II

o.a.._ .!
%
0

P,-

. .. -........ - ............
~

NJ
OUt :

. . ....
_Doub1"

rqo

laj

Ibl

qo

J(~

1T~ ~ ~ 1"77~ II.

~i
1'780:

I f1 0'i7 8!la- I
Fig. D3.23

II

1 t%
Ibl
(Shear Resist Web)

(e) Final Distribution

101

Pa

P,

% jr t-

'10

11

qQ......

f 7/8 qo 11 7/8 '10 t qo_ n '!8-

- jI U% IIqO - qo

Fig. D3.25

'10

lal

1/8 %

- 1 j ~ 1~ i t J

As shown in (a) the doubler whose thickness is yet to be determined is made to fit around the cut-out as shown. Reasor~ble internal radii are in the cut web and doubler at the corners to keep stresses due to curved beam bending reasonable (see Chapter Cll). Attachments are prOVided as shown to piCk up the basic shear flow in the web. The loading imposed on the doubler is shown in (b), namely the shear flow qo. Strictly speaking, the doubler should be analyzed as a f~-me. riith reasonable symmetry the loading in (c) can be assumed at the center at the frame. That is, one ralt of the total Shear, qo x h/Z is resisted in the top of the frame, one half in the bottom and a pin joint (no bending moment) exists at the cut. The bending moment axial loads and shears at any section of the frame follow as a ~tter or statics. For example, At A-A,

Ibl

'Io

'10

f l.3o. 1

II r

(e) Final Dtstrtbuttcn

Fig. D3.24

always the case when a set of self-balancing shear flows are applied to a flat panel str~cture or to a 3 dimensional box structure wit~ a cut-out on any side. The reader should st~dy the examples closely. Although the ~ethod is shown only for a ~lat be~ it is also applicable to any structure nith a cut-out, such as the box beam of Article A21.3. This has actually been illustrated in Solution No.2 of ttat article and the reader stould review it at ~his time.
Someti~es trs~i~g ~8mbers for a Gut-out are not cc~venlently available as Nere the stiffeners and flanges of the beam used in :he previous examples. In such cases they must, of course, be provided.

h W X=V qo 4 (there may also be a little relieving moment due to qo)

Fz = V =

xi.

~ 2

(0)
~ust

2raming Cut-Outs

~ith

Doublers or Bents

Frequently a cut-out in t~e web or a beam be so jeen that it removes nearly all of the Neb. In this case the ~ethod previously described. cannot be used. Insueac the "bruteforce~ approach is necessarJ and a heavy 1o~bler, or bent, is provided around ~he cut-out 'to carry ~~e shear. This is illustrated

The thickness of doubler required to t~~e the loads can thus be determined using standard methods of stress analYSiS. The doubler should have sufficient out of plane s~it!ness, alSO, to provide stmple suppor~ tor the be~~ web, as discussed in Chapter CIO. This will normally be prOVided by the thickness required tor strength purposes. Sometimes the nature of ~he ~ut-out is such that the :rame (doubler) can be deeper at the top (or bottom) . In such a case, the

D3.1O

SOME IMPORTANT
~art c~n j~

DETAILS

IN

STRUCTURAL

DESIGN

Qeeper

aSSI~ed to carry a greater of the total shear, V, ane the lower part a smal Ler- pcr-t i on . Th1s ts illustrated in ?i~. D3.26 where ~he cut-out cGes not extend ~3 tigh as in Fi 6 , D3.Z5 and the uD~er )art is ass~~ed to carry 4/5 of the total shear and the lower part only 1/5. ~ortion

iRing ~ .. ,. 'Double ', \ -

Basic Beam

The cases illustrated are for shear resistant webs. If a tension field is present the Goubler ~ust also be designed for the endbay effects disc~ssed in Chapter Cli.

Basic Ring Beam ~ ~ DOUbler

" "

'10

i ~
.
qo

~ r ~Y~.
V=4/5 'loh
~

'10

'10

t
View A-A Fig, 03.27

qo

I-

'10

Cross Section of Ring Doubler hal> same or Greater OUt-of- Plane Stiffness (t.ci!i) as Does 45 Flange of Required Round Lightening Hole.

V-l!5 lloh

Fig, D3.26

heavier fr~~e, as in ~eth0d (o), that c~n take the tension ~leld er:ects should be used.
D3.7 Special Cases of Beam Design.

(::

Providing ACCeSS ~1th Standard Round F'langed Holes or "Donut Doublers"

Frequently a cut-out size requ1rem rnt is such that a standard round flanged hole will provide the needed open space and stre~gth. In such cases either of the tc l Lowtng can be dene when the beam is of tte shear re~1stan~ type: 1) replace the web, locally, ~ith a panel having a standard. round hole that nas a 45 0 flange, as c~scussed in Chap~er
810.

There are several cases ir.volving be~ design not disc~ssed in c~~p~ers :la and Cll, since these chapters are intended pr1mari~y to ?resent funda~ental deSign princ1ples. Tee deSigner ','1111 encounter in practice, however, the following situations ~hich include ordinary straight beams and beams in the for.n of bulkheads or fr~es: a) b) Curved
bea~s

2)

Cut the required diameter hole into the web and attach a "donut." doubler (ring; that has the sam& (or greater) stiffness in a direc~icn nor.nal to the web as does the flange of (1) above.

rlanges with local cranges in direction ("bent" fl~n~es)

whlc~

c) d)

rjanaes subject to "norma.I" loads tend :0 b~nd tt~m.

In either case the allowable shear can be calculated as disc~ssed in Chapter 810. Th se allowables apply only to round holes, not to elll~tical or rectangular ~oles wi~~ :lar.~ed ~d~es, which are weaker. Item (2) above is 11lus~rated in ?ig. D3.27, Not.e that the :101e spac tng "b" 0:: :~~pter :10 will je quite large if only 0ne ~ole is ~resent and is not near the end of ~he team. The 'ioj,e epac Ing "b" 15, of course , used in 1ete~inlr~ the allowable shear i:: ~o s~i!feners are ~resent. If a beaded web 15 c~t, a panel c~n be inserted locally, Containing a ~ole with a beaded flange, as discussed in c~apter Cla ana these allowaoles ~sed ~onser~atively. If the be&~ is of ~h~ tension field ty~e, ~ethod (el does not apply. In su~h cases, a

s-tances requt r tng "stabilization" against bU~klin~ out 0: the ~lane of the 'Neb.

(a ccns icer a ::cr-:1J:1 'JI' a cur-ved oeam as shown in :1~, J3.28. ~;,e curvat.uz-e is te.cen as :::onstant, as 1s tn? ber.Ging ~O~8nt, ~or pur-poses c : t Lluat rat t on ,
~hen :he o~ter (~;~er flange in (a)) 1s 1:1 t.ens i on and t t-e i nne r r Iange tn concress i cr :i~re will ~e ~ sollaps1ng (8Gm~ressicn) Ioactng on the weD as encvn in (b:. This is because a "':'.OO?" Ioac Ing is recut red :0 xeep the r Ianees rr-cm -acvt ng t nwa rc , towar-ds each c t he r . ?'J, ".:0TIStan: Ioad , ?, ~!1 tne rIance and ~~c~ s:~~le s:atiss.

zp : 2'sR
W

,
0

4/6

.....
-

_,

.,

'-

~=~ r ~ .:,

A~ALYSrS

A~D

DESIGN

OF

FLIGHT VEHICLE

STRUCTURES

D3.11

\" Outer Flange

or-

Outer Flange

p.....b:?i "" ,~ P O,D l ' W9~'" 1'")' ..........

Web

c--L Inner <, p.

Flange
(a)

"")"';0
1

~'web .
~ l'

internal load ~nalysis. Then tte stiffener loads will be, Pst o
2P o tan Q
~

Wi--<'t

Psti = 2Pi tan

Pi

Pi/lJd~ ... Inner -,

t.

'(oit.

Flange

Any ~i~ference in stif~ener end loads will je sheared into the web through the attaching fasteners. The flanges should be deSigned ~or the axial loads, P, in them. Some allowance ~or beLcin~ ~cment, tl, (Fig. D3.30) can be rr~de by takin~ M as P x e x 1/2 at the center and - ? x e x 1/2 at the stiffener junctioLs. ~hese secondary ef~ects are sometimes ignored when obviously small, in practice.

Fig. D3.28
W

si~ce

is, of course, larger ~t the inner ~lange, R 1s smaller, tczn at the outer flange.
Fig. D3.30

This cJmpreSSive loading tends to collapse t:'18 'Neb. I:- the' web is also -23.IT:t':'::1g shear, as i3 ~~e ~sual case i::1 ?~~ti2e, an interacticn fa~ul~ ~~lc~ lnclu~es compression and shear s~resses should be used ~o pre~ict j~ckllng. This is the reason that the allowable s~resses quote1 tor beaded panels and ;anels witt flar.ged holes in 0t~pter CIO do nct ~;ply. These ~llowables were ~or strai~ht webs With no no~al loadS, as ar~ ;resent in cur-ved oeane .
2urva~ure is enough to cause compression in the web, stiffeners shaald be provided ~o take this lo~d. T~is can je done as sho~m in Fig. 03.29.

"el' is the eccentricity between s t t r r enerjunctions due to the flanges being curved rather than straight. The same a~plies to ~he inner- r Lange , except tba t u t is in compr-ess i on in this discussion.
~nen ~he outer flange is in comp~ession and the ~nner one in tenSion, ~ll of the loadings are reve~sed, as ~r. rl~. D3.28(0) and (d). The web is then subjected to tension loads and there is no collapsing problem. No stif:ene~s ~re required, ether tr~n ~cr the "normal" reasons. However, in ?ract::.ce, most loadings are reversible to some degree, and some com~ress::.on on the webs will then, of course, be present.

If the

si~nif:'cant

(b)

1='13.:1;95

'.;~:n Local (Bent :1::11.;;:es).

'-':lT~:1';:es

in Direction

This is a s98cial case 0: (a) above. Instead of a ~ener~l c~~~ture the ~lange ~as ~ leeal C1ange ~n cirection ~s i:1 Fig. D3.31.

Po is deter mined by assuming the

flanges to be straight between

stiffeners. 'I'his

gives the value of 9.

Fig. 03.31
~ove

?~
fl~~ge

and Pi car. be

:a~en

as

tt~

calculatec

de:e~ined

loads at the 5~i~fener J~~ction as ~r~~ statics, or from a bulkhead

The out s tanatng legs ~': the :'12.:1:;;85 Nil:' u;wards ~~les5 2. ~tif~~r.s~ J~cvides :~e lead required to CCl2 thes ~c}m. This oan ~Q calculated trom s~atics at the :oint. The

D3.12

SOME

IMPORTA~T

DETAILS

IN

STRUCTURAL DESIGN

10a':1 in tne st i r rene r is then shear-ed into the Neb. 1.:' the at Lrrene r is net orovtcec , some st~er~th ~ill ce p~esent ~~less the kick load put s tne web .rn c ompr-eas t on , in whrch case it Nill ;:'.1:~{1e ~~ thin. If the beam is nac nt ned ~~o~ :ar stock the stiffener can be machined in p.lac e . If the beam is "but Lt up" from sepa~te :la~~es and web parts a stiffener can be ~de up from a tee ~ember attached to the :la~ge ane. an ar~le member attached to the web and to the tee. This is illustrated in Fig. D3.32. The kick load path 1s ~rom the flange to the tee to the angle to the web.

fuselage bUlkhead or fr~~e (the outer :lange is stabilized, uSU~lly, by the outer Skin). The inner flange will usually be subject to a compression load over ~uch o~ its length for some design condition. Since the ~l&~ge will have little sti~fness of its own, its Lip will be small and it will buckle at VerJ low com?ressive stress levels as a long column. It is then necessary to prOVide su?porting members, as shown in Fig. D3.33, to reduce the unsupported length and bring the bUCkling stress up to an efficiently high level, nearer to the local crippling strength of the flange.

Outer Skin

11 ..l---Frames with Inner

Fig. D3.32

The loading shown will procuce an '.1p load on the bolts, putting tension into the leg of the tee and the stiffener angle. A reversed loading would ;ush down on the tee producing :c~pression. The above tee and angle co~birztion ~culj, J! course, be replaced by a 5~r~le ~achined ~ltting. The important thing is ~hat an adequate stiffener, attached to the o~tstanding legs, b~ ?r~ser.t if all of the o~tstanding leg ~s to f~nction effectively.
(c; Flanges SUbjected to Nor.nal Loads Tending to Bend Them.

Flanges Stabilized Against Out-of- Plane Buckling USing Axial MemberS (Tubes) and Fittings or Inter costals , Flange Must be Attached to Supporting Member in Stiff Manner
<

Fig. D3.33

Fr'equent.Ly the flanges 0 f' a beam are Subjected to loads which pull out'Nard or pUSh inward. It is important in each case that the :lange be ffbacked Upff by a stiffener. The ef~ect is the same as in the case just discussed, (b), i~volving kick loads due to bent flanges. A si~ilar stif:ener arrangement :2.n be :;.sed.

The supporting member will not be subject to any appreciable load but it must have sufficient stiffness to prevent-corumn buckling. The required stiffnesS criteria will not be discussed here but the reader should consult Ref. (1) or similar textbooks to obtain such criteria. Since the su?porting members are sized by stiftness rather than strength reqUirements they can usually consist of light tubes or intercostals. Their ~eight is usually less than would result from beefing-up the bul~~ead flange for out-of-plane strength. The stiffness of the supporting ~ember must, of course, include the end fitting attaching it to the ~lange.
D3.8 Structural Skin Panel Details.
~he general ?rinciples :nvolvt~g the deSign ~! structU~l skin or floor panels are cOVered in Chapters C10 and Cli. Chapter :11 concerning buckling panels, in ~art:cular, should be thoroughly understood ~y the deSigner. In addition to this information, several details of deSign are presented below.

(i)

Flanges Requiring St2.bilization Against Cut-Df-Plane 3~ckling.

In the cases or ~lange deSign discussed en Chacters CIO and ell, it 'HaS ass~ed that any fla~e in compreSSion was stabilized, or ?revented frcm buckling as a col~~. jy some sUPporting member. ~his ~ember was usually shown there as a :loor or a skIn to which the flange was attached. Thus the flar.~e could not bUckle since it was restrained in one plane by the be~ web and in a plane nor.nal to this by some s~1n.
(n~rmal to ~he web) supporting ~emb3r ~s'not

a)
b)

Rec~angular

holes of long
axi~l ~embers

Recessed on paneLs

p~nels

Occasionally, however,

~his out-Qf-nl~ne

c) d) e)

Install~tton

inherently present ~nd must be nrovided. typical example ~s the 1~2r ~lan6e of ~

Spot welding sheet

~etal

doublers

Tension skin splices

ANALYSIS AND DESIGN

OF

FLIGHT VEHICLE STRUCTURES

D3.l3

(al

Rectangular Holes.

It is frequently necessary to cut rectangular holes into load carrying panels to ?rovide doors or to mount equipment. ~hese h~les must, of course, be framed as discussed in Art. D3.6. Equally important from a fatigue standpoint is that the internal corner radii not be too small or cracks will event~lly start there. As an arbitrary design requirement it is suggested that the corner radii be maintained at R = .35 inches or larger With the normally obtained flnish. In those cases where a smaller radius is absolutely necessary, it is suggested that the lower limit of R 10 be maintained, and that an f-40 finish be specified around the edge of the panel at the comer. (b) Recessed Panels.

It is sometilnes necessary to recess a structural panel locally in Order to mount equipment, as shown in Fig. D3.34.
Corner Members

Transferring Load P to Beams

become "effective". That is , it 'Ni11 develope axial stresses as it strains alon~ with the skin, particu:arly in tension. Xo~t of the load picked up will enter t~~ough the fasteners near the ends of the me~bers. The member tends to strain the same amount as the ?anel it is attached to thou~h it never is as much due to the fleXibility of the fasteners. Thus the total load will be P = f ski n X Amember as an upper limit, but only 60% to 80% of the skin axial st~ess is nor.nally developed in the member. The larger the member and the greater its length the more axial load it will develope. Most of this load can be considered to enter through the outer (end) 20% of the fasteners at each end. High loads and relatively large bearing streSSes will thus be present here. When these are present in the skin panel along with high skin tension stresses, the fatigue life of the skin panel will suffer. Therefore to keep these induced fastener loads lower, axial members (other than prUnary structure) should be kept as short as is practical. If their area is large then their ends should be tapered. There are methods for evaluating these etfects more specifically, but they are beyond the scope of this article. The possible deleterious effect ahould be anticipated as it can sometimes caUSe cracks in panels. Whenever a member is installed on a buckling (diagonal tension) shear panel it must be strong enough to prevent the posSibility of Wtorced crippling" failure due to the action ot the buckling skin. The reader should consult Chapter ell in this respect for deSign criteria. (d) spot Welded Doublers.

P '" Out-of- Plane

Sell
Balancing . / ,--_Kick Loads arried to Re-

Kick Load

'" 2'10 x d. (Self

(d '"

Balancing) depth of recess)

Fig. D3.34

by Beams

distribution Structure Provided

The recessed panel can continue to carry its shear load but there will be out-of-plane Wkick" loads. TheSe are resisted by the framing ~embers waw and wb w as shown in the figure and carried over to some beam or frame that can redistribute them into the main side panels. If the recess must be very deep and its size is not too large, it may be better to omit the panel and Simply rrame the resulting hole. ThiS is more likely to be the case if hig~y buckling Skins are involved. The recessed panel C3IL~ot, of course, carry tension stresses, only shear. The hole in the basic skin at the recess must not have sharp corners, as discussed in (a) above. (cl Installation ot Panels.
_~lal

Frequently some bay or area of a skin or floor panel must be made thicker than other bays because of higher local shear flows. For example. a skin of .040 thickness might have to be made .064 in a bay becaUSe of higher loads as shown in Fig. D3.35. Rivets Along Edge of Doubler

.040 Skin

-.,...-~

Members on Skin

Whenever a local axial member, meaning one lying in the direction at the main bending stresses or the overall s~ructure (i.e. foreand-aft in a :uselage), 15 installed, care should be observed. This member will tend to

Fig. D3.35 One common way of achieving thiS, !or large panels, is to spot weld (or sometimes to bond) an .025 doubler to the .040 skin. The spot welds shoulc be put in a~ a close spacing to ~ke the combination act as unit rather tr~n as two separate skins, which would

03.14

SOME IMPORTANT DETAILS

IN

STRUCTURAL DESIGN

J:

It Is considered along the edge of ~art~cu~arlY 1~ the ~anel Is deSign wlt~ its accompanying
we~~ar. riv~ts

use

gOOd practice to the doubler, or a tension field buckles.

rows Should be ~ept as close together as is ?racticable).

l
II

(9)

~en81on

Skin Splices.

0 0

+ + + I -- - - 0- 0 0
0 0 0 0
0 0

------0 0 0

Skin splices should be kept to a minimum

number, but teey are common in aircraft str~ctureo. The splices of major structural skins 3~oulj be given due consideration since they al~ays contain stress concentrations and
11~lt

0 0

0 0 0 0 0

0 0 0 0

0
0

a a a 0 a a 0 0
0 0

0
0

the fatigue lIte.

~here

- -

- -,

are two factors present in tension

I'

--

0 0 0

0
0 0

a
0

a
I

[
II

--------

a o

splices which reduce the fatigue life:

Good Practice (Holes in Line)

Poor practtce (Holes S~ered)

a)

There Is the basic tension stress in the skin being spliced and the stress

I
I

Fig. 03.36

concentration due to the holes.

b) There Is also the bearing stress in the holes, due to the fastener splice loads, which worsens the situation. That is, ti~Q Combination of tens ton and high bearing stresses in the skins is worse than tension stresses only.

1
Splice Bolts

: ::p:::
t
Cross-Section of a Double Shear Tension Skin Splice With Skins Machined Thicker at the Splice Area. (Angle Shown Due to Wing Dihedral and Thickness Taper. )

To keep the fatigue life as large as Possible, the tollowing practices should be obServed; a) When more than one row of fasteners per side is required, as is the usual case: 1. Do not "stagger" the fastener pattern but keep the holes in line, as Shown in Fig. D3.36. This gives a lower stress concentration than staggered holes. It is an interesting fact that ~NO or more holes in line in the direction at the load gives a lower stress concentration factor than does a single hole. Avoid USing more than 3 rows of attachments per side. possible use a double shear splice

I
Ii

Fig. 03.37
D3.9 Additional Important Structural Details.

The following list of details 1s presented with a minimum amount of comment as representative of ngood practice". In addition to the following list, the details listed in Chapter 825 nFatigue AnalySiS and Fail Safe Design n should be observed.
1.

2.

T~en

(a splice plata on each side of the

skins). This is frequently seen in centerline splices of wing skins. ~nen the skins are machined or chem~illed they can be left thicker at the splice to reduce the tension and bearing stresses locally, as shown in Fig. D3.37. c) r.aintain a fastener spaCing in each row and between ~ows of approximately rour times the ~astener diameter (the

Avoid miXing hole-filling and ~cn-hol~ tilling fasteners in the same Dattern (i.e. aluminum and hi-shear type rivets or steel bol~s). ~hen this cannot be aVOided, ream the holes for the nonhole-filling fasteners to insure their picking up load better than a plain drilled hole, with its wslopn, would produce. 'Jhen USing less than ~our non-hale-filling rasteners in a pattern, use reamed holes to insure a better ~!stribution ~f load ~on~ the fasteners.
~en using !asteners in thi~ Sheets where the value of Dlt (fastener diameter to sheet thickness is 5Teater than 5.5) determine allowables tram tests or use

2.

3.

ANALYSIS AND DESIGN OF FLIGHT VEHICLE STRUCTURES

D3. 15

conservative extrapolations. The thin sheet tends to ~buckle" around the hole at relatively low bearing stresses. 4. 5. Do not use hi-shear type fasteners in jOlnts where the bottom skin is dimpled. Maintain an arbitrary margin of safety or .15 in shear jOints for fastener patterns to allow for uneven distribution of loads. Do not use spot welds to attach buckling skins to their supporting frames unless a "one_Shot" structure, such as a missile, is involved. Even in these cases do not use spot welds on either side at the joggled area of a joggled member; use rivets at the joggle. Do not use a "long string" of fasteners in a splice. In such cases the end fasteners will load up first and yield early. Three, or at most 4, fasteners per side Is the upper limit unless a carefully tapered, thoroughly analyzed splice Is used. Such a 1esign can be fashioned but is beyond the scope of this article. Care tully insure against feather edges in all fitting design. Re-entrant surface intersections must have their edges rounded o~ else fatigue cracks will ineVitably begin in such places. Anyangle between surfaces less tr~n 700 can be considered a feather-edge. These commonly occur in design if not watched, in drilling ~d other machining operation call-outs. The permanent buckling data presented in Chapter Cli applies to structures similar to fUselages where the sub-structure rings are closely spaced as compared to the radius of curvature of the skin. That is, the ratio of skin support spaCing to radius of curvature is small, conSiderably less than one. ~hen the spacing of the sub-structure becomes larger, permanent buckling Should be considered to occur at approximately the same shear stress that produces initial buckling; that is when the above ratio approaches unity. Avoid the use of "open section" members when torsion is present. Open section members are extremely flexible compared to closed sections as can be seen fram Chapter A6. For a given torsional stiffness, an open section member will be far heavier than a closed one. In the examole belOW, member ~bK is 65 times as s~if: as member ~a" for pure torSion applied at each e~d (see also ~he example in Chapter
Fig. 03.38

'"b"

6.

7.

A6). For even a closed section to operate efficiently the torsion should be ~apable Or being distributed into all sides of the closed section. This ~y require a "buL<head" type end fitting, or a ~uch thicker section locally where the torsion is put in. The action 1s similar to a fuselage bulkhead distributing a ~Nist loading into the skins of the ruselage or a wing rib distributing an applied twist into the skins and spar webs. 11. Compressive buckling does not necessarily mean failure. It means failure only if there is no other member to keep taking additior~l load. Shear buckling silnply indicates that additional load must be carried as diagonal tension as discussed in Chapter Cll. Ot course, the members supporting the web must be able to withstand the ensuing so-called "seccndaryK effects. When the compression skins of a fuselage or wing buckle they will carry no additional compression load but the stringers and tlanges are still capable of thiS, as discussed in Chapter Al9. The buckled skins can, however, carry additional shear load through tension field action. Thus to achieve a light efficient des1gn the deSigner should have a thorough understanding or the factors involved after members have buckled as covered in other chapters of this book.
12.

8.

9.

10.

Probably the ~ost single important item regarding detail structural deSign is the ~tter of eqUilibrium. If the designer will showthe load eqUilibrium for sveiJ part of his assembly, most errors ~ll1 be prevented. The majority of all struct~ral strength problems occur Si~ply because the laws of statics have not been observed, and it 1s usually in the ~ller detail parts that the time 1s not taken to do this. It takes a considerable amount of experience to safely substitute the ~eyeball" for the sllde rule and data book. The begln.~ing detail designer and ~ny others who r2ve been at it for a While simply do not have this experience

03.16

SOME IMPORTANT DETAILS IN

STRUCTURAL DESIGN

i~ all areas of structural design. The only safe course in such circumstances is always to show the member loads in static balance. ~or every part of a structural assembly. When this is done it will also give the deSigner a better feel as to how the structure is actually detlecting under load. This can be ot Significant help in anticipating problems where members are Joined together and therefore must PUSh. pull or pry on each other when loaded.

REFERENCES :

1.

~Theory of Elastic Stability~, Timoshenko. ~Buckling ot Metal Structures~, Bleich, F.