INTRODUCTION TO AERONAUTICS: A DESIGN PERSPECTIVE CHAPTER 4: WINGS AND AIRPLANES

After running the engine and propellers a few minutes to get them in working order, I got on the machine at 10:35 for the first trial. The wind, according to our anemometers at this time was lowing a little o!er "0 miles# "$ miles according to the go!ernment anemometer at %itt& 'awk. (n slipping the rope the machine started off increasing in speed to pro a l& $ or ) miles. The machine lifted from the truck *ust as it was entering the fourth rail.+ ,rom the -iar& of (r!ille .right for -ecem er 1$, 1/03

4.1 DESIGN MOTIVATION The Lift and D a! "f Win!# The study of airfoils in Chapter 3 gave insight into how wings generate lift, but it did not tell the whole story. The flow over a wing near the wingtips is very different from the two-dimensional flow around an airfoil. The differences have profound effects on the lift and drag generated by a wing. Understanding these effects is crucial to the aircraft designer who must shape an aircrafts wing to optimize its performance. ection !." discusses wing lift and drag theory and analysis methods. Wh"$e Ai % aft Lift C& 'e #ther components besides the wing contribute to an aircrafts lift. The lift contributions of the aircrafts fuselage, control surfaces, high-lift devices, stra$es, etc. must all be considered in order to accurately predict an aircrafts lifting capability. The aircrafts ma%imum lift coefficient is one of the governing factors in an aircrafts instantaneous turn capability, landing speed and distance, and ta$eoff speed and distance. ection !.3 describes a variety of devices for increasing an airplanes ma%imum lift coefficent, while ection !.! presents methods for estimating the lift curve slope and ma%imum lift coefficient of a complete airplane, including the effects of stra$es, high-lift devices, control surfaces, etc. Wh"$e Ai % aft D a! P"$a The drag of all aircraft components must also be included when estimating whole aircraft drag. The variation of an aircrafts drag coefficient with its lift coefficient is called the aircrafts d a! ("$a . The drag polar is the $ey information about an aircraft needed to estimate most types of aircraft performance. &ircraft ma%imum speed, rate and angle of climb, range, and endurance depend so heavily on an aircrafts drag polar that a '( change in drag can ma$e a huge difference in a )et fighters combat effecitivenes or an airliners profit potential. ection !.* presents a simple method for predicting an airplanes drag polar at low speeds, while ection !.+ describes how high flight ,ach numbers change an airplanes aerodynamics. ection !.+ also presents methods for estimating a complete airplanes lift curve slope and drag polar at high flight ,ach numbers. ection !.- is an e%ample of an aerodynamic analysis for a supersonic )et fighter aircraft. The analysis predicts aircraft lift and drag characteristics for ,ach numbers ranging from . to "... 4.) WINGS The Lan!&a!e /igure !.' illustrates a view of a wing planform with some of the important dimensions, angles and parameters used to describe the shape of an aircraft wing. The wing span , , is measured from *in! ti( to wing tip. The symbol c is used for the chord length of an airfoil at any point along the wing span. The subscript r indicates the chord length at the wing ""t or the aircraft centerline. The subscript t denotes the wing tip chord. The overbar denotes an average value of chord length for the entire wing. The symbol A0 is used for a parameter called a#(e%t ati". &spect ratio indicates how short and stubby or long and s$inny the wing is. The symbol is used for wing #*ee( an!$e with the subscript 12 denoting

0'

the wing leading edge. The subscript "5 denotes the line connecting the "*( chord positions on each airfoil of the wing. The symbol is used for the wing ta(e ati", or ratio of tip chord to root chord.

=
12 ."5 c

ct cr
"

(4.1) (4.2) (4.3)

3 3 = c

A0 =

cr

ct

b
+i!& e 4.1 +inite Win! Ge",et - Definiti"n# /igure !." shows a side view of the wing to illustrate the an!$e "f t*i#t. 1ings which are twisted so that the wing tip airfoil is at a lower angle of attac$ than the wing root airfoil are said to have *a#h"&t. 1ing twist in the opposite sense from washout is *a#hin. 1ing twist of this sort is also called !e",et i% t*i#t. &n effective twist of the wing can also be achieved by changing the airfoil shape along the wing span. 2f the airfoil at the wingtip has less camber than the airfoil at the root, this has much the same affect on the wing lift as if the airfoils were the same but the wingtip airfoil was at a lower angle of attac$ than the root. Changing airfoils along the wing span in this way is called ae "d-na,i% t*i#t.

Root

Angle of Twist

Tip

+i!& e 4.) Win! T*i#t Win!ti( V" ti%e# The flow around a wing section which spans the test section of a wind tunnel appro%imates the flow around a wing with an infinite span, no twist, and a constant chord length along its span . 2n Chapter 3, this type of flow was labeled two-dimensional, because flow properties did not vary in the direction. The flowfield around a finite *in!, or wing with a finite span is not two-dimensional. The ma)ority of differences between the flow around a finite wing and that around an infinite wing result from flow phenomena which occur at the wingtips. /igure !.3 shows a front view of the flowfield around a finite wing. 3ote that the differences between the pressures above and below the wing which produce lift also produce a strong flow around the wing tip. The arrows in /igure !.3 are intended to illustrate a front view of flow streamlines in the plane of the *.( chord point on the wing. The lengths of the tails of the arrows do not indicate the magnitude of the velocity vectors. #f course, the actual magnitudes of the velocity vectors must be such that there is no flow through the surface of the wing.

0"

"ownwas!
TOP SURFACE (relative low pressure

(relative !ig! pressure BOTTOM SURFACE

+i!& e 4.. + "nt Vie* "f Win! *ith +$"* A "&nd the Win! Ti(# &s shown in /igure !.!, these circular flow patterns around the wing tips become concentrated into very strong tornado-li$e swirling flows $nown as *in!ti( '" ti%e# or t ai$in! '" ti%e# . The trailing vortices generated by large aircraft persist for many miles behind them and can pose serious hazards to smaller aircraft which encounter them. &ir traffic controllers must allow sufficient spacing between aircraft so that the action of air viscosity and turbulence can dissipate a preceeding planes trailing vortices before the arrival of the ne%t one. This spacing re4uirement to allow vorte% dissipation is the limiting factor on traffic density at most commercial airports.

+i!& e 4.4 T ai$in! V" ti%e# D"*n*a#h &lso note in /igure !.3 that the circular flow pattern around the wingtips results in a downward component to the flow over the wing. This downward flow component is called d"*n*a#h. /igure !.* shows that downwash adds vectorially to the freestream velocity to change the direction of the flow velocity. 3ote that the resulting total velocity vector still results in flow parallel to the wing surface, but the orientation of the effective free stream velocity direction relative to the airfoil is altered.

03

Effective free strea# $irection in vicinit% of t!e wing

"ownwas!

+i!& e 4./ D"*n*a#h The change in flow direction due to downwash is called the d"*n*a#h an!$e, and is given the symbol . The angle between the airfoil chord line and the local flow velocity vector is called the effective angle of attac$, eff . 5ach individual wing sections lift, drag, and angle of attac$ vary with the airfoils orientation to this local flow direction, but the whole wings lift, drag and angle of attac$ must still be defined relative to the free stream direction. /igure !.+ reveals that, as a conse4uence of the change in effective flow direction caused by the downwash, the effective angle of attac$ of the airfoil is reduced, and the lift generated by each airfoil has a component in the wings drag direction. This component of lift in the drag direction is called ind&%ed d a!. The reduction in effective angle of attac$ due to the downwash causes the wing to produce less lift than it would if there were no downwash.

C!or$ line

&ift i 'n$uce$ "rag( " i

4
&ocal flow $irection

eff Parallel to c!or$ line

+i!& e 4.0 D"*n*a#h An!$e and Ind&%ed D a! /igure !.- ilustrates lift coefficient curves for an airfoil and for a finite wing with the same airfoil section shape. 3ote that cl denotes two-dimensional airfoil lift coefficient while6 51 7 1 6 73 8!.!9

is used for the three-dimensional finite wing lift coefficient. This same convention will be followed for cd and6 5- 7 - 6 73 8!.*9

0!

The reduction in effective angle of attac$ due to downwash decreases lift at any given and delays stall to higher values of . &s in Chapter 3, slopes of the lift curves are defined as6

cl

cl

and

51

51

8!.+9

c l and C&
Airfoil

cl C 1

Wing

+i!& e 4.1 T*"2Di,en#i"na$ and Th ee2Di,en#i"na$ Lift C"effi%ient C& 'e# S(an*i#e Lift Di#t i3&ti"n Unli$e the two-dimensional flow around an airfoil in a wind tunnel, the flow around a finite wing varies in the spanwise direction. This spanwise variation is primarily due to the inability of the wing to support a pressure difference at its tips 8the cause of trailing vortices9. 2t may be influenced by wing taper, wing twist, or even differences in airfoil shape at different spanwise positions on the wing. panwise variation of airfoil shape is called ae "d-na,i% t*i#t. :ut even an untapered, untwisted, unswept wing still has spanwise variation of the flowfield around it. This is because the trailing vortices on such a wing have a stronger effect and produce more downwash near the wing tips than they do far from the tips. &s a result, even though the wing is not twisted, increasing downwash reduces effective angle of attac$ and therefore lift near the wing tips. Tapering the wing or giving it wash out can help reduce this effect. 2n fact, a wing which is tapered and;or twisted to give an elliptical spanwise distribution of lift will have a constant downwash at every spanwise position. /igure !.0 shows an elliptical spanwise lift distribution. &n untwisted wing with an elliptical planform will have an elliptical lift distribution. &s shown in /igure !.<, the famous upermarine pitfire and =epublic >-!- Thunderbolt fighter aircraft of 1orld 1ar 22 both used elliptical wing planforms. uch wings are relatively comple% and e%pensive to build, so straight-tapered wings are much more common.

0*

&ift Per Unit Span

,b*+

)b*+

+i!& e 4.4 E$$i(ti%a$ Lift Di#t i3&ti"n

+i!& e 4.5 The S&(e ,a ine S(itfi e and Re(&3$i% P241 Th&nde 3"$t +i!hte Ai % aft "f W" $d Wa II 6"th Had E$$i(ti%a$ Win! P$anf" ,# 7Ph"t"# C"& te#- Nati"na$ Ai and S(a%e M&#e&,8

+inite Win! Ind&%ed D a! /igure !.+ shows that induced drag is a component of the three-dimensional lift in the drag direction6

-i = 1 sin

or 5-i = 51 sin

8!.-9

2t can be shown that the induced angle of attac$ everywhere along the span of wings with elliptical lift distributions is given by6

=
/or small, sin

51 *-.3 5 1 radians = degrees A0 A0

8in radians9 and 6

5-i = 51 = 51
S(an Effi%ien%- +a%t"

51 51 " = A0 A0

8!.09

0+

54uation 8!.09 applies only to wings with elliptical lift distributions. ?owever, it is possible to modify 8!.09 slightly to ma$e it apply to any wing by using a #(an effi%ien%- fa%t" , e, such that6

5-i =

51 " e A0

8!.<9

The value of e is ' for elliptical wings and between .* and ' for most common wing shapes. +inite Win! T"ta$ D a! The total drag of the wing is the sum of profile drag and induced drag6

5 - = cd +

51 e A0

"

8!.'.9

=ecall, however, from Chapter 3 that profile drag is composed of s$in friction drag and pressure drag. /igure !.'. illustrates the variation of each type of drag with lift coefficient.

Total @rag 2nduced @rag >ressure @rag $in /riction @rag

+i!& e 4.19 +inite Win! T"ta$ D a!

Win!$et# and Ti( P$ate# & variety of devices have been used on aircraft to reduce induced drag. /igure !.''8a9 shows three such devices. #f the three, the winglet is the most effective and most widely used. 2n addition, )et fighter aircraft which carry fuel tan$s or air-to-air missiles on their wingtips e%perience a small reduction in induced drag when such wingtip stores are in place. &ll of these devices inhibit the formation of the wingtip vortices and therefore reduce downwash and induced drag. /igure !.''8b9 shows a winglet on the wingtip of a ,c@onnell-@ouglas C-'-.

0-

Front -iew

Front -iew

Front -iew

la Tip P

te

"roope$ Tip

gl .in

et

7a8 Th ee Ind&%ed2D a!2Red&%in! Win!ti( De'i%e#

738 Win!$et# "n a C211 7USA+ Ph"t"8 +i!& e 4.11 Win!ti( De'i%e# f" Red&%in! Ind&%ed D a! #f course, )ust e%tending the wing to increase its span and aspect ratio will have a similar effect. ?owever, the increased lift far out at the end of the wing will increase the bending moment at the wing root and create greater loads on the wing root structure. The winglet increases wing span only slightly. 2t is preferred because it achieves an effective increase in aspect ratio without significantly increasing wing root structural loads. +inite Win! Lift ince the the induced angle of attac$ for a wing with an elliptical lift distribution is constant everywhere along the span, it is relatively easy to determine the lift of such a wing. 2f the wing has an elliptical planform and no geometric or aerodynamic twist, it will have an elliptical lift distribution for a wide range of angles of attac$. & twisted rectangular or tapered wing will normally achieve a true elliptical lift distribution at only one angle of attac$. The elliptical planform wings zero-lift angle of attac$ will be the same as for its airfoil section. &t an arbitrary positive angle of attac$ below stall, an elliptical wings effective angle of attac$ will be given by6

eff = =

*-.3 51 A0

8!.''9

&s shown in /igure !.-, the airfoil and finite wing lift curve slopes may be represented as6

00

cl =

cl 8 l = . 9

51 =

51 8 1 = . 9

8!.'"9

where is any arbitrary angle of attac$ in the linear range of the lift curves. 51 and cl are the lift coefficients at that arbitrary value of . /rom /igure !.- we recognize that6

51 = 51 8 1 = . 9 = cl 8 eff 1 = . 9 = cl 8

*-.3 51 1 =. 9 A0

8!.'39

Combining 8!.'"9 and 8!.'39, the e%pression for

5 1 becomes6
cl *-.3 cl

51 =

'+

A0

/ollowing the same convention as in 8!.-9 for non-elliptical wings, the e%pression can be written6

51 =

'+

cl *-.3 cl

8!.'!9

e A0

3ote that in general for a given wing, the value of e re4uired for 8!.'!9 is not the same as that re4uired for 8!.-9. The two values are typically 4uite close to each other, however. E:a,($e 4.1 & wing with a rectangular planform, a 3&C& "!'" airfoil, a span of * m and a chord of " m is operating in standard sea level conditions at a free stream velocity of !" m;s and an angle of attac$ of 0 degrees. 2f the wings span efficiency factor is ..<, how much lift and drag is it generatingA olution6 The aerodynamic properties of the airfoil may depend on the =eynolds number, which for standard sea level conditions and a free stream velocity of !" m;s is6

0e =

. slug ; ft 3 ( !" m ; s)( " m ) 4c '""* = = *,-*',0' ......'-0< $g ; m sec9

so the airfoil data curves for 0e 7 *.- million 8not standard roughness9 will be used. The values of 18. and the cl at 7 0o do not, in fact, vary with =eynolds number. Their values can be read from /igure 3."0 as6 18. 7 - "o, at 7 0o, cl 7 '..* ince the lift coefficient curve appears linear between 18. 7 - "o and 7 0o, the lift curve slope may be estimated as the change in lift coefficient divided by the change in angle of attac$6

0<

cl =

'..* . = .'.* . ;o o o 0 8 " 9

cd 7 ....<0

&lso from /igure 3."0, for cl 7 '..* and 0e 7 *.- million6

The dynamic pressure for the test is6

7 = ' ' " 4 " = '""* . $g ; m 3 ( !" m ; s) = ',.0. 3 ; m " " "

The airfoils planfor m are a is its chord multiplied by its spa n6 3 7

.

c 7 * m

" m 7 '. m "

2ts asp e c t ratio is det e r mi n e d using 8!."96

A0 =

"

( " m) "
'. m "

= ".*

and the finite wing lift curve slop e is predict e d by 8!.'!96

cl .'.* . ;o o 51 = = o = ..*+. ; *-3 . cl 8*-3 . ; rad98.'.* . ;o9 '+ '+ 8.< . 98"* .9 eA0
The lift coefficient is then calcul at e d using 8!.'396

51 = 51 8 1= . 9 = ...*+- ; o 80 o 8 " o 99 = ..*+2f the wing had an elliptical planfor m , the airfoil lift coefficient everyw h e r e on the wing would e4u al the finite wing lift coefficient, and a differe n t valu e of c d could be rea d from the airfoil chart for this lower c l valu e. ?owev e r, for a rect a n g ul a r planfor m , c l varies, and as a cons e rv a tiv e esti m a t e of the aver a g e value of c d , the valu e of c d rea d from the airfoil dat a chart for cl 7 '..* is used. The finite wing drag coefficient is then calculated using 8!.'.96

5- = c d +

51 ..*+- " = ....<0 + = ...** e A0 8 ..<9 8 ".*9

"

The lift, dra g, and mo m e n t abou t the aero d yn a m i c cent e r are the n given by6

<.

1 7 5 1 7 3 8 ..*+- 8',.0 . 3;m " 9 8'. m " 9 7 +,'"! 3 - 7 5 - 7 3 8 ...** 8',.0 . 3;m " 9 8'. m " 9 7 *<- 3 2t is inter e s ti n g to com p a r e thes e results with the force s gen e r a t e d by an airfoil in a wind tunn el with the sa m e geo m e t r y and free stre a m conditions, but purely twodim e n si o n al flow arou n d it, as calcul at e d in 5%am pl e 3.<. The decr e a s e in lift and incre a s e in drag caus e d by the thre e- dim e n si o n al flow arou n d the finite wings tips is significan t. 4.. HIGH2LI+T DEVICES =elatively thin airfoils with low camber generally give low drag at high speeds. Unfortunately, these airfoils also typically have relatively low values of ma%imum lift coefficient. ,ost aircraft are e4uipped with devices which can be used to increase lift when needed, at the e%pense of additional drag. These devices are of several types. T ai$in!2Ed!e +$a(# ,oveable surfaces on the rear portion of the wing which can be deflected downward to increase the wings camber are called t ai$in!2ed!e f$a(# or simply flaps. /igure !.'" shows four different types of flaps. The plain flap changes camber to increase lift, but its effect is limited by additional flow separation which occurs when it is deflected. The additional separation occurs because the upper surface of the deflected flap e%periences a stronger adverse pressure gradient . The split flap deflects only the underside of the trailing edge so that, while it creates a great deal of pressure drag, it avoids the strong adverse pressure gradient on its upper surface and therefore $eeps the flow attached slightly longer. This gives the split flap slightly greater lift.

+i!& e 4.1) T ai$in!2Ed!e +$a(# 7Ada(ted f ", Refe en%e 18 S$"tted f$a(# have a gap or #$"t in them to allow faster-moving air from the lower surface to flow over the upper surface. The higher-energy air from the slot gives the boundary layer more energy to fight the adverse pressure gradient and delay separation. & single-slotted flap creates the slot by moving away from the wing slightly when it is deflected. @ouble- and triple-slotted flaps are also used. 5ach slot admits more high-energy air onto the upper surface to further delay separation and increase lift. The +"*$e f$a( moves aft to increase the wing area before deflecting downward to increase camber. /owler flaps usually have one or more slots to increase their effectiveness. /igure !.'3 shows airfoil lift and drag coefficient curves for a typical trailing-edge flap. 3ote that in general the effect of flaps is to increase camber, moving the lift curve up and to the left. /or flaps other than /owler flaps the lift curve slope is unchanged. The angle of attac$ for zero lift is made more negative. 1ith the flap e%tended, the wing generates more lift at all angles of attac$ below stall. The

<'

ma%imum lift coefficient is greater, but it occurs at a lower angle of attac$. The amount of this shift in l80 and increase in 51ma9 is different for each type of flap. lots in flaps help delay the stall to higher angles of attac$ and higher values of 51ma9. The lift curve slope increases when /owler flaps are used. This is because /owler flaps increase the actual lifting area of the wing when they are e%tended, but the lift coefficient is defined using the same reference planform area as when the flaps are retracted.

Wing with Flap

CD

Wing with Flap

CL

Basic Wing Section

Basic Wing Section

CL

+i!& e 4.1. Lift and D a! C"effi%ient C& 'e# f" Win!# *ith +$a(# St a;e# and Leadin!2Ed!e E:tenti"n# /igure !.'! shows a #t a;e on an /-'+. & similar device on the /-'0 is referred to as a $eadin! ed!e e:ten#i"n 7LE<8. The stra$e has a sharp leading edge. 1hen the aircraft operates at high angles of attac$, the flow cannot stay attached as it flows over the sharp stra$e leading edge, and it separates. :ecause the leading edge of the stra$e is highly swept, the separated flow does not brea$ down into turbulence, but instead rolls up into a tornado-li$e '" te:. The vorte% generates an intense low pressure field which, since it is on the upper surface of the stra$e and wing, increases lift. The presence of the vorte% gives the rest of the wing a more favorable pressure gradient, so that stall is delayed. The stra$e also increases the total lifting area, but it is usually not included in the reference planform area. Therefore, the stra$e increases lift coefficient curve slope even at low angles of attac$ when the vorte% does not form. /igure !.'* shows lift and drag coefficient curves for a wing with and without stra$es. 3ote that at relatively high angles of attac$, the lift curve for the wing with stra$es is actually above the dotted line which is an e%tension of the linear region of the curve. 2t is at these angles of attac$ where stra$es are most effective.

Stra/es

Stra/e -ortices

<"

+i!& e 4.14 +210 St a;e#

Wing with Strake

CL

Wing with No Strake

+i!& e 4.1/ Lift C"effi%ient C& 'e# f" Win! A$"ne and Win! *ith St a;e

Leadin!2Ed!e +$a(#= and S$at# /igure !.'+ shows several devices which are used on wings to increase lift. >lain $eadin! ed!e f$a(# deflect to increase wing camber and move the point of minimum pressure further aft on the upper surface of the airfoil at high angles of attac$. The aft movement of the point of minimum pressure e%tends the region of favorable pressure gradient and delays separation. & fi%ed #$"t may be used to admit higher-speed air onto the upper wing surface to re-energize the boundary layer and delay separation. & #$at is a leading edge flap which, when it is e%tended, opens up a slot as well. &ll three leading-edge devices delay stall and e%tend the lift curve to higher angles of attac$ and higher ma%imum lift coefficients. :ecause angle of attac$ is defined using the chord line of the airfoil with no high-lift devices e%tended, e%tending a leading-edge device may actually decrease the lift coefficient at a particular angle of attac$. ome slats increase the lifting area when they are deployed, so they increase the lift curve slope li$e /owler flaps. /igure !.'- illustrates lift coefficient curves for a wing with and without a typical leading-edge slot, slat, or flap. The magnitude of the increase in ma%imum lift coefficient and stall angle of attac$ is different for each type of leading-edge device.

<3

+i!& e 4.10 Leadin!2Ed!e +$a(# and 6"&nda - La-e C"nt "$ De'i%e# 7Ada(ted f ", Refe en%e 18 6"&nda - La-e C"nt "$ ince flow separation and stall are caused by depletion of flow velocity in the boundary layer, several methods may be used to remove or re-energize this low-energy air and delay separation. #ne method is to drill thousands of tiny holes in the wing surface and use #&%ti"n to pull the low-energy air inside the wing. &nother method is to use 3$"*in! of high-velocity air tangent to the wing surface to reenergize the boundary layer and delay separation. &ir for tangential blowing is normally obtained as 3$eed ai from a )et engines compressor. :oth of these boundary layer control devices delay separation and stall to higher angles of attac$. Their lift curves loo$ similar to those for leading-edge devices shown in /igure !.'-. 5%amples of boundary layer suction and blowing are illustrated in /igure !.'+.

Wing with Leading-Edge Flap or Slat or Boundary Layer Control

CL

Basic Wing Section

+i!& e 4.11 Effe%t "f Leadin!2Ed!e +$a(# and 6"&nda - La-e C"nt "$ "n Lift C"effi%ient C& 'e#

P"*e ed Lift and Ve%t" ed Th &#t &n inte na$$- 3$"*n f$a( or >et f$a( has bleed air directed onto its leading edge and upper surface from the rear of the wing. The high-velocity air delays separation and increases lift. /igure !.'08a9 shows a typical internally blown flap configuration. 5ngine e%haust may also be used to increase or assist lift. /igure !.'0 shows three ways this may be done. The e%haust may be directed at the leading edge of a flap as on the ,c@onnell-@ouglas C-'-, or at the wing and flaps upper surface, as on the :oeing BC-'!. 2n either case, the vastly increased airflow over the flap increases lift. The engine nozzle may also be moveable to redirect or 'e%t" the engine e%haust downward. This re-orients the engine thrust vector so that it has a component in the lift direction to assist the lift generated by the wing. &lso note in /igure !.'0 the multiple slots in each /owler flap. everal high-lift devices are often used together on an aircraft. 5ach device adds to the total 51ma9. 2n some cases the devices complement each other so that the total increase in 51ma9 for several devices used together is greater than the sum of the 51ma9 increments for each device used alone.

<!

7a8Inte na$$- 6$"*n +$a(

738E:te na$$- 6$"*n +$a(

7%8U((e 2S& fa%e 6$"*in!

7d8Ve%t" ed Th &#t

+i!& e 4.14 +"& P"*e ed Lift C"nfi!& ati"n# 7Ada(ted f ", Refe en%e 18 4.4 WHOLE AIRCRA+T LI+T & complete aircraft will fre4uently generate significantly more lift than its wing alone. &n estimate of a whole aircrafts lift can be made by summing the lift contributions of its various components. The following is a simple method for ma$ing an initial estimate of an aircrafts lift. The method is suitable for use in the early conceptual phase of design. Win! C"nt i3&ti"n /or most aircraft, the ma)ority of the lift is generated by the wing. The finite wing lift prediction methods discussed in section !." give good initial estimates of wing lift curve slope, provided an

<*

appropriate value of e can be estimated. The results of e%tensive wind tunnel testing " of a vast variety of wing shapes suggest the following empirical e%pression for e6

e=

" " A0 + ! + A0 " 8' + tan " tma% 9

8!.'*9

where t ma9 is the sweep angle of the line connecting the point of ma%imum thic$ness on each airfoil of the wing. #ne effect of airfoil camber and wing twist on lift is to shift the zero-lift angle-of-attac$. & way to avoid the need for predicting zero-lift angle-of-attac$ early in the design process is to wor$ in terms of a3#"$&te an!$e2"f2atta%; 6

21.45

Aircraft #1 Design Parameters

a 7 - 17.

8!.'+9

WING VTAIL :ecause of the way a is defined, it always e4uals zero when lift is zero. Using absolute angle 2 of attac$ is AREA 680 FT 2 AREA (TRUE) 99 FT usually ade4uate for early conceptual design. SPAN 40 FT SPAN 21.45 FT ASPECT RATIO 2.4 ASPECT RATIO 2.4 methods. 5stimating wing ma%imum lift coefficient is difficult without more advanced analysis TAPER RATIO 0.07 TAPER RATIO 0.1 ?owever, a practical constraint of ta$eoffs and landings leads to a simple way to estimate the ma%imum SWEEP 47 DEG SWEEP 47 usable lift coefficient for LE those two phases of flight. /igure !.'<LE shows an aircraft with a t DEG i%-%$e DIHEDRAL 0 DIHEDRAL $andin! !ea on a runway. 1hen the aircraft accelerates to ta$eoff speed, it must "tate 40 to DEG the ta$eoff AIRFOIL NACA 64A204 airborne. AIRFOIL NACA 0004 angle of attac$ in order to generate enough lift to become The aircraft normally tips bac$ on its ,ain $andin! !ea as it INCIDENCE rotates. The amount which 0 the aircraft can rotate is limited by the tail stri$ing the ground. This limitation also applies to landing, will be at its landing angle of attac$ TWIST 0 since the aircraft LANDING GEAR when it touches down. /or many aircraft this angle is well below the wings AIN GEAR stall angle. Therefore, the ma%imum usable lift PROPULSION coefficient for ta$eoff or landing may be estimated as SI!E the wing lift curve slope, 51, TIRE "6#11 o ENGINES 2angle # G$%%&'( ) S*+,STRO.E 11of IN/igure !.'<. ama9 7 '* multipled by the ma%imum usable absolute of attac$, - 17., in the case F666/GLS/110 ROLLING RADIUS 14.7 IN 5 1ma9 = 5 1 ama9 = 5 1 8 ma9 1 = . 9 8!.'-9 A0 THRUST 281000 L2 NOSE GEAR IL THRUST 181450 L2 TIRE SI!E "0#7.7 /igure !.'< alsoDIA shows the pilots downward 'ie* an!$e over the nose. The ma%imum ETER "2 IN STRO.E 11 IN usable angle of attac$ of an aircraft may be limited, at least for landing, by the pilots visibility over LENGTH 1"" IN ROLLING RADIUS 12.7the IN nose. This visibility re4uirement is particularly important for aircraft which must land on an aircraft carrier. WEIGHT "1500 L2 This geometry constraint may also limit the ma%imum usable lift coefficient. o, as a result of limits on ma%imum rotation or ti(23a%; angles and pilot view angles, a good rule of thumb for a value for ma%imum usable angle of attac$ is about '* degrees.

!ilot "iew Angle Nose #ear $ain Landing #ear ip-Back Angle

+i!& e 4.15 Ti(26a%; An!$e and Pi$"t Vie* An!$e Hi!h2Lift De'i%e# &n appro%imate estimate for the effect of trailing-edge flaps on 51ma9 can be easily added to the wing 51ma9 prediction. ince most flaps change 17. but not 51 , their effect can be represented as an increment to the ma%imum usable absolute angle of attac$. /or flaps that span the entire wing, this increment in a is the same magnitude but of opposite sign as the increment in l7. in two-dimensional wind tunnel data for an airfoil with the flap system mounted on it. 2f flapped airfoil data is not available,

<+

the increment can be appro%imated by another rule of thumb. &ircraft often use partial e%tension of flaps for ta$eoff and full flaps for landing. &s a first appro%imation, a '.-degree increment in a for ta$eoff flap settings and '* degrees for landing flaps is acceptable. /or flaps which do not span the entire wing 8a much more common situation9, the increment in a is scaled by the ratio of f$a((ed a ea, 3f, to reference planform area , 3. 3f is the area of that part of the wing which has the flaps attached to it. /igure !.". depicts 3f 8shaded gray9 for a typical wing.

Are a

$ pe ap Fl

Fla ppe $

ea Ar

Flap

Flap

!0l0

+i!& e 4.)9 +$a((ed A ea and +$a( Hin!e Line S*ee( An!$e #nce 3f is determined, the change in ma%imum usable absolute angle of attac$, a, is given by6

a = a" -

3f 3

cos h.l .

8!.'09

where h .l . is the sweep angle of the flap hinge line, as shown in /igure !.".. 1ith a estimated, the ma%imum usable lift coefficient with flaps is appro%imated as6

51ma% 51ma% 8 no flap 9 + 51 a

8!.'<9

3ote that 8!.'<9 seems to disagree with the relationship between 51ma9 for flapped and unflapped wings presented in /igure !.'". The reason for the difference becomes apparent if the ma%imum usable angle of attac$ line is superimposed on the 51 vs curve, as shown on /igure !.".. &lthough this is not always the case, the situation depicted in /igure !."' is common. The aircrafts ma%imum usable angle of attac$ for ta$eoff and landing is significantly below it clean. 1hen flaps are deflected, the ma%imum usable angle of attac$ is still below stall with flaps, so the change in 51ma9 is correctly predicted by 8!.'<9.

<-

Wing with Flap

CL

Basic Wing Section

$a% &sa'le Angle of Attack

+i!& e 4.)1 Effe%t "f +$a(# and Ma:i,&, U#a3$e An!$e "f Atta%; "n

C &#a1

+&#e$a!e and St a;e# &n aircraft fuselage is usually relatively long and slender and therefore does not produce much lift. 2n the region of horizontal lifting surfaces, however, the lift being generated by those surfaces carries over onto the fuselage. This effect is modeled by treating the wing as if it e%tends all the way through the fuselage without any change in airfoil, sweep, or taper. 2n fact, the fuselage shape is significantly different from the wings airfoil shape and may be less effective at producing lift. ?owever, since the fuselage lifting area is generally larger than the portion of the wing in the fuselage, the two effects may be treated as canceling each other out, at least for early conceptual design. /or fuselages with stra$es or leading edge e%tensions, the effect should be included, even for a first estimate. /or angles of attac$ below '* degrees, the stra$e vorte% is not very strong, and e%tensive wind tunnel testing 3 has shown that the lift curve slope of the wing with stra$e may be modeled as6

51 8 with stra$e 9 = 51 8 without stra$e 9

3 + 3 strake 3

8!.".9

where 3strake includes only the e%posed surface area of the stra$e, not any portion inside the fuselage. ince 7 '* degrees is usually the ma%imum usable , 8!.".9 is ade4uate for the usable range. H" i?"nta$ Sta3i$i?e # and Cana d# The purposes of additional horizontal lifting and stabilizing surfaces on an aircraft will be discussed in Chapter +. /or a first estimate of the lift contributions of these surfaces, it is sufficient to treat them as additional wings. ?owever, the downwash created by the main wing will change the effective angle of attac$ of smaller horizontal surfaces in the wings. /igure !."" illustrates this effect. The figure also shows an &(*a#h field which increases the effective angle of attac$ of horizontal surfaces ahead of the wing. #f course, these smaller surfaces also create their own upwash and downwash. These upwash and downwash fields due to smaller surfaces will be ignored, because they are generally much wea$er than those of the main wing.

&pwash

Downwash

+i!& e 4.)) U(*a#h and D"*n*a#h To determine a horizontal surfaces contribution to the whole aircrafts lift curve slope, it is first necessary to determine the rate at which downwash 8or upwash as appropriate9, , changes with changing aircraft angle of attac$. 2n the e%treme case where the rate of change in the downwash angle e4uals the

<0

rate of change in angle of attac$ 8

surface in that downwash field is zero. That surface would ma$e no contribution to the whole aircrafts lift curve slope. 5stimates of the rate of change of downwash angle with angle of attac$ can be made using the following e,(i i%a$ 8based on testing rather than theory9 curve fit of wind tunnel " data6

= ' 9, the rate of change of the effective angle of attac$ of a

"'o 5 1 c a!g '. 3 : h = ' A0 ..-"* l h -

where6 ca!g is the mean geometric chord of the wing lh is the distance from the 4uarter chord point of the average chord of the main wing to the 4uarter average chord point on the horizontal surface, as shown in /igure !."3,

8!."'9

and :h is the vertical distance of the horizontal surface above the plane of the main wing, as shown in /igure !."3.
lc lh

."* croot

:h

."* croot

."* croot

+i!& e 4.). Ai ($ane Ge",et - f" D"*n*a#h P edi%ti"n

#nce

is predicted, the horizontal surfaces contribution to the aircrafts 51 is appro%imated as6

3t 5 1 8 due to horizontal tail 9 = 51 t ' 3
8!.""9

where the subscript t denotes parameters for the horizontal tail. Common 51 8 due to horizontal tail 9 values vary from almost zero to 3*( or more of

51 .

/or horizontal surfaces ahead of the wing, also $nown as %ana d#, the empirical e4uation " for predicting the rate of change of upwash with angle of attac$ for wings with ."5 C 3*o is6
( '..! ++ A0 u l = ( ..3 A0 ..3 ..33) c c
'. -

8!."39

<<

where u is the upwash angle and lc is the distance from the wings 4uarter chord to the canards 4uarter chord as shown in /igure !."3. #nce

u is estimated, the canards contribution to the aircrafts 51 is appro%imated as6 u 3 c 51 = 51 c 8!."!9 ' + 8 due to canard 9 3

where the subscript c identifies 4uantities related to the canard. Contributions of canards to the total aircraft lift curve slope are typically larger than those for horizontal tails. This is partly due to the canrd being in an upwash field rather than the downwash field surrounding most horizontal tails. 8#nce the contributions of canards and horizontal tails are estimated, the whole aircraft lift curve slope is given by6

51 8 whole aircraft 9 = 51 8 wing +

4./ WHOLE AIRCRA+T DRAG

od& +stra$e 9

+ 51 8 due to hori:ontal tail 9 + 51 8 due to canard 9

8!."*9

The drag polar for the complete aircraft is written somewhat differently than that for a wing alone. /or the whole aircraft, drag is identified as either (a a#ite d a! or d a! d&e t" $ift. The parasite drag is all drag on the aircraft when it is not generating lift. This includes both s$in friction and pressure drag, as well as several additional types of zero-lift drag which are associated with the complete aircraft configuration. The drag due to lift includes all types of drag which depend on the amount of lift the aircraft is producing. These include induced drag due to downwash, the pressure drag which increases with lift due to forward movement of the separation point, induced and pressure drag from canards and horizontal tails, and addition drag such as vorte% drag due to the leading-edge vortices on stra$es and highly swept wings. &ll of these types of drag may be appro%imated by the following simple e%pression for drag coefficient6

5- = 5- + k'51 + k"51
o

"

8!."+9

where6

k1 8 '6; eoA0<

8!."-9

and $" is chosen to allow modeling of wings with airfoils which generate minimum drag at some non-zero value of lift. 5-o is called the parasite drag coefficient. 2t represents all drag generated by the aircraft when it is not generating lift 8hence the Do subscript9. The variable eo in the e%pression for k' is called the O#*a$d@# effi%ien%- fa%t" . 2t is not the same as the span efficiency factor, e, used in 54uations 8!.<9, 8!.'!9 and 8!.'*9, because it includes all the other types of drag due to liftE. /or relatively straight wing 8F5 C 3. deg96 /or relatively swept wing 8F5 G 3. deg96
..'* eo = !.+'8' ...!* A0 ..+0 98cos 3' . 12 9

eo = '.-08' ...!* A0 ..+0 9 ..+!

2n order to model the common situation where minimum drag occurs at a positive value of lift coefficient, k" must be negative. This has the effect of shifting the entire 5- vs 51 curve to the right. /igure !."! illustrates this effect. The 51 for which 5- is a minimum is called 51min- .
E &lthough confusing, it is common to refer to the k'51" term in 8!."+9 as induced drag, though it is significantly different from the induced drag in 8!.<9

'..

CD

0.3

0.25 0.2 0.15 0.1 0.05 0 -1 -0.5 0

CD ( )*)//. , )*-. CL+ - )*)0 CL CD ( )*)+ , )*-. CL+

0.5
C & ,inD A 9..

1.5

CL

+i!& e 4.)4 E:a,($e "f D a! P"$a *ith Mini,&, D a! C"effi%ient at N"n2Be " Lift C"effi%ient Pa a#ite D a! Hust as lift predictions for the early stages of conceptual design rely heavily on the results of wind tunnel testing of similar configurations, so drag predictions rely heavily on drag data for similar types of aircraft. 2n later design stages, it is necessary to ma$e very precise predictions of the aircrafts drag, since )ust a '( difference in the drag at cruise conditions, for instance, can ma$e the difference between success and failure of a design. The methods used in ma$ing these precise predictions go far beyond the scope of this te%tboo$ and re4uire details of the design which are generally not available early in the conceptual design phase. 2t is important, however, to understand in a 4ualitative sense where the drag on an aircraft comes from. $in friction drag on a complete aircraft configuration is generally much greater than that on the wing alone, because the *etted a ea= 3wet, is greater. 1etted area of an aircraft is all the surface area over which air flows, and therefore to which the flowing air imparts shear stress. >ressure drag for the complete aircraft includes drag due to separation of the airflow around the aircraft fuselage, control surfaces, etc., in addition to the wing. Inte fe en%e d a! results from flow interactions between the various components of an aircraft which cause them to have more drag when assembled together than the sum of their drags when tested in a wind tunnel separately. ,iscellaneous drags include drag due to cooling air flowing through heat e%changers, air which lea$s through doors and fairings which dont fit perfectly and around moveable surfaces, plus the profile drag of antenae, gun barrels, sensors, etc. which protrude from the aircraft. The total of all these drags is the profile drag of the complete aircraft. To this must be added wave drag if the aircraft flies at or near the speed of sound. 1ave drag will be discussed in a later section. & very good initial estimate of subsonic parasite drag may be made from drag data for similar aircraft using the concept of an eC&i'a$ent #;in f i%ti"n d a! %"effi%ient , 5fe , which is defined as follows6

5 f e = 5 -o

3 3 wet

8!."09

Table !.' lists average 5fe values for several classes of aircraft. These values are based on historical data !,* for large numbers of each type of aircraft. 5fe is a function of such diverse factors as aircraft s$in materials and shapeI paintI typical flight =eynolds numbersI number of additional air scoops for ventilationI type, size, number, and location of engine air inletsI and attention to detail in sealing doors, control surface gaps, etc. 3aturally, these details vary significantly from aircraft to aircraft, but the data

'.'

in =eference * suggest that there is enough similarity among aircraft of a given class that useful average 5fe values can be established. Table !.' lists the most commonly used values of 5fe !,*. Ta3$e 4.1 C",,"n Cfe Va$&e# T-(e Het :omber and Civil Transport ,ilitary Het Transport &ir /orce Het /ighter Carrier-:ased 3avy Het /ighter upersonic Cruise &ircraft Fight ingle >ropeller &ircraft Fight Twin >ropeller &ircraft >ropeller eaplane Het eaplane C fe ....3. ....3* ....3* ....!. ...."* ....** ....!* ....+* ....!.

Using 5fe to predict 5-o for an aircraft which generates minimum drag when it is generating zero lift only re4uires selecting a 5fe for the appropriate category of aircraft and estimating the total wetted area of the aircraft concept. The value of 5-o is then obtained by solving 8!."096

5-o = 5 fe

3 wet 3

8!."<9

D a! D&e t" Lift >redicting drag due to lift must begin with predicting #swalds efficiency factor, eo. This is done with a curve fit of wind tunnel data " for a variety of wing and wing-body combinations. The e4uation for this curve fit is6

eo = !.+'8' ...!* A0 ..+0 98cos 12 9 ..'* 3' .

8!.3.9

3ote that increasing wing sweep tends to decrease the value of eo. &lso note that increasing A0 will tend to decrease eo. This is due to the fact that for high-aspect-ratio wings, that part of the airfoil profile drag which varies with lift is a larger part of the total drag due to lift which eo must model. Effe%t "f Ca,3e There are a number of reasons why an aircraft may generate its minimum drag at a positive 8nonzero9 value of lift coefficient. &s one e%ample, the profile drag on cambered airfoils is typically at a minimum at some small positive value of lift coefficient. &s another e%ample, the shape and orientation of an aircrafts fuselage may cause it to generate the least amount of drag at other than the zero-lift condition. 54uation 8!."+9 has an additional term, the k" 51 term, to model this effect. 2f, for instance, the minimum drag coefficient for an aircraft occurs at a lift coefficient signified by the symbol 51min-, then the necessary value of k" is given by6 k " = " k' 51min8!.3'9 The value of 51min- is determined by plotting the drag polar for the wing using actual airfoil data and 8!.39. 2f actual airfoil data is not available, as a crude appro%imation assume that the airfoil generates minimum drag when it is at zero angle of attac$, and that the effect of induced drag is to move 51min- to a value halfway between zero and the value of 51 when 7 .. The value of 51 when 7 . is given by6

'."

51
and6

=.

= 51 8 a 9 = 51 8 1=. 9 since a 7 - 180 and 7 .

8!.3"9

51

min-

1 = . = 51 "

8!.339

This value of 51min- is then used for the entire aircraft. This is done because it is assumed that the aircraft designer will design the fuselage, stra$es, etc. so that they also have their minimum drag at the angle of attac$ that puts the wing at its 51min-. 1hen this is done, the minimum value of 5-, which is given the symbol 5-min, must not be any lower than the 5-o predicted by 8!."09. =ecall that 5-o is the aircrafts zero-lift drag coefficient. /or aircraft with minimum drag at non-zero lift this leads to the following revised predictions6

5-min = 5 f e

3 wet 3
"

8!.3!9 8!.3*9

5-o = 5-min + k'51min4.0 MACH NUM6ER E++ECTS

The lift curve and drag polar predictions of ection !.* above are valid for relatively low speeds. &s with airfoils 8discussed in ection 3.+9, pressure changes around an aircraft are magnified by density changes at higher ,ach numbers. &s ,ach number increases to near unity and above, additional changes occur to the flow which have profound effects on the aircrafts lift and drag Consider an infinitesimally small body moving in the atmosphere. The body is ma$ing small pressure disturbances which are transmitted as sound waves. The bodys ,ach number indicates the relative speed between it and the sound waves it creates. 2f = 7 ., then the sound waves radiate outward in concentric circles from the body li$e ripples from the point where a stone lands in a pond. /igure !."*8a9 illustrates this situation.

8a9 = 7 .

8b9 = C '

8c9 = 7 '

8d9 = G '

+i!& e 4.)/ S"&nd Wa'e# Gene ated 3- a M"'in! 6"d2f the body is moving, then the sound waves upstream of it are closer together, because each successive wave is generated from a point further upstream, and the speed relative to the body at which each wave moves upstream is a > 4, since the body is moving the same direction as the wave. @ownstream of the body )ust the reverse is true. The spacing between the waves is greater and the waves are moving at a ? 4 relative to the body. The closer spacing of the waves upstream of the body causes the sound to have a higher fre4uency or pitch, while the sound downstream has a lower pitch. This is why the sound of an automobile horn or train whistle shifts to a lower fre4uency as the vehicle passes. The effect is called the D"(($e #hift. imilar shifts in the fre4uencies of reflected radio waves are the basis for radar speed detectors. The situation is illustrated in /igure !."*8b9.

'.3

The situation when = 8 ' is illustrated in /igure !."*8c9. 3ote that the body is moving at the same speed as the sound waves it emits, so all of the sound emitted by the body reaches a point ahead of it at the same time it does. The sound waves collect into a single pressure wave $nown as a Ma%h *a'e= which is perpendicular to the direction of movement of the body.. 1hen = G ', the ,ach wave trails bac$ from the body at an angle, as shown in /igure !.""8d9. &n e%pression for , the angle of the ,ach wave 8also $nown as the Ma%h an!$e9, may be derived from the relationship between the velocity of the body and the velocity at which the sound waves move out from their point of origin, as shown in /igure !."+.

+i!& e 4.)0 Ma%h Wa'e Ge",et #n the basis of the geometry of the /igure !."+, the e%pression for is6

= sin '

a ' = sin ' 4 =

8!.3+9

Sh"%; Wa'e# The pressure waves caused by a body moving through the air li$ewise influence the flowfield ahead of the body. Consider now a large body such as an aircraft or missile moving through the air. The influence of the high pressure at a stagnation point on the front of the body is transmitted upstream at the speed of sound, so that the flow slows down gradually rather than suddenly when encountering it. ?owever, as the speed of the body through the air e%ceeds the speed of the sound waves, this process of JwarningK the air ahead that the body is approaching becomes impossible. 2n such a situation, the pressure change occurs suddenly in a short distance. This sudden pressure change is called a #h"%; *a'e. &ir flowing through a shoc$ wave undergoes a rapid rise in pressure, density, and temperature, a rapid decrease in velocity, and a loss of total pressure. The angle of a shoc$ wave is usually different than the ,ach angle. 2t depends on the ,ach number and the angular change of the flow direction as it goes through the shoc$ wave. /igure !."- shows shoc$ waves around a model of the pace shuttle in the U &/ &cademys tri-sonic 8high subsonic, transonic, and supersonic9 wind tunnel. The waves are made visible by the bending of the light waves as they pass through the regions of rapidly changing air density.

'.!

+i!& e 4.)1 Sh"%; Wa'e# A "&nd a M"de$ "f the S(a%e Sh&tt$e at M A 1.1 and T*" Diffe ent An!$e# "f Atta%; in the USA+ A%ade,-@# T i2S"ni% Wind T&nne$ C iti%a$ Ma%h N&,3e hoc$ waves also occur around a body even when it is flying at speeds below the speed of sound. This happens because the air accelerates as it flows around the body. &n airfoil, for instance, may be moving at = 7 ..0 relative to the free stream, but it was shown in Chapter 3 that the shape of the airfoil causes the flow to be moving much faster over its upper surface. The local flow velocity over the upper surface of the airfoil may be greater than the speed of sound. This situation is described by saying the $"%a$ Ma%h n&,3e is greater than one 8= G '9, and the flow in this region is said to be #&(e #"ni%. The free stream ,ach number6

= =

83."09

at which the local ,ach number first e4uals unity is called the C iti%a$ Ma%h n&,3e , =crit. /igure !."08a9 illustrates this situation.
=8' Ter#inating S!oc/ =@' =G' =@' =G' Separate$ .a/e =8' Bow S!oc/ Obli2ue S!oc/ Obli2ue S!oc/s =G' =G' =G' =G'

=@'

8a9 = 7 =crit harp 3ose

8b9 =crit C

= C ' 8c9 = G ', :lunt 3ose

8d9

= G ',

+i!& e 4.)4 +$"*fie$d# at T an#"ni% and S&(e #"ni% S(eed#

= 7 =crit no shoc$ wave forms, because the local ,ach number only e4uals '.. at one point. &s = increases above =crit however, the region where = G ' grows. &s shown in /igure "08b9,
&t pressure waves from decelerating flow downstream of the supersonic region cant move upstream into that region, so they Jpile upK into a shoc$ wave. This shoc$ wave at the downstream end of the supersonic region is called a te ,inatin! #h"%; because it terminates the supersonic region and slows the flow abruptly to below the speed of sound. The strong adverse pressure gradient in the shoc$ wave which slows the supersonic flow also slows the flow in the boundary layer, and often causes it to separate. This phenomenon is called #h"%;2ind&%ed #e(a ati"n. 2t causes a significant increase in drag and decrease in

'.*

lift. The sudden rise in drag as = approaches ' was once thought to be an absolute barrier to higher speeds. 2t was called the #"&nd 3a ie . The ,ach number at which this rapid rise in drag occurs is called the d a! di'e !en%e Ma%h n&,3e , =--. &s = e%ceeds '.., another shoc$ wave forms a short distance in front of bodies with blunt or rounded leading edges. &s shown in /igure "08c9, air flowing through this shoc$ wave, called the 3"* #h"%;, is abruptly decelerated to = C '. The subsonic flow downstream of the bow shoc$ may accelerate again to be supersonic as it flows around the body, but it will e%ert a significantly lower pressure on the rear part of the body since it has lost so much total pressure. This low pressure on the rear of the body produces a great deal of pressure drag which is called *a'e d a!. The bow wave is perpendicular or normal 8it is also called a n" ,a$ #h"%;9 to the flow directly ahead of the body, but its angle to the flow becomes the same as the ,ach angle off to the sides of the bodys path. The terminating shoc$ moves to the trailing edge of the body, and no longer slows the flow to subsonic. 2f = is sufficiently greater than unity and the leading edge of the body is sharp, the bow shoc$ will touch the bodys point, as shown in /igure "08d9. The shoc$ is said to be atta%hed. 5%cept at the point of attachment, the flow no longer decelerates below = 7', but remains supersonic as it flows past the body. The shoc$ wave at the leading edge and the one at the trailing edge trail off at an angle which initially depends on the shape of the body. /urther from the body the shoc$ angles become the same as the ,ach angle. These shoc$ waves are referred to as "3$iC&e #h"%;#, because they are not perpendicular to the flow. The loss of velocity and total pressure in obli4ue shoc$s is less than for normal shoc$s. +$i!ht Re!i,e# The range of ,ach numbers at which aircraft fly is divided up into f$i!ht e!i,e# . The regimes are chosen based on the aerodynamic phenomena which occur at ,ach numbers within each regime, and on the types of analysis which must be used to predict the conse4uences of those phenomena. /igure !."< shows these regimes. ,ach numbers below =crit are grouped together as the #&3#"ni% flight regime. 1ithin this regime, compressibility effects are usually ignored for = C ..3.

Subsonic
'nco#, Co#pressible pressible

Transonic

Supersonic

3%personic

9..

Mcrit

1.9

).9 Mac! nu#ber( M

..9

4.9

/.9

+i!& e 4.)5 +$i!ht Re!i,e# /or freestream ,ach numbers greater than about '.3 8depending on aircraft shape9 the flow is entirely supersonic 8local ,ach number remains greater than ' everywhere in the flowfield9, as shown in /igure !."08d9. This is called the #&(e #"ni% flight regime. 3ote that while some local ,ach numbers in the flowfield are supersonic for speeds below the supersonic flight regime, an aircraft is only considered to be operating in the supersonic regime when all of the flow around it 8e%cluding flow entering the engine inlets9 remains supersonic. /reestream ,ach numbers above about *.. are considered h-(e #"ni%. The hypersonic flight regime is characterized by e%treme temperature changes and significant interactions between obli4ue shoc$ waves and the boundary layer. :etween the subsonic and the supersonic flight regimes lies the t an#"ni% regime. The transonic regime is characterized by a mi%ture of supersonic and subsonic flow, and in many cases there are also large areas of separation. /igure !."08b9 and 8c9 show two e%amples of this. Transonic flowfields are too comple% for accurate analysis by any but the most advanced methods, and the analysis often re4uires hours of computing time on the fastest supercomputers for a single flight condition. These methods are beyond the scope of this te%t.

'.+

Lift

&ll features of the lift curves of most aircraft vary with ,ach number. The most important of these effects is the change in 51. The relationship between 51 and cl does not change significantly with changes in subsonic ,ach number, so the >randtl-Llauert correction can be applied at subsonic speeds to 51 in the same way it was applied to cl in ection 3.+6

51 =
54uation 8!.3-9 is valid only for

51 = = .

' =

"

8!.3-9

= C =crit. &lso, the correction made by 8!.3-9 becomes trivial for

= C ..3. This fact is part of the basis for setting the dividing line between incompressible and compressible flow at = 7 ..3. /or the supersonic flight regime, if = G ';cos 12, the lift curve
slope is given by6

51 =

! = '
"

8!.309

:oth 8!.3-9 and 8!.309 yield an infinite value for lift curve slope at = 7 '... 2n fact, in the transonic regime, with so much shoc$-induced separation and comple% flowfields, 51 is difficult to predict. /or a well-designed supersonic aircraft, 51 levels off from the subsonic curve defined by 8!.3-9 and transitions smoothly to the supersonic curve of 8!.309. /igure !.3. illustrates a typical variation of 51 vs = .

Theoretical

&ift Coefficient Curve Slope( C&

Subsonic

Transonic

Supersonic

1.9

).9

Mac! 4u#ber( M

+i!& e 4..9 T-(i%a$ Va iati"n "f Lift C"effi%ient C& 'e S$"(e *ith Ma%h N&,3e 51ma9 and 180 also vary with ,ach number. 51ma9 initially increases due to compressibility effects and then decreases as shoc$-induced separation causes stall at lower angles of attac$. The zero-lift angle

'.-

of attac$ for cambered airfoils remains unchanged at subsonic speeds but becomes zero in the supersonic regime. /or early conceptual design, it is usually acceptable to ignore both these effects. D a! at Hi!h S&3#"ni% Ma%h N&,3e # @rag results from a comple% set of phenomena, and high ,ach numbers only add to the comple%ity. 2n the subsonic regime, the primary changes in 5-o and k' for a given aircraft are due to increasing =eynolds number as ,ach number increases. These changes are highly dependent on the relative importance of s$in friction drag and pressure drag for a particular aircraft configuration. /or many aircraft the changes are negligible. /or early conceptual design, it is fre4uently acceptable to assume 5-o and k1 do not vary with ,ach number below =crit. S&(e #"ni% Be "2Lift D a! /or the supersonic flight regime, wave drag is added to the other types of drag. Theoretical analyses and wind tunnel tests have shown that at supersonic speeds slender, pointed bodies whose crosssectional areas vary as shown in /igure !.3' have minimum wave drag for their size. These low-wavedrag shapes are $nown as Sea #2Haa%; 3"die# after the engineers + who initially studied them. The mathematical relationship for the area distribution which produces minimum wave drag is called the a ea &$e. The magnitude of the wave drag for these bodies varies as follows6

5-wa!e =

!.* Ama9 3 l

"

8!.3<9

where Ama9 is the ma%imum cross-sectional area of the body and l is its overall length. 2n order to achieve minimum wave drag for supersonic aircraft, designers strive to ma$e the cross-sectional areas of their designs vary li$e /igure !.3'. The process is called Japplying the area rule D or )ust Jarea ruling.K &rea ruling may re4uire reducing the area of the fuselage where the wing is attached to avoid a bump in the area plot. The result is a Jwasp waistK as is seen on such aircraft as the T-30 and /-'.+. /igure !.3" illustrates an e%ample of this, while /igure !.33 illustrates the T-30s wasp waist.

Cross-Sectional Area1 A2A3 a%

1.2 1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 A%ial Distance1 %2l

+i!& e 4..1 Sea 2Haa%; 6"d- A ea Di#t i3&ti"n

'.0

A ea Di#t i3&ti"n. A,a: A 4/ ft)

A ea Di#t i3&ti"n. A,a: A 41 ft)

.asp .aist

3ote6 :oth aircraft have the same internal volume

+i!& e 4..) The A ea R&$e A(($ied t" a S&(e #"ni% +i!hte Ai % aft

+i!& e 4... A ea R&$in! "f the T2.4 +&#e$a!e 76 andt %"$$e%ti"n8 5-wa!e for aircraft with reasonably smooth area distributions that conform appro%imately to /igure !.3' can be predicted using the following modification of 8!.3<96

5-wa!e
where6

!.* Ama9 = 2.- ( ..-! + ..3- cos 12 ) ' .3 = = 5-o ma9 3 l

"

8!.!.9

= 5-o ma% =

' cos
.."

12

8!.!'9

estimates the ,ach number where the ma%imum value of 5-o occurs. 54uation 8!.!.9 is only valid for = = 5-oma9 . 2.- is an empirical 8based on e%perimental data9 *a'e d a! effi%ien%- (a a,ete . 2t is a measure of how closely the area distribution for the aircraft appro%imates the smooth curve of /igure !.3', and how free the aircraft is of additional sources of wave drag 8antenae, lea$s, bulges, wing-body )unctions, engine inlets, etc.9 The magnitude of 2.- averages about ".. for typical supersonic aircraft. The modifications in 8!.!.9 are based on curve fits of wind tunnel and flight test data ". &ccurate estimation of the variation of 5-o through the transonic regime is e%tremely difficult. &s a simple appro%imation, a straight line is drawn between the subsonic 5-o at =crit and the 5-o

'.<

-o predicted at by adding the 5-wa!e from 8!.!.9 to the subsonic 5-o. The resulting error is typically acceptable for early conceptual design, provided the aircraft will not cruise in the transonic regime.

=5

ma9

=crit is determined either by the shape of the fuselage or the shape of the wing, depending on which component creates the fastest velocities in the air flowing around it. /or an unswept wing, the airfoil shape, especially its ma%imum thic$ness-to-chord ratio, determines how much the air accelerates as it flows around the wing, and therefore how high 4 can be before = 7 ' somewhere in the flowfield. &irfoil designers may e%pend considerable effort carefully shaping an airfoil to ma$e its =crit as high as possible, and to delay the development of strong shoc$ waves above =crit. 1hen actual airfoil data is not available, the following curve fit of =crit data for 3&C& +!- series airfoils may be used6

= crit

t 7 '.. - ...+* '.. ma9 c

. .+

8!.!"9

where tma9 is the airfoilss ma%imum thic$ness. &t this point it is interesting to remember the discussion in Chapter 3 about the advantages a thic$er airfoil with a larger leading edge radius gave to the /o$$er @M22 in 1orld 1ar 2. Thic$ airfoils were popular on all subse4uent types of aircraft until after the start of 1orld 1ar 22, when the fastest piston-engined fighters began reaching or e%ceeding the critical ,ach number for their wings airfoil. evere control difficulties fre4uently resulted. The phenomenon was not well understood at the time, but it was observed that fighter planes with thinner airfoils could fly faster before encountering the problem. Two similar fighter aircraft produced by the same company e%emplify the effect. The ?aw$er Typhoon and Tempest fighters had the same engine and fuselage, but the Typhoon had a smaller wing with a greater thic$ness-to-chord ratio. The Tempest had a ma%imum speed which was *.N $nots faster than the Typhoon, though it was nearly identical e%cept for its wing. &s ma%imum speeds of fighter aircraft have continued to increase, their airfoils have gotten progressively thinner, so that the thin, highly cambered airfoil sections of the outer wing panel of the /-'* are similar 8though far from identical9 to the airfoils of the 1orld 1ar 2 opwith CamelO Effe%t "f Win! S*ee( 2n addition to reducing airfoil thic$ness, aircraft designers can also raise a wings =crit by sweeping it either forward or aft. To understand how this wor$s, consider the untapered, swept wing in /igure !.3!. weeping the wing without changing its shape increases the effective chord length. /igure !.3! shows why this is true.

'm

'm

4
o 12 7 !*

+i!& e 4..4 The Effe%t "f Win! S*ee( "n St ea,*i#e Thi%;ne##2t"2Ch" d Rati" Chord is measured in the streamwise direction, since the airfoil shape the air must flow around is a streamwise slice of the wing. /rom the geometry of /igure !.3!, the relationship between the chord of the unswept wing and the chord of the swept wing is6

''.

c ;swept wing< 7 c ;unswept wing< ; cos 12 so that6

8!.!39

t t ma9 ma9 7 8cos 12 9 c 8 swept wing 9 c 8 unswept wing 9

8!.!!9 ubstituting the swept wing chord into 8!.!"9 yields an e%pression for critical ,ach number for swept wings6 =crit 7 '.. - ...+* cos..+ 12 '.. or, in terms of the unswept wings =crit6 =crit 7 '.. - cos..+ 12 8'.. - =crit ;unswept<9 8!.!*9

t ma9 c

.. +

/or tapered wings, the effect is modeled by using ."*c, the sweep angle of the line connecting the 4uarter chord points of the wings airfoils, and using the ma%imum value of

t ma9 on the wing6 c

8!.!+9

=crit 7 '.. - cos..+ ."5c 8'.. - =crit ;unswept< 9

+&#e$a!e C"nt i3&ti"n #nly fuselages with relatively blunt noses will produce a value for =crit which is lower than the one determined by the shape of the wing. Lenerally, ensuring the fuselage has a long pointed nose, so that the fuselage reaches its ma%imum area at least + fuselage diameters downstream of the point of the nose will ensure =crit due to the fuselage is higher than =crit due to the wings. The value for =crit for the entire aircraft will be the lowest of the two. The above methods give reasonably accurate predictions for 5-o for a wide variety of e%isting supersonic aircraft. /igure !.3* illustrates the actual variation of 5-o with ,ach number for the Convair /-'.+ @elta @art supersonic fighter, along with 5-o values predicted using the methods )ust described. /igure !.3* is a good e%ample of the type and magnitude of error which can be e%pected when appro%imating the transonic variation of 5-o with a straight line. S&(e #"ni% D a! D&e t" Lift &t supersonic speeds, all airfoils, regardless of shape, generate zero lift at zero angle of attac$. >ractical supersonic airfoil shapes also generate minimum drag at zero lift and zero angle of attac$, so in the supersonic regime, k" 7 .. The supersonic value of k' is given by6
k' =

( ! A0

A0 = " '
"

= ' "

cos 12

8!.!-9

/or well-designed supersonic aircraft, the transition from subsonic to supersonic values of k' and k" is gradual, so that the variation of these prameters through the transonic regime can be appro%imated with a smooth curve.

'''

0.025
C Do

0.02 !arasite Drag Coefficient1

0.015 Actual 0.01 Predicted

0.005

0 0 0.5 1 1.5 2 2.5 $ach Nu3'er1 M

+i!& e 4../ Va iati"n "f A%t&a$ and P edi%ted C"o *ith Ma%h N&,3e f" the +2190 T"ta$ D a! 2n summary, the total drag on an aircraft is the sum of profile drag 8the subsonic drag not due to lift9, wave drag, and drag due to lift or induced drag6 5-o 8 5-p ? 5-wa!e and 5- 8 5-o ? k1 51" ? k" 51 8!.!09

/igure !.3+ shows how these vary with ,ach number for a typical supersonic aircraft.

CD"

;1

Wa'e D a!

P "fi$e D a!

M% it

MDD

+i!& e 4..0 Va iati"n "f C"o and /1 *ith Ma%h N&,3e De#i!n C"n#ide ati"n# The foregoing discussion of the methods used to predict an aircrafts 51 , 51ma9 , and drag polar carries with it the basis for insight into how to ma$e wise design choices. 54uations 8!.'*9, 8!.3.9, 8!.!.9, and 8!.!+9 ma$e it clear that increasing wing sweep raises =crit and reduces wave drag, but degrades 51 and 51ma9, , and increases induced drag. 54uation 8!.!.9 emphasizes that the ratio of ma%imum cross-sectional area of an aircraft to its length has a larger effect on the aircrafts supersonic wave drag than does wing sweep or the JsmoothnessK parameter, 2.- . 2ncreasing aspect ratio has

''"

beneficial effects on induced drag at both subsonic and supersonic speeds. ?owever, it will be shown in Chapter - that using high-aspect-ratio wings in supersonic aircraft is impractical, because the structure of such wings would be far too heavy. & further consideration for the aspect ratio and sweep of wings for supersonic aircraft is the benefit to be gained by $eeping the wing inside the shoc$ wave cone generated by the aircrafts nose, as shown in /igure !.3-. This practice reduces the aircrafts wave drag because the ,ach number inside the cone is lower than = , and shoc$ waves are wea$er than they would be if the wing were e%posed to

=.

S!oc/ Cone

+i!& e 4..1 Win!# Sta- Within Sh"%; C"ne at De#i!n Ma%h N&,3e

4.1 WHOLE AIRCRA+T ANALESIS E<AMPLE Lift /igure !.30 shows a drawing of an /-'+ with the lifting surfaces and high-lift devices labeled. The /-'+ uses a 3&C& +!&-".! airfoil, which has its ma%imum thic$ness at *.( chord. The sweep angle of the line connecting the ma%imum thic$ness points of the airfoils, tma9 7 ".o. The flapped area for the trailing-edge flaps, as defined in /igure !."., is appro%imately '*. ft ". The flap hinge-line sweep angle, '1 7 '.o. The aspect ratios of the wing and horizontal tail are6

3. " A0 = = =3 , 3 3..

"

'0 " A0t = = =3 3 t '.0

t

"

''3

" .)/% A .9

49" Flap 49"

St 5879ft+

Stra/es

S 5 677 ft+

.9 ft. 14 ft.

S stra/es A )9 ft +

Flap

" tF% ,a: A )4

+i!& e 4..4 +210 Liftin! S& fa%e#. Then, using 8!.'*9 to estimate e 8the same for both surfaces since they have the same A096

e=

" " A0 + ! + A0 8' + tan t ma% 9

" "

" " 3 + ! + <8' + tan " "! o 9

7 .-.3 7 et .

The two-dimensional lift curve slope for the 3&C& +! &-".! is appro%imately ..' per degree, so:

51 =

cl '+

*-.3 cl 7 ...*3+ ;o 7

51

8for this uni4ue situation where A0t 8 A0 9

e A0

:ut this is before the effect of the stra$es on the wing is included. /or this6

51 8 with stra$e 9 = 51 8 without stra$e 9

3 + 3 strake 3.. + ". 7 8...*3+ ;o9 7 ...*-" ;o 3.. 3

The distance from the 4uarter chord of the main wings mean chord to the same point on the /-'+s horizontal tail, lh 7 '!.- ftI the wing taper ratio, = 3.*ft;'+.*ft = 0.21; and the distance of the horizontal tail below the plane of the wing, :h , averages slightly less than one footI so using 8!."'96

''!

o "' 5 1 c a!g '. 3 : h = ' 7 A0 ..-"* l h -

("') . *-"; ) '.ft'. 38."'9' 'ft

o o .-"*

3 '!.-ft - 3.ft

7 ..!0, and6

51

8 whole aircraft 9

= 51

8 with stra$e 9

N 51 t '

3 t 7 ...*-" ; o N ...*3+ ; o 8'-.!09 8'.0;3..9 7 . 3

.+- ;o

2t is interesting to compare these predictions with the /-'+s actual lift coefficient curve slope of ..+* ; o . @espite the comple%ity of the /-'+ configuration, the analysis result agrees 4uite well with the actual slope. The method does not achieve this degree of accuracy in every case, but its predictions are 4uite good for a wide variety of aircraft configurations. /igure !.3< compares actual aircraft lift coefficient curve slopes with slopes predicted by the method )ust described.

0.1

100 0 80 70 !ercent Error 60 50 40 30 20 10 0 !-16A "-5A !-106A #211 $-747 %-38A &ear'et !-15" Aircraft ype Predicted Actual ( )rr*r

0.0 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

+i!& e 4..5 P edi%ted and A%t&a$ Lift C"effi%ient C& 'e S$"(e# f" Se'e a$ Ai % aft #nce the lift curve slope is $nown, 51ma9 for ta$eoff and landing can be calculated6

Lift Coefficient Cur4e Slope1 C per degree

a = a" a = a" 3

3f 3

cos h.l . 7 !.<o, so 51ma9 7 ...+-;o 8'! o N!.< o9 7 '."- for ta$eoff

3f

cos h.l . 7 -.3+o, so 51ma9 7 .. .+- ;o 8'! o N-.3+ o9 7 '.!3 for landing

ince the /-'+ is e4uipped with plain flaperons which deflect a ma%imum of ". o for both ta$eoff and landing, rather than slotted flaps which deflect much further, it is no surprise that its actual usable 51ma9 for ta$eoff and landing is only '.", somewhat less than predicted. Pa a#ite D a! The first step in determining the /-'+s parasite drag is to estimate its wetted area. This can be done in a variety of ways. 2f the aircraft is drawn accurately on a %",(&te 2aided d aftin! 8C&@9 system, a function is probably available within the system to determine 3wet. & reasonably accurate

''*

estimate can be obtained with much less effort than is re4uired to ma$e a detailed drawing, however, by appro%imating the aircraft as a set of simple shapes as shown in /igure !.!.. The e4uations for the surface areas of these simple shapes are well $nown, and by ta$ing some care to avoid counting areas where two shapes touch, it is relatively easy to determine 3wet as shown in Table !.". 3ote that cross sectional shapes of the items labeled JcylindersK and Jhalf cylindersK in /igure !.!. may be circular, elliptical, rectangular, or any other shape whose perimeter is easily determined. The surface areas of these cylinders dont include the areas of the ends, since these generally butt up against another cylinder and are not wetted. 1hen the longitudinal flat face of a half cylinder or half cone touches another body, twice the surface area of that face must be subtracted, since it and an e4ual area of the other body are in contact with each other, and therefore not wetted. 2t is interesting to note that the actual wetted area of the /-'+ is '!<* ft", about *( more than was estimated by this simple model.
S& fa%e G1 S& fa%e G. C-$inde G1 S& fa%e G/ C"ne G1 C-$inde G) Ha$f C-$inde G1

Ha$f C"ne G1 Ha$f C-$inde G. Ha$f C"ne G) S& fa%e G0 Ha$f C-$inde G) S& fa%e G4 S& fa%e G)

S& fa%e G4

S& fa%e G1

Ha$f C-$inde G4

+i!& e 4.49 +210 Ge",et - A(( ":i,ated 3- Si,($e Sha(e#

1ith a value for 3wet in hand, and choosing 5fe 7 ....3* for a )et fighter from Table !.'6

5-min = 5 f e

3 wet 7 ....3* 8'!3.ft";3..ft"9 7 ..'+3

''+

Ind&%ed D a! The value of the #swalds 5fficiency /actor, eo , is estimated by 8!.3.9 as6

eo = !.+' 8' ...!* A0 ..+0 9 8cos 129..'* - 3.' 7 .<.+

so6 k1 8 '6; eoA0< 8 ..''Ta3$e 4.) +210 Wetted A ea E#ti,ati"n S& fa%e 'P" 3P! *P+ 0 < P '. C-$inde ' C"neH ' " Ha$f C-$ 'P" 3 ! Ha$f C"ne 1 ) S(an= ft '" + " '.! '.* Len!th= ft 3< Len!th= ft + ! Len!th= ft "! * 3. Len!th= ft " ! croot= ft '! -.0 <.+ '".* 0 * Hei!ht= ft ".* Hei!ht= ft ".* + Hei!ht= ft .0 " ".* Hei!ht= ft " " ctip= ft 3.* " . + 3 3 Width= ft * Width= ft * + Width= ft ' " * Width= ft " " !)= w) ., . !, ! t#a1*c ..! ..! ..+ .'. ..+ ..3 EC&ati"n 8'9 8'9 8'9 8'9 8'9 8'9 EC&ati"n 8"9 EC&ati"n 839 839 EC&ati"n *.( of 8"9 *.( of 8"9 *.( of 8"9 EC&ati"n *.( of 839 *.( of 839 Swet !'<.* ''-.* 30.+ "+.3 --.3 "3.< Swet **'.3 Swet !".! +".0 Swet +-.< '*."'".' Swet 3.' +.3 2 Sintersections . . . . . . 2 Sintersections . 2 Sintersections . . 2 Sintersections 30.! '. '0. 2 Sintersections " ! T"ta$ Swet
E Cone and;or Truncated Cone 8Tapered Cyllinder9 54uations6 8'9 1ing or tabilizing urface6 3 wet = 3 e9posed '.<-- + ..*"( t c)

net Swet !'<.* ''-.* 30.+ "+.3 --.3 "3.< net Swet **'.3 net Swet !".! +".0 net Swet "<.* *.3".' net Swet '.' ".3 '!'0 ft"

area of the surface e%cluding any portion which is inside another component, 8i.e. not e%posed9. 3ote that the surfaces for this e%ample were all drawn so that they were not inside any other component. 8"9 Cylinder ides6

where 3e9posed is the planform

h + w 3 wet = l for elliptical cross sections and 3wet 7 " l 8 h ? w9 for "

rectangular cross sections, where l is the length of the fuselage segment, h is its height and w is its width. The surface areas of the ends of the cyllinders are not included because they butt up against the ends of other cyllinders or cones. Typically, the only case where the end of a cyllinder does not butt up against another surface is at the end of a )et engine e%haust nozzle, where there is also no surface area. ome fuselages and e%ternal fuel tan$s also terminate without tapering to a point, but this is generally done because flow separation has already occurred ahead of this position on the body. The e4uivalent s$in friction coefficient method accounts for this design decision more accurately if the area of the end of the cyllinder is not included in the aircraft wetted area.

''-

839 Cone and Truncated Cone ides6 of the

h + w' + h" + w" 3 wet = l ' for elliptical and circular !

cross sections and 3wet 7 l 8 h' ? w' ? h" ? w"9 for rectangular cross sections, , where l is the length fuselage segment, h' is the height of the front end, w' is the width of the front end, h" is the height of the aft end of the segment and w" is its width. The surface areas of the ends of the cones and truncated cones are not included for the same reasons given for cyllinders.

The /-'+s average chord is c = A0 = 3. ft 3 ft = '. ft . /or standard sea level conditions and = 7 ..", 0e = 4c = ( ..."3-- slug ; ft 3 ) ( .." '''+.! ft ; s)( '. ft ) ; ( ..3-3- '. + slug ; ft s) ='!,"..,... . Using 8!.'.9 and airfoil data for the 3&C& +! &-".! 80e 7 <,...,...9, the following table of drag coefficient data for the wing alone is generated6 Ta3$e 4.. +210 Win! A$"ne D a! C"effi%ient Va iati"n 51 -.." -..' ... ..' .." from which 51mincd ...+" ...+ ...*3 ...!* ...!" k1 51" ...!...'" . ...'" ...!5- 8 cd N k1 51" ..'.< ...-" ...*3 ...*...0<

appro%imately the same 51 where the wing alone has its minimum drag6

Q =

...!. &ssuming a well-designed aircraft will have its minimum drag at

5-o = 5-min + k '51min- 7 ...'+- N ..''- 8..!9" 7 ..'+<

"

k " = " k' 51min- 7 -" 8.''-9 8..!9 7 -....<!

5- 7 ...'+< N .''- 51" - ....<! 51 S&(e #"ni% D a! The above analysis is valid for ,ach numbers less than =crit . Using the methods described in ection !.+ to predict =crit , 5-wa!e and k' 6 =crit 8unswept9 7 '.. - ...+* '..

t ma9 c

..+

7 '.. - ...+* 8!9..+ 7 ..0*

=crit 7 '.. - cos..+ ."5c 8'.. - =crit ;unswept< 9 7 '.. - cos..+ 3.o 8' - ..0*9 7 ..0+*

= 5-o ma9 =
"

' cos
.."

12

' cos
.."

!. o

= '.* .

5-

wa!e

!.* Ama9 = 2.- ( ..-! + ..3- cos 12 ) ' .3 = = 5-o ma9 3 l

The geometry built from simple shapes yields Ama9 7 "*.* ft" and l 7 !0.* ft, so6

''0

Ma%h n&,3e '..* '.* ".. &nd finally, the supersonic k' values6
k' =

C"wave .."+' .."'3 ..'0<

( ! A0

A0 = " '
"

= ' "

cos 12 /1 .'"0 ."*" .3+-

Ma%h n&,3e '..* '.* "..

Ta3$e 4.4 +210 D a! P"$a P edi%ted U#in! the Meth"d# "f Cha(te 4 Ma%h n&,3e ..3 ..0+ '..* '.* ".. C"o ..'+< ..'+< ..!"0 ..30. ..3*+ /1 .''.''.'"0 ."*" .3+/) -...<! -...<! -...!. .

#nce again it is interesting to compare these results with actual values for the /-'+6 Ta3$e 4./ A%t&a$ +210 D a! P"$a Ma%h n&,3e ..3 ..0* '..* '.* ".. C"o ..'<3 .."." ..!!! ..!!0 ..!*0 /1 .''.''* .'+. ."0. .3-. /) -...-...! -...' . .

The predicted values agree reasonably well with the actual values. The '"( lower subsonic 5-o values were to be e%pected in part because of the *( lower estimate of wetted area. & more accurate model of the aircraft made from more, smaller simple shapes could be e%pected to produce better 3wet and 5-o estimates. &lso, the fi%ed geometry of the /-'+s air inlet produces a great deal of additional wave drag and flow separation at high supersonic ,ach numbers. This e%plains why the /-'+s 5-o values actually increase at higher ,ach numbers when the model suggests they should decrease. 2t is best to view the drag values predicted by these methods as goals which can be achieved with careful design. The /-'+ is optimized for subsonic maneuvering, not supersonic cruise, so it should not be e%pected to achieve the lowest possible supersonic drag. #n the other hand, as shown by /igure !.3*, the methods predict very accurately the wave drag of the /-'.+, an aircraft optimized for supersonic flight. Ae "DENAMIC The computer program called &ero@B3&,2C which accompanies this te%t contains an aircraft design module which allows the user to appro%imate the shape of a conceptual aircraft design using simple shapes, in the same manner as was done in earlier in this section. The shapes are described by parameters such as height, length, width, span, thic$ness, etc. which are entered into a spreadsheet.

''<

&ero@B3&,2C draws the aircraft as the parameters are entered, and then performs an aerodynamic analysis similar to the e%ample in this section. This is all accomplished very rapidly, relieving the designer of many tedious calculations and drawing manipulations. The user is then free to e%plore many alternative design choices and to optimize the design. imilar, though much more capable computeraided design systems are used by virtually all modern aircraft designers. /igure !.!' shows an analysis screen from &ero@B3&,2C. 3ote that many more cylinders and cones have been used to appro%imate the /-'+ shape than in /igure !.!.. This gives the analysis greater fidelity.

+i!& e 4.41 Ae "DENAMIC Ana$-#i# "f the +210

RE+ERENCES '. 3icolai, F. ,., ,undamentals of Aircraft -esign , ,5T , 2nc. Renia, #?, '<-* ". /in$, =.@., A3A, 3ta ilit& and 5ontrol -AT5(=, 1right->atterson &/:, #?, '<-*. 3. Famar, H.5., and 3.T. /rin$, J&erodynamic /eatures of @esigned tra$e-1ing Configurations,K Bournal of Aircraft , Mol. '<, 3o. 0, &ugust '<0", pp +3<-+!+. !. =aymer, @. >., Aircraft -esign: A 5onceptual Approach , &2&& 5ducation eries, 1ashington, @.C., '<0< *. =os$am, H., Airplane -esign Cart 4I, =os$am &viation and 5ngineering Corp., #ttawa, S '<<.

'".

+. ears, 1.=., J#n >ro)ectiles of ,inimum 1ave @rag,K Duarterl& of Applied =athematics, Mol. !, 3o. !, Han '<!CHAPTER 4 HOMEWORI S-nthe#i# P "3$e,# -!.' &n aircraft design e%hibits e%cessive transonic wave drag and an unacceptably low =crit. :rainstorm five ways to change the design to improve these characteristics. -!." &n aircraft design has e%cessive induced drag for its high-altitude, low-speed surveillance mission. :rainstorm five ways to reduce its induced drag.

-!.3 & )et fighter aircraft cannot reach its re4uired ma%imum speed when carrying its air-to-air missile armament on underwing pylons and wingtip launch rails. Bou believe e%cessive parasite drag is the problem. :rainstorm at least five ways to allow the plane to carry its weapons with less parasite drag.

-!.! & large )et transport design is unable to achieve the value of 51ma9 it needs for acceptable ta$eoff and landing performance. :rainstorm five ways to increase its 51ma9.

-!.* & )et fighter aircraft design concept achieves its ma%imum lift coefficient at an angle of attac$ of "-o, but its ma%imum usable angle of attac$ for ta$eoff and landing is only '* o. :rainstorm at least * design changes which would allow the aircraft to use more of its ma%imum lift capability for ta$eoff and landing.

Ana$-#i# P "3$e,# &-!.' a. Consider a flying wing aircraft made using a 3&C& "!'" airfoil with a wing area of "*. ft " , a wing span of *. ft, and a span efficiency factor of ..<. 2f the aircraft is flying at a + angle of attac$ and a =eynolds number of appro%imately < % '. +, what are 51 and 5- for the flying wingA b. 2f the flying wing is flying at sea level at e%periencingA

4 7 "0. ft;s, how much lift and drag is it

&-!." @efine critical ,ach number.

&-!.3 1hat are two conse4uences when we have a wing 83-@9 instead of an airfoil 8"-@9A

&-!.! 2nduced drag is also called drag due to to A

. 1hen lift is zero, the induced drag is e4ual

&-!.* 1hen would induced drag be more prevalent6 during high speed flight 8such as cruise9 or low speed flight 8such as landing9A 1hyA

'"'

&-!.+ a. @raw a s$etch of a typical 51 versus angle of attac$ curve and a 5- vs 51 curve and show how increasing camber 8by putting a flap down9 would change these curves.

b. @raw the same curves and show how using boundary layer control or leading-edge devices would change these curves.

c. @raw the same curves and show how sweeping the wings bac$ 8as on the /-''' or /-'!9 would change the curves.

&-!.- & straight wing with a critical ,ach number of ..+* is swept bac$ 3*T. 1hat is its critical ,ach number in the swept configurationA

&-!.0 @raw a s$etch of a typical 5-o vs ,ach number curve for a )et fighter aircraft and show how sweeping the wings bac$ 8as on the /-''' or /-'!9 would change the curve. Fabel =crit and =-- on the curves.

&-!.< &n aircraft with 5-o 7 ...", k' 7 ..'", and k" 7 . is flying at = 7 ..0 at h 7 3.,... ft. 2f the aircraft has a wing area of 3-* ft" and is generating "*,... lb of lift, what is its drag coefficient and how much drag is it generatingA

&-!.'. &n aircraft with 5-o 7 ...'", k' 7 ..'0, and k" 7 . is flying with 51 7 .."+. The pilot attempts to temporarily reduce the drag on the aircraft by reducing 51 to zero. :y what percent will the drag be reduced in this situationA 1hy would it only be possible to do this temporarilyA

&-!.'' Fift-to-drag ratio is a measure of the efficiency of an aircraft. 1hat is the lift-to-drag ratio of an aircraft with 5-o 7 ...'0, k' 7 ..'3, and k" 7 -....< operating at 51 7 ..!A

'""