You are on page 1of 124

# CollegeofEngineeringandComputerScience

## MEC430 Machine Design

CHAPTER04THEORIES OF FAILURE

CollegeofEngineering&ComputerScience CECS

## MEC 430 Machine Design

Ch05FailureTheories

Introduction
WhatisFailure
SimpleTensionvs.Multiaxial Failure

BrittleFailureTheories
MaximumNormalStressTheory
MohrCoulombTheory

DuctilefailureTheories
MaximumShearStressTheory

DistortionEnergyTheory

ExperimentalverificationsandComparisons
ExperimentalStudies

Comparisons&Conclusions

Variousfailuretheories

Example

## Many theories have been put forth . some agree

reasonably well with test data, some do not

Warm-up Quiz
Acriticalsectioninashaftwithcircularsolidcrosssectionissubjectedto
torque,T,bendingmoment,MandnormalforceN.
Sketchthestressdistributiononthesection(useappropriatesymbolstomatch
withgivenaxes)andidentifycriticalpoint(s)fordesign.

## Multiaxial and Simple Tension Failure

xx

xx
Fails(Yields)when:
xx Sy

Fails(Yields)when:
F(xx ,yy ,zz ,xy ,xy,xy)f(Sy,Su,)

## Brittle Material Failure Summary

Mohr-Coulomb Theory (Internal Friction Theory)

Brittlematerialstypicallyhavesignificantlydifferentcompressiveandtensilestrengths.
TheInternalFriction(MohrCoulomb)Theorymaybeusedtoestimatethefailure
state.
ForsomematerialstheMohrCoulombTheorymayprovideaslightlymoreaccurate
estimate.

## Mohr-Coulomb vs Max Shear Stress Theory

PLANESTRESSCASE
Wheneverthestressstateiswithinthe
polygon,thematerialwillnotfail.
NotethattheInternalFrictiontheory(IFT)isageneralizationofthe
MaximumShearStressTheory(MSST).TheMaximumMSSTislimitedto
materialsinwhichthetensileandcompressiveyieldstrengthsare
approximatelyequal.

## Yield Functions for Slightly Porous Materials

Gadala et. al (1980) postulated that the
yield function for slightly porous
materials depends only on the porosity
of the material and the material yield
strength in tension.
The theory is tested and validated for
specific range of material porosity, :
0.95
0.60

Schematic presentation of
the yield surface

Postulated by:
Gadala et. al (1980) for slightly porous material that were initially ductile and for

## Yield Functions for Slightly Porous Materials (Cont.)

, and dimensionless constants dependent on the material porosity, . The
above equation requires testing for only one material property, the yield strength
while the other parameters , and are only function of
in tension
porosity. Based on finite element simulation of the problem, Gadala et al.
suggested the following values for these parameters , and as function of
the material porosity, , for 0.95
0.60:
25

22 , 10

0.51

0.417

10

1.46

1.65

## Maximum Shear Stress Theory or Tresca Criterion:

Relation Between Yield in Tension and Shear

## Maximum Shear Stress Theory or Tresca Criterion:

Shear Equations & Graphical Presentation

Special Cases
Biaxial Tension

Special Cases

## 2D Case & Graphical Representation

Example
Consider a thin-walled steel cylinder vessel shown in Figure (a). The ends of the cylinder
are closed with spherical caps, and it is subjected to an internal pressure p. Determine
the pressure under which vessel begins to yield. The material is made of SAE-1018 annealed
steel with yield strength Sy 225 MPa. (Note: consider only the vessel wall stresses).

t=6.35 mm
D=508 mm

## Thin-walled cylindrical vessel

Solution:
For an element at the thin walled pressure vessel, the state of stress
is biaxial with circumferential and axial stresses,
c

pD
,
2t

pD
4t

## Where D is the diameter and t is the wall thickness of the vessel.

Using Tresca Criterion: The order of the principal stresses is important for the
Tresca criterion. Therefore,

11 c ,

22 a ,

11 33 Y S y

3 0
py

2 t y
D

5.625MPa

## Using von Mises Criterion:

12 22 1 2 S yt2

c2 a2 c a S yt2

Which gives,
6.495
Comparing the two criteria, the Tresca is more conservative by approximately 14%.

Discussion Problem

(a) FindthebendingandtransverseshearstressatpointsAandBinthe
figure.
(b)Findthemaximumnormalstressandmaximumshearstressatboth
points.
(c)Forayieldpointof50,000psi,findthefactorofsafetybasedonthe
maximumnormalstresstheoryandthemaximumshearstresstheory.

## Distortion Energy Theory or

von-Mises Criterion: Hydrostatic Stress

Thedistortionalstresscomponentsareoftencalledthedeviatoricstress
components.

## Distortion Energy Theory or

von-Mises Criterion: Hydrostatic Effect
Thehydrostaticstresscausesachangeinthevolume.
evolumetricstrain
KBulkModulus
h = K*e
h = (1 +2 +3 )/3
Thecubegetsbiggerintension,
smallerincompression.

Theseunequalstressesacttodeformordistortthe
materialelement.
Thereisnochangeinvolume,butthereisachangein
shape.
Thesestressestrytoelongateorcompressthematerial
moreinonedirectionthaninanother.

## Distortion Energy Theory or

von-Mises Criterion

## Distortion Energy Theory or

von-Mises Criterion: Simple Tension Test

## Distortion Energy Theory or

von-Mises Criterion: Equations

## Distortion Energy Theory or

von-Mises Criterion: Equations

TheeffectivestressiscommonlyreferredtoasthevonMises
stresswhocontributedtothetheoryafterDr.R.vonMises

## Distortion Energy Theory or

von-Mises Criterion: Graphical Presentation

Thesurface,blueline,atwhichthematerialjustbeginstoyieldiscalledtheyield
surface.

## Tresca & von-Mises:

Experimental Results

## Tresca & von-Mises:

more experimental results

## Experimental Results (conclusions)

BoththeDistortionEnergyTheoryandtheMaximum
ShearStressTheoryprovidereasonableestimatesfor
ductile,homogeneous,isotropicmaterialswhose
compressiveandtensilestrengthsareapproximately
thesame.
BoththeDistortionEnergyTheoryandtheMaximum
ShearStressTheory predictthattheonsetofyieldis
independentofthehydrostaticstress.Thisagrees
reasonablywellwithexperimentaldataformoderate
hydrostaticpressures.

## Experimental Results Conclusions (cont.)

BoththeDistortionEnergyTheoryandtheMaximum
ShearStressTheoryunderpredictthestrengthof
materialsoftenhavemuchhighercompressive
strengthsthantensilestrengths.
TheDistortionEnergyTheoryisslightlymoreaccurate
thantheMaximumShearStressTheory.The
DistortionEnergyTheoryistheyieldcriteriamost
oftenusedinthestudyofclassicalplasticity.Its
continuousnaturemakesitmoremathematically
amenable.

## Failure Versus Yielding

The high stresses around stress concentration factors
are often very localized, and the local yielding will
cause a redistribution of stresses to adjacent material.
In many cases the local yielding will not cause a
conditions.
It is common to differentiate between local yielding
and gross yielding through the thickness of a member.
Local yielding may lead to early fatigue failure, and
stress concentration effects must always be
considered in fatigue calculations.

## Industry Standards & Codes

The American Society of Mechanical Engineers base the
ASME Boiler and Pressure Vessel Code on the Maximum
Shear Stress Theory.
The American Institute of Steel Construction does not use
either in the Manual of Steel Construction. Buildings,
bridges, etc. are dominated by normal stresses and
buckling type failures.
The American Society of Civil Engineers use the Distortion
Energy Theory in Design of Steel Transmission Pole
Structures.
There is no single standard that applies to the design of
machine components. Standard industry practice is to use
either the Distortion Energy Theory or Maximum Shear
Stress Theory with an appropriate safety factor.

Definitions

NotethatR=1fora
completelyreversed
stressstatewithzero
meanstress.

MaterialFailureMechanisms

## Ductile fracture: failure that involves a significant amount of plastic

deformation prior to fracture

## Brittle fracture: failure without a significant amount of macroscopic

plastic deformation prior to fracture.

## Fatigue failure: failure associated with slow crack growth due to

changing stress states.

## Corrosion-fatigue failure: failure due the combined actions of

changing stress and corrosive environments.

## Stress-corrosion cracking: failure in which a steady tensile stress

leads to the initiation and propagation of fracture in a relatively mild
chemical environment.

phenomena.

MaterialFailureMechanisms(cont.)

Liquiderosionfailure: typeofwearfailureinwhichliquidisresponsible
forremovalofmaterial.

Liquidmetalembrittlement: involvesthemateriallosingsomedegreeof
ductilitybelowitsyieldstrengthduetoitssurfacebeingwettedbya
lowermeltingpointliquidmetal.

Hydrogenembrittlement: notoriouscauseofcatastrophicfailureinhigh
ductility(fewpartspermillionofhydrogenisenough).

Creepandstressrupturefailures: failureduetocontinuedstraingrowth
Allofthesemechanismsareassociatedwiththefailureofthematerial.
Theydonotincludeoneofthemostimportantstructuralfailure
mechanismsthatmustbeconsideredincompressivestress
environments,Buckling.

MaterialFailureMechanisms(cont.)
alternating,orfluctuating
Maximumstresseswellbelowyieldstrength
Failureoccursaftermanystresscycles
Failureisbysuddenultimatefracture

## Stage I Initiation of microcrack due to cyclic plastic

deformation
Stage II Progresses to
macro-crack that repeatedly
opens and closes, creating
bands called beach marks
Stage III Crack has
propagated far enough that
remaining material is
and fails by simple ultimate
failure

## AISI 4320 drive

shaft
B crack initiation at
stress concentration
in keyway
C Final brittle
failure

## Fatigue Failure: Practical Examples (Cont.)

Fatigue failure
initiating at
mismatched grease
holes
Sharp corners (at
arrows) provided
stress concentrations

## Fatigue Failure: Practical Examples (Cont.)

Fatigue Failure:
Practical Examples
(Cont.)

Fatigue Failure:
Shaft Failures: You could be rich?? (Cont.)

## Despite decades of research and 1000s of Ph.D.

theses, even highly engineered shafts fail all too
frequently. Even NASA cant always get it right.
Often the connections are to blame: keys, splines,
couplings, and so on. Often fatigue wear failure is
the culprit.

## The world may beat a path to your door!

FatigueLifeMethods
Threemajorfatiguelifemodels
Methodspredictlifeinnumberofcyclestofailure,N,fora
Stresslifemethod
Leastaccurate,particularlyforlowcycleapplications

Strainlifemethod
Detailedanalysisofplasticdeformationatlocalizedregions

Linearelasticfracturemechanicsmethod
Assumescrackexists
Predictscrackgrowthwithrespecttostressintensity

Strain-Life Method

## Method uses detailed analysis of plastic deformation at localized

regions
Compounding of several idealizations leads to significant
uncertainties in numerical results
Useful for explaining nature of fatigue

Strain-Life Method

## Fatigue failure almost

always begins at a local
discontinuity
When stress at
discontinuity exceeds
elastic limit, plastic
strain occurs
Cyclic plastic strain can
change elastic limit,
Fig. 612 shows true
stress-true strain
hysteresis loops of the
first five stress reversals
Fig. 612

StrainLifeMethod:Relation

## Figure 613 plots relationship of fatigue life to true-strain

amplitude
Fatigue ductility coefficient 'F is true strain corresponding to
fracture in one reversal (point A in Fig. 612)
Fatigue strength coefficient 'F is true stress corresponding to
fracture in one reversal (point A in Fig. 612)

Fig. 613

StrainLifeMethod:RelationofFatigueLifetoStrain

## Fatigue ductility exponent c is the slope of plastic-strain line, and is

the power to which the life 2N must be raised to be proportional to the
true plastic-strain amplitude. Note that 2N stress reversals corresponds
to N cycles.
Fatigue strength exponent b is the slope of the elastic-strain line, and
is the power to which the life 2N must be raised to be proportional to
the true-stress amplitude.

Fig. 613

StrainLifeMethod:RelationofFatigueLifetoStrain

## Total strain is sum of elastic and plastic strain

Total strain amplitude is half the total strain range

## Applying Eq. (a), the total-strain amplitude is

StrainLifeMethod:RelationofFatigueLifetoStrain

## Known as Manson-Coffin relationship between fatigue life and

total strain
Some values of coefficients and exponents given in Table A23
Equation has limited use for design since values for total strain at
The equation may be easily used with numerical simulation of
design problems where all strains are calculated numerically.

StrainLifeMethod:TypicalvaluesforMansonCoffinrelationship

TypicalvaluesforMansonCoffinrelationship(Cont.)

FatigueLifeMethods:StressLifeMethod
Testspecimensaresubjectedtorepeatedstresswhilecounting
cyclestofailure
MostcommontestmachineisR.R.Moorehighspeedrotating
beammachine
Subjectsspecimentopurebendingwithnotransverseshear
Asspecimenrotates,stressfluctuatesbetweenequal
magnitudesoftensionandcompression,knownascompletely
reversedstresscycling
Specimeniscarefullymachinedandpolished

FatigueLifeMethods:StressLifeMethod,SNDiagram

Numberofcyclestofailureatvaryingstresslevelsisplottedonlog
logscale
Forsteels,akneeoccursnear106 cycles
Strengthcorrespondingtothekneeiscalledendurancelimit Se

StressLifeMethod: SNDiagramforSteel
StresslevelsbelowSe predictinfinitelife
Between103 and106 cycles,finitelifeispredicted
Below103 cyclesisknownaslowcycle,andisoftenconsidered
quasistatic.Yieldingusuallyoccursbeforefatigueinthiszone.

## The S-N curve gives

the cycles to failure
for a completely
reversed (R=-1)
uniaxial stress state.
What do you do if
the stress state is not
completely
reversed?

StressLifeMethod: SNDiagramforNonferrousMetals

Nonferrousmetalsoftendonothaveanendurancelimit.
FatiguestrengthSf isreportedataspecificnumberofcycles
FigureshowstypicalSN diagramforaluminums

## StressLifeMethod: The Endurance Limit

Simplified estimate of endurance limit for steels for the rotating-beam specimen, S'e

Fig. 617

StressLifeMethod: FiniteLifeCalculation

## For design, an approximation of the idealized S-N diagram is

desirable.
To estimate the fatigue strength at 103 cycles, start with Eq. (6-2)

cycles as

## Combine with Eq. (62),

StressLifeMethod: FiniteLifeCalculation

At 103 cycles,
f is the fraction of Sut represented by ( S f ) 3
10
Solving for f,

## To find b, substitute the endurance strength and corresponding

cycles into Eq. (69) and solve for b

Eqs. (611) and (612) can be substituted into Eqs. (69) and
(610) to obtain expressions for S'f and f

## StressLifeMethod: Fatigue Strength Fraction f

Plot Eq. (610) for the fatigue strength fraction f of Sut at 103
cycles
Use f from plot for S'f = f Sut at 103 cycles on S-N diagram
Assumes Se = S'e= 0.5Sut at 106 cycles

Fig. 618

StressLifeMethod: FiniteLifeCalculation

## Write equation for S-N line

from 103 to 106 cycles

## Two known points

At N =103 cycles,
Sf = f Sut

At N =106 cycles,
Sf = Se

## Equations for line (Use f

from figure 6-18 previous
slide):

Fig. 610

StressLifeMethod: FiniteLifeCalculation

## If a completely reversed stress rev is given, setting Sf = rev in

Eq. (613) and solving for N gives,

## Note that the typical S-N diagram is only applicable for

completely reversed stresses
For other stress situations, a completely reversed stress with the
same life expectancy must be used on the S-N diagram

## Low-cycle fatigue is defined for fatigue failures in the range

1 N 103
On the idealized S-N diagram on a log-log scale, failure is
predicted by a straight line between two points (103, f Sut) and
(1, Sut)

## Endurance limit S'e is for carefully prepared and tested specimen

If warranted, Se is obtained from testing of actual parts
When testing of actual parts is not practical, a set of Marin
factors are used to adjust the endurance limit

## Stresses tend to be high at the surface

Surface finish has an impact on initiation of cracks at localized
stress concentrations
Surface factor is a function of ultimate strength. Higher strengths
are more sensitive to rough surfaces.

## Larger parts have greater surface area at high stress levels

Likelihood of crack initiation is higher
Size factor is obtained from experimental data with wide scatter
For bending and torsion loads, the trend of the size factor data is
given by

## Applies only for round, rotating diameter

For axial load, there is no size effect, so kb = 1

## StressLifeMethod: Size Factor kb

For parts that are not round and rotating, an equivalent round
rotating diameter is obtained.
Equate the volume of material stressed at and above 95% of the
maximum stress to the same volume in the rotating-beam
specimen.
Lengths cancel, so equate the areas.
For a rotating round section, the 95% stress area is the area of a
ring,

Equate 95% stress area for other conditions to Eq. (622) and
solve for d as the equivalent round rotating diameter

to Eq. (622),

## StressLifeMethod: Size Factor kb

Table 63
A95 for common
non-rotating
structural shapes

## Accounts for changes in endurance limit for different types of fatigue

method (Sec. 614) when more than one load type is present.

## Endurance limit appears to maintain same relation to ultimate

strength for elevated temperatures as at room temperature
This relation is summarized in Table 64

## If ultimate strength is known for operating temperature, then just

use that strength. Let kd = 1 and proceed as usual.
If ultimate strength is known only at room temperature, then use
Table 64 to estimate ultimate strength at operating temperature.
With that strength, let kd = 1 and proceed as usual.
Alternatively, use ultimate strength at room temperature and
apply temperature factor from Table 64 to the endurance limit.

## A fourth-order polynomial curve fit of the underlying data of

Table 64 can be used in place of the table, if desired.

## StressLifeMethod: Reliability Factor ke

From Fig. 617, S'e = 0.5 Sut is typical of the data and represents
50% reliability.
Reliability factor adjusts to other reliabilities.
Only adjusts Fig. 617 assumption. Does not imply overall
reliability.

Fig. 617

Table 65

## Reminder to consider other possible factors.

Residual stresses
Directional characteristics from cold working
Case hardening
Corrosion
Surface conditioning, e.g. electrolytic plating and metal
spraying
Cyclic Frequency
Frettage Corrosion
Limited data is available.
May require research or testing.

## StressLifeMethod: Stress Concentration and Notch Sensitivity

Obtain Kt as usual (e.g. Appendix A15)
For fatigue, some materials are not fully sensitive to Kt so a
reduced value can be used.
Define Kf as the fatigue stress-concentration factor.
Define q as notch sensitivity, ranging from 0 (not sensitive) to 1
(fully sensitive).

For q = 0, Kf = 1
For q = 1, Kf = Kt

## StressLifeMethod: Notch Sensitivity

Then get Kf from Eq. (632): Kf = 1 + q( Kt 1)

Fig. 620

## StressLifeMethod: Notch Sensitivity

Then get Kfs from Eq. (632): Kfs = 1 + qs( Kts 1)
Note that Fig. 621 is updated in 9th edition.

Fig. 621

## StressLifeMethod: Notch Sensitivity

Alternatively, can use curve fit equations for Figs. 620 and 621
to get notch sensitivity, or go directly to Kf .

Bending or axial:

Torsion:

## Cast irons are already full of discontinuities, which are included

in the strengths.
Recommended to use q = 0.2 for cast irons.

## Use Kf as a multiplier to increase the nominal stress.

Some designers (and previous editions of textbook) sometimes
applied 1/ Kf as a Marin factor to reduce Se .
For infinite life, either method is equivalent, since
1/ K f Se

Se

nf

K f

## For finite life, increasing stress is more conservative. Decreasing

Se applies more to high cycle than low cycle.

Fig. 622

Fig. 622

## StressLifeMethod: Example 6-9

Fatigue Failure Curves

Whenthemeanstressis
zero,thealternating
componentisequaltothe
endurancelimit.
Theinteractioncurvesare
forinfinitelifeoralarge
numberofcycles.

Goodman Diagram
Anycombinationofmeanand
alternatingstressthatlieson
or
belowGoodmanlinewillhave
infinitelife.

FactorofSafetyFormat

Notethatthefatiguestress
concentrationfactorisapplied
onlytothealternating
component.
Nf: factorofsafety
kf: fatiguestressconcentration
factor

Gerber Diagram

Notethat:
Thefatiguestressconcentration
factorisappliedonlytothe
alternatingcomponent.
Anycombinationofmeanand
alternatingstressthatliesonor
belowtheGerberlinewillhave
infinitelife.
Nf : factorofsafety

kf :

fatiguestressconcentration
factor

Soderberg & ASME Elliptic

Summary of Equations

Fivecommonlyusedfailurecriteriaareshown
Gerberpassesthroughthedata
ASMEellipticpassesthroughdataandincorporatesroughyieldingcheck
ModifiedGoodmanislinear,sosimpletouse.ItismoreconservativethanGerber.
Soderbergprovidesaveryconservativesingle
Langerlinerepresentsstandardyieldcheck.
Itisequivalenttocomparingmaximumstresstoyieldstrength

Fig. 627

Summary of Equations (Cont.)

designequations
nisthedesignfactororfactorofsafetyforinfinitefatiguelife

Summary of Equations (Cont.)
Tables66to68summarizethepertinentequationsfor
ModifiedGoodman,Gerber,ASMEelliptic,andLangerfailure
criteria
Thefirstrowgivesfatiguecriterion
Thesecondrowgivesyieldcriterion
Thethirdrowgivestheintersectionofstaticandfatiguecriteria
Thefourthrowgivestheequationforfatiguefactorofsafety
Thefirstcolumngivestheintersectingequations
Thesecondcolumngivesthecoordinatesoftheintersection

Summary of Equations: ModifiedGoodman (Cont.)

Summary of Equations: Gerber (Cont.)

Summary of Equations: ASMEElliptic (Cont.)

Example610

Example610

Example610

Fig. 628

Example610

Fig. 628

Example610

Example610

Fig. 629

Example610

Example

1.0

1.0

1.0

Alternating Stress

(Cont.)

1.0

1.0

1.0

to fail

(Cont.)