You are on page 1of 37

Fatigue Failure

It has been recognized that a metal subjected


to a repetitive or fluctuating stress will fail at a
stress much lower than that required to cause
failure on a single application of load. Failures
occurring under conditions of dynamic loading
are called fatigue failures.
Fatigue failure is characterized by three stages

Ken Youssefi

Crack Initiation

Crack Propagation

Final Fracture
MAE dept., SJSU

Jack hammer component,


shows no yielding before
fracture.

Crack initiation site

Fracture zone
Propagation zone, striation
Ken Youssefi

MAE dept., SJSU

VW crank shaft fatigue failure due to cyclic bending and torsional stresses

Propagation
zone, striations

Crack initiation site

Ken Youssefi

Fracture area

MAE dept., SJSU

928 Porsche timing pulley

Ken Youssefi

MAE dept., SJSU

Crack started at the fillet

Fracture surface of a failed bolt. The


fracture surface exhibited beach marks,
which is characteristic of a fatigue failure.

1.0-in. diameter steel pins from


agricultural equipment.
Material; AISI/SAE 4140 low
allow carbon steel

Ken Youssefi

MAE dept., SJSU

bicycle crank spider arm

This long term fatigue crack in a high quality component took a


considerable time to nucleate from a machining mark between the spider
arms on this highly stressed surface. However once initiated propagation
was rapid and accelerating as shown in the increased spacing of the 'beach
marks' on the surface caused by the advancing fatigue crack.
Ken Youssefi

MAE dept., SJSU

Crank shaft

Gear tooth failure

Ken Youssefi

MAE dept., SJSU

Hawaii, Aloha Flight 243, a Boeing 737, an upper part of the plane's cabin
area rips off in mid-flight. Metal fatigue was the cause of the failure.

Ken Youssefi

MAE dept., SJSU

Fracture Surface Characteristics


Mode of fracture

Typical surface characteristics

Ductile

Cup and Cone


Dimples
Dull Surface
Inclusion at the bottom of the dimple

Brittle Intergranular

Shiny
Grain Boundary cracking

Brittle Transgranular

Shiny
Cleavage fractures
Flat

Fatigue

Beachmarks
Striations (SEM)
Initiation sites
Propagation zone
Final fracture zone

Ken Youssefi

MAE dept., SJSU

Fatigue Failure Type of Fluctuating Stresses

Alternating stress

a =

max min
2

Mean stress

m
=
Ken Youssefi

MAE dept., SJSU

max + min
2
10

Fatigue Failure, S-N Curve


Test specimen geometry for R.R. Moore
rotating beam machine. The surface is
polished in the axial direction. A constant
bending load is applied.

Typical testing apparatus, pure bending


Motor

Load

Rotating beam machine applies fully reverse bending stress


Ken Youssefi

MAE dept., SJSU

11

Fatigue Failure, S-N Curve


N > 103

N < 103

Finite life

Infinite life

Se

Se = endurance limit of the specimen

Ken Youssefi

MAE dept., SJSU

12

Relationship Between Endurance Limit


and Ultimate Strength
Steel
Steel

Se =

0.5Sut

Sut 200 ksi (1400 MPa)

100 ksi

Sut > 200 ksi

700 MPa Sut > 1400 MPa

Cast iron

Cast iron

Se =

0.4Sut

Sut < 60 ksi (400 MPa)

24 ksi

Sut 60 ksi

160 MPa Sut < 400 MPa


Ken Youssefi

MAE dept., SJSU

13

Relationship Between Endurance Limit and


Ultimate Strength
Aluminum
Aluminum alloys

Se =

0.4Sut

Sut < 48 ksi (330 MPa)

19 ksi

Sut 48 ksi

130 MPa Sut 330 MPa


For N = 5x108 cycle

Copper alloys
Copper alloys

Se =

0.4Sut

Sut < 40 ksi (280 MPa)

14 ksi

Sut 40 ksi

100 MPa Sut 280 MPa


For N = 5x108 cycle
Ken Youssefi

MAE dept., SJSU

14

Correction Factors for Specimens Endurance Limit


For materials exhibiting a knee in the S-N curve at 106 cycles
S = endurance limit of the specimen (infinite life > 106)
e

Se = endurance limit of the actual component (infinite life > 106)


S
103

Se
106

For materials that do not exhibit a knee in the S-N curve, the infinite
life taken at 5x108 cycles

Sf = fatigue strength of the specimen (infinite life > 5x108)


Sf = fatigue strength of the actual component (infinite life > 5x108)
S

Ken Youssefi

103

Sf
5x108

MAE dept., SJSU

15

Correction Factors for Specimens Endurance Limit


Se = Cload Csize Csurf Ctemp Crel (Se)
Load factor, Cload

Ken Youssefi

Pure bending

Cload = 1

Pure axial

Cload = 0.7

Pure torsion

Cload = 1 if von Mises stress is used, use


0.577 if von Mises stress is NOT used.

Combined loading

Cload = 1

MAE dept., SJSU

16

Correction Factors for Specimens Endurance Limit


Size factor, Csize
Larger parts fail at lower stresses than smaller parts. This is
mainly due to the higher probability of flaws being present in
larger components.

For solid round cross section

d 0.3 in. (8 mm)

Csize = 1

0.3 in. < d 10 in.

Csize = .869(d)-0.097

8 mm < d 250 mm

Csize = 1.189(d)-0.097

If the component is larger than 10 in., use Csize = .6

Ken Youssefi

MAE dept., SJSU

17

Correction Factors for Specimens Endurance Limit


For non rotating components, use the 95% area approach to calculate
the equivalent diameter. Then use this equivalent diameter in the
previous equations to calculate the size factor.

dequiv = (

Ken Youssefi

A95
0.0766

)1/2

MAE dept., SJSU

d95

18

Correction Factors for Specimens Endurance Limit


surface factor, Csurf
The rotating beam test specimen has a polished surface. Most
components do not have a polished surface. Scratches and
imperfections on the surface act like a stress raisers and reduce
the fatigue life of a part. Use either the graph or the equation with
the table shown below.

Csurf = A (Sut)b

Ken Youssefi

MAE dept., SJSU

19

Correction Factors for Specimens Endurance Limit


Temperature factor, Ctemp
High temperatures reduce the fatigue life of a component. For
accurate results, use an environmental chamber and obtain the
endurance limit experimentally at the desired temperature.
For operating temperature below 450 oC (840 oF) the
temperature factor should be taken as one.

Ctemp = 1

Ken Youssefi

for T 450 oC (840 oF)

MAE dept., SJSU

20

Correction Factors for Specimens Endurance Limit


Reliability factor, Crel
The reliability correction factor accounts for the scatter and
uncertainty of material properties (endurance limit).

Ken Youssefi

MAE dept., SJSU

21

Fatigue Stress Concentration Factor, Kf


Experimental data shows that the actual stress concentration factor is
not as high as indicated by the theoretical value, Kt. The stress
concentration factor seems to be sensitive to the notch radius and the
ultimate strength of the material.
Notch sensitivity
Kf = 1 + (Kt 1)q
factor

Ken Youssefi

MAE dept., SJSU

22

Fatigue Stress
Concentration Factor,
Kf for Aluminum

Ken Youssefi

MAE dept., SJSU

23

Design process Fully Reversed Loading for Infinite Life

Determine the maximum alternating applied stress, a, in terms of


the size and cross sectional profile

Select material Sy, Sut

Choose a safety factor n

Determine all modifying factors and calculate the endurance


limit of the component Se

Determine the fatigue stress concentration factor, Kf

Use the design equation to calculate the size

Se
Kf a =
n

Investigate different cross sections (profiles), optimize for size or weight

You may also assume a profile and size, calculate the alternating stress
and determine the safety factor. Iterate until you obtain the desired
safety factor

Ken Youssefi

MAE dept., SJSU

24

Design for Finite Life

Sn = a (N)b equation of the fatigue line


A

B
Sf

Se
106

103

Point A

Point B

Sn = .9Sut

Point A

Sn = .9Sut

N = 10

N = 103

Sn = Se

Sn = Sf

Point B

N = 10
Ken Youssefi

5x108

103

MAE dept., SJSU

N = 5x108
25

Design for Finite Life


Sn = a (N)b

log Sn = log a + b log N


Apply conditions for point A and B to find the
two constants a and b

log .9Sut = log a + b log 10

a=

log Se = log a + b log 106

b=

Sn = Se ( 106 )
Calculate Sn

Se
1
3

log

.9Sut
Se

Se
log ( .9S )
ut

and replace Se in the design equation


Kf a =

Ken Youssefi

(.9Sut)

Sn
n

Design equation

MAE dept., SJSU

26

The Effect of Mean Stress on Fatigue Life


Mean stress exist if the
loading is of a repeating or
fluctuating type.

a
Sy

Yield line

Gerber curve
Alternating
stress

Se
Goodman line

Sy

Soderberg line

Sut

Mean stress
Ken Youssefi

MAE dept., SJSU

27

The Effect of Mean Stress on Fatigue Life


Modified Goodman Diagram
a
Sy

Alternating
stress

Yield line

Se
Goodman line
C

Safe zone

Sy

Sut

Mean stress

Ken Youssefi

MAE dept., SJSU

28

The Effect of Mean Stress on Fatigue Life


Modified Goodman Diagram
a
Sy

Yield line

Se
Goodman line

Safe zone
- m

Ken Youssefi

Safe zone

Sy

- Syc

MAE dept., SJSU

Sut

+ m

29

The Effect of Mean Stress on Fatigue Life


Modified Goodman Diagram
Fatigue,

m 0

Fatigue,

a
Se

Se

a = n
f

Sn

+
+

m > 0

m
Sut

m
Sut

=
= 1

Se

Yield

a + m = n
y

nf
Finite life

Sy

Safe zone

Ken Youssefi

Infinite life

Yield

Sy

- m

Safe zone

a + m = n
y

Sy

- Syc

MAE dept., SJSU

Sut

+ m

30

Applying Stress Concentration factor to Alternating


and Mean Components of Stress

Determine the fatigue stress concentration factor, Kf, apply directly to

If Kf max < Sy then there is no yielding at the notch, use Kfm = Kf

the alternating stress Kf a

and multiply the mean stress by Kfm Kfm m

If Kf max > Sy then there is local yielding at the notch, material at the

notch is strain-hardened. The effect of stress concentration is reduced.


Calculate the stress concentration factor for the mean stress using
the following equation,

Kfm =

Sy

Fatigue design equation


Kf a
Kfmm
=
+

Se

Ken Youssefi

Sut

Kf a

m
1

nf

MAE dept., SJSU

Infinite life
31

Combined Loading
All four components of stress exist,

xa

alternating component of normal stress

xm

mean component of normal stress

xya

alternating component of shear stress

xym

mean component of shear stress

Calculate the alternating and mean principal stresses,

Ken Youssefi

1a, 2a = (xa /2)

(xa /2)2 + (xya)2

1m, 2m = (xm /2)

(xm /2)2 + (xym)2

MAE dept., SJSU

32

Combined Loading
Calculate the alternating and mean von Mises stresses,

a = (1a2 + 2a2 - 1a2a)1/2


2
2
m = (1m
+ 2m - 1m2m)1/2

Fatigue design equation

a
Se

Ken Youssefi

m
Sut

MAE dept., SJSU

nf

Infinite life

33

Design Example
12

A rotating shaft is carrying 10,000 lb force


as shown. The shaft is made of steel with

D = 1.5d

Sut = 120 ksi and Sy = 90 ksi. The shaft


is rotating at 1150 rpm and has a
machine finish surface. Determine the
diameter, d, for 75 minutes life. Use
safety factor of 1.6 and 50% reliability.
Calculate the support forces,

10,000 lb.
6
6

R1
r (fillet radius) = .1d

R2

R1 = 2500, R2 = 7500 lb.

The critical location is at the fillet,


Calculate the alternating stress,

MA = 2500 x 12 = 30,000 lb-in

a =

Mc

32M

=
3

305577

m = 0

Determine the stress concentration factor


r
= .1
d
D
= 1.5
d
Ken Youssefi

Kt = 1.7

MAE dept., SJSU

34

Design Example
Assume d = 1.0 in

Using r = .1 and Sut = 120 ksi,


q (notch sensitivity) = .85
Kf = 1 + (Kt 1)q = 1 + .85(1.7 1) = 1.6

Calculate the endurance limit

Cload = 1 (pure bending)

Crel = 1 (50% rel.)


Ctemp= 1 (room temp)
Csurf = A (Sut)b = 2.7(120)
0.3 in. < d 10 in.

-.265

= .759

Csize = .869(d)-0.097 = .869(1)-0.097 = .869

Se = Cload Csize Csurf Ctemp Crel (Se) = (.759)(.869)(.5x120) = 39.57


Ken Youssefi

MAE dept., SJSU

ksi
35

Design Example
Design life, N = 1150 x 75 = 86250 cycles
Se
log ( .9S )
86250
N
ut
Sn = 39.57 ( 6
Sn = Se ( 6 )
10

a =

305577

d3

10

= 305.577 ksi

n=

Sn
Kfa

39.57

log ( .9x120 )
= 56.5 ksi

56.5
1.6x305.577

= .116 < 1.6

So d = 1.0 in. is too small

Assume d = 2.5 in
All factors remain the same except the size factor and notch sensitivity.

Using r = .25 and Sut = 120 ksi,


q (notch sensitivity) = .9

Kf = 1 + (Kt 1)q = 1 + .9(1.7 1) = 1.63

Csize = .869(d)-0.097 = .869(2.5)-0.097 = .795


Ken Youssefi

MAE dept., SJSU

Se = 36.2 ksi
36

Design Example
Se = 36.2 ksi Sn = 53.35 ksi

a =

305577

(2.5)

= 19.55 ksi

n=

Sn

Kfa

53.35
1.63x19.55

= 1.67 1.6

d = 2.5 in.

Check yielding

n=

Ken Youssefi

Sy
90
= 2.8 > 1.6 okay
=
Kfmax 1.63x19.55
MAE dept., SJSU

37

You might also like