Fatigue Failure

Fatigue of Metallic Materials
Fatigue failures occur due to the application of fluctuating stresses that

are much lower than the stress required to cause failure during a single
application of stress. It has been estimated that fatigue contributes to
approximately 90% of all mechanical service failures. Fatigue fractures
are serious because they are insidious; that is, they are frequently
“sneaky” and can occur without warning. Obviously, if service is
abnormal as a result of excessive overloading, corrosive environments, or
other conditions, the possibility of fatigue fracture is increased.
Fatigue loading, is traditionally analyzed in design by considering the S-
N data without reference to the existence of cracks.
Fig. Types of fluctuating loads
𝜎𝑚𝑎𝑥 +𝜎 𝑚𝑖𝑛
𝜎𝑚𝑎𝑥 = 𝜎𝑚 + 𝜎𝑎 , 𝜎𝑚𝑖𝑛 = 𝜎𝑚 − 𝜎𝑎 , 𝜎𝑚 =
2
𝜎𝑚𝑎𝑥 −𝜎 𝑚𝑖𝑛
The stress range (Δσ) will be; ∆𝜎 = 2𝜎𝑎 = 𝜎𝑚𝑎𝑥 − 𝜎𝑚𝑖𝑛 , 𝜎𝑎 =
2
𝜎
The stress ration (R) will be; 𝑅 = 𝜎 𝑚𝑖𝑛 ,
𝑚𝑎𝑥
The S-N Curves:
Fig. Two types of S-N curves for different materials
A log scale is used for the number of cycles. The value of stress, S, can
be the maximum stress, 𝜎𝑚𝑎𝑥 , or the stress amplitude, 𝜎𝑎 . The fatigue
life is the number of cycles to failure at a specified stress level, while the
fatigue strength (also referred to as the fatigue or endurance limit).
Normally, the fatigue strength increases as the static tensile strength
increases. It should be noted that there is a considerable amount of scatter
in fatigue test results. Therefore, it is important to test a sufficient number
of specimens to obtain statistically meaningful results.
Much of the fatigue data has been determined for completely reversed
bending. However, the effects of mean stress are important, and an
increase in mean stress will always cause a reduction in the fatigue
life.
A number of mathematical models have been developed that allow the
effects of mean stress on stress amplitude to be predicted from fully
reversed-bending data.
a. Goodman: developed a linear model, while
b. Gerber: used a parabolic model.

Test data for ductile metals usually fall closer to the Gerber parabolic
curve; however, because of the scatter in fatigue data and the fact that
notched data fall closer to the Goodman line, the more conservative
Goodman relationship is often used in practice.
c. Soderberg: If the component design is based on yield rather than

ultimate strength, as most are, then the even more conservative Soderberg
relationship can be used.
Mathematically, the three relationships can be expressed by:
where x=1 for the Goodman line, x=2 for the Gerber curve, σu=σy for
the Soderberg curve, and σe is the fatigue limit for completely
reversed bending.
Fig. Comparison of Goodman, Gerber, and Soderberg models for relating

mean stress to stress amplitude
Measurement of fatigue:
a. Reciprocating-bending fatigue test machine
b. Rotating bending fatigue testing machine

c. tension-tension or tension–compression (push–pull)
Fatigue Fracture Mechanics:

Fracture mechanics deals with the fatigue problem, while considering the
presence of a crack-like defect of known size. This crack propagates
during the service life of the structure, reaching a critical size sufficient to
induce catastrophic failure as shown in the figure:
The Figure represents a series of tests on a number of identical specimens

having the same initial sharp crack ao. Each specimen is subjected to
cyclic loading with certain stress range    max   min as shown in Fig
(b). The figure shows that a higher stress fluctuation increment 
causes rapid crack growth, i.e., attaining a specific crack length ac after
smaller number of stress cycles.
( 1   2   3 )  ( N1  N 2  N 3 )
Hence, the crack growth rate expressed by (da / dN ) at a given crack length
is higher for greater stress fluctuation. Referring back to the stress
intensity factor relation:
K  K max  K min   a ( max   min )CCF   a . .CCF

Note that since the stress intensity factor is undefined in compression,
Kmin is taken as zero if σmin is negative.
The following figure is plotted representing (da/dN) vs K presented on

a log-log scale. The shape of this curve defines three regions, namely:
Region (i) Non-propagating cracks:
This region begins with a threshold value of stress intensity Kth, below
which no observable crack propagation occurs, or simply the crack
remains dormant.
Region (ii) Steady crack propagation:

In this region the crack propagates steadily under cyclic loading leading
to a fracture surface with beach marks. The relation is linear taking an
empirical form known as Paris law:
 C K 
da m
dN
Here C, and m are experimentally determined constants depending on

material and test conditions. The value of C is the value of ( da / dN ) at
1/2
K  1 MPa.m , and m is the slope of the line. The following table gives
typical values of Kth , C and m for some materials. Data are commonly
generated using standardized specimens, such as the compact tension
specimen to obtain crack growth under plane strain conditions. Fatigue
crack growth data are determined using a cycle of stress ratio
R   min /  max  K min / K max  0 , since compressive stresses close up the
crack.
Material ΔKth C (10)−11 m
MPa (m)1/2
Mild Steel 3.2-6.6 0.24 3.3
Structural Steel 2.0 – 5.0 0.07-0.1 3.85-4.2
Structural steel in seawater 1.0 – 1.5 1.6 3.3
Aluminum 1.0 – 2.0 4.56 2.9
Aluminum alloy 1.0 – 2.0 3 - 19 2.6 – 3.9
Copper 1.8 – 2.8 1.34 3.9
Titanium 2.0 – 3.0 68.8 4.4
Steel AISA 4340 - 4.85 2.64
Steel AISI 4340 in seawater - 1.23 2.98
Region (iii): Unstable crack growth rate:

In this region, the crack length ultimately attains its critical size ac,
leading to catastrophic fracture. This length is obtained from the fracture
toughness relation K I    ac CCF  K I c . It is this value of ac that is
used to determine the safe life of a structure during fatigue loading.
Safe life prediction:
The safe life prediction of a component with an initial crack, subjected to

fatigue loading is driven from the stable crack growth governed by Paris
law as follows:
af af
da da
Nf   
ao CK
m

ao C   aCCF
m

Where Nf is the number of cycles for the crack to propagate from an
initial length ao to a final length af . Since CCF varies with the size of the
crack, then an average value CCFav =  f (ao / W )  f (a f / W )/ 2 is assumed
to simplify integration.
For m  2 :
m m
(1 ) (1 )
1 af 2
 ao 2
Nf 
C CCFav   m / 2  
m m
m
(1  )
2
For m=2: ln a f / ao 
Nf 
C CCFav    
2 2
Typical Fatigue fracture surfaces (Fractographs):
This high tensile steel bolt failed under low stress high cycle conditions.
Fig. Transmission electron fractograph of aluminum alloy laboratory

spectrum loading fatigue test. Striation spacing varies according to
loading, which consisted of ten cycles at a high stress alternating with
ten cycles at a lower stress. The fracture surface exhibits bands of ten
coarse striations alternating with bands of ten fine striations.

Fatigue Failure

Uploaded by

Copyright:

Available Formats

Fatigue Failure

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fatigue Failure

Uploaded by

Copyright:

Available Formats

Fatigue of Metallic Materials

Fatigue failures occur due to the application of fluctuating stresses that

Fig. Types of fluctuating loads

Fig. Two types of S-N curves for different materials

a. Goodman: developed a linear model, while

b. Gerber: used a parabolic model.

c. Soderberg: If the component design is based on yield rather than

Fig. Comparison of Goodman, Gerber, and Soderberg models for relating

b. Rotating bending fatigue testing machine

Fatigue Fracture Mechanics:

The Figure represents a series of tests on a number of identical specimens

K  K max  K min   a ( max   min )CCF   a . .CCF

The following figure is plotted representing (da/dN) vs K presented on

Region (ii) Steady crack propagation:

Here C, and m are experimentally determined constants depending on

Region (iii): Unstable crack growth rate:

Safe life prediction:

The safe life prediction of a component with an initial crack, subjected to

Fig. Transmission electron fractograph of aluminum alloy laboratory

You might also like