Fa#gue Failure

By far, the majority of engineering design projects involve machine parts subjected to fluctuating or
cyclic loads. Such loading induces fluctuating or cyclic stresses that often results in failure by fatigue.
Usually, the stresses are fluctuating and their nominal magnitude is well within the elastic range of the
material. About 80% of all mechanical component failure under service are due to fatigue.

Necessary Factors for Fatigue Failure

A sufficiently high tensile stress. Other Factors

Large fluctuaDons of the applied stress. Stress concentrators, corrosion, combined
A sufficiently large number of stress reversals stresses, overload, temperature, material
factors, residual stresses.
Historical Overview
1829 Albert, W. A. J. (1838) ”Uber First to carried out a fatigue analysis by establishing a correlation between
Treibseile am Harz” Archiv fu r̈ the applied loads (amplitude) and their time/number of cycles to failure in
Mineralogie Geognosie Bergbau und iron chains.
Hu T̈ enkunde, vol. 10, pp 215-34
1839 Jean-Victor Poncelet , Introduction a First to use the term fatigue or metals getting tired into books on analysis of
la Mecanique Industrielle Physique ou mechanisms to explain the time-dependent failure of ductile cast iron axels
Experimentale, 1841)
1843 William John Macquorn Rankine He showed the detrimental effect of stress concentrations in the life of the
component

1853 F Braithwaite, First to introduce the term fatigue in a research paper “On the Fatigue and
Consequent Fracture of Metals#, Proceedings of the Institution of Civil
Engineers ICE v.13, pp.463-475, 1853-1854
1860 August Wohler Carried out a systematic investigation on the failure of steel rail axles
subjected to alternating bending moment. He introduced the stress
amplitude-life (S-N) curve (Wohler-curves) to characterize the fatigue
behavior of the steel, he showed that higher stress amplitudes lead to
shorter fatigue life, he described the concept of endurance limit that is a
stress amplitude value below which there is no fracture and he stressed the
importance of amplitude and mean stresses.
1886 J. Bauschinger, 1886 Change in the elastic limit by often repeated stress cycles. ”Uber die
Veranderung der Elastizitatsgrenze und der Festigkeit des Eisens und Stahls
durch Streckenund Quetschen, durch Erwarmen und AbkuJhlen und durch
oftmals wiederholte Beanspruchung”. Mitt.
Mech.-Tech.Lab. Munch. 13, 1-115.
1910 Basquin O. H.,1910 “The exponential law of endurance tests”.
Proc. Annual Meeting, American Society for Testing Materials,
Vol. 10, pp. 625-630
1903 J.A. Edwin and J.C.W. Humfrey “The Fracture of Metals under
Repeated Alterations of Stresses” Phil. Trans. Roy. Soc, London
v.CC, 1903 p241
1874 H. Gerber (Gerber, H. (1874), Bestimmung der zulassigen
Spannungen in Eisen-konstructionen. Zeitschrift des Bayerischen
Architeckten und Ingenieur-Vereins 6, 101),
1899 Goodman, J. (1899). Mechanics Applied to Engineering. London:
Longmans Green
1924 Palmgren, A. (1924). Die Lebensdauer von Kugellagern. Zeitschrift
1945 des Vereins Deutscher Ingenieure 68, 339-41.
Miner, M.A. (1945). Cumulative damage in fatigue. Journal of
Applied Mechanics 12, 159-64.
1954 Coffin, L.F. (1954). A study of the effects of cyclic thermal stresses
on a ductile metal. Transactions of the American Society of
Mechanical Engineers 76, 931 and Manson, S.S. (1954). Behavior
of materials under conditions of thermal stress. National
Advisory Commission on Aeronautics: Report 1170. Cleveland:
Lewis Flight Propulsion Laboratory.
1921 Griffith, A.A. (1921). The phenomenon of rupture and flow in
1957 solids. Philosophical Transactions of the Royal Society, London
1961 A221, 163-97.
Irwin, G.R. (1957). Analysis of stresses and strains near the end of
a crack traversing a plate. Journal of Applied Mechanics 24, 361-4.
Paris, P.C, Gomez, M.P. & Anderson, W.P. (1961). A rational
analytic theory of fatigue. The Trend in Engineering 13, 9-14.
Represented the finite life region of the "Wöhler curves"
in the form (log σa) on the ordinate (log N) on the
abscissa and describe
Used optical microscopy to study the onset of fatigue
cracks. They observed that localized slip lines and slip
bands lead to the formation of microcracks
Investigated the effect of mean stress on the life of the
component
Investigated the effect of mean stress on the life of the
component
Damage accumulation model developed by Palmgren and
later by Miner
Provided an empirical relationship between the number
of load reversals to fatigue failure and the plastic strain
amplitude.
Provided a mathematical tool for quantitative treatment
of solids with cracks.
Showed that the amplitude of the stress ahead of a crack
could be expressed in terms the stress intensity factor, K .
Suggested that the increment of fatigue crack advance
per stress cycle, da/dN, could be related to the range of
the stress intensity factor, AK, during constant amplitude
cyclic loading
Fluctuating Stresses
All fluctuaDng stresses can be decomposed into a stress amplitude and a stress mean.
From ASTM E1823
fatigue—the process of progressive localized permanent
structural change occurring in a material subjected to
conditions that produce fluctuating stresses and strains
at some point or points and that may culminate in cracks
or complete fracture after a sufficient number of
fluctuations.
cycle—in fatigue, one complete sequence of values of
force (strain) that is repeated under constant amplitude
loading (straining).
 test whose loading mimics the expected cyclical loading that will be experienced by the material in the
0.25 application. This means that if the part will experience cyclical uniaxial tension, then the test data should
be generated in uniaxial tension. For expected bending or torsional stresses, the test data should be
The0 fatigue characterization of a material must
S-N Diagram begenerated
related toorthe
in bending torsion,application
respectively. that is the data for
0 0.5 1 1.5 2 2.5
fatigue characterization must be obtained Stress Ratio
from aNext, type of loading
the stress ratio, or R ratio,that mimics
of the test the expected
must be determined. This is simply thecyclical
ratio of the minimum
The loading that will be experienced by the
most extreme case would be if each cycle consisted of loading the test specimenmaterial inthisFatigue
to
(R Ratio) a given stress levelthe application.
stress experienced during a cycle to the maximum stress experienced during a cycle. It is important to note
(in bending or torsion), and then loading to an equal and opposite stress level in the other direction. In that and Stress Ratios (continued) At other times, a part such as an electro-mechanical
the stress values can be positive (tensile stress) or negative (compressive stress). In the easiest case to
case,This means
the minimum that
stress would if negative
be the the part will experience
of the maximum cyclical
stress, and the R ratio would be exactlyuniaxial
relay maythebetest
visualize, tension,
preloaded
specimen to astarts
certain
outthen
stress levelthe
unloaded test
that isstress),
(zero isdata
never removed toshould
loadedduringthethe life of be
required the part. The
maximum part
positive
Unidirectional
equal to -1. This is known as fully reversed (R=-1) testing. A rotating axle or shaft subjected stress to a level, and then unloaded to start the next cycle. In this case, the R ratio would be exactly zero. the
may then be loaded to the maximum stress level and then released back to the preload. In this case,
generated in uniaxial tension. For expected
(R=0) Testing bending
constant bending stress would be an example of this condition, as each point on the circumference ofisthe or
minimum torsional
stress would be stresses,
some small positive the
value test
associateddata
with theshould
preload,
called unidirectional (R=0) testing. A spring contact in a modular jack would be an example of this and be
the maximum
This

shaftgenerated in bending
would alternately experience or torsion,
tension and compression as it rotatesrespectively.
around the center. stress would
loading be a(assuming
condition larger positive
that thevalue. In isthisnotcase,
contact the R ratio
preloaded in anywould
way.)be greater than 0 and less than 1.
Fully Reversed Another possibility is to have a minimum compressive stress of lower magnitude than the maximum
1.5 (R=-1) Testing stress. This would result in an R ratio between -1 and 0. This could be the case in some switch and
R=-1 Fully Reversed Stress relay1.25contacts. R=0 Unidirectional Stress
1
1
0.5
Mininum Stress
1.25
0.75
R=0.5 Fluctuating Stress
Mininum Stress
Cyclical Stress
Fatigue
0 and Stress Ratios (continued) At other
Cyclical Stresstimes, a part such as an electro-mechanical 1
Maximum Stress
0 0.5 1 1.5 2 2.5 0.5 Mininum Stress
relay-0.5
may be preloaded to a certain stress level that is never
Maximum removed during the life of the part. The part
Stress
may then be loaded to the maximum stress level and then released back to the preload. In this case, the 0.25 References:
0.75 Cyclical Stress
Maximum Stress
minimum -1 stress would be some small positive value associated with the preload, and the maximum 0.5
stress would be a larger positive value. In this case, the R ratio would be greater than 0 and less than 1. 0
-1.5
Another possibility is to have a minimum compressive stress of lower magnitude than the maximum 0.25 0 0.5 1 1.5 2 2.5

stress. This would result in an R ratio between -1 and 0. This could be the case in some switch and 0
relay contacts. The most 0
extreme
0.5
case 1would1.5be if each
2
cycle
2.5
consisted of loading the test specimen to a given stress level
(in bending or torsion), and then loading to an equal and opposite stress level in the other direction. In this
case, the minimum stress would be the negative of the maximum stress, and the R ratio would be exactly
ush Inc.
1.25
R=0.5 Fluctuating Stress equal 1.25
to -1. This R=-0.5
is knownFluctuating
as fully reversed
Stress(R=-1) testing. A rotating axle or shaft subjected to a
1
constant1 bending stress would be an example of this condition, as each point on the circumference of the
Mininum Stress
shaft would alternately experience tension and compression as it rotates around the center.
0.75 Mininum Stress
0.75 Cyclical Stress
0.5 Cyclical Stress
Maximum Stress 1.5
0.5 0.25 R=-1 Fully Reversed Stress
Maximum Stress

10
0.25 The next issue of -0.25
0 0.5 1 1.5 2 2.5
Technical Tidbits will 0.5
-0.5 Mininum Stress
0
0 0.5 1 1.5 2 2.5
continue the discussion Please
0
-0.75 contact your local Cyclical Stress
on Fatigue. sales
0 representative
0.5 1 for 1.5 2 2.5
Maximum Stress
-0.5
further information on
One important point to note is that the all sources of stress will factor into the fatigue performance of a
1.25 R=-0.5 Fluctuating Stress the
-1 fatigue
component. or other
This includes the design stress from mechanical sources (applied forces, pressures, or
 Manufacture
Geometry Load Environment

Fatigue Design Methods History

Fatigue design is performed through

synthesis, analysis and testing knowing with
the knowledge of many input items such as Figure 2.1 Fatigue design flow chart originated by H. S
geometry, load history, environment, design Steel Corp. and slightly modified by H. 0. Fuchs. It was
criteria, materials properties, processing of Automotive Engineers Fatigue Design and Evalua
effects. Fatigue design requires selecting theUniversity of Iowa’s annual short course on Fatigue C
configuration, material, processes,
performing the stress analysis, choosing a
fatigue life model and a cumulative damage Manufacture
model and making an estimation of the life of
the component. Choosing the fatigue life model is a significa
suchSteel
models
Figure exist for design engineers. These are
2.1 Fatigue design flow chart originated by H. S. Reemsnyder from Bethlehem
Corp. and slightly modified by H. 0. Fuchs. It was created for use by the Society
Fatigue Life Models of Automotive Engineers Fatigue Design and Evaluation (SAEFDE) Committee-
University of Iowa’s annual short course on Fatigue Concepts in Design.

1. Theaccepted
For life predicDon, four different models are generally nominal stress-life ( S - N ) model, first fo
and 1870s.
Choosing the fatigue life model is a significant decision. Currently four
such models exist for design engineers. These are:
The nominal stress-life model (S-N) (1850-1870)
The local strain-life model (e-N) (1960)
2. The localstress-life
1. The nominal strain-life ( Efirst
( S - N ) model, - Nformulated
) model, betweenfirst form
the 1850s
and 1870s.
3. The
The linear elastic fracture mechanics LEFM or fatigue fatigue
crack
2. The local strain-lifecrack
growth model growth
(1960)
( E - N ) model, ( d d dinNthe- A1960s.
first formulated K ) mo
3. The fatigue crack growth ( d d d N - A K ) model, first formulated in the
The two-stage model or a combination of the local1960s.
strain-life
1960s.
model and the LEFM model
4. The two-stage model, which consists of combining models 2 and 3 to
4. The two-stage
incorporate model,
both macroscopic fatiguewhich consists
crack formation of and
(nucleation) co
 Methods to Determine Fatigue Life
There are two different material behaviors (regimes)
(a) Low-cycle FaDgue (LCF – Number of life-cycles < 103) Material subjected to large stresses.
(b) High-cycle FaDgue (HCF – Number of life-cycles >103) Material subjected to low stresses.

One of the objecDve of failure analysis is to predict the faDgue life of a mechanical component, subjected to specific
operaDng condiDons. There are three major faDgue life methods used in design, analysis and component-life
esDmaDon:
(a) Stress-life method (S-N Curves): It is used in HCF. It is a stress-based model. The model uses nominal stresses and
it relates them to the local faDgue strengths for a notched and unnotched members.
(b) Strain-life method (ε-N Curve): It is used in LCF. It is a strain-based model used when the cyclic load is high enough
to produce localized plasDc deformaDon. It relates the local strain at a notch, with the faDgue behavior of a smooth
specimen under strain-controlled faDgue.
(c) Linear elasDc fracture mechanics method (LEFM): Used to model the propagaDon of an exisDng crack (during
stage II). It is used when there is an exisDng crack or when the stresses are high enough to cause the formaDon of
localized cracks. It requires the use of fracture mechanics concepts, and the integraDon of the faDgue crack growth
rate equaDon to determine the number of cycles required to growth a crack from a given length to another length
and/or to fracture.
 Fatigue failure of a bicycle crank spider arm

Crack Initiation (Stage I): Fatigue always begins at a crack

usually formed by a high concentration of stresses or small
inclusions where the interface matrix/inclusion is weak.
Crack Propagation (Stage II): There is a high stress
concentration associated with a sharp crack. Each tensile stress
cycle causes the crack to grow (from 1nm to 2.5µm/cycle)
Sudden Fracture (Stage III): Catastrophic failure occurs
with no warning when the remaining material can not support
the stresses according to fracture mechanics.
 Typical fatigue fracture surface

Each clamshell marking might represent

hundreds or thousands of cycles.
Stages I, II, and III of fatigue fracture process

Stage I: Initiation/nucleation
Stage II: Stable growth
Stage III: Final Fracture
 ethod, the specimen is sectioned along a plane which is oriented at a small
a (i.e. a few degrees) to the specimen surface. The profile of the surface, as
Crack Initiation (Stage I):
ed on the sectioned plane, is magnified by a factor of (1/ sin a), over and above
gnification obtained by other visual methods such as scanning electron micro-
From the point of view of a material scienDst, they are likely to consider the nucleaDon of flaws along
optical microscopy and optical interferometry. Hunsche & Neumann (1986)
persistent slip bands as the iniDaDon stage of faDgue failure, whilst a mechanical engineer may
efined this technique to obtain sections with sharp edges normal to the speci-
urface so that the surface features can be recorded to a resolution of 20 nm in
associate the resoluDon of crack detecDon with the threshold for crack nucleaDon.
anning electron microscope. Basinski & Basinski (1984), Hunsche & Neumann
, and Ma & Laird (1989a,b) have utilized this method to examine the details of
Wood's postulated (1958) is that repeated cyclic straining of the material leads to different amounts of
e roughening in fatigued Cu. Their results collectively indicate the following
net slip on different glide planes. The irreversibility of shear displacements along the slip bands then
l trends:

results in the 'roughening' of the surface of the material. This roughening is manifested as microscopic
The surface of the fatigued crystal is covered with PSB extrusions, intrusions
and protrusions. A protrusion is a surface uplift (a large extrusion), many
'hills' andin'valleys'
micrometers height, whereatasites where
macro-PSB, tens slip bands emerge
of micrometers wide and at the free surface. The valleys so generated function
as micro-notches and the effect of stress concentration at the root of the valleys promotes additional
containing tens of matrix and/or PSB lamellae, emerges at the free surface.
A protrusion may contain several intrusions and extrusions. Figure 4.2 is a
slip and
scanning fatigue
electron crackobtained
micrograph nucleation
using the sectioning method which
142 Fatigue crack initiation in ductile solids

Fig. 4.5. Fatigue crack initiation (denoted by an arrow) at a PSB-matrix interface in a Cu

Fig. 4.2. Protrusions with extrusions and intrusions on the surface of a Cu crystal fatigued at crystal fatigued for 60000 cycles at ypl = 0.002 at 20 °C. (From Ma & Laird, 1989b. Copyright
room temperature for 120000 cycles at ypl = 0.002. (From Ma & Laird, 1989a. Copyright Pergamon Press pic. Reprinted with permission.)
Pergamon Press pic. Reprinted with permission.)
Initiation of a new crack in smooth polished
metals under cyclic load is caused by irreversible
dislocation movement leading to intrusions and
extrusions.
Strain localization occurs when dislocation pattern
in a few veins or bundles becomes locally unstable
at a critical stress or strain thereby leading to
formation of thin lamellae of persistent slip bands
or PSB’s.
In poly-crystals, the PSB’s are generally found on
the grains which have suitable orientations for the
slip to occur.
Since large number of fatigue cycles is consumed Fig. 9. Schematic of slip under (a) monotonic load and (b) cyclic load [21, p. 13]

in crack initiation, removal of surface layer at New crack initiates from the surface. Firstly, because the surface grains are i
Fig. 10. Crack initiation sites [21, p. 16]
contact with the atmosphere, thus if environment is a factor in the fatigue damag
frequent intervals enhances the fatigue life
manifold.
The ASTM Committee E-08 recently adopted the
www.intechopen.com
use of either “nucleation” or “formation” to
describe the early stages of the fatigue process,
and dropped the use of the word “initiation”.
 Crack Propagation (Stage II):
Since growth mechanism and propagaDon rate of a crack
differs according to its size, it is essenDal to define the crack
based on its size in increasing order.
i) Metallurgically or micro-structurally small crack which is
small as compared to a metallurgical variable such as the
grain size.
ii) Physically small crack in which the resistanceMetal toFatigue
crackand Basic Theoretical Models: A Review 217

growth by micro-structural barriers is averagedthey outcan but

not be it iswith the principles of continuum mechanics iii) Long crack that has
dealt
stable closure and can be treated by continuum mechanics.
not long enough to be called a long crack. Length of physically
small crack is of the order of 3 to 4 grain size. 217
l Fatigue and Basic Theoretical Models: A Review
iii) Long crack that has stable closure and can be treated by
can not be dealt with the principles of continuum mechanics iii) Long crack that has
conDnuum
e closure mechanics.
and can be treated by continuum mechanics.

Fig. 11. Three regimes of crack size [21, p. 419]

Taken from: Metal Fatigue and Basic Theoretical Models: A

Review by S. Bhat and R. Patibandla
School of Mechanical and Building Sciences Vellore Institute of
11. Three regimes of crack size [21, p. 419] Technology, Tamil Nadu, India
Long Crack Growth
PlasDc blunDng and damage accumulaDon at the crack
Dp are two different mechanisms that describe the long
crack growth in high cycle (HCF) and low cycle (LCF)
regimes.
PlasDc blunDng:
a. by applicaDon of tensile load, highly localized
plasDc deformaDon takes place along the slip
planes of maximum shear stress.
b. Upon further increase of load, the width of slip
band increases and the crack Dp blunts to semi-
circular shape.
c. As a result of blunDng, the crack extends to about
half the crack Dp opening displacement.
d. During compressive loading, the direcDon of slip is
reversed and the verDcal distance between crack
surfaces decreases.
e. The new surface created during tensile load is
partly folded by buckling into double notch at the
Dp. At the maximum compressive stress, the crack
Dp is sharp again that facilitates further crack
growth.
Crack Propagation
a large stress concentration is developed around the
crack tip and each time the stress becomes tensile the
crack grows a small amount
when the stress becomes compressive, zero or to a lower
tensile state, the growth of the crack stops
(momentarily)
this process will continue as long as the stresses at the
crack tip cycle below and above the sy of the material
crack growth is due to TENSILE stresses and grows
along planes normal to the maximum tensile stress
Factors that affect fatigue life
Microstructure: Large grained materials have low yield strength and reduced
fatigue limit;
Manufacturing Process: Anisotropy and residual stresses are important factors on
the estimation of the life of the component. Tensile residual stresses promote
crack nucleation, while compressive residual stresses reduce the chances of crack
initiation.
Size Effects: Under bending conditions, if the diameter is less than 10mm then the
S-N behavior is independent of the diameter or thickness, for larger sizes, the S-N
fatigue resistance decreases with size.
Component Geometry: Discontinuities such as holes, notches and joints are a
source of stress concentration and facilitate crack initiation.
Surface Quality: Rough surfaces enhance crack initiation.
Type of Environment: Some environments enhance crack nucleation and crack
growth rates.
Loading Conditions : Tensile mean stresses and multiaxial loads reduce fatigue
life.
Solutions:
a) Polishing (removes machining flaws etc.)
b) Introducing compressive stresses (compensate for applied tensile stresses) into thin surface layer
by “Shot Peening”- firing small shot into surface to be treated. Ion implantation, laser peening.
c) Case Hardening - create C- or N- rich outer layer in steels by atomic diffusion from the surface.
Makes harder outer layer and also introduces compressive stresses
d) Optimizing geometry - avoid internal corners, notches etc.
 Fatigue data is highly
variable and must be
described in an
statistical manner.
Fatigue failure is an
statistical event.
S

The S-N Curves are really showing the

104 105 106 107 probability of failure.
N
 El049 Cycle Counting in Fatigue Analysis
Smax +
Stress-Life Approach El 823
s,
Standard Terminology
=
2
Relating to Fatigue and Fracture Testing
Smin

This empirical approach was introduced by WohlerThe stress range, AS, is twice the alternating stress. Tensile
in 1860.
Wide-spread use in fatigue stresses are taken algebraically asstresses
positive and negative, respec
mean,analysis in applications
S,, maximum, where
S, , minimum, low-amplitude
&in,
Standard and range,
fatigue cycled
AS, of and
test methods stress induce
are indi-
procedures primarily
for metals are availa
ing stress isthese
an[4].absolute value.standards
The stress ratio, R, and the a
elastic deformation, incated.
a TEST
4.1 FATIGUE LOADING, component
MACHINES, that
The algebraic
AND is designed
relationships
SPECIMENS 61 for a
among
from long
ASTM life (HCF).
terms
These are:
pertinent are under the direction of AST
ratio, A, are used frequently in fatigue literature, where
Committee E-08 and are given in Table 4.1. Both ASTM standard designatio
+ The stress range, AS,and is titles
twice the in
are given alternating stress. Tensile
Table 4.1. International Organization orforcompressiv
Standardizati
stresses are taken algebraically as positive
(ISO) draft standards and negative,
on fatigue respectively.
testing of metals are availableAlterna
R is thecurrently
through t
stressavaila
ISO, Geneva, Switzerland. Three separate draft standards
ing stress is an absolute value.
are given The4.2stress
in Table ratio,
with both R, and
designations andthe
ratioalternating
titles. and A is
Additional I S 0stres
fatig
ratio, A, are useds frequently
, = Smax +2 Smin
testing draftin fatigue
standards literature,
are being where
developed.
y;: the alternating
e R = -1 and R = 0 are two common reference test con
6i
4.1.3 Fatigue
obtaining fatigueTest SpecimensR = - 1 is called
properties. stressthe ratio
“fully reve
0
Time Smin
sinceCommon is test
equal to -Smax;
specimens R = 0 fatigue
for obtaining wheredata
Smin 0 is called
are=shown “puT
in Fig. 4.5.
YLx I
-I
Onespecimens
cycle isshown in Fig. 4.5~-f
the smallest have beenofused
segment thetostress
obtain total
versusfatigue
timli
Stress ratios: [ =
A = S IS,,, that includes crack nucleation life and crack growth life. These specime
R =AND-1SPECIMENS
4.1 FATIGUE LOADING, TEST MACHINES, and R 65repeated are periodically,
= 0usually two common
have finely as shown
polished surfacesintoFig.
reference 4.2.conditions
test
minimize Under variable am
usedeffec
surface roughness fo
The stress range, AS, is twice the the definition
alternatingofstress. Tensile or compressive of
Figure 4.2 Nomenclature for constant amplitudeobtaining
cyclic loading.fatigue[4] properties. = one
- 1 cycle
R negative, is not
is called theclear andreversed”
“fully hence reversals
conditio
TABLE 4.1 ASTM Standard Practices stresses are
Related to taken
Fatigue algebraically
Testing of Metals as positive
considered. and
In respectively.
constant amplitude Alternat-
loading, one cycle equal
E466 since Smin is equal
ing stress is an absolute value.while
Conducting Force Controlled Constant Amplitude Axial Fatiguc Tests to -Smax;
The stress R = 0 where Smin
ratio,amplitude =
R, and theloading0 is called
alternating “pulsating
stress tension
in variable a defined cycle may contain
E467
of Metallic Materials
ratio, A, areOneused cycle is
frequently the
Verification of Constant Amplitude Dynamic Forces in an Axial Fatigue insmallest
fatigue segment
literature, of
wherethe stress versus time history that
TABLE 4.2 I S 0 Draft Standards Related to Fatigue Testing of Metals
Testing System repeated periodically, as shown in Fig. 4.2. Under variable amplitude loadin
, S,,
E468maximum, S, ,of
Presentation minimum, andFatigue
Constant Amplitude
&in, range,Test of stress
AS, Results for are indi-
Metallic ISO/DIS 12106 Metallic Materials-Fatigue Testing-Axial Strain-Controlled
. The algebraic the
relationships among these terms are:
Materials definition of one cycle is not clear and hence reversals of stress are ofte
Method
E606 Strain-Controlled Fatigue Testing considered. In constant amplitude
ISO/DIS 12107 loading,
Metallic one cycle
Materials-Fatigue equals two
Testing-Statistical reversal
Planning and
E647 Measurement of Fatigue Crack Growth Rates Analysis of Data
E739 while in( Svariable
Statistical Analysis of Linear or Linearized Stress-Life - N ) and amplitude
ISO/DISloading
12108 a defined
Metallic cycle mayTesting-Fatigue
Materials-Fatigue contain many reversal
Crack Growth
Strain-Life ( E - N ) Fatigue
R Data
= -1 and R = 0 are two common referenceMethod test conditions used for
El012 Verification of Specimen Alignment Under Tensile Loading (under the
s, Smax
jurisdiction of Committee
=
+ Smin
obtaining fatigue
E-28 on Mechanical Testproperties.
Methods) R = - 1 is called the “fully reversed” condition
El049 2
Cycle Counting in Fatigue Analysis
El 823
since Smin is equal to -Smax;R = 0 where Smin= 0 is called “pulsating tension.”
Standard Terminology Relating to Fatigue and Fracture Testing
One cycle is the smallest segment of the stress versus time history that is
repeated periodically, as shown in Fig. 4.2. Under variable amplitude loading,
the definition of one cycle is not clear and hence reversals of stress are often
 Stress-Life Method
(a) based on stress levels only
(b) Based on the amplitude stress –number of cycles S-N diagram
(c) Not very accurate prediction for low-cycle applications.
(d) Accurate prediction at high-cycle applications
(e) Most traditional method with plenty of supporting data.

S-N Diagram
The strength of the materials under fatigue
load are determined in the laboratory

Most widely used rotating

beam fatigue testing machine
(R.R. Moore)
 S-N Curve
Semi-log curve
Two different behaviors one for N<103 (LCF) and other for
N>103 (HCF)
The faDgue strength (Sf) iniDally starts at a value of Sut at N=1
and declines logarithmically with increasing cycles
Materials with Endurance Limit (Steel, Casts, Ti)
In some materials at 106–107cycles, the S-N diagram plateaus
and the faDgue strength remains constant.
The material behavior exhibit a sharp “knee”
knee
This plateau is called the endurance limit (Se) and is very
important since stresses below this limit can be cycled
indefinitely without causing a faDgue failure.
For pracDcal purposes Se is only esDmated for steels.
Design for infinite life at stresses < Se
Materials with no Endurance Limit (Al, SS, Cu, Polymers)
In some materials there is no a sharp knee at 106–107cycles,
the S-N diagram does not have a plateau and the faDgue
strength conDnuously decreases with the number of cycles.
For pracDcal purposes the design life is esDmated at 5x108
cycles.
Design for infinite life at stresses < SfaDgue
It is important to menDon that the curve does not separate
the crack formaDon from the crack propagaDon and only the
total life to fracture is given.
S-N Curves
The number of cycles dividing the LCF and HCF behavior
depends on the material (usually between 10-105cycles).
Plot samp versus number of cycles to failure.
S-N data are usually ploTed in a log-log scale and the
actual line represent a 50% probability for failure.
BCC materials have an endurance limit (se) which is a
stress level below which the material has an infinite life.
For engineering purposes, the infinite life is usually
considered to be 106cycles.
There are several empirical relaDonships between the
faDgue properDes of steels and the monotonic tension
(uniaxial tensile properDes).
sf=0.5TS
sf,1000=0.9TS
In the finite life region, a power relaDonship can be used
to esDmate the S-N curve sf=aNb known as the
Basquin’s equaDon.
samplitude
240
TS = Tensile Strength
220 TS Sm Sm = 0.9TS if load is pure bending
200
Sm = 0.75TS if load is axial loading
180
sN=a(2Nr)b
160 Sm = 0.72TS if load is torsional loading
140 or Sm=fxTS (see graph below)
120
Se Se = Endurance limit (Corrected)
100
80
60 The fatigue strength of the materials
40 or the endurance limit of the material
20 is related to the tensile strength of the
Log(Nc)
0
0 1 2 3 4 5 6 7 8 9 10
material.

The endurance limit

become constant in
materials with high
tensile strength
Mean Stress Effects
Usually fully reversed stresses are employed for determining the
S-N diagram. Define the amplitude stress (samplitude).
However, in any situation of cyclic stresses, two stresses can be
identified, namely, samplitude and smean.

There is a relaDonship between the number of cycles (N), smean

and samplitude . As
Mean Stress Effects
In general tensile mean stresses are detrimental to the
fatigue life of the component while compressive stresses
are beneficial.
Several empirical relaDonships have been developed to generate the line defining the infinite-life
design region.

Where sf is the true fracture strength.

All these expressions have been used in fatigue design when modified for
notches or stress raisers, size effect, surface finish, environment effect, type of
loading, temperature, reliability and finite life.
 Modified Goodman (Germany, 1899)
σ amp σ mean 1
+ =
Se TS SF

Gerber (Germany, 1874)

€ SF ⋅ σ amp $ SF ⋅ σ mean ' 2
+& ) =1
Se % TS (

Soderberg (USA, 1933)

€ σ amp σ mean 1
+ =
Se σ Yield SF

Langer Criteria (First Cycle Yield) ASME Elliptic

€
σ amp + σ mean 1 $ SF ⋅ σ amp ' 2 $ SF ⋅ σ mean ' 2
=
σ Yield SF & ) +& ) =1
% Se ( % σ Yield (

€
The criterium for yielding can
be modified to take into
account the yield strength (Sy)
in monotonic tensile test and
the cyclic yield strength (Sy’).
 Miner’s Rule
Most of the fatigue testing has been carried out at a constant amplitude fatigue loading. The problem
faced by the designer is how to use constant amplitude fatigue data in the prediction of fatigue lives
when the component is subjected to a wide range of variable amplitude load history (load spectrum).

In predicDng the faDgue life of ball bearings Palmgren (Palmgren A (1924) Die Lebensdauer von
Kugellagern (The durability of ball bearings). Verein Deutscher Ingenieure, 68: 339–341.) assumed that
the damage accumulaDon due to faDgue was linear with the number of stress cycles. Miner (Miner MA
(1945) CumulaDve damage in faDgue. J. Appl. Mech., 12: A159–A164. ) also suggested that the faDgue
damage at a given stress level could be considered to accumulate linearly with the number of stress
cycles. (Palmgren-Miner rule).

If a component that is subjected to stress S1 has a life of N1 cycles, then the damage accumulated per
cycle is 1/N1 and the damage after n1 cycles is n1/N1
n1 n2
For a two stress level that ends with failure at the second level, then + =1
N1 N 2
The generalization of this approach is called Miner's Law, and can be
written : nj
where nj is the number of cycles applied at a load corresponding to a S =1
lifetime of Nj . NJ
 Example Example
Example7.7: 7.7: Cumulative
Consider a component Damage
subjected to the following loading:
Cumulative Damage
20% of its life at175MPa,
30% of its life at 220MPa…… ;
Loading 20% @ 175MPa 30% @ 220MPa 40% @ 250MPa 10% @ 275MPa
Loading Example
Example 7.7:
20% @ 175MPa Cumulative
30% @ 220MPa
7.7:268,564 40%Damage
@ 250MPa
Cumulative 46,048
Damage 10% @ 275MPa
Life @ Load ∞ 12,367
Life @ Load
Loading 20%∞@ 175MPa 268,564
30% @ 220MPa 40%46,048
@ 250MPa 10% 12,367
@ 275MPa
Loading 20% @ 175MPa 30% @ 220MPa 40% @ 250MPa 10% @ 275MPa
Life
S-N Diagram @forLoad
Steel Specimen with ∞ut ==440MPa,
withSS
268,564 46,048 12,367
Life
S-N Diagram @for
Load
Steel Specimen ∞ 440MPa, Axial
ut
Load
Axial 268,564
Load 0 . 2 L
0.2 L + 0.3L 46,048
0 .3 L 0+.4 L0 .4 L 0+.112,367
L0.1L = 1
0 450
+ + + =1
S-N Diagram for Steel Specimen with Sut = 440MPa, Axial Load
S-N Diagram for Steel Specimen with Sut = 440MPa, Axial Load
∞ ∞
0.2 L 268268
,564
0.3L
,564 46 ,
0.4 L
46
048 , 04812 ,
0.1L
12
367 ,367
450 0.+2 L 0 .3 +
L 0 .4 +
L 00.1.=1
L1
0 400 450 0.02∞.2 268 0+.30,.564
3 46 0.4 120+.,1367
0+.4,048 1=1 1
++ ∞ 268+,564 + 46,+048+ 12,367 = =
400 ∞0∞.2 268 0,564
268 .3 ,564460,.048 46
4 ,048 120,.367
112,367 1L L
0 350 400 0+.2 0.3+ 0.4+ 0.1= 1
∞ + ,564 46+,048 12+,367 L=
268 1
350 00+ 1.117E
∞ 268
+ 1.117E + 8.687E
- 06-,564
06 - 06 -+12
46,048
+ 8.687E 8.086E
,367
06 -L06 =- 06 = 1
+ 8.086E
300 350 1L
1 L
Fatigue Strength (Sf)

0 0 + 1.117E - 06 + 8.687E - 06 + 8.086E - 06 =

300 0 + 1.117E - 06 +
1 1 8.687E - 06 + 8.086E - 06
L =
Fatigue Strength (Sf)

300 1.78899E - 05 = L
Fatigue Strength (Sf)

250 1.78899E - 05L1=

0 1.78899E - 05 = 1
250
L = 55,897 - 05L= L
1.78899ECycles
250 L
200 LL==55,897
55,897Cycles Cycles
0 200 L = 55,897 Cycles
200
150
0 150
150
100 Mat’l: Steel with Axial Loading
Mat’l:
Mat’l:Steel withwith
AxialAxial
Loading
0
100
100 Mat’l:Steel
Sut=440MPa.
Steel with Axial Loading
Loading
Sut=440MPa.
50 Sut=440MPa.
Sut=440MPa.
SL’=75%Sut =330MPa.
50 SL’=75%Sut =330MPa.
0 50 SL’=75%Sut
Se’=45%Sut
SL’=75%Sut =330MPa.
= 200MPa
200MPa
=330MPa.
0 Se’=45%Sut =
1 0 10 100 1,000 10,000 100,000 1,000,000 10,000,000
Se’=45%Sut ==200MPa
Se’=45%Sut 200MPa
1 0 10 100 1,000
10,000 100,000 1,000,000 10,000,000
0 1 10 No. 100
of Stress1,000
Cycles (N)10,000 100,000 1,000,000 10,000,000
1 10 100 No.
1,000 of Stress Cycles 100,000
10,000 (N) 1,000,000 10,000,000
No. of Stress Cycles (N)
No. of Stress Cycles (N)