Fatigue Testing loading fixtures.

The load frame accommodates the

load train, consisting of the load cell, specimen grips
In simple terms, fatigue may be defined as the and the test specimen. The load frame also supports
progressive, localized, permanent structural change the actuator or loading device, as in the case of servo-
that occurs in materials when subjected to fluctuating mechanical or servo-hydraulic test systems. Figure 1
stresses and strains that may result in development of shows the schematic of a typical closed-loop servo-
cracks or fracture after sufficient number of cycles of controlled fatigue test system. The load cell is attached
fluctuations. to the rigid test frame on one side and to the load train
It is essential for the designers of safety-critical on the other side. The grips used to hold the specimen
structures subjected to dynamic loading to ascertain rigidly in the load-train can be of different type,
the soundness of their design, including predicted depending on the specimen.
design life, by evaluating the fatigue behavior of the Often, the type of grips that need to be used for a
material. There are a variety of fatigue testing systems specific test is specified by the corresponding ASTM
that can perform different types of tests on different test standard. Sometimes, special-type loading fixtures
types of specimens according to specified fatigue are employed to facilitate different types of testing,
testing standards, like those published by the In- such as the three-point\four-point bend test or a
ternational Standards Organization (ISO) and by the torsion test, by converting the axial force provided by
American Society for Testing and Materials (ASTM). the actuator in a uniaxial test system. The actuator or
load application unit is an electrical motor that
operates directly or through a gear-train as in the case
of the rotating beam fatigue test machine and in
1. Essential Components of Fatigue Test Systems electro-mechanical test system. In the servo-hydraulic
Fatigue test systems are designed to meet a wide test system, an electric motor drives an hydraulic
variety of requirements pertaining to fatigue data pump that generates high pressure for control of force
generation of materials and validation of structural and motion of a hydraulic actuator, fitted with a
integrity of new designs through fatigue testing. In servo-valve, depending on the demand from a test
physical appearance, various fatigue test systems may controller.
look different, but most of them have similar func-
tional components, such as a dynamic load or force
application mechanism, a control and test monitoring 1.2 Test Control and Monitoring Unit
or data acquisition unit.
A test system controller may be of analogue or digital
type, comprising electronic hardware and software to
control one or more of the following parameters of the
1.1 The Loading Mechanism
test system: stroke or position of the servo actuator,
This consists of a load frame, actuator or load load on the specimen and strain in the specimen. A
application device, load and displacement measuring servo-controller receives the desired inputs from input
devices, test specimen, specimen grips, and special devices such as a PC or a wave-form function genera-

Figure 1
Schematic of the essential components of a servo-controlled fatigue test system.

1
Fatigue Testing

tor and controls the test system through the servo- ents of such a test system have been described earlier.
actuator or the electric motor. It gets the feedback The closed loop electronic control system employed
on stroke or actuator position from the displacement allows such systems to function in load, stroke or
transducer, such as a linear variable differential trans- strain control mode. However, the loading frequency
former (LVDT), load feedback from the load cell and is normally restricted to about 1.0 Hz. Closed-loop
the strain feedback from an extensometer or a clip-on servo-mechanical fatigue test systems are sometimes
gage. The controller continuously monitors the sele- termed as ‘‘screw-driven test systems’’ and are pro-
cted test-control parameters during the conduct of a vided with specially designed vacuum chambers and\
test, compares the feedback signal with the desired or high temperature testing facilities along with axial\
control values and automatically adjusts the actuator diametric extensometers.
or servo-drive system to achieve the set values of the
control parameters. The controller is normally con-
nected to several output devices, including a PC, 2.2 SerŠo-hydraulic Fatigue Test System
through which the desired test data can be recorded on A servo-hydraulic test system is capable of executing a
a continuous basis or at specified time interval or at fatigue test under displacement control, load control
intervals in terms of number of cycles of loading. or strain control with high degree of resolution,
sensitivity, stability and reliability, at loading frequen-
2. Classification of Fatigue Testing Systems cies ranging from a fraction of 1 Hz to about 25–30 Hz.
The control system employed may be of analogue type
2.1 Electromechanical Test Systems or digital type that allows an output from the servo-
Electromechanical fatigue test systems are designed to hydraulic actuator that relates as closely as possible to
apply a large number of cycles of oscillating loads to a the programmed input to the controller. In the course
test specimen to determine the number of cycles to of a long duration fatigue test, it is likely that some
failure, under controlled cyclic loading conditions. electronic components may exhibit slight drift or their
The principal types of electromechanical fatigue test gain changes with time. Similarly, under the conditions
systems are briefly described below. of material cyclic hardening\softening or during fati-
gue crack growth, the compliance value of the speci-
men may change. The closed-loop negative feedback
(a) Rotating beam test system. The basic operating control system, generally employed in servo-hydraulic
principle of the rotating beam test system is the use test systems, is designed to make it insensitive to such
of an electric motor in order to rotate a shaft or a test variations and the system response depends primarily
specimen of specified dimensions around its longitud- on the system inputs and the feedback components.
inal axis. The specimen may be mounted either as a
simply supported beam or as a cantilever. On applica-
tion of a known static force through a set of bearings, 2.3 Ultrasonic Fatigue Test System
the resulting bending moment induces alternating High frequency fatigue testing can be performed by
tensile and compressive stresses of equal magnitude use of such a test system that involves cyclic loading of
on the outer surface of the test section in each revo- a test sample at frequencies in the range of 10 to
lution. Rotating beam test machines employ the 25 kHz. The principal advantage of using an ultrasonic
principle of complete stress reversal and are used for fatigue test system is greater reduction of time in
evaluating ‘‘fatigue limit’’ or ‘‘endurance limit’’. characterizing the fatigue limit. It consists of a sonic
Data recorded in such a test are essentially the maxi- energy converter, a series of acoustic amplifying horns,
mum stress applied on the rotating beam specimen computerized test control system, test specimen, a
and the number of elapsed cycles of loading until cycle-counter, test frequency display, and temperature.
failure of the specimen, Nf. The results from such a More details on ultrasonic fatigue test system and
test are normally presented in the form of a stress– their use for a variety of fatigue testing may be found
life (S–N ) curve, which is discussed in Sect. 3.8. in the handbook of the American Society for Metals
Often fatigue tests are conducted at very high (1985). The basic ultrasonic fatigue system is suitable
frequencies, typically in the order of 100–300 Hz with for generation of material stress–life (S–N ) data and
a view to reduce overall testing time, by use of fatigue crack growth, rate (da\dN) data for metals,
resonance test systems. These test systems are based composites, ceramics, glass and plastics within a short
on the use of spring–mass systems, in which the test time.
specimen is used like a spring and forms an integral
part of the oscillating mechanism. (see Ultrasonic
Fatigue). 2.4 Electromagnetic and Electro-dynamic Fatigue
Test Systems
(b) Closed-loop serŠo-mechanical fatigue test system. In axial fatigue test systems, the loading mechanism
The principles and functioning of essential compon- may consist of an electromagnetic or magnetostrictive

2
 Fatigue Testing

excitation system that can potentially reduce the high- predefined waveform. The total strain, representing
cycle fatigue (HCF) test duration by operating at the total span of the hysteresis loop, is kept within
higher frequencies compared to servo-mechanical or well-defined limits, and the resulting load is measured
servo-hydraulic test systems. However, these test to obtain the loop. If necessary a mean strain can be
systems are ideally suited to undertake such an axial superimposed over the above.
fatigue test that involves relatively lower magnitudes (ii) Plastic strain control: This test can be carried
of load and high fatigue life. Similarly, with an out keeping the total plastic strain, i.e., the width of
electrodynamic shaker employed as the drive system, the hysteresis loop, constant. A more modern scheme
HCF tests up to 10 million cycles can be completed in measures the plastic strain, by subtracting the elastic
less than 15 hours by operating at a frequency of strain from the total strain, and makes it follow a
200 Hz. Electro-dynamic shaker systems of fairly large predetermined waveform using the feedback control
force capacity of up to 50 kN are commercially loop.
available. However, their force rating decreases at
higher operating frequencies.
3.3 Stroke Control
2.5 Fretting Fatigue Test System In this case, the displacement of the actuator is selected
as the control variable, while load and strain data can
Fretting fatigue can occur in engineering assemblies be collected for the loop, or stroke itself can be used
when oscillatory loads are transmitted across a contact instead of strain.
interface between adjacent components. In a typical
fretting fatigue test system, the specimen is held
between one fixed and one movable jaw and is
subjected to sinusoidal varying stress. Typical ex- 3.4 Fatigue Life
amples of fretting fatigue include dovetail joints Fatigue life of any structure or component is the total
between aero-engine blades and disc, aircraft fuselage number of cycles of loading required to cause its
skin lap-splice owing to pressurization loading, etc. ultimate failure. The number of cycles to failure is a
However, fretting fatigue test systems may not be function of many variables, including nature of the
readily available off-the-shelf and may have to be wave form of cyclic loading, metallurgical condition of
custom-made to meet specific requirements (see the test article, magnitudes of maximum and minimum
Fretting Fatigue). stresses, test environment, etc.

3. Terminologies Related to Fatigue Testing 3.5 Applied Cyclic Stresses

3.1 Stress Control In constant amplitude fatigue testing the maximum
In stress control tests, the applied stress on the and minimum stress magnitudes applied to the test
specimen is controlled and the resulting strain is specimen are normally specified as Smax and Smin.
measured to obtain the hysteresis loop. Two types of Alternatively, the mean stress and stress amplitude or
stress control test may be performed: the stress ratio may be indicated. Mean stress Smean l
(i) Push–pull type: This type of test is used to study (SmaxjSmin)\2 and stress ratio, R l Smin\Smax. The
the response of a material under cyclic loading at amplitude of cyclic stress may be expressed as: Sa l
constant amplitude. A mean stress can be super- (Smax – Smin)\2 l Smax(1 – R)\2. In the case of fully
imposed on the waveform, to study the effect of mean reversed cyclic loading, the mean stress Smean l 0 and
stress in fatigue. R l –1. The frequency of cyclic loading is the re-
(ii) Ramp loading type: In this case, the specimen is ciprocal of time period of the applied waveform.
loaded with a sinusoidally varying stress, whose
amplitude is modulated by a ramp. This can be used as
a loading scheme, i.e., ramping the value of a maxi- 3.6 Fatigue Limit
mum stress until the desired level is reached and
continuing with a push–pull test. The fatigue limit or endurance limit is defined as the
maximum stress that a metallic material can withstand
for an infinitely large number of sinusoidal reversed
cycles of loading with 50% probability of failure
3.2 Strain Control
(American Society for Metals 1985). In a typical
The strain control test can be performed in two ways: stress–life (S–N ) curve, the fatigue limit strength is
(i) Total strain control: The strain on the specimen, represented by the horizontal portion of the curve,
as measured by an extensometer or a strain gage fixed whose magnitude varies for different materials (see
to it, is controlled by a feedback loop to follow a Sect. 3.8).

3
Fatigue Testing

versus number of cycles to failure data presented from

3.7 Low-cycle Fatigue
low-cycle fatigue testing (American Society for Metals
The low cycle fatigue regime is characterized by high 1985) is shown in Fig. 2 (see Fatigue and Thermo-
cyclic stress levels in excess of the endurance limit of mechanical Fatigue at High Temperature).
the material and is commonly accepted to be between
10% and 10& cycles. Generally, the method of presenting
low-cycle fatigue data is to plot in log–log scale the 3.8 High-cycle Fatigue
total strain range, ∆εtotal versus the number of cycles to
failure, Nf or the plastic strain range ∆εp versus Nf. In most structural components, designed for long
Low-cycle fatigue testing can be performed using operational life, the working cyclic stress levels are
servomechanical or servo-hydraulic test systems with generally below the yield strength of the materials. The
specially designed specimens and grips, as described in high-cycle fatigue data generated, either in a rotating
ASTM standard E-606 (American Society for Testing beam machine or in an axial push–pull machine, under
and Materials 1999). A sample plot of strain range fully reversed cyclic loading (R l –1.0) are graphically
presented as stress–life or S–N curves (American
Society for Metals 1985), as shown in Fig. 3. It can be
seen here that compared to a smooth specimen, a
total
specimen with a notch would fail at a much lower
number of cycles for a given stress amplitude. In the
Strain range

nearly horizontal portions of these curves, some data

points shown with arrows to the right are indicative of
‘‘run-outs’’, which implies test discontinuation at the
corresponding number of cycles. The asymptotes to
these nearly horizontal portions of the two curves
indicate the fatigue limit or the endurance limit.

No. of cycles to failure, Nf

3.9 Specimen Manufacturing
Figure 2 While manufacturing specimens for the HCF or LCF
A sample strain–life curve from low-cycle fatigue test. tests, it is important that the tolerances given for their
dimensions and the notch details in the relevant ASTM
Standards, such as E466 and E606 (American Society
for Testing and Materials 1999) are adhered to. For
flat specimens, end-flatness, end-perpendicularity, and
For un-notched end-parallelism should be maintained. Similarly, con-
steel specimens centricity of the outer surfaces in cylindrical speci-
mens, the surface finish in the notched region and
uniformity of dimensions in flat sheet specimens
Stress amplitude, Sa (MPa)

should be ensured.

4. Method of Specimen Gripping

In the low-cycle fatigue test with strain-control, the
cylindrical specimens are threaded at both ends, which
are gripped by the fixtures with the help of matching
nuts. For the tension–compression fatigue test, speci-
mens are held tightly in hydraulic grips with serrated
flat jaw-faces for flat specimens, or in hydraulic grips
with V-grooves for cylindrical specimens, so as to
avoid any backlash. The cylindrical specimens used in
rotating beam experiments are held in collet grips of a
rotating beam test machine.
No. of cycles to failure, Nf

5. Test System and Specimen Alignment

Figure 3
Sample stress–life (S–N ) curves for un-notched and In order to avoid large scatter in fatigue data, while
notched specimens. performing the fatigue test, it is necessary to check that

4
 Fatigue Testing

there is perfect alignment between the test system 7. Environmental Effects on Fatigue Testing
loading axis and the specimen axis, in the case of
Fatigue behavior of most engineering materials is
cylindrical specimens, and between loading axis and
adversely affected in the presence of any corrosive
mid-thickness plane in the case of flat specimens. In
environment, such as moist air or aqueous solutions,
the event of any error in the above alignment, the
saline or gaseous environment under ambient and
specimen may be subjected to bending loads that
elevated temperature conditions. It has been reported
would affect the fatigue life. As a rule, anti-buckling
in the literature (Barsom et al. 1971) that some high
guides are recommended for use in fatigue tests
yield strength steel specimens tested separately under
involving thin flat specimens under compressive load
vacuum, saline solution and gaseous hydrogen en-
cycles.
vironments exhibited two to three orders of magnitude
lower fatigue lives in the corrosive media, compared to
that under vacuum. Fatigue life degradation in steel
6. Fatigue Test Data Acquisition and Analysis owing to hydrogen embrittlement is well known. In
order to conduct a fatigue test under any given
It is necessary to perform the fatigue test on a test environment, one needs to use an environmental test
system with properly calibrated load cell, displacement chamber that permits the specimen to be completely
transducer (LVDT) and strain data acquisition trans- surrounded by the chamber along with heating or
ducer (extensometer or clip-on gage). Automated cooling facilities, while dynamic loading can be ap-
fatigue tests are normally conducted under stress- plied and specimen deformation or strain can be
control or strain-control conditions. It is therefore measured with the help of special-purpose exten-
necessary to acquire and store, in a test data file, useful sometry as schematically shown in Fig. 1 (see also
data in the form of load, displacement, strain, number Fatigue Cracks: Propagation of Short Cracks).
of cycles, crack length, temperature of the specimen, if
relevant, and test environment, etc. Analyses of the
acquired data are to be performed according to the
guidelines prescribed in the relevant test standards. 8. Fatigue Testing of Ceramics
The presentation of the data in the form of a stress– Structural ceramics are among the promising materials
life (S–N ) curve and a strain–life (ε–N ) curve for very high temperature applications. However, at
can be made as shown in Fig. 2 and room temperature they exhibit very low ductility, as
Fig. 3 respectively. Since fatigue data are likely to well as low fracture toughness. In view of the cost and
show considerable scatter, including ‘‘run-outs,’’ it difficult manufacturing process involved, it is not
is recommended to use a multiple specimen test matrix always possible to make flawless normal-size test
and carry out statistical analysis of test data. Further, specimens. Relatively small and smooth samples of
care has to be taken while predicting the life of any ceramics are normally employed for generation of
given structure, based on coupon level material data, high cycle fatigue or stress–life (S–N ) data (see also
so as not to over-estimate the structural life. Fatigue of Ceramics).

6.1 Scatter in Fatigue Data Bibliography

In the context of high cycle fatigue data generation, it American Society for Metals 1985 Metals Handbook, 9th edn.,
must be recognized that fatigue data are subject to Vol. 8: Mechanical Testing. ASM, Metals Park, OH
considerable scatter or variability. The scatter may American Society for Testing and Materials 1999 Annual Book
cause difficulty in interpretation of the results of any of ASTM Standards, Vol. 03.01: Metals—Mechanical Testing;
fatigue test programme unless the test results, in- EleŠated and Low Temperature Tests; Metallography. ASTM,
Philadelphia, PA
cluding ‘‘run-outs,’’ are properly handled through
Barsom J M, Imhoff E J, Rolfe S T 1971 Fatigue crack propa-
statistical analysis. If a large number of identical gation in high yield strength steels. Eng. Fract. Mech. 2,
fatigue specimens are tested at each of a number of 301–24
different stress levels, it is possible to draw a proba- Wirsching P H 1983 Statistical Summaries of Fatigue Data for
bility–stress–life plot. More information on this can be Design Purposes. University of Arizona, Tucson, NASA
found in the publication by Wirsching (1983) (see also Contractor Report 3697
Fatigue of Steels: Statistical Aspects, Fatigue Life
Predictions: Probabilistic Aspects). B. K. Parida

5
Fatigue Testing

Copyright ' 2001 Elsevier Science Ltd.

All rights reserved. No part of this publication may be reproduced, stored in any retrieval system or transmitted
in any form or by any means : electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or
otherwise, without permission in writing from the publishers.
Encyclopedia of Materials : Science and Technology
ISBN: 0-08-0431526
pp. 2994–2998