Page 434 - Mechanical Behavior of Materials
P. 434
Section 9.5 The Physical Nature of Fatigue Damage 435
up to 100 kHz are possible with some of these special techniques. At such very high frequencies,
active cooling of the specimen is required to avoid overheating.
Modifications and elaborations of simple mechanical devices, as just described, allow fatigue
tests to be run in torsion, combined bending and torsion, biaxial bending, etc. Test specimens made
of thin-walled tubes may be subjected to cyclic fluid pressure to obtain biaxial stresses. All of the
test equipment described so far is best suited to constant amplitude loading at a constant frequency
of cycling. However, additional complexity can be added to some of these machines to achieve a
slowly changing amplitude or mean level.
Closed-loop servohydraulic testing machines (Fig. 4.4) are also widely used for fatigue testing.
This equipment is expensive and complex, but it has important advantages over all other types of
fatigue testing equipment. The test specimens can be subjected to constant amplitude cycling with
controlled loads, strains, or deflections, and the amplitude, mean, and cyclic frequency can be set to a
desired value by the electronic controls of the machine. Also, any irregular loading history available
as an electrical signal can be enforced upon a test specimen. Highly irregular histories similar to
Figs. 9.7 to 9.10 can thus be used in tests that closely simulate actual service conditions. Closed-
loop machines are often controlled by computers, and the test results monitored by computers.
In most of the test apparatus described, the frequency is fixed by the speed of an electric motor
or by the natural frequency of a resonant vibratory device. This fixed frequency is usually in the
7
range from 10 to 100 Hz. At the latter value, a test to 10 cycles takes 28 hours, a test to 10 8
9
cycles takes 12 days, and a test to 10 cycles takes almost four months. These long test times
place a practical limit on the range of lives that can be studied. If very long lives are of interest,
one possibility is to use a special high-frequency resonant vibration testing device. However, the
frequency may affect the test results, so it is not clear that an S-N curve obtained at, say, 20 kHz
can be applied to service loading at a much lower frequency.
9.4.2 Test Specimens
Specimens for evaluating the fatigue resistance of materials are designed to fit the test apparatus
used. Some examples are shown in Fig. 9.14, and two fractures from fatigue tests are shown in
Fig. 9.16. The simplest test specimens, called unnotched or smooth specimens, have no stress raiser
in the region where failure occurs. A variety of specimens containing stress raisers, called notched
specimens, are also used. These permit the evaluation of materials under conditions more closely
approaching those in an actual component. Notched test specimens are characterized by the value
of the elastic stress concentration factor k t .
Actual structural components, or portions of components, such as bolted or welded joints, are
often subjected to fatigue testing. Structural assemblies, or even entire structures or vehicles, are also
sometimes tested. Examples are tests of aircraft wings or tail sections, or of automobile suspension
systems. A test of an entire automobile has already been illustrated by Fig. 1.13.
9.5 THE PHYSICAL NATURE OF FATIGUE DAMAGE
When viewed at a sufficiently small size scale, all materials are anisotropic and inhomogeneous.
For example, engineering metals are composed of an aggregate of small crystal grains. Within each

