Page 448 - Mechanical Behavior of Materials

448 Chapter 9 Fatigue of Materials: Introduction and Stress-Based Approach

Figure 9.31 S-N curves for zero-to-maximum bending, and residual stresses, for variously
shot peened steel leaf springs. (From [Mattson 59]; courtesy of General Motors Research
Laboratories.)

as many nonferrous metals, the stress–life curve is observed to continuously decrease as far as
test data are available. This can be seen for aluminum alloys in Figs. 9.4 and 9.27 and for brass
in Fig. 9.30.
The fatigue limit concept is widely employed in engineering design. Even materials that do not
have a distinct fatigue limit are sometimes assumed to have such behavior for design purposes. In
this case, the fatigue strength at an arbitrary long life is defined as the fatigue limit, where beyond
this life the stress–life curve decreases only very gradually. Such an assumption for a life of 5 × 10 8
cycles is the basis for the data of Fig. 9.25.
7
Due to the long test times that are required to reach lives beyond 10 cycles, most fatigue data
9
that are available are limited to this range. However, recent test data extending to 10 cycles and
beyond have exhibited the surprising behavior of a drop in the stress–life curve beyond the flat
7
6
region in the 10 to 10 cycles range. Such behavior has been observed in a number of steels and
other engineering metals, with one set of data being shown in Fig. 9.32. Detailed study reveals
that there are two competing mechanisms of fatigue failure: failure that begins from surface defects
and failure that begins from internal nonmetallic inclusions. The former dominates the behavior
7
up to around 10 cycles and exhibits the apparent fatigue limit, but the latter causes failures at
lower stresses and very long lives. Hence, where very large numbers of cycles are applied in
service, the concept of a safe stress may not be valid. See the book by Bathias (2005) for more
detail.