Evaluating the validity of rolling contact fatigue test results

The study described herein was conducted in order to validate the performance of a pair of two-headed rolling contact fatigue test machines acquired at a U.S. Navy facility. The purpose was to assess whether the machines yielded Weibull distributed fatigue lives with no systematic difference between the machines or between the heads within a machine. The test plan comprised of 24 runs on each machine yielding a total of 48 failure lives. The runs were balanced across machines, test heads, and specimens and thus allowed the identification of sources of possible non-Weibull behavior of the observed lives. The effects of these sources are not considered grave enough to invalidate test comparisons made in the usual way. They do raise the possibility, however, that if the design of the test machines could be improved to remove the sources of random variability, material comparisons could be made more precise, permitting material or lubricant effects on rolling contact fatigue to be detected in fewer tests.