Probabilistic character for S-N relations of high carbon-chromium bearing steel in super-long cycle life region

In order to accurately evaluate the fatigue S-N (stress-life) test data of high carbon-chromium bearing steel in super-long cycle life region, the entire S-N data and the S-N data based on differentiating fatigue crack initiation mechanisms were fitted by general maximum likelihood method respectively, its probabilistic S-N curve models were established. Analysis result shows that the test data can be well fitted and the S-N curve can be well depicted with differentiating fatigue crack initiation mechanisms. It is discovered that when load level is above the fatigue limit (about 1 200 MPa), the S-N curve that fatigue crack initiates from surface should be chosen, on the contrary, the S-N curve that fatigue crack initiates from subsurface should be chosen. It is proved that the material's probabilistic S-N curves, including survival probability-stress-life curves, confidence stress-life curves and survival probability-confidence-stress-life curves, can be determined, and the probabilistic characters for the S-N relations in super-long cycle life region can be depicted by general maximum likelihood method.