Sensitivity of step‐stress accelerated fatigue‐life tests on type I censored data—An engineering perspective

Abstract

Step‐stress accelerated life testing is often used to estimate the fatigue life or reliability of structural components. Various methods exist for designing the optimal accelerated life test and estimating the parameters of the fatigue‐life or reliability models. Although they are based on rigorous mathematical formulations, they sometimes result in inappropriate estimates of the model parameters if the experimental data consists of censored data. This article presents a robust engineering procedure for estimating the parameters of the fatigue‐life curve and its scatter, which combines an inverse power law and the two‐parametric Weibull distribution. The presented methodology is tested on use cases that stem from the experimental fatigue‐life data. The results show that the proposed approach can yield a good estimate of the fatigue‐life model. Moreover, it can be further concluded that the sensitivity of the fatigue‐life model to the type I censored data strongly depends on the data scatter.