Reliability assessment of few-failure data and zero-failure data for wheel axle of railway wagon

In order to evaluate the reliability of long-life high-reliability products, this paper proposed a framework for few-failure or even zero-failure lifetime data based on Bayesian method and Bootstrap resampling. The settings of hyperparameters in the prior distribution such as distribution and domain selection were detailed studied based on the E-Bayes method. Sensitivity analysis of these settings as well as the effects of censored time and sample size was carried out. The parametric Bootstrap resampling method is used to estimate the parameter interval. Simulation study and case study were employed to verify the feasibility and reliability of the proposed framework.