Bayesian method for reliability assessment based on zero-failure data under small sample size: Application for high speed railway vehicle

Abstract

In order to analyze the reliability of parts and components based on zero-failure data under small sample size, Bayesian method is applied with Weibull distribution to solve this problem. Gamma distribution and Uniform distribution are chosen as prior distribution and hierarchical prior distribution. And then, the hybrid MCMC algorithm is proposed to compute the estimators of posterior distribution. After that, taken air compressor in the braking system of railway vehicle as an example, the zero-failure data is assessed using the method given above. Computing result shows that the method can assess the reliability of air compressor based on zero-failure data under small sample size. © 2013 Asian Network for Scientific Information.