Bayesian analysis of fatigue life distribution including censored data
Nakagawa Takao; Yoshida Shou; Mura Junji
Abstract:In the present study, the Bayesian inference is applied to the fatigue life data including the run-out data (up to the specified number of stress cycles) in order to utilize the censored data for the estimation of fatigue life distribution. Assuming the prior distributions of the needed parameters and using the sample data of fatigue life from the Society of Material Sciences, Japan, the posterior distributions of the parameters and the predictive distribution of fatigue life are obtained. The effects of the sample size and the censored number of cycles on the predictive distribution are also discussed. Key Words:Bayesian analysis, censored data, fatigue life, prior and posterior distribution, predictive distribution