Bayes reliability demonstration test plan for series-systems with binomial subsystem data

Abstract

One reason that the Bayesian approach to reliability demonstration has not gained popularity in industry is the difficulty in establishing the prior. The problem becomes more complicated when only subsystem data are available. It has received little attention in the existing literature and this paper makes an attempt to do that. A method is proposed to derive the Bayesian reliability demonstration test plan for series systems with binomial subsystem data. The method makes use of Mann's approximately optimum lower confidence bound model to derive the system prior based on binomial subsystem data. The system Bayesian reliability demonstration test plan can then be derived using existing methods for meeting posterior confidence requirements. The proposed method is easy to apply and no complicated computation is involved in deriving the system prior distribution. It uses objective subsystem test data. No subjective judgement is required. This method is most beneficial for systems that already have substantial subsystem test data before the reliability demonstration.