A study of lifetime distributions for regularly replaced systems

Abstract

Usually lifetime distributions are used for systems that are not repaired or maintained. In practice, regular replacement is usually implemented and the actual distribution of time to failure in such a case is of great interest. In this paper, the use of exponential distribution assumption is investigated. For a number of commonly used lifetime distributions, their average failure rate expressions are derived under the assumption that the system is regularly replaced or maintained in such a way that it is considered to be good-as-new. Some analytical and numerical results are obtained. Such a study is useful for spare provision and maintenance decision-making. The interesting problem of optimal replacement policy is also investigated using Weibull distribution as an example.