Reliability modelling of stochastic wear-out failure

Abstract

This paper presents a reliability model for stochastic wear-out failure. The cumulative damage is modelled by a stochastic process and the time to failure is defined as the first passage time when the cumulative damage exceeds a certain prescribed threshold. Under some non-restrictive conditions, it was found that the time to failure can be approximated by an inverse Gaussian variate. This result has been confirmed by a comprehensive Monte-Carlo simulation. Thus, by evaluating the statistics of incremental damage over a fixed sampling time, the statistics of the time to failure and hence the reliability of the component under study may be deduced readily without having to resort to time-consuming life tests. A case study, based on the Archard wear theory in sliding-wear components, is presented to illustrate the model's applicability. © 1989.