Predictive component availability optimization with degradation correlated maintainability

Driven by practical maintenance experience and the desire for higher component availability, this article presents a predictive maintenance policy for maximizing the availability of a batch of components (or single-component machines) in one repair cycle based on their real-time degradation signals, where the maintainability of preventive maintenance acts is explicitly correlated with the degradation levels accumulated before the preventive maintenance acts. The degradation correlated maintainability is modelled by a recently proposed proportional repair model. Within the proposed predictive maintenance policy, each component’s reliability is calculated and updated using its real-time degradation signals. The results from a numerical experiment based on a typical degradation model show that, for a batch of components whose maintainability of preventive maintenance acts is indeed correlated with the damage/degradation level accumulated before the preventive maintenance acts, it will lead to suboptimal preventive maintenance schedules if the effect of damage/degradation level on maintainability is ignored. In addition, the experimental results provide evidence of the superiority of t