Periodic preventive maintenance planning for systems working under a Markovian operating condition

Systems commonly work under time-varying operating condition. Operating condition can influence the failure time of systems, which should be taken into account when do preventive maintenance (PM) planning. This study considers a one-component system working under time-varying operating condition, where the evolution process of the operating condition is governed by a Markov process. We assume that the failure time distribution of systems working under different operating condition can be modeled by an accelerated failure time model. A periodic PM policy is proposed to mitigate the increasing failure risk, where both imperfect maintenance (IM) and replacement are considered. The optimal combination of PM interval and number of IM actions before each replacement are determined by minimizing the long-term average cost, which is formulated based on the semi-regenerative properties of the system state. We also consider a special case where only two operating conditions exist. Numerical studies are conducted to validate the proposed maintenance policy.