Condition-based maintenance for a multi-component system in a dynamic operating environment

This paper develops a condition-based maintenance (CBM) model for a multi-component system operating under a dynamic environment. The degradation process of each component depends on both its intrinsic characteristic and the common operating environment. We model the environment evolution by a continuous-time Markov process, given which, the degradation increment of each component is described by a Poisson distribution. System reliability is firstly obtained, followed by a CBM policy to sustain system operation and ensure safety. In modelling the environmental effect on component degradation processes, two scenarios are considered. The first scenario considers renewable environment evolution while the second scenario on non-renewable environment evolution. The problem is casted into the Markov decision process (MDP) framework where the total expected discounted cost in the long-run horizon is utilized as the optimization objective to assess the policy. Structural properties of the optimal maintenance policy are investigated under mild conditions, which are further embedded into the value iteration algorithm to reduce the computational burden in calculating the maintenance cost. Applicability of the proposed model is illustrated through numerical examples.