Abstract

Maintenance plans aim to reduce the frequency of failures by maintaining components before their state of degradation reaches an undesirable level. These plans must balance the cost of performing maintenance against their benefits, i.e., a reduction in supply interruptions. This paper proposes a stochastic cost-benefit framework to develop maintenance plans that optimally balances these costs and benefits on an array of randomly generated scenarios in a power network. We implement this framework through optimization via simulation. We simulate the effect of various maintenance actions on the availability and degradation state of network components, failures due to human errors, and system operation (including demand curtailments). Based on these results, we assess the costs and benefits of these actions and find the best stream of Maintenance and Inspection (M) actions scheduled in time. We apply this approach to the power system supplying a large industrial facility to assess the impact of both offline and online inspections, as well as preventive and corrective maintenance actions. We take advantage of the flexibility of the proposed approach to quantify the potential benefits of online condition monitoring, analyze strategies to hedge against simultaneous outages, and study the relative value of asset redundancy and maintenance plans in enhancing reliability.