Maintenance strategy optimization for complex power systems susceptible to maintenance delays and operational dynamics

Maintenance is a necessity for most multicomponent systems, but its benefits are often accompanied by considerable costs. However, with the appropriate number of maintenance teams and a sufficiently tuned maintenance strategy, optimal system performance is attainable. Given system complexities and operational uncertainties, identifying the optimal maintenance strategy is a challenge. A robust computational framework, therefore, is proposed to alleviate these difficulties. The framework is particularly suited to systems with uncertainties in the use of spares during maintenance interventions, and where these spares are characterized by delayed availability. It is provided with a series of generally applicable multistate models that adequately define component behavior under various maintenance strategies. System operation is reconstructed from these models using an efficient hybrid load-flow and event-driven Monte Carlo simulation. The simulation's novelty stems from its ability to intuitively implement complex strategies involving multiple contrasting maintenance regimes. This framework is used to identify the optimal maintenance strategies for a hydroelectric power plant and the IEEE-24 RTS. In each case, the sensitivity of the optimal solution to cost level variations is investigated via a procedure requiring a single reliability evaluation, thereby reducing the computational costs significantly. The results show the usefulness of the framework as a rational decision-support tool in the maintenance of multicomponent multistate systems.