A Sensor-Driven Decision-Making Framework for Managing Degradation, Operations and Maintenance in Grid-tied PV Power Plants

This study presents an integrated sensor-driven framework for managing degradation, operations and maintenance in photovoltaic (PV) plants, with the objective of maximizing expected profit subject to maintenance costs. The model harnesses real-time sensor data that reflects degradation occurred in the performance of key components, including PV arrays, inverters, and transformers. It is formulated as a two-stage stochastic program in which power generation and the degradation levels of components are handled as uncertain parameters. The model simultaneously optimizes maintenance team routing, preventive and corrective maintenance scheduling, and decisions on the quantity of electricity dispatched to the grid. Its effectiveness is evaluated through a set of problem instances. The results highlight the usefulness of the sensor-driven OM model, showcasing that it can reduce total OM cost by at least 28.67% and increase total revenue by at least 9.74 % compared to the conventional periodic maintenance policy.