Comparative analysis of offshore wind turbine blade maintenance: RL-based and classical strategies for sustainable approach

Abstract

This study compares traditional methods like Corrective Maintenance (CM), Scheduled Maintenance (SM), and Condition-based Maintenance (CbM) with Reinforcement Learning (RL)-based offshore wind turbine (OWT) blade maintenance strategies. In order to address the dual challenge of minimizing carbon output while managing maintenance costs and operational efficiency, the study presents a mathematical model intended to estimate carbon emissions associated with OWT maintenance activities. The ability of the RL-based strategy to reduce the risk of fatigue failure in OWT blades and account for wind speed variability in maintenance schedule optimization is assessed. In order to provide a sustainable maintenance solution this strategy balances the trade-offs between economic profit and environmental effect. The findings demonstrate how RL can provide a balanced approach to maintenance that enhances both operational performance and environmental sustainability.