فهرست نشریات علمی دسترسی آزاد ایران

The efficient and reliable operation of power systems is of critical importance, as unplanned outages resulting from unit failures can cause substantial financial losses, disrupt electricity supply, and compromise grid stability. Maintenance plays a pivotal role in ensuring operational continuity. Traditionally, Generation Maintenance Schedules (GMS) have been either time-based or reactive, which can result in premature servicing or unexpected equipment failures. The prediction of Remaining Useful Life (RUL) addresses these limitations by enabling a more precise estimation of when a component is likely to fail. This paper proposes a comprehensive framework for GMS based on RUL. The degradation states and remaining operational lifespans of thermal units are estimated using a deep learning approach based on Long Short-Term Memory (LSTM) networks. Additionally, a Robust Optimization (RO) method is employed to account for uncertainties in electrical load demand and renewable energy generation. Simulation results demonstrate that the proposed RUL-based GMS framework achieves a 2.41% reduction in total system operation and maintenance costs, highlighting its effectiveness in improving system reliability and cost efficiency.