Advanced hybrid deep learning based framework for microgrid inverter predictive maintenance

Abstract

The increasing complexity of microgrids (MGs) demands sophisticated strategies for improved maintenance and reliable operation. The integration of artificial intelligence (AI) into microgrids allows for the analysis of system performance, anomaly detection, malfunction identification, and the generation of alerts through continuous monitoring in case of any unexpected drop in performance enabling reliable operations and maintenance, enhancing predictive maintenance capabilities and sustainable decision making. To improve system reliability and performance, accurate fault identification and the generation of maintenance alerts according to the performances over time are becoming crucial for enhanced predictive maintenance (PdM) of MGs. This study introduces an intelligent data-driven hybrid framework for microgrid PdM utilizing a hybrid deep learning (HDL) architecture, combined with advanced data analysis and a residual-based dynamic threshold technique. The loss function of the proposed hybrid algorithm has been optimized to enhance microgrid PdM. The results show significant accuracy in predicting the maintenance needs within MGs. This research offers valuable insights for designing and enhancing hybrid algorithms for advanced maintenance in MG systems, contributing to advancements in PdM technology and promoting a resilient and cost-effective operation of MGs.