Energy management in alternating current microgrids with renewable energy sources integration using giant trevally optimizer-self-adaptive physics-informed neural networks

This study addresses the challenges of energy management (EM) in alternating current (AC) microgrids (MGs) integrated with renewable energy sources (RESs), focusing on optimizing power balance, efficiency, and operational costs. Due to the alternation in renewable generation and demand, systems have to tackle the inefficiencies of power conversion and emissions related to the operation of backup generators. To overcome these issues, a novel hybrid approach, the Giant Trevally Optimizer-Self-Adaptive Physics-Informed Neural Network (GTO-SAPINN), is proposed. This approach aims to enhance system efficiency, minimize power loss, and reduce MG costs. In this method, the SAPINN forecasts demand and renewable generation patterns, ensuring stable energy supply. Meanwhile, GTO improves load balancing and distribution among RESs in AC MGs. The effectiveness of GTO-SAPINN is evaluated in MATLAB, compared against existing methods such as Beluga Whale Optimization, Flying Foxes Optimization-Deep Attention Dilated Residual Convolutional Neural Network, and Particle Swarm Optimization. Results reveal that GTO-SAPINN method achieves 99.1% efficiency at a total cost of €42 053, demonstrating superior cost-effectiveness and time efficiency over competing methods. This approach provides a promising, reliable solution for EM in AC MGs with RESs, optimizing energy distribution, and supporting sustainable MG operations.