Hybrid optimization tuned deep neural network-based wind power generation system for permanent magnet synchronous generator control

Wind energy, a cost-effective renewable source, has seen substantial growth. permanent magnet synchronous generator (PMSG) equipped wind turbines demonstrate superior performance in variable-speed applications. However, there remains a notable research gap in optimizing the overall system efficiency for such wind energy systems. Therefore, this research presents to develop a deep learning-based optimization technique that improves the efficiency of PMSG-based wind energy systems by minimizing overall system losses and maximizing energy output. Core loss and rotor speed data were fed into a deep neural network for various operating conditions ranging from 50 to 1000 rpm, to determine optimal system parameters. This work introduces a hybrid lyrebird-based coati optimization algorithm (LB-COA) to optimize the deep neural networks (DNN) classifier, combining two advanced optimization techniques to improve model performance. Simulation results validate that the proposed optimization strategy efficiently boosts the system's dynamic performance and overall power efficiency.