Deep Learning Optimized Global Adaptive MPC for Primary Frequency Synthetic Inertia Control in Microgrids with Wind Energy

As wind energy gains a larger footprint in global energy systems and demonstrates increasing utility in microgrid applications, the imperative for in-depth research into primary frequency regulation in microgrids incorporating wind power grows. Addressing this need, this paper proposes an innovative deep learning optimized Global Adaptive Model Predictive Control (MPC) algorithm, which is specifically designed to improve the primary frequency synthetic inertia control (SIC) in microgrids with integrated wind power. The paper first establishes a comprehensive response model for primary frequency SIC in such microgrids, incorporating mechanical energy compensation from wind turbines aimed at preventing secondary frequency drops. Then identify the limitations of the conventional MPC in controlling wind power’s contribution to primary frequency regulation. In response, the paper proposes the Global Adaptive MPC algorithm. This novel algorithm is engineered for real-time, adaptable adjustment of SIC parameters. Furthermore, the paper leverages cutting-edge deep learning techniques to expedite the MPC optimization process. This advancement significantly enhances the system’s real-time performance and stability. The efficacy and superiority of the proposed algorithms are substantiated through extensive experimental validation, which shows a marked improvement in the performance indices of primary frequency regulation.