Mechanism-Data Driven Improved ADP-Prediction Control for Renewable Energy Utilization in Microgrids

With the continuous development of renewable energy, enhancing its utilization efficiency has become a critical technical bottleneck restricting the optimal operation of microgrids. To address this challenge, this paper proposes a novel control algorithm for improving local renewable energy consumption. Firstly, an improved transformer network is adopted to establish a high-precision prediction model, which realizes accurate forecasting of photovoltaic (PV) output power and load demand. On this basis, an Adaptive Dynamic Programming (ADP)-based energy management controller is designed to optimize the charging/discharging scheduling of battery energy storage systems, thereby maximizing the absorption of local renewable energy. Furthermore, a mechanism-driven abnormal state regulation module is proposed to rapidly adjust the battery State of Charge (SOC) back to the safe operating range when abnormal conditions occur, while ensuring smooth mode switching of the microgrid to maintain stable system operation. Experimental validation results demonstrate that the proposed algorithm achieves a PV utilization rate of 99.41%, which effectively improves the economic benefits of the microgrid and enhances the overall system reliability.