METHOD FOR AUTOMATIC ADJUSTMENT OF POWER GRID OPERATION MODE BASE ON REINFORCEMENT LEARNING

A method for automatic adjustment of a power grid operation mode based on reinforcement learning is provided. An expert system for automatic adjustment is designed, which relies on the control sequence of thermal power units, enabling automatic decision-making for power grid operation mode adjustment. A sensitivity matrix is extracted from the historical operating data of the power grid, from which a foundational thermal power unit control sequence is derived. An overload control strategy for lines within the expert system is devised. A reinforcement learning model optimizes the thermal power unit control sequence, which refines the foundational thermal power unit control sequence and provides the expert system with the optimized control sequence for automatic decision-making in power grid operation mode adjustment. This method offers a solution to balancing and absorption challenges brought about by fluctuations on both the supply and demand sides in high-proportion renewable energy power systems.