Research on Active Power Corrective Control of the Power System Based on Deep Reinforcement Learning Model

The application of artificial intelligence (AI) technology in active power corrective control (APCC) within power systems can significantly enhance decision-making efficiency, high-dimensional data processing capability, and the intelligence level of the power grid. By establishing mechanisms for safety operation decision support analysis and autonomous knowledge learning, AI can improve the power grid’s response speed and handling efficiency for abnormal and accident situations, thereby reducing the workload of manual dispatching. Advanced digital technologies such as AI will play a crucial role in APCC. Deep reinforcement learning, as an emerging AI technology, has been gaining widespread attention. This paper combines the perception capabilities of deep learning with the decision-making abilities of reinforcement learning to propose a competitive Dueling Double Deep Q-network (D3QN) model. Furthermore, it introduces an APCC method for power systems based on the D3QN model. When faced with "N-1" faults, source-load fluctuations, and other situations, this method can re-distribute the active power among the various generator units in the system, effectively eliminating the transmission line overload issues.