Regional Power Grid AGC Control Strategy Research Based on Q-Learning Algorithm Under WACPB Mode

Wind and coal power bundling(WACPB) mode is an important operation mode for large-scale wind power access to regional power grid, which is widely used in northwest and northeast China. Compared with the traditional AGC control mode, AGC control strategy of regional power grid under WACPB mode is significantly different on control objectives, control accuracy and so on. Aiming at the power flow control problem in regional power grid under WACPB mode, reinforcement learning theory and Q-learning algorithm are introduced. Based on Q-learning algorithm, AGC control strategy in regional power grid under WACPB mode is proposed of which the action space, reward function and other model points are defined. Finally, a case study based on a region power grid is used to prove the outstanding performance of Q-learning algorithm on adaptability and robustness.