Optimal Inter-area Oscillation Damping Control: A Transfer Deep Reinforcement Learning Approach with Switching Control Strategy
Published in Get the paper arXiv preprint arXiv:2301.09321, 2023
Recommended citation: S.Y. Liang, L. Huo, X. Chen, P.Y. Sun, "Optimal Inter-area Oscillation Damping Control: A Transfer Deep Reinforcement Learning Approach with Switching Control Strategy," arXiv preprint arXiv:2301.09321, 2023.
In this paper, we propose to use DDPG, which is one of the most classical Deep Reinforcement Learning (DRL) methods, to suppress the wide-area oscillation in power grids, and a physical-information-aided reward is designed to help better train the model. Then a novel Switching Control Strateggy (SCS) is proposed to achieve hybrid control and boost the performance of the DRL control, which achieves great outputs. According to the experiments, the DDPG which is trained in a linear environment performs well in both linear and nolinear environments, which show the outstanding transferability of the proposed method.