We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
面向分布式电网的多区域协同控制方法研究.
- Authors
席磊; 孙梦梦; 陈宋宋; 朱继忠; 孙秋野; 刘宗静
- Abstract
To solve the stochastic disturbance caused by large-scale distributed power grid connection, from the perspective of automatic generation control, a deep reinforcement learning method is proposed, namely prioritized replay double deep Q network-B.ction weighting method (PRDDQN-A WM), which with the priority playback function and action space weighting optimization strategy. The method can speed up the optimization speed of the heuristic method in the process of balancing "exploration-utilization" and the sampling efficiency to obtain the optimal coordination of distributed multi-rngional grids in a strong random environment, thereby solving the random disturbance problem caused by large-scale grid connection of distributed energy, and making the distributed energy compatible with power system. Simulations on the improved IEEE-standard two--rngion load frequency control (LFC) model and Guangdong power grid model verify that the proposed method can obtain multi--rngion optimal coordinated control. Simultaneously, compared with other intelligent algorithms, it has faster convergence speed and better control performance.
- Publication
Electric Machines & Control / Dianji Yu Kongzhi Xuebao, 2021, Vol 25, Issue 12, p75
- ISSN
1007-449X
- Publication type
Academic Journal
- DOI
10.15938/j.emc.2021.12.009