Real-Time Cooperative Control Algorithm for Regional Power Grid Based on Long and Short-Term Memory Networks

With the increasing proportion of new energy resources connected, the power fluctuation of the regional power grid presents higher stochasticity property, and the real-time power balance faces greater challenges. To address this problem, a cooperative control model is constructed for real-time power balance in the regional grid with the high participation of multiple frequency regulation resources. In order to quickly generate the high- quality of cooperative control strategies, a long and short-term memory network (LSTM) is proposed to learn knowledge of the historical cooperative control strategies. Meanwhile, an infeasible solution correction method based on the ideal point method is proposed to modify the infeasible solution created by the network into a high-quality feasible solution. Finally, based on the extension model of the IEEE standard two-area frequency control model, the effectiveness and performance of the proposed method are verified by comparing various cooperative control optimization algorithms.