Multi-source Frequency Modulation Strategies Based on Stackelberg Game and Deep Adaptive Dynamic Programming

To address the issue of large frequency fluctuation in high penetration grid, a multi-source frequency regulation strategy of Stackelberg deep adaptive dynamic programming (SDADP) is proposed for smart generation control. Firstly, three deep neural network structures of the adaptive dynamic programming model are employed to acquire the characteristics from the historical data of the power grid and sequentially perform the prediction, evaluation, and execution actions for predicting the total frequency regulation power command, so as to minimize regional control error in a multi-source system. Based on the Stackelberg game, the power from wind, solar, hydropower, fire, and storage sources is subsequently reasonably scheduled to enhance the frequency regulation gain. Finally, the coordination optimization results of the proposed multi-source frequency regulation game strategy are verified through simulation analysis. The proposed SDADP multisource frequency regulation game strategy can mitigate the frequency deviation in the multisource system and improve the overall economic gain of frequency regulation.