Research on Automatic Generation Control Technology for New Energy Grid Connection

The power industry takes the lead to achieve its carbon peak and carbon neutrality targets. The construction of the new power system with new energy as the main body is of great significance to the realization of the "dual carbon" goal. Nonetheless, the substantial integration of renewable energy sources into the power grid is anticipated to induce considerable disruptions, while conventional control techniques struggle to mitigate challenges like frequency disparities and poor CPS performance. Based on this problem, from the standpoint of automatic generation control, this paper proposes an automatic generation control algorithm for new energy grid-connected, that is, Self-Adaptive Reinforcement Exploration Double Speedy Q-learning algorithm(SARE-DSQL). The presented algorithm incorporates a self-adaptive reinforcement exploration approach into Speedy Q-learning to effectively balance exploration and exploitation. Additionally, it integrates the Double Q-learning strategy to address the issue of action exploration value overestimation within the Q algorithm framework, thereby enhancing algorithm convergence and update efficiency. Our study conducts simulations using the enhanced IEEE standard two-area load frequency control model to assess the efficacy of the proposed algorithm.