Load Frequency Active Disturbance Rejection Control for Multi-Source Power System Based on Soft Actor-Critic

To ensure the safe operation of an interconnected power system, it is necessary to maintain stability frequency and tie-line exchanged power. This one hottest issues in system field usually called load control. overcome influences disturbances on multi-source systems containing thermal plants, hydropower gas turbine we design a linear active disturbance rejection control (LADRC) based bias mode. For LADRC, parameter selection controller directly affects response performance entire not feasible manually adjust parameters. Therefore, obtain optimal parameters, use Soft Actor-Critic algorithm reinforcement learning parameters real time, reward function according needs system. We carry out simulation experiments verify effectiveness proposed method. Compared with results other proportional–integral–derivative techniques using optimization algorithms LADRC constant method shows significant advantages terms overshoot, undershoot, settling time. In addition, by adding different areas demonstrate robustness strategy.