Transient Frequency Security Preventive Control Optimization based on A3C

Loss of massive generation caused by HVDC blocking or tripping of large power plants is a severe threat to frequency security. Only taking emergency control measures after the accident may lead to a large number of loads being cut off. Therefore, preventive control for transient frequency security is considered in this paper except the emergency control. In preventive control, new units are turned on to increase the system's spinning reserve. It's difficult to decide which units to turn on to reduce preventive control costs with regular optimization methods. In order to solve this problem, the reinforcement learning Asynchronous Advantage Actor-critic (A3C) method is applied in this paper. Firstly, Markov Chain is used to model the wind power time sequences. Then A3C is used to optimize the preventive control strategy at different time internals. A simplified provincial power system of China is adopted to verify the validity of the proposed optimization method.