The invention provides an
energy storage state-of-charge balance secondary control method based on security
reinforcement learning, which comprises the following steps: constructing a
state space, an action space and a reward function, establishing a strategy network and a
value network, designing an action space constraint mechanism through a
Nyquist stability criterion to meet the stability requirement of a
system, and obtaining a control strategy by utilizing
reinforcement learning interaction training, and deploying the control strategy in the
system to realize online control. According to the method, the depth deterministic strategy gradient agent is constructed, the control action is generated in combination with the real-time operation state information of the
system, the control target is to quickly realize the charge state
dynamic balance and the system frequency stability between the
energy storage devices, accurate modeling of a secondary controller is not needed, and the method has good adaptability and generalization ability and is suitable for large-scale popularization and application. The method can effectively improve the
energy storage operation efficiency and safety, and is suitable for the adaptive coordination control of the energy storage charge state in various
new energy power systems.