The Method of Power Grid Fault Disposal Auxiliary Decision Generation Based on Generative Adversarial Imitation Agent

It is difficult to deal with the randomness and volatility of the new power system in the fault disposal mode of the power grid, and this paper proposes a method of power grid fault disposal auxiliary decision generation based on generative adversarial imitation agent. Firstly, a generator network is constructed, and the fault disposal auxiliary decision is generated by observing the power grid operation status and fault condition. The expert dispatch experience and sensitivity calculation results are used as expert strategies to guide the learning of the generator network. Then, the discriminant network is constructed to identify the generation strategy and the expert dispatch strategy and output discrimination results to assist generator network in updating itself. In the game confrontation between the generator and the discriminator, both of them reach the Nash equilibrium state, and power grid fault disposal auxiliary decision generation agent is built based on generative adversarial imitation learning. Through the operation data verification of a certain regional power grid, the proposed fault disposal auxiliary decision agent converges quickly in off-line training and has high decision efficiency in online application, which can quickly generate the power grid fault disposal auxiliary decision.