A value network-based method for energy storage system scheduling optimization. The method uses a strategy of an energy storage system automatically adjusting an output plan thereof in the context of time variance of an energy value so as to achieve the maximum energy value: first performing rasterization on a 2D bounded state space enclosed by a time and an energy storage state; constructing a value network according to a time reverse sequence, wherein each unit in the network corresponds to a point of a state space; the value thereof calculating the maximum value from a state point to an ending state point of a scheduling cycle, an output plan corresponding to the maximum value chain recorded by a starting state point of the scheduling cycle being an optimal solution under rasterization accuracy; creating a state space with a finer granularity close to a low-accuracy solution of an output plan in a previous step, and repeating said process to drive solution convergence by means of repeated iteration until accuracy meets a requirement. The method has high solution accuracy, fast convergence, and good robustness, and may better ensure the economic efficiency and reliability of energy storage system regulation and control.
展开▼