Var-Voltage Control Capability Constrained Economic Scheduling of Integrated Energy Systems

Economic scheduling is an important solution to improve the operational efficiency of integrated energy system (IES). Active power is the majority being considered in economic scheduling, ignoring the influence of var-voltage effect on scheduling. In this article, an agent-based var-voltage control (VVC) capability is developed, and integrated in the economic scheduling modeling of IES. First, we build the VVC agent model based on the Markov decision process, and the economic scheduling model of IES is built with the decision function of the VVC agent as a constraint. Then, because the decision function is highly nonconvex and nonlinear, a deep learning enabled surrogate model is proposed to solve the scheduling problem. In the surrogate model, we train a pooling multilayer perception (P-MLP) to fit the mapping relationship between scheduling and maximum voltage deviation under VVC. And the trained P-MLP is used to surrogate the decision function constraint. After surrogating, the scheduling problem is transformed into a mixed integer programming problem, which can be solved directly by calling the solver. Finally, the effectiveness and accuracy of the proposed method are tested by cases. The results show that the scheduling integrated with the VVC capability model can strictly meet safety constraints and reduce the operation cost of IES as well.

Automated summary

This article introduces an agent-based var-voltage control capability to improve the economic scheduling efficiency of integrated energy systems and proposes a deep learning-enabled surrogate model to solve the scheduling problem. The results demonstrate that the proposed method can effectively and accurately meet safety constraints and reduce the operational costs of the system.