Data-Driven Electric Power Dispatching Equivalent Model Packing Method for Integrated Energy Systems

Abstract

In this paper, The electric power dispatching equivalent model (DEM) of the integrated energy systems (IES) and the global dispatching optimization issue with IES DEM integrated of the electric power system are studied. The electric power DEM of the IES is constructed as two models: one is the dispatching cost equivalent model (DCEM), and the other one is the dispatching feasibility equivalent (DFEM) model, considering that an electric power dispatching curve for IES is not always feasible. The DCEM is trained by deep neural network (DNN) model based on the data sets of the electric power dispatching curve and dispatching cost. The DFEM is trained by ensemble classifier model based on the data sets of the electric power dispatching curve of the IES and feasibility flag. Global dispatching optimization model is built with IES DEM integrated and solved by genetic algorithm. Case studies are carried out to verify the proposed electric power DEM and the global dispatching optimization model, and results show that the proposed IES DEM makes it possible for the power dispatching center to take advantage of the IES’s dispatchability so as to minimize the total operation cost of the entire electric power system.