Construction of an Optimal Scheduling Method for Campus Energy Systems Based on Deep Learning Models

Aiming at the problem of high cost and low efficiency of planning and scheduling caused by load uncertainty of campus energy system, a 3-layer planning and scheduling model based on multivariate load prediction is proposed, mainly including prediction layer, planning layer, and scheduling layer; a long-term and short-term prediction model of multivariate load is constructed based on random forest regression network and long and short-term memory network. With the objective of minimizing the comprehensive planning and scheduling cost and the scheduling operation cost, the optimal comprehensive system cost and configuration scheme are obtained by using improved particle swarm algorithm and CPLEX solver; the equipment status and system cost are analyzed by planning and scheduling under different scenarios. By comparing the planning and scheduling results of the constructed 3-layer model with the conventional two-layer model, the economy and reliability of the 3-layer planning and scheduling model are demonstrated.