Short-term Electric Load Forecasting Of Integrated Energy System Considering Nonlinear Synergy Between Different Loads

As an efficient form of energy utilization,the integrated energy system flexibly integrates primary energy such as coal,oil,and natural gas,and converts it into electric,cooling,and heating through energy conversion equipment for users to use.Different from the traditional primary energy system,there are different load demands in the integrated energy system,and different load demand changes are often not independent.A load change will be transmitted as a signal to the system side and affect other load changes.This brings challenges to accurately forecasting the load of the integrated energy system.This article has done the following research work on how to apply the coupling relationship between different loads of the integrated energy system to improve the accuracy of load forecasting:Firstly,this article analyzes the energy consumption characteristics of the integrated energy system from different time scales such as quarterly,weekly,and hourly in the form of box plots,including electric,cooling,and heating single load characteristics and REC,DEC,REH,and DEH coupling correlation indexes characteristics,mining electric cooling and heating load change patterns and coupling correlation characteristics.Secondly,the interaction mechanism of the supply and demand side of the integrated energy system is analyzed,and the reason of the coupling relationship between different loads in the integrated energy system is explained fundamentally.The relationship between the electric,cooling and heating load is analyzed by the method of scatter distribution diagram and the maximum information coefficient,and the correlation between the electric cooling and heating load is proved by the calculation of an example.Based on this,the quadratic synergistic forecasting formula of the electric load which reflects the nonlinear synergy between different loads in the integrated energy system is proposed.Subsequently,based on the Stacking ensemble learning method,fusion of BP neural network,SVR,RF and GBDT,a short-term electric load forecasting model of an integrated energy system considering the nonlinear synergy between loads was established.Finally,through the experimental analysis of the integrated energy system project in Tempe campus of Arizona State University,it is found that the accuracy of the quadratic synergistic electric load forecasting is higher than that of the primary forecasting.In addition,two other forms of synergistic electric load forecasting formulas are constructed,and their performance in the validation set is compared with the synergistic electric load forecasting formula proposed in this paper.The results show that the MAE,RMSE and MAPE of the forecasting results of the synergistic electric load forecasting formula proposed in this paper are lower than those of the other two forms.It shows that the synergistic electric load forecasting formula proposed in this paper can improve the accuracy of the electric load forecasting of the integrated energy system.