Short-Term Load Forecasting Model of the Power System Based on ICSO_RF_GRU Networks

Short-term load forecasting is one of the important bases for the operation of real-time power market. The improvement of forecasting accuracy is of great significance to revealing the characteristics of power consumption and providing real-time power system planning. Based on the rich historical data in power system, this paper proposed a method of short-term load forecasting model based on ICSO_RF_GRU network. The improved chicken swarm algorithm (ICSO) was used to optimize the number of decision trees and splitting characteristics in the random forest model, so that the performance of the random forest model is the best. Then, the random forest algorithm was used to fuse three GRU networks with different structures and perform group forecasting to get the load forecasting results of different groups. Finally, the forecasting results of each group were added to get the forecasting results. The historical load data of a prefecture level city power grid in Jiangsu Province is used for simulation analysis. Compared with the traditional prediction model, the prediction model proposed in this paper has better prediction accuracy.