Research and application of load forecasting method for park integrated energy system based on XGBoost algorithm

Abstract Load forecasting is the basis of energy management and optimal dispatching park-integrated system, its accuracy directly affects overall operation performance system. Therefore, this paper proposes a load method for system based on XGBoost algorithm. This first establishes simulation analysis model photovoltaic power generation electric vehicle charging discharging in park then carries out supply demand forecast characteristic modeling park’s Finally, constructs prediction optimization control flow uses mean absolute percentage error (MAPE) root square (RMSE) as two evaluation indicators. The validity proposed verified by data actual micro park.