Prediction of Spatio-temporal Distribution of Electric Vehicle Load based on Data-driven

Abstract:
The spatio-temporal distribution prediction of electric vehicles (EVs) charging load is the prerequisite and basis for analyzing the interaction between EVs and power grid. In order to improve the current deficiencies, which includes subjective parameter settings and user behavior analysis that is out of reality in the process of load forecasting, we use probability statistics to establish an analysis model for the charging behavior of different vehicle types, urban areas, and typical days. It bases on the actual data of the charging piles in Chongqing. Monte Carlo prediction model based on instantaneous charging probability is used to simulate the spatio-temporal distribution prediction of charging load. The results of comprehensive and regional load analysis are obtained, and the comparative analysis proves that the results of the regional forecasting method are more accurate. The vehicle charging behavior data fitting formula and load forecasting results in this article can provide a reliable basis for the construction of electric vehicle charging facilities and the optimal dispatching strategy of the power grid.