EV Charging Behavior Simulation and Analysis Using Real-World Charging Load Data

Abstract

Accurate short-term load forecasting (STLF) results for charging station charging load are essential to improve the reliability and safety of power distribution systems. In modern electric vehicle (EV) optimization dispatching, in addition to accurate STLF results, it is necessary to know the status of each user's EV, and user charging behavior promptly. In this paper, an EV charging STLF method is proposed and discussed. We first derive accurate STLF results based on a data-driven method. Using the driving distance probability function of vehicle and Monte Carlo method to generate a certain number of different states of EV. Using the two-stage filling method, the number of the EVs charging at each time point in the first step is set as the model's parameters, and then the model is enabled with multiple times to retain the results which are close to results obtained by the data-driven method. The number of the EVs charging at some time point in the second step are set as the model's parameters. Based on the first step, the load is filled, and the accurate model-driven STLF results and charging behavior information are finally obtained. Thus, this information is used to regulate the charging behavior of EVs.