Electric Vehicle Charging Load Prediction Based On Real-Time Road Traffic

As an environment-friendly transportation, electric vehicles have the characteristics of both transportation and mobile load, and they have great uncertainty in travel time and space, which are two important factors affecting the charging load prediction of electric vehicles. The prediction of EV charging load considering spatiotemporal distribution is of great significance for studying its interaction with the power grid and the location and capacity determination of EV charging stations. In this paper, an EV load prediction method based on residual neural network considering spatial and temporal distribution is proposed after analyzing the behavioral habits of three objects, namely, private cars, cabs and public vehicles, to start with. Firstly, a travel probability model is established based on the travel habits of users, then the real-time traffic flow predicted by the residual neural network framework taking into account the date, weather, temperature and other factors is used to invert the starting and finishing points of EVs, then the vehicle travel speed and power consumption are calculated based on the traffic flow, and finally the continuous change of battery charge state is considered, and the charging load model of EVs is established based on the Monte Carlo method.