Spatial-temporal Distribution Prediction of Electric Vehicle Charging Load Based on Charging Models

This paper proposes a method for spatial-temporal distribution prediction of electric vehicles (EVs) charging loads based on a charging decision model. Firstly, using the origin-destination (OD) probability matrix to identify the starting and ending points for the travel of private cars, taxis, and other urban functional vehicles, while the Dijkstra algorithm is applied to optimize driving paths. Subsequently, combining the node impedance and the road segment impedance to establish a dynamic road network model, and calculate the weights of each road. Simultaneously, constructing the energy consumption model to calculate the power consumption during their operation. Then, according to the remaining battery, users are classified into random and demand types, and a charging decision-making model of EV users is proposed. Then, based on the multi-source information mentioned above, the EV charging load forecasting framework is established. Finally, Monte Carlo method is used to simulate the travel processes and charging behaviors of EVs, resulting in the spatial-temporal Distribution of charging loads. Designed a typical and regional road network EV charging load forecasting case, verifying the feasibility and effectiveness of the proposed method.