Electric Vehicle Charging Demand Forecasting Model Based on Data-driven Approach

Electric vehicle (EV) charging demand forecast is an essential prerequisite for studying the interaction among electric vehicles and power grids together with transportation networks. Most of the existing researches have not utilized real-world traffic data to analyze the electric vehicle charging demand. On the basis of this, this paper presented an electric vehicle charging demand forecasting model based on data-driven approach. In this methodology, the original trip trajectory data of 'DiDi' online car-hailing, for the first time, were conducted to model via data mining and fusion. The processes of data analysis include region-scope selection, spatial grid modeling, trajectory data mapping, point of internet retrieval data identification and urban functional area clustering as well as traffic network modeling. Through the modeling and processing, the regeneration feature data of functional regional division, trip rule distribution and actual driving path were acquired. And then, the single EV model with driving characteristic parameters and charging characteristic parameters was established, considering the movable load feature of electric vehicles. Furthermore, the regeneration data obtained from modeling and analysis along with the determined single EV model were supported as data sources and models for the charging demand forecast architecture. At last, a certain area in Nanjing was selected as an example to conduct the path planning experiment and the charging demand load experiment under multiple scenarios. The results demonstrate that the proposed model is able to effectively predict the spatial-temporal distribution characteristics of charging demand loads in different date types and different functional areas, which lays a theoretical foundation for the subsequent research on charging control and charging guidance of electric vehicles as well. © 2020 Chin. Soc. for Elec. Eng.