电力建设2025，Vol.46Issue(6)：24-37,14.DOI:10.12204/j.issn.1000-7229.2025.06.003
融合多方信息的电动汽车充电负荷时空分布预测
Prediction of Spatiotemporal Distribution of Electric Vehicle Charging Load Based on Multi-Source Information
摘要
Abstract
[Objective]Factors such as road networks,temperature,and electric vehicle(EV)type affect the spatial and temporal distribution of EV charging loads.To improve prediction accuracy,a spatiotemporal EV charging load prediction model is developed by integrating multiparty information.[Methods]By introducing the model of temperature and vehicle speed on energy consumption,the impact of the external environment on EV range is quantified.A charging demand gravity model is also used,incorporating factors such as station size,electricity price,time cost,and gravity parameters.These are used to dynamically adjust user behavior in choosing charging stations.Additionally,the Dijkstra algorithm is improved to plan charging paths more effectively by including real-time road condition data.Finally,the total charging load is accumulated and superimposed.[Results]The MATLAB simulation results showed a significant difference between the charging load distributions of private cars and cabs.The charging loads of private cars in residential,working,and commercial areas are concentrated during nighttime,daytime working hours,and off-duty hours,respectively.In contrast,the charging loads of cabs are characterized by morning and evening peaks,valleys,and small peaks at noon due to operational demand.The proposed improved Dijkstra's algorithm improves the efficiency of path planning by dynamically adjusting road section weights,reducing driving time by 3.9％for the same destination.The proposed charging demand gravity model optimizes users' charging station selection behavior by integrating factors such as charging station size,electricity price,and user time cost,resulting in a more reasonable spatial and temporal distribution of the charging load[Conclusions]This study constructed a spatial and temporal distribution prediction model for electric vehicle charging loads by integrating information from multiple sources.It reveals the differences in the charging behaviors of different types of EVs,their temperature sensitivity,and the dynamic characteristics of user decision-making.The results provide theoretical support for grid load scheduling,charging station planning,and the development of an orderly charging strategy.关键词
电动汽车/充电站/充电需求引力模型/改进Dijkstra算法/充电负荷Key words
electric vehicle/charging station/gravitational modeling of charging demand/improved Dijkstra algorithm/charging loads分类
信息技术与安全科学引用本文复制引用
王强,毕宇豪,高超,宋铎洋..融合多方信息的电动汽车充电负荷时空分布预测[J].电力建设,2025,46(6):24-37,14.基金项目
国家自然科学基金项目(52077120) This work is supported by the National Natural Science Foundation of China(No.52077120). （52077120）
