基于充电需求时空分布预测的电动汽车充电站规划

Planning of Charging Stations for Electric Vehicles Based on Spatiotemporal Distribution Prediction of Charging Demand

电动汽车是实现能源消费侧清洁化的重要抓手。为了适应电动汽车的大规模推广,合理规划充电站布局,提出一种基于电动汽车充电需求时空分布预测的城区充电站规划模型。首先导入车辆、路网、交通等数据,利用蒙特卡洛法生成电动汽车初始电量、初始位置和起始时刻,引入转移概率矩阵,配合Dijkstra最短路径算法模拟电动汽车行驶过程。行驶过程中,结合单位能耗模型实时更新电动汽车电量,一旦达到充电阈值,就近充电并记录时段位置,据此生成城区充电需求的时空分布。在此基础上,综合考虑运营商和用户双方利益,以充电站投资成本和用户损失成本最小为目标,将Voronoi图与改进粒子群算法相结合,划分充电站服务范围,对其位置和数量进行寻优,并引入M/M/C排队论优化各站点充电桩数量。将上述模型应用到某市部分城区的充电站规划中,进行仿真分析。结果表明,建立的充电站规划模型能够有效表征充电需求时空分布规律,兼顾了运营商和用户双方利益,既考虑了投资的经济性,又考虑了充电的快捷性,可生成充电站最优数量、位置和容量,可为电动汽车充电站规划提供参考。

Electric vehicles are important for achieving clean energy on the consumer side.In order to adapt to the large-scale promotion of electric vehicles and rationally plan the layout of charging stations,a urban charging station planning model based on the spatiotemporal distribution prediction of electric vehicle charging demand is proposed.Firstly,vehicle,road network,traffic and other data are imported,and Monte Carlo method is used to generate the initial state of charge,initial location and initial time of electric vehicles.Transition probability matrix is introduced and combined with Dijkstra′s shortest path algorithm to simulate the driving process of electric vehicles.During the driving,the state of charge of the batteries for electric vehicles is updated in real time according to the unit energy consumption model.Once the charge threshold is reached,the electric vehicles are charged at the nearest charging stations and the time interval and location are recorded,so as to generate the spatiotemporal distribution of urban charging demand.On this basis,taking the interests of both operators and users into consideration and aiming at minimizing the investment cost of charging stations and the loss cost of users,Voronoi diagram is combined with improved particle swarm optimization algorithm to divide the service scope of charging stations,optimize their locations and numbers,and introduce M/M/C queuing theory to optimize the number of charging piles at each station.This model is applied to the planning of charging stations in some urban areas of a city for simulation analysis.The results show that the established charging station planning model can effectively characterize the spatiotemporal distribution of charging demand,take into account the interests of both operators and users,consider both the economy of investment and the rapidity of charging,and generate the optimal number,location and capacity of charging stations,providing reference for the planning of charging stations for electric vehicles.