| 引用本文: | 陈丽丹,张尧,Antonio Figueiredo.融合多源信息的电动汽车充电负荷预测及其对配电网的影响[J].电力自动化设备,2018,(12): |
| CHEN Lidan,ZHANG Yao,Antonio Figueiredo.Charging load forecasting of electric vehicles based on multi-source information fusion and its influence on distribution network[J].Electric Power Automation Equipment,2018,(12): |
|
| 摘要: |
| 电动汽车充电负荷具有时间和空间不确定性、随机性，提出一种融合路网、交通、电网、天气、车辆、充电设施等多源信息的考虑用户出行行为和充电需求的电动汽车充电负荷时空分布预测模型。由图论方法构建城市路网和电网信息模型及两者的耦合关系；引入出行链，以概率函数拟合车辆首次出行时间和行程目的地的驻留时间，采用Dijkstra算法规划车辆的出行路径以获得各段行程距离，由道路等级和各时段交通信息获得车辆的行驶速度，以计算行程行驶时间和荷电状态，再根据各行程目的地的充电需求判断条件，计算充电时长和充电负荷；采用蒙特卡洛方法对各功能区电动汽车出行的时间和空间充电负荷分布进行整体仿真；并根据耦合关系将充电负荷归算至对应电网节点，再通过时间序列潮流计算评估电动汽车接入电网后无序充电对电网负荷、电压和网损的影响。算例通过设置不同的场景预测了不同功能区和电网节点的充电负荷曲线，分析了不同因素对充电负荷分布及电网的影响，验证了所提模型的有效性。 |
| 关键词: 电动汽车 多源信息 充电负荷预测 路网-电网 时空模型 配电网 蒙特卡洛方法 Dijkstra算法 |
| DOI：10.16081/j.issn.1006-6047.2018.12.001 |
| 分类号:TM761;U469.72 |
| 基金项目:国家自然科学基金资助项目(61603141)；广东省普通高校青年创新人才自然科学项目(2015KQNCX229)；国家留学基金资助项目(201708440511) |
|
| Charging load forecasting of electric vehicles based on multi-source information fusion and its influence on distribution network |
|
CHEN Lidan1,2, ZHANG Yao2, Antonio Figueiredo3
|
|
1.School of Electrical Engineering, Guangzhou College of South China University of Technology, Guangzhou 510800, China;2.School of Electric Power, South China University of Technology, Guangzhou 510640, China;3.Department of Electronic Engineering, University of York, YO10 5DD, UK
|
| Abstract: |
| The charging load of EVs(Electric Vehicles) has spatial and temporal uncertainty and randomness. A spatial-temporal distribution forecasting model for EV charging load is proposed, which considers the EV users' travel behavior and their charging demands and fuses multi-source information such as road network, transportation, power grid, weather, vehicles, charging facilities and so on. The information models of urban road network and power grid and their coupling relationship are established by using the graph theory, the first trip time and the dwell time at destination of EVs are fitted by probability functions with the trip chains, the distance of each trip is obtained by app-lying the Dijkstra algorithm to plan the travel routes of EVs, the travel speed of EVs are obtained based on the road grade and the traffic information of each time period, then the traveling time and the state of charge are calculated. According to the judging condition of charging demand for each destination, the charging time and charging load are calculated. The spatial and temporal charging load distributions of EVs in each functional area are simulated by Monte Carlo method. The charging load is reduced to the corresponding power grid node according to the coupling relationship, and the influences of disordered charging of EVs in the power grid on the power grid load, voltage and power loss are evaluated by the time series power flow calculation. The case study verifies the effectiveness of the proposed model. By setting different scenarios, the EV charging load curves of different functional areas and grid nodes are forecasted, and the influences of different factors on the charging load distribution and the power grid are analyzed. |
| Key words: electric vehicles multi-source information charging load forecasting road network and power grid spatial-temporal model distribution network Monte Carol method Dijkstra algorithm |
