| 引用本文: | 王浩林,张勇军,毛海鹏.基于时空特征变量数据分析的共享汽车充电负荷预测方法[J].电力自动化设备,2019,39(12): |
| WANG Haolin,ZHANG Yongjun,MAO Haipeng.Charging load prediction method of shared vehicles based on data analysis of spatiotemporal characteristic variables[J].Electric Power Automation Equipment,2019,39(12): |
|
| 摘要: |
| 共享汽车大规模应用将会给电网运行和充电设施规划带来新的挑战。目前对共享汽车充电负荷的预测方法研究不够深入，为此提出了一种基于时空特征变量数据分析的共享汽车负荷预测方法。通过数据挖掘，构建了由时空特征变量支撑的二维动态交通行为模型。为了探讨共享汽车连续充电与集中充电的特性，设定了连续充电和集中充电2种充电情形，以此构建充电行为模型。通过蒙特卡洛法模拟共享汽车的交通-充电行为，计算得到不同时间、不同区域下共享汽车充电负荷的预测结果，并分析负荷对电网的影响。仿真分析结果表明，交互影响的时空特征变量能够合理描述共享汽车时空二维不确定变化的特点，所提方法能对随机分散的共享汽车充电负荷做出科学预测，为电网及用户共享汽车负荷管理策略的制定提供有效的依据。 |
| 关键词: 共享汽车 数据分析 交通行为 充电行为 负荷预测 |
| DOI：10.16081/j.epae.201911009 |
| 分类号:U469.72;TM761 |
| 基金项目:国家自然科学基金资助项目(51777077) |
|
| Charging load prediction method of shared vehicles based on data analysis of spatiotemporal characteristic variables |
|
WANG Haolin, ZHANG Yongjun, MAO Haipeng
|
|
School of Electric Power, South China University of Technology, Guangzhou 510640, China
|
| Abstract: |
| The large-scale application of shared vehicles will bring new challenges to power grid operation and charging facility planning. At present, the research on the prediction method of shared vehicle charging load is not thorough enough, therefore, a load prediction method based on data analysis of spatiotemporal characteristic variables is proposed. Through data mining, a two-dimensional dynamic traffic behavior model supported by spatiotemporal characteristic variables is constructed. In order to explore the characteristics of continuous charging and centralized charging of shared vehicles, two charging scenarios of continuous charging and centralized charging are set up, based on which, the charging behavior model is established. The Monte Carlo method is used to simulate the traffic-charging behavior of shared vehicles, and the predictive results of charging load at different times and areas are calculated, and the influence of the load on the power grid is analyzed. Simulation analysis results show that the spatiotemporal characteristic variables of interaction can reasonably describe the characteristics of time-space two-dimensional uncertain changes of shared vehicles, and the proposed method can make scientific prediction on the randomly dispersed shared vehicle charging loads, providing an effective basis for the formulation of load management strategies for power grids and shared vehicle users. |
| Key words: shared vehicles data analysis traffic behavior charging behavior load prediction |
