大数据背景下的充电站负荷预测方法

电动汽车负荷预测是充电站规划及调度的研究基础。相比传统的负荷预测,大数据背景下的负荷预测具有待预测数据可快速观测的特点,此时负荷预测方法需要相应调整。首先分析了充电站负荷预测所需数据及主要数据来源。其次,针对单辆电动汽车,基于大量、快速更新、多种类的数据分析电动汽车的充电习惯,预测每一辆电动汽车的充电开始时间、持续时间和充电地点,获取单辆电动汽车的负荷模型。该模型综合考虑电池状态、出行时间、行驶路径与速度、充电偏好等信息。然后,面向任意充电站,对与其相关的路网节点与交通线路上的所有电动汽车负荷求和,估算该充电站的总充电功率。最后,进行实例仿真,并与传统方法下的充电负荷预测结果进行了对比。

Load Forecasting Method for Electric Vehicle Charging Station Based on Big Data

The load forecast of electric vehicles(EVs)is the foundation of planning and scheduling of charging stations. Compared with the traditional method, the load forecast method under big data has the feature that the data to be forecast is quickly observable, real time, etc. Hence the need of the corresponding adjustments of the load forecast methods. This paper first analyzes the data demand for charging station planning and scheduling, and then the ways of main data acquisition. Based on volume, variety, velocity data, each EV''s start time, duration and location for charging, it will be possible to build the load model of a single EV. Furthermore, the total charging power of a charging station can be estimated by origin-destination(OD)flow statistics or adding up all the EV loads that are connected with its related transport line and node. Finally, a case study is given around the load forecast of EV station, and the load forecasting results from different load forecasting methods are compared. This work is supported by National Natural Science Foundation of China(No. 61104090)and National Key Technologies R&D Program(No. 2013BAA01B01).