基于出行链数据的电动汽车充电需求预测模型

An Electric Vehicle Charging Demand Forecast Model Based on Trip Chain Data

为解决电动汽车充电站规划过程中用户充电需求信息缺乏的问题,结合贝叶斯理论,构建了一种基于出行链数据的电动汽车充电需求预测模型。采用蒙特卡洛法,对路网用户充电需求参数进行模拟。研究结果表明:结合贝叶斯理论,可简便地获取用户产生充电意愿时的剩余电量与毗邻行程所需电量的概率分布;用户充电选择模型能模拟不同停车时长下的用户充电选择;基于历史出行链数据,所提出的充电需求预测模型能准确地预测用户充电需求参数,预测值概率误差处于±3%以内。该方法可为城市充电站选址和定容等领域的研究提供充电需求状态分布信息。

An electric vehicle charging demand forecast model based on trip chain data has been structured by integrating the Bayesian theory to address the lack of user charging demand information in electric vehicle charging station planning.The Monte Carlo Simulation is then used to simulate the charging demand profiles of users on road.The results showed that:When users’intention to charge develops,the joint probability distribution of remaining power and the power needed for the immediate trip could be simply obtained according to Bayesian theory.The user charging selection model can accurately simulate user’s charging selection under different parking duration.On the basis of historical trip chain data,the proposed model can quite accurately and simply predict the state of EV users’charging demand,and the probability error is within±3%.This method could provide charging demand ststus for research fields such as EV charging station location and sizing.