电动汽车换电需求时空分布的概率建模  

Probabilistic Modeling of Battery Swapping Demand of Electric Vehicles Based on Spatio-temporal Distribution

作　　者：段雪 张昌华 张坤 叶圣永 陈树恒 刘群英 吴云峰 DUAN Xue;ZHANG Changhua;ZHANG Kun;YE Shengyong;CHEN Shuheng;LIU Qunying;WU Yunfeng(School of Mechanical and Electrical Engineering,University of Electronic Science and Technology of China,Chengdu 611731,Sichuan Province,China;State Grid Sichuan Economic Research Institute,Chengdu 610041,Sichuan Province,China)

机构地区：[1]电子科技大学机械与电气工程学院,四川省成都市611731 [2]国网四川省电力公司经济技术研究院,四川省成都市611041

出　　处：《电网技术》2019年第12期4541-4549,共9页Power System Technology

基　　金：第50批国家教育部留学回国人员科研启动基金(计及配电网静态电压稳定性的电动汽车虚拟电厂运行管理策略研究);四川省科技计划项目(2019YFG0142);国家自然科学基金项目(51677020)~~

摘　　要：随着电力系统中电动汽车的高比例接入,换电作为电动汽车能源的重要补给形式受到广泛关注。电动汽车的移动具有时空随机性,换电需求也具有时空分布特性。针对这一问题,现有研究往往采用马尔科夫决策过程(Markov decision process,MDP)来计算汽车出行路径,即在每一个路口都以某一概率随机产生下一个目的地。但这种方式和人们的日常出行经验严重不符,即在熟悉的道路环境中,驾驶员路径的选择方式不是在每一个路口的MDP过程,而是事先有一条或多条候选路径,从中依概率选取一条。基于此,采用深度优先搜索(depth first search,DFS)和随机出行链确定了电动汽车1天的实际出行路径,完成了电动汽车出行空间分布规律建模;根据出行时间、停放时间等,确定了电动汽车在时间上的随机分布。通过时间和空间两个维度的结合,模拟电动汽车出行过程,为电动汽车的换电时刻、换电地点以及换电数量的确定提供了依据。最后,针对某一具体的交通网络和10000辆电动汽车,采用蒙特卡洛方法验证了所提模型和算法的有效性。研究成果可用于研究换电站的规划、交通规划以及对电网规划的影响等。With high proportion level of electric vehicles(EV) connected to power grid, battery swapping draws extensive attention as an important method for electric vehicle energy supply. The movement of electric vehicles is characterized with spatio-temporal randomness, and corresponding battery swapping demand also has the characteristics of spatio-temporal distribution. Therefore, to solve this problem, existing researches usually take Markov decision process(MDP) to build a vehicle travel route model, that is, when EV arrives at a road intersection, next destination is randomly generated with a certain probability. However, this method is seriously inconsistent with people’s daily travel experience,i.e. in a familiar road environment, the driver’s route choice is not consistent with MDP process at each intersection, but has one or more candidate paths in advance, from which one route can be selected according to certain probability. Thus, this paper introduces depth first search(DFS) method and random trip chains to determine travel route of electric vehicle in one day, modeling spatial distribution of electric vehicles. And then the random distribution in temporary dimension is determined by travel time, parking time, etc. Combining the two dimensions of time and space, this paper fully simulates travel process of electric vehicle, providing a basis for determination of battery swapping time, place and battery number. Finally, for a specific traffic network with 10,000 electric vehicles, the paper adopts Monte Carlo method to verify effectiveness of the proposed models and algorithm. The research results can be applied to study planning of battery swapping stations, traffic planning and EV charging load impacts on the grid schemes.

关 键 词：电动汽车 换电需求 深度优先搜索算法 蒙特卡洛法 出行链 

分 类 号：TM74[电气工程—电力系统及自动化]