基于多因素谱聚类算法的电动汽车群并网分析  

Analysis of Grid-connected Electric Vehicles Based on Multivariate Spectral Clustering

作　　者：谢义苗 余接永 张忠会[2] 叶惠芳 吴政国 何乐彰 

机构地区：[1]国网上饶供电公司,江西上饶334000 [2]南昌大学信息工程学院,江西南昌330031

出　　处：《华北电力技术》2017年第9期42-48,共7页North China Electric Power

摘　　要：针对包含各种类型的电动汽车群,将谱聚类方法应用于电动汽车群划分,提出了一种考虑多因素的电动汽车分类新方法。该方法以电动汽车群具有相同或相近运行特性为车组分群原则,应用谱聚类算法对电动汽车群的运行数据进行数据分析,找到电动汽车群之间实际运行特征的相似性,形成层次分明的分区方案,从而实现对电动汽车群的所有汽车的聚类划分。最后以IEE-RTS79系统为例,将不同类型的电动汽车并入电网,分析了电力系统电动汽车群并网对系统可靠性的影响,为今后实际电动汽车并入电网提供参考。For electric vehicles group which contains all types of electric vehicle, a method based on the spectral clustering considering multivariate technique is proposed. The spectral clustering algorithm analyses the electric vehicle operation data to capture the similarity among them in the dynamic operating process,and divides them into different groups. Finally,taking IEEE-RTS 79 system as an example, system reliability with grid-connected electric vehicles was analyzed,which can provide a theoretical basis for grid-connected electric vehicles.

关 键 词：电动汽车 谱聚类 多因素 

分 类 号：TM615[电气工程—电力系统及自动化]