基于相关性度量算法的台区线损异常判断及精准定位  

Judgment and precise location of abnormal line loss in station areabased on correlation measurement algorithm

作　　者：陈光宇 徐嘉杰 卢兆军 袁飞 张仰飞[1] 郝思鹏[1] CHEN Guangyu;XU Jiajie;LU Zhaojun;YUAN Fei;ZHANG Yangfei;HAO Sipeng(School of Electric Power Engineering,Nanjing Institute of Technology,Nanjing 211167,China;State Grid Shandong Electric Power Company,Jinan 250001,China;State Grid Shandong Electric Power Company,Taian Power Supply Company,Taian 271000,China)

机构地区：[1]南京工程学院电力工程学院,江苏南京211167 [2]国网山东省电力公司,山东济南250001 [3]国网山东省电力公司泰安供电公司,山东泰安271000

出　　处：《电力工程技术》2022年第4期67-74,共8页Electric Power Engineering Technology

基　　金：江苏省自然科学基金资助项目(BK20181021)。

摘　　要：针对台区发生线损异常时关联用户辨识困难的实际问题,提出一种基于相关性度量算法的台区线损异常判断及精准定位方法。首先,通过间隙统计-轮廓系数融合算法确定数据集的最佳聚类数,并在此基础上采用二分K-means++构建台区线损标准库;其次,基于标准库完成台区线损异常辨识,确定异常时间段;再次,计算异常时间段内各用户电量和线损的斯皮尔曼相关性系数(SCC)和欧式-离散弗雷歇距离(E-DFD),并基于SCC和E-DFD构造综合评判指标分析用户关联性;最后,采用逼近理想解排序法(TOPSIS)对综合评判指标值进行排序,实现异常关联用户的精准定位。算例采用某台区真实现场数据进行分析,结果表明文中所提方法在聚类有效性、计算时间以及辨识准确度等方面具有较好的性能。Aiming at the practical problem of the difficulty in identifying associated users when abnormal line loss occurs in the station area,a method for judging and accurately locating the line loss abnormality in the station area based on the correlation measurement algorithm is proposed.Firstly,the optimal clustering number of the data set is determined by the gap statisticscontour coefficient fusion algorithm,and on this basis,the dichotomous K-means++is used to construct the station area line loss standard library.Secondly,the station area line loss anomaly identification is completed based on the standard library and then the abnormal time is determined.The Spearman correlation coefficient(SCC)and Euclidean-discrete Fréchet distance(EDFD)of each user's power and line loss during the abnormal time is calculated.And based on SCC and E-DFD,a comprehensive evaluation index to analyze user relevance is estabilished.Finally,the technique for order preference by similarity to an ideal solution(TOPSIS)is used to sort the comprehensive evaluation index values to achieve precise positioning of abnormally associated users.The calculation example uses real field data in a certain area to analyze,and the results show that the method proposed in this paper has better performance in clustering effectiveness,calculation time,and identification accuracy.

关 键 词：台区线损 异常判断 精准定位 线损标准库 综合评判指标 逼近理想解排序法(TOPSIS) 

分 类 号：TM714[电气工程—电力系统及自动化]