基于变异系数法和K⁃means的电能表状态评价方法  

State Evaluation Method of Smart Meter Based on Coefficient of Variation and K⁃means Method

作　　者：汪世平[1] 邹志杨[1] 陈洪涛 蔡慧 WANG Shiping;ZOU Zhiyang;CHEN Hongtao;CAI Hui(NARI Technology Co.,Ltd.,Nanjing 211106,China;College of Mechanical and Electrical Engineering of China Jiliang University,Hangzhou 310018,China)

机构地区：[1]国电南瑞科技股份有限公司,江苏南京211106 [2]中国计量大学机电工程学院,浙江杭州310018

出　　处：《山东电力技术》2022年第3期29-35,共7页Shandong Electric Power

摘　　要：智能电能表周期轮换存在着工作量大、计划性不强、重复检修等问题,如何合理地进行智能电能表状态评价是智能电能表轮换的关键环节。基于此,提出一种新的智能电能表状态评价方法。首先得到智能电能表的地区因素、可靠度、全事件、计量异常事件、电量过载和时钟电池欠压这6个指标数据;随后,一方面通过变异系数赋值法得到每个电能表的状态评分,另一方面将这6个指标数据作为输入数据,通过K⁃means聚类算法进行分类,得到相应类别。最后将两种算法结合得到新的智能电能表状态评价方法,并输出最后的评价结果。试验结果证明了该方法的科学性。There are many problems in the periodic rotation maintenance of smart meters,such as heavy workload,poor planning and repeated maintenance.Reasonable state evaluation of smart electric energy meter is the key link of smart electric energy meter rotation.A new state evaluation method of smart meter was proposed.Firstly,the six index data of smart electric energy meter were obtained:regional factors,reliability,smart meter events,abnormal measurement events,electric quantity overload and clock bat-tery undervoltage.Then,the state score of each electric energy meter was obtained by the coefficient of variation assignment method.The six index data were used as the input data,and the corresponding categories were obtained by K-means clustering algorithm.Finally,the two algorithms were combined to get a new state evaluation method of smart meter,and the final evaluation results were output.The experimental results show that the method is scientific.

关 键 词：状态评价方法 K⁃means 变异系数法 

分 类 号：TM7[电气工程—电力系统及自动化]