基于改进K-means的电能表时钟异常状态智能检测方法  

Intelligent Detection Method for Abnormal Clock State of Electric Energy Meter Based on Improved K-means

作　　者：陈红 CHEN Hong(Hunan Provincial Key Laboratory of Intelligent Electrical Measurement and Application Technology,State Grid Hunan Power Supply Service Center(Measurement Center),Changsha Hunan 410000,China)

机构地区：[1]国网湖南供电服务中心(计量中心)智能电气量测与应用技术湖南省重点实验室,湖南长沙410000

出　　处：《信息与电脑》2023年第2期67-69,共3页Information & Computer

摘　　要：为了最大限度降低因时钟状态异常给电能表性能带来的影响,提出基于K-means的电能表时钟异常状态检测方法。采用K-means聚类算法对原始的时钟运行数据进行预处理,利用阈值参量将时钟信号划分为噪声信号和正常信号两类,再计算待检测时钟信号中噪声信号的含量,并通过比较计算结果与时钟允许噪声含量之间的关系,实现异常状态检测。测试结果表明,设计方法实现了对时钟不同程度异常状态的准确检测,可靠性较高。In order to minimize the impact of abnormal clock state on the performance of electric energy meters,this paper proposes a detection method of abnormal clock state of electric energy meters based on K-means.K-means clustering algorithm is used to preprocess the original clock operation data,and the clock signal is divided into noise signal and normal signal by threshold parameter.Calculate the content of noise signal in the clock signal to be detected and realize abnormal state detection by comparing the relationship between the calculation result and the allowable noise content of the clock.The test results show that the design method realizes the accurate detection of different degrees of abnormal state of the clock and has high reliability.

关 键 词：K-MEANS 时钟异常状态 预处理 阈值参量 噪声信号含量 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]