摘要
在线监测数据在判断变压设备运行状态中发挥着重要的作用,但传统监测方法无法及时了解变压设备异常情况,准确甄别噪声数据。针对这一问题,挑选K-means聚类算法引进相应的滑动窗口技术,设计变压设备出现异常情况的检测模型。以某变电站变压器异常数据为对象,提出K-means聚类算法进行验证。结果验证,这种方式能够实时检测变压设备数据流中的异常信息,从而有效去除少量传感器产生的噪音或者突变值的影响,有利于提升变压设备状态检测的精准性及实时性,应用价值较高。
Online monitoring data plays an important role in judging transformer operation status,but traditional monitoring methods cannot timely detect transformer abnormal conditions and accurately screen noise data.In view of this problem,the K-means clustering algorithm is selected below to introduce the corresponding sliding window technology,and the detection model of abnormal transformer situations is designed.Taking abnormal transformer data of a transformer substation as the object,the K-means clustering algorithm proposed in this paper is verified.The results show that this method can detect the abnormal information in the transformer data stream in real time,so as to effectively remove the influence of noise or mutation value generated by a small number of sensors,which is beneficial to improve the accuracy and real-time detection of transformer state,and has high application value.
作者
杜涛
王朝龙
朱靖
赵健勃
马麒
刘勃君
DU Tao;WANG Chaolong;ZHU Jing;ZHAO Jianbo;MA Qi;LIU Bojun(State Grid Qinghai Electric Power Company Hainan Power Supply Company,Xining 813000,Qinghai China;State Grid Qinghai Electric Power Company,Xining 813000,Qinghai China;State Grid Qinghai Electric Power Company Information and Communications Company,Xining 813000,Qinghai China;Shanghai Metro Energy Technology Co.,Ltd.,Shanghai 200000,China)
出处
《粘接》
CAS
2022年第12期137-140,共4页
Adhesion
