基于改进Kendall-f算法的发电机定子温度分析  

Improved Kendall-f algorithm based generator stator temperature analysis

作　　者：刘博伟[1] 周力行[1] 陈冬良[2] 范彬[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410004 [2]广西电网有限责任公司来宾供电局,广西来宾546100

出　　处：《电力科学与技术学报》2015年第1期47-51,共5页Journal of Electric Power Science And Technology

基　　金：湖南省教育厅科学研究重点项目(13A106)

摘　　要：在线监测发电机定子温度可有效发现定子部分过热性故障,如何从检测的温度数据中提取反映定子故障预警的信息是目前国内外研究的热点。该文采用可变窗改进Kendall-f算法处理监测温度数据,提取其变化趋势特征信息;采用偏度分析提取监测温度异常值,从而正确诊断定子过热性故障,仿真分析及实验结果表明该方法有效提高了定子温度监测与预警的准确率。Online monitoring of generator stator temperature can effectively detect stator partial overheating fault.And how to extract the stator fault information from temperature data is a hotspot.This paper used the improved Kendall-f algorithm with variable window to deal with temperature monitoring data and extract the change trend of feature information.Skewness analy-sis was applied to extract the abnormal value of monitoring temperature and then correctly diag-nose the stator overheating fault.The simulation analysis and experimental results showed that the generator stator temperature analysis method could effectively improve the precision of the stator temperature monitoring and early faults warning.

关 键 词：定子温度监测 趋势识别 Kendall-f算法 偏度分析 

分 类 号：TM307.1[电气工程—电机]