基于因子分析和灰色关联度分析法的变压器故障诊断  

Transformer fault diagnosis method based on factor analysis and grey relation degree analysis

作　　者：周涛[1] 胡海宁[2] 周力行[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]长沙电业局,湖南长沙410070

出　　处：《电力科学与技术学报》2013年第1期86-91,97,共7页Journal of Electric Power Science And Technology

基　　金：湖南省自然科学基金(10jj5059)

摘　　要：变压器油中溶解气体分析是目前有效的变压器状态诊断方法,应用灰色关联度算法处理变压器油色谱数据时,变压器故障标准模式的选取直接影响诊断结果.采用因子分析方法对大量确诊故障的油色谱数据进行分析计算,将变压器故障分为高温过热、高能放电等11类故障,并通过计算其标准差、方差以及M估计值,确定表征变压器各类故障的标准模式,该标准模式充分考虑了检测数据的分散性,减小了各类故障数据间的相关信息,使诊断具有更高的分辨率.结合灰色关联度分析法进行变压器故障诊断实例分析,结果表明该方法的诊断准确率高于传统的灰色关联度诊断法.Transformer oil dissolved gas analysis is currently an effective transformer fault diagno- sis method. Using the grey relation algorithm to deal with transformer oil chromatographic data, the diagnosis results are directly influenced by transformer fault standard model selection. In this paper, a large number of diagnosed oil chromatogram data was analyzed with the factor analysis method. The transformer faults were divided into 11 types, such as the overheating fault, the high-energy discharge fault. By calculating the standard deviation, variance and M estimate, the standard models of transformer faults were determined. Because of the standard model consider- ing the dispersibility of detection data and reducing, and reducing the related information between all kinds of fault data, the fault diagnosis resolution was improved. Combined with grey relation analysis method for transformer fault diagnosis, application examples showed that the diagnosisaccuracy of this method was higher than the traditional grey relation diagnosis method.

关 键 词：变压器故障 油色谱分析 因子分析法 灰色关联度分析 

分 类 号：TM41[电气工程—电器]