基于因子分析和基因表达式编程算法的变压器故障诊断  

Transformer fault diagnosis based on factor analysis and gene expression programming

作　　者：武中利[1] 董卓[2] 俞国勤[3] 朱永利[1] 

机构地区：[1]华北电力大学电子与电气工程学院,河北保定071003 [2]湖南省电力公司检修公司 [3]上海电力公司,上海200025

出　　处：《华北电力大学学报（自然科学版）》2012年第4期47-51,共5页Journal of North China Electric Power University：Natural Science Edition

基　　金：河北省自然科学基金资助项目(E2009001392)

摘　　要：将因子分析和基因表达式编程结合起来,对变压器油色谱数据进行处理和智能训练,并建立变压器故障分类诊断模型;收集到332组油色谱数据,选取150组作为GEP分类器的训练样本,182组作为测试样本,并将测试结果与三比值法和朴素贝叶斯分类器进行比较。大量诊断实例表明,所提出的变压器故障诊断模型性能优于另外两种方法。The factor analysis and gene expression programming are combined to go on with the transformer oil chromatography data processing and intelligence training, and the transformer fault diagnosis model classification is establish. 332 groups of oil chromatographic data are collected, and 150 groups are selected as the training samples of GEP classifier, 100 groups as the test samples. Test result is compared with those of three-ratio method and the naive Bayesian classifier. A large number of diagnostic examples show that the proposed transformer fault diagnosis model is superior to the other two methods.

关 键 词：变压器 故障诊断 基因表达式编程 因子分析 DGA 

分 类 号：TM411[电气工程—电器]