模糊数学结合BP网络的变压器故障诊断研究  

Researeh of transformer fault diagnosis based on fuzzy and BP network

作　　者：兰华[1] 郑广瑞[1] 高奥[2] 侯彩霞[3] 

机构地区：[1]东北电力大学,吉林吉林132012 [2]湖州电力局,浙江湖州313000 [3]内蒙古民族大学,内蒙古通辽028000

出　　处：《黑龙江电力》2011年第1期1-4,共4页Heilongjiang Electric Power

摘　　要：针对油中溶解气体分析(DGA)方法在诊断电力变压器潜伏性故障的研究中存在的不足,采用TOPSIS方法对原始样本集进行优劣排序和聚类,实现了对样本数据的优化,同时极大地缩减了样本集的容量。在此基础上,构建BP网络故障诊断模型。分析证明,此方法与传统的变压器故障诊断的方法相比具有明显的优越性。In view of the oil in the dissolved gas analysis（DGA） the method the insufficiency which exists in the diagnosis power transformer latent breakdown＇s research,this paper extends TOPSIS method under vague.Ordination and cluster the original sample not only optimize the sample data but also substantially reduce the capacity of sample set,to some extent making fault signature information and the problem of mapping space complexity balanced.By constructing diagnosis model of BP network fault,analyzes and shows that compared with traditional method of transformer fault diagnosis this method has obvious superiority.

关 键 词：DGA 模糊理论 TOPSIS 变压器 故障诊断 

分 类 号：TM401.1[电气工程—电器]