基于距离分析的邓式灰关联在故障诊断中的应用

Application of Den's Gray Relation Based on Distance Analysis in Fault Diagnosis

传统邓氏灰关联方法是采用等权重的形式求取关联度,其关联度带有较大的主观因素,易造成信息损失,造成错误的故障诊断。针对存在的问题,本文引入了距离分析法,利用非匀权重的形式求取关联度,体现五种油中溶解气体对判断变压器故障类型的不同程度影响。它的特点是权重的分配不受主观因素制约,完全由样本数据信息决定。将该方法用于变压器故障诊断,结果表明,该方法可行有效,有实用意义。

Traditional Den＇s gray relational method adopts the forms of uniform weight to calculate relational degree which has much more subjective factors and easily results in information losses as well as gives wrong fault diagnosis.Aimed at the existing problems,distance analysis method was introduced and non-uniform weight was used to obtain the relational degree which reflecting the different effects on judging default type of transformer caused by dissolved gas in five oil.Its characteristic was that weight distribution was not restricted by subjective factors,and it was decided by samples information.The method was applied to fault diagnosis for transformers.Results show that the method is workable and effective,and has practical significance.