基于改进灰关联分析的变压器故障诊断

Fault Diagnosis of Power Transformer Based on Improved Gray Relation Analysis

为提高油浸式变压器故障诊断精度以保障电力系统安全稳定运行,作者提出了一种基于改进灰关联分析的变压器故障诊断方法。首先,基于油中溶解气特征气体建立特征集合;随后采用Fisher Score法和Relief F法求解特征权重值并排序,保留重要特征并构建低维参考序列;最后,引入特征重要度作为权重计算序列间关联度,并确定测试样本故障类型。基于仿真测试结果及其对比分析表明:本文提出的基于改进灰色关联分析故障诊断方法较其他传统方法具有更高的诊断精度,对实际工程应用具有一定的参考借鉴意义。

In order to improve the accuracy of oil-immersed power transformer fault diagnosis for keeping power grid stable and safe,an intelligent fault diagnosis approach based on feature selection technique and Grey Relation Analysis(GRA)is present in this paper.First,a set of features extracted from crucial gases are applied to establish fault samples.Then,Fisher score method and Relief F method are used to assess importance of each feature,and low dimensional reference consequences are established with significant features.Finally,the importance of each feature is introduced as weight to calculate relation degree for determining the exact fault type of testing samples.and comparison of fault diagnosis results with different approaches suggest that the presented approach combined with feature selection technique and Grey Relation Analysis has better fault diagnosis accuracy than that of conventional methods,which has certain guiding significance for actual engineering application.