以模糊聚类标准谱与灰色关联序诊断变压器内部故障的方法研究

Study on Analysis Method About Fault Diagnosis of Transformer and Degree of Grey Inidence Based on Fuzzy Clustering

通过对变压器故障和以 DGA为特征量的症状之间难以量化的模糊和灰色特征的分析 ,采用模糊聚类分析方法获取 c个聚类中心 ,根据聚类中心的物理和数学意义 ,这 c个聚类中心组成变压器故障诊断的标准谱 ,利用灰色关联分析原理 ,求出待诊模式与各标准模式的关联序 ,实现故障诊断 ,进而提出了灰色关联分析和模糊聚类相结合的变压器故障诊断新方法。并经实例证明 。

Fuzzy and gray characteristic between transformer fault and symptom DGA characteristic are analyzed, which are difficult to determine. Then c kinds of cluster centers are computed by using fuzzy cluster analysis method. According to physical and mathematical significance of cluster center, these c kinds of cluster centers make up standard chart of transformer fault diagnosis. According to gray incidence analysis theory, the incidence order of diagnosing pattern with the standard pattern is computed. So the new fault diagnosis method to transformer is proposed, which is based on the combination of gray incidence analysis and fuzzy cluster. The tests show that diagnosis accuracy of the proposed method is higher than other traditional methods.