灰关联度在变压器故障诊断中的应用

Application of Grey Relative Degree to Fault Diagnosis of Transformer

为了提高传统油中溶解气体分析(DGA)的诊断能力,克服相对关联度诊断变压器故障的不足,探讨了基于关联分析的故障诊断理论,提出了三级关联算法,建立了一种定性与定量相结合的电力变压器故障诊断的三级关联度方法。以5种油中溶解气体:氢气、甲烷、乙烷、乙烯和乙炔的体积分数参数作为特征量,验证了新方法对变压器故障的实际诊断效果。结果表明,该方法不仅能找出故障类型而且能确定故障部位,其有效性超过了传统的相对关联度方法。

In order to enhance the ability of conventional dissolved gas analysis (DGA) and to overcome the disadvantages of relative dependent grade, a discussion about the principle of the faults diagnosis based on relative analysis theory is provided. And an algorithm is proposed for solving such problems. A novel method about three level degree of association is established for qualitatively and quantitatively describing faults diagnostic problem of power transformer. Experimental results show the effectiveness and the superiority of the proposed method using five kinds of dissolved gases in the oil, acetylene, hydrogen, ethane, methane, and ethylene.