粗糙集理论在变压器故障诊断中的应用

为了对变压器故障诊断过程中大量的冗余特征进行压缩或约简,提高诊断的效率,将粗糙集理论引入到变压器故障诊断中,提出了基于粗糙集理论的故障特征约简算法:即由故障样本构成信息表,组合表中不同的属性集,求取与全体属性集具有相同分类质量的最小属性集。对具体典型诊断实例进行了分析,结果表明:在保证故障分类结果不变的情况下,该算法能够剔除具有冗余信息的特征,找出对故障分类起主要作用的特征,从而达到了特征约简的目的,不仅大大减少了诊断信息提取的工作量,也为后续的智能诊断提供很大的便利。

Application of rough set theory in transformer fault diagnosis

In order to improve diagnosis efficiency and compress the redundant features in the transformer fault diagnosis, the rough set theory is introduced and the feature reduction algorithm based on the rough set theory is presented. Based on the samples, the information table is constructed. By arranging different attribute set and calculating its classification quality, the minimal attribute set having the same classification quality with the whole attribute set can be obtained. The representative samples in transformer fault diagnosis are analyzed. The results show that the redundant features can be reduced and the main features which are more important to fault classification can be gotten.