摘要
由于电力变压器故障诊断中数据信息不完备,存在一定的误差,不能完全正确分析、诊断故障。当变压器溶解气体分析复杂时,粗糙集诊断准确度降低,而支持向量机只适合小样本集。为此,提出了粗糙集与支持向量机相结合的变压器诊断方法,首先利用粗糙集对特征变量进行约简,去除冗余变量得到特征信息,应用支持向量机把该特征信息进行正确的分类,从而达到故障诊断的目的。与常规方法比较,该方法简单有效,诊断速度快,诊断正确率高。
Facing a lager number of incomplete fault date, the fault diagnosis methods cannot be effectively analyzed or accurately diagnosed. When the values of dissolved gas analysis（DGA） is complex, the accuracy of rough set （RS） fault diagnosis became more and more lower, and the support vector machine （SVM） is just suitable to the small sample date. The method is presented based on rough set theory integrated with support vector machine for fault diagnosis of transformer. Firstly, RS is used to reduce the attribute character, and then a key decision table is obtained after deducting redundant attributes. Finally, the key decision table is acted as a learning sample to train the constructed SVM, thus fault diagnosis of power transformer is realized. The experimental results indicate that the method is simpler, faster and more accurate compared with the traditional algorithm.
出处
《陕西电力》
2015年第7期78-82,共5页
Shanxi Electric Power
