基于多分类相关向量机的变压器故障诊断新方法  

New transformer fault diagnosis method based on multi-class relevance vector machine

作　　者：尹金良[1] 朱永利[1] 俞国勤[2] 

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003 [2]上海电力公司,上海200025

出　　处：《电力系统保护与控制》2013年第5期77-82,共6页Power System Protection and Control

基　　金：河北省自然科学基金资助项目(E2009001392)

摘　　要：变压器故障诊断本质为多分类问题,具有故障样本数据少,故障不确定因素多的特点。现有变压器故障诊断方法中,贝叶斯网络(BN)需要大量样本数据且计算量大,支持向量机(SVM)存在规则化系数确定困难的局限。针对此现状,提出基于多分类相关向量机(M-RVM)的变压器故障诊断新方法。该方法以变压器溶解气体含量比值作为M-RVM模型的输入,采用快速type-Ⅱ最大似然(Fast Type-ⅡML)和最大期望估计(EM)的方法进行模型推断,诊断输出为各故障类别的概率,以概率最大的故障类别作为诊断结果。实例分析表明该方法诊断速度较快,能满足工程需要,同基于BN和SVM的变压器故障诊断方法相比,具有较高的诊断正确率。The transformer fault diagnosis is naturally a multi-classification problem with few sample data and a lot of uncertainties. Among the existing transformer fault diagnosis methods, a large number of sample data and amount of computation are needed for Bayesian Network （BN）, and the adjustment of the coefficient is difficult for support vector machine （SVM）. So a new method of transformer fault diagnosis based on multi-class relevance vector machine （M-RVM） is proposed. The method takes ratios of feature gases as inputs and Fast Type-II ML and expectation maximization （EM） are adopted. Diagnostic outputs are probability for each fault category and fault type with the highest probability is taken as diagnosis result. Experimental results show that the diagnosis speed is sufficient for project needs and M-RVM shows higher diagnosis accuracy compared with BN and SVM.

关 键 词：多分类 相关向量机 贝叶斯网络 支持向量机 变压器故障诊断 

分 类 号：TM407[电气工程—电器]