改进的M-ary支持向量机模型及其在变压器故障诊断中的应用  

The Improved M-ary Support Vector Machine and Its Application in Power Transformer Fault Diagnosis

作　　者：肖燕彩[1] 陈秀海[2] 

机构地区：[1]北京交通大学机械与电子控制工程学院,北京100044 [2]北京电力公司,北京100031

出　　处：《上海交通大学学报》2008年第12期2033-2036,共4页Journal of Shanghai Jiaotong University

摘　　要：使用M-ary支持向量机进行基于油中溶解气体分析的变压器故障诊断研究.分析结果表明,M-ary支持向量机算法简单,与一对一支持向量机的诊断精度可以比拟.在此基础上,还对常用的M-ary支持向量机模型进行了改进,将各二类分类器的输出计算值直接使用另一支持向量机进行组合,可以更好地反映各分类器之间的非线性关系,从而使新模型具有更高的分类精度.变压器的应用实例证明了改进方法的有效性和优越性.The M-ary support vector machine was applied to the insulation fault diagnosis of power transformer based on dissolved gas analysis. The results show that the M-ary algorithm has the similar precision with the one against one method. To enhance the accuracy, an improved M-ary support vector machine model was proposed. In the built model, another support vector machine was used to combine the output vectors of all the two-kinds classifications. The given model can reflect the nonlinearity relations better. The effectiveness and superiority of the proposed model were verified with the results of the actual cases.

关 键 词：变压器 M—ary支持向量机 故障诊断 

分 类 号：TM855[电气工程—高电压与绝缘技术]