基于支持向量机的电晕放电模式识别方法  

Corona discharge pattern recognition based on support vector machine

作　　者：汪德才[1] 郑殿春[1] 张连星[1] 赵文君[1] 王新月[1] 赵大伟[1] 

机构地区：[1]哈尔滨理工大学电气与电子工程学院,黑龙江哈尔滨150080

出　　处：《黑龙江电力》2010年第6期420-425,共6页Heilongjiang Electric Power

摘　　要：应用自回归(Autoregression,简称AR)模型提取电晕放电信号的AR模型参数作为信号特征量,利用支持向量机(Support Vector Machine,简称SVM)作为分类器对放电模型进行识别,重点分析了核函数和支持向量机参数对分类器的影响。结果表明,AR模型参数作为信号特征量,并与支持向量机结合对电晕放电信号进行识别是有效的。AR model parameters of corona discharge signal extracted by AR model as signal characteristic quantity and support vector machine as classifier,discharge model recognition is conducted and influences of kernel function and support vector machine on classifier are analyzed.Result indicates that AR model parameters,as signal characteristic quantity,together with support vector machine is effective to corona discharge recognition.

关 键 词：电晕放电 AR模型参数 支持向量机 模式识别 

分 类 号：TM83[电气工程—高电压与绝缘技术]