摘要
对高压断路器动作过程中状态特征参量的提取与分析是状态识别与故障诊断的关键。高压断路器分合闸过程中的触头行程曲线蕴含着反映其内部机构机械状态的丰富信息。然而,仅依靠常规的时间和速度等参量无法对高压断路器的机械状态进行准确识别。文中提出了一种基于最优特征向量分类的高压断路器机械状态识别方法,通过计算各个特征与状态分类之间的互信息,根据最大相关最小冗余的准则筛选出最优特征子集,然后基于最优特征量构建支持向量机(SVM),利用分类准确度进行评价,确定出最优的特征向量和分类模型。对实验数据的分析结果表明,该方法可以有效提取触头行程曲线中蕴含的特征信息;基于最优特征向量集构建的分类模型的准确度高达97%,可以实现对高压断路器机械状态的识别。
In operation of high voltage circuit breakers,the status feature extraction and analysis are vital for status identification and fault diagnosis. The contact travel curve of a high voltage circuit breaker implies valuable information of operation process, which indicates the mechanical status of the high voltage circuit breaker. However,status of high voltage circuit breakers cannot be identified accurately only by the common parameters,such as time and speed. In this paper,a mechanical status identification method based on optimal feature vector classification is proposed for high voltage circuit breakers. Firstly,the mutual information between each feature and status classifications is calculated and sorted. According to the maximal relevance minimal redundancy,the optimal feature parameters are selected to form the optimal subset of features. Secondly,some support vector machines(SVMs) are built to identify the mechanical status of high voltage circuit breakers based on several features from the subset of optimal features. Then,the optimal feature vector and the optimal SVM classifier are selected according to the evaluation indicator of classification precision. Experimental data show that this method is effective to extract the feature parameters from the contact travel curve,and the accuracy of the SVM classifier with the optimal feature vector is 97%,so it can be used to identify the mechanical status of high voltage circuit breakers.
作者
杨景刚
吴越
赵科
李洪涛
腾云
张国钢
YANG Jinggang;WU Yue;ZHAO Ke;LI Hongtao;TENG Yun;ZHANG Guogang(State Grid Jiangsu Electric Power Company Research Institute,Nanjing 211103,China;State Key Lab of Electrical Insulation and Power Equipment,Xi'an Jiaotong University,Xi'an 710049,China)
出处
《高压电器》
CAS
CSCD
北大核心
2018年第6期60-66,共7页
High Voltage Apparatus
基金
国家电网公司总部科技项目~~
关键词
高压断路器
机械状态识别
最优特征向量
互信息
支持向量机
high voltage circuit breaker
mechanical status identification
optimal feature vector
mutual information
SVM
