【摘要】 针对钢铁企业传统高压线路行波检测器易出现误报警的问题,提出了采用支持向量机(SVM)的分类辨识方法,用于故障信号和扰动信号的辨识。首先,从行波法定位装置数据库中提取报警时的信号特征,构造支持向量的输入向量,建立基于数据驱动的支持向量机模型;然后采用仿真策略确定支持向量机径向基函数中的参数σ和惩罚系C的值,并分析了参数σ及惩罚系数C的值对故障和扰动分类准确率的影响。将所提出的方法应用到莱钢高压线路行波检测器中,结果表明:采用支持向量机的分类辨识方法,可以使行波检测器检测的准确率接近90%,大大提高了莱钢高压线路行波检测器故障检测的可靠性。更多还原

【Abstract】 A classification method based on support vector machine(SVM) is proposed to distinguish disturbance from failure on traveling wave failure detection device for traditional high-voltage circuit in metal industry.Initially recorded data for fault and disturbance are selected from database,and then input features and output feature are extracted,which are used for data driven classification model.After that,the model is established,and its parameters,i.e.,parameters of the radial basis function and the penalty factor C,are optimized by MATLAB simulation method.In addition,an analysis on influence of selected model parameters to model classification accuracy is also carried out.The proposed method is applied to traveling wave detection device in Laiwu Iron &Steel Company,and the classification accuracy reaches 90%,which enhances the reliability of failure detection compared to the traditional method and achieves the expected results.更多还原