支持向量机在模拟电路故障诊断中应用  

Support Vector Machine for Analog Circuit Fault Diagnosis

作　　者：孙永奎[1] 陈光[1] 李辉[2] 

机构地区：[1]电子科技大学自动化工程学院,成都610054 [2]电子科技大学空天科学技术研究院,成都610054

出　　处：《电子测量与仪器学报》2008年第2期72-75,共4页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金资助项目(编号:60372001,90407007);教育部博士点基金资助项目(编号:20030614006)

摘　　要：支持向量机是建立在统计学习理论基础上的机器学习方法,结构简单,泛化能力强,对小样本分类具有良好的识别效果。本文提出了基于支持向量机的模拟电路故障诊断新方法,描述了电路故障特征的选取过程,建立了以支持向量机为基础的模拟电路故障诊断模型。并以双二次滤波电路为诊断实例,实验结果表明,该方法故障诊断准确率大于96.5%,优于传统方法。Support vector machine （SVM） is a machine-learning algorithm based on statistical theory, which has the advantages of simple structure and strong generalization ability as well as classification ability for small sample. A new method for analog circuit fault diagnosis based on SVM is presented in this paper. The method for extracting the fault signatures of the circuit under test is pro- posed and the analog circuit fault diagnosis model based on SVM is established. The simulation results of a biquadratic filter testify that the proposed approach for analog circuit fault diagnosis is superior to conventional ones, the fault diagnosis accurate rate of the method is greater than 96. 5% and the fault diagnosis range extends from single-point failures to interval failures.

关 键 词：模拟电路 支持向量机 故障诊断 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TM930[自动化与计算机技术—控制科学与工程]