基于支持向量机的模拟电路故障诊断  

Fault diagnosis based on SVM in analog circuit

作　　者：王旭婧[1] 陈长兴[1] 王明芳[1] 任晓岳[1] 屈坤[1] 

机构地区：[1]空军工程大学理学院

出　　处：《计算机工程与设计》2014年第10期3584-3588,共5页Computer Engineering and Design

基　　金：陕西省电子信息系统综合集成重点实验室基金项目(201107Y16)

摘　　要：针对模拟电路的固有复杂性及其传统故障检测方法存在延时大和正确识别率低的问题,提出基于最小二乘支持向量机和Volterra级数的故障诊断方法。采用Volterra级数频域核对电路故障特征进行提取,利用最小二乘支持向量机进行模态分类,最终完成故障诊断。仿真结果表明,该方法与BP神经网络相比提高了系统故障辨识能力与系统故障诊断速度。To solve the problem of analog circuit complexity ,long time diagnosis and low correct recognition rate in traditional fault diagnosis ,a new fault diagnosis method was proposed which combined least squares support vector machine （LS-SVM） and Volterra series .The circuit feature extraction was completed by using Volterra frequency-domain core ,LS-SVM was used to do modal classification and finally the fault diagnosis was accomplished .Simulation results show that this method ,compared with traditional BP neural network ,is more efficient and accurate in system fault recognition and diagnosis .

关 键 词：故障诊断 最小二乘支持向量机 沃尔泰拉级数 模拟电路 特征提取 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]