基于SVD-WPA的模拟电路故障特征提取方法  

Fault Feature Extraction Method of Analog Circuit Based on SVD-WPA

作　　者：王振 马清峰 WANG Zhen;MA Qingfeng(School of Physics and Electronic Information,Huaibei Normal University,Huaibei Anhui 235000,China)

机构地区：[1]淮北师范大学物理与电子信息学院,安徽淮北235000

出　　处：《佳木斯大学学报（自然科学版）》2023年第3期31-34,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：2020年度安徽高校自然科学研究项目(KJ2020A0030)。

摘　　要：针对模拟电路中电路元件由于容差带来的故障诊断问题,提出了一种基于奇异值分解(SVD)结合小波包分析的模拟电路故障特征提取方法。首先对待测电路的输出信号进行奇异值分解重构,提高故障信号与无故障信号之间的区分度,然后利用小波包分析(WPA)对重构信号进行两层小波分解重构,得到故障特征。最后将故障特征采用网格搜索法改进的支持向量机中进行分类识别。仿真实验结果表明,奇异值分解结合小波包能量谱所提取的故障特征具有更好的区分度,验证了该方法的有效性和可行性。Aiming at the fault diagnosis problem caused by the tolerance of circuit components in analog circuits,a fault feature extraction method for analog circuits based on singular value decomposition(SVD)combined with wavelet packet analysis is proposed.Firstly,the output signal of the circuit under test is reconstructed by singular value decomposition to improve the distinction between the fault signal and the fault-free signal,and then the wavelet packet analysis(WPA)is used to decompose the reconstructed signal at two layers of wavelet decomposition to obtain the fault features.Finally,the fault features are classified and identified in the support vector machine(SVM)improved by the grid search method.Simulation results show that the fault features extracted by singular value decomposition combined with wavelet packet energy spectrum have better differentiation,which verifies the effectiveness and feasibility of the method.

关 键 词：模拟电路 故障诊断 奇异值分解 小波包分析 支持向量机 

分 类 号：TN707[电子电信—电路与系统]