基于方差和峭度的模拟电路故障诊断

The Analog Circuit Fault Diagnosis Based on Variance and Kurtosis

针对非线性容差电路故障诊断过程中存在的故障特征提取难的问题,结合小波包分析理论,提出了利用二阶统计量和高阶统计量来描述故障信息的故障特征提取方法,即方差和峭度法,并运用支持向量机作为分类器,形成了一种模拟电路故障诊断的新方法.仿真结果表明,该法能有效的提取故障特征,故障诊断率较高.

For the problem that it is difficult to extract features when diagnosis the nonlinear and tolerance fault circuits,combined with wavelet packet analysis theory,a fault feature extraction method which used the second-order statistics and the higher-order statistics to describe the fault information was presented,the variance and kurtosis method.By using support vector machine as classifier,it finally formed a new method for the analog circuit fault diagnosis.Simulation results showed that the method can be effective to extract the fault features and has a high accuracy in fault diagnosis.