基于LMD和SVM算法的模拟电路故障诊断

Analog Circuits Fault Diagnosis Based on LMD and SVM Algorithms

对模拟故障电路进行特征提取与分类是模拟电路诊断的两个重要环节。现有方法多对时域响应信号进行小波变换以提取故障特征,并用神经网络或支持向量机方法实现对故障进行分类。为提高模拟电路故障诊断率,提出一种局域均值分解(LMD)与SVM相结合的新算法。该算法运用局域均值算法(LMD),将其自适应地分解为一系列单分量调幅-调频信号(PF),通过提取电路正常和故障状态的特征,运用SVM对其分类,获得诊断效率。仿真实验结果表明,该方法对模拟电路的故障诊断精度达到98%以上,适用于模拟电路的故障诊断。

In the process of the fauh diagnosis of analog circuits, feature extraction and classifier design are two critical aspects. Most methods classified fault circuit via support vector machine（SVM） or neural network using ex- tracted time signals and wavelet transforms. A new algorithm based on LMD and SVM is proposed to improve the di- agnostic accuracy. The signal can be adaptively decomposed into a series of one-component AM-FM signal （PF） through using the LMD algorithm. The features of the normal or fault status of the circuit can be extracted. The fea- tures are classified using SVM to achieve the diagnostic accuracy. The result of simulation shows that the method is effective in the circuits fault diagnosis with an accuracy 〉98%.