基于差分进化杂草算法的容差模拟电路故障诊断

Fault diagnosis of tolerance analog circuits based on differential evolution invasive weed algorithm

为提高容差模拟电路参变故障的诊断率,提出了一种新颖的差分进化入侵杂草算法优化多核支持向量机参数的故障诊断方法。通过小波包变换提取被测电路时域响应信号的特征参量,并生成样本数据,经差分进化入侵杂草算法优化多核支持向量机参数,建立故障诊断模型。故障诊断结果表明,所提出的方法能较好地实现模拟电路故障诊断,与现有方法相比,此方法所建立的SVM模型表现出了更好的性能,获得了更高的故障诊断正确率。

In order to improve the diagnosis rate of analog circuits with tolerance,this paper proposed a novel differential evolution invasive weed algorithm to optimize the parameter of multi-core support vector machines.It used wavelet packet transform to extract the characteristic parameters of the time domain response signal of the tested circuit,and generated sample data.Through differential evolution optimized the invasive weed algorithm,it optimized the SVM model parameters,and established the fault diagnosis model.The fault diagnosis results show that the proposed method can better realize analog circuit fault diagnosis,compared with the existing methods,the proposed SVM model shows better performance,and gets higher accuracy of fault diagnosis.