基于粒子群算法优化支持向量机的模拟电路诊断

摘要

In order to improve the ccuracy of analog circuit fault diagnosis using support vectour machine ( SVM) network, this paper proposed the method based on particle swarm optimization( PSO) and SVM. It preproeessed the response signals of the analog circuit using multiwavelet transform and obtained the optimal fault feature with better classification capacity using energy normalization. Then, after training the SVM by PSO, inputted the features into the ensemble SVM to identify different fault cases. Simulation results indicate that this method can effectively enhance the analog fault diagnostis accuracy.%为了提高支持向量机网络(SVM)进行模拟电路诊断的准确率,提出了一种基于粒子群(PSO)算法和支持向量机的诊断方法.该方法首先对被测电路的响应信号进行多小波变换,通过归一化处理得到分类能力强的最优故障特征；然后用粒子群算法优化支持向量机的结构参数,实现对不同故障模式分类识别.仿真结果表明,此方法能有效提高模拟电路故障诊断准确率.