基于人工鱼群优化支持向量机的模拟电路故障诊断  

Analog Circuit Fault Diagnosis Based on Artificial Fish Swarm Optimization Support Vector Machine

作　　者：孙健[1,2] 王成华[2] 洪峰[2] 王蕾[1] 

机构地区：[1]南京信息职业技术学院,南京210046 [2]南京航空航天大学电子信息工程学院,南京210016

出　　处：《系统仿真学报》2014年第4期843-847,共5页Journal of System Simulation

基　　金：国家自然科学基金(50907033);南京信息职业技术学院科研基金(YKJ12-022)

摘　　要：为了提高模拟电路故障诊断的识别率,提出一种基于人工鱼群优化支持向量机的模拟电路故障诊断方法。该方法用主成分分析对采集到的模拟电路故障信息进行降维,将提取后的故障特征输入支持向量机进行故障诊断,同时采用人工鱼群算法对支持向量机的惩罚参数和核函数参数进行优化,避免参数选择的盲目性,提高模型的诊断精度。通过对Sallen-Key带通滤波器电路的故障诊断仿真结果表明,该方法是有效的,与BP神经网络和传统SVM等方法的诊断结果相比较能够诊断更多的故障类别,并具有更高的故障诊断率。In order to improve the accuracy of fault diagnosis in analog circuit fault diagnosis, an analog circuit fault diagnosis method based on artificial fish swarm optimization support vector machine was proposed. Principal component analysis was applied to reduce the dimension of the fault information to analog circuit. Then the extracted fault feature was inputted into a SVM network to identify different fault cases, and artificial fish swarm algorithm was used to optimize the penalty parameters and the kernel parameters of SVM, that avoided blindness of parameters selection and improved the diagnostic accuracy of the model. The simulation results of Sallen-Key bandpass filter circuit fault diagnosis show that this method is effective. It is able to diagnose more fault category and has higher fault diagnosis rate compared with the method of BP neural networks and SVM.

关 键 词：模拟电路 故障诊断 支持向量机 人工鱼群算法 参数优化 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]