摘要
针对复杂模拟电路故障诊断难的问题,提出了一种基于交流置换激励、同时结合灵敏度分析与支持向量机(Support Vector Machine,SVM)的模拟电路子网络级故障诊断方法。通过交、直流激励实现了目标子网络的隔离撕裂;根据子网络灵敏度优选测试节点,提取测试点信息,经蒙特卡洛分析与归一化处理后作为故障特征;选用SVM作为故障诊断分类器,实现了故障子网络的识别。最后,通过电路实例对该方法进行了验证,结果表明:与已有方法相比,该方法计算量小、精确度高,适用于容差和非线性电路。
In order to solve the difficulty of fault diagnosis in complex analog circuits,the paper presents a network-level fault diagnosis method combining with sensitivity analysis and Support Vector Machine( SVM) based on alternating replacement stimulus. The target sub-network is torn and isolated through the use of AC and DC stimulus; the test points are preferred according to the sub-network sensitivity and their information is extracted as the fault feature after Monte Carlo analysis and normalized processing;SVM is chosen as fault diagnosis classifier to realize the identification of the fault sub-network. Finally,the method is verified by the special circuit examples,the results show that: compared with the existing methods,this method has the advantages of small computation and high precision,and it is suitable for tolerance and nonlinear circuit.
出处
《装甲兵工程学院学报》
2017年第3期69-74,共6页
Journal of Academy of Armored Force Engineering
关键词
置换定理
交流激励
子网络灵敏度
支持向量机
replacement theorem
AC stimulus
sub-network sensitivity
Support Vector Machine(SVM)