灵敏度分析和SVM诊断模拟电路故障的方法

针对低可测性模拟电路中存在的模糊组问题，提出了一种模拟电路故障诊断的新方法。该方法首先计算被测电路网络传递函数零极点的灵敏度，利用零极点灵敏度提供的信息来对被测电路进行模糊组的划分，组成可诊断的元件集，并引入了支持向量机完成对故障的分类识别。零极点的灵敏度分析确定了被测电路可诊断的元件组，支持向量机结构简单、泛化能力强，实验结果证明了基于灵敏度分析和支持向量模拟电路故障诊断方法的有效性，故障诊断率大于99%。

Fault Diagnosis Method for Analog Circuits Using Sensitivity Analysis and SVM

A fault diagnosis method for analog circuits with low testability using sensitivity analysis and support vector machine (SVM) is presented in this paper. The transfer function poles and zeros sensitivity of circuit under test is analyzed. The ambiguity group is determined according to the poles and zeros sensitivity with respect to the components and the testable components set is obtained. The poles and zeros sensitivity analysis theoretically determines all of the diagnosable circuit components and SVM has advantages of simple structure and strong generalization ability. Experimental results prove that the proposed method for diagnosing analog circuit fault using poles and zeros sensitivity analysis and SVM is effective and the fault diagnosis accuracy of the method is more than 99%.