基于斜率特征的模拟电路故障诊断新方法

A New Method for Analog Circuit Fault Diagnosis Based on Slope Characteristics

针对低可测性模拟电路存在的模糊组问题,提出一种模拟电路故障诊断新方法。利用不同元件在电路可及节点间的斜率特征关系建立斜率测试矩阵,依据斜率测试矩阵准确找到电路中存在的模糊组和独立可测元件,从而确定可诊断集,然后利用支持向量机对故障进行诊断。斜率法优点在于不需要对电路拓扑结构进行分析,计算简便,支持向量机结构简单,泛化能力强。实验表明该方法具有良好的模拟电路故障诊断效果。

There is fuzzy set in analog circuit with low testability, a new method for analog circuit fault diagno- sis is proposed. The slope test matrix is determined by the different components＇ slope characteristics of the test points. According to the slope test matrix, the fuzzy set and testable elements are found accurately, and the diagnosable circuit components is also determined. After that the support vector machine （SVM） is used for fault diagnosis. This method does not need to analyze the circuit topology, and has simple calculation. The SVM has advantages of simple structure and strong generalization ability. Results show that the new method is effec- tive for analog circuit fault diagnosis.