基于敏感度因子与故障隔离度的模拟电路测点选择

Analog circuit test points selection based on inner-and inter-class sensitivity factor and fault isolation degree

提出了一种基于类内类间敏感度因子与故障隔离度的新方法对模拟电路测试节点进行优选。计算电路各测点采样数据的类内与类间距离离散度来定义测点敏感度因子,根据敏感度因子大小对待优选测点进行重新排序,利用KNN网络计算重排序测试节点的故障隔离组(度),最后优选出能辨识全部预设故障的最优测试节点集。实验结果证明:新方法得出的最优测点集包含的测点数量更少,该方法可以优选出相比其他文献方法故障诊断效率更高的同等规模的测试节点集合。

A new method for optimal analog test points selection based on inner-and inter-class sensitivity factor and fault isolation degree is proposed. The inner-and inter-class dispersion of sampling data is calculated to define the test points＇ sensitivity factor,The test points is reordered according to the value of the sensitivity factor, and fault isolation group and degree of the reordered test points is computed by KNN network. The optimal test points set can be selected by the guidance of the proposed algorithm. Experimental analysis proved that the optimal test points set derived by the method has less test points, compared with other reference, the proposed method can select test poins set with higher fault diagnostic rate.