基于波形相似度的容差模拟电路软故障诊断

摘要

针对容差模拟电路输出波形参数的随机性,构造了描述特定电路状态的特征区间向量,据此提出了一种基于多频响应波形相似度的软故障诊断方法。对每一种软故障,分别施加K种不同频率的激励信号,获取K个由输出响应构成的标准特征区间向量作为该故障状态的信号特征。对待测电路施加相同的K种频率的激励信号,获取K个待测样本,分别与每种故障模式的K个相应的标准特征区间向量进行相似度计算,获取K个波形相似度,在波形相似度融合中考虑了均方差权重系数和可信度权重系数,得到待测样本与该故障模式的总体相似度。最后依据一定的模糊规则进行故障定位。仿真实验证明,本方法在保证较高诊断准确率的情况下,对测点要求低,易于在自动测试系统中实现诊断的自动化。%Regarding to the random attribute of the parameters of tolerance analog circuit output waveform,the feature interval vector used to describe the tolerance analog circuit in a certain mode is constructed.Accordingly,a novel soft fault diagnosis method based on waveform similarity acquired by multi-frequency tests is proposed.For each fault mode of the circuit under test,K standard feature interval vectors are acquired by K tests with alternating current signal inputs with different frequencies respectively.In order to diagnose the fault mode of the circuit,it is input with the same signals with K different frequencies to form K output wave parameter samples.K wave similarities are calculated by the K output wave parameter samples with the K standard feature interval vectors of each fault mode.The integrated similarity is the fusion of the K wave similarities considering mean square deviation weight and credibility weight.Finally,fault location is realized by a set of fuzzy rules.The simulation results show that the proposed method gains satisfactory accuracy while less limited by accessible test points.Moreover,it is practical and easy to be realized in automatic test systems.