A Fuzzy Classification Approach for Analog Fault Diagnosis Applying Immune Algorithm

A novel fault diagnosis approach is presented for analog circuits where the accessible test points are limited and insensitive enough to some fault components. The wavelet analysis as a tool to extract the fault samples makes good use of the fault information from the test points. Applying immune algorithm to obtain the clustering center of corresponding fault mode is satisfactory on convergence and accuracy. A fuzzy approach localizes the fault component according to the membership degrees of the test sample to the clustering centers of the standard fault modes. The results of the simulation experiment show that the proposed approach is effective and practicable.