Abstract: A systematic approach combining fuzzy neural network, wavelet analysis and genetic algorithm was proposed for fault diagnosis of analogue circuits.The presented fuzzy neural network was developed with the improved fuzzy weighted reasoning method. The optimal feature sets was extracted to train the network by using wavclet analysis as a preprocessor.This can ensure a simple architecture for the neural network and minimize the size of the training set required for its proper training. And the adjusting of connection weights and optimization of membership functions were performed with genetic algorithms. The reliability of this method was experimented with active filter examples.The resups of experimental tests show that this method can satisfactorily detect and identify the faults. It not only distinguishes the ambiguity sets or some misclassificd faults that some other methods cannot identify, but also has faster speed in the training of network.

Key words: Fault diagnosis, Fuzzy neural network, Genetic algorithm, Wavelet analysis