Analog Circuit Fault Diagnosis Based On Wavelet Analysis And Neural Network

Recently, Neural networks (NNs), which is a typical representation of computational intelligences, provides a powerful way to diagnose faults of analog circuits. Now it is interested to researchers. In this paper, fault features are abstracted by wavelet analysis, then neural network and artificial immune algorithm are used to local fault competences. The main contributions in the paper are as follows:A systematic method for fault diagnosis of analogue circuits based on testability analysis is presented. The proposed method calculates the testability of the circuit, then uses wavelet analysis as a tool to extract the best optimal feature information, the features are applied to the proposed neural network and the fault pattern are classified. The simulation result shows that the proposed method can diagnosis the faults well.Inspirited from the vertebrate immune system, a new method for fault diagnosis of analog circuit based on artificial immune algorithm is presented in this paper. The proposed method uses wavelet analysis as a tool to extract the fault examples, then the fault examples are learnt to get the clustering centers of each fault class, finally the fault components are localized by computing the distance between the fault example and the clustering centers. The simulation result shows that the proposed method has the capability to diagnose faults in tolerance circuits and gains satisfactory accuracy. In order to solve the problems of insufficient test data and low diagnosis accuracy, a fault location approach for analog circuits is developed based on intelligent information fusion technology including data fusion and neural network. Input node current and output node voltage are abstracted by wavelet analysis, then feature fusion are used to local fault competences. The proposed approach makes full use of the fault information and has the capability to diagnose multiple and soft faults of analog circuits with satisfactory accuracy.