Fault Diagnosis of Analog Circuit Based on Wavelet Neural Networks and Chaos Differential Evolution Algorithm

A novel method for fault diagnosis of analog circuit based on chaos differential evolution wavelet neural networks (CDE-WNN) is proposed in this paper. In order to simplify network architectures and improve its learning accuracy and convergence rate, the architectures and parameters of wavelet neural networks are optimized by chaos differential evolution algorithm in the method. The fault dictionary is constructed in the weights of neural networks. The optimized WNN has the capability to detect and identify fault components in an analog electronic circuit. The simulation results show that the proposed method has not only the capability to reduce the effect on correct fault diagnosis due to components tolerance but also a small quantity of examples before test, fast diagnosis rate, and satisfactory accuracy of the diagnosis detection and location. A comparison of our work with WNN and BP algorithms, which reveals that our system requires a much smaller network and performs significantly better in fault diagnosis of analog circuits.