Analog Circuit Soft Fault Diagnosis Based on Chaotic Neural Network
Meirong Liu, Li Zeng, Liwei Zhang, Yigang He · 2016
In order to solve the fault feature redundancy problem in analog circuit fault diagnosis, a method of fault diagnosis is presented in this paper.This approach using the method of wavelet decomposition and fuzzy clustering on the fault signals to obtain test matrix; then inputted the test matrix into the neural network for fault diagnosis.The approach are combined the chaotic motion's ergodic, randomness and sensitive of the initial value to optimize the neural network for making the network have a better learning ability and have a more faster convergence speed to improve the efficiency of fault diagnosis.The simulation results verify the effectiveness of this approach.