Nonlinear Analog Circuit Diagnosis Based on Volterra Series and Neural Network

The Volterra kernels are the inherent characteristic of the system. This paper researched how to measure Volterra frequency kernels and used the second Volterra frequency kernels as the fault signatures in diagnosis nonlinear analog circuit. The fault dictionary of nonlinear circuits was constructed based on improved Back-Propagation neural network. Experiment result demonstrates that the method of this paper has high diagnose sensitivity and fast fault identification and deducibility.