Fault Diagnosis of Nonlinear Analog Circuits Using Neural Networks and Multi-Space Transformations

A systematic approach. based on piece-wise linear (PWL) models, a bilinear transformation ill multidimensional spaces and back-propagation neural networks (BPNN). for nonlinear analog fault diagnosis is proposed in this paper. The functions of input-output are applied for fault diagnosis to deal with the circuits Without Sufficient accessible nodes. Besides, We used the functions transformation ill multi-space to select fault features. which can decrease the ambiguity groups and improve the performance of fault diagnosis. Through preprocessing of the signals of test nodes from the analog circuits, the optimal features are selected. These features are then fed into the BPNNs for fault location. The single fault diagnosis is mainly discussed in this paper. Finally. all illustration to demonstrate the strength of our proposed method is given.