Tolerance Circuit Fault Diagnosis Of Bp And Sofm Neural Network Method

The analog tbuit diagnosis research is one of the forefront of the testing arena, one of the reason is that the reliability of e1ectrical syàtems.This paperependu on tbt of pnalog circuits and systems. This paper focuses Qfl the fault diagnosis of analog circuits with tolerances using the artificial neural networks method. The primary aim of tlie paper is to provide a mechanism to deal with the problem of tlement toleraneQs and reduce the testing time. Applying the classical pattern recogtiition theory anci ftiflcial neural networks method, this paper proposes the analog fault diahoes priricip1s with backward-propagation neural network (I3PNN) arid self-or~aniZing feature map (SOFM) neural network algorithm implementation. Tile robustne5s arfti associated memory of ANN make the method more advantageous over traditional mefliods. The supervised and unsupervised learning diagnosis methods are discussed and several improvements have been presented in the learning algorithms. The simulation results show that the proposed method can perforfti correct diagtioals iii the linear analog circuits with tolerances.