基于SOFM网络的模拟电路故障诊断方法研究

Simulation Studies of Analog Circuits Fault Diagnosis Based on SOFM Neural Networks

讨论了基于自组织特征映射神经网络SOFM实现交流字典的故障定位方法 ,该方法对单、多软硬故障的识别快速且有效。同时 ,利用神经网络的泛化性 ,有效地克服了容差因素对故障定位的影响。

In this paper,the construction of AC fault dictionary and its fast looking for based on SOFM are discussed. The method is very efficient in identification and location of single/many soft/hard fault. The effects correcting fault location due to components tolerance are reduced because of extensiveness of neural networks.