基于神经网络的模拟电路故障诊断

Analog circuits fault diagnosism based on neural network

探讨了基于故障字典法和基于神经网络的故障诊断方法 ,利用自组织特征映射神经网络进行模拟电路故障诊断 ,根据神经网络的输出结果可以判断发生故障的类型 .自组织特征映射神经网络聚类能力强、速度快 ,因此很适合复杂系统的故障诊断 .

The fault diagnosis method based on the fault dictionary and the neural networks was investigated and discussed,Self-organizing feature map (SOFM) neural network was used to diagnose faults in analog circuit,and the type of the faults can be judged by the means of the outputs of the neural network.SOFM neural network had a strong grouping capability,high speed and was therefore suitable for fault diagnoses of complicated systems.