小波神经网络在模拟电路故障诊断中的应用研究

对模拟电路提出了一种基于小波与神经网络辅助式结合的故障诊断方法．该方法用小波变换作为模拟电路故障信号的预处理器，大大减少神经网络的输入数目，简化神经网络结构和减少它的训练时间，提高辨识故障能力．在介绍该故障诊断方法的基本原理后。给出了小波函数及故障特征选择的方法．

Application of Wavelet and Neural Network in Analog Circuit Fault Location

A fault location method based on wavelet and neural network for analog circuit is presented. The proposed method uses wavelet decomposition as the preprocessor of analog circuit fault signals. This can dramatically reduce the number of inputs of the neural network, simplify its structure, minimize its training and processing time ,and promote the capability to detect and identify faults. After briefly introducing the basic principle based on wavelet and neural network for fault location, the method of selecting the wavelet and fault feature was given.