摘要
介绍小波神经网络的发展和基本原理。在此基础上,详细分析和推导了两种小波神经网络的结构和算法,使其把小波变换和神经网络有机地结合起来,充分继承两者的优点。重点阐述对小波神经网络进行的改进,即动量项的引入,证明了动量项的引入在一定程度上解决了局部极小的问题,并对小波神经网络在故障诊断中的应用研究现存的几个问题和未来的发展进行了深入的探讨。
This paper introduces the development and basic principle of wavelet neural network, on this basis, structures and algorithms of two different wavelet neural networks are analysed, which combine wavelet transform and nearal network emphatically set forth the improvement to wavelet neural network, the adoption of momentum item, which proves that the adoption solved the problem of part tiny on the fixed extend, several problems of wavelet neural network in fault diagnosis study are analysed,and settlement of these problem and development of future are discussed.
出处
《现代电子技术》
2009年第11期129-131,共3页
Modern Electronics Technique
关键词
故障诊断
小波分析
神经网络
小波神经网络
fault diagnosis
wavelet analysis
neural network
wavelet neural network
作者简介
冯楠女，硕士研究生。主要研究方向为电子设备故障诊断。
李文海男，硕士生导师。主要研究方向为电子设备故障诊断。
