轴承故障的自适应小波神经网络分类  

Fault Classification of Bearing Using Adaptive Wavelet-based Neural Network

作　　者：陈文戈[1] 赵学智[2] 

机构地区：[1]广东工业大学机电工程学院,广州510090 [2]华南理工大学机械工程学院,广州510640

出　　处：《轴承》2009年第3期37-40,共4页Bearing

基　　金：国家自然科学基金资助项目(50305005)

摘　　要：提出了一种用于故障分类的自适应小波神经网络,网络第一部分利用小波伸缩平移系把信号分解到不同频道上进行特征提取,第二部分对提取的特征信息进行学习或判断。推导了该网络的学习算法,并应用其对轴承进行了故障分类,结果表明该网络分类准确,可靠性高。One kind of adaptive wavelet neural network for fault classification is put forward. In the first component of network, a family of wavelet is used to decompose signal into different channels so that feature of signal can be extracted, while the second component of network is used to learn these information or judge fault type according to these feature. The learning algorithm is presented in detail and this network is applied to fault classification of beatings. The result demonstrates that this neural network can classify fault accurately and reliably.

关 键 词：滚动轴承 故障诊断 故障分类 自适应小波 神经网络 

分 类 号：TH133.33[机械工程—机械制造及自动化] TP206.3[自动化与计算机技术—检测技术与自动化装置]