摘要
提出一种基于小波包和BRF神经网络的智能故障诊断方法。对滚动轴承故障信号进行小波包分解,选择合适的小波基函数和尺度,将故障信号分解到八个不同的频段上,提取这八个频段上的能量信息,组成特征问量,作为RBF神经网络的输入;建立RBF神经网络模型并进行训练,对三种滚动轴承故障信号进行智能分类与识别。实验结果表明这种智能诊断方法有效可行。
A method of fault intelligent diagnosis based on the wavelet packet and the RBF neural network was proposed in the paper. First, the signal of the roll bearing was decompesed on the eight different frequency bands by the wavelet packet adopting a selected wavelet basis fanction and covel. The energy information on the eight frequency bands was extracted as the characteristic vectors which were the input of the RBF Neural Network. Finally, the RBF neural network model was established and trained, three kinds of fault signals of the roll bearing can be intdligent classified and recognized. The experiment result show that the intelligent method is availabe and effect.
出处
《计算技术与自动化》
2008年第4期115-117,127,共4页
Computing Technology and Automation
基金
江西省自然科学基金资助项目(0650099)
关键词
小波包
特征向量
RBF
智能分类
the wavelet packet
characteristic vector
RBF
intelligent classify