摘要
针对滚动轴承的故障诊断,提出了小波包分解与BP神经网络结合的诊断方法。对轴承振动信号进行3层小波包分解,构造其特征向量,输入神经网络进行训练和测试。Matlab仿真结果表明,该方法能有效地诊断出轴承的故障类型。
A method of combining wavelet packet decomposition and BP neural network is presented for fault diagnosis of rolling bearings. The eigenvector is constructed after decomposing vibration signal of bearings by three - layer wavelet packet, and then is inputted into the neural network to train and test. Matlab simulation results show that this method is effective to diagnose the fault types of bearings.
出处
《轴承》
北大核心
2012年第10期53-56,共4页
Bearing
关键词
滚动轴承
故障诊断
小波包分解
特征向量
BP神经网络
rolling bearing
fault diagnosis
wavelet packet decomposition
eigenvector
BP neural network