基于小波包分解和BP神经网络的轴承故障诊断  

Fault Diagnosis for Bearings Based on Wavelet Packet Decomposition and BP Neural Network

作　　者：叶瑞召[1] 李万红[2] 

机构地区：[1]洛阳轴研科技股份有限公司,河南洛阳471039 [2]河南科技大学机电工程学院,河南洛阳471003

出　　处：《轴承》2012年第10期53-56,共4页Bearing

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.

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 

分 类 号：TH133.33[机械工程—机械制造及自动化]