小波包分析和BP神经网络在滚动轴承故障模式识别中的应用

Fault pattern recognition of rolling bearing using wavelet package analysis and BP neural network

根据滚动轴承振动信号的不同故障模式在频域能量分布中的差异性,提出了基于小波包分解与重构和BP神经网络的轴承故障模式识别技术。论文首先对轴承振动信号进行小波包三层分解,完成了振动信号在空间的完整拆分,同时得到了第三层由低频到高频的小波包分解系数,再依据小波包分解系数分别重构各频段的信号,并提取信号各频段的能量。然后利用信号各频段的能量组成的特征矢量作为BP神经网络的输入样本,对BP神经网络进行训练,获得不同故障模式识别网络模型,最后利用测试数据对建立起来的BP神经网络模型进行检验,通过BP神经网络判别滚动轴承的故障类型。实验结果证明,采用小波包分解与重构和BP神经网络相结合的方法可以比较准确地识别滚动轴承的故障。

According to the frequency domain energy distribution differences of bearing vibration signal in the different failure modes,rolling bearing fault pattern recognition technology based on the orthogonal wavelet packet decomposition and BP neural network is proposed.The orthogonal three layer wavelet packet decomposition for rolling bearing vibration signal is carried out to get the third layer wavelet packet decomposition coefficients from low frequency to high frequency,then the different frequency band signal are reconstructed respectively to extract energy features by means of wavelet packet decomposition coefficients.Using the energy feature vector of different frequency band as the model input of the BP neural network model,a large number of samples are trained to get the network pattern recognition model for different bearing fault,then use several groups of test data are used to verify the BP network models to discrimination the type of rolling bearings fault.The test results proved that the method integrated the Wavelet packet decomposition with BP neural network can identify the fault of rolling bearings more accurately.