【摘要】 针对滚动轴承故障种类繁多,故障信号特征不明显的问题,提出了一种小波包能量与卷积神经网络相结合的滚动轴承故障判别方法。首先对原始振动信号进行小波包分解,其次求取分解后各个子带信号的能量,归一化后得到一组特征向量,最后将该特征向量作为卷积神经网络的输入,进而判断输入信号所对应的故障类型。为验证所提方法的有效性和优越性,采用美国凯斯西储大学轴承数据集,将所提出的方法与另外两种故障诊断算法进行对比。在不同工况情况下的对比试验结果表明,小波包能量特征提取方法,能够有效提取出原始信号故障特征。相较于常见的卷积神经网络的故障诊断方法,所提方法能够有效提高故障识别准确率,且速度快、稳定性好。更多还原

【Abstract】 In view of the various fault types of rolling bearings and the inconspicuous fault signal characteristics,a fault diagnosis method based on wavelet packet energy and convolutional neural network was proposed. Firstly,the original vibration signal was decomposed by wavelet packet,and the energy of each sub-band signal was obtained to a set of eigenvectors as the input of the convolutional neural network,and then the corresponding fault type was determined. To verify the effectiveness of the proposed method,the bearing data set of Case Western Reserve University was used to compare the proposed method with the other two fault diagnosis algorithms. The experimental results under different working conditions show that Wavelet packet energy feature extraction method can effectively extract the original signal fault diagnosis. Compared with the common method of convolutional neural network,this method can not only effectively improve the fault identification accuracy,but also has a high speed and good stability.更多还原