基于WPD-CNN二维时频图像的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based on WPD-CNN Two-Dimensional Time-Frequency Image

滚动轴承故障诊断是现代工业发展中的重要技术。针对滚动轴承信号特征提取与智能诊断问题,提出了一种基于WPD-CNN二维时频图像的滚动轴承故障诊断方法。首先通过小波包分解(WPD)将信号转换为二维时频图像;其次将时频图像输入VGG19卷积神经网络(CNN)模型自动提取有效特征,并输入Softmax分类器进行训练;最后使用训练好的分类器完成滚动轴承故障诊断任务。实验结果表明,10类故障数据的识别准确率均在98.3%左右,高于其他深度学习和传统方法,因此所提出的故障诊断模型能有效地进行滚动轴承复杂信号的特征提取以及分类任务。

Rolling bearing fault diagnosis is an important technology in the development of modern industry.Aiming at the problem of feature extraction and intelligent diagnosis of rolling bearing signals,this paper proposes a fault diagnosis method for rolling bearing based on WPD-CNN two-dimensional time-frequency images.First,the signal is converted into a two-dimensional time-frequency image by wavelet packet decomposition(WPD).Second,the time-frequency image is input into the VGG19 Convolutional Neural Network(CNN)model to automatically extract effective features,and input into the Softmax classifier for training.Finally,use the trained classifier to complete the fault diagnosis task of rolling bearings.The experimental results show that the recognition accuracy of the 10 types of fault data is about 98.3%,which is higher than other deep learning and traditional methods.Therefore,the proposed fault diagnosis model can effectively perform the feature extraction and classification tasks of complex signals of rolling bearings.