Aiming at the problems of low classification accuracy and weak model generalization ability of traditional fault diagnosis classification methods in the fault diagnosis of rolling bearings in rotating machinery,an intelligent fault diagnosis model based on signal processing technology combined with deep learning algorithm was proposed.Firstly,the original data set was repeatedly divided according to a certain proportion to realize data expansion.Secondly,the extended bearing vibration signal is converted into a two-dimensional wavelet time-frequency graph by continuous wavelet transform method.Then,the improved convolutional neural network model was used to train the divided two-dimensional image set to extract the deep features of time-frequency images.Finally,the extracted feature vectors were input into the support vector machine classification layer with optimized parameters by cuckoo search algorithm to realize the fault classification of rolling bearings.The fault diagnosis classification model outputs the highest classification accuracy of 100%after training,and the accuracy is better than the other five fault diagnosis models in the anti-noise experiment and the variable load experiment.The results show that the combination of convolutional neural network to extract fault features and parameters to optimize the classification model structure of support vector machine can not only improve the diagnostic accuracy,but also have strong generalization performance.
关键词
滚动轴承/故障诊断/深度学习/卷积神经网络/支持向量机
Key words
rolling bearing/fault diagnosis/deep learning/convolutional neural network/support vector machine
维普
万方数据