Bearing Fault Diagnosis Method Based On Improved Multi-mode Fusion Convolutional Neural Network

Bearings,as the basic parts of most machinery manufacturing,have an irreplaceable position in rotating machinery.The quality of bearings directly affects the working state of large mechanical equipment,so it is particularly important for bearing fault diagnosis research.At present,bearing fault diagnosis methods based on deep learning have attracted widespread attention,especially the bearing fault diagnosis method based on Convolutional neural network has achieved good results in the field of bearing fault diagnosis.This paper mainly studies the bearing fault diagnosis method based on multi-modal feature fusion technology and Convolutional neural network for rolling bearings..The main research contents are as follows:(1)In order to improve the accuracy of bearing fault diagnosis and build a bearing fault diagnosis model with certain anti-noise ability,we propose a Convolutional neural network fault diagnosis method based on multi-modal feature fusion,the residual network processing mechanism for feature extraction based on time-domain signals,the multi-channel processing mechanism for feature extraction based on frequency-domain signals,and the superposition convolution processing mechanism for feature extraction based on two-dimensional composite signals are used to realize the time-frequency domain and other three feature extraction of different modal features,and based on the feature fusion mechanism,a multi-modal feature fusion Convolutional neural network bearing fault diagnosis model was constructed.Experiments are carried out on the bearing data set of Case Western Reserve University,and the results show that the model has good applicability and good anti-noise ability.(2)In order to further improve the anti-noise ability of the above model,the original multi-modal feature fusion model is improved by adding an attention mechanism.Under certain parameter conditions,the model performs attention mechanism channel screening on the fusion feature map generated by the above model mechanism,assigns appropriate parameters to each different mode,and uses the fused feature map for subsequent experimental operations to improve the accuracy of the original model bearing fault diagnosis.