Research On Fault Diagnosis Method Of Rolling Bearing Based On Feature Extraction And Sparse Representation

Rolling bearing is one of the most widely used parts in mechanical equipment.It plays a role of supporting and transmitting loads in mechanical equipment.Due to the poor working conditions,it is also one of the most easily broken parts.Once the rolling bearing fails,it will affect the operation of the entire mechanical equipment.In serious cases,it will cause major safety accidents.Therefore,it is very important to carry out fault diagnosis and condition monitoring of rolling bearings.Fault diagnosis of rolling bearings mainly includes two parts:fault feature extraction and pattern recognition of rolling bearings.Based on this,three different methods for fault feature extraction of rolling bearings are studied in this paper.Variational Mode Decomposition(VMD)Manifold learning and deep learning are used to extract fault features of rolling bearings.A method for fault identification of rolling bearing based on sparse representation is proposed.The extracted fault features are input into the Sparse Representation Classifier(SRC)to diagnose the fault type of the rolling bearing.The main research contents of this paper are as follows:1.Aimed the fact that the collected vibration signal usually contains more background noise,a VMD denoising method is proposed.The simulation results show that the VMD denoising can effectively reduce the noise component in the signal,and the noise reduction effect is better Empirical Mode Decomposition(EMD)denoising,wavelet threshold denoising and other methods.2.The manifold learning method is used to extract the characteristics of rolling bearings.Studying the optimal embedding dimension of the manifold learning for extracting the feature of rolling bearing.The extracted features are combined with the sparse representation classifier for fault diagnosis of rolling bearings.Experimental result shows that this method can effectively identify various types of failures of rolling bearings.3.The vibration signal is decomposed into several modal components by VMD,and the kurtosis,energy moment,and fuzzy entropy of each component are extracted to form the combined features of the signals.Then,the optimal characteristics of the combined features are selected by the method of feature selection.The extracted features are combined with the sparse representation classifier for fault diagnosis of rolling bearings.Experimental result shows that this method can effectively identify various types of failures of rolling bearings.4.The Convolution Neural Network(CNN)and Deep Belief Network(DBN)network models respectively with training samples are trained.The Trained CNN and DBN are used to extract fault features of rolling bearings.The extracted features are combined with the sparse representation classifier for fault diagnosis of rolling bearings.Experimental result shows that this method can effectively identify various types of failures of rolling bearings.