| As a key component of rotating machinery,the running state of rolling bearing is not only related to significant economic benefits,but also has a profound impact on social security.As a result,it is very important to monitor the running state and diagnose the fault of rolling bearing.With the advent of artificial intelligence era,fault diagnosis has a tendency to develop towards intelligent diagnosis.In this paper,rolling bearing vibration signal is taken as the research object,combined with empirical mode decomposition,singular value decomposition,phase space reconstruction and spiral matrix correlation means,two fault diagnosis methods of rolling bearing based on deep learning are proposed.The main contents of this paper are as follows:(1)This paper systematically introduces the research background and significance of rolling bearing fault diagnosis,the development stages of rolling bearing fault diagnosis and the common fault diagnosis methods.In addition,this paper expounds the important role of the advanced machine learning and deep learning theories in bearing fault diagnosis along with the promotion of intelligent diagnosis.(2)A fault feature extraction method for rolling bearings based on phase space reconstruction and singular value decomposition is proposed.Firstly,the Hankel matrix is used to reconstruct the phase space of the signal,then the singular value decomposition of the constructed matrix is carried out,and the relationship between the number of eigenvalues of singular values and the number of rows and columns of the matrix is discussed.This method is simpler and more effective than the traditional feature extraction methods in time domain,frequency domain and time-frequency domain.(3)A rolling bearing fault diagnosis method based on EMD and SSAE is proposed.First,the vibration signals under different states of rolling bearing are decomposed by empirical mode decomposition.Second,to extract more representative high-level features,the obtained intrinsic mode functions are preprocessed with singular value decomposition to acquire singular value parameters,which are regarded as the inputs of the proposed SSAE network.The proposed method does not need the signal denoising processing,simplifying the traditional process of feature extraction of rolling bearing fault diagnosis.(4)A method of bearing fault diagnosis based on convolution neural network with characteristic spiral arrangement is proposed.In order to restrain the loss of important information in the process of information transmission,the feature extraction method of phase space reconstruction and singular value decomposition is applied to bearing vibration signals.The singular values obtained are regarded as the features of evaluating the health status of rolling bearings.Then these features form a spiral matrix from the center to the edge as the input of convolution neural network.The proposed method effectively resists the loss and loss of edge information in the process of network transmission.(5)A practical and easy-to-operate intelligent diagnosis system for rolling bearings is developed by using Python language and Pycharm IDE.The system takes the time-domain vibration signal of rolling bearing as input,and outputs the fault diagnosis results after system calculation and analysis,thus realizing the intelligent diagnosis of rolling bearing. |
