| As the core component of rotating machinery,deep groove ball bearing has a series of advantages such as high precision,good substitutability,low cost and easy batch manufacturing.It has a wide range of applications in the industrial field.At the same time,with the continuous advancement of the industrial field,many machines are moving toward high speed and heavy load,which inevitably increases the requirements for the performance of deep groove ball bearings.Therefore,research on deep groove ball bearings is of great significance to improve the stability of mechanical equipment operation and reduce the maintenance cost of equipment.In this thesis,the deep groove ball bearing is taken as the research object,and the fault feature extraction method and diagnosis method are researched,and the fault diagnosis system is developed.The main contents of the thesis are:(1)A fault signal feature extraction method based on nonlinear mode decomposition(NMD)and wavelet threshold denoising is proposed.Firstly,the wavelet signal is used to denoise the original fault signal.Then,the NMD is used to decompose the denoised signal,and then the envelope spectrum of the reconstructed signal is drawn.Finally,the relevant characteristic frequency of the deep groove ball bearing can be extracted.Then,the simulation experiment and the deep groove ball bearing fault signal experiment were carried out separately,and the experimental results were compared with the results of empirical mode decomposition(EMD)processing.At the same time,two parameters of root mean square error and signal to noise ratio are established to analyze the fault feature extraction ability of the method.(2)The fault diagnosis model of deep groove ball bearing based on wavelet packet transform for noise reduction processing and feature extraction and improved BP neural network is studied.The first is the noise reduction process,which uses the wavelet packet transform method to filter out the noise interference in the vibration signal.Then,the wavelet packet transform method is used to extract the energy characteristics of the denoised signal,and the feature value is used as the input matrix of the neural network;BP neural network realizes the fault classification of deep groove ball bearings.According to the related shortcomings of BP neural network,a BP neural network diagnosis model based on beetle antennae search algorithm is proposed and compared with other BP neural networks to verify the validity of the diagnosis results.(3)The principle of the algorithm is introduced,and a fault diagnosis method based on time-frequency diagram and convolutional neural network is proposed.First,the signal is subjected to ensemble empirical mode decomposition and reconstruction,and then converted into a time-frequency diagram,and gray scaled.After constructing the training set and the test set,the training set is first trained by the convolutional neural network,and then the trained set is used to identify and classify the fault of the test set.(4)According to the proposed deep groove ball bearing fault diagnosis method,the fault diagnosis system of bearing is designed by MATLAB software,including software registration,bearing fault diagnosis and bearing fault identification.It has functions such as system login,parameter setting,fault feature extraction,and fault diagnosis. |
PDF Full Text Request