Feature Extraction Of Rolling Bearing Based On Variational Mode Decomposition

Rolling bearing as one of the core components in the rotating machinery,often has far-reaching significance.Therefore,how to extract the fault feature effectively plays an important role in the operation of rolling bearing.In view of this,the rolling element bearing is taken as the research object in this thesis,and a series of related research work is conducted according to the characteristics of rolling bearing,as well as the extraction method.The main research contents are as follows:Firstly,as rolling bearing occupies an extremely important position in rotating machinery,the background and significance of the project are elaborated on the basis of theoretical analysis.Then a more comprehensive description of the rolling bearing is provided,which contains the current situation and the development trend of failure diagnosis method in fault diagnosis.Lastly,the bearing characteristics of each type fault,and the calculation of characteristic frequency in rolling bearing are introduced.Secondly,since the rolling bearing fault signal is susceptible to noise interference,resulting that the fault feature of rolling bearing is difficult to exact,To solve this,the signal analysis method of variational mode decomposition(VMD)combined with independent component analysis(ICA)is put forward in this thesis.At first,VMD is used to decompose the multi-component signal into a number of quasi-orthogonal intrinsic mode functions,and then the IMFs are reconstructed as the input matrices of ICA based on kurtosis criterion.After that,the source signal and noise signal can be separated with the help of FastICA algorithm.Finally,the signal is processed with envelope demodulation analysis,and then the theoretical value is compared with extracted frequency to identify the fault type.Thirdly,a fault diagnosis study based on variational mode decomposition and wavelet packet transform is put forward because the wavelet packed transform has several advantages,such as providing more detailed analysis for the signal,or ensure effective decomposition of both high and low frequency,so as to extract the fault characteristics accurately.Moreover,teager energy operator is also introduced to extract the fault features by analyzing the teager energy spectrum.Lastly,the proposed method is applied to actual signals.The experimental results show that the proposed method can not only reduce the effects of noise but also extract the fault features effectively.Finally,as an extension of wavelet transform,a new time-frequency analysis method called frequency slice wavelet transform is proposed.In this method,Fourier transform has the function of realizing time-frequency analysis and can realize the signal filtering and segmentation flexibly by introducing the frequency slicing function,estimating the frequency resolution ratio ?,the expected response amplitude ratio ?,and calculating the initial time-frequency resolution coefficient k.With this method combined with VMD method and the multi-correlation algorithm,the vibration signal is decomposed by VMD,and then the multi-correlation process is made for the signal in order to prominent fault signal characteristics.After that,FSWT is used to obtain the time-frequency distribution of the vibration signal in the whole frequency band,and the time-frequency distribution of the vibration signal in the whole frequency band is selected to divide the time-frequency region including the fault feature according to the energy distribution characteristics of the obtained vibration signal.Finally the signal component of the target region is reconstructed by inverse transformation and the fault feature information is extracted.The results show that the proposed method can effectively extract the fault characteristics of rolling bearings,and the validity of the proposed method is verified.