Singular Spectrum Decomposition Method And Its Application In Rolling Bearing Fault Diagnosis

The monitoring and diagnosis of rolling bearings is of great significance for ensuring the stable operation of mechanical and electrical equipment,reducing shutdown and avoiding major economic losses.Bearing signal is a non-stationary signal,and because of noise interference,aiming at the problem of early fault diagnosis of bearings,this paper studies the principle of singular spectrum decomposition method,and studies the bearing diagnosis method based on singular spectrum decomposition and envelope spectrum,minimum entropy deconvolution,etc.Finally,it is verified by experimental signals.The main contents of this paper are as follows:1.The related theory and decomposition effect of singular spectral decomposition are introduced and simulated.The main contents include: introducing the theory of singular spectral decomposition method,putting forward the singular spectral decomposition method based on mirror continuation,solving the end-point effect defect of singular spectral decomposition method;comparing the decomposition effect of singular spectral decomposition method and empirical mode decomposition by simulation signal decomposition,finding the singular spectral decomposition method.The method has more advantages in overcoming the defects of endpoint effect and mode aliasing,and has better adaptability to noise decomposition.2.Aiming at the difficulty of artificial selection of filtering frequency band in resonance demodulation method,a bearing recognition method based on singular spectral decomposition and envelope spectrum is proposed,and the effectiveness of this method is verified.3.Aiming at the problem that noise signal affects the effect of singular spectrum decomposition,a bearing diagnosis method based on singular spectral decomposition and minimum entropy deconvolution is proposed.The bearing signal is decomposed by singular spectral decomposition and the first eigenmode component is de-noised by minimum entropy deconvolution,which can improve the ability of spectrum analysis feature extraction,and the effect of rolling bearing fault feature extraction is good.Compared with envelope spectrum analysis,the result is better than that of single envelope spectrum analysis.4.Aiming at the problem that noise signal affects the effect of singular spectral decomposition,a bearing diagnosis method based on stochastic resonance and singular spectral decomposition is proposed.Stochastic resonance is used to de-noise the signal,then singular spectral decomposition is used to decompose the bearing signal,and the first component is analyzed by frequency spectrum.The effect of bearing fault feature extraction is good.5.The influence of bearing noise on signal singular spectrum decomposition is analyzed,and a bearing diagnosis method based on singular spectral decomposition and threshold noise reduction is proposed.Firstly,the signal is decomposed by singular spectral decomposition,then component noise reduction is selected.Finally,the method is validated by an example of bearing inner and outer ring diagnosis.6.Summarize the full text and put forward research prospects.The main summary of the paper and the future research work are prospected.