Fault Diagnosis Of Rotating Machinery Based On Empirical Wavelet Transform And Singular Value Decomposition

The fault signal of the rotating machinery is usually a non-stationary,non-linear noisy vibration signal.For the mechanical fault diagnosis,the time-frequency analysis method such as window Fourier transform,Wigner distribution and wavelet transform(WT)is widely used.Have some limitations,and are susceptible to interference.Since the advent of empirical mode decomposition(EMD)method has been widely used in the diagnosis of rotating machinery,EMD decomposition has the advantage of adaptive decomposition,but because of its own modal aliasing effect,end effect As well as their lack of a certain theoretical basis,so there are still some problems in the application.(EWT)is inherited from the empirical modal decomposition(EMD)and is based on the theory of wavelet.Compared with the empirical mode decomposition,it greatly reduces the existence of empirical mode decomposition Modal aliasing phenomenon,and has a reliable theoretical basis,in the rotary machine fault diagnosis has a high application value,the author of the use of empirical wavelet transform adaptive singular value decomposition of the noise characteristics proposed a new rotating machine failure diagnosis method.In this paper,we introduce the empirical wavelet transform theory,compare the empirical wavelet transform and empirical mode decomposition to the decomposition results of multi-modal aliasing signal,and verify the advantages of empirical wavelet transform over empirical mode decomposition.Then the singular value decomposition and singular value packet decomposition theory is expounded,and the characteristics of singular value and singular value package are verified by simulation signal.Finally,the empirical wavelet transform is combined with the singular value decomposition and the empirical wavelet transform combined with the realization of the singular value packet decomposition algorithm,and the validity of the algorithm is verified by the simulation signal.In this paper,the joint algorithm is applied to the analysis of the fault diagnosis of rotating machinery.By selecting the bearing,the rotor and the universal shaft as the research object,the inner fault of the bearing,the fault of the rotor and the dynamic imbalance of the universal shaft Signal extraction,the application of the proposed algorithm to analyze the fault signal to verify the joint algorithm for the practical application of the effectiveness of the project.EWT-SVD shows good adaptability and good filtering performance,and it can be used to analyze the fault signal and the actual engineering fault signal analysis.Based on the empirical wavelet transform combined with singular value decomposition(EWT-SVD)and the noise is decomposed into different frequency characteristics,and compared with the results of separate empirical wavelet decomposition.The joint algorithm shows good filter noise.By analyzing the bearing fault signal,rotor fault fault signal,universal axis Based on the empirical wavelet transform combined with the singular value decomposition,the original vibration signal can be decomposed into the component signals in different frequency bands,and the frequency values in the vibration signals are clearly decomposed into different component signals.Based on the empirical wavelet Compared with the empirical wavelet decomposition,EWT-SVDP has not only improved the self-adaptability,but also improved the performance of the actual signal.The processing power of the actual signal is also greatly enhanced.