Fourier Decomposition Method (FDM) And Its Application Research In Rolling Bearing Fault Diagnosis

As an indispensable transmission component,rolling bearing is widely used in rotating machinery.However,rolling bearings are very prone to failure in actual production,due to the high working intensity and harsh operating conditions of rotating machinery.Therefore,it is extremely important to real-time monitor and effectively diagnose of rolling bearings.The core to fault diagnosis of rolling bearing lies in extracting and identifying information related to fault characteristics from vibration signals.The popular signal decomposition methods in the field of mechanical fault diagnosis,such as Empirical mode decomposition(EMD),Empirical Wavelet Transform(EWT),and Variational mode decomposition(VMD)have been extensively studied by many scholars.However,these methods have their inherent defects,which limit their effective applications in practice.The Fourier decomposition method(FDM)is a newly proposed signal decomposition method that can decompose the analyzed signal into several single components in different frequency bands.The method has a complete theoretical foundation,no need to preset basis functions,and the decomposition is complete and adaptive.Under the funding of the National Natural Science Foundation of China(No.51975004),this article studied the principle of FDM algorithm.At the same time,corresponding solutions have been proposed to solve the theoretical problems existing in FDM.Through the comparative study of simulation and measured signals,the results demonstrated that the proposed method is superior to other existing methods.The main research contents and innovations of the thesis are as follows:(1)Research on FDM theory and algorithm,and compare FDM with other decomposition methods for the same simulation signal.The results show that FDM method overcomes the inherent shortcomings of EMD.At the same time the inherent components of the original signal can be better restored,which is superior to the compared methods in terms of signal fidelity.(2)Since the decomposition effect of FDM is easily affected by noise,by introducing the maximum correlation kurtosis deconvolution(MCKD)before decomposition,a fault diagnosis method of rolling bearing based on MCKD and FDM is proposed.By extracting periodic pulses related to the fault through MCKD,the noise interference in the vibration signal is reduced,thereby improving the decomposition effect of FDM.The superiority of the proposed fault diagnosis method is verified by the comparative analysis of simulation and measured signals.(3)To solve the problem that the FDM decomposition is prone to excessive components,an order-statistic filtering Fourier decomposition method(OSFFDM)is proposed using the advantages of order-statistic filtering(OSF)envelope processing.The proposed method is applied to filter and smooth the original frequency spectrum,which reduces the generation of false components.The diagnosis results and comparative analysis of rolling bearing vibration data show that the result obtained by OSFFDM meets the characteristics of rolling bearing faults and the diagnosis accuracy is high.(4)Inspired by the power spectrum of Burg algorithm,an adaptive power spectrum Fourier decomposition method(APSFDM)is proposed to solve the problem of FDM spectrum segmentation.The power spectrum is used instead of the Fourier spectrum to facilitate reasonable segment spectrum in APSFDM method.In the analysis of simulation signals,the decomposition results of APSFDM with EWT,EMD,VMD and FDM methods are compared,and the comparison results show the adaptability,completeness and superiority of the proposed method.Finally,the APSFDM method is applied to the actual collected fault signal diagnosis,and the adaptive multi-scale morphological filtering(AMMF)is used to reduce noise and improve the signal-to-noise ratio of the obtained components.The superiority of the proposed method combining APSFDM and AMM in bearing fault diagnosis is demonstrated through comparison of results.