Fault Feature Extraction Of Flexible Thin-Walled Bearings Based On EEMD And MCKD

The precision flexible thin-walled bearing is one of the core components of the harmonic reducer.The stability of flexible thin-walled bearing determines the performance of the harmonic reducer.Therefore,it is particularly important to study the fault diagnosis method of the flexible thin-walled bearing.Different from ordinary rolling bearings,the flexible thin-walled bearing is installed on an elliptical shaft,the pitch diameter changes during the shaft rotating,so its fault characteristic frequency has a time-varying characteristic,and it will also generate a periodic impact component under no fault conditions.Thus,when a bearing fails,the weak fault impact will drown,greatly increasing the difficulty of extraction.In view of the above problems,this paper takes flexible thin-walled bearings as the research object,study vibration signal fault feature extraction method.The main research contents of the paper are as follows:(1)Based on the kinematics principle of flexible thin-walled bearings,the formulas for calculating the characteristic frequencies of faults of the inner and outer rings and rolling elements of flexible thin-walled bearings are derived.A vibration signal acquisition device that simulates the actual working conditions of a flexible thin-walled bearing is expounded,and the characteristics of the vibration signal of a flexible thin-walled bearing are analyzed(2)According to the damage signal characteristics of flexible thin-walled bearings,a fault feature extraction method based on integrated index integrated empirical mode decomposition(EEMD)is proposed.A comprehensive index of kurtosis,fault characteristic ratio and related kurtosis is proposed to select IMF components for reconstruction to obtain fault characteristic signals.The results show that due to the effective selection of the decomposed IMF,the EEMD method based on comprehensive indicators accurately extracted the fault characteristic frequencies of the inner and outer rings of the flexible thin-walled bearing.(3)In view of the problem that the weak fault impact of flexible thin-walled bearings is easily submerged in the background interference component,the maximum correlation kurtosis deconvolution method(MCKD)is studied and improved to improve the calculation accuracy.The improved MCKD algorithm is used to extract the fault vibration signals of the inner and outer rings of flexible thin-walled bearings.The results show that the improved MCKD algorithm can extract fault features from complex background interference signals,effectively suppress noise components,and the extraction effect is better than ordinary MCKD methods.(4)Finally,the characteristics of the two algorithms EEMD and MCKD are analyzed,for the EEMD algorithm based on comprehensive indicators is easy to miss the IMF containing the impact component when the signal is reconstructed,EEMD-MCKD feature extraction method based on kurtosis is proposed.Taking the fault feature ratio and correlation kurtosis as indicators,the results show that this method can accurately extract fault features in the signal,and is superior to the EEMD algorithm based on comprehensive indicators.