Application And Research Of HHT In Vibration Signal Of Flexible Thin-wall Bearings

Flexible thin-walled bearings are an important part of precise harmonic reducer,and precise harmonic reducer plays an important role in industrial robots.The damage of flexible thin-walled bearings will affect the operation accuracy of industrial robots and the operation efficiency of machines.Compared with general rolling bears,flexible thin-walled bearings are more prone to failure.Therefore,it is necessary to find a way to detect the running state of flexible thin-walled bearings as a theoretical guide for engineering.The vibration signal is collected to analyze the health of the bearings.In the past studies,most of the spectrum analysis methods are used to observe the characteristics of the bearings,but these methods have the disadvantage of poor practicability.The Hilbert transform(HHT)method can observe the frequency change process of the signal well in the time-frequency domain.So HHT is a good tool for observing the actual process of signal changes.In this paper,the HHT algorithm is deeply studied and applied to the detection of vibration signals of flexible thin-walled bearings.The vibration process of vibration signals is observed indirectly through Hilbert spectrum,and it is applied to the fault identification of flexible thin-walled bearings.The main contents of this paper are as follows:(1)Research the loading mode of operating flexible thin-walled bearings during,and an experimental station is designed to simulate the running environment of flexible thin-walled bearings.Vibration signals of bearings during operation are collected by acceleration sensors.Research the motion model of flexible thin-walled bearings and the fault model of inner and outer rings,and reveal the shortcomings of analyzing these models by traditional frequency domain analysis method;(2)Deeply study the HHT algorithm,and apply it to the motion model of flexible thin-walled bearings and the fault model of inner and outer rings.And then analyzing the advantages of HHT algorithm in dealing with non-state and non-linear signals.At the same time,revealing the shortcomings of HHT algorithm in analyzing actual noisy signals.(3)Noise in the signal will affect the analysis results of HHT algorithm,so it is necessary to purify the collected signal.This paper proposes a PCA signal extraction algorithm based on SVR spectrum.compare with the wavelet,EMD and SVD algorithms,PCA signal extraction algorithm is good at extracting non-linear vibration process from noised signals.Then using the PCA algorithm and HHT algorithm to analyze the signals collected by the station.(4)Vibration signals of normal bearing,outer ring fault bearing and inner ring fault bearing are collected,three different signals are processed by PCA signal extraction algorithm and HHT algorithm,and the frequency variation law of flexible thin-walled bearing in different states is studied in Hilbert spectrum,which provides theoretical guidance for practical engineering application.