Research On Fault Diagnosis Of Rolling Bearings For Tamping Vehicles Based On Multi-wavelet Analysis

Because of its safety and large transport,railway becomes one of the major modes of transport,and tamping machine is mainly for the railway operations and maintenance.Due to the work environment and the strength of the work,tamping machine will have different problems,and rolling bearings as an important component,the failure of its will affect the tamping machine’s normal operation.Therefore,it is very important to carry out research on fault diagnosis of tamping machine rolling bearing.In the light of fault vibration signal,this paper firstly carried on fault information de-noising based on improved multi-wavelet threshold,and then studied a fault feature extraction method based on the combined empirical mode and energy entropy.At last,this paper adopted the classification method based on fuzzy neural network as far as the fault diagnosis concerned,and the simulation experiment is carried out.This paper introduces the theory of wavelet analysis and multi-wavelet analysis.On this basis,it proposed a multi-wavelet threshold de-noising with improved threshold function,and then studied a fault feature extraction method based on the combined empirical mode and energy entropy.Both of.which are applied to the tamping machine rolling bearing fault signal analysis.The multi-wavelet function and preprocessing method are firstly selected for the signal,and the correlation coefficient and the low frequency kurtosis ratio are taken as the evaluation criteria.Then noisy signal was adopted improved threshold function to carry on multi-wavelet layered de-noising.Finally,the de-noising signal used the overall empirical mode State decomposition to obtain a number of IMF components,and used energy entropy to form the fault feature vector.In this paper,the theory of fuzzy neural network is introduced.And the application of adaptive fuzzy neural network in fault classification of tamping machine was researched.The results are compared with BP neural network.The experimental results show that the improved multi-wavelet threshold de-noising analysis can achieve better de-noising effect,and the fault feature extraction method used in this paper can provide rich and accurate fault characteristic information for subsequent fault diagnosis.The simulation results show that the adaptive fuzzy neural network can be applied to the fault type of the tamping machine.