Fault Feature Extraction And Monitoring Analysis Of Crawler Walking Device

Tracked vehicles are widely used in coal mine,agriculture,construction machinery and construction industry with the advantages of good adhesion and slope-crossing ability.Because crawler walking device often works on the uneven road,the working conditions are bad,and bears a large impact load,it makes the driving wheel,track board,pin shaft and gear teeth of the reducer easy to wear,crack and break.Once walking device fails,it will affect normal production and even cause serious accident.Therefore,it is very important to monitor crawler walking device effectively and identify the fault in the early stage according to the running state of the equipment.Based on this,this article takes small tracked vehicle as the research object,carries on fault diagnosis research of theory,simulation and test for walking device,which have positive significance to assure work reliability of crawler walking device.Firstly,dynamic model of crawler walking device is established by Solidworks and Recur Dyn,the flexible treatment of the vehicle frame is carried out by using ANSYS,and the rigid-flexible coupling model of crawler walking device with broken teeth fault of driving wheel is established.The acceleration signal is extracted from the inner part of the vehicle frame,the beam and the corresponding node of the reducer shell to simulate the vibration of the vehicle frame.The state of driving wheel with broken teeth have a little influence on the transmission of power and the vibration of frame,and creat weak vibration impact on frame,which make extracting broken teeth signal get difficult.In order to solve this problem,the wavelet packet and cepstrum analysis are applied to extract fault feature,the meshing frequency range of the fault location is obtained by wavelet packet decomposition,and the rotation frequency of the fault location is obtained by cepstrum analysis.The results show that signal amplitude in meshing frequency 14 Hz and higher harmonic is obviously increased in fault state,inverted frequency which has a convex peak in many places was converted to frequency,which is close to rotational speed frequency or doubling frequency of driving wheel,and the energy in the first band of wavelet packet energy spectrum increases obviously,and meshing frequency and rotation frequency of driving wheel are located in this frequency band.Secondly,the simulation model of gear transmission system of the reducer is established and the crack,1/2 broken teeth and total broken teeth of the active wheel of the third-stage gear meshing are simulated.The dynamic simulation of gear transmission system of the reducer is carried out in ADAMS by combining the calculation method of gear engagement transmission of Hertz contact collision theory.The results show that the meshing force has an obvious periodic impact in the time domain under the fault condition,and with the fault degree increase of the crack,1/2 broken teeth and full broken teeth,the larger the megnitude of the meshing force impact is,the more the meshing frequency appears in the frequency domain.In the frequency domain,not only the meshing frequency appears,but also the fault frequency is used as the the edge band at the meshing frequency and its doubling frequency,and the amplitude of the side band increases with the fault degree increase of the crack,1/2 broken teeth and full broken teeth.Finally,road vibration test of normal and fault condition was carried out for the small tracked vehicle in the laboratory.By extracting diagnosis feature of the test vibration signal and comparing simulation results,it is found that test results and simulation signal characteristic are in good agreement,which proves that fault modeling and simulation of crawler walking device are really reasonable.Aiming at the vibration signal collected by the test,identified different vibration signals of crawler walking device by using state recognition method based on HMM.This method is that the energy distribution of frequency band by wavelet packet decomposition was regarded as feather vector of HMM,establishes HMM model library which contains two operating state of normal and broken teeth of driving wheel,then input the test samples into HMM model library,respectively test the matching results,so running state is identified according to the maximum likelihood logarithm value.The results show that the state recognition result is good and the recognition rate is as high as 90%,the diagnosis method based on HMM has achieved an ideal effect on the classification of two states: normal and broken teeth.In this paper,the dynamic modeling and numerical simulation of crawler walking device,fault composite diagnosis of crawler walking device,fault simulation analysis of walking reducer,typical fault test of crawler walking device and condition recognition of crawler walking device based on HMM are researched,which have certain reference value for fault diagnosis and condition monitoring of crawler walking device.