Fault Feature Extraction Method Of Rolling Bearing Under Variable Conditions Based On LMD And Bandpass Filter

The current detection of the running state of the bearing is mostly through the resonance demodulation technology.However,in the actual operation of the bearing,it is unavoidable to have the disturbance of variable speed and gear meshing,so that the vibration signal is no longer satisfied with the premise of resonance demodulation.At the same time,bearing is one of the most common standard parts in modern industrial equipment.Its running state is directly related to the performance of the whole equipment and the whole system.Thus,under the disturbance of variable speed and gear meshing,the research on fault diagnosis of bearing is becoming more and more important.In this paper,taking the rolling bearing as the research object,the feature extraction technology under variable speed and gear meshing interference is discussed and studied.The main contents are as follows:The advantages and disadvantages of spectral kurtosis and bandpass filtering technology are compared and analyzed.The results of simulations and experiments show that the bandpass filtering technology is better than spectral kurtosis analysis to a certain extent for the extraction of bearing fault characteristics.Aiming at the modal aliasing and end-point effect of the empirical mode decomposition(EMD)in the process of transient signal,in this paper,the ensemble empirical mode decomposition(EEMD)algorithm is used to solve the modal aliasing problem of EMD when dealing with interrupted signals.The local mean decomposition(LMD)algorithm largely inhibits the endpoint effect of EMD.Finally,through the comparison of the three algorithms,it is found that the LMD algorithm is superior to the EMD and EEMD algorithms in extracting the effective components of the signal and suppressing the noise components.Aiming at the influence of variable speed,the instantaneous dominant meshing multiply(IDMM)is equivalent to the bearing rotation frequency,in order to calculate the phase discrimination time scale with constant angular increment,then the original time domain non periodic signal is interpolated and resampled,and it is converted to a angular domain periodic signal with constant angular increment.The results of simulations and experiments show that this method not only eliminates the frequency ambiguity phenomenon of envelope spectrum under variable speed,completes the fault identification,but also gets rid of the dependence on the speed measuring device,overcomes the limitation of installation space and cost to the order ratio technology to a certain extent.For the influence of the gear meshing interference,this paper established the simulation model of the vibration signal of rolling bearing under variable speed and gear meshing.IDMM is used to resample the simulation model in the angle domain.The linear prediction algorithm(AR model)is adopted to predict and eliminate the gear meshing interference,The author finally eliminated the influence of the variable speed,and realized the effective separation of the bearing vibration signal and the gear meshing interference.The LMD algorithm can decompose the vibration signal into a series of production components(PF components).However,the extraction of bearing fault features for each PF component wastes computing resources.In this paper,the correlation coefficient is used to filter the PF components.Then the filter results are demodulated by resonance to obtain the characteristics of the bearing fault,and this method greatly reduces the amount of calculation.Based on the analysis of the existing bearing fault diagnosis technology,this paper presents a method for fault diagnosis of rolling bearing based on LMD and bandpass filter under variable working conditions.This method combines LMD,IDMM,order tracking,bandpass filtering technology and envelope demodulation.It is verified by simulations and experiments that the method is accurate and feasible.