Research On Fault Feature Extraction And Residual Life Prediction Of Rolling Bearings

Rolling bearing plays an important role in some large high-speed rotating and heavy-load machines,once it fail,it can likely cause paralysis of the entire system,leading to delays in production or shutdown which can even cause serious damage to people’s lives and property The irreversible loss.Therefore,it is necessary to study the fault diagnosis and the prediction of the remaining life of the rolling bearing.This paper does research on the fault diagnosis and the prediction of the remaining life of the rolling bearing,the details are as follows:(1)Based on the exposition of the Variation Mode Decomposition algorithm theory,the variation model decomposition algorithm is verified by the simulation signal of the outer ring fault of the rolling bearing,and it is compared with the decomposition results of the Empirical Mode Decomposition.In order to solve the problem of parameter selection of variable mode decomposition algorithm,a grid optimization algorithm aimed at envelope sparseness is provided.(2)Aiming at the components obtained by the variation modal decomposition,the entropy method is proposed to filter the optimal resonant frequency band.Firstly,four characteristic indexes are extracted as the basis of screening and the entropy method is used to assign the weights to the four characteristic indexes,the value of the information utility entropy of each variable modal decomposition component is calculated and the optimal resonant frequency band is screened out by its size.(3)The bearing life is closely related to the bearing degradation state,in order to accurately predict the remaining life,this paper presents an operating state partitioning method,which divides the entire life cycle of the bearing into stationary period,degeneration period and rapid failure period(retirement period),so as to establish the remaining life model for bearings with different states.It is proposed that the root mean square of the rolling bearing fault signal is used to divide the degradation state of the rolling bearing,the degradation process of rolling bearings under different working conditions is divided into different stages(stationary period,degradation period and scrap period);(4)According to the degeneration state of different rolling bearings,the remaining life prediction of the stated rolling bearing is put forward.In order to verify the effectiveness of the proposed algorithm,the residual life of the test bearing under the same working conditions is predicted by using the bearing life data of the bearing obtained under a certain condition.In view of the shortcomings of this paper,in the sixth chapter of the paper,the future workof the next step is in the planning to hope the future study can continue to improve.