Research On Prediction Of Remaining Life Of Rolling Bearing Driven By Degradation Data

Rolling bearings,as the most widely used basic parts in mechanical equipment,are called mechanical joints.Whether the mechanical equipment can operate normally,realize the predetermined function and reach the expected l ife depends largely on the performance,reliability and life of its basic components.Rolling bearings are prone to failures in high-speed,heavy-load or harsh environments,which leads to a significant reduction in their actual life compared with the expe cted life.Therefore,fault diagnosis and life prediction of rolling bearings have important theoretical significance and engineering practical value for the effective implementation of condition-based maintenance and health management of mechanical equipm ent.At present,there are many methods for fault diagnosis of rolling bearings,but most of the signal decomposition methods have problems in analyzing the multi-channel fault signals of the same bearing.For the study of rolling bearing life prediction,there also exists the problem that the reliability and error precision are low when only one single degenerate variable is extracted to predict the performance degradation trend of rolling bearing.On th e basis of the above statement,this paper studies the fault diagnosis and residual life prediction of rolling bearings,the main research work is as follows:1.Single channel damage assessment of rolling bearings based on CEEMD and Lempel-Ziv complexity.First of all,vibration signals of rolling bearings in a single channel are decomposed with CEEMD,t he effective IMF component is picked up based on kurtosis criterion so that the index of Lempel-Ziv complexity can be calculated.Finally,the damage degree of rolling bearing is judged according to the change of the index.2.Evaluation on the multi-channel fault of rolling bearing based on MEMD and RVM.First of all,vibration signals of rolling bearings in multi-channel are decomposed with MEMD,effective MIMF component is selected through correlation analysis method and the mutual approximate entropy is extracted.Finally,the RVM classifier is used to analyze and evaluate the multi-channel outer ring fault of rolling bearing.3.Residual life prediction of rolling bearing based on grey prediction model with multiple degenerate variables.First of all,the characteristic parameter set related to degradation vibration signals of lifetime bearing is extracted.Identification of the initial failure time of rolling bearing is completed combined with parameter set and early breakdown point,then a grey prediction model with multiple degenerate variables is established according to the mapping relationship between bearing life and characteristic parameters,the identification parameters of the model are calculated.Finally,the residual life of the bearing is predicted by the grey prediction model with multiple degenerate variables.Aiming at the two major problems: fault diagnosis and residual life prediction of rolling bearing,which is the most vulnerable core comp onent of mechanical equipment.The paper proposed three methods to study fault diagnosis and residual life prediction of rolling bearing from the angle of theoretical analysis and engineering application verification.The method s can achieve the expected research objectives and has a certain practical significance.