Research On Residual Service Life Of Rolling Bearing Under Variable Working Conditions

Rolling bearing is one of the indispensable parts in rotary machinery,Its main function is to support the mechanical rotating body,reduce the friction coefficient in rotary motion,and ensure its rotation accuracy.However,in the long-term operation process,due to improper initial assembly,manufacturing inaccuracy or inadequate lubrication and other reasons,defects and wear of rolling bearings are unavoidable.How to accurately identify the fault and residual life of rolling bearings has attracted a large number of scholars at home and abroad to study this topic for decades.Envelope analysis and demodulation techniques are widely used in fault diagnosis of rotating machinery due to the damage of defects affecting other parts of the rotating system and generating amplitude modulation vibration.However,in actual working conditions,the data of a single channel cannot fully reflect the signal characteristics.The full-vector spectrum technology adopts the homologous dual-channel information fusion technology,which can more comprehensively show the characteristic information of the vibration of rotating equipment,effectively avoiding the disadvantage that the single channel data cannot fully reflect the vibration information.In order to overcome the problem of modal mixing of EMD algorithm,based on EMD algorithm,EEMD algorithm is developed and widely used in signal noise processing.To sum up,according to the respective advantages of EEMD algorithm and full-vector spectrum technology,a new method combining EEMD and full-vector spectrum is proposed to extract fault feature information,and HMM is used for fault recognition of rolling bearing,and the effectiveness of the method is verified by experiments.Operating conditions refer to the operating conditions and environmental conditions of the equipment during operation.In the actual operation,different equipment of the same type cannot be in the same working environment,but different working environment and the running state of the machine will have a significant impact on the wear and degradation state of the equipment.According to different signal feature index reflects the ability of different residual lifetime of rolling bearing?In this paper,the vibration signal of the bearing isreconstructed by the fusion of all-vector spectrum method and all-vector EEMD method,and several typical degradation indexes that can reflect the failure degradation situation are analyzed and compared.Then,Several indicators with high correlation are selected and KPCA fusion algorithm is adopted.The value with a large maximum likelihood estimation(generally the first core principal component)is selected as the degradation index of rolling bearing failure.The problem that a single degradation index can not fully reflect the degradation characteristics of bearing life is solved.The residual life model of rolling bearing with different faults at different rotating speeds is established by using the method of bearing degradation index extraction proposed in this paper.Finally,the residual service life of rolling bearing is predicted by the classification method of support vector machine.The accurate prediction of the residual life of the rolling bearing at different speeds is realized.