Performance Degradation Evaluation And Remaining Life Prediction Of Rolling Bearings Based On Improved PSO And KELM

As one of the important parts of mechanical equipment,rolling bearing’s health and running state are very important to the reliability of mechanical equipment.Once the failure occurs,it will lead to unexpected consequences.Therefore,the research on the performance degradation evaluation and remaining life prediction of rolling bearing can provide feasible prevention for maintenance decision,and improve the safety and reliability of mechanical equipment,which has important application value and practicability.In this paper,rolling bearing is taken as the research object,and the vibration signal feature extraction,performance degradation evaluation and remaining life prediction are studied.Firstly,the feature extraction in time domain and frequency domain is carried out for the rolling bearing vibration signal,and the extracted features are analyzed by principal component analysis(PCA)and phase space reconstruction(PSR),combined with kernel extreme learning machine(KELM),a new performance degradation evaluation method,namely PAPRKM method,is proposed to realize the rolling bearing performance degradation evaluation.Because the parameters of KELM model are empirical parameters,which cannot achieve the optimal effect.In order to solve this problem,particle swarm optimization(PSO)algorithm is used,which is improved by adjusting inertia weight,introducing linear learning factor and adding disturbance term.The improved particle swarm optimization algorithm is used to optimize KELM model parameters.A new model of performance degradation evaluation,namely WCDPSO-KELM model,is constructed.Combined with PAPRKM method,the performance degradation evaluation is realized.In the case of limited samples,the traditional remaining life prediction method cannot effectively predict the remaining life of rolling bearing.In order to solve this problem,based on WCDPSO algorithm and KELM model,the remaining life prediction model is established.Based on PAPRKM method and WCDPSO-KELM model,the remaining life prediction of rolling bearing is realized.The performance degradation evaluation and remaining life prediction methods proposed in this paper are verified by the life cycle data of rolling bearings.The experimental results show that PAPRKM model has the best prediction performance and can effectively reflect the performance degradation process of rolling bearing;WCDPSO algorithm has strong optimization ability and can effectively improve the prediction accuracy of KELM model;The WCDPSO-KELM model can effectively improve the accuracy and stability of rolling bearings remaining life prediction and has practical application value.