Research On Remaining Useful Life Prediction Of Rolling Bearing Based On PSO Optimized SVM

Rolling bearings are one of the most common and important basic components in rotating machinery systems,and they are parts that come into contact with another component to do relative motion and bear loads,and their functions determine that rolling bearings are often in a harsh working environment.Under the influence of lubrication,temperature and mechanical vibration,rolling bearings are particularly prone to failure,and once a rolling bearing fails,it may trigger a chain reaction in the current trend of closer links between the components of the mechanical system,resulting in the paralysis of the mechanical system,resulting in unpredictable safety incidents and significant economic losses.The residual service life of rolling bearing is a comprehensive reflection of its damage degree,if the remaining service life of rolling bearing can be accurately predicted,then can detect and troubleshoot the fault in time,change the regular maintenance to the maintenance as appropriate,can effectively reduce the possibility of accidents.Therefore,accurate prediction of the residual life of rolling bearings is of great significance for the realization of visual maintenance.Taking rolling bearings as the research object,this paper establishes a prediction model for the remaining service life of rolling bearings based on improved particle swarm algorithm optimization SVM(Support Vector Machine,SVM),which is verified by experimental data.(1)The feature Index extraction and feature index reduction method based on the vibration signal of rolling bearing are studied.Aiming at the shortcoming that a single index can not fully describe the degradation trend of rolling bearing performance,a total of 23 indexes of common time domain and frequency domain are extracted to describe the degradation trend of bearing performance.Because of the phenomenon of data redundancy between 23 feature indexes and the high feature dimension,the generalization ability of predictive model will become worse,so the data drop-down method based on principal component analysis is introduced to reduce the redundancy and dimension of data,and the processed data is fed into the remaining service life prediction model of rolling bearing as a new characteristic index.To predict the remaining service life of rolling bearings.(2)The basic principles of support vector machine theory and particle swarm optimization algorithm are introduced.Because the selection of parameters in support vector machine has a great influence on the generalization ability of the model,and there is no fixed method for parameter selection,which often depends on the personal experience selection,which leads to the performance of the model,so this paper proposes a method to optimize the support vector machine parameters based on particle swarm algorithm.In view of the fact that the standard particle swarm algorithm often falls into the local optimal solution in the process of finding the optimal solution,an improved particle swarm algorithm is proposed in this paper and applied to the prediction model of the residual service life of rolling bearings.(3)The model established in this paper is verified by the full life experimental data of rolling bearings of the University of Cincinnati in the United States,and the results show that the prediction model of residual service life of rolling bearings based on particle swarm algorithm optimization support vector can accurately predict the remaining service life of rolling bearings.Comparing with the improved particle swarm algorithm to optimize the residual service life Prediction model of the support vector machine,the results show that the generalization ability of the remaining service life prediction model of rolling bearing is better and the prediction result is more accurate after the improvement of the particle swarm algorithm.The research work in this paper can provide some reference for the study of residual service life prediction of rolling bearings,and has certain theoretical value.