Research On Prediction Method Of Rolling Bearing RUL On Multi-change Point

Rolling bearings are one of the indispensable components in large-scale equipment.They are widely used in national defense and national economy,and have contributed to the modernization of the motherland.In order to ensure the smooth operation of large-scale equipment,it is necessary to infer the remaining use time of rolling bearings,which can not only ensure the safety of staff,but also greatly reduce costs.Therefore,it is urgent to carry out research and engineering practice on PHM of rolling bearings.Among them,RUL prediction has always been the key to PHM research.Based on the current remaining life prediction research,this paper takes rolling bearings as the object and uses rolling bearing data sets to study the degradation model in the presence of change points,and finally obtain a more accurate remaining life.For clarity,the main achievements are listed as follows:(1)Firstly,aiming at degradation process without change points,a single change point residual life prediction method based on Wiener process is presented.According to the failure time of the first arrival time,the parameter distribution of the RUL is obtained.According to the failure time of the first arrival time,the parameter distribution of the RUL is obtained,and then a parameter estimation method based on the EM algorithm is brought forward.(2)Secondly,considering that in the actual degradation process of the equipment,the number of change points is not a single one,but a state where multiple change points coexist.Aiming at the problems of multiple change points,based on the single change point model,this paper proposes Remaining life prediction method with variable points.According to the problem to be solved,a reasonable multi-change point degradation model is established,the estimated value of the unknown parameter is solved,and the prediction method of the single-change point degradation model is compared.(3)Finally,on the basis of the residual prediction model of multi-change points,considering the differences between individual equipment,a multi-change point prediction model under the condition of related model parameters is proposed.The model parameters obey the Gauss-inverse gamma distribution,and the parameter estimation method for the multi-parameter coupling in the model is proposed.After obtaining the estimated values of the model parameters,the Bayes formula is used to obtain the posterior distribution of the model parameters,and finally the updated remaining life is obtained.