Study On The Remaining Useful Life Prediction And Optimal Maintenance Policy For A Slowly Degrading System Subject To Discrete Condition Monitoring

The equipment is affected by wear,fatigue and corrosion during the long run process,equipment parts lost initial function,the performance and health condition of the equipment will be degraded even the final failure,this system with slowly failure phenomenon is called slowly degrading system.In the process of slowly degradation,the quality of equipment is reduced and the cost of run and maintenance is raised.The failure of parts will increase the probability of accidents which resulting in economic losses,casualties and other incalculable consequences.Therefore,the degradation data is monitored,and prognostics and health management are implemented during the running process of the system.The degrading parts should take effective maintenance policy timely which will play an important role in ensuring safety,reliability and economy of degrading systems.In the view of this,the objective of thesis is slowly degrading system,and a remaining useful life prediction method and control-limit maintenance policy is proposed.The main research work is as follows:Firstly,the acquisition of degradation data is affected by measurement errors,and the Kalman filtering method is selected to eliminate measurement errors of discrete condition monitoring data.The feasibility of the method which removes measurement errors of slowly degrading system is verified by an example.Secondly,Wiener process degradation model is built and the continuous Wiener process is transformed into discrete form.Then a hybrid degradation model based on Wiener process and Kalman filtering is established to predict remaining useful life.The feasibility of the method is verified by the monitoring data of certain brake pads.It lays the foundation for the following study of the maintenance policy of the slowly degrading system.Thirdly,the degradation process of brake pads is divided into finite degradation states under the framework of SMDP,and preventive control-limit maintenance policy model is established with the object of minimizing long-run expected maintenance cost per unit time.The transition probability is determined by Wiener process,and the policy iteration algorithm is used to solve the model.Finally,to compare with traditional age-based maintenance policy,and the correctness and effectiveness of the proposed maintenance policy are verified.The sensitivity analysis of the relevant parameters is carried out for the preventive control-limit maintenance policy.