|"Data-driven Maintenance Planning under Uncertainty" |
|Konuşmacı ||: || |
İpek Dursun, M.Sc., Eindhoven University of Technology
|Tarih ||: ||16 Aralık 2019 (Pazartesi) |
|Saat ||: ||14:30 - 15:20 |
|Yer ||: ||Bilgisayar ve Bilişim Fakültesi, |
İdris Yamantürk Konferans Salonu (1303)
Maintenance planning is important in order to minimize the total costs and machine downtime. This study focuses on a system with a fixed life span and a critical component. It is assumed that corrective maintenance action has a higher cost than preventive maintenance. In this study, a partially observable Markov decision process (POMDP) model is built in order to determine the optimal preventive maintenance action for the system. The time-to-failure model of the critical component is used in this model. It is assumed that there exists an uncertainty in the time-to-failure model. In order to take into account this uncertainty, Bayesian update is applied based on the new data obtained from the last maintenance action. The goal of this research is to show the effect of learning on total maintenance costs and to propose an optimal maintenance planning model for a multi-cycle system.
İpek Dursun is currently conducting her Ph.D. study at Operations, Planning, Accounting and Control (OPAC) group of Department of Industrial Engineering &Innovation Sciences (IE&IS) at Eindhoven University of Technology (TU/e). She is currently focused on data-driven maintenance planning and decision making under uncertainty for her research. She received her B.Sc. (2016) and M.Sc. (2018) degrees from Boğaziçi University Industrial Engineering department.