Proactive Computing in Industrial Maintenance Decision Making [article]

Alexandros Bousdekis, National Technological University Of Athens, National Technological University Of Athens
2018
Proactive event-driven computing refers to the use of event-driven information systems having the ability to eliminate or mitigate the impact of future undesired events, or to exploit future opportunities, on the basis of real-time sensor data and decision making technologies. Maintenance management can benefit from these advancements in order to tackle with the increasing challenges in today's dynamic and complex manufacturing environment in the context of Industry 4.0. To this end, the
more » ... thesis combines and brings together the research fields of Industry 4.0, Maintenance Management and Proactive Computing in order to frame maintenance management and information systems in the context of Industry 4.0. Therefore, it paves the way for the next generation of maintenance management in the frame of Industry 4.0, i.e. Proactive Maintenance. The focus of the current thesis is on proactive decision making. Consequently, it proposes proactive decision methods, capable of handling uncertainty, applicable to maintenance management and its interrelationships with other manufacturing operations, algorithms for continuous improvement of proactive decision making through the proposed Sensor-Enabled Feedback (SEF) approach and algorithms for context-awareness in proactive decision making. To do this, it utilizes methods and techniques for operational research, data analytics and machine learning. The aforementioned algorithms have been embedded in a proactive information system for decision making which was integrated with other tools in order to implement all the steps of the Proactive Maintenance framework. The system has been deployed and evaluated in real industrial environment, while further evaluation was conducted with extensive simulation experiments. Finally, the lessons learned and the managerial implications of the proposed approaches are discussed.
doi:10.26240/heal.ntua.2986 fatcat:aoc56z7qabcbpl5rew3xm4f3nm