00:00 Introduction
06:46 Benefits of a digital twin
09:46 Attributes of mechanistic models, machine learning, and digital twins
16:36 Scientific, technological & economic problems solved by digital twins
20:31 Components of a digital twin in additive manufacturing
45:39 Current uses of digital twins of additive manufacturing in the industry
47:06 Outlook
This talk was presented in the Department of Mechanical Engineering, University of Michigan. Speaker: Professor Tarasankar DebRoy, Host: Professor Wenda Tan.
Digital twins of additive manufacturing are virtual replicas of the additive manufacturing processes and are gaining increased interest in industry because of the scientific, technological, and commercial benefits they provide. This presentation explains the building blocks of a digital twin of additive manufacturing and explains the roles of mechanistic models, machine learning, statistical models, sensing and control, and data. The digital twins can avoid the current practice of qualifying metallic parts by trial and error with expensive printing equipment and feedstock materials. Detection and prevention of defects and diagnostic closed-loop process control using sensors that monitor predetermined features such as deposit geometry or peak temperature are discussed. Unlike closed-loop control, offline uses of digital twins for process planning can narrow the optimum process parameter window to obtain a predetermined outcome, thus reducing the likelihood of having to undertake extensive corrective actions during manufacturing. Finally, the outlook for its expanded use is discussed based on evidence.
#DebRoyResearchGroup, #digitaltwins #modeling, #machinelearning, #3Dprinting, #universityofmichigan, #seminar, #additivemanufacturing
