Presented By: Industrial & Operations Engineering
Digital Twin Calibration in the Era of Big Data
Cheoljoon Jeong
About the speaker: Cheoljoon Jeong is a Ph.D. candidate in the Department of Industrial and Operations Engineering at the University of Michigan, Ann Arbor. His research interests encompass industrial data science, quality and reliability engineering, system informatics, and the design and analysis of computer experiments. His current research focuses on developing digital twin calibration methodologies for several applications such as building energy and wind power systems. He has also been advancing industrial data science methodologies to enhance automation and improve process control in smart manufacturing. He has received multiple best paper recognitions, including the Richard C. Wilson Best Paper Prize from the University of Michigan, the IISE QCRE Best Student Paper Award, the IISE DAIS Best Paper Award Finalist, and the INFORMS QSR Best Paper Award Finalist. Throughout his academic journey, he has been awarded prestigious fellowships, including the Edward P. Fitts, Seth Bonder, and IES Graduate Fellowships.
Abstract: A digital twin is a digital representation of a physical system that is useful for monitoring, diagnostics, and prognostics in a virtual world. Due to advances in sensing technology, a massive amount of observational data is collected from various aspects of the system. Calibrating digital twins with the observational data is crucial to enabling them to accurately replicate the physical system. One way to calibrate digital twins is through parameter calibration. This study develops quantitative calibration methods that harness the full potential of big data, addressing research challenges posed by its scale and complexity. These include: (a) developing an approach that integrates nonlinear optimization with statistical theory to guide the sequential design of computer experiments, (b) devising a stochastic dimension reduction method for high-dimensional simulation parameters with explainability, and (c) extending the framework to accommodate a wide range of problems, including functional calibration. The approaches are validated through real-world case studies.
Abstract: A digital twin is a digital representation of a physical system that is useful for monitoring, diagnostics, and prognostics in a virtual world. Due to advances in sensing technology, a massive amount of observational data is collected from various aspects of the system. Calibrating digital twins with the observational data is crucial to enabling them to accurately replicate the physical system. One way to calibrate digital twins is through parameter calibration. This study develops quantitative calibration methods that harness the full potential of big data, addressing research challenges posed by its scale and complexity. These include: (a) developing an approach that integrates nonlinear optimization with statistical theory to guide the sequential design of computer experiments, (b) devising a stochastic dimension reduction method for high-dimensional simulation parameters with explainability, and (c) extending the framework to accommodate a wide range of problems, including functional calibration. The approaches are validated through real-world case studies.
Explore Similar Events
-
Loading Similar Events...