Event
Machine Learning for Enhancing and Complementing Computational Fluid Dynamics - Potentials and Limitations for Aerospace Applications| Title of the event | Machine Learning for Enhancing and Complementing Computational Fluid Dynamics - Potentials and Limitations for Aerospace Applications |
| Series | CIADS Colloquium |
| Organizer | Campus-Institut Data Science (CIDAS) |
| Speaker | Dr. Philipp Bekemeyer |
| Speaker institution | German Aerospace Center (DLR) |
| Type of event | Kolloquium |
| Category | Forschung |
| Registration required | Nein |
| Details | Machine learning and artificial intelligence techniques have transformed our everyday life within the past few years. In areas for which vast amounts of data are available the aforementioned techniques had a tremendous success, especially when mathematical models are lacking. Instead, engineering tools in general and computational fluid dynamics tools in particular rely on first-order principals that directly enable to describe and investigate system behavior. However, such tools are far from perfect and suffer several short-comings, e.g. computational bottlenecks once a massive amount of simulations is required or the problem of deriving accurate turbulence models to describe small scale turbulent behavior. Machine learning techniques are generally regarded as a possibility to enhance and complement firstorder based numerical simulation tools to circumvent these shortcomings. Following this ambition, the Center for Computer Applications in AeroSpace Science and Engineering department of the German Aerospace Center has investigated and industrialized scientific machine learning techniques within the past two decades always in close connection to established numerical simulation tools as well as industrial needs [1]. This presentation will provide an insight into previous and current activities within the department covering topics from purely data-driven approaches to the incorporation of physical knowledge into models. Namely, reduced order techniques [2], fusion of data from different sources and with different sparsity [3], data-driven turbulence modeling [4] as well as physics-informed neural networks [5] will be discussed. Highlights will be show-cased and existing limitations outlined. Moreover, open questions that the community has to address to further establish machine learning methods for industrial applications will be posed. REFERENCES [1] Bekemeyer, P. et. al., “Data-Driven Aerodynamic Modeling Using the DLR SMARTy Toolbox,” AIAA AVIATION 2022 FORUM, American Institute of Aeronautics and Astronautics, 2022 [2] Sabater, C., Stürmer, P., and Bekemeyer, P., “Fast Predictions of Aircraft Aerodynamics using Deep Learning Techniques,” AIAA Journal, Volume 60, Number 9, September 2022 [3] Bertram, A., Bekemeyer, P., and Held, M., “Fusing Distributed Aerodynamic Data Using Bayesian Gappy Proper Orthogonal Decomposition,” AIAA AVIATION 2021 Forum, American Institute of Aeronautics and Astronautics, 2021 [4] Jäckel, F., “A Closed-form Correction for the Spalart-Allmaras Turbulence model for Separated Flows,” AIAA SCITECH 2022 Forum, American Institute of Aeronautics and Astronautics, 2022 [5] Wassing, S., Langer, S. and Bekemeyer, P., “Parametric Compressible Flow Predictions Using Physics-Informed Neural Networks,” ECCOMAS Congress 2022, 2022 |
| Date | Start: 11.05.2023, 14:15 Uhr Ende: 11.05.2023 , 15:15 Uhr |
| Location |
Institut für Informatik (Goldschmidtstrasse 7) 2.101 |
| Contact |
0551 39-21289 imatthi@gwdg.de |
| External link | https://www.uni-goettingen.de/en/653203.html |