Department of Mechanics: Seminar: Abstract Lars Beex 2022: Difference between revisions
No edit summary |
No edit summary |
||
| Line 7: | Line 7: | ||
'''Abstract''': Projection-based model-order-reduction accelerates computations of physical systems in case the same computation must be performed many times for different load parameters (e.g. mechanical parameters, geometries, initial conditions, boundary conditions). It therefore finds its use in application domains such as inverse modelling, optimization, uncertainty quantification and computational homogenization. This presentation focuses on a number of extensions of reduced-order-models and their derivatives - some successful, others less so. The main focus is on elastoplasticity, since this is typically not straightforward to treat in reduced-order-models. Machine learning in the form of recurrent neural networks, k-means clustering, DBSCAN and the k-nearest-neighbour search is considered for a part of the extensions. | '''Abstract''': Projection-based model-order-reduction accelerates computations of physical systems in case the same computation must be performed many times for different load parameters (e.g. mechanical parameters, geometries, initial conditions, boundary conditions). It therefore finds its use in application domains such as inverse modelling, optimization, uncertainty quantification and computational homogenization. This presentation focuses on a number of extensions of reduced-order-models and their derivatives - some successful, others less so. The main focus is on elastoplasticity, since this is typically not straightforward to treat in reduced-order-models. Machine learning in the form of recurrent neural networks, k-means clustering, DBSCAN and the k-nearest-neighbour search is considered for a part of the extensions. | ||
'''Links''': [https://campuscvut-my.sharepoint.com/:v:/g/personal/doskamar_cvut_cz/ | '''Links''': [https://campuscvut-my.sharepoint.com/:v:/g/personal/doskamar_cvut_cz/Eb4b6NjtZatLuWlReIjCPw8BR82iHf-wzDqAZLNYRnbg7g?e=2vh32g Recorded lecture] (accessible after CTU login) | ||
Latest revision as of 11:58, 15 April 2022
Machine learning and model-order-reduction for elastoplasticity
Lars Beex, Department of Engineering, Faculty of Science, Technology and Medicine, University of Luxembourg
13th April 2022, 16:00-17:00 CEST, Room B-366 @ Thákurova 7, 166 29 Prague 6
Abstract: Projection-based model-order-reduction accelerates computations of physical systems in case the same computation must be performed many times for different load parameters (e.g. mechanical parameters, geometries, initial conditions, boundary conditions). It therefore finds its use in application domains such as inverse modelling, optimization, uncertainty quantification and computational homogenization. This presentation focuses on a number of extensions of reduced-order-models and their derivatives - some successful, others less so. The main focus is on elastoplasticity, since this is typically not straightforward to treat in reduced-order-models. Machine learning in the form of recurrent neural networks, k-means clustering, DBSCAN and the k-nearest-neighbour search is considered for a part of the extensions.
Links: Recorded lecture (accessible after CTU login)