Veranstaltung
TUCaN-Modulnummer: | TBD |
Veranstaltungsturnus: | Jedes Wintersemester |
Dozent: | Prof. Dr.-Ing. Dominik Schillinger |
Sprache: | Englisch |
TUCaN-Kursnummer der Vorlesung und Übung: | TBD |
Inhalt
Da die Vorlesung auf Englisch ist, ist das Inhaltsverzeichnis nur auf dieser Sprache vorhanden.
- Part I: Fundamentals and mathematical background
- Motivation of reduced order modeling (many-query, real-time, high-dimensional scenarios)
- Traditional engineering approaches: static condensation, modal decomposition
- Foundations of parametrized partial differential equations
- Proper orthogonal decomposition, snapshots, offline/online strategies
- Reduced basis methods, Galerkin projection and orthonormalization, sampling strategies
- Part II: Model order reduction in computational solid mechanics
- Computational homogenization of heterogeneous materials
- Generalized multiscale finite element methods
- Part III: Advanced topics
- Stability, system conditioning, empirical interpolation methods
Qualifikationsziele
Da die Vorlesung auf Englisch ist, sind die Qualifikationsziele nur auf dieser Sprache vorhanden.
Model order reduction techniques decrease the complexity of mathematical models in numerical simulations. They play a key role in dealing with parametrized systems that require fast and frequent model evaluation. This course provides an introduction to model order reduction with a focus on applications in computational solid mechanics. After successful completion of the first part, students know the foundations of parametrized partial differential equations and understand the challenges associated with their finite element approximation. They know the mathematical basis of different reduced order methods, including their specific advantages, and are able to decide in what scenario which method should be applied. After completion of the second part, students are able to bridge the gap between multiscale models in solid mechanics, discretization based on the finite element method, and model order reduction. They are able to implement different model order reduction techniques for linear problems and are able to critically assess their results in terms of accuracy and efficiency. After completing the third part, students understand limitations of model order reduction techniques and know about open questions and challenges related to current research.
Voraussetzungen
Gute Kenntnis der Finite-Elemente-Methode und der Kontinuumsmechanik sind empfehlenswert.
Hinweise
Prüfung
Die Prüfung findet in mündlicher Form statt.
Zeit und Ort der individuellen Prüfungen werden zu gegebener Zeit in bekannt gegeben.
Bringen Sie zur Prüfung bitte einen aktuellen Studierenden- und einen weiteren gültigen Lichtbildausweis mit.