In this webinar some advanced and highly up-to-date topics in the area of Computational Mechanics will be presented by international leading scientists in the respective fields. The spectrum of topics will cover solid mechanics, fluid mechanics, coupled problems, model reduction, inverse analysis / uncertainty quantification and the combination of machine learning and physics-based modeling. All lecturer will try to start from a basic introductory level for a generally interested audience but will also cover aspects of current scientific interest.
SCHEDULE
Monday, December 7
08.40 - 09.00 Welcome
09.00 - 10.30 Session I - Martin Kronbichler
11.00 - 12.30 Session I - Martin Kronbichler
14.00 - 15.30 Session II - Gianluigi Rozza
16.00 - 17.30 Session III - Alexander Popp
Wednesday, December 9
09.00 - 10.30 Session III - Alexander Popp
11.00 - 12.30 Session IV - Marek Behr & Wolfgang A. Wall
14.00 - 15.30 Session V - Paris Perdikaris
16.00 - 17.30 Session V - Paris Perdikaris
Thursday, December 10
09.00 - 10.30 Session IV - Wolfgang A. Wall
11.00 - 12.30 Session VI - Phaedon-Stelios Koutsourelakis
14.00 - 15.30 Session VI - Phaedon-Stelios Koutsourelakis
16.00 - 17.30 Session II - Gianluigi Rozza
Session IV - Lectures on "Deforming domain flow simulations and fluid-structure interaction" (with Wolfgang A. Wall)
Topics covered: Brief overview on approaches for handling flows on deforming domains; space-time method.
Session VI - Lectures on "Bayesian inverse problems and Multi-fidelity Uncertainty Quantification”
Topics covered: Brief overview of inverse problems and Bayesian formulations; Sampling- and non-sampling-based solution techniques; Black-box and grey-box approaches; Brief overview of uncertainty propagation and associated challenges; Single vs multi-fidelity techniques.
Session I - Lectures on "High-performance fluid dynamics simulations on the exascale threshold"
Topics covered: Incompressible and compressible Navier- Stokes equations and brief overview of turbulence; high- order discontinuous Galerkin methods; iterative solvers and multigrid methods; achieving high node-level performance via efficient matrix-free implementations; utilizing large-scale computers by domain decomposition; SIMD vectorization.
Session VI - Lectures on "Making neural networks physics-informed"
Topics covered: Brief overview of physics-informed neural networks; current capabilities, advantages and limitations; implementation aspects and common caveats; applications in design optimization, heat transfer, wave propagation, cardiovascular fluid mechanics, and modeling of COVID-19 spread dynamics.
Session III - Lectures on "Computational contact and interface mechanics"
Topics covered: Challenges of contact problems; nonlinear contact kinematics; constraint enforcement techniques; finite element discretization schemes; introduction to mortar methods; contact and high-performance computing (HPC); interface mechanics beyond classical contact and friction.
Session II - Lectures on "Reduced order methods for CFD"
Topics covered: Reduced basis method, reduced order modelling, parametric PDEs, proper orthogonal decomposition, stability, CF.
Session IV - Lectures on "Deforming domain flow simulations and fluid-structure interaction" (with Marek Behr)
Topics covered: CutFEM based fixed-grid approach; introduction to fluid- structure interaction (FSI); coupling schemes and solver for FSI; fluid-structure contact interaction.
The webinar will be hosted on the Microsoft Teams platform.
The registration fee is 100.00 Euro + VAT*, where applicable (bank charges are not included).
CISM will financially support students participating in the Advanced Webinar by providing free registration to a number of PhD student and early stage researchers.
A message of confirmation will be sent to accepted partici- pants.
The application deadline is December 3, 2020.
* Italian VAT is 22%.
