Beiträge von tmaric

Bumping this thread, we're still searching for a candidate. In case the Mathematical Modeling and Analysis Institute is scaring you off, this is a Computational Engineering topic to be organized at CE!

I am starting a master thesis project on Machine Learning in Computational Fluid dynamics in cooperation with Intel and BOSCH Corporate Research. We will work on coupling OpenFOAM with Physics-Informed Neural Networks. For more information and a small hands-on example, check out my Youtube talk at the OpenFOAM workshop. The minimal working example is on GitHub, this will be our starting point. We will collaborate with Intel's Advanced AI laboratory to see if we can train Physics-Informed Neural Networks in OpenFOAM efficiently, then use the models in microfluidics simulations in cooperation with BOSCH Corporate Research.

I'm looking for a Computational Engineer/Scientist with a strong affinity for research, interested in learning the details of the unstructured Finite Volume Method, OpenFOAM, and Physics-Informed Machine Learning. You will be working closely with me on this topic, and collaborating with industrial partners and PhD candidates in my research group. Email your grade transcripts and CV to maric@mma.tu-darmstadt.de

In (Computational) Fluid Dynamics lectures, we assume fluid gets trapped

within small cavities that define the surface roughness and use this

assumption to claim that the fluid's velocity at the surface equals the

velocity of the surface, i.e. zero-velocity boundary condition for fixed

surfaces. For large-scale CFD simulations, numerical methods cannot

resolve the surface roughness, so a zero-velocity boundary condition is

assumed. Surprisingly, the Darmstadt Tribometer experiment at TU

Darmstadt (FST Institute) indicates that the surface roughness introduces a

(small) partial slip at the surface, which can be modelled using the Navier-

Slip boundary condition, with the slip-length related to the characteristics of

surface roughness. This effect possibly plays a significant role in wetting

processes investigated in the Collaborative Research Center 1194 at TU

Darmstadt.

More details and contact information are in the attached PDF.