Ongoing research projects

IMMC

Ongoing research projects in iMMC (February 2023)


This a short description of research projects which are presently under progress in iMMC.
Hereunder, you may select one research direction or choose to apply another filter:

Biomedical engineering

Computational science

Civil and environmental engineering

Dynamical and electromechanical systems

Energy

Fluid mechanics

Processing and characterisation of materials

Chemical engineering

Solid mechanics


Research direction:
Listed keyword:
Other keyword:
Division:
Supervisor:

List of projects related to: computing




Efficient and scalable frameworks for PDE simulations
Researcher: Thomas Gillis
Supervisor(s): Philippe Chatelain

focuses his research on the development of efficient and scalable computational framework for the simulation of 3D PDEs on massively parallel and heterogeneous architectures.



A phase-field discrete elements model applied to granular material
Researcher: Alexandre Sac-Morane
Supervisor(s): Hadrien Rattez

The main goal of the research project is to combine a phase-field modelization with a discrete elements modelization. This new approach is then applied to granular material to investigate the effects of the environment. A model is built and will be calibrated by experiments.



2-phase CFD simulations of electrolyte-bubble interactions during alkaline water electrolysis
Researcher: Kevin Van Droogenbroek
Supervisor(s): Joris Proost

In today’s world, concern is growing about the future of energy. Despite very ambitious international climate goals by 2050, global energy-related carbon dioxide (CO2) emissions keep increasing. In order to tackle this problem, hydrogen (H2) seems to be the right solution since it is a way to produce, store, move and use energy in a clean way. However, 95% of the actual hydrogen production is made of grey hydrogen, e.g. H2 produced from fossil energies, which leads to high CO2 emissions in the atmosphere. One way to decarbonise this energy vector is to produce green hydrogen by means of renewable energies (solar panels, wind turbines, etc). This is where my research project funded by the Walloon region comes in, focusing on the production of green hydrogen by alkaline water electrolysis (AWE).

In general, AWE is characterised by the use of two planar electrodes separated by a certain distance and operating in a liquid alkaline electrolyte solution (e.g. KOH, potassium hydroxide). However, the efficiency of the process can be improved by the use of 3D electrodes in a zero-gap cell configuration. This configuration is the one that will be used in the scope of this research and it is depicted in Figure 1. The chemical reactions taking place at the cathode and at the anode are also highlighted.

More specifically, the work will consist in the fluid mechanical modeling of liquid and gaseous flows within alkaline electrolysis cells filled with 3D porous structures. The study of liquid electrolyte flow and of gaseous hydrogen bubble formation and escape will allow to optimise the performance of the electrolyser. Computational Fluid Dynamics (CFD) is a powerful numerical tool that will be used during this project to determine the optimal flow parameters required to homogenise the electrolyte flow (to take advantage of the full specific area provided by the electrodes) while favouring hydrogen bubbles removal from the electrolysis cell (to avoid bubble entrapment within the complex 3D structure).

As an example, the added value of a numerical simulation for a better understanding of the electrolyte flux distribution within an empty cell (e.g. without 3D structure) is shown in Figure 2. The velocity field of the electrolyte (in m/s) was simulated on the OpenFOAM software. Note that the geometry of the cell corresponds to the one of the pilot electrolyser used at UCLouvain (see Figure 3).



SLIM-G : GPU-accelerated SLIM
Researcher: Miguel De Le Court
Supervisor(s): Vincent Legat

I am developing a GPU version of the SLIM ocean model in order to significantly increase its performance, which unlocks previously unreachable resolutions.

Most current ocean models still use the CPU for all the computations, which makes them comparatively slow, and unable to use the next generation of supercomputers such as LUMI. Among the ocean models that are accelerated on the GPU, all of them either use finite differences, which lacks flexibility in the meshes, or finite volumes, which are often low order methods. Contrary to that, SLIM uses the discontinuous Galerkin Finite elements method, which is known for its low diffusivity in advective processes and maps very well to the massively parallel architecture of the GPU.

The current GPU version is still incomplete, but shows speedups of a factor ranging from 50x to 120x faster than a CPU. (R7 2700X vs RTX2080).