HPC simulations for PEM fuel cell H/F (Matériaux, physique du solide)

Job description

Domaine : Matériaux, physique du solide

Contrat : Post-doctorat

Description du poste :

The optimization design of proton exchange membrane fuel cells (PEMFC) and Proton Exchange Membrane Water Electrolyzer (PEMWE) to achieve the objectives of performance and durability remains the bottleneck for moving the technology to the market place. The numerical simulations require specific software to solve the multi-physics and multi-domains PEMFC and PEMWE system. HPC (High Performance Computing) codes must be used.
CEA-Liten, in parallel to numerical simulations with commercial software[1], starts to develop reference simulation at the cell level, using the application TRUST-FC, built on the CEA open source C++ TRUST platform[2]. TRUST can be viewed as a kernel which provides, among other things,time schemes, discretizations, access to PETSc solvers… Its structure supports massively parallel computations with a distributed memory model (MPI). TRUST-FC advantages are the robust numerical methods and the massive parallelism that allows to simulate coupled multi-physics phenomena on large scale domains. TRUST-FC can currently simulate a real CAD design containing tens millions of elements and numerous state variables in a few hours on cluster. SALOME[3] is used as the meshing and visualization tool.
To go further, a new collaboration between CEA and FZJ (Forschungzentrum Jülich) in Germany is built. The team in FZJ-IEK14 is a leader in modelling of fuel cells, electrolysers, hydrogen pumps, and other electrochemical devices on multiple scales from sub micro-scale through to industrial stacks and systems. They employ open source software suites, such as OpenFOAM[4] and cfMesh.
The objectives of this postdoc is to improve numerical simulations code of both PEMFC and PEMWE, using open-source code HPC oriented. FZJ is using OpenFOAM and CEA is using TRUST_FC. The main objective will be first to benchmark both the CEA TRUST code with the FZ-J OpenFOAM based-codes at the cell level design (meshing, domain decomposition, coupling between domains, scalability).
Second objective will be to start to develop state-of-the-art two-phase flow codes based on the most accurate approach (to be determined). Both codes can be used in a specific scale (front tracking approach at the microscale for TRUST-FC and OpenFOAM, 2-phase Eulerian-Eulerian approach at cell scale for OpenFOAM).
The possibility of module interchangability will be also investigated

The postdoc will be located at CEA Grenoble France, with several missions to FZJ, Germany.



You have a PhD and :

• Strong knowledge on numerical simulations, HPC oriented is an add value
• Strong formation on applied mathematics and physics
• Knowledge on multi-physics modelling
• Strong capacity to develop new numerical models
• Strong capacity to work in a collaborative environment, between two world class renewable energies institutes
• Capacity for autonomy and innovation

Ville : Grenoble