Location: Centrale Nantes, Fluid Mechanics Laboratory UMR6598, Nantes, FRANCE.
Contact: Michel.Visonneau [at] ec-nantes.fr.
Title: Discretization error estimate and automatic local grid adaptation for the computation of turbulent free-surface flows at high Reynolds number
Description: Many flows are characterized by the presence of localized structures (e.g. wing or propeller tip vortices with cavitation inception in water) or rapid temporal evolution of discontinuous regions (e.g. evolution of a free surface during the impact of a body, wave breaking phenomena, etc...), phenomena which justify the local unsteady adaptation of the grid on which the Reynolds-Averaged Navier-Stokes Equations are solved. This grid adaptation by successive refinement and/or coarsening steps is perfectly compatible with the unstructured face-based finite-volume discretization methods implemented in ISIS-CFD, the free-surface capturing flow solver developed since 1999 by the CFD Group of Fluid Mechanics Laboratory of ECN. Parallelized grid adaptation strategies are currently under development in the CFD group and this thesis will provide the rigorous tools which will be used to control the grid refinement. During this thesis, the Ph.D. student will develop innovative criteria to adapt the grid refinement to the characteristics of the flow field and he / she will propose strategies to evaluate the discretization error on a fully-unstructured single grid. The Ph.D. student will apply these various criteria to the simulation of turbulent free-surface flows at high Re around complex geometries (wave-breaking on a ship running in head waves, simulation of slamming, etc...) and will study the applicability of this approach to adaptive hybrid LES simulation.
Profile: Excellent knowledge in Computational Fluid Dynamics, expertise in algorithmics, FORTRAN 95 programming and MPI parallelization highly appreciated.