Cerfacs Enter the world of high performance ...

🎓Eloïse D'AYER Thesis Defense

  Wednesday 27 May 2026 at 14h30

  Phd Thesis       ONERA (Toulouse), Auditorium (Bâtiment A)    

Improving roughness models using high order direct numerical simulations and immersed boundary conditions

MEGEP (Mécanique, Energétique, Génie civil & Procédés)

https://rdv.onera.fr/soutenance-eloise-dayer

The presence of roughness on aeronautical surfaces can lead to an increase in drag and heat transfer. In order to define RANS models that includes roughness effect on the flow, database of flow over rough surfaces must be first obtained  from direct numerical simulations (DNS).

In this PhD thesis, the JAGUAR solver developed by ONERA and CERFACS is used  with this objective. This solver uses the spectral difference (SD) method to discretize the computational domain. Because of the large element size in the SD method, an Immersed Boundary Method (IBM) must be used to allow the inclusion of rough surfaces in the computations. In this regard, the volume penalization method was chosen. The combination of this IBM and the SD method is tested by simulating the steady flow over a cylinder. Interactions between the IBM and the SD method are analyzed from this first test case. Then, the turbulent flow over an academic rough surface is determined with and without the IBM. Those computations shows that the IBM is capable of reproducing the results obtained by the classical SD method for these academic rough surfaces. Finally, the turbulent flow over a more complex and realistic rough surface is computed. The final results shows that it is possible to build a DNS database of flow over complex rough surfaces with the IBM and the JAGUAR solver.

Jury

Koen HILLEWAERTUniversité de LiègeReviewer
Sébastien TORDEUXUniversité de Pau et des Pays de l'AdourReviewer
Éric GONCALVES DA SILVAENSMAExaminer
Guillaume PUIGTONERAExaminer
Héloïse BEAUGENDREENSEIRB-MATMECAExaminer
François CHEDEVERGNEONERAThesis supervisor
Arthur COLOMBIÉCERFACS, IPSAThesis co-supervisor
Omar DOUNIACERFACSThesis co-supervisor

No content defined in the sidebar.