Skip to Main content Skip to Navigation

Simulation d'écoulements fluides à partir de données réelles

Abstract : Point clouds are mathématical objects that allows to describe discretely multivariable functions. They are mainly used in the statistical domain but also in geometry to represent complexes geometrical manifolds. It is nowadays a real challenge to immerse those previous manifolds in finite element computation. Indeed the immersion of those point clouds requires the reconstruction of the manifold's surface and the generation of a surfacic mesh that has to be immerse in a volumic mesh. As those operations are often based on iterative process, they are extremely time-consuming since points clouds are usually massive. The method developped in this thesis allows to immerse points clouds in a meshed domain without the surface reconstruction and mesh generations steps. For that purpose, we use the Volume Immersion Method adapted to point clouds. We coupled this method with an adapted mesh generation technique. Then we are able to generate a monolithic anisotropic mesh, adapted around interest ares. We also use the variational multi-scale method to simulate fluid flow. The last part of this thesis introduce some applications cases in the aerodynamic, hydrodynamic and urbans domains.
Complete list of metadatas

Cited literature [102 references]  Display  Hide  Download
Contributor : Simon Santoso <>
Submitted on : Friday, November 29, 2019 - 10:46:22 AM
Last modification on : Monday, January 13, 2020 - 1:17:28 AM


Files produced by the author(s)


  • HAL Id : tel-02386089, version 1



Simon Santoso. Simulation d'écoulements fluides à partir de données réelles. Modélisation et simulation. Ecole Centrale Nantes; Institut de Calcul Intensif, 2018. Français. ⟨tel-02386089⟩



Record views


Files downloads