Three-dimensional real-life simulations are generally unsteady and involve moving geometries. Industries are currently still very far from performing such simulations on a daily basis, mainly due to their extensive computational cost. Anisotropic metric-based mesh adaptation, which has already proved its usefulness for steady and fixed-mesh unsteady CFD simulations, can certainly solve part of this issue by enhancing both the accuracy and efficiency of CFD simulations. However, before extending these techniques to moving mesh simulations, the fixed-topology constraint imposed by the classical Arbitrary-Lagrangian-Eulerian formulation (ALE) framework has to be released. This paper brings two new ideas. First, it demonstrates numerically that moving three-dimensional complex geometries with large displacements is feasible using only vertex displacements and connectivity changes. This is new and presents several advantages over usual techniques for which the number of vertices varies in time. Notably, it facilitates the handling of data structures on the solver part, thus favorably impacting CPU time. But it also answers the scarcely addressed question of spoiling interpolation errors in moving mesh simulations. The second novelty lies in the description of a new methodology extending well-known ALE schemes to changing-connectivity meshes. Even if for the moment this scheme has only be implemented in two-dimensions, our objective is clearly to use a similar methodology for three-dimensional connectivity changes. Finally, note that the complete description of the extension of multi-scale anisotropic metric-based mesh adaptation unsteady, and especially moving mesh simulations, is given in.
展开▼
