LEGI - UMR 5519

The aim of this work is to develop the creation and usage of anisotropic meshes for turbulence simulations using RANS, LES and DNS. By taking into account the flow’s preferential directions, anisotropic meshes allow for reduced mesh sizes, hence allowing to decrease the computational cost of these simulations.

Isotropic and anisotropic meshes in the boundary layer of a NACA airfoil. The ability to adapt the mesh allows to concentrate the cells and computational effort in the boundary layer region while the usage of anisotropic cells allows to reduce the overall computational cost of these simulations while ensuring the same degree of accuracy.

Throughout this project, anisotropic remeshing criteria have been developed and applied to turbulent configurations [4]. These criteria leverage the automatic mesh convergence procedure [1,2,3] that allows for an accurate discretization of complex turbulent flows without any prior knowledge of the flow. The extension of this procedure to anisotropic meshes that we have developed enables a dramatic reduction of computational cost while preserving accuracy.

Left : Q-criterion isovalues colored by the axial vorticity in a jet in co-flow configuration [4]. Right : anisotropic mesh used in this simulation.

These methods have been further extended to multiphase flow configurations where they are particularly interesting, as interfaces inherently showcase preferential directions.

Left : interface and the adapted mesh. Particular attention is paid to the region of high interface curvature, hence the mesh goes back to isotropic in these regions, while region of low curvature exhibits much more anisotropy. Right : application of this strategy to a jet-in-crossflow configuration similar to [5].

[1] Benard, P., Balarac, G., Moureau, V., Dobrzynski, C., Lartigue, G., et D’Angelo, Y., “Mesh adaptation for large-eddy simulations in complex geometries,” International Journal for Numerical Methods in Fluids, vol. 81, no. 12, pages 719–740, 2016. [Online]. Available : https://onlinelibrary.wiley.com/doi/abs/10.1002/fld.4204

[2] Grenouilloux, A., Leparoux, J., Moureau, V., Balarac, G., Berthelon, T., Mercier, R., Bernard, M., Bénard, P., Lartigue, G., et Métais, O., “Toward the use of les for industrial complex geometries. part i : automatic mesh definition,” Journal of Turbulence, vol. 0, no. 0, pages 1–31, 2023. [Online]. Available : https://doi.org/10.1080/14685248.2023.2214399

[3] Lam, H., Berthelon, T., et Balarac, G., “Non-dimensional meshing criterion of mean flow field discretization for rans and les,” Computers Fluids, vol. 291, p. 106572, 2025. [Online]. Available : https://www.sciencedirect.com/science/article/pii/S0045793025000325

[4] Barbera, R., Berthelon, T., Letournel, R., Ghigliotti, G., et Balarac, G., Anisotropic Mesh Adaptation for Large Eddy Simulations. [Online]. Available : https://arc.aiaa.org/doi/abs/10.2514/ 6.2025-1378

[5] Leparoux, J., Mercier, R., Moureau, V., et Musaefendic, H., “Primary atomization simulation applied to a jet in crossflow aeronautical injector with dynamic mesh adaptation,” in 14th Triennal International Conference on Liquid Atomization and Spray Systems, vol. 137, 2018.