[finzeo]

Hi all,


Currently I am doing simulations on heavy vehicle aerodynamics. When generating the mesh, very relevant and problematic topics are: (1) generating regular layers around the surface of the vehicle, and (2) correctly adapting the mesh in areas with small details, keeping a structured mesh away from the vehicle (for better results). I always used OpenFOAM (blockMesh + snappyHexMesh) to generate the mesh, because I have a lot of knowledge in the tools and their capabilities, but I see some inconsistencies when converting the mesh from OF to .med (needed for running in Code_Saturne); I use the unofficial foamMeshToMED utility (https://github.com/mortbauer/foamMeshToMED).
What mesh tool do you recommend for vehicle aerodynamics? Some colleagues recommended me to use GridPro and Star-CCM for these cases, but before managing a license through my institute and taking the trouble to learn a new software, I want to consult those who work in the same subject area.

[LuckyTran]

Star-CCM and snappy both use the same subsurface approach for wall layers so you will not be able to generate regular layers in Star any better or worse than foam.

Star's trimmed cell mesher is almost identical to snappy. The benefit of Star is for polyhedral meshes, nice unstructured ones! If you are a member of a meshing cult that believes meshes should still be structured in 2023 then I can assure you, Star-CCM is blasphemous and you should avoid it because it will never make you happy.

GridPro relies on you understanding where the location of topological defects in the mesh need to occur. If you already having trouble with snappy then I can also assure you that GridPro will be even more difficult to manage.

[finzeo]

Originally Posted by LuckyTran

Star-CCM and snappy both use the same subsurface approach for wall layers so you will not be able to generate regular layers in Star any better or worse than foam.

Star's trimmed cell mesher is almost identical to snappy. The benefit of Star is for polyhedral meshes, nice unstructured ones! If you are a member of a meshing cult that believes meshes should still be structured in 2023 then I can assure you, Star-CCM is blasphemous and you should avoid it because it will never make you happy.

GridPro relies on you understanding where the location of topological defects in the mesh need to occur. If you already having trouble with snappy then I can also assure you that GridPro will be even more difficult to manage.

Thank you for the information. Luckily I am not dogmatic in that sense, but pragmatic, although the choices of my work colleagues influence me. Perhaps the use of polyhedral meshes is a new and good option to consider. The purpose of my question was also, perhaps, to hear about new meshing tools (different from the ones I mentioned) that have emerged recently and are very convenient for this particular use (vehicle aerodynamics).

[JBeilke]

Originally Posted by LuckyTran

Star-CCM and snappy both use the same subsurface approach for wall layers so you will not be able to generate regular layers in Star any better or worse than foam.

CCM+ has 2 different methods for the sublayer. The better one is the advanving layer method. This gives much better results for difficult cases.

@finzeo: GridPro is phantastic when your case fits in. Check their blog to get an idea about geometries that work fine.

[JBeilke]

Originally Posted by finzeo

The purpose of my question was also, perhaps, to hear about new meshing tools (different from the ones I mentioned) that have emerged recently and are very convenient for this particular use (vehicle aerodynamics).

I nearly forgot about "Harpoon". It is a very nice mesher that I tried years ago with good success.

https://www.sharc.co.uk/index.htm

[finzeo]

Originally Posted by JBeilke

CCM+ has 2 different methods for the sublayer. The better one is the advanving layer method. This gives much better results for difficult cases.

@finzeo: GridPro is phantastic when your case fits in. Check their blog to get an idea about geometries that work fine.

Thank you. Right now I am testing, precisely, the possibility of using the advancing layer method.

The option to try Harpoon limits me because of the need to achieve compatibility with Code_Saturne.

[JBeilke]

When you can read a ccm+-mesh into Code-Saturn, then it should be easy to import something from Harpoon too.