[sfluck]

Hi there

I am trying to make a mesh with snappyHexMesh for a multi element airfoil chord 37 cm). I have the airfoil as an stl with three named regions (slat, main, flap).

I struggle mainly with resolving the trailing edges nicely with at least 1 adjacent cell, my inflation layers keep collapsing at the trailing edge. The trailing edges are cut to a width of about 0.5 mm. All feature edges are extracted with surfaceFeatures (OF V7), the generated obj of the feature edges shows that clearly. Attached are some images of the problem areas. The trailing edge cuts of the stl geometry are about perpendicular to the form, not as crooked as they appear in the mesh.
SnappyHexMeshDict can be found here: www.pastebin.com/63CMVxqq

I managed to do a case with pointy trailing edges (uncut) with a mere 4 and relatively thick inflation layers that wrap nicely around the trailing edge, but y+ values are too high like that. Making the layers thinner makes them collapse again, also increasing the refinement leads to a collapsing boundary layer. I played around with tolerances, feature snapping parameters, featureAngle, nothing seems to help, I started to get a bit annoyed. Especially because I don't really know what some parameters mean and don't find the documentation to be really helpful.

What are the parameters that steer the "wrapping-around-corner"-behavior in the sHMDict? Where would you start tuning? On my 8-core pc generating a new mesh takes (depending on n<...>Iterations) about 5-10minutes for castellated, snapping and addLayer-Phases. I've been playing around for about a week now trying different settings, good thing I'm a student and "free" workforce :-).

Also, my base mesh is generated with blockMesh, and it's quite coarse. I figured the refinementRegions in the sHM are enough to break the base mesh down a bit for surface snapping. The snapping step seems to work neatly though, except for the trailing edge snapping. Shouldn't
surfaceFeatures do the trick there?

Any help is appreciated! It's probably not that hard of a problem, I just seem to not see the solution anymore.

Kind wishes
Stefan

[linnemann]

This will be very difficult to obtain.

I would advice you to test out cfMesh.
It does a lot better in taking proper care of layers like this.

[sfluck]

Thanks, I'll look into it.

[mAlletto]

Hei hei. Did you find a solution how to get a few layers on a sharp trailing edge with either snappyHexMesh or cfmesh. I tried both and I was not successful. With snappyHexMesh (see attached figure) at the trailing edge no layer was added. (The picture is a section of the onera M6 wing).



For the one which is interested the corresponding case can be found in https://wiki.openfoam.com/OneraM6_by_Michael_Alletto.


I also tried cfmesh to get better resolved the wall but if i add layers at the wing I optain highly skewed faces close to the sharp trailing edge. Is there a way to avoid highly skewed faces at a sharp trailing edge?


I found a airfoil meshed with cfmesh (https://github.com/petebachant/foilInBox-cfMesh) but in this case the trailing edge is blunt. He got 10 layers around the airfoil and the mesh did not contain highly skewed faces.

[Brainwaves]

Has anyone found a solution?