Hybrid Mesh Contour Confusion
 

Hey Guys,

I am performing 3D wing simulations using a hybrid mesh with a hexahedral near-field region (y-plus 3) to capture the boundary layer and an unstructured tetrahedral mesh beyond that. The pictures attached are for a NACA 4415 profile aspect ratio 12 wing for a Re 3 million AoA 10 deg. I generate two different mesh files for the two different regions in ICEM and then combine them in FLUENT defining the interface as Matching interface. The contours for velocity magnitude and static pressure seem to be fine but when I plot vorticity magnitude I see some spurious regions where vorticity is not expected. Can anyone explain the reason behind this. The odd region is at the interface of the two meshes.

Any insight to mitigate the problem is appreciated.

Regards

Merge meshes?
 

When you import the second mesh from ICEM, does it have a 'merge' option?

I import the two meshes in FLUENT, rather than in ICEM. I define the common interface in fluent itself.

And this is why having a nice mesh is super important.

Try merging the zones instead of an interface. At interfaces, Fluent drops to 1st order upwind discretization. Alternatively, you can try switching to 1st order upwind discretiation everywhere to see if it goes away.

I can surely try switching to 1st order upwind to check if it gets rid of the problem. I am trying to merge the two meshes in ICEM but always get a runtime error. Is there some kind of cell limit in ICEM on merging two meshes? because I am trying the same on a practice mesh and there I have no issue merging hex and tets by a prism layer

I don't know about ICEM but you can merge them in Fluent using something like:

mesh/modify-zones/merge-zones fluid-1 fluid-2 to merge the cell zones so you have only 1 zone.

Then merge the overlapping boundaries (the interface).

/mesh/modify-zones/fuse-face-zones "" ""
This will convert them from boundary faces into interior faces.

Note that the fuse operation will dbl the memory needed.

Avoid merging the interior zones because it is a memory hungry process with known issues (188722). I wonder if the trouble you have in ICEM is similar.

I tried merging in FLUENT. The first step of merging the cell-zones executed normally but when i tried fusing the two overlapping surfaces / interface , which are individually defined as interface in the boundary conditions, it failed to do so. Any clue as to what might be causing this?