[ari003]

Hello everyone,
Hope you all are doing good. I'm stuck with an issue with gmsh which might be pretty simple for few. And also, I'm creating this thread in the main forum so that it reaches to more number of CFD enthusiasts.
The problem statement is as follows:-
1. I'm trying to do a numerical analysis in SU2 before creating the mesh in GMSH platform.

2. The scenario is a simple DFVLR airfoil simulation in transonic regime. My aim is to proceed with unstructured meshing concept.

3. The picture (pic 1) is attached below which represents the automatic meshing.

4. But I need to refine the region adjacent to the airfoil wall. So, I used 'Transfinite curve' option in GMSH (picture 2.)

5. By doing so when I export it in .su2 format and run it in SU2 it gives an error - Zero CV area found.

6. When I remove the 'Transfinite curve" option and proceed with normal automatic meshing it runs completely fine but the y+ is in the buffer region (which is undesirable for my case).

7. After encountering these error I tried to do some hit and trial case and noticed these as follows:-
a. When the transfinite curve option is set until 25 it runs fine until
convergence but in the demand of more refinement with 26, 30, 40,50
it is giving the same error.
b. Now, I tried to locally use the 'transfinite curve' option with 7 lines and
it gave satisfactory result (picture 3). But when it exceeds to more lines
on the wall the error of divergence was observed in SU2.

My aim is to refine the region around the airfoil but it seems like I'm running out of ideas. If anyone is aware with any kind of solution it will be very helpful. Please let me know if anyone needs more information on the problem.

[naffrancois]

Hello,

It is not very clear to me where you apply transfinite curve to get picture 2. Do you only apply it along the airfoil curves ? This way you get very small triangles close to the wall to meet y+ criteria. I am surprised anyway that the area of the smallest triangle falls below your cfd code preprocessor threshold...

I don't know su2 but I guess it can handle hybrid grids. You may try to construct blocks around the airfoil where you apply transfinite curve/transfinite surface + recombine with the aim of getting anisotropic quads in the boundary layer. Further away extend the grid using triangles up to farfield.

[ari003]

Originally Posted by naffrancois

Hello,

It is not very clear to me where you apply transfinite curve to get picture 2. Do you only apply it along the airfoil curves ? This way you get very small triangles close to the wall to meet y+ criteria. I am surprised anyway that the area of the smallest triangle falls below your cfd code preprocessor threshold...

I don't know su2 but I guess it can handle hybrid grids. You may try to construct blocks around the airfoil where you apply transfinite curve/transfinite surface + recombine with the aim of getting anisotropic quads in the boundary layer. Further away extend the grid using triangles up to farfield.

Thanks for your response Sir,
Yes I apply at the airfoil wall/surface to get the refines unstructured mesh around airfoil (pic 2).
I already used the hybrid mesh technique which gives satisfactory results but my problem will become more complicated at later stage and with hybrid mesh it will not be possible so need to make use of the unstructured mesh somehow.

I am surprised anyway that the area of the smallest triangle falls below your cfd code preprocessor threshold...

I didnt get this, can you elaborate a litle bit. Thanks in advance.

[naffrancois]

Well I am surprised but I do not have an explanation as I do not know su2. What seems to tell su2 preprocessor is that either:
- it found a badly oriented triangle, then calculating area may give negative value
- it found a very small triangle. I guess su2 checks area>tolerance.

I would discard first reason, as I would expect su2 throw a different, more specific message, but again I do not know.

Maybe you would have better luck posting in the su2 forum, as to me this does not sound like a problem coming from gmsh.

[ari003]

Originally Posted by naffrancois

Well I am surprised but I do not have an explanation as I do not know su2. What seems to tell su2 preprocessor is that either:
- it found a badly oriented triangle, then calculating area may give negative value
- it found a very small triangle. I guess su2 checks area>tolerance.

I would discard first reason, as I would expect su2 throw a different, more specific message, but again I do not know.

Maybe you would have better luck posting in the su2 forum, as to me this does not sound like a problem coming from gmsh.

Thank you Sir. The problem is sorted now. I used "Distance" along with "Threshold" to tackle this issue. But again as you said SU2 has tolerance limit with the area and excessive decrease of the cell size will lead to the same issue of zero control volume.