Assistance Needed with Overset Mesh Configuration in Nalu-Wind Using TIOGA

[physicsWorks42]

Dear Exawind-Community

I am currently running a simulation that utilizes the overset method with the TIOGA library, involving three distinct meshes: the background mesh, wake refinement, and a cylinder. My interest in this setup was piqued by the Nalu-Wind documentation on TIOGA, which discusses a similar case (see documentation). Unfortunately, I could not find the specific mesh described, so I created a new one (see attachment).

After setting up and running Nalu-Wind with this new mesh, I encountered some issues with how TIOGA handles the mesh interfaces. Ideally, I would like the background mesh to only have hole-nodes corresponding to the wake (a square), and the wake to only have hole-nodes around the cylinder (a circle), as illustrated in the documentation here:

However, my current mesh configuration appears as follows:

Included in the attachment are the mesh file and the input YAML file configured for the latest Nalu-Wind release (2.0.0).

I would greatly appreciate any guidance on how to achieve a more accurate hole-cell configuration similar to the one shown in the documentation. Has anyone successfully run the original case or a similar one with a "neat" hole-cell setup? Any advice on adjusting my YAML input file to improve the mesh handling would be very helpful.

Thank you in advance for any assistance you can provide.

bm_wake_cylinder.zip

[itopcuoglu]

Does your simulation run to completion?

[physicsWorks42]

Does your simulation run to completion?

Yes, the simulation runs to completion. However, the solution appears awkward because the nodes within the solid structure (the cylinder, in this case) are not designated as holes, as illustrated in my images.

[psakievich]

@itopcuoglu any updates on this one?

[itopcuoglu]

I haven't looked into this yet. This is potentially related to another bug we encountered, which I'm going investigate soon.

[itopcuoglu]

@physicsWorks42 You need to change the variable symmetry_direction under tioga_options from 2 to 3, or just omit that option since the default value is 3. The hole map functionality works differently when that value is not set to 3. The iblank fields look as expected when that is corrected.

Also, you do not need the set_resolutions: no option. That has recently been made obsolete, and it was not necessary even prior to that. Let me know if this resolves your issue.