Instability / strange behavior with increased resolution

[ryebread03]

First time using fluidX3D and I've been having issues increasing the resolution of the sample setups to even 10% of my GPUs available VRAM. I'm running a NVIDIA RTX PRO 6000 Blackwell and every time I try and increase the resolution of any of the setups beyond ~3500, I get what's in the picture below. No other lines in the sample setup have been changed besides the resolution. The simulation will converge with the entire domain filled with red. The most similar discussion I could find to this was solved by enabling subgrid. This was unsuccessful for me. Thanks for the help.

[ProjectPhysX]

Hi @ryebread03,

retaining stability at high Reynolds number is not always easy. Some general strategies:

• Enable SUBGRID model.
• Reduce lbm_u: 0.01-0.1 is suitable range.
• Make the model smaller relative to the simulation box size. Too large model will constrict the flow and create regions with too fast velocity / shear rate.
• If instability still persists, check where it originates from. Perhaps the model geometry contains a small hole where fluid is forced through.

Kind regards,
Moritz

[ryebread03]

Hi Thanks so much for the response :).
Unfortunately, I'm still struggling to get this to work. Subgrid is enabled. I Have tried numerous values of lbm_u within that range to no avail, and have also rescaled the model. The only thing that seems to affect this type of instability is the resolution specifically. below 3500, it runs fine. and at even higher resolutions, i.e. ~7000+, the model and wireframe just disappear when the simulation is run. With this NASA model in question, the instability may be originating from the small gap between the solid wing and the flap. However, I get the same problem when I import a simple solid teardrop model with no gaps/holes.