Different behavior in GEOM and OBST with prescribed VEL and TMP_FRONT

[johodges]

Describe the bug
I was running a small test case of a heat gun blowing hot air on a plate as part of evaluating the impinging jet model heat transfer model. I wanted to try using a circular nozzle to better align with the nozzle geometry, but was observing some oddities in the boundary conditions with GEOM. I put together the simple test case below to demonstrate. In this case the heat gun is represented as a rectangular prism for simplicity. In one mesh the heat gun is defined as an OBST and in the other as a GEOM. The attached video shows the different behavior (further mesh = GEOM, closer mesh = OBST).

Desktop (please complete the following information):

• OS: Windows
• Version: 6.9.1 release

Additional context

heat_gun.zip

heat_gun.txt

[marcosvanella]

Thank you Jon, I'll take a look.

[marcosvanella]

Jon do an update and try the case. Also, try the more complicated case you have and let me know.
Thanks

[johodges]

I can confirm the update fixes the first test case. Unfortunately, the more complicated case still fails. The attached version uses a cylinder rather than a rectangular prism for the jet. Similar behavior to the original issue.
bench_scale.txt

[marcosvanella]

On this case we you have 2 cells across the cylinder making all cells/faces across it cut-cells/faces. The simple method we use transferring vorticity and stresses information to cartesian edges breaks apart here, and it seems you would need substantial resolution on the heat gun tip. We have to come up with a scheme of representing sub grid scale cut-cell features for momentum. I'm thinking of discretizing these in a completely unstructured manner. I'll look into how to go about this for general polyhedrons.