Backing seems to be set to VOID with complex geometries

[ESett5555]

Hi everyone,

As I said in discussion #16349, I have been working with heat conduction using GEOM.

I tried to use a more complex geometry that consists of a vertical hollow cylinder with a dome on top. These two are approximately 10 cm thick. This geometry was created on a CAO software, imported on Blender using BlenderFDS, and then I created a test case with it.

I wanted to heat the fluid under the dome and to see how the heat would transfer through the dome and the cylinder and so I tried two different methods:

1. I used a simple propane fire under the dome
2. I used &INIT to specify a constant HRRPUV for the fluid under the dome

The other parameters (&MATL, &DUMP, &MESH, &VENT...) are the same as in discussion #16349, with the exception that I now use 8 meshes instead of one, because of the changes that have been made to address Backing='EXPOSED' in a separate mesh.

The results are really surprising to me :

case 1. :

igloo.cas1.mp4

case 2.:

igloo.cas.2.mp4

Then I have two questions :

• Does someone know why is a part of the geometry not heating at all (in both cases) ? It seems to me that a part of the geometry backing is set to VOID, but I do not understand why. It might be linked to the solid mesh, but i don't understand how.
• Then, in case 2., it looks like a horizontal "line" is cooling around the cylinder. You can see it clearly between seconds 2 and 3 of the video I made. Does someone know why ?

Thank you for your answers and for your time !

PS: I will give a .txt of case 1 to help to see the .fds I have made. It uses a .bingeom file for the GEOM line. If needed, I could send this .bingeom file, but not here as Github does not accept this type of file. If someone knows how I can share it here, I would be interested.

igloo case 1.txt

[marcosvanella]

@ESett5555 have you run the case with the latest nightly bundle? There are bundles from 4 days ago that have the fixes described in the other discusison.

https://github.com/firemodels/test_bundles/releases/tag/FDS_TEST

[marcosvanella]

Also, can you post the other case 2 and the geometry file? We'd like to try the cases here. Thanks!

[ESett5555]

The FDS version installed on my computer is FDS-6.11.0-122-g199afe5-nightly, I installed it last friday (03/07). It should have the fixes described in the other discussion.

I made two .zip to send the .fds and the .bingeom file together. The .bingeom files should be exactly the same, but to be sure I am sending you both in separate .zip.

Please tell me if you need more information, and have a great day !
igloo_case_1.zip
igloo_case_2.zip

PS: I also tried to run the same tests with an older version, and the results were similar

[marcosvanella]

I made this an issue. The back face selection is not working properly when the ray tracing algorithm hits several solid surfaces in the back. I'll let you know when the fix will be up.

[marcosvanella]

@ESett5555 Try the latest source code/nightly bundles in your cases. The issues should be addressed.
Let me know if you see any other problems. Thanks for reporting.

[ESett5555]

Hello @marcosvanella,

I tried the latest nightly bundle, the issues are not here anymore an heat conduction is happening as it should.
Thank you for the fix and have a great day !