svFSI error, Failed to project faces

[David Parker]

Hi Sara,

I have fixed a bug in the write FSI mesh files but svFSI is still failing with the files. I'm digging into the svFSI code to figure out what is going on. Stay tuned!

Cheers,
Dave

[Mattia Zanchi]

Hi David,
Working on this topic I've seen that there can be some trouble about the "global ID" numeration. In particular some nodes have the same "Global ID" value. I leave a picuture attached to be more clear and hoping it can be helpful. The mesh in the attached image is obtained following the sv tutorial.
Anyways if the mesh is create thx to meshmixer, or some similar software, all work correctly.

Regards
Mattia

[David Parker]

Hi Mattia,

Thanks for your observation! I think this is due to the bug I fixed but I will check to make sure.

Cheers,
Dave

[David Parker]

Hi Mattia and Sara,

I believe that I've fixed the problem with exporting mesh-complete files from SV, the node and element IDs were wrong. I will test a bit more before I check-in the code.

I also wrote a create-fsi-mesh-complete.py Python script (https://github.com/ktbolt/cardiovascula ... ete/python) that can create FSI mesh files from the SV Meshes .vtu and .vtp files, and a Models .mdl file. I tested this on some simple models and it seems to work. You may want to use the script until I can check in the fix and build new SV installers (end of December).

Cheers,
Dave

[Mattia Zanchi]

Hi dave,
Thx for your helpful advice and your work.
I will definitely try the script

Regards
Mattia

[Sara Zambon]

Hi David,
Thanks a lot for the replies and the support!
I will stay tuned for the new SV installer, meanwhile i'm going to use meshmixer or try your python script

Cheers!

[Sara Zambon]

P.S. to Mattia: suppongo tu sia italiano, se non fosse troppo disturbo (e dato che anche tu stai lavorando su Ubuntu) potrei chiederti un'altra questione su svFSI? (sono ancora agli inizi purtroppo e sul forum non ho trovato risposta)

[Mattia Zanchi]

Hi Sara,
I use english so everyone can read it.
I think that for the red line there is no problem, maybe just a visual difference between results of FS and results of MS ( I think you are solving a FSI ALE problem so you have 2 different domain).
Talking about mesh motion I haven't try nothing yet.

regards
mattia

P.S. if you need more info write me at m.zanchi15@studenti.unibg.it , can't find your email on the web.

[Sara Zambon]

Sorry everyone, i was in a bit of a rush yesterday, and didn't write in english!
I simply asked if the red text lines and exclamation marks "!" on the terminal interface stood for some underlying errors or warnings; and asked for some clues about imposing solid mesh motion as a boundary conditions for an svFSI simulation

Thanks Mattia! i surely will

[Vijay Vedula]

Yes, the colored terminal output is a visual effect to distinguish equations in a multi-equation system. You can turn it off in the input file using the command

Code: Select all

Colored terminal output: f

Brackets '[ ]' in the linear and non-linear solvers tabs indicate full convergence to the desired tolerances. Exclamation marks '!' in these tabs indicate non-convergence.