Perfect Periodic BC #266
Perfect Periodic BC #266
Conversation
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
from
693dae7 to
e5e8cf1
Compare
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
from
e64674f to
9e78b76
Compare
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
3 times, most recently
from
679e555 to
41beca2
Compare
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
2 times, most recently
from
d40d53e to
185a215
Compare
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
4 times, most recently
from
a0cfbf3 to
6671aa1
Compare
The boundary elements on one of the periodic pairs are added back with both MFEM and Gmsh files.
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
from
9861daa to
61b9b89
Compare
|
@hughcars , I don't seem to have necessary permissions to add reviewers but I think this is in good shape for review. I would still need to add support for lexographic elements but that shouldn't block the review. Thanks! |
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
4 times, most recently
from
0b81873 to
4ab35ea
Compare
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
from
4ab35ea to
6667c97
Compare
…er attributes pair.
LeilaGhaffari
force-pushed
the
lgh/periodic-bc
branch
from
08abef0 to
63bd460
Compare
TODO: add periodicity to the cavity example
… be used in MakePeriodic
There was a problem hiding this comment.
Almost done. The issue with the hex/prism/tet waveguide is going to mean the gmsh meshes need to be refined, cavity rebaselined and then the tables of the corresponding modes filled in. Additionally, do not need to remove the attributes on donor and receiver (and can patch mfem to not do the same as well) then we can add computing the power flux for each mode to the regression tests.
- Improves geometry fidelity and avoids issue in mfem where periodicity fails for meshes with too few elements in the periodic face. - Change geodata.cpp to not erase boundary attributes on periodic faces. Additionally to not fail if an incoming mesh has an interior boundary. - Regenerate baseline test data based on the updated meshes. - Update documentation to use Teflon in waveguide - Fixes issue with underspecified Q for waveguide - Makes documentation cleaner compared to cavity
|
I've addressed all the issues here in https://github.com/awslabs/palace/tree/hughcars/periodic-bc including merging main in. I'm not sure how to merge those commits onto this branch, given it's originally from a fork. You can either merge them in, or I can open another PR from that branch. It should just be a fast forward from this branch. |
|
Thanks @hughcars for fixing the issues. I just pushed your fixes to this branch. Is there anything else I can help with before merging? |
I just read through it again and found a few small things I missed, there's one more commit on |
Just pushed it, thanks! |
This PR is aiming at adding support for perfect periodic BC.
I have added two meshes in examples/cube/.
cube.meshis a regular cube andcube_periodic.meshis periodic.@sebastiangrimberg , are the boundary conditions specified here and here making sense?
Edit: Just noticed mfem/data/periodic-cube.mesh and it seems to be different than the one I created. This one fails too though.