Added sliding boundary nodes to VariationalMeshSmoother #4216

pbehne · 2025-07-17T19:52:00Z

Pull Request: Sliding Boundary Constraint Refinement and Dimensional Generalization

Summary

This pull request implements sliding external and subdomain boundary node constraints in the VariationalMeshSmoother
In addition to expanding test coverage to test sliding boundary nodes, 1D and 3D test cases were added to the unit test suite.

Highlights

Constraint Framework Refactor
- Introduced PointConstraint, LineConstraint, and PlaneConstraint classes to represent fixed nodes, nodes constrained to a line, and nodes constrained to a plane
- Implemented a unified intersect() interface for combining constraints using std::variant
- Added robust geometric logic for line-line, plane-line, and plane-plane intersections
- Fallback to PointConstraint if constraints fail to intersect cleanly
Sliding Boundary Enhancements
- Implemented sliding boundary node constraints: previously, boundary and subdomain boundary nodes were fully fixed and not allowed to slide
- Nodes on external boundaries internal subdomain boundaries now slide along planes and lines unless constrained by geometry
Expanded Tests and Validation
- Added new smoother test cases in 1D and 3D
- Replaced DistortSquare with DistortHypercube, supporting 1D, 2D, and 3D skewing
- Boundary nodes are now distorted within the faces and edges they belong to (while preserving corner vertices)
- Increased complexity of subdomain boundary test to cover more complex cases such a an element having neighbors of multiply different subdomain ids

References

Related to Revamp the variational mesh smoother #4082

Visualization of `testVariationalHexMultipleSubdomains` Test

The distorted mesh is on the left, smoothed mesh is on the right.

… 3D, not sliding edge nodes.

…re considered for sliding boundary constraints. This helps with adjacent corners nodes in triangular elements.

…ain tests to account for barely-moved subdomain boundary nodes that coincide with the mesh boundary.

… sides are sometimes not found.

…ks better, but both constaints should be conmbined for strict enforcement.

Implemented structs to represent Point, Line, and Plane constraints and logic to combine (intersect) them. This is more robust than the previous version and gives the correct constraints for nodes that are both subdomain boundary nodes and external boundary nodes. Ref libMesh#4082

Replaced DistortSquare class with the DistortHypercube class, which generalizes dimension and distorts boundary nodes within boundaries instead of leaving them unchanged. The helper also checks all applicable node dimensions instead of receiving the dimension to check as a parameter. Renamed the `center_distortion_is` helper function to `distortion_is` and updated to check sliding boundary nodes for distortion.

…ot intersect.

Instead of looking for the boundary between sub_id2 and sub_id2, we now look for the boundary between sub_id1 and "not sub_id1". This better handles cases where multiple faces of an element have neighbors of multiple different subdomain ids.

…LaplaceSmoother functions.

roystgnr

There's surprisingly little for me to complain about, for such a large PR.

include/systems/variational_smoother_constraint.h

src/systems/variational_smoother_constraint.C

moosebuild · 2025-07-18T03:22:25Z

Job Coverage, step Generate coverage on f8c37d4 wanted to post the following:

Coverage

	f5e204	#4216 f8c37d
	Total	Total	+/-	New
Rate	64.33%	64.42%	+0.09%	87.65%
Hits	75207	75559	+352	369
Misses	41701	41725	+24	52

Diff coverage report

Full coverage report

Warnings

New new line coverage rate 87.65% is less than the suggested 90.0%

This comment will be updated on new commits.

…tor==.

…s private.

roystgnr · 2025-07-23T02:43:59Z

From Slack, for public reference:

"Okay, just pushed the polymorphic approach. Let me know which one you prefer and we'll stick with that."

"I think I like the std::variant version better. Sorry to make you go to so much work! I'd be okay with either version."

"Yes, the polymorphic feels a bit more cluttered, doesn’t it? No worries, I was interested in how it would turn out"

I think my only remaining real complaint is that, even if we end up relying on exceptions here, at a minimum we need to disable the tests relying on them until the --disable-exceptions CI passes.

roystgnr

I'm thrilled with this now except for a couple of the whitespace changes; @pbehne is going to rewrite history now, and has sworn he won't edit anything else but whitespace in the process.

…exceptions.

jwpeterson

Looks good to me although I can't really comment on the technical details. I think some of the new std::variant and std::visit C++17 code could potentially be commented more. I know that it's possible to look up how these things work, but (to my knowledge) we don't use this anywhere else in the library so it would be handy to have a nearby comment explaining how these "modern" C++ concepts help to implement the overall design.

include/systems/variational_smoother_constraint.h

src/systems/variational_smoother_constraint.C

src/systems/variational_smoother_system.C

include/systems/variational_smoother_constraint.h

moosebuild · 2025-07-25T13:05:15Z

Job Min gcc on f8c37d4 : invalidated by @jwpeterson

Retry to see if "failed to get checksum" error is resolved automatically

jwpeterson · 2025-07-30T13:41:46Z

FYI @pbehne this PR seems to have broken the --disable-deprecated build (see here, for example: https://civet.inl.gov/branch/14145/). The error message is:

../src/systems/variational_smoother_system.C: In member function 'void libMesh::VariationalSmootherSystem::prepare_for_smoothing()':
../src/systems/variational_smoother_system.C:158:37: error: no matching function for call to 'libMesh::Tri3::set_node(int)'
               target_elem.set_node(0) = &node_0;
                                     ^
In file included from ../src/systems/variational_smoother_system.C:20:
./include/libmesh/elem.h:2571:6: note: candidate: 'virtual void libMesh::Elem::set_node(unsigned int, libMesh::Node*)'
 void Elem::set_node (const unsigned int i,
      ^~~~
./include/libmesh/elem.h:2571:6: note:   candidate expects 2 arguments, 1 provided
../src/systems/variational_smoother_system.C:159:37: error: no matching function for call to 'libMesh::Tri3::set_node(int)'
               target_elem.set_node(1) = &node_1;
                                     ^
In file included from ../src/systems/variational_smoother_system.C:20:
./include/libmesh/elem.h:2571:6: note: candidate: 'virtual void libMesh::Elem::set_node(unsigned int, libMesh::Node*)'
 void Elem::set_node (const unsigned int i,
      ^~~~
./include/libmesh/elem.h:2571:6: note:   candidate expects 2 arguments, 1 provided
../src/systems/variational_smoother_system.C:160:37: error: no matching function for call to 'libMesh::Tri3::set_node(int)'
               target_elem.set_node(2) = &node_2;
                                     ^
In file included from ../src/systems/variational_smoother_system.C:20:
./include/libmesh/elem.h:2571:6: note: candidate: 'virtual void libMesh::Elem::set_node(unsigned int, libMesh::Node*)'
 void Elem::set_node (const unsigned int i,
      ^~~~
./include/libmesh/elem.h:2571:6: note:   candidate expects 2 arguments, 1 provided

so the fix should be simple, just use the non-deprecated Elem::set_node() API instead.

pbehne · 2025-07-30T13:51:45Z

Thanks for the heads up, I'll add the fix to #4218!

pbehne · 2025-07-30T13:54:44Z

FYI, only the deprecated version shows up in doxygen. I had to go look at elem.h to see the non-deprecated version.

jwpeterson · 2025-07-30T13:58:11Z

FYI, only the deprecated version shows up in doxygen. I had to go look at elem.h to see the non-deprecated version.

OK, that's interesting... I guess Doxygen is somehow seeing the "pre-processed" version of the source (and the defines are somehow set such that it doesn't generate docs for the non-deprecated version)? Maybe that's a normal Doxygen thing, I don't know.

pbehne · 2025-07-30T14:08:21Z

Yeah, I figured I'd mention it since I recall @roystgnr mentioning that he and you typically just use the header files themselves for documentation. As someone less familiar with the library, I prefer doxygen so I don't have to dig through parent headers to see what a given class does and doesn't inherit.

roystgnr · 2025-07-30T18:31:21Z

I think the problem is that we only update the libMesh Doxygen manuallly, by which I mean "almost never", by which I mean "Oh my God the files still have copyright 2002-2018 headers". The libMesh Doxygen on the MOOSE site gets autoupdated, which I guess I thought was good enough ... but it doesn't matter if the MOOSE copy marks something as deprecated if that's not the copy everybody is reading.

pbehne added 23 commits July 17, 2025 13:09

Sliding boundary nodes working for 2D and 3D (coplanar edges only for…

4a50ef6

… 3D, not sliding edge nodes.

Moved logic to constrain node to plane into dedicated function.

ecad271

Added method to constrain nodes to lines in 3D.

eb7459c

documentation

d76ad06

Added logic where only nodes that share the same bid with neighbors a…

1262cf6

…re considered for sliding boundary constraints. This helps with adjacent corners nodes in triangular elements.

Added equilateral triangle as the target element for TRI3s.

405e516

Cleaned up constrints, added relative_fuzzy_equals to multiple subdom…

11331fb

…ain tests to account for barely-moved subdomain boundary nodes that coincide with the mesh boundary.

Nodes on internal subdomain boundaries are now allowed to slide.

a9f92bb

Fixed bug in constraints where the correct common element and matched…

3e11730

… sides are sometimes not found.

Eliminated a level of indentation.

ed5b413

Moved subdomain constraints ahead of boundary constraints. Result loo…

0ea66cb

…ks better, but both constaints should be conmbined for strict enforcement.

Saving current changes that almost fix constraint issues, but not quite.

1c3b0bf

Fixed bug causing 'node already constrained' error.

4c07077

Increased the complexity of the subdomain boundary test.

15cf24a

Added fallback to fixed node constraint in event that constrints do n…

c6e37e4

…ot intersect.

Added 1D and 3D tests for the VariationalMeshSmoother.

86eee47

Include variant.

c9973d1

Removed some assertions that don't apply in 1D.

db4b69d

Addressed maybe-uninitialized error.

92200b5

Separated testSmoother function into testVariationalSmoother and test…

7aeaa2b

…LaplaceSmoother functions.

roystgnr requested changes Jul 17, 2025

View reviewed changes

pbehne added 5 commits July 18, 2025 07:46

Swaped doxygen syntax from '///' to '/***/'

d4e397b

Added documentation for Point/Line/PlaneConstraint operators < and ==.

b32de64

Document the throwing of error in *Constraint structs.

3533633

Made *Constraint struct operator< symmetric and compatible with opera…

899e743

…tor==.

Changed try/catch to libmesh_try/libmesh_catch.

0720610

pbehne added 2 commits July 18, 2025 15:10

Changed Point/Line/PlaneConstraint structs to classes, made attribute…

4153e1e

…s private.

Added _tol attibute to Point/Line/Plane constraints.

cf08059

roystgnr requested changes Jul 23, 2025

View reviewed changes

pbehne added 4 commits July 23, 2025 15:23

Silenced variational smoother solver.

e6a353f

Added dummy InvalidConstraint class to eliminate dependent on thrown …

bca5f98

…exceptions.

Added explanatory comment.

73e927c

clang-format on variational_smoother_constraint.C

57c7c4e

pbehne force-pushed the variational_smoother_sliding_boundary_nodes branch from d2b4489 to 57c7c4e Compare July 23, 2025 21:29

clang-format on mesh_smoother_test.C

e2d026a

jwpeterson approved these changes Jul 24, 2025

View reviewed changes

pbehne added 3 commits July 24, 2025 10:22

Addressed John's review.

ea104ce

indented contents of switch statement.

505bb55

reindeneted some unindents.

f8c37d4

jwpeterson requested a review from roystgnr July 25, 2025 13:05

roystgnr approved these changes Jul 28, 2025

View reviewed changes

roystgnr merged commit 079a331 into libMesh:devel Jul 28, 2025
22 checks passed

roystgnr deleted the variational_smoother_sliding_boundary_nodes branch July 28, 2025 20:54

Added sliding boundary nodes to VariationalMeshSmoother #4216

Added sliding boundary nodes to VariationalMeshSmoother #4216

Uh oh!

Conversation

pbehne commented Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request: Sliding Boundary Constraint Refinement and Dimensional Generalization

Summary

Highlights

References

Visualization of testVariationalHexMultipleSubdomains Test

Uh oh!

roystgnr left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

moosebuild commented Jul 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Coverage

Warnings

Uh oh!

roystgnr commented Jul 23, 2025

Uh oh!

roystgnr left a comment

Choose a reason for hiding this comment

Uh oh!

jwpeterson left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

moosebuild commented Jul 25, 2025

Uh oh!

Uh oh!

jwpeterson commented Jul 30, 2025

Uh oh!

pbehne commented Jul 30, 2025

Uh oh!

pbehne commented Jul 30, 2025

Uh oh!

jwpeterson commented Jul 30, 2025

Uh oh!

pbehne commented Jul 30, 2025

Uh oh!

roystgnr commented Jul 30, 2025

Uh oh!

Uh oh!

pbehne commented Jul 17, 2025 •

edited

Loading

Visualization of `testVariationalHexMultipleSubdomains` Test

moosebuild commented Jul 18, 2025 •

edited

Loading