Hello,
I'm trying to solve a problem on a uniform mesh, refine part of it, and
then solve the problem again on the new mesh, using the previous solution
as an initial guess. To that end, I set
FEMSystem::project_solution_on_reinit() = true and call
EquationSystems::reinit() after each mesh refinement.
I set MeshRefinement::face_level_mismatch_limit() = 0 so that I can have
neighboring refined and unrefined elements like this:
o---o---o---o---o---------------o
| | | | | |
| | | | | |
o---o---o---o---o |
| | | | | |
| | | | | |
o---o---o---o---o |
| | | | | |
| | | | | |
o---o---o---o---o |
| | | | | |
| | | | | |
o---o---o---o---o---------------o
I mark some elements for refinement, then call
MeshRefinement::refine_elements() to refine them, like in Misc Ex 4.
However, when I call EquationSystems::reinit(), I get
"
Assertion `!libmesh_assert_pass' failed.
"
with the backtrace
"
#0 libMesh::MeshRefinement::test_level_one (this=0x7fffffffc950,
libmesh_assert_pass=true) at src/mesh/mesh_refinement.C:410
#1 0x00007ffff5e2f3fe in libMesh::MeshRefinement::coarsen_elements
(this=0x7fffffffc950) at src/mesh/mesh_refinement.C:605
#2 0x00007ffff615cec2 in libMesh::EquationSystems::reinit
(this=0x7fffffffd3f0) at src/systems/equation_systems.C:239
"
It appears that somewhere, MeshRefinement::_face_level_mismatch_limit
becomes 1 (although setting a breakpoint in
MeshRefinement::face_level_mismatch_limit() revealed no calls but that
which was used to set it to 0) and so MeshRefinement::coarsen_elements()
calls for the failed assertion.
I'm I missing something (a flag somewhere?), or does projecting on reinit
require the mesh satisfy the "level one" rule?
Thanks,
Harriet
------------------------------------------------------------------------------
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users
