[SD-LIE] Small deformation process with lower-dimensional interface elements #1452
Conversation
norihiro-w
force-pushed
the
sd-lie
branch
7 times, most recently
from
3261f97
to
e4bbeca
Compare
norihiro-w
changed the title
norihiro-w
force-pushed
the
sd-lie
branch
2 times, most recently
from
c0aaefa
to
ef84f37
Compare
|
Please rebase to 6.0.7 before merging. |
norihiro-w
force-pushed
the
sd-lie
branch
3 times, most recently
from
befefd4
to
ff6b970
Compare
norihiro-w
force-pushed
the
sd-lie
branch
6 times, most recently
from
2cb0071
to
ba3adbb
Compare
| "sigma_xy", 1, | ||
| makeExtrapolator( | ||
| getExtrapolator(), _local_assemblers, | ||
| &SmallDeformationLocalAssemblerInterface::getIntPtSigmaZZ)); |
There was a problem hiding this comment.
xy output does not match the getter for ZZ. But I have discovered this bug just a minute ago in SD.
There was a problem hiding this comment.
corrected following your PR
| std::unique_ptr<NumLib::LocalToGlobalIndexMap> | ||
| _local_to_global_index_map_single_component; | ||
|
|
||
| std::vector<MeshLib::Element*> _vec_matrix_elements; |
There was a problem hiding this comment.
It seems that these elements and nodes are kept untouched. Is it possible to use
std::vector<MeshLib::Element const*> or std::vector<MeshLib::Node const*>
There was a problem hiding this comment.
no it's not possible, because they are passed to functions which take the non-const versions as arguments.
| { | ||
| if (pv_name != "displacement" | ||
| && pv_name.find("displacement_jump")==std::string::npos) | ||
| OGS_FATAL("Found a process variable name '%s'. It should be 'displacement' or 'displacement_jumpN'"); |
There was a problem hiding this comment.
N at the end of displacement_jumpN does not match with find parameter.
| @@ -0,0 +1,17 @@ | |||
| /** | |||
| * \copyright | |||
| * Copyright (c) 2012-2015, OpenGeoSys Community (http://www.opengeosys.org) | |||
|
|
||
| // set fracture property assuming a fracture forms a straight line | ||
| setFractureProperty(DisplacementDim, | ||
| *_vec_fracutre_elements[0], |
There was a problem hiding this comment.
typo in fracture.
And above in the argument list too. ✅
| // set fracture property assuming a fracture forms a straight line | ||
| setFractureProperty(DisplacementDim, | ||
| *_vec_fracutre_elements[0], | ||
| *_process_data._fracture_property.get()); |
There was a problem hiding this comment.
While tracing where the get() is from (it's unique_ptr...)...
PL/SDwLIE/Common/FractureProperty.h is included in itself ✅
| getExtrapolator(), _local_assemblers, | ||
| &SmallDeformationLocalAssemblerInterface::getIntPtSigmaXY)); | ||
|
|
||
| auto mesh_prop_sigma_xx = const_cast<MeshLib::Mesh&>(mesh).getProperties().template createNewPropertyVector<double>("stress_xx", MeshLib::MeshItemType::Cell); |
There was a problem hiding this comment.
could you break the long lines? ✅
| std::unique_ptr<NumLib::LocalToGlobalIndexMap> | ||
| _local_to_global_index_map_single_component; | ||
|
|
||
| std::vector<MeshLib::Element*> _vec_matrix_elements; |
There was a problem hiding this comment.
In getFractureMatrixDataInMesh() the mesh elements are sorted in two sets those for fracture and those from matrix: One could also use std::partition_copy there.
There was a problem hiding this comment.
I am not sure what you mean. Later the type of _vec_fracture_elements will be changed to std::vector<std::vector<MeshLib::Element*>> for multiple fractures.
There was a problem hiding this comment.
Here all is fine. In getFractureMatrixDataInMesh() one could use std::partition_copy. Just an idea.
|
@endJunction let me know if you have further comments on this. if not, I'd merge this tomorrow morning. |
|
ufz/ogs/master locked now. |
# Conflicts: # Applications/CLI/Tests.cmake
|
OpenGeoSys development has been moved to GitLab. |
followup of #1455, #1456 and #1457
This PR adds the new process type "SMALL_DEFORMATION_WITH_LIE", which includes lower-dimensional interface elements for modeling fracture deformations. To make review easier, I would like to keep this first PR as simple as possible and focus on deformation with a single fracture cutting a domain. This is indeed enough for the DECOVALEX benchmark. Complexity can be increased later on step by step.