[GeoMechanicsApplication] Extend extrapolator to support lines #14030
Conversation
| expected_extrapolation_matrix <<= UblasUtilities::CreateMatrix({{ 0.923275, 0.444444,-0.367719}, | ||
| {-0.367719, 0.444444, 0.923275}}); |
There was a problem hiding this comment.
We should still check these values
There was a problem hiding this comment.
outdated, now an analytical result.
| // Used the integration method found in the 2d3 Timoshenko Beam element | ||
| constexpr auto integration_method = GeometryData::IntegrationMethod::GI_GAUSS_3; |
There was a problem hiding this comment.
I think this integration method is strange for a 2D2 (same for the 2D3 case), but for the moment I just took the same as the timoshenko beam (since that's what we'll use it for)
There was a problem hiding this comment.
Outdated, now using GAUSS_2
markelov208
left a comment
markelov208
left a comment
There was a problem hiding this comment.
Hi Richard, it looks the extension was very easy to do. ;)
| Matrix result(rInitializerList.size(), rInitializerList.begin()->size(), 0.0); | ||
| for (std::size_t i = 0; const auto& r_row : rInitializerList) { | ||
| std::ranges::copy(r_row, row(result, i).begin()); | ||
| ++i; | ||
| } |
There was a problem hiding this comment.
The function does not check the input for correctness. Is it on purpose? To have it fast, right?
| namespace Kratos::Testing | ||
| { | ||
|
|
||
| KRATOS_TEST_CASE_IN_SUITE(NodalExtrapolator_GivesCorrectExtrapolationMatrix_For2D3NLine, |
There was a problem hiding this comment.
There is no test when numbers of nodes and gauss points are equal. The output is expected. However, is it useful to have such a test to increase the code coverage? Or just in case?
avdg81
left a comment
avdg81
left a comment
There was a problem hiding this comment.
Hi Richard,
Thank you very much for extending the linear nodal extrapolator such that it can also handle line geometries. The changes are clear and nicely unit-tested. I only have a few minor suggestions, so I expect we can merge this soon.
| return result; | ||
| } | ||
|
|
||
| Matrix UblasUtilities::CreateMatrix(const std::initializer_list<std::initializer_list<double>>& rInitializerList) |
There was a problem hiding this comment.
The code suggests that we receive a list of rows. Perhaps we can rename the function parameter such that this is clearer? For instance:
| Matrix UblasUtilities::CreateMatrix(const std::initializer_list<std::initializer_list<double>>& rInitializerList) | |
| Matrix UblasUtilities::CreateMatrix(const std::initializer_list<std::initializer_list<double>>& rRows) |
| { | ||
| if (rInitializerList.size() == 0) return {}; | ||
|
|
||
| Matrix result(rInitializerList.size(), rInitializerList.begin()->size(), 0.0); |
There was a problem hiding this comment.
Since there are no checks on the sizes of the supplied rows, it may be safer to determine the maximum row length and use that as the number of columns of the matrix. What do you think? Or should we raise an error when the rows don't have one and the same size?
| nodal_extrapolator.CalculateElementExtrapolationMatrix(geometry, integration_method); | ||
|
|
||
| // clang-format off | ||
| Matrix expected_extrapolation_matrix = UblasUtilities::CreateMatrix( |
There was a problem hiding this comment.
Let's make this one const auto:
| Matrix expected_extrapolation_matrix = UblasUtilities::CreateMatrix( | |
| const auto expected_extrapolation_matrix = UblasUtilities::CreateMatrix( |
| {0.5, 0.5}}); | ||
| // clang-format on | ||
|
|
||
| KRATOS_EXPECT_MATRIX_NEAR(extrapolation_matrix, expected_extrapolation_matrix, 1e-6) |
There was a problem hiding this comment.
When we #include "tests/cpp_tests/test_utilities.h", we can use the default absolute tolerance in the check:
| KRATOS_EXPECT_MATRIX_NEAR(extrapolation_matrix, expected_extrapolation_matrix, 1e-6) | |
| KRATOS_EXPECT_MATRIX_NEAR(extrapolation_matrix, expected_extrapolation_matrix, Defaults::absolute_tolerance) |
📝 Description
Extends the extrapolator to be able to extrapolate lines and adds a regression unit test for it. Next to that, a minor addition is done to the ublas utilities to also be able to generate a matrix.