Implementation of nodal source terms #1977

TomFischer · 2017-11-02T09:31:17Z

Nodal source terms are setup in the project file in the process_variable section.

<source_terms>
    <source_term>
        <geometrical_set>geometry</geometrical_set>
        <geometry>inner</geometry>
        <type>Nodal</type>
        <value>1</value>
    </source_term>
</source_terms>

endJunction · 2017-11-02T14:55:36Z

...ectFile/prj/process_variables/process_variable/source_terms/source_term/t_geometrical_set.md

@@ -0,0 +1,4 @@
+The name of the geometrical set (see \ref ogs_file_param__gml__name) in which
+the \ref
+ogs_file_param__prj__process_variables__process_variable__source_terms__ource_term_geometry


ource_term_geometry -> source_term_geometry I guess... ✅

endJunction · 2017-11-02T14:56:52Z

MeshGeoToolsLib/CreateSearchLength.h

@@ -20,6 +20,7 @@

 namespace MeshGeoToolsLib
 {
+inline
 std::unique_ptr<MeshGeoToolsLib::SearchLength> createSearchLengthAlgorithm(
    BaseLib::ConfigTree const& external_config, MeshLib::Mesh const& mesh)
 {


Would you move it to cpp, please. ✅

endJunction · 2017-11-02T15:11:17Z

ProcessLib/SourceTerms/Tests.cmake

@@ -0,0 +1,98 @@
+AddTest(


I'd move the Tests.cmake content to the corresponding process' Tests.cmake.... ✅

endJunction

You have done extensive tests, but there is no comparison, which would be nice I think.
Few documentation comments should be fixed.

endJunction · 2017-11-03T10:45:41Z

MeshGeoToolsLib/CreateSearchLength.h


-#include "MeshGeoToolsLib/HeuristicSearchLength.h"
-#include "MeshGeoToolsLib/SearchLength.h"
+namespace MeshLib


Yes. This is much nicer because of the removed includes and the fwd-decls.

endJunction · 2017-11-03T10:51:03Z

...ectFile/prj/process_variables/process_variable/source_terms/source_term/t_geometrical_set.md

@@ -0,0 +1,4 @@
+The name of the geometrical set (see \ref ogs_file_param__gml__name) in which
+the \ref
+ogs_file_param__prj__process_variables__process_variable__source_terms__source_term_geometry


underscore missing before geometry: ...__source_term__geometry

endJunction · 2017-11-03T11:03:47Z

ProcessLib/SourceTerms/CreateNodalSourceTerm.cpp

+    config.checkConfigParameter("type", "Nodal");
+
+    //! \ogs_file_param{prj__process_variables__process_variable__source_terms__source_term__value}
+    auto const nodal_value = config.getConfigParameter<double>("value");


The <value> tag belongs to the "Nodal" type, so the ogs_file_param must be ..._source_term__Nodal__value. And then you need a new folder Nodal/ in the Documentation/.../source_term/. Compare to DirichletBC.cpp

wenqing

👍

endJunction · 2017-11-06T14:08:26Z

I think you didn't use 'git lfs' properly. Please enable the pre-commit hook.

TomFischer · 2017-11-07T11:23:37Z

If there are no objections I'd like to rebase and merge the branch.

wenqing · 2017-11-07T11:44:44Z

⏩ if you do not wait jenkins test works

chleh · 2017-11-08T10:41:11Z

ProcessLib/SourceTerms/NodalSourceTerm.cpp

+    MeshLib::Location const l{_mesh_id, MeshLib::MeshItemType::Node, _node_id};
+    auto const index =
+        _dof_table.getGlobalIndex(l, _variable_id, _component_id);
+    b.add(index, _value);


Does that mean that such source terms only work for single nodes? Is an extension to line sources/volumetric sources planned?

Yes to both questions.

chleh · 2017-11-08T10:43:08Z

ProcessLib/SourceTerms/SourceTermBuilder.cpp

+        ids.size(), variable_id, *config.component_id);
+
+    return ProcessLib::createNodalSourceTerm(
+        config.config, dof_table, mesh.getID(), ids[0], variable_id,


IMHO it's very bad to silently drop n-1 of n sound node ids.
More complying with our current standards would be to OGS_FATAL if ids.size() > 1. ✅

chleh · 2017-11-08T10:44:44Z

web/content/docs/benchmarks/elliptic/groundwater-flow-dirichlet-nodal-source-term.md

+$$
+G(x, y) = \frac{1}{2 \pi} \ln \sqrt{(x-\xi)^2 + (y-\eta)^2}.
+$$
+With a nodal source term at $(0.0, 0.0)$ the analytical solution is


How strong is the source? Unity? ✅

This avoids a linker error.

The reference files are created with the analytical solution. For the special setup with a dirac source term at position xi, eta the following analytical solution in 2 dimensions is valid: u(x,y) = ln(sqrt((x-xi)^2+(y-eta)^2))/(2 * Pi)

- Theory for analytical solution. - Comparison-images to analytical solution.

wenqing · 2017-11-09T16:29:06Z

ProcessLib/SourceTerms/NodalSourceTerm.cpp

+}
+
+void NodalSourceTerm::integrateNodalSourceTerm(
+    const double t,


@TomFischer t is not used. Would you please send a PR to remove this compilation warning?

ogsbot · 2020-06-19T12:49:03Z

OpenGeoSys development has been moved to GitLab.

See this pull request on GitLab.

TomFischer force-pushed the NodalSourceTerms branch from 03b5507 to 48c5e83 Compare November 2, 2017 13:09

endJunction reviewed Nov 2, 2017

View reviewed changes

endJunction approved these changes Nov 3, 2017

View reviewed changes

wenqing approved these changes Nov 3, 2017

View reviewed changes

TomFischer force-pushed the NodalSourceTerms branch from d2b009f to 2a95151 Compare November 6, 2017 10:28

chleh reviewed Nov 8, 2017

View reviewed changes

TomFischer force-pushed the NodalSourceTerms branch from 2a95151 to c4018dc Compare November 8, 2017 11:36

TomFischer added 11 commits November 8, 2017 12:52

[MGTL] inline createSearchLengthAlgorithm.

3280f04

This avoids a linker error.

[PL/SourceTerms] Impl. of NodalSourceTerm.

4e9ff82

[PL/SourceTerms] Impl. of createNodalSourceTerm.

164dbee

[PL/SourceTerms] Impl. of class SourceTermConfig.

57205b4

[PL/SourceTerms] Impl. of class SourceTermBuilder.

71018e2

[PL/SourceTerms] Put SourceTerms in ProcessVars.

fc36bb5

[PL/SourceTerms] Process vars.: Create source terms.

c3c3bb0

[PL/SourceTerms] Add tests for source term impl.

a90d44a

The reference files are created with the analytical solution. For the special setup with a dirac source term at position xi, eta the following analytical solution in 2 dimensions is valid: u(x,y) = ln(sqrt((x-xi)^2+(y-eta)^2))/(2 * Pi)

[PL/SourceTerms] Documentation.

5258b08

[PL/BoundaryCondition] Fix spelling in docu.

a3e3bc1

[Documentation] Add docu for nodal source terms.

1d621ca

- Theory for analytical solution. - Comparison-images to analytical solution.

TomFischer force-pushed the NodalSourceTerms branch from c4018dc to 1d621ca Compare November 8, 2017 11:53

Update test data.

07743a6

TomFischer merged commit 39a862e into ufz:master Nov 9, 2017

TomFischer deleted the NodalSourceTerms branch November 9, 2017 10:51

bilke added a commit to bilke/ogs that referenced this pull request Nov 9, 2017

[git] Converted benchmark files from PR ufz#1977 to lfs.

eb069e0

wenqing reviewed Nov 9, 2017

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Implementation of nodal source terms #1977

Implementation of nodal source terms #1977

TomFischer commented Nov 2, 2017

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

endJunction left a comment

endJunction Nov 3, 2017

endJunction Nov 3, 2017

endJunction Nov 3, 2017

wenqing left a comment

endJunction commented Nov 6, 2017

TomFischer commented Nov 7, 2017

wenqing commented Nov 7, 2017

chleh Nov 8, 2017

TomFischer Nov 8, 2017

chleh Nov 8, 2017 •

edited by TomFischer

Loading

chleh Nov 8, 2017 •

edited by TomFischer

Loading

wenqing Nov 9, 2017 •

edited

Loading

ogsbot commented Jun 19, 2020

Implementation of nodal source terms #1977

Implementation of nodal source terms #1977

Conversation

TomFischer commented Nov 2, 2017

endJunction Nov 2, 2017 • edited by TomFischer Loading

Choose a reason for hiding this comment

endJunction Nov 2, 2017 • edited by TomFischer Loading

Choose a reason for hiding this comment

endJunction Nov 2, 2017 • edited by TomFischer Loading

Choose a reason for hiding this comment

endJunction left a comment

Choose a reason for hiding this comment

endJunction Nov 3, 2017

Choose a reason for hiding this comment

endJunction Nov 3, 2017

Choose a reason for hiding this comment

endJunction Nov 3, 2017

Choose a reason for hiding this comment

wenqing left a comment

Choose a reason for hiding this comment

endJunction commented Nov 6, 2017

TomFischer commented Nov 7, 2017

wenqing commented Nov 7, 2017

chleh Nov 8, 2017

Choose a reason for hiding this comment

TomFischer Nov 8, 2017

Choose a reason for hiding this comment

chleh Nov 8, 2017 • edited by TomFischer Loading

Choose a reason for hiding this comment

chleh Nov 8, 2017 • edited by TomFischer Loading

Choose a reason for hiding this comment

wenqing Nov 9, 2017 • edited Loading

Choose a reason for hiding this comment

ogsbot commented Jun 19, 2020

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

endJunction Nov 2, 2017 •

edited by TomFischer

Loading

chleh Nov 8, 2017 •

edited by TomFischer

Loading

chleh Nov 8, 2017 •

edited by TomFischer

Loading

wenqing Nov 9, 2017 •

edited

Loading