[GeoMechanicsApplication] More improvements for the nodal extrapolation process #14093

avdg81 · 2025-12-24T13:41:44Z

📝 Description
Fixed some bugs that were found while testing the nodal extrapolation process.

🆕 Changelog

Each nodal extrapolation process now keeps track of which elements are connected to which nodes. Previously, this information was stored at the nodes themselves. This was, however, problematic when multiple nodal extrapolation processes were referencing the same nodes, since the element counters were shared across all processes. Also, we now keep a set of element IDs rather than a simple counter, to avoid incorrect smoothing when elements were repeatedly referenced.
As a result of the above changes, we no longer need to clear the element counter that was stored by the nodes. Member function ExecuteFinalize had become redundant and is now removed.
Replaced a raw for loop by a call to std::inner_product. This change required the introduction of member functions that return the zero value of the variable to be extrapolated.
Added a helper function that creates nodes on the given model part.
Attempted to make the unit tests of the nodal extrapolation process as consistent as possible in the way they are structured.
Removed several redundant retention law parameters from the settlement workflow test.
The settlement workflow test is now more specific about which model parts to consider for the nodal extrapolation processes.
Removed a duplicated model part from the settlement workflow test.
Removed several empty items from the mesh file of the settlement workflow test.
Modified and extended the documentation about the nodal extrapolation process where needed.

- The extrapolation process now keeps its own map from node ID to a set of element IDs rather than an element counter that is stored with the nodes themselves. In this way, we avoid interference of model parts that share nodes and/or multiple extrapolation processes per stage. - Removed member `ExecuteFinalize`, since we no longer need to reset the element counters that were kept by the nodes. - Two nodes were added to a different model part than the intended one. That has been corrected now. - Made the error messages a little bit more descriptive when the unit tests fail. - Use `EXPECT_NEAR` rather than `KRATOS_EXPECT_NEAR`. - In the unit test that uses two extrapolation processes, each process has its own extrapolation variable. - Removed a unit test that covered an implementation detail of the extrapolation process.

…anicsApplication's code base

In this way, we no longer need a loop counter in addition to the loop variable.

Also added some member functions that return the zero value of the given variable (e.g. a scalar, vector, or matrix).

Note that we have to be explicit about the type (`const T`), since using `auto` yields a type with lazy evaluation resulting in test failures. (Or, more accurately, that's what I expect is going on.)

…delPart from .mdpa Used model_part_name_list to reflect active modelparts for extrapolation to nodes. Stripped superfluous retention law parameters from MaterialParameters.json.

…ddition and the counting check if the element is active.

It now says that the second element in the test is inactive. For that test case, we now also use the default absolute tolerance for testing.

One unit test now uses this new utility function.

…or of matrices

…or of vectors

…or of arrays Also resolved a few ambiguous function calls.

Also extracted a function to assert matrices at a collection of nodes.

WPK4FEM

Thank you for the massive clean up of the unit test for the extrapolation process. I have some remarks, but none of them is blocking.

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp

- Added a simple sketch of the model created by `CreateModelPartWithTwoStubElements`. - Renamed a local helper function.

…ts-for-nodal-extrapolation-process

WPK4FEM

Thank you for the updates.

rfaasse

Thank you for these bugfixes, clean-up and reduction of duplications in the unit tests! I have some questions about the behavior and a couple of suggestions, but all in all a very neat PR

...nicsApplication/custom_processes/geo_extrapolate_integration_point_values_to_nodes_process.h

applications/GeoMechanicsApplication/test_setup_utilities/model_setup_utilities.cpp

...cations/GeoMechanicsApplication/tests/test_settlement_workflow/ProjectParameters_stage4.json

...cations/GeoMechanicsApplication/tests/test_settlement_workflow/ProjectParameters_stage3.json

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp

- Added `override`s of member `SetValuesOnIntegrationPoints` of the stub element. - Extracted a helper function that asserts nodal vector values (to reduce duplication).

- Added `override`s of member `SetValuesOnIntegrationPoints` of the stub element. - Extracted a helper function that asserts nodal vector values (to reduce duplication). - Removed a redundant unit test. - Changed the integration point values set on the integration points of the inactive element (to be more confident that inactive elements are not taken into account). - For two unit tests, used slightly more complicated fields to be extrapolated. Also added two comments that explain the expected nodal values. - Updated the documentation of the nodal extrapolation process.

…ts-for-nodal-extrapolation-process

rfaasse

Thanks for processing the review comments! It's good to go from my side

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp

...cations/GeoMechanicsApplication/tests/test_settlement_workflow/ProjectParameters_stage4.json

...cations/GeoMechanicsApplication/tests/test_settlement_workflow/ProjectParameters_stage3.json

...nicsApplication/custom_processes/geo_extrapolate_integration_point_values_to_nodes_process.h

...csApplication/custom_processes/geo_extrapolate_integration_point_values_to_nodes_process.cpp

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp

markelov208

Hi Anne, very nice improvements and, especially, helpers for the unit tests. I have only minor comments. Would you mind to create an issue to generalized the tests for other elements/interfaces?

...csApplication/custom_processes/geo_extrapolate_integration_point_values_to_nodes_process.cpp

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp

- Renamed a private member function to better reflect what it does. - Renamed a local variable to make it more consistent with other variable names.

WPK4FEM

ready for me.

markelov208

Hi Anne, thank you for processing the comments. The PR is ready to be merged

avdg81 and others added 30 commits December 19, 2025 11:36

Added two failing unit tests for the nodal extrapolation process

2d3b469

Removed the final reference to variable NODAL_AREA from the GeoMech…

7e7f172

…anicsApplication's code base

Use a node ID to local index map

c049588

In this way, we no longer need a loop counter in addition to the loop variable.

Renamed a local variable

f208252

Replaced a raw for loop by std::inner_product

22d223c

Also added some member functions that return the zero value of the given variable (e.g. a scalar, vector, or matrix).

Calculate the nodal value using a single statement

a5057cb

Note that we have to be explicit about the type (`const T`), since using `auto` yields a type with lazy evaluation resulting in test failures. (Or, more accurately, that's what I expect is going on.)

Cleaned empty SubModelPart things from .mdpa. Removed duplicate SubMo…

4d99e13

…delPart from .mdpa Used model_part_name_list to reflect active modelparts for extrapolation to nodes. Stripped superfluous retention law parameters from MaterialParameters.json.

Made extrapolation process symmetric in the sense that now both the a…

931b979

…ddition and the counting check if the element is active.

Added unit test for averaging variable with inactive elements.

cb83a29

Added a space to separate two model part names

1b68333

Extended the comment of one of the unit tests

01ff285

It now says that the second element in the test is inactive. For that test case, we now also use the default absolute tolerance for testing.

Use the default absolute tolerance for testing

246e66f

Renamed some variables to comply with the Kratos Style Guide

e2f68ac

Added a utility function that creates nodes on a model part

065a4ce

One unit test now uses this new utility function.

Simplified the test code a bit more

8b3a362

Extracted code to create extrapolation process settings

3b7b71c

Extracted code to build and run the extrapolation process

f64e031

Extracted code to check the nodal values of any double variable

f190e98

Extracted an overload that creates a process settings instance

997ea2e

Hopefully fixed the issue detected by SonarQube

36cd5f4

Renamed several variables to comply with the Kratos Style Guide

7ee53c5

Extracted a helper function to set integration point values

90ee1c4

Several minor stylistic changes

278f397

Only add the required test variable

c80e5d9

Added an overload of SetIntegrationPointValues that receives a vect…

60fc979

…or of matrices

Added an overload of SetIntegrationPointValues that receives a vect…

bbbb482

…or of vectors

Added an overload of SetIntegrationPointValues that receives a vect…

0240475

…or of arrays Also resolved a few ambiguous function calls.

Simplified the test code a bit more

bf6f19a

Extracted a generic function to get nodal values

9b77c3b

Also extracted a function to assert matrices at a collection of nodes.

WPK4FEM previously approved these changes Dec 24, 2025

View reviewed changes

github-project-automation bot moved this from In progress to Reviewer approved in GeoMechanicsApplication (Deltares) Dec 24, 2025

Processed Wijtze Pieter's review suggestions

50f6632

- Added a simple sketch of the model created by `CreateModelPartWithTwoStubElements`. - Renamed a local helper function.

avdg81 dismissed WPK4FEM’s stale review via 50f6632 December 24, 2025 14:59

Merge remote-tracking branch 'origin/master' into geo/more-improvemen…

7029b83

…ts-for-nodal-extrapolation-process

avdg81 added this to Kratos Product Backlog Dec 24, 2025

avdg81 moved this to 👀 In Review in Kratos Product Backlog Dec 24, 2025

WPK4FEM previously approved these changes Dec 24, 2025

View reviewed changes

avdg81 changed the title ~~[GeoMechanicsApplication] More improvements for nodal extrapolation process~~ [GeoMechanicsApplication] More improvements for the nodal extrapolation process Dec 24, 2025

WPK4FEM requested a review from rfaasse January 2, 2026 07:41

rfaasse reviewed Jan 2, 2026

View reviewed changes

avdg81 added 3 commits January 5, 2026 11:13

Processed Richard's review suggestions

8d595c5

- Added `override`s of member `SetValuesOnIntegrationPoints` of the stub element. - Extracted a helper function that asserts nodal vector values (to reduce duplication).

Merge remote-tracking branch 'origin/master' into geo/more-improvemen…

c112e10

…ts-for-nodal-extrapolation-process

avdg81 dismissed WPK4FEM’s stale review via c112e10 January 5, 2026 12:37

avdg81 requested review from WPK4FEM and rfaasse January 5, 2026 12:41

rfaasse previously approved these changes Jan 5, 2026

View reviewed changes

markelov208 reviewed Jan 6, 2026

View reviewed changes

...csApplication/custom_processes/geo_extrapolate_integration_point_values_to_nodes_process.cpp Outdated Show resolved Hide resolved

...pp_tests/custom_processes/test_geo_extrapolate_integration_point_values_to_nodes_process.cpp Outdated Show resolved Hide resolved

avdg81 dismissed rfaasse’s stale review via e4160af January 6, 2026 08:57

Processed Gennady's review suggestions

c564646

- Renamed a private member function to better reflect what it does. - Renamed a local variable to make it more consistent with other variable names.

avdg81 force-pushed the geo/more-improvements-for-nodal-extrapolation-process branch from e4160af to c564646 Compare January 6, 2026 09:26

avdg81 requested a review from markelov208 January 6, 2026 09:26

WPK4FEM approved these changes Jan 6, 2026

View reviewed changes

avdg81 enabled auto-merge (squash) January 6, 2026 09:49

markelov208 approved these changes Jan 6, 2026

View reviewed changes

avdg81 merged commit d82d666 into master Jan 6, 2026
10 checks passed

avdg81 deleted the geo/more-improvements-for-nodal-extrapolation-process branch January 6, 2026 12:00

github-project-automation bot moved this from Reviewer approved to Done in GeoMechanicsApplication (Deltares) Jan 6, 2026

github-project-automation bot moved this from 👀 In Review to ✅ Done in Kratos Product Backlog Jan 6, 2026

[GeoMechanicsApplication] More improvements for the nodal extrapolation process #14093

[GeoMechanicsApplication] More improvements for the nodal extrapolation process #14093

Conversation

avdg81 commented Dec 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

WPK4FEM left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

WPK4FEM left a comment

Choose a reason for hiding this comment

Uh oh!

rfaasse 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!

Uh oh!

rfaasse 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!

Uh oh!

markelov208 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

WPK4FEM left a comment

Choose a reason for hiding this comment

Uh oh!

markelov208 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

avdg81 commented Dec 24, 2025 •

edited

Loading