Tracer with Sharp IB #1185

oguevremont · 2024-07-04T18:02:04Z

Description

Previously, the tracer physics used with the Sharp IB solver could not consider material dependent physical properties: both fluid and solid had the same diffusivity.
Now, physical properties of immersed solids are defined using the solid 0 section, though only one solid can be used at the moment.
A tracer diffusivity model based on the signed distance function used in the Sharp IB has been added, which allows to smooth the diffusivity transition at the fluid/solid interface. The architecture is there for future models and could serve as a base for reactive physics models used in heterogeneous catalysis.

Testing

[tests/core/immersed_solid_diffusivity_tanh] Unit test for the tanh diffusivity transition model.
[applications_tests/lethe-fluid-sharp/tracer_around_sphere] 2D application test where the tracer physics is used with the Sharp IB solver. This also tests the new diffusivity model.

Documentation

[doc/source/parameters/cfd/physical_properties] The new model has been added in new section where immersed solids models will be added in the future.

Miscellaneous (will be removed when merged)

This PR was done with more generality than needed in preparation of the reactive physics that will be implemented soon.

Checklist (will be removed when merged)

See this page for more information about the pull request process.

Code related list:

All in-code documentation related to this PR is up to date (Doxygen format)
Lethe documentation is up to date
New feature has unit test(s) (preferred) or application test(s), and restart files are in the generator folder
The branch is rebased onto master
Changelog (CHANGELOG.md) is up to date
Code is indented with indent-all and .prm files (examples and tests) with prm-indent

Pull request related list:

Labels are applied
There are at least 2 reviewers (or 1 if small feature) excluding the responsible for the merge
If this PR closes an issue or is related to a project, it is linked in the "Projects" or "Development" section
If the fix is temporary, an issue is opened
The PR description is cleaned and ready for merge

blaisb · 2024-07-05T13:20:30Z

doc/source/parameters/cfd/physical_properties.rst

+      set number of solids = 1
+      subsection solid 0
+        set tracer diffusivity model = immersed solid tanh
+        subsection immersed solid
+          subsection tanh levelset
+            set tracer diffusivity inside    = 1
+            set tracer diffusivity outside   = 1
+            set tracer diffusivity thickness = 1
+          end
+        end
+      end


I'm not sure I understand. Why do you have a subsectoin solid and then inside the subsection solid have a subsection immersed solid? Shouldnt the immersed solid be its own subsection?

Furthermore, here you specify the diffusivity inside and outside. But outside of the solid is the fluid. Consequently, are you not duplicated a parameter which is the fluid diffusivity. I think that's not a very wise decision. Essentially, you are duplicating the fluid diffusivity model twice by having the solid also specify it....

Can the diffusivity at the outside edge of the solid be different from the fluid diffusivity?

The way I coded it yes, so we could have a fluid diffusivity, and a solid diffusivity that varies depending on the depth. The wording could be improved though.

I changed it to only a fluid model, that can be based also on levelset (for IB methods). I restructured the parameters as well in this way

blaisb · 2024-07-05T13:24:27Z

include/core/parameters.h

+    enum class TracerDiffusivityModel
+    {
+      constant,
+      tanh_levelset
+    } tracer_diffusivity_model;


That's not really a diffusivity model per say right, it's more like an interface model and we have classes for interface models (e.g. surface tension is an interface model)

include/core/physical_property_model.h

include/solvers/physical_properties_manager.h

include/solvers/tracer_scratch_data.h

tests/core/immersed_solid_diffusivity_tanh.cc

blaisb · 2024-07-05T13:28:42Z

tests/core/immersed_solid_diffusivity_tanh.output

+DEAL:: levelset = 0, diffusivity =  1.00000, dD/dlevelset numerical = 20.0000
+DEAL:: levelset = 100, diffusivity =  2.00000, dD/dlevelset numerical = 0.00000
+DEAL:: levelset = -100, diffusivity =  0.00000, dD/dlevelset numerical = 0.00000
+DEAL::OK


I am curious, in which situation do you think you will need the jacobian of the diffusivity coefficient with respect to the level set function?

It will not be used, I'll remove it

We don't have to test it, but the jacobian is virtual in the PhysicalPropertyModel parent class

legit, no worries then

hepap · 2024-07-05T12:59:50Z

applications_tests/lethe-fluid-sharp/tracer_around_sphere.output

+Tracer statistics : 
+	     Min : 0
+	     Max : 0.7887
+	 Average : 0.01367
+	 Std-Dev : 0.09216


Is it the statistic from the steady it 1? or the initial condition?

They were from the initial condition, good catch. postprocess() was not called in sharp

doc/source/parameters/cfd/physical_properties.rst

hepap · 2024-07-05T13:22:01Z

doc/source/parameters/cfd/physical_properties.rst

+
+* The ``tracer diffusivity outside`` parameter represents the desired diffusivity outside (or at the edge) of the solid. Setting it equal to the fluid diffusivity should provide a smooth transition.
+
+* The ``tracer diffusivity thickness`` parameter represents thickness of the applied tanh function.


Can you add the equation of the tanh function you use?

doc/source/parameters/cfd/physical_properties.rst

hepap · 2024-07-05T13:24:20Z

include/core/parameters.h

+   * @brief ImmersedSolidPhysicalProperties - Defines the parameters for immersed solids
+   * and more specifically how they behave close to their interface.
+   */
+  struct ImmersedSolidPhysicalProperties
+  {
+    TanhLevelsetParameters tanh_levelset_parameters;
+
+    void
+    declare_parameters(ParameterHandler &prm);
+    void
+    parse_parameters(ParameterHandler &prm, const Dimensionality dimensions);
+  };


This is because of the generalization you talked about in the PR description?

Yes that's it

include/solvers/tracer_scratch_data.h

hepap · 2024-07-05T13:58:45Z

doc/source/parameters/cfd/physical_properties.rst

+      set number of solids = 1
+      subsection solid 0
+        set tracer diffusivity model = immersed solid tanh
+        subsection immersed solid
+          subsection tanh levelset
+            set tracer diffusivity inside    = 1
+            set tracer diffusivity outside   = 1
+            set tracer diffusivity thickness = 1
+          end
+        end
+      end


Can the diffusivity at the outside edge of the solid be different from the fluid diffusivity?

source/solvers/gls_navier_stokes.cc

source/solvers/navier_stokes_base.cc

source/solvers/tracer_scratch_data.cc

doc/source/parameters/cfd/physical_properties.rst

blaisb · 2024-07-08T19:23:35Z

@oguevremont is this ready to review or not ? It seems to be in an intermediary state. If so please remove the ready to review label

oguevremont · 2024-07-08T19:27:43Z

@oguevremont is this ready to review or not ? It seems to be in an intermediary state. If so please remove the ready to review label

It is not ready yet, I forgot to remove the ready to review label.

doc/source/parameters/cfd/physical_properties.rst

include/core/tracer_diffusivity_model.h

hepap · 2024-07-09T19:24:34Z

include/core/tracer_diffusivity_model.h

+  void
+  vector_jacobian(const std::map<field, std::vector<double>> &field_vectors,
+                  const field                                 id,
+                  std::vector<double> &jacobian_vector) override
+  {
+    AssertThrow(field_vectors.find(field::levelset) != field_vectors.end(),
+                PhysicialPropertyModelFieldUndefined(
+                  "LevelsetDependentTracerDiffusivity", "levelset"));
+    vector_numerical_jacobian(field_vectors, id, jacobian_vector);
+  };


If the jacobian_vector is always initialized with zeros, this will work. But if it's not, I think you should have the same if condition as the jacobinan() function.

I added the condition to be safe

blaisb

Just a few minor things there and there, but after that this is good to go for me.

doc/source/parameters/cfd/physical_properties.rst

source/solvers/tracer.cc

blaisb · 2024-07-11T12:56:22Z

source/solvers/tracer_scratch_data.cc

@@ -71,15 +80,16 @@ TracerScratchData<dim>::calculate_physical_properties()
          diffusivity_models[0]->vector_value(fields, tracer_diffusivity_0);
          diffusivity_models[1]->vector_value(fields, tracer_diffusivity_1);



Maybe would be a good time to add an assert here instead so that it Throws.

For now we don't support tracer with 2 fluids at all right? So can I just throw an error instead of using an assert?

blaisb · 2024-07-11T12:56:45Z

tests/core/immersed_solid_diffusivity_tanh.output

+DEAL:: levelset = 0, diffusivity =  1.00000, dD/dlevelset numerical = 20.0000
+DEAL:: levelset = 100, diffusivity =  2.00000, dD/dlevelset numerical = 0.00000
+DEAL:: levelset = -100, diffusivity =  0.00000, dD/dlevelset numerical = 0.00000
+DEAL::OK


legit, no worries then

Co-authored-by: hepap <[email protected]>

Co-authored-by: Bruno Blais <[email protected]>

Description Previously, the tracer physics used with the Sharp IB solver could not consider material dependent physical properties: both fluid and solid had the same diffusivity. Now, physical properties of immersed solids are defined using the solid 0 section, though only one solid can be used at the moment. A tracer diffusivity model based on the signed distance function used in the Sharp IB has been added, which allows to smooth the diffusivity transition at the fluid/solid interface. The architecture is there for future models and could serve as a base for reactive physics models used in heterogeneous catalysis. Testing [tests/core/immersed_solid_diffusivity_tanh] Unit test for the tanh diffusivity transition model. [applications_tests/lethe-fluid-sharp/tracer_around_sphere] 2D application test where the tracer physics is used with the Sharp IB solver. This also tests the new diffusivity model. Documentation [doc/source/parameters/cfd/physical_properties] The new model has been added in new section where immersed solids models will be added in the future. Co-authored-by: hepap <[email protected]> Co-authored-by: Bruno Blais <[email protected]> Former-commit-id: e7bfe64

oguevremont added WIP When a PR is open but not ready for review Enhancement New feature or request Ready for review and removed WIP When a PR is open but not ready for review labels Jul 4, 2024

oguevremont requested review from hepap and acdaigneault July 4, 2024 18:22

blaisb reviewed Jul 5, 2024

View reviewed changes

include/core/physical_property_model.h Show resolved Hide resolved

blaisb reviewed Jul 5, 2024

View reviewed changes

include/solvers/physical_properties_manager.h Show resolved Hide resolved

blaisb reviewed Jul 5, 2024

View reviewed changes

include/solvers/tracer_scratch_data.h Outdated Show resolved Hide resolved

blaisb reviewed Jul 5, 2024

View reviewed changes

tests/core/immersed_solid_diffusivity_tanh.cc Show resolved Hide resolved

blaisb reviewed Jul 5, 2024

View reviewed changes

hepap reviewed Jul 5, 2024

View reviewed changes

oguevremont force-pushed the tracer_with_sharp branch 2 times, most recently from 740ad65 to 6b363fe Compare July 8, 2024 19:08

blaisb reviewed Jul 8, 2024

View reviewed changes

doc/source/parameters/cfd/physical_properties.rst Outdated Show resolved Hide resolved

oguevremont removed the Ready for review label Jul 8, 2024

oguevremont added the WIP When a PR is open but not ready for review label Jul 8, 2024

oguevremont force-pushed the tracer_with_sharp branch from 6b363fe to c20f7c0 Compare July 8, 2024 20:02

oguevremont added Ready for review and removed WIP When a PR is open but not ready for review labels Jul 8, 2024

hepap reviewed Jul 9, 2024

View reviewed changes

blaisb requested a review from hepap July 10, 2024 15:26

hepap approved these changes Jul 10, 2024

View reviewed changes

blaisb approved these changes Jul 11, 2024

View reviewed changes

blaisb added Reviewed and ready to merge and removed Ready for review labels Jul 11, 2024

oguevremont and others added 25 commits July 12, 2024 09:11

Added tanh as a diffusivity model that depends on distance

5aba7f2

Made solid diffusivity model responsible for the interface

9fb32e1

Prepare for PR

95b2150

Restructured parameters

0a78487

Added documentation for the new model

fe1c2ff

Added unit test for diffusivity model

ad08f26

Added tracer with sharp application test

c320323

Fixed typo

9898c89

Fixed typo in test

6c2271c

Added to changelog

13fbb5b

Fixed constraints happening when the number of solids > 0

3cc668c

Apply suggestions from code review

709b65e

Co-authored-by: hepap <[email protected]>

Indent

2bc03f0

WIP address review comments

629cfb3

Indent

0eaed55

WIP Address review comments

47d7fa5

Added equation of the tanh function used

c12ac0b

Fix typo

6a811e3

Apply suggestions from code review

67adecf

Co-authored-by: hepap <[email protected]>

Update include/core/tracer_diffusivity_model.h

1fcdd5f

Co-authored-by: hepap <[email protected]>

Address Helene review comments

0fd358b

Set vector to 0 without loop

1085529

Apply suggestions from code review

914fe5f

Co-authored-by: Bruno Blais <[email protected]>

Add throw for 2 fluids with tracer

9f4e118

Update changelog

0f01faa

oguevremont force-pushed the tracer_with_sharp branch from 15b498e to 0f01faa Compare July 12, 2024 13:23

blaisb merged commit e7bfe64 into master Jul 12, 2024
8 checks passed

blaisb deleted the tracer_with_sharp branch July 12, 2024 19:07

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Tracer with Sharp IB #1185

Tracer with Sharp IB #1185

oguevremont commented Jul 4, 2024 •

edited

Loading

blaisb Jul 5, 2024

hepap Jul 5, 2024

oguevremont Jul 5, 2024 •

edited

Loading

oguevremont Jul 8, 2024

blaisb Jul 5, 2024

blaisb Jul 5, 2024

oguevremont Jul 5, 2024

oguevremont Jul 8, 2024

blaisb Jul 11, 2024

hepap Jul 5, 2024

oguevremont Jul 8, 2024

hepap Jul 5, 2024

hepap Jul 5, 2024

oguevremont Jul 5, 2024

hepap Jul 5, 2024

blaisb commented Jul 8, 2024

oguevremont commented Jul 8, 2024

hepap Jul 9, 2024

oguevremont Jul 10, 2024

blaisb left a comment

blaisb Jul 11, 2024

oguevremont Jul 11, 2024

blaisb Jul 11, 2024


		* The ``tracer diffusivity outside`` parameter represents the desired diffusivity outside (or at the edge) of the solid. Setting it equal to the fluid diffusivity should provide a smooth transition.

		* The ``tracer diffusivity thickness`` parameter represents thickness of the applied tanh function.

		@@ -71,15 +80,16 @@ TracerScratchData<dim>::calculate_physical_properties()
		diffusivity_models[0]->vector_value(fields, tracer_diffusivity_0);
		diffusivity_models[1]->vector_value(fields, tracer_diffusivity_1);

Tracer with Sharp IB #1185

Tracer with Sharp IB #1185

Conversation

oguevremont commented Jul 4, 2024 • edited Loading

Description

Testing

Documentation

Miscellaneous (will be removed when merged)

Checklist (will be removed when merged)

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oguevremont Jul 5, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

blaisb commented Jul 8, 2024

oguevremont commented Jul 8, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

blaisb left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oguevremont commented Jul 4, 2024 •

edited

Loading

oguevremont Jul 5, 2024 •

edited

Loading