Integrated Heatflux for flow problems #1530

TobiKattmann · 2022-01-31T09:56:43Z

Hi,

Possibility to give the Heatflux over a boundary in [W] instead of [W/m^2]. This was already possible in the HeatSolver but now also in Inc and Comp flow solvers.
Prescribing a fixed Watt is especially useful in optimization scenarios with Temperature OF where otherwise the optimizer can exploit that a smaller boundary almost certainly yields smaller Temperatures.

The necessary SurfaceAreas are computed and stored in CGeometry instances which is computed once at program start by a call from the Driver.

I also added NEMO for this feature.

There is currently no Testcase which uses INTEGRATED_HEATFLUX= YES so one might want to change that 😬 🤫

For the HeatSolver this was originally introduced at some point by @oleburghardt

PR Checklist

I am submitting my contribution to the develop branch.
My contribution generates no new compiler warnings (try with the '-Wall -Wextra -Wno-unused-parameter -Wno-empty-body' compiler flags, or simply --warnlevel=2 when using meson).
My contribution is commented and consistent with SU2 style.
I have added a test case that demonstrates my contribution, if necessary.
I have updated appropriate documentation (Tutorials, Docs Page, config_template.cpp) , if necessary.

Therefore a SurfaceArea routine was added to CGeometry which is called once in the GeoPreproc of the Driver. This central Surface area comp makes the HF change straight forward and might be useful in some other cases. The ordering of the SurfaceArea vector of CGeometry follows the Global cfg ordering.

… centrally.

…computed centrally." This reverts commit 6452710.

Common/include/geometry/CGeometry.hpp

pcarruscag · 2022-01-31T10:11:05Z

Common/include/geometry/CGeometry.hpp

+   * \brief Compute the surface area of all global markers.
+   * \param[in] config - Definition of the particular problem.
+   */
+  void ComputeSurfaceArea(const CConfig *config);


Call it in CGeometry::UpdateGeometry or the sensitivities will be incorrect.
Might also be needed for the OF in multizone adjoints?

added the call in updateGeometry in 288a2eb . UpdateGeometry is only ever called for deforming meshes. When I moved the ComputeSurfaceArea routine only to UpdateGeometry my case segfaulted because ComputeSurfaceArea was never called and the vector never resized and filled. The call to ComputeSurfaceArea coming from the CDriver ctor (via Geometrical_Preprocessing) is always evaluated at startup (except for DG-FEM cases)

I am unsure where you think this would go wrong for multizone adjoints 🤔

Yes I said to call it, not to move it 😉
Just thinking of the places where Set_Heatflux_Areas was used.

I wanted to tackle Set_Heatflux_Areas at another point anyway as it is somewhat connected to Heat_Fluxes which computes OF's and hist-output for the HeatSolver. And in there I dont like 2 things:

On isothermal walls, the cfg T is used for HF computation. The Isothermal wall is weakly enforced so this could create quite a different value compared to a purely postprocessed (eg via paraview) result

AvgTemp is only reported for Heatflux-walls (not even for CHT-interfaces). Imo it should be users choice where he wants which quantities even if they are trivial

I reverted the cut-down of Set_Heatflux_Areas here e198acf because I would need to care about above points and that might be better done in a separate PR

SU2_CFD/src/solvers/CIncNSSolver.cpp

Co-authored-by: Pedro Gomes <[email protected]>

…e 'CfgFile' Co-authored-by: Pedro Gomes <[email protected]>

Common/include/geometry/CGeometry.hpp

Co-authored-by: Pedro Gomes <[email protected]>

Common/src/geometry/CGeometry.cpp

SU2_CFD/src/solvers/CNSSolver.cpp

…into feature_integratedHF

makes comparing cases a little less annoying

SU2_CFD/src/solvers/CNEMONSSolver.cpp

Used a 3D cht case and applied it in the solid zone. Residual stay exactly the same. This, to my experience is only the case because the surface is a plane rectangle. For e.g. a round shape the small mismatch in exact circle circumference and the discrete appriximation by the mesh is enough to alter the residuals (and final temperature) slighthly. The Residual change in this case for the Temperature is due to the error that was in the mesh (see Testcases #91)

TestCases/parallel_regression.py

TobiKattmann · 2022-01-31T15:24:46Z

Thanks already @pcarruscag for the quick suggestions 👍 Let me know if there are more :)

2 linked PR's go with this:

The testases PR needs to be merged before this to make sure the reg tests works (Edit: Done and reg tests work (except hybrid_par_AD...))
The website update (...has more time ;)

Otherwise I am happy so far with this PR

Common/include/geometry/CGeometry.hpp

Common/src/geometry/CGeometry.cpp

TestCases/incomp_navierstokes/streamwise_periodic/chtPinArray_3d/configSolid.cfg

TobiKattmann · 2022-02-02T15:21:39Z

Thanks for the reviewing @pcarruscag 💐 I am merging this in then. And then afterwards the website PR su2code/su2code.github.io#69

This is due to the previously merged PR #1530 where an NDIME error in the mesh was fixed (where the reg test values already changed) and as this PR makes the same case a restarted one... the values have to be adapted again

TobiKattmann added 2 commits January 31, 2022 10:37

Little cleanup in Heatsolvers heatflux bc

965f154

pr-triage bot added the PR: unreviewed label Jan 31, 2022

TobiKattmann added the changelog:feature label Jan 31, 2022

TobiKattmann added 2 commits January 31, 2022 10:57

Remove now superfluous SurfaceArea computation since that is computed…

6452710

… centrally.

Revert "Remove now superfluous SurfaceArea computation since that is …

e198acf

…computed centrally." This reverts commit 6452710.

pcarruscag reviewed Jan 31, 2022

View reviewed changes

TobiKattmann and others added 4 commits January 31, 2022 11:22

Move HF adaption to centralized CSolver and add NEMO HF-walls.

853f5d6

Change SurfaceArea to SurfaceAreaCfgFile

6f91db0

Co-authored-by: Pedro Gomes <[email protected]>

Move central Integrated HF comp to CGeometry from CSolver

716f445

Co-authored-by: Pedro Gomes <[email protected]>

Also call SurfArea comp in UdateGeometry and adapt function ti includ…

288a2eb

…e 'CfgFile' Co-authored-by: Pedro Gomes <[email protected]>

pcarruscag reviewed Jan 31, 2022

View reviewed changes

Common/include/geometry/CGeometry.hpp Outdated Show resolved Hide resolved

TobiKattmann and others added 2 commits January 31, 2022 12:22

Change UpdateIntegratedHF function to GetSurfaceArea.

8794f14

Co-authored-by: Pedro Gomes <[email protected]>

Add error catch in case string matching fails.

06a432e

pcarruscag reviewed Jan 31, 2022

View reviewed changes

Common/src/geometry/CGeometry.cpp Outdated Show resolved Hide resolved

SU2_CFD/src/solvers/CNSSolver.cpp Outdated Show resolved Hide resolved

TobiKattmann and others added 3 commits January 31, 2022 13:22

Apply clang-format in limited scopes.

a57f4a3

Update comment

f951a07

Merge branch 'feature_integratedHF' of https://github.com/su2code/su2 …

a4a3fef

…into feature_integratedHF

TobiKattmann mentioned this pull request Jan 31, 2022

Change NDIME in mesh to correct 3 instead of 2. su2code/TestCases#91

Merged

Screen output only written by MASTER_NODE

084727b

makes comparing cases a little less annoying

TobiKattmann commented Jan 31, 2022

View reviewed changes

SU2_CFD/src/solvers/CNEMONSSolver.cpp Show resolved Hide resolved

TobiKattmann commented Jan 31, 2022

View reviewed changes

TestCases/parallel_regression.py Show resolved Hide resolved

TobiKattmann mentioned this pull request Jan 31, 2022

Add Integrated Heatflux to Users Guide. su2code/su2code.github.io#69

Merged

TobiKattmann commented Jan 31, 2022

View reviewed changes

Apply suggestions from code review

6546dfb

TobiKattmann mentioned this pull request Feb 2, 2022

Specify C standard #1532

Merged

5 tasks

pcarruscag approved these changes Feb 2, 2022

View reviewed changes

pr-triage bot removed the PR: unreviewed label Feb 2, 2022

pr-triage bot added the PR: reviewed-approved label Feb 2, 2022

TobiKattmann merged commit 1502831 into develop Feb 2, 2022

TobiKattmann deleted the feature_integratedHF branch February 2, 2022 15:22

pr-triage bot added PR: merged and removed PR: reviewed-approved labels Feb 2, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Integrated Heatflux for flow problems #1530

Integrated Heatflux for flow problems #1530

TobiKattmann commented Jan 31, 2022 •

edited

Loading

pcarruscag Jan 31, 2022

TobiKattmann Jan 31, 2022

TobiKattmann Jan 31, 2022

pcarruscag Jan 31, 2022

TobiKattmann Jan 31, 2022

TobiKattmann commented Jan 31, 2022 •

edited

Loading

TobiKattmann commented Feb 2, 2022

Integrated Heatflux for flow problems #1530

Integrated Heatflux for flow problems #1530

Conversation

TobiKattmann commented Jan 31, 2022 • edited Loading

PR Checklist

pcarruscag Jan 31, 2022

Choose a reason for hiding this comment

TobiKattmann Jan 31, 2022

Choose a reason for hiding this comment

TobiKattmann Jan 31, 2022

Choose a reason for hiding this comment

pcarruscag Jan 31, 2022

Choose a reason for hiding this comment

TobiKattmann Jan 31, 2022

Choose a reason for hiding this comment

TobiKattmann commented Jan 31, 2022 • edited Loading

TobiKattmann commented Feb 2, 2022

TobiKattmann commented Jan 31, 2022 •

edited

Loading

TobiKattmann commented Jan 31, 2022 •

edited

Loading