[WIP] Composition dependent fluid properties for multicomponent flow #1332

mheimgartner · 2021-07-23T09:54:59Z

Proposed Changes

Hi all,

This draft PR contains all work on composition dependent fluid properties for multicomponent flow.

What I did so far:

• Implemented new CFluidScalar class (which inherits from CFluidModel) and can be accessed via a new fluid model: MIXTURE_FLUID_MODEL.
• Verified the current implementation by comparing with Fluent.
• Added regression test case for two component incomp_rans flow with variable fluid properties.

Future work:

• Fix variable specific heat capacity and convergence issues in case of axisymmetric flow.
• Add possibility to have more than a single transported scalar (i.e. mixture containing more than two components).
• Remove flamelet implementation from feature_multicomp in order to have this work self-contained. Note: Currently this is a PR into feature_flamelet to make the diff easier

In case there are comments, ideas, foreseeable stepping stones from experience please let me know.

Related Work

#1223 by @danielmayer @bigfooted which feature_multicomp currently branches from.
#1330 by @TobiKattmann which cuts the scalar transport necessary for this work out to merge it separately into develop

PR Checklist

Put an X by all that apply. You can fill this out after submitting the PR. If you have any questions, don't hesitate to ask! We want to help. These are a guide for you to know what the reviewers will be looking for in your contribution.

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.

…NSTANT_CONDUCTIVITY

…t calls mixing laws function every iteration.

…put argument.

…IVE_SCALAR.

…rties.

TobiKattmann · 2021-07-23T10:27:01Z

@mheimgartner first of all congrats to this openend PR

For the builds in the tests, compiler warnings are treated like errors. I.e. the code compiles and works on your side but the test-builds won't if certain warnings are present (like unused variables).
For more information where to look, see the Details button next to the individual builds. Click and wait a moment ... that will automatically scroll down for you.
The failing builds will be for the same code bits probably, so looking at BaseNoMPI for the start should suffice.

To look in the Details and change code from there might be tedious so you can see the warnings locally by setting ./meson.py <your build folder> ... --warnlevel=3 .... Work through the warnings until there are none left -> commit the changes and push. That will probably calm the compiler-warnings-are-errors gods and grant you access to the regression test

SU2_CFD/include/fluid/CUnityLewisDiffusivity.hpp

Common/src/CConfig.cpp

SU2_CFD/src/fluid/CFluidModel.cpp

WallyMaier · 2021-07-23T19:41:07Z

Hi @mheimgartner, thanks for the PR. I think this is a good use of the fluidmodel classes. There seems to be a lot of potential to generalize some aspects of Mixture_Fluid_Model and CNEMOGas. Im not sure what the best plan to approach this may be, but I think its something worth discussing.

bigfooted · 2021-07-30T07:30:19Z

Hi @mheimgartner, thanks for the PR. I think this is a good use of the fluidmodel classes. There seems to be a lot of potential to generalize some aspects of Mixture_Fluid_Model and CNEMOGas. Im not sure what the best plan to approach this may be, but I think its something worth discussing.

@WallyMaier I think there might be some overlap in fluid properties like mixing models as well, but I do not know what information you have available in a library like mutation++. I assume that it has general fluid property calculations and mixing rules available that could be used (wilke rule, kinetic models, etc)?

WallyMaier · 2021-07-30T16:07:54Z

Hi @mheimgartner, thanks for the PR. I think this is a good use of the fluidmodel classes. There seems to be a lot of potential to generalize some aspects of Mixture_Fluid_Model and CNEMOGas. Im not sure what the best plan to approach this may be, but I think its something worth discussing.

@WallyMaier I think there might be some overlap in fluid properties like mixing models as well, but I do not know what information you have available in a library like mutation++. I assume that it has general fluid property calculations and mixing rules available that could be used (wilke rule, kinetic models, etc)?

@bigfooted Thats correct! Both the CNEMOGas fluid models (SU2TCLib and CMutation++) have thermodynamic properties and transport properties. I imagine there would be a smart way to reduce overlap, though Im not sure what the best approach would be.

pcarruscag

Some comments below to keep in mind when you continue the implementation.

Common/include/CConfig.hpp

Common/src/CConfig.cpp

pcarruscag · 2021-08-01T15:46:46Z

SU2_CFD/include/fluid/CFluidScalar.hpp

+  unique_ptr<CViscosityModel> LaminarViscosityPointers[100];  
+  unique_ptr<CConductivityModel> ThermalConductivityPointers[100]; 


either std::vector or std::array. If you go for a fixed size define it as a constexpr member variable.

Yes I hardcoded this because I don't know how to set the size corresponding to the number of species in the mixture.
Right now, I changed it to e.g.:
std::vector<std::unique_ptr> LaminarViscosityPointers;

and kept void CFluidScalar::SetLaminarViscosityModel(const CConfig* config) {} in CFluidScalar.cpp the same. Is this the correct way?

SU2_CFD/include/fluid/CFluidScalar.hpp

pcarruscag · 2021-08-01T15:49:00Z

SU2_CFD/include/fluid/CFluidScalar.hpp

+
+#include "CFluidModel.hpp"
+
+class CFluidScalar : public CFluidModel {


missing override and or final specifiers in this class.

Thank you Pedro for all your comments, really appreciate the help! Question regarding the CFluidScalar class which I made final, is it better to use override or final for its members, e.g.:

void SetLaminarViscosityModel(const CConfig* config) override;
vs
void SetLaminarViscosityModel(const CConfig* config) final;

If a member function is declared without virtual and with final, it means that it already overrides a virtual function in base class, so I figure it might be preferred to use final ? Also, do I need to state final after every function member in CFluidScalar?

SU2_CFD/src/fluid/CFluidScalar.cpp

pcarruscag · 2021-08-01T15:57:46Z

SU2_CFD/src/fluid/CFluidScalar.cpp

+  wilkeNumerator.clear(); 
+  su2double viscosityMixture = 0.0;
+
+  massToMoleFractions(val_scalars); 


This is wrong, too many hidden things happening, make the method const and the interface as explicit as possible.
x,y go in, z comes out.
It should not be necessary to read massToMoleFractions to understand wilkeViscosity

Right now I only moved massToMoleFractions to SetTDState_T. I did not make the method const because laminarViscosity is altered by wilkeViscosity. I placed laminarViscosity within the function because the laminarViscosity value may change at every grid point, e.g. when using Sutherland law in case of non-isothermal flow. Is there a better approach to do this?

SU2_CFD/src/solvers/CIncEulerSolver.cpp

TestCases/incomp_rans/multicomponentflow_variablefluidproperties/c6_cms_venturi_conference.cfg

…icomp Conflicts: Common/include/CConfig.hpp Common/src/CConfig.cpp SU2_CFD/src/drivers/CDriver.cpp SU2_CFD/src/fluid/CFluidFlamelet.cpp SU2_CFD/src/numerics/CLookUpTable.cpp SU2_CFD/src/output/CFlowOutput.cpp SU2_CFD/src/output/output_structure_legacy.cpp SU2_CFD/src/solvers/CDiscAdjSolver.cpp SU2_CFD/src/solvers/CFlameletSolver.cpp SU2_CFD/src/solvers/CIncEulerSolver.cpp SU2_CFD/src/solvers/CPassiveScalarSolver.cpp SU2_CFD/src/solvers/CScalarSolver.cpp

…icomp Conflicts: .github/workflows/regression.yml

SU2_CFD/src/solvers/CIncEulerSolver.cpp

TobiKattmann · 2021-08-12T23:54:50Z

I hope to have fixed all regression testcases over at feature_flamelet so please merge and see whether we can get some ✔️'s here as well

…ulticomp

Most likely introduced by a63cb3b of #1332. A jacobian contribution was changed in order of computation but not its expression. This might have changed some floating point digits.

TestCases/multicomp_regression.py

…ulticomp

… feature_multicomp

…ulticomp

Had to add Linsol and Linsol_turb screen output to all of em.

This reverts commit 0561b77.

…ulticomp

See discussion 1403. Has to be implemented in a separate PR.

This reverts commit 33f9591.

…d not be multiplied by density.

…viour in volume output.

…to feature_multicomp

TobiKattmann · 2021-10-21T06:59:41Z

The SST-2003 changes are in 33f9591 and were reverted directly afterwards in 8208c9e such that we have a commit with the necessary changes but they are not applied in the latest version of this code. This was done to revisit and finish this at a later stage.

TobiKattmann · 2021-10-21T07:00:53Z

SU2_CFD/include/numerics/flamelet/scalarLegacy_sources.hpp

            Residual[iVar] += yinv*Volume*(Diffusion_Coeff_i[iVar]+Mass_Diffusivity_Tur)*ScalarVar_Grad_i[iVar][1];
          else
-            Residual[iVar] += yinv*Volume*Density_i*(Diffusion_Coeff_i[iVar]+Mass_Diffusivity_Tur)*ScalarVar_Grad_i[iVar][1];
+            Residual[iVar] += yinv*Volume*(Density_i*Diffusion_Coeff_i[iVar]+Mass_Diffusivity_Tur)*ScalarVar_Grad_i[iVar][1];


This little bugfix in the axissymmetric source term let the corresponding reg test fail (which was to be expected). reg test-vals need to be adapted

…ulticomp

1. Consistent number of species vals defined per option. This setup will see quite a bit more use with the multicomponent flow in #1332. 2. Boundness in [0,1] where applicable

stale · 2022-03-02T09:00:44Z

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If this is still a relevant issue please comment on it to restart the discussion. Thank you for your contributions.

mheimgartner added 12 commits July 22, 2021 16:26

Added possibility to add multiple inputs for fluid properties e.g. CO…

38c76a7

…NSTANT_CONDUCTIVITY

Added new CFluidScalar class which contains function SetTDState_T tha…

dd1b15d

…t calls mixing laws function every iteration.

Added new fluid model: MIXTURE_FLUID_MODEL.

6ee1122

Added unity Lewis diffusivity model.

66df854

Added new diffusion model option: UNITY_LEWIS.

d8c7d61

Already existing diffusivity models now comply with new therm_cond in…

1d7594d

…put argument.

Added therm_cond, spec_heat_cap, etc. to volume ouput in case of PASS…

76577fd

…IVE_SCALAR.

Changed residual trying to fix convergence in case of axisymmetric BC.

cac551b

Implemented SetTDState_T function which calls CFluidScalar class.

b84513d

Added new test case for multicomponent flow with variable fluid prope…

3d76d05

…rties.

Change regression config name multiple species.

fa07ad8

Regression pipeline triggered by PR to feature_flamelet.

da3083b

pr-triage bot added the PR: draft label Jul 23, 2021

TobiKattmann reviewed Jul 23, 2021

View reviewed changes

SU2_CFD/include/fluid/CUnityLewisDiffusivity.hpp Outdated Show resolved Hide resolved

TobiKattmann reviewed Jul 23, 2021

View reviewed changes

Common/src/CConfig.cpp Outdated Show resolved Hide resolved

TobiKattmann reviewed Jul 23, 2021

View reviewed changes

SU2_CFD/src/fluid/CFluidModel.cpp Outdated Show resolved Hide resolved

pcarruscag requested changes Aug 1, 2021

View reviewed changes

mheimgartner added 4 commits August 3, 2021 16:09

Fix build errors.

4c04417

Added multicomp_regression.py which only contains multicomp test case.

b62dc18

Fixed most comments on draft PR.

5227878

mheimgartner mentioned this pull request Aug 12, 2021

Multicomponent flow with variable fluid properties (feature_multicomp) su2code/TestCases#71

Closed

mheimgartner added 2 commits August 12, 2021 14:29

Changed default values to be static const.

e58c006

Merge remote-tracking branch 'su2/feature_flamelet' into feature_mult…

4721547

…icomp Conflicts: .github/workflows/regression.yml

TobiKattmann reviewed Aug 12, 2021

View reviewed changes

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

TobiKattmann added 5 commits September 16, 2021 09:39

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

8edb730

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

a63cb3b

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

def916d

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

6fac4d2

…ulticomp

Update axisymmetric reg vals. Very minor changes. Prob just FP-stuff.

9d439ce

Most likely introduced by a63cb3b of #1332. A jacobian contribution was changed in order of computation but not its expression. This might have changed some floating point digits.

TobiKattmann reviewed Sep 23, 2021

View reviewed changes

TestCases/multicomp_regression.py Outdated Show resolved Hide resolved

TobiKattmann and others added 2 commits September 23, 2021 13:25

Update TestCases/multicomp_regression.py

6c3a6fd

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

1b4a5a1

…ulticomp

TobiKattmann mentioned this pull request Oct 2, 2021

Species transport solver (passive for now) #1388

Merged

5 tasks

TobiKattmann and others added 14 commits October 2, 2021 13:18

Merge branch 'feature_flamelet', remote-tracking branch 'origin' into…

d94f5ae

… feature_multicomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

6dd9abf

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

217bd70

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

e490792

…ulticomp

That rans_naca0012 case is used for quite some reg tests!

0561b77

Had to add Linsol and Linsol_turb screen output to all of em.

Revert "That rans_naca0012 case is used for quite some reg tests!"

722d762

This reverts commit 0561b77.

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

7505426

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

46b1495

…ulticomp

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

47b53ef

…ulticomp

Change current SST-1994 to SST-2003. Directly reverted.

33f9591

See discussion 1403. Has to be implemented in a separate PR.

Revert "Change current SST-1994 to SST-2003. Directly reverted."

8208c9e

This reverts commit 33f9591.

Changed axisymmetric diffusion source term. Turb diffusion part shoul…

54642a1

…d not be multiplied by density.

Currently Cp from CFluidModel cannot vary much. Hardcoded mixing beha…

4af10ed

…viour in volume output.

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

8284550

…to feature_multicomp

TobiKattmann reviewed Oct 21, 2021

View reviewed changes

Merge remote-tracking branch 'origin/feature_flamelet' into feature_m…

cbf436f

…ulticomp

stale bot added the stale label Mar 2, 2022

stale bot closed this Apr 16, 2022

		unique_ptr<CViscosityModel> LaminarViscosityPointers[100];
		unique_ptr<CConductivityModel> ThermalConductivityPointers[100];


		#include "CFluidModel.hpp"

		class CFluidScalar : public CFluidModel {

[WIP] Composition dependent fluid properties for multicomponent flow #1332

[WIP] Composition dependent fluid properties for multicomponent flow #1332

Uh oh!

Conversation

mheimgartner commented Jul 23, 2021 • edited by TobiKattmann Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Proposed Changes

Related Work

PR Checklist

Uh oh!

TobiKattmann commented Jul 23, 2021

Uh oh!

Uh oh!

Uh oh!

Uh oh!

WallyMaier commented Jul 23, 2021

Uh oh!

bigfooted commented Jul 30, 2021

Uh oh!

WallyMaier commented Jul 30, 2021

Uh oh!

pcarruscag left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

pcarruscag Aug 1, 2021

Choose a reason for hiding this comment

Uh oh!

mheimgartner Aug 5, 2021

Choose a reason for hiding this comment

Uh oh!

Uh oh!

pcarruscag Aug 1, 2021

Choose a reason for hiding this comment

Uh oh!

mheimgartner Aug 5, 2021

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

pcarruscag Aug 1, 2021

Choose a reason for hiding this comment

Uh oh!

mheimgartner Aug 6, 2021

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

TobiKattmann commented Aug 12, 2021

Uh oh!

Uh oh!

TobiKattmann commented Oct 21, 2021

Uh oh!

TobiKattmann Oct 21, 2021

Choose a reason for hiding this comment

Uh oh!

stale bot commented Mar 2, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

mheimgartner commented Jul 23, 2021 •

edited by TobiKattmann

Loading