[WIP] Preconditioning for multicomponent flows #2426

Cristopher-Morales · 2025-01-18T11:31:19Z

Proposed Changes

This pull request aims to extend the preconditioning for incompressible multicomponent flows, specifically when FLUID_MIXTURE model is used and energy equation is solved. Changes introduced in this pull request aim to solve directly for the sensible enthalpy instead of temperature. Specifically, the energy equation that is going to be solved for multicomponent flows is given as follows:

and the sensible enthalpy is:

Currently $C_{p}$ does not depend on temperature for FLUID_MIXTURE, but it depends on the mixture composition. So the sensible enthalpy simplifies to $C_{p}*(T-T_{0})$. For problems where $C_{p}$ strongly depends on the temperature, Cantera library will be coupled to SU2. After this pull request is completed, a pull request with a fluid model using Cantera will be done where an energy equation for the total enthalpy will be solved.
A pdf document will be attached in the coming days for giving more details about this pull request.

Related Work

Currently fluid mixture does not work correctly when the energy equation is on, this is mainly due to the fact that the energy equation that must be solved for multicomponent flows is the one mentioned above for the sensible enthalpy and not the energy equation for a perfect gas that is currently solved in SU2.

PR Checklist

I am submitting my contribution to the develop branch.
My contribution generates no new compiler warnings (try with --warnlevel=3 when using meson).
My contribution is commented and consistent with SU2 style (https://su2code.github.io/docs_v7/Style-Guide/).
I used the pre-commit hook to prevent dirty commits and used pre-commit run --all to format old commits.
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.

SU2_CFD/src/solvers/CIncEulerSolver.cpp

+          lim_i = 1.0;  // nodes->GetLimiter_Primitive(iPoint, iVar);
+          lim_j = 1.0;  // nodes->GetLimiter_Primitive(jPoint, iVar);


SU2_CFD/src/solvers/CIncNSSolver.cpp

pcarruscag · 2025-01-19T19:30:22Z

SU2_CFD/include/fluid/CFluidModel.hpp

+  /*!
+   * \brief Virtual member.
+   * \param[in] val_enthalpy - Enthalpy value at the point.
+   */
+  virtual void ComputeTempFromEnthalpy(su2double val_enthalpy, su2double* val_temperature,
+                                       const su2double* val_scalars = nullptr) {}


This seems inconsistent with the SetTDState_* functions.
Should this be SetTDState_h and then you use GetTemperature to get the temperature?

In the pull request made by evert regarding preferential diffusion, he also used the same function for the flamelet solver. The only difference is that that function is in the CSpeciesFlameletSolver.cpp and I cannot access to that function when I set the primitives in SetPrimVar in CIncNSVariable.cpp

My point is again to keep an abstract interface.
Evert's function is a private function of a specific solver which is perfectly fine, and it uses the fluid model... If you need a similar function, move it somewhere higher up in the hierarchy to make it accessible. For example the scalar solver. Or even the fluid model but implemented in a way that keeps it useful for all fluids.

pcarruscag · 2025-01-19T19:31:36Z

SU2_CFD/include/fluid/CFluidScalar.hpp

+  /*!
+   * \brief Compute Enthalpy given the temperature and scalars.
+   */
+  su2double ComputeEnthalpyFromT(const su2double val_temperature, const su2double* val_scalars);


Why doesn't this use TDState_T

We compute the enthalpy from the temperature and the species mass fractions. That function is used for setting the enthalpy at the inlets.

And why don't you compute the enthalpy in that function too instead of adding another function?
The idea with generic classes like CFluidModel is to have a concise abstract interface. Having specific functions that only a few subclasses implement defeats the purpose of a generic interface.

pcarruscag · 2025-01-19T19:33:00Z

SU2_CFD/include/solvers/CFVMFlowSolverBase.inl

@@ -434,7 +434,7 @@ void CFVMFlowSolverBase<V, R>::Viscous_Residual_impl(unsigned long iEdge, CGeome
  const bool implicit  = (config->GetKind_TimeIntScheme() == EULER_IMPLICIT);
  const bool tkeNeeded = (config->GetKind_Turb_Model() == TURB_MODEL::SST);

-  CVariable* turbNodes = nullptr;
+  CVariable* turbNodes = nullptr; 


Please setup your IDE to remove trailing spaces.

pcarruscag · 2025-01-19T19:41:42Z

SU2_CFD/src/fluid/CFluidScalar.cpp

@@ -212,13 +212,43 @@ su2double CFluidScalar::ComputeMeanSpecificHeatCp(const su2double* val_scalars)
  return mean_cp;
 }

+su2double CFluidScalar::ComputeEnthalpyFromT(const su2double val_temperature, const su2double* val_scalars){
+  su2double val_Enthalpy = Cp * (val_temperature - 298.15);


Isn't T0 a config option? It is bad practice to hard code constants like this.

https://ansyshelp.ansys.com/public/account/secured?returnurl=////Views/Secured/corp/v242/en/flu_th/flu_th_sec_hxfer_theory.html this is the theory related to the enthalpy equation for multicomponent gases. T_ref is always 298.15 K for this equation

Then make it a constant, not a hard-coded value everywhere.

pcarruscag · 2025-01-19T19:45:40Z

SU2_CFD/src/numerics/CNumerics.cpp

+      val_Proj_Jac_Tensor[3][0] = val_scale * ((*val_temperature) * proj_vel / (*val_betainc2));
+      val_Proj_Jac_Tensor[3][1] = val_scale * ((*val_temperature) * val_normal[0] * (*val_density));
+      val_Proj_Jac_Tensor[3][2] = val_scale * ((*val_temperature) * val_normal[1] * (*val_density));
+      val_Proj_Jac_Tensor[3][3] =
+          val_scale * ((*val_temperature) * (*val_dRhodT) + (*val_density)) * proj_vel;


What is happening here? Is temperature being used for enthalpy?

the energy can be $C_{p}T$ or enthalpy. The enthalpy is used not the temperature.

then some renaming is in order and you need to come up with a better solution to switch between formulations than to have an if statement everytime Cp is used (for example making Cp=1, or encapsulating the computation of enthalpy in a function).

SU2_CFD/src/numerics/flow/convection/fds.cpp

Cristopher-Morales and others added 22 commits October 22, 2024 13:43

solving for enthalpy

ca9083e

fixing flux jacobian

0487a68

Merge branch 'develop' into feature_preconditioning

dd369bb

fix output error

5076f9d

adding enthalpy diffusion

09fe6f4

Merge branch 'develop' into feature_preconditioning

851b13f

moving enthalpy diffusion terms to CIncNSSolver.cpp

f258308

small fix

63a8f37

Merge branch 'develop' into feature_preconditioning

cb1d3ef

cleaning and rewritting some functions

30d0b88

Merge branch 'develop' into feature_preconditioning

dd22fcd

axisymetric source term and corrected flux jacobian at the wall

7f92730

updating BC_ConjugateHeat_Interface for fluid mixture

fdbb9b9

Merge branch 'develop' into feature_preconditioning

c0b10a6

Merge branch 'develop' into feature_preconditioning

911c996

activating muscl, consistent extrapolation for multicomponent

2af10c3

Merge branch 'develop' into feature_preconditioning

d0d0221

correct average enthalpy for multicomponent flows output

a1f526c

Merge branch 'develop' into feature_preconditioning

c469d32

changing boolean names

ef2dc5f

up to date to develop

94bd0b0

Merge branch 'develop' into feature_preconditioning

1c44044

Cristopher-Morales added the changelog:feature label Jan 18, 2025

Cristopher-Morales changed the title ~~[WIP] Preconditioning fro multicomponent flows~~ [WIP] Preconditioning for multicomponent flows Jan 18, 2025

github-advanced-security bot found potential problems Jan 18, 2025

View reviewed changes

Cristopher-Morales added 2 commits January 18, 2025 14:12

adding brackets

912642b

fixing warning

a6ce2d4

pcarruscag requested changes Jan 19, 2025

View reviewed changes

Merge branch 'develop' into feature_preconditioning

e9d1e1f

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] Preconditioning for multicomponent flows #2426

[WIP] Preconditioning for multicomponent flows #2426

Cristopher-Morales commented Jan 18, 2025 •

edited

Loading

pcarruscag Jan 19, 2025

Cristopher-Morales Jan 19, 2025 •

edited

Loading

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

Cristopher-Morales Jan 19, 2025 •

edited

Loading

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

Cristopher-Morales Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

Cristopher-Morales Jan 19, 2025

pcarruscag Jan 19, 2025 •

edited

Loading

		lim_i = 1.0; // nodes->GetLimiter_Primitive(iPoint, iVar);
		lim_j = 1.0; // nodes->GetLimiter_Primitive(jPoint, iVar);

[WIP] Preconditioning for multicomponent flows #2426

Are you sure you want to change the base?

[WIP] Preconditioning for multicomponent flows #2426

Conversation

Cristopher-Morales commented Jan 18, 2025 • edited Loading

Proposed Changes

Related Work

PR Checklist

Choose a reason for hiding this comment

Cristopher-Morales Jan 19, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Cristopher-Morales Jan 19, 2025 • 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

pcarruscag Jan 19, 2025 • edited Loading

Choose a reason for hiding this comment

Cristopher-Morales commented Jan 18, 2025 •

edited

Loading

Cristopher-Morales Jan 19, 2025 •

edited

Loading

Cristopher-Morales Jan 19, 2025 •

edited

Loading

pcarruscag Jan 19, 2025 •

edited

Loading