[WIP] Implementation of Amplification Factor Transport(AFT) 2019b tranistion model #2422

sun5k · 2025-01-15T13:05:57Z

Proposed Changes

Hi, all.

This PR is for implementing well known subsonic transition model (SA-AFT2019b).

Coder's SA-AFT2019b model is composed of the amplification factor and the logarithmic intermittency transport equation as shown below.

Current State

The primary implementation of the model has been completed and is now in the validation. The validation problems are refer : https://doi.org/10.2514/6.2023-3530 and https://doi.org/10.2514/6.2019-0039.

The transition onset location and the amplification factor field are slightly different. I'm still debugging and checking various things.

Validation Problem of Current State

Firstly, I'm trying the validation problem for 2-D problems (flat plate and NLF-0416 airfoil).
The current results are shown below.

S&K Flat plate from https://doi.org/10.2514/6.2023-3530

NLF-0416 airfoil from https://doi.org/10.2514/6.2023-3530

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 --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.

Primary back up

Test for SAAFT2019b

Renew_Min_Max

HL

Common/include/option_structure.hpp

SU2_CFD/include/numerics/turbulent/transition/trans_sources.hpp

SU2_CFD/src/output/CFlowOutput.cpp

@@ -1407,6 +1451,25 @@
      AddVolumeOutput("TURB_INDEX", "Turb_index", "PRIMITIVE", "Turbulence index");
      break;

+    case TURB_TRANS_MODEL::AFT:
+      //nodes -> SetAFT_Wonder_Func(iPoint, HL, H12, dNdRet, Ret0, D_H12, l_H12, m_H12, kv, Rev, Rev0, F_growth, F_crit, PAF, Pg, Dg);


SU2_CFD/src/solvers/CTransAFTSolver.cpp

AFT2017b

SU2_CFD/include/numerics/turbulent/transition/trans_sources.hpp

Test_Strain

pcarruscag

Thanks 👍 Looks good but we can re-use a lot of code from the LM solver and see the question about the aux var gradient which looks suspicious.

pcarruscag · 2025-01-19T22:47:43Z

Common/include/option_structure.hpp

+  NONE,         /*!< \brief No option / default. */
+  AFT2017b,       /*!< \brief using AFT2017b model. */
+  AFT2019b       /*!< \brief using AFT2019b model. */


Please align the start of the comments and add some links to reference papers.

pcarruscag · 2025-01-19T22:52:42Z

Common/include/option_structure.hpp

+ * \param[in] rank - MPI rank.
+ * \return Struct with AFT options.
+ */
+inline AFT_ParsedOptions ParseAFTOptions(const AFT_OPTIONS *AFT_Options, unsigned short nAFT_Options, int rank) {


Rank is not used, can you remove it from this function and maybe LMOptions if it is also not used there.

pcarruscag · 2025-01-19T22:55:22Z

Common/src/CConfig.cpp

+              switch (Kind_Turb_Model) {
+                case TURB_MODEL::NONE: SU2_MPI::Error("No turbulence model has been selected but AFT transition model is active.", CURRENT_FUNCTION); break;
+                case TURB_MODEL::SST: SU2_MPI::Error("k-w SST turbulence model has been selected but AFT transition model is active.", CURRENT_FUNCTION); break;
+              }


This block of code appears twice, please refactor this a little to avoid duplicating it.

pcarruscag · 2025-01-19T22:58:02Z

SU2_CFD/include/numerics/turbulent/transition/trans_correlations.hpp

+      case AFT_CORRELATION::AFT2017b: {
+        dNdRet = 0.028*(H12 - 1.0) - 0.0345 * exp(-pow(3.87/(H12 -1.0) - 2.52, 2.0));
+        break;
+      }
+
+      case AFT_CORRELATION::AFT2019b: {
+        dNdRet = 0.028*(H12 - 1.0) - 0.0345 * exp(-pow(3.87/(H12 -1.0) - 2.52, 2.0));
+        break;
+      }


See if there are other opportunities to avoid repeating formulas.

Suggested change

case AFT_CORRELATION::AFT2017b: {

dNdRet = 0.028*(H12 - 1.0) - 0.0345 * exp(-pow(3.87/(H12 -1.0) - 2.52, 2.0));

break;

}

case AFT_CORRELATION::AFT2019b: {

dNdRet = 0.028*(H12 - 1.0) - 0.0345 * exp(-pow(3.87/(H12 -1.0) - 2.52, 2.0));

break;

}

case AFT_CORRELATION::AFT2017b:

case AFT_CORRELATION::AFT2019b: {

dNdRet = 0.028*(H12 - 1.0) - 0.0345 * exp(-pow(3.87/(H12 -1.0) - 2.52, 2.0));

break;

}

pcarruscag · 2025-01-19T23:00:47Z

SU2_CFD/include/numerics/turbulent/transition/trans_diffusion.hpp

+  void FinishResidualCalc(const CConfig* config) override {
+    const bool implicit = config->GetKind_TimeIntScheme() == EULER_IMPLICIT;
+
+    /*--- Compute mean effective dynamic viscosity ---*/
+    const su2double diff_i_AF = Laminar_Viscosity_i + Eddy_Viscosity_i;
+    const su2double diff_j_AF = Laminar_Viscosity_j + Eddy_Viscosity_j;
+    const su2double diff_i_Gamma = Laminar_Viscosity_i + Eddy_Viscosity_i;
+    const su2double diff_j_Gamma = Laminar_Viscosity_j + Eddy_Viscosity_j;
+
+    const su2double diff_AF = 0.5*(diff_i_AF + diff_j_AF);
+    const su2double diff_Gamma = 0.5*(diff_i_Gamma + diff_j_Gamma);
+
+    Flux[0] = diff_AF*Proj_Mean_GradScalarVar[0];
+    Flux[1] = diff_Gamma*Proj_Mean_GradScalarVar[1];
+
+    /*--- For Jacobians -> Use of TSL (Thin Shear Layer) approx. to compute derivatives of the gradients ---*/
+    if (implicit) {
+      const su2double proj_on_rho_i = proj_vector_ij/Density_i;
+      Jacobian_i[0][0] = -diff_AF*proj_on_rho_i;  Jacobian_i[0][1] = 0.0;
+      Jacobian_i[1][0] = 0.0;                        Jacobian_i[1][1] = -diff_Gamma*proj_on_rho_i;
+
+      const su2double proj_on_rho_j = proj_vector_ij/Density_j;
+      Jacobian_j[0][0] = diff_AF*proj_on_rho_j;   Jacobian_j[0][1] = 0.0;
+      Jacobian_j[1][0] = 0.0;                     Jacobian_j[1][1] = diff_Gamma*proj_on_rho_j;
+    }
+  }


This only differs from LM by some constant factors, refactor it to use the same class.

pcarruscag · 2025-01-19T23:09:37Z

SU2_CFD/src/solvers/CTransAFTSolver.cpp

You need to find a way to share the common parts between the LM solver and this one, this way there is too much duplication.

pcarruscag · 2025-01-19T23:13:52Z

SU2_CFD/src/solvers/CTransAFTSolver.cpp

+  switch (config->GetKind_Gradient_Method()) {
+    case GREEN_GAUSS:
+      SetAuxVar_Gradient_GG(geometry, config);
+      break;
+    case WEIGHTED_LEAST_SQUARES:
+      SetAuxVar_Gradient_LS(geometry, config);
+    default:
+      break;
+    }


Why are you computing the AuxVarGradient after you use it above instead of before?

pcarruscag · 2025-01-19T23:16:26Z

SU2_CFD/src/solvers/CTransAFTSolver.cpp

+    HLGradTerm = nodes->GetAuxVarGradient(iPoint, 1, 0) * nodes->GetAuxVarGradient(iPoint, 0, 0) + nodes->GetAuxVarGradient(iPoint, 1, 1) * nodes->GetAuxVarGradient(iPoint, 0, 1);
+
+    if(nDim == 3) {
+      HLGradTerm += nodes->GetAuxVarGradient(iPoint, 1, 2) * nodes->GetAuxVarGradient(iPoint, 0, 2);
+    }


GeometryToolbox SquaredNorm

pcarruscag · 2025-01-19T23:17:29Z

SU2_CFD/src/solvers/CTransAFTSolver.cpp

+    su2double Temp3 = flowNodes->GetVelocity(iPoint, 0) * nodes->GetAuxVarGradient(iPoint, 0, 0);
+    su2double Temp4 = flowNodes->GetVelocity(iPoint, 1) * nodes->GetAuxVarGradient(iPoint, 0, 1);


Add some documentation please, if the initial aux var grad is 0 then you set the aux var to 0, then its gradient is 0 and this remains 0 forever? Looks like a bug.

pcarruscag · 2025-01-19T23:18:50Z

SU2_CFD/src/solvers/CTransAFTSolver.cpp

+
+    /*-- destruction term of Intermeittency(Gamma) --*/
+    const su2double Dg = c_2 * Density_i * VorticityMag * F_turb * (c_3 * exp(lnIntermittency) - 1.0);
+    nodes -> SetAFT_Wonder_Func(iPoint, HL, H12, dNdRet, Ret0, D_H12, l_H12, m_H12, kv, Rev, Rev0, F_growth, F_crit, PAF, Pg, Dg);


This is a lot of memory and these variables are only used for post processing (output) right?
If so these can be computed in the output class when setting the output values.
That saves doing this work every iteration and storing all these variables.

sun5k and others added 5 commits January 6, 2025 23:50

Primary SA_AFT_2019b push

e652a53

Primary back up

Merge branch 'su2code:master' into feature_AFT_2019_Model

e1e12ca

Test_SA_AFT2019b

058ee6e

Test for SAAFT2019b

Renew_Min_Max

448f5b9

Renew_Min_Max

HL update

a583d7c

HL

sun5k added the changelog:feature label Jan 15, 2025

github-advanced-security bot found potential problems Jan 15, 2025

View reviewed changes

Updat_AFT2017b

5349d97

AFT2017b

github-advanced-security bot found potential problems Jan 16, 2025

View reviewed changes

SU2_CFD/include/numerics/turbulent/transition/trans_sources.hpp Fixed Show fixed Hide fixed

SA_AFT2019b

0de4e65

Test_Strain

pcarruscag reviewed Jan 19, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] Implementation of Amplification Factor Transport(AFT) 2019b tranistion model #2422

[WIP] Implementation of Amplification Factor Transport(AFT) 2019b tranistion model #2422

sun5k commented Jan 15, 2025

pcarruscag left a comment

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

pcarruscag Jan 19, 2025

		su2double Temp3 = flowNodes->GetVelocity(iPoint, 0) * nodes->GetAuxVarGradient(iPoint, 0, 0);
		su2double Temp4 = flowNodes->GetVelocity(iPoint, 1) * nodes->GetAuxVarGradient(iPoint, 0, 1);

[WIP] Implementation of Amplification Factor Transport(AFT) 2019b tranistion model #2422

Are you sure you want to change the base?

[WIP] Implementation of Amplification Factor Transport(AFT) 2019b tranistion model #2422

Conversation

sun5k commented Jan 15, 2025

Proposed Changes

Current State

Validation Problem of Current State

PR Checklist

pcarruscag 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

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