Skip to content

[WIP] Multizone adjoints for turbomachinery #2446

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 15 commits into
base: develop
Choose a base branch
from
+638 −266
Open

[WIP] Multizone adjoints for turbomachinery #2446

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
15 commits
Select commit Hold shift + click to select a range
98989f7
Fixes from tagging tool
Feb 18, 2025
6082eac
Config catch + fix preacc in ComputeVorticityandStrainMag
Feb 18, 2025
b3267b6
Initial add turbo obj functions
Feb 20, 2025
c459ce9
Initial refactoring of turboperformance into solver for correct recor…
Mar 11, 2025
4a58b8c
Merge remote-tracking branch 'origin/fix_small_AD_issues' into featur…
Mar 21, 2025
1614784
Adding geometry to AD recording for turbo calcs. (NOT WORKING CURRENTLY)
May 1, 2025
d43fef5
Singlezone working with validated gradients and turbo obj funcs. Need…
Jun 9, 2025
6c9de92
Merge branch 'develop' into feature_mz_turbomachinery_adjoint
joshkellyjak Aug 25, 2025
0e03ead
Merge branch 'develop' into feature_mz_turbomachinery_adjoint
joshkellyjak Aug 31, 2025
ae2f001
Messy commit of refactoring
joshkellyjak Oct 6, 2025
32956c4
Cleaned up branch + Euler MZ working / Issues with RANS remain + Init…
joshkellyjak Oct 12, 2025
435bea6
Merge branch 'develop' into feature_mz_turbomachinery_adjoint
joshkellyjak Oct 12, 2025
912aeea
merge issue
joshkellyjak Oct 12, 2025
a0bfb53
Unused var
joshkellyjak Oct 12, 2025
1303145
Unused var
joshkellyjak Oct 12, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 10 additions & 1 deletion Common/src/CConfig.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3413,8 +3413,17 @@ void CConfig::SetPostprocessing(SU2_COMPONENT val_software, unsigned short val_i

/*--- Using default frequency of 250 for all files when steady, and 1 for unsteady. ---*/
for (auto iVolumeFreq = 0; iVolumeFreq < nVolumeOutputFrequencies; iVolumeFreq++){
VolumeOutputFrequencies[iVolumeFreq] = Time_Domain ? 1 : 250;
if (Multizone_Problem && DiscreteAdjoint) {
VolumeOutputFrequencies[iVolumeFreq] = nOuterIter;
}
else {
VolumeOutputFrequencies[iVolumeFreq] = Time_Domain ? 1 : 250;
}
}
} else if (Multizone_Problem && DiscreteAdjoint) {
SU2_MPI::Error(string("OUTPUT_WRT_FREQ cannot be specified for this solver "
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
SU2_MPI::Error(string("OUTPUT_WRT_FREQ cannot be specified for this solver "
SU2_MPI::Error("OUTPUT_WRT_FREQ cannot be specified for this solver "

"(writing of restart and sensitivity files not possible for multizone discrete adjoint during runtime yet).\n"
"Please remove this option from the config file, output files will be written when solver finalizes.\n"), CURRENT_FUNCTION);
Comment on lines +3479 to +3480
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmmm I'm pretty sure it is possible

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is one for @oleburghardt

Copy link
Contributor

@oleburghardt oleburghardt Mar 25, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes and no, writing the files is fine, but the continuation of the multi-zone discrete adjoint solver is erratic. My guess is that some indices aren't cleared properly before re-recording the tape. (Writing adjoint solution variables to file only, without re-recording at all and without sensitivities, might give us some hints.)
The debug mode could eventually pin down where exactly things go wrong.

} else if (nVolumeOutputFrequencies < nVolumeOutputFiles) {
/*--- If there are fewer frequencies than files, repeat the last frequency.
* This is useful to define 1 frequency for the restart file and 1 frequency for all the visualization files. ---*/
Expand Down
3 changes: 2 additions & 1 deletion SU2_CFD/include/solvers/CFVMFlowSolverBase.inl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -672,11 +672,12 @@ void CFVMFlowSolverBase<V, R>::ComputeVorticityAndStrainMag(const CConfig& confi
StrainMag(iPoint) = sqrt(2.0*StrainMag(iPoint));
AD::SetPreaccOut(StrainMag(iPoint));

AD::EndPreacc();

/*--- Max is not differentiable, so we not register them for preacc. ---*/
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What is the difference between ending the preaccumulation before or after?

strainMax = max(strainMax, StrainMag(iPoint));
omegaMax = max(omegaMax, GeometryToolbox::Norm(3, Vorticity));

AD::EndPreacc();
}
END_SU2_OMP_FOR

Expand Down
16 changes: 12 additions & 4 deletions SU2_CFD/src/iteration/CDiscAdjFluidIteration.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -471,16 +471,24 @@ void CDiscAdjFluidIteration::SetDependencies(CSolver***** solver, CGeometry****
solvers0[FLOW_SOL]->InitiateComms(geometry0, config[iZone], MPI_QUANTITIES::SOLUTION);
solvers0[FLOW_SOL]->CompleteComms(geometry0, config[iZone], MPI_QUANTITIES::SOLUTION);

if (config[iZone]->GetBoolTurbomachinery()) {
solvers0[FLOW_SOL]->TurboAverageProcess(solvers0, geometry0, config[iZone], INFLOW);
solvers0[FLOW_SOL]->TurboAverageProcess(solvers0, geometry0, config[iZone], OUTFLOW);
}
if (turbulent && !config[iZone]->GetFrozen_Visc_Disc()) {
solvers0[TURB_SOL]->Postprocessing(geometry0, solvers0,
config[iZone], MESH_0);
solvers0[TURB_SOL]->InitiateComms(geometry0, config[iZone], MPI_QUANTITIES::SOLUTION);
solvers0[TURB_SOL]->CompleteComms(geometry0, config[iZone], MPI_QUANTITIES::SOLUTION);
}
if (config[iZone]->GetBoolTurbomachinery()) {
if (config[iZone]->GetBoolGiles() && config[iZone]->GetSpatialFourier()){
auto conv_bound_numerics = numerics[iZone][iInst][MESH_0][FLOW_SOL][CONV_BOUND_TERM + omp_get_thread_num()*MAX_TERMS];
solvers0[FLOW_SOL]->PreprocessBC_Giles(geometry0, config[iZone], conv_bound_numerics, INFLOW);
solvers0[FLOW_SOL]->PreprocessBC_Giles(geometry0, config[iZone], conv_bound_numerics, OUTFLOW);
}
solvers0[FLOW_SOL]->PreprocessAverage(solvers0, geometry0, config[iZone], INFLOW);
solvers0[FLOW_SOL]->PreprocessAverage(solvers0, geometry0, config[iZone], OUTFLOW);
solvers0[FLOW_SOL]->TurboAverageProcess(solvers0, geometry0, config[iZone], INFLOW);
solvers0[FLOW_SOL]->TurboAverageProcess(solvers0, geometry0, config[iZone], OUTFLOW);
solvers0[FLOW_SOL]->GatherInOutAverageValues(config[iZone], geometry0);
}
if (config[iZone]->GetKind_Species_Model() != SPECIES_MODEL::NONE) {
solvers0[SPECIES_SOL]->Preprocessing(geometry0, solvers0, config[iZone], MESH_0, NO_RK_ITER, RUNTIME_FLOW_SYS, true);
solvers0[SPECIES_SOL]->InitiateComms(geometry0, config[iZone], MPI_QUANTITIES::SOLUTION);
Expand Down
6 changes: 3 additions & 3 deletions SU2_CFD/src/solvers/CDiscAdjSolver.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -219,14 +219,14 @@ void CDiscAdjSolver::RegisterVariables(CGeometry *geometry, CConfig *config, boo

if ((config->GetKind_Regime() == ENUM_REGIME::COMPRESSIBLE) && (KindDirect_Solver == RUNTIME_FLOW_SYS) && config->GetBoolTurbomachinery()){

BPressure = config->GetPressureOut_BC();
Temperature = config->GetTotalTemperatureIn_BC();

if (!reset){
AD::RegisterInput(BPressure);
AD::RegisterInput(Temperature);
}

BPressure = config->GetPressureOut_BC();
Temperature = config->GetTotalTemperatureIn_BC();

Comment on lines -224 to +231
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why are these wrong but not the others that follow the same pattern?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This threw a tagging error, and if I remember correctly it's because in the other sections the variables it accesses are directly used in the solver, whereas in the Giles implementation they are not, creating a mismatch in tags.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we fix it in a way that keeps a clear pattern for doing this type of thing?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can give it a go

config->SetPressureOut_BC(BPressure);
config->SetTotalTemperatureIn_BC(Temperature);
}
Expand Down
Loading