Custom Python wrapper marker BC values appropriately nondimensionalized

[j-signorelli]

Proposed Changes

In current release of SU2, having, for example, a MARKER_ISOTHERMAL=(marker, 300) and MARKER_PYTHON_CUSTOM=(marker) causes the BC temperature to be set to a nondimensional value of 300, which is inconsistent with what that marker's value is supposed to mean. This can be seen being set here:

SU2/Common/src/geometry/CGeometry.cpp

Lines 2441 to 2461 in b80e4ce

	/*--- Initialize quantities for customized boundary conditions.
	* Custom values are initialized with the default values specified in the config (avoiding non physical values) ---*/
	CustomBoundaryTemperature = new su2double*[nMarker];
	CustomBoundaryHeatFlux = new su2double*[nMarker];
	for(iMarker=0; iMarker < nMarker; iMarker++){
	Marker_Tag = config->GetMarker_All_TagBound(iMarker);
	CustomBoundaryHeatFlux[iMarker] = nullptr;
	CustomBoundaryTemperature[iMarker] = nullptr;
	if(config->GetMarker_All_PyCustom(iMarker)){
	switch(config->GetMarker_All_KindBC(iMarker)){
	case HEAT_FLUX:
	CustomBoundaryHeatFlux[iMarker] = new su2double[nVertex[iMarker]];
	for(iVertex=0; iVertex < nVertex[iMarker]; iVertex++){
	CustomBoundaryHeatFlux[iMarker][iVertex] = config->GetWall_HeatFlux(Marker_Tag);
	}
	break;
	case ISOTHERMAL:
	CustomBoundaryTemperature[iMarker] = new su2double[nVertex[iMarker]];
	for(iVertex=0; iVertex < nVertex[iMarker]; iVertex++){
	CustomBoundaryTemperature[iMarker][iVertex] = config->GetIsothermal_Temperature(Marker_Tag);

and called for example here:

SU2/SU2_CFD/src/solvers/CNSSolver.cpp

Lines 712 to 714 in b80e4ce

	else if (config->GetMarker_All_PyCustom(val_marker)) {
	Twall = geometry->GetCustomBoundaryTemperature(val_marker, iVertex);
	}

This PR nondimensionalizes any custom boundary temperatures or heat fluxes appropriately so that any setting of them does not need to be nondimensionalized manually before, both in MARKER_ISOTHERMAL and MARKER_HEATFLUX and in any Python code setting values of them.

Related Work

None that I am aware of.

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 have added a test case that demonstrates my contribution, if necessary.
• I have updated appropriate documentation (Tutorials, Docs Page, config_template.cpp), if necessary.

[pcarruscag]

Thanks, can you look for other places where the values are used, I think the incompressible solver might need the same fix.

[j-signorelli]

A grep -R CustomBoundaryTemperature . and grep -R CustomBoundaryHeatFlux . doesn't seem to show anything done for the incompressible solver, and also doesn't appear to be an option here at least:

SU2/SU2_CFD/src/solvers/CIncNSSolver.cpp

Lines 350 to 372 in db21a33

	/*--- Get the specified wall heat flux, temperature or heat transfer coefficient from config ---*/
	su2double Wall_HeatFlux = 0.0, Twall = 0.0, Tinfinity = 0.0, Transfer_Coefficient = 0.0;
	switch (kind_boundary) {
	case HEAT_FLUX:
	Wall_HeatFlux = config->GetWall_HeatFlux(Marker_Tag) / config->GetHeat_Flux_Ref();
	if (config->GetIntegrated_HeatFlux()) {
	Wall_HeatFlux /= geometry->GetSurfaceArea(config, val_marker);
	}
	break;
	case ISOTHERMAL:
	Twall = config->GetIsothermal_Temperature(Marker_Tag) / config->GetTemperature_Ref();
	break;
	case HEAT_TRANSFER:
	Transfer_Coefficient = config->GetWall_HeatTransfer_Coefficient(Marker_Tag) * config->GetTemperature_Ref() /
	config->GetHeat_Flux_Ref();
	Tinfinity = config->GetWall_HeatTransfer_Temperature(Marker_Tag) / config->GetTemperature_Ref();
	break;
	default:
	SU2_MPI::Error("Unknown type of boundary condition.", CURRENT_FUNCTION);
	break;
	}

However, I might have missed a spot here:

SU2/Common/src/geometry/CMultiGridGeometry.cpp

Lines 960 to 994 in db21a33

	void CMultiGridGeometry::SetMultiGridWallHeatFlux(const CGeometry* fine_grid, unsigned short val_marker) {
	struct {
	const CGeometry* fine_grid;
	unsigned short marker;
	su2double* target;
	su2double Get(unsigned long iVertex) const { return fine_grid->GetCustomBoundaryHeatFlux(marker, iVertex); }
	void Set(unsigned long iVertex, const su2double& val) const { target[iVertex] = val; }
	} wall_heat_flux;
	wall_heat_flux.fine_grid = fine_grid;
	wall_heat_flux.marker = val_marker;
	wall_heat_flux.target = CustomBoundaryHeatFlux[val_marker];
	SetMultiGridWallQuantity(fine_grid, val_marker, wall_heat_flux);
	}
	void CMultiGridGeometry::SetMultiGridWallTemperature(const CGeometry* fine_grid, unsigned short val_marker) {
	struct {
	const CGeometry* fine_grid;
	unsigned short marker;
	su2double* target;
	su2double Get(unsigned long iVertex) const { return fine_grid->GetCustomBoundaryTemperature(marker, iVertex); }
	void Set(unsigned long iVertex, const su2double& val) const { target[iVertex] = val; }
	} wall_temperature;
	wall_temperature.fine_grid = fine_grid;
	wall_temperature.marker = val_marker;
	wall_temperature.target = CustomBoundaryTemperature[val_marker];
	SetMultiGridWallQuantity(fine_grid, val_marker, wall_temperature);
	}

but my guess is that it's fine since values from here are re-nondimensionalized from the above commits. Please let me know if not however - I am not familiar with the specifics of multigrid and its implementation.

Note that the commits above presume that values specified in CustomBoundaryTemperature and CustomBoundaryHeatFlux in CGeometry are dimensional, regardless if SU2 is in nondimensional mode. It might make more sense to have these be nondimensional and just ensure that incoming default values from a MARKER_ISOTHERMAL and MARKER_HEATFLUX are nondimensionalized appropriately, but in my opinion it is more straightforward to have users set dimensional temperatures and heat fluxes. However, then it would also make sense to return dimensional heat fluxes and temperatures as well for Python custom walls which I am not sure is currently how it is done.

Apologies for the long explanation, but curious to hear your input on this.

I will look into why one of the regression tests failed and get back to you.

[pcarruscag]

You are right, custom temperature is not implemented in the incompressible solver.
@bigfooted the inlet profiles are always dimensional now right?

[pcarruscag]

You can relax the tolerance in the python script of the failing test @j-signorelli

[bigfooted]

You are right, custom temperature is not implemented in the incompressible solver. @bigfooted the inlet profiles are always dimensional now right?

Yes, the inlet profiles are dimensional.