[GaspA]

Hello,


I am trying to convert the code implementing the kOmegaSSTCC (curvature correction) model to Openfoam 11. The original code is intended for OF5.
Original code: https://github.com/ancolli/kOmegaSSTCC/tree/master
line 504 of kOmegaSSTCC/kOmegaSSTCC.C

// Compute rTilda 
    volScalarField D(sqrt(max(symInnerProduct, 0.09*omega_*omega_)));
    tmp<volSymmTensorField> divS =
    (
        fvc::ddt(tSymm())
       +fvc::div
        (
            alphaRhoPhi, tSymm()
        )
    );

The modification is going pretty well except that the fvc::ddt function doesn't seem to work.


The error obtained during compilation is:

 ../momentumTransportModels/lnInclude/kOmegaSSTCCBase.C:506:14: error: ‘ddt’ is not a member of ‘Foam::fvc’; did you mean ‘Foam::fvm::ddt’?
506 | fvc::ddt(tSymm())

The variable tSymm() = symm(fvc::grad(U)) and is of type tmp<volSymmTensorField>
What is surprising is that the function fvc::div works correctly (by compilation). Tested by commenting out the term fvc::ddt(tSymm())



Do you have an answer to this?


Many Thanks in advance.

[Tobermory]

I think I understand why the error occurs - from a cursory glance, there is no fvc::ddt( tmp<volField>), i.e. no ddt function that takes a tmp<volField> as a parameter. There is one that takes a volField, but not a tmp<volField> - check out https://cpp.openfoam.org/v11/fvcDdt_8H_source.html.

The fvc::div works because it has both - check out https://cpp.openfoam.org/v11/fvcDiv_8H_source.html.

[GaspA]

Hello Tobermory


Thank you very much for this answer.
This seems quite logical to me. I did not think to check if a tmp was accepted.

The code is here, line 504:
https://github.com/ancolli/kOmegaSST.../kOmegaSSTCC.C


And here is the problematic part:

 tmp<volTensorField> tgradU = fvc::grad(U);
    tmp<volTensorField> tSkew = skew(tgradU()); // tmp<volTensorField> tSkew = skew(fvc::grad(U)); 
    tmp<volSymmTensorField> tSymm = symm(tgradU());  // tmp<volSymmTensorField> tSymm = symm(fvc::grad(U));
    volScalarField S2(2*magSqr(tSymm()));         //volScalarField S2(2*magSqr(symm(tgradU())));
//
    volScalarField::Internal GbyNu(dev(twoSymm(tgradU()())) && tgradU()());
    volScalarField::Internal G(this->GName(), nut()*GbyNu);
    tgradU.clear();

///////////////// curvature correction

// Compute rStar
    volScalarField symInnerProduct(2.0*tSymm() && tSymm()); 
    volScalarField asymInnerProduct
    (
        max(2.0*tSkew() && tSkew(),
        dimensionedScalar("0", dimensionSet(0, 0, -2, 0, 0), 0.0))
    );
    volScalarField w
    (
        atan(dimensionedScalar("4",dimensionSet(0,0,2,0,0),1.0e-02)*asymInnerProduct)*2.0/(constant::mathematical::pi)*(asymInnerProduct-symInnerProduct)+symInnerProduct
    );
    volScalarField rStar(sqrt(symInnerProduct/max(w, dimensionedScalar("minw", w.dimensions(), SMALL)))); //avoiding dividing by zero    volScalarField rStar(sqrt(symInnerProduct/w));
     
    // Compute rTilda 
    volScalarField D(sqrt(max(symInnerProduct, 0.09*omega_*omega_)));
    tmp<volSymmTensorField> divS =
    (
        fvc::ddt(tSymm())
       +fvc::div
        (
            alphaRhoPhi, tSymm()
        )
    );

Do you have any idea how to get around this problem?
Can we simply declare tSymm() in volSymmTensorField and not in tmp<volSymmTensorField>?


Best regards

[Tobermory]

You could try replacing

fvc::ddt(tSymm())

with

0.5*fvc::ddt( fvc::grad(U) + fvc::grad(U).T() )

and see if that works.

[GaspA]

Hello Tobermory,

Thank you very much for this answer. Unfortunately it does not work. The error is identical.

After performing a search in the files invoked by this one, it appeared that the function fvc::ddt is never used!

This implies that this function is not known in this script, as surprising as it may seem.

Solution:
By adding #include "fvcDdt.H" in the header, everything works correctly with fvc::ddt(tSymm())

Thanks for your help.