[styleworker]

Hello foamers,

I'm playing around with MRconjugateHeatFoam, which is a conjugate heat transfer solver. I was able to modify this solver in a way, such that it can handle two phase flow. Unfortunately after some time steps (depends on the memory), I'll get errors like this:

--> FOAM FATAL ERROR: 
bad size -1317939304

    From function List<T>::List(const label size)
    in file /opt/openfoam220/src/OpenFOAM/lnInclude/List.C at line 49.

FOAM aborting

#0  Foam::error::printStack(Foam::Ostream&) at /opt/openfoam220/src/OSspecific/POSIX/printStack.C:221
#1  Foam::error::abort() at /opt/openfoam220/src/OpenFOAM/lnInclude/error.C:249
#2  Foam::Ostream& Foam::operator<< <Foam::error>(Foam::Ostream&, Foam::errorManip<Foam::error>) at /opt/openfoam220/src/OpenFOAM/lnInclude/errorManip.H:85
#3  Foam::List<double>::List(int) at /opt/openfoam220/src/OpenFOAM/lnInclude/List.C:54
#4  Foam::Field<double>::Field(int) at /opt/openfoam220/src/OpenFOAM/lnInclude/Field.C:50
#5  Foam::tmp<Foam::Field<double> > Foam::mag<double>(Foam::UList<double> const&) at /opt/openfoam220/src/OpenFOAM/lnInclude/FieldFunctions.C:180
#6  
 at ~/OpenFOAM/of-2.2.0/applications/solvers/CHTinterDyMFOAM/coupledFvPatchFields/regionCoupleHeatFlux/regionCoupleHeatFluxFvPatchScalarField.C:222
#7  
 at ~/OpenFOAM/of-2.2.0/applications/solvers/CHTinterDyMFOAM/convergenceCheck.H:112
#8  __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6"
#9  
 in "/home/willy/OpenFOAM/of-2.2.0/platforms/linux64GccDPDebug/bin/CHTinterDyMFoam"

After checking the error message, I had taken a look at regionCoupleHeatFlux/regionCoupleHeatFluxFvPatchScalarField.C at line 222:

const Foam::scalar&
Foam::regionCoupleHeatFluxFvPatchScalarField::maxResidual() const
{
    // Calculate interpolated patch flux
    const polyPatch& ownPatch = patch().patch();
    const polyPatch& nbrPatch = coupleManager_.neighbourPatch().patch();
    const fvPatchField<scalar>& Tnbr = coupleManager_.neighbourPatchField<scalar>();

    patchToPatchInterpolation interpolator(nbrPatch, ownPatch);

    scalarField qNbr2own =
        interpolator.faceInterpolate
        (
            refCast<const regionCoupleTemperatureFvPatchScalarField>
            (Tnbr).flux()
        );

    // Lookup transport properties
    const dictionary& transportProperties =
        db().lookupObject<IOdictionary>("transportProperties");

    dictionary phaseSolid(transportProperties.subDict("phaseSolid"));

    dimensionedScalar kS(phaseSolid.lookup("k"));

    // Calculate the maximum normalized residual
    const fvPatchScalarField& Town = *this; //Temperature of own face (Tsolid, if regionCoupleTemperature is set in solid bc)
    const scalarField& qOwn = kS.value()*Town.snGrad();
    const scalar& residual =
        gMax
        (
            mag(mag(qOwn) - mag(qNbr2own))/
            max(min(gMax(mag(qOwn)),gMax(mag(qNbr2own))), SMALL) //line 222
        );

    return residual;
}

The function flux() is defined as:

//- Return the patch flux
Foam::tmp<Foam::scalarField>
Foam::regionCoupleTemperatureFvPatchScalarField::flux() const
{
    //  Set up access to the thermaly conductivity field and mesh
    const volScalarField& k12 = db().lookupObject<volScalarField>("k12"); //Wird hier auch von runTime gelesen und nicht von dem gespeicherten Zeitordner???
    //const fvMesh& mesh = patch().boundaryMesh().mesh();
        
    //Get ID of local patch in T boundary file (fluid side)
    label patchID = patch().boundaryMesh().findPatchID(patch().name());

    //Get the k-boundaryField with same patch ID (fluid side)
    const fvPatchScalarField& kF = k12.boundaryField()[patchID];

    const fvPatchScalarField& T = *this; //Temperature of own face (Tfluid, if regionCoupleTemperature is set in fluid bc)

    return kF*T.snGrad(); //Fourier's law.
}

It looks like there is some problem with working memory. Maybe it could be due to bad use of pointers? I have to admit, I'm not really familiar with C++ programming and still learning this language.

The relevant source files are attached.

[styleworker]

I could't solve the problem directly, but I was able to run the case without errors, when I'm using a static mesh instead of dynamicFvMesh.

[emateo84]

Hi! could you give me please more information concerning the file createFields.H i,m also dealing with a similar problem MR solver + interFoam trying to couple them.