[Tobi]

Hi all,

yesterday a question arises about the divDevRhoREff() calculation (Calculating divDevReff)

Here, we have the following constellation

 tmp<fvVectorMatrix> laminar::divDevReff(volVectorField& U) const
{
    return
    (
      - fvm::laplacian(nuEff(), U)
      - fvc::div(nuEff()*dev(T(fvc::grad(U))))
    );
}

The question is as follows

• The divDevReff() function returns a fvVectorMatrix
• the fvm::laplacian() function returns a fvMatrix<Type> (so it is a fvVectorMatrix as U is a vector)

What I don´t get right now is the following:

• As the fvMatrix class do not contain an operator- that handles GeometricFields<Type>, the fvc::div() function should return a fvMatrix<Type> in order to substract it from the fvm::laplacian() term
• However, the fvc::grad(U) returns a tmp< GeometricField< typename outerProduct< vector, Type >::type, fvPatchField, volMesh > > (I cannot define what GeometricField it is related to; it should be a tensor as the dev function only handles tensors (matrix) fields)
• The dev() function removes the trace as it is known. Thus the result is a tensor (matrix) too.
• The fvc::div() function should return somehow a tmp< GeometricField< typename innerProduct< vector, Type >::type, fvPatchField, volMesh >> field (this I cannot evaluate right now) but basically it should be a tensor in order to be able to substract it from the fvm::laplacian term

So the question would be as follow: Can anyone give more insight into the fields that are returned by the fvc::grad(U), dev() and fvc::div() functions.


Thanks in advance.
Tobi

[mAlletto]

Hello Tobias,


to answer your questions:


fvc:grad(U) returns the velocity gradient tensor, see https://caefn.com/openfoam/tensor-operations


fvc:div returns the divergence of the tensor and dev is the deviation part of the tensor. how they are calculated is explained in the OpenFOAM programmer guide.



by the way the operator - which takes a fvMatrix and a geometricField as input is defined (see https://www.openfoam.com/documentati...8C_source.html)


what it does, it adds the Geometric field to the source of the matrix and returns this new matrix.


Hope it helps



Michael

[Tobi]

Hi Michael,


thanks for your answer. Most of the things were clear. The only interesting point was that I could not find the implementation of the operator- for a fvMatrix which takes the GeometricField<Type> as giben here: https://cpp.openfoam.org/v7/classFoam_1_1fvMatrix.html


However, if we jump into the source code of the fvMatrix.C file, it is there. I expected the operator- should be visible in the doxygen class too.

 template<class Type>

 Foam::tmp<Foam::fvMatrix<Type>> Foam::operator-

 (

     const fvMatrix<Type>& A,

     const tmp<GeometricField<Type, fvPatchField, volMesh>>& tsu

 )

 {

     checkMethod(A, tsu(), "-");

     tmp<fvMatrix<Type>> tC(new fvMatrix<Type>(A));

     tC.ref().source() += tsu().mesh().V()*tsu().primitiveField();

     tsu.clear();

     return tC;

 }

Thank you for your hint.