[Roie_E]

Hello all,

I want to add to my solver a diffusion equation for the field "C" which settles in a constant velocity V. In order to assign a constant velocity to the field I go by the thread:
How a scalar FIeld can be moved at constant velocity

So in my case:

    fvScalarMatrix CEqn
    (
        fvm::ddt(C) // changed - C
      + fvm::div(phiC, C) // changed - phiC & C
      - fvm::laplacian(kappaC, C) // changed - kappaC & C
     ==
        fvOptions(C) // changed - C
    );

where phiC is defined as

surfaceScalarField phiC
(
    IOobject
    (
        "phiC",
        runTime.timeName(),
        mesh,
        IOobject::READ_IF_PRESENT,
        IOobject::AUTO_WRITE
    ),
    fvc::flux(U+V)
);

and V is another vector field that has constant values all across the domain.

My problem is this:
If I define phiC, as a check, to be exactly like phi (meaning the last line only has fvc::flux(U)) I should get the same results as if I take phi instead of phiC in the diffusion equation itself - but I actually get different results!

Did I miss something on the way?
Thanks in advance,
Roie

[einstein_zee]

Originally Posted by Roie_E

Hello all,

I want to add to my solver a diffusion equation for the field "C" which settles in a constant velocity V. In order to assign a constant velocity to the field I go by the thread:
How a scalar FIeld can be moved at constant velocity

So in my case:

    fvScalarMatrix CEqn
    (
        fvm::ddt(C) // changed - C
      + fvm::div(phiC, C) // changed - phiC & C
      - fvm::laplacian(kappaC, C) // changed - kappaC & C
     ==
        fvOptions(C) // changed - C
    );

where phiC is defined as

surfaceScalarField phiC
(
    IOobject
    (
        "phiC",
        runTime.timeName(),
        mesh,
        IOobject::READ_IF_PRESENT,
        IOobject::AUTO_WRITE
    ),
    fvc::flux(U+V)
);

and V is another vector field that has constant values all across the domain.

My problem is this:
If I define phiC, as a check, to be exactly like phi (meaning the last line only has fvc::flux(U)) I should get the same results as if I take phi instead of phiC in the diffusion equation itself - but I actually get different results!

Did I miss something on the way?
Thanks in advance,
Roie

In principle you are right if the definitions are identical... But depending on your solver phi might get changed along the way... For example if you are using multiphase solvers like interFoam, CorrectPhi will change phi along the way... you may look here "https://github.com/OpenFOAM/OpenFOAM...i/CorrectPhi.C".

[Roie_E]

Thanks a lot Hosein!

So yes, I am using a modified solver built upon buoyantBoussinesqPimpleFoam - Does buoyantBoussinesqPimpleFoam use CorrectPhi? and if so, how can I add phiC to the correction procedure as well?

Thanks again,
Roie