[MaroneM]

Hello everyone!

I am trying to implement a personalized solver in openFoam in order to solve the transport of chemical species, and I am having some trouble to model a simple "cleaning" sink term on my equations. I based my model on what I would have done for an idealized heat-exchanger (since I am treating this chemical species as a volScalarField an therfore it is analog to the T field), so I am using a scalarSemiImplicitSource defined through the constant/fvOptions file as follows:

myElementSink
{
    type            scalarSemiImplicitSource;
    active          true;

     selectionModecell   Zone;
     cellZone                hx;
     volumeMode          specific;

     injectionRateSuSp
     {
        myElement	(0 -0.4);
     }
}

When I use values for the implicit term that are close to 0 (in the example it is -0.4) I obtain the expected behavior from the simulation. Nevertheless, if I go for strongly negative values (that I would need to model an efficient "cleaning" system) the simulation diverges and returns an increasing positive value (as if this was a source instead of a sink). If anyone has an idea of what might be happening and how to correct it, that would be great!

Just in case, the equation that I am solving with my solver looks like this:

        fvScalarMatrix myEq 
        (
              fvm::ddt(myElement)
            + fvm::div(phi, myElement)
            - fvm::laplacian(diffCoeff, myElement)
            ==
            + myExplicitSource
            + fvOptions(myElement)
        );

        myEq.relax(RELAX_FACTOR);
        fvOptions.constrain(myEq);

        initResidual += myEq.solve().initialResidual();

        fvOptions.correct(myElement);
        myElement.correctBoundaryConditions();

Thank you in advance ^^

[MaroneM]

Ok, so I did not find a "detailed" solution for this but I ended up working with much higher relaxation (even bellow 0.4 in some cases) to get the simulations through.