[fs82]

Dear Foamers,

hopefully, one can help me on a more ore less simple question: Why is the explicit part of the drag in twophaseEulerFoam, multiphaseEulerFoam, reactingMultiphaseEulerFoam and as well in the old bubbleFoam added to the pressure equation, in particular added to phiHbyA and HbyH?

phiHbyAs[phasei] += fvc::interpolate(rAUs[phasei]*K)*MRF.absolute(phase2->phi());

HbyAs[phasei] += rAUs[phasei]*K*phase2->U();

I do understand the reason for dividing the drag into an implicit part and an explicit part. Also, I do understand how the implicit part is added to the UEqns (more precisely via the momentumTransfer function in MomentumTransferPhaseSystem in the new reactingMultiphaseEulerFoam)

forAllConstIter(phasePair, pair, iter)
{
    const volVectorField& U = phase->U();

   *eqns[phase->name()] -= fvm::Sp(K, U);

    Swap(phase, otherPhase);
}

However, to my understanding it would be possible to add another term for the explicit part

K*otherphase->U()

to this function as well. But this is not done. Is there a specific reason for that?

Best regards
Fabian

[ahaley]

Maybe you've already figured this out, but I think I know what the deal is there.

Since K is a sum of the drag coeffs between all the other phases and that phase, you can only include the implicit part there (ie. for phase 1 (Kd12+Kd13+Kd14)*U1). The individual explicit drag terms between each phase pair have to be calculated separately (ie. Kd12*U2+Kd13*U3+Kd14*U4 etc.), which should be done in the Peqn.

[fs82]

I wouldn't agree with that. The pair includes all information about the other phase, so you can add Kd12*U2 piece by piece with a loop over the drag model table. I guess the reason is that one do not want to interpolate the explicit drag within the A matrix. One want to interpolate the drag separately, like the pressure in RhieChow algorithm.

[ahaley]

Yes it is set up so that you could calculate the explicit drag terms there, but that would be incorrect. The explicit part of the drag must be included in H, and therefore pEqn, as it is not a multiplier of the phase velocity in that cell or in any of the neighbours. The colocated grid is not the issue, in this case.

[fs82]

Ok, you are right, this was my mistake. The explicit part of the drag is added to H and not to A.