[NBad]

Hi everyone!

I am trying to simulate the high pressure injection of water into a vessel full of molten metal. My end goal would be to implement both phase change and a sonic model to simulate flashing of water when it its injected into the vessel. To do so, I am modyfying compressibleMultiphaseInterFoam (OF2212) to handle the sonic condition and phase change.

To achieve this, since I am no master of C++ I just looked into compressibleInterFoam and built from there.

I run into trouble when trying to implement the sonic condition into compressibleMultiphaseInterFoam. To do so, I added a branch in the code that modifies the pressure equation utilizing the one implemented into compressibleInterFoam.

What I came up with is the following pEqn:

{
    volScalarField rAU("rAU", 1.0/UEqn.A());
    surfaceScalarField rAUf("rAUf", fvc::interpolate(rAU));
    volVectorField HbyA(constrainHbyA(rAU*UEqn.H(), U, p_rgh));
    surfaceScalarField phiHbyA
    (
        "phiHbyA",
        fvc::flux(HbyA)
        + fvc::interpolate(rho*rAU)*fvc::ddtCorr(U, phi)
    //    + fvc::interpolate(rho*rAU)*fvc::ddtCorr(U, Uf)              //
    );

    surfaceScalarField phig
    (
        (
            mixture.surfaceTensionForce()
          - ghf*fvc::snGrad(rho)
        )*rAUf*mesh.magSf()
    );

    phiHbyA += phig;

    // Update the pressure BCs to ensure flux consistency
    constrainPressure(p_rgh, U, phiHbyA, rAUf);

    // Make the fluxes relative to the mesh motion                                      //
    fvc::makeRelative(phiHbyA, U);                                                      //

    PtrList<fvScalarMatrix> p_rghEqnComps(mixture.phases().size());

    label phasei = 0;                                                                   //
    if (pimple.transonic())
    {
        Info<< "Transonic equation construction" << endl;
        phasei = 0; 
        forAllConstIters(mixture.phases(), phase)
        {
            const rhoThermo& thermo = phase().thermo();
            const volScalarField& rho = thermo.rho()();

            surfaceScalarField phid("phid", fvc::interpolate(thermo.psi())*phi);
            surfaceScalarField rhof(fvc::interpolate(rho));  
            surfaceScalarField alphaPhi(phi*fvc::interpolate(phase()));
            
            p_rghEqnComps.set
            (
                phasei,
                (
                    pos(phase())
                    *(
                        (
                            fvc::ddt(phase(), rho) + fvc::div(alphaPhi*rhof)
                        )/rho
                        - fvc::ddt(phase()) - fvc::div(alphaPhi)
                        + (phase()/rho)
                        *correction
                        (
                            thermo.psi()*fvm::ddt(p_rgh)
                            + fvm::div(phid, p_rgh) - fvm::Sp(fvc::div(phid), p_rgh)
                        )
                    )
                ).ptr()
            );

            ++phasei;
        }
    }
    else
    {
        Info<< "Subsonic equation construction" << endl;
        phasei = 0; 
        forAllConstIters(mixture.phases(), phase)
        {
            const rhoThermo& thermo = phase().thermo();
            const volScalarField& rho = thermo.rho()();

            p_rghEqnComps.set
            (
                phasei,
                (
                    fvc::ddt(rho) + thermo.psi()*correction(fvm::ddt(p_rgh))
                    + fvc::div(phi, rho) - fvc::Sp(fvc::div(phi), rho)
                ).ptr()
            );

            ++phasei;
        }
    }
    // Cache p_rgh prior to solve for density update
    volScalarField p_rgh_0(p_rgh);

    while (pimple.correctNonOrthogonal())
    {
        Info<< "Incompressible matrix creation" << endl;
        fvScalarMatrix p_rghEqnIncomp
        (
            fvc::div(phiHbyA)
          - fvm::laplacian(rAUf, p_rgh)
        );

        tmp<fvScalarMatrix> p_rghEqnComp;

        phasei = 0;
        Info<< "Compressible matrix creation" << endl;
        forAllConstIters(mixture.phases(), phase)
        {
            tmp<fvScalarMatrix> hmm
            (
                (max(phase(), scalar(0))/phase().thermo().rho())
               *p_rghEqnComps[phasei]
            );

            if (phasei == 0)
            {
                Info<< "Base matrix" << endl;
                p_rghEqnComp = hmm;
            }
            else
            {
                Info<< "Other phases" << endl;
                p_rghEqnComp.ref() += hmm;
            }

            ++phasei;
        }
        
        Info<< "Solving the equations" << endl;
        solve
        (
            p_rghEqnComp
          + p_rghEqnIncomp,
            mesh.solver(p_rgh.select(pimple.finalInnerIter()))
        );

        Info<< "Final nonortho iteration" << endl;
        if (pimple.finalNonOrthogonalIter())
        {
            phasei = 0;
            for (phaseModel& phase : mixture.phases())
            {
                phase.dgdt() =
                    pos0(phase)
                  *(p_rghEqnComps[phasei] & p_rgh)/phase.thermo().rho();

                ++phasei;
            }

            phi = phiHbyA + p_rghEqnIncomp.flux();

            U = HbyA
              + rAU*fvc::reconstruct((phig + p_rghEqnIncomp.flux())/rAUf);
            U.correctBoundaryConditions();
        }
    }

    {                                                                               //
        Uf = fvc::interpolate(U);                                                   //
        surfaceVectorField n(mesh.Sf()/mesh.magSf());                               //
        Uf += n*(fvc::absolute(phi, U)/mesh.magSf() - (n & Uf));                    //
    }                                                                               //

    p = max(p_rgh + mixture.rho()*gh, pMin);

    // Update densities from change in p_rgh
    mixture.correctRho(p_rgh - p_rgh_0);
    rho = mixture.rho();

    // Correct p_rgh for consistency with p and the updated densities
    p_rgh = p - rho*gh;
    p_rgh.correctBoundaryConditions();

    K = 0.5*magSqr(U);

    Info<< "max(U) " << max(mag(U)).value() << endl;
    Info<< "min(p_rgh) " << min(p_rgh).value() << endl;
}

the structure of the equation is equal to the one implemented in compressibleInterFoam, modified to utilize the field of multiphaseCompressibleInterFoam. The non-sonic branch was left untouched and it works fine.

When executing the sonic branch instead, the solver is able to create the equation set, however, when building the matrices I get the error:

[3] --> FOAM FATAL ERROR: (openfoam-2212 patch=230612)
[3] Incompatible dimensions for operation
    [p_rgh[-1 3 -1 0 0 0 0] ] + [p_rgh[0 0 -1 0 0 0 0] ]
[3] 
[3]     From void Foam::checkMethod(const Foam::fvMatrix<Type>&, const Foam::fvMatrix<Type>&, const char*) [with Type = double]
[3]     in file /usr/lib/openfoam/openfoam2212/src/finiteVolume/lnInclude/fvMatrix.C at line 1915.
[3] 
FOAM parallel run aborting

I am aware that probably the error is caused by the term

- fvc::ddt(phase()) - fvc::div(alphaPhi)

, but after many hours of tests I can't really figure it out.

Any help would be appreciated!

[NBad]

I had just implemented the equation wrong, the right one is:

                    fvc::ddt(rho) + 
                    correction
                    (
                        thermo.psi()*fvm::ddt(p_rgh) 
                        + fvm::div(phid, p_rgh) 
                        - fvc::Sp(fvc::div(phi), rho)
                    )
                    + fvc::div(phi, rho)