[jammy4536]

Hi all, I'm trying to simulate a liquid jet in gaseous crossflow, using compressibleInterFoam, since the flow speeds are reasonably high. I've been wrestling with it in various forms for quite some time, and just run out of options that I think might work.
The geometry is set up so that I have a large box for a distant freestream, and a short length of pipe that the jet comes out of (I've attached some images of the mesh and pipe with fluid). The mesh is created using Pointwise, joined into a single domain, and has been renumbered. Gas enters from the left (inlet) and the top is the atmosphere, and right face is the outlet. This example is rotated, just to see if that would have any effect.
My issue is that for some reason the fluid will come up the pipe a little bit, but then gets stuck at that point. I've done some simulations before where this hasn't happened, but I have no idea what's changed between them to inspire this issue. Any advice on usual suspects would be appreciated, as I'm at a loss about what to try!



U.orig:

internalField   uniform (0 -40.62 0);

boundaryField
{
    atmosphere
    {
        type            zeroGradient;
    }
    bottom
    {
        type            slip;
    }
    inlet
    {
        type            fixedValue;
        value           uniform (0 -40.62 0);
    }
    outlet
    {
        type            zeroGradient;
    }
    pipe
    {
        type            fixedValue;
        value           uniform (10.633 0 0);
    }
    pipe_wall
    {
        type            slip;
    }
    BaseAndTop
    {
        type            empty;
    }
}

p_rgh

internalField   uniform 98800;

boundaryField
{
    atmosphere
    {
        type            freestreamPressure;
        freestreamValue uniform 98800;
        value           uniform 98800;
    }
    bottom
    {
        type            fixedFluxPressure;
        gradient        uniform 0;
        value           uniform 98800;
    }
    inlet
    {
        type            fixedFluxPressure;
        gradient        uniform 0;
        value           uniform 98800;
    }
    outlet
    {
        type            freestreamPressure;
        freestreamValue uniform 98800;
        value           uniform 98800;
    }
    pipe
    {
        type            fixedFluxPressure;
        gradient        uniform 0;
        value           uniform 98800;
    }
    pipe_wall
    {
        type            fixedFluxPressure;
        gradient        uniform 0;
        value           uniform 98800;
    }
    BaseAndTop
    {
        type            empty;
    }
}

p

internalField   uniform 98800;

boundaryField
{
    atmosphere
    {
        type            calculated;
        value           uniform 98800;
    }
    bottom
    {
        type            calculated;
        value           uniform 98800;
    }
    inlet
    {
        type            calculated;
        value           uniform 98800;
    }
    outlet
    {
        type            calculated;
        value           uniform 98800;
    }
    pipe
    {
        type            calculated;
        value           uniform 98800;
    }
    pipe_wall
    {
        type            calculated;
        value           uniform 98800;
    }
    BaseAndTop
    {
        type            empty;
    }
}

alpha.water

internalField   uniform 0;

boundaryField
{
    atmosphere
    {
        type            inletOutlet;
        inletValue      uniform 0;
        value           uniform 0;
    }
    bottom
    {
        type            zeroGradient;
    }
    inlet
    {
        type            fixedValue;
        value           uniform 0;
    }
    outlet
    {
        type            inletOutlet;
        inletValue      uniform 0;
        value           uniform 0;
    }
    pipe
    {
        type            fixedValue;
        value           uniform 1;
    }
    pipe_wall
    {
        type            zeroGradient;
    }
    BaseAndTop
    {
        type            empty;
    }
}

T

internalField   uniform 298;

boundaryField
{
    atmosphere
    {
        type            inletOutlet;
        inletValue      uniform 298;
        value           uniform 298;
    }
    bottom
    {
        type            zeroGradient;
    }
    inlet
    {
        type            inletOutlet;
        inletValue      uniform 298;
        value           uniform 298;
    }
    outlet
    {
        type            inletOutlet;
        inletValue      uniform 298;
        value           uniform 298;
    }
    pipe
    {
        type            inletOutlet;
        inletValue      uniform 298;
        value           uniform 298;
    }
    pipe_wall
    {
        type            zeroGradient;
    }
    BaseAndTop
    {
        type            empty;
    }
}

fvSchemes

ddtSchemes
{
    default         Euler;
}

gradSchemes
{
    default         Gauss linear 1;
    grad(U)         cellLimited Gauss linear 1;
}

divSchemes
{
    div(rhoPhi,U)  Gauss limitedLinearV 1;
    div(phi,alpha)  Gauss vanLeer;
    div(phirb,alpha) Gauss interfaceCompression;

    energy          Gauss upwind;

    div(phi,p)     Gauss upwind;
    div(phid1,p_rgh)     Gauss upwind;
    div(phid2,p_rgh)     Gauss upwind;

    div(rhoPhi,K)   $energy;
    div(rhoPhi,T)  $energy;

    div(((rho*nuEff)*dev2(T(grad(U))))) Gauss linear;
}

laplacianSchemes
{
    default         Gauss linear orthogonal;
}

interpolationSchemes
{
    default         linear;
}

snGradSchemes
{
    default         orthogonal;
}

wallDist
{
    method meshWave;
}

fvSolution

solvers
{
    "alpha.water.*"
    {
        nAlphaCorr      1;
        nAlphaSubCycles 4;
        cAlpha          1;

        MULESCorr       yes;
        nLimiterIter    5;

        solver          smoothSolver;
        smoother        symGaussSeidel;
        tolerance       1e-5;
        relTol          0;


    }

    ".*(rho|rhoFinal)"
    {
        solver          diagonal;
    }

    p_rgh
    {
        solver         GAMG;
        tolerance       1e-06;
        relTol          0;
        smoother        DILU;
        // smoother        DIC;
        nPreSweeps      0;
        nPostSweeps     1;
        nFinestSweeps   1;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 700;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

    p_rghFinal
    {
        solver         GAMG;
        tolerance       1e-06;
        relTol          0;
        smoother        DILU;
        // smoother        DIC;
        nPreSweeps      0;
        nPostSweeps     1;
        nFinestSweeps   1;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 700;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

    "pcorr.*"
    {
        solver         GAMG;
        tolerance       1e-06;
        relTol          0;
        smoother        DILU;
        // smoother        DIC;
        nPreSweeps      0;
        nPostSweeps     2;
        nFinestSweeps   2;
        cacheAgglomeration true;
        nCellsInCoarsestLevel 700;
        agglomerator    faceAreaPair;
        mergeLevels     1;
    }

    "(U|T|K|UFinal|TFinal|KFinal).*"
    {
        solver          PBiCG;
        preconditioner  DILU;
        tolerance       1e-08;
        relTol          0;
    }
}

PIMPLE
{
    transonic yes;
    momentumPredictor   yes;
    nOuterCorrectors    1;
    nCorrectors         3;
    nNonOrthogonalCorrectors 0;
}

relaxationFactors
{
    equations
    {
        ".*" 1;
    }
}

[geth03]

how many inlets/outlets do you have?
can it be possible that you overdefine your velocities across your domain?
you have a velocity for the internal domain and also for boundaries.

[jammy4536]

Thanks for replying!

I have one air inlet on the far left, and one liquid inlet at the base of the pipe, and then one outlet for the domain on the far right, and atmosphere on the top so closer to inletOutlet.
I set the velocity in the domain just to set a starting state and avoid transience, since a real case would already have a steady air state before starting the jet.

[geth03]

did you check if signs are set correctly?
i don't really know if you overdefine your velocity field, so there is nothing to move.
i once had the problem that i didn't correctly adjust my signs so there was nothing exiting the domain and my fluid stuck just like yours.

[jammy4536]

I have certainly checked the signs of the velocity fields, as that would be a rather mundane mistake, but a classic.
I'll have a play with different outlet conditions later in the week, and report back. I did have it working previously, and the outlet/atmosphere conditions may be one of the only things that is somewhat different. I changed them since I was getting either velocity or pressure artefacts at the outlet and atmosphere boundaries, but could have broken the simulation in the process.

[jammy4536]

Had a chance to look at this again, and while I have changed the settings in the 0 folder, I think the real culprit is the tolerance on alpha in the solution file. I did tests on a coarse mesh with all other settings the same and fluid got stuck again, but when I added the increased tolerance, it stopped being stuck. A short test on the finer mesh suggested that it would simulate fine. Very odd!
My modfication to the 0 folder were just to change the atmosphere/outlet to:
U

type pressureInletOutletVelocity;
value            uniform (0 0 0);

p_rgh

type totalPressure;
p0        uniform 98800;
value    uniform 98800;

[geth03]

when you changed your BC for U to pressureInletOutletVelocity, you gave your system some free space to adjust everything. so your flow right there is adjusted to the pressure there. in the other case, where you had velocities, i think a correct solution couldn't be found bc pressure-velocity coupling didn't work well.

[jammy4536]

Hi Geth,

I just tested with everything exactly as I posted originally (freestreamPressure for p_rgh, and zeroGradient for velocity on outlet and atmosphere) but with a higher tolerance on the alpha in the fvSolution file and the problem is solved, so I think that was the only issue.
Since the inletOutlet conditions apply zeroGradient on outlets anyway, it is basically the same boundary condition. Truly the only factor in the effect of the alpha phase was the tolerance.
Using total pressure I see a low velocity patch at the atmosphere boundary in the simulation, but I don't tend to see that with a zeroGradient boundary hence why I use it.