[Elvin]

Hello, dear OF users. I am investigating the flow in the pipe shown in the above file. The flow leaving the inlet goes to the outlet. But as you can see, in outlet the flow drops down. It seems to "drop" down instead of flowing straight and maintaining its level. The inlet is divided into two parts. I have tried many variants of boundary conditions (inletOutlet, outletInlet, etc.), but everywhere the same result. Right now my geometry is in the positive coordinate region, because when the geometry crossed the coordinate axes, then the flow was reflected from the outlet. Also if you change the direction of free fall acceleration (g), then the reflection from the outlet boundary reappears. Please help with this. Thanks.

U:

dimensions      [0 1 -1 0 0 0 0];

internalField   uniform (1 0 0);

boundaryField
{
    inlet_bottom
    {
        type            fixedValue;
        value           uniform (1.5 0 0);
    }
    inlet_top
    {
        type            fixedValue;
        value           uniform (0.1 0 0);
    }
    walls
    {
        type            noSlip;
    }
    outlet
    {
        type            zeroGradient;
    }
}

p_rgh:

dimensions      [1 -1 -2 0 0 0 0];

internalField   uniform 0;

boundaryField
{
    inlet_bottom
    {
        type            zeroGradient;
    }
    inlet_top
    {
        type            zeroGradient;
    }
    walls
    {
        type            zeroGradient;
    }
    outlet
    {
        type            fixedValue;
        value           uniform 900000;
    }
}

alpha.water:

dimensions      [0 0 0 0 0 0 0];

internalField   uniform 0;

boundaryField
{
    inlet_bottom
    {
        type            fixedValue;
        value           uniform 1;
    }
    inlet_top
    {
        type            fixedValue;
        value           uniform 0;
    }
    walls
    {
        type            zeroGradient;
    }
    outlet
    {
        type            inletOutlet;
        inletValue      uniform 0;
        value           uniform 0;
    }
}

1.jpg

[vatavuk]

Hi Elvin,
The recommended practice for setting up a new simulation is to use a tutorial similar to the case you want to study. I suggest you study the weirOverflow tutorial.

[Elvin]

Originally Posted by vatavuk

Hi Elvin,
The recommended practice for setting up a new simulation is to use a tutorial similar to the case you want to study. I suggest you study the weirOverflow tutorial.

Thanks for the reply! But it didn't help, I don't understand why primitive BCs don't help (such as fixedValue, zeroGradient etc).

[wkernkamp]

Is the solution converged?

[Elvin]

Originally Posted by wkernkamp

Is the solution converged?

Hello, wkernkamp! Yes, solution converged. The problem, it seems to me, is that the pressure gradient doesn't converge correctly. Inlet and outlet are kind of disconnected.

123133.jpg

I also feel like it's physical on one hand. Because when the flow rate increases, this level drop at the outlet is not noticed. And when the flow velocity is low, this "drawdown" appears to the left (meaning closer to the inlet). There seems to be a minimum velocity for a certain liquid level in the pipe. If anyone knows of any articles dealing with this, I would be glad to read them.

[VRN]

Hello Elvin,

I noticed you have assigned a fixed value pressure boundary for the outlet at 900000Pa. That might be causing the dip near the outlet.

Best
VRN

[vatavuk]

Originally Posted by Elvin

Thanks for the reply! But it didn't help, I don't understand why primitive BCs don't help (such as fixedValue, zeroGradient etc).

HI Elvin,

One of the reason why complex boundary conditions must be used have to do with the formulation of SIMPLE and PISO algorithms. They use pressure as fundamental variable but in the Navier-Stokes equations (for incompressible flow) there is no explicit equation for pressure.

So these methods create a pressure equation using the momentum transfer and the continuity equation. This pressure equation must have boundary conditions that are compatible with the two previous equations (momentum transfer and continuity), so sometimes complex boundary conditions like the fixedFluxPressure boundary condition must be used for pressure.

There is a good explanation about this in chapter 4 of the book of Greenshilds and Weller that is available at:
https://doc.cfd.direct/notes/cfd-gen...iples/contents

I hope this helps.