[eric.m.tridas]

Hey Foamers,

I'm working on a project simulating the air flow into a vacuum chamber through a controlled leak from atmosphere. I was just wondering if anyone else has tried something similar using OpenFOAM.

I'm somewhat stuck at the moment as my simulation using rhoSimplecFoam fails after a few iterations. I've also tried rhoCentralFoam but obtain strange results after a time (the fluid begins to flow from the center of the leak to both the vacuum chamber and atmosphere.) If anyone has any experience with a case like this your input would be greatly appreciated! My BC's are:

U
inlet - pressureInletOutlet
walls - slip
outlet - fixedValue, 18.67 m/s (simulating flow rate of vacuum pump)

p
inlet - fixedValue, 101325
walls - zeroGradient
outlet - zeroGradient

T
inlet - fixedValue, 293
walls - zeroGradient
outlet - zeroGradient

Again, thanks in advance for your help!

-Eric

[eric.m.tridas]

Hey everyone,

Here is an image of the results that I get when performing this simulation using rhoCentralFoam. Notice in the narrow region on the left that the flow is moving outward from the center. (Note: in this image the slip boundary was replaced with a fixedValue (0 0 0) at the wall. The effect is the same with the slip boundary condition)

Any help with this is really appreciated as I've been stuck with this issue for a little while now. Thanks!

-Eric

[eric.m.tridas]

This is an image of the 2-D case I started with before working with the 3-D case shown in the previous posts. Although the same outward flowing phenomenon is not occurring, there are still very strange velocities observed. The flow in the narrow region is vertical (as shown in the overall image with glyphs). This simulation was performed with rhoCentralFoam as well.

[eric.m.tridas]

There may be a problem with my fvSchemes dictionary. Here is the code:

fluxScheme      Kurganov;

ddtSchemes
{
    default         Euler;
}

gradSchemes
{
    default         Gauss linear;
}

divSchemes
{
    default         none;
    div(tauMC)      Gauss linear;
}

laplacianSchemes
{
    default         Gauss linear corrected;
}

interpolationSchemes
{
    default         linear;
    reconstruct(rho) vanLeer;
    reconstruct(U)  vanLeerV;
    reconstruct(T)  vanLeer;
}

snGradSchemes
{
    default         corrected;
}

[jiez]

Hi Eric,

Have you solved this problem?

Jie

[Linse]

While I do not yet have a solution for this point, at least I can give some advice on this thing:
- Try with rhoPimpleFoam as well. For a similar case some knowledgable people from support had suggested that to me.
- Check if you are in the acceptable range of densities for the FVM-method. At some point the density in the suction part gets to low as to use a method working with parameters averaged over the control volumes. Search for "mean free path" and "Knudsen number" for more information.
- See if you need some wave transmissive BCs, particularly at a potential outlet. Flow with extreme pressure differences can cause shock waves which can cause real problems in simulation if not treated correctly.

[JR22]

Did any of you find a solution to your problem?

My problem is very similar: I also have a vacuum chamber with vacuum decay through a needle valve. I am using rhoPimpleFoam, a 3D geometry, and have not yet settled on a turbulence model. I started using rhoSimpleFoam, and later on realized this cannot be steady state. My vacuum is at ~ 300 Pascal and a flow of 2.7x10-8 m3/s and connects to atmospheric pressure. Although the simulated leak is supposed to be 10um, the orifice I use can be bigger than that because this is a needle valve and I am interested in the vacuum decay (in Pascal).

Thanks.