[val46]

Hello,

i would like to use a negative velocity as BC for an outlet (the volume flow is known).
Unfortunetaly the solution crashes or shows huge values.

I'm using the simpleFoam solver with k-epsilon.
My guess is I have wrong settings:

0/U:

 in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform (-6 0 0);
    }

0/p:

 in
    {
        type            fixedValue;
        value           uniform 0;
    }
    out
    {
        type            zeroGradient;
    }

0/epsilon:

in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform 14.855;
    }

0/k:

in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform 0.375;
    }

0/nut:

in
    {
        type            calculated;
        value           uniform 0;
    }
    out
    {
        type            calculated;
        value           uniform 0;
    }

0/nuTilda:

in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform 0;
    }

0/R:

in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform (0 0 0 0 0 0);
    }

Does anyone have suggestions for that?
By the way, the solution looks good when I switch to blowing (positive velocity at inlet).


Regards,
Toni

[sega]

You are talking about an outlet at the beginning, about an inlet at the end.

What is it now?

Can you tell us how it is aligned in respect to your coordinate system?

[val46]

Sorry, I know it's confusing.
I attached a slice through the modell.
The outlet called "in" is on the left side. The inlet called "out" is on the right side.
Inlet and outlet is switched because i changed from blowing to suction.

In the meantime i found out the solvers crashes because of the high velocity. I tried a smaller value and it worked.

The value also has to be positiv. Otherwise it would be an inflow.

The code for 0/U:

in
    {
        type            zeroGradient;
    }
    out
    {
        type            fixedValue;
        value           uniform (0.5 0 0);
    }

Know I have to find out why the solver is instable for higher velocities...

[tfuwa]

Hello the Forum and a nice week to you all,

I would like to specify a negative inlet velocity B.C. in icoFoam solver and am encountered a problem.

I enclose the case, which is quite simple, in the attached file and also give a short description here. It's 2D case for flow simulation around a cylinder and the b.c.s are as follows,

inlet: U fixedValue -1 or 1, p zeroGradient
outlet: U zeroGradient, p fixedValue 0
at two sides, symetryPlane is applied and empty b.c. for in cylinder span direction.

fvSchemes are specified as, Euler for ddt, Gauss linear for gradSchemes, Gauss cubic for divSchemes and Gauss linear corrected for laplacianSchemes.

For U=1, icoFoam gives reasonable result. The force coefficients are shown in the attached Fig U=1. Strouhal number is 0.2, which agrees with experimental results. Drag coefficient is around 1, which is good.

But when you change U=-1, which is the case in the attached case, the simulation crashes at 95s. The force coefficients, as shown in Fig U=-1, behaves normal at the beginning, but increases to a very large value when the simulation stops.

I have no idea why I am experiencing such different results with only one change of the direction of the velocity. Can you please guide me to find out the reason? Any help would be greatly appreciated.

The case can be found here, http://dl.dropbox.com/u/62722643/ico...donline.tar.gz
And can be run directly with command

icoFoam

in OpenFoam 1.7.1

[ybapat]

Hello,

You can try to use lower relaxation for momentum equation. Also another option is to ramp velocity at outlet in steps.

Regards,
-Yogesh

[tfuwa]

Originally Posted by ybapat

Hello,

You can try to use lower relaxation for momentum equation. Also another option is to ramp velocity at outlet in steps.

Regards,
-Yogesh

Hi Yogesh,

Thanks for your answer. Can you explain why and where I should lower the relaxation for momentum equation?

Can you please also point out why there is such a great difference when the velocity direction changes?