[FlyBob91]

Hello to all,
i've just met some problem with scalarTransportFoam, because i find the scalar is larger than 1 in some points and smaller than zero. I think this is up to the schemes used.
I'm working on OF1612+ and this is my fvSchemes file (the original from pitzDaily case).
I run the direct simulation to find the U field then i copy it in the 0 folder where i run scalaTransportFoam.
Why is my scalar out of bounds? Any Solution?

/*--------------------------------*- C++ -*----------------------------------*\
| =========                 |                                                 |
| \\      /  F ield         | OpenFOAM: The Open Source CFD Toolbox           |
|  \\    /   O peration     | Version:  plus                                  |
|   \\  /    A nd           | Web:      www.OpenFOAM.com                      |
|    \\/     M anipulation  |                                                 |
\*---------------------------------------------------------------------------*/
FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    location    "system";
    object      fvSchemes;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

ddtSchemes
{
    default         Euler;
}

gradSchemes
{
    default         Gauss linear;
}

divSchemes
{
    default         none;
    div(phi,T)      Gauss linearUpwind grad(T);
}

laplacianSchemes
{
    default         none;
    laplacian(DT,T) Gauss linear corrected;
}

interpolationSchemes
{
    default         linear;
}

snGradSchemes
{
    default         corrected;
}


// ************************************************************************* //

Thanks for help

[floquation]

Use a bounded scheme instead.

Replace your "divScheme" (so the scheme for the convective term) by banana.
Then choose one of the schemes that start with "limited". For example (disclaimer: typed from memory, could contain a syntax error):

div(phi,T) limitedVanLeer 0 1;

[FlyBob91]

Originally Posted by floquation

Use a bounded scheme instead.

Replace your "divScheme" (so the scheme for the convective term) by banana.
Then choose one of the schemes that start with "limited". For example (disclaimer: typed from memory, could contain a syntax error):

div(phi,T) limitedVanLeer 0 1;

Many thanks Kevin. I wrongly took for granted the solver bounded the scalar between 0 and 1.
Maybe can you suggest me a second order scheme (i noticed that the linear scheme flattens too much the solution) that works good in channel flow?

[floquation]

Did you mean "linear" or "linearUpwind"? Since you said "linear" in your last post, but your first post showed that you were using "linearUpwind".

Either way, according to the OF user guide, both "linear" and "linearUpwind" are already second order.

Personally, I am using limitedVanLeer, but what works for my case needs not work for your case. I am performing multiphase simulations with interFoam.

[FlyBob91]

Hello,
I tried to use some bounded schemes they seem to limit correctly the passive scalar but, unfortunately, the solution is very far to what i have to obtain.

If you compare the image of what i have to obation with my solution without the bounded condition you can see they are qualitatively similar except for the scalar which is unbounded

What I have to Obtain.png

My Solution Without Buonded Condition.png

The fvSchemes i used are those i reported in the first post.
Then i used these schemes

ddtSchemes
{
    default         Euler;
}

gradSchemes
{
    default         Gauss linear;
}

divSchemes
{
    default         none;
    div(phi,T)      Gauss limitedLinear01 1;
}

laplacianSchemes
{
    default         none;
    laplacian(DT,T) Gauss linear corrected;
}

interpolationSchemes
{
    default         linear;
}

snGradSchemes
{
    default         corrected;
}

fluxRequired
{
    default         no;
    T               ;
}

And here the results. It seems Gauss limitedLinear01 flattens to much the scalar field.

My Solution With Bounded Condition.png




I state my case is a mixing problem in a channel with 2 inlets. At the first Inlet i impose T=1 and at the second T=0 and the images show the outlet of the channel. I'm working on ofv1612+
Another question. Reading the OF user guide i noticed that Gauss limitedLinear01 is followed by another 1. What is his meaning?

Any Idea?
Thanks for help

[FlyBob91]

Hello to all,
i'm here again. I literally tested alle the limited method as suggested but the problem continues to appear.

Here all the files of my case. Maybe the problem is not the scheme.

ddtSchemes
{
    default         Euler;
}

gradSchemes
{
    default         Gauss linear;
}

divSchemes
{
    default         none;
    div(phi,T)      Gauss linearUpwind grad(T);
}

laplacianSchemes
{
    default         none;
    laplacian(DT,T) Gauss linear corrected;
}

interpolationSchemes
{
    default         linear;
}

snGradSchemes
{
    default         corrected;
}

FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    location    "system";
    object      fvSolution;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

solvers
{
    T
    {
        solver          PBiCG;
        preconditioner  DILU;
        tolerance       1e-06;
        relTol          0;
    }
}

SIMPLE
{
    nNonOrthogonalCorrectors 0;

}

dimensions      [0 0 0 0 0 0 0];

internalField   uniform 0.5;

boundaryField
{

    inlet1
    {
        type            fixedValue;
        value           uniform 1;
    }

    inlet2
    {
        type            fixedValue;
        value           uniform 0;
    }

    outlet
    {
        type            zeroGradient;
    }

    walls
    {
        type            zeroGradient;
    }

    movable
    {
        type            zeroGradient;
    }

}

thanks for help

[FlyBob91]

Little update!
The Gauss linearUpwind grad(T) for the divergence term seems to be the scheme which, qualitatively, gives best results. Is there a manner to make to mantain it applaying a limit? Os is better if i start to work to the other term as the gradient and laplacian?