[matzbanni]

Hello,
I'm running a channel-case with a smaller cross-section in between. The little box in the middle is filled with a porous zone. Furthermore there are also other geometric details, which shouldn't influence the results.

The problem is, that the pressure values which I'm recording right in front of the porous zone are fluctuating considerable, aroung +/- 50 Pa. The values behind the zone are pretty continuous.

Any Ideas?

My fvSchemes:

ddtSchemes
{
    default         steadyState;
}

gradSchemes
{
    default         Gauss linear;
}

divSchemes
{
    default         none;
    div(phi,U)      Gauss GammaV 0.5;
    div(phi,k)      Gauss Gamma 0.5;
    div(phi,omega)  Gauss Gamma 0.5;
    div((nuEff*dev(T(grad(U))))) Gauss linear;
}

laplacianSchemes
{
    default         Gauss linear limited 0.333;
}

interpolationSchemes
{
    default         linear;
}

snGradSchemes
{
    default         corrected;
}

fluxRequired
{
    default         no;
    p;
}

fvSolution:

solvers
{
    p
    {
        solver           GAMG;
        tolerance        1e-9;
        relTol           0.1;
        smoother         GaussSeidel;
        nPreSweeps       0;
        nPostSweeps      2;
        cacheAgglomeration on;
        agglomerator     faceAreaPair;
        nCellsInCoarsestLevel 10;
        mergeLevels      1;
    }

    U
    {
        solver           smoothSolver;
        smoother         GaussSeidel;
        tolerance        1e-11;
        relTol           0.1;
        nSweeps          1;
    }

    k
    {
        solver           smoothSolver;
        smoother         GaussSeidel;
        tolerance        1e-11;
        relTol           0.1;
        nSweeps          1;
    }

    omega
    {
        solver           smoothSolver;
        smoother         GaussSeidel;
        tolerance        1e-11;
        relTol           0.1;
        nSweeps          1;
    }
}

SIMPLE
{
    nNonOrthogonalCorrectors 4;
}

potentialFlow
{
    nNonOrthogonalCorrectors 10;
}

relaxationFactors
{
    fields
    {
        p               0.3;
    }
    equations
    {
        U               0.6;
        k               0.3;
        omega           0.3;
    }
}

cache
{
    grad(U);
}

[Lieven]

Hi MB,

A simple test you can perform to see whether the fluctuations are caused by numerical instabilities (wiggles), is to set all 'Gauss gamma 0.5' to 'Gauss upwind'. From accuracy point of view, this is certainly not interesting. But if the wiggles disappear you know it is related to the discretization and not the case setup (BC, initial conditions, ...). If this solves the problem, replace the upwind scheme by a 2nd order one other than the gamma 0.5.

Cheers,

L

[matzbanni]

Originally Posted by Lieven

Hi MB,

A simple test you can perform to see whether the fluctuations are caused by numerical instabilities (wiggles), is to set all 'Gauss gamma 0.5' to 'Gauss upwind'. From accuracy point of view, this is certainly not interesting. But if the wiggles disappear you know it is related to the discretization and not the case setup (BC, initial conditions, ...). If this solves the problem, replace the upwind scheme by a 2nd order one other than the gamma 0.5.

Cheers,

L

Hi Lieven,
thanks four your reply. I already did a simulation with 'Gamma upwind' for k and omega, but the fluctuations are the same. So I have to check my BC, initial conditions etc.? I don't think that I made a mistake there...

Regards,
MB

[Lieven]

Hi MB,

Do it also for div(phi,U). Since the velocity field is directly coupled with the pressure, I even expect this to have a much bigger effect than k and omega.

Cheers,

L

[matzbanni]

Originally Posted by Lieven

Hi MB,

Do it also for div(phi,U). Since the velocity field is directly coupled with the pressure, I even expect this to have a much bigger effect than k and omega.

Cheers,

L

Just got the results. A little bit better, but still fluctuating a lot. Another suggestions?

Regards,
MB

[Lieven]

Did you do set all three to upwind at the same time?

What about the quality of the mesh? Hexahedral? Non-orthogonality? Coarse/fine?
Somewhere conflicting boundary conditions? All steady BC?

Cheers,

L

[matzbanni]

Originally Posted by Lieven

Did you do set all three to upwind at the same time?

What about the quality of the mesh? Hexahedral? Non-orthogonality? Coarse/fine?
Somewhere conflicting boundary conditions? All steady BC?

Cheers,

L

All three are upwind. Regarding the mesh, mostly hexahedral, and a few prisms. Non-orthogonality check is OK, but 3 highly skew faces. Could this be a reason?

Regards,
MB

[Lieven]

This is not so trivial to say but if the fluctuations and/or the prisms or skew faces are near each other this might indeed be the cause...

[matzbanni]

Originally Posted by Lieven

This is not so trivial to say but if the fluctuations and/or the prisms or skew faces are near each other this might indeed be the cause...

Right now I'm creating a new mesh. Thanks for your help so far, I'll let you know how it's working...

[matzbanni]

Originally Posted by matzbanni

Right now I'm creating a new mesh. Thanks for your help so far, I'll let you know how it's working...

Checked a few other meshes, still fluctuating even with upwind. A bit confusing because of the nice values after the porousZone...?!

Regards
MB

[Lieven]

Mmm, do you include the porous zone explicitly or implicitly in the momentum equation?

[matzbanni]

Originally Posted by Lieven

Mmm, do you include the porous zone explicitly or implicitly in the momentum equation?

So far I always included it explicit, but I read that the implicit solver is a bit more robust, let's see...

[Lieven]

Ok, I'm kind of running out of ideas. Anyone else maybe?

[fredo490]

From my experience on rhoPimpleFoam, you can try to set the pressure relaxation factor to 0.5 and also change the laplacian to 0.5.

I often face "oscillations" on the pressure field that disappear when I rise the relaxation factor. Maybe you can try it.

[matzbanni]

Originally Posted by fredo490

From my experience on rhoPimpleFoam, you can try to set the pressure relaxation factor to 0.5 and also change the laplacian to 0.5.

I often face "oscillations" on the pressure field that disappear when I rise the relaxation factor. Maybe you can try it.

Okay, I'll give it a try. What do you mean by change the laplacian to 0.5?

[fredo490]

laplacianSchemes { default Gauss linear limited 0.5; }

[NickolasPl]

Hello everyone,

I' m experiencing exactly the same problem. When I plot the pressure inside the porous region everything looks fine but A LOT of wiggling happens in the non-porous region. I experimented myself with both implicit and explicit posousSimpleFoam solver but without any luck. I also tried different schemes in the fvSchemes dictionary but the wiggling still exists. I do not have any turbulence models on (laminar flow) and I simply cannot understand why this wiggling takes place. The mesh was generated via Gmsh with prismatic cells filling the computational domain which very dense cell arrangements in some regions. Can anybody shed some light regarding this issue? From the BC's point of view I don't think I 've made any mistake. The flow behaviour generally looks..."realistic" short of, and as expected. But the pressure produces oscillations.

Thanks in advance,

Nickolas