[pedroxramos]

Hello,

I need a reference (article, thesis or manual) where explicits shows how postProcess utility from OpenFOAM simulates the wallShearStress, more specifically when I use a wall model in a LES simulation.

Can someone tell me any document where is that information?

Thank you for your kind cooperation!

[wyldckat]

Greetings Pedro,

I'm late, but here is my answer just the same - In code documentation for OpenFOAM 5.0, see: https://cpp.openfoam.org/v5/classFoa...s.html#details
Quoting from there:

where:

• = stress tensor
• = patch normal vector (into the domain)

From the source code file "/src/functionObjects/field/wallShearStress/wallShearStress.C" at line 190:

    typedef compressible::turbulenceModel cmpModel;
    typedef incompressible::turbulenceModel icoModel;

    volVectorField& wallShearStress =
        mesh_.lookupObjectRef<volVectorField>(type());

    tmp<volSymmTensorField> Reff;
    if (mesh_.foundObject<cmpModel>(turbulenceModel::propertiesName))
    {
        const cmpModel& model =
            mesh_.lookupObject<cmpModel>(turbulenceModel::propertiesName);

        Reff = model.devRhoReff();
    }
    else if (mesh_.foundObject<icoModel>(turbulenceModel::propertiesName))
    {
        const icoModel& model =
            mesh_.lookupObject<icoModel>(turbulenceModel::propertiesName);

        Reff = model.devReff();
    }
    else
    {
        FatalErrorInFunction
            << "Unable to find turbulence model in the "
            << "database" << exit(FatalError);
    }

    calcShearStress(Reff(), wallShearStress);

Therefore what you need to search for is the "devReff()" and "devRhoReff()" methods for the turbulence model that you are currently using... although it seems like it's best to simply search for "devRhoReff()", because "devReff()" is just when "rho" is simply 1.0.

Since you are looking for LES, I went straight for Smagorinsky and there: https://cpp.openfoam.org/v5/classFoa...agorinsky.html - you can find that it's in the public member functions for this inherited class:

linearViscousStress< LESModel< BasicTurbulenceModel > >

The source code is shown here: https://cpp.openfoam.org/v5/linearVi...ce.html#l00070
The calculation one there is this:

this->rho_  *  this->nuEff()  *  dev(twoSymm(fvc::grad(this->U_)))

Which effectively is the product of the effective dynamic viscosity (muEff) with the deviatoric operator applied to "2.0 * symmetric tensor" of the gradient of the velocity field... which is related to this: https://en.wikipedia.org/wiki/Cauchy...eviator_tensor


Mmm... OK, if this still looks gibberish to you, take a look into Tobias Holzmann's "Mathematics, Numerics, Derivations and OpenFOAM" document: http://www.holzmann-cfd.de/index.php...s-and-openfoam - this is just so I don't need to go into more detail myself on the mathematics


edit: I forgot to mention that it looks like this calculation is common to any of the turbulence models.

Best regards,
Bruno

[albet]

I found the answer,

wallShear = (-mesh.Sf().boundaryField()[patchID]/(mag(mesh.Sf().boundaryField()[patchID])) ) & (turbulence->devReff()().boundaryField()[patchID])


this piece of code gives the wall shear stress equal to wallShearStress utility.
Kind regards,
Amir

[Tibo99]

Hi everyone,

I don't know if I'm on the right post for my following question but, I hope someone here will guide me trough it.

Im trying to simulate a neutral Atmospheric Boundary Layer with a LES tuubulence model in Openfoam. The thing is, when your mesh is not small enough (which is often the case with limited resources) there is a "correction" that you need to do at the bottom wall. The option is to use a "synthetic wall shear stress" at the bottom wall.

One of the most popular "synthetic wall shear stress model" used is shown in this following paper (also see the formula in attachment):

Journal of Wind Engineering & Industrial Aerodynamics 82 (1999) 189-208
T.G. Thomas, J.J.R. Williams
Generating a wind environment for large eddy simulation of bluff body flow

Is there someone who can guide me on how I could code this "wall model", suggesting me a tutorial in Openfoam programming to achieve this? Maybe starting with a code similar to this one or there is maybe something online that I could find close to this?

Thank you very much!

Best Regards,

[wouterremmerie]

Hi everyone,

perhaps related to this post:
when we run k-omega SST sims (steady state with wall functions) using a snappyHexMesh without layers, we see the meshing pattern reflected in the surface friction (see attached image). We tried blending for the wall functions and non-orthogonal correctors, but they don't solve the issue. When we add enough layers to the mesh, this mesh-related pattern disappears.

So probably, the "edgy" pattern has something to do with non-orthonogal cells close to the surface. "Nut" also seems to have this pattern in the cells touching the surface, so it's not just a matter of an error in the "cell-to-surface" calculation that wallShearStress performs, but really a problem with the way Reff is calculated perhaps.


Could it be that Reff is not calculated properly for distorted cells? And if so, is there a way to improve the quality somehow?

Thank you very much for your input!