[rr3245]

Dear all,

Before calculating the Q field for snapshots in time (vortex identification), I would like to first filter the velocity field on some length scale. I was beginning to think this would be possible by taking the Q post-processing function, and adding an 'LES type' filter operation in before the gradU is calculated. For example, in myQ.C:

        const volVectorField& U = lookupObject<volVectorField>(fieldName_);
        const tmp<volVectorField> filtU(filter_(U));
        const tmp<volTensorField> tgradU(fvc::grad(flitU));
        const volTensorField& gradU = tgradU();

        return store
        (
            resultName_,
            0.5*(sqr(tr(gradU)) - tr(((gradU) & (gradU))))
        );

where filtU is my new, filtered field. I have managed to compile a new Q field in OF version 4.1, but I have no idea how to first filter the velocity field. I have tried to pick apart the code from different LES SGS models, but this hasn't been successful at all. Any help greatly appreciated.

[Andrea1984]

What is the filter_(U) doing in const tmp<volVectorField> filtU(filter_(U)) and where is it defined?

I am not sure what you are trying to do makes sense, especially in the context of implicit LES used in OpenFOAM. The filtering operation in implicit LES is not performed explicitly, but rather given by the numerical grid and discretization procedure following the FVM, with a filter width equal to your mesh size.

Andrea

[HPE]

Could you please which file you refer to for `filter_(U)`?

[H.Tofighian]

I think you'd better use "proper orthogonal decomposition" and remove high frequency modes.

[john_alm]

Originally Posted by rr3245

Dear all,

Before calculating the Q field for snapshots in time (vortex identification), I would like to first filter the velocity field on some length scale. I was beginning to think this would be possible by taking the Q post-processing function, and adding an 'LES type' filter operation in before the gradU is calculated. For example, in myQ.C:

        const volVectorField& U = lookupObject<volVectorField>(fieldName_);
        const tmp<volVectorField> filtU(filter_(U));
        const tmp<volTensorField> tgradU(fvc::grad(flitU));
        const volTensorField& gradU = tgradU();

        return store
        (
            resultName_,
            0.5*(sqr(tr(gradU)) - tr(((gradU) & (gradU))))
        );

where filtU is my new, filtered field. I have managed to compile a new Q field in OF version 4.1, but I have no idea how to first filter the velocity field. I have tried to pick apart the code from different LES SGS models, but this hasn't been successful at all. Any help greatly appreciated.

Hello rr3245 and others,

It would be nice to call filter_(U), e.g. from a coded function object. If anyone knows how to access this object, that would be a great help.