[cyln]

Hello, I have a computational domain which is a quarter cylinder. On the sides, I defined cyclic (periodic) BCs. I have run a transient simulation, however, on the cyclic interfaces (boundaries), there is an unexpected gradient on my flow parameters such as velocity. A screenshot of velocity contour plot is attached.

I have assigned the BCs are as follows:

0/p:

    SIDE11
    {
        type            cyclic;
    }

    SIDE22
    {
        type            cyclic;
    }

0/U:

    SIDE11
    {
        type            cyclic;
    }

    SIDE22
    {
        type            cyclic;
    }

constant/polyMesh/boundary :

    SIDE11
    {
        type            cyclic;
        inGroups        1(cyclic);
        nFaces          54131;
        startFace       16644366;
        matchTolerance  0.001;
        transform       rotational;
        neighbourPatch  SIDE22;
        rotationAxis    (1 0 0);
        rotationCentre  (0 0 0);
    }
    SIDE22
    {
        type            cyclic;
        inGroups        1(cyclic);
        nFaces          54131;
        startFace       16698497;
        matchTolerance  0.001;
        transform       rotational;
        neighbourPatch  SIDE11;
        rotationAxis    (1 0 0);
        rotationCentre  (0 0 0);
    }

Could anyone please tell me why I get such a gradient gradient very next to the cyclic interface?


Cheers

[piu58]

It looks if there is a wall function in action.

[cyln]

Hello Uwe,

Thanks for your reply. This is a laminar simulation, so there are only U and p within the 0 directory. The boundary conditions for the cyclic boundaries are as above. I do no use any wall function in the simulation.

I just cannot find the reason behind this problem :/

Thanks

[cyln]

Does anyone have another suggestion?