[jaaybee]

Hello Everybody,

i have a question due to the cyclic boundary condition. I want to combine the transform option: translational and rotational. But if i wirte it together in one file (blockMesh.dict or createPatch.dict) Openfoam only read the last command. My back patch is shifted in X and Y direction. Does OpenFoam have any command to match my front and back patch with a rotational transform and a translational transform?

Many thanks in advance for your support!

Jens

[taalf]

Hi all,

I am facing the same issue.

I need to connect two patches whose relative position can be defined only by a combination of a rotation + a translation.

Cyclic boundaries are perfect to do that, but the built-in solution allows only pure rotation or pure translation.

Does anybody have any idea how to implement rotation+translation cyclic boundaries ? Or is there some alternative built-in solution ?

Best regards,

[GRAUPS]

I am also interested in the answer to this question. It seems to only allow one or the other.

cycout
    {
        type            cyclicAMI;
        inGroups        1(cyclicAMI);
        nFaces          0;
        startFace       79747;
        matchTolerance  0.0001;
        neighbourPatch  cycin;
        transform       translational; // rotational       
        separationVector (0 -0.048568 0.096656);
        //rotationAxis (1 0 0);
        //rotationCentre (0 0 0);
        //rotationAngle 30; // optional
    }

[taalf]

Hi,

In that time I asked to some OpenFOAM code expert. He said this is just not possible in OpenFOAM...

Can I ask a bit more detail about what you want to do? To see if the alternative I used can work also in your case?

All the best,

[GRAUPS]

Hi William,

Thanks for your response. Did the OpenFOAM expert say whether it was simply a current functionality limitation, or was it a deeper more difficult to solve limitation? The functionality appears to be there, it just needs to be combined in a way so that it works.

I'm currently looking to simply connect two water flow pipes of the same diameter with different meshes and at a position that can't be represented by a single transform.

[taalf]

Hi Brock,

As surprising as it is, yes, it's a deeper problem.

My issue was quite similar. What I did was to connect pipes with physical meshes that overlap the main mesh.

In OF you can make cells overlapping each other. They just ignore each other.

So, I meshed the domain, then the linking pipe(s) separately, then use the mergeMesh command (just puts all cells in a single polyMesh) and cyclicAMI patches (with "noOrdering" transform - meaning no transformation) to connect everything.

Like this you can link your pipes with original mesh untouched (you just add to it overlapping mesh linking pipes).

Hope this helps...

All the best,

[alfiogn]

Hi,

I have the same issue... I have to make cyclic an inlet and outlet.

My case is a single screw extruder and I don't want to simulate an entire geometric periodicity of the screw. Hence I consider only a fourth of it and wanna map the inlet rotational velocity profile to the outlet BUT rotated by 90 degrees.

Does someone have found a solution to this issue or it is also an open issue?

Thanks a lot
Giorgio

[mpiecuch]

Hi, has anyone tried to use cyclicAMI for the described problem in two stages, for example in the first cyclicAMI using translational transformation and in the second cyclicAMI using rotational transformation?