[Maff]

Hi all,

I'm having some problems with cyclicAMI patches (OpenFOAM 2.1.1).
I'm working on a 3-bladed wind turbine, meshing a third of the domain (120-degree slice, single rotating reference frame).

The model is SpalartAllmaras and the solver is MRFSimpleFoam (by the way, MRFSimpleFoam is far more stable than SRFSimpleFoam in my opinion).

At the beginning I had problems with the convergence of nut and nuTilda.
After trying a lot of different schemes and relaxation factors, I noticed that "y" (wall distance in yPlusRAS) was calculated wrong. And I thought: that's why nut is not good.
With "cyclic" patches the wall distance is ok, so cyclicAMI seemed to be the problem.

Then I found this thread,
I added the cyclicAMI patches to "nonRotatingPatches" in MRFZones dictionary and now my results are ok (at least they seem ok).
But the wall distance y is still not good. I don't know if this can cause any problem, maybe with the calculation of y+.

The attached images show the problem with the calculation of y on the cyclicAMI patches (purple arrow).
The contour of y shows where the problem starts (yellow arrow), i.e. where it thinks there's a wall but there isn't.

I created the cyclicAMI patches with the following createPatchDict.

pointSync false;
patches
(
    {
        // Name of new patch
        name cyclic_half0;
        // Type of new patch
        patchInfo
        {
            type cyclicAMI;
   neighbourPatch cyclic_half1;
   transform rotational;
   rotationAxis (0 1 0);
   rotationCentre (0 0 0);
   matchTolerance 0.0001;
        }
        // How to construct: either from 'patches' or 'set'
        constructFrom patches;
        // If constructFrom = patches : names of patches. Wildcards allowed.
        patches (periodic1);
        // If constructFrom = set : name of faceSet
        set f0;
    }
 
    {
        // Name of new patch
        name cyclic_half1;
        // Type of new patch
        patchInfo
        {
            type cyclicAMI;
   neighbourPatch cyclic_half0;
   transform rotational;
   rotationAxis (0 1 0);
   rotationCentre (0 0 0);
   matchTolerance 0.0001;
        }
        // How to construct: either from 'patches' or 'set'
        constructFrom patches;
        // If constructFrom = patches : names of patches. Wildcards allowed.
        patches (periodic2);
        // If constructFrom = set : name of faceSet
        set f0;
    }
);

I tried to specify the "rotationAngle", but it doesn't seem to have any effect.

Is this a bug or am I doing something wrong? Do I have to worry about this? (will y+ be calculated right?)

Thank you in advance

[Maff]

there's an easy way to reproduce this.
in the "mixer" tutorial (incompressible/SRFSimpleFoam/mixer) replace the cyclic patches with cyclicAMI patches.
run yPlusRAS and you will get strange "y" contours.

[kiddmax]

Hi Maff,

Although I know it is a quite old thread. But I am still wondering that you mean MRFSimpleFoam is more stable than SRFSimpleFoam. Can you give some suggestions about the tips to run SRFSimpleFoam? I also want to model a 3-bladed wind turbine.

The other question is can you use one third of domain with MRFSimpleFoam solver? not the whole domain for MRF ?

Thanks a lot!

Best regards
Ye

[Maff]

well, it's been a while. this is what I remember: the problem I had with SRFSimpleFoam started at the axis of rotation, nut started increasing without reason.
with MRF (simpleFoam with fvOption dictionary in the last versions of openfoam), I didn't have that problem.

you can try MRF for your case, it behaves as SRF if you specify just one region.
as for the stability, it may depend on the case, there's probably a reason why you have both options in openfoam. (for one, with MRF you have to calculate Urel yourself)

[kiddmax]

Hi Maff,

Thanks a lot for your reply.

I want to save computational time. For SRFSimpleFoam I can use one channel blade mesh for sure with periodic boundary conditions , but I am not sure whether I can use it for MRFSimpleFoam.

Best regards,
Ye

[Maff]

I've used simpleFoam with MRF on a periodic single-blade mesh (2-bladed wind turbine). So I guess the answer is yes.