[pardoa]

Hello,

I have successfully coupled two solvers to solve several fields on two separate meshes. Both meshes have the same resolution and share one patch which works as coupling interface. The flow solution from each mesh at the coupling interface is mapped on the other mesh at each timestep. This means, e.g, the U field resulting at the interface patch of one mesh is mapped on the the same patch of the second mesh at each timestep. To achieve this, I have used the function PatchToPatchInterpolation. However, the mapping doesn't work properly. The solution is correct at most of the faces, but somehow, the values are not transferred at some. Since, initially, the nodes are at the same location for both meshes at the interface patch, I assume that the lack of mapping at some faces could be the result of rounding errors? I am just guessing here. Else I don't see why the faces are not correctly mapped from one patch to the other, as their location is exactly the same.

This is the piece of code I am using for the mapping:

PatchToPatchInterpolation<primitivePatch,primitivePatch> bottomToTopInter
(
    bottomPatchMesh1, //fromPatch
    topPatchMesh2 //toPatch
);

PatchToPatchInterpolation<primitivePatch, primitivePatch> topToBottomInter
(
    topPatchMesh2, //fromPatch
    bottomPatchMesh1 //toPatch
);

U2.boundaryFieldRef()[topPatchMesh2ID] == bottomToTopInter.faceInterpolate<scalar>(U1.boundaryField([bottomPatchMesh1ID]);

U1.boundaryFieldRef()[bottomPatchMesh1ID] == topToBottomInter.faceInterpolate<vector>(U2.boundaryField([topPatchMesh2ID]);

Is there a way to define an interpolation distance or interval? Maybe that could help. Any hint on this matter would be greatly welcome. Thanks!!

Alvaro

[pardoa]

Hello again,

I tried to directly assign the values from one patch to the other based on the face center coordinates, but some faces remained unmapped.

This is the piece of code that I implemented:

forAll (faceCtr_meshSub,i)
{
    forAll (faceCtr_mesh,j)
    {
        if (round(mag(faceCtr_mesh[j])*10e6) == round(mag(faceCtr_meshSub[i])*10e6))
        {
            hSub.boundaryFieldRef()[topPatchMesh2ID][i] = hSur.boundaryField()[bottomPatchMesh1ID][j];
        }
    }
}

The coordinates are rounded to avoid potential small differences (smaller than 10e-9) between the coordinates of the face centers of both patches.

By doing this, the differences between both patches decrease dramatically, but the mismatch persists at some cells. I really have no idea why this happens, as now the assignment is straight. There is no interpolation.

Any help will be very welcome!

Alvaro