[Elliptic CFD]

Hello,

I am having problems mapping fields between two regions within my solver in parallel. I am using OpenFOAM+ 3.0 and the meshToMesh0 utility.

When I run a serial case, the mapping function works well. I will give a brief description of the serial code within the solver.

The source and target meshes are constructed as srcMesh_ and tgtMesh_ respectively. The mapping function for a volume vector field is defined as:

void interpToMesh0::interpField(const volVectorField& srcVec, volVectorField& tgtVec)
{
meshToMesh0 interpolatorObj_(srcMesh_, tgtMesh_ , patchMap_, cuttingPatches);
interpolatorObj_.interpolate(tgtVec, srcVec, mapOrder_, eqOp<vector>());
}

This code works fine in serial within the solver. The problem starts when the solver is run in parallel. Modifications are made to the vector mapping function in line with the mapFields utility already supplied in OpenFOAM 3 or 4. For instance, addProcessorPatches(tgtMesh_ , cuttingPatches) replaces cuttingPatches.

My question is how to access the decomposed source and target meshes during runTime and incorporate the submeshes into the meshToMesh0 object of the solver? mapFields in OpenFOAM works by creating the submeshes before activating the mapping utility. In my case, the meshes have already been declared at runTime, and I will like to know how to access the different decomposed parts of the source and target meshes during runTime.

Thank you