Add OpenFOAM partitioned heat equation

[MakisH]

This is an attempt to port the 2x-laplacianFoam case from the OpenFOAM adapter repository and shape it to simulate the same problem as the FEniCS and Nutils cases that are already here.

Change with side effects: In the precice-config.xml, the data Flux needs to be renamed to Heat-Flux, as the OpenFOAM adapter uses the data name (prefix) to determine the data type. Naming it Flux in the first case sounds like a historical residue and does not agree with our flow-over-heated-plate case. This change means that the data needs to be renamed in the (case-specific) FEniCS and Nutils examples.

This is only an incomplete attempt, as I cannot manage to set the heat source and sink to be a single point/cell in OpenFOAM. I have tried different (less or more wrong) approaches:

• Define an L-shape faceSet and create a patch from it using createPatch. This is the setup I include here. It runs, but not as expected: the additional patch does not get any temperature values and ParaView complains. A problem may be that there is a conflict of boundary conditions for the same faces.
• Define a cellSet and set its internal field with setFields. Somehow this also gave weird results.
• Define a smaller box around the corner in blockMeshDict. I guess the problem there was that the blocks were overlapping.

A trivial workaround is to set the complete bottom & left walls of the Dirichlet participant as a heat sink and the complete top & right walls of the Neumann participant as a heat source (I think I even swapped the names accidentally). But this is then a different setup and the isolines have completely different shape.

@davidscn if you want, feel free to start from this and debug it.

TODO:

• Make the case work with point source & sink.
• Check the case parameters for consistency. I have set T_sink = 1, T_source = 8, k = 3 (which should be alpha). I have no clue what to do with beta.
• Make the same changes also to the Neumann participant.
• Cleanup any unused files: topoSetDict, createPatchDict, unused settings in blockMeshDict/T, etc.
• Update FEniCS / Nutils for Flux/Heat-Flux
• Add entry in the README.md

The original motivation for this case was to study precice/openfoam-adapter#93. The issue is observable also in the current state of the case.

This also closes #204.

[davidscn]

After some investigations:
I don't think we can use laplacianFoam for this test case, since laplacianFoam solves the heat equation with a zero RHS whereas the described test case has a non-zero RHS. We could adjust the existing test case by applying the constraint RHS=0 using alpha and beta (set beta=2+2*alpa), but I'm not sure if this would be intended.

[MakisH]

Closing as duplicate of the now merged #223.