[JCollins]

OpenFOAM seems have a lot on inconsistencies and overconstraints over the data specified within it. I am trying to set up a case where an internal object creates a flow velocity on one side and controls the domain pressure on the other side. The mesh is a large volume and the flow creator is entirely enclosed within it. I have its shape defined in one STL file, and its inlet and outlet faces in two other STLs that are subsets of the main file.

It seems that the only way for snappyHexMesh to generate a patch definition is to use a named STL it as a refinement surface. Without using refinement, the STL is ignored. What do I do if I do not want the object to control the mesh density around it?

When specifying the inlet and outlet patches, they also need to be defined using refinement surfaces to be recognised and given their type. This refinement level is also redundant and can conflict with the already-defined refinement on the overall shape.

This all creates a problem because it seems that all named patches need to be referred to in the boundary conditions files in the '0' folder. Here I can freely set the inlet and outlet conditions, but simpleFoam will not allow the main shape to go undefined. Clearly I cannot have one definition for both sides of the object, as they have different roles, and it seems that overconstraining for instance the velocity, leads to the solver failing to start.

As a result of the above, I cannot seem to progress with this model.

Update:

I seem to have been able to get simpleFoam running with the shared patches, however have discovered in paraview that the shared patch faces aren't actually shared, they are randomly distributed between the two patches. It means there is no consistent boundary condition across the faces that I would like to define. Is there a way to define a patch on a surface cut out with a 'superset' STL?

Thanks in advance.