interFoam problems with refinement at fluid surface
When running an interFoam case with two phases (water and air) with a refinement that crosses the surface as shown in picture 3, immediately after start a strange "explosion" like "dynamite fishing" occurs one cell above the surface within the refined cells along the refinement boundary, see picture 4.
The case has been derived from tutorials/multiphase/interFoam/RAS/floatingObject, simplified (no floating body any more, no water in the corner to fall into the basin at start), and modified with the said refinement.
A workaround for this simple case is to avoid local z-refinements close to the water surface.
But if I add a complex floating body with snappyHexMesh in a much larger case, this inescapably leads to refinements close to the surface resulting in high velocities in the smallest cells and thus high courant numbers. The only workaround is to add an initial calm down phase before the real action starts, but this obviously eats up a lot of CPU time.
Steps to reproduce
Run the example case with openFoam 2106. In allrun script different refinements may be selected by commenting/uncommenting: "none" (leading to pictures 1&2), "center" (leading to pictures 3&4), "block" and "all". Refinement "center" is pre-selected.
What is the current bug behaviour?
We get U values other than (close to) zero in cells one cell above the surface within the refined cells along the refinement boundary causing something that looks like a compression wave spreading from there. See Picture 4.
What is the expected correct behavior?
Nothing happens. Picture 4 should be all blue just as Picture 2.
Relevant logs and/or images
- OpenFOAM version : 2106 (docker image opencfd/openfoam2106-run)
- Operating system : Ubuntu 20.04.3 LTS