Faces incorrectly oriented DURING the simulation
 

Hi Foamers,
I'm running a free decay test in OF2.2 with interDyMFoam applied to a floating sphere constrained to move only in the heave motion, created with snappyHexMesh, simulating only a quarter of the sphere with two symmetry planes.

The simulation crashes with a FPE after a while. Looking into the log file I noticed that alpha1 that should be between 0 and 1, suddenly falls to an enormous negative value (-7e+283) right in the last time step and then crashes.

Running a checkMesh in the beginning everything is fine, but running a checkMesh - latestTime after the crash, some faces incorrectly oriented pop out. This faces are written into a set called wrongOrientedFaces that I visualized in paraFoam. They are two one-cell-width-stripes in the bottom of the sphere next to the two symmetry planes.

Anybody can help me to understand why is this happening and what to do to avoid this problem?

Thank you all in advance,

Giuseppe