Is this steady or transient case? You mention physical time step.
If the interactions between stationary and rotating components are significant, I would recommend switching to sliding-mesh model. Although, you may still want to get a stable solution with MRF as your initial solution.
You are navigating away from the AIS Discovery experience