Not an expert on multi-phase DOF simulation but it seems like the issue you’re encountering are a result of a setup that still has issues rather than say a non ideal CFL number that causes your simulation to go into instability instantly.
Referring back to what Jousef referenced with regards to the 6DOF boat project, there are some irregularities that are in your simulation as compared to that.
For a start, your phase fraction sub-domain is being applied to a small area rather than the entire water domain which is occupied. This is what the 6DOF boat simulation project has done. You can start with that change and see if it helps.
The second irregularity I noticed was the boundary condition related to the object in the water, which is this case is your geometry. In the 6DOF boat simulation the geometry is defined as a moving wall velocity with zero values in the xyz direction, fixed flux pressure and a phase fraction value of gradient set to zero. Yours is a fixed value. Maybe this change might help.
Generally speaking if you’re running into issues like this I don’t think it is a CFL number issue. I would recommend looking into a successful simulation like the 6DOF boat project and to replicate it as closely as you can to every detail. Of course your case I believe is slightly different due the presence of the inlets having a set velocity so adjust as accordingly.
Hopefully this helps! Do let me know if it works.