Divergence DrivAer model
 

Hello
I am running the DrivAer model with DDES k-w SST in fluent. What I do is to run URANS k-w SST for some time steps and then switch to DDES. I run DDES for 1500 time steps and then i enable data averaging and run it for 2000 more.
The problem is that it runs apparently well untill some time after those 1500 time steps when the message 'turbulent viscosity was limited to ...' appears and some iterations later 'floating point exception', so I lose a lot of time.

I am using Coupled p-v, second order for all the equations and bounded implicit second order for the time.

Once I have solved it running more iterations per time step and lowering the relaxation factors even more (0.3) but it was in another mesh and I dont know if that is a general solution because now I have the same relaxation factor as the ones in the case that ran well.

Any suggestion?

Thanks!