URANS not converging - flow in channel
 

Dear all,
I would like to ask your opinion about a divergence problem I am encountering in Fluent.
I am simulating the flow in a bifurcating channel (there is 1 inlet and 2 exit channels) by setting up inlet total pressure and outlet static pressures. The RANS is converging well, but when I start the URANS with the same model (realizable k-epsilon) it diverges after more than 5000 time steps, which I believe is quite weird.
What do you think the problem might be? I am used to convergence issues which originate from the very first time steps (due to mesh and BCs) but this seems not to be the case. I was thinking that maybe there is a pressure wave propagating from the outlet to the inlet channel which causes solver issues, but I am not sure. What do you think about that?
Thanks a lot