problem of convergence
 

Hello everyone,
My cas is of turbumence, a "T" form tube with two cylindres. it has one inlet and two outlets, with the distribution of water, 50% mass flow to each outlet.

My problem is that I always have the convergence problems and there are also warnings of reversed flows.

I tried the standard wall function, enhanced wall function and realizable model of k-epsilon model, and I also reffined the mesh, extended the outlet, but the situation hasn't been improved.

By the way, the Y+ is always between 5 and 30.

Do you have any suggestions ?

Thank you.
Steven