CONVERGENCE HELP

giogio

I study a 3d turbulent flow, I have a porous media and a flow with swirl, i have problem with convergernce with a 2nd order discretization ( PRESTO is suggested by Fluent wit a model with a porous media, I use a k epsilon standard model with standard wall function (I have dimensions of the order of meter)and a pressure based. I try to remake my mesh, changing relaxation factors, initialing wit a good solution but convergence doesn't get better(Residuals are about 10^-3 for continuity and 10^-4 for others parameters). Can you give me any suggestions (multigrid parameters....?

Thanks

Soryy for my English (I'm italian).