I'm studying a radial fan (7500 rpm) using mrf-implicit method, but I have a lot of convergence problem.I'm using wall function K-E high Reynolds turbulence model....can someone help me?????

use MIXASI and run it moving mesh transient, it isn't as expensive as you think, it is almost guarenteed to work and it'll give you a better answer.

B.

How much...expansive????? Plase, if you can give me more details!!!!

Thank you, Lorenzo

start with a low rpm, then make repeated restarts while increasing rpm. values for rpm could be 50, 250, 1000, 3500, 7500. some iterations, say 20, per rpm will do. save you postfiles in between ! as soon as you get into trouble reduce the rpm increment. most often the rotational speed of wall boundaries can be fixed at the final value during this process.

something else you can try is to alter the pressure dependence on density during the first iterations.

hope this helps