[Bharatt]

Hello everyone,

i am running a simulation for different turbulence models for an airfoil. I am facing AMG divergence or sometimes FLOATING ERROR: OVERFLOW when i am shifting to RSM turbulence model.
the procedure i am following is achieving convergence in K-epsilon at first place and then shifting to RSM, the following parameters i used in initial(K-epsilon) stage.

Solver:- pressure based, Green-gauss node based
Viscous:- Standard K-epsilon
Boundary conditions(under turbulence):- K and epsilon
Scheme:- SIMPLE 1st order for 100 iterations then 2nd order(URF:- momentum reduced to 0.5 from 2nd order)

By about 4000 iterations convergence is achieved.
NOW, i go back to viscous:-
under solver-->viscous:- RSM, linear pressure strain etc and rest kept the same(parameters which i used for k-epsilon).

By about 100 iterations AMG divergence occurred

P.S:- please correct me if i am doing anything wrong!
-do i have to initialize it again?

thanks in advance

bharat

[meangreen]

You don't have to initialize again, you can use the solution from k-epsilon to converge with RSM.

I wonder if you are running into issues with your boundary conditions. Are you checking your Y+ values at the walls?

[Bharatt]

Hey Casey,

Thanks for your comments. i'll keep that point in mind.
and yes i am checking for y+ values.

bharat

[Rami]

Smells like a Fluent simulation to me. If so - why don't you post it in the Fluent-dedicated forum?