what to do when solution doesn't converge?
 

Hi everyone,

I'm using SST model. My problem is a steady state fully developed turbulent curved duct flow. Outlet is set as opening with zero pressure and inlet with the velocity profile. The solution seems to stop converging after a certain no. of iteration. Can the separation in flow cause the SST model to fail?

http://www.angelfire.com/ab7/cfdwork/RMS.jpg http://www.angelfire.com/ab7/cfdwork/RMS2.jpg http://www.angelfire.com/ab7/cfdwork/RMS3.jpg

Re: what to do when solution doesn't converge?
 

Hi Jenny

I doubt this is the case. I would suggest looking at your residuals in Post first to see where the maximum values are. I would then examine the resolution of your mesh - very often you may only be able to achieve a certain level of convergence with a certain level of grid refinement.

Johnson

p.s. do you have re-circulation at your outlet ?

Re: what to do when solution doesn't converge?
 

Nice residual curves! Try to define convergence! When I look at your residual curves I would be satisfied, I cant see your problem. Look in your outfile and check the global balances and see if they also are OK. (22 days left as a CFX user :)

Re: what to do when solution doesn't converge?
 

Hi Jenny,

This doesn't look too bad. You may just be moving through a slow transient, increasing your timestep may help accelerate things.

Regards, Robin

Re: what to do when solution doesn't converge?
 

Hi Jan,

22 days left? Are you moving across to the Dark Side?

Glenn

Re: what to do when solution doesn't converge?
 

Hi Glenn,

I have decided to sacrifice my self for a period and investigate what is going on at the Dark Side :) (21 days left)