CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   converge problem (http://www.cfd-online.com/Forums/fluent/31786-converge-problem.html)

Li Jingping July 7, 2003 23:10

converge problem
 
Hi,

I encounter a very strange problem in my simulation. I made several models each of them are similar to the cylinder or tube. However, I found it is most difficult for FLUENT to converge for the most common cylinder,i.e. no bend, no curvature, just a cylinder with velocity inlet on one end and outflow on the other end. After 200 iterations for steady flow, it can only converge to 3*10^-4.

Because limited insight to the CFD, I am confused with it. Anybody can explain this phenomenon for me ?

thanks a lot.

Best, Li Jingping

mateus July 8, 2003 01:23

Re: converge problem
 
Hi!

Jou must realize that value of residuals is not nesseserily a good indicator for convergence. It is possible that you have a good initial sollution (in noncurved cilinder flow, that is not so hard to do), that means residuals can not get any smaller. Also I think that 3*10^-4 is allready a good convergence criteria. try to monitor pressure or velocity on outlet, or at a certain point - you wiil get a converged solution when value doesnt change any more... It may be also wise to define outlet bc as pressure outlet - it ussually works better...

Regards

MATEUS


Li Jingping July 8, 2003 02:54

Re: converge problem
 
thanks a lot!


All times are GMT -4. The time now is 17:54.