CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Number of iterations (https://www.cfd-online.com/Forums/fluent/86905-number-iterations.html)

Nora April 5, 2011 15:02

Number of iterations
 
Dear All,

I'm working with a 3D model, a cylinder with tangential air inlet. I've meshed the model for 3 million cells. When I run the simulation, the number of iterations only arround 400 - 500 iterations to get a convergence. Is it possible something like this happen or there is a problem with my model?

Please help.

Thanks

-mAx- April 6, 2011 03:04

in solve/monitor/convergence...
disable the check convergence, and let iterate.
Consider a convergence history (monitoring) of your goal quantity (massflow at inlet, pressure at inlet or whatever)

Nora April 6, 2011 06:07

Thank you for your reply. How disable the check convergence make any different? Sorry I still don't understand. Please explain more, thanks.

-mAx- April 6, 2011 06:14

For instance if you want to compute the drop pressure in a valve (steady state): your residuals are falling down (converging), but if the monitoring of pressure at your inlet is still oscillating, then you don't have a converged solution (although all your residuals are below teh target you specified).

Nora April 6, 2011 06:52

Hi, thank you very much for your help. I think I understand now what do you mean by diasble the check convergence, and let it iterate. But how do I know when to stop? Or should I make the absolute criteria smaller, let say 10 to the -6? And one more question, how to the monitor the pressure at the inlet? Thanks again.

-mAx- April 6, 2011 06:58

you can monitor with surface/monitor...
And to stop, a nice CTRL-C will do it perfectly ;)

taff224 April 6, 2011 07:00

The definition of convergence can vary.

If the residuals have flattened off and are not changing over time then more iterations are not worth doing, you will not get any better convergence. Whether this is an acceptable level of convergence is a different matter.

The easiest measure of convergence is to get all the values of mass flux for the solution - then work out the difference between input and output. If this is less than 1% of the smallest individual input/output stream then you are probably at a sensible level of convergence.

Remember convergence doesn't mean correct, it just means the calculation has gone as far as it can with the setup you have.

Nora April 6, 2011 07:05

Thank you very much for all replies. Its really help.


All times are GMT -4. The time now is 06:43.