CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   How to get a convergence solution? (https://www.cfd-online.com/Forums/fluent/62727-how-get-convergence-solution.html)

 Conan March 18, 2009 02:40

How to get a convergence solution?

In my case, I use Realizable k-e, and model is filled withTetrahedral grids (unstructure grid) but without any highly skewness.

The problem is, the continuity equation error does not drop down at 1.E-4, it is fluctuating around 5.E-3. The others are well.

Is there any methods to change the situation?

Thanks.

 panda March 18, 2009 04:24

You may choose to see if other global variables under interest converge or not. Or, you may try to locally refine the mesh where the fluction is too apparent.

 paka March 20, 2009 02:53

Try to increase number of internal iterations. If it is too low (this is relative), for example 20, after reaching 20 Fluent will assume this is the most accurate obtained solution and will proceed to the next iteration. Try to increase it to 40, but keep in mind the more iterations obviously the computational time will increase. Depending on your case I would say 60 is maximum, but again... depends what you are solving.

Krst

 Conan March 20, 2009 04:06

where is the pannel for changing the internal iteration? I can not find. How to change?

 paka March 20, 2009 04:28

In the top menu go to Solve, then Iterate.

I don't know how familiar you are with CFD. If you are not familiar with specifics of this panel, I would advise you to go to Fluent Documentation and read more about it: Try to find info about Iterate panel. This relates to theoretical basics of solution methods.

I hope this helps.

 Conan March 20, 2009 05:35

My calulation is steady-state. NOT unsteady. So,there is no the corresponding setting pannel. In >Solve>Iterate, we can only set the total iteration number for outer cycle, rather than internal cycle.

 All times are GMT -4. The time now is 23:49.