CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Convergence Problem with two phase flow (

challenger85 December 29, 2009 08:01

Convergence Problem with two phase flow
Hi All

I am simulating now some two-phase air-water flow in different geometries (rectangular rod bundle geometry and square channel geometry).

Right now, i am just doing simple monodispersed calculations on that topic; the aim is the use of the inhomogeneous MUSIG model, which allows to specify multiple velocity and bubble size groups which is needed to account for the influence on the behavior having different bubble sizes (e.g. Lift Force reversal etc.).

My problem now is a quite general one: the value of the converged residuals, especially for the voulme fractions. I tried to avoid figure out the reason of the bad behavior, but i could nod find a "general" solution on this question.

Time is running now, and i am quite frustrated and helpless how to continue. I also changed different boundary conditions, models etc., of course also the investigated meshes, but there is no consens between the different things but the one, that the values are most probable not sufficient for a comparison between measurement and simulation.

My first question is:
ANSYS guide mentions, that solutions which converge below RMS residual vlaues around 10E-4 are quite sufficient. In the most of the cases, i can nod reach these values at all.

Also the hints given in the documentaion are not helping that much... Of course, all the things which were mentioned were carried out (as far as possible).

Is there a standard approach which you use to get better results? Just some points would be helpful (in fact, they would bring hope, and hope is always a good thing).

Of course, I can also be more precise on the models used for the investigations, if you have some suspicion on possible reasons on that topic.

Torsten Betschart

P.S.: just one question about the outlet condition: I tried to apply an outlet condition to the problem as standard (means outlet b.c. using specified static pressure). The ANSYS solver mentioned afterwards some upcoming cross flow and recommended the use of an opening condition. Is that the way its done (i did like this...).

All times are GMT -4. The time now is 03:40.