|
[Sponsors] |
May 16, 2008, 06:00 |
Convergence Problem in Star-CCM+
|
#1 |
Guest
Posts: n/a
|
Hi guys,
I have a heat transfer model consist of fluid and solid. I used the segregated solid and fluid models for simulation - solution converged but temperature would increase if I continued the simulation! in other words, Energy residual is a straight line and if I continue the simulation temperature of solid and fluid will increase slowly. They should stop after convergence, don't they?! Does this make sense?! I switched to Coupled solver and solution doesn't converge at all any help/comment?! |
|
May 16, 2008, 07:13 |
Re: Convergence Problem in Star-CCM+
|
#2 |
Guest
Posts: n/a
|
Well don't rely on residuals for convergence, they aren't the be all and end all of judging convergence. They are a good indicator of numerical convergence but it doesn't necessarily mean that your engineering values are converged, think about the simulation as an experiment, you wouldn't stop the test until the readouts you are interested in are stable would you....
|
|
May 16, 2008, 08:19 |
Re: Convergence Problem in Star-CCM+
|
#3 |
Guest
Posts: n/a
|
Yes, you are absolutely right but in that case I have to wait for about 8 days to get the results that I expect!
Don't you think it's a long period of time for a model with about 0.5 million cells?! |
|
May 16, 2008, 08:24 |
Re: Convergence Problem in Star-CCM+
|
#4 |
Guest
Posts: n/a
|
Try changing the underelaxation for the solid to 1.0 that should make it converge quicker
|
|
May 16, 2008, 08:27 |
Re: Convergence Problem in Star-CCM+
|
#5 |
Guest
Posts: n/a
|
In the case of Segregate solver yes - what if I use Coupled Solver?!
|
|
May 16, 2008, 09:01 |
Re: Convergence Problem in Star-CCM+
|
#6 |
Guest
Posts: n/a
|
Well unless it is buoyancy driven flow (natural convection) or high speed you shouldn't need to use the coupled solver, if you do need to use the coupled solver make sure your courant number is set correctly.
|
|
May 16, 2008, 09:37 |
Re: Convergence Problem in Star-CCM+
|
#7 |
Guest
Posts: n/a
|
When you think that the fluid flow doesn't change any more (velocity, pressure and turbulence) just freeze these solvers and increase the time step to 1 second. It should help to speed up the calculation.
|
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
star-ccm+ convergence criteria | star | Siemens | 2 | January 14, 2009 04:57 |
the problem of running star-cd after pro-star | liu-jinsong | Siemens | 0 | November 20, 2008 20:58 |
convergence, star-cd | rsi | Siemens | 7 | October 2, 2007 05:35 |
random convergence failures (STAR-CD on Linux-PC) | Wolfgang Binsteiner | Siemens | 2 | December 29, 2000 12:08 |