CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Time Steps ? (http://www.cfd-online.com/Forums/fluent/27972-time-steps.html)

Prateep Chatterjee October 20, 2000 14:14

Time Steps ?
 
Is it possible to have different flow and reaction time steps in a case ? I'm using Fluent 5.3, single step CH4-Air finite rate chemistry, laminar flow, convective losses to the atmosphere from the geometry and radiation. I need to have variable time steps because chemical and flow time scales in case of laminar flows are many orders of magnitude different.

Greg Perkins October 22, 2000 18:46

Re: Time Steps ?
 
Prateep,

if you get anywhere with this, please let me know. I have a similar problem with a chemical reaction in porous media.

Regards

Greg

A. Hassaneen October 23, 2000 04:39

Call for Maurizio Barbato
 
Maurizio, Please try to answer when you have time to read the above message.

Maurizio Barbato October 24, 2000 04:41

Re: Time Steps ?
 
Hi Prateep,

this is a very tough problem. I had to face it with an internal combustion engine where the time step I was using was limited by the chemistry (I was forced to use even 1.0e-7 seconds !). Using a time step "too" long combustion just became an explosion and temperatures went to the hell. Some time strong under-relaxation helped me, but was some rare occasion. In general I think that, unless you set up a sistem to make inner iterations for the chemistry, these phenomena will be the limit for the time step. I know that this means to wait for years in front to your screen, but up to now with Fluent I didn't succeded in having a different solution.

May be this is a challenge to rise for the Fluent people.

Ciao

Maurizio


A. Hassaneen October 24, 2000 06:25

Re: Time Steps ?
 
Maurizio, I think that Prateep needs to know how (technically) he can change the time step during calculation at certain time, am I correct Prateep?.

Prateep Chatterjee October 24, 2000 11:35

Re: Time Steps ?
 
Hi Maurizio & Hassaneen !

You are right Hassaneen. I'd like to know if I can have two independent time steps in Fluent. Like 10^-7 s for the chemistry and 10^-2 s for the flow. This way I can solve the species and energy eqns separately and couple them to the momentum and continuity after every 10^-2 s.

I hope there is a way of doing it, perhaps by writing a UDF. Maurizio, at the moment I'm using a large time step of 10^-4 s and it takes me days to get a solution out (unsteady 1 s data for a FFT). Also, yesterday I tried using 10^-5 s and found that the results come out to be quite different. The selection of the time step has been a headache. On one hand getting results out fast is of paramount importance (otherwise the advisor is not happy!), and on the other hand accuracy is lost.

Thanks anyways.

Prateep


All times are GMT -4. The time now is 13:55.