CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   LES flow through time? (https://www.cfd-online.com/Forums/main/14995-les-flow-through-time.html)

kasim March 28, 2008 08:46

LES flow through time?
 
Hi

I am running a LES case on a duct with repeated obstcales inside. I have initialised my case with RANS model. I was wondering how many flow through times (residence time) would I need to obtain converged statistics. What excatly would indiacte that my flow has achieved stability and I do not need to run it any longer. I know its a very basic question but I dont wanna get it wrong. Thank you

kasim

Tedy March 28, 2008 21:07

Re: LES flow through time?
 
Hello Kasim, Typically you need about 5-10 flow-through time to obtain a good initial LES solution, then you should run a similar time to accumulate you data for LES statistics. This will give you statistically converged LES results.

Success!

Ted

kasim March 29, 2008 07:33

Re: LES flow through time?
 
Thank you Ted for your answer. I am collecting statistics in my current run. Do you mean I should disable these statistics and then run the problem with new statistics for 5-10 flow through times.

Kasim

kasim March 29, 2008 13:31

Re: LES flow through time?
 
One more thing, when I re-start the simulation should I interpolate the existing solution on new run OR should I continue running existing solution and only remove the history from old statistics. Thank you

kasim

Ted March 29, 2008 14:11

Re: LES flow through time?
 
Hello Kasim,

Yes, you should first run to obtain a statistically converged LES solution (5-10 flow through is an empirical value based on experimenting), THEN you should start accumulating your statistics, that is if you want to look say at your flow mean and turbulent quantities, ex: Ui_av, Ui'_rms, Ui'Uj'_av, etc.

Best, Ted


kasim March 29, 2008 15:06

Re: LES flow through time?
 
Thank you Ted. In order to collect new statistics should I just disable my current statistics and continue with the LES run OR should I interpolate my existing solution on new run to collect new statistics?

kasim

Ted March 29, 2008 15:31

Re: LES flow through time?
 
Kasim, Just continue your LES run, and when you consider that your initial LES solution is ok (5-10 flow through's), start accumulating LES statistics. Before you LES becomes fully statistically converged, the statistics you accumulate are more-or-less meaningless. Good luck! Ted

kasim March 29, 2008 18:41

Re: LES flow through time?
 
Thanks Ted. Have a nice week end.

Kasim


All times are GMT -4. The time now is 16:49.