CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Difference in result due to time step (http://www.cfd-online.com/Forums/cfx/79390-difference-result-due-time-step.html)

prateek August 23, 2010 03:32

Difference in result due to time step
 
Hi all,

In high resolution scheme steady state, when I changing the time step results or different.
Why it is happening ?

ghorrocks August 23, 2010 07:37

Are you asking when you change the time step size in a steady state simulation why are the results different?

prateek August 23, 2010 23:50

hi,

Yes change in result due to change in time step

ghorrocks August 24, 2010 06:31

It should not make a difference. If the time step gets small enough the Reynolds Averaging gets unreliable, but as long as you are larger than a certain amount it should not make any difference.

prateek August 24, 2010 06:47

Thanks.
How I can adjust/ decide the time step.

ghorrocks August 24, 2010 18:50

Do the tutorials, read the documentation. It is all discussed there.

CCTech_Pune August 25, 2010 08:04

Hi

As said above

Your results should not change with change in the timestep for steady state. Please check if your simulations are well converged before comparing two results.

1 Residuals at least 10-3
2 Imbalances less than 1%
3 Monitored point is stable with respect to iterations

Thanks
CCTech Pune

prateek August 25, 2010 23:18

Hi,

Thanks,
All three points are achived.

But results are different.

What should I do?

ghorrocks August 26, 2010 06:44

Usually this is because your time step is too small. Your time step needs to be big enough to "average" out the turbulent fluctuations into a mean flow. To do this, your time step needs to be larger than the eddy turnover time of the largest eddies.

So, in short, increase the time step size until your result converges.

prateek August 26, 2010 23:26

Hi, Thanks

Can you give me one example for that.

ghorrocks August 27, 2010 06:44

This is a general principle learnt from doing CFD for decades. I can't give you a specific example for it. Just trust me. Or even better, fiddle with the time step and see what happens - then you will prove it to yourself.

prateek August 30, 2010 00:03

Thanks I will try it.


All times are GMT -4. The time now is 20:40.