CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Time step size, number of time steps and max iterations per time step (http://www.cfd-online.com/Forums/fluent/70746-time-step-size-number-time-steps-max-iterations-per-time-step.html)

 guido_88 December 4, 2009 06:07

Time step size, number of time steps and max iterations per time step

I'm modeling a flow in an open-open pipe with a sinusoidal UDF. My frequency is 2000 Hz. How I can set time step size, number of time steps and max iterations per time step ? I want that it resonate so i find that it has to be 8.43cm long and 1cm of diameter

 thecfduser December 4, 2009 13:34

Hi
Time step = 1/2*2000.
Number of time steps: u must continue ure simulation untill u obtain a steady (oscillatory of course) solution. U will need at least that ure fluid traverse the pipe for 2 times , i mean 2*V/L
Max iteration by time step: try to monitor variables...choose the max iter. as the number of iterations after what ure solution does not change anymore within a time step....(u can also add 1 iteration for security, this ure unsteadiness is not uniform)

If those condirtions are so strict (take a lot of time to finish ure simulation), use NITA solver: it will give u fairly accurate solution within 1 iter by time step

P.S be4 asking a question, u can say hello and at the end, u can say thanks no?

 guido_88 December 4, 2009 13:40

Thanks a lot,you're absolutely right !

 thecfduser December 4, 2009 14:11

Anyway i forgot to tell u, for the time step i told u 1/4000 to stay conform to shannon theorem...but if u have more unsteadiness, u must use a smaller time step.
URANS models are not sensible to high frequencies, so this will not be a big problem but if u are performing DNS or LES, u will need maybe to refine ure time step...
There is no universal formula: the best way is to perform ure simulations with 2 different time steps and compare so u can be sure.

With smaller time step, u will need less ietrations by time step. Sometimes using a smaller time step only influences rms values without influencing the mean field...

Good Luck