|
[Sponsors] |
February 23, 2006, 19:11 |
Setting time steps and no. of iterations
|
#1 |
Guest
Posts: n/a
|
Hi all,
I want to model an unsteady velocity condition. For this, is there any specific criteria to set up the number of iterations, time step size, max no. of iterations per time step etc? Or do we just play around, starting with a small number like 0.01 and then adjust accordingly? How do we evaluate which is the best combination of these numbers? One other question- I hooked my unsteady velocity profile to Fluent, and then interpreted it. But when I ran the case, I do not see the waveform on the screen (along with the velocity vectors, etc). Is there any reason for this? How do I get the waveform pattern to show up? Thanks in advance for any advice/ suggestions. Regards, Vidya |
|
February 23, 2006, 22:54 |
Re: Setting time steps and no. of iterations
|
#2 |
Guest
Posts: n/a
|
If you have no idea about the period of the oscillating flow than you can start with a time step and check whether the convergence is achieved in few iterations like 20-30.
|
|
February 24, 2006, 00:06 |
Re: Setting time steps and no. of iterations
|
#3 |
Guest
Posts: n/a
|
OK, Muhammad, Thanks........... I'll see how the solution behaves and then adjust accordingly.
|
|
February 24, 2006, 04:08 |
Re: Setting time steps and no. of iterations
|
#4 |
Guest
Posts: n/a
|
hii, the time step size should be taken as min size of a cell in the domain divided by average flow field velocity and regarding no of iterations u give any number like 1000etc max no of iterations start with more iterations like 100 and there is no problem if u give more iterations it will automatically switch to other time step when the solution is converged bye
|
|
February 25, 2006, 12:00 |
Re: Setting time steps and no. of iterations
|
#5 |
Guest
Posts: n/a
|
1- What is your reference velocity? 2- What is your reference length scale? These two will give you the reference time scale Your time step should be less than this time scale
|
|
February 25, 2006, 16:20 |
Re: Setting time steps and no. of iterations
|
#6 |
Guest
Posts: n/a
|
Ahmed,
I do not understand what you mean by reference velocity and reference length cale. My velocity is an unsteady defined by the UDF. It is actually a pulsatile flow waveform. So now, how do I calculate the time step and the number of oterations per time step? Thanks, Vidya |
|
February 25, 2006, 18:13 |
Re: Setting time steps and no. of iterations
|
#7 |
Guest
Posts: n/a
|
1- Reference Velocity What velocity would you use if you want to non-dimensionalise your governing equations? In your case, the peak value of the pulsating wave, am I right? 2- Reference Length The same comment as before, but select the proper one. 3- Number of iterations per time step this has been answered in a previous posting, here it is again use a number that will ensure convergence each time step. You see, if you use, say 1000, and it converges after, say 100 or 140 the solver will automatically switch to the next time step. Be aware that Fluent Documentation mention critical velocity, critical length and critical time instead of reference....
|
|
|
|