CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > CFX

Transient Simulations

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   July 9, 2013, 20:23
Default Transient Simulations
  #1
Senior Member
 
---------
Join Date: Oct 2010
Posts: 302
Rep Power: 8
saisanthoshm88 is on a distinguished road
Could some one please suggest on how to judge the convergence in transient simulations, I mean in a steady state simulation it is possible to judge the convergence using monitor points. But in a transient simulation, the monitor points also show a time dependent trend so what criteria could be used to judge convergence in transient simulations.

And is it always fine to initialize a transient simulation with a steady state simulation for better convergence.
__________________
Best regards,
Santhosh.
saisanthoshm88 is offline   Reply With Quote

Old   July 9, 2013, 23:49
Default
  #2
Senior Member
 
hmasenger's Avatar
 
hamed
Join Date: Apr 2009
Posts: 148
Rep Power: 8
hmasenger is on a distinguished road
yes.it is always a good way to run a steady state simulation before transient one to have an accurate initial condition.
for the convergence question, i know every time step should be converged just like a steady state run .you should let results in all sub steps converge to a constant value
hmasenger is offline   Reply With Quote

Old   July 10, 2013, 02:17
Default
  #3
Senior Member
 
---------
Join Date: Oct 2010
Posts: 302
Rep Power: 8
saisanthoshm88 is on a distinguished road
Hamed, thanks for your response but monitoring a quantity for each time step could be difficult because as time progresses each time step may even converge within 2 iterations
__________________
Best regards,
Santhosh.
saisanthoshm88 is offline   Reply With Quote

Old   July 10, 2013, 02:46
Default
  #4
Senior Member
 
hmasenger's Avatar
 
hamed
Join Date: Apr 2009
Posts: 148
Rep Power: 8
hmasenger is on a distinguished road
why don't you set the RMS convergence criteria to a small value like 10e-5 or 10e-6 to let the solver go further iterations to converge?
hmasenger is offline   Reply With Quote

Old   July 10, 2013, 03:02
Default
  #5
Senior Member
 
---------
Join Date: Oct 2010
Posts: 302
Rep Power: 8
saisanthoshm88 is on a distinguished road
No I think it's not a recommended practice, because the simulation then takes longer.
__________________
Best regards,
Santhosh.
saisanthoshm88 is offline   Reply With Quote

Old   July 10, 2013, 04:13
Default
  #6
Senior Member
 
hmasenger's Avatar
 
hamed
Join Date: Apr 2009
Posts: 148
Rep Power: 8
hmasenger is on a distinguished road
you can set maximum iteration loops to a specific number to prevent from consuming time in another word for example,by setting RMS to 10e-6 and a maximum iteration to 50 you can run the solver up to 50 iteration and maybe the RMS equal to 10e-4.I hope you get the point.
hmasenger is offline   Reply With Quote

Old   July 10, 2013, 05:43
Default
  #7
Senior Member
 
Join Date: Jul 2011
Location: Berlin, Germany
Posts: 120
Rep Power: 6
monkey1 is on a distinguished road
I would monitor a related value that is supposed to reach convergence.
E.g. when I simulate the flow and dispersion through a domain I always monitor the pressure difference between in and outlet. That one beeing related to the required result variables / flow solution and supposed to reach a "steady state".
monkey1 is offline   Reply With Quote

Old   July 13, 2013, 06:57
Default
  #8
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 10,954
Rep Power: 85
ghorrocks has a spectacular aura aboutghorrocks has a spectacular aura aboutghorrocks has a spectacular aura about
The concept for assessing convergence in transient simulations is the same as for steady state - you pick some important points and see how it converges with different variations on the thing you are controlling (mesh size, convergence tolerance, time step size etc). It is just in a transient simulation the thing you are assessing does not necessarily converge to a single value, but is a function of time. This complicates things a little but is still pretty easy to handle.

The technique I usually use is to run the transient simulation for long enough for the important physics to start doing something, but as short as possible. Let's say that is after 1s of simulated time. Then you do a mesh convergence study using the result of important monitor points at the 1s time step. Then you have a single value just like a steady state simulation to do a normal convergence assessment on (including Richardson extrapolation if you wish).
ghorrocks is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Transient simulations: how to tell its converged (I've read the FAQ & user guides!) RicochetJ CFX 10 March 12, 2013 10:02
Transient simulations with translational periodic flow kriver FLUENT 11 April 18, 2012 05:56
Best practice for transient simulations? siw CFX 5 October 30, 2010 05:45
How are a transient simulations performed?? Luk Main CFD Forum 0 October 19, 2007 10:09
URANS and Transient Simulations bob Main CFD Forum 0 October 1, 2003 03:54


All times are GMT -4. The time now is 03:56.