CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > ANSYS > FLUENT

How pseudo-transient method helps convergence in fluent

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

Like Tree2Likes
  • 2 Post By LuckyTran

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   March 12, 2019, 12:37
Default How pseudo-transient method helps convergence in fluent
  #1
Senior Member
 
ali
Join Date: Oct 2009
Posts: 318
Rep Power: 13
alinik is on a distinguished road
Does anyone have any documentations discussing how pseudo-transient method helps convergence in fluent, from theoretical perspective?

Thanks,
alinik is offline   Reply With Quote

Old   March 12, 2019, 13:29
Default
  #2
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 3,681
Rep Power: 45
LuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura about
Do you have a specific question?

The pseudo-transient method is a form of implicit relaxation (as opposed to explicit relaxation). In explicit relaxation you solve the equation but limit the variable change a posteriori. In implicit relaxation, you don't solve the original equations but relaxed equations. The relaxed equations limit the variable change because you inject the current solution into the equations.

You can think of implicit relaxation as lowering the local cell flow courant number. You can think of a steady state solver as a transient solver with a very wonky but large (almost infinite) time-step and a transient solver as having a small finite time-step.

Last edited by LuckyTran; March 12, 2019 at 19:33.
LuckyTran is offline   Reply With Quote

Old   March 12, 2019, 16:23
Default
  #3
Senior Member
 
ali
Join Date: Oct 2009
Posts: 318
Rep Power: 13
alinik is on a distinguished road
Thanks LuckyTran.
Yes, I had seen that, but as you can see it does not give a good discussion why this would help.
Reading you comment was helpful.

Can we think of it the following way too?
Adding transient term, makes the equation hyperbolic and it is no longer elliptic, hence it would be easier for the solver to reach convergence.
Is this reasoning correct?
alinik is offline   Reply With Quote

Old   March 12, 2019, 19:32
Default
  #4
Senior Member
 
Lucky Tran
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 3,681
Rep Power: 45
LuckyTran has a spectacular aura aboutLuckyTran has a spectacular aura about
Quote:
Originally Posted by alinik View Post
Can we think of it the following way too?
Adding transient term, makes the equation hyperbolic and it is no longer elliptic, hence it would be easier for the solver to reach convergence
Not really. Hyperbolic/elliptic/parabolic is a classification referring to the characteristics in one dimension only (either only time or only space). Introducing a transient term doesn't change the characteristics in space.

Also the characteristics has nothing to do with convergence. A parabolic/hyperbolic/elliptic equation can all be solved. It also doesn't explain how under-relaxation relates to stability, which is what the pseudo-transient option does. The pseudo-transient solver applies implicit under-relaxation to a steady solver.

What's hidden-under-the-hood is that the solution update of a steady solver vs iteration has a correspondence in time. So that a steady solver can be interpreted as a transient solver with a very wonky time-step. It's not so wonky when you realize that it is related to the local flow Courant number of each cell. But rather than the entire solution advancing the same time-step, each cell advances at its own local time-step. You can experience this most readily if you use the COUPLED pressure-velocity coupling scheme which is available in steady or transient solvers and has a flow courant number as an input.


Implicit relaxation slows down the solution so that it advances in time slower. It reduces the effective time-step of the simulation. As I mentioned already, a steady solver is like transient solver with a wonky time-step. The pseudo-transient option makes this wonky time-step smaller.
alinik and Darko like this.
LuckyTran is offline   Reply With Quote

Reply

Tags
convergence, fluent, pseudo-transient

Thread Tools Search this Thread
Search this Thread:

Advanced Search
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 Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
looking for a smart interface matlab fluent chary FLUENT 20 November 6, 2017 10:12
Transient simulation in Fluent khaham FLUENT 13 March 15, 2016 09:18
Running UDF with Supercomputer roi247 FLUENT 4 October 15, 2015 14:41
Animation of Fluent transient data saisanthoshm88 FLUENT 2 December 18, 2012 10:51
Force can not converge colopolo CFX 13 October 4, 2011 23:03


All times are GMT -4. The time now is 15:51.