Switching transient and steady or having an initial phase

Numerical methods and mathematical models of Elmer
Post Reply
Franz Pichler
Posts: 196
Joined: 29 Sep 2011, 12:25
Antispam: Yes

Switching transient and steady or having an initial phase

Post by Franz Pichler »

Hello my dear Elmer friends,

I was wondering if there is a way to switch between steady and transient simulation.

The idea is that a solver could need an initialization (think of a mix of elliptic and parabolic equations that give DAE instead of ODEs in the transient case).
This initialization could should be iterating until steady state is reached and than we switch to normal transient behaviour.

I ma not sure if the solver can have an initialization function.

any hints are highly appreciated,

thanks
best regards
and a nice weekend
Franz
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Switching transient and steady or having an initial phase

Post by mzenker »

Hi Franz,

what you can do is to save the steady state output in a file and use that as restart file for the transient simulation (see "Restart from existing solutions" in the Solver Manual). That is what I have done in a similar case.
Thinking about it, there may even be some possibility to automate that using the RUN keyword (see "Running several sequences" in the Solver Manual), but I have never tried it.

HTH,

Matthias
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Switching transient and steady or having an initial phase

Post by raback »

Hi Franz

I just added a keyword > Steady State Condition < to the Solver section. When it is positive the solver is enforced to run in steady-state even though simulation would otherwise be transient. There is a corresponding test case named "heateq_steady_start" that provides a working example.

-Peter
Franz Pichler
Posts: 196
Joined: 29 Sep 2011, 12:25
Antispam: Yes

Re: Switching transient and steady or having an initial phase

Post by Franz Pichler »

wow thats awseome, thanks.
looking forward to try it
nguyent
Posts: 32
Joined: 17 Apr 2012, 21:19
Antispam: Yes

Re: Switching transient and steady or having an initial phase

Post by nguyent »

I just did an svn update to rev. 6035. I didn't see any difference adding 'Steady State Condition = Logical True' to the solver section. Then I went to look for heateq_steady_state in tests and didn't find it.

Not a big deal, since I'm already doing what Matthias suggested. I was just wondering what was going on.
tzwinger
Site Admin
Posts: 99
Joined: 24 Aug 2009, 12:20
Antispam: Yes

Re: Switching transient and steady or having an initial phase

Post by tzwinger »

Hello,
I think we are way past revision 6035 (check out http://sourceforge.net/p/elmerfem/code/HEAD/tree/) - actually Peter's commit was r6158. Perhaps something wrong with your subversion settings?

Best Regards,
Thomas
nguyent
Posts: 32
Joined: 17 Apr 2012, 21:19
Antispam: Yes

Re: Switching transient and steady or having an initial phase

Post by nguyent »

Ah ha! Thanks for the update - I was still using https://elmerfem.svn.sourceforge.net/svnroot/elmerfem.
Pisolino
Posts: 43
Joined: 11 Jul 2011, 11:32
Antispam: Yes

Re: Switching transient and steady or having an initial phase

Post by Pisolino »

nguyent wrote: 01 May 2013, 00:44 I just did an svn update to rev. 6035. I didn't see any difference adding 'Steady State Condition = Logical True' to the solver section. Then I went to look for heateq_steady_state in tests and didn't find it.

Not a big deal, since I'm already doing what Matthias suggested. I was just wondering what was going on.
Dear elmer users, is this option still available ?

i have a calculated field named A that i would like to reuse in a MATC equation ( f(A) ), such field is steadystate since the field doesn't vary during time but only in coordinates (XYZ).
The simulation that will reuse such A field (as constant... it will not change its value) will be transient.


Andrea
Post Reply