Menu

Discussion Forum

New Discussion
Filtrer

Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Solution not converged

Posted 26 avr. 2012, 13:17 UTC−4 RF & Microwave Engineering, Studies & Solvers Version 3.5a, Version 4.3 6 Replies

Pu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi, all

Recently I'm encountering convergence problems. Physically the problem has decent solution. So I assume the problem lies in numerics. However it does not seem easy to find the reason for the divergence...

The complication of the problem may be one boundary condition. It's a model coupling an RF module (2D, solving Hz) and a PDE. In the PDE module, a boundary condition has electric fields in the expression. For some direct solver, and a particular value of a parameter in the model, it can converge. When the parameter takes other values, it always diverge, leading to non-physical solution with several singularly large values.

For some simple iteration methods, divergence may be avoided by tuning certain parameters. Similarly can we change something in the solver settings to secure convergence?

Thanks a lot!
6 Replies Last Post 7 mai 2012, 05:25 UTC−4
Pu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 4 mai 2012, 05:48 UTC−4
Does anyone have clue? I'm still struggling with this problem.

Thanks!
Does anyone have clue? I'm still struggling with this problem. Thanks!
Pu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 4 mai 2012, 05:54 UTC−4
Since I'm sure the physical setting has proper solutions, I suppose there may be two sources of the divergence:

1. The way to assign the boundary condition.

2. The solver itself.

I'm using version 3.5a. Is it possible the new versions give converged solutions? Thinking about try 4.2...
Since I'm sure the physical setting has proper solutions, I suppose there may be two sources of the divergence: 1. The way to assign the boundary condition. 2. The solver itself. I'm using version 3.5a. Is it possible the new versions give converged solutions? Thinking about try 4.2...
Jim Freels mechanical side of nuclear engineering, multiphysics analysis, COMSOL specialist
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 5 mai 2012, 23:11 UTC−4
It is difficult to help you without knowing more about the problem. You might supply your model.

Yes, v4.2a is much improved over v3.5a. That is why it was released. Version 4.3 is coming out soon.
It is difficult to help you without knowing more about the problem. You might supply your model. Yes, v4.2a is much improved over v3.5a. That is why it was released. Version 4.3 is coming out soon.
Pu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 6 mai 2012, 12:54 UTC−4
Hi, James

Thanks a lot for your reply!

I only have 3.5a version model files. Is it ok if I upload a 3.5a file?
Hi, James Thanks a lot for your reply! I only have 3.5a version model files. Is it ok if I upload a 3.5a file?
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 6 mai 2012, 13:52 UTC−4
Hi

as often, newer versions can read older ones (but not earlier than 3.5a), but 3.5a cannot read v4 files, unfortunately

--
Good luck
Ivar
Hi as often, newer versions can read older ones (but not earlier than 3.5a), but 3.5a cannot read v4 files, unfortunately -- Good luck Ivar
Pu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 mai 2012, 05:25 UTC−4
Hi, Ivar and James

Thanks for your replies!

Here's my model file attached. I modified subdomain and boundary equation systems to introduce coupling and boundary condition.

Thanks!
Hi, Ivar and James Thanks for your replies! Here's my model file attached. I modified subdomain and boundary equation systems to introduce coupling and boundary condition. Thanks!

Attachments:

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content