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 The information provided may be out of date.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.
Frequency domain study for Joule heating physics
Posted 2013年3月28日 GMT-4 18:20 Studies & Solvers Version 4.3a 1 Reply
Please login with a confirmed email address before reporting spam
I am trying to do a frequency domain study for the physics "Joule heating and thermal expansion". However, I realized that comsol does not support this study for the physics. When I select frequency domain as my study, I get a yellow triangle with an exclamation mark under the menu "physics and variables section" for the joule heating physics. Could anyone help me with a work around for this?
1 Reply Last Post 2013年3月29日 GMT-4 04:03
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I would rather say it's not that COMOSL doesnt support that solver set-up, but rather that COMSOL cannot decide "automatically" how to solve such coupled physics without that you end user decides how you want to couple the physics.
One can also categorise the physics after their order of time derivatives, Solid as ACDC RF and other have 2nd order time derivatives and are of the type "wave equations" you can imagine sustained standing waves. But diffusion type equations (including HT) are rather 1st order in time or arranged such that you cannot have "standing waves".
What one have to decide in a case of joule dissipation (ACDC) and heat equation (HT) is how to couple them, as HT act, in the frequency domain a an extremely "low pass filter" (not fully true in MEMS size due the surface to volume ratio) so normally one solves with two solvers, a frequency domain (or stationary and harmonic excitation) for the ac (respectively DC+ac) current EC and a time dependent transient, or stationary for the HT part, by using the RMS value of the frequencya domain solver results as loads.
So far I have noticed that "pre-cooked" multihysics in COMSOL are mixed in such that you do not split up the solver sequences when setting up your solver methodology (except for a few cases).
So to make the long stor much shorter: set up your model with separate physics, then you decide how to couple and the granularity of the interaction, it comes to the same as the multi-physics case, but with a bit more clicks, and it forces you to think really over how these physics interact, in your case.
But check carefully and validate the results as it's very easy to miss some couplings, when working in this way, just as when using multiphyiscs, the coupling hypothesis used might not be the way you intendend !
--
Good luck
Ivar
I would rather say it's not that COMOSL doesnt support that solver set-up, but rather that COMSOL cannot decide "automatically" how to solve such coupled physics without that you end user decides how you want to couple the physics.
One can also categorise the physics after their order of time derivatives, Solid as ACDC RF and other have 2nd order time derivatives and are of the type "wave equations" you can imagine sustained standing waves. But diffusion type equations (including HT) are rather 1st order in time or arranged such that you cannot have "standing waves".
What one have to decide in a case of joule dissipation (ACDC) and heat equation (HT) is how to couple them, as HT act, in the frequency domain a an extremely "low pass filter" (not fully true in MEMS size due the surface to volume ratio) so normally one solves with two solvers, a frequency domain (or stationary and harmonic excitation) for the ac (respectively DC+ac) current EC and a time dependent transient, or stationary for the HT part, by using the RMS value of the frequencya domain solver results as loads.
So far I have noticed that "pre-cooked" multihysics in COMSOL are mixed in such that you do not split up the solver sequences when setting up your solver methodology (except for a few cases).
So to make the long stor much shorter: set up your model with separate physics, then you decide how to couple and the granularity of the interaction, it comes to the same as the multi-physics case, but with a bit more clicks, and it forces you to think really over how these physics interact, in your case.
But check carefully and validate the results as it's very easy to miss some couplings, when working in this way, just as when using multiphyiscs, the coupling hypothesis used might not be the way you intendend !
--
Good luck
Ivar