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 The information provided may be out of date.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.

Current in a coil, time domain, error

Posted 19 avr. 2012, 08:41 UTC−4 Low-Frequency Electromagnetics Version 4.2 3 Replies

Zhiqiu Lu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello,

I'm simulating a simple model according to the tutorial of induction currents from circular coils. I want to use square wave signal as the current in the coils and do the triansient analysis.

the signal is defined as,
Globe definition->waveform1

and in the single-turn coil domain, I changed the coil potential to "wv1"

in the Time domain study, there is alwayse an error indicating
Failed to evaluate variable Jacobian.
- Variable: wv1
- Geometry: 1
- Domain: 3 4

what's the problem here? The file is attached.
3 Replies Last Post 23 avr. 2012, 14:28 UTC−4
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 19 avr. 2012, 09:53 UTC−4
Hi

I'm not behind my COMSOL WS but I suspect, from reading you that yuo have "dirac" type spikes at the "square" transition of your current. If so, at these points COMSOL canot define any derivative for the Jacobian estimation, used for the solver process. Have you tried the "rect1()" operator with smooth transitions, start with long transition imes then reduce them, and perhaps you need to adapt the time stepping intervall too.

But it might be someting else too

--
Good luck
Ivar
Hi I'm not behind my COMSOL WS but I suspect, from reading you that yuo have "dirac" type spikes at the "square" transition of your current. If so, at these points COMSOL canot define any derivative for the Jacobian estimation, used for the solver process. Have you tried the "rect1()" operator with smooth transitions, start with long transition imes then reduce them, and perhaps you need to adapt the time stepping intervall too. But it might be someting else too -- Good luck Ivar
Zhiqiu Lu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 avr. 2012, 08:21 UTC−4
Hi Ivar,

thanks! I have to smooth it, then for a one-step time dependant analysis, it works.

However, it comes to another question. I have seen in one of the tutorials, "magnetic brake", stationary analysis is used before time dependant. And also in "Induction Currents from Circular Coils", a frequcency with time dependant. Besides, I have a file from someone with both stationary and time dependant analysises as step1 and step2 respectively.

So my question is can I just use time dependant analysis?
Hi Ivar, thanks! I have to smooth it, then for a one-step time dependant analysis, it works. However, it comes to another question. I have seen in one of the tutorials, "magnetic brake", stationary analysis is used before time dependant. And also in "Induction Currents from Circular Coils", a frequcency with time dependant. Besides, I have a file from someone with both stationary and time dependant analysises as step1 and step2 respectively. So my question is can I just use time dependant analysis?
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 23 avr. 2012, 14:28 UTC−4
Hi

As often, "it all depends". Often its good practive to run a stationary solver first to set all unknown/initial condition values to a knwn sate, and tehn start a smooth gentle time series solving sequence. If not you might have a cick start with quite some transient ringing, that even might avoid the solver to converge.

So if you know how you want to start your time series, a stationary case can often be beneficil to start from a known state

--
Good luck
Ivar
Hi As often, "it all depends". Often its good practive to run a stationary solver first to set all unknown/initial condition values to a knwn sate, and tehn start a smooth gentle time series solving sequence. If not you might have a cick start with quite some transient ringing, that even might avoid the solver to converge. So if you know how you want to start your time series, a stationary case can often be beneficil to start from a known state -- Good luck Ivar

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