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 The information provided may be out of date.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 help needed.

Please login with a confirmed email address before reporting spam

I am attempting to do a frequency domain study in order to find what my resonant frequencies are. I keep encountering an error and do not know how to resolve it. the error reads:

"Failed to find a solution for the initial parameter.
Out of memory during LU factorization.
Returned solution is not converged."

it is a piezoelectric material sandwiched by a top and bottom electrode. The top is assigned a voltage while the bottom is ground. Please help, I have included the model in hopes that someone can see what i may be missing.


3 Replies Last Post Oct 6, 2012, 4:28 p.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 5, 2012, 1:30 a.m. EDT
Hi

I'm not by my WS so I cannot check your model, but if you use the frequency sweep to identify modes one often get into trouble with the limited damping that is in the model (by default, except if you add some explicitely) that tend to make the solver fail at the first sharp peak.
Another issue is that the peaks are far appart and often narrow and the solver must either take small steps to not jump over any, or you must know and guide it where the peaks are.

Therefore I always run an eigenfrequency scan, and then generate a set composed of a series of

{range(,,,) range(,,) ...}

lines for my frequency scan values,

Unfortunately, I do not know how I can easily define in COMSOL such ranges as "parameter" and quickly plot them out over an eigenfrequency scan to ensure I have enough points, so often I miss something and have to resolve, which is ennoying as it takes a long time

I have suggested some improvement here already a few years ago, this I have not seen arrive, but many other things yes (COMSOL developpers are really reactive on our suggestions),
Even a small script that detect the eigenmodes of a scan and proposes a time stepping could be made. It looks like the COMSOL developpers do not really model many true engineering cases of complex mode shape, under tighht time stress, which is our job, as end user ;)

--
Good luck
Ivar
Hi I'm not by my WS so I cannot check your model, but if you use the frequency sweep to identify modes one often get into trouble with the limited damping that is in the model (by default, except if you add some explicitely) that tend to make the solver fail at the first sharp peak. Another issue is that the peaks are far appart and often narrow and the solver must either take small steps to not jump over any, or you must know and guide it where the peaks are. Therefore I always run an eigenfrequency scan, and then generate a set composed of a series of {range(,,,) range(,,) ...} lines for my frequency scan values, Unfortunately, I do not know how I can easily define in COMSOL such ranges as "parameter" and quickly plot them out over an eigenfrequency scan to ensure I have enough points, so often I miss something and have to resolve, which is ennoying as it takes a long time I have suggested some improvement here already a few years ago, this I have not seen arrive, but many other things yes (COMSOL developpers are really reactive on our suggestions), Even a small script that detect the eigenmodes of a scan and proposes a time stepping could be made. It looks like the COMSOL developpers do not really model many true engineering cases of complex mode shape, under tighht time stress, which is our job, as end user ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 5, 2012, 10:45 a.m. EDT
So how would you plot impedance in 4.2. I have done a frequency domain sweep and it shows some displacement after I apply 5V to the terminals.
So how would you plot impedance in 4.2. I have done a frequency domain sweep and it shows some displacement after I apply 5V to the terminals.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 6, 2012, 4:28 p.m. EDT
Hi

a comment:
in your model, add a Definition View, then under the sub node Camera deselect Preserve aspect ratio, it helps for MEMS cases and it's easier to select in this view, then you can easily switch between the true and scaled view

For admittance, have you checked the COMSOL predefined parameters linked to the terminal port ?
See Results Derived Values - Global Evaluation - select variables (>< icon, or type CNTRL SPACE in the Expression field)


--
Good luck
Ivar
Hi a comment: in your model, add a Definition View, then under the sub node Camera deselect Preserve aspect ratio, it helps for MEMS cases and it's easier to select in this view, then you can easily switch between the true and scaled view For admittance, have you checked the COMSOL predefined parameters linked to the terminal port ? See Results Derived Values - Global Evaluation - select variables (>< icon, or type CNTRL SPACE in the Expression field) -- 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.