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 The information provided may be out of date.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.

Vacuum properties

Please login with a confirmed email address before reporting spam

Hi,

While doing eigenfrequency analysis on rf module of a cylinder, if the geometry consists of only a cylinder, the cylinder is assumed to be standing is vacuum. Can I change the surrounding medium to be say "water" or "glass" without increasing the computation volume size?

10 Replies Last Post 3 juil. 2014, 09:33 UTC−4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10 oct. 2012, 09:11 UTC−4
That's not correct. If the geometry is only a cylinder and nothing around, your simulation is confined to the cylinder and no fields around the cylinder are taken into account.
In order to include the space around and its properties you MUST define a respective domain. You can use PML (perfectly matched layer) to reduce the required size of the domain around your device.

Cheers Edgar
That's not correct. If the geometry is only a cylinder and nothing around, your simulation is confined to the cylinder and no fields around the cylinder are taken into account. In order to include the space around and its properties you MUST define a respective domain. You can use PML (perfectly matched layer) to reduce the required size of the domain around your device. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10 oct. 2012, 09:28 UTC−4
Thanks for the propt reply and correcting my mistakes.

But, doesn't this imply that one has to do a parametric sweep to find the modes of the cylinder i.e. taking an excitation port or working with a background field and solving for the scattered field? Am I correct on this?
Thanks for the propt reply and correcting my mistakes. But, doesn't this imply that one has to do a parametric sweep to find the modes of the cylinder i.e. taking an excitation port or working with a background field and solving for the scattered field? Am I correct on this?

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 11 oct. 2012, 01:00 UTC−4
Hi

I would answer : it all depends,
you can either do an eigenfrequency analysis, that solves for the energy modes, and rather easily,
or do a frequency domain sweep, hence solving a harmonic development of the equations (with a source of given amplitude and phase) and observe the modal response, which takes usually longer to solve

--
Good luck
Ivar
Hi I would answer : it all depends, you can either do an eigenfrequency analysis, that solves for the energy modes, and rather easily, or do a frequency domain sweep, hence solving a harmonic development of the equations (with a source of given amplitude and phase) and observe the modal response, which takes usually longer to solve -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 oct. 2012, 09:30 UTC−4
Thanks for that information. I've been doing the same actually.

I have another doubt regarding the eigenfrequency analysis method. Some of the solutions that are returned have complex frequencies. Why does this happen?
Thanks for that information. I've been doing the same actually. I have another doubt regarding the eigenfrequency analysis method. Some of the solutions that are returned have complex frequencies. Why does this happen?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 oct. 2012, 09:52 UTC−4
Eigenmodes with damping have complex Eigenfrequencies. So in the real world all modes should be complex. I don't know if COMSOL has a threshold for displaying the imaginary component. Or maybe the purely real Eigenfrequencies are assumed without damping, not sure here.

Cheers
Edgar
Eigenmodes with damping have complex Eigenfrequencies. So in the real world all modes should be complex. I don't know if COMSOL has a threshold for displaying the imaginary component. Or maybe the purely real Eigenfrequencies are assumed without damping, not sure here. Cheers Edgar

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 11 oct. 2012, 10:31 UTC−4
Hi

have have seen sometimes complex values (if very small imaginary parts for cases when I would have expected only real, so these are probably "eps" tricks leftover. Anyhow more than 5-6 digits with todays double floating is illusory, so very small values, compared to much larger related ones, also betwen real andimaginary, I always consider as numerical noise.

Binary representation of real number is a sampling theory with more holoes than continuum ;)

--
Good luck
Ivar
Hi have have seen sometimes complex values (if very small imaginary parts for cases when I would have expected only real, so these are probably "eps" tricks leftover. Anyhow more than 5-6 digits with todays double floating is illusory, so very small values, compared to much larger related ones, also betwen real andimaginary, I always consider as numerical noise. Binary representation of real number is a sampling theory with more holoes than continuum ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 oct. 2012, 11:16 UTC−4
This might be the case most probably as I'm solving on lossy media. I just wanted to confirm if this was the case. I donot know about COMSOL's threshold, but interestingly, it also give purely real eigenfrequencies.

Thanks a lot for all the responses ....
This might be the case most probably as I'm solving on lossy media. I just wanted to confirm if this was the case. I donot know about COMSOL's threshold, but interestingly, it also give purely real eigenfrequencies. Thanks a lot for all the responses ....

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11 oct. 2012, 11:18 UTC−4
Hi

yeah .. this was something I used to see earlier when I worked for a short while on COMSOL. Then, the imaginary part was quite small. so, I was not concerned. This time around I was working on lossy media, so, just wanted to confirm the reason for the large imaginary parts.

Thanks Ivar for the quick responses...
Hi yeah .. this was something I used to see earlier when I worked for a short while on COMSOL. Then, the imaginary part was quite small. so, I was not concerned. This time around I was working on lossy media, so, just wanted to confirm the reason for the large imaginary parts. Thanks Ivar for the quick responses...

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 juil. 2014, 05:29 UTC−4
Hello to everyone,
I have the same problem this discussion started with, and unfortunately I am quite new to the use of comsol.

For brevity, I want to solve an eigenfrequency analysis of a box filled with air, surrounded by water. How can I implement this kind of analysis?

I mean, how can I make the software understand that inside the box there is air( for which I do not need to calculate the eigenfrequencies) and water outside?

Can anyone help me?
Hello to everyone, I have the same problem this discussion started with, and unfortunately I am quite new to the use of comsol. For brevity, I want to solve an eigenfrequency analysis of a box filled with air, surrounded by water. How can I implement this kind of analysis? I mean, how can I make the software understand that inside the box there is air( for which I do not need to calculate the eigenfrequencies) and water outside? Can anyone help me?

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 juil. 2014, 09:33 UTC−4

Eigenmodes with damping have complex Eigenfrequencies. So in the real world all modes should be complex. I don't know if COMSOL has a threshold for displaying the imaginary component. Or maybe the purely real Eigenfrequencies are assumed without damping, not sure here.

Cheers
Edgar


I was assuming that, whenever there is no damping, the eigenfrequency calculation of Comsol should always return real values, since there is no damping, there are no shift of the locations of the nodes/peaks with respect to time. It may be an attention point if the imaginary part is greater than the real part though, or if you have no damping in your system but you are still getting imaginary values with small imaginary parts, which may indicate a poor (lossy) mesh.

/Onur
[QUOTE] Eigenmodes with damping have complex Eigenfrequencies. So in the real world all modes should be complex. I don't know if COMSOL has a threshold for displaying the imaginary component. Or maybe the purely real Eigenfrequencies are assumed without damping, not sure here. Cheers Edgar [/QUOTE] I was assuming that, whenever there is no damping, the eigenfrequency calculation of Comsol should always return real values, since there is no damping, there are no shift of the locations of the nodes/peaks with respect to time. It may be an attention point if the imaginary part is greater than the real part though, or if you have no damping in your system but you are still getting imaginary values with small imaginary parts, which may indicate a poor (lossy) mesh. /Onur

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.