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.

Imaginary eigenfrequencies despite no damping in the model?

Please login with a confirmed email address before reporting spam

I have created a very simple model of a cylinder in the solid mechanics module and have run an eigenfrequency study on it. There is no damping whatsoever in my model so I'm very confused about why I'm getting imaginary eigenfrequencies in the results. Can anything other than damping produce imaginary eigenfrequencies?

7 Replies Last Post 30.06.2017, 16:00 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 28.06.2017, 16:45 GMT-4
I've attached my model, if anyone is willing to have a look and tell me where the imaginary eigenfrequencies are coming from I'd greatly appreciate it.
I've attached my model, if anyone is willing to have a look and tell me where the imaginary eigenfrequencies are coming from I'd greatly appreciate it.


Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 28.06.2017, 17:05 GMT-4
Updated: 7 years ago 28.06.2017, 17:17 GMT-4
Henry,

your model is not fully defined. You need to add constraints to avoid translations and rotations of the cylinder.

To see the free eigenmodes without constraints, increase the 'Desired number of eigenfrequencies' above 6. The first 6 are the trivial ones.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Henry, your model is not fully defined. You need to add constraints to avoid translations and rotations of the cylinder. To see the free eigenmodes without constraints, increase the 'Desired number of eigenfrequencies' above 6. The first 6 are the trivial ones. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 28.06.2017, 18:05 GMT-4
Thank you Edgar for the prompt reply!
Thank you Edgar for the prompt reply!

Magnus Ringh COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 30.06.2017, 06:55 GMT-4
Hi,

Just to complement Edgar's reply: If you know the approximate eigenfrequencies, you can change the value in the "Search for eigenvalues around" field to a value other than 0.

From the "Eigenfrequency search method" list, you can also select "Region" to search for eigenvalues in a specified eigenfrequency range only.

As Edgar pointed out, for a 3D object, the first 6 eigenfrequencies are 0, corresponding to three translations and three rotations (rigid body modes), so what you see as small and possibly imaginary numbers are just "0 with some numerical noise".

Best regards,
Magnus Ringh, COMSOL
Hi, Just to complement Edgar's reply: If you know the approximate eigenfrequencies, you can change the value in the "Search for eigenvalues around" field to a value other than 0. From the "Eigenfrequency search method" list, you can also select "Region" to search for eigenvalues in a specified eigenfrequency range only. As Edgar pointed out, for a 3D object, the first 6 eigenfrequencies are 0, corresponding to three translations and three rotations (rigid body modes), so what you see as small and possibly imaginary numbers are just "0 with some numerical noise". Best regards, Magnus Ringh, COMSOL

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 30.06.2017, 12:56 GMT-4
This is great information Magnus, thank you.

A follow-up question: if the first 6 eigenfrequencies correspond to translational and rotational motion, which eigenfrequencies will correspond to longitudinal, transverse, and torsional modes of vibration? I have been trying to figure out how to automate comsol to extract a list of those three frequencies only, and tell me which frequency is which. Are they listed in a specific order? Or is there a way to run a global evaluation on the Derived Values node and select those specific eigenfrequencies as values to export to a table?

Thank you,
Henry
This is great information Magnus, thank you. A follow-up question: if the first 6 eigenfrequencies correspond to translational and rotational motion, which eigenfrequencies will correspond to longitudinal, transverse, and torsional modes of vibration? I have been trying to figure out how to automate comsol to extract a list of those three frequencies only, and tell me which frequency is which. Are they listed in a specific order? Or is there a way to run a global evaluation on the Derived Values node and select those specific eigenfrequencies as values to export to a table? Thank you, Henry

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 30.06.2017, 13:03 GMT-4
I've adjusted my model to display 15 eigenfrequencies, and I'm still imaginary results beyond the first 6 frequencies, and am getting ones that display a real and imaginary part despite no damping. I've attached the adjusted model here.

Possibly useful information: for this project my supervisor wants me to model a completely free solid cylinder to compare to our theoretical benchmark. Basically I am trying to replicate this comsol example but for a solid cylinder instead of a hollow one: www.comsol.com/model/eigenfrequency-analysis-of-a-free-cylinder-271

Thank you,
Henry
I've adjusted my model to display 15 eigenfrequencies, and I'm still imaginary results beyond the first 6 frequencies, and am getting ones that display a real and imaginary part despite no damping. I've attached the adjusted model here. Possibly useful information: for this project my supervisor wants me to model a completely free solid cylinder to compare to our theoretical benchmark. Basically I am trying to replicate this comsol example but for a solid cylinder instead of a hollow one: https://www.comsol.com/model/eigenfrequency-analysis-of-a-free-cylinder-271 Thank you, Henry


Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 30.06.2017, 16:00 GMT-4
Henry,

the imaginary components are already 3 - 4 orders of magnitude smaller than the real components. Try to use a finer mesh and the imaginary components will become even smaller or disappear.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Henry, the imaginary components are already 3 - 4 orders of magnitude smaller than the real components. Try to use a finer mesh and the imaginary components will become even smaller or disappear. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

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.