Menu

Discussion Forum

New Discussion
Filtra

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 The information provided may be out of date.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.

Eigenfrequency seems right, but displacement units are totally wrong.

Posted 7 feb 2013, 23:18 GMT-5 MEMS & Nanotechnology, MEMS & Piezoelectric Devices, Structural Mechanics Version 4.1, Version 4.2, Version 4.2a, Version 4.3, Version 4.3a, Version 4.3b, Version 4.4 14 Replies

Erik Bauch
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Dear Comsol Community,

in the model attached I have a 20 micron long beam using the Solid Mechanics module to get the Eigenfrequencies. I calculated them manually using Euler-Beam theory and the first Eigenfrequency should be around 5 MHz.. Comsol calculates 4.8 MHz. I thought great result, but if you look carefully at the screenshot, you will notice that Comsol claims the displacement is on the order of meters, which is totally wrong (it is rather pico meters). I feel I am missing something fundamentally, but I cannot get Comsol to spit out the correct displacement as well. (Scaling or anything of that sort is disabled).

Thanks for advices!

Erik

Attachments:

14 Replies Last Post 26 gen 2015, 18:49 GMT-5
Nagi Elabbasi Facebook Reality Labs
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 feb 2013, 00:12 GMT-5
Hi Erik,

In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity.

Nagi Elabbasi
Veryst Engineering
Hi Erik, In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity. Nagi Elabbasi Veryst Engineering
Erik Bauch
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 feb 2013, 13:34 GMT-5

Hi Erik,

In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity.

Nagi Elabbasi
Veryst Engineering


Thank you! With your help I found a lot of more information on the forum about this and I think I do get it now.

:)
[QUOTE] Hi Erik, In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity. Nagi Elabbasi Veryst Engineering [/QUOTE] Thank you! With your help I found a lot of more information on the forum about this and I think I do get it now. :)
Amey Revandkar
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12 feb 2013, 02:16 GMT-5
Respected sir,
How to do modal analysis on comsol4.2a with simulation of electrostatic voltages as loads for respective geometries of beam>
Moreover can you help me out in knowing the path to analyse stress results on beam.
Regards!!!
Respected sir, How to do modal analysis on comsol4.2a with simulation of electrostatic voltages as loads for respective geometries of beam> Moreover can you help me out in knowing the path to analyse stress results on beam. Regards!!!
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 12 feb 2013, 11:12 GMT-5
Hi

you could try a stressed modal analysis, use the static force of the ACDC and Solid (stationary solver) as initial conditions for your solid modal analysis (assuming the oscillation is small and not influencing the ACDC force values
Check the doc of the prestressed modal analysis, or see by similarity on the "buckling load that is _almost_ identical in settings.

--
Good luck
Ivar
Hi you could try a stressed modal analysis, use the static force of the ACDC and Solid (stationary solver) as initial conditions for your solid modal analysis (assuming the oscillation is small and not influencing the ACDC force values Check the doc of the prestressed modal analysis, or see by similarity on the "buckling load that is _almost_ identical in settings. -- Good luck Ivar
Amey Revandkar
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 feb 2013, 08:45 GMT-5
Thanks sir,
i'M NOT ABLE TO FIND THE PRESTRESSED MODAL ANALYSIS OF CANTILEVER. IN WHICH MODULE I CAN FIND IT? IS IT AVAILABLE IN COMSOL4.2A VERSION?

REGARDS!!
Thanks sir, i'M NOT ABLE TO FIND THE PRESTRESSED MODAL ANALYSIS OF CANTILEVER. IN WHICH MODULE I CAN FIND IT? IS IT AVAILABLE IN COMSOL4.2A VERSION? REGARDS!!
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 13 feb 2013, 08:53 GMT-5
Hi

have you searched your doc, in particular the Structural user manual for "Prestressed Analysis" ?
There is quite a lot in my 4.3a

--
Good luck
Ivar
Hi have you searched your doc, in particular the Structural user manual for "Prestressed Analysis" ? There is quite a lot in my 4.3a -- Good luck Ivar
Amey Revandkar
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 feb 2013, 07:06 GMT-5
hi,
Thanks for the reponse.
i have located and read the [prestressed analysis,frequency] subnode data on my 4.2a version.

i STILL DOUBT HOW TO USE IT TO FIND THRESCHOLD dc Voltage at which cantilever vibrates?

Ihave obtain simulation results for fixed-fixed beam .
Following conclusions i draw from that data,

1) If width is varied keepng length and thickness constant , there is no significant change seen in eigenfrequency of beam.

2) If length of the beam is varied keeping width and thickness constant the eigenfrequency has a significant change.

May i know if these observations are authentic?

Regards
hi, Thanks for the reponse. i have located and read the [prestressed analysis,frequency] subnode data on my 4.2a version. i STILL DOUBT HOW TO USE IT TO FIND THRESCHOLD dc Voltage at which cantilever vibrates? Ihave obtain simulation results for fixed-fixed beam . Following conclusions i draw from that data, 1) If width is varied keepng length and thickness constant , there is no significant change seen in eigenfrequency of beam. 2) If length of the beam is varied keeping width and thickness constant the eigenfrequency has a significant change. May i know if these observations are authentic? Regards
Nagi Elabbasi Facebook Reality Labs
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 feb 2013, 08:19 GMT-5
Hi Amey,

The observations you stated are correct. For “thin” (Euler-Bernoulli) beams the bending natural frequencies do not vary with beam width and vary with the inverse square of the beam length.

Nagi Elabbasi
Veryst Engineering
Hi Amey, The observations you stated are correct. For “thin” (Euler-Bernoulli) beams the bending natural frequencies do not vary with beam width and vary with the inverse square of the beam length. Nagi Elabbasi Veryst Engineering
Amey Revandkar
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22 feb 2013, 03:19 GMT-5
thanks sir,
may i know the path in comsol4.2a version for simulating the actuation voltage at which cantilever will resonate at its natural frequency? i'm trying it in prestressed analysis,eigenfrequency study node.
thanks sir, may i know the path in comsol4.2a version for simulating the actuation voltage at which cantilever will resonate at its natural frequency? i'm trying it in prestressed analysis,eigenfrequency study node.
Nicolas Tsagarakis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 feb 2014, 17:18 GMT-5

Hi Erik,

In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity.

Nagi Elabbasi
Veryst Engineering


Hello Nagi,

I understand that, but what if you wanted to find out the actual displacement? Whould you do it elsewise rather than extracted it from the mode shapes, I suppose?

Regards,
Nicolas T.
[QUOTE] Hi Erik, In a mode shape the relative magnitude of displacements matter not the absolute value. If you scale the mode shape by any factor it is still a mode shape. COMSOL selects by default RMS (root mean square) normalization to scale mode shapes and that results in displacements of the order of unity. Nagi Elabbasi Veryst Engineering [/QUOTE] Hello Nagi, I understand that, but what if you wanted to find out the actual displacement? Whould you do it elsewise rather than extracted it from the mode shapes, I suppose? Regards, Nicolas T.
Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 feb 2014, 18:04 GMT-5

There is no such thing as actual displacement in an Eigenmode study and the respective shape. It is a mathematical property of the model.
If you want to study displacement you have to run a frequency domain study and you have to excite the model some way according to your application.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
There is no such thing as actual displacement in an Eigenmode study and the respective shape. It is a mathematical property of the model. If you want to study displacement you have to run a frequency domain study and you have to excite the model some way according to your application. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com
Nagi Elabbasi Facebook Reality Labs
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 feb 2014, 18:07 GMT-5
Hi Nicolas,

You should do a frequency domain analysis. In that case you put the harmonic loads (magnitude and frequency) and COMSOL calculates the harmonic displacements with magnitude (and phase if necessary).

Nagi Elabbasi
Veryst Engineering
Hi Nicolas, You should do a frequency domain analysis. In that case you put the harmonic loads (magnitude and frequency) and COMSOL calculates the harmonic displacements with magnitude (and phase if necessary). Nagi Elabbasi Veryst Engineering
Nicolas Tsagarakis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 feb 2014, 03:36 GMT-5
Thanks a lot guys. You've been very helpful.

Regards,
Nicolas T.
Thanks a lot guys. You've been very helpful. Regards, Nicolas T.
Marc Serra
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 26 gen 2015, 18:49 GMT-5
I don't fully agree on that. Comsol does output displacements (and stresses, and strains) in eigenmode analysis, which I guess they correspond to a mode coordinate of 1, and a mode profile normalized depending on what normalization you choose.

What I find very confusing is that the "mode shape" plot and the normalized output don't seem to match. There should definitely be a way to plot the mode profile corresponding to the mode coordinate of your choice.
I don't fully agree on that. Comsol does output displacements (and stresses, and strains) in eigenmode analysis, which I guess they correspond to a mode coordinate of 1, and a mode profile normalized depending on what normalization you choose. What I find very confusing is that the "mode shape" plot and the normalized output don't seem to match. There should definitely be a way to plot the mode profile corresponding to the mode coordinate of your choice.

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