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.
modal analysis - to large displacement
Posted Sep 18, 2009, 3:58 p.m. EDT 2 Replies
Please login with a confirmed email address before reporting spam
Hello.
I think I have a simple problem, but I do not really know how to solve this issue. The problem is as follows. I am simulating a micromechanical structure (MEMS Module - Solid, Stress-Strain). The geometry is in 1e-6. After defining material properties (Silicon), meshing, boundary conditions (fixed suspensions of the springs), a modal analysis is carried out. The eigenfrequency of the result is very close to the expected value, but the calculated displacement of the movable structure is toooo large. How can it be, that a micrometer scale suspendet structure can have a displacement in a 2m range?
Does anyone of you had a similar problem and could help me how to solve this problem or could anyone give me a hint what I made wrong or forgot?
Many thanks in advance for helping me.
I need this simulation for my master thesis.
I think I have a simple problem, but I do not really know how to solve this issue. The problem is as follows. I am simulating a micromechanical structure (MEMS Module - Solid, Stress-Strain). The geometry is in 1e-6. After defining material properties (Silicon), meshing, boundary conditions (fixed suspensions of the springs), a modal analysis is carried out. The eigenfrequency of the result is very close to the expected value, but the calculated displacement of the movable structure is toooo large. How can it be, that a micrometer scale suspendet structure can have a displacement in a 2m range?
Does anyone of you had a similar problem and could help me how to solve this problem or could anyone give me a hint what I made wrong or forgot?
Many thanks in advance for helping me.
I need this simulation for my master thesis.
2 Replies Last Post Oct 15, 2009, 6:27 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi
if you are doing a eigenfrequency analysis, the frequencies are what you are looking for, the displacement sare arbitrary, that is are normalised in different ways, Comsol is using it's way, other FEM tools have different norms. So the 2[m] displacement is not really relevant.
Normally, one should look at the "mode participation factors" to see the importance of a given mode w.r.t. a given coordinate variable, but this requires a specific normalistion of the eigenvectors (common in NASTRAN and ANSYS).
This is no:1 on my list of lack/missing features for use of structural analysis in COMSOL.
For simple geometries its possible to renormalise the matrices in Matlab, but for MEMS and any more complex geoemtry than a canteliever this gets far too heavy, slow and error prone.
I hope COMSOL will add the mass partcipation factors in a futur release, for me they are mandatory to be able to tell and sort out the modes, the absolute displacement given is just a nice to have.
By adding the "effective masses" of the modes you consider and relate them to the total mass of the object you also have a tool to say you are considering most of the important modes, without this you are purely guessing.
If you are interested in amplitudes you must be sure you have damping values correctly set, and that you input a given noise spectra as a force or an imposed displacement, with a limited energy, if not, at any resonance your mode amplitude should go to infinity
I hope this brings you somewhat further.
Good luck
Ivar
if you are doing a eigenfrequency analysis, the frequencies are what you are looking for, the displacement sare arbitrary, that is are normalised in different ways, Comsol is using it's way, other FEM tools have different norms. So the 2[m] displacement is not really relevant.
Normally, one should look at the "mode participation factors" to see the importance of a given mode w.r.t. a given coordinate variable, but this requires a specific normalistion of the eigenvectors (common in NASTRAN and ANSYS).
This is no:1 on my list of lack/missing features for use of structural analysis in COMSOL.
For simple geometries its possible to renormalise the matrices in Matlab, but for MEMS and any more complex geoemtry than a canteliever this gets far too heavy, slow and error prone.
I hope COMSOL will add the mass partcipation factors in a futur release, for me they are mandatory to be able to tell and sort out the modes, the absolute displacement given is just a nice to have.
By adding the "effective masses" of the modes you consider and relate them to the total mass of the object you also have a tool to say you are considering most of the important modes, without this you are purely guessing.
If you are interested in amplitudes you must be sure you have damping values correctly set, and that you input a given noise spectra as a force or an imposed displacement, with a limited energy, if not, at any resonance your mode amplitude should go to infinity
I hope this brings you somewhat further.
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi again
I would to propose some good litterature on modal analysis (and modal mass participation factors, see §4.5)
"Vibration Dynamics and Control", by Gienacarlo GENTA, Springer, 2009, ISBN 978-0-387-79579-9
Have a nice reading ;)
Ivar
I would to propose some good litterature on modal analysis (and modal mass participation factors, see §4.5)
"Vibration Dynamics and Control", by Gienacarlo GENTA, Springer, 2009, ISBN 978-0-387-79579-9
Have a nice reading ;)
Ivar