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.

Sensitivity Eigenvalues analysis to geometric model parameter

Please login with a confirmed email address before reporting spam

Dear

I would like to create a Sensitivity Comsol Model of a Structural Mechanics structure in order to evaluate the sensitivity, who the eigvalues of the structure could change (in absolute/relative) to some variation of the geometric , material inputs.


I am not very familiar of Sensitivity module but the most important step is the definition the "Objective function".

The natural frequency of simple beam structure, or plate coulbe are evaluated using analytical function. For example the beam eingftequencies are proportional to sqrt(E*I/(rho*A*L^4)) with E= Young's models, rho = density, A= section Area of the beam, L= lenght of the beam, I= inertia section.

For a genral FEM model the Objective function is no so obvious. New ideas are welcome.

Michele

2 Replies Last Post 7 juin 2012, 09:38 UTC−4
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 7 juin 2012, 02:36 UTC−4
Hi

hat about using a parametric Sweep to change your geoemtry shape (not topology your BC will get changed) and then perhaps a COntinuation sweep to change some damping or other material properties for a given geoemtry ?

--
Good luck
Ivar
Hi hat about using a parametric Sweep to change your geoemtry shape (not topology your BC will get changed) and then perhaps a COntinuation sweep to change some damping or other material properties for a given geoemtry ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 juin 2012, 09:38 UTC−4


Dear Ivar

the parametric sweep could be a way but I think that It is not the most effective.
If you have big models and several parameters you must have a lots of time.
In theory you could do the derivative of the stiffness and mass matrix in function of your geometric/material parameters and evaluate how the eigenvalues change.
I think that could be done exporting in to matlab the Mass/stiffness matrix., I am not familiar with matlab/comsol, but I have seen that comsol doen't export the material and proprieties of the Fem model in to matlab.

so I was thinking to create a general function that contains the material and geometric proprieties and doing on this function the sensitivity.
But i think that it is not so simple.
Michele
Dear Ivar the parametric sweep could be a way but I think that It is not the most effective. If you have big models and several parameters you must have a lots of time. In theory you could do the derivative of the stiffness and mass matrix in function of your geometric/material parameters and evaluate how the eigenvalues change. I think that could be done exporting in to matlab the Mass/stiffness matrix., I am not familiar with matlab/comsol, but I have seen that comsol doen't export the material and proprieties of the Fem model in to matlab. so I was thinking to create a general function that contains the material and geometric proprieties and doing on this function the sensitivity. But i think that it is not so simple. Michele

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.