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 The information provided may be out of date.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 analysis of a simple poly-si structure
Posted Apr 11, 2011, 9:55 a.m. EDT Mesh, Studies & Solvers, Structural Mechanics Version 3.5a 1 Reply
Please login with a confirmed email address before reporting spam
Hi to all,
I am new in Comsol ( Comsol 3.5a ) and I am facing problem to find out eigenmodes in a simple structure.
It will be very helpful if someone can give some advice.
I am trying to perform an Eigenfrequency Analysis,in order to find out eigenfrequencies of Poly-Si surface(2D) in vacuum , so to get acquainted with that kind of problem I started solving a very simple problem , I drew a rectangle of poly-Si ( width = 500e-6 , height= 5e-6 ) , I fixed the upper and lower boundary and after meshing I solved the model.
I selected in the solve/solver manager window the "initial value expression" and "Use settings from initial value frame", and in order to get a more correct result I gave also , in the "search for eigenfrequencies around" dialog box , the frequency of 1st mode I expected.
So I have to deal with 3 main problems:
1 . When I solve the same problem ( without changing anything ) I get different results, in a considerable way for my purpose. ( I set in the " eigenfrequencies tolerance " dialog box 0.0 and also reducing the " relative tolerance " I get no result at all )
2. Results change in a very considerable way if I refine the mesh.
3. I woud like to know how to increase the number of digits I get as result.
To overcome that problem I tried using "adaptive mesh refinement" or extruding my geometry, but I got no result.
Finally I tried changing solver from the default linear system solver "direct(UMFPACK)" to "direct(SPOOLES)"
( I first tried the PARDISO direct solver as suggested in the guide, but I had some problems), solving 1st of my problems, but not the others.
Thanks in advance for the answer.
Kind regards
Alessandro
I am new in Comsol ( Comsol 3.5a ) and I am facing problem to find out eigenmodes in a simple structure.
It will be very helpful if someone can give some advice.
I am trying to perform an Eigenfrequency Analysis,in order to find out eigenfrequencies of Poly-Si surface(2D) in vacuum , so to get acquainted with that kind of problem I started solving a very simple problem , I drew a rectangle of poly-Si ( width = 500e-6 , height= 5e-6 ) , I fixed the upper and lower boundary and after meshing I solved the model.
I selected in the solve/solver manager window the "initial value expression" and "Use settings from initial value frame", and in order to get a more correct result I gave also , in the "search for eigenfrequencies around" dialog box , the frequency of 1st mode I expected.
So I have to deal with 3 main problems:
1 . When I solve the same problem ( without changing anything ) I get different results, in a considerable way for my purpose. ( I set in the " eigenfrequencies tolerance " dialog box 0.0 and also reducing the " relative tolerance " I get no result at all )
2. Results change in a very considerable way if I refine the mesh.
3. I woud like to know how to increase the number of digits I get as result.
To overcome that problem I tried using "adaptive mesh refinement" or extruding my geometry, but I got no result.
Finally I tried changing solver from the default linear system solver "direct(UMFPACK)" to "direct(SPOOLES)"
( I first tried the PARDISO direct solver as suggested in the guide, but I had some problems), solving 1st of my problems, but not the others.
Thanks in advance for the answer.
Kind regards
Alessandro
1 Reply Last Post Apr 11, 2011, 10:32 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
try to do it simple, keep the default settings first, check that
dimensions are correct (3D or 2D + thickness) (start with 1:50 slender ratio, certainly not a 1:1000'000 to begin with;)
Materials data are ok
meshing is at least a few elements across the thin beam dimension, but do not overdo it, to start with
Ask for a simple eigenfrequency analysis, default settings
Compare to analytical values (mode, validation)
Then start to shape it in your way, step by step.
This will make you progress quicker even if it appear somewhat slowish.
Once you master well eigenfrequency analysis you will be able to go quicker to the point, but be aware, even after a few years with intensive COMSOL I often have to back up, because I wanted to go straight to the final, not doing my homework ;)
--
Good luck
Ivar
try to do it simple, keep the default settings first, check that
dimensions are correct (3D or 2D + thickness) (start with 1:50 slender ratio, certainly not a 1:1000'000 to begin with;)
Materials data are ok
meshing is at least a few elements across the thin beam dimension, but do not overdo it, to start with
Ask for a simple eigenfrequency analysis, default settings
Compare to analytical values (mode, validation)
Then start to shape it in your way, step by step.
This will make you progress quicker even if it appear somewhat slowish.
Once you master well eigenfrequency analysis you will be able to go quicker to the point, but be aware, even after a few years with intensive COMSOL I often have to back up, because I wanted to go straight to the final, not doing my homework ;)
--
Good luck
Ivar