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.
Minimal element size
Posted 15 févr. 2011, 22:01 UTC−5 2 Replies
Please login with a confirmed email address before reporting spam
Hello,
I am working with 3D acoustics model that calculates resonant radial frequencies in the cavity. The highest one I am expecting is around 9kHz(scaled model), so with it the wave length would be ~4cm. The minimal element size for my model was set to 1.2 mm, which I assume is sufficient (~31 elements per wave length). One thing I am still puzzled is about element size. I kind of new to FE/Acoustics so that might be something trivial. The thing is, edges of the cavity are fillet with 0.2mm radius, and from what it looks like Comsol puts ~ 2 elements "into" the fillet edge. Besides that, the mesh is unstructured, especially inside of the cavity. This brings up the question, is 1.2 mm is really a minimal element size? And is there a way to select an element of interest, and check its size?
Would really appreciate any help to clarify this for myself
I am working with 3D acoustics model that calculates resonant radial frequencies in the cavity. The highest one I am expecting is around 9kHz(scaled model), so with it the wave length would be ~4cm. The minimal element size for my model was set to 1.2 mm, which I assume is sufficient (~31 elements per wave length). One thing I am still puzzled is about element size. I kind of new to FE/Acoustics so that might be something trivial. The thing is, edges of the cavity are fillet with 0.2mm radius, and from what it looks like Comsol puts ~ 2 elements "into" the fillet edge. Besides that, the mesh is unstructured, especially inside of the cavity. This brings up the question, is 1.2 mm is really a minimal element size? And is there a way to select an element of interest, and check its size?
Would really appreciate any help to clarify this for myself
2 Replies Last Post 16 févr. 2011, 02:00 UTC−5
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I usually start with some 5 elements per smallest wavelength (but I have little experience with acoustics, even I do not see why it should differ from RF or optics) then I refine smaller on need.
COMSOL has advanced meshing features, if you look closely there are several parameters defining how to "behave" in curved regions, in small gap regions min/max element size, gradients etc. It's worth taking some time reading through the doc on meshing, and trying out several examples to get used to these parameters, you will gain a lot of time alter when you come with more complex cases. Comsol organises also some excellent courses worth to follow, but I would suggest once you have tried out a little to not to e a full novice
--
Good luck
Ivar
I usually start with some 5 elements per smallest wavelength (but I have little experience with acoustics, even I do not see why it should differ from RF or optics) then I refine smaller on need.
COMSOL has advanced meshing features, if you look closely there are several parameters defining how to "behave" in curved regions, in small gap regions min/max element size, gradients etc. It's worth taking some time reading through the doc on meshing, and trying out several examples to get used to these parameters, you will gain a lot of time alter when you come with more complex cases. Comsol organises also some excellent courses worth to follow, but I would suggest once you have tried out a little to not to e a full novice
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thanks for the input! The growth rate along with curvature distribution was sufficient to determine largest element in the cavity and number of them. Should of read that doc earlier ;) As for the 30/w.length, that was not intentional ;) it just happened that during mesh independence study, e'frequencies converged to 3 decimals on 0.5mil elements, hence, as a consequence of that I got a gazillion elements per wave length ;)
Thanks again!
Oleksandr
Thanks for the input! The growth rate along with curvature distribution was sufficient to determine largest element in the cavity and number of them. Should of read that doc earlier ;) As for the 30/w.length, that was not intentional ;) it just happened that during mesh independence study, e'frequencies converged to 3 decimals on 0.5mil elements, hence, as a consequence of that I got a gazillion elements per wave length ;)
Thanks again!
Oleksandr