Modal Analysis of Spherical Shell
- Claus
-
- Offline
- Moderator
-
Less
More
- Posts: 670
- Thank you received: 34
15 years 3 months ago #3924
by Claus
Code_Aster release : STA11.4 on OpenSUSE 12.3 64 bits - EDF/Intel version
Replied by Claus on topic Re:Modal Analysis of Spherical Shell
Btw., as far as I know, it is not necessary to secure the structure when doing modal analyses - in this case, my study just produces invalid results for the first 3 or so frequencies. That has something to do with my mesh I think.
Think of it as throwing a steel ball up in the air and hitting it with a BB gun or something, that'll vibrate freely without any boundary conditions.
/C
Think of it as throwing a steel ball up in the air and hitting it with a BB gun or something, that'll vibrate freely without any boundary conditions.
/C
Code_Aster release : STA11.4 on OpenSUSE 12.3 64 bits - EDF/Intel version
- sbrockingtonhyperv
- Topic Author
- Offline
- New Member
-
Less
More
- Posts: 5
- Thank you received: 0
15 years 3 months ago #3944
by sbrockingtonhyperv
Replied by sbrockingtonhyperv on topic Re:Modal Analysis of Spherical Shell
Yeah, kicking a blob in freespace is what i expect the answers to be like
I finally got something similar,
If you assume frequency is in kHz because of my units, the first 3 modes could be 0 frequency or DC displacement of the bulk, and the negative numbers are occurring simply because they are essentially zeros. mode 4 is a breathing mode at about 55htz and the spherical harmonics start at mode 5 with ~200htz on up.
number
3 1.8021e-8
2 -2.3197e-8
1 -3.1513e-8
4 5.4681e-2
5 1.9647e-1
6 1.9934e-1
7 4.3854e-1
.
.
.
cool, I'm changing my mesh scale so that i can use MKS units, and i'll do a convergence test over the week
thank you for all the great advice, i'll be sure to let you know how the convergence test goes
I finally got something similar,
If you assume frequency is in kHz because of my units, the first 3 modes could be 0 frequency or DC displacement of the bulk, and the negative numbers are occurring simply because they are essentially zeros. mode 4 is a breathing mode at about 55htz and the spherical harmonics start at mode 5 with ~200htz on up.
number
3 1.8021e-8
2 -2.3197e-8
1 -3.1513e-8
4 5.4681e-2
5 1.9647e-1
6 1.9934e-1
7 4.3854e-1
.
.
.
cool, I'm changing my mesh scale so that i can use MKS units, and i'll do a convergence test over the week
thank you for all the great advice, i'll be sure to let you know how the convergence test goes
- kwou
-
- Offline
- Moderator
-
15 years 3 months ago #3945
by kwou
Interest: structural mechanics, solar energy (picture at 'my location' shows too little pv panels)
--
kind regards - kees
Replied by kwou on topic Re:Modal Analysis of Spherical Shell
Hoi
I would expect 6 rigid body movements, 3 displacements and 3 rotations, and hence 6 zero eigenvalues for a floating object. So why do you only have 3 frequencies zero?
edit later:
Sorry, I didnot read the full contribution. You locked one node to ''ground'' (I presume 3 translations), so what remains is the rigid body movement around this point, being 3 rotational dofs. So 3 zero eigenvalues (due to machine precision you may get negative results here, but at least order of magnitude smaller than the first 'real' eigenvalue).
If you need [Hz] in stead of [kHz] use for density [ton/mm3] (thus water has density 1.0e-9 ton/mm3) if the are dimension in [mm] adn pressure, Young's module in [MPa]. See www.caelinux.org/wiki/downloads/docs/PCa...00180000000000000000 Paul Carrico 's overview.
BTW: your eigenvalues will change depending on whether you connect one or more nodes to ''ground'' or not. Boundary conditions have great impact on the eigenvalues and modes of the construction.
Kind regards - kees<br /><br />Post edited by: Kees Wouters, at: 2010/03/10 10:42
I would expect 6 rigid body movements, 3 displacements and 3 rotations, and hence 6 zero eigenvalues for a floating object. So why do you only have 3 frequencies zero?
edit later:
Sorry, I didnot read the full contribution. You locked one node to ''ground'' (I presume 3 translations), so what remains is the rigid body movement around this point, being 3 rotational dofs. So 3 zero eigenvalues (due to machine precision you may get negative results here, but at least order of magnitude smaller than the first 'real' eigenvalue).
If you need [Hz] in stead of [kHz] use for density [ton/mm3] (thus water has density 1.0e-9 ton/mm3) if the are dimension in [mm] adn pressure, Young's module in [MPa]. See www.caelinux.org/wiki/downloads/docs/PCa...00180000000000000000 Paul Carrico 's overview.
BTW: your eigenvalues will change depending on whether you connect one or more nodes to ''ground'' or not. Boundary conditions have great impact on the eigenvalues and modes of the construction.
Kind regards - kees<br /><br />Post edited by: Kees Wouters, at: 2010/03/10 10:42
Interest: structural mechanics, solar energy (picture at 'my location' shows too little pv panels)
--
kind regards - kees
- sbrockingtonhyperv
- Topic Author
- Offline
- New Member
-
Less
More
- Posts: 5
- Thank you received: 0
15 years 2 months ago #4086
by sbrockingtonhyperv
Replied by sbrockingtonhyperv on topic Re:Modal Analysis of Spherical Shell
Alright, I did the classic fixed-free beam problem, where you have a square cross-section beam. I used a 1 cm x1 cm x1 meter long beam with one face anchored and performed modal analysis. I performed this simulation with both a mm scale and meter scale mesh. I got an elongation mode in the direction of the beam length, a torsional mode about the long axis of the beam, and all the longitudinal oscillation modes of in the length of the beam, in both short dimensions. Using a millimeter scale length mesh does necessitate that the units of the density and Young's modulus be appropriate scaled, and does produce results in kHz which code aster mislabels as Hz.
using MKS units and a mesh scaled in meters will produce frequencies in Hz
and i even got results very close to an analytical model.
Number Frequency(Hz) Description
1 8.446 n=1 x mode
2 8.446 n=1 y mode
3 52.90 n=2 x mode
4 52.90 n=2 y mode
5 148.0 n=3 x mode
6 148.0 n=3 y mode
...
13 754.9 torsion about z axis
...
16 1293 Elongation down the z axis
17 1321 n=8 x mode
18 1321 n=8 y mode
I re-did the ball tank modal analysis and had a easier time getting simulations to finish if i "grounded" a group of faces as an anchor rather then a single point. Thank you for noting boundaries influence the results, I'm going to work making a more physical representation of the tank. Thank you again
using MKS units and a mesh scaled in meters will produce frequencies in Hz
and i even got results very close to an analytical model.
Number Frequency(Hz) Description
1 8.446 n=1 x mode
2 8.446 n=1 y mode
3 52.90 n=2 x mode
4 52.90 n=2 y mode
5 148.0 n=3 x mode
6 148.0 n=3 y mode
...
13 754.9 torsion about z axis
...
16 1293 Elongation down the z axis
17 1321 n=8 x mode
18 1321 n=8 y mode
I re-did the ball tank modal analysis and had a easier time getting simulations to finish if i "grounded" a group of faces as an anchor rather then a single point. Thank you for noting boundaries influence the results, I'm going to work making a more physical representation of the tank. Thank you again
Moderators: catux
Time to create page: 0.152 seconds