User login

Navigation

You are here

Modal analysis

Hello
I am doing a Modal analysis on a structure made up of 3D-Beam4, Mass21 and Shear panel 28. To be more clear, the structure looks like a cube made up of Beam element on all 12 edges with Mass21 element at all corners of the cube and also with shells (shear panel 28) at all six faces of the cube.

I first constructed the beam structure with keypoints and lines, then i added the point masses and meshed them, then i created the shell element using Modelling-create-area and meshed it. Finally i merged them using "NUMMRG, ALL" in the command field.

After solving, the Eigen frequency i obtain is about 0.24271e-5 Hz. But the same problem when solved using ABAQUS gives the Eigenfrequency as 27 Hz which is the correct value.

Somebody please let me know where i went wrong.... Couldn't proceed any futher....

Thanks in advance
Prem

Comments

ramdas chennamsetti's picture

Hi,

Since the frequency is very low, it looks like the body is undergoing rigibody motion. If there are some unmerged nodes beyond the merging tolerence, then issuing the command NUMMRG, NODE will not help. You may check the following.

1. Check the nodes between the shear panel and beam. All the nodes should be merged on all 12 edges.

2. Check real constants and they are properly attributed.

3. Just try a simple static run by applying some load and BCs symmetrically. If some nodes are not merged properly, you may make out.

4. I suggest you to follow some other modelling approach (Top-Down). It is as follows.

(a) Create a cube - Volume

(b) Delete volume only. Still areas, lines and keypoints are left.

(c) Give element size on lines.

(c1) Define material props and attribute them

(d) Mesh areas with shells

(e) Mesh lines with beams

(f) Mesh keypoints with mass elements

(g) Issue NUMMRG, NODE command

(h) Just check all corners nodes - there should be only one node at each corner. Randomly check nodes on the edges. At each location, there should be only one node.

(i) Apply proper BC and run....

Best regards,

- Ramadas

Subscribe to Comments for "Modal analysis"

Recent comments

More comments

Syndicate

Subscribe to Syndicate