Simulation of a 3D Beam Element in contact with a 3D Shell element 1

[Marquiavel]

Hi everyone,
I’m trying to simulate a Wall solicited by a distributed load, supported by a Beam, as is presenting on the image (1):

Image 1

I created a 3D Beam Element to represent the beam, and a 3D Shell element to represent the wall. Then, I assignment the sections of each element and in the beam element I assign a beam orientation (Image 2):

Image 2

Them I made the assembly of the instances, apply the boundary conditions on the Beam and apply the load:

Image 3

To finish, I apply the tie constraint in the elements:
Image 4

The mesh, I made with the same seed size for the 3 elements.
There is something wrong in my model? Because the results were 15 X higher comparing with the model I made using solids elements. Do I made correctly the assign of the beam orientation? Could be the error the application of the tie constraint?

 

[Mustaine3]

Upload the .cae with the two models or the two .inp and maybe someone can find the reason for different results. Just with your pictures I would guess, that your solid model could be wrong, since you could have Hourglassing or Shear Locking in your Hex-Elements.

 

[Marquiavel]

Hi Mustaine3, I'm sending the .cae of the models. The results of the solid model are more consistent when compared with what is expect for this type of structure.
Thank you!

http://files.engineering.com/getfil...a4-49fa-9404-a7534b5a4143&file=Wall&Beam2.cae

http://files.engineering.com/getfil...3-4ca8-a796-d3348afab431&file=SolidModel3.cae

 

[Marquiavel]

I'm sorry, I attach the models, but is work only the solid model:

http://files.engineering.com/getfil...3-4ca8-a796-d3348afab431&file=SolidModel3.cae

I'm attaching the .inp of the another model.

 

[Mustaine3]

a) You solid model creates bad results because of bad element selection. Use C3D8I or C3D20R in your case. The other linear elements run into Hourglassing or Shear Locking because of the coarse mesh.

b) I strongly recommend not to run the shell model. Shell elements are not suitable for your situation. They are not good with in-plane loadings and bad ratio of thickness to edge length. Think about running the problem 2D (plane stress or plane strain) instead.

 

[Marquiavel]

Thank you Mustaine3. I will apply this changes in the solid model. Do you think I should apply a different mesh? For example, more seeds or a different format.
Do you think the interaction I made between the beam and shell element is good? I will think about running the problem 2D.