Transfer Slab Modelling Issue

[ETET001]

Hi,

I have a few questions regarding modelling in FEM software.
Let's say I have a tranfer slab which is gonna support the walls above. So I have modelled it in two different ways like below.
The first one is using multi shells around the columns below and walls above. The second one, on the other hand is one single shell with null lines (or dummy beams). Their meshes are the same.

However, the stress looks different. The stress of the first method looks not right. I would like to know what seem to be problem for the first method?

Also, the wall above will become two point load applied in the slab in FEA instead line loads. Should we and can we do any amendment for wall objects and why?


Thank you.

 

[Trenno]

"Also, the wall above will become two point load applied in the slab in FEA instead line loads. Should we and can we do any amendment for wall objects and why?"


You're right, with an unmeshed shell the wall will deposit it's reactions onto the slab at the corners. To get closer to a line load, you'll need to mesh both the slab and wall finer, ensuring the wall/slab mesh align and are properly connected at common joints.

 

[ETET001]

Trenno thanks for your reply. That's what I was trying to do but before that I would like to know what the difference is between dividing shell into multi shells & simply assigning meshes. As you can see from the slab above using these two methods, the stress results differ. Thank you.

 

[Trenno]

What stress are you showing (S11, S22, S12)?

Are the local axes of all the shells aligned the same? When the 3 axis on shells are in opposite directions, this can affect what stress is shown when you use 'visible face' option.

Are the column reactions the same?

 

[ETET001]

Hi Trenno, sorry they were shell internal forces. I just chose F11 & Mmax as examples.

Yes all the local axes of all the shells aligned are the same. And the column reactions are the same too.


Also, even if I divide the wall into more areas, from the F11 results of the walls I can see that at the two ends the figures are much greater than others, which means they still behave like two point loads applied on the slab.

 

[Trenno]

I'm not sure why you're getting the discrepancy in results. But vertical load in a wall, despite how fine it's mesh is, will tend towards the edges / stiffer parts of the slab below.