SAFE - Different Shells

[EngDM]

Hello,

I was wondering if anyone here had a rule of thumb or guidelines on when to use Shell-Thin, Shell-Thick or Membrane for SAFE. I know that the thick shells are for transverse shear transfer, but I'm not sure how thick your concrete member needs to be before you should transition over to a thick shell. Further, membrane appears to transfer based on trib area instead of the finite mesh, but I'm not sure where the use of this would be appropriate.

For instance, do you typically treat slabs as shell-thin, walls as membrane, pile caps as etc...

 

[Luceid]

The shell-thin/shell-thick question always bothers me too - and I don't have a good rule of thumb for it either. You need to judge when shear deformation may become more substantial that it's worth getting involved (longer span, thicker plates? not sure) You can always test both options to see what happens and if the difference is even worth worrying about. But there's others who will know better than I do how to judge.

Shell vs membrane depends on if you want your element to have any out of plane stiffness at all. If you are using your slabs as rigid diaphragm elements that spread tributary loading to beams nearby, then you can use membranes no problem since they don't need any bending stiffness. But if you are doing a typical slab analysis in SAFE where you are looking at deflection, bending moments in the slab itself, you'd want a shell since membranes only have in-plane stiffness and therefore only provide in-plane results.

 

[271828]

The following should be helpful:

Thin vs. Thick shells - Technical Knowledge Base - CSI Knowledge Base

web.wiki.csiamerica.com

 

[milkshakelake]

Easy rule for me: thin shells for ETABS, thick shells for SAFE. The thick shell is more precise, but takes longer to run. In SAFE, you're only running one slab at a time, so it doesn't matter as much. In ETABS, you might be running lots of slabs for various stories, and it's more used to get an overall picture anyway.

 

[Luceid]

So using a thick-shell element is never "incorrect" when modelling slabs - it may just be unnecessary?

 

[milkshakelake]

@Luceid It can be incorrect for very thick slabs with small spans or two columns next to each other. That's kind of a rare case, though. And honestly, thin shell will give you just as many headaches as thick slab in those cases. It's just that a thin shell will be slightly more accurate in those specific cases. I end up doing stuff by hand when it comes to those special cases, and using the FEM model for the rest of the area.

Edit: Just read the CSI article and what I said isn't true at all, sorry. I'd just refer directly to that article and make your own judgment.