CSA A23.3 two-way slabs as elastic frames

[STpipe]

I've been working through the design of a two-way slab based on the elastic frame method in the CSA A23.3 standard, and there is one part of the code that is unclear to me and I don't have the commentary or reference in terms of the rationale behind these provisions.

Based on my understanding, the standard provides two methods to establish the member stiffness - there is the non-prismatic modelling of member stiffness outlined in 13.8.2, and the prismatic modelling of member stiffness outlined in 13.8.3. The non-prismatic method seems to be the same methodology that's provided in ACI 318 for the equivalent frame method, so it's unclear why one would choose one method over the other given that ACI 318 does not have a second methodology to establish the member stiffness.

While digging a bit more into the theory behind these provisions, I also found this particular part of the PCA notes to ACI 318-11 interesting:

Reading that paragraph, it implies that the calculation of the equivalent column stiffness is only required for hand calculation methods like the moment distribution method. Does that mean that if you're using a frame analysis program, then it is not necessary to calculate the equivalent column stiffness and simply using the cross-sectional properties of the columns is sufficient?

 

[Celt83]

With general frame programs like RISA-3D, RAM Elements, etc. it is not possible or incredibly difficult to reduce the column stiffness for the effect of the torsional members.

In theory the torsional members still need to be modeled, and you'd need to distribute the moment from those elements into the slabs for design purposes. In practice I believe the most folks are doing is the rigid end offsets for the columns and maybe the slab/beams in these general frame programs.

Programs like SpSlab and Adapt I believe have a method programmed in to include the torsional member stiffness reduction on the columns.

 

[STpipe]

That makes sense, and I would be curious to know how common/prevalent it is to follow the procedure outlined to the letter, or whether the rigid offset simplification is considered good enough.

It's still unclear to me what the second methodology is about.

 

[STpipe]

Ok, as an update, I was able to find a source that addresses my original question now that I was able to get back to the office. I'll post it here for everyone's awareness. From Brzev, S., & Pao, J. (3rd Edition):

"A non-prismatic approach is used in conjunction with the moment distribution method, which has been used since the EFM (equivalent frame method) was first introduced in North American design codes in the 1960s (...). Each non-prismatic slab and column member with variable cross-sectional properties along the span can be modelled as a single member (CSA A23.3 CL.13.8.2). Correction factors are applied to modify member stiffnesses, carry-over factors, and fixed end bending moments for each frame member."

"A prismatic approach considers members with constant cross-sectional properties. Variation in slab properties within a span can be accounted for by considering prismatic segments with different gross cross-sectional properties. This approach is used in conjunction with the Direct Stiffness Method and it is suitable for computer applications (...) A column is usually modelled as a single member (...) CSA A23.3 Cl. 13.8.3.3 accounts for a reduction in column stiffness due to the attached torsional member through the column stiffness modification factor."

I guess what threw me off was the terminology of prismatic vs. non-prismatic. Typically prismatic for structural analysis means a single member, and when you have a member that has cross-sectional dimensions that vary within the span, then that's considered non-prismatic. So the approach appears to be simplified where the column stiffness in the prismatic method is easily calculated based on spans of the slab in both directions as opposed to calculating the Kec.