Model load assumptions for supports of continuous slabs

[dhgr8]

Hello, I'm a bit confused and I hope someone can clarify this:

When designing continuous one or two-way slabs supported on all edges, with ACI (any method), there is a distribution of moments. This causes the loads transferred to the supports of the slab to be different than the ones assumed by simple tributary areas, because of the continuity.

In a complete structure model such as one made in ETABS, there doesn't seem to be a way for continuous slabs to transfer this "realistic" load distribution to the beams supporting the slabs (well, not even a way to indicate that the slab is continuous). The program always transfers the load via tributary area.

Is it common practice to assign the realistic loads manually to the beams, while preventing the program from taking the tributary loads? To me, this would seem to be the right way to analyze and design the beams; however, it can become a complex and tedious task.

 

[centondollar]

I am not quite sure I understand what you mean. However, I assure you that ETABS and similar software (assuming that they are programmed correctly) do not deal with "tributary areas" when calculating internal forces in slabs. Slabs (structurally modelled as Kirchoff-Love or Reissner-Mindlin plates) have moments in two directions (Mx, My) and twisting moment (Mxy) in each part of the slab, and this must be accounted for when designing reinforcement.

The common practice (or at least the correct one, assuming linear elastic analysis) is to model slabs, beams and columns together, and to assign appropriate releases to the connections in order to portray boundary conditions (is the beam/slab corner a moment connection or a pin connection, for example) realistically and thus receive realistic results.

PS. In reality, all slabs are two-way, but if geometry and boundary conditions are restricted to a certain degree (e.g., continuous supports on two sides of the slab and/or very large aspect ratio), slabs are sometimes treated as "one-way", i.e., as beams with significant major axis bending moment (Mx or My) and negligible twisting moment.

 

[Tomfh]

It gets done both ways, and I believe ETABS allows both depending on how you model the slab (membrane vs shell?)

Read this thread:

https://www.eng-tips.com/viewthread.cfm?qid=479388

 

[dhgr8]

Thank you so much for the insight, that thread was very helpful. I'm currently working with joist systems (waffle slabs) supported on all edges.

So I guess my next question would be if makes sense to design a continuous slab on its own, for example, using the Equivalent Frame Method; and after that, model the slab as a shell along with the beams and columns, and assume that the loads will be distributed correctly by ETABS to the corresponding edge supports?

It just seems to me that the distribution of loads on the supports from the continuous slab analysis methods doesn't really match the one in the software. A clear example is one backspan, or two supports very close to each other next to a long span, so I'm trying to work out which approach is the conservative or the realistic one.

As a side question, is moment distribution in two-way slabs the same as shear (loading direction)? For example, I remember in ACI 318-63 there used to be shear and load direction coefficients that actually transferred the loads to the supports depending on the edge continuity. So would the same proportion apply to how Mx and My distribute the load direction?

 

[centondollar]

"It just seems to me that the distribution of loads on the supports from the continuous slab analysis methods doesn't really match the one in the software. A clear example is one backspan, or two supports very close to each other next to a long span, so I'm trying to work out which approach is the conservative or the realistic one."
The equivalent frame method (I assume you refer to it when you write "continuous slab analysis method") is not realistic, nor is it always conservative. The linear elastic solution from a FEM-software (or from a hand-calculation) is the "realistic" solution in the sense that it satisfies equilibrium: the slab is modeled as a plate, and it includes both bending and twisting.

"So would the same proportion apply to how Mx and My distribute the load direction?"
A slab always has Mx, My and Mxy moment components. If you want to make an accurate structural model for a slab in bending, you should use plate model bending and twisting moments to design reinforcement in ULS, calculate crack width in SLS by e.g. analyzing the slab as a beam in two directions (conservative, since "D > I_x" and "D>I_y") and calculate deflection in SLS by applying a reduced stiffness (e.g., by tweaking the Young's modulus, or by analyzing the slab as a beam in two orthogonal directions).

 

[MotorCity]

It depends on how you define your slab boundary. Say you have a 100'x100' floor. To model continuity, define your floor using 4 nodes in a 100'x100' square. Alternatively you could model several 10'x10' slab areas which would negate continuity.