How do you model RC slab? 1

[mats12]

I want to hear some opinions on modeling RC slabs in FEM softwares.

I'm asking this since I have seen more than one way to do it:
(1) I have seen some calculations where RC slab was modeled in some FEM software as 2d plate with simple (rigid) supports instead of concrete/masonry walls.
(2) I have seen some 3d models - whole storey with RC slab above it.

In case (1) supports are defined as rigid/stiff,
In case (2) walls are defined by material (concrete, msaonry, wood) which mean they are not really rigid - that has some influence on internal forces in a slab based on my experiences (larger values),

In a case where we have a RC beam that is partly integrated in RC slab - do you model it as an actual beam with actual cross section or do you replace it with simple support (which is fixed).
I have seen both and based on my experiences on designing slabs - internal forces are signif. larger when you consider actual stiffnes of a beam instead of rigid support.


Id like to disccus how do you model slabs in FEM because some times I get questionable results when I consider actual stiffnes of supports/beams.

 

[DCBII]

I often get questionable results when modeling 2-way slabs. FEM can produce the most convincing wrong answers you'll ever see. I'd never trust any software without comparing it to other calculations first to make sure I know what it's doing. I use RAM Concept, but before ever using it I ran a textbook problem in it, and I built an "Equivalent Frame Method" spreadsheet to compare it to as well.

Stiffness of elements connecting to the slab such as columns/walls/beams will have a big effect on your final results and load distributions. For example, modeling a beam element integrated with your plate elements will overestimate the width of the slab, since the plates overlap the beam. You also need to consider that the beam is not usually centered in the slab, but below the slab. This will stiffen the beams and draw more load to them. Are you modeling columns as hinged at the far end or fixed? ACI lets you assume they are fixed at the far end in the "Equivalent Frame Method", so that's what I usually do when using FEM.

If your model doesn't include framing elements like walls and columns, you'll have to assume some type of fixity at the supports. To say a slab is "Pinned" or "Fixed" at a support is a big simplification. I've always thought that if you assumed pinned edges, at least you'll have a secondary load path in the slab should a fixed connection fail at the edges. But then you get moments in the wall or columns below you didn't plan for unless you detail it as a true pin (who does that?). You could model the slab both ways: pinned for slab design, and fixed for support design. Then it can behave however it wants to and your covered - maybe?.

As a minimum the engineer should fully understand the hand methods before designing with FEM. Definitely use judgement.

 

[MIStructE_IRE]

I like FEM and its capabilities. However I also really like old school design and calculation methods.

Doing a slab by hand you would of course consider a slab to span between either beams or walls. In that case they would be considered spanning from a pin to a roller support. The IStructE’s ‘aims of structural design’ suggest that ‘no engineer should be more clever than is necessary’. Its a line that has always stuck with me. Sometimes I think we lose the run of ourselves with fancy software.

 

[hokie66]

No FEM modelling of slabs for me. I use all supports as pinned for the slab design, then go back and consider how much rigidity should be factored into the slab to support connections. Usually, this just requires nominal good practice reinforcement.

 

[TehMightyEngineer]

I typically use FEA for slab for things that traditional methods don't do well; generally precast panels/slab with openings and vehicle wheel loads. Once I'm away from the weird stuff I default back to traditional methods or at the very least FEA backed up with traditional methods for double checks.

It's been a while since I looked at it but ACI 318-14 has some guidelines/rules regarding using FEA for concrete design.

Ian Riley, PE, SE
Professional Engineer (ME, NH, VT, CT, MA, FL) Structural Engineer (IL, HI)
American Concrete Industries

https://americanconcrete.com/

 

[n3jc]

But when using traditional methods - hand calculations with lets say Hahn tables you assume that walls and BEAMS as rigid/stiff supports. But in FEM your beam has a stiffness based on cross section dimensions and its position in a slab - that results in larger inner forces - more reinforcement. Sometimes FEM gives results that are just off and they differ from hand calc. quite a lot. The question is should you neglect or consider this extra reinforcement (larger moments) you get from FEM because of actual stiffness of a beam??

Would like to see more thoughts on this.

 

[formi]

It's more tricky than you might think...

When doing hand calcs, you normally assume that the beam acts as a stiff support to a slab, whilst it doesn't... The beam will deflect and the load will be re-distributed somewhere else (to a stiffer element). When you model the beam in your FEA software you need to take into account the fact that your beam will be stiffer (being below the slab) and hence your beam's moment of inertia need to be increased (adjusted). Also, your FEA software uses elastic analysis which means that the long-term deflections (incl. creep) are not taken into account. This is another thing that should be considered when adjusting your beam's stiffness.

Re pinned/fixed slab edge: by default I assume that the slab is pinned when connecting to the wall, unless I need it to be fixed for some other reasons (transferring high lateral forces, etc.). In reality it's neither pinned nor fixed. If you assume it's pinned it will take only as much hogging moment (over support) as much top reinforcement you have (U-bars, top mesh). Then it will start to crack and the loads/forces will redistribute so that the sagging moment will be larger. This is how I see it. There's no need for 'tru pins' when dealing with standard buildings. it's different when designing bridge or really long span beams (when even small fixity might induce huge moment in the supporting column/wall due to high reactions).

I design slabs using FEA, but am careful and always check the outputs. Although my older colleagues keep saying to me that in the old days there were no computers and buildings still stand... Which is true...