Two-way or one-way concrete slab system? 2

[GalileoG]

Dear all,

I have a concrete floor system with 8m x 7m bays with a 2.5m slab cantilever at the perimeter. I can only run beams parallel to the 8m span.

I am wondering if by simply adding a 'stiff' beam parallel to the 8m span, I can design the slabs as a one-way system perpendicular to the beam.

What would constitute a 'stiff' beam and wouldn't that mean I can expect significant cracking before the two-way distribution would turn into a one-way distribution?

I would prefer to have slab-bands instead of a deeper beam (headroom restriction), but surely I can not design for a one-way slab condition with a shallow slab-band albeit wider than the concrete beam, can I?

Also, any special considerations when having a 2.5m long cantilever at the perimeter of the floor system?

Would really appreciate some guidance on this topic.

Thanks!

 

[hokie66]

Band beams and one way slab sounds like the ideal system to me. I would start with bands 350 to 400 deep and 2400 wide.

 

[slickdeals]

That brings me to another question. How stiff is this band that it precludes two-way behavior? I have modeled a couple of these in a FE program and more often that not there is no predominant one way behavior when such slab bands are modeled, especially for a square grid.

 

[GalileoG]

hokie66,

Thanks for the advice.

slickdeals,

I absolutely agree. I've done a FEA analysis of this and can confirm that even with a reasonably sized slab band, two-way distribution is still very dominant. Can someone elaborate how we can justify a one-way slab design?

 

[hokie66]

By not using FE programs. As long as you have a reliable load path...one way slab, one way band...why do you need another? I have done lots of these with one way action assumed. If in practice there is some two way action, all the better, as that just gives you some redundancy.

 

[slickdeals]

Hokie, I agree with you on ultimate moments, but what about serviceability? Would that be an issue if you don't reinforce based on the *true* moments? A one way slab would have been designed for the same positive and negative moments along the length of the band, based on a unit width, but in reality the negative moments in the one-way slab are going to be higher near the supports.

 

[hokie66]

Maybe. If you are concerned about more flexural cracking in the slab in the vicinity of the columns than elsewhere, then by all means add some bars. But I have never seen that phenomenon exhibit itself "in reality". The main crack control issue with banded slabs is normally the restraint cracks which form perpendicular to the bands.

 

[nicam]

What are the support conditions? and where are the supports?

 

[ron9876]

Why not just design a two way slab system?

 

[JAE]

I think I agree with hokie66 in that the way you reinforce concrete can dictate the load path in terms of structural strength and the cracking that results from "real" behavior may show up. The trick is to visualize how that inherent 2-way action will actually occur and add reinforcement to control cracking, knowing you have a secure load path elsewhere.

You see this in your beam-slab system designs where engineers will include:
1. Traditional transverse temp/shrinkage reinforcement
2. Heavier T/S reinforcement near ends of beams where 2-way cross stresses occur
3. Diagonal bars extending from the outer beam faces at corners towards the center of the slab to control diagonal cracking at the corners.

 

[hokie66]

The system of band beam and one way slab is very common in Australia, probably the dominant system for floors in commercial buildings. Builders prefer it because of the simplicity of formwork and reinforcement, thus the economy, and engineers like it because of simplicity of design. The bands provide a stiffer floor than a flat slab...the bands are just drop panels connected, so they provide room for bottom reinforcement with a greater effective depth than available in a flat slab or flat plate. The span of the slab itself is also reduced, by the width of the band.

The band beam and slab system is used here either conventionally reinforced or post-tensioned (generally with partial prestressing). It also lends itself to the use of steel form deck for both the bands and slabs. Another version uses precast form panels (one proprietary system is called Humeslab), which have bottom reinforcement cast in.

 

[rowingengineer]

Can I discuss here why FEA doesn't catch this situation correctly? or have people had enough of my FEA dribble at every chance.

The reason that Finite element doesn't catch the one-way slab behavior correctly without manipulation can be due to a number of items. The most common is the stiffness of the beam compared to the slab. If the beam and slab are not modeled with the correct stiffness due to reo than the results will be skewed. Generally the beams should be stiffer by a factor of 4ish. This stiffness should also include the stiffness due to T and L beam framing. This is often ignored if plate elements are used with centroids matching and should be adjusted by experience.

I could also go into the other assumptions into the creep and shrinkage differential ect but I am sure that will bore most people.

in other words I agree with Hokie66.

http://www.nceng.com.au/

"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

 

[GalileoG]

Great input folks. Really good stuff.

hokie66, I'm a bit hesitant to take the slab span as the clear distance between the bands as opposed to the center to center distance of the bands. Am I out to lunch on this?

 

[hokie66]

The span is the centre to centre distance, but the thickness of the band will generally mean your controlling negative moment is at the edge of the band, and the controlling positive moment is greatly reduced by redistribution to the thick band section. Just do a simple analysis, and you will see.

 

[rowingengineer]

I have been asked to provide some modelling examples showing an example of my above post.

So please find attached screen shots from three different model types.

from the pdf you can see with the 1'st model that the My moments are very small in the model compared to the beam moments, this is to be expected and shows that the finite model can be corrected to show the one way action. The second model dose not capture the reduction in the my moment as well, this is due to the beams being modeled as beam elements, beam elements are really only good is you include offsets to the edge of the beam or the beam is narrow. The 3'rd model is a straight flat slab.

For the 1'st model which is the model I would recommend is modelling in Finite elements package.

I have not looked at any of my assumptions in this modelling for example the column moments are assumed fully fixed. I have only shown the My because that is what we are interested mostly.

http://www.nceng.com.au/

"Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."