Requirements of Drop panel in flat slab 1

[bdlc2k]

ACI states that a drop panel dimension should be L/6 in both directions. Is this required if the drop panel is being considered for punching only, not for flexural design (say when a two way slab has shearwalls)? Has anyone ever taken this approach?

 

[dik]

KootK... I seem to recall (way back when) checking the As required at the edge of the drop with that at the centreline and the As requirements at the edge were slightly less than centre, but of a similar magnitude. Used to use 2 cycle moment distribution for the analysis using stiffness coef from haunched beam coefficients and used the full strip width for calculating the properties.

Dik

 

[dik]

take it that it's your opinion that middle strip top reinforcing should remain in our designs

It was common practice when I first started out... a bottom mat with with top mats only at column locations

Dik

 

[hokie66]

That's a new one on me. Must be a Canadian thing, and the loss of your business card shows that it is a poor idea.

I have seen a few flat slabs and flat plates, more than I want to remember, but have never seen zero reinforcement at the top, C-M strip. The only place you have zero reinforcement top is in the internal M-M area, and sometimes not even there.

 

[BAretired]

When top mats are used only at column locations, some of the bars should extend into the middle strip in each direction to serve as tie bars for middle strip top bars which are placed in the outer quarter of the middle strip. The total top reinforcement in the middle strip should satisfy the requirements of the code.

The advantage of this arrangement is it provides a clear path half the width of the middle strip between columns, allowing concrete to be placed without stumbling over top mats.

BA

 

[rapt]

The Australian code now (since 2001 I think) allows the top reinforcement to be concentrated over the column strip for ultimate strength. And it still requires crack control to be checked on the elastic moment distribution, which then requires reinforcement in the top in the middle strip. So you end up with extra reinforcement if you design properly for both Ultimate ans Service.

I know of one consultant did the column strip only detail for top reinforcement when it was first introduced into the Australian code, without checking crack control in the middle strip (that provision was spelled out in a later release of the code). We were discussing the code clause later and I suggested it was not logical as crack control would require reinforcement there. He then told me what happened in his slabs. He said he would never do it again as he ended up with a perfect crack pattern in the middle strips at the top with uncontrolled crack widths (similar to what Dik saw).

And if you had no middle strip reinforcement, or too little reinforcement so that it fractures at the support, then it is likely that the crack will be full depth and that it will be rotating as he suggests. Especially as the bottom reinforcement is probably not lapped at the support line!

But if you put sufficient reinforcement in the top in the middle strip, the cracks will be controlled and rotations controlled and the middle strip reinforcement will have the effective depth of the slab in the middle strip.

The other thing that I think we do wrong in this case (and Codes do not say anything) is that the middle strip check should be done at the line of the centre of the supports, not at the face of the supports. Dik will probably confirm that the cracks were on the lines of the centre of the supports. I have seen this once before in a PT slab that cracked at transfer.

 

[dik]

BART. The mat width was larger than the column strip...

Dik

 

[dik]

a perfect crack pattern in the middle strips at the top with uncontrolled crack widths (similar to what Dik saw).

It may be the common practice here to use thicker drop panels... drop panel depth typically 3/4" formply + dimensioned lumber, so no trimming of dimensioned lumber required.

Dik

 

[hokie66]

I am all for constructability, but don't see leaving walkways as good reason for omitting middle strip top bars. Concreters soon learn where the reinforcement is.

 

[dik]

Hokie... just ease of construction... and likely economy. It's like putting PT strand distributed and over column strips... works... but only makes limited sense.

Dik

 

[hokie66]

Very limited, IMHO. I think it is just wrong.

 

[KootK]

Two other factors that tend to encourage the omission of middle strip top steel:

1) It's often used on modern residential towers where the column layout is irregular and it's pretty hard to even define a proper strip in the first place. The use of the middle strip top steel leads to reinforcing that is difficult to detail and place appropriately (skews, non-orthogonality, etc). That, or you resort to effectively top steel everywhere which isn't a popular choice with the build team.

2) In many condo farm cities, like Vancouver BC, the method has been successfully used on zillions of square foot of tower without causing enough problems to deter anyone from doing it on future projects. As always, it's tough to argue with success, especially the $$$ kind.



I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[dik]

In many condo farm cities, like Vancouver BC, the method has been successfully used on zillions of square foot of tower without causing enough problems to deter anyone from doing it on future projects. As always, it's tough to argue with success, especially the $$$ kind.

Per Christoffersen, an original partner of RJC did some model studies about 5 decades back to determine the distribution of reinforcing in top mats. There were different distributions for flat plates, flat slabs, and column shapes. The total strip width reinforcing was determined and for negative moment mats, there was an portion that went across the entire mat, and a portion that was concentrated over half the drop panel width. With flat plates, there was a concentration over the column area, with a greater concentration over 'narrow' columns. I no longer recall the different distributions.

Dik

 

[rapt]

Dik

A drop panel would normally be 1.5 - 2 times the slab thickness.

Dik and Hokie

Here is a new one. RAPT accepting banded distributed!

It works for ultimate strength with a lot of redistribution. Even on Yield line theory you can probably justify it. But neither of those consider service stresses and crack control and deflections. You are providing a load path to the supports, just one that is not the elastic load path. And you cannot do a fully redistributed design or a yield line design based on an elastic FEM analysis. The analysis has to reflect the load path adopted. For rectangular column grids there is no real problem, but as soon as you have significantly offset column locations, you cannot use elastic FEM for the analysis.

In the early days of banded/distributed, the idea was to keep the stresses very low at service so the concrete was un-cracked (based on real stresses. not averages). Everything stayed elastic so there was no redistribution at service. That works as it is un-cracked and and deflections can be estimated reasonably well.

But USA designers saw that elsewhere we were doing partially prestressed design with much higher tension stresses and reinforcement to control cracking and wanted the benefits of that design logic. Once you get cracking at service with banded/distributed and especially with average moments, then it is not logical. You are getting redistribution at service that is being ignored in deflection calculations and cracking because real stresses are not being estimated as the ultimate load path is being assumed while service design must be based on the elastic load path. That is what causes it to crack in the first place.

Unfortunately some of the "experts" who have been advising the PTI and ACI on all of this over the years really did not understand what they were doing (in my opinion) and you now have the design logic that exists today all based on "I have been doing it for years and there have been no problems" logic. There was a very long "discussion" on the SEASOC webpage many years ago where I had an in depth argument with one of those people where he was justifying a multi span edge beam parallel to the distributed tendon direction carrying a 4m high cavity brick wall and used an effective flange width of about 5m (full half panel) and included all of the distributed tendons in the slab over that width in estimating the beam flexural capacity, with those tendons having the effective depth off the beam at the supports. His final justification was that he had done it on hundreds of projects and millions of sqft of floor systems, so it must be ok.

Same argument has been used to justify unbonded prestress and the design methods I mention in an earlier post for drop panel and band beam slabs. In the drop panel case, for the cases I have checked, I would estimate the design is under capacity by 15-20%. So nothing outwardly bad happens. But the client has paid for 100% capacity and has been given 80-85%.

The PTI continually says there is no inherent problem with unbonded prestress slabs and banded/distributed designs. But ask the people doing reviews and repairs on them after 10-30 years.They will tell you a different story. But they probably do not want to go into print on it as they are making too much out of doing the reports and repairs.

 

[dik]

A drop panel would normally be 1.5 - 2 times the slab thickness.

In these environs they are usually slightly thicker than the slab and confirmed by the As required at the edge of drop vs centerline of drop. I don't think I've ever seen drop panels approaching twice the slab thickness.

Dik

 

[rapt]

Dik,

are we talking total thickness of drop or depth below the slab? I am talking total.

 

[dik]

Rapt... we're talking the same thing.

Dik

 

[rapt]

Dik,

I just ran a 5 span flat slab, 8m * 8m spans RC with drop panels 1300mm each side of the column in both directions. Slab 250mm deep and drop panels 500 or 375 for the 2 cases.

- With drop depth = 2.0 * slab depth, the column strip reinforcement required at the face of the drop is about 90% of that required at the column critical section. So pretty close to your optimum!
- With drop depth = 1.5 * slab depth, the column strip reinforcement required at the face of the drop is about 54% of that required at the column critical section. And deflection of the column strip is a little over 50% higher!

So if you are using a lot shallower drop panels than that, the reinforcement at the column critical section must control by a lot more.

 

[dik]

rapt... yup.

Dik

 

[rapt]

Dik,

So you are happy to have a small reduction overall in concrete resulting in a large increase in deflections and more than double the amount of top reinforcement!

Even when the reduction in concrete is not resulting in an overall building height reduction as the depth is required for services.

 

[dik]

rapt: nope I check that the centre As governs to determine my As and proportion my drop accordingly.