Concrete slab on metal deck cracking. Any insight? 7

[MWPC]

We have a client who's elevated slab on metal deck and bar joists is cracking. Thankfully it was not our design. It appears to be shrinkage cracks but I cannot find much information on control joint locations for elevated slabs. Any help or direction to authoritative source would be appreciated. Ken

 

[271828]

Lin, Yiching, J.Y. Yen, and Chen-Fun Chen "Tracing Initiation and Propagation of Cracks in Composite Slabs," ASCE Journal of Structural Engineering, July 1996.

Lin, Yiching, J.Y. Yen, and Chen-Fun Chen "Composite Beams Subjected to Static and Fatigue Loads," ASCE Journal of Structural Engineering, June 1997.

These guys discussed the subject a little, but it's not very helpful. There are lots of refs in those papers if anybody cares enough to go digging.

 

[haynewp]

I had {heard} from a good source that some testing that was done did not show much of a difference, but I would not bet my license on it. I need to call that guy out in the Midwest that does testing for Vulcraft, Lloyd I think? And see if he knows anything. I have come across a couple of composite cracked slabs over girders myself though not of my design. Or I may decide to research it out on my own.

 

[271828]

haynewp, I'd be interested to see what he says.

The next time I talk to Dr. Easterling at VT, I'll ask him too. He's a composite slab guru.

As for betting a license on it, I don't think one has much of a choice other than to not assume composite action at those locations. That one paper even said that rebar performed WORSE than WWR, so there's no help there!

 

[hokie66]

Curious to know why you think the beams supporting the joists may be composite. What would be the detail for the depth of the joist seat? We didn't do composite beams with bar joists when I worked in the US, and in Australia, we don't have bar joists at all, so I may be in the dark. If the beam is composite, there must be a web of concrete over the flange and surrounding the joist seats, in which case I would not expect the cracks to align with the studs.

 

[271828]

The beam almost certainly isn't composite. If the floor is screeded level and the joists deflected, then I'm assuming that the slab is thicker away from column lines and thinner at them. The principle would be similar--that cracks would probably form at the thinnest locations.

 

[Boiler8019]

Joist construction doesnt typically space supporting members as much as beam-girder construction. Therefore pattern cracking can be attributed to the joist span more likely than deck span. It's obvious that the joists will deflect due to the weight of the wet concrete, though, I'd recommend reading into how quickly the deflection occurs. I'll have to dig and find it, but I have read that there is time lag. If the joists were somewhat underdesigned then he'd may want to research that avenue.

The post above regarding the lab test on the 3 span deck tested the deck capacity of a narrow walkway is quite different than a longer spanning support system. Just something to ponder.

 

[hokie66]

Boiler,
Why would there be a time lag? There is no creep component of deflection in steel at atmospheric temperatures.

 

[JAE]

yes...I wondered that too.
Boiler8019 - I don't think there'd be a time lag with steel. The deflection would be instantaneous wouldn't it?

 

[271828]

I don't see how there could be a time lag either, but perhaps Boiler will dig up a ref.

One idea is that the concrete at the top shrinks faster than the concrete on the deck. This would cause curvature and perhaps negative moments over supports.

I'd have to think about the moment diagram with the top of the continuous beam trying to shrink more than the bottom.

The concrete is so weak during those early times that it might crack due to this.

I don't see any way that it can be due to deflection of slabs, joists, beam, whatever, though.

 

[Boiler8019]

My bad guys, now I remember my reference. It came from a 30yr+ senior engineer I work with. Though, not everything he says makes sense and I think he's making a bogus claim on this one. Boiler Materials Class 101 says steel has minimal creep...shouldve stuck to the basics.

I'm interested about hearing the final outcome on this.

 

[MWPC]

Wow, this has sparked a lot of replies.

From a quick site visit and conversations with the contractor here is what I understand the system to be:
-3" normal wt concrete slab, total thickness.
-WWF 6x6-W2.9xW2.9
-3/4"+- corrugated metal deck, non-composite
-Steel joists at 2' o.c. spanning between beams
-Steel beams spanning between columns, non-composite and simple spans.

As noted in my earlier post, the contractor made a poor choice of construction joint locations. He located them parrallel with the beams at roughly the quarter points of the joist span with the 1/4 placed first with adjacent bays. Then when he placed the remaining 3/4 of the slab, the dead load created negative moment in the hardened slab over the beams. Loading the beams also created negative moment over the columns and caused cracking in the orthogonal direction. Throw in a few hairline shrinkage cracks and you can see why the owner questioned the product.

It seems that placing the construction joints on the column lines parrallel to the joists would have eliminated most of the cracks, especially the larger ones.

Most concrete in the area contains fly ash. All else equal, and it never is, how does fly ash effect the shrinkage as compared to a straight portland mix?

Ken

 

[hokie66]

The location of the construction joints doesn't prove whether the cracks formed due to bending or shrinkage. The section over the beam is still the thinnest, so where the shrinkage cracks would occur. I suggest directly over the beam is the best location for construction joints in this system. Makes a nice straight crack.

Properly designed fly ash mix should not shrink more than all portland cement mix. If anything, a bit less. More likely culprit is inadequate (or zero) curing.

I did a lot of floors like this one years ago and never had problems with this type cracking. Wonder if the floor is as stiff as it should be. What does the engineer who designed it say?

 

[MWPC]

I sent the designer a courtesy email and did not get a reply.
Casting the slab in strips parallel to the joists would eliminate negative moment cracking over the beams since the concrete would not have set and over the columns there would be a nice straight joint.

 

[hokie66]

MWPC,
Maybe you didn't read all the posts, especially those by 271828. He and I, and perhaps others, believe that your cracks are probably not due to negative moment but rather to drying shrinkage which occurs above the beam because that is where the slab is thinnest. If we are correct and you pour strips parallel to the joists, the cracking over the beam will still occur. Thus my suggestion to pour the strips parallel to the beam.

 

[JAE]

This discussion is interesting because it stimulates the mind in figuring out why the slab may have cracked.

But I wouldn't worry about the cracks in suspended slabs. They aren't all that critical. Control joints aren't, and in my opinion shouldn't, be cut into concrete on metal deck slabs.

 

[MWPC]

hokie66 and 271828,

I believe I had read the posts carefully and agree that the slab is thinnest and would be most likely crack over the supports both parallel to the joists and perpendicular to the joists assuming that the beams deflect too. However one of the cracks over the beam was within about 8 feet of a construction joint and it was one of the larger cracks. If it was only shrinkage induced, I would not expect it to be that close to the end and that large. As BEFORT and the references stated "typically no control joints in slab on deck" since the deck acts as resistance to shrinkage. With this in mind, I would not have expected the large crack.

If the joists deflect enough to make the slab significantly thinner then they deflect enough to cause negative moment in the hardened concrete over their end. If a lightly reinforced 3" concrete slab was cantilevered 8 feet, I would expect it to crack. If there was a little bonding of the concrete to the deck then the concrete was trying to stop the end rotation of the joists to boot.

While I agree that thinner slabs over beams are typically the cause of the cracks, in this case I stand by my assumption of negative moment being the major cause.

Thanks for the references and discussion. If I had more time I would measure all the crack widths and document their location along with the construction joint locations, joist sizes, beam sizes, spans etc and we could look at it analytically, but that doesn't pay the bills.

Ken

 

[BSE05]

MWPC

A very interesting post. I have seen a great number of cracked slabs with composite steel deck slabs. Some were office floors and some 125 psf mezzanines.

In two separate cases, one office and one storage mezzanine slab cracked a couple of years after the initial pour with a very load noise that alarmed the occupants and called for our investigation. It appeared that the cracks were negative moment stresses that belatedly let go.

Recently a viewed a large mezzanine slab framed with open web joists about 3 ft oc. The slab crack at almost every joist location, most were hair line, but some started to unravel under pallet jack wheel traffic.

I think the best solution to eliminate or at least keep the cracks tight is to limit beam deflections to L/480 to L/600 if you can.

An tell your Clients that this is the beauty of concrete "it cracks where it wants to!"

Any other thoughts on limiting deflection?

 

[jike]

I have always had fairly good luck achieving uncracked slabs by using draped WWF in commercial construction where I require continuous bolsters over all joists and beams. I have had similar good luck achieving uncracked slabs with draped WWF over composite beams plus #4 rebar over the composite girders. Make sure the As of the draped WWF can resist the calculated negative moment. I calculate the required negative moment over the composite girders using Appendix A, Design of Two Way Slabs, Method 2 from ACI 318-63. The effective slab thickness over the composite girder is just the slab above the deck which does not give much to work with.

 

[MWPC]

jike,

Good point. Keep the reinforcing in the top of the slab to limit cracking. I am not sure you can drape it in both directions with the steel joists at 2' o.c. but just keeping it in the top should help reduce crack width.