Technical Question for Architects, Engineers, Project Managers

[Kelvin]

>>> message(s) received <<<

Conventional deck & beam systems are difficult to make economical because you need to build extra capacity into the beam to carry the concrete while it is still wet. Once the concrete does set up, you have often times many times what you need for your remaining floor load (some Additional DL and the eventual LL). In most cases, your "worst case" will be your "construction case"!

The "temporary PT" system gives temporary stiffness to the girder/beam while the deck is still wet. Once it sets up and you achieve a composite section, the system then has sufficient strength to carry it's own dead load quite efficiently.

I ran a few numbers to see if I could get a basic concept to work. Take an 18.3 m span (60') x 1.524 m spacing (5') with a 101mm deck (4"), plus 51mm non-structural topping (2"). I tried a few different W's and settled on a W530x74 (meaning 27" deep at 50 lb/ft) which ended up at 91% of flexural capacity during construction (also a small construction LL in place which later is removed). My span-to-depth is 28 w/o the deck, 22 with. Not bad. Oh yes, I'm assuming a 45 MPa (6500 psi) concrete which is common enough these days - esp. in the Pacific NW known for 19,000 psi concretes! The steel is the most common variety 345 MPa (50 ksi)

Now, I run an external PT on a custom built plate girder, although a stock W might be cheaper to use even if it does have a few extra kilos. The section I ended up was effectively a W520x44 (30 kg/m less than the first example). I build the same concrete deck (101mm thick) and let it get to 70% strength (3 days max.) The concrete is compositely bonded to the steel beam so I can release the PT but the stresses imposed in the section are now effectively "locked in". When the PT is released, it is like adding a new axial tension to offset the previously imposed axial compression, so while the two forces negate each other and lead to a state of zero axial force, the flexural and axial stresses are acting on two different sections (one non-comp, the second is fully composite). Long story short = there is still plenty of locked-in stress that has to be overcome by new floor loads before failure occurs. This system also has a max stress of 90% capacity, but it occurs under full D+L, not the construction scenario.

The LL that I used is 2.4 kPa (50 psf), somewhere between a lightly loaded room (bedroom @ 25 psf) and a congested assembly area (theater or mall at 100 psf). As you might expect, the first floor is basically stiffer, so LL deflection at midspan is ~12 mm (1/2"). For the second system, the deflection is greater, ~16mm (5/8"). For the record, neither of these values are especially disturbing and are at least 3x what the code would prescribe as a maximum.

Another question is - will it flex to the point of being annoying to human occupation? The first flexural frequency is 3.1 Hz and under normal walking, a person would expect to feel approx. 0.0025g's of acceleration. The threshold of human comfort varies, but 0.005g's is sometimes considered a reasonable point of tolerance.

The only real difficulty is making this system work is that the girder is very flimsey laterally and the PT force that I need to use is approx. 10x the Euler Buckling criterion. One can always develop methods for restraining the lateral torsional failure mode, so it does become a very intriguing possibility.

For this example then, we could conceivably chop the DL of the beam used from 74 kg/m to 44 kg/m, or 550 kg per 18.3 m span. Just in material cost, I would price that savings as between $600 and $1000. I would wager that it could be an economical solution, eh!

Fireproofing would likely have to be the old fashioned type (spray-on) but that could be done at the factory too (better QA/QC, more efficient productivities, etc.).

Seismic details would be easy enough to incorporate as it is essentially already a very common construction technique (slab on beam) so it could also be integrated into IFC walls or steel/concrete columns too.

Now all I have to do is figure out how much it would cost to retool a old fabrication shop to produce these units to see if the bottom line is workable!

[Dr. Smoke]

I am glad to find people who know their stuff, here.

An 18.3m span would cover any of my requirements, although 6,500psi concrete could not have a high percentage of recycled materials. (LEED-R)

I am not clear what you mean by:
"Now, I run an external PT on a custom built plate girder, although a stock W might be cheaper to use even if it does have a few extra kilos."
This seems to be a W form system with concrete beams, as part of the W form? If so, they are prestressed up, so as you say when PT is released they are in tension at the bottom. This seems necessary, and like it would reduce the tension on the bottom of the beams, while maintaining a flat floor above, as opposed to no camber? If so, seems like we could put alot of camber in it, to simultaneously increase strength and reduce weight?

I presume there'd be a plate on the ends, to spread the PT forces, as the ICF walls would not be cast at this point. Idea is to cast the walls and deck together, to save time. Would it be possible to cast the deck monolitically? IOW, why do you recommend 51mm topping? Hm, maybe a 6" ICF wall couldn't support such a deck system. Likely need to be 8" or 10" of concrete.

I'm a bit concerned that this custom solution would need to be stamped, and that plan review may be delayed. Are there prescribed methods for non-bonded PT? (Best to fly under the radar, as you know, my brother)

[Kelvin]

I'm not sure what you need to make your concrete LEED, but a 6500 psi mix (or at least my preferred batching) is an 8-10% Silica Fume blended cement and it could also include some Fly Ash as well. Both these are otherwise "waste" materials from the steel industry, so there is a basic recycled component.

The beam I'm using is a steel wide-flange, or W. These are commonly available and easy to fabricate to desired lengths etc. Concrete could just as easily be used, but then I would suggest using permanent prestressing as opposed to temporary pre-tensioning. I'm not sure how much recycled steel makes its way back into the system these days, but I know it is not insignificant. Use of rolled steel shapes, plate girders, or just common OWSJ's are likely to have some "green" value too.

The pre-tensioning would be carried out externally - that is to say, without actually connecting to the beam (kind of like a bow and string). I only grab hold of the ends and then wrench a tension into the bar(s) which are below the beam. During the tensioning operation, the beam would camber up 152mm (6"), but under the weight of the concrete placed the expected deflection would be 122mm (4-3/4"). You would therefore have a net camber (after the concrete hardens) of 30mm/1.25". Barely noticeable on a 60' span! Another factor not discussed at this point, is long-term creep & shrinkage effects in the system. Shrinkage will occur in this system, so that small camber may eventual disappear, but creep will be almost non-existant. Creep is the process whereby materials "creep" under sustained loading - meaning that they actually get a little weaker over time. What we do to account for this process is to keep the material properties constant but add a pretend load to achieve similar deformations (believe a lot of guess work goes into this one!!). In our case, the pre-tension disappears around day-3, so it's not a real concern.

The deck could be poured semi-monolithically. I would see units being installed in 10' widths (easy to carry on the highway) and then, once installed, you simply grout a small preformed gap between adjoining panels and you're done. The topping is sometimes used to get prefab units all with the right elevation in a floor system. For example, if all the units have some large unwanted camber, you then place a monolithic topping in the field to build a nice flat level floor. Or perhaps it's a roof and you want to build drainage lines into the surface. It's entirely optional and could be taken out just as easily.

I'm not sure what kind of loads you can carry with IFC walls, but additional reinforcement could be placed, or perhaps build a pilaster (rather than thickening the whole wall) to carry the extra load. If you don't use 60' spans (perhaps you are at 20' spans), and still use 5' beam spacing, then end reactions might only be 20 - 30 kN (6700 lbs max.) at each end. Placed on a 6" x 6" bearing, it would not amount to anything unworkable.

The floor units would also perhaps a end-regions left without concrete so that they would slip right into your wall and then the floor and wall develop some fixity (this may not actually be structurally desirable - so some thought would be needed here). But you are right, there would be some end plates or anchorages at the ends of the beams (left of from the tensioning op's) that could work to become anchorages in the wall system too.

And also, you are right about the requirement for a PE stamp. This little scheme is a bit unorthodox, but certainly not without precedent. I wouldn't advocate renting a buch of hydraulic jacks and experimenting on your next project, but with some planning it is doable! Precast prestressed concrete girders (for bridges) are very common and externally post-tensioned systems have been used as well.

[zilfondel]

From what I've seen (and I'm not an architect yet, nor an engineer) from an existing project U/C here in Portland using the Hambro system, it allows slabs of only 4" thickness, which allows you to embed all of your radiant floor heating and other cables in it. However, it can be tough to build, and some sections crack and need to be fixed during construction... apparently it's not a big deal, however.

The problem is that it needs a 24" deep area in the ceiling for the joists... requiring a drop ceiling for residential, which isn't too bad, since most of the high-rises in Portland do that anyway to run their air ducts and the other mechanical through it.