Ponding of Concrete Roof 2

[theonlynamenottaken]

I am evaluating a circular concrete tank roof. It is (was) a flat slab sloped in one direction at an initial (design) slope of about 1/16" per foot. It now has 9.5 inches of sag at the center, equating to L/67.

Without putting pencil to paper I know something has to be done. However, just to have calculations backup I started researching. I can find no guidance for ponding on concrete. There's plenty on timber and steel framing, but these don't translate well since they integrate the idea of individual members.

Does anyone know a way to do this other than an incremental analysis?

 

[MiketheEngineer]

What are you asking?? What the load can possibley be in this condition or the force to "reset" the concrete or actaul stress in concrete??

 

[csd72]

I would say this is definately a ponding issue at this level of sag.

 

[mijowe]

I am interested in more of the details, what is the diameter, how thick is the slab stc.

 

[theonlynamenottaken]

I realize there's a problem. Thats not being questioned. I'm asking if anyone knows of any short cut equations, Taylor series, etc - that can be used in lieu of an incremental analysis. I can approximately determine the internal forces and stresses at rest using the known sag dimension. However, the ponding loads to then add are parabolic and I'd prefer not to do an incremental analysis of these.

The clients know that it is a problem and needs attention, because I told them while I was standing on it with them. They want to know at what depth of water it will fail, regardless of the fact that such information is useless.

It is 53' O.D. and thickness varies from 11" to 14". Asymmetrical reinforcing to accommodate openings and concentrated loads. Built in 1978.

 

[msquared48]

Incremental loading on a plate seems to be the most realistic way to me to analyze this problem, unfortunately for you.

I am concerned though with his question as to what depth it will fail. As I see it, there are too many unknowns to properly answer this question without putting yourself and your company in a lot of avoidable litigation.

Personally, I would recommend a temporary roof drain be installed at the lowest point of the roof, perhaps a small electrically operated sump pump, and immediate measures be undertaken to effect a permanent repair. This could work for a short time for rain water while the fix was being designed.

My other concern is ponding and freezing due to melting snow. The temporary sump pump solution would not work well here.

Some questions:

What is happening at the connection of the roof to the side walls?

What is happening at the underside of the slab inside the tank?

What is in the tank, if anything?

Should the roof fail, what would be the effects on the surrounding area?

Does the local BD jurisdiction know about this yet?

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[cvg]

Structurally, it has already failed and is in the process of falling. What is their definition of failure?

 

[theonlynamenottaken]

They have already denied all confined space entry permits to inside the tank and are lining up a P.O. for me to design a replacement. I've suggested that they install a drain and they will either do that next week or begin demo. Until then they've restricted all access.

This is an industrial client and all the client contacts are mechanical engineers. They want to know the depth at failure merely because they're curious. They were all making their guesses when I left. These were the same guys that wanted me to finish doing wind and seismic calcs to determine how to connect three existing pressure vessels, stacked vertically, to become a tower - even after they had given up and ordered one fresh from a fabricator due to discovered material flaws.

To answer your questions:
- The simple supports at the top of shell/wall have rotated significantly. Hence my recommendation to replace rather than repair.
- No clue about the condition of the underside of slab.
- Tank does stay about half full of Compound C87R or something like that. Secret and proprietary. SG = 1.20
- If the entire tank were to collapse the spill would be the worst consequence.
- The BD only wanted to ensure they'd pay for a permit when replacing.

 

[lexpatrie]

Well, the old school way to do this was to presume a sag and apply the sag-induced load over the whole plate. If your new design met the "no deflection checks needed" slab thickness-span ratio, I suppose you could presume the deflection at center is something like L/240 and use that as the initial applied load. Or you could design it for a flat load of X and state that on the Con Docs. And show a drain. You might also put a requirement that no more than X psf of snow be allowed to accumulate on the cover.

Presuming a uniform slab load from ponding is very conservative, but it does have the advantage of simplicity (and did I mention it's conservative?).

As to checking in-field deflections, all I can do is point you at the long-term deflection calculations in ACI (i.e. creep et. al.)

True failure load? True failure deflection? Perhaps a hinge mechanism can be presumed and a failure load derived from that, with the presumed plastic hinge rotation of 2x the theta needed to get Mp in the plate? If those mech engineers are as technical as they sound and as curious as they sound, you might point them in that direction, since at large deflections the plate is no-way elastic throughout, and the elastic deflections can be largely ignored once the hinge mechanism forms. But the embedment and development and hooks would all have to be adequate to develop the full force in the bars, and so on. At that point, you'd probably get concrete crushing and eventually the bars have to snap at something like 200 ey. I imagine the extreme yielding would be limited to the bar that's exposed to air at the hinge, but you're making a fair leap from what's been researched and studied, after all. But so long as everyone's doing it as an exercise / betting pool, it's an approach. Oh, yeah, you're also neglecting any potential arch / wall action, which I believe can be significant but also, very hard to quantify in a practical sense.

 

[paddingtongreen]

If they mean "when will it fall?", they are asking the impossible. When all the moment capacity is gone, there will still be the tension network provided by the reinforcing. I suspect that long before it falls, the water will find a way to drain into the tank. Simply performing incremental loading will not work because the support conditions will change with each iteration. As the lid sags more, the walls will be put in compression. This is not my area of expertise, but with the effect of openings spoiling a simple analysis, I wonder if some form of yield line theory could simplify the iterative process.

Michael.
Timing has a lot to do with the outcome of a rain dance.