"As for the balcony having redundancy, I don't see any."
Let's suppose that the balcony was only supported by the outriggers, and further suppose that a sensible design would have made each of the outriggers capable of supporting the entire load of the balcony. Then, by definition, there is redundancy. Since the balcony actually has 5 joists in addition to the outriggers, I think that meets the literal definition of redundancy. The fact that they all failed at the same, or nearly the same, time does not void the redundancy. Even if the railings had been a tertiary support, the same rot that affected the joists could have just as easily been located on whatever the railing might have been anchored upon, since each fastener requires penetration of the outer wall, and runs the risk of water intrusion.
My point is that there are always failure points; they're unavoidable. Even the redundancy professionals get it wrong sometimes, such as in the case of the DC-10 hydraulic systems, which were fully redundant, except at one specific spot in the tail, which was tolerably protected, until the rear mounted turbine developed a catastrophic blade failure, and took out all the hydraulics which had no cable backup, since it was nearly impossible, statistically, for all redundant hydraulics to simultaneously fail.
The issue is that redundancy is a slippery slope; where do you draw the statistical line in the sand, and can you live with the consequences when the line is crossed. Much of redundancy reliability is tied to the statistical analysis of probability of occurrence. So, even in the best case, an unknown probability can be assigned a lower value than what is real. Or, equally likely, the probability is correctly remote, but that probability card just shows up in the deck.
Case in point, and even considering the other balcony failure just reported, what is the statistical likelihood of a balcony failure? Is it one failure per million balcony-hours, or one per billion, or even trillion? Would any seriously contemplate potentially doubling the cost of a balcony to prevent a one in a billion balcony-hour failure probability? What about all the other one in a billion balcony-hour failures? Do we protect against them as well? The balcony-hours as I use it here is the number of actual balconies multiplied by the number of hours, so assume the the US as 70M housing units and 1/100 have a balcony, and assume there is one balcony failure every 2 years would result in 1 per 12 billion balcony-hours as the failure rate.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
I too am a big fan of redundancy -- especially in places where you can do so easily.
Take a car. It makes a lot of sense to me to have a traction-control system. Divert power from wheels that are slipping to those that aren't. Good redundancy.
But you can't do it everywhere. You wouldn't mandate a fifth wheel just to provide redundancy in case the four wheels stop working. That doesn't make sense. Those first four wheels just need to be engineered to work.
Same with this balcony. Providing a second load path is good when you can do it smartly. But this cantilevered system just needs to be engineered to work.
I agree with much that's said here. A few thoughts:
[ol 1]
[li]Making cantilevered joists PT. Why not? Makes sense. [/li]
[li]Watertesting some percentage of the balcony water proofing before covering it with finishes. Preferably with a third party. Won't help deferred maintenance items but it should at least keep people on their toes during construction. I don't see why this doesn't make it into Masterspec, if not the codes. [/li]
[li]Drilling holes in the sheathing under the waterproofing. This could at least let water drain should it get trapped under a lapped flashing joint. The leak would at least be visible underneath.[/li]
[/ol]
"We shape our buildings, thereafter they shape us." -WSC
PT does not last forever. Just means the collapse will be some years down the road. Yes PT but that is only a small part of what can never be an entirely satisfactory solution. Annual inspection is another part. It is pretty hard to build redundancy into a cantilever. I like Hokie66's response, namely do not build cantilevered balconies in wood. Don't forget that PT when the lumber is incised as it generally must be, has a big strength reduction factor, but at least that can be accounted for in the design. When the wood eventually does rot, there seems no alternative at that point than to cut the balconies off.
On the basic principle, if not the application to a particular class of structures.
I do note however that there is a good deal of what I call either-orism in this thread.
Those who think additional redundancy in balconies is a good idea think the cost would be trivial, and those who think it would be a waste of time think it would be horribly expensive.
For a structure to be redundant to prevent this type of collapse, it has to last longer than the primary structure, not be affected by whatever might cause the primary to fail, and possibly take a higher load due to energy added to the structure during the failure of the primary support (as in dropping and being suddenly stopped by the back-up.)
If the designer knows all about this, why not use this to design the primary structure so it doesn't fail?
The belt-and-suspenders concept is appealing until the realization that neither belt or suspenders are under much load and that well fitted pants will stay up without either.
Regardless of cost, designing and building a structural back-up is very difficult, particularly to offset a basic deterioration unrelated to structural use. Even designing a structure to fail gracefully under overload when the parts are in like-new condition is hard.
Given sufficient time and money, we could all generate a balcony that would not have failed in the same way as this one. But they each may fail in a different way due to the same or different causes. What we do all the time is to make our designs "good enough" because that is the economics of our profession. In this case, we all probably agree that given the outcome, the design and/or construction was not good enough.
It is easy enough to look a failure and determine the cause. How we react as an industry is what matters. There were a series of things leading to the final catastrophe. A fault tree would be useful, but I don't feel like drawing right now, so below is a ludicrously over-simplified assessment of how things got to where they are now:
failure: balcony fails due to deterioration of multiple structural members arising from undetected (or uncontrolled) water intrusion over a period of time.
Anywhere along the road to failure, this could possibly have been avoided. Some controls work better than others:
If the water intrusion had been properly prevented
If the effects of water intrusion had been mitigated
If the materials used were not subject to water damage (treated lumber or other material selection)
If the engineer had recognized that water intrusion was likely and specified appropriate materials
If the water intrusion had been detected and fixed
If the builder and his personnel had known about this issue and been more careful
If the owner or architect had required special inspection of the waterproofing
If the engineer had added separate redundant load path not subject to water damage from the same leak(s)
(aka If the structure was redundant enough, in the right ways)
If the building had not been built because they could not afford to construct a building costing more than one designed to minimum requirements.
