Stupid Engineering Mistakes 592
lee1 writes "Wired has bestowed on us a list of the ten worst engineering mistakes of all time. We have the St. Francis Dam designed by 'self-taught' engineer William Mulholland, which burst and wiped out several towns near LA; the Kansas City Hyatt walkway collapse; the DC-10, and more, but my favorite is the one I'd never heard of: a giant tank of molasses that ruptured in 1919 and sent 'waves of molasses up to 15 feet high' through Boston, killing 21."
one comment, one addition (Score:5, Informative)
The Kansas City Hyatt was a disaster, but it wasn't because of bad design, but actually, "Construction issues led to a subtle but flawed design change that doubled the load on the connection between the fourth floor walkway support beams and the rods carrying the weight of the second floor walkway. This new design could barely handle the dead load weight of the structure itself, much less the weight of the spectators standing on it [wikipedia.org]". The original design would have been safe but what seemed an innocuous change completely changed the dynamics of load bearing, a result easily derived by any first year physics student.
Also, while a "top ten" list is always subjective, I think it'd be instructive to at least include Galloping Gertie [nwrain.com] as honorable mention, another design which had been identified as flawed. This Tacoma Narrows suspension bridge began swaying wildly as it set up its own harmonic resonance in a typical Puget Sound winter wind storm and eventually ripped apart and collapsed into the Sound. Interestingly the original Galloping Gertie could and would have sustained the fatal winds by strategically placed holes in the beams.
Vasa (Score:1, Informative)
Great out of print book (Score:5, Informative)
Amazon.com
The moral of this book is that behind every great engineering success is a trail of often ignored (but frequently spectacular) engineering failures. Petroski covers many of the best known examples of well-intentioned but ultimately failed design in action -- the galloping Tacoma Narrows Bridge (which you've probably seen tossing cars willy-nilly in the famous black-and-white footage), the collapse of the Kansas City Hyatt Regency Hotel walkways -- and many lesser known but equally informative examples. The line of reasoning Petroski develops in this book were later formalized into his quasi-Darwinian model of technological evolution in The Evolution of Useful Things, but this book is arguably the more illuminating -- and defintely the more enjoyable -- of these two titles. Highly recommended.
DC-10 Worst Engineering Disaster hardly... (Score:5, Informative)
And even those accidents the safety defects were quite minor, nothing major that one could claim that it was poorly engineered. Outward opening doors have been used on all aircraft, Douglas was the first one to make one as a baggage door for a production airliner, improper servicing lead to issues with the locks and finally two accidents, the final resulting in a bulkhead failing that sliced the control cables.
United 232 was a result of a failure of imagination, no one imagined that there would be a failure that massive that would severe all there hydraulic lines, even though they weren't placed next to each other (just near each other as they would have be as they have to run to similar areas of the aircraft). The engineer that designed it probably reasoned, that any failure that would result in all three being severed would be large enough that the aircraft would be lost.
Re:15 foot high waves of molasses (Score:4, Informative)
Re:15 foot high waves of molasses (Score:5, Informative)
Mullholland wasn't always wrong (Score:5, Informative)
Re:one comment, one addition (Score:5, Informative)
Re:Number 3, the Vasa (Score:1, Informative)
Re:Number 3, the Vasa (Score:4, Informative)
Now, back in the day good wood carvers were relatively cheap, so hiring a crew to gussy your ship up was, all things considered, pocket change. Think of it as the 1%-for-art stipulation that is built into many civic construction projects today. The result was your ship looked shu-weet, and so when it sailed into port everyone noticed, and talked, and generally got your nation some good press.
By the way, that's still a big deal in navel circles, visiting ports and showing the flag. These vessels have to do something, keep in training, and so doing diplomatic/PR duty is as good as many other things. Part of that is looking the part - now we go for angular grey steel & exotic weaponry, back then it was "I can afford to pimp-out-my-ship" gilding.
As to the decoration being heavy, the whole freakin' ship was "heavy", a layer of pretty painted bits was about negligible in effect.
Finally, your considered expert opinion on historical wooden sailing ships aside, the hull was perfectly fine for it's needs. Yes most i^Hg^Hn^Ho^Hr^Ha^Hn^Ht^H unsophisticated folks look at these ships and wonder "however did they stay upright" but they did. Much of the misapprehension comes from not understanding the weight distribution on these craft, the rest comes from not respecting the skills of it's sailors.
And, as has been doubtless pointed out several times already, the ship sank due to late-added lower gunports that were left open and effectively scuppered them.
4. Northeastern US power grid, 1965 (Score:3, Informative)
A single protective relay tripped in Ontario, overloading nearby circuits and causing a cascade of outages that left 30 million homes without power for up to 13 hours. A fragile, redundancy-free design ensured that it would happen eventually. After decades of repairs and upgrades, it happened again in 2003."
Although this point implies that the 2003 outage originated in Ontario as well, a joint U.S. and Canadian investigation found that it originated in Ohio due to several failures of FirstEnergy corporation, among them the failure to keep trees near high voltage power lines adequately trimmed! When the Eastlake generating plant in Ohio went offline during a period of high demand, other high voltage power lines in the area experienced increased demand to pick up the slack. The increased current across these HV lines caused them to sag and short-out when they came into contact with said trees. HV lines heat up and sag as current increases, and this is accounted for in both their design and in guidelines for keeping trees near HV power lines trimmed, which were apparently not adhered to by FirstEnergy.
This wasn't the only thing that FirstEnergy did wrong however. In total, they were found to be in violation of *seven* NERC standards. Although more reliability and redundancy could be built into the North American power grid, blaming the 2003 outage on poor engineering is not accurate. It was FirstEnergy's failure to adhere to standards that precipitated the cascade failure. As such, it would be more accurate to blame greedy corporate management that was too cheap to shell out adequate funds for operation.
For more on this, check out the report found here:
https://reports.energy.gov/BlackoutFinal-Web.pdf [energy.gov]
Re:Number 3, the Vasa (Score:3, Informative)
But at least it got Stockholm a pretty impressing museum.
Re:No Asian disasters? (Score:3, Informative)
The Toronto Skydome beat them by 8 years.
And Montreal's Olympic Stadium by at least 5 more years. But the important point (as a former SkyDome employee) is that SkyDome was the first retractable roof stadium *which actually worked*.
Re:DC-10 Worst Engineering Disaster hardly... (Score:2, Informative)
Lo and behold the door blows and the floor partially collapses. Fortunately only a couple of people get sucked out of the plane before the airplane bar lodged in the hole. You might think that they would fix the problem after this first incident but nope, Turkish Air 981 loses its door and the floor catastrophically collapses severing the control cables within it and the plane crashes killing everyone on board.
As to the design of the hydraulic lines, well Lockheed got it right in the L-1011, they put locks on the lines so that in the case of loss of pressure the lines were sealed so control could still be maintained. The reason they did this is because all of the hydraulic lines in the L-1011 like the DC-10 pass right next to the tail engine. Why did Lockheed know they needed this? Because they knew that ENGINES WILL INEVITABLY FAIL and tear apart and the loss of a single engine should NOT bring down a plane. So Hmm, sending all of the unprotected hydraulic lines right by an engine which will most certainly fail on some flight is stupid. The worst part isn't the flaws in the DC-10, the worst part is the criminal negligence of McDonnell douglas to not fix problems they knew were there and to not acknowledge their responsibility for the disasters.
On the DC-10 (Score:2, Informative)
It's very unfair to group the DC-10 with these disasters. McDonnell Douglas was actually very little at fault [wikipedia.org] for the 3-4 accidents that unfortunately occurred right near each other. The most spectacular crash of the American Airlines flight was actually caused by an AA maintenance crew being dumb and cracking the pylon holding the engine. But thanks to the American sensationalistically hostile TV media, the only thing that everyone saw was the engine falling off the wing, which led everyone to assume it was the DC-10's fault, and led to huge cancellations on flights on the actually safe DC-10. It was a good airplane destroyed by bad press and bad luck.
(If any of you have read Airframe [amazon.com] by Michael Crichton, you'll know what I'm talking about...from the NYT review [nytimes.com] of that very good book:
"And, Casey explains, when something goes wrong, a media industry that has grown hostile and shallow with the ascendancy of television always jumps to the wrong conclusion. Why, just look at what happened to the DC-10, ''a good aircraft . . . destroyed by bad press,'' because the crash of an American Airlines flight from Chicago to Los Angeles in May 1979 was misreported and misunderstood. ")
Re:one comment, one addition (Score:3, Informative)
While the original design may have been safe in theory, it was unbuildable. The supporting rods would have needed to be threaded for their entire lower half (which wasn't in the original design) in order for the loadbearing nuts for the higher walkway to be put in place. And that threading would have been damaged to the point of uselessness when the top walkway was raised into place. The original design was flawed. The disastrous change was made to fix it.
Re:15 feet high? (Score:5, Informative)
Re:DC-10 Worst Engineering Disaster hardly... (Score:3, Informative)
Entirely possible, but that would have had nothing to do with the accident. It was the *tail* engine that threw a compressor disc, and severed the hydraulic lines where they ran through the tail.
The less obvious lesson of that disaster is to have multiple ways to let the operator know what's going on. The pilot lost some sensors and instruments when the engine peeled off.
I'm not sure what flight you're talking about. When you lose all hydraulic controls, you notice instantly. The engine didn't peel off, it essentially exploded. The only way they could steer the aircraft was by differential throttle inputs to the left and right engines. That anyone survived at all, let alone something like half the people on board, was purely because of the skill of the folks who were on the aircraft that day.
Re:Lake Peigneur (Score:3, Informative)
Not enough to matter.
The dominant source of angular momentum in the water of a lake will be the currents from the entry to the exit channels, which will have some offset from dead-on toward each other and the center of the lake, along with the other currents (such as half-lake-sized eddies) they cause. The momentum from the earth's rotation will be orders of magnitude down.
Seems somewhat arbitrary to me (Score:2, Informative)
The deHaviland Comet. Stress concentrations and metal fatigue resulted in the loss of several aircraft.
HMS Titanic. Inadequate watertight compartments (IMHO the bulkheads should've extended higher, and/or been closer together). Also too few lifeboats to accomodate everyone on board.
Hubble telescope. Nno loss of life, just extremely bad press on a very expensive engineering program. 100% avoidable too.
Denver airport luggage system. No loss of life, unless one of the engineers jumped. The automated system was very expensive, late, and never worked correctly. To the point that the airport is using a normal manual labor system and has given up on ever using the automated system. (but is still paying for it)
Chernobyl (sp?) and/or Three Mile Island. Safety equipment, procedures, and training obviously not up to the task.
Any one of several early Soviet nuclear submarine designs. That more of them didn't sink or irradate their crews (more) is a credit to the bravery and dedication of their crews.
The main thing to look for in a "worst engineering mistakes" list would be something that not only seems obviously a bad idea in retrospect... But that should've been recognized as a really bad idea, even with the technology and education levels available at the time.
Re:one comment, one addition (Score:3, Informative)
No; that was true of both the original and the assembled plans. What failed was (a) the original plan could not be put together as designed and (b) the suggested change seemed innocuous to the guy on-site.
The plans were for one rod to carry the weight from the ceiling through to all the walkways, being threaded at each level and bolted on. Problem was, you can't fit a threaded rod through a straight hole. So it was changed to make it possible; only now rods weren't simply under tension, but I-beams were under torsion too. So they tore out.
One error + another error = 144 dead and 200+ injured. But not the way you said.
Source: Why Buildings Fall Down, pp. 224-229, Levi & Salvadori. If you don't have this book on your shelf, why don't you?
Re:15 foot high waves of molasses (Score:4, Informative)
A large molasses (treacle) tank burst and a wave of molasses ran through the streets at an estimated 35 MPH (56 km/h), killing twenty-one and injuring 150 others.
The collapse unleashed an immense wave of molasses between 8 and 15 ft (2.5 to 4.5 m) high, moving at 35 mph (60 km/h) and exerting a pressure of 2 ton/ft (200 kPa).
Google calculator shows:
35 miles = 56.32704 kilometers
Re:one comment, one addition (Score:1, Informative)
The original design was also a box beam created out of two C-sections. The biggest problem was that the design of the support rods was changed. In the original design, the support rods were continuous from the ceiling, to the upper bridge (4th floor), and then to the lower bridge (2nd floor). The upper bridge would have been supported by a nut threaded all the way up from below the lower bridge.
As built, there were two separate rods. One rod from the ceiling to the upper bridge. A second rod went from the upper bridge to the lower bridge. As designed, the upper bridge support beams only supported the weight of the upper bridge. As built, the upper bridge support beams now supported both the upper bridge AND the lower bridge. In other words, the design change DOUBLED the stress on the upper support beams.
An additional design problem was that the connections at the support beams should have had a cover plate welded over the box beam. The designed connection was vulnerable to a ripping failure of the weld.
I understand why the walkway was not built as designed -- the original design was basically unbuildable. The change from one continuous rod to two rods made sense. The problem was that the support beams and connections were not resized as required.
Comment removed (Score:5, Informative)
Truth on front-loaders (Score:5, Informative)
1. It's actually your great-grandmother's suffering you're reliving. You see, the way to wash the sweat and human oils out of clothes was to take the big pot (like a witch's cauldron) and make Clothes Soup over an open fire. So good job on advancing yourself to 1890.
2. If you went back to freshman chemistry, you'd learn that water and oil do not mix. Which means, if you want to get the human soils out of your underwear, and the human sweat/grease out of your clothes, you're going to have to use soap. Water won't do it. Or, if you don't believe me, just stop buying laundry detergent. You do use it, right, hypocrite? FYI: The water is the medium for the soap, and removed soils. It all has to go somewhere - the soap alone won't carry it.
3a. A liberal arts guy, huh? 'Nuff said.
3b. Just for general info, did you ever see what your top-loader does with your Clothes Soup? The paddle in the middle spins a turn clockwise, then a turn counter-clockwise....and so forth. It also has to spin the drum for the spin cycle (you know, the only major moving part on a front-loader). So you have 2 major moving parts, one of which has to support counter-movement. So you're actually on the WRONG END OF THE SIMPLICITY ARGUMENT. Duh.
You do have the efficienty argument down, though. Front-loaders use 40% less water and much less soap, along with being much easier on the actual clothes because there is no paddle-like implement used to pummel your clothes. Gravity and water do that for the front-loader, off that one mono-dirctional moving part.
4. So...you do change the water in your washing machine from time to time, right?
How do you get it out?
Could it be...........a cute little rubber seal? At the bottom of the drum? Under way more standing water pressure than a front-loader sees?
PS: Check into how long Mankind has been making watertight seals. I bet you'll be suprised. We've had time to actually get kinda good at it.
How the hell did your particular brand of idiocy get modded up?
Re:Number 3, the Vasa (Score:2, Informative)
The original design had two rows of cannons. The King insisted on a third row, placing the new row of ports far closer to the waterline (and hence limiting the heel of the ship under sail).
As a result of the additional weight above the waterline (from all the extra cannons), extra ballast was required below the waterline to prevent the entire ship from becoming top-heavy. This merely exacerbated the problem of the lower row of gun ports by raising the waterline.
In the end a 5 knot breeze was sufficient to heel the ship enough so she began taking on water through the lower gun ports, with the expected result.
Oh, and the waters of Stockholm harbour are brackish, with salinity levels below that favoured by woodworm. Hence the preservation.
Re:one comment, one addition (Score:2, Informative)
What? Field changes are not the same thing as plans. I'm not sure where you work but at my job if someone asks me for "the plans" and I hand them a folder of field changes I will probably hear some select four-letter words and be sent back to get the actual "plans". Now it's true both will be used in construction, but trying to use the two terms interchangeably is wrong and will only confuse people.
Correct... (Score:5, Informative)
We studied this case, as well as many on the list above, in detail. In particular, the box beams in question ran horizontally to support the walkway, while the vertical rod was the support for the end of the box beams. The beams could have been made better, but they were good enough for their design loads.
The problem was that the original design called for one continuous vertical rod, with several levels of walkway hanging from it at different heights. However, due to construction issues, the installation was changed (for the worse) so that separate vertical rods were used. This unfortunately got written approval, and shouldn't have. Instead of the successive loads being applied to the rod, the box beam was then holding the weight of all the floors below it, which it was not designed to do.
Imagine one rope hanging from a ceiling, with 3 people hanging at various heights on the rope. The rope can hold the total weight of the 3 people easily, but each climber needs only enough grip to hold up his own weight. Now imagine due to "construction issues" you can't get one long rope, so you get 2 shorter lengths. Ideally, you'd tie the ropes together to create a nearly identical scenario, but in this case, it's like they tied the bottom rope to the middle guy's ankle, and expected him to hold on with the added weight of the guy below him.
Unfortunately, it was just strong enough to hold a few people, but let go when it was fully loaded.
=
Re:Number 3, the Vasa (Score:3, Informative)
A major factor was that the king ordered another row of cannons added to the design to increase firepower and make it look more impressive. They did do stability tests by having sailors run en masse back and forth across the deck, but it started tipping so dangerously they had to stop. Even so the people in charge didn't dare to go against the kings wishes. And down it went...
If you ever get to Stockholm, the Vasa museum is defenitely worth a visit.
Re:No Asian disasters? (Score:5, Informative)
The Romans beat you by almost 2000 years. The Flavian Amphitheater had a retractable roof.
Re:Number 3, the Vasa (Score:2, Informative)
They have a full featured movie that explains how it sank, and how it was brought back up. How it sank was the really interesting part. It's something all programmers and engineers will relate to: last minute changes to the design.
The king at the time (I think it was Gustav Vasa) decided he wanted the biggest ship in the world. And bigger meant more guns, so he asked for a second level of guns when the ship was already half built. That's a completely new deck of iron cannons on a gun that was designed for only one. Since it was a request from the king, nobody dared say no.
So the second row of guns was added, pushing the boat far lower in the water than was originally planned. So far down that the water line was only a few feet over the lower gun ports. Worse, because the boat was already so low in the water, they couldn't add additional ballast (ballast is the weight at the bottom of a boat that keeps it pointing up). Ballast is critically important to sail ships, since it counters the rolling effect of the wind. So sure enough, the first gust of wind to hit the sails caused it to tip far enough that water came through the already too low gun ports, and sure enough, it capsized and sank.
The reason it was kept in good condition is because of the silt, and also the salinity of the water. I don't remember if it's because it's too salty or not salty enough, but either way, woodworms don't like the salinity at that area, and so there aren't any there to eat the wood, so it kept really well.
The thing that amazed me the most at the museum was the main sail. Sails were kept in boxes at the time, to help protect them. One of the main sails was still in it's box when the ship sank. When the ship was brought back up, the sail was discovered, laid out on a huge piece of glass, and it's now on display at the museum, in remarkably good shape.
Phemur
Re:PEPCON rock fuel factory (Score:4, Informative)
The storage drums were plastic.
And wow, one of the two who died was in a wheelchair. Those folks must have some serious survivor's guilt for not helping that guy when they all ran and drove away.
Vasa (Score:2, Informative)
The ship sank some 10 to 15 minutes into it's maiden voyage. The exact location was forgotten. It was found as the result of one old fellow who spent years and years looking for it by taking core samples of the bottom of the sound every meter or so. The ship was then dug out of the mud by (now archaic-looking) dive teams, raised and then brought to dry dock where it is today.
IIRC the shipmaster died partway into the construction without a trained or skilled replacement. Unmodified, it would have not been noteworthy and maybe a little under armed. Adding the extra gun deck made the ship too tall and unstable. So to compensate, extra ballast was added, bringing the lowest gun deck about inline with the water.
Before launching, it failed the stability test of the time in which 40 men where to run in unision from one side of the ship to the other 40 times (or something like that). It was launched anyway, sliding nicely into the water, some sails were set and when it rounded the end of the island and caught it first breeze, it tipped and sank.
The sinking roughly co-incided with the end of Sweden as a feared superpower, thought it was only one factor of many.
Re:one comment, one addition (Score:1, Informative)
This is the original plan, note the continous support rod:
--[|]--
--[|]--
--[|]--
This is what got built, not how the structure of the first floor holds up the bottom ones when it wasn't designed to do so:
[|]----
--[|]--
----[|]
Re:Interesting factoid about the "Galloping Gertie (Score:5, Informative)
I read the Wikipedia article and found it very interesting:
The vibration had nothing to do with the resonance frequency of the bridge as a structure, but with the fact that it was wind (as opposed to some other form of energy input, e.g. sound) that was exciting the bridge. At a certain wind speed, the bridge enters a positive feedback loop - when the small motion induced by the wind changes the angle of attack in a way that makes the bridge absorb more and more energy from the wind, eventually increasing the amplitude of the oscillation to a point where structural failure occurs.
To make it short: The bridge did not oscillate at one of its resonant frequencies - aerodynamics caused it to vibrate at an entirely different frequency but managed to pump enough mechanical energy into the bridge to break it anyway.
Kansas City Hyatt walkways, not negligent design (Score:2, Informative)
the design specified that a for each support, a single rod would run vertically down, and each walkway sat on a nut on the rod. The rod was strong enough, each Nut could support a single walkway.
The incorrect interpretation meant that the rod terminated at the first walkway, and a new rod went down to the next level which then terminated, and a third rod then ran to the next walkway down and so on. With 3 walkways suspensed from a nut that was designed to handle the load for 1 walkway it's no surprise it collapsed.
a
Re:No Asian disasters? (Score:3, Informative)
Re:Three Gorges Dam (Score:2, Informative)
Re:Three Gorges Damn (Score:3, Informative)
http://www.reference.com/browse/wiki/Banqiao_Dam [reference.com]