June 4, 2002

9:17 AM

What do a 17th-century Swedish warship, an opulent Chicago theater and a Kansas City hotel "skyway" have in common? All met catastrophic ends—and they have important lessons to teach today's innovators.
posted by Irontom (24 comments total)

I lived in Kansas City during that walkway collapse. Years later, in the audience of the Rocky Horror Picture Show, it was memorialized in a joke - during the scene in Frank N. Furter's lab, as the camera pans the assembled Transylvanians on the upper level, the audience would yell "get off the skywalk, don't you know this is Kansas City?!"
posted by dnash at 9:29 AM on June 4, 2002

In Stockholm, there is a great museum devoted to the Vasa--the 17th century ship that sank within 30 minutes of its initial launch. They have recovered the vessel, which is very nearly intact, and built the museum around it, filled with information about both the technological hubris that led to the ship's faulty construction, and lots of horrific background about what it actually meant to be a sailor at that time.
posted by Rebis at 9:46 AM on June 4, 2002

Good article. Engineering failures are almost always more interesting than engineering successes.

For other disaster afficionados, I recommend the book "Inviting Disaster" by James R. Chiles. Excellent reading, and provides a lot of fascinating details of many well-known and not-so-well-known engineering failures that ultimately led to disaster. There are also some examples of disasters that were avoided.

And there's also a website.
posted by groundhog at 9:51 AM on June 4, 2002

This reminds me of the PBS special on how the WTC collapse happened. It was consluded that the crash impact blew the insulating agent off the steel supports making them vulnerable to heat from the ensuing fire. As the supports weakened, the bolts holding the floors in place gave out and eventually one of the floors fell. Since the floors were made of poured concrete they were very heavy and it only took one floor falling the distance of one story to trigger the entire collapse. Although the WTC collapse was (rightfully) not seen as an engineering mistake, it was known that the force of a single concrete floor falling to the one below it would cause every floor below it to collapse. Thousands of buildings across the country are built with the same floor construction.
posted by plaino at 10:27 AM on June 4, 2002

Another good book is "Engineers of Dreams" by Henry Petroski. All about bridge design, and how big disasters tend to fuel revolutions in bridge construction. I have always been fascinated by the "design by disaster" ethos -- after the Hindenburg, no more zeppelins; after Challenger, no more o-rings, etc. And won't all the new skyscrapers be 747-proof now?
posted by drinkcoffee at 10:32 AM on June 4, 2002

Thousands of buildings across the country are built with the same floor construction.

Not true, actually. The WTC was fairly unique in its design, which allowed for a column-free interior between the inner core and the outer wall. Pretty much every other skyscraper uses more conventional column-beam construction.
posted by laz-e-boy at 10:42 AM on June 4, 2002

For a short look at the Vasa failure from a software project management perspective, see Going Down with the Ship.
posted by dws at 11:07 AM on June 4, 2002

And there are still O-rings in the shuttle boosters. Now they respect the temperature parameters for launch.
posted by NortonDC at 11:07 AM on June 4, 2002

The earlier thread about the Molasses Disaster of 1919 popped into my mind.
posted by MJoachim at 11:14 AM on June 4, 2002

plaino - no building (except maybe some military bunkers and the like) are strong enough that a single floor can fall and the building will survive. The forces involved are just too great. For some good news, it looks like the changes needed to make buildings airliner-proof aren't fundamental structural design changes, but rather more attention to details like more damage-resistant fireproofing, redundant escape paths, and redundant sprinkler systems. It's just a matter of tweaking building codes and spending a bit more money.
posted by jaek at 11:16 AM on June 4, 2002

Don't forget "Why Buildings Fall Down" and "Why Buildings Stand Up".

Engineers graduating from Canadian universities go through a ceremony known as "The Ritual of the Calling of an Engineer". At the end of the ritual and their oath (written by Rudyard Kipling), they bear a wrought iron or stainless steel ring on the little finger of their working hand that is meant to symbolise pride in their profession and to remind them of their responsibilities.

There's an urban legend that the original set of engineers' rings were constructed from the steel of a poorly constructed railway bridge in Quebec that collapsed in 1907, killing 75. It has also been asserted that wearing the enginerer's ring is meant to remind them of that tragedy, but there's no evidence to support this story. Nonetheless, practically every Canadian can spot an engineer in a crowd (even before s/he starts spouting bad puns or discussing SF) and most still believe the bridge story.
posted by maudlin at 11:31 AM on June 4, 2002

One of my engineering instructors had helped with the Kansas City walkway investigation, and developed a slide show for the first day of class that included some really gruesome pictures of that and other engineering failures.

His closing point after walking us through 50 minutes of assorted mayhem:

"As a surgeon you hold a person's life in your hands, and people consider this an awesome responsibility. As an engineer, you hold the lives of potentially thousands of people in your hands. So pay attention in class."

Yikes. I paid attention.
posted by dglynn at 11:39 AM on June 4, 2002

There's a chapter on the Vasa in Evan S. Connell's The White Lantern; the amount of buck-passing that went on after it sunk will be familiar to anyone who reads the newspaper. It was an embarassment to the Swedish navy, the king, and all involved, so after salvaging as many of the cannons as they could using seventeenth-century technology (a diving bell, which I believe was a new invention), they just left it there.

In 1956, some clever maritime detective work rediscovered the hulk of the Vasa; efforts to raise it began in 1959, and in 1961 it was hoisted to the surface; the low salinity and oxygen level in the water where it landed had preserved a great deal of the organic material (including the sail, amazingly enough), and the Vasa artifacts now live in a museum.
posted by snarkout at 11:45 AM on June 4, 2002

Which is, you know, just to expand on what Rebis said -- sorry, Rebis, I missed your comments when reading the thread.
posted by snarkout at 11:48 AM on June 4, 2002

I often wonder about the psychological effects of having caused one of these disasters. Is there an 80-year-old former draftsman or engineer, sitting in a bungalow someplace, thinking how he contributed to the Kansas City walkway collapse? Is the manager of the construction project, or whomever else may have authorized the fatal cost-cutting step still alive? Does he or she hear the screams in his or her sleep? We are all guilty of lazy negligence some time at work. How would we feel it had such spectacularly deadly consequences? Gives me the chills...
posted by Faze at 12:03 PM on June 4, 2002

I remember reading this New Yorker article on the collapse of the WTC and feeling so sorry for Leslie Robertson.
posted by maudlin at 12:22 PM on June 4, 2002

That's awesome, maudlin -- it's too bad this tradition doesn't exist here. Although, I suppose as a computer engineer, I probably won't be responsible for thousands of lives. "The poor pipeline implementation and abysmal branch prediction of this processor went on to inconvenience HUNDREDS of innocent lives."
posted by Eamon at 12:53 PM on June 4, 2002

Sorry, Eamon, but software jockeys can kill, too, although not always in staggering numbers:

"Twenty-eight patients at the Panama National Institute of Oncology were overexposed to radiation during radiation therapy for colon, prostate, and cervical cancer. The overexposures ranged from 20 to 100 percent over the prescribed dose. Reports to date indicate that nine of the patients have died, with five of the deaths attributed to radiation overexposure. Many of the remaining patients are expected to develop serious radiation related complications. ... the following factors contributed to the overexposures: ... the method of entering beam block data into the Multidata software; [the] interpretation of beam block data by the Multidata software ..."

The equipment -- the Theratron 780-C -- was built by a Canadian firm and the software was developed by an American company. A smaller number of deaths have been reported from software errors introduced into the Therac-25 linear acceleration used for radiation therapy (another Canadian device, developed by Atomic Energy Canada Ltd.).

But we also gave you insulin, so don't hold it against us, OK?
posted by maudlin at 1:12 PM on June 4, 2002

snarkout, Let me second your recommendation of Evan S. Connell's "The White Lantern". It's a great, great book of historical quests and misadventures that I think would appeal to everybody on this thread.
posted by Faze at 1:36 PM on June 4, 2002

Actually, Engineers in the US can join the Order of the Engineer. I was sworn in at Vanderbilt University (Link 14) last year when I graduated in ChemE. It's somewhat obscure here, but does exist.
posted by antimony at 1:50 PM on June 4, 2002

Faze: as the article notes, Mulholland was devastated by the dam failure (though it was not actually his personal responsibility, he had publicly assured of its safety the day before). And Andre Coyne, one of the designers of the Malpasset Dam, was personally shattered and died six months later. (Another engineer on the project was indicted for negligent homicide, though he was later acquitted.)

Another approach is taken by people like Swede Momsen, a submarine officer who was so affected by his boat's failure to rescue the crew of another that sank in 1925, he devoted years to solving the engineering challenges of submarine rescue -- and in 1939, his techiques were used in the only successful such rescue in history, of the Squalus.

Engineering disasters have always fascinated me, though I didn't follow that career path myself. A while back a blogger noticed an item about engineering students who had cheated on an exam in their ethics class -- and he wondered why they were being made to study ethics, which he connected with foggier pursuits such as philosophy. I pointed out that, in fact, Engineering Ethics is a pursuit unto itself. Some of the most famous case studies are the Challenger disaster -- Roger Boisjoly, the O-ring whistleblower, is held up as an emblematic hero -- and LeMessurier's Citicorp Tower design, which was (like the Hyatt Regency walkway) changed during construction such that a catastrophic collapse appeared possible if a hurricane were to strike Manhattan. LeMessurier worked with the owners to retrofit the tower after it was already occupied, strengthening it against dangerous wind shear, and it is believed safe today.

I used to link to that example, shuddering at the thought of a skyscraper collapse -- never imagining I would watch it live on television. Twice.

NortonDC: The O-ring is necessitated by the segmented design of the Solid Rocket Boosters (which is itself necessitated by a variety of other factors). The original design was vulnerable to temperature extremes in ways that were not anticipated. The modified design is less vulnerable by intent to several specific failure modes. But the real problem with Challenger wasn't a bad design per se or the envelope-pushing, it was a faulty feedback mechanism within the organization. There were people in this case who knew there was a danger, and there was a specific decision made not to communicate that information.
posted by dhartung at 4:24 PM on June 4, 2002

Okay. How is that different from what I said?
posted by NortonDC at 5:01 PM on June 4, 2002

Norton, you barely touched on the relevant issues. For example, they launched Discovery in temperatures in the lower 40s just over a year ago. (In this case, though, it was a night launch, not a mid-morning launch after a freeze. And nowadays they use heaters on the pad to make sure the rings stay flexible.)

In point of fact, at the time of 51-L, there were no set "temperature parameters". To say that there were is to misunderstand the situation. The only line that was crossed, from the FD perspective, was the line of previous coldest launch, a line that had been crossed several times already. They didn't have the data that would tell them that there was a relationship here to O-ring damage. If Thiokol had told them "The SRB is only rated to 50 F" then your statement might have some relevance. That wasn't the case. Indeed, NASA contributed by putting Thiokol in the infamous position of "proving it wasn't safe to fly" rather than "proving it's safe".

I'm not really trying to rag on you -- I was only trying to clarify. But your statement was something like saying that the Johnstown dam failed because the lake was too full, which only addresses the physical failure mode. The causality for that one goes to modifications that reduced the dam's overflow -- and which wasn't a problem until an unusually heavy rainstorm.

Put it this way. Engineering is not just about respecting limits. Engineering is about reaching to our maximum grasp, and that reach is constantly extending. Many new bridges are cable-stayed, for instance, and it seems that almost every new cable-stayed bridge means a handful of patents on new ways of attaching, threading, or interlocking the cables, or strengthening the pylons while using less steel, and so forth -- all while building bridges safe enough to carry thousands of cars a day for the next half-century. To say we should simply respect known limits is to end that innovation.

The entire literature of engineering failures, the Levy and Petroski books, and the ethics case studies, and even the post-collapse studies of the WTC, are all attempts to provide a feedback mechanism to the engineering community, so that bridge or 'scraper builders everywhere will have a clearer sense of what mistakes to avoid. One of Challenger's most important lessons is that this feedback was short-circuited.

I mean, we already knew that things freeze.
posted by dhartung at 1:15 AM on June 5, 2002

There were people in this case who knew there was a danger, and there was a specific decision made not to communicate that information.

That's incorrect. The danger was communicated directly to NASA, as laid out in this article I blogged a year and a half ago.

The previous evening Boisjoly and Ebeling had spent six hours in teleconference with Nasa managers arguing that the Challenger launch should be delayed. The two men had been told that the temperature in Florida was plummeting to below freezing and had been instantly concerned about whether their rockets would perform properly in such conditions.

They appeared to be winning the argument - until their own managers turned against them and gave Nasa the recommendation they appeared to want: to launch.

NASA was ultimately given given a faulty recommendation, but they also had the danger directly communicated to them.

Additionally, the engineers at M-T did have temperature parameters: "Thiokol's bottom line was that it didn't want to fly outside its data base - which meant not launching below 12C (53F), the coldest launch temperature to date, and the temperature responsible, according to Boisjoly, for the damaged joints he had inspected a year before."

Additionally, if I had said "weather parameters" then your statement about heaters would be a relevant criticism. I didn't. I said "temperature parameters." The heaters are about the way they stay within temperature parameters, not the existence or absence of temperature parameters.
posted by NortonDC at 6:55 AM on June 5, 2002