Join 3,497 readers in helping fund MetaFilter (Hide)


November 23, 2001
4:52 AM   Subscribe

Apparently, the crash of Flight 587 was an accident. The official explanation claims that the American A300-600 encountered the wake turbulence of a JAL 747 which tore off its tailfin. Now, I am no fan of conspiracy theories, but the lack of historical problems gives food for thought (Tripod link via RobotWisdom). [more inside]
posted by costas (35 comments total)

 
The usual measure of an aircraft's wake turbulence strength is its weight and speed --the energy it puts back in the air, so to speak. Check out the JAL and Airbus links and the detailed specifications for the Boeing 747-400 (which according to the JAL website should be the type that took off ahead of flight 587). The A300 is about half the weight (~170 tonnes versus 390), while speed capabilities look much closer (0.82 Mach maximum for the Airbus, versus 0.85 cruising for the 747).

It is not intuitive to this former aero engineer that an aircraft will lose a major piece of structure if it encounters the wake turbulence of an aircraft that's just double its size in weight and 1.5x in wingspan (64m vs 45m).

Further, the BBC and other sources quote that the reason the vertical tailfin failed was because it was made of composite materials which can have hard-to-detect internal damage. Even if you ignore the fact that composites have been used in aerospace extensively for decades now, there is one thing about composites that is common knowledge: when they do fail, they fail badly (i.e. major cracks, tears, delamination, etc.). For example, wood and paper are a type of composite material: when wood fails, it splinters and cracks.

Look at that tailfin in both the BBC and Tripod links. That is not a cracked or delaminated piece of composite. It looks like it was detached from the root, where it meets the airframe. The frame, including the tailfin's own frame is made of aluminum.

I am not saying that 587 was an act of sabotage. However, I find the official evidence lacking.
posted by costas at 5:12 AM on November 23, 2001


hmmm. well even though i have worked in the industry and understand that critical failures do (rarely) happen in aviation, i just can't help but be ruled by statistic.

I work in engineering and so failures mean alot to me. If a customer phones with a failed unit, then I will usually accept this and carry on day-to-day. Last week a customer came back with 4 failed units. 'Hang-on' I said. 4 ? Later investigation found he had put 240Volts through 12V equipment.

I know it's a tenuous link, but the bottom line is 3 downed aircraft in the same city in fewer months. YES, the first 2 were undeniably intentional, but how strong has the fear been of planes downing in cities? I live in London so I often consider this.

My base theory is: another terrorist attack - on the same country - in the same city - on a now sufffering industry (last night in the UK we had Dubya speaking on an ad for holidaying in the US!)

Moo says...........cover-up.
posted by Frasermoo at 5:26 AM on November 23, 2001


I'd agree with the poster, the explanation doesn't quite seem to fit. What would be most concerning is if there wasn't any problem at all, and the crash happened simply because the plane was not capable of dealing with that level of turbulence.

If that were the case then the lack of margin for error that appears to be available to pilots on take off is a serious concern.

'To avoid running into this serious hazard planes are required to leave a gap of at least two minutes - or four nautical miles - before taking-off after the plane ahead of them.

Air traffic controllers did follow this protocol correctly, leaving a gap of two minutes 20 seconds in between clearing two planes to begin their journeys.

But a Japan Airlines jumbo jet seems to have paused slightly before take-off, leaving just one minute and 45 seconds between the two planes.'

posted by RobertLoch at 5:31 AM on November 23, 2001


A story on New Scientist suggests that maneuvers performed by the pilots or a failed autopilot attempting to recover after encountering wake turbulence might have led to the structural failure.

Can our engineers suggest a BS rating on this?
posted by rocketpup at 6:08 AM on November 23, 2001


Probably stating the obvious here, but it's possible (reasonable?) to believe that it was both sabotage and the result of a turbulent wake. Whatever secures the tailfin to the main body could have been tampered with so that although strong enough for takeoff, it would fail under the stress of turbulence (after all, turbulent wake isn't a closely guarded secret - I heard of this problem way before this happened, when it was used by a lecturer to spice up a fluid dynamics lecture).

Which, of course, makes any cover-up particularly effective, as only the degree of the effect needs to be "adjusted".

The biggest argument against conspiracy (imho) is the lack of a strong claim of responsibility (that I know of). If it was an Al Qaeda act, then you'd expect them to milk it for propoganda. That means that it's more along the lines of the anthrax scare, which lowers (in my eyes) the strength of the conincidence argument - instead of it being "just one more act" from Al Qaeda, it's a new kind of act, from someone else.
posted by andrew cooke at 6:13 AM on November 23, 2001


Oh, one other (obscure? interesting?) point. The "it can't just be coincidence" argument has a scientific basis in Bayesian probability theory. After 911 your "prior" (a mathematical thingummy that represents your view of the world) is strongly in favour of aircraft disasters in New York being man-made. So when another one happens, the probability of it being a "natural" accident is much less.
posted by andrew cooke at 6:18 AM on November 23, 2001


Robert: a) I don't believe that the FAA-mandated limit of 2:20 doesn't have a built-in safety margin, and b) the fact that NTSB found out about the JAL take-off delay belatedly indicates to me that it's not an unusual phenomenon; if the planes were that close the control tower would have delayed 587, or register the event as an incident (which would have raised a red flag much sooner).

Again, this is a well-educated guess, but I am not an air traffic controller.

Andrew: I agree; If it was sabotage, the turbulence was just the 'trigger'. As for claim of responsibility: well, there wasn't any for 9/11 itself, was there?

I am beginning to sound like a conspiracy nut... Well, I am off to board a plane and travel a coupla thousand miles: I commute across Europe every weekend...
posted by costas at 6:22 AM on November 23, 2001


I just read the New Scientist article too: it basically just points the finger at the loss of the tailfin as the cause (which everyone accepts as reasonable) but doesn't try to explain why this loss occurred. Of course the aircraft would have banked after losing the tailfin, as it would be unstable and the crew would have tried to use the rudder to compensate, which would have been ineffective (no rudder!) so it would induce more instability.

On the other hand, they quote Dr. Kroo of Stanford who doesn't offer any alternate theories. I've studied with Dr Kroo and he is the best aerodynamicist I know, so I now feel humbled and much more of a conspiracy nut...
posted by costas at 6:29 AM on November 23, 2001


conspiracy just works. although it is not unreasonable to take the view that America is itself in a time of War, therefore all actions taken to keep morale high and protect the interest of the country are acceptable - this argument opens a rather large and unsightly can of worms.
posted by Frasermoo at 6:38 AM on November 23, 2001


Before I entertain any conspiracy theories, I'd like to hear the results of the FAA-mandated inspections of the rest of the US A300 fleet's stabilizers. Were any defects found? If so, you'd expect it to be all over the news. When that Alaskan Airlines MD-83 went down in the Pacific a while back, you couldn't turn on the tv for weeks without hearing something about stripped jackscrews turning up in other planes.
posted by MrBaliHai at 6:47 AM on November 23, 2001


As for claim of responsibility: well, there wasn't any for 9/11 itself, was there?

But you'd think they might have leaked something if this was being covered up. The whole idea would presumably be to scare people.
posted by walrus at 6:47 AM on November 23, 2001


I'm sure that in a cave somewhere Allah is being given the credit.
posted by RobertLoch at 6:50 AM on November 23, 2001


Those pictures don't show undamaged composite spars; they show the undamaged aluminum skin. The composite structure would be underneath the skin and I can't see any of it sticking out the end of the tail, either.
If anyone knows of more pictures of the recovered tail, it would be helpful.
posted by cardboard at 6:55 AM on November 23, 2001


Were any defects found?

Note to self: must read all links in FPP before posting comment...
posted by MrBaliHai at 7:02 AM on November 23, 2001


After 911 your "prior" (a mathematical thingummy that represents your view of the world) is strongly in favour of aircraft disasters in New York being man-made. So when another one happens, the probability of it being a "natural" accident is much less.

Not sure I follow you here. How does one's "view of the world" affect the probability of an event? (We're not talking Schroedinger's cat, here....)
posted by rushmc at 7:17 AM on November 23, 2001


The conjecture is that the composites of the tail itself didn't fail, but the composites of the six attachment lugs did -- or that one of them did. Had that happened, it could have led to a catastrophic cascade failure of the other attachment lugs leading to a completely intact and undamaged tail separating from an equally intact and undamaged fusilage.

I'm really rather surprise that there are so few attachment points; prudence would suggest use of many more than that. The weight increase would not be huge but it would add considerable redundance. As it stands, if one fails then the attachment loses nearly 20% of its structural strength -- and is offbalance too. That's not good design.

Rush, your view of the world doesn't affect the probability of the event. But it affects how you would evaluate the probability of the event.
posted by Steven Den Beste at 7:27 AM on November 23, 2001


rushmc - What I was trying to argue against was the ("mathematically sound") view that if you throw two 6s in a row, it doesn't alter what happens next throw - it's still 1 in 6 that you'll see a 6.

That's true for a fair dice, but what Bayesian statistics do is let you handle a messy world in which things aren't equally likely. If someone keeps throwing 6s, there's also a probability that they're cheating - the more they continue to throw them, the more likely it is that the dice is biased.

My impression is that lots of people with a moderately good science education (Metafilter readers?) have had the first example (or its coin-tossing equivalent) drummed into them at school and so think that the "coincidence argument" is just our silly human minds finding patterns that don't exist, when it is, in fact, quite reasonable.

Or maybe I'm just paranoid.
posted by andrew cooke at 7:49 AM on November 23, 2001


im coming up 6's...It could be sabotage. a small bit of plastique...mix it in with a noise...thing is, if it was, few if any will ever know...I think Mr.Cooke has a good take on this.
posted by clavdivs at 8:15 AM on November 23, 2001


That's true for a fair dice, but what Bayesian statistics do is let you handle a messy world in which things aren't equally likely. If someone keeps throwing 6s, there's also a probability that they're cheating - the more they continue to throw them, the more likely it is that the dice is biased.

Nice arguement, but... In your example it's the same person throwing the dice, thus linking all those odd 6s together. For 587 though, you've linked it to 9/11 because they both happened in NYC, even though it could be coincidence. In other words there is a firm link between your crooked dice man throwing one time and the next (it's the same person!) but for 9/11 and 587 you've got a more fuzzy 'they're nearby'
posted by iain at 9:11 AM on November 23, 2001


I don't believe in chance or coincidences, not when it comes to things like this. I suspect sabotage and have from the beginning.

I think the authorities came up this this explanation because of the consequences for the real thing:
- Chaos and panic in the USA and very likely, the rest of the world
-The effect on the US economy and the world economy, would be too much to handle.

I also think we 'll never know what really happened.
posted by ginz at 9:15 AM on November 23, 2001


Chaos and panic in US? Maybe. As for the rest of the world, well, errr, no. Worried, a bit. Maybe.
posted by iain at 9:21 AM on November 23, 2001


The biggest argument against conspiracy (imho) is the lack of a strong claim of responsibility (that I know of). If it was an Al Qaeda act, then you'd expect them to milk it for propoganda.

Ah yes, if it is was an Al Quaeda act. But seeing mr. Vialls main site, I don't think in his ideas this conspiracy leads to Al Quaeda. For some reason all of the other conspiracy theories on mr. Vialls site seem to point to Israel...
posted by Berend at 9:37 AM on November 23, 2001


It's amazing how people see things the way they want to see them. With regards to the 9/11 plane that crashed in Pennsylvania--which I'm inclined to believe was shot down by US fighters--everyone's so quick to poo-poo such a "conspiracy theory," even though the White House has admitted that it would do exactly that if it were possible.

But here, where all the hard evidence points to an accident, everyone wants to blame terrorism.

I don't believe in chance or coincidences...

They happen whether you believe in them or not.
posted by jpoulos at 9:45 AM on November 23, 2001


But here, where all the hard evidence points to an accident

The trouble is, I don't believe "the hard evidence" . Turbulence? I think it's a fabricated story, but a difficult one to contest.
posted by ginz at 10:03 AM on November 23, 2001


Steven: Structural attachment points for wings and stabilizers are not supposed to fall apart when only one fastener lets go. Beyond the standard 1.5 safety factor on all aviation design, a fastened joint at a critical place like this would have a fair bit of thought put into it, and would have come under a lot of scrutiny by regulating authorities, who tend to be quite conservative, especially with regard to the use of advanced composites in primary structures, especially at the time the A300 was certified, when far less was known about composite behaviour.

The design strength of the joint would be determined with fatigue effects in mind (not static loads) and so the joint would be built to withstand many times any anticipated in-flight loads so as to keep part stresses low. The loss of one fastener might weaken the joint, but the joint would still be operating well below material limits, and would be designed to operate that way until the next scheduled inspection that would detect the failure.

The use of even more fasteners would complicate manufacture, assembly, maintenance and inspection of the joint, with minimal effect on its reliability. As you say, there may be design flaws in the airplane that led to this crash, but the numbers of bolts used to hold on its important pieces would not be one of them.


Ginz: shouldn't we leave the rampant speculation and conjecture to media professionals?
posted by cardboard at 10:12 AM on November 23, 2001


right, I'm no expert.
Ginz: signs off
arriverderci
posted by ginz at 10:17 AM on November 23, 2001


For a readable introduction to structural failure, both on aircraft and on larger things like building, bridges, and drilling platforms, I recommend Henry Petroski's To Engineer Is Human : The Role of Failure in Successful Design. Planes have been falling out of the sky for a long time, and the reasons aren't always obvious. (Did Icarus fall from the sky because his wax wings were melted by the sun? Or did the cooler temperature at higher altitudes cause the wax to get brittle and fail?)
posted by dws at 12:23 PM on November 23, 2001


It's possible to get cascading failures beginning with small vulnerabilities.

The DC-10 cargo door was an example of that: the bolts holding it on were just a little too weak, and once in a while they'd let go while the plane was at altitude. That depressurized the cargo hold of the plane, but not the passenger compartment, which was separated from it by a flat floor which would not have been able to stand up to multi-pound-per-square-inch pressure.

So part of the floor would collapse, and when it did so it would part a channel through which cables and other controls passed from the cockpit to the tail, causing the pilots to lose control of all the control surfaces on the tail. That made them lose control of the jet. Two different DC-10's tumbled out of the sky and crashed because their cargo-hold doors failed in flight.

As to the design of the six fasteners, I have no doubt that the designers intended what you say. But the Airbus A-300 was the first commercial jet to use those materials in that way, and the possibility exists that they truly didn't understand the materials that well. For example, long term chronic damage to the fasteners may not have acted the way they thought they did, and it may be that all of them had already weakened. When one let go the rest may not have been strong enough after all (because of micro-fractures) to actually hold the tail on.

I'm not an ME, but I'm an experienced engineer, and I know that no competent engineer says that the design will have no unforseen hazards. And they can pop up after years of faithful service. That's particularly true when you're pioneering and entirely new way to do things.
posted by Steven Den Beste at 1:05 PM on November 23, 2001


no competent engineer says that the design will have no unforseen hazards. And they can pop up after years of faithful service

e.g. o-rings ...
posted by walrus at 1:14 PM on November 23, 2001


What struck me was the possible conceptual similarity between this and the WTC towers' disaster in a sort of domino (butterfly?) effect, from the fatal weakening of a small area multiplying into a total collapse. The concept of undermining the weakest link that breaks up the whole.

Some interesting read about the structural issues with the towers:
http://www.newyorker.com/FACT/?011119fa_FACT
posted by semmi at 1:18 PM on November 23, 2001


Sorry, forgot to build the link, but here it is:
http://www.newyorker.com/FACT/?011119fa_FACT
posted by semmi at 1:45 PM on November 23, 2001


These two booke are materials/engineering classics and easy to read, if anyone is interested. Oldies but goodies.... (the second is more relevant here - if it was structural failure ;-)
posted by andrew cooke at 1:52 PM on November 23, 2001


Steven: my engineering days are gone (them databases have seduced me) but I did work on aviation safety and preventive maintainance for 1.5 yrs. I agree with you on the DC-10 or the Tacoma Narrows (?) bridge, but in aviation there is one solid rule when it comes to design flaws: you have historical symptoms. Maybe not a catastrophic failure but some symptom.

What the Aviation Week link on the FPP says is that *none* has been found. Moreover, AA which has a largish A300-600 fleet (~35) inspected all their other A300s and came up with nothing. This is not typical. It is also not typical that any article I have come across (and I have looked) mentions any other sort of historical problem with these fasteners.

That leaves only one plausible explanation that could justify a design flaw: that these fasteners have fatigued either under extraordinary circumstances before (like the incidents this particular aircraft was in before) or that their flaws only show at approximately this number of flight hours (or take-off/landing cycles). This is unlikely --as the inspection of other A300s, including older non -600s showed nothing-- but possible.

I am hopeful that the French would rather blow the cover of a potential cover-up rather than tarnish the reputation of their planes --which admittedly, have a better safety record anyway.
posted by costas at 2:14 PM on November 23, 2001


In the DC-10 case, the door failed because it was incorrectly secured. To answer why the door wasn't correctly latched will reveal a number of design flaws stemming from the use of a weight-saving latching system, but it does not appear that understrength bolts was one of them. My point here is not to pick nits, but to point out that structural engineers know how to pick bolts (and in my experience the regulatory process will correct them when they don't), even when everything else gets screwed up.

Mistakes and compromises happen and unforseen secondary effects are the challenge that every engineer must deal with. The tail fell off, so obviously something or many things went wrong that weren't anticipated by what we assume was a competent team of engineers. My contention is with the statement that the low numbers of fasteners in the joint was an inherently bad design. This does not refer to unforseen hazards, but to fundamental design principles, and generally dismisses a design practice that has been used successfully on all recent aircraft I know of, not just the A300. I felt obligated to disagree.

Beyond that, time will tell if a weakness in either the design or maintenance of the aircraft led to this incident, but a convincing argument on how saboteurs could sneak onto a plane in the post 9/11 aviation security environment and weaken the tail structure without anyone noticing it escapes me.
posted by cardboard at 3:46 PM on November 23, 2001


(Did Icarus fall from the sky because his wax wings were melted by the sun? Or did the cooler temperature at higher altitudes cause the wax to get brittle and fail?)


Cool.
posted by rushmc at 10:36 PM on November 23, 2001


« Older Al Jazeera english language summary....  |  CellManners.com:... Newer »


This thread has been archived and is closed to new comments