Boeing 777 Crashes At San Francisco Airport

Asiana Flight 214 from Seoul crashed while landing at San Francisco Airport today. Early reports suggest the plane was unstable as it touched down, which led to the tail of the plane breaking off. There are no official casualty reports yet, but passengers were seen walking off the plane. Preliminary estimates say one or two dead and 75 being transported to area hospitals. (Others are reporting two dead and several dozen injured.) Eyewitness report: "You heard a pop and you immediately saw a large, brief fireball that came out from underneath the aircraft," Anthony Castorani said on CNN. "At that moment, you could see that that aircraft was again starting to lift and it began to cartwheel [Ed: he likely means spinning horizontally, like a top]. The wing broke off on the left hand side. You could see the tail immediately fly off of the aircraft. As the aircraft cartwheeled, it then landed down and the other wing had broken." The media has estimated about 290 people were on board the plane. The top of the cabin was aflame at one point, but it's not known yet whether that affected the passengers. "Federal sources told NBC News that there was no indication of terrorism." Some images from the news make it look like the plane may have tried to touch down too early, hitting the seawall just before the runway.

Comments

Wrong week . . .

[cashman73]

Looks like I picked the wrong week to quit smoking!

Re:Wrong week . . .

[binarylarry]

Looks like I picked the wrong day to stop sniffing glue!

Re:Wrong week . . .

[Grog6]

Looks like I picked the wrong week to quit amphetamines

.

Open airplanes

[Anonymous Coward]

This is the problem with non-free airplanes. If the blueprints had been free under a freedom preserving license I'm sure the problem that caused the hiccup had been found.

Re:Open airplanes

[tibit]

Unfortunately, most airplane accidents and incidents are due to pilot error, ATC error and maintenance mechanic error (I think in this order). Problems with hardware or firmware that are unrecoverable in spite of following proper procedures are pretty damn rare. For example, AF447 was not directly caused by any hardware failing - it was due to the pilots not following procedures and good practice.

Re:Open airplanes

[Jodka]

This is the problem with non-free airplanes. If the blueprints had been free under a freedom preserving license I'm sure the problem that caused the hiccup had been found.

. . . and the plane could have been printed on an off the shelf 3D printer . . .

. . .and from the MakerPlane website:
"MakerPlane is an open source aviation organization which will enable people to build and fly their own safe, high quality, reasonable cost plane using advanced personal manufacturing equipment such as...3D printers."

Re:Open airplanes

[evilviper]

Unfortunately, most airplane accidents and incidents are due to pilot error

Repeat after me: "human factors"

Almost any accident can be prevent by a prescient pilot always making PERFECT decisions.

Passing accidents off as pilot error in all but the most egregious cases, is massively disingenuous, and something airlines and manufacturers like to do to shield themselves from all liability that they deserve.

Airlines trained pilots to do something stupid? Pilot error.
Airlines failed to train pilots on the new systems? Pilot error.
Counter-intuitive controls resulted in a pilot throttling down instead of up, and crashing? Pilot error.
Stall warning systems were non-functional, and the pilot wasn't fastidiously checking sensors? Pilot error.
Airline was juggling pilot schedules around, making them work with little sleep? Pilot error.

Re:Open airplanes

[Colonel Korn]

Unfortunately, most airplane accidents and incidents are due to pilot error, ATC error and maintenance mechanic error (I think in this order). Problems with hardware or firmware that are unrecoverable in spite of following proper procedures are pretty damn rare. For example, AF447 was not directly caused by any hardware failing - it was due to the pilots not following procedures and good practice.

If you read the black box transcript from AF447 the most notable single problem was that two of the three qualified pilots in the cockpit believed that the stick was being pushed forward, when in fact it was being pulled back.

When a plane stalls, a common panicked reaction is to pull back on the stick in an attempt to point the plane back into the sky. A lot of training goes into eliminating this instinct, because the solution is actually the opposite - point the noise toward the ground to regain speed. IIRC, the copilot was pulling back on the stick for the last several minutes even as the pilot and the alternate pilot (or a flight engineer or something, I forget) were saying things like, "we're pushing on the stick, why aren't we gaining airspeed?"

The communication problem was largely caused by an major Airbus design flaw: the sticks between the left and right seats aren't linked. In other planes, the pilot would have known the copilot had the stick pulled back because the action would make his own stick move back as well. On AF447, the pilot saw nothing other than the copilot's hand on the stick and assumed he was doing the right thing, and in the understandable confusion as they struggled to gain control of the plane the copilot never verbally corrected the misconception.

Certainly pilot error in response to some external factors that aren't that uncommon (like a pitot tube freezing) was the direct cause of the crash of AF447, but a more sensible flight control design would have likely prevented the pilot error.

Re:Open airplanes

[Kell Bengal]

The always-pilot error belief is a mild form of conspiracy theory. People familiar with aircraft systems can tell you that aircraft are frightingly well-engineered precisely because the aircraft manufacturers, airlines and associated industries know fully well that their livelihoods live and die by the public perception of safety. There is a very strong motivation to getting the engineering right. That's why it literally takes almost a decade to bring a new aircraft to market.

It also turns out that an aircraft in flight is one of the most predictable engineering systems - it's a single self-contained structure where there are only two unknowns: weather and human factors. Outside of those, manufacturers develop good working models of their systems and then test them extensively until they find failure points. Considering the uncertain factors, weather was a major cause of accidents early in aviation before weather radar was a standard safety feature, and aviation forecasts were fine-tuned. Now that weather is largely an avoidable risk factor (excepting extremes such as flying into thunderstorms or freak clear air turbulence cells), it should hardly be surprising that the leading cause of accident is the one thing that is most difficult to design around. If you can find a way to make any complex system (aircraft, car, nuclear reactor, etc) idiot-proof without taking away control from the idiot then I have a prize for you.

But this is a conspiracy theory, so let's see if we can falsify it by testing specific consequences we might expect, if it were true. We would predict that there should not be events admitted to design fault that could not be pilot rectified, because then the manufacturers would be liable - the FAA/EASA/CASA/etc should bury any such case. These regulators would never stop aircraft from flying, since obviously no fault could be admitted to. Rather, a quick search shows a long history of documented design flaws: de Haviland Comet square windows, DC-10 cargo door, 787 battery fires, 747 cargo door electrical fault (AL182, PA103, UAL811, TWA800), 747 aft galley electrical bus being under refridgerator drip pan (QF2), many many engine failures (Delta 1288, QF32, CA786 to name a few). Not all of these flaws were fatal, but they are officially acknowledged as design and manufacturing flaws nonetheless. If the conspiracy were real, they should never have been owned up to. Instead, we see well-documented explanations of what the flaws are, and FAA directives on how they must be fixed. In fact, there have been many cases where the FAA has forbidden whole classes of aircraft from flying until design flaws are rectified - most recently with the 787 battery problem.

When you consider that there are 10s of thousands of aircraft in the air at any one time and the failure rate is so low, it's obviously that engineers have gone to great lengths to make the design-build-maintenance-operation process for modern aircraft very reliable. Obligatory car analogy: ask yourself how often cars crash because the driver messes up, and how often cars crash because of a design flaw. Again, you'll find human error completely dominates.

Like most conspiracy theories, a little research and a bit of common sense goes a long way to shooting down bogus claims.

Photo

[Black Parrot]

http://thinkprogress.org/wp-content/uploads/2013/07/SFO_crash-e1373139561971.png

Shows it upright, with at least one wing still attached.

Re:"Crashes in"?

[sarysa]

Just to elaborate even more (edit function please), anyone who lives around SF and flies knows that the airport is 10 miles south of the city. So an airplane crashing within city limits would be extremely bizarre. Very little air traffic is actually routed through the city...in fact I think most planes that fly overhead are ads for car insurance companies and tires during baseball games.

Survivor Story

[XiaoMing]

Samsun Exec. David Eun survives, posts pic. Then proceeds to teach CNN some manners.

Re:Survivor Story

[XiaoMing]

Then proceeds to teach CNN some manners

They wanted to talk to him about the crash and he said he didn't want to divert attention away from the crash. I'm not even sure what that means.

If you've watched any of what passes for "news" at all today, it's full of talking heads speculating on every possible thing, from the myriad of ways people could have died (ranging from blunt trauma to smoke inhalation because gee fucking golly, the plane carries so many people it must have taken forever to get everyone off and who knows what happens to your lungs in that type of environment!) to just who's fault it could have been that it went down in the first place. And that's all from aerial helicopter footage and an interview with an idiot who used to in some way work with traffic control.

Now can you just imagine what would happen if they got even the slightest tidbit of first-hand information? Oh wait you don't have to, there are half a dozen 5-star "informative" threads on here already discussing just why the plane's wing was or wasn't sheared off while doing some kind of barrel roll a-la Starfox64.

So yeah, when this guy posts as much information as he feels confident doing, including a very uplifting and hopeful picture immediately after the crash showing survivors leaving what looks to be a mostly intact plane, and then doesn't feed the media's desperate attempt to capitalize on the situation any more than they already have been, I am kinda grateful.

No, the wings didn't break off

[Animats]

The tail broke off, not the wings. And the aircraft didn't "cartwheel". There are many good pictures of the wreckage. The wreckage is sitting on the ground alongside the runway, right side up, wings intact, on its belly. The tail assembly is completely detached from the plane. Much fire damage to the top of the fuselage, which is puzzling.

There are pictures of the passengers evacuating, including, inevitably, one of the passengers who just evacuated taking pictures of the plane.

Too early to discuss causes. Reports indicate the plane landed short in an nose-up attitude, but it's too early to say why.

Re:No Cartwheeling

[sribe]

Pictures show the aircraft sat on the ground with the tail missing and the forward roof burnt out but it certainly did NOT cartwheel...

I happened to check news just as this story was breaking. The word "cartwheel" came from the first eyewitness report. The next two eyewitnesses said it "spun". So I'm guessing that the guy who said "cartwheel" doesn't really know what the word means, and that instead it spun on its belly.

Re:No Cartwheeling

[girlintraining]

certainly did NOT cartwheel or bits would be scattered down the runway. It seems that all passengers and crew have been accounted for with no fatalities.

The term "cartwheel" has different meanings to different people. Unfortunately, just like with the Boston Bombing, CNN rushed a story out without getting its facts straight, though at least this time it was somewhat more substantial than pure speculation.

At this time, it appears the plane's air speed was too low on final approach, and the pilot may have over-corrected by throttling up and then (mistakenly) putting the nose further up as a panic measure; This resulted in a severe tail strike on the sea wall, and the plane would have become aerodynamically unstable immediately after.

Typically in these scenarios, the plane (appears) to shoot upwards briefly due to the sudden change in weight distribution, and comes down on angled heavily to one side (having lost any ability to control lateral movement). The wing will typically sheer off, as they're actually designed to break away from the fuselage in such an event, and the plane will roll onto its roof then (if speed is high enough) or the nose will take a digger, break off, and the whole thing will flip in the air and then promptly "face plant" in the dirt in one piece.

Either way, the plane did exactly what it was designed to do -- separate the flammables from the fuselage where the passengers were, and maintain integrity until all motion stops. The emergency crew's prompt response is what saved everyone's lives -- most people don't die due to the impact or fire, but rather smoke inhalation.

This is a text-book crash landing, and the investigation will now focus on whether a mechanical fault caused the plane to lose speed at the last moment (bird strike on engine is common), or whether the pilot neglected to flare correctly. Judging by the debris, it looks like it would have been a steep descent with flare at the end -- which results in a faster landing and is preferred at high-volume airports, over a shallower approach, with less flare. If the pilot is inexperienced, distracted, or any number of a dozen other things go wrong (one plane crash I know of was due to a circuit breaker trip-out which meant the captain did not have 'stick shake' or stall alarm warnings in this exact scenario) -- there's very little time to react, and even going to full power take off speed will not prevent disaster due to the steep descent angle, lack of altitude, and lack of speed.

Any airplane pilot knows the key to a successful crash landing is speed and altitude -- they add precious seconds to react to an emergency. This plane had neither.

Interesting Photos & Data on Twitter

[McGruber]

Some interesting photos and data have shown up on twitter today. First, here is (I think) the original source of that photo taken by a passenger: https://twitter.com/Eunner

Second, here is a photo, taken across a small bay, showing the plane crashing: https://twitter.com/stefanielaine/status/353591123958173696/photo/1

And, most interesting, a comparison of flightpath data (from flightaware.com) of yesterday's flight against today's flight: https://twitter.com/sbaker/status/353611787750494208/photo/1

While I am no expert, it looks like it hit the ground short of the runway, like the previous crash of a 777 (BA 34).

Re:Not geek news...

[hedwards]

The real news here is that this happened today and we're reading about it today. I would have expected to have to wait at least a fortnight for the initial report to show up here. Followed by a week of dupes.

Re:silly eyewitness

[msauve]

He said "cartwheel" when he meant "spin." I suspect he meant "rudder" when he said "tail," and "elevator" when he said "wing."

Casualties doesn't mean fatalities

[Sycraft-fu]

It means serious injuries or deaths. In military speak, which is where it comes from, it means a soldier hurt to the point they can't go back and fight. So someone who's dead: casualty. Someone who has a compound fracture in both legs: casualty. Someone who has a surface cut on their arm: not a casualty.

There's not as hard and fast a civilian definition, but it is just if the injury is serious. It is a useful number for determining how bad something is. Number of injuries period is irrelevant, number of fatalities while relevant doesn't tell the whole store. Number of fatalities and casualties gives a good idea of the human damage that happened in an incident.

Re:No Cartwheeling

[Solandri]

Pictures show the aircraft sat on the ground with the tail missing and the forward roof burnt out but it certainly did NOT cartwheel or bits would be scattered down the runway.

There are actually bits of debris on the runway starting almost with the rocks separating the runway from the bay. The integrity of the fuselage says it did not cartwheel (objects this big don't move in one piece like the movies - they'll disintegrate with just moderate lateral forces). But the debris trail and missing tail suggest it came down at a high angle of attack hitting tail-first possibly from a stall (in a regular landing you hit landing gear-first), then hit the ground hard enough to collapse its landing gear and skid off the runway. The jagged yellow partial dome you see at the tail end of the fuselage is the plane's aft pressure bulkhead - the end of the pressurized section of the fuselage. So nobody was in the tail portion which broke off.

The high AOA suggests the pilot was pulling up trying to gain altitude (or at least decrease the rate he was losing it). Possible reasons are an engine problem (with inadequate thrust, pilot was trying desperately to glide a little further to make the runway) or some failure of the flaps (if they retract, they increase the plane's stall speed possibly causing the plane to drop out of the air). Or wind shear (sudden tailwind deprives the plane of lift and pilot pulls up to try to maintain altitude - unlikely given the weather). Or pilot error (was coming in too high and tried to bleed altitude too quickly, instead of declaring a missed approach and trying again), though the tail striking short of the runway makes this unlikely unless the pilot accidentally put the plane into a stall.

Re:Pilot error?

[tibit]

I'm not aware of any passenger aircraft that has such a fly by wire system.

Well, that's your problem, then, since autoland has been around for a while and I have been through a zero visibility autoland landing all the way down to the runway. On exiting the plane, I've asked the first officer if they did it manually or using autoland, she said autoland. It was a by-the-book landing, by the way, as far as I could tell. Very smooth.I could tell it was a bit of a crab landing since the nose swayed right as soon as the main gear touched down. So, it was autoland with side wind, too.

Why always mention "terrorism?"

[Sir Holo]

TFA: ...there was no indication of terrorism..."

Why is this still included in any US media article about any aviation accident, or similar event, in the news?

As an ordinary citizen, the question of terrorism is not anywhere near the top of my list of questions regarding "how" or "why" an accident may have occurred. Not at all. Now, the question of "who screwed up? Maintenance, pilot, management, etc.?" is the kind of question that springs to mind.

Or, perhaps, maybe the problem is with me? Should I learn to be more afraid?

Re:Pilot error?

[TubeSteak]

It is not used during take off or landing, and although either could be handled by computer, I'm not aware of any passenger aircraft that has such a fly by wire system. All of them are on the drawing board.

Autoland systems were developed in the 40s and perfected in the 60s by the Brits.
Developed for military purposes and then perfected for commercial purposes because England had endless problems with zero visibility due to their fog + pollution.

Autoland systems are so accurate that a fudge factor was added in, since multiple aircraft will all land on the exact same patch of runway and destroy the surface.
I can't say why you're "not aware of any passenger aircraft that has such a fly by wire system."

http://en.wikipedia.org/wiki/Autoland
http://en.wikipedia.org/wiki/Instrument_landing_system#Special_CAT_II_and_CAT_III_operations

Re:news for nerds

[PPH]

He posted a picture of the crash:

We warned you: Turn off the damned phone!

Re:"Crashes in"?

[osu-neko]

For anyone confused by this comment, the original title of this article (before an editor stepped in and fixed it) read "Boeing 777 Crashes In San Francisco". The current title (at the time I'm posting this comment), "Boeing 777 Crashes At San Francisco Airport", is a much better description of the event without taking the mind in some horrific directions before giving the important details.

Re:Not geek news...

[hawguy]

This doesn't really sound like geek news.

One rumor I heard was that ILS (or some portion of it) wasn't functioning on the runway the plane was landing on (28L) so the pilot was making a manual approach without the automated glidepath alerts he'd normally have. If this is true, then this gives the story a technology/geek tie-in, and touches on issues like whether or not humans (pilots in particular) have become too reliant on machines and when the machines fail, humans don't have enough experience without them to be an adequate back up.

Though I haven't seen the ILS issue reported in any official reports, so maybe it's not true.

Re:Pilot error?

[AJWM]

I am a pilot.

You're wrong.

Re:Pilot error?

[Anaerin]

The pilot HAD to manual land, the ILS system and PAPI glidepath height assistance for runway 28L (and 28R) at SFO is down, as reported in the current NOTAMs(Check for SFO)

!SFO 06/005 (KSFO A1056/13) SFO NAV ILS RWY 28L GP OTS WEF 1306011400-1308222359
!SFO 07/046 (KSFO A1326/13) SFO RWY 28L PAPI OTS WEF 1307062219

That means he was relying on nav beacons and glidepath estimates to come in. Given that SFO's beacons are approx 1.2 miles apart, if he picked the wrong beacon to guild his descent he would have been too high, dropped steeper than usual to get down once he noticed the discrepancy, and didn't have the necessary power to flare and ascend at the end of the runway, so he tailstruck. That makes it pilot error, but confounded and mitigated by most (if not all) the regular guidance and assistance systems they rely on being out of commission.

Re:Pilot error?

[dgatwood]

Citation: takeoffaviationweather.com. The relevant bits:

KSFO

28 NAVAID Instrument Landing System Runway 28L Glide Path out of service started about 1 month ago ending in about 1 month
28 NAVAID Localizer Type Directional Aid Runway 28R Glide Path out of service started about 1 month ago ending in about 1 month
28 NAVAID Instrument Landing System Runway 28R Glide Path out of service started about 1 month ago ending in about 1 month
23 NAVAID Instrument Landing System Runway 28R Inner Marker out of service until Aug 22 23:59
20 NAVAID Instrument Landing System Runway 28R Category 2/3 Not Authorized started about 1 month ago ending in about 1 month

Emphasis mine.