Air France 447: The ‘Coffin Corner’ and a ‘Mesoscale’ Maw | Hacker News Reader

Air France 447: The ‘Coffin Corner’ and a ‘Mesoscale’ Maw

trueslant.com

71 pointsold-gregg53 comments

Miles is a pilot and aviation enthusiast, and I've admired his work in the past.

I hope this bit of writing is more public-serving than self-serving. After all, we still know very, very little about what happened, aside from the data transmissions. Even speculating on structural disintegration is problematic because the obvious next question was: why wasn't the plane configured for heavy turbulence?

So you're left with this article that goes over some cool new facts (for those of you outside the aviation community) but really doesn't shed much light on the incident at all. It's not a news piece, it's not an opinion piece. It's kind of a science piece hanging off a news story -- one that puts the author in the role of expert without facts and one in which hundreds of people get killed.

Not my cup of tea.

The article suggests that a stall is a horrible outcome.

Old style biplanes used to do "stall landings" all the time. If you know the characteristics of the wing, then a stall can be anticipated and is not necessarily cause for alarm.

How is it that the autopilot system (or other systems on the plane) would not be programmed to handle a stall gracefully?

In a stall the wing loses lift, but as the plane falls the air thickens and the angle of attack decreases and the wing ought to regain lift sufficient for the pilot to be able to control the plane.

I'm not suggesting that handling a huge jet in a stall would be easy, but I would guess that knowing how to do it is part of the requirements for flying one.

This was my question too. But I think stalls in large airliners are a very different thing than in GA aircraft. I'm pretty sure full stalls are still part of the test regimen, but I don't think normal pilots ever do a full stall. It may also be other factors playing in at high altitude, and of course in severe turbulence a recovery might be even more difficult.

Glider pilot here. I have decent knowledge of aerodynamics, but have never studied large jets in detail. This is my personal theory - essentially what lies behind the words of a few articles I have read over the last few days, but the authors haven't been able to formulate it clearly.

Assuming that the autopilot and fly-by-wire systems were somehow compromized, a full stall in the "Coffin Corner" (Q-corner) would be a very, very serious condition. A stall causes the nose to drop and speed to accumulate. Since the difference between the stalling speed and the never-exceed speed is so small at the altitude in question, this accumulation of speed could easily cause the plane to exceed its critical Mach number, causing supersonic shockwaves around certain parts of the airfoil. In planes not designed for supersonic flight, these shockwaves will eventually cripple the effect of the control surfaces. If the plane is in a nose-down position when it looses effect on its elevator rudder, the plane will overspeed and eventually suffer structural failure (essentially meaning the loss of some critical part) due to flutter (resonance in parts of the airframe) or aerodynamic stresses.

A stall during normal flight conditions doesn't have to be a serious problem, but these weren't normal flight conditions. And the information that we have suggests that certain Airbus models have had problems with their pitot (speed measurement) systems in icing conditions, leading to erroneous speed readouts. The fly-by-wire system, which prevents the pilot from flying the aircraft into stall or overspeed, cannot function correctly without a correct speed measurement. And as I have detailed, a proper stall might be enough to doom the airplane.

The weather was never considered "horrible" and still isn't. [edit:] At least 12 planes took the same route that night without any incident.

http://ibnlive.in.com/news/12-planes-shared-sky-with-af-447-...

From the very last paragraph of the same article (which originates from CNN, incidentally):

> Although none of the other flights are known to have reported weather problems en route, aviation experts said weather can change suddenly and vary over short distances, so one plane might experience conditions far worse than another.

Insightful but it seems that there's a lot of speculation about this. Theres an article from timesonline that tied some pieces together and answered some questions for me (like: "How do they know this much detail of the flight without the black box?" A: ACARS).

I hope this is helpful to someone else: http://www.timesonline.co.uk/tol/news/world/us_and_americas/...

Thanks. I assume that the forum mentioned in the article is http://www.pprune.org.

The current thread about this accident is http://www.pprune.org/rumours-news/376433-af447.html

Apart from the current tragedy, it's a great window into another professional sphere.

Q: ACARS? A: automatic communications and reporting system

Why when internet is becoming available in planes is something as simple and and essential as updated weather reports not possible during a flight?

Why are these practices seemingly stuck in a stone age of paper weather reports, printed before the flight leaves, that could have been outdated even before the planes takes off.

Maybe this accident will speed up a much needed needed technological update in this aspect of air-travel. We should stop wasting dollars on security theater measures and invest it in real safety measures that can save lives.

They get weather updates en-route sent to them by their home airline (via the radio system) - not to mention that jets of that size carry along with them a better weather radar system than most local news channels have.

I'm talking about satellite imagery like it is shown in the article, which is the only kind that would provide reliable and usable info for storm evasion. You need a view from above to see a storm size and not just a surface look at it from the sides (with a radar). Weather radar only provides limited information, like the article says.

"interpreting its display is a bit of a black art." - "It is akin to a blind man with a cane;" - "For instance, the radar mostly detects rain and hail – and if that first layer of storm cells was particularly heavy, it might have acted like a curtain – hiding the reinforcements from radar beams."

You can't know what's inside a walled garden or know how big it is just by looking at one side of the outside walls, you need something that allows a top view.

More raw data to the cockpit is unlikely to help. Pilots have a lot of things going on and varying levels of weather knowledge and experience (even professional pilots - weather is but one thing you train on).

A professional meteorologist on the ground who can interpret and summarize the relevant information is of way more use.

I agree that it's also useful but this is not black and white. It's not one ore the other, it should be both. Because when your flying into a storm front in the middle of the atlantic you are on your own and it's not that interpreting a satellite picture is a black art like radar. Besides that's why there are 2 of them in the cockpit. If a pilot can't handle that he shouldn't be flying a plane. It's like saying we shouldn't invest in new CT machines or other better medical imagery, doctors have more than enough data to handle.

Er .. why should the pilot interpret it, can't we have a specialist inside the cockpit to interpret the meteorological data ?!

because you can't fit in the extra guy in all cockpits :-) Would be unpractical and expensive to put a 3rd person there just for weather. Good updated meteo from the specialists on the ground, and basic training in and availability of satellite images in the plane.

Like it's said you might loose contact with any ground station when flying in the middle of nowhere or you might lose imaging data of the storm. So both should be there to provide a better standard than the current limited info of printed weather maps in preflight briefings, especially for long flights.

So while you are napping, eating or watching a movie on that flight to LAX, you should know the plane you are flying is cruising along at the ratty edge of its capabilities. Why? Money. The higher an airliner flies, the better gas mileage it gets.

As libertarian I prefer an airline which will charge me premium for flying more expensive but safer routes.

However, as a realist I realize that the barrier to entry for airlines is very high and competition is sub-optimal.

So I'd settle for a government mandate of how high commercial airlines can fly.

That's a bit of a jump. How'd you go from "As a libertarian..." to "So I'd settle for a government mandate" in 3 lines?

He's that smart.

However, as a realist...

I would assume HN would favor people who looked beyond knee jerk responces. Just because you like to use the free market when it's possible does not mean you need to apply it in all cases.

Well simple.

I really do want the market to work, but I know in reality it very often does not work at all. Most often with big industries where the startup costs are high.

Do you prefer to continue to fly in a needle's eye rather then have the government step in?

Ideally at least one airline would offer the safer option as a way to gain customers, but what if no one does?

Is your libertarianism that pure? If so, that's admirable, I guess my ideals are just not quite as strong.

Seems like a rather suicidal market strategy. So you invite people to pay 2x to fly with FlySafe (perhaps an existing airline with n outstanding safety record that rebrands itself), milk the resulting controversy to sell tickets for a while, only to later have every untoward incident, no matter how trivial, spun into a damning indictment of your business philosophy? No thanks.

Furthermore, the first time such a company did have a fatal or obviously dangerous incident, they'd go insolvent immediately. I see the point you're trying to make but it's simply impractical in a mass market like commodity air travel. Uber-expensive executive flying, perhaps.

it's simply impractical in a mass market like commodity air travel. Uber-expensive executive flying, perhaps.

If it wasn't for the coverage of this incident, I would never have known of the "coffin corner." How would market forces operate on factors of which the public is not even aware?

And until that pilot landed in the Hudson, I'm betting you never thought too much about birds getting sucked into jet engines (I certainly didn't). Perhaps the conclusion from this is that there isn't really much utility in your being aware of it, since you're not a pilot. That is to say, the knowledge is of little benefit to you as a consumer because the weight of this particular factor in the overall safety of air travel is not that big.

Like say you design software for something critical like medical devices, and one of these fails and somebody famous dies or a bunch of people die in different places. Suddenly everyone is talking about unit testing and software verification as it pertains to your field, someone with a talent marketing starts up safercoding.com and so on. I'm not sure the stock of human knowledge has really gone up.

I have to disagree. For one, I think birds flying into plane engines was common knowledge. It's certainly something I've know for as long as I can remember.

And when I got my CS degree we specifically studied accidents where bad software in medical devices killed people. I recall an X-ray machine where the hardware was upgraded and a bunch of software assumptions about when the fuses would blow became wrong. That ended up killing people.

And as someone who has written software for medical instruments I can tell you unit test coverage is part of the contracts we sign.

The coffin corner only played a small part in the accident but that's how accidents always happen. It's always several small things that go wrong together.

A safe airline would not flaunt its safety record, but it's safety practice. Not flying the coffin corner would remove that one part that can go wrong from any future accident scenario.

The safe airline having an accident would not automatically bankrupt it, every other airline has had accident and this one would continue with its safer then average practices.

However, I agree with you this will never happen because as you say the airlines are a crap commodity business. Even Warren Buffet has mentioned how bad a business airlines are for investors.

That's why even though I consider myself a libertarian I want a government mandate to stay out of the coffin corner.

We already have much more complex and difficult to enforce regulations on airlines, this one is simple and would be easy to enforce as we already track all planes on rader closely all of the time.

>> As libertarian I prefer an airline which will charge me premium for flying more expensive but safer routes.

How would you even know? Saying "I'd pay more for a safer routing" is rather reminiscent of Steve Yegge's Shit's Easy Syndrome - you say you want a safer routing, but what does that mean?

You'd pay an extra ten percent on your ticket to fly ten percent lower? What about other factors - would you demand more reinforcements on the wings of the planes you fly on? How about longer flight times due to having to stop for fuel? What about how this would impact global route scheduling?

I could keep going with that, but the bottom line is that you don't understand what makes a plane safe. I don't understand what makes a plane safe. Few people - unless they're Airbus or Boeing design engineers - on this site know what makes a plane safer, and how much that would cost.

Airlines have a huge interest in making their flights safe - think off all the shitty PR that Air France (one of the world's most respected airlines) and Airbus (one of two real manufacturers of jumbo jets) are getting, and how much it will directly and indirectly cost them. The last people died on an American airline was in February. Before that, it was in August 2006. The last time an American jumbo jet crashed and killed people was in 2001. Flying is an incredibly safe way to get from point A to point B - but like everything, there are some times when nature throws us a curveball, and we get hit by it.

>> So I'd settle for a government mandate of how high commercial airlines can fly.

I wouldn't trust Congress to determine how high an airline can fly. If I want to know how high an Airbus can safely fly, I'd ask Airbus - they know the planes better than anyone, and have a huge vested in interest in keeping them safe.

Flying lower = flying slower. You can already do that, just pick a plane that flies lower, like a commuter jet...

But in general, when flying it's safer to be higher and faster. By far the most dangerous thing you might encounter when flying is the ground...

Instead of flying a lot lower and slower, how about marginally lower and slower? You could give the plane 50 knots leeway in speed and still fly over most of the weather out there.

For conditions like the ones that the Air France flight was going through, this would seem prudent. You can't fly above thunderheads like that. Might as well get yourself into a safer part of the flight envelope in case something bad happens.

There is a maximum safe speed in turbulence (it's called "design maneuvering speed") and I really doubt they would exceed that in these conditions. (You want the wing to stall in case of severe turbulence, because the alternative is to overstress it. That's why you don't want to fly fast.)

Yes, but that depends on the amount of turbulence, yes? Perhaps the margin of safety needs to be bigger for that region of the ocean?

Weakly so. The idea is that you want the wing to stall and lose lift before it generates that design load limit you don't want to overstress. You generate lift by increasing angle of attack, but the idea is that the wing should stall, and thus stop generating lift, before. The amount of lift a wing can generate depends on its airspeed, so if you keep the airspeed down, the only situation in which it could generate a dangerous amount of lift would be if you entered a region of very high horizontal windshear such that the airspeed all of a sudden increases to above maneuvering speed. (So in principle you should be at least the gust speed below maneuvering speed to be safe, so here is where the margin of safety enters. But it's a second order effect.)

you don't understand what makes a plane safe

Agreed. Now if the captain Sully (the guy who landed in the Hudson) says flying out of the coffin corer is safer would you believe him?

Airlines have a huge interest in making their flights safe

Yes, they also have a huge interest in being profitable, they are trying to optimize both.

There's nothing wrong with that, but as a consumer I'd like a bit more choice.

But as other in this thread have pointed out the particular nature of the airline business makes it impossible for that kind of market differentiation to happen. That's known as a market failure.

Hence my desire for a government mandate.

Alternatively, do you think we need any of the current government regulations on airlines?

> However, as a realist I realize that the barrier to entry for airlines is very high and competition is sub-optimal.

Is the barrier actually high? (I see new airlines every so often.)

What are the actual barriers? If they're govt, why isn't reducing the barriers the right response?

> So I'd settle for a government mandate of how high commercial airlines can fly.

And govt would get it right because that's what they do.

Didn't we just have a nice discussion of regulatory capture? (My favorite comment was by someone who thought that regulatory capture was the act of promotiing govt power grabs.)

Is the barrier actually high?

Besides the obvious cost of acquiring multi-million dollar jetliners (which, if bought new, typically have to be ordered 5-10 years in advance of delivery) and signing leases for gate space and hiring ground crew at airports and so on?

You also need FAA and TSA approved operations plans, adding for instance the cost of screening all your employees and equipping your planes with things like black boxes and weather radar.

Then you end up like Virgin America, who can't fly to Chicago because the incumbent carriers there claim there are no available gates at O'Hare (despite the number of flights in and out of the airport being down like 10-20% from previous peak).

Doesn't seem that high to me...

You don't need to buy new jetliners to run an airline. You can go to ILFC and rent them. Even big guys like American and Continental do this for some of their aircraft.

One webpage I found suggests the lease price for a 737 is between $50,000 - $100,000 per month (which seems reasonable, or maybe a little low). Wikipedia says YouTube's bandwidth costs are estimated at $1M per day, so the hardware cost of a couple aircraft doesn't seem all that expensive.

(Disclaimer: I know almost nothing about commercial aviation.)

While there is a lot of good information in this article, the tone is a bit too tabloid-sensationalist for my tastes.

"So while you are napping, eating or watching a movie on that flight to LAX, you should know the plane you are flying is cruising along at the ratty edge of its capabilities. Why? Money. The higher an airliner flies, the better gas mileage it gets."

Boo those big corporations for trying to pick optimal flight paths, save customers flight time, or otherwise make a buck (and who cares about all the environmentalists breathing down their necks about burning so much kerosene). How DARE they? Think of the children!!

The one thing common after all these crashes is the amount of looney speculation that goes on right after, and then we magically forget about it. Does anybody even remember the British Airways 777 from Beijing that landed short at Heathrow a couple of years back? The reason: entirely mundane - faulty fuel pump.

Reports seem to indicate that they almost made it through the stormcell.

Very unfortunate. Whatever it was probably happened extremely fast. Perhaps a (ridiculously unlikely) wing snapped off? Or they stalled and spinned violently?

Boeing tests its wings to breaking by bending them behind the back of the plane until they are pointed up at ridiculous angles. The kind of force that would break the wings of a modern jetliner would also break the necks of a large number of the passengers.

http://www.youtube.com/watch?v=pe9PVaFGl3o

EDIT: I stand corrected. As pointed out below, that huge deflection doesn't represent neck-breaking forces. But imagining the force of an updraft that can accelerate an airliner upwards at 2G's -- that's still a heck of a lot of power.

Cool video. I wonder how much of an engineering/structural difference there is between the heavier 777 and the Airbus A330-200 that crashed.

Both are tested to 150% of their maximum rated load.

Not at all. The A340 limit load factor is something like +2.5 / -1.0g. This is not a condition injurious to seated passengers. It is unrestrained passengers who are injured in turbulence -- but that is due to the passenger/cabin collision, not the initial acceleration. The wings in the video flex a long way under these relatively light load factors because the wings are designed to be only a little bit stronger than they need to be.

Aerobatic airplanes and pilots routinely experience transients to +/-6g with no ill effects.

A very insightful examination. A lot of things I didn't know.

True. Too bad about the tabloid-style typography and literary style, which does serve a purpose of highlighting the takeaways but also gives it an air of sensationalism.

I agree, but he's pretty glib about it all (quoting star-trek? really?). Strikes me as a pretty odd tone to take.

I know its all very serious and its sad commentary on the mental baggage I now carry from watching too much TV when I was 15, but I couldn't help but do a double take when a guy named "Miles O'Brien", quoted Scottie describing engines.

I will always say it was a bomb, and the more they try to make it look like it wasn't, the more I believe they are just trying to hide it from the public not to scare them to fly like in 2001.

It was a bomb.

For a hundred years we've been flying planes thru storms, lightnings, adverse conditions of every kind, so fuck the coffin corner.

The pilot and copilots had thousands of flight hours under their belt.

Not a single mayday call.

Radio communication fail? Weird electrical/magnetic phenomena? Meteor? These are just some of the more plausible ideas than "covert terrorism".

Why has no one stepped forward to take credit for downing the plane with a bomb? Isn't that the MO for terrorist organizations?

Accidents happen, unfortunately. Sometimes planes crash. Our technology isn't perfect, and all we can hope for now are useful lessons to be learned from the crash itself.

If I was a terrorist org, I wouldn't take credit. The new modus operandi is stealth mode, the days of TV shows about terror are over.

Terror to create havoc on the world economy and air traveling will be the first hit.

Expect more "accidents" in the future.

if a terrorist group does not claim it, the world would not know it was a bomb (per your assumption) and that fear would not translate to the population....therefore your theory dies

Terrorist organizations use terror to promote a cause. If you're just making people generally a little nervous, you're not doing a good job of advancing an agenda.

Both pilots of an Air Comet flight from Lima, Peru, to Lisbon, Portugal, sent a written report on the bright flash they said they saw to Air France, Airbus and the Spanish civil aviation authority, the airline told CNN.

"Suddenly, we saw in the distance a strong and intense flash of white light, which followed a descending and vertical trajectory and which broke up in six seconds," the captain wrote.

http://edition.cnn.com/2009/WORLD/americas/06/04/plane.crash

I think you were voted down unfairly. We don't know what happened, but whatever it was, it was instantaneous and caused a bright flash.

Reply on news.ycombinator.com