This article in WAPO and presumeably others elsewhere relates that the passengers in Air France 447 didn’t know they were flying into the sea since there was not some kind of announcement from the cockpit. Did the airspeed have something to do with this? Everytime I’ve landed in a plane I believe I have had a sensation of descent. Were the authorities trying to comfort/shield the passengers’ relatives or is it possible they didn’t know anything was wrong?
I suspect that the sensation of falling you feel when landing is partially due to the airplane’s nose lowering as it intentionally descends. Air France 447 never lowered its nose, it was in a nose-high stall for the entire descent.
Here are some interesting pilot training exercises that show how easily our bodily sensations are fooled in flight.
On the CBS Evening News tonight, the NTSB chief described what it was like as “horrific.” So maybe their senses weren’t fooled.
I am not buying it speaking as a perpetual flight student and long-time aviation buff. It is true that are airliner can do a perfect one g roll all the way around and neither the plane nor the passengers will know unless they are looking out the window. The cabin attendants could serve drinks during the whole thing and never notice if it is done perfectly. Here is a 707 roll maneuver done by a test pilot that is perfectly safe.
It is also true that human senses are terrible at detecting an airplanes attitude along any axis without visual references. You can think you are flying straight and level or flying upwards when you are about to pull a JFK Jr straight into the ocean.
That isn’t what we are talking though with this disaster. The Air France flight lost control because of air control surface aerodynamic stalling. That results in wild attitude swings along multiple axises. People may not be able to tell exactly which way they are headed at any given moment but the body is basically a large bag of fluid and feels those changes in attitude sometimes violently as they shift. That means that no one peacefully read their book until they were completely surprised by a smack to the ocean. It makes a nice story for the families but it isn’t true. The only way that would have been true would have been if the plane a straight downward path towards the ocean in an attitude that had a steady 1 g rate all the way down and that doesn’t happen with general loss of attitude control.
I can simulate this disaster in Flight Simulator X just like many amateurs can complete with calculated G forces. I may do that later tonight. The black boxes give the same data and it probably isn’t that pretty during peak forces.
Airline pilot …
What **Shag **said. When the accident first happened I got lambasted in the long running thread here What would a plane crash like the Air France one be like from inside the plane? - Factual Questions - Straight Dope Message Board by arm chair experts for asserting it was a wild & nasty ride for the passengers while the French authorities were saying everybody died happy.
I have *not * read any reports on what the data recorders now tell us. Having said that …
There is almost no way in hell this was an easy ride where everybody was clueless until impact with the sea. If nothing else the large rate of descent would have produced rapid pressure changes in folks’ ears. They would have known they were descending rapidly and known based on the time since departure that they were out over the ocean.
Swept-wing airplanes do not stall smoothly. At high AOA there would have been a feeling like aggressive pounding turbulence all the way down. Maybe the magic computerized flight controls would have kept the wings relatively level. In a less computerized aircraft a sustained stall like that is usually characterized by rapid wing rock from 30-45 degress wing down on one side to the other & back again. Over & over & over.
I’ll be curious to see what the recorder readout says. My gut now is DGAC (French air safety agency) is saying whatever they think will make the families of the people that were killed feel good.
Nitpick: That’s not true. Someone, somewhere, sometime, has said this and it has become gospel among pilots who don’t quite know enough to know it’s wrong. Any manoeuvre involving pitch changes must have associated g changes. A barrel roll can be positive g all the way around, but it can’t be 1 g, it must be greater than 1 at some points.
I honestly appreciate you correcting me on my aerodynamic mistakes a few times. I know what you say is literally but I try to learn. In this case though, the main point is that planes can go through major maneuvers as it appears from the ground without the passengers feeling much. However, that is a point the media is lilkley to play up in this story and it simply isn’t true for the Air France flight. Violent changes in attitude can be felt be anyone even if they can’t tell which direction they are coming from.
I read the preliminary report from the French equivalent of the NTSB. It wasn’t pleasant or peaceful, m’kay? Based on what was in that report, I’m pretty sure I myself would be screaming or crapping my pants or something, it reads as a horrific ride.
Here’s the link
The black boxes recorded a descent speed of nearly 11,000 feet per minute, and rolls of 40 degrees.
Um… no, I wouldn’t call it that. They weren’t lurid, it’s all in a dry “just the facts” format, but they didn’t seem to sugar-coat it.
One of the local news stations just did a report saying that most people wouldn’t have noticed, because it was nighttime and they would have been asleep. :rolleyes:
Yeah your main point is definitely valid, the 1 g roll thing is just a pet peeve of mine.
So why couldn’t the pilots recover? From what I have read it seems the pitot tubes might have been blocked and giving faulty air speed readings.
That said I thought pilots are trained extensively on the characteristics of a stall (plane rolling, etc.) and know full well where the envelope for a stall is on any given plane they are flying and certainly know the proper procedures to recover from a stall. Presumably other instruments were working showing the pitch of the plane and altitude and such. (Is the cockpit all glass that could theoretically fail or are some things like the attitude indicator still mechanical and not something that can break?)
Even with faulty readings I would think it’d be abundantly clear to any commercial pilot they were in a nose-up stall (sounds like it from what was said above). Unless the plane simply cannot (or is unlikely) to be recoverable in that situation (which would surprise me) you’d think they’d have figured it out.
IANAPilot
I will leave it up to the airline pilots to respond definitively but they should have been able to get out it unless there is more to the story. There are redundant instruments for most things that should be enough to keep a competent pilot in the air. Pilots are trained to deal with multiple instrument failures. Air speed indicators are a primary instrument and it is bad news if they fail but that wouldn’t effect the attitude indicator which shows where the nose is pointed in relation to the ground. It also wouldn’t affect a GPS reading for ground speed. Airplanes fly based on air speed and not ground speed but an experienced pilot can use ground speed and some rough math to keep the plane from stalling. Pitot tubes also have heaters that can clear most blocked ice rather quickly once the pilot realizes that they have gotten blocked by ice but you have to turn the heaters on manually.
Then again, pilots of all experience levels sometimes fly straight into the ground for little reason other than confusion. Look at the crash of Steve Fossett, the aviation record setter, who crashed a small plane on a routine flight or the pilots of a Lear Jet that disappeared for years in the New Hampshire woods for little apparent reason. It happens sometimes but thankfully it is extremely rare on commercial flights.
On commercial flights you have at least two pilots so presumably a lot less likely for both to become confused in the same way at the same time. Shit happens anyway I know but this is probably why it is less likely compared to, say, JFK Jr.
This is correct.
Even on airplanes that are entirely glass cockpit there are back-up instruments that will continue to work even if the airplane loses all power and “glass”.
That doesn’t mean those back up instruments are equally easy to use - they are not centrally located, and it’s a different sort of display than the pilots are accustomed to using. They certainly work, and I can name off the top of my head to instances in which an Airbus that had lost all power except emergency was successfully landed with no one hurt. An important difference, however, is that neither of those two airplanes were slogging through a massive thunderstorm at the time. Weather was undeniably a factor here.
Remember, big thunderstorm - the airplane would have been shoved around by the weather and they were almost certainly in turbulence. Between that, and it being night and overcast so they couldn’t see out of the cockpit it actually is entirely plausible they were nose up and didn’t realize it, at least at first. 10 degrees pitch up, for example, might not be terribly obvious under such circumstances. Again, you have to factor in the weather.
That is, in fact, one of the reasons airplanes have stall warnings - because while “nose-up and slow” is often obvious it’s not always obvious, especially when there are other distractions like turbulence, lightning, and, oh gosh, we’re getting multiple airspeed readings, WTF?
I wish to refer you to what one of our resident Big Airplane guys, LSLGuy, said upthread:
I’m told you can recover from stalls in a big airliner, but it takes time and altitude to do it. And they ran out of both quickly. If you get a severe stall in a situation where you’re already distracted, have poor or diminished information, and for whatever reason don’t get on top of it immediately this is the result. Being in a thunderstorm with the associated turbulence will only complicate a stall recovery.
Yeah, but rough math while tossing around in weather turbulence, alarms going off all over, and having to wrestle with an airplane that suddenly went off autopilot? Just a few distractions there, on top of the fact you’d have to do such math quickly and accurately. Possible, yes, but not a given the pilot can pull that off adequately. Math skills often deteriorate under such conditions.
And even if you do, there are weather conditions where the heaters can be overwhelmed. Even before the black boxes were found it was speculated that the very thing happened in this accident. There have been other instances of that model of pitot icing over despite having the heaters on.
That model of pitot is being replaced by a model less prone to ice-over. All fleets have accelerated the changeover after this accident. Thus, it is less likely to happen again in the future but it’s not absolutely guaranteed.
Well … I read the BEA (French equivalent to US NTSB) interim report linked to by **Broomstick **above.
This is a classic “deep stall” scenario. Due to some mechanical failure, they lost all usable airspeed info. It was around dusk and they were in the clouds at high altitude.
The computers which are really flying the airplane also lost AOA inputs. That meant the aircraft got harder to fly & may have done some stupid stuff caused by stupid non-fail safe automation.
The problem took almost exactly one minute to develop from *everything is 100% normal *to they’re screwed. Then it took about 3.5 minutes to fall to the sea while they tried ineffectually to remedy the situation. During which the ride was about as I described in my earlier post.
During that one minute the crew had a bunch of conflicting instrument readings, warning signals, etc. For whatever reason, the pilot on the controls did exactly the wrong thing at first. Or more accurately, he let a second priority consideration (maintaining altitude) override a first priority consideration (maintaining control without usable airspeed indications).
For some reason the pitch trim then went full nose up. I have never flown any Airbus product, nor do I have their manuals refer to, but from what I’ve read about them elsewhere, this was likely to have been the computer trying to help the pilots. Instead it killed them.
A deep stall is well-explained here Stall (fluid dynamics) - Wikipedia . The wiki article talks about it being mostly an issue with T-tailed designs, which the Airbus is not. However there have been a couple of deep-stall accidents involving both Boeing & Airbus non-T-tailed designs. So it can happen to any type.
The critical thing from a laymans’ POV about deeep stalls is the nose is pointed up pretty steeply (ie. like just-after-takeoff steep), the airplane is descending like a brick, and there is damn near nothing the crew can do to recover once the situation gets this far. All the flight controls are mostly useless. Each of us has our creative ideas about what we’d try in extremis. Whether anything will work is unknown. Tellingly, the manufacturers don’t offer any advice on what to do next.
[Soap box on (just a bit)]IMO, the real *coup de grace *here was the pitch trim. Once that had run full nose up there was nothing they could do to overpower it. And my bet is that they never *noticed *the pitch trim had run full nose up. Maybe there is/was a beeper to tell them this was happening; maybe not. When stuff gets exciting, there are a lot of noisemakers in the cockpit which can cause more confusion than anything else. And in the “smarter” aircraft they’re prioritized by the computer, muting the less important. That starts from a good impulse; reduce conflicting & confusing inputs to the crew. But in this case it *may *have silenced the one warning they really needed.
[/soap box]
A few years ago in a 727 I lost one of the three airspeed systems. At night in the weather. It got very busy & noisemaker-y & confusing there for a minute or so. Our situation worked out just fine, but we didn’t have even 10% of the problem these guys had. So my point here is not “I’m cool; they’re goofs”. Rather it’s “From direct personal experience I can say they were confronted with a real rattlesnake. And they failed to grab it successfully on the first attempt in the first 10-15 seconds. So it bit & killed them.”
I had no idea Jimmy Stewart was a test pilot.
Without trying to get into a tit-for-tat …
Folks are right that big jets have redundant airspeed and AOA systems. Typically 3 independent systems. We know the primary system for the pilot who was flying and the backup system both failed simultaneously. We don’t directly know what happened to the system feeding the other pilot’s instruments, but they were recorded as saying it was out too.
Folks are right that pilots *should *be able to hand fly without any airspeed indication. That’s much harder to do in an Airbus than a Boeing due to the design of the flight controls. And they were at the max possible altitiude for their weight. At that point, making even a small error by climbing even a smidgen would rapidly deplete what little excess energy they had.
It’s also a self-reinforcng problem. The slower you get, the more quickly the airspeed bleeds away. If you lose even 10%, which can be gone in seconds, you’re committed to a pretty aggressive dive to just maintain speed, much less get back what you’ve lost. This is very, very different from how a light plane behaves.
Pitot heat is on at all times in big jets. As **Broomstick **said above, due to other incidents in the Airbus fleet it appears the design standards for the pitot heaters wasn’t good enough. The heaters simply don’t put out enough heat to prevent icing in conditions which can readily occur at altitude.
That’s also what this accident’s present understanding of the circumstantial evidence points to. In any case, new & improved heaters are being retrofitted as fast as the manufacturer can produce them.
Sorry to get on a roll here …
Anticipating a question alluded to by somebody above …
Pitch trim in a big jet is very different from a small airplane. The trim system pivots the horizontal stabilizer up and down while the normal flight controls move the elevator. The stab provides about 5-10x the control power the elevator does. For somebody with lightplane experience, it’s almost as if the roles of elevator & pitch trim have been reversed.
If the pitch trim is set inappropraitely the airplane is not controllable, period. The elevator is nowhere near powerful enough to offset a mis-set pitch trim. The airplane *is *going to go where the stab wants it to go.
Pitch trim runaway is one of the very few “fix it in 5-10 seconds or you’re fighting for your life.” scenarios.
Why would pilots these days need “creative ideas” about this?
I would think this is something that could be covered in a simulator. Pilots can find out for themselves, in safety, what works. I am surprised it is not covered at all in the simulations or manual. Is it something considered to be so unlikely it has not been worth exploring ways to recover and train pilots on it? Or are the simulators, good as they are, just not capable of re-creating faithfully what happens in this situation?
Note I am not suggesting it is something “easy” to recover from. Just that pilots would have had training on this (or one would think so anyway).
Also, as to magic flight controls maintaining some control and thus denying clues to the flight crew I think it was said above the plane was rolling 40 degrees side-to-side. It was suggested this is exactly what happens when a big jet stalls. I’d think a glance at the AoA indicator combined with that roll would be sufficient to clue in the pilot what was happening. What else would make the plane roll violently like that?
Again, IANAPilot and I am not second guessing here but asking. I have no doubt things in the cockpit got hairy real fast and it was a complicated problem for the pilots under highly stressful conditions. Pilots are well trained and good at what they do but they remain human.