AF447 wreckage found

[The Ancient Geek]

the stall was effectively unrecoverable

It would have been easily recoverable if the stall horn had worked properly.

[J-Class]

If you think about what could have been done to prevent this accident, rather than about whether the crew were competent/incompetent or whatever, it seems as though if they had had available the information that we have from the FDR, ie alpha, airspeed and descent angle, they would have recognized the problem immediately and recovered. But that info was not provided evidently because the system designers considered this scenario impossible. But three identical pitots known to be able to be overwhelmed by ice are not triply redundant; they are a single point of failure. The airplane -- any airplane whose systems rely so heavily on airspeed -- needs a truly alternate airspeed source, a full-time angle of attack indicator, and a stall warning that does not go to sleep and then wake up at inopportune moments.

Garrison, I'm with you. Indeed, this view hasn't changed much since I made a similar point on p.24 of this thread (before the BEA report was published) where I said:

"Surely AoA (and perhaps a GPS speed indication for good measure) would be very helpful in instances where airspeed indicators have gone doolally and the aircraft computers are assaulting with pilots with a bunch of not necessarily consistent error messages and alarms?

All modern aircraft rely on system redundancy, but can any system be deemed truly redundant if it relies on the same components on each of its legs? (I'm imagining that more than one pitot tube iced up). Given the problems of producing 'true' redundancy in a single measurement system, why not admit that visible workarounds should always be available to the pilot?"

[maynardGkeynes]

the stall was effectively unrecoverable

It would have been easily recoverable if the stall horn had worked properly.

It was working properly, which is part of the problem. The captain was not in the cockpit at the time of the first two stall warnings, and only heard the third. One must question, therefore, the procedure to call him back into the cockpit and then, further, to allow him to direct the crew's response.

[ST27]

As a simple-minded Twotter driver there is one factor here which seems critical to me. If the stall warning sounds I know to get the nose down in a hurry then grab for the apehangers to add power.

In this case, the first question that needs to be answered is why the PF decided to pull the nose up when the aircraft was in otherwise stable flight at the time the autopilot disconnected. That might explain the mindset that had to be overcome to get things under control again.

In this case the stall warning only sounded (twice?) briefly which was probably the critical misleading clue which convinced the crew that they were not stalled.

While the report isn't perfectly clear, the way I read it, the stall warning sounded at least three times: Twice immediately after the PF initially pulled the nose up, while they were in the steep climb, and once after the climb peaked and the aircraft regained airspeed as it began to fall. The report says the aircraft was stalled for the remainder of the flight, so I suspect the stall warning was sounding continuously thereafter.

The PF reacted to the first stall warnings by trying to lower the nose during the climb, but then later applied TOGA power and pulled back before they had recovered. At some point near the top of the climb, he dropped the throttles to idle, but still maintained the nose-up input.

As they started to fall, he again tried putting the nose down, but when the speed picked up, and the stall warning sounded again, returned to full nose-up input, which he retained to the end, even though the attitude was something like +15 degrees, and the AOA+35 degrees. By then, the speed sensors were working again, so all his instruments should have been working as well.

If the stall warning had behaved as expected by any sane pilot and sounded continually during the stall then any sane pilot would have realised that the aircraft was indeed stalled and applied the correct recovery.

That's a good question, which brings me back to what the PF's midset was, and why he didn't react correctly to the stall warning in the last 2 1/2 minutes of the flight, even though the warning must have been sounding continuously. That should have been enough time to reconsider what he was doing, and try something different.

Why the hell did the stupid automation silence the stall horn ?
THIS SHOULD NOT HAPPEN

From the BEA news release:

"When the measured speeds are below 60 kt, the measured angle of attack values are considered invalid and are not taken into account by the systems. When they are below 30 kt, the speed values themselves are considered invalid."

Perhaps they consider the instrumentation to be unreliable at those low speeds? The lack of wind noise past the cockpit should have been a clue that they weren't making much headway.

Even if they didn't suppress the alarm when the speed dropped, would it have sounded when the AOA, which is the primary input, dropped to only 4 degrees, as it did near the top of their climb?

[Me Myself]

The amount of rubish posted on this site is just bewildering !
Some professional Bus drivers have posted over and over again what the recovering procedure is, was at the time ! What else do you need ?
Attitude and thrust ! Period !


Had the skipper been at the controls, this would never have happened and yes, it is a shocker to see a captain go to the bunk while the aircraft is toying with a TCZ. This is another no brainer and all this politically mumbo jumbo won't hold for very long.
BEA carefully chosen words in order to control the leaks is not going to change the fact that the next report will be absolutly shattering.
More than meet the eyes doesn't apply in this case but rather what you see is what you get. Plain, simple and brutal.

[aguadalte]

Torquemada60

Reading through the comments on this web site, it appears that 3 pilots made wrong decisions. They chose to fly through a storm when many airlines had deviated their course by miles. The Captain went for a nap knowing there was a storm ahead. The two copilots appeared to be poorly trained and unable to cope with issues.
I hope Air France will be punished and the families compensated for what appears to be recklessness, arrogance and a total lack of professionalism from the pilots.

Don't worry Torquemada, AF is already "paying" and the pilots have already paid, with their lives. With AF, and its pilots, its a question of "company culture". We all know that...

But regarding the Report,
I must say, that I find the (3 pages) BEA report quite insufficient and vaguely accurate. Its quite tendentious, because it doesn't give all the relevant information. One has to read between the lines to figure out that a lot of relevant information was not aired, yet.

First of all, because it doesn't give us a clue (apart from a couple of phrases between the two co-pilots, that point to the general idea that they have lost valid indications and that they were in bed weather) of what was really happening in the cockpit during those 3 or 4 minutes:

- What kind of information was being "shown" on Pilot Flying's #2 PFD/ND (yes I know it is not registered but, are there any conversation clues(?), other than the ones selected by the BEA to be transcripted to the report). Was he first responding to an overspeed indication,(2h10m16s) and only after he has reached 37500ft, (2h10m51s) he has triggered TO/GA in response to another "Stall, Stall" warning?
- What was the role played by the 13º Trimmable Horizontal Stabilizer (THS) plus the 4.9Ton of fuel stuck on the Trim Tank?
- What was the role of the "systems invalidations" design that below 60kts and 30kts, cancelled the "Stall, Stall" Automatic Call-Out Warning, and may have lead the pilots think they were out of it during precious seconds?
- We know that the Captain was able to reach the cockpit. It seems that by that time the speed read-outs were so low that the Stall Warning was out...but, what was his assumption of what was going on? By then, the aircraft was at about 35000ft, Pitch at about 15º and thrust 100%. Did he tell anything? Did he try to help? Didn't he notice the high pitch? Didn't he notice, the wings bouncing and the PF saying (only a minute after his entrance in the cockpit), that they were reaching FL100?
- Was the PNF so occupied, handling the ECAM, that he didn't notice the vertical speed rate, until FL100 was called off?

This report has a lot of unanswered questions and a lot of information gaps.

I think we should render the benefit of doubt to the crew, before crucifying them on the public arena. I do know its not "natural" to pull-up during a Stall, but I'd like to know more about what has really happened in that cockpit, before judging them.

Again, as to "Company Culture", as to why enter the thunderstorm, as to lack of formation, etc. I think we should get back to the times when, Safety was never compromised in lieu of commercial revenues, "Fuel Saving" programs, low cost tickets, passenger protection programs, etc.
Do you find any logical explanation to why is it possible to pay less for a two hour flight ticket, than for a 15 minute ride taxi? Well, millions of passengers do!

[awqward]

(Peter?) Garrison, could a full tail stall develop with severe icing build-up (combined with an aft CG)....It seems the pitot tubes iced up...the met conditions were highly abnormal...is it feasible that there was also a failure of the tail de-icing system? Although I have to agree with Torquemada60: the Capt should not have left the FD with the impending CBs ahead and they should have made a larger diversion (fuel margins notwithstanding)

[armchair_pilot]

Not So Simple ?

IMO there's too much over simplification going on here, I don't believe the pilots were incompetent - even 20 years after I last flew anything the recovery from a stall is ingrained. Hence:

(a) Are we sure that the nose up trim was added manually rather than automatically by the aircraft ?

(b) Why did the pilot keep pulling up when, with our near perfect knowledge, the opposite would have been more successful ? I'd suggest that the cockpit's displays had completely confused the sufficiently experienced crew and they were tracking a plausible (but wrong) hypothesis based on what they concluded from the displays.

(c) Perhaps there is something wrong with the Airbus assumption of how crews process data, particularly conflicting data combined with lots of visual and aural warnings, set against a system (=the plane) whose outputs depend upon which mode the system is in ?

(d) It would seem that the Airbus approach to automation works in 99.99% of the time and during that time is probably the most efficient and safe approach. However in the remaining 0.01% of the time a 'deep s***' mode is needed whereby only the basics are displayed & communicated. If these had included a height, an angle of attack and a (perhaps GPS calculated) speed, things would have become clear to the crew.

(e) How much height is needed to recover from an A330 near stall, stall and deep stall ?

[bearfoil]

Garrison

Your post is a fine starting point to unwind this accident. The downside, is that what you have written is well known, and to this moment, unaddressed.

How does one tell a Butterfly from a moth?

[smith]

Could altitude and speed data from gps not be entered into the flight computers and compared to pitot static data with any major deviation between the the two setting off alarms? SLF here

[robdean]

From my perspective (psychology qualified) the statement by the PF of "no valid indications" is significant. It strongly implies to me that he was sufficiently disoriented by invalid readings from SOME instruments to become incapable of distinguishing whether he had any useful information in front of him at all.

This reaction that 'it has all gone haywire' is lethal because it undermines the best (and often only) means you have to regain situational awareness, which is to work the data. This has been true since the days of doped fabric, and has led to a consistent litany of stalls and CFIT.

Complex modal fbw systems add the peculiar complication that the aircraft itself can, in a manner of speaking, lose situational awareness itself. Of course this is purely algorithmic, but can lead to 'assisting' the pilot in a task at odds with his actual intentions, or inhibiting a warning as a false alarm, despite the fact that given the current flight phase and circumstances such a warning is probably urgently appropriate.

That said, it looks like plain dumbness rather than a side-effect of complexity when a stall warning seems to activate whilst approaching stall but then to quit in full stall, only to trigger again upon attempted recovery.

[thermostat]

stall warning

The ancient Greek.

The stall warning discontinued when the speed was BELOW 60 Kts. How in blazes do you fly a large jet at 35000 ft at a speed of 60 Kts ?? At that height the speed has to be above 270 Kts. The stall warning is of no use to the pilot below 60 Kts when your plane should be at 272 Kts IAS at 35,000 ft, and 0.80 mach speed. The normal speed was 0.82 M and they slowed to 0.80 because of the turbulence. Flying into CBs close to the "coffin corner" is nuts. Remember too that the 'plane rose to 38,000 ft which meant they had no room for error.

[Garrison]

There are all sorts of possibilities, including an unreported (by ACARS) tail deice failure, but a good Occamite would say that what we have here is sufficient unto itself. The vulnerability of the Thales pitots was known in advance; and the proneness of flight crews to become utterly confused and to do seemingly insane things when confronted with inconsistent airspeed/attitude information and uncertainty about systems behavior is well known from previous incidents, notably ones involving pitot icing. So even the fragmentary information we now have is not inconsistent with history (though it seems to be inconsistent with what a lot of pilots think they would have done if they had been there) and shouldn't astonish us. What would be astonishing is if there is no rethinking of pilot/flight data interfaces, especially with regard to AoA, after this -- though once the dust settles this, like previous crew-befuddlement accidents, will probably come to seem more and more like a freak outlier and less and less like a call for systems improvement.

[bearfoil]

Exactly so. The failure of complex systems used to be less emergent than recently. I remember flicking instruments with my index finger, long after I could have expected a result.

This will more than likely be an easily reducible stream of data leading to insufficient flying speed, and no readily available method to increase it.

Is there an Unreliable AoA in the FCOM ? Will there be?

[ST27]

Reading through the comments on this web site, it appears that 3 pilots made wrong decisions. They chose to fly through a storm when many airlines had deviated their course by miles. The Captain went for a nap knowing there was a storm ahead. The two copilots appeared to be poorly trained and unable to cope with issues.

You are rushing to judgement, prior to the release of the final report. Don't let the media spin on things warp your views of what really happened. Nothing so far says that the crew was poorly trained. You have simply made that assumption. There are too many inconsistencies in the report to understand what was going on, what the crew saw, and why they reacted the way they did.

As far as the captain going for his rest period, why do you see that as a problem? Flying across the equator through less-than-perfect weather is fairly common. The First Officer, who was presumably flying at the time, had more than twice the experience on that type of aircraft as the captain. The captain may have had full confidence in his abilities.

I hope Air France will be punished and the families compensated for what appears to be recklessness, arrogance and a total lack of professionalism from the pilots.
228 people died as a result.

Now you are truly jumping to conclusions before the incident has been fully explained. Judge, jury, and now you want to assign punishment even though you don't understand what transpired and why. I hope you never get on a jury, as it doesn't appear you would be swayed by any evidence that didn't fit your fixed opinion.

Arrogance? Recklessness? Lack of professionalism? Don't be absurd.

[bearfoil]

Although it will undoubtedly be a very minor finding in the Final, in France it is considered tres gauche to question a man who is doing his job. I cannot say I have a problem with that. These were all eminently qualified men.

[ST27]

Could altitude and speed data from gps not be entered into the flight computers and compared to pitot static data with any major deviation between the the two setting off alarms? SLF here

The critical speed is airspeed, not physical speed over the ground. Consider a simple case where there is a 200 knot tailwind. The result will be a 200 knot variance between the GPS and true airspeed, yet the aircraft would be perfectly happy in the air. No warning needed.

[thermostat]

AoA

Garrison

Is it possible that the super cooled water that made the 3 different pitot tubes block with ice simultaneously, would also have prevented the AoA from moving thereby rendering it useless?
I don't think that some folks on this site realize the amount of chaos super cooled water can cause. At that altitude the air is so pure that the water cannot freeze. It needs some impurities for the ice to form. When this super cooled water came into contact with the aircraft it immediately turned to ice. It may also have formed on the nose and wings of the the 'plane making it heavier and changing the C of G.

[Rananim]

crazy design

The autotrim and stall warning inhibit were lethal traps here.I love it when Airbus guys come on here and say its easy,just follow the procedure.The Airbus isnt so easy in non-normals is it?Its not really an intuitive a/c.How many different laws are there?Even an AIrbus guy didnt know all of them and had to be corrected.And this autotrim givesfull ANU in one LAW but a second later autotrim is no longer available because we're now in a new degraded law.And so the poor pilot who is used to autotrim is now stuck with full ANU trim.Hes got to second-guess what these Flight Computers are doing.In real time.And then the stall warning..on one minute,off the next even though the stalled condition hasnt changed.So when he did push down and the stall warning re-activated,he reacted instinctively and reversed his decision.My God,it would confuse the hell out of me too.And all of this going on in the middle of the night over the ocean with no tactile/visual feedback.No trim wheel.No control wheel.Just two sidesticks.The PF could be performing the direct opposite of the required recovery procedure and the PNF wouldnt be any the wiser.Myhat is off to you guys who fly this thing,itreally is.

Still,I expect they will still go with pilot error as probable cause.Technically,it was.But system design didnt help them any on this dark stormy night.

[thermostat]

Basics

182 Flyer

Basics unfortunately have nothing to do with flying a jet at high altitude close to the "coffin corner" and then having supercooled water (ice) block all 3 pitot tubes simultaneously, build up on the wings increasing the weight and changing the C of G. This is a totally different kettle of fish. The way to prevent this type of incident is to stay away from CBs when close to the coffin corner. That's what they didn't seem to have done.