AF447 wreckage located...
Discussion
Simpo Two said:
They had artificial horizons and night flying in WW2; the standard 6-dial instrument panel. That seemed to work.
They clearly had doubts about the accuracy of their instruments (particularly airspeed) and it seems that they had severe information overload as well.The operating manual states a combination of engine rpm and aircraft attitude setting that would keep the airraft at a safe flight condition but they appear not to have used it.
It seems they basically did not recognise they were in a stall and therefore didn't manage the aircraft accordingly.
I'm certainly not going to judge the crew however - it's one of the hazards of modern cockpits that a sudden deluge of warnings, alerts and bongs on a previously relaxed crew in the middle of the night, can severely affect crew performance.
MarkK said:
They clearly had doubts about the accuracy of their instruments (particularly airspeed) and it seems that they had severe information overload as well.
I read that the pitot tubes were giving false data - so they were either going too fast into turbulence or too slow. Either way the pilots did not have an accurate sense of their speed.A dreadful accident, whatever the cause.
NDA said:
I read that the pitot tubes were giving false data - so they were either going too fast into turbulence or too slow. Either way the pilots did not have an accurate sense of their speed.
A dreadful accident, whatever the cause.
Yeah I agree - I suspect the air data probes had a degree of ice formation hence the incorrect readings.A dreadful accident, whatever the cause.
DamienB said:
There's bugger all excuse for not carrying out stall recovery actions correctly. One of the first things you learn to do in training and by the point in your career that you're flying passengers around it should be instinctive. Something desperately wrong with AF's training I think!
A pilot is only going to commence stall recovery if they realise they are in a stall in the first place. At night, in turbulent conditions, with dubious instrumentation readings and sensory overload it would be relatively easy to overlook the stall condition.An iced up pitot-static system can cause incorrect airspeed, altitude and vertical speed indications amongst others. It can also cause the instruments to show the exact opposite of what the aircraft is actually doing. This has caused the loss of many other aircraft over the years.
When you look at the pilot inputs, i can't believe they knew that the aircraft had entered a deep stall!
I wonder if being fixated on a rapidly unwinding altimeter, and maybe hearing a lot of wind rushing past the cockpit (AOA > +60deg !!) fooled the pilots into believing they were in a fast dive (for which an idle throttle setting and prolonged pitch up request (via the side stick) would be normal.
Unfortunately, in this case, it looks like the horizontal trim system took this constant (over about 3min) request for nose up and translated it into a 13deg THS value, a setting that will override any nose down elevator authority.
What is not answered however, is why 3 pilots came to this conclusion, in what appears to be direct opposition to the aircrafts pitch (roll and horizontal) data from the inertial system, and with GPS ground speed data to suggest a forward speed of only ~100kts.
Unfortunately it rather looks like had they just gone back to "seat of the pants" pitch + power flying they would have successfully recovered the upset with sufficient altitude.......
Not that i'm blaming them, as they found themselves as the last link in a long chain of events, all of which conspired to make critical actions a necessity.
What is certain however, is that the BEA must establish beyond doubt exactly what each flight display system was showing throughout, and then rewrite the SOP book on high speed stalls in FBW aircraft a bit
I wonder if being fixated on a rapidly unwinding altimeter, and maybe hearing a lot of wind rushing past the cockpit (AOA > +60deg !!) fooled the pilots into believing they were in a fast dive (for which an idle throttle setting and prolonged pitch up request (via the side stick) would be normal.
Unfortunately, in this case, it looks like the horizontal trim system took this constant (over about 3min) request for nose up and translated it into a 13deg THS value, a setting that will override any nose down elevator authority.
What is not answered however, is why 3 pilots came to this conclusion, in what appears to be direct opposition to the aircrafts pitch (roll and horizontal) data from the inertial system, and with GPS ground speed data to suggest a forward speed of only ~100kts.
Unfortunately it rather looks like had they just gone back to "seat of the pants" pitch + power flying they would have successfully recovered the upset with sufficient altitude.......
Not that i'm blaming them, as they found themselves as the last link in a long chain of events, all of which conspired to make critical actions a necessity.
What is certain however, is that the BEA must establish beyond doubt exactly what each flight display system was showing throughout, and then rewrite the SOP book on high speed stalls in FBW aircraft a bit
Max_Torque said:
Unfortunately it rather looks like had they just gone back to "seat of the pants" pitch + power flying they would have successfully recovered the upset with sufficient altitude.......
That was the angle behind my posts... too much complexity means the pilots can never work through the clutter of electronics or know which instruments are right and which are wrong. First principles tell me all you need is an artifical horizon and airspeed indicator, no?Or was it in a flat spin?
It will be interesting to see what happens when they plumb the data into a simulator and see if other pilots can save it.
Simpo Two said:
It will be interesting to see what happens when they plumb the data into a simulator and see if other pilots can save it.
I believe this was shown in a documentary either this year or last year on tv. They put two experienced pilots in a simulator with the exact condition the AF aircraft experienced. They succeeded numerous times with manual flight control to fly through the storm with no problem at all. Obviously pilots in a simulator and it happening for real 3.5 hours in to a flight are two different things. It is all good and well us speculating and saying they could of saved the aircraft, they could of done this they could of done that etc, but at the end of the day they were faced with an awful situation with everything against them i.e being a night flight (no visual reference), a pretty awful and unique storm with unusual super cooled water droplets (which was shown in the documentary to have iced up the pitot tubes.DamienB said:
MarkK said:
A pilot is only going to commence stall recovery if they realise they are in a stall in the first place.
What, like a stall warner going off? Or perhaps the altimeter unwinding at a frightening rate? No excuses.These days pilots of modern automated aircraft are very much like security guards. For most of the time (99.99%) you hope you don't need them, so they just have to pass the time(boredom & routine etc become overiding), but then, for a few brief moments where peoples lives hang in the balance, you need them to response accurately and quickly (clear, fast, logical thinking) without error. Those two demands are very much at opposite ends of the human spectrum.
It's also easy to look at the accident report, with the hindsight knowledge that actually the plane was operating without serious fault, and cast doubt on their actions. In reality, faced with a barrage of warnings, possibly turblent and unexpected flight dynamics, they had no way of determining, in the little time they had, the real status of the aircraft etc.
DamienB said:
What, like a stall warner going off? Or perhaps the altimeter unwinding at a frightening rate? No excuses.
I don't think you can comment on this in all fairness. Like already said, blocked/iced up pitot static tubes can cause opposing and conflicting info.The altimeter may not be winding off at all if blocked. The airspeed indicator could be showing increasing airspeed if the pitot system is blocked, the aircraft may be climbing and slowing down but the airspeed indicator still going up.
The pitot tubes blocked with ice would report the airspeed and altitude wrong. In any case they won't work properly in such a deep stall. Falling out of the sky at such an angle of attack is going to feel like slowing down rather than speeding up anyway, since the plane is effectively going backwards. 
El Capitano said:
I don't think you can comment on this in all fairness. Like already said, blocked/iced up pitot static tubes can cause opposing and conflicting info.
The altimeter may not be winding off at all if blocked. The airspeed indicator could be showing increasing airspeed if the pitot system is blocked, the aircraft may be climbing and slowing down but the airspeed indicator still going up.
Your "what if" scenarios are irrelevant - the altimeter was unwinding; the stall warner was going off. It's pretty clear from the BEA report that the erroneous speed readings were transient - lasting less than a minute - all else from that point on was a total failure to read the important instruments and comprehend what the aeroplane was telling them. Yes, nose-up trim could make life difficult but no attempt to re-trim? No attempt at all to lower the nose when a stall warner is blaring? It's not as if they had the excuse that they were at low level where some pilots will be scared to lower the nose - they knew they were at 30K+ before the upset began.The altimeter may not be winding off at all if blocked. The airspeed indicator could be showing increasing airspeed if the pitot system is blocked, the aircraft may be climbing and slowing down but the airspeed indicator still going up.
If experienced pilots cannot filter out the important information and fly the aircraft then clearly their experience is next to worthless - it's just hours boring holes in the sky. The basic stick and rudder flying techniques aren't being emphasized enough in their training - and their currency regime is failing badly.
The what if scenarios are highly relevant - they did have conflicting data during the incident and that could easily cause the crew to doubt the information that the instruments were giving them.
It's all very well saying the stall warning had gone off and the airspeed was low but what if the other displays say the airspeed is fine and there is no stall condition - which do they believe? What if the airspeed drops sharply on one display but jumps up again on the standby?
What if everytime they lowered the nose the stall warnings reactivate (with good reason but still enough to confuse the crew)?
Especially as previous incidents such as the following were known to have occurred?
"An Air France A340-300 (F-GLZN) en route from Paris to New York encountered turbulence followed by the autoflight systems going offline, warnings over the accuracy of the reported airspeed and two minutes of stall alerts"
Would you stick you nose down and build speed knowing you are about to enter the turbulent conditions of a thunderstorm if you were not absolutely certain you were stalled.
What if you then see the speed building above the stall speed and the altitude descending rapidly (easily possible with a screwed air data system)? Would you not then assume you were out of the stall and begin to raise the nose before you increase speed towards dangerous limits?
Are you seriously suggesting that this whole accident was simply caused by the crew not hearing a stall warning and failing to lower the nose?
There is ALWAYS more to these accidents than meets the eye and I'll be surprised if the full report doesn't highlight serious failings in the human-machine interface that contributed severely to this accident along with a failure in the air data system.
I agree that in the cold light of day if the crew had just set a safe attitude and engine setting (as the pilots in the sim did), the whole accident could have been avoided. However, that is a very simplistic view and I'm sure many other factors combined to confuse a crew with many thousands of hours between them.
It's all very well saying the stall warning had gone off and the airspeed was low but what if the other displays say the airspeed is fine and there is no stall condition - which do they believe? What if the airspeed drops sharply on one display but jumps up again on the standby?
What if everytime they lowered the nose the stall warnings reactivate (with good reason but still enough to confuse the crew)?
Especially as previous incidents such as the following were known to have occurred?
"An Air France A340-300 (F-GLZN) en route from Paris to New York encountered turbulence followed by the autoflight systems going offline, warnings over the accuracy of the reported airspeed and two minutes of stall alerts"
Would you stick you nose down and build speed knowing you are about to enter the turbulent conditions of a thunderstorm if you were not absolutely certain you were stalled.
What if you then see the speed building above the stall speed and the altitude descending rapidly (easily possible with a screwed air data system)? Would you not then assume you were out of the stall and begin to raise the nose before you increase speed towards dangerous limits?
Are you seriously suggesting that this whole accident was simply caused by the crew not hearing a stall warning and failing to lower the nose?
There is ALWAYS more to these accidents than meets the eye and I'll be surprised if the full report doesn't highlight serious failings in the human-machine interface that contributed severely to this accident along with a failure in the air data system.
I agree that in the cold light of day if the crew had just set a safe attitude and engine setting (as the pilots in the sim did), the whole accident could have been avoided. However, that is a very simplistic view and I'm sure many other factors combined to confuse a crew with many thousands of hours between them.
Munter said:
So you'd believe a stall warning over the reading saying you're going too fast? Where do you draw the line on which warnings you believe and those you don't?
Absolutely. They had no idea how fast they were going, no idea which way the nose was pointing, and probably no idea what altitude they were at. pulling up rather than down seems like a sensible idea in the circumstances.I imagine the manual will get rewritten in these circumstances though.
Gassing Station | Boats, Planes & Trains | Top of Page | What's New | My Stuff