Air France 447 - On topic only!

[Gabriel]

It doesn't seem likely that such a pitch input could be accidental, but nothing else seems likely here either.

That's why my puzzlement (if that word even exists).

Here I can recognize that the pilot could have thought "Ok, I pushed down and It didn't work, I pull up and the warning stop".

Why would a pilot with basic PPL knowledge think this?

Because that is what actually happened. He pushed down (briefly) and it didn't work, and unfortunately when he pulled up again the stall warning stopped, not because of the pull-up itself, but because the speeds became invalid and with that the AoA became invalid too, but that was probably not known to the pilot.

There's a natural tendency to associate "causation" with "correlation": If I pulled up AND the stall warning stopped, then probably the stall warning stopped BECAUSE I pulled up. Which was wrong in this case.

Didn't it?

Properly trained pilot: "UAS, ok... pitch to 5° up, lets get a hand on the TL's and move them back into CL... ok, UAS checklist please.

(everybody lives)

Freestyle pilot: "Hmmm, time to use my well-honed PPL skills from all my time in the 172. Ooops, stall warning... let's try this. Oh good, no stall warning... lets keep that up then... wait, what's it doing now?!

(everybody doesn't live)

Don't fit the facts to your discourse. The event with the stall warning going on and off happened one minute after it was continously on. It was a tad too late by then. If the pilots worked so hard to get in that situation to begin with, its unlikely that at that point they will suddenly be "illuminated" to find the solution.

I'm restricting the analisys to the point of the autopilot disconnect through the climb, the stall warning at 37500 ft, and the initial reaction to that that made the plane climb again to 38000ft and fully stall.

PPL pilot: Hmmm, I've lost my speed but the plane was flying well before, so I'll try to keep it stable just as it was flying. In particular, I would not perform a maneuver that I would not never perform regardless of what's working and what not, and if I find myself accidentally in that situation, I will correct for it, and by no means will I apply and hold pull-up inputs upon the triggering of the stall warning"

Or lets go to the classroom:

Instructor: "Ok future Airbus pilots, this is the memory procedure for UAS... Pilot A, I see you have your hand up?

Pilot A: "Why do we have to do these exact things? Why can't we just fly at attitudes, power settings, and vertical speeds that give us familiar, healthy airspeeds?"

Instructor: "Because this works. Because the engineers at Airbus who have designed this airframe and tested it extensively in wind tunnels and test flights and performed exotic calculations you can't begin to understand have determined that this is the best course of action to stabilize the aircraft while you work the checklist to find a more exact solution."

Pilot A: "Sure, but I bet none of them ever flew a 172 inverted in a snowstorm with a broken arm".

Instructor: "I'm sorry, this is the Airbus class. The Boeing class is down the hall."

The depiction of events here speaks so blatantly of poor training (lack of CRM has nothing to do with type) that I do sincerely hope it results in the death of Air France.[/quote]
And the grounding of the first instructors these pilot used in their PPL course.

Think of this:

Why do you think that a pilot trained to do it would apply the UAS memory items when he cannot apply the much more basic stall memory item????

I'm not defending Air France. On the contrary. What I'm saying is that Air France alone could have not done it, no matter how crappy their instruction (unless they instrucet to do exactly what they did, which wasn't the case)

And while grounding Air France, please also ground all those other airlines where there was an UAS event and the procedure was not followed (that is, all other UAS events surveyed by the BEA), even when in those other cases the pilots got through it with their common sense PPL skills. The lack of application of the correct procedure is equally unacceptable regardless of the final outcome.

[Fear_of_Flying]

Which is why memorized procedures MUST be trained and memorized and religiously followed in upset situations. Not to mention CRM. The depiction of events here speaks so blatantly of poor training (lack of CRM has nothing to do with type) that I do sincerely hope it results in the death of Air France.

You must have said this twenty times, so I guess why stop now? To me, the problem with this "solution" to what happened - using memory items - is that it would not require a better trained pilot at all. Just program the plane to do your memory items when UAS occurs for a number of seconds.

The advantage that CA's approach has is that while he still accomplishes what you want, stabilizing the aircraft in order to diagnose and correct the underlying problem, he retains his ability to use his judgement and assess the specific variables at hand.

Maybe we just have different ideas of what "better trained" means. To me, a better trained pilot is one who understands and can fly the aircraft better, not someone who has memorized memory items and checklists, and acts without thinking, with no comprehension of why he is doing what he's doing.

As for what this pilot did, I completely agree that he "just tried something", and it seemed to work. Stall warning on, we're stalling. Stall warning off, everything's ok. The behavior of the stall warning was very unfortunate (though does not explain pulling up in the first place), and no doubt will be fixed as a result of this accident.

[Evan]

Why do you think that a pilot trained to do it would apply the UAS memory items when he cannot apply the much more basic stall memory item????

I'm not excluding that! Both UAS and approach to stall require instincts derived from practical repeated training that this pilot obviously was not getting. But if the pilot had a strong training in the first item, he would never have needed his strong training in the second.

But an important distinction must be made between instinctive reactions to sudden events and rational thought-out responses based on cognitive knowledge. I thought the lesson we learned from Colgan (crashed by a former PPL instructor) was that cognitive processes can be warped by temporal compression, i.e. the shock factor. You can't overstress the importance of pilot instincts over pilot knowledge in these situations. There is a distinction there.

Keep in mind that there are only about 7 of these memory items for the A330. Learn these 7 tenets of the Airbus religion and thee shall not perish.

And while grounding Air France, please also ground all those other airlines where there was an UAS event and the procedure was not followed (that is, all other UAS events surveyed by the BEA), even when in those other cases the pilots got through it with their common sense PPL skills. The lack of application of the correct procedure is equally unacceptable regardless of the final outcome.

I totally agree about grounding all of them if they have not made changes to their training requirements. Please do. But of these, only Air France is now guilty of involuntary manslaughter, and I hope it takes them down (or at least removes those responsible for these training deficiencies, all the way up the ladder).

[Evan]

To me, the problem with this "solution" to what happened - using memory items - is that it would not require a better trained pilot at all. Just program the plane to do your memory items when UAS occurs for a number of seconds.

Actually, I'm all for that. I already suggested that a provisional autoflight mode might be a good idea.

The reason the memory items are better than a pilot's accumulated knowledge is that they do not require further rational thought to arrive at them; if they are embedded in memory, they are available instantly to the mind. Once stabilized, the pilot can assess the situation more clearly and fly by his correct instincts. So why 5° and CLB power? These values are calculated by the manufacturer to provide the best safe compromise until the situation can be assessed and weight can be calculated and a more exact set of values can be derived. For AF447, the actual optimum values were something like 3.5° pitch and 93% N1 for flight above FL30.

[Highkeas]

..............................
Here I can recognize that the pilot could have thought "Ok, I pushed down and It didn't work, I pull up and the warning stop".
.................

I used to work with a system that used pitot tubes designed for a high AOA. Most 'conventional' pitot tubes do not work at high AOA and also are unreliable at low speeds.

If the aircraft was at a 40 degree AOA could the computer have cut-off the airspeed reading resulting in your comment above?

[Highkeas]

..........wow dude....look at the deck angle.

It is possible to have a high AOA and a level deck (and that might have been part of the problem)

[Gabriel]

I used to work with a system that used pitot tubes designed for a high AOA. Most 'conventional' pitot tubes do not work at high AOA and also are unreliable at low speeds.

If the aircraft was at a 40 degree AOA could the computer have cut-off the airspeed reading resulting in your comment above?

That's easy to unswer: I don't know

[3WE]

It is possible to have a high AOA and a level deck (and that might have been part of the problem)

So, the transcript where the Pinnacle guys cite a 20 degrees ANU attitude did not contribute to the stall?

(nor the 30 degree ANU attitude of Colgan, nor the relatively aggressive climb before, during and after a stall warning on Air France)?

It all happened from level attutudes?

[3WE]

Properly trained pilot: "UAS, ok... pitch to 5° up, lets get a hand on the TL's and move them back into CL... ok, UAS checklist please.

(everybody lives)

Freestyle pilot: "Hmmm, time to use my well-honed PPL skills from all my time in the 172. Ooops, stall warning... let's try this. Oh good, no stall warning... lets keep that up then... wait, what's it doing now?!

(everybody doesn't live)

Totally Wrong:

Scenario 1: Gee, airspeed readings are fubared. Let me reduce to 50% power go 5 degrees nose down and start a 1000 ft/min descent.

(Please explain how this kills people)

Scenario 2: Gee, there's a stall warning and instead of continuing to pull up in a fairly steep climb, let me go to zero degrees nose up and 75% power and zero rate of climb where I normally cruise.

(Please explain how this kills people)

We've explained this a zillion times- It's great to follow THE procedure but if it's 5 degrees up or 3 degrees up or 5 degrees down and you chose a power setting that normally works with that attitude, the plane is extremely likely to fly just fine!

[Evan]

Totally Wrong:

Scenario 1: Gee, airspeed readings are fubared. Let me reduce to 50% power go 5 degrees nose down and start a 1000 ft/min descent.

What?! I don't know what scenario that is, but it sure the hell isn't scenario 1.

Scenario 2: Gee, there's a stall warning and instead of continuing to pull up in a fairly steep climb, let me go to zero degrees nose up and 75% power and zero rate of climb where I normally cruise.

75% N1 at FL 350? And 0° pitch? Low thrust and low angle of attack at 205t in low density air. Interesting, in a freestyle kind of way. I will offer my RIP in advance.

[Highkeas]

So, the transcript where the Pinnacle guys cite a 20 degrees ANU attitude did not contribute to the stall?

(nor the 30 degree ANU attitude of Colgan, nor the relatively aggressive climb before, during and after a stall warning on Air France)?

It all happened from level attutudes?

One has to consider the flight path angle.
For example during the last moments of AF447 it had a pitch angle of 18 deg and an AOA of around 45 deg (derived from vert and horizontal velocities). If it had a zero pitch angle it would still have an AOA of 27 deg.

[Gabriel]

75% N1 at FL 350? And 0° pitch? Low thrust and low angle of attack at 205t in low density air. Interesting, in a freestyle kind of way. I will offer my RIP in advance.

What would go wrong? I tell you that if I'm going to aim at a pitch of 0º I'd use even less thrust for fear of verspeeding.

Oh, and 0º pitch is NOT 0º AoA, for God sake! It's time we stop confusing the two.

[Gabriel]

One has to consider the flight path angle.
For example during the last moments of AF447 it had a pitch angle of 18 deg and an AOA of around 45 deg (derived from vert and horizontal velocities). If it had a zero pitch angle it would still have an AOA of 27 deg.

I fully agree with what you say from a technical point of view: AoA = pitch - path (save for vertical air currents), but I also understand 3WE's possition:

If the crew would have kept a normal cruise or even climb attitude for that altitude, the stall would have never happened in the first place, even with the engines at idle, and you would have never had your 45 deg of AoA to work with.

[Black Ram]

BEA Interim Report

Hi all,

The real interim report has been published on BEA's site. It is currently available only in French, with an English translation coming soon.

http://www.bea.aero/en/enquetes/flight.af.447/flight.af.447.php

[Evan]

What would go wrong? I tell you that if I'm going to aim at a pitch of 0º I'd use even less thrust for fear of verspeeding.

I'm no pilot of course, but it seems to me that with an AoA in the < 3° range, 75% N1 is way too low for a 205t A330 to remain in the flight envelope at FL350. But since your airspeed is also now unknown, couldn't reducing pitch also result in overspeed before the thrust reduction takes effect? The QRH wants 90% N1 for 3.5° pitch at this weight and altitude (95% for FL370). Even if they had been below 160t, it prescribes 84% N1 to fly at 2° pitch (90% at FL370).

If this is a best guess, it just proves my point about the importance of not guessing.

Why is there such pilot resistance to following the manufacturer's instructions? Just do the checklists so we can all live the next time.

Oh, and 0º pitch is NOT NECESSARILY 0º AoA, for God sake!

Fixed.

If the crew would have kept a normal cruise or even climb attitude for that altitude, the stall would have never happened in the first place, even with the engines at idle, and you would have never had your 45 deg of AoA to work with.

So, with the engines at idle, and your airspeed falling off, and the vertical speed increasing as the plane descends, and the pitch held at 0°, you are never going to reach critical AoA?