Air France 447 - On topic only!

[Evan]

So, let me get this right - in a hudson (or similar land) situation, a pilot may unintentionally leave himself without flare authority?

Full backstick was applied, and the aircraft protected against the stall (which is a good thing). However, upon reaching the flare, if you already had full backstick, and you would have no further movement available. Indeed, if the airfract actively pitched down, to create a "normal" flare environment, it could actually pitch you further down?

That seems to be it in a nutshell. This is why you need to follow green dot and F speed and be stabilized in pitch descending below 50'. I question if the 2° nose down order occurs when in the Alpha Protect law, which is an AoA demand law that might override this feature. Since the stick was already at the back stop in this condition, there is no point to the nose down order.

One thing we can't really say, however, is how successful it would have been on a 737. Sully had the protections, and he KNEW he had the protections. A 737 pilot faced with the same situation would, hopefully, have been aware to not stall, and the stall warnings would have been significant compared to the Airbus, where it isn't so critical.

Very true. He may have flown the 737 much differently, by 'feel' and been much more cautious with the AoA. The question is: on the A320, was he aware of the ramifications that not following green dot and/or flying at alpha max would have in limiting his ability to flare? This is what I keep getting back to. The Airbus has its own piloting requirements and one of those is a deep understanding of the interplay between systems and control law. Airbus pilots need to know WHY it is so important to fly at green dot in this situation and not to get into a high AoA situation in the first place.

Sully has something like 4700 hours in the A320, and I think he is a hell of a pilot in any case, and I'm sure he had his reasons for not lowering the nose and taking on airspeed. Maybe he was concerned with floating too long in ground effect and running out of clear 'runway'. But I do want to know that Airbus pilots are being taught a deeper understanding of how the aircraft differs from what they may have been used to it the past.

[Black Ram]

If it was a 737, I like to think he would have tried landing with more speed. But I think in the end, as I said before, things worked out for the better. Low speed, safe water landing.

[Black Ram]

Anyway, can the 737 stay controllable at 125knots and at that weight? Does it make a difference if it's a 737NG (which I think US Airways doesn't have)?

[Evan]

If it was a 737, I like to think he would have tried landing with more speed. But I think in the end, as I said before, things worked out for the better. Low speed, safe water landing.

Based on the certification requirements and the recommendations of the NTSB, I think the best thing to do here is come in fast, flare and then let ground effect work its magic until you settle into the water at both a low airspeed and a low vertical speed. Ditching certification assumes a rate of 3.5fps. Sully impacted at 12.5fps. or almost 3x that rate. This caused extensive damage to the lower fuselage and caused water to enter, and the plane to eventually sink. However...

Although the airplane impacted the water at a descent rate that exceeded the Airbus ditching parameter of 3.5 fps, postaccident ditching simulation results indicated that, during an actual ditching without engine power, the average pilot will not likely ditch the airplane within all of the Airbus ditching parameters because it is exceptionally difficult for pilots to meet such precise criteria with no power.

Another thing to consider is that once you get below 34th St, you hit traffic. Sully got it down early at around 59th street, which was ideal. Bleeding off airspeed in ground effect might have taken them too far into harms way.

And he also started the APU as a memory item. It was not on the part of the checklist they performed. This was critical because it kept the aircraft in normal law. So, point Sullenberger.

This high rate of vertical speed was due to his inability to arrest the descent rate with flare, because the AoA protection maintains a 3° margin below critical angle and he had used up the rest due to insufficient airspeed and flaps 2. So, point Airbus.

However...

The NTSB concludes the captain’s difficulty maintaining his intended airspeed during the final approach resulted, in part, from high workload, stress, and task saturation.

But I think, and the NTSB seems to think, that this is mainly due to pilot training shortcomings.

Postaccident flight simulations indicated that attaining the Airbus ditching parameters without engine power is possible but highly unlikely without training.

Recommendation: Require Airbus operators to expand the angle-of-attack-protection envelope limitations ground-school training to inform pilots about alpha-protection mode features while in normal law that can affect the pitch response of the airplane. (A-10-74)

Now, back on topic: It's time to stop blaming the aircraft just because it requires different pilot instincts. The focus should be on things like this instead:

Airbus’ training curricula does not contain information on the effects of alpha-protection mode features that might affect the airplane’s response to pilot sidestick pitch inputs.

I mean wow. There's a serious problem there. Really Airbus, you can't certify a new type of flight control system if you don't train the pilots on a systems level.

[MCM]

The problem then becomes, Evan, how complicated should you make your systems, if they need that level of knowledge?

You CANNOT expect your pilot to be trained in thousands of specific situations. It just isn't practical. It isn't even practical to train them to understand so many different modes and submodes. It just isn't good design! Thats why aircraft need to be as simple, and pilot friendly, as possible. This is particularly so in the case of the basis of all flight - the flight controls.

This is EXACTLY what many on the "Boeing" side of the fence are arguing, and what is constantly being missed - While YES, almost all pilot error incidents can be fixed with training, it is a fault of a design when the training required becomes so significant that it is impractical to train and retain that level of knoweldge.

Pilots have limited memory and recall ability, no matter how much you train.

I have been trying to follow the argument over the different submodes over the last few pages, and it is pretty evident that the actual function of these small submodes modes is NOT made clear at all. If a pilot needs to know this level of detail to safely fly the aircraft, then the system should NOT be designed like this.

If the pilot does not need to know that level of detail, but just the basics, then thats another thing entirely. But you can't then criticise him for not knowing the detail which is not required.

[Black Ram]

I think everybody and everything did right. A normal flare is great when you have speed. In this case, when speed has been bled off and there is no way to get it back, and when the ground effect technique has not been applied, alpha protection comes in to give the max performance, though not necessarily the highest AoA permissible. According to the NTSB report, some of the back-pressure was attenuated in the final seconds, but for the better. At least that's my understanding.

[Evan]

The problem then becomes, Evan, how complicated should you make your systems, if they need that level of knowledge?

MCM, I see your point and I agree with you on that point: we can't expect pilots to have an overly complex awareness of systems and all of their sub routines and interactions. If the aircraft requires that level of technical knowledge to fly in abnormal situations, then it is unsafe.

But I still fail to see where this is the case. In the case of Cactus 1549, the only thing missing was a basic instruction on alpha protection limits during flare (respect green dot / F speed, manage your AoA budget, use flaps strategically, provide margin for flare), and a more practiced procedure for dual engine failure (not flameout) and ditching in normal law. In the case of AF447, (aside from the universal issue of stall avoidance procedure) it was only a matter of memorizing a procedure to stabilize airspeed and manage AoA without static stability feedback. It was two items long!

But, if you teach the basic control law concepts, you don't need to teach procedures for every combination of events. Pilots can then extrapolate, just as you do with conventional aircraft knowledge. I wonder if the problem doesn't lie there. Are they training only for rote procedural memorization and not a greater conceptual understanding?

If I didn't have such a high impression of commercial pilot intelligence I would agree that this conceptual knowledge is unreasonable. But if pilots can be taught the laws of aerodynamics and gravitational physics, the laws of Airbus shouldn't be too much to ask. I think it is the standard of intelligence that needs to be focused on more than the operating system requirements.

[Evan]

I think everybody and everything did right. A normal flare is great when you have speed. In this case, when speed has been bled off and there is no way to get it back, and when the ground effect technique has not been applied, alpha protection comes in to give the max performance, though not necessarily the highest AoA permissible. According to the NTSB report, some of the back-pressure was attenuated in the final seconds, but for the better. At least that's my understanding.

In the post accident interviews, Sully said that he thought he had stayed near green dot speed and always above VLS, and he hadn't. He was probably focused out the windscreen more than on the airspeed tape. "Tunnel vision" is common under stress, especially in a navigational emergency like this. He tried to extend the initial glide too much with pitch and didn't leave enough margin for flare. He calculated wrong. Green dot was there to guide him and was providing the optimum speed for the approach with margin for flare, but the investigators think he was not focusing on it. He stayed at flaps 2 when that might not have been the best move. In the final moments he played the AoA demand law to his benefit, commanding alpha max and holding it there, and letting the plane determine the actual pitch without losing roll stability.

But he got it down wings level in the right place at the right airspeed at the wrong rate of descent. Maybe that was the compromise this situation required.

[TeeVee]

The problem then becomes, Evan, how complicated should you make your systems, if they need that level of knowledge?

You CANNOT expect your pilot to be trained in thousands of specific situations. It just isn't practical. It isn't even practical to train them to understand so many different modes and submodes. It just isn't good design! Thats why aircraft need to be as simple, and pilot friendly, as possible. This is particularly so in the case of the basis of all flight - the flight controls.

This is EXACTLY what many on the "Boeing" side of the fence are arguing, and what is constantly being missed - While YES, almost all pilot error incidents can be fixed with training, it is a fault of a design when the training required becomes so significant that it is impractical to train and retain that level of knoweldge.

Pilots have limited memory and recall ability, no matter how much you train.

I have been trying to follow the argument over the different submodes over the last few pages, and it is pretty evident that the actual function of these small submodes modes is NOT made clear at all. If a pilot needs to know this level of detail to safely fly the aircraft, then the system should NOT be designed like this.

If the pilot does not need to know that level of detail, but just the basics, then thats another thing entirely. But you can't then criticise him for not knowing the detail which is not required.

1,000,000 x kudos MCM. very well stated indeed.

evan's problem is that he appears to believe, in his heart of hearts, that in the "God only knows how many hours" he spends poring over the manuals and what-nots published by airbus, that since he finally found the right procedure, all airbus pilots should, in a matter of literally seconds or maybe less, be able to recall and implement the procedure while simultaneously monitoring the myriad instruments and, oh yeah, trying to figure out what "law" and mode or sub-mode the damn plane is in to begin with, since the proper procedure to employ will be based upon, law, mode, sub-mode and paint scheme! oh and not to forget those niggling things like thinking about dying, family, passengers etc etc. sure! why not. pilots after all are smart folks. they can do it without a problem!

i can't even fly a flight sim. just looking into a modern airliner cockpit amazes me! it's is simply incredible that anyone can successfully monitor and interpret all the displays, gadgets and doodads, let alone really understand the extremely fluid and dynamic goings-on behind the instrument faces, or should i say, behind the glass panels?

this particular point leads me to this: either FULLY and I mean absolutely fully automate flying and eliminate human pilots (e..g., HAL 9000), or simplify the system so that the humans IN CHARGE can successfully handle the SYSTEM without being overloaded by information that requires multiple CPU's and absolutely zero distractions such as those pesky thoughts we humans tend to have every once in a while, to process.

since the former aint happening anytime soon, my vote is for the boeing philosophy which seems to be the latter.

[Highkeas]

................this particular point leads me to this: either FULLY and I mean absolutely fully automate flying and eliminate human pilots .................

I attended a presentation by an aircraft company conducting studies associated on this very concept and their prediction is it may well happen sometime in the future. One of the points mentioned was that passengers at one time had real big concerns with removing the attendant from elevators and drivers from certain rail systems (I think BART was mentioned).
[Sorry for the diversion from AF447.]

[EconomyClass]

Did Sully ever weigh in on AF447? Did he say, as some here do, "the pilots could have saved it but made errors that doomed everyone on the plane"? It is easy to praise Sully and then ignore his verdict if he made one. Again, those pointing fingers at the pilots haven't flown that route in that plane. What ever happened to "walking a mile in the other man's moccasins"?

[Gabriel]

I think this point has been overlooked repeatedly, and posing the counterfactual "What if it had been a 737?" doesn't make much sense. Sully wasn't on a 737, he was on an Airbus, and he flew the Airbus in precisely the manner necessary, with protections in place and knowing how they would respond, to have a safe landing.

I don't agree.

I don't think that Sully intentionaly flew in a way that left him without further nose-up authority knowing that the plane would touch down with more than 3º of path angle.

And if he did, it was not wise.

It's much better to level of at just a few feet above the water, and with enough speed to level off to begin with, and hold it there bleeding off speed until the plane smoothly touches down.

[Evan]

It's much better to level of at just a few feet above the water, and with enough speed to level off to begin with, and hold it there bleeding off speed until the plane smoothly touches down.

Yes, if you have that luxury, that's the way to do it, they proved that in the follow-up, but once you get below the 34th St ferry terminal there can be a lot of traffic on the 'runway'. I have no idea what was there when they were coming down, but touchdown was around 59th St and the first civilian rescuers arrived very shortly after touchdown.

[Gabriel]

Yes, if you have that luxury, that's the way to do it, they proved that in the follow-up, but once you get below the 34th St ferry terminal there can be a lot of traffic on the 'runway'. I have no idea what was there when they were coming down, but touchdown was around 59th St and the first civilian rescuers arrived very shortly after touchdown.

There are other ways to manage your glide.

First, you have flaps and spoilers if you want a steeper descent without getting too close to a too slow speed. In fact, the flaps will lower the definition of what is a too slow speed so you can both have a steeper angle and land slowly (warning: too much flaps can make the glide too steep and complicate the flare).

Then, I'm quite sure that there is much more free water than ferries at any time in the Hudson, which is a very long and wide river (compared with a runway, I mean). I'm quite sure that ferries are well visible from the cockpit at a good distance too. It should not be difficult to steer a bit left or right to a safe spot. The plane is not going to skid a long distance anyway. And I don't think that the float in ground effect will last too long wither, especially if you added a bit more of flaps (I think that the config that they used for landing was a slats only one).

I really don't think that Sully said "I better spoil my glide and the float at flare by flying too slow so I don't reach the ferry terminal". It's too dangerous. How in the Earth does a pilot calculate the right speed as to reach the ground at Alpha max with the stick all back, without any nose-up authority left, but with a still low enough vertical speed as to survive the ditching? There is no "purple dot" to guide you through this. Had Sully flown maybe 10 kts slower or started the flare 10 ft higher or a little more markedly, the result could have been very different. In fact, it looks that he was already a tad on the "airspeed too slow / flare too high" side.

[guamainiac]

Face palm.