AF447

[surplus1]

To M(F)S and safetypee

Thank you gentlemen for your very informative replys. I now have a much better understanding of what you mean/meant by a “stable stall.” While I would not use that terminology myself it doesn’t matter so long as I comprehend what you intended. I’ll try to address your key points with a single post to both of you. I'm too verbose as it is.

BTW, neither one of you answered my question with respect to the pilot’s ability to deliberately take the aircraft out of Alternate Law and force it into Direct Law if he so desires and I don’t know the answer myself. Although someone posted a fine diagram indicating which protections were retained in Alternate Law and which were lost, I do not recall them exactly. Safetypee you said you believe that stall protection is lost in ALTN Law but what I want to know is: Can the pilot go to Direct Law and remove all “protections” if he chooses to do so? From what I've read, stall protection is lost but stall warning is available from the AOA source.

I ask this question because the entire concept of the automated control system seems to be that it will prevent undesirable positions/attitudes, speeds, etc., as long as we are not in Direct Law. If we carry that concept to its limits - upsets, stalls and overspeeds simply can’t occur – so there is no need to waste time or money training for recovery from them. I do not believe that to be truth.

Every effort and tons of money have been made and spent to keep the Space Shuttle safe yet the vehicle has been lost twice. We build things that we say are ‘fail safe” but nothing is truly fail proof.

While there is nothing at all wrong with fully automated flight control systems per se - they are the wave of the future - if this thread is any indication, there seems to be precious little precise understanding or solid interface between the pilots and the current systems.

As an old timer, I subscribe to Murphy’s Law: “Anything that can go wrong, will go wrong.” While I much appreciate the advances in FCS technology applied to airline transports, we have not reached a level of technology that is “fool proof” and we will not in my lifetime. Systems do malfunction or fail. This is not an Airbus problem nor is it limited to airplanes. It is systemic to all high-tech and low-tech man made systems. If I am correct then, in the case of transport aircraft, I concur with the findings of IFALPA in their June 2009 white paper. If you’re interested, you can read the complete paper here:

About IFALPA

Sorry, I digress; back to stalls.

M(F)S – I basically agree with what you said in your first paragraph, with some qualifications. a) I’m not considering deep stalls of the type that occur in T-tail designs – they don’t apply. b) I’m fairly certain that the A330 is not susceptible to them; otherwise it would have a stick pusher; unless my information is incorrect – it does not. It follows I believe that there is enough elevator authority at all times (where the pilot is controlling the aircraft) to both recover from a stall or to keep the aircraft in a fully stalled state. [I can’t envision why the pilot would want to do the latter in scheduled operations. It does not compute.]

The one exception that I see to that would be stabilizer trim, which could so reduce elevator authority as to render it barely useful. I do not envision that as being related to AF447. Reason: the aircraft was in level flight and trimmed for cruise somewhere around M 0.8 or 0.82. - just before the event. I see stabilizer trim as a factor in low speed flight and a scenario in which the AP is holding an altitude or a glide path and A/THR is disengaged and in flight idle. I think that is what you are refering to safetypee re the UK A320 and the 737 AMS accident/incident. If a stall is approached in that mode, when the AP disconnects and power is applied there will be a severe pitch up that the elevator may not have the authority to overcome in a timely manner unless and until the trim is removed by the pilot.

Like you I do not have specific knowledge of A330 aerodynamics at high AOA but I presume them to be quite conventional. Again, this is an airliner, not a fighter, and they all have to meet the same general certification criteria. There will probably be some device(s) to regulate spanwise flow and guarantee that roll control is retained up to and in a stall (even if reduced). Rudder to control skids and slips will also be available if it is not limited by lockout through the limiter and the FCS.

Of course if during an upset the pilot elects to trim or the automatic trim continues to function and to input trim change, it could indeed be possible to create a situation where there is not enough elevator authority to prevent a stall, or on the other side to effect a recovery from a dive. I doubt that’s a factor in this accident. Since I don’t know how much artificial feel is present in the A330 I can’t guess as to whether the pilot would decide to trim during and upset. I also don’t know at what point automatic trim stops functioning. Is it available without the AP?

Now we come to the meat and potatoes of your “stable stall” theory.

If the pilot is intentionally stalling the aircraft he will be hand flying, not the AP – speed will be reduced with back pressure at a steady rate, in flight idle, until buffet ensues or the shaker activates. If he then wishes to reach a full stall he will continue to increase back pressure until it happens. At that point, under normal conditions, the nose will drop. [If that didn’t happen, the aircraft would be considered ‘unstable’ and fitted with a stability augmentation device to restore the minimum stability required for the certification category.]

The pilot can continue to apply full back pressure and the wing will remain stalled. A high sink rate will ensue and AOA will increase further. However, when the pilot runs out of back pressure (as the speed and elevator effectiveness decay),, the nose will drop again and the aircraft will begin to oscillate in pitch. [Whether side stick or yoke, there is only so much physical movement available – and only so much elevator authority.By design that is not normally allowed to exceed required inherent stability] Aircraft designs change but the laws of Physics do not. In the transport category (as in others) there are established requirements that have to be met. I am quite sure that all ABI transports meet or exceed those requirements. At this point in time what you define as a “stable stall” will end. When I say that the nose will “drop” that does not necessarily mean that it will go below the horizon (although it probably will). It does mean that the body angle or pitch attitude will decrease significantly.

During these oscillations one wing will certainly become more stalled that the other in short order and the aircraft will begin to roll in the direction of the wing that is more stalled. When up aileron/spoiler is applied on the high wing there will be a degree of yaw, which will tend to induce rotation. If this continues roll/bank angles will get progressively higher with each oscillation and the pilot will always be just a little behind the aircraft [he can’t really anticipate control inputs – he merely reacts to attitude changes that are already in progress.] Eventually he will be using full aileron back and forth and bank angles are quite likely to exceed 90 degrees eventually. He will also be using whatever rudder is available and will also be behind in its application. If there is enough altitude available and the pilot never releases the back pressure, the aircraft will enter a spin. If full rudder is available and used, it is highly probable that the VS may fail before the normal spin occurs [side loads will get pretty high with continual erratic control inputs]. If that happens the spin is likely to be “flat” since the VS is the only thing that is preventing rotation about the cg. Should the fin fail, one or more engines will most probably leave the airframe – due to severe yaw.

This url will take you to a scenario that occurred in a DC8-63. It demonstrates what I have said. They didn’t start with enough altitude to actually spin.

http://www.apstraining.com/article9_fci_training_aug03.htm

It is my belief that in line operations – no pilot is going to intentionally stall his aircraft - or if it stalls accidently attempt to keep it in what you call a “stable stall”. The stall will very rapidly become “unstable” in the extreme in any case.

In my opinion, this negates the idea that AF447 could have descended 35,000 ft in a prolonged wings level – body level – flight attitude – with a vertical trajectory - still in one piece. I just can’t buy that. If my analysis is even close to correct and the aircraft did descend – wings level body level – until impact, something else had to make that happen.

Whatever happened to them [that caused the upset] happened very fast and didn’t last very long before things were completely out of control. Most probably seconds – at best a few minutes.

At this point I don’t buy the idea that the anomalies reflected in the ACARS messages were the cause of a sudden upset. If they did cause a complete and prolonged loss of control at all, the aircraft would likely have descended in a steep spiral at very high speed or a normal spin –stalled but at relatively low speed. Either one of those would seem to preclude contact with the ocean in a “level flight attitude” or anything close to it. However, a ‘flat spin’ could produce a “level” impact of the remaining structure. I don’t believe the a/c would flat spin unless there was some structural failure at altitude. That’s not fact, it is just my opinion.

At their altitude they were not very much above the stall speed for starters, and not much below overspeed either. If they got the shaker it is quite probable that they would have added power if they could. How much is only a guess. Intense vertical gusts could prevent that. An extreme up draft could have rapidly produced an altitude increase of severl thousand feet - placing the aircraft outside of the envelope. Follow that with an equally violent downdraft and repeat the sequence several times. I'll leave the "g" to your imagination - but I respectfully submit they are going to be way in excess of +2.5 or -1.5.

If they got a high speed warning they most probably would have reduced power. With normal law – AP and A/THR engaged, if I understand correctly, those things would have happened automatically as the FCS tried to keep the aircraft within the flight envelope. The aircraft would not have stalled or experienced an overspeed – if all of that worked as intended. For what ever reason the automatic stuff wasn’t playing by the rules that night.

Extreme turbulence would most probably exceed by far the limits of the AP – causing it to disconnect. IMO, it could also exceed the parameters of the automated FCS – which would probably trigger multiple failure modes and warnings. I don't know that but it seems logical.

Everything has its parameters and limits. When uncontrollable forces exceed those limits by wide margins, things begin to fail.

I’m not suggesting that there is anything wrong with its design – I am suggesting that whatever happened very rapidly exceeded its limits. In the process multiple errors and failures occurred – triggering the automated messages. In other words I am suggesting that the system went berserk after something extreme and unusual happened to the aircraft – not before it did.

As soon as the ACARS messages were leaked everyone began to focus on the messages and to assume that whatever their meaning (we now know pretty much) they must have been the cause of what ever happened to AF447 – the trigger. Most of us remembered the Qantas and Air Caraibes incidents and looked for ways to make the messages fit the cause of the accident. I was on that bandwagon as well.

Then I looked at the excellent wx analysis of Mr. Vasquez a few more times and began to wonder. I’ve flown in that part of the world. Back then we didn’t have satellites to show us pretty pictures. We had radar in the airplane but it wasn’t in living color, and yes, my airplane was the badest jet in town at the time. I had a 3-man cockpit crew – all very experienced.

Well, one dark night I inadvertently flew my airplane into a cloud. We knew the weather would be bumpy, it always was on that route, but we didn’t expect what happened next and we never saw it on the tube. There was no lightning in our immediate vicinity. It was at night and very dark outside. The canary flew into the fan and the result was feathers everywhere all at once. It got my attention and I will never forget that experience. I’m sure there are others on this forum with similar experiences. That is why old timers are afraid of Cb’s.

We didn’t upset or lose control but at times that really wasn’t because of anything that we did. I couldn’t begin to tell you how we managed to keep it right side up for the truth is the airplane went wherever it wanted to go, whenever it felt like it, and there was nothing we could do about it. Altitude gains and losses were not hundreds of feet, they were thousands of feet. The violence was more than I’d ever thought could exist. Most instruments were unreadable due to the vibrations but we didn’t lose the old steam gauges, they just weren’t very useful at all. We never even tried to change the power since it was already set for turbulence penetration – if we had wanted to I doubt we could have since keeping a hand on the throttles would have been virtually impossible. We just hung on for dear life and had our individual conversations with the Great Spirit. Somehow the airplane managed to make it through. The entire experience was only about 4 minutes long but it seemed like eternity (no pun intended). Since I’m writing this you know the airplane stayed in one piece but I have no idea how that happened. Were we scared? Yeah, you could say that; never more before or since. I vowed I would do anything to avoid a repeat of that experience. And you know what - It never did happen again. By coincidence we were over Brasil at the time, in cruise, on the way to Sao Paulo. In the future I got yelled at a couple of times for what the “suits” called ‘unnecessary diversions’ but you know what – my kids still had a dad.

If the crew inadvertently flew into a deadly cell it does not mean that they were negligent and it does not mean that they were not deviating for the weather they could see. Sometimes those things just happen.

With respect to the BEA report I have some opinions too. First of all it is a Preliminary Report, nothing more. There are lots of “interested parties” on those Boards and they are not going to agree to any “release” that might be interpreted as putting them in a ‘bad light’ when the evidence is not definitive. They will seek to include anything they can that would reduce their individual exposure.

A “break up” in flight would not be a good omen for ABI. It would reflect adversely on AF’ operational controls, it might be adverse to the flight crew and imply pilot error. Each of those “interests” will protect its own. The last thing they need is images of parts coming off the airplane in flight questioning its structural integrity or suggesting pilot error (the union won’t sit still for that).

That does not speak ill of the BEA or any of those entities. It happens everywhere. Boeing will fight tooth and nail to eliminate anything in a report that might adversely reflect on its designs. Remember what they did re: the rudder hard-overs in the 737? It took many dentists and more than one fatal accident to get them to finally admit to a flaw in the rudder PCU. Anything that suggests that the automatic systems of the ABI design failed or were unable to do their job will be resisted until there is no way out. Find a reason for that floating VS – doesn’t matter what it is – as long as you don’t imply that it may have come off in flight. That’s all very natural. The Board came up with a reason. It’s only preliminary and they can change it later if necessary. Don't read too much into a Preliminary Report.

But, somebody tell me how you can determine the failure mode of the VS by examining photos? How do you know the entire tail didn’t come off? You can’t – it will take long-term, in-depth engineering analysis and maybe even recovery of the tail section (very unlikely). They didn’t have that when they issued the Preliminary Report, just as they didn’t have the full pathology on the recovered victims of the tragedy. Those things take time – but in 24/7 hungry media you have to say something, so they did.

As the investigation proceeds, interested parties will continue to protect their interests to the extent possible. They always do in all investigations. If the recorders are recovered there will be hard evidence and the truth will eventually come out for the most part. That is how the process works and we will just have to wait. It may well take a couple years or more before we have a final report. The media frenzy will have long ended and they’ll put it on page 16 – if the bother to publish it at all. The public span of attention is very narrow.

We don’t know what happened gentlemen – but do not rule out a sudden encounter with extreme turbulence. The forces of nature are far greater than those of men and they are NOT fully predictable. There is nothing that we put together that nature is incapable of disassembling – in any variety of modes or parts.