Announcement

Collapse
No announcement yet.

Air France 447 - On topic only!

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Originally posted by elaw View Post
    I actually disagree. It may be disorientation, but not spatial in the classic "non-instrument-rated PPL flies into clouds and freaks and crashes because the horizon disappeared" sense.
    We probably agree more than we disagree....

    You throw the word "classic" in, and I cannot argue.

    Also, they were aware of their roll status, which may be one of the bigger things in PPL disorientation.

    So Elaw 2, 3BS 1

    However, I think my angle is that the horizon on the flat screen basically did disappear (or maybe it was still there, but with the computer going nuts and giving a strong indication that lots of bad data are coming in, can you trust the AI???)

    So- it's not exactly classic, but you loose your primary attitude indicator (Windshield for the PPL, and Flat Screen for the Bus driver), and then you get overwhelmed with an unfamiliar situation and YOUR MIND lacks the proper data inputs (although the data may be available).

    The PPL needs to know to look at the mechanical AI and forget the GA (Gluteal accelerometer). The AF trio needed to know to look at the mechanical AI (if there was one) and forget the "Fly-the-Big-Plane App (With Genuine AI emulator)" on the high-dollar I-Pad.

    I guess I'm saying that I see a lot of parallels- especially as I'm realizing that three guys were unable to figure out just what the attitude situation was because they mentally lacked performance data and were overwhelmed with a strange situation...not perfectly the same, but some similar mechanisms.
    Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

    Comment


    • Originally posted by elaw View Post
      they were flying based on instrument indications (once the autopilot decided not to), but the instrument indications were confusing and conflicting, and they didn't know which ones to believe.
      Between the point where the AP/AT gave up and the point, almost 2 minutes later, stall warning stopped (after some 30 seconds of uninterrupted shouting) due to a low indicated airspeed limitation: Can you please tell me exactly what instrument indications where contradictory and confusing, other than the airspeed that they knew about 2 seconds after the AP gave up when they announced "we've lost the speeds"?

      Let me remind you what GOOD, clear and non-confusing / non-contradictory information they had:
      3 perfectly good and agreeing attitude indicators.
      3 perfectly good and agreeing altimeters.
      2 perfectly good and agreeing vertical speed indicators
      All the engine instruments.
      The stall warning.

      Even the speed indication was recovered somewhere in the beginning of the stall, and started to show good, agreeing and real values, before it started to fail again once in the deep stall, this time not because of the ice but because the pitots are simply not designed to work at an angle of 40°.

      They pulled up, the G increased, the attitude indicators showed the nose going up, the VSIs went up to 7000 fpm, the altimeter started to wind up, then they pushed down, the nose went down on the attitude indicators, the VSIs went to zero and the altimeters leveled 2500 ft higher than where they had started and too high to sustain flight, so they pulling up to hold the altitude, the nose kept going up in the attitude indicator, the stall warning sounded, they added full thrust (of which there was not much to add at that altitude) and pulled full up in response to that warning, the nose went up in the attitude indicator, the VSI increased again, the altitude started to increase again and they gained another 1000ft, the stall warning kept sounding, they kept pulling up, the plane started to descend in a stall, they recovered one by one all the 3 speed indications, that all showed a real very low airspeed, the stall warning kept sounding, they kept pulling up, the stall went deeper and deeper, and then we reach the point where the airspeed indication failed again due to excessive AoA and the stall warning started to be inconsistent.

      I agree that this was no spatial disorientation. They were able to keep the wings level (on average) despite the mushy controls and the plane being highly unstable in roll, they managed to keep the nose up as they wanted. This was a "how a wing is flown(*)" disorientation.

      (*) This is how Wolfgang Langewiesche used in lieu of "how airplanes fly" in his famous 1944 book Stick and Rudder.

      --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
      --- Defend what you say with arguments, not by imposing your credentials ---

      Comment


      • Originally posted by Gabriel View Post
        Let me remind you what GOOD, clear and non-confusing / non-contradictory information they had:
        3 perfectly good and agreeing attitude indicators.
        3 perfectly good and agreeing altimeters.
        2 perfectly good and agreeing vertical speed indicators
        All the engine instruments.
        The stall warning.


        They pulled up, the G increased, the attitude indicators showed the nose going up, the VSIs went up to 7000 fpm, the altimeter started to wind up, then they pushed down, the nose went down on the attitude indicators, the VSIs went to zero and the altimeters leveled 2500 ft higher than where they had started and too high to sustain flight, so they pulling up to hold the altitude, the nose kept going up in the attitude indicator, the stall warning sounded, they added full thrust (of which there was not much to add at that altitude) and pulled full up in response to that warning, the nose went up in the attitude indicator, the VSI increased again, the altitude started to increase again and they gained another 1000ft, the stall warning kept sounding, they kept pulling up, the plane started to descend in a stall, they recovered one by one all the 3 speed indications, that all showed a real very low airspeed, the stall warning kept sounding, they kept pulling up, the stall went deeper and deeper, and then we reach the point where the airspeed indication failed again due to excessive AoA and the stall warning started to be inconsistent.
        Gabe's black font:

        You know that because you read the final report from your arm chair at FL 0.5 and 0 kts with the horizon clearly ahead of you and no 3-dimensional accelerations to confuse you.

        I really don't think THEY knew that...

        ...I hear you that maybe the should have, but it was a dark, stormy night, and all hell broke lose from their computer system.

        They lost their SA and never regained it.

        Gabe's red font:

        I really think we need to forget 'everything' technical and go 'totally' psychological here.

        "Good, clear, non-contradictory" you say (with some all-caps).

        So damn good, clear and non-contradictory that three (repeating three) ATP, type-rated pilots could not figure out WTF it was doing now.

        Don't take my comments as all or nothing and don't take my comments as absolution for the pilots.

        However, I do not think the answer is in what instruments were working and that three stooges ignored them.

        The answer is how three incredibly well-trained guys got confused and overwhelmed and likely distracted on a dark and stormy night when lots of alarms went off...something insidious and psychological.


        That's what we need to figure out (and may never know for sure).
        Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

        Comment


        • Two stooges. Dubois arrived too late, by the end of my description of the events.

          I think that they understood the instruments. They were able to control the airplane at their will following them, of course with the limits imposed by the laws of Physics. They wanted up? Up they went. They wanted to level? They did. They wanted up again? Up they went again, but not for too long because thy stalled. And even during the stall they managed to keep the nose up and the wings level as they wanted (I've been pulling up all the time).

          Now, I don't think that they were simply stupid. I think that they were very very VERY badly trained in manual flight and the basic principles of flight mechanics.

          Of course, startle, confusion, nervousness, maybe even panic, where all factors. Factors that should have not kicked in (not so strongly to leave them useless) if they had had proper training.

          Yet, among all the things that they did wrong, there is one that I really don't understand in any terms: The initial reaction to pull up 1.x Gs and put the plane in a steep 7000fpm climb during about 30 seconds. That's like the Colgan crash. I will never understand them. They were not "burst inputs" caused by a sudden startle. They were sustained wrong inputs over a long time. (although in the Colgan crash I have the excuse that, although he had no good reasons to do so, the pilot thought that he was in the middle of a tail stall, case in which his "recovery" procedure was not the proper one either).

          --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
          --- Defend what you say with arguments, not by imposing your credentials ---

          Comment


          • Originally posted by Gabriel View Post
            I can't prove it, but I feel very confident that the plane would have stabilized at a mild pitch and mild VSI (say, for example, 7 degrees nose-up and -2000 fpm) and then it would have kept going down until the altitude where that 75% N1 was enough to keep the altitude...
            Well, it's all theoretical but we ARE talking about pilots in the cockpit, flying the straight and level as before the UAS, just not watching the actual N1 levels (or donut positions... or FMA... or ECAM...). I think at some point, if the control law ceases to hold flight path angle, the pilots are going to apply increasing pitch... and with no airspeeds to alert them to the upcoming stall warning. That's what I'm getting at. I can't recall if we ever figured out what the stall warning speed would be at FL350 for that flight but I think it was at least 180kts. Could they remain at flight level 350 above 180kts with 75% N1?

            I think you are right about the Vco speed for the C* algorithm being around 250kts although it is a blended transition. But I am assuming pilots in control would override any deviations from level flight, being unaware of the thrust deficiency.

            It is often stated that the C algorithm provides flight path stability. This is based on the fact that, without any pilot stick input the controller tends to stabilize the aircraft at a constant vertical speed or flight path angle assuming that airspeed is kept constant by thrust control. However, the algorithm is not capable of guaranteeing a zero steady-state flight path angle deviation, because it does not include integral control of the flight path angle resulting in flight path deviations due to external disturbances and configuration or thrust changes. As a result, the pilot must periodically re-enter the control loop to keep the airplane on the desired earth referenced flight path.
            But, as I keep trying to point out, FBW simply requires a different understanding of piloting skills. For one thing, as the pilots of AF447 maintained a constant upward pitch in a deep stall, the elevators and stabilizer trim were almost constantly at the upward stop. In a FBW aircraft, it will be more difficult to reduce pitch in this condition due to the behavior of FBW in a situation that is never supposed to occur. Therefore, a forceful downward pitch command is needed where a gentler one might be needed for a conventional aircraft. This is the kind of thing that FBW pilots must understand before you can say that they have "stick-and-rudder" skills.

            So even 'stick-and-rudder' is not a universal skill set.

            Comment


            • Originally posted by Gabriel View Post
              Yet, among all the things that they did wrong, there is one that I really don't understand in any terms: The initial reaction to pull up 1.x Gs and put the plane in a steep 7000fpm climb during about 30 seconds.
              I've have two theories for this. 1) Having failed to turn off the FD's (memory procedure for AP failure) he might have been following an erroneous FD prompt. 2) the blue ECAM message MAX SPEED.......330/M.82 may have been mistaken for an overspeed warning (blue items are actionable). Or may a little of both. Aside from these: 3) basic stoogery.

              Comment


              • Originally posted by Evan View Post
                1) Having failed to turn off the FD's (memory procedure for AP failure) he might have been following an erroneous FD prompt.
                Would the flight directors command a pitch up at all, and such a pitch up in particular, in a case like this? I guess that the vertical mode was alt hold so why would it command 7000fpm climb?

                2) the blue ECAM message MAX SPEED.......330/M.82 may have been mistaken for an overspeed warning (blue items are actionable).
                If that was the case, I would have expected a pull back on the throttles instead (or at least together with) the pull up.

                3) basic stoogery.
                Even then, how would be the bio/psycho mechanics of such a reaction? Just "I don't know what to do so let's pull up like I've never done in my life"?

                --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                --- Defend what you say with arguments, not by imposing your credentials ---

                Comment


                • I do like the way Piper Cubs and 787's respond.

                  ...that natural 'desire' to keep flying and hunt for it's trimmed vertical and forward speeds...

                  ...Set it up fat dumb and happy and it's most likely going to stay fat dumb and happy...

                  ...of course, it requires some moderately strong traditional airmanship skills if you really want to nail an altitude and speed...

                  ...since the plane is always wanting to do it's own correcting thing whenever speed, or power or attitude gets out of whack...Might even oscillate up and down....slower and faster on you.

                  ...Fortunately, you can get an autopilot to help out if you need a break (well maybe not in the Cub)...

                  ...Of course, if the autopilot blue screens on you, you get an airplane back that behaves like...

                  ...an airplane.
                  Les règles de l'aviation de base découragent de longues périodes de dur tirer vers le haut.

                  Comment


                  • Originally posted by Gabriel View Post
                    Let me remind you what GOOD, clear and non-confusing / non-contradictory information they had:
                    3 perfectly good and agreeing attitude indicators.
                    3 perfectly good and agreeing altimeters.
                    2 perfectly good and agreeing vertical speed indicators
                    All the engine instruments.
                    The stall warning.
                    I'm going to avoid the urge to make a social/political "agreement does not equal correctness" joke here...

                    ...but in the cockpit that principle sometimes does apply. Remember the accidents a few years back when static ports had been taped over? The instruments agreed then... they were wrong, but in agreement. It's possible, even likely, that the AF447 pilots remembered those accidents when they were looking at instruments that were giving readings consistent per instrument type but inconsistent between types.

                    Also re the stall warning: it was definitely not GOOD, and non-confusing the whole time. When the airspeed got really low and the stall warning stopped, that information was BAD as the airplane was definitely still stalled. Unfortunately the info it was giving at that time was although wrong, quite clear: it was telling the pilots the plane was not stalled. If I were in their shoes I'd find that confusing.
                    Be alert! America needs more lerts.

                    Eric Law

                    Comment


                    • Originally posted by Gabriel View Post
                      Now, I don't think that they were simply stupid. I think that they were very very VERY badly trained in manual flight and the basic principles of flight mechanics.
                      I'm going to disagree in a subtle way.

                      You've spent some hours flying a light plane, no? I'd bet some money that the AF447 pilots had too, and were reasonably good at it. Then at some point they started flying larger (but not FBW) planes, and presumably were at least "good" at it... it's pretty tough to get an ATP certificate if you can't fly an airplane.

                      What I'm going to toss out is that somewhere along the line in certain key respects they got UNtrained in manual flight. They received training that led them to believe that the machine they were sitting in at the time of the accident did not fly like a normal airplane, due to the FBW and other automation. And when this incident happened that made the automation decide to go take a smoke break, and presented them with some confusing instrument indications, they couldn't throw off that training and switch to pretending they were in a Cub.
                      Be alert! America needs more lerts.

                      Eric Law

                      Comment


                      • Originally posted by Gabriel View Post
                        Would the flight directors command a pitch up at all, and such a pitch up in particular, in a case like this? I guess that the vertical mode was alt hold so why would it command 7000fpm climb?
                        WIthout pitot-static data altitude can be erroneous. There are interactions between the ADIRU and the FD that render the FD unreliable and make it first on the list of things to kill on the UAS checklist.

                        (procedures...)

                        The investigation determined (somehow) that the FD bars on AF447 showed pitch-up commands repeatedly (when they were present) during the time between the loss of AP and the first stickshaker activation. Under the circumstances, pitching up may not have brought the FD bars down and may have resulted in a sustained and aggressive pitch command despite the stall warning.

                        Also, I think the instinctive response to overspeed near cruise ceiling is to add pitch, no? You can't be over-thrusting if your normal N1 there is around 100% and you are in CLB, right?

                        But I'm favoring theory #1. Or #3

                        Comment


                        • Originally posted by elaw View Post
                          I'm going to disagree in a subtle way.

                          You've spent some hours flying a light plane, no? I'd bet some money that the AF447 pilots had too, and were reasonably good at it. Then at some point they started flying larger (but not FBW) planes, and presumably were at least "good" at it... it's pretty tough to get an ATP certificate if you can't fly an airplane.
                          Many (in aviation terms) private pilots are quite bad at understanding flight mechanics. They can fly very nice and be very precise because they have a lot of practice and have the know-how, but not the know-why. You see them quite often (in aviation terms) loosing control of the plane, stalling, spinning and what not.

                          I was quite bad too. There were things that I didn't know, thinks that I didn't know that I didn't know, and things that I knew I knew but I didn't (ok, a few things I had right too). Even when I was training for the instrument ticket (that I never got) things were very "recipe-like", much more than for the PPL.

                          The vast majority of what I've learned (well) about the Physics of flight came from the university, not from the flight instructors or ground school.

                          Unfortunately, we see relatively often (in aviation terms) pilots that crash after botching the "stick and rudder" part (or that fail in using the stick and rudder skill to save the day after a problem cause by another issue). Air France, Colgan, Austral, Pinnacle are just a few that come to mind without thinking. You can have a dozen of them in my "stall" rant-thread, and these were only stall-related.

                          Plus... I agree with the DEtraining issue. Even without brainwashing them with that the automation will save the day, imagine a long-haul pilot flying perhaps 5 flight per week, with 2 minutes of manual flight at take off and 5 minutes at landing. 7 minutes per flight, 5 flights per week, say 45 weeks per year account for 26 hour per year. A weekend pilot will have more manual flight per year than these ATPs.

                          I remember in the 90s an article in Flying that was titled "the hours that count". The author suggested that having a lot of hours in the logbook is not necessarily a synonym of experience. That flying straight and level from A to B doesn't count as experience. And taking of and landing in long, wide paved runways aligned with the smooth breeze counted but not that much. How many hours practicing slow flight, stalls, emergencies, recoveries? How many hours in challenging flight conditions, where you needed to assess and manage risks and decide a course of action? Those are the hours that count. And the article didn't even mention the hours monitoring the autopilot.

                          --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                          --- Defend what you say with arguments, not by imposing your credentials ---

                          Comment


                          • Originally posted by Evan View Post
                            WIthout pitot-static data altitude can be erroneous.
                            Yes, but not without pitot data only. Static was Ok as shown by all three altimeters and VSI showing agreeing (and correct) values.

                            But if the BEA found that the command bars went up, there must be an explanation for that and that can be very well the reason why a child of the magenta blindly followed suit and pulled up.

                            Acknowledged what you say about the procedure. Again, by all means, do follow the procedure. But again, even then, that the pilot didn't turn the FD off as required doesn't mean that he needs to follow the flight director commands if it tell him to roll the plane inverted.

                            Also, I think the instinctive response to overspeed near cruise ceiling is to add pitch, no? You can't be over-thrusting if your normal N1 there is around 100% and you are in CLB, right?
                            I think that the instinctive response for too much speed is reduce thrust.
                            Yes, you cannot overthrust while cruising at your ceiling, but whatever thrust you have is what keeps you cruising at 600 MPH instead of zero, so there is a good bunch of forward force to reduce by pulling back on the thrust levers.

                            But maybe, who knows.

                            The flight director conjecture seems much more plausible to me.

                            --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                            --- Defend what you say with arguments, not by imposing your credentials ---

                            Comment


                            • Guys, this is a 200MB animation video prepared by the BEA that you HAVE to download and watch.


                              [EDIT]
                              Now available in youtube. Watch in HD:
                              Dec 12th 2014 the French BEA have released a 200MB video showing an animation of the accident sequence stating: This animation has been developed by the BEA ...

                              [/EDIT]

                              It's just unbelievable. After watching it I can understand even less the actions of this crew.

                              By now I will just say that, while the flight director displayed some pull-up cues, it also displayed push-down cues, and there is no correlation between the flight director's and the pilot's commands. I don't think that the pilot was in any way misled by the flight director (that by the way he should have turned off by procedure), particularly in the initial crazy take-offish pull up and during the initial response to the stall warning.

                              So pilots my ask "and what is it doing now?", but I will keep asking "and what the hell where they doing?". I have not the slightest idea.

                              Non adherence to the procedure, child of the magenta line, lack of training in UAS at cruise altitude, lack of practice in hand-flight at cruise... All those are good things to work on that might have prevented this accident. But none of them, by themselves or all together, can explain the reaction of these pilots.

                              I am just astonished and in disbelief.

                              --- Judge what is said by the merits of what is said, not by the credentials of who said it. ---
                              --- Defend what you say with arguments, not by imposing your credentials ---

                              Comment


                              • Amazing.

                                Two things that strike me (other than the overall stupidity of what happened):
                                1) Almost from the start, the right stick is showing *large* deflections - often over 50% of max. Maybe a FBW aircraft is different, but in general during high-speed cruise aren't you supposed to make *small* corrections with the controls? Almost from the start it seems like this guy thinks he's flying an Extra or a Stearman in an airshow... not a transport-category jet in commercial service.

                                2) I love the point around 4:07 where the narrative says "...it was very difficult for the captain to make a diagnosis.". Okay kiddies: airspeed varying from 0 to around 80 knots, aircraft "wallowing" for lack of a better term, and a very large constant negative vertical speed. How could you interpret those instrument indications other than to say the a/c is stalled? Not a difficult diagnosis at all IMHO.
                                Be alert! America needs more lerts.

                                Eric Law

                                Comment

                                Working...
                                X