Announcement

Collapse
No announcement yet.

Air France 447 - On topic only!

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

  • Originally posted by EconomyClass View Post
    I'm wondering what significance there is at this time, more than a year and a half after the event, that a court is officially beginning an investigation. (...)

    Hmmmm... well, I guess the significance is that a court is beginning a criminal investigation a year and a half after the event.

    Comment


    • Originally posted by Fear_of_Flying View Post
      I guess for me the question I would like to have answered by pilots on this forum is how controllable do they think the plane would have been given events as we know them (the loss of speed data followed by the other failures at nighttime, etc.). An Airbus pilot would have the greatest insight here, of course, but to me the answer makes all the difference in understanding which factors may have been most important. "Yes, the pilots should have been able to manage the aircraft despite the loss of systems" versus "Controlling this plane would have been exceptionally difficult even for the best pilots" would be very telling to me. Surely this has now been tried repeatedly in simulators. I wonder how pilots have done.
      Well, since nobody has taken this lead, I will risk my opinion.

      The answer is the highlighted one above.
      Of curse, this assumes two things:
      1- That there weren't other failures beyond unreliable airspeed and the related cascade. For example, I'm assuming that they had working attitude indicators (and that there was no ball lighting that turned the computers crazy).
      2- That the pilots knew and were trained to deal with this kind of emergency, and do it accordingly.

      My opinion is based on several things:

      First, there had been several unreliable speed events before and since this accidents in A330s/A340s, and all were survived.

      Second, the procedure itslef doesn't look difficult, it goes pore or less like this:

      1- Establish a prescribed pitch attitude and thrust setting. This is a memory item which, while it won't hold the latitude (it will probably make it climb a bit), it will keep the airplen within it's flight envelope and prevent it from falling from the sky.
      2- Based on the altitude and airplane weight, look up on a table the fine-tunned pitch attitude and thrust setting that will hold the altitude.

      Ok, flying within stormy clouds at night in the middle of nowhere with turbulence and the like won't help, but it still holding an attitude and a thrust setting, and I'm quite sure that if the prescribed attitude ios held "on average" while the plane oscilates around it due to the turbulence, it will still work well enough.

      Now, like I've similary said for stalls a couple of pages earlier in this thread, Who sais that a controllable situation will be controlled?

      Regarthing this, I find amazing and very worrying the investigation finding that Evan showed us: that in all other unreliable speed events, the pilots DID NOT apply the prescribed procedures. So Ok, maybe if you are in a situation with no or little turbulence, holding the attitude and thrust setting that kept the airplane flying before the unreliable speed, will keep the aiprlane flying after it too, and so not applying the procedure is not so critical (but still unforgivable). Now, in turbulence, where the A/P and A/T (now disconnected) was constantly adjusting the attiude, the elevator, the trim and the thrust in an effort to keep the airplane flying straight and level against the will of the turbulence, might have left the attitude, speed, trim and thrust in a setting that were ok to correct the upset it was fighting when they disconnected, but that are not what the airplane needs on average to keep flying safely.

      Evan again gave an example: Suppose that the plane was hit by a stron headwind, the airspeed increased and the A/T retarded the throttles to loose airspeed, now the A/T disconnects. If the pilot's don't add thrust, the plane will keep loosing speed.

      So, since the situation was not controlled, there are three possibilities here:
      1- There were other failures. Assumption supposition 1 is wrong.
      2- The pilos did not what it took to keep the plane under control, either because lack of knowledge/training or because other factors (maybe human factos) kicked in, (for example, panic). Assumption 2 is wrong.
      3- All my analisys is wrong, and the situation was very hard to control even for the best pilots applying the proper procedures.

      I forecast that training will be a big issue in the final report.

      Finally, for what it's worth (if anything), I will humbly share my experience with in an incident that involved lack of reliable speed, and reliable anything else for the matter.

      I was my first instruction flight at night when the instruccotr issued a simulated "total electrical failure and you forgot the flashlight" emergency. What he did was simply to switch the instrument pánel light off and to forbid the use of anything fed by elctricity. Of course I "lost" all com and nav radios, transponder and landing lights. But the darkness also meant that I had lost every other instrument, just because I could not see and read it. I lost airspeed, altitude, climb rate, attitude, RPM and other engine gauges, the "stick and ball" and even the clock and stopwatch.

      On one hand:
      - I lost no only airspeed but everything else and was left only with the noise of the wind, the sound of the engine, the sight through the window and the seat of the pants as my only refernces.
      - I was (and still am) a low time PPL.
      - I had never trained this emergency. In fact, I had never discussed it with the instructor, or seen a procedure about how to handle it. C'm on, I had never enven thought of it!
      - And I managed to navigate back to the airport, apporach and land.
      - And it wasn't hard! My instant procedure was be conservative by keeping the plane a little higher and a little faster than what it felt normal, the runway was long anyway.

      On the other hand, I concede that:
      - A Piper Tomahawk flying at 1000ft and 90kts is not an A330 flying at FL330 and M 0.82 (the A330 should have had that covered with propperly trained pilots).
      - It was night, but it was a nice clear night, with good visibility and no turbulence. The sky full of stars give a good idea of attitude, and you can easily see when you are pitching up or rolling (the A330 should have had that covered with a good pair of big-screen working attitude indicators and instrument rated pilots).
      - And I was not in the middle of nowhere. In fact, I was around populated areas so I had a lot of ground lights to see out of the window, including the airport lights that was close (the A330 should have had that covered with a working navigation system).

      --- 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


      • What of the effects of altitude on stall speed?
        Live, from a grassy knoll somewhere near you.

        Comment


        • Originally posted by guamainiac View Post
          What of the effects of altitude on stall speed?
          None directly.

          Indirectly:
          - At higher Mach numbers you stall at lower IAS, and you fly at higher Mach numbers up there.
          - At a constnat IAS the Mach number increases with altitude so you approach a poin, known as coffin corner, where too slow IAS and too high Mach number can meet at the same speed (you can simultaneously stall and overspeed).

          None of those should have been very a too big factor.

          --- 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


          • None of those should have been very a too big factor.
            I would suggest the altitude played a very large factor by the significant reduction in manoeuvre margin, albeit not to the point that high speed and low speed problems occur simultaneously. What are the stats on controlled airspeed problems down low compared to at altitude?

            Lets consider for a minute what the pilot is faced with. He is sitting, at night, in a cockpit possibly avoiding weather, and in IMC. It is a dark night, with no visibile horizon anyway. Suddenly, a cascade of warnings and loud noises commence - messages start scrolling through the screen, and simultaneously the stick shaker and stall warning go off. The autopilot and autothrottle disconnect, and due to some light turbulence at the time neither is trimmed and the aircraft commences a shallow climb (or descent).

            Now the pilot must first actually identify what has gone on. Is his attitude indicator working? Airspeed indicator? If so, which one on the flight deck is accurate? Are the thrust level indicators accurate? This is a particularly quiet aircraft with limited ability to get thrust noise feedback, and so relies on the instrumentation.

            I'm just trying to build a picture so we can understand exactly how difficult it is to make accurate assessments in these situations. You do not sit there, and get one message saying "airspeed problem"... you get a barrage of messages, often contradictory and seemingly irrelevant, drawing your attention and confusing you. To just say "oh, he's had an airspeed failure and should set x and y" is simplistic and doesn't fully take into account the complexities of the environment.

            Anyway, our smart pilot realises what is going on and attempts to set a given thrust and attitude value, despite still being told he is both stalling and overspeeding. However, the prescribed value is only a ballpark one, and in the 15 seconds it has taken him to realise what is happening, and set the value, his speed has already increased/decreased by 15kts (it would take less than 1deg pitch change to increase at over 1kt per sec... this involves moving the symbol about 5mm from the perfect position), and given the altitude involved he is now approaching a genuine stall/overspeed.

            Lets also not forget that these aircraft are NOT designed to be hand flown at altitude. Yes, you can do it, of course, but the control surfaces are so sensitive at altitude that it is difficult to precisely control. That is why RVSM airspace requires autopilots.

            I'm just trying to ensure everyone is aware that while the *theory* of airspeed failure management is sound, in practice it is not the easiest thing to do.

            Comment


            • Originally posted by MCM View Post
              I would suggest the altitude played a very large factor by the significant reduction in manoeuvre margin, albeit not to the point that high speed and low speed problems occur simultaneously.
              I didn't say it wasn't. I've said it shouldn't have been.
              Of course margins are reduced, but they are still wide.

              You'll get an overspeed when you exceed Mmo, which is M 0.86 (about 530 KTAS), but the plane is demonstrated to be good to fly up to Md, which is M 0.93 (about 575 KTAS). That alone gives you about 45 KTAS of margin, and with all probabilities the plane was flying quite below Mmo to begin with.

              On the other hand, the plane was probably flying with an attitude of just a couple of degrees nose up (typical attitude at cruise), so the margin above the stall was probably not too small either.

              IIRC, the plane was at FL330, and the service ceiling of the A330 is 41000ft (again IIRC), so it was not that close to the service ceiling either, where you'd be closer to the coffin corner (but not that close anyway).

              What are the stats on controlled airspeed problems down low compared to at altitude?
              I don't know, but I'd guess that it'd be more frequent at low altitudes, particularly during approach.

              ... messages start scrolling through the screen...
              "Unreliable airspeed" (or something to that effect) being one of them and remaining high in the list.

              ... and simultaneously the stick shaker and stall warning go off
              Why would that happen? In particular in this case, I think we have no evidence that that happened, and since the AoA vane was still working properly (I assume) I see no reason for a stall warning. I also see no reason for an overspeed warning.

              Now the pilot must first actually identify what has gone on.
              I agree. Finding out what's wrong would be the hardest part. But what would the pilot do in the meantime? Someone MUST be avaiating while the other one is troubleshooting. I don't know what "aviating" would be in this case, though.

              I'm just trying to build a picture so we can understand exactly how difficult it is to make accurate assessments in these situations. You do not sit there, and get one message saying "airspeed problem"... you get a barrage of messages, often contradictory and seemingly irrelevant, drawing your attention and confusing you.
              Again, I agree. And the human-machine interface of new complex and sophisticated airplanes, especially during abnormal situations, is something of concern to me, as I've already mentioned in another thread. If the CVR is ever recovered, it won't be surprising to find a "Que fait-il maintenant?" (What is it doing now?)

              Anyway, our smart pilot realises what is going on and attempts to set a given thrust and attitude value, despite still being told he is both stalling and overspeeding. However, the prescribed value is only a ballpark one, and in the 15 seconds it has taken him to realise what is happening, and set the value, his speed has already increased/decreased by 15kts (it would take less than 1deg pitch change to increase at over 1kt per sec... this involves moving the symbol about 5mm from the perfect position), and given the altitude involved he is now approaching a genuine stall/overspeed.
              And the problem with that is...?

              I'm not expecting them to keep the plane within +/- 100ft and +/- 5kts. +/- 2000ft and +/- 25kts is perfect for me.

              I'm just trying to ensure everyone is aware that while the *theory* of airspeed failure management is sound, in practice it is not the easiest thing to do.
              Sorry. It wasn't my intention to say it was easy. I just intended to say that they should have been able to control the plane. It's not so easy, ok, but it should be doable without exceptional skill, just with the skill of a properly trained and rated pilot.

              --- 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


              • G'day Gabriel -

                The bulk of that message wasn't specifically directed at you - just a general comment - but thanks for your answers.

                The service ceiling is all good and well, but at the weight the aircraft was likely flying at it would never have had suitable performance ability to climb anywhere near that. It may have well been very close to its maximum altitude for its cruise weight. Also, in cruise it is common to sit quite near MMO - while I do not fly an airbus, in the boeing it is not uncommon to sit within .03 or so of MMO - certainly not a long way, and it can be oversped within seconds.

                I mentioned the warnings because many unreliable airspeed events/accidents have featured both overspeed and stall warnings simultaneously. The Boeing version of the checklists include notes that "N1, Altitude and VSI information may be unreliable". It is very hard to maintain a power setting and pitch attitude if you have unreliable N1 indications! Again - have you accurately identified what has failed? Don't forget that this followed an unreliable pitch problem on an A330 (QF72), and that would certainly have a pilot questioning exactly what has failed.

                It also mentiones that you may get overspeed indications, as well as "simultaneous overspeed and stall warnings". Depending on how the airspeed system fails, if it reads too low or high, you will get one or the other. If both fail, you might get both. You may also be actually in one situation, while the airspeed has failed in the other.

                To give you an indications of the messages, Boeing list 10 messages that can be related to this kind of failure - and that is JUST for the airspeed problem. It doesn't include ones for the Airbus including flight laws, and any other system failures that occur at the same time.

                Unreliable Airspeed in IMC at night would be one of the more difficult things a pilot has to contend with.

                Comment


                • Originally posted by MCM View Post
                  I mentioned the warnings because many unreliable airspeed events/accidents have featured both overspeed and stall warnings simultaneously. The Boeing version of the checklists include notes that "N1, Altitude and VSI information may be unreliable". It is very hard to maintain a power setting and pitch attitude if you have unreliable N1 indications!
                  Why does unreliable airspeed affect RPM indications?

                  Comment


                  • Originally posted by Evan View Post
                    Why does unreliable airspeed affect RPM indications?
                    Since airspeed affects N1 itself, I can see why unreliable airspeed might screw up the engine controller (or some equivalent thereof).

                    Comment


                    • MCM I really appreciate your contributions.

                      Comment


                      • Originally posted by Deadstick View Post
                        MCM I really appreciate your contributions.
                        Indeed. Me too.
                        Thanks MCM.

                        --- 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 Curtis Malone View Post
                          Since airspeed affects N1 itself, I can see why unreliable airspeed might screw up the engine controller (or some equivalent thereof).
                          I also fail to see why an unreliable airspeed would affect the N1 indication, which should be based on some encoder or similar actually sensing number of turns of the fan per unit of time.

                          Airspeed may affect N1 and may screw up the engine controller, but the N1 indication?

                          --- 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 also fail to see why an unreliable airspeed would affect the N1 indication, which should be based on some encoder or similar actually sensing number of turns of the fan per unit of time.

                            Airspeed may affect N1 and may screw up the engine controller, but the N1 indication?
                            Well, Boeing seems to think it can have an effect on indication, and they build airplanes, so I'm inclined to believe them, but that's just me.

                            I suppose getting to the bottom of this would require knowledge of the intricacies of the indication system, such as whether or not the signal from the sensor (or encoder, or whatever) gets somehow amplified or corrected before it gets to the DCU (or something similar) before it gets to the EICAS screen. Unreliable airspeed indication may have an effect on said amplification/correction.

                            I, for one, rather doubt that this is just a matter of a simple optical sensor counting the number blades passing by it.

                            Comment


                            • Originally posted by Curtis Malone View Post
                              Well, Boeing seems to think it can have an effect on indication, and they build airplanes, so I'm inclined to believe them, but that's just me.
                              I'm also inclined to believe MCM. My question was not if, but why. I always thought N1 was simply an RPM measurement, independent of anything else. Perhaps I'm wrong there.

                              I don't want to appear to have missed the larger point MCM was making though. It is instructive to note that, in the previously reported cases of UAS, controlling the plane was described as being very challenging to veteran pilots, with the distraction of cascading failures impairing situational awareness. I remember reading a first-hand account on pprune posted by an 320 pilot, testifying to the unnerving difficulty involved in turbulent conditions. A lot can be learned (and has been learned) by interviewing these pilots.

                              The fleetwide pitot refit was a good move. I would really like to see a similar fleetwide upgrade to the BUSS option (standard on the A380). This at least gives pilots an instrument needed to fly within a safe airspeed range while they are working the problem. Thus far, I've seen no recommendation of this.

                              Comment


                              • G'day All,

                                Spot the deliberate error?

                                A closer look probably answers your questions, and reveals that the actual N1 indication should indeed be accurate. A case of RTFQ on my behalf I'm afraid!

                                It should in fact be that the calculated data will be inoperative - ie the Limit, Reference and Bug N1 values - not the N1 data itself which should be valid.

                                Good to see you are all reading and analysing what I say though

                                Comment

                                Working...
                                X