Announcement

Collapse
No announcement yet.

Air France 447 - On topic only!

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

  • Originally posted by Theoddkiwi View Post
    I know this was a while back, but anyway.
    Our fleet get hit by lightening often due to climate that we fly in. Aircraft are well capable of withstanding most lightening strikes. The belief that all the avionics will explode and the aircraft fall apart is crap.

    Electricity will always take the path of least resistance, everything on the aircraft is bonded to the airframe with metal braid straps. What these straps do is ensure that electrical charge from one part of the aircraft can pass easily through the metal structure and skin and out to the static dischargers then out to the atmosphere. The path of least resistance.

    What electrical charge won't do is work down into the airframe and into the electronics boxes which are themselves metal boxes which are also bonded to the airframe.
    Even if a lighting strike charge did try to pass though the avionic bays it will pass around the metal boxes and not through the electronic cards inside as they pose higher electrical resistance.

    Lightening Strikes will almost always cause some surface damage to the external aircraft structure, typically burnt rivet heads and some damage to the skin where it entered and exited.

    It will sometimes cause damage to some structural and composite components and this is often due to less than perfect bonding.

    It will occasionally damage susceptible radio systems typically the HF radio system being most common due the big antenna. But very rarely if ever all communications as each system is independent and often duplicated so the same fault occurring to multiple systems is almost impossible.

    It can very very rarely cause disruption to the Navigation, Flight control and Engine Control systems. But almost never all at the same time.

    Sorry but aircraft are just to well designed to be at any real risk to lightening.

    Lightening Strikes are the most overrated hypothetical cause of aviation accidents and most frequently blamed by the media and witnesses.
    The fact is that there have only been a handful of accidents where a lightening strike led to the loss of an aircraft. I believe the last commercial airliner was a Pan Am 707 in 1968 when fuel vapor ignited and the aircraft exploded.

    Lightening did not cause AF447 to crash nor any other airliner in the last 40 years.
    Well, the Do228 between Röst and Norwegian mainland had a complete blackout due to lightning but landed safely anyway.
    I personally dont rule out terrorism but its highly unlikely.
    "The real CEO of the 787 project is named Potemkin"

    Comment


    • My votes are almos identicla to those of Peter:
      Originally posted by Myndee View Post
      Do you believe the crash was caused by a pitot failure of some type?
      Yes. Not that alone, though.

      Do you believe the crash was caused by a structural failure?
      I have no reason to suspect that at this point.

      Do you believe the crash was caused by pilot error (not configuring the aircraft correctly after the airspeed could not be determined?)
      Yes. This is the other part of answer 1. Airplanes can be flown without working pitots. It's harder, but it can be done.

      However, I never (well, almost) blame the "pilot error" per se. Pilots are human. Humans do err, period. The only way to eliminate the pilot error is to eliminate the human pilot, but we are way far from that and I don't want to get into this discussion. However, this takes us to your next question.

      Do you have some other theory?
      Not another theory really, but the previous theory is incomplete without adding what caused the pilot error and their failure to detect and/or correct such error. But I have more questions than answers here:

      Was the training appropriate? I find it disturbing that, in similar cases in similar planes, the unreliable speed procedure was systematically NOT followed (yet, the pilots managed to get rid with that). So I think it wasn't.

      Is the unreliable airspeed condition reliably recognizable before it's too late? I find it disturbing that, in the Qantas A380 case (blown engine), the pilots spent 40 minutes just in going through the ACARS message and addressing them. Would an "airspeed disagree" be at the top of the dozens of ACARS messages that that condition would trigger? I don't know. I have doubts.

      Once the condition is recognized, is the unreliable speed procedure reliably performable? Was it realistically assessed and trained in a simulator? I mean, I can write a book with a detailed procedure about how to do juggling, and you can study it until you learn it by hart and become so knowledgeable that you are able to teach that procedure to others, but when your time to do juggling comes all that knowledge will not give you the skills, and you won't be able to do any juggling at all. I think that the procedure is doable. It's not like you have to simultaneously grab a lever with your right hand, push a button with your left hand and turn a knob with your other hand. It's all about setting a given thrust and keeping a given pitch attitude (memory values first, and when time permits you fine tune them by looking up more precise values in a table)

      Thanks for your replies.
      You are welcome.

      --- 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 Joe H View Post
        Look at the Colgan issue, both were cockpit sim pilots in reality, and I think the result was because the pilot THOUGHT he was in a tail stall, although he wasn't..
        A small OT comment here. I think that too. But then his reaction was inappropiate even for a tail stall. The taill stall procedure is to pull up enough to recover the 1G level flight, not pitch up at 1.5Gs and 30 deg nose up.

        --- 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 Joe H View Post
          Absolutely Agreed Evan !

          I was using that as an example of the Zero to Hero training,

          You are totally correct, it is not approriate for all and I did not mean it that way, I stand corrected, but the concept is correct, the more time passes, the more that time served pilots in the real sense fall out of the loop, hence the Zero to Hero ! the days of the 'real' pilot who can actually fly an aircraft are numbered, we just now have sim trained button press heroes who will and do panic the moment a crisis hits. and they can do no more, because they were not in charge of a Lancaster Bomber or B17 etc... they are a completly new breed, trained to manage the instruments, and only fly the aircraft in a real shit hit situation, previously, they were true airmen, who were then trained to use the additional instruments.

          Joe.
          Joe, I think the skill set has changed so much since the B-17 days that you really can't compare pilot skill this way. A piston prop transport pilot never had to contend with the sensitivities of transonic flight at high altitude, or swept wing lift/drag issues at approach speeds or thrust lag or mach tuck. He didn't have to place such importance on stabilized approach. All of these things require a partnership between the computer (PF) and the human (PM, PIC). The pilots are still 'real pilots', on T/O and landing they need solid flight skills, but the controls are now more of an interface. Stick-flying is just an obsolete thing in the new realm of ultra-efficient, envelope pushing jet aircraft. Certainly, when the automation fails (see AF447) the pilots need to have quick, rehearsed procedures and a clear head, and when there are no procedures, a deep understanding of both aerodynamics and systems. That training and experience, not stick-and-rudder time, is what defines a good pilot these days. With the A330 for instance, I would feel much safer with a pilot who had immense experience with a high-fidelity, A330 sim than a pilot with immense turboprop stick time (this assumes both also have the minimum actual time on type to be A330 certified). But I also want my PIC to have a lot of experience in the air on type.

          Comment


          • Originally posted by Gabriel View Post
            Is the unreliable airspeed condition reliably recognizable before it's too late? I find it disturbing that, in the Qantas A380 case (blown engine), the pilots spent 40 minutes just in going through the ACARS message and addressing them. Would an "airspeed disagree" be at the top of the dozens of ACARS messages that that condition would trigger? I don't know. I have doubts.
            Well, not to be difficult Gabriel, but I'm pretty sure I was the lone voice on that thread expressing concerns at how long it took the pilots to rummage through the error messages while who-knows-what deteriorating condition was transpiring.

            Part of my post 276:
            ...In light of this report, I now wonder how much time was spent calculating weight/landing distance and everything crucial to getting the plane down, and how much time was spent troubleshooting a huge list of computer error messages which may or may not have had immediate relevance to landing the aircraft safely.
            Part of your reply, the following post:
            What systems are working fine, which one are affected but still working in a degraded mode, which systems are not working, what system feeds what, which sub-systems are affected?

            This is the info needed to answer all of the above.

            Now, where, do you think, would you get that info from if not from trubleshooting the list of "computer error messages"?

            Comment


            • Originally posted by MCM View Post
              We will be far safer when we actually take that step back from total automation to put the human right where he belongs... at the centre of the loop. Give the required training, and ensure the information he needs is provided in the best way possible for human interpretation and awareness.
              I agree with Gabriel that this could lead to a protracted discussion vis a vis the relative merits of automation vs. the human pilot. However, perhaps that is the discussion that will turn out to be most relevant to the Air France accident.

              The logic as I see it is that if increased automation has led to fewer accidents, then the proper course of action is to continue to improve the automation, identify the anomolies, and make aircraft even more foolproof. Putting the pilot back in the middle of the loop to me would be a regressive step. Even though we've seen an increase in accidents caused by glitchy technology and human-computer interface issues in recent years, I would still venture a guess that these make up a small percentage of accidents, while most are still caused by old fashioned human error and bad weather. Given the accidents that automation has probably prevented, we have a significant net gain in safety, and should continue steadfastly in that direction. Don't worry, I wouldn't ever want to be a passenger on an RPV, but the less there is for the pilot to screw up, the better as far as I'm concerned.

              Comment


              • Originally posted by Fear_of_Flying View Post
                Well, not to be difficult Gabriel, but I'm pretty sure I was the lone voice on that thread expressing concerns at how long it took the pilots to rummage through the error messages while who-knows-what deteriorating condition was transpiring.

                Part of my post 276:


                Part of your reply, the following post:
                And? You are worried about the length of the list. I am worried about the order of the items in that list. I don't care if the list has 50 items, but "You have a speed disagree" must be at the top. AP off, AT off, rudder limitter off, alternate law, etc would probably follow in the list. But the pilot, if propperly trained, would already know all that from reading just the first item, and would be already propperly executing the unreliable speed procedure by when he moves to reading the second and following items. Or better, the PF will be executing the unreliable speed procedure while the PNF does the rest of the troubleshooting.

                --- 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 guamainiac View Post
                  Has there ever been developed a list and formula for all of the variables that would list the "relevant" and the "spurious" variables possible.

                  The things TeeVee and Theo have discussed for instance? Airline training, type of oversight by regulators and such?
                  Since the total numbers of a/c losses is not large, you will easily end up with only very few or even no accidents at all in each category when you split up by too many variables. No need to try and do statistics then...

                  Comment


                  • Originally posted by Evan View Post
                    We have learned that, in the 13 other reported instances of UAS studied in the investigation, there was not a single case where the memory items were actually performed. It's reasonable to believe that the crew of AF447 also did not have these memory items properly instilled in their training. But, you have to draw a fine line here between 'pilot error' and 'pilot limitation'. Under the conditions they were probably encountering (turbulence, darkness, loss of airspeed data, loss of autoflight, loss of flight augmentation and a barrage of confusing ECAM messages, all occurring near the limit of their flight envelope), I think we may find that it would have been extremely challenging to avoid errors (by the estimations I've seen (205t, ASI +13) their service ceiling would have been around 37,500ft, before any G-loading or gust-loading).
                    Since I'm clearly one of the non-experts here, would everybody forgive me my possibly naive question...

                    I've now read many times here about the proximity of the coffin corner or its variants as a contributing factor in the difficulty of the situation that AF447 encountered.

                    Now, wouldn't it be a safer procedure - safer than e.g the manual selection of thrust and AoA from tabulated values - to immediately initiate a descend and get deeper into the flight envelope as soon as anything is indicated to not be functioning properly?

                    Would it all have been much less of aproblem - let alone disaster - if the altitude had been different, i.e. lower?

                    m.

                    Comment


                    • Originally posted by Gabriel View Post
                      And?
                      And I'm pretty sure that's part of what I was saying, it was a long, non-prioritized list of information that they apparently slogged through. Critical items at the top sounds like a good idea, so does grouping items (for example if failures 2-4 were a result of failure 1.). But I wouldn't know.

                      Comment


                      • Originally posted by mfeldt View Post
                        ..........................
                        Would it all have been much less of aproblem - let alone disaster - if the altitude had been different, i.e. lower?
                        ................
                        I don't know the answer to your question but flying at a lower altitude increases fuel consumption and flight time plus the weather is often worse at lower altitudes.

                        Comment


                        • I guess since the envelope is a bit more generous for stall and control inputs and such the answer would be yes regarding coffin corner.

                          Move the stick or yoke and induce pressure and you may induce a stall but the same at a lower altitude would result in an increase or decrease of speed but that is unknown regarding the disaster yet. What is more likely at that altitude but the cause is not known? What was the service ceiling and such for that weght?
                          Live, from a grassy knoll somewhere near you.

                          Comment


                          • Originally posted by mfeldt View Post
                            Since I'm clearly one of the non-experts here, would everybody forgive me my possibly naive question...

                            I've now read many times here about the proximity of the coffin corner or its variants as a contributing factor in the difficulty of the situation that AF447 encountered.

                            Now, wouldn't it be a safer procedure - safer than e.g the manual selection of thrust and AoA from tabulated values - to immediately initiate a descend and get deeper into the flight envelope as soon as anything is indicated to not be functioning properly?

                            Would it all have been much less of aproblem - let alone disaster - if the altitude had been different, i.e. lower?

                            m.
                            I think the 'coffin corner' aspect has been overly dramatized. I haven't seen any definitive graph showing the speed margin AF447 would have had at that moment, and I'm not sure if I have this right, but the way I calculate it based on the graphs I have for the A332, 'coffin corner' for AF447 at 205t and ASI +15 would have been above 41,000ft (max operating alt of 37,500 + 1.3g margin), so cruising at 35,000ft should have given a reasonable margin for error. Thrust has much less effect at high altitudes, and maintaining cruise altitude requires high power settings to begin with. I think it is safe to assume that CLB thrust is not going to be enough to overspeed in that thin air while holding 3-5° pitch, and a decent pilot is always going to detect that nose down moment when approaching Vmo. AS long as you keep that pitch up, I think the greater danger lies in underspeed. Descending would mean slowing down. In turbulence, any gust loading would speed up the slow end of the range so it would be safer to stay on the high side. I think this is why the procedure calls for maintaining altitude. And the weather, obviously.

                            How did I do there Gabriel?

                            Comment


                            • I'm certainly not an expert here, but from the research I've done it would appear to be quite tricky to get this exactly right. Simply lowering the nose would increase airspeed and hence the chance of an overspeed. Retarding throttle on the other hand decreases airspeed and hence the chance of a stall. I believe a I read somewhere that they basically had a 6kt "sweetspot" that they had to remain within! While the computers were talking to each other and working properly the automation was able to maintain this, but when the fan became soiled they had to go manual. It's my understanding that this is why the SOP dictates power 85% and a prescribed attitude will hold things as straight and level as possible while the crew can run down what's going on.

                              As I said, I'm not an expert, but if my understanding of this is wrong I'd welcome an explanation from one who is.
                              Yet another AD.com convert!

                              Comment


                              • Originally posted by Evan View Post
                                I think the 'coffin corner' aspect has been overly dramatized.
                                I agree. There is a design speed, caled Vd/Md (not known to the pilot), which in itself has margin over the max speed the plane can fly before suffering destructive flutter or becoming uncontrollable. Not only that, but at Vd/Md the plane must still comply with all the usual controllability and loads requirements.

                                The max allowed speed, Vmo/Mmo, is an oprative speed well below Vd/Md such as, given a certain prescribed upset that puts the plane past Vmo/Mmo, the recovery can be performed without exceptional skills before reaching Vd/Md.

                                Finally, planes cruise say a dozen of knots IAS below Vmo/Mmo, which translates in a couple dozens of knots TAS below Vmo/Mmo, which then is well below Vd/Md, and then there is still a margin. So there is a significant margin between typical cruise speeds and the point where the high speed will be a real problem.

                                Regarding the low speed margin, well, suffices to say that an airplane at cruise flies level with some 2 to 4 degress of nose-up attitude, and that's well below the AoA (= attitude in level flight) needed to stall.

                                All tht said, of course that the margins are higher when flying lower, especially with the high speed limit.

                                (rant about ISA + X, margins with 1.3Gs, etc...) I think it is safe to assume that CLB thrust is not going to be enough to overspeed in that thin air while holding 3-5° pitch
                                I think that if Airbus says to set CLB and 3-5º it's because you are NOT going to put the plane in danger by doing so. Maybe, until you fine-tune the values with the table, you are not to hold the altitude (as I discussed several weeks ago), maybe you will climb or descend a bit, but for sure the plane will remain controllable and within it's operating envelope.

                                and a decent pilot is always going to detect that nose down moment when approaching Vmo.
                                I don't think that you'll detect much of a nose down moment when approaching Vmo. Especially in an Airbus and especially while riding in considerable turbulence.

                                AS long as you keep that pitch up, I think the greater danger lies in underspeed. Descending would mean slowing down.
                                Why?
                                How did I do there Gabriel?
                                I'd say Ok.

                                I'd add that I see another reason to keep the altitude: You fix one variable. When the plane flies level the trajectory is horizontal, so there is a direct relationship between pitch and AoA. And the pilot has an easier feedback of how things are going. If he thinks he is keepinc the prescribed values and finds himseld descending say 3000fpm, something is obviously wrong. Time to double check everything. If a descent was prescribed the expected rate of descent should be also in the table, together with the attitude and thrust settings, and that's another parameter to worry about. We are descending 500fpm: Is it too little? Maybe we are going too fast and are about to overspeed? We are descending 3500fpm: Is it too much? Maybe we are going too slow and about to stall? Zero is zero, easy to remember, easy to understand, easy to monitor.

                                That reminds me something (OT) from my work. We make some kind of products where it's shape is mostly a matter of cosmetic design, not something functional, and sometimes it's difficult to get the exact designed shape. We in production love designs with curves, lines that are not completely horizontal or vertical, angles that are not 90º, and mat finish.

                                Why? Any customer will tell you "hey, this stright line is curved, this angle is 2 degrees off 90º, this side is tilted a bit off the horizontal, this mirror-finished surface has a not-as-bright spot". Nobody will tell you "hey, this curve is supposed to have a radius of 50cm, not 48, this angle is supposed to be 37º, not 39, this side is supposed to be tilted 12º off the horizontal, not 15", and the mat surface doesn't has the exact level of gloss.

                                --- 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

                                Working...
                                X