Its well known that the French investigation is focused on avoiding any allocation of blame. With all the inter-linkages between Airbus, Air France and the French government it is clear that the BEA is actually investigating the French government. The level of disclosure tolerated by the French is abysmally low in comparison to what the NTSB would be revealing.
The above paragraph from an earlier posting on this blog is being repeated in response to some inquiries I've had regarding the recent publication of more complete transcripts in a book that focuses clearly on pilot error and pilot standards.
The French view is that only pertinent conversations get published. The American view is that everything, including the final expletives, get published. It is a question of openness and a dedication to seeing that absolutely nothing is to prevent the deaths that have taken place from being deaths that contribute to future air safety.
There are serious issues as to whether the initial climb was the result of pilot commanded inputs or not. The book fails to address this issue. The fact that irrelevant conversations were taking place is meaningless. The plane was not at an altitude where rules impose a "sterile cockpit" devoid of non-pertinent conversations or activities.
The pilots were suddenly confronted with a situation where the computer suddenly said "your plane". As young pilots with an instructor beside them they've heard that phrase a zillion times but suddenly they are "hearing" it without warning and they are no longer in some small two-seated training aircraft. The computer gave up when limits were exceeded and dumped them into alternate law amidst a flurry of warning messages and loss of instruments coupled with loss of any sense that remaining instruments were or even might be reliable. Its pure hell when a complex system cuts out.
Other, non fatal, incidents indicate a serious likelihood that that initial climb of a few thousand feet was not selected by pilot initiated control inputs. After that the cascade of events proceeds with confusion and unreliable information. The start of the fatal event was that initial climb. Blame the pilots for it? Not yet!!!
Showing posts with label AF447. Show all posts
Showing posts with label AF447. Show all posts
Friday, October 14, 2011
Wednesday, June 8, 2011
AF 447 The real causes...
Sure one can blame the bean counters for throwing up roadblocks to the replacement of the pitot tubes but no one really knows if the new pitot tubes had they been installed would not have iced over also.
One can blame a pilot for taking a rest break but I'd sure prefer an well rested pilot be at the controls during the later and more demanding portions of the flight. It is certainly clear that the pilots in the cockpit for fully qualified and experienced.
The real cause is that the pilots were simply not capable of deciding which instruments to rely on and which to mistrust. The pilots were simply too dependent on a mind set of trusting their instruments and relying on a computer to keep the airplane from stalling. Once it went into Alternate Law they lost Angle of Attack protection.
The airspeed was indicated as very low. They should have rejected that data but didn't. However, that is forgivable. They added power and with the engines slung low under the wings of course they shot up real fast into a nose up position and gained altitude.
The pilot may have become distracted by a non-silenceable stall warning horn. Well, its a critical situation and its probably good that it can't be turned off. The trouble is that any pilot knows to get the nose down instantly and to add power in or near a stall attitude. Yet the pilot kept the nose up and allowed it to go even higher and higher and although power was initially set to TOGA, they quite promptly lowered it to flight idle and although it takes time for a jet engine to spool down, they were soon at fifty-five percent power when they should have stayed at Take Off/Go Around power.
Even when the airspeed suddenly indicated a more comforting value of 215 they were keeping the nose up and lowering power.
Its simple. Pilots who've done nothing but input commands into a computer were unable to suddenly start flying the plane. They didn't know what to accept and what to reject. They didn't know that buffeting in a low speed stall and high speed stall is pretty much the same thing. They didn't know to lower the nose and add power. They were overwhelmed by a flurry of error messages and unreliable instruments when the static air port was iced over and the computer put the nose up but soon detected too great a discrepancy between the data from the two pitot tubes and suddenly turned the plane over to the pilots without warning.
Pilots who deal only with computers and don't learn to actually fly the airplanes become too reliant on instruments and don't know how to actually fly the airplane themselves. They had three minutes to learn how to fly, but they seem to have forgotten the lesson every student pilot learns on his third flight: lower the nose immediately and add full power immediately.
One can blame a pilot for taking a rest break but I'd sure prefer an well rested pilot be at the controls during the later and more demanding portions of the flight. It is certainly clear that the pilots in the cockpit for fully qualified and experienced.
The real cause is that the pilots were simply not capable of deciding which instruments to rely on and which to mistrust. The pilots were simply too dependent on a mind set of trusting their instruments and relying on a computer to keep the airplane from stalling. Once it went into Alternate Law they lost Angle of Attack protection.
The airspeed was indicated as very low. They should have rejected that data but didn't. However, that is forgivable. They added power and with the engines slung low under the wings of course they shot up real fast into a nose up position and gained altitude.
The pilot may have become distracted by a non-silenceable stall warning horn. Well, its a critical situation and its probably good that it can't be turned off. The trouble is that any pilot knows to get the nose down instantly and to add power in or near a stall attitude. Yet the pilot kept the nose up and allowed it to go even higher and higher and although power was initially set to TOGA, they quite promptly lowered it to flight idle and although it takes time for a jet engine to spool down, they were soon at fifty-five percent power when they should have stayed at Take Off/Go Around power.
Even when the airspeed suddenly indicated a more comforting value of 215 they were keeping the nose up and lowering power.
Its simple. Pilots who've done nothing but input commands into a computer were unable to suddenly start flying the plane. They didn't know what to accept and what to reject. They didn't know that buffeting in a low speed stall and high speed stall is pretty much the same thing. They didn't know to lower the nose and add power. They were overwhelmed by a flurry of error messages and unreliable instruments when the static air port was iced over and the computer put the nose up but soon detected too great a discrepancy between the data from the two pitot tubes and suddenly turned the plane over to the pilots without warning.
Pilots who deal only with computers and don't learn to actually fly the airplanes become too reliant on instruments and don't know how to actually fly the airplane themselves. They had three minutes to learn how to fly, but they seem to have forgotten the lesson every student pilot learns on his third flight: lower the nose immediately and add full power immediately.
Thursday, May 26, 2011
AF447 Circumstances but not causes:
The French have announced the information to be released tomorrow will give "Factual elements on the operation of the flight that will establish the circumstances of the accident but not the causes" to correct "The partial and contradictory information published in the media".
I am all in favor of establishing circumstances and certainly think it should be done prior to laying any blame. However, dealing with forty to fifty major alarms in a four minute sequence is inherently dangerous particularly at night near turbulence in a situation that is neither taught nor simulated during the pilot's training. Pilot Error is one term that should often be prefaced with the phrase Design Induced. In the case of AF447 it should be Design Induced and Training Re-inforced Pilot Error.
The French announcement implies that the plane had already entered turbulent air prior to the throttle being reduced and the nose pitching up based on instructions derived from the pitot tube data. This has not necessarily been established. Partial and contradictory information is often the fault of the authorities, not the journalists.
...
What fools these journalists are! The press statement was just more incomplete data with obvious omissions.
Why do journalists keep discussing the fact that the chief pilot was on a rest break. So what? Each of the two pilots in the cockpit was a fully qualified and experienced pilot. The Chief Pilot arrived back in the cockpit about one minute after the event commenced but wisely did not immediately assume command of the aircraft.
Today's statement focused on actions taken by the Pilot Flying but never once compared those actions as to whether they were proper for what the instruments displayed or what was actually happening. A pilot immediately adding full power when he thinks there is a stall is fine and dandy. Student pilots learn that by their third lesson. Yet the pilot simultaneously commanding a nose-up pitch increase? In a stall? Of course the essential point is that a pilot for Air France is only trained on low altitude stalls at low power. Yet any pilot should know to get the nose down not up. Perhaps it was a question of which instruments to believe and which to disregard but soon each pilot agreed that they had "no indications" (of speed). They never escaped a deep stall and it seems never went to a sixty degree bank in an attempt to escape the deep stall. They were doing what they had been trained to do if they were "low and slow" and inadvertently stalled the aircraft. Everything in their training revolves around the FlyByWire system that will not allow the pilot to stall the aircraft but once the wildly divergent pitot sensor data caused the computer to trip out, the plane was in Alternate Law mode and while some protections are in place, stalls are quite possible. Computer control over the stabilizer trim tab is lost upon entering Alternate Law but pilots appear not to have known that which is understandable since they never really train for Alternate Law.
So weather really played no role in this as they made a simple deviation around the weather, it was the non-turbulent icing that was a problem for them. They had erroneous indications of dangerously slow speed and then no reliable indications of speed at all. The initial climb of several thousand feet was due to the added thrust when they went to TOGA power, this also increased the severe nose-up attitude due to the pilot's failure to consider stabilizer settings being now manually controllable but not computer controllable.
The first incident was icing of the pitot and static air sources. Soon thereafter the computer became aware of unreliable data. The problem is that at the point in time wherein the static air source became iced and therefore the airplane's speed supposedly fell from 275 to 60, it is clear that such an event could not have taken place without some sort of deceleration forces having been applied. Prior to sending a series of critical warnings to the pilot, how about a critical warning to the computer: airspeed is clearly erroneous so make no immediate adjustments.
The pilot seeing such an abrupt air speed change should have known it was erroneous and probably related to static air data and the pilot hearing the stall warning should not have assumed that it was already a stalled aircraft. The FlyByWire system had pitched the nose up but not gravely, speed was decaying but not gravely.
I am all in favor of establishing circumstances and certainly think it should be done prior to laying any blame. However, dealing with forty to fifty major alarms in a four minute sequence is inherently dangerous particularly at night near turbulence in a situation that is neither taught nor simulated during the pilot's training. Pilot Error is one term that should often be prefaced with the phrase Design Induced. In the case of AF447 it should be Design Induced and Training Re-inforced Pilot Error.
The French announcement implies that the plane had already entered turbulent air prior to the throttle being reduced and the nose pitching up based on instructions derived from the pitot tube data. This has not necessarily been established. Partial and contradictory information is often the fault of the authorities, not the journalists.
...
What fools these journalists are! The press statement was just more incomplete data with obvious omissions.
Why do journalists keep discussing the fact that the chief pilot was on a rest break. So what? Each of the two pilots in the cockpit was a fully qualified and experienced pilot. The Chief Pilot arrived back in the cockpit about one minute after the event commenced but wisely did not immediately assume command of the aircraft.
Today's statement focused on actions taken by the Pilot Flying but never once compared those actions as to whether they were proper for what the instruments displayed or what was actually happening. A pilot immediately adding full power when he thinks there is a stall is fine and dandy. Student pilots learn that by their third lesson. Yet the pilot simultaneously commanding a nose-up pitch increase? In a stall? Of course the essential point is that a pilot for Air France is only trained on low altitude stalls at low power. Yet any pilot should know to get the nose down not up. Perhaps it was a question of which instruments to believe and which to disregard but soon each pilot agreed that they had "no indications" (of speed). They never escaped a deep stall and it seems never went to a sixty degree bank in an attempt to escape the deep stall. They were doing what they had been trained to do if they were "low and slow" and inadvertently stalled the aircraft. Everything in their training revolves around the FlyByWire system that will not allow the pilot to stall the aircraft but once the wildly divergent pitot sensor data caused the computer to trip out, the plane was in Alternate Law mode and while some protections are in place, stalls are quite possible. Computer control over the stabilizer trim tab is lost upon entering Alternate Law but pilots appear not to have known that which is understandable since they never really train for Alternate Law.
So weather really played no role in this as they made a simple deviation around the weather, it was the non-turbulent icing that was a problem for them. They had erroneous indications of dangerously slow speed and then no reliable indications of speed at all. The initial climb of several thousand feet was due to the added thrust when they went to TOGA power, this also increased the severe nose-up attitude due to the pilot's failure to consider stabilizer settings being now manually controllable but not computer controllable.
The first incident was icing of the pitot and static air sources. Soon thereafter the computer became aware of unreliable data. The problem is that at the point in time wherein the static air source became iced and therefore the airplane's speed supposedly fell from 275 to 60, it is clear that such an event could not have taken place without some sort of deceleration forces having been applied. Prior to sending a series of critical warnings to the pilot, how about a critical warning to the computer: airspeed is clearly erroneous so make no immediate adjustments.
The pilot seeing such an abrupt air speed change should have known it was erroneous and probably related to static air data and the pilot hearing the stall warning should not have assumed that it was already a stalled aircraft. The FlyByWire system had pitched the nose up but not gravely, speed was decaying but not gravely.
Monday, May 16, 2011
AF447 FDR and CVR data readable.
The data from the Flight Data Recorder and Cockpit Voice Recorder is readable.
I'm sure that as the French finally get around to releasing transcripts more precise information will emerge and much of the speculation will have been proven to be baseless. A good deal of it, however, will have been shown to be quite sensible.
Its well known that the French investigation is focused on avoiding any allocation of blame. With all the inter-linkages between Airbus, AirFrance and the French government it is clear that the BEA is actually investigating the French government. The level of disclosure tolerated by the French is abysmally low in comparison to what the NTSB would be revealing.
The fact that controllers in Senegal and also Air France lost track of the flight entirely and that therefore Air France was woefully late in launching any search effort has absolutely nothing to do with the cause of any difficulties the plane suffered but it is a clear example of the differences in the type of investigation that is being conducted. The BEA will make no statement critical of Air France or commenting on the elapsed time prior to the initiation of search efforts.
I do so hope the French attitude changes and that the BEA will adopt a more open and informative orientation.
ADDENDUM: After the initial downloading and review of the FDR and CVR data, Airbus has sent a No Immediate Issuance of Any Service Bulletin telex to the aviation community. All the telex really means is that there is no immediately discernible indication of a design fault that would require immediate or extremely prompt servicing. Le Figaro appears to be encouraging the viewing of this Airbus action as some sort of determination that it was pilot error and nothing but pilot error. This is not the proper viewpoint. It will probably be next year before we learn a more complete sequence of events. The computerized flight control system detected a severe difference in pitot indications and dumped the pilots into Alternate Law. What the pilots did as far as attitude, throttle and angle of attack adjustments becomes critical. What the weather impact on the plane was will be critical. Probably the most important question is whether the CVR will reveal a reason for a lack of storm related course deviations. Also of interest will be any indications of whether the radar was operational. The important thing to remember at this point in the process is that the action recently taken by Airbus simply indicates no immediately discernible indication of a design fault that must be attended to immediately. It is not an accusation against the pilots and it is not a declaration that a design defect or deficiency may not later appear.
I remind everyone of the Gimli Glider incident in Canada. An airliner full of passengers suddenly became a glider. Most airline pilots today could not deal with that situation because they not only are not glider pilots they are button pushers who don't really have much "hands on" experience in flying an airliner. The don't know the "feel" of the controls in normal situations much less in emergency situations. The pilot of the flight that suddenly ran out of fuel in mid-air near Gimli happened to be glider pilot. He flew unpowered planes for fun. He suddenly found himself flying an unpowered airplane full of passengers. It was not fun, but he knew what to do and could do it. If Airbus dumps you into Alternate Law mode there no time to figure out what to do and get into Direct Law mode by pressing the overhead buttons. You know how to do it instinctively or you don't. Most pilots don't fly airplanes, they push buttons yet when it comes to the computer having given up and the pilots being forced to fly the plane out of a thunderstorm's downwash you can't call out "checklist", you can't grab for a manual and madly flip pages to find the directions, you can't break open the safe and find the secret code numbers to enter. You have to know what overhead buttons to press and where they are. You have to be able reach them in an unusual attitude with annoying alarms going off, you have to be able to fly the plane when you have to suddenly fly the plane. A simple act of leaving one altitude and climbing to another is done by pressing buttons because it saves fuel for a computer to make hundreds of extremely minute control inputs than to have some heavy-fisted human pilot actually try to level out precisely at a certain altitude. The problem is that when the pilots only have the experience of pressing the buttons and don't have any experience in flying the aircraft, they have no time to learn how to fly. When the computer gives up, they have to know how to get into direct law mode and fly the darn plane by feel even if its the dead of night and they suddenly find themselves in a super cell. The real problem is that any passenger selected at random probably has an much experience actually flying the plane as any of the pilots do. All the pilots have done is push buttons on a computer panel. The pilots do not really get any experience in flying the darn aircraft.
ADDENDUM: One item I've been considering for quite some time is the concept of latency of state. If ice accretion on the pitot tubes did send wildly disparate data to the computers and a program therefore cut out and dropped the pilots into Alternate Law, what happens if the fluctuations in the various readings have moments wherein the coincide within the accepted reference range. Is there a resumption of a Fly By Wire state wherein pilots are restrained from certain control inputs? Even in Alternate Law mode there are some systems working and a tendency to have a pronounced pitch-up. Its night, its stormy, the pitot condition may not have been noticed until there was the sudden cut-out and the computers that had been trying to deal with the speed signals suddenly realized the signals were spurious. Pitch and throttle settings had been changed to deal with the erroneous data during the time such erroneous data was well within the acceptable parameters, but once the data discrepancy exceeded the established parameters the pilots were in Alternate Law mode and in an unusual attitude with an unusual power setting.
Further ADDENDUM: A formal announcement is expected near the end of this week that will reflect a preliminary statement regarding the pilot's failure to react to a stall indication as automatic adjustments to pitch and thrust were made pursuant to erroneous pitot data prior to the computer declaring the pitot data to be unreliable and putting the plane in Alternate Law mode. Essentially its an instantaneous mental overload for the pilots who don't know that there has been a dangerous pitch up due to declining static pressure and the iced-up erroneous pitot readings. That reduced thrust and high angle of attack situation needs immediate attention and the pilots were probably so overwhelmed with a flurry of error messages and unreliable instruments that they just didn't have time to sort things out before they simply totally ran out of time and altitude. When you train people to push buttons on a computer rather than really flying the airplane its difficult for them to do what is right when that computer suddenly turns itself off. Increasing thrust with underslung engines and a somewhat aft center of gravity won't help the situation much. If they didn't realize what had happened immediately going into alternate sixty degree banks should have worked but it would take time. They had 35,000 feet to sort things out. It turned out not to be enough.
I'm sure that as the French finally get around to releasing transcripts more precise information will emerge and much of the speculation will have been proven to be baseless. A good deal of it, however, will have been shown to be quite sensible.
Its well known that the French investigation is focused on avoiding any allocation of blame. With all the inter-linkages between Airbus, AirFrance and the French government it is clear that the BEA is actually investigating the French government. The level of disclosure tolerated by the French is abysmally low in comparison to what the NTSB would be revealing.
The fact that controllers in Senegal and also Air France lost track of the flight entirely and that therefore Air France was woefully late in launching any search effort has absolutely nothing to do with the cause of any difficulties the plane suffered but it is a clear example of the differences in the type of investigation that is being conducted. The BEA will make no statement critical of Air France or commenting on the elapsed time prior to the initiation of search efforts.
I do so hope the French attitude changes and that the BEA will adopt a more open and informative orientation.
ADDENDUM: After the initial downloading and review of the FDR and CVR data, Airbus has sent a No Immediate Issuance of Any Service Bulletin telex to the aviation community. All the telex really means is that there is no immediately discernible indication of a design fault that would require immediate or extremely prompt servicing. Le Figaro appears to be encouraging the viewing of this Airbus action as some sort of determination that it was pilot error and nothing but pilot error. This is not the proper viewpoint. It will probably be next year before we learn a more complete sequence of events. The computerized flight control system detected a severe difference in pitot indications and dumped the pilots into Alternate Law. What the pilots did as far as attitude, throttle and angle of attack adjustments becomes critical. What the weather impact on the plane was will be critical. Probably the most important question is whether the CVR will reveal a reason for a lack of storm related course deviations. Also of interest will be any indications of whether the radar was operational. The important thing to remember at this point in the process is that the action recently taken by Airbus simply indicates no immediately discernible indication of a design fault that must be attended to immediately. It is not an accusation against the pilots and it is not a declaration that a design defect or deficiency may not later appear.
I remind everyone of the Gimli Glider incident in Canada. An airliner full of passengers suddenly became a glider. Most airline pilots today could not deal with that situation because they not only are not glider pilots they are button pushers who don't really have much "hands on" experience in flying an airliner. The don't know the "feel" of the controls in normal situations much less in emergency situations. The pilot of the flight that suddenly ran out of fuel in mid-air near Gimli happened to be glider pilot. He flew unpowered planes for fun. He suddenly found himself flying an unpowered airplane full of passengers. It was not fun, but he knew what to do and could do it. If Airbus dumps you into Alternate Law mode there no time to figure out what to do and get into Direct Law mode by pressing the overhead buttons. You know how to do it instinctively or you don't. Most pilots don't fly airplanes, they push buttons yet when it comes to the computer having given up and the pilots being forced to fly the plane out of a thunderstorm's downwash you can't call out "checklist", you can't grab for a manual and madly flip pages to find the directions, you can't break open the safe and find the secret code numbers to enter. You have to know what overhead buttons to press and where they are. You have to be able reach them in an unusual attitude with annoying alarms going off, you have to be able to fly the plane when you have to suddenly fly the plane. A simple act of leaving one altitude and climbing to another is done by pressing buttons because it saves fuel for a computer to make hundreds of extremely minute control inputs than to have some heavy-fisted human pilot actually try to level out precisely at a certain altitude. The problem is that when the pilots only have the experience of pressing the buttons and don't have any experience in flying the aircraft, they have no time to learn how to fly. When the computer gives up, they have to know how to get into direct law mode and fly the darn plane by feel even if its the dead of night and they suddenly find themselves in a super cell. The real problem is that any passenger selected at random probably has an much experience actually flying the plane as any of the pilots do. All the pilots have done is push buttons on a computer panel. The pilots do not really get any experience in flying the darn aircraft.
ADDENDUM: One item I've been considering for quite some time is the concept of latency of state. If ice accretion on the pitot tubes did send wildly disparate data to the computers and a program therefore cut out and dropped the pilots into Alternate Law, what happens if the fluctuations in the various readings have moments wherein the coincide within the accepted reference range. Is there a resumption of a Fly By Wire state wherein pilots are restrained from certain control inputs? Even in Alternate Law mode there are some systems working and a tendency to have a pronounced pitch-up. Its night, its stormy, the pitot condition may not have been noticed until there was the sudden cut-out and the computers that had been trying to deal with the speed signals suddenly realized the signals were spurious. Pitch and throttle settings had been changed to deal with the erroneous data during the time such erroneous data was well within the acceptable parameters, but once the data discrepancy exceeded the established parameters the pilots were in Alternate Law mode and in an unusual attitude with an unusual power setting.
Further ADDENDUM: A formal announcement is expected near the end of this week that will reflect a preliminary statement regarding the pilot's failure to react to a stall indication as automatic adjustments to pitch and thrust were made pursuant to erroneous pitot data prior to the computer declaring the pitot data to be unreliable and putting the plane in Alternate Law mode. Essentially its an instantaneous mental overload for the pilots who don't know that there has been a dangerous pitch up due to declining static pressure and the iced-up erroneous pitot readings. That reduced thrust and high angle of attack situation needs immediate attention and the pilots were probably so overwhelmed with a flurry of error messages and unreliable instruments that they just didn't have time to sort things out before they simply totally ran out of time and altitude. When you train people to push buttons on a computer rather than really flying the airplane its difficult for them to do what is right when that computer suddenly turns itself off. Increasing thrust with underslung engines and a somewhat aft center of gravity won't help the situation much. If they didn't realize what had happened immediately going into alternate sixty degree banks should have worked but it would take time. They had 35,000 feet to sort things out. It turned out not to be enough.
Thursday, May 5, 2011
AF447 Recovery of corpses from seabed.
It seems the families are divided as to attempts to recover corpses. It would be difficult but one has already been recovered so it is certainly not impossible. I doubt the corpses have much evidentiary value. The Brazilian mortuary was overwhelmed but would not permit the French doctors to perform autopsies. The Brazilian medical examiner made comments and drew conclusions well beyond his sphere of expertise even suggesting that the passengers had assumed crash positions prior to impact. Given the large number of passengers who were not belted-in and the fact that the galley had not been secured it was quite likely that there had been no instructions to even fasten seat belts much less an instruction concerning severe turbulence and certainly no instruction to assume crash positions.
The damage and injuries on the earlier recovered corpses all show major impact forces on the underside of an intact fuselage. Nothing reflects instructions to cabin crew or passengers regarding imminent impact.
There has been alot of discussion regarding overspeed conditions and underspeed conditions. At 35,000 feet that "coffin corner" is still a pretty wide window. And a stall could begin at six or seven degrees at that altitude. Presumably the iced-up pitot tubes would create a differential airspeed reading that was well beyond the computer's parameters. Error messages would be sent as the flight director computers cut out and placed the plane in Alternate Law mode. If speed either decayed too far or increased to far, a stall would take place but it seems crew awareness and crew actions should have been able to deal with it. Surely from 35,000 to 0 would have taken some time. Initially the French searchers felt they should estimate continued flight along intended course for some time beyond the Last Known Position. Although this may have taken place and then been followed by a course reversal that was either commanded or most likely uncommanded, it is quite possible that descent was rapid and largely devoid of forward motion. I can only envision direct flight into a powerful supercell as doing that.
I look forward to the CVR showing us why there were no WX related course changes to thread the plane around the intense storms. I look forward to the FDR analysis even more, but I don't see the corpse recovery as providing critical information.
The damage and injuries on the earlier recovered corpses all show major impact forces on the underside of an intact fuselage. Nothing reflects instructions to cabin crew or passengers regarding imminent impact.
There has been alot of discussion regarding overspeed conditions and underspeed conditions. At 35,000 feet that "coffin corner" is still a pretty wide window. And a stall could begin at six or seven degrees at that altitude. Presumably the iced-up pitot tubes would create a differential airspeed reading that was well beyond the computer's parameters. Error messages would be sent as the flight director computers cut out and placed the plane in Alternate Law mode. If speed either decayed too far or increased to far, a stall would take place but it seems crew awareness and crew actions should have been able to deal with it. Surely from 35,000 to 0 would have taken some time. Initially the French searchers felt they should estimate continued flight along intended course for some time beyond the Last Known Position. Although this may have taken place and then been followed by a course reversal that was either commanded or most likely uncommanded, it is quite possible that descent was rapid and largely devoid of forward motion. I can only envision direct flight into a powerful supercell as doing that.
I look forward to the CVR showing us why there were no WX related course changes to thread the plane around the intense storms. I look forward to the FDR analysis even more, but I don't see the corpse recovery as providing critical information.
Tuesday, May 3, 2011
AF447 CVR located and recovered.
The Cockpit Voice Recorder has been both located on the sea floor and successfully recovered from it by a robotic vehicle.
The device is under BEA seal and the data, if any, will be available to the BEA in about two weeks, though its release will surely be several months at the earliest.
The CVR will provide us with information regarding whether the lack of WX related course deviations was truly due to crew fatigue. There are many who believe this is not a WX related event and who interpret the ACARS data as showing two pilots present and alert in the cockpit. There seems little to indicate a malfunctioning or non-functioning WX radar aboard the aircraft and it would seem that if any such event took place the crew would not have engaged in continued flight into the ITCZ but would have returned and landed. It is likely that had there been any such non-functioning onboard radar system the crew would have at the very least issued a turbulence warning to passengers and crew. This would have resulted in an activation of the seat belt sign and the securing of the galley. Yet damage inspection and autopsy reports indicate that the galley was not secured and a high number of passengers did not have their seat belts fastened.
The device is under BEA seal and the data, if any, will be available to the BEA in about two weeks, though its release will surely be several months at the earliest.
The CVR will provide us with information regarding whether the lack of WX related course deviations was truly due to crew fatigue. There are many who believe this is not a WX related event and who interpret the ACARS data as showing two pilots present and alert in the cockpit. There seems little to indicate a malfunctioning or non-functioning WX radar aboard the aircraft and it would seem that if any such event took place the crew would not have engaged in continued flight into the ITCZ but would have returned and landed. It is likely that had there been any such non-functioning onboard radar system the crew would have at the very least issued a turbulence warning to passengers and crew. This would have resulted in an activation of the seat belt sign and the securing of the galley. Yet damage inspection and autopsy reports indicate that the galley was not secured and a high number of passengers did not have their seat belts fastened.
Monday, May 2, 2011
AF447 FDR's Data Module recovered.
The Data Module that apparently was dislodged from the Flight Data Recorder upon impact was located and roboticly recovered. It appears to be intact based upon gross examination but it is likely to be two weeks before it is known if data can be recovered from it. It is likely to be considerably longer, however, before any recovered data is released to the aviation community or the public. Shear forces on the bolts of the data module and the ultrasonic pinger are unknown but separation of the data module from the chassis of the FDR has taken place in prior crashes involving impact with water. The BEA has not released definitive data on the pinger but it is believed that signals were recorded by the French submarine involved in the search but not detected at the time by sonar operators. Later data scrubbing is said to have provided a computerized detection of the pings.
There are so many questions as to why the plane went into an unusual attitude and most particularly as to why there was no recovery despite there being 35,000 feet in order for their to be recovery from a stall. It is possible that a pilot's attempt to re-establish computerized flight control created an additional upset since there is a difference between the plane's attitude and the computer systems processing of the attitude information. Essentially it would be a physical recovery from the stall followed by a systems induced control input that re-established the stalled condition as an unrecoverable deep stall.
There has been much ado about extreme turbulence and the plane's presumed entry into a supercell whose location was masked by a radar detection of a smaller but nearer cell. The trouble with all this turbulent weather assumptions is that other planes transited the area without incident and nothing in the wreckage or bodily injury data indicates that the plane ever encountered severe weather. Galley equipment was not secured and dogged-down as it would be had there been a severe turbulence warning nor was there any indication that all passengers had seat belts securely fastened.
The ACARS system indicates pitot icing problems causing a discrepancy in the speed data the flight computers were receiving. As the computer put the plane into Alternate Law mode and therefore no computer would override pilot input, there would have to have been prompt and proper action by the crew. I can not envision a pilot attempting to re-engage computer controlled flight until the plane itself was known to have recovered from any stalled situation. There was massive loss of data to the pilots and massive cognitive overload as whatever data that was available was unreliable.
As yet, there has been no word concerning the other "black box", the Cockpit Voice Recorder.
There are so many questions as to why the plane went into an unusual attitude and most particularly as to why there was no recovery despite there being 35,000 feet in order for their to be recovery from a stall. It is possible that a pilot's attempt to re-establish computerized flight control created an additional upset since there is a difference between the plane's attitude and the computer systems processing of the attitude information. Essentially it would be a physical recovery from the stall followed by a systems induced control input that re-established the stalled condition as an unrecoverable deep stall.
There has been much ado about extreme turbulence and the plane's presumed entry into a supercell whose location was masked by a radar detection of a smaller but nearer cell. The trouble with all this turbulent weather assumptions is that other planes transited the area without incident and nothing in the wreckage or bodily injury data indicates that the plane ever encountered severe weather. Galley equipment was not secured and dogged-down as it would be had there been a severe turbulence warning nor was there any indication that all passengers had seat belts securely fastened.
The ACARS system indicates pitot icing problems causing a discrepancy in the speed data the flight computers were receiving. As the computer put the plane into Alternate Law mode and therefore no computer would override pilot input, there would have to have been prompt and proper action by the crew. I can not envision a pilot attempting to re-engage computer controlled flight until the plane itself was known to have recovered from any stalled situation. There was massive loss of data to the pilots and massive cognitive overload as whatever data that was available was unreliable.
As yet, there has been no word concerning the other "black box", the Cockpit Voice Recorder.
Saturday, April 16, 2011
AF447 Search: Phase 5.
The FDR was found but only the chasis was located the Data Memory Module was not attached and is thought to have separated at impact. There are some other crashes wherein this separation has taken place. 04/27/11.
..................................................
The final phase of the search for the "black boxes" will be getting underway. The Alucia considered its mission complete and departed the area on April 9th. Any delay in getting the successor search vessel on scene was due to the high level of breaks in submarine cables after recent seismic activity.
The debris field is very close to the Last Known Position despite all the earlier reliance on mathematical modeling of winds and currents.
The tail section has been located. It is thought that this presently contains the FDR and CVR that would normally be found there. Recovery of the black boxes is highly probable, recovery of information from them is less certain but there is a generally high expectation that data will be recovered. Its not yet known just why the ultrasonic pingers were not detected during earlier searches.
So it becomes the usual: wait and see.
Addendum: I've been informed that the initial attempt will be tomorrow April 22nd. BEA representative is aboard even a psychologist to deal with ships crew who have difficulties dealing with the images.
..................................................
The final phase of the search for the "black boxes" will be getting underway. The Alucia considered its mission complete and departed the area on April 9th. Any delay in getting the successor search vessel on scene was due to the high level of breaks in submarine cables after recent seismic activity.
The debris field is very close to the Last Known Position despite all the earlier reliance on mathematical modeling of winds and currents.
The tail section has been located. It is thought that this presently contains the FDR and CVR that would normally be found there. Recovery of the black boxes is highly probable, recovery of information from them is less certain but there is a generally high expectation that data will be recovered. Its not yet known just why the ultrasonic pingers were not detected during earlier searches.
So it becomes the usual: wait and see.
Addendum: I've been informed that the initial attempt will be tomorrow April 22nd. BEA representative is aboard even a psychologist to deal with ships crew who have difficulties dealing with the images.
Tuesday, April 12, 2011
AF447 and the One Legged Stool
I have long been a believer in the one-legged stool theory of aviation safety. Now, I know have very few readers but since some are not in the USA let me set forth with some particularity what a one-legged stool is. Most stools have three legs and are therefore quite stable when someone sits on them. A stool with one leg allows a person to sit on it and rest just as much but does require that he remain alert. Drifting off to sleep on a three legged stool can be easy. Drifting off to sleep while on a one legged stool is just as simple but has obvious consequences: you fall and that wakes you up.
One consequence of the ultra reliability of any electrical system is that monitoring it becomes so very boring. The aircraft systems are reliable. If something fails, there is a primary backup, if that fails there is a secondary backup, etc. It is so very boring. The Fly By Wire takes this even one step further and declares that the computers are so reliable that they should even be able to override pilot input since pilots are less reliable. Only in Alternate Law Mode does the computer surrender control of the airplane to the pilot and even that is done piecemeal with only certain systems surrendered to pilot control during the various Alternate Law modes.
But what about the Pilot's mode? Pilots have various modes too. Alert, drowsy, asleep, "in the loop", "out of the loop", etc. Qantas used to force pilots to work out a position report using pen and paper rather than simply pressing a button. This would keep the pilots mentally in the loop should trouble occur. We all know that on long flights there are a variety of breaks some of which are mandatory and some of which are optional.
Physical presence on the flight deck is one thing. Mental presence is another. A pilot who is sleeping in the cockpit or merely dozing there requires some time to get back into the loop.
Some have alleged that failure to deviate around weather indicates a dozing flight crew. All we can really say is that we have no indication that the pilots were in their seats and alert. We have some indications that fatigue may have overcome a sole occupant of the cockpit. While we can think that a pilot would not nap as the plane was penetrating the convergence zone it is possible and he might have asked to be roused and the pilot who was to have him awakened fell asleep.
Dealing with suddenly being dumped into Alternate Law Mode is difficult when the pilot is awake and fully aware of his situation. If a stall developed we all wonder why but mostly we wonder why was there no reaction to it. After all, by the third lesson student pilots are practicing stalls and having it drummed into them to lower the nose instantly and add power instantly. Nothing about that changes as one progresses from student pilot in a Cessna 172 to the left-seat of a multi-engined airliner. What changes is the excitement. A student pilot on his third lesson is "in the loop". He has full situational awareness. The professional pilot in an airliner has thousands of hours of sitting there punching buttons into a computer that does the flying. Its boring. We need to bring back that one-legged stool. Keep the pilots from getting complacent.
We all wonder how the airframe could become involved in a deep stall but the question that also must be addressed is how did it all begin. If it takes too long to wake up and gain situational awareness, a brief nap can be fatal. Airline pilots really do not practice stalls. They practice approaches to stalls. Once the airspeed deteriorated and a stall started to develop action can be taken. Once a stall is fully developed and the plane is in a deep stall wherein control surfaces are functionally ineffective there is far less that can be done and the pilot is in unknown territory. It is possible the pilot woke up to a nightmare. Often an instructor will have a student practice suddenly being required to select an emergency landing zone and make the approach to a landing in some pasture rather than on an airport runway. The practice usually ends just above the farmer's field but a good instructor will sometimes actually have a student make such a landing. A recently awakened pilot might not have been able to cope with a deep stall because he had only been given simulator time for an approach to a stall. There is no Emergency Checklist for a deep stall. You either know what to do or you don't. It either works or it doesn't. There is no time to do aught but act. If there was a double flame-out (doubtful) differential thrust could not be used.
One consequence of the ultra reliability of any electrical system is that monitoring it becomes so very boring. The aircraft systems are reliable. If something fails, there is a primary backup, if that fails there is a secondary backup, etc. It is so very boring. The Fly By Wire takes this even one step further and declares that the computers are so reliable that they should even be able to override pilot input since pilots are less reliable. Only in Alternate Law Mode does the computer surrender control of the airplane to the pilot and even that is done piecemeal with only certain systems surrendered to pilot control during the various Alternate Law modes.
But what about the Pilot's mode? Pilots have various modes too. Alert, drowsy, asleep, "in the loop", "out of the loop", etc. Qantas used to force pilots to work out a position report using pen and paper rather than simply pressing a button. This would keep the pilots mentally in the loop should trouble occur. We all know that on long flights there are a variety of breaks some of which are mandatory and some of which are optional.
Physical presence on the flight deck is one thing. Mental presence is another. A pilot who is sleeping in the cockpit or merely dozing there requires some time to get back into the loop.
Some have alleged that failure to deviate around weather indicates a dozing flight crew. All we can really say is that we have no indication that the pilots were in their seats and alert. We have some indications that fatigue may have overcome a sole occupant of the cockpit. While we can think that a pilot would not nap as the plane was penetrating the convergence zone it is possible and he might have asked to be roused and the pilot who was to have him awakened fell asleep.
Dealing with suddenly being dumped into Alternate Law Mode is difficult when the pilot is awake and fully aware of his situation. If a stall developed we all wonder why but mostly we wonder why was there no reaction to it. After all, by the third lesson student pilots are practicing stalls and having it drummed into them to lower the nose instantly and add power instantly. Nothing about that changes as one progresses from student pilot in a Cessna 172 to the left-seat of a multi-engined airliner. What changes is the excitement. A student pilot on his third lesson is "in the loop". He has full situational awareness. The professional pilot in an airliner has thousands of hours of sitting there punching buttons into a computer that does the flying. Its boring. We need to bring back that one-legged stool. Keep the pilots from getting complacent.
We all wonder how the airframe could become involved in a deep stall but the question that also must be addressed is how did it all begin. If it takes too long to wake up and gain situational awareness, a brief nap can be fatal. Airline pilots really do not practice stalls. They practice approaches to stalls. Once the airspeed deteriorated and a stall started to develop action can be taken. Once a stall is fully developed and the plane is in a deep stall wherein control surfaces are functionally ineffective there is far less that can be done and the pilot is in unknown territory. It is possible the pilot woke up to a nightmare. Often an instructor will have a student practice suddenly being required to select an emergency landing zone and make the approach to a landing in some pasture rather than on an airport runway. The practice usually ends just above the farmer's field but a good instructor will sometimes actually have a student make such a landing. A recently awakened pilot might not have been able to cope with a deep stall because he had only been given simulator time for an approach to a stall. There is no Emergency Checklist for a deep stall. You either know what to do or you don't. It either works or it doesn't. There is no time to do aught but act. If there was a double flame-out (doubtful) differential thrust could not be used.
Tuesday, April 5, 2011
Undercarriage lowered on AF447?
Even though Phase 4 is still underway and the FDR and CVR have not been located, it seems the suspense is growing. A very small debris field, high vertical descent component suspected. Did the crew fixate on a problem and not realize they were in a gentle descent? Or did they attempt to increase drag by lowering the undercarriage? Why would they attempt such a thing? I just can't wait for the Flight Data Recorder to be recovered and analyzed, though I don't know how complete the information would be. Its possible that the data was changing so rapidly the recording of averages will be deceptive.
It just doesn't seem reasonable that the pilots would descend all the way from Level 35 to sea level without realizing they were descending, no matter how fixated on some other item they might have been. Also during any such gentle descent, the lateral travel from the LKP would be greater no matter what the direction of that travel was, even if they did a 180 whether it be intentionally or by happenstance. Being dumped into Alternate Law mode suddenly and losing so many instruments without warning may have overloaded the crews ability to cope with sensory input, but the plane would travel in some direction even if they lost engines.
I had thought a deep stall but this appears to be so very unlikely since it is not known how to put that plane into a deep stall and yet have it strike the water in a mainly downward direction consistent with a flat spin. So perhaps the answer is simply that there was no flat spin. Yet extending the undercarriage would indicate pilot attention and desperation. It would also indicate an over speed condition, not an under speed condition, though it is possible that a desperate crew could have extended the gear even if the airspeed was minimal simply as an attempt to lower the nose to re-stablish laminar flow over the wing and gain an attitude wherein an engine re-start might take place. However, I do think that only impact forces affected the main landing gear and the main landing gear doors. I doubt there was any inflight attempt to lower the landing gear.
Some are focusing on WX and pilot induced oscillations leading to a loss of longitudinal dynamic stability so that the plane was tail heavy and both engines had flamed out. No control surfaces effective, no way to restart, not much flight from LKP, high vertical component, low horizontal one. Not a good situation to be in at all and certainly not at night while inside a thunderstorm. Of course the problem is that other aircraft flew pretty much the same track not long before and not long after AF447 and had no problems at that flight level. So while its hard to imagine longitudinal stability problems its sort of harder to imagine anything else.
If the initial data from the iced pitot tube resulted in the throttles being retarded sufficiently to induce an underspeed stall and that stall progressed to a fully stalled situation with the nose down sixty degrees and neither the wings nor the horizontal stabilizer generating lift it is possible that leading edge vortexes impinged on the vertical stabilizer. Such vortexes when combined with commanded inputs from a desperate flight crew could result in a catastrophic failure at altitude but I would expect the fuselage to impact in a more vertical rather than horizontal attitude. So it seems that the vertical stabilizer did indeed remain attached until impact.
At impact the transmission would cease but the intermittent interruptions in the ACARS transmissions and unavailability of the signal is thought by some to have been caused by the antenna being occluded by an inverted fuselage but this is undetermined.
It just doesn't seem reasonable that the pilots would descend all the way from Level 35 to sea level without realizing they were descending, no matter how fixated on some other item they might have been. Also during any such gentle descent, the lateral travel from the LKP would be greater no matter what the direction of that travel was, even if they did a 180 whether it be intentionally or by happenstance. Being dumped into Alternate Law mode suddenly and losing so many instruments without warning may have overloaded the crews ability to cope with sensory input, but the plane would travel in some direction even if they lost engines.
I had thought a deep stall but this appears to be so very unlikely since it is not known how to put that plane into a deep stall and yet have it strike the water in a mainly downward direction consistent with a flat spin. So perhaps the answer is simply that there was no flat spin. Yet extending the undercarriage would indicate pilot attention and desperation. It would also indicate an over speed condition, not an under speed condition, though it is possible that a desperate crew could have extended the gear even if the airspeed was minimal simply as an attempt to lower the nose to re-stablish laminar flow over the wing and gain an attitude wherein an engine re-start might take place. However, I do think that only impact forces affected the main landing gear and the main landing gear doors. I doubt there was any inflight attempt to lower the landing gear.
Some are focusing on WX and pilot induced oscillations leading to a loss of longitudinal dynamic stability so that the plane was tail heavy and both engines had flamed out. No control surfaces effective, no way to restart, not much flight from LKP, high vertical component, low horizontal one. Not a good situation to be in at all and certainly not at night while inside a thunderstorm. Of course the problem is that other aircraft flew pretty much the same track not long before and not long after AF447 and had no problems at that flight level. So while its hard to imagine longitudinal stability problems its sort of harder to imagine anything else.
If the initial data from the iced pitot tube resulted in the throttles being retarded sufficiently to induce an underspeed stall and that stall progressed to a fully stalled situation with the nose down sixty degrees and neither the wings nor the horizontal stabilizer generating lift it is possible that leading edge vortexes impinged on the vertical stabilizer. Such vortexes when combined with commanded inputs from a desperate flight crew could result in a catastrophic failure at altitude but I would expect the fuselage to impact in a more vertical rather than horizontal attitude. So it seems that the vertical stabilizer did indeed remain attached until impact.
At impact the transmission would cease but the intermittent interruptions in the ACARS transmissions and unavailability of the signal is thought by some to have been caused by the antenna being occluded by an inverted fuselage but this is undetermined.
Subscribe to:
Posts (Atom)