[Update – Podcast with an A330 type rating instructor discussing the BEA AF447 information]
Update: MSNBC has a three minute video with two former NTSB investigators discussing the accident.
Citing inaccuracies and errors in media reports on the crash of AF447 off the coast of Brazil, the Bureau d’Enquêtes et d’Analyses pour la Sécurité de l’aviation civile in charge of the investigation released a note on the crash this afternoon.
Their note covers the preliminary results from the flight data recorder and cockpit voice recorder for the ill-fated Airbus A330-203 registered F-GZCP that departed Rio for Paris on 31 May 2009 with 12 crew members and 216 passengers on board that crashed in the South Atlantic Ocean off the coast of Brazil.
As many have suspected, the aircraft stalled. Their findings were summarizes as follows:
? The composition of the crew was in accordance with the operator’s procedures
? At the time of the event, the weight and balance of the airplane were within the operational limits
? At the time of the event, the two co-pilots were seated in the cockpit and the Captain was resting. The latter returned to the cockpit about 1 min 30 after the disengagement of the autopilot
? There was an inconsistency between the speeds displayed on the left side and the integrated standby instrument system (ISIS). This lasted for less than one minute.
? After the autopilot disengagement:
? The airplane climbed to 38,000 ft.
? The stall warning was triggered and the airplane stalled
? The inputs made by the PF were mainly nose-up
? The descent lasted 3 min 30, during which the airplane remained stalled. The angle of attack increased and remained above 35 degrees
? The engines were operating and always responded to crew commands.
? The last recorded values were a pitch attitude of 16.2 degrees nose-up, a roll angle of 5.3 degrees left, and a vertical speed of -10,912 ft/min.
These results clearly indicate that the aircraft went into a stall, and the nose-up, rather than nose down inputs that are required for stall recovery, were incorrect, exacerbating the problem and resulting in a loss of control of the aircraft.
The full text of the statement is available in English here: http://www.bea.aero/fr/enquetes/vol.af.447/point.enquete.af447.27mai2011.en.pdf .
To me, a lay person who has had no training, weren’t the pilots who were on deck at the time awfully young. It looks like their inexperience may have contributed to this crash. There were storms in the area and maybe the pilot should have waited until they were clear of these storms before taking a break. In crashes it is usually not 1 thing but a series of events that occur. Just a shame because this looks avoidable.
I can’t think of anything more terrifying than not knowing which instruments to trust. Your in a situation where a history of false data was already a known concern with this aircraft. Whether that was the case here has yet to be determined though. My heartfelt sympathies to the families.
lack of basic stick and rudder skills. In a stall lower the pitch and add power. In the case of unreliable instruments, set cruise power and level pitch and regain normal performance.
I agree completely with Duane. This is a by-product of overreliance on technology to fly the airplane. There’s a lack of basic airmanship, the stick and rudder skills alluded to that have been in a pilot’s toolkit since the Wright brothers. Same types of issues were involved in the Colgan air accident outside of BUF a few years ago FLY THE PLANE FIRST !
High altitude stalls in a swept wing airplane are nothing to mess with… but as they descended into thicker air, yes, they surely needed to reduce the angle of attack.
Yes I agre with prior comments
I just can’t believe that 3 pilots, couldn’t apply basic stall recovery procedures. Has anyone ever considered loss of hydraulics? Maybe they thought they were applying the right techniques when in fact it didn’t make a difference on the plane. I don’t care how young they were, stall recovery is one of the first things you learn.
I do hand carry delivery all over the world and I can safely say that I was relieved to learn it was pilot error – just sayin
With 3 pitot tubes available for redundancy, were they all iced up, what happened to the heaters? With 2 angle of attack vains giving attitute imformation throught seperate channels, were they also iced over and inop, what happened to the heaters?????
Maybe its a Airplane problem and not the pilots fault, maybe they didn’t have any attitude or pitot system imput. Also, what happened to the backup steam gages artificial gyro horizons which gets its power from the emergency bus, why wasn’t it available for attitude imformation. It doesn’t appear that all the imformation provided by the Flight Data Recorder is being reported.
What is missing in these reports is that the flight was in total darkness over water. No lighs on the ground, and stars invisible in the clouds. JFK’s son had a similar experience.
Sorry, forgot the t in lights. With no instruments and no ground reference (at 38,00 feet they could have been above or below the clouds and lightning) and have no idea of up and down or left and right. Flying by the “seat of one’s pants” is a myth. I learned this during my pilot training. My instructor blindfolded me and did a number of manuevers and I never felt like anything but level flight.
Ah, so much more info remaining to be made public. This model of Airbus has envelope protection to keep pilots from making the inputs reported and inducing the deep stall. 16 years on the AirBus taught me somrthing, those planes they make, every model do the strangest damn things totally unexpected and beyond what can be taught in the simulator.
If the truth be told and the culprit not covered up by the political power of AirBus, you may have a very different opinion from the one’s already expressed here.
I love all the armchair quarterbacks. Put in the same situation, you don’t know what you would do, until you do it. In the whole scheme of things, 3 1/2 minutes isn’t a long time to recover the aircraft. It took the captain a minute and a half just to get back to the flight deck. If the aicraft had accumulated severe icing, there may have been nothing they could have done to stop the descent. Ice on wing surfaces can have a myriad of results, such as restricted control surfaces and lack of lift. God rest their souls.
In my early flight training, I was told that if all else failed, let go the controls and the bird will find it’s own way home.
You must be very naive. These were airline pilots
for Air France. It is inconceivable that these qualified pilots with thousands of hours flying time, did not know how to recover from a stall!!
I agree completely. There must have been some major malfunction, or pilot incapacitation, to cause such a
situation.
I think Brian G Mortin is closest here. Airbus seems to be the biggest culprit, but all these very, very computerized planes have had odd responses. Manual flying is no longer manual, so just pushing forward on the yoke won’t necessarily put the nose down. The plane feels it knows better.
As far as redundant sensors, that’s been an issue before, on a lot of air/space vehicles. Two in agreement and one disagreeing seems, to humans, to be obvious: The third one is failing. Doesn’t always get programmed that way. Even if the aircrew did have a visual frame of reference, if the plane thinks it knows better, but has improper data, you might not get out of it no matter how good a stick and rudder man you are.
First, it cannot be a hydraulics failure because the cockpit recorder system is supposed to provide the info from the controlling system, not hydraulics output system. In other words, it must have been a failure to follow the standard stall recovery procedure. However, it may be suggested that the storm conditions and lack of trainings in the co-pilots in charge contributed to the crash.
Also for the captain who was on a brake, it should have been very reasonable for him to take a break because it should have been an ordinary day in the cockpit.
Instruments must be trusted when flying at night or in cloud – but regardless of the “glass displays” that the Airbus aircraft have they also have an artificial horizon – they have these in the glass display as well as a gyro instrument below the glass display. For over three minutes the PF would have scanned the panel, and was confused by a) lack of outside reference (it was night) b) the initial incorrect airspeed shown due to the iced up pitot tubes.
The fact that the aircraft was in a 35 degree nose up attitude is appalling to me – these guys should have known how to ‘fly the plane’ – and while the air speed indicator would have been showing a higher than true air speed due to the pitot icing, there is no excuse for the PF maintaining a high nose attitude. It was just stupid beyond comprehension. Remember, even if the glass display was showing aberrant data, the gyro artificial horizon below the glass instruments would have shown that he was 35 degrees nose high.
Based on the reported voice comments, the PF was confused by the aberrant data on the screen – yet he didn’t resort to the basic gyro instrument – if he had, they would have been OK.
adrian says May 29, 2011 at 8:47 pm
I just can’t believe that 3 pilots, couldn’t apply basic stall recovery procedures.
———
Earlier reporting on flightglobal concluded that this same situation has occurred multiple times in recent years and that several of the flight crews took well over a minute to establish pitch/power. This wasn’t an isolated freak incident – this crew just took too long and didn’t recognize the situation before they ran out of altitude.
I think you said it right. This happens more than we would like to know if the truth be said.
Indeed it seems clear to me that there are serious programing errors at play here. Brian G. Mortin points out again the envelope protection on the 300 models so theoretically it is impossible to stall the plane. Also apparently the the whole system shuts down when the airspeed is not available, but why would it not, if programed well, automatically pitch the plane and set thrust to the requisite N settings? based on the last altitude readings to main straight and level flight?
I’m not sure i want to fly on the 300 models anymore but wonder if that isn’t the case on all fly by wire aircraft.
As an idiot passenger sitting in the cabin, I can tell a steep nose-up at flight altitude is not normal, and I can certainly tell when a descent is too fast. The cockpit crew was incredibly inept and stupid to boot. Never, ever fly Air France (or British Airways, either, but that is another story).
This is just an opinion as I have no idea if Airbus has this system on the A330 or not but it seems that if they used the all-in-one type of pitot-static system (static ports and aoa sensor all incorporated in the pitot tubes), when those tubes iced up, it very well may have caused all of their indications to be erratic (ie: airspeed, altitude and attitude). Again, this is purely speculation as I don’t know if they use that type of system but those systems are out there on a lot of newer planes.
I fly a lot from Spain to the UK & back to see family.
I am only a passenger and do not know the first thing about Aircraft, but for a qualified pilot not to be able to control a plane in stall is very frightening. Are they not taught do apply any proceedures during emergencies like this ? , or do they just rely on Auto Pilot to get them from A to B. Not very reasuring is it, I might change to driving overland instead, at least I will be in charge of my own destiny.
Gravity never fails. After all , it is still a priviledge to be in the air. Every flight is a risk. Those poor pilots did their best , I’m sure. The engineers, programmers and the rest of the industry will take lessons from this.
As a forty year pilot with an airline transport rating, I have been watching these Airbus accidents with great interest, There is a pattern of pilots telling plane to do one thing and the plane doing something different. The common thread in these airbus accidents is the “fly by wire” systems that is unique to the airbus. When the pilot moves the controls, he is only telling the computer what he wants to happen and the servo motors actually make the control surfaces move. There seems to be a pattern with the Airbus accidents where the pilot is moving the controls and the plane responds differently than the input. Airbus (and the FAA) can easily say that the pilot made the wrong moves and caused the plane to stall (or roll). This seems to be a pattern with Aibus aircraft and not with Boeing where the controls are directly wired to the surfaces (ie aliorons, elevators and rudders)
It is hard to believe that with two experienced pilots in the cockpit, they make the wrong moves to recover from a stall.
@Matt – “Relieved it was just pilot error”? Is the emphasis there on the “just”? I hope I never have to sit in the back of an aircraft that you may be the pilot of. Air France didn’t spend millions of Euros training these pilots for them to “just make an error”. Having said that, the A340 overrunning the rain-soaked runway in Toronto was also pilot error, so maybe Air France’s training methods need serious reviewing.
It is simply not true that fly by wire is unique to Airbus. The Boeing 777 is fly by wire as is the Falcon 7X business jet and many military aircraft.
There seems to be a big confusion here between ‘pitch attitude’ and angle of attack. the angle of attack was 35 degrees and the pitch attitude about 16 degrees. This means the angle of decent was about 19 degrees. So if the crew had set zero degrees pitch attitude they would still have had an angle of attack of 19 degrees.
Ok…even if they had false info on airspeed from the frozen pitot tubes…how about ground speed from their GPS systems…that should’ve raised some eyebrows
It is also inconceivable, to a reasoning pilot’s mind, that the PIC of the AF 340 accident @ CYYZ would attempt to touchdown and stop under a thunderstorm with only only about 5000 ft of flooded runway remaining. There are guidlines based on long years of experience for what to do in the presence of TRW+. Circumnavigate them enroute and avoid attempting to land or take-off when one is near the runway. These guidlines were not heeded apparently.
Right on Mike. In military training of long ago they had to practice those flat spins (wings stalled). Recovery required a negative pitch attitude to get airflow again, then pull out of the descent. AF447 would have likewise required a negative pitch attitude until there was some indication of an increase in speed. Interesting that no one else raised the point. Parhaps the software prevented a pitch down attitude in their situation.
Mike Groves says June 3, 2011 at 4:03 pm
In that case Mike that would depend on their forward speed … and seems to me they didn’t have any, they stall at 35% nos up and in that position unles you manage to put nos down you are becoming a brick….It doesn’t appear that all the imformation provided by the Flight Data Recorder is being reported.
As much as I hate to second guess / Monday morning quarterback I must agree with your assessment, the sad thing is that we pilots train for that sort of thing, perhaps they were also severely distracted. The standby instruments are vacuum and/or electrically driven and do not have air data computer inputs like the primary and secondary “glass” displays which were in this case in error. Whenever I flew (Emb-145) it was a natural part of my scan to include the standby attitude indicator.
Luke
Ehay is so far the only person who may have correctly interpreted these findings. If in fact the Airbus was embedded in TRW+, it may not have mattered how good or poor the pilots’ scans were.
However, I wouldn’t go blaming the pilots solely just yet. There is enough meteorological evidence to scatter the blame.
Firstly, nowhere have I read what the indicated airspeeds were during the descents. We only know what the vertical descent rates were; and the stall determination has only been made from the AOA indicators
Secondly, Massive TRWs can have downdrafts of several thousand feet per second. Remember the 1985 Delta tragedy at KDFW? Humans have really only been studying the potential energy in TRW downdrafts in the last 25 years. Much remains unknown.
Third, yes all glass cockpits have backup instruments independent of the primary AHRS–which usually includes an electric AI, but based upon my years in glass cockpits, I can safely say there aren’t backup IAS indicators save for the co-pilot’s. It’s unlikely that with such redundancy, a catastrophe like that described in AF 447 could occur.
Fourth, the instruments indicated the jet was in nose high attitudes as severe as 35 degrees and as shallow as 16, AND in a 5 degree roll. These could be indications of a stall, flat spin entry and subsequent flat spin. Remember, an airplane can stall at any airspeed and any angle; therefore an airplane can also stall into a flat spin at any airspeed and any angle, too. Most flat spins are unrecoverable.
Finally, the three flight crew members on the flight deck during this event probably felt and saw their amazing descent rate. So as evidenced by the data, probably tried to use their still operational motors to power out of the situation. If three pilots, and two massive engines could not get the nose down, nor return the aircraft to straight and level flight, it is safe to conclude whatever occurred to create this abnormal flight situation was beyond their control; not that it was beyond their competence.
If in fact the Airbus was embedded in TRW+, it may not have mattered how good or poor the pilots’ scans were.
However, I wouldn’t go blaming the pilots solely just yet. There is enough meteorological evidence to scatter the blame.
Firstly, nowhere have I read what the indicated airspeeds were during the descents. We only know what the vertical descent rates were; and the stall determination has only been made from the AOA indicators
Secondly, Massive TRWs can have downdrafts of several thousand feet per second. Remember the 1985 Delta tragedy at KDFW? Humans have really only been studying the potential energy in TRW downdrafts in the last 25 years. Much remains unknown.
Third, yes all glass cockpits have backup instruments independent of the primary AHRS–which usually includes an electric AI, but based upon my years in glass cockpits, I can safely say there aren’t backup IAS indicators save for the co-pilot’s. It’s unlikely that with such redundancy, a catastrophe like that described in AF 447 could occur.
Fourth, the instruments indicated the jet was in nose high attitudes as severe as 35 degrees and as shallow as 16, AND in a 5 degree roll. These could be indications of a stall, flat spin entry and subsequent flat spin. Remember, an airplane can stall at any airspeed and any angle; therefore an airplane can also stall into a flat spin at any airspeed and any angle, too. Most flat spins are unrecoverable.
Finally, the three flight crew members on the flight deck during this event probably felt and saw their amazing descent rate. So as evidenced by the data, probably tried to use their still operational motors to power out of the situation. If three pilots, and two massive engines could not get the nose down, nor return the aircraft to straight and level flight, it is safe to conclude whatever occurred to create this abnormal flight situation was beyond their control; not that it was beyond their competence.
Your comment really only applies to non-swept wing aircraft operating considerably below mach:
Ehay is so far the only person who may have correctly interpreted these findings. If in fact the Airbus was embedded in TRW+, it may not have mattered how good or poor the pilots’ scans were.
However, I wouldn’t go blaming the pilots solely just yet. There is enough meteorological evidence to scatter the blame.
Firstly, nowhere have I read what the indicated airspeeds were during the descents. We only know what the vertical descent rates were; and the stall determination has only been made from the AOA indicators
Secondly, Massive TRWs can have downdrafts of several thousand feet per second. Remember the 1985 Delta tragedy at KDFW? Humans have really only been studying the potential energy in TRW downdrafts in the last 25 years. Much remains unknown.
Third, yes all glass cockpits have backup instruments independent of the primary AHRS–which usually includes an electric AI, but based upon my years in glass cockpits, I can safely say there aren’t backup IAS indicators save for the co-pilot’s. It’s unlikely that with such redundancy, a catastrophe like that described in AF 447 could occur.
Fourth, the instruments indicated the jet was in nose high attitudes as severe as 35 degrees and as shallow as 16, AND in a 5 degree roll. These could be indications of a stall, flat spin entry and subsequent flat spin. Remember, an airplane can stall at any airspeed and any angle; therefore an airplane can also stall into a flat spin at any airspeed and any angle, too. Most flat spins are unrecoverable.
Finally, the three flight crew members on the flight deck during this event probably felt and saw their amazing descent rate. So as evidenced by the data, probably tried to use their still operational motors to power out of the situation. If three pilots, and two massive engines could not get the nose down, nor return the aircraft to straight and level flight, it is safe to conclude whatever occurred to create this abnormal flight situation was beyond their control; not that it was beyond their competence.
The “co-pilots” weren’t necessarily young or inexperienced. At least one and probably both were captain qualified. Either may have been more experienced than the “captain.”
Retired Pan Am pilot with 17 years on 747 and only 300+hrs. on the A-320. Two experiences taught me to not speak too quickly. Mod/severe turb. will shake the a/c so badly you can’t read the instruments. At least the 747 had a ADI that was large enough that even in severe you had some idea wheather you were right side up or not! After Pan AM went belly up I flew a couple of contracts with Korean Air and had the chance to fly the 747-400. Great airplane, but the Boeing rep told me on the q-t “stay out of thunderstorms if at all possible”. He didn’t have to tell me that- I tried my best to not enter the “shag nasty” if at all possible. In my experience, all computer controlled A/C have a punch button on the control stick which cut out the “fly by wire” and gave you manual control if needed. (much the same as the auto-pilot release sw.). Like Brian says, the computer controled airplanes all seem to have their own personality and some are quite “sneaky” in their individual characteristics. I learned not to allow the autopilot to stay engaged after an auto landing in a 400–we got down to aprox.40kts and the damm thing tried to go around! It had been a night flight and my breakfast that morning consisted of a stiff drink and a laugh or two with the first officer! I love airplanes but, like women, there are some you can’t trust!
Sorry Dain, but you are mistaken. Both the 767 and 747-400 have limited “fly by wire” (on these A/C there is a “punch sw.” that allows reversion to manual control but you are still only moving push/pull systems on the central control actuator. Even if you punch and hold the “manual reversion switch” the computer won’t accept the command until your stick position is loosly coordinated with the control position indicator. This will keep a overcontrol condition from “breaking” something you might need!
I have been flying on Air France several times a year for almost 50 years. I have entrusted the lives of my children to flying alone on Air France back and forth between France and various destinations in the US. I always justified the higher cost of flying Air France with the notion that it was a highly professional airline. Now I am thinking otherwise. There are other things that make me nervous. Once I remember seeing the Air France pilot of a 777 going through security with his crew at LAX (they flew my kids back to CDG). The pilot flicked off his loafer, failed to catch it, it sailed up in the air, and came down — you guessed it — right on top of a TSA security guard. It didn’t do much to increased my confidence level. More little incidents like that spring to memory these days, as I wonder whether I should rebook my daughter’s flight this summer from Air France to … oh, wait a minute … who can we trust nowadays? A US carrier? Ha!
If an EMP was the cause for the computer flight controls to fail and not pilot error, would we ever find out or told the truth?
If you are a pilot you know how many times the crew never resets the AHI at the gate or at altitude to display level flight. Still they had over 3 minutes to figure it out with airspeed but you and I were not there. Very sad.
You may be right about the spin. I have done an approach to stall in the A300 (in the simulator and knowing it was coming) without aggressive nose down attitude and max power it does not recover, it will porpoise, wings level in a flat attitude at a rate of 2-4 thousand ft/minute and even with full power and a few degrees nose down we regularly lost an average of 4000-6000 ft. before recovery. Then if it isn’t done correctly a secondary stall is a possibility.
Hard to believe you say? What about that birginair 757 incident. Also lost the IAS.
The captain nearly stalled and all of the sudden reduced thrust, the aircraft plunged down.
The copilot was shouting ADI ADI!!!!
WHY DIDNT THE CAPTAIN LISTEN? and you know why? He was “OVERWHELMED” Gtfo, those shouldn’t be in the cockpit, and probably the same thing happened to those dudes at the A330.
Dont you pilots get to trained on how to recover from a stall withouth the IAS and on IMC conditions? Seriously.
Regards,
An airbus enthousiast and an active and skilled flight simmer.
Boeing planes all allow direct pilot control to override the ‘fly by wire’ features. They are mechanically connected. Even in the 787. .. someday it will fly on commercial routes.
when the only connection to the flight controls is thru a computer, can anyone say for sure what actually happened at the flight control surface?