Air France 447 Update on Investigation
The Bureau d’Enquêtes et d’Analyses today released an Update on Investigation regarding flight Air France 447. AF 447 was an Airbus A330-203 on a scheduled flight from Rio de Janeiro Galeão to Paris Charles de Gaulle on the 1st of June 2009. Contact with the flight was lost in the early hours of the morning. A week later, a few pieces of the aircraft were found.
The crash made international headlines because the aircraft appeared to simply disappear: there was no mayday call nor radar information and initial attempts to find the crash site and wreckage has failed.
Last month, an expedition ship, the Alucia, discovered the wreckage at the bottom of the South Atlantic Ocean and was able to recover the black boxes. This Update on Investigation is a result of the data recovered. It is not a final report.
The BEA have released an English translation of the document which I am using for this post. However, it includes a foreword to state that the original text in French should be considered the work of reference.
You can read the full document online: Update on Investigation in PDF format. The French original is also available: accident survenu a l’Airbus A330-203. The report is also available in German and Portuguese translations.
I’ve written this to help people keep to the facts in mind when reading news coverage, which traditionally gets somewhat confused and over-excited when reporting on this incident.
Here are the highlights of the document, hopefully in an easy-to-follow format.
The report confirms that there were no initial issues with the crew nor the weight and balance of the flight.
22:29 The flight took off.
Initially, the Captain was Pilot Not Flying and one of the co-pilots was Pilot Flying which was in accordance with standard procedure. During the flight, the Captain took a rest break and the other co-pilot took his place as Pilot Not Flying.
This is important to understand as I’ve already seen media headlines screaming that the Captain was not in the cockpit when the problems began. This is a non-issue. There were three flight-crew members on board, all fully trained, in order to ensure that the pilots can have rest breaks.
SKYbrary – Pilot Flying and Pilot Not Flying
When two pilots fly a fixed-wing aeroplane which requires a two-person flight crew, the aircraft commander, who must be appropriately qualified and hold the rank of Captain, occupies the left hand seat and the First Officer or Co-Pilot occupies the right hand seat.
Before the commencement of each flight sector, the aircraft commander decides which pilot will take direct responsibility for flying the aircraft for the complete flight or for particular parts of it such as the Descent/Approach and Landing and they become ‘Pilot Flying’ (PF) for that sector or the specified part of it. The other pilot is then designated for that sector or relevant parts of it as ‘Pilot Monitoring’ (PM) or alternatively as ‘Pilot Not Flying’ (PNF) and in that role must monitor the flight management and aircraft control actions of the PF and carry out support duties such as communications and check-list reading. The Operations Manual will specify fully the roles for the PF and PM/PNF, but one of the most important aspects of the duties of any PM/PNF is the cross-check of the actions of PF. Indeed, this part of the role represents one of the most important single reasons why a two-pilot flight crew is specified.
01:35 The crew were in contact with air traffic control. Everything seemed normal.
01:55 The Captain woke the second co-pilot to take his place as Pilot Not Flying.
Between 1 h 59 min 32 and 2 h 01 min 46, the Captain attended the briefing between the two co-pilots, during which the PF said, in particular “the little bit of turbulence that you just saw […] we should find the same ahead […] we’re in the cloud layer unfortunately we can’t climb much for the moment because the temperature is falling more slowly than forecast” and that “the logon with Dakar failed”. The Captain left the cockpit.
At this stage, everything was normal. Ten minutes later, the incident begins.
02:06 The Pilot Flying warns the cabin crew about upcoming turbulence.
02:08 The crew make a slight turn to the left. The turbulence increased and they reduce the speed to about Mach 0.8.
At this point, it’s worth watching the timeline including the seconds.
02:10:16 The autopilot and then the auto-thrust disengages. The Pilot Flying says, “I have the controls.” The plane begins to roll to the right. The Pilot Flying rolls left and raises the nose. The stall warning sounds. The plane appears slows right down.
The recorded parameters show a sharp fall from about 275 kt to 60 kt in the speed displayed on the left primary flight display (PFD), then a few moments later in the speed displayed on the integrated standby instrument system (ISIS).
The autopilot and auto-thrust remained disengaged for the rest of the flight.
Now in my opinion, the notes in the Update on Investigation are very interesting:
Note 1: The angle of attack is the angle between the airflow and longitudinal axis of the airplane. This information is not presented to pilots.
Note 2 : In alternate or direct law, the angle-of-attack protections are no longer available but a stall warning is triggered when the greatest of the valid angle-of-attack values exceeds a certain threshold.
Alternate Law means that there is a disruption to the aircraft flight control system(s). As the data input could not be trusted (the speed displays in this case), some of the monitoring protective functions of the control system are not available.
It’s interesting that this is pointed out: the angle-of-attack was critical to the incident and that the Airbus A330 does not display this information as flight information for the flight crew.
At any rate, the angle of attack increased and the plane began to climb.
The recordings show that the Pilot Flying made nose down control inputs. The Pilot Not Flying attempts to call the Captain back to the cockpit.
The vertical speed, which had reached 7,000 ft/min, dropped to 700 ft/min and the roll varied between 12 degrees right and 10 degrees left. The speed displayed on the left side increased sharply to 215 kt (Mach 0.68). The airplane was then at an altitude of about 37,500 ft and the recorded angle of attack was around 4 degrees.
If I understand this correctly, the speed displayed on the left side increased but the integrated standby instrument system had not updated yet, so would have shown 60 knots. I’m not quite clear on this though.
02:10:51 The stall warner triggers again.
The thrust levers are moved to the Take Off/Go Around position. The Pilot Flying maintains nose-up inputs on the controls. The recorded angle of attack continues to increase.
Around fifteen seconds later, the speed displayed on the ISIS increased sharply towards 185 kt; it was then consistent with the other recorded speed. The PF continued to make nose-up inputs. The airplane’s altitude reached its maximum of about 38,000 ft, its pitch attitude and angle of attack being 16 degrees.
Note: The inconsistency between the speeds displayed on the left side and on the ISIS lasted a little less than one minute.
02:11:40 The Captain re-enters the cockpit. The report says that the recorded speeds “became invalid” which means that the measured speeds were shown as below 30 knots. The stall warning stopped.
The altitude was then about 35,000 ft, the angle of attack exceeded 40 degrees and the vertical speed was about -10,000 ft/min.
This means the plane was falling out of the sky. The Pilot Flying holds the sidestick to full left and nose-up position for around 30 seconds.
There is a lot of speculation that this is what caused the crash. Based on the information given here, it would seem that the plane was in a stall and the nose needed to come down. Pitching the nose-up to the full extent would be a bad response in a stall situation.
However this is a preliminary report and no such conclusion is drawn in the report. It is not stated in the report that the crash was caused by this action, simply that the action occurred.
02:12:02 The Pilot Flying says, “I don’t have any more indications” and the Pilot Not Flying responds with, “we have no valid indications”.
By now, the thrust levers are in the idle position and the engine power is at 55%.
The Pilot Flying pitches the nose down, decreases the angle of attack and the speeds become valid again. The stall warning sounds again.
02:13:32 The pilot says “we’re going to arrive at level one hundred”. Fifteen seconds later, both pilots manipulate the sidesticks and the Pilot Flying passes control to the Pilot Not Flying.
02:14:28 The recordings stop.
The document includes a list of new findings. The penultimate point is chilling:
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 information in this report would seem to imply that the pilot became overwhelmed and incorrectly interpreted the situation. I recommend reading the full Update on Investigation. The initial problem remains the pitot and the bad speed data. The pilot was handed control of a plane (when the autopilot and auto-thrust failed) that was incorrectly configured. There are many aspects to any incident and AF 447 is no exception. It is possible that the BEA will conclude that pilot error was a primary cause but, as yet, we do not have a full report.
So remember: be wary of news reports that state any cause as fact until the BEA release their conclusions!
can I have permission to link to your article on my blog site? It gave a great simple summary and allows me to talk about my thoughts without filling the whole blog article with the facts of the crash.
Steve: Absolutely, you can. :)
Thanks, here is my article with a link to yours
Am I confused here or is there an alteration in the speeds based on altitude or something because One Knot equals 1.14 miles per hour (and I’m suing North American/US speeds, not kilometers), but the point is the speeds noted here are inaccurate.
We know that the speed of sound is over 700 mph (as a general rule or from http://en.wikipedia.org/wiki/Speed_of_sound “In dry air at 20 °C (68 °F), the speed of sound is 343.2 metres per second (1,126 ft/s). This is 1,236 kilometres per hour (768 mph), or about one kilometer in three seconds or approximately one mile in five seconds.”)
And here the notes (and I’ve not read through the reports yet, but will) put mach .68 as 215 kts x 1.14 = 245 mph (which with further calculation, would make the speed of sound merely 360 mph, which we know it in fact is not).
NASA has created a speed of sound at altitude calculator to make this easy for us at:
Using what is given here: “The speed displayed on the left side increased sharply to 215 kt (Mach 0.68). The airplane was then at an altitude of about 37,500 ft” ANd then plugging this as a speed of 245 mph into the calculator, we show a speed of sound of actually 660 mph and a mach of .371
Regardless, this was an absolutely horrid tragedy and initial reports made it out that the plane came apart in mid-air in heavy turbulence, but these reports show it was entirely because of pilot input at the controls–and I say this with no degree of expertise, but a stall configuration and holding the controls there and the nose up instead of down to pull out of the stall, created a resounding hold of the stall and thus the crash.
Thank You, regardless for putting this information out here.
Douglas, you are absolutely right in your calculations. I wondered if it was a transcription error but the original French report says:
So the error is absolutely there in the original. My guess is that the 215 kt is correct and simply the Mach number is wrong. But it’s interesting.
Excuse my typos btw…I’m a horrible typist and see a couple gross mistakes of my own…for example when I wrote “suing” instead of “using.”
the airspeed indicator reads exactly that, airspeed.
The speed of sound changes as a function of temperature and mach number is a ratio of dynamic and static pressure. Using my crp5, at isa temp at 36000 feet this gives 390kts true air speed, adjusting for compressibility this gives 220kts indicated, so it must have been colder than -56.5c which is consistent with supercooled water droplets freezing the pilot static ports which caused the asi to read incorrectly anyway. See my blog on frozen asi, i was travelling at 200mph but the asi read zero bc the port froze.
What’s the URL for your blog?
I spotted an article that suggest that the first officer was focusing on the flight director telling him to pitch up:
This would go some way into explaining why the aircraft went down to such a slow speeds without the pilot doing anything. If the FD is telling the pilot to go one way and the speed is telling him to go another, then I can see where the confusion might come from. I can imagine that some pilots might use the FD in the misguided notion that it could get them out of trouble.
Really I think the Flight Director should turn off in a stall, or maybe show a pitch down attitude, or change colour, or SOMETHING!
I tapped AF447 sidesticks into Google today to see if there was any sign yet that Airbus were eliminating the way that interaction between inputs could contribute to accidents like this. I found nothing but did come across this very insightful report on the unwanted effects of automation.(see below) As it says in the article, as we work towards autonomous vehicles we appear to be blindly ignoring the adverse effect that increased automation has on pilots and car drivers concentration. Eg. if the need to focus our attantion on the vehicle in front as we drive is removed by adaptive cruise control which brakes and accelerates for us, to my mind this is dangerous! Insurance companies and manufacturers however would have us believe it is a safety enhancement!
I tapped Airbus sidesticks into Google today to see if there was any sign yet that Airbus were eliminating the way that interaction between inputs could contribute to accidents like this. I found nothing but did come across this very insightful report on the unwanted effects of automation.(see below) As it says in the article, as we work towards autonomous vehicles we appear to be blindly ignoring the adverse effect that increased automation has on pilots and car drivers concentration. Eg. if the need to focus our attantion on the vehicle in front as we drive is removed by adaptive cruise control which brakes and accelerates for us, to my mind this is dangerous! Insurance companies and manufacturers however would have us believe it is a safety enhancement!