The riddle of Flight AF447
As bodies of crash victims are recovered from the south Atlantic, John Lichfield examines what we know – and may never know – about the fate of the Air France jet
Ten days after flight AF447 vanished over the south Atlantic, the mystery of what happened to the Air France Airbus and the 228 people on board is starting to become clear. Perhaps. All the indications are that the calamity was influenced – if not entirely caused – by a malfunction of the three tiny tubes in the nose of the plane which measured its speed.
Air France voiced its concerns to Airbus months ago about the reliability in icy conditions of the model of speed sensors, or "Pitot tubes", fitted to the A330-200 which crashed 400 miles north-east of Brazil.
The French national airline, after reassurances from Airbus, started in April an unhurried programme to replace the tubes with an improved version. Air France, under pressure from its own pilots, yesterday ordered that at least two of the three Pitot tubes on all its 34 remaining A330 and A340 aircraft should be upgraded immediately.
On Monday, one of the smaller Air France pilots' unions told its members to refuse to take the controls of any Airbus A330 or A340 which had not had its speed sensors changed.
But how can the malfunction of something so simple and basic as a speed sensor explain how a hyper-modern aircraft, only four years old and packed with sophisticated control and communication equipment, could fall from the sky without broadcasting a mayday message?
The aircraft issued 24 automatic, emergency signals in the space of a few minutes in the early hours of Monday 1 June, indicating the collapse, one by one, of most of its electrical and computer systems. Can that cascade of mechanical disasters all be traced to a problem with speed sensors?
Initially, the aircraft was said to have flown into a severe storm and to have "probably" been crippled by a lightning strike. At the weekend, the French meteorological office said that the weather on the night of 31 May and 1 June, while turbulent, was fairly normal for the south Atlantic.
The Brazilian authorities added to the confusion early last week by announcing that they had found debris from the missing aircraft, and a large slick of aviation fuel. The location of this wreckage was nearer to the Brazilian coast, and further south, than the last known position of the Airbus had suggested the crash site would be. Had the aircraft tried to turn back to Brazil? If so why had it broadcast no mayday message? The presence of the aviation fuel also appeared to rule out the possibility of a mid-air explosion and, therefore, of a terrorist attack.
The shamefaced Brazilian authorities admitted on Friday that they had spoken too soon. The wreckage and oil they had found came from unknown ships, not from the Airbus. Within a day, genuine pieces of Airbus wreckage – and by yesterday 24 bodies and part of the tailplane – had been recovered to the north-east, along the plane's scheduled flight path toward Senegal and onward to Charles de Gaulle airport in Paris.
French investigators now say that terrorism has not been ruled out, and was never ruled out, but "all the indications" remain that the Airbus suffered some kind of catastrophic equipment failure.
Airline disasters are meat and drink to conspiracy theorists. Several alternative explanations still exist for the Concorde disaster in Paris in 2000 and the terrorist bombing of a Pan Am jumbo over Lockerbie in 1988. The confusion and misinformation surrounding Flight AF447 will inevitably lead to similarly fevered speculation.
French investigation sources said yesterday that such confusion was difficult to avoid. "You have a three-way muddle here," one official said. "There is the understandable desire of the media and public for a rapid explanation and the unfortunate fact that most of this aircraft has vanished into deep ocean. On top of that, there are extremely complex, technical questions involved which are difficult for even experts to grasp fully."
A French nuclear submarine, the Emeraude, and a naval vessel containing robot submarines should reach the crash site within the next two days. Investigators hope that the electronic equipment aboard the nuclear submarine will allow it to pick up the radio signals from beacons attached to the aircraft's black boxes, or flight recorders, which could be up to 15,000 feet below the surface. Without the records contained in those boxes, it might be impossible to be certain about what happened to Flight AF447.
For the time being, attention is on the aircraft's speed sensors.
In November last year, Air France issued a warning note to its pilots about a "significant number of incidents" and "anomalies" with the functioning of the Pitot tubes, or speed detection devices, on A330 and A340 aircraft. The tubes, fitted in or behind the nose, measure dynamic air pressure against the fuselage and indicate the speed to the human pilot and co-pilot but also, crucially to the automatic, computerised "fly-by-wire" systems which take over the piloting of modern aircraft.
The previous "incidents" on Air France Airbus craft, starting in May last year, suggested an earlier model of Pitot tubes fitted to some planes could seriously malfunction in icy conditions. The problems included wrong, or fluctuating, indications of speed; the automatic switching off of the automatic pilot and even a false warning that a plane was about to fall from the sky. In the early hours of 1 June, when Flight AF447 was 400 miles out into the Atlantic, one of the first of the 24 automatic distress messages that it broadcast to Air France headquarters indicated that the speed of the aircraft was fluctuating wildly. Soon after, the computerised "fly-by-wire" system was switched off – possibly automatically, possibly because the pilots were wrestling to save the plane. At that point its fate and that of its passengers seem to have been rapidly sealed.
Still, French investigators say that they are not fully convinced that the speed sensor problem is the main explanation for the disaster. The jigsaw puzzle is incomplete, they say.
However, senior French pilots point out that the correct speed is crucial to any aircraft's capacity to fly, especially in turbulent conditions. "Speed is a fundamental value for piloting a plane," said Patrick Magisson, a member of the technical committee of the main French pilots' union, SNPL. "It is speed which makes an aircraft fly. It is speed which determines its aerodynamic balance."
In other words, the wrong speed – either too slow or too fast – can be calamitous. If false speed readings are proven to be the cause of the Atlantic crash, both Air France and Airbus will have awkward questions to answer.
View all comments that have been posted about this article.
Offensive or abusive comments will be removed and your IP logged and may be used to prevent further submission. In submitting a comment to the site, you agree to be bound by the Independent Minds Terms of Service.
- Print Article
- Email Article
-
Click here for copyright permissions
Copyright 2009 Independent News and Media Limited


Comments
The reliance on a computer to fly an aircraft will eventually be found to be the reason. One only has to consider the recent accident in Australia where someone programmed the wrong weight into the computer resulting in the aircraft not being able to take off safely and competently almost causing a crash.
If the Airbus relies so heavily on the Pitot tubes for reading the air speed and if the air speed is so critical to the aircraft's safety how was it ever given a safety certificate to fly when the pitot tube is known to be affected dirt debris ice etc? (I expect that a computer somewhere said it was OK.)
I think you'll find that all aircraft rely on pitot tubes to measure airspeed, whether or not the aircraft is fly-by-wire or uses computers. They are very simple pressure instruments, in common use for 100 years, normally very reliable and accurate, and they are heated to prevent icing. However, they can ice up and here is an account at http://aviation-safety.net/database/rec
The bizarre thing is that most icing-related incidents occur around take-off and landing -- when the plane is travelling near stall speed, and has critical pieces (landing gears and flaps) extended, that are prone to damage at higher speeds. If the plane is flying in a storm, surely the pilot can play it safe and fly at high trust.
Also if there is a sudden blast of turbulence or wind are the jet engines powerful enough to quickly increase power and stabilize the plane? Considering such a big plane must be flying at quite a speed just to stay in the air, even if its going slower than optimal speed it must still be moving well over the speed of winds and airstreams at that altitude, so I wonder how great the turbulence or crosswind must be to make it lose complete aerodynamic balance as is suggested. Judging by how hard so many presumed ex-airforce pilots land commercial airliners, it seems that these big planes can be handled quite roughly without losing control. The leaf in the the wind analogy does`nt fit the impression I have of these transcontinental heavyweights.
Regarding pitots and fly by wire... as somebody pointed upthere, yes, wrong speed reading will affect actually most of the system, but the critical thing here is the fly-by-wire... why: well, modern comercial aircraft fly pretty close to the Transonic Regime (0.82 Mach) this is to say... if you push the pedal a bit more, you'll get in to supersonic regime... ofcourse, comercial Aircraft are not designed for supersonic... if they dare to venture themselves in to the vecinity of Mach1 (speed of sound) they lose control. flight stability is lost (handled by fly by wire) and the A/C goes in to what is called the high speed stall... the aircraft begins an accelerated free fall which not necessary ends up on the ground (or sea) but with the airframe braking in to bits due to high speed. All this horrific things are prevented by the fly by wire system, by having the right air speed inputs.
first was the weather and the decision of the crew to penetrate a line of severe storm cells - this could have been the result of misleading weather radar data or the failure of the crew to properly analyse the display or simply the fact that the storm was developing and changing too fast for the crew to keep up...
second was the potential failure of the pitot system to provide reliable speed data to the autopilot, which has been well documented, such failures have been cited in numerous incidents and the inability to measure airspeed makes it impossible for the autopilot to keep the aircraft flying in severe turbulence and also makes it extremely difficult for the crew to recover the aircraft once control is lost...
third is the design of the Airbus family which takes much of the decision making control away from pilots and relies on a set of complex systems to manage the flight - this has the potential effect of delaying the crew's awareness of a critical and deteriorating situation until it becomes too late to correct...
finally is the potential structural weakness of the A300 and A330 vertical stabiliser (tail fin) which has been a factor in a previous fatal crash...the fact that the stabiliser was found largely intact suggest a catastrophic and sudden in-flight break-up due to a structural overload the onset of which was likely due to the above cited contributing causes...if the stabiliser is determined to have separated from the aircraft at the beginning of structural failure, this may be a fundamental design flaw in this family of aircraft...
obviously this is all speculation until the flight data and cockpit voice recorders are recovered...however I would say that there is enough information to warrant a serious look at all these issues concerning the vulnerability of all modern aircraft when cruising through severe weather...
caveat: I am not a pilot nor an engineer but I have long experience in the aerospace industry and an interest in flight safety...
It is because flying is so safe that every individual incident becomes a significant test case for the entire system: crew, aircraft, company policy and support systems such as meteorology.
So, let's not let a rush to defend the manufacturer get in the way of our search for the whole truth. I don't want my family to be exposed to any preventable cause and I'm sure you feel the same way.
FDR externally as well to avoid situations like these? (smartest reader!)...
There seems to have been no second system apart from the Pitot Tube to establish airspeed. Despite the fact that Airbus must use GPS to establish position. Thus there has been a gigantic failure of system architecture that the highly computerised facilities of the plane were not used to combine position with time to calculate airspeed net of wind conditions. Even now it is not too late to issue a new software version which will do precisely that. And high-level personnel should also examine which other potential future Elephants are lurking in Airbus's Cupboard.
Mr Alex Weir, ex-Rolls-Royce Aero Engines, Harare, Zimbabwe
Anyone know the safety record of Boeings vs Airbuses? No? Why not? Why does them media feast on air crash stories but invariably fail to give us the most important piece of information about them: Air safety statistics? Ever wonder why they aren't published and updated after each crash? Could it be because it would make the Yanks' plane maker look like Tupolev?
Fatal Event Rates for Selected Airliner Models
Fatal Event Rate Per Million Flights
Your statistics are selective and somewhat misleading.
Readers should look at http://www.airsafe.com/events/models/ra
Note that many of the figures given are ESTIMATES, whilst others are way out of date.
Flying a huge aircraft at high altitude in clear calm weather is a complex enough task. Factor in really lousy weather and dodgy mechanics, and it can quickly become a marginal activity.
To get a vague idea of the level of complexity, for example, see
http://en.wikipedia.org/wiki/Coffin_cor
Fortunately most plane drivers are extremely well qualified and highly experienced -- and, to be honest, they are probably underappreciated and taken for granted.
I'm not one, by the way.
But, the next time you get on a budget flight, just think of the guys up front and meditate on how little they are probably paid to do the job, and give a silent vote of thanks.
Since the AF447 pilots are sadly no longer with us, get ready to read an accident report that follows tradition and lays a good part of the blame for the crash on pilot error.
Pilots refer to each others as drivers, as it happens. Go read PPRUNE or one of the other airline professional sites before you post something else that you know nowt about.
If that's all you can come up with to criticise my post, that's pretty foolish. Apart from your ugly sexist claptrap, of course, which the moderators should ban from this web site. I'm actually a rather muscular bloke who isn't impressed that a troll thinks that my nose is cute.
Selective quotation of estimated statistics in support of your argument is otherwise known as lying. OK, I was trying to be overly polite. I only wished to demonstrate that what you write can't be trusted.
posting simple statistics is an asinine obfuscation of the real issue - we first need to isolate the causes of fatal accidents to make any comparison between aircraft
since airbuses are much more recent designs than most of the other models it's only logical that they would have a lower loss rate, benefiting both from what was learned from previous designs and also benefiting from a progressive reduction in fatal crashes DUE TO ALL CAUSES
we still need to isolate and correct any current, preventable causes, yes? or, are you arguing against that?
Calculations and algorithms correlating Groundspeed (from GPS), Airspeed (from Pitot Sensors), instantaneous Fuel Consumption (from instruments), and Altitude changes (again from GPS and/or sensors) should be able to indicate whether or not the Pitot Sensors are accurate, and possibly they could provide an airspeed recalibration estimate which might at least get the plane back to base or to its destination without crashing.
Such a mini-system or system patch could be described as a Heath Robinson measure but it might alert pilots and ground control to the existence and/or development of problems and it might also save lives.
Mr Alex Weir (ex-Rolls Royce Aero Engines), Mechanical Engineer and Software Developer, Harare, Zimbabwe
Also, why doesnt the black box data get streamed up in real time via a sattelite phone to the airlines control centre? Sort of like F1 cars have their telemetry constantly monitored.
Neither of these ideas seem very expensive.
Oh, and while I write, why not two black boxes, one that is ejected before impact and floats, one that stays with the wreckage and sinks?
Until the flight recorders are found we can only speculate.
This article is presented as a summary of what's known, but is nothing more than completely unfounded speculation.
At the moment nobody knows whether or not the pitot heads had anything to do with it.
At the moment, nobody actually knows anything.
If there isn't capacity then spend a bit of money devoted to some arcane military satellite to allow this. Even a 56kbit data stream would be far more info than required to piece together almost any accident and very quickly as well. Even CVRs could be compressed and sent up as well.