Industry Opinion

« back to blog listings

Cascading computer - the mystery of AF447

It was only in the last issue of Aerospace Testing International that I wrote about a recent air crash. The US Airways A320 that ditched in New York’s Hudson River prompted a vitriolic response from some Boeing pilots, who ranted about whether Airbus aircraft are technically flawed. Whatever theories people have about the incident, it is a courageous tale of a captain who glided the plane to safety, ensuring that all who flew survived.

   
The disappearance of Air France 447, meanwhile, is shrouded in an air of mystery that sets it apart from other airplane disasters. Nearly all air crashes take place around the time of take-off or landing. But the Airbus 330 came down four hours out of Rio de Janeiro, over the Atlantic Ocean, on the edge of a tropical storm. There was no distress signal; it just disappeared off the radar.  

There has been an enormous amount of press speculation about the cause of the crash, and by the time you read this, even more ‘experts’ will have thrown their speculation into the maelstrom. The conjuring up of theories is rife and I really don’t want to buy a first-class ticket onto the bandwagon, but… In this issue there is a feature on wire failure (p36), the ensuing hazards, and a direct link to the crash of TWA 800 in 1995. At this point, and I am saying this very tentatively, there could be a correlation between the theory of lightning, stalls, weak wiring, and computer failure. (So here I am bandwagon ticket in hand…)

There has been so much debate, covering topics from terrorists to icing updrafts, that it would be stupid to list all the theories about the cause. So I got in touch with a very experienced pilot who was not keen to join me on the wagon, but did say he would forward a comment about the crash from a friend-of-a-friend, a very experienced Airbus pilot from Cathay and Virgin. This is what he said (it’s interesting, and sounds realistic; I have taken out the suggestion that the captain was not even in the cockpit): “The aircraft (an A330) has five computers that handle the operation of the aircraft, with No 1 computer designated as the controlling entity for the other four. If No 1 fails, the controlling task is passed to No 2 and the aircraft remains fully functional. If Nos 1 and 2 fail, No 3 takes over and starts load shedding (by that I mean that the aircraft flight functions start to be downgraded). The aircraft can still be flown with only one computer, but the functions are greatly reduced so you have to know what you are doing. The big fear of course is that you get a cascade failure of the entire computer system – and then you have problems. The aircraft can be controlled, just, with rudders (mechanical linkage), I think possibly tail plane trim, but I am hazy now, and I have a strong feeling one needs the engines to give pitch control.

“During flight testing, in perfect conditions, the chief test pilot did an approach and landing with all five computers turned off but with a bunch of guys with fingers poised over the ‘ON’ buttons. During training in the simulator, captains had to display an ability to keep the aircraft in some form of flight with all computers off while the co-pilot rebooted the computers one by one. From experience I can tell you it is difficult enough in smooth conditions. In the middle of a thunderstorm or any heavy turbulence I would categorically say that it would be practically impossible. From what little I have read, if all five computers went into meltdown, the pilot would be stuffed, with basic standby flight instruments only and a marginally controllable aircraft in the worst possible flight conditions. If any pilot was faced with this scenario I genuinely think total loss of control would happen exceedingly quickly.

“My only conjecture is that the aircraft computers were working overtime controlling the aircraft in extreme conditions, one failed and they went into cascade failure mode and the pilot lost control to such an extent that it fell out of the sky and broke up. Toulouse desperately needs that flight recorder; the last thing anyone wants is an inexplicable incident.”

Several days after the crash, I got in touch with Edward B. Block, who wrote the wiring article in the issue. He said, “The scenario that brought down Air France Flight 447 fits exactly with that of faulty wiring. There have been instances lately related to Qantas aircraft, when the aircraft suddenly plunged without warning. And this is the exact scenario described on the F-14s that had picked up spurious signals from cracks in the wiring insulation. If turbulence shook the wire bundles, causing a pre-weakened spot in the insulation to break, this would lead to a 10,000°C arc-tracking event as happened three times on the Space Shuttle Columbia. If a lightning bolt with millions of amps was sent along the outside skin to be exited at the tail or wing tips, and it came upon a pre-existing crack in the wiring insulation, it would be like a hemorrhage in an artery.” The speculation bandwagon rolls on… 

 

 

There are currently no comments.