No matter what the AF447 air crash investigation discovers, there are already implications for airlines.
They are going to have to reconsider how pilots manage the highly computerised flight control systems of all types of passenger jets.
Or according to some very experienced Airbus and Boeing pilots, to reaffirm the need to fly the plane first, rather than trouble shoot the systems.
Many airlines are run by managements that no longer include anyone with piloting experience. They tend to see flying their fleets as an activity for which a semi-automated set of procedures, aided by computer intensive cockpits, can be bought as a solution off the shelf in Seattle or Toulouse, rather than as a human resource that can destroy brand value and reputation if it isn’t properly utilised.
Airline managements are about as connected to their pilots as many newspaper companies are to the journalists they call content providers.
This is without prejudice to Air France, or Qantas, which is a major A330 user and an intensely interested party in the incomplete multinational investigation being lead by the ATSB into the QF72 emergency, when an aircraft made an emergency landing at Learmonth in WA on 7 October 2008 after serious malfunctions affected its flight control functions.
For all that anyone knows, the pilots on the Air France jet were doing everything they could as a team to retain physical control of a jet that was in diabolical trouble for reasons that are far from being fully understood.
But the questions about how high technology jets are best flown has never gone away and a bad accident always brings them prominence.
These questions are of increased relevance because the most computer reliant airliner yet built, the electric-plastic Boeing 787 Dreamliner, now seems close to making its first flight after many delays.
Those Dreamliners, of which 800 are on order will be followed by a slightly larger capacity Airbus family, the A350s. These are also designed around the latest iteration of the concept of pilots managing computerised flight systems, which the European consortium pioneered more than any other manufacturer with its A320 single aisle family that has now been in service for 20 years.
Airbus itself emphasised the importance of flying the plane first in the bulletin it send to all A330 users less than two days after the Air France A330-200 crashed into the mid Atlantic ocean — after sending a puzzling stream of service messages that appear to document the aftermath of events that caused a serious deterioration of systems and controllability of the jet in its final minutes early on 1 June.
Boeing has also referred in detail to this need in the design of the so-called gauntlet tests which the prototype Dreamliner completed a day ago. In those tests the jet was spoofed into thinking it was in flight while sitting on the ground with engines running and electrics on, testing the failure scenarios that pilots could encounter by shutting down various combinations of systems to validate the responses required on the flight deck.
The airlines and safety authorities need to ask if the culture of professional piloting is being degraded inadvertently by an automated approach to operating procedures, or unduly influenced by pressure to fly the most direct route available, meaning closer to severe weather than would have been considered prudent 25 or more years ago.
It is true of most airlines that a pilot who makes a 200 kilometre detour to avoid a belt of storms will be grilled about his or her decision by airline management. It is a process that can intimidate career minded pilots into flying places where traps can lie hidden in the radar imagery of a cluster of storms.
In part of an immense debate on these issues in the comments section to Tom Vasquez’s Weathergraphics site one A330 pilot says of his Air France peers:
I would think that they could have picked their way through the line with just some moderate turbulence. I will comment however, that there seems to be increasing pressure to deviate as little as possible from the proposed flight plan, and never to do so without a prearranged clearance (which takes time) unless the captain declares an emergency. The tracking system aboard the latest aircraft automatically reports the slightest deviation to oceanic ATC, and the captain will be explaining himself or facing violation if that occurs. Perhaps this incident will result in giving a little more latitude to the flight crews for weather avoidance without consequence.
It needs emphasising that this may prove irrelevant to the fate of AF447, but the questions merit discussion.
At last count 41 bodies have been recovered from the crash zone, and US and French naval equipment which can locate and recover the all important ‘black box’ flight data and cockpit voice recorders will arrive at the scene between tomorrow and Friday.
I have worked in the Industrial Automation Industry and have developed a healthy distrust of computers and allowing them to control anything mission critical. Even after extensive validation of code things that make one say “I wonder what happened then” occur. Why then do we think that it is OK to let a computer, albiet a very well tested and redundant systems computer fly an aircraft. I also refuse to rely entirely upon GPS navigation and other forms of FRED’s (that is Fairly Ridiculous Electronic Device’s)!
Consistently excellent reporting on this story Ben, and on aviation in general.
Ben, I think it’s interesting to hear that experienced pilots (among others, no doubt the mainstream press will jump on board too at some point if AF447 turns out to be due to tech failure) are critical of the use of high-tech fly-by-wire flight control systems. I’m not a pilot and I’m not in the industry, but surely these systems
a) have directly prevented infinitely more incidents, and saved far more lives, than they’ve cost; and
b) are inherently more reliable, all things being equal, than humans, notwithstanding that they may still be capable of failure .
Two recent (and excellent) articles in Vanity Fair by William Langewiesche (www.vanityfair.com/magazine/2009/01/air_crash200901) and (http://www.vanityfair.com/style/features/2009/06/us_airways200906) play on this theme of technology in the cockpit, and are worth a read. The latter article discusses the recent A320 that ditched in the Hudson River, demonstrating how pilot skill combined, perhaps crucially, with the Airbus flight control system to allow a flawless ditching of the aircraft that had lost power to both engines, something that may not have been possible without the technology.
Mark, with respect, relying upon a GPS in a car is in a completely different ball park to the type of reliance modern airliners place upon flight control / navigation computers and I’m not sure it’s a valid comparison. You in fact do rely on computers every time you step into an airplane (assuming you fly).
Good point Ben re commercial pressure on pilots. On many occasions in recent years I have endured ‘bumpy’ rides across the Inter Tropical Convergence with different airlines and wondered why some ‘evasive’ action wasn’t taken as after all passengers have been assured for years that modern aircraft are weather radar equipped assuring passenger comfort [and safety]. Well perhaps that is not completely true.
Andrew,
Both Vanity Fair articles were outstanding. And the contribution of the Airbus systems to stability in the Hudson River ditching was foreshadowed by the Mulhouse-Habsheim crash, the very first A320 crash, on 26 June 1988.
In my view of that controversial crash the flight envelope protections kept the wings level when the jet hit the trees, instead of cartwheeling into a stall and killing all on board as would have been the case without them. As a result more than 130 people escaped and only three died. However others disagree, arguing that the computers didn’t allow prompt engine response after the pilot, who had descended to 30 feet and was waving at the crowd, suddenly realised the trees were very close.
AF447 has aroused the debate once more. If there is a rising culture of depending on automated solutions rather than a piloting solution then this event may cause some important changes.
There is a longer article on this in Plane Talking.