I just saw a documentary on National Geographic about Air Transat flight 236 (wikipedia story). This was a flight from Toronto Pearson which nearly ditched in the atlantic following an incident in which a faulty fuel pump caused one of the fuel tanks to empty and the flight to become a gimly glider. The Air Line Pilots Association, International (ALPA) noted in October 2004, “The competence and professionalism the two pilots exhibited in safely landing the aircraft and protecting the crew and passengers under exceptional circumstances is clearly confirmed in the Portuguese government’s report”.
Official news stories focus on the fact that these pilots skillfully managed to glide a plane for over 15 minutes for the furthest this has ever been done before and that they successfully landed. While the recounts also state that the fault was due to a replacement of a fuel pump being botched due to unavailability of the correct part, they don’t seem to adequately note what appears to have been the pilot error factor.
Seriously. If the computer says “you’ve got a problem” then it might actually mean it. In this case, the computer system apparently issued a few somewhat confusing errors (oil pressure was high but the recorded temperature was low in the right engine) which caused the pilot(s) to distrust what it was saying and consequently only believe there was a problem when the engines finally stopped minutes later. What I want to know is, how often do pilots on planes I fly on decide to ignore serious warnings on the flightdeck computer and fly on regardless. How many people reading this would be happy for a plane to divert if there was any suspicion that there could be any potential problem? I think that’s everyone.
This story annoyed me because it appears that a major reason for this incident was the fact that the pilots ignored the computer system. While they saved 306 people on board and were commended for their actions, I think it’s also valid to question why they would ever take it upon themselves to think it’s acceptable to ignore serious safety warnings. What’s the point in writing that software if the pilots refuse to believe the warning? I’m not saying you should blindly follow a computer, but if it thinks there’s a problem, land the plane and check it out properly.
Of course, I’m not completely naive (contrary to what some readers may think, I’m an idealist not an idiot) and I realise that if we actually followed everything by the book then nothing would get done at all. Yet I feel that too much attention has been paid to the gliding aspect of the flight and not enough attention has been focused on whether the Airbus software was worthwhile.
In other news, I returned a loaned copy of a Rolls Royce technical reference I’ve been reading on modern jet engine design. I didn’t get chance to read the whole book but I found bits I did look at to be very enlightening as to the complexity and durability of modern design techniques – from the single crystal fan parts (or multiple even axes, depending upon the usage) to double skinned tubing for fire supression and containment purposes, a lot goes in to a jet engine. This is part of my not trusting something I don’t understand philosphy.
Jon.