Pulling-Up in a 777

If that’s true about 777’s then it’s pretty scary and not what you’d expect – (and quite independent of the bizarre mix of trainee and trainer pilot experience on the flight deck of a commercial flight operation, and whether they were relying on glide-path systems on the ground that were actually switched off for maintenance. Like most major accidents there may eventually appear to have been one fatal mistake, but the situation is inevitably complex). Anyway …

In the aftermath of the Air France Airbus mid-Atlantic loss there was much web commenting about the distinctly different control automation philosophies of Airbus and Boeing. The former moving to greater and greater active / positive automation of aircraft controls – computer systems being more reliable than humans – whereas the latter seemed to have a policy of no-matter how much automation in protective / preventive systems, active controls needed positive pilot inputs. (There is a huge amount of automation in both of course, and it’s a complex philosophical debate about, which kinds of risks are best managed by which kinds of system design responses – a system of systems with humans in the loop.)

In this latest Asiana crash at San Francisco, it seems the experienced pilot is able to issue a “pull-up” command – to the first officer pull the stick back – and rely on the aircraft systems to boost the throttles automatically. As any schoolboy knows, “pulling up” doesn’t pull the aircraft up, it lowers the tail, and the rotation pulls the nose up, increases the angle of attack and slows the plane down, and if you do nothing else – the plane falls. Climbing priorities are all about engine power and speed. Why dumb down the pilot and give him the secondary task – like, don’t you worry your pretty little head there, the systems will look after the most important stuff – mad?

And, talking of complex situations involved in accidents – what about the Lac-Meganic oil-train disaster. How unfortunate that an earlier fire in the one out of five functioning – brake-maintaining – locos, should lead to events that lost braking in the 72 tanker cars that ran-away back into town. Though strangely, that loco doesn’t seem to be amongst the wreckage ? Were the cars decoupled to put distance between the loco fire and the oil cargo – as part of the earlier fire-fight, would seem sensible. How come such trains don’t have brake cars any more – that last wagon looks just like all the others. Seems amazing that the original fire-fight wouldn’t involve rail-operator expertise. More to this story yet. [Post Note – yep, it was a rail engineer that decoupled and – apparently – failed to set the tank-car hand-brakes effectively. Still baffled however. The train had passed through Lac-M already and was parked for the night west of the town at Nantes, so presumably the locos were at the front, west of the cars, so that the cars were able to roll back east to the accident ? Baffling because the rail line comes from Dakota, passing north of the Great Lakes with full oil-cars bound for the St. John Refinery in New Brunswick – the opposite direction to the arrangement described . Were the contents naphtha product or some such going somewhere else in the opposite direction ? Or had there been more complex shunting operations involved in moving the loco(s) out of the way of the cars? Will need to see a more complete report. [2014 Update]].

Ironic also, to be writing this the morning after “Fire in the Night” the BBC televised documentary on Piper Alpha. Excellent programme. Focussing on the survivors, what it took to survive, and what they experienced in their own words of those around them who didn’t. Whatever the tragic circumstances that led to the event, from which industry has learned many lessons, there is a lot worth learning about humans.

And when it comes down to it, humans rely on humans, not objects. (Must look at the recent Arizona fire-fighting team tragedy in that context too.)

[Post Note – thanks to Smiffy on Facebook for pointing out links to Professional Pilots Rumour Net on the topic of the very interesting problem with the Boeing 777 auto-throttles, particularly when converting from Airbus, the trainee / trainer pilot relations and more, including one plea that pilots be allowed to be pilots, not systems managers.]

[Post Post Note – and this Spanish (Santiago de Compostella) rail crash. Sure it was going too fast, and sure some component has to fail first first under the lateral rolling loads on the carriages that experienced the loads earliest, but why was it going too fast, is a wider systems engineering question, a system involving the two (?) guys in the cab – and it seems a system of transitioning between multiple different systems – hmm. (Incidentally, multiple people on the bridge was a feature of the Italian cruise ship crash too?)]

[Post Note : More on “complexity” of automated controls – Asiana 777 crash report.]