Apparently idea that computers exclusively run the AirBus has been debunked:
[...] THE SYSTEM WILL NOT FIGHT THE PILOTS. The system just gives me the best the airplane can do at the moment without me having to stop and consider my conditions.
If I as the pilot deflect the stick left, if the autopilot is engaged, it will be disengaged for me. The load factor is considered and the airplane will begin a maximum-rate roll in the direction I deflected the stick. I keep holding the stick and the airplane will keep rolling up to the limiter. That's when Normal Law is active. If we're in Alternate or Direct law, there is no roll limit and I can roll the airplane onto its back and crash it if I desire. [...]
But this isn't what the article is asserting. They are not saying the autopilot couldn't be overridden. The article is saying that it appears that while flying into a very bad storm the pilots disengaged the AP and attempted to make an emergency, beyond-limits, turn, which the plane 'allowed', and then at some point after that the disaster struck (structural failure, for example). Whatever happened, did so before (presumably very busy) pilots could radio anything.
The article possible alludes to maybe pilots used to making very small changes to the very capable AP might loose their skills over time, but that's like saying cruise control does the same to drivers (and it might, I have no idea)
I didn't get that from the article. From what I understood, the point he was making was that the AP systematically relinquishes control of the plane as conditions deteroriate. By the time human pilots get hold of the plane, the conditions are for too difficult for them to handle. My issue with the article is the implications he makes. He compares the Boeing to the Airbus and then talks about the philosophy behind the Airbus. As I replied to sweetdreams, his statements about the plane (eg. Who's in charge here?) are misleading. It's as if he's saying that pilots are in full control of Boeing, but not in an Airbus.
I don't think he was saying that computers exclusively run the Airbus. He's saying that the failure of the autopilot in a difficult situation leads to a cascade of problems that can be harder to deal with in the more simplified Airbus cockpit.
Yeah, I should have made it clear that I was talking about autopilot. I have read other stories about the disaster that toyed with the idea that the Airbus autopilot cannot be overriden, and that the computers will always decide whats best for the plane.
For example, this statement from his article is misleading:
The computers are programmed with some strict rules (in fact, Airbus calls them “Laws”) designed to assess the human commands from the flight deck – and veto them if they would put the plane in harm’s way. Point the nose too high or too low – or bank to steeply and the computer will correct your bad airmanship. Who’s in charge here?
I used to design nuclear reactor control systems. One of the never ending debates was how much to automate and how much to leave for the reactor operators (or pilots in this case). There is no clear cut answer since with today's technology its possible to automate everything. However, if you over-automate you get a detached, bored reactor operator who suddenly has to take control if things go wrong. On the other had human error is one of the most common causes of industrial accidents and plane crashes. Tough engineering choices.
I don't understand why a thunderstorm would have affected this plane. Wouldn't it have been cruising by that time at an altitude immune to weather? Perhaps this was discussed in another thread; I haven't been following the discussions too closely.
"Thunderstorms are the "tallest" weather phenomena that we experience here on the ground. An "average" height of thunderstorms is 30,000, however, depending on atmospheric conditions (i.e., the height of the tropopause) some thunderstorms can reach as high as 50,000 to 60,000 feet over the central U.S. and in the tropics near the equator. In some stronger thunderstorms, the upward vertical motion in them can "punch" through the tropopause and will actually penetrate into the stratosphere before stopping their upward growth."[1]
"The Airbus SAS A330-200, with 228 on board, appears to have flown into or near a large cluster of thunderstorms northeast of Fernando de Noronha, located off Brazil’s coast, according to AccuWeather.com.
Updrafts may have reached 100 mph, and the storms, stretching for over 400 miles (644 kilometers), towered as high as 50,000 feet, according to the weather service." [2]
Airbus A330 Service Ceiling: 39,370 ft (12,000 m) [3]
According to weather services, the storms at the region towered as high as 50,000 feet, an altitude higher than maximum ceiling that most commercial jetliners can climb.
What I don't understand is, if they lose altitude because of being at the edge of the corner, wouldn't the pilots be able to steady the aircraft at a lower altitude?
You're correct in that being in the "coffin corner" is not dangerous in itself. In straight-and-level flight, if you go any faster, the extra drag will slow you down. If you reduce your thrust from your engines, you will slow down and lose altitude.
The problem comes about if you enter into any unusual attitudes, which is certainly possible in severe turbulence. The most important part would be a safe recovery to straight and level flight. An unusual attitude can put you in a situation where you overspeed the aircraft (e.g. nosedive). Above the maneuvering speed of the aircraft, you must be very careful to regain control without overstressing the airframe.[1] If your speed sensor(s) are giving you or the autopilot faulty readings, it would be very difficult to know the safest way to regain control of the aircraft after being put into an unusual attitude due to turbulence.
[1] American Airlines Flight 587 crash (NY, 2001) taught us that even below the manuevering speed of the aircraft, the airframe can be overstressed due to dynamic loads from pilot control inputs.
The description of Airbus Fly-By-Wire reminds me of programming with Eclipse!
EDIT: Okay, you thin-skinned Eclipsers, note that while the Airbus has a horrendous failure mode in extreme conditions, according to the article, it's a much more efficient way to fly a complex aircraft, and much less prone to human error.
The point is, you still need to know your craft, for when the automation fails you!
[...] THE SYSTEM WILL NOT FIGHT THE PILOTS. The system just gives me the best the airplane can do at the moment without me having to stop and consider my conditions.
If I as the pilot deflect the stick left, if the autopilot is engaged, it will be disengaged for me. The load factor is considered and the airplane will begin a maximum-rate roll in the direction I deflected the stick. I keep holding the stick and the airplane will keep rolling up to the limiter. That's when Normal Law is active. If we're in Alternate or Direct law, there is no roll limit and I can roll the airplane onto its back and crash it if I desire. [...]
http://tech.slashdot.org/comments.pl?sid=1261521&cid=282...