Posted on 01/04/2015 5:39:05 AM PST by C19fan
Without doubt the great automation and reliability of modern flight control systems has made the sky much safer. In fact the safest it has ever been.
My question is do we have enough training (particularly in foreign airlines) for the crew to recognize when the system has gone bad such as in the Air France crash off Brazil? Also relative to that crash, why did they not go back to just hand flying the craft by attitude via their artificial horizon as opposed to letting the computer fly it into the ocean due to false information generated by instrumentation and feed to the computer control system? It is my assumption that all modern aircraft have backup mechanical gyros that are separate from the glass cockpit that is computer driven.
“I am an airline captain, and I train airline pilots. You are woefully ignorant.”
I know nothing about what happened so I am not speculating on what may or may not have happened.
I would like to hear from you about how the newer technology planes operate and manual procedures by pilots which over ride these systems.
I would never question the hands on experience of professional pilots but, perhaps, a comment upon the systems physics and the design choices of Airbus and Boeing is appropriate.
Airbus fly-by-wire design is computer centric. The aircraft computer sets a flight envelope beyond which the pilot cannot go without significant effort. Worse, Airbus designs are dead stick designs which yield no stick feedback to the pilot.
Boeing, on the other hand, uses a pilot centric design. If the pilot requires a certain out of envelope maneuver then the computer is overridden immediately. Their cockpit design tries to emulate the physical feedback (proprioception) of traditional cockpits.
The implication of these choices is profound. The dead stick A300 accident was simply due to no proprioceptic feedback to the pilot from flipping the stick back and forth (trying to compensate for wake turbulence) eventually ripping off the vertical stabilizer.
From a control systems point of view the Airbus choice was poor. Add this to the over reliance upon automation generally and, specifically, to European Airbus training and you shift the probability of error negatively.
The Air France accident likely occurred due to very poor ergonomics in the cockpit and the fact that a computer reboot or override took too long once an error was detected.
These data are somewhat old (several years) so engineering changes may have been made since the Air France accident. However, I try to avoid Airbus equipment when the flight service station indicates any weather issue.
Thank you and agreed. It’s way too easy to assume “we’ve gone too far” with the technology when something goes wrong...only to forget how many lives have been saved by them, avoiding human mistakes.
It’ll be the same thing once we have self-driving cars. They’ll eventually be better than average (human) drivers...but *when* an accident occurs everyone will be suing and blaming them.
Very interesting article. Thanks for posting.
Tzfat, perhaps you’ve keyed on the problem:
“All US airline pilots train for automation failures. The training involves reversions, layering down, and relayering automation. No one can get a job flying an airliner in the US without flying skills.”
These pilots and those of Air France crash several years ago, were NOT trained to US standards.
“No one can get a job flying an airliner in the US without flying skills. Period.”
But they can in Korea, and I’ll wager other SEA countries as well. The Asiana crash at SFO illustrates that fact all too well. When faces with an inoperative ILS, they were simply unable to stabilize the a/c on the glide path and execute a successful manual landing. These guys are lucky if they can taxi to the active before turning the rest of the flight over to the plane’s automatic systems. And they get to log ten or twelve hours of flight “time” for sitting on their a$$es drinking coffee.
Moral of the story: fly American.
Fourth generation airliners are first and foremost airplanes. Most of these breathless articles make it sound like these airplanes fly themselves. They don’t. In most regards, pilots flying a fourth generation airliner take the same basic steps to go from point A to point B as an early DC-9 (I’ve flown them all, first gen to fourth gen).
You flight prep at the gate in both. You taxi to a runway the same way. You take off the same way. You turn on an autopilot the same way (and the same place). You land the same way.
The differences are in the tools used to do the same tasks. In today’s airliner the preflight prep at the gate is extremely time intensive. There is no “kick the tires, light the fires” attitude. That intensive programming then transfers tasks that previously were done enroute, namely navigation, to on the ground, at the gate. It makes sense: move some tasks to a less critical phases of flight.
Bottom line: a pilot is intimately involved THROUGHOUT a flight, in new and old airliners. The automation permits the pilots to transfer tasks to less critical phases of flight, thereby freeing them up for the most important task: monitoring.
It is a myth that Airbus are computer centric and Boeings are pilot centric. The fact is, the Boeing B787 is currently the most high tech airliner flying. Fly by wire. And no, a pilot on an Airbus CAN MOST CERTAINLY override the design limits imposed by the computers (Boeings have flight computers as well). Airbus makes NORMAL operation to remain within limits, but override is possible, and we train pilots to know how.
As for ergonomics causing AF447, that is ridiculous. It was caused by a barometric system failure not a computer failure. It was caused by almost the same set of events that caused a North Eastern B727 to crash 40 years ago.
Your basic premise is outdated, and easily proven to be false. Modern US fighters follow the same “ergonomic” patterns you mistakenly think are Airbus’ idea. If the “pilot centric” myth you think exists at Boeing, why is the configuration and operation of the B787 essentially the same as an Airbus A350XWB from a pilot’s perspective? Why did Boeing abandon all those “great ideas” they had in cockpit design found in the B676?
Sorry Capt, but you are wrong. 2 and possibly 3 recent airline disasters prove you wrong.
I, too, have been associated with the airlines, and I am quite proficient in partial panel, and total elect failure flying. NO current pilots of recent minting are proficient in these areas in real life. And, ex- Military pilots rarely understand the civilian air traffic system until they have worked inb it a while.
The American Airlines A300 crash was not caused by the pilot “flipping the stick back and forth.” It was caused by his feet on the RUDDERS (not fly by wire) moving them with the aggressive movements taught to FIGHTER pilots. The forces applied to the rudder would have ripped the tail off most FAR Part 25 airliners.
Bad training, and most certainly not automation, or aircraft design caused that crash. And just so you know: the aggressive rudder hard over actions he learned from U.S. military training concepts.
But if you had read the NTSB findings you would know that.
I have not simply been “associated” with airlines. I have been an airline pilot for 34 years. 15 of those training pilots to fly the most advanced airliners in history.
I think your self-educated opinion is just that, and your use of “NO current pilots...” is illogical on its face. Have you flown with ALL pilots to make such an emphatic statement? I didn’t think so...
The other pilot on the right was pushing the joystick forward, but the joystick on the left overrode him, and he didn't realize it. (Boeing uses wheels that are mechanically linked, so pushing one steering wheel forward always pushes the other.)
The official determination was pilot error due to failure to understand how the fly-by-wire A300 behaved whenever the computer kicked offline and went into alternate (mostly manual) flight mode.
BFL
As for ergonomics causing AF447, that is ridiculous. It was caused by a barometric system failure not a computer failure.
...
Initially, but all three sets of instruments had good airspeed indications after one minute. From that time until the crash, the problem was extremely bad piloting skills and resource management.
The crew did poorly, but they had conflicting indicators. Very similar to the North Eastern B727 crash 40 years ago.
This is a good question. They should know enough to put the nose down and add power when they hear the stall warning, but I'm not sure pilots can really override the computers.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.