Air France Flight 447


#1

Heavy weather can cause previously hidden problems to surface. However the focus post-accident tends to be biased to problems closer to the last moment when disaster can avoided.

An example is Air France flt 447. On the pilot forums the discussions focused on the actions of the third officer and the set-up of the flight controls. But there’s this:

FromPopular Mechanics:

Other flights crossing this stretch of ocean around the same time have gone out of their way to avoid the storms; evidently the co-pilots believe they can pick their way through the worst of it with their weather radar.


Hydrostatics
#2

Yes, put a lot of stress on any system and the faults will quickly come to the
surface.

Other flights crossing this stretch of ocean around the same time have gone out of their way to avoid the storms; evidently the co-pilots believe they can pick their way through the worst of it with their weather radar.

Where have I read this before? There certainly is a high level of Davidson reasoning in it!

As ever so often, a matter of human errors again. Unbelievable that an experienced pilot can make such a beginners mistake, pulling the nose up in a stall… Where was that captain?


#3

I saw a simulator re-enactment of the iced up pito tubes condition that they think caused this accident. The pilots in the simulator were simulator instructors. They weren’t told what error to expect. They did manage to control and land the simulation safely, however. Tempting to say it was a training problem.


#4

Thank you! I understand that the Pitot tubes were frozen for about one minute. Wikipedia furthermore says:

The BEA’s final report, released at a news conference on 5 July 2012, concluded that the aircraft crashed after temporary inconsistencies between the airspeed measurements – likely due to the aircraft’s pitot tubes being obstructed by ice crystals – caused the autopilot to disconnect, after which the crew reacted incorrectly and ultimately caused the aircraft to enter an aerodynamic stall, from which it did not recover.


#5

It was a series of small errors William Langewiesche has a good article here.

The crew had a “profound loss of comprehension”. I know how that feels, that how I felt when the ship started rolling in head seas.

Langewiesche talks about the way the flight controls work on a Airbus 330. A better test for the instructors would have had one jr pilot panic and see if the rest could trouble shoot that system.


#6

Langewiesche is one of the best aviation writers around but I am not alone in thinking this was less than his best effort. In particular, many of us think he was way too easy on the flaws in the Airbus cockpit design, specifically the absence of an angle of attack indicator (just ask any Navy pilot of your acquaintance what they think about doing without that.)

The excessive level of automation led to a system where the robot delivered a conclusion (STALL!) without giving the pilot the basic flight parameter (AoA) which led it to conclude that. This forced the pilot to, in effect, play “twenty questions” with the system: “if I do this, does it go away?” “OK, that didn’t work, how about this?” “Are you really sure we’re in a stall?” and so forth. You can’t cross-correlate instruments if one of them is missing. No wonder the poor guys got confused.

This, and other shortcomings of the Airbus system (sidestick and throttle logic) are why I hate to fly on the things. I can never go to sleep on one, whereas I climb on a Southwest 737 (whose crew does more landings and takeoff a day than other airline pilots do in a month) I zonk out as soon as the wheels are up. Makes me, like most of us, a walking example of actual vs. perceived risk :sunglasses:

Cheers,

Earl


#7

Modern pilots cant fly, get over it, all the crashes in the last 20 years have been unskilled pilots flying a perfectly good aircraft into the ground/sea.
AirAsia 8501, another one.

If the pilots have grey hair, you will arrive at your destination


#8

Automation definitely played a serious role in that accident. The captain was the ONLY one with manual flying experience on that particular model AC. The co-pilots were all button pushers.


#9

next time you meet a pilot ask them if they are even allowed to do high altitude manual flying in the sim


#10

#11

Pat Smith of “Ask the Pilot” had a post about that article: Automation and Disaster with a reply from Langewiesche.

Langewiesche focuses on the roles cockpit automation may have played in the accident. His conclusion is a familiar and, to somebody who flies for a living, irritating trope: pilots have become so reliant on automation that we no longer know how to fly. It was a degradation of basic flying skills, more than anything else, that was responsible for the Air France pilots committing such a basic and unforgivable blunder.

But the crash of flight 447 was brought on by a combination of things. Poor airmanship was only one of them. Pilot inexperience was another factor, and possibly crew fatigue as well. Langewiesche also gives short shrift to certain Airbus design quirks that may have played a significant role. For example, the fact that the control sticks are not inter-linked. When first officer Pierre Bonin, sitting in the right seat, first pulled the jet into a stall, the extreme inputs he was making weren’t apparent to the pilot in the left seat. This contributed much to the ensuing confusion.

When the pilot in the left seat took control at one point Bonin took it back without saying anything. In hindsight we know that there was nothing wrong with the plane but how did the pilot know it wasn’t pyscho like flight QF 72?


#12

And here we get into the Airbus vs Boeing controversy, fueled by the marketing departments of the respective manufacturers. Airbus fans love that space age cockpit and think Boeing cockpits are stupidly retro. Boeing fans think the Airbus cockpit is dangerously weird and Boeing cockpits reflect the realities of flying. Smith is an Airbus pilot.

What he refers to as a “design quirk” would be called by a Boeing fan (like me) a flat out design flaw. As is the throttle logic on the Airbus, where the throttles don’t move to reflect what the robot is doing. So you can look down at the throttles and they are sitting at full power but the robot has cut back to idle. A similar disconnect exists for the sidesticks on the Airbus: the position of the stick does not necessarily reflect the position of the control surface. All part of a design philosophy that the crew’s role is to act as advisors to the robot. This “robot is in charge” approach, according to the (obviously biased) Boeing engineers I’ve talked to, was motivated by the desire by Airbus to reduce flying cost by improving fuel efficiency. As always, it all boils down to the Benjamins.

Cheers,

Earl


#13

This is an interesting article Dealing With Unexpected Events on the Flight Deck: A Conceptual Model of Startle and Surprise

The article uses the concept of “frames”

Pilot perception and actions are conceptualized as being guided by “frames,” or mental knowledge structures that were previously learned. Performance issues in unexpected situations can often be traced back to insufficient adaptation of one’s frame to the situation. It is argued that such sensemaking or reframing processes are especially vulnerable to issues caused by startle or acute stress.