What Really Happened Aboard Air France 447

[roo_ster]

http://www.popularmechanics.com/technology/aviation/crashes/what-really-happened-aboard-air-france-447-6611877

Green pilot, over-automation, lack of clear command authority.  Toss in night & t-storms and you get what happened on AF447 a couple years back.

[Fly320s]

That is effed up.

[AZRedhawk44]

"Alternate Law" sucks.

Or, more accurately, "Normal Law" sucks that it doesn't act like Alternate Law.

No disrespect towards TallPine intended, but we let software monkeys do too much with aircraft.  Autopilot is one thing, but software that overrides control commands, written and accepted as a means of invalidating human control input, is a bad idea.

Pilot fault is one thing.

Code error is another.

Technically there is no code error and this is pilot error on the part of the junior co-pilot... but his assumption that it was impossible to stall the plane because he "shouldn't" be able to induce a climb steep enough to do so... is asinine.  Like an 18 year old kid on a motorcycle.

But I fault the industry for even allowing such a computer to exist, and for allowing a junior co-pilot with poor yoke control to kill 230 people via assumptions made about that computer.

[Nick1911]

The thing that got me about this is that there is no feedback from what the other pilot is doing at the controls.  The computer just averages differences of control inputs.  At least with mechanical controls that are tied together, if one pilot is doing something stupid with the flight surfaces, the other would have some knowledge of it by way of his control stalk moving!

[Fly320s]

AZ, I disagree. The laws work well, when understood by the pilot.  They even work well when the pilot has no idea how the PFM works.  In this case, the pilot had no idea how the laws work and no idea that the A330 is still an airplane and needs to be flown like one.

Nick, that shouldn't make a difference if both pilots are paying attention and have some SA. The sidestick has a take-over button that allows one pilot to override the other pilot's input.  For some reason, neither of these pilots ever thought to lower the nose and trade altitude for airspeed before it was too late.

I think a contributing factor is that there are fewer "plane-guys" out there. Many pilots today have no interest in flying, just an interest in pushing buttons and earning a paycheck. 

[tokugawa]

This was horrifying to me.  Absolute panic and incompetence. They burned through 38,000 feet of air with the stick held back hard. They had a attitude indicator, altitude, rate of climb, ground speed, power setting, and were unable to figure out what the aircraft was doing?  (apparently when they got the pitot tube ice free they were stalled so hard the airflow over the tube would not register )  They taught me  this stuff in primary flight training.
 
 And unbelievably, they had that rarest of things in any emergency- time!- yet they blew it anyway. How in the hell does someone get to the place of flying commercial passenger jets with so little understanding of flying?

[AZRedhawk44]

This was horrifying to me.  Absolute panic and incompetence. They burned through 38,000 feet of air with the stick held back hard. They had a attitude indicator, altitude, rate of climb, ground speed, power setting, and were unable to figure out what the aircraft was doing?  (apparently when they got the pitot tube ice free they were stalled so hard the airflow over the tube would not register )  They taught me  this stuff in primary flight training.
 
 And unbelievably, they had that rarest of things in any emergency- time!- yet they blew it anyway. How in the hell does someone get to the place of flying commercial passenger jets with so little understanding of flying?

This, in eleventy-spades.

I just quit a job due to coworker incompetence, starting a new job where it seems people actually have intellectual integrity and accountability.

I fail to comprehend how a PILOT with 230 souls on board, can be this stupid.  Quite literally... terminally stupid.  And not just 1 pilot.  3 of them.

And with the blanket assumption that "the software makes this plane un-crashable!  Lalalalalala-BOOMSplash...."

[MicroBalrog]

We're now at the point the software is far more reliable than the pilots.

[De Selby]

We're now at the point the software is far more reliable than the pilots.

This - the tech wasn't the problem; its assumption that it needed to hand over control to a human was.   Said human ignored what his machines told him in a panic. 

[RoadKingLarry]

The closest I've ever come to flying an actual airplane was about 20 minutes in the front seat of a small Cessna as a passenger and I think I may have been able to pull that one out of the crapper.

[Tallpine]

No disrespect towards TallPine intended, but we let software monkeys do too much with aircraft.

This stuff scares me to death.  It's one thing to automate the control of a turbofan engine, but another to fly a plane.

And like sausage, I've seen how it's made 

[AJ Dual]

We're now at the point the software is far more reliable than the pilots.

Yep, brain-lock and panic. The problem here was more that the software didn't/couldn't override the pilot.

Are there scenarios where a human may recognize a special set of circumstances where the default safe actions of an automated system are "wrong"? Sure, however, playing the odds, the smart money is to bet on the machine.

At least presuming that the automated solution has it's origins in the first world, and was designed with lawyer-proofing it in mind.

[Hawkmoon]

It seems at least that a contributing factor was that the itty-bitty joysticks in the Airbus planes are not linked together like the traditional yokes in other aircraft. I got the sense reading the transcript that the more senior first officer was not really aware that his doofus right seater was pulling back on the "stick" the entire time he was wondering why the plane was descending. I don't trust the Airbus system, and I didn't even know about this "feature" until I read this article. I don't even like to fly these days, courtesy of the TSA, but when I do have to fly, I'll always choose a Boeing plane if there's any possible option to do so.

[AJ Dual]

It seems at least that a contributing factor was that the itty-bitty joysticks in the Airbus planes are not linked together like the traditional yokes in other aircraft. I got the sense reading the transcript that the more senior first officer was not really aware that his doofus right seater was pulling back on the "stick" the entire time he was wondering why the plane was descending. I don't trust the Airbus system, and I didn't even know about this "feature" until I read this article. I don't even like to fly these days, courtesy of the TSA, but when I do have to fly, I'll always choose a Boeing plane if there's any possible option to do so.

There's a fair number of things about Airbus that have been good for Boeing the past few years.

America has had it's share of commercial aviation blunders, but it seems that Europe has a slight statistical lead on us over that.

[MicroBalrog]

It seems at least that a contributing factor was that the itty-bitty joysticks in the Airbus planes are not linked together like the traditional yokes in other aircraft. I got the sense reading the transcript that the more senior first officer was not really aware that his doofus right seater was pulling back on the "stick" the entire time he was wondering why the plane was descending. I don't trust the Airbus system, and I didn't even know about this "feature" until I read this article. I don't even like to fly these days, courtesy of the TSA, but when I do have to fly, I'll always choose a Boeing plane if there's any possible option to do so.

Not Superjet? 

[Fly320s]

This - the tech wasn't the problem; its assumption that it needed to hand over control to a human was.

Wrong. The tech did need to give the control to the humans because the input to the computer had failed. No data input = no computer.  Once disconnected due to failure, the computer can not auto-reconnect. Nor should it.

[roo_ster]

Wrong. The tech did need to give the control to the humans because the input to the computer had failed. No data input = no computer.  Once disconnected due to failure, the computer can not auto-reconnect. Nor should it.

This.  Iced over airspeed sensors.  Software can't chew on good intentions.

[MicroBalrog]

Question: Is it technologically possible to use a GPS/GLONASS receiver as a backup tool to estimate the speed of an aircraft?

[roo_ster]

Question: Is it technologically possible to use a GPS/GLONASS receiver as a backup tool to estimate the speed of an aircraft?

Likely latency precludes it from being useful for flying a plane, while still being useful for navigation.

[Fly320s]

Question: Is it technologically possible to use a GPS/GLONASS receiver as a backup tool to estimate the speed of an aircraft?

No.  roo_ster got part of it.  The big problem is that GPS measures ground speed, not airspeed.  Airspeed is used for flying; ground speed is used for planning.

[AJ Dual]

No.  roo_ster got part of it.  The big problem is that GPS measures ground speed, not airspeed.  Airspeed is used for flying; ground speed is used for planning.

I agree, however, unless you're in a hurricane or tornado, and flying with the wind, some sort of ballpark estimate, at least above stall speed for your particular jet could be done when the pitot tube isn't working or plugged.

I mean, ground speed does have to matter sooner or later, as it does have direct bearing on ETA, fuel consumption (in light enough aircraft under the right wind conditions it is possible to "stand still" and burn fuel gaining no ground) and on take off an landing, when ground speed is suddenly going to matter... a lot.

It kind of bleeds over into my  about the UAV/Iran thread over in politics.

[MicroBalrog]

Help me here.

GPS/GLONASS can give you altitude and location. If we can calculate it's changes in realtime or near-realtime, why can't that also give you airspeed, i.e. tell you how fast you are descending/ascending? Isn't GPS/GLONASS already used in missile and shell guidance like this?

[Nick1911]

Help me here.

GPS/GLONASS can give you altitude and location. If we can calculate it's changes in realtime or near-realtime, why can't that also give you airspeed, i.e. tell you how fast you are descending/ascending? Isn't GPS/GLONASS already used in missile and shell guidance like this?

Airspeed is the speed relative to the air around the aircraft.  So, if you're flying into a strong wind, you might have very good airspeed, but not being going very fast relative to the ground.  (In this case, you'd be well within your flight envelope, even though a GPS would say you are going too slow.)  The amount of air going over the wings dictates lift - and it can't be measured by absolute velocity of the aircraft relative to ground.

[roo_ster]

Airspeed is the speed relative to the air around the aircraft.  So, if you're flying into a strong wind, you might have very good airspeed, but not being going very fast relative to the ground.  (In this case, you'd be well within your flight envelope, even though a GPS would say you are going too slow.)  The amount of air going over the wings dictates lift - and it can't be measured by absolute velocity of the aircraft relative to ground.

Also, GPS has some latency issues.  Use for navigation?  Yes.  Use to fly the plane?  Not likely.

[MicroBalrog]

Airspeed is the speed relative to the air around the aircraft.  So, if you're flying into a strong wind, you might have very good airspeed, but not being going very fast relative to the ground.  (In this case, you'd be well within your flight envelope, even though a GPS would say you are going too slow.)  The amount of air going over the wings dictates lift - and it can't be measured by absolute velocity of the aircraft relative to ground.

Thank you. I didn't know that.