Air France Flight 447: 'Damn it, we’re going to crash’

Very good article on the Air France crash of a few years ago.

http://www.telegraph.co.uk/technology/9231855/Air-France-Flight-447-Damn-it-were-going-to-crash.html

Good article indeed.


Now let me get this straight:

In an Airbus, the movement of one of the "joysticks" (sry) is not repeated to the other one ? No synchronized movement, or some force-feedback ? So if the pilot pulls, the copilot's stick does nothing and remains in neutral position and vice versa ?

And two experienced pilots do not look at the artificial horizon to verify (or not) the plane's nose attitude, or angle of attack ?
Or was the artificial horizon also turned off - is this even possible ?

I must say as soon as this fly-by-wire turned up i had a very bad feeling .. when it comes to controllers, managers and programmers never flying a plane but setting the standards following economical reasons ... gawd. Give the pilot full control in an emergency and keep some analogue instruments to be able to fly and land a plane with complete electronics/avionics failure !
But Boeing also flies by wire now, there is no plane left i think that still uses direct hydraulic or cable control.

Anyone remember this Airbus that used the whole runway and crashed into a ditch in Poland back then due to not able to break ? My theory is the computer did not allow counterthrust for breaking, because one side of the main landing had no ground contact yet - but in bad sidewinds sometimes you HAVE to land the plane on one leg if necessary - or am i wrong ?

Thanks and greetings,
Catfish

Together with Ernst Udet I also had a bad feeling when they began to make glass-enclosed cockpits, Catfish... How do they expect pilots to fly if they can't feel the on-rush of air?

Airliner safety is much better now than it ever has been.

They may have believed that all their instruments were unreliable. You can't know what they were thinking without asking them. You can't know where they were looking and why.




Why would you want direct hydraulic control when you can have fly by wire? There's no reason for it.
This wasn't a fly-by-wire issue. This seems to have been an aircrew training issue.




Without seeing the report your theory doesn't mean much.

^^ lol Freycenet,
You are certainly right about overall safety, however .. in a real bad situation with bad weather and avionics failing, i would feel most safe in an older Iljushin, with russian pilots at the controls not spoiled by electronic gadgets

Udet had nothing against glass cockpits ? He just had some kind of despise for pretty blinking lights without real information for the pilot - as he proved when he took up Heinkel's 178 jet for a test

Greetings,
Catfish

I see your point, but no synchronized joysticks ?
And if the pitot tubes froze and the computer shut off the displays, what about the artificial horizon.

You are right, no evidence without the report which we will most likely never see.

Here is what I don't understand:

"Once again they began to shout: “Stall, stall, stall.” Tragically, as Bonin did the right thing to pick up speed, the aircraft seemed to tell him he was making matters worse. If he had continued to descend the warnings would eventually have ceased. But, panicked by the renewed stall alerts, he chose to resume his fatal climb."


When you get a "stall" warning, why would you pull back on the stick?

I suppose its easy to get overloaded by the warnings, especially if they don't seem to be making sense. I think there was a programme made about this crash, 'Air Crash Investigation' or something. Watching it, you think 'WTF how could they not realise!'

I'm not saying there's no room for improvement. In this case, synching the sticks may or may not have helped. The person at the controls had clearly said he had control, and I don't recall if the other guy tried to take control without saying so (it might be possible to determine this from black box data).

The displays did not shut off. There's no such silly things happening. The reporting of the airspeed seized, as you would expect to have happen with any instrument if the pitot tube freezes. The autopilot correctly shut off at that point.

There was an erroneous overspeed indication, which is why the pilot pulled up to slow the plane down.

This is a tale of mounting confusion in the cockpit. The pilots couldn't see outside (it was dark, no sight of horizon), apparently didn't trust their instruments any longer (can't see horizon + don't trust instruments ... see where this is going?) - but that's speculation on my part - not trusting the instruments. The captain figured it out the moment he walked in and the guy told him he was holding the stick back all this time. It was too late.

Edit: I read the article (I have read some of the report before) and I see what you are referring to now, I think. Once the aircraft was stalled, and the airspeed fell below 60kt, several readings were blanked out because they were now useless. This isn't turning off the displays'. It simply gives you an indication of bad/unavailable data for some things. The attitude indicator is operated by a gyroscope, so it would not have been affected. Things such as stall warnings, overspeed warnings, generally things affected by reading airspeed and AoA would have been affected. Pressure altitude readout may or may not have been affected.

Hard to tell. Initially they had an overspeed warning (Due to the frozen pitot). This could have ended up becoming a self-fulfilling scenario.

Have you ever crashed your virtual plane in broad simulated daylight while in a turning fight?

All the time!

Something to remember for those who seem to think that one manufacturer is better than another (or nervous flyers!) is that aviation accident investigation units around the world are in mini-crisis.

This is because there are no accidents for them to investigate. There is concern that junior staff are not getting enough investigative experience. Its amazing how safe air travel has become.

'Mayday', about airplane crashes, on the Discovery Channel is a good show to watch.

Same deal, different stimuli

Correct.
There's no mechanical inter-connect between the two sticks, and the flight computers average the position of the two sticks. Now this sounds like Bullsh** but it's true; if one pilot pushes one stick full nose-down and the other simultaneously pulls full nose-up, nothing will happen until one of them lets go.
I did most of a full type-rating on the A-330 and became familiar with it. I was a proud Boeing Bigot before that, but after knowing how they work I am quite happy to say they are a good thing. However, I do not agree that having unconnected joysticks is a good idea at all. All the Airbus people, when quizzed on this, say it's fine and never a problem. When I ask them why they don't do the same thing then with the rudder pedals, a blank stare always occurs ...



Edit - I am also of the same opinion with the throttles, as they are not always representing the actual power commanded by them. For example in the cruise setting the actual engine power could be literally anywhere from flight idle to full power and the only way to tell is by looking at the clocks. Other settings (takeoff, flex, idle) give what they say though. I think the idea of removing the tactile feedback from moving throttles with changing power is also a bad idea.

I stopped reading the article. It appears to me to be full of assumptions, some facts, some general personal attitudes and some some questionable statements.

From what we can tell from the story.

1. It becomes apparent that they flew into a adverse weather (possibly severe) - is this the 1st problem in the chain of events.
2. the article mentions a smells and st-elmos fire, but later mentions some screens went down. So there were losses of systems.
3. Its pratcically impossible to stall an airbus with autopilot and autothrottle engaged. the fact the aircraft stalled and the article stating the autopilot dropped out to me also points to exptreme adverse weather and loss of systems. Aircraft possibly in alternate law ?

4th and 5th are the most 2 important factors to me. The article mentions stall warning went off 75 times, yet pilot was trying to climb, also mentions toga power (most power you are asking from the engines.)

So firstly trying to climb whilst in a stall. and secondly with underslung engines, toga power will be providing a nose up momentum adding to the problem.

To me, stating that if it was a boeing the accident may not have happened, is absolutely ludicrous.

Stall recovery was not implemented. Is that the fault of fly by wire ?

And the discussion of the sticks not reflecting the others position is not really relevant here i think. At no time was the big red button pressed which locks out the other controller (providing the audible "priority left/right") for the other pilot to rectify the situation.

I would give this article and its assumptions a wide birth and leave it up to the findings of the appropriate investigation board.

Actually the stuff they printed is mostly from the official crash report.



No screens went down and no systems were lost. The article is pretty clear that certain instruments were blanked out due to inconsistent data. It was a proper response to the sensors not feeding correct data (they were not ABLE to feed correct data once the aircraft became slow enough)



Pitot tube froze, autopilot popped off because it didn't have a good airspeed indication. Exactly as it should be.



Actually initially they got an overspeed warning, and the PIC climbed to slow down. He got a stall warning, so he dove, and then he started trying to climb again.



It may or may not have been relevant. You can't tell if the other pilot would have recognized the problem earlier if he saw what was happening with the stick.



Take your own advice. The article repeats findings from such an investigation

One thing I'll agree with though - this article shouldn't be trying to draw conclusions on cockpit design. None of the author's business.

Actuall the limited facts in the article may have been from the official report but the article was also filled with a lot of speculation and some questionable statements. Which is what i said before.



This is a contradictory statement from you due to what you have written below. If a system is frozen, not performing accurately, disabled, unreadable or in any state where you cannot rely upon it wether in a temporary or permanent time frame... the system is lost. What were the screens displaying then Greyghost? Quite simply systems were lost.


This is the contradictory statement. in any event "exactly as it should be". Is it???? What technical background do you have to base this on? What were the ECAM warnings, what was the trigger for the autopilot disconnect? was it the weather (uncontrolable flight path) was it loss of systems ???? With the pitot and AofA sensors heated, what weather did they fly into to freeze them up like that ?



Actually was it the pilot or was it an effect of the protection systems (depending on what law the aircraft is in) or a combination of both ? at high altitude and a heavy aircraft you can have not much of a gap between overspeed and stall (coffin corner). How did they get that overspeed warning ? what was the cause from it ? what gap did they have between cruise IAS/mach and overspeed ?





And thats the point. With 2 PFD's and an ISIS surely there was ONE accesible reference of attitude. You can tell if the levers are in Climb, flex or toga dentent (written in the FMA's on the PFD) also written on the lever deflection and the pilot asked at one point for confirmation that he was in toga. So i cannot see any relevance at all that the deflection of the stick was an overiding factor here. Seems to me the problems were laid elsewhere. So with knowledge of where the thrust levers are and some form or attitude reference which should have been accesible on the flight deck, one pilot should see if the actions taken by the other pilot are adding to the problem. Instead of a stall warning going of a reported 75 times (well it would have gone off for a bit in a deep stall)



Yes i will take my own advice. While there may be factual items reported in the article, the article alone is far from conclusive in the information it provides and it also makes assumptions and also includes boeing bias.

So that is not journalistic integrity and more journalistic sensationalism. And i picked that out of it pretty quickly.

Ok i can agree with some misgivings with fly by wire (i use flyby wire) but i cannot see its relevance here. And to bash airbus fly-by-wire over boeing is pretty bloody rediculous as if it wasn't safe, the military wouldn't use it, the government wouldn't allow it, boeing wouldn't have moved to it and i wouldn't fly it.

A tragic chain of events... 3 seasoned pilots losing it? to their credits, you cannot really reproduce the violence and the unpredictability of a thunderstorm in a simulator, so for pilots usually relying on computers and autopilot most of their flying time, to find themselves in a sudden and very stressful environment must have shaken them pretty bad... maybe enough to not analyze the situation in a clear and basic way... It's easy to find faults and look for culprits but it's another to find yourself in the same situational hyper stress...

Reading this article is shivering and scary... even when the aircraft have the most advanced securities ever in the history of aviation, and that you have a real experienced crew of 3 pilots... #%&*$# can still happens which end your life....

The one faulty thing on the aircraft was the pitot tube (They were known to be prone to freezing, and I believe Air France had been replacing them). Beyond that, what you're seeing is graceful failure of systems, not loss of systems.

The systems are still there, they will still work, but they won't show you info that they don't have. When the sensor input becomes unreliable, the system will not show you anything or otherwise it will inform you that the sensor input has become unreliable.



Are you familiar with the concept of systems that fail gracefully? It was exactly as it should be. The autopilot was deprived of input that it needs to maintain safe flight, so it disengaged. It says so in the report.





Actually it was the pilot. Read the report.



Right, and the stick thing was the point of sensationalism.



Do you mean about the stick? I agree. But there's nothing in here about systems going down, screens turning off, star trek red alert blaring with sparks flying in the cockpit.

Well, i am glad that no one i fly with would agree. PF to PNF, "unreliable speed,".....PNF.... "Don't worry its not lost its just being graceful". lol.... !!!!



LOL>>> !!


If a system does not function correctly it is unreliable and lost or failed. Until it pings back operational or prooved reliable (recovered and operational), then unfortunately the system is lost until recovered. I don't know how you would explain fail passive and fail operational.... Fail graceful ? lol




And this is my point you are missing. The news article (Not a report) is what we are talking about here and it catagorically does not provide nearly enough facts for it to be a factual complete cover of the events in the crash. It surmises, it assumes, it uses bias. Hell the article is more about fly by wire being a devil in disguise which has been agreed by its hellish minions operating places of power inside ministries, industries, and aeronautical governing agencies.

The article does not explain the icing and the article does not explain a wide number of factors that could make the pilots job harder. It has been proven that an accident is a chain of events. The swiss cheese model, and this article does not cover all the holes in the cheese but places scrutniy anyway.

(setting aside how they got into such icing conditions) from my perspective, i can already envisage additional problems they may have encountered not mentioned in the article.

Wether or not the article rightly put blame in the correct place, is not what i am debating. The article is poorly presented with some ridiculous statements.



I beleive so. We practice proceedures for windshear, stall, unreliable speed and a host of other sweat inducing moments in the sim. all are accomplishable with the use of the side stick. MY point that i have been trying to make all along is that the news article does not cover all bases.
Yet asserts a negative connotation towards fly-by-wire, and then adds fuel by saying, if in a boeing, they may have lived...... Crikeys, absolutely fricken rediculous. If so, airbus would be grounded, and we would all be happy in a boeing.

STUPID !



Well i don't know what article you have been reading but if its the same one, then i think your rear end has been writing on this thread more than your brain....


AND I QOUTE FROM THE NEWS ARTICLE



Lets pick another really really REALLY REALLY stupid comment in the article.

Flying towards a tropical thunderstorm. He was experienced enough to handle it.... You don't fricken handle it, you AVOID IT. Especially a torpical one as the troppopause is so much higher in the tropics those storms have a #%&*$# load more energy. The article does not mention a squal line or a line of tropical thunderstorms.

Mate, the ARTICLE IS $HIT and its attempts to sensationalise the old fly-by-wire boeing vs airbus debate is just rediculous.

Losing the ASI is no big deal if you know it's happening.

Those are my biggest beefs with the Airbus design philosophy. WHY ON EARTH would it ever, in any situation, be a good idea to degrade decades of gradually acquired (often through accident analysis) cockpit cooperation philosophy to the point where one pilot doesn't know what control inputs the other one is making?!

Yep.
The other problem I have with them is that I think they tend to rely too much on the reliability & fidelity of the various sensors.

interesting billzilla, i would like to know more

I just got back from our training center in Denver after completing my Proficiency Check (PC). My Sim partner and I finished the program early and the Instructor Pilot asked if we wanted to try any thing "different" or just go home. I saw that they had the "Air France scenario" in options so that is what I requested.

All I can say is "Wow". It was a hand full and required the epitome of team work to accomplish correctly. We had the advantage of knowing what was coming and it was still tough and made us sweat more than anything we got in the PC. An Airbus at a high flight level at a heavy weight in Alternate law with no altitude and airspeed indications in the dark in the clouds with turbulence and convective activity with non-stop ECAMS was fun in the Sim. I doubt it would have been as fun in real life.

That is a great experience. Going from FDH (Fat, Dumb, & Happy) to trying to sort out a myriad of annunciators and failure displays and warnings, at night, in a thunderstorm must be quite the challenge. At high altitude the difference between overspeed and stall can be a bit confusing in a chaotic situation. I'm sure revised training will stress going to the backup attitude indicator (all the jets I fly have them / separately powered electrically driven) and putting the nose on the horizon and setting a mid range "flying" power setting and then sorting out all the conflicting indicators.

Poor bastards.

BeachAV8R

Hell yeah - In the 747 all you have to do is set about 2.5° nose-up on the AI and about 90% - 95% on the fans and you have quite a long time to sort out whatever problem you have.

Sounds interesting. I have my check coming up this month, we don't usually get extra time, but it would be interesting if we could have the option to replicate such a scenario.

To be honest, i wonder how they flew into the storm in the first place? surely wx radar would have shown up the cell and win direction would have given you the side the anvil is on??

From what I understand it was a large "airmass" area of almost tropical-like thunderstorms. I don't know what the feasibility of circumnavigating it was.. Looks like it could have been done but one of those things were they've driven through it a hundred times before with no problem.. Until there was a problem..

"Air France Flight 447 crossed through an area of tropical showers and/or weak thunderstorms with weak to moderate updrafts and a high likelihood of turbulence. The flight penetrated one cell at about 0150 UTC and then entered a cluster of cells beginning at 0158 UTC. The suspected zone of strongest cells was reached at 0208 UTC, which corresponds with the beginning of a track deviation, and another cell appeared to be reached at 0210 UTC, which corresponded with the time of autopilot disconnect. The flight was suspected to be within areas of showers and precipitation up until the time of impact, and the descent below FL250 into the critical -10 to -20 deg C zone probably involved some degree of clear icing on control surfaces, though it is uncertain whether this affected recovery of the aircraft, especially due to the short accumulation time that would be involved."

From This analysis..

And that is the scariest part.

I wonder how many aircraft accidents follow in the same footsteps.

[edit] the report beach is very detailed.. thanks.