The Last Four Minutes of Air France Flight 447

Hi!

Have just been reading the latest analysis of the failure according to Der Spiegel, it makes for interesting reading. If already posted I couldn't find it.

'One tiny technical failure heralded the impending disaster. But the measurement error was so inconspicuous that the pilots in the cockpit of the Airbus A330 probably hardly noticed it.'

Link

www.spiegel.de...

Peace!

reply to post by The Wave


G'day The Wave

Thank you for posting that very interesting report.

It will be intriguing to watch how all this is handled by Airbus & their competitor, Boeing.

Kind regards
Maybe...maybe not

Originally posted by Maybe...maybe not
reply to post by The Wave

 

It will be intriguing to watch how all this is handled by Airbus & their competitor, Boeing.

Kind regards
Maybe...maybe not

They will do what they always do - fix the reported problem, issue a maintenance directive and move on.

reply to post by The Wave


Hi!

Thanks for the comments - it's interesting that Air France has resumed looking for the flight recorder after quite a time - there's obviously some flaw that Air France/Airbus are worried about. And I agree - they'll find it and fix it.

I did like the tone of the report though - factual and plausible.

PS Took a while for anyone to respond - guess I should have titled the thread; 'Magnitude 11 gay, disgruntled tax paying alien didn't bring down Flight 447' :-)

Peace!

Originally posted by RichardPrice

Originally posted by Maybe...maybe not
reply to post by The Wave

 

It will be intriguing to watch how all this is handled by Airbus & their competitor, Boeing.

Kind regards
Maybe...maybe not

They will do what they always do - fix the reported problem, issue a maintenance directive and move on.

G'day RichardPrice

I was also thinking about how each Co might use this to communicate a competitive edge to prospective purchasers of their planes & the general public.

Kind regards
Maybe...maybe not

Personally I've always thought Airbus relied to much on automation. At the very least the pilots should be able to take manual control of the aircraft. I find it outrageous and sickening that these experienced aviators were reduced to amatuer computer repairmen desperately trying to reboot the flight computer as they fell helplessly from the sky instead of being able to take control of the plane.

Originally posted by danwild6
Personally I've always thought Airbus relied to much on automation. At the very least the pilots should be able to take manual control of the aircraft. I find it outrageous and sickening that these experienced aviators were reduced to amatuer computer repairmen desperately trying to reboot the flight computer as they fell helplessly from the sky instead of being able to take control of the plane.

You completely and totally misunderstand the situation in both this flight and general Airbus aircraft operations.

The myth that pilots have no manual control over an Airbus aircraft is just that - a myth.

Airbus control law has 4 laws, and 5 states.

The laws are:

Normal Law - everything operating as it should. All aircraft protection limitations are in place in this law, meaning the pilot cannot make a manoeuvre which will damage the aircraft.

Alternate Law - multiple system failures. Most aircraft protection limitations are disabled, meaning the pilot has direct control of the aircraft in 99% of manoeuvres.

Abnormal Alternate Law - aircraft has somehow gotten into an abnormal manoeuvre or attitude. All protections are disabled, pilot has full control to recover the aircraft.

Direct Law - no flight control systems in place, no functioning computers. Pilots commands are held as a direct relationship between controls and command surfaces.

The fifth 'state' is in addition to the above:

Mechanical Backup - flight controls not responding to input. Thrust, pitch and yaw have direct mechanical backups on the flight deck.

In this flights case, they would have degraded down to Direct Law, as the flight computers became unavailable but they still had power to transmit control signals. This means that the aircraft was never not in the pilots control.

The reason the pilots rebooted the computers was because they were seeing bad information, and were trying everything to reconcile that with the real world. How many times have you rebooted your desktop computer to see if it will fix a problem, only to find the problem is not solved by said reboot?

The problem with this flight was not that Airbus aircraft take control away from pilots, but that bad data being reported to the flight deck, whether interpreted by a computer or a human, is fatal.

Without reliable pitot static data, the pilots do not know quite a bit of information, and there is no current backup on either Boeing or Airbus aircraft for this as a standard feature. In a IFR (instrument flight rules) situation, bad pitot static data will probably result in any aircraft crashing.

Take a look at the Boeing 757 crash which involved taped up pitot static ports - the pilots had no chance, couldnt work out where they were or what they were doing, and crashed the aircraft into the ocean. And that was in an aircraft with no flight control computers.

[edit on 28/2/2010 by RichardPrice]

S+F cause this is good stuff!

I would have preferred 'Magnitude 11 gay, disgruntled tax paying alien didn't bring down Flight 447' as a title myself.

RichardPrice above this post has some real good insight into this too, maybe even enough to reboot (sorry) the whole debate.

Good info, let's see how it turns out.

-m0r

Have bumped the thread as BBC is reporting that the black box from AF447 has been located within a 5 k area (but not retreived)

news.bbc.co.uk...


Will be interesting to see if they can retrieve them and what findings are published.

Peace!

[edit on 6-5-2010 by The Wave]

Well, here we are, several years after the event itself and also after all of the speculation.

Today, we have loads more information, thanks to Airbus, the BEA and the French governments efforts in continuing to search for those black boxes.

And in the light of that information, boy does that Spiegal article look like a hatchet job!

So, as it turns out, the computers didn't crash, the aircraft didn't "have a stroke" and infact remained eminently flyable throughout the entire incident.

It was the human element that failed those passengers. Those three pilots uhmed and ahhed for several minutes in that cockpit, and still couldn't solve the issue - not because it wasn't solvable, but because they weren't working together.

reply to post by RichardPrice


Grossly unfair to the pilots. Certainly the problem WAS solveable....but as even Air France said:

A brief bulletin by Air France indicated that "the misleading stopping and starting of the stall warning alarm, contradicting the actual state of the aircraft, greatly contributed to the crew’s difficulty in analyzing the situation."

- pilots untrained to deal with stall warning

The aircraft was excessively nose-up - at which point the stall warning ceases because the aircraft considers the attitude invalid and/or the airspeed too low.

it started again when they pushed the nose down - and so they essentially becamse conditioned to kep the nose too high which meant they weer never going to recover.

The initial problem was a pilot error - not applying the incorrect airspeed procedure - but after that they were probably toast due to this design "feature".

interim report #3 - english version is a good read - but it is hard to argue with the Air France Statement above too.

The evidence that points to faulty information and late reaction is that they understood that they were losing altitude, most probably due to the loss of airspeed, and presumably countered with the standard procedure of increasing thrust..but, was the engine stalled or did the pilot(s) just react too little, too late? The angle of impact seems to show they were reacting with incorrect data, for if they knew they were going in the correct manuver would have been to slow airspeed as much as possible for the best chance of an emergency water landing, yet evidence shows no warnings to the passengers (and especially none to the stewardesses).
I don't understand how they couldn't have used visual sighting of correct altitude as the moon was out, the turbulence and poor visibility was at higher altitude. The article only guesses that the automated warning system was telling them "terrain, pull up!", it's unknown until the black box is recovered.
The main flaw of the airbus may be an unrecoverable stall when the lift is reduced due to faulty sensors. Given only four minutes until disaster, this leaves almost no time to react and get that 200 ton elephant enough lift again to recover.
Flying is still MUCH safer than Florida drivers however..any day of the week and especially in bad weather.

Originally posted by Aloysius the Gaul
reply to post by RichardPrice

 

Grossly unfair to the pilots. Certainly the problem WAS solveable....but as even Air France said:

A brief bulletin by Air France indicated that "the misleading stopping and starting of the stall warning alarm, contradicting the actual state of the aircraft, greatly contributed to the crew’s difficulty in analyzing the situation."

- pilots untrained to deal with stall warning

The aircraft was excessively nose-up - at which point the stall warning ceases because the aircraft considers the attitude invalid and/or the airspeed too low.

it started again when they pushed the nose down - and so they essentially becamse conditioned to kep the nose too high which meant they weer never going to recover.

The initial problem was a pilot error - not applying the incorrect airspeed procedure - but after that they were probably toast due to this design "feature".

interim report #3 - english version is a good read - but it is hard to argue with the Air France Statement above too.

I've followed this extensively - of course Air France are going to say that, because it shifts the blame (see the reactions from the pilots unions and other crew related sources after the announcement earlier this year - especially when Airbus announced that it had no further amendments to make to issued guidance, the pilots union and Air France when nuts on that alone!), which shifts the culpability.

The crew put the aircraft into a fully developed stall - every aspect of crew training in all modern large civil airliners is to avoid a stall at all costs, an airliner in service should never actually enter a stall.

The stall warnings are just that - warnings of an impending stall. They are not announcements of an actual stall, because you should never be in a developed stall - your training should take you away from there before you are ever in that situation. There is absolutely no situation at all where a stall warning when pushing the nose down is a valid one - the crew would have known that, especially as they were already dealing with known invalid readings elsewhere. For Air France to try and say "the crew were untrained for this particular stall warning" is a load of tosh.

In this case, the pilots reacted wrongly to the original problem and against published Airbus guidance on the matter (incidentally, also against Boeing guidance as well) - the pilot flying should never have tried to climb the aircraft, and it was the continual attempt to push that climb that forced the airframe into a fully established stall.

The stopping and starting of the stall warning is a nice distraction for Air France to move the publics attention elsewhere, but in reality the air crew ignored pretty much every indication on the flight deck instrumentation - including those which were never invalid. Then the pilot flying reacted wrongly to the situation, and never corrected himself - in an almost constant nose up position, the pilot flying forced the aircraft into an aerodynamic stall, when all he had to do is push the nose down...

So no, I don't consider my comments unfair to the pilots, the crash was immensely avoidable if they had followed established procedure - but they didn't, and infact they got it so wrong that the BEA are taking long hard looks at the crews relationship with an eye to seeing if it can be avoided again.

Originally posted by rlever1
The evidence that points to faulty information and late reaction is that they understood that they were losing altitude, most probably due to the loss of airspeed, and presumably countered with the standard procedure of increasing thrust..but, was the engine stalled or did the pilot(s) just react too little, too late? The angle of impact seems to show they were reacting with incorrect data, for if they knew they were going in the correct manuver would have been to slow airspeed as much as possible for the best chance of an emergency water landing, yet evidence shows no warnings to the passengers (and especially none to the stewardesses).

I suggest you go and read the reports on the matter, they answer all of your questions.

In short, the pilot flying reacted to the initial instrumentation issues wrongly, and never corrected his actions during the entire descent. The interaction between all three pilots on the flight deck indicate that no one was actually "working the problem", all of them were almost literally just sitting there going "uhm, wtf?"

The constant nose up attitude put the aircraft into an aerodynamic stall, meaning that the angle of attack of the air over the wings no longer produced any lift - which produced the very high descent rate of the aircraft. The pilots basically didn't believe this was happening, right up to the end.

I don't understand how they couldn't have used visual sighting of correct altitude as the moon was out, the turbulence and poor visibility was at higher altitude.

Because "visual sighting" is useless in such a situation - you are over a featureless ocean, with no datum points. There is nothing for you to make any sort of judgement on.

But then again, the altitude indicator was never showing an invalid reading. Neither was the rate of descent indicator. But the pilots were ignoring both.

The article only guesses that the automated warning system was telling them "terrain, pull up!", it's unknown until the black box is recovered.

Both black boxes have been recovered - read the reports.

The main flaw of the airbus may be an unrecoverable stall when the lift is reduced due to faulty sensors. Given only four minutes until disaster, this leaves almost no time to react and get that 200 ton elephant enough lift again to recover.

The stall was recoverable from (multiple senior pilots have suggested ways, and its been done in the simulator), the lift was never reduced due to faulty sensors, it was reduced because the pilots reduced it through their actions.

Modern large civil aviation pilots are not trained to recover from stalls, because you should never ever put a large aircraft into one - pilots train regularly and thoroughly for recovery from the approach to a stall, there is significant amounts of guidance on the topic and significant amounts of help in the cockpit. But an approach to a stall should never be allowed to develop into a full stall - that is left to the test pilots.

In this case, a fully developed stall occurred because the pilot reacted wrongly to the original problem - he should have pushed the nose down and increased power to the engines (the established procedure for invalid readings at altitude - "fly the aircraft by AoA and thrust"), but instead he pitched the nose up and continued to pitch the nose up, and thus the aircraft entered the stall.