(Historic) French Airliner Crash into Sea, TV Prog yesterday.

Read the forum code of contact

Member for

14 years 6 months

Posts: 2,536

This was a program about the French Airbus that stalled and crashed into the sea when the Pitot tubes iced up. The crew were confused. It wasn't entirely their fault. They died too.

Two things I find amazing about the aircraft.

1) Why don't airliners have airspeed relayed from satellite positioning as a backup? True, not as accurate but better than nothing? Cars have this, my IPOD has this. Why not aircraft?

2) In a stall, where stall is defined as out of control, recover or crash and die. Why doesn't the computer fly by wire override the crew and re-stabilize the aircraft. It could also give a verbal `voice` update about the problem one stabilized ("you were in a stall condition. You were too slow and held up attitude causing the stall. To computer did the following...')? What's the point in fly by wire if everybody dies?

Passing 20'000 feet the pitot tubes started to work again but the aircraft was still stalled. IMHO the computers could have taken control. Stabilized the aircraft and then handed back control of the aircraft with an update for the crew. This seems to me one of those situations where you want the computer to take control rather than everybody dies. It doesn't make sense for the computer to just watch an aircraft fall from the sky without even trying to intervene.

Original post

Member for

13 years 5 months

Posts: 919

1) GPS can be switched off by the Americans (who own the satellites)... at any time...

2) What would signal the fact that the pilot is wrong? The FBW computers would see a controlled pilot input as a manoeuvre which can be carried out. The aircraft did everything that was asked of it.

Remember the A320 at Paris? The pilot put the aircraft in a landing configuration so the aircraft landed. This seems like a similar situation but with complications - the pilot put the aircraft into a stall configuration, which the aircraft held. I'll be utterly honest and know little about the airbus-into-the-atlantic crash, but I'm sure I recall that the aircraft DID give verbal and illuminated warning that it had stalled. The pilot viewed this with suspicion and ignored the advice of his rather clever aeroplane... I could be wrong, so don't take it as gospel.

The best thing you could do to improve the safety of aircraft is to remove the pilot......

Member for

24 years 2 months

Posts: 9,780

Hampden -the auto pilot had clicked out when it stopped receiving airspeed input from the probes. As soon as the aircraft gave an airspeed input -the crew could have switched the autopilot on and the aircraft would have recovered itself without any aircrew input. Simply put the aircraft was trying to save itself but with it in manual it cannot do anything to override the input that is holding the aircraft at a steep angle of attack - i.e the pilot holding the stick back!

Member for

19 years 5 months

Posts: 9,821

1) GPS can be switched off by the Americans (who own the satellites)... at any time...

So?
The question is still valid, why not use it as a back up system?

Member for

13 years 5 months

Posts: 919

A back up system is generally a simple fail safe in case the main system fails. If the backup can be switched off, then it is no longer any use!

There is nothing wrong with taking your i-pad (other GPS apps may be available...) in the cockpit with you, but EASA / FAA / CASA etc, are not going to allow you to fly a commercial airliner basing its emergency procedures on a system which is not internal to the aircraft.... Manufacturers are not going to fit an internal GPS when EASA et al won't certify it because it is under an individual's control... particularly when that individual is US Military....

EDIT: Apparently using GPS in EASA regulated aircraft is perfectly safe! (Something to do with EggNog or something...) It seems to be limited to lateral guidance only. http://ec.europa.eu/enterprise/policies/satnav/egnos/files/safety-information-bulletin-easa_en.pdf

So to recap.... US Military GPS (what your I-pad uses) can't be used, BUT you can use EASA Eggnog with EASA regulated aircraft. No doubt they'll charge a fortune to fit it, which most airlines won't do.....

Member for

12 years 11 months

Posts: 390

The documentary said a single pitot de-iced. As there are three an auto-flight system would use the value from the two best readings. If only one of the three pitots were giving an airspeed indication I would expect the auto-flight system to drop out.

The thing that disturbed me most was that there was no clear agreement about who was the handling pilot at any time after the aircraft stalled. Both co-pilots were flying it at the same time. And they were unaware of this. In a standard cockpit aeroplane it would be very obvious if both pilots were hands on at the same time. But with the stupid little Airbus joystick down by the pilot's side, it wasn't clear that the right-seater was holding the nose up in a deep stall, even though the guy in the left seat was attempting to recover the aeroplane.

I feel pilots of complex transport aeroplanes need to go back to basics when things go wrong...

Attitude and power = airspeed.

Stall warner sounding = lower the nose.

Also after the Teneriffe 747 collision (PanAm and KLM 747s collided on runway in fog) the standard 'Captain = god' mode of crew operation was changed to the 'All Pilots are members of a Team' mode. We saw the down-side of this beaurocratic 'lets discuss things together' approach in this tragic accident. The captain should be the active commander of the flight IMHO.

Member for

24 years 2 months

Posts: 9,780

The crew had the horizon providing attitude indication -that should have been enough to determine that something was wrong. As for the side sticks - the Airbus A330 has been flying perfectly well with them for a good while.
If its a case that a pilot is holding it back against all reason and the crew cannot realise this its hardly a fault of the aircraft!

Member for

15 years 10 months

Posts: 268

Why did the pitots ice-up? I thought that they were heated against icing.

Member for

16 years

Posts: 1,059

There was a problem with faulty pitot tubes on A330/340 aircraft at the time, I think. They are supposed to heat up, yes.

Member for

24 years 2 months

Posts: 9,780

Airbus say the pitot tubes were working within their design parameters -the was a programme underway to change the type as there had been failures in the previous years of operation on the A330. Whilst the new types would have been desirable on this machine -the lack of airspeed information should not have meant the loss of the aircraft.

Member for

15 years 10 months

Posts: 268

There was a problem with faulty pitot tubes on A330/340 aircraft at the time, I think. They are supposed to heat up, yes.

Airbus say the pitot tubes were working within their design parameters..

Are you saying that the heating on the pitots was insufficient to prevent ice build up in the conditions encountered during that flight? That this was therefore an inherent design specification deficiency?

Steven

Member for

24 years 2 months

Posts: 9,780

Steven -Airbus say that the probes in use exceeded the design specification for the aircraft. However at least one probe did ice up -the probes did work again on the decent and the aircraft itself sent a fault code to France .
There was a programme underway to replace probes with a different type- it doesnt indicate that the probes themselves didnt meet the specification -it tends to indicate to me that they found a better type.

Member for

15 years 10 months

Posts: 268

Steven -Airbus say that the probes in use exceeded the design specification for the aircraft. However at least one probe did ice up -the probes did work again on the decent and the aircraft itself sent a fault code to France .
There was a programme underway to replace probes with a different type- it doesnt indicate that the probes themselves didnt meet the specification -it tends to indicate to me that they found a better type.

The probes may well have exceeded the design specification, but that is no good if the specification was set at a too low standard, which could be the case here. Either that, or the probes were prone to the heating system failing.

Steven

Member for

15 years 9 months

Posts: 1,684

Here lies the crux of the matter as VeeOne puts it.

Here lies the crux of the matter as VeeOne puts it.

More senior pilots have said this about the AF447 tragedy than have said anything else.

I feel pilots of complex transport aeroplanes need to go back to basics when things go wrong...

A rating license should mean the holder should be able to fly the aircraft with or without many automatic FBW systems by using their experience, brains and working as a team (flight crew, cabin crew, ground voice support, other aircraft in vicinity).

It takes 'dosh' - training in simulators is cheaper for airlines/pilots than taking a real 'bird' up - maybe bending it and even in extreme cases breaking it beyond repair.

Simulators can only provide systems designed for events that have 'gone before'. Unfortunately modern aircraft fly in all weathers, all altitudes and are made of different materials piloted by humans (with varying skills/experience) and the ensuing combinations of error are humungous.

In Sully Sullenberger's own words - "The AF447 story will be a seminal case in airline, pilot and aircraft history" - here is just one extract from Sully on AF447 http://www.cbsnews.com/2100-500202_162-20067860.html

And here is the original quote from Sully from http://www.cbsnews.com/2100-18563_162-20067739.html ""This Air France accident is going to be a seminal accident that will be studied for years, and we need to ask ourselves as an industry tough questions about the way we're designing airplanes, the way we're displaying information to the pilots in the cockpit. And about whether or not making airplanes more complicated, more technologically advanced makes it more difficult for pilots to very quickly intervene and very effectively act when things go awry," Sullenberger said. "

My take on this, IMHO (as a former pilot and mission critical systems tech now turned 'bean counter') is that until Systems Designers / Software Engineers stop living in their own coccooned world and actually fly aircraft they will continue to produce fallible systems (infallible in their opinions only).

Member for

24 years 2 months

Posts: 9,780

The aircraft was displaying a wealth of information just the crew for whatever reason could not figure out what was going on. The horizon clearly was working -therefore no matter what training you put the pilots through -if you cannot understand that your in a climb and at the same time loosing height from your instruments I cannot see what more a manufacturer could do. Much is made about how automation is making the pilots job harder -in reality the aircraft are giving you lots of information and helping you ! All that was needed was a pilot to switch on the autopilot as soon as airspeed information came back on line and the aircraft would have sorted itself out.
The aircraft was flown into the sea -the crew got themselves in a box rather than trying different ideas -a simple hands off controls would have shown what was happening early.

Member for

15 years 9 months

Posts: 1,684

Flying absolutely serenely at 35,000 ft, then ......

Flying absolutely serenely at 35,000 ft, then ......

1. Altitude 35,000 ft -FACT
2. No evidence of anything untoward and Captain is on routine rest break. FACT
3. Weather radar set to sense at Medium Threshold - FACT(one to be reviewed by all manufacturers)
Other aircraft in the area avoided the storm
4. AF447 flew into a small storm but due to weather radar being set to sense at Medium Threshold did not see the monster storm they actually were flying in to. - FACT
5. Pitots iced up - FACT
6. Auto pilot dis engages - FACT
7. Left Seat should have taken controls, hand on stick, feet on rudder pedals and hand on throttle levers (minor movement of to dis engage auto throttle) - No evidence of all of this being done while keeping her straight and level.
8. Issue a Radio Message to any aircraft in vicinity stating predicament. - No evidence of this
9. Aircraft engines were performing, all electrical systems were performing and absolutely no evidence of any explosion - FACT
10. Keeping her as straight and level as humanly possible through a horrendously 'bumpy' ride through the 'monster' storm would have helped as they had 35,000 ft to play with over ocean (no mountains).
Possibly a lot of spilt coffee and drinks in the cabin and maybe some injuries to pax and crew not belted up but hardly likely to have had any fatalities through the extremely 'bumpy ride'.

A fix for weather radar sensing systems to warn that once into a storm that the weather radar sensing was automatically going to maximum sensing threshold unless manually overidden, would help greatly.

Hence among millions of other reasons why 'softies' in FBW design must also learn to fly and gain real experience not just design from textbook logic and code.

And most of all pilots being capable of really flying their aircraft.

Member for

12 years 11 months

Posts: 390

Yes! This is the old issue come back to haunt us all. Should a computer fly the plane or should the human do it? I think Airbus has stepped over the line into dangerous lands with their systems (designed by NON-pilots as nJAYm points out!).

These auto-flight systems are subtly taking away the flying instincts from airline pilots that they grew up learning on Pipers and Cessnas.

Like

... a quick and positive response to a stall warning system
(all three pilots ignored the stall warner for over a minute).

... and a clear verbal understanding about who is the handling pilot at any time
(both 'first officers' were hands on and when the left seater attempted to lower the nose he was unaware that the other guy was still pulling the stick back and holding the aeroplane in a fully stalled condition).

And they should have jumped up and taken action when they heard the radio altimeter react at 4000 feet. Admittedly the 330's RA voice alert said "pull-up pull-up"; not helpful when in a nose-up stall condition. Older Radio Altimeter systems would have said "Terrain Pull-up" and this would give a different message over the ocean (eg: we are going into the sea).

But any pilot understands that low/no airspeed, nose high and in a fast descent = deep stall.

I can't see why this accident happened except to think nobody was actually the pilot-in-command, everyone was 'discussing' like a good team crew.

Airline pilots need to fly hands-on more and fly auto-flight systems less. Computer systems are not yet so reliable that pilots are 'stand-by' systems. I doubt they ever will be in my lifetime. Until computers are more reliable and safer than humans we should safeguard those flying skills of the pilots rather than teaching them to become systems operators who are over-reliant on the FMS.

Maybe I am just getting old but this is one of the most disturbing crashes I have heard about. No engine fell off, no stabilator runaway happened, no structural failure. A perfectly servicable aeroplane and three pilots doing their best to recover from a stall (!!) and 220 people die for what? No meaningful reason I can see. It is really sad.

Member for

24 years 2 months

Posts: 9,780

The ironic thing is that by the crew flying 'hands on' they actually lost the aircraft . Flown straight and level it would have been simple to reingage autopilot once they had defrosted -there was no need for any panic.
The pilots didnt do their best to recover from a stall - a pilot put it in a stall and held it there . Eventually the aircraft ran out of height -it wasnt an aircraft fault - if you choose to do the opposite to what the aircraft is designed for it will eventually crash.

Member for

14 years 6 months

Posts: 2,536

This isn't a question of having the computers fly the plane.
Give the pilots 100% control while the aircraft is in safe, controlled flight.
But, have an emergency system that, should the aircraft depart from controlled flight (as far as I am aware a stall should never happen in a passenger airliner) the computer cuts in and restores control.

Member for

18 years 5 months

Posts: 472

This isn't a question of having the computers fly the plane.
Give the pilots 100% control while the aircraft is in safe, controlled flight.
But, have an emergency system that, should the aircraft depart from controlled flight (as far as I am aware a stall should never happen in a passenger airliner) the computer cuts in and restores control.

And that's just how the fly by wire works when in protection mode! There is also a second mode with no protection (i.e. just like a non fly by wire) which can be selected if the pilot wishes or if a vital sensor failure degrades the protection available.

Member for

18 years 5 months

Posts: 472

Hence among millions of other reasons why 'softies' in FBW design must also learn to fly and gain real experience not just design from textbook logic and code.

And most of all pilots being capable of really flying their aircraft.

With respect what utter tosh.

You cannot name a single A330/340 flight control system engineer let along say for certain that not one was a qualified pilot - FACT

Your comments show close to zero understanding of the flight
Control design process - FACT

Your ill informed fantasy stated as expert comments is really insulting to professional engineers who are also qualified pilots like myself.