AF 447 report out

[Carjockey]

Automation Complacency

What worries me as your average SLF, is that the FBW / computer systems on commercial aircraft are apparently almost totally dependent on a single instrument, namely the pitot tube.

Is their no backup system available in the event of pitot tube failure?

With the technology available today, surely there is an alternative means available for an aircraft's computer system to establish it's airspeed without simply disconnecting the AP?

This situation began because the AP disconnected, and the situation degraded because the pilots did not understand what was going on.

Pilot error is surely a factor in this case, but a more significant factor is the failure of aircraft manufacturer's to recognise that automated systems should be designed as an adjunct to a pilots flying skill and training, and not as a substitute for it.

[MountainBear]

I find it hard to believe the hypothesis -- because that's all it is, unfortunately we cannot ask the pilots what was actually going through their mind but can only formulate hypothesis -- that they went back to pulling on the stick because when they pushed the previously suppressed STALL warning became active again.

Your assumption that in a crisis situation a pilot will revert to training is false. The PF pulled back on the stick not because he was confused but because he was biased.

The 'startle effect' is an example of the over arching mental heuristic known as the recency effect. A different but related error is the primacy effect. The report makes clear that the pilot was primed to take the plane up because of the anxiety he had expressed about the plane's flight level prior to the incident. He went up because he wanted to go up; his training be damned.

The transcript makes clear the consistent incomprehension the PF had regarding the situation. Despite this incomprehension he still acted. He may not have acted on his training. It makes perfect sense that when confronted with conflicting and confusing stimuli he fell back on his pre-existing albeit erroneous mental model that the plane needed to go up.

You're correct, of course, that this remains a hypothesis since we cannot read his mind. Yet all hypothesis are not created equal and based upon what an extensive amount of science teaches us about human behavior it is a rational, reasonable, and quite plausible hypothesis.

[stepwilk]

archaic pitot tubes for airspeed reading

Please explain how your 10-times-as-accurate "solid state sensors" will determine airspeed--the salient part of that word being "air."

[NigelOnDraft]

Is their no backup system available in the event of pitot tube failure?

Well, yes there is. The other 2 pitot tubes. Trouble is when there is a common problem affecting mroe than 1

Pilot error is surely a factor in this case, but a more significant factor is the failure of aircraft manufacturer's to recognise that automated systems should be designed as an adjunct to a pilots flying skill and training, and not as a substitute for it.

This situation began because the AP disconnected, and the situation degraded because the pilots did not understand what was going on.

Your 2 statements do not quite add up

The AP disconnected precisely because 'that automated systems should be designed as an adjunct to a pilots flying skill and training' occurred - the AP was not able to determine what was happening to a sufficient degree - the system designers / regulators decided that was a point to hand it back to the crew.

Now as you say 'and the situation degraded because the pilots did not understand what was going on' and that is an issue for training / skill / practice. Now I fly "made in Tolouse" products for a major European airline, LHS for nearly 10 years. I do not recall one sim training session where I was "persuaded" to be distracted / looked down / pretend dozed off, and then given the aircraft in an unusual attitude, with some systems / insts failed, and told to resolve it *. I do recall being given some very gentle stall recoveries to perform, precisely, with a detailed brief in the minutes before "to address the AF issue" (all systems working except in Altn law)

* The sort of UP recovery every RAF student pilot started doing at ~25hrs and in every IF training trip / test thereafter.

[vovachan]

I think (as a SLF) -- this is a problem inherent in multitasking. You cannot fly a plane and troubleshoot a plane at the same time. Humans just are not very good at this kind of thing. Not with a million confusing messages popping up.

One moment they are bored half to death next thing you know all hell is breaking loose...

[SadPole]

@stepwilk

Please explain how your 10-times-as-accurate "solid state sensors" will determine airspeed--the salient part of that word being "air."

Already did, here:


http://www.pprune.org/rumours-news/4...ml#post7280726

Airplane application of such a sensor would basically require a little metal peg or even simply piece of metal foil stuck on a wing. The principle of operation is such that you heat a piece of metal to a specific temperature and measure the amount of energy required to keep it at that temperature. The bigger air flow cools it faster so more energy is required to keep it at constant temperature. Damaging/disrupting/contaminating such a sensor is orders of magnitude less likely than Pitot tube.

Pitot tube is really an archaic sensor that is used in aviation (in my view) only because of historic and bureaucratic reasons. They continue to use it in spite of constant problems with it - ice and bugs clogging it and causing crashes.

To begin with, Pitot tube does not even measure directly any kind of flow but pressure difference. To calculate some sort of flow/speed out of it you need to know current air density.

Pitot tube - Wikipedia, the free encyclopedia

On top of it all, the pressure difference measured by Pitot tube depends on angle of attack, which adds even more problems.

On the other hand, hot-wire mass flow sensor reading is directly what you need for better control of flying condition as it it the mass of the air flown around the wings that gives the lift. Or in other words - thinner air gives you smaller lift than thicker air at the same speed. Using this sensor in vacuum would give you zero flow reading and zero aerodynamic lift even if you were going at a very large speed.

Meaning, if this technology was known from the start, that's what would have been used - because it directly measures what is needed to maintain required aerodynamic lift.

Hot wire mass flow sensors are used in every car produced today. It is a well-known technology.

Mass flow sensor - Wikipedia, the free encyclopedia

[overthewing]

You cannot fly a plane and troubleshoot a plane at the same time.

As SLF also, I understand that standard aviation practice recognises this and that this is why the PF is supposed to fly the plane while the PNF does the troubleshooting. In the case of AF447, this division seems to have been muddled.

[VC10man]

Is it normal to have 3 pilots on a trip of this length?

[hetfield]

Is it normal to have 3 pilots on a trip of this length?

YES, about 12h flight time.

[fireflybob]

Sorry, i don´t agree.
Why were they at that position at that time?
Other flights, e.g. IB, diverted miles to the east.

That´s when/where it started!!

testpanel, good point - or you could say is started by not fitting the new pitot heads. Aircraft have been flying around for decades with pitot heads fitted that don't ice up - I find it difficult to believe that a modern aircraft of this complexity was not fitted with pitot heads that did not ice up!

Is it normal to have 3 pilots on a trip of this length?

VC10man, yes - it's a function of the duty time - above a certain figure (depends a bit with operator and Flight Time Limitation scheme) you are required to have a "heavy" crew.

[flapsjml]

I'm a private pilot with a visual rules rating. Having watched programmes on TV about air disasters, I don't understand why so many accidents have happened to commercial pilots not recognising a stall situation, or were unable to recover from the situation.
Does relying on instruments, when piloting commercial flights, take away the skills that were learnt during basic training?

[jcjeant]

so much so that when pilots are faced with something "unusual" they become lost

Is that "unusual" to lost Pitots indications ?
Is that "unusual" to lost an engine ?
Is that "unusual" to lost part of hydraulic power ?
Or is that "unusual" to fly the aircraft manually ?

[Mike744]

Pitot tube is really an archaic sensor that is used in aviation (in my view) only because of historic and bureaucratic reasons. They continue to use it in spite of constant problems with it - ice and bugs clogging it and causing crashes.

To begin with, Pitot tube does not even measure directly any kind of flow but pressure difference. To calculate some sort of flow/speed out of it you need to know current air density.

I was wondering how many alternatives to the Pitot tube sensor there are and what would be the pros & cons? Apart from the hot wire method there are also ultrasonic sensors similar to the Curtiss-Wright type. Considering the reliability inherent in modern electronics surely there must be a cost effective & viable alternative to the Pitot tube.

[Northbeach]

TOGA bury the stick & everybody else

Airbus makes a fine, technologically advanced and safe jet. I've never flown one, never had the opportunity.

I had an Airbus captain in the jumpseat and asked this individual about his high altitude stall recovery training. The answer I got was that the Airbus will always beat my jet's performance because to escape a ground proximity warning they can engage TOGA and apply maximum pitch input (bury the stick) and the airplane will not stall. After all you can't stall an Airbus.

The answer did not address my question about Airbus' high altitude stall recovery training. Perhaps there was/is no Airbus high altitude stall training. With all due respect; selecting TOGA and burying the stick will not solve every problem. This leads us to the post two positions aove mine where a pilot in training asks the question if becoming a professional pilot causes one to forget/lose their basic flight training background in stall recognition and recovery. How do we answer the question?

Every aircraft in flight is capable of being stalled. It takes the right combination of events and/or circumstances. The idea that " XYZ" aircraft is incapable of being stalled should be eliminated, it is a dangerous error in thinking.

Advanced aircraft, avionics, and ratings do not invalidate basic aeronautics. If the airfoil is stalled then the angle of attack must be decreased to regain laminar airflow to get that airfoil " flying" again.

There are not two sets of basic aerodynamics: one for Private pilots and the other for us commercial pilot " professionals".

Thou shalt maintain thy flying speed, lest the ground (or sea) rise up and smite thee.

[xcitation]

@jcjeant
Indeed this can happen.. only if you don't check your airspeed ... (who was under 60 knots) and if you don't know the specifications of your aircraft ...
Anyway .. when you read 60 knots (all parameters indicated that this speed was valid) on the speed indicator of an A330 .. you can deduce two things:
Or you are rolling on ground
Or if you are in the air ... your plane is no more flying .. he is falling like a stone ... (check altimeter)

For airspeed a qualitative 'sanity' check would be Ground Speed from GPS. If both airspeed and ground speed are around 40kts then it supports a diagnosis of stall unless you believe in 300kt wind. To that end the PAX could have seen the GPS speed decay from 400kts to 40kts in the GPS in the seat back - very sad.
The PF had less than 2 years ATPL, I recall his wife was on board. Perhaps many with similar training/experience would do the same or worse in this nightmare scenario. The ultimate responsibility lies with the capt.

[hetfield]

The answer did not address my question about Airbus' high altitude stall recovery training. Perhaps there was/is no Airbus high altitude stall training.

Sad but true.

Personally, I flew A300/320family and A340 for a major EU carrier.

NEVER EVER had high altitude stall recovery training.

Only in the beginning training on A300 some approaches to stall below 10k feet were done.

[SadPole]

@Mike744

Pitot tube was invented in year (get this) 1732 so it is quite safe to assume we did learn a thing or two about better ways to measure gas flows since then;

Pitot tubes allowed to measure approximate air-speed without need of electronics or even electricity (except for heating the damn thing) and thus were a great tool from the very beginning of aviation.

Original Pitot tube provided two pipes with static and dynamic/impact pressures that were connected to a mechanical gauge where the different pressures acted upon a membrane that was driving a air-speed gauge needle.

See this Popular Mechanics article from 1944.

Popular Science - Google Books

This isn't to say that everything that is that old must automatically be bad - but the problem is that all countries have aviation laws that specifically require Pitot Tubes and other such archaic things and it simply seems easier to throw some electronics into 300 year old sensor than to change all the laws, I think.

So, "modern" Pitot tubes no longer output their readings in form of two hoses but have a silicone pressure sensor instead - and this is the extend of the "innovation" in the aviation sensor field.

On the other hand, in cars you can now find all different types of mass flow electronics.

[rgbrock1]

SLF here.

I've read and reread the CVR several times now. There are a couple of things that were occurring on the flight deck that I really do not understand.

On several occasions the PNF points out to PF that he is ascending and that he should push the nose down to enable a descent but to do so "gently!". PF then responds that he's descending. (PNF also points out the decaying forward speed of the aircraft to PF.) But PF really isn't descending as he's merely eased up the back pressure on the stick. PF then states "we're in a climb." At this point the PNF attempts to summon the Captain back to the flight desk.

After announcing that he no longer has control of the aircraft the PNF takes control and proceeds to pull back on the stick. Even after suggesting a descent - but "gently!" - to the PF.
ter PNF took control the PF took control back and the Captain appeared on the flight deck with the words "What the hell are you doing?"

PNF states "Climb... climb.... climb" to the PF whereupon, finally, PF states that he's had the stick back the entire time.

The captain then says "no... no... no. Don't climb. No... no... no"

After this seeming command PF continues pulling back on the stick anyway.

So, after having received several suggestions or seeming commands to put the nose down, PF continues pulling back anyway. This is a huge interconnect between what he's being told and what he's actually doing. Can it be that at this point in the situation the PF is actually in panic mode and is merely doing anything he can think of without forethought as to what it is he was actually doing? Similar to the flight or fight method of dealing with extreme/dangerous situations? It almost seems as if the PF has temporarily lost all semblance of situational awareness, including what is being suggested/commanded to him.

[TRW Plus]

Hey, not a pilot here, just curious ... is there any technology in existence or under development that could relay air speed to pilots from GPS technology? For example, a GPS device on board an aircraft would interact with either fixed ground stations or satellites in defined locations, and this interaction would yield an air speed? I would imagine (in my rather non-technical haze) that such technology could also be adapted to read out the pitch or AOA, all of which would render these pitot tubes obsolete?

[Sandy Swan]

I would just like to urge all Airbus pilots to read BEA's AF447 Final Report. It is an education, plugs holes in our understanding and illuminates many areas left grey or fuzzy by our training. It took me about 7 hours to read and made me more aware than ever that aviation is still very much a work-in-progress, that there are still major problems or deficiencies in design, regulation, company procedure, training, checking, simulators, flight safety reporting, ATC infrastructure, ergonomics, even in meteorological knowledge. Once again good and honest pilots and people have perished so that the rest of us can live and learn.

If the PF in the right seat, once the AP disconnected, had done nothing except hold the wings level and the pitch at 2.5 degrees until the ADRs kicked in again with some sensible readings, for about 60 confusing, ECAM flashing seconds, we probably wouldn't have even heard about it. All the PNF had to do was silence the bloody aural, disconnect the FDs and monitor the pitch and roll on his PFD like a hawk.

It took me 7 hours to read the Report and discover why that didn't happen. Of course anyone on PPrune who has followed the threads and read the Preliminary and Interim Reports would already know most of it but the Final puts a lot of flesh and insight on those bones.

My own area of concern is the absence of a Captain on the flight deck and the Report has quite a lot to say on the subject, including a CRM Training recommendation. Personally, I don't think it goes far enough. I believe the travelling public, once the Long Haul stuff started, has been short changed by the operators. For economic reasons, they deemed, with the blessing of the Authorities, that **** only happens below FL200 and that there were no status and role problems with two F/Os sitting together in the cruise. I am afraid AF447 calls that comfortable assumption into question. **** certainly can happen above FL200 and, for the AF skipper, it happened very fast. One and half minutes, to be precise.
I believe a pilot with the status, role, rights and duties, and salary, of a Captain should be in the left seat of an aircraft at all times. Period. Try convincing those surrounding the President of even the most impoverished banana republic that it is OK, in cruise we don't have the money to pay for a full Captain to be on the watch. Flying 300 passengers safely, to the best working practice possible, is flying 300 Presidents. This argument has nothing to say about the proficiency or professionalism of F/Os, which is taken for granted and may well exceed a Captain's. It is simply about responsibility, about having a legitimate person in the hot seat for when and where the buck stops. CRM improvements, more emphasis on briefings, more training for F/Os, all suggest some mystification at work to conceal a less than optimum crew configuration.

I was a bit surprised by how the Report quickly covered the analysis of the failure of the crew to recover from the stall once the horizontal stabiliser had reached its 13 degrees position, presumably still in Alternate 2B law. What could have been done, were the Elevators enough, was Auto Pitch trim still available? I would have expected a recommendation for some very defined jet upset training procedures.

But my main purpose here is not to initiate debate or even to ask any questions. I would simply urge you to read this Report if you are an Airbus pilot. For what it's worth, I have 12 years flying A330/340, 27000+ hours, and 2 months to go to retirement at 65. As a pilot I have always strived for excellence and only ever achieved average. And this Final Report has taught me a lot.