Pilot handling skills under threat, says Airbus

[Microburst2002]

nitpicker

the airbus does not fly as you might expect from reading the manuals. According to these, it seems that is like CWS, or even like ALT HOLD or even that if you are in glide and loc and just keep stick neutral the airplane will make a perfect ILS, even in turbulence. pure bull****.

It is easy to fly, that's for sure, but the only difference is that you don't feel anything in the stick, you don't have to trim and it is very stable in bank durign turns. You still have to adjust pitch and bank and thrust to obtain the desired performance, and do this continuously. And thrust still affects pitch, and gusts will still affect bank angle. The fact that they are meant to be flight path stable rather than speed stable has pros and cons, and I think it is a necessity when you have no feeling in the stick.

conventional airplanes are easy to manoeuvre when you get used to trim properly. If you are fighting against stick forces you are not using the correct technique. there is no need for this technique in the bus, that's true. Nor there is need for an artificial stick force system. the bus is flown very easily, but you stil have to fly it.

[Centaurus]

however, I do not believe that a practical solution can be found either within the manufacturers’ use of technology or necessarily with ‘more training’ in hand flying skills.

Safetypee. Excellent reply and thanks for going to the trouble of writing. Would it be true to say that all pilots should be equally competent at automation skills and pure flying skills? After all, you would think that is a given. But loss of control accidents have proved otherwise - and that is the present day problem.

[Microburst2002]

Very interesting presentation

I'm afraid that the ETTO in aviation is the norm (or is it the TETO?), from the point of view of flight deck crews, at least. Efficiency always wins the battle, but lack of thoroughness will be always blamed when things go wrong.

And things going well are very seldom studied. Aviation learns from crashes, and rarely from successful flights.

In my opinion, the problem underlying here is liability. He who dares to encourage hand flying is taking a risk, should one day happen an accident where the pilot was hand flying. However, it is accepted that airplanes are normally flown on AP, besides if they crash you can still blame the pilot (maybe for not taking over manually, or doing it catastrophycally).

If I had to do Resilient Management for air safety I would make sure that pilots are skillful hand flyers first and then skillful automation flyers. Because in this manner they would be much better automation flyers (they would understand automation and its limitations and shortcommings better) and they would be much safer in case of AP disconnection or malfunction (either apparent or subtle). Focusing in only one thing is dangerous.

A pilot who will not feel confident if he has to hand fly is a swisscheese slice with too big a hole in it. A symptom of this is reluctance to use manual thrust after A/THR (during a simulator) disconnects on its own or is not functioning properly (occasionally in real flights). Another one is the use of AP for visual approaches.

[AvMed.IN]

So, we are back to the basics of Human Error and Human Factors!

[EIDWSkypilot]

Can we see the evolution towards the flight deck of the future, crew of one with dog. Handler to feed dog, dog trained to keep Handler from touching controls. Both there to take the blame when anything goes wrong. SLFs reassured that dog has security training.

[safetypee]

Loss of control or lack of control – which end of the piece of string?

Centaurus,
Would it be true to say that all pilots should be equally competent at automation skills and pure flying skills? Mmm … a challenging thought.

This depends on what is meant by competency – “A collection of related abilities, commitments, knowledge and skills that enable a person (or an organisation) to act effectively in a job or situation.” This is a vague definition, like saying ‘how long is a piece of string’.

Competency, like string, depends where the ends are – the initial training and the final objectives, and where the competency is to be used. What are the abilities, knowledge and skills; these depend on the context, the situation.
Furthermore, I would hesitate to compare automation skills with flying skills. Removing the physical coordination aspects, then cognitive skills predominate. Superficially, these would not differ between old (manual) and new (auto) aircraft, however the sensed information and display media may be different, changes in tactile feedback, information content, and computer aided decision making.

If these changes are significant for cognition, I think they are, then pilots require a different set of skills and competencies for automatic flight, but then you still need back-up for the occasions when automation is not providing assistance. In these circumstances, the manual flight skills in auto aircraft are similar to older aircraft, but the level of competency may not have to be the same, only enough to cope with the foreseeable situations. As is realized with AF 447 not all situations are foreseeable.

In other LOC accidents, foreseeable circumstances were encountered but were not recognized or they were handled incorrectly (also factors in AF 447).
IMHO this suggests weaknesses in situation assessment and decision making skills. Thus the issue is whether these skills are absent or need to be tailored for automated aircraft operation, including failures; probably some of each.
These would not have to be specific skills, but generic abilities to assess and think about situations, using different / novel information sources, act without undue rush, and keep an open mind as situations develop. These stem from experience; so another conclusion may be that the industry is not providing an appropriate level of experience for automated flight, but not necessarily the same as for manual flight, i.e. we have been using the wrong training context.

Thus at this time I would conclude that it is not necessary to have equal competencies for manual and automatic flight, but it is necessary to have an appropriate set of competencies; and some accidents suggest that the existence of these is questionable.

[safetypee]

Tie the ends of the string together.

Microburst2002, “but lack of thoroughness will be always blamed when things go wrong”, possibly, but as with ‘Human Error’ (AvMed.IN) it need not be.
The industry has to move away from error based assessment in order to learn about accidents.
We have to understand what happens in daily operation, why pilots don’t hand fly when they could, why they apparently see and act on aspects which with hindsight appear obvious. We must look at every day operations for this understanding and for the precursors to accidents; this is not entirely FOQA or even LOSA, it is something which every pilot can do independently – a self line check – ‘why did I do that’, ‘what did I see’, ‘what do I understand and why’. And of course we require time and some ability (competency as above) to do that.

“If I had to do Resilient Management for air safety …”, why not use self initiated Resilient Management.
We could use How resilient is your organization for guidance. The essential items can be applied to individuals:-

• The ability to Respond. – Knowing what to do and being capable of doing it.
• The ability to Monitor. – Knowing what to look for.
• The ability to Anticipate. – Finding out and knowing what to expect.
• The ability to Learn. – Knowing what has happened.

These too are flying skills, we must take care not confuse the apparent lack of ‘hands-on’ (physical hand / eye co-ordination) flying skills as an indicator of the problem.

For guidance other than FOQA / LOSA, consider Day-2-Day Safety Survey (page 10).
What would be the key aspects to observe in the flight deck?
Would these help tie the ends of the string together – normal operations // accident contributions?

[mm43]

Originally posted by bubbers44 - post #444 ...

I'm not sure what you are trying to say but a bad pilot getting too slow and approaching a stall above the water would set up a significant sink rate impacting the Hudson, Sully knew how to control the energy and zero out his sink rate at touch down. Had nothing to do with Airbus engineering.

No one is saying that Sully didn't do a great job, and the alpha protections that the A320 has in Normal Law also aided him by attenuating his final NU commands during the flare as RoD was over 3 times higher than desired due to the aircraft having descended at VLS or less. The VLS warning was over-ridden by the GPWS, and any further NU would most likely have resulted in structural failure at impact.

So, Sully and Airbus had a good compromise, though I doubt Sully at the time recognized that alpha protections were assisting.

[stepwilk]

There is a book written by a US journalist - I'm sorry but forget the name of both - which tells the story of the Hudson airbus.

That writer is William Langewiesche, whose father, Wolfgang Langewiesche, wrote "Stick and Rudder." The title of his book is "Fly by Wire." William and I, and my wife, spent many hours flying together, back when all three of us were baby pilots.

[CONF iture]

and any further NU would most likely have resulted in structural failure at impact.

mm43,
To the contrary, a further NU, as requested by the pilot, would have allowed the aircraft to reach the 11 degrees of positive attitude as recommended by Airbus for ditching, and would most likely have resulted in a better touchdown and no or less structural failure at the aft underside part of the fuselage.
The potential was there in the AoA value, but the aircraft refused to deliver up to alpha max ...

[Centaurus]

This depends on what is meant by competency –

The answer is easy. If an airline pilot (Boeing/Airbus) was told to undertake his instrument rating sequences solely on raw data, and no automatics, then at an educated guess he would probably fail to remain within regulatory tolerances in most areas. In other words he has failed the skills test and by definition is incompetent.

Currently, instrument rating tests in sophisticated airline simulators are 90 percent flown on full automatics with an occasional nibble at hand flying. That does not necessarily add up to manual flying competency. At an Asia Pacific Flight Safety Forum in recent times, a senior management speaker made the point that practicing a manually flown ILS once a month on line does not ensure currency at manual flying.

[safetypee]

Centaurus, thanks, your post (#459) puts ‘competency’ in context; in this instance it relates to modern operations using highly automated aircraft.
There appears to be an assumption (big assumption, but by who) that normal operations will be flown using the automatics and thus appropriate skills (competency) for these are required. What these are, or how are they checked requires further explanation.

This ‘big assumption’ also appears to restrict manual flying skills to non-normal or abnormal conditions, - when the autos are unavailable or unable to cope with conditions. If the assumption, as is likely, considers the frequency of encountering these conditions is low, then risk management might allow for a lower skill level; but even so, the skill (a subset of manual flying) and the level of competency have to be defined. Who does this, how is it achieved?

An important aspect of safety is to write down all of the assumptions, particularly the situations in which they apply. Also state the justification for the assumption including action to be taken by operators.
A glaring omission might be the prohibition of manual flight in normal operations on the basis that the skills are not checked, or if checked they are of a lower standard, thus involve higher risk.
If this is so, then it appears contrary to current thoughts on cognitive skills training / currency required for automatic flight.

What assumptions are published on this issue?

[vaneyck]

Langewiesche

I enjoy William Langewiesche's reportage very much. I used to subscribe to the Atlantic, and now I subscribe to Vanity Fair, largely due to his presence as a staff writer. That said, bubbers44 points up something I've noticed frequently in Langewiesche's writing: He tends to identify with the point of view of his main sources. The latest example of this is a Vanity Fair article in which he makes a hero of an extreme surfer whom most other surfers would probably go miles to avoid.

So I wasn't surprised that he bought the Airbus line to a possibly excessive degree. He admired Bernard Ziegler (rightly), and fell in love with his way of looking at matters, resulting in a lack of balance in his article and book on the Hudson River incident.

[mm43]

The potential was there in the AoA value, but the aircraft refused to deliver up to alpha max ...

CONF iture,
Why did the aircraft fail to deliver up to alpha max ??

It didn't because it would have stalled!

That's the NTSB's take on it in their Final Report page 97, 2.7.2 High-AOA Envelope Limitations;

The airplane’s airspeed in the last 150 feet of the descent was low enough to activate the alpha-protection mode of the airplane’s fly-by-wire envelope protection features. The captain progressively pulled aft on the sidestick as the airplane descended below 100 feet, and he pulled the sidestick to its aft stop in the last 50 feet, indicating that he was attempting to raise the airplane nose to flare and soften the touchdown on the water. The A320 alpha-protection mode incorporates features that can attenuate pilot sidestick pitch inputs. Because of these features, the airplane could not reach the maximum AOA attainable in pitch normal law for the airplane weight and configuration; however, the airplane did provide maximum performance for the weight and configuration at that time.

.... The flight envelope protections allowed the captain to pull full aft on the sidestick without the risk of stalling the airplane.

They say that the alpha protections prevented any further NU.

[Morden]

Just read the NTSB report on Flight 1549, very interesting, thanks for the link!


mm43:
There is a AoA threshold from the activation of the alpha protection to the maximum flyable AoA (p. 11). So in a sense, the aircraft did failed to attain the maximum alpha.

Here is one point (p. 96) from the difficulties mentioned by the NTSB report on ditching the Bus:

• Deliberately or inadvertently slowing the airplane into the alpha-protection mode may result in an attenuation of pilot nose-up stick inputs, making it more difficult to flare the airplane, even if AOA margin to alpha maximum exist

!

"They say that the alpha protections prevented any further NU."

Exactly. And it is NOT necessarily a good thing in this case.

And I firmly believe that the Captain (even if he didn't kept up the airspeed quite correctly), would have not pulled the stick like a mindless monkey if there weren't any protection.

I believe he would have stalled the airplane just a few inches above the water

[PEI_3721]

US Congress report (by a group of experts).

The soon to be published Congressional report on airline pilot training is reported to call for commercial pilots needing enhanced manual flying skills, improved leadership abilities, and greater access to advanced simulators.
Airlines and regulators should rethink how U.S. flight crews are chosen, trained and checked and recommends changes in initial and recurrent training for airline pilots.

WSJ UPDATE: Industry Study Urges Big Shifts In Airline Pilot Training

[DozyWannabe]

@mm43, CONF iture

Indeed - Alpha Max was attained, but the AoA that represented Alpha Max was a shallower angle than usual because of the low-energy configuration due to double engine failure (in mathematical/software/engineering terms, Alpha Max is a function of several constantly changing parameters, not a hard-coded value or constant). It looks like Sullenberger used his knowledge of the systems as a backstop while he attempted the optimal AoA possible. Ultimately in Normal Law the FCS will do it's best to prevent the aircraft from stalling with the tools at it's disposal. If it doesn't have thrust as part of its toolset (in the case of A/THR disable, low altitude inhibition or, indeed, engine failure), it will use the flight controls until either the thrust is increased to compensate, or in the case of complete dual engine failure, you run out of enough forward motion to keep you in the air. This behaviour has been seen before, but unfortunately it was not as well-understood at the time, and that story did not have a happy ending. However in that case the aircraft was new, and by the time of the Hudson incident it had been in service for around 21 years and was well understood by crews.

[CONF iture]

Alpha Max

It didn't because it would have stalled!

No it would have not.
Do you remember what is alpha max mm43 ?

The System behaved as designed.
That design is not described in the manual.
That system thinks it knows better than an experienced pilot.

Originally Posted by Dozy

Alpha Max was attained

No it was not on the Hudson.
As it was already not in 1988 ... but keep Habsheim for the PM, no one needs to know how the BEA was already in damage control mode.

[silverstrata]

Please correct me if I am wrong, but I understand that the system will deliberately limit body angle below 50ft (ie, not achieve alpha max), to prevent a tail strike on landing.


.

[hetfield]

Engineering skills under trheat, say Pilots.

e.g. AF447