Flight International and Future A350 pilots `learn by doing`

[Centaurus]

As a future passenger I will always want my flight operated by a senior Captain and First Officer who have the knowledge to get us to our destination safely but without the need for them to use skill.

I don't know about "senior" captains. The term "senior" means nothing. Any captain will do as long as he is qualified and competent.

The name of the game is redundancy. There are airworthiness system redundancy requirements for ETOPS and in the design of aircraft in general.
Pilots are also part of the redundancy systems - they just happen to be human beings suitably skilled and certified competent (or should be) in all facets of automatic and manual flying when an operation does not go according to plan.

[CONF iture]

Originally Posted by PJ2

These things are taught, trained and checked against standards.

But UAS in cruise has never been taught, trained and checked against standards.
Airbus had everything to do something about it before AF447 happened but failed to act ...

The necessary was obvious and simple.
I would have loved to practice this.

I'm still waiting ...

[aguadalte]

J. F.:

In reply to “What are you going to do when the autopilot fails?” my answer is that future automatic steering systems will not fail in a critical way. Unlike today’s autopilots which disconnect themselves in the event of a problem, future automatics will be designed to fail safe and carry on performing their functions. Just like today’s wing structures. Autoland thanks to special certification standards has not caused a landing accident since it was first used with passengers in the 70s. Sadly there have been quite a few steering errors by aircrew over the same period.

In the following case, reported in David Learmount's blog, the auto-pilot didn't disconnect. In fact, even with AP disconnected the pilots wouldn't be able to counter-act the Alpha-Prot "stall-protection system" without changing its flight controls rules to ALT LAW.

How would a redundant/fail safe future system cope with this particular situation?

I think human errors on accident statistics should be compared to every day accident/incident avoided cases due to pilot awareness and subsequent actions...

Airmanship lives - Learmount

[John Farley]

aguadalte

By definition a fail safe system is that. Fail safe. Current autopliots are not designed to that spec. So in my view it is not reasonable to have thoughts about future ops based on present systems. A fail safe system will have to be developed from a clean sheet of paper not a few mods to today's kit.

How many years it will take for fail safe systems to evolve and become cost effective in production I have no idea. But I am sure they will in the end because as time goes by it will become clearer and clearer that there is a need for them.

I was 17 when I started a five year apprenticeship at the RAE in 1950 so I grew up surrounded by some of the best boffins in the business. I apolgise if that upbringing makes me lean towards what R&D is capable of producing given time and need.

[Mac the Knife]

"A fail safe system will have to be developed from a clean sheet of paper not a few mods to today's kit."

An AP which on receiving UAS data continues flying pitch and power which annunciating same to pilots does not need a "clean sheet".

"Attention! Unreliable pitot-derived airspeed detected. Maintaining current pitch and power. You must evaluate other AS indications."


In my field, this would be like an anaesthetic machine that just stops ventilating if it finds an error in the blood oxygen saturation detector circuit (which it doesn't, tho' it does generate a warning).

[PJ2]

CONF iture;

But UAS in cruise has never been taught, trained and checked against standards. Airbus had everything to do something about it before AF447 happened but failed to act ...

You're in the business and should already be well aware that training, standards and checking as well as SOPs are up to the individual airline within their OpsSpecs and the JARS/FARS/CARS etc.

The manufacturer supplies a basic SOPs, Abnormal and Emergency drills and checklists as well as the aircraft's essential requirements. Airlines are free to modify them and as you would already know, do modify them, sometimes extensively.

That the UAS event was never taught in cruise in the face of known industry events from 1996 onwards is a question with which individual airlines must come to terms. You are also, or should be aware that UAS events in cruise as well as increased focus on hand-flying skills are now part of recurrent training scripts as a result of the AF447 accident, like Colgan, 'watershed accidents' in my view. Whether such training could have actually prevented an accident like AF447 cannot be determined conclusively.

Regarding the ECAM which had to be dealt with, there exists already an example of a far more complex set of circumstances in QF32, handled successfully by the crew in a real and very serious emergency which actually did threaten the safety of the aircraft.

Re, "Learn-by-doing"...manufacturers aren't stupid so I'll wait and see the details of what Airbus means by this. Some systems knowledge is required but self-diagnosing / self-fixing systems are not an impossibility.

Some of the comments here are, I think, quite thoughtful and prescient and will prove "ordinary" at some future point.

For those raised on nuts-and-bolts courses and high standards for airmanship and flying the airplane this seems a step too far, not, as some have pointed out here, because it can't be done but because with individually-directed play one tends to run out of road very swiftly if all a crew member knows is which button to push to make the airplane go.

The old must give way to the new, often in surprising and unanticipated ways.

[John Farley]

An AP which on receiving UAS data continues flying pitch and power which annunciating same to pilots does not need a "clean sheet"

Couldn't agree more. I have suggested this very thing earlier earlier on one of the 447 threads. It is such a minor re-write of a bit of code that I really hope it happens soon.

[CONF iture]

Originally Posted by PJ2

You're in the business and should already be well aware that training, standards and checking as well as SOPs are up to the individual airline within their OpsSpecs and the JARS/FARS/CARS etc.

How an Airline could be aware of a recurrent issue if the information stays in the Manufacturer headquarters ?

Air Caraibes meeting took place nowhere else than in Toulouse - That one had to be a wake up call where pilots were clearly stating how uncomfortable they had been with the overall situation.

What did Airbus ?

That was 6 months before AF447.

[aguadalte]

John Farley,

Fail Safe philosophy is being used by all manufactures since the Comet accidents. I believe AI was having Fail Safe in mind when building an aircraft created "against pilot error". And it was with that philosophy in mind that Airbus designers decided to add Alfa Prot (and all other protections) as a fail safe device to prevent pilots to over stress their airliners.
Ultimately Fail Safe is a unachievable goal due to the extreme complexity of the aviation world. One can always dream about perfection and science fiction, but the "real world" always finds ways to relegate us all to our humbleness.
And by the way, I was not talking only about AP, I was also addressing AI's FBW flight controls design/laws. And to that regard, I was asking how would it be possible to solve a problem, with a (supposed to be fail safe) design that was receiving wrong inputs by the failure of all AOA probes...
In that particular case, it was up to the pilots to save the day and act as the real "failsafe device".
We will have indeed to wait for a long time, until a complete failsafe system is born. But until then, human intelligence (and experience) will have to be in the top of the chain.

[Clandestino]

Originally Posted by PJ2

Happily, very safe indeed. But we both know that to be meaningful, these must be long-term trends and that isn't the case yet.

Trend is pretty clear: overall airlines safety has been radically improved compared to 90-ies, not to mention 80-ies. In fact, Joe Public of thirty years ago was every couple of months faced with smoking, bodies filled crater of air disaster that happened in the first world, while (despite fantastic growth of air traffic) his modern equivalent is so seldom faced with the ugly truth of air travel being very hazardous that he is easily lulled into belief it is not dangerous at all. So when disaster strikes, attempts to reconcile very wrong picture of the way aeronautical world works with the stark reality result in some spectacularly funny mental gymnastics and that's what made bulk of AF447 threads.

Now, some of us have the privilege of really flying the airliners and with it comes attached access to boring pieces of information contained in internal safety publications that never catch the public eye. There you can see that instruments or autopilot failures are not oh-so-horrible-once-in-a lifetime-events-that-can-get-the-best-of-crews but rather many-times-a-day occurrences that most of the times get handled in a way that makes passengers not realize something is wrong with their aeroplane.

Personally, I don't think flying the bus sim fully manually increased my manual skills a bit; I rather suspect it was intended to weed out those manually inept at the early stage. Unsurprisingly, no one ever failed it at my gang; those who didn't make it were usually unable to get to grips with FMGS.

Originally Posted by PJ2

Even as the experience and set of skills are called upon less and less, knowing how to fly an airplane is a fundamental and required skill that can't be replaced.

Exactly! Overwhelming majority of pilots when faced with trouble aloft performs safely, even if it is their first time they see such a thing or they were never trained for such an eventuality. If you think you'd read about it on Avherald, chances are small. Not even every tenth safety occurrence we had was reported there and we are not operating in some backwaters - unless FRA, MUC, CPH, LHR, FCO, CDG could be considered as such.

Originally Posted by John Farley

To be the Captain or First Officer of a future airliner is an ambition that any youngster should be proud to have.

Hopefully so. Present is pretty bleak, unless one has some private means and flies half the normal hours for fun.

Originally Posted by PJ2

I disagree that "luck" has anything to do with the AF447 accident or for that matter most accidents.

Yup. We call it "fate". It is still the hunter.

Now I have your attention, I'd like to say that world does not conform strictly to Newtonian mechanistic cause-effect view. It's nature is so complex, our best chance of making sense of it is to think of it in statistical terms. Pilot who is dedicated to his profession stands far better chance of making it to retirement than some ignorant risk-taker but this is still a chance; he might be abruptly terminated if the circumstances conspire against him while the second one could have long life of minima-busting, scud running, overweight take-offs etc. Folks who have problem understanding this are those who believe that if they got away with e.g. taking-off with snow on the wings, they will get away every time.

Originally Posted by PJ2

I don't know anyone who has flown transports and also done work in flight safety who believes in "luck".

Must be a cultural thing. I don't know anyone who believes otherwise but then every pilot I know 1) knows it should not be pushed under any circumstance 2) would turn away if he suspects that it would take to have luck involved even in ppms to execute the original plan successfully. Presence of luck is oft acknowledged post festum.

Originally Posted by PJ2

Thirty crews were not "lucky", they responded in a manner which maintained control of their aircraft,

Yup. Even as most of them did not responded in prescribed manner. Makes one think.

Originally Posted by PJ2

It is not "just plain unlucky" that one does not execute the standard operating procedures as one has been trained to do in responding to an abnormality nor is it unlucky that in an abnormality one does not communicate with the other crew member(s) as to what one is doing with the airplane and instead launches on one's own without the other crew member knowing or understanding what one is doing.

Correct. This is not unlucky. This is suicide through panic.

Originally Posted by Ventus45

UAV technology has arrived.
Pilot(s) are not (operationally) required, at any level, any more.
Groundies, dispatchers, atc, can all do the necessaries.

If you have a slightest idea how to make this statement of yours true, USAF would like to have a word with you.

Originally Posted by Mac the knife

May I venture that the reason is that it is simply too expensive in terms of time and money to train flight crew to "steer", as he puts it, modern large airliners.

Pilots are cheap. Lots of them nowadays come with negative price. Steering is easy. Knowing where to steer and when is a bit more difficult.

Originally Posted by Mac the Knife

Gaining (and maintaining) the requisite manual flying skills to cope with what is becoming an extremely rare eventuality is just no longer practical.

Was always. Still is. It's the matter of perspective - and those booooooring safety bulletins.

Originally Posted by Gretchenfrage

Clandestino's gems ....

Ah, just the usual balanced and well supported arguments we are used to receiving from this source.

Originally Posted by Gretchenfrage

... is best put back into oblivion by a quote of a REAL professional:

Now you mention it; both Habsheim and Hudson have one more common trait besides both involving A320: both aeroplanes hit the deck in alpha floor (EDIT: I had a brainf@rt here, I meant alpha prot, not alpha floor, alpha floor is autothrust mode which was not applicable in either case) with significant aft stick displacement. If flight controls were classical or brand B FBW with overridable limits, outcome would be much worse.

Originally Posted by PJ2

The point being made is that what kept the thirty-odd crews who also experienced this abnormality from becoming flummoxed was SOPs, CRM and ensuring the priorities of "aviate, navigate, communicate", were adhered to.

Not always. A lot of the crews were utterly puzzled but they did nothing - which turned out to be the safe thing to do.

Originally Posted by Artie Fufkin

Presumably the conversion course will start in the sim, not in the classroom? Seems all very sensible to me.

Could be. On my 320 course we didn't follow classic "first groundschool then sim" but before each abnormal session dedicated to specific system, we'd go through CBT.

Originally Posted by CONF iture

But UAS in cruise has never been taught, trained and checked against standards.

So? Other AF crews negotiated UAS successfully.

Originally Posted by CONF iture

Airbus had everything to do something about it before AF447 happened but failed to act ...

Act upon what? Non-events? In every case there was pitot blockade, then there wasn't. Everyone passed through it without a scratch. It was reasonable to require pitot replacement but there was no reason to make it high priority.

Originally Posted by aguadalte

I think human errors on accident statistics should be compared to every day accident/incident avoided cases due to pilot awareness and subsequent actions...

That would be smart thing to do if we are really interested in meaningful aviation safety discussion.

Originally Posted by John Farley

I apolgise if that upbringing makes me lean towards what R&D is capable of producing given time and need.

And financing, of course. That's why we still fly ILS and not far superior MLS.

Originally Posted by Mac the Knife

An AP which on receiving UAS data continues flying pitch and power which annunciating same to pilots does not need a "clean sheet".

Unfortunately, it needs not just clean sheet, it needs computer vastly superior to anything we have today.

Every computer can recognize failed speed. None can unreliable. Comparing so radically different categories as pitch, power and speed to see which one is wrong cannot be done without applying intelligence and real intelligence is what every computer so far lacks.

Originally Posted by John Fairley

It is such a minor re-write of a bit of code that I really hope it happens soon.

It would be pretty major as, unlike ATR or DHC-8, Airbus has no attitude hold mode. Her autopilot is actually simpler than those of typical turboprop. I haven't flown many types but A320 was smoothest and easiest to fly of any, even in direct law where there is direct stick to controls displacement and only manual trim is available through trim wheel. It was very easy to precisely set the desired attitude on her. Trick (very easy one) was knowing which attitude was the right one.

Originally Posted by aguadalte

Ultimately Fail Safe is a unachievable goal due to the extreme complexity of the aviation world.

Absolute fail safe is indeed unachievable. When something aeronautical is certified and referred to as fail safe, it means it is fail safe except when one-in-a-couple-of-millions occurrence happens. HTBJ is still very much relevant.

[DozyWannabe]

Originally Posted by CONF iture

How an Airline could be aware of a recurrent issue if the information stays in the Manufacturer headquarters ?

Air Caraibes meeting took place nowhere else than in Toulouse - That one had to be a wake up call where pilots were clearly stating how uncomfortable they had been with the overall situation.

What did Airbus ?

You mean apart from setting a programme in motion to replace the problematic pitot tubes which was underway, and publishing bulletins to airlines notifying them of recent UAS issues and of the importance of UAS procedures - both of which happened prior to AF447?

You can argue whether they were slow off the mark, but it's not like they did nothing.

[CONF iture]

Originally Posted by Clandestino

Now you mention it; both Habsheim and Hudson have one more common trait besides both involving A320: both aeroplanes hit the deck in alpha floor with significant aft stick displacement. If flight controls were classical or brand B FBW with overridable limits, outcome would be much worse.

Pure disinformation on both reports.

Ah, just the usual balanced and well supported arguments we are used to receiving from this source.

Sarcasm applied to yourself.

Originally Posted by Dozy

publishing bulletins to airlines notifying them of recent UAS issues and of the importance of UAS procedures

Where are they ... ?

[Slasher]

Gaining (and maintaining) the requisite manual flying skills to cope with
what is becoming an extremely rare eventuality is just no longer practical.

Somehow I think Sully would disagree with that.


Rare events, "isolated incidents" (how I hate that bloody phrase!), "freak accidents",
etc - only need to happen once. And its THAT which I diligently maintain my manual
flying skills, even on the 320 suck-squirt.

Show me an event where the FMC, A/P or A/T saved an aircraft and I'll

[KAG]

Show me an event where the FMC, A/P or A/T saved an aircraft and I'll

It will be a non event by definition, so impossible to show you.
Anybody on pprune knows A/P, A/T... saved the day many times.
That's obviously not a reason for giving up on flying skills, but ignoring the role and task of the A/P in the thousands daily worldwilde flights is simply amazing.

[Clandestino]

Originally Posted by CONF iture

Pure disinformation on both reports.

1. it's not disinformation, it is my mistake; I meant alpha prot but wrote alpha floor. Anyone knowing a bit about Airbus FBW would recognize the error. Alpha prot is AoA limiting feature of FBW, alpha floor is automatic application of TOGA thrust at high AoA which wasn't applicable in either case. At Habsheim through autothrust being deactivated and aeroplane flying below 100 ft RA anyway while engines being wrecked is what precipitated the Hudson ditching. Alpha floor is autothrottle mode also available on B737NG and there it's called... alpha floor.

2. if you substitute "floor" for "prot", does it contradict the official reports?

Now sit and watch the outburst of usual tales how investigating bodies hate the pilots and protect the evil Airbus.

3. Ah, just the usual balanced and well supported arguments we are used to receiving from this source.

[Centaurus]

Airbus plan to use hands-on learning process for pilots to familiarise themselves with the aircraft and the manual handling characteristics,

It will be instructive to view the additional training syllabus that Airbus plan to use to teach pilots how to fly the Airbus by hand.

Until that syllabus is published we will not know whether it is just another motherhood statement (hot air) or something that really is revolutionary () like practice circuits and landings with and without strong crosswinds, raw data manual throttle instrument approaches and circling approaches without waypoints spread all over the MAP mode. Unless this revised training syllabus means flying without the flight director and autothrottles, it is doomed to being a waste of time as far as cadet pilots and those new to type are concerned

[Slasher]

Anybody on PPRuNe knows A/P, A/T... saved the day many times.

Ok KAG - let's see how many Pruners can quote events where the fancy stuff
saved the day. You didn't provide any examples which doesn't surprise me in
the least.

[KAG]

If you beleive the A/P and A/T have never saved anybody's day, so there is no need to talk to you, at that level I am afraid there is nothing we can do for you.

[deptrai]

I am still calling you in to provide even just one example where automatics saved an aircraft

most likely you won't hear much about such examples, exactly because they became non-events. What qualifies as "automatics"? Does ILS count as automatics? I wouldn't fly with the eagle-eyed supernatural hero pilot who thinks he can do better without, in low visibility.

Columbus didn't cross the Atlantic using ancient Roman technology either. It's easy to vilify (new) technology, but learning to use what you have at your disposal it in a safe way, and maximise it's benefits will probably get you farther.

[Slasher]

If you beleive the A/P and A/T have never saved anybody's day, so there is
no need to talk to you, at that level I am afraid there is nothing we can do
for you.

Who's "we"? There's no "we" when it comes to you Mr KAG, or haven't you realised
that yet?

Anybody on PPRuNe knows A/P, A/T... saved the day many times.

I am still calling you in to provide even just one example where automatics saved
an aircraft.

When you find you can't I suggest you retract your statement.