"Pilots are no longer trained to fly planes..." - David Learmount

FWVT

And I totally agree with you, David.

In today's Indo

http://www.independent.ie/opinion/comment/pilots-are-no-longer-trained-to-fly-planes-they-cannot-cope-if-the-computers-fail-30872720.html

We now know that Air Asia flight QZ8501 crashed. And while we cannot be absolutely certain why it did so, the recent history of aviation disasters, which I have studied as a safety analyst for the last three decades, gives us a very good idea.

SHARE
GO TO
Over the past 20 years, 1,800 airline passengers and crew have been killed because pilots simply lost control of their aircraft.

Loss of control in-flight, not caused by any structural or technical problem, has become established as the aviation accident category responsible for more passenger deaths than any other.

Back in the 1970s and earlier, airliners used to crash far more often than today, but for different reasons. Engineering and design then were less advanced, so technical failures of the engines, airframes, or the pilots' instruments (avionics) were far more common, and the aircraft themselves were more difficult to fly. But in those days, loss of control by the pilots of a sound aircraft hardly ever featured.

The US Federal Aviation Administration has spent five years examining why perfectly airworthy planes now just fall out of the sky. The FAA's John Hickey explained the rationale behind the research: "An evolution has occurred in that pilots, with their training, have become somewhat incompatible with the way airplanes are built and designed today.

"When we look at accidents over the last 10 years, there is a common theme in the US and globally. There is a fundamental misunderstanding going on between man and machine. We have to fix that."

The FAA's study, published last year, was called 'The Operational Use of Flight Path Management Systems'. That rather turgid title refers to the computers that manage today's autopilots. The pilots have to manage the computers. Pilots touch the traditional flight controls only for a couple of minutes in every flight, and spend much of their time programming and monitoring the FMS - the flight management system. The FMS is their main computer interface with the aircraft, and directs its flight via the autopilot.

The FMS flies and navigates the aircraft more accurately and smoothly than a pilot could, and hardly ever fails. The pilots learn to trust it. The FAA sums up the difference between the 1970s and now: "Managing the flight path of an aircraft is a basic pilot responsibility, but the means for accomplishing it have evolved considerably."

In other words, pilots used to fly by brawn and brain. Now they do it by computer, with the brawn and brain as a backup. That, at least, is the theory. In practice, most pilots have not been trained for the modern piloting task. Regulations regarding what they must regularly practise - by law - are way out of date. A few airlines are voluntarily adopting evidence-based training - tailored to pilot needs rather than out-of-date regulations - but only a few.

Not that pilots can be completely replaced by automation. This is because only one in 10 flights is completed as originally planned; as events change, so pilots are able to mitigate the risks.

The FAA working group established that today's pilots have a number of "vulnerabilities". The prime one is that if the automatics fail, the pilots are no longer practised in managing without them. This leads pilots to lose confidence in their own traditional flying abilities, so when things go wrong they have a tendency to try to restore failed automatic systems when, in fact, they should be flying the aircraft to keep it safe.

Now back to the AirAsia accident. Control was lost shortly after the pilots requested clearance to climb and alter course to avoid storms. This has not been the only such fatal crash recently: three others have taken place under remarkably similar circumstances.

On July 24 this year, an Air Algerie flight over West Africa also went missing shortly after the pilots reported changing course to avoid storms in their path.

About two days later the wreckage was found in the desert in Mali. In June 2009, an Air France Airbus A330 - flight AF447 - in which three Irish doctors were travelling, was lost on a Rio de Janeiro-Paris route over the South Atlantic. The pilots were manoeuvring to avoid storm clouds at the time. The aircraft's "black boxes" were eventually recovered, revealing that the pilots had been confused by a momentary loss of airspeed information.

The autopilot relinquished control to the pilots, but they quickly lost control of the aircraft. The icing that had temporarily blocked the external airspeed sensors was related to the storms close by.

In another weather-related loss in August 2005, a West Caribbean Airways Boeing MD-80 flying from Panama City to Martinique went out of control while the pilots were discussing how to deal with engine icing the aircraft seemed to be suffering in high-level cloud. The crew told air-traffic control that the plane could not maintain altitude. The stall warning - telling the pilots the aircraft was flying too slowly to generate lift - operated during the descent, which took just 210 seconds from cruising at 33,000ft to hit the ground.

In all three cases, there was nothing wrong with the aircraft that would have prevented the pilots retaining control if they had been practised at operating without the automatics. But they were not. The loss of the AirAsia A320 and all 162 people on board may well add to this sad catalogue of avoidable accidents. (© Daily Telegraph, London)

David Learmount is operations and safety editor for Flightglobal

sailing

David Learmount, the Eamon Dunphy of Aviation.

airbuspilot

Not true. Its a bit of a generalisation to say pilots cant fly their aircraft without automation. All he has done is name out 3 accidents where pilots made mistakes. It is not that they dont know how to fly their aircraft they made a mistake which unfortunately does happen occasionally. In my opinion the problem wasn't their lack of flying ability without automation.

Also, there are a lot of examples which show that pilots can fly. Chesley sullenberger us1549, northwest 85, taca 110, Qantas 32 to name a few

EchoIndia

Crashes directly from the en-route phase of a flight used to be very rare (unless caused by deliberate action such as a hijacking or bombing) and approach or take-off have been the phases when most accidents occurred. While every accident is attributable to multiple factors, what Learmount says is not a view only held by him. I would place him, as an aviation professional for several decades (ex-RAF C-130 pilot) well above the Dunphy category; he is able to support his assertions with evidence. The extremely high levels of safety that we experience today are due in large part to learning from past events. No doubt there will be further lessons from the AirAsia crash as well.

FWVT

airbuspilot wrote: »

Not true. Its a bit of a generalisation to say pilots cant fly their aircraft without automation. All he has done is name out 3 accidents where pilots made mistakes. It is not that they dont know how to fly their aircraft they made a mistake which unfortunately does happen occasionally. In my opinion the problem wasn't their lack of flying ability without automation.

Also, there are a lot of examples which show that pilots can fly. Chesley sullenberger us1549, northwest 85, taca 110, Qantas 32 to name a few

Look at the Turkish Airlines at Amsterdam, AF447, The Korean Airlines at San Fran this summer, etc. Especially that last one, a complete failure to remember the basics of flying an aeroplane, watch your airspeed. Automation does it for you, but when it fails to...

airbuspilot

FWVT wrote: »

Look at the Turkish Airlines at Amsterdam, AF447, The Korean Airlines at San Fran this summer, etc. Especially that last one, a complete failure to remember the basics of flying an aeroplane, watch your airspeed. Automation does it for you, but when it fails to...

Yes I agree that all the mentioned accidents were bad piloting but the statement that pilots can't fly is untrue. The majority of pilots can.

FWVT

airbuspilot wrote: »

Yes I agree that all the mentioned accidents were bad piloting but the statement that pilots can't fly is untrue. The majority of pilots can.

Of course not all, but he is refering to the TREND in training. Basic flying skills are on a downward trend, and as older, more experienced traditional pilots become replaced by younger computer programmers, this trend must surely be set to continue.

airbuspilot

FWVT wrote: »

Of course not all, but he is refering to the TREND in training. Basic flying skills are on a downward trend, and as older, more experienced traditional pilots become replaced by younger computer programmers, this trend must surely be set to continue.

Young pilots are trained the same if not better than before.

FWVT

airbuspilot wrote: »

Young pilots are trained the same if not better than before.

But trained to do what? Fly an Airbus by the seat of their pants? Not too well, I imagine.

airbuspilot

FWVT wrote: »

But trained to do what? Fly an Airbus by the seat of their pants? Not too well, I imagine.

No their are first taught the basics of flying in cesnas and small multi engine planes like a piper Seneca and then learn to fly a jet which depending on the company may be an airbus. I fail to see how you are concluding that they don't learn to fly the airbus without automation because they do.

Icepick

1800 dead in 20 years. It's almost as serious as choking on a peanut. Almost.

FWVT

airbuspilot wrote: »

No their are first taught the basics of flying in cesnas and small multi engine planes like a piper Seneca and then learn to fly a jet which depending on the company may be an airbus. I fail to see how you are concluding that they don't learn to fly the airbus without automation because they do.

What I'm saying is they don't get to practice it, so whatever bit of hands-on training they do get in their rating how many times do they put it into practice?

Of course you seem to be one of these pilots so you do know the score a lot better than me (a PPL), but I also see this trend in my line of flying too. It's almost laughed at now in the States to fly a non-glass cockpit, and reading publications like Flying, etc., this ethos is reinforced. Use the autopilot as much as possible, buy that G1000, etc. That's fine, but what about in a failure?

airbuspilot

FWVT wrote: »

What I'm saying is they don't get to practice it, so whatever bit of hands-on training they do get in their rating how many times do they put it into practice?

Of course you seem to be one of these pilots so you do know the score a lot better than me (a PPL), but I also see this trend in my line of flying too. It's almost laughed at now in the States to fly a non-glass cockpit, and reading publications like Flying, etc., this ethos is reinforced. Use the autopilot as much as possible, buy that G1000, etc. That's fine, but what about in a failure?

I'm not an airline pilot either just an enthusiast(hopefully a pilot in the future though)I agree with your point that pilots are encouraged to use automation as much as possible but I still think that they do know to fly apart from a few incidents already outlined.

Grandeeod

I'm not a pilot or a particular enthusiast, but as a regular flyer and someone who takes an interest in aviation due to my location, I am concerned about this.

concerns about degradation of pilot knowledge and skills” because modern aircraft are so technologically advanced.

I am very interested in hearing opinions on it. I appreciate that it can be offensive to actual pilots and indeed enthusiasts to an expected lesser degree. However from experience in railway forums, I'm fully aware of how actual train drivers and train enthusiasts are equally defensive of their trade/hobby. Is it the same in the airline industry or is there a more open dialogue to be had regarding this issue?

Tenger

FWVT wrote: »

Of course not all, but he is refering to the TREND in training. Basic flying skills are on a downward trend, and as older, more experienced traditional pilots become replaced by younger computer programmers, this trend must surely be set to continue.

airbuspilot wrote: »

No their are first taught the basics of flying in cesnas and small multi engine planes like a piper Seneca and then learn to fly a jet which depending on the company may be an airbus. I fail to see how you are concluding that they don't learn to fly the airbus without automation because they do.

The article gives a sensationalist spin to it but the AF447 crash really brought the hands-on piloting experience vs FMC automation discussion to the fore.

No-one here would argue that new flight crew are not trained to fly. The question that arise's is the overall level of hands on experience that modern commercial pilots receive in their career. Look back 30-40 years, most commercial pilots were ex-military. The aircraft they flew needed to be handflown too.

With the dominance of the safety at the lowest cost mentality amidst the aviation industry the financial/accountancy depts are looking at how they can meet training requirements in the most cost effective way. At the same time airline fleets are becoming more rationalised and most aircraft are being designed in a similar way.

To take a modern example.......a Euro airline with a single type fleet starts a cadet program. They engage a training academy to train their successful applicants. The 30-40 cadets (selected rigoursly from among 5000+) go through 18 months of training and within 2 years of selection find themselves at the controls of an A320/B737.
Now no rules have been broken or safety corners taken......but the overall experience of these new recruits is less than was the norm 30 years ago.

Even looking nearby about 20 years ago. I can remember when EI operated a fleet with approx 5-6 types. (F50, Saab, Shorts, B737, Bac-111, B747)
So any new pilot out of the academy worked his way through the fleet and got lots of hands-in experience over his first 10-15 years.

Since the AF447 incident airlines have been addressing some of the issues highlighted. But it will take a while to filter through the industry

Tenger

Icepick wrote: »

1800 dead in 20 years. It's almost as serious as choking on a peanut. Almost.

The mods have been quite busy over the last 3-4 days. Don't make us impose another ban.

While I acknowledge your statistical input we do not appreciate you belittling the deaths of 100's of people.

Grandeeod

Tenger wrote: »

The article gives a sensationalist spin to iy but the AF447 crash really brought the hands-on piloting experience vs FMC automation discussion to the fore.

No-one here would argue that new flight are not trained to fly. The question that arise's is the overall level of hands on experience that modern commercial pilots receive in their career. Look back 30-40 years, most commercial pilots were ex-military. The aircraft they flew needed to be handflown too.

With the dominance of the safety at the lowest cost mentality amidst the aviation industry the financial/accountancy depts are looking at how they can meet training requirements in the most cost effective way. At the same time airline fleets are becoming more rationalised and most aircraft are being designed in a similar way.

To take a modern example.......a Euro airline with a single type fleet starts a cadet program. They engage a training academy to train their successful applicants. There 30-40 cadets (selected from among 5000) go through 18 months of training and within 2 years of selection find themselves at the controls of an A320/B737.
Now no rules have been broken or safety corners taken......but the overall experience of these new recruits is less than was the norm 30 years ago.

Even looking nearby about 20 years ago. I can remember when EI operated a fleet with approx 5-6 types. (F50, Saab, Shorts, B737, Bac-111, B747)
So any new pilot out of the academy worked his way through the fleet and got lots of hands-in experience over his first 10-15 years.

Since the AF447 incident airlines have been addressing some of the issues highlighted. But it will take a while to filter through the industry

You have somewhat answered the questions I asked in my last post,

Turbulent Bill

In my own (entirety unqualified) opinion, increasing automation and then expecting pilots only to intervene in emergencies is a recipe for disaster. Humans are terrible at monitoring automatic systems, so it shouldn't be any surprise that they perform poorly when shocked into action. It's a systems design issue more than a piloting one.

That said, I think the level of deaths has plummeted with increased automation, so I'd be happier flying now than in the 70s.

eatmyshorts

More attention seeking from Learmouth, the self-appointed expert.

airuser

I agree with what Tenger had to say say about the different types of aircraft flow by pilots many years ago. Crews had to work up to flying larger aircraft. So by the time they were flying the largest aircraft they had plenty of experience.

Now, many flying crews are only flying one type of aircraft.

But is what has happened in the recent losses akin to "too much familiarity" with the aircraft and when some thing happens suddenly. Pilots who are trained to react they are overloaded with information from the cockpit instruments and may not react correctly.

I have been interested in aircraft and flying all my life.

Have been in Flight Simulators in both the U.S. and UK and see what faces pilots everyday.

The amount of paperwork before and during a flight. Flight Management Systems etc.

Many of us drive cars every day. However, when something goes wrong can we react quickly to avoid that accident. Taking care of the road conditions and traffic.

airuser

I should have added, was that not a wonderful landing of an aircraft at Gatwick recently VS 43.

MoeJay

Of course Learmount is only repeating what has been discovered, also by the UKCAA, that loss of control is now the biggest cause of fatal accidents in commercial aviation. Loss of control is a pretty wide definition but how else do you define it all?

Airline pilots, given the commercial nature and pressure of the business, are at the mercy and whim of managers who seek to mitigate risk to the operation. After all, headlines involving your aircraft are bad for business! So it is said to maximize use of automation to minimize the chance of error. What Learmount alludes to in his article is that the mandated recurrent check in the simulator is centered on a failure that in other circumstances is so unlikely to happen, the ETOPS or EDTO number just keeps going upwards...

So when something does go wrong, when was the last time, if ever, they saw that failure? The so called "startle" factor is something that we should all be aware of, yet seems to have been ignored until recently. The monitoring aspect of the operation is something that is almost taken for granted yet as has already been pointed out, humans are particularly bad at staring at things for a long period of time and maintaining concentration!

Is this the fault of the operating crew? The management? The regulator? People expect tickets for free yet this argument about training keeps coming back. Airlines only have to achieve the minimum to get all the ticks in the boxes, and then internally create rules to yet further mitigate risk e.g. Reduced crosswind limits for first officers, autolands in CAT I conditions etc etc.

So nothing is learnt from the Colgan accident, and while the passenger is comfortable with the seats for the crew operating their aircraft having basically gone to the highest bidder, nothing will be learnt. Pilots will keep asking for more training but it will take the highest authorities to force it, as it keeps coming back to cost....what cost???

And the managers will be comfortable with the risk they take.

Irish Steve

I could go in to a lot of detail here, but it will only provoke a fire storm of criticism.

Davis Learmount has access to information and people that are very much in the sharp end of the industry, and he is raising concerns about a very specific and serious problem that goes back more than a few years, and is not being well addressed by the industry.

There are issues across all aspects of the industry, and the race to the bottom being orchestrated by bean counters is the underlying issue, but it's not just about cost reductions, it's also about the manner in which skills are being diluted.

Many years ago, simulator sessions were managed by senior pilots, or retired pilots, who knew the aircraft well, and were able to provide instruction and scenarios that required the students to think about what was happening, and why.

Then, to reduce costs, the senior and experienced people were moved out of the loop, and increasing levels of "standard operating procedures" were introduced, and the simulator instructors were replaced by line pilots with much less experience of flying in general and of the type they were instructing on, and the combined result of these changes is that the experience levels are diminishing, and most of the time, that's not an issue, but, and this is a very big BUT, when things go wrong, the crew are left wondering how to resolve the problem, because they've never encountered anything like it before, and there is a very real likelihood that what's happening is not in the "standard operating procedures", because it's NOT standard, so the books, and the training, and all the other aspects of the carefully rehearsed and closely scripted training are suddenly blown out of the window by something that's way beyond the experience of the crew.

This scenario is then made worse by the issue that crew are no longer encouraged to learn more about their aircraft, the opposite is sometimes the case, an enthusiastic crew member seeking to use "spare" simulator slots to learn more about the extremes of the aircraft's capability is likely to be frowned on by senior managers, or even suspected of trying to hide a performance issue they are having with normal operations.

From my own experience of working with simulators and developing training scenarios, there are huge areas of the capabilities of the airframe that modern crews do not know how to use, and they never go near the limits of the envelope of the performance that the aircraft is capable of, partly because doing so could be uncomfortable for the passengers, and partly because they don't know what the airframe performance limits are.

Well, my attitude towards that is very simple, if I have a choice of doing something that is outside of "standard operating procedure", but is within the performance envelope of the airframe, and it will result in being able to recover the situation, or (for example) get the the airframe on the ground more rapidly, then discomfort to the passengers is a lot more preferable than losing those passengers, and there are times when they are the alternatives. In order to do that, the crew have to know what the limits are, and right now, the harsh reality is that large numbers of modern crew do not know how to safely use the limits of the envelope of the type they are flying.

A long time ago, 2 of us were using an A320 simulator to get performance information for some research work that was being done by BAE, (neither of us were 320 type rated) and as part of that, we did a manual reversion go around. We had to go to Denver to get into a simulator that was actually capable of doing that, most of the European simulators were not programmed with that level of capability, as it had been deemed as a scenario that was so unlikely, it was not worth training for. Having done it, and landed successfully from the second approach, we were surprised to discover that we'd just flown a procedure that the airline concerned had been told was not possible. The instructor was somewhat surprised that we'd managed to do what we had, and it provoked some very serious discussion in the instructor group at the airline.

There are very clear and complex issues that need to be addressed by the industry and the regulators, but it is going to be difficult, in that there will be massive problems to be faced if the training concepts are not agreed across the industry, which will require that the regulators have to be more actively involved in setting the standards, and monitoring how they are being implemented.

It could be the cynic in me, the fact that bean counters are never involved at the sharp end when things go wrong may have a lot to do with this, they are always a long way from the action, and have more than enough justifications for why they have done what they have, and perhaps that's the real problem, if the bean counters were more accountable for their actions, maybe some of the changes we've seen in the last 20 years would not have happened.

And, yes, that accountability might have resulted in less of a reduction in the cost of air travel. That might not have been a bad trade off, given some of the results that have been talked about in this thread.

airuser

Irish Steve,

Very much enjoyed reading you post.

You are not cynical. The bottom line has become too important in every industry.

We have seen that with our Banking Crisis over the last few years. Anglo were making huge profits. So the others followed suit and jumped on the band wagon.

Yes, we love the cheaper and cheaper flights, but at what cost.

Recently when travelling back from Spain watched the pilot after they landed, could see the cockpit. They got up to go to the toilet and then back to the cockpit.

Back to flying again.

OzCam

If it's any comfort to the pilots here, computer engineers have been saying the same thing for at least a decade, probably much longer... Automation saves money, but it cannot be built to handle all eventualities and humans are actually really bad at handling exceptional situations unless they have trained and practiced for such regularly.

The theoretically ideal way to drive an aircraft is to have the pilots doing most of the work (so they get the practice) while they're watched over by the computers (which are much better at that stuff - they don't get bored or tired). Unfortunately human nature - me included, if I was a pilot - doesn't like that idea at all, and autopilots save fuel and reduce costs so the bean counters are happy.


Anyone interested in cockpit automation issues will find lots of pointers to discussions on the Risks Digest, especially the contributions of Peter B Ladkin.

Furp

Great discussion on this thread, I'm keen on a caution mostly because I Fly a good bit for work and being an engineering geek. This reminded me if the excellent article in vanity fare recently about these very issues, it goes into great detail about the Air France flight that crashed in the Ocean. And the politics business behind how pilots are now trained.

http://www.vanityfair.com/business/2014/10/air-france-flight-447-crash

bronn

Great article from Vanity Fair, Furp and thanks for sharing. Interesting bit towards the end:

The best pilots discard the automation naturally when it becomes unhelpful, and again there appear to be some cultural traits involved. Simulator studies have shown that Irish pilots, for instance, will gleefully throw away their crutches, while Asian pilots will hang on tightly. It’s obvious that the Irish are right, but in the real world Sarter’s advice is hard to sell. The automation is simply too compelling.

You have to ask, is it nature or nurture with Irish pilots? Why the Irish in particular?

Tenger

bronn wrote: »

Great article from Vanity Fair, Furp and thanks for sharing. Interesting bit towards the end:

You have to ask, is it nature or nurture with Irish pilots? Why the Irish in particular?

mad bastards all of us!



Seriously though, I would disregard such a sweeping statement.......can you blindly state "that the Irish are right"..........but at the same time make one of my own.......the Irish have what I feel is a healthy disregard towards authority, not a disrespect but a nonchalance towards it.
Perhaps this is a cultural bias which promotes the lateral non-conformist thinking that allows people in problem solving situations to look outside the SOP's, whether they are flight crew, cabin crew or software developers.
At the same time of course such an attitude is still subject to the same pressure as elsewhere in the aviation industry, perhaps the Irish are better at CRM and cross cockpit comms? It is very hard to quantify until the SOP's go out the (cockpit) window.

*To reference the cabin crew name drop, (in a flight crew related thread) I have it on very good authority that when the international medical alert facility in PHX get a call from an EI crew about an issue onboard they know it is very serious, EI apparently have the lowest call factor of any longhaul airline who are signed up to it.
[FR do not use the facility for any who felt I slighted the other major Irish airline]

Bog Man 1

Can the computers get the plane out of a stall at high altitudes ? or does it require the intervention of the pilot .
With automation and the crew just spending time observing is there a brain process that when a light comes on to indicate a fault the brain tries to rectify the fault rather than fly the plane and does the present pilot training reinforce this .

smurfjed

Can the computers get the plane out of a stall at high altitudes ?

My limited understanding of the Airbus software is that it isn't supposed to allow the aircraft stall in the first place, however i doubt that it would be capable of recovering from something that was outside its programmed envelope.

On a Boeing, there is no software to protect you in the first place, they are equipped with stick shakers and pushers.

Irish Steve

Looking at the last few comments, I've been aware for some considerable time that my background is somewhat unusual, in that I spent 20 or so years working in the computer industry before getting involved in aviation, and a CPL/IR etc, and it was only the bad timing of Saddam that I didn't end up flying commercially, so I ended up becoming very involved with PC based simulation, some of it very complex and used by manufacturer (BAE) research facilities.

One of the spin offs of that experience was that I got to spend a lot of time on the jump seat of a lot of different aircraft and airlines on flights around Europe, and that was incredibly revealing in terms of the manner in which different airlines operated.

MCC and the pressure of "stanadard operating procedures" was only starting to rear it's head ( a long time pre 9/11), I was involved in designing one of the early MCC courses for Aer Lingus, their cadets in Vero Beach didn't do MCC as part of their training, so I'm very much aware of the sort of issues that it threw up, and we put together a very simple low cost Flight Sim based course that made it possible to do the necessary training without using the full flight simulators.

To make the specific point that comes out of this, and related to the comments earlier, because I was working a lot in Ireland, I spent more than a few hours with Aer Lingus crews, mainline and commuter (back in the S360 days), and I have commented on more than a few occasions to people that were interested that the Aer Lingus flight deck atmosphere was twice as relaxed as any other flight deck, but the level of professionalism was every bit as high and good as any other airline, and it made for a much better operating environment.

As part of some of the other work I did with Aer Lingus, I got to talking with some of the training captains about that very aspect, and it was something that they worked on, without working on it visibly, if that makes sense, and the result was very noticeable.

Some of the airlines of Europe were very formal, and rigid, and it made for a very different operating environment, and I know from discussions with other professionals that are involved with flight crew training that the problem is even more apparent and visible in the Far Eastern cultures, where even suggesting that the captain is not making the right decision can be seen as a huge offence. There are also issues where the airline has a large number of ex military captains, due to the much more rigid nature of military authority structures.

For sure, some of it is cultural, but some of it is very much down to the manner in which the senior people in the training regime determine their operations standards. Tenger's comments are also valid, maybe some of it is down to the cultural attitude of the Irish to authority, though I don't think it's that simple.

As for things like the capabilities of the computers, the problem with high level operations is that if there are technical problems, the computers may well not be operating, so the crew have to handle the aircraft. That can be very tricky, at high levels, the aircraft can be very sensitive to the controls, due to the speed and how thin the air is at those levels, and there is a very small difference between the mach buffet high speed limitation and the stall speed, and both have to be avoided.

To be blunt, it's even more difficult with the Airbus style side stick, as there's no real feedback to the crew of the sensitivity of the controls, and the total movement distance on a side stick is a lot less than on a control yoke, so being sensitive with a side stick is a lot harder than with a traditional yoke, and to make it harder, it's very rare that a crew has to hand fly at high levels, and if something has failed, which means the crew has no choice, it's possible that there are then other issues that make the handling even more difficult, the controls may have degraded in terms of their mode of operation, and that then removes some of the protections that would normally prevent things like stalls.

In some cases, based on research work I did related to degraded flight, I think it's safer to leave the side stick alone, and make small adjustments using trim, but that's not necessarily "official" Airbus strategy and operating procedure, it's just something that I've tried with success when researching extremes of the envelope handling when in different simulators, as part of the procedures to validate the handling we were building in to the research simulator for BAE.

I suspect that over the next couple of years, there will be wide ranging changes in the training regime to "encourage" crews to be more aware of high altitude handling issues, and I suspect that there will also be changes in the computer programming to make the aircraft better able to deal with some of the issues that have been revealed in recent years, and to make it easier for the crew to analyse what's gone wrong. In that respect, I was horrified to read just how long it took for a "heavy crew" to work through all the check lists and error messages that were thrown at them when the A380 uncontained engine failure happened, the crew were faced with a massive information overload, and the computers were not a lot of help, they reported all the issues, yes, but they didn't prioritise them in a way that reduced the crew workload, which in an emergency is essential, the crew need to know what's urgent in order to allow them to make informed decisions about how to deal with the emergency, the most important being, can we continue to fly safely, or do we need to get this thing on the ground NOW, or sooner, because of the severity of the issues. The computers on most aircraft right now are incredibly bad at assisting with that decision making tree, partly because their power is still very limited due to all manner of certification issues, and partly because the programmers don't have flying experience, and their priorities as programmers may not be the priorities that the flight crew would like to see to assist them deal with the issues.


Then there's the problem that modern aviation training for the airlines does not have the same degree of emphasis on "flying" that used to be the norm, and the result of that is crews that have not had to develop the raw data handling skills that were second nature to the crews of older aircraft, as that was all they had, so interpreting the information from the instruments, and a "scan" of all the instruments was second nature, but with the way that things are managed and displayed on the modern EFIS, even getting access to the raw data is not always possible, and if things have gone badly wrong, the raw data may be the only thing that will allow a timely recovery to normal flight.

Don't get me wrong, modern flying is inherently safe, the issues that we're seeing and talking about here are the very few and rare occasions when something out of the ordinary happens, and circumstances then combine to provide the Swiss Cheese scenario that results in too many holes lining up at the same time and resulting in an incident or accident that should have been preventable.

Part of the answer to that is (unfortunately) down to reversing the trend to remove as much cost as possible from the operation, when it comes to flight training, my personal view is that training needs to cover areas that the crew may NEVER see in their experience, in order to make sure that if they have the misfortune to become involved with a rare and unusual incident, they have the underlying knowledge and skills to be able to analyse the problem and come up with a resolution. That directly contradicts the bean counter and cost benefit analysis concept that has become all too prevalent in the industry, the fact that something is very rare, and unlikely does not mean that it should be discounted as being too expensive to train for.

Yes, I've ranged over a lot of areas in this message, but I'm pretty sure that there are a lot of people out there in the industry that have similar feelings to me, but saying so for them is dangerous, as it puts their position at risk. I'm in the fortunate position that I don't have to be afraid to speak out, as doing so doesn't threaten my position.

bronn

That was a really great post, Irish Steve. Very informative. Thanks!

smurfjed

Are you legally allowed to practice hand flying at high levels due to RVSM requirements ?

As for using trim to recover, we practice wake turbulence encounters at altitude, i have found myself banked 70 degrees, full aileron and rudder, nose down without the autopilot or yaw damper. There is no way on earth that my mindset would allow me to try to trim out of that situation, but there again i don't fly a FBW aircraft with a stick

There are reports attributed to some Asian airlines, especially those that allow P2F, saying that the first officer isn't even allowed to land the aircraft, so i don't see them allowing them to do high altitude hand flying.

Irish Steve

smurfjed wrote: »

Are you legally allowed to practice hand flying at high levels due to RVSM requirements ?

As for using trim to recover, we practice wake turbulence encounters at altitude, i have found myself banked 70 degrees, full aileron and rudder, nose down without the autopilot or yaw damper. There is no way on earth that my mindset would allow me to try to trim out of that situation, but there again i don't fly a FBW aircraft with a stick

There are reports attributed to some Asian airlines, especially those that allow P2F, saying that the first officer isn't even allowed to land the aircraft, so i don't see them allowing them to do high altitude hand flying.

Re RVSM, I don't know what the latest rules are, as I'm out of the direct loop now, but I suspect that even if it were legally allowed, the bean counters have put pressure on flight ops departments along the lines of Automation ON at (by the latest) FL180, and off again as late as possible, all of which dilutes the hands on stick abilities of the modern pilot, which is one of the reasons we are seeing more "incidents" during landing in difficult conditions, as the vast majority of pilots now have probably less than 20 minutes hands on time per sector, with most of that being on the ground, or close to it. If hand flying at high levels is not legal, then in my book, it should be a mandatory requirement on a regular basis in the simulator, to ensure that the crews have the ability to deal with flying by hand at high levels if the need arises.

AF447 came down as result of a relatively minor failure combined with a serious lack of analytical and flying skills on the part of the crew that were on the flight deck at the time. Yes, it was and is that simple, despite the operating airline being the flag carrier of the country that makes a significant percentage of the Airbus aircraft.

It should have been a massive wake up call for all involved, especially the software side of the industry, and it may have been, but it's been very quiet on that front as far as any public pronouncements are concerned, which doesn't altogether surprise me, given the competitive pressures from Boeing, (and others) and the very real potential for public disquiet if the real issues were being more widely discussed in the media.

Some of the issues are that the media don't have the level of information required to fully understand the issues, and they are notorious for going into a feeding frenzy when issues that could be problematic are exposed in the open.

The Far Eastern airlines have some very real culture issues, and their reluctance to allow First officers to perform landings even is well known, and not helpful to the overall concept of crew resource management and operation.

I'm not suggesting using trim to recover from a gross upset, that requires some very positive and aggressive control inputs, (albeit with care to avoid over stressing things or going outside of the envelope) but things like a failure of the pitot system (a la 447,) and the degradation of the FBW as a result means that the aircraft has lost things like Alpha floor protection, and stall protection, and possibly roll protection, and has to be handled with extreme care to avoid being taken out of the flight envelope.

If you then add to that a crew that's never seen degraded flight modes in real life or even in the simulator, there are some real issues staring them hard in the face, and if the protections are no longer operating, it is all too easy to end up departing from controlled flight, especially if you've not realised that the protections are off line.

I'm going to live dangerously here, based again on experience with a number of different airlines, and I'm saying this because I do NOT want anyone to think I'm aiming this at Aer Lingus, i most definitely am not doing so. Some of the issues with the lack of handling skills is because the instructors also have not had the experience of dealing with the more extreme issues, so can't even provide adequate guidance on possible recovery options, and in some cases, to make it worse, the simulators can't adequately simulate them either.

A number of Airbus simulators in Europe don't correctly model manual reversion, (we had to go to a sim in the States to do that research) and that's because the relevant risk assessments have been made, and the decision has been made that it's not worth training for manual reversion because the percentages are so high against it happening that the cost of training to that level is not justified, as the high level decision makers didn't believe that even if the crews were trained, they would be capable of achieving an acceptable outcome.

Remember that I mentioned research, this work was going WAY beyond the normal levels of involvement and training that is the normal airline flight ops scenario. That meant we did some serious digging into areas that are not normally gone into by instructors or programmers, because we had to make sure that what we were going to produce was not going to invalidate the research findings. We didn't like some of what we found, but we couldn't change it, because to have done so would have invalidated the research criteria. Catch 22.

I have to admit I have a problem with the attitude of discounting low risk issues where flying is concerned, and there are certain flag carriers in some countries that i choose not to fly with as a result.

The most insidious problem is ending up at extreme bank angles, and pulling full up stick when the aircraft has for whatever reason gone into a degraded mode, in that situation, the aircraft ends up with the speed decaying rapidly, and eventually stalled and at very steep bank angles, and the automation doesn't know how to recover from that if it's gone into one of the degraded modes as a result of another problem.

If the crew don't understand instinctively what the different degraded modes are, and respond accordingly, it's all too easy for them to make the problem worse, rather than recovering from a relatively benign situation.

While he may not have said it in the comment he made, I'm pretty sure that David Learmount had these sorts of issues in mind when he put pen to paper, and addressing the issues he's raised will have to happen, to ensure that we don't see an increase in the number of "incidents" across the industry.

And yes, this is more of an issue for Airbus than it is for Boeing, in that the Airbus was designed from the outset to be a very integrated aircraft, and there are things that are connected that make sense if you know about them, but if you don't, they can be serious gotchas, whereas the Boeing systems have evolved over time to become what they are today, so there is not the same degree of interdependence in the Boeing as there is with the Airbus, and the result is that a failure in one area on a Boeing does not have quite the same effect as it does on Airbus. That said, the 767 gear up landing in Poland was because no one realised that a particular breaker had tripped or not been reset, so the problem is not just related to Airbus.

Sorry it's so complex, the line pilots in the community will (hopefully) recognise that this area is one that does need to be addressed, and that the sooner it happens, the safer things will be for everyone.

FWVT

Great post Steve. Re. Airbus v Boeing, let's not forget it was a Boeing that the Koreans flew into the ground at San Fran.

Irish Steve

FWVT wrote: »

Great post Steve. Re. Airbus v Boeing, let's not forget it was a Boeing that the Koreans flew into the ground at San Fran.

Indeed, and it was a Boeing that crashed at Amsterdam because of a faulty Radalt,

Airframes have no respect for bad handling, or for crews are not aware of what's going on, and neither major manufacturer has bragging rights in this respect, both have their issues that result in accidents that should not have happened.

The most important issue is for the industry to stop listening so much to the bean counters where there are safety considerations being looked at, and to dig their heels in and tell them that cutting costs on safety is not an option. If that happens, then hopefully, we will see an end to this sort of incident.

MoeJay

Steve, you make a lot of sense but while people expect flights to cost less than the bus into town, things won't change a great deal...

National Aviation Authorities could take the lead, but they wouldn't want to make themselves uncompetitive for their biggest customers I.e. the airlines themselves.

I'm not sure what it would take to tackle the challenges you have described so well...

Irish Steve

I suspect that the pressure for change is going to have to come from ICAO, so that ALL the regulators have to respond in the same way and at the same time.

I am well aware that anything less than this is likely to have problems, and getting ICAO to take this on board is likely to be a significant fight, the one thing that will change this much faster will be the unthinkable, the loss of a large (747/380) size airframe as a direct result of the sort of issues that we've been discussing.

AF447 should have provoked some significant responses, but because it took so long to find the black boxes and then work out what happened, a lot of the momentum was lost. If I am to be brutal, the French Regulators also have a reputation for blaming crews for problems even when there may have been "issues" with the airframe that were a major contributor to the incident, especially where Airbus are concerned.

Changing the rules to require different training won't make the regulators uncompetitive, but the costs for the operators of more simulator time, or more comprehensive regular training could well be an issue for the airlines, unless they are all required to meet the same standard worldwide, regardless of the state of registration.

I even wonder if some of the training required to get a type rating needs to be looked at, to make sure that all members of the crew have enough in depth understanding of the aircraft and it's systems, so that they are prepared for being able to deal with the worst case scenarios that can happen. If that's appropriate, it will put some severe pressures on a number of areas, like crew resourcing, training, simulators, and training organisations, so it won't be a quick or an easy change.

I'd like to see the software on the aircraft being changed to make it much more interpretive of the issues, so that the crew get more help to analyse and resolve the problems, the hardware is powerful enough now to make this possible, or to be blunt, it should be, but it's amazing and slightly scary how many systems on aircraft are still using 386 processors, because the later processors have not passed some of the specialist validations that are required before they can be used in aircraft applications.

What's very clear is that this issue is not going to go away any time soon, and there are some very complex issues in this area that are going to make for some challenging times for the regulatory authorities.

cml387

Excellent post Irishsteve, the kind of thing that used to be in PPRUNE before the lawyers stepped in (not referring to your posts btw just an observation about PPRUNE).

One thing I would ask is, can a simulator be totally accurate when it's asked to go outside the normal flight envelope?

Irish Steve

cml387 wrote: »

Excellent post Irishsteve, the kind of thing that used to be in PPRUNE before the lawyers stepped in (not referring to your posts btw just an observation about PPRUNE).

One thing I would ask is, can a simulator be totally accurate when it's asked to go outside the normal flight envelope?

Know exactly where you are coming from on PPRUNE, I've not posted there for a long time, they became a bit too "precious" about the qualifications required to be able to make a valid contribution there, especially where commercial pilots were concerned.

Simulator flight models are an interesting subject, I spent a lot of time working very closely with a developer of a PC based flight model a number of years ago, and it was both interesting and challenging, and if 9/11 hadn't caused all manner of problems for simulator makers, we'd have had a very powerful and successful product available, but that's history now.

They vary, significantly, depending on the structure of the software that's driving the flight model, which is the core of the simulation. If they are using fluid dynamic modelling, then there is a pretty high level of fidelity, if they are using a table based model, where the values of fields are interpolated from a range of look up tables, then there can be significant issues with the way the simulator responds in comparison to the full size aircraft.

The newer the aircraft, the more likely the simulator will be accurate, and in some situations, the aircraft computers may be "tweaked" slightly to make sure that the aircraft responds in the same way as the simulator, and with most new types the simulator is built first and the crew that will fly the new prototype spend many hours in the simulator before ever going near the full size aircraft. The vast majority of the electronic equipment used on the flight deck of a simulator is identical to that found in the aircraft, and in dire emergency, it's been known to take a component from a simulator to get an aircraft back into the air, though it's rare for that to happen.

That said, it all depends on the simulator manufacturer, 2 identical simulators to look at may operate completely differently, as we found out some time ago when exploring the extremes of the A320 envelope for a research device for BAE.

Some of the A320 simulators in Europe could not model the extremes of the envelope because the airline that owned them had decided that it was not necessary to train crews to fly manual reversion or in the extremes of the envelope, as their assessment of that scenario was that the cost of upgrading the simulator package to cope with the scenario, and the additional training would not be worth it as they deemed the risk of the event happening to be so small.

I am not sure that I would want to have to live with the realisation that a decision I made about training had resulted in the loss of a hull and the lives of the passengers, in my book, a crew should know about all the possible degradations of control systems that can occur, and how to fly the aircraft in that mode. To leave a crew with a severely degraded aircraft without the knowledge of how to handle it in that situation seems to me to be a dereliction of duty.