Automation Related Accidents - History Repeats itself

[Centaurus]

Flight International editorial 10-16 March 2009, revives earlier editorials about the need for pilots to improve their manual flying skills. This in the wake of the Turkish Airlines B737 accident at Amsterdam where the crew allowed the aircraft to stall on final.

One telling paragraph in the latest editorial says regulators and the airlines have a lot to answer for; the regulators for their tendency to cling to current training tradition; and the airlines for assuming the latest aeroplanes should deliver not just improved reliability, but lower crew training costs. It goes on to say that increased systems complexity, combined with a lack of pilot line flying practice at manual approaches, demands a different training approach.

The most cost-effective solution to poor hand flying ability is blindingly obvious. If the Flight International editorial is accepted as valid criticism - and the accident record of some airlines indicates it is - then the emphasis on recurrent simulator and type rating training programmes, must lean more toward hand flying exercises than the current accent on full use of automation.

Ok - some recent accidents can be partially sheeted home to enthnic culture mores. But even the world's safest most reputable airlines need to heed the warning inherent in the latest Flight International. Every airline has it's share of lazy pilots content to sit back and rely on the automatics to dominate. In in many cases, it is the operator via the FCOM that actively discourages even the occasional manual flying - apart from the last few seconds from the outer marker.

In the simulator, LOFT in real time with its accent on automation, CRM and crew "management" of the various scenarios, is certainly useful to a degree. But it also has the disadvantage of using up simulator time best directed to more urgent priorities. And that is keeping up a pilot's basic flying ability. The Flight International editorial should be a wake up call to regulators and airlines. One only has to study most automation related accidents to confirm history often repeats itself.

[411A]

Yup, couldn't agree more.
We don't do LOFT to any great degree, the time is used for manual flying and abnormals. And, as I have the say so, it will not change.
The only time the autopilots are engaged is for CAT II.

[733driver]

Wouldn't it be nice if everything was either black or white. It is obvious that that's how 411A likes to see the world.

I think there are a couple of issues. And flying every sim session manually is not the solution.

I will admit that I have frequently turned off the A/P and F/D during OEI approaches in the sim to make things a bit more challenging for me even when not called for in the exercise. I would always mention that I am doing it for proficency purposes only. And if it went too well the instructor would have me fly a NPA instead and/or increase the CWC or 'fail' the F/O.

However, there is also the concept of 'train as you fly and fly as you train'. That to me means that we shoud really use the A/P after an engine or system failure if there isn't a good reason not to do so. It is expected from us in real life and should be no different in the sim for obvious reasons.

I do agree manual flying must be practiced both in normal line ops and in the sim but to not train for appropriate use of automation in the sim and in real life is not the solution to the problem in my opinion.

[AerocatS2A]

Use of automation is as much of a skill as hand flying so we need to be careful not to go too far in the wrong direction. Of course we normally get plenty of practice using the automation in normal line flying but it is necessary to use it intelligently in abnormal ops which should mean that you use it appropriately in the sim (train as you fly as mentioned above.)

Ultimately it is necessary to have balanced training that gives adequate focus on all aspects of flying. The challenge will always be that there is never enough simulator time booked and so compromises must be made and it's the role of the training department and the individual check/trainers to get the best value out of the time available.

In our operation a certain balance is forced upon us as the Dash 8 autopilot is not certified for use in OEI approaches. Also we don't use a cyclic training system and so the test follows CASA's instrument renewal test form which requires a hand-flown, OEI, raw data approach. So we all typically get to fly a hand-flown, raw data, OEI, NDB with a circling approach to land every six months. In addition we are encouraged to use the automation at anytime it is available.

[ALK A343]

I have to agree with 733driver. Today's aviation is not simply black or white but lots of different shades of gray. The important thing is striking a healthy balance between retaining manual flying skills and the proper use of automation.
I guess most of us would agree that trying to do a raw data ILS without flight directors and autothrust into a busy place like LHR after a 12 hour flight is certainly not very smart. There are way to many things going on at that time, e.g. 160 knots until 4DME and late landing clearance etc..
On the other hand pilots should be encouraged to switch everything OFF and fly a manual ILS without autothrust and flight directors or a visual approach as and when weather conditions, traffic and crew fatigue permits.
More than anything common sense should prevail. This is when company corporate culture plays a very important role in stressing the need to practice manual flying when appropriate, but also teaching the proper use and monitoring of automation when required.
I think that most the time something goes wrong with the automatics it is our lack of understanding and monitoring which gets us into trouble and we sometimes tend to get carried away along with the automation instead of taking over manual control when necessary.
With regard to simulator sessions it should be stressed that you should treat it as a real flight and I would certainly make use of all resources including automation when it comes to an emergency. Yet again even in the simulator you have to strike a healthy balance and manual flying should be encouraged, e.g. when you do a visual approach with an engine on fire. The main thing with automation is to use an appropriate level depending on the circumstances.

[scambuster]

There is a very good video produced by American Airlines called "children of the magenta". Over reliance on automation is the subject. An address by the chief pilot to maybe 300 pilots and staff on the subject highlights the dangers inherent in todays culture. He says....."its not their fault, we made them like that, we trained them that way.....it has to change!" ( not quite an exact quote but that's the gist.

[A37575]

What well written, balanced replies.

Pilots spend probably 95 percent of all actual line flights on automatic pilot. What is the point therefore of spending 95 percent of all simulator flying on automatic pilot?

Of course no one in his right mind is going to throw away the automatics in seriously bad weather just to practice hand flying a real time Cat 3 ILS. Practice circuits and landings in the simulator without automatics is an excellent way of keeping up basic flying skills. The fact that this type of flying does not occur in line flying is missing the point. It is one way of improving basic handling skills. With a bit of imagination by the instructor an hour in the simulator expressly for keeping up manual flying skills might just have saved the day in the Turkish Airlines accident.

[TopSwiss 737]

Does anybody have a link to this "children of the magenta" video?

Sounds interesting... :ok:

Tnx, TS

[Old Smokey]

When I flew DC3s, I flew them as the manufacturer intended. They were meant to be hand flown, with a very primitive Auto Pilot to relieve the pilot's work load during cruise, and not much else.

When I flew later generation aircraft in the 70's and 80's, I flew them as the manufacturer intended. They had quite good "automatics", but a low state of redundancy. Thus, I remained proficient in both the use of the automatics, and hand flying.

Now that I fly modern generation aircraft, I fly them as the manufacturer intended. They have extremely competant "automatics", with a level of redundancy that is astronomical, with the chances of significant failure on a level with the chances of an in-flight collision with a meteorite.

Of course I can still fly the aircraft competantly by hand, it's fun! The manufacturer didn't intend that I have fun, he intended that I become more of an aircraft MANAGER than a hand flyer, and that's what I do most of the time.

It's despicable of me to make comment upon the recent Amsterdam tragedy whilst it is still being investigated, but is hand flying the issue? I would think, on the information SO FAR available, that improved MANAGEMENT and MONITORING might have saved the day. I'm sure that the crew could have hand flown the aircraft for the entire flight if they'd had to.

Now, does anyone have statistics of Hand Flown Accidents Vs Automation accidents available?:confused:

Regards,

Old Smokey

[A37575]

Use of automation is as much of a skill as hand flying

I disagree. Automation requires very basic computer skills and apart from exercising the forefinger nothing much else. Hand flying requires not only situational awareness skills but hand-eye coordination. Watch a pilot in a simulator who is a whip on the FMC but out of touch with hand flying raw data and you will see the difference.

[FullWings]

Use of automation is as much of a skill as hand flying
Well, I'll have to agree with that. To be good at flying on "automatics" you need to understand what they can and can't do, how to set them up to do what you want, how to recognise that things are not quite right and what to do about it when this happens. That's a comparable skillset to that needed for manual flying - slightly more cerebral, maybe but still comparable.

When you do a type conversion course, you are expected to become proficient in the operation of the aircraft and that generally includes all of the autoflight features. Many SOPs/regulations require the use of automation for certain phases of flight, e.g. low RNP approaches, RVSM, lo-viz ops. To lack the knowledge or ability to fully utilise these features is not something to be proud of, IMHO.

I'm very much with Old Smokey on "MANAGEMENT and MONITORING", both in manual and automatic flight. In an emergency it doesn't have to be pretty - just safe. In modern aircraft (FBW Airbus & B777 onwards) there is always some degree of automation engaged, unless you go around pulling CBs and flipping guarded switches, and a good working knowledge of the limitations/modes of these systems would seem to be helpful.

When looking at some of the recent airline incidents/accidents it is tempting to come to the conclusion that a lack of hand flying practice/competency contributed to the outcomes. However, there is also a common thread of mode confusion/perplexity with automatics/lack of system knowledge leading up to where it went wrong...

[A37575]

What we occasionally see in the simulator is the marked aversion by some pilots to promptly revert to basic manual flying (depending on the aircraft type) in order to initially recover a serious flight path discrepancy. In these cases it is common to observe pilots going instant heads down to the CDU, including frantic button pushing accompanied by changing of various pages when the aircraft is apparently flying headless towards rising terrain.

This aversion to reducing the stages of automation to manual flying is often caused by lack of confidence in one's own ability to actually fly the aeroplane. In turn, one can speculate that perhaps simulator training has accented automation to such a degree that the pilot is "brainwashed" into thinking that manual flight is a very last resort equating only to a Mayday situation. I jest a little here of course.

While some operators may sneer at those pilots who wish to keep current with manual flight, these operators should take time out to note the Boeing advice in the 737 FCTM under the heading "Automatic Flight." The extract reads thus:

"Early intervention prevents unsatisfactory airplane performance or a degraded flight path. Reducing the level of automation as far as manual flight may be necessary to ensure proper control of the airplane is maintained. The pilot should attempt to restore higher levels of automation only after airplane control is assured."

The Flash Air B737 crew rolling inverted at night with the captain shouting to his first officer to engage the autopilot, is a case in point. None survived.

[4dogs]

A37575,

IMHO, AerocatS2A was quite correct in his statement that "use of automation is as much of a skill as hand flying" for many reasons. I found your response quite shallow and dismissive, despite the fact that you probably would be in heated agreement about the problem if you were open to the debate.

I think that the current levels of automation require considerable skill to understand and to effectively utilise the nuances that software engineers keep throwing into the mix in their belief that it will make our job easier. Most importantly, just like many hand flying decisions, it takes skill to know when to give up what you are doing or trying to get the automatics to do and adopt a simpler but safer course of action.

I think it is generally true to say that you need to know how to do something manually before you can monitor the automatics doing it for you. But knowing how you used to do it does not mean that you can still do it if required - because that, as has ever been the case, requires practice! In my team, every recurrent sim requires a successful demonstration of manual flying as part of the training sequences. We are also looking at a monthly self-monitored manual approach and landing in suitable conditions in order to embed the philosophy and skills required if manual flight is unexpectedly required.

Bit hard sometimes when you are not the masters of your own destiny and have the boffins in QF safety recommending that we never train with any automatics off - that very French Airbus engineer philosophy of "minimise pilot intervention in flying the aircraft" is obviously very catching, particularly with those folks who don't actually fly!

Stay Alive....

[Iceman49]

Some of the problem could be mitigated in intial training on the aircraft, there is so much pressure on the cost of training, that the ability to hand fly the aircraft is minimized... it would be better to start off hand flying, than teach how to add and subtract the automation vs having only one period devoted to hand flying.

[extreme P]

Does anybody have a link to this "children of the magenta" video?

Sounds interesting...

It's called Automation Dependency and the speaker is Warren Vanderberg. "Children of the magenta" is a quote from the video and it refers to pilots following that magenta line at all cost and not flying the airplane. There are three levels of automation and he stresses that it is perfectly acceptable to use the lowest level (ie looking out the window and hand flying) when required. About thirty minutes long and well worth watching if you can track it down.