Company Lip service to Go-around flying skills

[A37575]

Thomson Airways has modified its Boeing 757 pilot-training programme to include a refresher on all-engine go-around following a missed approach incident at Bristol.

The aircraft (G-OOBE) entered an increasingly steep climb - pitching almost 31° nose-up at one point - during a go-around in gusting winds, during which its airspeed bled away from around 160kt to a minimum of 110kt.

UK investigators state that the 757's flight directors had commanded a high nose-up attitude as a result of the speed at which the go-around was initiated.

The aircraft entered a high rate of climb, says the Air Accidents Investigation Branch, with an "ever increasing" nose-up attitude owing to the combination of wind speeds, turbulence and full go-around thrust. This pitch attitude exceeded that indicated by the flight director.

As the 757 closed on the missed approach altitude of 3,000ft its automatic flight-control system reduced thrust as it tried to level the aircraft.

Investigators found that, although the captain had noticed the declining airspeed and encouraged the first officer to lower the nose, this was ineffective because the captain was unaware that the first officer had engaged the autopilot.

Despite the airspeed bleed and high attitude, the aircraft's stick-shaker did not activate.
Investigators determined that the crew had been "startled" and suffered degraded situational awareness during the high workload of the go-around. The probe into the incident, on 1 February 2016, adds that the winds had been "challenging" and that the first officer, at the time, had been undertaking his first line-training flight after a near half-year absence on medical grounds.
............................................................ ...................................

Chasing FD needle indications in what should have been a basic go-around procedure. And why the blazing hurry to plug in the autopilot in the middle of a go-around?

[Widger]

Thanks for that. Nice if you could spell Ascent properly. Shoddy journalism!

Or were you writing about his plummy pommy voice on the R/T during go-around?

[DaveReidUK]

Or perhaps the OP was just reflecting on the increasingly blind accent on automation nowadays ...

Like wot it said, in fact.

[Widger]

Ah ..... I see. Well like many laws...it was open to interpretation. Pun intended!:ouch:

[peekay4]

I'm going to play devil's advocate here. This might not be about automation but about crew mental overload and loss of situational awareness due to an unexpected event ("startle").

I'm reminded of another 757 incident a few years back, in Olso, where the crew also completely botched what should've been a routine G/A procedure after an unstable approach.

The Captain (PF), flying manually with ATHR engaged, started to apply huge control column inputs. The 757 pitched to 21 degrees nose up, before the Captain reversed input and plunged the aircraft to an extreme -49 degrees nose down steep dive at low altitude.

With the GWPS blaring both pilots yanked their control columns back and they somehow recovered at just 321 ft radio altitude, 251 kts, after a sustained 3.59G pull. The 757 zoomed back up at 40 degrees nose up until they finally leveled off at 4,000 ft.

This routine G/A was botched in manual flight, and in fact in this case automation could have helped reduce the workload in a period of mental overload. In the Thomson Airways case it's also possible that the FO turned on the A/P in the middle of the go around because he was similarly overloaded and wanted help from the A/P.

So I think it's too easy to blame "automation". The problem to solve may be a bit more fundamental than that.

[+TSRA]

I agree with peekay4 on this one. Blaming automation has almost become the pilots' scape goat. We push crew to use a balance of manual skills and automation, but when anything goes wrong we immediately say they should have used the other method.

I think this gets to the heart of it:

The probe into the incident, on 1 February 2016, adds that the winds had been "challenging" and that the first officer, at the time, had been undertaking his first line-training flight after a near half-year absence on medical grounds.

Perhaps rather than the automation being at fault, it was the decision to allow the FO to attempt a landing in such conditions after so much time away that should be questioned. Not saying he couldn't handle it, but you don't throw "new guys" to the wolves on their first flight or, in this case, first flight back. Line Indoc is meant to train for the line, not to scare the hell out of you.

[A37575]

Second thread first post

https://assets.publishing.service.gov.uk/media/57ff8e8d40f0b67135000006/Boeing_757-28A_G-OOBE_11-16.pdf

This incident report on a botched go-around in a Boeing 757 reveals somewhat feverish activity in the cockpit for a seemingly straight forward visual go-around. The haste at which the PF tried to engage the automatic pilot a few seconds after initiating the go-around, suggests either he lacked faith in his own ability to fly a manual go-around; or the company policy was to engage the automatics as quickly as possible after a go-around and he merely followed a company recommended procedure. That procedure is flawed if this incident is any example.

Either way, we are seeing yet another situation where the accent on rapid engagement of the autopilot and associated automatics such as the flight director, over-rides common sense and good airmanship. If this is what is taught in the simulator during type rating and recurrent training, then is it any wonder incidents such this will continue to happen.

Time and again, automation dependency has proved a threat to the safe operation of a jet transport aircraft. A manually flown raw data go-around, whether visual or IMC, should offer no problem to a well trained competent pilot. The trouble is there are not many of these people around, anymore. That being so, it could be argued the blame lies squarely with training departments who consistently disregard the evidence that degradation of manual flying skills needs to be addressed by more than just lip service and a quickie manually flown visual ILS.

[DaveReidUK]

The haste at which the PF tried to engage the automatic pilot a few seconds after initiating the go-around, suggests either he lacked faith in his own ability to fly a manual go-around; or the company policy was to engage the automatics as quickly as possible after a go-around and he merely followed a company recommended procedure. That procedure is flawed if this incident is any example.

The report suggest the latter:

"The PF then engaged the autopilot, with the aircraft climbing and in trim, using the company mnemonic 'TAGL'"

(TAGL = Trim, Autopilot, Go-Around, and Lateral Mode)

[RAT 5]

There is also some confusion whether the F/O was PF on both sectors or not. Las Palmas is generally a more benign airport then Bristol in the winter; confirmed by the weather. If the F/O was on a series of refresher line training there is no problem maximising PF duties, If his only sector at the end of a long day was into BRS, that could be challenging.
There was no comment about why the a/c ended up requiring a G/A, other than it drifted off and became high.

However, to understand better root causes, it would require answers to A37575's questions about company philosophy and use of automatics; otherwise we can surmise, going round in circles, from an incorrect starting point. This topic, about automatic dependency, has been debated from all points of the compass and from every angle on many occasions here.

[Capn Bloggs]

Whoever designed that thing needs their head read. How on earth can an aeroplane, with AP engaged, be doing 6000fpm when trying to level off at 3000ft during a GA?? And what was the ATS doing?? 110kts?? I'll bet that wasn't the selected speed or Vapp.

No wonder the crew "lost the plot".

Dark and bouncy. Dodgy for an FO who hadn't flown for 5 months?

Unless I'm missing something...

[cosmo kramer]

Brits and surrounding Islands loves "mnemonics", whether it's for doing a briefing or just flying the aircraft.

Maybe it's time to revert to skills and common sense? This reduce-the-pilot-to-ape-like approach doesn't seem to work all that well.

[Jwscud]

Altitude capture modes can be very poorly programmed at high rates of climb, and often there is no way out of them apart from manual intervention.

The programming of ALT* was the root cause of the A330 flight test accident.

[sleeve of wizard]

ALT CAP submode can be a dangerous animal.
With a high rate of climb going into ALT CAP the AFDS then follows a parabolic curve to capture the desired altitude and MAY disregard the speed set on the MCP.
If memory serves correctly on the 757 the first press of TOGA commands 2000fpm climb and with another it gives full TOGA thrust, hence the possibility of 6000fpm :ok:

[Groundloop]

Whoever designed that thing needs their head read. How on earth can an aeroplane, with AP engaged, be doing 6000fpm when trying to level off at 3000ft during a GA??

I once chatted to an ex-BA 757 skipper about this and his comment was "How on earth was it certified?"

[Peter G-W]

Wasn't the root cause of the A330 flight test accident the decision to shut an engine down during ALT* manoeuvre? Hardly a fault of the flight mode.

[vilas]

At least in airbus that is not the procedure for putting AP on. First select FDs (if Off), get them in correct flight mode (OP CL/DES/ALT)and centre them manually before putting the AP. AP should not be asked to make corrections to flight path.

[Lantirn]

Vilas I think in the boeings the AP wouldn't engage unless the aircraft is trimmed.

But in a go around (320), considering a raw data approach, FDs and ATHR would engage automatically, leaving only the action of autopilot engagement if required

[vilas]

The procedure I stated is applicable in general. In a GA although the FDs would appear automatically still you need to centre it manually before engaging the AP.ATHR will be armed (blue).

[john_tullamarine]

Two threads merged by OP request. A bit too messy to attempt to tidy up the post sequences .. hopefully it will not be too difficult to follow the discussion prior to this post. Post relating to the start of the second thread is noted.

[vapilot2004]

Investigators found that, although the captain had noticed the declining airspeed and encouraged the first officer to lower the nose, this was ineffective because the captain was unaware that the first officer had engaged the autopilot.


The captain had a column between his legs that clearly indicates his FO's pitch commands and all that FO needed to do was put the nose where he wanted it, AP engaged or not.

I understand how the automation failed to do its job properly, but fail to see why the crew were unable to take control of the aircraft using basic flying skills and if need be, a little muscle.

[RAT 5]

I understand how the automation failed to do its job properly, but fail to see why the crew were unable to take control of the aircraft using basic flying skills and if need be, a little muscle.

That question has been asked & debated about on here for numerous scenarios; the question is oft repeated, but a definitive answer is still awaited, or rather implemented policy. The answer may be well understood, but not applied by operators. However, with inside and historical knowledge of Thomson, they are one of the more 'manual flight encouraged' airlines.

[No Fly Zone]

If the reports are accurate, it sounds like the crew (entire company?) needs to revisit the subject of COMMUNICATION and COORDINATION in a multi-crew environment. IMO, there is NO EXCUSE for one pilot - even the commander - doing something without informing the other pilot. Got to TALK boys and girls. TALK to each other! That is what Multi-Crew is all about!

[Amadis of Gaul]

Reminds me of one of my very first sim instructors. He used to say "talk to me, son, you gotta TALK to me!"

[Dan Winterland]

Full power low weight GAs are a dangerous manoeuvre. We don't practice them often and they often go wrong. LOSA data showed that one in 10 result in an undesired aircraft state, but I think the rate is probably close to 1 in 5. peekay4 mentions the B757 incident at Oslo which was so very nearly a crash. The crew were subject to the somatogravic illusion which is human limitation. We have not evolved to assess prolonged acceleration and can easily misinterpret it as pitch up. In the high energy GA, the AP is your friend.

In the Thompson case, it was sensible to have the AP engaged. The question that needs to be asked is why the AT didn't command the power required. Intervention from the PF would have resolved the issue, but in the high workload situation of the GA, it would have been best if it had responded as it should.

[Centaurus]

The crew were subject to the somatogravic illusion which is human limitation. We have not evolved to assess prolonged acceleration and can easily misinterpret it as pitch up

Somatogravic illusion is often cited by investigators as a contributory cause of many accidents of this nature. That has applied to Cessna 150 go-around accidents as well as jet transports.

If that were true, then one must ask why aircraft carrier pilots who are catapulted at night or in IMC or conduct go-arounds or "bolters," are not crashing in their thousands. The acceleration of after-burner take offs are ten times more powerful than a go-around or take off in an airliner.

No. I believe the problem is simply poor instrument flying skill which needs to be ironed out during simulator training. We see this often during simulator training. Crashes during go-arounds occur in simulators due generally to poor instrument flying skills and inappropriate operation of the automatics. These are not excused as the result of somatogravic illusions. Yet in real aircraft accidents poor instrument flying ability is often covered up in favour of the more politically correct and mysterious sounding `Somatogravic Illusion.`

While there is medical evidence this illusion is a human limitation, it is practically impossible to prove after a crash that it was the primary cause of the accident. That is why the illusion is such a convenient scapegoat in accident reports because it is impossible to prove with any certainty.

[PEI_3721]

Dan, 'dangerous', a bit strong; don't want to frighten the poor dears. Challenging, yes, for the reasons outlined.
Challenges are there to be met, anticipated, and managed; beware of surprise - "a landing is an approach without a GA", plan for workload, and simplify procedures (including Nav and ATC - see ASAGA report).
Centaurus, yes illusions are a threat, but more often an illusion can be managed having been previously experienced - basic training?

The 757 and similar aircraft have strong nose up pitching moments with GA power, thus reduce power and manage the pitch control. Unfortunately many pilots have been influenced or even trained by the mantra to fly the aircraft in trim; in the extreme to fly the aircraft with trim, and thus are reluctant to use input large control deflections or experience high control forces.
A significant issue in this event, and others, is expecting the autopilot to take over with the aircraft out of trim. Autopilots have no brain - no situation awareness, limited anticipatory ability, and restricted application of control forced and trim rate (safety features).
If a pilot handed over the aircraft to me out of trim, there was a severe cuffing around the ears; perhaps autopilots should have a similar feature, or at least pilots be aware of the limitations of handing over control to a dumb system.

http://i.ebayimg.com/images/g/SK8AAOxygLxSZVmf/s-l300.jpg

[RAT 5]

[I]thus reduce power and manage the pitch control.

Indeed; as I've often emphasised to students, "power can be your friend, or very unfriendly."

Some instructors were taught to teach a G/A is full power at all times. Ouch. Thrust is a variable when needed, and over thrust during G/A is a real gotcha.

[WindSheer]

Gonna stick my neck out here, but I feel incidents such as this would be avoided if the SOP was simply "fly the a/c up to the G/A altitude".
Without the confusion of following infrequently practised procedures, it would be a walk in the park to apply a moderate amount of thrust, level the a/c, tweak the thrust as appropriate and simply hand fly it back up, calling for gear and flap as appropriate.

I guarantee any competent pilot would make a much cleaner job of it rather than trying to obey flight directors and rigorous memory items with a massively overpowered light a/c like the 757.
In the world we live in, this would never happen though...........

[peekay4]

If that were true, then one must ask why aircraft carrier pilots who are catapulted at night or in IMC or conduct go-arounds or "bolters," are not crashing in their thousands. The acceleration of after-burner take offs are ten times more powerful than a go-around or take off in an airliner.

Happens more often than one might think:


F/A-18C HORNET NIGHT CATAPULT LAUNCH
The Mishap Aircraft (MA) crashed into the water after night catapult launch. The Mishap Pilot (MP) was
well rested and mentally prepared for the Mishap Flight (MF). MP spent significant time troubleshooting
several discrepancies while on deck, all of which were satisfactorily resolved prior to MA launch. Weather
conditions were overcast at 600-1000 ft, creating an extremely dark night under the low overcast. MP
conducted a normal catapult shot with sufficient airspeed for flyaway. Almost immediately after launch, MP
grabbed the stick and easily countered a slight roll to the right due to MA asymmetric condition. MP
gradually applied forward stick during the climb out. After peaking in altitude at 224AGL, the MA
responded to the forward stick by accelerating and following a nose down flight path toward the water. Just
prior to water impact, MP realized he was in extremis and attempted to eject, but was already out of the
ejection envelope resulting in an unsuccessful attempt. MP lost at sea.

Official Cause Factor:
AIRCREW: MP applied improper forward stick inputs during climb out due to the effects of somatogravic
illusion.

...

The F/A-18 mishap presented in this report is not a rare type of mishap. For the past several years, NAMRL has
assisted on at least one case per year of somatogravic illusion in the "fast mover" communities (three such mishaps
in FY2001). Somatogravic illusion mishaps are not always associated with catapult launches, but may also occur in
high performance takeoffs, landings and bombing runs over land.


Somatogravic illusions and other spatial disorientation mishaps are a leading killer of military pilots worldwide.

Source: www.dtic.mil/cgi-bin/GetTRDoc?AD=ADP013854

[Centaurus]

Peekay 4. Thanks for including the US Navy accident data. Certainly food for thought.
Cent.

[PEI_3721]

Simplicity is a good objective, but there are limits.
An SOP to 'fly the aircraft' overlooks the need to adapt the 'standard' to the situation; do pilots have knowledge of the 'when and how to adapt'. These aspects can be primed with a good briefing by considering the differences between the expected situation and that trained for, particularly if training amalgamates GA with engine failure.

Many SOPs are written for the exceptional circumstance; GA, + engine fail, + max wt, + terrain limited. How many accidents have involved some or all of these aspects in combination, vs incidents during a normal GA. Many SOPs are not 'standard', they are abnormal procedures; a GA is just part of normal operation.
SOPs have to be written for the normal, we must train for the normal; brief for the abnormal. The industry needs to develop skills of awareness, planning ahead, knowing when and how to adapt, then review and learn from debriefing.
However there are significant obstacles to this in the training mantra, to always follow SOPs, which may adversely influence and restrict pilots' flexibility and adaptability in those situations when most needed.

Procedure writers need to review SOPs for ambiguity, unpublished assumptions, or specific situations where 'standard' does not apply.
Provide explanation and context; as indicated in #29 'moderate', 'tweak', 'as appropriate'; these also require knowledge of how and when.

... the best laid plans ..... frequently have to adapt to reality.
Things are 'simple' before thinking, but simplicity is only enabled after thought.

[PEI_3721]

PK 4, beware the 'Americanism' of using 'a leading killer', - one of many causes, in the context of 'the leading cause', i.e. top of the list. The report does not support the view that this illusion(SD) is top of the list. The discussion relates to a model, computer simulation, as an aid to investigation. Take care not to over rate the problem of SD in commercial aviation.
However, it does warrant attention, but it may be more beneficial to focus on real aircraft demonstration vs forum discussion and false claims of training simulators.

[RAT 5]

Many SOPs are written for the exceptional circumstance;
SOPs have to be written for the normal, we must train for the normal; brief for the abnormal.

Is there not a slight contradiction there? SOP's can be written for the norm, and can also be written for the simple non-normal and some expected emergencies. These are trained and checked ad infinitum, so help me god. The accidents happen when events happen outside these scenarios and pilots are left grasping at straws in the dark.
It could be that you can become 'over-SOP'd' & 'under-airmanshiped & under-trained & under-educated' in your a/c.
IMHO pilots should be trained in all the aspects of how their a/c was designed and what it can do, and its systems. There will be basic simple SOP's for the basic simple manoeuvres. Outside those the crews should have full confidence and discretion to use whatever weapons the a/c & its systems offer them at their disposable to achieve the task in hand. Not all G/A's are the same. The crew should be able to act as pilots and not trained monkeys where there is a 1 size fits all SOP. That is what I find too much in todays TR courses. Too many SOP's not enough education. Too much who, what & when and not enough how & why and other options.

[WindSheer]

PEI......a very informed, well written response to my ideal world scenario.

[peekay4]

I wrote "a leading killer" for military pilots and I mean it. During Desert Storm the US Navy lost 15 aircraft in non-combat accidents. 7 of those 15 involved SD.

From the US Navy (http://www.med.navy.mil/sites/nmrc/sitepages/NewsStory.aspx?StoryID=32):

​Spatial Disorientation (SD) is an aviator’s misperception of the attitude, position, or motion of his/her aircraft relative to the Earth’s surface and gravitational vertical. SD is a serious threat to flight safety and is the leading aeromedical cause of Class A mishaps in Naval aviation. SD is also a leading killer across the DoD and in civilian aviation as well.

From the US Air Force (http://static.e-publishing.af.mil/production/1/af_a3/publication/afi11-2f-16v3/afi11-2f-16v3.pdf):

7.8. Spatial Disorientation (SD). SD has proven to be a leading killer of F-16 pilots. Although SD is most common at night or in IMC, it can and has happened in day VMC. Reference AFPAM11-417 for information on the causes of SD, how to avoid it, and how to mitigate its consequences.

[Centaurus]

the US Navy lost 15 aircraft in non-combat accidents. 7 of those 15 involved SD.

From the US Navy:


My problem is what facts were produced by the authors of the USN paper to confidently state with absolute surety that these crashes were the result of the pilot being affected by a somatogravic illusion during a take off or go-around? Spatial disorientation coves a whole gambit of accidents including those caused by disorientation at high altitude in IMC. The subject is specifically go-around accidents. Maybe some pilots are more medically prone to these type of illusions than others? A hang over from the night before or lack of a decent night's sleep?

Over many years as a current pilot I have seen countless accident reports where the investigators opined the most likely cause of similar accidents involving take off or a go-around (night/IMC)was SI . They make an educated guess it was an SI problem; but that is not factual as it cannot be proved beyond doubt.

[galaxy flyer]

Having known a number of carrier pilots, including my brother, everyone would say, if you aren't deeply SD'd on a bolter back in dark clouds, you're inhuman. It takes a lot of self-discipline and experience to handle it well. I've sworn the tanker was doing all kinds of weird maneuvers while on the boom in and out cloud at night.

[IcePack]

IMHO part of the problem is over pedantic SOP. For instance taking an A 330 out of xxxx with a 2000ft level off & a sop for power reduction of 1500ft. Me says to pf please feel free to select clb pwr once wheels up. Answer from pf NO that is not sop. Me says ok whatever you wish to do you are not to bust the level or the flap speed. Anyway atc changed height clearance on the take off roll. Would have. Power is part of the flying controls & sometimes total adherence to sop is dangerous but the fear of the QAR is causing lack of intelligent thought and the ability to actually control the aeroplane.

[A37575]

the fear of the QAR is causing lack of intelligent thought and the ability to actually control the aeroplane
So true - especially in the Middle East and some Asian operators

[PEI_3721]

... a slight contradiction there?
Yes, but who is expected to resolve the contradiction, and when. Most likely the crew at the time of an event, and in real time.
A well crafted SOP could reduce the mental workload; also appropriate training should enable greater thinking capacity in demanding situations.

Is it easier for pilots to relate to a change of action when moving from 'a norm' - the SOP for an 'everyday' GA, to that of judging the need for increased thrust in a more extreme situation; or
the alternative of having an SOP which covers all situations, and then having to judge the need to change activity for the more likely encountered 'normal' situation? (But not to fool ourselves that extremes cannot happen.

A danger with the latter option, which may apply to many SOPs, is that pilots may be tempted to fit a situation to the SOP because the SOP is known to cover 'all' of the extreme combinations in a situation (as defined by someone else). This bias can be reinforced if training influences the 'expected' norm (extreme failure case), thus reducing the need to understand the situation, e.g. GA training 'always' involves engine failure.

There are probably good psychological arguments for one view or the other. One powerful argument is that of framing a situation, or 'nudging' people toward a particular behaviour. However, do the so called experts of behaviour - or self styled managers, actually consider that the line pilots are the real experts, particularly those who have experienced situations with the use of 'non-normal' SOPs in normal situations.
Do SOPs reflect the real, practical world which pilots operate in, or are they written based on theoretical premise or regulatory need; especially if the latter is biased by the theory.

Re considering line pilots as a source of expertise, see the ASAGA study * which identifies problem areas and solutions from a pilots perspective (section 3., and page 38 onwards). Also note the differences in severity rating between pilots and instructors for the same problems. Are instructors being overly sensitive to pilot performance, or are pilots operating so close to the limit of their mental performance they are unaware of the high workload, and thus more liable to sudden loss of mental capacity enabling awareness and control ... surprise.

An interesting analogy is that choosing to use SOPs to define a situation might involve similar mental processes to those involved in illusions.
SD, SI, relate to 'confused' senses; perhaps poorly crafted (weakly framed) SOPs add to mental workload and confusion, which results in an illusion of awareness, understanding, and being in control.

"The first principle is that you must not fool yourself, and you are the easiest person to fool." Richard Feynman.

* ASAGA, https://www.bea.aero/etudes/asaga/asaga.study.pdf Aeroplane State Awareness during Go-Around.

[Lantirn]

Very interesting study.

Sadly the training is simply not enough for everything. 2 sim days per 6 months, is not enough. Its not about go arounds, its a dozen of subjects.

SOPs are the only reason behind those thousand succesful tasks by humans, the "experts", in everyday flights. When "experts" fail to comply due to system breakdown they are judged because they didnt comply as expected. Because they are "experts".

There are thousands of tasks and demands with combinations leading to errors that individuals have never experianced before. Its a matter of luck if this avalanche of errors will lead to an incident, accident, or a normal continuation of a flight.

But a system accounting for everything is impossible. Enjoy aviation.

[piratepete]

I have flown hundreds of base training flights on a very large jet, V1 cuts, touch and goes go arounds one or two engines you name it.I learnt very early on to only allow a relatively low thrust setting in any GO case to make the handling easy, moderate rate of climb etc, works extremely well for all concerned.Mind you we always had a light airplane.I apply a similar policy when flying the line, with a payload.Peter.

[safetypee]

Lantirn, I agree with the sentiment, but offer an alternative view of "SOPs are the only reason behind those thousand successful tasks by humans".

Pilots have been successful despite SOPs because they are able to adapt by using their 'expertise' in situations which are not as assumed by SOPs. Unfortunately in some situations the adaptation does not work, leading to the blame and train cycle.

An increase in SOP violation was forecast by Amalberti *. As an industry approaches a high level of safety there will be more violations due to a tendency for over regulation as adding more regulations and procedures because they are an easy safety response for rare complex accidents.


My take-away point from the BEA study is that in situations where human performance is limiting then it's the situation which should be changed, opposed to more training, as blame and train may not achieve any improvement.
Alternatively changing the situation could be more effective, and also cover a wider range of human issues.
The situational aspects aspect in the study were noted by several pilots; they identified complex navigation routing, unnecessary ATC calls or re-routing, and complexity and workload when using workload 'alleviating' automation. The study also noted that some crews focussed on the FMA and achieving the correct sequence of call outs as required by the SOP, opposed to flying or monitoring the aircraft flight path.


The safety investigation and responses in this incident, both from the operator, and presumably the regulator, focused on even more training, which could imply some blame or failure to learn from the incident.
This was a training flight, the supervising pilot like many instructors was faced with the difficult judgement of how far to let the other pilot continue, if too late, more training (blame), but train for what.
The 'student' might benefit from more training, but the essence of the event was the difficult conditions requiring adaptive judgement, both from the instructor and the operator, who perhaps expected training to be completed irrespective of conditions.

Several aspects of aircraft certification refer to an average pilot (no definition given), the assumption might be that with two crew both would not be below average (having a bad day) in a given situation. Of course this is a probabilistic assumption, but rarely considered in increasingly high pressure operations which may assume the best performance from everyone all the time. Amalberti also noted this - slide 20 in http://ihi.hamad.qa/en/images/Keynote_Haraden.pdf

Ref: * Amalberti: https://www.irit.fr/SIGCHI/old/docs/debat/Bad%20homburg.book2000.doc