Flydubai crash at RVI final report out

[UltraFan]

https://www.mak-iac.org/upload/iblock/3d1/report_a6-fdn_eng.pdf

[atakacs]

https://www.mak-iac.org/upload/iblock/3d1/report_a6-fdn_eng.pdf

Still reading it but overall I'm fairly impressed by the quality of the report. Lots of efforts went into it.

[Greek God]

That makes sobering reading!

[Austrian Simon]

https://www.mak-iac.org/upload/iblock/3d1/report_a6-fdn_eng.pdf

Is this official? The report has not been linked to from any of the investigation pages, the news or any other page, and the link can not be found via search? How did you get this link?

[Hogger60]

A very well written report. As detailed as I have ever seen. The one thing I would have liked to have seen was the roster patterns that these pilots flew prior to the accident. As the report touched on fatigue was a factor and I believe the rostering patterns played a big part in cumulative fatigue.

[Bravo Delta]

A very disturbing report for me. Educational for those coming up through the.ranks.
Seems more of a coroner’s inquest vs A plane crash report. Very disturbing

[UltraFan]

What amazes me is that they even checked (or at least tried to check) what the pilots could see on their HUDs when the plane was in negative Gs. They tried to get Rockwell-Collins to emulate it and make a video recording of what happens when the pilots are displaced from their seats by the negative G-forces, but the company said it was impossible. SPOILER ALERT: One of the recommendations to Flydubai is to replace the HUDs.

[UltraFan]

Is this official? The report has not been linked to from any of the investigation pages, the news or any other page, and the link can not be found via search? How did you get this link?

I'm connected. :cool: To the Internet. If you reduce the link to just the .org part, you will find yourself on the official website of the investigating authority. This is the link to the whole investigation:
https://www.mak-iac.org/en/rassledovaniya/boeing-737-800-a6-fdn-19-03-2016/

[73qanda]

Personally I think that Somotagravic probably played a significant role in the crash. I don’t agree with the conclusion that it wouldn’t have been a factor just because they were VMC.When you set 100% N1 in a light NG at night there’s not always much of a horizon being sensed from the outside world. I’ve had it once and all I wanted to do was push forward and when the instruments told me the pitch was low I found it very difficult to believe.

[DaveReidUK]

I'm connected. :cool: To the Internet. If you reduce the link to just the .org part, you will find yourself on the official website of the investigating authority. This is the link to the whole investigation:
https://www.mak-iac.org/en/rassledovaniya/boeing-737-800-a6-fdn-19-03-2016/

It's also included in the chronological listing at https://www.mak-iac.org/en/rassledovaniya/

[alf5071h]

The report is an exceptional example of a thorough investigation and well considered analysis of an accident involving human factors. This should be referenced for all investigators as how it is possible to identify significant contributions to adverse (but as might be expected) human performance, yet not conclude by blaming the human - as we might do.

Although the conclusions focus on the human aspects, which in many circumstances could be labelled ‘error’ and blame, this is skilfully avoided with balanced discussion of many diverse potential contributing factors.

The safety recommendations identify many significant issues, contributing factors, which deserve consideration by all operators and regulators. Particular issues relating experience of other aircraft types and differences in trim systems, depth of training vs information available in manufacturers manuals, both aircraft and HUD, assumption re PM being able to identify rare situations requiring intervention - incapacitation, and the limits and limiting influence of SOPs in rare situations. The HUD observations should concern everyone who uses them, questioning what is taught, display formats over the range operational situations.

Yet again issues of thrust-trim coupling during GA are identified, applicable to many types, but specifically to the 737 in this accident.
Additionally, the reoccurring issues of trim awareness, assumptions about human contribution in abnormal situations, and design philosophy (737).

The regulators involved with the 737 Max recertification should read the last page of this report; paras 5.21, 5.24 and 5.25, and footnote #38.

[Paul852]

I have to agree. I am not a pilot, but with an interest in these things I have read many reports from the USA, UK and other authorities, and this for me is the most through analysis leading to useful conclusions that I have read. Kudos are due to the team producing it.

[Twiglet1]

A very well written report. As detailed as I have ever seen. The one thing I would have liked to have seen was the roster patterns that these pilots flew prior to the accident. As the report touched on fatigue was a factor and I believe the rostering patterns played a big part in cumulative fatigue.
Agreed, a scientific evaluation of the rosters attached to the report would have given an insight. It seems strange that this has been missed.

[Euclideanplane]

Agreed, a scientific evaluation of the rosters attached to the report would have given an insight. It seems strange that this has been missed.

Based on this quote from page 105 of the report it may not have been deemed necessary:

"The analysis of compliance of the work and rest schedule within a record period (28
consecutive days) did not identify any violations. The crew had a sufficient amount of the preflight
rest. As per the submitted data, the Fatigue Management System is implemented in the airline. The
system encourages the fatigue-related confidential reports by the crewmembers for any stage of
the flight operations (the preflight, in-flight, post-flight one). For a number of quantitative
indicators the system goes beyond the national aviation legislation (that is it ensures the improved
conditions for the crewmembers). Since 2009, the airline has accumulated 450 000 flights with a
total flight time of more than 1 million hours. Within the period, 70 fatigue-related confidential
reports were submitted. The majority of them were proactive by nature – as the crewmembers
reported the fatigue presence and were removed from duty until they felt fit for flight operations."

[Stuka Child]

The report is an exceptional example of a thorough investigation and well considered analysis of an accident involving human factors. This should be referenced for all investigators as how it is possible to identify significant contributions to adverse (but as might be expected) human performance, yet not conclude by blaming the human - as we might do.

Although the conclusions focus on the human aspects, which in many circumstances could be labelled ‘error’ and blame, this is skilfully avoided with balanced discussion of many diverse potential contributing factors.

The safety recommendations identify many significant issues, contributing factors, which deserve consideration by all operators and regulators. Particular issues relating experience of other aircraft types and differences in trim systems, depth of training vs information available in manufacturers manuals, both aircraft and HUD, assumption re PM being able to identify rare situations requiring intervention - incapacitation, and the limits and limiting influence of SOPs in rare situations. The HUD observations should concern everyone who uses them, questioning what is taught, display formats over the range operational situations.

Yet again issues of thrust-trim coupling during GA are identified, applicable to many types, but specifically to the 737 in this accident.
Additionally, the reoccurring issues of trim awareness, assumptions about human contribution in abnormal situations, and design philosophy (737).

The regulators involved with the 737 Max recertification should read the last page of this report; paras 5.21, 5.24 and 5.25, and footnote #38.

And we're back to the 737!

What's with the obsession on this forum about "not blaming the pilots"? Hell yes, blame the pilots! Did you read the autopsy section? Did you read about what happened to those passengers? They found the highly fragmented remains of 63 people. Only 62 were on the manifest, they eventually identified the 63d as a foetus - one of the women on board was pregnant. 63 beautiful lives turned into mince meat.

The dead pilots are not with us anymore, and it doesn't matter what we say about them. They were the first to die. But it sure as hell should not be conceived by anyone that this was in any way an acceptable level of performance from a trained, professional flight crew. Illusions and fatigue are always going to be there - that is not an excuse to smash an airplane nose-first into the ground, along with all its contents.

On the contrary, we need to blame pilots a lot more, training and selection of crew should be a million times more restrictive and intense. It shouldn't be anyone who has money and can pass some ridiculous tests gets to be crew. Flying an airplane is not a joke.
The heavy burden of responsibility from flying living beings around at speeds generating energies massive enough to disintegrate their bodies cannot be overstated.

F/O should have yelled out "I HAVE CONTROL! LET GO OF EVERYTHING NOW!", worse case punched the Cap's lights out and immediately taken control. With 40 deg pitch down, speeding towards the ground with engines at full power is not a time to be giving advice and "no, no, no"ing or pulling half halfheartedly on the yoke at the same time the other guy is pushing down. It's time to knock out the dude that's going to kill us all, engines idle and pull out of the freaking dive.

[Contact Approach]

Who actually has the time to read 175 pages?

[safetypee]

LookingForAJob, ‘absence’ - equally you will find what ever you look for.
‘Operational’ fatigue was discussed in the report.

Contact Approach, ‘who has time’ - those with an appropriate safety culture, a willingness to learn, prepared to consider this report and other viewpoints, to debate and argue as required.

This report has a lot to learn from; if only we, the industry, is willing, able to learn.

[Double Back]

Dreadful accident. Magnificent report.

[Hogger60]

Based on this quote from page 105 of the report it may not have been deemed necessary:

"The analysis of compliance of the work and rest schedule within a record period (28
consecutive days) did not identify any violations. The crew had a sufficient amount of the preflight
rest. As per the submitted data, the Fatigue Management System is implemented in the airline. The
system encourages the fatigue-related confidential reports by the crewmembers for any stage of
the flight operations (the preflight, in-flight, post-flight one). For a number of quantitative
indicators the system goes beyond the national aviation legislation (that is it ensures the improved
conditions for the crewmembers). Since 2009, the airline has accumulated 450 000 flights with a
total flight time of more than 1 million hours. Within the period, 70 fatigue-related confidential
reports were submitted. The majority of them were proactive by nature – as the crewmembers
reported the fatigue presence and were removed from duty until they felt fit for flight operations."

What it doesn't say is that the board reviewed the actual FRMS data for the rosters of the two pilots in the weeks before the crash. How many back side of the clock flights did they do in the preceding weeks? One may be legal with regards to rest and still be suffering from the affects of cumulative or chronic fatigue. Those who have done back to back to back overnight flights know what one feels like after two or three in a row. I don't see this detailed in the report.

[F-16GUY]

Who actually has the time to read 175 pages?

He who wants to learn and do better...

[OldnGrounded]

Who actually has the time to read 175 pages?

With respect, I would say that quite a number of people should make time.

[OldnGrounded]

Dreadful accident. Magnificent report.

Yes and yes.

[73qanda]

How many back side of the clock flights did they do in the preceding weeks? One may be legal with regards to rest and still be suffering from the affects of cumulative or chronic fatigue. Those who have done back to back to back overnight flights know what one feels like after two or three in a row. I don't see this detailed in the report.
​​​​​​​I agree. Being ‘legal’ and being unaffected by fatigue are two very different things.

[golfyankeesierra]

The report is an exceptional example of a thorough investigation and well considered analysis of an accident involving human factors. This should be referenced for all investigators as how it is possible to identify significant contributions to adverse (but as might be expected) human performance, yet not conclude by blaming the human - as we might do.

Alf, and others, thanks for the heads up. Saved it and will read it when I have some more time available. This really makes pprune worth it!

[Oasis]

Who actually has the time to read 175 pages?

He/she who is crossing the pacific at night on a flight deck.

[RickNRoll]

F/O should have taken control? He seemed to know what was happening. CPT should have admitted to himself he had no idea what was happening.

[Bravo Delta]

F/O should have taken control? He seemed to know what was happening. CPT should have admitted to himself he had no idea what was happening.



CORRECT limited time though.

[Double Back]

Been long enough in the RH seat to know that to physically take over control is maybe the most difficult decision to take as a F/O . However in this case... he saw death approaching, In my view there was some kind of "blurry" steep authority gradient, caused by different cultures/backgrounds/languages.

[Mr Optimistic]

(pax) I read the report and found it apparently thorough but a bit ' discursive'. It is as if they wanted to make the report a ' home' for everything they considered or photographed. The -1.4g excursion got my attention, as did the observation that uncontrolled leg movements are considered a typical sign of loss of personal control, which I assume means what a layman may call ' panic'. I think it harsh to blame the F/O given the timeline and collapsing situation with the PF hands locked on the column, pressing the trim switch and pumping the pedals. The fact that the PF kept harping on about the speed excursion, including engaging with cabin crew on this, perhaps indicates his mental calm was fragile before the long hold which was indicative of his anxiety to avoid a diversion. A dreadful read and a very great pity. The F/O was getting there but there wasn't time for him to recover the situation given the dynamics.

[Fox_one]

A really tragic accident. Lots of lessons to be learned though. A couple of things that stood out for me:

Fatigue

If it’s not reported then there is no problem.

Between 2009 and 2016, flydubai performed 450,000 flights and had 70 fatigue reports. Seriously, 70 reports, that’s a report rate of 0.015 %. There’s your problem right there, clearly a culture where people do not feel able to report fatigue.

It really highlights why FTLs are not fit for purpose and offer no protection against the effects of fatigue and tiredness. How can a safety critical industry just ignore scientific and medical studies? Oh yeah, because if you are tired, or fatigued you just say so. Well that doesn’t work in all airlines (or any airlines IMO) does it? See above!

Pilot Training

I still find it unbelievable that professional airline pilots do not need to do stall training in a jet or upset training (in an aircraft) or be subjected to G force. We could all be faced with a situation like this. At the end of a long duty, on a dark, crappy night either our control inputs or some upset expose us to confusing and previously unknown psychological effects such as G force and within seconds we are completely disoriented. Already tried and under high stress and now we subjected forces and feelings we’ve never experienced, and guess what we probably won’t do such a great job.

This was an experienced crew who had been making decent decisions and within a few seconds, the PIC lost all spacial awareness and flew a perfectly good aircraft into the ground. But oh yeah, training like that would be pretty expensive so let’s just tick the box in the simulator and pray.

It’s a bit like training soldiers at laser quest and then sending them to war. “Don’t worry guys, there’s a little bit more noise and the odd explosion but your training will kick in, good luck!”.

And no, fully autonomous aircraft are not the answer. Properly trained crews, with sensible rosters and the best tech is.

[rb14]

Who actually has the time to read 175 pages?
Me. A newly-qualified PPL (Piper Warrior) who just wants to know.

[alf5071h]

The investigation did consider this aspect; there were indications that the PM had the better understanding of the rapidly developing situation. However, whilst he could ‘see’ what the aircraft was doing, not knowing why might have delayed the choice of ‘drastic’ intervention after repeatedly alerting the PF to the abnormal condition (CVR).

The report discusses control force from the FDR which suggests that there was dual control input, even to the point of left / right control breakout - PF pushing, PM pulling, but the situation at that time was significant nose down attitude, speed increasing, and the trim moving fully forward.
The above indicates that the concept of monitoring and crew intervention cannot be assured in all situations.

The difficulty in identifying trim operation and position (as in the Max accidents) is explained, as is the mechanics of control forces and impossibility of effecting a recovery at low altitude, again as in the Max accidents. This point should be noted by regulators / operators if the 737 Max post MCAS mod, were to revert to a ‘near NG’ configuration, in that any assumption requiring quick identification and pilot action to intervene for a trim malfunction is misjudged given the findings in this accident.
Also see footnote #38 page 175.

[excrab]

Fox_one perfectly highlights the problem with flydubai's fatigue reporting system. If you reported fatigued you would be taken off flying whilst the company decided, in consultation with an AME, if fatigue was an issue. This could take a considerable time, and if they decided that it was not then the time you spent not flying was recorded as sick leave. Bearing in mind that you were only allowed paid sick leave for fifteen days per year, if they decided that there was nothing wrong with your roster and you were therefore not fatigued you could potentially spend a long period not being paid until you returned to flying, and living in Dubai is expensive, especially if you have a family. To be fair to the company they did go through the motions of improving things, but even two years after the accident one of the second officers was called into a meeting with the chief pilot after he had not felt fit to fly due to lack of sleep having been kept awake by his newly born baby, to be told by the chief pilot that in his (the chief pilot's) opinion, if he could not manage his sleep with a new baby in the house he was not only unfit to be a professional pilot, but unfit to be a parent. That, by the way, was the new chief pilot who was put in place when the previous one was promoted out of the way following the accident. With that sort of thing to contend with it is hardly surprising that people didn't report fatigued, instead it was not uncommon for occupants of both seats to be taking controlled rest of one or two hours on long night flights.

There were also issues with training. I joined the company as a type rated DEC two years before the accident happened and left two years afterwards. For my OCC I was paired up with a type rated F/O, and despite what was quoted in the report regarding the training to use the HUD we never swapped seats, so that was at least one F/O who didn't get to use the HUD during his training. Also, despite the syllabus quoted in the report, shortly after the accident I asked an F/O who I knew had joined non type rated how many standard two engine go-arounds he had done in training and the answer was one, flown by the captain during LVO training. He had done a couple of others due to systems failures, also in the LVO training, but had never done one himself. He had also never flown using the HUD as his training had been done with two F/Os and a stand in captain, so also no seat swapping, although he had seen the hud repeater image in the back of the sim whilst observing. Also, up until the accident there was no SOP to brief how a go around was to be flown, indeed when I did this on my first day of line training I was told by the LTC not to do so, as it wasn't SOP and might confuse F/Os who did not have English as their first language.

On the subject of the HUD, I can't help thinking that flydubai did the accident crew, and indeed all the crew members who went through the type rating training with them, a great miss service. The SOP, as high lighted in the accident report, was for the HUD to be used for all stages of flight, and as mentioned in the report their are many stages of flight where this is not an advantage. Flying the HUD, for those who haven't done it, is just like being in a giant computer game, and takes away a lot of situational awareness, as you don't have the map display to look at. Also, flydubai got everyone into a mind set of flying all approaches manually from 1000ft, as that was what was required for a cat3a approach. With the weather conditions these pilots experienced at Rostov I9and this is just personal) would have left the automatics engaged for as long as possible, right down to cat1 minimas if I felt it neccessary. There was no reason for the captain to disengage the autopilot and auto-throttle except that it was what he was used to doing. Many of the captains who joined the company from other airlines would ignore the SOP and only use the HUD for cat2 or 3 approaches. Unfortunately flydubais training in bad weather ops was limited, as most line training flights were done on short sectors around the Gulf to make it easier for the line trainers.

As a caveat to all this, I left two years ago. Maybe things have changed, but I'm surprised that this isn't being discussed on the ME forum or that there don't seem to be more posts from flydubai pilots. Having said that I am lucky not to have to access pprune through a VPN anymore, and wonder if it is still blocked in the UAE?

[wiggy]

Also, flydubai got everyone into a mind set of flying all approaches manually from 1000ft, as that was what was required for a cat3a approach.

Apologies if I'm being thick, probably am, but I'm a non HUD user and that comment has confused me. Are you saying the SOP in CAT 3a was autopilot out at 1000' and then a hand flown approach to (?) a hand flown landing ("through" the HUD??)...

Other than that question I thought the whole report was a very sobering read and provides a classic example of how quickly things can deteriorate in aviation from being essentially "Ops normal though a bit challenging", to coming completely undone..

[Chesty Morgan]

Apologies if I'm being thick, probably am, but I'm a non HUD user and that comment has confused me. Are you saying the SOP in CAT 3a was autopilot out at 1000' and then a hand flown approach to (?) a hand flown landing ("through" the HUD??)...


Different type (E195) and airline but that's exactly what we did for CAT3A approaches.

[misd-agin]

Who actually has the time to read 175 pages?
It doesn't require reading every single line on every single page. Thirty minutes should provide enough time to review the portions most interesting to a pilot.

[excrab]

wiggy - yes that’s exactly what I’m saying, for a cat3a approach A/T and A/P disengaged by 1000 ft Rad alt, and hand flown to a manual landing following the HUD cues for RETARD and flaring by following the steering cue (target circle in the doughnut). It works well, but not the best thing to do when you’re knackered after a 12 hr night duty; and because of this the normal teaching of leaving the automatics in until you get low down beneath the cloud base with increased peripheral vision of the runway environment of flying a cat 1 or non-precision approach was not taught.
And let’s not be mistaken here, there was one reason for flydubai having the HUD. It wasn’t to
make the operation safer or give the captain increased situational awareness, it was because it was cheaper than keeping the autopilots certified for dual channel
approaches and auto land.

[misd-agin]

"Experienced pilots". Both pilots had roughly 6,000 hrs TT. I upgraded with about 6,000 hrs and wouldn't have called myself "experienced." 15,000 hrs? 20,000 hrs? Yes. 6,000? No.

[wiggy]

wiggy - yes that’s exactly what I’m saying, for a cat3a approach A/T and A/P disengaged by 1000 ft Rad alt, and hand flown to a manual landing following the HUD cues for RETARD and flaring by following the steering cue (target circle in the doughnut)....




Ah OK many thanks to you and Chesty for the explanation and also the extra info.

We're (T7) very much automatics in in poor wx, and the full use of the automatics are very much recommended for as long as possible if there's prospect of windshear. As a result I had a :confused: the first time I read about the autopilot disconnect in the report, but can now see why perhaps the crew were "spring loaded" to knocking the automatics out....

[Twiglet1]

Based on this quote from page 105 of the report it may not have been deemed necessary:

"The analysis of compliance of the work and rest schedule within a record period (28 consecutive days) did not identify any violations. The crew had a sufficient amount of the preflight rest. As per the submitted data, the Fatigue Management System is implemented in the airline. The system encourages the fatigue-related confidential reports by the crewmembers for any stage of the flight operations (the preflight, in-flight, post-flight one). For a number of quantitative indicators the system goes beyond the national aviation legislation (that is it ensures the improved conditions for the crewmembers). Since 2009, the airline has accumulated 450 000 flights with a total flight time of more than 1 million hours. Within the period, 70 fatigue-related confidential reports were submitted. The majority of them were proactive by nature – as the crewmembers reported the fatigue presence and were removed from duty until they felt fit for flight operations."

Euclideanplane I hope you have seen from some of the posts since you posted that your comments with respect show a lack of understanding about FRMS. I'm no sleep scientist (how many Airlines engage such folk) but with the early/late/night FDP's at FZ - there will be a fatigue issue. The FRMS was clearly not functioning.

[bounce'em all]

Every FRMS I’ve seen so far is absolute bull****. We all know how some (all) airlines actually manage it. But hey, most of us need the job, right ?

[Euclideanplane]

Euclideanplane I hope you have seen from some of the posts since you posted that your comments with respect show a lack of understanding about FRMS. I'm no sleep scientist (how many Airlines engage such folk) but with the early/late/night FDP's at FZ - there will be a fatigue issue. The FRMS was clearly not functioning.
I posted the quote vertabim because another poster had wondered why the fatigue aspect appears relatively downplayed in the report. It is unusual for a final report on an accident of this character to treat the possible issue with just one short paragraph like this.

[vaneyck]

As to the question of the F/O taking control - A comment from one of his examiners at a training and proficiency check (p 39) says:

Needs to be quite a bit more assertive in
what is needed from the Captain. Tell
him/her what you want done and do not
wait for the Captain to enquire with you or
direct you in this regard. Need to be more
decisive in taking actions when needed.

[OldnGrounded]

As to the question of the F/O taking control - A comment from one of his examiners at a training and proficiency check (p 39) says:

Needs to be quite a bit more assertive in
what is needed from the Captain. Tell
him/her what you want done and do not
wait for the Captain to enquire with you or
direct you in this regard. Need to be more
decisive in taking actions when needed.

And it was less than one minute between the (second) go-around and the moment the CVR clearly indicates that the FO knew they were in trouble. Indeed, it was about 20 seconds from the exchange: "Go around" -- "Go Around" to the CVR recording the PM saying, "Keep it to fifteen degrees, nose up." 25 seconds after that, we have the PF exclaiming, "Oh sh*t!."

[UltraFan]

"Experienced pilots". Both pilots had roughly 6,000 hrs TT. I upgraded with about 6,000 hrs and wouldn't have called myself "experienced." 15,000 hrs? 20,000 hrs? Yes. 6,000? No.

Yeah, those 50-year-old puddle-jumpers... who the hell do they think they are with their meagre 10k hours! Pfft... amateurs!

However... Let's do some math. A pilot is legally allowed to fly (depending on the country) anywhere between 70 and 90 hours a month. Let's take 80 as an average (because otherwise I'll need a calculator). Even if the pilots gets no vacation or sick leave, 20,000 hours would take 20+ years of legal-limit flying. Some countries limit the minimum age of commercial pilots to 21 or even 23 years old. Therefore, your experienced-title-worthy age doesn't come until at least 40, likely 45. Would you want to call 40+ years old captains with 10-15 years of flying under their belts INexperienced? Or are you just having one of those grass-was-greener-in-my-day moments? :)

PS And OH, YES, you called yourself experienced! Many, many years later you may have realized that you probably hadn't been, but back then... LIKE HELL YOU DIDN'T!!!

[Phantom Driver]

In the old classic 737 ( I'm talking -200 ) many many years ago ( nothing fancy like AT ) we were taught on GA , after establishing initial 15 degree pitch , next thing you did pronto was pull those thrust levers back a notch or two , otherwise you'd have a ( generally light ) aircraft going like a bat out of hell , with all attendant consequences regarding flap limits / altitude busts .

On my later Boeing types ( 744 / 777 ) , the auto thrust was programmed to do it for you on GA ; brought the power back sufficient to maintain 2000 fpm rate of climb . Always nice to slow things down a bit while brain cells were reorganised .

(p.s. did a spell on A 330/340 ; great aircraft , but one complaint was that , on GA , the system gave full thrust , no matter what . Now this was ok on the 340-300 ( the one with the 4 CFM hairdriers ), but on the 330 (RR donks) it could be a handful ... Our company lost a 320 on GA ; for sure all that thrust contributed to the somatogravic illusion that sadly is still all too common) .

[400guy2]

Or are you just having one of those grass-was-greener-in-my-day moments? :)


Yes, it was; thank you for asking. But it is not just the grass: the older I get, the better I used to be.

[CurtainTwitcher]

I find this study quite useful as a review of previous Go-Around incidents & accidents: BEA on Aeroplane State Awareness during Go-Around (https://www.bea.aero/etudes/asaga/asaga.study.pdf), Published August 2013.

Phantom Driver, one of the incidents in the BEA study is an A320 Go-Around event in Melbourne in 2007, which the crew tried to manage the anticipated excessive rate of climb by reducing thrust: Go-around event Melbourne Airport, Victoria, 21 July 2007, VH-VQT, Airbus Industrie A320-232 (https://www.atsb.gov.au/media/793232/ao2007044.pdf).

[misd-agin]

Yeah, those 50-year-old puddle-jumpers... who the hell do they think they are with their meagre 10k hours! Pfft... amateurs!

However... Let's do some math. A pilot is legally allowed to fly (depending on the country) anywhere between 70 and 90 hours a month. Let's take 80 as an average (because otherwise I'll need a calculator). Even if the pilots gets no vacation or sick leave, 20,000 hours would take 20+ years of legal-limit flying. Some countries limit the minimum age of commercial pilots to 21 or even 23 years old. Therefore, your experienced-title-worthy age doesn't come until at least 40, likely 45. Would you want to call 40+ years old captains with 10-15 years of flying under their belts INexperienced? Or are you just having one of those grass-was-greener-in-my-day moments? :)

PS And OH, YES, you called yourself experienced! Many, many years later you may have realized that you probably hadn't been, but back then... LIKE HELL YOU DIDN'T!!!

Maybe we use the English language differently on this side of the pond. Where did I call myself "experienced" as an airline pilot with 6,000 hrs TT? When I had that little experience? You might have but I certainly didn't.

This has come up with coworkers and we discussed this privately when a report called a 12,000 hr pilot 'highly experienced.' None of my peers thought that was highly experienced. It might be semantics but we disagree on what experienced, or highly experienced, means. Does a person who only flies into one city get to call himself highly experienced? What about the guy who has limited experience on a lot of aircraft? Does he get to claim 'experienced', or 'highly experienced', when his experience in his new position is relatively limited (500 hrs? 1000 hrs?)?

[misd-agin]

(p.s. did a spell on A 330/340 ; great aircraft , but one complaint was that , on GA , the system gave full thrust , no matter what . Now this was ok on the 340-300 ( the one with the 4 CFM hairdriers ), but on the 330 (RR donks) it could be a handful ... Our company lost a 320 on GA ; for sure all that thrust contributed to the somatogravic illusion that sadly is still all too common) .

There's an option to reduce power to CLB after selecting TOGA on the AB. TOGA gets you the go-around guidance and then reducing to CLB on the Airbus reduces the rate of climb to a rate similar to just pushing the TOGA button once on a Boeing. Full/max power, and a landing weight, can induce vertigo issues, time comparison, high rates of climb, etc. Selecting V/S 1000' FPM is also a way to reduce the rate of climb and make the go-around less hectic.

[JPJP]

On the subject of the HUD, I can't help thinking that flydubai did the accident crew, and indeed all the crew members who went through the type rating training with them, a great miss service. The SOP, as high lighted in the accident report, was for the HUD to be used for all stages of flight, and as mentioned in the report their are many stages of flight where this is not an advantage. Flying the HUD, for those who haven't done it, is just like being in a giant computer game, and takes away a lot of situational awareness, as you don't have the map display to look at. Also, flydubai got everyone into a mind set of flying all approaches manually from 1000ft, as that was what was required for a cat3a approach. With the weather conditions these pilots experienced at Rostov I9and this is just personal) would have left the automatics engaged for as long as possible, right down to cat1 minimas if I felt it neccessary. There was no reason for the captain to disengage the autopilot and auto-throttle except that it was what he was used to doing. Many of the captains who joined the company from other airlines would ignore the SOP and only use the HUD for cat2 or 3 approaches. Unfortunately flydubais training in bad weather ops was limited, as most line training flights were done on short sectors around the Gulf to make it easier for the line trainers.


Your points regarding the HUD are well made. The GA mode is poorly thought out. Apart from a HUD takeoff there is no other mode that uses the same cues. As a result, it’s the least used.

The ‘bore-sight’ and pitch target aren't intuitive, and the transition to the normal symbology shortly afterward is an unnecessary distraction. Especially given the other factors involved (low altitude, high thrust etc.) The GA mode also has a distractingly large perceptual distance between the bore sight and the dotted line. Which adds to any somatographic illusion.

Regarding an FO and there interaction with a CA flying either a GA, or a low vis takeoff on the HUD - they should understand (or be briefed, as you alluded to) the threat associated with the modes. ie. ‘if you see the CA pitching above 15 degrees ANU, be aware that they may be going down a slippery slope’.

[UltraFan]

Flight Global goes Daily Mail on flydubai 981:
https://www.flightglobal.com/news/articles/flydubai-737-could-have-been-saved-6s-before-impact-462590/

[Jwscud]

I can’t help but wonder given that the crew hours were “legal” that there may have been some politics at play in downplaying fatigue. I seem to recall the roster of one of the pilots being posted on here just after the accident and it was pretty bloody awful.

I also found the HUD and guidance cues deeply confusing. It was clear the FO had far better SA throughout, which although was a consequence of his not manually flying the aircraft was also I think because he had traditional instruments in front of him. I suspect with a blue/brown PFD and the traditional crossbar flight directors this would just have been a messy go around.

[CargoOne]

excrab
I am confused with the purpose of HUD on 737NG and performing CATIIIA landing with A/P and A/T disengaged. There are aircraft types with no autoland (like CRJ) where HUD option gives you CATIIIA capability for manual landing, so it is clear. But why would anyone use HUD & manual while 737NG can perform CATIIIA autoland?

[SOPS]

Cargo One.. I have been thinking the same thing after reading the report.

I would also love to see their rosters prior to this flight. Legal does not mean safe. And airlines in this area ( and many others), are well known for making the legal limit, not a limit, but a target.

[wiggy]

excrab
I am confused with the purpose of HUD on 737NG and performing CATIIIA landing with A/P and A/T disengaged. There are aircraft types with no autoland (like CRJ) where HUD option gives you CATIIIA capability for manual landing, so it is clear. But why would anyone use HUD & manual while 737NG can perform CATIIIA autoland?

excrab made this comment in #38 :

"there was one reason for flydubai having the HUD. It wasn’t to make the operation safer or give the captain increased situational awareness, it was because it was cheaper than keeping the autopilots certified for dual channel"
approaches and auto land."

[Double Back]

Retired 10+ Years ago after having commanded 17 Years on relative "simple" A/C like the 744, I do not envy pilots having to deal with those "stuck or parked" throttles. I could fly an auto GA with the 744 with my eyes closed, NO FMA indications needed at all, all "natural" feedback from pitch movements from the controls, INCLUDING forward A/T driven movement of the throttles to the approximate correct position, G-forces, sound of engines. The position of the throttles in GA or T/O was etched into my stretched arm's muscle memory, NO way I could overlook that.
It took a while though before the company accepted that in a GA it was allowed to leave one hand on the throttles, because officially one had to release them as soon as forward movement was detected. A bit curious as during the approach one was supposed to follow up on the A/T by resting one hand on those throttles.
Muscle memory is way more reliable than inputs obtained by vision, especially in stressful situations when the brains gets saturated with vision inputs.

Studying the presentations of the possible HUD indications make me feel glad I never had to deal with such a system.
The GA procedure should be an extremely simple one, in my career as long hauler I made maybe only 3 actual, so to speak a kind of "emergency". All in good weather. One was F$#*ed up by ATC by requesting us to level off at an altitude we just were zooming through on our way to the published one. Really helpful :( :( :(
MAYBE we are on our way to full pilot less flight and the industry is now coping how to use the developing automatics to "help" the pilots in this era in between. But I am not impressed with the progress that is made because of the arrogance of the automatics.
I just heard the car industry is slowly loosing the idea of ever reaching full automatic control of cars, at least not in the current road infrastructure. Inside concrete tubes, yes, might be possible :)

[Twiglet1]

A really tragic accident. Lots of lessons to be learned though. A couple of things that stood out for me:

Fatigue

If it’s not reported then there is no problem.

Between 2009 and 2016, flydubai performed 450,000 flights and had 70 fatigue reports. Seriously, 70 reports, that’s a report rate of 0.015 %. There’s your problem right there, clearly a culture where people do not feel able to report fatigue.

It really highlights why FTLs are not fit for purpose and offer no protection against the effects of fatigue and tiredness. How can a safety critical industry just ignore scientific and medical studies? Oh yeah, because if you are tired, or fatigued you just say so. Well that doesn’t work in all airlines (or any airlines IMO) does it? See above! .

70 reports in 450,000 flights - wonder what Ryanair had for same amount. Then throw them up against an Airline such as Easyjet who has an FRMS which is up there at the front

[autobrake3]

70 reports in 450,000 flights - wonder what Ryanair had for same amount. Then throw them up against an Airline such as Easyjet who has an FRMS which is up there at the front

The problem at Easyjet, is that thousands of fatigue reports have been submitted and yet nothing has been done about the issues contained in these reports. However, at least you can report fatigued for a duty and then blow off steam by submitting the required report without being invited to a serious grilling and bullying exercise at head office. This is probably one of the reasons why there is not a fatigue issue in Harp land.

As for this particular sad accident, it has fatigue written all over it. Their rosters were generally known to be punishing, they had already been operating for 7 hours in very difficult flying conditions, they had already flown a go around in ice and windshear which the Captain was subsequently fretting about and it was about 4 in the morning for their body clocks. No wonder the Captain ran out of cognitive capacity.

[Centaurus]

Interesting discussion indeed. As a contractor flight simulator instructor for many years I see some quite scary moments while training pilots on the 737 simulator. Low altitude go-arounds on instruments in particular. Automation addicted pilots seem to have the most trouble because of their lack of manually flown no flight director basic raw data instrument flying skills.

One exercise in particular is a reproduction of the Turkish Airlines 737 crash at Amsterdam where the autothrottles closed to idle prematurely while on a coupled ILS approach. Google the accident if more details are required.
For the simulator exercise the autothrottles are set to idle at 1500 feet on final and as the autopilot attempts to maintain the ILS glide slope, eventually the stick shaker actuates. By then the stabiliser trim has wound a long way back under the influence of the autopilot. A go-around is initiated with the IAS in the region of VREF minus 20 knots.

The thrust levers are manually positioned to max thrust for the go-around and the ensuing severe pitch up is further exacerbated by the aft position of the stabiliser trim. The pilot can only contain the pitch up that occurs by forward control column and holding constant forward stabiliser trim for approximately six seconds which permits fairing of the stabiliser and elevator thus allowing normal elevator control to keep the pitch attitude within reasonable limits.

Flown in IMC it takes very careful handling to keep out of the stall regime and on the other hand to prevent a dangerously high nose attitude. A common factor seen is failure by the pilot to stop forward trimming of the stabiliser and this causes the initial pitch up to turn into a dive. In other words the high workload in the few seconds of the go-around results in the pilot forgetting to release the stabiliser thumb switch.

Another factor often noticed is the pilot forgetting to double click the autopilot disconnect button on the control column. In the stress of the go-around he may disconnect the autopilot with one click of the AP disconnect button but then there is the continuous noise of the wailer which is distracting and only adds to the perceived urgency of the situation. Low altitude manually flown go-arounds in IMC without use of the autothrottles and without use of the flight director need to be practiced regularly during simulator training in order to keep instrument flying skills up to scratch. Automatic pilot go-arounds then become a doddle.

[alf5071h]

Humans are simple creatures in a complex world. We crave understanding, we simplify issues and discuss them in isolation (we are lazy - cognitive misers), and detesting uncertainty conclude ‘black or white’, not wishing to accept the reality of a grey world.

The report is ‘grey’ - a very good shade, without ‘definitive conclusion’, yet able to identify and discuss contributing factors which could have resulted in the observed behaviour (hindsight).
Factors were reviewed in isolation, enabling each to be examined for their potential to improve safety, but not excluding more realistic or unidentifiable combinations - a subjective task requiring skills of critical thought and acceptance that there is no single solution - systems thinking - foresight.

The first GA might not have been totally unexpected, but still surprising - different levels and duration of mental stress. A Wind-Shear GA is different, how often practiced vs a normal GA.
Aircraft with large thrust-pitch coupling may use reduced thrust for a ‘normal’ GA to reduce pitching tendency and trim effort, but for WSGA the SOP mandates max power, no matter what else - taught and trained by rote.

The second GA, 2 hrs less fuel, light weight aircraft, max thrust, overshoots the potentially misleading pitch target in the HUD. The HUD approved bypassing normal certification and by default accepted by other regulators. Forward stick is maintained until trim catches up; but in this aircraft stick force only reduces with opposite stick movement; it is possible to fly the aircraft with trim.
The HUD format changes with ‘excessive’ attitude to retain a horizon line, but unlike EFIS it compresses the pitch scale with potential for inducing false sense of pitch rate, also ‘upwards’ cues are presented, not related to pitch-roll convention (90deg) and might be misinterpreted as a roll indication.

We are unable - unwilling to consider a graduated response between GA and WSGA, too many variables, complexity - thus KISS (of death), ultimately resorting to the crews judgement, judged after the fact in a mandated SOP culture, with display information based on static format designs - not considering the dynamics of the situation, man, machine, and environment, all assumed (canned in SOPs) for simplicity.
Systems thinking starts by accepting that there is no ‘solution’ - we are challenged by a complex adaptive issue which at best might be contained, and with careful consideration and small interventions, safety might be improved - but we may never know.

In a highly reliable industry we must be careful not to disturbed the fine balance which has enabled this level of safety; don’t jump to conclusions or rush to yet more training; list and check assumptions.
Start with the individual, ourselves; would the issues affect us, if not why not.
There is opportunity in either viewpoint, the key is to think about the issues and form a viewpoint and understand why (you, not others).
.

[PGA]

Experience simply isn’t related to hours flown, as was alluded to previously to by misd-agin

I just happened to pick up D.P. Davies his book again, “only” 5000+ hours, but I doubt I’ll ever reach the experience this gentleman had in my whole career.

Flying a widebody in cruise teaches one very little about flying. I now teach how to fly a widebody, and although my hours now number well in to the 5 digits, the hours in cruise didn’t matter all that much..

In my humble opinion exposures and practise is the key, not flight hours. If you get the opportunity to learn how to properly fly your aircraft in varied conditions, you’ll be ever so much better then if you fly 20k hours across the Pacific with the autopilot engaged.

[FlightDetent]

excrab made this comment in #38 :

"there was one reason for flydubai having the HUD. It wasn’t to make the operation safer or give the captain increased situational awareness, it was because it was cheaper than keeping the autopilots certified for dual-channel approaches and auto land." Not enough for me. The -3/4/500 had no dual autopilot neither HUD and was CATIIIA. The question stands, what kind of assessment yielded the result, not to use a certified AP/AT for LVP.

@misd-agin: The present wording on AB, formally interpreted, now instructs the crew to immediately reduce to CLB thrust.

---

The report raises an issue with force-on-yoke vs. trim technique, saying (my wording) to the investigators it was not a new find ... applicable to A/C with movable horizontal stabiliser. Massive truth, and also the note about trim rocker switch being prone to cognitive lock (and alternative technical solutions).

[excrab]

[QUOTE=FlightDetent;10629190]Not enough for me. The -3/4/500 had no dual autopilot neither HUD and was CATIIIA. The question stands, what kind of assessment yielded the result, not to use a certified AP/AT for LVP.

Flight detent

I don’t know what type of 300/400 or 500 you flew. I’ve not flown a 400, but all the 737 300s and 500s that I flew certainly had dual autopilot, both of which had to be armed for a dual channel approach and auto-land exactly the same as the 800.

I don’t know what assessment was made by flydubai when they decided to use the HUD, as the decision was made before I joined. However, the following is a direct quote from an email I received from the safety department a year before the accident, when I queried why dual channel approaches weren’t flown, as doing so would allow automatic go arounds to be flown with a reduction in crew workload :

” Dual channel approaches would involve a lot of engineering/ maintenance/ servicing/ paperwork and increase training costs. As yet, the costs probably outweigh the benefits”.

So from that it is evident that at some point prior to the accident, someone in Flydubai had done a cost analysis and decided not to use dual channel approaches and instead to use the HUD. If that still isn’t enough for you there is nothing else I can add.

[Bradley Hardacre]

OK, the HUD is designed to allow the pilot to see through the window while superimposing guidance. Why would you need to look out of the window during a go around?

The FO, monitoring the PFD could see exactly why it was going pear shaped, the PF was experiencing a somatogravic upset exacerbated by his view of the landing lights rushing into the overcast.

[Phantom Driver]

Centaurus--
A common factor seen is failure by the pilot to stop forward trimming of the stabiliser and this causes the initial pitch up to turn into a dive. In other words the high workload in the few seconds of the go-around results in the pilot forgetting to release the stabiliser thumb switch.

Ahh , TRIM . After 5 years on the Airbus , it was strange having to retrain the left thumb to start blipping the trim switch again on the jumbo .
Reminds me of my training on the 737 back in the dark ages . Favourite trick of the crusty old trainer , somewhere on finals , was to say " take your hands off the stick " just to see if you had the aircraft in trim or not .; you soon learned to make sure you were ...

[FlightDetent]

So from that it is evident that at some point prior to the accident, someone in Flydubai had done a cost analysis and decided not to use dual channel approaches and instead to use the HUD. If that still isn’t enough for you there is nothing else I can add. excrab, sincere apologies. I could not have been more wrong. You are correct on all accounts including my previous type setup. Baffling, both in the original and quote of your post, I mistakingly read it as if the airline decided against dual-rudder (fail-operational) option electing to use the HUD instead, and that also for CATIIIA approaches. :*

Ironically, in the imaginary world where I had a say on how to sketch an airline, purchasing HUDs that would make the approval, trainings and maintenance of autolands obsolete is an idea I'd wish to explore.

[Uplinker]

I have read most of the official report and about half of this thread, so apologies if I am going over old ground.

Two observations about HUDs, (which I have not used myself):

The one in this report is monochromatic. There s no blue for sky, brown for ground. The user cannot glance at the HUD to see if one is pitching up or down, the actual + or - of the pitch graduations have to be read and interpreted? The barbers pole on the speed scale is not red but the same colour as everything else, (albeit a fuller hashed line), so an impending overspeed does not stand out either.

The HUD presentation is different to the ADI. We all started by flying single engined Cessna’s or equivalent, with the standard “artificial horizon” instrument. The vast majority of our total flying hours have been spent looking at the familiar blue for sky, brown for ground presentation, and bank information to ascertain our attitude. FlyDubai’s SOP was to always use the HUD, but as I understand it, they only flew a small number approaches one day in a SIM to learn the HUD, and that was it? “OK, you’re fully qualified and practised to use the HUD for all flight operations in all weathers from now on”. Really?


Hell yes, blame the pilots!.....................should not be conceived by anyone that this was in any way an acceptable level of performance from a trained, professional flight crew. Illusions and fatigue are always going to be there - that is not an excuse to smash an airplane nose-first into the ground, along with all its contents.

On the contrary, we need to blame pilots a lot more, training and selection of crew should be a million times more restrictive and intense. It shouldn't be anyone who has money and can pass some ridiculous tests gets to be crew

Totally agreed, but it is not necessarily a given pilot’s fault. No pilot starts their training wanting to be a bad pilot.

I agree, and have always said that training does need to be vastly improved. I have been amazed to witness some very poor flying in the SIM from experienced Captains of wide-body aircraft. Much of recurrent SIM checks are a tick in the box if you were vaguely within limits and move on to the next exercise. Very tellingly, if the CAA are sitting in during a SIM, one has to absolutely get it right - near enough is not a pass in those situations.




F/O should have yelled out "I HAVE CONTROL! LET GO OF EVERYTHING NOW!", worse case punched the Cap's lights out and immediately taken control. With 40 deg pitch down, speeding towards the ground with engines at full power is not a time to be giving advice and "no, no, no"ing or pulling half halfheartedly on the yoke at the same time the other guy is pushing down. It's time to knock out the dude that's going to kill us all, engines idle and pull out of the freaking dive.

Absolutely, of course he should yes, and it quite obviously should have been an automatic reflex in this particular situation. All I would say is that we are not trained in taking control. We are just told to do it - in text somewhere in the manuals, but during training, F/Os are never actually put into the situation of having to take control from the Captain*. It can be a big deal to take over from a Captain, somebody maybe very senior to oneself, and whom the company has deemed to be good enough to be a Captain, while they’ve decided you’re not.

*The only time we do practise taking control is for an incapacitation, or instrument failure, which is easy because then there is no argument from the Captain - you don’t have to shout at him and wrestle the controls from him and have a disagreement; because he is ‘unconscious’ or willing.

Cabin crew are given actual instruction on how to deal with angry or dangerous passengers - how to deconflict the tension, how to physically act, and how to speak, and they practise it. If the other pilot was going to kill me, I would most certainly take over, (and have done so on two occasions), but what I am saying is; we have never been trained in this, merely told. It is a big deal to become a Captain - many don’t make it - so the image of a Captain is of one who has proved they have got what it takes. Therefore it is also a big deal to take command over the commander, so this should be trained, not merely written somewhere in the manuals.

.

[Denti]

All I would say is that we are not trained in taking control. We are just told to do it - in text somewhere in the manuals, but during training, F/Os are never actually put into the situation of having to take control from the Captain*. It can be a big deal to take over from a Captain, somebody maybe very senior to oneself, and whom the company has deemed to be good enough to be a Captain, while they’ve decided you’re not.

*The only time we do practise taking control is for an incapacitation, or instrument failure, which is easy because then there is no argument from the Captain - you don’t have to shout at him and wrestle the controls from him and have a disagreement; because he is ‘unconscious’ or willing.


That was true for my training on the 737 as well, taking control was only trained for the instance of an incapacitation. When i switched to the bus it was actually trained against a colleague trying to kill you. State loud and clearly „I have control“, press the red button and keep it pressed to lock the other out who still continues to make control inputs and recover the situation.

It is still trained pretty much every simulator event as part of the upset recovery training, for both the FO and the CPT. Very good training to be had in my view.

[Twiglet1]

The problem at Easyjet, is that thousands of fatigue reports have been submitted and yet nothing has been done about the issues contained in these reports. However, at least you can report fatigued for a duty and then blow off steam by submitting the required report without being invited to a serious grilling and bullying exercise at head office. This is probably one of the reasons why there is not a fatigue issue in Harp land.

As for this particular sad accident, it has fatigue written all over it. Their rosters were generally known to be punishing, they had already been operating for 7 hours in very difficult flying conditions, they had already flown a go around in ice and windshear which the Captain was subsequently fretting about and it was about 4 in the morning for their body clocks. No wonder the Captain ran out of cognitive capacity.
Autobrake From what I know of Easyjet plenty is done - sometimes Aircrew forget they are a highly paid asset......

[alf5071h]

The text below is copied from an HF web forum. There are some important lessons involving certification, and for operators resulting from unforeseen situations perhaps not considered in the SOP: to use HUD at all times. A Windshear GA becoming an upset, the dynamics of equipment and situations vs static evaluation, and erroneous choice of training as a solution.

In normal flight, P1 and P2 use different mental models. When the HUD upset mode appears automatically and unexpected, the Captain is forced to change the mental view and instrument references.
Although a dual HUD installation might alleviate some of the problems, the unconventional format and surprising change during a Windshear GA adds to the surprise and confusion during the GA manoeuvre - controlling pitch and trim.
The overall deficiencies in the HUD format, cf EFIS # 69, and increased workload, suggests that the better, safer option is not to use the HUD all the time.

“The HUD format (certification) and mandated use (operator SOP) would influence awareness.
Normal HUD use is flight director based; a vector display - winged aircraft symbol and circular target, which depicts where the aircraft is going. The ‘command’ task is to match the symbols (video game - skill based) reducing mental workload.

An automatic change to the ‘upset recovery’ format depends on manoeuvre.
This format is attitude based, without the FD, only depicting where the aircraft is pointing. The mental task is ‘interpretive’ - different symbology and format (knowledge based). Mentally demanding, sudden unexpected change, further limits mental resource.

Thus “The F/.O appeared to have a much better grasp of the issue, but was unable to change the outcome,”
… influenced by a better mental model / task, (head down EFIS with FD), but unable to recover due to limiting control - stick forces, trim, conflicting pilot inputs.

HUD approval via STC by-passed formal HF certification and design evaluation, only assessing operations, procedures, and training (OEB).
An opportunity to mask system deficiencies by requiring ‘more training’ (cf 737 Max).
Change the system, not the pilot.”

[fdr]

A really tragic accident. Lots of lessons to be learned though. A couple of things that stood out for me:

Fatigue

If it’s not reported then there is no problem.

Between 2009 and 2016, flydubai performed 450,000 flights and had 70 fatigue reports. Seriously, 70 reports, that’s a report rate of 0.015 %. There’s your problem right there, clearly a culture where people do not feel able to report fatigue.

It really highlights why FTLs are not fit for purpose and offer no protection against the effects of fatigue and tiredness. How can a safety critical industry just ignore scientific and medical studies? Oh yeah, because if you are tired, or fatigued you just say so. Well that doesn’t work in all airlines (or any airlines IMO) does it? See above!

Pilot Training

I still find it unbelievable that professional airline pilots do not need to do stall training in a jet or upset training (in an aircraft) or be subjected to G force. We could all be faced with a situation like this. At the end of a long duty, on a dark, crappy night either our control inputs or some upset expose us to confusing and previously unknown psychological effects such as G force and within seconds we are completely disoriented. Already tried and under high stress and now we subjected forces and feelings we’ve never experienced, and guess what we probably won’t do such a great job.

This was an experienced crew who had been making decent decisions and within a few seconds, the PIC lost all spacial awareness and flew a perfectly good aircraft into the ground. But oh yeah, training like that would be pretty expensive so let’s just tick the box in the simulator and pray.

And no, fully autonomous aircraft are not the answer. Properly trained crews, with sensible rosters and the best tech is.

It is not possible to recreate the forces that occur at the seat station In simulator. The arm from the cg is considerable and high rate pitch or yaw associated accelerations are not reflected within constraints of 6DOF motion systems. They are good, they are not exact. Doing such manoeuvres in a light jet also gives a considerably different qualitative outcome to the real aircraft.

Undertaking stalls in a large transport adds substantial buffet related loads to the tail of the aircraft, physical maintenance inspections are called for in most types due to theses loads on the structure.

Sim training is a good start, but the crews need to be aware of the limitations in fidelity as to loads experienced in the seat.

Having stalled transports in test, and having used light military jets, heavy pistons and 1 or 2 seat Pitts etc for continuity, I'm not sure that anything more than a decathlon, CAP or RV is needed to maintain a semblance of proficiency in handling. The sim gives benefits but only within the constraints of fidelity.

To achieve -1. 4g recorded at the cg requires a considerable pitch rate.

[Fursty Ferret]

Who actually has the time to read 175 pages?


I read all of it. Make time.

Some thoughts...

The HUD symbology on both the 737 and 787 is pretty terrible. Predictive and reactive wind shear generate the same “WINDSHEAR” overlay text. In a missed approach you switch from flying flight path guidance to pitch attitude, then back again fairly quickly. Everything is green, the same thickness, and either solid or dashed.

FOs are not trained to take control. Captains are trained to wrestle control from an uncooperative FO. It’s also impossible for another pilot to take control in a Boeing aircraft if the other pilot doesn’t want them too - you can see in this case that it just activated the jammed controls system.

[seven3heaven]

My 2 cents worth having read the report. Yes, its long and detailed but certainly worth reading for any pilot.

Massively disappointed by the lack of info on Fatigue and the Rosters of both pilots in the proceeding weeks. I know that the FO's roster was brutal leading up to the crash.
The number of fatigue reports submitted is eye opening and tells you all you need to know about the safety culture. I left FZ before the crash but if my blurry memory serves me correct, if you were brave enough to call in fatigued there was the threat of been sent to the AME for a "consultation" with the possibility of your medical being suspended while your fatigue was investigated.

The first discontinued approach was a windshear escape, the second was a standard go around procedure. With the first approach still fresh in the mind its so easy for any pilot who is absolutely shattered, dealing with awful weather conditions and staring through a bright HUD to get things wrong. 101% N1 was excessive but there is no N1 indication on the HUD as far as i can remember. You get so fixated on the HUD that your scan can breakdown especially when things start to become "non-normal".

The commercial pressure from ops's to "get in" is also a massive issue that doesn't really get interrogated. For sure they would have left DXB with somewhere close to full tanks of fuel knowing that the weather was marginal. Op's would always say "load it up with fuel and get in when you can", knowing that a divert will leave the operation an aircraft short for the next 12/15 hours. Normally as pilots we like to carry as much fuel as we can, but in this instance a bit of time pressure resulting from less fuel could have meant a divert earlier, maybe even after the first approach resulting in a totally different outcome.

Pretty much every pilot I ever flew with at FZ always feared that something like this might happen one day due to the fatiguing rosters and the nature of the destinations. Tragically in March 2016 all the holes in the cheese lined up.

[gtseraf]

A really tragic accident. Lots of lessons to be learned though. A couple of things that stood out for me:

Fatigue

If it’s not reported then there is no problem.

Between 2009 and 2016, flydubai performed 450,000 flights and had 70 fatigue reports. Seriously, 70 reports, that’s a report rate of 0.015 %. There’s your problem right there, clearly a culture where people do not feel able to report fatigue.

It really highlights why FTLs are not fit for purpose and offer no protection against the effects of fatigue and tiredness. How can a safety critical industry just ignore scientific and medical studies? Oh yeah, because if you are tired, or fatigued you just say so. Well that doesn’t work in all airlines (or any airlines IMO) does it? See above!

Pilot Training

I still find it unbelievable that professional airline pilots do not need to do stall training in a jet or upset training (in an aircraft) or be subjected to G force. We could all be faced with a situation like this. At the end of a long duty, on a dark, crappy night either our control inputs or some upset expose us to confusing and previously unknown psychological effects such as G force and within seconds we are completely disoriented. Already tried and under high stress and now we subjected forces and feelings we’ve never experienced, and guess what we probably won’t do such a great job.

This was an experienced crew who had been making decent decisions and within a few seconds, the PIC lost all spacial awareness and flew a perfectly good aircraft into the ground. But oh yeah, training like that would be pretty expensive so let’s just tick the box in the simulator and pray.

It’s a bit like training soldiers at laser quest and then sending them to war. “Don’t worry guys, there’s a little bit more noise and the odd explosion but your training will kick in, good luck!”.

And no, fully autonomous aircraft are not the answer. Properly trained crews, with sensible rosters and the best tech is.

Fox one, well said on both counts.

I work at a mob where the approach to fatigue is pretty cavalier, to say the least. Only in the past 2 years has the regulator required FRMS etc. I was involved for a short while is liasing with management about scheduling issues. Despite attempts to promote fatigue reporting, very few reports were submitted, with reasons varying from, "I was too tired" to "I am scared I will be victimised" being given. Many verbal complaints were made, but these are useless. Pilots are their own worst enemies when it comes to this and airlines KNOW it and use it.

I reckon the airlines will very quickly change their strategies when they receive a massive number of legitimate fatigue reports, AND they know copies are being collected by interested parties.

Is it possible that fatigue would have received a far greater emphasis in this report IF there had been a huge number of reports submitted. I am very sure the answer is "YES"

[gtseraf]

Every FRMS I’ve seen so far is absolute bull****. We all know how some (all) airlines actually manage it. But hey, most of us need the job, right ?

we also need to be alive to do the job! this kind of attitude facilitates the bull**** you refer to. As a professional pilot, you have a responsibility to maintain a certain standard and keep the operation honest!

[Uplinker]

Not being an expert, it seems to me that factors contributing to this accident were attempting a go-around using the unfamiliar, different, monochromatic symbology and presentation of an HUD, instead of the normal ADI, IAS, V/S and N1 gauges - leading to disorientation? Operation with one pilot on a HUD and the other on conventional instruments had not been sufficiently trained, nor allowed to be sufficiently practised - particularly in the case of an F/O taking control from a Captain - both using different instruments - during a high workload situation, such as a go-around, which is going pear shaped.

Add to this, a possibly fatigued crew during their WOCL, and a horrible weather situation. The Captain clearly felt pressure that he had to land at the destination, and also not go out of hours, whereas from one’s armchair it would seem reasonable to divert after the first two landing attempts.

I think I have experienced good company FRMS systems, but allowing airlines to monitor their own fatigue reporting and deal with it in-house could potentially be a case of the foxes looking after the hen house?

If fatigue reports had to be sent to and dealt with by the equivalent CAA, then fatigue would be officially examined and might actually have to change?

(I know.......dream on !)

[Denti]

If fatigue reports had to be sent to and dealt with by the equivalent CAA, then fatigue would be officially examined and might actually have to change?

(I know.......dream on !)

Is that not the case in the UK? I thought it was actually a requirement under EASA rules (which of course do not apply to FlyDubai). At least when working in a FRMS Safety Action Group in a previous airline all actual fatigue reports were sent on to the authority, however there was a dispute if pre-emptive fatigue reports ("I am concerned this roster will induce fatigue") had to be send on. Reports were actively encouraged, calling in fatigued for a duty then required a report, so there was plenty to work through.

[I'mbatman]

First approach at roughly 230AM body clock after descending through area with severe turbulence reports.
Wind gusting to over 40Kts.
Wind shear Go-Around.

Just on that I'm exhausted and jittery entering my WOCL.

Now, hold for two hours and attempt landing #2 in pretty much identical conditions.
Go around and an accident.

Their contingency plans were well thought out and appropriate and the report backs that up in terms of assessing hold fuel, diversion airports, and potential FDP issues. The one thing the PIC didn't brief at all was their level of fatigue and/or mental state after an extremely taxing evening and the very real possibility that despite a great plan(shoot another approach and divert if necessary), they physically and mentally may not have been able to cope.

So sad.

[bounce'em all]

we also need to be alive to do the job! this kind of attitude facilitates the bull**** you refer to. As a professional pilot, you have a responsibility to maintain a certain standard and keep the operation honest!

exactly right. That’s why, when I realized that in order to keep it honest I’d have had to call “fatigued” every other week (at best), I decided to take my services elsewhere. Where, for one, I fly no nights. Huge difference. I feel honest again.
Nevertheless, a couple of intense years longer, and I’ll get the f@#! out of this pointless grind.

[Uplinker]

Is that not the case in the UK? I thought it was actually a requirement under EASA rules.....

You might be right, but all the fatigue reports I have ever completed were required to be sent to the company, not the CAA.

[excrab]

And don't forget, as this thread is specifically about flydubai, that you are surmising that fatigue reports should be sent to the GCAA, the chairman of which is also the chairman of the Emirates group of which flydubai is a part. Don't expect the regulator in this case to take significant action over anything which may harm the profitability of the Emirates group. Like many things in Dubai saying something is being done and it actually being done is not always the same thing, the truth is not necessarily at the top of their list of priorities.

[G SXTY]

Since 2009, the airline has accumulated 450 000 flights with a total flight time of more than 1 million hours. Within the period, 70 fatigue-related confidential reports were submitted.

In the latter half of 2015, flight crew fatigue ASRs at Flydubai were running at around 20 per month. Some of them submitted by me. There were 21 reports in December 2015 alone. I believed then (and still do) that there was a corporate indifference to fatigue at FZ, and that the holes were starting to line up. It was one of the main reasons I left.

When the crash happened, the only thing that surprised me was the location - I could name many FZ destinations that were far more challenging and/or hazardous than Rostov. But that's the thing about fatigue - keep on sending tired crews up night after night, and the destination doesn't make much difference. Sooner or later, two equally fatigued pilots will encounter nasty weather and/or an emergency, and run out of capacity.

What an utter waste of 62 lives.

[alf5071h]

The discussion drifts back to fatigue, a single focus. Although important, it isn’t the dominant or overriding issue in this accident; that is about how it is possible for all of the factors to come together at that time.

#62 titled GA, could equally be titled fatigue, or system design, or SOPs, etc; the quote below still applies. Perhaps this viewpoint is the better item to debate.

Humans are simple creatures in a complex world. We crave understanding, we simplify issues and discuss them in isolation (we are lazy - cognitive misers), and detesting uncertainty conclude ‘black or white’, not wishing to accept the reality of a grey world.
The report is ‘grey’ - a very good shade, without ‘definitive conclusion’, yet able to identify and discuss contributing factors which could have resulted in the observed behaviour (hindsight).
Factors were reviewed in isolation, enabling each to be examined for their potential to improve safety, but not excluding more realistic or unidentifiable combinations - a subjective task requiring skills of critical thought and acceptance that there is no single solution - systems thinking - foresight
Consider how each of the agencies cited in the reports recommendations might act:- the airline, the regulator, the manufacturer, ICAO. At best, individually, they might “consider” the issues.
And anything that they or we might consider would be ineffective without action; thus the prime questions are what to act on, and who, and how is this to be actioned, what do we change - joined up thinking.

Thinking … https://ta-tutor.com/sites/ta-tutor.com/files/handouts/thinkng.pdf

[Euclideanplane]

What an utter waste of 62 lives.
Not nitpicking, but, 63. RIP little unborn one.

[olster]

Very sad event and a classic example of the Swiss cheese holes lining up. I was there for a while and there were many challenges. Not least a toxic and bully boy management culture backed up by minimal or no human rights as is typical in the ME. Pilots were understandably afraid to be fatigued or even sick. Chief pilot at the time embraced the toxic culture until being moved sideways to another management job in a training organisation having no previous training experience. You could not have made it up. Very tragic and RIP to all involved.

[Starbear]

The thrust levers are manually positioned to max thrust for the go-around and the ensuing severe pitch up is further exacerbated by the aft position of the stabiliser trim. The pilot can only contain the pitch up that occurs by forward control column and holding constant forward stabiliser trim for approximately six seconds which permits fairing of the stabiliser and elevator thus allowing normal elevator control to keep the pitch attitude within reasonable limits.

With all due respect ( and I do mean that sincerely as we are apparently both in the same field) but depending upon when you were teaching this, I would take issue with your statement (bolded by me) as the pilot does in fact have another tool with which to contain the pitch up and that is the reduction of thrust in conjunction with pitch inputs. Certainly not intuitive in such conditions but probably vital. If there is insufficient pitch control or authority thrust reduction to the point of restoring that authority MUST be considered at least.

It was exactly this type of event, as well as the likes of the Thomsonfly B737 GA at Bournemouth (UK 2007)
in very similar circumstances which led to a complete review of the stall and approach to stall procedures globally, which changed the priorities of pitch (AoA) over thrust.


Thomsonfly B737 2007
https://assets.digital.cabinet-office.gov.uk/media/5422eab5e5274a131700001f/3-2009_G-THOF.pdf

[safetypee]

Starbear, “- the pilot does in fact have another tool with which to contain the pitch up and that is the reduction of thrust in conjunction with pitch inputs.”

You appear to overlook that the GA was in response to a Windshear alert, which in general (overwhelmingly) requires maximum thrust.
At some point there may be a conflict between reducing thrust as judged by the crew in the actual conditions, and the operators SOP, - cognitive dissonance - mental effort, confusion, distraction.

What do operators teach and mandate by SOP ?
What advice do operators provide for reducing thrust after a Windshear GA ?

[OldnGrounded]

You appear to overlook that the GA was in response to a Windshear alert . . .

That's true for the first go-around, but the second appears to have been initiated because of overspeed.

At the point of One thousand advisory callout activation (00:40:37) the aircraft was nearly stabilized for the approach (the flaps at a landing position 30), landing gear down, the deviations off the beam on localizer and glideslope within tolerance), with that the PIC uttered: «Stabilizing now», most probably, speaking about speed that was equal to 163kt (and trended to reduce), which was 10kt higher than the approach speed, determined by the crew.

The aircraft was flown a little bit higher of glideslope (0.3...0.2 dots), and the PIC was applying the corrective “pushing” movements on the control column to maintain the glide path descent more precisely, along with that the thrust (N1) was increased from 65% to 70%. Over the same moment the aircraft encountered wind gust. The combination of these three factors resulted in the IAS, after decreasing to the target value of 153kt, increase within a second for15kt (from 153 to 168kt), in 2 next seconds it additionally increased up to 176kt. In such a way the actual speed exceeded the target one (153kt) for more than 20kt.

This overspeed was responded by the F/O at 00:40:49: «Check the speed». It is the overspeed for a considerable value that, most probably, was the reason for the PIC to make decision on go-around. The PIC took the decision right away, called it out to the F/O and similarly was responded immediately:

00:40:49,7 00:40:50,4 CPT (Ok), go around.
00:40:50,500:40:51,1 F/O Go around.

At 00:40:50 the TO/GA mode was activated with the power levers advanced to full thrust.

[Twiglet1]

Is that not the case in the UK? I thought it was actually a requirement under EASA rules (which of course do not apply to FlyDubai). At least when working in a FRMS Safety Action Group in a previous airline all actual fatigue reports were sent on to the authority, however there was a dispute if pre-emptive fatigue reports ("I am concerned this roster will induce fatigue") had to be send on. Reports were actively encouraged, calling in fatigued for a duty then required a report, so there was plenty to work through.
Hi Denti - not in the UK don't forget a fatigue report is a confidential report normally so SMS protocols kick in - the biggest test of FRMS however is the balls of the Chief Pilot to ensure his crew are not "named and shamed" to the DFO and MD - sometimes difficult if a flight is cancelled. CAA get top level stats and invites to FSAG.
In my experience of a mature AOC with FRMS fatigue issues are split down the middle between AOC and crew. AOC are the usual roster issues, stability of changes, Hotels etc and crews poor sleeping habits and lack of education and the C word (and that's not Christmas)
There is something missing though in that whilst the likes of Easyjet can provide loads of data on early and late combo's and DHL UK can provide loads of data on nights the ME airlines operate a mix of early, late and nights (and particularly long nights) and with 70 fatigue reports - sadly no data.

[Twiglet1]

Not being an expert, it seems to me that factors contributing to this accident were attempting a go-around using the unfamiliar, different, monochromatic symbology and presentation of an HUD, instead of the normal ADI, IAS, V/S and N1 gauges - leading to disorientation? Operation with one pilot on a HUD and the other on conventional instruments had not been sufficiently trained, nor allowed to be sufficiently practised - particularly in the case of an F/O taking control from a Captain - both using different instruments - during a high workload situation, such as a go-around, which is going pear shaped.

Add to this, a possibly fatigued crew during their WOCL, and a horrible weather situation. The Captain clearly felt pressure that he had to land at the destination, and also not go out of hours, whereas from one’s armchair it would seem reasonable to divert after the first two landing attempts.
I think I have experienced good company FRMS systems, but allowing airlines to monitor their own fatigue reporting and deal with it in-house could potentially be a case of the foxes looking after the hen house?
If fatigue reports had to be sent to and dealt with by the equivalent CAA, then fatigue would be officially examined and might actually have to change?

Fatigue reports are the same as confidential reports - managed within the SMS. If names get leaked then the Pilots lose trust in the SMS and stop reporting. Again some AOC's have sleep scientists that assess the rosters individually. However an aweful lot of Airlines also have (commercial) bio-mathematical modelling systems latched onto their scheduling systems which they use to convince their CAA's they are managing fatigue e.g. system says green so good to go. Are they validated, do they have data on that particular AOC - you know the answer.

[safetypee]

Oldn, my interpretation of the report was that windshear caused the speed change. The Captain reported that the approach was stable, higher speed than normal, but then “increase within a second for 15kt (from 153 to 168kt)” a change most unlikely to be a commanded change in the thrust setting or vertical speed.

15kt/sec represents a gust, whereas aircraft are more sedate, approx max 3kt/sec, depending on type.

[OldnGrounded]

Oldn, my interpretation of the report was that windshear caused the speed change. The Captain reported that the approach was stable, higher speed than normal, but then “increase within a second for 15kt (from 153 to 168kt)” a change most unlikely to be a commanded change in the thrust setting or vertical speed.

15kt/sec represents a gust, whereas aircraft are more sedate, approx max 3kt/sec, depending on type.

I understand, Safety, and the report does speculate that the PF may have interpreted the reported gust to be evidence of windshear. On the other hand, there was no windshear warning on that second attempt; PF was pushing the nose down a bit and thrust was increased a bit in the moments leading up to the overspeed; the CVR recording doesn't include any evidence that PF told PM they were executing a windshear escape GA; and the PM's offer to retract flaps back to 15 suggests that he didn't think they were flying a windshear escape. I guess we're all guessing.

This is a desperately sad story.

[Double Back]

In "my" days it was promoted to specify the type of wind shear. If during the APP a "negative" shear or performance decreasing shear, was reported, a drop in speed (IAS) could be expected. The confusing about it that the same (TWR) controller had to report this same shear as a "positive" shear to departing A/C.
In this case the arriving A/C should have been told to expect a "positive" shear.
Is this still needed to/practiced?

[Starbear]

Starbear, “- the pilot does in fact have another tool with which to contain the pitch up and that is the reduction of thrust in conjunction with pitch inputs.”

You appear to overlook that the GA was in response to a Windshear alert, which in general (overwhelmingly) requires maximum thrust.
At some point there may be a conflict between reducing thrust as judged by the crew in the actual conditions, and the operators SOP, - cognitive dissonance - mental effort, confusion, distraction.

What do operators teach and mandate by SOP ?
What advice do operators provide for reducing thrust after a Windshear GA ?

No, I don't believe that I did but perhaps did not make myself clear enough. Perhaps I should have included the full quote from his post but I was specifically referencing Centaurus' tale of teaching recovery from a severely out of trim/low speed situation such as the Turkish B737 at Amsterdam and offered the Thomsonfly B737 incident as a parallel to the Turkish one.I was not or not intending to link to the Flydubai incident. There have been quite a few others but am a little hazy on precise details now, without research but a couple of A310s, I think Tarom (?) but again severely out of trim for a variety of reasons. And the modifcations to a stall/approach to stall recovery were changed long before the FlyDubai case.

With regard to reactive windshear recovery and your posed questions, I would suggest that sadly the answer is "not very much" and is often as much use as the manufacturer's instruction to "smoothly adjust the pitch to follow the guidance". However what I can confirm from observation that many pilots continue to maintian an inapproriate high pitch attitude even when well clear of the windshear and/or terrain. Often as high as 20 deg (and more)for a sustained period with 2 or even 3,000 feet terrain clearance and reducing IAS.
They do not appear to have a reference for confiming they are now in a safe situation and able to revert to a normal GA type scenario. They always get there in the end, (so who can knock that?) but it is often very "untidy".

I would not comment on the pilots' reactions in the Dubai case, as I believe the report has already covered that very well.

[Hogger60]

Fatigue reports are the same as confidential reports - managed within the SMS. If names get leaked then the Pilots lose trust in the SMS and stop reporting. Again some AOC's have sleep scientists that assess the rosters individually. However an aweful lot of Airlines also have (commercial) bio-mathematical modelling systems latched onto their scheduling systems which they use to convince their CAA's they are managing fatigue e.g. system says green so good to go. Are they validated, do they have data on that particular AOC - you know the answer.


I agree wholeheartedly. The FRMS software systems that I have experience with are pretty easy to manipulate. The company can make almost any roster appear to be in compliance with their regulator's standards. At most companies only lip service is paid to fatigue. I believe this is the case in the Rostov crash, with brutal rosters playing a huge part in the Captain's mistakes.

[IcePack]

I used to fly for a very well respected UK airline. They wrote a “propaganda” pamphlet to the crews on how wonderful their FRMS system was on making the rosters. Interesting in that I wrote to the author pointing out on release the rosters were fine, but they were never flown as the rosters were always completely changed by day to day crewing. ( to knackering patterns). Funny old thing I never got a response. But bet the pamphlet looked great to the CAA.
Am absolutely certain fatigue played a huge part in this crash.

[safetypee]

Starbear, our interpretations of this report differ, as might be expected by the human condition.

Similarly with other examples; accidents involving human-automation interaction, indications, alerting, and physical interaction with aircraft trim or autopilot. I recall that one of the examples and possibly two others (A310, Avro RJ) involved overpowering the autopilot causing the trim to run. The latter two resulted in aircraft modification, but other manufacturers argued, as you might, that the human should manage systems, where others would see them as weak / flawed designs.

Certification regulations now require improved human - autos interaction, but not retrospectively. [for info, see the AMS 737 report - annex of comments, manufacturer / FAA / investigator. Chilling similarities with recent events.]

A problem with this line of thought is that it might exclude the wider picture, to consider other factors and interactions, which will also differ according to viewpoint, within or after the fact.

Your views on WS GA are useful, thanks. However, with the advent of sophisticated detection systems, differentiating predictive or reactive WS, and possibly excess energy, result in more, and case sensitive alerts and procedures; general or specific SOPs. This adds complexity to operations and training, increasing demands on situation awareness an interpretation. So claimed technical safety improvement, may be diminished by more dependence on stretched human ability in rare and surprising situations.

[787PIC]

If not represented by a strong union, "fatigue reports" are treated by the management as nails into pilot's own coffin!
Try calling fatigued or writing SMS reports, twice in a row at an airline like Fly Dubai. Heck, even at Emirates or Ryanair!
Although this very well prepared crash investigation report paints a perfect picture of a total incapacitation due to pilot pushing and chronic fatigue.

[retired guy]

If not represented by a strong union, "fatigue reports" are treated by the management as nails into pilot's own coffin!
Try calling fatigued or writing SMS reports, twice in a row at an airline like Fly Dubai. Heck, even at Emirates or Ryanair!
Although this very well prepared crash investigation report paints a perfect picture of a total incapacitation due to pilot pushing and chronic fatigue.

Currently a high court case in Dublin Ryanair v Peter Bellew form COO is quoting "a climate of fear among pilots" denied by Mr O' Leary. But it is being mentioned in court. Watching with interest. Court case going on 3 days already and three to go next week.
Cheers

[alf5071h]

From the accident report, when the HUD reverts to upset mode, the mental reference has to change from FPA to attitude; - this requires significant mental readjustment. Also the FD is removed.

This was noted by Fursty Ferret in https://www.pprune.org/showpost.php?p=10630380&postcount=74 (https://www.pprune.org/showpost.php?p=10630380&postcount=74) , and that the “The HUD symbology on both the 737 and 787 is pretty terrible”, but how is this reported to the regulators, considers equipment reassessment or associated SOPs, checks training implications, etc, …

The change of ‘mental model’ - switching from one format to another could be similar to reverting from head down instruments to visual, which has been shown to take 2-4 secs, i.e. a delay in re-establishing sufficient awareness for the task - upset recovery.

Other questions about the HUD; the accident report shows views of the HUD during the approach, it is unclear if these are either a mathematical simulation or from a training simulator.
Assuming that the top-right dial display is AoA, then the adjacent digital values appear not to correspond with the aircraft manoeuvre and speed.
Nose up, fig 65, 1.4. Alternatively in fig 66, -0.6.
Nose down, fig 70, 3.6.
Why should this be ?

And for all those who demand “we must have AoA displayed - upset recovery”; the HUD Upset / de clutter mode removes the AoA display !!!
Anyone able to confirm this via training simulator / real aircraft / HUD.

[Double Back]

Why is it that in the airline industry the HUD seems to be a cheepo version of what the mil guys use? AFAIK their flying relies a lot on HUD, no matter what attitude they are in.
Again, I can't relate to own experience, never flew in mil A/C nor used a civvy HUD.
At least this thread learned me why airlines choose for a HUD to make CAT 1+ landings, iso a dual A/P... They are way cheaper!

[safetypee]

Double Back.
Civil HUDs are not cheap; flight operations are significantly different from military use.
HUD certification can be just as expensive as a fully capable AP, and then involve additional training costs, and currency requirements.

A previous comment that the accident operator traded AP servicing cost for HUD use, appears to show how easy it is to misjudge relative cost and safety.

Other issues relating to HUD operations are discussed in
https://www.pprune.org/tech-log/627839-use-hud-visual-approach-landings.html

[alf5071h]

Section 2.4 of the report discusses the possible effect of somatogravic illusions (page 166) - longitudinal acceleration and false pitch perception.

Discussion in the 737 Max thread raises associated issues of sensed pitch rotation with reduced vertical acceleration (g). The report does consider ‘negative g’ perhaps using a wider definition of illusion as a false sensation, and separately considers issues of pitch control and trim which contribute to the pilots sensation of flight path and aircraft feel - control feedback.

Normal flight conditions range 0.8 to 1.3 g, generally in a steady sate between manoeuvres; greater g values can be encountered for short duration, e.g. turbulence. Aircraft motion (sensed acceleration) relates to control movement and stick force, and ‘seat of the pants’ feedback referenced to general experience.

Abnormal, infrequently encountered conditions, may require larger stick inputs which could contribute to the assessment that additional trim is required. Alternatively the aircraft trim changes quickly due to speed increase, large change in thrust (pitching moment), or configuration change, thus requiring more pilot control input and longer duration trim change.

In pushover manoeuvres the pilot could experience reduced g more than generally experiences (often reported as negative g); in this condition pilots can encounter combined reduction in g, and negative pitch rate simultaneously. Also with the need for nose down trim and large stick defection the flying task is beyond normal expectation which could be interpreted a false sensation - like an illusion - even surprise. Large stick inputs in GA configuration at low speed, differ significantly from normal flight experiences - only requiring small stick input at higher speeds.

Thus during a GA the pilot may be reluctant to use large stick defections to avoid unusual g conditions; or with larger control input quickly seek to reduce the higher stick force with longer periods of trim.
In either case the pilot is experiencing less g than normally experienced, simultaneously with the unusual pitch change and control feel in a very short time scale. The combination could contribute to poor understanding of the aircraft flight condition, which with false sensation reduces the ability to correctly interpret the situation and apply corrective control inputs and trim - particularly where the duration of abnormal control change and sensations are longer than normal experience suggests.

Any thoughts on this view ?

[Gipsy Queen]

In either case the pilot is experiencing less g than normally experienced, simultaneously with the unusual pitch change and control feel in a very short time scale. The combination could contribute to poor understanding of the aircraft flight condition, which with false sensation reduces the ability to correctly interpret the situation and apply corrective control inputs and trim - particularly where the duration of abnormal control change and sensations are longer than normal experience suggests.

Any thoughts on this view ?

I think it probably is necessary to distinguish between vertical accelerations experienced by the aircraft as a whole and those perceived by the pilot as exaggerations created in the lateral axis by the long moment arm between the centre of rotation and the cockpit. But don't ask me how that might be done!

Just an observation rather than a useful thought.

[vilas]

a5701
When learning instrument flying it is repeatedly drilled into you to ignore bodily sensations and feelings and only react to what instruments tell you about the situation. Some where this message seems to be lost. Whether it is AF447, QZ8501, or Rostov OD in all of these large scale pitch inputs are given without any reference to attitude display on PFD or HUD. In the present RoD case despite the PM's excellent support the PF continued to push and trim without checking the result on the attitude. Even the increasing speed should have triggered pitch up and not a pitch down reaction. It appears to be case where the pilot was overwhelmed and mentally incapacitated and the resulting control inputs were without any thought or rationale. The PM was monitoring and verbalizing everything wish he had taken over.

[alf5071h]

Gipsy Queen, thanks for your observation; it is a very useful thought - to avoid the assumption that pilots will experience the same accelerations as recorded by the FDR.

vilas, as much as we train pilots to ‘mask’ illusionary effects by focussing on the instruments, we cannot guarantee success in every situation; not for every pilot, not now nor future.

Thinking about your point “where the pilot was overwhelmed and mentally incapacitated”, this could be related to an effect of surprise. Yet the regulators are now considering methods of training pilots to be ‘resistant’ to surprise, with the belief that basic human reactions can be sufficiently suppressed, again, with what confidence.
The industry really has to accept that there are limits to what can be trained, or what pilots can recall / replicate in today’s complex situations. We must re-examine the situations and equipment in them which have to be managed.

Also, similar issues apply to monitoring and intervention (CRM). Instead of citing human error, failure, more training, it is time to accept that pilots are doing their best in the situation as they understand it. Either the situations are more complex, humans less experienced in understanding, or more likely both.
The modern safety phrase - work as imagined - not being the same as work as done, - as identified in accidents. We need to change the way we think about these issues.

A thought ‘out of the box’: pilots suffer illusions, but are they only consequential during manual flight. With autos engaged there are no adverse effects on motor skills irrespective of wether the pilots believe the instruments or not.

[vilas]

as much as we train pilots to ‘mask’ illusionary effects by focussing on the instruments, we cannot guarantee success in every situation; not for every pilot, not now nor future. The point is although the g forces, sensations are same for both pilots the PM was not affected by them and was guiding the PF correctly. Despite clear cut guidance by PM to maintain the pitch PF wasn't able to do it. Should there be some guideline as to how much deviation the other crew member should accept before physical intervention otherwise we learn nothing from these loss of lives. Also
Thinking about your point “where the pilot was overwhelmed and mentally incapacitated”, this could be related to an effect of surprise. In taking a decision and executing a go around which is a well practiced manoeuvre where is the surprise?

[wiggy]

The point is although the g forces, sensations are same for both pilots the PM was not affected by them and was guiding the PF correctly. Despite clear cut guidance by PM to maintain the pitch PF wasn't able to do it. Should there be some guideline as to how much deviation the other crew member should accept before physical intervention otherwise we learn nothing from these loss of lives.

very very valid point...

The somatogravic illusion isn’t something new, it hasn’t just been invented, the sims actually don’t do a bad job at replicating the same- those of us used to flying some geriatric but over powered military types in a previous existence did exotic manoeuvres at night and/or in IFR and will have experienced the over whelming sensation of feeling they are pitched up at some horrible angle whilst in actuality being S&L and telling themselves. “For f**** sake whatever you do don’t push”...as you correctly said earlier Vila’s it was drummed into us that you flew the instruments, not the seat of your pants..,....so what happened here?

so why didn’t the P1 do so that night on the second go - around..?fatigue.? And is there a case for the guy who is heads heads down in such circumstances (in this case the P2) doing the actual GA (not a HUD user myself).

[Euclideanplane]

The elephant in the room right now seems to be the question, is the onset of or a particular susceptability to somatogravic illusion heavily linked to fatigue issues?

[alf5071h]

Surprise, J Reason; “ a profound discrepancy between one's perception of the world and the reality” (which could include illusions)

Having decided to fly a GA - the second occasion, perhaps not a significant surprise even though it conflicted with what was planned - expected - hoped for.
A larger surprise effect could emerge from the difference between the lighter weight GA and perhaps slower pitch response in following the FD. A difference in ‘the conditioned expectation of normalcy’.
Also increased or ‘new’ surprise from the unexpected, even unknown automatic change of HUD format to ‘upset’, with additional problems in changing mental focus and assessing the flight path within HUD - previous discussions.
The effect of surprise together with the stick feel/position trim characteristics could have contributed to the excessive nose down pitch response and continued trim operation (‘2.2. On the peculiarities of the trim (relief) of forces’: page 159).

If the 737 is normally flown with the combination of small stick movement, quickly followed up with trim ‘blips’ (note Boeing cautions not to use fly the aircraft with trim). Then depending on personal technique it is possible that with the stress of the situation that the trim blips merge into one, which together with slower pitch response than anticipated - slow speed, high nose up attitude, and max GA thrust, the aircraft is inadvertently flown with trim. If the stick is not centred to judge the effectiveness of trim - reduced stick force, then the aircraft continues to pitch nose down with trim still applied to counter the erroneous stick force from pushing the stick forward.

Recent ref for surprise - https://pure.tudelft.nl/portal/files/55707836/dissertation_startle.pdf
Good background on the mechanism and effects of surprise, that it can occur without startle, and build up relatively slowly during a developing situation - several layers of surprise.
However the investigations into training solutions are less convincing - can you ever be really surprise in a simulator - fear. Even if possible can you be sure that everyone has been ‘conditioned’ or will respond appropriately in every ‘surprising’ situation.

Another description of different trim characteristics - https://www.skybrary.aero/bookshelf/books/2627.pdf
Could a 737 operator please confirm that Safety Recommendation 2009-045 (page 4) has be actioned ?

[Arkroyal]

Who actually has the time to read 175 pages?

it’s ok Contact. There are lots of pictures for the children 😂

[fizz57]

From the accident report, when the HUD reverts to upset mode, the mental reference has to change from FPA to attitude; - this requires significant mental readjustment. Also the FD is removed.


An X-15 was lost when the attitude target indicator on the artificial horizon was left in yaw mode for ascent rather than in sideslip mode for re-entry (or was that the other way round?) due to high pilot workload because of other issues. The re-entry was thus flown at the incorrect angle resulting in loss of control and the breaking up of the aircraft. The NASA Dryden historical document I read some time ago says something along the lines of "although not mentioned in the final report, several engineers came away with the message that the same instrument should not be used to display conflicting information at different times".

Seems that some lessons have to be re-learnt every now and then.

[airpasty]

I lost my job in flydubai for banging the drum about fatigue, the culture there was toxic. This report is incredibly disrespectful to the flight crew who where leaving the company due to the culture and fatigue. People especially the cabin crew would be called into meetings for refusing discretion, calling in fatigued and calling in sick.

Im so glad to be out of there. The report should have investigated the culture of fear. When we had a meeting post crash, the head of OPS who is now running the show said “so we had a crash, crashes happen all the time, get over it” when we asked for a minutes silence in the same meeting to remember those who died the chief pilot started talking 20 seconds into it. Abhorrent.

Despite promising to take care of the crews families, to this day I speak with some of the the crews family members, some are still struggling financially. They where given some money after the crash but only if they signed a piece of paper saying that they wouldn’t sue the company later on when the report came out. Profit over people time and again.