Kaiarahi From Canada, joined Jul 2009, 3209 posts, RR: 38 Posted (3 years 4 months 3 days 10 hours ago) and read 3498 times:
Since the latest AF447 thread in Commercial and Civil Aviation has been hijacked and in any event has almost 250 posts, I am starting this thread in the hope that it will be an evidence-based discussion. To that end, I would respectfully request that posters read the existing facts and evidence as established in the BEA reports, which can be found in English here:
* Mandala499 has also created an excellent spreadsheet integrating the FDR and CVR data from the BEA reports which can be found here http://bit.ly/uWMF08
In the last two threads, there was a general consensus amongst informed posters that human factors lie at the root of the accident. In September, 2011, the BEA established a "human factors" working group whose mandate is to analyze:
- Crew actions and reactions during the last three phases of the flight described in the third Interim Report, in particular in relation to the stall warning;
- Cockpit ergonomics;
- Man-machine interfaces.
The working group is expected to complete its work by the end of December. BEA has not indicated if the working group's findings and recommendations will then be made public or if they will only be incorporated into the final report, which is expected to be made public in the first half of 2012.
Kaiarahi From Canada, joined Jul 2009, 3209 posts, RR: 38
Reply 1, posted (3 years 4 months 3 days 7 hours ago) and read 3445 times:
Just to get the discussion going, I'm copying a few posts from the other thread.
Yes, this is the simple irony of it all... the cause of the crash could be as simple as doing the wrong pitch and power, followed by the wrong stall recovery procedure, leading to task saturation rendering them unable to recognize the situation and perform a correct recovery post entering stall.
Wrong initial reaction to a problem, to which not getting the anticipated reaction increases stress and workload, and opens the door to wider erroneous reaction on every subsequent warning/failure indication. In other words, classical task saturation, reduced situational awareness, and subsequent spatial disorientation. The reason is simple, they did not do the task sharing (part of the golden rule) properly!
I don't know if it will be just one reason, as human dynamics and interaction is a complex web, but I'm still thinking back to the composition of the working group looking at the human factors. This working group is made up of seven experts:
* Three BEA investigators specializing in human factors;
* A psychiatrist specializing in risk analysis;
* A human factors aviation consultant;
* A type-rated A330 pilot;
* An A330 test pilot."
I find the inclusion of a psychiatrist "specializing in risk analysis", rather than a psychologist, interesting. Psychiatry is a medical specialty focused on the diagnosis and treatment of mental disorders and pathologies. Psychology is the science of behavioural and mental processes. Obviously there's cross-over, but psychologists focus on why individuals or groups do things or think in a certain way (behavioural / cognitive science), while psychiatrists focus on mental disorder.
Are they looking at the risk that people with certain behavioural or cognitive *disorders* will act/react in a certain way? Are they looking at the risk that flight crew intake profiling and training does not disclose certain relevant *disorders*? All speculation, obviously, but the inclusion of a medical specialist rather than a behavioural/cognitive scientist could be indicative.
From my experience, it would be useless for someone to try to understand that if they're adamant that human dynamics, communication and interaction in a multi-crew environment, has nothing to do with the multi-crew environment itself. It is also futile to persuade anyone to read and understand those if that person refuses to treat the problem without also understanding (or at least acknowledging) the differences between standard and/or static environment decision making process and control-monitor flow process, and aviation and/or dynamic environment decision making process and control-monitor flow process, because the sequence of logical gateways and sequences of logical gateways are not understood, or refused to be understood.
canoecarrier From United States of America, joined Feb 2004, 2844 posts, RR: 12
Reply 2, posted (3 years 4 months 3 days 2 hours ago) and read 3380 times:
Quoting Kaiarahi (Reply 1): The reason is simple, they did not do the task sharing (part of the golden rule) properly!
I think it's evident from the CVR that they didn't share any of the tasks. I keep coming back to the PF taking priority on the joystick a few seconds after the PNF said he had the airplane. And, even when the PM was in the cockpit, from what we know, he didn't really contribute to solving the situation they were in.
Even if they did fly the correct pitch and power (for the wrong situation), that doesn't answer why the PF climbed 2k feet after the AP disconnected. That only stopped when the PNF told the PF to stop climbing, but at that point they lost speed when they gained altitude, and then exited the flight envelope.
Quoting Kaiarahi (Reply 1): I find the inclusion of a psychiatrist "specializing in risk analysis", rather than a psychologist, interesting. Psychiatry is a medical specialty focused on the diagnosis and treatment of mental disorders and pathologies. Psychology is the science of behavioural and mental processes.
I know you've mentioned this a couple times. Is there a cultural difference between what we consider a psychiatrist and what the French consider a psychiatrist? Something else we've talked about that may be pertinent to the discussion, what kind of relationship did the PM and the PNF have? I know you can't answer that or my next question, but I'm sure BEA is trying to understand that by including the human factors working group in the investigation. Why would he pass command off to the most junior flight crew member when he left the cockpit?
I think we can understand all the mechanical/aerodynamic reasons for this accident, but trying to get into someone's mind that's not with us anymore so you can understand why they reacted the way they did is the most difficult task this team has ahead of them.
Starlionblue From Greenland, joined Feb 2004, 17266 posts, RR: 67
Reply 3, posted (3 years 4 months 2 days 20 hours ago) and read 3332 times:
Quoting canoecarrier (Reply 2): Quoting Kaiarahi (Reply 1):
I find the inclusion of a psychiatrist "specializing in risk analysis", rather than a psychologist, interesting. Psychiatry is a medical specialty focused on the diagnosis and treatment of mental disorders and pathologies. Psychology is the science of behavioural and mental processes.
I know you've mentioned this a couple times. Is there a cultural difference between what we consider a psychiatrist and what the French consider a psychiatrist?
Actually Kairahi's description works on both sides of the Atlantic. While there is overlap, psychiatrists are licensed medical physicians while psychologists are behavioral specialists. In general, psychiatrists use medical models and rely on medication while psychologists use psychotherapy and rely on behavioral modeling and modification.
I do understand Kairahi's question. Since as far as anyone knows there was nothing medically wrong with the crew, a psychologist, in other words a behavioral specialist, would seem like a more logical choice.
"There are no stupid questions, but there are a lot of inquisitive idiots."
A post on an earlier thread by someone who knew both said the Captain was not an authoritative leader, and that he was apparently awed by the PNF number of hours in the A330. The Captain deferred to the PNF / PNF experience on flying decisions in the past.
Quoting Starlionblue (Reply 3): a psychologist, in other words a behavioral specialist, would seem like a more logical choice.
But the team already has three behavioral specialists on the team - the BEA investigators - and a fourth behavioral person - the human factors consultant.
Human factors in aviation investigation is a psychological function, and most people who specialize in that area have psychological training, some are licensed psychologists.
Kaiarahi From Canada, joined Jul 2009, 3209 posts, RR: 38
Reply 5, posted (3 years 4 months 2 days 13 hours ago) and read 3242 times:
Quoting rfields5421 (Reply 4): A post on an earlier thread by someone who knew both said the Captain was not an authoritative leader, and that he was apparently awed by the PNF number of hours in the A330. The Captain deferred to the PNF / PNF experience on flying decisions in the past.
Which would suggest that there was something going on here between the crew:
1. The captain passed off command to the (much) less experienced F/O.
2. When the F/O PNF took priority, the PF (who was nominally in command) took it back again. The PNF also faced the hurdle of never having been trained to fly from the right seat.
3. When the captain re-entered the cockpit, the much more experienced PNF did NOT brief him (unless that part of the CVR has not been released for some reason, but there doesn't seem to be a significant gap).
Note also in terms of psychological factors that the P/F's wife was a pax.
WingedMigrator From United States of America, joined Oct 2005, 2275 posts, RR: 56
Reply 6, posted (3 years 4 months 2 days 12 hours ago) and read 3228 times:
The evidence also shows that this crew inadvertently discovered that an airliner could be flown at an angle of attack of 45 degrees. Before the FDR was recovered, few would have agreed this was even possible-- least of all professional airline pilots such as the three unfortunate souls trying to make sense of their instruments that night.
mandala499 From Indonesia, joined Aug 2001, 7010 posts, RR: 77
Reply 8, posted (3 years 4 months 2 days 11 hours ago) and read 3193 times:
Quoting canoecarrier (Reply 2): Even if they did fly the correct pitch and power (for the wrong situation), that doesn't answer why the PF climbed 2k feet after the AP disconnected. That only stopped when the PNF told the PF to stop climbing, but at that point they lost speed when they gained altitude, and then exited the flight envelope.
I think the 2000ft climb can be explained by picking the wrong pitch and power.
Maintaining level for pitch and power would be (examples taken from RR Trent 772 numbers, therefore to be used for description purposes only):
FL200-370: 260kts Abv190T: 3.5/70.9%, 160-190T: 2.5/67.7%, Blw160T: 2.0/66.1%
AbvFL270: M0.80 Abv190T: 3.5/79.5%, 160-190T: 3.0/76.7%, Blw160T: 2.5/73.1%
The above are not memory item actions... the memory items are for "if the safe conduct of the flight is impacted", the pitch/power values are:
- Below Thrust Reduction Altitude: 15deg/TOGA
- Above Thrust Reduction Altitude and below FL100: 10deg/CLB
- Above Thrust Reduction Altitude and above FL100: 5deg/CLB
So, these guys "doubted" safe flight unless they did something... but which value did they pick?
5deg/CLB would put the aircraft in a a shallow climb...
10deg/CLB would put them in a much bigger climb because it's not what they were supposed to do.
But it appears that the PF was targetting 10deg/CLB... (with the late thrust to CLB application).
Then when the stall warning came (the 54s worth of warnings, not the 2 short warnings previously)... The correct procedure was to select TOGA and reduce pitch (wings level, speedbrake check retracted) until the warning goes away...
Instead, the PF did the "stall warning at lift off" procedure, which is 12.5deg/TOGA.
Quoting Kaiarahi (Reply 5): Which would suggest that there was something going on here that had the captain pass off command to the (much) less experienced F/O. And then when the captain re-entered the cockpit, the much more experienced F/O PNF did NOT brief him (unless that part of the CVR has not been released for some reason, but there doesn't seem to be a significant gap).
There doesn't seem to be any significant gap... What I would like to know is, what are people's guesses (and explanations) to:
2:11:51, PF: What's happening? I don't know I don't know what's happening
2:11:52, CP: So here take take that
2:11:58, PF: I have a problem it's that I no longer have vertical speed
2:11:5X, CP: OK
2:11:5X, PF: I no longer have any indication
- "So here take that" by the Captain, what would he be pointing to? Speed? Pitch? Altitude? I don't think it's the VSI...
Uuurgh... research into this is still a long way from completion for me... On the behaviour and control-action-reaction, I only got so far as 2:11:42...
Further complications with even looking at all the data available, is why was the Thrust Levers put from TOGA to MCT at 2:11:43, and then to idle at 2:11:47?
OK, nose was falling from 2:11:47 (15deg pitch) to 0deg pitch at 2:12:15...
The sequence of actions during that period leads me to suspect that they were so preoccupied in wondering "why are we not getting the correct reaction to the TOGA/12.5pitch", that task saturation isolated their minds from thinking "then why is our nose falling?".
I suspect that the crew did not fully trust whatever speed indications were given to them beforehand since the A/P disconnect.. and therefore did not realize the speeds that were given to them from 2:11:07 to 2:11:45 was correct. Therefore when the nose went down, they thought, "oh no, we're going to overspeed now!"
What's interesting is that I do suspect everyone in the cockpit was focusing on "get the wings level"... PF did it, CPT commented it, and PM also.
It was only after wing roll oscillation stabilized between 10L and 20R (2:12:19 until 2:13:40) were they able to think beyond the roll aspect of the aircraft... 2:12:33 showed indications that they were all trying to think through various parts of the stall reaction procedure, but not as a team.
CP was focused on "wings level", PM was focused on "get the nose down", and PF was trying to listen to the other two... "OK, now wings are level, so, nose down?"
There are some very interesting possible explanations to why he pulled the nose up (2:13:04) after a period of nose down... but this has to wait as it needs preparation, especially on how to best present the information (I'm thinking if snapshots of the PFD during this sequence would help... but would be time consuming to construct).
When losing situational awareness, pray Cumulus Granitus isn't nearby !
2:11:45: 32 degree right bank and increasing, stall warning stops, PF already thinks they're overspeeding - fear of spiral dive?
PF at 2:11:58: "I have a problem, I have no VSI" and then "Now I don't have any indications" (the meaning of this is ambiguous).
PF at 2:12:04: "I feel/think we're extremely overspeed, what do you think" and extends the spoilers.
Could he have remained fixated on a spiral dive, so that he was focused on getting the wings level, reducing thrust and spped, and pitch-up (all despite the stall warning, which nobody mentioned). In that case, he would tend to build "impressions" that matched his diagnosis.
Thrust Levers were put to CLB at 2:12:10 - but by whom?
Quoting mandala499 (Reply 8): "So here take that" by the Captain, what would he be pointing to? Speed? Pitch? Altitude? I don't think it's the VSI...
Back-up attitude? If he were referring to something on the PFD, it's strange he didn't name it because he'd have to lean a long way over the console to make it clear what he was pointing to.
Which leads me to a hypothesis that needs a lot more testing.
The FO/PNF believed they were stalled, but never called it out (maybe thinking it was obvious from the stall warning)
The FO/PF believed they were in a dive (panic, relying on "impressions" of overspeed, eventually calling out rapidly descending altitude - perhaps not believing that a stall could result in the alti unwinding so fast).
Whenever the FO/PNF took priority and commanded nose down, the PF took back priority and commanded nose up.
The captain never called out a diagnosis.
In summary, two FOs with opposed assumptions of what was happening, neither of which was communicated, and a captain who was reactive and "behind the curve".