AF 447 Thread No. 8
Moderator
Thread Starter
.. as a sideline, should anyone be wondering why I haven't started the next thread, I'm leaving it until the report comes out .. as I suspect that the posts will be coming fast and furious for several weeks after that ..
Join Date: Jul 2009
Location: Not far from a big Lake
Age: 82
Posts: 1,454
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by HN39
In my search for a correlation between side stick angle, pitch rate and normal acceleration, based on the C* control law, I started comparing the elevator positions to the side stick commands, and am surprised to find that the elevator seems to be leading the side stick:
Join Date: Aug 2005
Location: fl
Posts: 2,525
Likes: 0
Received 0 Likes
on
0 Posts
I have never used a HUD so don't know if it would have helped them or not. Attitude indicators, altimiters, Vertical speed indicators and stall warnings didn't, so doubt if HUD would have helped either. They just needed an experienced pilot flying, not them. Sorry. That is how I see it. Monitoring an autopilot for thousands of hours does not make you a pilot.
Join Date: Aug 2011
Location: Grassy Valley
Posts: 2,074
Likes: 0
Received 0 Likes
on
0 Posts
HN39
In the first extreme, the stick follows acquired angle of elevator by eight seconds?
What is wit that?
Mach... Or a lead in the elevator data? If it is real, what could cause such a discrepancy?
Was the a/c hunting, the pilot chasing? Sounds like an abnormal.
In the first extreme, the stick follows acquired angle of elevator by eight seconds?
What is wit that?
Mach... Or a lead in the elevator data? If it is real, what could cause such a discrepancy?
Was the a/c hunting, the pilot chasing? Sounds like an abnormal.
Last edited by Lyman; 29th Jun 2012 at 03:01.
Join Date: Oct 2011
Location: Lower Skunk Cabbageland, WA
Age: 74
Posts: 354
Likes: 0
Received 0 Likes
on
0 Posts
In resp. to mm43
Thanks for the link to the Desdemona simulator project. I noted a link hidden in the article leads to an EU Research & Innovation page that includes a comprehensive write-up and video. Some may be interested.
Mr. 43,
(Since nobody has remarked on this)-- Thanks, that was superb. The soundtrack of the video is the same as the article text. A large download people, but highly recommended! The new simulator really is a centrifuge.
Stupendous.
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Machinbird
That sounds like a data reduction problem with the original data to me. Something likely caused a delay artifact in the side stick position data.
Join Date: Jul 2002
Location: UK
Posts: 3,093
Likes: 0
Received 0 Likes
on
0 Posts
Join Date: Oct 2009
Location: berlin
Posts: 152
Likes: 0
Received 0 Likes
on
0 Posts
HN In the simulation also?
the time delay can and will be a non accurate start-time in the simulation or in the transfer from the simulation to the report,
they searched for bigger effects than you
Last edited by grity; 29th Jun 2012 at 13:23.
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes
on
0 Posts
I don't get it. If you do a simulation with the side stick as input and the elevator as output, how can the elevator get ahead of the side stick? If there was a delay artifact in the elevator or side stick position data, wouldn't the simulation have revealed it?
Yes, I'm looking forward to seeing that explained in the final report.
Yes, I'm looking forward to seeing that explained in the final report.
Last edited by HazelNuts39; 29th Jun 2012 at 14:37.
In one specific hypothetical case, over a given period of time nose high somewhere in that 6-15 degree pitch range, airspeed decaying, if you were to graph a hands-off 'neutral' sidestick, would that not be a 0 degree deflection integrated over a particular time period (flat line)?
But you would end up with a constantly increasing value of nose up elevator deflection over time attempting to hold the FP which I guess in a way could be considered as 'elevator leading sidestick'.
Now the SS is physically moved nose-up...
The same existing pitch rate (=elevator position) required for airspeed decay and in effect for the flight path is continually input over a range of increasing SS angles as the deflected elevator holds the FP until the SS rate command catches up? Would this give the appearance of elevator leading SS?
This may not be what you're getting at.
But you would end up with a constantly increasing value of nose up elevator deflection over time attempting to hold the FP which I guess in a way could be considered as 'elevator leading sidestick'.
Now the SS is physically moved nose-up...
The same existing pitch rate (=elevator position) required for airspeed decay and in effect for the flight path is continually input over a range of increasing SS angles as the deflected elevator holds the FP until the SS rate command catches up? Would this give the appearance of elevator leading SS?
This may not be what you're getting at.
Last edited by OK465; 29th Jun 2012 at 15:33. Reason: added 'rate command'
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes
on
0 Posts
grity,
No, it's the blue line on page 42, section 1.16.5.1.
With thanks to free Engauge Digitizer software.
No, it's the blue line on page 42, section 1.16.5.1.
With thanks to free Engauge Digitizer software.
Last edited by HazelNuts39; 29th Jun 2012 at 16:20.
Join Date: Oct 2009
Location: berlin
Posts: 152
Likes: 0
Received 0 Likes
on
0 Posts
HN
"Airbus conducted a simulation of the aircraft behaviour based on
the theoretical model and on the actions of the PF (sidestick and thrust)" (s.41)
"2 h 10 min 23 The N1 start to increase and reach around 104 % in 12 seconds." (s.90) and (s.113)
is there any automatismus for the elevator to reakt against the moment (--->pull) of the increasing engines ???
"Airbus conducted a simulation of the aircraft behaviour based on
the theoretical model and on the actions of the PF (sidestick and thrust)" (s.41)
"2 h 10 min 23 The N1 start to increase and reach around 104 % in 12 seconds." (s.90) and (s.113)
is there any automatismus for the elevator to reakt against the moment (--->pull) of the increasing engines ???
Join Date: Jun 2009
Location: somewhere
Posts: 451
Likes: 0
Received 0 Likes
on
0 Posts
Final AF447 report set to re-ignite safety feud
French investigation authority BEA appears likely to rekindle the smouldering conflict over the loss of Air France flight AF447 when it releases the final report into the accident on 5 July.
Air France and the main French pilots' union, SNPL, have previously clashed with Airbus over the circumstances of the crash and whether the fundamental reason for the loss centred on pilot competence or the design of the Airbus A330's flight-control and warning systems.
As expected, Airbus has submitted substantial comment to the inquiry following the release of the draft version of the report.
SNPL has already reiterated its concerns over aircraft functions and the alerts given to the crew, in a document published in February.
Airbus declines to comment on the AF447 report ahead of publication, but has previously strongly defended its aircraft and pointed out that three pilots appeared unable – despite clear warnings – to recognise the aerodynamic stall which downed the jet in June 2009.
Although the draft did not include recommendations it featured analysis by a human factors panel established to look into the crew's response to the stall.
But a source familiar with the situation indicates that the airframer is concerned whether the conclusions will focus too narrowly on the human-machine interface.
"If there are things to improve on the aircraft, [Airbus] won't try to escape in any way," says the source, but adds that the manufacturer “would like to see a report in which all the issues are being dealt with”.
The source also suggests that Airbus is likely to "become vocal" if it feels the breadth of the report is too narrow.
Former BEA deputy chief Jean Pariès - who heads human factors consultancy Dédale and took part in an Air France safety review - told an operations forum in Oslo in April that current safety models assume pilots will recognise and identify abnormal situations, then implement relevant procedures.
However in reality, he said, emergency situations generate surprise, causing momentary loss of cognitive control as well as resistance to recognising a loss of comprehension.
Pariès cited 16 events similar to AF447, all of which showed poor understanding, rare implementation of unreliable airspeed procedures and stall warnings which were "perceived but mostly not believed".
He suggests the problem cannot simply be reduced to "automation complacency" or loss of basic skills. Pariès claims crew training aims to prepare pilots for anticipated emergencies, not the unexpected, and highlights the irony that the competencies needed to cope with the unexpected "are those that are lost in a continuous effort to anticipate and respond to all potential threats".
Investigations into a strikingly similar event to AF447, involving an Air France A340 in July 2011, recommended that pilot training include shock and surprise elements.
Air France and the main French pilots' union, SNPL, have previously clashed with Airbus over the circumstances of the crash and whether the fundamental reason for the loss centred on pilot competence or the design of the Airbus A330's flight-control and warning systems.
As expected, Airbus has submitted substantial comment to the inquiry following the release of the draft version of the report.
SNPL has already reiterated its concerns over aircraft functions and the alerts given to the crew, in a document published in February.
Airbus declines to comment on the AF447 report ahead of publication, but has previously strongly defended its aircraft and pointed out that three pilots appeared unable – despite clear warnings – to recognise the aerodynamic stall which downed the jet in June 2009.
Although the draft did not include recommendations it featured analysis by a human factors panel established to look into the crew's response to the stall.
But a source familiar with the situation indicates that the airframer is concerned whether the conclusions will focus too narrowly on the human-machine interface.
"If there are things to improve on the aircraft, [Airbus] won't try to escape in any way," says the source, but adds that the manufacturer “would like to see a report in which all the issues are being dealt with”.
The source also suggests that Airbus is likely to "become vocal" if it feels the breadth of the report is too narrow.
Former BEA deputy chief Jean Pariès - who heads human factors consultancy Dédale and took part in an Air France safety review - told an operations forum in Oslo in April that current safety models assume pilots will recognise and identify abnormal situations, then implement relevant procedures.
However in reality, he said, emergency situations generate surprise, causing momentary loss of cognitive control as well as resistance to recognising a loss of comprehension.
Pariès cited 16 events similar to AF447, all of which showed poor understanding, rare implementation of unreliable airspeed procedures and stall warnings which were "perceived but mostly not believed".
He suggests the problem cannot simply be reduced to "automation complacency" or loss of basic skills. Pariès claims crew training aims to prepare pilots for anticipated emergencies, not the unexpected, and highlights the irony that the competencies needed to cope with the unexpected "are those that are lost in a continuous effort to anticipate and respond to all potential threats".
Investigations into a strikingly similar event to AF447, involving an Air France A340 in July 2011, recommended that pilot training include shock and surprise elements.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
Shock and surprise are virtually impossible in a simulation cession ... unless you make suddenly appear on top of the dashboard .. a tarantula or a poisonous snake or even a mouse when female pilots on training ...
Investigations into a strikingly similar event to AF447, involving an Air France A340 in July 2011, recommended that pilot training include shock and surprise elements.
Last edited by jcjeant; 29th Jun 2012 at 18:49.
Join Date: Dec 2010
Location: Middle America
Age: 84
Posts: 1,167
Likes: 0
Received 0 Likes
on
0 Posts
Hi HN39,
I was reading an old post by Chris Scott (#1060 7/9/11 AF447 Thread #4)
I wonder if this could explain your graph findings of the elevator leading the SS?
I was reading an old post by Chris Scott (#1060 7/9/11 AF447 Thread #4)
I don’t think much would happen to the THS during the initial rotation from level flight, as it would have required little up-elevator to enter the climb. Once the 7000ft/min had been achieved, the trajectory would be maintained by the EFCS even with no back-stick. As the speed started to drop, more up-elevator would need to be introduced by the EFCS to maintain 1g, and it would then start to trim the THS a bit to retain full elevator authority. Once the aircraft got on to the back end of the drag curve, however, this process would proceed rapidly.
In fact the PF arrested most of the climb at FL375, by “nose down control inputs...”, so some down-elevator may have been used briefly by the EFCS at this point. This partial recovery was to be short-lived, as the PF seems to have reacted to a stall warning by selecting TOGA (causing a pitch-up) and resuming “nose-up inputs”. Resuming the climb at that point, on the wrong side of the drag curve, was when the EFCS had to start using up-elevator and nose-up trim (THS) in earnest, in order to try and maintain the trajectory it thought the PF wanted.
In fact the PF arrested most of the climb at FL375, by “nose down control inputs...”, so some down-elevator may have been used briefly by the EFCS at this point. This partial recovery was to be short-lived, as the PF seems to have reacted to a stall warning by selecting TOGA (causing a pitch-up) and resuming “nose-up inputs”. Resuming the climb at that point, on the wrong side of the drag curve, was when the EFCS had to start using up-elevator and nose-up trim (THS) in earnest, in order to try and maintain the trajectory it thought the PF wanted.
Last edited by Turbine D; 29th Jun 2012 at 19:37. Reason: Spelling correction
Join Date: Dec 2010
Location: Middle America
Age: 84
Posts: 1,167
Likes: 0
Received 0 Likes
on
0 Posts
Hi jcjeant,
Actually, I have a family member who is working on the development and installation of two new nuclear reactors for electrical power generation here in the US. In order to take in account human psychological factors when things go wrong, they have a computer program they follow as it relates to S & S.
She calls it Human Risk Analysis, and they use an HRA calculator from EPRI as the tool. They go through plant operating procedures and identify human actions that could be a pre-initiator to an event or a post-initiator that could inhibit mitigating systems. They also identify operator actions necessary to mitigate an accident. They can quantify Human Error Probability (HEP) with the HRA calculator knowing various environmental factors, the approximated stress level of the operator/operators, the complexity of the procedures that must be followed, and the time window available to perform the action. All of these inputs are combined using various statistical methods in the HRA calculator, resulting in in a human error probability. If the analysis depicts a high HEP, the controls, control panels and procedures are reviewed and reworked to lower the HEP within acceptable levels.
She thinks that responses to aircraft accident scenarios are also highly proceduralized and imagines that similar methods could also be used in the aviation industry.
Just some food for thought....
Shock and surprise are virtually impossible in a simulation cession ... unless you make suddenly appear on top of the dashboard .. a tarantula or a poisonous snake or even a mouse when female pilots on training ...
Actually, I have a family member who is working on the development and installation of two new nuclear reactors for electrical power generation here in the US. In order to take in account human psychological factors when things go wrong, they have a computer program they follow as it relates to S & S.
She calls it Human Risk Analysis, and they use an HRA calculator from EPRI as the tool. They go through plant operating procedures and identify human actions that could be a pre-initiator to an event or a post-initiator that could inhibit mitigating systems. They also identify operator actions necessary to mitigate an accident. They can quantify Human Error Probability (HEP) with the HRA calculator knowing various environmental factors, the approximated stress level of the operator/operators, the complexity of the procedures that must be followed, and the time window available to perform the action. All of these inputs are combined using various statistical methods in the HRA calculator, resulting in in a human error probability. If the analysis depicts a high HEP, the controls, control panels and procedures are reviewed and reworked to lower the HEP within acceptable levels.
She thinks that responses to aircraft accident scenarios are also highly proceduralized and imagines that similar methods could also be used in the aviation industry.
Just some food for thought....
Join Date: Sep 2009
Location: Around the World
Age: 74
Posts: 87
Likes: 0
Received 0 Likes
on
0 Posts
Things that have never happened before happen all the time
#1453
Some details here
Former BEA deputy chief Jean Paries - who heads human factors consultancy Dedale and took part in an Air France safety review - told an operations forum in Oslo in April that current safety models assume pilots will recognise and identify abnormal situations, then implement relevant procedures.
However in reality, he said, emergency situations generate surprise, causing momentary loss of cognitive control as well as resistance to recognising a loss of comprehension.
However in reality, he said, emergency situations generate surprise, causing momentary loss of cognitive control as well as resistance to recognising a loss of comprehension.
Some details here
Last edited by NeoFit; 29th Jun 2012 at 21:50. Reason: title
Join Date: Jun 2011
Location: West of Offa's dyke
Age: 88
Posts: 476
Likes: 0
Received 0 Likes
on
0 Posts
Hi HN39
I think TurbineD has it right. Surely the whole point of the C* law is that there is no direct relationship between sidestick and elevator deflections? (and therefore no correlation of their integrals either).
A sidestick movement commands a 'g' level and the control system moves the elevators as much as it thinks is needed to fulfill that command. If the stick is held at a constant position the elevators will still move until that 'g' level is acheived - i.e. they will 'lead' the sidestick. If the 'g' is not what the pilot wanted then he would add more stick and the process would continue. Seems to me that is what was happening here.
In my search for a correlation between side stick angle, pitch rate and normal acceleration, based on the C* control law, I started comparing the elevator positions to the side stick commands, and am surprised to find that the elevator seems to be leading the side stick:
A sidestick movement commands a 'g' level and the control system moves the elevators as much as it thinks is needed to fulfill that command. If the stick is held at a constant position the elevators will still move until that 'g' level is acheived - i.e. they will 'lead' the sidestick. If the 'g' is not what the pilot wanted then he would add more stick and the process would continue. Seems to me that is what was happening here.