AF 447 report out
Join Date: Jul 2012
Location: California
Posts: 6
Likes: 0
Received 0 Likes
on
0 Posts
Clandestino wrote...
Let's assume that the 1 out of 37 failure data that you quote was a true experiment (it wasn't, it's an observational quasi-experimental design, but let's ignore that for now).
1 out of 37 is about 2.7%, meaning that based on this data, roughly 3% of crews in identical situations would crash the plane.
Of course, we know from basic statistics that small sample sizes are problematic. Plugging the #s into a confidence interval calculator gives us a margin of error of roughly +/- 7%, at the 99% confidence level.
In plain English, with a sample size of 37, we can only say with confidence that actual rate of crashing in this situation will be between 0% and 10%.
Care to explain your "Bad Science" comment? If you don't have doctoral-level statistics knowledge and don't understand this, I'd be happy to explain in more detail?
I think your point that "the study...has already been done..." is valid, only if you accept that results between 0% and 10% are acceptably precise. Do you?
Soylentgreen wrote...
Those who have been paying attention know that between Nov 12 2003 and Aug 07 2009, there were 37 recorded cases of unreliable airspeed on A330/340 worldwide. 36 of them ended without damage to aircraft or injury to anyone. One ended up in airframe write-off and death of all on board.
Study you proposed has already been done. Results are in the final report. Your notion that:
...is not confirmed.
You are welcome, even as I liked it as a very good example of bad science.
Given an identical situation, what % of professional pilots (or perhaps '3 man groups of pilots') would flub it and crash the plane?
Study you proposed has already been done. Results are in the final report. Your notion that:
A big part of this was the human-machine interface, which did an extremely poor job of letting the pilots know what was actually going on.
Thanks to those who liked my post
1 out of 37 is about 2.7%, meaning that based on this data, roughly 3% of crews in identical situations would crash the plane.
Of course, we know from basic statistics that small sample sizes are problematic. Plugging the #s into a confidence interval calculator gives us a margin of error of roughly +/- 7%, at the 99% confidence level.
In plain English, with a sample size of 37, we can only say with confidence that actual rate of crashing in this situation will be between 0% and 10%.
Care to explain your "Bad Science" comment? If you don't have doctoral-level statistics knowledge and don't understand this, I'd be happy to explain in more detail?
I think your point that "the study...has already been done..." is valid, only if you accept that results between 0% and 10% are acceptably precise. Do you?
Last edited by soylentgreen; 9th Jul 2012 at 19:54.
For anyone interested:
If you go to the tech log forum, you will find the following point excerpted from the BEA report.
The UAS (Unreliable AirSpeed) procedure memory item of turning off the FD's was not accomplished. (FD = Flight Director).
There are even some handy pictures and graphs that show when the FD was receiving good data, and when it wasn't. Mind you, neither pilot was aware of when it was good and when it was bad, but given the procedure as stated, they should not have needed to.
This opens the question: was the PF using FD as a reference, or a primary reference for his instrument scan, even though it was not being reliably sourced by the data it is usually fed?
The reasons behind that might be habit, might be training, might be a lot of things. One of the things CVR's and FDR's don't do is read minds, so one can only guess at that. I think he at various times tried to follow the FDR. I am not on perfectly solid ground in that estimation.
For Sassypilots wife:
The Tech Log discussion features no small number of pilots who post there: gums, Machinbird, PJ2, Retired F4, CONFiture, Clandestino, bubbers44, and numerous others. I discuss things there now and again, though I have not flown much of anything in a few years, and never flew heavy metal ever. I have flown and taught flying in multipiloted aircraft, taught instruments, and instructed in simulators where you try to work a crew very hard to see how strong their systems knowledge and airmanship are. I have also investigated mishaps, including mishaps with fatalities.
I would point out that Tech Log is now where ALL of AF447 threads resides, which John T has kindly explained in the preface to the various threads.
The first three IIRC started on the R & N forum.
Rockhound
Not a good idea. See below.
Not so much. He missed a fundamental detail: the difference between pitch angle and angle of attack.
Not quite.
According to the FDR, the nose wasn't 40 degrees up, ever. It was somewhere around 15 degrees up, more or less, at its peak. The AoA, while there was more airspeed, began somewhere between 4 and 6 degrees as the aircraft neared and entered stall (it seems to have been a dynamic event so maybe stall AoA was passed a few degrees higher). Note that even with a roughly fixed pitch, (more or less 15 degrees) the plane slowed AND the AoA increased as airspeed decreased. This should not surprise any pilot who has done stall training.
Based on the analysis of the pitch angle and airspeed, AoA got upwards of 40 degrees as the plane went more deeply into the stall.
Sooo, Rockhound .... maybe that pilot you referenced ought to make sure he understands the problem before he suggests a solution.
Case One
He was explaining how to answer a multiple choice test on flight physics.
I laughed.
I have found the tech log discussions, though full of some chaffe, to have a lot of good nuggets.
I find this thread, after three years of learning about AB, AB330, FBW, sidesticks, control laws, and a few other things, somewhat entertaining.
Mind you, it is also filled with chaffe.
Fly safe. Fly the plane.
If you go to the tech log forum, you will find the following point excerpted from the BEA report.
The UAS (Unreliable AirSpeed) procedure memory item of turning off the FD's was not accomplished. (FD = Flight Director).
There are even some handy pictures and graphs that show when the FD was receiving good data, and when it wasn't. Mind you, neither pilot was aware of when it was good and when it was bad, but given the procedure as stated, they should not have needed to.
This opens the question: was the PF using FD as a reference, or a primary reference for his instrument scan, even though it was not being reliably sourced by the data it is usually fed?
The reasons behind that might be habit, might be training, might be a lot of things. One of the things CVR's and FDR's don't do is read minds, so one can only guess at that. I think he at various times tried to follow the FDR. I am not on perfectly solid ground in that estimation.
For Sassypilots wife:
The Tech Log discussion features no small number of pilots who post there: gums, Machinbird, PJ2, Retired F4, CONFiture, Clandestino, bubbers44, and numerous others. I discuss things there now and again, though I have not flown much of anything in a few years, and never flew heavy metal ever. I have flown and taught flying in multipiloted aircraft, taught instruments, and instructed in simulators where you try to work a crew very hard to see how strong their systems knowledge and airmanship are. I have also investigated mishaps, including mishaps with fatalities.
I would point out that Tech Log is now where ALL of AF447 threads resides, which John T has kindly explained in the preface to the various threads.
The first three IIRC started on the R & N forum.
Rockhound
What about the following remedy for the situation?:
throttling back one engine to force a wing drop, to drag the nose out of its 40-degree up angle into a dive.
throttling back one engine to force a wing drop, to drag the nose out of its 40-degree up angle into a dive.
SLFinAZ, Indeed,
Actually he was entirely correct.
jcj - Rockhound's 'pilot' was talking out of his seat cushion.
So...in the early stages of the upset the statement is entirely correct...
The pilot in question was advocating this method of recovery for the last two minutes of flight, i.e. when AF447 was deep into the stall.
The pilot in question was advocating this method of recovery for the last two minutes of flight, i.e. when AF447 was deep into the stall.
According to the FDR, the nose wasn't 40 degrees up, ever. It was somewhere around 15 degrees up, more or less, at its peak. The AoA, while there was more airspeed, began somewhere between 4 and 6 degrees as the aircraft neared and entered stall (it seems to have been a dynamic event so maybe stall AoA was passed a few degrees higher). Note that even with a roughly fixed pitch, (more or less 15 degrees) the plane slowed AND the AoA increased as airspeed decreased. This should not surprise any pilot who has done stall training.
Based on the analysis of the pitch angle and airspeed, AoA got upwards of 40 degrees as the plane went more deeply into the stall.
Sooo, Rockhound .... maybe that pilot you referenced ought to make sure he understands the problem before he suggests a solution.
Case One
Could any of you professionals here explain what knowledge of flight physics is taught to pilots in their training
What?
Originally Posted by Clandestino
This: CDBDA AADCD BAADC CADCB BABCD AACDB BACDA. I kid you not
I laughed.
I have found the tech log discussions, though full of some chaffe, to have a lot of good nuggets.
I find this thread, after three years of learning about AB, AB330, FBW, sidesticks, control laws, and a few other things, somewhat entertaining.
Mind you, it is also filled with chaffe.
Fly safe. Fly the plane.
Last edited by Lonewolf_50; 9th Jul 2012 at 21:00.
Join Date: Feb 2009
Location: Worldwide
Posts: 579
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Lonewolf_50
...full of some chaffe...
20000+ turns of a broken record.
The reasons behind that might be habit, might be training....
Read the CVR transcript from long before the stall and you'll get an idea of the level of professionalism displayed by this so-called 'crew'. Remarkably reminiscent of AF4590 crew behaviour.
The captain couldn't even readback or recall simple HF allocations as PNF.....
Join Date: Jun 2001
Location: UK
Posts: 27
Likes: 0
Received 0 Likes
on
0 Posts
No Clandestino, the radar had not been set up correctly:
<<<02:07:00 (Bonin) On est apparemment ŕ la limite de la couche, ça devrait aller.
We seem to be at the end of the cloud layer, it might be okay.
In the meantime Robert has been examining the radar system and has found that it has not been set up in the correct mode. Changing the settings, he scrutinizes the radar map and realizes that they are headed directly toward an area of intense activity. >>>
If they had not entered the area of severe icing the incident would not have happened:
<<<At 1h 36m, the flight enters the outer extremities of a tropical storm system. Unlike other planes' crews flying through the region, AF447's flight crew has not changed the route to avoid the worst of the storms. The outside temperature is much warmer than forecast, preventing the still fuel-heavy aircraft from flying higher to avoid the effects of the weather. Instead, it ploughs into a layer of clouds.
At 1h51m, the cockpit becomes illuminated by a strange electrical phenomenon. The co-pilot in the right-hand seat, an inexperienced 32-year-old named Pierre-Cédric Bonin, asks, "What's that?" The captain, Marc Dubois, a veteran with more than 11,000 hours of flight time, tells him it is St. Elmo's fire, a phenomenon often found with thunderstorms at these latitudes. >>>
On the Airbus it is possible to have the radar on but not displaying anything - that was my point.
<<<02:07:00 (Bonin) On est apparemment ŕ la limite de la couche, ça devrait aller.
We seem to be at the end of the cloud layer, it might be okay.
In the meantime Robert has been examining the radar system and has found that it has not been set up in the correct mode. Changing the settings, he scrutinizes the radar map and realizes that they are headed directly toward an area of intense activity. >>>
If they had not entered the area of severe icing the incident would not have happened:
<<<At 1h 36m, the flight enters the outer extremities of a tropical storm system. Unlike other planes' crews flying through the region, AF447's flight crew has not changed the route to avoid the worst of the storms. The outside temperature is much warmer than forecast, preventing the still fuel-heavy aircraft from flying higher to avoid the effects of the weather. Instead, it ploughs into a layer of clouds.
At 1h51m, the cockpit becomes illuminated by a strange electrical phenomenon. The co-pilot in the right-hand seat, an inexperienced 32-year-old named Pierre-Cédric Bonin, asks, "What's that?" The captain, Marc Dubois, a veteran with more than 11,000 hours of flight time, tells him it is St. Elmo's fire, a phenomenon often found with thunderstorms at these latitudes. >>>
On the Airbus it is possible to have the radar on but not displaying anything - that was my point.
Join Date: Dec 2002
Location: Europa
Posts: 612
Likes: 0
Received 0 Likes
on
0 Posts
Thot's on stall recovery
Well done Clandestino (and others with a sound understanding of flying and the perilous state of affairs that this industry has become). At last some comments that help destroy the myths and pointless rants that sometimes appear on these threads.
Ten years before this accident, Airbus Chief Test Pilot Capt W Wainwright wrote an article on Stall recovery - it is doubtful many Airbus pilots have read it. Had the AF447 crew been taught and applied these techniques history would be different:
http://www.google.co.uk/url?sa=t&rct...NBc9_0_SR-U9Sg
Adding power when stalled is poor stall recovery advice. For FAR 25/JAR25 machines with underslung engines, stall recovery for flight test: Unload by reducing AoA - no thrust is added until around 20% above Vs due to the negative effects of adding thrust. It also happens on light propellor aircraft where adding power alone also adds yaw which can induce incipient spin.
How does a glider recover from a stall?
An aircraft can be stalled at any airspeed depending on wing loading - so speed alone is not the only factor.
NASA carried out a stalling project from FL450 with a B757 some years ago. The machine was held at the stalling AoA (only a few degrees nose/pitch up) with full TOGA thrust. It continued to decend at over 5000fpm until standard stall recovery (reduce AoA) technique was used.
Whilst Deep Stall recovery requires a very large AoA reduction (not just lowering a few degrees of pitch), the reason the A330 was heading for the ocean floor almost tail first was because of high thrust vector and aft control column inputs.
I have to completely disagree with the concept of adding full thrust and/or banking to break a stall whatever the altitude. All that is doing is creating a worse Upset senario for both the pilots and (if a high speed dive ensues) the structural integrity of the aircraft. Using roll control when stall buffet is present can also lead to a spin.
The only time to add full TOGA thrust would be in a low level pre full stall senario in conjunction with an AoA reduction. Without auto trim, adding power alone in both an underslung engine and a tractor propellor aircraft will cause a nose up (relative to pilot) pitch which is traditionally negative longitudinal stability.
Whilst W&B have their part, in general (unlike straight wings), swept wings also trend to provide further negative longitudinal stability at the stall with a nose up (relative to pilot) pitching moment as the spanwise flow is modified and the boundary layer thins from the wing tips first. Without wing fences/ Vortex Generators, this also reduces any outer wing/aileron control effectiveness.
We have already seen in the past few years several major incidents and accidents where pilots just adding power to try to recover form a stall:
Thomson Fly B737 Bournemouth UK
:http://www.skybrary.aero/bookshelf/books/875.pdf
Norwegian Merlin Student Stall:
Fatal Merlin crash puts spotlight back on stall recovery
Even a tractor prop aircraft at high IAS with a sudden reduction in power will (for stable design) result in a pitch down - this is positive longitudinal stability which helps to reduce AoA and therefore would unload the wings.
Clearly, what is permissable within the flight envelope of a Fighter jet (some of which are capable of 10G per sec wing loading changes) is not so with an airliner.
Banking an aeroplane to a high AoB is fine for aerobatic and military machines where a high Rate of Decent is required. For larger hardware it can be fatal - not only at low level in the following examples but due to in flight breakup in the subsequent recovery form a high speed dive.
B52:
C17:
The other issue is Va (manoeuvre speed) which is the maximum certified wing loading at the stall (stalling at max "g") and permits ONE control to maximum deflection, ONCE - not simultaeous or multiple roll and pitch control inputs:
American Airlines Flight 587 - Wikipedia, the free encyclopedia
Ten years before this accident, Airbus Chief Test Pilot Capt W Wainwright wrote an article on Stall recovery - it is doubtful many Airbus pilots have read it. Had the AF447 crew been taught and applied these techniques history would be different:
http://www.google.co.uk/url?sa=t&rct...NBc9_0_SR-U9Sg
Adding power when stalled is poor stall recovery advice. For FAR 25/JAR25 machines with underslung engines, stall recovery for flight test: Unload by reducing AoA - no thrust is added until around 20% above Vs due to the negative effects of adding thrust. It also happens on light propellor aircraft where adding power alone also adds yaw which can induce incipient spin.
How does a glider recover from a stall?
An aircraft can be stalled at any airspeed depending on wing loading - so speed alone is not the only factor.
NASA carried out a stalling project from FL450 with a B757 some years ago. The machine was held at the stalling AoA (only a few degrees nose/pitch up) with full TOGA thrust. It continued to decend at over 5000fpm until standard stall recovery (reduce AoA) technique was used.
Whilst Deep Stall recovery requires a very large AoA reduction (not just lowering a few degrees of pitch), the reason the A330 was heading for the ocean floor almost tail first was because of high thrust vector and aft control column inputs.
I have to completely disagree with the concept of adding full thrust and/or banking to break a stall whatever the altitude. All that is doing is creating a worse Upset senario for both the pilots and (if a high speed dive ensues) the structural integrity of the aircraft. Using roll control when stall buffet is present can also lead to a spin.
The only time to add full TOGA thrust would be in a low level pre full stall senario in conjunction with an AoA reduction. Without auto trim, adding power alone in both an underslung engine and a tractor propellor aircraft will cause a nose up (relative to pilot) pitch which is traditionally negative longitudinal stability.
Whilst W&B have their part, in general (unlike straight wings), swept wings also trend to provide further negative longitudinal stability at the stall with a nose up (relative to pilot) pitching moment as the spanwise flow is modified and the boundary layer thins from the wing tips first. Without wing fences/ Vortex Generators, this also reduces any outer wing/aileron control effectiveness.
We have already seen in the past few years several major incidents and accidents where pilots just adding power to try to recover form a stall:
Thomson Fly B737 Bournemouth UK
:http://www.skybrary.aero/bookshelf/books/875.pdf
Norwegian Merlin Student Stall:
Fatal Merlin crash puts spotlight back on stall recovery
Even a tractor prop aircraft at high IAS with a sudden reduction in power will (for stable design) result in a pitch down - this is positive longitudinal stability which helps to reduce AoA and therefore would unload the wings.
Clearly, what is permissable within the flight envelope of a Fighter jet (some of which are capable of 10G per sec wing loading changes) is not so with an airliner.
Banking an aeroplane to a high AoB is fine for aerobatic and military machines where a high Rate of Decent is required. For larger hardware it can be fatal - not only at low level in the following examples but due to in flight breakup in the subsequent recovery form a high speed dive.
B52:
C17:
The other issue is Va (manoeuvre speed) which is the maximum certified wing loading at the stall (stalling at max "g") and permits ONE control to maximum deflection, ONCE - not simultaeous or multiple roll and pitch control inputs:
American Airlines Flight 587 - Wikipedia, the free encyclopedia
Join Date: Dec 2002
Location: Europa
Posts: 612
Likes: 0
Received 0 Likes
on
0 Posts
DP Davies on Deep Stalls
My understanding of Deep/Super Stall is the stage after the initial stall when the aircraft is not recovered immediately - i.e: held in the stall (in AF447 case through aft control column, and full aft tailplane trim along with full thrust). This prolongs deceleration (due drag rise) and despite fairly benign looking attitude (perhaps 8 deg pitch up) leads to a massive increase in the incidence (effective AoA) as the flight path becomes more of a sinking than forward motion.
DP Davies devotes around 6 pages to the subject in his book Handling the Big Jets. He was chief test pilot for the UK Airworthiness Authority (pre CAA) and tested all manner of machines from ultra lights to the B707, B747 and Concorde. He categorically states that Deep/Super stall recovery is only possible where elevator authority exceeds the nose up pitching moment post stall (caused by wing sweep, change in chordwise lift distrubution after initial stall, forward fuselage lifting body effect and now add in AF447 aft stick, trim and TOGA underslung engines).
To quote DP Davies:
"...aircraft that have been lost to such manouevres finally reached the ground substantially level laterally, having defied all efforts to roll or spin them out of this stabilised condition,....with little or no forward speed, rotating only slowly in yaw, ... exposed to massive angles of incidence (AoA) and enormous vertical velocity"
"At one time it was thought that an attempt to roll or spin the aeroplane would offer the best chance. This idea has now been withdrawn (1971), because even assuming that this (additional) upset could be achieved, the resulting very steep nose down attitude, the lack of proof of spin recovery capability, he very high rate of decent and the large height range required (for subsequent) recovery makes it unlikely this method offers any advantage at all"
" The best recovery technique (super stall) is now considered.......persist in full forward elevator control, put the flaps to the postion recommended by the maufacturer and wait for the aeroplane to pitch down and recover from the stall"
"...... pitch attitude is not enough guide to the recovery.Too early a recovery from a gentle dive, following assumed recovery from a super stall, will again increase incidence and ensure the aeroplane remains locked in."
"Without incidence information, a nose down pitch with an increasing speed is no guarantee that recovery has been effected and an up elevator movement at this stage merely serves to keep the aircraft stalled"
DP Davies devotes around 6 pages to the subject in his book Handling the Big Jets. He was chief test pilot for the UK Airworthiness Authority (pre CAA) and tested all manner of machines from ultra lights to the B707, B747 and Concorde. He categorically states that Deep/Super stall recovery is only possible where elevator authority exceeds the nose up pitching moment post stall (caused by wing sweep, change in chordwise lift distrubution after initial stall, forward fuselage lifting body effect and now add in AF447 aft stick, trim and TOGA underslung engines).
To quote DP Davies:
"...aircraft that have been lost to such manouevres finally reached the ground substantially level laterally, having defied all efforts to roll or spin them out of this stabilised condition,....with little or no forward speed, rotating only slowly in yaw, ... exposed to massive angles of incidence (AoA) and enormous vertical velocity"
"At one time it was thought that an attempt to roll or spin the aeroplane would offer the best chance. This idea has now been withdrawn (1971), because even assuming that this (additional) upset could be achieved, the resulting very steep nose down attitude, the lack of proof of spin recovery capability, he very high rate of decent and the large height range required (for subsequent) recovery makes it unlikely this method offers any advantage at all"
" The best recovery technique (super stall) is now considered.......persist in full forward elevator control, put the flaps to the postion recommended by the maufacturer and wait for the aeroplane to pitch down and recover from the stall"
"...... pitch attitude is not enough guide to the recovery.Too early a recovery from a gentle dive, following assumed recovery from a super stall, will again increase incidence and ensure the aeroplane remains locked in."
"Without incidence information, a nose down pitch with an increasing speed is no guarantee that recovery has been effected and an up elevator movement at this stage merely serves to keep the aircraft stalled"
Last edited by angelorange; 10th Jul 2012 at 13:19.
Join Date: Dec 2010
Location: Middle America
Age: 84
Posts: 1,167
Likes: 0
Received 0 Likes
on
0 Posts
angelorange, Just to clarify something:
A "Deep Stall/Super Stall is confined to an airplane design having a "t-tail". What happens is, under a severe pitch-up and speed decay situation the wings stall and the turbulent air coming off the top surfaces of the stalled wings renders the horizontal stabilizers and elevators useless in terms of recovering from the stall condition. The C-17 crash is a prime example of this phenomenon as was the NW B-727 accident out of JFK in December 1974. I am only aware of one successful recovery from a deep stall on a t-tail aircraft. It occurred during the B-727 test program where the test pilots had enough altitude and were able to rock the wings to get the nose down and regain speed and lift on the wings.
AF447 never suffered from this sort of condition. The THS and elevators still had lift in the stall and a prolonged forward stick probably would have recovered the aircraft. The design of the aircraft matters a lot.
My understanding of Deep/Super Stall is the stage after the initial stall when the aircraft is not recovered immediately - i.e: held in the stall (in AF447 case through aft control column, and full aft tailplane trim along with full thrust).
AF447 never suffered from this sort of condition. The THS and elevators still had lift in the stall and a prolonged forward stick probably would have recovered the aircraft. The design of the aircraft matters a lot.
Join Date: Mar 2010
Location: South Korea
Age: 62
Posts: 115
Likes: 0
Received 0 Likes
on
0 Posts
Simple status feedback.
I am an electrical engineer with considerable experience in highly automated equipment that can kill people when things go wrong. Hence I am highly interested in these discussions about the plane’s automation. I have very limited flying experience.
A reason for the AF447’s pilot illogical behavior has been given as the “startle factor” or “information overload” The key trigger for this was the mode change to alternate law, At this point recognition of a stall was unnecessary because the plane was not stalled. The UAS was the trigger for the mode change but the UAS was recognized by the pilots. What appears to have been misunderstood was the consequences of the mode change. If the pilot recognized that the mode change disabled the AOA protection and the consequences having no AOA protection I do not think he would have pulled the stick fully back prior to the plane becoming stalled.
The plane was in a non optimum state, the computer recognized this state and implemented some counter measures to supposedly keep the plane flying. The key counter measure (from my very limited understanding) was removing the AOA protection and handing the elevator control over to the PF who supposedly with his superior senses and analytical skills can continue flying the plane with feedback from various other sources including the computer. However his ability to cope with the situation is dependent on how well he understands the situation from his prior training, what he can decipher from the environment and what the computer is telling him.
A modern computer has a lot of processing power and lots of available memory allowing for a complex program. With a complex computer program, when there is a reversion to “manual mode” better feedback indicating the status of the equipment (an aeroplane in this case) as seen by the computer is required. Note, I am referring to the computer status. While feedback of the PF’s physical body status can be determined better with control columns what I am talking about is the status of the plane as seen or determined by the computer. With simpler programs indicator lights, aural warnings and text messages suffice but as the computer actions get more complex better feedback to the operator (pilot in this case) is required.
In industry we now have better tools for providing such feedback. In complex automated applications we now implement touch screens with pictorial representations of the equipment. A picture contains 1000 words. There appears to be limited pictorial representation of the plane’s status as determined by the computer. In the AF447 case we get messages indicating UAS and a mode change. The pilot has to understand the consequences of the mode change, one of which is the removal of the AOA protection. He must also understand the consequences of the removal of the AOA protection and that the side stick inputs now control the surface deflection and not the G (as far as I understand) and a plethora of other things crucial to the survival of the plane and its contents. Good, simple computer feed back of the status of the equipment from the computer’s viewpoint enables better understanding of the situation and better pilot ability to cope with the situation. In this case a message indicating the mode has changed was announced and it is up to the pilot to know the resulting complex changes to the flight controls. There is a lot of understanding here that the pilot is required to bring to the front of his analytical thinking in a short space of time. What is required in such complex computer programs is good simple feedback to help the operator understand the status of the equipment as determined by the computer. Rather than announce a vague description such as a mode change the pilots need to know exactly what is happening with the controls. In this case the AOA protection (among other things) has been lost. A display with a pictorial view of the plane with the elevators color changing from green to orange would be a good start. This would indicate there is some abnormality with the elevator control and further to this there would be some indication showing the AOA protection is lost etc. This is an example, I am sure a pilot would have better ideas as to how this should operate.
Note, I may have errors in my aeronautical descriptions, which I apologize for but please understand the main point of my post is to highlight the need for better computer feed back as computer programs get more complex when there is a reversion to “manual mode” (alternate law).
A reason for the AF447’s pilot illogical behavior has been given as the “startle factor” or “information overload” The key trigger for this was the mode change to alternate law, At this point recognition of a stall was unnecessary because the plane was not stalled. The UAS was the trigger for the mode change but the UAS was recognized by the pilots. What appears to have been misunderstood was the consequences of the mode change. If the pilot recognized that the mode change disabled the AOA protection and the consequences having no AOA protection I do not think he would have pulled the stick fully back prior to the plane becoming stalled.
The plane was in a non optimum state, the computer recognized this state and implemented some counter measures to supposedly keep the plane flying. The key counter measure (from my very limited understanding) was removing the AOA protection and handing the elevator control over to the PF who supposedly with his superior senses and analytical skills can continue flying the plane with feedback from various other sources including the computer. However his ability to cope with the situation is dependent on how well he understands the situation from his prior training, what he can decipher from the environment and what the computer is telling him.
A modern computer has a lot of processing power and lots of available memory allowing for a complex program. With a complex computer program, when there is a reversion to “manual mode” better feedback indicating the status of the equipment (an aeroplane in this case) as seen by the computer is required. Note, I am referring to the computer status. While feedback of the PF’s physical body status can be determined better with control columns what I am talking about is the status of the plane as seen or determined by the computer. With simpler programs indicator lights, aural warnings and text messages suffice but as the computer actions get more complex better feedback to the operator (pilot in this case) is required.
In industry we now have better tools for providing such feedback. In complex automated applications we now implement touch screens with pictorial representations of the equipment. A picture contains 1000 words. There appears to be limited pictorial representation of the plane’s status as determined by the computer. In the AF447 case we get messages indicating UAS and a mode change. The pilot has to understand the consequences of the mode change, one of which is the removal of the AOA protection. He must also understand the consequences of the removal of the AOA protection and that the side stick inputs now control the surface deflection and not the G (as far as I understand) and a plethora of other things crucial to the survival of the plane and its contents. Good, simple computer feed back of the status of the equipment from the computer’s viewpoint enables better understanding of the situation and better pilot ability to cope with the situation. In this case a message indicating the mode has changed was announced and it is up to the pilot to know the resulting complex changes to the flight controls. There is a lot of understanding here that the pilot is required to bring to the front of his analytical thinking in a short space of time. What is required in such complex computer programs is good simple feedback to help the operator understand the status of the equipment as determined by the computer. Rather than announce a vague description such as a mode change the pilots need to know exactly what is happening with the controls. In this case the AOA protection (among other things) has been lost. A display with a pictorial view of the plane with the elevators color changing from green to orange would be a good start. This would indicate there is some abnormality with the elevator control and further to this there would be some indication showing the AOA protection is lost etc. This is an example, I am sure a pilot would have better ideas as to how this should operate.
Note, I may have errors in my aeronautical descriptions, which I apologize for but please understand the main point of my post is to highlight the need for better computer feed back as computer programs get more complex when there is a reversion to “manual mode” (alternate law).
Join Date: Dec 2010
Location: Middle America
Age: 84
Posts: 1,167
Likes: 0
Received 0 Likes
on
0 Posts
CG of AF447
Originally posted by Lyman:
You are correct, the Tail tanks were full at the time of a/p loss (2:10:04.6). 10.5 thousand pounds of fuel. Initially the Cg computed by BEA was released to be right at the margin for flight (37.6?). Later, that value was retracted and replaced with ~ 28.7. In turbulence and experiencing rolling and pitching moments, the Cg may have been well aft, and exacerbated the inability to recover from the initial upset at a/p quit.....
You are correct, the Tail tanks were full at the time of a/p loss (2:10:04.6). 10.5 thousand pounds of fuel. Initially the Cg computed by BEA was released to be right at the margin for flight (37.6?). Later, that value was retracted and replaced with ~ 28.7. In turbulence and experiencing rolling and pitching moments, the Cg may have been well aft, and exacerbated the inability to recover from the initial upset at a/p quit.....
Here is a statement regarding this:
The use of wing sweep and stability augmentation on modern commercial airplanes makes them more fuel efficient. However, flight crews must understand the effects of CG and altitude on performance and handling qualities. For example, operating at an aft CG improves cruise performance, but moving the CG aft reduces static longitudinal and maneuvering stability. Many modern commercial airplanes employ some form of stability augmentation to compensate for relaxed stability. However, as long as the CG is in the allowable range, the handling qualities will be adequate with or without augmentation. An understanding of static and maneuvering longitudinal stability is an essential element of flight crew training.
Join Date: Dec 2010
Location: Middle America
Age: 84
Posts: 1,167
Likes: 0
Received 0 Likes
on
0 Posts
Lonewolf 50 in Post#265 made an excellent suggestion:
To make it even easier for those interested, here are the memory items:
There has been considerable discussion in the Tech Log regarding whether or not the safe conduct of the flight was impacted or not in the case of AF447 at cruise. Ultimately, flying pitch and power and/or getting out the pitch and power tables after checking and completing the memory list items would have saved the day.
Additionally, a great deal of information regarding the A330-200 is given in this presentation:
http://www.smartcockpit.com/data/pdf...t_Controls.pdf
An interesting article by Boeing can be found here:
Aero 08 - Erroneous Flight Instrument Information
Happy reading !!!
For anyone interested:
If you go to the tech log forum, you will find the following point excerpted from the BEA report.
The UAS (Unreliable AirSpeed) procedure memory item of turning off the FD's was not accomplished. (FD = Flight Director).
There are even some handy pictures and graphs that show when the FD was receiving good data, and when it wasn't. Mind you, neither pilot was aware of when it was good and when it was bad, but given the procedure as stated, they should not have needed to.
If you go to the tech log forum, you will find the following point excerpted from the BEA report.
The UAS (Unreliable AirSpeed) procedure memory item of turning off the FD's was not accomplished. (FD = Flight Director).
There are even some handy pictures and graphs that show when the FD was receiving good data, and when it wasn't. Mind you, neither pilot was aware of when it was good and when it was bad, but given the procedure as stated, they should not have needed to.
There has been considerable discussion in the Tech Log regarding whether or not the safe conduct of the flight was impacted or not in the case of AF447 at cruise. Ultimately, flying pitch and power and/or getting out the pitch and power tables after checking and completing the memory list items would have saved the day.
Additionally, a great deal of information regarding the A330-200 is given in this presentation:
http://www.smartcockpit.com/data/pdf...t_Controls.pdf
An interesting article by Boeing can be found here:
Aero 08 - Erroneous Flight Instrument Information
Happy reading !!!
Or was PF trying to implement the wrong UAS procedure, namely 15 degrees pitch up and TOGA, given that is what he would have mostly trained on in the sim for a UAS event? With Normal Law protection, the plane would not have stalled following this procedure (or even adding full pitch up with no thrust change) but in the stress of the moment he didnt register that it had reverted to Alt 2 Law.
Join Date: Mar 2009
Location: Malaysia
Posts: 112
Likes: 0
Received 0 Likes
on
0 Posts
@Organfreak
Mr.Organ,
I am not confused re your position.
I am not being dismissive re 'piloting experts', but I am entitled to question their opinions.
As for being 'uncool', it is not really something that bothers me. I leave that to those aged around 14 or 15.
Mr.Organ,
I am not confused re your position.
I am not being dismissive re 'piloting experts', but I am entitled to question their opinions.
As for being 'uncool', it is not really something that bothers me. I leave that to those aged around 14 or 15.
Join Date: Jun 2001
Location: Surrounded by aluminum, and the great outdoors
Posts: 3,780
Likes: 0
Received 0 Likes
on
0 Posts
NW B-727 accident out of JFK in December 1974.
Another classic case in failure to unload the wings, the pitch in this case was actually quite pronounced nose down, but since the FPA was almost straight down, the aircraft remained stalled, and in the end, the horizontal stab failed structurally, and the aircraft crashed...airspeed unreliable was also the initial cause of this accident...(frozen pitots)
http://www.fss.aero/accident-reports...4-12-01-US.pdf
Another classic case in failure to unload the wings, the pitch in this case was actually quite pronounced nose down, but since the FPA was almost straight down, the aircraft remained stalled, and in the end, the horizontal stab failed structurally, and the aircraft crashed...airspeed unreliable was also the initial cause of this accident...(frozen pitots)
http://www.fss.aero/accident-reports...4-12-01-US.pdf
Last edited by ironbutt57; 10th Jul 2012 at 06:00.
Join Date: Jun 2001
Location: Surrounded by aluminum, and the great outdoors
Posts: 3,780
Likes: 0
Received 0 Likes
on
0 Posts
On the Airbus it is possible to have the radar on but not displaying anything - that was my point.
Thats true on almost any type of weather radar in any aircraft...other info includes a prog chart, assumed to be provided to the crew, and of course...LOOKING OUT THE WINDOW.....one might have seen the wall of lightening ahead and adjusted the radar sooner.....(what a novel idea)...
Thats true on almost any type of weather radar in any aircraft...other info includes a prog chart, assumed to be provided to the crew, and of course...LOOKING OUT THE WINDOW.....one might have seen the wall of lightening ahead and adjusted the radar sooner.....(what a novel idea)...
Join Date: Apr 2008
Location: .
Posts: 309
Likes: 0
Received 0 Likes
on
0 Posts
1. Don't fly into a storm
Join Date: Jun 2001
Location: Surrounded by aluminum, and the great outdoors
Posts: 3,780
Likes: 0
Received 0 Likes
on
0 Posts
Air France 447 - AFR447 - A detailed meteorological analysis - Satellite and weather data
Interesting stuff, survivable in itself, but with airspeed unreliable in the middle of it??...
Interesting stuff, survivable in itself, but with airspeed unreliable in the middle of it??...
Join Date: Sep 2005
Location: Oslo, Norway
Posts: 189
Likes: 0
Received 0 Likes
on
0 Posts
Bad software.
Originally Posted by Cool Guys
I am an electrical engineer with considerable experience in highly automated equipment that can kill people when things go wrong. Hence I am highly interested in these discussions about the plane’s automation. I have very limited flying experience.
I think your post was spot on with regard to the lack of sound HCI principles in the design of the autopilot system. If this kind of automation is going to be in place, it should not completely crumble upon the loss of one input parameter (albeit a very important one).
The flight computer is aware of the weight, the altitude, the AOA, the last indicated reliable airspeed, and almost everything else the pilot will use to perform his own calculations to decide on the best course of action. If the autopilot just hands over control as it did in this situation, and the PF is supposed to go 85% N1 and 5% AOA (or whatever it is), then I think the autopilot should do that for him, and just display a very big warning. The pilot always has the option to disconnect the autopilot if it starts doing something stupid, this would at least give them more time to think and react.
Heck, can't the autopilot even handle stall recovery? It had the vertical speed parameter intact and it knew AOA, thrust settings, elevator position, trim, as well as weight etc. so it should be able to preempt a stall even without knowing airspeed. Am I wrong? I'm tempted to write the code for that and send it to Airbus, though most of it probably already exists in flight simulator software.
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Cool and PPY - these are indeed valid queries you have raised. The basic answer is that the 'task' presented to the crew was a very simple one as far as we can tell from the report - fly the a/c, set a 'safe' attitude and if necessary then adjust power while the rest of the actions were completed. As you can read to the point of complete boredom and disillusionment on all these threads, both pilots failed this task.
The idea of a 'straight and level' 'fallback' autopilot function is a good one, except that as you will both doubtless know, once you start coding for this you will find there will probably be situations where that is not the best solution.
However, the idea has merit and is worthy of discussion here. Something needs doing as we all recognise. The ideal is having pilots that can undertake a relatively simple task, but if we fail there your system may have to come into play.
We need to take it step further, of course, to the ?eventual? pilot-less a/c and when this mode would be triggered and under which parameters. Not easy!
The idea of a 'straight and level' 'fallback' autopilot function is a good one, except that as you will both doubtless know, once you start coding for this you will find there will probably be situations where that is not the best solution.
However, the idea has merit and is worthy of discussion here. Something needs doing as we all recognise. The ideal is having pilots that can undertake a relatively simple task, but if we fail there your system may have to come into play.
We need to take it step further, of course, to the ?eventual? pilot-less a/c and when this mode would be triggered and under which parameters. Not easy!