AF447 wreckage found
Join Date: Jun 2008
Location: 31000FR
Posts: 60
Likes: 0
Received 0 Likes
on
0 Posts
No, we do not get all the details, and unfortunately whoever did the translation into English is not a professional pilot. "Assiette" in French means Pitch, not AOA (which is "incidence" in French).
So, initially the PF raised the nose to 10 dg+ (with less than 7 dg AOA which would have caused an immediate accelerated stall) and at 38000`the nose lowered to around +5 dg. The momentum + the TOGA thrust (which at this altitude is only a modest MCL) then took the aircraft into a deep stall.
At this time the only survivable strategy would have been to pitch the aircraft down to below the horizon.
For the record: even with aft trimmed stab at VS, the A330 does have sufficient nose down authority for this. Only in a fully developed deep stall at very low EAS may this not be available; goes for almost any aircraft type.
As a TRI/TRE I do these hi-altitude stall recovery exercises on a regular basis on A330 and believe me; in a confusing scenario like the actual one, only the most proficient pilots prevail, so please stop the "pilot-factor" blame game.
If the majority of pilots cannot cope with a given situation it is either "natural causes" (like all-engine failure + loss of instruments in ash scenario) or "system failure" (if an operator or manufacturer does not ensure adequate training for a particular situation).
The 747 "sales video" is nice, but misleading. It is true that you can power your way out in level flight (and the A330 actually does this better than the 74) provided you have not stalled. In that case you must regain airflow by lowering the nose and thus decrease the AOA and regain speed.
When building an aircraft, the various design offices try to envision all kinds of contingencies and design a safe aircraft accordingly, but there are limits to human imagination (like when a 74 lost part of the rudder in Japan, or when the same happened to an A330 in the US ).
The events so far seem to be logical but unfortunate. We can only hope that the operators will use this wake up call to go beyond the "minimum required training" policy that is so common today in our cost saving environment. Few pilots come to the airlines with 10 years of fighter experience and the rest have never seen a 10 dg nose down recovery at FL 380.
And for the stupid West Coast vs. TLS comparison: I have stalled wide bodies from both sides: Marginally prefer TLS for its better wing design (and the good red wine you get after the flight testing
.
So, initially the PF raised the nose to 10 dg+ (with less than 7 dg AOA which would have caused an immediate accelerated stall) and at 38000`the nose lowered to around +5 dg. The momentum + the TOGA thrust (which at this altitude is only a modest MCL) then took the aircraft into a deep stall.
At this time the only survivable strategy would have been to pitch the aircraft down to below the horizon.
For the record: even with aft trimmed stab at VS, the A330 does have sufficient nose down authority for this. Only in a fully developed deep stall at very low EAS may this not be available; goes for almost any aircraft type.
As a TRI/TRE I do these hi-altitude stall recovery exercises on a regular basis on A330 and believe me; in a confusing scenario like the actual one, only the most proficient pilots prevail, so please stop the "pilot-factor" blame game.
If the majority of pilots cannot cope with a given situation it is either "natural causes" (like all-engine failure + loss of instruments in ash scenario) or "system failure" (if an operator or manufacturer does not ensure adequate training for a particular situation).
The 747 "sales video" is nice, but misleading. It is true that you can power your way out in level flight (and the A330 actually does this better than the 74) provided you have not stalled. In that case you must regain airflow by lowering the nose and thus decrease the AOA and regain speed.
When building an aircraft, the various design offices try to envision all kinds of contingencies and design a safe aircraft accordingly, but there are limits to human imagination (like when a 74 lost part of the rudder in Japan, or when the same happened to an A330 in the US ).
The events so far seem to be logical but unfortunate. We can only hope that the operators will use this wake up call to go beyond the "minimum required training" policy that is so common today in our cost saving environment. Few pilots come to the airlines with 10 years of fighter experience and the rest have never seen a 10 dg nose down recovery at FL 380.
And for the stupid West Coast vs. TLS comparison: I have stalled wide bodies from both sides: Marginally prefer TLS for its better wing design (and the good red wine you get after the flight testing
.
Join Date: Jun 2011
Location: Shoreham
Age: 72
Posts: 35
Likes: 0
Received 0 Likes
on
0 Posts
Cruise stall recovery
In our company we developed the following cruise stall recovery by trial and error in the B747-400 simulator due to the fact that the aircraft would not accelerate in level flight with full power...
Select VS -3000 ft/min
(stick shaker stops almost at once)
When IAS is half way up amber caution range...
Reduce VS in 100 ft/min stages with caution...
When IAS is at top of amber caution range...
Select FLCH to recover in protected mode to cruise alt.
Typical height loss is 2000-2500 feet.
Training Captains with plenty of time in the simulator could with trial and error finesse this to a height loss of 1000 feet but often nibbled the stick shaker...the above method works every time for regular line pilots.
Select VS -3000 ft/min
(stick shaker stops almost at once)
When IAS is half way up amber caution range...
Reduce VS in 100 ft/min stages with caution...
When IAS is at top of amber caution range...
Select FLCH to recover in protected mode to cruise alt.
Typical height loss is 2000-2500 feet.
Training Captains with plenty of time in the simulator could with trial and error finesse this to a height loss of 1000 feet but often nibbled the stick shaker...the above method works every time for regular line pilots.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
FDR record in real time (no delay or blank timing)
So the BEA use of word "around" is deceiving
Not being familiar with the specific FDR output characteristics, it could conceiveably be the FDR data sampling intervals, for instance- write every 5 seconds...
So the BEA use of word "around" is deceiving
Join Date: Apr 2007
Location: Aberdeen
Posts: 56
Likes: 0
Received 0 Likes
on
0 Posts
jcjeant
With respect, I do not think that the use of the word "around" is "deceiving". My background is IT, and when you look at log files from certain software applications the timestamps will be accurate to hundredths or thousandths of seconds.
Therefore, it is not unreasonable to assume that the log files from the DFDR are recording in hundredths or thousandths of seconds intervals, which would mean that for ease the time would be rounded up to the nearest second hence the use of the word "around".
Or maybe it's just been lost in translation from French to English...
With respect, I do not think that the use of the word "around" is "deceiving". My background is IT, and when you look at log files from certain software applications the timestamps will be accurate to hundredths or thousandths of seconds.
Therefore, it is not unreasonable to assume that the log files from the DFDR are recording in hundredths or thousandths of seconds intervals, which would mean that for ease the time would be rounded up to the nearest second hence the use of the word "around".
Or maybe it's just been lost in translation from French to English...
PEI, thanks, well said. 
The question you refer to that I learned to ask about mishaps "were they set up to fail?" describes a generic way to look at the human to system to machine interface, and was not directed at the crew of AF 447 personally. What is behind that question is overcoming the throwaway of "OH, it was pilot error" in one mishap or another, and to lead to identifying the influences of systems, conditions, training, and other inputs (like from ATC or unergonomic cockpit set ups) that contribute to things going wrong.
Considering the numerous extended threads on this topic, some common themes keep cropping up: there are some arguments to be made that the crew was set up for a surprise, and worse.
If you look at it from the perspective of Team Air France, the flight failed to do as intended: deliver plane and pax to Paris. Would a more useful form of that question be: "where were the seeds of mission failure sown?"
I suggest that's the driving question behind BEA's difficult task.
The question you refer to that I learned to ask about mishaps "were they set up to fail?" describes a generic way to look at the human to system to machine interface, and was not directed at the crew of AF 447 personally. What is behind that question is overcoming the throwaway of "OH, it was pilot error" in one mishap or another, and to lead to identifying the influences of systems, conditions, training, and other inputs (like from ATC or unergonomic cockpit set ups) that contribute to things going wrong.
Considering the numerous extended threads on this topic, some common themes keep cropping up: there are some arguments to be made that the crew was set up for a surprise, and worse.
If you look at it from the perspective of Team Air France, the flight failed to do as intended: deliver plane and pax to Paris. Would a more useful form of that question be: "where were the seeds of mission failure sown?"
I suggest that's the driving question behind BEA's difficult task.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
I disagree this to justify the use of word "around" by the BEA
Read again the BEA note please.
For all the other timings BEA use **H**Min**Sec
Why use the word around for another important timing ?
Something lost in the translation ? .. that's possible .. but not acceptable from a organization like the BEA ....
They are not writing romances .. they are writing reports and notes that can have in the futur .. consequences on the life of many people...
With respect, I do not think that the use of the word "around" is "deceiving". My background is IT, and when you look at log files from certain software applications the timestamps will be accurate to hundredths or thousandths of seconds.
Therefore, it is not unreasonable to assume that the log files from the DFDR are recording in hundredths or thousandths of seconds intervals, which would mean that for ease the time would be rounded up to the nearest second hence the use of the word "around".
Or maybe it's just been lost in translation from French to English...
Therefore, it is not unreasonable to assume that the log files from the DFDR are recording in hundredths or thousandths of seconds intervals, which would mean that for ease the time would be rounded up to the nearest second hence the use of the word "around".
Or maybe it's just been lost in translation from French to English...
Read again the BEA note please.
For all the other timings BEA use **H**Min**Sec
Why use the word around for another important timing ?
Something lost in the translation ? .. that's possible .. but not acceptable from a organization like the BEA ....
They are not writing romances .. they are writing reports and notes that can have in the futur .. consequences on the life of many people...
Join Date: Mar 2001
Location: us
Posts: 694
Likes: 0
Received 0 Likes
on
0 Posts
PEI 3721
Without the release of the full CVR transcript, there is no way of knowing what they saw, or didn't see painted on their radar. We do know what the Lufthansa, Iberia, and AF459 crews saw on their radars that night because they have told the BEA.
As for the Cb, the satellite imagery of 02h07 as analyzed by Meteo France (included in the first interim BEA report) indicates at 02h10 off their starboard wing there was a rather cold cloud top:
I assume you don't need a translator for "cumulonimbus".
Tim Vasquez has this to say in his June 1 2011 re-analysis,
Air France 447 - AFR447 - A detailed meteorological analysis - Satellite and weather data
CP, #1558
1. ‘They were in a CB’’. There is no evidence that the aircraft was in a CB; the crew had seen and planned / turned to deviate around weather.
1. ‘They were in a CB’’. There is no evidence that the aircraft was in a CB; the crew had seen and planned / turned to deviate around weather.
As for the Cb, the satellite imagery of 02h07 as analyzed by Meteo France (included in the first interim BEA report) indicates at 02h10 off their starboard wing there was a rather cold cloud top:
On constate qu’à 2 h 07 les températures les plus froides sont de l’ordre de -75 °C à -80 °C, alors que la tropopause se situe entre les FL500 et FL520, avec une température voisine de - 80°C : certains des cumulonimbus de l’amas ont atteint l’altitude de la tropopause et leur stade de maturité, mais l’imagerie ne révèle aucun développement vertical exceptionnel du point de vue climatologique, qui serait caractérisé par un « overshoot ».
Tim Vasquez has this to say in his June 1 2011 re-analysis,
Based on the soundings above, my conclusion is that the maximum cumulonimbus tops were 56,000 ft with an equilibrium level of 47,000 ft, representing the tops of most parts of the MCS except near the edges. This agrees fairly well with the observed METEOSAT thermal data.
.....
This indicates that the aircraft was flying through convective clouds at about 0150 UTC and again from 0158 UTC onward.
.....
This indicates that the aircraft was flying through convective clouds at about 0150 UTC and again from 0158 UTC onward.
Not likely, Vovachan.
Type training typically includes systems understanding and working knowledge. The "all or nothing" idea is at odds with how each instrument is fed data by different systems, something pilots who fly professionally have to understand to pass their exams and type ratings.
Type training typically includes systems understanding and working knowledge. The "all or nothing" idea is at odds with how each instrument is fed data by different systems, something pilots who fly professionally have to understand to pass their exams and type ratings.
Join Date: Jun 2009
Location: California
Age: 55
Posts: 203
Likes: 0
Received 0 Likes
on
0 Posts
Bad weather
I would add that the pilot reduced cruise speed from mach .82 to .80 which is condusive with passing through bad weather. The purpose of this slowing is to reduce stress forces on the air frame. Secondly the BEA clearly states that the pilots advised cabin crew of turbulence. Also stated was air temperature increase which indicates updrafts and turbulent weather.
Join Date: Jun 2011
Location: Italy
Age: 64
Posts: 3
Likes: 0
Received 0 Likes
on
0 Posts
I am not a pilot but I have an MS in Aerospace Engineering and an MS in Aeronautical Engineering. I have worked for over 15 years in the aerospace industry. My main occupation has been control law design for satellites - this is different from flight control SW for atmospheric flight but there are many common issues. Today I work in a company that measures the complexity of systems (air traffic systems, power plants, IT systems and SW).
Modern airliners - especially those that fly-by-wire - are full of SW (4+ million lines of code is a commonly cited figure). In practice, pilots "train to fly the SW", not the aircraft. I agree, it is a strong statement to make but when I see pilots posting in this thread that the AoA is not something you necessarily want to be displayed in a cockpit it kind of confirms my point.
But the point is this. I work in measuring the complexity of (SW) systems and I can state that when you have 4+ million lines of code:
1. There is a huge amount of circumstances (combinations of operating conditions) are never tested.
2. To test a SW system of that size you need another SW system that is at least as large as the one you're testing.
3. Because high complexity implies the capacity to deliver surprising behavior, SW systems of that size are almost bound to do so.
4. There are some misconceptions when it comes to systems of systems:
- if you have 100 components then you can get at the most 100 headaches. In actual fact, the number is orders of magnitude larger.
- the worst condition for a given system is NOT that which corresponds to all variables operating each at its operating limits. Sometimes, combinations of values well within the design bounds correspond to the worst-case scenario
- if you have 100 great components a system made of these components is also great.
5. High complexity implies high fragility. If we continue to manufacture more complex SW and more "intelligent" aircraft, these will cause an increasing number of accidents.
My opinion, based simply on my own professional experience and knowledge, is that we are in the hands of computers and that the trend will be to go in that direction even more. I used to work for a computer HW company in the late 1990s. We had a problem with our operating system on one of our models (which was already on the market). We put in one meeting room all our experts on OUR operating system. Their knowledge, when added up, was estimated to cover about 99% of OUR own product! Now 1% of a complex product/system is still something terribly huge. That was in the 1990s, now things are even worse!
Modern airliners - especially those that fly-by-wire - are full of SW (4+ million lines of code is a commonly cited figure). In practice, pilots "train to fly the SW", not the aircraft. I agree, it is a strong statement to make but when I see pilots posting in this thread that the AoA is not something you necessarily want to be displayed in a cockpit it kind of confirms my point.
But the point is this. I work in measuring the complexity of (SW) systems and I can state that when you have 4+ million lines of code:
1. There is a huge amount of circumstances (combinations of operating conditions) are never tested.
2. To test a SW system of that size you need another SW system that is at least as large as the one you're testing.
3. Because high complexity implies the capacity to deliver surprising behavior, SW systems of that size are almost bound to do so.
4. There are some misconceptions when it comes to systems of systems:
- if you have 100 components then you can get at the most 100 headaches. In actual fact, the number is orders of magnitude larger.
- the worst condition for a given system is NOT that which corresponds to all variables operating each at its operating limits. Sometimes, combinations of values well within the design bounds correspond to the worst-case scenario
- if you have 100 great components a system made of these components is also great.
5. High complexity implies high fragility. If we continue to manufacture more complex SW and more "intelligent" aircraft, these will cause an increasing number of accidents.
My opinion, based simply on my own professional experience and knowledge, is that we are in the hands of computers and that the trend will be to go in that direction even more. I used to work for a computer HW company in the late 1990s. We had a problem with our operating system on one of our models (which was already on the market). We put in one meeting room all our experts on OUR operating system. Their knowledge, when added up, was estimated to cover about 99% of OUR own product! Now 1% of a complex product/system is still something terribly huge. That was in the 1990s, now things are even worse!
Lonewolf, re "were they set up to fail?" # 1582
In that sense, they (us, the industry), were ‘set up’ to fail. However, ‘set up’ could imply forethought / previous knowledge, which might be, or its absence might be, a major contributor to the accident.
With hindsight, many posts question “why did/didn’t they” … etc, etc; whereas if the industry had foreseen these issues then the accident might have been avoided, e.g. a complex interaction of system design and human reaction, changing economic climate, social changes in operations and training, expectation, peer pressure.
Alternatively, where the industry considers issues, human judgement could still decide not to act – the risk is acceptable, e.g. multiple pitot icing resulting in LOC. Much of this is driven but public expectation; - and we are a very safe industry, but beware the seeds of complacency.
The above reflects normal human behaviour; humans are irrational. We don’t foresee every eventuality, systems cannot be tested rigorously, particularly where there is human interaction, and where problems are identified safety vs practicality are often judged to balance.
In this sense, AF 447 may have been an accident too far. Yet there have been other, similar situations where the humans rescued the situation (DC-10 Sioux City, A300 Baghdad). The human behaviour in all of these was identical – the humans did their best in the circumstances – as they saw the situation; only with hindsight is an event judged, and on occasion, ‘best’ is not good enough.
I prefer to avoid hindsight, instead look to see how and why human performance varies, and hope to establish what might be done to improve human performance or change the nature of circumstances to be faced (technical / social); but this probably requires some forethought, which is where the ‘seeds of failure’ might be found.
In that sense, they (us, the industry), were ‘set up’ to fail. However, ‘set up’ could imply forethought / previous knowledge, which might be, or its absence might be, a major contributor to the accident.
With hindsight, many posts question “why did/didn’t they” … etc, etc; whereas if the industry had foreseen these issues then the accident might have been avoided, e.g. a complex interaction of system design and human reaction, changing economic climate, social changes in operations and training, expectation, peer pressure.
Alternatively, where the industry considers issues, human judgement could still decide not to act – the risk is acceptable, e.g. multiple pitot icing resulting in LOC. Much of this is driven but public expectation; - and we are a very safe industry, but beware the seeds of complacency.
The above reflects normal human behaviour; humans are irrational. We don’t foresee every eventuality, systems cannot be tested rigorously, particularly where there is human interaction, and where problems are identified safety vs practicality are often judged to balance.
In this sense, AF 447 may have been an accident too far. Yet there have been other, similar situations where the humans rescued the situation (DC-10 Sioux City, A300 Baghdad). The human behaviour in all of these was identical – the humans did their best in the circumstances – as they saw the situation; only with hindsight is an event judged, and on occasion, ‘best’ is not good enough.
I prefer to avoid hindsight, instead look to see how and why human performance varies, and hope to establish what might be done to improve human performance or change the nature of circumstances to be faced (technical / social); but this probably requires some forethought, which is where the ‘seeds of failure’ might be found.
Saturn V, I suspect that we are interpreting words differently.
What I posted was “the crew had seen and planned / turned to deviate around weather”, which was based on details released so far.
I agree we do not know what exactly the crew saw, but circumstantial evidence suggests that the aircraft was not ‘in’ the core of a CB, although it was in an area of convective weather.
In this regard I disagree with Tim Vasquez’s conclusion “This indicates that the aircraft was flying through convective clouds at about 0150 UTC and again from 0158 UTC onward.” Implying, (my interpretation) flying through a Cb core.
From personal experience – ‘in’ and around Cbs – I would be very surprised if the reported conversations and aircraft parameters originate from a situation in a Cb core, but transiting an area of convective clouds would be understandable.
IMHO, the aircraft was below a cloud layer or in high level cirrus (anvil) in a situation similar to that shown in VH-EBA Incident (picture on page 11) where there were high cloud tops in an area of convective weather – a classic ice crystal situation.
What I posted was “the crew had seen and planned / turned to deviate around weather”, which was based on details released so far.
I agree we do not know what exactly the crew saw, but circumstantial evidence suggests that the aircraft was not ‘in’ the core of a CB, although it was in an area of convective weather.
In this regard I disagree with Tim Vasquez’s conclusion “This indicates that the aircraft was flying through convective clouds at about 0150 UTC and again from 0158 UTC onward.” Implying, (my interpretation) flying through a Cb core.
From personal experience – ‘in’ and around Cbs – I would be very surprised if the reported conversations and aircraft parameters originate from a situation in a Cb core, but transiting an area of convective clouds would be understandable.
IMHO, the aircraft was below a cloud layer or in high level cirrus (anvil) in a situation similar to that shown in VH-EBA Incident (picture on page 11) where there were high cloud tops in an area of convective weather – a classic ice crystal situation.
Vovachan; lonewolf, re “assumed ALL instruments were suspect/wrong”, but ‘not likely’.
Not likely, but don’t discard the thought. See another A330 Incident where unreliable airspeed (pitot probes freezing) resulted in an apparent change in altitude (assumed low speed correction factor). Also, note the effects in the erroneous change in TAT (TAT probe freezing). What haven’t been established so far are the effects of ‘rapidly’ changing speed and temperature on other computations in the ADC and possibly in the IRS module.
If the crew’s perception was of a ‘failure’ of airspeed, sudden drops in altitude, and perhaps rapid changes in VS, then concluding that there was a ‘display’ failure (the computation behind displayed parameters) would be logical.
xcitation “Also stated was air temperature increase which indicates updrafts and turbulent weather”.
The ‘increasing’ temperature could have been due to the TAT probe blocking with ice crystals, again see the A330 incident above.
Not likely, but don’t discard the thought. See another A330 Incident where unreliable airspeed (pitot probes freezing) resulted in an apparent change in altitude (assumed low speed correction factor). Also, note the effects in the erroneous change in TAT (TAT probe freezing). What haven’t been established so far are the effects of ‘rapidly’ changing speed and temperature on other computations in the ADC and possibly in the IRS module.
If the crew’s perception was of a ‘failure’ of airspeed, sudden drops in altitude, and perhaps rapid changes in VS, then concluding that there was a ‘display’ failure (the computation behind displayed parameters) would be logical.
xcitation “Also stated was air temperature increase which indicates updrafts and turbulent weather”.
The ‘increasing’ temperature could have been due to the TAT probe blocking with ice crystals, again see the A330 incident above.
Second Law
Join Date: Jan 2008
Location: Wirral
Age: 77
Posts: 113
Likes: 0
Received 0 Likes
on
0 Posts
Steed
1581 Agree
Spurious accuracy does not make one data set innately superior to another
"All time data here rounded to x dp" could be held to be a clumsy use of English and "around" could be held to be held to be more succinct. I personally prefer the former.
I suspect different authors were at work rather than Machiavelli.
1581 Agree
Spurious accuracy does not make one data set innately superior to another
"All time data here rounded to x dp" could be held to be a clumsy use of English and "around" could be held to be held to be more succinct. I personally prefer the former.
I suspect different authors were at work rather than Machiavelli.
http://www.pprune.org/rumours-news/4...ml#post6505825
safetypee:
Roger your post. I had asked, in this thread or the other, about TAT probes being ice obstructed, and you provide me an answer. Report saved FFR. Thanks.
300 feet off. (The ATSB document you linked to) Not quite the altimeter unwinding, is it? 
However, perhaps the point you are making is that seeing alt low a few hundred feet (from an iced TAT and iced pitot probe) prompts pilot to make what is initially a small correction to climb back to FL350? Is that what you were getting at?
But not the attitude indicator, which is what I had in mind primarily when I made my response. Sorry for the ambiguity, I had thought I was addressing how pilots tend to know which systems influence which cockpit gauge ... but I guess I did a poor job.
ADIRU includes IRU. IRU does not use airmass data to display attitude. Have I missed something?
I may have asked that in the other thread. Thanks!
safetypee:
Roger your post. I had asked, in this thread or the other, about TAT probes being ice obstructed, and you provide me an answer. Report saved FFR. Thanks.
Vovachan; lonewolf, re “assumed ALL instruments were suspect/wrong”, but ‘not likely’.
Not likely, but don’t discard the thought.
See another A330 Incident where unreliable airspeed (pitot probes freezing) resulted in an apparent change in altitude (assumed low speed correction factor).
Not likely, but don’t discard the thought.
See another A330 Incident where unreliable airspeed (pitot probes freezing) resulted in an apparent change in altitude (assumed low speed correction factor).
However, perhaps the point you are making is that seeing alt low a few hundred feet (from an iced TAT and iced pitot probe) prompts pilot to make what is initially a small correction to climb back to FL350? Is that what you were getting at?
If the crew’s perception was of a ‘failure’ of airspeed, sudden drops in altitude, and perhaps rapid changes in VS, then concluding that there was a ‘display’ failure (the computation behind displayed parameters) would be logical.
ADIRU includes IRU. IRU does not use airmass data to display attitude. Have I missed something?
The ‘increasing’ temperature could have been due to the TAT probe blocking with ice crystals, again see the A330 incident above.
Join Date: Jan 2008
Location: UK
Age: 69
Posts: 292
Likes: 0
Received 0 Likes
on
0 Posts
Lonewolf 50
Not all seeds are immediately obvious. Nearly impossible to disprove, parts of the system soft/firmware might have been altered either by mistake or or purpose, maybe with the deliberate intention of causing this crash. Now that systems behave according to rules laid down in software they must be a prime candidate for interference by those who intend harm. Perhaps it hasn't yet happened somewhere, but for certain it will.
Oh, yes, it HAS happened. Erebus. TE901 November 1979. And, as Gordon Vette pointed out, there were a number of factors that came together to turn an incident into a tragedy.
If you look at it from the perspective of Team Air France, the flight failed to do as intended: deliver plane and pax to Paris. Would a more useful form of that question be: "where were the seeds of mission failure sown?"
I suggest that's the driving question behind BEA's difficult task.
I suggest that's the driving question behind BEA's difficult task.
Oh, yes, it HAS happened. Erebus. TE901 November 1979. And, as Gordon Vette pointed out, there were a number of factors that came together to turn an incident into a tragedy.
Lonewolf_50, TAT probe freezing.
This is (was) a ‘relatively’ common problem occurring on several aircraft types. It was most prevalent on the BAe146 which suffered engine roll back due to ice crystal icing in the compressor – all engines since modified.
The most usual erroneous TAT indications, as with the A330 incident, are where the temperature stabilises at zero (ice or water ice) due to the blockage. The temperature change is relatively slow in comparison to the assumed ‘near instantaneous’ pitot blocking.
“…seeing alt low a few hundred feet … prompts pilot to make what is initially a small correction to climb back to FL350?”
Yes, very much my point; this would provide a reason for the initial nose up input, but not answer the subsequent control actions.
Amongst this we should consider that the FDs had probably failed (cf VH-EBA), thus the crew might be looking for a cued parameter to follow. Also, consider what the VS might show after an ‘instantaneous’ 300 ft descent and attempted recovery (inertial / air data mix IIRC).
What happens to Mach No? cf VH-EBA.
What happens to MMO/VMO computation? It’s most likely to be independent of the airspeed display as the value is also used by the FWC for audio warning; would the warning still work?
As for IRU, I knew of one design which had an air data correction in the unit (I forget why), but during testing, a sharp pull up with a rapid loss of speed and quickly increasing altitude caused the ADC software to detect a ‘failed’ (out of tolerance) condition and shutdown both ADCs, which in turn shut down both IRUs. The system design was changed before certification!
The point is that ADC seems to get into many dark corners of modern systems, thus we might not be sure of what the effects of simultaneous ADC malfunction might be.
This is (was) a ‘relatively’ common problem occurring on several aircraft types. It was most prevalent on the BAe146 which suffered engine roll back due to ice crystal icing in the compressor – all engines since modified.
The most usual erroneous TAT indications, as with the A330 incident, are where the temperature stabilises at zero (ice or water ice) due to the blockage. The temperature change is relatively slow in comparison to the assumed ‘near instantaneous’ pitot blocking.
“…seeing alt low a few hundred feet … prompts pilot to make what is initially a small correction to climb back to FL350?”
Yes, very much my point; this would provide a reason for the initial nose up input, but not answer the subsequent control actions.
Amongst this we should consider that the FDs had probably failed (cf VH-EBA), thus the crew might be looking for a cued parameter to follow. Also, consider what the VS might show after an ‘instantaneous’ 300 ft descent and attempted recovery (inertial / air data mix IIRC).
What happens to Mach No? cf VH-EBA.
What happens to MMO/VMO computation? It’s most likely to be independent of the airspeed display as the value is also used by the FWC for audio warning; would the warning still work?
As for IRU, I knew of one design which had an air data correction in the unit (I forget why), but during testing, a sharp pull up with a rapid loss of speed and quickly increasing altitude caused the ADC software to detect a ‘failed’ (out of tolerance) condition and shutdown both ADCs, which in turn shut down both IRUs. The system design was changed before certification!
The point is that ADC seems to get into many dark corners of modern systems, thus we might not be sure of what the effects of simultaneous ADC malfunction might be.