AF447
My QHI had a gentle titter.
sevenfive;
In any fatal accident, there are always interested parties for whom the existence of data is "inconvenient" and perhaps even harmful to their interests. Safety investigations go where they must of course, regardless of "harm".
I certainly am not alone in experiencing this kind of ostrich-like corporate behaviour when confronted with data that paints an ugly operational picture. Denial's a powerful thing.
However, it is the nature of the process under ICAO Annex 13 that a thorough investigative process will, notwithstanding the capacity and abilities of the investigators in charge under the leadership of the IIC, unfold in a way that will at least ensure that reasonable outcome which is non-conspiratorial in nature and goes where it will regardless of who or what it harms, will obtain.
I am well aware that such outcomes are somewhat ideal and that investigators can "get it wrong" in parts. There are plenty of dissenting views and assessments which go to this point. But that is different than "hoping the recorders won't be found", which I think, while it is not an unrealistic statement, goes to unethical and dangerous thinking.
I was wondering when "Habsheim" would arise in this thread again...
Maybe they dont want to find the recorders...
I certainly am not alone in experiencing this kind of ostrich-like corporate behaviour when confronted with data that paints an ugly operational picture. Denial's a powerful thing.
However, it is the nature of the process under ICAO Annex 13 that a thorough investigative process will, notwithstanding the capacity and abilities of the investigators in charge under the leadership of the IIC, unfold in a way that will at least ensure that reasonable outcome which is non-conspiratorial in nature and goes where it will regardless of who or what it harms, will obtain.
I am well aware that such outcomes are somewhat ideal and that investigators can "get it wrong" in parts. There are plenty of dissenting views and assessments which go to this point. But that is different than "hoping the recorders won't be found", which I think, while it is not an unrealistic statement, goes to unethical and dangerous thinking.
I was wondering when "Habsheim" would arise in this thread again...
Too Close To The Coffin Corner?
I'm not a widebody pilot, and in fact I've never had my hands on anything bigger than a B-47, (steam gauges and steam powered controls) so I'm asking this in all innocence: Are ABs routinely being flown too close to the coffin corner? If so, is it for reasons of cost? Do they need to change their cruise procedures to avoid perilous circumstances?
Join Date: Jan 2002
Location: USA
Posts: 286
Likes: 0
Received 0 Likes
on
0 Posts
From what I saw in the French TV news, I assume the highlighted seat chart indicated seat assignments of paxs recovered. It would appear they were random locations, but mostly the right side of the airplane, beginning in FRCL and progressing rearward. Is there a more detailed visual of the seating chart.
Last edited by wes_wall; 25th Jul 2009 at 23:22.
Smilin',
No, they're not flown too close to "coffin corner", precisely because of the high cost in fuel flow. That said, one takes what altitude one can get on the ocean crossing. Taking a high altitude early protects one's position while taking one too low may not permit one to climb later, depending upon traffic. Westbound, er, no, eastb..n...yeah, eastbound out of Hong Kong/Beijing etc across Japan to N.A. can be a problem. I think out of Hong Kong with the 340's hair dryers we always had to cross ELATO (if I recall the name correctly) at 290 because of weight and even that was a stretch. Three hours into the flight however, reduces the weight and improves the MAX ALT and Optimum Alt substantially.
FWIW, I don't think, (never did) that "coffin corner" had anything to do with this accident. The airplane's flight planning package, (fuel flow, weights, temps) and digital presentations (the actual lowest and highest speeds available are presented right on the Primary Flight Display) and the cost of fuel all mitigate against climbing too early in the flight notwithstanding all the above - FL350 is not an extraordinary altitude for the 330 at that point in the flight - pretty normal stuff, really. The stall speed and the Mach buffet speed at that altitude would be, guessing, about 100 kts apart perhaps slightly less, say 190ks to 290kts IAS, (the stall charts in the AOM only go to FL200 and there is no Max Dive speed indicated but maintenance isn't even interested in overspeeds until they're 20kts above the VMO).
Are ABs routinely being flown too close to the coffin corner? If so, is it for reasons of cost?
FWIW, I don't think, (never did) that "coffin corner" had anything to do with this accident. The airplane's flight planning package, (fuel flow, weights, temps) and digital presentations (the actual lowest and highest speeds available are presented right on the Primary Flight Display) and the cost of fuel all mitigate against climbing too early in the flight notwithstanding all the above - FL350 is not an extraordinary altitude for the 330 at that point in the flight - pretty normal stuff, really. The stall speed and the Mach buffet speed at that altitude would be, guessing, about 100 kts apart perhaps slightly less, say 190ks to 290kts IAS, (the stall charts in the AOM only go to FL200 and there is no Max Dive speed indicated but maintenance isn't even interested in overspeeds until they're 20kts above the VMO).
Last edited by PJ2; 26th Jul 2009 at 06:04.
Stall Speed vs. Mach Buffet
The stall speed and the Mach buffet speed at that altitude would be, guessing, about 100 kts apart perhaps slightly less, say 190ks to 290kts IAS,
Join Date: Jun 2009
Location: I am where I am and that's all where I am.
Posts: 660
Likes: 0
Received 0 Likes
on
0 Posts
PJ2, with regards to "maybe not wanting to find the recorders" perhaps we should start a real live rumor that there was secretly about $30,000,000 in gold aboard the plane.
JD-EE (Did *I* say that? I deny everything!)
JD-EE (Did *I* say that? I deny everything!)
Smilin'
I don't know. My AOM, (not for a 330 with center tanks), shows an optimum weight for a 4000ft climb from 310 to 350 for a M0.80 cruise as 237k kg's. I believe they were about 205k kgs if I recall so there is lots of room between stall and Mach buffet speeds - Mach buffet isn't an issue here nor is a stall between 0200 and 0214Z. We cannot say what happened after.
My "100kts" is an educated guess and not from the books, so it's in quotes. It could be 70 to 90 kts but it isn't 30 to 40kts. A spurious stall warning would likely be recognized because of the absence of buffet and air noise, (if the warning was false) but there was no stall warning in the time frame we have information for.
In my view that feeling is unfounded. A 205k kg A330 at 350 even in heavy turbulence is a stable platform with lots of headroom. Been there.
This being the case, why do I get the feeling that many on this thread are concerned about either stalling or entering mach buffet and experiencing a loss of control?
With a 100 knot difference, I wonder why a spurious stall warning wouldn't be easily recognized.
I get the feeling from many here that flying under these conditions is like trying to balance on a bowling ball.
Last edited by PJ2; 26th Jul 2009 at 02:55.
Stall Warning Recognition
A spurious stall warning would likely be recognized because of the absence of buffet and air noise, (if the warning was false)......
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes
on
0 Posts
It is also speculated that the A/THR had possibly already narrowed the upper aerodynamic margin just before it was auto-disengaged along with the A/P when the cascade of ECAM alerts began.
However, if this is a credible scenario, how would the a/c gain speed? More thrust - that's where any crew monitoring function would/should have (been able to) pick it up?
1) How would you see that increasing Mach as displayed thrust change - or would you the steady rise against varying thr. levels (assuming some turbulence) ?
2) Is there any reporting of discrepancy between Inertial and ASI rates of change (at least in the longitudinal axis) - e.g. unaccountable apparent accelerations
===== On the general drift of inherent dangers at cruise at M .8/.82 @35k & 205 tonne). ===
PJ2 paints a very different picture than the linked (& somewhat scary) analyis from UNCTUOUS. The incipient speed-rise situation above might explain it, as perhaps would extreme turbulence. It does sound like it would have to have been a combination of turbulence + poor speed control, from his judgement on the matter?
Ed; found some M0.82 cruise numbers:
Some reasonable but hypothetical speeds:
At FL350, with an MAC of 37% at 205k kgs, the IAS is 279kts, 271kts for M0.80.
Green dot, high and low speed buffet onset - A330, no center tanks:
Best lift/drag ratio speed is "Green dot", named for its symbol on the Primary Flight Display airspeed tape for 205k kgs at 350 would be 245kts, 26kts below cruise speed.
VLS, (lowest selectable speed) known as "the hook" because of the shape of the symbol, is 1.23 VS in the clean config and would be approximately 236kts or 35kts below cruise speed.
For a M0.82 buffet onset it would take a bank angle of 54deg (in a coordinated turn) to produce a 1.75g load factor for a 330 weighing 200k kg with a 40% CG
The buffet onset charts indicate that at 1g, mach buffet is well above MMO at about M0.87 perhaps a bit less, for an IAS of about 296kts or about 25kts above cruise speed.
The low speed buffet would be M0.58 at 1g, (level flight) or an IAS of about 192kts at FL350 or about 80kts below cruise speed.
...or about 100kts or so between the two buffet boundaries.
singpilot made a salient comment the other day - "trying to get into the cockpit" or something to imagine what was going on. I think a lot are trying. I'm trying to imagine what circumstances would cause the loss of 80kts or so. From 272kts, (roughly) to 192kts, (roughly), it would take a very long time even with engine thrust at idle, maybe a knot every second or two, or about a minute and a half to slow from 272 to 190.
Speed reduction occurs much more quickly with speed brakes out. With the boards up, both the VLS and stall speeds go up with speed brake use. In my experience, I have seen VLS increase by 30kts when the boards are at their maximum deflection, (usually in descent - they are rarely used if ever, in cruise unless one needs to slow the airplane down very quickly - it's not an issue, it's just rare). There is a slight pitch-up with use of speed brakes and a corresponding pitch down when they are stowed. The rate of deployment and retraction are computer-controlled and quite slow, (3-5" for up, 5-8" for down), giving lots of time to adjust pitch attitude.
While the speed calculations are from the AOM and are not unreasonable, what was done with them is entirely speculative and not knowable at present.
HarryMann;
In the article to which "UNCTIOUS" supplied the link, I saw no direct analysis of the A330 performance numbers or direct references to speeds. While I think the article is correct in its statements about "mach crit" and other notions the statements which are generic, do not bear a solid (causal) relationship to the A330 or this accident. A 100kt "distance" between the two buffet boundaries bears this assessment out. I think the statement contained in the article about "surprise" is a reasonable statement, perhaps understatement, but that alone does not account for the loss of control.
The article does make some statements that cannot be verified and are pure conjecture,
"It's becoming apparent that whichever pilot was PF in AF447 misinterpreted the ADIRS symptoms as an aerodynamic stall and added power (possibly also increasing AoA) - with a resultant coffin corner encounter with Mach crit (which rapidly leads to uncontrollable roll and pitch excursions - see definitions below in next box)."
and that, and the absence of any analysis as it relates to the performance numbers of an A330 regarding high and low speed buffet boundaries leads one to suspect other "conclusions".
Nor am I "arguing a case"! I'm just trying to use the best numbers and a bit of thinking to see "what fits" and what isn't reasonable.
Some reasonable but hypothetical speeds:
At FL350, with an MAC of 37% at 205k kgs, the IAS is 279kts, 271kts for M0.80.
Green dot, high and low speed buffet onset - A330, no center tanks:
Best lift/drag ratio speed is "Green dot", named for its symbol on the Primary Flight Display airspeed tape for 205k kgs at 350 would be 245kts, 26kts below cruise speed.
VLS, (lowest selectable speed) known as "the hook" because of the shape of the symbol, is 1.23 VS in the clean config and would be approximately 236kts or 35kts below cruise speed.
For a M0.82 buffet onset it would take a bank angle of 54deg (in a coordinated turn) to produce a 1.75g load factor for a 330 weighing 200k kg with a 40% CG
The buffet onset charts indicate that at 1g, mach buffet is well above MMO at about M0.87 perhaps a bit less, for an IAS of about 296kts or about 25kts above cruise speed.
The low speed buffet would be M0.58 at 1g, (level flight) or an IAS of about 192kts at FL350 or about 80kts below cruise speed.
...or about 100kts or so between the two buffet boundaries.
singpilot made a salient comment the other day - "trying to get into the cockpit" or something to imagine what was going on. I think a lot are trying. I'm trying to imagine what circumstances would cause the loss of 80kts or so. From 272kts, (roughly) to 192kts, (roughly), it would take a very long time even with engine thrust at idle, maybe a knot every second or two, or about a minute and a half to slow from 272 to 190.
Speed reduction occurs much more quickly with speed brakes out. With the boards up, both the VLS and stall speeds go up with speed brake use. In my experience, I have seen VLS increase by 30kts when the boards are at their maximum deflection, (usually in descent - they are rarely used if ever, in cruise unless one needs to slow the airplane down very quickly - it's not an issue, it's just rare). There is a slight pitch-up with use of speed brakes and a corresponding pitch down when they are stowed. The rate of deployment and retraction are computer-controlled and quite slow, (3-5" for up, 5-8" for down), giving lots of time to adjust pitch attitude.
While the speed calculations are from the AOM and are not unreasonable, what was done with them is entirely speculative and not knowable at present.
HarryMann;
The incipient speed-rise situation above might explain it, as perhaps would extreme turbulence. It does sound like it would have to have been a combination of turbulence + poor speed control, from his judgement on the matter?
The article does make some statements that cannot be verified and are pure conjecture,
"It's becoming apparent that whichever pilot was PF in AF447 misinterpreted the ADIRS symptoms as an aerodynamic stall and added power (possibly also increasing AoA) - with a resultant coffin corner encounter with Mach crit (which rapidly leads to uncontrollable roll and pitch excursions - see definitions below in next box)."
and that, and the absence of any analysis as it relates to the performance numbers of an A330 regarding high and low speed buffet boundaries leads one to suspect other "conclusions".
Nor am I "arguing a case"! I'm just trying to use the best numbers and a bit of thinking to see "what fits" and what isn't reasonable.
Last edited by PJ2; 26th Jul 2009 at 03:41.
Join Date: Jul 2005
Location: btw SAMAR and TOSPA
Posts: 566
Likes: 0
Received 0 Likes
on
0 Posts
What a useless chat
There are many cases in which over-controlling is an issue in the AB for exactly these reasons., (I note the thread on the Hamburg accident).
Ed; found some M0.82 cruise numbers:
threemiles;
Well, for a given SAT rise and a maintained Mach and CAS, only the TAS increases. ISA+40 would yield about a 39kt TAS increase but the Mach remains at M0.82, the CAS at 279kts.
ISA+40 is pretty rare and it would seem you're not likely going to be staying at 350 at ISA+40. The Environmental Limitations (ATA90) doesn't list a maximum TAT but the graph shows 350 and a SAT of -20 as right on the FL350 line. The AOM doesn't have cruise performance charts for anything other than ISA and the Optimum weight for 4000ft step climb has ISA+10 and ISA+20. ISA+20 @ M0.82 shows a climb from 310 to 350 at an optimum weight of 226k kg's however. So ISA+40 may not present the serious issues we might have first thought. You were thinking of a column of much warmer air, as per the earlier conversation in re Tim Vasquez' work could possible stall the aircraft? Otherwise, not sure what you're thinking.
Your notion brings to mind an accident that occurred on the Vancouver coast - Fraser river delta, when a heavily-laden fishboat sank, costing five lives, if I recall. The ship was in salt water and when it entered the less bouyant fresh water of the Fraser's outpouring, it became unstable.
Pointless numbers unless you know what the actual ISA deviation was. Can be very very significant in a tropical storm, maybe ISA+40 and then?
ISA+40 is pretty rare and it would seem you're not likely going to be staying at 350 at ISA+40. The Environmental Limitations (ATA90) doesn't list a maximum TAT but the graph shows 350 and a SAT of -20 as right on the FL350 line. The AOM doesn't have cruise performance charts for anything other than ISA and the Optimum weight for 4000ft step climb has ISA+10 and ISA+20. ISA+20 @ M0.82 shows a climb from 310 to 350 at an optimum weight of 226k kg's however. So ISA+40 may not present the serious issues we might have first thought. You were thinking of a column of much warmer air, as per the earlier conversation in re Tim Vasquez' work could possible stall the aircraft? Otherwise, not sure what you're thinking.
Your notion brings to mind an accident that occurred on the Vancouver coast - Fraser river delta, when a heavily-laden fishboat sank, costing five lives, if I recall. The ship was in salt water and when it entered the less bouyant fresh water of the Fraser's outpouring, it became unstable.
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes
on
0 Posts
About stall and AoA
Some reflections to invite comment:
- An airplane stalls when it exceeds a critical angle of attack, regardless of airspeed.
- Stall warning is an important protection against stalling, especially when airspeed is unreliable.
- Pilots should be able to rely on stall warning and should respect it when it occurs.
- Stall warning should be based on angle of attack and should not be corrupted by false airspeed information.
- If there is a need for an additional warning based on airspeed to cater for unreliable AoA, that warning should say something like "SPEED SPEED SPEED" rather than "STALL STALL STALL".
That being said, a stall is normally not catastrophic as long as there is sufficient height available to recover from it. 35000 ft should normally be sufficient.
regards,
HN39
- An airplane stalls when it exceeds a critical angle of attack, regardless of airspeed.
- Stall warning is an important protection against stalling, especially when airspeed is unreliable.
- Pilots should be able to rely on stall warning and should respect it when it occurs.
- Stall warning should be based on angle of attack and should not be corrupted by false airspeed information.
- If there is a need for an additional warning based on airspeed to cater for unreliable AoA, that warning should say something like "SPEED SPEED SPEED" rather than "STALL STALL STALL".
That being said, a stall is normally not catastrophic as long as there is sufficient height available to recover from it. 35000 ft should normally be sufficient.
regards,
HN39
Last edited by HazelNuts39; 26th Jul 2009 at 08:45. Reason: Rephrased 'loss of airplane'
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by PJ2
The AOM doesn't have cruise performance charts for anything other than ISA and the Optimum weight for 4000ft step climb has ISA+10 and ISA+20.
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes
on
0 Posts
ISA +40
Think I'm right in saying that if you suddenly ran into +40, then your MCrit margin would in short term increase, thrust decrease, and stall margin decreases (?)
... until A/P A/T restabilises at cruise Mach.Which would take some time.
Think I'm right in saying that if you suddenly ran into +40, then your MCrit margin would in short term increase, thrust decrease, and stall margin decreases (?)
... until A/P A/T restabilises at cruise Mach.Which would take some time.
Join Date: Jun 2009
Location: in a plasma cocoon
Age: 53
Posts: 244
Likes: 0
Received 0 Likes
on
0 Posts
Data extracted from the France 2 TV news
I have just colored each seats on an A330 plan from the stills. With a thought for each of these. I only numbered 33 individuals. Available in private.
Jeff
- about the CVR/FDR, if they are located in the tail of the aircraft (probably thought to be the safest place in most cases ?), in this case they may have been in the area absorbing most of the impact energy. The flight data are armored but the ULB/pinger (little cylinder on the armored box) does not seem very hardenned: could it be that these ULB were damaged at the impact and have never pinged at all ?
- about the false stall alarm recognition : is it easy to discriminate between the pre-stall buffeting and the effects of turbulence ?
can a rapid loss of 300 ft (like in the Air Caraïbe case) reinforce the idea of an on going stall ?
- about the ACARS not transmitting possible stall alarms : we cannot be certain that these have occurred, but they did in several past cases involving Pitot freezing (with little delay after the A/P & A/THR off), and we know that they could be a natural consequence of corrupted airspeeds (using AoA and airspeeds to generate the stall alarm in alternate law 2). If the "stall - stall - stall" was generated in a similar manner to the other cases, then this may have occured as early as 02:10 or 02:11Z, and the overspeed may have been initiated as early as 02:10Z through the A/THR chasing airspeeds that were roling back (increased/max thrust from 02:10Z).
- about the colored seats: after correction, there are 38 colored seats on the BEA map (38 passengers + the 2 crew members)
Jeff
- about the CVR/FDR, if they are located in the tail of the aircraft (probably thought to be the safest place in most cases ?), in this case they may have been in the area absorbing most of the impact energy. The flight data are armored but the ULB/pinger (little cylinder on the armored box) does not seem very hardenned: could it be that these ULB were damaged at the impact and have never pinged at all ?
- about the false stall alarm recognition : is it easy to discriminate between the pre-stall buffeting and the effects of turbulence ?
can a rapid loss of 300 ft (like in the Air Caraïbe case) reinforce the idea of an on going stall ?
- about the ACARS not transmitting possible stall alarms : we cannot be certain that these have occurred, but they did in several past cases involving Pitot freezing (with little delay after the A/P & A/THR off), and we know that they could be a natural consequence of corrupted airspeeds (using AoA and airspeeds to generate the stall alarm in alternate law 2). If the "stall - stall - stall" was generated in a similar manner to the other cases, then this may have occured as early as 02:10 or 02:11Z, and the overspeed may have been initiated as early as 02:10Z through the A/THR chasing airspeeds that were roling back (increased/max thrust from 02:10Z).
- about the colored seats: after correction, there are 38 colored seats on the BEA map (38 passengers + the 2 crew members)
Last edited by Hyperveloce; 26th Jul 2009 at 16:47.