AF 447 Search to resume
bearfoil;
It was observed and stated somewhere shortly after the crash in 2009 that there wouldn't be sufficient dynamic pressure above FL200 or so, to cause the VS to fail as it did and that upset would be the only event which could produce sufficient forces which means loss of control could only precede any loss of the VS before impact.
Another way to examine it is, (and this too, has been said before, by myself and others in the thread), if the fin broke off due to upset, entry into a thunderstorm etc, then almost certainly loss of control at high altitude followed by a high-speed impact would have been the result and the available evidence does not lead us to such a conclusion. If the loss of the VS occurred lower, then such loss was not the precipitating event.
The wreckage in the photographs is "relatively" intact, even "pristine" when compared to that of Swissair 111 after an impact of 300kts. The recovered wreckage, mostly from the cabin, is similarly absent of any evidence of high-speed collision with water or other parts during the breakup sequence. Ejected overhead bin parts, the defibrilator (red box) and doctor's medical kit (yellow box), are all in pristine condition, relatively speaking, again, evidence against a high-speed impact.
In fact, the larger parts such as galleys, FA seats, washroom doors etc came from those areas which are major joins in the barrels of the fuselage - just behind the cockpit, just in front of and behind the wing box and just ahead of the tail section aft of the pressure dome.
We can't say any more about it other than that's where the larger pieces came from but it is reasonable to believe that this is how the wreckage will be distributed.
It's been mentioned many times...the track that AF447 took in comparison with others on the same two tracks, especially the number of diversions left and right of track taken by the aircraft behind AF447. I wonder if 123.45 was as busy that night as it usually is on the Pacific and Atlantic?...
It was observed and stated somewhere shortly after the crash in 2009 that there wouldn't be sufficient dynamic pressure above FL200 or so, to cause the VS to fail as it did and that upset would be the only event which could produce sufficient forces which means loss of control could only precede any loss of the VS before impact.
Another way to examine it is, (and this too, has been said before, by myself and others in the thread), if the fin broke off due to upset, entry into a thunderstorm etc, then almost certainly loss of control at high altitude followed by a high-speed impact would have been the result and the available evidence does not lead us to such a conclusion. If the loss of the VS occurred lower, then such loss was not the precipitating event.
The wreckage in the photographs is "relatively" intact, even "pristine" when compared to that of Swissair 111 after an impact of 300kts. The recovered wreckage, mostly from the cabin, is similarly absent of any evidence of high-speed collision with water or other parts during the breakup sequence. Ejected overhead bin parts, the defibrilator (red box) and doctor's medical kit (yellow box), are all in pristine condition, relatively speaking, again, evidence against a high-speed impact.
In fact, the larger parts such as galleys, FA seats, washroom doors etc came from those areas which are major joins in the barrels of the fuselage - just behind the cockpit, just in front of and behind the wing box and just ahead of the tail section aft of the pressure dome.
We can't say any more about it other than that's where the larger pieces came from but it is reasonable to believe that this is how the wreckage will be distributed.
It's been mentioned many times...the track that AF447 took in comparison with others on the same two tracks, especially the number of diversions left and right of track taken by the aircraft behind AF447. I wonder if 123.45 was as busy that night as it usually is on the Pacific and Atlantic?...
PJ2...
Your question re "123.45" is interesting. Many of us have postulated that a lack of comms from AF447 to "fellow travellers through the ITCZ" that night is indicative of something. Quite what, we aren't sure of course.
What we can say -- as you have alluded to -- is that exchanges of some sort on 123.45, in that area at that time of day, are the rule rather than the exception. For there to have been no participation by AF447 is, to me at least, a possible hint that things weren't quite "normal ops" up front. I'm choosing my words carefully as I have no idea what the apparent failure to use that resource specifically means. There are so many possible reasons for it: one or more radios inop (we have an indication pre-takeoff of a possible problem in this area); something else on the flight deck taking priority (for a longer time than we may have been supposing); and on and on...
At this point this tragedy must be accepted as a puzzle with too many pieces missing to solve it. All we can hope is that the DFDR and the CVR, along with the aircraft remains of course, will provide the investigators enough pieces to solve this puzzle. I, for one, am optimistic.
Your question re "123.45" is interesting. Many of us have postulated that a lack of comms from AF447 to "fellow travellers through the ITCZ" that night is indicative of something. Quite what, we aren't sure of course.
What we can say -- as you have alluded to -- is that exchanges of some sort on 123.45, in that area at that time of day, are the rule rather than the exception. For there to have been no participation by AF447 is, to me at least, a possible hint that things weren't quite "normal ops" up front. I'm choosing my words carefully as I have no idea what the apparent failure to use that resource specifically means. There are so many possible reasons for it: one or more radios inop (we have an indication pre-takeoff of a possible problem in this area); something else on the flight deck taking priority (for a longer time than we may have been supposing); and on and on...
At this point this tragedy must be accepted as a puzzle with too many pieces missing to solve it. All we can hope is that the DFDR and the CVR, along with the aircraft remains of course, will provide the investigators enough pieces to solve this puzzle. I, for one, am optimistic.
Join Date: Jan 2008
Location: Atlanta, GA, USA
Posts: 349
Likes: 0
Received 0 Likes
on
0 Posts
PJ2,
Your vast experience and knowledge is compelling, but so is the lack of any flotsam around the fin and spoiler panel when they were found. How do you explain that? Surface tension tends to keep such debris islands intact.
Your vast experience and knowledge is compelling, but so is the lack of any flotsam around the fin and spoiler panel when they were found. How do you explain that? Surface tension tends to keep such debris islands intact.
Join Date: Jul 2009
Location: France - mostly
Age: 84
Posts: 1,682
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by bearfoil
An airload that could be responsible for the removal of 447's fin could be drag alone, in overspeed, with an overswept Rudder. Such drag would put a great load on the aft bracket downward into the fuselage
The side to side is captured by a narrow breadth of lug to lug span, the longitudinal by the distance of the saddles fore to aft.
Last edited by HazelNuts39; 20th Apr 2011 at 09:47.
Join Date: Jan 2008
Location: uk
Posts: 857
Likes: 0
Received 0 Likes
on
0 Posts
From the wreckage photos and descriptions, Air-NZ Perpignan crash looks to me to be much higher engery that AF447 - but still the VS didn't "explode".
How would you compare these two crashes (with what info we have) as regards impact energy (and condition of other wreckage), departure of VS before impact or not, and condition of VS after impact ?
Join Date: Apr 2011
Location: UK
Posts: 13
Likes: 0
Received 0 Likes
on
0 Posts
desitter:
Water drag and wind are more important. Since the tail was largely submerged it's going to act like a sea-anchor, while lighter debris (such as lifejackets) are going to be exposed to wind.
Therefore, you wouldn't expect them to stay together.
On another note. Bearfoil:
I'm still looking for evidence of exploding composites. Erm... nope. Still not found any. (Lucky for me or that Reliant Robin could have gone up like a Ford Pinto.)
Bearfoil, you can't have it both ways. The VS shows that the composites were stronger than the metal lugs and yet you theorise that the spoilers would be atomised on impact and leave no trace.
Surface tension tends to keep such debris islands intact.
Therefore, you wouldn't expect them to stay together.
On another note. Bearfoil:
Originally Posted by bearfoil 
In my experience, FRP has a signature failure in high energy impacts that is unmistakable. It explodes.
In my experience, FRP has a signature failure in high energy impacts that is unmistakable. It explodes.
Bearfoil, you can't have it both ways. The VS shows that the composites were stronger than the metal lugs and yet you theorise that the spoilers would be atomised on impact and leave no trace.
Join Date: Mar 2001
Location: us
Posts: 694
Likes: 0
Received 0 Likes
on
0 Posts
If the VS departed during flight, then, and I'm theorizing, it presumably would have been nearer the LKP, and immediately buoyant on the ocean surface. The French overflew the LKP and areas west and east of the LKP (perhaps as much as 10 NM west) at about 1530 hours on June 1, and saw nothing. The Brazilian AF also overflew the LKP on June 1, although their search box was not as far to the west as that done by the French Dassault Atlantique 2. (Have not found times for the Brazilian search.) Brazil saw nothing either.
The question to be answered is why nothing was seen on June 1, but perhaps part of the answer involves the law of buoyancy, and whether objects detached from a relatively intact airplane as it was sinking, and gradually made their way to the surface.
The question to be answered is why nothing was seen on June 1, but perhaps part of the answer involves the law of buoyancy, and whether objects detached from a relatively intact airplane as it was sinking, and gradually made their way to the surface.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
Good question indeed.
BTW .. the VS even if released from the plane from under the surface (any depth) will no take days for emerge .......
If the VS departed during flight, then, and I'm theorizing, it presumably would have been nearer the LKP, and immediately buoyant on the ocean surface. The French overflew the LKP and areas west and east of the LKP (perhaps as much as 10 NM west) at about 1530 hours on June 1, and saw nothing. The Brazilian AF also overflew the LKP on June 1, although their search box was not as far to the west as that done by the French Dassault Atlantique 2. (Have not found times for the Brazilian search.) Brazil saw nothing either.
The question to be answered is why nothing was seen on June 1, but perhaps part of the answer involves the law of buoyancy, and whether objects detached from a relatively intact airplane as it was sinking, and gradually made their way to the surface.
The question to be answered is why nothing was seen on June 1, but perhaps part of the answer involves the law of buoyancy, and whether objects detached from a relatively intact airplane as it was sinking, and gradually made their way to the surface.
BTW .. the VS even if released from the plane from under the surface (any depth) will no take days for emerge .......
Psychologist on Board
The last underwater recoveries, Swissair 111 and TWA 800, were done from land; so, the shrinks could be seen on shore the same day if the need was felt.
It's quite bad enough dealing with a recently dead intact body; in this case the bodies have been on the bottom for a considerable time and I expect the bodies are already colonised by a number of creatures
It's quite bad enough dealing with a recently dead intact body; in this case the bodies have been on the bottom for a considerable time and I expect the bodies are already colonised by a number of creatures
Join Date: Jul 2009
Location: IAH
Posts: 58
Likes: 0
Received 0 Likes
on
0 Posts
Maybe they should have medically qualified personnel onboard, as I think there were seven children and a baby on that flight. I get choked up just thinking about it, eventhough I have seen dead children in the past. I do not envy those on their way to recover the remains and wreckage. Brave people.
PJ2 Wrote:
You're talking about two completely separate
systems. Flight data recorders have nothing to do with the ACARS
messages - that's been the point all along. The recorders don't see
these messages, don't/can't record these kinds of messages, don't
"timestamp" data and don't hold them in memory. That's not how flight data recording works.
From your original post 3640:
The key point is, parameters can't be recorded all at once. If there are
1800 parameters (in binary form) coming into the system, the system must have a way of "listening, parsing and recording". The data frame software is how that process is handled.
And then:
Without getting more complicated, (because it has to if we go any
further and everyone will be asleep), the nature of recording can give
the appearance of a 'lag', when there "may, or may not" have been one.
Which is what I was disagreeing about. A few general data logging and
comms system basics may help here, but apologies if preaching to the
choir :-).
To start, there may be many sources of data/parameters to log, in different
formats (ie: "language") and at different sample rates, (ie: how often a
parameter is examined). The data format, sample rate and accuracy of
measurement depend, of course, on the requirements of each parameter to
be measured.
Sampled data is typically sent across high speed data buses (ie: highways),
at microsecond rates, encapsulated in frames, or "packets". A fair data bus
analogy is a road full of houses, each with a unique address, containing
people with different names. A fair analogy for a frame is a posted letter,
which will have a destination address, the sender's address, perhaps time
and date, with the measured "data" on the page inside the envelope. For
frame based data comms, all the above information would normally be
included in one form or another, as well as checksum data to allow
detection of errors in transmission. (ie: a torn envelope) Where letters can
take an indefinate amount of time to deliver, avionics buses are designed to
provide guaranteed delivery within very strict time limits.
Between the data source and recorder there may be, depending on the
application, a "data aquisition unit". The function of this is to provide a
central "clearing house" for the various sources and to measure and / or
translate to a common format, or language, to send to the recorder. It may
also timestamp the incoming data, and filter or manipulate it in other ways.
In some systems, acquisition and logging are combined into a single
physical unit, but the functionality remains the same. Such systems often
use queueing techniques to avoid date loss at peak rates, but the data
would be timestamped *before* insertion into any queue. This avoids time
displacement errors where the actual data is stored some time later. By
“some time”, I mean milliseconds, not seconds or minutes.
The key thing i'm, trying to get across here is that accurate timestamping
is fundamental to any data logging application such as this. It is always
designed to have sufficient accuracy in terms of measured value and timing
to meet the needs of the application...
You're talking about two completely separate
systems. Flight data recorders have nothing to do with the ACARS
messages - that's been the point all along. The recorders don't see
these messages, don't/can't record these kinds of messages, don't
"timestamp" data and don't hold them in memory. That's not how flight data recording works.
I don't know where acars comes into this, because I understood from your
post that the context was flight data recording ?.
post that the context was flight data recording ?.
From your original post 3640:
The key point is, parameters can't be recorded all at once. If there are
1800 parameters (in binary form) coming into the system, the system must have a way of "listening, parsing and recording". The data frame software is how that process is handled.
And then:
Without getting more complicated, (because it has to if we go any
further and everyone will be asleep), the nature of recording can give
the appearance of a 'lag', when there "may, or may not" have been one.
Which is what I was disagreeing about. A few general data logging and
comms system basics may help here, but apologies if preaching to the
choir :-).
To start, there may be many sources of data/parameters to log, in different
formats (ie: "language") and at different sample rates, (ie: how often a
parameter is examined). The data format, sample rate and accuracy of
measurement depend, of course, on the requirements of each parameter to
be measured.
Sampled data is typically sent across high speed data buses (ie: highways),
at microsecond rates, encapsulated in frames, or "packets". A fair data bus
analogy is a road full of houses, each with a unique address, containing
people with different names. A fair analogy for a frame is a posted letter,
which will have a destination address, the sender's address, perhaps time
and date, with the measured "data" on the page inside the envelope. For
frame based data comms, all the above information would normally be
included in one form or another, as well as checksum data to allow
detection of errors in transmission. (ie: a torn envelope) Where letters can
take an indefinate amount of time to deliver, avionics buses are designed to
provide guaranteed delivery within very strict time limits.
Between the data source and recorder there may be, depending on the
application, a "data aquisition unit". The function of this is to provide a
central "clearing house" for the various sources and to measure and / or
translate to a common format, or language, to send to the recorder. It may
also timestamp the incoming data, and filter or manipulate it in other ways.
In some systems, acquisition and logging are combined into a single
physical unit, but the functionality remains the same. Such systems often
use queueing techniques to avoid date loss at peak rates, but the data
would be timestamped *before* insertion into any queue. This avoids time
displacement errors where the actual data is stored some time later. By
“some time”, I mean milliseconds, not seconds or minutes.
The key thing i'm, trying to get across here is that accurate timestamping
is fundamental to any data logging application such as this. It is always
designed to have sufficient accuracy in terms of measured value and timing
to meet the needs of the application...
Join Date: May 2010
Location: Boston
Age: 73
Posts: 443
Likes: 0
Received 0 Likes
on
0 Posts
syseng68k to PJ2:
There may be confusion here between time stamps,synchronous sampling and lower than nyquist sample rate artifacts
Even with accurate time stamps there can be artifacts such as phantom split control surfaces if the data is not sampled (and time stamped) at the same instant or at a high enough rate to allow full fidelity.
Phantom split control surfaces can be caused by the time (up to 1/2 the sampling interval) difference between A and B samples.
If a short "jog" (less or near sample interval) is sent to A and B only A might show movement with B appearing stationary.
If the data for 2 surfaces are sampled at the same instant the artifact will not happen.
Eihter both will be captured moving or the "jog" will be missed entirely.
Note though that no more "real" data will be available, if anything slightly less since sampling related data with a time offset may help in filling in the time gaps.
From your original post 3640:
The key point is, parameters can't be recorded all at once. If there are
1800 parameters (in binary form) coming into the system, the system must have a way of "listening, parsing and recording". The data frame software is how that process is handled.
And then:
Without getting more complicated, (because it has to if we go any
further and everyone will be asleep), the nature of recording can give
the appearance of a 'lag', when there "may, or may not" have been one.
Which is what I was disagreeing about. A few general data logging and
comms system basics may help here, but apologies if preaching to the
choir :-).
The key point is, parameters can't be recorded all at once. If there are
1800 parameters (in binary form) coming into the system, the system must have a way of "listening, parsing and recording". The data frame software is how that process is handled.
And then:
Without getting more complicated, (because it has to if we go any
further and everyone will be asleep), the nature of recording can give
the appearance of a 'lag', when there "may, or may not" have been one.
Which is what I was disagreeing about. A few general data logging and
comms system basics may help here, but apologies if preaching to the
choir :-).
Even with accurate time stamps there can be artifacts such as phantom split control surfaces if the data is not sampled (and time stamped) at the same instant or at a high enough rate to allow full fidelity.
Phantom split control surfaces can be caused by the time (up to 1/2 the sampling interval) difference between A and B samples.
If a short "jog" (less or near sample interval) is sent to A and B only A might show movement with B appearing stationary.
If the data for 2 surfaces are sampled at the same instant the artifact will not happen.
Eihter both will be captured moving or the "jog" will be missed entirely.
Note though that no more "real" data will be available, if anything slightly less since sampling related data with a time offset may help in filling in the time gaps.
syseng68k, now I understand - my thanks for the explanation on timestamping - it makes sense to me now...the "ACARS" item turned up because of previous discussions on timestamping of messages and I had incorrectly assumed your reference.
MurphyWasRight, thanks - much clearer; - that's what my rudimentary "strobe" example was trying to convey. I know about flight data analysis but much less about data transmission protocols so I appreciate all explanations.
I considered the notion of "lag", (which it isn't, really...), to be important for the very reasons cited and in response to Machinbird's concern about "elevator split"...that what is in the minutae of the data may not re-present exactly what occurred - I've seen such disparities lots of times and the context (did other systems, or the aircraft, exhibit behaviour commensurate with the split?), often supports the notion that (for example) the flight controls weren't out of sync. It is offered as a caution to be mindful of the process of recording and interpreting...that reading recorders is not at all straightforward.
MurphyWasRight, thanks - much clearer; - that's what my rudimentary "strobe" example was trying to convey. I know about flight data analysis but much less about data transmission protocols so I appreciate all explanations.
I considered the notion of "lag", (which it isn't, really...), to be important for the very reasons cited and in response to Machinbird's concern about "elevator split"...that what is in the minutae of the data may not re-present exactly what occurred - I've seen such disparities lots of times and the context (did other systems, or the aircraft, exhibit behaviour commensurate with the split?), often supports the notion that (for example) the flight controls weren't out of sync. It is offered as a caution to be mindful of the process of recording and interpreting...that reading recorders is not at all straightforward.
Salute!
Will keep beating the horse, 'cause when the FDR's are decoded we will be able to see what actually happened with much higher resolution than the "maintenance" messages sent via ACARS. Could and probably will reflect the whole story.
I go with Chris, another systems engineer re: data recording.
Wrote the spec for a data reduction facility, managed it during test missions and had to provide the "system experts/designers", as well as managers, with good representations of what happened and when for a recce system. Unlike space vehicles, UAV's and manned aircraft experience turbulence that lasts for fractions of a second. This poses problems primarily for video sensors versus RF sensors, as you can imagine.
I look at the VS separation accident ten years ago and realize that many parameters were captured/recorded at rates too slow for the stuff I had worked on, but still good enough to determine the cause of the crash. So comparing the FDR data with the ACARS messages will help all figure out what happened.
I am saddened to see some of the frame rates BEA references, as I cannot rule out a similar scenario where pilot inputs may have aggravated the situation, especially in severe turbulence. And BTW, I have flown thru CB's and it isn't fun, tho I wasn't concerned about structural integrity for the Viper and Sluf. Was more concerned with not getting the jet into a stall or spin or inverted, heh heh.
Bottomline for our discussion, way I see it, is to square away the ACARS transmission protocol/timing, etc. with what we shall see with the decoded FDR data.
Data acquisition units are damned cheap - some even tens of $$$ for precision in terms of tenths of a gee, or fractions of a degree of surface movement, or tenths of feet per second at 100 hz for velocities. It's the final "mother" recorder that has to store it all for later analysis. And much data can be acquired and transmitted with tiny messages, like 8 bits to get 256 th's resolution of a parameter. Only thing I ever had to work with requiring ultra-high resolution and timing was vibration data to characterize optical sensor performance.
Great thread here, and I am tickled we are not getting into conspiracy theories and simplistic suppositions concerning the aerodynamic and structural aspects of the event.
Will keep beating the horse, 'cause when the FDR's are decoded we will be able to see what actually happened with much higher resolution than the "maintenance" messages sent via ACARS. Could and probably will reflect the whole story.
I go with Chris, another systems engineer re: data recording.
Wrote the spec for a data reduction facility, managed it during test missions and had to provide the "system experts/designers", as well as managers, with good representations of what happened and when for a recce system. Unlike space vehicles, UAV's and manned aircraft experience turbulence that lasts for fractions of a second. This poses problems primarily for video sensors versus RF sensors, as you can imagine.
I look at the VS separation accident ten years ago and realize that many parameters were captured/recorded at rates too slow for the stuff I had worked on, but still good enough to determine the cause of the crash. So comparing the FDR data with the ACARS messages will help all figure out what happened.
I am saddened to see some of the frame rates BEA references, as I cannot rule out a similar scenario where pilot inputs may have aggravated the situation, especially in severe turbulence. And BTW, I have flown thru CB's and it isn't fun, tho I wasn't concerned about structural integrity for the Viper and Sluf. Was more concerned with not getting the jet into a stall or spin or inverted, heh heh.
Bottomline for our discussion, way I see it, is to square away the ACARS transmission protocol/timing, etc. with what we shall see with the decoded FDR data.
Data acquisition units are damned cheap - some even tens of $$$ for precision in terms of tenths of a gee, or fractions of a degree of surface movement, or tenths of feet per second at 100 hz for velocities. It's the final "mother" recorder that has to store it all for later analysis. And much data can be acquired and transmitted with tiny messages, like 8 bits to get 256 th's resolution of a parameter. Only thing I ever had to work with requiring ultra-high resolution and timing was vibration data to characterize optical sensor performance.
Great thread here, and I am tickled we are not getting into conspiracy theories and simplistic suppositions concerning the aerodynamic and structural aspects of the event.
Guest
Posts: n/a
infrequentflyer789
Hey. Perpignan was a nose plant at flying speed: 14 degrees ND??
We are told 447 was a buttdrag at nominal speed, certainly not flying.
So the impact aspects were totally different. The 320 assumedly lost its VS fully forward, clear of the impact debris, while 447 is said to have landed on its tail, then belly. This is the upshot of my disagreement: 447 drove her VS into the tail cone, then clear?. No evidence is apparent that the skin of the fin spent any time whatsoever in a decelerating trajectory into the mass of structure that is the tail: APU, HS, Elevators, Pressure cap, tankage, etc. Nor is there evidence on the inner perimeter of the torn out fuselage frames of rupture that would have occurred as the energy of the Fin was dissipated into the axis of the fuselage.
machaca
At twenty nine feet tall and three feet thick, The Vertical Stabiliser has perhaps the best strength to weight ratio of any structure on any swept area of the aircraft. Of course it is silly to call it fragile, it is not. However, it has diminished survival futures in a debris field because of its light 'weight' relative to other structures, due to its strength from 'spread' rather than 'mass'. It is susceptible to puncture and asymmetric torsional stress, something it likely would not have seen had it been lost prior to impact. It is reasonable I think to posit that it would survive a flat impact with the water at some terminal velocity in free fall, but would show immediately any puncture or insult from thousands of pieces of structure with which it would have shared the impact site.
Of course FRP does not 'explode', that was a poor term choice. Better "Shatter, or Disintegrates".
chrs
Hey. Perpignan was a nose plant at flying speed: 14 degrees ND??
We are told 447 was a buttdrag at nominal speed, certainly not flying.
So the impact aspects were totally different. The 320 assumedly lost its VS fully forward, clear of the impact debris, while 447 is said to have landed on its tail, then belly. This is the upshot of my disagreement: 447 drove her VS into the tail cone, then clear?. No evidence is apparent that the skin of the fin spent any time whatsoever in a decelerating trajectory into the mass of structure that is the tail: APU, HS, Elevators, Pressure cap, tankage, etc. Nor is there evidence on the inner perimeter of the torn out fuselage frames of rupture that would have occurred as the energy of the Fin was dissipated into the axis of the fuselage.
machaca
At twenty nine feet tall and three feet thick, The Vertical Stabiliser has perhaps the best strength to weight ratio of any structure on any swept area of the aircraft. Of course it is silly to call it fragile, it is not. However, it has diminished survival futures in a debris field because of its light 'weight' relative to other structures, due to its strength from 'spread' rather than 'mass'. It is susceptible to puncture and asymmetric torsional stress, something it likely would not have seen had it been lost prior to impact. It is reasonable I think to posit that it would survive a flat impact with the water at some terminal velocity in free fall, but would show immediately any puncture or insult from thousands of pieces of structure with which it would have shared the impact site.
Of course FRP does not 'explode', that was a poor term choice. Better "Shatter, or Disintegrates".
chrs
Last edited by bearfoil; 20th Apr 2011 at 16:43.
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes
on
0 Posts
Hi,
Despite all show the contrary (so far in official interim reports .. VS separated from plane at impact) a VS separation in flight is very seducing .. as a consequence of the loss of VS in flight is that you get a plane that is no more controllable ... what appears to have been the case for AF447
Despite all show the contrary (so far in official interim reports .. VS separated from plane at impact) a VS separation in flight is very seducing .. as a consequence of the loss of VS in flight is that you get a plane that is no more controllable ... what appears to have been the case for AF447
Join Date: Jan 2008
Location: Atlanta, GA, USA
Posts: 349
Likes: 0
Received 0 Likes
on
0 Posts
jcjeant - that's right - the great unknown in all of this is the behavior of composite structures under atypical stresses - there are many competing models, often mutually wildly divergent, all highly non-linear, of composite failure under stress.
sd666 - not quite sure what you mean about anchoring, but the water moves as a body, any flotsam would be well inside the boundary layer of air at the water interface, and any islands of debris, large and small, would tend to remain together until dispersed by divergent currents of water, not air. The Perpignan fin was just the largest piece in a huge swarm of debris, and I would have expected something similar here.
sd666 - not quite sure what you mean about anchoring, but the water moves as a body, any flotsam would be well inside the boundary layer of air at the water interface, and any islands of debris, large and small, would tend to remain together until dispersed by divergent currents of water, not air. The Perpignan fin was just the largest piece in a huge swarm of debris, and I would have expected something similar here.
Join Date: Jul 2009
Location: SUSSEX UK
Age: 76
Posts: 57
Likes: 0
Received 0 Likes
on
0 Posts
@bearfoil
To add to the many suggestions attempting to answer your question regarding how the VS became detached, it is worth considering just how water impacts vary from those on hard terrain. During impacts with rigid ground the undercarriage, if deployed, absorbs a significant portion of the impact energy, with the remainder being transferred to the stiffest structural members such as the energy absorbing subfloor beams. These are generally designed to progressively collapse in order to limit to G load on the occupants. For a water impact, the loading mechanisms differ significantly. The landing gear is unable to absorb the impact energy and instead the impact loads are distributed as a transient dynamic pressure load over the fuselage skin. This initial absorption by the skin momentarily slows the rate that force is applied to the structural members, and has the effect of inhibiting the buckling process to the extent that energy absorbing subfloor components become ineffective. This is why in recent years Navy helicopters have been designed with additional features to improve their crashworthy response over both hard terrain and water.
If the skin is compromised, and the fluid is uncontained, as is the case with a high velocity impact, constant interactions occur between fluid and structure, inducing continuous changes in the loads applied throughout the crash event. In addition to the previous mechanism, the structure now experiences considerable hydraulic shock allowing large pressure forces to act directly upon the cabin floor and interior and as a consequence, increases the vertical accelerations experienced by the occupants.
Drop tests on passenger sized airframes at NASA Langley Research Center have shown how the cabin section experiences quite a pronounced deformation even for a 10m/s drop. For the impact suffered by AF447, which likely descended at a rate in the region (60 to 80m/s), the deformation would have looked more like that for the Trident accident (Papa India in 1972), where the appearance of the main cabin sections are very much more flattened . Given the likely fuselage distortion under AF447's VS, even with the rear bulkhead support, the tendency will be for the fixings under the VS to suffer serious deformation and fracture. Now, some milliseconds later, add hydraulic shock imparted by HP water meeting the underside of the cabin roof, and one can easily imagine how the downward movement of the upper cabin shell is arrested, while at the same time, the inertia of the VS imposes itself on the support fixings with the resulting failure and separation. The evidence for significant hydraulic damage is apparent in the photo of the fuselage/door section where the inner lining has been scoured away.
What happened next to the VS is open to some conjecture. It has been suggested that the VS flipped forward over the airframe and splashed down in front of the aircraft. I find this scenario less than convincing since that pathology report gives no indication of any forward injuries, which would suggest minimal forward motion. If the VS had just fallen over onto the airframe, then surely there should have been some noticeable damage to the leading edge. Mechanical structures tend to break at the weakest points, which in aircraft impacting horizontally, is for cabin failure forward and aft of the wings, and near door openings. One possible visualisation is for the rear section, around 73/74, to become detached at impact (rotate to the rear), to rapidly submerge, followed by the buoyant VS in turn becoming disentangled and returning to the surface.
With regard to whether or not impact was horizontal or vertical; aside from the relatively intact galleys etc, and the deformation evidence in the recovered wreckage, the pathology tells its own story. In studies on the analysis of the pattern of injuries (Nato study - Cullen et el 1980), it was demonstrated that the pattern of injuries in water impacts strongly correlated with aircraft attitude at the time. In one example, those in the rear of an aircraft incident suffering a tail down impact were likely to be more severely injured than those in the front - the implication here being that a typical passenger carrying aircraft crash is likely to result in either a pattern of uniform injuries or a steady logical gradation in severity (TWA800 being the exception). Given that the 43 recovered bodies were distributed between rows 1 to 42, the forensic pathology would have provided similar clues as to the nature of AF447's attitude at the time of impact - and as the BEA have concluded, was probably horizontal. The fact that 43 appear to have suffered pelvic fractures strongly suggest that they were seated at the time and in their designated positions.
http://rpmedia.ask.com/ts?u=/wikiped..._Felthorpe.jpg
...my strong conviction that the Vertical Stabilizer and Rudder were not on the fuselage at impact...
If the skin is compromised, and the fluid is uncontained, as is the case with a high velocity impact, constant interactions occur between fluid and structure, inducing continuous changes in the loads applied throughout the crash event. In addition to the previous mechanism, the structure now experiences considerable hydraulic shock allowing large pressure forces to act directly upon the cabin floor and interior and as a consequence, increases the vertical accelerations experienced by the occupants.
Drop tests on passenger sized airframes at NASA Langley Research Center have shown how the cabin section experiences quite a pronounced deformation even for a 10m/s drop. For the impact suffered by AF447, which likely descended at a rate in the region (60 to 80m/s), the deformation would have looked more like that for the Trident accident (Papa India in 1972), where the appearance of the main cabin sections are very much more flattened . Given the likely fuselage distortion under AF447's VS, even with the rear bulkhead support, the tendency will be for the fixings under the VS to suffer serious deformation and fracture. Now, some milliseconds later, add hydraulic shock imparted by HP water meeting the underside of the cabin roof, and one can easily imagine how the downward movement of the upper cabin shell is arrested, while at the same time, the inertia of the VS imposes itself on the support fixings with the resulting failure and separation. The evidence for significant hydraulic damage is apparent in the photo of the fuselage/door section where the inner lining has been scoured away.
What happened next to the VS is open to some conjecture. It has been suggested that the VS flipped forward over the airframe and splashed down in front of the aircraft. I find this scenario less than convincing since that pathology report gives no indication of any forward injuries, which would suggest minimal forward motion. If the VS had just fallen over onto the airframe, then surely there should have been some noticeable damage to the leading edge. Mechanical structures tend to break at the weakest points, which in aircraft impacting horizontally, is for cabin failure forward and aft of the wings, and near door openings. One possible visualisation is for the rear section, around 73/74, to become detached at impact (rotate to the rear), to rapidly submerge, followed by the buoyant VS in turn becoming disentangled and returning to the surface.
With regard to whether or not impact was horizontal or vertical; aside from the relatively intact galleys etc, and the deformation evidence in the recovered wreckage, the pathology tells its own story. In studies on the analysis of the pattern of injuries (Nato study - Cullen et el 1980), it was demonstrated that the pattern of injuries in water impacts strongly correlated with aircraft attitude at the time. In one example, those in the rear of an aircraft incident suffering a tail down impact were likely to be more severely injured than those in the front - the implication here being that a typical passenger carrying aircraft crash is likely to result in either a pattern of uniform injuries or a steady logical gradation in severity (TWA800 being the exception). Given that the 43 recovered bodies were distributed between rows 1 to 42, the forensic pathology would have provided similar clues as to the nature of AF447's attitude at the time of impact - and as the BEA have concluded, was probably horizontal. The fact that 43 appear to have suffered pelvic fractures strongly suggest that they were seated at the time and in their designated positions.
http://rpmedia.ask.com/ts?u=/wikiped..._Felthorpe.jpg
gums;
And compact - memory and processing power is cheap.
The expensive parts as you likely know, are the STCs, the cooperation from Boeing and Airbus as the case may be, to obtain dataframe information on older aircraft to be able to verify/update/change DFLs to capture more parameters.
It isnt very often that it is a physical sensor and/or wiring problem, but a lot of this information is proprietary and therefore cannot be legally shared even just to help someone out with a dataframe issue such as conversion factors or knowing where the words are put for a certain parameter. THAT is where the time, money and even political problems lie because the airlines lobby against any legislation which can increase costs without a "direct, demonstrable" benefit.
Refitting DFDAUs/FDIMUs (not even recorders), is an extremely expensive proposition...new DFDAUs in older airplanes can run, all in, over US$60,000 per airframe even though the equipment itself is comparatively inexpensive. Why should we take that route when "88 parameters at the legal sampling rates is legally good enough"? I have heard that a number of times, along with "we'll just take the 'statistical approach' and equip a few airplanes in the fleet." The joke was, at the end of the day we all had the "GE" trademark from the FOQA lab fridge embossed backwards on our foreheads.
Obtaining certification for changes to DFDAU software, which is essentially a licensing matter and not an engineering/certification matter, puts such changes out of reach in terms of making the business case for most airlines. While some know the value of data and support the work wholeheartedly, you wouldn't believe the gyrations some airline managements go through to avoid such "unnecessary" changes. One can't expect a blank cheque, but one can expect the understanding that not all departments can be "profit centers". I'll leave that particular item at that.
Precisely.
The International Working Group on Flight Data Recovery, (link on the BEA AF447 website) submitted their final report on "lessons learned", called "Triggered Transmission of Flight Data Working group". The need was stated quite succinctly in the working group's first report: "The difficulties encountered have raised questions about the adequacy of existing flight data recovery technology, when considering accidents over oceanic or remote areas."
I don't want to discuss AA587 nor the larger issues now outlined here by many contributors, but here is another problem highlighted by yet another accident: - locating the recorders in hostile, essentially inaccessible environments.
If the industry wants high data granularity of the very kind you're describing for SMS and accident investigations, it is from the regulator, not private enterprise under a de-regulated economic environment, that the mandates to do so must come.
I am saddened to see some of the frame rates BEA references,
. . . .
Data acquisition units are damned cheap -
. . . .
Data acquisition units are damned cheap -
The expensive parts as you likely know, are the STCs, the cooperation from Boeing and Airbus as the case may be, to obtain dataframe information on older aircraft to be able to verify/update/change DFLs to capture more parameters.
It isnt very often that it is a physical sensor and/or wiring problem, but a lot of this information is proprietary and therefore cannot be legally shared even just to help someone out with a dataframe issue such as conversion factors or knowing where the words are put for a certain parameter. THAT is where the time, money and even political problems lie because the airlines lobby against any legislation which can increase costs without a "direct, demonstrable" benefit.
Refitting DFDAUs/FDIMUs (not even recorders), is an extremely expensive proposition...new DFDAUs in older airplanes can run, all in, over US$60,000 per airframe even though the equipment itself is comparatively inexpensive. Why should we take that route when "88 parameters at the legal sampling rates is legally good enough"? I have heard that a number of times, along with "we'll just take the 'statistical approach' and equip a few airplanes in the fleet." The joke was, at the end of the day we all had the "GE" trademark from the FOQA lab fridge embossed backwards on our foreheads.
Obtaining certification for changes to DFDAU software, which is essentially a licensing matter and not an engineering/certification matter, puts such changes out of reach in terms of making the business case for most airlines. While some know the value of data and support the work wholeheartedly, you wouldn't believe the gyrations some airline managements go through to avoid such "unnecessary" changes. One can't expect a blank cheque, but one can expect the understanding that not all departments can be "profit centers". I'll leave that particular item at that.
... some even tens of $$$ for precision in terms of tenths of a gee, or fractions of a degree of surface movement, or tenths of feet per second at 100 hz for velocities. It's the final "mother" recorder that has to store it all for later analysis. And much data can be acquired and transmitted with tiny messages, like 8 bits to get 256 th's resolution of a parameter.
The International Working Group on Flight Data Recovery, (link on the BEA AF447 website) submitted their final report on "lessons learned", called "Triggered Transmission of Flight Data Working group". The need was stated quite succinctly in the working group's first report: "The difficulties encountered have raised questions about the adequacy of existing flight data recovery technology, when considering accidents over oceanic or remote areas."
I don't want to discuss AA587 nor the larger issues now outlined here by many contributors, but here is another problem highlighted by yet another accident: - locating the recorders in hostile, essentially inaccessible environments.
If the industry wants high data granularity of the very kind you're describing for SMS and accident investigations, it is from the regulator, not private enterprise under a de-regulated economic environment, that the mandates to do so must come.