Thanks for the clarification, mm - language problems as always - spit!

PJ - the only comment I have is my surprise when I was firmly informed by a couple of BA guys that in a lot of cases 'breaks' are rostered and mostly not interfered with.

Probably not your way or mine, but 'it' appears to be out there.

BOAC;
Well, there may be some justification for this approach, based in the research on fatigue, sleep patterns and brain activity and the fact that none of us are the best judge of when we are fatigued or too groggy (after a deep-REM sleep during the break - we know that is to be avoided) to be in the seat right away. So I can understand if an organization has rostered breaks based upon these understandings if that's why and how it's done.

That said, just like any scheduling activity, by definition such does not take into account variations in circumstances and I think, within reason, crews should have the freedom to alter rostering schedules, followed by an air safety report so that reasons can be trended and modifications made to such rostering requirements.

Diversification

Yes, the Arm failed while in tension, ostensibly "picking up" Stress produced in the Rudder's "Down Axis". Yes, again, for the Rudder to have endured its damage from the tail cone, the vector would have been opposite, in the "up" vector, as the Rudder was pinned to the upper Fuselage and the VS kept going. This holds true with the Rudder in a variety of deflection angles. So the "prying off" theory needs another look.

It may have been weakened by corrosion, these Hinges are open to the Air. Also, the Bolt seems to have remained in its position, meaning that there was little tension damage (if any) to the remaining Arm. Also notice that the Pivots of the Hinge, the two plates bracketing the joint from the Aft Spar of the VS, seem unscathed. There is no deflection or cracking of these plates nor the pin carriage on the Rudder. The Hinges must remain oriented in a straight line, if there is deviation from straight, the Hinge can "Bind" and cause problems with the actuator.

Something does not square in the description of Rudder Hinge failure. If the Vertical Arm breaks before the Hinge does, it is apparently not intended to prevent catastrophic failure and Rudder separation, and if the Rudder Hinge is not damaged, why then is this Arm present? Alas, the Arm failure occurred at the place of minimum material, so by interpretation, the length of the Arm that is composed of additional Metal is overengineered. One wonders why the Bolt/Arm mate is not built with additional structure, eg, an "Eye" (a circular pan with its hole in center of sufficient diameter to preclude a failure here; the Hinge end of the Arm is tapered, and I assume the "Barrel face" is radiused to conform to the circumference of it's round bed).

The simple way to see this is to say that in spite of the horrific load the VS and Rudder endured, it was the VS that failed, and separated from the a/c, not the Rudder. Also if there was sufficient energy to shear the VS, that same energy failed to separate the two surfaces. As you say, the bit of Arm that snapped off looks like a corrosion crack/fail.

The engineering consideration of the VS join is lateral energy, obviously for the ability to handle very serious loads imparted by the Rudder. The collateral consideration is simply to remain with the Airframe. BEA'S theory of vertical failure happens to pinpoint a stress for which the assembly is built exceptionally strong, but not as a result of consideration of failure in this direction. It is a bonus that derives from the VS' layout atop the Fuselage. Given that there was upset, loss of control, and impact in the Water, I simply maintain that it is strange, no, highly strange that the vertical Stabiliser was torn out virtually "over itself" and in a mode that is resisted in the VS structure here, with the presenting energy defying great odds. Possible.

bear

If it is not been said before, there was a gap of two weeks before a nuclear sub was sent to locate the recorder. Given the approx thirty days of recorder life a major error.

4Greens

The Nuclear Attack Submarine(SNA) "L'Emeraude" was on location 9 days after the crash, and remained for another 30 days. From where-ever the sub came from takes time, and primary functions such as storing and crew changes etc.. need to be made before such a commitment can be made.

That's accepting that there may well have been other "lurkers" in the area earlier - such is the intrigue of international espionage.

mm43

Bearfoil, "So the "prying off" theory needs another look."

Definitely.

AF 447
 

Cause? Read the report done by KIRILL MINOVALOV flying his Citation X into Moscow on a stormy night. He got away with it. Google it!! He has just given evidence at the inquiry in Paris and it sounds very very interesting.

Arm 36 g again
 

After a lot of thinking about how the badly corroded end of arm 36 g could have played a role in the accident, I now think that it is irrelevant. It points more to a potential danger in future flights, had the aircraft escaped this time. A total failure of both sides of this arm could have a rather bad effect on rudder movability. Probably Airbus will have to look into the inspections required in order to find if other aircrafts have the same corrosion problem.

PJ2 (page 20)

Dear PJ,

In relation to your first comment, that a loss of the VS would necesarily result in a high forward speed impact, I doubt that this is sustainable as an outcome. On loss of the VS/Rudder assembly, the aircraft will be at best left with marginal lateral stability, and at worst will be divergent in yaw, which we have seen in the failure of AA587.

In relation to the second comment re pilot handling, do you not consider that a design certification criteria and subsequent design that can be overloaded by sequential torsion-bending loads within the limits of the rudder force (ratio) limiting system is not a factor rather than merely pilot handling? Note as well, that the functional design approach taken by AI on the rudder limiting effectively increases the sensitivity of the rudder in respect to the operators input. ( the system limits throw of control, and thereby rudder deflection, not reducing rudder deflection but maintaining the control throw). Further, the AA587 loss highlighted the massive misunderstanding of operators generally as to the application of Va.

In the event of a compound ADC failure, all bets are off as to control gain and FBW stability systems, and the potential in moderate severe turbulent conditions to result in a control system initiated structural failure remains possible. This is not limited to pilot only control inputs, as the FBW system would until reverting to direct law be acting in 2 channels to retain attitude. Whether this did in fact occur to 447 remains an open question, but is probably inconsistent with the BEA's observation of a failure of the VS/rudder on water impact. I personally find that opinion, and the available data inconsistent, suspect that the loss of data airborne and the preceding system degradations indicative of structural failure airborne, with VS separation a potential consequential failure following the ADC and turbulence encounter.

Rather than having a high speed impact following an inflight VS separation, I would be expecting a relatively low speed impact, potentially with moderate vertical rate. Note that even with JAL103, which retained some lateral stability, the KCAS varied enormously as a consequence of the phugoid, from near stall to high speed. On AA587, the VS failure resulted in divergence in yaw (with beta dot much higher than encountered during the control input phase pre VS separation), and rapidly the separation of both engines under excessive lateral load of the pylon attachments. The aircraft at this point had no thrust, hydraulics, and was pretty much limited to uncontrolled directionally unstable descent as a falling object with path modified by the remaining aerodynamic appendages operating at varied relative airflows.

Just a thought...

FDR

Having thought a bit about how an "intact" hull would react on impacting with the water in the following configuration:-

Pitch Attitude +5°
Bank 0°
Rate of Descent 9,120 ft/min (90KTS)
Ground Speed 50KTS
Heading +10° relative to direction of flight probably in a flat spin clockwise

Initial contact of the empennage would have caused a rapid reduction in pitch, the engines (if still attached) would bite into the water, and as the main wing span impacted, the g forces between the engines and their pylon mounts would have resulted in them shearing off and heading straight down.

Deformation of the empennage and the fuselage aft of the wing will have resulted in a vertical fracture at this point, and similarly to the fuselage forward of the wing. These fractures would have opened fairly cleanly as the result of impact g forces rapidly reducing to zero, then followed by the reciprocal moment due to initial buoyancy and the water rebounding and overshooting its original level before dampening down over a few more cycles. The forward, middle and aft sections of the fuselage all acted differently, with the largest g forces being sustained by the wing section, due to its much bigger volumetric area.

If you need a little help in imagining this, then get a small round plastic pill bottle, add about third of it's volume in soft brown sugar (gets rid of any free surface effects if water is used) to it and replace the cap. Lay it on its side, tap it to ensure the sugar settles along its length, then drop into a bucket / bath etc.. of water on its side and watch what happens!

This accident bears no relationship to the Ethiopian 767 ditching in 1996, where forward moments were reduced by the left wing slicing into the water first, catching a coral reef, and the aircraft rolling and breaking up in the swell.

The point I am making in respect of AF447, is that hitting terra firma in the configuration described above, would have resulted in an impact crater, major disintegration of wings and fuselage followed by fire. This didn't happen, and the reciprocal moments generated by this sort of impact with terra oceania need to be taken into account when discussing the manner in which the vertical stabilizer could have left the empennage anchor points. Not to put too finer point on it, the seat belts didn't cause the damage they are reported to have done due to the impact acceleration to zero, but the substantial reciprocal moment caused by the sea and initial buoyancy, will have. Just think of being belted in on the wrong side of a slingshot that had been tensioned down, then released.

mm43

wizele . . . if you're a pilot, you should know that journos often misinterpret aviation jargons; and Der Spiegel is hardly an aviation journal.

The "Bordeaux" en route alternate airport is just for planning purposes. It doesn't mean that if the flight couldn't make it to CDG that it must land at BOD.

If you study your big atlas you will see that there are many airports in NW Africa that could be used for diversion if there were a significant over-burn and adequate fuel to BOD became an issue.

In any case, pilots do not forego weather deviations to conserve fuel. :ooh:

Deformation
 

MM43, presume the empennage slowed rapidly on contact, rest of structure carries on progressively being decelerated, rear fuselage shears early on owing to the differential accelerations with VS detaching some time in the sequence owing to differential accelerations between it and the (residual) fuselage. Rear section then sinks taking its witness marks and downward twisting with it. Are there any signs of compression at the base of the VS ?

mm43

With respect, your last entry is perhaps rather too harsh on the 330 airframe. In recent years, given the sparse quantity of accidents to assess, one can come up with parallel impacts in which the Tail/Vertical stayed together. This presumes, of course that the speeds are reasonably accurate, as you quote: 100knots vertical, 50 knots horizontal. I believe these speeds are quite low, that the impact involved a great deal more energy than is commonly perceived.

Turkish. In a Stall, the 737 struck the turf Tail first, and the fuselage pancaked and separated into the sections you describe.

Colgan. Here, there is a strong parallel to 447's "impact". Stalled, pancaked, destroyed in subsequent fire.

DFW. Caught in shear, the 1011 hit tail first, broke up and burned

UAL 187. At 180 knots, the DC-10 hit in a "landing" aspect, broke up and burned.

Consider that all these a/c impacted in similar fashion. Tail first, with a velocity of >50 knots, and with varying vertical speed.

In all of the above, the Tail survived attached to the fuselage.

447's impact (as presumed) is not in itself radically different from others.

With a horizontal component, there would be some "bounce, skip" of the Tail, immediately reducing the full force of the vertical, iow, it hit at an angle.

Look at the mounting bed of the VS per 587. Singed but intact. This 300 hit nose first at high vertical acceleration, the VS/Rudder, and Fuselage underneath the VS were intact, though not attached.

447: I wouldn't discount the energy involved, it was almost certainly higher than proposed by BEA. "High vertical acceleration". As a conclusion, it is a poor one, it is a guess without fact, a statement w/o supporting evidence, unless one adds two guesses up to equal a "Finding". I will include Mr.Optimistic's point re: Compression/VS, it is a sound one.

Too much depends on which areas of the aircraft impacted with wave tops. 3 meter waves depicted.

http://i337.photobucket.com/albums/n...30waves3mz.jpg

So Bearfoil
 

Isn't everyone on this thread dancing around the very frightening notion that the VS separated in flight due to unexpected aerodynamic loads or pilot-induced loads? Such would certainly produce a flat spin. How could this happen?
I can accept one incident (587) in which pilot inputs break the airplane (though I don't buy this explanation at barely above maneuvering speed), but two, with the same sort of failure?
I am a former round-motor driver who finds himself SLF on the 330 at least once a month on non-aviation related business, and am regrettably not an engineer. I like the Airbus' comforts and hold no grudge for EADS nor any prejudice in favour of Boeing, whose 737 rudder hard-over issue was neatly covered up. And I confess I couldn't navigate myself around a glass cockpit for 2 minutes. So, ignoramus, yes... but!
When the second Lockheed Electra L-188 shed a wing in flight the manufacturer pulled out all stops, including enlisting NASA, to discover the cause (an unfortunate harmonic in whirl mode that broke the main spar). They fixed, at their own cost, every L-188 in revenue service. The speed with which Lockheed responded was commendable.
Now we have at least two Airbii that have died in-flight with vertical stabs found a bit of a distance from the rest of the airframe. Odd coincidence here: the errant L-188 wing was found a bit of a distance from the rest of the airplane.
Lockheed swallowed their pride and fixed their mess in quite a public way. Will EADS do the same?

Not at all:=

There are a whole bunch of that don't dance but stand square shoulder against drawing conclusions prematurely without data.

Lot's of folks try "what'ifs" but lacking the ability to convince the majority these also fade into the darkness as we all await another tidbit of real data rather than hypothesis alone.

If you understand investigative techniques please feel free to contribute. If you don't then don't be swayed by someone elses arguments on the internet.:ok:

fdr;

Thanks for your input and observation. Your posts on PPRuNe may not be many but they are worth reading. I may have some B777 questions for you.

With loss of the VS, (I am imagining your scenario at high altitude, not lower down) I wasn't thinking a longitudinally stable trajectory as a scenario in a high speed descent - it isn't possible without lateral stabilization. I would expect strong roll due to yawing, loss of the engines due yaw/roll and a high speed descent of what remained of the airplane. For me, the parts that remained of the cabin structure, (from cockpit to rear galley) either indicated an intact-fuselage as described in the BEA report or a mid-air break up of the fuselage. Perhaps your scenario, (low speed descending spin after loss of the VS at high altitude, I think?) may have obtained. I don't think so, but that doesn't matter; as with every scenario, it is at this point where we must wait.

As far as AA587 goes, I know there is substantial controversy regarding the conclusions and reasons for the loss of the vertical stabilzer and that APA has provided significant input. My thinking and therefore my opinion on the design of the AI VS is unimportant as I am not an engineer and cannot validate design. I believe however that this fashion of attaching the VS to the fuselage is the way it is done throughout the industry.

Regarding use of the rudder, I can offer the following:

I have to say that I knew very early on that use of the rudder on a jet transport was discouraged except to keep the aircraft straight during an engine failure and to align the aircraft with the runway in de-crabbing during a cross-wind landing. Otherwise one kept one's feet on the floor.

In the very early 70's, "Jet Upset" did not exist as a concept or a cause of an accident and so was never discussed. In any later (post 2000 or so), jet upset literature we had access to, (we were not, and to my knowledge, still are not, taught recovery from jet upset in the simulator), use of the rudder was specifically warned against.

As a result of all this I would never have expected the use of the rudder in the manner indicated in the AA587 data, not because I thought it might break the VS but simply because one just never did that with the rudder in a jet transport...I never thought there were ever any circumstances in which it was required and while I've certainly experienced wake turbulence at low and cruise alititude, use of the rudder to control the airplane just wasn't a consideration, not from AA587 on but right from the beginning of the career. I think that I wasn't unusual in this understanding.

That said, I wasn't on the flight deck of AA587 so don't know what they faced.

I read with interest both bearfoil's and mm34's input, especially the notion of a sacrificial rudder vs the loss of the VS. The merits of such notions may already be under examination, we don't know.
As has been pointed out somewhere, (can't recall), the PRIM failure on the QF A330, which did cause a violent manoeuvre (but did not break or damage the airframe) was not the same failure as was recorded in the series of ACARS messages. While I am not qualified to say that ADC failure would not behave similarly and move flight controls violently, would an ADC system failing in designed-for manner command airframe-breaking control deflections? Regarding the limitation on the actual rudder, vice the pedals, I believe the rudder itself is limited but it will take some time go go through the documents and manuals I have to determine same. I know there have been discussions on various aspects of this limitation in the other thread.

My one question, and I must emphasize that this and all comments regarding what happened after the loss of control are, for me, far less important than what caused the loss of control in the first place; I realize that structural break-up may be part of the latter point but given intact cabin material, some of it pristine, I question any high-speed scenario which also includes the loss of the VS before impact.

That is the fine thing about this forum. Informed and thoughtful participants bring their theories for others to respond to and we get to think on things we had heretofore possibly never considered.

PJ2

Can we draw breath here in lomapaseo's 'dance', and perhaps mm, who's contribution here has been superb!) could save me treading loads of pages and refresh us on the locations/distance apart of the tail empennage and the other debris? This 'in-flight separation' which is being toted here would most likely produce a wide separation, would it not?

Only a vertical distance, in the sense of top of the ocean and bottom. 447 VS was found in the middle of the debris field (see back of first BEA report). 587 was found a few miles back from the (land) impact point.

There are two airbus water impacts that I know of where the VS has floated free - 888T off Perpignan and 447. In the former case there is no evidence of a VS detachment and witnesses to the (intact) impact. In the case of 447 there are no witnesses, but finding the VS floating free doesn't mean it came off before impact (as 888T proves)

The failure mechanisms shown by the recovered VS in 587 and 447 are also different. 587 failed as the VS was torn from the mountings, whereas in the 447 case, those same mountings were ripped out of the rear fuselage with the VS. I haven't been able to find good photos of 888T VS, but what there is seems to show it looking more like 447 than 587 (as in detached with parts of fuselage).

To me, none of this proves the VS was still attached at impact, however it does point in that direction, and there is no evidence pointing to it detaching at altitude - at the moment that remains conjecture.

See eg. http://www.pprune.org/rumours-news/3...ml#post5541860

VS is the yellow diamond. See also page 37 of the first BEA report, and also annex 4 of the same. In short - VS was found in the debris field, not separated. The wide spread of debris in the overall diagram is down to drift due to currents and wind - the spread of items found on each day is less, as would be expected.

Troubling image
 

The recovered galley section didn't show evidence of applied inertial forces anywhere near '36g': the other evidence still has to be accounted for and no doubt the full account will eventually explain all.

[QUOTE-iff789]VS is the yellow diamond.[/quote]-thanks for both posts - saves me a lot of trawling!

Can we now consider in-flight separation at high or medium altitude to be most unlikely and move on?

BOAC,

You may have got the wrong end of the stick. Most long haul flights carry three pilots - ie 2 FOs, and 1 captain. To that end, one of the FOs will be rostered "heavy" on each flight. This is not set in stone, and is negotiable, if for some reason, someone wanted to swap on the day (tired/commuter etc.).

The role of the "heavy" FO is to do any odds and sods the operating crew want him to do (walkarounds/security checks), and then go to sleep for the first leg of the journey. He then replaces each of the operating pilots, in each of their seats, whilst they sleep, and retires, at top of descent.

The "rostered" breaks are done on the day, so not set in stone, either, although all will be given a perusal, to confirm that they are fair.

This whole "the captain would never leave 2 FOs in charge in the ITCZ", is frankly, bollocks. Why not? They've all got ATPLs, and thousands of hours, and the captain needs his rest. If they're not capable of using a weather radar set, and avoiding thunderstorms, then frankly, they shouldn't be there. But the default position is that they can.

Hi Slickie - I asked this question at the time a fair way back when 'Captain on the flight deck' was being discussed and I was quite specific/careful with the question to 2 (retired) long-haul Captains with reference to ITCZ crossing and AF447, and they confirmed they would not expect to alter the rostered breaks (emergency situations excluded, of course, which this particular ITCZ crossing should not have been). With no 'rostered break' experience I can only go on 'reliable' info. Maybe it has changed in the last few years - I guess they could have been described as 'old-school' guys:)?

I agree with your last para, naturally. It shouldn't really matter whether the Captain was 'resting' or not. My post was releated to PJ's post and I think both he and I would assess the 'abilities' before taking a rostered breal with a 'busy' ITCZ ahead.

IF789's post (currently #460), which begins "only a vertical distance...", is an excellent short summary of all that can be said about the VS -- from what we (ppruners) currently know.

As PJ says, we have vast (combined) experience on this site and it can make for interesting reading -- the kind that results in one re-examining the "known" data and then re-examining one's own current bias. That said, IF789's post reflects where any investigator would currently be, if he/she had only the info we have.

Good and thought provoking stuff from many regulars the past few days. Thanks!

grizz

Mr> Optimistic, unless I misread a posting a very few days ago 36 g acceleration is off the table. It came about due to a mistranslation of a sentence that included reference to a part number that contained 36g.

That aside, the pertinent question would be, "How long was the 36 g sustained?" 36 g for a millisecond would cause severe damage to the particular portion of the plane subjected to that acceleration. That damage would instantly mitigate the 36 g making it much less in other parts of the plane as they compressed the damaged areas.

Either way, 36 g acceleration on impact or something less, you have to deal with energy. How much energy change took place over what length of time. Some parts of the plane will get one acceleration. Other parts will get another acceleration. It's not a rigid bar of unobtanium.

JD-EE
 

Thanks: didn't see that (it is easy to get confused when every argument dips, twirls and returns ).

Well, say the a/c struck the sea at a significant vertical speed.

The deceleration was sufficient to disrupt the metal structure. As one part of a structure gets contact (eg bottom part of lightweight galley 'box') a deceleration/compression wave transmits through the body and it is this wave which slows the rest of the structure. So the thin walls and weak joints of the galley have to transmit a force of the order of the mass of (the yet-to-be-slowed) galley panels times 'x' g, whatever 'x' was. If this force exceeds the strength of the material, or the joints, they will fail.

Now I don't know what the impact speed was or the time to rest (ie to find x) but it was sufficient to disrupt the VS at least. Seems odd the VS couldn't stand it but the galley structure could !

The galley and some other recovered pieces (eg the panel section with jump seats) are interesting witnesses. I wonder if any testing is going on to reverse engineer the impact conditions. If a number of the galley units were dropped at various orientations over water from say 40 feet (~30mph by my rough calc) it would be informative to see the deformations. Then increase in 40 ft increments until a match.

The absence of evidence of secondary impacts or crushing, the remaining integrity of the weak right angle joints, the large panel section with jump seats still intact, don't look consistent with a high speed vertical catastrophic impact. Partially distributed load from fore-aft deceleration maybe but not smashing vertically downwards. Doesn't immediately suggest 'non survivable impact' either. If other evidence (eg medical) suggests otherwise it just adds to the whole puzzle.

AF447 - The Numbers
 

Let's put some numbers on it.

9120 ft/min = 152 ft/sec, and 50KTS = 84.44 ft/sec, resulting 173.88 ft/sec or 53 meters/sec.

The mechanics of determining the time over which the acceleration was reduced to zero is not easily calculated due to the cylindrical shape of the fuselage, plus the area of the main wings and elevators come into play. On top of this, the aircraft is initially buoyant, and the forces cancelled out will reciprocate as buoyancy moments. If the aircraft impacted terra firma, the time taken to dissipate the impact moments would be about 100 milliseconds and the structural damage would be extreme. In the case we are dealing with, the shape and area of the fuselage combined with the large area of the wing will provide a dampening effect and probably the time to accelerate to zero is around 250 milliseconds, with half the remaining velocity being absorbed each 50 milliseconds.

Density of air at sea level and 25°C is around 1.185 kg/m3, whereas sea water is 1,025 kg/m3 or 865 times denser than air. Not quite solid, but at the velocities we are talking about, its close to it. On top of that we have an aircraft weighing in at about 210 tonnes, but the total volumetric area is about 1260m3, and that is potentially the water that could be displaced during impact. The force at impact will be around 210,000kg x 53m/sec = 11,130,000 m kg/s, and that energy either has to be dissipated by the aircraft or transmitted into the water. Water does not compress, therefore the energy gets turned into a wave with amplitude and length, e.g. the stone in the pond principle.

http://i846.photobucket.com/albums/a...330-impact.png

Looking at the force vectors drawn through the V/S, it can been seen that there would definitely be compression on the forward end of the V/S, and a combination of compression on the aft end caused by the THS forcing framing upwards, later reverting to tension at the aft clevis as the cancelling of the forward moment caused the V/S to rotate off in that direction and to port. All these forces will have created their own local tsunami and the effects of that will most likely take a minute to oscillate down to the background sea and swell conditions.

http://i846.photobucket.com/albums/a...0-impact-2.png

Likely points of fracture through the fuselage have been marked, and discussion around the items recovered may help determine if there was another fracture near the aft pressure bulkhead.

However, the initial parting of the waves will result in a violent return of the water, and the wing spar section will pop to the surface, and fractures already formed at its fore and aft ends will be flexed in the opposite direction, causing complete separation of the fuselage ends. Water will invade those ruptured spaces, eventually permeating through linings etc.. and buoyancy will be lost.

What bothers me in the above example, is that if the actual time taken to arrest the impact moments was say 1 sec, then the impact could have been survivable. The potentially large volumetric area of the aircraft is the reason I believe that the time to arrest the impact forces was nearer 250 milliseconds. What little we know of the pathology reports tends to describe spinal and pelvic injuries that point to terminal velocities greater than represented above. So the aim of this exercise is to come up with some velocities that truly represent the damage represented. My feeling is that the terminal vertical velocity was about 30 percent greater than shown. Deformation of the galley sides, toilet doors etc.. was on the narrow sides from bottom to top. In fact the deformation was mostly near the bottom and relatively small in length, which could imply that the "g" forces were high and of a very short duration.

Accident damage patterns
 

I can see there is confusion on the part of some about the characteristics of the damage on AF447. The process of tearing apart an aircraft in a crash is a complex process but there are some generalities.
If AF447 had stacked into the water on its nose, there would be some deceleration of the bits from the rear of the AC by virtue of whatever columnar stiffness the fuselage provided and they would generally be less damaged than bits from the nose, but there is another factor in play that would increase the average degree of damage. The volume of water receiving the aircraft's energy would be much smaller and thus the average energy per unit volume of water would be higher than for a belly flop. This higher energy density would take longer to dissipate and the bits entrained in this volume of water would smash on each other with more energy and for a longer time and break each other down in size more.
BEA says the aircraft pancaked on its belly and this seems reasonable to me based on the damage pictures I've seen. In a belly flop, the energy is dissipated very quickly over a larger volume of water. Also in a belly flop depending on aircraft touchdown attitude, there will likely be a number of fuselage breaks. In this environment, aircraft wreckage will be larger and more varied with pieces from the bottom of the aircraft compressed and/or torn, and pieces from higher in the aircraft experiencing varying damage depending on what is in the immediate vicinity. From the number of large parts, it is clear that this was a low energy impact (for a jet).
In the case of the vertical stabilizer, it is pretty clear that the aircraft structure below it was pretty fractured when it commenced it's independent flight.
And why did it peel off forwards? AF447 probably hit with somewhere around 35 to 40 degrees AOA. Since it was basically in a level attitude, that gives a forward speed over 100 knots relative to the water.
Now this is a "professional" opinion from one who has participated in a few military accident investigations involving jet aircraft impacting both over land and over water and I could be really wrong in some of the details I've postulated. I hope they find the wreckage and settle these details.
From the pictures I've seen of the AF447 wreckage, I don't think anything from around the THS has been found that would indicate its trimmed position. I'm betting it was trimmed full nose up. Does anyone have any info that would bear on this detail?

36 g again
 

The correct translation (see post #440) of the french report shows that 36 g is a part name and not acceleration.

IFF 789
 

Thank-you! I confessed ignorance up front but appreciate the clarification (also from BOAC and a couple of others), and would agree with the lot of you that conclusions drawn prior to FDR recovery are idiotic, even though I ventured one, based on the fact that the 300-series VS floats and so looks to be the obvious culprit even when it ain't.
But I'll stand by my point about the L-188 (Electra). Lockheed turned to everyone (!) which meant NASA, and all their competitors, Boeing, Convair and Douglas, to figure out if something were broke in the design of the thing. With only two makers of big transports left on the planet I wonder of EADS has the same quality or quantity of knowledge pool to draw on than Lockheed had back in the 1960s.
Again, thank-you and I'll shut up now.

Hi,

Air France condamné à payer une forte indemnité à une victime de l'AF447

Start of researches delayed ......

Liberation (French newspaper)
Google translation

Brazilian press :
G1 > Edição Rio de Janeiro - NOTÍCIAS - Air France é condenada a pagar R$ 2 milhões a família de vítima do voo 447

Bear in mind that the upper portion of the radome was found, (Interim report 1, p35-36) and by the photo, while heavily damaged, it isn't fragmented/shattered but in a contiguous section. That it was even found would be remarkable if not impossible in any high forward speed entry.

Search Delay - BEA Press Release
 

Flight AF 447 on 1st june 2009
A330-203, registered F-GZCP

Information, 11 March 2010

Administrative and technical difficulties on departure from the United States, together with unfavourable meteorological conditions, are delaying the arrival of the « Anne Candies » in the port of Recife (Brazil), where it should join the « Seabed Worker » before sailing to the zone to undertake the search for the wreckage of the A 330. The beginning of the sea search operations has been put back accordingly.

The BEA will release further information on Monday 15 March 2010.
-------------------

The "Anne Candies" was last located using the marine Automatic Identification System (AIS) in Flotation Canal, Port Fourchon, Grand Isle, Louisiana on 01 March 2010.

mm43

Machinbird
If you're betting on that, then what is behind your assumption?

mm43

Mr Optimistic
I think you described "shredded airplane" for a plane that is largely composite materials. If that is so consider the shredding to the way the various galley carts and other items, including bodies, exited the plane. The scenario is more like the plane vanished from around the contents.

The postulated angle of attack, however, suggests the whole thing flopped down onto the ocean with very little time, milliseconds, between tail hitting and the rest of the plane hitting.

That, of course, is predicated upon my memory of discussions here indicating that AF447 was largely built of composites which fracture and shred rather than bend, stretch, and then tear.

PJ2 & BOAC; Vertical Stabiliser separation
 

The following is in answer to the posts of PJ2 and BOAC; I doubt that the question of the tail separation being airborne or on water entry will be resolved readily, however it is the domain of the BEA’s investigation, and hopefully in due course will be answered satisfactorily.

The current BEA position is the tail was attached at impact, and the airframe did not depressurise. This may be the case, however there is the matter of the last CMC message that needs to be explained, which otherwise would indicate a potential structural defect in flight. The BEA observes that the O2 masks were not deployed, however also notes that the covers are off some O2 units. (The RTLU state corresponding to normal conditions is surprising given the airdata failure...)

The airdata failure confronting the crew in turbulence is conducive to developing conditions of high loads, including on the tail, and may result in in flight failure. High yaw rates and angles have not only been a problem on AA587 for Airbus, but were the cause of the structural failure for Lauda 1, over Thailand, resulting from the PW4060 being in reverse in flight, resulting in failure of the VS, and of at least one of the horizontal stabilisers through excessive loads. The failure of the horizontal tail in that case caused a rapid nose down pitch change, and a subsequent failure of the wings in negative overload.

A feature of the Airbus tail is that the secondary load paths in the VS will fail from the same loads that cause a primary load path structural failure, due to the lever arm that is generated at the secondary yoke attachment. While this is interesting where the failure of the primary is due to a sub design load component failure, where the loads applied exceed the ultimate design loads, this is moot; any system will fail in that case.

With a commencement altitude of FL350, if the aircraft remained within the normal envelope until the loss of data, the data loss would occur above around FL300. If the operational envelope was exceeded, then any altitude above sea level would be possible. A possible point of structural failure is at the time where pressurization faults are noted. If a VS failure occurs around this time through one assumes oscillatory divergent torsion-bending, then the aircraft would be commencing the event around Vs-Mmo, approximately 180-280Kts roughly. The Thales pitot tube failure of a single channel would be recognised as a degradation of a PRIM & or SEC computer, and the control laws would be degraded. An identified failure of all pitot tubes would result in the control laws going to direct law. In all failure cases, rudder limiting degrades, which increases the potential for developing high loads on the VS.

If a failure of the VS occurs airborne, the aircraft yaw divergence will result in a very high deceleration rate, rough guess is in the range of 2-3g, developing to around 40-60kts sec-1 reduction. the yaw will result in roll, however induced roll lags yaw, but will be significant, and a rapid roll would develop, until IAS has washed off, at which case the aircraft will likely oscillate in all axis due to local flow effects. If the engines are thrown off the pylons, the yaw divergence will be reduced at that time. The wings and fuselage loads may be within design loads, but would be still fairly high.

The VS if separated in flight as a single component, has the potential to develop lift through auto-rotation about its center of mass, and will not just freefall. The likely offset of the VS from the remainder of the airframe impacts is able to be modeled, but is complicated by ocean drift of the VS post impact. Roughly, for a separation at 30-35000', the impacts CEP's would be in the range of ~3-5nm +/-2 (very rough estimate). The VS will drift at near the current rate, with little effect from wind. The engines if separated at high altitude and moderate to high forward speed, would have a ballistic trajectory, modified by drag variations from tumbling, but would have a CEP of around ~5-6.5nm +/-1, and a flight time of a ~46-52 secs, with a very high water entry speed. Low density components will remain close to the point of separation, with wind determining their final impact. Wind will not affect the engine CEP significantly. The fuselage and the VS time of flight is going to be ~4-8 minutes for the fuselage, and the VS around 10-15 minutes.
The hydrodynamic effects post water entries are dependent on the structural integrity of the component. The engines will sink near vertically post water entry, but the fuselage if retaining some integrity may well sled away from the entry point to its final resting place.

In respect to jet upset, in approximately 1997, as a result of USA427 and UA585, and other events, both Boeing and AI worked on a jet upset program. This program included a comprehensive video presentation, which included advice on application of rudder for roll control in high AoA conditions, due to the reduced effectiveness of ailerons at high AoA. This advice was and remains reasonable, however, the flight control characteristics of various jets means that there is a potential for excessive loads to be generated. The worst case for flight control application is a rapid control reversal, and where coupled with high angular rates, the resulting forces are extreme, where oscillatory inputs are made, the cyclical bending-torsion loads are basically divergent, as long as the control inputs are maintained. The VS and horizontal stabiliser are the most susceptible surfaces, as they are a short rigid structure. The ailerons of cycled, will result in large flexing of the wings, which will absorb some of the loads. On the B767, B747 and B777, a very small amplitude doublet pulse of the ailerons will feel very uncomfortable, but is fairly well damped. At high altitude damping is reduced, and for the B767, I know of one case where an accidental small amplitude doublet was made which caused a very poorly damped oscillation to occur, resulting in the autopilot tripping eventually.

The training programs supported by AI and Boeing were appropriate, but the application of this training is problematic. While aerobatic training may be assumed to assist in assessing what is “judicious” in relation to a control input, the majority of aerobatic aircraft do not have swept wing effects or even large dihedral effects, tending towards low lateral-directional stability. Military aerobatic aircraft which do have swept wings, have significantly different B/A inertial coupling effects to a commercial jet…

Until the submissions of AI were received by the NTSB on AA587, the potential to develop divergent responses in the tail from moderate control inputs below Va was at best poorly understood, or little known, or probably completely unknown. The generic advice to pilots to treat their aircraft kindly with control inputs doesn’t imply knowledge existed that the empirical effect on the structure is known, it just means it is wise to avoid unknowns where possible. To place that in perspective, an unlimited aerobatic aircraft with rapid cyclical control reversal of a rudder may well result in structural loads that exceed any testing or design. The timing of the control input and the aircraft response is important in determining whether divergent response occurs.

Page 197-198 of NTSB/AAR-99/01
1.18.9.2 Postaccident Activity

On August 16, 1995, the FAA disseminated Flight Standards Handbook Bulletin
for Air Transportation (HBAT) 95-10, entitled “Selected Events Training” (SET), to its principal operations inspectors (POI). The HBAT contains “…guidance and information on the approval and implementation of ‘Selected Events Training’ for operators training under 14 CFR Part 121, who use flight simulation devices as part of their flight training programs.”

The HBAT states that the SET is “voluntary flight training in hazardous inflight
situations which are not specifically identified in FAA regulations or directives.” Some of the examples of these selected events include false stall warning in rotation, excessive roll attitude (in excess of 90°), and high pitch attitude (in excess of 35°). The HBAT further states that the SET program was developed jointly by the FAA and the aviation industry in response to previously issued Safety Board recommendations addressing the need for unusual events and unusual attitude training for Part 121 and 135 air carrier pilots.

In 1996, USAir implemented SET as a required recurrent training element for all
of its pilots. The training program at USAir included simulator training in recovering from nose high, nose low, and inverted airplane attitudes. Also, many air carriers began implementing SET/Advanced Maneuvers Package programs patterned after the guidelines of the FAA’s HBAT 95-10 and United Airlines’ program, respectively.

On October 18, 1996, the Safety Board issued Safety Recommendation A-96-120. This recommendation asked the FAA to require 14 CFR Part 121 and 135 operators to provide training to flight crews in the recognition of and recovery from unusual attitudes and upset maneuvers, including upsets that occur while the aircraft is being controlled by automatic flight control systems and unusual attitudes that result from flight control malfunctions and uncommanded flight control surface movements.

In a January 16, 1997, letter to the Safety Board, the FAA stated that it was
considering an NPRM proposing to require that air carriers conduct training that will emphasize recognition, prevention, and recovery from aircraft attitudes that are normally not associated with air carrier flight operations. In its July 15, 1997, response, the Safety Board stated that it was not aware of any training programs that specifically addressed unusual attitudes that resulted from a control system failure or for which some flight controls would not be available, or would be counterproductive to, the recovery. (This recommendation is discussed more fully in section 1.18.11.5.)

In a November 2, 1998, letter to the FAA, the Safety Board listed those safety recommendations, including A-96-120, for which no recent action had been taken by the FAA. In a January 13, 1999, letter to the Safety Board’s Director of the Office of Aviation Safety, the FAA’s Associate Administrator for Regulation and Certification stated that “14 CFR part 121, subparts N and O (Training Program and Crewmember Qualifications, respectively), are being extensively rewritten. The rulemaking is expected to contain specific requirements addressing the NTSB’s concerns.” (See section 2.7 for the Safety Board’s review and evaluation of the FAA’s action in response to Safety Recommendation A-96-120 and the recommendation’s current classification.)

During 1997 and 1998, a working group composed of representatives of aircraft manufacturers, air carriers, pilot associations, training organizations, and government agencies (including the FAA) developed the Airplane Upset Recovery Training Aid. This publication and video program provided background information for air carrier pilots and managers on jet aerodynamics, stability, control, and upset recovery. The training aid also provided a model curriculum for classroom and flight simulator training in recovering from unusual flight attitudes. As of late 1998, the Airplane Upset Recovery Training Aid publication and video program were being distributed by two major air transport manufacturers (Boeing and Airbus) to their customers. This training aid, however, does not include simulator training in unusual attitudes resulting from flight control malfunctions and uncommanded flight control surface movements.

NTSB/AAR-99/01
Recommendation #6:
Require all 14 Code of Federal Regulations Part 121 air carrier operators of
the Boeing 737 to provide their flight crews with initial and recurrent flight
simulator training in the “Uncommanded Yaw or Roll” and “Jammed or
Restricted Rudder” procedures in Boeing’s 737 Operations Manual. The
training should demonstrate the inability to control the airplane at some
speeds and configurations by using the roll controls (the crossover airspeed
phenomenon) and include performance of both procedures in their entirety.
(A-99-25)


The NTSB's recommendations were sound, as was the intent of the training that followed, but there was missing information in respect to the certified design. The training (version 1) possibly didn't result in sufficient attention to the sensitivity of the rudder although this was mentioned in passing, and the potential for a structural failure was not fullt comprehended.

The figures are rough assessments only, but the Cd of the VS is going to approximate 1.0 in unsteady flight, about the same as a parachute... the engines Cd are dependent on the structure that may separate with them and the resultant stabilised attitude, (I get the engine stabilising in an oscillatory exhaust forward attitude after mid flight...) around Cd of 0.15-0.6. The airframe Cd is dependent on final attitude, but is going to be around 0.5-0.8 or thereabouts.

Whether the airframe is is one piece is moot; the control problem confronting the crew in adverse weather/IMC was sufficient to lose flight path control, and has at the least ended in impact of the aircraft at relatively low energy state with the water. If the aircraft remained intact, then a recovery without airdata is problematic from a stalled condition, and requires at the very least the use of other data such as AoA (ATT-FPV). Flight crew are not trained to recover using such data, and IIRC on the A330 you would need to change MCP modes to get the FPV displayed, whereas on the Boeing it can be selected separately. AoA also may be displayed full time on Boeing's by operator pin selection. (I gave an ex F16/Hawk IP a UA with unreliable AS recently, and the recovery was not achieved until such time as he was talked through using the ATT/FPV cues to detect he was in a stalled condition while indicating a Mmo overspeed. Once provided the cue, the state recognition and recovery was possible. The difficulty of properly identifying and rectifying such a failure in turbulent conditions cannot be overstated, IMHO)

"doubt is not a pleasant condition, but certainty is absurd"
Voltaire (1694-1778)

mm43, I am not sure from reading your scenario and from the picture that you factored in the mass and viscosity of the water. It takes time for the water to get out of the way. I rather envision the plane was somewhat flattened by the time you have the plane in the trough it makes in the water. That would violently compress the air in the cabin adding to the shredding effect on the body of the plane.

Water hit at 50 MPH is not too unlike hitting a wall. It does get out of the way. But that takes enough time to break things that hit at that speed, especially with a er ah "belly flop."

That said, I'm with you in regards to the impact being enough to cause the pelvis and back injuries cited. I'd expect it to have hit closer to 100MPH vertical and horizontal both. But I suspect BEA has the engineers and data to make the reverse calculations leading to their estimates.

Why the vertical stabilizer didn't come off in flight.
 

It would sure be nice if someone would run an actual dynamic simulation on youtube of a swept wing airliner missing all of its vertical stabilizer at cruise airspeed. It would be very educational.
Swept wing induced roll rate is very powerful with even with relatively low yaw angles creating significant roll rates. I would anticipate that an airliner in this situation would randomly but inevitably develop significant yaw, and with any angle of attack would rapidly build roll rate. The increasing roll rate would cause an increase in angle of attack and yaw which would further build g while increasing the roll rate further. As angle of attack increases, the fuselage gets more out of line with the roll axis, and centrifugal force is the driver of increasing angle of attack and yaw. In seconds, g loads would exceed structural limits of the entire airframe.
The only way you might hope to avoid the immediate breakup would be to zero angle of attack (but that has its own bad consequences if held for very long).

JD-EE
Well, in a way I have, but haven't shown compression on the bottom side of the fuselage. Though I have implied that when I said that energy involved at impact has either got to be absorbed by the aircraft or it gets imparted into the water in the form of a wave with amplitude and horizontal length - corresponding speed and vector of the energy inputs, i.e. the medium is not compressible. Check again and you find I made reference to the air/seawater densities and the potential mass to be displaced. Newtons 2nd Law works only in a vacuum and consequently varies with the medium density - 1/865 in the air/seawater example.

Photographic evidence would indicate that vertical moment at impact was much greater than the horizontal one, but that could be misleading when the area presented in the vertical plane was many times that in the horizontal plane. This factor alone needs to be taken into account in dealing with the forces involved, and perhaps the the entry vector needs to be broken up and the accelerations to zero reassessed based on the square area presented in each plane.

It really boils down to - a bigger surface area present at impact results in higher rates of acceleration to zero, equaling higher g forces. The high-board diver enters the water quietly and with only a ripple, a belly flop creates a large noise and a splash (also a hurt "belly").

PJ2 has mentioned the radome top half was found virtually intact, and that could even be partially explained away by the predominate force acting on it was upwards and the bottom half was shattered by the equipment within.

Machinbird has ventured that the AoA was in the region of 35 ~ 40 degrees. The AoA used in my earlier example was 61 degrees and as I have already indicated, the numbers using a vertical moment of 90KTS and a horizontal moment of 50KTS do not result in the "g" forces I perceive were involved.

I'll run some other numbers, but need to consider what sort of drag vortex could result in whatever terminal velocities we are dealing with. Just doubling the horizontal moment to 110KTS gives an AoA of 42 degrees, which results in an impact velocity of 69.2 m/sec, or (69.2m/s / 0.25sec) / 9.8N = 28.2g. Now that is no where near a high enough "g" force to cause the damage it did. So, I am now looking at not only the upward thrust on the underside of the fuselage and main wing spar, but also the drag factor (lift) induced by water flowing over the top of the wing. Comment has been made by the BEA suggesting that it was an upward vector that tore off the Port Outer Spoiler and attached framing from the wing, but what was happening back at the wing root could well be another matter. The THS would probably be responsible for over-stressing the framing in the empennage.

My initial calculation assumed that the a/c was arrested in about 7 meters along the entry vector, but on this later consideration I am more inclined to believe it was only half that, and the forces involved on the intact aircraft were around 28.2 * 2 = 56.4g.

I'm still inclined to believe the terminal vertical moment was more than 90KTS, more likely 120KTS, and when combined with the 50KTS horizontal moment would give an AoA of 67.4 degrees and an entry vector of 130KTS, or (66.9m/s / 0.125sec) / 9.8N = 54.6g. Note: I halved the time to zero acceleration.

BTW, in an earlier post I mentioned that the deformation of the fuselage on impact would probably have reduced the cabin volume and avoided an implosion due to the low cabin pressure. A very minor point at that stage.

As always, I am liable to be found wrong, so feel free to challenge my assumptions.

mm43