Putting some things together
 

Gentlemen,

my thanks to oldengmkr for posting some links of high interest.

To the attention of those interested in my preferred subject which is flight controls and their potential for lethal mischief, if I may, I would like to gently suggest reading these highly theoretical, mind-numbing documents :

AG RVS - Computer-Related Incidents with Commercial Aircraft

Byzantine fault tolerance - Wikipedia, the free encyclopedia

http://www.ann.ece.ufl.edu/courses/e...l%20System.pdf

In very short : Byzantine faults are those special failures where faulty data gets propagated. Relevant ?

The general effect of these faults is erratic or unpredictable system behaviour. I am talking flight controls sytem here. Relevant ?

The theoretical response to these particularly obnoxious faults lies in the direction of 'distributed' and 'reconfigurable' architecture for a computer system. The last link points to a paper that was published in October 2009, and produced by a team of scientists and a select few engineers from the thread subject matter manufacturer. Relevant ?

At least, I for one find this concept of Byzantine fault very much relevant to our present mystery. Applied to flight controls, it certainly gives me the creeps.

Svarin

Edit : first link was posted by oldengmkr, thanks again, very enlightening.

SaturnV

I don't disagree, have I intimated somehow she was South/LKP?? At any time??

What I do say is if takata is right, re: the "Direction of Flight", and the indexing of the engines, then AF447 was 'heading' ENE when she hit. This in no way conflicts with the "direction" of the debris, his thesis is workable either orientation of the a/c at impact. The engines will be first down, and if the current didn't alter its direction, (much) then the lighter debris merely overtook (horizontally) the heavy stuff and landed "beyond" it, later. The fact is, since the ENE direction puts the lighter stuff further East than the heavy things, it gives some leeway (timewise) for the light things to travel further than the engines in the horizontal.

I am thinking the pilots deviated left/right of the airway, and at some point beyond it (the LKP), made the decision to return to land back in Brazil, (or other, dependent on severity of the circumstances). Having completed the turn, they were on a heading back when LOC happened, and she went down quickly. We don't know her heading at LKP, merely that she flew through it.

Even if on the airway, they may have turned back, and the "new" heading was WSW, but displaced the width of the diameter of the turning circle (something like 4nm). This pretty cleanly falls in line with the Impact Point.

If BEA is holding fast to "no position" South, fine by me, how is that fatal to the discussion?? Except to say, they may have evidence not released. Oh, and the fact that AF447 was S of the LKP the entire flight, until her passage through LKP.

A/C direction* when hit sea surface
 

In my post #598 i put:

Is the picture showing the debris field an indication of the plane´s TRAJECTORY? ("heading")

Observe i put TRAJECTORY (and "heading" is under quotes)

Posting a similar question again:

The picture:

http://takata1940.free.fr/debris_field1.jpg

Is coherent with a/c hitting surface in the attitude BEA said?

(With a trajectory estimated as approx. 75 degrees, based on BEA pic)

If not, WHY?


(*) DIRECTION necessarily the same of HEADING during unusual attitude

I have read every post on AF447 from day one, so apologies if I have missed the answer to the following question - to those with 330 experience.

You are happily up high in the cruise, on AP, Aft CofG etc..

Time shifts somehow and you are now at 5,000 ft, still in cruise trim on AP, and have to get onto a stable 3 degree approach to land. Would you have time to re-configure the aircraft completely for an SOP landing?

If not, could you get it set up for a GA?

etc...
All the ex Air Inter craft were AoA indicator equipped.
See this this pic of an A320
which has both an AoAI and an HUD.

HN39, quote:
I'm wondering about the possibility that Alpha-Prot engaged due to turbulence at, say, 2:09:50. Too much of a coincidence perhaps?

Can Alpha-protection be activated while the AP is engaged?

Re the NAT-TRK A340, think you and NOD are right to consider the possibility that the initial pitch-up was caused by overspeed protection, which also disconnected the AP. With speed back in limits, Normal Law would be restored. As NOD implies, absence of sidestick movement by the PF would allow a 1G (approx) trajectory, with speed decaying (and AoA rising) until Alpha-Prot was activated.

The report highlights the plethora of aural warnings (overspeed continuous-repetitive-chime plus TCAS resolution call), which it says would have initially caused the AP-disconnect "cavalry charge" to be inhibited. There is no separate W/L for AP disconnect. It uses the same Master W/L as the overspeed warning.

EDIT
I should add that there are other indications of AP status:
(1) the right-most FMA (flight-mode "annunciator") on each PFD;
(2) the green LED on the respective AP push-button (or absence of) on the FCU (glareshield).


PS
Thanks, Lemurian, re AoA indicators on all Air Inter A/C, and the link to what is otherwise my old office...
Thanks, ChristiaanJ: just checking you were still paying attention. ;)

Putting some things together
 

Svarin,

Your valuable post makes me remember "Black Swan" of Nicolas Nassim Taleb (NNT), Professor of Risk Engineering at NYU 's Poly Institute.

May be a more complex way to talk about Murphy´s law

PS

And using a phrase Bearfoil put in Post #597:

"Nature is relentless in her punishment of those whom are ill-prepared."

PS2

There is a link of a NNT paper on Fukushima tragedy.

I think that the original comment by Jean-Paul Troadec, was intended to mean that the crash position was not south of the LKP. By that I believe he, or anyone else for that matter, had no idea where the aircraft was going after the final AOC report at 0210z.

The point I am making is, that as posited by takata, the aircraft may well have been airborne for some time after 02:14:30z, - and could well have ventured south of the LKP. On the other-hand, as pointed out by the MAK/IAC, - dives, spirals and spins can result in rapid loss of altitude and the impact position is often close to the point of upset. I don't think we have determined whether the UAS (if it was) resulted in an immediate LOC, and therefore it is difficult to understand that an effective distance of less than 5NM was covered between 0210z and the Impact Point, at whatever time that occurred.

In the debris map, which way is North ?

Edit:oK, thanks

Chris;
I must apologize for airing a couple of thoughts which I quickly withdrew after re-reading the AAIB Bulletin, the Fcom and the AF447 ACARS messages.

The NAT incident shows that Alpha-prot can be activated with AP engaged, but causes AP disconnect. Therefore this scenario doesn't fit the AF447 ACARS messages, which indicate that AP disconnect was immediately followed by reversion to ALT law.

In the NAT incident M=0.882 was reached at 14:21:43, but the pitch up began at 14:21:50 when Alpha-prot law was engaged. Once Alpha-prot is engaged, it remains active until a nose-down command is given with the sidestick (see the FCOM and the AAIB Bulletin). Until that occurs, "the AoA protection law seeks to hold the angle of attack constant at alpha prot".

The vertical railway lines are spaced 5 meters apart. They run north/south and north is at the top.

Thanks Svarin for the interesting post.
In your first link, the article previous to that was very enlightening and sounds similar to what the AF447 crew may have faced.
AG RVS - Computer-Related Incidents with Commercial Aircraft
It relates to an eerily similar scenario on a B777 off Perth in 2005. (faulty airspeed indications, autopilot 'doing the wrong thing' etc)
Here is a snippet fron the page....Makes you think about FBW and 'Byzantine fault' propagation and / or tolerance...

Help!
 

Bonsoir Lemurian,
I'm very glad to see you here tonight, Sir!

Could you, please, explain specifically to our friend Bearfoil, what Air France procedure should have been during this particular flight concerning the pilots flying legs.

I remember your explanation showing that the commandant de bord should have taken the first leg and could have been relieved shortly before the first incident happened.

What would be the theoretical timming on the flying deck replacements?
Or, am I wrong?

Heading
 

Scpeculating on the a/c heading when she hit the ocean ... any bank left or right would yield a sharp turn, right or left.

Hi Chris,My poor eyes are still able to distinguish #1 (up) & #2 (bottom) the way I pictured them. The possibility is that those numbers may have been used for wreckage identification rather than port/starboard correspondance if they were unable to indentify their correct place under the wings at this early point.

Pylons, capots (or part of them) are definitively elsewhere -in Toulouse- as they were recovered floating -it was pictured in the BEA annexe to the second report showing those aircraft parts already recovered.

Or, they were not numbered for this purpose (possibly from the early high scan passes as this document dated from the REMUS search before the first ROV dive)...

A very rough model of the wreckage spreading at the bottom related to surface ending place will take into account three categories of wreckage:
1. pieces buoyant;
2. pieces buoyant for some time, then sinking;
3. pieces going down immediately.

* Category 3 should be a very close carbon copy of the surface impact spreading (before sinking) considering what little time it will take before reaching the deep where the currents are barely nil.
* Category 2 should drift on surface as long as it takes to lose all its buoyancy, then sink also mostly straight down. Some pieces may consequently end at some distance from their initial place at impact.
* Category 1 should be recovered far away if spoted at the surface several days later.

Overall, this is not very surprising that the engines are lying at the edge of the debris field (excepted one large piece) considering it is presumed that they were the first to hit the surface.

That was true at the time, but things have been slightly (and discretely) modified since, maybe partly due to that event.
The text in the FCOM is a bit ambiguous but says that it is now also possible to leave the AoA protection if the sidestick is at neutral or pushed forward, for at least 0.5 seconds, when alpha is below alpha prot.

Nothing is that simple with the airbus, a very complex machine indeed.

We'll be told (or not ... ) if some bizarre coincidences mixing unreliable airspeed indication + protections played a trick on AF447 ...

Possible model to explain debris field shape
 

Looking to the debris field of pictures below,

The model could be of a/c hit surface at approx. 75 deg. direction:

1) Eng #1 hit surface, detaches and dive decelerating from a/c trajectory speed, reducing it´s horizontal speed gradually.
2) LH wing separates, breaks and it´s "water speed braking" capabilities (due it´s still present profile despite broken) decelerates faster than fuselage.
3) Fuselage breaks and its pieces "showers" to ocean floor. It´s parts decelerating as a function of it´s characteristics (mass and geometry) creating the debris shape.
4) Eng #2 detaches, "fly" some meters or so and dive reducing it´s horizontal speed gradually (traveling a litle bit further than #1)
5) RH wing separates, "fly" further than eng #2, hit surface, breaks and dive accordingly it´s "water speed braking" capability (RH LG mass and shape of still attached parts and panels)

http://takata1940.free.fr/debris_field1.jpg

http://www.seattlepi.com/mediaManage...628&height=471

IMO is more logical the a/c trajectory of 75 deg than 255.

Sorry to have introduced an error in that legend but the western "wing" position is false (and I have no clue about which one is the RH or LH).

In fact, it seems that one wing is largely broken in smaller parts, east of the squared area, all over the square, and also west of it. I'm trying to roughly compile that from several other frames from those three videos.

I'm not sure also if the Main Landing Gear (like I captioned it) is really with the southern wing part. On the other hand, the Nose Landing Gear (NLG) is also broken in several parts roughly in the middle of the squared area and north of it.

Supposing a/c hit water with some horizontal speed (*)
 

Hi, Takata

Still supposing they hit surface first:

1) When they separate from a/c IMO they would travel further than other a/c pieces. Reasons:
a) A/c horizontal speed
b) it´s inertia (mass)
c) Less "water braking capability" (better hydrodynamics)

Hence it seems to me that they are in the "end" of the debris field farther from the point a/c hit surface.

* And no 'important" local submarine currents, what is probable due the debris concentration.

PS

My objective now is to have some insight on the trajectory of a/c at the end of it´s "flight" from FL350. I am anxious to learn at least how far the debris field is from LKP. I understood they did not revealed yet the recovery ship position. Who knows it´s PSN?

Pylons, capots (or part of them) are definitively elsewhere -in Toulouse- as they were recovered floating -it was pictured in the BEA annexe to the second report showing those aircraft parts already recovered.


I can't seem to find the report. Could somebody please link me!

Thanks