AF447

[augustusjeremy]

auto flight, etc

e) she had an FCPC1 reset when a-d conditions were still valid.

Takata,

I beg to differ. I think this is at least controversial.

And I remeber you saying something about failed engines. Maybe that's why Will Fraser is coming up with that AD now.

PJ2,

I understand your point. I thought realignment could be done in flight if you had the other two IRs ok.

Of course if you don't have any NAV information to realign from you will have to land and do it on the ground anyway.

But remember that if a pitch down ever occurred - something equally controversial - the crew would be put in extreme stress - not understanding what really was going on.

Let's also recall that between 2:11 and 2:12 (2:13 ?) zulu that night something happened that caused a brief interruption in the ACARS message transmission.

[PJ2]

Will;

See "PRIM1" specifically. Inop or selected off, the chain is interesting relative to HS movement and the possibility of overspeed leading to upset.

Though some will know this, I think just to be absolutely clear for all, it is the elevator movement, not the "HS movement" which was commanded in the QANTAS upset. Although the horizontal stabilizer movement is controlled by PRIM1, it is a trim function not a flight control function. It is possible that ND trim motion occurred but not only have we no evidence to believe that, but we do not have evidence of a high-forward-speed, high-rate-of-descent impact with the sea, which an uncontained/uncorrected ND HS trim condition would almost certainly cause. (Note: HS trim does have manual reversion backup).

augustusjeremy;

Just so all are clear on IRS alignment procedures, an IRS cannot be aligned when in motion. Its alignment process uses the angular acceleration sensed by the laser gyros and a mathematical Euler pendulum (theoretical center of the earth perpendicular to the aircraft attitude at rest, if I recall - someone can correct me here, it's been a while), to "null" the only acceleration affecting attitude. The accelerations are "resolved" to a position from which the IRU calculates latitude.

I'm only trying to convey a sense of the alignment process and not a blow-by-blow detailed description, (unnecessary), and why alignment can't be done if the aircraft is at all in motion. There is no such thing as "entering one's position" or taking present positions from the other IRSs - the potential for gross errors are obvious. In fact, a good stiff breeze pushing the tail around while the aircraft is parked at the ramp will upset or delay the alignment process as will an inadvertent towing of the aircraft from one gate to another, (it's been done...!). Forgive me for droning on, but I think this technical aspectd needs to be clear.

[takata]

Will Fraser:
Since it is a flight controls issue relative to the computer, and involved similarities with ACARS from 447, I'm not convinced it doesn't obtain.

All we have in AF447 case is "F/CTL PRIM 1 FAULT" @ 0213
It's a bit fast to talk about "similarities" when at this point (ALT2), pitch was only protected to g-loads, and the flight control computer only ability was to correct any pitch imput in excess off structural limits.

See "PRIM1" specifically. Inop or selected off, the chain is interesting relative to HS movement and the possibility of overspeed leading to upset.

I'm trying to give credit to the engineers here, and as well, I'm trying to give credit to the pilots: what would be the pitch setting at FL350? something like +2.5 deg and, even if (very unlikely) the pitch was unexpectidly reseted to 0 (whilch system would do that in Alternate 2?), this would hardly introduce a critical upset outside the pilots skills to recover from. The speed margin would be about Mach 0.04-0.06 under overspeed, and such an overspeed would still be far from the real structural limits (more than 50+ kts). By the way, if an overspeed really occured at one point, I guess that it will show some traces on the tailfin leading edge; then, the BEA will tell us about it but even so, it won't tell us if an overspeed was due to any computer or crew action.

Last, I'm hardly an authority you have to defer in any way. We are only exchanging personnal point of views in a public forum.

augustusjeremy:
I beg to differ. I think this is at least controversial.
And I remeber you saying something about failed engines. Maybe that's why Will Fraser is coming up with that AD now.

Sorry if I may have induced some of us in error, but the possibility of a (double) engine failures seems totally excluded until the ACARS stopped. And by this point, everything would be in direct law (manual) due to the limited power supply. So, it is very difficult to add any computer mis-behavoir after a situation like that would have developped.
I just don't believe in the scenario: "fast fall from FL350 in one piece and four minutes" I may be wrong, but it looks to me like a false lead to follow until we'll get much more facts to explain it.

S~
Olivier

[Hyperveloce]

Initial loss of control

Originally Posted by augustusjeremy

Takata,
But remember that if a pitch down ever occurred - something equally controversial - the crew would be put in extreme stress - not understanding what really was going on.
Let's also recall that between 2:11 and 2:12 (2:13 ?) zulu that night something happened that caused a brief interruption in the ACARS message transmission.

Are you refering to a pich down in the Qantas fashion (under AP control), or to a nose down induced by the flying pilot (under human control) ? The Qantas nose down has probably nothing to do with the AP¨447 case (except for possible weird reactions of the AP just before it throws the towel), but for a good model, I would rather refer to the Air Caraïbe report (of great safety interest). Take the same Air Caraïbe situation, manageable using pich and thrust (or by switching off the A/THR & fixing a N1 value by applying the "severe turbulence" proc. but it doesn't seem it was the case for AF 447) and concentrating on the trajectory, but just assume that the AF 447 took the opposite decision that was made when stall alarms sounded in the cockpit (the stall alarms were ignored by the Air Caraïbe CPT, although the procedure was to take them in account). It is just about making the assumption that the AF SOPs were applied strictly in this case (in a similar context to the Air Caraïbe incident, with stall alarms): what happens next, at high altitude, if the crew attention was also focussed on the failing nav, on the SOPs implementation, on failures very difficult to understand and isolate, and on the AP reanimation ?
Jeff

[Will Fraser]

takata/PJ2

Respectfully, a trim function would have more relevance to "upset".
At 2 +/- degrees, a down trim to "zero" would have a steady and insidious
effect on stability relative to airspeeds. With a fault as addressed in the pre-AD testing, trim wouldn't be correctable, and/or noticeable to pilots or computers wrestling with Unreliable a/s data to begin with. Only a few hundred feet (Qantas) can produce a substantial increase in speed. Any established descent, left alone, could be disastrous at flight limits experienced by 447: high, hot, and heavy.

It isn't necessary to impute any other entry to the Sea than BEA's, following an overspeed recovery. Seven miles of altitude to "recover" in, (Assumed). A high speed recovery in Direct Law wouldn't cost a bit of Elevator, VS or other control surface?

[PJ2]

Will;

Yes, I agree with your comment, (as stated in the earlier post) - trim would have an enormous and initialy insidious effect if left uncontained/uncorrected. The QANTAS accident occurred through elevator deflection, however. Caught quickly, it is controllable through pilot input. I think what takata stated regarding giving credit to the engineers and pilots is the correct view and coincides with my own experience in general - it is an easily controlled set of circumstances if caught quickly - and it's what pilots do - just like the QANTAS guys did. It's just that I don't think that's what occurred here.

[Hyperveloce]

Vertical rate at impact

There are known past cases of losses of control with very high vertical rates: 20 000 fpm, 30 000 fpm, even more (>40 000 fpm) before a mid air break up occured. In the Adams Air case, involved in a 60° high dive, the structural integrity of the airframe was compromised at ~475 kts, under a vertical load factor exceeding +3G (over 40 000 fpm). Those irrecoverable dives lead to short scenarii (<1min30sec: compatible with ACARS being emitted more than 3mn after the multiple FLRs and AP disengagement ?), very high impact speed and incidence angles, very small debris and no body to recover as such. Looking to the debris as shown by the BEA report (more spec. the big furnitures/Galley G2), knowing that the Airbus was in "ligne de vol" when it impacted on its belly (or on its tail, with a slight pich up), knowing that the 30 bodies recovered by the Ventose Frigate were all in a relatively good overall state (with their clothes), it seems difficult to think that the impact occured in such a way (with such high vertical rates). If we need a longuer scenario (>3mn), maybe we have to assume that the control initially lost at high altitude was somewhat regained but either not enough, or overcorrected ? and that the max. vertical rate at the impact did not exceed 10 000 fpm ?
Jeff

[takata]

Will Fraser:
Respectfully, a trim function would have more relevance to "upset".
At 2 +/- degrees, a down trim to "zero" would have a steady and insidious
effect on stability relative to airspeeds. With a fault as addressed in the pre-AD testing, trim wouldn't be correctable, and/or noticeable to pilots or computers wrestling with Unreliable a/s data to begin with. Only a few hundred feet (Qantas) can produce a substantial increase in speed. Any established descent, left alone, could be disastrous at flight limits experienced by 447: high, hot, and heavy.

As I said before, this is going to be very stretched to fit this hypothesis:
- we don't know whilch system would act "by itself" to do that in Alternate 2;
- we don't know why trim and elevators control "wouldn't be correctable", why wouldn't they?
- we don't know why the pilots, trying to fly "pitch and thrust" wouldn't notice any pitch issue, being due to some weirdo computer trim settings?
- it doesn't look like F-GZCP was so "high, hot or heavy"; she still had a good margin in all her flight parameters at this point;
- everything (beside the technical root which is very undetermined) may be assumed only if the crew would be totally panicked.

It makes a lot of assumptions and a lot of unadressed root causes in a system already proven by more than 13 millions flight hours...

S~
Olivier

[Will Fraser]

PJ2

The engine failure causing power interruption in the reasearch pre-AD is not at issue here. Neither is the TO mode. The linkage is what happens when PRIM1 is inop/unselect, and the trim is set to zero, and an inability to correct it with any means, until the second computer resets. To me, this isn't more than just another serious problem to deal with on a FD with serious issues of mysterious severity presenting to pilots who have (ostensibly) had the a/c handed back to them (automatically)

I will keep in mind what ACARS has reported (demonstrably after the fact), and assess what the avenues of handling were, as available to the PIC. (Excuse the bolding, it isn't meant to appear confrontational).

takata

I'm not assuming, I use the Data in the AD. I am however, noting that ACARS is not the FDR. If pitots are involved (it is said they are), Ua/s in combination with a PRIMinop state, unknown a/s or attitude (IRU) a nose down increasing a/s would be unknown to the crew, but if on determining a nose up was necessary and initiated with the ELEV inop (see AD), an input that was correct could be enhanced by a returning ELEV function, to the value of as much as an additional 3 degrees NU notwithstanding a lack of Pitch authority at all. At an already seriously overspeed conditon, further pitchup (by a reentering FC) could cause serious damage even without an upset. This is actually suggested by ACARS. If in Alt Law 2 but no elevator function is available, try to transit to Direct Law, only to find that the HS trim (manual) is woefully insufficient to command recovery.

Will

[takata]

Will:
I'm not assuming, I use the Data in the AD

So lets see the data.
http://atis.casa.go.kr/asms_ad/file/...008-0010R1.pdf
IMHO, this doesn't match with your explanation of the failure:

Reason:
On A330/A340 aircraft, the Flight Control Primary Computer 2 (FCPC2)
and FCPC3 are supplied with power from the 2PP bus bar. Electrical
transients on the 2PP bus bar occur, in particular during engine n.2 failure
on A330 aircraft or engine n.3 failure on A340 aircraft. Such electrical
transients lead to a FCPC2 reset. FCPC3 reset does not occur thanks to
the introduction of second electrical power supply to FCPC3 from 1PP bus
bar associated to Electrical Contactor Management Unit (ECMU) standard
5.
During the evaluation of specific engine failure cases at take-off on Airbus
flight simulators, it has been evidenced that with FCPC1 inoperative, in the
worst case, when FCPC2 and FCPC3 resets occur during rotation at take
off, a transient loss of elevator control associated with a temporary
incorrect flight control law reconfiguration could occur. It leads to a
movement of the elevators to the zero position, which induces a pitch
down movement instead of a pitch up movement needed to lift off. In
addition, it leads to a limitation of the pilot authority in pitch axis and limits
the capacity to counter the pitch down movement during this flight phase,
which constitutes an unsafe condition.

[Will Fraser]

takata

I have corrected my post to read Elevator inop. Bonehead mistake.
The material is in the AD. I see it, honestly. Law degradation requiring reset, inop elevators, Pitch unavailable, etc. Due to inop FCPC1. A Pitch problem at TO is very dangerous, no less so at altitude near the regime of 447 at loss of control.

[lomapaseo]

Vertical rate at impact

Hyperveloce

Thank you for that reasoning. It adds an important boundary (if accepted) on our speculations)

[takata]

Will:
I see it, honestly. Law degradation requiring reset, inop elevators, Pitch unavailable, etc. Due to inop FCPC1. A Pitch problem at TO is very dangerous, no less so at altitude near the regime of 447 at loss of control.

Well, if you see it "honestly", what can I add to make you change your point?
For me, the problem described is not relevant to any "Law degradation", nor specific to a PRIM1 reset, but only inside the above configuration described or the AD wouldn't be so specific with this issue. Moreover, it is involving both FCPC2/3 resets (in the worst case) in normal law situation (where FCPCs are managing "protections") with a transient issue concerning elevators reset position at the moment they are needed the other way with a larger deflection. It doesn't imply something which lasted and any "pitch unavailability", it is said that the zero pitch could not be countered by imputs, neither that it would happen in other configurations.
Again, it is taking a few bits from this AD for making something else, ignoring what doesn't fit with theory.

S~
Olivier

[augustusjeremy]

speculation

I just don't believe in the scenario: "fast fall from FL350 in one piece and four minutes" I may be wrong, but it looks to me like a false lead to follow until we'll get much more facts to explain it.

takata,

I would accept the blocked pitot theory without further questioning if it was not that IR2 fault.

I am just trying to find a rational relationship between those ACARS messages.

But I agree with you. We need much more facts to tell the whole story.

Anyway several simulator tests are probably being performed right now at AI to check all the possibilites that could lead to those ACARS messages.

Forgive me for droning on, but I think this technical aspectd needs to be clear.

PJ2,

I just have to thank you for your technical explanation.

[aguadalte]

I just wonder what in this case is convincing you to believe that an A330-200, fitted with HONEYWELL ADIRUs, and showing problems with unreliable airspeed probes freezing, would be comparable with an A330-300, fitted with NORTHROP GRUMMAN-LITTON ADIRUs, and showing AOAs spikes unfiltered?

What in AF447 ACARS sequence is similar to QF72 case?
- answer: nothing.
So, maybe its time to drop this claim for lack of any substantial or relevant fact to come with.

Takata,

Apart from the fact that there are reports of allegedly unreliable speed indication problems with all sorts of equipment, there is no relationship between AF447 and QF72.
But regarding the speculation about what really happened to AF447, your and everyone's speculative ideas, have the same value of mine...
And having in consideration the statistics regarding causes for air accidents, I think human error is well ahead of equipment failure.
I'm convinced that the Air France pilots are just men, not Gods, and that they may have committed a mistake by disconnecting the wrong equipment while dealing with a dramatic situation.


Coming back to basics of the investigation, its easy for me, to understand why they didn't deviate from the storm. They have taken-off from SBGL limited by MTOW. The Captain could not have ordered any extra fuel without prejudice of payload due to the ETOW of 233,757Kg.
According to BEA prelim report page 17, fuel weight was 70,400Kg.
In the next page of the report:

The on-board fuel weight corresponded to forecast trip fuel of 63,900 kg, route factor fuel of 1460 kg, final reserve of 2200 kg, fuel to alternate airport reserve of 1900 kg and 940 kg additional fuel

.
According to EU OPS 1:

(12) OPS 1.255 Fuel Policy:
(i) For A to A Flights — An operator shall specify the minimum fuel contents at which a flight must end. This
minimum, final reserve, fuel must not be less than the amount needed to fly for a period of 45 minutes.
(ii) For A to B Flights — An operator shall ensure that the pre-flight calculation of usable fuel required for a flight
includes;
(A) Taxi fuel — Fuel consumed before take-off, if significant; and
(B) Trip fuel (Fuel to reach the destination); and
(C) Reserve fuel –
(1) Contingency fuel —
Fuel that is not less than 5 % of the planned trip fuel or, in the event of in-flight replanning, 5 % of the trip fuel for the remainder of the flight; and
(2) Final reserve fuel —
Fuel to fly for an additional period of 45 minutes (piston engines) or 30 minutes (turbine engines);
and
(D) Alternate fuel —
Fuel to reach the destination alternate via the destination, if a destination alternate is required; and
(E) Extra fuel —
Fuel that the commander may require in addition to that required under subparagraphs (A)-(D) above.

The term Additional Fuel is used in EU OPS 1, when needed due to ETOPS especial fuel needs:

Additional fuel, if required by the type of operation (e.g. ETOPS);

I believe the term used by BEA, route factor fuel, meant contingency fuel, (a value of 5 % of the trip fuel (that may also be reduced to 3%, when a fuel alternate aerodrome is available under certain circumstances). It intrigues me that the value of contingency fuel of AF447 is only 1460 kg. Additional fuel is supposed to cope for ETOPS minimum fuel needs at ETPs and is added to Estimated Total Fuel that must already contemplate all the items of Part 1.255 of EU OPS 1.
I'm not saying that, because they were limited on fuel, they could not deviate, but I'm willing to accept the fact they didn't. What I don't understand is why, after almost 3700 replies on this second thread, I'm almost the only one here writing about human factors.

[PJ2]

aquadalte;

What I don't understand is why, after almost 3700 replies on this second thread, I'm almost the only one here writing about human factors

Well, if I can offer this view as a pilot (now retired), I think many colleagues here are only too keenly aware of the human factors possibilities but have remained essentially silent because, like the accident itself there is only so much one can say and it's all from the ACARS messages and what little the BEA adds to the thinking. At this point there is absolutely nothing to speculate meaningfully upon regarding human factors. We can delve tentatively into a few possibilities that you have touched upon but I'm not sure the discussion has anywhere to go at the moment.

You mention fuel load - most pilots who have contributed to this thread have faced the issue of offloading freight for more fuel or risking an enroute landing/diversion due to weather etc - those flight planning problems are all in a day's work. Whether the two guys up front didn't want to risk a weather diversion because they only had 900+kg of fuel and weren't aware how little it takes to go a hundred nm's off course and back, or whether they knew how to read and interpret their radar returns are all human factors and are all moot points for the time being.

Analyzing "human factors" at this point risks building a series of straw man arguments and then knocking them down - the arguments and their possibilities, mean nothing in the sense that they explain the accident in a way that other ways have not.

Yes, the crew could have run across an extremely rare triple failure - pitot sensors, ADIRU 1 or 2, and a PRIM1, these equipment failures being all independant of one another for the theory to work, the failure of the PRIM with subsequent ND bias requiring action under the AD, (assuming they knew that action off by heart, which I am almost 100% certain both would not know), - there is NO way they would have had time to look it up and run the checklist from the QRH OEB section.

And if they launched into shutting an ADR and/or IRU off "by memory", unless they were following an ECAM drill that got interrupted by a serious event, that takes us into extremely rare and difficult territory -in the realm of rogue actions outside memory and checklist territory in a last-ditch panic, again another human factors matter but in my view given the airline involved and the circumstances the flight was in, not very probable if not all but impossible. Also, we still dont' know what manufacturer's ADIRUs were installed.

[kilomikedelta]

This thread

3700+ posts (including the deleted ones) based on some photos, cryptic BEA reports and archival meteorological satellite data? I would say that 99% of the postings are egocentric masturbations by people with lots of hubris, too much time on their hands and regrets about their contribution to society. Let's wait a year or so while the experts accumulate the real data and sort out what's possible to deduce. Then we can revisit it. Time to close the thread and time for the frequent posters to discuss their behavior with their therapists. KMD M.D.

[Will Fraser]

And now yours makes 3,701, mine 3,702. You are wanting to join? You have authority issues? Your Munchhausen's is unresponsive to treatment? Lonely in Canada? I think by your definition you have wasted some of your valuable time. WF CPL

[HarryMann]

I would say that 99% of the postings are egocentric masturbations by people with lots of hubris, too much time on their hands and regrets about their contribution to society.

I would say about 50% might fit that bill, including yours...

[Lemurian]

Aguadalte :

Apart from the fact that there are reports of allegedly unreliable speed indication problems with all sorts of equipment, there is no relationship between AF447 and QF72.

I agree

Coming back to basics of the investigation, its easy for me, to understand why they didn't deviate from the storm.

Whether they did or not is a red herring. The position at 0210Z, reported on either the MeteoFrance map or Tim's 0215Z sat infrared picture puts the airplane out of the convection area. (something that most posters still haven't accepted, I wonder why ? ).

The on-board fuel weight corresponded to forecast trip fuel of 63,900 kg, route factor fuel of 1460 kg, final reserve of 2200 kg, fuel to alternate airport reserve of 1900 kg and 940 kg additional fuel

This is the flight plan that the Captain signed :

• Trip fuel ...63940
• Alternate fuel ORY 1900
• Contingency 1460
• Final reserve 2200
• Extra fuel 900 (at captain's decision)

What additional fuel, for ETOPS / loss of pressurisation, ETPs is included in the route/flight study and is inside the fuel load. Please note that on this particular flight, the en-route alternates were Natal and Sal, which are on the planned flight track.

It intrigues me that the value of contingency fuel of AF447 is only 1460 kg.

It's because AF fuel policy takes into account the burn-off caused by the need to carry a given load. The fuel plan shows the quantities left at the end of the flight : Basically, the alternate / contingency / final reserve /...are carried to the landing at destination, hence they are part of the landing weight and the burn-off they cause is included in the trip fuel.
In this case, and considering that the K factor on this flight was 1.350, a contingency fuel of 1460 kg at destination is worth 1460 x 1.350 = 1970 kg which is in fact slightly more than the required 3% of the trip fuel (1918 kg ).

I'm not saying that, because they were limited on fuel, they could not deviate, but I'm willing to accept the fact they didn't.

You are not saying it....true....Just strongly suggesting...

I'm almost the only one here writing about human factors.

And what is the basis for your discussion ?

• All aspects of the flight prior to 0210Z show a functioning aircrew, doing a serious job of keeping ahead of the aircraft, proposing some strict R/T procedures...
• A crew who was kept abreast of any changes on the route weather (the despatch message confirming some active convection inside the ITCZ)

Against that, a lot of hot air about "other aircraft have deviated...bla..bla bla ..." and not one of these posters has taken the effort of plotting those deviations on the sat pictures. Most of these flights were still smack inside the red area, except maybe the IB flight on their 70 Nm east of track who would have flown between the two most active zones.
I find it quite arrogant to judge on a crew's action without a modicum 1/-of experience of flying through the ITCZ and 2/-in the comfort of an armchair.
These guys had one hour and six minutes of flight worth in fuel and you tell me that they were concerned about any deviation or diversion ?
As to what happened after 0210Z, we are into speculations and conjectures.
And if we start the game about how this crew reacted to a wholly abnormal situation, we need both to consider what happened as the ACARS messages don't tell the whole story, and who was in the flight deck at that moment, taking into account AF policy of "no left-hand seat qual for F/Os".
But that's the sort of speculation I would be very careful in discussing.