The acars message said the pitot system failed causing AP and AT to disengage probably because of their recurring icing problem. Any speculation about weather other than icing taking the plane down needs to be proven. So far there is no evidence of that being the problem since none of the other flights with normal deviations had a problem.

Those who insist it was turbulence and flying into a thunderstorm that caused the accident have absolutely no proof and are speculating. We know BEA knows exactly why they crashed by now. When they write the report we will know too. Be patient.

Cbs
 

When sitting my F/Nav. examination, one of the questions was " You are the Chief Navigator of a Trans Atlantic Airline. What radio equipment would you require and why?"
I specified pretty well everything...
I asked another candidate what he had replied. He said that he had written " I have been in this position myself. I would require 2 ADFs. Two for reliabilty. With them I could get time checks for Astro. I have flown across the Atlantic 90 times with just this equipment and am quite satisfied with it." ( He had held this position with a South American airline. This would have been pre Weather radar being around.)

It might be helpful to look at what had happened to some of the other airlines' experiences with loss of airspeed. NWA, near Tokyo were some 25 miles from a Cb which was off to one side and went into some cirrus. It was daylight. I think that they lost their Pitot etc on two occasions in quick succession, both recovered in a few minutes. TAM was over the Caribbean, at night. Again the ASI came back after a few minutes. Both of the flights were prior to AF. Jetstars episode came after AF.

More details must be available of these (and others ?) from places other than my armchair.

What went right ? ( for them)

Frames of reference
 

Dutch M
Hi,
I don't really want to discourage an ambitious attack on the AF447 energy question, but let me add a few additional questions to your approach for an answer.

Have you added in the Earth's velocity vector around the Sun? How about the velocity of the solar system around the Milky Way galaxy? The answer should be that it is not necessary, these are negligible factors. For the same reasons, the rotation of the earth does not need to be included (unless you are flying ballistic like an ICBM). It exerts a force on the aircraft all right. We call it Coriolis force, but that too is negligible for our purposes.

There is no absolute inertial frame of reference then, is there? The height gain only reflects the change in position relative to the Earth's center of mass.

Will an aircraft flying 100 knots into a 100 knot head wind be able to convert some of that velocity into altitude gain? Of course it can. (Assuming it can fly slower than 100 knots)

Dutch M
HN39 has done an energy matching simulation (segment by segment) combined with aerodynamic assumptions. (Believe it was in the prior Tech thread.) His conclusion was that the final climb from FL375 to FL 380 was ballistic, i.e. the aircraft started upward on the last dregs of its lift and decelerated below actual level flight stall speed in the climb. This would allow the aircraft to fly below actual level flight stall speed and continue trimming nose up because the AOA had not yet become high enough to shut down the automatic pitch trimming. When gravity caught up with the aircraft and it began to fall, it then went rapidly well past the stall AOA and shut down the automatic pitch trim at 13 degrees ANU.

HN39 is very good with this type of thing, and we Physics duffers should respect his knowledge.

HN39, if I have gilded the Lilly too badly please correct me.:}

Do we read a different version of the first interim report ?
But what's the point after all, still have to work with ACARS messages + BEA note when actually all FDR data have been retrieved ...

up to AoA 45 deg the move is very low,
even with AoA 60 deg it has moved not to much aft.....

look at the curve of the pitching moment (Cm) against angle of incidence posted by machinbird

http://www.pprune.org/tech-log/39510...ml#post5569978

Hi Grity,
That was posted well over a year ago, and is mostly a regurgitation of Davies excellent book. That pitching moment curve Davies provided was no doubt derived from a T-tailed aircraft and had a "sticking spot" that kept the aircraft from pitching back down after a deep stall. Probably the more recent pitching moment curves from the FAA generic airliner stall study are more appropriate to AF447.

hi Machinbird, as time goes by......exist a link? for a Cm and AoA up to high deg (90)

>>Have you added in the Earth's velocity vector around the Sun? How about the velocity of the solar system around the Milky Way galaxy?

No, I didn't add this, though I did consider and is the reason I did add the phrase "Due to lack of time, I didn't have the time to think this through for 100%, though I do expect this to be relevant."

And indeed at first glance those aspects do have to be taken into account, though I -think- this is not relevant because the airplane is not acting in the sun's gravitational field. The earth is, but the airplane isn't (because of its low mass).

>>up to AoA 45 deg the move is very low, even with AoA 60 deg it has moved not to much aft.....

>>look at the curve of the pitching moment (Cm) against angle of incidence posted by machinbird

Yes with a higher AoA it'll become less, though still the same effect.

And on this plane the tail is not in the main wing shadow/wash.

>>HN39 has done an energy matching simulation (segment by segment) combined with aerodynamic assumptions. (Believe it was in the prior Tech thread.) His conclusion was that the final climb from FL375 to FL 380 was ballistic, i.e. the aircraft started upward on the last dregs of its lift and decelerated below actual level flight stall speed in the climb. This would allow the aircraft to fly below actual level flight stall speed and continue trimming nose up because the AOA had not yet become high enough to shut down the automatic pitch trimming. When gravity caught up with the aircraft and it began to fall, it then went rapidly well past the stall AOA and shut down the automatic pitch trim at 13 degrees ANU.

>>HN39 is very good with this type of thing, and we Physics duffers should respect his knowledge.

>>HN39, if I have gilded the Lilly too badly please correct me.http://images.ibsrv.net/ibsrv/res/sr...s/badteeth.gif

Yes, I've read that and as explained, those calculations is missing something. As such I started my article. Only for vertical movements, the trade-off in energy can be applied without further side conditions. When changing a horizontal speed into height, other aspects start to act up. If you want more explanation, let me know. I'm aware it's written down pretty condensed.

Maybe I should have mentioned before: I'm not a pilot, though do have a degree in Applied Physics (which includes Control systems behavior at the University I went - some years ago- ).

And: My description also suggests why there are around 10 reported similar upsets. Unexplained upsets without any plausible technical cause. And no, no conspiracies .....

Dutch M..

That sounds very unlikely to me. You are suggesting that the HS is producing lift in the downward (negative) direction despite being at a severe POSITIVE angle of attack.

Imagine your explanation applied to a wing.. You're suggesting that a wing flown at a severe negative angle of attack (so severe that the lower surface is stalled) can still produce positive lift. I'm aware most cambered sections have a zero lift angle of attack that is slightly negative but your suggestion goes way beyond that.

3rd report will be released Friday 29th July...
 

communiqu 25 juillet 2011

Press release only in French (at the time of writing)... the third report will be published this Friday (29th July), with press conference at 14.30 local time.

"Ce rapport présente les circonstances exactes de l'accident avec des premiers points d'analyse et de nouveaux faits établis à partir de l'exploitation des données des enregistreurs de vol."

The report will present new information established from the flight recorders...

@Dutch M

I politely suggest you go a read the Wikipedia article on Kinetic Energy. It is correct to use the airmass as reference and airspeed to calculate the KE relative to the axes (its not an absolute quantity). Newton's 2nd law is all about momentum, and momentum is linear in velocity (for constant mass). The equation for Kinetic Energy comes from integrating the momentum from one velocity to another. You do not have to take in account the KE of the system relative to the earth, sun, milkway etc - just use the same axes set before and after. So IMHO it is correct to just use 1/2m(v2^2 - v1^2) = mgh, you do not need intertial equivalents for v. 1/2m(v1-v2)^2 is wrong - unless v1 or v2 are zero!

Better get ready to begin Thread nbr 6. This board is about to get real busy.

Low/high speed protection availability in ALT2
 

Hello there
Had the low speed protection been available, would the A/C have been able to zoom climb (or later stall) like it did ?

Hi wes wall,
Certainly. And, as it's another interim report, maybe Threads 7, 8 ,9... will follow!

This communiqué means something like this:
This report will present the exact circumstances of the accident, the first analysis pointed, and the new facts established from the flight recorder data's processing.

Hi Hyperveloce,
It's hard to tell as it could be overriden by the pilot, but with VSw displayed, Stall alarms workings... certainly.
Check it by yourself:

http://takata1940.free.fr/lowspeed.jpg
*ALTERNATE LAW 2:
- Low speed stability is lost with double ADR fault.
- High speed stability is lost with triple ADR fault.

Going round in circles
 

As a student I piloted gliders. These routinely fly tight circles to stay in thermals. In a strong wind there is no airspeed loss turning downwind, nor gain turning upwind. I've never quite figured that out, may be as a physicist you can.

Hi Takata. Ok, the gentle ND introduced by the low speed prot. can be overriden by the PF (and maybe the PF persistant NU inputs were to counteract the low speed prot. ND orders, but he did so much more than that). But if we consider the past UAS incidents analyzed by the BEA, do we see a single occurrence of a low/high speed prot. activation when the ALT2 was triggered by a NAV disagreement ? In a few past cases, a descent (never more than 5000 ft) was decided when the stall warnings sounded but it was a PF action. Can we assume that these low and high prot. were not available in all these UAS incidents and in the AF 447 case ?

Hi Confiture,
First, yes, I guess that we are reading the same report, but I'm trying to include every available informations from it, not only a single quote about the general case above that you underlined from it.

Second, the point is mostly to understand what we already have, before even going further with new data when they will be available (soon).

So, my initial point was to link those ACARS with real cockpit events:

1. Event started at 0210:05 with AP/FD and A/THR faults;
2. Event is due to 2 or 3 PITOT-PROBEs (total pressure) faults affecting some of their respective ADR outputs (barometric altitude, speed, Mach, angle-of-attack, temperature and overspeed warnings);
3. Those PITOT-PROBE faults caused all the following 0210 ACARS to be reported;
4. ALTERNATE LAW 2 was consolidated after monitoring, it was not transient;
5. 2 or 3 ADRs were rejected by FCPCs and FMGCs;
6. There was no "ADR DISAGREE" triggered during the 0210 sequence;

Hence, it may be deduced that it was due to a "triple ADR" fault:
=> ADR 1 + 2 / ADR 1 + 3/ ADR 2 + 3 (ECAM)
=> ADR 1 + 2 + 3 (not displayed by ECAM)
Consequently, Low and High Speed protections were lost while "SPD LIM" was flagged on both PFDs without VSw and VMax displayed.

- ADR 1 & 3 were coherent and valid again before 0211:06.
- Another triple ADR fault sequence started "after 0211:40";
- All speeds became invalid again;
=> It may be assumed that aircraft's airspeeds went below 30 kt, they would not be displayed anymore on both PFDs between 0211:40+ and 0212:19+
=> It may be assumed that "ADR DISAGREE" was triggered after this second triple fault was cleared.

Ground speed is affected, but not airspeed.
Once airborne, the glider is in the air mass that is moving at the speed of the wind, but for the glider itself regarding that air mass, there is no notion of tail or head wind.

See my post above. Low speed Stability, VSw, (and associated speed stall warnings) would only be available with two ADRs being declared valid, which might not have been possible before 0211:40+. Moreover, if it has been available before or after this point, it would have triggered "DIRECT LAW", whatever imputs made by the PF in order to counter those active ND protection imputs. Quite frankly, there is no trace of "Direct Law" and this never happened.

No (first question). Yes (second one). All events involved 2 or 3 ADRs faults, this ruling out "Low Speed Protection" (other have been transient, hence, the system would revert to NORMAL LAW and this protection doesn't apply). Moreover, not a single aircraft ever made a single excursion out of its flight envelope.

After reading post #661 ...

Following so far, these protections are features of Normal Law.

Caveat: the following informed by sources that are between 5 and 10 years old.

When the pilots noted "Alternate Law," speed stability functions should have been in play if they were in Alt 1, but not Alt 2.

The system needs two valid ADR for stall warning and low speed stability.

I look at the FMA, and try to figure out if alert to "Alternate Law" is in one of the five fields above the attitude display, or on the ECAMS page. I cannot. (Don't have anything that tells me.)

Stability function can make inputs but can be overridden by the pilot's inputs.

This makes me wonder: did the crew know if their state was Alt 1 or Alt 2? (There may be more on CVR that will clear this up from next report, but there may not. Likely their work load was up, see below, and as such precluded any discussion as to which Alt mode was in play).

If PF didn't recognize alt 2 versus alt 1, would that have made a difference in how PF made and monitored nose inputs? My guess is no, but it's a guess.
(Question based on not having flown A330 ...)
This leads to the question: were the pilots aware of this?

Perhaps the next reprt, and CVR analysis, release will shed light on that question. Perhaps task load precluded such a realization.
From June 2005 revision of the Flight Crew Training Manual (8.110.3)

"... if all three ADRs provide consistent and erroneous data, the PRIMs and FEs will reject the “good” outlier ADR and will continue to operate normally using the two consistent but faulty ADRs."

Interestingly, in another section of the manual (2.60.11)
DIRECT LAW - In most cases of triple failure, e.g. triple ADR failure, direct law is triggered. Autopilot and auto-trim are not available.

One hopes that the various cases are handled in the simulator training sessions.

I also note that around pages 2.30.5 through 2.30.8 some ECAMS crew coordination is disucssed.

The process by which a crew clears ECAMS alerts is organized and formalized, but I see a source of possible scan disruption for the PF in this process, based on standard procedures for clearing ECAMS alerts.

HN39 and CONFiture,
Your aircraft is not purely a feather blowing in the wind, it has a certain amount of inertia. Usually more of an issue during windshear, rather than while turning from downwind onto finals....
I would suggest the subject is mostly O/T here, and could maybe do with a separate thread on TechLog to kick the relevant notions around?

New BEA report next Friday
 

The BEA will publish a new report next Friday 29th of July in the afternoon.

henra, grity, jd_ee,

Thanks for the additional comments on your previous posts.

I am looking forward for the next BEA report, and see how more it will contribute to the theoretical side of the accident understanding and interpretations.

Dutch M,

Interestingly, I am among those that went through the same reaction, from an initial, simplistic, trivial "pilots fault" to understanding better and better the complexity of a rare intersection and confluence of multiple causes, that have a much wider span, and involve so many more players. Consequently, I believe, the AF 447 is one of those cases - the allocated resources were a big help as well - which will contribute to improving multiple aspects of the safety of commercial air transport.

I think the spacial 3D/3axle character of the Kinetic Energy requires a consideration in its conversion to Potential Energy, regardless of it being linear or non-linear in terms of a mathematical function. However - if I understand correctly your sentence - I could not consider each of the 3 axle components "independent" of each other, by virtue of them all depending on the same spacial element, of which they are a component of.

I have struggled with the use of airspeeds, including TAS, for the Kinetic to Potential Energy conversion/conservation, and I came to the same conclusion that the speed relative to the ground is the one that is relevant. Two elements are relevant, IMO, in this very order: (i)the height is relative to the ground, and therefore, that's the reference system, and (ii) air can transition from adding, diminishing, or being a null factor in the A/C speed relative to the ground, and thus Kinetic Energy, which is conserved from one (air) transition to the next, and relative to the ground.

I would not go that far though, as to using the ground rotational speed.

Based on the info that we have so far - this may change, once we have the next BEA report, and more info about the state of awareness in the cockpit - the Stall Warning or the lack of it is, IMO, a very, very important contributor to the accident. It was the lack of a Stall Warning while in the middle of the "stall" and the presence of a Stall Warning, when the A/C was close to "exit" from the "stall"..

One or another, is ultimately less relevant as the fact, that in adverse weather condition, all fail the same way, reducing this part of the system's "redundancy" to zero. This is an industry wide element, not only Airbus. The redundancy built in may work for individual electrical or mechanical failure, but such events are arguably less frequent than adverse weather conditions, and thus a weak point in passenger A/C design industry wide.

Interesting....

I appreciate your post :D

Hi airtren,

Simply use the air mass as your frame of reference. It may be moving horizontally over the earth's surface, but the conservation of KE + PE still holds.
IAS is just a measure of dynamic air pressure, it must be converted to TAS.

flight protections in ALT2 (end)
 

Thank you Takata for your patient answers. A last question: after the apogee at FL 380 of the quasi-ballistic trajectory, the AoA rapidly exceeded 30°, and the abnormal attitude law could have been triggered on this criteria (AoA>30°), but it has been said that it was inhibited since the airspeeds were invalid (below 60 kts) at that time (which is strange since IAS<60 kts is also an abnormal attitude criteria), and this also inhibited the stall warnings. But when the airspeeds became valid again, and when the stall warnings were allowed to sound again (after the ND inputs and thrust reduction), shouldn't the abnormal attitude protection law have triggered ? (since the AoA always remained above 35° according to the BEA)

There have been many well thought-out and well explained hypotheses in this amazing thread. However, I still think that the basic problem was one that could have been solved by Power and Attitude.

Decades ago I flew an aircraft that had a beaut, coloured Standby Horizon, powered by a 28v DC battery, which was in turn kept charged by the aircraft's main electrical system. In the event of a Total Electrics Failure, the aforesaid 28v battery would provide power for a guaranteed 30 minutes. It was completely independent of any other power source, and would get you out of trouble when all else had failed.

Why can't we have such a simple and effective solution today?

To what problem?

More effective than what?
There's no indication AI was ever lost...

JD-EE

wallybird7 said, "Apparently this crew didn't."

That is an assertion not in evidence. How do you know the storm showed on their radar? Was their radar working? Was it employed properly?

The weather forecast showed a line of thunderstorms about 100 miles wide over the equator. See AF447 Meteorological Analysis.

By the time the problem appeared on their radar any escape from the storm would have required extreme deviations from their flight plan with no communications to DAKAR.

They did note it at that time and made a deviation towards an area that appeared less risky than plowing on straight ahead.

Now, one might argue that they should have called Atlantico about this. But apparently this happens so often the pilots simply noted the fact and plowed onwards. I wonder what would have happened had they tried to check in at the follow-on control site after DAKAR since their flight plan was not properly forwarded.

Deviations happen all the time and you don't need clearance, just announce it on Guard Channel.

But the pesky detail stands - moving aside from the storm at the time they saw it would have required some serious course deviations. At the time they hit it the storm was a very wide solid band.

Agree. They should have deviated sooner. They knew the turbulence was coming.
Jcjeant
Hi,
Quote:
Was their radar working? Was it employed properly?

Again we return to the basic questions and answers ..
Radar working .. plane defect ?
Employed properly .. crew error ?
Again .. nothing between ...

Except a large build-up. And then all hell broke loose.

Bubbers44
The acars message said the pitot system failed causing AP and AT to disengage probably because of their recurring icing problem.
"Any speculation about weather other than icing taking the plane down needs to be proven."

Proven? How? Do you want a precise measurement of potential up and down drafts? What caused the ice build-ups? If not strong up and down drafts. Does turbulence exist in towering cumulous? Hell yes. Would it affect someone trying to hand-fly the plane?

So far there is no evidence of that being the problem since none of the other flights with normal deviations had a problem.

Just because the others skated through, is not "proof" that this one didn't get caught.

Those who insist it was turbulence and flying into a thunderstorm that caused the accident have absolutely no proof and are speculating. We know BEA knows exactly why they crashed by now. When they write the report we will know too. Be patient.

All of it is mere speculation. All the BEA has to come up with is "Probable" cause. Not "Exact" cause.

@cwatters
Yes, indeed, and it does look very strange.

Though, don't forget, in the main wing situation, the lift is destroyed because the airflow over the wing is not able to "carry" the weight of the airplane any longer and the boundary layer lets go.

In the HS situation, the weight is zero, because the direction of movement is already downwards.

And yes, in some situations the physics does look bizarre.

@sensor_validation
Yeah, suppose the windspeed goes from 100 knots back to zero, does this change the Ekin of the airplane ? If not, then you will need to take the groundspeed as speed reference.

The energy exchange Ekin vs Epot, is in general only valid for speeds in the direction of the change of Epot (ie Height) and not when trading speed between perpendicular axis.

Dutch_M, you need to be aware that you have to determine the kinetic energy always with regard to the relvant reference system.
For an aircraft that is the air surrounding it.

To make it easier to understand let's take an example:

You have a wind speed of 200kts. Your aircraft is travelling at 250kts.
Minimum speed of your AC is 150kts.
How much speed can you trade for altitude now?

Correct ! You can trade 100kts.
You will be flying backwards then with regard to mother earth, but aerodynamically who cares ? You will still be flying.

If the relevant reference system changes its state beyond what the AC can bear, well then: bad luck. but you would have the same problem when you are flying at 100kts and suddenly a tailwind of 100kts hits you. Won't make a difference except for the horizontal component during impact with mother earth..

@henra
Hmmm, headwind, tailwind, airspeed, groundspeed ?

And yes, the speed bleed-off is limited by the aerodynamics to keep the airplane flying. Though speed bleed-off is not the same as loss in inertial Ekin.

Waouh!
If a/c weight was rero, then she didn't fall!
Weight can't be cancelled. Balanced, perhaps (when the a/c flies), but cancelled, not!
"Weight is the name given to the force on an object due to gravity." (Wiki). And excepted a gravity failure (:}), it can't be zeroed on Mother Earth

Hmm, I don't want to sound rude, but did you really understand my example ?

And yes, the square of the speed before the bleed minus the square of the speed after the bleed is exactly the kinetic energy you can trade for altitude.
What is speed bleed if not redution of kinetic energy ???

Edit:
You really need to understand the concept of reference systems to get a grip on this.
Let me ask:
What do you think is the kinetic energy of the aircraft in my example after the speed loss of 100kts?
Please note: It is flying backwards now. Does that mean it has a negativ kinetic energy by now ?

Other question: Would the aircraft care if mother Earth instantly decided to change its sense of rotation by 180° from one second to the next?
You probably guessed it: It wouldn't give a sh't as long as the air around it keeps moving unchanged. The only force which the Earth itself applies to the aircraft is gravity and that is vertical.

Edit3:
Yes you can calculate the total kinetic energy of the AC with regard to the sun but that would be absolutely pointless. You can also calculate the kinetic energy with regard to the ground but for the question of how much altitude can I gain by giving up airspeed it is also irrelevant. It is only speed through the air which I can trade. When speed through the air is Zero there is excactly 0kts to trade. You can't get 'slower' than 0kts TAS/IAS/CAS. No matter what the surface below does. Or how the sun turns around this point.