AF447

[Mad (Flt) Scientist]

Originally Posted by surplus1

.....

Over time I’ve learned a few things about stalls but unfortunately for me that does not include any knowledge of anything known aerodynamically as a “stable stall”. If you would be kind enough to tell me what that is, I would be grateful.

In those four decades of flying the line I’ve never heard of anything known aerodynamically as a “stable stall” to which you referred in all three of your posts. Of course that doesn’t mean that it does not exist, it could well be my ignorance. Again, would you please tell me what a “stable stall” is?

.....

I am extremely confused as to how you could define it by the term “stable” as well as by how that aircraft would be kept in a full stall, of both wings, wings level, over a descent of some 35,000 feet in a [near] vertical trajectory until impact. Please help me to understand what would keep one of the wings from gaining some lift and the other from rolling off during the descent. Something other than ALTN law, please.

If what you’re going to tell me is “ALTN law did it”, then tell me also why, when ALTN law makes the required control input to keep a wing from rolling off, would that not be likely to induce rotation of some type [what we call a flat spin]. What control surface(s) would you expect the computer to move in its effort to keep the wings level?

.....

If I may be so bold as to attempt to answer these points.

An aircraft can be in a "stable stall" if it possesses adequate elevator power to hold the AOA above the critical value, retains enough roll control authority to allow the crew to maintain wings level, retains enough directional stability and/or rudder power to keep sideslip relatively constrained and, in such circustances, if the crew elects to remain in the flight condition. (I'm excluding a "locked-in" stalled condition, such as a T-tail deep stall, where the crew may not have the option, at least in the pitch axis).

Without specific knowledge of A330 high-AOA aerodynamics, and just looking at the configuration, with a low set tailplane I would expect there to be plenty of elevator power - both to pitch down if desired, but also to stay post-stall as well. Assuming AB tried to get reasonable handling characteristics for the natural stall (always a good thing to do, aerodynamically, whatever "protections" you plan on building in) then there may well be a inboard/midwing stall - perhaps from the pylon/wing area? - which would imply that some (outboard) roll control should still exist.

The question then becomes, why would the crew not recover, since I believe that the configuration shouldn't be vulnerable to the locked-in case. The answer might lie in the nature of the upset and the pitch/power unreliable airspeed procedure.

Speculation: once in the post stall regime, with a high AOA, a pitch attitude near the horizon (as the procedure calls for) might not have been enough to reduce the AOA if the flight path had dropped 30 degrees or more below the horizon. Similarly, the intermediate power (I believe) called for might not have been enough for the very high drag of the developed stall. In that scenario the crew actions of flying pitch and power and maintaining wings level could, conceivably, lead to a long, stable, stalled condition with a near-zero pitch attitude, a relatively high rate of descent and a fairly low forward speed.

In the context of that speculation I'll note that fixed pitch, apply power, is the typical low altitude "recovery from stall warning" FAA training but does absolutely nothing for you at high altitudes, where you don't have the power to just drive out of the stall and need to lower the nose. The unreliable airspeed procedure is more designed to maintain you in a stable flight condition and avoid an upset, not necssarily to provide recovery from an upset once it develops.

[JD-EE]

Will Fraser, I suspect the news media should do a little explanation to their readers that the BEA reports "pretty good condition" is probably a declaration of a crewman on the vessel that picked it up with perhaps a remark or two from the medic. Broken bones but basically all in one piece is pretty good condition for somebody used to people coming apart and bleeding heavily.

edit - besides, as noted, they said "relatively well preserved" not "pretty good condition". That moots most of my comment just above.

I am beginning to wonder if the BEA formally asked to view the autopsies. I get the impression that the Brazilians are sticklers for procedure. The French may have asked informally relying on their formal status and been "rightly" rebuffed by the Brazilians. A formal request might have turned the trick.

edit- Another issue that I personally have with the report is that it's formally organized and reasonably complete but appears to be "sloppy" with critical informational clues, such as SatCom antenna and satellite used tossed in as after-thoughts. The particular SatCom transceiver used might shed more light on the ACARS communication pattern.

Upon rumination I've concluded that 1 month is a nice time limit for a plane that has faced a deadly mishap under conditions that allow its quick location and recovery of parts. The BEA could have been under such time constraints to meet the 1 month deadline that they didn't have the time needed to tighten up the report with more details on silly things that would not normally affect an investigation. -/edit

I'm willing to see this as a misunderstanding created by cultural differences.
Differences between the legal systems of the two countries involved could have added to the issue.

This is simpler than a growing conspiracy in an internet savvy world. If there is a conspiracy people like the pilots here, the engineers here, the experienced investigators here, and other specialists here would dig out the truth from almost anything they concoct to protect their backsides.

JD-EE

[JD-EE]

Squawk_ident (and others)

You folks into the no proof of somebody alive from 0135 or so onwards please consider page 34 paragraph 1.16.2.1 and let me and others know of the plane itself might have tried to initiate contact with Dakar or the pilots had to interact with the equipment to make it happen.

JD-EE

[takata]

JD-EE:
Will Fraser, I suspect the news media should do a little explanation to their readers that the BEA reports "pretty good condition" is probably a declaration of a crewman on the vessel that picked it up with perhaps a remark or two from the medic. Broken bones but basically all in one piece is pretty good condition for somebody used to people coming apart and bleeding heavily.

Well... about 30 out of 51 bodies were recovered by the French Navy. Most were transfered from Ventôse to BPC Mistral which is fully equiped and teamed for a very serious examinations of the "body state". Look here:
Mistral Class ? Amphibious Assault, Command and Force Projection Ship - Naval Technology
[...]
The 69-bed, 750m˛ hospital is equipped with two operating theatres. If additional hospital or medevac space is required, the hangar can also be converted into a modular field hospital.

I would consider as a fact that the BEA forensic team (barred by the Federal Police) would be fully informed of any "medical observations" made by the French Navy hospital team, no matter if the Brazilian are still dragging their feet and do not fully cooperate with them.

The Brazilian Federal Police is in charge of the victims identification in cooperation with Interpol (including the French police team which is not allowed to do anything). They stated that this task was their primary one and a report would follow once the identification work will be completed. They don't want the BEA killing the suspense by revealing too much about it.

But anyway, I would consider that if the BEA mentioned this observation, it should be taken as a fact backed by a serious medical advice.

[mm43]

Squawk_ident

INTOL to TASIL is 364NM and irrespective of the SALPU ORARO intermediate reporting waypoints, the ETA TASIL was 0220z based on M0.82 FL350 OAT 45C. This gives a GS of 477KT -10KT headwind = 467KT

The distance above hasn't been taken from the sector distances on the Enroute Chart, but is the calculated distance. The distance from INTOL to the 'derniére position connue' - last known position at 0210z is 293NM, and equates to an average GS of 475KT which puts the A/C about 1 minute ahead of estimate, based on the above quote.

EDIT:-
My apologies, I now realize you are raising your question based on the 0133z position report with ETO SALPU 0148 and ETO ORARO 0200z.

No wonder ATLANTICO were requesting an ETA for TASIL! They had already checked AF447 on 6649 and if they couldn't raise them you'd think they'd have tried the secondary HF frequency of 5565.

Going back to the HF frequencies, at 00:36:40 RECIFE CENTER on VHF had already told AF447 to contact ATLANTICO on PRI 6535 SEC 5565 at INTOL. This was later retracted at 01:31:44 when RECIFE CENTER changed the HF frequencies to 6649 /5565 at INTOL and 6535 at TASIL. Once again there seems to have been confusion.

BEA do not appear to have made a comment regarding this discrepancy - it being politic not to say anything pointing the finger in the Preliminary Report.

I suspect the Captain was already taking his rest period when the transfer from RECIFE CENTER to ATLANTICO took place. END EDIT

Agree but Atlantico had the ETO given by the crew. Respectively 0148 and 0200 but not TASIL and it was needed to relay the ETO to DAKAR.
From where does 0220 come from? ADS-C may be or estimation by Atlantico from the latest radar reception, but if the position is correct until SALPU (0148 under radar contact and the same given by 447) the ETO ORARO is not respected according to what the ACARS has send as we can see from the published position report.
The speed was reduced at or just after SALPU. Why?

ATLANTICO ~ DAKAR audio with the TASIL time of 0220z is on this tape:-

http://countjustonce.com/af447/audio_020709.mp3

But how ATLANTICO derived the time is another matter.

mm43

[safetypee]

surplus1, thanks for your views and the comprehensive explanation. (#3133)

First the stall: my hypothesis is that with the loss of airspeed procedure the aircraft was likely to slow down (power, attitude), which in the circumstances of weight and altitude would progress towards the stall.

Modern swept wing aircraft, in general, tend to pitch up into the stall (usually gently) – I don’t know if the A330 does this. If the nose-up pitching moment remains at post stall AOAs then without any nose down elevator the stall may be described as ‘stable’ (not the best of terms, as the condition may involve pitch and/or roll oscillations). Like a stall with full back stick, the wings may rock, but some lateral stability usually remains – why should the wings ‘roll off’.

My understanding is that in ALTN law (no speed protection) a stall is possible – AOA protection may be available, but I await an answer to my previous questions on this.
It is also my understanding that the normal control law in manual flight provides a ‘follow up’ trim function – at least with the stick free the system holds attitude. If ALTN law is the same/similar, then the aircraft may be trimmed into the stall, or at least at a very slow speed before the stall. The alternative is to use manual trim.
Without trim stick-force feedback the crew might be unaware of the trimmed state, they would have to look at the trim indicator – I have asked validating questions on this elsewhere; note similarities with conventional auto trim (also covered by TheShaddow), but the A330 systems appears not have the same mechanism as conventional auto trim, which disengages with the autopilot disconnect.

The slow speed trimmed state, together with any nose-up pitching moment could maintain the stalled condition (stable stall), which the crew may not deduce, and even with full nose down control input (not an instability), the situation is not recoverable without applying nose down trim: c.f. 737 AMS acct & UK incident. A conventional aircraft has stick force proportional to displacement from trim, whereas ALTN appears only to have to have force proportional to stick position – validation required please.

The important issues are; as stated by TheShaddow (#3106)… “But what was happening to the THS (trimmable horizontal stabilizer)” and “…. mask the back-trim and leave the pilot totally flummoxed?”; our difference is that as explained above, the nose remains high in a stalled condition.

Re turbulence: in general terms, the certification requirements require that an aircraft can withstand the ‘worst case’ gust (turbulence) from both aerodynamic and structural perspectives when flying at Vra; at slower speeds, a stall (or roll upset) is possible. Engines have similar requirements, but also like aeros/structues not necessarily considering all contributing issues. Furthermore, “what is worst case” and how is it affected by crew action; the industry still has to learn, unfortunate from accidents like those you cite.

Our views don’t differ on Cb encounters, my “inadvertent” is in the context of not premeditated or not deliberate.

Re pitot balance; I have corrected my post, and in other replies explained that the point is still relevant. Turbulence is most unlikely to cause the abnormalities reported – consider that a gust is more likely ‘alpha’ opposed to IAS change.

MFS #3136, thanks for the parallel view. In addition to your post-stall description I would appreciate comment on the trim issues.

[JD-EE]

safetypee,

I'm not a pilot and don't pretend to be one. I know commincations. That's simply where I come from.

Your description of the deep stall and coming out of it is reasonably clear. I'll add that my understanding of any stall is that you get buffeting. I further understand the aircraft had its CG shifted aft by some amount. (It sounds large to me, but what do I know?)

They were apparently or at least possibly flying into a really nasty set of conditions. Their 0210 position seems to plot to inside the West edge of the most severe of the conditions, if I am interpreting the temperature contours as being of interest relative to the severity of conditions.

Entering that region may have upset the applecart leading to a stall that, due to buffeting from the "rough air" (as pilots often say in announcements to the passengers), they were unable to notice. The arrived, somehow, at a condition that required mitigation by precisely the opposite of what they were doing. This persisted, due to no horizon reference, until they bottomed out of the clouds and got low enough the rain did not obscure vision. At that time recovery would be impossible.

Doesn't recovery with a pilot flying the plane require that the pilot actually be aware that the plane is in a stall condition? How possible is it that the poor fellows simply were not aware of the plane's true flight configuration until too late?

JD-EE trying to learn something that MIGHT explain the plane coming in relatively flat quite close to their last reported position if not behind it.

[Tailspin Turtle]

Why is everyone fixated on a stall/spin?

If the pilots lost attitude reference or couldn't reconcile which one to use, the default maneuver in instrument conditions is a spiral dive even if the loss of control began with a stall. A spiral dive will get the airplane from 35,000 feet to the water in less than a minute. The airplane won't come apart even if it goes supersonic if the computer is restricting load factor. If that's in fact what they got into, they could still hit wings level and flat at a high rate of sink if they finally got attitude reference back but too late to completely pull out.

[galaxy flyer]

MfS

We now practice high altitude (FL 430) stall recoveries. It is amazing, with a slow degradation of speed into the shaker, how long and how much altitude is required to regain M.75, the min to start pulling. Basically, 5-10 degrees nose down at MCT thrust and 4000-5000 feet, IIRC.

GF

[einhverfr]

How would the A330 react to a negative G stall?

[Hyperveloce]

BEA report

Just read it. frustrating. no forensic data available to the BEA at the date of the report ? just visual inspection made on 30 bodies on board the Ventose frigate. In the absence of the recorded data of the flight, forensic data and associated identities when available seems of prime importance. Analysis on the debris would indicate that the structural integrity of the aircraft was not compromised before it impacted the sea surface with a high vertical speed (a high rate vertical *acceleration* is mentionned in the report). The VS would have been detached with a forward motion by the impact, but the damages on the lower part of the rudder does not seem adressed: did the aircraft impact the surface tail first, then belly ? It is said that the connecting brackets between the floor and the walls were bent backwards: is it the result of the impact with a high vertical speed and tail first/a positive pich ? does it suggest that the aircraft had also a significant horizontal speed ?
No analysis of the distribution of the debris versus their drift/date of recoverage. If the aircraft impacted the surface in one piece, the debris/bodies were initially collocated: between the debris collected on June 6 and those collected on June 9, there is a differential drift between groups of debris/bodies (approx. 15 NM on 6 days for the ones, approx 60 NM on 9 days for the others) which is very difficult to account with the NOAA data derived from satellite observations. Furthermore, if the AF 447 had been flying at mach 0.82 from the last ACARS position at 22:10Z, to 22:14:30Z the aircraft would be too much north on its way to TASIL to explain the position of the first debris collected on June 6 (it would require a drift toward the west).
About the ACARS temporary interruption at 22:13Z, a minute before the definitive interruption: no explanation offered (severe met conditions degradating the satcom link budget ? unusual aircraft attitude and possible fuselage masking ?).
It is a report a bit frustrating because about established facts only, giving the feeling that sometimes it would require a bit of interpretation or explanation (of the established facts) or even expert assumptions/opinions about possible causes. I have learnt more by reading the contributions of the good, experiented flying people here on PPruNe than by reading this report, despite less data available to them.And the main question remains unanswered: how did the pilots come to lose the control of an Airbus which is said to be fully controlable even in alternate law 2 and pich & thrust, moderate turbulences, and procedures (unreliable speed) to implement.

How the crew reacted to the cascade of fault reports, alarms and applied the procedures is surely part of the key. Stall or overspeed warnings are not recorded by the ACARS but if the crew were subjected to stall alarms whil flying in ALTN 2/pich & thrust, and instructed by the procedures to take them in account: how should they have reacted to these stall alerts ? (the Air Caraibe captain chose to ignore them despite th instructions). Would they alter the pich & thrust parameters on their own (decrease pich and/or augment thrust) to gain speed ?
Sorry for this long contribution.
Jeff

[Mad (Flt) Scientist]

Originally Posted by ttcse

It seems to me a standard stall from flight-levels to the sea is unlikely. Either the crew would recover or the natural stability of the aircraft wouldn't allow it to be sustained.

That presumes that the crew would be aware of the stalled condition. With no reliable or trusted airspeed information, and with perhaps serious doubt in the crews' minds as to the validity of altitude data also, and thus perhaps no indication of the actual flight path, I'm not sure we can assume that a stall would be easily identified from the available data.

Originally Posted by ttcse

Only a deep-stall would be a stable stall.

I disagree. Deep stalls are a form of stall where the aircraft cannot be recovered to normnal angles-of-attack by means of the aerodynamic controls, due to (typically) blanking of the empennage.

Stability merely requires that there be a combination of angle of attack and flight controls which will maintain the post-stall condition. Any aircraft with a post-natural-stall stick pusher system has to be capable of a stable stalled state if the crew so elects. Such aircraft are not, in that state, deep-stalled, because they are recoverable. (NB The A330 is not such an aircraft in terms of architecture, simply pointing out that stability /= deep stall)

[Mad (Flt) Scientist]

Originally Posted by ttcse

MFS
As for crew hindering recovery, I am a believer that all pilots can make mistakes dealing with critical urgent situations.

Otherwise, you disagreed to my "only a deep-stall would be a stable stall" but in talking it out appeared to discount it for the A330. And I am trying to be specific to the A330 and not talk for all aircraft in general.

No, I wasn't discounting the possibility of being able to control the A330 in a post stall regime - what I was pointing out that there are definite counter examples that prove that not all stable stalls are deep stalls. Therefore there exists the possibility that such is also the case for the A330. Basically deep is not the same thing as stable. What I was illustrating is that if you have enough pitch control power you can hold an aircraft at an AOA above the stall, without it being deep stalled.

If the cg was towards the aft end of the range then it's definitely possible, in my mind, that there was enough control authority in the pitch axis to hold the aircraft in the stalled condition. Depending in detail on what the (natural) stall characteristics of the aircraft are like of course - if it's a root stall I'd expect a fairly large pitch down that might not be controllable. But if it's mid wing, then it might have a post-stall solution. If its a tip stall then there'd be a pitch up, but in that case I'd expect the roll axis to be a handful which would preclude stability in that axis in all likelihood.

[ClippedCub]

Any aircraft with a post-natural-stall stick pusher system has to be capable of a stable stalled state if the crew so elects.

Stick pushers aren't designed for post stall, but an accelerated stall can preclude stick pusher activation. They are intended to prevent stall. The A330 configuration won't exhibit a 'stable', or deep stall characteristic, ccrew desire or not. It will buck and roll and yaw when stalled and will not find a naturally stable position in the baseline configuration unless there is a spin element, i.e., normal or flat.

It could have been bucking all the way down, or the aft cg, and/or, reconfiguration of slinging engines and/or breaching fuel baffles could have set it up for a flat spin. The tail in this case wouldn't be enough to overcome the stalled condition. A normal spin would have imparted too much energy into the airframe on impact.

Somebody mentioned the spoiler removal as perhaps reverse flow when extended, either in the air, or somehow the airplane was traveling backwards on water entry. Another explanation would be retracted spoiler would have been ripped off by the hydraulic shock of a high AOA impact.

As an example of a similar configuration in stall, take Flight 903, an A300,

DCA97MA049

NTSB Identification: DCA97MA049 .
The docket is stored in the Docket Management System (DMS). Please contact Records Management Division
Scheduled 14 CFR operation of AMERICAN AIRLINES
Accident occurred Monday, May 12, 1997 in WEST PALM BEACH, FL
Probable Cause Approval Date: 2/11/2000
Aircraft: Airbus Industrie A300B4-605R, registration: N90070
Injuries: 1 Serious, 1 Minor, 163 Uninjured.The flight was assigned an airspeed of 230 knots and cleared to descend from FL240 to 16,000 feet in preparation for landing at Miami. The FDR indicated that while the autopilot was engaged in the descent, the power levers moved from the mechanical autothrottle limit of 44 degrees to the manual limit of 37 degrees. As the aircraft leveled at 16,000 feet the airspeed decreased. The F/O began a right turn to enter a holding pattern and added some power, which stabilized the airspeed at 178 knots. However, the right bank and the resultant angle of attack (AOA) continued to increase, despite left aileron input by the autopilot. As the autopilot reached the maximum input of 20 degrees, bank angle increased past 50 degrees, and the AOA increased rapidly from 7 degrees to 12 degrees. At this point the stick shaker activated, the autopilot independently disconnected, the power was increased, and full left rudder was used to arrest the roll. The bank angle reached 56 degrees, and the AOA reached 13.7 degrees at 177 knots. The aircraft then pitched down, and entered a series of pitch, yaw, and roll maneuvers as the flight controls went through a period of oscillations for about 34 seconds. The maneuvers finally dampened and the crew recovered at approximately 13,000 feet. One passenger was seriously injured and one flight attendant received minor injuries during the upset. According to wind tunnel and flight test data the A300 engineering simulator should adequately represent the aircraft up to 9 degrees AOA. Unlike the accident aircraft; however, the simulator recovered to wings level promptly when the lateral control inputs recorded by the FDR were used. The roll disagreement between the simulator and accident aircraft began at 7 degrees AOA, and it appears that some effect not modeled in the simulator produced the roll discrepancy. Just prior to the upset the accident aircraft entered a cloud deck. The winds were approximately 240 degrees, 35 knots, and the ambient air temperature was approximately minus 4 degrees C. An atmospheric disturbance or asymmetric ice contamination were two possible explanations considered, but unproven.
The National Transportation Safety Board determines the probable cause(s) of this accident as follows:
The flightcrew's failure to maintain adequate airspeed during leveloff which led to an inadvertent stall, and their subsequent failure to use proper stall recovery techniques. A factor contributing to the accident was the flightcrew's failure to properly use the autothrottle.

Don't be surprised by the sym not matching the flight trace. This is a non-linear flow regime and the wind tunnel data would be all but useless when extrapolating full scale Reynolds numbers, and the flight test data would be limited, and taken in controlled conditions.

[Mad (Flt) Scientist]

Originally Posted by ClippedCub

Stick pushers aren't designed for post stall, but an accelerated stall can preclude stick pusher activation. They are intended to prevent stall.

Not true for all aircraft. On some types they are used either to provide stall identification and/or to prevent excursions to high AOA where a deep stall IS a concern, but they are set to fire post CLmax (i.e. with the wing stalled) in the interest of maximizing aerodynamic performance. But as the A330 doesn't have them anyway thats tangential.

Originally Posted by ClippedCub

The A330 configuration won't exhibit a 'stable', or deep stall characteristic, ccrew desire or not. It will buck and roll and yaw when stalled and will not find a naturally stable position in the baseline configuration unless there is a spin element, i.e., normal or flat.

That really depends on the details of the stall location and progression. I can conceive of it doing as you say; I can also conceive of it being relatively stable/controllable. It all depends where the wing stalls, how much of it stalls, and how sensitive it is to AOA shifts in terms of progression.

[ClippedCub]

I can also conceive of it being relatively stable/controllable.

The stick pusher can be post stall for accelerated stalls. The evidence, and experience points to sustained stalls on conventional airplanes being an unstable event unless stabilized by gyroscopic forces.

[RatherBeFlying]

Re turbulence: in general terms, the certification requirements require that an aircraft can withstand the ‘worst case’ gust (turbulence) from both aerodynamic and structural perspectives when flying at Vra; at slower speeds, a stall (or roll upset) is possible.

I would add that the 'worst case' gust is a certification standard that nature can and does exceed at times, perhaps at the 10^^-9 frequency.

[philipat]

VS torn away from AFT?

surplus1:

Perhaps most important – at this point in time – we cannot prove how or when the VS separated from the aircraft. We only know that it did. The rest is theoretical.

I thought we also knew from the BEA that the VS was ripped off FROM AFT. I thought I also read that one of the spoilers had been ripped away by forces in the wrong direction, but maybe I'm wrong. This begged the question of how the aircraft could, apparently, have entered the water tail first and hence my earlier question about whether this was consistent with a spin or flat spin. I guess an attendent question would be how else the VS could be torn away from AFT?

Again, I am just trying to learn and these are questions directed to the better informed herein. I stand more than prepared to be corrected on any of these observations and mean no disrespect if my comments appear to be contradictory in any way.

[ClippedCub]

My theory is the after body was ripped away from the VS on impact, not the other way around. Either that or the horizontal separated downward in flight ripping the after body away from the vertical. The spoiler was ripped forward if extended, or upwards if retracted. A belly flop would be consistent with a stowed spoiler.

[ClippedCub]

Regarding that damaged spoiler. It could've been due to secondary flow of water after impact rather than initial flow. If at impact water was pushed away from under the wing there would then be a reversal toward the wing. Also possibly a wave might've done that. The spoiler was made to take forces from the front, who knows what it was or wasn't capable from the rear. Significant strength from backside forces probably not a requirement.

Would have to take exception to that, built pretty tough from the pictures and the design condition. Wouldn't be much weaker in the deployed, reverse flow case. The spoiler was yanked out by it's roots, and would most likely be a consequence of high rate of descent impact.