I think it's fair to say my focus on control surfaces might be annoying. If so, I simply find BEA's explanation wanting. I would like nothing more than to simply say, that's it!

The colored surfaces in red of the missing controls say quite a bit. It is possible of course that the airliner just sat in, with the VS equilaterally driving vertically down into the tail's support structure, then being spit out, with only a leading edge squash left instead of shrapnel, (spoiler). Then too, takata points to a lack of hydraulics problems on the ECAM. As the linked essay states, the comparable "peace" of a deep stall, potentially cueless, (compared to buffet and airstream noise of other ua's) might have been the last attempt to establish controlled flight. Controls that require airflow parallel to their surface are useless in vertical mush. Especially when trying to move a stubborn 150 tons

I have some folks working on animations of possible paths 447 took downward.
I'll look at them and decide if they are appropriate.

bear

Sensor_Validation:

Here's a graphic overlay of the latest BEA Search Area Update, 17 May, 2010, centered on the Last Known Position. As you can see, no part of the central 20nm diameter area has been searched. Some search traverses were made northward and eastward about 12 miles out from the LKP. The blue area directly north of the LKP is very deep.

The bodies seemed to drift in a cluster along a course setting due northward from this central area (see post 1120).

No reason to believe the aircraft didn't end up southward of the LKP, either. Once the upset developed, heading became irrelevant.

http://i958.photobucket.com/albums/a...earchZones.jpg

GB

Agreed. So unless there was sufficient forward motion through the air, the V/S would have no stabilizing effect, and neither would +/- 4° of rudder. Likewise HN39's comment regarding a yaw to starboard and the right wing dipping slightly because of the V/S moving to port would have caused a small roll effect, also comes unstuck.

The scenario of a high vertical acceleration with a small nose up pitch and wings nearly level, equates more to a slowly rotating flat spin where all the control surfaces had no affect because of the lost lift and the vortex created. The fact that the BEA reported that damage to the hull was symmetrical leads me to believe that the aircraft's heading nearly matched the real direction of "flight" [travel] at the point of impact.

HN39;
Your description of the forces involved on the V/S at the point of impact is appreciated. They have of course exceeded the magical "36 g"!

mm43

I wonder if Dr. Ballard will be called in, if anyone can find something on the bottom, he's the man for the job. With Titanic, Bismark, Yorktown and several others, I'm sure he would do well. I hope that it will be found

jsfboat
I wonder if Dr. Ballard will be called in, if anyone can find something on the bottom, he's the man for the job. With Titanic, Bismark, Yorktown and several others, I'm sure he would do well. I hope that it will be found

I think this gentleman was the one who found the location of the HMAS Sydney off Western Australia.


Mike

The FAB screen shot shown in post #1127 shows the early search was a grid search along the route from the LKP to TASIL and it commenced 10NM before the LKP and included 10NM each side of projected track to about 42NM from LKP. A more intensive search was done from LKP to TASIL at about 3NM each side of track. The finding of some unidentified debris to the SE of the LKP disrupted any further searching in what we now believe was the area of real interest.

mm43

A simple SLF asks
 

Has anybody thought about the main body of the aircraft not sinking to the bottom of the ocean?

Some composite boats reach a depth in the water where the hull is just the same density as the water and stop there,and just float with the tide,or so I have been told.???:confused::confused:

crippen
 

The mainly composite part broke off, floated, and was fished out.

The rest of the structure is metallic. It's not airtight, so it will sink.

Where? Where? Where?
 

Hello GreatBear,
In fact, this is the first place the BEA searched with its best assets (USN's dedicated TPLs = pinger locators) when the recorder's balises where still working.
You should remember that the search operation is in phase 3 and that the previous two phases were detailed in BEA 2nd report, annexe 2, p. 77-86.

In short: they defined the most probable crash zone called "Alpha zone" starting about 10 nm South of last know positition and extending to 10 nm South of Tasil. This is the black box pictured in the map below:
http://takata1940.free.fr/search1.jpg
Each grid box is 10 x 10 nm.
When pingers stopped to work, they assessed the above "quality" map to continue into phase 2 where other means were deployed by oceanic survey vessel Pourquoi pas? consisting of deep sonar robots (i.e. J-K-L-M 24 previously skipped boxes were searched this way). The goal was to systematically finish the Alpha zone and eventually also the 40 nm radius defined circle centered at LKP.

As you can see it now, phase 3 is completing the previous work using better reverse drift models newly developped, but they were also scanning already searched boxes having low quality scores (mostly from Emeraude's assigned zone).

Taking into account the very high probability that TPLs would have picked any signal from the recorders along the flight path to TASIL (where the relief is much more gentle), it is very probable that AF447's wreck is not located into this zone... or, if it is really there, that both pingers were destroyed during the crash.

You are right considering two big 'IF'.
- If upset did take place at 0210... [nothing is telling us it did].
- If end of flight was ~0215... [there is still very good chances it was not].
About the time it could take for AF447 to go down, something as fast as 1 minute and half is possible from cruise level. Then if ~0215 was the end of flight, an upset could have occured as late as ~0214, making quite some distance between 0210 (LKP) and crash site.

Consequently, the next major step into this investigation would have to be the localization of the wreckage which, by itself, will reveal a lot more about this upset/crash sequence. Last declarations (yesterday) are telling us hoppefully that the search will not stop here and big efforts will continue until they find it.

S~
Olivier

Boeing Aeromagazine on Upsets
 

In his post #1358 sensor validation offers a link to an article by "Dark Matter". While that article is mainly the usual mix of truths, half truths, and misconceptions, it contains a link to an interesting article in the Boeing Aeromagazine on Upsets, that I would like to recommend to many readers of this thread:

Boeing Upset article

HN39

HN39

Very interesting.

Well, I´m new here, so forgive me if this was already asked by someone, but what are the prerequisites to becoming an airline pilot and, specially, to piloting large aircraft like an A-330 or a 7x7?

The article you linked to stress the need of pilots to be familiar with the laws of physics. It seems to me that this would constrain the pool of would be pilots to engineers, physicists and mathematicians. Maybe some economists also. :)

What is the usual extraction of the pilot workforce in the airlines?

ps: I´m a brazilian that for profissional reasons fly over the south atlantic quite often, including in the ill-fated AF447. I have read this thread from the beggining, and also the old one. I´m an engineer, but have no training in the aeronautics industry, and must confess that I´m learning a lot from the very competent commentators here.

Is this where ignorance is bliss ?
Have a look at this site, for instance Smartcockpit flight ops docs

or check John Tullamarine's sticky thread " Useful websites and document references" in TECH LOG.
some knowledge in these two places.

and , of course, your hunch is wrong.

Regards

Bearfoil
Bearfoil,
If I may ask : What is your profession / personal connection to this accident ?

Are you a lawyer or journalist or something completely different that you have people working on this?

Hello Centrosphere,
It is not the same depending on countries. Concerning France, the main courses are ENAC (École Nationale de l'Aviation Civile, in Toulouse, one of the French prestigious "grandes écoles") and, for older generations also l'École de l'Air (Armée de l'Air). Both course are following a very serious scientific baccalaureat plus preparation and will provide a diploma at least equivalent to an engineer.
See ENAC page here: ENAC - Ecole - Accueil
S~
Olivier

Lemurian;

Thank you for your useful and very welcome link to the Smartcockpit site. I wasn't aware of it and certainly still have to learn a lot about sources of information freely available on the web. On second thoughts, I have removed the remark from my post.

Regards,
HN39

certification
 

The second BEA interim report into AF447, (p62-67), identified 32 other incidents caused by or related to freezing of pitots on A330-A340 and also stated that the test/certification regime (JAR25) for those pitots was for a maximum altitude of 30,000ft.
Why is it that one of the most heavily regulated industries in the world allows almost all of its commercial operations to be conducted using a critical speed sensing technology that is neither tested nor certified for critical icing factors at the very altitude that almost all of their commercial flights are conducted.
Maybe they should limit all flights to within the certified envelope (30,000ft) and wait for the business out-cry to put funds into solving the problem.

Where?
 

Takata,

Aaaaah yes, there are so many Unknowns! So many IFs. So many Possibilities. So many Questions.

My bad, not to clarify "BEA have not searched there yet" by adding "with side-scanning sonar." Appendix 2 of the 2nd Report includes a graphic ("Lanes defined in the search zone") and states "The SAR was operated on line 24, squares J,K, L and M." That's eight to ten nautical miles southwest of the Last Known Position, so it seems that IFREMER’s Towed Acoustic Sonar (TAS), with its 1 pixel for 25 cm resolution, only explored a tiny corner of the area within a 10nm radius of the LKP during Phase 2.

It's unfortunate that the very best assets (including a nuclear submarine) and TPL technology could not locate the pingers during Phase 1 in the Alpha Zone or elsewhere before their predicted battery life expired. As you say, if the wreckage is really in a tight circle near the LKP, both pingers could have been destroyed during the crash. Recommendations are already in place for lowered beacon frequencies, and pinger specs and CDR methodologies are undergoing a broad re-evaluation by the industry.

The timing of events with clues only in the ACARS messages is very tricky. My thinking is that for the 02:14:26 cabin vertical speed message to be sent, it had to be triggered by a rapid descent at least a minute or two PRIOR to its transmission. As the BBC experiment showed, a fall of 19,000 feet in 50 seconds might be a worst-case possibility. So the actual falling event must have begun well before that last message was timestamped and sent. In any case, it would seem that loss of control (the upset) occurred closer to 02:10, well before the cabin vertical speed message was sent. The idea of straight-and-level cruise flight much beyond AP Disconnect at 02:10:10 is difficult for me, but I am continually trying to simplify a very complicated event.

Aaaaaah, the unknowns.

Without doubt, the BEA search effort has gathered the best minds and resources to recover the hull and black boxes. It's simply one of the most difficult researches so far undertaken in this 21st Century.

Thank you for your considered feedback.

GB

The French are already familiar with Dr. Ballard, and Paul-Henri Nargeolet has an opinion:

http://www.titanicfiles.org/Essays_e...%20Ballard.pdf

yes, where!?
 

GreatBear,

Good point! my opinion too as one really need to keep open the right possibilities and discarding the wrongs only very very carrefully without being subject to tunnel vision.

There is still a good hope that the Navy picked the right signal. As a former member of this institution, I'm very confident about it, even if MOD's communication nearly ruined its credibility against BEA (to be verified) opinion of "no aircraft South of LKP".
This cabin vertical speed message is still under high scrutinity. It will be very difficult to fully understand it without knowing precisely what was the exact configuration of the aircraft when it was triggered, but also, because it was the last one of the sequence while few more ACARS could have told a lot more about it. We can't know for sure that the flight ended strait after or a while after.
In fact, it left only three choices open: a) crash in the following seconds or b) ACARS failure; c) EMER-ELEC config.

This advisory can be triggered both in PRESS auto or manual mode (crew action). If PRESS manual mode (for whatever reason) was selected when the valve was open, cabin vertical speed could climb quite fast. All we know is that during 5 seconds (at least), the cabin moved (up or down) by 150 ft (at least). If PRESS was in auto mode and no ADR was valid, it is also possible, beside other faults, that the security will work as to keep the DeltaP (internal/external) constant based on other data (then move at the same rate at the aircraft from whatever altitude). AIRBUS FLIGHT SAFETY DEPARTMENT communicated about it that its meaning was not obvious:
But, as it was the last ACARS sent, and like EMER-ELEC could have followed it - without being transmitted due to ACARS shut down - a wider range of possibilities could even occur (i.e. was it the consequence of engine roll back on PRESS/VENT systems before flameout?, etc.).

The BBC show was overall good (I feared worse) but there was still some fair amount of drama over reality; The third pilot would not have been PF if the commandant was resting (which is still unproved but quite possible). More annoying was the fact that they used an A320 simulator showing wrong settings (mach 0.76 in place of min. 0.80) and did not follow the real SOP -quite conveniently- for addressing their "non-moving throttles have been forgotten" theory (not adjusted to right settings), then the aircraft was stalled in heavy turbulences.

But as ATHR was ON, then kicked OFF with AP, it would also trigger a warning every 5 seconds until throttles would be adjusted into the right settings manually. Consequently, it would be pretty hard to forget about thrust and let the airspeed drop down to any dangerous level without noticing it. If turbulences were already an issue, ATHR would be OFF and the correct parameters of pitch and thrust already set before this pitot issue... In short, this is unlikely enough to produce an upset at this stage unless unknown issues or maybe, the appearance of all factors at exactly the same time (freezing, multiples faults, and unexpected heavy turbulences).

Concerning the stall demo, it is hard to tell as important details about the military simulator settings are left vague but it might be possible to go down in about 90 s without breaking appart the airframe. In this case horizontal distance covered would be quite limited and aircraft terminal attitude unlikely level.

S~
Olivier

I am not sure whether any references have previously been made to the Swissair Flight 111 MD-11 crash of 2 September 1998. I would have thought a lot of knowledge was gained in underwater search and recovery operations and techniques from this accident. It took TSB of Canada four years and cost CAD 57 million (approx USD 38million ) to get the answers. The aircraft had crashed closer inshore and the debris lay relatively in much shallower waters.
In the case of AF447 the cost to date is EUR 9 million and the depth and extent of the search area is considerably larger. Should the FDRs and the CVR be eventualy recovered it is not an absolute certainty that the answers will be provided. Should the wreckage be found what would be the challanges and difficulties in recovering it from the great depths of the ocean. What may the final cost amount to.

How much should be spent on accident investigation for the sole benefit of flight safety. Surely in these days of global economic austerity there must be a limit.

auv-ee

The French are already familiar with Dr. Ballard, and Paul-Henri Nargeolet has an opinion:

http://www.titanicfiles.org/Essays_e...%20Ballard.pdf

Thank you for that insight and reference.

In return, may I recommend you have a look at the following site regarding the HMAS Sydney and in particular the Volumes 1-3 of the outcomes of the search and find.

The Search for the HMAS Sydney II Report Available for Download - Official Press Releases

They were also searching in deep water but no where near the depth of water AF447 is resting .

Thanks again for providing that reference - it enables one to get a better understanding of matters.

Mike McInerney

Anyone aware of an alternate site hosting this program? The video streaming from that site...well, at least here, isn't.

"Without doubt, the BEA search effort has gathered the best minds and resources to recover the hull and black boxes. It's simply one of the most difficult researches so far undertaken in this 21st Century."

I'm sorry but I could not disagree more strongly. The stipulation within the last tender issued that AUV's would be used for this search raised eyebrows amonst many involved in the subsea search & salvage industry.

An AUV is designed to fly over ideally a flat sandy bottom flying a pre programmed route using waypoints, ideal for a lengthy cable / pipeline / route survey. The terrain involved in the AF447 is about as far from ideal as you could get for the use of AUV's. The second factor being the battery life of an AUV and the time taken to submerge to 4000m, basically by the time it gets down there there is very little time left to actually do any searching before it then needs to start the journey back to the surface.

Basically the use of AUV's are a great way to spend a lot of time messing about without actually getting much searching done, I'll leave it at that.

Currents and debris NE of LKP
 

Failure to so far locate F-GZCP to the N, NE, NW, W and SW of LKP prompted me to revisit information on the currents and surface debris locations.

I'm well aware of mm43's analysis suggesting a location SW of LKP, relatively close to the area resulting of Meteo France's reverse-drift analysis, and until recently reinforced by the French Navy supposed pinger data.

A SW of LKP location is of course still not to be excluded, but one should beware of tunnel-thinking. Although BEA's recent stance that F-GZCP "has never been south of LKP" may equally be qualified as such, please consider the following:

Unless there was an immediate catastrophic event @ 0210 (other than iced pitots) there is little reason to assume that F-GZCP would be instantly that much off course that it could within 4 minutes end up at a location that would require (significantly) more than a 90° turn. To maintain heading would normally be a priority for a crew dealing with a series of events as it was confronted with according to the ACARS messages.

And if weather ahead would have required a change of course under those circumstances, it would make sense to turn the least possible. Given the N/NE course of F-GZCP, that would imply a turn to E rather to W.

Not only have two US reverse drift analyses suggested an impact point to the NE and NE/E of LKP respectively, also data of the Mercator model (used by Meteo France) do not exclude that.

I'm not privy to all data available to Meteo France used for their reverse drift calcs, but the Mercator based graphs I got hold of show that although on 27 May (closest I could get to 1 June) the prevalent current NE of LKP was E/NE, on 2 June the current was W/NW, while turning N on 9 June. US model based data are even more clear on a prevalent W/NW current.

Taken together this would IMHO not rule out that F-GZCP after its LKP initially pursued along its heading followed by a (relatively shallow/controlled) turn right and possibly ended up fairly E/NE near the 40nm search perimeter. We know that this area wasn't covered that well initially and certainly has not been covered recently.

Such an E/NE point of impact could also better explain the three debris items found in that area (see map page 37 of BEA's 1st interim report). Strangely enough these three debris items do not figure in the 1st interim report's annex 4 where a day by day location of the finds is given. So, unfortunately it is impossible to say when they were found and, equally important, what they are (in terms of how high or low they would float).


Dutch

Gwillie
 

I managed to download it and have it shared in Winmx..
Also, you could watch/download it using the BBC iPlayer from their site
If you are not in the UK you will have to use a proxy IP..

Dutch, the area E and NE of the LKP was searched extensively between June 2 and June 5, and nothing was found. How do you reconcile your hypothesis with the first body and wreckage being found W and NW of the LNP on June 6?

It's available on usenet, for those of you familiar with that concept. In alt.binaries.multimedia.

Hi,

Just tested and the streaming fom Zshare still available ....

I'm curious to know what alternative you propose and what its search rate and percent coverage would be in mountainous terrain. All other systems I can think of are tethered and have issues of their own.

While any search system works better on a flat bottom, AUVs have characteristics that make them well suited to this work:

1. I covered the search rate parameters in post http://www.pprune.org/5692187-post1029.html
It works out to about 40 square kilometers/day for double coverage by one vehicle. [BEA: 100sq-km/day/AUV, maybe for single coverage.] A claim of much time lost in vertical transit is simply incorrect. REMUS-6000 descends and ascends at an average rate of about 60m/min, so it can make a one-way excursion to 3500m in 1 hour. With a 20-22 hour mission time, 2 hours lost in descent/ascent, and a 2-4 hour turn-around at the surface, that yields 18-20 bottom hours out of every 20-26 hours. Even with the worst of those values, the vehicle is on the bottom 70% of the time.

2. When surveying the bottom, the AUV transits at 4 knots. All the towed systems I know of are limited to 1-2 knots in deep water, to prevent the cable drag from lifting the vehicle off the bottom. [BEA: Orion system is listed at 2kts. BEA also lists 100sq-km/day/towed sonar, probably using lower frequency, lower resolution, but longer range to offset the slower tow speed.] Perhaps if the entire cable is faired, then the speed could increase, but that would significantly increase the size and complexity of the cable handling system.

3. AUVs spend no time in turns (other than the few minutes it takes to cover the offset distance between survey lines). [BEA: 15min for AUV.] Towed systems follow the ship track with 30-60min delay in deep water, and it can take 6-12 hours to align with the next track. There are perhaps operating skills that could shorten that time, but it won't be minutes. [BEA: Orion system turn time listed as 3hr.] Turn time is to the towed system, as vertical transit/servicing is to the AUV.

4. AUVs can follow the bottom more closely in rugged terrain than towed systems, at least more closely than I have seen towed operators willing to attempt. Contact of a towed system with the bottom in rough terrain carries the risk of fetching the vehicle under a ledge and breaking the cable if the winch operator and bridge watch don't react in time. For an AUV, collision with a cliff is a survivable event. While no one plans a mission to include a collision, it does happen and it is recoverable, usually with the mission continuing as planned. Aside from a collision, an AUV can follow terrain at about +/-30deg (depends on the AUV). Certainly there are steeper parts in the AF447 search area, but a lot of it is within +/-30deg, especially with "terrain aware" route planning.

5. Multiple AUVs can be operated from a single ship, further increasing search rate for a nominal increase in day rate (cost).

6. Navigation of track lines is generally more precise with an AUV, because there is no influence from ship motion. This improves coverage (fewer holidays) and becomes especially important if the debris field of AF447 is ever found, when the AUVs can be used to photo mosaic the field.

After the wide area search and initial photo work are complete, then tethered systems will be needed to finish any detailed survey or recovery.

Certainly, towed/tethered systems have advantages too. The most obvious being that the sensor data is available in real time, and similarly there is much higher bandwidth for control of the vehicle.

Enough said. I'm ready to be educated about the system(s) you have in mind.

Pitots freeze
 

How about using secondary navigational inputs on GPS based data?

The weather models suggest AF447 was inside the storm system @0210. They had reduced speed from M0.82 to M0.80 at some point. The final ACARS position report which probably reports the position at 02:10:30, shortly after the incident started, shows deviation West from the planned route, but with a precision that the BEA report does not appear to believe significant

AF447 ACARS MESSAGES
The full set of ACARS position (and turbulence?) reports do not appear to be in the Public Domain. There are clearly false reports which appear to give a location for the 2:10:14z message.

The BEA 40NM search radius would appear to assume less than 1 minute level flight plus 10,000 ft/min rapid descent - keeping airframe substantially intact, triggering pressure advisory below 8,000ft and end of flight less than minute later.

It seems to me most likely that loss of control occurred almost immediately, with one cascade of problems starting with the pitots - detected at 2:10, but possibly giving faulty data before?

Hi,
a) sorry, but which weather model is suggesting that?
BEA (position) + Méteo France (0215 weather sat map) data -see reports- are showing that AF447 was close to the coldest part of this active CB system @ 0210 (and roughly half way - but previous meteo analysis from Tim Vasquez was showing the wrong position on the right map).
b) Where did you get this 10 nm W of flightpath?
0210 last know position was exactly recorded 2.94 nm W of flightpath with an error margin inferior to +/-0.3 nm (eurocockpit source):
http://www.eurocockpit.com/images/UN873.jpg
http://www.eurocockpit.com/images/UN873Z.jpg

... way too much reading into this estimation!
This 40 nm circle was drawn from day 1 barely without any data analysis and absolutely no wreckage to study: 5 mn of flight (ACARS transmitted) at speed inferior to 500 knots = 40 nm (= 5.000 square nm). Alpha zone (based on higher probability) was even extended near TASIL and the Southern part, passed 10 nm from LKP, was discarded.

It doesn't seem most likely at all -until someone will find where the wreckage is.

Probability that it was not discovered by sea scans inside already searched zones is currently very low (both pingers destroyed).
The most unreliable part of the search so far was the airborne SAR operation during the first six days due to the lack of adequate means (site distance, very few aircraft, no helicopters), inadequate weather (low clouds, poor visibility, rain, sea state), disturbance by floating garbage (high seas are full of floating debris), dispersion with time (floating stuffs -nearly fully submerged - finally covered a large area that made it very hard to be spotted from the air).

The probability that all three sensors were giving simultaneous faulty data before 0210 is many many times more unlikely (such an issue would more likely start from the ground). More precisely, it is almost ruled out by the fact that this pitot issue was detected in flight at 0210 and that relevant systems seems to have reacted as per design.

S~
Olivier

Certification
 

Hi Geoff,
You've got no answer as new post from new posters do not show up until a while and may be left unoticed. You did bring in a valid point:
It is hard to make a very simple answer to this point as this is a very complex issue. I have read few papers on the subject and this may be summarized like that:

- Until not so long ago, about 10 years, most scientists thought that no ice/water particules of any dangerous size for aircraft sensors could be encountered at altitudes above 30,000 ft due to very low atmospheric temperature. There was several long haul airliner pilots testimony saying otherwise during specific flight conditions: oceanic flight over seasonal tropical thunderstorms or at proximity of them. The problem was that it could not be reproduced in laboratory. It has to be 1) to prove the phenomenon, 2) to develop some new probe certification process.

- In fact, almost all pitot events are taking place at lower altitudes on every model of pitots, causing a lot of accidents, mostly in general aviation. Airliners are supposed to be safe due to the constant care about this known issue as well as by multiplying their sensors and backup systems. But the risk still remain that those systems could be overwhelmed by conditions not reproductibles due to actual knowledge and experimental means. It is just not possible to take a sample of one particular atmosphere and to bring it into our labs without changing it. Conditions have to be re-engineered and this seems to be the core of the problem.

- Until AF447, known pitot issues encountered in A330-340 fleet (only) lasted from 5 to 20 seconds. After AF447, other incidents occured, including the other probe models considered less sensitive than Thales AA (Thales BA & Goodrich/Rosemount probes). Actually, it seems that the probe makers are working empirically on this issue, without exactly knowing what to change in order to fix the problem, and it is the same for the regulator. Nowaday, a lot of ressources seems to be affected in order to study very seriously this phenomenon. Some scientists are pointing that it may be caused by climate changes and that it is much more frequent, others that it always existed and was dismissed, but as traffic increased, its frequency increased.

S~
Olivier

Slightly off the current topic, is it possible that they have found some of the wreckage with their current search, but because of logistical restraints they are unable to explore those sights?

BS

Ignore me I'm sure Takata is correct re facts re location/storm - there was much discussion in previous thread and elsewhere about how quickly you can get down in one piece from 35,000ft, clearly not a gentle 20:1 glideslope?

As stated in the BEA report, pitots on this type of aircraft have a history of problems. The AA probe itself was introduced to replace inferior earlier models, and then early examples had manufacture QC issue with drain hole

http://ftp.resource.org/gpo.gov/regi.../2004_5788.pdf

This earlier incident on an Airbus quantifies the effect of pitot drain hole blockage for the probe in use at that time - is much higher/ faster much different in dynamic pressure?

http://www.aaib.gov.uk/cms_resources...pdf_501829.pdf

So drain hole blockage alone is sufficient to disengage autopilot/ thrust.

Having identical probes makes it easier to detect one outlier - they should all read the same - but why would they then not fail in same way when subject to same ice cloud?

Hi Bluestar51,
Quite on topic!
They might have found something without actually knowing it (further data analysis may later reveal some possible traces of wreckage) but I don't think that logistical restraints would have been an issue. Those vessels were fully equiped for recovering any wreckage, or simply to verify any doubtfull spot at will.

For those interested, there is also some good videos about the search means used during this last campaign posted on BEA's site (three parts):
1.
http://www.bea.aero/fr/enquetes/vol....es.navires.mp4
English subtittles:
http://www.bea.aero/en/enquetes/flig...es.navires.srt
2.
http://www.bea.aero/fr/enquetes/vol.af.447/videos /premieres.plongees.operationnelles.mp4
English subtittles:
http://www.bea.aero/en/enquetes/flig...tionnelles.srt
3.
http://www.bea.aero/en/enquetes/flig...donnees.en.mp4
English subtittles:
http://www.bea.aero/en/enquetes/flig...donnees.en.srt
[in order to read the subtittles file with the video, save it in same folder with video, use vlc, mediaplayer classic, or any player with vobsub filter.]

sensor_validation,
Airflow related to probe has to be perfectly symetrical, which is barely never the case in flight, because they are on both side of the fuselage. The value displayed is a mean value of two or three pitot readings.
If one probe is failling to drain one kind of particule, it doesn't mean that it will fail to drain another kind. You'll have to know what kind of particule you are dealing with.
S~
Olivier

Median or mean?
 

Hi takata,

My understanding is that the PRIMs use the median value of the three ADIRU's, not the mean, except for the AOA where it uses the mean of #1 and #2, as explained in the ATSB report on QF72.

Regards,
HN39

EDIT:: As to displays: no averaging, see tubby linton post #1399 below

median of 3 (or bigger odd number) values is generally interpretted as you take the middle one when sorted in order of magnitude, without any averaging. This is sensible as it automatically ignores extremes and spikes values on any one channel. For even number you do average the middle pair. No difference between mean and median if you only have 2 values! A median filter also be applied to multiple samples over a time period, linear/ averaging filters tend to turn spikes into bumps.

Hazelnut -efis switching for capt is -normal(sys1),then sys3 then sys2.
FO switching is -normal(sys2) then sys3 then sys1.
Reference fcom 1.31.50 p3