AF447

[Graybeard]

Pitot Ice

Pitot probes have a liquid drain (bleed) hole, whose air bypass has to be accounted for in the airspeed calcuation. Clog the drain, and pitot pressure rises, giving erroneous high airspeed. Indeed, per reading in this thread, some of the earlier Airbus pitot malfunctions were with poorly manufactured drain (bleed) holes.

Add more ice, and you clog the pitot head itself, and then pitot pressure will lock, or will decrease if there is any drain opening at all.

GB

[Graybeard]

Wx Radar Factor

As posted earlier, the head of AF is wondering aloud if the crew of 447 had not operated the Wx radar in a way to see the storm they must have flown into.

Spend $200 Million for a new airplane, but not $200.1 Million for one with the latest and greatest Wx Radar. . Why? Because the 25-30 year old technology Wx Radar meets the minimum requirement for airworthiness. Besides, you would have to re-train the pilots (if they get any Wx training now), and retrofit the rest of that type fleet.

The "new" (~7 year old) multi-scan Wx radar would have had a far better chance of accurately displaying that storm system to the pilots. Weather radar operation and interpretation is one of the few pilot tasks that haven't been automated. Why not spend a few extra bucks and give pilots the best tools available?

GB

[takata]

pitot's holes

Graybeard:
Pitot probes have a liquid drain (bleed) hole, whose air bypass has to be accounted for in the airspeed calcuation. Clog the drain, and pitot pressure rises, giving erroneous high airspeed. Indeed, per reading in this thread, some of the earlier Airbus pitot malfunctions were with poorly manufactured drain (bleed) holes.

Add more ice, and you clog the pitot head itself, and then pitot pressure will lock, or will decrease if there is any drain opening at all.

Hi,
I'm with Hyperveloce on this one. Considering the drain's hole dimension (I've got this detail somewhere) which is very small on Thales probes, an overspeed would be barely noticeable and certainly won't reach this 20-30 kts overspeed range that this simulation would imply (maybe only few kts at max). When drains are blocked, ice will build up much faster because turned into water by the probe heaters, then the pressure will be under reported.

S~
Olivier

[Hyperveloce]

Pitot freezing: over/underestimation

I see GB, indeed, with both the Pitot tube ram air and the drain blocked, a machmeter should work like an altimeter, largely overestimating the airspeed if altitude increases (seems in the order of +40kts/+1000ft @FL350). But if altitude remains constant, it can't be an overestimation ?
Jeff
PS) following an accident wich began by Pitots freezing, obstructed or polluted Pitot drains were found on a large fraction (~20%) of the French fleet of Mirage 2000 (with a high probability of occurrence, ~70%, in humid/hot areas). the quality of the manufacturing was also a problem for a fraction of them.

[Graybeard]

OK, guys, maybe you can explain the overspeed warning had by the NW A330 into Japan recently? The static ports haven't been implicated, just the pitot probes.

I suspect it's not just the size of the bleed orifice compared to the main one, but also its alignment. If it is at right angle to the airstream, it will draw a vacuum, meaning that it is a much larger factor than mere size relationship.

[Rananim]

It's not important how they crashed but why.

http://boeing.com/commercial/aeromag...7_article3.pdf

These additional conditions are also typically
found during engine ice crystal power-loss events.
■ No pilot reports of weather radar returns at
the event location.
■ Temperature significantly warmer than
standard atmosphere.
■ Light-to-moderate turbulence.
■ Areas of heavy rain below the freezing level.
■ The appearance of precipitation on heated
windshield, often reported as rain, due to
tiny ice crystals melting.
■ Airplane total air temperature (TAT) anomalyreading
zero, or in error, due to ice crystal
buildup at the sensing element
■ Lack of observations of significant airframe icing.

Convective clouds can contain deep updraft
cores that can lift high concentrations of water
thousands of feet into the atmosphere, during
which water vapor is continually condensed
and frozen as the temperature drops. In doing
so, these updraft cores may produce localized
regions of high ice water content which spread
downwind. Researchers believe these clouds
can contain up to 8 grams per cubic meter
of ice water content; by contrast, the design
standard for supercooled liquid water for
engines is 2 grams per cubic meter.

Another possibility that some have mooted on the internet is that the aircraft was being blown around a lot in severe to extreme turbulence, but I don’t see how thereby one would get discrepant readings: rather, all probes would vary wildly, but coordinated, as individual gusts hit all three at more or less the same time. So I really don’t see that as a plausible reason for the P-S system issues.

Ladkin

Severe icing alone overwhelming the sensor systems, though, does not by itself lead to an accident. The AC could be controlled with pitch and power, and the Aircraft Operating Manual explains exactly what pitch and what power setting in some detail, if one has an “ADR disagree” warning.

Ladkin

“Of several acceptable explanations for a phenomenon, the simplest is preferable, provided that it takes all circumstances into account.”

Occams Razor Theory

Multi-sensor blockage combined with dual flameout?Doesnt this fit like a glove with the least assumptions(ie.just one assumption:high-alt ice crystal icing alone ) and no contrivance(experienced crew flew into cb etc).Its been posited before I know.Is it a match for the ACARS messages,esp the last advisory re cabin alt?ie CPC affected by flameout/stall/surge prior ACARS disabled due total main AC outage?
Im just thinking that all this talk of control laws,jet upset,wx radar gain/tilt/brightness,rudder etc may be too convoluted and unnecessary.

A 22h22 et45s,les "ENGINE ANTI-ICE" sont places sur "ON"

Air Caraibe report
This was the first pilot action in response to the weather phenomenon they experienced that night and occurred only 9 seconds after the TAT registered -5C and 14s before erroneous airspeed.

[ttcse]

Rananim

Im just thinking that all this talk of control laws,jet upset,wx radar gain/tilt/brightness,rudder etc may be too convoluted and unnecessary.

Your double-flameout is less convoluted only because you ommited the 'and then.....'.

[takata]

Conclusion need time

Graybeard:
OK, guys, maybe you can explain the overspeed warning had by the NW A330 into Japan recently? The static ports haven't been implicated, just the pitot probes.

I suspect it's not just the size of the bleed orifice compared to the main one, but also its alignment. If it is at right angle to the airstream, it will draw a vacuum, meaning that it is a much larger factor than mere size relationship.

Difficult to take NW incident as an example as very few details are actually available yet. It is missing most of the details: "large airspeed fluctuations, small altitude fluctuations, and an overspeed alert." That's all.
It looks like the fluctuation is due to sequential probes clogging/cleaning/clogging... then, at some point, over-pressure also. But from all other more detailed reports, when the probes are blocked for few minutes, the speed is always under-reported. And it is very difficult to know exactly where and how they were blocked as, once landed, no ice remains. The same issue is reported from laboratory test as it is difficult to re-create some unknow conditions in order to simulate and study this particular failure.

S~
Olivier

[safetypee]

Graybeard, (# 3409) IIRC it was static pressure that was implicated – altimeter fluctuation.
But does the altitude computation have an airspeed correction for static error proportional to speed?
If so, then static pressure need not be considered.

[vapilot2004]

Hyperveloce:

The situation as described in the real world would have brought at the least an ADR Disagree and would require cross checking your AOA and A/S data. Meanwhile, the prudent gents would have their eyes on the indicated altitude & VS during all of this and adjust the pitch and power appropriately.

The story credited to "pilotaydin" lays out an uncertifiable behavior for any aircraft. I say the sim was unrealistically programmed, likely on how pitot/air data failure modes were setup.

To go back to AF447, an ADR disagree/unreliable airspeed at that altitude, in moderate to severe turbulence and possible vertical air currents, could be a tough nut to crack on any aircraft, flight envelope protections properly working/fitted/designed or not.

[HarryMann]

Rananim

Indeed, a sequence to consider?

Some things are on your side
No real evidence of 'extreme' turbulence in that area (others avoided it)
Combined effects of Flight Instrumentation/degrading systems AND engines tp diagnose/respond to e.g even higher cockpit workload

You would have to make a case for:

No Maydays, ELT
Inability to re-start lower down
What ACARs tx this would or wouldn't have generated

Are you suggesting the combined effects of lost FI and thrust loss would have created a rapid stall/upset?
Or a gliding descent into possibly worse conditions?

[takata]

ttcse:
Your double-flameout is less convoluted only because you ommited the 'and then.....'.

I'll try it:
When the pitots-probe/TAT were assumed frozen by the pilots, they would have read in the QRH that any "stall warning" should be respected. As they experienced undue stall warnings from the begining of the sequence, without noticing any buffeting nor decreasing aerodynamic noise; they decided to ignore it. But, in order to keep a safer margin, they decided to reduce their altitude, just in case. Thrust was set to idle for a safe descent, and then, an engine flameout occured because some very small ice particules in this saturated Anvil had accumulated over the engine for a too long time.

S~
Olivier

[takata]

flameout

HarryMann:
You would have to make a case for:
No Maydays, ELT
Inability to re-start lower down
What ACARs tx this would or wouldn't have generated

1. No Maydays heard, doesn't mean they didn't try; radio com problems reported all over this zone (weather?) and radio problems with this particular aircraft.
2. or engine refight too late. The process is not easy, based on speed and altitude which will lack in the first place; need to troubleshoot many other problems at the same time; need to make distress call with unreliable radioset due to lack of main power.
3. NIL...no more ACARS in case of double engines flameout.

S~
Olivier

[lomapaseo]

1. No Maydays heard, doesn't mean they didn't try; radio com problems reported all over this zone (weather?) and radio problems with this particular aircraft.
2. or engine refight too late. The process is not easy, based on speed and altitude which will lack in the first place; need to troubleshoot many other problems at the same time; need to make distress call with unreliable radioset due to lack of main power.
3. NIL...no more ACARS in case of double engines flameout.

You can't push a theory unless you also consider the cons against the theory.

But just to keep the discussion alive, give a thought to what happens to the flight after the engines flame out and ACARS goes off line with nary a clue that the individual electrical generators preceeded it.

[takata]

lomapaseo:
You can't push a theory unless you also consider the cons against the theory.

Sure, but 99.99% of this thread is only about the cons against this theory. The basic cons are just that we don't know what happened for sure and it will be really difficult to prove anything without finding the FDR/CVR. But we also need to use the few facts we already have and be able to fill the huge gaps in between.

Rananim basically used what we know for sure: ice was there and certainly in enough quantity to make something ugly to happen to engine like in certain cases already seriously studied because it could harm any one of the most modern engines (there is plenty of papers about severe icing, at previously unknown levels, during the last four years as it appears to be an increasing threat for high flights above tropical weather - read what NW pilots said about the conditions they met in Japan).

But just to keep the discussion alive, give a thought to what happens to the flight after the engines flame out and ACARS goes off line with nary a clue that the individual electrical generators preceeded it.

ACARS should not report flameout because it takes some time to trigger an ACARS and no time to switch to EMER ELEC once engine flameout. Then Satcom will go off-line no matter if any ACARS is on the pipe.

What happens then? with an unreliable airspeed, altitude, direct law, etc. The pilots will be busy for quite some time in order to sort things out. They'll need to maintain 300 kts for a refight, one engine at a time. At this speed, they would be at sea level in less than 15 mn. Considering they'll have no reliable airspeed until much lower, they could have missed the conditions and ended too low. Then hit the water (it's night and certainly under heavy rain) in an attempt to refight an engine. For example, hitting hard with the tailcone when making a ressource, killing horizontal energy, then pancacked. Another possibility will be a stall at low altitude after such an attempt, successfull or not.
The other point which make me think this kind of scenario a serious possibility (flameout) is that I'm pretty sure they turned back: if not, it is just not possible to reconciliate the actual drift of the wreckage with any fast upset at altitude.

S~
Olivier

[ttcse]

takata

What happens then? with an unreliable airspeed, altitude, direct law, etc. The pilots will be busy for quite some time in order to sort things out. ...

I wasn't arguing against power failure, but against the notion that it would offer an even simpler Occams Razor. To make my point I was fishing for the 'What happens then?"

The other point which make me think this kind of scenario a serious possibility (flameout) is that I'm pretty sure they turned back: if not, it is just not possible to reconciliate the actual drift of the wreckage with any fast upset at altitude.

I would like to re-offer another theory on crash point, debris drift, and course deviation.

To cut to the chase, they came down west of course but near the zone of eastward current flows. Then they drifted eastward/east-north-eastward before then currents took debris more north-north-eastward. Somewhere in the interface between the eastward zone and what's north of it is a region of very slow moving current. Just to put that card on the table as a potentially valid thing.

[philipat]

Kind of off topic

The Mods do a remarkable job of staying on top of such a fast moving thread. I wonder if it might be desirable/feasible to delete posts without automatically re-numbering the thread? This would then make references to earlier post numbers in message text easier to locate (And would also provide a measure to the reader of how much cr*p gets deleted!)

[takata]

Occams Razor

ttcse:
I wasn't arguing against power failure, but against the notion that it would offer an even simpler Occams Razor. To make my point I was fishing for the 'What happens then?"

I know you didn't but also most people don't really understand what is the real meaning of this overused reference to Occams Razor "law". In short, it means that the simpliest explanation which agree with all the facts is certainly the good one. The difficulty behind this theory is much more about understanding all the facts than putting them together into the simpliest explanation possible.

For most people here, the simpliest interpretation of the end at 0214 of ACARS is to say: obviously, no more ACARS = aircraft crashed. This is a fact and lets build a theory starting with this postulat without even verifying if another possibility exist. But this is not the only possibility as no more engine imply no more ACARS also.

Another example was the severe/extreme turbulence theory which postulate: proximity with an active CB = severe turbulences. Then, because this aircraft was experiencing severe turbulence at 0210, it could have been only worst later, consequently it crashed at 0214. Until one say: "hey, at 0210, the autothrust setting doesn't fit with severe turbulence from Air France SOP and, by the way, the initial report of turbulence has never been confirmed by the BEA. Why do you always mention turbulences when the only factor confirmed about the weather is ice?"

And so on.
If the criminal investigators used this Occams Razor principle the same way to arrest people, they would always arrest inocent people which looks like criminals rather than criminals which looks like inocent people. As I said, the main difficulty is to find what are the real facts and to cross-check everything before building "simple" hypothesis based on unverified postulats. Others will start from their own conclusion and then will try to bend any relevant fact around, no matter if others doesn't fit at all.

To cut to the chase, they came down west of course but near the zone of eastward current flows. Then they drifted eastward/east-north-eastward before then currents took debris more north-north-eastward.

An approximate distance crash and zone with a map will explain it much better.
If you based this claim on OSCAR, let me tell you that this is a good tool for an average zonal current but in no way it should be considered as an ultimate reference for such a study as it is based on sattelite data extrapolation. See the validation method here compared to drifting buoys and the limitation in the granularity of the dataset (about 20 days average mean with filtering):

Method:
In these comparisons, as in the rest of the validation study, the drifting buoy data are minimally processed. The principle of this validation study is to compare OSCAR to the drifter data at the locations of the buoys.

For each individual buoy, the drifter data (locations and derived velocity vectors) are initially sampled every 6 hours, and are 20-day low-pass filtered along the drifter trajectory. Hence this validation study only concerns surface current variations on time-scales on the order of, or larger than, 20 days. This is consistent with the OSCAR surface current processing, which, by construction, is relative to periods larger than 10-20 days. The filtered drifter data are then sub-sampled onto the ~5 day OSCAR time base. For a set of drifter vectors defined by locations, times and velocity components, the collocated OSCAR vectors are obtained through linear interpolation at the same locations. The two collocated datasets can then be compared.

Moreover, the actual drift is not a simple function of surface currents. Drift may be approximated by buoys but it is related to the drifting object (the reason why, after several days, a wreck field will be 10 time larger than it was the first day, because objects are drifted differently. The main components of a drift vector are the near surface currents and the surface winds (here waves height may be neglicted) which are function of the object.
I did something much simplier than that. I reconstructed the real drift from 6-10 June (0.37 m/s, hdg 357 deg). then, I compared it to different currents maps (including OSCAR) to check versus the zonal currents and for extracting the wind factor during that period. The main purpose was to get an approximation of the distance covered during 1-5 June which is in the range between 90-115 NM. I'm still working on the wind part to have a better estimation of the distance and direction. But between 1-10 June, all the objects have to come closer and end at roughly the same place. From your estimation, I just can't understand how it is possible.

S~
Olivier

[ArthurBorges]

To Narval

If Mr Gourgeon...seems to be very happy with his company (some people never learn) ... Mr Spinetta...

Narval, when Mr. Spinetta was appointed to head Air France over a decade ago, nobody there or at Air Inter, the domestic carrier, expected him to last very long. The job was a political plum that came on plush one- or two-year ejection seat and he turned it into a real job. From your pseudonym, I infer you already know this but it might be news to at least some readers here.

Reports of his speech at the AF/KLM annual general meeting reflect his ability to communicate sincere sober emotion along with the frank statement "We were caught out on safety."

On Mr. Gourgeon, the same interview which has been served up here in bits and pieces contains another morsel:

Pourquoi ne pas être allé immédiatement au Brésil ?
Je ne suis pas allé immédiatement au Brésil pour pouvoir assister à la cérémonie en hommage aux victimes à Notre-Dame de Paris. Le lendemain, j'étais à Roissy pour le rassemblement en mémoire des douze membres d'équipage et de quatre autres salariés disparus dans le vol AF 447. C'était très émouvant, il y avait 10 000 personnes. Le silence était total. Le moment très difficile. Nous avions la mission d'amener nos passagers à bon port et nous ne l'avons pas fait. Chacun pense dans la compagnie : «Air France, ce n'est pas ça». La charge émotionnelle est immense. Quant au Brésil, j'en reviens. Là-bas, le 30ème jour après la disparition d'un proche est un jour de deuil très important. J'y étais. Les familles que j'ai rencontrées en étaient très touchées.

This obtains:

Why didn't you go out to Brazil right away?
I didn't go right out to Brazil so I could attend the memorial service for the victims at Notre Dame Cathedral. I was at Roissy (Airport) for a ceremony to commemorate the 12 crew members and four other AF staff lost on AF447. It was very moving; there were 10,000 people. There was total silence. It was a difficult moment. Our mission was to deliver passengers to destination safely and we didn't do that. Everyone in the company was thinking: 'This is not what Air France is about.' The atmosphere was heavy. Getting back to Brazil, there, the 30th day after someone passes away is a very important day of mourning. I was there. It was very touching for the families I met.

Looking at the whole interview, I'd qualify Mr. Gourgeon's attitude as defensively shaken, responding with a barrage of detail to project an attitude of transparency. Another question was about the rumor of a spate of cabin crew resignations, which he took as an insult to the commitment of AF staff, adding that "not one" had opted out. A preceding question asked about delays in contacting victims' families. He just said nobody asks you for your next of kin anymore when you buy a ticket, you just indicate your mobile phone number: "But obviously the mobile phones were with the passengers."

On a last detail, Spinetta chairs the AF/KLM board; Gourgeon is AF/KLM CEO.

[vapilot2004]

The Mods do a remarkable job of staying on top of such a fast moving thread. I wonder if it might be desirable/feasible to delete posts without automatically re-numbering the thread? This would then make references to earlier post numbers in message text easier to locate (And would also provide a measure to the reader of how much cr*p gets deleted!)

Philipat, the numbering scheme may be a limitation or setting of the board software. There is a permalink at the upper right hand corner of each post that lives up to it's name and will always connect to a post still up.

I agree with you. Sometimes posts are more than a page or two away from their original location due to all of the bad that requires Mod Pruning.