Sorry to say chaps, but such a faith in our ability to totally comprehend the atmosphere and design solution to every problem it throws at us is excessively optimistic.

Icing conditions that AF447 encountered a couple of minutes before its destruction are very rare and seldom cause trouble. Being elusive, we can't analyze them as much as we would want. All we know about them that they usually occur near "deep convections" and unlike every other known type of icing, they need the surface they hit to be heated or they won't stick, so they clog heated probes and engine compressors. I am exaggerating here but if we approach the problem narrow-mindedly, there is obvious design solution: cut heat to probes as they pick-up ice. That would make probes resilient to exotic high level ice crystals but would make them more vulnerable to every other icing condition there is and that's a lot of problems created for just one particular solved. In real world, there are no designs without downsides and designing solution for problem one doesn't even know what it is can be successful by pure chance.

Order for replacement of Thales probes was based just on empirical evidence (which is more than enough when dealing with high hazard activities such as aviation) that they are ten times more susceptible to getting blocked by high level ice crystals than rival ones. We have no idea what is in the design of probe or complete installation that makes Goodrich probe ten times less likely to get blocked than Thales but we can be sure that different performance was completely unintended. Also Goodrich probes are not completely immune and we better make sure the pilots know what to do when they lose airspeed information, like keep on flying regardless.

Probably not. 30something crews got it right in the real life even without resorting to prescribed procedures. Some of them were AF.

It is remarkable that aviation culture will wax hysterical only to be followed by a rapid version of "Nothing here, let's move along."

When a passenger packed 777 lands short of the runway with fuel starved engines (starved, not exhausted), we end up with an urban myth of "mysterious fuel characteristics", and a band aid sheet metal workaround.

Here, it seems, we are on the cusp of similar dismissal of anomalous circumstances.

Whistling in the cemetery? "We are going to crash...It cannot be." "Odd Pitots" and crap pilots? Keep whistling.

"It cannot be..."

Oh, yes it can. Merely unusual Pitot behaviour? Definitely. If one believes in real red herrings, and not virtual.

paull;

Re, "1> I know that prior to AF447 the pitot-icing had already been identified and there were steps in place to change the sensors, but was any action taken to introduce this scenario into sim. flights so that we could assess what proportion of pilots would get it right?"

Yes, such training was planned and done as part of AFs recurrent briefings and simulator training curriculae for 2008/2009.

The BEA Second Interim Report, Section 1.17 Information on Organisations and Management, provides specific information on this and the previous section, 1.16 Tests and Research discusses the thirteen previous UAS events which had sufficient data such that adequate studies could be undertaken of the phenomenon.

Oddly, of the 36 UAS Events listed in Appendix 7 of this Report which occurred to the A330 up to the day of the accident and during the recurrent training period, only two were not in cruise* yet the UAS script was for the right-after-takeoff phase which of course is the most critical phase of flight.

The briefing notes for the exercise indicate that if the ADRs are not rejected and the FCPCs and FCSCs continue to use the incorrect information, the crew will have to:

" Trigger the emergency manoeuvre(15) if they consider control of the flight
is dangerously affected (initial climb, go-around, etc.);

" Trigger the Flight QRH procedure with UAS / ADR check if the trajectory
has been stabilised and flight is under control." **

*one event was on descent (FL290) and the other event was on approach.
** p54, BEA Second Interim Report, English version

PJ2

A quick question. You refer to "UAS" procedures in the literature from Airbus. I may be mistaken, but I believe the term "UAS" was coined here at PPRUNE, post AF447. Prior, the nomenclature was not specific, generally referencing other than "Unreliable Air Speed" (UAS). ??

Lyman;

The term "UAS" was used in the BEA Second Interim Report but not the first. In researching the eight or so AF447 threads the term was used first by Will Fraser on the 23rd of September 2009 and thence by bearfoil on the 24th of December, same year. Whether the BEA adopted the abbreviation or not can't be known. I know some from the BEA have monitored the threads in the past.

The evolution of the UAS drill and QRH checklist is discussed in the BEA Reports. I first saw it around 1998 and it showed up in the FCTM around 2003/05. At that time it was clarified when to apply the memorized items and when to not do the memorized items and go straight to the QRH procedures which prescribed setting pitch and power according to the QRH tables. Typically such circumstances were cruise flight; the memorized items were intended for low-level, during/after-takeoff circumstances where immediate information was required due to the immediate high risk to the flight, (as per Birgenair & Aeroperu accidents). The aircraft is not in immediate danger with a loss of airspeed information in cruise.

Unreliable airspeed has been called unreliable airspeed for many years. It is not a new term, neither is the abbreviation.

I recall checklists called "unreliable airspeed" from at least the early 90's, possibly the late 80's, and expanded checklists and explanatory material with the heading "speed control with unreliable airspeed".

The expanded checklists and explanatory material were more detailed in the A300 days than for later models, particularly concerning the various failure modes and effect on indications and systems.

PJ2, Hamburt Spinkleman.

Thank you both for your input. The reason I suggest that "UAS" is a "new" acronym has to do not with aviating, but with marketing. At the time, in 2009, it occurred to me that "Unreliable" is not a term to be found on any documents emanating from the community, manufacturer, Line, Pilot's, etc. Not only does it conjure up "risk", it could be conflated to mean, erm...."Unreiiable". It could also have a negative effect in "product liability".

So Hamburt, I am curious if you have copies of these "checklists from years ago".

The reason I coined the term was its self-explanatory nature, and its 'honesty' in an industry that can at times be over obsessed with "marketing".

I meant it without malice, but was attracted to its inception merely because it was....descriptive.

me

Especially at low level, PJ2, would it not also be important to include "Point" after Pitch/Power? Heading after T/O could be construed as important as the first two, CFIT?

A history of the Unreliable Airspeed drill and checklist can be found in the BEA Interim Reports.

PJ2, Lyman, Hambert Sprinkleman,

I did a little research on UAS and came up with some interesting results. It would appear the accident investigation and reporting agencies tend to use the term "Erroneous Airspeed Indications". This is true in the Aeroperu and Birgenair reports as well as one issued by the US NTSB on an incident over Kansas City involving a B-717 which did not crash. These reports were issued prior to or around the time of AF447.

However, there is an FAA document that was issued to discuss this phenomenon:

I would say, in terms of the English language, errorneous airspeed meaning is pretty clear, the airspeed was wrong. Unreliable airspeed is a little "softer" as its meaning is "don't count on it to be accurate or uncertain of accuracy".

Lyman, In a court of law, I suspect the terminology of either poses a problem for the defendant.

Unreliable Airspeed
 

As QRH section 2.21 is titled "UNRELIABLE SPEED INDICATION/ADR CHECK PROCEDURE", it looks like the "air" has been added at some point of time to the "speed".

But does it change something to the AF447 story (except adding almost one page to this thread:}) ?

From the BEA July 2011 report
 

Memory item the following procedures are to be applied without referring to paper: immediate actions of UNRELIABLE SPEED INDICATION/ADR CHECK PROC

"Air France a traduit le titre de cette procédure par « Vol avec IAS douteuse »."

Like llagonne66, I don't know what the point is, but.

"Unreliable Air Speed" is the term Airbus has been using for decades in Flight Manuals, FCOM's, QRH's and other material.

Boeing, by the way, use the same term. Or actually they use both "Airspeed Unreliable" and "Unreliable Airspeed".

Hamburt Spinkleman

The point is a small one. I got tired of writing out the various phrases used to identify a problem that was getting much attention. In creating (?) the acronym, UAS, I found a way to simplify and shorten the nomenclature. If it was used before, fine, but I believe the industry was calling the problem different names, and spelling them out, as if it was this ill-defined, and rare anomaly. I thought it deserved its own acronym. That BEA started to use it also meant I may have been correct in codifying it. It also meant that BEA might have been reading PPRuNe, which means PPRuNe has some standing in the discussion/investigation. So Hamburt, it means whatever it means to you, nothing more. At all.

If it is a matter of credit and accolades then it appears those should go to a Will Fraser. See PJ2's post #86.

But even he is some 30+ years too late.

Will Fraser c'est "Nom d'Aero". Nom de Plume.

The acronym, "UAS", not the text. Let's drop it.

Total pressure tube heater:
 

Heater is 2 elements of resistance wire, ~34 Ohm @ 20°C.
Uses full 115VACrms wave in flight and half wave (~81Vac) on ground.

PHC warnings levels <0.9A and >6A in flight; <0.4A and >4A on ground.

http://i474.photobucket.com/albums/r...Pt_tubeHtr.jpg

Similar idea posted here.

I think you will find that Claudette is a bot of some sort, roaming the forums reposting snippets of previous posts.

There have been a few of them recently.

KBPsen

Quite. "She" happened to land just ahead of the accident flight? At the age of 21?

Claudette:
More useful would be, "TILT". (And...you'd lose your quarter.):p

First of all, "PULL" is TAUGHT as a partial in approach to STALL. It is a version of "Maintain Altitude".

Secondly, on the PFD already was ALTERNATE LAW in AMBER. Piloting an a/c in the conditions 447 found herself takes more than second grade reading skills.

Failing more data, I will never understand the PF's actions. Never. Without context, and perhaps there is none, all that aft stick is bizarre. Equally mystifying is the utter lack of "conversation" on the flight deck. I do not believe it was so, there is more to hear. If wrong, I will apologize in time. The Human Factors group is stuck with the hot potato, No? To make conclusions, they will have to support them. In this support I think more shall be revealed.

PFD messages.
 

You confuse PFD with ECAM....
ECAM message is F/CTL ALTN LAW(PROT LOST)

The only messages you will see on PFD is RED "MAN PITCH TRIM ONLY"
or AMBER "USE MAN PITCH TRIM" or a FMGS related or EFIS switching message.

One mo' time
 

Good grief.

I thot most here agreed that the crew

- Pulled versus "neutral or even down command on the stick
- Had poor coordination
- Couldn't figure out they were stalled because the jet didn't shake, rattle and roll all that much

The pitot heat needs to be examined, for sure. But it is not going to be "found" as a primary "cause" by the investigators. The jet can fly real fine with only attitude and "manual" power settings while working out the UAS problem, right?

It's gonna be all pilot/crew error.

Lyman;
Re, "Without context, and perhaps there is none, all that aft stick is bizarre."

Without spending too much time reviewing old ground, I had some time ago posited the notion that the PF was executing one of the memory items in the UAS drill.

In the BEA Second Interim Report pgs 54 & 54, Section 1.17.2.4, it is noted that the UAS event was included in the 2008-2009 recurrent training season. The training included a booklet, the contents of which are briefly described in this section.

The training script for the recurrent simulator session in which the UAS event was included and which the PF First Officer had taken on February 2 2009, (the PM First Officer had received this training a couple of months earlier), required that the aircraft remain in Normal Law with no warnings triggered.

The UAS scenario used in this simulator session was a loss of airspeed information just after takeoff.

In such a case, the safe conduct of the flight is impacted and the memory items of the UAS drill are to be executed.

After ensuring that the autopilot, flight directors and autothrust are set to OFF and if below thrust reduction altitude (nominally either 1500ft AGL or more recently, 1000ft AGL), TOGA thrust is set and the initial "safe" pitch attitude is set to 15deg NU.

At slightly higher altitudes (thrust reduction altitude to FL100) the thrust levers are set to the "CLB" (climb) detent and pitch is set to 10deg NU. Above FL100 and if the safe conduct of the flight is still impacted, pitch is set to 5deg.

The last item in the memorized group states that once the aircraft is above circuit altitude or MSA, (Minimum Safe Altitude), the aircraft is leveled off for troubleshooting.

The FCTM (Flight Crew Training Manual) dated February 2007, though not the legal operational document (only the FCOM is) states that if the safe conduct of the flight is not impacted, the memory items are not to be executed and the crew must reference "part 2" of the QRH UAS checklist for correct pitch and power settings. In other words, present pitch and power settings are maintained while the QRH is brought out and the checklist read by the PM.

As pointed out by Owain Glyndwr, the failures resulting from the loss of airspeed information are graceful and do not present controllability problems. Loss of airspeed information does not require immediate action when in stable, level flight. So, what explains the immediate action?

I think the PF concluded that with the loss of airspeed information the safety of the flight was at immediate risk and was simply executing what he recalled from the earlier simulator session.

As the BEA Third Interim Report states, neither cockpit discipline nor SOPs were followed, and in combination with the effects of the immediate, strong pitch-up commands and subsequent aircraft response taking the aircraft out of stable, controlled flight, I think this prevented them from perceiving and then assessing what was happening.

The situation became immediately confusing as to why the airplane was behaving the way it did, (two momentary stall warnings, due strong pitch-up commands) and the returning airspeed information after about 28 seconds, and a likely clearing of ECAM faults as the airspeed data returned to normal, presented confusing signals to the crew. By this time the airplane was approaching the stall at the apogee of the climb.

Once the pitch-up occurred, a cognitive dissonance rapidly emerged from the differences between expected aircraft behaviour and what they were seeing/hearing. Putting it differently, the pitch-up was expected to be the "correct" response but it wasn't producing the expected stability. Normal perception of the stall warning was blocked because they were trying to resolve cognitive expectations with what was happening).

There seemed to be the expectation that the aircraft would respond at cruise altitude the way it would right after takeoff. This speaks to training and experience issues of manually flying transport aircraft at cruise altitudes and of knowledge and understanding of high altitude, high Mach number flight.

I'm speaking here as a pilot of course, but I think these are reasonable, if not at the moment theoretical, human factors.

gums;

Precisely.

Issues arising from this may be the loss of stall-warning below 60kts and some form of audible signalling when the trim is moving but I think the designers/engineers had a reasonable expectation that the aircraft would not be flown in these regimes by competent line crews, and that every contingency imaginable could not be designed against, not, at least, without prohibitive and difficult-to-justify cost.

I think the one very significant issue is however, the design of the UAS drill and QRH checklist which does not reflect with sufficient clarity what is stated in the FCTM. All we need do is review the PPRuNe contributions on this to see that the matter is indeed conflated even among those who fly the airplane. The fact that 36 other crews did not do what many here have claimed to be the "correct" response in cruise, (if above FL100, pitch up to 5deg) is evidence that there is a significant disconnect here. In fact I think even Airbus is confused about this.

Gratuitous Bravo!
 

PJ2 wrote, among many sensible things:
Bravissimo! Well-done. :D

Without in the least dissenting from what Gums and PJ2 have written, and accepting that the drills were not exactly clear, it is I think worth reminding ourselves that 5deg pitch/CLB is not, of itself, deadly.

At 02:10:50 or thereabouts the aircraft was at 6 deg pitch, 5 deg AoA, climbing gently. The bank had been substantially zero for 15 secs past. The airspeed was 216 kts; which is about 1.17Vs. No stall warning. of course.

Had they stopped at that point or gone to level flight we would not be having all these discussions.It was the second pull up to 17 deg pitch that mattered and which is, for me, completely incomprehensible.

Well, this is entirely speculative and unsubstantiated! It is only one of a number of reasonable explanations of why the PF almost-instantly pitched the aircraft up to an attitude that, in a transport aircraft at cruise altitude, I still find frightening, and like everyone else I'm just trying to explain/understand. Greatly-reduced damping effects of much thinner air in high altitude flight requires that one be very gentle with the airplane, just like Davies* says. It is no wonder that the stall warning blipped twice, as the AoA at which the stall warning occurs (in Alternate and Direct Laws only) is between 1 and 2 degrees higher than cruise flight AoA's.

Aircraft inertia will continue the physical trajectory of the mass and "fly through" any immediate response by the aircraft to increased lift resulting from an increase in pitch and thus the AoA increases in any such swift changes in pitch, here, to the point of triggering a couple of stall warnings.

Frankly, while such handling is an abuse of the airplane which exhibits a lack of understanding of high altitude flight it's no big deal because the airplane isn't stalled. But none of this explains the continued pull on the stick after the stall warning continued sounding. Any response should have been a full-forward stick to achieve about a 12 to 15 degree nose-down attitude to unstall the airplane...that results in a descent rate of around 15,000fpm.

When the speed begins to increase indicating exit from the stall, a very gentle pull on the stick is required to avoid another high AoA resulting from the higher positive g's during recovery and subsequent second entry into the stall.

This is, or should be, all pretty basic stuff, thus the questions surrounding the continued pull. I think the confusion surrounding the ceasing of the stall warning and then its reappearance is understandable but the extremely high descent rate and unwinding altimeter, (the (older) standby altimeter was going around the dial once every 2 seconds in the sim), should be sufficient to have indicated a fully-stalled aircraft but they only realized that at around FL100 or so.

It's all been thoroughly discussed in seven previous lengthy threads but sometimes a review is helpful.

Owain Glyndwr;

Thank you for this reminder...I fully agree with you that a 5deg pitch attitude (about 2.5deg above normal cruise pitch) is not itself a problem, (I had originally thought and had posted that it was but again learned and changed my views!).

The second pull to 17 was accomplished by a pilot who had done that before, and not experienced any fatal result. He was repeating what the a/c and the UAS drill taught and were teaching. Pull, TOGA. It's in the book, and not contra indicated. Remember the "second Pull to 17" happened just after the STALL WARN blared, as it had before. Duff speeds? Is it Real?

Because it defeated a stable attitude is IRRELEVANT. You and I know that, but PF's screen was not recorded, remember? Did he purposely leave a stable attitude, only to STALL? I suggest, not on purpose.

A33Zab. Thank you, yes the ECAM. What is more important to a flying pilot, his instruments? His duty to the ECAM and fault resolution? An interesting discussion? Where should the cue appear? On ECAM, along with dozens of other bitches, or on the life and death of what he knows the a/c is doing?

In a very real way, there is a built in conflict on this flightdeck, if only "philosophical". Get out the book, address each warning, alert, team the return to unannunciated complaints? Or handle the a/c? There was little CRM, if any, shall the PF do it all, or must he choose? PFD? The fact that at any given time in this four minutes a plan and some teaming would have helped, the confusion was inconquerable, each opportunity ignored, misunderstood, or wasted.

imo

Lyman;

As has been reviewed on numerous occasions, the SOPs to deal with an abnormality come under the heading (as you probably know) of aviate, navigate, communicate then deal with the abnormality or emergency. Control of the aircraft is first assured, then ensure terrain clearances, then communicate internally (announcing the emergency/abnormality and calling for the drill or checklist, then when all is done, communicate with ATC then the flight attendants/passengers.

Except for the rejected takeoff and TCAS/GPWS/Stall warnings, no emergency and certainly no abnormality (which this was), requires immediate action and never without crew coordination.

In the RTO for example, crew coordination may be just the announcement "Reject!", followed by the simultaneous closing of thrust levers, selection of reverse, monitoring/applying full brakes and ensuring ground spoilers are up. At a minimum, announcing the drill or checklist ensures awareness and subsequent coordination and appropriate action.

In abnormalities, on the Airbus the PM calls the ECAM name of the abnormality, the PF calls for "ECAM Actions" and the PM reads (silently) and carries out the ECAM checklist items, confirming any non-reversible actions with the other crew member before actioning them. Then the STATUS page is then checked for system/aircraft unserviceablilities and possible performance considerations.

Where/when required either by the ECAM or company SOPs, the QRH is then brought out for further responses to the abnormal. The entire process including high levels of CRM is heavily emphasized in all initial and recurrent simulator sessions, check flights and in the aircraft manuals.

It is against all this that the response of both the PF and PM is astonishing and incomprehensible even given cockpit gradient and lack of command assignment and consequent assessment and decision-making issues. It may be unfair to observe (because I wasn't there, not because it isn't reasonable to expect), but upon initial pitch-up without announcement from the PF which is non-standard and highly irregular behaviour I would have expected the PM to announce, "I have control", level the airplane, call for the UAS checklist and sort it out with the captain later, on the ground.

Hi PJ2. Again I must salute your patience and wisdom. I do think, though, that in reiterating the proper procedures we can be distracted by just how emergent was the crisis, helped not in the least by the apparent thick skulled behaviour of the pilots, which we both acknowledge. We have the same target, but I think we approach the understanding from different perspectives. Your words describe "best case, best practice." Here, clearly, such was not the case, not with the information we have to hand. I cannot know why the PF started to handle the a/c so quickly, but he did, and right or wrong, I think he had a motive, one he thought was sound. That he should not do that is patent, as we know, again your patience trumps my volatility. I do not have the experience you have, so it may be easier for me to achieve confusion than even our PF. My sense of the cockpit is that there were two opinions on board, that of the PF, whose behaviour we know was incorrect, and the PM, whose only mistake appears to have been that he allowed the PF's opinion to prevail.

Throughout this three year long thread the evolution appears to be that the pilots were not up to it, and the a/c was wasted due to precipitous and unfounded actions. It has been far too long since anyone has written what these three pilots were up against. We cannot be at all sure that the conditions extant at the twenty seconds before and after were conducive to standard action, imo.

grateful to you

Interim Report #2 para. 1.6.11.5 (page 47 in the English edition) shows the changes on the PFD between Normal and Alternate law.

With all due respect,
 

Lyman:
:eek:
I beg to differ. As was just recently posted (and many times before), that procedure is strictly reserved for UAS near to the ground and for windshear (near the ground). It is patently wrong at high-altitude cruise. That's why it's now being trained, I do believe.

The only practice that would have saved them. :*

as
 

"That's why it's now being trained....." qed.

Point being, it was the second time he made that "mistake". Was there a Windshear warning? W/S appeared on ACARS, as did TCAS. Monkey hear, monkey do.... Besides, Pitch is filtered through the computer in AL2, none of his aft stick had as robust a result as had his ROLL. This in itself could have been the confusion's source, one of. Twitchy in Roll, molasses in PITCH. Acute Roll inputs, and chronic in Pitch? Was he treating them the same, respect wise? If he had forgotten and expected both axes to respond the same, his overcontrol in Roll could have incested his PITCH commands? Vice versa?

@HazelNuts39:
 

Absolutely agree! Thanks for the addition:ok:

Lyman, Re: UAS

Perhaps this will help refresh your memory, a 2006 Airbus presentation that contains some readable data.

http://www.iag-inc.com/premium/Airbu...ableSpeeds.pdf

Hey TD

Thanks for the link. Despite the cartoon format, I see a serious side, and my attempt to shorten and more specifically define the challenge remains. "Speed" is not as descriptive as "Airspeed", and at each opportunity to quantify the problem the manufacturer continued to use full script. I think the acronym is an apt one. As PJ2 points out, BEA picked it up. Whether concurrent or original, the acronym remains descriptive and has become part of the dialogue, I think in a good way. No?

:E:E:E lawyer, asked to define the difference between "unlawful" and "illegal" by the judge in a criminal court case could not define the difference. "Ah", said the judge, "I have always thought that the one was "against the law" and the other was a sick bird!"

Lacking information
 

Hi,

We are with limited factual information on the case. There are some reasons, we understand.

Trying to understand the "non understandable" reasons of PF (persistent NU) and in a simulation exercise (in doing so, trying to cover
ALL POSSIBLE reasons) and "last but not least" considering that RHS was not recorded (one of "inputs" to PF) i would ask: What kind of information he could see to justify his actions?

The SS sitck to the extreme left was coherent. (bank angle indications were reliable). What kind of possible misleading information could be coherent with (some) PF actions with respect to pitch?

RR NDB

Reselected PFD's to acquire the BIRD, coupled to reselected autoflight, which drank in 90 knots and showed profound NoseDown? There was a great deal of mystery and yellow tarps on the avionics bay (recovered).

V/sel: 5000 fpm DOWN, etc. The fact remains PF should have relinquished control. The fact also remains he did not. What was he looking at, surely not all blue? Mostly Brown? Why would he ignore PM: "TJIS< here here, we are climbing, go down." He Did Not descend. He saw all blue and continued to pull?

What did Air Caraibes show on the selected PFD during its uncommanded climb?

Down? Nothing?

Lyman,

In my post to you, I was not referring to the acronym relative to speed. The message I was conveying was the fact that in 2006, it was documented exactly what to do if one encountered unreliable or erronous air speeds at high altitudes. The correct procedure was available then. It is what PJ2, gums and others have been saying. 2006 is prior to AF447.