barry54

Hi folks, first post. I've been following PPRuNe for a couple of years now, Spanair , Colgan 3407, and AF447. Not a pilot but I am an interested frequent flyer. Thank you to all pilots and aviation people for the many contributions to these forums. Better training for pilots and having a clearly visible AOA indicator do seem like common sense improvements.
I do have an idea, a plan maybe, that I have not heard anyone discuss. It is this ---- Provide a shield system for the pitot tubes in potential icing conditions----. How do you do that ? The speed sensors on modern big jets are state of the art . If a more advanced design were available it would be used. Someday scientists may create a new material called unobtanium . It will be immune to ice buildup at all speeds and conditions. That would be great but that day is not here yet. In the meantime what planes need is an Apollo 13 type of a solution.

One way this could be done is to design small aerodynamic bullet shaped shields for all 3 pitot tubes. The two part shields would normally be left in an open retracted position along the sides of the pitot tubes. When a plane is about to enter bad weather one or maybe two of the shields would be moved forward and together protecting the speed sensors from ice buildup. There are a lot of different ways a shield system could be engineered. When flying through an area like the ITCZ one shield could be left open and one shield closed , that pitot tube being held in reserve , and one shield could cycle open for a few seconds, just long enough to get an accurate reading and then closed for maybe 7 seconds It would be better to get intermittent but accurate readings then none at all. You could also add a smaller diameter 4th pitot tube with no shield. It would be designed to fail in half the normal time in icing conditions and would serve as an early warning system for the pilots and computers that all speed indications and the AP may soon be lost.

Sometimes bad weather can't be avoided. If airlines are going to fly people across the ocean at night in stormy weather then the speed sensors need better protection from icing , pilots should get an AOA display like the military guys have, and some kind of real practice in hand flying at high altitude in difficult circumstances. I am waiting to see what the BEA will recommend.

henra

I find the report quite telling. Working in a bigger company myself I háve the feeling I don't need too much reading between the lines to get an idea of the general 'culture'.

Surely it does not explain the AF447 accident in any direct way.
On the other hand it conveys a slight sense of a relatively strong hierarchy within an 'elitist' group and certain mistrust in other's abilities.
Statements coming from Pilots about Captains being 'immune' to their line managers implicitely point to a rather steep hierarchy in the cockpit.
In general such a spirit does not help a good workload sharing in case of high workload/stress.
Also mentioned is a slight tendency to 'Not go by the books', cut corners, etc.
All this doesn't explain any of the direct actions of the PF of AF447.
It might however be related indirectly to a potentially sub- optimal workshare during the crisis and maybe even reluctance to openly express any discomfort with the actions of the colleague or different impressions of the actual situation. This would especially apply if the PF was the more experienced of the two F/O's.
Knowing who actually was PF could indicate if this might have been a factor or rather not.

Edit: As a hint that CRM seems to be a very important topic in this report: They have added Sub- Items to go more into the detail for that and used up most of the alphabet in doing so. The only other items where they also chose to go a bit into details was Instruction/Training and Flight Data Montitoring, albeit to a lesser extent.
Take that for what it's worth....

Lonewolf_50

barry54, that is an Interesting idea.

If you look at the discussions in the three earlier AF 447 threads, and posts that address the improved Goodrich pitot tubes, the "build a better pitot" approach seems to have in part succeeded. That this hull had not yet gotten the new probes is an unfortunate factor in this crash. The new design looks to have been begun after some years of the industry reporting issues with various pitot tube malfunctions.

From the FAA airworthiness E9-21368 of 02 September 2009:
Federal Aviation Administration, 14 CFR Part 39, [Docket No. FAA-2009-0781; Directorate Identifier 2009-NM-111-AD; Amendment 39-16004; AD 2009-18-08] RIN 2120-AA64// Airworthiness Directives; Airbus Model A330-200 and -300 Series Airplanes, Model A340-200 and -300 Series Airplanes, and Model A340-541 and -642 Airplanes
A shortcoming I see in the "shielding" solution is that if you get it slightly wrong, you interfere with airflow around the pitot tube and thus interfere with its normal function. (A fix worse than the problem makes the engineer a sad fella ... as he has to do it again and now gets funny looks due to the budget overrun ... )
Note also, from the same document:

jcjeant

Hi,

But ..
So .. this system worked for Apollo 13 but can't for Airbus ?
If I understand well .. the "Apollo 13" pitot type .. is NOT working (not in the measures loop) when shielded .. so it's not interference at all as the shielded pitot tube is neutralized.... IMHO

barry54

Lonewolf 50 -- The Goodrich pitot tubes are somewhat of an improvement, but they can still fail like the Thales tubes. Any shielding device would clearly need to be several inches behind the pitot tube opening when in the retracted position.
You are right about possible problems. If a shield system was not well engineered and thought out it could cause problems of its own. It would need to be as simple and foolproof as possible.
jcjeant -- I was not referring to the Apollo 13 pitot tubes . There was a problem during the mission with the ships air supply .The carbon dioxide scrubbers were inoperable because of loss of power in the main module .The astronauts could not wait for a perfect solution .They needed an immediate fix. It it involved duct tape and cardboard, but it worked

xcitation

@Takata

Thanks for the superb graphic. Do you recall the posts that AF did not install all of the available backup steam guages option on this a/c?

The report does not explicitly state attitude failure. However it does not state that attitude indactions were all displayed faithfully. FDR did not record speed on right PFD, so maybe not the attitude.

The BEA report does show errors on both ADIRU and ISIS. Having read some other incidents I recall that Airbus can give bogus stall warnings and PFD degradation under certain ADIRU failures (those 2 Airbus incidents near the Australian radar station?). IMHO the scant BEA reports do not eliminate the possibility of a degraded attitude indicator on the right PFD.

The list of 7 reasons are purely speculative. Probably a better title is "possible contributing factors to a general nose up/stick back". I was struggling with imagining the crew spending nearly 4 minutes looking at the +15 deg pitch and giving it generally stick back - sometimes to the stops. Which scenario(s) would you envisage to explain this?

I am trying to get a list of all factors either likely or not. Add any new ones that are missing. Identify any that are false (e.g. attitude indicator).

PJ2

xcitation;

There is no evidence either in the ACARS series or any of the BEA Reports and Update that failure(s) of the attitude display(s) occurred.

Initially, (before the pitch-up), this was a simple loss of airspeed data, from which all messages from 02:10:05 until the apogee of the climb may ultimately be traced. In other words, there are no ACARS messages which indicate failure of the IRUs, which supply all attitude information, before the pitch-up.

The ADIRU to which takata's schematic shows is a dual "ADR"/"IRU" installation. As shown in the schematics below, the ADR part may be separated and otherwise shut off separately from its IRU, (as per standard fault actions set out in the FCOM or on the ECAM).

At the time you claim the "possible" effect (of loss of attitude information) to have occurred (which "causes" the pitch-up), the 3 IRUs remained unaffected by the pitot failure/airspeed loss and would have continued to display correct attitude information after the pitch-up. A change in altitude information would require that the Static ports were affected which would in turn affect both the ADR and the IRU incoming data. By all indications, this did not occur.

Therefore the "original cause" that is claimed to have caused the PF to pitch the aircraft up, is not present.

By the time the IR1 & IR2 FLR messages appear in the ACARS series, the aircraft is seriously stalled and descending at > 10,000fpm.

In different words, if IRU #1 or IRU #2 or both were functioning (we have no messages before the pitch-up that they weren't), and DMC #1 or DMC #2 or both, were functioning (again, no messages, no comments in the BEA Reports), then attitude information would be displayed normally, very likely on all three indicators.

From this, there is no basis for believing or even positing that any attitude indicator failed, causing the PF to pitch the aircraft up.

Merely saying that it is possible has no relation to whether it actually occurred or not, but the statements above show that normal attitude indications were very likely available.

I'm not dismissing your ideas outright here. I am demonstrating the way an investigative process might function.

A claim that any or all attitude indications were lost requires that the above statements be refuted.


PJ2

Here are the schematics:


ADIRS Control Panel:





ADIRS Schematic - information flow:




For comparison to the ADIRS schematic - the Pitot-Static ADM/ADIRU system. Pitot-static data is fed to the ADIRS for use by both the IRU (baro altitude, and rate of baro altitude change) and the ADR:

airtren

What's this supposed to mean?

Based on the BEA text, the A/THR was OFF at the time of the climb.

The BEA text does not mention manual power/throttle changes during the climb time interval from FL350 to FL375. So, that makes the climb unintentional?

A kinetic versus potential energy conservation calculation shows that the BEA indicated height delta (2500 ft from FL350 tgo FL375) checks against the BEA indicated delta speed of 60 knots (from 275 to 215 knots). Same is true for FL380, and 185 knots.

dEk = Ek (275knots) - Ek(215knots) = dEp (2500ft)

That's clear. Thanks for the detailed explanation.

But the connecting to Dakar failed, and the contact with the previous ATC was no longer active, so there was no online ATC control/contact at the time.

A deviation of 12 degrees to the left started already prior to to the A/P disconnect, without ATC approval, which would seem to indicate that the PF (and PNF) were already in ATC bypass mode when the climb started...

PJ2

Hello airtren;
I'm not clear on what is meant by "ATC bypass mode". Can you help me out? - I've never heard of it.

Just to review, from the First BEA Report, (all reports found here), the crew had established contact with ATLANTICO on HF and had acknowledged a successful receipt of their SELCAL by the ATLANTICO aeradio operator, (ATLANTICO and all radio stations contacted on HF are not ATC controllers, remember - they relay messages from overseas flights to ATC controllers who then provide clearances/instructions to be relayed back to the flight. This often takes a lot of time so one has to anticipate needs early).

A deviation off course requires an ATC clearance. As I've stated, an aircraft cannot just be taken off course without such clearance unless an emergency diversion is required, (I've done it once, on the Pacific - one lights up the aircraft, broadcasts intentions on 123.45 and if necessary on 121.5 and one might climb a few hundred feet 'just in case'). The CPDLC is a direct link to ATC but it had not successully logged on. (I have my theories on why) but communications were established and a clearance to deviate or climb is required unless, as described, an emergency exists.

airtren

Perhaps I am posting under a misunderstanding on my part: the PF & PNF had no radio communications from the time when the event started, all the way to FL 0.

They deviated to the left at FL 350, and climbed from FL350 to FL375 and then FL380 without any radio communications and thus ATC permission/approval - if I understand correctly.

JD-EE

In most situations roughly analogous to AF447 the crews have visual clues of one sort or another. That means the plane leaves it to the pilots to discern "too fast" from "too slow" from those clues.

Um, that brings up a probably silly question. Might the pilots be better able to guess AoA from simply turning on the landing lights or watching wing tip lights to see if they can get any hints from the variations in the clouds through which they are flying?

There were two of them who apparently learned it wrong.

--

I love PJ2's pages from Davies. Even I understood most of it. And it explains a lot. Thanks.

Now I have a question:
The plane turned to the right despite NU-L inputs on the stick. Did we ever hash this out? (If so, excuse me please. I somehow missed it.) Did the pilot fixate on the plane not being able to turn back to a straight course and forgot about stall? And why did the plane apparently never respond to the left part of the inputs?

Rolling the plane to the left with up elevator is, I believe, a left turn input. If I'm right why'd the plane go into a fairly tight right turn?

bearfoil, you should have mentioned aside from one set of alternating roll inputs all inputs were either neutral roll or left roll to make your point better. I don't suggest a reason. I simply ask "why?" Then you launched into what seems to me to be silliness.

PJ2 - in message 515 of http://www.pprune.org/tech-log/45687...ml#post6582663 you suggest you're out of ideas. Does the notion of turning right with left stick inputs key any ideas?

bearfoil

There are several sources of data to support these last ponderings. As above, PNF would alert by radio as to the deviation, and also the climb, and if it was inadvertent, it may have been accompanied by a M'aider.

Simply because communication is problematic, the crew would follow procedure and make the calls. The CVR, depending on ambient sound level, will support this, so I am not worried.........

The inference that could be made if notification to local traffic of a climb was not made, would mean that the PF's original input will not have been intended as a 'climb', per se, but a recapture of FL350 if low.....

For about two years, various protestations of a/c fidelity and reliability have been made by quoting design considerations.

When it comes to actual BEA data "PF input RLNU.....at PITCH +10 the a/c began to climb", the immediate response is the PF commanded an absurd ascent which led to the loss of a/c. Various adjectives are selected out of thin air to savage his training, and hold in suspicion his competence.

I find that outrageous. Is there another way to see these slanders here?


JD-EE I have no opinion on what you consider silly. If you think I made a compelling statement, acknowledge and leave the gossip in the computer. I have been hammering on the initial ten seconds for months, actually from the beginning...........

What part of airframe/EFCS failure do you seem reluctant to address? A fall off of the a/c to the right is noted by BEA, as you state. It happens again, and BEA states same. Rolls and reversals are stated on the way up the climb. The a/c is behaving unusually, notwithstanding the absurd climb.

Rudder for Roll. Rudder issues. If Left Rudder was unavailable, any out of trim excursion to the right would be un reversed, and additive. Likewise if Right Rudder was in, and not corrected (or correctable).

mm43

airtrenThey were on UN873, an unidirectional northbound airway, and even though there is comment in the BEA Note that they hadn't managed to logon to DAKAR via ADS-CPDL, they were still (12 mins) inside the ATLANTICO FIR. There are currently no known records of them attempting to call either ATLANTICO (which had their guard) or DAKAR on HF.

There were a number of "off track" deviations made that night by other aircraft passing through the ITCZ, and it is not clear how many requested clearance, or how many just deviated.

Adjacent parallel airways are spaced at 90 NM.

DozyWannabe

Bear, I honestly don't see that being either implied or said outright anywhere - in fact the only person to say words to that effect throughout the running threads has been you!

Even elite pilots make mistakes - sometimes fatal ones. *If* the PF in this case made a mistake, then there may have been extenuating circumstances. *If* his training was part of the cause through being substandard, then it must be corrected. This is not about blame or slander, it's about trying to make sure that it doesn't happen again.

bearfoil

I agree since you state such, that you do not see an imbalance of invective. Much of it has become tacit, part of the Thread Line Culture.

Part urban myth, at this point. The time just before a/p unlatch is not well addressed by BEA. Therefore Not mentioned=Cannot have happened? The part that is disclosed is parsed such that we look for faults in the Pilotage, and accept the a/c "AS DESIGNED". "The aircraft won't do that". "What was the pilot Up to".

Doze: Do "Elite Aircraft" make mistakes?

PJ2

Hi airtren;
I'm just going by your post #519 where you say:

Perhaps we're talking past one another!

For the record, it is stated in the BEA Update that the pitch-up was the result of an aft movement of a sidestick.

To your point, all I'm saying is, you can't just "decide" to climb, or descend to a new altitude or deviate off course for weather or anything else, without an ATC clearance unless there is an emergency.

They didn't get a clearance to climb but just pitched the aircraft upwards.

Therefore we have to conclude that the PFs intention, unchallenged by the PNF, was not to just climb to a higher altitude and level off.

Therefore his reasons for the pitch up lie elsewhere and that is what we need to find out.

And to drive a point home which I have been stating for some time now, when one loses the airspeed indications, one does NOT change pitch or power. As soon as one does that, one loses the pitch and power settings in which the aircraft was stable immediately prior to the loss of airspeed indications and very quickly loses situational awareness.

Without careful attention to attitude and power, loss of control can quickly result. A pitch-up of 15degrees in a transport aircraft operating at FL350, if held and not reduced, is, for all intents and purposes, a loss of control.

Whether the PNF knew about and understood what the PF was doing is not known and not discussed in the BEA Update. We will know, I hope, in the upcoming Report.

Does this help?

A33Zab

All A330 Flight Controls servos are individually monitored, including spoiler servo jacks.
If one 'refuse' to act as commanded the FCPC/FCSC will take corrective
actions to prevent such an asymmetry; e.g. remove symmetric hydr.
pressure to aerodynamically stow the spoilers and at the same time
locking the extend movement (excl. 4 & 6 for roll support) to prevent
floating spoilers.

This will be notified to crew by an appropriate ECAM message:
"F/CTL SPLR FAULT"
or in case of aileron fault:
"F/CTL L (R) OUTR (INR) AIL FAULT"

Both faults were NOT present at the time.

I did mention it before, could MLA be any factor?

Manoevres Load Allevation.

The purpose of MLA is to distribute the lift over the wing to relieve
structural load on the outer wing surfaces (bending moment).
The demanded load factor is maintained.
MLA utilises spoilers 4,5 & 6 and the ailerons.
The MLA becomes active when the side stick is pulled more than 8°,
and the load factor is more than 2g, in which case:
- The ailerons are deflected symmetrically upwards:
Max 11° added to Roll demand, if any.
-Spoiler 4,5 & 6 are symmetrically deflected:
Max 9° added to Roll demand, if any.
- Deflection is proportional to load factor in excess of 2g.

An elevator demand is simultaneously applied to compensate for the
pitching moment induced by spoilers and ailerons.

The load allevation is only available:
CAS > 250 Kts
FLAP LVR = 0 position
In NORMAL or ALTERNATE LAW.

MLA has priority over the speedbrakes.

DozyWannabe

I disagree - this is and has always been a technical discussion regarding systems design and the man/machine interface, any attempt to deviate from that has generally been avoided.

Well, there's two likely reasons for that:

1) Everything appeared relatively normal up until that point, so mentioning events during that time period were unnecessary
2) They didn't know exactly what was going on during that time period, so could not put anything conclusive in the note

Not at all - that is how you seem to be choosing to interpret it. Coming at it with no preconceptions it reads as a very dry statement of events that happened at specific points in time without any implication of anything, as you would expect of what was essentially a press release to get the more lurid speculation out of the way so they could do their job.

With all due respect, you seem to choose to be very definite on occasion and then swing wildly into evasiveness and conjecture if anyone asks anything of you directly, so I'm going to ask you directly and see what - or if - you answer:

Do you believe in a conspiracy on the part of the BEA and Airbus to blame the pilots in the case of AF447?

Either way I think you're too emotionally involved with this, and I suspect a few days away from the thread might do you some good and get you some perspective - it certainly did me good a few weeks back!

bearfoil

A33Zab

Thanks. If aerodynamically stowed, and absent hydraulic pressure, can they flutter?

Also, RTLU. Is Rudder centered before limit is ennabled? Any chance of (R) Rudder staying with the a/c all the way, without annunciated ECAM or ACARS rpt? If jammed at (Right) would there be any ECAM, since the Rudder has (limited) authority anyway? Would the deflection be accepted as an input, not an anomaly?

Hi Dozy. Conspiracy is defined as an effort to collude with others to effect an outcome, generally for ulterior motives. A "Surprise Birthday Party" is also a conspiracy, a good one! Few people have left an ability to accept that sometimes motives are unconscious, or inadvertent. That means we can say bias instead of conspiracy. With bias, it is even more difficult to pin down, since it frequently lacks acknowledgmwent among 'buds'.

The silence of the lost pilots is a sort of vacuum. Nature abhors a vacuum, and the firstest with the mostest will fill it up. Yes, there is an element of 'blame the crew'. I think it does not rise to the definition of criminal, but merely to an excusable 'birds of a feather' sort of interdisciplinary sort of thing. I take your recommendation seriously, I have once again become too passionate, and will take your advice.

DozyWannabe

*le sigh*...

Or the CVR that is now in the hands of the BEA - the "lost" pilots are silent no more!

I'd like to see precise examples of where you're getting this idea from.

See you when you come back!