Tech Log The very best in practical technical discussion on the web

AF447

Old 29th Jun 2009, 10:00
  #2481 (permalink)  
 
Join Date: Jun 2009
Location: Earth
Posts: 79
Likes: 0
Received 0 Likes on 0 Posts
Quote:
emu787:
did the AFCS detect low speed - about to get a stall or low speed buffet boundary approaching (erroneous airspeed information) and the computer pitched the aircraft nose down and applied full power thinking this is the way to recover. This would almost certainly result in instantaneous excessive speed that would test the integrity of the airframe to its limit.

takata
It seems very improbable considering A/THR and AP OFF => Alternate Law2 (protection lost) due to unreliable Air Data (ADR disagree). Moreover, Air France ADIRUs (Honeywell) are different from Quanta's (Northrop-Grumman).

S~
Olivier
to Takata :

How do you know the aircraft was actually aware of ADR DISAGREE at 0210Z ? The sequence of ACARS messages says NAV ADR DISAGREE (which sounds somehow something else than F/CTL ADR DISAGREE ?) at 0212Z, giving us a full minute at least between A/P OFF A/THR OFF and your proposed going into ALTN LAW 2.

I would rather suggest that the aircraft was, at 0210Z, just after AP disconnect, in ALTN LAW 1, not yet aware of ADR DISAGREE condition.

Which would mean that Low Speed Protection was active. Precisely, while the pilots were thinking "we're hand flying now", the aircraft applied "gentle progressive nose down input" without any kind of warning other than... STALL STALL STALL... Interesting ?

Anyone could disprove this, and do not hesitate to go very technical.
Svarin is offline  
Old 29th Jun 2009, 10:01
  #2482 (permalink)  
 
Join Date: May 2009
Location: France
Age: 59
Posts: 14
Likes: 0
Received 0 Likes on 0 Posts
Culture

I can tell you from inside AFR that 'cultural' issues would indeed lead to opposite behaviours,like 2 FO's initiating a diversion without waking the commander in the bunk..(yes,it's been done already..)
Cheers

Wil
Wilbur60 is offline  
Old 29th Jun 2009, 10:17
  #2483 (permalink)  
 
Join Date: Aug 2005
Location: fairly close to the colonial capitol
Age: 55
Posts: 1,693
Likes: 0
Received 0 Likes on 0 Posts
I see a possibility for misunderstanding underway

The rudder travel limits per the chart posted last page are for excursions in each direction starting at center. So at 350kts, the A330 rudder, as limited by the flight control software will be restrained to a 4 degree deflection. While not a typical figure of discussion, the total deflection arc is then equal to 8 degrees.
vapilot2004 is offline  
Old 29th Jun 2009, 10:39
  #2484 (permalink)  
 
Join Date: Mar 2005
Location: 41N12E
Posts: 80
Likes: 0
Received 0 Likes on 0 Posts
Culture

--------------------------------------------------------------------------------
I can tell you from inside AFR that 'cultural' issues would indeed lead to opposite behaviours,like 2 FO's initiating a diversion without waking the commander in the bunk..(yes,it's been done already..)
Cheers


Well when I'm resting I certainly do not expect a bunk call asking for a permission to deviate for weather.
There are 2 competent F/O's in front, you've briefed them, I would expect a call only for a major problem i.e. tech, medical. Not for weather deviations, even crossing the equator.
If you dont trust the other guys or you feel unconfortable when turbulence starts, just stay in the cockpit.
sleepypilot is offline  
Old 29th Jun 2009, 10:59
  #2485 (permalink)  

Sun worshipper
 
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes on 0 Posts
Svarin :
How do you know the aircraft was actually aware of ADR DISAGREE at 0210Z ? The sequence of ACARS messages says NAV ADR DISAGREE (which sounds somehow something else than F/CTL ADR DISAGREE ?) at 0212Z, giving us a full minute at least between A/P OFF A/THR OFF and your proposed going into ALTN LAW 2.
Takata is right.
You don't take a "WRN" message, coming straight from the flight warning computer to the CMC then the ATSU with the same degree of urgency as the "FLR" message which is an advisory to the maintenance department.
Look again at the reversion page already posted several times on this forum and see that , pertinent here, the condition for an auto pilot disconnect is a dual ADR fault and the reversion is to ALTN 2 PROT LOST.

emu787:
did the AFCS detect low speed - about to get a stall or low speed buffet boundary approaching (erroneous airspeed information) and the computer pitched the aircraft nose down and applied full power thinking this is the way to recover. This would almost certainly result in instantaneous excessive speed that would test the integrity of the airframe to its limit.
Nice try. You just forgot that the autothrust had disengaged, too and if anything, we have here a thrust lock. You need another theory.
Lemurian is offline  
Old 29th Jun 2009, 13:12
  #2486 (permalink)  

Sun worshipper
 
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes on 0 Posts
ttcse,
WillFrasier
Quote:
I won't exclude that the deflection seen on the Rudder at discovery and recovery is not the deflection that caused the separation, iow, full deflection to starboard. It took massive hydraulic power to put it there, and at sever, with no hydraulics or command to return it to neutral (or other), it may be there just as it was. The damage to the bottom of the Rudder may have been air induced, that portion, in the airflow, would be taking the brunt of the air.
and
And you should've also noticed that in at least one photo of the fin for AF447 while it was on the ship, the rudder was much more deflected than when the fin lay on the surface of the ocean(tell him Graybeard). Required 'massive hydraulic power ' ??
Pictures better than words :



Lemurian is offline  
Old 29th Jun 2009, 13:18
  #2487 (permalink)  
 
Join Date: Jun 2009
Location: ATL
Age: 67
Posts: 131
Likes: 0
Received 0 Likes on 0 Posts
The rudder is overbalanced. Trailing edge will point up regardless of side of rest. The focus on rudder deflection would be better spent elsewhere.
ClippedCub is offline  
Old 29th Jun 2009, 13:29
  #2488 (permalink)  
Cool Mod
 
Join Date: Apr 1998
Location: 18nm N of LGW
Posts: 6,185
Likes: 0
Received 0 Likes on 0 Posts
A polite reminder - this time - to STAY on topic, meandering this way and that will get deleted.
PPRuNe Pop is offline  
Old 29th Jun 2009, 14:01
  #2489 (permalink)  
 
Join Date: Dec 2002
Location: UK
Posts: 2,451
Likes: 0
Received 9 Likes on 5 Posts
Milt, I do not know if the specific pitot head has been tested in the relevant icing conditions (#2511).
Some pitot heads, OAT, and TAT probes have been assessed in icing tunnels, and all installed systems should have been exposed to conventional icing during aircraft type certification.

There appears to be many facets to the problem of pitot icing in this instance; first it is not conventional icing.
The atmospheric conditions and likely mechanism of icing (as you describe) are discussed in http://www.pprune.org/tech-log/37634...ml#post5027449 and the linked paper "The Ice Particle Threat to Engines in Flight".
There does not appear to be any industry standard against which a pitot could be tested in these specific conditions. Furthermore, in addition to the possibility that every design of pitot head will produce different results, so too could each location on differing aircraft types.
Ice particle induced icing appears to be greatly affected by the airflow characteristics prior to the probe – very small particles probably follow the airflow, but larger ones (and water drops) might cross the flow due to their higher mass.

Fortunately the conditions are rare, or that pilot’s previous avoidance of large Cbs has contributed to infrequent hazardous encounters. Changes in these aspects appear central to recent incidents and this accident – so what is different? The training and use of weather radar has been discussed elsewhere, as has pilot knowledge, attitudes, and behaviour. These probably have contributed, but in previous incidents the outcomes were significantly different and thus the range differing piloting standards should not have been a critical factor in this accident.

The above also appears to support the view that the loss of all pitot information is no greater risk that hazardous – Instrument External Probes, and that the aircraft can be flow relatively safely for an appropriate period (average exposure to the conditions 5nm, max 135 nm - engine ref above). However, there may be ‘hidden’ assumptions about crew competence in severe atmospheric conditions, checklist design and use, lack of primary inst, aircraft condition, lack of radar, etc, – what should the safety requirements consider for an extreme adverse combination of factors?
Other certification standards define an adequate level of structural strength for inadvertent flight in severe weather. If (IF) the aircraft broke up in flight then at this time we do not know why, nor apparently do we understand any mechanism for such a failure – we must wait and see.
Meanwhile avoidance of large storms by a wide margin would be a prudent safety action.
safetypee is offline  
Old 29th Jun 2009, 14:41
  #2490 (permalink)  
 
Join Date: Jun 2009
Location: in a plasma cocoon
Age: 53
Posts: 244
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by Svarin
to Takata :

How do you know the aircraft was actually aware of ADR DISAGREE at 0210Z ? The sequence of ACARS messages says NAV ADR DISAGREE (which sounds somehow something else than F/CTL ADR DISAGREE ?) at 0212Z, giving us a full minute at least between A/P OFF A/THR OFF and your proposed going into ALTN LAW 2.

I would rather suggest that the aircraft was, at 0210Z, just after AP disconnect, in ALTN LAW 1, not yet aware of ADR DISAGREE condition.
The ADR DISAGREEment is time stamped 02:12Z (with the uncertainty associated to ACARS time stamping). If two ADR are reported in disagreement, this imply that one ADR had already been voted out (but it was not reported as faulty like in an NAV ADR FAULT) ? The switching to ALTN LAW is time stamped 02:10Z, so you suggest that they were in ALTN LAW 1 and not ALTN LAW 2. This seems logical to me too.

Which would mean that Low Speed Protection was active. Precisely, while the pilots were thinking "we're hand flying now", the aircraft applied "gentle progressive nose down input" without any kind of warning other than... STALL STALL STALL... Interesting ?

Anyone could disprove this, and do not hesitate to go very technical.
In ALTN LAW 1, the LOW SPEED STABILITY protection is still active (pitch in direct law) so if the AP monitor a CAS that is close to/under the stall speed (in the Air Caraïbe case, the CAS plunged from 273 to 75 kts), it can order a "gentle progressive nose down input" to regain a margin versus the stall speed, an order which can be overriden.
Is it possible to simulate or compute the evolution of the (real) airspeed from that event ? (for example using the Air Caraïbe observed biases or the airspeed).
It also raises the reaction of the crew. Did they start to apply the procedure about "unreliable airspeeds" from 02:10Z ? if so did they believed to the instruction that says to "rely on the stall warning that may be triggered in ALTN or direct law because it is not affected by unreliable airspeeds but based on the AoA " ? (and let AP (re)gain speed leading to overspeed). Which part of their attention did they devote to control the aircraft, to try to understand what was the problem with their NAV and instruments & to apply the procedure ? Did they try to restart PRIM1 to regain the normal law during the first 3mn from 02:10Z ? Did they let a critical overspeed situation to unfold and exited the flight enveloppe this way ? (in the very center of a cluster of Cb). Under this scenario, the upper range of allowed airspeeds should have been exceeded rapidly after the initiation of the Low Speed Stability protection ? Is it a matter of 20 s, 30 s, one minute if the pilots let the AP do ? Would it mean that at 02:11Z or 02:12Z max, the aircraft was already in a critical overspeed ? If we have at least 2 mn of ACARS/flight left, would it mean that the overspeed was somewhat contained ?
After the ADR disagreement, they were in ALTN 2 with a very limited protection for their control surfaces (if the AP reads 75 kts instead of 275 kts like for Air Caraïbe, the rudder can be allowed 31.6° instead of ~7° max on the graph given by PJ12). If the VS was lost in mid air, the things must have unfolded very quickly from then.
Jeff
Hyperveloce is offline  
Old 29th Jun 2009, 16:12
  #2491 (permalink)  

Sun worshipper
 
Join Date: Dec 2001
Location: Paris
Posts: 494
Likes: 0
Received 0 Likes on 0 Posts
For the umpteenth time :
Conditrions for loss of Autopilot :
-THS jam
-THS position lost
-One elevator fault
-Yaw damper actuator lost.
In the above cases, the reversion is to ALTN 1

But for :
-Both engines out
- Double IR fault
-Double ADR fault,
The autopilot will disengage and the law reversion is for ALTN 2.

To go a bit further,the lost protections are :
- Pitch attitude
- High AoA in the case of a dual ADR fault
-Bank angle
- Low energy.
The low speed stability is also lost still in case of a dual ADR fault.

So your ideas about ALTN 1 and low speed stability, etc...etc... are not valid.
To rely on the CMC--> ATSU-->SatCom--> ACARS reception for determining a chain of events is faulty reasoning as the CMC is -if I may - just a snitch reporting on events, computations that have been achieved elsewhere in the systems.
Lemurian is offline  
Old 29th Jun 2009, 16:13
  #2492 (permalink)  
 
Join Date: May 2000
Location: London, UK
Posts: 90
Likes: 0
Received 0 Likes on 0 Posts
@ sleepypilot: if you re-read Wilbur60's post you'll see he referred to a diversion, not a deviation round WX
Beanbag is offline  
Old 29th Jun 2009, 16:56
  #2493 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 76
Posts: 2,484
Received 0 Likes on 0 Posts
Lemurian;

Perhaps these will help the situation. I posted this a while back along with a host of other schematics to help those who otherwise don't fly or know the Airbus A330 systems and rely on web-based generic information.





PJ2 is offline  
Old 29th Jun 2009, 17:08
  #2494 (permalink)  
 
Join Date: Jun 2009
Location: in a plasma cocoon
Age: 53
Posts: 244
Likes: 0
Received 0 Likes on 0 Posts
Lemurian, I understand (sorry, I am a bit slow). Thank you. Only an ALTN 2 could disengage the AP at 02:10Z (even if the ADR disagreed two minutes later according to the time stamping of the ACARS). And the "FLAG ON CPT PFD SPD LIMIT" (02:10Z) would be the elimination of the CAS by the CAS MONITORING function of the EFCS. All the protections were lost except the LOAD FACTOR one, the flight enveloppe protections of the two FMGEC were lost.
Jeff
Hyperveloce is offline  
Old 29th Jun 2009, 17:12
  #2495 (permalink)  
 
Join Date: Jun 2009
Location: VA, USA
Age: 58
Posts: 578
Received 0 Likes on 0 Posts
No, it wasn't the computer, it was the sensor

In the USA, near Washington, DC, a computer has been accused of failing and allowing two trains (like subways) to collide, killing 9.

computers, HA.
This is incorrect - the sensor in one section of track failed to report that a train happened to be occupying the track, hence the computer system did not "know" there was anything to slow down for.

Sure there are a number of questions (like "why, if the system knew there was a train passing through the preceding sections of track, did the system not question where that train went...?").

Different topic, but similar. A computer can only react to the data presented (or not presented), and the trick is provide some level of useful behavior when any one one piece (or more) of data (for whatever reason) is not available.

-GY
GarageYears is offline  
Old 29th Jun 2009, 17:24
  #2496 (permalink)  
 
Join Date: Jun 2009
Location: Earth
Posts: 79
Likes: 0
Received 0 Likes on 0 Posts
Gentlemen, I take this thread as an opportunity to learn.

In the schematic kindly posted by PJ2, we can see the phrase AUTOPILOT LOST. Please confirm this means the AP cannot be put back in operation.

Surely, this is not the same as AP disconnect at the time of failure, or is it ?

Thanks.
Svarin is offline  
Old 29th Jun 2009, 17:40
  #2497 (permalink)  
 
Join Date: Jun 2009
Location: in a plasma cocoon
Age: 53
Posts: 244
Likes: 0
Received 0 Likes on 0 Posts
according to PJ12's document the AP should not have disengaged only with an ADR Disagreement (and ALTN 2). It would take two ADR failures or two IR failures. Though in the Air Caraïbe sequence, you get the same ECAM FLR (and irreversible AP loss) from an ADR disagree only.
Jeff
PS) and similar problems (A/P, A/THR loss, ALTN, ADR Disagree) in... 1994:
see the documents published today at Eurocockpit - Accueil

Last edited by Hyperveloce; 29th Jun 2009 at 18:18.
Hyperveloce is offline  
Old 29th Jun 2009, 17:43
  #2498 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 4,569
Likes: 0
Received 1 Like on 1 Post
Meanwhile avoidance of large storms by a wide margin would be a prudent safety action.
unfortunately this is often subjectively interpreted.

Above the storm has been interpreted as good-enough yet is that not a greater risk to pitot, engines etc.?

And then we have the descent through heavy rain?

Somewhere the FCOM etc. needs to be standardized in this respect so at least we know after the fact what side of the judgement curve (man or machine) needs to be looked at.
lomapaseo is offline  
Old 29th Jun 2009, 18:04
  #2499 (permalink)  
 
Join Date: Jun 2009
Location: Colorado
Age: 74
Posts: 48
Likes: 0
Received 0 Likes on 0 Posts
My (limited) experience of flying in icing conditions has shown me that ice accretion is incipient rather than sudden. I doubt that the pitots 'blocked' due ice but may well have become restricted. A blockage would be easy to spot, a restriction not so.

Various AB/AF docs posted on this forum have shown concerns with blockage of pitot drains and have queried the blocking agent; is corrosion/erosion a suspect, perhaps aided by the scouring effect of ice crystals?

The AD systems would not be concerned about different readings from the pitots (due position etc.) as long as the differences were small and relatively static (forgive the pun), the system may then permit a slow build up of ice cristals in the pitot tubes.

A restriction of the drain holes would cause an increase of pressure in the pitot tube. The AP would see this as a potential overspeed and slow the aircraft nearer to the stall.

Could this have been the start of the upset?

Editted:Spelling

Last edited by EGMA; 30th Jun 2009 at 03:59.
EGMA is offline  
Old 29th Jun 2009, 18:15
  #2500 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 76
Posts: 2,484
Received 0 Likes on 0 Posts
Svarin;
Please confirm this means the AP cannot be put back in operation.

Surely, this is not the same as AP disconnect at the time of failure, or is it ?
I have found Airbus to be very careful in using such terms and take them at their word. The word "LOST" means that the autopilot cannot be re-connected until the conditions which caused it to be "lost to use" no longer exist. Your assumption is therefore correct: This is not just an autopilot disconnect.

The reason for this is, the autopilot no longer has a normal airframe or normal systems' operation. If we examine the conditions which cause the loss of the AP, we can see why.

lomapaseo;
Somewhere the FCOM etc. needs to be standardized in this respect so at least we know after the fact what side of the judgement curve (man or machine) needs to be looked at.
Failing the original opportunity to fulfill this philosophical need, ostensibly because the power and promise of "automation" was intoxicating to so many, what is now to be a re-examination has been a long time coming. In my view both hubris (not on the part of the manufacturer) and said intoxication with technology, (not entirely on the part of the manufacturer), without ever asking the question, "just because we can, should we?...", have played significant roles in the development and complexity of autoflight over the years.

That said, it works extremely well and is a significant advancement in flight safety, notwithstanding the present issues being discussed.
PJ2 is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.