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

AF447

Old 27th Jun 2009, 19:00
  #2421 (permalink)  
 
Join Date: Apr 2009
Location: Petaluma
Posts: 330
PJ2

You have claimed that I, as a member of the group of 'all other posters', hold that the computers somehow prevented the pilots from controlling the a/c in some way.

My position is exactly the opposite. Given the regime at the beginning of the ACARS, a/p flying, cruise alt, pilots monitoring, etc. I have held that given the parameters allowed the a/p, the flight crew were satisfied with the situation (this is of course assumed). The a/c experiences some turbulence or other challenge the a/p is unable to track, and it trips off.
Following directly its disconnect is entry into Alternate Law. It is at this point things begin to go pear, or so we can entertain, given the messages on the data sheet.

In some as yet unknown sequence, multiple failures are reported, and experienced as warnings, flags, and multiple displays. I would suggest that it appears the PF and PNF were 'startled', first by the disconnect, then by the computers programmed responses such that they were possibly overwhelmed; it is at this juncture that an untrained for and (perhaps) deliberately ignored result may have eliminated any possibility of recovery. As per the other (and recent) unreliable AD incidents show, to recover demands a crew be incredibly sharp, profusely trained, and able to selectively ignore some of the computer's conclusions. On this night it wasn't to be, or so we can propose.

As I have said, the Airbus330 is a phenomenal a/c, no one can deny it. What might be lacking is a recognition of, and training for, some blind spots in programming, and a hard look at the imperfections in the interface between FBW and the pilots. As far as that goes, the only real difference so far making the AB unique is Alternate Law; it is as you say quite possible no a/c could have escaped the fate of 447.

Loss of all pitot/statics has been discussed, it might have caused an overreaction in the software (perhaps not), and an underreaction in the pilots. If this is true, we need to put pilot and FBW back on the same page.

cheers, Will

Last edited by Will Fraser; 27th Jun 2009 at 19:21.
Will Fraser is offline  
Old 27th Jun 2009, 19:32
  #2422 (permalink)  
 
Join Date: Jan 2008
Location: Los Angeles
Posts: 27
PJ2;
I think there are good reasons to base criticism of the 320/330/340 design philosophy upon, but citing notions that "the pilot has no control over the airplane", or that " 'the computers' have taken control" is not a solid, nor informed basis for such criticism.
Thoughtful post and thank you for it. The A330/340 series represent the pinnacle of achivement in control automation. Essentially, the design in its entirety is a "system of systems" and this area is still very much a work in progress in all control systems philosophy. It's not the control systems themselves the represent the "problem", as you rightly point out, but the manner in which they interact both between themselves and with the operating crew.

Will (IIRC) brought forward the concept of "supremacy" in terms of the relationship between the operational philosophy of the avionics suite in its complex totality and that of the essentially passive crew, whose operational role is now more of observers, rather than participants, viewing events through myopic PFDs .

Nothing I've read suggests anyone posting here is a Luddite and I believe it's generally accepted that the level of automation will continue to increase, to the benefit of all. ABI are very much at the forefront of this and, rightly or wrongly, will take the flak when the holes line up.

It seems, to me at least, that the industry may have got ahead of itself in using the speed of digital electronics to marry functions which traditionally had little or no direct interdependance without paying sufficient attention to the need for an elegant regression to manual control.

Thanks again.

24V
24victor is offline  
Old 27th Jun 2009, 19:47
  #2423 (permalink)  
 
Join Date: Mar 2008
Location: At home, retired 2012
Age: 71
Posts: 14
stepwilk
Well, somebody finally said what I've been wondering for 123 pages of posts. Back in the Dark Ages when I learned to fly (mid-1960s), we were taught that what you did if you flew into Cunim was 1/maintain attitude and wings-level, 2/hold pretty much the power you've got, and 3/ignore altitude. Nobody ever said much about airspeed, which took care of itself if you did 1, 2 and 3.


That is absolutely true and it will stay that way… but…
It is very difficult to perform step #1 in alternate law, as the “protection” against too low or too high speed – be it erroneous or not, will demand a considerable permanent force on the stick, until the computers, not the pilots, switch to direct law.
Quote, FCOM, my underlining,
PROTECTIONS
Low speed stability
At low speed, a nose down demand is introduced in reference to IAS, instead of angle of attack, and alternate law changes to direct law.
It is available, whatever the slatslflaps configuration, and it is active from about 5 knots up to about 10 knots above the stall warning speed, depending on the aircraft's weight and slatslflaps configuration.
A gentle progressive nose down signal is introduced, which tends to keep the speed from falling below these values. The pilot can override this demand.)

I wish I had a DDL (Directly to Dorect Law) Switch (not a push button).
jcarlosgon is offline  
Old 27th Jun 2009, 20:01
  #2424 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 72
Posts: 2,429
24victor;

And thank you. I realize that in saying the things I have, I risk treading on people's views and do not wish to do so simply to disagree and leave. I think it was important to offer clear reasons why I think suggesting a re-focussing or a re-informing of what autoflight is and what it isn't as it relates to this accident and I hope I have done this. I know it won't please everyone.

If all systems are working as intended/designed, the 330/340 types are a dream to fly; the entire autoflight system is nothing short of brilliant, but where we may have issues is in that what we could call the "corpus callosum" ...that intellectual space between a very complex machine and an even more complex human operator.

Thank you for your response.
PJ2 is offline  
Old 27th Jun 2009, 20:03
  #2425 (permalink)  
 
Join Date: Jun 2009
Location: Iowa
Posts: 9
As per the other (and recent) unreliable AD incidents show, to recover demands a crew be incredibly sharp, profusely trained, and able to selectively ignore some of the computer's conclusions.
I think this is what is making me confused. The computers conclusions? The computer doesn't produce conclusions. This is as far as I know at least. The pilots interface to the airplane remains the sidestick. Throughout the entire episode, the sidestick remains the only control input. According to one of the published papers on the Airbus systems, the failure of any of the FEP systems should only result in an increase in the pilots workload (which would be equivilant to the pilot operating as he did without any FEP system).

As far as I can see, if FEP starts failing, the pilot just continues steering from the sidestick. S/he just needs to be aware that if you make bad choices for inputs they could damage the aircraft. Is this not right?
jeremiahrex is offline  
Old 27th Jun 2009, 20:06
  #2426 (permalink)  
 
Join Date: Jun 2001
Location: East of the Sun & West of the Moon
Posts: 286
lomapaseo;

Quote:
One thing that sticks with me so far is that the pilot comments on this thread seem to have a pretty good idea of what they would have done in the the loss of air-data turbulence scenario.

If my interpretation is correct then this is a pointer for searching for other factors.

PJ2;

Quote:

Yes. That is one of the reasons I wished to take some time to write this rather longish post. I think looking towards "the computers" for a cause is a rabbit trail; we will not be able to determine what happened by following it.
lomapaseo, PJ2

Thank-you for bringing the discussion back down to earth.


Will Fraser;
Quote:

I proposed early on that the interface between Otto and flying pilots is weak and in an out of balance "command" way. It is too quick and too absolute, seemingly wonderful traits, but deadly under certain circumstances. Pilots are comfortable (too much so) when things go pear and the computers make command decisions that are precipitous, and leave the ultimate pilot out of "the loop".

Imagine flying along fat and happy, and in less time than one can belch, the box cascades decisions and reports to be assimilated immediately with the a/c and its contents at risk!! Dark, turbulent, seven miles up, and the box hands two people two folders of data, to be understood and acted upon whilst the seats under your primitive "cheeks gyro" are telling you you have one half second to input a control.

The previous reports of similar loss of data flow involved less than harrowing conditions, and were successfully escaped; conclusions about those incidents are being debated even now, and you are at ease with the highly possible loss of data, panel, as, etc. to be dumped unceremoniously on the pilots laps???

This interface is the 400 kg gorilla no one seems to be discussing.
Will,

Frankly I'm getting tired of your dramatic, near hystrionic imaginings. Based on your comments you don't really seem to have much of an idea how this airplane actually works or what the situation that presents itself in the cockpit will look like or how it is dealt with by either pilot or aircraft.

I do. And it is nowhere near "the a/c and its contents at risk!!" "in less time than one can belch" scenario you keep trying to conjure. Now, before you come up with some "No, you don't really, you just think that based on sim exercises" type remark, let me state that I really do. This is the result of an event in the real aircraft (in this case an A320) at night in bad weather. My experience was the second (at that time, 14 years ago) ever recorded near-simultaneous dual TAT probe failure on an Airbus family aircraft (later traced to non-common failure modes with an incredibly low probability of occurence). Because it was such a rare experience it was fully investigated by Airbus even to the point of sending update telexes directly to me as the PIC involved.

The incident, which included a very similar set of failures to that which I have seen from the AF ACARS print-out other than the assumed loss of airspeed indication, did not involve the "box cascading decisions" or "two folders of data, to be understood and acted upon whilst the seats under your primitive "cheeks gyro" are telling you you have one half second to input a control." (do you happen to write for the Sun when you're not posting here?).

What did happen was first the loss of AP1, then of AP2 and Auto-thrust. This was followed by reversion to Alternate Law and ECAM messages related to changes and degradations in a number of other sub-systems. This occured while climbing at night through cloud with turbulence somewhere in the mid-twenties. None of the losses as they occurred caused any significant effect on how the aircraft handled, and in that sense the transition from autopilot control to manual control by the pilot was virtually seemless. The number of ECAM messages to be dealt with, though significant, did not cause "cheek gyro clenching panic" or even a great deal of confusion for the crew in terms of how to configure or operate the aircraft. The biggest challenge was that the source failure (a dual one as it turned out) was difficult to correlate to the system failures observed. This turned out to be a result of the failure message for the STBY TAT probe failure not being sufficiently unambiguous to be differentiated from the message for the CAPT TAT probe failure. Hence it was only immediately obvious that one source had failed, not two, which caused a good deal of FCOM investigation as we went along to try and match what we were observing to that which had to have failed.

Shortly after we left our departure airfield it had closed due to weather and likewise the weather had reduced all nearby airfields to CAT II conditions or lower. Given our situation we did not consider a return or diversion for landing in such conditions to be an optimal choice and hence we then flew the airplane (yes, all manual flight at high-altitude) for over an hour across the Rockies to get to a suitable diversion airfield. As the situation evolved, it was the more mundane aspects that proved a bit more challenging, the aircraft handling itself was a complete non-issue. The dual TAT probe failure locked some of the FMGC functions such as flight phase changes, adding a few tricks to how to navigate to our diversion airport and make approach preparations. When we arrived at the diversion airport the approach was conducted hand flown in alternate law with an uneventful landing in direct law once the gear was extended.

The moral of the story, Will, is that the probe faliures that have been described, and which we believe may have occurred to the AF aircraft, do not result in things going immediately and dramatically pear shaped or the aircraft making "command decisions that are precipitous, and leave the ultimate pilot out of "the loop". They are a challenge to deal with, as they would be on any aircraft, but they do not materially affect the controlability or stability of the aircraft. Trying to cast this as "the 400 kg gorilla no one seems to be discussing", is a venture into over-wrought hyperbole at a point where there is no information that supports how a loss of control may have occurred and whether or not it was from forces that would have caused an upset in any aircraft, modern or geriatric.

ELAC

Last edited by ELAC; 27th Jun 2009 at 20:37.
ELAC is offline  
Old 27th Jun 2009, 20:06
  #2427 (permalink)  
 
Join Date: Jan 2008
Location: Pretty far away
Posts: 316
IF (just if) the Capt was not on the flight deck, is there any part of the AFR pilot culture which would make the two folks up front hesitant or reluctant to make a significant deviation (time & fuel implications) around severe weather without the Capt's prior knowledge?
Right on the money ! There is and it is mainly due to the differences in reactions from skippers once they discover the aircraft miles away from the track. It's happened to me when I was a F/O and being rather pig headed and a woos, I couldn't care less about what the guy said as long as we were out of the CB's. Not the case for every copilot.
I remember being amazed............and scared as a F/O at how close to yhe CB's some of these older guys would take us. As to the radar which I was adament to use with full gain, it was a constant battle with the captain turning it back to auto telling me " You're going to scare yourself boy. See, it's all green now ". I swear I am not kidding.
Today, I really have to brief the F/O's not to hesitate to deviate as much as it takes..........and I don't always get the results I want.
There is both a lack of training and a lack of awareness. A lot of younger chaps have no clue, or had no clue how dangerous it can be and 447 was a brutal wake up call. Even now, there is a huge denial about the possibility that the crew might have flown into this weather, and this despite all evidence shown by Tim Vasquez.
One thing I can tell you is, a lot of us are going to stay on the flight deck as long as it takes before getting our well deserved r&r's.
My opinion is that, regardless of the talentof the chaps sitting up front, there is a breach in the chain of command. No one is really in charge and in a dire situation like the one 447 encountered, there is no way you can have ONE single person in command who makes the life saving call. This is why the captain has to be in his seat until the situation allows him to take his rest safely.
Not a put down for F/O's, just a hard fact.

Some of the posts here are full of very simple wisdom. Show respect to mother nature and go back to basic. Unfortunatly, keeping attitude and thrust isn't rehearsed in the sim. we're supposed to be happy with what's in the books. Lack of time, lack of money..........etc
I think the system has just gone cookoo.
Me Myself is offline  
Old 27th Jun 2009, 20:08
  #2428 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 72
Posts: 2,429
As far as I can see, if FEP starts failing, the pilot just continues steering from the sidestick. S/he just needs to be aware that if you make bad choices for inputs they could damage the aircraft. Is this not right?
Yes, that is correct, and the cautions are in the AOM and QRH. Handling the aircraft is completely conventional in Direct Law.
PJ2 is offline  
Old 27th Jun 2009, 20:22
  #2429 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,583
To bring this back to the thread topic - because of the limit on 3 letter search here I cannot easily find if we established whether the a/c used EPR or N1 on the engines - anyone know? If EPR, is there any way the 'faulty' ADUs could affect the readings?
BOAC is offline  
Old 27th Jun 2009, 20:25
  #2430 (permalink)  
 
Join Date: Jun 2001
Location: East of the Sun & West of the Moon
Posts: 286
BOAC,

The GE engines use N1.

ELAC
ELAC is offline  
Old 27th Jun 2009, 20:29
  #2431 (permalink)  
 
Join Date: Aug 2007
Location: Alabama
Age: 54
Posts: 366
jcarlosgon

Is it correct than in case of dual ADR failure (ADR disagree) the aircraft reverts to alternate law 2 and therefore the low speed stability is lost?
FrequentSLF is offline  
Old 27th Jun 2009, 20:30
  #2432 (permalink)  
 
Join Date: Jun 2009
Location: in a plasma cocoon
Age: 49
Posts: 244
How can we be sure that all the 3 inertial references (attitude) were faulty/lost ? Does the 02:11Z FLR : 341234 IR2 EFCS1X,IR1,IR3 point toward a triple inertial failure or just that the IR2 has been voted out by IR1 and IR3 ? Has aviation history ever recorded a triple inertial failure ?
In fact, how can we even be sure that one IR actually failed ? Even in the Qantas "NAV IR 1" fault report, the IR part of the ADIRU1 was probably not faulty, the culprit was the ADR1 delivering spiky AoA measures (which were not voted/filtered out by the AP despite multiple "ADR1 fail" indications, even after ADIRU1 was switched off).
Jeff
Hyperveloce is offline  
Old 27th Jun 2009, 20:50
  #2433 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,583
Thanks ELAC - that's good news!

I commented a while back on 'crew rest' issues and as a (retired) short-haul w***er [with NO experience of flight-crew 'relief(')] I expressed the opinion that I would not expect a Captain to 'take his rest' with the ITCZ to cross and 6-7 hours of relative peace to follow. I was somewhat surprised yesterday to hear from a 5 year retired BA 747 trainer friend that in his life the system was very 'rigid' and that rest would be taken 'on schedule'. I had always assumed that as Captain and 'responsible' I would, in that position, have dispatched a co to rest in my rest slot. Certainly with the known 'history' of probe problems in weather I would be pretty certain I would have stayed. Anyone to comment?

Hyperv - I don't think anyone has suggested that? AD failure is all I think we assume although there have been questions over ISIS? As far as we know, I think, it is assumed that IRS systems would be unaffected UNLESS denied by the software.

FrequentSLF - armed with my new knowledge from the Tech forum and with acknowledgement to SmartCockpit, High AoA protection does appear to be lost with dual ADR failure.
BOAC is offline  
Old 27th Jun 2009, 20:58
  #2434 (permalink)  
 
Join Date: Jan 2008
Location: PARIS FRANCE
Age: 73
Posts: 52
Upset and managing the plane

This is not a comment on the Air France crew, but a remark on the way training in airlines (the one I know anyway) is becoming just a quick dip in the books.I have been retired for five years now, but I believe that not only has there been no strengthening or widening the scope of rating qualification on the big liners, but to the contrary : many necessary enlightments that were given at the time by ground instructors have disappeared…with them. Young pilots, with well-made brains, are trained quickly to an acceptable standard, that means dealing competently with the aircraft faults in the simulator, clearing them, reading the status…Then they fly the A320 with very low flight hours at the beginning, and after a short experience on that remarkable plane (no joke intended, I liked it) they naturally have a tendency to believe that they are airline pilots. The simulator sessions are programmed to deal with what is in the regulations (engine failure, engine fire etc…) and not much else as there is no spare time. Economic pressure all around prevents the Chief Pilots from dealing with the « nice to know ». And when you are on the A330 for example, you might land the plane three times in the month as a copilot, and eventually land it in the simulator to get the necessary « recent experience ». Some fly sailplanes or do aerobatics during their spare time, but they are not the majority, far from it.
I just read a paper by an Airbus Test Pilot (Airplane Upset Recovery) which sums up what he sees as most important, after participating in the very exhaustive and remarkable « Airplane Upset Recovery » (see FAA Airline Operator Training) in which Boeing, Airbus, Flight Safety worked together. What does Captain Wainwright, Chief Test Pilot Airbus, tell us ? The study (which I feel is a need to know for every airline pilot) « is aimed at preventing loss of control on conventional aircraft. It is not aimed at Fly by Wire aircraft. There is no need for this type of continuation training on protected aircraft, although a general knowledge of the principles involved is useful »
I have the utmost respect for Captain Wainwright, and he is of course, himself, fully ready to deal with any upset…but he has apparently no idea of the way this knowledge slowly disappears in some airlines (large ones). This attitude might even be dangerous for Airbus as we can see that there is an evident need to train for such upsets…Well, I may be wrong, but his statement frightens me a little…
NARVAL is offline  
Old 27th Jun 2009, 22:39
  #2435 (permalink)  
 
Join Date: Jun 2009
Location: France
Age: 72
Posts: 193
Hyperveloce: I agree with your interpretetion of the message about IR2. A number of people (including me) suspect other problems, beyond the Pitot probes, interfered.

And I disagree about your statement that only the AoA probe was spiky in the QF072 accident. The FDR and The QAR reported spikes from many parameters, mostly air data but also pitch and roll.

The upset only occured because the master PRIM failed to correctly filter the AoA spikes. ATSB found that IR1 was flaged inop first by FMGEC1.
DJ77 is offline  
Old 27th Jun 2009, 22:45
  #2436 (permalink)  
 
Join Date: Jun 2009
Location: Toronto
Age: 75
Posts: 117
Management of the unexpected

For Narval and others who have expressed similar opinions, I can empathize with your sentiments as a very similar degradation is occurring in my profession - medicine. Experience (however expressed: gut feeling, seat-of-the-pants, I know, I don't know how, etc), is discounted as being unquantifiable and therefore invalid. I've encountered many situations were health care professionals have responded only to the numbers. Fortunately, I was able to point out that the numbers didn't fit the situation so no harm was done. I've been designing and programming computers since 1964 and practicing medicine since 1978. I still don't trust computers but I do trust my gut. Computers certainly do make life easier but they are not professionals.

Last edited by kilomikedelta; 27th Jun 2009 at 22:55.
kilomikedelta is offline  
Old 28th Jun 2009, 00:17
  #2437 (permalink)  
 
Join Date: Jan 2002
Location: USA
Posts: 286
Why does the question continue to rise relative to what the Captain was doing at the 0214z time frame. Either the crew as assigned was qualified or they were not. Whether the Captain was at rest, or sitting in the left seat should not be a factor. If it was, then AF has some serious problems with its bid lines, and more importantly, its company policy in how it staffs it flights.
wes_wall is offline  
Old 28th Jun 2009, 01:16
  #2438 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 5,247
The simulator sessions are programmed to deal with what is in the regulations (engine failure, engine fire etc…) and not much else as there is no spare time
An interesting but soft point (little data).

On the one hand I read opinions from pilots who claim to understand and know how to handle loss of air data in turbulence. This seemingly contradicts the claim of inadequate training.


If something is missing in the training we need understand where and how to fix it.

It's probably too early to conclude this without more data including what happened in this accident.

I'm a fan of simulators but having lost arguments in the past about nice-to-have, training we have to face the reality of prioritization and availability of pilots for unique training.
lomapaseo is offline  
Old 28th Jun 2009, 01:55
  #2439 (permalink)  
 
Join Date: Mar 2008
Location: At home, retired 2012
Age: 71
Posts: 14
FrequentSLF

Is it correct than in case of dual ADR failure (ADR disagree) the aircraft reverts to alternate law 2 and therefore the low speed stability is lost?
Yes it is.
jcarlosgon is offline  
Old 28th Jun 2009, 02:42
  #2440 (permalink)  
 
Join Date: Jan 2009
Location: alameda
Posts: 1,053
if the captain was in back and the other pilots were at the controls and the airline and french authorities deemed them fit for those positions, THEN all was ok.

perhaps, all of us super experienced jet jockeys should think for just a second that: these pilots were as good as we are and maybe something really unexpected happened.
protectthehornet is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information

Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.