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

AF447

Old 8th Aug 2009, 22:21
  #4161 (permalink)  
 
Join Date: Nov 2000
Location: Sol, sector ZZ9 plural Z alpha
Posts: 80
Likes: 0
Received 0 Likes on 0 Posts
2 out of 3 aint bad

Will, I think part of the logic behind the recommendation to replace only 2 of the 3 probes is due to the emerging idea that 3 identical units are more likely to suffer the same problem at the same time than differing units; coupled with the lack of firm hard data characterising the exact nature of icing / blockage at this time.

It is not known as fact whether the BF Goodrich units are impervious to similar issues, so the recommendation allows some additional redundancy in the form of component variation.
Clear_Prop is offline  
Old 9th Aug 2009, 00:28
  #4162 (permalink)  
 
Join Date: Apr 2009
Location: Petaluma
Posts: 330
Likes: 0
Received 0 Likes on 0 Posts
Clear Prop

If as you say the theory behind leaving a suspect Thales probe on speaks of divergent design, bravo. DD is not additional redundancy, it's a form of fault isolation due to engineering approach. Keep in mind that the new Thales has a failure on record already. To my knowledge, the Goodrich replacements are unblemished, thus far.

I'd like to believe the remaining Thales is related to 'improved' safety considerations, unfortunately, it is a face saving crumb thrown to France.

I would not rule out that the Thales is vulnerable due to drain block. The inspect/clean/reinstall may be the fly in the ointment. If the drain is blocked even a little, water plugs the tube. If ICE was the issue, the Goodrich would likely have logged some failures, (they may have).
Will Fraser is offline  
Old 9th Aug 2009, 03:30
  #4163 (permalink)  
 
Join Date: Jun 2009
Location: Georgia
Posts: 169
Likes: 0
Received 0 Likes on 0 Posts
Post

Automation can mitigate the effects of fatigue because it doesn't get distracted, lose situational awareness or suffer from any of the other very elementary human frailties such as vertigo or fear. Automation is a fine and "loyal" servant, used wisely and intelligently.
The problem is that MOST automation has NO situational awareness, as situational awareness is recognizing not only where you ARE, but remembering where you WERE and knowing the difference. The way auto pilots and most flight systems are written now is as if you jump out of a plane on a skydiving trip. While freefalling, your two feet for a moment are firmly on the back of the fellow sky diving below you. as he spreads his arms and decelerates, you gain 'gravity' relevant to his speed and thus, UNBUCKLE AND DISCARD YOUR PARACHUTE *good luck with the remaining 10,000 feet, sucker!
-----
Ludicrous? (not the singer) but that's how A/P ,F/C software is written today! Sure, there are minor bits of pseudo intelligence , what I call 'digital common sense' like storing the last 15 seconds of a probe's readings to be able to discard 'noise' of intermittent fluctuations but that's not applied across the board. Most sampling and decision making is using that very instant of data and until corrected will allow a plane coming in to land: 800ft, 700ft, 600ft, -82 ft!!! to instantly cut engines (we already landed, whoopee!) and fall short of the runway. Obviously, if the autopilot had situational awareness, it would discard the sudden drop from 600ft to -82 ft as being unsustainable and invalid and do something else (automatic TOGA?) handing over to the pilot WOULD not be a good option, I think as there would be no time.
cessnapuppy is offline  
Old 9th Aug 2009, 04:40
  #4164 (permalink)  
 
Join Date: May 2003
Location: USA
Posts: 83
Likes: 0
Received 0 Likes on 0 Posts
Angel Don't just do something, stand there !

Excellent summary. As I get older and more experienced in my profession, sometimes I realize that the best response to a dire computer generated warning is to simply do nothing: scan instruments and perform a quick "reality check" but not take aggressive action unless warranted....

Cheers !
averow is offline  
Old 9th Aug 2009, 12:57
  #4165 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 4,569
Likes: 0
Received 0 Likes on 0 Posts
The problem is that MOST automation has NO situational awareness, as situational awareness is recognizing not only where you ARE, but remembering where you WERE and knowing the difference.
An example of situational awareness would be the FADEC systems on engines.

They compare multiple inputs and when they disagree (presumed fail) the FADEC latches at the last known reliable condition of flight.
lomapaseo is offline  
Old 9th Aug 2009, 14:40
  #4166 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 75
Posts: 2,482
Likes: 0
Received 0 Likes on 0 Posts
lomapaseo;
An example of situational awareness would be the FADEC systems on engines.
I think I understand what you mean and I think we understand that there may be as much a philosophical meaning as a semantic one here. No software is "aware". My very simple understanding of software is, while software can mimic learning, software doesn't "learn" in the meaning of the term we usually understand. While I am aware that those that work in AI will likely take issue with such views, the kind of sofware that guides an airliner is by comparison, pretty basic, an algorithm, without the ability to anticipate or recall in a human way. cessnapuppy touches on this in an interesting way. The processors and memory capacity are similarly pretty basic as are the displays. The trade is perhaps in robustness.

For a number of reasons, I am not sure that "fuzzy logic" solutions are suitable to airline work. I may be missing something but I don't think the problems of flight are that complex that such a sophisticated approach is necessary. The problems driving change and "improvement", (sometimes known as "progress"...), are all about cost, not physics.
PJ2 is offline  
Old 9th Aug 2009, 16:16
  #4167 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes on 0 Posts
The great benefit the human brain has over the artificial is the capacity to 'think outside the box' which by definition AI cannot do. I fear also that the human pilots are on the way to losing that faculty too.
BOAC is offline  
Old 9th Aug 2009, 16:52
  #4168 (permalink)  
 
Join Date: Apr 2009
Location: Petaluma
Posts: 330
Likes: 0
Received 0 Likes on 0 Posts
BOAC

I'm seeing some overlap between this thread and the Colgan accident.
Your last post focuses directly on the interface of Automation and manual flight. It isn't a stretch to say that the heightened hazard is at handoff. AF447 and Colgan both had a/p drop in a charged and atypical F/D setting. It is unknown what the exact chain was in 447, except to say that the a/c was lost in/at the time frame in question; with Colgan the a/p disconnect happened just prior to shaker.

With 447, the challenges appear to be IAS and/or weather related, with Colgan pilot error(s). Either way, training 'to the moment' is under review.
In 447 the 'errors' in question appear to be mechanical in origination, with Colgan, likely PE. Either way, it's of interest where the training in both instances takes the discussion. Pitch and Power.

It's inescapable that the default is manual flight. 447 aside, I've noticed broad similarities in many of the more recent accidents. Turkish, Perpignan, Colgan, Birgenair, even Continental off runway, these are accidents whose direct cause was a loss of Pitch awareness/control. 447 may have been. It may be simplistic, especially for the more detail oriented, but to the extent that tragic consequences have resulted, a focus and consequent discussion of basic flight issues seems to need to be had.

Personally, I don't think it's necessary to start Stall training ATP's. It seems to be something else that is present, or lacking. If a disease can be prevented, it doesn't need to be treated.
Will Fraser is offline  
Old 9th Aug 2009, 17:09
  #4169 (permalink)  

Do a Hover - it avoids G
 
Join Date: Oct 1999
Location: Chichester West Sussex UK
Age: 90
Posts: 2,206
Likes: 0
Received 0 Likes on 0 Posts
Some (me and my 330 mate) see it as a pity that when airspeed info is lost the automatics are not programmed to hold the appropriate attitude and power thus allowing the crew to concentrate on comms and possible fixes without the need to handle the aircraft at the same time.
John Farley is offline  
Old 9th Aug 2009, 17:24
  #4170 (permalink)  
 
Join Date: Dec 2007
Location: Cornwall-on-Hudson, New York
Posts: 876
Likes: 0
Received 0 Likes on 0 Posts
But isn't that why there are two of you? What I learned--okay, 40 years ago--was that if you had a problem, the captain gave either himself or his first officer the job of flying/controlling the aircraft, while the PNF got out the emergency checklists, sent necessary comms and did the troubleshooting. Try to both do it and you can put an entire L-1011 into the Everglades without even knowing it....
stepwilk is offline  
Old 9th Aug 2009, 17:32
  #4171 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 4,569
Likes: 0
Received 0 Likes on 0 Posts
PJ2


I don't think the problems of flight are that complex that such a sophisticated approach is necessary. The problems driving change and "improvement", (sometimes known as "progress"...), are all about cost, not physics.
I suspect that there is room for understandings here. Let's put aside the cost aspect as it tends to divert these kinds of discussions.

Generally the idea behind a product design function is to provide what the pilot wants (cost is a later gate that it must pass). So IMO the discussion should be arround the area of what the pilot needs.

IF "improvement and progress" are the desired output. Then the idea is to agree on what constitutes this and proceed from there. For this problem area I was tending along the thinking that the crew needed the confidence to take the time to sort out the changing situation. So I was following the thread hint (my read may have been wrong here) that a computer could not help much here. Thus my response was aimed at the example of how the computer can best be trusted vs the rapid responses of man.
lomapaseo is offline  
Old 9th Aug 2009, 18:13
  #4172 (permalink)  
 
Join Date: Aug 2003
Location: FR
Posts: 223
Likes: 0
Received 0 Likes on 0 Posts
There was some discussion about what is the appropriate reaction to a stall warning. Now, it is not clear at this time whether or not there was a stall warning in this case. But even if a stall warning can not be issued in ALTN2, I think it remains possible that it happened earlier? For example, in case the initial CAS drop has been below the 30kt/second threshold?
pax2908 is offline  
Old 9th Aug 2009, 18:26
  #4173 (permalink)  
 
Join Date: Aug 2007
Location: inmysuitcase
Posts: 208
Likes: 0
Received 0 Likes on 0 Posts
It's inescapable that the default is manual flight. 447 aside, I've noticed broad similarities in many of the more recent accidents.
Im with you, bufalo and amsterdam crash; they both were able to crash a, for the rest, perfect a/c!
I use those 2 crashes during my profchecks, and yes I am no crash-investigator nor expert, but i do have "some" hours both up-front and behind in the sim. And i have seen a tendency among pilots to fixate instead of AT LEAST 1 (ONE) PILOT FLYING!
testpanel is offline  
Old 9th Aug 2009, 19:09
  #4174 (permalink)  
 
Join Date: Mar 2009
Location: us
Age: 63
Posts: 206
Likes: 0
Received 0 Likes on 0 Posts
I remember reading some impressions by a russian test pilot of, among other things, an A310. He was rather down on the stall behavior. The stall came on at low AoAs with little or no warning, and the ac lost lateral stability rather dramatically and pitched up making the situation worse. He described the ac as a case of the mfg trying to compensate for the ills of an aerodynamically questionable platform with electronic doodads.

Granted that AF447 is a different plane, although sharing the same gene pool as the A300/310, how sound IS the airframe? When the electronic protections cut out under less than optimal conditions and the pilots are required to manually fly what may be an evil-handling airframe?
vovachan is offline  
Old 9th Aug 2009, 19:10
  #4175 (permalink)  
 
Join Date: Apr 2009
Location: Petaluma
Posts: 330
Likes: 0
Received 0 Likes on 0 Posts
At cruise, a/p is at its most practical and important aspect. To select a/p only to sit on the edge of the seat and birddog every bump and discrepancy is to eliminate the need for it. On a sophisticated transport, autoflight is a form of commanded complacency. So its disconnect (uncommanded) of course comes at an inopportune moment. After a series of warnings, chimes, and an uncertainty in the F/D, one pilot needs to retrieve the checklists and troubleshoot, solo ? Yes, because the PF needs to update his scan, 'get' the panel and the 'feel' starting with Pitch and Power, from a degraded (of course) SA. The PNF is at an immediate disadvantage, all the alerts are after the fact (of course), and there are many. Both Pilots are behind the a/c. The a/p was behind enough to need to disconnect. No matter the reason, an unexpected a/p disconnect demonstrably causes problems, some of which seem to have gone unaddressed.
Will Fraser is offline  
Old 9th Aug 2009, 19:20
  #4176 (permalink)  
 
Join Date: Sep 1999
Location: United Kingdom
Posts: 926
Likes: 0
Received 0 Likes on 0 Posts
Un-Substantiated rumor doing the rounds that the a/c may have reached 47000 ft during it's upset. (Convective lift?)
Sorry if this has already been posted. I believe it is a result of further analysis of the recieved ACAR's reports. However I repeat only a rumor.
IcePack is offline  
Old 9th Aug 2009, 19:29
  #4177 (permalink)  
 
Join Date: Nov 2006
Location: UK
Age: 62
Posts: 37
Likes: 0
Received 0 Likes on 0 Posts
Vovachen a310 stall

Can you back up your statement regarding the A310 stall? I don't think what you say makes any sense, details please?
proteus6 is offline  
Old 9th Aug 2009, 21:31
  #4178 (permalink)  
 
Join Date: May 2002
Location: Canada
Posts: 155
Likes: 0
Received 0 Likes on 0 Posts
The Russian test pilot's comments reminded me that I am not enamoured with the FCC's on AI products making rudder inputs without the pilot's knowledge or desire (yaw damper inputs excepted). Watching any video of the 380 shows substantial rudder inputs and reversals on short final.
Tree is offline  
Old 9th Aug 2009, 21:55
  #4179 (permalink)  
 
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes on 0 Posts
The stall came on at low AoAs with little or no warning, and the ac lost lateral stability rather dramatically and pitched up making the situation worse.
Unlikely indeed if we're talking subsonic well short of Mcrit
HarryMann is offline  
Old 9th Aug 2009, 22:14
  #4180 (permalink)  
 
Join Date: Jan 2008
Location: Herts, UK
Posts: 748
Likes: 0
Received 0 Likes on 0 Posts
IF...

IF 35,000 > 47,000 ft

4 minutes at 50 ft/s (a v. high thermal climb rate in the troposhere, 20 ft/s is quite normal)

If a/t out, manual flight, the throttles would need to be cut quick if encountering that sort of thermal bubble. And going 'over the falls' out the backside of that could be quite an experience.

IF...
HarryMann is offline  

Thread Tools
Search this Thread

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

Copyright © 2023 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.