Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

AF 447 Search to resume

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

AF 447 Search to resume

Old 10th Nov 2009, 14:00
  #41 (permalink)  
 
Join Date: Nov 2000
Location: not a million miles from old BKK
Posts: 494
Likes: 0
Received 0 Likes on 0 Posts
In several earlier posts reference was made to aircraft constantly broadcasting data and the bandwidth considerations that resulted from that.
I agree. But, what's wrong with doing it the other way round? The aircraft does not have to broadcast what is, essentially, routine data all the time but rather let the aircraft broadcast a trigger signal to say 'I'm in trouble' which in turn triggers a detailed interrogation from the base station. A lot less bandwidth used up there methinks.
Xeque is offline  
Old 10th Nov 2009, 14:26
  #42 (permalink)  
 
Join Date: Jan 2005
Location: France
Posts: 2,315
Likes: 0
Received 0 Likes on 0 Posts
Xeque,
Read some of the previous posts, in particular PJ2's on the previous page.

By the time the aircraft "knows it's in trouble", it may already be diving vertically, with the satellite aerial blown off the roof by an explosive decompression or such-like.

CJ
ChristiaanJ is offline  
Old 10th Nov 2009, 14:30
  #43 (permalink)  
 
Join Date: Mar 2009
Location: us
Age: 63
Posts: 206
Likes: 0
Received 0 Likes on 0 Posts
By the way, I don't think any such system (deploying recorders) exists but that is beside the point.
To the point, jettisoned buoys containing recorders are in fact used on military naval ac, so the technology exists. While it would be overkill for most medium-short range planes flying over land, for the long-range planes which spend most of their life over oceans there is no reason not to have it.
vovachan is offline  
Old 10th Nov 2009, 15:02
  #44 (permalink)  
 
Join Date: Jul 2007
Location: Switzerland
Age: 78
Posts: 109
Received 7 Likes on 2 Posts
ChristiaanJ about satellite upload:
The continuous data stream into the FDR includes far, far more parameters, llike attitude (all axes), airspeed, engine parameters, control surface positions, etc., sampled every few seconds, sometimes even more often.
It's way, way beyond what could be transmitted via a satellite link, or stored and managed at the receiving end, for the tens of thousands of commercial flights in the air at any one time.
Agreed. But it is not necessary to stream the full data. Data may be compressed by a method called "delta modulation" whereby only changes of values are transmitted once their base value is established (as in modern video and audio streams). In level flight, compressed FDR data could be reduced to a manageble data stream. Data uploaded this way may not give a precise picture when data changes rapidly (like an FDR) but might still be helpful to establish the sequence of events.
clearedtocross is offline  
Old 10th Nov 2009, 17:44
  #45 (permalink)  
 
Join Date: Jul 2009
Location: Surrey, UK
Posts: 130
Likes: 0
Received 0 Likes on 0 Posts
Re: auto-jettison of black box: How does the aircraft predict when it's going to crash and let the black box go? Interesting idea.
As to other comments, it took time to mobilise the submarines, ships with ROVs etc and send them to the suspected crash site. We had the RovHomer ready in a day or two and despatched it in collaboration with the BEA by air to Africa where it had to be transported to a vessel with a ROV on it, which had to steam to the site etc etc.
Detection relies on "line of sight" to the black box signal emitter, so if box is in a ravine etc, the sound cannot get out and hence you cannot detect it.

Last edited by HamishMcBush; 10th Nov 2009 at 19:36.
HamishMcBush is offline  
Old 10th Nov 2009, 19:25
  #46 (permalink)  
 
Join Date: Jan 2005
Location: France
Posts: 2,315
Likes: 0
Received 0 Likes on 0 Posts
clearedtocross,
ANY form of data transmission adds a whole string of imponderables, the data link being lost at the critical moment being only one of them.
Having the FDR and CVR "listen" to the aircraft until the last possible second, then try to protect and retrieve those data, has been the method of choice so far.

"Delta modulation" only needs the 'base value' to be corrupted briefly for the data to become unusable, until the new base has been transmitted (video is a good example... we're seeing it here locally right now, with the introduction of digital TV and dodgy reception....)

To me, if something is going to be done, it should be an upgrade of the ULB, with louder pings, and maybe a transponder function after a certain time. Not sure if it can be done within the same "form, fit and function" envelope?
It may be a "sellable" idea for the FAA, ARINC, etc. because it would improve matters in all cases where a recovery from underwater is required.

As to the data transmission idea, one could say we already have a primitive kind of that in the form of ACARS.... without the ACARS messages from AF447 we wouldn't even be able to make any wild guesses.

Originally Posted by HamishMcBush
Re: auto-jettison of black box: How does the aircraft predict when it's going to crash and let the black box go? Interesting idea.
Much what I was saying, really. You want those boxes on board to the last second.

CJ
ChristiaanJ is offline  
Old 10th Nov 2009, 20:27
  #47 (permalink)  
 
Join Date: Apr 2008
Location: Durham
Age: 62
Posts: 187
Likes: 0
Received 0 Likes on 0 Posts
(I'm anticipating the SLF - what do you know about it comments but hey)

The technology may be available to do wonderful things with data trasmission which may be extremely useful. There is no doubt about that. There are concerns about security/bandwidth accessibility/signal conflict etc and all are valid, but a black box tells you what has happened. It is a totally retrospective instrument.

Bandwidth, etc does not come cheap and in pure cost terms alone may I suggest that the cost of developing an all singing all dancing black box may be offset by investing in technology which may prevent an accident in the first place?
mercurydancer is offline  
Old 10th Nov 2009, 21:19
  #48 (permalink)  
 
Join Date: Aug 2009
Location: Germany
Age: 67
Posts: 1,777
Likes: 0
Received 0 Likes on 0 Posts
Hi,

all singing all dancing black box may be offset by investing in technology which may prevent an accident in the first place?
Yes indeed
BTW the unsinkable Titanic is still waiting to be invented
jcjeant is offline  
Old 10th Nov 2009, 21:25
  #49 (permalink)  
 
Join Date: Mar 2009
Location: us
Age: 63
Posts: 206
Likes: 0
Received 0 Likes on 0 Posts
How does the aircraft predict when it's going to crash and let the black box go? Interesting idea.
Like a car airbag - G forces
vovachan is offline  
Old 10th Nov 2009, 23:05
  #50 (permalink)  
 
Join Date: Jan 2005
Location: France
Posts: 2,315
Likes: 0
Received 0 Likes on 0 Posts
Originally Posted by vovachan
Like a car airbag - G forces
Could you use your imagination for a moment, rather than try to find a solution to one extremely rare accident??

On the one hand... we know nothing really about what brought AF447 down, but the G-forces at 40,000ft plus at the top of a thunderstorm can be enough to start a break-up of an aircraft. Do you want to eject the FDR at that moment, before the rest of the sequence? No, obviously not.

On the other extreme, you ditch relatively cleanly, then the aircraft sinks. What G-force are you going to select as a trigger?

If there was a "simple" solution it would long since have been implemented.

CJ
ChristiaanJ is offline  
Old 11th Nov 2009, 00:01
  #51 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 76
Posts: 2,484
Received 0 Likes on 0 Posts
vovachan;
Like a car airbag - G forces
How would that have worked in the AF447 accident? (I am assuming for a moment that the BEA Interim Report has it right and that the aircraft, intact, struck the sea in a relatively flat attitude with little forward speed, but fairly high vertical speed. Regardless, the question still applies to such a scenario.)

Regarding deployment based upon 'g' loads and air bag deployment sensing and times, let's take a look at some numbers.

I don't know the design metrics for airbag systems in cars but I think I can offer some reasonable parameters which others with the actual engineering knowledge can modify if my estimates are way off.

I theorize that vehicle airbag deployment design would reasonably contemplate 'g' forces which would spike over time periods of less than one half a second which obtain in collisions of the order of 15fps to possibly 200fps. As an aside, I suspect that for most vehicles, a collision at 200fps would not likely be survivable given the compromising of the passenger compartment but we're talking about an airbag system response time as it may apply to deploying a DFDR and CVR.

Takeoff and approach speeds of most transport category aircraft are in the neighbourhood of 130 to 180kts or between 200 and 300 fps. As another aside, nerve impulse speeds can reach about 160fps so the brain would not likely perceive any unfolding event which occurred at higher speeds.

Speeds below 10,000ft MSL are restricted to 250kts with exceptions to this speed limit on departure; that is about 420fps. That's 250kts IAS, so the TAS could be as high as 300kts, or 500fps. Above 10,000ft or FL100, climb, cruise and descent speeds are in the neighbourhood of 450kts TAS, give or take, which is about 750fps.

At the above-described transport speeds, even on the approach, most aircraft travel their entire fuselage length in slightly less than a second, and at climb, cruise and descent speeds, in about a fifth of a second.

By now you will see where I am going with this... The technical requirements to deploy a DFDR and CVR in the time frames described would require substantial design and manufacturing engineering not to mention certification and regulatory standards which must be tested and enforced. If we think about the "how to", the deploying mechanism must sever both recorders cleanly from their mounts and harnesses and fire them away from a structure which would literally be in the middle of an accident, in the time described, (sensing, arming, firing all in, say, a tenth of a second, so as to clear the failing structures). There is the weight penalty to consider as well - fuel burn increases by about 4% of the added zero fuel weight per flight hour - over the life of the airframe, that is a substantial amount of increased burn for what may be little practical return.

I return to my point made in post #36, which offers the notion that the expense and resources required to do this likely do not meet the risk-reward "thought process" let alone a formal analysis. The number of safety initiatives which must compete for industry and government-limited resources would place this kind of solution well down the list I should think. This isn't to say it isn't doable or isn't worth examining but the realities of doing this must be taken into account.

I'm not saying it isn't worth it. What I'm outlining here are some of the considerations which must be addressed when contemplating such a system. I will have missed some things because I am not an engineer or a mathematician but I do know about airplanes and accidents and flight data analysis solutions and off the top of my head these would be some but not all of the considerations.

The other deployment parameters which would continually assess the "health" of the entire system, (airplane, airplane systems, environment, 'g' loadings) and which could possibly be compared to algorithms built to determine when the aircraft is likely in distress, could form an alternative to the notion of "airbags and 'g' forces". Again, this "solution" must compete for limited resources.

Of course, no two accidents are the same; how do we design a system which is "sure it's time to deploy the recorders", without missing something critical?

Let's examine a practical case: - Investigators could only guess what happened in the last six minutes of SW111 because the fire on board destroyed the main electrical distribution system. The batteries likely continued to supply power however. The blunt reality today is, we dont' even have flight data and voice recorders that continue to work when the aircraft's main electrical system fails. Even getting that fixed, so that the recorders continue to function during serious electrical disruptions providing valuable last-second flight data which may hold the key to the entire accident, is a monumental political task and THAT solution is, and should be, already many positions ahead of "deploying recorders".

PJ2

Last edited by PJ2; 11th Nov 2009 at 00:13.
PJ2 is offline  
Old 11th Nov 2009, 00:35
  #52 (permalink)  
 
Join Date: Feb 2009
Location: Virginia
Posts: 2,058
Received 28 Likes on 22 Posts
I'm not convinced a jettisonable recorder is worth the expense and effort. But it seems to me the FBW computer would know when the aircraft was no longer recoverable, or at least when it was no longer in controlled flight. It might activate prematurely on occasion, but data on what caused the original upset would be a lot better than no data at all. It shouldn't be that hard to set up the recorder to also jettison if the computer stopped working altogether.

Or you could just mount the recorder in Monty Python's tomato.
Chu Chu is offline  
Old 11th Nov 2009, 03:37
  #53 (permalink)  
PJ2
 
Join Date: Mar 2003
Location: BC
Age: 76
Posts: 2,484
Received 0 Likes on 0 Posts
Or you could just mount the recorder in Monty Python's tomato.
(Sigh.)

Why bother.
PJ2 is offline  
Old 11th Nov 2009, 04:23
  #54 (permalink)  
 
Join Date: Jan 2008
Location: above it all
Posts: 367
Received 0 Likes on 0 Posts
Canīt remember the specifics now, but the FAA has new rules on flight data recorders now - and the 787 will have an Enhanced Airborne Flight Recorder which, at least, is supposed to have an independent power supply:

GE - GE Begins Delivery of Flight Recorders for Boeing 787
Finn47 is offline  
Old 11th Nov 2009, 08:06
  #55 (permalink)  
 
Join Date: Jan 2008
Location: Denver, CO
Posts: 265
Likes: 0
Received 0 Likes on 0 Posts
It would an incridible strike of luck and a more than welcome one.
Perhaps. Finding the recorders would certainly be welcome in many areas, but should they provide "ugly" information, not so much in other quarters.


Who is paying the cost of the extended (new) search?
Rumor has it, the NTSB has kicked in a few bucks as well. Many US airlines have a vested interest in this, as it will help them assure that many of the AB-I planes in their fleets are, in fact, safe.


The prototype Concordes (well at least G-AXDN) had a jettesonable flight recorder. Why hasn't that idea been developed?
There are several problems with the recorder leaving the frame - not the least of which is finding it after it has. Some military airframes have a "deploy" switch for recorder and "pinger" packs, but these are used mostly in the hopes of retrieving crew, not flight data. And, they're not 100% reliable, as even a crew-triggered device can land far away from the actual site.


It's still an elaborate solution for an extremely rare problem.
Most assuredly - most flights don't crash. The sticky wicket is determining why those which do did so.


Here's a thought - the pilot community considers downloading of FDR data to be some sort of invasion of their right to privacy while carrying forth the duties which they have agreed to discharge, and unions argue that the information could be used to "rate" pilots at their "desks."

However, a running "concordia" of a plane's performance and how it has been flown might be very helpful in the event recorders cannot be recovered, or recovered in a timely fashion.

Do we know for sure that pitot changes have fixed the problem?

Sort of, but not really.

Were any specific problems with the airframe responsible?

Don't know.

Seems to me that one could "sequester" FDR data for an indefinite period of time, unless the data was needed for the investigation regarding an incident or an accident.

But having the last few flights' worth of data could prove valuable.

At this point, who's to say that particular tail number didn't experience an odd combination of rougher-than-normal landings, well-meaning but less than mfg maintenance on a few items, and rougher than normal handling of cargo hatches, et cetera?

We've been looking for a way for all the "holes" to line up, but, perhaps, we haven't been looking far enough back into the frame's history - and there is nothing to clue us in.

A bump here, a scratch there, throw it all into a magnificent storm, and see what happens...
rottenray is offline  
Old 11th Nov 2009, 14:12
  #56 (permalink)  
Pegase Driver
 
Join Date: May 1997
Location: Europe
Age: 73
Posts: 3,657
Likes: 0
Received 0 Likes on 0 Posts
Retrieving FDR/CVRs in deep waters has always been difficult. If the orography of the sea bed is what has been reported ( the " Alps" with 4000m deep crevasses) , then the task will be even more difficult. Pingers or no pinger.

Incidentally I was reading yesterday the (excellent) interviews of both the pilot of the Cessna 210 that ditched with 6 POB off the coast of Corsica on 12 October last , and that of the Puma helicopter Pilot who finally found them .
(the C210 pilot was an AF 777 Co and his partner a Flight attendant with some experience ) The C210 was on radio contact until 1min prior ditching, , called Mayday, was under radar contact on SSR transponder, and had a new 406 Mhz ELT coupled with GPS. Rescue SAR knew exactly where the a/c went down. A first helicopter went above that point only 15 min later but failed to spot the survivors.,despite all on orange life vests , but drifting. This was 1400 Loc , full daylight. 3 rescue helicopters and a Breguet Atlantic joined the rescue. They never spotted them , finally when night fell 5 hours later, the C210 pilot activated the small light on his life jacket and that was picked up by an infra red camera of the military helicopter , who got them all 6 out.
Hypothermia, small bruises but all OK.
So even with all the new gear , ELT and all, at sea, it seems extremely difficult to retrieve something floating ,even knowing the exact GPS position 15 min before. let alone FDRs/CVRS at 4000m deep in an area 30 x30 NM months afterward...
ATC Watcher is offline  
Old 11th Nov 2009, 17:37
  #57 (permalink)  
 
Join Date: Mar 2009
Location: us
Age: 63
Posts: 206
Likes: 0
Received 0 Likes on 0 Posts
Some additional comments - an airbag reacts in abt 1/25th of a second to a crash. Now let's assume the airliner slams into the sea at full cruise speed of 800 km\h which is waaay too fast. In 1/25th of a second it will have moved 8m or 24 ft if my calculations are correct. Most planes out there are considerably longer than that giving an ejectable black box in the tail plenty of time to eject to safety.

PS If you look at the condition AF447 debris was found in, some of the cabinets floating out there were almost intact, and they look no sturdier than IKEA furniture, so my guess is the terminal speed at impact wasn't that great.
vovachan is offline  
Old 11th Nov 2009, 20:26
  #58 (permalink)  
 
Join Date: Aug 2009
Location: Riga
Posts: 1
Likes: 0
Received 0 Likes on 0 Posts
just an idea

Another option - after one month still send the signal, but only once a minute or even hour, so bateries would last MUCH longer. And also - set pulse frequency depending on presure/depth, seems not too hard technically, seems that for underwater search knowing exact depth can be very important clue.Or it is already considered and already working this way?
megla is offline  
Old 11th Nov 2009, 21:15
  #59 (permalink)  
 
Join Date: Jul 2009
Location: Surrey, UK
Posts: 130
Likes: 0
Received 0 Likes on 0 Posts
Yes, you could add a depth sensor for not much money in overall cost of a commercial passenger aircraft (maybe USD10,000) BUT, and this is the big BUT.... you still need a direct "line of sight" to the FDR to be able to pick the signal up. If it's stuck in an undersea crevass etc, you will never detect it (see earlier posts).

Regarding recent new point raised: I used to work for a company that made automotive crash sensors. Typically these trigger at 14 to 16g. They had to withstand 8g without triggering; every one we made was tested for performance.

Edit: Thinking about it, the RovHomer calculates distance and angle IIRC, so it automatically calculates the FDR depth (or the control software does)
HamishMcBush is offline  
Old 11th Nov 2009, 21:31
  #60 (permalink)  
 
Join Date: Oct 2000
Posts: 570
Likes: 0
Received 1 Like on 1 Post
Responder Mode for Black Boxes

djp,
Very sensible & as so often happens a very simple idea, well done. Now let's see who can be the first to come up with just such a device.
Hope the search does re commence & continue for as long as it takes, yes it's a long shot, but surely worth the effort.
kaikohe76 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.