Tuesday, May 3, 2011

Air France Find Could Alter the Future of the Black Box

How excited do you think these guys are viewing the most significant break-through in the investigation into the crash of Air France Flight 447? On the monitors, accident investigators are watching the end of a very long and very expensive phase; recovering the black boxes from the plane that crashed while en route to Paris from Rio de Janeiro in June 2009.  Two hundred and twenty eight people were killed in the mishap.

Over the weekend, the men saw, via remote camera, a robotic arm fetch the flight data unit of the Honeywell flight data recorder off the ocean floor and drop it into a basket where it was then hauled out of the Atlantic. An activity repeated on Monday with the cockpit voice recorder. I'm just guessing here, but I'm thinkin' this monumental development nearly two years after the search began, is being greeted with the popping of champagne corks.

The news from the French Bureau of Investigation and Analysis (let's just call it the BEA) may may be eclipsed in the headlines, this being the same week the world learned that US special forces in Pakistan killed Osama Bin Laden. So one particular airplane crash may pale in comparison to finally "getting" the man who plotted the use of four airliners to attack American landmarks and kill thousands of people. 

While I won't argue with that, I will chime in with this thought. The mysterious crash of Air France Flight 447 will change commercial aviation significantly. (If I didn't loath the cliché, I might call it a "paradigm shift".) But this metamorphosis might not have happened without those years of expensive and frustrating searches for the airplane's black boxes.

When the  flight data capsule produced by Honeywell was finally dredged up from the ocean, BEA investigator Olivier Ferrante, was cautiously optimistic.  "Cautiously" because while the black boxes are poked, soaked, baked in a flame and fired from a cannon to earn the moniker "crash survivable memory unit", the experience of this particular flight recorder has grossly exceeded at least one of the test parameters. Submerged in the Atlantic for 700 days, it has been in the drink, 670 days longer than the salt water submersion survivability test, as described in this March 2010 article in Avionics News.

One month ago, when it appeared likely that the black boxes might be found, Ferrante and another investigator, Arnaud Desjardin,  spoke to the European Society of Air Safety Investigators about what they learned from their pursuit of those elusive devices.  The BEA made several recommendations based on their experience.  

Ferrante "hopeful" at April's safety meeting
Now, in possession of the soggy recorders, Ferrante is hopeful that data can be read, even after being in the ocean for so long. 

"In previous operations, the same type of recorder was damaged but BEA has always been able to read out and analyze this data," he said. "We will try to push the limits to read them."
He must hold on to this hope and he is not alone. After five search efforts and $135 million spent scouring inhospitable terrain and ocean depths, what will be the return on the investment if the data is unusable?

In the days after the crash of Flight 447, I faced what I called the twitter paradox; in which I essentially asked this question, "How can my 16-year old son using his cell phone, know his friends' every move, while multi-million dollar airliners are unable to transmit critical information?" Turns out I was not the only one asking that question. 

At DRS Technologies, executives say before the year is out, they will demonstrate how airliners can compress and encrypt flight information and cockpit communication and transmit this information off the airplane as a continuous stream of data in real time for less money than airlines are now spending on airtime. Money, bandwidth and a lack of necessity are the three persistent arguments I've heard against the feasibility of this development. 

"We’ve looked at the number of aircraft flying at any one time. We've looked at flight profiles as they fly into major hubs. We've looked at congestion of transmitted data and the available satellites and we believe it is possible to do this," Scott Newbold, an executive with a Florida subsidiary of DRS, told me in a phone interview Monday afternoon.  Air France Flight 447 and another fatal crash that same month, Yemenia Airway Flight 626 may address the third argument, "Where's the need?"

I learned about DRS while reporting on the twitter paradox for The New York Times. Then, I was writing about the company's deployable recorder, a nifty product that encases a recorder in an airfoil so that the entire device flies away from the airplane in a crash and floats if it lands on water. It pings its location and the last known position of the aircraft, through the international search and rescue satellite network. It would have saved the French a bundle, had it been on Air France Flight 447. And in the case of the Yemenia A310, which crashed off the coast of Grand Comore in the Cormoros Islands, June 30, 2009, it could have saved lives. A 14-year old on the flight reported hearing the voices of other passengers as she clung to floating wreckage in the Indian Ocean. But when rescuers finally arrived 13 hours later she was the only one found alive. 

"They were 10 miles off the coast, in a normal flight pattern. It wasn’t like they were hundreds of miles off the coast," said Blake van den Heuvel an executive with DRS.

No airline uses deployable recorders, though. "They’re not mandated so typically what you see in civilian aviation, those things that are not mandated don’t get put on aircraft," van den Heuvel said.

Crashes like Yemenia where passengers survive the disaster and perish waiting for rescue, may be rare. Cases where recorders have been difficult to find, not so unusual. Over the past 30 years, there have been 26 accidents in which underwater searches were required to recover CVRs or FDRs.

Arnaud Desjardin speaks of onboard data transmission
That little factoid comes from the BEA, which gets us back to the ever-hopeful and I hope the presently-champagne-quaffing investigators. They obviously believe that black box data is a non-redundant system. There's no backup when the recorders are too hard to find or can never be found, as was the case on three of the airliners flown into buildings on 9/11.

And so while waiting for the day the BEA worried might never come, the investigators started working with DRS, Airbus, satellite manufacturers and dozens of others to come up with new ways to bring airliners into the digital age by giving them the same capabilities as the modern teenager.  An airplane that takes its secrets to the bottom of the ocean is a paradox, yes. It can also be an opportunity to push ahead with what's possible in the future.


Anonymous said...

Not all that is technically possible is reasonable or desirable.

The advantages of sending all data of a flight recorder realtime via satellite somewhere on the ground are relatively small. It will only help in these very rare cases where FDRs are not accessible after one of the (thank god) rare crashes. It will not help to avoid accidents primarily, but instead it will reduce safety on board.
A lot of questions arise with the idea of realtime datastreaming via satellite:

- Who is willing to work in an environment where every move one makes will be recorded and sent to the ground including every word one speaks? This is not only a complete negation of one's personal rights but creates an atmosphere where one cannot move or speak freely.

- The permanent watch under which pilots would have to work will create and absolute contraproductive uneasy feeling of beeing watched, which increases the fear of a pilot to do something wrong. Before a pilot will act he will think five times, wether he will make a "fool out of himself" if he says or does something. If he does a mistake, which is human and happens from time to time unavoidably he will rather be occupyied with the idea if somebody did see this and what consequences it might have, than with what is needed from him right at this moment. This is against all we learned in the last 30 years about CRM, which is absolutely necessary in the cockpit to avoid accidents.

- Anybody could receive the data send to the ground. It would enable a complete supervision by the employer on the ground which is inacceptable to any human being. There would be no privacy at all. Even terrorists could receive these data unless you find a way to protect these datastreams. We still can't make the internet safe yet. How can we guarantee safety of these sensible data?

- Who will store these data and where? Who will have the right to access these data and when?

- The cost will be enourmous and for what? Just to have one or two cases in 50 years where the data could not be retrieved from the FDR because it is sunken in the sea. We might learn a little more about an accident. The gain does not justify the costs which would be enourmous due to the extensive amount of data need to be sent via expensive satellite connections. This would have to be paid by the passengers. If we would take this money and put it where it really is needed instead, we would have a much greater effect in flight safety.

What might look "sexy" on the first hand, looks different once you take a closer look. The idea of realtime streaming is anything else than helpful for flight safety and will be rejected by the pilots all over the world, I am quite sure.

The only one's who are aggressively promoting this idea, are the one's who sell the equipment.

Joerg Handwerg
German Airline Pilots' Association

Anonymous said...

Thanks for this article, Christine. I am not an aviator, but I have watched this drama unfold with great interest for two years. I, too, cannot understand why data is not transmitted, with the CVR and FDR used as back-ups. I think the real reason, as stated in your article, may be that in civil aviation (as in many other profit-driven industries), unless it is government mandated, it doesn't happen.

God knows where we'd be without government-mandated MPG ratings on automobiles.

Jeff said...

I agree with the first anonymous person who said that there is no need to send all routine data via a satellite link. But my reasons are different. I do not share the concerns about pilot fears over the streaming the flight data. There is little difference between doing that and the current method which records the data onto a tape or hard drive for downloading later on.

Most major and mid level airlines around the world are currently engaged in flight data analysis programs where this data is used to improve procedures and training and raise pilot awareness to flight hazards.

That said, I do believe there would be value in having a system that would send real time data in critical situations that could either be triggered automatically by a significant event (such as an engine failure of control difficulty) or manually by the crew by pressing an "event" button when they were experiencing difficulty and in need of assistance.

The only other concern I would have with this is that such real time data would need to be stored in highly secured servers so that they could only be accessed by an investigative authority when it became obvious that the normal on board flight recorders would not be available in a timely manner, such as Air France 447. If the regular on board recorders were available and useable, I would expect the real time data to be erased immediately.

Georges said...

The BEA has published a working group report that talk about data transmission. They have studied the technical feasibility of triggering the transmission of data in case of an emergency. So it would not be permanent data streaming of all FDR data of all aircraft all the time!


FLYHT said...

FLYHT, the company I work for, has developed a real-time emergency data streaming technology called FLYHTStream. We were involved in the BEA working group and provided our solution, which is available and working on aircraft now. There is a whitepaper about the streaming on our website, http://www.flyht.com/index.php/investors/media/ and there are also articles from media.
To address some of the concerns above, the streaming is triggered based on a set of parameters identified by the airline. The streaming can also be started by the operations on the ground or the pilots. The data is sent to a secure location on the ground, determined by the airline.

Dick Newman said...

The week before the two black boxes were found, there was a news item in the Pax River NAS paper about a Navy crash recorder found after five years. The box, which was ejected from the F-18 drifted some 4000 miles and was found by a man walking on the beach. Why don't international airline aircraft carry an ejectable, flotable recorder?

Randi said...

'So one particular airplane crash may pale in comparison to finally "getting" the man who plotted the use of four airliners to attack American landmarks and kill thousands of people.' -- What makes you think Osama Bin Laden had anything to do with the 9/11 attacks? Even assuming he was still alive, and didn't die many years ago as many have claimed. After all, as the FBI's chief of investigative publicity explained: "The FBI has no hard evidence connecting Bin Laden to 9/11."


Christine Negroni said...

The sentence following the one you quote from my post says, "I won't argue that." I'm gonna stick to that position.

Stuart Johnson said...

I should note that I worked in the crash recording business for many years and have seen all these diverse, mostly negative comments before. Of course a pilot union member would include the "Who is willing to work in an environment where every move one makes will be recorded and sent to the ground including every word one speaks? This is not only a complete negation of one's personal rights but creates an atmosphere where one cannot move or speak freely." In reality, the same restrictions that currently apply to CVR and DFDR playback could be applied to streamed data so this is nothing but flak!
Streamed data has many advantages: 1) It is immediately available, an incredible advantage in any crash investigation.
2) My guess is that the overall cost is lower considering how memory costs keep plunging and satellite bandwidth requirements are not a strain on current capacity. This compared to recorder development, qualification, acquisition, and installation costs as well as the recorder installation contribution to aircraft unproductive empty weight.
3) Many airline provide world-wide internet access; quite a useful medium for streaming data!
4) Data could be safeguarded quite readily from people hell-bent on seeing how their pilots perform.
5) Things that make sense should be achieved, and streaming data to improve the accuracy and timing of incident and accident evaluation is a very sensible thing.

Stuart Johnson

Anonymous said...

I find it quite disgusting when so many companies literally rejoice about a plane crash while trying to open the need for a totally useless "new product".
- why useles: because one more time the CVR and FDR were found and are readable... like in almost 100% of the cases. So tell me why again we need such new products?
- "new product": this is the most amusing part: this all technology has been available for well over 20 years now. It's called ACMS (Aircraft condition Monitoring System), part of the DFDAU nowadays (a mandatory equipment on most carriers planes : this is the box that feeds the FDR with the parameters) coupled with ACARS (satellite) and is installed in literally every modern plane today!
Some people use Iridium already for this and tomorrow (see Gogo) you get get several Mb/s of bandwidth anywhere on the planet with new satellites to which a DFDAU can talk to through its Ethernet output...
And regardless, keep in mind that everyone is already going slowly but surely with Iridium and GPS for GPS-based navigation (signed by President Obama: NextGen, all 35 major US airports to be RNP-ready by 2020) - the same is happening in Europe too. So again, where is the need for those so-called new product?

Also, there is already a redundancy for most of those data as a flash card that records even more data in the ACMS system. Funny story again: even though they were never designed to be crash-survivable they almost always survive crash and fire and water!

Tell to me about some GPS sensors that costs pennies today to be installed all over the plane so at least one will always survive to locate the crash site in no time and I may be game.
But please, by all mean do not perpetrate the dirty marketing practices and complete BS (sorry, but there is no other word) of those companies you discussed with and the government agencies using our tax payers money to fund such stupid projects so they can pretend they are actually working and that we need so many people in those agencies. Sic :o(

Finally, as it was noted in another commment, through good fligth data monitoring practices (mandatory as per ICAO for all aircraft with a GTWO > 27T) one learn much more and prevent much more incident / accident that any crash investigation has ever managed to produce when it comes to crew mistake. Hardware failure is another story yet covered by the solution above mentioned.