Better Black Boxes on Airliners

[Spyman]

What land?

Malaysia:

 

Route: Kuala Lumpur - Beijing. Inserted: initial search areas and known path near a navigational waypoint called Igari, Vampi, and Igrex. Small red squares: radar contacts. Small circles: claimed spotting of debris.

http://en.wikipedia.org/wiki/Mh370#Disappearance

 

Estimated route

 

On 11 March, it was reported that military radar indicated the aircraft had turned west and continued flying for 70 minutes before disappearing off the Malaysian radar near Pulau Perak, and that it was tracked flying at a lower altitude across Malaysia to the Malacca Strait. This location was approximately 500 kilometres (310 mi) from its last contact with civilian radar. The next day, the Royal Malaysian Air Force chief denied the report. A Vietnamese transport minister said Malaysia was informed on the day that they "noticed the flight turned back west but Malaysia did not respond."

 

US experts, assigned to assist with the investigation in a low-key manner consistent with conventions of responsibilities, analysed the radar data and subsequently reported that the radar data did indeed indicate that the aircraft had headed west back across the Malay Peninsula, with Reuters and The New York Times saying that the route changes suggested that the aircraft remained under a trained pilot's control. The New York Times also said the aircraft experienced significant changes in altitude.

http://en.wikipedia.org/wiki/Mh370#Estimated_route

[John Cuthber]

Thanks for that.

It's entirely possible that the military did see it, but they recognised it as a lost civilian jet so they didn't do anything about it.

Odd, but possible.

A spot of googling suggests the the plane's track was over two areas that aren't very wealthy so it's possible the authorities just didn't care much.

[faslan]

i'm also thought a better black box. black box with some signal emitting system. so we can locate where is the plane

[Delbert]

i'm also thought a better black box. black box with some signal emitting system. so we can locate where is the plane

Doubtless there'll be a number of recommendations and changes following this sad event.

 

I understand a cyber attack hasn't been eliminated. Perhaps infiltrating control software of on-board computers not like the one that reportedly supposedly damaged centrifuges in a middle eastern country a while ago. Apparently a cleverly designed computer virus wandered about the cyber world until by chance it found itself in the control software of said centrifuges. Following which it set to work and at an appropriate time destroyed said devices by speeding them up.

[John Cuthber]

Doubtless there'll be a number of recommendations and changes following this sad event.

 

I understand a cyber attack hasn't been eliminated. Perhaps infiltrating control software of on-board computers not like the one that reportedly supposedly damaged centrifuges in a middle eastern country a while ago. Apparently a cleverly designed computer virus wandered about the cyber world until by chance it found itself in the control software of said centrifuges. Following which it set to work and at an appropriate time destroyed said devices by speeding them up.

As far as I can tell, an attack by flying (radar transparent) unicorns hasn't been eliminated.

 

But I don't plan to speculate on what colour they were or how they orchestrated the attack.

In fact, until someone comes up with some convincing reason to think otherwise, I'm going to assume that the crash was due to a combination of cock ups.

[Delbert]

In fact, until someone comes up with some convincing reason to think otherwise, I'm going to assume that the crash was due to a combination of cock ups.

I think that's guaranteed since it is designed and built by humankind, operated under a manmade protocol with a human at the controls.

 

And as for the cyber attack you appear to associate with unicorns, I suggest that's a human decision without evidence; which would undoubtedly end up as a cock-up if it does turn out to be a cyber attack.

 

Anyway, this appears to be a link reporting on the referred cyber attack: http://en.wikipedia.org/wiki/Stuxnet . So I don't think the possibility of similar attack on aircraft is on a par with mythical creatures.

[John Cuthber]

What I'm saying is that we have no actual evidence so the unicorn guess is as good as the cyber attack guess.

Both are baseless.

[Spyman]

What I'm saying is that we have no actual evidence so the unicorn guess is as good as the cyber attack guess.

Both are baseless.

While I agree that we don't have any indication of a cyber attack and that it's a very unlikely scenario, I still think that technically there exists a possibility to influence airplane systems and cause them to crash. However "flying (radar transparent) unicorns" doesn't even have any evidence of existance, so these two guesses are far from equally good.

 

(IF we had evidence of "flying (radar transparent) unicorns" then the guesses would have been equally good and baseless.)

 

The Federal Aviation Administration has special condition rules for Boeing Model 777-200, -300, and -300ER series airplanes, which shows that there is real concern about vulnerabilities, so it's not some random conspiracy theory:

 

The integrated network configurations in the Boeing Model 777-200, -300, and -300ER series airplanes may enable increased connectivity with external network sources and will have more interconnected networks and systems, such as passenger entertainment and information services than previous airplane models. This may enable the exploitation of network security vulnerabilities and increased risks potentially resulting in unsafe conditions for the airplanes and occupants. This potential exploitation of security vulnerabilities may result in intentional or unintentional destruction, disruption, degradation, or exploitation of data and systems critical to the safety and maintenance of the airplane. The existing regulations and guidance material did not anticipate these types of system architectures. Furthermore, 14 CFR regulations and current system safety assessment policy and techniques do not address potential security vulnerabilities which could be exploited by unauthorized access to airplane networks and servers. Therefore, these special conditions are being issued to ensure that the security (i.e., confidentiality, integrity, and availability) of airplane systems is not compromised by unauthorized wired or wireless electronic connections between the airplane information services domain, aircraft control domain, and the passenger entertainment services.

https://www.federalregister.gov/articles/2013/11/18/2013-27343/special-conditions-boeing-model-777-200--300-and--300er-series-airplanes-aircraft-electronic-system

Edited April 4, 2014 by Spyman

[John Cuthber]

All the "unexplained" crashes are, of course, produced by unicorns.

 

What we have is evidence that cybercrime is possible.

What we don't have is any evidence that it happened in this case.

 

Unicorn induced failure might have been a poor example, but the point remains that while cybercrime is one possibility, there are plenty of others which are equally plausible.

 

A discussion of cyber attack would be premature until we have evidence that can rule it in or out.

[Delbert]

What I'm saying is that we have no actual evidence so the unicorn guess is as good as the cyber attack guess.

Both are baseless.

I'm sorry, but I don't know that you're on about. All I said was a cyber attack hasn't been eliminated, like all the other scenarios until they can be eliminated or alternatively identified as a cause. But there's no words available to describe the ridiculousness of including a unicorn scenario.

[davidivad]

All the "unexplained" crashes are, of course, produced by unicorns.

 

What we have is evidence that cybercrime is possible.

What we don't have is any evidence that it happened in this case.

 

Unicorn induced failure might have been a poor example, but the point remains that while cybercrime is one possibility, there are plenty of others which are equally plausible.

 

A discussion of cyber attack would be premature until we have evidence that can rule it in or out.

 

i would like to point out the fallacy that is portrayed here.

the inability to logically debate different scenarios is important in discussing possibilities.

any discussion about anything is premature if the reasoning is not known yet.

this type of argument is not much different than me telling you that unicorns are scientifically known not to exist therefore do not belong on the forum.

Edited April 4, 2014 by davidivad

[Enthalpy]

One acoustic locator that presently searches for the blackboxes' pingers of MH370 is said to detect them from about 6km range, which takes time at boat speed.

A hypothetical alternative: drop listening buoys from a plane, each with 6km range and evenly spaced, which record the sound and later radio it back to the plane. That was done decades ago and looks useful with the small range of a pinger.

The searched area in the Indian Ocean is 64,000 sq nm (1nm = 1852m). One doesn't need to drop one buoy for each 20 sq nm, because if a buoy drifts at 1 knt in the current or the wind, it covers 120 sq nm in 1 day, so it takes only 533 buoys.

 

Each must have hydrophones and signal-processing capability. I imagine the acoustic part hangs deep under the float. The transmission could perhaps look like Gsm - transmission to a plane has a longer range than on the ground. The buoy should best weigh very few kg, with a float as a big bottle. Depending on the search area, the cable could be >1nm long, since cold water and silence help. The plane needs an automatic dispenser. It's not cheap development, but I see nothing expensive in the series production, and the search for MH370 has already cost over usd 50M
http://www.smh.com.au/national/most-expensive-aviation-search-53-million-to-find-flight-mh370-20140404-36463.html
And, so sorry for that, but the development takes more than one or two weeks.

If developed without time pressure, the floats would rather be tiny sailboats with autonomous steering, to keep aligned and properly spaced, and head to a desired direction.

Marc Schaefer, aka Enthalpy

[Cap'n Refsmmat]

Sonobuoys are already commonly carried by military antisubmarine warfare aircraft:

 

https://en.wikipedia.org/wiki/Sonobuoy

 

The question is whether they listen on the right frequencies. A P-3 could cover a wide area with sonobuoys faster than a ship could trail a locator device, but perhaps the ship-mounted locator is more sensitive.

[Enthalpy]

Yes. Sonobuoys were old technology when I developed a helicopter-borne sonar two decades ago, but in this search they seem to fit the task better.

 

I suppose that existing buoys listen to much lower frequencies, because 37.5kHz propagates badly in the Ocean. Active sonars work at very few kHz or less, passive ones even target boat screws (probably their flow noise).

 

As for sensitivity: this data is hard to grasp... Ships tow a winged sonar, far from the screw and hull, and deep in cold water away from the noisy surface. A towed sonar has the advantage of length and possibly width, that permits to filter out noise from unwanted directions. But an immobile buoy has less flow noises, has no towing ship, and can be kilometer deep. My gut feeling favours the buoy, by far.

 

It's also much a matter of signal processing. If the buoys were developed 30 years ago as expendable items, they don't have subtle software.

[John Cuthber]

i'm also thought a better black box. black box with some signal emitting system. so we can locate where is the plane

They do, and it has been detected if this report is correct

http://www.foxnews.com/world/2014/04/05/china-ship-detects-pulse-signal-in-search-for-missing-malysia-airlines/

 

Still no evidence (yet) of a cyber attack or unicorns being involved.

[Enthalpy]

If the absolutely thinkable cyber attack did occur, be 100% confident that Boeing and its governments will do everything they can to hide it. Even misleading the search if they feel it necessary.

 

To be clear: Airbus' governments will do the same, if Airbus as well was stupid enough to merge all data networks on its future planes. After the fires on the Dreamliner, Airbus realized they had the same wrongly chosen lithium battery chemistry on the A380, so propagandists tried to explain that only the battery size on the Dreamliner is dangerous.

[Enthalpy]

The best methode to locate a pinger may rely on Doppler effect.

 

A ship sailing at 10m/s (20knt) observes 0.67% Doppler shift on sound, far from the source. Provided that it hears the ping at that speed, and so on.

 

The operation is then: sail at constant speed, record the ship's position and the received frequency against time. Measure the received frequency long before and long after passing at the maximum amplitude. Compute the mean of both frequencies, and deduce where the boat passed through thes zero-shift received frequency. This is the minimum distance, more accurate than the amplitude would tell.

 

Then, take a new route to pass over the previous point, but at 90° from the previous route. Repeat the recording and position estimate. The two minimum distance lines intersect at a point, the best position.

 

You may repeat the first pass, this time right over the already estimated position.

 

 

Well, there are many conditions for that method to work...

• The emitted frequency must be stable enough - if the pinger oscillator goes on when the piezo doesn't emit, then far better than 0.1% minute-term stability results from any RC oscillator.
• The received signal must be clean enough to measure a frequency.
• The ping duration must suffice to measure the frequency accurately. But if the oscillator goes on between ping emission, one could correlate the phases from one ping to an other. 0.67% shift over 37.5kHz already makes 25 cycles over 0.1s.
• Multipath propagation would be a worry.
• And of course, it takes special apparatus, or possibly software. Nothing complicated, but time runs out.

Marc Schaefer, aka Enthalpy

Edited April 7, 2014 by Enthalpy

[Enthalpy]

After a Chinese ship heard pings at 37.5kHz, so did an Australian ship yesterday, and much louder, but at locations very distant from the first one, and today it heard none at that place.

 

I wonder. The pings are extremely caracteristic, and yesterday they were heard for a very long time, needing margin on the signal strength. It has been proposed that today the batteries are empty. Though, I read previously that the signal strength would first decay over several days. How could both pingers run out of power at the same night, and so abruptly? And since losing the black boxes hundreds of miles apart isn't believeable, could one ship mistake a noise for the ping, which has an accurate and very uncommon frequency, at a regular 1s repeat beat? That's obscure to me.

 

----------

 

I've also read some proposals to improve plane wreckage recovery at sea in the future. Some even propose specialized radar satellites for that. And, despite understanding the importance for the relatives, my position is that one billion dollars are completely out of proportion with the usefulness of finding a wreckage and explaining a crash. With very few millions, you would equip avalanche rescue helicopters with radars that see the victims through the snow, and save living persons.

 

----------

 

Anyway, there would be cheaper methods to locate black boxes more easily at sea, the first one being a sound frequency better chosen than 37.5kHz. For instance here, you find the absorption versus frequency (and depth) in the Ocean:

http://resource.npl.co.uk/acoustics/techguides/seaabsorption/

At 37.5kHz its about 65dB over 10km distance at 3,000 depth (and impractical 130dB over 20km distance)

At 3kHz it's 15dB over 100km distance

that's why sonars and mammals use low frequencies. A submarine hull reflects less than 1mW active sonar power (less than a pinger can emit) but is detected very (very) far away under neutral conditions.

 

3kHz would also be audible by humans (almost the highest note on a Piccolo flute) without technology or just with a stethoscope, which I consider a big advantage. In an emergency situation, no-tech or ubiquituous low-tech is better. And with technology: that's a frequency military sonars would hear, and one ship-based helicopter would cover the search area quickly. My first estimates indicate that compact transducers are feasible.

 

As a note, one shouldn't evaluate a sound range at sea from open-space propagation and losses. Sound velocity increases at depth because pressure makes water stiffer, and the resulting refraction reduces a signal emitted from the Ocean floor. Temperature differences at water layers also create refraction. This explains why towed hydrophones are flown deeply.

 

----------

 

In an other thread I suggest planes and helicopters with hydrogen tanks and fuel cells. For the present search far from the coast, the several days flight time would be an advantage. Have more pilots on board, or a remote control - or build them as drones, since both Sonar and optical survey is an easy task for them.

Edited April 8, 2014 by Enthalpy

[Delbert]

If the absolutely thinkable cyber attack did occur, be 100% confident that Boeing and its governments will do everything they can to hide it. Even misleading the search if they feel it necessary.

Yes. Although I'm not too sure about misleading the search.

To be clear: Airbus' governments will do the same, if Airbus as well was stupid enough to merge all data networks on its future planes. After the fires on the Dreamliner, Airbus realized they had the same wrongly chosen lithium battery chemistry on the A380, so propagandists tried to explain that only the battery size on the Dreamliner is dangerous.

As for the battery business, I postulated my idea elsewhere on this site that there's a fundamental flaw in all batteries. But it was apparently dismissed some others - but as far as I could tell their objections never managed to contradict or counter my postulate.

 

And from what I can understand Boeing haven't resolved the battery problem. Placing it in a steel box is not a resolution.

[Enthalpy]

I'm not sure at all that the search is mislead - and having no reason to believe so, I prefer to believe it's done properly by everyone. What I'm sure about is that if a company or government believes that showing a design flaw on the plane hampers the sales, then they will easily leave the wreckage and bodies where they are to favour the sales. As the Airbus 320 got its infancy crashes, investigations were deliberately hampered and mislead.

 

Lithium batteries: some are reasonably safe but heavier. On the Dreamliner, designers and supervisors at the company and certification agencies took a chemistry that had already caught fire at laptops. That was an error to my opinion, by successive people, but mass constraints are just brutal at an aeroplane.

 

Once these batteries caught fire as they just do sometimes, designers couldn't replace them by a heavier chemistry on the Dreamliner, which relies on huge electricity storage. They put a box, which is an answer, if not the ideal one.

 

What is unthinkable is that Airbus took the same chemistry for the A380 and, instead of replacing it - as the A380 needs less electricity storage - they tried to explain that the size of the batteries, not the chemistry, made them dangerous on the Dreamliner. To my knowledge, the A380 hasn't even added boxes around its batteries. That's bad behaviour.

 

----------

 

More pings heard today, supposedly from MH370's black boxes. Fainter, but underwater propagation isn't reproducible. According to the Beeb

http://www.bbc.com/news/world-asia-26950387

sonobuoys will be dropped in the more accurate zone, with hydrophones hanging 300m below the surface.

[Enthalpy]

Some geologists took advantage of the search for MH370 wreckage to encourage the complete mapping of the Ocean floor by sonar, for a few billion bucks.

http://www.bbc.com/news/science-environment-26956798

 

Of course, any very hypothetical help in this search from better maps isn't worth the money. For other purposes, I don't know.

 

But if someone was to make this exhaustive mapping, he should reconsider the method. Sending surface boats and crews for years costing billions, there must be better methods.

 

One first alternative to consider would be, still by sonar, a swarm of aquatic drones. If sailing in close group, they could get a better resolution, just as the MiG-31 does with its radar - provided sound propagation doesn't blurr it. Submarine drones could be advantageous: no separate towed sonar, and unmanned ones would easily dive to 800m or more. Electric engines are more silent; range may need to surface and charge batteries - but consider fuel cells, which give present German U-Boots over one week underwater operation.

[Enthalpy]

A correlation receiver has two or more antennas (could be hydrophones, microphones...) followed by separate preamplifiers, mixers, amplifiers, converters. It computes many time-shifted products of the signals caught by the antennas and integrate them over a long time. The time shifts correspond to different listening directions, and the integration time accumulates the desired signal, which is correlated at the antennas, but less so the uncorrelated noise, improving the sensitivity.

I suppose one can compute products or something over more than two antennas; ask a radio-astronomer. With more hydrophones, irregular patterns reduce the sidelobes; for instance the VLT (that's interferometry, but it holds for correlation) has no square baseline. Hydrophones at varied depths would discriminate noise from the surface.

This integration gain is paramount in radio-astronomy, whose signals are wideband random noises remarkable only by their provenance and correlation. A correlation receiver would have advantages as well with a narrow-band pinger:

• Reduce the local aquatic noise, including flow noise.
• Reduce the noise created by the hydrophones and preamplifiers.
• Integrate many pings.
• Reduce the background noise not in the listened direction.
• Identify the pinger's direction.
• Discriminate multiple propagation paths.

If existing pingers at 37.5kHz have a ceramic buzzer and an oscillator powered on and off, the phase of successive pings is random, and a correlation receiver improves over a simple narrow-band (Fourier) receiver - but keep the narrow band. Same if the aquatic propagation introduces jitter whose difference is smaller between the paths to the hydrophones.

How far apart should the hydrophones be? Directivity would accept very close hydrophones, but distant ones discriminate local noises better. The limit depends on how much and quickly the propagation conditions to the hydrophones differ, and should be experimented.

I suppose different correlation receivers exist already at sonars, which correlate noise sources at the Ocean's surface with their echo at the silent target, but here we search an emitting target. Anyway, much knowledge and know-how is common.

The hydrophones are expected to move over the Ocean floor: even 1m/s with 10m baseline and 3000m depth would shift a nadir signal after only 6 pings, so the software should identify the speed and compensate its effect over time for various depths and directions.

The swarm of aquatic or subaquatic drones can use a correlation receiver, either when moving or by stopping when they listen (future muscles and a tail will be more silent than a propeller). A helicopter, quadrocopter or multicopter drone looks attractive to carry quickly a wide frame of hydrophones and dip it low (800m+) from place to place: fuel cells would give it range before it refuels on the continent or on its boat. Folding or retracting a bigger frame before landing would be a funny endeavour for mechanical designers.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Here's my black box pinger design (click for full size), meant to radiate isotropic 40mW at 3kHz for long range, and from 8km water depth to include all Ocean floor.

 

 

The hull and resonators are of Ti-Al6V4 treated for 1030MPa yield, 3mm thick at the caps, 4mm at the cylinder. A long cylinder would need 5.5mm to resist oval buckling, and a short cylinder as well may need a few azimuthal stiffening rings - theory fails, so check Nasa's Sp-8007 instead:
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19690013955.pdf (thanks!)

Vibration is small at the seals; if these damp too much, one may try rings of unalloyed titanium, squeezed at assembly and used once, or reduce the vibration by a longer cylinder and sturdy stiffening rings separated from the flange. Air at 1atm inside is fine for the function, but if altitude moves the seals which leak then in deep water, a low pressure and greased rings would help - or grooves narrow in both directions if possible.

The pre-stress at the flanges shouldn't have to exceed the force of an external pressure. Stainless screws in titanium seize horribly (material pairing theories are wrong), but the Bulten company claims to have the proper coating. An variant at smaller pre-stress might be a spring wire, pre-formed in zig-zag, bent elastically for assembly, that passes many times at each part alternately.

--------------------

Materials and shapes that resist the water pressure hardly change their volume to radiate as a monopole, but resonance helps us. Here 0.68µm peak radius oscillation over equivalent 39mm length shall provide 40mWrms acoustic power and need radial 2.7kN peak, essentially as reactive power. Poisson's coefficient (µ=0.31 for Ti-Al6V4) loses volume only as 1-µ/2 if we act at the radius.

The radial amplitude decreases with axial distance as exp(-x/x₀) with (x₀/R)²=G/E=39GPa/104GPa if the walls only shear; add lengthwise stiffeners if the walls bend. I computed no azimuthal stiffener, but added the wave reflected at the flanges for the force and subtracted it for the displacement.

The radiation resistance (Pa*s/m³) of a small source is 0.5*pi*rho*F²/c: at 3kHz in water 9.4Mohm - and in air 50kohm, which still radiates 210µW or 52dBa at 10m, ouch.

Inertia of the resonator provides the big reactive power that deforms the hull, and the actuator the small active power for radiation and losses. Mass at the walls would be too heavy for 3kHz, so I use instead 2*50g connected by beams tilted to tg=0.14.

Speed makes voice coil motors efficient. To claim peak 0.19N from each at 0.21m/s peak, I added compliance at the tilted beams (two sets of eight, each 16mm 30mm²) from 41MN/m by the hull to 18MN/m. This necessarily increases the reactive energy and raises Q=440 to 1000. The thinner beam ends must bend easily.

Ti-Al6V4 dissipates little. The walls, beams and masses are better turned from one piece, and the beams then separated, say by laser cutting. The weld seam must be perfect; consider laser or electron beam. Ageing needs thin sections and is best done after welding, but dissipation may need to de-scale the surfaces. A cage should protect the hull against damping objects.

Unsymmetric masses or stiffnesses keep a single resonance but introduce a lengthwise vibration of the hull. If its fastenings create any damping then, symmetrize.

--------------------

The coils must be stiff; glass fiber between the copper layers should help - check as well the printed coils I describe there
http://www.scienceforums.net/topic/74606-blinds-stick-with-laser-tactile/#entry743275
The neodymium magnets could be hold and wrapped in thin aramide strands and create 0.53T at the D=29mm W=6mm L=5mm gap. The coils drift by 1.6mm through outer 80MPa. For peak 5.6A*turn, they dissipate 2mW rms each.

Laminar viscous losses are P=S*F*pi*rho*V²*e where e²=eta/(pi*F*rho) with eta in Pa*s. At 3kHz, 1m² and 1m/s, this amounts to 39W for water and 0.12W for air. They are around 200µW in air here and tiny in water.

The design is flexible, and piezoelectric or magnetostrictive actuators would find a matching impedance at some position of the beams. A clapper can also ring the masses, beams and hull like a bell; the compressive and flexural stiffnesses and the masses are easily tweaked for increased Q then; consider vacuum.

100g of Li-MnO₂ would provide 22Wh, enough for 3 months of 50mWe pulses with 0.4+1.6s cycles. The powered pinger weighs ~1kg. As the hull for 100g batteries adds 200g, a magnesium-seawater battery is even more advantageous.

Marc Schaefer, aka Enthalpy

[Delbert]

Here's my black box pinger design (click for full size), meant to radiate isotropic 40mW at 3kHz for long range, and from 8km water depth to include all Ocean floor.

Unless I'm way off the mark and completely misunderstood the calculations, it seems that a 20 amp/hour battery would only last 21 days. And if my car battery is anything to go by, a 20aH battery is a fair size and quite heavy.

[Enthalpy]

Where did you pick 20Ah? (Ah, not aH, amp/hour even less so).

Watt=amp*volt.

Lithium is not lead.

 

22Wh / 50mW = 440h. 20% duty cycle make 2200h or 92 days.