Loading summary
Transcript1 lines
- [00:02]
A
There are over 5 million route kilometers of optical fiber around the world, both under the sea and ground. And over the past six to eight years, telecoms and tech companies have laid down over half a million kilometers such a broad and growing network of simple, boring infrastructure. Now imagine a new technology that can turn all that into a massive network of listening sensors. In today's video, the recent ongoing rise of distributed acoustic sensing, or das. An optical fiber has at least two components. First, a central core of very pure glass about 10 to a few tens micrometers wide. And second, a thick cladding made from a material with a slightly lower refractive index than that of the core. Propagation depends on a mechanism called internal reflection. A laser, or led, encodes data into a series of pulses and flashes it into the fiber core. As it propagates, the light hits the boundary between core and cladding. If it hits at an angle larger than that of a critical value, which is determined by the difference in the refractive indices of that core, and cladding then reflects entirely back into the core. No loss. Though as the light travels through the fiber, it does suffer a gradual transmission loss. We call it attenuation. Most attenuation is due to scattering. Light collides with atoms or otherwise inside the glass and gets sent off their trajectory. They might either go backwards at the emitter or into the cladding at less than the critical angle where it gets absorbed. Three types of scattering are relevant here. Briluin, Raman and Rayleigh scattering. Let me explain the first two, which are subtle and happen due to factors like temperatures and strain affecting the fiber. Briluin scattering happens due to interactions with low frequency acoustic waves or, or to use a phrase, phonons moving through the glass medium. Such waves often occur from temperature or strain. Raman scattering describes light scatter from molecular vibrations in the glass, that is phonons, but a different type of phonon from that which creates Brilloan scattering. Raman type phonons largely come from thermal energy, which makes it useful for temperature sensing. The discovery of this scattering, by the way, won its namesake Sir C.V. rahman, a physics Nobel in 1930, the first Asian to win a Nobel in the sciences. There's your Asian angle. The third and most significant category is Rayleigh scattering, first described by the very same Rayleigh, whose name adorns the ASML formula. Rayleigh scattering is when light gets scattered by something smaller than its wavelength. Rayleigh scattering does not transfer energy to the instigator, nor does it shift the light's frequency. Most famously, it is known as the reason why the sky is blue inside an optical fiber. Rayleigh scattering happens when the light hits small random density fluctuations inside the fiber's amorphous glass. You can imagine these fluctuations as acting like little mirrors sprinkled across the fiber crossing, reflecting light backwards. Such fluctuations are formed as the molten glass inside the fiber cools down, making Rayleigh scattering an inherent fundamental property of these glass fibers. It is a more prominent effect than the other two. DAS began as a simple diagnosis tool for optical fiber cables. In the 1980s, people laying down and operating subsea fiber cables used something called optical Time Domain reflectometry, or odtr, to scan for breaks or disruptions along an optical fiber line. It worked by first sending a pulse of light down through the cable and then measuring with a photodetector or photodiode the light that gets scattered back at it via Rayleigh. Backscattering defects in the fiber, such as breaks or porous splices, can create localized reflective points or sections of high loss that can be picked up. And by timing, when these reflections or loss drops arrive, we can roughly locate where their points of origin are along the whole cable's length. Pretty helpful, as it can be challenging to locate such a fault otherwise. But an ODTR can also pick up light scattered back at the sender via Rayleigh scattering. In 1981, A.J. rogers of the Central Electricity Research Laboratories in the United Kingdom proposed to take advantage of this with a concept called polarization Optical Time Domain reflectometry. Various external factors like temperature, strain, Sound, electric fields, etc. Can affect the polarization or orientation of light traveling through the cable. This can be leveraged to find and map physical conditions affecting the cable along its length. In a way, it works quite similarly to radar or sonar, except rather than sending probing signals into the sky or water, we send them into a glass fiber. We parse and analyze what comes back polarized. ODTR garnered attention in the 1980s, but ultimately petered out. The problem was that polarization changes happened for various reasons, and it was hard to tie a specific cause for any individual change. Efforts in the 1980s to use Rayleigh scattering for temperature sensing fell short for similar reasons. Since Rayleigh depends on many flecks of fluctuation randomly dispersed throughout the whole fiber. The backscatter profile came back as the sum of all those flecks. It was just too generalized to discern a temperature signal without employing specialized optical fibers. This led to a switchover to Raman and Brilloin type scattering. These two scattering signals are not as strong as Rayleigh. But as I earlier said, they can be more conclusively tied to acoustic vibration, strain and temperature factors. Moreover, along with the scattered light, we had the presence of two things called stokes and antistokes signals. These are scattered photons that had either lost or gained energy and due to impact with a phonon scatterer. When that happens, their frequencies either move up or down. The size of that frequency shift helps figure out the fiber's temperature acoustic strain. So Rayleigh Backscatter remained a simple OTDR diagnosis tool for fiber until the 1990s when vendors finally solved these noise issues with several new developments. First was the arrival of better equipment tighter lined with lasers, allowing us to send through more spectrally purer light. Such items arrived as telecom growth matured the fiber industry in the decade. The second was the addition of more powerful compute capable of sampling tens of thousands of times each second. This allowed users to actually track how the Rayleigh scattering profile might change due to strains in the fiber. The other big motivator was demand. While some preliminary work was done by the US Navy and remained somewhat shrouded, DAS did not find its big break commercially until it was adopted by the oil and gas industry in the mid-1990s. Oil and gas companies adopted fiber optic to monitor their wellbores. Conditions inside these boreholes are harsh with temperatures of some 200 degrees Celsius and and pressures of up to 2000 bar. Not to mention the various corrosive chemicals and heavy vibrations down there. And Rodin. Don't forget him though. Technically, Rodin's emergence in 1956 was out of a collapsed coal mine rather than oil and gas borehole. But whatever electronics need to be packaged and protected against these conditions. Even so, after some time they wear down and break and must then be replaced, which is inconvenient. Fiber optic based systems, on the other hand, do not require us to put any expensive electronics down the hole. The sensitive equipment all sits on the surface. Another advantage is that fiber is small, resilient and does not have to be regularly changed out. The first optical fiber based sensors focused on temperature sensing using Raman scattering as well as anti stokes signals. They're referred to as distributed temperature sensing or dts. Over the years, the technique evolved from taking measurements at single points along the fiber to multiple points and then finally to taking continuous measurements. So long as we know the fiber's relative length inside the bore, we can take thermal measurements along the fiber's whole length. After DTS proved the technology's resilience, the oil and gas Industry slowly began to adopt optical fiber based sensing for a process called vertical seismic profiling, or vsp. During vsp, well operators image the rock surrounding the borehole using sound waves sent down into the ground and listening to what gets reflected. Operators use it to appraise the reservoir and just get a general lay of the rock. The listening part was traditionally done by lowering and installing a string of sensors called geophones. And inside the borehole. These geophones then listen to the vibrations traveling through the bedrock, hearing things that you wouldn't otherwise hear if the listening devices were on the surface. Again, the issue was that the geophones are delicate electronics. It is expensive to harden them against the rough conditions inside the borehole, and they still often break and they can only sit at discrete locations. So to do one seismic survey, the team lowers seven or so geophones strung up together on one steel armored cable. They do the measure and then if necessary, they uninstall the geophones and reposition the line for a second survey. This is inconvenient, unproductive and potentially unsafe. In situations where time really costs money, like on a deep water rig, which are rented at a daily rate. VSPs are too expensive to run at all. But starting in the late 2000s, Shell, BP and other oil and gas companies adopted optical fiber based VSP utilizing Rayleigh scattering to sense acoustic waves. These fibers are often wrapped around the casing. The signal to noise ratios with these optical scans are not as high as that of traditional wired geophone surveys, though it can get pretty close. But the real killer app is that fiber makes it far faster and easier to run a seismic survey. In some cases just minutes as compared to up to 10 hours for a geophone based survey. No need to lower and reposition a string of geophones and you get far more coverage. One early challenge was figuring out how to interpret all the data. A fiber can Send up to 10,000 times more data every second than a regular temperature sensor. And data that size wasn't easy to interpret at first. This form of VSP can be considered the first commercially viable use of DAs. Oil and gas investment throughout the mid 2010s, particularly with the rise of tight oil, matured the DAS technology to such an extent that academics realized that there were far broader applications out there. In the early 2000 and tens, two teams at Berkeley and Stanford began applying DAS to seismic monitoring. Traditionally, earthquake monitoring is done with seismic stations installed on the ground. While they work, they also cost a good amount of money, which makes it very expensive to install in large numbers on the Other hand, there is a massive glut of dark fiber, meaning optical fiber laid down underground but not being used. This might be left over from the telecom bubble, but also be extra fiber laid down by the telecom company for future capacity reasons. While it's waiting for demand, it can be used. Jonathan Aho Franklin at Berkeley Lab estimated in 2019 that there were 1 million kilometers of this dark fiber around the world, with another 100,000 to 200,000 kilometers being added each year by telecoms and tech firms. All estimates because nobody has actual numbers. If made usable for das, then dark fiber can power a broad and comprehensive earthquake monitoring network. Moreover, just like with the BSP situation, you can take measurements in distance increments as small as a few meters. The work was not easy. The two teams detailed some of the problems with using telecom laid dark fiber. Unlike DAS fibers for oil and gas, dark fiber is not affixed to anything. It sits free floating inside a plastic tube. Moreover, earthquakes are very low frequency sounds. It can be challenging to identify them amidst a general cacophony of surrounding noise. Systems have to differentiate between seismic events and things like passing cars and trains. In these systems, there's an interrogation unit and a fiber cable. The unit fires laser pulses into the cable, which gets scattered via Rayleigh backscatter. The unit interprets changes in the backscatter profile to see how the cable's surrounding soil is being strained, either via temperature or sound. One of the researchers, Eileen Martin of Stanford, said in an interview, people didn't believe this would work. They always assumed that an uncoupled optical fiber would generate too much signal noise to be useful. But tasks using both specialized and actual telecom laid dark fiber networks for first at campus, but later in the field were able to track hundreds of small earthquakes, some local, but in other cases as far away as Mexico. Pretty remarkable. At around this time, other parallel work emerged demonstrating similar concepts. One team in Europe, led by Giuseppe Mara at the National Physical Lab in the United Kingdom, notably reported to have used laser interferometry related to das, though not exactly the same, to detect undersea earthquakes, which was significant as there are even fewer earthquake monitoring networks located on the sea bottoms as there are on the land. Mara's network was able to detect a large earthquake in central Italy from all the way in the United Kingdom, again using fiber that was not installed for seismic detection purposes. A few years later, the Berkeley team, led by Nate Lindsey and took advantage of a four day maintenance period for a cable in the Monterey Bay to run full on DAS scans. They turned a 20 kilometer stretch into a string of 10,000 sensors and recorded a minor earthquake, revealing several fault zones on the seabed. It was not very long before people found that they can detect things other than just earthquakes. I mean, the idea's been around all the way back. In 1993, Henry Taylor and Chung Li of Texas A and M patented an intrusion detection security tool that worked by monitoring changes in light backscattered by a Rayleigh scattering. I'm imagine like they bury an optical cable at a perimeter, then if someone steps on it, the Rayleigh backscatter profile changes, which triggers an alarm. Such systems seem to have found a niche in covering sprawling perimeter security zones like airports and high speed rail tracks. One recent interesting use case reported by India Railways is a DAS fiber system to detect elephants walking onto railroads. Darn elephants. But the work that has really caught people's attention has been using DAS for tracking whales, storms and boats. By 2022, DAS techniques have improved enough to detect and track the songs of baleen whales in the open ocean. With just stretches of cable already installed onto the seafloor, the sonic environment on the seafloor is even more challenging than an underground one with all kinds of noise from internal waves, sediment transport and storms. Despite this, they were able to track tens of singing whales as they traverse fjords and open ocean. Very quickly, the techniques for scanning and finding whales improved with filtering techniques and special algorithms like grid search and Bayesian filters. Today we have something roughly like an automated workflow for finding and tracking whales as they sing their way through life. Once you are able to pick out the big citizens, you can track other things across the ocean like storms and undersea landslides, the latter of which can trigger devastating tsunamis or damage offshore infrastructure. In both cases, DAS based monitoring can help people organize faster disaster response or provide early warning and ships. In 2021, a team in France led by Diane Rivet published a study finding that it was possible to track tankers with fiber installed at up to 2000 meters depths. Though they got far better results at shallower depths of 85 meters. Due to attenuation, the accuracy and range are impressive. At depths of 85 meters, they were able to track one tanker's engine and equipment sounds from up to two kilometers away from that cable. Pretty remarkable. Wonder if it can hear subs? Generally, maritime surveillance is done using things called passive sonar arrays, special listening systems with hydrophone nodes installed onto the seafloor. Civilian or military, these things cost a lot, reportedly hundreds of millions or billions of dollars. For military systems, DAS fiber systems offer far broader range. And as compute and AI technologies improve, they will get ever better at identifying what exactly is passing nearby. The ramifications of turning dark fiber into the world's largest and most comprehensive acoustic sensor network are pretty large. Relating to science, DES can help track ocean conditions across massive stretches in deeper detail than ever done before. It can discover, track and count large animals like whales with unprecedented accuracy, granting us the chance to track the recovering populations in real time. The geopolitical ramifications are significant too. Ships can turn off their transponders to evade satellite tracking, but the cables can still hear their acoustic signatures and even make out what they are doing. It can listen for and warn of potential threats to infrastructure, like an anchor dragging across the seafloor, giving time to scramble countermeasures. And it presents a powerful surveillance tool. Perhaps before long, DAS techniques can make the ocean, and maybe even land, a far less private place than it ever has been before. It is a game changer. Alright, everyone, that's it for tonight. Thanks for watching. Subscribe to the channel. Sign up for the Patreon and I'll see you guys next time.