In 2019, a storm was keeping Tieyuan Zhu up at night. It wasn’t a dog’s howl chasing the cracks of thunder or the thought that water might be creeping into the basement. He had just installed his novel acoustic sensing technology in some of the fiber optic cable lines beneath Penn State’s University Park campus, and he was curious if the data was being logged on a tiny server near his Deike Building office.
In the morning, he rushed into the office of David Stensrud, then head of the Department of Meteorology and Atmospheric Science, to compare data.
It matched.
Even better, in something that even surprised Zhu, associate professor of geosciences, the acoustic data showed the directional strength of the storm. This meant that the spatial resolution was even stronger than he thought.
Distributed Acoustic Sensing (DAS) technology
Zhu is an expert on what happens below ground. He’s a trained geoscientist who often takes out-of-the-box approaches to solving traditional problems. DAS is one such approach.
“I like to think of myself as an untraditional earth scientist,” Zhu said. “I’m interested in advancing science but with a goal in mind of advancing society. I’m always asking how can these tools be used to solve societal issues.”
Because things happening below ground create motion—simple physics—they also create acoustic signals in the same way the stereo speaker physically moves to create sound as a turntable needle rides the grooves of a record, again creating sound from the motion. DAS technology assigns acoustic signals to all those physical motions.
Using DAS, Zhu’s team taps into the fiber optic cable and fires a laser down the line. It returns, recording all movement as well as where the movement occurred along the way. It even records how far away the sound was in all directions. Because the cable indiscriminately records all motion, there’s a lot of extraneous data. Using machine learning, Zhu’s team has worked for years applying filters to isolate only the data they want.
Penn State, Pittsburgh, and beyond
Since then, Zhu’s team has shown how DAS can spot a whole range of underground issues. It’s effective at real-time tracking of storms, traffic patterns, sinkholes, water leaks, flooding – just about anything that creates a low frequency sound wave.
For a fun proof of concept, the team tracked the scoring of a Penn State football game and identified four songs by the band Grouplove during Movin’ On, Penn State’s student-run music festival.
Pilot work that started at Penn State recently led to a large-scale project in Pittsburgh. Zhu’s team secured one of only nineteen Civic Innovation Challenge grants – $937,000 – sponsored by the National Science Foundation. Zhu is principal investigator on a team that includes Carnegie Mellon University and researchers across Penn State. Civic grants are awarded to rapidly transitioning emerging technologies that address community challenges.
Zhu’s team installed DAS in roughly 18 linear miles of fiber optic cable in sections that cover the residential area of Oakcliffe, downtown Pittsburgh and the city’s West End. The locations were chosen for the variation in terrain and development.
Early results are promising.
“The beautiful thing is that the data has been so spatially accurate,” Zhu said. “We can see within meters what is happening.”
Pittsburgh is a perfect analog for most east coast cities. It’s heavily populated, topographically diverse, and littered with an aging infrastructure. According to the Franco Harris Pittsburgh Center at Penn State, a partner on the project, geohazards are very costly for the city.
Zhen Lei, professor of energy and environmental economics, has been tasked with looking at the social impact of the project. Often, higher income areas are equipped with more resources. Yet, fiber optic cable can be an equalizer. It’s throughout the city. The team hopes to show DAS is effective at tracking hazards in all parts of the city, even lower income areas.
Zhu said DAS wins on cost and accuracy. For example, he said, a single water leak sensor costs about $1 million annually to operate. Because each sensor only tracks data in the spot it’s located, many sensors are needed. However, fiber optic cables are everywhere. In the test area, DAS is currently recording at a resolution of about four meters. The only barrier, Zhu said, is dealing with the data. In 4 months of operation, about 200 terabytes of data have been sent back to Penn State. However, the cost of managing that data is less than the cost of a single water sensor, Zhu said.
In that time, researchers relied on existing sensors and ground-truthing to calibrate and corroborate DAS.
“Using existing fiber optic networks is an innovative and cost-effective way to gather information on water flows underground, as well as changes in flow, without having to deploy and maintain a large number of costly flow monitors throughout the underground system of pipes,” said Beth Dutton, senior project manager of stormwater at Pittsburgh Water. “We welcome exploring new technologies such as this to more effectively leverage our ratepayers’ resources to address Pittsburgh’s challenges.”
Visualizing the data
Making sense of the mounds of data is something that comes easy for people like Zhinong Wang, a postdoctoral scholar hired this spring to help on the project. But it’s useless if city officials can’t make sense of it.
He’s been working with Shouvik Majumder, an undergraduate student in computer science in the College of Engineering, to develop filters so that city planners can visualize and isolate data.
In just a few months, Majumder created a graphical user interface so that DAS data can be filtered with the click of a mouse.
Thinking bigger
Zhu’s goal is bigger than one city. He wants to show DAS is cheaper and more effective at tracking geohazards in any city. He said all of the technology scales. There are many cities with an aging infrastructure, rising costs, and increased challenges to dealing with geohazards, traffic, and even yet-to-be-developed uses.
A few of Zhu’s students are exploring novel uses for DAS while earning their doctorates. So far, DAS has proved helpful in oil and natural gas monitoring, geothermal assessments, and carbon sequestration monitoring.
Joe Miller, a graduate student, is using DAS to track things like mine blasts, traffic, and earthquakes near the University Park campus. He points to research that shows DAS improving warning times for a slew of geohazards.
“If you really want to make an impact, you need to ask yourself the ‘who cares’ question,” Miller said. “And people care about better predicting these hazards. Revealing when hazards are going to happen is the holy grail of geophysics.”
As he watches the news during hurricane season, Zhu’s still spotting new opportunities for DAS to help emergency responders. For example, flash flooding occurs when the ground can’t absorb any more water. Yet, until DAS, he said there were no cost-effective means for tracking this.
“DAS can show when the ground is below its threshold and when it’s surpassed it,” Zhu said. “If we have fiber optic cable, we now have a way to track this. The list of things we can accomplish is continuing to grow and seems limitless.”