Earth and Mineral Sciences

Using machine learning, existing fiber optic cables to track Pittsburgh hazards

A $937,000 NSF grant will fund the project led by Penn State researchers

Using a $937,000 grant from the National Science Foundation and a novel, low-cost approach, researchers at Penn State and Carnegie Mellon University will use Pittsburgh's fiber optic telecommunications infrastructure to monitor a section of Pittsburgh for geological and infrastructure hazards. Credit: Pixabay. All Rights Reserved.

UNIVERSITY PARK, Pa. — Existing fiber optic cables used for high-speed internet and telecommunications, in combination with machine learning, may be able to help scientists track ground hazards in Pittsburgh. The National Science Foundation awarded a $937,000 grant to a team of Penn State and Carnegie Mellon University (CMU) researchers to further develop the low-cost monitoring approach.

The effort, which is led by Tieyuan Zhu, associate professor of geosciences at Penn State, relies on prior research that shows hazards such as flooding, landslides, sinkholes and leaking pipes can be monitored at a fraction of the cost of existing methods.

The distributed acoustic sensing (DAS) approach developed by Zhu uses nanoscale vibrations captured through data cables to diagnose hazards. The technology assesses hazards over large areas, and costs about one-tenth as much as existing hazard sensors used in the Pittsburgh area.

 

“Cities around the world are dealing with the compounding challenges of geohazards, aging infrastructure and climate change,” Zhu said. “This manifests as issues such as leaky sewer pipes, storm flooding and geotechnical failures like sinkholes and landslides. Many of these issues are related to the subsurface environment that plays many critical roles for cities but for which there is a lack of real-time information.”

The grant — one of just 19 offered nationwide — is part of the Civic Innovation Challenge, a national research and action competition focused on using rapidly transitioning emerging technologies to address community challenges.

According to Zhu, the research team chose to further develop and test their approach in Pittsburgh because of its aging infrastructure, challenging terrain and susceptibility to geological hazards. Collaborator David Himes, sustainable communities manager at the Penn State Center Pittsburgh, will help the team leverage and expand upon existing partnerships with local municipalities and utilities. In addition, collaborator Karen Lightman, executive director of CMU’s Metro21: Smart Cities Institute, will guide the team’s interactions with community partners with the goal of facilitating diverse, equitable and inclusive engagement. 

DAS can deliver city-wide signals along the cable in nearly real time, Zhu said, enabling researchers to identify and localize urban environmental and infrastructure hazards across different communities. The data could allow cities, utilities and communities to take targeted and timely interventions in a more cost-effective and equitable manner tailored to the communities affected.

The pilot project, which builds on preliminary research in Pittsburgh, will target just one area of the city to test the approach. DAS will be calibrated with temporary conventional sensors and input from civic partners. Researchers said the challenges Pittsburgh offers could show proof of concept that the science would work in other areas.

Lauren McPhillips, assistant professor of civil and environmental engineering at Penn State, will contribute by validating the DAS against existing sensors. Zhen Lei, associate professor of energy and environmental economics at Penn State, will lead the economic assessment of the technology.

“With extensive community engagement, this project will demonstrate and calibrate this new real-time, high resolution, easy-to-implement and economic sensing system that has potential to truly revolutionize how we manage geo-environmental and infrastructure challenges and create smart, sustainable and equitable cities,” Zhu said. “While Pittsburgh is a natural fit for this initial implementation, we anticipate that the insights and lessons learned could facilitate much broader implementation of this potentially very impactful tool.”

Seed grant funding from the Institute of Energy and the Environment led to this NSF-funded study.

Last Updated November 13, 2023

Contact