Leveraging Levees with Digital Twins: Tufts Duo Receives NSF Grant

Levees—raised structures that are built parallel to bodies of water to prevent flooding—are an essential tool for mitigating flooding, which is occurring more frequently and intensely all over the globe. By diverting excess water to protect communities, agricultural land, and infrastructure, levees help protect more than 23 million people (over 700,000 in Massachusetts alone), millions of acres of farmland, and trillions of dollars in property. 

But with 23,000 miles of levees across the United States, ensuring the structures are working properly proves to be quite difficult. Not to mention, many levees were first built by farmers hundreds of years ago to prevent seasonal flooding and need significant upgrades.  

Farshid Vahedifard, professor and Louis Berger Chair in the Department of Civil and Environmental Engineering, recently received a $1.15 million three-year grant from the National Science Foundation (NSF) as principal investigator for a project titled “NSF TTP-T: Transforming Levee Monitoring and Risk Assessment by Translating Hybrid Digital Twin Innovation into Practice.” The project aims to advance levee monitoring and risk assessment through the development of digital twins—virtual models of physical objects or systems developed using real-time data—which could help inform and accelerate levee maintenance. The grant is funded through the Translation to Practice Program, a new NSF initiative to translate research discoveries into practical tools that benefit communities, industry, and society.  

Vahedifard is a leading expert in geotechnical engineering and resilient infrastructure, most recently winning a 2026 ASCE State-of-the-Art of Civil Engineering Award from the American Society of Civil Engineers for a book he co-authored titled, “Effects of Climate Change on Life-Cycle Performance of Structures and Infrastructure Systems.”  

Proactively protecting communities and land 

“The first generation of levees were initially made to protect farmland a long time ago,” explained Vahedifard. “As land use changed and developed, they sort of accidentally became important, slowly protecting more valuable land and infrastructure over time.”  

The average age of a levee is 62 years old, but the current approach to upgrading levees is reactive: fixes are prioritized only after levee failure, in which human, ecosystem, and infrastructure health is already damaged.  

Developing digital twins of levees could help shift to a proactive approach that identifies and addresses levee issues before they cause a larger problem—or even before issues arise at all. Vahedifard will collaborate with structural modeling expert Babak Moaveni, associate chair and professor of civil and environmental engineering, to develop digital twins that will simulate anticipated levee function, continuously updated with real-world data about levee performance.  

“Sensors are great for monitoring levees,” Vahedifard said. “But you can’t use sensors for 23,000 miles of gigantic structures. The digital twins will help us apply our understanding of more heavily studied and sensored levees to structures with little to no data, so we can better understand, monitor, and improve levees everywhere.” 

Research that reaches the real-world 

To ensure the capabilities of levee digital twins reach the communities they could help protect, the Tufts duo will collaborate with experts from industry, federal and state agencies, and local levee districts, including the U.S. Army Corps of Engineers. Incorporating the knowledge and perspective of those who actually work to monitor and fix levees can assure that the digital twins provide useful, practical information. 

With more accurate representation of levee function and failure, digital twins are a promising tool to better monitor these important, vast structures essential to mitigating damages from the world’s intensifying floods and precipitation. 

The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Science Foundation. Research reported in this article was supported by the National Science Foundation under the following award number: 2552816.