When the levee breaks

University of Pittsburgh

With flooding becoming more frequent and severe as a result of climate change, the stakes are rising. Recent estimates place global flood-related damage at over $50 billion annually, and experts predict an increase in damage to U.S. communities by the end of the century without new interventions.

Researchers from the University of Pittsburgh and Vanderbilt University received more than $729,307, with $317,811 coming to Pitt from the National Science Foundation (NSF) for a three-year project to address one of the most critical threats to flood protection infrastructure: backward erosion piping (BEP). This phenomenon, a major cause of levee and dam failures, occurs when water seeps through and erodes sand beneath flood barriers, potentially leading to catastrophic failures.

Alessandro Fascetti, assistant professor of civil and environmental engineering and Roberta Luxbacher Faculty Fellow at Pitt’s Swanson School of Engineering, and Caglar Oskay, professor of civil and environmental engineering and professor of mechanical engineering at Vanderbilt, are developing a novel computational, artificial intelligence-driven model designed to predict BEP and help mitigate risks. Their project aims to revolutionize how flood protection systems are designed, maintained and monitored.

“Flooding is the most common and costly disaster in the U.S., and BEP is one of the least understood threats to levees and dams,” Fascetti said. “By developing a model that simulates BEP progression, we can provide engineers with the tools to predict when and where failures might occur, enabling them to take preventative action.”

The research will use Dual Random Lattice Modeling (DRLM) to capture the complex, three-dimensional behavior of BEP as it evolves beneath levees and other protective structures. This model is capable of describing the unique characteristics of soil and hydraulic conditions, allowing for more accurate predictions of when and how BEP will lead to system failure. The model will also be validated through experiments and real-world data from past flood events.

The Researchers’ Three Key Objectives: 

• Understanding the relationship between water pressure, soil conditions, and BEP initiation
• Developing a novel computational framework for predicting BEP-related failures
• Quantifying the risk of levee failure over time due to BEP progression

The project also includes a public outreach and education component. The team will engage with K-12 students and the public through interactive demonstrations, including an Augmented Reality Sandbox that simulates flood scenarios and demonstrates the importance of infrastructure in flood protection.

“We want to not only advance scientific knowledge but also raise awareness about flood risks and encourage the next generation of engineers,” Fascetti said.