Projects with ‘real-world relevance’ of priority to ICDS co-hire

Hands holding tablet on blurred automation machine see machine learning tools for project.

Posted on July 23, 2024

UNIVERSITY PARK, Pa. — A combination of machine learning and classical physics methods allow Xiaofeng Liu, Penn State Institute for Computational and Data Sciences (ICDS) co-hire and associate professor, Civil and Environmental Engineering, to conduct research that “touches everybody’s life.”

Liu’s current works look at constructing and maintaining bridge infrastructure, river hydraulics, flooding and water quality. The researcher uses ICDS computational tools and resources to help develop computer models to predict and solve these civil engineering, real-world problems.

In a seed-grant funded project using artificial intelligence (AI), Liu’s research team is helping United States Geological Survey (USGS) improve flow measurement, especially during dangerous floods. USGS uses a combination of tools including cameras and sensors to do measurements, and Liu incorporates machine learning and AI to automate the process.

“We combine [the methods] together with the hope to have better and more accurate measurements and data,” Liu said.

High-performance computing methods have also been used in research to restore river functions and prevent floods. Liu is working on a U.S. National Science Foundation (NSF)-funded project doing just that.

“We are heavily using high-performance computing methods and engineering… looking at the complexity of river dynamics and functions,” Liu said. “In the past, civil engineers preferred clearing rivers and paving with concrete to make the river bottoms smooth — that’s traditional thinking. We must also think about ecology, the fish and their environment. The industry [civil engineering] has been restoring the full functionality of rivers and making it more natural.”

A sustainable and safe habitat and environment for organisms like fish living in rivers is a concern in restoration.

In a research project for the state, Liu and his research team looked at nature-based solutions to study how we can mimic nature and fully restore rivers and waterways combining high-performance computing and field work.

“There are boulders, fallen trees and debris in natural rivers which make good habitat for fish,” Liu said. “We are using computer model results to better understand how we should design a structure to mimic nature. There are many interesting projects like this that I am working on.”

In a National Cooperative Highway Research Program-funded project that Liu led, researchers used fundamental flow physics and computer models to quantify obstacles and difficulty for water to travel in floods, determining a key parameter named Manning’s roughness coefficient within all flood models.

“It’s important to figure out that parameter,” Liu said. “Practitioners need guidelines. We are also very glad to have a project like that because it has national and international impact. In flood models, that parameter plays a key role in predicting things like what areas will be inundated and for how long and what the extent of the damage is.”

The project has produced a lot of visibility for Penn State and the research team. Liu has been invited to give talks at conferences about the details of this work.

Liu and his group often work on research funded by state agencies and collaborate with agencies like the Department of Transportation and local government.

“I really like dealing with problems that have real-world relevance,” Liu said. “Many of our projects are the best examples of that. Civil engineering is a very old and traditional profession, but it has a lot of new meaning, new demand, new problems and new technologies … We are doing something tangible, and it potentially touches everybody’s life.”

Liu, who started at Penn State in 2014 as an ICDS co-hire and assistant professor, has been a part of the conversations of AI across the University — the use of high-performance computing machine learning methods and artificial intelligence is increasing and not dependent on discipline.

“I’ve been involved with ICDS in many aspects,” Liu said. “I really enjoy interacting with other researchers there. I want to contribute to how ICDS and Penn State position themselves in AI across the University and beyond. We need to look at how AI is shaping our future in both research and higher education.”

Liu is leading an effort in the civil and environmental engineering department to modernize curriculum with AI and machine learning, adding a certificate with a computing and data science focus in civil engineering.

“The idea is that we want to combine the traditional courses with new tools, ideas and methodologies from AI,” Liu said. “We need to educate our students with both established and cutting-edge knowledges. I want to get ICDS involved as well. ICDS plays a critical role for cross-campus coordination, sharing expertise and providing computing resources.”

ICDS News

Projects with ‘real-world relevance’ of priority to ICDS co-hire

Share

Related Posts

News & Events