A civil engineering researcher at The University of Texas at Arlington will utilize a $122,000 grant to develop an automatic crack evaluation (ACE) system that could save time and money and prioritize maintenance for roads, bridges and runways.
For the project, Suyun Ham, assistant professor in the Department of Civil Engineering, will develop an internal damage identification model that leverages deep-learning technologies.
“This ACE system will be like using a sonogram on bridges and pavement,” Ham said. “We’ll use the system to find internal cracks—ones that aren’t always visible to the naked eye—without a closed lane on a highway. We’ll use a wave-scattering theory approach to determine the severity of those internal cracks.”
The wave-scattering method shoots stress waves through the pavement to determine the location of cracks. Ham said it can be challenging to identify cracks in need of repair because they are concealed by different layers and types of bridge and road materials.
“It requires you to look through sometimes several different materials before getting to the actual crack,” Ham said.
The initial data collected will serve as a baseline so deep learning and artificial intelligence can predict what might happen to that infrastructure in the future. Much of the deep-learning analysis will be done via both computer simulation and field test results.
The Transportation Consortium of South-Central States, known as Tran-SET, funded the project. Tran-SET is Region 6’s University Transportation Center and includes Arkansas, Louisiana, New Mexico, Oklahoma and Texas.
Ali Abolmaali, chair of the Department of Civil Engineering, said Ham’s work has the chance to streamline costly repairs across the five-state region.
“The region’s member states are very different in that they have different soils, topography and climates,” Abolmaali said. “That’s why the use of artificial intelligence within the ACE system is so vital. Ham’s project could end up saving the region a lot in pavement and bridge repair and replacement.”