A new inspection method developed by engineers at the University of Houston could transform the way building owners and engineers assess hidden structural damage in modern steel-framed buildings.
By combining ground-penetrating radar with artificial intelligence, the technology allows inspectors to detect possible problems in concealed cold-formed steel framing without tearing open walls, ceilings or cladding.
The breakthrough comes at a time when cold-formed steel is being used in an increasing share of commercial and institutional construction across North America.
Lightweight, cost-effective and more environmentally efficient than traditional hot-rolled steel, cold-formed steel now accounts for roughly one-third of non-residential buildings in the United States.
But while the material has become increasingly common, inspecting it after installation has remained a major challenge.
Traditionally, assessing concealed steel framing required inspectors to physically expose the structure by cutting into drywall or removing cladding systems. That process is expensive, disruptive and time-consuming, particularly in occupied buildings or in structures requiring rapid post-disaster assessment.
Researchers led by Vedhus Hoskere, Kaspar J. Willam, assistant professor of civil and environmental engineering at the university, believe they have found a better way.
“To address these limitations, we introduce a new framework that combines a quick radar scan with AI that reads the radar images and points to where the steel is, where damage is likely and the severity and type of damage,” explained Hoskere.
“That lets inspectors verify only the flagged spots instead of opening up everything – saving time, money and disruption and helping maintenance or post-disaster assessments scale.”
The research was published in the Civil Engineering Journal under the title Concealed Cold-Formed Steel Structural Members and Damage Assessment Integrating Ground Penetrating Radar with Vision Foundation Model.
At the heart of the system is ground-penetrating radar, or GPR, a technology more commonly associated with locating buried infrastructure underground. In this application, the radar device is scanned across a wall surface, sending electromagnetic pulses through drywall or other cladding materials.
When the radar waves encounter steel framing behind the wall, they bounce back and create distinctive patterns in the radar image.
“The radar sends pulses into the wall and listens for echoes from what’s behind it,” Hoskere said. “Hidden steel creates a recognizable pattern in the radar scan image. If the steel is damaged – for example, buckled – it can create a small gap or void that changes the echo pattern in a consistent way.”
The second part of the innovation is the AI system trained to interpret those radar signals automatically.
“The AI is trained to recognize these patterns and draw boxes around them, labeling what it thinks it sees,” said Hoskere.
In simple terms, radar captures the hidden image while AI interprets what the image means.
The researchers used a large-scale vision foundation model known as InternImage to analyze the radar data. To train the system, the team created a specialized dataset containing radar scans of concealed cold-formed steel members behind a variety of common wall coverings and under multiple damage conditions.
That dataset includes different member orientations, cladding combinations and damage types, allowing the AI system to learn how hidden steel behaves under real-world conditions.
The researchers also developed what they call “GPR-CutMix,” a new AI training technique designed to improve the model’s ability to handle variations encountered in actual buildings. That includes differences in stud spacing, wall assemblies and field conditions that are often messy and inconsistent compared to laboratory settings.
According to the researchers, one of the most important findings was the model’s ability to generalize from controlled laboratory data to real buildings with different wall systems and concealed framing configurations.
That capability could make the technology particularly valuable after natural disasters such as earthquakes, hurricanes or severe storms, when engineers must quickly assess structural integrity across large numbers of buildings.
Instead of conducting destructive inspections throughout an entire structure, inspectors could rapidly scan walls, identify areas of concern and focus invasive investigations only where needed.
The system could also have important applications in ongoing building maintenance and rehabilitation projects, especially as aging commercial buildings increasingly require condition assessments.
Muhammad Taseer Ali, lead author of the research, said the findings demonstrate how advanced sensing technologies and AI can modernize traditional building inspection practices.
“These findings highlight the potential of our framework to advance the concealed cold-formed steel structural inspection methods by providing a rapid, reliable and scalable approach for damage detection, ultimately improving building maintenance and rehabilitation.”
Before beginning his doctoral studies, Ali spent a decade working in industry on cold-formed steel structure design, experience that helped shape the practical focus of the research.
While the technology is still in the research phase, the implications for the construction and building management sectors could be significant.
