Concrete culverts, which channel water under roads and embankments, are an essential component of urban and rural infrastructure. They’re commonly constructed using pre-made segments connected on-site, a technique that lowers costs and shortens road-closure times, among other benefits.
The joints where the segments come together, however, can separate over time. Separations can let soil enter the pipe — potentially clogging the culvert, damaging the embankment, or even creating a sinkhole in the road above.
A MnDOT-funded research project aimed to determine what contributes to joint separation. Even though concrete culverts have been common for decades, the factors that lead to joint separation have not been thoroughly studied, says Brock Hedegaard, former associate professor in the Department of Civil & Environmental Engineering at the University of Minnesota Duluth (UMD), the project's principal investigator. Based on the study findings, the research team recommended installation practices to avoid significant joint separation — and keep culverts and roads open.
The study focused on round precast concrete culverts. Roughly 75 percent of the highway culverts owned by the state are concrete, nearly 87 percent of which are round pipes, according to a 2012 study. That study also found that joint separation affects nearly 20 percent of the MnDOT round concrete culvert inventory.
In this project, Hedegaard’s team took a three-pronged approach: examining a MnDOT database of culvert and storm drainage assets, conducting a field survey, and performing computational modeling of soil-culvert systems.
The database examination showed that separation is complex. “No single variable stands out as the primary cause,” Hedegaard says, and some pipes with minor joint separation may not show other problems or harm.
Using the database, the researchers created a model for predicting the likelihood of joint separation, looking for commonalities. They confirmed that the condition of a culvert was the most important predictor of joint separation. “Joint damage regularly led to further issues such as soil infiltration and cavities forming on the embankment inslope,” Hedegaard says.
Another notable finding: Separation was most common in southeastern and southwestern Minnesota. The counties with the highest rates of joint separation often had many intermittent streams and rapidly changing landforms. Geographic features were more important for predicting joint separation, he says, than geometric features such as pipe size or installation depth.
In the field survey, the team inspected 86 concrete culverts throughout Minnesota during the summer of 2024. “We saw the most separation at joints near culvert ends and at joints without pipe ties,” Hedegaard says.
Pipe ties are steel rods — essentially long U bolts — that lock segments together. A common practice has been to tie only a few joints at the ends of concrete culverts, based in part on design and installation requirements as well as on cost. The field survey found, however, that “partial tying” tends to push joint separation to the first joint just past the ties. “Partially tied pipes very often had some kind of separation,” he says.
To augment the inspections, researchers surveyed MnDOT construction and inspection personnel. The most common suggestion from the survey? Tie all concrete pipe joints. Other possible causes of separation, the staff reported, include insufficient compaction of backfill material around the pipe and freeze-thaw damage at the ends of pipes. Compaction increases the “stiffness” and density of soil, making it less prone to sinking and water infiltration.
The team’s modeling focused on the impacts of traffic loads and the weight of the embankment. Results showed that both forces put their highest demands under the center of the road. And compared with embankment self-weight, Hedegaard adds, “Traffic load is a pretty minor player.”
The database analysis and field inspections, however, showed the most separation at culvert ends. “This contradiction indicates that traffic loading and embankment self-weight likely do not lead to significant joint separation by themselves,” he explains. Instead, stresses that build up at the ends of the culverts — caused by factors such as soil freezing and thawing and changes in the water-table level — may have greater effects on separation.
The overall recommendations of the study, both from the research team and from the survey of MnDOT personnel, are to tie all joints during installation, properly compact the backfill, and limit freezing expansion of the embankment materials. “Soil freezing is our most likely cause of separation at the ends,” Hedegaard says. “All soils near culverts must be frost-heave-resistant.”
“This project identified probable causes of joint separation to help improve standards and construction practices for culverts while also offering new ways to leverage our hydraulic infrastructure database to support future asset management needs,” says Nicholas Olson, state hydraulic engineer with MnDOT’s Bridge Office and the project’s technical liaison.
— Pam Snopl, contributing writer
