Researchers at a Purdue University lab are using AI to improve concrete used in highways and bridges to cut costs and save taxpayer dollars.
One technique involves installing sensors that help engineers know when new concrete is ready to take on heavy traffic loads. With that information they can prevent tiny stress cracks that lead to expensive repairs if roads are opened too soon. Fewer repair projects to replace concrete, in turn, cut down on traffic held up by the road work.
Luna Lu, an associate professor at Purdue’s Lyles School of Civil Engineering, and a research lab team have embedded sensors they developed into three Indiana interstates near Indianapolis. Data from the sensors helps the team know the best time to open traffic after a patching or new pavement project. They can also continuously track concrete development.
Lu collaborated with the Indiana Department of Transportation to embed the sensors in the highways near Indianapolis and her team is now working with the Federal Highway Administration to use the concrete sensors in other states, including California, Texas, Kansas and Missouri.
The sensors the Purdue lab used for the project are piezoelectric sensors, which can generate an electric charge in response to mechanical stress. In this case, the sensors are attached to rebar about every 11 inches inside the concrete. On command, each sensor vibrates the concrete to detect stiffness and the mechanical energy of the vibration wave is converted back to electrical energy. That charge is first measured within an hour after the concrete is poured. The results can be interpreted to show the strength of the concrete at certain times as it dries.
One benefit of relying of the electrical measurement is accuracy over conventional mechanical measurements that require construction workers to cast a heavy concrete beam onsite and transfer the beam offsite for analysis. Another conventional method relies on temperature testing, but because there are so many different concrete mixes there can be thousands of different temperature-strength correlation curves which can be influenced by outdoor temperature.
In a short video, Lu said that road construction projects are often done in a hurry and concrete is opened to traffic too soon to hold up well. Opening too soon can cause premature failure, which leads to frequent repairs. Collecting more data on how soon concrete cures properly can help highway departments plan more efficiently.
The sensors being installed are protected from corrosion and can be used to collect data for years. “Being able to track concrete strength over a longer period of time would help engineers to know if they’ve over- or under-designed roads and better determine when to replace the concrete,” Lu said in a statement.
Tommy Nantung, research manager for Indiana DOT research and development, said the sensors will help a contractor “to know exactly when the concrete is mature enough to accommodate heavy truck loads and the curling and warping stress.”
Lu’s lab is also working on developing self-healing concrete that frequently crack during winter freeze and thaw cycles. When temperatures drop below freezing, water droplets on the surface freeze and when thawing occurs, they expand beneath the surface, causing cracks and fissures over time.
The lab is investigating different types of porous, sand like materials to act as internal curing agents that would be mixed into concrete. When the concrete cracks, the curing agents absorb the water and begin a chemical reaction that produces substances that seal off a crack. The healing process prevents more water from seeping into the concrete where it can corrode steel or rebar.
A third area of research for the lab is intelligent infrastructure that relies on AI and big data to identify underused traffic lanes and divert traffic to them in order to cut down on the need to add extra traffic lanes.
The theory of intelligent traffic infrastructure has been around for decades, going back to when Chicago and other cities installed reverse traffic lanes. In one example, a divided highway like Lake Shore Drive in Chicago with four lanes in either direction could be converted to six or all eight lanes in a single direction in a heavy rush hour when nearly all the traffic is headed the same way.
The trick with such technology is being precise with when to reverse traffic on a stretch of highway and for how many lanes. Also, it requires an automated mechanism for lane controls, gates and signage that might also communicate with vehicles wirelessly.
“Traffic is always directional. Conventional thinking is to add extra lanes, but artificial intelligence and big data could identify an underused lane and shift traffic into that direction,” Lu said. “We’re developing technology that would allow for better control of traffic without adding extra lanes.”
Big data can also help find the best ways to make infrastructure safer and more resilient. “We can developer algorithms that map out vulnerabilities in infrastructure going forward,” Lu said.
Last year, Purdue created the Center for Intelligent Infrastructure that will bring together researchers to work on ways for infrastructure to directly communicate with humans on hazards such as bridges that have failed and could fall, among other topics. Lu is helping establish a Midwest intelligent infrastructure consortium working with several state transportation departments.
Bridge collapses are a major concern of many transportation departments. Sensor suppliers like ST Micro have seen prices for MEMS sensors decline to make it possible to monitor more bridges with sensors, some that are wirelessly connected to a transportation data center.
Michigan’s Mackinac Bridge first had prototype sensors installed in 2016 to monitor traffic vibrations and to act as early warning technology.
The American Road and Transportation Builders Association said in a recent report than more than 33% of U.S. bridges need repairs. The group tallied 46,100 structurally deficient U.S. bridges with Iowa, Pennsylvania and Illinois with the largest number.
RELATED: