By T.R. Witcher
Walking the grounds at the CES in Las Vegas—dodging the 175,000 attendees, 4,000-plus exhibitors, 6,000 or so reporters, and innumerable screens—can be an overwhelming experience. Tech nerds come to this premier technology exhibition event for the 8K televisions from Samsung and LG or the futuristic concept cars shown by Mercedes-Benz and Sony. Others come for visionary glimpses of pie-in-the-sky transportation solutions, like aerial taxis by Hyundai/Uber and Bell. But away from the flashy consumer electronics are plenty of firms and researchers working on smaller, more focused technologies that may prove to have significant impacts on civil engineers, builders, and city planners. Contributing editor T.R. Witcher reports from the show floor.
The 2020 CES—sponsored by the Consumer Technology Association, based in Arlington, Virginia—was full of products designed for a future filled with objects connected and communicating through the 5G wireless network protocol. Among the show's many "smart city" offerings, most of these devices focused on advanced sensors in automobiles that can communicate with other automobiles or with traffic signals or streetlights.
But one company, Pittsburgh-based RoadBotics, is instead developing a novel way to analyze the roads themselves. RoadBotics began as a spinoff from the Robotics Institute at Carnegie Mellon University. President and cofounder Benjamin Schmidt, Ph.D., had the idea of using smartphones, which are already packed with sensors, to study the health of roadway infrastructure.
The predominant method of analyzing roads is laborious, inefficient, and time consuming. According to Schmidt, most owners send inspectors to conduct manual surveys of a specific area's road network. Assessors pull their vehicles over to the side of the road and make notes in their notebooks—a tedious and dangerous task. It can take weeks to inspect a 100 mi section.
At the other extreme, Schmidt says, the Federal Highway Administration deploys big, expensive sensor-laden vans to inspect federal highways—but these can cost thousands of dollars per mile, and the time it takes to process the data they generate can stretch to as long as nine months. If a sensor van completes an inspection of a road network in September, for example, the data might not get into the hands of transportation officials until spring or later—meaning any deterioration of the roads that occurred over the winter goes unaccounted for.
RoadBotics is attempting to split the difference by leveraging the sensors in smartphones to analyze roads. Mounted on the interior of the windshields of vehicles, the phones can combine videos, GPS data, and machine learning to find potholes, "alligator cracking," and longitudinal cracking cracking. The company is teaching the phones to distinguish between these anomalies, which indicate problems, and such common features as manhole covers and lane markings.
A RoadBotics phone takes photographs of 10 ft sections of pavement and generates a map, cataloging individual examples of "road distresses," and combines them into a rating that is both numerical (1 to 5) and color-coded. This allows the agencies that operate the roads to get a macro-level view of the general state of their roads and identify individual problems at the more local level.
RoadBotics's solution also streamlines what is currently a fragmented process. State-owned roads may be analyzed once per year, while smaller, locally owned roads may only be inspected every three years. And it may take a staggering five to seven years to analyze all the data from a given jurisdiction. RoadBotics promises it can collect all that data in just two days and deliver it in meaningful form within 30. This could give infrastructure owners and operators data that is not only more comprehensive but far more up to date than ever before, enabling them to better allocate scarce resources quickly.
"For any piece of road, you can usually calculate a degradation curve (due to) weather, material, traffic," and other factors, says Schmidt. But real-world complexities mean "we very rarely have a good understanding of traffic or where [we] are on that degradation curve."
This can lead to agencies adopting what he calls the "worst-search strategy," which involves patching those parts of the road network that are in worst shape first and hoping the moderately degraded parts don't get any worse in the meantime. Schmidt considers this "the worst way to manage any asset. You want to focus your resources on those roads in the middle, just starting to go bad," he explains. By doing this, owners can make simpler, less costly fixes to more roads and lane-miles before they reach crisis states.
Savannah, Georgia, recently completed a full survey of 700 mi of roadway using the RoadBotics system, shaving a three-year process down to three months and saving $80,000. (Read "Street Smarts,"
, February 2020, pages 60-67.) The results helped the city secure $28 million in state funding to target a specific array of road improvements.
Currently, Schmidt says, roughly 200 governmental entities, including Detroit; South Bend, Indiana; and London, are using the company's technology for road assessment. "Every government is stretched way too thin," he says. The money and personnel time saved by using the system can be spent on "higher value" projects, he points out.
What's next for the firm? Schmidt envisions that RoadBotics phones could be attached to street sweepers, combining a service that already exists with the new technology to save even more money. And there's plenty more civil infrastructure that can be digitized and analyzed by the system in the future-curbs, sidewalks, signage, hydrants, utility poles, and streetlights are all infrastructure assets that must be assessed and maintained.
Yet Schmidt is determined to not lose the human element. "If I take ten civil engineers and have them look at the same street, I'll get eleven different answers," he quips. Data collection, he says, is just a tool, even if it is a powerful one. "You take the data, then go visit the street," he says. "We won't replace civil engineering judgment, but we will support [those] decisions."
According to the U.S. Department of Transportation, there are more than 600,000 bridges across the United States—and about 56,000 of them are considered structurally deficient. The primary way to inspect them to determine which are most in need of repair is through nondestructive evaluation (NDE). But current NDE technologies can't keep up with the demand, and bridge inspections are notoriously challenging for human beings. "It's very difficult to do that job," says Hung La, Ph.D., A.M.ASCE, an assistant professor of computer science and engineering at the University of Nevada, Reno (UNR). It's dangerous to climb a bridge to inspect it up close. Only a select group of engineers are trained to do so, and it can be difficult to hire and retain them. "They only work for a couple of years and then leave," La says. "It's a very hard job."
Inspections are also time consuming; it can take 10 days or more just to inspect the towers of a large bridge like the Golden Gate. La, who is also the director of UNR's Advanced Robotics and Automation Lab, has been searching for a technological solution. He's closing in on one with his plans for a bridge inspection robot.
The 3 ft long, 24 lb robot looks like a tiny Mars rover—a computer wrapped in a steel cage wearing a pair of roller skates—or, more accurately, chunky, magnetized wheels. It runs on a flat or cylindrical surface, and when it transitions from one surface to another—making a 90-degree turn at a bend, for instance—it can lift one set of wheels, like a foot, and plant it on the new surface.
The biggest challenge for La and his fellow researchers is keeping the robot on the bridge. The magnets need to be strong enough to hold the robot in place, but that makes the device heavier, which means larger motors are required to move it. So, for now, it moves more slowly and less nimbly than La would like.
A combination of GPS, lidar, and inertial measurement unit (IMU) technology—which handles orientation, velocity, and gravity—helps the robot stay on course. The device is equipped with a camera for visual inspection and an NDE sensor that can check for cracks inside steel and also determine the integrity of welds and bolts.
The robot rovers, La says, have the potential to work more quickly and safely than human workers. And he argues they will work better than aerial drones because drones can conduct only superficial examinations of bridge surfaces; they can't attach to the steel or see what's within it.
There's one other, somewhat more subtle, benefit to these miniature robotic vehicles. When motorists see inspectors dangling from bridges or drones flying around, they tend to stop and stare, creating accident hazards. La's robots are generally too small for drivers to notice.
Currently the robots are controlled remotely by users, but the team is working on autonomous operation. The robots are still being tested and refined, but they have undergone trials on more than 30 bridges. La founded a start-up, Automated Inspections Robots (AIR) Corp., in Reno, and has received interest from the city of Calgary, Alberta, Canada, as well as the Georgia Department of Transportation. "My goal is to be able to sell prototypes for bridge inspection within a year," La says.
Meanwhile, a second research team in Nevada, this one based at the University of Nevada, Las Vegas (UNLV), is trying to develop a way to better maintain another piece of critical transportation infrastructure—railways.
All rail track eventually deteriorates, explains Zhiyong Wang, Ph.D., an associate professor of mechanical engineering at UNLV. "With all rail, just like all people, there will be a time when it gets old," he says.
Wear is inevitable, and it's more serious on curved track. Wear on the rails means more vibration, which stresses the trains, affecting their safety. Degraded rails also slow down trains and cause delays, and those delays can be extensive when sections of track have to be pulled out and replaced.
Even wear that creates just an inch or two of difference in elevation from one element to another can affect train performance; Wang likens it to walking around with one leg a bit shorter than the other. Wang's objective is to quickly fix the wear and tear of rails in place without cutting out pieces of worn track and swapping in new ones.
His proposed solution is to use 3-D printing to rebuild worn-out rails in place. "Traditionally, train rails are made by forging," he says. "You squeeze the beam from a red-hot point into shape. Once it has been made, what can you do [to change it]? The traditional manufacturing process doesn't have a good solution. But 3-D printing really opens up a new approach."
There's more than one way to engage in 3-D printing, including using lasers and electric arcs. Both work but in different ways. With a laser, pieces of metal are dropped over the rail—Wang likens them to snowflakes falling from the sky. When they hit the laser, they turn to liquid and fall onto the surface of the rail. A metallurgic process starts to quickly bond the new material to the base metal.
An electric arc, on the other hand, sends a current through a piece of metal and melts it; the melted metal flows onto the rail and bonds with it, becoming solid.
Laser printing works better for adding a thin layer to the rail; electric arc printing works better for thick layers. Both technologies show potential for fixing light-rail, subway, and even heavy freight rail.
In a paper Wang coauthored with two UNLV colleagues and delivered at an American Society of Mechanical Engineers conference in 2018 ("Thermal-Mechanical Study of 3D Printing Technology for Rail Repair"), the researchers examined the challenges of adding heated metal to existing, cooled rail. The hope is that "the interface of the railhead and additive materials should conserve high stresses to prevent any crack initiation," they wrote. "Otherwise, the additive layer would likely shear off the rail due to crack propagation at the rail/additive interface."
Their research suggests that preheating the existing rail is the key to significantly reducing any residual stresses—by about 40 percent—at all points along the transversal direction of the interface.
Another trick is to find the right material; it has to match the underlying rail. But Wang cautiously predicts that his method should be able to cut in half the eight hours it usually takes to replace a piece of worn rail.
He is currently seeking investors to bring his team's technology out of the lab and onto test sites and from there into the field. "We hope one of those companies would be interested in [building] a larger-scale testing site," he says. "We're doing this at a university-site scale. We need a company scale."
One of the most compelling pieces of infrastructure-related technology at CES was shown by a company trying to reduce industrial water usage. Somerville, Massachusetts-based start-up Infinite Cooling was founded a few years ago by a research team at the Massachusetts Institute of Technology (MIT) that was exploring how to capture small fog droplets in the air more efficiently. The answer, says chief technology officer Karim Khalil, Ph.D., involved electrostatic forces—using an electric charge to attract or repulse particles. In seeking practical applications for its breakthrough, the company began to look seriously at cooling towers. Cooling towers at power plants (as well as data centers, manufacturing plants, food and beverage processing plants, and other industrial facilities) emit a staggering amount of water in the form of vapor—that's the striking plume of fog that can be seen surrounding those units in action.
A 600 MW power plant—big enough to power a city the size of Boston—will lose 3,000 gal. of water per minute, an astonishing 60 to 70 percent of that through evaporation, says Khalil. The rest will be lost, as water is purposefully "bled" from the system to ensure the water chemistry remains at an acceptable level and doesn't get too salty.
Khalil says 39 percent of the fresh water used in the United States goes toward cooling such industrial and manufacturing plants. Infinite Cooling claims it can offset about 20 percent of that by reducing the amount of water lost to evaporation and plugging that newly collected water back into the plant, which would reduce the bleed rate. That could mean a savings of 100 million gal. of water a year. And that's just for one plant; there are thousands of such facilities around the country.
"We're talking astronomical amounts of water," he says. Infinite Cooling's technology does two things. First, it charges droplets coming out of the plumes with electric charges, and second, it creates an electric field that pushes those charged droplets toward a collector above the tower. The collector is an array of meshes that attract the microscopic droplets. As these collect and form larger droplets, the droplets drain down the mesh into a series of guttering systems, which then merge to feed a common pipe.
The meshes would need to be situated several feet above the towers, held in place by a structure—perhaps made of steel, fiberglass, or aluminum. This is why Infinite Cooling is hoping to partner with structural and civil engineering firms to refine the plans for the retrofit of tower structures, taking into account variations in tower size as well as plant layout and even local wind conditions.
Power plants operate at fairly thin profit margins, Khalil says; reducing operational costs is always a priority. Infinite Cooling promises to do so in three ways: by reducing the amount of water plants have to purchase for cooling—which would also be a net gain for other water users in those communities; by reducing the amount of water released into local wastewater treatment systems; and by reducing the amount of chemicals required to make sure the water is properly balanced with respect to salt content.
And an added benefit may be more aesthetic in nature: plumes are often seen as, at best, nuisances or unattractive and, at worst, health concerns because people often don't know that they are only water vapor. Khalil's system can help power plants be better neighbors.
At the end of 2018, Khalil and his colleagues completed a pilot project at MIT's 20 MW gas plant as a proof of concept, and they continue to refine prototypes in the lab. Last year they achieved their first commercial installation at a small nuclear facility near Boston.
For now, the main target, especially in the United States, is working with structural and civil engineers to retrofit existing plants with this collection. Down the line, Infinite Cooling may be able to collaborate with engineers during the design of new facilities, Khalil says, by "embedding our system in the structure of the cooling tower itself."
-T.R. Witcher is a contributing editor for
Civil Engineering, April 2020, © American Society Of Civil Engineers. All Rights Reserved