Despite Stimulus Money, Most U.S. Bridges Might Stay Dumb

Sensors are starting to prove themselves in the biggest, most complex bridges, but the technology isn't ready for the hundreds of thousands of smaller ones

3 min read

The 2.9-kilometer Rion-Antirion Bridge in Greece, with its 300 sensors, is a testament to how smart a piece of infrastructure can be. It routinely tells operators when an earthquake, frequent in those parts, or high winds warrant shutting down traffic.

”The bridge tolls are meant to collect thousands of euros per day,” says Alexandre Chaperon, an engineer at the company that designed the system, Advitam, in Vienna, Va. ”Without the monitoring system, the bridge would be closed after every earthquake, more than three days in some cases, instead of 5 minutes.”

Dozens of the largest and most complex bridges in the world are already studded with strain and displacement gauges, three-dimensional accelerometers, tiltmeters, temperature sensors, and other instruments. They are wired to central data-acquisition units—though some newer bridges have wireless systems—which collect and analyze the information and relay it to engineers, in hopes of catching signs of distress before human inspectors could. With the United States injecting US $27.5 billion into revamping the country’s roadways and bridges as part of an $800 billion economic stimulus effort, it might seem like a perfect opportunity to add smarts to more bridges.

But monitoring system costs are too high and the benefits unproven for most of the nearly 600 000 bridges and overpasses in the United States, experts say. Sensor systems seem to make sense only for big, complex bridges and for those that are already known to be in trouble. And it will be many years before these systems can be used on other bridges.

An installation of sensors on an existing bridge that would help save the owners money and mitigate disasters has not been demonstrated yet, says Emin Aktan, director of the Intelligent Infrastructure and Transport Safety Institute at Drexel University, in Philadelphia. ”There is no case where we can say sensors have saved a lot of lives,” he says. For most bridges and overpasses ”it will take 5 to 10 more years before we have sufficient fundamental knowledge to start deploying such systems and expect returns.”

John DeWolf, a civil and environmental engineering professor at the University of Connecticut, in Storrs, who has equipped six bridges with up to 50 sensors each over the past decade, has some of the longest-term data in the country. At today’s costs and level of knowledge, he says, a monitoring system might be of most use on ”an older bridge that is nearing the end of its life cycle or a bridge with known problems.” However, there are potentially an awful lot of those. In 2008, the U.S. Federal Highway Administration listed about 152 000 bridges as structurally deficient or obsolete.

The systems also make sense for bridges with new designs that need to be tested, such as the Rion-Antirion, or for structures like the long suspension and cable-stay bridges that span waterways and cost hundreds of millions of dollars to build. ”With these very complex bridges there is a benefit, and there are enough resources available to apply that level of effort and technology,” says Andrew Foden, a supervising engineer with infrastructure consulting firm Parsons Brinckerhoff.

The monitoring system for the 630 million Rion-Antirion should be able to tell bridge operators when and where maintenance is needed, potentially saving 15 to 25 percent of long-term maintenance costs, says Advitam’s Chaperon. The system has already identified two abnormal vibrations in the nearly five-year-old bridge. One of them, if not treated, could have led to damage to the stay cables. So engineers have installed additional dampers on cables in certain locations to prevent the problem.

Foden says structure monitoring is costly mainly because there is no ”one size fits all” system. Every bridge is unique and is exposed to different traffic, weather, and terrain conditions. The instruments must be tailored to each structure and placed at crucial spots. And the more sensors, the better.

When the I-35 bridge collapsed in Minneapolis in 2007, it was due to a gusset plate failure, Foden says, and a strain gauge at that particular gusset might have given a warning. ”But there can be thousands of gusset plates on a bridge, and to put sensors everywhere becomes impractical,” he adds. Analyzing data from thousands of sensors would only add to the cost.

The National Institute of Standards and Technology recently released new funding for research on lowâ¿¿cost wireless sensors and smart materials that can be built into a structure to measure changes, as well as other technologies that could bring down the cost of smart bridges. But results from that research are likely years away.

The Conversation (0)

The Future of Deep Learning Is Photonic

Computing with light could slash the energy needs of neural networks

10 min read

This computer rendering depicts the pattern on a photonic chip that the author and his colleagues have devised for performing neural-network calculations using light.

Alexander Sludds

Think of the many tasks to which computers are being applied that in the not-so-distant past required human intuition. Computers routinely identify objects in images, transcribe speech, translate between languages, diagnose medical conditions, play complex games, and drive cars.

The technique that has empowered these stunning developments is called deep learning, a term that refers to mathematical models known as artificial neural networks. Deep learning is a subfield of machine learning, a branch of computer science based on fitting complex models to data.

Keep Reading ↓ Show less