 |
|
PaintSquare Daily News is available in three editions: Industrial Only Commercial Only Combined Industrial & Commercial Your free subscription also includes: Special topical editions and Weekend Brief—a recap of the week’s top news and some great in-depth reads. Subscribe now: paintsquare.com/psdn |
CTDOT Evaluating Bridges Using AI Research
The Connecticut Department of Transportation and the Federal Highway Administration recently awarded a research grant to a University of Hartford professor to evaluate and predict infrastructure needs for the state’s 5,000 bridges using artificial intelligence.
Clara Fang, professor of civil engineering in UHart’s College of Engineering, Technology and Architecture, received a $238,000 research grant last month for the two-year project.
About the Research
Using AI, Fang’s team will reportedly be able to better predict the bridge performance and deterioration process. The university reports that creating a more advanced prediction model will allow the state to further plan for bridges for years and decades to come.
“The CTDOT has ‘big data’ on all of our bridges in a large database from their inspections and ratings. There are 20 million inspection records from the past 30 years,” Fang explained. “Using AI will allow us to understand how a bridge performs, and comprehend patterns to try and see how it will perform in the future and learn information about its lifespan, while predicting its next major rehab needs.”
Fang adds that their work will allow the state to take a more proactive approach, rather than a reactionary one, to ensure bridges are safe and that work done to the infrastructure is efficient and cost-effective.
“The model will monitor the bridge deterioration process and predict when maintenance and rehabilitation will be required, reducing the need for reactive maintenance and helping to reduce the risk of bridge failure,” Fang said.
According to UHart, the project will use AI to acquire knowledge from the state’s existing large data sets and then create algorithms the state can use in the future to predict bridge conditions and needs.
The program model will also take into account bridge features such as geometry, design, construction, service, cost, weather, traffic dynamics and many other characteristics. The model will then study the intricate connections between various bridge features and bridge performance.
The FHWA reports that there are about 600,000 bridges in the United States, and more than 25% of the bridges are either structurally deficient or functionally obsolete and in need of maintenance, rehabilitation or replacement.
The research team includes Daniel Jimenez Gil, a UHart civil engineering student who has been working for the project in data processing and machine learning modeling; co-investigator Saleh Keshawarz, professor and chair of the Department of Civil, Environmental, Biomedical Engineering in CETA; Yang Yang, associate professor of civil engineering in CETA in a consultant role; and two research fellows from University of Auckland, New Zealand.
Recent Bridge Monitoring Research
Last month, an Iowa State University researcher created a multifunctional stretchable strain sensor that changes colors when covering a crack in a steel component.
Simon Laflamme, a professor in the Department of Civil, Construction and Environmental Engineering, worked with Harvard University to develop the stretchable polymer that changes color from white to blue when a crack in a structure is identified.
According to Iowa State, the polymer was also modified into a piece of flexible electronics that changes electrical properties upon a change in geometry, therefore augmenting the optical color feedback with an electrical feedback.
In November, a study conducted by a team involving researchers from the Massachusetts Institute of Technology suggested that mobile phones could potentially collect useful integrity data while crossing bridges. Equipped with special software, mobile devices can reportedly capture the same kind of information about bridge vibrations that stationary sensors compile.
To study the “modal frequencies” of natural bridge vibrations in many directions, engineers can typically place sensors, such as accelerometers, on the structures. Changes in these frequencies over time may indicate changes in a bridge’s structural integrity.
The first part of research was reportedly conducted on the Golden Gate Bridge, using an Android-based mobile phone application to collect accelerometer data when the devices were placed in vehicles passing over the bridge. Researchers could then see how well those data matched up with data record by sensors on bridges themselves, to see if the mobile-phone method worked.
Researchers drove over the Golden Gate Bridge 102 times with their devices running, and the team used 72 trips by Uber drivers with activated phones as well. The team then reportedly compared the resulting data to that from a group of 240 sensors that had been placed on the Golden Gate Bridge for three months.
As a result, data from the phones converged with the bridge’s sensors, with 10 particular types of low-frequency vibrations a close match and five cases showing no discrepancy between the methods at all. Researchers estimate that mobile-device monitoring could add from 15-30% more years to the structure’s lifespan depending on the age of the bridge.
According to MIT, results found up to a 2.3% divergence between methods for certain modal frequencies over all 280 trips, and a 5.5% divergence over a smaller sample. The data suggests that a larger volume of trips could yield more useful data.
