Queensland project aims to monitor bridges with AI

A new Queensland-based project plans to use artificial intelligence (AI) to update engineers in real-time about how each bridge is coping under its load.

Australia has almost 60,000 bridges, with nearly 70per cent of them being older than 50 years. With vehicle-use exploding, these ageing structures are often carrying loads far beyond what they were originally expected to carry.

As part of a recent CSIRO entrepreneur program, the Skins Alive Infrastructure team placed high-tech sensors across the bridges, which immediately uploads data to the cloud, where an AI assesses the information and sends immediate alerts to engineers.

The eight-week ON Prime5 program was aimed at creating links between science entrepreneurs and real-world problems and several Brisbane entrepreneurs took part.

Skins Alive Infrastructure’s Govinda Pandey told Fairfax Media that nations have been assessing their bridges after the August tragedy in the Italian city of Genoa, where a bridge collapsed during a sudden, violent thunderstorm, sending vehicles plunging 45m into a heap of rubble below, killing 43 people.

“We thought it was really important that we use technology to solve this integrity problem associated with infrastructure,” Pandey continued.

“The problem with bridges was that they were such critical pieces of infrastructure and so expensive to replace, that they were frequently used far beyond their predicted lifespans. As a society we have to keep using them. We just don’t have the money to replace them.

“If you can actually measure the performance using sensors then you can get a better sense of what a bridge’s capacity is, as well as having access to information in real-time if there’s an anomaly or overloading event, or if there’s going to be an imminent collapse.”

According to Pandey, his company was not just limited to bridges, but could be applied to mining and any other large-scale infrastructure if the cost of high-tech sensors could be brought down.

“A single high-tech sensor could cost $500 and as an example, several hundred would be needed to place across the Story Bridge,” Pandey added.

“But using AI could reduce the need for multiple high-tech sensors and instead replace most with lower cost sensors that would then be coordinated by the AI.

“This could also prevent high-cost infrastructure maintenance from being conducted unnecessarily.”