AI improves crack detection in nuclear reactors

June 17, 2019
System in development at Purdue University uses AI to detect cracks from videos

As I write this, Futurama’s Bender is on my TV expressing his opinions about the flaws of us humans. Although he may take it a little farther than I would, it’s true that we don’t have the best natural detection capabilities. And when you’re talking about detecting structural flaws in something like a nuclear reactor, human error isn’t something with which I’d want to take a chance.

Luckily, technology is able to help us with this, and it’s sure to be much more helpful than Bender the Robot. A system in development at Purdue University is poised to help operators detect cracks and their severity in nuclear reactors, according to a recent article by Chris Adam.

Instead of technicians reviewing a video taken within the reactor frame by frame, the system uses artificial intelligence (AI) to do a frame-by-frame analysis. The AI system detects cracks and can track them from one frame to the next.

“This is a giant leap for inspection technology and being able to reduce accidents, deaths and maintenance costs,” said Mohammad R. Jahanshahi, assistant professor leading the research team at Purdue’s College of Engineering, in the article. “It lets the computer do the hard work, and then provides a human operator with quantitative information about the crack such as the thickness and the length of the crack.”

The operator can then make a decision based on the data and referenced frames within the video.

The system has been tested on 20 nuclear power plant inspection videos with positive results, Jahanshahi said in the article. It is also being considered to monitor buildings, bridges, dams and the like.

“Our system is smart and adaptive to allow an operator to use their own data,” Jahanshahi said in the article. “The computer can be reprogrammed based on that data to detect cracks within various structures and different materials.”