Ultrasonic sensors can safeguard gas lines, says UBCO

Photo: UBCO

UBC Okanagan researchers are investigating a new method to monitor underground gas pipelines with high-tech sensors that can make it easier to find weaknesses, discrepancies and even a diversion in residential natural gas lines.

University student Abdullah Zayat says there's has been considerable research into diagnosis methods for steel pipes such as radiography, ultrasonic testing, visual inspection and ground penetrating radar, but little work has been done on the commonly used high-density polyethylene pipe, which is responsible for carrying natural gas to homes.

“Early detection of structural degradation is essential to maintaining safety and integrity. And it lowers the risk of catastrophic failure,” Zayat explains.

The student and his supervisor tested a technique that allows for inspection of HDPE pipes with ultrasonic sensors. This method limits the likelihood a gas diversions, not being measured in the metre.

“This tampering with the pipe poses many risks since it is unrecorded, violates pipeline quality standards and can lead to potential leaks and possibly explosions," explained Dr. Anas Chaaban.

"This can pose a significant risk to public safety, property and the environment in the vicinity of the altered gas line. Such diversions have been discovered in the past through word of mouth, leaks or unexpected encounters with an unrecorded natural gas pipe in a construction site.”

Zayat says that given the concealed nature of underground pipes, it's challenging to inspect them.

Using a non-invasive strategy, this method allows for the inspection of buried, insulated and underwater pipelines using ultrasonic sensors.

"UGW sensing is getting a lot of attention from the industry because of its long-range inspection capabilities from a single test location. They can inspect more than 100 metres of pipeline from a single location,” added Zayat.

The technology is still in the early stages, but Dr. Chaaban notes the majority of this current research involved the development and assessment of a deep-learning algorithm for detecting diversions in pipes.

According to research results, this method has 90 per cent accuracy when one receiving sensor is used and nearly 97 per cent accuracy when using two receiving sensors.