Leak Detection System for Pipelines in Real-Time

A research team from the Korea Institute of Machinery and Materials, an institute under the Ministry of Science and ICT’s jurisdiction, has developed a noble technology that uses acoustic sensors to locate pipeline leaks in real-time from a distance.

The creation of a new device that can detect pipeline leaks in real time and locate their position is a substantial national achievement. This technique is critical for preventing severe mishaps in manufacturing sites.

The KIMM research team has created a device that can quickly locate a pipeline leak by strategically placing four-channel acoustic sensor modules capable of determining the direction of the sound. By combining location data from a three-dimensional virtual world with security cameras within the system, facility workers may quickly detect leak locations and respond efficiently within the critical first hour of an incident.

Previously, leakage detection was based on acoustic pressure, which measures sound magnitude. Determining the exact position of a leak requires proximity, which might jeopardize worker safety. This strategy also required costly equipment with over 100 channels of sound sensors per module.

The KIMM's novel technology is built around a signal-processing approach that uses acoustic intensity. This approach significantly minimizes detection errors in situations with strong reflections or reverberations by using a statistical algorithm to remove anomalous data.

It enables leak detection with an error margin of less than one meter (estimated error within three degrees). The position can be validated in a three-dimensional virtual area within the system, allowing for a faster reaction to leakage-related incidents.

This revolutionary approach minimizes the size of the required equipment, reducing the number of acoustic sensors per module from more than 100 to just four. When compared to typical systems, this change reduces equipment costs by more than 30 %. Furthermore, with the capacity to monitor leak detection results in real-time from a distance, leaks can be recognized early in an accident, allowing for swift intervention.

This three-dimensional acoustic intensity-based leak detection technology is incredibly valuable for pinpointing the origins of noises at various sites and analyzing their characteristics, thus aiding in determining the causes of accidents. The technology can expanded to include preemptive response to incidents and accidents, and can also be applied to a broad spectrum of industrial sectors, including alarm and surveillance systems for crime prevention.

Jung-Han Woo, Senior Researcher, Korea Institute of Machinery and Materials

This research was made possible by the “dArtificial intelligence based predictive diagnosis and damage management technology” project, which is one of the KIMM’s basic projects.