Enabling Automatic Pipeline Leak Detection

Hifi Engineering is currently monitoring nearly 3,000 km of pipeline using Fiber-Optic Cables placed along their length. Tiny scratches, called Fiber Bragg Gratings (FBG), are inscribed into the fiber every 25m. Since the optical length of the fiber depends on the surrounding pressure, these gratings allow the fiber-optic cable to be used as a series of microphones, one for each grating. Hifi Engineering currently uses a series of threshold-based alarms, each of which triggers an assessment by a human operator, to detect alarm conditions. They are actively developing machine-learning based methods to automate this process as much as possible. This project will improve several of the fundamental signal processing steps used to reconstruct the sounds present at each of the detectors. This will provide the later leak detection algorithms with better data, hence increasing the reliability and accuracy of the system.

Faculty Supervisor:

Arne Dankers;David Westwick

Student:

Partner:

Hifi Engineering Inc

Discipline:

Engineering

Sector:

Mining

University:

University of Calgary

Program: