Surface pipeline leak detection using realtime sensor data analysis

Abstract

The most common type of Leak Detection System (LDS) is designed to detect leaks that generate a sufficient pressure variation which can be detected at either the inlet or the outlet sensors. However, the pressure variation from low-pressure leaks at locations far away from the inlet and outlet is dissipated before they arrive at these sensors. Thus, these leaks can continue for weeks before they are detected. This work developed a leak detection architecture which comprised of a pressure sensor installed in the middle of the pipeline segment. This sensor was found to be more sensitive to leak-induced pressure variations from leaks far away from the inlet and outlet and this was because leak-induced pressure variation was higher at the midpoint even when it had diminished to zero at the inlet or outlet. The work also developed a Leak Detection as a Service (LDaaS) platform, which utilizes the leak detection algorithm developed from this research and pressure values from the inlet and midpoint sensor to detect real-time pipeline leaks. The midpoint sensor utilizes an exception-based transmission protocol capable of extending the sensor's battery life. Operators can subscribe to the Leak Detection service by installing the midpoint sensor and transmitting the inlet and the midpoint pressure values to the platform. This platform will monitor the pipeline in real-time and detect both the high-pressure and low-pressure leaks which would have ordinarily been missed by the traditional LDSs.