The paper addresses the specific issues of leak monitoring in transportation pipelines under transient operating conditions. Various pipeline leak detection methodologies are reviewed in the context of transient leak detection.
The model-based leak detection system (LDS) incorporates a continuously running real-time simulation based on the enhanced fluid mechanical and thermodynamic model („virtual pipeline“). If a leak occurs, alarm is usually triggered by comprising the measured and calculated (mass) flows. Since the pipeline operation is an inher-ently transient process the transient model delivers considerably better results than simple solutions based on mass flow balance.
Still, leak monitoring under non-stationary operating conditions represents a chal-lenge for any leak detection system. Sharply changing operating conditions produce short-time fluctuations nearly in all measured characteristics making it difficult to trace the hydraulics of the pipeline. In case of large leaks or ruptures, the reaction time is critical. The leak must be indicated as soon as possible to avoid extensive losses.
Recently, a new version of the pipeline application software LEO-Pipe® of MAGNUM Automatisierungstechnik GmbH was installed on several pipelines typically operating under unsteady flow conditions. In this version, several enhanced leak detection methods run independently and parallel to each other, thus providing reliable and sensitive leak monitoring under all operation conditions.
A set of on-site experiments were conducted on one of these pipelines to demon-strating the ability of the implemented LDS to detect and locate leaks in the presence of transients of various types.