The performance of pipeline leak detection systems has been a hot subject for discussions in the past few years . In response to the public pressure and regulatory requirements, API published its recommended practice (RP) for continuous improvement of leak detection program . While the main objective of API 1175 is to provide guidance to pipeline operators for developing and maintaining management of pipeline leak detection programs, it encourages them to "go beyond" in order to achieve advancement in culture and strategy, overall performance, control center procedures, alarm management, roles and responsibilities, training and equipment maintenance. One important part of the leak detection program is the ongoing improvement.
Leak detection systems range from non-continuous to continuous, physical inspection to sensor-based monitoring, manual observation to computational pipeline monitoring (CPM) . The more widely applied systems are CPM ones based on volume balance, pressure monitoring, real-time transient model and statistical analysis . One common feature of the CPM methods is the use of pressure and/or flow measurements and SCADA (Supervisory Control And Data Acquisition) system. It is also the common limitation of such systems. To improve the leak detection performance of CPM systems, non-intrusive sensors have been developed.