Abstract:
While there are currently several methods available for pipeline leak detection, very few can be implemented on large complex gas pipelines in an intuitive and cost-efficient way. This is due to extensive measurements and additional instrumentations being rudimentary for such implementations to reach the expected real-Time performances. This study explores a leak detection technology built for large, complex gas pipelines, created as a virtual simulation and a digital twin for real pipeline systems, that parallels the dynamic behaviors of the transient flows over time and space. The virtual system works as a mirror of the real systems to reveal the occurrences of abnormal events. Thus, leaks can be observed and located by analyzing discrepancies in the simulated-measured pressure profiles of real-Time transient models (PPRTM). The key assumption of this method is that discrepancies arising from the simulated-measured pressure profiles imply signatures of pipeline leaks. Such an assumption is verified by in-lab experiments as well as field trials. The underlying principles, assumptions, experiments, simulations, trials, and implementation of the PPRTM will be discussed in this study in detail. The application of this method on the YUJI natural gas pipeline system demonstrates that PPRTM is suitable as an efficient and effective implementation on large complex pipeline systems. The method provides a new way of continuously monitoring and locating the occurrences of leaks. It overcomes limitations of existing leak detection systems (LDS) since the pipeline system is monitored as whole and no extra measurements and instrumentation are required. Likewise, it is compatible with standard configurations of the supervisory control and data acquisition (SCADA) system and it is also capable of detecting and locating multiple leaks on the same pipeline. Β© 2023 ACM.
