Multiphase Leak Detection: Moving Beyond API 1130 Metrics

Presentation Overview

Robust and Reliable Leak detection is difficult in pipelines carrying fluids of multiple phases (e.g. oil and water and gas). This difficulty is due to the continually changing nature of fluid properties with time and the pipeline's length and other reasons such as inappropriate models or fundamentally wrong approaches. Most complexities in the leak detection process arise when a leak happens in liquid pipelines in the presence of phase change or the lines carrying fluid congaing multiple phases- originate from wrongful application of models or approaches with common issues and predictable failure scenarios. 

The presentation will review pipeline leak detection challenges under degraded operating conditions such as pipeline operation under slack or column separation, which involves phase change and multi-phases in the domain simultaneously.  The presentation shows the importance of the phase change modelling ability of the CPM to look beyond API 1130 performance metrics and showcase studies to shed light on different complexity aspects of multi-phase leak detection and their treatment. For example, the presentation highlights what can go wrong with common approaches using machine learning algorithms even when trained with pure data while performing leak detection on multi-phase pipelines."