Effective monitoring and anomaly detection are fundamental prerequisites for safeguarding the efficiency, integrity and reliability of pipeline systems. Here, we explore both physics-based and machine-learning approaches for operational asset monitoring and anomaly detection, as well as evaluate their performance and appropriateness across a selection of analytical challenges.
Specifically, we look at identifying and quantifying anomalies in pump performance and orifice plate alignment accuracy relating to work completed for British Pipeline Agency (BPA) and a prominent UK gas operator.
In this paper, we assess each method and its trade-offs and present their effectiveness as monitoring and anomaly detection approaches. We conclude that machine-learning in isolation is no replacement for engineering and physics expertise, so delivering physics-based insights overlaid with machine-machine learning is the best and most practical approach.