ISSN 2360-7955
Abstract
Asset integrity management is a complex task in the oil industry, particularly in surface production facilities. Pressure vessels, crucial for storing and transporting gases and liquids under high pressure, are susceptible to various damage mechanisms. Ensuring their integrity is vital to prevent catastrophic failures. This study aims to develop a multidisciplinary approach to asset integrity management (AIM) by integrating expert knowledge, survey questionnaires, and data analytics (non-parametric and machine learning). Two sets of questionnaires were developed in which the Kendall Coefficient of Concordance (W) ranked the features of AIM and the Random Forest Classifier ranked the NDT techniques. W showed that Corrosion (79), Pressure (84), Temperature limits (87), Vibration (93), Maintenance Strategies (96), Inspection Techniques (106) are the most important ‘AIR’ parameters while the Phased Array ultrasonic testing, Time of flight Diffraction, Acoustic Emission Testing, Eddy current testing, Pulsed Eddy current are the most important NDT techniques. The study provides a practical framework for controlling and minimizing incidents in oil and gas operations, ultimately contributing to improved safety and efficiency by providing insights into which features are most influential as regards AIR and NDT techniques.
Keywords: Asset Integrity, Data-Driven, Non-parametric Test, Machine Learning