Rethinking Well Integrity for Sustainability: Adopting a Risk Based Approach

Abstract

Loss of containment (LOC) issues are every oil and gas operator's nightmare. These low frequency, high consequence spill-events have the potential to adversely affect the environment, an operator's financial health and public perception. For operators with mature aging assets and a sizeable well count, the probability of a LOC event is more likely. As the industry moves towards increased efficiency, there is a growing need to get more well integrity projects completed despite competition from oil-generating projects in the company-portfolio. This has necessitated re-thinking our response to the fundamental question of "how can we work within the existing constraints to ensure a robust well integrity program that protects people, the assets and the environment from a LOC event?" We show in this paper, a risk-based approach to solving this challenge. With over 1000 well strings, the starting point was having a comprehensive well integrity management system underpinned by a robust database that contained the test records of all well-related safety devices. A swiss-cheese model was then applied to analyze each device (Downhole safety valve, Wellhead valves and Casing valves) as a layer of protection considering possible hydrocarbon flow-paths. With this new methodology, each safety device was assigned a risk-factor denoting its relative importance in preventing the occurrence of a LOC event. Multiple safety device failures (e.g. combined failure of downhole and wellhead valves) had extra penalty assigned since holes lining up in the swiss-cheese would allow a catastrophic event pass through undetected. The risk factors were then summed up for each well to generate a risk-index that was used to compare wells and prioritize barge intervention activities accordingly. The key message is that catastrophic events typically require multiple safeguard failures and we can significantly reduce the chances of its occurrence by applying a risk-ranked approach to well integrity.

A novel machine learning model for autonomous analysis and diagnosis of well integrity failures in artificial-lift production systems

“…Zhao et al (2019) used hierarchical Bayes analysis to assess WI failures. Adeyinka et al (2020) assessed safety devices related to WI using a Swiss-cheese model. Recently, Yakoot et al (2021a) have reviewed a large body of literature.…”