Deliverable 5.4: Workflow for reliability assessment

Abstract

This report presents a workflow for probabilistic fault stability assessment in CO2 storage, emphasising uncertainty quantification, parameter ranking, and Bayesian updating for critical inputs. Using North Sea's Horda platform cases, including the Vette fault zone in Smeaheia, this study demonstrates the value of the suggested methods. This study highlights the usefulness of the probability of failure (Pf) as a reliable measure for stability assessment, particularly when traditional methods present conflicting results. It also highlights the effectiveness of the Sobol sensitivity analysis for input ranking and understanding parameter interactions, which is crucial for resource prioritisation and monitoring strategy in early field development. Furthermore, the report illustrates how the Bayesian approach can enhance the accuracy of stress prediction by leveraging data from geologically similar sites and acknowledging site-specific heterogeneity, significantly aiding in geomechanical risk assessments.