Ai - Powered Fracture Gradient and Caprock Evaluation to Define Optimal Plug Depths for Efficient Well Abandonment

The growing number of wells approaching end-of-life has intensified the need for robust, standardised, and efficient solutions for plug and abandonment (P&A). As operators transition mature assets toward permanent abandonment, there is increasing emphasis on ensuring that subsurface barriers are designed and verified with the same level of technical assurance as drilling and completion operations. Achieving this requires precise knowledge of formation behaviour under abandonment conditions, particularly in identifying competent caprock intervals and determining the Minimum Setting Depth (MSD) for long-term well integrity and environmental containment. A key challenge in this process lies in the complexity and variability of subsurface data. Conventional workflows for determining MSD and caprock suitability depend heavily on manual interpretation of petrophysical and geomechanical datasets. This traditional approach is often time-consuming, labour intensive, and prone to inconsistencies among interpreters, especially when dealing with multi-field or legacy data where log quality and data density vary significantly. Such variability can lead to non-uniform decisions on plug placement and barrier design, increasing the uncertainty in post-abandonment well performance. To overcome these limitations, the Artificial Intelligence (AI) P&A framework was developed as a data-driven digital platform that integrates machine learning (ML) models with petrophysical and geomechanical data to deliver a consistent, automated, and scalable solution for abandonment design. The framework employs a hybrid approach that combines geomechanical equations with ML-based pattern recognition to automate critical tasks such as lithology classification, rock property prediction, and MSD determination within a unified AI-assisted workflow.