Abstract Technology has become an essential part of the oilfield industry, with Oil Operator companies focusing on the deployment of innovative initiatives and leveraging the benefits of their implementation. The demand for automation has increased significantly in both the Upstream and Downstream segments of the industry. Artificial Intelligence (AI)-driven processes are particularly relevant in this context. This work highlights the application of AI in characterizing and understanding oil reservoir rock properties. Traditionally, the integrated analysis of well log data is done manually and can take days to complete, but a fully automated machine-learning-driven approach can perform this task in a much shorter time. In the Petrophysics industry, Well-Bore Image (WBI) logs require extensive analysis due to their detailed geological content, which makes them more time-consuming to interpret compared to other logs. Ensuring the timely availability of high-quality outputs is crucial for the efficient design and execution of subsequent well operations, as demonstrated in the case study presented in this paper. AI-generated results are both time-efficient and detailed, providing sharp and clear outputs for operational needs. Comprehensive log data acquisition, including Well Bore Imaging (WBI), was performed in candidate horizontal lateral wells situated in heterogeneous carbonates of Lower Cretaceous age. The captured information-spectrum encompassed triple-combo, sonic, and nuclear magnetic resonance logs, as well as drilling, mud-log, and core data from the pilot offset and horizontal producer well-bores. Machine learning was utilized to execute WBI-based textural heterogeneity analysis. AI-driven key deliverables included quantifying diagenetic porosities and extracting rock permeability indices. These results were subsequently integrated with other down-hole information to develop a thorough understanding of dynamic flow outcomes. This comprehensive approach was crucial feed for designing an effective Lower Completion strategy within the required timeframe. AI-driven analysis has reduced the time required for tasks while maintaining quality. What used to take days or weeks can now be completed in hours. Cost optimization has also been achieved as a result of reducing dependency on contractors. The results have provided better insights into production flow profiles. AI-based outputs showed stronger correlation with flow logs compared to conventional rock property outputs, enhancing understanding of reservoir behaviour. Utilizing AI-derived analytical results in operational decisions has become the best practice in our organization. One of the practical applications includes aiding in the design of Lower Completion for newly drilled wells, where actual or test data has not yet been obtained. It is important to note that AI machine learning is an iterative process, expected to improve over time with more testing and feedback. The refined workflow utilizes supervised machine learning and computer vision to analyse electrical well-bore images at each depth, enhancing texture analysis and image porosity evaluation for permeability prediction. Through finite element Darcy experiments, the system simulates permeability variations along the wellbore under different boundary conditions. This AI-driven approach assesses the impact of various pore structures on overall permeability, applying machine learning algorithms to classify and quantify these effects accurately. This methodology provides depth-specific permeability predictions, supporting exploration and management in carbonate reservoirs.
Girinathan et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: