Enhancing Facies Classification in Geological Studies Through Artificial Neural Networks: A Review
Year: 2024
Type: article
Abstract: Geological studies rely heavily on facies classification since it offers vital information for reservoir characterization and hydrocarbon exploitation. Because facies are inherently complex and heterogeneous, traditional approaches frequently struggle to categorize them effectively. Artificial Neural Networks (ANNs) have shown great promise in recent years for improving the efficiency and accuracy of facies classification. This review assesses ANN applications for facies classification in geological investigations critically and it begins by delineating the essential principles of facies classification and the constraints of traditional methodologies. Then ANNs' theoretical underpinnings and applicability to tasks involving the classification of facies was explored. The different architectures and configurations of ANNs used in geological research were also examined, as well as the benefits and difficulties of their use. The several ANNs architectures and configurations used in geological research are examined, as well as the benefits and difficulties of putting them into practice. In order to enhance the efficacy of ANNs in facies classification, the paper also addresses the integration of auxiliary data sources, such as well logs, seismic, and core data. Furthermore, the application of new developments in Deep Learning methods, including Recurrent Neural Networks (RNNs) and Convolutional Neural Networks (CNNs), to facies classification were discussed. To guarantee solid and trustworthy classification results, factors including feature selection, data preparation, and model assessment metrics were also taken into account. Lastly, the review highlights possible avenues for future research and breakthroughs in leveraging ANNs for enhanced facies classification, precision and effectiveness in geological studies.
Authors: Ofoh Juliana, Onyekuru Okechuwu, Ikoro Diugo, Opara Iheanyichukwu, Njoku I.O, Okereke Chikwendu, Akakuru Chigozie
Institution: Federal University of Technology Owerri
Language: English
Subfield: Mechanics of Materials
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6891712834