Search

Manual interpretation of geophysical logging data can be a tedious and time‐consuming task in the case of the nonlinear behavior of well‐logging signals. This study aims to enhance lithology classification of reservoir formations through advanced machine learning (ML) and deep learning (DL) techniques, introducing and comparing three novel algorithms, GrowNet, Deep‐Insight, and blender, against traditional models like random forest (RF) and support vector machine (SVM). Data from the South and North Viking Graben regions, encompassing 12 lithological facies, was preprocessed through cleaning, normalization, transformation, and imputation of missing values using regression models. The data set was enhanced with physic‐based features and balanced using SMOTE and NearMiss algorithms. Deep‐Insight converted tabular data into images for a convolutional neural network (CNN), significantly improving classification accuracy compared to conventional models such as decision trees (DTs). GrowNet and blender models leveraged hybrid approaches for enhanced performance. These hybrid approaches successfully addressed data imbalance and enhanced model learning, outperforming classic methods. The GrowNet and blender models for lithology classification successfully increased the penalty score and accuracy compared to the FORCE 2020 competition. Additionally, introducing the class prediction error plot visualizes multiclass classification performance more effectively than using a confusion matrix. These novel models in multiclass classification contribute to the petroleum industry by providing more accurate and reliable lithology classification, thereby improving reservoir characterization and exploration efficiency. [ABSTRACT FROM AUTHOR]

Gas reservoir development and the estimation of rock properties heavily rely on lithology classification, which can be difficult, time-consuming, and prone to errors. In this study, a novel deep learning-based approach has been developed for the rapid, accurate, and efficient prediction of lithology in a gas field from conventional well-log data. The well-logs, referred to as numerical well-logs (NWLs), are transformed into two-dimensional images through two proposed approaches: shallow images (SIs) and deep images (DIs). In these images, the pixels effectively represent the relationships between different logs. For this purpose, we developed residual convolutional neural networks (ResCNN) named SIs-ResCNN 2D and DIs-ResCNN 2D. The feed data for DIs-ResCNN 2D are images created and referred to as DIs, which are initially formed from a vector in which the order of logs is somehow repeated, ensuring that each pairwise combination occurs only once. This resulted in the incorporation of the connection between the logs within the pixels of the generated images, alongside the integration of unique binary combinations of the logs. We compared the proposed models, including DIs-ResCNN 2D, DIs-ResCNN 2D, and NWLs-ResCNN 1D with baseline methods such as random forest (RF), K-nearest neighbor (KNN), and support vector machine (SVM). Based on the evaluation metrics, DIs-ResCNN 2D outperformed the other proposed and baseline methods on the test dataset. A balanced DIs-ResCNN 2D model achieved 93% accuracy and F1-score of 80% on a test well, highlighting the importance of data balancing during CNN model training. [ABSTRACT FROM AUTHOR]