Search

Reservoir characterization struggles with identifying fractures, a typical imbalance classification problem. To handle this issue, a novel approach called fracture identification by sliding windows and you only look once (YOLO) (FISY) is proposed. FISY utilizes YOLO, a fast and accurate object detection neural network, to transform fracture identification into an abnormality detection task. Conventional well logs from specific interval formations are used as the detected image, and adjacent depth intervals enhance the analysis. Image samples are generated via a sliding window approach. YOLO detects the minority abnormality class (fractures), addressing data imbalance issues. Each depth sample undergoes multiple scans, and the probability superposition of the sliding window approach reduces forecast uncertainty. To assess FISY, a dataset from carbonate reservoirs in the Asmari Formation, Middle East, was used. The mean average precision (mAP) evaluation metric, combining precision and recall, assesses FISY’s performance in object detection. Among five-scaled FISY, XLarge demonstrates the highest mAP at 99.16%, with a rapid prediction time of 4 ms/image. In blind well tests, FISY surpasses fracture identification by sliding window (FIS) + Transformer (TF) (TF replacing YOLO in the FISY), synthetic minority oversampling technique (SMOTE) + support vector machine (SVM), and SMOTE + random forests (RFs) by over 13% in F1 score, showcasing its efficacy in fracture identification. To further assess FISY’s generalization across various lithologies, an experiment with a clastic reservoir dataset in the Ordos Basin in China reveals FISY’s good performance (mAP 87.5%).

Due to the capacity to quantify the complexity of a fracture network, fractal dimension (D) is widely applied in analyzing fracture network-related issues, such as connectivity and permeability. While the relationship between Dand individual properties of a fracture network has been extensively studied, Dis influenced by a combination of multiple attributes of the fracture network. Therefore, this work utilizes multivariate analysis to establish an equation for predicting D, taking into account various properties of the fracture network, namely fracture length, number, and orientation. Monte Carlo simulation is employed to generate a substantial number of fracture network models. Subsequently, relationships between the fractal dimension (D) and various properties are derived for three types of fracture networks with (1) invariant fracture length and random orientation, (2) exponential fracture length and random orientation, and (3) exponential fracture length and von-Mises orientation. The initial analysis focuses on the simplest relationship, wherein the fundamental formula of fractional expression is determined. Then the second and third relationships are obatiend though replacning by the fixed parameter in the first relationship by the distribution parameters of fracture properties. Correlation analyses between the predicted Dand the actual values reveal a remarkably high correlation (>0.99). To validate the established relationships, a fracture network obtained from geological outcrops is utilized. The results demonstrate the validity of the derived relationships. The utilization of these equations enhances the efficiency, practicality, and convenience of estimating fractal dimensions from fracture properties. As a result, the analysis of fracture network-related issues becomes more feasible and accessible.