Natural Resources Research最新文献

The primary objective of this study was to evaluate the effectiveness of three machine learning techniques in the confidence categorization of mineral resources within a copper deposit in Peru: extreme gradient boosting (XGBoost), random forest (RF), and deep neural network (DNN). To achieve this, geostatistical and geometric datasets were employed to categorize mineral resources into measured, indicated, and inferred categories. The dataset included ordinary kriging estimates, kriging variance, average distances, the number of composites, the kriging Lagrangian, and geological confidence. This dataset was used to train the models, followed by the application of smoothing techniques to the initial classification results to ensure a spatially coherent representation of the deposit. The results indicate that the RF model achieved the highest overall accuracy (94%), categorizing 1403.70 million tons (Mt) as measured resources (average grade of 0.43%), 2230.58 Mt as indicated resources (average grade of 0.33%), and 2225.08 Mt as inferred resources (average grade of 0.31%). XGBoost classified a slightly higher tonnage of measured resources (1412.35 Mt) with average accuracy of 91%, while DNN excelled in inferred resources, classifying 2254.64 Mt with accuracy of 93%. Smoothing improved the transitions between categories, reducing discontinuities and providing a more coherent representation of the deposit. The study concluded that machine learning techniques are robust and accurate tools for mineral resource categorization, particularly in geologically complex deposits.

Coal deformation-induced by adsorption/desorption is dynamic and anisotropic, influenced by various factors, such as pressure, temperature, and gas type. This paper investigates the dynamic deformation of coal during the adsorption–desorption process and analyzes the anisotropic and hysteretic characteristics. Results show that maximum deformation is reduced by approximately half with every 10 °C increase above 40 °C, and nearly doubles with each 1 MPa pressure increase. The swelling of CO₂ at adsorption equilibrium is twice that of CH₄, and almost 4 × that of N₂. During desorption, shrinkage and desorption gas are approximately linear. Anisotropy coefficients increase initially, then decrease with adsorption, stabilizing around 2. During desorption, anisotropy coefficients generally decrease. The anisotropy coefficient of CO₂ is higher than that of CH₄ and N₂, and all show a tendency to increase with equilibrium pressure. Cumulative hysteresis deformation decreases with the increasing temperature, even reversing at higher temperatures. CO₂ exhibits significantly higher hysteresis than CH₄ and N₂. These findings offer valuable insights for engineering applications.

The efficient characterization of heterogeneous carbonate reservoirs remains a significant challenge due to complex depositional environments and diagenetic alterations. While traditional methods like electrofacies analysis and clustering techniques offer inherent benefits, they often yield incomplete or conflicting results if used solely. This paper suggests an integrated study using petrophysical, geological, and statistical analyses to improve reservoir characterization. The proposed approach was applied to a carbonate reservoir case study of a gas field in South Iran. Well-log data and core samples were employed for detailed petrographic and petrophysical analyses. Electrofacies analysis using multi-resolution graph-based clustering (MRGC) identified five distinct electrofacies. Clustering techniques, including K-means and Gaussian mixture models (GMMs), were applied to petrophysical data to delineate similar zones. The Silhouette coefficient was used to evaluate the quality of the clusters. Results showed strong correlation between electrofacies 5 and clusters 4 (from K-means) and 5 (from GMMs), implying the best reservoir properties. This integrated approach suggested a more accurate assessment of reservoir quality attributes (e.g., porosity and water saturation) and highlighted the importance of dolomitized ooid grainstone in controlling hydrocarbon accumulation. This study provides a comprehensive framework for efficiently characterizing heterogeneous carbonate reservoirs by combining petrophysical, geological, and statistical methods. This integrated approach, validated through its successful application in similar reservoir studies, enables a more accurate assessment of reservoir quality attributes such as porosity and water saturation. By leveraging the complementary strengths of these methods, the approach ensures a comprehensive understanding of reservoir heterogeneity and its impact on hydrocarbon accumulation. Additionally, it is beneficial for improving reservoir modeling, enhancing hydrocarbon recovery, and reducing exploration risks.

Hot dry rock undergoes cyclic temperature variation during an enhanced geothermal system (EGS) operation, resulting in variations in reservoir rock’s transport properties and subsequently influencing the heat extraction efficiency of EGS. Therefore, the subject of this study was to systematically investigate the effect of cyclic heat treatment on the transport properties of granite, commonly employed in EGS, through the analysis of P-wave velocity, density, and scanning electron microscopy images. Besides, the effect of changes in the granite transport properties on EGS operation was also comprehensively discussed. The results indicated that the cyclic heat treatment led to an increase in granite permeability and a reduction in thermal conductivity. These changes primarily occurred due to the initiation and propagation of microcracks within the granite. Notably, higher-temperature heat treatments exhibited a more pronounced impact on granite properties. Additionally, a significant shift in the granite properties was observed within 450–550 °C, serving as a threshold temperature in this study. Due to the Kaiser memory effect and the blocking effect of the pre-microcrack on the subsequent microcrack, the effect of heat treatment on the properties of granite mainly came from the first heat treatment. Finally, the relationship models between heat treatment temperature and transport properties damage factors were obtained by fitting literature data.

The deterioration of coal strength caused by geological conditions of high gas in deep mines and disturbance from mining operations is one of the elements that influence the incidence of dynamic disasters like gas outbursts and rock bursts. To study how gas pressure and cyclic loads interact to determine the mechanisms and phenomena of coal dynamics, the split Hopkinson pressure bar apparatus was used to perform cyclic impact test on coal samples to investigate the mechanical behavior of gas-bearing coal samples under cyclic dynamic load and gas pressures. The findings indicated that there are three stages in the stress–strain evolution of gas-bearing coal: linear elastic stage, plastic stage, and post-peak stress attenuation. As cycle time grows, the peak stress and attenuation stress of the coal samples decrease, while the maximum and peak strains exhibit a general increasing trend. Under the impact of dynamic load, the macroscopic damage form of the coal sample is mainly a macroscopic crack, and the microscopic examination revealed that the coal samples interior crystal was primarily a trans-granular fracture. By considering dynamic load, gas pressure, and number of cycles, the test results can be more accurately verified by the mechanical damage constitutive model. Finally, based on cyclic dynamic load and gas pressure, the proposed fatigue prediction model of gas-bearing coal can better anticipate coal samples dynamic load-bearing capability.

Optimizing hydrocarbon recovery in the Illizi Basin requires precise reservoir characterization. Traditional methods face challenges in efficiently handling large datasets from multiple wells. This paper employs principal components analysis (PCA) to evaluate the petrophysical properties of the reservoir intervals (IV-3, IV-1b, IV-1a) using wells P8, P4, and P6, situated in the northern, center, and south of our reservoir, respectively. PCA reduced the dimensionality of the data, while preserving original information, facilitating the analysis of the reservoir's geological and sedimentological features. The results showed that unit IV-3 has the highest average porosity (average NET porosity) and the lowest average water saturation (average PAY log sw) across all wells, indicating significant hydrocarbon production potential. In contrast, units IV-1b and IV-1a exhibited higher water saturations, suggesting less favorable conditions for hydrocarbon extraction. Strong negative correlations between petrophysical properties and water saturation in unit IV-3 highlighted its potential for hydrocarbon production. PCA correlation circles illustrated these relationships, with unit IV-3 showing predominantly hydrocarbon saturation, Unit IV-1b exhibited mixed saturation, whereas unit IV-1a was characterized by high water saturation. These findings demonstrate the effectiveness of PCA in guiding hydrocarbon resource management and exploitation strategies in the Illizi Basin; therefore, we recommend prioritizing drilling in zones with optimal reservoir properties, as identified through PCA. These zones are likely to have higher porosity, permeability, and lower water saturation, we also recommend Considering implementing suitable enhanced oil recovery techniques, such as waterflooding, polymer flooding, or gas injection, to improve recovery factors, especially in low-permeability zones. Finally, we recommend implementing a robust monitoring system to track reservoir performance and adjust production strategies as needed. This may involve real-time monitoring of pressure, temperature, and flow rates. These recommendations, can significantly enhance hydrocarbon recovery from unit IV-3, maximizing economic benefits, while minimizing environmental impact. This study demonstrates the practical application of PCA in reservoir characterization and provides valuable insights for optimizing field development and production strategies in the Illizi Basin.

Mineral prospectivity mapping (MPM) can be deemed a binary classification task, with classifiers trained and validated on labels indicating the presence or absence of the targeted mineralized zones. Using economically viable mineral deposits as positive labels could, in theory, yield prospectivity models with geometallurgical reliability, thereby aiding land management and decision-making. The inherent scarcity of economically viable deposits, however, ultimately affects MPM products. The positive class label, therefore, often requires augmentation with either mineral occurrences (i.e., mineralized sites lacking economic viability) or synthetically generated labels. This paper examines how augmented positive labels and different negative label selection procedures geospatially represent economically viable mineral deposits and affect deep learning-based MPM’s classification performance and its spatial selectivity (i.e., MPM’s capability to efficiently narrow the exploration search space). To achieve this objective, large ensembles of deep learning classifiers were trained and validated with diverse combinations of positive and negative labels. Two positive class label sets were created by augmenting mineral deposits with either synthetic labels, generated using generative adversarial networks, or mineral occurrences, paired with distinct negative label sets selected based on (1) locations distant from known mineral deposits, (2) areas geospatially dissimilar to known mineral deposits, and (3) mineralized areas unrelated to the targeted style of mineralization, resulting in six unique class configurations. This study ultimately provides insights into how different label sets affect MPM's classification performance and spatial selectivity. The results indicate that selecting negative class labels from geospatially different localities enhances classification performance and MPM's spatial selectivity compared to other negative label selection procedures.

Machine learning (ML) is increasingly being used in geosciences for complex classification tasks. Mica minerals are commonly found in deposits of precious metals, rare metals, and rare earth elements, including tungsten, tin, lithium, and copper, among others. These minerals can provide insights into the formation environment and age of various deposits. While ML has been applied mainly for optical recognition and compositional analysis of mica, its use for classification of deposit types and mineralization types remains underexplored. This study aimed to fill this gap by developing a stacking multi-classification model, which integrates multiple ML algorithms, and logistic regression as the meta-model. Trained with a dataset of 3479 and 4005 mica major element compositions, both models achieved 0.99 accuracy on the test set. Precision, recall, and F1-scores were all reported at 0.99, indicating excellent classification performance. Feature importance analysis revealed that elements such as F, MgO, FeO, MnO, and Al₂O₃ are crucial for classification, reflecting distinct geological conditions across different types of ore deposits. Copper and gold deposits typically form around 700 °C under high oxygen fugacity and low fluorine fugacity, while W and Sn deposits form in the temperature range of 600–700 °C with varying oxygen fugacity. Lithium and beryllium deposits form at temperatures ranging 500–650 °C, exhibiting moderate oxygen fugacity and a wide range of fluorine fugacity. This paper presents a robust model for classifying deposit types and mineralization types based on mica composition and emphasizes the strong link between ML outcomes and geological characteristics.

Shale gas reserves represent a significant source of natural gas, but unlocking their full potential depends on effective hydraulic fracturing. This research investigates the application of machine learning (ML) techniques to predict fracability index (FI), offering a faster and more cost-effective alternative to traditional experimental methods. Focusing on the Upper Ordovician Wufeng to Lower Silurian Longmaxi Formation in the Weiyuan shale gas field, Sichuan Basin, China, this study employed deep neural networks that integrate two metaheuristic algorithms—genetic algorithm (GA) and particle swarm optimization (PSO)—with the back-propagation technique. These combined algorithms—termed GABPNN and PSOBPNN—were utilized to predict the FI. Model performance was assessed using three metrics: R², RMSE, and MAE. The GABPNN achieved R², RMSE, and MAE of 0.97531, 0.024754, and 0.0042875, respectively, while the PSOBPNN yielded values of 0.97494, 0.024938, and 0.0048962, respectively. Notably, when predicting FI values for the test well, the PSOBPNN model attained a R² of 0.99848, and the GABPNN model achieved a R² of 0.9993, indicating exceptional predictive accuracy. Both models demonstrated nearly perfect prediction accuracy for FI in the testing dataset, underscored by their high R² values. Importantly, the GABPNN model exhibited superior capability in mitigating overfitting, a common challenge in ML applications. Overall, the GABPNN and PSOBPNN models offer effective alternatives for assessing the fracability of shale gas reservoirs. By facilitating the identification of sweet spots for fracturing, these ML-based approaches have the potential to optimize operations in shale gas reservoirs.

Handling the micro-occurrence mechanisms of fluids is vital for the exploitation of shale gas. As the research hotspots shift towards the deep strata, the gas storage and transport capacity in shale relies to a great extent on the nanostructure. In this work, the grand canonical Monte Carlo and molecular dynamics simulations were performed to systematically study the adsorption and diffusion behaviors of water and methane in illite pores of marine shale. We aimed at providing a molecule-level insight into the thermodynamic and kinetic properties of fluids. The results demonstrate that water molecules tend to form two adsorption layers on each side of the illite surface in micropores. Specifically, the adsorbates are preferentially distributed between K+ and adsorbed above the tetrahedral silicon oxide layer through the hydrogen bonds. With the addition of methane in the system, the second adsorption layers of water disappear. Meanwhile, the density of free water at the pore center decreases and displays some small fluctuations. The variation in burial depth is mainly manifested by the controlling effects of temperature on the fluids. In general, it is manifested as a decrease in the adsorption capacity and an increase in the diffusion ability under the deep geological conditions. In this paper, the molecular dynamics simulation is shown to be an efficient and effective tool to further improve microscopic theory of the gas–water enrichment in shale nanopores.