A Robust Strategy of Geophysical Logging for Predicting Payable Lithofacies to Forecast Sweet Spots Using Digital Intelligence Paradigms in a Heterogeneous Gas Field

Natural Resources Research, Journal Year: 2024, Volume and Issue: 33(4), P. 1741 - 1762

Published: May 14, 2024

Language: Английский

---

Application of an Optimized PSO-BP Neural Network to the Assessment and Prediction of Underground Coal Mine Safety Risk Factors

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(9), P. 5317 - 5317

Published: April 24, 2023

Coal has played an important role in the economies of many countries worldwide, which resulted increased surface and underground mining with large coal reserves, such as China United States. However, is subject to frequent accidents predictable risks that have, some instances, led loss lives, disabilities, equipment damage, etc. The assessment risk factors mines therefore considered a commendable initiative. Therefore, this research aimed develop efficient model for assessing predicting safety using existing data from Xiaonan mine. A evaluating was developed based on optimized particle swarm optimization-backpropagation (PSO-BP) neural network. results showed PSO-BP network most reliable effective, MSE, MAPE, R2 values 2.0 × 10−4, 4.3, 0.92, respectively. study proposed mine assessment. can be adopted by decision-makers mines.

Language: Английский

---

Machine Learning Algorithms for Lithofacies Classification of the Gulong Shale from the Songliao Basin, China

Energies, Journal Year: 2023, Volume and Issue: 16(6), P. 2581 - 2581

Published: March 9, 2023

Lithofacies identification and classification are critical for characterizing the hydrocarbon potential of unconventional resources. Although extensive applications machine learning models in predicting lithofacies have been applied to conventional reservoir systems, effectiveness clay-rich, lacustrine shale has yet be tackled. Here, we apply well log data automatically identify Gulong Shale Songliao Basin. The were classified into six types based on total organic carbon mineral composition from core analysis geochemical logs. We compared accuracy Multilayer Perceptron (MLP), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Random Forest models. mitigated bias imbalanced by applying oversampling algorithms. Our results show that ensemble methods (XGBoost Forest) a better performance than other do, with accuracies 0.868 0.884, respectively. siliceous proposed best can identified F1 scores 0.853 XGBoost 0.877 Forest. study suggests effectively clay-rich logs, providing insight sweet spot prediction reservoirs. Further improvements model performances achieved adding domain knowledge employing advanced data.

Language: Английский

---

Prospect Evaluation of the Cretaceous Yageliemu Clastic Reservoir Based on Geophysical Log Data: A Case Study from the Yakela Gas Condensate Field, Tarim Basin, China

Energies, Journal Year: 2023, Volume and Issue: 16(6), P. 2721 - 2721

Published: March 14, 2023

This paper evaluated the oil and gas potential of Cretaceous Yageliemu clastic reservoir within Yakela condensed field lying in Kuqa Depression, Tarim Basin, China. The petrophysical properties interest zones area were characterized using geophysical logs from five wells. results reveal that gas-bearing are by high resistivity, good permeability (K) effective porosity (Φeff), low water saturation (Sw), shale concentration (Vsh), reflecting clean sand. distribution model showed these shales have no major influence on fluid saturation. average volume, porosity, hydrocarbon indicate Formation studied contains prospective properties. spatial parameters, rock typing (RRT), lithofacies analyzed cross plots litho (volumetric analysis), iso-parametric representations characteristics, cluster analysis, self-organizing feature maps, respectively. southeastern northeastern regions research should be ignored because their concentrations. sediments southwest northwest include most intervals considered for future exploration development fields study area.

Language: Английский

---

Quantitative Characterization of Shallow Marine Sediments in Tight Gas Fields of Middle Indus Basin: A Rational Approach of Multiple Rock Physics Diagnostic Models

Processes, Journal Year: 2023, Volume and Issue: 11(2), P. 323 - 323

Published: Jan. 18, 2023

For the successful discovery and development of tight sand gas reserves, it is necessary to locate with certain features. These features must largely include a significant accumulation hydrocarbons, rock physics models, mechanical properties. However, effective representation such reservoir properties using applicable parameters challenging due complicated heterogeneous structural characteristics hydrocarbon sand. Rock modeling sandstone reservoirs from Lower Goru Basin fields represents link between seismic diagnostic models have been utilized describe sands two wells inside this Middle Indus Basin, including contact cement, constant friable The results showed that sorting grain coating cement on grain’s surface both affected cementation process. According levels in ranged 2% more than 6%. established study would improve understanding for relatively high Vp/Vs unconsolidated under study. Integrating prediction elastic estimated data. velocity–porosity moduli-porosity patterns zones are distinct. To generate template (RPT) Early Cretaceous period, an approach based fluid replacement has chosen. ratio P-wave velocity S-wave (Vp/Vs) P-impedance can detect cap shale, brine sand, gas-saturated varying water saturation porosity Rehmat Miano fields, which same shallow marine depositional characteristics. Conventional neutron-density cross-plot analysis matches up quite well RPT’s expected detection sands.

Language: Английский

---

Knowledge-based machine learning for mineral classification in a complex tectonic regime of Yingxiu-Beichuan fault zone, Sichuan basin

Geoenergy Science and Engineering, Journal Year: 2023, Volume and Issue: 229, P. 212077 - 212077

Published: July 1, 2023

Language: Английский

---

Machine-Learning-Based Deformation Prediction Method for Deep Foundation-Pit Enclosure Structure

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(3), P. 1273 - 1273

Published: Feb. 3, 2024

During the construction of deep foundation pits in subways, it is crucial to closely monitor horizontal displacement pit enclosure ensure stability and safety, reduce risk structural damage caused by deformations. With advancements machine-learning (ML) techniques correlation analysis engineering, data-driven methods that combine ML with engineering monitoring data have become increasingly popular. These offer benefits such as high prediction accuracy, efficiency, cost effectiveness. The main goal this study was develop a method for predicting deformation pits. This achieved analyzing factors influencing foundation-pit incorporating historical cases reports. performance each model systematically analyzed evaluated using K-Fold cross validation. results revealed random forest outperformed other models. result test showed an R2 0.9905, MAE 0.8572 mm, RMSE 1.9119 mm. Feature importance identified depth structure, water level, surface settlement, axial force, exposure time most critical accurate prediction. structure had especially significant impact on deformation.

Language: Английский

---

Three‐dimensional structural modelling of structurally complex hydrocarbon reservoir in October Oil Field, Gulf of Suez, Egypt

Geological Journal, Journal Year: 2023, Volume and Issue: 58(11), P. 4146 - 4164

Published: April 18, 2023

The October Oil Field is structurally complex due to its presence in the system of Gulf Suez Rift Basin area, with last updated structural model developed 2012. Although 2012 defined general framework and reservoir architecture, many challenges arose during field development. current study focusing on elements affecting this giant update using newly processed 3D seismic survey, acquired data from drilled wells, associated different logging techniques. Several geological structure contour maps cross‐sections were generated help delineating understanding reservoir's extension. Based detailed correlation study, we able detect faults that affected detail, define their throw amounts directions, identify missed sections across studied area. This introduces a comparison between old scenarios show differences effect development plan recommendations. shows study's modified number, extension, location faults: has 17 faults, while 13 faults. main clysmic fault “F1” significant impact entire because it affects all wells. Furthermore, F3 F4 have ability create add compartmentalization within area study. revealed can support plans for Nubia motivate drilling, workover, dynamic operations assign opportunities proper location. model, there are at least three attic areas could increase oil production reserves avoiding any more failures.

Language: Английский

---

Porosity prediction from prestack seismic data via deep learning: incorporating a low-frequency porosity model

Journal of Geophysics and Engineering, Journal Year: 2023, Volume and Issue: 20(5), P. 1016 - 1029

Published: Sept. 1, 2023

Abstract Porosity prediction from seismic data is of considerable importance in reservoir quality assessment, geological model building, and flow unit delineation. Deep learning approaches have demonstrated great potential characterization due to their strong feature extraction nonlinear relationship mapping abilities. However, the reliability porosity often compromised by lack low-frequency information bandlimited data. To address this issue, we propose incorporating a based on geostatistical methodology, into supervised convolutional neural network predict prestack angle gather inversion results. Our study demonstrates that inclusion significantly improves predictions heterogeneous carbonate reservoir. The can be compensated enhance network's capabilities capturing background trend. Additionally, blind well tests validate considering constraint leads stronger generalization abilities, with root mean square error two wells reduced up 34%. incorporation training also remarkably enhances continuity prediction, providing more geologically reasonable results for characterization.

Language: Английский

---

Harnessing Advanced Machine-Learning Algorithms for Optimized Data Conditioning and Petrophysical Analysis of Heterogeneous, Thin Reservoirs

Energy & Fuels, Journal Year: 2023, Volume and Issue: 37(14), P. 10218 - 10234

Published: July 6, 2023

Petrophysical analysis is an industry-standard practice for reservoir evaluation as it provides critical inputs characterizing subsurface formations and estimating resource potential. Khadro/Ranikot Formation sands are proliferous producers in the Central Indus Basin, Pakistan. The demarcate potential intercalated sand shale layers that thin heterogeneous makes a challenging reservoir. Conventional petrophysical interpretation laborious does not produce up-to-mark results due to complexity, data limitations, associated uncertainties. Hence, emerging delicate machine-learning (ML) approach has been comprehensively applied analyze robustly interpret well log while addressing challenges. This case study entails thorough of quality, assessing several algorithms such least-squares support vector machines (one-class SVM), Random Forest Regressor (RFR), Extra Tree (ETR), Gradient Boosting (GBR), Decision Classifier (DTC), etc. compare their efficacy reliability. One-class SVM helps reduce outliers with great certainty, missing logs sonic (DT) density (RHOB) precisely predicted via GBR ETR 0.66 0.88 R2, respectively. providing reliable optimized quality suitable ML-based petrophysics. ML worked on these augmented by dividing into 60% training 40% testing. outperformed rest models correlation 0.99 0.91 among conventional results. Likewise, RFR performed exceptionally water saturation modeling, expressing highest 0.93 correlation. Finally, DTC modeled facies best 91% accuracy 0.935 F1 measures at blind well. Excellent calibration >85% met estimates obtained predictive model compared methods. comprehensive offers cost-effective robust workarounds modern formation minimal uncertainty resource-efficient multiwell within complex reservoirs sets stage further research ecosystem.

Language: Английский

---