Predicting green hydrogen production using electrolyzers driven by photovoltaic panels and wind turbines based on machine learning techniques: A pathway to on-site hydrogen refuelling stations

International Journal of Hydrogen Energy, Journal Year: 2025, Volume and Issue: 101, P. 1421 - 1438

Published: Jan. 8, 2025

Language: Английский

---

Research on seismic performance prediction of CFST latticed column-composite box girder joint based on machine learning

Construction and Building Materials, Journal Year: 2025, Volume and Issue: 460, P. 139811 - 139811

Published: Jan. 1, 2025

Language: Английский

---

Enhanced Data-Driven Machine Learning Models for Predicting Total Organic Carbon in Marine–Continental Transitional Shale Reservoirs

Sustainability, Journal Year: 2025, Volume and Issue: 17(5), P. 2048 - 2048

Published: Feb. 27, 2025

Natural gas, as a sustainable and cleaner energy source, still holds crucial position in the transition stage. In shale gas exploration, total organic carbon (TOC) content plays role, with log data proving beneficial predicting reservoirs. However, complex coal-bearing layers like marine–continental transitional Shanxi Formation, traditional prediction methods exhibit significant errors. Therefore, this study proposes an advanced, cost- time-saving deep learning approach to predict TOC shale. Five well records from area were used evaluate five machine models: K-Nearest Neighbors (KNNs), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Extreme (XGB), Deep Neural Network (DNN). The predictive results compared conventional for accurate predictions. Through K-fold cross-validation, ML models showed superior accuracy over models, DNN model displaying lowest root mean square error (RMSE) absolute (MAE). To enhance accuracy, δR was integrated new parameter into models. Comparative analysis revealed that improved DNN-R reduced MAE RMSE by 57.1% 70.6%, respectively, on training set, 59.5% 72.5%, test original model. Williams plot permutation importance confirmed reliability effectiveness of enhanced indicate potential technology valuable tool parameters, especially reservoirs lacking sufficient core samples relying solely basic well-logging data, signifying its effective assessment development.

Language: Английский

---

Total organic carbon content estimation for mixed shale using Xgboost method and implication for shale oil exploration

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Sept. 6, 2024

Language: Английский

---

Reservoir evolution and property prediction of the Paleogene Apollonia Formation, Western Desert, Egypt

Journal of African Earth Sciences, Journal Year: 2025, Volume and Issue: unknown, P. 105631 - 105631

Published: March 1, 2025

Language: Английский

---

Prediction of Lithofacies in Heterogeneous Shale Reservoirs Based on a Robust Stacking Machine Learning Model

Minerals, Journal Year: 2025, Volume and Issue: 15(3), P. 240 - 240

Published: Feb. 26, 2025

The lithofacies of a reservoir contain key information such as rock lithology, sedimentary structures, and mineral composition. Accurate prediction shale is crucial for identifying sweet spots oil gas development. However, obtaining through core sampling during drilling challenging, the accuracy traditional logging curve intersection methods insufficient. To efficiently accurately predict lithofacies, this study proposes hybrid model called Stacking, which combines four classifiers: Random Forest, HistGradient Boosting, Extreme Gradient Categorical Boosting. employs Grid Search Method to automatically search optimal hyperparameters, using classifiers base learners. predictions from these learners are then used new features, Logistic Regression serves final meta-classifier prediction. A total 3323 data points were collected six wells train test model, with performance evaluated on two blind that not involved in training process. results indicate stacking predicts achieving an Accuracy, Recall, Precision, F1 Score 0.9587, 0.959, respectively, set. This achievement provides technical support evaluation spot exploration.

Language: Английский

---

Data-driven total organic carbon prediction using feature selection methods incorporated in an automated machine learning framework

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 27, 2025

An accurate assessment of shale gas resources is highly important for the sustainable development these energy resources. Total organic carbon (TOC) analysis thus becomes fundamental understanding distribution and quality hydrocarbon source rocks within a reservoir. The elevation TOC often associated with presence rocks, indicating potential oil production. performed using laboratory methods, which can be time-consuming costly. Data-driven models have been successfully applied to model relationship between other constituents predict content. However, methods depend on extensive parameter adjustments that must carefully conducted in different sedimentary environments. In this context, Automated Machine Learning (AutoML) an alternative accurately predicting TOCs, saving fine-tuning steps development. This study aims develop AutoML strategy estimating well log data. procedure automatically preprocesses search best method parameters, reducing execution time. Among evaluated, Extremely Randomized Trees (XT) (R = 0.8632, MSE 0.1806) test set. proposed provides powerful data-driven method, allows real-world use assist data subsequent decision-making.

Language: Английский

---

Machine Learning-Based Sweet Spot Prediction for Lacuscrine Shale Oil in the Weixinan Sag, Beibu Gulf Basin, China

Marine and Petroleum Geology, Journal Year: 2025, Volume and Issue: 179, P. 107436 - 107436

Published: April 29, 2025

Language: Английский

---

An interpretable ensemble machine-learning workflow for permeability predictions in tight sandstone reservoirs using logging data

Geophysics, Journal Year: 2024, Volume and Issue: 89(5), P. MR265 - MR280

Published: May 30, 2024

Tight sandstone reservoirs exhibit strong vertical heterogeneity and complex pore structures, challenging conventional permeability evaluation methods based on well-logging data. Although rising machine-learning (ML) techniques have demonstrated excellent accuracy for industrial applications, the physics rationality within such a powerful “black box” remain less clear. Hence, reliable prediction would benefit from an interpretable ML-based workflow that could reveal controlling factors. To compare models examine underlying features, 16 different ML submodels are tested after data preprocessing, feature selection, hyperparameter optimization. By comparing fitting tuning time, light gradient boosting machine optimized by whale optimization algorithm, referred to as LGB-WOA, is determined be optimal model with best relatively short time. A field application demonstrates even in highly heterogeneous reservoir sections, LGB-WOA outperformed petrophysical being most consistent directly measured core samples ([Formula: see text]). The Shapley additive explanation values then used interpret predictions of our model. As expected, porosity curve exhibits highest importance among all input significantly contributing predictions. Conversely, wellbore diameter compensated neutron log contribute least need not subsequent improvements. These experiments provide method accurately assessing broader understanding characterization, paving way establishing more models.

Language: Английский

---

Deep learning for predicting porosity in ultra-deep fractured vuggy reservoirs from the Shunbei oilfield in Tarim Basin, China

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Nov. 28, 2024

Deep and ultra-deep carbonate reservoirs in China, which account for 34% of the country's oil gas reserves, pose significant challenges porosity prediction due to their complex geological features, including extensive burial depth, weak seismic signals, high heterogeneity. To address these challenges, this study develops an advanced deep learning approach specifically designed ultra-deep, fault-controlled, fractured-vuggy Tarim Basin. The utilizes a three-dimensional dataset applies Principal Component Analysis (PCA) select five key features from eight attributes. Additionally, phase-controlled constraints are incorporated into model. Using technology, model has been constructed. Validation using blind wells Shunbei oilfield shows that achieves 76% reduction Mean Square Error (MSE) compared traditional impedance inversion techniques, highlighting its predictive accuracy. Through SHapley Additive exPlanations (SHAP) analysis, attributes LAMBDA_AAGFIL PHASE_ANT identified as most influential, importance representing karst cave fracture structures within reservoir. These findings underscore innovation substantial improvement proposed method over conventional offering robust high-precision reservoirs.

Language: Английский

---