Machine-learning modeling of pore pressure variation in geological CO2 storage reservoir

Ensemble Machine-Learning-Based Horizontal Permeability Prediction and Hydraulic Flow Unit Characterization of the Hugin Sandstone

Understanding permeability is essential for evaluating reservoir quality and field development planning. Reliable permeability estimation can reduce the uncertainty in reservoir characterization, particularly in intervals where core data are limited. As the industry relies on log-based interpretations and empirical correlations, the limitations of these approaches become apparent. Data-driven approaches offer a promising alternative to conventional empirical methods. The data set in this study comprises 252 samples with seven features derived from conventional well logs. Data preprocessing includes handling missing values, smoothing logs, feature engineering to add an extra input, and transformation with the Yeo-Johnson technique. A center moving average filter was used to reduce variance and improve data consistency. Ensemble machine-learning (ML) and baseline models were developed and evaluated using a 75-25 train-test split, four-fold cross validation, and model complexity assessment. Ensemble methods outperformed baseline models, with extremely randomized trees (ET) and random forest (RF) emerging as the most stable, achieving a mean R² of 0.93 and 0.89 and a low R² standard deviation (0.3). Multilinear regression (MLR) and artificial neural networks (ANNs) show limited accuracy, while gradient boosting (GB) and extreme gradient boosting (XGBoost) methods exhibit overfitting despite perfect training scores. Predicted kh values were compared with core data. Both linear and nonlinear empirical equations were derived using MLR, polynomial regression, and a power-law model. The power-law model (empirical equation) achieved an R² value of 0.79 and can therefore be used to estimate permeability. Additionally, a Gaussian mixture model (GMM) was used for unsupervised classification of hydraulic flow units (HFU) using the flow zone indicator (FZI), computed from the continuous permeability curve obtained from the best ML model. Thus, ML-based permeability prediction is an indispensable component of HFU modeling. The model identified three distinct flow zones, enabling HFU clustering and defining their corresponding petrophysical properties and depositional environments.

Performance and Limitations of Machine Learning Models for Groundwater Level Prediction Using Hydro-Climatic Variables

Accurate groundwater level (GWL) prediction is essential for sustainable groundwater management and resource planning. However, it is challenging in heterogeneous hydrogeological settings for physics-based models, particularly under limited subsurface characterisation. Machine learning (ML) techniques can capture complex spatio-temporal groundwater dynamics, complementing conventional modelling approaches. This study systematically evaluates the performance and limitations of ML models for GWL prediction in sandstone and mudstone formations, with emphasis on the influence of local hydrogeological conditions. Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), Support Vector Regression (SVR), and Random Forest (RF), along with their wavelet-enhanced counterparts, were applied to monthly hydro-climatic data from 11 observation wells in the Lower Otter Catchment, UK, covering 2011–2023. The final four years were reserved for validation. Time-series predictors were used because time-invariant and sparsely available geological parameters provide limited explanatory power for local-scale GWL dynamics. Their influence is implicitly reflected in observed GWL responses. Model performance was assessed using statistical criteria, including the coefficient of determination (R 2 ). Additionally, a new metric, the Data Difference and Trend Index (DDTI), was introduced to measure the proportion of simulated values matching observed trends within a predefined threshold (e.g. 0.5 m). Model performance was site-specific, with validation R 2 ranging from < 0.1 to > 0.9. This indicates the dominant influence of local hydrogeological conditions, with lower accuracy observed in partially confined, non-recharge-dominated, and river-disconnected wells. A 2-month time lag produced optimal model performance, reflecting the catchment’s characteristic response time to infiltration processes. Individual models occasionally outperformed ensemble averages, which showed fewer outliers. Wavelet transforms did not consistently enhance performance. Model efficacy varied seasonally, with validation R 2 markedly lower in summer (e.g. < 0.1) and higher in autumn (e.g. > 0.9). This emphasises key limitations of ML-based GWL prediction, including reduced reliability near lithological boundaries and strong sensitivity to hydro-climatic conditions, constraining model transferability. Overall, the findings highlight the value of moving beyond performance benchmarking to explicitly identify hydro-climatic and hydrogeological conditions under which ML models lose reliability, informing groundwater modelling and sustainable water management. Graphical Abstract This study evaluates the performance and limitations of machine learning (ML) models for predicting groundwater levels (GWL) in sandstone and mudstone formations using hydro-climatic variables across 11 observation wells in the Lower River Otter Water Body, UK. Four ML models, i.e. Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), Support Vector Regression (SVR), and Random Forest (RF), along with their wavelet-enhanced versions, were applied to monthly hydro-climatic data from 2011 to 2023, with the last four years reserved for validation. Model performance was assessed using the coefficient of determination (R 2 ), Root Mean Square Error (RMSE), and Nash-Sutcliffe Efficiency (NSE), and a novel Data Difference and Trend Index (DDTI), which quantifies the proportion of simulated data following observed trends within a defined threshold. Results indicate that predictive accuracy is highly site-specific, with local hydrogeological conditions strongly influencing outcomes. No model consistently captured GWL near interior boundaries where sandstone is confined by mudstone, and neither wavelet transforms nor model ensembles reliably improved performance. Seasonal variability also affected model efficacy, with the highest accuracy in autumn and the lowest in summer. Overall, the workflow highlights the limitations of ML for GWL prediction and provides insights for future hydrogeological modelling.

Prediction of Sand Production in Vertical Oil Well Using Supervised Machine Learning Models

Sand production has become a significant concern in the hydrocarbon recovery process from unconsolidated reservoirs which may result in equipment damage, flow restrictions, and costly operational downtime in vertical oil wells. Accurate prediction of sand production is vital to optimize well integrity. Conventional geomechanical and empirical models frequently fail to capture the highly non-linear interactions among reservoir pressure, multiphase flow rates, rock mechanical properties, and dynamic operating conditions. This study addresses the identified research gap by developing and systematically comparing four supervised machine learning classifiers for binary prediction of sand production occurrence using routine well-test data from a single vertical oil well in the Niger Delta basin. A total of 235 validated well test observations consisting of 19 recorded variables which include date and operational parameters such as production rates, pressure conditions, choke size, and fluid properties were pre-processed and analyzed. The target variable was formulated as a binary classification problem with the operational threshold sand rate > 0 lb/1000 bbl to enable early detection of any sanding event. Four machine learning algorithms including Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine were developed and evaluated following feature optimization and model tuning. Model performance was assessed using accuracy, precision, recall, F1-score, and ROC–AUC metrics. The results show that ensemble and kernel-based methods significantly outperform linear and single-tree models, with the Random Forest classifier achieving the best model prediction accuracy of 93.62%. This strong performance demonstrates the model's robustness in capturing the complex, nonlinear interactions governing sand production behavior. This study demonstrates that machine learning classifiers can be effectively utilized in a manner that enables proactive sand management strategies, including choke adjustment, artificial lift optimization and selective sand control deployment, ensuring a minimal risk of equipment failure and increasing overall well productivity and hydrocarbon production.

Permeability prediction through machine learning regression in gas hydrate bearing sediments of Krishna-Godavari basin, using NGHP-02 well log data

Prediction of Porosity and Permeability Using Well Log and Core Data: A Data-Driven Approach

Accurate prediction of porosity and permeability is very important for reservoir characterization and hydrocarbon extraction. Traditional workflow in the form of empirical correlations is usually difficult, time-consuming, spatially limiting, and entirely dependent on formation geology. The current study examines a different approach, which uses machine learning (ML) regression models based on well-log and core data. Three regression architectures including Random Forest, CatBoost, and K-Nearest Neighbors (KNN) were trained and validated based on a dataset consisting of 340 samples of shaly sand gas reservoirs. The gamma ray (GR), resistivity (RLLD), spontaneous potential (SP), bulk density (RHOB), neutron porosity (NPHI), and depth were used as the input variables with the core-derived porosity (CPHI) and permeability (CKHG) being used as the targets. The quantitative measures of performance of the models included R2, Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The findings indicated that KNN regression was better than its counterparts as it achieved R2 = 0.8933 and R2 = 0.9340 in terms of porosity and permeability prediction, respectively, and more acceptable metrics of errors showed. Comparatively, the traditional empirical methods showed a significantly lower accuracy rate. The findings highlight that machine learning has the potential to provide precise, scalable, and low-cost predictions of the reservoir properties which could lead to better choices for exploration and production activities.

Prediction of Water Saturation Using Physics-Guided Machine Learning in Deep Silurian Shale Gas Reservoirs

Accurate water saturation estimation in deep shale reservoirs is complicated by clay-related additional conductivity and coupled pore, organic-matter, and structural effects. This study develops a feature-level physics-guided machine-learning framework, termed PhysML-Hybrid. Five mechanism-derived descriptor groups representing clay–water interfacial behavior, low-resistivity correction, pore connectivity, organic-pore development, and structural stress were integrated with conventional reservoir variables in a validation-weighted ensemble of random forest, XGBoost, and Bayesian neural network models. The framework was evaluated using 153 depth-matched samples from five wells in the Dingshan area of the Sichuan Basin. The data were divided into 107 training, 16 validation, and 30 independent test samples, and target-stratified five-fold cross-validation was conducted exclusively within the training set. Mean cross-validation R2, MAE, and RMSE were 0.907±0.009, 1.69%±0.10%, and 2.25%±0.14%, respectively. On the independent test set, the corresponding values were 0.902, 1.77%, and 2.34%. PhysML-Hybrid outperformed Archie, SVM, ML-only, and Phy-XGB. SHAP and statistical analyses identified clay content, the curvature–clay interaction, TOC, pore connectivity, and structural descriptors as influential variables; candidate transitions were interpreted as dataset-specific rather than universal thresholds or causal relationships. Three blind-well cases provided supplementary evidence of cross-well applicability, although larger independent multi-basin datasets are required to assess transferability.