Robust prediction of advanced reservoir petrophysical parameters using conventional logs: Insights from XGB and neural network models

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.

Core-Calibrated Machine Learning–Based Permeability Estimation: A Western Niger Delta Case Study

Accurate permeability estimation is essential for optimum reservoir characterization, simulation, and development planning; however, its prediction remains challenging due to the limited availability of core-derived permeability measurements, often constrained by acquisition cost and data coverage. This study presents an integrated machine learning workflow for permeability prediction calibrated against core permeability data using wireline log information from wells located in western Niger Delta. The workflow involved systematic data preprocessing, depth-based alignment of core and well log data, and feature engineering to derive additional petrophysical attributes relevant to fluid flow behavior. Key input variables used in the model include Gamma Ray (GR), Bulk Density (RHOB), Resistivity, Neutron Porosity, and several derived parameters such as shale volume, resistivity index, effective porosity, neutron–density separation, and bulk volume water. Permeability values were transformed into logarithmic space to improve modeling stability and capture the wide range of permeability values typical of heterogeneous clastic reservoirs. Five (5) supervised machine learning algorithms; Random Forest (RF), Extreme Gradient Boosting (XGB), Extra Trees Model (ETM), AdaBoost (ADB) and Decision Tree (DT) were developed and evaluated using an 80–20% train–test split, with blind test well excluded for validation against measured core permeability. The Extra Trees model demonstrated the highest predictive performance, achieving an R² value of ~ 0.9164, Mean Absolute Error (MAE) of ~ 1.5392, and Root Mean Squared Error (RMSE) of ~ 4.1011, indicative of high correlation between predicted and core-measured permeability. The result visualized predicted versus actual permeability along the depth axis, providing a vertical reservoir-scale understanding of permeability distribution. The results indicate that machine learning models can effectively capture complex petrophysical relationships and provide reliable permeability estimates in intervals lacking core measurements.

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

Advances in the reservoir characterization from well logs using ML and rock physics analysis: a case study from the West Offshore Nile Delta, Egypt

The limited availability of shear-wave velocity (Vs) logs in deep-water reservoirs presents a significant challenge for reliable reservoir characterization and elastic property analysis. This study aims to develop and validate a robust machine learning (ML)-based workflow for predicting Vs from conventional well logs and integrating the results into rock physics–driven reservoir characterization. The dataset comprises five wells from two geologically analogous gas-bearing marine reservoirs: four wells from the West Offshore Nile Delta, Egypt, and one well from the Scarborough Gas Field, Australia. All wells include measured Vs data, enabling a structured training and validation strategy in which three wells are used for training, and two wells are reserved for blind validation across different fields. Input logs include gamma-ray (GR), bulk density (RHOB), deep resistivity (LLD), neutron porosity (APLC), and compressional velocity (Vp). Four ensemble ML models, gradient boosting (GB), extreme gradient boosting (XGB), light gradient boosting (LGB), and categorical boosting (CB), were evaluated. The CB model achieved the best performance, with R2 values of 0.912 for testing and 0.904 for blind validation, along with the lowest RMSE and MAPE. The predicted Vs was subsequently used to derive elastic attributes, including Lambda-Rho and Mu-Rho, and to construct rock physics templates for lithology and fluid discrimination. These elastic properties were further integrated into a three-dimensional geostatistical modelling framework using Gaussian Random Function Simulation (GRFS), constrained by seismic-derived structural surfaces, to generate volumetric distributions of porosity, acoustic impedance, and Lambda-Rho across the reservoir. The results demonstrate that the ML-predicted elastic properties consistently identify gas-bearing sands, brine sands, and shale intervals. Furthermore, the successful application of the workflow across two distinct but analogous reservoir settings highlights its robustness and transferability. This integrated ML–rock physics framework offers an effective solution to enhance reservoir characterization while reducing reliance on costly and limited Vs measurements.

Application of supervised machine learning for predicting and classifying the lithology type of reservoir rock: the case study of Iraq

Accurate lithology identification during subsurface exploration is essential for efficient hydrocarbon production, drilling optimization, and reservoir management. This study investigates automated lithology classification using supervised machine learning (ML) to reduce uncertainty and support operational decision-making in hydrocarbon reservoirs. A dataset from Majnoon oil field in Iraq was used, including sonic (DT), gamma ray (GR), resistivity (LLD, LLS), neutron porosity (NPHI), and bulk density (RHOB). Four ML algorithms were evaluated: decision tree (DT), support vector machine (SVM), K-nearest neighbors (KNN), and neural network (NN). All models achieved high predictive performance. The accuracy values for KNN, SVM, NN, and DT were 98.62%, 99.08%, 99.39%, and 99.69%, respectively. Among these, DT achieved the highest classification performance, with 99.69% accuracy on validation data and 100% on test data. Notably, for the sandstone class, DT reached a precision of 99.7% and a recall of 99.2%, outperforming the other models. Feature importance analysis revealed that depth and GR logs were the most influential predictors. In addition, error analysis was performed using Percent Bias (PBIAS), confirming the minimal deviation of predicted values from observed data across all models. This work offers a high-accuracy, low-cost framework for real-time lithology prediction, providing a novel integration of interpretable ML techniques in a domain traditionally reliant on expert interpretation. Unlike previous efforts, this study quantitatively compares multiple models and highlights class-specific performance, enhancing practical applicability in field operations.

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.