Reservoir characterization integrates geological and petrophysical information to estimate key subsurface properties and reduce uncertainty in field development. However, the workflow remains challenging because reservoir systems are inherently heterogeneous and nonlinear, while field datasets are often limited, noisy, and inconsistent across wells and formations. This review synthesizes state-of-the-art machine learning (ML) methods for reservoir characterization using commonly available petroleum datasets, including core and petrophysical measurements, well logs, seismic attributes, and production/well test data. The literature is organized by learning objective (regression, classification, and optimization) and major model families, emphasizing artificial neural networks (ANNs), support vector machines (SVMs), and fuzzy/hybrid frameworks for key tasks such as porosity-permeability prediction, lithofacies identification, saturation estimation, and reservoir connectivity analysis. Several studies have demonstrated that machine-learning-based well log interpretation workflows can achieve high predictive performance for reservoir characterization tasks, with reported R2 values frequently exceeding 0.90 and, in some optimized field-specific applications, approaching 0.98. Optimization-assisted ANN, ensemble-learning, and hybrid machine-learning approaches have further enhanced prediction accuracy, model convergence, and error reduction for applications including porosity, permeability, saturation, and missing-log prediction. Nevertheless, model performance remains highly dependent on reservoir heterogeneity, data quality, feature engineering, and validation strategy, highlighting the need for rigorous model development and independent validation prior to field deployment. Beyond summarizing applications, this review identifies key causes of degraded field performance, including data leakage during validation, and domain shift across wells and fields. The key novelty is that prior work is fragmented across data types and isolated case studies; this review consolidates those results into a unified, task-driven framework that synthesizes methods, typical performance metrics, limitations, and deployment guidance—providing a practical reference for selecting and validating ML workflows for reservoir characterization. • Comparison of experimental and ML methods enhances reservoir characterization. • Machine learning advances production forecasting and well test interpretation. • Hybrid modeling and data integration significantly improve reservoir accuracy.
Porous gas reservoirs are abundant in reserves and represent a significant proportion of discovered gas reservoirs worldwide. During the middle to late stages of development, these reservoirs often face the risk of water invasion, which reduces displacement efficiency and gas recovery rates. However, the gas-water two-phase flow behavior in porous media remains a persistent challenge. To address this issue, a pore-scale gas–water two-phase flow model was established by coupling fluid flow equations with two-phase interfacial transport equations. Using micro-computed tomography technology, a digital core model representative of a porous gas reservoir was reconstructed. The gas–water distribution and migration characteristics were simulated during both the gas-driven water accumulation stage and the water-invasion depletion extraction stage. The effects of injection rate, wettability (contact angle), and interfacial tension on flow behavior and displacement efficiency were analyzed. The results show that wettability governs the two-phase distribution: while water displaces gas along pore walls under hydrophilic conditions, it occupies pore centers under hydrophobic conditions, thereby increasing flow resistance. Increasing the injection rate significantly enhances the driving force and substantially improves gas recovery. In contrast, higher interfacial tension strengthens capillary forces, leading to greater gas retention and reduced ultimate recovery. These results provide an important theoretical foundation and key technical support for understanding gas-water seepage mechanisms in porous gas reservoirs, predicting development dynamics, and formulating rational development strategies.
The lack of permeability in hydrocarbon-bearing formations can significantly hinder the flow of hydrocarbons from the formation to the wellbore, thereby limiting oil and gas production. To address this challenge, a novel method has been developed that utilizes laser technology in conjunction with enablers, such as activated carbon, to effectively stimulate wellbore production. In a laboratory-scale study, a continuous-wave 1 kW laser was applied for 30 s to limestone under three scenarios: dry rock without activated carbon, dry rock with activated carbon, and water-wet rock with activated carbon. The measured peak surface temperature increased from 888 °C for the first scenario to 1795 °C with dry activated carbon, corresponding to a 102
Predicting post-acidizing porosity and permeability in sandstone reservoirs remains a fundamental challenge due to complex, nonlinear acid–rock interactions. This study presents a physics-augmented machine learning and deep learning (ML/DL) framework that integrates 24 laboratory core-flooding measurements from the Lower Indus Basin, Pakistan, with 500 synthetic training samples generated by the Panga–Lemerle–Balakotaiah (PLB) two-scale continuum simulator. Eight ML/DL architectures are benchmarked on the augmented test set (n = 75): Support Vector Regression (SVR) achieves the highest porosity R ^2 = 0.927 and Random Forest the highest permeability R ^2 = 0.907. A systematic domain-gap analysis reveals that naive median imputation of the unknown HCl concentration causes catastrophic external validation failure (porosity R ^2 =−2.34). Five targeted remediation strategies are evaluated. Strategy S4—Optuna-tuned LightGBM with Leave-One-Out cross-validated joint augmented+experimental training—raises external R ^2 to 0.918 (porosity) and 0.841 (permeability), representing improvements of +3.25 and +2.18 R ^2 units over baseline. SHAP interpretability analysis identifies HCl concentration, initial permeability, and the Damköhler number as collectively governing >75
High-pressure high-temperature (HPHT) wells present significant challenges for conventional water-based drilling fluids (WBDFs), including thermal degradation of additives, excessive fluid loss into formations, rheological instability, poor cuttings suspension, and weak, permeable filter cakes. These limitations result in increased non-productive time (NPT), wellbore instability, elevated equivalent circulating density (ECD), and higher operational risks in deep and ultra-deep hydrocarbon reservoirs. This study systematically investigates the application of nickel oxide (NiO) nanoparticles as a functional additive to enhance thermal stability, rheological properties, and filtration control of bentonite-based WBDFs under HPHT conditions. NiO nanoparticles with an average size of 10–20 nm were dispersed in 7 wt
The present study focuses on the flow and heat transfer dynamics of a Carreau nanofluid flowing through a micro-channel. The micro-channel is filled with a porous media and the micro-channel walls are assumed permeable, allowing for fluid injection and suction. The flow is subjected to the heat transfer under effects of thermal radiation and viscous dissipation. In addition to being shear-rate dependent under the Carreau viscosity constitutive model, the fluid viscosity is also considered temperature dependent. A uniform magnetic field is also applied transverse to the main flow direction. The Carreau model is an example of a non-Newtonian fluid model which. The flow and heat transfer dynamics of non-Newtonian nanofluids through porous channels or micro-channels finds wide application, say to, petroleum engineering, heating and cooling processes in engineering and industry, biomedical and biotechnological phenomena, etc. The governing equations form a complex system of coupled and nonlinear partial differential equations (PDEs). The governing system of PDEs is solved via direct numerical simulation using robust and efficient numerical algorithms based on semi-implicit finite difference methods (SIFDM). Qualitative analyses, based on changes to the values of the embedded parameters, are performed on the profiles of the field variables, namely, the main-flow velocity, fluid temperature, and nanoparticle concentration. The principal insights derived from the research reveal that the qualitative behaviour of the main-flow velocity mirrors that of the fluid temperature, i.e., these quantities either both increase or both decrease in response to changes in the values of the embedded parameters. The results illustrate that the following effects are drag-inducing, i.e., lead to a decrease in the main-flow velocity; porous media intensity, transverse magnetic field strength, increased injection/suction cross-flow motion, and higher fluid viscosity. The study also demonstrates the effects of Brownian motion and thermophoresis on the main-flow velocity and fluid temperature. Specifically, it is observed that an increase in either thermophoresis or Brownian motion lead to increases in temperature and velocity. A particularly important result of the study illustrates that shear-thickening Carreau nanofluids are less susceptible to thermal runaway phenomena than Newtonian fluids. The results lead to important conclusions with regards to heat transfer applications. An increase in nanoparticle concentration is linked to a increase in thermophoresis but to a decrease in Brownian motion. Similarly, a decrease in nanoparticle concentration corresponds to a decrease in thermophoresis but to an increase in Brownian motion. It is therefore important in applications to carefully balance the nanoparticle concentration in order to optimize the combined effects of thermophoresis and Brownian motion. For applications which may demand thermal runaway phenomena mitigation, it is crucially important to utilize rheologically appropriate fluids such as sher-thickening Carreau fluids. The present study significantly adds to the existing scientific body of knowledge with regards a systematic analysis of multiple heat transfer effects simultaneously, instead of these effects being examined in isolation as is largely presented in the existing literature.
Vibration in drilling systems is the fundamental challenge for the oil-and-gas sector impacting productivity, accident prevention, and economic feasibility. By and large, vibrations are axial, lateral, and torsional oscillations resulting from the complicated interaction between a drill bit, rock mass, drill string dynamics, and process parameters followed by the accelerated tool deterioration; decreased rate of penetration (ROP); and bore instability. The research examines such vibration action to devise effective strategies improving both drilling performance and safety. While applying a multivariable approach, the study combines analytical modeling, numerical simulation through ANSYS and MATLAB, and empirical verification by means of lab-based and full-scale testing in a shale-quartzite stratum which thickness is 2000-m. The analysis determines axial vibrations peaking at the frequencies of 10–20 Hz with up to 6 mm amplitudes resulting from high weight on a bit and hard rocks; at the same time, transverse oscillations being 5–15 Hz up to 4 mm happen because of rotational velocity- bottom assembly disordered orientation. Torsional behavior characterized by nonuniform motion at frequencies being 0.5–2 Hz and angular displacements of up to ± 120 are actuated by high-friction fluids with and the reduced greasy properties. Parameter-oriented analysis shows the following: a bit pressure being up to 15 kN as well as rotating velocity being up to 100 rpm reduces longitudinal and torque vibration level by 50
High water saturation and complex pore-throat structure commonly restrict gas flow in tight gas reservoirs by increasing the threshold pressure gradient. This study experimentally evaluated the effect of water saturation on the gas-phase threshold pressure gradient using the microflow displacement pressure-difference method. Matrix-type and fracture-vuggy cores were tested at 80 ℃ with 30,000 mg/L brine and nitrogen. The measured water-phase threshold pressure gradient ranged from 0.0633 to 1.2860 MPa/m and showed negative correlations with permeability and reservoir quality index, but little correlation with porosity. Fracture-vuggy cores exhibited threshold pressure gradients 58–72
Heavy oil upgrading remains an important challenge because of the high viscosity, high content of resins and asphaltenes, and elevated sulfur concentration of heavy crude feedstocks, which limit their production, transportation, and refining. In this work, the catalytic performance of commercially available transition metal oxide nanoparticles, namely NiO, Co3O4, CuO, Fe2O3, MoO3, CeO2, and MnO2, was comparatively evaluated for the hydrothermal upgrading of high-sulfur Ashalcha heavy crude oil under identical reaction conditions. Upgrading experiments were conducted in a high-pressure high-temperature batch reactor at 300 °C for 24 h using an oil/water mass ratio of 70:30 and a catalyst loading of 0.3 wt
Cement integrity evaluation, a critical step in the well plug and abandonment process, is traditionally performed using acoustic logging techniques. However, conventional analyses often focus on early-time reflections, leaving a gap in the literature regarding the diagnostic information contained in later portions of the signal. In this work, a laboratory-scale experimental setup is presented to investigate pulse-echo acoustic wave propagation in cemented well configurations, together with a data-driven framework for signal processing and interpretation. The setup enables acoustic logging with excitation frequencies from 400 to 1000 kHz across 11 sample configurations combining different materials and thicknesses to emulate field-representative conditions. We investigate the effect of signal-processing techniques, including temporal filtering, Hilbert transform, PCA, and data augmentation, for training machine-learning models aimed at classifying sample conditions. The results show that the Reverb Zone, a signal region commonly underexplored, exhibits pronounced sensitivity to damage, particularly in the low excitation-frequency range (400–500 kHz), where reverberation phenomena are enhanced. The proposed framework, combining Reverb Zone–focused processing with a K-Nearest Neighbors classifier at 500 kHz, achieved an overall prediction accuracy of 91
Electrofacies classification is essential for reservoir characterization, as it links well-log responses to lithological and petrophysical properties. A data-driven workflow was developed for electrofacies classification using routine well logs from an Iranian oil field. The dataset comprised 44,719 depth-indexed data points from eleven wells, using seven routine well logs (Caliper, True Formation Conductivity, Sonic Transit Time, Gamma Ray, Compensated Gamma Ray, Bulk Density, and Neutron Porosity). The preprocessing workflow involved quality screening, missing-value handling, caliper adjustment, selective outlier treatment, and class balancing applied to the training data. Three tree-based classifiers, Random Forest, Extreme Gradient Boosting (XGBoost), and LightGBM, were trained individually and integrated through soft voting and stacking-based ensemble strategies. Generalization was assessed via three blind-test experiments in a leave-one-well-out design, each time training on eight wells and predicting electrofacies in the unseen well. Model performance was evaluated using three complementary metrics: Accuracy, Macro F1-score, and Cohen’s kappa to account for the pronounced class imbalance in the dataset. Based on Macro F1-scores, the hybrid stacking approach achieved values of approximately 50.7, 81.1, and 62.9 for Wells B, E, and H, respectively. Relative to the average performance of the individual tree-based models, hybrid stacking resulted in an approximate improvement of 2–4 percentage points in Wells E and H, while showing a decrease in Well B due to the absence of Type 5 (Anhydrite) in that well, which limits the model’s ability to generalize to all facies classes. These results demonstrate that ensemble learning can enhance electrofacies prediction in typical wells and provide quantitative insights into subsurface heterogeneity. Beyond predictive performance, improved electrofacies classification helps reduce uncertainty in reservoir evaluation, supports more cost-effective development planning, and strengthens operational decision-making. The proposed workflow addresses a key limitation of prior single-algorithm approaches by demonstrating validated cross-well generalization with measurable performance improvements.
The lower sub-member of the fourth member of the Eocene Shahejie Formation (LSm-ESF) in the southern Bohai Bay Basin hosts red-bed sandstones that are increasingly targeted as deep hydrocarbon reservoirs, yet the quantitative controls of diagenesis on their reservoir quality remain poorly constrained below 3000 m. This study integrates core description from 20 wells (406.72 m), petrographic point-counting (92 samples, 27,600 points), well-log and mud-log data from over 120 wells, and 450 km² of 3D seismic data to quantify the relative contributions of compaction, cementation, and dissolution to porosity evolution. Compaction is the dominant porosity-reducing process, accounting for an average porosity loss of 13.3
Abstract The search for unexplored unconventional reservoirs has intensified to meet the worlds expanding energy needs. The present research creates the first non-toxic oil-in-water (O/W) emulsion drilling fluid by combining bentonite and tragacanth gum with β-cyclodextrin as a green emulsifier and corn oil as the dispersed phase. The designed corn oil O/W emulsion mud’s rheological, filtration, and shale inhibition properties are contrasted with those of the traditional diesel oil O/W emulsion mud. The corn oil mud exhibited an excellent efficiency with a 21% improvement in rheological characteristics and a 41% decline in filtration loss volume as compared to diesel mud. The emulsion mud made of corn oil showed exceptional shear-thinning behavior as well as outstanding resistance to salt contamination. The shale dispersion and slake durability experiments were used to assess how formulated mud interacted with drill cuttings. The primary shale recovery of the maize oil mud was 88.50%, suggesting that the mud is harmless and does not interact with other materials. Likewise, the slake durability index of corn oil mud was 73.2%, suggesting exceptional inhibitive qualities in corn oil mud compared to diesel oil. According to the core flooding investigation, the proposed corn mud system (81.5%) would have a low formation damage effect compared to diesel oil mud (74.2%). The environmental evaluation of corn oil mud showed that it had a higher LC50 and no aromatic components, which indicated that it was less hazardous and more biodegradable than diesel and making it suitable for application in drilling unconventional reservoirs.
The Sembar Formation is a key petroliferous unit in the Lower Indus Basin of Pakistan; however, its resource potential remains uncertain due to limited and discontinuous Total Organic Carbon (TOC) and Rock-Eval pyrolysis data. Accurate assessment of TOC, volatile hydrocarbons (S1), remaining hydrocarbons (S2), carbon dioxide yield (S3), maximum pyrolysis temperature (Tmax), and Vitrinite Reflectance (
The complexity of diagenesis and its evolutionary processes in the Chang 2 reservoir complicates the identification of high-quality reservoirs, thereby directly affecting reservoir evaluation and subsequent oilfield development decisions. The diagenesis process and pore evolution of Chang 2 reservoir in the Caoduowan area in the Ordos Basin were analyzed by using the methods of cast thin section, grain size analysis, mercury intrusion porosimetry, cathodoluminescence, X-ray diffraction (XRD) and fluid inclusion analysis. The findings indicate that the Chang 2 reservoir is predominated by feldspar sandstone, which primarily develops dissolution pores, intergranular pores and micro-fractures. Diagenetic modifications, including compaction, cementation, and dissolution, have collectively controlled the porosity evolution of the Chang 2 reservoir. In light of the disparity in diagenesis, the reservoir can be categorized as the chlorite-cemented intergranular pore type, the feldspar-dissolution intergranular pore type, and the carbonate-cemented composite pore type. The feldspar-dissolution intergranular pore type reservoir underwent strong compaction at an early stage, but the significant dissolution effect resulted in better reservoir physical properties. The chlorite-cemented intergranular pore type and the carbonate-cemented composite pore type reservoirs exhibit relatively fine intergranular porosity during the early and middle diagenetic stages. Nevertheless, the carbonate-cemented composite porosity reservoirs are impacted by late cements, leading to the deterioration of their physical properties. By quantitatively characterizing the diagenetic evolution of three representative reservoir types, this study clarifies the controlling factors for the development of high-quality reservoirs. These results provide a geological basis for predicting favorable reservoir intervals and guiding development adjustments in the Caoduowan area.
Running sand screen completion strings to target depth is challenging in horizontal and extended-reach wells because drag, string stiffness, borehole curvature, cuttings accumulation, and borehole shrinkage can jointly reduce passability and cause sticking. This study develops an integrated passability prediction and diagnostic workflow that combines a torque-and-drag model considering cuttings-bed-enhanced friction and buckling-induced contact force, a deformation-energy-based stiffness-matching model, and an elastic geometric-compatibility model for large-diameter downhole attachments. The workflow was applied to Well H in the Kenli 10-2 Oilfield, where the sand screen completion string became obstructed at a measured depth of 3316 m and could not pass the interval even under a maximum set-down load of 35 t. Initial analyses indicated that the string should theoretically reach the target depth, with a predicted hook-load margin of 62.67 t, a stiffness ratio greater than 1.0, and a minimum sliding-sleeve passable curvature of 8.18°/30 m. However, the apparent open-hole friction coefficient inverted from field hook-load data reached approximately 3.5, whereas the cuttings-bed-corrected friction coefficient increased only from 0.50 to 0.59, indicating that ordinary friction and cuttings accumulation were not the dominant causes. Further shrinkage analysis showed that a 2.2
Unlocking the untapped gas potential of the Nile Delta requires moving beyond conventional seismic interpretation, which often struggles to distinguish between economic gas accumulations and non-commercial lithological variations. This study addresses this challenge by implementing a quantitative Amplitude Variation with Offset (AVO) workflow to delineate new gas prospects within the Pleistocene El Wastani and Pliocene Kafr El Sheikh formations of the offshore Baltim Field. The methodology integrates three-dimensional partial-stack seismic data (Near, Mid, and Far offsets) with petrophysical logs from six exploratory wells. A calibrated workflow was applied, utilizing super-gathers and trim statics, seismic-to-well calibration (correlation coefficient > 0.85), and the extraction of primary (intercept A, gradient B) and secondary AVO attributes (A × B , A-B , A+B ). Quantitative classification was achieved using AVO cross-plotting to isolate gas-bearing anomalies from the background lithology trend. Results reveal Class III AVO anomalies at the reservoir tops, characterized by a distinct "brightening" of amplitude with offset. The gas sands exhibit a strong positive AVO Product response and a negative Scaled Poisson’s Ratio, successfully mapping the 3D geometry of the reservoirs. Specifically, the analysis identified high-potential gas zones with negative shear reflectivity ( A-B ) anomalies, confirming a decrease in shear impedance. The novelty of this work lies in the successful application of a calibrated, attribute-driven AVO workflow to the often-overlooked shallow Pliocene section, challenging the traditional focus on deeper targets. By providing a quantitative "stoplight" mechanism for prospect ranking, this approach offers a transferable model for risk-mitigated exploration in similar deltaic environments, potentially reducing drilling risks and enhancing discovery rates in mature basins.
Abstract The Miocene-aged fluvio-deltaic Bhuban Formation in the Bengal Basin is generally regarded as a poor hydrocarbon producer due to the dominance of low-quality reservoir facies. This study investigates the occurrence and characteristics of higher-quality reservoir facies within the formation, leading to the identification of sand-filled incised valleys in outcrop. In addition to this valley-fill sandstone (RF-III), alternating sandstone and shale layers (RF-I), and fine-grained channel sandstones (RF-II) were identified through detailed fieldwork. Petrographic, X-ray Diffraction (XRD), and Scanning Electron Microscope (SEM) analyses indicate that RF-III possesses the most favorable reservoir properties, with coarser grains and high porosity observed in thin section. By contrast, RF-I represents the poorest reservoir quality, while RF-II has moderate potential that becomes significant when multiple units are stacked. Seismic data from the adjacent Srikail gas field were interpreted to map incised valleys within the Bhuban Formation and used for volumetric modeling. Results suggest that a single sand-filled incised valley could contain up to 40 BCF of Gas Initially In Place (GIIP), with substantially greater volumes possible if the valley and field dimensions are larger. These findings demonstrate, for the first time, the quantitative reservoir potential of sand-filled incised valleys in the Bhuban Formation, highlighting them as promising but underexplored exploration targets in the Bengal Basin.
Gas flow in shale reservoirs is strongly affected by adsorption-induced deformation, stress sensitivity, and gas slippage in micro–nano scale pores, which makes permeability prediction challenging. In this study, a theoretical apparent permeability model is developed by coupling a capillary bundle representation with adsorption deformation, stress/strain–permeability relationships under constant confining pressure, and the slippage effect. The model is validated using published pulse-decay permeability data for Posidonia shale with CH₄ and He. The predicted results agree well with experimental measurements, with average absolute deviations generally below 0.15. In addition, non-adsorptive gas (He) exhibits higher apparent permeability than adsorptive gas (CH₄). Compared with existing models, the proposed approach explicitly accounts for coupled adsorption deformation and slippage effects within a unified framework, providing improved accuracy and clearer physical interpretation of shale gas transport.
Reliable estimation of the minimum miscibility pressure (MMP) is a critical requirement for the successful design of gas injection processes in enhanced oil recovery, as miscibility strongly controls displacement efficiency and recovery performance. Direct laboratory measurements of MMP are expensive and time-intensive, while traditional empirical correlations often fail to account for the complex interactions between fluid composition and reservoir conditions, leading to limited predictive reliability. To overcome these challenges, an interpretable data-driven framework based on a stacking ensemble learning strategy is developed for MMP prediction. The framework combines Random Forest, Gradient Boosting, Support Vector Regression, and XGBoost models and incorporates polynomial feature expansion together with recursive feature elimination to explicitly capture nonlinear and interaction effects among gas composition, oil composition, and thermodynamic variables. Model performance was assessed using an independent test dataset and evaluated using the root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R²). The optimized ensemble model demonstrates a substantial improvement over individual base learners, achieving an RMSE of 2.21, an MAE of 1.317, and an R² of 0.912. Beyond predictive accuracy, model interpretability was enhanced through SHapley Additive exPlanations and partial dependence analysis. The results reveal that the interaction between reservoir temperature and the molecular weight of heavy oil fractions exerts the strongest influence on MMP, followed by key compositional interactions involving nitrogen and carbon dioxide. These trends are physically consistent with established miscibility and phase-behavior principles. The main contribution of this work lies in the integration of an optimized stacking ensemble with interaction-focused feature engineering and explainable artificial intelligence techniques. This combination enables accurate, transparent, and physically meaningful MMP predictions, advancing beyond previous studies that emphasize either accuracy or interpretability alone. The proposed framework offers a practical and reliable tool for supporting gas injection design and decision-making in diverse reservoir systems.
