In the Jiyang Depression, sand, coal, and mudstone are interbedded, and the sand and coal layers contain high hydrocarbon content. This study addresses the optimal selection of fracturing layers for multi-gas co-production and simulates fracture propagation across interbedded coal and rock. A finite-discrete element model was established to analyze the effects of the fracturing layer, pumping rate, liquid viscosity, and cleat connections on the longitudinal propagation of hydraulic fractures in interlayer coal formations. The results show that when the fracture initiation layer is sandstone, hydraulic fractures can effectively propagate through interlayer interfaces to fracture the coal seams. When the fracture initiation layer is coal rock, hydraulic fractures in the vertical direction are confined within the coal seam and cannot propagate through the interlayer interface to fracture the sandstone layers. Hydraulic fractures are mainly connected by cleat fractures in the horizontal direction, forming complex network of fractures. Based on the research results, it is recommended that sandstone layers be used as the fracturing treatment layer for interbedded coal-bearing strata in the Jiyang Depression. The pumping rate is 22 m3/min, and the liquid viscosity is 15 mPa & centerdot;s. Through on-site production data collection, the highest daily gas production of Well A in the Jiyang Depression exceeded 20,000 m3/d, with an average daily gas production exceeding 10,000 m3/d, which significantly improved the production effect.
This paper summarizes a machine-learning competition, which is the first public benchmark comparing machine-learning (ML)-based automatic depth-shifting methods for well logs. Accurate depth determination of subsurface formations is critical for various applications in the oil and gas industry. However, achieving accurate depth registration of well logs can be challenging due to inherent ambiguity and subjectivity. Conventional depth-shifting methods relying on manual bulk shift based on peaks and troughs can be tedious and time consuming. The 2023 Machine-Learning Competition aimed to evaluate and compare various data-driven techniques for automatically aligning well logs to a reference log and correcting depth misalignments. It was organized by the Petrophysical Data-Driven Analytics (PDDA) Special Interest Group (SIG) of the Society of Petrophysicists and Well Log Analysts (SPWLA). Various machine-learning algorithms and deep-learning architectures from the top five teams were explored, including techniques like dynamic time warping (DTW), convolutional neural networks (CNNs), recurrent neural networks (RNNs), and ridge regression, known for their ability to learn complex patterns from data. The performance of the submitted solutions was evaluated using metrics like root mean squared error (RMSE) and mean absolute deviation (MAD). The contest results, presented in this paper, highlight the effectiveness of data-driven methods in automatically rectifying well logs and provide insights into the strengths and weaknesses of different approaches for this task. The findings contribute to the advancement of automated depth-shift techniques and their potential application in subsurface characterization and reservoir management. The code and data set are available at https://github.com/ pddasig/Machine-Learning-Competition-2023. This paper only gives a high-level summary of this competition, and we refer readers to the GitHub repository for a more indepth understanding of the works.
In adjustment well cementing of old areas, underground high-pressure water layers can invade the cement slurry, leading to poor cement-stone bonding and compromised cementing quality. Previous studies had primarily focused on developing cement-slurry systems with anti-water-invasion (AWI) materials; however, the evaluation of the AWI ability of the cement slurry was often neglected. This study proposes a method to rapidly evaluate the AWI performance of cement slurries, along with an experimental approach for assessing the water invasion during coagulation. The mechanisms are investigated using Fourier transform infrared (FTIR) spectroscopy and permeable water conductivity, X-ray diffraction (XRD), and thermogravimetric (TG) analyses. The FTIR results indicate the presence of functional groups such as methylene and hydroxyl in the AWI material, which demonstrate adsorption properties, and can improve the cohesion of cement particles and thereby, the AWI ability. The results of the rapid evaluation of the AWI capability of cement slurry show a sudden increase in the conductivity of the AWI cement slurry, but at a later stage than that of the conventional cement slurry. The results of experiments conducted for assessing the water invasion during cement coagulation demonstrate that the conventional cement slurry in high-permeability cores exhibits higher water-invasion rates, wider second-interface cracks, and greater hydraulic conductivity than the AWI cement slurry in low-permeability cores under identical intrusion pressures. This indicates that combining low-permeability cores with AWI materials can effectively reduce the cement-slurry ion losses. XRD and TG analyses reveal that formation-water invasion significantly affects the conventional cement-slurry components in high-permeability cores, while showing minimal impact on AWI cement slurries in low-permeability cores. The proposed method can evaluate the AWI capability of different cement-slurry systems, laying the foundation for improving the AWI capability of adjustment well cementing slurry systems and cementing quality of adjustment wells.
Acid fracturing technology in coal seams has been proven to be an efficient coalbed methane (CBM) mining method. With the development of research, acid fracturing technology has become a hot research topic. Here, the deep No. 8 coal seam in the Daning-Jixian of the Ordos Basin is taken as the object of study. By using a series of physical simulation experiments of triaxial hydraulic fracturing and field tests, we systematically investigate the fracture propagation behavior of acid fracturing in the deep coal seam and mainly examine the acid concentration and perforating location. We discover that the direct acid fracturing effect of deep primary structural coal seam is significant, and the fracturing ability is stronger than that of shallow broken and soft coal seams. During the acid fracturing process, there is no large filtration phenomenon, which makes it easy to form complex, long fractures. With the increase of acid concentration in the same perforated location, the complex acid fracturing fractures are more likely to form. The field test shows that 10% sulfamic acid concentration is preferred, with the consideration of fracture complexity, fracture height control, fracture pressure, and construction safety. Moreover, we observe that the perforation location is set at the top of the coal seam, which is conducive to the formation of a complex seam network. The work provides a guiding significance for the design of acid fracturing in deep coal reservoirs.
Groundwater significantly impacts the stability of rock engineering and mineral resource exploitation. Mechanical parameters and fracture characteristics of rocks are crucial in theoretical analysis, experimental testing, and numerical simulation. To investigate the weakening mechanism and fracture energy characteristics of water-saturated rocks, sandstone specimens with varying water-bearing states were prepared through a drying and soaking process, and uniaxial compression acoustic emission (AE) experiments were conducted. The results show that with the increase of soaking duration, the compressive strength and elastic modulus exhibit a negative exponential decreasing trend and eventually tend to stabilize. Based on the theory of energy accelerated release, the characteristics of the AE energy release rate are quantitatively analyzed. The AE energy of dry specimens is more likely to be released in the form of accelerated, abrupt changes, while that of water-bearing specimens is more likely to be released in a slow and gradual manner. During the loading process, the values of AE energy follow a normal distribution. The weakening of the internal structure of rocks and the enhanced attenuation of wave propagation result in the AE energy expectation of water-bearing specimens being significantly lower than that of dry specimens. Moreover, the upward trend of the expectation for dry specimens near failure is more significant than that of water-bearing specimens. This study concludes that water soaking softens mineral particles and weakens microstructure, resulting in the degradation of the mechanical parameters and the increase of the attenuation. The reduced and gradual AE energy release in saturated specimens reflects a transition from sudden brittle failure to progressive plastic failure in the rock.
assessment of sponge core data through parameterization, uncertainty analysis, and Monte Carlo modeling of critical variables influencing lab-derived saturation results. This work examines lab data and reservoir information that could impact final saturation results in sponge coring. We analyzed ranges of standard raw data from Dean-Stark and spectrometric analysis, including gravimetric weights, distilled water volumes, pore volumes, and sponge absorbance. We evaluated the input variables of fluid and rock properties, such as water salinity, formation volume factors, plug dimensions, and stress corrections. In addition, we examined the governing equations, including the saltcorrection factors and laboratory mass-balance equations, as well as factors such as sources of water salinity, bleeding due to gas liberation, and water evaporation.
Wettability is a critical parameter that controls oil recovery, water production, and formation damage in unconventional reservoirs. It directly influences fluid distribution, capillary forces, stimulation outcomes, and drawdown performance. The fine grain size and heterogeneity of many tight formations make pore-scale wettability a key factor in designing completion fluids, flowback strategies, and production plans. This study introduces a quantitative framework for evaluating pore-type partitioning and wettability using nuclear magnetic resonance (NMR) T2 relaxation and mineralogical characterization across 12 lithologically diverse core samples, with gravimetric monitoring applied only as quality control. Single-and dual-fluid saturation tests were performed to assess fluid behavior under representative reservoir conditions. Results show that mixed-wet pores dominate silica-and clay-rich samples, accounting for up to 85% of the pore network, while carbonate-rich samples exhibit a higher proportion of oil-wet pores. Dual-fluid displacement experiments reveal that, although mixed-wet pores accept both oil and brine, brine more effectively displaces oil, whereas oil struggles to displace brine, demonstrating a clear brine preference under competitive conditions. A key finding of this study is the quantification of an oil saturation choking threshold (26 to 71%), defined as the residual oil-filled porosity at which no further oil is displaced during brine forced imbibition, offering a screening metric to gauge the potential benefit of brine-based choking prior to flowback operation. These insights highlight the need for lithology-driven reservoir production strategies, recommending moderate drawdowns and brine-based recovery for mixed-wet zones, whereas carbonate-rich, oil-wet intervals require higher pressures or wettability-altering agents (e.g., surfactants, CO2) to enhance hydrocarbon recovery.
This paper addresses the difficulty of accurately and quickly determining the true resistivity of tight sandstone reservoirs in horizontal (HZ) wells, due to the coupled effects of dipping angle, mud invasion, and anisotropy on array induction logging (AIL) responses. A novel data processing method based on a hierarchical correction strategy is proposed. By analyzing the response characteristics of AIL in HZ wells, the study decouples the effects of anisotropy and invasion, decomposing the complex three-dimensional (3D) response simulation problem into a combination of horizontal layered (onedimensional (1D)) and cylindrical layered (1D) models, significantly reducing the complexity of 3D inversion. Furthermore, based on the horizontal layered model, a library of true thickness correction coefficients is established to eliminate the influence of thickness, simplifying the horizontal layered model into a homogenous (zero-dimensional (0D)) model. Based on this, a direct focusing method was developed for the 0D model to achieve rapid simulation of radial focused responses. Subsequently, the responses corrected for true thickness were input into the cylindrical layered model to correct for invasion effects. Next, anisotropy coefficients were determined using measurement responses from the HZ well and its vertical profile. Finally, the invasion correction results were input into the 0D model to complete the anisotropy correction. Field cases indicate that the new correction method achieves an inversion speed of 4 points per second, with accuracy errors controlled within 3%, meeting realtime correction requirements. It effectively extracts the resistivity of tight sandstone reservoirs and accurately characterizes invasion features, providing a reliable basis for reservoir evaluation. This method assumes that invasion profiles are axially symmetric and relies on logging data from adjacent vertical wells, limiting its applicability when reference wells are lacking or when invasion is severely asymmetric.
In the shale development of the Qiongzhusi Formation in southern Sichuan, double fracturing is typically used to improve the complexity of hydraulic fractures by utilizing the induced stress field generated during fracturing. However, there is limited research on the disturbance of the stress field caused by pumping rate and injection volume during a single fracturing process, which leads to significant uncertainty in field design. In this study, a two-dimensional (2D) plane model assuming isotropic conditions was established using the finite discrete element method and combined with the design parameters of Well ZY-1. The influence of pumping rate and fluid volume on the formation stress field during single fracturing was simulated and analyzed with the aim of reducing the formation stress difference. The results show that during single fracturing, the induced stress in the direction of the minimum horizontal principal stress is distributed in a "strip" pattern, whereas the induced stress in the direction of the maximum horizontal principal stress exhibits an "X"-shaped distribution. As a result, the induced stress difference displays a zonal distribution pattern that gradually weakens with increasing injection volume. Based on the influence of single fracturing on the formation-induced stress difference under different pumping rates and injection volumes, the pumping rate of the single fracturing operation in Well ZY-1 was optimized to 18 m3/min, and the injection volume was optimized to 1,500 to 2,000 m3. Field application shows that fracture complexity was significantly improved after implementing double fracturing.
Focusing on the Sulige area of the Ordos Basin in China, this paper integrates a data-driven approach with petrophysical models to achieve high-precision prediction of key reservoir parameters, including porosity, water saturation, and permeability. An augmented data set is constructed using conventional logging data combined with computed results from Archie's formula, the Timur formula calibrated with regional statistics, and the neutron porosity logging interpretation model. Core experimental data are used as training labels for evaluation. After data preprocessing, multiple mainstream machine-learning models are comprehensively evaluated, revealing that the deep neural network (DNN) delivers the best overall performance. To further enhance feature extraction and generalization ability, this paper innovatively introduces dilated convolutional neural networks (DCDNN) into the DNN architecture, constructing a DCDNN model. Ablation experiments confirm that this strategy significantly improves the model's representational ability. To boost model robustness and prediction accuracy, the multi-player dynamic game (MPDG) algorithm and Bayesian optimization are applied for hyperparameter tuning of the DCDNN. Experimental results demonstrate a substantial improvement in the optimized model's predictive performance. Finally, based on ensemble learning theory, an expert committee-based decision-making mechanism is established, and an optimal prediction model is selected using multiple comprehensive metrics as evaluation criteria. Compared with the standalone DNN model, the ensemble model reduces the mean absolute error in porosity, water saturation, and permeability predictions by 1.066, 12.711, and 1,661, respectively. The R2 values improve by 3.33, 2.95, and 23.9%, while the relative percent difference values improve by 1.32, 0.48, and 0.87, respectively. Applied to a blind well, the model achieves excellent predictive performance, providing reliable technical support for geological structure analysis and resource assessment in the region.
During downhole operations in complex geological formations, wireline logging instruments are susceptible to operational anomalies, including downhole obstruction and jamming. Under such conditions, significant changes occur in the cable tension, and this tension monitoring is of great significance for anomaly diagnosis. Nevertheless, the challenging downhole environment and operational constraints significantly compromise tension monitoring accuracy, thereby exacerbating the complexity of anomaly diagnosis. Therefore, this paper proposes a dual-signal fusion analysis method for anomaly diagnosis based on cable tension and winch vibration. The anomaly diagnosis is initially performed through independent feature extraction and analysis of both tension and vibration data sets. Building upon this foundation, we propose a novel dual-signal fusion method for enhanced anomaly diagnosis. This approach elucidates the mechanical-vibration coupling mechanism, leading to the development of a synchronized acquisition and fusion strategy for cable tension and winch vibration signals. Both laboratory and on-site experiments show that this method can accurately identify downhole instrument anomalies, effectively reducing missed alarms and false alarms. Field tests on four actual wells achieved an accuracy of 99.3% and a false-alarm rate of 0.15%. The vibration-tension dual-signal fusion approach offers a more robust technical support for downhole condition diagnosis of wireline logging instruments.
Due to the complex composition of grain types and pore types, lithological classification of bioclastic limestones has long posed significant challenges. However, such classification is crucial for reservoir prediction and waterflood development in bioclastic limestone reservoirs. This study presents a novel lithological classification method that integrates geological facies and petrophysical facies characteristics, focusing on the Mishrif Formation in the H oil field. First, based on thin-section observations and hydrodynamic conditions in depositional environments, grain types in the Mishrif Formation were categorized into three groups: high-energy, mixed-energy, and low-energy grains. Subsequently, the 16 lithofacies types in the study area were consolidated into seven geological facies based on the proportional relationships among low-energy grains, high-energy grains, and micritic matrix. Next, petrophysical parameters-including porosity, permeability, R35 (median pore-throat radius), discharge pressure, and median radius-were measured through core analysis, porosity-permeability testing, and mercury injection capillary pressure experiments. By performing cluster analysis on these five parameters, the petrophysical characteristics of the 16 lithofacies have been compared. The results revealed that lithofacies within each geological facies exhibited similar petrophysical properties. Ultimately, the lithofacies in the study area were classified into seven types: mudstone (Type A), wackestone (Type B), limestone-bearing low-energy grains (Type C1), limestone-bearing mixed-energy grains I and II (Types C2-1 and C2-2), and limestone-bearing high-energy grains I and II (Types C3-1 and C3-2). Comparative analysis of production performance across different reservoir types demonstrated that limestone-bearing mixed-energy grains (Types C2-1 and C2-2) and high-energy grains (Types C3-1 and C3-2) exhibited higher productivity than wackestone (Type B), with Types C3-1 and C3-2 showing the highest productivity. This indicates that the presence of high-energy grains promotes the formation of high-quality reservoirs. The rock type classification developed in this study can provide valuable insights for reservoir development and enhanced oil recovery in fields with similar geological settings.
Recent advancements in deep and ultradeep azimuthal resistivity tools (DAR and UDAR) have widened the scope of electromagnetic (EM) measurements available at various depths of investigation. These technologies have successfully enabled geosteering decisions based on a combined assessment of both azimuthal resistivity measurements. UDAR provides valuable insights into formations located farther from the wellbore, whereas DAR excels at identifying complex structures in close proximity. While UDAR and DAR can yield differing resistivity measurements due to different measurement sensitivity, resolution, and detection range within a transition zone (Li et al., 2020; Yan et al., 2020), typically spanning 15 to 30 ft from the wellbore, experienced geosteering engineers can enhance geological decisions by considering the inversion uncertainties and potential detection ranges associated with UDAR and DAR at specific formation profiles. Nonetheless, this approach involves numerous manual steps and can be time consuming, potentially leading to subjective interpretations if conducted by engineers with limited experience and knowledge in azimuthal measurements and formation geology. To reduce manual interpretation and improve consistency, we explored approaches to incorporate UDAR and DAR measurements into an automated workflow. A robust joint inversion technique incorporating both UDAR and DAR measurements has been developed to address the limitations of the traditional human-guided approach. The inversion technique accounts for the significant differences in sensitivity between UDAR and DAR measurements by employing weighted functions tailored to individual measurement types and specific inversion zones. T hese weighted functions are determined by the sensitivity of individual measurements to formation resistivity contrasts and boundaries, individual inversion outcomes, and their associated uncertainties. The joint inversion approach on the basis of different measurements at multiple steps is employed to fine-tune separate inversion zones, at the cost of longer calculation times or increased computational resources. This paper evaluates various synthetic and field examples where both DAR and UDAR measurements are acquired within the same wellbore. Initially, independent inversion results are obtained using DAR and UDAR measurements separately. Subsequently, ajoint inversion is conducted utilizing a multilayer, multitransition formation model to integrate both measurement types. Results demonstrate that the proposed joint inversion approach not only enhances the accuracy of formation property characterization but also extends the detection range into deeper formation layers. These improvements are achieved consistently without the need for manual intervention.
As the number of new wells is continuously increasing across oil fields to maintain and enhance hydrocarbon production, there is an ever-growing challenge to safely drill near existing wellbores, given their downhole location inaccuracy caused by well surveys with an ellipse of uncertainties up to hundreds of feet. An emerging well ranging solution, known as active resistivity ranging and built from the measurements of the ultradeep azimuthal resistivity (UDAR) tool, can be the key technology to accurately determine the relative distance and direction of adjacent wells to avoid the risk of potential collisions during drilling operations. According to Chen et al. (2022), the electromagnetic measurements of the UDAR tool have high sensitivity to a close, near-parallel well that exhibits a high-resistivity contrast with the surrounding formations. Therefore, a novel workflow was developed to determine in real time the distance to the target well using a simulated tool response table, while the measured angles of UDAR's second harmonics can derive the azimuthal orientation of such a wellbore. Motivated by the successful initial results, this ranging technology was then applied to distinct scenarios with the objective of evaluating its performance when the wells are close to vertical and in a more complex case of high-angle wells. In the vertical scenario, the formation layers are nearly perpendicular to the wells, leading to the formation effect being primarily due to inhomogeneity and thus a relatively moderate impact in the UDAR measurements. Consequently, the ranging workflow can accurately determine the distance to the target cased well. However, in the case of high-angle wells, the measurements are very sensitive to both casing and formation layers. Besides the formation effect caused by resistivity anisotropy, other subsurface features can also interfere in the UDAR measurements, such as bed boundaries proximity, high structural dip, and formation discontinuities. Thus, an additional step is applied to first remove the formation-only response from the measured data, which substantially improves the ranging results in terms of relative distance accuracy. The respective case studies provide technical support to this innovative ranging to be capable of deploying at any well inclination, from vertical all the way to horizontal, under near-parallel scenarios. By effectively detecting nearby wells while drilling, such an application can largely increase the safety in the drilling operations and thus potentially enable well construction even within the zones that are nowadays not permitted due to the collision risks. Additionally, it can be applied beyond the anti-collision solution, such as well twinning and borehole interception.
Two well logs from the same section of a well often have an unknown depth shift between them. If the well logs are corrected for this depth shift, the correlation between the logs will increase. We formulate an optimization problem that maximizes the correlation between the logs, while maintaining a physically reasonable depth shift. We implement a solver for this optimization problem and test it on real well logs from the Norwegian Continental Shelf. We demonstrate that changing the trade-off between the correlation term and the smoothness term in our cost function also changes the trade-off between correlation and smoothness in the resulting estimated depth shift. The automatically estimated depth shifts were confirmed by expert opinion to be similar to the depth shift that would have been estimated by an expert on data from 12 real wells, illustrated in Appendix 1.
Ultradeep azimuthal resistivity (UDAR) measurements provide advanced subsurface imaging capabilities by bridging the gap between traditional well logging and seismic interpretation. However, interpreting UDAR measurements in complex, spatially heterogeneous, or anisotropic formations remains challenging. Traditional two-dimensional (2D) and three-dimensional (3D) inversion procedures are computationally intensive and unsuitable for real-time applications. We describe our recent progress in the development of fast modeling and inversion algorithms for the interpretation of UDAR measurements. These algorithms are designed for multi-CPU clusters and tailored for integration with local 3D geological models and arbitrary well trajectories. Modeling employs a finite-volume solution of Maxwell's equations implemented on an adaptive Lebedev grid, effectively accounting for diverse formation complexities and tool configurations. A new block-based solver computes multi-input multi-output (MIMO) responses efficiently, significantly reducing CPU runtime. Inversion is performed with a gradient-based approach (Occam-type) and incorporates regularization to account for measurement noise and non-uniqueness in the estimation. Furthermore, inversion complexity grows progressively from zero-dimensional (0D) to 3D, based on the localized dimensionality of both measurements and the spatial distribution of electrical conductivity. The method is successfully validated with challenging synthetic and field measurements acquired in the North Sea by various service companies. It is confirmed that the modeling and inversion algorithms are efficient, stable, and reliable for estimating the spatial distributions of anisotropic electrical conductivity around the well trajectory. The Jacobian matrix is efficiently calculated and updated by the modeling algorithm, which is as accurate as the traditional perturbation method but five orders of magnitude faster. In addition, we observe that both data misfit and model uncertainty decrease as inversion dimensionality increases. Our adaptive inversion method also significantly reduces computational costs by triggering expensive 2D and 3D inversions only when necessary along the well trajectory. The current modeling algorithm has the potential to be further optimized for hybrid GPU-CPU implementation to enable real-time 2D and 3D inversions.
Permeability assessment in rocks with complex pore structure requires reliable quantification of pore-body/ throat-size distribution as well as tortuosity. Pore body/ throat size can often be estimated within each rock type using nuclear magnetic resonance (NMR) T2 and mercury injection capillary pressure (MICP) measurements. Tortuosity, however, is challenging to quantify. Tortuosity can be quantified in the pore-scale domain through modeling of electrical current flow, fluid flow, heat transfer, and molecular diffusion, which can be challenging in the presence of microporosity. These methods also provide non-unique estimates of tortuosity. It is also challenging to upscale these tortuosity estimates to the well-log domain for depth-by-depth assessment of tortuosity and permeability. The objectives of this paper are (a) to quantify hydraulic and electrical tortuosity in the pore-scale domain in the presence of microporosity, (b) to quantify the impacts of hydraulic and electrical tortuosity on permeability assessment, (c) to upscale flow-based tortuosity estimates to the well-log domain, and (d) to quantify depth-by-depth permeability in the well-log domain honoring flow-based tortuosity. To achieve these objectives, we developed a workflow that integrates NMR and MICP measurements, as well as pore-scale images to estimate depth-by-depth permeability in the well-log domain. The introduced permeability model takes as inputs the pore-body-size distribution, constriction factor, and tortuosity of the rock. We use NMR data corrected for the impacts of reservoir fluids and mud for depth-by-depth assessment of the pore-body-size distribution. Next, we use MICP measurements within each rock class to estimate pore-throat-size distribution. Then, the estimated pore-body and pore-throat-size distribution are adopted to assess the constriction factor in each rock class. We conduct numerical simulations of electrical potential distribution and fluid flow to estimate the electrical and hydraulic tortuosity, respectively, in each rock class, honoring a representative elementary volume (REV). We successfully applied the proposed workflow to a complex carbonate formation in the Brazilian presalt reservoirs. By the integration of fluid-based tortuosity with NMR and MICP measurements in the well-log domain, permeability estimates indicated 61.1% improvement compared to those obtained from the integration of NMR, MICP, and electrical tortuosity. Results demonstrated the importance of hydraulic tortuosity assessment for reliable permeability quantification in rocks with complex pore structure. The novelty of this workflow is the integration of multiscale formation data (i.e., pore-scale image analysis and well-log-scale measurements) for enhanced interpretation of fluid flow in spatially heterogeneous carbonate formations. The introduced method also enables depth-by-depth assessment of hydraulic tortuosity in the well-log domain.
The understanding of fluid phases in reservoirs poses significant challenges, particularly when dealing with depletion, moving contacts, gas injection, or any physical phenomenon. Such conditions can complicate conventional petrophysical evaluations, and neutron-density separation and borehole resistivity responses can lead to ambiguous interpretation. To address these complexities and make informed reservoir development decisions, results from various techniques and sources, such as advanced mud gas analysis, fluid sampling, and petrophysical data, are integrated to gain additional insights while drilling. In the example of a depleted oil field, zones of interest were identified based on the analysis of mud gas and conventional logs. However, the presence of both oil and gas phases in the transition zone made it difficult to get an accurate fluid interpretation. Consequently, a sampling tool is deployed to confirm the flowing phase, and new cutoffs for the mud gas measurement (as per Cely et al., 2023 and Bravo et al., 2024b) in this reservoir were established, allowing calibration of subsequent wells in the same field. Post-well pressure-volume-temperature (PVT) analysis showed consistent fluid properties with the realtime measurements taken by the sampling tool. Moreover, this paper presents additional cross-comparison plots to reinforce confidence in technology adoption and its added value.
The unconventional reservoirs in the mixed carbonate/ clastic basinal facies of the Vaca Muerta Formation (Tithonian-Valanginian) located in the Neuqu & eacute;n Basin (Argentina) are considered one of the most prolific shale-oil and shale-gas reservoirs in the world. In combination with the Quintuco Formation, they represent prograding sequences composed of proximal carbonate facies (Quintuco) and distal shale facies (Vaca Muerta). With an extensive surface area of over 30,000 km2, the Vaca Muerta Formation is a key focus for hydrocarbon exploration and production. Drill cuttings from two horizontal wells (one targeting the Org & aacute;nico Inferior member and the other the Cocina member) were subjected to a comprehensive and rapid reservoir characterization analysis using a new cuttings-based, standardized workflow. This workflow integrates high-resolution (HR) imaging, artificial intelligence (AI) algorithms using a customized image analysis model, and X-ray fluorescence (XRF) elemental data, allowing for a consistently measured and detailed characterization of all drill-cuttings samples from both wells. Additionally, X-ray diffraction (XRD) analyses were conducted on selected samples to enhance the identification and classification of specific intervals of interest. The study successfully identified 22 distinct lithotypes based on a combination of physical and geochemical parameters, including RGB, calibrated brightness, elemental volumes, and key ratios (e.g., Si/Ca). These lithotypes were further classified into lithofacies associations and lateral well sedimentary packages, with detailed analysis highlighting significant vertical and lateral variations within the Org & aacute;nico Inferior and Cocina members. Tuffaceous, heterolithic, and calcareous components were distinguished using this approach, providing a cost-effective methodology for reservoir characterization. Variability in clay speciation, detrital silica content, and anoxia indicators was also identified, which were linked to deviations in the bit trajectory and geosteering during drilling. High mineral luminance peaks identified from image analysis of the ultraviolet (UV) light images were interpreted as volcanic tuff layers and corroborated by XRD data showing the presence of igneous minerals and clay minerals formed from volcanic glass weathering. These tuffaceous intervals, commonly observed in the Vaca Muerta Formation, are critical for understanding their effects on bit performance and potential trajectory deviations during drilling. This new drill-cuttings-based workflow offers an efficient, quantitative, consistently measured, and standardized approach to reservoir characterization, automatically carried out in a closed and controlled environment using readily available drill cuttings. It has the potential to optimize well planning by defining target zones, improving the understanding of the formation's depositional framework, and refining the static model as more wells and data become available.
A vast amount of petrophysical information can be inferred from continuous longitudinal (T1) and transverse (T2) relaxation time distributions of nuclear magnetic resonance (NMR) data from laboratory and well-logging settings. As an NMR post-processing step, numerous methods, like manual cutoffs or machine learning, partition one-dimensional (1D) and two-dimensional (2D) distributions to quantify pore systems and fluids. Challenges remain, caused by the continuous distributions derived from the inverse Laplace transform (ILT) of raw NMR data. This study develops new post-processing workflows based on the discrete inversion method, demonstrated through applications to fluid typing and quantification for synthetic and shale data sets. Due to its ill-posed nature, the ILT method applies constraints to obtain stable, continuous solutions. While often treated as aproxy for ground truth, such solutions pose challenges for fluid partitioning. The smoothing effects inherent to these solutions are demonstrated with synthetic data sets of known ground truth, as true distributions of actual samples are unknown. We present a framework using a discrete inversion approach, detailing its implementation across various laboratory data sets, its compatibility with ILT, and its enhancement through component optimization. This methodology shows promising results for challenging tasks, including the characterization of shale fluids and the analysis of logging data with low signal-to-noise ratio (SNR). Notable limitations of conventional ILT methods include: (1) excessive smoothing, resulting in substantial peak overlap and diminished spectral resolution, and (2) considerable uncertainty in estimating short relaxation times, particularly at low SNR conditions. These oversmoothing challenges subsequent partitioning methods, as smoothed features cannot be distinguished from the true distribution, and inversion inaccuracies are propagated into post-processing. In the discrete inversion approach presented, the number of components is estimated using the Bayesian information criterion (BIC), and neighboring components can be merged to capture broad distributions. These discrete components differ fundamentally from ILT-based results, as they inherently eliminate a separate partitioning step. With this method, we demonstrate that fluid classification and quantification are simplified once fluid relaxation time ranges are calibrated for a reservoir. A further advantage is the method's low bias when analyzing low SNR logging data; while the variance of individual results can be high, this resilience to systematic error enhances the reliability of the average result.
