The morphology of natural micro-fractures in shale reservoirs is a key factor controlling their fluid flow capacity and mechanical stability, while the microscopic distribution of minerals significantly influences the development characteristics of local micro-fractures. Accurately extracting the geometry of micro-fractures and establishing its relationship with mineral types and spatial distribution are essential for a deeper understanding of wellbore instability mechanisms in shale formations. However, due to the strong heterogeneity of the shale matrix, conventional threshold-based segmentation methods struggle to precisely distinguish micro-fractures from mineral boundaries, leading to considerable uncertainty in the extraction of fracture morphological parameters. To address this issue, this study proposes a TS-LSTM fracture extraction method based on scanning electron microscopy (SEM) images, which combines threshold segmentation with a long short-term memory neural network to achieve high-precision segmentation and completion of micro-fractures. Using the extracted fracture morphologies, the width and tortuosity of the fractures are quantitatively characterized. To quantify the mineral distribution around the fractures, different distances outward from the fracture boundaries are defined, and the area percentage of a specific mineral within each distance zone is designated as the threshold mineral percentage content. On this basis, correlation analysis is applied to investigate the statistical relationships between the local content of three major minerals-quartz, albite, and illite-and fracture width and tortuosity. The results show that the TS-LSTM fracture extraction method can effectively extract micro-fracture regions from complex shale SEM images, with strong completion capability particularly for discontinuous fractures. Using the threshold mineral percentage content at different distances, the mineral distribution around fractures can be quantitatively described. Illite content exhibits a negative correlation with fracture width and a strong positive correlation with tortuosity, indicating that fractures in illite-rich zones are narrower and more tortuous. Quartz content is positively correlated with fracture width and overall negatively correlated with tortuosity, which favors the formation of wider and straighter fractures. However, in local areas with dense quartz grains, fractures may propagate around the grains, leading to increased local tortuosity near quartz. Although albite content shows a certain positive correlation with fracture width, its relationship with tortuosity is more complex. In summary, the type and spatial distribution of minerals collectively shape the complex propagation paths of fractures. This study establishes, through an intelligent approach, the relationship between minerals and micro-fracture morphology, providing a new pathway for developing micro-scale models of wellbore stability in shale formations.

The immediate undrained poroelastic response after drilling is critical for short-term wellbore stability in low-permeability formations. Despite its significance, existing analytical undrained wellbore stress solutions in transversely isotropic formations are limited to special cases where the borehole axis is either perpendicular or parallel to the plane of transverse isotropy. By extending Amadei's elastic framework into the undrained poroelastic regime, this study presents an analytical undrained solution for arbitrarily inclined boreholes in transversely isotropic formations, considering fully anisotropic undrained poroelastic effects. Validation against undrained isotropic and transversely isotropic solutions, as well as anisotropic Biot's effective stress coefficients, demonstrates the robustness of the proposed model. The derived Biot's effective stresses are subsequently applied to evaluate short-term collapse pressure with varying wellbore trajectories. Results show that undrained wellbore stress distributions deviate significantly from drained Amadei solution with varying elastic property anisotropy and wellbore pressure, particularly in terms of axial and shear stress components. Furthermore, estimating the undrained pore pressure response using drained Amadei stress solution introduces substantial errors, highlighting the necessity of adopting a fully anisotropic undrained poroelastic framework. Unlike isotropic formations, the anisotropic undrained pore pressure response exhibits an asymmetrical distribution, reflecting coupled shear-tension and solid-fluid couplings. Polar plots of collapse pressure that incorporate the undrained effect exhibit significant differences compared to those derived from previous models, indicating that the optimal drilling direction may not coincide with the orientation of the horizontal principal stress. This analytical solution establishes a benchmark for fully anisotropic undrained wellbore stress distribution and holds substantial implications for enhancing early-stage wellbore stability assessments in strongly anisotropic poroelastic formations.

Petroleum science stands at a historic turning point: exploration extends to ultra-deep and deepwater frontiers, unconventional resources become dominant, and the dual imperatives of carbon neutrality and artificial intelligence are reshaping the industrial landscape and disciplinary boundaries. Building upon the “100 Grand Challenges in Petroleum Science” initiative launched by Petroleum Science in 2021, this paper presents a curated list of 100 fundamental scientific problems through systematic literature synthesis, expert deliberation, and multi-criteria selection. The challenges span six domains: exploration and development, transport and pipelines, refining and petrochemicals, materials science, carbon capture, utilization and storage-geothermal-hydrogen, and energy economics and digital transformation. In-depth analysis reveals three archetypes: mechanism-oriented problems that seek unified theories of multi-scale coupling, technology-oriented problems that push engineering limits under extreme conditions, and system-level problems that address full-chain integration and energy transition. A key insight is that artificial intelligence has evolved from an enabler to a core driver, and together with the low-carbon transition it is redefining the very paradigm of petroleum science. This roadmap is intended to guide research planning, investment, and talent development over the coming decades and to foster interdisciplinary collaboration on a global scale.

The temporal evolution of reservoir rock microcracks and other defects in the chemical-mechanical environment is the key issue causing rock strength deterioration and even macroscopic fracturing. It is of great significance to explore the development process of the cross-scale behavior of damage evolution induced by microcrack growth and destabilization damage induced by damage accumulation to ensure the stability of rocks surrounding the wellbore to efficiently explore and develop oil and gas resources. Firstly, three basic behaviors and characteristics of microcrack growth in rocks are introduced, and it is proposed that the hydration process of shale is the mechanism by which the internal microcracks continuously undergo more active subcritical growth, and the collapse instability of rocks surrounding the wellbore is more often manifested as the time-sequential damage evolution of microcracks from subcritical growth to dynamic fracture. Based on this understanding of the study of wellbore stability, in order to reveal the subcritical fracture mechanism of microcracks in rock under chemical-mechanical coupling, and to quantitatively characterize the damage evolution process and destabilizing mutation behaviors, the research team proposes the concept of "Chemical Fracture", which focuses on the fracture mechanics theory of microcrack growth in a chemical-mechanical environment, and the response of rock mechanical parameters to the structural damage evolution. Then the current research status of the subcritical fracture mechanism of rock chemical-mechanical coupled microcracks is summarized from the perspectives of theoretical modeling and experimental testing, respectively. Finally, several specific problems and challenges faced in the current research work are listed, and new thoughts and perspectives into microcrack growth evolution and fracture mechanisms of brittle rocks in chemical-mechanical environments are elaborated.

Hydraulic fracturing of highly deviated wells is an effective means to develop ultra-deep oil and gas reservoirs. The transportation law of proppant becomes more complicated due to the complex nonplanar morphology formed by the distortion and extension of fractures in three-dimensional space. To investigate the transportation of proppant in nonplanar fracture, a fluid-solid coupling model of proppant transportation in the 3D nonplanar fracture was established and solved based on computational fluid dynamics (CFD) and discrete element method (DEM). The effects of displacement, proppant density, proppant size and fluid viscosity on proppant transportation and distribution were studied. The results show that: The shape of nonplanar fractures has a significant influence on proppant transportation. Compared with the flow form in planar fracture, the flow in nonplanar fracture appears eddy current. The collision frequency between proppant and fracture wall and other proppants increases, which increases the energy loss during sand-carrying fluid flow. By increasing the displacement, sand plugging in the near-wellbore area can be avoided and the length of sand bank in the fracture is increased. As proppant density and proppant size decrease, the length of the sand bank increases and the height decreases. High-viscosity fracturing fluid can effectively carry proppant for transportation and avoid proppant settlement near the wellbore to occur sand plugging. This study clarifies the laws of proppant transportation and distribution in the nonplanar fracture, and helps guide the design of proppant pumping construction parameters for ultra-deep highly deviated well fracturing.

Offshore drilling costs are high, and the downhole environment is even more complex. Improving the rate of penetration (ROP) can effectively shorten offshore drilling cycles and improve economic benefits. It is difficult for the current ROP models to guarantee the prediction accuracy and the robustness of the models at the same time. To address the current issues, a new ROP prediction model was developed in this study, which considers ROP as a time series signal (ROP signal). The model is based on the time convolutional network (TCN) framework and integrates ensemble empirical modal decomposition (EEMD) and Bayesian network causal inference (BN), the model is named EEMD-BN-TCN. Within the proposed model, the EEMD decomposes the original ROP signal into multiple sets of sub-signals. The BN determines the causal relationship between the sub-signals and the key physical parameters (weight on bit and revolutions per minute) and carries out preliminary reconstruction of the sub-signals based on the causal relationship. The TCN predicts signals reconstructed by BN. When applying this model to an actual production well, the average absolute percentage error of the EEMD-BN-TCN prediction decreased from 18.4% with TCN to 9.2%. In addition, compared with other models, the EEMD-BN-TCN can improve the decomposition signal of ROP by regulating weight on bit and revolutions per minute, ultimately enhancing ROP.

Numerical simulation is an essential technique for CO₂ geological storage operations. However, high-resolution geological models typically consist of a large number of grid blocks, making numerical simulations computationally expensive and time-consuming. Upscaling methods are commonly used to coarsen the fine-scale geological model, with global flow-based upscaling methods generally demonstrating the highest accuracy. However, since these methods require solving flow equations over the global domain, which is still time-consuming, their applications are typically limited to cases where the coarse model is reused repeatedly (e.g., history matching or optimization). To overcome these limitations, this study develops a novel deep learning (DL)-based upscaling framework for the simulation of CO₂ injection into saline aquifers. The framework incorporates convolutional neural networks (CNNs), Transformer encoders, and Fourier neural operators (FNOs) to construct surrogate models for upscaled well index, permeability, relative permeability, and capillary pressure. A preprocessing procedure is first applied to address the issue of inaccurate upscaled parameters, which are typically caused by weak flow conditions in traditional upscaling computations. Then the surrogate models are trained using relevant local information, and the trained surrogate models are used to replace traditional numerical upscaling computations, enabling instantaneous and parallel predictions of upscaled parameters. Two representative flow patterns (left-to-right and bottom-to-top) are considered to evaluate the framework's performance. The results demonstrate that the DL-based framework significantly improves computational efficiency, achieving a speedup factor of approximately 1133 times compared to traditional upscaling methods. Additionally, it maintains or even enhances simulation accuracy, as the surrogate models correct inaccurate upscaled parameters.

Wellbore stability is a key factor affecting safe and efficient drilling. At present, it is difficult to conduct real-time and accurate analysis of wellbore stability in related research. To address the current research shortcomings, this study proposes a real-time analysis model of wellbore stability integrating image recognition and an expert system, which mainly includes caving image segmentation and recognition, and a wellbore stability expert system. The caving image recognition proposes a new dynamic threshold segmentation method based on simple linear iterative clustering superpixel segmentation and visual geometry group 19-layer image classification. After completing the segmentation of the caving image, the geometric features of the caving are calculated, and the multi-source feature fusion GoogleNet model is established by integrating the geometric features with the convolution features extracted by GoogleNet to identify the caving types efficiently. After segmentation and recognition of caving images. The wellbore stability expert system uses the caving features to establish an expert system model to determine the mechanism of wellbore instability and provide reasonable solutions. Finally, the wellbore stability integrating image recognition and an expert system model was applied to a well in field production, accurately determining the mechanism of wellbore instability in real time and effectively solving the corresponding wellbore instability problem based on the measures provided by the model.

Deep shale gas reservoirs have geological characteristics of high temperature, high pressure, high stress, and inferior ability to pass through fluids. The multi-stage fractured horizontal well is the key to exploiting the deep shale gas reservoir. However, during the production process, the effectiveness of the hydraulic fracture network decreases with the closure of fractures, which accelerates the decline of shale gas production. In this paper, we addressed the problems of unclear fracture closure mechanisms and low accuracy of shale gas production prediction during deep shale gas production. Then we established the fluid–solid–heat coupled model coupling the deformation and fluid flow among the fracture surface, proppant and the shale matrix. When the fluid–solid–heat coupled model was applied to the fracture network, it was well solved by our numerical method named discontinuous discrete fracture method. Compared with the conventional discrete fracture method, the discontinuous discrete fracture method can describe the three-dimensional morphology of the fracture while considering the effect of the change of fracture surface permeation coefficient on the coupled fracture–matrix flow and describing the displacement discontinuity across the fracture. Numerical simulations revealed that the degree of fracture closure increases as the production time proceeds, and the degree of closure of the secondary fractures is higher than that of the primary fractures. Shale creep and proppant embedment both increase the degree of fracture closure. The reduction in fracture surface permeability due to proppant embedment reduces the rate of fluid transfer between matrix and fracture, which has often been overlooked in the past. However, it significantly impacts shale gas production, with calculations showing a 24.7% cumulative three-year yield reduction. This study is helpful to understand the mechanism of hydraulic fracture closure. Therefore, it provides the theoretical guidance for maintaining the long-term effectiveness of hydraulic fractures.

Due to the complexity and variability of carbonate formation leakage zones, lost circulation prediction and control is one of the major challenges of carbonate drilling. It raises well-control risks and production expenses. This research utilizes the H oilfield as an example, employs seismic features to analyze mud loss prediction, and produces a complete set of pre-drilling mud loss prediction solutions. Firstly, 16 seismic attributes are calculated based on the post-stack seismic data, and the mud loss rate per unit footage is specified. The sample set is constructed by extracting each attribute from the seismic trace surrounding 15 typical wells, with a ratio of 8:2 between the training set and the test set. With the calibration results for mud loss rate per unit footage, the nonlinear mapping relationship between seismic attributes and mud loss rate per unit size is established using the mixed density network model. Then, the influence of the number of sub-Gausses and the uncertainty coefficient on the model's prediction is evaluated. Finally, the model is used in conjunction with downhole drilling conditions to assess the risk of mud loss in various layers and along the wellbore trajectory. The study demonstrates that the mean relative errors of the model for training data and test data are 6.9% and 7.5%, respectively, and that R² is 90% and 88%, respectively, for training data and test data. The accuracy and efficacy of mud loss prediction may be greatly enhanced by combining 16 seismic attributes with the mud loss rate per unit footage and applying machine learning methods. The mud loss prediction model based on the MDN model can not only predict the mud loss rate but also objectively evaluate the prediction based on the quality of the data and the model.