Laminae play a critical role in controlling the enrichment and high yield of shale oil through the pores within individual laminae and the microfractures between different laminae. Due to the lack of effective methods, research on the contribution of laminar fractures to shale reservoirs has been delayed, primarily relying on indirect evaluation through comparisons of the reservoir properties of laminar shale and massive shale. This has hindered a comprehensive understanding of the role of laminar fractures in the storage and migration of shale oil. Using Shahejie shale from the Jiyang Depression as a case study, this research employs various imaging technologies, including scanning electron microscopy (SEM), Modular Automated Processing System (MAPS), quantitative mineral evaluation using scanning electron microscopy (QEMSCAN), and focused ion beam scanning electron microscopy (FIB-SEM), to conduct both two-dimensional and three-dimensional imaging-based quantitative assessment of the pore structure within individual laminae. The surface porosity in various laminae is ranked as follows: felsic lamination > granular calcite lamination > mixed lamination > clay-organic matter lamination > dolomite lamination > sparry calcite lamination > organic matter lamination. A similar trend in pore size, porosity, and pore connectivity within laminae was observed using SEM and FIB-SEM imaging. Matrix porosity was quantified through binary image processing, and the storage capacity of laminar fractures was assessed by combining this with the measured porosity. Laminated fractures contribute approximately 36.03% to shale storage capacity under laboratory conditions. However, under in-situ conditions, the porosity of laminated fractures decreases from 1.51% to 0.44%, reducing their contribution to total porosity to 10.61%. This study, through extensive image analysis and comparisons, fills the gap in directly quantifying the contribution of laminar fractures to storage capacity. It provides valuable insights into why laminar shales are viable exploration and development targets, offering new perspectives on the differences in hydrocarbon enrichment across various types of laminar shales.
Diffusion models have achieved breakthrough progress in single-object video editing. However, existing methods still suffer from two limitations when applied directly to multi-object editing: (1) These methods struggle to understand the spatial relationships and layouts of multi-object in the video, leading to semantic confusion and misalignment. (2) They lack explicit semantic consistency supervision in complex scenarios and editing targets, resulting in generated results deviating from text prompt. To address the above limitations, we propose SemEdit, a novel semantic-aware content alignment scheme for fine-grained multi-object video editing. Specifically, SemEdit comprises two key components: Semantic Adaptive Modulation and Semantic Prior Modeling. The former utilizes region-aware and text-aware attention modulation to achieve feature decoupling and precise object localization by adjusting the spatial semantics distribution. Collaboratively, the latter leverages semantic prior to model high-level content correlations, thus ensuring alignment between the edited video and input prompt. Benefiting from the above designs, our method achieves precise spatial object alignment while maintaining semantic consistency across the entire video. Extensive experiments show that our SemEdit outperforms existing video editing methods in both editing precision and semantic alignment, providing a new perspective for multi-object video editing.
Accurate identification of microscale pore structures in shale reservoirs is crucial for oil and gas reserve evaluation and recovery strategy optimization. This task faces the dual challenges of scarce labeled data and strong geological heterogeneity. We incorporate semi-supervised learning methodologies to leverage large volumes of unlabeled data, thereby mitigating the challenge of data scarcity, but existing methods typically treat all image regions equally, failing to fully exploit the discriminative features of heterogeneous areas. To overcome this, we propose HB-Net, a geologically informed semi-supervised model that explicitly guides the model to focus on geologically heterogeneous regions, achieving high-precision pore identification with low annotation dependency. We propose a heterogeneity representation method that effectively locates highly heterogeneous and pore-enriched image regions by integrating average grayscale, entropy, and pore density information. Additionally, we design a heterogeneity-aware contrastive learning approach to enhance the model’s discriminative capability in complex areas. Experiments conducted on the shale SEM dataset demonstrate that with only 10% labeled samples, HB-Net achieves a pore intersection over union of 75.02% on the validation set. This represents a 7.11% improvement over the supervised baseline and a 1.34% gain over the advanced semi-supervised baseline, with a particularly significant enhancement (1.54%) in heterogeneous regions. This study marks the first integration of geological heterogeneity as an explicit guiding signal within a semi-supervised learning framework, providing a high-precision, low-data-dependent solution for micropore identification. This approach holds practical significance for the intelligent characterization of unconventional oil and gas reservoirs.
Deep coal-rock reservoirs are considered excellent carriers for CO2 sequestration due to their widespread distribution and substantial storage capacity. However, the sealing performance of these reservoirs is compromised by the development of natural fractures and pronounced heterogeneity, making them susceptible to CO2 leakage. To address this challenge, this study developed a novel CO2-enhanced sequestration system specifically designed for deep coal seams. Sodium silicate was selected as the primary agent in the system by using the interaction energy between the agent and the initiator as the screening criterion. The system formulation was then optimized through experimental design. The solidification mechanism was elucidated, revealing the formation of a three-dimensional cross-linked network via dehydration condensation between silanol groups. Silicon-oxygen bonds and the aromatic skeleton serve as key structural units, imparting high strength and adaptability to the system. Injection and plugging experiments demonstrated that the optimized system exhibits excellent and broadly applicable plugging performance. In fractured coal-rock masses with permeability ranging from 60 mD to 1600 mD, the residual resistance coefficient remained high, decreasing only from 31 to 10, indicating that the system effectively blocks seepage channels across a wide permeability range. These results confirm that the developed system can significantly enhance reservoir CO2 sequestration by sealing fractures, provides a promising material approach and theoretical basis for CO2 sequestration in deep coal seams.
Dry reforming of methane (DRM) provides a sustainable route for converting CH4 and CO2 into syngas, yet the high-temperature reforming environment often accelerates Ni nanoparticle sintering and carbon accumulation, leading to rapid catalyst deactivation. Stabilizing highly dispersed Ni species within a robust interfacial environment is therefore essential for durable DRM catalysis. Herein, we report a silane-assisted confinement strategy using 3-aminopropyltrimethoxysilane to construct a thermally robust Ni/A-SiO2 catalyst. The silane coupling chemistry induces the in situ formation of Si-O-Ni interfacial linkages, which strengthen the metal-support interaction and confine ultrafine Ni nanoclusters within an amorphous SiO2 matrix. This confined interfacial architecture effectively suppresses Ni migration, coalescence, and carbon nucleation during DRM. Consequently, the optimized 4Ni/A-SiO2 catalyst delivers CH4 and CO2 conversions of 93.06 % and 95.19 %, respectively, at 700 degrees C and a GHSV of 10,600 h-1, and remains stable for 400 h with negligible coke deposition. These results demonstrate that silane-mediated interfacial confinement provides an effective strategy for stabilizing nonprecious metal catalysts under harsh DRM conditions.
