Deepwater gas field development faces complex challenges such as ultra-deep water, strong multi-field coupling, and high operational risks, including hull stability risks, difficulties in reservoir characterization, limited accessibility of monitoring data, and the complexity of integrated production management. Traditional approaches often rely on fragmented single-module simulations and manual decision-making, resulting in delayed model updates, isolated information, and the inability to achieve end-to-end collaborative optimization across reservoirs, pipeline networks, and platforms. Digital-twin technology overcomes these limitations by breaking down data silos, enhancing model coupling, and reducing decision latency. It enables real-time interaction, closed-loop control, and full-chain integrated management, thereby eliminating information barriers among reservoirs, wellbores, pipelines, and platforms and providing coordinated, efficient production control and safety assurance for deepwater gas fields. Focusing on the “Deep Sea No. 1” production platform, this study explores the construction of a production digital-twin system that spans the entire business chain of reservoir-wellbore-pipeline network-platform-operation-finance. First, the research progress of digital twins in oil and gas production is systematically reviewed, including typical modeling methods, technical frameworks, and engineering practices. Secondly, a modular hybrid modeling approach integrating physical mechanism models and data-driven models is proposed, establishing a complete modeling workflow comprising system decomposition, model construction, data integration, optimization solving, and feedback control. Third, based on the actual application scenario of the “Deep Sea No. 1” platform, digital-twin modules are developed for mooring and hull management, flow assurance management, intelligent reservoir management, and 3D visualization, enabling early warning, predictive maintenance, decision support, immersive visualization, and full-chain closed-loop control. Field application results demonstrate that the system significantly improves the automation and intelligence of the platform, reducing production allocation calculation time from 4-5 days to less than 1 hour with prediction accuracy exceeding 90%. Finally, in response to current issues such as limited model transferability and heavy manual intervention, this paper suggests establishing a linkage framework of large and small models, strengthening integration with subsea control systems, and building a full-lifecycle digital-twin system. The research results provide a feasible technical pathway and engineering reference for the intelligent and efficient development of deepwater gas fields.

With the continuous advancement of oil and gas exploration technologies, unconventional hydrocarbon reservoirs have emerged as a pivotal domain for global energy resource augmentation and production enhancement. However, the inherent characteristics of low permeability, dense rock matrix, and complex heterogeneity in these reservoirs pose substantial challenges to conventional logging interpretation methodologies, particularly in constructing theoretical models, deriving empirical formulas, and inverting reservoir parameters, thereby hindering accurate reservoir identification and efficient development. The recent breakthroughs in artificial intelligence (AI) technologies have provided innovative solutions for logging interpretation in unconventional reservoirs. Through systematic analysis of cutting-edge research achievements worldwide, this paper first elucidates the core geological characteristics and evaluation challenges of unconventional reservoirs. Subsequently, it comprehensively summarizes the implementation modalities and operational efficacy of AI techniques, including machine learning and deep learning algorithms, in critical logging interpretation processes such as lithology identification, porosity prediction, permeability estimation, and hydrocarbon-bearing potential assessment. The study particularly highlights the transformative capabilities of convolutional neural networks in processing multi-scale logging data, recurrent neural networks in handling time-series measurements, and ensemble learning approaches in enhancing prediction accuracy under high-dimensional parameter spaces. The research demonstrates that AI-driven approaches achieve remarkable performance improvements compared to conventional methods, with reported accuracy enhancements of 25%~40% in lithofacies classification and 15%~30% reduction in mean absolute error for porosity estimation across various case studies. Furthermore, advanced deep learning architectures have shown exceptional capability in capturing nonlinear relationships between logging responses and reservoir properties, effectively addressing the “low signal-to-noise ratio” dilemma common in unconventional reservoir evaluation. A critical evaluation is conducted from multiple dimensions, including data quality requirements, algorithmic adaptability, computational efficiency, and model interpretability. The analysis reveals that while data-driven models excel in pattern recognition, their physical consistency and generalization capability require further improvement, particularly when dealing with spatially heterogeneous formations and limited training datasets. To address these challenges, the paper proposes three strategic development directions: (1) Hybrid modeling frameworks integrating physical constraints with data-driven approaches. (2) Transfer learning schemes for small-sample learning scenarios. (3) Multi-modal data fusion architectures incorporating logging, core, and seismic information. Moreover, the study emphasizes the necessity of establishing standardized workflows for feature engineering, model validation, and uncertainty quantification in AI-based logging interpretation systems. Special attention is given to emerging technologies such as graph neural networks for 3D reservoir characterization and physics-informed neural networks for incorporating petrophysical laws into machine learning architectures. This comprehensive review not only synthesizes the current state-of-the-art in intelligent logging interpretation but also provides a strategic roadmap for future research endeavors. The findings offer valuable theoretical references and methodological guidance for optimizing AI-based interpretation techniques in unconventional reservoir evaluation, ultimately contributing to more reliable reservoir characterization and enhanced hydrocarbon recovery in complex geological settings.

The rapid advancement of generative artificial intelligence (AI), exemplified by models like the DeepSeek series, has substantially lowered the application barrier for large language model (LLM) technology. This progress injects new intelligent capabilities into the field of oil and gas exploration and development, a domain highly dependent on expertise and data-intensive analysis. However, the practical capabilities and implementation pathways of large language models in vertical industry scenarios remain unclear. This study comprehensively and systematically evaluates the multidimensional application capabilities of the DeepSeek model series within this specific domain. A comprehensive six-dimensional quantitative assessment framework was designed to evaluate core competencies, including foundational domain knowledge, complex reasoning, computational proficiency, multimodal processing, performance on open-ended and innovative problems, and professional task execution capabilities. The testing results indicate that LLMs demonstrate exceptional performance in terms of breadth of foundational knowledge coverage and in handling open-ended, innovative questions, revealing strong domain-specific comprehension and application and significant potential for interdisciplinary knowledge integration. However, several critical limitations were identified. The models exhibit hallucination risks when processing specific instances and data, display a lack of sufficient granularity in logical reasoning within complex problem-solving scenarios, and show deficiencies in the accuracy and efficiency of intricate numerical computations. Furthermore, distinct capability boundaries were observed, particularly in multimodal processing - especially the generation and interpretation of professional diagrams and images -, in the operation of specialized software, and responsiveness to real-time engineering demands. To address these identified limitations, this paper proposes an integrated four-dimensional technical pathway to facilitate intelligent transformation. This cohesive strategy comprises: 1) a dynamic knowledge fusion mechanism based on Retrieval-Augmented Generation (RAG) to mitigate knowledge obsolescence and data hallucinations; 2) a knowledge-graph-driven reasoning engine designed to enhance logical reasoning precision for complex problems; 3) a specialized software collaboration architecture that extends the model’s operational boundaries via API gateways integrating domain-specific tools; and 4) an Agent-empowered engineering system for the automated decomposition and execution of complex tasks. The research further delves into key technical challenges, such as the construction of vertical domain knowledge graphs, software ecosystem interoperability, and real-time decision-making by AI agents, proposing targeted directions for technological breakthroughs. In conclusion, the deep integration of LLMs into the oil and gas sector necessitates tight coupling with domain knowledge engineering, specialized software ecosystems, and edge computing technologies. The transition from point solutions to systemic intelligence should be gradual, starting with focused scenario development, overcoming core technical bottlenecks, and ultimately realizing the synergistic application of “Data-Knowledge-Tools.”

The ongoing Fourth Industrial Revolution, characterized by artificial intelligence (AI), is driving a wave of intelligent transformation across the oil and gas industry. International oilfield service and operating companies in regions such as the United States, Norway, and the Middle East are investing heavily in digital and intelligent transformation to secure a competitive advantage in the future landscape. As a cutting-edge technology in oil and gas engineering, intelligent drilling and completion is poised to yield disruptive and leapfrog innovations, empowering the development of new quality productive forces. This paper elaborates on the conceptual framework of intelligent drilling and completion technologies and provides a comprehensive review of the current domestic and international development status and technical disparities across five key areas: surface equipment, measurement-transmission-conduction tools, drilling fluids, cementing, and software. It further identifies four major existing problems and four key challenges. Furthermore, it proposes the “1244” development direction: focusing on one goal—empowering technological innovation and industrial upgrading through “Drilling and Completion + AI”; concentrating on two major fields—intelligent drilling and intelligent cementing; tackling four key technological directions—equipment, tools, fluids, and software; and realizing four typical application scenarios—fully automated wellsite operations, autonomous drilling control, intelligent drilling fluid regulation, and intelligent cementing operations. Simultaneously, the article advocates building a technical system with the geology-engineering knowledge base as the foundation, large/small AI models as the backend support, and “Drilling and Completion + AI” as the front-end application carrier. This aims to accelerate the transition of intelligent drilling and completion technology from unit-level closed-loop to global closed-loop and remote-controlled operations.

In the process of petroleum exploration and development, long-term accumulated documents contain a large amount of engineering knowledge and practical experience, and these materials are of great significance for the scientific development of oilfields and production decision-making. However, such information is mostly preserved in multimodal and unstructured forms such as textual descriptions, data tables, and illustrative figures, lacking a unified structured representation, which leads to low efficiency in retrieval and utilization, and makes the knowledge difficult to be systematically applied. Traditional information retrieval methods have limitations in dealing with complex cross-paragraph and multimodal corpora, and relying only on large-scale language models for question answering is prone to hallucinations and context fragmentation, which cannot meet the requirements of professional fields for accuracy and interpretability. To solve this problem, this paper, based on Microsoft’s open-source graph retrieval-augmented generation framework, constructs a graph retrieval-augmented intelligent question answering system for reservoir geology. Aiming at the linguistic complexity, hierarchical diversity, and structural heterogeneity of oilfield documents, three optimization methods were applied: a logical structure-based segmentation method was used to identify heading hierarchies and numbering rules to achieve reasonable division of semantic units, thereby avoiding semantic fragmentation in entity and relation extraction; a prompt optimization mechanism combined with the terminology system of reservoir geology was applied to improve the accuracy and completeness of entity and relation recognition and extraction, and to reduce errors and omissions; and a multimodal output mechanism was employed to realize the linkage of textual answers with relevant figures and tables through embedding matching, so that the results not only have linguistic coherence but also obtain visual evidence support, enhancing the interpretability and credibility of the answers. In the experimental part, a comprehensive report of about ninety thousand characters from a typical offshore oilfield was used as the data source to construct a knowledge graph and carry out system evaluation. Compared with unoptimized methods and the original framework, the results show that the optimized system has achieved significant improvements in factuality, answer relevance, context precision, and context recall. The improvement in factuality and answer relevance indicates that the system can more accurately generate answers that conform to facts and question intent, while the improvement in context indicators shows that it has greater advantages in cross-paragraph integration and multimodal association. The research results show that this system exhibits higher accuracy and reliability in knowledge extraction, organization, and application, has good engineering adaptability and scalability, not only provides a feasible solution for the structured management and intelligent utilization of complex oilfield knowledge, but also offers references and practical experience for the application of large language models in petroleum engineering and other highly specialized fields.

Driven by the “dual carbon” strategy, nuclear energy is increasingly recognized as a vital component of the energy mix due to its low-carbon and high-efficiency advantages. As the key raw material for nuclear power, the green extraction of uranium is of great significance. In-situ leaching (ISL) has emerged as a mainstream method for uranium extraction owing to its environmentally friendly and efficient nature. However, low-permeability uranium deposits present significant challenges during ISL, such as poor lixiviant flow and low uranium recovery rates. As a mature permeability enhancement technique, hydraulic fracturing can effectively create fracture networks through the injection of fracturing fluids, thereby improving formation permeability and enhancing lixiviant flow, which ultimately increases uranium recovery efficiency. This paper systematically reviews various ISL processes and their applicability, and thoroughly analyzes the mechanisms and current research status of advanced hydraulic fracturing technologies, including multi-stage fracturing, diversion fracturing, intelligent fracturing, acid fracturing, foam fracturing, and supercritical CO₂ fracturing. Based on the resource characteristics and extraction demands of low-permeability uranium deposits, this study proposes a synergistic application of ISL and hydraulic fracturing technologies. It further explores how hydraulic fracturing can enhance formation permeability and lixiviant efficiency. A development pathway integrating multiple technologies and intelligent optimization is suggested to achieve high-efficiency resource recovery and minimized environmental risks, providing robust support for the green and sustainable development of uranium mining.

This paper systematically reviews the current technical framework and implementation approaches for drilling digital twin modeling, focusing on key challenges such as multi-source heterogeneous data fusion during wellbore drilling operations, the coupling of physics-based and data-driven models, and intelligent decision feedback mechanisms. First, targeting the objects and levels of drilling data fusion, the fusion mechanisms for multi-temporal and multi-spatial scale data and conflict resolution methods were explored. Second, a comprehensive digital twin architecture suitable for drilling operating conditions was established, and methods for implementing joint physics-based and data-driven modeling were summarized. Then, diagnostic methods for drilling anomaly data features were proposed, enabling the establishment of multi-perspective decision feedback mechanisms between the physical and digital entities. Finally, the application potential of wellbore digital twin technology was prospected in the areas of high spatiotemporal resolution cognition, edge deployment, and model credibility assurance. The research results can provide theoretical support and methodological guidance for achieving drilling state recognition and efficient control under complex operating conditions.

The deep coalbed methane (CBM) resources (buried deeper than 1500 m) in North China exceed 30 trillion cubic meters, with significant breakthroughs in exploration, making it a new frontier for natural gas production growth. However, it faces key issues, including complex preservation conditions, unclear controls on enrichment and high productivity, and large production variations. In order to clarify the control factors of preservation conditions of deep CBM, the coal quality, lithology combination, and structural preservation characteristics of typical deep CBM exploration wells in North China are systematically analyzed with multidisciplinary analysis methods of geology, geochemistry, and geophysics. This study reveals that the differential preservation of deep CBM is controlled by the dynamic coupling mechanism of adsorption self-sealing, physical property sealing, and structural preservation system: (1) Favorable coal quality is the basis of strong adsorption. Coal with low ash content and a high evolution degree is characterized by a large Langmuir volume, a great critical depth, and strong self-sealing ability. (2) Tightly lithology combination is a barrier to deep free gas dissipation. Limestone and mudstone exhibit low porosity, a small throat radius, and high breakthrough pressure, resulting in good physical sealing conditions and a high total retained gas content. (3) Continuous and stable structural preservation is the key to free gas enrichment. Late-stage uplift with low amplitude and weak tectonic activity is conducive to free gas accumulation, resulting in a high-pressure coefficient. In summary, areas characterized by deep burial, low-ash coal, limestone/mudstone roof rocks, and structural stability are favorable for deep CBM preservation. The study of differential preservation of deep CBM deepens the understanding of enrichment and high productivity, and provides a theoretical basis for optimizing favorable exploration and development targets.

Carbonate laminae in saline lacustrine shale sequences are widely developed, and their diverse combinations and spatial stacking patterns represent a key geological factor controlling the differential enrichment of lacustrine shale oil. These laminae, composed of micro- to millimeter-scale thin layers of minerals such as calcite and dolomite, serve as preferred targets for saline lacustrine shale oil “sweet spots” due to their significant economic potential and development prospects. The laminae exhibit various structures, including cryptocrystalline, sparitic, and fibrous structures, which originate from diverse formation mechanisms—mechanical, physicochemical, and biochemical processes—leading to distinct mineralogical and structural heterogeneities. In terms of pore-fracture systems, carbonate laminae develop advantageous pore types such as intercrystalline, dissolved, and intergranular pores. Their high brittleness promotes the formation of heterogeneous fracture networks, including bedding-parallel and structural fractures. During diagenesis, organic acid dissolution and recrystallization further enhance pore connectivity, significantly improving in storage capacity and permeability. This facilitates the micro-migration and enrichment of light hydrocarbons within the laminae, forming movable oil. The control of carbonate laminae on shale oil sweet spots is manifested in four aspects: extensive distribution facilitates segmental and areal sweet spot development; high organic matter content provides the material basis for hydrocarbon generation; high-quality source-reservoir configurations promote near-source migration and enrichment; high brittleness enhances reservoir fracability and development potential. In conclusion, detailed study of carbonate laminae provides critical guidance for shale oil exploration and sweet spot prediction in saline lacustrine basins.

Carbonate fractured-vuggy reservoirs in the Middle-Lower Ordovician of the Tarim Basin are strongly controlled by deep strike-slip faults. Their pronounced heterogeneity has become a key challenge to the efficient exploitation of ultra-deep oil and gas resources. In this study, a representative strike-slip fault within a carbonate outcrop on the northwestern margin of the basin was selected as the research target. By integrating multiple analytical approaches, including field structural measurements, petrographic thin-section observations, high-pressure mercury intrusion porosimetry, and rock physics experiments, the study systematically characterizes the multi-scale heterogeneity of fault-controlled carbonate fractured-vuggy reservoirs and identifies their dominant controlling factors. The development patterns of high-quality reservoir zones are also summarized. At the macro scale, fracture distribution along the fault strike is highly uneven, with the overlap segments exhibiting the highest fracture densities. Within these segments, the boundary fault zones are characterized by small breccias with high roundness, indicating strong interconnectivity. Along the dip direction, fault core zones are distinguished by abundant large fractures and vugs with only minor vein filling. These zones have porosities approximately 4~6 times those of the damage zones and compressive strengths only 25%~50% as high, making them prime sites for high-quality reservoir development. Vertically, zones of high porosity and low strength alternate with sealing layers, resulting in a discrete vertical distribution of high-quality reservoirs. At the micro scale, fault cores exhibit diverse pore types, including intracrystalline pores, intercrystalline pores, semi-filled microfractures, and microvugs. In addition, they display high surface porosity, large pore aspect ratios, numerous interconnected pore nodes, and well-developed throat channels, which together contribute to their significantly higher permeability compared with the damage zones. On this basis, the Lorenz curve method is combined with the entropy weight method, which is applied for the first time to evaluate the heterogeneity of fault-controlled carbonate fractured-vuggy reservoirs. The evaluation results indicate that these reservoirs exhibit overall strong heterogeneity, with the micro scale showing a higher degree than the macro scale. This heterogeneity is primarily governed by the coupling of fault structures and diagenetic fluid-driven dissolution-precipitation processes. Integrating these findings, three types of “sweet spot” zones are identified within the fault-controlled reservoirs: boundary fault zones in the overlap segments along the fault strike, fault core zones along the dip, and dolomitic limestone intervals in the vertical sequence. This study fills a gap in understanding the heterogeneity of fault-controlled carbonate fractured-vuggy reservoirs and provides theoretical support for improving the recovery efficiency of ultra-deep oil and gas resources.

Accurate and reliable production forecasting is a critical component for the efficient development of oil and gas fields and supports informed scientific decision-making. Although machine learning methods have achieved significant progress in this domain, existing models are typically trained from scratch using limited historical production data, making it difficult to effectively capture the complex nonlinear dynamics, long-term temporal dependencies, and high-dimensional interactions among variables inherent in production time series. This often leads to insufficient generalization capacity and limited predictive robustness. To address these challenges, this study proposes a novel gas well production forecasting method based on large language models (LLMs). The approach builds upon a pre-trained GPT-2 architecture and incorporates several key adaptations to enable effective time-series prediction. First, the input data—including daily gas production rate, tubing pressure, casing pressure, and production time—are subjected to instance normalization to facilitate knowledge transfer. Second, a trainable embedding layer is designed to map numerical time-series data into the semantic embedding space of the LLM, thereby achieving cross-modal alignment between continuous signals and the discrete representation format required by the model. Third, a parameter-efficient transfer learning strategy combining freezing and fine-tuning is implemented: the core self-attention and feed-forward network layers of the LLM are frozen to preserve general-purpose knowledge acquired during pre-training, while the positional encoding and layer normalization modules are selectively fine-tuned to enhance the model’s ability to characterize temporal patterns specific to production dynamics. The resulting model, termed GPT4TS, is systematically evaluated on real-world production data from a marine carbonate gas reservoir in the Sichuan Basin. Experimental results show that for wells with long production histories, GPT4TS significantly outperforms the conventional LSTM model. Under univariate input, the mean absolute percentage error (MAPE) is reduced by 18.573% on average; under multivariate input, the MAPE reduction reaches 35.610%, demonstrating its superior capability in modeling complex trends and leveraging multi-variable synergies. However, for newly commissioned wells with short production histories, insufficient data hinders effective fine-tuning, leading to lower prediction accuracy compared to LSTM. This study not only validates the potential of large language models in petroleum production forecasting but also highlights their strong dependence on historical data length, providing both theoretical insights and practical guidance for model selection in real-world engineering applications.

The distribution characteristics of carbon isotopes in sedimentary organic matter are often used as indicators for oil-source correlation, determination of organic matter origins, and paleoenvironmental analysis. However, when organic matter reaches the over-mature stage, cracking processes can lead to isotopic fractionation, resulting in anomalies such as carbon isotope reversal. This phenomenon, particularly in deep and ultra-deep petroleum systems, has long been a key and challenging issue in hydrocarbon geology research. This study examines the carbon isotopic distributionsof various solid and liquid organic materials from source rocks and reservoirs in the Anyue Gas Field, the central Sichuan Uplift, Sichuan Basin, based on existing data and relevant research. The findings indicate that paleo-oil reservoirs experienced thermal alteration, causing the early-formed liquid hydrocarbons in both source and reservoir layers to crack into gaseous hydrocarbons, residual liquid hydrocarbons, and pyrobitumen. The pyrobitumen can be categorized into in-situ pyrobitumen in source rocks and reservoir pyrobitumen. In present-day source and reservoir layers, an overall inversion is observed where kerogen has lower δ¹³C values than liquid hydrocarbons, along with a localized inversion in which saturated and aromatic hydrocarbons show higher δ¹³C values than non-hydrocarbon components and asphaltenes. Additionally, the δ¹³C values of reservoir pyrobitumen are lower than those of both kerogen and liquid hydrocarbons. Comprehensive analysis of relevant data and simulation experiments indicates that carbon isotopic enrichment during hydrocarbon cracking is the primary cause of the observed isotopic inversions in both source and reservoir samples. After high-temperature cracking, the residual liquid hydrocarbons derived from original liquid hydrocarbons in source rocks and paleo-reservoirs exhibit an overall increase in δ¹³C values of approximately 4‰. Due to varying thermal exposure in source rocks and paleo-reservoirs, the extent of δ¹³C increase differs among various group components of the residual liquid hydrocarbons. Although reservoir pyrobitumen largely inherits the isotopic signature of the original crude oil, thermochemical sulfate reduction (TSR) during its formation may also contribute to its anomalously light carbon isotopic values. In summary, the carbon isotopic inversions observed in the source and reservoir layers of the study area are primarily attributed to carbon isotopic fractionation during liquid hydrocarbon cracking—a phenomenon that may be common in deep and ultra-deep petroleum reservoirs that have experienced high temperatures.

The tight conglomerate reservoir in the Junggar Basin exhibit complex multi-scale pore-throat structures, which obscure the dominant controls on fluid occurrence, hinder reservoir classification, and complicate the quantitative evaluation of graded reserves. To clarify the relationship between microscopic pore-throat characteristics and fluid occurrence states and achieve effective evaluation of highly heterogeneous tight conglomerate reservoirs, this study takes the Upper Urho Formation (P₃w) in the CO₂ flooding pilot area of the Mahu 1 well block as an example, conducting comprehensive workflow of “pore-throat structure-fluid occurrence-reservoir classification-graded reserves”. Integrating methods including cast thin sections, high-pressure mercury intrusion, and nuclear magnetic resonance (NMR) logging were integrated to characterize the lithomechanical properties and pore-throat structure parameters of different sub-layers within the main reservoir interval. Grey relational analysis further identified movable-to-total porosity (weight 0.913) and clay mineral content (weight 0.805) as the key factors controlling effective oil saturation. Based on NMR T₂ spectral morphology, three-component index, IB value, and permeability, a classification standard dividing the reservoir into three types (Ⅰ~Ⅲ) was established, A methodology for calculating tiered pore-volume reserves was proposed, achieving a systematic evaluation from pore-throat structure to classified reserves. Results indicate that in the main producing interval P₃w₂², the lower sublayer exhibits significantly better reservoir quality than the upper. The lower section has a micron-scale pore proportion of 25.48%, an average permeability of 5.59 mD, and an effective oil saturation 15%~20% higher than the upper section. Nano-scale pores (<100 nm) represent the dominant storage space in the pilot area, containing reserves of 784.3 thousand tons (61.7% of the total). From the perspective of reservoir classification, type Ⅰ reservoirs, with the highest proportion of micron-scale pores (25.48%) and excellent oil-bearing capacity. Type II reservoirs show reduced micron-scale pores and moderate oil-bearing capacity; Type III reservoirs suffer from poor oil retention due to strong confinement by nanopores. The results of this study reveal the controlling factors of fluid distribution, establish a classification standard and tiered reserve characterization system for tight conglomerate reservoirs constrained by pore-throat structures, and provide theoretical support and technical support for the optimal selection of CO₂ flooding target zones and the identification of “sweet spots” in the Mahu conglomerate oilfield.

The goal of “carbon peak and neutrality” is driving China’s energy system to accelerate transition to clean, low-carbon development. As an important new clean energy source, natural gas hydrate (NGH) exhibits high energy density, wide distribution, and substantial resource potential. Therefore, accelerating its industrial production is key to achieving a reduction in pollution and carbon emission. NGH production involves Thermal-Hydrological-Mechanical-Chemical (THMC) multi-physical field coupling. Traditional experiments and production tests fail to fully reveal underlying mechanisms, making numerical simulation—with high functionality, flexible methods, and low cost—an essential research tool. This study systematically reviews theories, technologies, and applications of numerical simulation for natural gas hydrate production to provide theoretical support for safe, efficient extraction and advance the translation of simulation technologies to engineering practice. Specifically, it clarifies evolution laws of seepage parameters (e.g., porosity, permeability) during hydrate dissociation; identifies evolution of mechanical parameters (e.g., shear strength, cohesion) with hydrate saturation, revealing the core mechanism by which hydrates dominate reservoir mechanical property evolution via decomposition behaviour; outlines approaches to constructing THMC multi-physical field coupling models; summarizes functions and advantages of major global simulators (e.g., TOUGH+Hydrate, SuGaR-TCHM), and validates applications at typical pilot sites. Current numerical simulation research has limitations: multi-phase flow models insufficiently account for continuous pore structure evolution and impacts of hydrate saturation on relative permeability; characterization of mechanical properties and sand production risk responses in unconsolidated clayey silt sediments is inadequate; and capacity to predict long-term mechanical stability risks (e.g., land subsidence, submarine landslides) induced by production is limited. Future work should establish “micro-macro” cross-scale parameter models, refine elastoplastic constitutive models for clayey silt sediments, and develop integrated geological engineering simulation tools to advance simulation technologies from mechanistic interpretation to engineering decision support.

China possesses vast and widely distributed oil and gas resources, characterized by significant development potential and rapid growth in their exploitation. However, the exploration and development of unconventional resources such as shale gas and tight oil remain at an early stage. As domestic demand for oil and gas continues to rise, the extraction of deep unconventional resources has consequently become a major focus in national energy engineering. A common practice in unconventional oil and gas production involves large-scale fluid injection into deep geological formations. Such injection disturbs the in-situ stress field and alters the stress state of subsurface faults, which may lead to fault instability, slip, and subsequently induced seismicity. Therefore, evaluating and mitigating such anthropogenic seismic risks has become a critically important issue for achieving safe and sustainable resource development. In recent years, seismic events have been monitored during the production phases of multiple deep energy projects worldwide. Post-earthquake analyses indicate a clear spatiotemporal correlation between these events and fault activation or instability triggered directly by fluid injection. Notable examples include geothermal projects in Pohang, South Korea, and Basel, Switzerland, which were suspended due to significant induced earthquakes. In fact, geological conditions in China’s oil and gas fields are often more complex, making it difficult to accurately predict fault slip conditions, specific slip patterns, and displacement, or to determine optimal operational parameters near fault-developed zones. To balance benefits and risks in unconventional resource development, predicting and preventing the associated environmental geological issues and seismic hazards caused by deep energy extraction has become an urgent challenge. This study reviews typical cases of induced seismicity in energy projects globally, with a focus on analyzing the underlying stress conditions and the detailed triggering mechanisms of fault instability and slip at development sites. It systematically summarizes fault slip instability modes, established criteria for induced seismicity, estimated affected ranges, and current magnitude prediction models. The influence of both anthropogenic engineering factors and inherent environmental geological conditions on induced seismic events during development is also thoroughly discussed. Finally, this study summarizes the major unresolved issues and key technical challenges that need to be addressed in current research, and outlines potential future research directions through three primary approaches: numerical simulation, experimental methods, and enhanced field monitoring. The work contributes to a deeper fundamental understanding of human induced seismicity from oil and gas operations, and holds practical implications and considerable engineering value for mitigating or preventing potential seismic hazards.

To systematically review the research progress of intelligent fracturing technology in unconventional oil and gas reservoirs, by integrating machine learning algorithms (e.g., random forest and gradient boosting), the embedded discrete fracture model (EDFM), fiber-optic/microseismic monitoring technologies, and smart equipment, this study analyzes technological breakthroughs and application cases in key aspects such as reservoir parameter prediction, fracture propagation simulation, and real-time control. Results demonstrate that the random forest and gradient boosting models demonstrated optimal performance in permeability prediction (R²>0.92). The EDFM-AI workflow reduces fracture parameter calibration errors to 6.8%. Fiber-optic monitoring technology achieves sub-millimeter resolution in fracture detection. The intelligent early-warning system predicts sand plugging risks 30 seconds in advance (accuracy above 85%). Intelligent fracturing technology significantly enhances reservoir modification efficiency and production, but challenges such as small-sample generalization, multi-source data fusion, and equipment autonomy require further resolution. Establishing a closed-loop technical system encompassing “reservoir evaluation, optimized design, fracture monitoring, anomaly prediction, and equipment control”, and promote the development of intelligent and precise fracturing processes.

The enrichment patterns of natural gas in the tight sandstone reservoirs of the Xujiahe Formation in the northeastern Sichuan Basin are complex, and the fault system exerts a crucial control on hydrocarbon accumulation. To clarify the structural characteristics of the faults and their impact on natural gas enrichment in this area. Based on three-dimensional seismic data and drilling data, integrated with artificial intelligence fault identification technology and structural analysis, this study identifies a NE-SW trending en echelon strike-slip fault zone in the Xujiahe Formation of the Wubaochang Area, northeastern Sichuan Basin. The structural deformation characteristics, evolutionary history, and reservoir-controlling and accumulation-controlling effects of this fault zone are investigated. The results indicate that the fault zone exhibits vertically layered deformation characteristics: it manifests as thrust faults in the strata from the Upper Ordovician to the Lower Triassic Jialingjiang Formation, while in the Upper Triassic Xujiahe Formation, it appears as a strike-slip fault zone composed of a series of small NNW-SSE trending thrust faults arranged in a right-stepping en echelon pattern. The fault zone underwent two stages of tectonic evolution. During the Indosinian-Yanshanian period, under continuous SE-NW oriented tectonic compression, it exhibited thrust fault activity. In the Himalayan period, the regional tectonic stress field shifted, and under a NE-SW oriented horizontal compressional stress field, the pre-existing major faults underwent sinistral strike-slip movement, forming a series of right-stepping en echelon strike-slip faults in the Xujiahe Formation. The faults in the Xujiahe Formation are interconnected with the deep-seated major faults, linking the Xujiahe Formation reservoir, the Longtan Formation source rocks, and deep fluids. This connection establishes a “strong below and weak above” transport system and a “dual-source hydrocarbon supply” model. Simultaneously, it exerts a dual effect on the tight sandstone reservoirs of the Xujiahe Formation, namely “dissolution and porosity enhancement” and “cementation and destruction,” providing favorable dynamic conditions for the differential evolution of tight reservoirs and the development of sweet spot reservoirs. Consequently, the fault zone serves as a favorable site for natural gas enrichment and preservation in the tight sandstone reservoirs of the Xujiahe Formation. The research demonstrate that strike-slip fault zones are advantageous enrichment belts for natural gas in Triassic tight sandstone reservoirs in the northeastern Sichuan Basin, and greater emphasis should be placed on the evaluation and exploration of such target areas.

With the increasing demand for real-time monitoring and efficient management in offshore oilfield development, Distributed Acoustic Sensing (DAS) has emerged as a powerful tool for downhole fluid monitoring and production interpretation. Despite its growing application, the understanding of DAS acoustic response characteristics and underlying mechanisms during well production remains insufficient. Moreover, conventional sound velocity-flow interpretation methods typically require relatively high flow rates, which limits their applicability in domestic wells with lower production rates. Addressing these challenges, this study presents a systematic investigation of the DAS response mechanisms associated with downhole oil-water two-phase flow. A theoretical analysis was first conducted to elucidate the acoustic response mechanisms of multiphase flow within the wellbore. Building on this analysis, a fully coupled physical simulation apparatus was designed, comprising a reservoir analog, a full-scale wellbore, and a test tubular string, replicating realistic downhole production conditions. Experiments were performed to monitor the DAS response during the flow of oil-water mixtures from simulated reservoir cores into the wellbore, considering varying flow rates and water cut conditions. DAS data were processed with advanced denoising techniques, and Frequency Band Energy (FBE) metrics were computed to quantitatively characterize the acoustic response. The experimental results reveal that the viscosity of oil-water mixtures is a key factor influencing DAS responses. Variations in the oil-water ratio generate complex multiscale interfacial dynamic processes and flow regime instabilities, which are identified as the primary causes of changes in the spectral characteristics. The frequency spectrum of two-phase flow within the wellbore was found to be concentrated below 500 Hz. Specifically, when one phase predominates, the spectrum typically exhibits a single dominant peak; whereas for near-equal oil-water ratios, multiple peaks emerge in the frequency spectrum, reflecting the complex dynamics of the multiphase flow. To enable quantitative interpretation, the Reynolds number (Re) was selected as a representative characteristic parameter integrating both fluid properties and pipe geometry. An FBE-Re relationship model was subsequently established, showing a highly correlated logarithmic relationship. This model provides a robust framework to quantitatively link DAS responses to downhole two-phase flow characteristics. Field validation demonstrates that the observed DAS frequency bands closely match the laboratory-determined ranges, and the interpreted total liquid production and water cut achieve an accuracy exceeding 85%. These results confirm the reliability and practical applicability of the proposed FBE-Re model for field production monitoring. Overall, this study significantly advances the understanding of DAS acoustic response characteristics and governing mechanisms in oil-water two-phase downhole flow. The findings provide a theoretical foundation and practical guidance for accurate well production profiling, high water-cut zone identification, and enhanced field-level production monitoring, offering substantial potential for improving offshore oilfield management and operational decision-making.

Inter-well connectivity has become one of the important criteria for guiding the development of water drive reservoirs. Traditional methods for predicting inter-well connectivity, such as tracer analysis, well testing analysis, numerical simulation, etc, have problems such as computational difficulties, cumbersome processes, and high costs. However, deep learning based methods suffer from issues such as data sensitivity and poor adaptability. To address the aforementioned challenges and problems, this paper proposes a prediction method for predicting inter well connectivity by fusing numerical models with graph neural networks. On the one hand, this method fully considers the physical parameters related to the injection and production well network in the production process, which derives a numerical model of inter-well connectivity considering multiple factors, and solves the problem of single form in previous numerical models. On the other hand, the similarity between well network structure and graph structure was reasonably utilized, a graph neural network model based on long and short-term memory neural network was designed. Then, a fusion method with deep learning network model was proposed, which was fused with numerical model. This solves the problem of traditional artificial intelligence methods ignoring physical parameters, and introduces self attention mechanism to optimize the model under the long and short-term memory neural network framework. The mechanism model and actual reservoir production data were used to predict inter-well connectivity and fluid production on the established model, and new development plans were formulated based on the predicted results. The model’s performance was verified through several sets of experiments. The results show that when the model established in this paper is used to predict inter-well connectivity, and the liquid production is calculated based on the connectivity prediction results, the prediction accuracy is as high as 98%, indicating that the model’s prediction results are reliable. Using the fusion model to predict inter-well connectivity of different sub-layers in the reservoir, it is found that the model’s prediction accuracy reaches over 95% in well patterns of different scales, showing strong applicability. Finally, the production plan was adjusted based on the connectivity prediction results: liquid reduction was implemented for both wells with high connectivity and those with low connectivity. Comparison shows that the predicted recovery degree after 10 years of the adjusted development plan is 6.8% higher than that of the original plan. This method balances physical interpretability and computational efficiency, providing reliable technical support for judging the development effect of water-flooded reservoirs and designing secondary development plans.

Against the backdrop of the full advancement of China’s “Dual Carbon” strategy and the impending inclusion of the oil and gas industry in the national carbon emissions trading system (ETS), traditional optimization models for reservoir development, which mostly adopt a static carbon emission accounting mode for the net present value (NPV) maximization problem, cannot accurately capture the real-time impact of production strategy adjustments on carbon emissions, and thus struggle to address the dual challenges of economic benefits and emission reduction brought by carbon costs. This study develops a carbon-embedded NPV optimization model that establishes a coupling system between subsurface reservoir simulation and surface facility energy consumption simulation. An improved particle swarm optimization (PSO) algorithm with a constraint repair operator is adopted to solve the global optimal control strategy of the model with production well liquid rates as decision variables. Three carbon quota allocation mechanisms are designed for quantitative analysis: free allocation, mixed allocation, and paid auction. Numerical experiments demonstrate that, compared with the baseline scenario, the NPV improvements under the three mechanisms are 7.63%, 5.56%, and 3.86%, respectively, indicating that stricter carbon constraints lead to greater compression of profit margins. Furthermore, the paid auction mechanism exhibits the strongest water-cut control and the most significant emission reduction effects. The study verifies that the carbon-embedded NPV optimization model can achieve the synergistic optimization of production and emission reduction, and effectively transmit carbon price signals to development enterprises. Carbon quota mechanisms are ideal transitional policy tools for the low-carbon transformation of the oil and gas industry.

Rate of penetration (ROP) prediction is of great significance to drilling engineering and can provide important reference basis for drilling optimization, resource allocation, safety guarantee. In recent years, artificial intelligence has sparked a new round of intelligent transformation, promoting the intelligent transformation and upgrading of drilling engineering and giving rise to a large number of new ROP prediction methods. However, there is still a lack of systematic summary and analysis of these new ROP prediction methods at present. This paper, through systematic research and refinement of the main ROP prediction models and methods at home and abroad, summarizes and expounds the development background and theoretical principles of three types of ROP prediction methods: explicit ROP equation, numerical simulation, and artificial intelligence model. It also deeply analyzes the key technical problems and challenges faced by various ROP prediction methods in practical applications at present. It is also pointed out that the ROP prediction method integrating mechanism and data is an important direction to break through the existing bottlenecks and also the mainstream trend of future technological development. Based on this, combined with the development trend of intelligent drilling and the main bottlenecks of the ROP prediction model, five future development directions are proposed: ① Automated and intelligent drilling equipment; ② A dedicated mechanism and data fusion model for ROP prediction; ③ Environmental response mechanism of ROP prediction model based on embodied intelligence, swarm intelligence, reinforcement learning and online learning; ④ A general ROP intelligent prediction model based on large models and transfer learning algorithms; ⑤ Closed-loop optimization of ROP prediction model based on scientific knowledge discovery.

Formation drillability assessment is crucial for drilling operations, as it directly influences operational efficiency and cost-effectiveness. Traditional three-dimensional (3D) assessment methods often face challenges due to the unstable integration of multi-source and cross-scale data, resulting in limited spatial generalization and suboptimal prediction performance. To address these limitations, this paper proposes a multi-source data fusion method based on a gated recurrent unit (GRU) network to enhance intelligent formation drillability assessment and improve drilling efficiency in a study area in eastern China. The method consists of two phases: well data training and 3D application. In the first phase, pseudo-depth domain seismic records synthesized from seismic average wavelets and well logging data serve as the foundation. Sensitive attributes related to formation drillability are further extracted as network inputs. These sensitive attributes include a velocity model incorporating geological information and a seismic frequency-fraction attribute that captures multi-scale stratigraphic structure. A corrected drillability index (D_c) is used as a label for model training, ensuring that the network learns to establish an accurate mapping relationship between input attributes and drillability indicators. This training method leverages the temporal and sequential learning capabilities of the GRU network to effectively model complex relationships in the data. In the second phase, the pretrained network was extended to 3D applications, constructing a 3D input dataset by extracting the corresponding attributes. This dataset was then fed into a pretrained GRU model to predict formation drillability in the study area. Analysis of five representative wells in the study area validated the effectiveness of D_c in characterizing rock drillability in the study area. Furthermore, experiments using the Marmousi numerical model demonstrated that the method outperformed traditional intelligent prediction methods, such as those relying solely on raw seismic data or a combination of raw seismic and well logging data. Practical application in the study area further confirmed the method’s ability to effectively capture variations in formation drillability. By providing reliable predictions, the method becomes a powerful tool for optimizing drilling operations and enhancing drilling engineering decision-making.

The extraction of shale oil and gas is confronted with numerous complex geomechanical issues. As a core factor determining extraction efficiency and safety, the mechanical properties of shale urgently require in-depth research and exploration. Against this backdrop, machine learning, with its powerful capabilities in data processing and pattern recognition, has opened up new avenues for research on shale mechanical properties. This paper focuses on the application of machine learning in the study of shale mechanical properties, systematically elaborating on the current status, challenges, and prospects in this field. Firstly, it details the application achievements of current machine learning algorithms in the prediction of shale mechanical parameters and the recognition of failure modes, demonstrating their significant advantages over traditional research methods in processing complex data and mining potential patterns. Subsequently, it summarizes the multiple challenges faced in the application of machine learning to the study of shale mechanical properties. Shale sample data exhibits high-dimensional and small-sample characteristics, which easily lead to overfitting in models. Meanwhile, the internal operating mechanisms of most machine learning models are difficult to interpret, restricting their popularization and application. In addition, the geological conditions of shale are complex and variable, with significant differences in mineral composition and pore structure of shale in different regions. The existing models show obviously insufficient universality when applied across regions and geological conditions. Finally, the future is prospected based on the development trends of cutting-edge technologies. Machine learning has broad prospects in the field of shale mechanical properties research. By integrating multi-source data such as geological, geophysical, and logging data, it can provide more abundant information for models and reduce the negative impact caused by data dimensionality. Optimizing algorithm architectures and combining technologies such as transfer learning and ensemble learning can improve the generalization ability of models. Constructing physics-constrained machine learning models can not only enhance the interpretability of models but also improve their adaptability under complex geological conditions. These strategies are expected to break through existing bottlenecks, promote the in-depth application of machine learning in the study of shale mechanical properties, and provide solid theoretical and technical support for the efficient development of shale oil and gas resources.

Faults serve as crucial pathways and sites for hydrocarbon migration and accumulation, making their identification a key task in the interpretation of seismic data. However, the diversity of fault types, extensive distribution, and complex characteristics pose significant challenges to fault identification. To address this issue, this paper proposes a fault identification method using a 3D TransUnet model. Constructed based on 3D CNN and transformer modules, this model adopts an end-to-end structural design of the 3D Unet architecture. By learning the spatial relationships of three-dimensional faults in synthetic seismic data, it directly predicts fault information in actual seismic data. The method has been successfully applied to seismic work areas in the F3 block of the Dutch North Sea and the Halahatang area of the Tarim Basin, achieving excellent results. The research findings demonstrate that the 3D TransUnet model combines the high local accuracy of CNN and the global attention mechanism of Transformer, enabling inference and prediction of faults in complex regions based on global fault information. Compared with the 3D Unet model and other traditional fault identification methods, the 3D TransUnet model achieved a recall rate of 0.87 and a precision rate of 0.83 on the validation set, significantly outperforming other approaches. In practical applications within three-dimensional seismic work areas, the 3D TransUnet model accurately identifies fault information across different regions. For faults with subtle features, the incorporation of the Transformer module equips the model with a global attention mechanism, allowing it to infer the presence of faults by analyzing distribution trends across the entire work area. By applying the trained fault identification model to different practical seismic work areas (the F3 block and the Halahatang area in this study), the universality of the method is demonstrated, indicating that the trained fault identification model can be effectively utilized across seismic data from various regions. This study finds that the method can effectively identify microfracture information within formations. In oil and gas fields where microfractures serve as reservoirs, since microfractures primarily develop along major faults, well locations are typically deployed near these large faults. However, during the middle and late stages of oil production in such fields, well placement decisions rely more heavily on the development degree of microfractures. Therefore, this fault identification method provides valuable guidance for well placement in oil and gas fields where microfractures act as reservoirs.

Under the “dual-carbon” target, geothermal energy has become an important direction for energy transition due to its characteristics of being clean and efficient, having abundant reserves, and being low-carbon and environmentally friendly. Aiming at the seepage-heat transfer evolution characteristics and well-pattern optimization of geothermal fields under multi-well production and reinjection conditions, this study takes the geothermal field in the Xiong’an high-speed railway area as the research object. A three-dimensional coupled seepage-heat transfer numerical model was constructed based on COMSOL Multiphysics, and parameter calibration was carried out by combining field monitoring data of production-well water temperature and groundwater level to ensure consistency between numerical results and actual operation conditions. On this basis, twenty additional geothermal wells were designed, including ten production wells and ten reinjection wells. By simulating three well-pattern schemes—linear, staggered, and triangular—the variations in production-well water level, temperature, and dynamic recoverable geothermal resources under different schemes were analyzed, and the influences of different schemes on reservoir stability and heat-exchange efficiency were systematically compared. The results show that the linear well pattern has a simple structure and is easy to implement; however, its thermal mobilization range is limited, the overall recoverable geothermal resources are the lowest, and boundary production wells are prone to forming localized groundwater drawdown cones. The staggered well pattern exhibits a more uniform distribution of production and reinjection, which can effectively suppress localized overexploitation, delay cold-water breakthrough, and maintain relatively high reservoir thermal stability, resulting in the most stable evolution of dynamic recoverable geothermal energy. The triangular well pattern shows the strongest thermal mobilization capacity and the highest recoverable geothermal resources during the early stage of operation, but the risk of cold-water intrusion increases in areas with high production-well density, leading to a tendency of localized thermal attenuation in the later stage of operation. On a 50-year timescale, the total recoverable geothermal resources of both the staggered and triangular well patterns are more than 1% higher than those of the linear well pattern. Among them, the staggered well pattern achieves the best balance between resource utilization efficiency and system stability and is recommended as the preferred well-pattern scheme for the efficient and sustainable development of the geothermal field in the study area. The research results can provide theoretical references for well-pattern optimization, development performance evaluation, and sustainable operation of multi-well production and reinjection systems in mid- to deep-depth hydrothermal geothermal fields under similar geological settings and operating conditions.

Fractured-vuggy carbonate reservoirs represent a major domain in global hydrocarbon exploration and development. However, their efficient development is constrained by strong heterogeneity, characterized by spatially discrete distribution of fractures and vugs, complex connectivity, and diverse filling types. This paper systematically reviews research advances in reservoir characteristics, experimental methods, remaining oil distribution, and water/gas injection mechanisms. Based on genesis and controlling factors, the reservoirs are classified into three types: (1) Epikarst: dominated by fractures and dissolution pores (depth < 70 m), with a filling ratio of up to 60%; (2) Subterranean river karst: characterized by large-scale dissolution caves (extending 0.8 ~ 3.5 km), which are further divided into four types based on cave height: hall caves, main channel caves, tributary caves, and terminal caves; (3) Fault-controlled karst: forming vertically connected “fracture-cavity complexes” along fault zones, with full filling ratios ranging from 20% to 38%. Physical experimental models have evolved from conceptual ones (e.g., spliced marble/acrylic plates) toward realistic structures: marble models reveal fundamental flow mechanisms; transparent acrylic models enable visualization of displacement processes; laser-etched models replicate complex fracture-vug networks; full-diameter leached core samples support high-temperature and high-pressure simulations; and 3D printing allows precise control of fracture-vug morphology. However, limitations remain in simulating diverse filling media. After water flooding, remaining oil occurs in various forms, including attic oil, blind-end oil, bottom-water coning oil, oil shielded by high-conductivity channels, and filling-dependent oil. Field analyses, numerical simulations, and physical experiments consistently indicate that attic oil accounts for the largest proportion, making it a key target for enhanced oil recovery. Development strategies have progressed from single-well huff-and-puff to multi-well water/gas flooding, and further to optimized synergistic gas-water injection. Specifically, single-well water flooding mainly reduces water-coning residual oil; gas huff-and-puff mobilizes attic oil via gravity segregation; bottom-water injection suppresses coning and expands sweep efficiency; nitrogen flooding effectively recovers attic oil but requires well pattern optimization to control gas channeling; cyclic water injection enhances water diffusion through pressure fluctuations; flow-reversal injection alters flow paths to enlarge sweep volume; and synergistic gas-water injection couples gravity segregation with mobility control to improve displacement efficiency. Based on these insights, this paper proposes an integrated framework for potential tapping, termed “reservoir-specific adaptation, remaining oil targeting, stage-wise optimization”. This framework entails precisely matching injection-production techniques to reservoir types, implementing targeted controls based on remaining oil patterns, and dynamically optimizing strategies according to development phases. Future efforts should address challenges in three-dimensional modeling of complex fracture-vug systems, suppression of gas channeling, and dynamic prediction of remaining oil. Integrating artificial intelligence and high-precision detection technologies will enable accurate characterization of fracture-vug architectures. Developing novel experimental materials and model fabrication techniques will enhance simulation authenticity. Constructing numerical models that account for multi-physics coupling and optimizing synergistic gas-water flooding strategies are crucial for advancing the development of fractured-vuggy reservoirs from an “experience-driven” to a “precision-controlled” paradigm.

This study addresses the challenges of poor real-time performance and low accuracy in drilling condition identification by introducing an innovative intelligent recognition method. The proposed approach integrates a one-dimensional convolutional neural network (1dCNN) for local feature extraction, a bidirectional gated recurrent unit (BiGRU) to capture sequential dependencies, and a multi-head attention mechanism to emphasize critical information. This fusion enables efficient discrimination among 13 drilling conditions, including rotary drilling, slide drilling, whipstocking, and reverse whipstocking. In the model design phase, comprehensive ablation studies were conducted to evaluate the contributions of each module—1dCNN, BiGRU, self-attention, and multi-head attention—as well as their serial and parallel configurations. The performance was further optimized using the Optuna framework for automatic hyperparameter tuning. Experimental results demonstrated that the model achieved an accuracy of 96.22% on time-domain data from a single well. Additionally, in both intra- and inter-block transfer tests, the overall accuracy ranged from 94% to 97%, with each drilling condition exceeding an 80% recognition rate. Real-time testing on field data also showed a high degree of consistency with actual operational conditions. Overall, the proposed method provides a robust technical framework for real-time monitoring and optimization of drilling operations.

To investigate the triaxial compression creep deformation behavior and failure mechanisms of shale, this study conducted stepwise incremental loading creep tests on Wufeng Formation shale of Changning under confining pressures of 10 MPa, 30 MPa, and 50 MPa. The characteristics of axial, radial, and volumetric creep curves were analyzed. Combining computed tomography (CT) and scanning electron microscopy (SEM) techniques, the macroscopic and microscopic creep failure mechanisms of shale were systematically examined. A damage model was employed to fit and predict the long-term creep behavior of shale. The results indicate that: Volumetric creep effectively characterizes the evolutionary features of shale creep, with creep crack initiation stress and damage stress threshold corresponding to approximately 60% and 80% of peak strength, respectively. Shale creep failure is predominantly governed by microcrack initiation and coalescence, forming spatially distributed shear fracture networks characterized by shear-compaction bands and branched cracks. Multiple subparallel fractures develop within the shear-compaction bands, while extensive branching cracks readily form in surrounding regions. The damage model accurately predicts the long-term creep behavior of shale. Parameter A, associated with loading stress and elastic modulus, serves as the primary controlling factor for creep. This research can be applied to the analysis of the influence of shale creep on wellbore stability and artificial fracture closure during the drilling and fracturing processes of shale oil and gas reservoirs.

In the Haiba block of the Zhaotong demonstration area, shallow shale gas reservoirs are characterized by shallow burial depth, low formation pressure, and large variation in formation dip angle. The flowback process during well testing in these reservoirs is marked by a long gas breakthrough time, a high gas breakthrough flowback rate, and low stabilized gas testing rates. Improper flowback can lead to wellbore proppant settling, liquid loading, and loss of fracture volume, requiring an optimized flowback strategy to ensure safety and efficiency. This study presents a systematic analysis of flowback data from 71 wells in this block, identifying three distinct flowback stages: pure liquid flowback, gas breakthrough, and stabilized gas flowback. Using a fracture compression model and rate-transient analysis, the variations in effective fracture volume and fracture half-length under different flowback strategies were examined. The results show that a fast flowback strategy can cause an irreversible loss of up to 30% of the effective fracture volume, whereas a slow flowback strategy can reduce this loss to about 12%. To minimize this irreversible fracture loss, key flowback issues and controlling factors for these shallow shale gas wells were addressed. The well-opening pressure was determined by analyzing the effective stress within fractures. A wellbore proppant settling model and a fracture proppant flowback model were developed, yielding the critical flow rate in the wellbore required to carry proppant and the critical flow rate in fractures that triggers proppant flowback. Based on multiphase flow theory, the minimum bottomhole pressure for fluid unloading and the critical gas rate for liquid lifting were also determined. Accordingly, a comprehensive optimization strategy was proposed, encompassing appropriate well opening timing, wellbore sand cleanout, proppant flowback prevention, and assurance of fluid drainage, and the flowback optimization diagram was developed. Field application results indicate that this diagram enables a synergistic optimization of fracturing effectiveness and production efficiency: under similar reservoir conditions and stimulation measures, the estimated ultimate recovery increased by 20%. This study marks a transition of flowback management from an empirical approach to a quantitative one, and the proposed critical criteria and optimization diagram have broad applicability, providing a technical reference for flowback optimization in other unconventional oil and gas wells.

As a new strategic replacement for shale gas development in China, the Cambrian Qiongzhusi Formation shale gas reservoir in the Southern Sichuan Basin is characterized by great burial depth, high in-situ stress, well-developed natural fractures, and notable inter-well interference. These characteristics pose severe challenges to the evaluation and optimization of hydraulic fracturing performance in new production platforms. Meanwhile, conventional hydraulic-fracturing monitoring methods are limited in spatial continuity, inter-well-scale characterization, and quantitative evaluation of stage-to-stage stimulation differences, making it difficult to continuously monitor complex fracture behaviors such as fracture propagation, fracture-offset-well interaction, and communication with natural weak planes in ultra-deep shale reservoirs. Address these challenges, this study conducted cased-hole offset-well fiber-optic strain monitoring using low-frequency Distributed Acoustic Sensing (DAS) during zipper fracturing operations on a production pad in this area. Strain-rate data were acquired throughout the complete treatment process of 52 fracturing stages from two treatment wells, and the far-field strain-response characteristics associated with hydraulic-fracture propagation were systematically analyzed. A field-data interpretation workflow was established, including raw DAS data quality control, low-frequency information extraction, synchronized analysis of strain-rate waterfall plots and treatment parameters, fracture-arrival-time statistics, shut-in response analysis, and cumulative-strain-based evaluation of reservoir stimulation effectiveness. The results show that the monitoring responses are generally characterized by widespread response, clear depth correspondence, complex response patterns, and significant differences between earlier and later treatment stages. Hydraulic fracture propagation mainly exhibits three dominant modes: competitive propagation of multiple fractures, coexistence of vertical and inclined fractures, and large-scale communication with bedding-parallel fractures. The fracture propagation velocity toward the offset well and the strain response characteristics at shut-in are closely correlated with fracture propagation behavior. The average fracture propagation velocity is jointly controlled by local in-situ stress and natural fracture conditions, the vertical offset between the treatment well and the monitoring well, the degree of communication with previously created fractures, and local stress-field modification induced by zipper fracturing.The strain response during shut-in reveals differences in closure behavior among different fracture types. Fracturing stages dominated by opening-mode vertical fractures commonly exhibit obvious strain reversal after shut-in, indicating relatively sufficient fracture closure. In contrast, intervals affected by bedding-plane communication or shear slip along weak planes generally show slowly decaying tensile strain that is difficult to fully reverse, suggesting that residual fracture opening or shear displacement may remain on the fracture surfaces. This study pioneers the use of hydraulic-fracturing far-field strain monitoring results for quantitative evaluation of reservoir stimulation effectiveness, enabling quantitative assessment of stimulation intensity and the spatial extent of stimulation for different fracturing stages. Furthermore, it documents the monitoring of seismic events during operations and discusses their impact. The study provides a data-driven framework for optimizing fracturing design, mitigating well interference risks, and enabling collaborative development in deep shale reservoirs, while establishing a foundation for future fracture parameter inversion and quantitative evaluation.

To address the challenges of poor sorting, strong heterogeneity, and unclear overburden pressure-dependent variations in porosity and permeability in glutenite reservoirs, where existing conventional sandstone correction methods are inadequate, this study investigates the Paleogene Kongdian Formation glutenite reservoirs (burial depth: 3800~4100 m) in the BZ area. Using overburden pressure porosity-permeability experiments and rock mechanics test data from 13 core plugs covering the main porosity-permeability range, we reveal the unique variation patterns and establish a targeted correction model. The results show that this is a low-porosity, ultra-low-permeability reservoir, with matrix porosity of 5%~15% and permeability of 0.1~6 mD under ambient conditions. The pore space is dominated by intergranular and intragranular dissolution pores, with occasional microfractures, indicating significant heterogeneity. Unlike conventional sandstone reservoirs, the porosity and permeability of these glutenite reservoirs decrease following a power-law function with increasing overburden pressure. The decline gradient varies with the initial porosity and permeability at ambient pressure. Thus, the overburden pressure-dependent porosity/permeability is a bivariate function of the overburden pressure and its initial value at ambient conditions. By fitting the experimental data, we derived the quantitative relationship between the power-law coefficients and the ambient-pressure porosity/permeability, constructing a triaxial overburden pressure correction model. Incorporating rock mechanics data, a uniaxial overburden pressure bivariate correction model for in-situ conditions was also established. Accuracy verification shows that the triaxial model achieves an MSE of 0.369% for porosity and 0.281 mD for permeability. Compared to the traditional univariate model, accuracy improves by 67% and 43%, respectively, providing reliable characterization across both high and low porosity-permeability ranges. The proposed bivariate correction method fills a gap in overburden pressure porosity-permeability correction for glutenite reservoirs and solves the problem that traditional methods cannot adequately account for their strong heterogeneity. This method provides an essential correction tool for accurate reserve evaluation and productivity prediction, and offers a theoretical basis for designing efficient development strategies.

Accurate prediction of collapse and fracture pressures is crucial for well trajectory design, wellbore stability control, and efficient drilling operations. Traditional numerical and analytical methods are often computationally complex and inefficient, while purely data-driven models, although faster, suffer from pronounced black-box characteristics and lack interpretability, which limits their engineering applicability. To overcome these challenges, this study proposes a hybrid-driven prediction method that integrates wellbore stability mechanistic models with a multi-task learning framework (MW-MMoE). In this approach, stress coordinate transformation is embedded as physical prior knowledge at the input stage, while the output targets are reconstructed by first predicting key stress components and then converting them into equivalent densities of collapse and fracture pressures through physical formulations. The Mohr-Coulomb criterion is further incorporated into the loss function as a physical constraint. The model architecture leverages a multi-gated mixture-of-experts network combined with the GradNorm algorithm to dynamically adjust task weights and balance gradients during training. Ablation experiments demonstrate that the proposed MW-MMoE achieves mean absolute errors as low as 0.0019 g/cm³ and 0.0033 g/cm³ for collapse and fracture pressure equivalent densities, respectively, significantly outperforming both single-task and conventional multi-task models, while achieving over a hundredfold improvement in computational efficiency compared with analytical methods. Case studies further validate its engineering applicability: the model can rapidly generate collapse and fracture pressure equivalent density curves for individual wells, produce high-resolution contour maps under arbitrary well inclinations, azimuths, and stress conditions, and perform large-scale three-dimensional predictions across the entire study area. These results highlight that the MW-MMoE model combines high accuracy, efficiency, and interpretability, providing a novel and practical solution for intelligent wellbore stability prediction with broad application prospects.

Under the “Double Carbon” strategy, the high-energy-consuming and high-emission petroleum industry faces urgent demands and pressures for green transformation. However, traditional evaluation models focusing on financial performance are ill-suited for current high-quality development requirements, and existing general ESG evaluation systems lack targeted characterization of petroleum industry traits. To address this issue, based on a macro-micro dual perspective, this paper constructs a comprehensive evaluation index system covering economic, environmental, social, and governance dimensions, specifically incorporating industry-characteristic indicators such as carbon emission intensity and low-carbon technology investment. Using the AHP-Entropy Weight-TOPSIS combined evaluation model, the sustainable development capability of 32 A-share listed petroleum enterprises from 2018 to 2023 is evaluated. The results indicate that: Macroscopically, constrained by cyclical fluctuations in international oil prices and initial green transformation costs, the industry’s overall sustainable development performance declined in 2019 and subsequently exhibited a fluctuating adjustment trend. Microscopically, while the average comprehensive performance of state-owned enterprises (SOEs) is generally superior to that of private enterprises, the gap demonstrates a significant convergence trend, with the score difference narrowing from 0.113 in 2018 to 0.002 in 2023. Further analysis reveals that this disparity primarily stems from SOEs’ institutional advantages in environmental compliance and social responsibility, whereas private enterprises have achieved rapid catch-up by improving governance efficiency and increasing social responsibility investment. This study reveals the differentiated evolutionary paths of petroleum enterprises under different ownership structures in the “Double Carbon” context, providing a quantitative basis for the government to improve industry regulatory policies, and offering scientific decision-making references for SOEs to consolidate their green benchmark status and for private enterprises to optimize ESG governance strategies.

Deep drilling operations involve complex and variable working conditions, where surface-based identification methods suffer from signal propagation delays and high misclassification rates, while limited downhole computing resources preclude the direct deployment of complex algorithms. This paper proposes a Data-driven Modified Analytic Hierarchy Process (DM-AHP) that replaces expert scoring with SHapley Additive exPlanation (SHAP) values for constructing AHP judgment matrices. The mathematical properties of SHAP values—efficiency, symmetry, linearity, and the null player property—theoretically guarantee the objectivity and multiplicative consistency of weight allocation. A “surface training-downhole identification” distributed architecture is established: model training, feature interpretation, and judgment matrix generation are performed on the surface, while only a lightweight 160-parameter matrix is deployed downhole, coupled with a boxplot-based dynamic threshold mechanism for real-time condition identification and adaptive updating. Validation on field data from 11 wells demonstrates that the model achieves 95.5% accuracy and an F1-score of 0.954 in classifying seven normal drilling conditions, with an anomaly detection F1-score of 0.931. Memory consumption is 48 KB and single inference latency is 42 ms, satisfying downhole real-time constraints. Cross-well generalization tests show that the dual dynamic update strategy—combining normalization parameter updates with adaptive threshold adjustment—improves the F1-score from 0.864 to 0.954. Compared with conventional machine learning models with parameter counts on the order of 10⁴~10⁶, the proposed method reduces parameters by over 99% with only a 1.7% decrease in recognition performance. The method can be directly embedded into measurement-while-drilling systems, offering a practically deployable solution that simultaneously achieves interpretability, lightweight deployment, and high accuracy for intelligent condition monitoring in deep drilling operations.

In-situ stress is a fundamental parameter for seismic research, deep resource development, and underground engineering. However, in soft rock formations with strong rheological behavior, the hydraulic fracturing(HF) method cannot directly determine the maximum horizontal principal stress(σ_H), and the anelastic strain recovery(ASR) method often suffers from large uncertainties due to the difficulty of accurately determining the anelastic strain recovery compliance. To address these limitations, this study proposes a novel approach for in-situ 3D stress in soft rock formations that integrates the hydraulic fracturing method with the anelastic strain recovery method. The minimum horizontal principal stress(σ_h) is obtained through hydraulic fracturing tests, while the anelastic strain recovery compliance is back-calculated from ASR data on recovered core samples, enabling the complete in-situ 3D stress tensor to be reconstructed. This combined method has been successfully applied in the Ningjin salt cavern gas storage project, confirming its reliability and applicability in deep soft rock formations. The proposed approach provides a new technical solution for acquiring in-situ stress in soft rock formations and offers valuable support for salt cavern gas storage construction, unconventional hydrocarbon development, and wellbore stability evaluation.

Distributed fiber-optic monitoring technology has become an important tool for hydraulic fracturing fracture diagnosis and fracturing performance evaluation. This technology can provide continuous and high-spatial-resolution sensing responses along the wellbore, and therefore provides an effective method for identifying fracture initiation, fracture propagation, fracture communication, fluid distribution, and production contribution. However, due to the diversity of technical schemes and the different monitoring objectives, together with the continuous development of the understanding of fiber-optic monitoring mechanisms and data acquisition, processing, and interpretation methods, the effective application of this technology in hydraulic fracturing monitoring still faces certain challenges. To promote the standardized design and reasonable application of this technology, this paper reviews and summarizes the monitoring methods and technical experience of fiber-optic monitoring technology in hydraulic fracturing.

Based on field practices from 11 typical hydraulic fracturing test sites at home and abroad, this paper systematically summarizes the monitoring scheme design and data acquisition strategies of fiber-optic monitoring technology in fracturing process monitoring and post-fracturing production evaluation. First, regarding the selection of hydraulic fracturing fiber-optic monitoring schemes, this paper discusses the possible monitoring results, corresponding interpretation techniques, and technical difficulties in field implementation from three major application scenarios: in-well monitoring during fracturing, offset-well monitoring, and post-fracturing production profile monitoring. The differences among different monitoring schemes in monitoring objectives, applicable conditions, and interpretation logic are mainly analyzed, including fracture hit identification, near-wellbore fracture initiation, fracture propagation characteristics, interwell fracture communication, cluster efficiency evaluation, and post-fracturing production contribution analysis. Second, in terms of the selection of important technical parameters for hydraulic fracturing fiber-optic monitoring, based on various field practices, the selection and influence of technical parameters such as fiber deployment mode, spatial resolution, sampling interval and frequency, and monitoring well spacing in offset-well monitoring are summarized and analyzed. Finally, the methods of fiber-optic monitoring data acquisition and storage are discussed. In addition to the basic requirements of field data recording, this paper focuses on standardized data storage formats, quality control of acquired data, full lifecycle management of massive datasets, and future development directions involving integration with edge gateways, cloud platforms, and artificial intelligence-based interpretation methods.

The analysis shows that a single fiber-optic monitoring configuration is difficult to meet all monitoring objectives, and the monitoring scheme should be specifically designed according to geological and engineering conditions, the target information to be obtained, and the available interpretation methods. Therefore, clear monitoring objectives, a reasonable fiber deployment strategy, and a standardized data workflow are important prerequisites for improving the reliability and applicability of fiber-optic monitoring results. This paper provides a practical reference for the design of hydraulic fracturing fiber-optic monitoring schemes, the selection of technical parameters, and the development of interpretation methods, and also provides support for the further improvement and field application of fiber-optic monitoring technology.

Drilling digital twin technology, as a core method for virtually mapping actual drilling processes and enabling visual monitoring and intelligent decision-making, has become one of the inevitable trends in the future development of intelligent drilling. Drilling digital twins primarily involve two aspects: surface drilling rigs and subsurface wellbores. While surface rigs have entered systematic research phases globally due to their relatively stable nature and abundant mature experience, subsurface wellbores present significant challenges. These include invisibility, tangibility issues, high uncertainty, complex operating conditions, and the involvement of multi-medium and multi-physics coupling phenomena. This paper systematically elaborates on the core theories, key technologies, and development trends of constructing and applying digital wellbore models for drilling operations. It aims to provide a reference for subsequent research and engineering practice in this field.

The Lucaogou formation of the Permian System in the Jimsar Sag, Junggar Basin, represents a typical example of mixed siliciclastic-carbonate continental shale oil in China. After more than a decade of exploration and pilot development, remarkable technological and production breakthroughs have been achieved, confirming its substantial resource potential. Nevertheless, the current depletion-based development mode continues to face critical challenges such as rapid formation energy loss, sharp production decline rates, and persistently low recovery factors. The key technical direction for improving oil recovery at the later stage of development therefore remains unclear and urgently requires systematic investigation. To address these issues, a coupled numerical model of hydraulic fractures and the shale matrix was established based on the actual geological model of a well platform located in the lower sweet-spot interval of the Jimsar Sag. The model was validated through history matching by optimizing key reservoir and fracture parameters. On this basis, an inter-well gas injection displacement scheme involving three horizontal wells was designed, in which gas is injected through the central well and oil is produced from the two adjacent wells following an initial depletion phase. Simulation results show that, for Jimsar shale oil characterized by an ultra-tight matrix and poorly developed natural fractures, the degree of communication between horizontal wells through hydraulic fractures is the key factor determining the effectiveness of gas injection development. Comparative analyses of different injected gases (CO₂ and CH₄), injection timing, and injection pressure were performed to clarify their influence on the displacement process. Based on these results, an optimized operation strategy was proposed, involving cyclic reinjection of produced gas after the initial breakthrough of the injected gas at the production wells. Compared with depletion development, the designed inter-well gas injection scheme increases the cumulative oil production by approximately 45%, while the oil utilization ratio-an indicator of carbon and energy efficiency-nearly doubles. The proposed cyclic reinjection of produced gas further enhances carbon utilization efficiency and reduces the external CO₂ supply demand, thereby achieving a dual objective of improving oil recovery and promoting effective carbon recycling. Overall, the developed coupled model and the derived optimization strategy provide a technically feasible and scientifically grounded approach for enhancing oil recovery in continental shale oil reservoirs. The findings clarify the mechanisms and governing parameters of inter-well gas displacement under ultra-tight conditions, and the proposed cyclic reinjection mode offers a practical pathway for synergistic oil production enhancement and CO₂ geological utilization. This study provides important theoretical support and engineering reference for large-scale application of gas injection displacement and CO₂ sequestration in the late-stage development of continental shale oil fields in China.

Accurate in-situ stress determination is vital for evaluating the mechanical state of subsurface rock and supports key petroleum engineering operations such as wellbore stability, hydraulic fracturing design, and fracture propagation control. Increasingly complex geological settings expose the limitations of conventional measurement methods in resolution, accuracy, and continuity. Fiber-optic sensing provides a means for continuous and high-precision monitoring under such conditions. This work presents an in-situ stress measurement method that combines fiber-optic strain sensing with a reverse differential strain mechanism. By applying internal pressure to drive microfractures around the borehole from closure to reopening, the method infers in-situ stress from the mechanical response of t he rock. Laboratory physical simulation experiments were conducted to establish the theoretical framework and measurement workflow. Results show that fiber-optic strain data effectively capture rock anisotropy and heterogeneity, enabling estimation of in-situ stress ratios through the reverse differential strain model. Compared with the Kaiser acoustic emission method, the approach yields an 8%~24% error range, averaging 16%. Although still in its early phase, the method offers a promising technical pathway for in-situ stress determination in oil and gas engineering and provides a foundation for future refinement and application.

Geological carbon storage (GCS) is one of the key technologies to achieve Carbon peak and carbon neutrality goals. Since industrial-source CO₂ usually contains various impurity gases, and its purification is costly and technically challenging, these impurities are often injected into the subsurface together with CO₂ in practical engineering applications. Based on the geological characteristics of the Shiqianfeng Formation in a CCS demonstration project area, a two-dimensional geological model was established, and reactive transport simulations of impure CO₂ were conducted using the CMG-GEM compositional simulator. N₂ and H₂S were selected as representative impurity components to investigate the migration pathways, occurrence forms, and spatial distribution characteristics of impure CO₂ in the saline aquifer. The study systematically analyzed the dominant trapping mechanisms, including structural trapping, residual gas trapping, and solubility trapping, at different storage stages, and explored the role of capillary pressure in the geological storage process of impure CO₂. The simulation results indicate that during the early injection stage, CO₂ predominantly accumulates in the supercritical state at the top of the reservoir. Over time, the capillary trapping effect gradually emerges, more CO₂ is retained in the reservoir pores, promoting the continuous conversion of supercritical CO₂ into bound and dissolved forms, thereby significantly enhancing the overall storage stability and safety. CO₂ migration exhibits pronounced spatiotemporal heterogeneity: the injected gas rapidly rises under buoyancy and accumulates beneath the caprock, then gradually migrates downward under density and concentration gradients, promoting dissolution. After injection ceases, the gas spreads laterally along the base of the caprock, forming a tongue-shaped migration front with a maximum diffusion distance of approximately 650 m. The transport behaviors of the impurity components differ significantly. Due to its low solubility, N₂ tends to accumulate at the leading edge of the gas-liquid displacement front, whereas CO₂ and H₂S, with higher solubility, form dissolution-enriched zones near the injection well, reaching peak solubilities of 1.4 mol/kg-H₂O and 0.53 mol/kg-H₂O, respectively. Capillary pressure plays a crucial role by suppressing the rate of gas-phase migration, enhancing dissolution trapping efficiency, and inducing reverse imbibition of formation water during the post-injection stage, thereby promoting greater retention of CO₂ in pore spaces in the form of residual gas and effectively increasing the proportion of residual trapping. Comprehensive analysis demonstrates that the storage behavior of impure CO₂ in saline aquifers is jointly governed by impurity properties, trapping mechanisms, and capillary pressure effects. The findings provide scientific support for optimizing injection strategies in impure CO₂ geological storage projects and offer important guidance for ensuring long-term storage security and improving storage efficiency.