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.

Natural fractures in middle–deep shale reservoirs experience continuous stress evolution under the combined influence of hydraulic stimulation and long-term production. Their stability evolution critically affects wellbore integrity, stimulation effectiveness, and the safety of infill-well deployment. Focusing exclusively on natural fractures, this study develops a stability evaluation framework by coupling a 3D geomechanical model with a discrete fracture network (DFN). The approach maps the strike and dip of natural fractures onto the 3D grid and constructs a spatially continuous fracture-orientation volume through geostatistical interpolation. This enables the unified coupling of natural fracture geometry with the regional 3D stress field and rock mechanical attributes, providing a continuous 3D quantification of natural-fracture stability. Results show that natural-fracture slip risk is governed by the combined effects of fracture orientation and tectonic stress regime, with distinct high-risk orientations under normal-faulting, strike-slip, and reverse-faulting conditions. Fluid injection may trigger natural-fracture instability through reduced effective normal stress and lowered frictional strength, whereas long-term production enhances effective stress and generally improves natural-fracture stability. In the YS108 block, the stress regime evolves toward a typical normal-faulting state after nearly ten years of production, leading to significantly reduced slip risk of natural fractures. The proposed 3D evaluation framework provides a practical basis for post-production stability assessment, well-trajectory optimization, and stimulation-risk management in middle–deep shale reservoirs.

The lacustrine shale oil reservoirs in China exhibit pronounced heterogeneity and anisotropy. Conventional approaches for characterizing shale mechanical properties—such as plug-scale mechanical testing and log-derived elastic estimates—provide only coarse resolution at meter- to decimeter-scale. These methods lack the precision required to capture the fine-scale heterogeneity required for advanced geomechanical analysis. To address this limitation, we conducted scratch testing, uniaxial compression testing, and CT scanning on full-diameter core samples from the Yingxiongling area in the Qaidam Basin. This study investigates the principles and analytical methods of scratch testing and validates the approach for calculating continuous mechanical parameters along full-diameter cores. Key findings demonstrate significant strength variations among different lithofacies in the Yingxiongling area. Scratch testing successfully provides continuous profiles of uniaxial compressive strength and Young's modulus across full-diameter cores. Furthermore, it visually identifies transition zones between lithofacies and reveals interlayer fracture locations within the core. This study establishes an efficient method for obtaining continuous, high-resolution mechanical parameters from full-diameter cores, offering theoretical and methodological support for enhancing the accuracy of mechanical property characterization in shale oil reservoirs.

In view of the scientific problem of non-uniform initiation of hydraulic fracturing perforation clusters in unconventional anisotropic reservoirs, this paper considers the anisotropy of weak plane and the superposition of stress concentration effect of perforation hole and wellbore, and establishes the stress model of arbitrary directional hole and perforation hole and the analytical model of hydraulic fracture initiation morphology prediction. The initiation conditions of vertical and horizontal fractures are studied. the mechanical analysis of two wells with the largest difference in production in Luobu block is carried out. The results show that the stress distribution around the perforation hole has a significant influence on the fracture initiation pressure, which affects the geometry of the fracture near the wellbore. In well Y1, which is arranged along the direction of the minimum horizontal principal stress and is mainly in the stress state of the strike-slip partial normal fault, the main mechanical reason for the absence of gas-liquid is that the two stress concentrations of the borehole and the perforation channel lead to the reversal of the horizontal principal stress of the in-situ stress field near the well, and the stress state becomes reverse fault. The near well is mainly horizontal fracture, and the gas-liquid migration channel is blocked. In the Y2 well, which is in the stress state of the strike-slip normal fault, the stress state becomes a strike-slip reverse fault after two stress sets, and the vertical fracture initiation range accounts for more than 75%. The connectivity of the near-well fracture channel is good, which is helpful for the migration of gas and liquid in the reservoir. This study offers insights for the optimal design of perforation holes in hydraulic fracturing, providing information on the minimum fracture initiation pressure and the optimal perforation direction for the wellbore-reservoir connecting channel in practical field applications.

The development of tight oil reservoirs in the Mahu conglomerate formation poses formidable challenges, marked by pronounced reservoir heterogeneity, significant horizontal bidirectional stress difference, the absence of natural fractures, and poor reservoir physical properties. In pursuit of a more economically efficient development strategy, a field trial involving tight well spacing and stereoscopic development was initiated for the first time in the Ma131 well area. Despite achieving a relatively high overall recovery rate in the trial area, the economic benefits fell short of expectations, prompting an urgent need for well spacing optimization.This study adopts an integrated approach that combines geological and engineering principles to establish a comprehensive method and process for optimizing the well spacing of stereoscopic well networks. Additionally, a method for rapidly fitting the hydraulic fracture network model under conditions involving multiple wells and fracturing stages is proposed, combining fracturing parameters and treatment pressure response characteristics. The key steps involve a systematic reservoir engineering analysis, utilizing unconventional fracture models based on detailed geological and geomechanical models to simulate complex fracture networks and match the history treatment pressure. The process further includes coupled reservoir numerical simulation for historical production matching. Well spacing optimization in the Ma131 stereoscopic development demonstration area was achieved through an integrated simulation of hydraulic fracturing and production, with validation conducted through a large scale well spacing field trial.The research results show that the analytical fracture length and the simulated fracture length can be mutually verified. The fractures in horizontal wells of the Bai3 section are relatively long, with an average propped half fracture length of 70.1 m and an average hydraulic fracture height of 24.6 m. In the Bai2 section, horizontal well fractures are relatively short and exhibit layer-penetrating effects, with an average propped half fracture length of 61.1 m and an average hydraulic fracture height of 28.3 m. Under conditions of certain permeability in the formation, the well spacing for both development layers can be appropriately expanded to a range of 200~300 m, ensuring an increase in individual well productivity and economic benefits while maintaining a high recovery rate.The proposed optimized well spacing range can be extended to similar development layers within the same well area, based on the well spacing field trial validation results. And the well spacing optimization method and process proposed in this study, utilizing a stereoscopic well network, can serve as a valuable reference for other unconventional oil and gas reservoir types.

Tight sandstone and shale reservoirs have gradually become the main sources of global oil and gas production growth. Accurate reservoir evaluation is the key to achieving efficient development of oil and gas fields, and experimental measurement of the fluid flow capacity of oil and gas reservoirs is one of the important aspects of reservoir evaluation work. Traditional permeability measurement methods are based on the steady-state method, that is, the permeability is calculated based on Darcy’s law when the pressure drop at the inlet and outlet of the core reaches equilibrium. However, for ultra-low permeability and shale reservoirs, the steady-state method is difficult to apply. Therefore, transient method needs to be used instead of the steady-state method, and the pressure pulse decay experimental device has become a commonly used means to study ultra-low permeability. This paper comprehensively analyzes the development process of pressure pulse decay experiments in the past 50 years, and focuses on the representative research results. In view of the deficiencies of the existing technical means, this paper proposes the research direction of strengthening the fundamental research into fluid flow in ultra-low permeability porous media. Currently, one-dimensional analytical solutions and numerical models can calculate ultra-low permeability with high precision, and two-dimensional and three-dimensional models are also rapidly developing. However, the efficient embedding of complex-shaped and multiscale fractures into pressure pulse decay models is still a key problem to be solved. For adsorptive gases such as shale gas, it is a fundamental scientific problem that needs to be further explored to clarify the difference between static adsorption and dynamic adsorption and effectively incorporate the adsorption curve into the flow model. Based on this, the rock physics characteristics under multi-physical effects can be evaluated through pressure pulse decay experiments. In addition, the systematic study of non-Darcy flow behavior under relatively low pressure conditions is also an important way to promote scientific calculation of pressure pulse decay experiments. Over the past few decades, significant progress has been made in the experimental and simulated evaluation of the storage and permeability of ultra-low permeability rock cores. Further theoretical research and the development of advanced experimental devices will be important. The progress of related research has important strategic significance for the efficient exploration and development of China’s tight and shale oil and gas reservoirs and the guarantee of national energy security.

Lithic-rich tight sandstone is one of the most enrichment lithofacies in the Sulige gas field. Clarifying the enrichment mechanism of high-quality lithic-rich tight sandstone is important to economic and efficient development of the tight gas reservoir. This paper introduces a new classification method, which is based on the origin of particles and interstitial materials and their control on reservoir pores growth. Lithic-rich tight sandstone can be subdivided into three types: sedimentary lithic sandstone, diagenetic lithic sandstone and event-type lithic sandstone. The genetic mechanism of a high-quality reservoir is studied by this new method. Research shows that the sedimentary lithic sandstone has high contents of plastic lithics, strong compaction effects of early diagenesis, large porosity reduction and almost no dissolution-induced porosity. The diagenetic lithic sandstone has high contents of rigid lithics and strong compaction effects. Organic acids promote alteration of a large amount of feldspars into kaolinite, while such sandstones are highly cemented. It is seen with moderate porosity reduction and moderate dissolution-attributed porosity growth. Event-type lithic sandstone also has high contents of rigid debris and strong compaction effects. Synsedimentary volcanic dust materials of subaerial deposition are altered into illite through smectite and illite-smectite mixed-layer clay under the effects of acids, which generate many pores and results in large dissolution-attributed porosity growth. Research shows that the sedimentary lithic sandstone has poor physical properties and is identified as the unfavorable reservoir; the diagenetic lithic sandstone having medium physical properties, as the relatively favorable reservoir; the event-type lithic sandstone having good physical properties, as the favorable reservoir. The research route and results have laid a solid geological foundation for better development of lithic-rich tight sandstone reservoirs.