非常规油气在全球能源中扮演的角色越来越重要。致密砂岩储层作为非常规储层的典型代表，对其实现大规模经济开采一直是能源行业关注的焦点。但致密砂岩通常具有低孔隙度和低渗透率的特点，所以需要在储层中构建人工缝网来增加油气从孔隙向井筒运移的能力。因此研究致密砂岩微裂缝网形成的力学机制至关重要。本文聚焦于以脆性矿物为主的致密砂岩，基于二维离散元法构建了数值模拟模型，模型考虑了多组分矿物特性以及岩石微结构的影响，实现了多种加载方式下的缝网生长模拟。同时，针对当前受困于原位高分辨率观测压裂缝网的难题，搭建了水力压裂在线扫描平台（HFCP），并设计了相应的实验流程和分析方法，探讨了致密砂岩原位水力压裂的一般规律。具体的研究工作如下：
探究了拉应力下岩石缝网生成和扩展机制。采用巴西劈裂的加载方式，研究了岩石模型中心形成微裂缝网的规律。通过实验结果验证了模型的力学性能和缝网生长计算的准确性。并采用局部拉伸模型证明了原生微裂缝具有比微孔更加明显的优先取向效应。随后进一步讨论了原生微裂缝和微孔对模型宏观断裂和应力-应变曲线的影响。计算结果表明孔密度升高会使得岩心发育扭转拉伸裂缝，而裂缝密度升高则没有出现这种现象。在相同的中心拉伸应力下，受优先取向效应影响的原生微裂缝比微孔更容易诱导岩石微裂缝的形成。当裂缝密度超过0.324  或孔密度超过0.041  时，生长微裂缝的数量主要由微缺陷密度控制，此时可以忽略弹簧刚度比的影响。
提出了原位水力压裂实验及结果分析方法。针对在微米CT中观测水力压裂缝网遇到的难点，设计了新的压裂零件、并提出了适配的安装方式和功率调整方案。最终搭建了开展原位水力压裂实验的HFCP，设计了相应的实验流程。利用HFCP开展了致密砂岩水力压裂实验，并在实验过程中实现注入流体压力的全程监测以及缝网图像的后处理。随后基于压裂实验结果讨论了注入速度和围压对流体压力曲线的影响。实验结果表明，在不同围压下，随注入速度的增加，流体的破裂压力和平均裂缝开度都有所升高，但分支裂缝体积分数变化呈现多样化特征。
明确了水力压裂对岩石缝网生成和扩展的影响机制。基于真实岩石的物理性质构建了水力压裂数值模型，模型的加载与室内HFCP压裂实验保持一致。在数值理论解和压裂实验验证的基础上，讨论了缝网生长中的注入速度和围压变化的最优方案。随后在岩石模型中添加原生微裂缝和微孔，比较了两类微缺陷对流体压力曲线、颗粒云图和缝网分布的影响。计算结果表明随着围压升高，破裂压力呈线性增加，而裂缝平均宽度呈指数函数单调减少。当应力系数在0.532-0.572之间时，压裂缝网的分支裂缝面积分数最高。与不同矿物和微裂缝相比，微孔能引起模型内流体压力曲线更强烈的波动以及颗粒位移和应力分布的显著变化。此外，各向应力能改变模型的裂缝生长方向并对缝网分布产生影响。
本文的研究结果有助于压裂施工中进一步控制人工缝网的生长，同时为提高致密砂岩储层油气采收率提供了理论基础。

The Unconventional oil and gas resources are playing an increasingly important role in global energy. Tight sandstone reservoirs, as typical representatives of unconventional reservoirs, have been a focal point of the energy industry in achieving large-scale economically viable exploitation. However, tight sandstone formations typically exhibit low porosity and permeability, necessitating the construction of an artificial fracture network within the reservoir to enhance the migration of hydrocarbons from the pores to the wellbore. Therefore, studying the mechanical mechanisms of microfracture network formation in tight sandstone is crucial. This paper focuses on tight sandstone dominated by brittle minerals and presents a numerical simulation model based on a two-dimensional discrete element method. The model takes into account the influence of multi-component mineral characteristics and rock microstructure, enabling simulation of fracture network growth under various loading conditions. Furthermore, to address the current challenge of in situ high-resolution observation of hydraulic fracture networks, a hydraulic fracturing online scanning platform (HFCP) is established, accompanied by corresponding experimental procedures and analysis methods, to investigate the general patterns of in situ hydraulic fracturing in tight sandstone. The specific research contributions are as follows:
Investigating the mechanisms of fracture network generation and propagation under tensile stress. The Brazilian splitting loading method is employed to study the formation patterns of microfracture networks at the center of the rock model. The experimental results validate the mechanical performance of the model and the accuracy of fracture network growth calculations. Furthermore, a local tensile model is utilized to demonstrate that primary microfractures exhibit a more pronounced preferential orientation effect compared to micropores. Subsequently, the influence of primary microfractures and micropores on macroscopic rock failure and stress-strain curves is further discussed. Computational results reveal that an increase in pore density leads to the development of twisting tensile fractures within the rock core, while an increase in fracture density does not exhibit such behavior. Under the same central tensile stress, primary microfractures affected by the preferential orientation effect are more likely to induce the formation of rock microfractures compared to micropores. When the fracture density exceeds 0.324  or the pore density exceeds 0.041  , the number of growing microfractures is primarily controlled by microdefect density, and the influence of the stiffness ratio of springs can be neglected.
An in situ hydraulic fracturing experiment and corresponding result analysis method are proposed. To address the challenges encountered in observing hydraulic fracture networks using micro-CT, new fracturing components are designed, along with compatible installation methods and power adjustment schemes. The Hydraulic Fracturing Online Scanning Platform (HFCP) is ultimately constructed to conduct in situ hydraulic fracturing experiments, accompanied by a designed experimental procedure. Utilizing the HFCP, hydraulic fracturing experiments on tight sandstone are conducted, enabling real-time monitoring of injected fluid pressure and post-processing of fracture network images. Subsequently, based on the experimental results, the influence of injection rate and confining pressure on the fluid pressure curve is discussed. The experimental results demonstrate that, under different confining pressures, an increase in injection rate leads to higher fluid breakdown pressure and average fracture aperture, but the variation of branch fracture volume fraction exhibits diverse characteristics.
The mechanisms of hydraulic fracturing impact on fracture network generation and propagation are clarified. A numerical model for hydraulic fracturing is constructed based on the physical properties of actual rock, with loading conditions consistent with the laboratory HFCP fracturing experiments. Building upon the theoretical solutions and experimental verification, the optimal strategies for injection rate and confining pressure variations during fracture network growth are discussed. Subsequently, primary microfractures and micropores are introduced into the rock model, and the impact of these two types of microdefects on the fluid pressure curve, particle cloud images, and fracture network distribution is compared. The computational results demonstrate that as the confining pressure increases, the breakdown pressure linearly increases while the average fracture width monotonically decreases exponentially. The highest fraction of branch fracture area in the fracturing network occurs when the stress coefficient is within the range of 0.532-0.572. Compared to different minerals and microfractures, micropores cause more significant fluctuations in the fluid pressure curve within the model, as well as notable changes in particle displacement and stress distribution. Additionally, the presence of isotropic stresses alters the direction of fracture growth and affects the distribution of the fracture network.
The research findings of this study contribute to further controlling the growth of artificial fracture networks in hydraulic fracturing operations, while providing a theoretical foundation for enhancing the oil and gas recovery efficiency in tight sandstone reservoirs.