Conclusions and Future Research

In this book, we introduce 3D digital shale rocks reconstructed by combining experimental data (e.g., pore size, shape, porosity, roughness, etc.) and a random algorithm. This approach ensures the representativeness of the digital sample’s heterogeneity while avoiding the time-consuming and costly full imaging process typically required by physical experimental methods. However, the simplification of pore or mineral particle morphology in these stochastic algorithms may introduce uncertainties in fluid modeling results. It is necessary to strike a balance between calculation resolution and characterization dimension. Therefore, the current trend in characterizing nanoporous shale at the pore scale focuses on improving the representativeness of digital rocks while reducing experimental efforts. This involves incorporating more realistic morphological data, considering burial and diagenesis processes, generating more realizations based on different SEM images to cover variabilities, and utilizing advanced deep learning algorithms for rock generation (Tahmasebi et al., 2020; Yang et al., 2023). Although this book primarily discusses pore-scale modeling in the nanoscale, it is essential to acknowledge that fluid transport in shale formations is a typical multiscale process. Thus, establishing multiscale digital rock models is also a highly relevant topic. For example, combining serial sectioning imaging with FIB technology enables the creation of high-resolution small-scale digital rock models, while micro-CT can generate lower-resolution large-scale digital rock models. Integrating these two scales of digital rock models allows for obtaining large-scale high-resolution representations (Wu & Tahmasebi, 2023). This highlights the importance to develop new algorithms for integrating pore-throat structures across different scales.The success of the proposed pore-scale LBM for modeling water sorption in nanosystems lies in its ability to accurately reproduce scenarios from the literature, including water sorption hysteresis inside nanopores (Li et al., 2017) and water adsorption in shale containing both hydrophilic and hydrophobic sites (Sang et al., 2020). Through the combination of accurate shale digital rocks, the LBM emerges as one of the best candidates for visualizing water distribution in nanoporous media, eliminating the need for indirect methods like water sorption experiments. However, one drawback of this model that still needs to be addressed is the lack of a direct transfer of lattice density into vapor pressure (relative humidity), making it inconvenient to directly compare modeling data with experimental data. Additionally, coupling the clay swelling induced by water adsorption is another possible investigation direction for the future. This is because condensed water is usually free of ions, making the dissociation of exchangeable cations such as sodium ions much easier. As a result, negatively charged clay layers separate from each other under electrostatic repulsion, leading to clay swelling. Clay swelling commonly occurs during water adsorption experiments, significantly altering the porous structure of shale (Li et al., 2016). Local wettability heterogeneity is also crucial in determining water adsorption distribution in 3D pore spaces. In addition to accurate digital rocks, conducting more pore-scale contact angle experiments (Deglint et al., 2017) is necessary to characterize the local wettability of shale and improve the prediction reliability of water distribution.Currently, methane sorption research primarily focuses on the adsorption process, with limited attention given to the desorption process. However, the desorption hysteresis can significantly influence gas production predictions during late-stage development, warranting separate study. In recent years, deep shale gas reservoirs (with burial depths greater than 3500 meters) have emerged as a newly explored field in China. These deep shale gas reservoirs are characterized by high temperatures (120–150 °C) and high pressures (>70 MPa) (Wu et al., 2021). Due to experimental limitations (with maximum pressure capabilities below 70 MPa), the adsorption/desorption research on shale gas predominantly concentrates on shallow and intermediate reservoirs. Therefore, it is recommended to combine MD simulations and LBM simulations to establish a mathematical model for adsorption/desorption at high temperatures and high pressures, suitable for engineering applications. Given the unique physicochemical properties of supercritical CO2, CO2 is considered an excellent fluid medium for injection into shale gas reservoirs. This approach can not only enhance the gas recovery rate but also sequester CO2. The proposed LBM can be easily calibrated to fit the physicochemical properties of CO2, enabling competitive adsorption studies between CH4 and CO2 and providing more in-depth implications for CO2 injection in shale gas reservoirs. Specifically, competitive adsorption between CH4 and CO2 is a dynamic process during CO2 injection into shale reservoirs. It would be interesting to couple the microscopic gas transport mechanisms with the sorption LBM to capture this dynamic process in the future.Because the shale matrix is extremely tight, conducting conventional laboratory experiments is difficult. Therefore, modeling tools have become a popular alternative for capturing gas transport in nanoporous shale. Over the last 20 years, extensive investigations have been conducted to model gas transport in nanoporous shale, using a range of analytical/semi-analytical methods and simulation methods (LBM, PNM, MD, etc.). The improvements in gas modeling include the incorporation of more transport mechanisms (Knudsen diffusion, surface diffusion, real gas effect, etc.) and more accurate representation of nanoporous shale. In this book, we introduce a semi-analytical method that considers all the known transport mechanisms to quantify the gas apparent permeability of nanoporous shale. Actually, significant success has been achieved in the LBM community to simulate gas transport in nanoporous shale. The state-of-the-art pore-scale LBM can incorporate all gas transport mechanisms (Zhao et al., 2023). Moreover, the REV-scale LBM, considered a Darcy-type equation solver, has been proposed to simulate gas transport at a larger scale, allowing for the consideration of the influence of microscale fractures along with the nanoscale pores (Chen et al., 2015). Currently, researchers are still working on unsolved problems, such as further improving the pore-scale LBM to adapt to gas transport at higher Knudsen numbers (Zuo et al., 2019), using machine learning to accelerate simulations (Wang et al., 2021), and clarifying the contribution of surface diffusion, among others.The proposed single-phase water nanoscale LBM has wide applicability across various industries and significantly improves computational efficiency compared to MD simulations. However, the model’s computational cost for calculating flow within REV-scale porous media still remains unacceptable. Future efforts to reduce the model’s computational cost will involve implementing techniques like adaptive meshing and non-uniform grid methods (Yu & Fan, 2009). This approach allows for setting grid accuracy near the wall, influenced by nanoscale effects, to the size of a single molecule’s diameter, while using coarser grids in regions farther from the wall. During the fracturing process, fracturing fluid migration is influenced by various physical and chemical processes, such as particle transport and stress sensitivity, which affect the geometric space of fluid transport (Singh, 2016) and, consequently, the apparent permeability. To address these mechanisms, a new LBM-DEM (discrete element method) coupled method is recommended for future research. Additionally, the interaction between fracturing fluid and original reservoir water results in a certain degree of salinity. The pore-scale LBM in this book assumes deionized pure water, and incorporating short-length forces (e.g., electrical force) into the model is a potential direction for future research. In clay nanoporous media, the strong attractive forces on the pore walls and the electroviscous effect resulting from the diffusion double layer increase flow resistance (Cheng et al., 2020), necessitating further investigation.In this book, the pore-scale LBM for two-phase transport at the nanoscale can capture phenomena such as fluid films/gas films near the wall, spatial fluid density, and disjoining pressure. However, the model does not account for the slip effects of gas at the gas-solid boundary and gas-liquid interface. When the water saturation of the shale matrix is low and the water phase primarily flows in the form of films, the gas phase flow channels exhibit good connectivity. Neglecting the interfacial slip effects would underestimate the gas flow capacity. Therefore, a key problem that needs to be addressed in the future is how to identify the gas-water interface and propose a reasonable mathematical representation for slip at this “dynamic interface” within the LBM framework. The REV-scale single-phase LBM has been well established, enabling fluid flow simulations at the core scale. Characterizing the coupling relationship between multiphase fluids in the matrix and microfractures and extending the core-scale single-phase LBM to a multiphase LBM are critical and challenging problems in the future. Additionally, multicomponent (e.g., CH4 and CO2) and multiphase LBM considering the nanoscale effects are also very promising directions for future development. The interactions between different components and different phases should be carefully calibrated with experimental data to properly model this complex flow system.