The formation mechanism and propagation behaviors of a three-dimensional hydraulic fracture network in fractured shale reservoirs remain unclear, especially when the scale of hydraulic fractures is much larger than that of natural fractures. In this study, taking the well XH in the Longmaxi shale reservoir in the Sichuan Basin, China as an example, we develop a fully three-dimensional numerical model for hydraulic fracturing coupled with microseismicity based on the discrete lattice method. We introduce a randomly generated discrete fracture network into the proposed model and explore the formation mechanism of the hydraulic fracture network under the condition that the hydraulic fractures are much larger than natural fractures in scale. Moreover, microseismic events are inversely synthesized in the numerical model, which allows the evolution of the fracture network to be monitored and evaluated quantitatively. In addition, we analyze the effects of injection rate, horizontal stress difference, and fluid viscosity on fracture propagation. Our results show that when the scale of hydraulic fractures is much larger than that of natural fractures, the fracture morphology of “main hydraulic fractures + complex secondary fractures” is mainly formed. We find that a high injection rate can not only create a complex fracture network, but also improve the uniform propagation of multi-cluster fractures and enhance far-field stimulation efficiency. Optimizing the horizontal wellbore intervals with low horizontal stress differences as the sweet spots of hydraulic fracturing is also beneficial to improve the stimulation efficiency. For zones with a large number of natural fractures, it is recommended to use an injection schedule with high viscosity fluid early and low viscosity fluid late to allow the hydraulic fractures to propagate to the far-field to maximize the stimulation effect.

Recently, the Changning shale gas field has been one of the most outstanding shale plays in China for unconventional gas exploitation. Based on the more practical experience of hydraulic fracturing, the economic gas production from this field can be optimized and gradually improved. However, further optimization of the fracture design requires a deeper understanding of the effects of engineering parameters on simultaneous multiple fracture propagation. It can increase the effective fracture number and the well performance. In this paper, based on the Changning field data, a complex fracture propagation model was established. A series of case studies were investigated to analyze the effects of engineering parameters on simultaneous multiple fracture propagation. The fracture spacing, perforating number, injection rate, fluid viscosity and number of fractures within one stage were considered. The simulation results show that smaller fracture spacing implies stronger stress shadow effects, which significantly reduces the perforating efficiency. The perforating number is a critical parameter that has a big impact on the cluster efficiency. In addition, one cluster with a smaller perforating number can more easily generate a uniform fracture geometry. A higher injection rate is better for promoting uniform fluid volume distribution, with each cluster growing more evenly. An increasing fluid viscosity increases the variation of fluid distribution between perforation clusters, resulting in the increasing gap between the interior fracture and outer fractures. An increasing number of fractures within the stage increases the stress shadow among fractures, resulting in a larger total fracture length and a smaller average fracture width. This work provides key guidelines for improving the effectiveness of hydraulic fracture treatments.

The mechanical properties of multi-lithologic reservoir rock masses are complex, and the failure mechanism is not clear. This research belongs to the field of oil and gas exploration and development. Brazilian splitting, and digital image correlation (DIC) tests were performed to study the mechanical properties and failure mechanism of assemblages containing sandstone, shale, and limestone. In the Brazilian splitting test, the tensile strength of the rock mass first increased and then decreased with an increasing angle. As the interface angle increased, The morphology of the splits is also changing. In the DIC test, according to the deformation characteristics at different angles, the Brazilian splitting test process was summarized into three stages: interface deformation, elastoplastic, and failure, the interface has a great influence on the fracture propagation, and the fracture tends to expand along the interface. Interface characteristics were studied based on 3D scanning technology and fractal method, with an increase in the angle, the fractal dimension D showed first decreased, then increased, and further decreased, which was identical to the variation trend of the maximum load in the Brazilian splitting test and the maximum principal strain in the DIC test with the angle. This study provides reference for predicting fracture propagation trend in field hydraulic fracturing work.

The injection volume and the distribution of a proppant inside a fracture have a direct impact on the stimulation effect of fracturing. In this study, a new proppant transport model was established based on the Euler method. In this model, the proppant plugging element allows fluid to pass through. Furthermore, the proppant plugging process was successfully simulated based on this model. The proppant transport and ultimate injection concentration under different injection modes were discussed. The numerical simulation results indicate that compared with the strategy of constant concentration, the strategy of a stepwise increasing concentration can make the proppant distribution in the fracture more uniform. The strategy of injection with a stepwise increasing concentration and a periodic injection with a stepwise increasing concentration can increase the injection volume of the proppant by 25%. In the fracture network, a 67% increase in the number of branch fractures resulted in a 17% increase in the maximum proppant injection volume. If the branch fracture width is reduced by 50%, the maximum proppant injection volume is reduced by 17%.

Deep-learning-based (DL-based) seismic inversion has emerged as one of the state-of-the-art research areas in exploration geophysics with the development of artificial intelligence technology. Due to its good portability and high computational efficiency, this method has emerged as a data-driven approach for estimating subsurface properties. However, most of the current DL-based methods rely solely on seismic data, lacking the incorporation of prior information. In addition, these methods are usually performed trace-by-trace, resulting in insufficient horizontal constraints. These limitations make traditional methods less robust, particularly when dealing with high noise levels or limited data. To address these challenges, we propose a multi-input deep learning network for pre-stack inversion, which combines data-driven and model-driven approaches for optimization. The proposed method separately extracts features from the model and data, merging them to improve feature utilization. Moreover, by adopting a 2-D training unit, rather than a trace-by-trace approach, the method improves the horizontal continuity of the results. Tests on synthetic and real seismic data confirmed the robustness and improved stability of the proposed method, even under challenging conditions. This dual-driven approach significantly enhances the reliability of seismic inversion.

The stress threshold of deep reservoir shale subjected to fracturing fluid immersion is an important factor affecting fracture initiation and propagation during fracturing. However, little information has been reported on the effect on shale of soaking at high temperature and high pressure (HTHP). In this study, immersion tests and triaxial compression tests were carried out at reservoir temperature and in-situ stress on the downhole cores with different mineral compositions. The characteristics of stress thresholds, i.e., crack initiation stress (${\sigma }_{\mathrm{c}i}$), crack damage stress (${\sigma }_{\mathrm{cd}}$), and peak deviator stress (${\sigma }_{\mathrm{p}}$), of shale affected by the different times of soaking with low-viscosity fracturing fluid (a) and the different viscosity fracturing fluids (a, b, and c) were investigated. The results show that hydration at HTHP has a significant softening effect on the stress thresholds (${\sigma }_{\mathrm{c}i}$, ${\sigma }_{\mathrm{cd}}$, ${\sigma }_{\mathrm{p}}$) of reservoir shale, but the softening rate varies for samples with different mineral compositions. The crack initiation stresses of quartz-rich and clay-rich shales treated with different soaking times and different soaking media remain almost unchanged in the range of 47 to 54% of the corresponding peak strength, while the crack initiation stresses of carbonate-rich shales are significantly affected. The ratio ${\sigma }_{\mathrm{cd}}$/${\sigma }_{\mathrm{p}}$ of quartz-rich shale is significantly affected by the different viscosity fracturing fluids (a, b) and the different times of soaking with low-viscosity fracturing fluid (a), while clay- and carbonate-rich shales are less affected. The results of this study can provide a reference for the fracturing design of deep shale gas development.