Shear-slip and hydraulic fracture penetration in laminated shale reservoirs: discrete element method (DEM) and explainable machine learning
Year: 2026
Type: article
Abstract: A complex fracture network is critical for enhancing hydrocarbon recovery and deep geo-energy utilization in shale reservoirs with nano- to micro-scale pores. However, extensively developed bedding planes with relatively weak mechanical properties can impede hydraulic fracture penetration, leading to inadequate vertical fracture height growth and a reduced stimulated reservoir volume. To address this issue, a three-dimensional simulation study based on the discrete element method (DEM) was conducted to investigate the penetration behavior of hydraulic fractures in shale reservoirs. The results indicate that, under a strike-slip fault stress regime, bedding-plane shear strength dominates fracture penetration, whereas tensile strength exerts a minor effect. As bedding-plane cohesion and internal friction angle increase, fractures become more inclined toward vertical propagation rather than bedding-plane deflection. Furthermore, there is a critical shear strength threshold for evaluating hydraulic fracture penetration under constant normal stress, wherein higher cohesion for vertical fracture penetration lowers the critical internal friction angle requirement, and vice versa. To predict fracture penetration behavior, an explainable machine learning approach integrating XGBoost and SHAP was applied, achieving both high accuracy and clear insights into the controlling parameters for various fracture behaviors. These findings elucidate fracture propagation mechanisms in laminated shale and similar clay-rich formations under in-situ stress conditions. This provides critical insights not only for hydrocarbon recovery but also for broader geo-energy applications, such as engineered barriers, CO 2 sequestration, and underground storage.
Source: Environmental Earth Sciences
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,290
Award: G8248674989
OpenAlex ID: W7140182394