Discrete fracture matrix modeling: From CT imaging to upscaling fracture properties for subsurface applications

This thesis focuses on modeling fractured porous media to address critical challenges in fluid flow and transport, which are essential for subsurface applications such as carbon capture and storage (CCS), underground hydrogen storage (UHS), and geothermal energy production. The research presented in this thesis combines CT imaging, discrete fracture-matrix (DFM) modeling, and computational methodologies to advance our understanding of fractured porous systems.

The thesis is structured in three parts, comprising five chapters reproduced from peerreviewed journal articles, two chapters reproducing manuscripts that are currently under review, and one chapter with some preliminary results from an ongoing work. The first part addresses the upscaling and modeling of fluid flow in fractured porous media. It introduces a machine-learning framework for upscaling fracture permeability tensors, discusses the tensorial nature of hydraulic apertures, and investigates matrix flow effects on fracture permeability. An empirical correction for upscaled models is proposed for matrix-driven permeability enhancements, validated across single fractures and fracture networks.

The second part of this thesis focuses on deriving fracture networks and developing DFM models using computed tomography (CT) imaging. A novel workflow for generating DFM models from CT scans is presented, enabling accurate characterization of fracture geometry and aperture distributions. This framework is extended to study the impact of confining stress on flow and transport properties, revealing stress-dependent anisotropy and solute transport dynamics in fractured chalk formations.

The third and last part presents an upscaling approach for different effective apertures of rock fractures. Analytical solutions for the different macroscale aperture tensors of fractures with heterogeneous microscale apertures are derived and used for validating the different upscaling approaches. The method is used to explore the relationship between the effective macroscale apertures and other fracture parameters.
The findings summarized in this thesis contribute to robust and efficient modeling of fractured porous media for subsurface applications. By addressing the complex interactions between fractures and porous matrix, this thesis provides insights into mitigating risks associated with fractured rocks in subsurface operations.