3-D fracture network reconstruction and quantitative fractal analysis of subsurface rock fractures via integrated CT scanning and box-counting dimension methodology

Abstract

ABSTRACT The inherent challenges posed by natural fracture networks and lithological disintegration in tight oil reservoirs significantly impede drilling integrity, completion efficiency, and hydrocarbon recovery rates. This study introduces a novel multi-scale quantitative characterization framework that fundamentally distinguishes itself from prior works through three key innovations. First, it establishes a pioneering integration of micron-resolution computed tomography (CT) imaging with fractal geometry analysis, enabling unprecedented precision in determining fracture aperture distributions (≤ 1.2 μm accuracy), anisotropic propagation patterns, and spatial density correlations—a significant advancement beyond conventional single-scale fracture characterization methods. Second, this research is the first to develop a coupled poroelastic-fracture model incorporating Analytic Hierarchy Process (AHP)-based size deviation indices , achieving permeability prediction errors of ≤ 4.84% compared to experimental data, substantially improving upon existing models' typical 8-12% error margins. Third, we propose a breakthrough dual-parameter fractal system combining box-counting dimension with fracture concentration degree values, effectively resolving the long-standing limitation of single-fractal-dimension approaches in quantifying complex fracture networks. Unlike previous studies focusing on static fracture characterization, our framework uniquely captures hydraulic fracturing-induced dynamic transformations, revealing a 35-40% decrease in fracture network concentration post-stimulation despite local expansion. Furthermore, the establishment of a morphology-driven natural fracture development index provides the first quantitative link between wall roughness coefficient (JRC) distributions and reservoir fracturability indices , offering actionable insights for stimulation design. Validation across carbonate and tight sandstone reservoirs demonstrates this methodology's ability to reduce completion efficiency uncertainties by 28% compared to conventional approaches. These innovations collectively advance fracture network analysis from qualitative description to true multi-parameter predictive modeling, addressing critical gaps in current practice.

Topics & Concepts