Cross-scale crack evolution in pre-flawed sandstone under high temperature: Insights from a composition-aware thermal-mechanical grain-based model

Abstract

In deep energy extraction, thermal damage induced by high temperatures critically degrades the mechanical integrity of rock, posing substantial risks to engineering safety. To systematically clarify the damage mechanisms under coupled thermal-mechanical-fracture conditions, this study develops a novel Composition-Aware Thermal-Mechanical Grain-Based Model (CATM-GBM).This model explicitly integrates the rock's quantitative mineralogical composition and spatial heterogeneity, implementing a temperature-dependent thermal expansion function to capture differential mineral degradation and employing parallel bond model and smooth joint model to simulate intergranular tensile and intragranular shear failure modes. Results demonstrate pronounced mechanical degradation under high-temperature exposure, characterized by marked reductions in uniaxial compressive strength and elastic modulus, significant increases in peak axial strain, and a coherent transition from brittle to ductile post-peak behavior. Crack analysis reveals tensile-dominated failure throughout loading and identifies a critical transition near 500 °C: intergranular cracking prevails below this threshold, while intragranular cracking dominates above it due to intensified thermal mismatch stresses and micro-damage within mineral grains. Pre-existing flaws control macroscopic failure patterns, promoting through-going “X”-shaped conjugate shear bands in flawed specimens, in contrast to localized “V”-shaped failures in intact ones. Microscopically, deviatoric loading induces particle-scale anisotropy, triggering progressive grain rotation, slip, and contact network destabilization. Macroscopically, cracks initiate predominantly at flaw tips and propagate to form specimen-scale through-fractures. This research provides a systematic, multi-scale decoupling of thermal cracking evolution in fractured sandstone, establishing a robust predictive numerical framework for assessing thermomechanical stability in critical deep geological engineering applications, including enhanced geothermal systems and deep geological repositories for nuclear waste.