Size Effects on Crystalline Rocks Under Uniaxial Compression and Brazilian Tests

ABSTRACT The tensile strength of crystalline rocks in Brazilian tests (BT) shows a more significant size dependency than the uniaxial compression strength in uniaxial compression tests (UCT), while the micro-mechanism is still unclear. This study seeks to elucidate the micro-mechanism by comparing different microcracking processes under UCT and BT. We adopt a grain-based model in the discrete element method (DEM) to reproduce the interlocked microstructure of crystalline rocks and perform UCT and BT with different sizes ranging from 0.5 to 3 times the standard sample size. It was found that peak stresses in UCT are nearly insensitive to sizes, which is consistent with experimental results measured at very low loading rates. Due to the homogenous stress distribution, no clear cracking paths are observed; instead, randomly distributed microcracks appear at failure. In contrast, numerical results indicate that the tensile strength measured in BT decreases as size increases and continuous cracking paths are observed at failure. Larger specimens have a higher possibility to include weaker cracking paths, hence tensile strength shows stronger size dependency. The findings of this work offer microscopic explanations for different size effects observed in UCT and BT, which can bridge the gap between laboratory-scale strength data and field-scale applications. INTRODUCTION Rock engineering applications often require assessing the strength and failure characteristics of rock masses. These rock masses are composed of intact blocks across various scales. However, the strength and cracking mechanism of these field-scale intact blocks are usually indirectly measured and studied through laboratory-scale tests on intact specimens (Duan et al., 2017; Duan and Kwok, 2015; Fei et al., 2021; Mahabadi et al., 2014; Martin and Chandler, 1994; Wong et al., 1996). Uniaxial compression test (UCT) and Brazilian tests (BT) are standard tests that measure the uniaxial compression strength (σf) and tensile strength (σt) of rocks (Eberhardt et al., 1999; Li and Wong, 2013; Martin, 1994). The sizes of UCT and BT specimens in the laboratory usually vary from 5-10 cm in diameter, much smaller than in fields. However, a dependence of strength on size is widely known in rocks (Bažant, 1984; Choo et al., 2023; Paterson & Wong, 2005). The strength determined by these laboratory-scale tests may overestimate the loading capacity in fields, and the cracking process observed in the laboratory-scale specimens may also not accurately manifest the failure mechanism in the field scale. Therefore, to apply laboratory-scale experimental data to practical applications, it is vital to investigate the influence of specimen size on strength and cracking patterns.