Stress relaxation tests on brittle rocks have shown that the stress drop associated with the relaxation period is related to the propagation of internal cracks but the fracturing processes that occur when the stress is relaxed is not well understood and requires further investigation. This study focuses on the analysis and interpretation of the micro-cracking processes that occur in rocks under different levels of constant axial strain through systematic stress relaxation experiments. Prismatic Barre granite specimens with two pre-existing flaws were subjected to unconfined compression and the temporal evolution of acoustic emission (AE) were evaluated. Two-dimensional digital image correlation (2D-DIC) was also used to evaluate the accumulated in-elastic strain and damage in the specimens. The image-based strain profiles along with the AE moment tensor inversion were used to identify the mode of deformation in the relaxation stages. As the strain level for relaxation increased, the stress dropped and correspondingly the number of AE events increased. The temporal analysis of the different cracking processes detected through AE and DIC revealed the higher contribution of tensile cracks in relaxation experiments.
Time-dependent behavior of rocks can be described as a phenomenon caused by the weakening of rocks with time, such as creep and stress relaxation. Detailed knowledge of the time-dependent behavior of rocks is important to predict the long-term stability of the underground structures such as mines, tunnels (Diederichs and Kaiser, 1999) and nuclear waste repositories (Nara et al., 2010). With this goal in mind, many experimental and numerical studies have been conducted to study the creep behavior of rocks under uniaxial, biaxial, and triaxial loading conditions (Hudson and Harrison, 1997; Zhang et al., 2013; Özsen et al. 2014; Mishra and Verma 2015; Liu et al., 2015; Yu et al. 2020). However, there have been very few studies in the past on the relaxation behavior of brittle rocks (Hudson and Harrison, 1997; Paraskevopoulou et al. 2017; Wang et al., 2019).