The pre-peak cyclic mechanical shear properties and damage characteristics of rock joints are of great significance for the safety of underground engineering. This study employed granite specimens with average mechanical properties, including a uniaxial compressive strength of 121.12 MPa and elastic modulus of 37.27 GPa. The research reconstructed natural rough joints via 3D laser scanning and engraving, integrating digital image correlation (DIC) and acoustic emission (AE) techniques to synchronously monitor deformation fields and microcrack activity. Pre-cyclic shear tests were conducted on both smooth and rough joints under controlled variables, including normal stress, loading frequency, static/dynamic amplitude, and cycle number, with a loading rate of 0.4–40 kN/s (strain rate of 10⁻ 3 –10⁻ 2 /s). Machine learning-based feature analysis (XGBoost, Gini value) identified normal stress as the predominant governing factor. Experimental results demonstrate that increased normal stress significantly enhances shear strength, average stiffness, and damping ratio while suppressing cumulative displacement. AE parameters show characteristic patterns, where count and energy rates peak within initial 10 cycles followed by rapid decay. For both joint types, elevated normal stress conditions lead to increased cumulative AE counts and energies, along with rising variations in b values and declining S values. Local strains show that the thickness of rock joints increases as the normal stress raises. Notably, rough joints formed discrete clusters under dominant principal strain, while smooth joints exhibited continuous shear bands with uniform strain distribution. This study reveals the mechanical shear properties and damage accumulation mechanism of rock joints under pre-peak cyclic shearing, providing experimental evidence for the stability assessment of rock engineering.
Xie et al. (Mon,) studied this question.