True-triaxial simulation of sandstone with full range of ₂ based on the Rigid-Body-Spring method

Accurately simulating the mechanical properties of rock under true-triaxial stress conditions (sigma(1) > sigma(2) > sigma(3)) is crucial for understanding deep underground rock masses. Although discrete approaches have advantages in describing rock fracturing behavior, further development seems necessary for quantitatively simulating the intermediate principal stress sigma(2) effects on rock strength. In this paper, we study the failure mechanisms of microscopic contacts of a discrete approach, the modified Rigid-Body-Spring Method (mRBSM), in response to sigma(2). Numerical results indicate that the model calibrated by axisymmetric-triaxial tests underestimates the strength of Yunnan sandstone when the stress state approaches sigma(2) = sigma(1). To address this discrepancy, we propose a new microscopic contact failure criterion that takes into consideration not only the normal and shear stresses on the contact but also the average Cauchy stress tensor around the contact. Results show that the mRBSM with the new failure criterion is able to reproduce the true-triaxial strength of Yunnan sandstone more accurately than the original model, especially under high sigma(2) and sigma(3) conditions. In addition, the corresponding stress-strain curves and failure modes under the influence of sigma(2) still show reasonable trends under true-triaxial stress conditions. This study indicates that when simulating rock failure under three-dimensional stress conditions using the rigid-body-spring system, the contact failure criterion might be influenced by the local stress state.