Perpustakaan Digital ITB

Rock failure is usually controlled by tensile strength. Commonly used means to determine the tensile strength of rock include the direct tension test (DTT), the Brazilian test (BT) and three-point bending test (TPBT). However, previous laboratory tests showed a considerable difference in tensile strength measured by those three types of tests, even for the same rock type from the same rock core. In addition, the relative magnitudes among them were found to be indefinite. In this paper, the FEM-based three-dimensional numerical simulations of the DTT, BT and TPBT were performed, aiming to determine the difference and relative magnitudes of rock tensile strength measured via those three methods and to investigate the physical mechanisms lying behind. The simulation results demonstrated that the tensile strength derived from the TPBT is much higher than that determined from the DTT and BT, and that the tensile strength determined by the BT is different from that from the DTT. Through stress analysis and parametric study, we found it is the rock heterogeneity and size effect that contribute to the distinction in rock tensile strength determined from different testing methods. The rock heterogeneity results in an overestimation of rock tensile strength obtained from the BT and TPBT due to the additional energy consumption in weak rock and defects in the sample, which does not contribute to the formation of macro fracture. We found that rock tensile strength determined in the TPBT increases with decreasing beam span because of the stress concentration above the supporting. The size effect leads to underestimation of the rock tensile strength determined in the BT, because the non-uniform distribution of the tensile stress along the centerline of rock disc. In addition, with increasing thickness of rock disc, the derived Brazilian strength decreases.