Time-Dependent Behaviour of Brittle Rocks Based on Static Load Laboratory Tests

Cumulative elastic and inelastic strain and associated internal stress changes as well as damage evolution over time in brittle rocks control the long-term evolution of the rockmass around underground openings or the land surface settlement. This long-term behaviour is associated with time-dependent deformation and is commonly investigated under static load (creep) conditions in laboratory scale. In this study, low Jurassic and Cobourg limestone samples were tested at different static load levels in unconfined conditions to examine the time to failure. Comparisons are made with longterm testing data in granites and limestones associated with the Canadian nuclear waste program and other data from the literature. Failure typically occurred within the time limits of the test program (4 months) with axial (differential) stress levels near or above the crack damage threshold (CD) estimated from baseline testing. The results also suggest that the time to failure of limestone is longer than that of granite at a given driving stress. Further insight into samples that did not reach failure was investigated and it was found that there was a clear division between failure and no failure samples based on the Maxwell viscosity of the samples tested (indicating that viscosity changes near the yield threshold of these rocks. Furthermore, samples showed a clear tendency towards failure within minutes to hours when loaded above CD and no failure was shown for samples loaded below CI (crack initiation threshold). Samples loaded between CD and CI show a region of uncertainty, with some failing and other not at similar driving stress-ratios. Although such testing is demanding in terms of setup, control of conditions, continuous utilization of test and data acquisition equipment and data processing, it yields important information about the long-term behaviour of brittle rocks, such as the expect time to failure and the visco-elastic behaviour. The information presented in this paper can be utilized for preliminary numerical studies to gain an understanding of potential impact of long-term deformations.