No abstract

No abstract

Several tailings impoundments have failed as a result of earthquakes in the last few decades. A majority of these failures were due to direct seismic loading, tailings liquefaction during shaking, or the post-seismic behavior of the tailings as it relates to the dissipation of excess pore water pressures that were generated during shaking. Previous work has indicated that the UBCSAND model (Byrne et al. 1995) is capable of simulating the cyclic simple shear testing response of low plasticity tailings over a range of consolidation stresses and cyclic shear ratios. However, the ability of the model to simulate the dynamic and subsequent behavior of such tailings for other conditions, such those induced by shakings table tests, has not yet been evaluated. In this regard, previous work has shown that the main components of the UBCSAND model cannot realistically simulate some specific responses, including the post-seismic volumetric strains related to excess pore water pressure dissipation. This paper presents numerical modeling results of the dynamic behavior of tailings from hard rock mines. It introduces a method for simulating their postseismic behavior by including an updating scheme for the elastic moduli into the UBCSAND model. The results of cyclic simple shear testing, seismic table testing and complementary experimental relationships were used to calibrate and validate the model with its new component.The simulated response of tailings during cyclic simple shear testing and for a complete seismic table test shows that the proposed approach simulates the experimental observations well. Level ground, seismically-induced liquefaction and post-seismic behavior of a 15-m-thick tailings deposit are also simulated, leading to post-liquefaction settlements that are in agreement with empirical relationships. résidus d'une épaisseur de 15 m soumis à un séisme mène à des tassements post-liquéfaction qui montrent une bonne corrélation avec ceux obtenus de relations empiriques.

Over the years, seismic activity has been a relatively common cause of tailings impoundment failure. The flow of liquefied tailings from such ruptures can result in very severe consequences, including loss of life and environmental damage. A co-disposal technique consisting of placing waste rock inclusions in tailings impoundments prior to and during tailings deposition was proposed by the authors. The waste rock is placed to create continuous inclusions within the impoundment, which provide a number of environmental and geotechnical benefits, particularly with respect to seismic stability. The results of numerical simulations previously performed have shown that the UBCSAND model can predict the seismic response of tailings. The UBCSAND constitutive model was used to conduct simulations to evaluate of the use of waste rock inclusions to improve the seismic stability of a tailings impoundment. The evaluation consists of numerical analyses of an actual tailings impoundment as constructed (without inclusions), and then assuming that it was constructed with inclusions, subjected to earthquake loads of various energy contents and with different predominant frequencies. The analyses were conducted in static, seismic, and post-shaking phases. The displacement of the surface of downstream slope of the tailings dyke was recorded during the analyses. The results indicate that the presence of waste rock inclusions can significantly improve the seismic behavior of the impoundment by reducing the displacements of the surface of the downstream slope and the extent of potential failure zones. Also, the results show that in most cases, the influence of a low-frequency earthquake on the displacement of the downstream slope of the tailings dyke is more important than that of a high-frequency earthquake. The performances of the tailings impoundment with different configurations of waste rock inclusions (varying width and center-to-center spacing) were classified based on the average normalized horizontal displacement of the downstream slope (ARx) for a range input ground motions. Charts were then developed to show how ARx is influenced by the total width of inclusions, their spacing, and the input ground motions.