A numerical investigation of the seismic response of tailings impoundments reinforced with waste rock inclusions.

摘要

Aubertin et al. (2002b) considered placing waste rock in tailings impoundments in tandem with tailings deposition to improve the environmental and geotechnical performance of the impoundments. The waste rock would be placed to create relatively narrow, continuous inclusions along planned routes in the impoundment. It was postulated that these waste rock inclusions would provide a number of benefits, particularly by accelerating the consolidation of the tailings and acting to reinforce the impoundment with respect to static and seismic loads.;James (2009) conducted a series of cyclic simple shear tests on tailings over a range of confining stresses and cyclic stress ratios. Then, using numerical simulations, he showed that the UBCSAND model (Beaty and Byrne 1998) can predict well the cyclic behavior of tailings. James (2009) used numerical analyses to simulate the seismic response of a conceptual tailings impoundment with and without waste rock inclusions. It was shown that the use of such waste rock inclusions (WRI) could considerably improve the seismic response of an impoundment.;One of the cyclic simple shear tests performed by James (2009) on tailings was simulated using the UBCSAND model (version of 904aR). The results of the simulation, presented in this thesis, indicate that the model can represent well the shear strain, excess pore water pressure generation, and initiation of liquefaction under these conditions.;In the next step, the capability of the UBCSAND model, as implemented in FLAC (Itasca, 2008), to simulate the seismic behavior of tailings in seismic table tests was evaluated using tests conducted by Pepin et al. (2012a, 2012b). The two tests (Nos. 3 and 7) were conducted on lightly tamped tailings without and with a sand wall inclusion. The results of modeling showed that the UBCSAND model can predict fairly well the generation of excess pore water pressure in seismic table testing of tailings with and without an inclusion.;Tailings impoundments may fail after earthquake shaking due to the dissipation of excess pore water pressures. To model this phenomenon realistically, one can simulate post-shaking consolidation (excess pore water pressure dissipation). Two approaches were used to do so. In the first approach, a cap (Vermeer, 1980) was added to the UBCSAND model (version 904aR). The modified model was tested by simulating hypothetical isotropic compression tests, oedometer tests and level ground liquefaction of tailings. Reasonable results were obtained, but the UBCSAND model with a cap developed some mathematical instabilities (possibly due to a singularity on the yield surface). Therefore, another approach was used. In this second approach, the values of the elastic moduli of the tailings were updated in the model based on the consolidation curve of loose tailings. This approach combines the Sento et al. (2004) methods and the UBCSAND model to predict the post-shaking settlement of tailings due to pore water pressure dissipation.;The capability of this model to simulate the post-shaking settlement of level ground liquefaction in tailings was also compared with the empirical relationships of Wijewickreme and Sanin (2010) and the results were encouraging.;After verifying that the UBCSAND model can simulate the seismic and post-seismic behavior of tailings, the seismic response of a tailings impoundment (based loosely on an actual site located in New Brunswick), with and without waste rock inclusions was simulated numerically. The impoundment with various configurations of inclusions was subjected to earthquake loads of various energy contents and with different predominant frequencies. The results of the numerical analyses showed that the tailings impoundment without waste rock inclusion undergoes excessive deformation and failure under the seismic loads considered. The results of simulations with inclusions showed that WRI can significantly decrease the deformation of the downstream slope and prevent failure of the impoundment. Additionally, the response of the tailings impoundment was evaluated based on the average normalized horizontal displacements (ARx) of the downstream slope and of only the WRI. The performance of the tailings impoundment with different configurations of WRI (width and center-to-center spacing) was classified based on the ARx values at the end of shaking. Graphs that show the ARx (total and WRI) values as a function of the total width of different configurations are presented. These graphs can be used to obtain (preliminary) optimum configurations of WRI for the seismic loads considered. The results also indicate that in most cases, the low frequency ground motions produce significantly more deformation than the high frequency ground motions.;The effect of WRI on the seismic response of a tailings impoundment was further evaluated using numerical simulations of another tailings impoundment. The seismic response of the tailings impoundment was again evaluated using ARx (total and WRI) and the critically displaced volume of tailings at the end of shaking, for different configurations of the inclusions. (Abstract shortened by ProQuest.).