The objective of this research is to investigate potential triggers and failure mechanisms of flow and sliding failures of tailings dams and loose fill slopes. This research consists of three components. The first component is the investigation of triggers and failure mechanisms of tailings dams. Two model tailings dams were constructed from gold tailings using the moist tamping technique. The model tailings dams were accelerated using a geotechnical centrifuge and subjected to a rising groundwater table using a viscous pore fluid until failure was triggered. Two triggers and failure mechanisms were identified. In the first centrifuge test, a retrogressive flow slide failure was triggered by large seepage forces at the toe of the slope. The large seepage forces caused sloughing at the t...[ Read more ]

The objective of this research is to investigate potential triggers and failure mechanisms of flow and sliding failures of tailings dams and loose fill slopes. This research consists of three components. The first component is the investigation of triggers and failure mechanisms of tailings dams. Two model tailings dams were constructed from gold tailings using the moist tamping technique. The model tailings dams were accelerated using a geotechnical centrifuge and subjected to a rising groundwater table using a viscous pore fluid until failure was triggered. Two triggers and failure mechanisms were identified. In the first centrifuge test, a retrogressive flow slide failure was triggered by large seepage forces at the toe of the slope. The large seepage forces caused sloughing at the toe, which ultimately resulted in a rapid retrogressive failure. The rapid rate of shearing caused the generation of significant excess pore pressures within the toe region of the slope, with sufficiently large positive excess pore pressures to trigger localised liquefaction. The localised liquefaction resulted in the tailings at the toe of the slope not being able to support the tailings upstream of the toe, which triggered the retrogressive failure. In the second centrifuge test a drained slip failure triggered at least partially undrained shearing of the soil within the slope, which transitioned into a slide-to-flow failure. The rate of shearing was reduced compared to the first test, which resulted in smaller positive excess pore pressures being generated within the slope. The positive excess pore pressures generated were smaller compared to the first test due to the slower rate of shearing. The slide-to-flow failure was predicted using the finite element method. During the numerical back-analysis of the slide-to-flow failure, similar computed excess pore pressures were generated within the model slope when compared with the measured excess pore pressures. Thus, the slope only experienced instability. This confirms that a drained slip failure triggered by large seepage forces caused the slide-to-flow failure observed in the centrifuge test.

Buttresses constructed at the toe of the model tailings dams were illustrated to improve the stability of tailings dams. However, if a buttress is too small, a slide-to-flow failure can still be triggered in the slope. In addition, if a buttress extends too far upstream of the tailings dam, it can also trigger a slide-to-flow failure, as a section of the buttress acts as a destabilising weight. A design method for buttresses for tailings dams was developed. An optimum height of the buttress can be determined by identifying the centre of rotation of the critical slip surface and designing the buttress to be constructed on the downstream side of the centre of rotation.

In the final component of this study a numerical back-analysis of a previously conducted centrifuge test on loose fill slopes was conducted. It was previously hypothesised that a drained surface failure was triggered by large seepage forces. The drained surface failure triggered the generation of significant positive excess pore pressures within the slope, triggering a degree of localised static liquefaction of the soil within the slope. The numerical back-analysis confirmed the hypothesised triggering mechanism. During the strength reduction phase of the numerical back-analysis a drained slip failure caused the generation of significant positive excess pore pressures within the slope. Localised liquefaction was triggered resulting in the failure of the entire slope. Soil nails were also illustrated to be able to prevent the static liquefaction failure of the loose sand fill slope. Soil nailing can prevent the peak shear strength of the soil from being mobilised. If the peak shear strength is not mobilised, strain softening cannot be triggered in the soil. Thus, soil nailing limits the excessive strains required to trigger static liquefaction.