Some considerations when preparing thickened tailings for shear strength testing in the laboratory from a slurry

Abstract

Preparation of tailings in the laboratory to simulate the likely range of densities and resulting behaviour that will occur under load in situ is challenging. A number of issues of relevance include whether effects such as subaqueous deposition can be reproduced at all in the laboratory, and the potential importance of in situ layering of segregating tailings. These issues may have (thus far, largely unquantified) effects on mechanical behaviour in ways that are difficult to predict with element tests that are usually intentionally prepared homogeneously. Thickened tailings testing is conceptually easier, as issues such as segregation and the potential for subaqueous deposition are often not relevant. Despite this relative simplicity, a number of important considerations remain. While it is trivial to pour a thickened slurry into various forms of moulds or vessels to prepare for element testing, questions remain over whether the density that is reproduced will be relevant to in situ conditions. In particular, preparation of triaxial samples from thick slurries is particularly challenging as a number of steps are required to enable such samples to be ‘free-standing’, with each one of these steps potentially leading to slight disturbance and thus densification of the sample. Where such densification occurs, it would result in the element test results being non-conservative. This could have important implications with respect to expectations of the contractive (and potentially liquefiable) or dilative response of the tailings in situ. To investigate these issues, a series of slurry-deposited triaxial tests was carried out using a non-segregating slurry. As test methods were refined during the program, the amount of disturbance applied to the specimen was reduced. However, comparison of the triaxial tests to slurry consolidometer tests indicated that, regardless of efforts made, the triaxial tests achieved denser states at a given amount of consolidation stress. This was found to be the case using any conceivable range of assumed geostatic stress ratio to interpret the slurry consolidometer results. This outcome is speculated to be a result of the quiescent conditions used in the preparation of a slurry consolidometer specimen, which only requires pouring and then application of vertical load, first using weights, then a load frame. The implications of the increased density seen in triaxial tests compared to the likely more realistic value seen from the slurry consolidometer are discussed. Alternative preparation methods to target this looser density range are briefly discussed.