Abstract
In the last decade, the use of filtered tailings has steadily increase in the mining industry in South America as it is an innovative solution for the construction of tailings storage facilities (TSF).
This technology allows increasing the safety of tailings dams since it eliminates the slurry pond from the impoundment. Moreover, filtered tailings enhances the recovery of water for mining processes and enables the application of co-disposal schemes.
Many TSFs are located in zones where major earthquakes may occur such as the countries located within the fire belt. Thus, the behavior of filtered tailings when subjected to large dynamic shear stresses must be assessed. The state-of-practice recommends the application of numerical modelling with advanced constitutive models capable of reproduce the behavior of tailings to assess the performance of TSFs against large earthquakes.
There are many models already implemented in commercial software; nevertheless, these constitutive models have disadvantages when dealing with dynamic loading. For instance, the widely used PM4Sand model can only model boundary problems in 2D conditions. The HS Small model does not take into account explicitly the theory of critical state soil mechanics. The NorSand model is not suited for dynamic problems. Hence, there is a need for another constitutive model that cover these gaps.
In this regard, the family of SANISAND constitutive models is a new option to be applied in mining. SANISAND-Sf was implemented in a finite difference software and applied in a real project located in zones where peak ground accelerations of 0.60g and higher are expected. This constitutive model is an improvement of the original SANISAND. The SANISAND-Sf includes a new set of parameters that allows reproducing the undrained cyclic behavior of granular soils when the effective confining stress is low and the material behaves in a semi-liquid state or semi-fluidized (SF).
