The assessment of tailings dam vulnerability to flow liquefaction has evolved from deterministic factor-of-safety approaches to sophisticated numerical deformation analyses.
Following the proposal introduced by the authors at PCSMGE 2024, this study refines the framework by integrating recent advancements in constitutive modelling, Eurocode 7 (Second Generation) reliability principles, and performance-based design concepts.
The methodology aims at replacing traditional trigger analyses with a structured vulnerability assessment framework. It prescribes a uniform set of deteriorating actions—crest loading, toe contraction, and phreatic surface rise—regardless of site-specific conditions. These actions are applied systematically in pushover-style analyses to quantify the dam’s susceptibility to flow liquefaction, moving towards a limit state-driven design philosophy akin to the Reliability Based Design approach of Eurocode 7.
The methodology further introduces the construction of an Ultimate Limit State (ULS) surface, derived from multiple numerical simulations, and interpolated using Radial Basis Functions (RBF). A fuzzy transition zone is incorporated to account for modelling uncertainty, enabling the computation of a continuous probability of failure (PoF) across the action space using a Gaussian cumulative distribution.
The framework includes the use of empirical or parametric probability density functions (PDFs) for each trigger, combined through Latin Hypercube Sampling in a Monte Carlo simulation to estimate annual PoF. By standardising vulnerability assessments, this methodology facilitates more objective risk management, improving dam resilience and regulatory compliance.
The paper concludes with a case study illustrating the use of this procedure in a hypothetical tailings storage facility.
First presented at Life of Mine | Mine Waste and Tailings Conference 2025 and available in the conference paper proceedings (Australasian Institute of Mining and Metallurgy).
