Fully coupled flow-liquefaction in tailings storage facilities: induced shear bandwidth on partially drained triggers

Abstract

Flow liquefaction remains one of the most critical threats to the stability of Tailings Storage Facilities (TSFs). Yet, its reliable prediction has long been hindered by mesh-dependent localization in numerical models. This study demonstrates that fully coupled hydro-mechanical formulations overcome this limitation by embedding pore-pressure diffusion directly into the governing equations. Unlike uncoupled or purely mechanical approaches, the coupled system introduces a physically consistent internal length scale that suppresses pathological mesh sensitivity and stabilizes strain localization during undrained or partially drained softening. We use spectral analysis to deduce a dispersion relation that governs perturbation growth, showing that stability requires short-wave modes to be damped by increments exceeding a critical threshold. Therefore, the resulting localization bandwidth scales with permeability, fluid and skeleton compressibility, and the time step increment expressed as a function of the load increment and its rate of application. Our numerical simulations, from axisymmetric triaxial tests to TSF-scale scenarios, confirm these predictions: large time steps suppress spurious mesh-aligned bands, while permeability reduction - or skeleton compressibility increases - narrows shear zones and lowers triggering loads until the undrained limit is reached. Load-rate effects are likewise clarified, vanishing beyond a threshold where the system behaves as fully undrained. This work bridges analytical derivations, laboratory-scale validations, and TSF-scale modeling, establishing fully coupled formulations as a rigorous and predictive framework for assessing partially drained liquefaction triggers. The results advance both the theoretical understanding of strain localization and the practical reliability of tailings dam safety analyses.