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

Flow liquefaction remains one of the most critical threats to the stability of Tailings Storage Facilities (TSFs) and its reliable prediction has long been hindered by a fundamental numerical pathology: mesh-dependent localization in computational models. This study demonstrates that fully coupled hydro-mechanical formulations overcome this limitation by embedding pore-pressure diffusion directly into the governing equations, thereby introducing a physically consistent internal length scale absent in uncoupled or purely mechanical approaches. Our spectral analysis derives a dispersion relation governing perturbation growth in the coupled system. Stability emerges when short-wave modes are sufficiently damped—a condition requiring time step increments to exceed a critical threshold. The resulting localization bandwidth is not arbitrary but scales systematically with material and loading parameters: permeability, fluid and skeleton compressibility, and the time step increment expressed as a function of load increment and its rate of application. Thus, the coupled formulation transforms an ill-posed mesh-dependent problem into a well-regularized one via an intrinsic diffusion length scale. Our numerical simulations span scales from axisymmetric triaxial tests to full TSF configurations, systematically validating our theoretical predictions. Large enough time steps suppress spurious mesh-aligned bands that plague uncoupled models. Reducing permeability—or increasing skeleton compressibility—narrows shear zones and lowers the triggering load for localization, eventually converging to the undrained limit. Load-rate effects follow a predictable pattern: at sufficiently high rates, the system transitions to undrained behavior regardless of permeability, as pore-pressure diffusion becomes negligible compared to loading timescales. We bridge analytical derivations, laboratory-scale validations, and engineering-scale modeling to establish that fully coupled formulations not merely as a numerical convenience but as a rigorous and predictive framework for assessing partially drained liquefaction triggers in tailings dams. We advance both the theoretical understanding of strain localization in fluid-saturated porous media and the practical reliability of safety analyses for critical infrastructure. We clarify the physical mechanisms that control shear band formation under partially drained conditions to provide engineers with a principled basis for evaluating liquefaction susceptibility in TSFs.