LBM-DEM simulation and flow regime analysis of two-phase geophysical flows

Abstract

For large-scale geophysical flows, such as debris flows and submarine landslides, the flow velocity and the runout distance are the critical parameters for risk assessment and hazard control. Such catastrophic events usually involve the fast transport of densely packed granular materials in an ambient fluid. Many previous studies focus on statistical correlations between the runout distance and the source volume or the fall height of the initial mass. However, some geophysical flows with similar sizes may run faster and longer than others, which, in giant submarine landslides, could be the origin of potentially devastating tsunamis. One possible reason behind these different behaviors could be the pore pressure effect when the complex fluid-particle interactions play a non-negligible role. In this regard, the pore pressure effect on the granular flow dynamics and its influence on the runout scaling are yet to be fully understood.

Geophysical flows usually have transient dynamics following the sequence of initiation, runout, and deposition. At different stages, excess pore pressure develops due to the dilatancy of granular materials and dissipates through the pore space. In this thesis, a high resolution three-dimensional numerical model coupling the lattice Boltzmann method (LBM) and the discrete element method (DEM) is employed to accurately capture the pore fluid flow and its feedback on the mechanical behavior of the granular material. The LBM-DEM model is first validated against benchmark cases, from which a guideline for the simulation of densely packed granular flows in an ambient fluid is proposed. The LBM-DEM model is applied to simulate immersed granular column collapses, which are typical model cases for geophysical flows, with various packing densities to investigate the influence of pore pressure on flow mobility. Supplementary experiments are performed to verify the role of ambient fluid and its interplay with the column size.

According to the numerical results, dense cases dilate and induce negative excess pore pressure. A strong arborescent contact force network is formed to prevent particles from sliding, resulting in a delayed collapse. By contrast, loose cases contract and induce positive excess pore pressure. The granular phase is liquefied substantially, leading to a rapid collapse. The influence of pore pressure is enhanced with the increasing column size.

Depending on the dominant resistance, granular flows can be classified into different regimes. The dense and loose granular collapses show similar dynamics to flows with high viscous and inertial effects, correspondingly. In loose cases with high inertia, hydroplaning of the granular front is observed, which significantly reduces the basal friction and promotes a longer runout distance. The influence of hydroplaning is quantified by the densimetric Froude number, which is linearly correlated to the normalized runout distance.

This study provides fundamental cornerstones to two future works. First, unified scaling laws for the prediction of runout distance can be proposed once various flows are classified into proper regimes with similar dynamics. Second, the micro-mechanical data from coupled simulations provide essential information for the development of constitutive models, which contributes to the continuum simulation of large-scale geophysical flows.