Multi-Phase Soil Modeling for Refinement of Slope Stability Analysis

Abstract

Failure of slopes is often hydrodynamic in nature. Most of the failures are due to water infiltration and movement in the matrix of a relatively dry soil mass. In such cases, the presence of air in the soil voids has to be taken into consideration and suitably accounted for analyzing the mechanism of slope failure. The conservative approach of assuming soil to be a two-phase material may often produce results that may err in the prediction of a suitable time frame of occurrence of failure. Hence, multi-phase soil modeling presents a novel approach for prediction, analysis and possible prevention of slope failures. In the present study, the mesh-free Discrete Element Method has been employed for understanding soil behavior. By defining the soil as a collection of discrete solid particles in a matrix of pore air and pore water, the soil behavior will be a cumulative response of its constituent components. Various particle sizes (variation of ±35% from mean particle diameter) and packing have been considered for analysis. An appropriate capillary contact law has been adopted to account for unsaturated conditions. Results in the form of force networks, strain deviators, evolution of different energy components have been extracted and interpreted and their usefulness in predicting slope failure has been highlighted. Findings show that suction has a significant effect on the behavior of soil at the micro scale by manifesting as an additional cohesion and arresting the deformations due to applied load, thereby contributing to the stability of the soil mass. The influence of packing arrangement of particles has been discerned. Any changes to the arrangement of the soil particles on a micro level shall disturb the stability and the possibility of attainment of unstable conditions shall be enhanced by external loading. The insights on the patterns of energy (plastic dissipation energy, elastic strain energy and kinetic energy) distribution and evolution emphasize on the advantage of tracking and monitoring energy components for management of vulnerable slopes. The variations in kinetic energy can be an indicator of system instability. Apart from the conventional view of stresses, strains and displacements, the current study presents a microscopic energy-based particle-scale technique as a novel, alternative approach towards stability of unsaturated slopes for enhancing the capabilities of predictions of slope failures