Study Of Debris Flow Behavior PDF Download

ABSTRACT Rapid flow-like landslides occur commonly in different parts of the world and present a serious threat to communities and infrastructures, especially in mountainous areas. Due to their high flow velocities, large impact forces, long-runout distance, and poor temporal predictability, flow-like landslides are among the major geohazards. Human safety and infrastructures can be endangered by large impact forces in unexpected flow-like landslide events. Hence, reliable estimates of the runout of flow-like landslides and their resultant impact force on structures are essential for both hazard assessment and the design of protective measures. However, it's challenging to capture, analyze and interpret the behavior of flow-like landslides as real full-scale observations and experiments are often difficult or impossible to conduct because of dangers involved with in-situ field experimental studies, uncontrollable geophysical conditions, and unpredictable time and locations of landslide events. Alternatively, scaled experimental studies of rapid debris flows can provide fundamental and valuable insights to better understand the underlying physics of landslides or to develop and validate a robust computational framework for landslide modeling. Although the scaled laboratory and field experiments of debris flow are useful, they cannot be directly utilized for landslide prediction and designs. Hence, a numerical model capable of simulating the debris flow and its interaction with structures, and reliably estimating the debris flow impact force would be an ultimate alternative for landslide designs. Numerical modeling of debris flow, however, is exceptionally complex since debris flow is known to exhibit both solid-like and fluid-like behaviors from grain inertia to micro viscous material which are responsible for the complex flow dynamics of debris flow. Therefore, numerical models capable of simulating the debris flow are naturally complex. Currently, most of the numerical methods proposed for modeling flow-like landslides or debris flows are developed by means of simplified relationships which are based on the assumption that the granular material is to behave like an incompressible fluid. These single-phase equivalent fluid models rely on manipulation of the rheological properties of landslide mass and, hence, have limited predictive power, and are mainly useful for modeling landslides runout, propagation and spreading which are mostly related to the fluid-like behavior of the granular flow. Solid-like behaviors of a sliding mass, however, can be inferred from its impact forces on a rigid obstruction. Therefore, a physically-based numerical model is required which is capable of taking into account the non-linear behavior of the granular material considering both fluid-like and solid-like behaviors, suitable for modeling of large displacement and deformation of the granular mass and taking into account frictional soil-structure interactions. Such numerical models need to be validated through systematic experimental studies at the well-controlled laboratory scale. In this research, the behavior of a granular sliding mass down an incline from collapsing, to sliding, and to impacting a rigid obstruction was investigated. The main objectives of this research are: (1) to develop and validate a computational framework capable of addressing the following four main challenges in landslide modeling and granular flow namely: (A) suitable for modeling of the large deformation of the granular material, (B) capable of considering the frictional soil-structure interaction, (C) equipped with suitable constitutive models to capture the complex behavior of a granular sliding mass in a rapid fluid-like landslide, and (D) computationally efficient to be practical for medium to large-scale granular modeling; (2) to study the governing mechanisms of granular collapse on a frictional surface; and (3) to study the governing mechanisms of granular flow and investigate the transition between solid-like and fluid-like behavior in a granular flow and its impact force. These research objectives are achieved by developing an efficient parallel numerical model based on the Smoothed Particles Hydrodynamics (SPH) method. Two frictional boundary conditions in the framework of SPH method were developed in this study and implemented into the SPH computational tool and their performance was examined in collisional and sliding problems. An elasto-plastic model, a critical hypo-plastic model and a collisional Bagnold-type rheology constitutive model were implemented in the SPH computational tool to investigate the granular flow impact force and mass front velocity. Results of this study revealed that the granular flow is governed by a combination of sliding, collapsing, and spreading mechanisms, depending on the inclination angle, sliding distance, basal friction, and the material initial condition. It is shown that the elasto-plastic models can predict the granular behavior and impact force relatively well in sliding-dominant flows only. The critical hypo-plastic model can consider the effect of dispersive pressure and shearing on the material stiffness and energy dissipation, it shows better predictions of the impact forces compared to the elasto-plastic model, especially in spreading-dominant flows. It is observed that the unified critical hypo-plastic and collisional Bagnold-type model predicted more accurate results of the impact force than the critical hypo-plastic model did, as the unified model accounts for both fluid-like and solid-like behaviors within a sliding granular mass.