Fluid-Dynamics Based Analysis of Landslides, Debris Flow, and
Inclusive Dates: 10/01/09 Current
Background - Landslides and debris flows are severe natural disasters that are a source of societal hazard throughout the world. The occurrence of a landslide depends on a number of factors including bedrock geology, geotechnical properties of surface materials, rheology and soil conditions. The large spectrum of landslide phenomena makes it difficult to define a single standard technique to evaluate landslide hazards and risk. However, a detailed analysis of the relationship between landslides and their various causes not only provides insight into landslide mechanisms, but also can form a basis for predicting the occurrence of future landslides and assessing landslide hazards. A similar class of natural geotechnical hazard is the liquefaction of loose, saturated, cohesion-less soils and other granular materials during large-magnitude earthquakes. Lateral spreading induced by seismic liquefaction causes large ground displacements and shear strains that can cause extensive damage and disruption to pile foundations of buildings and bridges, embankments, river dikes, pipelines and waterfront structures. Most prior analyses of debris and landslide flow and liquefaction-induced lateral spreading employed numerical techniques such as the finite element method (FEM) and discrete element method (DEM). Compared to FEM and DEM, a mesh-free computing method such as smoothed particle hydrodynamics (SPH) provides a significant advantage in handling large deformation and postfailure analyses.
Approach - The proposed work uses a SPH-based computational framework for predicting the size, shape, and runout length of debris flows, landslide flows and liquefaction-induced lateral ground displacement. The computational framework is supported by limited experiments related to the rheology of the debris material and liquefied soil. The computational analysis is primarily based on SPH; however, supporting computations are made with the Navier-Stokes volume of fluids (N-S VOF) to validate the SPH computations. Simulations are carried out for cases where experimental data and analog site data are available. The broad overall objective of the research project is to establish a generic SPH-based computational framework capable of solving problems in geomechanics that involve both small and large deformations. The SPH approach provides a significant advantage over existing numerical models because it is a mesh-free computing method and the rheological characteristics are amenable to changes in the flow characteristics, geometry and material properties for flows with free surface deformation.
Accomplishments - Several simulations for mud type flows were carried out using the Herschel Bulkley, Bingham-Plastic power law models. These models were implemented in the SPHYSICS code and validated. These are the classical non-Newtonian models used in the simulations. The implemented models were validated against problems involving non-Newtonian fluids that have experimental data. These simulations highlighted both the importance and complexity of rheology, simulation parameters, and geometry on the predicted flowfield. The simulation results with the non-Newtonian matched well with experimental data. These include a dam break problem that involved the experimental data for the collapse of a water-clay mixture. The project team is now working on experiments related to the viscosity and non-Newtonian rheology of landslide and debris materials. In addition, the team has initiated preliminary work on the numerical analysis of the time-dependent lateral spreading of liquefied soil. Different rheological models also were compared.