Three-dimensional Material Point Method Modeling of Runout Behavior of the Hongshiyan Landslide

Xiaorong Xu,Feng Jin,Qicheng Sun,Kenichi Soga,Gordon G. D. Zhou
DOI: https://doi.org/10.1139/cgj-2017-0638
IF: 3.6
2019-01-01
Canadian Geotechnical Journal
Abstract:This study presents a field-scale simulation of the Hongshiyan landslide in China. It uses an advanced numerical approach (material point method (MPM)) and a constitutive model (the Drucker–Prager model + μ(I) rheological relation) for the three-dimensional (3D) simulation. The performance of the developed MPM model is validated with laboratory-scale experimental data on granular collapse before being applied to field-scale analyses. ArcGIS data are used to create a 3D MPM model of the soil body with complicated geometry. Although the developed model can describe the multiple phases of granular flow, it focuses on the runout behavior of the landslide in this work. The landslide is assumed to have occurred suddenly due to an earthquake, and global sudden failure rather than progressive failure is modeled. The MPM simulation results match reasonably well with the measured post-earthquake topography (e.g., deposit height of about 120 m and stretch length of about 900 m in the river) and landslide duration of about 1 min. The velocity of the sliding mass increases rapidly during flow, especially in the first 20 s. The velocity profiles along the depth direction at different locations of the sliding body exhibit an exponential distribution similar to that of a Bagnold-type profile, indicating that the sliding body is fully mobilized. The rate-dependent dissipation parameter β used in the model significantly influences the runout behavior (e.g., flow speed, velocity distribution, and deposit shape).
What problem does this paper attempt to address?