Large deformation failure analysis of the soil slope based on the material point method

Abstract

The failure analysis of the soil slope is a very important topic in the field of geomechanics. Being a fully Lagrangian particle method, the material point method (MPM) has distinct advantages in solving the extremely large deformation problem. For both cohesive and noncohesive soil slopes, the large deformation failure problems are analyzed using MPM and the Drucker-Prager constitutive model. For verification of the numerical method, the comparison between MPM and analytical solutions of the dam break problem is presented. Moreover, the numerical results by MPM are compared with the experimental results for the collapse of the aluminum-bar assemblage. Simulations reveal the cohesive soil slope under gravity has a shear band failure mode. Computational results show the reposed angle of non-cohesive soil slope is less than the internal friction angle, and the reason for this phenomenon is presented. The purpose of this study is to give a further understanding of the slope failure in different soil types and provide a computational tool for the failure analysis of soil slopes.

Fast and efficient MATLAB-based MPM solver (fMPMM-solver v1.0)

“…These numerical results demonstrate the solver to be in agreement with both previous experimental (Bui et al (2008)) and numerical results (Huang et al (2015)) and confirms its ability to solve elasto-plastic problems such as granular collapses https://doi.org/10.5194/gmd-2020-183 Preprint. Discussion started: 20 July 2020 c Author(s) 2020.…”

“…We compare our MPM solver with a non-associated plasticity based on a Drucker-Prager model with tension cutoff (Huang et al (2015)) to the experimental results of an elasto-plastic collapse of a material (e.g., an aluminium-bar assemblage, see Bui et al (2008)). The numerical implementation is detailed in Huang et al (2015), and we therefore suggest the interested reader to directly refer to their contribution since we do not describe the elasto-plastic model in this manuscript.…”

“…Our elasto-plastic MPM solver closely follows the implementation of Huang et al (2015), except our solver relies on an MUSL procedure, whereas Huang et al (2015) selected the USF procedure. The elasto-plastic problem is solved by i) an elastic trial, ii) corrected by a return mapping when the material yields either in shear or in tension or both, assuming a nonassociated (the dilation angle ψ = 0) perfectly plastic behaviour of the material.…”

“…The elasto-plastic problem is solved by i) an elastic trial, ii) corrected by a return mapping when the material yields either in shear or in tension or both, assuming a nonassociated (the dilation angle ψ = 0) perfectly plastic behaviour of the material. Since the MPM variant in Huang et al (2015) was not clearly stated, we use the uGIMP variant. The reason is the collapse results in extreme deformations, for which a domain update based on the deformation gradient systematically resulted in a failure during the simulation.…”

“…We present an application of the MPM solver to the case of landslide mechanics. Considering a CPDI2q variant coupled to an elasto-plastic constitutive model based on a Mohr-Coulomb (M-C) non-associated plasticity Simpson (2017), we i) analyse the geometrical features of the slump and ii) compare the results (the geometry and the failure surface) to the numerical simulation of Huang et al (2015), which is based on a Drucker-Prager model with tension cutoff (D-P).…”