A computational periporomechanics model for localized failure in unsaturated porous media

This article is devoted to a computational plastic periporomechanics model for localized failure in unsaturated porous media assuming passive atmospheric pressure. The computational periporomechanics model is formulated based on the hypothesis that unsaturated porous media can be represented by a collection of two-phase material points with infinitesimal volumes interacting with each other within finite distance. The effective force state concept for solid skeleton deformation and the peridynamic state concept for unsaturated fluid flow are utilized in the coupled integro-differential field equations that have an intrinsic length scale. The unsaturated fluid flow state and effective force state are determined by the fluid potential state and the deformation state of the solid skeleton, respectively. The periporomechanics model is numerically implemented by an implicit two-phase meshfree method in which the message passing interface technique is used for parallel computing. The numerical implementation is validated by simulating classical coupled poromechanics problems and comparing numerical results with analytical solutions and experimental data. Numerical examples under different scenarios are conducted to demonstrate the robustness of the implemented computational periporomechanics model for simulating localized failure in unsaturated porous media as a coupled solid deformation and unsaturated fluid flow problem. (C) 2021 Elsevier B.V. All rights reserved.

Automated summary

This article presents a computational plastic periporomechanics model for localized failure in unsaturated porous media assuming passive atmospheric pressure. The model is based on the hypothesis that unsaturated porous media can be represented by a collection of two-phase material points with infinitesimal volumes interacting within a finite distance. The numerical implementation is validated by simulating classical coupled poromechanics problems and demonstrating the robustness of the model under different scenarios.