SUMMARYIt is well established that the mechanical behavior of granular media is strongly influenced by the media's microstructure. In this work, the influence of the microstructure is studied by integrating advances in the areas of geostatistics and computational plasticity, by spatially varying the porosity on samples of sand. In particular, geostatistical tools are used to characterize and simulate random porosity fields that are then fed into a nonlinear finite element model. The underlying effective mechanical response of the granular medium is governed by a newly developed elastoplastic model for sands, which readily incorporates spatial variability in the porosity field at the meso-scale. The objective of this study is to assess the influence of heterogeneities in the porosity field on the stability of sand samples. One hundred and fifty isotropic and anisotropic samples of dense sand are failed under plane-strain compression tests using Monte Carlo techniques. Results from parametric studies indicate that the axial strength of a specimen is affected by both the degree and orientation of anisotropy in heterogeneous porosity values with anisotropy orientation having a dominant effect, especially when the bands of high porosity are aligned with the natural orientation of shear banding in the specimen.

A nonlinear ground response analysis is conducted for the Niigata-ken Chuetsu-oki earthquake recorded at a free-field vertical array near the Kashiwazaki-Kariwa Nuclear Power Plant in Japan. A bidirectional site response analysis is carried out using LS-DYNA which allows user defined stress-strain relationships to dictate soil behavior subjected to dynamic loading. Dynamic soil behavior is characterized using a two-stage hyperbolic backbone curve implemented with modifications to consider the peak strength of soil layers as well as the strain at which the peak strength is fully mobilized. The effects of bidirectional input motions, strain rate, and the shape of the shear modulus degradation curves are investigated, and it is demonstrated that each factor can have a significant influence on the results.

In this technical note, an evaluation of the robustness and predictive ability of a constitutive model for sands is performed. The model is shown to capture the main features of sand behavior under both drained and undrained monotonic loadings for a wide range of relative densities and stress paths. The main contribution of this technical note is to evaluate a robust, yet simple, constitutive framework based on a solid theoretical basis that fulfils the most fundamental requirement of any useful constitutive law: accurate predictions.

A complete formulation of the BRICK soil model in general strain space is presented herein for the first time. Like all elasto-plastic constitutive models, BRICK exhibits some anisotropic behaviour, owing to the development of plastic strains once its yield surfaces are engaged. However, an abundance of laboratory and field evidence demonstrates that stiffness anisotropy is also significant within the elastic domain. Because of the inseparable nature of strength and stiffness in BRICK, the simple use of an anisotropic elastic stiffness matrix would result in an unrealistically high degree of strength anisotropy. Therefore, in addition to the established BRICK formulation, this paper also presents a novel framework to introduce stiffness anisotropy by transforming the coordinate system in which the model is based. The transformed coordinate system evolves to enable a constant-volume condition during shearing at critical state, reflecting the reorganisation of the soil fabric. The superior performance of the enhanced BRICK model over the classic model is demonstrated by a variety of conventional and non-conventional laboratory tests on London Clay.

Recent criteria have been developed to describe the onset of static liquefaction in constitutive models. This paper expands the theory to a finite-element framework in order to predict potentially unstable regions in granular soils at the engineering scale. Example simulations are presented for two plane strain tests and a submarine slope to demonstrate the applicability of a proposed liquefaction criterion to boundary value problems. In addition, loading rate and mesh size effects on the liquefaction prediction are examined. The methodology presented herein shows promise as a means of predicting soil liquefaction based on solid mechanical theory rather than empiricism.