Linking inherent anisotropy with liquefaction phenomena of granular materials by means of DEM analysis

Abstract

The liquefaction phenomena of sands have been studied by many researchers to date. Laboratory element tests have revealed key factors that govern liquefaction phenomena, such as relative density, particle size distribution, and grain shape. However, challenges remain in quantifying inherent anisotropy and in evaluating its impact on liquefaction phenomena. This contribution explores the effect of inherent anisotropy on the mechanical response of granular materials using the discrete element method. Samples composed of spherical particles are prepared which have approximately the same void ratio and mean coordination number (CN), but varying degrees of inherent anisotropy in terms of contact normals. Their mechanical responses are compared under drained and undrained triaxial monotonic loading as well as under undrained cyclic loading. The simulation results reveal that cyclic instability followed by liquefaction can be observed for loose samples having a large degree of inherent anisotropy. Since a sample having initial anisotropy tends to deform more in its weaker direction, leading to lower liquefaction resistance, a sample having an isotropic fabric potentially exhibits the greatest liquefaction resistance. Moreover, the effective stress path during undrained cyclic loading is found to follow the instability and failure lines observed for static liquefaction under undrained monotonic loading. From a micromechanical perspective, the recovery of effective stress during liquefaction can be observed when a threshold CN develops along with the evolving induced anisotropy. Realising that the conventional index of the anisotropic degree (a) is not effective when the CN drops to almost zero during cyclic liquefaction, this contribution proposes an alternative index, effective anisotropy (a×CN), with which the evolution of induced anisotropy can be tracked effectively, and common upper and lower bounds can be defined for both undrained monotonic and cyclic loading tests.

Keywords

• Liquefaction
• Discrete element method
• Anisotropy
• Fabric
• Triaxial tests