Numerical modelling of the NGI-DSS test and cyclic threshold shear strain for degradation in sand

Abstract

Ability of laboratory determination of a sand behaviour in static and dynamic loading conditions are influenced by (among other things): sample preparation, number of tests, size of strains, speed of loading and averaging of the errors during examination. Dynamic load per se causes accumulation of the pore water pressure and the phenomenon of stiffening-threshold-degradation which makes the understanding of the sand behaviour more difficult. The complex behaviour of granular material, i.e. sand, is caused by chemical and physical properties of individual particles and their mutual interaction. Obviously, these interrelationships could not be analysed on the basis of laboratory testing. One way to analyse it is numerically, with algorithm that takes into consideration the characteristics of individual particles as well as their interaction in the sand matrix. Discrete element method (DEM) is a numerical method which takes into consideration the discrete nature of sand and shape of particles and is used as an elementary research tool for sand behaviour. Program package PFC3D is based on DEM and allows the modelling of the laboratory equipment, materials and calibration of its micro-characteristics, based on experimental results. The research of cyclic threshold shear strain for degradation in sand includes observation and visualization of the sample preparation (creation of the material skeleton), pouring of the material (transition from liquid to meta-stable state), influence of the particle shape (interlocking, arching), consolidation (deformation of the skeleton) and development and braking of force chains through the sample. This paper explores the suitability of the selected numerical method for modelling of NGI-DSS device, calibration of the tested granular material (Nevada sand) and preparation of the sample for testing and presentation of the stiffening-threshold-degradation phenomenon.