Quantifying void fabric using a scan-line approach

Abstract

The internal topology or fabric of granular materials including soil is known to influence their mechanical behaviour. Discrete element method (DEM) simulations, physical experiments using photo-elastic materials or Schneebli rods, and high resolution computed tomography scans can all give detailed particle-scale information for these materials. Data from these simulations and experiments can then be used to develop quantitative measures of the soil fabric. Considering DEM simulations in particular these measures of fabric have focussed on the particle and contact orientations; quantitative analysis of the topology of the void space is less common. This paper revisits a method for quantifying void fabric that was proposed by Oda et al. (1985) [1]. An approach for applying this method to 2D digital images of particle systems is outlined. The algorithm is then applied to analyse the results of a 2D DEM simulation of a biaxial compression test in which two complementary shear bands develop. This application highlights the viability of using digital image analysis to interpret DEM simulation data. The void anisotropy induced during shearing is significantly more pronounced in the shear bands. There is a clear link between the volumetric strain and the anisotropy, showing that the voids are becoming both larger and more elongated during shearing.