Comparison between controlled laboratory experiments and discrete particle simulations of the mechanical behaviour of rock

Discrete particle modelling offers the possibility for performing grain-scale modelling of rock mechanical behaviour. An important step in turning this kind of numerical modelling into a more quantitative tool is to perform a calibration of the numerical model against controlled laboratory experiments. In this paper, a strategy for calibration is outlined and results from tests with artificial rock-like materials are presented. The rock-like materials consist of spherical grains (glass beads) while the numerical simulations were performed with PFC3D (PFC: Particle Flow Code), which uses spherical particles as its basic elements. Stress dependent P- and S-wave velocities in uncemented assemblies of glass beads were computed using a Hertzian contact law, and a satisfactory correspondence with laboratory data was obtained without any fitting parameters. The discrete particle model was able to reproduce the observed dependence of unconfined strength and Young's modulus on cement content. The initial experiments presented here demonstrate the potential Of using discrete particle modelling as a numerical laboratory for studies of rock mechanical behaviour. (c) 2005 Elsevier Ltd. All rights reserved.