Abstract
In this paper, the three dimensional Particle Flow Code (PFC3D) with a newly developed contact model which can properly consider the contribution of moment to contact normal and shear stresses and the condition at which the contact bond fails was used to investigate the cracking process of rocks containing single flaws and under uniaxial compression. The new contact model was first validated by using it to simulate the experimental cracking process of gypsum containing pre-existing single flaws at different inclination angles. Then the influence of flaw shape (length and thickness) on the cracking process was systematically studied and the key features were identified based on the simulations. The results indicate that the first cracks (usually called primary cracks) initiate from the boundary of the pre-existing flaw and are always caused by tensile failure, and the secondary cracks first emanate from the tips of the pre-existing flaw due to shear failure and then develop to a mixed shear and tensile cracking zone.
1. INTRODUCTION
The initiation, propagation and coalescence of cracks dominate the deformability and strength of rock. Therefore, extensive experimental and theoretical research has been conducted on the cracking process of rock. The cracking process of samples containing flaws under uniaxial compressive loading provides insightful understanding of the cracking of natural rocks and thus has been studied extensively by researchers using rocks and manmade materials. The term flaws refer to preexisting cracks or holes in the rock specimens. The results indicate that tensile wing cracks are always found to be the first cracks to initiate from the flaw boundary and continued loading leads to initiation of a secondary group of cracks that may be a combination of tensile and shear cracks. However, the types and mechanisms of the new cracks that developed after the tensile wing cracks are often hard to be categorized, and the stress field inside the specimen is never fully revealed [1]. Most researchers simply describe them as secondary cracks without implying the mode of crack initiation. In this way the first cracks are commonly referred as primary cracks. So the words "primary" and "secondary" simply imply a temporal relationship.
