Rock mechanical problems are often governed by the shear strength of joints, but assessing it theoretically is difficult because of the many underlying factors. Consequently, expensive and time-consuming shear tests must be performed either in laboratory or in situ. Artificial shear tests based on numerical models would therefore be a valuable complement. In this paper, results from an initial study on real shear tests are compared with numerically simulated shear tests performed with the computer software PFC2D. The results from the analyses are good from a qualitative view, but also revealed the need for further research. In this paper, the results from the performed analyses are presented and the current limitations and requirements of further development are discussed.
Rock mechanical problems such as sliding stability of rock slopes, the stability of caverns located in rock, or foundations on rock are in many cases governed by the joints that run through the rock mass. Therefore, it is necessary to gain knowledge about the shear strength of these joints. The shear strength of rock joints is influenced by several factors such as the surface roughness, the compressive strength of the joint surface, the normal stress, the degree of weathering, the matedness, the infilling material, and the scale. This makes it difficult to theoretically determine the shear strength of the joints. Therefore, shear tests are often performed, both in laboratory and in situ. These tests are often expensive and also time consuming. Thus, it would be valuable if artificial shear tests could be performed using numerical models.
The shearing of rock joints involves both sliding on asperities as well as crushing, shearing and tensile failure. To successfully model this complex behavior, the numerical model must be able to reproduce these failure modes. According to Cundall (2000), using a micro-mechanical model to numerically represent rough rock joints could lead to simulated behavior that is similar to that observed in real joints. Similar observations were also made by Asadi & Rasouli (2010, 2011), who presented an application of the contact bond model for rock joints with symmetric triangular profiles and with irregular rough profiles. They showed PFC2D (Particle Flow Code) is able to numerically reproduce the progressive shear behavior of rough fractures. Park & Song (2009) performed a study in PFC3D using the contact bond model on the relationship between the macro-response of a joint and the micro-parameters. They analyzed the influence on the macro-response by changing the following micro-parameters; the particle size, the joint friction coefficient, the joint roughness and the joint contact bond strength. According to their observations, both the shear behavior and the failure process were successively modelled and corresponded well to those observed in shear tests.