Towards an understanding of joint roughness

Summary It is argued that the currently available joint models are incapable of accurate predictions of joint shear behaviour without resorting to substantial levels of empiricism. This is because these models fail to adequately quantify joint roughness, or appreciate the importance of scale. A novel approach, which uses fractal geometry to investigate joint roughness, is described. This approach goes beyond describing the symptoms of roughness and seeks to find the cause. In the application of this approach, the concepts of fractal geometry, fractal dimension and self similarity are described and used as a framework to formulate a statistically based and practical model for the characterisation of rock joint roughness. Important relationships between the fractal dimension and the more useful statistical parameters of standard deviation of both asperity angles and asperity heights are derived. These relationships not only provide a useful, working method for quantifying joint roughness, but are also shown to provide a basis for understanding the Barton empirical JRC-JCS model. In addition, the fractal model is able to provide conceptual models for the effects of normal stress on the shear behaviour of joints and the scale-dependence of joints.