This paper presents the results of ongoing research on the characterization of rock mass structure from discontinuity data. Multivariate clustering analysis represents a relatively recent development in characterizing the structure of rock masses. Multivariate clustering allows characterization of discontinuities into subsets according to multiple parameters, such as orientation, spacing, and roughness, where, rather than considering one variable at a time, a number of parameters can be treated simultaneously, so that the interactions between parameters are taken into account. The comprehensive algorithm has been developed into a software package called CYL. It enables fully automated multivariate clustering analysis and offers various visualization tools, such as a three-dimensional stereonet, a stereoscopic view, and a statistical table. In this paper we focus on the implementation of the algorithms and the application of this method to field data, both from oriented core and mapped road cut discontinuity data.
The characterization of rock discontinuities is an essential requirement for most of the rock engineering projects. Discontinuities within rock dominate the mechanical and hydraulic behaviors of rock masses (Figure 1). Because of lack of effective and comprehensive tools for the interpretation of the data, the discontinuity analyses suffer from the inability to incorporate more than one parameter simultaneously. Multivariate clustering analysis represents a relatively recent development in characterizing the structure of rock masses. It characterizes discontinuities into subsets according to multiple parameters, such as orientation, spacing, and roughness, where rather than considering one variable at a time, a number of parameters can be treated simultaneously, so that the interactions between parameters are taken into account (Dershowitz & Einstein 1988). Rock characterization using oriented bore-hole data or linear mapping data is often more useful because it is cost effective, and can target the exact location of a proposed underground structure. But because of the lack of effective tools for the interpretation of this data, it is more often than not underutilized.
In the previous papers (Maerz & Zhou 1999; Maerz & Zhou 2000) a new approach to the analysis of oriented borehole discontinuity data was introduced. Rather than considering parameters such as orientation, spacing, infilling, wall rock strength, roughness and mineralization individually, a multivariate approach was used. The new approach is designed around utilizing multivariateclustering algorithm, considering both spherical orientation data and spatial data. The data can be visualized in a "three-dimensional" stereonet where joint poles are plotted on individual "stacked" stereonets, where each pole is plotted with respect to its own stereonet, and each stereonet is plotted in a linear position that corresponds to the position where the joint corresponding to that joint normal intersects the bore hole (Figure 2). This idea was first used by Wenk (Wenk et al. 1987) to represent the pattern of lattice preferred orientation in deformed rocks.