Calibration of 3D, discontinuum, strain softening, dilatant (SSD) models of mines often yields different rock mass scale (representative elementary volume, or REV) properties to empirical methods. In this paper, some steps towards a calibrated empirical scheme for estimating material properties for some types of 3d, discontinuum non-linear models targeting larger than (REV) scale phenomena are described. The scheme uses typical pre-mining rock mass classification date and strength tests to estimate the peak and residual yield, softening and dilatancy parameters. The scheme is a purely empirical device, derived from UCS and GSI field data and calibrated model results, and is a work in progress. The underlying nature of the Hoek Brown-GSI scheme (Hoek, E., and E. T. Brown., 1997; Hoek et al., 2002), is validated by the work.
An essential task of modelling of rock masses is to establish the length scale below which each particular rock mass can be treated as a continuum. On this scale the medium is said to be homogeneous (Witherspoon et al., 1981). Witherspoon et al. (1981) illustrated this with a diagram for permeability similar to the one shown in Figure 1.
The REV of rocks is ordinarily large enough to preclude laboratory measurements of REV scale properties, so calibration or an empirical estimate are necessary. Calibration is usually preferred.
Quantitative calibration is the process of adjusting model inputs to achieve a like-for-like match between measured data and model results. The quality of calibration qualifies the model for use and is measured by the resolution, precision and coverage of the models match to real measurements across space, time and length scales.
When calibration is not possible, REV scale properties must be empirically estimated. The most commonly used empirical scheme for estimating ‘rock mass scale’, or REV strengths for non-linear geotechnical models is the Hoek-Brown Geological Strength Index (HB-GSI) after Hoek, E., and E. T. Brown (1997). That scheme uses the laboratory scale unconfined compressive strength (UCS), a qualitative classification of the nature of the discontinuity network (GSI) and a parameter obtained from triaxial testing of laboratory specimens (mi) to establish the parameters of the yield potential function.
