Upscaling of fractured rock mass properties - An example comparing Discrete Fracture Network (DFN) modeling and empirical relations based on engineering rock mass classifications

In this study a comparison of Discrete Fracture Network (DFN) modeling and empirical relationships based on conventional engineering rock mass classifications is carried out in order to determine the elastic properties of fractured rock masses. Uncertainties are considered by a stochastic description of the fracture network geometry and a probabilistic approach, utilizing Monte Carlo simulation techniques, for the various other input parameters. The proposed workflow is applied to two sandstone outcrops. Geometric information on the fracture systems is obtained from terrestrial laser scanning providing the input parameters (orientation, size and intensity) for each fracture set including probability density functions required for DFN modeling. After adding the mechanical properties of intact rock and fractures obtained by laboratory tests, the DFN approach allows to calculate values for rock mass deformation modulus and Poisson's ratio assuming either isotropy or vertical-transverse isotropy. These estimates are compared to values derived from empirical relationships based on engineering rock mass classifications. The comparison of the two approaches and, in particular, of the different empirical relationships reveals a significant scatter in the estimated rock mass properties. Only those empirical equations which incorporate intact rock properties in addition to the classification score yield plausible results, i.e., a pronounced weakening due to the fractures, and are in agreement with the estimates of the DFN approach. The additional effort of generating a DFN model shows its benefits if an anisotropic rock mass and/or spatially varying properties have to be considered, thus, providing a comprehensive hydromechanical characterization of anisotropic fractured rock masses.

Automated summary

This study compares Discrete Fracture Network (DFN) modeling with empirical relationships based on conventional engineering rock mass classifications to determine the elastic properties of fractured rock masses. The results show that incorporating intact rock properties in addition to classification scores in empirical equations yields more reliable estimates for predicting fractured rock mass properties.