Geostatistical simulation to map the spatial heterogeneity of geomechanical parameters: A case study with rock mass rating

Abstract

A better characterization of complex rock masses is essential in geotechnical engineering, as the empirical systems widely used for this purpose have significant limitations and do not provide adequate answers for risk analysis. Geostatistics offers a set of tools that allow not only predicting the rock mass properties, but also mapping their heterogeneity at different spatial scales and quantifying the uncertainty in their actual values. In this paper, two geostatistical approaches are compared for modeling the Rock Mass Rating (RMR), which is used to geomechanically characterize the rock mass in geotechnical works. The first approach consists of the direct simulation of the RMR values, based on a Gaussian spatial random field model. In contrast, the second approach uses the truncated Gaussian model to separately simulate the individual parameters of the RMR, which subsequently are summed to obtain the final RMR value. The computation time, practical implementation, level of details and post-processing outputs that can be obtained from both approaches are analyzed. Besides the RMR mapping and associated uncertainty, the deformation modulus is subsequently obtained based on these maps together with empirical expressions.