Stochastic Approaches To Rock Characterization For The Design Of Underground Structures

A site characterization strategy formulated for the probabilistic design of underground structures is presented, which takes into account the variability of rock properties and the stochastic nature of the rock failure process. It entails determining probability distributions of rock properties by utilizing the theory of extremes and Monte Carlo simulation. Laboratory and field test data are analyzed and presented to illustrate the methodology. The results indicate that the distributive characteristics of rock mass properties can be estimated using empirical relationships currently available. This results of analyses can be incorporated into the design of underground structures, and then, the relative uncertainties of the failure probability of the structures can be assessed.