| A sound understanding of rock mass characteristics is critical for the engineering prediction of tunnel stability and deformation both during construction and post-excavation. The rock mass modulus of deformation is a necessary input parameter for many numerical analysis methods to describe the constitutive behavior of a rock mass. Tests for determining this parameter directly by in situ test methods are inherently difficult, time consuming and expensive, and these challenges are more problematic when dealing with tunnels in weaker, softer rock masses where errors in modulus (stiffness) estimation have a profound impact on closure predictions. In addition, rock masses with modest structure can be candidate sites for highly sensitive structures such as nuclear waste repository tunnels. For these generally stiffer rock masses, the correct modulus assessment is essential for prediction of thermal response during the service life of the tunnel.;Numerous empirical relationships based on rock mass classification schemes have been developed to determine rock mass deformation modulus in response to these issues. The empirical relationship provided by Hoek & Diederichs (2006) based on Geological Strength Index (GSI) has been determined from a database of in situ test data describing a wide range of rock masses with GSI values greater than 25 and less than 80. Within this range of applications there is a large variation in measured values compared to the predicted relationship and predictive uncertainty at low GSI values. In this research, a practical range of rock mass quality, as defined by GSI, including "BlockyDisturbedSeamy" rock masses, "Very Blocky" and relatively competent rock masses are analyzed using discretely fractured numerical models. In particular the focus is on tunnel response. Tunnel closure in these simulations is compared to predictions based on modulus estimates. The proposed refinement to the Generalized Hoek-Diederichs relationship is made on the basis of these simulations for tunnelling applications. |
