ABSTRACT
A probabilistic approach is undertaken to incorporate uncertainty into the geotechnical design of a tunnel through gneissic environments in Northern Greece to determine optimum support measures. The gneissic unit, part of the Pelagonian formation, varies significantly from fresh to completely weathered and massive intact to disintegrated rock mass. Such geological conditions result from intense tectonic compression and ongoing structural activity. Firstly, the various lithological types for gneissic environments are defined based on literature. Regression analysis is then employed on the raw primary gneissic data to determine the relationship of the intact rock strength parameters (Young Modulus and intact strength). The most probable rock strength parameter values are assessed based on the results of intensive probabilistic analyses. Based on these values, rock mass strength parameters are then determined utilising the generalised Hoek-Brown failure criterion. At the same time, the second series of probabilistic analyses are performed to reduce uncertainty. Finally, the lowest, most probable and mean rock mass parameters, coupled with intact strength parameters, are tabulated through a lognormal distribution. At this stage, various analyses of the structural data (i.e. discontinuity sets, weathering grade, RQD) are carried out to determine Geological Strength Index (GSI) values to assist in defying Rock Mass Types. Moreover, the Rock Mass Behaviour Types are determined based on the expected tunnel behaviour due to the structural and stress environments. These behavioural conclusions and parameters are then used as input data for numerical analysis (2D) using RS2 and Unwedege. Finally, based on the most probable total displacements and the input parameters, a preliminary design of support measures to determine the most optimum combinations possible.
