On the Influence of Direction-Dependent Behavior of Rock Mass in Simulations of Deep Tunneling Using a Novel Gradient-Enhanced Transversely Isotropic Damage–Plasticity Model

Abstract

In engineering practice, numerical simulations of deep tunneling are commonly based on isotropic linear–elastic perfectly plastic rock models. Rock, however, commonly exhibits highly nonlinear and distinct direction-dependent mechanical behavior. The former is characterized by irreversible deformation, associated with strain hardening and strain softening, and the degradation of stiffness; the latter is due to the inherent rock structure. Nevertheless, the majority of the existing rock models focuses on the prediction of either the highly nonlinear material behavior or the inherent anisotropic response of rock. The combined effects of nonlinear and direction-dependent rock behavior, particularly in the context of the numerical simulations of tunnel excavation, have rarely been taken into account so far. Thus, it is the aim of the present contribution to demonstrate the influence of both effects on the evolution of the deformation and stress distribution in the rock mass due to deep tunnel excavation on the example of a well-monitored stretch of the Brenner Base Tunnel (BBT). To this end, the recently proposed gradient-enhanced transversely isotropic rock damage–plasticity (TI-RDP) model, is employed for modeling the surrounding rock mass consisting of Innsbruck quartz-phyllite. The material parameters for the nonlinear transversely isotropic rock model are identified by means of three-dimensional finite element simulations of triaxial tests on specimens of Innsbruck quartz-phyllite, conducted for varying loading angles with respect to the foliation planes and different confining pressures. Subsequently, the results of the nonlinear 2D finite element simulations of tunnel excavation are presented for different anisotropy parameters and different orientations of the principal material directions with respect to the tunnel axis. The capabilities of the TI-RDP model are assessed by comparing the numerically predicted results with those obtained by the isotropic version of the RDP model.