عنوان مقاله [English]
Estimation of stress and deformation in rock mass around deep tunnels is of great concern in geotechnical engineering, rock mechanics and mining engineering. Due to large stresses in these tunnels, the surrounding rock mass may experience failure in some points. Hence, in these tunnels, the failure process is of great importance. There are different rheological models to simulate the failure behavior of rocks, including perfect plastic, strain softening and brittle models. Since many rocks in nature are hard, with brittle behavior, it is required to select a suitable mechanical model for simulating such behavior. In this research, the post-peak behavior of hard rocks is studied using the numerical methods of FEM and FDM, with $Phase^2$ and FLAC codes.
The effect of brittle behavior is evaluated on deformation and the extent of the plastic zone in deep tunnels excavated in hard rock. Then, these results are compared with those obtained based on perfect plastic assumption. Due to the fact that the rock strength after failure decreases to a residual level, different residual values are considered, in order to show the importance of selecting the correct residual parameters. Firstly, a circular tunnel is analyzed by both numerical and closed form methods and the obtained stresses and displacements are compared. In addition to tunnel wall deformation, the radius of the plastic zone is evaluated. In this circular tunnel, deformation of the tunnel wall with brittle assumption is obtained twice greater than the
results based on perfect plastic assumption. This ratio is decreased at higher radial distances. Moreover, radial stresses in brittle condition are lower than values obtained in perfect plastic condition. Also, tangential stresses in the yielded zone are lower than those stresses in the elastic region, because of stress redistribution. In this paper, a parametric study is accomplished on a horse-shoe tunnel, with both Mohr-Coulomb and Hoek-Brown failure envelopes. The rock is assumed homogenous and isotropic with a non-associated flow rule. The
Cai et. al. approach was adopted to estimate the peak and residual strength values of the rock mass. The importance of selecting correct residual strength is again presented. This study illustrates that numerical and closed form methods submit the same results in perfect plastic assumption, but different results in brittle assumption. Specifically, the obtained deformations and plastic zone are highly dependent on post-peak residual parameters. The strain softening behavior, as an intermediate condition between perfect plastic and
brittle conditions, was considered as well. In this research, it was observed that the available numerical codes usually exhibit non-symmetric results in very low residual parameters, even for symmetric loading. This can be
attributed to the weak points of these numerical codes.