Effect of soil–infilled joints on the stability of rock wedges formed in a tunnel roof

Abstract

The use of an equivalent continuum for a rock mass is not always suitable for situations, where the failure is structurally controlled by discontinuities as in the case of wedges in the tunnel roof. In these instances, discontinuum approaches are usually preferred. Rock joints that are filled with soft infill are likely to be the weakest planes in a rock mass, having a dominant influence on its overall shear behaviour. In this case, the joint material model adopted for the discontinuities should be able to describe important mechanisms, such as asperity sliding and shearing, post-peak behaviour, asperity deformation, and the effect of the soft infilling. The latest version of a soil–infilled joint model is discussed here. It describes more comprehensively than previous models the occurrence of dilation and compression with lateral displacements, and also represents the hardening mechanism related to asperity interference as observed in the laboratory that cannot be readily captured by the existing joint models. An analytical approach for the analysis of rock wedges structurally controlled by soil–infilled joints and a numerical simulation based on a metro station collapse which occurred in Brazil in 2007 are presented.

Introduction

One of the most challenging tasks in the design of geo-structures (tunnels, dams, foundations, etc.) in a rock mass is to simulate the mechanical behaviour of the rock mass, which usually consists of an interlocking assembly of discrete blocks. The shear and normal displacements of rock discontinuities and their shear and normal stiffnesses control the distribution of stresses and displacements within the discontinuous rock mass. In conditions where an equivalent continuum-based approach is not applicable, the joint material model should be able to describe important mechanisms, such as the asperity sliding, deformation and shearing, post-peak behaviour, and the effect of a possible soft infilling.

Bandis [1] discussed a range of studies of the shear behaviour of rock joints as being theoretical or empirical. The theoretical approaches usually utilise numerical methods coupled with advanced constitutive laws to model the behaviour of the joint interface. Although advanced mathematically, these constitutive laws are still empirically inspired from the results of laboratory or field tests and do not explicitly model the kinematics of shear development at the interface [2]. As a result, many practical solutions are still based on empirical approaches which involve the analysis of test data to obtain a correlation between the joint and material characteristics, and the observed behaviour. Consequently, many analyses may only be valid for limited rock formations (individual projects), but not for a universal application as is the JRC–JCS model developed by Barton [3], [4] and Barton and Bandis [5], [6].

The complex and irregular geometries of rock joints can only be estimated statistically, which means that the simultaneous interacting processes of sliding, shearing, dilation, and elastic-plastic deformation are also of a complex nature. In addition, rock joints may present different conditions of weathering and contact surfaces, varying from clean, fresh surfaces to clay covered and slickensided surfaces. Amongst these contact surfaces, soil–infilled rock joints are likely to be the weakest planes because of the low frictional properties of the infill [7], [8]. Because of lack of models, it has been common practice to assume that their shear strength is equal to that of the infill material alone, usually described by a Coulomb slip model. This assumption (i.e. joint strength equal to that of the infill) may lead to unsafe conditions, where the rock–infill contact has lower strength [29] or uneconomical designs, where the overall strength of the infilled joint is not fully accounted for.

The most pronounced effect of an infill is to reduce the friction between the discontinuity boundaries or joint walls below that for the rock-to-rock contact condition. The shear behaviour is influenced by the nature of the infill material and the characteristics of the wall–infill interfaces. In addition, parameters such as the boundary condition, roughness of the joint, thickness of the infill material, pore-pressure, and the over-consolidation ratio may influence the shear strength of an infilled joint [8], [9].

Because of the lack of a generalised model that correctly simulates the behaviour of soil–infilled joints, limited attempts to capture the numerous factors affecting infilled joints have been made in recent years [9], [10], [11], [12], [13]. For this purpose, it is important to understand the governing mechanisms occurring during the shearing of infilled rock joints.