Evaluating the effect of scale and heterogeneity on the mechanical behaviour of rock blocks

Rock block strength is a significant factor controlling the rock mass behaviour and the rock-support interactions in fractured rock masses. Especially when the design relies on discontinuum analysis, the adopted block properties are a dominant driver influencing the results. A series of 2D UDEC grain-based models were performed on samples of different sizes and qualities to simulate the results of lab- and block-scale experiments. The effect of pre-existing defects was simulated either in a smeared sense by adjusting the grain micro-properties or by explicitly modelling micro-Discrete Fracture Networks (DFN) that were previously generated within FracMan. Relationships that link the rock block strength with its volume and in-situ heterogeneity were proposed for the estimation of scaled Mohr–Coulomb and Hoek–Brown parameters. The UCS of blocks was expressed as a function of scale, defect intensity, persistence and strength. The quantified scale/condition dependant reduction of block strength was then linked with a block-scale Geological Strength Index parameter named micro GSI (mGSI). Special focus was also given on the selection of appropriate constitutive relationships and discontinuum modelling techniques when simulating tunnel-scale problems. For continuum blocks in-between DFNs the traditional Hoek–Brown approach does not capture realistic behaviours and the modified Damage-Initiation and Spalling-Limit approach is needed to predict the expected damage near the excavation boundaries. When blocks are simulated as a packing of grain elements, considerably reduced damage, stress relaxation and deformation is predicted as the Voronoi skeleton creates a well-interlocked structure that clamps the pre-existing joints. The research highlights that the estimation of representative block properties is of equivalent importance with the selection of appropriate modelling approaches.