In this study acoustic emission (AE) techniques were employed to investigate the process of fracture formation in large, shear-critical, reinforced concrete beams and to gain improved insight into the mechanisms of shear failure. Large sensor arrays were specially designed to study various aspects of failure by observing the fracturing processes throughout the load history of the beams. Smaller, more concentrated sensor arrays revealed complex spatial and temporal fracture development at the slow quasi-static and spontaneous dynamic stages of propagation. When AE events are related to the micro-structure of concrete through computed tomography images and surface fracture measurements they show that coarse aggregates play important roles in shear fracture propagation in reinforced concrete.

Most of the hydraulic fracturing experiments by the mining industry in hard rocks were conducted to precondition the rockmass with the aim of improving caveability and fragmentation for block caving mining operations through the creation of hydraulic fractures (HF). Based on an extensive literature survey and models, it is suggested that successful preconditioning could be obtained through hydraulic treatment of the rockmass. This paper discusses the interaction between hydraulic fluid injection and the pre-existing discrete fracture network (DFN) in a rockmass subject to in-situ stresses. Three-dimensional numerical studies have been used in an initial attempt towards understanding how the rockmass and the pre-existing natural fractures response to fluid injection is affected by some of the DFN characteristics and borehole length. Results indicate that DFN characteristics control fluid percolation in low-permeability formations and influence stimulated rock volume. When injection pressures are lower than pressures required for hydraulic fracturing, borehole length does not influence significantly fracture surface area stimulated by slip. It is shown that representing the fractures explicitly in the numerical models and adopting a fully coupled hydromechanical modelling approach provide promising capabilities in the prediction of rockmass responses to fluid injection.

Rock mass classification provides fundamental data for a numerical stability analysis of rock structures. Among rock mass classification systems, the RMR and Q systems often are used for rock support system selection and the Geological Strength Index (GSI) system for estimating rock mass strength and deformation parameters. Moreover, the GSI system is the only rock mass classification that is directly linked to engineering design parameters such as the Mohr-Coulomb or Hoek-Brown strength parameters or the rock mass modulus. However, the original application of the GSI system requires long-term experience and a careful approach because of the fact that its use is a subjective decision. A quantitative approach to assist a less experienced engineer in assigning representative GSI values was presented. It employed the rock block volume and joint conditions as quantitative characterisation factors. Their approach is founded on the linkage between descriptive geological terms and measurable field parameters, such as joint spacing and joint roughness. In this study, a discrete fracture network (DFN) model incorporated with stochastic simulation is applied to characterise rock block size distribution for determination of the GSI. The fracture frequency obtained from the core logging data is analysed and provided to the DFN model as input data. Realisation of the DFN and its verification are conducted to establish the joint systems corresponding to the original fracture frequency. As a result, the stochastic simulation can successfully provide the information on the rock block size distribution to the procedure of the GSI determination.

The hybrid REBOP-FLAC3D approach allows prediction of the limits of the geomechanical zones defining the cave as a function of production. The results of the model can be used to derive estimates of: (i) caveability and caving rate, (ii) abutment stresses and cave loads, (iii) recovery and dilution entry, (iv) fragmentation, and (v) breakthrough timing and subsidence. The approach simulates the caving process by explicitly modelling each isolated movement zone derived from REBOP into FLAC3D to determine the yielded zone and cave back associated with mass drawn. After one cycle of extraction, REBOP informs the location of the movement zones and the presence of air, if it exists, to the continuum FLAC3D model. FLAC3D solves stresses associated with the presence of these zones and estimates the yielded zone surrounding the cave. FLAC3D informs REBOP which zones (initially inactive) could now be mobilised. The procedure is repeated until the draw schedule used as an input in REBOP is finished. Two of the main advantages of the hybrid approach include the capability of studying the potential impacts of isolated draw on cave growth and point loading on the extraction level, as well as the effect of including, explicitly, the airgap and mechanisms of fines migration and rilling on cave growth and subsidence. Two case studies are presented showing the capabilities of the hybrid approach.

No abstract