Ojeda, P, Fogel, Y, Roy, J, Valerio, C & Rogers, S 2024, 'Probabilistic three-dimensional kinematic analysis to improve design reliability of complex underground excavations', in P Andrieux & D Cumming-Potvin (eds), Deep Mining 2024: Proceedings of the 10th International Conference on Deep and High Stress Mining, pp. 1103-1114, https://doi.org/10.36487/ACG_repo/2465_71
(https://papers.acg.uwa.edu.au/p/2465_71_Roy/)
Abstract: Conventional approaches to evaluate the stability of wedges in underground excavations include identification of dominant joint sets and deterministic stability checks on possible wedge geometries for a simplified excavation profile. However, these conventional approaches tend to produce extremely conservative results or require several assumptions as well as engineering judgement to filter and scale the wedges to achieve reasonable wedge sizes and shapes. Further, when the excavation geometry is complex, many different kinematic models need to be constructed to capture individual geometries.
An improved approach to assess kinematic stability in complex underground excavations has been developed using a discrete fracture network (DFN)-based method. The DFN method has a number of advantages over conventional kinematics; importantly, the fabric orientation pattern is applied fully without restriction, allowing more complex structural patterns to be evaluated including both deterministic structures as well as stochastically generated ones; and complex three-dimensional (3D) excavation and blocks shapes can be considered in a single analysis. Multiple realisations of the excavation-scale DFN can be generated to simulate block formation around the whole excavation surface, with unstable blocks being identified from each iteration. Then, the blocks are used to generate heatmaps of the probability of unstable block occurrence, volume, apex height and required support pressure to make the blocks stablealong the excavation surface. Once the analysis has been conducted, a design wedge can be established for each key excavation surface which can be carried over to ground support design. 
An example of the improved method is presented for an underground crusher complex which has complex excavation geometry and areas of concerns (e.g. multiple mine levels, large brows, and large spans). The example shows the significant advancement in the probabilistic assessment of 3D kinematics, including improved visualisation of kinematic stability checks, less arbitrary engineering judgement, and a single model that can capture complex geometries.

Keywords: discrete fracture network, underground excavation, kinematics, risk assessment