In this study, the displacement discontinuity boundary element method (DD-BEM) is used to simulate crush- and shear-type failures associated with observed seismic events at two deep South African gold mines: mine X, near Carletonville, and mine Y near Welkom. Crush-type failures associated with a 12-month extraction sequence at mine X are simulated using the limit equilibrium method, and shear-type failures associated with flooding of a geological structure at mine Y are simulated using a simple shear-displacement (or ride) method; the input parameters of the DD-BEM models are calibrated according to the cumulative potency associated with observed seismic events. In both cases, the calibrated DD-BEM model adequately describes the observed seismic response in terms of the cumulative potency, number of large seismic events and the corresponding potency-frequency distribution.


A recent study by Malovichko and Basson (2014) used the framework proposed by Salamon (1993) and Linkov (2005) to model seismicity in mines, and the corresponding methodology is now used by the Institute of Mine Seismology (IMS) to routinely estimate the seismic hazard at several mines in Australia. However, only shear-type failures were considered by Malovichko, and it was necessary to extend the developed algorithm to simulate on-reef, crush-type failures associated with face- and pillarbursting in South African mines. These crush-type failures are simulated using the limit equilibrium method (LEM) suggested by Napier and Malan (2007, 2012). The LEM uses a simple force-balance model to adjust the boundary conditions of a displacement discontinuity boundary element model (DDBEM) to accommodate the residual strength of the crushed rock mass around a tabular mining excavation. Previously, LEM has been used to simulate pillar failure in platinum mines (Spottiswoode and Drummond, 2014), and the preliminary results of this study have shown that the LEM may also be appropriate for estimating the seismic hazard associated with on-reef, crush-type failures in deep South African gold mines. The latter requires calibration of the developed LEM-based algorithm according to the observed seismic response such that the potency release of recorded seismic events is correlated with the cumulative potency of simulated crush-type failures. In addition, shear-type failures on geological structures are simulated using a simple shear-displacement method, and the corresponding algorithm is also calibrated (in terms of potency) according to the observed seismic response. These calibrated models are then used to estimate the potency release of future (or planned) mining sequences and provide an efficient, robust estimation of the associated seismic hazard.

This content is only available via PDF.
You can access this article if you purchase or spend a download.