Mponeng

Abstract: On 27 December 2007, a $$\text{M}_w$$ 1.9 seismic event occurred within a dyke in the deep-level Mponeng Gold Mine, South Africa. From the seismological network of the mine and the one from the Japanese–German Underground Acoustic Emission Research in South Africa (JAGUARS) group, the hypocentral depth (3,509 m), focal mechanism and aftershock location were estimated. Since no mining activity took place in the days before the event, dynamic triggering due to blasting can be ruled out as the cause. To investigate the hypothesis that stress transfer, due to excavation of the gold reef, induced the event, we set up a small-scale $$(450\times 300\times 310\;\text{m}^3)$$ high-resolution three-dimensional (3D) geomechanical numerical model. The model consisted of the four different rock units present in the mine: quartzite (footwall), hard lava (hanging wall), conglomerate (gold reef) and diorite (dykes). The numerical solution was computed using a finite-element method with a discretised mesh of approximately $$10^6$$ elements. The initial stress state of the model is in agreement with in situ data from a neighbouring mine, and the step-wise excavation was simulated by mass removal from the gold reef. The resulting 3D stress tensor and its changes due to mining were analysed based on the Coulomb failure stress changes on the fault plane of the event. The results show that the seismic event was induced regardless of how the Coulomb failure stress changes were calculated and of the uncertainties in the fault plane solution. We also used the model to assess the seismic hazard due to the excavation towards the dyke. The resulting curve of stress changes shows a significant increase in the last $${\sim}50\,\text{m}$$ in front of the dyke, indicating that small changes in the mining progress towards the dyke have a substantial impact on the stress transfer.

Abstract: A multi-disciplinary study of the response of the rock mass to mining and mining-induced earthquakes has been conducted in six gold mines in South Africa at sites deeper than 3 km, or with equivalent stress. More than 80 holes (with a total length of more than 2.8 km) were drilled in earthquake-prone areas to locate faults and install instruments. Microfracturing activity associated with an ML2.1 event at 3.3 km depth in Mponeng Mine and a highly stressed rock mass at 1 km depth in Ezulwini Mine were finely delineated and analysed. A hole drilled through the hypocentre of the ML2.1 event at Mponeng Mine allowed the stress and strength in the seismogenic area to be constrained. Previously published in situ stress measurements had been limited to depths or stresses smaller than 2.7 km or 100 MPa, respectively. We successfully measured stress at depths and stresses up to 3.4 km and 146 MPa, respectively, at four mines. These in situ measured stresses were used to calibrate elastic stress modelling and yielded better estimates of stress and strength on the rupture planes of seven earthquakes (ML2.1 to 4). Comparison of the elastically modelled stress with the constrained stress in a hole drilled across the ML2.1 fault and the strain change monitored in situ at a close distance showed that the elastically modelled stress was smaller but correctable. The South African National Seismograph Network was enhanced by installing 10 surface strong-motion seismometer stations in the Far West Rand mining district. Two other such dense surface networks are currently operated in mining areas by the Council for Geoscience: 25 stations in the Klerksdorp region, supported by the Mine Health and Safety Council; and 17 stations in the Central Rand to monitor fluid-induced seismicity. Research on the routine processing of spectral parameters was carried out using seismicity recorded by these 52 new surface stations.

Abstract: Secondary Minerals from the Carletonville Gold Mines: Witwatersrand Goldfield, South Africa Dave Kershaw a , Bruce Cairncross b , Brenda Freese c & Pierre De Vries a Mineral Resource Department Mponeng Mine , P.O. Box 8104 Western Levels, 2501, South Africa E-mail: b Department of Geology , Afrikaans University , P.O. Box 524 Auckland Park, 2006 Gauteng, South Africa E-mail: c Mineral Resource Department Mponeng Mine , P.O. Box 8104 Western Levels, 2501, South Africa E-mail: Published online: 08 Jul 2010.

Abstract: The re-interpretation of the old 3D seismic data from the world’s deepest gold mine of the Witwatersrand basin, have proven to be effective in enhancing the detection of fault zones and unstable lithologies that are potential mining hazards. Most of the structures identified seem to offset the gold-bearing horizons such as the Ventersdorp Contact Reef (VCR) and Carbon Leader Reef (CLR) with throws below seismic resolution limit (equivalent to one-quarter dominant seismic wavelength (λ/4)). To derive or ghost the seismically transparent Carbon Leader Reef model, we used more than 2000 drillholes that intersect the CLR (within the Central Group), VCR (top of the Central Rand Group) and Crown lava (top of the West Rand Group). The final ghosted CLR shows a very consistent dip with the VCR and Crown lava. The model also shows that 60% of the structures imaged at the VCR level at 2-3 km depth continue and displace the ghosted-CLR at 3.5-4.5 km depth in a similar manner. This model represents the gross-structural architecture that was formed following the massive Platberg age extension on the first-order scale structures.