This paper constitutes part of the work undertaken to determine if there are economic and practical ways of safely mining stabilising pillars in South African gold mines. These pillars represent an enormous resource of high-grade gold ore with an estimated value of approximately US$18 billion. Stabilising pillars were first introduced into South African gold mines in the 1960s in order to reduce the mining spans in longwall layouts which would result in reducing volumetric closure in stopes. The consequence of this was to reduce the energy release rate (ERR) in mining panels and decrease the seismicity, leading to improved mining conditions and safety. One of the assumptions made was that the rock mass behaves elastically. However, this was, at best, an approximation. In reality, inelastic deformation around these deep gold mining excavations is common and complete closure of the stoping excavation can occur twenty metres or less from the mining face. This paper describes observations that indicate effects of stresses in the rock surrounding the supposedly destressed tunnels beneath the mined-out longwall stopes. In order to confirm the inferences from the observations, stress measurements were undertaken at three separate sites in tunnels beneath the mined-out longwall stopes and between two lines of strike pillars. The measurements revealed that significant stresses were being transmitted through the old mined-out areas. The implication of this finding is discussed in the paper. Using the input from the stress measurements, numerical models were run to determine the likely stress on the pillars. It was found that the stabilising pillars in the longwall would be less stressed than a simple elastic analysis would predict. Drawing on other work, the depth of the fracture zone in the pillars is estimated and it is speculated that the deep fracture zone would decrease strong ground motion at any breast mining face in a strike pillar. In conclusion, the paper then considers the possibility of mining stabilising pillars in South African gold mines.
Longwall mining was introduced in the Witwatersrand goldfields early in the 20th century. The concept was simple, the mining of the 1 m wide ore body was the advanced cut with tunnels and other excavations lagging in the stress-relieved zone beneath the mined ore body. This type of mining was undertaken to mitigate the effects of rock fracture and rockbursts due to high stresses. Strike stabilising pillars were first introduced into these longwalls at the East Rand Proprietary Mine in the 1960s, in order to reduce the mining spans which would result in reducing volumetric closure in stopes . The consequence of this would be to reduce the energy release rate (ERR) at the mining faces. As was expected, seismicity decreased in proportion to the decrease in ERR  and mining conditions improved. Figure 1 shows a schematic plan of a longwall with strike stabilising pillars. The strike stabilising pillar concept was then introduced at other mines employing the longwall mining method. It was later realised that the combination of stabilising pillars and backfill could, in an additive manner, both contribute to the reduction in volumetric closure and hence a reduction in ERR at the mining faces.