Advanced techniques for the monitoring of pillar and excavation behaviour at a deep level massive mine

Abstract

The mining methods used at South Deep started with conventional mining (pre-1998), which later evolved into mechanized mining with various layouts for drifting and benching, low-profile horizontal destressing with crush pillars and long hole stoping (LHS). The mining method was then changed during 2015-2016 to the current high profile destress with LHS. Pillars in the destress cuts are designed to yield and not pose a significant rock burst risk. High profile development was initially conducted using smaller pillars (4.5 m x 10 m and 6 m x 10 m) which completely yielded however, as these were prone to excessive scaling and extensive rehabilitation requirements. Numerical modelling was conducted by Lilly, (2016), to optimise the destress pillar dimensions. The study concluded that a larger yield pillar is viable in the high profile destress cuts. The new pillar design was still not proven in situ owing to lack of information regarding to yielding, closure rates, convergence and fracturing through the core as well as the transition between the old and new mining layout. As a first approach, borehole camera surveys were conducted to assess the pillar conditions. Ground penetrating radar (GPR) technology was then used to determine fracturing around the yield pillars and where possible, the results where compared to the borehole camera surveys. A Zeb Revo scanner was then used to obtain accurate closure and pillar yield performance. This data was used to calibrate actual vs model-predicted displacements and to calibrate numerical models, reconciling pillar and also pro-actively identify areas that were undergoing extensive deformation. This paper will describe the outcome of the above-mentioned monitoring and reconciliation programmes and the interpretation of the results.

Ground support and strata monitoring: what is needed?

“…Mobile scanners are less accurate but they have the advantage of covering larger areas in a short time. They are usually carried around by hand (Andrews et al 2019), connected to a vehicle (Lynch et al 2017) or even attached to drones (Jones et al 2019). A typical output from a LiDAR scanner is shown in Figure 6.…”