FLAC3D was employed to assess pillar and panel stability for a partial stooping operation at a coal mine in South Africa, where the main requirement was that ground surface stability be ensured. The stability of square and rectangular pillars was assessed using a strain criterion linked to laboratory testing, as well as Salamon and Munro's empirical pillar strength criterion. The relationship obtained from the initial analysis was then applied to predict potential failure of the irregular shaped pillars created during the stooping operation.
A project was commissioned to determine whether partial stooping of coal pillars may be done to increase the extraction ratio of Mine A. The main pre-condition stipulated was that ground surface stability be ensured during the stooping process. A minimum factor of safety of 1.6 was set as a criterion for pillar stability. Subsidence criteria were chosen based on available published data for buildings, roads, streams, railways, etc. (Bakker, 1992).
It was decided to carry out numerical modelling, which would be compared to the empirical factors of safety calculations to determine whether irregular shaped pillars may remain stable or would be classified as unstable. All underground information such as pillar sizes, mining heights, depths below surface, etc. was collated and it was decided to use the average values for mining height, pillar width, depth below surface, etc. in the models, which was assumed to be representative of the underground bord and pillar layout.
No partial pillar extractions have previously been carried out at this particular mine, therefore no static or seismic deformation data was available to calibrate the model. The approach was therefore taken to carry out the numerical modelling to predict potential deformation and strains, which would be confirmed through the use of instrumentation and monitoring data collected during the pillar extraction process (back-analysis using static and dynamic (seismic) rock mass response data). This paper therefore does not discuss the back-analysis of partial pillar extraction, but rather the predicted rock mass and pillar behaviour prior to pillar extraction.