:
Maximum unsupported spans for highwall mining at an Australian coal mine have been estimated using UDEC numerical models with explicit bedding planes in the siltstone roof strata and a Voronoi tessellation to represent potential cracks in the siltstone material. The tensile strength for siltstone is shown to be a critical parameter, with a small but non-zero value being required for the models to produce the expected mode of roof failure, initiated by separation on bedding planes. The predicted stability of single and multiple spans then matches well results from an empirical method based on precedent experience.
When strip coal mines reach the maximum economic depth, further coal may be produced by highwall mining. This involves remote extraction of coal from a succession of parallel, unsupported drives from the base of the highwall. Methods to predict the maximum span that can be mined unsupported for a variety of geomechanical conditions will be critical to the success of highwall mining. This paper reports part of a collaborative research project between BHP Australia Coal Pty Ltd (BHPAC) and the CSIRO Division of Exploration and Mining, which aimed to develop geomechanical design guidelines for highwall mining under Australian conditions. Present highwall mining technology limits roadway widths to that of conventional continuous miner drum widths, which range between 2.8 and 3.6 m. The determination of maximum unsupported spans is nevertheless useful in that it enables a choice to be made within this range and provides a qualitative assessment of potential problems that could occur at a chosen span. Knowledge of the maximum span also leaves open the possibility of using wider drums in the future, which could lead to increases in both productivity and percentage extraction. Trueman (1993) developed a method to predict unsupported spans and stand-up times using rock mass classification systems linked to an extrapolation of precedent experience. From a review of the open literature, he concluded that previous numerical modelling studies had failed to give a quantitative assessment of unsupported spans. The present work investigates the application of numerical models to the quantitative prediction of maximum unsupported spans in environments similar to those at BHPACs Moura Mine in Queensland. There are strong links between the empirical model and precedent experience for the prediction of unsupported spans in fully underground coal mine tunnels (Trueman, 1993, and references therein). Thus a partial validation of the numerical models is possible by comparing numerical and empirical predictions. Nevertheless a full validation of both methods, specifically for highwall mining, will need to be carried out as case studies are documented.
In order to give realistic estimates of unsupported spans, numerical models need to simulate the correct mechanism of roof failure. In general the failure of coal mine roofs is initiated by bed separation, followed by bending and tensile yield of the rock "plates" between bedding planes. The need to allow this mechanism to develop influenced both the choice of the computer program and the way in which models were formulated.
