Open stope design; beyond the Stability Graph

The Australian Centre for Geomechanics (ACG) held two workshops in 2019—one in Australia and one in Canada—to discuss processes and methods currently used in the mining industry for stope design and reconciliation. In practice, this work strives to optimise stope performance and aims to maximise ore recovery (minimise underbreak), while minimising dilution (overbreak). One of the key findings from the workshops was that current stope design practices, almost exclusively, rely on some versions of the empirical Stability Graph methods developed in 1980s. Although robust and useful at the feasibility study stage, there are a number of inherent shortcomings with this approach. The Stability Graph’s limitations first lie in that it cannot assess underbreak, which is a very important measure of performance. Furthermore, the method doesn’t consider a number of factors which often have a critical influence on stope stability, for example, drilling and blasting or the presence of faults. Given the current industry standard practice, there is a great opportunity to reinvent the stope design process to optimise performance. A finding of the workshops was that every mine quantifies stope performance in terms of volume reconciliation with a focus on dilution and ore losses, and reconciling metal tonnes recovered from the mill. Many mines further investigate stope performance by looking at individual stope faces and the causes for the deviations between the planned stope shape and the measured cavity. Recent research at the ACG has shown that investigating overbreak and underbreak at a higher resolution is a powerful way to identify the root causes of overbreak and underbreak. Finer scale assessment is achieved using an octree analyses, which effectively divides the relevant stope volume into cubes on the scale of a metre and produces many tens of thousands of data points. The octree data structure and subsequent analysis underpins a data-driven stope design methodology. The ACG is currently developing a design approach based on multivariate modelling methods that can account for the many variables that influence stope performance and forecast stope overbreak and underbreak results. Back-analysing stopes based on octree reconciliation results provides a rich database for training the model. Once the model can achieve reliable forecasting of overbreak and underbreak, an interactive drill-and-blast design will allow for the assessment of different firing strategies to optimise stope performance.