Cannington mine data provides new opportunities to reduce dilution. Dilution occurs during mining when waste rock gets mixed in with the ore. Whilst some dilution is unavoidable, reducing the handling and processing of waste improves a mine’s productivity. South32’s Cannington Mine already has excellent stope performance compared to industry benchmarks. This means it isn’t easy to identify further opportunities to reduce dilution.
Laser cavity monitoring survey data (CMS) enables mines to measure average dilution (ELOS in metres). Stope design shapes are compared to actual mined shape. A “stope” is the excavation created by mining ore. Analysis of CMS stope data, together with rock and design data provided new opportunities to reduce dilution.
PRODFINDERTM deep analytics identified three new interacting dilution factors not included in standard design methods. Specifically, factorial ANOVA provided insights into the interaction between how long the stope is open, and the effectiveness of cables in managing fault related instability.
The factorial ANOVA shows that in the case of stopes with faults, dilution can be reduced by a combination of cable bolting and filling stopes within 28 days. It is important to note that the factorial ANOVA isn’t biased by rock conditions. Additionally, neither total paste dilution, nor minimum principal stress (σ3) explained the results. Whilst a p-value of 0.13 isn’t considered scientifically proven for academic purposes (typically requiring p < 0.05), these results are consistent with site observations of a relationship between stope open duration, and the extent of failure when unravelling on a fault, even when cables are present.
PETRA would like to take this opportunity to thank and acknowledge South32 for providing the Cannington mine data, and for permission to publish this work and supporting Kimberley Maher’s award winning undergraduate thesis: “A data science approach to identifying and quantifying causes of dilution at Cannington mine”: Underground Metalliferous Prize winner Kimberley Maher
This article is a summarised excerpt from the MassMin2016 paper by K. Maher, P. C. Stewart and S. Robotham entitled “A Data Science Approach to Identifying and Quantifying Causes of Dilution at Cannington Mine”. UQ Logan MassMin2016 Paper
Visit our PRODFINDERTM case studies to view more examples of how deep analytics identifies opportunities to improve productivity: