Abstract

To reduce energy intensity across the mining value chain, the largest lever an operation can pull is blasting. Even though it may account for less than 10% of the overall total cost, it directly drives the other 90% of processing costs. Moreover, the blasting results impact downstream processes such as loading, haulage, crushing, and grinding.  The challenge for traditional blasting practices, is it typically utilises low resolution data from block models where blasthole samples were taken with significant separation, translating into a manual and subjective definition of blasting domains with less data from extrapolations of block model samples. This low resolution available which is used to define rock hardness domains, eventually can lead to over blasting practices, under the assumption that the chemical energy cost of the explosive is less than electricity costs of the comminution plant. The combination of this low resolution of the orebody data, combined with more heterogenous orebodies, concludes in inconsistent blasting outcomes as drill and blast designs are built on Blast Domains. Along with this, there are significant impacts from applying both more or less explosive energy that the rock in fact requires affecting fragmentation, excavations rates, and environmental outcomes. Through the combination of Measure While Drilling (MWD) data, processed through Machine Learning (ML) models and a domaining algorithm, Orica has successfully developed an integrated workflow to enable automated high resolution rock hardness segmentation to enable operations to refine blast loading designs to more closely match rock mass variability with the required explosive energy in a production workflow. This workflow enables a tailored explosive energy allocation at the blasthole level based on more granular blasting domains, in order to maximise the use of this energy translating in a direct reduction of explosive consumption while improving blasting performance.   This presentation will review three different successful deployment of this workflow, combining automated rock hardness domaining using MWD, and using axial delay as a proxy of Unified Compressive Strength (UCS) from an external sensor in the drill string, in situations when there are limitations with MWD data quality. This workflow has been instilled from a Proof of Concept, to becoming the established drill and blast design process today at Roy Hill iron ore mine in Australia, Antucoya copper mine in Chile and a large open cut mine in Tierra Amarilla, Chile. A high-level introduction to the workflow will be reviewed as well as the results at all three operations which: successfully reduced their explosive consumption by up to 10 %, increased excavation rates, improved particle size distribution and maintained a more homogenous particle size distribution (PSD) along with significant improvement in environmental outcomes including reduction in ultra fine materials and scope 1 emissions. This is an open workflow which enables operations to utilise the delivered Blasting Domains in any blast design package and to further ensure the execution of specific loading designs by utilising a digitised drill and