When an iron orebody nears its end of life, the extracted material can become increasingly problematic to handle. Higher moisture content occurs as larger portions of originally-below-water-table material is mined, which in some deposits is accompanied by further changes in physical and/or chemical properties present along the edges of the orebody. Plant delays related to problematic material have affected all of BHP’s mining hubs to various degrees.
Following the recent successful work in Yandi, where the development of an index to quantify the handleability of channel iron ore deposits resulted in an increase in extracted volumes and an extension to the expected life-of-mine, the value of implementing similar indices for bedrock iron ore deposits has been recognised.
A materials handling index has been developed by combining observed empirical relationships between material properties and processing plant performance. Indices for each deposit were further refined and optimised by implementing a back-testing strategy which facilitated an understanding of what weighting or scaling each of the contributing components should carry to yield the highest success rate of predicting actual delays at a given plant.
Currently materials handling indices are deployed to all mine sites and are made available in block models as well as calculated for individual blasthole and production samples. Values in block models are obtained by predicting the required properties using machine learning models, whereas hyperspectral data is used to ‘measure’ some of those properties on a sample level.
This paper discusses the work that has been undertaken to obtain materials handling indices, the testing that has been done to optimise them, and what processes have been developed to ensure the best metrics are being used in the different planning and scheduling horizons to optimise extraction.