@inproceedings{2025_38_Hayman, author={Hayman, LJ}, editor={Dight, PM}, title={Utilising data science to test similarity of rock mass unit strength distributions in the Pilbara}, booktitle={Slope Stability 2020: Proceedings of the 2020 International Symposium on Slope Stability in Open Pit Mining and Civil Engineering}, date={2020}, publisher={Australian Centre for Geomechanics}, location={Perth}, pages={625-636}, abstract={Rio Tinto Iron Ore (Rio Tinto) undertakes diamond drilling within the Pilbara to geotechnically characterise the rock mass pertaining to proposed pit walls prior to implementation every year. Data from the diamond core is logged, tested in the field and in the laboratory; then committed to a local database and analysed for that particular pit or deposit when the need arises. Currently there is the assumption that each deposit within the Pilbara has independent strength properties and is diamond drilled extensively as a result. Within the Pilbara rock masses there are a series of continuous, stratified units whose genesis are from the same depositional, folding, faulting and weathering events. This paper aims to utilise diamond core logging data to statistically verify that there are similarities of strength properties for some of the rock units across the Pilbara in neighbouring deposits. Proving this to be the case potentially will provide statistical evidence and the impetus to enable the Rio Tinto geotechnical department to reduce the metres of diamond drilling undertaken with a zero net effect on slope stability and safety in implemented slope designs. A total of 39,815 test observations from 2008 to 2018 from five separate neighbouring deposits were used in this study; and were derived from three different tests including field estimated strength (FES), point load (PLT) and unconfined compressive strength (UCS) which are key parameters in slope stability design. The data was interrogated using state-of-the-art data visualisation and statistical software appropriately chosen based on the size of the samples and nature of the distribution. This research follows on from recommendations made by Maldonado & Haile (2015) applying non-parametric statistical methods to the skewed data for a more appropriate approach. This study found sufficient statistical evidence that there is similarity in rock strength properties between five proximal deposits with 95% confidence in the vast majority of cases. }, keywords={drilling}, keywords={statistics}, keywords={optimisation}, doi={10.36487/ACG_repo/2025_38}, url={https://papers.acg.uwa.edu.au/p/2025_38_Hayman/} }