Making the most from resource modelling
By Industry Queensland
Numeric block modelling, otherwise known as interpolation, is a huge part of the resource reporting lifecycle of any mineral deposit.
This week’s JKMRC lecture, organised by the University of Queensland-based Sustainable Minerals Institute, will show how rarely it is used to its true potential.
It’s in speaker Dr Alexander Wilson’s address, ‘Automatic Block Modelling Using Locally Adaptive Machine Learning’, which describes how most projects assay for 30-40 elements, collect mineral percentages, measure density, magnetic susceptibility, RQD, recovery and others, but they rarely block model all these data attributes.
In his current role at Minerva, Dr Wilson is the technical product owner and lead data scientist creating DRIVER – a 3D modelling software for mineral systems that uses machine learning geostatistics to model drilling data rapidly and automatically.
Dr Wilson will outline how the DRIVER system can quickly and accurately estimate all attributes into block models so they add to improved ore deposit knowledge.
“DRIVER’s unique technology works to extend block modelling to all data simultaneously, assuming little about the deposit, the AI tools are capable of automatically identifying and locally adapting to many of the complex geological challenges that these natural systems present – including folded and geometrically non-stationary deposits,” he said.
Specialising in numerical and geostatistical modelling with machine learning, Dr Wilson has worked in a variety of field mapping-based roles in iron ore (Western Australia) and porphyry exploration (Canada) and has experience in volcanology natural hazard assessment with Natural Resources Canada.
It’s on tomorrow morning at 9.
