As miners look to new strategies to exploit valuable resources, issues such as stakeholder demands, talent shortages and risk management must be addressed. In an increasingly digital industry, information is demanded at a faster rate, and changes must be responded to instantly. The strategic planning cycle can no longer be tied up by cumbersome approaches to resource modelling.
A resource model should portray the best understanding of geological observations and facts. The resource does not change over time, just the understanding of its characteristic idiosyncrasies. Computer-based modelling has traditionally replicated hand-drawn methods. Recently, other approaches - simulation of categorical variables and implicit techniques - have been adopted.
To survive, companies will need to embrace new technologies and advances in digital transformation. Automated resource modelling underpins the digital mine platform. New methods can contribute to new understanding.
This paper introduces using machine learning to revolutionise geological modelling. The resource modelling process is a bottleneck in most operations, taking weeks or months to complete. Machine learning will assist in debottlenecking processes at all mining operations, improving the timeliness of geological data delivery and its potential impacts on upstream applications, such as mine planning, scheduling and process plant performance.
CITATION: Sullivan, S, Green, C, Carter, D and Sanderson, H, 2019. Machine learning a new paradigm for resource geology, in Proceedings Future Mining 2019, pp 2324 (The Australasian Institute of Mining and Metallurgy: Melbourne).