Creating tomorrow today – DomainMCF devours Big Data

As a geologist, you look at some surface or downhole data and intuitively ‘know’ what is going on. You build a sub-surface model in your head and try to prove it with the data. You hope that when it comes to mining the plan will be close enough to reality.

The traditional paradigm of modelling is more data = better models = greater confidence. The reality is more data = bigger models = longer time. The upshot is less time for thinking = hasty decisions = lower confidence.

Machine learning improves on the old paradigm. More data is still better. Clean data is even more important. Intuition is still valid. But the model should be driven by the data, not the bias of the geologist.

You can trust data-driven modelling

Get your data right, then you can sit back and relax. The results may even be available before you’ve finished your coffee. Your geological knowledge lets you choose the best option from a range of potential models using the same data. An inherent measure of uncertainty attached to the model allows mine planning to proceed with confidence.

DomainMCF has been tested with more than 100 million data attributes from one of the largest
ore deposits in the world and delivered a 3D domain model with grade trends
for 25 different variables in less than four hours.

The interplay and correlation between grade trends is preserved in the machine learning workflow. In areas of uncertainty, grade trends can be used to target where to drill and collect data to improve confidence in the geological model.

Challenging the old paradigm

I don’t challenge the old order lightly. There’s nothing intrinsically wrong with some traditional methods. We just don’t have time to use them properly with big data.

Our industry has 60 years of confidence in kriging and its variants and 30 years of experience in different forms of simulation for resource estimation and reporting. A new technique requires extensive comparison with and validation against accepted practice before becoming the new standard.

This comparison process is happening exponentially. I’m amazed at the response – we already see companies subscribing to DomainMCF for use in domain modelling for resource reporting.

Grade trends are not yet grade estimates

However, we must remember that grade trends predicted using machine learning are not to be confused with grade estimates for resource reporting.

Using DomainMCF, geological domain boundaries are predicted in 3D and then the prediction of grade trends is constrained within these geological domains. Currently, DomainMCF grade trends can be input to locally varying anisotropy (LVA) estimation techniques. The grade trend vectors in x, y and z replace the standard sample search ellipsoid orientations and facilitate grade estimation around fold hinges and curved domains.

When we released grade trends in March 2022, I noted that further R&D is required to tune the underlying processes to convert these grade trends to grade predictions. Industry feedback will help DomainMCF do that.

What people are saying

‘I joined Maptek in 2021 after working as a geologist around Australia for more than a decade. As a veteran of Vulcan, I’ve been using traditional methods to help customers with resource modelling, so my introduction to DomainMCF blew me away.’
Andrew Sanggaran, Maptek Technical Services Consultant – Geology

‘DomainMCF is fast and simple to set up and run. Processing 3750 samples and 2 million blocks took less than 2 minutes. Comparing the marble quality classification produced by DomainMCF to the conventional method it was clear that the machine learning classifications appear more uniform. The machine learning engine also provides a measure of uncertainty for predictions. This is useful for identifying areas where more sampling may be required.’
Dr Ioannis Kapageridis, Professor, Department of Mineral Resources Engineering, University of Western Macedonia

‘Machine learning will become particularly attractive if the process can not only model geological domains, but also return reliable grade estimates for mine planning across the full range of mineralisation styles. A well-understood confidence measure can assist in risk quantification of both geology and grade.’
Fletcher Pym, Senior Mine Geologist, IGO Limited

Watch this space

The industry has been struggling to find experienced personnel during the current mining boom, so embedding years of experience into smart systems helps get the job done on time and under budget. DomainMCF is a standout among smart systems, and smart miners can prove it for themselves.