First prize in this year’s Maptek Geology Challenge has been awarded to Miguel Aliaga Oblitas, who showed how Maptek DomainMCF improved model consistency and reduced processing time in advanced multivariable geological modelling.

Senior Geomodeller with Newmont Mining Corporation, Oblitas receives a personal prize of $500 and a six-month DomainMCF subscription for his company.

Entries displayed a range of modelling techniques and applications, and judging was tight, resulting in a tie for second place awarded to Ed Lynch, Superintendent Exploration Geology, SIMEC, and Danielle Karbishev, Senior Resource Estimation Geologist, Fortescue.

Oblitas highlighted the chance to work with real data from an active mine and apply innovative methodologies to enhance geological modelling. His report defined the challenges of accurately predicting vein behaviour at depth, especially with limited data, and noted significant improvements in model accuracy and efficiency.

“The ability to integrate various geological inputs – such as lithology, structural data, mineralogical information and vein intensity – was invaluable,” Oblitas said. “The reduction in processing time from weeks to just a few hours greatly facilitated model updates and more timely decision making in exploration and resource estimation.”

Oblitas noted that the flexibility of the Domain Manager in GeologyCore, which enabled the creation of custom rules and rapid testing of different structural scenarios, was key to overcoming the deposit complexity.

“The integration of multivariable inputs, including lithological, vein intensity and mineralogical data, allowed for a more accurate and detailed representation of the deposit’s structural framework – something that was challenging to achieve with traditional methods,” Oblitas commented.

Danielle Karbichev jumped at the opportunity to trial GeologyCore and DomainMCF to increase efficiency for the Fortescue Resource Modelling and Estimation team and assess the potential applications of machine learning.

The biggest surprise for Karbishev was the ability of DomainMCF to rapidly generate grade estimations comparable to those produced via established estimation methods such as kriging, Maptek says.

“Geological volume model outputs and grade predictions drastically improved with more detailed geological input data, however purely data-driven models can also be used to identify trends and structures prior to interpretation and domaining,” Karbishev said. “More testing is required but it is clear that machine learning could revolutionise resource modelling and estimation as technology advances!”

In terms of GeologyCore, Karbishev found that the 3D drill hole visualisation options were particularly useful for validation of drill hole coding against modelled surfaces, as well as for stratigraphic domain interpretation using multi-element geochemistry and downhole geophysical data, Maptek says.

“Domain Manager allowed rapid flagging, processing and compositing of the drillhole database while Sample Manager enabled direct transfer of validated data points to DomainMCF for modelling,” Karbishev said. “DomainMCF provided almost instantaneous 3D block models, granting the ability to rapidly analyse vast databases and test alternate scenarios.”

Superintendent Exploration Geologist for SIMEC, Ed Lynch, was keen to apply DomainMCF to specific situations in the company’s hematite and magnetite operation, to test whether it could make things easier for geologists on site and unlock more time to go out and ‘kick the rocks’.

“The complexity of the geological setting we work in presents significant challenges to our geologists when it comes to 3D modelling and grade control,” Lynch explained. “DomainMCF was simple to use and incredibly fast. When given enough data it was able to produce similar results to more traditional human-driven modelling processes. I think it is particularly suited to grade control type modelling scenarios.”

Now in its fourth year, the Maptek Geology Challenge provides an opportunity to experiment with cutting-edge technology to create models directly from raw data, that are accurate and transformative for resource modelling and production applications.

The theme for 2024 – geological control for geological models – encouraged participants to combine their expertise with the power of machine learning to create models that accurately reflect geology.

Maptek provided geological modelling tools including GeologyCore and AI-assisted DomainMCF for up to four weeks, supported with documentation and technical assistance from our global team.

Steve Sullivan, Senior Geology Specialist and Technical Lead for DomainMCF, said that the Geology Challenge was founded to inspire geologists to engage with new approaches. The challenge provides a low-risk environment to test real data with the latest technology and explore avenues for improving existing practices.

“The winning entries were strong examples of tackling problems that are difficult to solve with traditional methods, and demonstrated the use of novel techniques to control their geology,” Sullivan said. “Recommendations for software improvements have already been passed on to our development teams.”