Seequent Driver helps OceanaGold identify additional gold, and points to 2000+ more ounces

By Colleen O’Hanlon

By adopting a new, artificial intelligence-driven technology platform, Toronto Stock Exchange-listed OceanaGold solved critical data-sharing problems between its geology and mine planning teams, leading to the identification of more than 2000 ounces of gold (US$10 million at the time of writing) with a portion of this already recovered from a single area at their Waihi mine in New Zealand.

The discovery of a previously unmodelled vein splay was made possible by implementing Seequent Evo and its structural intelligence tool, Driver. Connecting Deswik mine planning tools through Evo and augmenting geological interpretation with Driver allowed the company to move from manual, less reliable workflows to a more strategic, insight-driven approach. The results included significant time savings, reduced risk, improved collaboration, and more gold recovered.

Willi Vigor-Brown, Superintendent of Mine Geology, at the company’s Waihi Operation, talked about the challenges his team faced and how new technology unlocked significant value. The 2000 ounces indicated represents a 2% increase for the year for the site, and Vigor-Brown said there was potential for more gold to be found as Driver was applied to other areas at Waihi.

OceanaGold’s Waihi Operation includes the Martha Underground project, a long‑life, high‑grade ore body that has revitalised mining in the region and extended Waihi’s production future.
Source: OceanaGold

Related content:
* Watch: Mining challenged by global competition and accelerating change
* Watch: Technology can shift mining mindsets, ways of working
* Watch: Better targeted mining a step towards wider public acceptance

The geological challenge: The subjectivity of modelling

The Waihi underground mine is a geologically complex site where teams are exploring for new veins while also mining around historic workings, which in some cases are more than a century old. This presents a significant modelling challenge.

‘Any epithermal deposit can be challenging,’ Vigor-Brown said. ‘We’ve got multiple mine areas here, each with their own geological and structural complexity. On top of that, we’re mining remnant material…around historic workings. It makes it a pretty complex ore body to model and estimate.’

These intricacies introduce subjectivity and risk into the modelling process.

‘It’s possible that individual geologists will come up with different interpretations of data,’ Vigor-Brown said. The biggest risk is misinterpreting the geology before a major investment is made.

‘Robust geological domaining is crucial. Without that knowledge, there’s a risk of modelling discontinuous structures as a single vein. You can mine out to that interpreted vein and then things don’t match up.’

Willi Vigor-Brown, Superintendent of Mine Geology, at the OceanaGold’s Waihi Operation
Credit: OceanaGold

The workflow challenge: The friction of data

Before implementing a fully integrated workflow, the process of sharing data between the geology team (using Seequent’s Leapfrog Geo) and the mine planning team (using Deswik) was manual and inefficient.

Vigor-Brown said that the old method for sharing data between the geology (Leapfrog Geo) and engineering (Deswik) teams was time consuming. It required manually exporting and importing various files, and critically, the process of returning mine plans to the geology software meant losing all associated data, creating a significant bottleneck.

This created friction. The primary issue, Vigor-Brown said, was ‘the consumption of time, and it required you to open different software packages’.

This not only slowed down the geology team but also tied up software licences which could have been used by engineers for other technical work.

The turning point: finding a new way forward

The implementation of Seequent’s cloud-based platform, Evo, and its machine learning structural analysis tool, Driver, marked a turning point. Vigor-Brown said he could immediately see Driver’s potential and decided to put the new technology to the test.

‘I set myself a goal,’ he said. ‘I want to run my data through it, and I want to identify or find an unmodelled vein. I want to model it, estimate it, and deliver it into the life-of-mine model for the engineers to run Deswik Stope Optimiser and see what comes back.’

Driver’s artificial intelligence is capable of learning structural features such as the orientation and distribution of individual veins, directly from the raw drilling data. It’s a machine generating an objective structural interpretation that can be used to guide analysis and inform deposit models. Vigor-Brown began by validating Driver’s interpretation of the veins against known geology.

‘It matched up really well with all the underground ore control data, the underground mapping data,’ he said. ‘That gave me confidence in Driver.’

Driver ellipsoids (blue and orange) showing vein splay offshoots from the mapped vein. (red)
Source: OceanaGold

Vigor-Brown then explored the data and found ‘a group of clustered ellipsoids’ where no vein had been modelled. This insight was the key.

‘The accuracy and speed of Driver when it comes to identifying structural trends simply means that smaller veins, which may once have been overlooked, are now easily found. In this case the whole process took me one hour which is a pretty good return on the investment of my time,’ Vigor-Brown said. ‘Without Driver, I probably wouldn’t have seen it. I wouldn’t have modelled it.’

The figure on the left shows where Driver ellipsoids have highlighted an unmodelled mineralised structural trend in the drillhole data set. The figure on the right shows the splay wireframed after identifying with Driver ellipsoids.
Source: OceanaGold

The impact: a streamlined workflow and quantifiable results

The discovery of this new vein splay, a secondary branch off the main deposit, had an immediate and quantifiable impact. From the initial data-driven insight to having a fully estimated block model ready for mine planning took just one hour.

Vigor-Brown said the company had already begun recovering gold from the splay, and had further drilling planned to prove the remaining 2000+ ounces that the model indicated.

Beyond the discovery, the new integrated workflow connecting Deswik and Leapfrog Geo via Evo has fundamentally improved the team’s efficiency. ‘Having the ability to quickly and seamlessly upload fully attributed mine plans into Evo and bring them into Leapfrog has made reviewing short- and long-term mine plans easier and quicker. The review is quicker and it is more thorough.’

The above Leapfrog model shows the identified vein splay at OceanaGold’s Waihi underground site. The pink area indicates a high grade gold channel sample.
Source: OceanaGold

Precision and sustainability: Minimising environmental footprint

As well as the immediate financial gains and workflow efficiencies, the precision offered by this new integrated approach points toward a more sustainable future for the Waihi operation. The ability to accurately sub-domain large, complex ore bodies has significant environmental benefits, primarily through the reduction of waste rock.

In mining, dilution refers to non-valuable waste rock that is extracted alongside valuable ore. This waste must be hauled, processed, and stored, all of which consumes energy and increases the mine’s overall environmental footprint. By being more precise in their models, an operation can minimise dilution and thereby reduce its impact overall.

This is where Vigor-Brown sees the next major opportunity for Driver. By using machine-learning to analyse wide vein zones, his team can hone in and identify the specific, high-grade strands within the larger structure.

‘What I want to do with Driver is investigate using the application to guide sub-domaining of these lodes to minimise dilution,’ he explained. ‘I think Driver could be really powerful to try and highlight those actual mineralised zones, create sub-domain veins from that main vein, and then hand that to engineers so we extract it the best we can and minimise taking dilution.’

This targeted approach means less waste rock is disturbed. The environmental payoffs included:

• Less waste: By leaving more barren rock in place, the mine reduces the total volume of waste material that needs to be managed on the surface.
• Reduced energy consumption: Hauling and processing less waste rock leads to lower fuel and energy consumption, directly reducing the operation’s carbon footprint.
• More efficient blasting: More precise geological models can also lead to more efficient use of explosives, targeting only the rock that needs to be extracted.

Seequent Senior Product Manager Ryan Lee said that ultimately, this technology allows OceanaGold to align its production goals with its sustainability commitments, proving that efficient mining can also be responsible mining.