A screenshot from one of Maptek's mining software programmes. Image: Maptek
Business Technology

From model, to mine to market… and back again

How graphics technology that originated in the 90's film industry has enabled advanced mine design, scheduling and planning capabilities. Carly talks spatial modelling in mining with Maptek’s Peter Johnson

One of the biggest challenges when writing about a topic as vast as 3D modelling is simply where to start. However, on this occasion, my task was made infinitely more straightforward by an email:

“I saw your post on Linkedin about data modelling and visualisation… I’m available for an interview if you like?”

I did indeed ‘like’. And so Peter Johnson, chairman of mining software firm Maptek, and I sat down for a virtual glass of wine, for him, coffee for me (due to the difference in time zones), in late August to talk about the origins of 3D modelling technology and its place in the mining industry.

Peter Johnson is chairman of Maptek
Peter Johnson is chairman of Maptek

“Spatial modelling has been applied in mining since the late 1970s, but the real application of what we now recognise as 3D modelling in mining came about during the 90s due to advances in computing capabilities, and the ability of software to display data and allow users to interact with it in the 3D space,” explained Johnson.

“A lot of that technology actually came out of the computer graphics world. Maptek pioneered the use of 3D modelling in geological mapping and mine design using Silicon Graphics technology that was developed for computer generated effects for movies and entertainment back in the mid-90s. The first orebody ever viewed in 3D on a screen was Escondida in Chile (using Vulcan).”

Prior to that (brace yourselves Generation Z), geologists and engineers relied on 2D drawings and tables of data to map and model orebodies.

“There was still quite a bit of software involved but, initially, geology and mine planning were looked at as tables of raw data, or as sections and plans in 2D space,” said Johnson with a grin as my eyes boggled. “You can imagine the relative difficulty reading that compared to in 3D?”

As someone who tends to baulk when faced with numbers, I certainly could.

Linking the value chain

Topics like geological modelling, mine planning, data analysis or scheduling are often treated separately, but they are in fact quite closely related and often hang off of common datasets based around the orebody, its value and how it’s to be extracted.

“The common thing between all of those aspects of a mining operation, is that they’re all spatial considerations,” Johnson told me.

“The size and shape of an orebody defines things like the haul cycle costs and times. The geotechnical parameters and structures within an orebody define the shape of the pit, which determines where the ramps can go, which defines what sort of trucks can be used.

“Then the orebody characteristics dictate the type of crushing, grinding and recovery processes that are required. And that’s all variable within the orebody in a spatial sense as well.

“When you look at mine planning, for instance, the size and shape and distances heavily affect the economic viability of a mining operation. So, the ability to understand all of those aspects in 3D is critical.”

To me, one of the greatest values of mine modelling technologies is the ability to allow geologists, engineers, executives even, to see, interact with and, ultimately, understand different facets of an orebody that is likely buried beneath the ground, at some depth, and cannot otherwise be seen or touched.

They give us control over scale when planning operations, both in time and space, which is incredibly useful in an industry that operates on a completely different and, to some, inconceivable, magnitude than any other.

That capability, when combined with other spatial tools, like simulation, allows us to run multiple iterations of a process, design or operation over different time schedules and examine the outcomes long before a shovel even hits the ground.

It allows us to guard against, or prepare for, a great deal of uncertainty. Of which there is much in mining.

“A huge amount of effort goes into orebody modelling using geostatistics and trying to predict, from a relatively sparse set of drill samples, the entire composition of the orebody,” said Johnson. “But there’s still a lot of learning that goes on at the mining stage and when the ore is processed.

“Using the data that’s available now, there are techniques and capabilities that can help us understand the behaviour of an orebody and how the entire mine performs as a result of its variability.

“That’s why one of the main reasons that Maptek has partnered with [Australian data science specialist] PETRA Data Science. Because Maptek products contain a lot of information about the orebody and mine design, and PETRA has solutions that can inform us how the plant will perform in response to that.”

Enter data science

Which was a nice segue into the topic of data science…

I asked Johnson to talk in more depth about some of those data collection and analysis techniques.

“We’re seeing a lot more mining companies take a holistic, less siloed, view of their data,” he said. “Companies are looking at entire mining operations as a system, and are trying to understand what the data means, and what data they need to properly control or understand the variants in that system, so they can link things like short-term scheduling to marketing, customer demand, product prices and profitability at the mine.

“There’s a real business incentive to link these parts of the value chain.

“One of the things that AI and data science gives the industry an opportunity to do, is to cut through some of the natural shortcomings and variations in data quality and to understand, overall, the relationships between their data and what it’s telling them.

“For example, most mines have a very rich block model, and spend a huge amount of time and effort making it as accurate as possible. They do the same tracking and measuring the performance of their machinery.

“There are huge amounts of data collected, but what’s the relationship between those two things? That’s the tricky bit.

“It’s not something you can give to a good engineer in Microsoft Excel and they’ll come up with a result in a few days. Being able to apply machine learning techniques to link the two and identify trends adds a huge amount of value.”

Techniques like machine learning and their integration with spatial modelling technologies are going to become even more valuable in the coming years as the mining skills shortage, which is already being felt in some areas, really hits home. But also as more executives and engineers from different backgrounds come to work in the mining industry.

If those people don’t have technical backgrounds, how else can we make these vast quantities of data manageable and understandable for them?  

“Without these technologies, how are they going to know what ‘good’ looks like?” Johnson countered.

“Being able to use historical data and understand, amongst all the variation and seemingly random mine performance data, how to identify what the best results are and how to replicate or cause them… that’s where the real potential lies.”

“The size and shape of an orebody defines the haul cycle costs and times. The geotechnical parameters and structures define the shape of the pit, which determines where the ramps can go, which defines what sort of trucks can be used.” Image: Maptek

One source of the truth

I’ve noticed a growing trend among technology providers towards creating platforms that provide ‘one source of the truth’. (A term that I’ve always placed in a box with ‘blue sky thinking’ and ‘big data’. I hate the term big data. But that’s a story for another day.)

What does it even mean to have ‘one source of the truth’?

“Most mining companies manage quite well day-to-day with cross functional collaboration and management, good leadership etc.,” explained Johnson. “But certainly, having one ‘version of the truth’ or one source for all your data and one way of considering and evaluating it helps a lot. That’s a trend we’re seeing more and more.

“The data that’s recorded is quite unique to the mining industry, and so one of the challenges has been getting collaborative platform capabilities delivered to be able to properly treat and use that data in the unique way the mining industry needs to.

“The ability to bring all that spatial data together… there’s a validation and context step required to be able to compare two data sets that might be related but will never match up. To build a data framework to be able to do that, and then relate that data and what it’s telling us to the mine business model and control the outcomes in production… there’s been a lot of work on that in the last five years.”

The ability to handle large quantities of data in a timely fashion is particularly important. I’ve spoken to numerous experts who’ve described future mining operations as having a virtual throttle lever that allows managers to instantly speed up or slow down production in response to market demand and commodity pricing.

“That’s not, in concept at least, a terribly long way away,” said Johnson.

“We’re working on a programme called MaterialMRT which tracks and reconciles production from the geological model through loading and hauling, stockpiling and blending, to enable customers to have a dashboard that shows what stock they have available, what they’re planning to deliver as product, what lead time they need and what’s being produced next. It gives them an end-to-end view of the mining process to help inform decision making.

“The mining supply chain is relatively complex when you start including things like rail and port and shipping, but it’s not so complex that it can’t be done.

“The agent-based simulation capability in Maptek’s Evolution software allows mines to run different simulations to optimise the net present value (NPV) of an orebody from the life of mine, all the way through to short term scheduling. By breaking it down into phases so you can start to understand the impact of short-term decisions on long-term goals.

A screenshot from Maptek’s MaterialMRT programme which provides mines a virtual dashboard to manage their operation from end to end. Image: Maptek

“We’re also doing a lot of work using measurement-while-drilling (MWD) technology to generate data from production drilling which can help provide a better understanding of the orebody and its characteristics which, in turn, allows timely decisions in blasting.

“There is a lot of real-time data being used in automation. A lot of systems are ‘open’. It’s just a matter of connecting the data and using it to make a difference.”

The automatic mine

Nowhere is data making a bigger difference than in mine automation.

“There’s been a lot of publicity and noise made around automation in mining,” said Johnson, with a hint of scepticism. “Autonomous trucks and drills are great, they offer a lot of benefits in terms of efficiency and safety… but the real challenge lies in automating decisions about where those trucks go, what they’re loading or dumping, which haul route that they take etc.

“The connection between the 3D spatial information describing the mine, the orebody, the performance of the mine and those decision processes… those will become more prominent as mines become more automated. That’s probably where we’ll see the greatest advances in the next five to ten years.

“There’ll be a lot more movement towards… I guess the analogy is in the manufacturing world, in the computer aided design of a machined part. You can send that design to a CNC-controlled mill and, as long as you’ve got stock in that machine, it’ll make the part. The rest is all automatic.

“The more knowledge you have about the components going into the process and the behaviour and performance of the process, the more you’re able to automate it.

“I think, long term, we’ll see mining moving in that direction. There’s a very good business case for it, and technology will be the enabler.”

An analogy from manufacturing was the perfect way to end the interview. Because if there’s one thing to be learnt from the success of spatial modelling technology in mining, it’s that this sector has a lot to gain from venturing outside of its bubble.

In the late 80s, if someone had told geologists they would soon be using graphic capabilities from the film industry to create mine plans, they would probably have laughed. But look at where we are today.

It’s proof that an open mind is one of the most powerful tools of all.

0 comments on “From model, to mine to market… and back again

Leave a Reply

%d bloggers like this: