Operations managers across the resources sector are pulling together plans to create so-called intelligent mines, where all assets are networked together and capable of making decisions themselves.

The intelligent mine uses a fully integrated system to connect driverless trucks, trains and drills and using data analytics to optimise production, improve safety and cut downtime, using just a fraction of the regular workforce.

Defining intelligent mining

Professional services company Deloitte describes intelligent mining as a broader organisation transformation and not just a digital mine.

Intelligent mining will impact the way decisions are made, the skills you require, how you engage with your workforce and communities, and how to optimally utilise resources, such as energy.

For Darren Kwok, General Manager of Technology & Innovation at Barminco, intelligent mining can be defined as a technology enabler, but not the focus.

“We're very fortunate in Barminco. We're a mature 30-year-old business. We have many longstanding employees that have worked with us for a long time—my job is to augment their experience and their knowledge with the best in class technology.

“When I think about intelligent mining, I think about providing them with the right insights at the right time to make the best decision possible for the situation they are in," said Mr Kwok.

Brendon Cullen, Product Manager for Automation & Control at RCT, defines intelligent mining as successfully implementing advanced and technologically agnostic solutions into every step of the mining process.

Cullen says that the company has pioneered agnostic automation technology in the mining industry for more than 48 years.

“Our technology is designed to be highly interoperable and integrate with any mining machine regardless of original equipment manufacturer and other mine management systems. RCT also specialises in protection products, fleet management systems and mine communication networks,” said Mr Cullen.

Dr Penny Stewart, CEO and founder of mining AI software company Petra Data Science echoes those thoughts saying that it invests in making its technology fit within existing workflows.

The Use of Data

Born in Kalgoorlie, Dr Stewart started in the resources sector as a Mining Engineer, much like her father and grandfather before her.

“I worked on mine sites mainly around Western Australia after university and found that geology at any one mine was the cause of the greatest uncertainty," said Dr Stewart.

This experience led her back to university and a PhD in rock mechanics and blasting.

"During my studies, I received formal training in statistics and performed a lot of data analysis and modelling, and through this experience saw the potential for data science in mining. My PhD was very empirical and proved operations could build their own data models, and through statistical analysis also provided a new method for estimating how much data is required to build an operational data model," said Stewart.

After a brief period consulting and working with Newcrest Mining in Brisbane, Dr Stewart went on to start Petra Data Science in 2015 with a view that there was a lot of value that companies could extract from the data they were already collecting, including vast amounts of geological data. Even the name Petra means rock in Greek.

Dr Stewart says that Petra is one of the first – if not the first – companies in the world to apply machine-learning in a mining environment successfully.

“In 2016, we collaborated with Newcrest Mining at their Lihir mine in Papua New Guinea where we showed that you could mine historical data to develop algorithms and machine learning to provide insights into what causes downtime,” said Dr Stewart.

The company’s early success was highlighted in a presentation by Newcrest CEO Sandeep Biswas in 2017, where he said that Petra Data Science had eliminated mill overloads at the mine.

“At Lihir, our three mills had been experiencing multiple overload events each year, resulting in significant downtime and traditional engineering-based approaches had failed to adequately identify when overload events would occur.

"Over 360 million lines of data across 130 variables were collected, and cross-referencing this big data, the Petra team was able to identify causal factors and develop machine-learning algorithms which could predict future outages,” said Mr Biswas at the 2017 mining luncheon.

According to Stewart, once you have a good model of how a plant operates, you can then identify ways that individual operators achieve peak performance.

“We call this best day, every day. In unison with the machine learning and the data mining, companies can learn from their best operators and then give feedback to other team members so that they too can achieve the best performance,” said Stewart.

The Digital Twin

A digital twin can be thought of a time machine. It enables companies to look back in time to discover how they achieved our best performance. It also allows us to simulate the future by experimenting.

Stewart is quick to admit that Petra’s model of a digital twin was inspired by NASA. While the term itself was only coined by the space agency in 2002, the concept was first applied in the 1970s during the Apollo 13 program, where engineers on the ground needed to be able to rapidly account for changes to their vehicle while exposed to the extreme conditions in space, and with lives on the line.

When life support failed, NASA found they could no longer base corrective decisions on the original model because the actual module had undergone.

For Petra, NASA’s interpretation of the concept is most relevant to mining.

"Like mining, both aviation and NASA are heavily reliant upon spatial data. Weather and climate wreak havoc on aviation. Similarly, geological uncertainty adversely affects mining," says Dr Stewart.

The company’s digital twin software is the result of five years of intensive research and development into data integration across the mine value chain, including proprietary ore tracking technology.

Automation in Mining

When it comes to vehicles, five levels of autonomy have already been clearly defined.

At the lowest level of autonomy, level 1, we see automation like cruise control, where the automation does a single task and always requires human supervision. On the highest level, level 5, the vehicle is fully autonomous, and no human intervention of any kind is necessary, allowing even the steering wheel to be removed.

While becoming more and more prevalent in mining the levels are yet to be clearly defined in the resources sector.

Barminco’s Darren Kwok who previously worked in the automotive industry, says this is because the two sectors are different animals.

“It’s the tale of consumer versus industrial. Take the example of a software or applications developer. You can put an app on the market for people to download, which is very much a beta version, and over time it will improve to a point where it'd be a usable system.

“Whereas in the industrial space, we want production-ready products and you don't get the opportunities to try and learn as you do in the consumer space because you want a robust system and, often that's an excuse for low adaptation of new technologies. It's a bit of a vicious cycle, but things are changing,” said Mr Kwok.

According to Mr Kwok, mining companies are getting better at building sandboxes where they can trial new things saying that the appetite for innovation is increasing at a rapid pace.

For Barminco this was best represented in a recent trial with IGO at their Nova operation where the Barminco Remote Operating Centre or BROC successfully operated a machine at an underground mine.

The new system was trialled in the early morning at Barminco’s Head Office in Hazlemere, just outside of Perth on a machine located at IGO's Nova Minesite, almost 1000 kilometres away.

“The challenge from Barminco's standpoint is that being a service provider in the contracting world; we often have relatively short term contracts with our clients, and we have a many concurrent projects which are geographically spread out. This means that we don't have the luxury of running a fibre optic cable to each one of those separately.

"The challenge was, how do some, and potentially all, of the remote operations work with those constraints? And so BROC was put together to investigate how we would do that. Our focus is that we want to provide more flexibility to our workforce and work closely together with our clients.

From the technical standpoint, the first test for Barminco was, how to get past the fibre optic challenge.

While it may be something that we take for granted, Mr Kwok says in the company’s extensive research revealed that no one had ever operated an underground machine remotely over the internet.

"The first trial was run at Nova, where we utilised Sandvik's existing tele-remote system called Automine. We had an area underground at the Nova operations where we were able to build a secure connection via the internet to the control systems, with Sandvik ensuring that we had all of the normal failsafe's available,” said Mr Kwok.

They discovered that it was possible to use the internet to operate an underground machine at a mine in Western Australia’s goldfield from just outside Perth’s CBD, via the internet. However, they found that there was zero tolerance for any loss, which is incumbent to any internet traffic, which meant that every 10 minutes or so, they lost connection for a split second, which would switch the machine to a failsafe mode.

"After the first trial, we worked with Sandvik on decoupling the unnecessary systems and focused on only moving the control systems over the internet.

"We were able to then run a second trial with those changes in hardware and software for an entire shift with no downtime at all. On top of that, we were as productive, if not more productive than a standard shift,” said Mr Kwok.

According to RCT, all of these solutions are continually evolving in line with mining industry expectations and client’s technology roadmaps.

RCT’s flagship product is the autonomous solution, ControlMaster Guidance Automation which is used by major mining companies around the world in a variety of underground and surface mining environments.

Currently, ControlMaster Guidance Automation features include map-less autonomous tramming between loading and dumping points.

“Just this year the platform was equipped with its newest AutoDump function enabling an underground loader to autonomously tram to a dump point and empty its bucket without operator intervention,” said Mr Cullen.

While not working in the automation space, Petra closely aligns its software functionality to the level's automation, saying that the business started in the prediction space – level 1.

"The second area we work is in simulation or the ability to ask your data questions because you have a good model. For example, if I change this lever in my process, whether it be drill and blast or how you operate the crusher, how does that change the outcome. The third area we operate in is optimisation.

"Petra operates between level one and three. The reason we do that is because we think that engineers across the value chain are looking for operational decision support, and no black box solutions. Our digital twin software is a glass box solution where engineers have visibility on model inputs, relationships and accuracy," said Dr Stewart.

Power to the People

According to Deloitte, many companies have focused their efforts on technology and not the people who need to use it.

Companies that have been successful in their digital journeys have often been ones that recognise the importance of change management, and a good portion of realising the value can come from changing people's behaviour.

Dr Stewart says that Petra is aware of this problem and invested a lot in integrating its software into existing workflows and processes to enable the end-user to extract the most value in the shortest time.

“We believe that if we can make our applications fit within existing workflows such as the software or the control system that they see every day in the mine and the processing plant, the better.

"For example, when you buy a car with autonomous features, you don't need to get a new drivers’ licence to drive one. This is the underlying principle that guides our software development, and we believe it contributes to successful implementations.

“There's a lot of technology out there. But what there is a shortage of is technologies that are designed to generate value quickly and are designed with that human factor in mind.

"I really think that the companies that get that right and make decisions about technology based upon how fast they can get it into operations and how fast they can extract value are the ones who get the greatest competitive advantage” said Dr Stewart.

For Mr Kwok, companies must first define change management, saying that most people think of the documentation required to execute change when you mention change management. When in reality, it’s the entire process of capturing and understanding the problems and challenges, finding the right stakeholders, all the way to executing the technology.

“It's a broader conversation than I think most people define change management to be. It's cultural, and for us, we are focused on mining outcomes, and broadly that's about safety and production unit cost.

"We think that by understanding these outcomes first and understanding the challenges that our people face, then it gathers a little bit of buy-in, which is often missing in the technology story," said Mr Kwok.

Mr Cullen agrees, saying that technology implementation needs to be looked upon holistically.

"RCT ensures that change management is thought about in the buying process. We also ensure that our customers are aware of the changes required and helps them through the stages which are necessary for successful integration.

"It has been said many times that technology applied to a well-run operation will enhance and can help optimise the operation. Likewise, when it is applied to a poorly ran operation, it can fail spectacularly.

“At times the latest tech isn't necessarily. The best fit for the operation and a scalable approach to implementation can help the change management and provide the optimal solution for the customer's operations," finished Mr Cullen.