Tag Archives: machine learning

SensOre and CSIRO to clean up exploration datasets for AI algorithms

SensOre has commenced a joint project with the Commonwealth Scientific and Industrial Research Organisation (CSIRO) looking at automation and efficiency in big data cleaning and processing solutions for the mineral resource sector.

Access to high-quality datasets has, in the past, been an obstacle to applying cutting-edge predictive analytics to solve geoscience and mineral exploration problems, according to SensOre.

The joint project, which will confront this problem, will see several CSIRO data science experts embedded within SensOre over the next six months.

SensOre says it aims to become the top performing minerals targeting company in the world through the deployment of artificial intelligence and machine-learning technologies, specifically its Discriminant Predictive Targeting® workflow.

“SensOre is committed to world-leading mineral exploration research and development,” Richard Taylor, CEO and Director of SensOre, said. “This is the second time SensOre has worked with CSIRO and the engagement has led to order of magnitude improvements in our approach. Australian government support, such as the Kick Start initiative, is important for Australian technology companies looking to grow globally.”

The joint project benefits from CSIRO’s Kick Start initiative for innovative Australian start-ups, providing funding support and access to CSIRO’s research expertise. The program offers eligible businesses matched funding of up to A$50,000 ($34,681) to undertake research activities.

Eclipse Mining’s SourceOne can help miners prepare for the unexpected

While all businesses should have a risk plan, the very nature of mining presents a unique set of problems and opportunities for consideration, according to Abinash Moharana of Eclipse Mining Technologies.

Whether an operation is contained in one location or spread throughout the world, strategic plans impact physical activity, which, again, impacts the strategic plan. The optimisation of this continuing cycle is imperative to success, Moharana says.

Even when physical operations are forced to shut down, data analysis and other planning must continue to meet revenue projections, and to prepare for physical operations to resume. This is increasingly relevant during today’s COVID-19-related lockdowns.

“In the event of a physical shut down, all existing plans become irrelevant,” Moharana, Technical Product Manager at Eclipse, says. “New plans are needed quickly. And the most significant risk (or opportunity) at this point, is the quality and timeliness of your data, and access to that data for employees working in remote locations.”

Eclipse’s SourceOne solution, which features a collaborative platform to connect data from different sources, and a datahub to store historical and contextual data, can help here. SourceOne renders this data serviceable for analytics and for adoption of tools, such as artificial intelligence and machine learning.

The ability to create new reactive plans and multiple scenarios is imperative to surviving shutdown events, Moharana says. “And, doing so accurately may provide an opportunity to emerge on the other side even stronger than before so that when physical operations resume, you can jump right into it without wasting time or resources.”

Moharana added: “To achieve this, you must turn your biggest risk into opportunity using high-quality data that is easily accessible with real-time updates. And SourceOne can make this happen for you.”

As a mine platform designed with multi-users in mind, SourceOne is made to host multiple remote users.

It accommodates concurrent on and off-site users, while also handling off-line users with automated merges to a clean state, according to Moharana. “A geographically disparate team can work seamlessly together, transfer data and messages as part of a workflow, always be able to work with the latest data, and know the genesis of each data.”

Moharana explained: “With all the existing mine plans rendered redundant, the mine planning engineers can start working on new mine plans with updated assumptions and requirements. The requirements may have changed, such as volume becoming an immediate priority, rather than profitability, to be able to supply the end user. While doing this, the goal is also to not deviate too far from the strategic plan.

“SourceOne maintains the complete history of each project and data. Historical plans can be used to generate the differentials between the existing plans and the strategic plans. These can be weighed against the new goals, and a sub-optimal plan with smaller recovered metal may be considered as it does not stray too far from the strategic plan.”

With so many plans being made, the chances of errors magnify, as does the need to be able to audit the results with an internal or external auditor.

SourceOne maintains a complete audit trail of the entire project, so every plan can be traced back to the assumptions that were used at each step, according to Moharana. This speeds up the process to validate the suitability of the plan to be implemented in the field.

Mine planners must realise that for their corporate management, there is a big difference between risk and uncertainty, Moharana says. Risk is something that can be measured, while uncertainty cannot. The difference lies in measurable information.

“A good Business Continuity Plan (BCP) for mine plans allows the enterprise to convert some of these inherent uncertainties into calculated risks, which then can be properly weighted by the management,” Moharana said.

Mining personnel may think that the COVID-19 pandemic is an unprecedented event, but they need to be prepared for such a unique occurrence, according to Moharana.

“This preparedness allows mines to better manage the risk by being able to make mine planning a part of your BCP and ensuring that the mine is well prepared for any disruption, rare or not.

“De-risking the mine planning process is one of the many ways SourceOne can help your organisation become more resilient and be prepared for ordinary (and extraordinary) events.”

Datarock machine learning drill core analysis tool hits major milestone

DiUS and Solve Geosolutions have leveraged PyTorch-based image analysis techniques to help automate the analysis of drill core imagery and provide greater insight that can be used to influence decision making on mine sites.

Mine sites often produce between 100 and 1,000 m of drill core per day, generating hundreds of images a week at a single drill site, according to PyTorch, an open source machine learning platform.

Historically, these images have been kept as a record of the job and a resource for geologists to refer to, rather than being used as a quantitative dataset that adds value to a mining operation.

Tapping into this rarely used data source of drill core imagery, DiUS – an Australia-based technology services organisation with a strong focus on machine learning and deep learning image segmentation analysis – joined forces with Solve Geosolutions – a mining-focused data science and machine learning consultancy – to build a machine learning-powered, cloud-based platform to automate the analysis of this drill core imagery using image segmentation technology.

Together, DiUS and Solve Geosolutions worked on applying a range of PyTorch-based image analysis techniques, including image classification, object detection and both semantic and instance segmentation, to a range of geological problems.

In particular, the team wanted to understand how different models performed in terms of training and inference speed, training requirements and overall accuracy of prediction to inform how they could be deployed in a production environment.

One model they used extensively is Mask R-CNN.

“This model can be applied to a range of segmentation tasks, however it can also demand large training datasets that are sometimes not available,” PyTorch said. “To support this, the team developed novel ways to increase the initial, often sparse training dataset through data augmentation techniques such as rotation, flipping, contrast, saturation, lighting and cropping.”

Following the initial discovery period, the team set about combining techniques to create an image processing workflow for drill core imagery. This involved developing a series of deep learning models that could process raw images into a structured format and segment the important geological information, according to PyTorch.

Their first productionised process was a metric referred to as RQD, otherwise known as rock quality designation. “RQD is a difficult and monotonous dataset to collect manually,” PyTorch said. “It’s also well suited to automation, and of high value to a mining operation. RQD is used by engineers to understand the strength of a rock and is used in the design and engineering of a mine.”

With the release of Detectron2 – a PyTorch-based computer vision library released by Facebook in October 2019 – the team made the decision to switch from the previous model implementation on TensorFlow to the next-generation platform to help improve instance segmentation tasks, PyTorch said.

The team found Detectron2 to be four times faster in training the models (using GPUs) and three times faster in inference (using CPUs) than the previous model implementation, PyTorch said.

Building the models on PyTorch-based frameworks meant the team was able to reduce valuable training time across the board. This increased the number of experiments and, as a consequence, improved model accuracy on an identical dataset. The PyTorch Dynamic Graph also made it much easier for the team to debug and investigate any issues that arose, PyTorch said.

The resultant Datarock platform is a software as a service offering that applies machine learning – image segmentation technologies – to drill core imagery and delivers information about a mineral deposit’s geology at scale, and at a resolution that’s not been previously economically viable, PyTorch says.

Since launching Datarock in 2019, the team has extended the platform to turn drill core imagery into high-quality datasets to support decision making throughout the entire mining cycle.

“The models perform optical character recognition, instance and semantic segmentation, as well as geological statistical analysis on a dataset,” PyTorch said. “This allows a geologist to inspect the model prediction and check for quantity and quality in unmatched datasets.”

Mining and exploration companies can now get consistent geological information from their rock core imagery in a matter of minutes, according to PyTorch.

“This near real-time power is enabling more intelligent decisions to be made further down the mining chain – saving time and money that can be put towards other business-critical projects – and freeing up geologists to do higher value tasks,” PyTorch said.

To date, the Datarock platform has processed more than 1 million metres of drill core images – that’s enough core to cover the distance between Sydney and Melbourne – over 800 km.

Yamana lets GoldSpot loose on Cerro Moro exploration database

Following recent successes at El Peñón, GoldSpot Discoveries Corp has been reengaged by Yamana Gold  to use machine learning to identify new drilling targets at the Cerro Moro gold and silver mine, in Argentina.

Yamana has commissioned GoldSpot’s team of geologists and data scientists to examine its entire database and look for previously unrecognised data trends to identify areas of potential mineralisation at depth and on a regional scale, it said. By engaging GoldSpot, Yamana seeks to minimise exploration risk and mitigate exploration and drilling costs, the company added.

“GoldSpot will use its geoscience and machine science expertise to clean, unify and analyse exploration data from Yamana’s Cerro Moro mine and produce 2D and 3D targets for the exploration program,” GoldSpot said. “GoldSpot will also deliver new geophysical, geochemical and geological products produced through the reprocessing of the satellite images and other relevant layers which will help interpretations and mineralisation models.”

Denis Laviolette, Executive Chairman and President of GoldSpot, said the new contract with Yamana validates its work, thus far. “Yamana has been an incredible supporter of GoldSpot and we are proud to be a part of their digital transformation,” he said.

GoldSpot was previously commended for its use of machine learning technology to improve exploration targeting and also contribute to the meaningful increases in mineral resource inventory at Yamana’s El Peñón mine.

Henry Marsden, Senior Vice President, Exploration, at Yamana, said in February: “The collaborative AI process undertaken with GoldSpot has allowed Yamana’s exploration team to leverage many years of multidisciplinary exploration data and is playing a significant role in the current exploration targeting process at El Peñón. We are pleased with the progress that our partnership with GoldSpot has yielded so far and look forward to continued success.”

Freeport to invest in data science, AI programs at North/South America mines

After carrying out a successful pilot at its Bagdad copper operation, Freeport McMoRan says it is rolling out a program across its North America and South America mines involving the use of data science, machine learning and integrated functional teams.

The program, aimed at addressing bottlenecks, providing cost benefits and driving improved overall performance, was announced in its December quarter results this week.

It said: “During 2019, FCX (Freeport) advanced initiatives in its North America and South America mining operations to enhance productivity, expand margins and reduce the capital intensity of the business through the utilisation of new technology applications in combination with a more interactive operating structure.”

It said the Bagdad mine (Arizona, USA) pilot program, initiated in late 2018, was “highly successful” in utilising these innovative technologies and it would build on this for the implementation across its other mines in North and South America.

According to a report in the Financial Times, the system at Bagdad found that the mine was producing seven distinct types of ore and that the processing method, which involves flotation, could be adjusted to recover more copper by adjusting the PH level.

The company didn’t provide any details on who it was working with on this project, but confirmed at the back end of 2019 that the Bagdad trial was carried out with management consulting firm McKinsey.

In its investor presentation announcing its December results, the company provided a little more colour on these initiatives.

On the processing/concentration side, it was using a digital twin for processing plant, in tandem with a machine-learning algorithm. These used historical data to predict results and optimise throughput and recovery. In addition to this, the solutions were able to provide “quality recommendations”, aiding real-time data-driven decisions. This allowed the processing teams to target “best performance every day”, while unlocking bottlenecks and providing more consistent operations.

It was a similar story on the mine side. Data is being aggregated from multiple systems to help inform the data-science algorithms to predict the most efficient setups. It also sends commands to dispatch to adjust mining equipment and resource execution, allowing for clear visibility of the best possible performance for shift/day, again, effectively providing real-time decision making.

Under the title “agile way of working”, Freeport said it was promoting a more interactive organisational structure that will challenge norms and identify and prioritise opportunities as part of these initiatives.

CREDIT: Freeport McMoRan

Freeport continued: “A series of action items have been identified, prioritised and are being implemented. Based on the opportunities identified to date, FCX has incorporated higher mining and milling rates in its future plans, resulting in estimated incremental production of approximately 100 MIb (45,359 t) of copper in 2021 and around 200 MIb in 2022.”

Freeport said capital expenditures associated with these initiatives are expected to be “attractive” in relation to developing new copper supply, with the company estimating capital costs – principally associated with mining equipment and ongoing development of data science and machine-learning programs – of some $200 million.

Looking back at the quarterly production figures, it is easy to see the impact this trial had on Bagdad. In the March quarter of 2018, the mine produced 49 MIb of copper, with 48 MIb coming out in the June quarter of that year. It dropped to 45 MIb in the September quarter before stepping up to 57 MIb in the last quarter of that year (when the trial commenced). In the March quarter of 2019, output dipped slightly to 55 MIb, before heading back to 57 MIb in the June quarter and surpassing that (58 MIb) in the September quarter. Output fell back to 48 MIb in the most recent December quarter.

The Bagdad operation consists of a 75,000 t/d concentrator that produces copper and molybdenum concentrate, an SX/EW plant that can produce up to 32 MIb/y of copper cathode from solution generated by low-grade stockpile leaching, and a pressure-leach plant to process molybdenum concentrate.

IMDEX symposium sets the exploration scene for AME Roundup

IMDEX recently held its fifth annual Xploration Technology Symposium in Vancouver, Canada, in which virtual reality, machine learning and new exploration technologies all received significant airtime.

The event, held on January 17, came ahead of AME’s annual Mineral Exploration Roundup, also held in Vancouver, on January 20-23. It saw 16 speakers and 160 attendees turn up.

IMDEX, which has a suite of drilling optimisation products to improve the process of identifying and extracting mineral resources globally, said the event covered multi-element data, artificial intelligence for mineral exploration and exploration instrumentation, along with a series of case studies. The focus was on improving and driving innovation in the mining industry and providing a platform to share big ideas, new technologies and new processes in exploration.

International consulting practice, SRK, had Principal Structural Geologist, Wayne Barnett, present on virtual-mixed reality, where he discussed augmented visual powers to automatically measure surface orientations and how this technology is changing best practices in data collection and analysis, IMDEX said.

Professor Bern Klein, of the University of British Columbia, meanwhile, discussed industry research to optimise value and ensure worker safety in deep underground mass mining operations.

The use of machine learning for mineral exploration in greenfield areas was discussed by GoldSpot Discoveries Corp Chief Operating Officer, Vincent Dube-Bourgeois, during the session on artificial intelligence.

Among the exploration case studies was one from Chris Gallagher, Rogue Geoscience President, a company that has been instrumental in developing several exploration technologies and geological data management systems used in the industry today, according to IMDEX.

And, Nick Payne, Global Product Manager Structural Geology at IMDEX, in his presentation ‘A New Wave of Drilling Optimisation’, discussed new technologies IMDEX COREVIBE and IMDEX XTRACTA – which, he says, offer substantial improvements in safety and productivity.

Minerva to show AME Roundup crowd what TERRA AI software can do

Minerva Intelligence says it will be showcasing its TERRA mining artificial intelligence software at the 2020 AME Roundup Conference next week in Vancouver, Canada.

Minerva’s core competency is combining machine intelligence with human intelligence to produce explainable, rapid conclusions that enable cost-effective decision-making, it says.

Its TERRA suite uses this knowledge to put together a range of software applications that helps “clients harmonise and utilise poorly-structured or legacy data, produces new and precise auditable geological targets for 92 different mineral deposit types, optimises underused 3D drilling data, and provides rapid, intelligent discovery of documents”, it said.

Minerva says it has carried out a number of projects for government agencies focused on generating public domain exploration targets to promote mining within their jurisdictions. It recently updated a project from 2004 carried out in Canada’s Yukon territory, with analysis of the exploration areas highlighted by the project showing a very good correlation with claims held for exploration today, 15 years after the study.

Minerva will be showcasing this technology at Roundup’s Innovation Hub, an area reserved for conference invitees to display the latest innovations in the mineral exploration sphere. Minerva will be demonstrating its advanced augmented reality technology as well as the TERRA product suite at the hub, it said.

The 2020 AME Roundup Conference will be held on January 20-23 at Vancouver’s Convention Center.

Startups Seglico and Miqrotech win I’MNOVATION awards

Startups from Uruguay and the US are due to provide innovative safety and environmental solutions for mining as part of Ennomotive’s Acciona I’MNOVATION program.

The program, which aims to create an impact in industries such as mining, renewable energy, and Smart Cities by solving innovation challenges with the help of startups, closed on November 28, with Miqrotech and Seglico chosen to build a pilot of their technologies after winning.

Over 160 technological startups from all over the world submitted entries, with 20 of them selected as finalists from countries such as Chile, Uruguay, Brazil, Germany, Spain, Canada, the US, and Australia.

Two mining challenges particularly stood out, and their goal was to improve worker’s safety and protect the environment in this industry, Ennomotive said.

The Uruguay-based startup, Seglico, with its occupational safety management solution, was the winner of the challenge about monitoring the health parameters of mining and construction workers. This company has an app that registers in a smartphone the vital signs captured by the worker’s wearables, according to Ennomotive.

The US-based startup, Miqrotech, is to provide a sensorisation solution for tailings and copper concentrates pipes, which has already been successfully implemented in the oil and gas sector. This company, headquartered in Tampa, Florida, uses IoT devices to monitor different parameters such as pressure, temperature, or humidity in the pipes to predict leakage using an AI system, Ennomotive said.

Currently, the winning startups are in the middle of a piloting process that will go on until May 2020 where they will adapt their technologies and undergo real tests on site, according to the organisation.

To read more about the winning startups, follow this link: https://www.ennomotive.com/winning-startups-acciona/

Sandvik showcases digital mining developments in Brisbane

Last week, close to 300 leaders from the mining, construction and quarrying industries from Australia, Japan and Indonesia met in Brisbane, Australia, for a two-day summit, hosted by Sandvik, to showcase best practice examples of digitalisation.

The Digitalization in Mining event, on December 3-4, allowed Sandvik to demonstrate its latest digital offering and introduce participants to the latest innovations across its product portfolio, including process optimisation with OptiMine®, information management through My Sandvik digital services and autonomous operation with AutoMine ̶ together with the latest equipment in underground and surface drilling, loading and hauling, crushing and screening and the rock tools management system.

During the event Sandvik also announced two product launches: AutoMine Access API, which gives mines the power to connect non-Sandvik equipment to AutoMine, and its first Stage V compliant underground loaders for hard-rock mining applications.

Jim Tolley, Vice President, Sales Area Australia Pacific, Sandvik Mining and Rock Technology, said digitalisation is helping companies to grow and optimise their operations. “Our partners were keen to join us at this event because they know that digitalisation has a critical part to play in making their mines sustainable for the future.”

Day one of the event featured speakers from mining companies across Australia, as well as leaders in mining technology, process optimisation and automation. They explained the benefits their organisations have gained by implementing automation and process optimisation solutions, as well as the accompanying change in mindset, according to Sandvik.

The following presentations set the program for the day, followed by a panel discussion:

  • Shaping the Industry Digital Ecosystem (Sandvik);
  • Holistic Perspective, Focusing on Productivity, Safety and Optimised Machine Performance (Byrnecut);
  • Developing the Mine of Tomorrow (Barminco Ltd);
  • Machine Learning  ̶  Keeping it Real with Case Studies from across the Mine Value Chain (PETRA Data Science);
  • Capturing Opportunities for Digital and other Product Technology Solutions (Rio Tinto);
  • Automation Technology to Improve Efficiency and Consistency in Longwall Development Operations (Glencore);
  • Direction of Technology and Automation (Newcrest); and
  • Data Privacy, Rights and Control (Sandvik).

Pat Boniwell, Managing Director, Byrnecut Australia, said the industry will improve productivity, safety and optimise machine performance through a more “fundamental understanding” of the individual processes that make up our operations.

“New technology, automation, data transfer and analysis will all assist us in increasing the utilisation of our resources,” he said. “Data is essential, but if it is not being looked at then we are just gathering data for the sake of it. We need to continue to increase the levels of engagement between all stakeholders.”

He concluded: “We are doomed to failure unless we take our people with us and are prepared to question and be challenged.”

PETRA CEO, Penny Stewart, meanwhile, homed in on machine learning, which, she said, powers “digital twin prediction, simulation and optimisation to increase mine productivity, efficiency and yield, by showing engineers and supervisors how to reproduce their ‘best performance’ 24 hours a day, seven days a week”.

She added: “PETRA’s MAXTA™ Suite digital twin applications provide platform agnostic software-as-service operational decision support across the mine value chain ̶ from resource engineering through to processing plant set point optimisation.”

Day two of the event began with a presentation on sustainability by Henrik Ager (pictured), President, Sandvik Mining and Rock Technology, explaining how critical it is for long-term performance.

“Driving productivity and greenhouse gas efficiency together is going to be key for us at Sandvik, improving productivity and greenhouse gas efficiency will be the best way for us to add value for our customers,” he said. “My view is that the more we link our sustainability targets to normal business targets and find ways to combine them to achieve a common good, the better chance we have to deliver on them.”

Also, during the second day, delegates had the opportunity of a virtual visit to several Sandvik customers, including: Northparkes Mine (Australia), Resolute Mining Syama mine (West Africa), RedBull Powder Company (New Zealand) and Aeris Resources Tritton mine (Australia).

Harry Hardy, General Manager Customer Accounts, Applications Engineering and Marketing, Sandvik Mining and Rock Technology, Sales Area APAC, said the company often gets asked for reference cases and data to illustrate the value and payback of digital solutions. “Over the two days of the conference, our customers were able to share their own experiences and quantitatively demonstrate how our solutions have helped increase their productivity, reduce their production costs and increase their safety.”

GMG helping miners leverage machine learning

The Global Mining Guidelines Group (GMG) has published a new whitepaper that, it hopes, will better equip mining companies to leverage artificial intelligence (AI) and machine-learning technologies.

The Foundations of AI: A Framework for AI in Mining offers an overview of the process of planning for and implementing AI solutions for mining companies, GMG said.

GMG explained: “AI-based innovation is being used increasingly in the mining industry as a means to improve processes and decision-making, derive value from data and increase safety, but the levels of operational maturity are variable across the industry.

“Though many mining stakeholders are adopting AI, there is still uncertainty about the technology and how it can be harnessed in the mining industry.”

This white paper – developed collaboratively through workshops, conference calls and online collaboration tools – addresses a variety of concerns, such as the challenge of establishing data infrastructure, apprehensions about the effect on the workforce and worries about failure after investing substantial time and funds into an AI project, GMG said. “It offers a realistic strategy for building a foundation for planning, implementing and moving forward with AI.”

The primary audience is those in charge of introducing or expanding the use of AI in mining companies, according to GMG.

Rob Johnston, Project Manager at CITIC Pacific Mining and GMG AI Project Leader, says: “There has been a recent explosion in the application of AI in industry, and this document aims to assist mining companies to fully embrace this exciting technology and drive business value.”

Having this information available will also help cut through the hype that surrounds AI, according to GMG.

Andrew Scott, GMG Vice-Chair Working Groups and Principal Innovator at Symbiotic Innovations, said: “Although mining stakeholders generally recognise the value of understanding the technology, many are intimidated by the concept and see expertise in AI as a very specialised knowledge set, so this will help them start off on the right foot.”

This document will also be useful for those who are part of the ecosystem that surrounds mining companies, which comprises those assisting in applying the technology, culture and safety considerations and regulatory frameworks that are necessary for a successful AI strategy, according to GMG.

Speaking from his perspective as a solution provider, Kevin Urbanski, CTO at Rithmik Solutions, says the white paper will provide “current and future customers with a macro view of artificial intelligence and related solutions”, while helping mining operations to “identify opportunities to apply these powerful algorithms within their organisations.”

He added: “Mining companies know best what their needs are, and this document will help them match those needs with what’s possible.”

Urbanski thinks the document will also help to standardise communications around the technology, saying it will “provide great level-setting, ensuring that we and our customers are speaking the same language when talking about AI”.

Johnston, meanwhile, says that while this publication is an important step, the document will be reviewed and updated as needed: “The field of AI moves so fast that this will be a document that will be updated regularly in order to remain relevant to the industry.”