14 minute read
AUTOMATION IN ESG FRAMEWORK:
HOW AUTOMATION CAN CONTRIBUTE TO THE SUSTAINABILITY OF MINERAL RESOURCES SUPPLY
Technological advancements are paving the way for the mining industry to adopt smart technology and automation, ultimately leading to the adoption of Industry 4.0. Professor Mohsen Yahyaei is an expert in process modelling, optimisation, and control of mineral processing circuits working as the director of Julius Kruttschnitt Mineral Research Centre (JKMRC) to bring new technology to the major mining companies he works with.
Advertisement
Professor Yahyaei gave insight on the autonomous systems utilised in the industry, the potential risks involved with employing this technology, and some of the work he’s undertaking through JKMRC.
What is automation in the mining industry?
Professor Yahyaei said that when it comes to the topic of autonomous mining it is important to understand and distinguish between the various levels of automation.
“Automation at its basic level starts with some function-specific tasks being automated, and decision support is provided while the operator maintains control over the process.
“On the other extreme, a fully autonomous system is the state that, in all situations, an automated system makes all decisions and deals with unusual conditions without human operators overseeing the system or process.”
He said that a fully autonomous system is the ultimate target when talking about automation, and that even though the mining industry has implemented automation for the past 25 years in its various business processes, the state of automation varies between operations and companies.
“Some companies are well and truly advanced in automating their processes to the extent that they use autonomous systems and have the infrastructure to manage and operate their processes remotely.
“However, the level of automation has yet to progress evenly across all sections of the operation and also it has a long way to go to a fully autonomous system.”
Despite this, Professor Yahyaei said mining and extraction are ahead of the other sections of the operation regarding automation.
“There has been an acceleration in the adaptation of automation technologies in the mining industry with increased attention to Industry 4.0.
“In mineral processing – which is my main area of focus – Advanced Process Control (APC) and Model Predictive Control (MPC) have developed significantly over the past decade.
“Most vendors have equipped their process control platforms with APC, and MPC features.”
According to Professor Yahyaei, many companies and operations have progressed in their journey toward advanced process control for their mineral processing plants.
The value of automation
Professor Yahyaei said there is a general misconception that autonomous systems are primarily a means for reducing the cost of the workforce and, consequently, the cost of production.
He also detailed some of the ways automation of processes could deliver value to the industry, including:
♦ Stable operation, leading to higher overall production, consistent product quality and improved productivity
♦ Reliable and predictable processes, enabling targeted and accurate economic optimisation of the operation
♦ Optimum operating conditions, increasing equipment life and consequently reducing operating costs and downtime
♦ Reliable data on water, energy, emissions and consumables usage, which can be used for managing and forecasting the costs and optimising utilisation
♦ Autonomy of more mundane tasks, allowing operators to focus on more rewarding and higher-skilled tasks
♦ Use of autonomous units in hazardous environments, improving safety
Automation and ensuring a sustainable supply
Professor Yahyaei expects that automation will have a significant impact on the sustainability of mineral resources production and supply.
“The key aspect of the sustainable supply of minerals to the developing society is ensuring that all impacts of the mining activities on the environment and society are measured transparently and managed according to the frameworks developed as part of the United Nations’ Sustainable Development Goals.”
He emphasised that reliable and consistent data and transparency are inherent features of automation because it will not be trusted and achieved without them. As such, automation could be the foundation for providing consistent and transparent data needed to measure and manage a sustainable supply of resources.
“On top of that, automation is the means for reducing the losses and inefficiency of the mining process and ensuring maximum utilisation of environmental resources.”
Automation’s influence on mining’s ESG
One of the other areas mining in Australia can expect to see changes created by automation is in the industry’s environmental, social, and governance (ESG) goals.
“Process autonomy is a key tactic for the minerals industry to achieve its ambitious environmental, social, and governance (ESG) goals set for 2030 and 2050,” Professor Yahyaei said.
“The mining process is energy intensive. Mining processes utilise environmental resources significantly and generally have a large carbon and environmental footprint.
“Therefore, autonomy for mining operations presents a considerable opportunity based on the advantages mentioned earlier.”
In regard to the social impact of automation, however, we still have a long way to understand it fully.
“We still need to investigate and fully understand all aspects of the impacts of automation on social aspects and then incorporate those in all aspects of conceptualisation, design and implementation of automation in the industry.
“In my opinion, this is one of the big gaps in the current state of automation in our industry.”
Changing the landscape of the future workforce
Over the last few decades the mining industry has become more complex and multi-disciplinary as the scale of the industry has increased.
According to Professor Yahyaei, the automation of the mining industry has added to the complexity of the business on the one hand, but has also reduced the complexity of most routine tasks. He said these changes will significantly impact the landscape of the future workforce and have already impacted the current workforce.
“The skills and capabilities required for the workforce in the industry are transitioning from specific and targeted technical knowledge and experiences toward diverse and wide technological capabilities and skills and capabilities to perform in a multi-disciplinary and complex environment with a strong emphasis on soft skills.
“Given that most manual and routine tasks benefit from automation, operators need to know how to run and manage automated technologies rather than the mining process itself.
“I understand many of the scholars in the mining industry might disagree with this view, but our industry is changing the same way that computer has changed the landscape of science education. For instance, manually conducting mathematical calculations or statistical analysis is no longer required, and we use computers for most of those.
“One should only understand the concept and know what tool to use. In fact, some software packages are smart enough to identify and suggest an appropriate analysis tool based on the problem's definition.”
He said the landscape in the business is changing in the same fashion.
“In most cases, a detailed understanding of the processes and interdependency between parameters is not essential, and most of those are taken care of by process automation.”
According to Professor Yahyaei, it is, however, essential for the workforce to implement high-level thinking to identify bottlenecks and manage the available resources to remove them.
“This aspect was one of the operator’s duties before automation as well, but by automating most processes, the nature of this task has changed, and the focus of debottlenecking efforts is across the whole operation rather than parts of the process, and also has a significantly faster pace to cope with the speed of automated systems.”
Professor Yahyaei said the current and future workforce should work comfortably with technology, only focusing on the capabilities and features that advanced technologies offer and without understanding the complexity of what is beneath.
“A holistic view and understanding of the operation is also a key capability that the future workforce should demonstrate. The workforce level will influence the essential skills.
“The skills of people developing advanced technologies will differ from those who will manage the technology, and people who will operate the technology will require different skill sets. There will be some skill overlaps, but skill sets will differ for each role.”
This is an aspect that Professor Yahyaei said our education system is missing.
“Recognising the skill sets required for various roles and bringing a multidisciplinary approach to training and education of the future workforce beyond the traditional disciplines is a must for a successful transition to an automated industry, and our education system is lagging behind.”
According to Professor Yahyaei, one of the underlying reasons for the inefficiencies observable in the implementation and adaptation of automation in the mining industry is how the current workforce is educated.
“There is a big gap in this aspect, and it requires a collective effort of the industry, private and public education sectors and governments to address the gap.“
The inherent risks of automation
In contrast to automation’s advantages of reducing the risks for mining operation by removing the workforce from hazardous environments and reducing risks through consistency and coherence in the decision-making process as well as accessibility to necessary data, Professor Yahyaei said automation imposes new risks.
“Automation is a double edge sword when it comes to risk.
“One of the major risks is when automation is done partially, and there is a mix of manual and automated processes within an operation. This situation will increase the risk of wrong decisions and incidents due to high uncertainty and a more dynamic environment.
“The uncertainty decreases when the operation is fully automated, and the process becomes more predictable.
“Another aspect of risk with implementing automation is the scale and size of damage in case of incidents which is expected to be larger due to the speed and scale of automated processes.”
Professor Yahyaei said that although automation has the potential to reduce risk, the risks must be fully evaluated and managed during the design, implementation, and integration into the process.
“When a process is automated, it will change the processes of the automated section and many other sections affected. Therefore, a thorough review of the processes and risk assessment and management is essential to make the risk of automation manageable.
“The risk aspect of automation also has many gaps and requires better understanding and we are actively working on this topic in collaboration with world-leading experts.”
The Advanced Process Prediction and Control research group at JKMRC
The Advanced Process Prediction and Control research group was established in 2017 at Julius Kruttschnitt Mineral Research Centre (JKMRC).
Professor Yahyaei said the goal of the group was to build on a long history of JKMRC in developing phenomenological process models and develop solutions for real-time process prediction and control.
“Most of the process models developed in JKMRC are in application for process design, process diagnosis and optimisation by plant metallurgists, consultants and engineering firms and they are tested and validated extensively with industrial data.
“However, we realised that with the increased emphasis on transitioning to Industry 4.0 in mining, we need to develop and offer solutions to enable advanced process control.
“Our research group pioneers developing solutions for real-time process prediction and advanced control based on validated semiphenomenological models.”
Developing dynamic models for critical processes such as materials handling and storage – which significantly impact process stability – is another aspect that the research group is leading in the industry.
The dynamic materials storage model that Professor Yahyaei’s research group has developed is unique in the sense that it calibrates to a small-scale laboratory test, and it can predict material size segregation. The model is also validated with industrial data.
“We also took a unique approach in leveraging the power of Machine Learning (ML) and Artificial Intelligence (AI) to guide our mathematical models and avoid using them as black box models for modelling mineral processing processes for which we have robust and validated semi-phenomenological models.”
In the past two years, the research group has commenced research in utilising ML for advanced process control by learning the complexity of interactions between various parts of the process and guiding operators for best debottlenecking strategies.
Developing soft sensors – which are mathematical models to calculate and infer parameters that are not practically possible to measure directly – is another area of focus and an aspect which Professor Yahyaei said the group has been “leading the topic by developing and offering soft sensors to the industry”.
“Our Mill Filling Inference Tool (JKMRC Mill FIT) is a soft sensor which has been very successful and its application in the industry is growing rapidly.
“We have a long list of soft sensors under development which provide more visibility across the mineral processing circuit and not only assists plant metallurgists in their decision-making, but also improves reliability of the data collected from hard sensors.”
According to Professor Yahyaei, addressing the issue of trust in automation by filling the gaps in data accuracy, risk analysis and management, and improving decisionmaking support systems by using semiphenomenological models is a new line of multidisciplinary and collaborative research in the research group.
Unpacking the JKMRC Mill FIT
The JKMRC Mill Filling Inference Tool is a collection of validated models that use the existing measurements for the operation of tumbling grinding mills to infer the total and ball filling inside the mill. Those two parameters are not practical to measure directly.
Operators usually use their experience to infer the mill total filling based on the power draw and trunnion’s bearing pressure or load cell measurement (if it exists).
“For measuring the ball load, operations should do a regular mill grind out – a process in which the feed to the mill is stopped, and the mill rotates until all rock charge is grounded and leaves the mill. Then the mill is stopped, and the ball charge inside the mill is measured directly,” Professor Yahyaei said.
The traditional method heavily relies on operators’ experience and requires frequent interruption of the process to correct the assumptions using direct measurements.
Professor Yahyaei said the JKMRC Mill FIT provides the total and ball filling of the mill in real time using the power draw and the mill bearing pressure or load cell.
“The tool as a soft sensor has many advantages, including no need for installation of new hardware, and the possibility of developing and integrating the tool into the process control system remotely and without interruption.
“Operations realise the tool’s value through more stable operation, which is because of accurate knowledge of the mill total and ball filling, less process interruption because of eliminating the need for frequent mill grind outs.
“Furthermore, because operators can accurately estimate the mill total and ball filling, they can operate the mill without any conservation, due to the fear of damaging the mill liners or excessive liner and grinding media wear.”
Professor Yahyaei said the group’s data demonstrates that operations could reduce the mill throughput ramp-up time after each mill shell reline and reduce production loss, which is very common after mill shell relines.
The JKMRC Mill FIT is one of the examples of the JKMRC approach in converting semiphenomenological process models into soft sensors for inferring important operating parameters that are not practical to measure but are important for real-time optimisation.
“We have developed a soft sensor for bins and stockpiles, which provides a real-time estimation of live capacity above each feeder and also an estimation of the average size distribution of materials above each feeder,” Professor Yahyaei said.
“There is also a soft sensor for monitoring hydrocyclone’s s performance (i.e. JKMRC CycloPS) which is developed to provide a real-time estimation of the mass split, water split to overflow and underflow streams. The soft sensor can also estimate the size distribution of overflow and underflow.”
JKMRC’s role in the future of the mining industry
Professor Yahyaei described JKMRC as “a world-leading research centre that has developed and delivered process models for comminution, classification and flotation processes to the minerals industry. All models are developed based on understanding the fundamentals of each process and are validated with a wide range of laboratory and industrial data.”
Building on its established track record, JKMRC researchers and students will work with world-renowned academics at Sustainable Minerals Institute (SMI) and the University of Queensland (UQ)schools, as well as international scholars on emerging challenges of the mineral industry.
The JKMRC’s aim is to develop and deliver technological innovation within a responsible ESG framework to the minerals industry.
The centre is uniquely positioned to pursue its aims thanks to JKMRC’s connections with SMI’s five other Research Centres and UQ’s Schools and Faculties.
The centre’s research focuses on identifying new processing technologies and developing semi-phenomenological process models.
“Moving on from the steady-state models previously developed at the JKMRC, the centre creates and applies new techniques that make greater use of data generated on-site and sensor technologies in combination with advanced process control, computational analytics and modelling techniques,” Professor Yahyaei said.
“The JKMRC research will offer practical tools to the minerals industry that are essential to achieve a sustainable supply of minerals to the developing society.”