PROCESS CONTROL
the best algorithms and refine how these algorithms work to deliver a unique AI system that has been designed to control and optimise industrial processes . This is very complex and requires some of the latest technologies in AI to be able to achieve the necessary control . We have also developed a system that can uniquely be applied in stages as data quality improves and operator confidence increases – going from making recommendations to fully autonomous control at the speed our customers want . This allows customers to build confidence before committing to fully autonomous control . We have chosen not to patent our solution as it would have exposed the technology , however , we would argue it is very unique .”
He said the system is also capable of selfadaptation – a form of prescriptive analysis - which can be achieved when you have both the right data and the right AI models . “ The AI creates a dynamics model that is made from both historical and ‘ live ’ operational data . It can then choose the set of control parameters that deliver optimum performance towards a specific goal , eg maximum production for minimum energy usage . In comparison to traditional supervisory control systems it adapts to changes in the process . For example , if the process changes for some reason , leading to new data points , the AI incorporates these into its model , new predictions are made , and new control parameter settings are used . New goals can also be set , resulting in SentianController choosing the best control parameters to achieve those goals .”
Metso ’ s new Performance Center capabilities
In April 2023 , IM toured Metso ’ s state of the art Performance Center in Santiago , Chile with Pablo Zuniga – Metso ’ s Remote Services Manager . The facility is just a small part of Metso ’ s office space in the city but an increasingly important one as mining houses look to maximise efficiency using the latest available technology .
Metso started to build the Performance Center back in 2019 , with all the construction completed between April and mid-year 2020 . Zuniga joined in January 2020 – having spent about 12 years working with mining companies in processing plant maintenance roles – first with Codelco and then with Antofagasta Minerals . He told IM : “ I am an electronic engineer – and started in mining as an instrumentation supervisor in the Codelco Andina concentrator plant . I then moved into a reliability management role at that operation – before moving to AMSA ’ s Antucoya where I held the role of Reliability Superintendent . So , my experience has involved working with concentrator plant equipment and specifically maintenance and reliability aspects . This involved looking at solutions related to preventive , predictive and prescriptive maintenance .”
He explained that in mining concentrator plants you can have the control systems , advanced process control ( APC ), model process control and overall automation systems . “ But you also have individual areas of automation – plus additional tools – such as using machine learning or instrumentation to monitor asset health in order to try and predict what is going to happen based on historical data as well as the original equipment specifications , set points and limits which of course as the equipment supplier we are the most knowledgeable in . You can start to map the behaviour of a piece of equipment in this way .”
What you are trying to do is detect specific failures and for that you need also expert information that includes knowing the limits for each kind of signal . For predictive maintenance you must find a way to predict failures – to some extent you know how it is going to behave as you have experts with deep equipment experience as well as the manuals with all the information and values related to the equipment .
He continued : “ But you can then add a layer – using machine learning and cloud computing to get information from the site with different IoT tools . You get signals from the site , upload those signals to the cloud and to some kind of hardwired infrastructure to process that raw data . At that point you do need the historical data to show examples of the behaviour of the equipment in different situations , such as a reaction to changes in ore chemistry , and build up algorithm-based models that are going to predict a failure without having to have a human expert monitoring it . But it doesn ’ t rely on algorithms alone – you also have to almost ‘ teach ’ the model for each mode of failure .”
Metso is building these models for diverse types and models of equipment at the Performance Center . Once the model is in place it has two things to offer the customer – it of course has human experts remotely monitoring data to predict and prevent problems and failures . “ Then you also have the machine learning based models looking for potential failures 24 / 7 . The expert is doing the prescriptive part – you can predict a failure but then the expert validates it and then based on their experience decides what they believe is happening and why , and then opts for the right action or response for the site to avoid that failure . This is the essence of prescriptive maintenance .”
So why did Metso need a dedicated Performance Center facility and was it driven by Metso or by the customer ? “ I would say initially it was Metso-driven – as we were looking to develop these kinds of remote services based on machine learning and cloud computing . We wanted to use digital intelligence to help the customer increase efficiency in a better way – use more technology to assist the client combined with our expert knowledge in our own equipment . The establishment of a Performance Center was needed as it would be hard to conduct prescriptive maintenance on the customer ’ s own site . They are more focused on operations and scheduled maintenance rather than detailed signal analysis .”
Also having a centrally located Performance Center in the city enables Metso to be more inclusive – it can have expert engineers based there that for whatever reason are not physically able to go to the actual sites due to factors such as age , gender , or disability . “ For example , we have engineers here that are in their 70s but with unique experience . At the other end of the scale young people are much more likely to stay with Metso long term if they are not having to live on remote mining sites .”
Santiago ’ s Performance Center is complemented by Metso Performance Centers in Espoo , Finland and Changsha , China . In Santiago , there are currently 11 engineers supporting the Chile Center – some of them are not based in the Performance Center but provide remote infrastructure and software development or programming support . The human expert part of the monitoring is during normal Chilean day shift hours , not 24 / 7 , but the technical platform and tools are operating all the time . The experts are divided by technology , not by mine – so currently there are crusher and screen experts , plus a mill expert for both horizontal and vertical mills , as well as an analyser expert monitoring Metso Courier and PSI on-stream analysers . The Center has also begun monitoring Metso pumps and grinding mills at customer mines .
IM asked Zuniga for examples of failures or failure predictions that have been identified for mining clients . “ A common one is related to dust and fines in the crushing area which can start to cause oil contamination in the crusher if the hydraulic system is not very well sealed . This contamination can eventually damage the equipment . Using the pressure of the hydraulic system , and the differential pressure at the filter , we can detect when this type of contamination is increasing before it gets to a critical stage . The customer can then change the filter to avoid equipment stoppages later . A failure is always more costly than a planned shutdown .” From the Performance Center , Metso is monitoring large cone crushers including the MP800 , MP1000 , MP1250 and MP2500 .
For Courier analysers , the Performance Center is doing online calibration . “ When you are operating a flotation plant you need to know what the percentage of copper is in the cell feed and the overflow froth . This information allows
66 International Mining | JANUARY 2024