PROCESS CONTROL
The IT-OT convergence
Anglo American ’ s IROC in Santiago for Los Bronces
Mining is converging the board room and the control room , as Remote Operations Centres integrate not just operational people but also market people . Paul Moore looks at how this part of mining is developing
More integrated remote operations centres were opened in 2021 – notably in Latin America such as Anglo American ’ s IROC in Santiago for Los Bronces , Antofagasta Minerals ’ IROC in Antofagasta for Centinela plus Codelco ’ s new CIO-E , also in the Santiago , which acts as the central ‘ brain ’ of the company ’ s tactical mining control centres , located in Calama , Los Andes and Rancagua respectively .
But in the bigger picture , how far advanced is the mining industry on its journey towards real time process plant ( and mine ) control , and what still needs to be done to be able to achieve plants that are autonomous and only monitored from these IROCs ?
Jeannette McGill , VP and General Manager of Metals and Mining at AspenTech told IM : “ Each mining company ’ s journey to autonomy looks different . It depends on a lot of factors , including how digitally mature they are . Many are still in the process of migrating away from point solutions that were used for large , fixed assets , toward fully intelligent systems applied across their operations . Isolated tools will need to be replaced with more pervasive options that are able to collect data across all asset classes , offer full coordination across the entire workflow and allow for the integration of moveable and fixed assets that are ‘ talking ’ to the plant .”
She adds : “ Migrations to these systems have accelerated due to pressure driven by COVID-19 and the resulting remote workforce . Automation is now critical and mining companies are increasingly looking for reliable , strategic software partners that have deep expertise in the mining value chain to help them . In some instances , elements of intelligent systems , comprised of technologies like prescriptive maintenance , analytics , and AI are already being used successfully to prevent downtime . For example , prescriptive maintenance on a mobile equipment fleet could proactively identify a problem and allow operators ample time to address it remotely from an on-site control room or a remote centre miles away . With this setup , costly downtime is minimal and , most
importantly , a safety incident is avoided .”
Recently , Emerson and AspenTech have announced that the companies have entered into a definitive agreement to contribute Emerson ’ s industrial software businesses – OSI Inc and its Geological Simulation Software business – to AspenTech to create a diversified , highperformance industrial software leader with greater scale , capabilities and technologies being referred to as “ new AspenTech .” AspenTech is already a major player in mining in terms of its machine learning and predictive analytics offering . Its Mtell software recognises patterns in mineral process equipment operating data that predicts degradation and impending failure – well before it happens . Emerson ’ s industrial automation and control solutions are used in mining to merge scalable , high-speed deterministic controls with analytics .
At a process plant level , machine learning has
come on leaps and bounds . Zeljka Pokrajcic , Technical Director at PETRA Data Science stated : “ For a number of years PETRA has been working on processing decision support solutions using AI and machine learning . At the heart of these solutions are the geological properties of the ore being processed and how it impacts process performance .”
PETRA ’ s flagship product , MAXTA , harnesses the insights and features in historical operating data . It merges geology information from the block model , the fundamental depiction of ore variability , with blasting , load and haul data as well as processing data to form a digital replica , or digital twin , of the whole mining value chain ensuring ore properties are carried through all stages . This digital twin forms the basis of MAXTAProcess prediction and optimisation machine learning models .
“ To better support operators and plant metallurgists in processing set point decisions , MAXTAProcess was implemented in the flotation circuit of a large open pit copper-gold porphyry mine to recommend reagent addition and mass pull rates to maintain low copper tail grade and thereby maximise copper recovery . The flotation copper tail grade was predicted six hours in advance even though the flotation circuit residence time is less than 30 minutes . The machine learning prediction model was heavily dependent on the geological properties of the ore upstream of the flotation circuit and operational set points were matched to the properties of this ore .”
In order to recommend operational set points for best flotation performance , a conventional mathematical optimisation technique was overlaaid on the machine learning prediction model . An historical view of the operating
PI trends with PETRA MAXTAProcess predictions for flotation tail copper grade six hours in advance
28 International Mining | JANUARY 2022