The value a ‘true’ digital twin could provide the mineral processing and metallurgical industries has been spoken of continuously throughout the last few years, and Metso Outotec believes it is on the cusp of realising such value with its “science-based” Metso Outotec Geminex™ digital twin.

Designed to manage variability and optimise resources, the solution simulates and optimises seamless sustainable operations in minerals, pyro- and hydro-metallurgical processes by combining operational data from both internal and external data sources, the company says.

For simulation and production, Geminex uses the renowned HSC process models that already have over 20,000 users worldwide. These models – which combine versatile chemical, thermodynamic and mineral processing features – have been successfully implemented in hundreds of minerals and metals processing flowsheet development cases, with the same models used in plant run-time optimisation, according to Metso Outotec.

There’s more to this digital twin than the HSC process models alone, according to Veli-Matti Järvinen, Vice President, Automation at Metso Outotec.

Geminex uses “first-principle” dynamic process and equipment models for calibrated performance to help provide an ‘accurate’ digital twin for existing operations, he says.

“Here, both quality and availability of key process data are important,” he told IM, mentioning that normally available instrument and laboratory data tend to be sufficient inputs for generating the digital twin’s simulations.

And Geminex, through a series of “soft sensors”, also produces a vast amount of data that cannot be measured directly yet is leveraged through the dynamic process models, according to Järvinen.

“These soft sensors, as they are called, give an insight to the process that was not available before,” he explained, with examples including element or mineral content by particle size fractions in the process streams. This goes one step further than the feedback from, for instance, Metso Outotec’s Courier® on-stream analysers, which provide real-time elemental analysis measurements.

This makes the digital twin that much more powerful than existing solutions on the market, as it can introduce a new dataset to the equation.

And, as with all ‘true’ digital twins, the models are adapted to live data and continuously improved by machine-learning algorithms, Järvinen said.

“An example of this is model adaptation for variable ore types with different processing performance to help manage variability,” he said. “The resulting metallurgical digital twin is accurate, and its behaviour reflects the physical process.”

All this means Geminex can simulate and test alternative operational scenarios and parameters based on accurate process models and real data, providing the sort of decision-making tools the industry has been after for decades.

While the resource sustainability angle is key here – reinforced by the fact Metso Outotec has labelled Geminex as a Planet Positive product: a collection of the company’s most environmentally efficient technologies – the digital twin’s ability to use resources in an optimal way while considering both impacts and constraints is only a fraction of the industries’ value case.

The ability to incorporate new equipment and tools in a ‘live’ flowsheet that considers the specific characteristics of the orebody at hand and the conditions in which each processing stage receives material is very powerful. One can see it easily aiding the incorporation of new technology in the plant, with the simulations able to provide operators with the confidence to leverage innovations.

Järvinen says the digital twin’s use will also enable the company to start process flowsheet design that much earlier in the mine exploration/development stage.

“As soon as the customer has metallurgical data, the process can be designed and optimised to match the required economic optimum,” he said.

Metso Outotec’s strong experience in developing metallurgical processes, as well as incorporation of the process models of Metso Outotec’s HSC simulation package for minerals processing, pyro- and hydro-metallurgical processes, enables this early analysis.

“These models provide great value in the exploratory phase by enabling scenario analysis, which will help find alternatives for process flowsheets, equipment selection and even blending of different types of ore, if needed,” Järvinen said.

At the same time, Järvinen expects the Geminex digital twin to reduce the plant ramp-up time and “time to market” in the later mining project stages as operators will be that much better prepared for the likes of cold and hot commissioning.

“Great value and impact towards a positive change can be achieved in the run-time of the mining operation and the operating strategy by enriching decisions with the help of Geminex,” he said.

Geminex may be Metso Outotec’s own proprietary digital twin, but, thanks to an extensive back catalogue of process plant modelling references, it is able to run dynamic simulations that incorporate competitor equipment, according to Järvinen.

Similarly, while the company has sustainability and Planet Positive aims for Geminex, the equipment to feature in simulations does not have to be classified as Planet Positive, Järvinen says.

“The processes or assets do not need to be Planet Positive equipment, but those can become such with proper control and optimisation,” he said. “With Metso Outotec Geminex, the full value chain of mineral processing plants and hydro- and pyro-metallurgical plants are considered for Planet Positive production.”

Able to be implemented in modules, Geminex can be deployed piece by piece as part of customers’ digital transformation and continuous improvement projects, he added, opening the possibilities for Metso Outotec to leverage its capabilities beyond full flowsheet design.

In the development phase of Geminex, Metso Outotec carried out several successful pilots with universities and, now commercially available, the company has three ongoing projects in the delivery/commissioning phase with early adopters.

“These projects are already providing good results,” Järvinen said.

“With the good and growing process expertise, we are well prepared to support numerous customers with various optimisation targets in mind.”