Like most companies in the process industry, Boliden also sees great potential in being able to use simulations of its complex processes. In order to be able to do this, it is necessary to create faster and more cost-effective models over all or parts of the production processes, which can also be updated in a simple and efficient way.

- The project's idea has therefore been to streamline process modeling by creating a digital twin, which consists of a combination of mathematical formulas and AI. The intention is to be able to use this digital model of the process in the long term to optimize process control, says Johannes Holmberg, CEO of Calejo Industrial Intelligence.

Focused on flotation The project, that was funded by vinnova and PiiA, focused on a limited part of the process step flotation in Boliden's Aitik mine. Flotation is a complex process part, which includes many different types of physical events. The process is continuous, but the conditions change over time linked to shifts in the composition of incoming ore.

The flotation technique is used here to separate the valuable minerals by making them float (float up) in a foam of air bubbles with the help of chemical additives, agitation and air additives.

In order to be able to regulate the flotation cells with a more advanced control, a functioning and fair model is required, which is not only traditionally physical.

- Having a reliable model of your processes is an important step towards a more comprehensive process optimization. Due to its complexity, the flotation process is difficult to model with traditional physical modeling. Parts of the process are also difficult to measure, which limits the possibilities of creating a pure AI model based on historical data, says Johannes Holmberg, Calejo.

Successful combination The company Optimation has for a long time helped Boliden to build physical models over the Aitik mine's various processes.

- As part of the project, Boliden's existing Dymola models, which we at Optimatjon have already developed, were therefore used in combination with Calejo's AI solution. The result was a Dymola model, which makes calculations on input data from the AI ​​model and reproduces information about the flow distribution to the AI ​​model, says Tomas Eriksson, CEO of Optimation.

The strength of the gray box model The project showed that the model with good results can predict the concentration of copper in the foam and the remaining flow two hours ahead. As the model handles both the copper concentration and the flows, this means that it is also possible to obtain a value for the amount of copper that comes out of the flotation bank.

The results clearly show the potential of combining physical (white box) and database (black box) modeling into one and the same model, a so-called gray box model.

- We have together succeeded in demonstrating that it is possible to combine physical and database modeling in one and the same model, a so-called gray box model, and that this can be used for a better understanding of the process and in the long run also contribute to increased optimization when applies to the extraction of copper, says Johannes Holmberg, Calejo, and continues:

- The finished model can also be used to advantage to gain a better understanding of the process, as well as as a basis for optimizing the process control in order to maximize the extraction of copper.

David Degerfeldt, section manager at Boliden, concludes:

- We think that the results from the gray box modeling are very exciting and we are now evaluating how we will be able to use this technology in other process sections as well.

• DIGITAL TWIN
  over the flotation process.
• BETTER UNDERSTANDING
  of the process.
• ENABLES
  increased optimization.