������

The Finnish Center for Artificial Intelligence FCAI has started a new industrial collaboration with mining and minerals giant Metso, data visualization systems manufacturer LightningChart and AI analytics platform Quva. The AIMODE project addresses several challenges in the mining industry, such as energy efficiency, sustainability and raw material quality, with artificial intelligence solutions. The three-year project, backed by Business Finland, aims to create new AI methods and tools to better track and optimize each step along the mine-to-market pipeline.

The process from excavating the earth to producing products is long, but for FCAI professor Simo Särkkä, it resembles familiar industrial control systems in factories. “We are dealing with the same multivariate optimization problems, process models and control systems that could be applied to mining or anything else,” says Särkkä. “We are trying to understand the process, and want to do it more efficiently, sustainably and by saving energy.”

Another benefit of artificial intelligence in the mining industry is risk reduction. Intelligent automation can increase safety and reliability, especially in mining where high turnover of staff can create skill and knowledge gaps and monotonous tasks can lead to accidents. AI can also be used to better identify minerals and boost recovery, leading to purer ore and reduced waste.

Mining is a new area for Särkkä and his group members at Aalto University, including project manager Lauri Palva and postdoctoral researcher and mathematician Christos Merkatas, but they are excited to adapt their expertise to an unfamiliar industry. Along with members of the research team of Anssi Laukkanen from VTT, the collaborators are visiting Metso’s customers’ mines, where there are typically hundreds of sensors on-site monitoring every step of the process. New hardware isn’t needed, explains Särkkä. “We can perform more computations from existing sensors, to get better estimates of process performances and other quality indicators.”

The goal is a real-time, AI-infused dashboard of the mining process. Metso already has sophisticated simulators, effectively ‘digital twins’ of the mine, in which different conditions can be tested before real-world changes are made. FCAI’s contribution will be improved AI methodology and implementations. LightningChart and Quva will integrate FCAI’s work into Metso’s existing systems, creating fusion tools.