Advancing Process Understanding through Particle-Based Separation Modelling

Mineral beneficiation involves particle-specific processes, where performance is strongly influenced by the microstructural characteristics of individual particles. Ores currently being mined often exhibit high complexity, with fine-scale mineral intergrowths and diverse gangue associations that hinder separation. Within the field of process mineralogy, automated mineralogy has advanced the characterisation of such features, providing detailed particle-level data. However, these data are frequently summarised into bulk distributions (e.g., mineral composition, particle size, or liberation degree of the ore mineral) resulting in the loss of critical particulate information. Since particles are unique in size, shape, and mineral association, reliance on aggregated properties may oversimplify the relation between processing behaviour and ore microstructures, leading to incomplete or biased process interpretations.

In a keynote lecture at next year's Process Mineralogy '26, Lucas Pereira, of the Helmholtz Institute Freiberg for Resource Technology, Germany.  will highlight the use of data science techniques to exploit full-resolution particle datasets from automated mineralogy, namely particle-based separation models. The usefulness of this technology, both for forecasting and for understanding separation processes, will be illustrated by a series of case studies dealing with complex ores.

Lucas Pereira is Group Leader at Helmholtz Institute Freiberg. He develops particle-based separation models that transform high-resolution mineralogical data into practical solutions for industry. His research interests lie at the intersection of digitalisation, data analytics, and process modelling in the minerals sector and he is passionate about exploring the potential of machine learning and data-driven approaches to improve our understanding of complex systems and drive technological advancements.

#ProcessMineralogy26