Machine learning for particle size prediction in iron ore grinding process

The mining industry is finally embracing the new opportunities brought by Industry 4.0, and one opportunity to be explored is the use of machine learning and artificial intelligence tools. To meet the continuous global demand for iron ore and improve the performance of facilities, process optimization is essential and may revolutionize the industry with data-based decision-making and increased efficiency and profitability. In iron ore beneficiation plants, the potential gains from machine learning tools are significant, and tend to increase in the milling process, where the grinding phase incurs the highest energy costs among mineral processing operations. This article presents the application of machine learning techniques for the prediction of the main product quality parameter of a milling plant. CRISP-DM, a well-established approach for data science projects, was applied, and the results demonstrated the effectiveness of the model, achieving satisfactory predictive performance. The deployment reduces reliance on laboratory testing and provides energy savings through proactive process control optimization.