Yerzhan Amanbek

Nazarbayev University School of Mining and Geosciences

1st year Msc student in Mining Engineering, School of Mining and Geosciences, Nazarbayev University.
Previously, worked as analyst in finance and IT fields, 2008-2019.
Bachelor degree in Mathematical and Computer modeling, Kazakh National University named after al-Farabi, 2008.

Geometallurgical Domaining in an Iron Deposit based on Machine Learning Algorithms

Conventional approach of 3D block modeling is mostly based on geological description that divides ore body into sub-domains with similar homogeneous properties. However, this might be impractical when the target is to divide the ore body into the domains based on machine-based parameters, applicable for mineral processing plant optimization. To overcome this impediment, one of the machine learning approaches, so-called "k-means" is applied in this study, to obtain these type of domains (cluster) in an Iron deposit. The aim is to define the areas that demonstrate high content of Iron while their Aluminum and Phosphorus are as low as possible. The results showed that the algorithm is capable of defining these domains, for which they can be used as a framework for identifying the destination of the block to be sent either to mill, stockpile or dump. This will, however, impact saving energy consumption and subsequently will increase the NPV of a mining project.