This is a Preprint and has not been peer reviewed. This is version 4 of this Preprint.
Computing single-particle flotation kinetics using automated mineralogy data and machine learning
Downloads
Authors
Lucas Pereira 
,
Max Frenzel,
Duong Huu Hoang,
Raimon Tolosana-Delgado,
Martin Rudolph,
Jens Gutzmer
Abstract
Flotation kinetic studies are essential for predicting, understanding, and optimizing the selective recovery of an ore through flotation. Recently, much effort has been put into incorporating intrinsic ore properties in the understanding of their flotation behavior. Particle-based characterization systems (e.g. automated mineralogy) drove much of this development. However, the currently available methods for flotation kinetic studies cannot accommodate single-particle data, and most of the available data end up not being used. Here we demonstrate a method to fit flotation kinetic models to each single particle characterized in a sample. Our method, based on the lasso-regularized multinomial logistic regression, allows for an in-depth understanding of a particle’s flotation behavior according to every particle-descriptive variable available. We validate the efficiency of our new method in an apatite flotation case study that had been previously studied following traditional approaches. With the proposed method we can show the joint influence of particle size, shape, and modal and surface compositions in the recovery of single particles – such holistic understanding could not be captured before. We expect our method to help developing the field of flotation further and ultimately assist the implementation of more efficient mineral recovery plants – key for a more sustainable use of raw materials.
DOI
https://doi.org/10.31223/osf.io/5tghq
Subjects
Earth Sciences, Engineering, Geology, Mining Engineering, Physical Sciences and Mathematics
Keywords
automated mineralogy, flotation kinetics, geometallurgy, machine learning, particle-tracking
Dates
Published: 2020-08-21 02:29
Last Updated: 2020-10-24 17:20
Older Versions
License
CC BY Attribution 4.0 International
Additional Metadata
Data Availability (Reason not available):
The data used in this study is from the manuscript: "Hoang, D.H., Kupka, N., Peuker, U.A., Rudolph, M., 2018. Flotation study of fine grained carbonaceous sedimentary apatite ore – Challenges in process mineralogy and impact of hydrodynamics. Miner. Eng. 121, 196–204. https://doi.org/10.1016/j.mineng.2018.03.021".
There are no comments or no comments have been made public for this article.