Investigating the multidimensional separation behavior of particles in a cyclosizer setting -- A case study on calcite, fluorite and magnesite

TL;DR

This study explores how particles with multiple properties, like shape and size, behave in a complex separation process, revealing that different particle types are separated based on their combined characteristics.

Contribution

It introduces a multidimensional analysis of particle separation behavior using advanced characterization and statistical modeling in a cyclone-based separator system.

Findings

Coarse particles with high roundness are mainly recovered in initial cyclones.

Fine particles with varying roundness are separated in later cyclones.

Multidimensional analysis reveals property-dependent separation patterns.

Abstract

Particle separation is typically investigated regarding one particulate property only. Virtually all separation processes, however, act on various particle properties in different ways. Modern particle analytical modalities enable a statistically meaningful multidimensional particle characterization. Within this study, individual particle fractions of magnesite, calcite and fluorite (-71 $\mu m$) are processed via the turbulent cross-flow separator cascade Cyclosizer (M16, MARC Technologies Pty Ltd), consisting of 5 hydrocyclones, thus producing 5 different product streams. Particle characterization is achieved via dynamic image analysis from which information on the particle shape and size is obtained. Using this data, bivariate Tromp functions are computed, which show the combined effect of the particle descriptors of roundness and area-equivalent diameter on the separation behavior. While the first cyclones recover predominantly coarse particles with high roundness values, fine particles with varying roundness are recovered in the latter cyclones.