Modelling and optimization of hydrocyclone for iron ore fines beneficiation — using optical image analysis and iron ore texture classification

Abstract

A new modelling technique for simulating hydrocyclone performance has been developed, in which particles in every size fraction of the feed ore are classified based on ore texture type, taking into account that the same ore texture types in every size fraction of the feed ore have similar mineral contents and densities. Mineral tracking by optical image analysis and newly-developed texture classification software was used in this technique to classify the feed ore particles by texture type and to determine the average particle density of each class in every size fraction. Particle density calculations took into account the reduction of porosity with reduction of particle size and the effect of different imaging magnifications for different size fractions. The data obtained about each class in every size fraction was used to create a virtual feed which was input to the hydrocyclone model to simulate the ore processing performance. For model validation, pilot-scale hydrocyclone beneficiation experiments were performed on an iron ore blend, using different hydrocyclone pressures and percent solids in the feed pulp. Model parameters were determined from one set of experimental results and the calibrated model was then used to predict the outcomes of the two subsequent experiments. Comparisons of the model and experimental results are presented and discussed. This new approach enables prediction of the recovery of each mineral and texture type in the products, calculation of the total product iron grade and recovery, and optimisation of the hydrocyclone performance for a given ore.