Abstract:
Anglo American’s Minas Rio iron ore beneficiation plant located in Conceição do Mato Dentro, Minas Gerais, Brazil uses the reverse flotation technique to recover ore while ensuring that majority of the silica (SiO2) is discharged into the tailings dam. Circulating load (CL) samples were sent for testing at the Derrick Screen facility in the USA using the SuperStack™ screening machine with three different aperture sizes i.e. 150 µm, 180 µm and 210 µm. The aim of these tests was to determine the best conditions for the maximum removal of coarse SiO2 in the oversize (O/S) stream. The results from these tests were used for the development of single component (mass) and multicomponent (Fe, SiO2, Gangue) model that predicts the removal of SiO2 from the CL. The use of these models will in turn save money and time for the business. An empirical model development technique was used to develop multi-variate linear regression equations for calculating the following model parameters i.e. alpha, beta, water split and d50c. The beta parameter was fitted because all the efficiency curves for all tests displayed the fishhook effect at finer particle size end. The model parameters for the single component model were: alpha – 5.64, beta – 0.19, d50c – 0.15 mm, & water split – 90.81%. The model parameters for the multi component model were: alpha,Fe – 4.27, alpha,SiO2 – 4.44; beta,Fe – 0.21, beta,Fe – 1.04; d50c,Fe – 0.20 mm, d50c,SiO2 – 0.16 mm. The developed models were tested validated and they closely predict experimental performance.