Multi-response and Misplacement Optimization of Upgrading BMQ Ore in Liquid-Solid Fluidized Bed Separator Using Taguchi-based Grey Relational Analysis

Abstract

Liquid-solid fluidized bed separators (LSFBS) are particles handling equipment where solid particles are caused to exhibit liquid-like behavior. Segregation and separation of suspended particles are achieved by maintaining superficial velocity of upward flowing fluid between settling velocity and terminal velocity of particles. Rapid increase in the consumption rate of metals and ores in the last few decades necessitates effective utilization of marginal, low-grade ores. This work focuses on the ability of LSFBS in upgrading lean-grade banded magnetite quartzite (BMQ) ore. To gain an understanding of segregation, separation, and layer inversion phenomena inside the fluidization column, hydrodynamics characteristics, process efficiency, and misplacement of solid particles have been studied qualitatively and quantitatively. Feed, concentrate, and tailing samples were analyzed using wet chemical method, scanning electron microscopy (SEM), and X-ray diffraction (XRD) analysis. Taguchi-based grey relational analysis (GRA) and one-way ANOVA were employed to statistically analyze the fluidization process and provide ranking of input parameters for individual and multiple responses synchronically. In order to characterize misplacement, fractions of iron-phase minerals and gangue minerals reporting in overflow (tailings) and underflow (concentrate) have been calculated. The best rank indicating multi-response optimization and the highest normalized misplacement index were obtained at superficial velocity: 1.41 cm/s, mean particle size: 125 mu m, overflow tap height: 62 cm, and bed height: 20 cm. The highest values of normalized misplacement index and misplacement index obtained are 28.82975% and - 0.5979, respectively. This suggests that minimum misplacement and maximum process efficiency can be achieved at this optimized set of experimental conditions.