Improvement in Roller Screening of Green Iron Ore Pellets by Statistical Analysis and Discrete Element Simulations

Author(s)
B.B. Silva, E. R. da Cunha, L.M. Tavares, R.M. de Carvalho
Publisher
Taylor & Francis
Source
Mineral Processing and Extractive Metallurgy Review
Keywords
classification, Discrete element method, factorial analysis, green iron ore pellets, response surface methodology, roller screen, Simulation
Year
2019

In a recent study by the authors, green pellets motion behavior and interactions among themselves and with the equipment during classification were successfully modeled through simulations with the discrete element method using the Hertz-Mindlin with JKR contact model. The present work initially deals with a parametric simulation study of five important variables in a roller screening operation, namely, gap between the rolls for removal of undersize, rolls rotational frequency, rolls diameter, deck angle, and throughput. The effect of the individual variables and their interactions are analyzed through factorial simulation experiments. Following these initial simulations, a response surface methodology coupled to the application of multiobjective optimization was used to identify a setting of the various variables that would allow improving classification performance. Such results led to a proposition of a new configuration for the roller screen, which consisted in a significant reduction in roll diameter and deck slope, an increase in rolls rotational frequency and a reduction in undersize gap. Simulations predicted, for the studied material, an increase in fines removal efficiency from 91.4% for the base case configuration to 97.1% for the optimized case, with a reduction to one-third the contamination of the on-size material, and a moderate increase in the efficiency of properly sized material reporting to the on-size product. Although at the expense of a modest increase in loss of properly sized pellets to the oversize, such optimized configuration is also predicted to be significantly more robust to surges in the feed than the base case.

Keywords: Discrete element method, green iron ore pellets, factorial analysis, response surface methodology, simulation, classification, roller screen

Access Full Text