Fine Iron Particle Recovery Using the Reflux Classifier: Assessment of the Technology at Pilot and Lab-Scale

"In the iron ore industry, fine particles recovery is an important concern in terms of process performance and production costs. Continuous efforts are thus devoted to increase the efficiency in handling these fines.The Reflux Classifier (RC) is a relatively new technology that is mainly used in the coal and industrial minerals industries. The operating principle of this system is based on three mechanisms: hindered settling, autogenous dense medium and lamella settling. A preliminary evaluation of the Reflux Classifier technology for the beneficiation of iron ore fines was formerly done at the pilot scale using a RCTM300 unit. The good results obtained, using iron ore fines from screened spiral stream, made this technology very attractive for the iron industry. However, the high capacity of the RCTM300 unit (1–3 t/h), makes it unsuitable for R&D purposes. Thus, a smaller unit, the RCTM100, was assessed to validate the performances obtained with the RCTM300.This paper summarizes the performances obtained with the RCTM100. The challenges of the scale-down related to the adaptability of this technology to the iron ore industry are also presented. The results are compared to the RCTM300 performances, as well as benchmarked against sink-float tests. The metallurgical results on feed and products issued after the test work with the RCTM100 unit are supported by mineralogical characterization using a Mineral Liberation Analyzer.INTRODUCTIONFine particles recovery by physical separation means is a historical concern (Holland-Batt, 1992; Peukert and Wadenpohl, 2001; Srivastavaa et al., 2001; Majumder et al., 2003). Attempts to improve fine particle recovery at the laboratory, pilot and industrial scales adopting different strategies with traditional equipment or assessing new promising devices such as the MGS (multi-gravity separator), the Kelsey jig, the Yang jig, etc. (Kim, 2002; Goudreau, 2008; Kerr, 2012) have yielded variable success."