Diagnosis of the Operation of a Hydraulic Classifier for Iron Ore Concentrate

"ArcelorMittal Mines Canada operates hydraulic classifiers to reduce the silica content of an iron oxide concentrate from 4.5% to less than 1.2% SiO2. Sampling campaigns were conducted to develop a preliminary model for the hydraulic classifier. Results show that iron oxide and quartz are respectively concentrated in the coarse and fine size fractions of the classifier feed. The mineral partition curves in the classifier and the size distribution of the minerals in the feed are used to simulate the classifier operation in order to identify the source of variability observed in the daily plant production data. Results indicate that most of the variation is probably caused by unsteady operating conditions of the classifiers.INTRODUCTION Hydraulic classifiers are a class of gravity separation devices (Wills, 1992) used to classify particles according to their size and specific gravity. Hydraulic classifiers also known as density separators (Hearn, 2002) combine hindered settling and upward flow of water to selectively entrain light particles to an overflow stream. Denser particles settle to an underflow stream. Hydraulic classifiers are used for various purposes ranging from cleaning of coal (Das, Sarkar, & Mehrotra, 2009), recovery of chromite from plant tailings (Kumar, Tripathy, & Rao, 2009) to upgrading iron oxide concentrate (Hearn, 2002). ArcelorMittal Mines Canada operates several hydraulic classifiers to upgrade an iron oxide concentrate by removing fine quartz and other light minerals from hematite that is the main iron oxide mineral. A project was initiated to analyze the operation of the classifier in an attempt to identify the source of the variability in the production results and to assess the potential of increasing iron recovery in the classifier. One possible disturbance for the classifier operation is the change in the classifier feed composition and size distribution, while the other considered disturbance is the operation of the classifier. The paper is divided into five sections. The first section summarizes the ore processing sequence for the production of iron oxide pellets to identify the position of the hydraulic classifier in the process. The second section describes the hydraulic classifier. The third section discusses the sampling campaigns conducted on the hydraulic classifier. The method used to process the data from sampling, the results of the data processing and simulation of the classifier are discussed in the remaining sections."