Estimation of Ore Mineralogy from Analytical Analysis of Iron Ore

"A pressing issue that has often been the topic of discussion in iron-making operations in India is the high content of alumina in the country’s iron ore deposits. The presence of alumina has hampered production by affecting slag fluidity, and thereby the hot metal quality. Iron ores from Indian mines are generally rich in hematite, with traces or small amounts of goethite/limonite, kaolinite, gibbsite and quartz reported from time to time. In day-to-day production, emphasis is placed on checking the chemical analysis of product samples, but the true solution to the alumina problem lies in estimating the mineralogy of the ore samples. We developed a mathematical model, based on singular value decomposition, to predict mineralogy from classical chemical analysis. We then used X-ray diffraction patterns, scanning electron microscope analysis, and grain counting with an optical microscope to determine actual mineral compositions and to test the values predicted by the model. The results indicate that with regular calibration, the model can be used as a simple, fast and effective tool to predict on a regular basis the mineralogy of multiple samples.IntroductionIndian blast furnaces use high-grade iron ores (as lumps or as sinter) and coke (obtained from the destructive distillation of coal) as raw materials for the production of high-quality hot metal to be converted into different steel products through the Linz- Donawitz process. The iron ores are mainly sourced from the Singhbhum region of Odisha and Jharkhand states, the Bailadialla region of Chattisgarh state and the Bellary region of Karnataka and Goa states, and are mainly classified based on size as (1) lump ore (fed directly into blast furnaces), (2) fines (converted into sinter/pellets and fed to blast furnaces) and (3) slimes (< 0.15 mm; currently dumped in huge slime ponds). Any variation in the quality of the ores supplied results in variations in the production trends of hot metal."