Modelamento do índice de tamboramento de pelotas de minério de ferro para redução direta

• Ano de defesa: 2016
• Autor(a) principal: Luis Alberto Sfalsin Passos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Materiais e de Minas; Engenharia Metalúrgica; Engenharia metalúrgica
• Link de acesso: http://hdl.handle.net/1843/BUBD-AC8H5G

Over the years, a change is observed in the mineralogical characteristics of iron ore mined in new mines of Samarco SA. According to surveys, changes can be observed in the percentage of porous hematite, specular hematite and goethite in the constitution of iron ore concentrate. We observed a trend of increased percentage of goethite and LOI (loss on ignition) intrinsic characteristics of the iron ore. Therefore, these facts imply greater difficulty to concentrating and ensuring the quality of the iron ore slurry pumped to the pellet plants, since factors such as particle size (particles smaller than 325 #) and the specific surface of ore are affected by mineralogy. Consequently, there is a greater difficulty in the filtering process to obtain the pellet feed with a moisture range required for the process of pelletizing and firing. On the other hands, when increasing the LOI it can be difficult the drying process of pellets in the pelletizing furnace, requiring a greater thermal input to remove the residual water contained in the crystals that do not evaporate at 1000. All these factors contribute to the quality of burned pellets, and characteristics as cold compression and tumbler. However, the fired pellets of iron ore should resist handling during the transport process to the end customer, without suffer degradation. In this context, the tumbler index is one characteristic that must be monitored during pellet production process. The tumbler index measures the percentage of fines generated by a given sample of burned pellets when subjected to bearings within a tumbler of standardized dimensions, and for this reason is called as tumbler test. This study aimed to identify the variables that most influence the tumbler index through developed a dynamic model based on the metallurgy of the pelletizing process and the statistical theory of multiple linear regression. Based on this principles and application of DMAIC methodology, the process variables candidates for the model, were identified, treated and standardized. The database formed was used to generate a model for the tumbler index called as field model. The field model was validated according as the assumptions assigned to the errors associated with the linear regression model. The direct analysis of the magnitude of the model coefficients showed the importance of variables such as: downdraft drying wind box temperature, travelling grate speed and the pellets size distribution, for the tumbler index control. Subsequently a new model with fewer variables was generated in the laboratory from the pelletizing tests realized in pot grate. The laboratory model was built to assert the main relationships identified by the field model however in a controlled environment. In addition, the laboratory model allowed the observation of other factors that directly influence the growth of the green pellets changing the size distribution therefore suggesting future studies.