Abstract:
The depletion of high-grade lump ore and increased blast furnace productivity with the use of prepared burden have resulted in increased use of iron ore sinter and pellets. There is considerable interest in including fine iron ore concentrate and micropellets into raw material mixtures. These materials need to be accommodated in the mixture without adversely affecting the permeability of the sinter bed. Addition of fine concentrate and micropellets to the raw material mixture will significantly affect the particle size distribution of the granules in the sinter feed. Finer materials form less permeable sinter beds and consequently can reduce the productivity of the sinter plant. The most convenient technique of converting fine ores into a useful sinter feed is granulation. This process consists of the mixing of raw materials (iron ore fines, return fines, fluxes and coke breeze) in a drum granulator for several minutes with the addition of moisture. The granules produced are considered semi-products, whose properties must be monitored in order to achieve improved bed permeability in the sintering process. In this thesis, the granule size distribution was predicted via the mathematical model of Litster. The auto-layering process was found to be the main mechanism of granulation, whereby finer material adheres around large particles. Litster’s model was applied to the ore mixtures with four mass fractions of concentrate and micropellets. The results obtained indicated that the predicted size distributions of granules are in good agreement with the experimental data. The analysis of the granule shape is a challenge in iron ore sinter. X-ray microtomography was used to capture the three-dimensional shape of the granules. Different shape parameters were used in the characterisation of the granule shape. The results showed that the shape factor and sphericity increase with the addition of concentrate and micropellets. Zingg’s diagram was also used as a tool to estimate the overall shape of granules from granulated mixtures. The pressure drop through the green granule bed was modelled by coupling Rocky DEM (Discrete Element Method) and Fluent CFD (Computational Fluid Dynamics). Systems of mono and bi-sized glass beads were used to validate the applicability of the developed DEM – CFD model to packed beds in a permeameter. Comparisons resulted in good agreement between the predicted and measured pressure drops for both systems of glass beads. The deviations were within ± 10%. Ergun’s equation could not satisfactorily describe the experimental results, with deviations beyond ± 20%. Coupled DEM – CFD simulations were extended to granulated mixtures that contain concentrate and micropellets. The importance of the stiffness, friction, size distribution and shape of granules was investigated. The size distribution was truncated at 0.5 mm size fraction to reduce the number of DEM particles and computational costs. A parametric analysis of the DEM parameters was also investigated to determine the reliable values of stiffness and adhesion force fraction and particle shape. The stiffness was adjusted for mixtures with addition of concentrate and micropellets. A good agreement between the predicted and measured pressure drops was achieved.
