Computational Modelling of Melting, Solidification, and Slag Dynamics in Sustainable Ironmaking

Abstract

This thesis supports the transition towards fossil-free steelmaking by investigating two innovative processes: steelmaking with hydrogen direct reduced iron (H-DRI), and the IronArc process. Both processes aim to replace the blast furnace based primary steel production with less CO₂ intensive alternatives, H-DRI production in shaft furnace coupled with electric arc furnace melting, and the IronArc process with electrical heating and melting of iron ore by plasma generators and reduction with liquid natural gas. To facilitate the implementation of these processes, five supplements use experimental work and numerical modeling to examine flowing slag to elucidate the effect of its material properties on refractory wear, formation of freeze-linings, melting of hydrogen-direct reduced iron, its infiltration into porous structures, heat transfer, and gas blowing. The IronArc process involves a flowing slag with 90 % FeO, which is expected to be very corrosive to most common refractory materials. To facilitate the process the possibility of formation of a freeze-lining over the refractory wall was investigated. It was found that it is possible to protect the refractory from the corrosive effect of the slag by a freeze-lining, and that it requires a high amount of cooling and a slow flow of slag of a maximum of 3 kg s^-1 to maintain proper coverage. This limits the production capacity to ca 100 000 t/year. Additionally, the resistance to wear from such a slag was evaluated in high temperature experiments for a set of refractory materials. The experiments were combined with thermodynamic calculations of the stability of the refractories, and it was found that thermodynamic equilibrium calculations can be used to predict the stability of refractory materials in contact with corrosive slags by studying if any of the constituent phases of the refractory are stable at the experimental temperature in a system with the slag composition added. The work found that only a MgO-spinel type refractory was able to offer partial resistance to the slag, and that a high temperature sintering of the refractory was beneficial to its resistance and structural integrity. The behavior of the superheated gas jet from a plasma generator was also evaluated using compressible numerical simulations. It was found that a submerged jet requires a Froude number of at least 300 in the IronArc slag to be mostly stable, and that the limit of the Froude number is dependent on the density ratio of injected gas to liquid. To ensure stability of a jet and predictability of the penetration length using empirical equations, the Froude number of the flow must be sufficiently high, which can be achieved by reducing the inlet diameter of the gas injection. If the stability of the jet is not maintained, significant pulsations and bubbling may occur, which risks bringing the injected gas in contact with the refractory wall, which would increase the risk of wear and breakthrough. The melting of H-DRI in the EAF relies on short melting times to avoid the buildup of large agglomerates of unmelted material called ferrobergs. To facilitate quick melting, the effect of slag flow, temperature, infiltration, and material properties was evaluated. It was found that the melting time of H-DRI is significantly affected by the slag temperature in proportion to its superheat, and the forced flow of slag. The shortest melting time is achieved when combining high temperature slag (>1950 K) with low viscosity foam (<0.1 Pa s), high thermal conductivity (>0.9 W m^-1 K^-1), and high flowrates (>0.1 m s^-1). This results in melting times as low as 4.56 s, which would allow for complete melting of H-DRI during its fall through a foaming slag. The metallization and porosity of H-DRI appear to only have a minor effect on melting time as compared to the slag conditions, as it is the heat transfer coefficient to the H-DRI which is the controlling factor for melting time. The infiltration of slag into the porous structure was simulated as capillary rise of hot slag and was found to be limited by the temperature of the iron matrix to such a degree that the slag infiltration cannot propagate further than the isotemperature line of 1523 K into the H-DRI without solidifying. Additionally, the convective contribution of heat from the inflowing slag was found to be a very small fraction of the heat required for melting, and that slag infiltration does not necessarily result in short melting time. For infiltration to shorten H-DRI melting time the porosity must be high (65%) and combined with a high heat transfer coefficient (> 2500 W m^-2K^-1) from slag to H-DRI. The combined results of this work highlight the importance of accurate data on the material properties of slag for numerical simulations and reliable prediction of process performance. It is also shown that the IronArc process is a viable alternative as a low carbon alternative for primary steelmaking at smaller scale plants, while the melting of H-DRI can be used for medium to large scale primary iron production. The models developed in this work can be used to further evaluate how different types of H-DRI with varying properties will behave in melting processes

urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-369667