A Study of Bubble and Inclusion Behaviors in a Liquid Steel Bath

Abstract:     The bubble formation processes in a water model experiment were measured by a high-speed camera. Nozzle diameters of 0.5 mm, 1 mm and 2 mm were investigated under both wetting and non-wetting conditions. The bubble sizes and formation frequencies as well as the bubbling regimes were identified for each nozzle size and for different wettabilities. The results show that the upper limits of the bubbling regime were 7.35 L/h, 12.05 L/h and 15.22 L/h under wetting conditions for the 0.5 mm, 1 mm and 2 mm nozzle diameters, respectively. Meanwhile, the limits were 12.66 L/h, 13.64 L/h and 15.33 L/h for the non-wetting conditions. In addition to experiments, numerical modeling was performed. The Volume-of-Fluid (VOF) method was used to track the interface between the gas and liquid. The simulation results were compared to experimental observations from an air-water system. The comparisons show a satisfactory good agreement between the two methods (maximum difference is 0.029 s during a bubble formation period). Simulations from the argon-steel system show that the effect of the nozzle size on the bubble formation is insignificant for the current studied metallurgical conditions. The upper limits of the bubbling regime were approximate 60 L/h and 80 L/h for a 2 mm nozzle for wetting and non-wetting conditions, respectively. In addition, a poor wettability leads to a bigger bubble size and a lower frequency compared to a good wettability, for the same gas flow rate.    The fundamental aspects of rising argon bubbles in molten metal flow were investigated by numerical simulations. The results show that 3~10 mm bubbles rise in a spiral way with strong instabilities which cause them to change their instantaneous shapes. In addition, 10~20 mm bubbles rise rectilinearly and their shapes are kept almost steady. All these bubbles’ terminal velocities are around 0.3 m/s, which are in accordance with literature data. The simulation results of bubble bursting at the liquid surface show that when the surface tension is 1.4 N/m, the critical bubble size is 9.3 mm. Also, the ejection is found to increase with an increased surface tension value, unless a critical bubble size is reached. The single bubble passage through the liquid-liquid interface was numerically simulated. The calculation results show that the passing patterns at the steel-slag interface are oscillation-pass, oscillation-breakup, oscillations-pass, pass and pass-breakup for the 3, 5, 7, 10 and 15 (20) mm bubbles, respectively.  For a 5 mm bubble, it was found that an increase of the interfacial tension from 0.04 to 0.8 N/m results in a delayed bubble passage time. The results also show that the bubble experiences an oscillations-breakup process if the interfacial tension value is up to 1.15 N/m. However, a higher interfacial value (1.8 N/m) can make the bubble pass again but with a longer passage time.    The inclusion removal mechanism due to a bubble wake flow was studied using a water model and using a three dimensional numerical model. The individual particle motion was tracked by the discrete phase model (DPM). The average bubble velocity and particle velocity from simulation are of 2.5% and 28.9% differences compared to the water model experimental results. The predictions in an argon-steel-inclusion system show that the removal rate per bubble is increased with an increased bubble size. However, the inclusion removal rate per unit bubble volume can be improved by decreasing the bubble size. Also, the particle rising zone was found to be 1.625 and 5 times of the bubble size in width and height, respectively. It is also shown that the bigger inclusions are more easily removed compared to the smaller ones.    The Euler-Euler two-phase simulation model was used to investigate gas stirring in a ladle. In addition, water model experiments were carried out to validate the predicted flow field by using the UVP velocity measurement method. The simulation results show a maximum 0.04 m/s difference in axial velocity compared to the experimental observations. The flow patterns under different gas flow rates were obtained. Simulations for 120 s without the stochastic turbulent motions show that more than 30% of the 400 μm light inclusions can be removed. Also, that the 1~100 μm light inclusions have their highest removal percent for a 3 m3·h-1 gas flow rate. The size (1~100 μm) of heavy inclusion shows little effect on the inclusion removal percent. The 400 μm heavy inclusions are harder to remove compared to the smaller inclusions when the gas flow rate is or higher than 2 m3·h-1. Also, the inclusion transport behavior with the stochastic turbulent motions shows that both light and heavy inclusions are removed by up to 94% with a 1 m3·h-1 gas flow rate after 40 s. When the gas flow rate is 3 m3·h-1, the inclusion removal percent can be improved from 77% to 90% after 20 s and it reaches to almost 99% after 40 s. A further increase of the gas flow rate reduces the inclusion removal percent slightly.