Flow field characterisation of AusIron top submerged injection system

In this study the flow field characteristics within an elliptical liquid bath, agitated by a top submerged gas injection lance system, namely AusIron, were investigated experimentally and numerically. Two types of possible furnace configuration were examined: elliptical and elliptical with an inclined off take. The experimental study used a laboratory model where isothermal air was injected into the liquid bath. The experimental scale model consisted of an elliptical Perspex vessel and two vertically supported lances. These lances were connected to an air supply and helical inserts in each one to impart a swirling motion to the gas flow in the annulus. Water was used as the working fluid. The laboratory model was constructed according to the geometric and dynamic similarities, where the experimental operating conditions were determined based on the Froude number. Although the model was simple, it captured the main physical features of the overall flow characteristics and mixing within the furnace. The experimental work commenced by first carrying out a qualitative flow visualization study followed by a quantitative experimental investigation, whereby a measurement technique namely Laser Doppler Anemometry was utilized to measure the instantaneous velocity components within the bath. The data obtained was used to determine the overall flow characteristics and mixing under various operating conditions. These include parameters such as top gas injection rate, the submergence levels and the level of the liquid bath as well as lance arrangements. In the numerical part of the study, the fluid flow structure within the liquid bath agitated by a top submerged gas injection system for the elliptical geometry was carried out using the commercial Computational Fluid Dynamics (CFD) software package CFX 5.1. A two-dimension gas-liquid transient numerical model was developed by using the Eulerian-Eulerian approach. The standard k-ε model was employed to predict the turbulence structure of the liquid phase. A new empirical correlation was developed based on Froude number similarity to calculate liquid axial velocity and jet penetration depth. The agreement between experimental data and the published results were found to be satisfactory. The results indicated that several different flow structures can occur in the bath depending upon the buoyancy and momentum of the jet. The jet penetration height fluctuates around a mean value significantly lower than the maximum height of penetration. The downward annular flow turns radially outward as it reaches the bottom of the vessel, then flows along the bottom of the vessel forming an annular recirculation zone with a vertical motion opposite to the vorticity of the jet. This recirculation zone moves along the bottom of the vessel to the vessel wall, where it remains. Moreover, with respect of the gas flow rate, a higher gas-flow rate caused the liquid to be ejected deeper into the bath therefore a better mixing effect at the bottom of the bath can be achieved. However, the lower gas-flow rate resulted in a moderate and adequate agitation of the liquid bath. On the effect of the submergence the lower the level of the submergence the more rapidly the gas jet spreads at the top section of the bath stratifying above the heavier bulk liquid then simply penetrating to the free surface. At deeper submergence levels the high initial momentum of the gas jet transfers to the liquid bath which improves agitation and mixing within the bath. In addition, the distance between the lances play a major role in relation to the liquid behaviour at the center of the bath. For the case of the lance located at 115 mm from center, the two ejected plumes spread out and left a stagnant area due to the high pressure between them. When the lances moved closer to each other, 90 mm from the center, the two ejected plumes almost coupled into one. This was beneficial to reduce the stagnation around central region and therefore desirable mixing at the central region can be achieved. The comparison of the numerical predictions with experimental data under various operating conditions showed a reasonable agreement on both mean and turbulent fluid flow characteristics. These two studies were complementary to each other and provided the necessary information required for the successful optimisation of the AusIron process.