The behavior of a bubble in collision with a poor wettability plate was experimentally investigated. Equilibrium shape of a sessile bubble beneath a horizontal plate was determined from the condition minimizing the bubble energy being the sum of the potential energy and the surface energy. Rebound motion of an impacting bubble was observed for a poor wettability plate as well as good wettability plate. For the poor wettability plate, small bubbles (dB≈ 2×10-3m) attached to it up to an inclination angle θp of 60°C. Large bubbles being attached to an inclined plate of poor wettability ascended along the plate. The limit between stopping and sliding of an attached bubble was determined from a force balance including the advancing contact angle θa and receding contact angle θr. There is not difference of bubble rising velocity between two flat plates.
The fluid flow and gas-liquid mass transfer in a water model of gas-injected ladle were analyzed by a numerical simulation of turbulent flow and mathematical models on mass transfer. In the fluid flow simulation, a single-phase model was adopted, in which the turbulent diffusion of bubbles was assumed. The computed average dissipation rates agreed well with the energy-input rates due to the work done by bubbles. The volumetric coefficients in the bubble dispersion zone and free surface zone were estimated by the eddy-cell model and penetration model, and compared with the observed results. The eddy-cell model gave better prediction of volumetric coefficients in the bubble-dispersion zone than the penetration model. On the other hand, both models did not predict well the volumetric coefficients at the free surface; the eddy-cell model gave larger values and the penetration model gave smaller values than the observed values. This was thought to be attributable to the fluctuating motion of the free surface, which consumed turbulence energy and decreased the energy dissipation rate required for the surface renewal.
Model experiments were carried out to investigate the bubble and liquid flow characteristics in a bottom blowing bath covered with a thick slag layer typical of in-bath smelting reduction processes. An aqueous ZnCl2 solution and silicone oil were used as the models for molten metal and molten slag, respectively. The density ratio of the solution to the silicone oil was 1.7, being close to a steel/slag density ratio of 2.0 to 2.2 in practice. The diameter of a vessel containing the two liquids was changed over a wide range. The holdup of the solution carried up by bubbles into the upper silicone oil layer was measured with a suction tube. The volume of the solution, Vm, was dependent mainly on the density difference. Empirical correlations of Vm and the penetration height of the solution were derived.
RH degassing vessels having one up-leg and one down-leg have been used for secondary refining of molten steel. One of the causes that determine their lives is the melting of the firebrick on the side surface of the lower vessel, which is induced by the slag infiltration into the brick. To increase the life of the vessel, the authors have proposed an RH degassing vessel with arranged four legs on the bottom of the lower vessel. A down-leg is positioned at the center and three up-legs are arranged around the down-leg. Both the circulation characteristics and the floating behavior of the slag have been examined using a water model with four legs (proposed model). The slag behavior was simulated using colored plastic particles. To induce water circulation, the air was blown into the up-leg. The results were compared with those of a water model with two legs (conventional model). (1) In the conventional model, the side surface of the lower vessel restricts the water inlet into the down-leg. In the proposed model, the water flows uniformly into the down-leg because of the down-leg position. (2) The circulation flux (i. e. the water circulation rate per unit area of the up-leg) in the proposed model was not less than that in the conventional one. In both models, the circulation flux is increased as an increase in the blowing air rate. (3) In the proposed model, the slag gathers the central part of the surface of the water in the lower vessel.
Molten steel flow and inclusion removal behavior in tundish were studied by using numerical simulation and water-model experiment. The effects of tundish characteristics such as width, depth, length on inclusion outflow ratio were evaluated and a new inclusion outflow model was deduced using tundish Width/Depth ratio. The results are as follows: (1) By a bottom short-circuit flow in a tundish, inclusion outflow ratio is extremely increased in the case of Width/Depth ratio<0.6. (2) With the enlargement of tundish width in same volume, inclusion outflow ratio is decreased. (3) When Width/Depth ratio exceeds 1.5, cross-sectional circular flow is enlarged and inclusion outflow ratio attains to a constant value.
A new process was devised for promoting inclusion separation from the steel in a tundish. The process utilizes electromagnetic force to rotate molten steel in a cylindrical chamber. This process is refered to as the Centrifugal Flow Tundish (CF Tundish). In this work, oxygen content and the number of inclusions in the tundish are examined in an industrial plant to estimate the performance of deoxidation and inclusion separation. A new mathematical model was developed and the deoxidation rate constant in the rotation chamber of the CF tundish is estimated at as much as 0.7 min-1. This large deoxidation rate is considered to be due to the large turbulence energy of the CF tundish. Large inclusions over 10μm in diameter reduce with the CF tundish during the ladle exchange as well as during the steady state casting. 2-phase fluid dynamic simulation using the VOF method revealed that this large inclusion separation in the CF tundish can be explained by the restraint of the short circuit flow. Thus, this process has been commercially used to produce clean Al-killed stainless steel with high quality.
In order to evaluate the flow velocity of molten steel driven by electromagnetic stirrer in the mold, not only deflection mechanism of dendrite due to fluid flow but also fluid dynamics promoted by Lorentz force have been taken into account. Deflection angle of dendrite in cross section have been analysed in various steels which have been cast with various width and promotive forces. The relation between deflection angle and promotive force is not well correlated (correlation factor: 0.63). It is found from the numerical analysis of fluid dynamics that the flow velocity increases with increasing width of stab, at a constant promotive force. Thus, promotive forces are normarized with respect to width. Since the deflection angle of steel dendrite is affected by carbon content, experimental data on deflection angle are also modified by carbon content. The relation between normarized promotive force and modified deflection angles is then quite well correlated (correlation factor: 0.97). Since the Lorentz force is a body one, the flow velocity (Us) can be simply derived. The ratio between flow velocity calculated from modified deflection angle and Us is approximately 0.4, which may be constant with usual casting condition of slab. This indicates that 60% of imposed electromagnetic force disappears due to the interaction of 3D flow, due to the interference between submerged entry nozzle etc. To estimate the flow velocity, it is important to take the effect of carbon content on deflection angle into consideration.
In order to understand the formation mechanism of equiaxed grains due to forced convection, in-situ observation using succinonitrile, which is a transparent organic material, has been carried out. A new experimental technique, using a fiber scope, has been developed in this study to observe in a 3-D mold. Molten succinonitrile was poured into the mold and was unidirectionally solidified from a chill plate. During this experiment, molten succinonitrile was stirred by a rotor at various rotation rate. Phenomena in the mold were monitored by a fiber scope and recorded by video tape. Thermal field was also measured by some thermocouples. Equiaxed grains formed in the presence of forced convection caused by the rotor, on the other hand, no equiaxed grain formed without forced convection, which agree with the industrial data. The number of equiaxed grain (Ne) as well as the diameter of equiaxed grain (De) increased with time (t). Ne and De are approximately proportional to t2.5 and t0.3 respectively. The rate of increase of Ne increases and the diameter of equiaxed grain decrease with increasng rotation rate. Furthermore, the temperature gradinet in the liquid was found to decrease with increasing rotation rate. Therefore, it can conclude that finer and more equiaxed grains form and they may continue to grow in the bulk liquid with higher rotation rate.
The characteristics of molten steel jet flowing out from the immersion nozzle in the continuous casting mold control the flow pattern in the mold, thereby strongly influencing the quality and productivity of the cast steel slabs. We proposed a new method to establish a reasonable flow pattern in the mold by imparting a swirling motion to the flow in the immersion nozzle without the bottom. The following results were obtained from a water model study. (1) A quite stable swirling flow being established in the immersion nozzle without the bottom when the swirling velocity exceeded a critical value of 0.8m/s, under this condition there existed no separation on the inner wall of the immersion nozzle. (2) When the swirling velocity was higher than 0.8m/s, the fluid on the symmetry plane of the immersion nozzle moved along the curved inner wall of the nozzle. Accordingly, the outlet-flow was directed outwards as well as downwards, while a weak upward flow, i.e., inflow was observed around the vertical nozzle axis near the outlet of the nozzle. As a whole, the fluid flow near the guide plane of the nozzle was directed downwards. (3) The fluctuation of the surface flow, i.e., the flow on the meniscus of the mold was strongly suppressed due to the appearance of the S-shaped flow pattern in the transverse sections. As a result, swirl motions around the immersion nozzle were also suppressed. In addition, both the fluctuations of the mean velocity components and the turbulence components of the flow in the mold became very small, being preferable conditions for continuous casting.
A sensor for flow velocity measurement of molten steel was developed which measured the shedding frequency of Karman vortex streets behind a refractory cylinder, and then the absolute value of flow velocity below the meniscus in a high speed continuous casting mold was made clear. Experiment for calibration of this sensor was done using both molten steel and molten metal which having low melting point. It was found that the flow velocity was described as a function of both the shedding frequency of Karman vortex streets and the diameter of the refractory cylinder. On the basis of this relation, the measurement of the shedding frequency of Karman vortex streets could give the absolute value of flow velocity of molten steel below the meniscus in the mold. It was possible to determine optimum conditions of high speed continuous casting from the result of measured flow velocity in the mold using this sensor.
A flow control of molten steel in a continuous slab caster mold has been carried out with travelling magnetic field. The aim of flow control is prevention against mold powder entrapment into molten steel. The travelling magnetic field was imposed to spouting stream from a submerged entry nozzle with the linear-motor-type actuator. This actuator has two travelling modes of magnetic field. One is for braking the velocity of spouting stream (EMLS: Electromagnetic Level Stabilizer), and another is for acceleration (EMLA: Electromagnetic Level Accelerator). With these braking and accelerating performance, control of steel flow velocity just below meniscus was carried out. By velocity measurement below molten steel surface at a commercial caster mold, performance of EMLS and EMLA was investigated. Especially for EMLS performance, change in meniscus velocity profile along mold width direction was examined. When increasing the intensity of EMLS, at first the meniscus velocity around a submerged entry nozzle was reduced and then the meniscus velocity around the mid of half width of a mold, hereafter the velocity around narrow face of a mold was reduced. Moreover formularization of the braking performance of meniscus velocity by EMLS was carried out. In this formula the braking performance was divided into two factors. By using the EMLS and EMLA with a computer control system, the surface defect frequency on cold rolled coil was reduced drastically.
Sequential casting of different steel grades with a level DC magnetic field (LMF) has been performed by 8-ton scale pilot casting machine. According to the results of casting experiment it has been clarified that the mixing length at the ladle change region decreased with LMF and was affected by density difference of molten steel between the first and the second ladle. The fluid flow phenomena in strand pool with LMF and the mechanism of the difference of mixing behavior by density difference have been elucidated by numerical analysis. Moreover the criteria of stability in the strand pool has been proposed in consideration with temperature and solute difference of molten steel between the former and the latter ladle.