The main function of the converter furnace lining was to provide a durable container for the high-temperature molten bath. During the steelmaking process, the occurrence of melting corrosion led to the destruction of the furnace lining structure and a subsequent change in the shape of the furnace. Hence, the dynamic condition of the molten bath was altered. In this paper, both water experiment and numerical simulation have been adopted to analyze the flow field characteristic of molten bath by various oxygen lance parameters, in both the initial and late stages of the furnace lining structure of a top-blowing converter. The results revealed the late furnace lining structure improved the average velocity of the molten bath, thereby reducing the mixing time and volume of the low-velocity dead zone, comparing with the furnace lining structure. In the late furnace lining structure, the larger furnace diameter expanded the impaction area of the molten bath, resulting in an enhanced contact area between the liquid slag and the molten steel. Consequently, the FeO in the liquid slag rapidly reacted with the C element in the molten steel, leading to a decrease in the fluidity of the liquid slag and a decline in dephosphorization efficiency. Based on the results generated by water experiment and numerical simulation, two types of new oxygen lances were investigated in the industrial application research. Subsequently, it was determined that the new oxygen lance with an inclination angle of 12.3° was deemed suitable for the 35t top-blowing converter.
The basic refractory has been extensively utilized in the construction of furnace linings, including the permanent lining and working lining, for the basic oxygen furnace (BOF) and Electrical Arc Furnace (EAF).1,2) In order to effectively mitigate the rate of melting corrosion, primary constituents of basic refractory materials consist of MgO and CaO. During the steelmaking process, the high-temperature molten bath keeps impacting the furnace lining. This, combined with the phenomenon of mass transfer from the furnace lining to the liquid slag due to variations in saturation concentration, ultimately leads to structural damage of the furnace lining.3,4) Although the slag splashing technology is adopted to prolong the lifespan of the furnace lining, the volume of the furnace gradually increases, resulting in changes to the dynamic conditions of the molten bath.5,6)
The main sources of kinetic energy for the molten bath are the top-blowing oxygen lance and bottom-blowing nozzle, which also play a crucial role in determining the mixing effect of the molten bath.7,8,9) However, during the later stage of the converter’s life cycle, non-standard operation may lead to the partial plugging of the bottom-blowing nozzle, resulting in a decreased dynamic condition of the molten bath.10,11) Therefore, optimizing the parameters of the oxygen lance is an effective approach to enhance the dynamic condition of the molten bath during the later stage of the converter’s life cycle.
There are a series of numerical simulations and experiment tests have been carried out to investigate the effects of Laval nozzle structure and arrangement on the behavior of the molten bath. As a result, several effective methods have been proposed to enhance the impaction ability of supersonic oxygen multi-jets and improve the mixing effect of the molten bath.12,13,14) Odenthal and Liu et al.15,16) investigated the effect of Laval nozzle shapes on the behavior of the supersonic jet, and pointed out the Laval nozzles designed by the characteristic-line method were able to suppress the velocity diffusion of oxygen jet in the radial direction, which increased the utilization rate of oxygen gas. Feng et al.17) analyzed the flow field characteristic generated by a six-hole oxygen lance using a two-parameter nozzle arrangement, and found the centreline velocity formed by the two-parameter oxygen lance was greater than that by the traditional oxygen lance at room and high ambient temperature. Sambasivam et al.18) designed an Laval nozzle layout by inserting a subsonic nozzle at the center of the traditional oxygen lance tip, and proposed the central nozzle could increase both radial and axial multi-jets velocity, which effectively increases the impaction area of the molten bath. Liu et al.19) indicated the main jet and secondary jet converged in advance with the increase of the secondary nozzle number, which increased the kinetic energy of oxygen multi-jets due to a reduction in the entrainment phenomenon between the main oxygen jet and ambient gas. Lv et al.20) researched the influence of the shrouding gas flow rate on the velocity flow field of the supersonic jet, and adopted a coherent oxygen lance without combustion gas in the BOF steelmaking process, which proved the shrouding gas technology could improve the metallurgical indexes. It is important to note that the majority of these studies are carried out in the initial furnace lining structure condition, with limited research conducted on the furnace lining structure condition.
The Panshi Jianlong Iron and Steel Company is restricted by the structural limitations of the 35-ton top-blowing converter. As a result, the bottom-blowing technology is not adopted to enhance the mixing efficiency of the molten bath in this converter. Additionally, the furnace life of this converter is approximately 20 thousand heats. When the furnace reaches 15 thousand heats, the furnace lining undergoes significant changes. Consequently, there is an increase in the occurrence of molten steel spilling. This leads to a worse viscosity of liquid slag, a lower dephosphorization rate and higher M. Fe content in the end-point liquid slag. In order to achieve the better metallurgical and economic effects, Panshi Jianlong iron and steel company decided to use a suitable oxygen lance at the late furnace lining structure. Therefore, it is necessary to conduct both a water experiment and numerical simulation to propose appropriate oxygen lance parameters in the industrial application.
In this research, there are three kinds of oxygen lance have been tested at both initial and late furnace lining structure conditions. The mixing time, impaction area and depth have been measured in the water experiment. Additionally, the average velocity, volume of low-velocity zone and contact area between liquid slag and molten bath would be calculated by the simulation model. Consequently, the optimized oxygen lance will be tested in the industrial application research to evaluate the smelting time, dephosphorization rate and T. Fe content in slag.
The gas flow rate of oxygen lance has a great influence on the flow field characteristic of molten bath. In order to satisfy requirement of dynamic similarity decided by viscous force, gravity, and inertia force, the relationship of gas flow rate between water model and industrial prototype should be calculated by the modified Froude number (Fr). The following equations, which take the temperature variation into consideration, are used to achieve the gas flow rate in water experiment, on the basis of the similarity of dynamic and geometry.21)
| (1) |
| (2) |
| (3) |
where, subscript m and p present the water model and industrial prototype, respectively. ρg and ρl are the gas and liquid phase density (kg/m3), respectively. u and L are the gas velocity (m/s) and feature size (m), respectively. Q, T and H are gas flow rate (Nm3/h), temperature (K) and depth of molten bath (m), respectively. g and C are the gravity force (m/s2) and geometry ratio. The ρ° is the density (kg/m3) at the standard state.
2.2. Experimental ApparatusFigure 1 showed the experimental apparatus of the water model. The top-blowing converter was built in plexiglass with the geometric ratio of 1:3, and the prototype and model parameters were shown in the Table 1. In the experiment, compressed air and water were adopted for representing gas flow and molten steel, respectively. Figure 2 represented the prototype parameter variation for the furnace at the initial and late period of the converter life-cycle.
| Parameters | Prototype | Model | ||
|---|---|---|---|---|
| Initial period | Late period | Initial period | Late period | |
| Molten bath depth (mm) | 1165 | 658 | 388 | 219 |
| Furnace diameter (mm) | 2820 | 3720/3680 | 940 | 1240/1227 |
| Bottom diameter (mm) | 2435 | 2655 | 812 | 885 |
| Total height (mm) | 5403 | 1801 | ||
| Density of molten steel (kg/Nm3) | 7200 | 1000 | ||
| Oxygen top-blowing rate (Nm3/h) | 9200/9600/10000/10400/10800 | 128.7/134.3/139.9/145.5/151.1 | ||
| Lance height (mm) | 1000/1100/1200/1300 | 260/280/300/320 | ||
Based on the operational condition of the 35t top-blowing converter, the design oxygen flow rate for new oxygen lance has been increased from 9100 to 10000 Nm3/h. At the same time, the inclination angle of Laval nozzle has also been improved, due to a bigger furnace diameter. Table 2 shows the structure parameters of top-blowing lance.
| Label | Mach number | Flow rate (Nm3/h) | Inclination angle (°) | Diameter | Prototype | Model |
|---|---|---|---|---|---|---|
| Original oxygen lance | 2.00 | 9100 | 11.5 | Throat diameter (mm) | 28.5 | 9.5 |
| Exit diameter (mm) | 37.0 | 12.3 | ||||
| New oxygen lance | 2.00 | 10000 | 12.0/12.3 | Throat diameter (mm) | 29.5 | 9.8 |
| Exit diameter (mm) | 38.3 | 12.8 |
In the initial condition, the water has been added into the water model, and then the compressed air would be blown into oxygen lance. After the compressed air flowed for 3 minutes, the tracer element concentration and mixing time would be measured. For calculate the mixing time, there were three conductivity electrode have been arranged at different region to trace the tracer concentration (KCl concentration) variation. Hence, the mixing time was defined as that when the tracer concentration at the monitor points reached 99% of the average tracer concentration in the water layer. The average mixing time (Tm) was achieved by Eq. (4), and T1, T2, and T3 were the mixing times measured by three conductivity electrodes, respectively.
| (4) |
To determine the radial and depth of impaction, two graduated scales were positioned on the water model, and a high-speed camera was positioned at a fixed location beside the water model to record impaction data, as represented in Fig. 3. One minute after initiating the water experiment, the camera recorded the shape of the impaction cavity within a 30-second timeframe. Following this, these measurements of the impaction cavity were used to calculate both the impaction area and impaction depth.
The following assumptions have been adopted to in the simulation model:
(1) Both the molten steel and liquid phases were considered to be incompressible Newtonian fluids.
(2) The foam process of the liquid slag, generated by the carbon-oxygen reaction between the slag and metal,22,23,24) was not taken into consideration, which meant the density and volume of liquid slag were constants.25,26,27)
(3) The oxygen was considered as a compressible ideal gas.
(4) No chemical reactions in the molten bath were taken into consideration, and the simulation model was assumed to be an isothermal process, which is same as the water experiment.
3.2. Governing EquationsDue to the obviously density variation, the interface between the gas, liquid slag, and molten bath was clear visible under the force of gravity. Therefore, the Eulerian-multifluid Volume of Fluid (VOF) model was used to track the free surface zone (gas/liquid slag/molten steel interface) in the computational domain. The governing equation for the volume fraction of tracked phases can be written as Eqs. (5) and (6) in the sharp interface method.28)
| (5) |
| (6) |
where, αi and ρi are the volume fraction (%) and density (kg/m3) for ith phase, respectively;
The Reynolds-Averaged Navier-Stokes (RANS) formation was used to solve the continuity, momentum and energy equations, and the SST k-ω turbulence model was adopted to calculate the viscous effects and vortex structure for coherent jet. Equations (7) and (8) shown the turbulence kinetic energy (k) and its dissipation rate (ε) model, respectively, referring to transport equation.29)
| (7) |
| (8) |
where, Gk and Gω are the generation of turbulence kinetic energy (m2/s2) and specific dissipation rate, respectively. Yk and Yω are the dissipation of k and ω, respectively. Γk and Γω are the effective diffusivity (m2/s) of k and ω, respectively.
3.3. Computational DetailTo investigate flow field of molten bath generated by various operational conditions, two full-scale top-blowing converter models have been established, according to the furnace parameters at the initial and late furnace lining structure. The computational domain included the oxygen lance, molten bath, and the region extending from oxygen nozzle tip to molten bath surface. Figure 4 depicted the mesh profile of furnace with boundary conditions. The mesh density has been improved near the oxygen lance exit, due to the greater velocity and pressure gradient.
The mass flow inlet condition (blue plane) was used at the inlet face for the Laval nozzle. At the oxygen lance wall, both non-slip condition and enhanced wall treatment methods were adopted for the wall boundary condition (gray plane). The pressure outlet condition (red plane) was adopted for the outlet face of the simulation model. Initially, the computational domain started with no gas passing through the Laval nozzle, and the furnace gas and molten bath remained stationary.
The flow field of molten was achieved by the transient solution in the simulation process, and the PISO (pressure implicit with splitting of operators) scheme was selected to solve the pressure-velocity coupling process. The Body Force Weighted and Geo-Reconstruct models were used to calculate the pressure and volume fraction variations for the multiphase flow under gravity condition, respectively. The other equations were solved by the second-order upwind scheme. At the initial condition, time step size was set as 10−5s by the fixed type, and then the adaptive method was used to calculate the time step with the global Courant number of 1.0. Moreover, in all the numerical simulation cases, the residuals of dependent variables would be defined as convergent with the energy residual and other variables being less than 10−6 and 10−3, respectively.
The mixing effect of a molten bath can be effectively demonstrated by the mixing time in a water experiment, and a shorter mixing time leads to an enhancement in the mixing effect of the molten bath. Figure 5 depicts the mixing time profiles in water experiment generated by various operational conditions. Hereafter, in the figure legend, the terms “origin”, “12.0” and “12.3” are represented the original oxygen lance, new oxygen lance designed by inclination angle of 12.0 and 12.3°, respectively. Additionally, the terms “ini” and “late” are addressed by the initial and late furnace lining structures, respectively. For instance, the Origin-late presents that the original oxygen lance is tested at the initial furnace lining structure. In this research, the lance height is defined as the distance from the tip of Laval nozzle exit to the surface of molten bath. This means that when the lance height is 1000 mm, the distance from the tip of Laval nozzle exit to the surface of molten bath is also 1000 mm, for both initial and late furnace lining structures.
On the basis of the results, both oxygen lance parameter and furnace lining structure have a great influence on the mixing time. The mixing time reduces with a higher top-blowing flow rate, as shown in Fig. 5. The average mixing time generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 81.8, 73.5, 63.5, and 67.8 s, respectively. Hence, the mixing time achieved by the initial furnace lining structure is larger than that by the late lining structure, which means the late furnace lining structure would reduce the mixing time. Meanwhile, the mixing time generated by the original oxygen lance is longer than that by new oxygen lance, and the new oxygen lance with an inclination angle of 12.0° further decreases the mixing time. This implies that both types of new oxygen lances exhibit enhanced stirring capability compared to the original oxygen lance, while the late furnace lining structure also enhances the dynamic condition of the molten bath.
In the industrial production process, the oxygen flow rate is ranged from 9200 to 10800 Nm3/h. Hence, for the original oxygen lance designed by the 9100 Nm3/h, its variation between actual and design flow rate is ranged from +0.1% to +18.7%. For the new oxygen lance designed by 10000 Nm3/h, its variation between actual and design flow rate is ranged from −8.0% to +8.0%. Therefore, the maximum variation between actual and design flow rate of original and new oxygen lance are 18.7% and 8.0%, respectively.
When there is a disparity between the actual and design flow rates of the Laval nozzle, a noticeable shock wave forms at the exit of the Laval nozzle, which ultimately reduces the stirring ability of the supersonic jet.30) To investigate the effect of flow rate variation on the mixing time, the slopes of the linear regression for the mixing time at various conditions have been calculated, as presented in the Table 3.
| Flow rate range | Origin-ini | Origin-late | 12.0-late | 12.3-late |
|---|---|---|---|---|
| 9200–10000 Nm3/h | −0.051 | −0.065 | −0.033 | −0.033 |
| 10000–10800 Nm3/h | −0.017 | −0.013 | −0.035 | −0.038 |
The result indicates for the original oxygen lance, the slopes of the linear regression for mixing time achieved at flow rate ranging from 9200 to 10000 Nm3/h are much smaller than those at flow rate ranging from 10000 to 10800 Nm3/h. Thus, the decrease rate of mixing time reduces with a higher oxygen flow rate, for the original oxygen lance. This indicates that the stirring ability of the original oxygen lance would be significantly suppressed, when its actual flow rate exceeds the design flow rate.
For the new oxygen lances, their slopes of the linear regression of mixing time show only a slight change. Hence, in this research, the variation between actual and design flow rate has limited impact on both the stirring ability of the new oxygen lance and the mixing time of molten bath.
The kinetic energy of the oxygen multi-jets decreases as they move further away from the Laval nozzle exit. This is attributed to the entrainment phenomenon that occurs between the oxygen multi-jets and the surrounding gas flow. Consequently, the total kinetic energy of the oxygen multi-jets at the surface of the molten bath is reduced with a higher lance height.
However, kinetic energy source for molten bath is solely provided by the oxygen multi-jets in this converter. Meanwhile, the fluid collision formed by various vortexes of molten steel stream also has a great influence on the mixing effect of molten bath. Thus, when the lance height is reduced, although total kinetic energy of the oxygen multi-jets is enhanced, the mixing effect of molten bath may also be unsatisfactory under certain specific conditions, as demonstrated by previous literatures.31,32,33)
In this research, the mixing time reduces with a higher lance height. The maximum and minimum mixing time are achieved at lance heights of 1000 and 1200 mm, respectively. Furthermore, the average slopes of the linear regression for mixing time are −0.025 for lance heights ranging from 1000–1100 mm, and −0.01 for lance heights ranging from 1100–1200 mm. The result indicates that the reduction rate of mixing time would be evidently suppressed by a higher lance height, at the tested condition.
Figure 6 shows the impacting diameters profiles in water experiment at a series of flow rates and lance heights. A bigger impaction diameter leads to a bigger impaction area, which improves the contacting area of oxygen multi-jets and molten bath.
The average impaction diameters generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 0.159, 0.194, 0.215, and 0.221 m2, respectively. Figure 2 shows the diameter of furnace at the initial furnace lining structure is obviously smaller than that at the late furnace lining structure. This indicates that, for the liquid slag, the flow obstruction effect of furnace in the radial direction at the initial furnace lining structure is smaller than that at the late furnace lining structure. Therefore, all the impaction area formed at the furnace lining structure is smaller than that at the late furnace lining structure.
The impaction area of molten bath increases with a greater inclination angle, under the same conditions of flow rate, lance height and Mach number. Meanwhile, the inclination angle of original oxygen lance is smaller than that of two new oxygen lances. Hence, the ascending of impaction area of molten bath formed by different oxygen lance is: original oxygen lance, new oxygen lance with inclination angle of 12.0° and new oxygen lance with inclination angle of 12.3°.
Increasing the flow rate improves the total kinetic energy and stirring ability of oxygen multi-jets, resulting in a larger impaction diameter. As mentioned, the reduction rate of the mixing time formed by the original oxygen lance gradually reduces with an increase in flow rate, due to the variation between design and actual flow rate. In order to analyze the relationship between the oxygen flow and impaction area generated by various oxygen lances, the slopes of the linear regression for impaction area at different flow rates have been calculated, as shown in the Table 4.
| Flow rate range | Origin-ini | Origin-late | 12.0-late | 12.3-late |
|---|---|---|---|---|
| 9200–10000 Nm3/h | 0.109 | 0.081 | 0.047 | 0.051 |
| 10000–10800 Nm3/h | 0.027 | 0.015 | 0.036 | 0.035 |
Based on the results, for the both original and new oxygen lances, the slopes of the linear regression for impaction area are larger in the flow rate range of 9200 to 10000 Nm3/h compared to the range of 10000 to 10800 Nm3/h. This indicates that the increasing rate of impaction area suppresses with a higher flow rate. When the oxygen flow rate is ranged from 9200 to 10000 Nm3/h, the increasing rate of impaction area generated by the original oxygen lance is bigger than that by the new oxygen lance. However, the increasing rate of impaction area generated by the original oxygen lance is smaller than that by the new oxygen lance, when the oxygen flow rate is ranged from 10000 to 10800 Nm3/h. This further proves that the impaction area formed by oxygen multi-jets reduces with a larger variation between actual and design oxygen flow rate.
With a higher lance height, the top-blow multi-jets could contact with a large surface area of molten bath at the same flow rate and inclination angle, which increases the impaction diameter, as previous literatures reported.34,35) Based on the data presented in Fig. 6, the average slopes of the linear regression for impaction area are 0.054, 0.040, 0.022 and 0.021 for lance heights ranging from 1000–1050 mm, 1050–1100 mm, 1100–1150 mm, and 1150–1200 mm, respectively. In this research, with a higher lance height, the increasing rate of impaction area gradually decreases.
The impaction depth of molten bath is an indicator of the penetration ability of oxygen multi-jets. A larger impaction depth signifies a greater penetration ability. With a larger impaction depth, there are more oxygen multi-jets would be contact with molten steel, which improves the direct reaction rates among the oxygen gas and elements in the molten bath. Figure 7 is the contrast of impacting depth at different operational conditions.
The average impaction depths generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 0.129, 0.121, 0.113, and 0.107 m, respectively. Based on the furnace parameter, the molten bath depth at the initial furnace lining structure is significantly larger than that at the late furnace lining structure. With a lower molten bath depth, the flow obstruction effect at the furnace bottom is increased, resulting in a smaller impaction depth. As a result, all the impaction depths formed at the initial furnace lining structure are bigger than those at the late furnace lining structure.
Comparing with a bigger inclination angle, it takes less time to reach the surface of molten bath for oxygen multi-jets with a smaller inclination angle, which suppresses the consumption of the kinetic energy for oxygen multi-jets due to the entrainment effect between oxygen multi-jets and ambient gas. Thus, the descending of impaction depth of molten depth formed by different oxygen lance is as follows: original oxygen lance, new oxygen lance with inclination angle of 12.0° and new oxygen lance with inclination angle of 12.3°.
Table 5 presents the slopes of the linear regression for impaction depth at different flow rates. With a bigger flow rate, the initial kinetic energy of oxygen multi-jets is improved, resulting in a greater penetration ability. However, for the both original and new oxygen lances, the slopes of the linear regression for impaction depth generated within a flow rate range of 9200 to 10000 Nm3/h are bigger than those at flow rate range of 10000 to 10800 Nm3/h. This indicates that the increasing rate of impaction depth also decreases with a higher flow rate, which is same with the rule governing the change in impaction area.
| Flow rate range | Origin-ini | Origin-late | 12.0-late | 12.3-late |
|---|---|---|---|---|
| 9200–10000 Nm3/h | 0.101 | 0.090 | 0.074 | 0.084 |
| 10000–10800 Nm3/h | 0.022 | 0.016 | 0.060 | 0.063 |
As the supersonic jet flows through the ambient gas, the low-velocity ambient gas keeps absorbing the kinetic energy of supersonic jet, which suppresses its impaction depth. Increasing the lance height extends the distance between the nozzle tip and the surface of the molten bath. Consequently, more kinetic energy is removed from the oxygen multi-jets by the ambient gas.
As represented in Fig. 7, the impaction depth reduces with a higher lance height. Furthermore, the average slopes of the linear regression for impaction depth are −0.051, −0.056, −0.061 and −0.068 when lance heights are range from 1000–1050, 1050–1100, 1100–1150 and 1500–1200 mm, respectively. This means that the increasing rate of impaction depth would gradually decrease, with a higher lance height.
The melting corrosion phenomenon in the BOF steelmaking process leads to a noticeable increase in furnace diameter and a consequent reduction in molten depth after 15000 heats. Hence, the dynamic condition of molten bath is changed for the initial and late furnace lining structure. The mixing time measured by water experiment demonstrates the dynamic condition of the molten bath at late furnace lining structure is better than that at initial furnace lining structure. This means that a larger furnace diameter can improve the stirring effect of the molten bath. It indicates a higher radial velocity of molten bath may further enhance the stirring effect of molten bath, comparing with the axial velocity of molten bath.
Based on the result of mixing time, the new oxygen lance achieves a greater stirring effect and impaction area, comparing with the original oxygen lance. However, the impaction depth generated by original oxygen lance is greater than that by new oxygen lance.
Thus, for the original oxygen lance, the stirring effect of molten bath has been improved at the same oxygen flow rate and lance height. This improvement has led to an increased mass transfer rate between the liquid slag and the molten bath, resulting in a lower FeO content in liquid slag. However, the fluidity of liquid slag is also suppressed, which may impair the metallurgical effect in the steelmaking process.
4.2. Numerical Simulation AnalysisFor the all simulation models, their lance height and top-blowing flow rate are 1000 mm and 10800 Nm3/h, respectively. Besides, the slag thickness with the initial and late furnace lining structure are 180 and 103 mm, respectively, with a slag density of 3000 kg/m3. That means the operational condition are same in various simulation models, and the oxygen lance and furnace lining structures are the only variables. Figure 8 is the flow velocity profile on the longitudinal section using various operational conditions.
Based on the result, both oxygen lance and furnace lining structures have a great influence on the flow field of molten bath. The depth of molten bath formed by the initial furnace lining structure is deeper than that by the late furnace lining structure. Thus, for the initial furnace lining structure, the high-velocity molten steel at the surface of the bath needs to traverse a greater distance before reaching the bottom. During this process, the kinetic energy of high-velocity molten steel is absorbed by the low-velocity molten steel. As a result, there is an obvious velocity stratification between surface and bottom of molten bath with the initial furnace lining structure.
Meanwhile, Fig. 9 displays four distinct flow regions labeled as A, B, C, and D, respectively. To provide a more comprehensive understanding of the flow characteristics within these specific regions, Fig. 9 represents the vector distributions of multiphase on the longitudinal section at different regions.
The A region corresponds to the area near the point where the top-blowing oxygen jets descend onto the surface of the molten bath. It extends from the descent point to the side-wall of the furnace. At the boundary of impaction cavity, the flow directions between the eddy of oxygen jet backflow and the eddy of furnace gas are totally opposite, as shown in Fig. 9(a). This indicates that the backflow of the oxygen jet or the furnace gas does not maintain a continuous acceleration of the molten bath stream at the surface. Consequently, within the A region, an eddy is generated within the molten bath, resulting in a reduction in velocity of molten bath.
Both the B and A regions are symmetrically distributed along the axis of the oxygen lance centerline. Meanwhile, the extended line of Laval nozzle can directly impact the A region, which is in contrast to the B region. Figure 9(b) depicts the flow directions between the molten bath stream and the furnace gas are basically same, at the surface of molten bath. As a result, the furnace gas eddy keeps accelerating the velocity of molten bath within the B region, resulting in a larger high-velocity volume in the B region compared to the A region.
The C region is located beneath the center of the impaction cavity. Because of none bottom-blowing arrangement, the flow stream of molten bath directly ascends towards the impaction cavity, which is opposite to the flow direction of top-blowing oxygen jet, as shown in Fig. 9(c). The flow stream of the molten bath initially decelerates as it approaches the impaction cavity, and then be accelerated by the top-blowing oxygen jet, due to the opposite flow direction of steel stream. Consequently, a low-velocity zone has formed in the C region.
The D region distributes in the bottom center of furnace. Figure 9(d) represents the flow directions of molten bath streams have been changed due to the collision among the eddies at the bottom of molten bath, which suppresses kinetic energy of molten bath and forms a low-velocity region.
Based on the result, the average velocities of molten bath generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 0.087, 0.145, 0.175, and 0.168 m/s, respectively. Compared to the initial furnace lining structure, the late furnace lining structure exhibits a noticeable increase in furnace diameter and enhanced interface area between the molten bath and gas flow.
Furthermore, the late furnace lining structure allows for the absorption of a greater amount of kinetic energy from high-velocity gas flows, such as the oxygen jet and furnace gas, by the low-velocity molten bath. Consequently, as the same result proposed by water experiment, the late furnace lining structure increases the average velocity of molten bath, which improves the dynamic condition of molten bath.
In order to further investigate the effect of operational conditions on the flow field of molten bath, both axial and radial velocities of molten bath have been analyzed, as shown in Table 6. In this research, the flow region for liquid slag and molten steel were marked by the Fluent software in the simulation process, and the volume-average method was adopted to calculate the average velocity of the molten bath. As mentioned, the late furnace lining structure could suppress velocity stratification between the surface and bottom of the molten bath due to its shallower depth. Meanwhile, a larger furnace diameter enhances the rate at which the molten bath absorbs kinetic energy from oxygen multi-jets or furnace gas, due to a greater interface area between molten bath and gas flow. Hence, the late furnace lining structure distinctly improves the axial and radial velocities of molten bath, comparing with the initial furnace lining structure.
| Lable | Origin-ini | Origin-late | 12.0-late | 12.3-late |
|---|---|---|---|---|
| Average axial velocity m/s | 0.045 | 0.073 | 0.070 | 0.064 |
| Average radial velocity m/s | 0.058 | 0.091 | 0.113 | 0.121 |
The Table 6 indicates that when the original oxygen lance was used, the average axial and radial velocities of molten bath with the late furnace lining structure are 1.62 and 1.57 times higher than those with the initial furnace lining structure, respectively. This means that the late furnace lining structure can further improve radial velocity of molten bath compared to the axial velocity of molten bath. At the test condition, the average axial velocity of molten bath decreases with a larger inclination angle, which is opposite to the radial axial of molten bath with the late furnace lining structure.
In this paper, a velocity threshold of 0.018 m/s, which corresponds to approximately 10 percent of the average velocity of the molten bath under different operational conditions, has been selected to calculate the volume of the velocity dead-zone. Figure 10 is the distribution of the relationship between the average velocity of molten bath and volume of velocity dead-zone in the molten bath.
The result presents the oxygen lance using inclination of 12.3° achieves the smallest volume of velocity dead-zone of 0.019 m3, which means it could evidently improve the mixing degree of molten bath in steelmaking process. Meanwhile, the volume of velocity dead-zone reduces as the average velocity increases, and the volumes of velocity dead-zone formed by the late furnace lining structure are consistently smaller than that by the initial furnace lining structure. This further proves that the late furnace lining structure improves the dynamic condition of molten bath.
Comparing with the origin-late, the variations in the average velocity of the molten bath by 12.0-late and 12.3-late are 0.030 and 0.023 m/s, respectively. The volumes of velocity dead-zone formed by 12.0-late and 12.3-late are reduced by 0.004 and 0.003 m3, compared to the origin-late. Hence, as the average velocity increases, the reduction rate of average volume of velocity dead-zone is substantially decreased.
The average velocity at the interface between liquid slag and molten steel is an important indicator to evaluate their mass transform rate, and Fig. 11 depicts the velocity distribution at the interface between liquid slag and molten steel.
The average velocities at the interface, between liquid slag and molten steel, generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 0.192, 0.206, 0.221, and 0.215 m/s, respectively. This means that the average velocity at the interface is increased with a larger furnace diameter. Meanwhile, for the late furnace lining structure, ascending of average velocity at the interface formed by different oxygen lance is: original oxygen lance, new oxygen lance with inclination angle of 12.3° and new oxygen lance with inclination angle of 12.0°.
In this study, the quasi-stable state of the molten bath is defined as its average velocity fluctuating within a specific range. When the molten bath achieves quasi-stable state, the interface area between the liquid slag and molten steel is assessed within a 30 s timeframe. Subsequently, these interface area measurements are utilized to calculate the average interface area. Based on the result, the average interface area, between liquid slag and molten steel, generated by origin-ini, origin-late, 12.0-late, and 12.3-late are 6.32, 9.36, 9.11, and 9.02 m2, respectively. Hence, the average interface is also enhanced with a larger furnace diameter. Additional, the average interface area is further improved with a smaller inclination angle of the oxygen lance.
Based on the result, when the original oxygen lance is adopted, both average velocity at the interface and the interface area generated by the initial furnace lining structure are obviously smaller than those by the late furnace lining structure. Thus, the mass transform rate of FeO between the liquid slag and molten bath is improved with a larger furnace diameter. This means that the FeO in the liquid slag would be also quickly reacted with the [C], [Si] and [Mn] in the molten bath, which enhances viscosity of liquid slag.
Therefore, the adoption of the original oxygen lance in the late furnace lining structure leads to a higher occurrence of molten steel spilling phenomenon. This, in turn, results in a lower dephosphorization rate and higher M. Fe content in the end-point liquid slag, compared to the initial furnace lining structure, as mentioned in the introduction section.
For the new oxygen lances with an inclination angle of 12.0° and 12.3°, although the average velocity of molten bath at the late furnace lining structure would be further increased, their impaction areas are improved and interface areas are suppressed, comparing with the original oxygen lance. Therefore, the utilization of these new oxygen lances in the industrial application may lead to the maintenance of a suitable FeO content in the liquid slag.
4.3. Industrial Application ResearchThe results of the both the water experiment and the numerical simulation proves that the new oxygen lances, with inclination angle of 12.0° and 12.3°, achieve the greater mixing and impaction degrees, comparing with the original oxygen lance. In order to further investigate the metallurgical effects of two types new oxygens lance in the steelmaking process, the new oxygen lances with inclination angle of 12.0° and 12.3° were used in a 35t top-blowing converter, at the late furnace lining structure.
The mass balance of oxygen has been calculated before the industrial application research, to achieve the theoretical melting time. And the validity of the industrial application research has been confirmed. A total of 360 heats were collected in the industrial application research, with 120 heats using original lance, 120 heats using 12.0° new oxygen lance and 120 heats using 12.3° new oxygen lance, and the same operation mode has been adopted for these three types of oxygen lances.
This paper considers the liquid iron as the initial conditions prior to the blowing and steelmaking process, while the molten steel represents the final condition after the blowing and steelmaking process. Table 7 provides detail of the average components and temperatures for liquid iron and molten steel. Moreover, the initial conditions of the liquid iron are stable, indicating that they have minimal influence on the results of the industrial application test.
| Top-blowing Lance | Original oxygen lance | 12.0° new oxygen lance | 12.3° new oxygen lance | |
|---|---|---|---|---|
| Liquid iron | C (mass%) | 4.86 | 4.87 | 4.87 |
| P (mass%) | 0.087 | 0.086 | 0.087 | |
| Si (mass%) | 0.590 | 0.586 | 0.587 | |
| Temperature (°C) | 1361 | 1360 | 1360 | |
| Molten steel | C (mass%) | 0.06 | 0.06 | 0.06 |
| P (mass%) | 0.024 | 0.019 | 0.018 | |
| Temperature (°C) | 1641 | 1644 | 1643 | |
| Melting time (min) | 11.1 | 10.5 | 10.6 | |
During the industrial application research, the flow rates for all three types of oxygen lances remained constant at 10800 Nm3/h, and their lance heights ranged from 950 to 1200 mm. Additionally, the same slag charging method was employed for all types of oxygen lances. The late furnace lining structure was chosen as the testing condition for these oxygen lances, and the furnace life of the converter reached approximately 16 thousand heats.
To sum up, all the melting conditions remain the same and have limited impact on the results of the industrial application research. This means that the metallurgical indexes of the end-point for molten steel and liquid slag are determined solely by the structure of the oxygen lance.
The Sherwood number (Sh) is a dimensionless number used in mass-transfer operation, which represents the ratio of convective to diffusive mass transport. That means the mass transfer rate between molten bath and oxygen gas is enhanced by a bigger Sherwood number.36,37)
| (9) |
where, Re and Sc are the Reynolds number and Schmidt number respectively. k1, k2 and k3 are the positive constants depended on the dynamic conditions of molten bath, respectively. ρ and μ are the density (kg/m3) and viscosity (Pa·s) of molten bath, respectively. D and d are diffusion coefficient and characteristic length of molten bath, respectively. v is the velocity (m/s) of molten bath steam.
As mentioned, the initial condition of the liquid iron is basically the same for different oxygen lances. Thus, the initial parameters of molten bath, such as density, viscosity, diffusion coefficient and characteristic length, also have little influence on the steelmaking process. Consequently, increasing in velocity of molten bath would lead to a higher mass transfer rate, as indicated by a larger Sherwood number (Sh) value.
Based on the results of the water experiment and simulation, the stirring effect of molten bath using the 12.0° new oxygen lance is better than that using the original and 12.3° new oxygen lances, due to a shorter mixing time and a higher velocity of the molten bath. Hence, the ascending order of the both Sh and mass transfer rate between molten bath and oxygen gas generated by the different oxygen lances is as follows: original oxygen lance, 12.3° new oxygen lance, and 12.0° new oxygen lance.
The utilization rate of oxygen gas is increased with a greater mass transfer rate between molten bath and oxygen gas, which reduces the melting time in the steelmaking process. As a result, both 12.0° and 12.3° new oxygen lances achieve shorter melting time compared to the original oxygen lance. Additionally, the melting time generated by the 12.3° new oxygen lance is the shortest, as shown in Table 7.
The dephosphorization rate can be calculated from the difference of [P] content in liquid iron and molten steel, referring to the Table 7. Figure 12 illustrates the distribution of dephosphorization rate of molten bath using three kinds of oxygen lances.
Based on the results obtained from the water experiment and numerical simulation, the ascending order of the impaction area of molten bath generated by the different oxygen lances is as follows: original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lance. This order is opposite to the interface area between the liquid slag and molten steel.
An increase in impaction diameter causes an elevation in FeO content in the liquid slag. When the interface area between the liquid slag and molten steel is improved, FeO in the liquid slag would be reacted quickly with [C], [Si] and [Mn] in the molten bath, thereby increasing viscosity of the liquid slag. Moreover, a higher FeO content in the liquid slag decreases the viscosity of the slag, which is beneficial for improving the dephosphorization rate.
Based on the result, when the operational conditions are the same, the ascending order of FeO content in the liquid slag formed by the various oxygen lances is as follows: original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lance. Thus, the dephosphorization rate of the molten steel achieved by the original oxygen lance, 12.0° and 12.3° new oxygen lances are 72.0, 77.7 and 79.5%, respectively. Figure 12 represents the variation in the dephosphorization rate of molten bath using original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lances are ±11.5, ±6.3 and ±4.8%, respectively. This indicates that the new oxygen lance obviously improve the dephosphorization rate and stability, comparing with the original oxygen lance. The both dephosphorization rate and stability are further enhanced by the inclination angle of 12.3°, for two kinds of new oxygen lances.
Because of the suitable viscosity of liquid slag, the occurrence of molten steel spilling is reduced with the use of the new oxygen lances. A reduction in occurrence of molten steel spilling leads to a lower ferrous charge consumption, during the steelmaking process. Based on the result, the ferrous charge consumption generated by the original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lances are 1063.3, 1059.3 and 1058.8 kg/t, respectively. Thus, the ferrous charge consumption is significantly decreased using the new oxygen lances, comparing with the original oxygen lance.
Figure 13. depicts the main content of end-point slag. The result shows the basicity of the end-point slags for all three types of oxygen lances is approximately 2.5, as a result of using the same slag charging method. The FeO contents in the end-point slag formed by the original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lances are 10.7, 12.0, 12.3 mass%, respectively. This proves that the 12.0° and 12.3° new oxygen lances could increase the FeO content in the liquid slag, which achieves a suitable viscosity of liquid slag, comparing with the original oxygen lance.
The decreasing in FeO content in liquid slag leads to a reduction in the fluidity of liquid slag. Additionally, the sticky slag suppresses the transport efficiency of metallic iron (M. Fe) from liquid slag into molten bath, resulting in a higher M. Fe content in endpoint slag. Consequently, the M. Fe contents in the endpoint slag formed by the original oxygen lance, 12.0° new oxygen lance, and 12.3° new oxygen lances are 6.2, 4.6, 4.4 mass%, respectively.
The [P] content requirement for 95 percent of the finished steel produced by the 35-ton top-blowing converter is less than 0.035 mass%. Both the 12.0° and 12.3° new oxygen lances can meet this requirement. Ultimately, the 12.3° new oxygen lance is selected as the appropriate oxygen lance for the late furnace lining structure, due to lower ferrous charge consumption, for the Panshi Jianlong Iron and Steel Company.
In this paper, the dynamic conditions of the molten bath generated by various oxygen lances were investigated by the water experiments and numerical simulations, at both initial and late furnace lining structure. In order to determine the appropriate oxygen lance, the metallurgical effects of original and optimized oxygen lances in the steelmaking process also were tested at the late furnace lining structure. The main conclusions can be described as follows:
(1) At the same oxygen flow rate and lance height, the impact area of the molten bath formed by the late furnace lining structure is larger than that formed by the initial furnace lining structure, due to the larger furnace diameter. However, because of the smaller depth of the molten bath, the impact depth of the molten bath formed by the initial furnace lining structure is bigger than that formed by the late furnace lining structure, under the same operational conditions.
(2) Comparing with the initial furnace lining structure, the mixing effect of the molten bath is significantly improved when using the original oxygen lance with the late furnace lining structure. This improvement is further enhanced by new oxygen lances with inclination angles of 12.0° and 12.3°, based on the mixing time, average velocity, and volume of the velocity dead-zone.
(3) The flow directions between the eddy of oxygen jet backflow and the eddy of furnace gas are completely opposite, which reduces the average velocity at the A region. Meanwhile, the furnace gas eddy keeps accelerating the velocity of molten bath within the B region, resulting in a larger high-velocity volume in the B region compared to the A region.
(4) Due to a larger furnace diameter, the average interface is significantly enhanced in the later furnace lining structure. This improvement in the mass transfer rate of FeO between the liquid slag and molten bath leads to a rapid reaction of FeO with [C], [Si], and [Mn] present in the molten bath. Consequently, the viscosity of the liquid slag increases, and there is a higher occurrence rate of molten steel spilling in the steelmaking process.
(5) From the results of the industrial plan test, the 12.3° new oxygen lance achieves the highest dephosphorization rate at 79.5%, and the lowest ferrous charge consumption at 1058.8 kg/t, comparing with the original and 12.0° new oxygen lances. As a result, the 12.3° new oxygen lance is determined to be the appropriate oxygen lance for the late furnace lining structure.
The authors would like to express their thanks for the support by National Natural Science Foundation of China (NSFC 52322407, NSFC 52074024 and NSFC 52274313).
On behalf of all authors, the corresponding author states that there is no conflict of interest.
