A CFD Study on Refractory Wear in RH Degassing Process

Abstract

In order to assess the wear damage of the lining refractory in the RH degasser, a transient 3D numerical model has been established using volume of fluid approach-discrete phase model (VOF-DPM) technology. The gas-oil-water three-phase flow in a RH degasser water model was evaluated. The breakup and coalescence of gas bubble was taken into account, and moreover the bubble diameter changed with static pressure. The wall shear stress and turbulence intensity were employed to predicate the erosion rate of the lining refractory, while the diffusion coefficient of the refractory material and the slag property at high temperature were used to consider the corrosion rate. The effects of the operational parameters on the refractory wear rate were clarified. A careful comparison between the experimental and the numerical results was conducted for the model validation. The results show that the wear behavior of the lining refractory at the up snorkel wall is the most severe due to the rapidly rising bubble. The vacuum chamber wall that near the up snorkel is also subjected to a serious wear damage. Besides, a higher wear rate is observed at the ladle wall that close to the oil/water interface, since both the physical erosion and chemical corrosion contribute to the wear damage of the lining refractory here. The developed model could help smelters to estimate the remaining thickness of the refractory in the RH degasser under different operational conditions.

1. Introduction

The Ruhrstahl-Heraeus (RH) degasser is widely used for the production of the ultra-clean steel, since it could efficiently reduce carbon and sulfur contents in molten steel.^1,2,3) It is well known that the complicated flow of the molten steel in the RH degasser, caused by the vacuum and the lifting gas, would create a significant wear and tear of the lining refractory.^4,5,6) Besides, the slag impregnation, depended on the chemical composition and the smelting temperature, also induces the degradation of the lining refractory.^7,8,9)

In order to improve the service life of the lining, it is valuable to gain a deeper understanding of the refractory wear behavior during the RH degassing process. Running experiments on a real device is the most intuitive approach for studying the refractory wear in the RH degasser. However, in the majority of cases, the measurement limitation and the harsh process environment result in the information provided by the experiments being inadequate.¹⁰⁾ The critical phenomena such as the flow pattern of the molten steel, the floating motion of the gas bubble, and the fluctuation of the slag/steel interface, are difficult to be accurately assessed. With the continuous increasing of the computation resource and numerical technique, the computational fluid dynamics (CFD) method, which is able to afford a comprehensive data, becomes a potent tool to investigate the performance of the lining refractory in the RH degasser.^{11,12,13,14,15,16)} Furthermore, the simulated results could be verified through the application of the water model experiment.

Many numerical models have been established to figure out the multiphase flow in the RH degasser by using Eulerian-Lagrangian or Eulerian-Eulerian approach.¹⁷⁾ The movement of the interface between the gas bubble and the molten steel could be satisfactorily evaluated. The fluctuations of the molten steel free surface in the vacuum chamber and the slag/steel interface in the ladle however were not taken into account. A coupled volume of fluid approach-discrete phase model (VOF-DPM) approach has received increasingly attention, since it is capable of simultaneously describing the surface fluctuation and the bubble rising.^14,16,18,19) The VOF method was invoked to define the motion of the continuous phases such as the molten slag and molten steel, meanwhile the DPM model was used to illustrate the movement behavior of the discrete phase such as the gas bubble and non-metallic inclusion. Due to the presentation of the fluctuation of the slag/steel interface in the ladle, it is able to numerically clarify the damage of the lining refractory around the refractory/slag/steel interface.²⁰⁾ The wear rate of the lining refractory was supposed to be associated with the molten slag corrosion and the molten steel erosion. Besides, the erosion of the refractory, caused by the floating gas bubble, could also well be recognized.²¹⁾ The distribution of the wall shear stress in the up and down snorkels, as well as the vacuum chamber was predicated by the DPM technique.

A thorough search of the relevant literature indicates that few efforts have been devoted to show the wear behavior of the lining refractory, including the molten steel erosion and molten slag corrosion, during the RH degassing process. Because of this, the authors were motivated to establish a transient 3D numerical model to figure out the gas-oil-water three-phase flow in a RH degasser water model using the VOF-DPM technology. The breakup and coalescence of the gas bubble was taken into account during the flotation process, and moreover the bubble diameter was supposed to change with the static pressure. The wall shear stress and turbulence intensity were employed to predicate the erosion rate of the lining refractory, while the refractory diffusion coefficient and the molten slag property at high temperature were used to assess the corrosion rate. The effects of the operational parameters such as the volume flow rate of the lifting gas, vacuum degree, and snorkel immersion depth on the wear rate of the lining refractory were clarified. To examine the model accuracy, a careful comparison between the experimental and the numerical results was conducted.

2. Mathematical Model

2.1. Assumptions

The following assumptions have been adopted to simplify the model:

(1) The gas in the vacuum chamber, oil, and water were assumed to be continuous phase, while the gas bubble was treated as discrete phase. The fluctuations of the free surface of water and oil/water interface were tracked using VOF method. The movement of the rising bubble was described by DPM technique.

(2) Since the water model experiment was performed at room temperature, the fluid flow was assumed to be an isothermal process.

(3) The density of the continuous phase was assumed to be constant, while it was a function of the static pressure associated with the gas bubble. The viscosities of the continuous and discrete phases kept constant.¹⁸⁾

(4) The gas bubble was idealized as a non-sphere with a 0.3 shape factor.¹⁹⁾

(5) Since it is difficult to find a reasonable and accurate numerical module for the quantification of the interaction between the erosion and corrosion, the final wear rate was assumed to be the sum of the erosion and corrosion rates. It should be stated that the calculation method of the final wear rate is questionable, which is needed to be improved.

2.2. Flow Pattern of Continuous Phase

As mentioned above, there were three continuous phases in the RH degasser, i.e. the gas in the vacuum chamber, oil, and water. The VOF approach was invoked to represent the three-phase flow, which was well demonstrated in Refs. 18, 19, and 22. Only a basic description was given here. The volume fraction of each phase was tracked in the entire computational domain, and was updated at every time step. The properties of the continuous phase were related to the volume fraction. A single momentum equation, which was dependent on the volume fractions of the three phases through the density and viscosity, was solved throughout the domain. Furthermore, a source term that describes the momentum exchange between the molten steel and the gas bubble was included, since the two-way coupled Euler-Lagrange approach was adopted in the DPM technique. The realizable k-ε turbulence model was employed to calculate the turbulent viscosity, because it incorporates a substantial improvement over the standard k-ε model in which the flow structure includes strong streamline curvature, vortices, and rotation.²³⁾ An enhanced wall function was invoked to work in conjunction with the realizable k-ε turbulence model.

2.3. Bubble Tracking

Figure 1 represents the forces acting on the gas bubble when it moves upward in the up snorkel. Prediction of the trajectory of each non-spherical gas bubble was made by integrating the following equation, which considered the contributions of five different forces:   

$${\rho }_b\frac{\pi }{6}{d}_b^3\frac{d{\overrightarrow{v}}_b}{dt}={\overrightarrow{F}}_g+{\overrightarrow{F}}_b+{\overrightarrow{F}}_d+{\overrightarrow{F}}_v+{\overrightarrow{F}}_{𝓁}$$(1)

where ρ_b was the density of the gas bubble, which varied with the static pressure according to the ideal gas law. d_b and ${\overrightarrow{v}}_b$ were the equivalent diameter and velocity of the gas bubble, respectively, and t was the time. The terms on the right-hand side were the gravity, buoyancy, drag, virtual mass, and lift forces, respectively. Besides, the random walk module was invoked for including the chaotic effect of the turbulence.

Fig. 1.

Schematic of the gas bubble motion in the up snorkel of RH degasser. (Online version in color.)

Stochastic collision module was utilized to consider the collision and coalescence between two gas bubbles in the present work. If the center of the bubble i passed within a flat circle centered around the bubble j with the area of $\raisebox{1ex}{$\pi $}\!\left/ \!\raisebox{-1ex}{$4$}\right.{(d_{bi}+d_{bj})}^2$ perpendicular to the trajectory of the bubble i, a collision thus may be assumed to take place between the two bubbles i and j. The stochastic estimate method, O’Rourke algorithm, would subsequently be switched on to calculate the chance of collision.^22,24)

2.4. Erosion and Corrosion

As mentioned above, the wear of the lining refractory is composed of physical erosion and chemical corrosion. The physical erosion is caused by the motion of the molten steel and the floating gas bubble at high temperature, while the chemical corrosion is induced by the high-temperature molten slag. The overall wear rate is therefore the sum of the erosion and corrosion rates.

The erosion of the lining refractory could be regarded as the plasticity erosion. The shearing stress was generated because of the relative movement between the molten steel and the lining refractory. The erosion rate of the lining refractory therefore could be estimated by the shearing stress and turbulence intensity:^6,25)   

$$\begin{array}{l}{R}_{erosion}=-2.061\times {10}^{-3}{(\tau \cdot I)}^2\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+2.642\times {10}^{-2}(\tau \cdot I)+5.636\times {10}^{-1}\end{array}$$(2)

where R_erosion implied the erosion rate of the lining refractory, expressed by the loss of the refractory thickness per time, τ was the wall shear stress, and I was the turbulence intensity.

The corrosion of the lining refractory was first created by the dissolution of the refractory oxide in the molten slag at the air/slag/steel interface and was promoted by Marangoni convection, resulting from the surface tension gradient.^5,8,26,27) The corrosion rate could be obtained by the following equation:^8,28)   

$$R_{corrosion}=k_{eff}{(\epsilon -{\epsilon }_c)}^{1.2}$$(3)

  

$$\epsilon =2.27\times {10}^{-7}\frac{h_s{\eta }_s}{D^2}{\beta }^2$$(4)

  

$$\beta =\sqrt{\frac{4D{u}_m}{\pi h_s}}{\left(\frac{D{\rho }_s}{{\eta }_s}\right)}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$6$}\right.}$$(5)

where k_eff was the coefficient of the corrosion rate. ε_c and D were the threshold value of the dissolution rate and the diffusion coefficient of the refractory material, which was considered as the magnesia-carbon refractory in the present work. u_m denoted the Marangoni flow rate, which was determined by the concentration and temperature gradients around the refractory/slag/steel interface.^29,30) h_s was the thickness of the molten slag layer, η_s and ρ_s were the dynamic viscosity and density of the molten slag, respectively. The numerical values of all the parameters used in the calculation are listed in Table 3.

Table 3. Numerical values of the parameters used in the calculation.

Parameter	Value
Coefficient of the corrosion rate	0.548
Density of molten slag, kg/m3	2780
Viscosity of molten slag, Pa∙s	0.154
Threshold value of the dissolution rate of the magnesia-carbon refractory, kg/s	6.24×10−5
Diffusion coefficient of the magnesia-carbon refractory, m2/s	4.82×10−7

It should be declared that the wall shear stress and turbulence intensity used for the estimation of the erosion rate were calculated based on the RH degasser water model as mentioned above. However, the parameters such as the dissolution rate and diffusion coefficient of the refractory material used for the evaluation of the corrosion rate were associated with the actual refractory material and slag, and were irrelevant to the RH degasser water model. The final wear rate of the lining refractory was the sum of the erosion and corrosion rates, which was supposed to be more close to the wear behavior of the lining refractory serviced in the industrial RH degasser.

2.5. Boundary Condition

According to the experimental observation, the initial equivalent diameter of the gas bubble at the inlet of the up snorkel was assumed to be 2.0 mm and the motion of the bubble was in a direction that perpendicular to the inlet. A mass flow rate was adopted at the inlet with a uniform value of the turbulent kinetic energy and dissipation rate. The motion of the gas bubble was arbitrarily terminated once the gas bubble floats up to the free surface of the water in the vacuum chamber. While the bubble would lose 60% momentum when it impact the solid walls.

A pressure outlet was applied to the top surface of the vacuum chamber, and the reference pressure was set to be the atmospheric pressure at the top surface of the ladle. In order to better define the free surface flow, zero shear stress was employed at the top surface of the ladle. As for the rest of the solid wall, a non-slipping was applied with a roughness equal to 1×10⁻⁵ m.³¹⁾

3. Solution Procedure

The numerical simulation was carried out using the commercial software ANSYS-FLUENT 14.5. The governing equations for the turbulent flow and bubble movement were integrated over each control volume and simultaneously solved using an iterative procedure. We developed an in-house code for evaluating the trapping of the bubbles and the erosion and corrosion of the lining refractory. For the trapping of the gas bubble at the water free surface, the volume fraction of the water was tracked. If the volume fraction of the water was larger than zero, the gas bubble would keep moving, otherwise the motion of the bubble was arbitrarily stopped. For the wear of the lining refractory, the turbulence intensity and the wall shear stress were first obtained from the commercial software main program. The physical erosion and the chemical corrosion were then calculated. The widely used SIMPLE algorithm was employed for calculating the Navier-Stokes equations. In order to ensure greater accuracy, all the equations were discretized using the second order upwind scheme. Before moving on to the next step, the iterative procedure continued until all the normalized unscaled residuals were less than 10⁻⁶. A grid independence test was also conducted. Three grid models were generated, with the respective sizes of 8 mm, 14 mm, and 20 mm. After a pre-simulation, the magnitude of velocity of some points in the domain were compared carefully. The deviation of predicated results between the first and second mesh was about approximately 4.8%, while it was approximately 8.7% between the second and third meshes. Furthermore, the values of y+ within the three different grid models adjacent to the wall were equal to ~1. Considering the high computational cost, the second mesh, as shown in Fig. 2, was therefore retained for the simulation. A time step of 0.01 s, was employed for the iteration of the simulation, so as to ensure that the above aforementioned convergence criteria were fulfilled. Using eight cores of a basic frequency of 3.50 GHz, the calculations for a typical case took approximately 152 CPU hours.

Fig. 2.

Boundaries of the computational domain and the mesh used in the simulation. (Online version in color.)

4. Water Model Experiment

Figure 3 displays the RH water model experimental device, built in polymethyl methacrylate with a scale factor λ = 1:6. The nitrogen gas was injected at the up snorkel through 16 nozzles equally distributed along the circumference of the up snorkel in upper and lower rings. The vacuum chamber was connected to the vacuum pump, and the level of water in the vacuum chamber was remained constant during the experiment. Besides, an ultrasonic flowmeter was attached to the down snorkel for measuring the circulation flow rate of water. To understand the mixing characteristics, conductivity experiment has been carried out. Saturated NaCl solution was injected from the top of the vacuum chamber just above the up snorkel. A conductivity sensor was positioned at the bottom of the ladle close to the wall for continuously measuring the variation of the electrical conductivity. The detailed dimensional, property, and operational parameters of the water model experiment are listed in Tables 1 and 2.

Fig. 3.

Photo of experimental device. (Online version in color.)

Table 1. Dimensions of RH water model.

Parameter	Value
Up/down diameter of ladle, mm	683/623
Height of ladle, mm	708
Diameter of vacuum chamber, mm	403
Height of vacuum chamber, mm	325
Diameter of snorkels, mm	125
Height of snorkels, mm	283
Diameter of gas inlet, mm	4
Height of lower inlets from the bottom of snorkel, mm	90
Distance between the lower inlets and up inlets, mm	25
Number of gas inlet	16

Table 2. Property and operational parameters of RH water model.

Parameter	Value
Density of water, kg/m3	998.2
Viscosity of water, Pa∙s	1.003×10−3
Density of oil, kg/m3	900
Viscosity of oil, Pa∙s	0.058
Density of gas, kg/m3	1.38
Viscosity of gas, Pa∙s	1.663×10−5
Surface tension between oil and water, N/m	0.041
Surface tension between gas and water, N/m	0.073
Volume flow rate of lifting gas, NL/min	20.28/30.42/40.56
Vacuum degree, Pa	3136/3626/4116
Immersion depth of snorkel, mm	60/80/100
Oil thickness, mm	10/30/50

5. Results and Discussion

5.1. Flow Characteristics

Figure 4 illustrates the flow pattern and phase distribution in the RH degasser. It is clear that a circulation flow is observed between the vacuum chamber and the ladle. An upward flow is first created in the up snorkel under the action of the lifting gas, while a downward flow is generated in the down snorkel. Because the mixture density in the down snorkel would increase, resulting in a greater gravity, after the escaping of the lifting gas from the free surface of the water in the vacuum chamber. Besides, the downward flow is divided into two flow streams when it enters the ladle. Some water moves to the bottom of the up snorkel by a shortest path, creating a noticeable short-circuit flow. The rest part of water could reach the bottom of the ladle, which promotes the stirring efficiency.

Fig. 4.

Flow pattern and phase distribution in the RH degasser. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Figure 5 indicates the motion of the gas bubble in the RH degasser. The velocity of the gas bubble along the horizontal direction rapidly decays due to the resistance of water. In the meantime, the velocity of the gas bubble along the vertical direction gradually increases under the effect of the buoyancy. The diameter of the gas bubble changes all the time during the flotation process as a result of the breakup and coalescence, as well as the decreasing static pressure as shown in Fig. 6. The equivalent diameter of the gas bubble in the up snorkel and the vacuum chamber varies from 0.54 mm to 46.8 mm. Besides, due to the influence of the lift force, the larger bubble tends to shift to the center of the up snorkel, while the smaller one inclines to adhere to the wall. The volume fraction of the gas at the center of the up snorkel would therefore increase. It can be seen from Fig. 7 that there is no gas at the center of the lower part of the up snorkel such as planes 1 and 2, which position is demonstrated in Fig. 8. With the migration of the larger bubble, the gas volume fraction at the center of the upper part of the up snorkel gradually increases. The gas volume fraction at plane 6 obviously becomes more uniform.

Fig. 5.

Morphology of gas bubble in the RH degasser. The diameter of the sphere is proportional to the equivalent diameter of the gas bubble, and the equivalent diameter of the gas bubble varies from 0.54 mm to 46.8 mm. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Fig. 6.

Distribution of static pressure on the vertical section Y = 0.0 m. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Fig. 7.

Distribution of the gas volume fraction on the observation planes in the up snorkel of the RH degasser. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Fig. 8.

Schematic of the observation planes and points. (Online version in color.)

5.2. Model Validation

As mentioned above, the saturated NaCl solution was injected from the top of the vacuum chamber just above the up snorkel in the experiment. The electrical conductivity of the water was then examined at point 2 as shown in Fig. 8. In the simulation, a dimensionless scalar was released at point 1 after the flow field achieves stability in the RH degasser. The variation of the scalar concentration of point 2 over time thus could be observed.

According to the experimental observation and the simulation results, it is believed that flow regime in the up snorkel of the RH degasser is the bubbly flow. Because there is a distribution of bubbles of various sizes throughout the water inside the up snorkel. As the gas flow rate increases, the average bubble size increases, resulting in a larger buoyancy. The larger buoyancy would then promote the water movement and also the turbulence intensity.

Figure 9 indicates the comparison of the dimensionless concentration of the tracer between the measurement and simulation. The mixing time was supposed to be the time when the dimensionless content achieved the range of ±3 pct of the final dimensionless content. The calculated mixing time is 43.4 s which is close to the measured mixing time 39.8 s as illustrated in Fig. 10. The discrepancy between the measured and simulated circulation flow rate is negligible small. The agreement between the model results and the experimental data gives confidence in the fundamental validity of the established model.

Fig. 9.

Variation of the tracer dimensionless concentration versus time. (Online version in color.)

Fig. 10.

Comparison of the mixing time and circulation flow rate between the measurement and the simulation. (Online version in color.)

It should be stated here that the comparison of the mixing time and the circulation flow rate could well verify the macroscopic flow in the RH degasser. The linear velocity, turbulence intensity, and shear stress on the wall however is closer to the predicating of the wear rate of the lining refractory. It is therefore necessary to study the micro-flow situation of the water on the wall, which would be focused on in our next work.

5.3. Wear Behavior of Refractory

Figure 11 represents the distribution of the overall wear rate on the walls of the RH degasser. It is obvious that the wear of the lining refractory at the up snorkel wall is the most severe, in which the highest wear rate reaches 13.41 mm/h. The rapidly rising bubble in the up snorkel could be responsible for the serious wear damage. Owing to the drive of the bubbles, the velocity magnitude of water in the up snorkel is higher than other places in the RH degasser, resulting in a greater wall shear stress as described in Fig. 12. The wall of the vacuum chamber that near the up snorkel is also subjected to a serious wear damage because of the higher shear stress. Additionally, a higher wear rate is observed at the ladle wall that close to the oil/water interface. Because both the physical erosion and the chemical corrosion contribute to the wear behavior of the lining refractory here. Furthermore, due to the weak flow, the effect of the physical erosion lags behind that of the chemical corrosion as displayed in Fig. 13. The intensity of the corrosion above the slag line is higher than that below the slag line.

Fig. 11.

Distribution of the overall wear rate on the walls. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Fig. 12.

Distribution of the wall shear stress on the walls. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

Fig. 13.

Distribution of the corrosion rate on the walls. The immersion depth, vacuum degree, gas volume flow rate, and oil thickness are 80 mm, 3626 Pa, 30.42 NL/min, and 30 mm, respectively. (Online version in color.)

5.4. Effects of Operational Parameters

The influences of the operational parameters such as the volume flow rate of the lifting gas, vacuum degree, immersion depth of the snorkel, and oil thickness on the wear rate of different walls have been clarified in detailed. From Fig. 14, it can be seen that the area weighted average wear rates of the up snorkel wall, down snorkel wall, vacuum chamber wall, and ladle wall become greater with the increasing gas volume flow rate. The increase in the gas volume flow rate promotes the movement of water in the RH degasser, resulting in a larger circulation flow rate as indicated in Fig. 15. As a result, the shear stresses of all the walls mentioned above become higher, and thus the lining refractory are more susceptible to wear degradation. In all cases, the wear damage of the up snorkel wall is the most serious, followed by the vacuum chamber and ladle walls, and the wear damage of the down snorkel wall is the slightest. On the other hand, the wear rate of all the four walls would increase with the increasing snorkel immersion depth. The main reason is that the pressure difference between the gas inlet and the free surface of water in the vacuum chamber becomes greater, which would improve the driving force for the rising of the bubble. The faster bubbles encourage the movement of water as well as the wear damage of the lining refractory in the RH degasser. The contributions of the increasing gas volume flow rate and immersion depth on the average wear rate however gradually diminish. The average wear rate of the up snorkel wall increases by 12.45% with the gas volume flow rate ranging from 20.28 NL/min to 30.42 NL/min and a constant immersion depth of 80 mm, while only increases by 4.77% if the gas volume flow rate continuously increases to 40.56 NL/min. The increasing rate of the average wear rate of the up snorkel wall is 13.76% when the immersion depth increases from 60 mm to 80 mm and the gas volume flow rate is kept at 30.42 NL/min, and drops to 7.84% while the immersion depth ranging from 80 mm to 100 mm.

Fig. 14.

Effects of the gas volume flow rate and snorkel immersion depth on the average wear rate of different positions in the RH degasser: (a) up snorkel wall, (b) down snorkel wall, (c) vacuum chamber wall, and (d) ladle wall. (Online version in color.)

Fig. 15.

Effects of the gas volume flow rate, snorkel immersion depth, vacuum degree and, oil thickness on the circulation flow rate. (Online version in color.)

It is well known that the vacuum degree significantly affects the motion of the bubble and water in the vacuum chamber. A larger vacuum degree definitely improves the driving force for the rising of the bubble. The flow of water in the vacuum chamber therefore becomes increasingly turbulent. The average wear rate of the vacuum chamber wall is inclined to increase with a higher vacuum degree as shown in Fig. 16. Also note that the increasing rate of the wear rate reduces if more lifting gas is injected. The average wear rate of the vacuum chamber wall increases by 9.88% with the vacuum degree ranging from 3136 Pa to 4116 Pa and a constant gas volume flow rate of 20.28 NL/min, while the increasing rate reduces to 2.83% with the same change of the vacuum degree and a constant gas volume flow rate of 40.56 NL/min. Because the influence of the vacuum degree on the water movement lags behind that of the lifting gas. It can be inferred that the increase in the gas volume flow rate is more efficient than that in the vacuum degree if we want to improve the circulation flow rate and shorten the mixing time.

Fig. 16.

Effect of the vacuum degree on the average wear rate of the vacuum chamber wall. (Online version in color.)

Figure 17 represents the effect of the oil thickness on the average wear rate of the ladle wall. The wear damage of the ladle wall becomes more serious with a thicker slag layer as expected due to the aggravated corrosion. Furthermore, the increasing of the wear rate with the thicker oil layer is more pronounced at a larger gas volume flow rate. The average wear rate increases by 5.35% with the oil thickness ranging from 10 mm to 50 mm and a constant gas volume flow rate of 20.28 NL/min, but the increasing rate increases to 11.29% with the same change of the oil thickness and a constant gas volume flow rate of 40.56 NL/min. Because a larger gas volume flow rate would promote the water flow and the oil/water interface fluctuation in the ladle, resulting in an increase in both physical erosion and chemical corrosion.

Fig. 17.

Effect of the oil thickness on the average wear rate of the ladle wall. (Online version in color.)

In the present work, a transient 3D numerical model has been developed to figure out the wear behavior of the lining refractory in the RH degasser. The influences of the operational parameters such as the gas volume flow rate, vacuum degree, snorkel immersion depth, and oil thickness on the wear rate of different walls have been quantitatively analyzed. The detailed results could help smelters to estimate the remaining thickness of the refractory in the RH degasser. Also the present numerical model is supposed to be a powerful tool for further investigations concerning the high-temperature chemical interaction between the refractories and the molten slags with different composition, which is expected to exert a great impact on the damage of the lining refractory in the RH degassing process.

6. Conclusions

In order to assess the wear damage of the lining refractory in the RH degasser, a transient 3D numerical model has been developed using the VOF-DPM technology. The gas-oil-water three-phase flow in a RH degasser water model was evaluated. The contributions of the physical erosion and chemical corrosion to the wear rate of the lining refractory were included. The effects of the operational parameters such as the gas volume flow rate, vacuum degree, snorkel immersion depth, and oil thickness on the wear rate were clarified. To examine the model accuracy, a careful comparison between the experimental and the numerical results was conducted.

Our study shows that the wear of the lining refractory at the up snorkel wall is the most severe due to the rapidly rising bubble. The vacuum chamber wall that near the up snorkel is also subjected to a serious wear damage. Besides, a higher wear rate is observed at the ladle wall that close to the oil/water interface, since both the physical erosion and chemical corrosion contribute to the wear damage of the lining refractory here.

The increase in the gas volume flow rate and snorkel immersion depth, creating a larger circulation flow rate, promotes the wear degradation of the lining refractory, especially at the up snorkel wall. The contributions of the increasing gas volume flow rate and immersion depth on the wear rate however gradually diminish. The average wear rate of the vacuum chamber wall is inclined to increase with a higher vacuum degree. Besides, the increase in the gas volume flow rate is supposed to be more efficient than that in the vacuum degree if we want to improve the circulation flow rate and shorten the mixing time. Due to the aggravated corrosion, the wear damage of the ladle wall becomes more serious with a thicker slag layer.

The present model could help smelters to estimate the remaining thickness of the refractory in the RH degasser under different operational conditions, and is a powerful tool for further investigations concerning the high-temperature chemical interaction between the refractory and the molten slag, which is expected to exert a great impact on the damage of the lining refractory.

Acknowledgements

The authors’ gratitude goes to National Natural Science Foundation of China (Grant No. 51974211) and the Special Project of Central Government for Local Science and Technology Development of Hubei Province (Grant Nos. 2019ZYYD003 and 2019ZYYD076). Thanks are also given to Prof. Yongxiang Yang at Delft University of Technology for his very helpful advice on numerical simulation and Baoshan Iron & Steel Co., Ltd. for providing the supporting data.

References

• 1)   Y. G.  Park and  K. W.  Yi: ISIJ Int., 43 (2003), 1403. https://doi.org/10.2355/isijinternational.43.1403
• 2)   M. A.  van Ende,  Y. M.  Kim,  M. K.  Cho,  J.  Choi and  I. H.  Jung: Metall. Mater. Trans. B, 42 (2011), 477. https://doi.org/10.1007/s11663-011-9495-4
• 3)   C. Y.  Zhu,  P. J.  Chen,  G. Q.  Li,  X. Y.  Luo and  W.  Zheng: ISIJ Int., 56 (2016), 1368. https://doi.org/10.2355/isijinternational.ISIJINT-2016-124
• 4)   P. H. R.  Vaz de Melo,  J. J. M.  Peixoto,  G. S.  Galante,  B. H. M.  Loiola,  C. A.  da Silva,  I. A.  da Silva and  V.  Seshadri: J. Mater. Res. Technol., 8 (2019), 3764. https://doi.org/10.1016/j.jmrt.2019.06.036
• 5)   V. A.  Rovnushkin,  L. M.  Aksel’rod,  L. A.  Smirnov,  S. A.  Spirin,  T. V.  Yarushina,  I. G.  Maryasev,  É. V.  Visloguzova and  S. Yu.  Fefelov: Refract. Ind. Ceram., 56 (2015), 113. https://doi.org/10.1007/s11148-015-9794-4
• 6)   A.  Huang,  H. Z.  Gu,  M. J.  Zhang,  N.  Wang,  T.  Wang and  Y.  Zou: Metall. Mater. Trans. B, 44 (2013), 744. https://doi.org/10.1007/s11663-013-9805-0
• 7)   E.  Blond,  N.  Schmitt,  F.  Hild,  P.  Blumenfeld and  J.  Poirier: J. Am. Ceram. Soc., 90 (2007), 154. https://doi.org/10.1111/j.1551-2916.2006.01348.x
• 8)   P. F.  Lian,  A.  Huang,  H. Z.  Gu,  Y. S.  Zou,  L. P.  Fu and  Y. J.  Wang: Ceram. Int., 44 (2018), 1675. https://doi.org/10.1016/j.ceramint.2017.10.095
• 9)   V. A.  Rovnushkin,  É. A.  Visloguzova,  S. A.  Spirin,  E. V.  Shekhovtsov,  V. V.  Kromm and  A. A.  Metelkin: Refract. Ind. Ceram., 46 (2005), 193. https://doi.org/10.1007/s11148-005-0083-5
• 10)   D.  Mukherjee,  A. K.  Shukla and  D. G.  Senk: Metall. Mater. Trans. B, 48 (2017), 763. https://doi.org/10.1007/s11663-016-0877-5
• 11)   B. K.  Li and  F.  Tsukihashi: ISIJ Int., 40 (2000), 1203. https://doi.org/10.2355/isijinternational.40.1203
• 12)   D. Q.  Geng,  H.  Lei and  J. C.  He: Metall. Mater. Trans. B, 41 (2010), 234. https://doi.org/10.1007/s11663-009-9300-9
• 13)   F. S.  Qi,  Y.  Jiang,  L.  Yang,  F.  Wang and  B. K.  Li: Adv. Mater. Res., 396–398 (2011), 2266. https://doi.org/10.4028/www.scientific.net/AMR.396-398.2266
• 14)   H. T.  Ling,  F.  Li,  L. F.  Zhang and  A. N.  Conejo: Metall. Mater. Trans. B, 47 (2016), 1950. https://doi.org/10.1007/s11663-016-0669-y
• 15)   K. F.  Feng,  A.  Liu,  K. J.  Dai,  S.  Feng,  J.  Ma,  J. Y.  Xie,  B.  Wang,  Y. W.  Yu and  J. Y.  Zhang: Powder Technol., 314 (2017), 649. https://doi.org/10.1016/j.powtec.2017.01.043
• 16)   W. X.  Dai,  G. G.  Cheng,  S. J.  Li,  Y.  Huang,  G. L.  Zhang,  Y. L.  Qiu and  W. F.  Zhu: ISIJ Int., 59 (2019), 1214. https://doi.org/10.2355/isijinternational.ISIJINT-2018-826
• 17)   C.  Liu,  S. S.  Li and  L. F.  Zhang: Acta Metall. Sin., 54 (2018), 347 (in Chinese). https://doi.org/10.11900/0412.1961.2017.00429
• 18)   G. J.  Chen,  S. P.  He and  Y. G.  Li: Metall. Mater. Trans. B, 48 (2017), 2176. https://doi.org/10.1007/s11663-017-0992-y
• 19)   H. T.  Ling and  L. F.  Zhang: Metall. Mater. Trans. B, 50 (2019), 2017. https://doi.org/10.1007/s11663-019-01583-3
• 20)   L. M.  Aksel’rod,  T. V.  Yarushina,  A. A.  Platonov,  A. O.  Migashkin,  I. V.  Anoshkin,  A. A.  Bessol’nikov and  M. A.  Eroshin: Refract. Ind. Ceram., 59 (2018), 115. https://doi.org/10.1007/s11148-018-0192-6
• 21)   Q. X.  Yang,  B.  Björkman and  G.  Carlsson: Steel Res., 68 (1997), 107. https://doi.org/10.1002/srin.199700549
• 22)   L. M.  Li,  B. K.  Li and  Z. Q.  Liu: ISIJ Int., 57 (2017), 1980. https://doi.org/10.2355/isijinternational.ISIJINT-2017-069
• 23)   H. J.  Zhu,  Y. H.  Lin,  D. Z.  Zeng,  Y.  Zhou,  J.  Xie and  Y. P.  Wu: Eng. Fail. Anal., 22 (2012), 83. https://doi.org/10.1016/j.engfailanal.2012.01.007
• 24)   J.  Zhang,  J. C.  Mi and  H.  Wang: Aerosol Sci. Tech., 46 (2012), 622. https://doi.org/10.1080/02786826.2011.649809
• 25)   Z.  Dai,  X. G.  Niu and  S. M.  Shen: J. Petrochem. Univ., 20 (2007), 85 (in Chinese).
• 26)   B.  Ertuğ: Adv. Eng. Forum, 26 (2018), 9. https://doi.org/10.4028/www.scientific.net/AEF.26.9
• 27)   I.  Pérez,  I.  Moreno-Ventas and  G.  Ríos: Ceram. Int., 45 (2019), 9788. https://doi.org/10.1016/j.ceramint.2019.02.015
• 28)   S.  Vollmann and  H.  Harmuth: Proc. Unified Int. Technical Conf. on Refractories (UNITECR 2013), ed. by The American Ceramic Society, John Wiley & Sons, Hoboken, NJ, (2014), 857.
• 29)   Y.  Chen: Ph.D. thesis, McMaster University, (2011), https://macsphere.mcmaster.ca/bitstream/11375/11677/1/fulltext.pdf, (accessed 2018-05-11).
• 30)   Y. N.  Wang,  Y. P.  Bao,  H.  Cui,  B.  Chen and  C.  Ji: J. Iron Steel Res. Int., 19 (2012), 1. https://doi.org/10.1016/S1006-706X(12)60064-8
• 31)   Y. G.  Park,  K. W.  Yi and  S. B.  Ahn: ISIJ Int., 41 (2001), 403. https://doi.org/10.2355/isijinternational.41.403