Babak Ghazian Tafrishi
Study of Argon Shrouding in Ingot Casting, with Focus on Improving the Operation at
Scana Björneborg Steel Plant
Master thesis (30 credits)
Engineering Materials Science Program (TTMVM), Industrial Materials (IMTA)
Date: March 2014 Supervisor: Mats Söder
R&D manager,
Scana Steel Björneborg
Examiner: Pär Jönsson
Professor,
Royal Institute of Technology
(KTH)
ABSTRACT
This thesis has been carried out as a development project at Scana Steel Björneborg with the purpose to study the influential parameters in argon shrouded ingot casting during the manufacturing of low-‐alloy steels.
In the first stage, a literature study was conducted in order to investigate the theoretical background of the procedure and the importance of protecting the melt during ingot casting.
Next, a computer model of the shield was designed using COMSOL Multiphysics® with regard to the process conditions at Scana Steel Björneborg. The effect of various parameters on the process was examined through simulations of the argon gas flow pattern, heat transfer between the gas and the melt stream, and the chemical species transport in the gas around the melt stream.
Based on the simulation results, two different shapes of shield were proposed for the argon shrouding operation. A set of implementation tests was executed in order to check the installation and usage conditions of the two new shields.
After deciding the proper shape of the shield, a full-‐scale ingot-‐casting test was performed with the selected shield to investigate the protection behavior. Moreover, the impact of the new casting-‐protection shield on the nitrogen and oxygen contents of steel was examined through sampling and analyzing the steel before and after casting.
It was found that the use of the new shield during the uphill ingot casting is an effective way to reduce the final nitrogen and oxygen contents of the casted ingot. Therefore, the new design of the shield can be used as a developed substitute for the protection of the melt stream in the ingot casting operation.
ACKNOWLEDGMENTS
I would like to express my sincere gratitude to my supervisor, Mats Söder, for all the guidance and supports. Your advices and the experience of working with you have provided me with invaluable inspirations.
I am grateful for all the help from Johan Lundin, Production Leader at Scana Steel, who trained me for sample preparation and analysis at the laboratory.
I also want to thank the operators at Scana Steel plant, who helped me with sampling during the project, and gave me practical tips and a good insight in the processes.
Thanks to all the people as Scana Steel Björneborg who were not only colleagues, but also friends. It has been a very interesting and worthwhile experience for me at Scana Steel in Björneborg.
I would like to specially thank Professor Pär Jönsson at Royal Institute of Technology (KTH), for all the supports and encouragements and also the beneficial discussions, which have motivated me through the project.
CONTENTS
1. Introduction ... 1
1-‐1. Background ... 1
1-‐2. Plant Description ... 5
1-‐3. Estimation of oxygen absorption ... 6
2. Methodology ... 11
2-‐1. Simulations ... 11
2-‐1-‐1. Assumptions ... 11
2-‐2. Parameters ... 12
2-‐2-‐1. Distance between the ladle and the shield ... 12
2-‐2-‐2. Distance between the shield and the trumpet ... 13
2-‐2-‐3. Argon gas flow rate ... 13
2-‐2-‐4. Argon inlets ... 14
2-‐2-‐5. Shield shape ... 14
2-‐3. Sampling ... 15
2-‐3-‐1. SP Samples ... 15
2-‐3-‐2. GP Samples ... 15
2-‐4. Measurements ... 16
2-‐4-‐1. Total oxygen content ... 16
2-‐4-‐2. Nitrogen content ... 16
3. Results and Discussions ... 17
3-‐1. Simulations and investigations of effective parameters ... 17
3-‐1-‐1. introductory simulations ... 17
3-‐1-‐2. the effect of argon gas flow rate ... 20
3-‐1-‐3. the effect of number of argon inlets ... 22
3-‐2. Simulations of recommended shield forms and the effect of distance between the shield and the trumpet ... 24
3-‐3. Implementation test of the proposed shield ... 29
3-‐4. Chemical analysis of the steel ... 31
4. Conclusions ... 34
References ... 35
1. INTRODUCTION 1-‐1. Background
Production of high quality steel has always been of concern to the steelmaking industry.
Quality of steel in today’s steel industry mainly refers to the cleanliness of steel. Reducing the amounts of the impurities in steel is among the most important means of improving the steel quality. During the past decades much research have been done on production of cleaner steels with better characteristics. It has been showed that the properties of steels are highly dependent on the presence of non-‐metallic inclusions and the size and distribution of them. Lower amounts of inclusions have a direct influence on the improvement of mechanical properties such as ductility, formability, fatigue resistance, as well as the corrosion resistance in low-‐alloy and high-‐alloy steels. Moreover, inclusions can have a detrimental effect on the machining performance and also the casting operations. In order to reduce the amount and size of the inclusion, efforts have been spent to improve the secondary steelmaking processes rather than the casting operations. But since it would be difficult to remove inclusions during casting, it is better to prevent the formation of inclusions in the first hand and avoid the inclusions’ carry over to the final product.
The non-‐metallic inclusions are typically categorized into oxides, sulfides, and nitrides.
Among them, oxide inclusions such as alumina (Al2O3) and silica (SiO2) are the most important ones in production of low-‐alloy clean steel. Zhang et al. [1] performed a comprehensive study on inclusions existing in industrial bottom-‐teemed ingots of carbon steel. It was showed that 59% of inclusions larger than 20μm were pure alumina or alumina/FeO.
These inclusions can be formed in steel during ladle refinement, deoxidation operations when aluminum or silicon deoxidants are added, and reoxidation during teeming and mold filling. If it is the case, they are called indigenous inclusions. In addition, there is another type, which is called exogenous inclusions. These come from incidental chemical and mechanical interaction of liquid steel with its surroundings, such as entrapment of ladle slag in the molten steel or lining erosion [2]. There are also other possibilities to generate exogenous inclusions, which happen during ingot casting operations, such as inclusions eroded from ladle nozzles and runners, and exogenous inclusions from casting powders.
Many studies were conducted to recognize the sources of these different inclusions and identify the major types. [3] [4] [5]
In the work of Tripathi et al. [6] inclusions in the steel were studied at different stages of steel making. The results showed that alumina based inclusions (namely Al2O3 and Al2O3-‐SiO2 inclusions) are formed during mold filling and that they are present after a casting operation.
Examining the chemical composition of the casting powder and considering the position of Al2O3-‐SiO2 inclusions, which is close to the mold wall, supports the idea that these inclusions are entrapped casting powder. In case of Al2O3 inclusions, a part are introduced into steel when the melt passes through the pouring gate and nozzles with a high volumetric flow rate
and flushes off small pieces of refractory. Another part was generated during casting, which is an evident for reoxidation during casting.
A study by Doostmohamadi et al. [7] was performed on identifying the inclusions left in runners after ingot casting. The study showed that oxide inclusions found in the samples contained the following elements: Al, Ca, Mg, Si and O. Complex inclusions with oxide composition such as oxide-‐sulfide were also found in the runner and Al2O3-‐MgO-‐MnS inclusions were the most frequent among them. Additionally, the study indicated that the composition of almost all of the inclusions with sizes larger than 10μm was Al2O3-‐SiO2-‐MgO and SiO2-‐MgO. Figure 1 shows the size classification of inclusions based on the Swedish Standard SS111116. The study also presented a comparison between the type and size of the inclusions present in samples from different states of steel making, namely from ladle, in runner, in the mold. Considering that comparison, it can be interpreted that the majority of small oxide inclusions were generated during casting and that can be associated to reoxidation.
Figure 1. Size groups of inclusions in runner (based on the work of Doostmohammadi et al.) [7]
A review study on inclusions existing in steel ingot casting has referred to a survey of 25 heats of ingot casting which has found that 77% of the inclusions which form during casting (exogenous inclusions including reoxidation inclusion) are reoxidation inclusions. [8]
Therefore, reoxidation during teeming can account for a major source of inclusions in ingot casting.
The contact of the liquid steel with the oxygen from the surrounding air is the main cause of reoxidation. The surface of the stream of the pouring melt and the top surface of
3≤A≤7μm 7≤B≤11μm 11≤C≤22μm D≥22μm
should be noted that a very small part of reoxidation reactions could be caused by the exchange reaction at melt/slag or melt/refractory interfaces. [2]
Figure 2. Percentage of affected heats by different kind of exogenous inclusions in the work of Thomas et al. [8]
Another disadvantage of inclusion formation, which is noteworthy, is waste of alloying elements. It has been shown that the consumption of oxidizing elements is linearly proportional to the partial pressure of oxygen around the steam during casting [10]. A loss of about 0.73 kg Mn, 0.41 kg Si, 0.41 kg Al, and 0.14 kg C per ton of steel has been reported for unprotected tapping and ladle-‐filling operations in wrought steel production. [10]
In addition to reoxidation, the exposure of liquid steel to air can lead to other problems, namely nitrogen pickup. N2 gas evolution may occur during solidification of molten steel with high nitrogen contents and leads to formation of bubbles or pinholes. Also, the physical properties of steel are greatly affected by the precipitation of nitrides. (The later phenomena can sometimes be beneficial. For example in the case of TiN precipitants that helps in grain refinement.) It should be noted that nitrogen absorption is inversely dependent on surface-‐
active solutes such as oxygen and sulfur. Therefore, the extent of nitrogen pickup increases at final stages of steelmaking, namely casting, at which the concentration of dissolved oxygen is low. [11]
The idea of using a protective shroud around the stream has been utilized from many years ago in industry. This could be done by a submerged nozzle in continuous casting or by an inert gas in case of ingot casting.
77%
46%
38%
8%
23%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Percentage of affected heats
****
The teeming operation in the ingot casting is of great importance for the quality of the final steel product. It might be underestimated because the diameter of the stream is not very large. However, actually this diameter is enough to provide a large free surface for the melt in the duration of casting. Hence, a considerable amount of oxygen can be absorbed by the melt, which leads to reoxidation. On the other hand, in bottom-‐pour teeming, a massive amount of air entrainment takes place due to high casting velocities (more than 4 m/s) [12].
During contact of the melt with the atmosphere, oxide films form on the surface of stream and are then folded into liquid and left behind as weak planes in the solidified ingot. The oxidization is more critical for killed steels since they contain elements with high oxygen affinity such as Aluminum, Calcium or Silica. It is therefore very important to protect the melt stream from the surrounding air during casting to minimize air entrainment in molten steel. [2]
The best way to prevent reoxidation is to restrict the contact of liquid steel and atmosphere. Industrial investigations have shown that using a shroud during transferring operations at different stages of steelmaking, i.e. tapping or teeming, could protect the liquid steel against air and prevent its contamination. [3] [13]
The most conventional way of protection is to cover the stream with an argon gas curtain, which is injected from around the teeming nozzle. Another way is to use a physical shroud between the ladle and the casting trumpet, which is filled by inert gas. It has been reported that a good sealing can reduce nitrogen content by 30ppm and lower the number of large Al2O3 inclusions. The protection is more effective for bottom-‐pouring ingot casting.
[8] A couple of examples for the physical shrouding method are shown in Figure 3. A two-‐
third reduction in oxides in steel was reported in casting operations using shroud with inert gas environment [12]. Another alternative, which is costly and more complicated and used only for special purposes, is casting under vacuum conditions.
The focus in this thesis is finding effective parameter in argon shrouding. The study consists of a theoretical part, a computerized modeling and plant trials based on the simulations results. Plant trials were carried out at Scana Steel Björneborg.
Figure 3. Different devices for protection of melt stream against air adsorption [8]
1-‐2. Plant Description
Scana Steel Björneborg is a scrap-‐based steel plant that mainly produces low-‐alloy steels.
Recycled steel is melted in an electric arc furnace (EAF) of a 55-‐ton capacity. After melting, steel is phosphor refined and heated to a desired temperature. The steel is deoxidized upon tapping into the ladle and thereafter transferred to the ladle furnace station. There, various alloying elements are added in regard to the chemical specification of the desired steel grade. During the ladle treatment, the melt is heated and stirred with argon bottom injection in order to homogenize the temperature and composition. In the next stage, the ladle is transferred to the vacuum tank with the purpose of sulfur removal and hydrogen degassing and inclusion separation with assistance of argon gas bottom injection.
Thereafter, the ladle is transported to the casting section and the steel melt is poured from the bottom of the ladle into ingot using uphill casting method. The surface of the melt in the molds is covered by casting powder for thermal insulation and to prevent reoxidation. Also, for some specific steel grades, the steel stream between ladle and the vertical runner is protected by argon gas flow during teeming. After stripping the ingots, they are transferred
to the forging shop to form them into required shape. Then, they are heat-‐treated and sent to machining shop to prepare the final product.
Figure 4. Schematic picture of the steelmaking operations at Scan Steel Björneborg
1-‐3. Estimation of oxygen absorption
A couple of studies have been carried out to explain the oxygen pickup behavior of steel.
It is shown that the oxidation rate during teeming is primarily controlled by the transfer of oxygen from the atmosphere to the surface of the steel. [3] [10] [14]
Choh et al. [14] studied oxygen and nitrogen absorption in the melt stream during teeming. The absorption occurred by dissolution of the gas, which is introduced to the melt by two mechanisms, namely, by gas entrainment and through the surface of stream. The volume of the entrained gas increases with increasing the length and the velocity of the stream and with decreasing the nozzle diameter. It was also showed that at a short teeming height (less than 50cm), gas entrainment is independent of the physical properties of molten steel such as viscosity, density and surface tension. With an increasing casting rate, less oxygen is absorbed in the melt during teeming. A mathematical model has been presented based in another work [15] with the intention to calculate the amount of dissolved gas during teeming in top casting, as well as teeming from ladle to tundish.
The same approach has been taken in the present study to obtain equations to estimate the absorbed oxygen into the liquid steel through the surface of the stream. But since this study is interested in uphill ingot casting and there is no pool of liquid steel under melt stream and thus, no plunging region is formed, the effect of gas entrainment is neglected and only absorption through surface of stream is considered.
It was indicated in a work of Sasai et al. that there is a good conformity in the experimental results and the calculated values by a mass transfer-‐controlled model for oxidation of liquid steel during teeming. [16]
When a liquid is poured from the bottom of a container i.e. free-‐fall stream, the diameter of the stream becomes smaller. This decrease continues as the liquid travels down.
The factor ξ accounts for the change of melt stream diameter, depending on the distance from nozzle. [15]
Eq. 1: 8
1
2 0
) 2 1
( +
= U ξ gZ
Where g is the constant of the gravity, Z is the teeming height and U0 is the teeming flow rate at the nozzle. Therefore, the radius of the stream (a) at the teeming height can be calculated by Eq. 2.
Eq. 2: a=a0ξ−2
The area of the surface of the stream (As) which represents the interface between gas and liquid phase, is therefore expressed by Eq. 4:
Eq. 3: As =2π
∫
0ZadZEq. 4: ⎥
⎦
⎤
⎢⎣
⎡ + −
= 2 1) 1
3 (
4 43
2 0 2 0 0
U gz g
U
As πa
Because of a continuing renewal of the liquid phase, the rate of oxygen absorption is assumed to be controlled by the rate of mass transfer in the surrounding gas phase, which is calculated by Eq. 5 [15]. Because of immediate consumption of dissolved oxygen due to activity of aluminum in steel, the interfacial pressure of oxygen (!!!!) was taken to be zero.
But PO2 depends on surrounding atmosphere, which in the case of air as surrounding gas it is 0.21 and in the case of argon protection, it is assumed to be 0.07 based on the results showed in work of Zinchenko et al. [17]. Specially, they state that the nitrogen content and the oxygen content of the atmosphere around the stream could be reduced nearly by a threefold reduction when argon-‐blowing is used during the teeming.
Eq. 5:
( )
RT P k P
n
i O O
g 2 2
= −
Eq. 6: 3
1 2 1
Re 664 . 0
2 2
D Sc k L
O N
g =
−
Where R is the gas constant, T is the temperature of the liquid steel and kg is the mass transfer coefficient in the gas atmosphere. Also, in Eq. 6, Re is Reynolds number and Sc is Schmidt number for the surrounding gas.
It should be noted that Eq. 6 is valid for a one-‐dimensional flux and a fluid on a flat plate.
However, it is used in the present study as an approximation. Also, the tapping height is taken as the characteristic length (L).
Fuller, Schettler and Giddings [18] have presented a simple and very useful correlation for estimation of the interdiffusion coefficient. A modification to that correlation was then recommended by Reid et al. and by Danner and Daubert (Eq. 7) [19]. This equation was used for two conditions of the surrounding gas, namely air and argon.
Eq. 7:
gas O
gas O
gas
O M M
V V
P
D T 1 1
) ( ) (
10 0 . 1
2 2
2 2
3 1 3
1
75 . 1 3
+
⎟⎟
⎠
⎞
⎜⎜
⎝
⎛ ∑ + ∑
×
= ×
−
−
Table 1. Diffusion Volumes of Simple Molecules [19]
Gas ΣVi
N2 18.5
O2 16.3
Air 19.7
Ar 16.2
Equations 8-‐11 and Table 2 were used in order to calculate the properties of the surrounding gas at the desired temperature.
Eq. 8:
gas gas gas
LU µ
0ρ
Re =
Eq. 9:
gas O gas
gas
Sc D
−
= ρ 2
µ
Eq. 10:
gas gas
gas RT
= PM
ρ
Eq. 11: 2 0 Sutherland'sCorrelation
3
0 ,
0 T S
S T T
T
gas gas
gas
gas +
+
⎟⎟⎠
⎞
⎜⎜⎝
=µ ⎛
µ
Table 2. Constants for Sutherland’s Correlation [20]
Gas μ0 (g/cm.s) T0 (K) S (K)
Air 1.176×10-‐4 273.11 110.56
Ar 2.125×10-‐4 273.11 144
N2 1.664×10-‐4 273.11 107
Assuming 1000˚C for the average temperature around stream and by using Eq. 6, equation below was obtained for prediction of mass transfer coefficient (kg).
Eq. 12: 2
1 0 3 1 2 1
) ( 664
. 0
2 1 2
Z Sc U k D
gas gas gass O g
µ
− ρ
×
=
Finally, the absorbed oxygen in melt is calculated by:
Eq. 13:
[ ]
Steel O s
m tM n
O A 6
100 10Δ 2
=
Δ
It should be noted that Eq. 13 refers to a uniform oxygen concentration in steel. But, during teeming, the absorbed oxygen does not distribute evenly to the center of stream and it is sounder to consider smaller amount of liquid steel instead of mSteel. In order to apply this consideration, the oxidation is assumed to be limited to the outer section of the stream (dox=a0/3). Then, the corrected mass of liquid steel is calculated regarding to the cross section of the assumed oxidation layer.
Eq. 14: a U t
mcorrected = − × 0Δ
2 6 0
9
10 5π ρ
Figure 5. Schematic illustration of melt stream and the effective oxidation layer
Therefore, by using Eq. 14 and replacing the calculated mass of liquid steel in Eq. 13, the amount of absorbed oxygen through the surface of the stream can be estimated by:
Eq. 15:
[ ]
ρ π 0
2
5 0
100 9 2
U a
M n
O As O
×
=
Δ
The density of the melt can be estimated using the equation proposed in work of I.
Jimbo and A. Cramb. [21]
Eq. 16: ρ =(7.10−0.0732[%C])−(8.28−0.874[%C])×10−4(Tmelt −1823)
Table 3 summarizes the parameters that have been used in above-‐mentioned calculations.
dox
a0
Table 3. Parameters existing in oxygen absorption calculations
Parameter Description Unit
a0, a radii of stream at the nozzle exit and at the teeming height Z cm DO2-‐gas Interdiffusion coefficient of oxygen and gaseous atmosphere cm2/s
As area of the interface between gas and the stream cm2 dox thickness of effective layer of stream prone to oxidation cm
g gravity constant (=981) cm/s2
kg mass transfer coefficient in gas atmosphere cm/s
mcorrected Corrected weight of casted steel regarding to dox ton
msteel weight of casted steel ton
MO2, Mgas molecular weight of oxygen and the gas around the stream g/mole
! flux of oxygen in the surrounding gas mole/cm2s
∆ ! amount of absorbed oxygen through surface of the stream %
P atmospheric pressure atm
!!!, !!!! oxygen partial pressure and it interfacial value atm
R gas constant (=82.05746) cm3.atm/mol.K
Re Reynolds number for surrounding gas -‐
Sc Schmidt number for surrounding gas -‐
te exposure time of stream to gas s
Tmelt, Tgas temperature of the liquid steel and the surrounding gas K
U0 teeming flow rate at nozzle exit cm/s
ΣVO2, ΣVgas special atomic diffusion volumes cm3/mole
Z teeming height cm
δ film thickness cm
ξ coefficient of growth of stream radius -‐
ρ, ρgas density of liquid steel and surrounding gas g/cm3
2. METHODOLOGY
There are two distinct sections for the methodology for this study. The first is a computational modeling of the argon protection during the ingot casting in order to examine the influential parameters. The second section is industrial experiments and sampling in order to compare with the modeling results and to study the effectiveness of the new shield design.
2-‐1. Simulations
In order to fulfill the lean production criteria, the manufacturing process must be done correctly the first time, to eliminate the waste. Back in the time, product development was mainly done by trial and error, which was very time and money consuming. On the other hand, the experimental of the gas flow in the protective shield during casting, if possible, is very difficult and dangerous. Today, with the help of computer simulation, it is possible not only to reduce the steps and cut the costs (both materials and labor), but also to design better and optimized products. Therefore, because the simulation process is fast, inexpensive and safe, it encourages creativity and innovation.
The simulations in this project consist of three different physics in two steps. First, the flow pattern of argon gas in the shield was calculated by using the fluid flow module (CFD) and the heat-‐transfer module. Then in the second step mass transfer of oxygen, nitrogen and argon was simulated considering the forced convection by the gas flow and diffusion.
In the pre-‐study simulations, the fluid flow pattern in the shield chamber and the velocity magnitude at the cross sections and stream vectors of the gas were examined. Using the simulation results, the oxygen concentration of the gas atmosphere in the shield was studied for verification and comparisons, as well as to establish a baseline for modifications and suggesting a new design for the shield. The simulations were executed with software COMSOL Multiphysics®.
2-‐1-‐1. Assumptions
A 3D geometrical-‐model of argon protection of the liquid steel stream between the ladle and the casting trumpet was created regarding to the dimensions existing in Scana Björneborg steel plant. Then a series of preliminary simulation was carried out in order to decide about the assumptions and the mesh dimensions.
The following assumptions were used in the calculations in the present study:
• The numerical calculations for prediction of the flow were obtained by solving the equations of continuity and momentum conservations and Navier-‐Stokes equations.
• The K-‐ε model was used for simulating the turbulent flow.
• The gases (argon and air) considered as compressible Newtonian fluid with Mach number less than 0.3.
• It was assumed that no chemical reactions take place.
• The stream of the liquid steel considered as a sliding wall condition with a constant velocity corresponding to the casting velocity.
• The standard temperature for argon gas was taken as 25˚C, which is increased along the inlet’s length to reach to the temperature of the environment around the melt stream in the shield.
• The temperature of the melt stream was taken as 1400˚C.
• Based on real situation measurements, it was assumed that the outer side of the shield is 300˚C.
• Heat exchange occurred via radiation from melt stream, natural convection in the chamber and natural convection to the outer atmosphere.
• For heat capacity values, argon was considered as a monoatomic ideal gas. (γ=5/3)
• A “mixture-‐averaged” diffusion model was chosen for mass transfer simulations. The diffusion coefficient in this model is computed by using multicomponent Maxwell-‐Stefan diffusivities (Eq. 7).
• The outer atmosphere of the shield is considered to be air with pressure of 1 atm and following constitutes mole fraction: 0.93% argon, 78.12% nitrogen and 20.95% oxygen.
2-‐2. Parameters
One of the purposes of the present study is to investigate the effect of different parameters on the flow of argon gas around the pouring steel and the degree of protection against oxygen adsorption.
One way to protect the stream is to blow an inert gas to the stream in order to purge the active atmosphere away from the liquid steel. But the drawback of shrouds without a physical enclosure is that they require large amounts of the inert gas to ensure the protection [22]. Therefore, to overcome that problem, a physical shield is used around the stream to provide a semi-‐closed environment for the inert gas.
In the present work, a cylindrical casing has been used around the melt stream and argon gas was injected into it. That leads to a slight positive pressure which confines uptake of air through gaps and joints. So the air entrainment could be efficiently reduced.
Moreover, because the casing enfolds around the nozzle, it helps in preventing reoxidation in nozzle and protecting it from clogging.
The factors, which were considered in the study of the effectiveness of protection, are introduced in this section.
2-‐2-‐1. Distance between the ladle and the shield
Although a layer of 25-‐millimeter insulating pad had been used between the shield and the ladle, there was still a leakage gap of around 3 cm at the upper part of the shield. The reason of that was poor contact between the shield and the protective plate of the frame of the sliding nozzle. This gap can be seen in Figure 6.
The effect of omitting the gap between the shield and the ladle was studied in the first series of simulations.
Figure 6. Existing shield for protected ingot casting
2-‐2-‐2. Distance between the shield and the trumpet
Although there is a short distance between the ladle nozzle and top of the casting trumpet, this space plays an important role in the amount of oxidation. The gap between the shield and the trumpet depends on the height at which the ladle is held during casting.
L. Ragnarsson et al. [5] simulated the flow of argon gas injected near the ladle furnace nozzle, using a physical model coupled with Particle Image Velocimetry (PIV) technique. It was found that the most influential factor on the amount of available oxygen around melt stream is the distance between the ladle and the trumpet.
This can be regarded to the drag of air due to high velocity of the stream and also the pumping action due to the conical trumpet. These two phenomena suck air into the runner.
In another study [15] has been showed that the volume of the entrained gas in liquid steel increases with increasing the length of the stream.
2-‐2-‐3. Argon gas flow rate
Argon can be considered the best protective gas because it is completely inert and does not react with or form compounds with other elements. Argon functions as a curtain to
hinder the transfer of air to the stream of melt. Moreover, in case of physical shroud, air is simply replaced by argon inside the shield space.
It was shown in the work of Ragnarsson [5] that the flow patterns (both the shape and the velocity of the flow) at different gas flow rates are similar. Also, based on the oxygen content measurements of the gas around the stream, it was concluded that the flow rate has little effect on distribution of oxygen content in the shroud. However, no physical shield was used in that study and shrouding had been done only by argon blowing.
Hence, if the gas inlet is located near the openings as the case in work of Ragnarsson et al. [23], higher flow rate would suck in air into the region surrounding the stream and cause more air flushing and oxygen entrainment.
2-‐2-‐4. Argon inlets
The flow pattern in the shield is also affected by design characteristic such as number and position of the argon inlets. Shields with 1, 2 (with two alternatives for their vertical positions) and 3 inlets were investigated in the simulation studies.
2-‐2-‐5. Shield shape
In the final stage of the simulation studies, the flow pattern of argon using a new shape for the shield was examined. This alternative shape was recommended based on the results of the previous simulations as well as existing conditions in the Scana steel plant.
2-‐3. Sampling
With the intention to investigate the air uptake during teeming, the samplings were done at two stages. The first series was before casting (SP) in which samples were taken at the end of vacuum treatment. The other was after casting (GP) where the samples were taken from solidified vertical runner. It this way, the isolated nitrogen and oxygen uptake in the teeming stream could be measured.
2-‐3-‐1. SP Samples
To obtain SP samples, a specific type of lollipop sampler was used to take samples from liquid steel from ladle at the vacuum station before transferring it to the casting station. The sampler is called Björneborg lollipop sampler and is shown in Figure 7. The samplings were performed manually and its position was aimed to be more or less identical. To prevent entering the slag into the sampler, argon is blown into sampler during immersion through the melt. After samples were quenched in air and water, three pieces with specific dimensions were cut out from the pin part of the sample.
Figure 7. Björneborg sampler; green sections show the selected parts for analyzing of the oxygen and nitrogen contents
2-‐3-‐2. GP Samples
After casting and when the steel was solidified, the steel rod from the vertical runner in the casting bottom-‐plate was sawed and machined in order to prepare the GP samples. Figure 8 demonstrates the schematic picture of GP samples.
Figure 8. a) position of the GP samples in the horizontal runner of the casting system, b) the procedure of
preparing GP samples , c) the selected parts for oxygen and nitrogen analyses from GP samples
2-‐4. Measurements
The total oxygen content and the nitrogen content in the prepared samples were measured by a LECO elemental analyzer, model TC-‐600. It utilizes an inert gas fusion method and thermal conductivity detection to measure the content of desired element (oxygen or nitrogen).
2-‐4-‐1. Total oxygen content
The total oxygen (T.O.) in the steel is the sum of the dissolved oxygen and the combined oxygen in the oxide inclusions. Despite the fact that the size of the samples is small (1 gram) and it is rare to capture inclusions due to their small population, there is a direct correlation between probability of large oxide inclusions and total oxygen content. Therefore, the T.O.
can serve as an index of steel cleanliness. [8]
Figure 9. Relationship between T.O. and macroinclusions in steel [8]
2-‐4-‐2. Nitrogen content
Sasai et al. [16] showed that there is direct correlation between the amount of oxidation of molten steel by air and the change in nitrogen content of steel, during teeming of steel from ladle to tundish.
Nitrogen pickup of steel inversely depends on amount of oxygen and sulfur in steel because of their effect on surface tension of liquid steel. In the steel with low oxygen content, nitrogen absorption happens faster and the difference in nitrogen content before and after casting of low oxygen steel can be an indicator for amount of air adsorption.
3. RESULTS AND DISCUSSIONS
3-‐1. Simulations and investigations of effective parameters 3-‐1-‐1. Introductory simulations
The initial digital model based on the conventional shield used in shrouded casting at Scana Steel Björneborg is shown in Figure 10.
Figure 10. Schematic of the 3D model of the existing shield
Two parts of the model, namely the distance between the ladle and the shield at the upper part and the distance between the ladle and the trumpet at the lower part, were set as outlet with pressure constraint condition.
Figure 11 demonstrates the temperature distribution of gases in the shield space around the melt stream obtained from steady state simulations.
Figure 11. Temperature distribution calculated by steady state simulations considering the assumptions in section 2-‐1-‐1
It is seen that temperature of the stream is less than the assumed initial temperature of the melt in section 2.1.1 due to heat transfer to neighboring atmosphere. (Although it is below the solidification temperature of the steel alloy, it is assumed the stream is in liquid phase. This assumption does not affect the gas flow simulations results.) The highest temperature in the model occurs in the space in the sprue that is made of refractory material.
The flow pattern of gas flow in the shield chamber during the protected casting with conventional conditions at Scana Steelmaking is demonstrated by streamline in Figure 12.
The color and the thickness of the streamlines are proportional to the gas flow velocity.
Figure 12. Streamlines of gas flow in the existing shield
The flow pattern shows that there is relatively high gas flow in the upper part of the shield near the gap at the connection of the shield to the ladle’s sliding-‐gate. The reason of that can be explained by overall flow pattern, which is caused by a downward stream of the melt as well as by the swirling effect the argon gas flow.
In the following simulation results, the molar concentration of the oxygen gas around the melt stream is taken as the study subject in the simulations of the various conditions for the shield-‐protected casting. The unit of these measurements is mole/m3.
Table 4 can be used as a guide for comparing the results with available oxygen in the air at different temperatures.
Table 4. Oxygen content of air at different temperatures Temperature
(˚C) Air molar concentration
(mol/m3) Oxygen molar concentration
(mol/m3)
25 40.87 8.56
300 21.27 4.46
1000 9.57 2.01
Figure 13 presents the oxygen content results for Scana’s conventional shield with a 100 l/min flow rate of argon and a 3cm gap between the shield and the ladle. The existing conditions of the shield installation are shown in Figure 6.
The simulation result of the identical shield with better sealing at the ladle connection, namely with 1cm gap, is depicted in Figure 14.
Figure 13. Oxygen concentrations for existing shield with 100 l/min argon flow and 3cm upper gap
Figure 14. Oxygen concentrations for existing shield with 100 l/min argon flow and 1cm upper gap
