Research on process parameters of homogenised continuous casting for GCr15 bearing steel bloom

Abstract

A solidification and heat transfer model was established to analyse quantitatively the influence of process parameters on solidification behaviour during bloom continuous casting (CC), and then the appropriate process parameters for homogenised CC were determined through systematically considering soft/heavy reduction (SR/HR), final electromagnetic stirring (F-EMS), shell thickness and temperature re-rising of secondary cooling zone (SCZ). Results indicate that solute element is prone to enrich towards the bloom's centre when liquid core is a slender conical shape. And casting speed needs to be controlled within 0.450 m/min < V < 0.478 m/min for making seven HR rollers be in liquid core reduction and the corresponding solidification rate of F-EMS is F_S = 0.65∼0.69, namely, both key processes are in suitable operating ranges. Moreover, the shell thickness of mould outlet is about 13.51∼17.81 mm and the maximum re-rising temperature in the SCZ is 121°C, and consequently neither of them will affect the strand quality. Eventually, the appropriate control strategies of process parameters for homogenised CC are 0.46 m/min or 0.47 m/min in casting speed, superheat of 20∼30°C, 0.125 L/kg in water ratio and running one reduction rollers, and so corresponding solid fractions f_s = 0.43∼1.0 (0.46 m/min) and f_s = 0.41∼1.0 (0.47 m/min) in the SR/HR region are more reasonable. Finally, centre segregation was reduced from 1.18 to 1.09 in industrial production.

Keywords

GCr15 bloom homogenised continuous casting solidification and heat transfer process parameters reduction

Introduction

Bearing steel with high quality and large cross-section has become one of the key materials that supports the rapid development of high-end manufacturing, as a result, higher technical requirements have been put forward for the bloom continuous casting (CC), which is the main production process of bearing steel. At present, the principal problem faced by bearing steel bloom is the presence of homogenised defects such as centre segregation, porosity and shrinkage cavity.^1,2 In order to solve them and enhance the strand's quality, low superheat,³ the control of casting speed and SCZ's system,^4,5 mould electromagnetic stirring(M-EMS),⁶ final EMS (F-EMS),⁷ soft/heavy reduction (SR/HR)^8,9 and other processes have been proposed and applied.

Aiming to various process technologies, extensive research had been carried out in the past few decades. Choudhary et al.¹⁰ found that reducing superheat of molten steel can increase the equiaxed grains rate and further improve centre segregation of the strand through statistical analysis of lots of industrial data. Hou et al.¹¹ also discovered a similar changing rule when studying the solidification structure of the strand by numerical simulation. Chen et al.⁵ analysed the influence of casting speed and cooling water distribution of SCZ on the strand's centre segregation via a three-dimensional multi-field coupling mathematical model. Results showed that high casting speed is not conducive to improving segregation, yet the centre segregation could be lowered to a certain extent by optimising the SCZ's system. M-EMS can change the flowing pattern of molten steel and accelerate the dissipation of superheat, thereby making equiaxed grains rate increase and central segregation mitigate in the strand.^12,13 An et al.¹⁴ and Yang et al.¹⁵ thought increasing the current and frequency of M-EMS could contribute equiaxed grains to grow up and centre segregation of the strand to improve based on production experiments. Zhang et al.¹⁶ and Wang et al.¹⁷ simulated and analysed the effect of M-EMS on the strand's segregation. Results indicated that strong M-EMS would aggravate the subcutaneous negative segregation of the strand. Owing to F-EMS can break the ‘bridging’ dendrites and eliminate the uneven composition distribution in mushy zone, it has developed into one of main processes for homogenised CC.¹⁸ In this regard, Sun et al.¹⁹ and Jiang et al.²⁰ investigated the influence of different stirring modes of F-EMS on the internal quality of the strand based on numerical simulation. The former believed that alternating stirring mode was more conducive to reducing central segregation, while the latter thought that central carbon segregation was the lowest under unidirectional stirring mode. Luo et al.²¹ and Chao et al.¹⁸ optimised stirring parameters of F-EMS via industrial experiments to depress central segregation. As another important process to realise the homogenised CC, SR/HR also plays an irreplaceable role.²² By conducting several group tests, Ogibayashi et al.²³ found that although the implementation of SR could significantly improve centre segregation of the strand, the central solid fraction f_s must be in an appropriate value. When f_s < 0.25, adding reduction amount would worsen the centre segregation. For the HR, Ji et al.²⁴ combined simulation and experiments to study the impact of HR on the internal quality of GCr15. And it was found that the strand's centre consistency was obviously upgraded when 28 mm in reduction amount. The research results of Zhang et al.²⁵ showed that compared with SR, HR was more conductive to improve the central defects. Zhao et al.'s²⁶ research indicated that raising reduction amount could significantly alleviate the strand's centre segregation under adopting the convex reduction rollers. Zhao et al.^27,28 found that solid fraction f_s = 0.8∼0.86 was the key for solve internal defects of the strand via HR. If it exceeded the critical value, the HR's effect would be greatly weakened.

Current investigations and production practices have displayed that F-EMS and SR/HR processes are optimal to improve centre segregation, porosity and shrinkage cavity of the strand, and both control cores are the stirring timing and reduction interval.^18,22 In actual production, due to that relative devices of F-EMS and SR/HR are basically stationary, the most common method is to adjust superheat, SCZ's water ratio and casting speed to change the strand's solidification process and further regulate the position of F-EMS and SR/HR indirectly. However, based on the analysis of the literature,^3–28 it is found that the existing research mostly focuses on the design and optimisation of M-EMS, F-EMS and SR/HR parameter for improving the strand quality, and less involves process parameters of CC such as casting speed, superheat and water ratio of SCZ. Even if there is a small amount of relevant research, there is a lack of systematic investigation on process parameters of homogenised CC, especially for special steel of bloom. In view of this, through established the solidification and heat transfer model of 390 mm × 480 mm bloom CC, this article analysed the influence of superheat, SCZ's water ratio and casting speed on the solidification behaviour of the strand, and determined the suitable process parameters for homogenised CC via systematically considering SR/HR, F-EMS, shell thickness at mould outlet and temperature re-rising in the SCZ.

Model description

Assumptions

In order to simplify mathematical model, appropriate assumptions are proposed without affecting the reliability of modelling. The main contents are as follows:

Only considering heat transfer in the horizontal direction of the strand. Because that the heat transfer amount along the casting direction accounts for about 3%∼6% of the total, which can be ignored.

The strengthening effect of molten steel's convection on heat transfer in the liquid and mushy zone is addressed by artificially magnifying the thermal conductivity coefficient.

The solidification latent heat is considered by adopting the equivalent specific heat method.

Neglecting the influence of mould oscillation, electromagnetic stirring, reduction, and the arc-shaped structure of CC machine on heat transfer.

Geometric model and meshing

Figure 1(a) is the schematic diagram of 390 mm × 480 mm bloom CC machine with an arc radius of 16.5 m, equipped with F-EMS and SR/HR that contribute to achieve homogenised CC. The two-dimensional(2D) slice method is used to model the CC of GCr15 bearing steel from meniscus to shear position. This process is 32 m in length, and its detailed parameters are shown in Table 1. Figure 1(b) shows the 2D slice geometric model established and meshing. Considering the symmetry, only one-fourth of the strand cross-section is selected. Moreover, the minimum size of 5 mm is used to divide grid, and the total number of grids is 1872. Finally, this model is run by using the commercial software of Abaqus.

Table 1.

Parameters required for simulating during CC process of bloom.

Parameters	Values	Parameters	Values
Cross-section, mm × mm	390 × 480	Zone 4, m	5.2
Mould length (effective length), mm	800 (730)	Air cooling, m	16.675
Zone 1, m	0.345	Casting speed, m/min	0.30∼0.50
Zone 2, m	4.25	Superheat, °C	20∼60
Zone 3, m	4.8	Water ratio, L/kg	0.07∼0.16

Governing equations

The 2D slice method is used to simulate the solidification and heat transfer process of bloom CC, and the governing equation is shown in equation (1).

ρ c \frac{\partial T}{\partial t} = \frac{\partial}{\partial x} (k_{e f f} \frac{\partial T}{\partial x}) + \frac{\partial}{\partial y} (k_{e f f} \frac{\partial T}{\partial y}) + S

(1)

where T is temperature, °C; ρ is density, kg/m³; c is specific heat, J/(kg·°C); k_eff is thermal conductivity coefficient, W/(m·°C); and S is inner heat source, W/m³.

The latent heat released from mushy zone is the inner heat source, which is expressed in equation (2).

S = ρ L \frac{\partial f_{s}}{\partial t}

(2)

where L is solidification latent heat, J/kg. And then substituting equation (2) in (1) to obtain equation (3).

ρ c_{e f f} \frac{\partial T}{\partial t} = \frac{\partial}{\partial x} (k_{e f f} \frac{\partial T}{\partial x}) + \frac{\partial}{\partial y} (k_{e f f} \frac{\partial T}{\partial y})

(3)

where c_eff is equivalent specific heat, J/(kg·°C). Specifically, as shown in equation (4).

c_{e f f} = c - L \frac{\partial f_{s}}{\partial T}

(4)

where f_s is solid fraction.

Boundary conditions

During numerical simulation, the slice moves with starting from the meniscus, passing through the mould, SCZ and air cooling region successively. The heat transfer's boundary conditions applied to the mathematical model in each stage are as follows.

Initial boundary

At the initial moment, boundary condition sets as casting temperature T_c of molten steel, namely, T_c = T_L+ΔT. Where T_L is liquidus temperature, °C and ΔT is superheat, °C.

Mould zone

The average heat flux is used as boundary condition, which can be got by the following equation (5).

{\bar{q}}_{m} = \frac{ρ_{w} c_{w} Q_{w} Δ T_{w}}{F}

(5)

where ${\bar{q}}_{m}$ is average heat flux of the strand, W/m²; $ρ_{w}$ is density of cooling water, and it is 1000 kg/m³; $c_{w}$ is specific heat, and it is 4181 J/(kg·°C); $Q_{w}$ is flow rate, and wide and narrow faces of mould are 0.031 m³/s and 0.025 m³/s, respectively; $Δ T_{w}$ is temperature difference of cooling water at the inlet and outlet of water gap, and the measured value is 5.5°C; and F is contact area between mould's copper plate and cooling water, m².

It is noted that there exist air gap in the mould corner, which will significantly affect the circumferential solidification of the strand and cannot be ignored. Based on this, this article will address as follows. When the distance from the meniscus is 0∼0.1 m, heat flux of corner $q_{c o r n e r} = {\bar{q}}_{m}$ ; and the distance is 0.1∼0.23 m, $q_{c o r n e r} = 0.8 {\bar{q}}_{m}$ . And 0.23∼0.73 m in distance, $q_{c o r n e r} = 0.7 {\bar{q}}_{m}$ .

Secondary cooling zone

The SCZ's heat transfer is relatively complex, mainly including spraying water convective, evaporation, thermal conductivity of pinch roller and radiation.⁴ For solving it, the comprehensive heat transfer coefficient (HTC) is applied to each zone of SCZ in this article,¹⁷ as shown in equation (6).

h = 5849.9 \times W^{0.451} \times (1 - 0.0075 T_{e})

(6)

where h is comprehensive HTC, W/(m²·°C); W is flux rate of cooling water, L/(cm²·min); T_e is ambient temperature, and °C is the range of HTC's values in each zone and are shown in Table 2.

Table 2.

HTC's values in each zone of bloom's SCZ.

SCZ	HTC, W/(m²·°C)	SCZ	HTC, W/(m²·°C)
Zone 1	245∼346	Narrow face of zone 3	51∼73
Wide face of Zone 2	65∼128	Wide face of zone 4	56
Narrow face of Zone 2	72∼111	Narrow face of zone 4	49
Wide face of Zone 3	58∼84

SCZ: secondary cooling zone; HTC: heat transfer coefficient.

Air cooling zone

The strand mainly exchanges heat with surrounding environment through radiation heat transfer at this stage, and the corresponding formula is as follows.

q_{r a d} = σ \cdot ε \times [(T_{s} + 273)^{4} - (T_{e} + 273)^{4}]

(7)

where q_rad is heat flux of radiation heat transfer, W/m2; $σ$ is Stefan–Boltzmann constant, and it is 5.67 × 10⁻⁸ W/(m²·°C)； $ε$ is emissivity of the strand, and it is 0.8; and T_s is surface temperature of the strand, °C.

Thermal properties

The main chemical composition and content of GCr15 bearing steel are shown in Table 3. And its thermal properties required in the research include solidus T_S = 1337 °C, liquidus T_L = 1453°C and solidification latent heat L_f = 215,000 J/kg, respectively. And these were derived from the equilibrium phase diagram of GCr15 bearing steel. Other parameters, such as the solid fraction, density, thermal conductivity coefficient, and specific heat, need to be calculated by equations (8)–(11).

Table 3.

Composition and content of GCr15 gearing steel.

Elements	C	Si	Mn	P	S	Cr
wt.%	1.0	0.21	0.33	0.013	0.0072	1.47

Equation (8) is the relationship between solid fraction f_s and temperature T in mushy zone during the solidification process of GCr15.

f_{S} = \sqrt{\frac{T_{L} - T}{T_{L} - T_{S}}}

(8)

Liquid density $ρ_{L}$ and solid density $ρ_{S}$ are 7800 kg/m³ and 7000 kg/m³ respectively, while the density of mushy zone $ρ_{S L}$ is solved by equation (9).

ρ_{S L} = ρ_{S} \cdot f_{S} + ρ_{L} \cdot (1 - f_{S})

(9)

The thermal conductivity coefficient k can be solved by equation (10). Obviously, the thermal conductivity coefficient k_s in solid is linearly related to temperature, k_L in the liquid is enlarged by four times to consider the effect of molten steel's convection on heat transfer, and k_SL in the mushy zone is in between the both.

k = {\begin{matrix} k_{S} = 18.4 + 0.0096 \times T T < T_{S} \\ k_{S L} = k_{S} (1 + (n - 1) (1 - f_{S})^{2}) T_{S} \leq T \leq T_{L} \\ k_{L} = n \cdot k_{S} T > T_{L}, n = 4 \end{matrix}

(10)

As shown in equation (11), the specific heat c_S in the solid is linearly related to temperature, c_L in the liquid is a constant value, and c_SL in the mushy zone is treated by using the equivalent specific heat method, which includes two parts of the specific heat form original phase change and the increase of specific heat caused by latent heat released.²⁹

c = {\begin{matrix} c_{S} = 550 + 0.0952 \times T T < T_{S} \\ c_{S L} = f_{S} \cdot c_{S} + c_{L} \cdot (1 - f_{S}) - L_{f} \cdot \frac{\partial f_{S}}{\partial T} T_{S} \leq T \leq T_{L} \\ c_{L} = 842 T > T_{L} \end{matrix}

(11)

Mesh independence

Figure 2 shows the comparison of central solid fraction of the strand under grid sizes of 2 mm, 5 mm, 8 mm, and 10 mm, respectively. Results show that their variation patterns are similar with different grid sizes. From the local enlarged view, it can be seen that in the early stage of solidification, several distances of initial solidification position from the meniscus are 4.976 m, 4.968 m, 4.940 m, and 4.919 m successively with grid sizes increasing from 2 mm to 10 mm. Among them, the grid sizes of 2 mm and 5 mm are relatively close, and there is only an 8 mm gap. However, the difference between 2 mm and 8 mm, 10 mm increase evidently, and they are 36 mm and 57 mm, respectively. In the middle stage, both solid fractions are very similar for 2 mm and 5 mm in grid sizes, while other grid sizes’ are relatively high. At the end stage, when the grid size is 2 mm, the solidification endpoint is located at 25.798 m, while they are all at about 25.805 m for other grid sizes, and only existing slight difference. To sum up, it can be seen that simulation results are similar under 2 mm and 5 mm in grid sizes, while there will be obviously deviations when grid sizes are 8 mm and 10 mm. Therefore, for ensuring the accuracy of simulation results and saving computational costs, a grid size of 5 mm is chosen in this article.

Model validation

In this article, the solidification and heat transfer model is verified via comparing the centre temperature of inner arc surface measured by an infra-red thermometer under steady CC(0.43 m/min, 30°C, 0.125 L/kg) with numerical simulation's, and corresponding results are plotted into Figure 3. The measured values are slightly lower than the simulation's, mainly because infra-red temperature measurement is a non-contact measurement that is affected by factors such as water vapour and iron oxide scale in the SCZ. The error of the strand's temperature between both methods is 3.16%∼8.36%, which is within the allowable range. In addition, this study also verified the reliability of the liquid core length of the GCr15 strand calculated via the constructed mathematical model in this article. Due to the current inability to directly obtain experimental data through nail shooting experiments, which is illegal, this paper uses data from the literature for verification. Based on the relevant conditions of 380 mm × 450 mm cross-section GCr15 bearing steel given in the literature,³⁰ a new mathematical model was established according to the modelling ideas in this article. The simulation results are then compared with the literature's data. The calculation conditions are casting speed of 0.44–0.46 m/min, superheat of 26°C, and SCZ's water ratio of 0.11 L/kg. The solidification endpoint position was shown in Table 4. Results indicate that solidification endpoint calculated using the model in this article is basically consistent with the data in the literature. In summary, the mathematical model constructed in this article is reliable.

Table 4.

Comparison of solidification endpoint obtained from literature and numerical simulation.

	0.44 m/min	0.45 m/min	0.46 m/min
Literature's data	22.62 m	23.13 m	23.64 m
Simulation results	21.93 m	22.30 m	23.02 m
Error	3.05%	3.59%	2.62%

Results and discussion

Solidification process of bloom under different process parameters

Solidification morphology

Figure 4 shows the solidification morphology of GCr15 bloom at the solidification end with casting speeds of 0.40 m/min, 0.43 m/min, and 0.46 m/min, superheat of 30°C and water ratio of 0.125 L/kg, where the grey area represents liquid core and the remaining is solidified shell. According to the figure, three kinds of liquid core's shapes are basically the same, and they are all slender conical shape along the casting direction. This indicates that under current conditions, the cooling uniformity of 390 mm × 480 mm bloom at the circumferential direction is well. And this solidification process can achieve a relatively uniform shell thickness of the strand, which is beneficial for preventing bulging defects and even leakage accidents. Nevertheless, a significant drawback associated with this solidification behaviour is the severe centre segregation observed in GCr15 (high carbon steel) bloom. This is due to the enrichment of solute elements towards solidification front under selective crystallization, and then the enriched solute elements gradually concentrate towards the strand core with the solidification progressing, ultimately forming centre segregation after complete solidification of the strand.

To bloom centre segregation, the most commonly used solution is to apply F-EMS and SR/HR processes. Besides, in order to achieve well improving effect, it is necessary to regulate the process parameters of CC based on bloom's solidification process. Figure 4 displays the influence of casting speed on the strand's solidification behaviour. The solidification endpoints at three different casting speeds are 23.99 m, 25.81 m, and 27.56 m respectively, and it is far away from the meniscus with the raising of casting speed. In addition to casting speed, superheat and SCZ's water ratio are also key control parameters, and for facilitating fine control, it is also necessary to further clarify the solidification behaviour of the strand under different process parameters. It is worth noting that solidification morphology under different superheats and SCZ's water ratios is basically consistent with that shown in Figure 4. Therefore, the figure is not being drawn repeatedly in the subsequent chapters, and only some key information such as central solid fraction and solidification endpoint will be extracted for studying.

Changing of central solid fraction

Solid fractions under different casting speeds, superheats and SC's water ratios were plotted into Figure 5. It is known that as casting speed raises from 0.40 m/min to 0.46 m/min, the solidification endpoint moves backwards by 3.57 m along the casting direction. For that every 0.01 m/min increases in casting speed, it moves by about 0.60 m on average. When the superheat rises from 20°C to 50°C, the solidification endpoint changes from 25.44 m to 26.49 m from the meniscus, moving backwards by 1.05 m. In other word, superheat increases by 1°C, and the solidification endpoint moved backwards by 0.035 m correspondingly. Contrary to casting speed and superheat, as the SCZ's water ratio increases, the strand will solidify completely in advance. With the rising of water ratio from 0.076 L/kg to 0.150 L/kg, the solidification endpoint moves from 27.62 m away from the meniscus to 25.04 m, changed by 2.58 m. On average, for every increasing 0.001 L/kg in water ratio, it alters 0.036 m. Comparing three kinds of process parameters of CC, it can be seen that the casting speed has the most significant effect on the solidification process of GCr15 bloom, followed by the SCZ's water ratio, and superheat is the smallest.

For the convenience of quantitatively controlling of casting speed, superheat, and water ratio in practice, Figure 6 was plotted by extracting solidification endpoint data from Figure 5 in this article. On the basis of this, the quantitative relationships between casting speed, superheat, water ratio and endpoint position were fitted and determined, as shown in equations (12), (13), and (14).

L = 0.2232 + 59.4643 V

(12)

L = 24.7521 + 0.0350 Δ T

(13)

L = 30.1820 - 34.7166 Q

(14)

where L is distance of solidification endpoint from the meniscus, m; V is casting speed， m/min;

Δ T

is superheat of molten steel, °C; and Q is SCZ's water ratio, L/kg.

Analysis of reduction system and controlling of process parameters

Based on the solidification process calculated, the implementing condition of reduction process under casting speed of 0.43 m/min, superheat of 30°C, and SCZ's water ratio of 0.125 L/kg in a steel plant was determined, as shown in Figure 7. There are 10 pairs of reduction rollers, of which 6 and 7 rollers perform HR, while the rest are SR rollers. In actual production, the factory only starts 2∼8 reduction rollers, and the total reduction amount is 32 mm, correspondingly the central solid fraction of strand is f_s = 0.62∼1.0. From test results of the strand, it can be known that central porosity and shrinkage cavity of GCr15 bloom has been promoted significantly under current reduction system. Although the centre segregation has also been improved to a certain extent, the improvement is limited and cannot meet the acceptance requirements. So that, it is imperative to adjust the current reduction system for further reducing the centre segregation of the strand and assisting homogenised CC.

According to the analysis of current reduction system, there exist two main problems: (1) central solid fraction f_s = 0.62 at the initial reduction position 2# rollers is relatively large. Because the central solid fraction selects generally around f_s = 0.3∼0.4^25–28 when improving effectively centre segregation through the reduction process. (2) 7# HR rollers is in solid core reduction, namely f_s = 1.0. The sudy³⁰ has shown that the delivering effect of deformation to the strand centre will be greatly weakened when adopting solid core reduction, which is not conducive to the improvement of central segregation, porosity, and shrinkage cavity. Aiming at these issues, changing the reduction system is the most effective way, and detailed operations are as follow: (1) Starting the 1# reduction rollers and implementing reduction in advance. (2) Resetting the reduction interval to decrease the central solid fraction of initial reduction rollers and also place 7# HR rollers in a liquid core reduction, thereby maximising SR/HR effectiveness. However, reduction rollers are fixed in actual, and consequently the direct method to alter reduction interval is to regulate process parameters such as casting speed, superheat, and SCZ's water ratio, and then realising that the solidification process moves back as a whole.

The research in the changing of central solid fraction section indicates that casting speed has the most remarkable impact on the strand's solidification process, and it is also easy to adjust in practice, so it should be given priority consideration. Similarly, increasing superheat can also promote solidification endpoint to move back, and achieving the aforementioned reduction control strategy. However, this conflicts with the existing research results,¹⁰ namely, low superheat casting is beneficial for reducing centre segregation. After comprehensive consideration, reasonable superheat should be controlled at 20∼30°C. Although reducing SCZ's water ratio can realise the backward movement of solidification process, its influence is relatively small and likely causing excessive temperature re-rising and surface crack in the SCZ. Therefore, it is best to maintain the original water ratio. In summary, a reasonable strategy for adjusting process parameters is to increase casting speed, control superheat at 20–30°C, and adopt the original SCZ's water ratio of 0.125 L/kg.

For the detailed adjustment of casting speed, it must satisfy the following requirements: firstly, ensuring 7# HR rollers to be in liquid core reduction. Secondly, the solidification endpoint of GCr15 bloom needs to be appropriately moved back and controlled between the 7# rollers (26.993 m from the meniscus) and the 8# rollers (28.643 m from the meniscus). Based on this, combined with the quantitative relationship (equation 12), the casting speed should be controlled at 0.450 m/min < V < 0.478 m/min to achieve this goal.

Influence of process parameters on stirring region of F-EMS

The appropriate stirring region is an important guarantee that F-EMS realises well improving effect to centre segregation of the strand. In production, process parameters of CC such as casting speed, superheat, and SC's water ratio are usually adjusted to ensure that F-EMS is in the appropriate range. However, it is worth noting that changing process parameters to control F-EMS is not independent, and its impact on the subsequent SR/HR process also needs to be considered. Therefore, associated with results in Chapter 3.2, the main process parameter that can be applied to the control of F-EMS is casting speed.

The solidification rate F_S of the strand, which is the area proportion of solidified shell in cross-section of the strand (solid phase fraction f_s greater than 0.7 is considered as solidified shell³⁰), is usually used as the evaluation standard for stirring position. Figure 8 shows solid phase fraction's distribution of the strand cross-section at the inlet (12.725 m from the meniscus) and outlet (13.560 m from the meniscus) of F-EMS device under steady production conditions (0.43 m/min, 30°C, 0.125 L/kg). On the basis of this, the solidification rate can be obtained by equation (15). Under the steady CC, the solidification rate of F-EMS's action region is about 0.71∼0.73.

F_{S} = \frac{S_{0} - S_{1}}{S_{0}}

(15)

where S₀ is area of bloom cross-section, m²; S₁ is area of the incomplete solidified region in the strand, m². Due to this area presents an elliptical shape, it can be computed by using the corresponding area formula.

Similarly, the solidification rate at the inlet and outlet of F-EMS was obtained by using the same method under other casting speeds, and then these data were plotted in Figure 9. By linearly fitting the solidification rate data at different casting speeds, equations (16) and (17) can be gained.

F_{S, E n t r a n c e} = 1.3273 - 1.4417 V

(16)

F_{S, E x i t} = 1.3663 - 1.4774 V

(17)

where F_S,Entrance and F_S,Exit are solidification rate of the strand cross-section at the inlet and outlet of F-EMS.

Research on the position of F-EMS has shown that a cross-sectional solidification rate of 0.6∼0.8 for high carbon steel is a reasonable stirring range.^18–21 Based on this, using equations (16) and (17), it can be concluded that the corresponding casting speed should be controlled within 0.383 m/min < V < 0.504 m/min. Combined with results of the control of SR/HR, and considering that the minimum variation unit of casting speed in actual operation is only 0.01 m/min, it can be inferred that the available casting speeds for achieving the goal of homogenised CC are 0.46 m/min or 0.47 m/min, respectively. The corresponding solidification rates in the stirring zone of F-EMS are 0.66∼0.69 and 0.65∼0.67.

Effect of process parameters on smooth running of CC

Adjusting casting speed not only changes the action region of F-EMS and SR/HR, but also affects the shell thickness of mould outlet and the surface's temperature re-rising of the strand in the SCZ. If the shell thickness is too thin, bulging or even leakage may occur, similarly, excessive temperature re-rising may cause surface cracks. As a result, this article analysed the changes of shell thickness at mould outlet and surface temperature in the SCZ at casting speeds of 0.46 m/min and 0.47 m/min.

Figure 10 shows the morphology of solidified shell at the mould outlet with 0.46 m/min and 0.47 m/min in casting speed. The solidified shell along the circumference presents basically uniform distribution, and the thinnest part appears near the corner. This is because there exists air gap in the mould corner, which remarkably weakens the heat transfer, eventually resulting in the remelting of solidification front and further the decrease of shell thickness. The shell thicknesses at casting speeds of 0.46 m/min and 0.47 m/min are 14.00∼17.80 mm and 13.51∼17.56 mm, respectively. And the latter reduces by about 0.3∼0.5 mm compared with the former, which is relatively smaller. There is relative research that indicate the safety shell thickness at mould outlet is above 10∼15 mm.³⁰ Therefore, it can be concluded that increasing the casting speed from 0.43 m/min to 0.46 m/min or 0.47 m/min can achieve smooth CC without leakage accidents.

The temperature variations of the strand's surface centre under two casting speeds are shown in Figure 11. Results display that changing laws are similar along the casting direction at 0.46 m/min and 0.47 m/min. There is a sharp decline in the mould. Subsequently, after entering the SCZ (Zone 1, Zone 2, Zone 3, Zone 4), there exists an alternating increase and decrease's variation in the temperature of the strand, which is because the dynamic alternation between the heat carried away by spraying water and the heat transferred from the strand core to the surface. And at the end of SCZ's Zone 4, the strand's surface temperature remains stable at around 1250°C. Afterwards, the strand moves into air cooling zone and is governed by radiation heat transfer, leading to a continuous decrease in surface temperature, ultimately reaching about 860∼900°C. In addition, because of the difference of shell thickness and HTC, the temperature of wide and narrow faces of the strand differs by 10∼40°C.

On the other hand, according to Figure 11, it is known that the maximum temperature re-rising in the SCZ occurs in Zone 2, and they are respectively 107°C and 121°C at the wide and narrow faces. Both temperatures are below the reasonable range 100∼150°C.³¹ Therefore, when the casting speed is adjusted to 0.46 m/min or 0.47 m/min, the strand's temperature re-rising is suitable and cannot cause surface cracks.

Control strategy of CC's process parameters

According to the research above-mentioned, under the premise of shell thickness is above the safety thickness and temperature re-rising of the SCZ cannot result in cracks as well as the stirring position of F-EMS is suitable, for realising the moving towards the direction away from the meniscus of solidification process of GCr15 bloom and further ensuring that 7# HR rollers is in the liquid core reduction, the control strategy of process parameters of CC should be 0.46 m/min or 0.47 m/min in casting speed, 20∼30°C in superheat, and 0.125 L/kg in SCZ's water ratio. In addition, in order to further raise the effect of SR/HR on improving centre segregation, a reasonable operation is to start 1# reduction rollers, thereby reaching the destination of lowering central solid fraction of initial reduction rollers. Figure 12 is a schematic diagram of the implementation of SR/HR after adjusting the CC's process parameters. At the casting speed of 0.46 m/min, the central solid fraction of the strand in 1#∼8# reduction range is f_s = 0.43∼1.0, and correspondingly it is f_s = 0.78∼0.91 for HR rollers. While at 0.47 m/min, the solid fraction is f_s = 0.41∼1.0, and HR rollers corresponds to f_s = 0.74∼0.86. Both of them are within the reasonable central solid fraction f_s = 0.4∼1.0 and corresponding to f_s ≥ 0.7 for the HR roller,^23–28 which can effectively improve centre segregation, porosity and shrinkage cavity. This is very significant to guiding the homogenised CC of GCr15 bearing steel.

Centre segregation results in industrial production

According to the research results of this article, the CC process was adjusted for industrial production, and then the transverse samples of strand were taken for low magnification characterisation and carbon element detection, and the centre carbon segregation index was calculated.

Figure 13 shows the results of low magnification and centre carbon segregation of the bloom before and after the improvement of the CC process. From Figure 13, it can be seen that by adjusting CC process parameters such as casting speed and superheat, the reduction process has achieved the effect of improving bloom segregation. The industrial production results indicate that the centre segregation index of GCr15 bloom has decreased from 1.18 to 1.09, and the effect is significant.

Conclusions

This article investigated solidification and heat transfer behaviour of GCr15 bloom via numerical simulation and determined the homogenised process parameters of CC. The main conclusions are as follows.

The morphology of liquid core at the solidification end is a slender conical shape, which easily causes solute element enrichment towards the core, and resulting in central segregation.

Casting speed has the most significant impact on the solidification process, followed by water ratio, and superheat is the smallest. For increasing every 0.01 m/min in casing speed, the movement of solidification endpoint is about 0.60 m on average.

When the casting speed is in 0.450 m/min < V < 0.478 m/min, 7# HR rollers can be in a liquid core reduction, thereby ensuring the effect of improving centre segregation through reduction process.

After increasing casting speed, the solidification rate of F-EMS's stirring region is F_S = 0.65∼0.69, shell thickness at mould outlet is about 13.51–17.81 mm, and the maximum temperature re-rising in the SCZ is around 121°C, all of which are within a suitable range.

The appropriate process parameters for homogenised CC of GCr15 bloom is in casting speed of 0.46 m/min or 0.47 m/min, superheat of 20∼30°C, SCZ's water ratio of 0.125 L/kg and running 1# reduction rollers. And central solid fractions f_s = 0.43∼1.0 (0.46 m/min) and f_s = 0.41∼1.0 (0.47 m/min) of SR/HR range after improving process parameters are more reasonable. After improving the process, the centre segregation of the strand was reduced from 1.18 to 1.09, with a significant effect.

Footnotes

Acknowledgements

The authors sincerely thank Bingjun Lu and Deyong Wang for provision of device and production data.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Group Major Projects in Ansteel Beijing Research Institute and the Regional Company Projects in Ansteel Beijing Research Institute (Grant Nos. 2023AGB1005-2 and 2022BJB-18BG).

ORCID iD

Pei Xu

References

Sun

Zhang

. Study on the macrosegregation behavior for the bloom continuous casting: model development and validation. Metal Mater Trans B 2014; 45: 1133–1149.

Kim

Kang

, et al. Feasibility of using continuously cast round bloom as a substitute to cast ingot in the manufacture of heavy forgings. Steel Res Int 2020; 91: 2000079.

Kenzo

Hido

Kazuyuki

, et al. Low superheat teeming with electromagnetic stirring. ISIJ Int 1995; 35: 680–685.

Long

Chen

. Study on mitigating center macro-segregation during steel continuous casting process. Steel Res Int 2011; 82: 847–856.

Chen

Long

Chen

, et al. A novel flux splitting scheme with robustness and low dissipation for hypersonic heating prediction. Int J Heat Mass Transfer 2018; 127: 126–137.

Tzavaras

Brody

. Electromagnetic stirring and continuous casting — achievements, problems, and goals. JOM 1984; 36: 31–37.

Ayata

Mori

Fujimoto

, et al.

Improvement of macrosegregation in continuously cast bloom and billet by electromagnetic stirring.

ISIJ Int 1984; 24: 931–939.

Moon

Lee

, et al. Effect of the roll surface profile on centerline segregation in soft reduction process. ISIJ Int 2012; 52: 1266–1272.

Zhu

. Closure of internal porosity in continuous casting bloom during heavy reduction process. Metal Mater Trans B 2019; 50: 2867–2883.

10.

Choudhary

Ganguly

. Morphology and segregation in continuously cast high carbon steel billets. ISIJ Int 2007; 47: 1759–1766.

11.

Hou

Cheng

Jiang

, et al. Compactness degree of longitudinal section of outer columnar grain zone in continuous casting billet using cellular automaton-finite element method. ISIJ Int 2013; 53: 655–664.

12.

Maurya

Jha

. Influence of electromagnetic stirrer position on fluid flow and solidification in continuous casting mold. Appl Math Model 2017; 48: 736–748.

13.

Lan

Tang

, et al. Study on the electromagnetic field, fluid flow, and solidification in a bloom continuous casting mold by numerical simulation. Steel Res Int 2018; 89: 1800071.

14.

Bao

Wang

, et al. Effects of electromagnetic stirring on fluid flow and temperature distribution in billet continuous casting mould and solidification structure of 55SiCr. Metall Res Technol 2018; 115: 103.

15.

Yang

Wang

Zhang

, et al. Effect of electromagnetic stirring on molten steel flow and solidification in bloom mold. J Iron Steel Res Int 2014; 21: 1095–1103.

16.

Zhang

Luo

Chen

, et al. Numerical simulation of fluid flow, heat transfer, species transfer, and solidification in billet continuous casting mold with M-EMS. Metals (Basel) 2019; 9: 66.

17.

Wang

Zhang

Yang

, et al. Effect of mold electromagnetic stirring and final electromagnetic stirring on the solidification structure and macrosegregation in bloom continuous casting. Steel Res Int 2021; 92: 2000661.

18.

Xiao

Zhang

Luo

, et al. Control of macrosegregation behavior by applying final electromagnetic stirring for continuously cast high carbon steel billet. J Iron Steel Res Int 2013; 20: 13–20.

19.

Sun

, et al. On the alternate stirring mode of F-EMS for bloom continuous castings. Metal Mater Trans B 2018; 49: 1909–1918.

20.

Jiang

Zhu

. Center segregation with final electromagnetic stirring in billet continuous casting process. Metal Mater Trans B 2017; 48: 444–455.

21.

Sen

Piao

Jiang

, et al. Numerical simulation and experimental study of F-EMS for continuously cast billet of high carbon steel. J Iron Steel Res Int 2014; 21: 51.

22.

Zeze

Misumi

Nagata

, et al. Segregation behavior and deformation behavior during soft-reduction of unsolidified steel ingot. Tetsu-to-hagané 2001; 87: 71–76.

23.

Ogibayashi

Kobayashi

Yamada

, et al.

Influence of soft reduction with one-piece rolls on center segregation in continuously cast slabs.

ISIJ Int 1991; 31: 1400–1407.

24.

Zhu

. Thermo-mechanical behavior of the continuous casting bloom in the heavy reduction process. JOM 2016; 68: 3107–3115.

25.

Dong

Zhang

Wang

, et al. Shrinkage porosity and its alleviation by heavy reduction in continuously cast strand. J Mater Process Technol 2016; 238: 81–88.

26.

Zhao

Liu

Wang

, et al. Effects of heavy reduction technology on internal quality of continuous casting bloom. Ironmaking Steelmaking 2019; 46: 227–234.

27.

Zhao

Zhang

Lei

, et al. The position study of heavy reduction process for improving centerline segregation or porosity with extra-thickness slabs. Steel Res Int 2014; 85: 645–658.

28.

Zhao

Zhang

Lei

, et al. Dynamic recrystallization (DRX) analysis of heavy reduction process with extra-thickness slabs. Steel Res Int 2014; 85: 811–823.

29.

Lally

Biegler

Henein

. Finite difference heat-transfer modeling for continuous casting. Metall Trans B 1990; 21: 761–770.

30.

Yang

Wang

Zhang

, et al. Optimization of heavy reduction process on continuous-casting bloom. Metals (Basel) 2022; 12: 1873.

31.

Chaudhuri

Singh

Patwari

, et al. Design and implementation of an automated secondary cooling system for the continuous casting of billets. ISA Trans 2010; 49: 121–129.