High Temp. Mater. Proc.

2017; 36(7): 741–753

Kun Dou, Zhenguo Yang, Qing Liu*, Yunhua Huang and Hongbiao Dong

Influence of Secondary Cooling Mode

on Solidification Structure and Macro-segregation
Behavior for High-carbon Continuous Casting
Bloom
DOI 10.1515/htmp-2016-0022 Introduction
Received January 27, 2016; accepted May 31, 2016

Abstract: A cellular automaton–finite element coupling As an important type of high-carbon steel, solidification
model for high-carbon continuously cast bloom of GCr15 structure and macro-segregation of GCr15 steel in
steel is established to simulate the solidification structure continuous casting process have obvious influence on
and to investigate the influence of different secondary the quality of the final rolled products [1]. Desired
cooling modes on characteristic parameters such as solidification structure can be obtained by optimization
equiaxed crystal ratio, grain size and secondary dendrite of solidification parameters, for which numerical
arm spacing, in which the effect of phase transformation simulation is an economic and desirable way.
and electromagnetic stirring is taken into consideration. Over the past decades, several models for predicting
On this basis, evolution of carbon macro-segregation for the solidification structure formation have been devel-
GCr15 steel bloom is researched correspondingly via oped, including stochastic models such as the Monte
industrial tests. Based on above analysis, the relationship Carlo method [2] and cellular automaton (CA) method
among secondary cooling modes, characteristic para- [3, 4] and deterministic models like the phase-field
meters for solidification structure as well as carbon method [5, 6] Many attempts have been carried out to
macro-segregation is illustrated to obtain optimum sec- simulate solidification structure formation of alloys using
ondary cooling strategy and alleviate carbon macro-seg- these methods. However, applications of these models
regation degree for GCr15 steel bloom in continuous are mainly focused on ingot casting [7, 8] and few
casting process. The evaluating method for element attempts have been made in simulating solidification
macro-segregation is applicable in various steel types. structure formation in continuous casting. Two reasons
account for this. First, forced flow in melt during
Keywords: carbon macro-segregation, cellular automaton– solidification is too complex to be simulated, such as
finite element method, solidification structure, secondary electromagnetic stirring (EMS), bulging, soft reduction
cooling mode, continuous casting and so on. Second, complex transformations such as
peritectic reaction occur during the solidification of
steel, which make modeling of solidification structure in
continuous casting very difficult. The formation mechan-
ism of carbon macro-segregation is closely associated
*Corresponding author: Qing Liu, State Key Laboratory of Advanced
Metallurgy, University of Science and Technology Beijing, Beijing with solidification structure for continuous casting of
10083, China, E-mail: qliu@ustb.edu.cn steel while the variation of secondary cooling mode
Kun Dou, State Key Laboratory of Advanced Metallurgy, University of would have obvious impact on the solidification structure
Science and Technology Beijing, Beijing 10083, China, such as columnar-to-equiaxed transition (CET), dendrites
E-mail: dou_kun@163.com
orientation, grain size (GS) [9, 10]. Hence, it would be
Zhenguo Yang, State Key Laboratory of Advanced Metallurgy,
necessary to illustrate the impact of various secondary
University of Science and Technology Beijing, Beijing 10083, China;
Special Steel Plants, Laiwu Iron and Steel Co., Ltd., Laiwu 271104, cooling modes on solidification structure characteristics
Shandong, China, E-mail: lwyzg@163.com and describe the formation mechanism of carbon macro-
Yunhua Huang, Institute for Advanced Materials and Technology, segregation in GCr15 steel continuous casting process.
University of Science and Technology Beijing, Beijing 100083, In this paper, the cellular automaton–finite element
China, E-mail: huangyh@mater.ustb.edu.cn
(CA–FE) coupling model [11, 12] is used to simulate the
Hongbiao Dong, Department of Engineering, University of Leicester,
Leicester LE1 7RH, UK, E-mail: hd38@leicester.ac.uk solidification structure of GCr15 steel bloom under
742 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

different secondary cooling modes in continuous casting process. The CA–FE model simulating the solidification
process, in which the variation of thermophysical para- structure mainly includes heat transfer model, nucleation
meters during solidification and the effect of EMS are model and dendrite tip growth kinetics.
taken into consideration. On this basis, industrial tests
under the same casting condition are conducted to obtain
specimens for carbon segregation and solidification Heat transfer model
structure analysis. Finally, the relationship between car-
bon macro-segregation and solidification structure evolu- A two-dimensional unsteady state heat transfer equation
tion under different secondary cooling modes is is available as eq. (2) [14]:
discussed.
 ∂T  ∂  ∂T 
ρc ∂T ∂
∂t = ∂x k ∂x + ∂y k ∂y (2)

where ρ is density, c is specific heat, keff is thermal

Thermophysical properties conductivity. The evolution of latent heat during solidifi-
calculation cation is incorporated to the calculation by using the
effective specific heat method, shown in eq. (3).
A simple pair-wise mixture model (shown in eq. (1) [13])  
dfs
based on the thermodynamic database from Procast soft- Cp ′ = Cp − L (3)
dT
ware is used to calculate the thermophysical properties of
GCr15 steel at different temperatures, including thermal where Cp ′ is the effective specific heat, L is the latent heat
conductivity, density, enthalpy, etc. The calculated and fs is the solid fraction.
results are coupled into the FE model later.[13] The heat transfer model based on the moving slice
 V method is established to simulate the solidifications of
P PP P
P= Xi Pi + Xi Xj ΩV Xi − Xj (1) GCr15 steel. Figure 2 shows the boundary conditions of
i j>i v
the heat transfer model during the continuous casting
where P and Pi are the thermophysical properties of one process. The section size of slice is the same as the
phase and pure element, respectively. ΩV is the binary bloom, 260 mm × 300 mm and the thickness is 10 mm.
interaction parameter, Xi and Xj are the mole fractions of In this research, the secondary cooling zone is divided
elements i and j. into five segments according to water flow rate, such as
Chemical compositions of GCr15 steel are given in foot roller zone, I zone, II zone, III zone and IV zone.
Table 1. Temperature dependences of enthalpy, density, The length, boundary conditions and calculated
thermal conductivity and fraction of solid are calculated formula are listed in Table 2. In Table 2, T, Tw, Te are
by importing them into the thermodynamic database, as the temperature of the bloom surface, the water tem-
depicted in Figure 1. It could be seen that variation of perature in the secondary cooling zone, the environ-
enthalpy, density and thermal conductivity with tempera- ment temperature of the air cooling zone, respectively,
ture indicated several turning points, e. g. the range of K; qm, qk,i and qa are the heat flux of mold, segment i
1,613–1,723 K and 893–1,043 K corresponds to the phase in secondary cooling zone and air cooling zone, respec-
transformation l → γ and γ → α, respectively. tively, kW/m2; t is the holding time in the mold, s; β is
a coefficient about the conditions of the mold, kW/
(m2 s1/2); hi is the heat transfer coefficient of segment i
Numerical model in the secondary cooling zone, W/(m2 K); Wi is the
sprayed water density of segment i in secondary cool-
In the present model, the CA method is combined with ing zone, L/(m2 min); σ is Stefan–Boltzmann constant,
heat transfer calculation during the continuous casting W/(m2 K4); ε is the radiation coefficient. The heat flux

Table 1: Main chemical compositions for GCr15 steel (wt. %).

Steel C Si Mn P S Cr Ni Cu Al Mo Ti

GCr . . . . . . . . . . .
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 743

Figure 1: Calculated
thermophysical proper-
ties of GCr15 steel.

of mold is assumed to be a function of the casting

speed and the distance from the meniscus. The most
important parameter that affects the mold heat flux qm
is the mold dwell time (equivalent to distance below
the meniscus) neglecting the steel grade, entry nozzle
geometry, mold flux, etc.
In the secondary cooling zone, the surface heat flux
qk,i is proportional to the heat transfer coefficient hi as
well as the difference between the surface temperature
and the temperature of cooling water. The radiative heat
transfer to the atmosphere is neglected in water cooling
zone.

Nucleation model

Nucleation can be divided into homogeneous nucleation

and heterogeneous nucleation during solidification pro-
cess. In the present study, the continuous heterogeneous
model [17] is applied. A continuous nucleation distribu-
tion function, dn/d(ΔT), is used to describe the grain
density change, dn is induced by increase of the under-
Figure 2: Schematic illustration of boundary conditions and the slice cooling, d(ΔT). The distribution function is expressed by
moving method. eq. (4) [17].
744 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

Table 2: Boundary conditions and the calculated formula of GCr15 steel.

Section Length/m Boundary condition Calculated formula

 pffiffi
Mold . qm qm = 2.68 − β t × 103 [15]
Secondary . qk,i = hi(T − Tw) hi =  + .Wi. []
cooling zone
Air cooling zone . qa = εσ[(T + ) − (Te + )] σ = . × −ε = . []

 2 Givoanola, Trivedi) model [18, 19] is used as the model

ΔT − ΔTmax
dn
dðΔT Þ = pnffiffiffiffi
max
2π ΔT
exp − 1
2 ΔTσ (4)
σ of growth kinetics of a dendrite tip in the GCr15 steel.
Based on the marginal stability criterion, eq. (5) is
where ΔT is the calculated local undercooling, K; ΔTmax is
obtained.
the mean undercooling, K; ΔTσ is the standard deviation,
K; nmax is the maximum nucleation density which can be V 2 A + VB + C = 0 (5)
reached when all the nucleation sites are activated while
where A = π 2 P 2 D2 , B = mC0 ð1 − k0 Þc D½1 − ð1 − k0 ÞIvðP Þ,
cooling, m−3.
C = G, Γ is the Gibbs–Thomoson coefficient (Γ =
Gaussian distribution is used to describe heteroge-
1.9 × 10 − 7 ), V is the growth velocity of a dendrite tip, P
neous nucleation both at the mold surface and in the
is the Peclet number for solute diffusion, D is the diffu-
bulk of the melt. Two such functions are shown in
sion coefficient in the liquid, m is the liquidus slope, C0 is
Figure 3 [3].
the initial concentration, k0 is the partition coefficient, Iv
(P) is the Ivantsov function, ξc = π2/(k0p) and it closes to
unity at low temperature gradient, G is the temperature
Dendrite tip growth gradient. For the dendrite growth regime, G has little
effect on the growth velocity V and can be regarded as
The growth kinetics of both columnar and equiaxed
zero.
morphologies can be calculated. The KGT (Kurz,
Ttip is the temperature at the dendrite tip, T0 is the
melting point of GCr15 steel (T0 = 1,630 K), ΔT is the
undercooling temperature at a tip of dendrite
(ΔT = Ttip − T0), ΔT is expressed as eq. (6).
h i
ΔT = mC0 1 − 1 − ð1 − k10 ÞIvðPÞ + 2r (6)

where r is the dendrite tip radius. The relationship

between the undercooling ΔT and growth velocity V can
be calculated by substituting an arbitral value of the
Peclet number into eqs. (5) and (6). The material proper-
ties of GCr15 steel used in the simulation are given in
Table 3.
In order to accelerate the computation velocity dur-
ing the simulation process, KGT model is fitted and the
following equation (eq. (7)) is gained.

vðΔTÞ = a2 ΔT 2 + a3 ΔT 3 (7)

where a2 and a3 are the fitting coefficients of the

KGT model, both of them reflect the growing velocity of
the dendrite tip. ΔT is the total undercooling of
the dendrite tip. Using the simulation and based on
Figure 3: Nucleation site distributions for nuclei formed at the mold
Table 3, the calculated values of a2 and a3 are 0 and
wall (indexed as “s”) and in the bulk of the liquid (indexed
as “v”) [3].
1.354 × 10−5m/(s K3).
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 745

Table 3: Materials properties of GCr15 steel used in the simulation.

Composition C Si Mn P S

Partition coefficient, k [] . . . . .

Liquidus slope, m − − − − −
Diffusivity in liquid, D, m/s [] . × − . × − . × − . × − . × −

Effect of EMS

EMS is generally used to reduce the centerline segrega-

tion in the continuous casting process. To evaluate the
effect of the fluid flow due to EMS on heat flow in the
bloom, accurate flow pattern in the molten steel melt
should be known. However, at present, it is difficult to
combine accurate calculation of fluid flow with heat
transfer calculation and the CA procedure due to very
large computational load. Two methods have been
reported for incorporating the effect of fluid flow into
heat transfer calculation.
(1) Changing thermal conductivity of liquid during EMS
[22–24]. By this means, it is considered that the
thermal conductivity of molten steel with EMS is
larger than that without EMS. Based on the solid– Figure 4: Schematic illustration of solid–liquid coexisting zones
liquid coexisting zone model proposed by Takahashi during the solidification of steel.
et al. [24], thermal conductivity of solid is used in
the region of solid and the mushy zone with solid
fraction (fs) larger than 0.7 while the effective ther- the complex mechanism of fragmentation caused
mal conductivity of liquid with fluid flow is used in by fluid flow. In this simulation, the effective crys-
the liquid region and mushy zone with fs < 0.3. In the tal formation rate is set to 25 cm−2 s−1 higher than
mushy zone with 0.3 ≤ fs ≤ 0.7, thermal conductivity that of heterogeneous nucleation in liquid.
is assumed to change with fs linearly, as shown in
Figure 4. The exact value of the thermal conductivity
in liquid with EMS is not clear since it varies with Model validation
the flow velocity of the liquid. Hence, several testing
are required to find an appropriate one to describe To prove the accuracy of the established model, indus-
the real casting conditions in simulation. Mizikar trial test and numerical simulation for GCr15 steel contin-
[22] used a 7.5 times higher thermal conductivity uous casting blooms are conducted under the same
while a 10 times higher thermal conductivity was casting condition, shown in Table 4. Bloom surface tem-
used in Yamazaki’s study [25]. In the present simula- peratures at typical positions in the continuous casting
tion, an 8.5 times higher thermal conductivity is process are measured using infrared radiation thermo-
assigned after many tests. meter, comparisons between measured and calculated
(2) Changing crystal formation rate in bulk liquid bloom surface temperatures are shown in Figure 5,
[8, 26]. In this way, it is considered that the crystal which shows well accuracy in temperature distribution
formation rate would increase due to the contribu- calculation.
tion of fragmentation of dendrites caused by EMS, Besides, bloom samples are sliced and etched by
which should be incorporated into the nucleation H2O–50 %HCl reagent for solidification structure observa-
model. However, accurate effect of the dendrites tion and equiaxed crystal ratio (ECR) calculation. The
fragmentation is difficult to describe because of calculation method for ECR is illustrated in eq. (8).
746 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

Table 4: Casting parameters used for model validation. tion structures include three parts: outer chill zone, inter-
mediate columnar zone and central equiaxed crystal
Parameters Value zone. It can be seen from Figure 6 that the simulated
Steel grade GCr solidification structures by CA–FE model are compatible
Casting speed . m/min with the actual results and the experimentally observed
Superheat  K and simulated ECR are about 23 %. This result indicates
Mold water amount  m/h that the selected nucleation parameters are reasonable
Mold EMS status  A/ Hz
and the present model can be used to simulate the
Secondary cooling water flow rate . L/kg
solidification structure during continuous casting of
GCr15 steel.

Results and discussion

Solidification structure evolution under

different secondary cooling modes

In the actual continuous casting process of GCr15 steel

bloom, five different types of secondary cooling modes
are usually applied. The water distribution profiles
under five different cooling modes are illustrated in
Figure 7. Values of the five axes (Qi) represent the
water flow rate of segment i in the secondary cooling
zone, L/min, which is calculated as eq. (9).
Figure 5: The calculated temperature profiles with measured data Qi
Wi = (9)
points. Ai
where Wi is the sprayed water density of segment i in
SE
ECR = × 100% (8) secondary cooling zone, L/(m2 min) (seen in Table 2); Ai
ST
is the sprayed area of segment i, m2.
where SE is the area of equiaxed crystal zone, ST is the Here, in this paper, the water distribution profiles
total area of bloom cross section. from Figure 7 are used in the boundary condition calcu-
The observed and calculated solidification structures lation for secondary cooling zone as in Table 2 while
are shown in Figure 6(a) and (b). The typical solidifica- other parameters are the same as Table 4.

Figure 6: Morphology of actual/simu-

lated GCr15 steel bloom (cross section).
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 747

Influence of secondary cooling mode on secondary

dendrite arm spacing

The evolution of secondary dendrite arm spacing (SDAS)

at the cross-sectional direction of the GCr15 steel bloom
under different secondary cooling modes is shown in
Figure 10(a)–(e).
It could be deduced from Figure 10 that under all types
of secondary cooling mode, SDAS at the cross-sectional
direction of the GCr15 steel bloom increases from bloom
edge to center. During the steel continuous casting process,
the chill zone of the continuous casting bloom forms due to
rapid cooling in the mold, after which large quantities of
dendrites grow toward bloom center with the dendrite tip
undercooling and dendrite growth velocity decrease gradu-
ally until the CET occurs. SDAS in typical positions of the
Figure 7: Water distribution profiles in five different secondary
cooling modes. bloom (Figure 11) is analyzed statistically as in Figure 12.
It could be observed from Figure 12 that under certain
secondary cooling mode, SDAS value increases from bloom
edge to center. Take the super weak cooling mode as an
Influence of secondary cooling mode on equiaxed crystal example, with the sample point change from C11 to C0, the
ratio and grain size SDAS value increases from 67.27 µm to 96.57 µm. With the
secondary cooling mode changes, SDAS value of the same
The water distribution of various secondary cooling modes sample point decreases with the strengthening of cooling
are input into the CA–FE model and the cross-sectional intensity. For sample point C0, with the secondary cooling
morphology of solidification structure for GCr15 steel con- mode varying from super weak cooling to super strong cool-
tinuous casting bloom is shown in Figure 8(a)–(e). It could ing, the SDAS value decreases from 96.57 µm to 93.02 µm.
be observed that with the change of cooling intensity, the Considering above results, the secondary cooling
central equiaxed crystal area varies evidently. The mode for GCr15 steel continuous casting could be suitably
increased cooling intensity leads to rise of undercooling at controlled to obtain finer grains and favorable equiaxed
solid/liquid phase interface, which contributes to growth of grain structures regarding the final blooms.
dendrites thus inhibiting the formation of central equiaxed
crystals. To quantitatively describe the influence of second-
ary cooling mode on solidification structure of GCr15 steel Carbon macro-segregation evolution under
bloom, ECR and average grain diameter (Dm) are adopted as different solidification structure
quantify indexes. Dm is calculated as eq. (10) based on the characteristics
assumption that all dendrites are spherical.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
A The formation of carbon macro-segregation for GCr15 steel
Dm = 2 (10) bloom could be clarified in both macro- and microscale. In
ðN  π Þ
micro perspective, liquid steel solidifies as dendrites and
where A is the area of calculation, N is the nuclei num- the dendrite arm spacing determines the diffusion distance
ber in the calculation area, which can be obtained from of solute elements which causes the micro-segregation
CA–FE calculation. behavior of elements. Relative research [27–29] have
Figure 9(a) and (b) describes the variation of ECR and proved that minimizing SDAS would contribute to alleviat-
Dm at different secondary cooling modes. ing the micro-segregation degree of elements thus improv-
It could be known from Figure 9(a) and (b), when the ing the uniformity of element distribution in macroscale.
secondary cooling mode varies from super weak cooling In macroscale, strong and developed columnar dendrites
to super strong cooling, ECR drops from 35.4 % to would stimulate cracks between dendrite tips and cause
10.44 % while Dm enlarges from 1.56 mm to 1.99 mm, the occurrence of central segregation. Properly enlarging
correspondingly. the ECR would be useful in eliminating such defects.
748 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

(a) (b)

(c) (d)

(e)

Figure 8: Morphologies of solidification structure for GCr15 steel bloom at various secondary cooling modes (one-fourth cross section):
(a) super weak cooling, (b) weak cooling, (c) mild cooling, (d) strong cooling and (e) super strong cooling.

Figure 9: Equiaxed crystal ratios and average grain diameters of GCr15 steel bloom under different secondary cooling modes.
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 749

Figure 10: Evolution of SDAS for

GCr15 steel bloom under differ-
ent secondary cooling modes
(µm): (a) super weak cooling,
(b) weak cooling, (c) mild cool-
ing, (d) strong cooling and (e)
super strong cooling.

Figure 12: Evolution of SDAS at different sample points for GCr15

Figure 11: Sample points for SDAS analysis of GCr15 steel bloom. steel bloom.
750 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

Considering above facts, it is necessary to study the influ-

ence mechanism of solidification structure characteristics
on carbon macro-segregation.

Influence of secondary dendrite arm spacing (SDAS)

on carbon macro-segregation

Samples of GCr15 steel bloom are obtained for carbon

macro-segregation analysis. The sample positions areillu-
strated as in Figure 13. In the meantime, solidification
structure characteristics for GCr15 steel bloom under the
same casting condition are calculated using the model
established above.
Figure 14: Evolution of SDAS and carbon segregation index at
different locations.

27.94 µm and Kc is 0.95. At sample point C0 (130 mm

from bloom wide side), the SDAS value is 95.23 µm and
Kc is 1.24. Above all, increase of SDAS promotes the
diffusion distance and stimulates the nonuniform distri-
bution of carbon element. In the GCr15 steel continuous
casting process, proper secondary cooling mode could be
selected to control the growth of dendrite and refine
SDAS to improve bloom quality.

Influence of equiaxed crystal ratio on carbon

macro-segregation

In order to study the influence of ECR on carbon macro-

segregation behavior of GCr15 steel bloom, industrial
Figure 13: Sample points for carbon segregation and SDAS analysis.
tests are conducted to obtain bloom samples under five
different secondary cooling modes shown as Figure 7.
The samples are then etched by H2O–50 %HCl reagent
Carbon segregation index is defined as eq. (10). for structure observation and ECR estimation.
Kc = C
(10) Morphologies for central region of the bloom cross sec-
C*
tions under various secondary cooling modes are illu-
where C * is the average carbon content of the bloom, C is strated in Figure 15.
the local carbon content of the sample point. It could be clearly observed from Figure 15 that with
The evolution of SDAS and carbon segregation index the increase of secondary cooling intensity, the porosity
at various sample points from bloom side to center is and segregation in central region has become more
illustrated in Figure 14. severe and micro-cracks even occur under the strong
It could be observed from Figure 14 that the carbon cooling mode. These phenomena indicate that stronger
segregation index increases with the development of cooling intensity enlarges the area of columnar dendrite
SDAS from bloom wide side to center. At sample point and reduces the compactness degree of central region of
C20 (10 mm from bloom wide side), the SDAS value is the bloom.
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 751

Figure 15: Morphologies for

horizontal direction of GCr15
steel bloom under various sec-
ondary cooling modes (a) super
weak cooling, (b) weak cooling,
(c) mild cooling, (d) strong
cooling and (e) super strong
cooling.

On this basis, the ECR and the carbon segregation index at increasing to 1.18 under super strong cooling mode.
different locations (seen in Figure 16) of the above five Moreover, under certain cooling mode, Kc increases from
samples are measured as in Figures 17 and 18, respec- bloom side to center, which explains the formation of
tively. When the secondary cooling intensity transforms central segregation of solute elements.
from weak to strong mode, the measured ECR drops from Considering above results, it is evident that when the
33.2 % to 10.05 % and the carbon segregation index cooling intensity becomes stronger, the ECR decreases
decreases to some extent but then gradually increases. while Kc decreases first and then increases. Apparently,
For instance, at point C0, Kc under super weak cooling there exists the optimum secondary cooling mode in the
mode is 1.27 then it stabilizes at about 1.109 before transformation process of cooling intensity for the bloom
752 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation

Figure 18: Carbon segregation index at different positions.

Figure 16: Sample points for carbon segregation index analysis.

It could be clearly observed from Figure 19 that

the optimum secondary cooling water intensity is
between 0.26 L/kg and 0.30 L/kg, in which the GCr15
steel bloom would have relatively higher ECR, finer
grain as well as SDAS while the carbon segregation
degree is low. Above all, for GCr15 steel bloom studied
in this paper, the secondary cooling process should be
controlled within weak/mild cooling mode (secondary
cooling water intensity between 0.26 L/kg and 0.30 L/
kg) to keep low carbon macro-segregation degree and to
improve the comprehensive performance of final
products.

Figure 17: Measured results for equiaxed crystal ratio of GCr15

bloom.

to keep low-carbon segregation level while obtaining

relatively high ECR and finer grains.

Summary
In order to determine the optimum cooling mode for
improving the quality of GCr15 steel continuous casting
bloom, the evolution of ECR, GS, SDAS as well as car-
bon segregation index at typical sample position with
the change of secondary cooling intensity is summar- Figure 19: Summary of characteristic parameters under different
ized as in Figure 18. secondary cooling mode.
 K. Dou et al.: Influence of Secondary Cooling on Solidification and Segregation 753

Conclusions [5] R. Kobayashi, Physica D, 63 (3) (1993) 410–423.

[6] Y. Natsume, K. Ohsasa and T. Narita, Mater. Trans., 44 (5)
(2003) 819–823.
A coupled CA–FE model has been used to simulate the [7] J.L. Wang, F.M. Wang and C.R. Li, ISIJ Int., 50 (2) (2010)
solidification structure in continuous casting process of 222–230.
high-carbon GCr15 steel bloom. The present model is [8] Y. Natsume and K. Ohsasa, ISIJ Int., 54 (2) (2008)
validated by experimental data. This study simulated 415–421.
[9] Y.Z. Luo, J.M. Zhang, X.D. Wei et al., Ironmak. Steelmak, 39 (1)
the effect of secondary cooling intensity on the solidifica-
(2012) 26–30.
tion structure of GCr15 steel bloom. Besides, industrial [10] K. Dou, L. Wang, Q. Liu et al., Advances in the Science and
test is conducted to investigate different secondary cool- Engineering of Casting Solidification: An MPMD Symposium
ing water intensity on the macro-segregation of GCr15 Honoring Doru Michael Stefanescu, John Wiley & Sons, Inc.,
steel. Based on the CA–FE simulation and industrial Orlando, (2015), pp. 173–180.
[11] M. Rappaz, C.A. Gandin, J.L. Desbiolles et al., Metall. Mater.
test, the following conclusions can be drawn:
Trans. A, 27 (3) (1996) 695–705.
(1) It is found that the ECR and bloom compact degree [12] M. Rappaz, C.A. Gandin and R. Tintillier, Metall. Trans., 25 (3)
(GS, SDAS) changes in the opposite direction at (1994) 629–635.
different secondary cooling intensities. The ECR [13] N. Saunders and A. Peter Miodownik, CALPHAD (Calculation of
would be improved with weak cooling intensity. Phase Diagrams), A Comprehensive Guide, A Comprehensive
Intensive cooling intensities could improve the com- Guide, vol. 1, Elsevier, (1998).
[14] S.K. Choudhary and S. Ganguly, ISIJ Int., 47(12) (2007)
pactness of the bloom.
1759–1766.
(2) For the high-carbon GCr15 steel bloom in this [15] S.K. Choudhary and D. Mazumdar, Steel Res. Int., 66 (5) (1995)
research, the optimum secondary cooling water 199–205.
intensity is between 0.26 L/kg and 0.30 L/kg to [16] Z.B. Hou, F. Jiang and G.G. Cheng, ISIJ Int., 52 (7) (2012)
reduce the carbon macro-segregation and to 1301–1309.
[17] P. Thevoz, J.L. Desbiolles and M. Rappaz, Metall. Trans. A,
improve bloom compact degree in the continuous
20 (2) (1989) 311–322.
casting process. [18] W. Kurz, B. Giovanola and R. Trivedi, Acta. Metall., 34 (5) (1986)
823–830.
Funding: The authors would like to acknowledge the [19] W. Kurz and R. Trivedi, Acta. Metall. Mater., 38 (1)
financial support for this study provided by Key (1990) 1–17.
Research & Development Foundation (No. 41614014) [20] M.C.M. Cornelissen, Ironmak. Steelmak, 13 (4) (1986)
204–212.
and Independent Research & Development Foundation
[21] C.L. Jing, X.H. Wang and M. Jiang, Steel Res. Int., 82 (10) (2011)
(No. 41602023) of State Key Laboratory of Advanced 1173–1179.
Metallurgy, University of Science and Technology [22] E.A. Mizikar, Trans. Metall. Soc. AIME, 239 (11) (1967)
Beijing, China. 1747–1758.
[23] T. Takahashi, K. Ohsasa and N. Katayama, Tetsu-to-Hagané, 76
(5) (1990) 728–734.
[24] T. Takahashi and I. Hagiwara, J. Jpn. Inst. Met., 29 (1965)
References 1152–1159.
[25] M. Yamazaki, Y. Natsume, H. Harada et al., ISIJ Int., 46 (6)
[1] Z.G. Yang, B. Wang, Q. Liu et al., J. Iron Steel Res. Int., 21 (12) (2006) 903–908.
(2014) 1095–1103. [26] L. Nastac and N. El-Kaddah, J. Manuf. Sci. Prod., 15 (1)
[2] J.A. Spittle and S.G.R. Brown, Acta Metall. Mater., 37 (7) (1989) (2015) 3–12.
1803–1810. [27] M.C. Flemings, ISIJ Int., 40 (9) (2000) 833–841.
[3] M. Rappaz and C.A. Gandin, Acta Metall. Mater., 41 (2) (1993) [28] G. Lesoult, Mater. Sci. Eng. A, 413 (2005) 19–29.
345–360. [29] G. Guillemot, C.A. Gandin and H. Combeau, ISIJ Int., 46 (6)
[4] L. Nastac, Acta Mater., 47 (17) (1999) 4253–4262. (2006) 880–895.