Solidification Analysis in Continuous Casting Process - Barman Tambunan
Solidification Analysis in Continuous Casting Process - Barman Tambunan
Solidification Analysis in Continuous Casting Process - Barman Tambunan
Barman Tambunan
ABSTRACT
Continuous Casting is a manufacturing process for producing billets, slabs and flat
products. It is used for casting non-ferrous and steel production to produce high quality
products of slab or billet cast at reduced cost. A thin solidified metal shell is formed
initially during continuous casting due to the intense cooling from the water cooled mould.
Once the cast leaves the mould it goes through various cooling processes and in solid state
will proceed to the further forming processes. To exploit the benefit of continuous casting
process it is essential to develop close control of the production process. This is achieved
through a good understanding of the influence of the various compositional and
operational variables. The objective of the mathematical analysis presented here is to
model the heat transfer during continuous casting of low melting point Bismuth alloy cast
billets material for predicting the surface temperature. Good quality of the cast in
conjunction with the surface quality of the cast is relied on the solidification temperature
during continuous casting. In fact by simulating the solidification temperature of the
continuous casting process through applying different operational casting parameters, the
result are crucial for identifying the best operational casting parameters.
33
Berkala MIPA, 15(3), September 2005.
Thin solidified metal shell is formed parameters have been considered during
due to the intense cooling from the water simulation to determine solidification
cooled mould. Inside the shell large molten temperature of the strand and to predict the
core remain and solidify during the shell thickness of the billet from the
withdrawal of the strand. The operating meniscus region throughout the mould exit
conditions for the mould during casting are in continuous casting process.
the most critical for the whole process for
producing a crack-free surface cast. Several
parameters controls the shell development in 2. HEAT TRANSFER PROCESS IN
the mould such as the temperature of the THE MOULD
incoming hot liquid material, the dimensions
of the mould, the oscillating mould, mould The first continuous casting analysis
lubrication, the cooling water flow rate and concerns the heat transfer solidification of
the speed of the solid strand. The quality of the molten metal in the mould region. During
the slab casting production depends on the operation of the casting various variable can
solidification phase during casting process. influence the solidification of the product
Therefore, by developing a mathematical such as the incoming metal temperature,
heat transfer model of the solidification molten properties, mould geometry, mould
process, one may identify the optimal lubricant and the mould cooling system. The
operational parameters during continuous formation of the casting shell is shown in
casting to obtain a high quality slab. Fig. 1. V m is the mould oscillation speed, Vs
In order to gain further insight into is the casting speed and CL is the centre line
continuous casting it is important to of the mould geometry since the profile of
understand the solidification process which the mould is always symmetry. As depicted
is strongly influenced by the temperature and in Fig. 1, the Meniscus zone where the shell
flow of molten metal. This can only be begins to form the liquid metal is separated
obtained by carefully study and pinpoint by the lubricant layers. Three different
various operational casting parameters that phases of metal cast exist in the mould zone
can affect the quality of the continuous during solidification process of the cast
casting product. Most of theoretical consists of solid, liquid and mushy zone
modeling available (Takeuchi, et.al, 1991; (Huang, et.al., 1992; Shy, et.al., 1992). The
Takeuchi and Brimacombe, 1984) have been mushy zone is the solid-liquid interface
compared either with data from a where part of the metal becomes solid and
commercial production process using part of them is liquid. Casting factors such as
sophisticated equipment or using water flow heat extraction process accompanying by the
to verify their analysis (Sheng and Jonnson, solidification has shown an important role to
2000). Economically, it is not feasible to perform a good quality cast product. The
obtain a wide range of data from an casting design and operational parameter
industrial continuous casting to build a better during continuous casting at the meniscus
understanding of the process. To identify mould region is important since the surface
parameters that improve the process require of the cast is initiated and built in this region
changes in production parameters, addition (Mahapatra, et.al., 1991).
of new sensors or measurement devices to a Heat from the liquid metal is
commercial casting process. This may cause extracted to the mould trough various heat
interruption in production and flow sinks consisted of conduction,
inconveniences for the organization. convection and radiation heat transfer,
The present study is concerned with initiated from the liquid pool to the mould-
the review of mathematical model and water cooling interface. To help prevent the
development of numerical simulation of heat strand sticking to the surface of mould,
transfer solidification of the molten low lubricant is poured from the top of the mould
melting point Bismuth alloy in order to and flows into the gap with the help of the
predict the surface temperature of the billet oscillation motion (Darle, et.al., 1993).
cast product. Different operational casting Samarasekara et al., 1978 have reported that
34
Barman Tambunan, Solidification Analysis…
in steel continuous casting, the heat transfer strand. Inside the mould liquid steel
at the flux channel constitutes the largest solidifies against mould surface controlled
contribution to the heat extraction in the by the water cooling mechanism at the back
mould while the mould wall has the lowest of the surface of the mould. In addition
contribution. Thus heat removal in the casting speed and the geometry of the mould
continuous casting mould is dependent also influence the solidification process
largely to the gap between the solidified during the formation of the shell. The steel
strand and the water cooled copper mould. shell must be thick and strong enough to
contain the molten steel during cast
withdrawal.
y
Mould Lubricant layer
CL
x
Molten Metal
Vm
Shell
Mushy
Vs
35
Berkala MIPA, 15(3), September 2005.
36
Barman Tambunan, Solidification Analysis…
casting has been reported by Huang et al., 1980; Cliff and Dain, 1967) the density ρ(T)
1992. Thomas et al., 1997 have investigated is assumed constant in this analysis The
numerically the heat transfer and temperature density is less dependent on temperature
distribution in the solidifying shell of steel including in the region where the phase
slab using the 1-D transient heat conduction transformation occur. The specific heat cp (T)
equation by assuming negligible axial heat and the conductivity k(T) of bismuth alloy at
conduction. Mizikar, 1967 described a a range of temperature from the liquid phase
mathematical model of heat transfer to the solid state are strongly varied. The
solidification in continuous casting mould temperature dependent thermophysical
and predicted the slab shell thickness during properties are the subject of further
casting. His approach was to solve a 1-D discussion in the next section.
unsteady state heat conduction equation by When solidification is occurring, the
using an “artificially” high effective thermal associated latent heat L constitutes a heat
conductivity and a “simplify” effective source which is taken care by the internal
specific heat versus temperature relationship heat generated term q * . This heat source term
over a wide range of temperature up to the q * as indicated by Laitinen and
liquid region. Neittaanmaki, 1988 can be obtained from the
From Mizikar, 1967 analysis it is time rate of change of solid fraction f s during
evident that he has assumed a constant value solidification which is expressed as,
for the latent heat release in the mushy zone.
In this work a constant value for the latent ∂f ∂T (2)
q * = ρ L s
heat release in the mushy zone is applied to ∂T ∂t
enhance more accurate prediction of the
strand thickness in the mould region of where the parameter L is the latent heat of
continuous casting process. fusion and f s denotes the fraction of solid.
The fraction of solid f s = f s (T) describes how
4.1 Theory of Thermal Analysis the solid phase fraction varies with
temperature T. The release of latent heat is
Due to the latent heat release during difficult to describe mathematically because
solidification the heat transfer equation is of the complexity of solidification in multi
non-linear. The basic equation for two component alloy system. However, it was
dimensional nonlinear heat conduction assumed in this model that no convective
equations in an isotropic medium can be mass and energy transport in the liquid phase
written in the form and mushy zone was considered. The liquids
and the solidus temperatures were defined at
∂T
div (k (T )∇T ) = ρ (T )c p (T ) − q* (1) an exact range of temperature to permit
∂t solidification process at the mushy region to
be specified on the basis of point
where ?(T) is the density of the medium, ∇ is temperature.
the Laplace differential operator, cp (T)
denotes the specific heat, T is the 4.2 The Solid Fraction Relationship
temperature which is a function of position
and time, and k(T) is the thermal In order to solve Eq. (2) the fraction
conductivity. Furthermore q* denotes the of solid f s and temperature T relationship
internal heat generated term involves the must be defined. In practice, the solid
energy removed due to the latent heat of fraction f s versus temperature curve depend
fusion during solidification in continuous on local solute redistribution and this
casting process. The internal heat generated relationship exhibit a strong non-linearity as
term q * is equal to zero everywhere except in indicated by Clyne , 1984. Different methods
the mushy region where the latent heat is have been introduced to describe the fraction
released. As has been reported by many of solid and temperature relationship (Chen
workers (Mizikar, 1967; Davies and Shin, and Tsai, 1990). On equilibrium solidify-
37
Berkala MIPA, 15(3), September 2005.
cation it is assumed that that complete 4.4 Material Property Data of Bismuth
diffusion occur between the solid and liquid Based Alloy Material
states during solidification. In non-
equilibrium solidification the liquid is A low melting point Bismuth based
assumed to reach complete diffusion while alloy, called Wood’s metal (Handbook of
diffusion in the solid state is negligible. The Industrial Materials, 1992) was chosen as the
easiest way to descript this relationship is by casting material to suit our laboratory
assuming that latent heat is released linearly facilities. The melting point temperature of
between the solidus and liquids temperature. the Wood’s metal is 71 °C. During
continuous casting Wood’s metal have
4.3 The Enthalpy Formulation successfully been used to cast a billet shape
material. It was noticed that Wood’s metal
A popular variation for solving Eq. 1 satisfies the condition of a low melting point
with the latent heat release during temperature of 71 °C. This is regarded as the
solidification has been reported by Lally et proper material for conducting the
al., 1999; Voller and Swaminathan, 1991; experimental work. Properties of the bismuth
and Voller and Peng, 1994. They have based alloy are shown in Table 1 (Igor, et.al.,
suggested including the latent heat as an 1997).
enthalpy source term and then considering
the latent heat as a dependent variable which Tabel 1. Thermal Properties of Bismuth
is known as the enthalpy method. By Eq. 1 Based Alloy
and Eq. 2 we get the heat transfer model with No. Properties Parameter
the latent heat and fraction of solid variables, 1. Thermal Conductivity Conductivity
∂f ∂T
at Various Temperature (W/m °K)
ρ cp (T ) − L s = div(k (T )∇T ) (7) ( °K)
∂T ∂t
4 4
The relationship between the enthalpy and 10 11
temperature the is defined by 20 18
40 20
80 23
H = c p (T )T + ∆H , 0 ≤ ∆H ≤ L (8) 100 24
2 Specific Heat at Specific Heat
where H is the total enthalpy, cp (T)T is the Temperature (°K) (J/gr °K)
sensible heat and ?H is the latent heat. At the 278 – 423 0.147
solid and liquid interface the total Enthalpy
H is determined as, The thermal conductivity of bismuth
alloy in continuous casting is temperature
∂f s
T T
H = ∫ ρ c dT − ∫ ρ L dT (9) dependent and its value varies according to
T0 T0
∂T the material composition and thermal
The heat transfer equation in terms of history. Representations of material
enthalpy after applying the chain rule to properties data as a function of elevated
?H/?t can be written as temperature are one of the most severely
limiting constrains for a successful
∂H ∂ T
ρ = div(k (T )∇ T ) (10) simulation and usually these data are hard to
∂T ∂t find or sometimes conflicting. The specific
The sensible heat H1 in Eq. 8 can be obtained heat per unit mass cp is required to solve the
by integration the specific heat while to enthalpy formulation. The latent heat has
determine the heat latent heat H2 the fraction been treated using an apparent
of solid and temperature relationship must be enthalpy/specific heat method where the heat
defined which has been described in generation term has been eliminated. During
previous section. this transformation the latent heat fusion and
the sensible heat (specific heat time’s
temperature) are released simultaneously. It
38
Barman Tambunan, Solidification Analysis…
is thus necessary to treat the sensible heat metal, the mould heat flux as a function of
and latent heat of solidification mould dwell time is required to be
simultaneously which has been included in established. This function covers a wide
the enthalpy formulation. range of casting conditions for different
mould lubricants. This function is applied as
4.5 Enthalpy - Temperature the heat flux boundary condition for the
Relationship finite element analysis. The lack of boundary
condition information will undermine the
An enthalpy-temperature relation- ability of any previous or current analysis to
ship is used to simulate the temperature predict the solidification temperature
distribution of the solidification cast. To variation during the process of billet
define the enthalpy of the material at solidification.
different range of temperature and account
the latent heat during solidification the
Billet Top
enthalpy which has units of heat/volume is
determined by the integral of density times
specific heat with respect to temperature 16.5
defined in Eq. 10. mm
Symmetry
Line
Symmetry Line
Billet Cast
400 mm
16.5 mm
5. SOLUTION METHODOLOGY Heat Flux
Boundary
Conditions
This section present the solution
methodology of the heat transfer
solidification temperature of the billet during
2 Dimensional Modeled
continuous cast Wood’s metal with the finite Transverse Slice
element analysis simulation. The details of
the Wood’s metal billet model simulated Billet Bottom
using the finite element method is shown in
Fig. 4. 33
During billet cast withdrawal, heat
from the billet surface is conducted to the Figure 4. Transverse slice area of the Wood’s metal
surface copper mould’s wall. From the round billet cast model to be analyzed
experimental work, it was noted that the
temperature decreases sharply when the The boundary condition for the
liquid metal is poured into the mould. The mould cooling zone requires a knowledge of
sudden drop of temperature is hindered after the longitudinal variation of heat flux in the
the temperature reaches the melting point mould, preferably, as a function of time.
temperature of the Wood’s metal. The Unfortunately, insufficient data from
temperature decreases in the mushy region at literatures are available to formulate a mould
a lower rate in comparison to the liquid zone heat flux boundary condition. This has
region. During billet withdrawal, the heat been also reported by Savage and Pritchard,
from the billet surface is conducted to 1954 in their analysis for continuous casting
surface copper mould’s wall. The process of steel. Therefore, the mould heat
temperature fall when the cooling water is flux function for continuous casting of
circulated around the mould wall which will Wood’s metal is calculated from
take place in about 20 second depends on the measurement of mould wall temperatures
casting withdrawal speed. Outside the mould using embedded thermocouples as shown in
region, the radiation will enhance the Fig. 5.
completion of billet solidification. It involves the placement of pairs of
To apply the solidification analysis closely spaced thermocouple pairs at each
for the continuous casting process of Wood’s measurement position in the mould wall. At
each position, the two temperature
39
Berkala MIPA, 15(3), September 2005.
measurements are used to calculate the local Tabel 2 . Input data for finite element
heat flux with the integrated form of analysis of Wood’s metal billet
Fouries’s heat conduction equation (Holman,
1976) given as,
No. Parameters Values Unit
qm = mw (T1 − T2 )
k (11) 1. Wood’s Metal Igor et al., 1997
Lt
Parameters and Bejan, 1993
Conductivity (Kw) 12.8 W/m.°C
where kmw is the copper mould thermal Specific Heat (cp ) 147 J/Kg.°C
conductivity, Lt , is the distance between two Latent Heat ( Lf ) 32567 J/Kg
thermocouples, T1 and T2 are the two Density (?) 1056 Kg/m3
measured temperatures. The thermal Fraction of Solid (Fs ) Linear -
conductivity of the commercial copper Pouring Temperature 77 °C
material at 20 °C is 372 W/m °K. By taking ( Tp )
the value of the heat flux, q, at each position Melting point 69-71 °C
across the length of the mould, a longitudinal temperature ( Tm )
heat flux profile can be generated.
2. Casting Parameter Data from
Experimental Rig
Mould Top
Billet Diameter ( D ) 33 mm
Withdrawal time ( t ) 35 sec
Mould Length (Lm ) 400 mm
Length between 0.01 m
Thermocouple (Lt )
Thermocouple 0.2 °C (Fig.
Temperature (T1 – T2 ) 6.15)
Casting Speed (Vc) 3.7 mm/sec
Hot Face
Cold Face
40
Barman Tambunan, Solidification Analysis…
Temperature ( C )
progressed towards the middle of the billet. 70.5
71.8 °C for time t is 20 seconds, which is Distance from the middle of the billet (mm)
REFERENCES
41
Berkala MIPA, 15(3), September 2005.
Chen J. H., Tsai H. L., 1990, “Comparison gical Transactions B, Vol. 21B,
on Different Modes of Latent Heat August, pp. 761-770.
Release for Modelling Casting Mahapatra R. B. Brimacombe J. K.
Solidification”, AFS Transac-tions , Samarasekara I. V., Walker N.,
pp. 539-546. Paterson E. A., Young J. D., 1991,
Cliff K., Dain R. J., 1967, “Computerized “Mold Behavior and Its Influence on
Calculation of Operating Conditions Quality in the Continuous Casting of
in Continuous Casting Machines”, Steel Slabs: Part 1. Industrial Trials,
Journal of The Iron and Steel Mold Temperature Measurements, and
Institute , pp. 278-284, Mathematical Modelling”, Metallur-
Clyne T.W., 1984, “Modelling Heat Flow in gical Transactions B, Vol. 22B, pp.
Solidification”, Mat. Sci. and Eng., 861-874.
Vol. 65, pp. 111-124. Mizikar, E. A., 1967, “Mathematical Heat
Darle T., Mouchette A., Nadif M., Roscini Transfer Model for Solidification of
M., Salvadori D., 1993, “Hydraulic Continuously Cast Steel Slabs”,
Oscillation Of The CC Slab Mold At Trans. Met. Soc., AIME, Vol.
Sollac Florange: First Industrial 239(11), , pp. 1747 – 1753.
Results, Future Developments”, McDavies R. M., Thomas B. G., 1996, “Flux
Steelmaking Conference and Thermal Behavior of the Top
Proceedings , pp. 209-218. Surface Flux/Powder Layers in
Davies G. J., Shin Y. K., 1980, Continuous Casting Molds”,
“Solidification cracking of continuous Metallurgical and Materials
casting steels”, Solidification Transactions B, Vol. 27B, pp. 672-
Technology in the foundry and cast 685,
house, University of Warwic k, Najjar F. M., Thomas B. G., Hershey D. E.,
Conventry, pp. 517-523. 1995, “Numerical Study of Steady
Handbook of Industrial Materials, 1992, 2nd Turbulent Flow through Bifurcated
Edition, Elsevier Advanced Nozzles in Continuous Casting”,
Technology, Oxford, UK. Metallurgical and Materials Tran-
Holman J. P., 1976, Heat Transfer, Fourth sactions B , Vol. 26B, pp. 749-765.
Edition, McGraw-Hill Book Company, Samarasekara, I. V., Brimacombe, J.K.,
New York, 1978, “The Continuous Casting
Huang X., Thomas B. G., Najjar F. M., 1992, Mould”, Int. Metals Review, Vol. 6,
“Modelling Superheat Removal during pp. 286-300.
Continuous Casting of Steel Slabs”, Savage, J., Pritchard, W. H., 1954, ‘The
Metallurgical Transactions B, Vol. Problem of Rupture of the Billet in the
23B, pp. 339-356. Continuous Casting of Steel’, Journal
Huang X., Thomas B. G., Najjar F. M., 1992, of Iron and Steel Institute, Vol. 178,
“Modelling Superheat Removal during pp. 269 – 277.
Continuous Casting of Steel Slabs”, Sheng D. Y., Jonsson L., 2000, “Two-Fluid
Metallurgical Transactions B, Vol. Simulation on the Mixed Convection
23B, 339-356. Flow Pattern in a Noniso-thermal
Igor S. Grigoriev, Evgenii Z., Meilikhov, Water Model of Continuous Casting
1997, Handbook of Physical Tundish”, Metallurgical and
Quantities, CRC Press, Boca Rotan. Materials Transactions B, Vol. 31B,
Laitinen E., Neittaanmaki P., 1988, “On pp. 867-875.
Numerical Simulation of the Shy, W., Pang Y., Hunter G. B., Wei D. Y.,
continuous casting process”, Journal Chen M. H., 1992, “Modeling of
of Engineering Mathematics , Vol. Turbulent Transport and Solidification
22, pp. 335-354, During Continuous Ingot Casting”,
Lally B., Biegler L. Henein H., 1999, “Finite Int. J. Heat Mass Transfer, Vol. 35,
Difference Heat-Transfer Modelling No. 5, pp. 1229-1245.
for Continuous Casting”, Metallur-
42
Barman Tambunan, Solidification Analysis…
Takeuchi S. Miki Y., Itoyama S., Kobayashi Voller V. R., Swaminathan C. R., 1991,
K., Sorimachi K., Sakuraya T., 1991, “General Source-Based Method for
“Control of Oscillation Mark Solidification Phase Change”,
Formation during Continuous Numerical Heat Transfer, Part B,
Casting”, Steelmaking Conference Vol. 19, pp. 175-189.
Proceedings , pp. 73 – 33, Voller V. R., Peng S., 1994, “An enthalpy
Takeuchi E., Brimacombe J. K., 1984, “ The formulation based on an arbitrary
Formation of Oscillation Marks in the deforming mesh for solution of the
Continuous Casting of Steel Slabs”, Stefan problem”, Computational
Metallurgical Transactions B, Vol. Mechanics, Vol. 14, pp. 492-502.
15B, pp. 493-509.
Thomas B. G., Lui D., Ho B.,1997, “Effect
of Transverse Depressions and
Oscillation Marks on Heat Transfer in
the Continuous Casting Mould”,
Sensor and Modelling in Materials
Processing: Technique and
Applications , S. Viswanathan, R.G.
Reddy, and J. C. Malas, eds,. The
Minerals, Metals, & Materials Society,
Warrendale, PA.
43