Metallurgist, Vol. 43, Nos.
9-10, 1999
DESIGN OF MOLDS FOR E L E C T R O S L A G
REMELTING
K. N. Vdovin, A. A. Podosyan, UDC 669.154.9
M. I. Gorodetskii, and A. G. Alekseev
The metallurgical equipment repair shop at the Magnitogorsk Metallurgical Combine uses ESP-1.25 - 11 furnaces to
make rollers for continuous casters. The copper ingot molds of the crystallizing tanks (CMTs) undergo rapid wear during ser-
vice, which has made the method used to obtain the molds an important concern. To provide the shop with serviceable crys-
tallizing tanks, first the combine mastered a technology for refining copper by electroslag remelting (ESR) and using the cop-
per to make the CMT. A modification was subsequently proposed for the method used to calculate the thickness of the walls
of the CMT.
One method used to design CMTs and check their quality is based on the critical heat flux and the temperature of the
inside working surface of the ingot mold (which should not exceed 500~ [1]. The shop successfully introduced this method
and proposed an algorithm for designing CMTs that were to be made in the shop. However, the calculation of the critical heat
flux in the method in [1] is based on the sampling of a large number of process parameters, which complicates the design pro-
cess. As a result, well-known relations [2, 3] were used to simplify the heat-flux calculation in the design of CMTs.
The calculation and design are based on determination of the necessary dischm~e of coolant water and the temperature
on the inside surtace of the CMT. This temperature should be lower than the allowable value. The method used to pertbrm the
thermal calculations for the CMT is described below.
We determined the amount of heat released into the slag bath:
Q= GQt (1)
1 - (qs + q,-dn)'
where G is the productivity of the remelting operation, kg/h; Qt is the theoretical amount of heat needed to melt 1 kg of metal
(for most grades of steel, at the initial moment of time we take Q --- 1210 kJ/kg); qs is the fraction of the heat of the slag bath
which is lost through the wall of the CMT, this fraction depending on the ratio of the area of the contact surface between the
molten slag and the wall to the surface area of the metal bath [2]; Tlrdn is the fraction of heat radiation lost from the surface of
the slag (assumed to be equal to 0.06).
We determined the average heat flux reaching the wall of the CMT:
Qqs
qwa ~c.t ' (2)
where F f t is the area of contact of the slag with the wall of the CMT, m2:
Ff t = 3.14Dct h, (3)
where Dot is the working diameter of the CMT, m; h is the height of the slag bath, m.
We determined the ratio FS/Fmb (Fmb = 3.14D2t/4 is the area of the metal bath). We then used the method in [2] to
determine qs and we calculated Q from Eq. (1).
Magnitogorsk State Technical University; Magnitogorsk Metallurgical Combine. Translated from Metallurg, No. 9,
pp. 39-41, September, 1999.
402 0026-0894/99/0910-0402522.00 9 1999 Kluwer Academic/Plenum Publishers
Critical heat flux, qcr, MW/m2
3
5
4
3
2 2
0
0.1 0.3 0.5 0.7 0.9
V, n a J s c C
Fig. 1. Dependence of critical heat flux on water flow velocity v at different
water pressures: 1)2.5 atm; 2)2.0 atm; 3)1.5 atm; 4)1.0 atm; 5)0.5 atm.
We calculated the average heat flux in the water inside the cooling chamber:
qw = q w a F f t / F e f , (4)
where Fef is the effective cooling surface, m 2.
The shop uses crystallizing tanks with rectangular cooling channels, for which
Fef = ncl c ( 2 h f k f + b c ), (5)
where n c is the number of channels in the heat-transfer region; b c and Ic are the width and length of the cooling channels, respec-
tively, m; hf is the height of the fins on the channels, m; kf is a coefficient characterizing the efficiency of the fins (the value of
this coefficient is chosen from [2]).
We determined the maximum heat flux in the water:
qmax = kqw, (6)
where k is the safety factor (k = 1.3).
Heat removal during the cooling of the mold should occur in the nucleate boiling regime (the sheet boiling regime is
not allowable). In accordance with this, we compared the resulting value of heat flux with the critical heat flux (qcr), at which
nucleate boiling turns into sheet boiling. Safe operation of the mold is possible under the following condition (Fig. 1):
qmax < qcr,
where
qcr = 4.2' 104 -vl/2p 1/3 , (7)
where v is the water flow velocity, m/sec; p is the water pressure, Pa.
We determined the average temperature of the water in the mold twin':
twaV= twin +(At w/2), (8)
where twn is thewater temperature at the inlet of the mold, ~ At w is the temperature to which the water is heated in the mold, ~
At w = 0.75Q / cwG w . (9)
403
ill'
The subcooling temperature is found from the formula Ats = t c - t w [1], where c w is the heat capacity of water; G w is the
mass rate o f water flow for a specified flow velocity (determined from readings of special transducers or calculated), kg/h.
In the final step, we determined the temperature of the wall of the CMT on the melting-chamber side and compared it
with the maximum allowable temperature (500~ 9
~wa (10)
twa.mt = twWa+ qwa )~wa '
where 8wa is the minimum distance between the working wall and the cooling channel, mm; ~wa is the thermal conductivity of
the material of the mold (1.2 MJ/m-h.~ for copper); t wwa is
9 the average temperature of the wall on the water side, ~
w qw (11)
lwa = t w + ,
~w
where t w is the boiling point, ~ c~w is the coefficient of heat transfer to the water. During boiling
., , , 0.7~0.15
CZw =J'J~+qw rw 9 (12)
For walls with fins and channels, the heat-transfer coefficient is equal to [2]:
o~f = o~w [ ( 2 k f h f + fif') [(2hf + b c)]. (13)
The calculated temperature of the copper wall was multiplied by the safety factor k = 13 and compared with the max-
imum allowable temperature.
Hydraulic calculations were performed to obtain results for the case when there are no instrument readings, as well as
to refine the main dimensions and parameters o f the equipment. The calculations were performed in the following sequence:
9 water discharge was determined
G = v.F, (14)
where v is the water flow velocity, m/sec (the optimum velocity is 1.5-3.0 m/sec); F is the area o f the given section or the sum
of the sections o f parallel channels, m2;
9 water velocity in individual sections o f the system was determined:
v =G/Fci, (15)
where Fci is the cross-sectional area of a channel in the ith section of the system, m2; G is the discharge of water, m2]h;
9 pressure on a specific cooling section was determined as the sum of atmospheric pressure and the pressure losses in
all elements o f the cooling system from the given section to the outlet:
Pct = Z k'Pi + Patm , Pa; (16)
i=I
9 the pressure loss (resistance) was determined:
V2 . Z r ]'
ka~i - , (17)
2g
where 7 is the density of the water, kg]m3; g is the acceleration due to gravity, m/sec2; Z~ is the total resistance coefficient:
404
Z ~ = ~1 +~fr, (18)
where ~l is the local resistance coefficient of the given element of the tube or channel determined from graphs and tables [5];
~2~ is the frictional resistance coefficient in the given element of the tube or channel:
~fr = ~..1/d, (19)
where d is the equivalent diameter of the tube or channel, m; I is the length of the tube or channel, m; ~, is the frictional resis-
tance coefficient of a unit of the relative length of the given element of the tube or channel:
1
~,: (20)
1.82 l g R e - 1.64'
where Re is the Reynolds number during the movement of water in the given element of the tube or channel.
Values of local resistance on the diffuser section were calculated in accordance with [5]:
~d = k~out(in)" (21)
where ~out(in) is the resistance coefficient with a sudden change in cross section (corresponding to the velocity in the smaller
cross section) and was determined from graphs in [5]; k is a correction factor used in calculating the resistance of the diffuser
(k = 1 with an apex angle ~ > 40~
The value of the resistance coefficient when there is a abrupt change in the direction (reversal) of the flow was deter-
mined from the expression:
~a.rvr = 1.2B, (22)
The resistance coefficient for smooth reversals (reversal with both the inside and outside edges being curved and the
radius of curvature being greater on the outside edge than on the inside edge) was calculated from the formula:
~s rvr = ~0"B'C' (23)
where ~0 is the initial resistance coefficient for smooth reversals [5]; B is a correction factor for the resistance coefficient char-
acterizing reversals that depend on the angle of rotation, the value of this factor having been taken from [5]; C is a correction
factor for the resistance coefficient characterizing smooth reversals, which depends on the form of the cross section. In the case
of a circular or square cross section, C = 1.
The combine has used the above algorithm to design and build copper ingot molds for the crystallizing tanks of ESR
furnaces that make continuous-caster rollers.
The critical heat flux was calculated by the method in [1-3]. An analysis of the data showed that the method in [3]
gives more accurate results than the method in [1].
REFERENCES
1. B. E. Paton and B. E. Medovar (eds.), Electlvslag Furnaces [in Russian], Nauk. Dumka, Kiev (1976).
2. G. V. Samokhvalov and G. I. Chemysh, Electric Furnaces in Ferrous Metallurgy [in Russian], Metallurgiya, Moscow
(1984), pp. 106-110.
. A. V. Egorov and A. F. Morzhin, Electric Furnaces [in Russian], Metallurgiya, Moscow (1975), pp. 214-215.
4. A. I. Veinik, Technical Thermodynamics and the Principles of Heat Transfer [in Russian], Metallurgiya, Moscow
(1965), pp. 203-205.
5. S. S. Kutateladze and V. M. Borishanskii, Handbook of Heat Transfer [in Russian], Gosen6rgoizdat, Leningrad (1959),
pp. 183-186, 292-307.
405
