See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/361207880

Optimization of Aluminum Deoxidation Practice in the Ladle Furnace

Article in Iron & Steel Technology · June 2022

DOI: 10.33313/TR/0722

CITATIONS READS

2 422

2 authors:

Kanan Sahoo Ranjan Kumar Pathak

Visakhapatnam Steel Plant Engineers India Limited (EIL)
21 PUBLICATIONS 10 CITATIONS 22 PUBLICATIONS 11 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Kanan Sahoo on 10 June 2022.

The user has requested enhancement of the downloaded file.

162 AIST Transactions Vol. 19, No. 4 DOI 10.33313/TR/0722

Optimization of Aluminum Deoxidation Practice in the

Ladle Furnace
Authors Deoxidation is the key step for steel refining and generally aluminum is commonly used as
the strongest deoxidizer for liquid steel in the steel industry. However, in order to optimize
Kanan Kumar Sahoo
Assistant General Manager, WRM II its consumption and assess its efficiency under ladle metallurgy furnace (LMF) conditions, an
Operations, Visakhapatnam Steel Plant, effective deoxidation practice needs to be established in which the quantification of oxygen
India
ingress into the steel melt at each stage must be taken into account as well as the amount of
[email protected]
oxygen transfer from carryover slag to the steel. However, a negligible amount of oxygen is
Ranjan Kumar Pathak
Manager R&D Center, Visakhapatnam
introduced into steel via ferroalloys, which are ignored in the present investigation. The liquid
Steel Plant, India steel reoxidation in the ladle and tundish during casting is also not considered in the present
investigation. To achieve this objective, an investigation has been carried out at Visakhapatnam
Steel Plant (VSP) to modify Al addition pattern mainly for Al-Si-killed steel. VSP follows a complex
deoxidation practice, thus there is a competitive reaction between all the elements — Al, Si, Mn
and C — based on their degree of lowering the dissolved oxygen in the liquid steel, and this results
in a different share of oxygen by each element, which is calculated by the Gibb’s Free Energy
Minimization technique. Similarly, different models were developed to calculate reoxidation
during tapping and argon gas stirring. Due to the presence of carryover slag in the ladle, the
activity of oxygen is controlled by FeO and MnO due to the transfer of oxygen from slag into the
bulk of the liquid steel. Slag deoxidation by Al shots, therefore, poses its importance to reduce
the oxygen transfer. Finally, a deoxidation mechanism for Al-Si-killed steel was proposed and
a comparison between measured and calculated percentage Al opening at LMF was presented
during industrial trials. It was revealed from the study that by incorporating an Al addition pattern
as per the present technique, Al consumption reduced from 2.6 to 1.93 kg per ton of liquid steel.

D eoxidation of liquid steel is

one of the main processes dur-
ing steel refining that substantially
during tapping operation of liq-
uid steel by lowering the dissolved
oxygen in the steel melt which is
affect the metallurgical steel quality. measured by CELOX probe at end-
This process removes the dissolved of-blow liquid steel. Generally, Al
oxygen from the liquid steel via is used for lowering the oxygen
fixation of oxygen into stable oxides (single deoxidation) in Al-Si-killed
when deoxidizers are introduced steel along with alloying by bulk
into the steel melt. Aluminum, in addition of other ferroalloys while
steelmaking, is a powerful deoxi- Al-wire and trimming ferroalloys are
dizer as well as an alloying element. added at the ladle metallurgy fur-
As an oxidizing element, it ensures nace (LMF) refining operation. In
lower oxygen concentration in the this practice, deoxidation product
steel melt for subsequent smoother Al2O3 is in solid phase at steelmak-
continuous casting operation, and ing temperature. Visakhapatnam
I IRON & STEEL TECHNOLOGY I AIST.ORG

as an alloying element, it is used to Steel Plant (VSP) is also following a

increase the toughness of steel by similar practice of deoxidation and
controlling the grain size by form- alloying. Deoxidation practice in
ing AlN in the final product for which a deoxidizer such as Al and
Al-Si-killed steel.1 Excessive alumi- ferroalloys such as Si-Mn, Fe-Si and
num additions during tapping of coke are added almost simultane-
liquid steel increases the cost of ously leads to complex deoxidation
steel; therefore, in steel production process. As per K.V. Grigorovich et
practice, reduced aluminum content al.,3 in the complex deoxidation of
JUL 2022

is always beneficial.2 The deoxida- a metallic melt, the activities of the

tion practice generally commenced forming oxides are lower than unity,
 163

which makes it possible to produce a melt with a lower out by the researchers on this aspect. The estimated
oxygen concentration with the same content of the results are in agreement with the data available in
primary deoxidizer. Thus the present study uniquely the literature.6 Once the quantification of all possible
emphasizes the share of oxygen consumed by the sources of oxygen is over, it is quite easy to calculate
deoxidizing and alloying elements (Al, Si, Mn and C) and optimize the amount of Al required in different
added to the steel melt simultaneously. stages of refining.
In steelmaking operations, oxygen is in-filtered into The commercial aluminum consumed in VSP comes
the steel melt during various stages such as during in two forms: (a) Al bar and (b) Al wire. Al bar is typi-
tapping, in which oxygen is in-filtered into the liquid cally used for primary deoxidation purposes, particu-
steel stream from air during the LMF refining opera- larly during tapping, when the oxygen concentration
tion of liquid steel from the easily reducible oxides is on the order of 800–1,200 ppm. Al wire is injected
(FeO+MnO) of carryover slag (COS) and from plume in the steel melt for fine tuning the Al content of steel
eye during argon gas stirring at on-line rinsing station at the ladle furnace. Another form of Al, aluminum
(OLR) and during LMF refining of liquid steel. Fig. 1 lumps/shots/nuggets, has been recommended during
represents the possible sources of oxygen ingress into the present investigation to deoxidize the slag which
the steel melt during different stages of refining of generally contains the easily reducible oxides such as
steel. FeO and MnO.
Oxygen ingress in the metal stream during tap-
ping is mostly ignored in the previous studies4 as its
amount is not so significant and liquid steel already Industrial Trials
has a high concentration of oxygen. However, in the
present investigation for Al consumption optimiza- VSP has three LD converter and two ladle furnace
tion in liquid steel melt, the oxygen entrained from facilities each with a nominal capacity of 150 tons
the surrounding air into the liquid steel stream has of liquid steel. During the present investigation, a
been incorporated for better and accurate quantifica- total of 43 industrial trials were conducted. Once the
tion. COS contains high concentrations of reducible primary steelmaking by oxygen blowing of hot metal
oxides such as FeO and MnO, which are among the was over, CELOX probe was used to measure the dis-
major source for supplying oxygen from slag to steel, solved oxygen content and temperature of liquid steel
thereby reoxidizing the steel. As per Z. Deng et al., 5 while the slag and steel samples were also collected
wt.% FeO in the ladle slag in the range of 0.4–1.0 can at end-of-blow and on-line rinsing station (OLR) for
supply 21–63 ppm of oxygen at the steel-slag interface. further analysis and estimation of COS. The sampling
Thus the amount of COS tapped in the ladle is the procedure at the converter, OLR and LMF is shown in
important parameter that affects Al consumption. It Fig. 2. The arrows (↓) represent the time that a sample
is a general phenomenon that, as the taphole life of is taken.
an LD converter increases, the amount of COS also A two-phase industrial trial work has been under-
increases. Hence, the estimated amount of COS can taken in order to define the effect of the initial
be correlated with the taphole life for industrial pur- conditions of slag and liquid steel on the Al recovery.
pose. Apart from reoxidation from the easily reduc- In the first phase, a total of 13 Al-Si-killed, low- and
ible oxides present in COS, another major contribu- medium-carbon steel grade industrial trials were con-
tion of this paper is to estimate reoxidation due to ducted with the existing practice of no addition of Al
oxygen ingress from air into the plume eye on the lumps/shots/nuggets onto the top slag for reduction
top slag surface of the ladle during bottom argon gas of FeO and MnO, which is the standard practice of
stirring. However, very little work has been carried steel refining treatment at the ladle furnace process.
JUL 2022

Figure 1
I IRON & STEEL TECHNOLOGY I AIST.ORG

Oxygen transfer
Oxygen ingress in from carryover
Dissolved oxygen Reoxidation
the liquid steel slag to liquid
during blowing at during argon gas
stream during steel dissolved
BOF stirring
tapping oxygen during
blowing at BOF

Source of oxygen in the refining of steel from LD converter to ladle metallurgy furnace (LMF).
164 AIST Transactions
Figure 2
Coke, lime, Fe-Si, OLR
Si-Mn, Al LMF-Entry LMF-Middle LMF-Exit

EOB slag Avg tapping Ar(g) purging Slag and steel Slag and Slag and steel
CELOX time 6 min. for 4 min. sample, CELOX steel sample
sample (after Al and
(without additions) Ca-Si wire additions)

Sampling procedure at end of blow, on-line rinsing station (OLR) and ladle furnace.

Slag samples were taken along with measurement of transport ladle is popularly referred to as tapping.
oxygen activity as soon as the heat arrived at the LMF, Fig. 4 shows schematically the various phenomena tak-
then addition of 200–300 kg of lime and 150–200 kg ing place during tapping. A considerable amount of
synthetic slag was carried out. For correcting the Al nitrogen and oxygen find their way into molten steel
concentration in steel composition to meet the grade and also a significant drop in temperature occurs
requirement, Al in the form of Al wire was inserted as high as 70–100°C during the liquid steel tapping
into the liquid steel and the injection of calcium- operation. Deoxidation or “removal of oxygen” is gen-
silicide (Ca-Si) wire for inclusion modification com- erally facilitated by the addition of elements such as
menced. Thereafter at the end of the ladle furnace Al, Si, Mn and C, etc., either individually or in combi-
treatment, final steel and slag samples were taken nation, which have higher affinity toward oxygen than
and heat dispatched for continuous casting of billets. iron does. During tapping, air is entrained into the
Analysis of the first phase trial of 13 heats revealed liquid steel stream, causing reoxidation as shown in
that the oxygen transfer from COS to steel and oxy- Fig. 4. The stream of liquid steel is reoxidized during
gen ingress from air into the liquid steel during argon its fall under gravity from taphole of BOF to the ladle.
stirring was quite high, which resulted in lowering of The falling liquid steel stream at 1,873 K heats the
the Al content in steel. Thus in the modified trial of surrounding air up to 1,173 K, which is the boundary
30 heats of Al-Si-killed steel, lowering of the Al con-
tent was minimized by adding Al in the form of lumps/
shots/nuggets to slag for deoxidation and to take care
of oxygen ingress from air. All 13 heats’ data were ana- Figure 3
lyzed methodically to develop several models to opti-
mize the overall consumption of Al in steel starting
from converter tapping of liquid steel, where primary Mass
deoxidation was carried out with addition of Al bar, balance
and LMF refining of steel, where Al wire was added to module
meet grade requirement to achieve minimum of 0.02 [ppm O]
wt.% Al in Al-killed steel. The developed models were ingress Reoxidation
integrated as a global model and used for optimiz- due to Ar during tapping
gas stirring module
ing the overall Al consumption for the second phase, module Global model
which is known as modified practice, and consists of for optimization
industrial trials of 30 Al-Si-killed steel grades with of Al
I IRON & STEEL TECHNOLOGY I AIST.ORG

addition of Al lumps/shots/nuggets for slag deoxida- consumption Sharing of

tion as shown in Fig. 3. Subsequent sections discuss Slag [ppm O]
different model development and its application in deoxidation among Si,
optimizing Al consumption optimization. module Mn, Al and
module
Carryover
slag
Model Development module

Reoxidation Due to Air Ingress in the Liquid Steel Stream During

JUL 2022

Tapping — Transfer of molten steel from a primary steel-

making basic oxygen furnace (BOF) vessel to a steel Global model for Al consumption optimization.
 165

layer air film temperature. The pouring rate of liquid Rt = the radius of liquid steel stream which is assumed
steel is nothing but tonnage of steel tapped per unit to be equivalent to the radius of taphole (m)
time; in this case taking an average tapping time of 6 (refer to the Appendix for the relation between
minutes and tapping weight of liquid steel of 155–165 diameter of tapping stream of steel melt and
tons, the pouring rate and amount of oxygen in- taphole life),
filtered into the liquid steel stream during tapping H = the height of liquid steel stream exposed to atmo-
can be estimated as given in Eqs. 1 and 2:7 sphere (m) and
ϑk,a = the kinematic viscosity of air at 1,173 K.
m· = ηρsu 0At
Module for the Estimation of Activity of Slag Components —
(Eq. 1) The steelmaking slags are oxide solutions that obey
the Raoultaian law for its solid oxide as standard state.
 1
 Hence they can be modeled on the basis of regular
 ρa    ϑ k ,a H   ϑ k ,a H  2 
[ ppmO ]ingress
= 0.21 3
 ρs Rt2    u 0 
+ 3.44Rt 
 u 0  
and subregular solution models. In the present inves-
  tigation, the activity of metallic oxide Mx Oy in oxide
systems was estimated from Subregular Solution
(Eq. 2) Model as follows:

where x[M] + y[O] = (Mx Oy)

m· = the mass flowrate of tapping stream of steel (kg/ (Eq. 3)

second),
η = the liquid steel discharge coefficient to allow for a where
vena contracta effect, which results in overestima-
tion of area of the tapping stream and frictional [M] and [O] = the dissolved deoxidizing elements
losses within the nozzle, which is taken as a con- and soluble oxygen in the liquid iron, respectively,
stant value of 0.9 in the present investigation, x and y = the stoichiometric coefficients, respectively,
At = the cylindrical area of liquid steel stream during and
tapping, which is assumed to have a linear varia- Mx Oy = the oxide product formed from reaction
tion with taphole wear rate (m2), given in Eq. 3.
u0 = the velocity of liquid stream during tapping (m/
second), The equilibrium constant of the reaction given in
[ppmO]ingress = the amount of oxygen entrained into Eq. 3 can be expressed as given in Eq. 4:
the liquid steel stream (ppm),
aMO  ∆G M0 −O 
ρa and ρS = the density of air and liquid steel, respec- ( x y)
K M-O = = exp  − RT 
tively (kg/m3),
[hM ]x [hO ]
y

(Eq. 4)

Figure 4 where

KM–O = the equilibrium constant of reaction given in

Eq. 3,
a(Mx Oy) = the activity of Mx Oy in the oxide system
referred to pure solid standard state,
[hM] and [hO] = the Henrian activity of deoxidizing
JUL 2022

element M and soluble oxygen O at 1 wt.% stan-

dard state in liquid iron, respectively,
I IRON & STEEL TECHNOLOGY I AIST.ORG

R = the universal gas constant and

T = the absolute temperature.

 G exM O − G 0M O 
( x y) ( x y)
aMO =XMO  
( x y) ( x y)  RT 


Schematic diagram of tapping operation of liquid steel in the (Eq. 5)

ladle.
 166 AIST Transactions

where supersonic oxygen gas in the iron bath, a significant

amount of oxygen gets dissolved in the steel in the
X(Mx Oy) = the mole fraction of an oxide Mx Oy in the range of 600–1,400 ppm depending on its [C]*[O]
oxide system, equilibrium value. This dissolved oxygen is deoxi-
 G exM O − 0
 partial molar free energy of the
( x y ) G (Mx Oy )
= the excess dized/killed to get a lower oxygen concentration by
=XMO  
oxide in reference to the pure liquid state and adding deoxidizers and the liquid steel is alloyed
( x y)  free energy of phase transforma-
 G exM O − G 0M O RT = the molar by adding ferroalloys which are performed almost
( x y) ( x y)
 tion of the oxide from liquid to solid, simultaneously by adding ferroalloys such as Al,
 RT  Fe-Si, Si-Mn and coke in a preferred order within
In the present investigation G exM O −isGcalculated
0
 by the stipulated time of tapping operation. Thus there
( ) (M O )
=  
x y x y
using subregulara Msolution
X model as shown below. The will be sharing of dissolved oxygen in the liquid
( O) (M O )  ex RT 0 
  G M O (M = 1 to 3)
x y x y
excess partial mole free energy ( ) − Gx(O
M yO )
steel melt among the elements Al, Si, Mn and C,
=  a multi-
x y x y
can be expressed a Mwith aX set of formulae in where the major share is contributed by strongest
( O) (M O )  xRT y
 x y
component system, in which 1 to  3 denote three oxide deoxidizer, which is calculated on the basis of Gibb’s
compounds in a MnO-Al2O3 -SiO2 ternary oxide sys- Minimization Technique. The deoxidizer and ferroal-
tem, respectively.8–10 loy addition sequence followed at VSP is lime, petro-
j′ k′
coke with Fe-Si, Si-Mn and Al, where the deoxidation
G1ex = ∑ ∑ A jkY j Z k and alloying occurs simultaneously. Therefore, at any
2 0 instant of time the liquid steel that was tapped in the
ladle, all the deoxidizers, i.e., Al, Si, Mn and C, are
(Eq. 6) available for the dissolved oxygen present in the melt,
which results in a competitive deoxidation reaction
j′ k′  among the oxidizing elements present in the liquid
j′
A jk  j − k 
G 2ex = − ∑ + ∑ ∑  A jkY j Z k 1 +  steel melt. The standard Gibb’s free energy values for
2 (1 − j ) 2 0   Y (1 − j )  the deoxidation reactions are given in Table 1.11
However, during the stipulated time interval
(Eq. 7) between liquid steel tapping into the ladle and LMF
refining, the deoxidation reactions shown Table 1
j′ k′ A jk are not in the equilibrium, thus a change in Gibb’s
G 3ex = − ∑ ∑ + energy should be taken into account, which is given
2 0 (1 − j ) in Eqs. 10–12:
j′ k′   j − k  −k 
∑ ∑ A Y j
Z k 1 + +  a(Al2O3 )
 Y (1 − j )  YZ (1 − j ) 
jk
2 0  ∆G Al = ∆G Al0 + RT ln
( f Al ) ( fO ) [W Al ]2 [WO ]3
2 3

(Eq. 8)
(Eq. 10)
X 
Y = 1 − X 1 and Z = 1 −  2 
Y  a(SiO2 )
∆G Si = ∆G Si0 + RT ln
( f Si )( fO ) [WSi ][WO ]2
2
(Eq. 9)

where (Eq. 11)

j' and k' = the order of the subregular solution model

a(MnO )
and it is 3 in the present model and ∆G Mn = ∆G Mn
0
+ RT ln
I IRON & STEEL TECHNOLOGY I AIST.ORG

Y and Z = variables in the molar fractions. ( f Mn )( fO )[WMn ][WO ]

A set of Ajk parameters in the above equations can (Eq. 12)
be evaluated based on the known boundary condi-
tions. Once Ajk parameters of an oxide system are p(CO )
evaluated, the oxide activities in the oxide system can ∆GC = ∆GC0 + RT ln
be calculated with Eqs. 4–7 and Eq. 3. ( fC )( fO ) [WC ][WO ]
Sharing of Dissolved Oxygen Among the Added Oxidizing (Eq. 13)
JUL 2022

Elements — After the end of primary refining of carbon–

saturated liquid iron in the LD converter by impinging where
 167
Table 1
Deoxidation Reactions, Equilibrium Equations and Standard Gibbs Free
log f i = ∑ e ij [Wi ]
Energy Values Used in the Present Study
(Eq. 14)
Deoxidation reactions Keq equations Values Reference
a(Al2O3 ) where
2[Al] + 3[O] = (Al2O3) K eq = ∆G0Si = –594,230 + 229.73 * T 11
[h ] [h ]
2 3
Al O
fi = the activity coefficient of element i,
a(SiO2 )
K eq = e ji = the first order interaction parameter
[Si] + 2[O] = (SiO2) ∆G0Al = –1,201,860 + 323.22 * T 11
[h ][h ]
2
Si O and
a(MnO ) [Wi] = the wt.% of dissolved element i in
K eq = the liquid iron.
[Mn] + [O]
[h ][h ]
Mn O
∆G0Mn = –288,120 + 128.26 * T 11

p(CO ) Now by plugging all the values in Eqs.

[C] + [O] = CO K eq = ∆G0C = –48,323 + 27.374 * T 11
[h ][h ]
C O
10–13, the values of ∆GAl, ∆GSi and ∆GMnfor
the deoxidation reaction can be estimated.
Tohgeet al.17 reported that the dissolved
∆GAl, ∆GSi and ∆GMn = the Gibbs free energy change oxygen present in the liquid iron is distributed among
per mole of Al, Si and Mn as shown in the reac- the oxidizing elements Al, Si, Mn and C present in the
tion given in Table 1(kJ/mol), liquid proportional to the Gibbs free energies of their
∆GAl 0 , ∆G 0 and ∆G0 = the Gibbs free energy change
Si Mn oxidation reactions at the interface. From this consid-
per mole of Al, Si and Mn in the reaction for eration, the amount of shared quantity of oxygen hold
unmixed reactants and products at standard con- by each of the oxidizing element can be estimated as
ditions as given in Table 1(kJ/mol), mentioned in the relationship given in Eq. 15:
a(Al O ), a(SiO ) and a(MnO) = the activity of Al2O3, SiO2
2 3 2
νk ∆Gi
and MnO in the slag, respectively, ωi = ∗ [ ppmO ] ;0 < νk ≤ 1
fAl, fSi, fMn and fO = the activity coefficients of dissolved ∑ νk ∆Gi
Al, Si, Mn and O, respectively, and
[WAl], [WSi], [WMn] and [WO] = the concentration (Eq. 15)
of dissolved Al, Si, Mn and O in the liquid steel,
respectively (wt.%). where

The activity of oxides and present in the corre- ωi = the fractional share of oxygen used by the oxidiz-
sponding slag can be estimated from the previously ing elements Al, Si, Mn and C in the liquid steel,
discussed module of subregular solution model from νk = the reciprocal of the number of oxygen atoms
Eqs. 3–9. In the present study, the partial pressure participating in the reactions and
of CO gas (p CO) is taken as 1 atm. The activity coef- [ppmO] = the dissolved oxygen in terms of parts per
ficient of dissolved elements in the liquid iron can be million present in liquid iron.
estimated by Wagner’s interaction formalism for the
first order as given in Eq. 14: Estimation of Carryover Slag Tapped Into the ladle From LD
Converter — After the completion of tapping of liquid
steel into the steel transport ladle, subsequent refin-
ing operation starts at ladle furnace. The ladle top
slag at the start of LMF operation
generally contains burnt lime
Table 2 (600 kg during tapping addition),
First Order Interaction Coefficients (eij) of Elements Dissolved in Steel Melt Used Al2O3 formed during tapping
JUL 2022

in the Present Work12–16 by Al bars/ingots addition for

j→ deoxidation, ladle slag remained
I IRON & STEEL TECHNOLOGY I AIST.ORG

i↓ Al Si Mn O C from previous heat, carryover

slag (COS) from the BOF vessel
Al (63.0/T) + 0.011 0.056 0.0 (–34,740/T) + 11.95 0.091
in tapping operation and eroded
Si 0.058 (34.5/T)+0.08 –0.015 –0.119 (380.0/T)–0.023 refractory from the ladle walls.
Mn 0.0 0.033 0.0 –0.083 –0.054 For calculating the amount of
for [WC] < 0.5; eOC = –0.421 COS from the LD converter into
O (–20,000/T) + 7.15 –0.066 –0.021 (–1,750/T)+0.734 the steel transport ladle, Eq. 16
for [WC] >1.0; eOC = 0.1
can be used for estimating the
C 0.043 0.08 –0.012 –0.34 0.14 COS entrained in the ladle top
168 AIST Transactions

slag by knowing the fact that Al is being used as the MnO content present in the COS and the LMF slag
primary deoxidizer in Al-Si-killed steel:18 before start of refining treatment, as given in Eq. 18:

W SC = WSt – (Wf + WSAl2O 3 + W SL + W Sr )

2M Al (m FeO + MnO )cos − (m FeO + MnO )LMF in 
m Al −lumps =
(Eq. 16) 3M RO

where (Eq. 18)

WSt = the total weight of slag in the ladle (kg), where

Wf = the weight of burnt lime added during tapping
(kg), mAl–lumps = the amount of Al lumps/shots/nuggets
WSAl2O 3 = the weight of Al2O3 formed in the deoxida- required for the slag deoxidation (kg),
tion process (kg), MRO = the molecular mass of reducible oxide
W SL = the weight of slag remained from the previous (FeO+MnO) which can be approximate to molec-
heat (kg), ular mass of FeO (kg/mol) and
W SC = the weight of COS from the BOF vessel (kg) and 2M Al (m FeO + MnO )cos − (m FeO + MnO )LMF in = the quantity of reducible
W Sr = the weight of refractory eroded from m the ladle
Al −lumps = oxides (FeO + MnO) present in the COS and the
3M RO
lining (kg). LMF slag before start of refining treatment (kg).

Furthermore, weight of total slag in the ladle WSt Reoxidation From Atmosphere Due to Argon Gas Purging —
can be calculated by performing the Al balance During ladle furnace (LMF) operation, arcing is
between steel and slag as given in Eq. 17: done to increase the temperature in the ladle; some
ferroalloys may also be added to adjust the composi-
 (W Alt − W AlS ) M Al O  L (wt .%Al 2O3 )   100  tion of the liquid steel, but these activities result in a
WSt =  2 3
+ WS ∗ A
 ∗  
  ( 2 3 )t 
 2M Al  100 wt .%Al O 
change in temperature and composition locally. To
make this change throughout the ladle, it is important
(Eq. 17) to homogenize the steel bath, and this homogeniza-
tion is done by purging argon gas in the ladle from
where the bottom porous plug, as shown in the schematic
diagram Fig. 5.
WAlt = the amount of Al added to liquid steel during As the injected argon gas escapes to the surround-
tapping (kg), ings through the overlying slag, it leads to the for-
WAlS = the amount of Al that dissolved in the liquid mation of a “plume eye” or “slag eye.” The eye is a
steel by analyzing the initial sample of LMF potential site for steel reoxidation affecting phenom-
before the start of refining (kg), ena such as lowering of Al content in steel, thereby
MAl2O3 = the molecular mass of Al2O3 (kg/mol), affecting steel cleanliness. A relatively larger eye also
MAl = the atomic mass of Al (kg/mol), exacerbates heat loss to the surroundings, leading to
(wt.%Al2O3)A = the average content of Al2O3 in the noticeable drop in melt temperature during argon
slag after the completion of refining process at stirring.19 It is therefore desirable to know and control
LMF (wt.%) and
(wt.%Al2O3)t = the Al2O3 content in the slag after
deoxidation in the tapping process which is sam-
pled from OLR just after tapping (wt.%).
Figure 5

Since the weight of slag remaining from the pre-

I IRON & STEEL TECHNOLOGY I AIST.ORG

vious heat (W LS) is difficult to estimate, this value

was assumed as a constant mass of 100 kg for every
heat. Furthermore, the average weight of refractory
eroded in every heat, which was taken as the differ-
ence between the weights of newly refractory lined
ladle and the circulation life of ladle before the first
slag zone repair, is considered as an average of 651 kg.
Once the amount of COS in ladle is known, the quan-
tity of Al lumps/shots/nuggets estimation for slag
JUL 2022

deoxidation can be established based on the FeO and A schematic sectional view of two-plug argon-stirred ladle
system and its characteristic flow pattern.
 169

the extent of slag eye formation during industrial

 Q 1/3 H 1/4 
ladle-refining operations. The estimation of oxygen u p = 4.78  Ar 1/4 
ingress into the liquid steel in the LMF refining via  R 
plume eye is estimated from Eqs. 19–25:
(Eq. 23)
1
O2 ( g ) = [O ]1wt .%Fe In the present study, the mass transfer coeffi-
2
cient is estimated from Higbie’s penetration theory.23
(Eq. 19) However, for the bubble-agitated system, the esti-
mation of contact time is uncertain and as per the
6100 authors’ knowledge, there is no method available to
log K eq = + 0.3220
T estimate it. Hence, the contact time can be assumed
to be the ratio of diameter of vessel to velocity of
(Eq. 20) plume as:

where DO DO u p
km = 2 =2
πtc πR
[O]1 wt.%Fe = the dissolved oxygen in liquid iron at
1 wt.% standard state (wt.%) and (Eq. 24)
Keq = the equilibrium constant for the reaction given
in Eq. 19. where DO is the diffusion coefficient of oxygen atom in
the liquid iron (m2/second) and is the contact time (s).
Flux of oxygen ingress into the plume eye of liquid In the present investigation, the value of has been
steel in the ladle due to Ar(g) stirring can be calcu- adopted from Sujuki and Mori et al.24 Incorporating
lated as per Eq. 21: all the required parameters in the flux Eq. 21 and
assuming a negligible dissolved equilibrium oxygen
JO = Aekm(CO – COeq ) concentration in the steel melt, Eq. 24 is derived for
estimating the entrained oxygen from the surround-
(Eq. 21) ing air into the plume eye of the liquid steel melt.

where
d [ ppmO ]  K eq ρm PO2 DO u p 
= 1.13Ae  
JO = the flux of oxygen ingress (moles/second), dt  Wm R 
CO, COeq = the concentration of oxygen ingress and
equilibrium concentration in the liquid steel (Eq. 25)
respectively (moles/m3),
km = the mass transfer coefficient (m/second) and where
Ae = the plume eye area for bottom gas-stirring ladle
d [ ppmO ]  K eq ρm PO DO u p 
system (m2), which was estimated by Iguchi et ==1t.13
heAerate
 of change
2
 of
dissolved oxygen due
dt  Wm R 
al.21 as: to air entrainment into plume eye (ppm/
minute) and
  2g (ρm − ρs ) H s   PO2 = the partial pressure of oxygen gas present in
Ae = H m2 0.6 − 0.652   atmosphere (Pa).
  ρm u p2  
The average estimated value of oxygen concen-
(Eq. 22) tration calculated by present model is 77.5 ppm as
JUL 2022

compared to the measured value of 58.5 ppm for the

where measured conditions of a 151-ton LD converter with
I IRON & STEEL TECHNOLOGY I AIST.ORG

argon gas flowrate of 107 Nm3/minute as given by

Hm and Hs = the height of liquid steel and liquid slag Peter et al.6 It can be observed that there is a reason-
in the ladle, respectively (m), able agreement between the present developed model
g = the acceleration due to gravity (m/second2), and the measured data reported in the literature.
ρm and ρs = the density of liquid steel and slag, respec-
tively (kg/m3), and
up = the average velocity of liquid steel in the plume,
which can be estimated from Mazumdar et al.22 as:
170 AIST Transactions

Results and Discussion 65.4 and 9, respectively. The order of sharing percent-
age of oxygen among the oxidizing elements, Al, Si
The deoxidation practice followed at Visakhapatnam and Mn, is Al > Si > Mn, thereby corroborating with
Steel Plant is complex in nature due to the addition thermodynamic evidence of the reaction based on
of C, Si, Mn and Al simultaneously, which results in the oxygen potential of individual elements. However,
a competitive reaction according to their Gibb’s free the oxidizing element C forms CO(g) as the reaction
energy. Thus the concept of sharing of oxygen by each product at the tapping temperature, which has a
of the elements present in the liquid steel is evolved. faster reaction rate, thereby increasing its percentage
The percentage share of deoxidizers with respect to share of oxygen as comparable to percentage share of
dissolved oxygen present in the liquid steel during oxygen by Si. It can also be observed from Fig. 6a and
the present investigation is shown in Fig. 6a and 6b 6b that, in some of the low-carbon Al-killed (LCAK)
for both existing and modified practice. From Fig. 6a grade steels heats, the percentage share of oxygen by
and 6b it is evident that the average of maximum and the oxidizing element C is 0 due to the fact that all the
minimum shared percentage of dissolved oxygen in oxidizing elements Al, Si and Mn were added except
the liquid steel is deoxidized by Al and Mn, which are that of petrocoke during tapping of liquid steel.
From Fig. 6a and 6b it can also be
observed that there is no significant
variation in the sharing of oxygen by
Figure 6 Al as deoxidizer and other deoxidiz-
ing elements present in the ferroalloys
such as Si and Mn both for existing and
modified practice. It is understood that
once the share amount of dissolved
oxygen taken by the deoxidizer and
the oxidizing element present in the
alloying element is known, the quantity
of deoxidation products such as Al2O3,
SiO2 and MnO in the ladle top slag can
be estimated accurately. Therefore, in
the present investigation, the amount
of Al used in a heat changes from
250–300 kg in the existing practice to
140–160 kg in the modified practice
in order to fulfill the purpose of Al
as deoxidizer only during the tapping
operation.
(a) In the existing practice, the addition
practice during tapping of liquid steel
into the ladle includes the addition of
250–300 kg of Al in the form of Al bar,
150–200 kg of Fe-Si and 1,800–2,000 kg
of Si-Mn, while the addition of petro­
coke depends on the grades of steel,
i.e., for low carbon (0–200 kg) and
medium carbon (400–600 kg). The
recovery of Al in the liquid steel melt
I IRON & STEEL TECHNOLOGY I AIST.ORG

can be defined from the mass balance

as follows:

 m Al
oxdn
+ m[WAl ] + m Al
loss

R Al =  tot 
 m Al 

(b) (Eq. 26)

JUL 2022

Percentage share of oxygen taken by each deoxidizer in complex deoxidation where

for existing practice (a) and for modified practice (b).
 171
Figure 7
RAl = the recovery of Al in the steel
melt (%),
oxdn = the mass of Al used for deoxi-
mAl
dation in the steel melt (kg),
m[WAl] = the mass concentration of Al in
the steel melt (kg),
loss = the amount of Al lost in reac-
mAl
tions other than for deoxidation
and alloying (kg) and
tot = the total amount of Al added
mAl
to the steel transport ladle (kg)
at the stages of tapping and LMF
refining.

The oxygen entrainment in the liq-

uid steel stream from the surrounding
air during tapping operation is esti-
mated from Eq. 2 of the reoxidation
model. It was found from Fig. 7 that,
as the liquid steel stream diameter Amount of oxygen ingress into the liquid steel stream as a function of taphole
increases due to increase in the taphole sleeve life and tapping time.
sleeve life, less oxygen ingresses into
the tapping stream of liquid steel due
to the decrease in tapping time. The amount of oxy- results in a way to establish a relationship between the
gen ingress into the liquid steel ranges from 12.5 to amounts of COS and taphole sleeve life.
23 ppm with the mean quantity of 16.5 ppm oxygen, During the tapping of liquid steel, all the deoxi-
which is negligible as compared to dissolved oxygen dizer and ferroalloys are added almost simultaneously
present in the steel and it only consumes approxi- within the stipulated tapping time into the ladle to
mately 2.8 kg of Al dissolved in the steel. The variation deoxidize and alloying the liquid steel. In the exist-
in the quantity of oxygen ingress into the liquid steel ing practice, 250–300 kg of Al bars are added during
stream is due to the variation in contact time of liquid the tapping operation of liquid steel irrespective of
steel stream and surrounding air during the tapping amount of end-of-blow (EOB) dissolved oxygen con-
operation, i.e., tapping time, which corresponds to centration and COS in the ladle. However, at the EOB
the taphole sleeve life of LD converter
as illustrated in Fig. 7.
During tapping of liquid steel into
the steel transport ladle, a significant Figure 8
quantity of COS is always tapped into
the ladle. However, VSP uses a DART-
plug slag arrester to promote slag-free
tapping, but frequent failure leads to
high COS. The average quantity of
COS as estimated by the COS model
as given in Eq. 16 during the pres-
ent investigation was 1,304 kg with
minimum of 31 kg and maximum of
JUL 2022

2,500 kg. However, the amount of COS

depends on the taphole life and the
I IRON & STEEL TECHNOLOGY I AIST.ORG

manual intervention due to insertion

of DART-plug to restrict the slag flow
from LD converter into the steel trans-
port ladle during the tapping process.
From Fig. 8, it can be observed that, as
the taphole sleeve life increases, the
amount of COS also increases, which

Amount of carryover slage (COS) as a function of taphole life.

172 AIST Transactions
Figure 9
condition, the dissolved oxygen for
existing and modified practice is in the
range of 800–1,200 ppm. The share of
oxygen deoxidizes by Al from the pres-
ent model estimation is 65.4%, which
requires 88.3–132.4 kg of Al for deoxi-
dation and about 15 kg Al for alloying
to achieve 0.007–0.010 wt.% Al in the
liquid steel at LMF entry, which are the
average wt.% Al contents as sampled by
both the data of existing and modified
practice for the maximum and average
quantity of COS which is estimated by
the present model as 2,500 kg and 1,304
kg, respectively. In order to achieve the
aforementioned deoxidation and alloy-
ing, only 104–147 kg of Al is required.
Therefore, the differential amount of
147–153 kg of Al is an unnecessary addi-
Relation between %Al in LMF in sample and quantity of COS.
tion which results in material loss when
250–300 kg of Al is used during tap-
ping. The concentration of Al in liquid
steel at LMF entry must be restricted to
maximum of 0.01 wt.%. Increasing the Al concentra- On the basis of present oxygen share model of
tion more than 0.01 wt.% will reoxidize to form Al2O3 oxidizing elements Al, Si, Mn, C, reoxidation model
when argon gas is stirred to homogenize the liquid during tapping and from the COS model, the concen-
steel thermally. The problem associated with the tration of Al in liquid steel at LMF entry samples are
existing practice is that more than 50% of the added predicted and compared with actual data, which is
amount of Al is not taking part in any useful reaction. illustrated in Fig. 10 as given in Eqs. 27–29 in the form
Therefore, the existing practice of adding Al is being of simple mass balance.
modified by adding 120–150 kg of Al during tapping It can be observed from Fig. 10 that the variation
instead of 250–300 kg to minimize the losses in terms in the wt.% Al in the liquid steel at LMF entry for the
of Al2O3 formation, keeping the amount of other existing practice (red diamond) is similar as com-
deoxidizer unchanged. As the variation of COS is very pared to modified practice (blue diamond), which
high, i.e., from 31 kg to 2,500 kg, addition of higher confirms that higher Al bar addition during tapping
quantity of Al bar during tapping can-
not ensure higher concentration of Al
at the LMF entry steel sample. However,
in some cases of existing practice, wt.% Figure 10
Al in liquid steel at LMF entry sample is
high, i.e., 0.020–0.025, which is further
reduced to 0.01 wt.% or even less due to
argon gas stirring and interaction with
easily reducible oxides, i.e., FeO and
MnO of COS, as shown in Fig. 9, which
gives the relation between Al concen-
I IRON & STEEL TECHNOLOGY I AIST.ORG

tration at LMF entry and the amount of

COS. However, in the case of the modi-
fied practice, the amount of Al used for
deoxidizing and alloying is optimum;
hence, losses regarding formation of
Al2O3 are reduced to 2.6–16.7 kg of
Al. This ensures concentration of Al in
liquid steel in the range of 0.007–0.010
wt.% with a lesser amount of reoxida-
JUL 2022

tion during argon stirring.

wt.%Al in liquid steel at LMF entry (actual vs. predicted).
 173
Equations 27–29

[WAl]actual = [WAl]LMF Entry

(Eq. 27)

[WAl]predicted = [WAl]Al bar – [WAl]deox – [WAl]loss due to reduction of RO in COS

(Eq. 28)

 2M Al ω Al [ ppmO ]t 10 −6Wsteel WSt 2M Al  WSC 

m
 Al bar − − (wt .%FeO ) EOB t
− (wt .%FeO )LMF Entry 
η × 100  3MO 100 3M RO  WS 
[W ]
predicted
=  
 + (wt .%MnO ) WS + M Mn ω Mn [ ppmO ]t 10 Wsteel − (wt .%MnO )
Al
Wsteel  C −4
 
LMF Entry 
  EOB
W t
M 
 
 S O 

(Eq. 29)

where

[WAl]actual and [WAl]predicted = the actual and predicted value of Al content in the liquid steel (wt.%),
[WAl]LMF Entry = the Al content at LMF before start of the refining operation (wt.%),
[WAl]Al bar = the content of Al in the liquid steel due to Al bar addition,
[WAl]deox = the Al content liquid steel used for deoxidation (wt.%),
[WAl]loss due to reduction of RO in COS = the Al content in the liquid steel used for the reduction of reducible oxides
(RO = FeO + MnO) present in COS,
η = the efficiency of Al due to oxidation, volatilization which depends on the method of addition into the
steel melt (refer to Table 3) (%),
Wsteel = the weight of liquid steel (kg),
mAl bar = the mass of Al bar added during the tapping operation,
MAl, MMn, MRO = the atomic and molecular mass of Al, Mn and reducible oxide (RO) (kg/mol),
[ppmO]t = the total quantity of dissolved oxygen in the liquid steel at EOB and due to oxygen ingress from
air into the liquid steel stream (ppm),
(wt.%FeO)EOB and (wt.%MnO)EOB = the FeO and MnO content at EOB slag (wt.%),
ωAl and ωMn = the % share of oxygen by Al and Mn during competitive deoxidation (%), and
(wt.%FeO)LMF Entry and (wt.%MnO)LMF Entry = the content of FeO and MnO at LMF before start of the refin-
ing operation (wt.%).

Figure 11 does not always guarantee a higher wt.% Al in the

initial steel sample at LMF. In addition to that, the
increasing Al bar addition during tapping will reduce
the Al bar recovery significantly due to oxidizing and
volatilization losses as shown in Fig. 11.
JUL 2022

Generally, the composition of COS is the same as

that of EOB slag of LD converter, which contains a
I IRON & STEEL TECHNOLOGY I AIST.ORG

significant quantity of wt.% FeO and wt.% MnO as

easily reducible oxides (RO). These reducible oxides
will react with dissolved Al present in the liquid steel
and oxidize it according to the Eqs. 30 and 31 as slag-
metal reaction.

3FeO + 2[Al] → Al2O3 + 3[Fe]

Amount of Al bar added during tapping vs. its recovery. (Eq. 30)
174 AIST Transactions

3MnO + 2[Al] → Al2O3 + 3[Mn] During LMF operation, argon gas stirring is done
from two off-centered segmental porous plugs from
(Eq. 31) the bottom in order to homogenize the composition
and temperature throughout the liquid steel bath.
In the present investigation, FeO is of a significant In this process, reoxidation takes place through the
amount as compared to MnO in the COS due to the plume eyes formed on the liquid steel surface due to
steel grades during study being of lower Mn content. argon gas stirring and the atmospheric air present in
Generally, the amount of reducible oxides in the ladle the ambient. The amount of oxygen ingress into liq-
slag needs to be neutralized and brought down to less uid steel can be estimated from the present developed
than 4 wt.% (FeO+MnO) as they continuously feed reoxidation model as given in Eq. 25. In the present
oxygen to the steel at the slag-metal interface. In order investigation, the average value of oxygen ingress into
to achieve this neutralization, Al lumps/shots/nug- the liquid steel during LMF refining comes out to be
gets are added in the modified practice
to ladle top slag layer after 4 minutes
of start of arcing at the LMF, as this
will ensure there is a liquid slag. It is Figure 12
always be beneficial to add Al lumps/
shots/nuggets to the slag layer as early
as possible probably at the OLR station
just after tapping to reduce the FeO
and MnO content of the ladle top slag.
However, this concept is not applicable
at VSP as there is a frequent case of on-
line argon gas purging failure, which
leads to a viscous ladle top slag, result-
ing in non-mixing of Al lumps/shots/
nuggets with slag layer and added Al
lumps/shots/nuggets being unutilized
and burnt in the air, forming Al2O3.
After 4 minutes of arcing at the LMF,
the slag liquefies and adequately mixes
with Al lumps/shots/nuggets. A higher
amount of COS results in a higher
amount of FeO in the ladle top slag,
i.e., 16 wt.%, as observed during the
present investigation. This is one of Amount of Al lump required as a function of %FeO in OLR slag.
the reasons for lowering of Al con-
tent in the liquid steel during LMF
refining. The required amount of Al
Figure 13
lumps/shots/nuggets for slag deoxida-
tion can be predicted by using Eq. 18,
which is presented in Fig. 12. It can
be observed that the Al lumps/shots/
nuggets requirement increases with an
increase in FeO content of the COS,
resulting in a 100 kg Al lumps/shots/
I IRON & STEEL TECHNOLOGY I AIST.ORG

nuggets requirement for 14 wt.% of

FeO. On average, 72 kg of Al lumps/
shots/nuggets are used per heat for slag
deoxidation for the modified practice
trial of 30 heats. This slag deoxidation
practice takes care of reducible oxides
(FeO and MnO) in the COS, thereby
protecting the soluble Al from deoxi-
dizing in between the LMF exit and
JUL 2022

tundish. The amount of oxygen ingress as a function of argon gas flowrate and
processing time.
 175
Table 3 77.5 ppm, as shown in Fig. 13. It can be observed from
Efficiency of Al Recovery for Different Types of Additions Fig. 13 that as the processing time at LMF increases,
Types of Al addition η the amount of oxygen entrainment increases for a
Al-wire 98 98 similar argon flowrate. On the other hand, for a high-
er argon gas flowrate, the amount of oxygen ingress
Al bar 95 95
into the liquid steel will be on the higher side.
Al-lumps/shorts 90 90 Tables 4 and 5 show the output of the different
developed models in the present investigation and the
comparison between existing and modified operating

Table 4
Output of the Parameters of the Models Developed and Assumed Values
Sl No. Model Unit Model results
1. (a) Total dissolved oxygen shared by Al % 65.4
(b) Total dissolved oxygen shared by Si % 14.2
(c) Total dissolved oxygen shared by C % 11.3
(d) Total dissolved oxygen shared by Mn % 9.0
2. Average oxygen ingress from air into the liquid steel stream during tapping ppm 16.5
Average oxygen ingress from air into the liquid steel surface during argon gas stirring with both the plugs at the rate of
3. ppm 13.7
50 Nm3/hour at on-line argon rinsing station (OLR)
4. Average oxygen ingress from air into the liquid steel surface during argon gas stirring with both the plugs at LMF ppm 77.5
5. Assumed efficiency of Al due to oxidation, volatilization which depends on the method of addition into the steel melt (η) % 95

Table 5
Operating Parameters Observed in Existing and Modified Practice
Operating parameters Unit Existing practice Modified practice
Dissolved oxygen content at EOB ppm 718 876
Average Al bar addition during tapping kg 293 150
Average reoxidation/oxygen ingress from air due to tapping stream ppm 17.17 15.67
Average wt.% Al content at LMF entry steel sample wt.% 0.014 0.016
Average dissolved oxygen content of steel at LMF entry ppm 12.9 11.0
Average amount of COS kg 1,799 1,089
Average content of RO (FeO + MnO) in the LMF entry slag sample wt.% 7.5 6.4
Average Al lumps/shots/nuggets addition for slag deoxidation kg 0 62
Average content of RO (FeO + MnO) in the LMF exit slag sample wt.% 5.2 4.7
Average Al wire addition to meet grade requirement Kg 93 88
Average content of Al in LMF exit steel sample wt.% 0.023 0.035
Average arcing time per heat min 17.6 20.0
JUL 2022

Average processing time at LMF min 40 40

Average content of Al in tundish sample wt.% 0.013 0.031
I IRON & STEEL TECHNOLOGY I AIST.ORG

0.01 0.004
Average content of Al that reoxidized between LMF exit and tundish steel sample wt.%
(0.023–0.013) (0.035–0.031)
Total amount of Al consumption per heat kg 386 (293+0+93) 300 (150+62+88)
Al consumption per ton of liquid steel kg/TLS 2.6 2.0
Average Al saving per heat kg/TLS — 86
176 AIST Transactions

practice with the model calculated values which shows reducible oxides (FeO+MnO) are reduced by
that, by the application of the global Al optimization adding Al lumps/shots/nuggets as a slag deoxi-
model, on average 86 kg of Al can be saved per pro- dizer. It was found that the average amount of
cessed heat, thereby reducing down the specific Al Al lumps/shots/nuggets required for slag deox-
consumption from 2.6 to 2 kg per ton of liquid steel. idation for such a high COS requires 62 kg.
e. The content of Al at the LMF entry is predicted
and compared with actual LMF entry data,
Conclusions which shows a fair agreement between the two.
f. Oxygen ingress during argon gas stirring at
Al is a strong deoxidizer, thus its optimization is the LMF is also quantified in the present study,
quite complex in the steel refining process due to which is mostly ignored in the previous studies
multiple sources of oxygen ingress into the liquid available in the literature. It was found that the
steel. The only way to optimize Al consumption is to average estimated oxygen ingress due to argon
quantify all oxygen sources. In this study, an attempt gas stirring is 77.5 ppm with a processing time
has been made to quantify all the sources of oxygen of 40 minutes at the LMF.
by developing mathematical models. These models
are developed on the basis of 13 heats’ data collected
as existing practice and the outcome of these models References
are applied on the subsequent 30 heats processed with
slag deoxidation by addition of Al lumps/shots/nug- 1. L.A. Erasmus, J. Iron Steel Inst., Vol. 202, 1964, pp. 32–41.
gets. The models were discussed and the findings are 2. G. Stolte, Secondary Metallurgy, Fundamentals, Processes and Applications,
illustrated as follows: Verlag Stahleisen GmbH, Düsseldorf, Germany, 2002, p. 146.
3. K.V. Grigorovich and A.K. Garber, Russian Metallurgy (Metally), No. 9, 2011,
p. 934.
a. VSP follows a complex deoxidation practice.
4. A.N. Conejo and D.E. Hernandez, Materials and Manufacturing Processes,
The share of soluble oxygen present in the Vol. 21, 2006, pp. 796–803.
steel melt deoxidized by Al is estimated along 5. Z. Deng and M. Zhu, ISIJ International, Vol. 54, No. 7, 2014, pp. 1498–1506.
with other elements such as Si, Mn and C based 6. J. Peter, K.D. Peaslee, D.G.C. Robertson and B.G. Thomas, AISTech 2005
on the Gibbs free energy minimization due Conference Proceedings, 2005, p. 972.
to the simultaneous addition of Al and Fe-Si, 7. J. Szekely, Fluid Flow Phenomena in Metals Processing, Academic Press
Fe-Mn and coke. Al has the maximum share Inc., 1979, p. 105.
as it is the strongest deoxidizer, followed by Si, 8. X. Zhang, G. Jiang, K. Tang, W. Ding and D. Xu, Calphad, Vol. 21, No. 3,
1997, pp. 301–309.
C and Mn, which are 65.4, 14.2, 11.3 and 9%,
9. X. Zhang, G. Jiang and K. Xu, Calphad, Vol. 21, No. 3, 1996, pp. 311–320.
respectively.
10. X. Zhang, G. Jiang and K. Xu, The International Conference on Molten Slags,
b. Reoxidation from the tapping stream is also Fluxes and Salts, Sydney, Australia, January 1997, pp. 259–262.
considered in this study, which is mostly 11. A. Ghosh, Secondary Steelmaking, Principles and Applications, CRC Press,
ignored in the previous studies. The average New York, N.Y., USA, 2000, pp. 105–146.
reoxidation comes out as 16.5 ppm. This reoxi- 12. J.F. Elliott, M. Gleiser and V. Ramakrishna, Thermochemistry for Steelmaking,
dation amount is directly proportional to the Addison-Wesley, Reading, MA, Vol. 2, 1963.
tapping time and as the tapping time increases, 13. G.K. Sigworth and J.F. Elliott, Metal Sci., Vol. 8, 1974, pp. 298–310.
oxygen ingress into the stream of steel melt is 14. The Japan Society for the Promotion of Science, Steelmaking Data
Sourcebook, Gordon & Breach Science Publishers, New York, N.Y., USA,
also likely to increase. 1988, pp. 280–291.
c. During tapping in the existing practice, an 15. N.H. El-Kaddah and D.G.C. Robertson, Metall. Trans. B, Vol. 8B, 1977,
addition of 250–300 kg of Al bar found to be pp. 569–579.
very high with low recovery will result in a high- 16. H. Ohta and H. Suito, Metall. Trans. B, 1996, Vol. 27B, pp. 264–270.
er amount of losses. Therefore, the consump- 17. T. Tohge, Y. Fujita, and T. Watanabe, Proc. 4th Process Technology Conf.,
tion of Al addition has been reduced in the Iron & Steel Society, Chicago, Ill., USA, 1984, pp. 129–36.
I IRON & STEEL TECHNOLOGY I AIST.ORG

modified practice to 140–160 kg to increase Al 18. R. OnurAydem, M.S. thesis, Middle East Technical University, 2007, p. 46.
recovery in the liquid steel. 19. D. Mazumdar and J.W. Evans, Modeling of Steelmaking Processes, CRC
Press, 2010, p. 368.
d. In the present study, COS is also estimated,
20. O. Kubaschewski, E. Evans and C. Alcock, Metallurgical Thermochemistry,
which is the main source of reducible oxides 4th ed., 1967, Pergamon Press, London.
(FeO+MnO) present in the ladle top slag. It is 21. M. Iguchi, K. Miyamoto, S. Yamashita, D. Iguchi and M. Zeze, ISIJ
a continuous source to supply oxygen to the International, Vol. 44, No. 3, 2004, pp. 636–638.
steel melt and is thereby responsible for lower- 22. D. Mazumdar and R.I.L. Guthrie, Met. Trans., Vol. 15B, 1985, p. 83.
ing the Al content present in the steel melt. 23. R. Higbie, Trans. Am. Inst. Chem. Eng., Vol. 35, 1935, pp. 36–60.
The maximum and minimum amounts of COS 24. K. Sujuki and K. Mori, Tetsu-to-Hagane, Vol. 57, 1971, pp. 2219–2229.
JUL 2022

found are 3,777 and 31 kg, respectively, with an

average of 1,304 kg. In the present study, these
 177

Appendix Dtapping stream = 1.011 × 10 –3Ltaphole

• The internal diameter of taphole sleeve = Di = (Eq. A-3)

129 mm
• The external diameter of taphole sleeve = DO = where Dtapping stream is the diameter of the steel melt
220 mm during tapping (mm) and Ltaphole is the life of taphole
i.e., the number of heats taken by the taphole before
Considering the average life of the taphole sleeve = changing to a new sleeve. ✦
90 heats and the erosion rate of the sleeve is uniform
with taphole life, the variation of diameter as a func- This paper was presented at AISTech 2021 — The Iron & Steel Technology
Conference and Exposition, Nashville, Tenn., USA, and published in the AISTech
tion of taphole life can be calculated by assuming a 2021 Conference Proceedings.
linear variation as:

 DO − Di 
Dtapping stream =  Ltaphole
 Total life of taphole 

(Eq. A-1)

 220 × 10 −3 − 129 × 10 −3 
Dtapping stream =   Ltaphole
 90
(Eq. A-2)

JUL 2022
I IRON & STEEL TECHNOLOGY I AIST.ORG
178

AIST Transactions AIST Transactions seeks to publish original research articles focusing on
Key Reviewers scientific or technical areas relevant to the iron and steel community. All
articles are peer-reviewed by Key Reviewers prior to publication.

The scope of the Transactions spans all fields of iron and steel manufactur-
ing, from extractive metallurgy to liquid metal processing, casting, thermo-
Andre Assis mechanical processing as well as coating, joining/welding and machining.
Vallourec Star Articles related to structure and properties, i.e., characterization, phase
transition, chemical analysis and testing of creep, corrosion or strength are
accepted. In addition, papers on process control, testing and product perfor-
mance are also encouraged.
S. Thomas Britt
Carnegie Mellon
University The Transactions will publish papers on basic scientific work as well as
applied industrial research work. Papers discussing issues related to specific
products or systems can be considered as long as the structure follows a
scientific investigation/discussion, but not for the aim of advertising or for
Emmanuel De Moor the purpose of introducing a new product. Papers should be no longer than
Colorado School of 5,000 words, or 12 pages including figures.
Mines
Papers will either be accepted unconditionally, conditionally based on
reviewers’ comments, or rejected. The review process should be completed
in less than two months. Upon acceptance, authors will be asked to complete
a form to transfer copyright to AIST.
Neslihan Dogan
McMaster University
A PDF proof of the accepted manuscript will be sent electronically to the
author(s) approximately one month prior to publication for final approval.
Reprints of published manuscripts can be ordered through AIST.

Megha Jampani Please contact me with any questions regarding AIST Transactions.
Linde Inc.

Il Sohn
Neo-Metallurgical Process-
ing Lab, Dept. of Materials
Science and Engineering, P. Chris Pistorius
Yonsei University POSCO Professor of Materials Science and Engineering
Co-Director, Center for Iron and Steelmaking Research
Carnegie Mellon University, 5000 Forbes Ave., Pittsburgh, PA 15213 USA
Phone: +1.412.268.7248
[email protected]

Dai Tang
Nucor Steel–Decatur LLC To submit a manuscript for consideration, visit:
aist.org/publications-advertising/iron-steel-technology/aist-transactions/submit-a-manuscript

Ming Tang
I IRON & STEEL TECHNOLOGY I AIST.ORG

ArcelorMittal Global AIST Transactions are indexed through Chemical Abstracts Service,
R&D – East Chicago
Columbus, Ohio, USA.
JUL 2022

View publication stats