JOM, Vol. 66, No.

9, 2014
DOI: 10.1007/s11837-014-1092-y
Ó 2014 The Minerals, Metals & Materials Society

Review of Innovative Energy Savings Technology

for the Electric Arc Furnace

BAEK LEE1 and IL SOHN2,3

1.—POSCO Research Laboratories, Gwangyang Works, Gwangyang 545-090, Korea. 2.—Mate-

rials Science and Engineering, Yonsei University, Seoul 120-749, Korea . 3.—e-mail: ilsohn@
yonsei.ac.kr

A review of the energy innovations for the electric arc furnace (EAF) steel-
making route is discussed. Preheating of scrap using vertical and horizontal
shafts that have been commercially successful in lowering the energy con-
sumption to as much as 90 kWh/t reaching almost the operational limit to
heating input scrap materials into the EAF is discussed. Bucket-type and
twin-shell preheaters have also shown to be effective in lowering the overall
power consumption by 60 kWh/t, but these have been less effective than the
vertical shaft-type preheaters. Beyond the scrap preheating technologies, the
utilization of waste heat of the slags from the laboratory scale to the pilot scale
has shown possible implementation of a granulation and subsequent heat
exchange with forced air for energy recovery from the hot slags. Novel tech-
niques to increase metal recovery have shown that laboratory-scale testing of
localized Fe concentration into the primary spinel crystals was possible
allowing the separation of an Fe-rich crystal from an Fe-depleted amorphous
phase. A possible future process for converting the thermal energy of the CO/
CO2 off-gases from the EAF into chemical energy was introduced.

steelmaker. However, the EAF process route, which

INTRODUCTION
typically uses scrap as its input material and elec-
According to the World Steel Association data1 on tricity for melting and heating, is less capital
steel production in 2012, approximately 1.5 billion intensive and is more flexible in materials utiliza-
tons of steel were produced globally, and this rate tion, making the EAF a viable alternative that can
was expected to surpass 1.6 billion tons in 2013, of overcome the shortfalls of the integrated steel route.
which 50% is accounted for by China. The produc- The EAF process has become increasingly cost and
tion amount of steel through the electric arc furnace quality competitive to the integrated steel mills
(EAF) route has surpassed more than 400 million through process and technology innovations, which
tons and is expected to be approximately 650 million have significantly lowered power consumption and
tons by 2020,2 as depicted in Fig. 1a. This would be increased productivity while satisfying the custom-
approximately 30% of the world’s steel supply ers quality needs of steels.3–5 According to Fig. 2, the
according to Fig. 1b and continues to be an impor- power consumption during the 1970s was approxi-
tant sector of the steel industry, as higher scrap mately 600 kWh/t-steel, but today’s highly efficient
recycling and increased energy efficiency in the EAF EAF furnaces consume roughly 300 kWh/t-steel.
is realized. The tap-to-tap time, which determines the produc-
Although the integrated steel route continues to tivity of the EAF, has decreased from 180 min to
be one of the most efficient furnaces currently being 40 min over comparable sized heats in the past.
used in any materials processing in terms of energy Overall, recent advances in technology and its
utilization and processing costs, its initial capital implementation to the production process have sig-
investment, high CO2 emissions per ton of steel, and nificantly improved the efficiency and productivity of
lower flexibility with materials input and operations the furnace, making the EAF a cost-effective option
brings about significant challenges to the continued than the integrated steel route. In addition, as
expansion and investments to the integrated greenhouse gas regulations become stricter and

(Published online August 20, 2014) 1581

1582 Lee and Sohn

carbon tax globally mandated,6–8 the EAF process, Fig. 3. According to the estimates, internal scrap
which emits indirect CO2 through electrical con- supply from the Chinese steel industry will total
sumption, is considered environmentally more approximately 40 million tons and obsolete scrap
advantageous than the integrated steel route. will exceed 160 million tons by 2020.
Furthermore, the expected steel scrap available As technology advances continue to support the
in the near future seems to indicate at least for the EAF route, higher-value-added steels such as
low-quality scraps that the supply is more than automotive grade steels are being produced in the
sufficient to satisfy the EAF demands. In particular, EAF,9–11 and with the expansion of DRI facilities
the scrap exports from China are expected to be using shale gases,12–15 the product range of the EAF
pronounced from 2020 and will likely stabilize the is expected to expand and compete with the inte-
scrap market for the near future, as depicted in grated steel mills.
In this study, some of the recent advances and
technology improvements in energy minimization
(a) and productivity enhancements for the EAF steel-
making route have been reviewed. The various
energy savings realized through process changes,
implementation of innovative technologies, and
increased efficiency are discussed. EAF develop-
ments for the near future and novel process meth-
ods for the EAF steel route is also addressed.

EFFECT ON ENERGY SAVINGS THROUGH

PREHEATING IN THE EAF
For the typical EAF process, the amount of energy
input and output stream balance is shown in Fig. 4.
To produce 1 ton of steel in the EAF at 1873 K
(b) (1600°C) from steel scrap at 298 K (25°C), approxi-
mately 700 kWh/t is needed of which electrical
power input comprises more than 60%. This
approximates to 640 kg-CO2/t-steel emissions com-
pared to the 2000 kg-CO2/t-steel for the integrated
steel mills.16,17 Much of the input heat is trans-
ferred to the steel, but a significant amount of heat
is lost through the exhaust gas and the slag, which
comprises more than 28% of the input energy. Thus,
Fig. 1. (a) Current and predicted EAF production according to the improvements in the energy efficiency of the EAF
WSD. (b) Global steel production percent through the EAF pro- require engineers to use this wasted heat. Fur-
cessing route. thermore, slags can contain up to 30% FeO, which

Fig. 2. Major technology developments in the EAF as a function of time (year). Adapted from Toulouevski and Zinurov.4
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1583

Fig. 3. Estimated scrap supply in China by 2020.

Fig. 4. Energy input and output stream balance for a typical EAF process. Adapted from Atkinson and Kolarik.10

can be significant iron losses in the production emissions decrease of about 0.13 t-CO2/t-steel.
stream and should be minimized. Considering that the global production of steels
The recent technology implementation of furnaces through the EAF was 400 million tons in 2012,
has attempted to use the off-gas heat through pre- scrap preheating technology if fully used can lower
heating the scrap before charging into the fur- CO2 by 52 million t-CO2.
nace.2,18–22 Preheating input scrap to 1073 K
(800°C) from room temperature is estimated to save TYPES OF SCRAP PREHEATING SYSTEMS
approximately 90 kWh/t-steel in the EAF, which is FOR THE EAF
20% of the electrical input energy. Considering an
Vertical Shaft Scrap Preheating Systems
average electricity cost to be approximately 0.15 $/
kWh, savings of $13.50/t-steel can be realized, The ECOARC (ecologically friendly and economical
which is a significant portion of the overall costs arc) illustrated in Fig. 5 developed by JP Steel Plan-
associated with the EAF. A 20% reduction in the tech Co. and partially funded by New Energy and
electrical input energy also results in indirect CO2 Industrial Technology Development Organization
1584 Lee and Sohn

(NEDO) in Japan directly connects a preheating for a 100% scrap-based EAF process, but the oxida-
vertical shaft to the alternating current (AC) EAF tion rate of the scrap with increased postcombustion
melting chamber, where the high-temperature pro- in the melting reactor may be severe and optimization
cess exhaust gas during melting, superheating, and for lower postcombustion and slag foaming may be
refining exchanges heat with the scrap.23–25 The off- necessary. A total of five commercial plants are in
gas beyond the preheater is combusted in a post- operation in Japan, Korea, and Thailand with the
combustion temperature above 1073 K (800°C) to first 70-ton capacity facility installed in Kishiwada,
decompose the dioxins (C4H4O2) and furans (C4H4O), Japan. A sixth 200-ton capacity facility is also cur-
and it is rapidly cooled in a spray cooling chamber. rently under construction in Japan for special steel
Scrap at the bottom of the preheating chamber con- grades. However, with any new advances in tech-
tacts the molten steel constantly during the flat bath nology, further improvements are warranted as
operation. The scrap is preheated to approximately issues with scrap partial melting and fusing causing
600–700°C and has been limited to these tempera- sticking, premachining of input scrap due to the
tures due to the downstream off-gas control systems limitation of the shaft dimensions consuming time
to minimize emissions of dioxins and furans from the and costs, cooling-water leakage at the scrap gate
process. Recent data from Dongkuk Steels (Seoul, increasing process downtimes and maintenance costs
South Korea) 120-ton furnace suggest scrap pre- can limit the economic viability of the process. In
heating temperatures to reach nearly 1073 K (800°C) particular, the scrap quality and size is of primary
resulting in a power consumption of below 300 kWh/t- concern to the technology because effective heat
steel, which is one of the lowest energy consumptions transport is maximized by the exposed surface area of
the scrap and convection of the exhaust gas can
determine the kinetics of the heating. An optimized
scrap size promotes good flow control of gases and
minimizes scrap fusing and sticking within the shaft.
Another issue could be the continuous charging of the
scrap, where chemistry homogenization can be a
problem with scrap sources significantly different
from the tap chemistry of the steel. Thus, the supply
chain of the scrap may restrict the raw materials
sources, requiring increased controls compared to
other EAF processes provided in the following sec-
tions.
Similar to the ECOARC, Fuchs used a vertical
shaft preheating system termed continuous opti-
mized shaft system (COSS) depicted in Fig. 6, but
the preheater is not directly connected to the EAF,
which reduces the total height of the system and is
absent of water-cooled parts above the liquid steel
unlike other EAF systems.27,28 Shaft-furnace scrap
Fig. 5. Schematic of the ECOARC process with vertical scrap pre- preheating technology was actually pioneered by
heater26. Fuchs in the late 1980s. The shaft preheating

Fig. 6. Schematic of the Fuchs COSS process with separated retractable vertical scrap preheater.
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1585

advertised COSS to be cheaper than the comparable

Fuch’s Finger Shaft Furnace and the Tenova Con-
steel process.
Another shaft furnace type scrap preheating sys-
tem is the Finger Shaft EAF developed by Fuchs
shown in Fig. 7.21,29–32 Unlike the COSS, the
charging shaft is located directly above the melting
furnace, but the scrap charge is not continuous and
employs a unique scrap-retaining system with wa-
ter-cooled fingers. The operational temperatures for
scrap preheating have reached 773 K (500°C) and
higher depending on the type and size of the scrap
charge used and the degree of postcombustion. A
maximum energy savings of 100 kWh/ton can be
realized in ideal conditions. Tap-to-tap times of
35–40 min have been said to be possible with
appropriate heel in the furnace, which increases the
productivity of the furnace by about 20%. Northstar
Bluescope in the United States operates an efficient
twin-shell twin shaft AC furnace producing
approximately 190 tons of steel per heat. Issues
with the water-cooled fingers and robustness of the
finger components have yet to be completely re-
solved and can increase maintenance costs. To re-
tain the scrap and provide single charging of the
Fig. 7. Schematic of the Fuchs finger-type vertical scrap preheater. scrap, the height of the vertical shaft is higher than
other preheating units currently in operation. The
loss of heat through the scrap-retaining fingers can
furnace uses the off-gas exiting from the side walls limit the maximum preheating temperature of the
of the EAF, and a horizontal scrap pusher directs scrap, which will be a future challenge for the next
the input materials into the steel bath. The move- version of the finger shaft furnace. Other steel
able preheater has less restrictions in the size of the companies have also adopted this new technology,
scrap input, but the typical preheating tempera- including HYLSA in Mexico.
tures of the scrap is approximately 773 K (500°C), Other vertical shaft furnaces for preheating have
resulting in tap-to-tap times of roughly 45–48 min been developed, and this study does not reflect all of
and electrical energy consumption of approximately the vertical preheating types of EAF systems
325–350 kWh/t for a 140-ton capacity using a 40 ton available today. For the vertical scrap preheating
heel according to vendor specifications. The first shaft EAF processes, continuous charging and flat
commercial COSS started up in Shagang ZSJ China bath operation such as those of the ECOARC and
in 2008 and used a feed of 40% hot metal and 60% COSS can have homogeneity issues in the steel
scrap, tapping approximately 105 tons per heat. quality as undermelted scrap can develop within the
There are currently five COSS units globally with furnace. Thus, additional chemical heating near the
the most recent facility installed in Liepajas Met- charging zone of the scrap is typically required.
allurgs Latvia, but is currently inoperative. There Furthermore, as the steel bath and refractory is in
have been issues with the pusher and partial continuous contact, increased refractory wear oc-
maintenance issues with the charging system, but curs for these furnaces, therefore increasing main-
these seem to be improved. The separated scrap tenance and prolonging the downtime for
preheater seems to provide more robust operations permanent refractory damage. In addition, with the
of scrap charging and feeding compared with the vertical position of the shaft, convective heat
ECOARC, but the preheating temperature of the transfer by the gas is fast. With higher tempera-
scrap is lower according to operational data. It can tures of the off-gases and an increased oxygen par-
be speculated because of the separation of the tial pressure, scrap can be oxidized if the CO2/CO
charging vessel and the melting furnace that the ratio is thermodynamically high enough for iron
homogeneity of the steel can be reached faster than oxide to form, which can limit the amount of post-
the ECOARC when continuously charging into a flat combustion in the EAF. The limited amount can
bath, but the heat-transfer efficiency seems to be lower the heat efficiency of the off-gas. The slag
slightly lower. Similar to other water-cooled gate- foaming operation in the EAF can also be an issue
operating systems for scrap charging, water leaking for these continuously fed operations when the in-
and increased maintenance have been experienced put materials change because the thermophysical
for COSS. Vendor-supplied studies comparing properties of the slags including the density, vis-
the maintenance costs per ton steel basis have cosity, and surface tension changes with the slag
1586 Lee and Sohn

Fig. 9. Schematic of a bucket preheater arrangement connected to

the EAF.

Fig. 8. Schematic of the Consteel horizontal scrap preheater continues to be improved, as addressed in many
arrangements with AC EAF. published literature.38,39

Bucket Type and Twin-Shell Scrap Preheating

chemistry and because of the decrease in the free Processes
board height of the furnace when scrap is being
Figure 9 shows a schematic of scrap preheated in
charged from the side of the furnace. Thus, process
the charge bucket with the exhaust gas of the fur-
optimization is continuously needed for these highly
nace.4,40–44 The efficiency of the heat transfer
efficient EAF processes.
within the bucket was not as high as the afore-
mentioned technologies, but the system is simple
Horizontal Shaft Scrap Preheating Systems
and does not require elaborate feeding systems or
Compared to the vertical shaft type preheating controls. However, due to the operational limita-
processes, the only major horizontal shaft scrap tions of the bucket, the preheating temperatures are
preheating process today, as illustrated in Fig. 8, is typically limited to below temperatures for safe
the Consteel process developed by Tenova of Ita- handling of raw materials for transport by cranes,
ly.33–37 An oscillating conveyor system, typically but it has been measured to be between 573 K and
longer than 100 m, continuously charges preselect- 723 K (300°C and 450°C). This amount of preheat-
ed scrap into the EAF. Since 1989, there have been ing in the scrap surmounts to about 40–60 kWh/t-
more than 25 units of Consteel installed globally steel reduction in power consumption and decreased
including two newly installed 160-ton furnaces at tap-to-tap times by 5–8 min. In addition, there has
Dongbu Steels in Korea. The continuous feeding of been increased premature failure of the scrap
the scrap horizontally and increased operationally bucket charge with increased sticking of the scrap
stability of the gas atmosphere within the furnace and nonuniform preheating and deformation in the
due to the absence of furnace top cover openings for bucket. Heat transfer within the past bucket design
scrap charges results in lower nitrogen pick-up and took longer to reach the economical threshold of the
decreased flicker of the electrical power. However, target temperature than the tap-to-tap times of
the horizontal shaft has less heat transfer efficiency conventional EAF operations and was not viable in
than the vertical shaft type because the heat the long term.
transfer mechanism was dominated by radiation A more recent development of the bucket-type
heat transfer from the refractory walls. The exhaust preheater described in Fig. 10 has been developed
gas during operations travels near the roof of the by Danieli as the DANARC,45 which continues to
horizontal shaft while the scrap is oscillated back use a moveable scrap bucket-type preheater. The
and forth at the bottom of the horizontal shaft. preheater bucket station consists of a water-cooled
Thus, direct convective heat transfer between the retention bucket positioned on a retractable car for
scrap and the gas is limited, resulting in a pre- single-bucket charging operations into the melting
heating temperature between 473 K and 573 K furnace. Mean temperatures of 873 K (600°C) can
(200°C and 300°C). This lowers the heat-transfer be reached from exchange with the hot off-gases
efficiency resulting in a power consumption of about from the EAF with gas exiting at approximately
380 kWh/t. In addition, because of the feeding sys- 723 K (450°C). Energy savings of 57 kWh/t have
tem type of the oscillating conveyor, the size of the been realized with a tap-to-tap time decrease of
scrap is limited and homogeneous feeding of the maximum 4 min. However, with the bucket-type
scrap using the oscillating conveyor has been an arrangement, it is speculated that the temperature
issue. Additional maintenance costs associated with uniformity of the scrap is not exceptional and con-
refractory wear and oscillating mechanisms have sidering the exit gas temperatures, the efficiency of
also been a problem for the Consteel process, but it heat transport is not optimal. There have been
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1587

observations of possible CO rich pockets that may option during economic constraints. A comparison of
instigate safety issues, but process controls can be the various scrap preheating processes described
installed to eliminate these issues. previously is summarized in Table I.
Unlike the bucket-type preheater arrangement,
the twin-shell EAF developed jointly by NSC, NKK, SLAG USE AND METAL MAXIMIZATION
SMS-Demag and CLECIM uses an additional EAF TECHNOLOGIES
vessel instead of a preheating bucket with a com-
Slag Waste Heat-Recovery Technologies
mon arcing electrode and power supply sys-
tem.4,43,46–49 By connecting the two EAF vessels, Heat recovery through preheating scrap using the
one in operation and the other on standby, the high- high-temperature exhaust gas from the EAF has
temperature exhaust gases are directed toward the actively been studied and commercial-scale pro-
standby vessel and preheats the scrap. Compared cesses have been developed since the 1980s.50–59
with a single vessel, the operation power consump- The use of these high-energy off-gases have signifi-
tion savings are estimated to be approximately cantly improved the energy efficiency of the EAF,
17 kWh/t-steel. Although the amount of preheating but it would seem that the technological limitations
is not as substantial as the aforementioned pro- of scrap preheating has reached maximum heating
cesses, the preheating system is relatively simple capabilities of 1073 K (800°C) considering the eco-
and complex feeding systems are not required. nomics and scrap fusing and melting at higher
However, preheating of scrap to higher tempera- temperatures as well as the reoxidation of the scrap.
tures comparable with the ECOARC and COSS Further energy minimization and increased effi-
result in emission control issues and downstream ciencies may be reached by utilizing the metallur-
gas scrubbing systems and cooling systems maybe gical slags of the EAF and tapping into its vast
required for optimal scrap preheating operations. thermal energy currently being ignored.
Furthermore, additional capital costs are needed for An estimated 69 million tons of EAF slags with a
an additional EAF vessel, which may not be a viable speculated heat value of 35 TWh/year is laid waste
in the typical slag pits, where slag is poured and
slowly cooled for either landfill or road construc-
tion.60 Because of their high FeO contents, these
steelmaking EAF slags cannot be used for higher
value-added Portland cements unlike the blast fur-
nace slags. Three types of technology are being
developed for possible energy recovery from slags
including energy recovery as hot air or steam, con-
version to fuels through endothermic chemical
reactions, and thermoelectric power generation.
Large-scale pilot trials and a few semicommercial-
scale plants have been developed for the energy
recovery as hot air or steam and will be reviewed in
detail in this section. The major constituents of the
EAF slag compositions are provided in Table II.
EAF slags are typically tapped at approximately
1823 K (1550°C) and approximately 150 kg-slag/
t-steel is formed during the EAF process. Thermal
conductivity of the slag on average is less than
Fig. 10. Daniel DANARC PLUS M2 melting and preheating furnace
0.5 W/m K,50,61–63 which makes slag cooling and
arrangement adapted from Michielan and Fior.45 heat transfer a significant problem for engineers,
and thus limited commercial-scale processes have

Table I. Summary of scrap preheating characteristics for high-efficiency EAF processes

Published
preheat Approximate
Preheating type Description temperatures (K) energy savings (kWh/t) Remarks
Vertical shaft ECOARC Max. 1073 90 Heat transfer high
Vertical shaft COSS 773 60 Heat transfer moderately high
Vertical shaft Finger shaft 773 60 Heat transfer moderately high
Horizontal shaft Consteel 573 40 Heat transfer moderate
Bucket charge DANARC 723 57 Heat transfer moderate
Twin shell Twin shell EAF 473 17 Heat transfer low
1588 Lee and Sohn

Table II. EAF slag compositional range for the major components in low carbon steel production

Component CaO SiO2 Al2O3 MgO FetO

Amount (mass%) 35–60 9–20 2–9 5–15 15–30

been developed. However, considering the issues

with the lower thermal conductivity of the molten
slags technology, developments for efficient heat
exchange have focused on obtaining higher surface
area and smaller particle sizes through atomization,
as is the case for most of the technology described in
this article.
In the review by Barati et al.,50 thermal energy
recovery methods from molten slags have been well
described. Rotary cup atomizers, spinning disk,
rotating drum, and solid slag impingement, as well
as air blast granulation, are considered to be the
most commercially viable methods for energy
recovery and some are discussed in the following. In
the rotary cup atomizing technique devised by
Pickering et al.53 shown in Fig. 11, a rotating cup
combined with an air blast atomizes the molten
slag, and heat recovery occurs in subsequent fluid-
ized beds of the amorphous slag particles providing
an energy recovery speculated to be approximately
59%. However, the operational durability of the
rotating cup and fluidized bed control makes the
operation moderately complex. In addition, the wa-
ter-cooled walls in the laboratory-scale setup may
result in additional heat losses, which would likely
be replaced with an insulated wall for better heat Fig. 11. Rotary cup atomizing process schematics initially devel-
transport to the heat-exchanging air. oped by Pickering et al.53 Note the air blast supply at the rotary cup
for heat exchange and solidification.
A simpler process could be the spinning disk slag
granulation technology and heat-exchange method
driven by Sumitomo Metal54 and further developed
by Akiyama et al.64 and Xie et al.65 Slag is spread accumulate in a cooling chamber and exchange heat
onto a rotating disk and scatters the liquid slag and with air. The temperature of the air reaches
cooled in-flight as it interacts with the dispersed approximately 773 K (500°C) and was speculated to
cooling air similar to the rotary cup atomizer but have heat recovery rates of approximately 60%.
substituted with a rotating disk. The slag granules However, because of the slower rotating speeds of
are collected onto a fluidized or packed bed and air the drum, a thin layer of slag could possibly adhere
passes through the hot particles resulting in hot air to the surface of the drum, which can significantly
at approximately 873 K (600°C). The spinning disk lower heat transfer and result in inefficient opera-
method seems to result in increased uniformity of tion of the rotating drum process. Furthermore, the
the granules compared with the rotating cup effective contact between the slag and the drum
atomizer and smaller particles. Considering the may require significant diameters and length to
angular momentum to propel the liquid slag and meet the needs of the amount of slags formed from a
control the particle sizes, the spinning disk may 100-ton and higher EAF furnace similar to a rotary
require a controlled steady-state feeding and in- kiln in the cement industry, which could be an
creased rotation speeds for commercial-scale appli- excessive capital investment for the steel industry.
cations. However, excessive rotation speeds can Compared with the abovementioned centrifugal
result in irregular shaped glassy silk instead of a type of dry granulation technologies, granulation by
uniform particle.51 air blasts similar to the Ecomaister and solid slag
A continuous rotating drum process for dry impingement processes have also been developed.59
granulation and subsequent heat recovery uses a The liquid slag is poured into a blast of air and
rotating drum and disintegrates the liquid slag atomized into particles that are projected toward
solidifying the small particles as it travels down the the gutter of the waste heat boiler, where heat ex-
inclined drum.50,51 The particles, which are less change between the hot particles is transferred via
uniform than the rotating cup or the spinning disk, convective and radiative heat to the tubes. A second
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1589

boiler located at the bottom of the pile of slag chemical energy could be stored in the CO and H2
granules extracts the heat, and it is estimated that gases.
the waste heat boiler recovers about 40% of the
energy and another 40% is transferred to the hot air Improvements in Slag Compositional Control
exhausted at 773 K (500°C), which could be poten- and Lower FeO
tially used for increased heat-transfer efficiency.
Beyond the energy recovery of the exhaust gas
and slag, another significant issue for increased
Production of Hydrogen Through Thermal
energy efficiency in the EAF is concerned with the
Cracking of Gases During Slag Rapid
slag composition itself. As described in Table II, the
Quenching
FeO content depending on the operation and feed
Although significant developments to exchange materials is between 20% and 30% FeO. If the
heat after slag granulation have been realized and average FeO is used, then approximately 25%FeO is
pilots to commercial-scale plants have been devised contained in the EAF waste slag that is typically
particularly in Japan and Sweden. The potential dumped on the ground. That constitutes approxi-
energy of the molten slag by heat transport and mately 19.7 million tons of FeO and subsequently
subsequent heating of the air will likely have limi- 15.3 million tons of Fe lost after significant energy
tations because of the successive transport of the has already been supplied to produce the highly
heat from multiple stages. Instead of transferring valued metal component. Thus, energy savings in
the thermal energy of the slag from one medium to the EAF must also maximize the return of Fe and
another, this section discusses the possible use of minimize losses to the slag by process developments
slags by transitioning the thermal energy to chem- in the EAF before tapping the steel.
ical energy, which can be used effectively later for Recent work has shown that additions of appro-
potential fuel applications. priate reductants such as Al, C, SiC, and combina-
By using the thermal energy of the slag, the tions of these reductants can significantly lower the
endothermic thermochemical decomposition of wa- FeO content of the slag and increase the metal
ter to hydrogen can be realized. By mixing H2O and production of the EAF process. In the work of Joo
C in a 10:1 mass ratio with the molten slag, the et al.,68 EAF steelmaking slags containing T. Fe of
energy provided by the slag is used directly to form 23% could be reduced with SiC-Al-CaO and SiC-C-
H2 and CO, which requires a heat of enthalpy of CaO composite reductants without significantly
approximately 131 kJ/mol. Matsuura and Tsukih- increasing the slag volume. Additions of CaO con-
ashi66 described the oxidation of FeO in molten trolled the viscosity of the slag during the reduction
steelmaking slag and the formation of H2. With 20– of FeO for continued operation in the EAF. Thus, as
30% FeO in typical EAF slags, thermodynamic the recovery of metal increases from the slag, the
predictions suggested hydrogen could be produced thermophysical properties of the slag could change
according to reaction (1). and affect the process parameters. According to Kim
and Min,69 as the FeO is substituted with Al2O3 in
2FeOðlÞ þ H2 OðgÞ ¼ Fe2 O3ðsÞ þ H2ðgÞ the CaO-SiO2-FeO-Al2O3-MgO EAF slag system at
(1)
DG ¼ 82; 130 þ 102:7T ðJ/molÞ 1823 K and fixed basicity of unity, the MgO solu-
bility for the slag system increases,

but beyond a
certain XAl2 O3 XAl2 O3 þ XFeO , the primary struc-
Although forward reaction (1) is not spontaneous ture of the slag changes from a magnesio-wustite to
in its standard states, the control of the activities of a spinel structure. This change in structure could
the above constituents may allow the equilibrium increase the breakpoint of the slag and thus signif-
distribution of the H2 and H2O to occur and thus icantly decrease the fluidity of the slag making
promote the formation of hydrogen as calculated by operations difficult. Before tapping the steel, addi-
Matsuura and Tsukihashi.66 For the FeO-CaO- tions of Al and other reductants within the slag
SiO2-Al2O3-MgO-P2O5 system containing 32–50% layer could reduce the overall FeO content in the
FeO at 1773 K (1500°C), an Ar-0.086H2O feed gas slag as depicted in Fig. 12. An increase in yield of
for 120 min resulted in a maximum production of 2% has been realized with the new tapping proce-
0.59 mol-H2/kg-slag. dure. However, continued work is required to en-
Maruoka et al.67 using the rotary cup atomizer sure complete separation of the metal and slag and
designed a system injecting methane and steam to compensate for the changes in thermophysical
toward the hot granules of the cup and reformed the properties of the slag with a lower FeO content.
CH4 to form CO and H2 according to reaction (2).
CH4ðgÞ þ H2 OðgÞ ¼ CO(g) þ 3H2ðgÞ (2) METAL RECOVERY FROM WASTE SLAGS
Fe Separation and Enrichment in EAF Slags
Using Controlled Crystallization
The expected theoretical heat recovery was esti-
mated to be approximately 83% using this Slags with high concentrations of metal-contain-
endothermic reaction, and significant amounts of ing compounds have been speculated to release
1590 Lee and Sohn

Fig. 12. Effect of reductant types on the reduction rate and slag volume of FeO-containing EAF slags at 1823 K.

metals or harmful elements causing water and soil operations and the Fe-deficient phase could be
pollution beyond the toxicity characteristic leaching directly used as an aggregate for clinkers in Port-
procedure criterion.70 Small particulates can be a land cement.
risk to humans through inhalation. Although both Although only laboratory-scale experiments have
Proctor et al.71 and Geiseler72 have shown that steel been done, pilot scale experiments are currently in
slags should not be considered hazardous wastes, progress and further applications to this novel pro-
there have been increased restrictions and tougher cess could be expanded to Ni, Cr, and other non-
regulations that make it difficult to dump slag ferrous materials and processes. The advantage of
without decreasing the metallic species. In addition, this process is that no additional heating is required
maximizing metal recovery from the slags can and it uses only the sensible heat that is currently
increase productivity and decrease environmental available in the slag melt itself.
treatment costs for the steel industry. From the Weiss et al.76 discussed methods of adding min-
works by Jung and Sohn,73–75 instead of dumping erals to steelmaking slags followed by crystalliza-
the molten slag and allowing slow cooling, a con- tion to recover metal values from steel slags. The
trolled cooling pattern could concentrate the Fe mineral additive acts as a heterogeneous nucleation
cations into the spinel crystal structure (MgAlFeO4) site for dicalcium silicate formation and allows
enriching local areas of the primary spinel crystals separation between a slag and metal chips for
and separating the Fe from the amorphous phases increasing metal recovery.
of the slag, as illustrated in Fig. 13. This novel
method would cool the slag to 1473 K (1200°C) from
1823 K (1550°C) and isothermally hold the slag at FUTURE DEVELOPMENTS IN INCREASED
the target temperature for a defined period and then ENERGY EFFICIENCY FOR EAF STEEL-
continuously cool to room temperature. At a defined MAKING
temperature, the thermal energy provides the Improved Exhaust Gas Utilization Technology
driving force for diffusion and partial substitution of and Preheating
the Al3+ cation with the Fe3+ cation contained in the
slag within the primary spinel crystal structure. During the EAF operation of boring, melting, and
When the subsequent slag is pulverized and mag- heating, the amount of gases and composition of
netically separated, a highly enriched Fe-containing gases vary. In addition, with exothermic and endo-
spinel phase and an Fe-deficient amorphous phase thermic chemical reactions taking place during
could be obtained. This Fe-enriched phase could these periods, a dynamic fluctuation of both the
be reused back into the primary steelmaking composition and off-gas temperatures can be typi-
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1591

Fig. 13. SEM-EDS mapping of EAF slags containing FeO in the CaO-Al2O3-MgO-FeO slag system isothermally cooled to 1473 K. Note the
localized concentration of Fe into the primary spinel crystals.73

cally observed. Postcombustion ratio of the off-gases required in much of the operational know-how for
can directly affect both the temperature in the furnace these new generations of EAF processes. Companies
and the degree of preheating of the scrap. To obtain have attempted a trial-and-error examination of the
an optimum heat balance that considers and inte- optimal scrap feeding procedure, but room for
grates these reactions would likely be the next po- enhancements is possible.
tential step for increased savings for the EAF process.
Compared with the top cover removal and
Durability Improvements to the Supplemental
charging practices of typical EAFs, the preheating Equipment of the EAF
of feed materials through a preheating vessel will
constrain the input scrap size. Excessive scrap sizes Most of the operators of preheating-type EAFs
could physically get stuck during operations and have concerns on the cooling-water leakage at gates,
heat transfer is known to be decreased because of panels, fingers, and pushers. Considering the ideal
the lower surface contact area for comparable scrap gas law of water vapor entrapped within liquid
charge amounts. Innovative preheating design and steel, significant internal pressures can build up
charging concepts that could overcome the scrap and result in explosions within the furnace that
shape limitations is expected to be another area for could seriously damage equipment and harm per-
further improvements. Depending on the particular sonnel, which is of primary concern. Thus, addi-
size and type of scrap, the fluid flow dynamics of the tional engineering specifications for leak-free
exhaust gas and optimal heat transfer is still designs must be one of the primary concerns for
1592 Lee and Sohn

(a) CRACCK process for chemical fuels

Feed Material
CO CO2 Preheater
Offgas Post Combustion Reactor

(1) CO + 1/2O2 = CO2

(2) CO2
CO CO2
Offgas
EAF

CO Fuel Burning
C + CO2 = 2CO

Carbonaceous
C
Materials

(b) CRACCK process for pre-reduction

Iron Ore
or
Composite DRI
C Carbonaceous
Materials
CO CO2
Offgas
C + CO2 = 2CO

Post Combustion Reactor Feed Material

Preheater
(1) CO + 1/2O2 = CO2
(2) CO2 CO CO2
EAF/Smelter
Offgas

Fig. 14. Cyclic use of CO2 for chemical energy in the proposed CRACCK (CO2 recycling and conversion to CO in Korea) process as (a) chemical
fuels and (b) prereduction.

engineering and continued expansion of the most Utilization of Alternative Iron Units
efficient EAF processes available today. Recent to Substitute Scrap for Higher Quality
developments to omit water-cooling devices within and Increased Heat Efficiency
the hot gas traveling path have been implemented.
As the shale gas revolution takes hold and higher
Excessively high temperatures of scrap pre-
value-added steel production through the EAF
heating can result in partial melting and fusion of
becomes economically feasible, direct reduced iron
scrap, causing materials to stick. Although higher
(DRI) has become a possible source to substitute
temperatures and increased heat transfer to the
high grade scrap in the EAF. Although continuous
scrap decreases the overall power consumption
charging of DRI into the EAF is a regular practice in
and tap-to-tap times in the furnace, a maximum
many of the typical EAF and also in the Consteel
temperature of 1073 K (800°C) is likely consider-
process, other vertical shaft furnaces have yet to
ing the process issues that could occur at higher
expand on the uses of DRI in the preheaters. It can
temperatures. It is speculated that for some
be speculated that DRI could be a clean substitute
alternative iron units such as hot briquetted iron
and dilutant to scrap containing significant copper
(HBI), higher preheating temperatures can be
and tin, which is difficult to remove through
achieved.
Review of Innovative Energy Savings Technology for the Electric Arc Furnace 1593

oxidizing slags. Furthermore, depending on the partial melting and fusion of the scraps within the
carbon content, the melting rate of the raw mate- preheating vessels, it is speculated that for scraps a
rials can be decreased, thereby lowering the amount maximum heating temperature is expected to be
of power for melting. Thus, further work with close to 1073 K (800°C), resulting in approximately
alternative iron units for the preheating EAF pro- 90 kWh/t-steel energy savings. However, it should
cesses is needed. be mentioned that the adverse effects of dioxin and
furan formation during preheating of scrap require
Conversion of Thermal Energy to Chemical strict environmental regulations in some countries
Energy that make it difficult to widely implement these
energy saving technologies to the steel industry.
The exhaust gas from the furnace contain signif-
This is particular true in Europe and less in Asia by
icant energy that could be used for producing high-
comparison.
temperature chemical energy such as CO from car-
Beyond the scrap preheating technologies, the use
bonaceous materials similar to the production of H2
of waste heat of the slags is likely to be the next
and CO from thermal cracking and methane gas
technological breakthrough that must be achieved
reformation reactions of waste slags, as described
in the mid to long term for continued viability of the
previously. The CO2 gases could react with carbon
EAF, as power consumption exceeds the supply and
in the carbonaceous materials according to the
electricity costs continue to rise. Very few commer-
endothermic Boudouard reaction (3) using the sen-
cial-scale heat-exchanging processes have been
sible heat of the exhaust gas.
developed, and significantly more research and
CðsÞ þ CO2ðgÞ ¼ 2COðgÞ (3) development funding in this area is critical. In
addition, increasing metal recovery using novel
methods of controlled cooling and localized concen-
CO2 gas could be recycled within a circulating tration of Fe and subsequent separation of the
system and produce CO gases for either fuels for amorphous phase for use in the cement industry can
chemical heating back into the EAF or as a reducing provide benefits to both the steel and cement
agent for HBI again producing CO2 to react again industry. Developing commercial-scale process
with incoming C, as illustrated in Fig. 14a and b. parameters is the next hurdle for this novel process.
This conceptual proposed process is termed CO2 The conversion of thermal energy to chemical en-
recycling and conversion to CO in Korea (CRACCK) ergy via reactions of the hot exhaust gas containing
but has yet to be experimentally verified. Fig- CO2 with carbonaceous materials to form CO may
ure 14a shows the high-temperature CO/CO2 off- also be a possible alternative to use the high-tem-
gas mixture, which is postcombusted after the pre- perature off-gas. The CO gas can be used as a fuel for
heater providing additional exothermic heat. The chemical energy back into the EAF or as a prere-
sensible and additional postcombustion heat sup- ducing gas. This CO2 recycling and conversion to CO
plies the endothermic energy for reaction between could be effective in mitigating greenhouse gas evo-
CO2 and carbonaceous materials such as steaming lutions within the steelmaking shop as well.
coals to form CO gases at relatively high tempera-
tures. This CO gas is recycled back into the furnace ACKNOWLEDGEMENTS
with oxygen-rich burners for chemical heat supplied This work has been partially supported by the
to the furnace producing CO2 and recirculating BK21 (Brain Korea 21) PLUS Project in the Division
within the loop. In Fig. 14b, the CO gas is used as a of the Eco-Humantronics Information Materials and
reducing gas for iron ore, which prereduces the Ministry of Trade, Industry, and Energy (2014-11-
oxides for partial reduction. Additional reduction 070).
can be completed either in a smelter or by addition
of carbon into the ore as a composite pellet for pro- REFERENCES
duction of DRI into the EAF. 1. World Steel Association, Crude Steel Production (2012).
http://www.worldsteel.org/statistics/crude-steel-production.
SUMMARY AND CONCLUSION html. Accessed 30 May 2014.
2. World Steel Dynamics, World Steel Dynamics Report-Scrap
With preheating technology development, the Outlook: A New Bundle (2012). http://www.worldsteeldynamics.
past 30 years of research and development efforts com/pg/reports. Accessed 15 Sept 2013.
have resulted in viable commercial applications for 3. C.P. Manning and R.J. Fruehan, JOM 53(10), 36 (2001).
4. Y.N. Toulouevski and I.Y. Zinurov, Innovation in Electric
increased energy efficiency for the EAF process. In Arc Furnaces (New York: Springer, 2010), pp. 1–23, 81–92.
particular, the vertical shaft and horizontal pre- 5. J.P. Birat, Revue de Me´tall. 97, 1347 (2000).
heaters have gained widespread acceptance for fu- 6. OECD, OECD Environmental Outlook to 2030 (Paris: OECD
ture greenfield EAF installations. An average Publishing, 2008).
energy savings using preheaters of 50 kWh/t-steel 7. L. Mathiesen and O. Maestad, Energy J. 25, 91 (2004).
8. A.W. Wyckoffa and J.M. Roop, Energy Policy 22, 187 (1994).
have been realized, and continued optimization of 9. S.H. Anderson, G.E. Metius, and J.M. McClelland, 60th
the individual processes now are at the brink of Electric Furnace Conference Proceedings (Warrendale, PA:
maximum operational capabilities. Because of the ISS, 2002), pp. 175–191.
1594 Lee and Sohn

10. M. Atkinson and R. Kolarik, AISI Steel Industry Technology 41. K.E. Sewald, 44th Electric Furnace 957 Conf. Proc. (War-
Roadmap Report (Oak Ridge, TN: American Iron and Steel rendale, PA: ISS, 1987), pp. 243–250.
Institute, 2001), pp. 23–28. 42. A.B. Steblov, A.V. Mateiko, and É.A. Steblova, Metallurgist
11. J.P. Birat, Ironmak. Steelmak. 28, 152 (2001). 50, 571 (2006).
12. N.A. Godinskii, N.N. Kushnarev, D.S. Yakhshuk, V.I. 43. E. Worrell, L. Price, and N. Martin, Energy 26, 513 (2001).
Kotenev, and E.Y. Barsukova, Metallurgist 47, 16 (2003). 44. V. Portanova, JOM 36, 37 (1984).
13. C.W. Forsberg, Int. J. Hydrogen Energy 32, 431 (2007). 45. A. Michielan and A. Fior, 6th European Electric Steelmak-
14. K.S. Yoshiki-Gravelsins, J.M. Toguri, and R.T.C. Choo, ing Conf. Proc. (Warrendale, PA: ISS, 1999), pp. 204–209.
JOM 45(5), 15 (1993). 46. D. Neuschutz, High Temp. Mater. Process. 4, 309 (2000).
15. O. Ostrovski and G. Zhang, AIChE J. 52, 300 (2006). 47. M. Gojic, Metalurgija 43, 163 (2004).
16. T. Emi and S. Seetharaman, Scand. J. Metall. 29, 185 48. B. Butcher, B. Laroy, K. Vanover, Y. Krotov, and S. Meyer,
(2000). Iron Steel Technol. 8, 42 (2011).
17. G.H. Geiger, Proc. International Symposium on Global 49. H. Trenkler, ABB Rev. 9/10, 18 (1996).
Environment and the Iron and Steel Industry (Beijing, 50. M. Barati, S. Esfahani, and T.A. Utigard, Energy 36, 5440
China: CMS, 1998), pp. 191–196. (2011).
18. D. Yanez, M.A. Pedroza, J. Ehle, and H. Knapp, Proc. 6th 51. H. Zhang, H. Wang, X. Zhu, Y.-J. Qiu, K. Li, R. Chen, and Q.
European Electric Steelmaking Conference (Dusseldorf, Liao, Appl. Energy 112, 956 (2013).
Germany: Verlag Stahleisen mbH, 1999), pp. 24–28. 52. P. Berger, Stahl Eisen 50, 1775 (1930).
19. I.A. Madugu and M. Abdulwahab, Aust. J. Basic Appl. Sci. 53. S.J. Pickering, N. Hay, T.F. Roylance, and G.H. Thomas,
2, 835 (2008). Ironmak. Steelmak. 12, 14 (1985).
20. T. Wang, M. Kawakami, K. Mori, and S.H. Shahidan, Mater. 54. M. Yoshinaga, K. Fujii, T. Shigematsu, and T. Nakata,
Sci. Forum 449, 329 (2004). Trans. Iron Steel Inst. Jpn. 22, 823 (1982).
21. H. Ludwig, SEAISI Q. 30, 16 (2001). 55. J.R. Donald and C.A. Pickles, Proc. 77th Steelmaking Conf.
22. S.S. Baker, A.M.W. Briggs, P.J. Lewis, D. Capodilupo, E. (Warrendale, PA: ISS, 1994), pp. 681–692.
Repetto, S. Gonthier, Y. Zbaczyniak, B. Kleimt, S. Kohle, M. 56. T. Akiyama, K. Oikawa, T. Shimada, E. Kasai, and J.-I.
Knoop, G. Mosel, and P.G. Oberhauser, European Com- Yagi, ISIJ Int. 40, 286 (2000).
mission Report—Ecological and Economical EAF Steel- 57. T. Shimada, V. Kochura, T. Akiyama, E. Kasai, and J. Yagi,
making EUR 19480, 2001, pp. 1–442. ISIJ Int. 41, 111 (2001).
23. Y. Fujiwara, Mater. World 9, 17 (2001). 58. F. Sieverding, Steel Times 208, 469 (1980).
24. R. Yamaguchi, H. Mizukami, T. Maki, and N. Ao, 58th 59. J. Ando, T. Nakahar, H. Onous, S. Tchimura, and M. Kondo,
Electric Furnace Conference Proc. (Warrendale, PA: ISS, Technical Review (Japan: Mitsubishi Heavy Industries,
2000), pp. 325–336. 1985), pp. 136–142.
25. H. Mizukami, R. Yamaguchi, T. Nakayama, T. Maki, N. Ao, 60. World Steel Association, World Steel Association Yearbook
and Y. Sato, NKK Tech. Rev. 82, 14 (2000). and Factsheets (2012). http://www.worldsteel.org/Steel. Ac-
26. T. Sugasawai, H. Kato, and T. Nagai (Paper presented at the cessed 30 May 2014.
2013 SEASI Conference and Exhibition, Dusit Thani Pat- 61. Verein Deutscher Eisenhuttenleute, Slag Atlas, 2nd ed.
taya, Thailand, 3–6 June 2013). (Dusseldorf: Verlag Stahleisen, 1995), p. 597.
27. J. Ehle and G. Fuchs, U.S. patent 5153,894 A (1992). 62. M. Susa, K. Nagata, and K.S. Goto, Trans. JIM 29, 133
28. United States Environmental Protection Agency, Available (1998).
and Emerging Technologies for Reducing Greenhouse Gas 63. K. Nagata, M. Susa, and K.S. Goto, Tetsu-to-Hagane´ 69,
Emissions from the Iron and Steel Industry (Research Tri- 1417 (1983).
angle Park, NC: U.S. Environmental Protection Agency, 64. T. Akiyama, T. Shima, T. Kasai, and J. Yagi, China-Japan
Office of Air and Radiation, 2012). http://www.epa.gov/nsr/ International Academic Symposium (2000), pp. 53–65.
ghgdocs/refineries.pdf. Accessed 30 May 2014. 65. D. Xie, Y. Pan, R. Flann, B. Washington, S. Sanetsis, and J.
29. M.A. Pedroza, R.G. Gonzalez, and M.A. Herrera, Ironmak. Donnelley, Proc. 1st CSRP Annual Conf. (Melbourne, Aus-
Steelmak. 24, 83 (1997). tralia: CSIRO, 2007), pp. 29–30.
30. M. Haissig and W. Gebert, 56th Electric Furnace Conference 66. H. Matsuura and F. Tsukihashi, ISIJ Int. 52, 1503 (2012).
Proc. (Warrendale, PA: ISS, 1998), vol. 55, pp. 391–405. 67. N. Maruoka, T. Mizuochi, H. Purwanto, and T. Akiyama,
31. J. de Beer, Potential for Industrial Energy-Efficiency ISIJ Int. 44, 257 (2004).
Improvement in the Long Term (Dordrecht: Kluwer Aca- 68. S.-W. Joo, J.-D. Shin, D.-K. Shin, S.-H. Hong, J.-S. Ki, J.-I.
demic Publishers, 2000), pp. 145–147. Hwang, and B.-D. You, Korean J. Met. Mater. 50, 753 (2012).
32. A. Villar, J.J. Arribas, and J. Parrondo, Clean Technol. 69. Y.J. Kim and D.J. Min, Steel Res. Int. 83, 852 (2012).
Environ. Policy 14, 29 (2012). 70. H. Shen, Waste Manag. 23, 933 (2003).
33. P. Argenta and M.B. Ferri, La Metall. Ital. 1, 41 (2005). 71. D.M. Proctor, K.A. Fehling, and E.C. Shay, Environ. Sci.
34. M.H. Adjei-Sarpong, M. Fox, B. Trumble, and J. Powers, Technol. 34, 1576 (2000).
Iron Steel Technol. 3, 65 (2006). 72. J. Geiseler, Waste Manag. 16, 59 (1996).
35. C. Giavani, E. Malfa, and V. Battaglia, SEAISI Q. J. 41, 16 73. S.S. Jung and I. Sohn, Environ. Sci. Technol. 48, 1886
(2012). (2014).
36. F. Memoli, M. Guzzon, C. Giavani, M. Zanforlin, and P. 74. S.S. Jung and I. Sohn, Mater. Metall. Trans. B 43B, 1530
Galbiati, Iron Steel Technol. 7, 56 (2010). (2012).
37. J. Tang, M.B. Ferri, and P. Argenta, Ironmak. Steelmak. 32, 75. S.S. Jung and I. Sohn, J. Am. Ceram. Soc. 96, 1309 (2013).
191 (2005). 76. F.J. Weiss, M.A. Goksel, J.L. Coburn, and G.E. Metius, The
38. J. Simmons, A. Manenti, and K. Shoop, 60th Electric Fur- Recycling of Steel Plant Waste Oxides Using the PTC (Pellet
nace Conference Proc. (Warrendale, PA: ISS, 2002), vol. 60, Technology Corporation) Cold Bond Carbon Bearing Pellet
pp. 377–386. Technology. Disposal, Recycling and Recovery of Electric
39. J.A. Vallomy, Scand. J. Metall. 28, 277 (1999). Furnace Exhaust Dust (Pittsburgh, PA: Iron and Steel
40. A. Michielan, A. Fior, and G. Lavaroni, 6th European Elec- Society, 1987), p. 115.
tric Steelmaking Conf. Proc. (Warrendale, PA: ISS, 1999),
pp. 78–83.