Chapter 13

Chapter 13
Alternative Oxygen Steelmaking

Processes
R. J. Fruehan, Professor, Carnegie Mellon University
C. L. Nassaralla, Assistant Professor, Michigan Technological University
13.1 Introduction
Virtually all of the steel in the United States and the world is produced either in an oxygen steel-
making converter such as a BOF, LD, OBM (Q-BOP) etc. or an electric arc furnace (EAF). The
only exception are the few remaining open hearth shops and a negligible amount in other processes
such as the Energy Optimizing Furnace (EOF). This is understandable since much of the steel-
making capacity was built or rebuilt from 1955 to 1975; the BOF and EAF were the best tech-
nologies available at the time and were compatible with large blast furnace production and
relatively inexpensive scrap. However, these conditions are changing. It may be possible to produce
iron economically on a small scale, 0.5 million tons per year, in new processes such as COREX
and bath smelting. Steelmakers desire more flexibility to use scrap and hot metal and other forms
of iron such as direct reduced iron (DRI) and iron carbide. Therefore, it is of value to re-examine
steelmaking options.
There is currently a large amount of development work related to the electric arc furnace includ-
ing scrap preheating, continuous melting and the use of fossil fuels and hot metal. These are dis-
cussed in detail in Chapter 10. In particular electric arc furnaces will use more oxygen to oxidize
carbon and for post-combusting CO to CO2. The EAF will act more like a BOF using fossil fuel
such as carbon in iron or in direct reduced (DRI) products. They will also use more virgin (non
scrap) iron such as DRI and liquid hot metal. The alternative steelmaking processes examined in
this chapter are those using no electrical energy.
Many alternative oxygen steelmaking processes were being developed in the 1970s such as IRSID
continuous steelmaking, WORCRA and the Bethlehem continuous process. However, these never
were commercialized, in part because most companies invested in conventional oxygen steelmak-
ing and no new capacity was required. In recent years, other processes have been developed or pro-
posed such as the EOF, the AISI continuous process, iron carbide continuous processes and
IFCON. In this chapter, the general principles of process types are examined, selected steelmaking
processes are technically evaluated and an economic analysis of selected processes is made. This
has been the subject of a 1998 publication and much of this chapter comes from that publication.1
13.2 General Principles and Process Types

The steelmaking processes considered in this chapter can be classified into two general types:
batch and continuous. The continuous processes can be classified as a continuous stirred tank reac-
tor (CSTR) or plug flow reactors (PFR). The process types are shown in Fig. 13.1 along with their
characteristics with regards to concentration and reaction rates.
Copyright © 1998, The AISE Steel Foundation, Pittsburgh, PA. All rights reserved. 743
Steelmaking and Refining Volume
dV
Batch CSTR PFR
Mixing } complete complete only in dV
Concentration none none longitudinal

Gradient
Concentration with time constant with distance

Changing
Reaction
speed with time constant longitudinal
charging
Fig. 13.1 Classification of steelmaking process types and their characteristics.
By performing a simple mass balance in a reactor, one can see that the degree of dispersion in the
system defines the type of reactor where the process takes place. The contributions to backmixing
of fluid flowing in the x direction can be described by the following equation2:
¶C ¶ 2C
=D 2 (13.2.1)
¶t ¶x
where the parameter D is called the longitudinal dispersion coefficient, which characterizes the
degree of backmixing during flow. Equation 13.2.1 can be written in a dimensionless form where
z = (ut + x)/L and Q = t/t = tu/L
¶ C æ D ö ¶ 2C ¶ C
=ç ÷ - (13.2.2)
¶ Q è uLø ¶z 2 ¶ z
where the dimensionless group (D/uL) is the inverse of Peclet number, and measures the extent of
longitudinal dispersion. The Peclet number represents the ratio of material transferred by bulk flow
to material transport by Eddy diffusion. L is the length of the reactor and u is the velocity in the
direction of the flow. Therefore,
D
® 0 negligible dispersion, (PFR)
uL
D
® ¥ large dispersion, (CSTR)
uL
It means that when the diffusion, plus convection inside the vessel, is small the reactor approaches
an ideal plug flow reactor. On the other hand, when the diffusion and convection are large, the reac-
tor approaches a continuous stirred tank reactor.
Szekely3 has shown that for values of (D/uL) greater than 0.2 there would be extensive dispersion.
He further concluded that furnaces of more than 100 meters in length may be required for minimal
dispersion. Measurements of the (D/uL) from an actual operation of the WORCRA furnace give
value of 0.1 indicating shorter furnaces may be satisfactory4. For plug flow Taylor8 has shown that
the length to width ratio, L/W, should be around 50 for this type of reactor to be a PFR. For highly
744 Copyright © 1998, The AISE Steel Foundation, Pittsburgh, PA. All rights reserved.
Alternative Oxygen Steelmaking Processes
stirred steelmaking systems it is difficult to maintain a PFR. The closest such reactor in steelmaking
is a continuous casting tundish. Even for a tundish where turbulence is kept to a minimum it deviates
significantly from a perfect PFR. In some PFR steelmaking processes, such as WORCRA, the slag
runs counter-current to the metal. This has advantages when refining hot metal with high phospho-
rus contents. A further discussion of trough concurrent and counter-current steelmaking is given by
G. Brooks et al 5.
The normal oxygen steelmaking processes, the EAF, and the EOF are batch processes. The AISI
and IRSID processes are CSTR while the WORCRA, Bethlehem and the initial Iron Carbide
processes should approach plug flow reactors; these processes are discussed later.
In steelmaking, the primary reaction is decarburization. At high carbon contents the rate is simply
controlled by the oxygen input rate6. Below a critical carbon content, typically about 0.3%C for
normal OSM, the rate is controlled by mass transfer of carbon, is first order with respect to car-
bon, and given by:
d %C
dt
[
= - k %C - %C e ] (13.2.3)
where %C and %Ce are the carbon content and the equilibrium content with the slag respectively
and k is the overall decarburization rate constant given by:
Amr
K = }} (13.2.4)
W
where
A = reaction surface area
m = mass transfer coefficient
r = density of steel
W = weight of steel
It has been shown that for a CSTR the steady state or tap carbon content, %Ct, is given by7:
F [%Co ]
% Ct =
F+ k (13.2.5)
where, F is the specific feed or production rate which is the production rate divided by the weight
of steel in the vessel and %Co is the initial carbon content.
In attempting to partially convert a reactor with intense stirring into one with concentration differ-
ences like a PFR, barriers are sometimes installed with a small opening for metal flow as shown
schematically in Fig. 13.2. In this case, the tap carbon content is given by:
( F + b ) % CI
% Ct = (13.2.6)
F +b+ k
where b is the backmixing flow rate through the opening and %CI is the carbon content of steel in
Reactor I. In this case the W used in computing F is that of Reactor II.
13.3 Specific Alternative Steelmaking Processes

In this section, selected alternative steelmaking processes will be reviewed. A brief description,
current status and an evaluation of the processes is given. The technical evaluation addresses issues
such as productivity, yield and refining capability. First batch processes are discussed followed by
CSTR and PFR. The processes considered are summarized in Table 13.1 with respect to type, typ-
ical charge make-up and current status.
O2 O2
Metal Feed Off

gas
Slag Slag
Slag
tap
O2 O2 O2 O2
Metal
Reactor I Reactor II tap
Metal flow
1%C Þ P¢ + QB
0.05% Ü QB
Fig. 13.2 Schematic diagram of a trough process with a barrier.
Table 13.1 Classification and Status of Steelmaking Processes.
Process Type Typical Charge Current Status
BOF batch 80% hot metal 65% of world production

20% scrap
EAF batch 70–100% scrap 35% of world production
0–30% other
EOF (Brazil) batch 50% hot metal commercial
50% scrap
IRSID (France) CSTR* 80% hot metal pilot plant in 1970s
AISI (U.S.) CSTR hot metal pilot plant in 1980s
WORCA (Australia) PFR** hot metal pilot plant in 1970s
Bethlehem (U.S) PFR hot metal pilot plant in 1970s
NRIM (Japan) PFR hot metal pilot plant in 1970s
COSMOS (U.S.) PFR hot metal pilot scale in 1980s
IFCON (South Africa) PFR ore, coal pilot plant in 1990s
Carbide to steel CSTR; PFR iron carbide concept
* CSTR = continuous stirred tank reactor; ** PFR = plug flow reactor.
13.3.1 Energy Optimizing Furnace (EOF)

The EOF is essentially a batch oxygen steelmaking process with high post-combustion, with coal
additions and extensive scrap preheating; a schematic diagram is shown in Fig. 13.3.9 The process
theoretically can charge up to 100% scrap, but typically 40–60% scrap is used. Hot metal is
charged into the vessel followed by scrap from the lower preheating chamber. In some cases, a sec-
ond scrap charge is added about five minutes into the blow. The scrap is preheated in a series of
preheat chambers to about 800 to 1200°C. For a 50/50 scrap to hot metal mix about 70 m3/ton of
9
1
10
3 1. Scrap charging
2. Scrap, cold
3. Scrap, preheater
4. Scrap, preheated
5. Combustion air, preheated
4 11 6. Oxy fuel burners
7. Submerged O2-Tuyeres
8. Coal injection
9. Combustion air, cold
10. Recuperator
12 11. Combustion air, preheated
12. Water cooled elements
13. Additional oxygen
5 14. Furnace vessel
15. Bottom tapping
13
6
7
14
8
15
Fig. 13.3 Schematic diagram of the EOF process.
oxygen, with about one third into the metal and two thirds above the bath for post-combustion, and
about 20 kg/ton of coal is used. The post-combustion gas is used to preheat the scrap and, there-
fore, the energy from post-combustion does not have to be completely transferred to the bath. The
blowing time is about 27–32 minutes and tap-to-tap is less than one hour. Refractory consumption
increases with scrap usage and is approximately 2.5 kg per ton for a 50/50 mix.
An EOF furnace has been operating in Pains Brazil for over 10 years and has produced up to 0.5
mtpy from two 30 ton furnaces. There is an 80 ton furnace at Tata Steel in India and a 60 ton fur-
nace at AFS in Italy. The process has been reasonably well proven using a scrap charge up to 60%.
The productivity of an EOF is similar to that for a state of the art EAF and can produce a 100 ton
heat in an hour. The yield with a 50/50 charge mix is 92–94% which is similar to a BOF. Nitrogen
contents of 30–40 ppm can be achieved and phosphorus levels are similar to a BOF. Residuals
depend on the amount and type of scrap and can be computed from a simple mass balance. Sulfur
can be a significant problem as coal is used in the process. Hot metal desulfurization alone is not
effective because of the high scrap rates and the use of coal. For the following conditions, per ton
of steel, the final sulfur level is 0.07% for a 50/50 mix.
(%S )
=4 25 kg/ton coal
[%S]
70 kg slag 0.7% S in coal
0.10% S in hot metal
0.025% S in scrap
Therefore, significant steel desulfurization is required. If the hot metal is desulfurized to 0.02%,
the final sulfur is 0.033% which is still high for most quality applications.
The EOF is a proven technology which allows for higher scrap melting than a BOF and reasonable
flexibility with respect to the charge mix. With a 400,000 tpa COREX, or bath smelter, an 80 ton
EOF with a 50/50 mix could produce over 750,000 tpa of steel.
Mention should be made of a number of converter based processes which have been developed to
melt more scrap than conventional oxygen steelmaking. The first of these was developed by
Klockner.10 The KMS melted increased amounts of scrap by injecting coal into an OBM (Q-BOP)
converter while the KS process could melt 100% scrap using a liquid metal heel. Nippon Steel 11
and Kawasaki Steel12 investigated similar types of processes. The Tula or Z-BOP process which
originated in Russia and used by ISCOR at its Newcastle plant and tested by Bethlehem Steel, uses
lump coal in a conventional top blown BOF to melt up to 100% scrap.13 When melting large
amounts of scrap there was excessive iron oxidation. However, none of these processes continued
on a commercial basis and do not appear as efficient as the EOF which maximizes post-combus-
tion and uses scrap preheating.
13.3.2 AISI Continuous Refining

As part of the AISI-DOE direct steelmaking project a continuous steelmaking process was evalu-
ated and a limited number of pilot scale tests were run. The process is described and evaluated in
detail by Abel et al 7. A schematic diagram is shown in Fig. 13.4. Hot metal and possibly scrap
along with the required fluxes are added continuously to the top of the converter, oxygen is blown
through the bottom using OBM (Q-BOP) type tuyeres and tapped continuously.
The AISI converter is essentially a CSTR and equation 13.2.5 can be used to relate the final carbon
to the production rate. For example, for a 100 ton reactor processing 80% hot metal and 20% scrap
producing 0.05%C could produce about 100 ton per hour. A value of 60 hr–1 (1.0 min.–1) was used
for k which is typical for bottom blowing processes. However, as discussed in detail elsewhere7 since
decarburization is carried out at low carbon contents a large percentage of the oxygen oxidizes iron
resulting in large yield losses (>25%). For this reason, the process development was terminated.
Flux estimates
1.65 kg/min CaO
1.65 kg/min Dolomite
Slag chemistry estimates
50% FeO
32% CaO
512 Liters metal 8% MgO
91 Liters slag 5% MnO
Aim input conditions
1.5 - 4.5 %C 5663 Liters/min O2
0.040 %P Aim tap conditions
0.095 %S < 0.100 %C
1454 °C 33.9cm Slag 0.010 %P
106-227kg/min 0.005 %S
O2 gas 1650 °C
136-227 kg/min
88.9cm
Metal
86.4cm
Fig. 13.4 Schematic diagram of the 4 ton AISI continuous refining process.
As part of the AISI project a two zone smelter was investigated both in the pilot plant and theoret-
ically. The basic concept is similar to that shown in Fig. 13.2. The concept is to reduce the carbon
content in Reactor I to approximately 1% followed by final decarburization in Reactor II with the
excess FeO in the slag flowing to Reactor I where it is reduced. A critical concern is the size of the
opening in the barrier to avoid excessive back mixing. As an example, consider a 300 t/hr furnace
consisting of two 100 ton CSTR separated by the barrier and the carbon content is reduced to 1%
in Reactor I. To produce steel with 0.05% C, backmixing according to equation 13.2.6 must be lim-
ited to 20 t/hr (b = 0.2) Pilot plant work and water modeling by Zhang et al (14) indicates the area
of the opening must be limited to about 0.1 m2 or for a square opening 31.6 cm on a side. To main-
tain the barrier with such a small opening would be extremely difficult to do in practice.
Another problem is that there is excessive iron oxidation in Reactor II. In theory it is reduced in
Reactor I supplying 75% of the oxygen required for decarburization. The reduction reaction 13.3.1
is endothermic and with the reduced oxygen gas requirement slag-metal mixing may be insuffi-
cient for rapid refining.
(FeO) + C = CO + Fe (13.3.1)
Reactor I would be operating more like an open hearth type furnace. Based on the anticipated prob-
lems with the barrier, control and the nature of the reactions in Reactor I work on the development
of this process was also terminated.
13.3.3 IRSID Continuous Steelmaking

Of the existing continuous steelmaking processes the IRSID process was the most extensively
developed and is one of the most promising15. The process was tested on a 10–12 , ton per hour
scale at IRSID at Maizieres-Les-Metz and up to 25 tons per hour at Hagondange. A schematic dia-
gram of the process is shown in Fig. 13.5. The process is described elsewhere in detail.15 Briefly,
liquid hot metal is poured into the reactor at a controlled rate and an oxygen jet impinges onto the
stream forming an emulsion of gas, slag and metal. The emulsion overflows into a decanter in
which slag-metal separation takes place.
Exhaust
gas
O2 and lime
Decanting Slagging off

vessel
Cooling
additions
Reactor
Hot metal Raw

inlet Pumping steel
effect
Complex Phase:
Slag-Metal-Gas
Fig. 13.5 Schematic diagram of the IRSID continuous steelmaking process.
In normal batch oxygen steelmaking, such as the BOF, the oxygen jet and reaction gases cause an
extensive slag-metal emulsion with about 30–35% of the metal in the emulsion. The rate of decarbur-
ization at high carbon levels is limited simply by oxygen blowing rates. Below some critical carbon
content, about 0.3%, the rate is limited by liquid phase mass transfer of carbon in the metal. The
rate of mass transfer is first order with respect to carbon content. The metal drops in the emulsion
are decarburized to some extent and then enter the liquid metal pool in the bottom of the furnace.
The average carbon content of the metal drops is slightly less than in the bulk metal and, therefore,
the rate of decarburization can be related to the bulk metal concentration. The production rate for
a given furnace is limited by the rate of gas removal. That is, there is a free volume requirement
for a given rate of gas generation.
The reaction mechanism for the IRSID process is not known. As the metal enters the reactor it is
hit with an oxygen jet causing nearly all of the metal to be emulsified. Observations indicate there
is no liquid metal pool. In this case the metal drops remain in the emulsion during their entire
retention time in the reactor. The rate depends on how long the drops are in the reactor since their
carbon content is changing in time. The metal which overflows into the decanter is the average car-
bon content of the metal in the reactor. The retention time of the metal drops in the reactor depends
on the reactor size.
Obviously the IRSID process is more complex than a normal BOF and no model exists for the
process. In order to extrapolate the results obtained by IRSID and evaluate the process a reason-
able model is essential. The assumptions for the present model are given below:
1. The rate of decarburization of the metal drop is controlled by liquid phase mass
transfer of carbon.
2. There is always sufficient oxygen present as oxygen gas or FeO in the slag.
3. The metal drops react independently and do not combine.
4. It is recognized that there is a distribution of metal drops in the emulsion that are
decarburized at different rates due to differing area to volume ratios. However, it is
assumed that the average rate of decarburization for the weighted average size
droplets adequately describes the system. The Sauter diameter d32 should be used in
such an analysis.
å d 3L
d 32 = i
(13.3.2)
å d 2i
i
5. The carbon content of the metal leaving the reactor is the average of all the drops in
the emulsion.
6. There is sufficient slag to hold all of the metal in the emulsion.
The rate of decarburization of a single droplet is controlled by mass transfer and the carbon con-
tent is given by:
CD = Co e- kt (13.3.3)
where,
CD = carbon content of a drop after time t
Co is the initial carbon content
k is the overall rate constant for mass transfer given by:
A D rm
k= (13.3.4)
WD
AD = area of drop
m = mass transfer coefficient
r = density of the metal
WD = weight of the drop
The reactor is not plug flow but rather a continuous stirred tank reactor (CSTR). That is, some of
the particles only react for a short time and may have a high carbon content while others are very
low in carbon. For a CSTR the residence time distribution f(t) is given by:
1 -t /t r
f (t ) = e (13.3.5)
tr
where f(t) is the fraction of drops in the bath for time t and tr is the average residence time. The
average or final carbon content of the metal leaving the reactor is given by:
¥
C= ò0 C ( t )f ( t )dt (13.3.6)
CO æ 1ö
C= òC -ç k + ÷ t (13.3.7)
O
e è trø
Co
C= (13.3.8)
tr k + 1
Equation 13.3.8 is a very simple but useful expression for evaluating the IRSID process. It is inter-
esting to note that the rate of decarburization is reasonably fast and most of the drops are fully
decarburized. The carbon in the final metal results primarily from the drops which only spend a
short time in the reactor.
For typical oxygen blowing rates and slag volumes in a BOF the value of k is about 0.017 s–1. In
comparison to the BOF in the IRSID process all of the metal is in the emulsion. This causes the
reaction area (A) to increase correspondingly. With all the metal rather than 30% in the emulsion
the value of k is estimated to be a factor of about 3.3 higher or about 0.55 s–1.
It is also necessary to estimate the retention time (tr). It has been found that there is a required
reactor volume to hold the slag metal emulsion in a BOF which depends on the rate of CO for-
mation. The reactor volume at any time is filled with slag, metal and gas. In a normal BOF, 70%
of the emulsion is gas. In the IRSID process, it may be even higher due to the faster rate of gas
generation. If the gas occupies 70% of the volume, the retention time of the metal and slag is
given by:
0.3 VR
tr = (13.3.9)
Vs + V
˙ ˙m
where,
V. R = reactor volume
V.s = volume flow rate of slag
Vm = volume flow rate of metal
For example, consider a 10 m3 reactor producing 100 t/hr of steel and 75 kg of slag per ton, the
retention time is about 10.75 minutes (645 s). Using equation 13.3.7, and 0.055 s–1 for k and an
average initial carbon of the scrap plus metal of 4%, the predicted final carbon content is about
0.1%. It should be noted that this is only a crude estimate. However, it does agree with the pilot
plant results at IRSID in which in a smaller scale unit the final carbon content was about 0.08%.
The productivity of the IRSID process obviously depends on the reactor size but 200 t per hour
should be easily achieved. The yield and refining for sulfur phosphorus, etc., should be similar to
a BOF. Scrap can be melted but must be added continuously and sized so it melts rapidly. Metal
feeding rate is critical in order to match the oxygen input. IRSID used a fairly crude but effective
metal feeding system over 20 years ago. With the advent of slide gates and other metal flow con-
trol devices adequate metal flow control should be possible. Due to the high gas generation rate
there could be a large amount of dust generation. However, some of the dust should settle out in
the decanter. Refractory wear in the reactor may be a concern due to the violent reactions for the
emulsion overflow. Refractory consumption is expected to be significantly higher than for a BOF
with slag splashing.
13.3.4 Trough Process

From about 1970 to 1983 a number of continuous steelmaking processes were suggested or tested
on a pilot scale in which a long trough was used in which liquid hot metal was added to one end,
a series of oxygen lances or tuyeres, placed longitudinal, were used for decarburization and the
metal was tapped on the other end. The slag in some cases flows counter current to the metal. The
processes which were tested on a pilot scale included the Bethlehem process16, the NRIM (Japan)
process17, the WORCRA process18 and the COSMOS (U.S.Steel) process19. These processes all
reached the pilot scale at production rates generally of about 5 to 12 tons per hour.
The processes were all somewhat similar and a schematic of the WORCRA process is shown in
Fig. 13.6. The NRIM used a series of trough furnaces with desiliconization and partial dephos-
phorization carried out in the first reactor, decarburization and final dephosphorization in the sec-
ond and grading and alloying in the third reactor. The slag from each furnace is removed separately.
COSMOS used bottom OBM (Q-BOP) tuyeres instead of lances.
As noted earlier, to avoid backmixing the length to width ratio (L/W) should be around 50. None
of the pilot facilities had this ratio and, therefore, did not act as a perfect PFR, but had significant
backmixing. Even with modest values (L/W) there was excessive heat losses due to the large sur-
face to volume ratio and relatively low production rates. Calculations indicate that even at reason-
able production rates the heat losses for these processes would be significantly higher than for
other processes.
Flux O2
additions lances
Inspection and
sampling doors
Hot metal
from Cupola
Steel
Refining Slag dam

Bowl zone
Scrap
coolant
Slag
Fig.13.6 Sectional plan and elevation of L-shaped pilot plant WORCRA steelmaking furnace.
Other problems associated with these processes include metal feed control, the necessity to size
scrap and other coolants, lower phosphorus removal due to inadequate slag metal mixing and the
difficulty in producing low carbon steels (<0.06%).
It should be noted that these type processes had some success in nonferrous metallurgy. Noranda20
developed a process for copper using a 21 m long reactor with 16 tuyeres. An analysis of the process
indicates that the reactor worked as 6–7 CSTR reactors in series and was approaching a PFR. The
QSL process21 for lead has also been commercialized. These processes operate at lower tempera-
tures than for steelmaking and, therefore, the high heat losses and refractory with this geometry is
less critical. Also, the production rate is low compared to that required for steelmaking.
13.3.5 Other Steelmaking Alternatives

Several other alternative steelmaking processes have been proposed or reached a pilot stage. Two
iron carbide continuous steelmaking processes have been proposed. In the original conceptual
process, iron carbide is refined in a trough type reactor with extensive post-combustion supplying
much of the energy22. This process has not been tested on any scale. It would suffer from all of the
drawbacks of the other trough processes and the need for very high degrees of post-combustion
which have not been achieved in any other process. A second process based on iron carbide, “Car-
bide to Steel” has been proposed by G. Geiger23 in conjunction with U.S.Steel. They concluded that
the original concept of a QSL type trough reactor would not be feasible because the CO evolution,
from the unreacted FeO in the carbide and from decarburization, would be so great it would be
impossible to maintain a carbon concentration gradient. They concluded it was more attractive to
divide the process into two stages shown schematically in Fig. 13.7. In the first vessel, the Fe3C
would be injected and melted with carbon oxidation providing part of the energy. In the first reac-
tor, the carbon content is reduced to about 2% and then tapped into a second reactor which is a
trough reactor in which final decarburization takes place. The CO from the second reactor is trans-
ferred to the first and along with the CO from the first reactor is post-combusted to CO2 to pro-
vide most of the energy for reduction of the remaining FeO in the carbide and melting. They
Lime
feed
bin
O2
Iron CO
carbide
Liquid
Fe-C Stage II Reactor
metal
Stage I Reactor
Launder
O2 Slag pit
O2
Fig. 13.7 Conceptual flow sheet for carbide to steel.
Ore
coal
O2 for post O2 for post

combustion combustion
Coal-char-ore
Slag
Steel
Channel inductors
Fig 13.8 Schematic diagram of the IFCON process.
concluded that in order to avoid excessive heat losses the smallest feasible plant would be 45
tons/hour. They invited Nucor to help design such a plant. Whereas the process is theoretically
attractive, it is only a concept and to date (1998) no steel has been produced. It requires higher lev-
els of post-combustion than have even been achieved in other processes and the availability of iron
carbide at a reasonable cost. If such a process was possible it could be used for other feeds such as
HBI or DRI plus carbon, granulated pig iron or liquid hot metal and DRI or HBI.
Another extremely interesting, and possibly revolutionary process, is the IFCON process under
development by ISCOR in South Africa. Little public information is available except for its patent.24
A schematic diagram is shown in Fig. 13.8. The process utilizes a channel type induction furnace
similar to that used in hot metal mixers. According to the patent, the process can use hot metal and
partially reduced sponge iron (direct reduced iron), sponge iron only and, even more interestingly,
iron ore and coal to produce steel directly. The latter version of the process would be truly revolu-
tionary and satisfy the long time dream of steelmakers to go from ore to steel in a single reactor.
According to the ISCOR patent24, when using ore, the ore and coal are added to the furnace con-
tinuously. Oxygen is added primarily for post-combustion of the CO from reduction and hydrogen
from the coal. Post-combustion supplies a significant amount of energy for reduction. Electrical
energy is supplied by the channel inductors for melting the reduced iron. ISCOR has operated a
pilot plant and is considering a commercial/demonstration plant. No operating information from the
pilot plant is available. However, a small pilot plant has been in operation at Pretoria and has run
continuously for several campaigns of over one month each and produced steel with 0.03–0.1%C.
A semi-commercial plant producing approximately 300,000 tpy is scheduled to begin operation in
1998 in South Africa. If such a process proves feasible it could drastically reduce the capital cost
associated with steelmaking by eliminating the coke and sinter plants and combining iron and steel-
making in one vessel.
13.4 Economic Evaluation

For a new process to be commercialized it not only must be technically feasible, it must have a clear
economic or strategic advantage over existing proven processes. Also, when comparing processes
using different input materials it is the only way a meaningful comparison can be made. All eco-
nomic evaluations for steelmaking processes are site specific. Furthermore, other considerations
beyond cost alone, such as the availability of coke, hot metal and electricity and the type or qual-
ity of steel for the market application are important. Nevertheless, an economic evaluation is valu-
able as long as the assumption and unit costs used are reasonable.
The processes which appear to be the most promising are the EOF and the IRSID continuous
process. All of the other processes considered have significant technical drawbacks. These two
processes are compared to a BOF and an EAF.
A major concern is capital costs which are also site specific and certain assumptions must be made.
The assumptions in the present case are:
1. Annual production of 800,000–1,000,000 tons.
2. The building exists.
3. Hot metal and steel ladles and cranes exist.
4. Oxygen and other utilities are at the battery boundary.
It is beyond the expertise of the authors to make accurate capital cost estimates. The capital costs for
the various processes given in Table 13.2 were made by experts for the American Iron and Steel Insti-
tute (AISI)1. There is a relatively large range of capital cost for the EOF and IRSID processes because
there is greater uncertainty since few or no units have been built. Whereas actual capital costs are very
site specific the values given in Table 13.2 are for a similar site. Therefore, whereas the actual capi-
tal costs may be significantly different, the relative cost for the processes should be reasonable.
Table 13.2 Estimated Capital Cost for Selected Steelmaking Units in Dollars per
Annual Ton of Production.
Process
________ Dollars per annual ton
_____________________
BOF 80–86
EAF(1) 90–100
EOF 40–60
IRSID 40–60
(1)When using large quantities of hot metal the capital cost for the EAF will be about 10% higher due to the oxygen lances and gas
cleaning system.
Estimated cost of production was made for several cases. Production costs are sensitive to the unit
cost of the inputs including scrap, hot metal, labor, etc. The unit costs will be a function of plant
location and time and therefore can vary greatly. The analysis used was for the United States in
1998. However, as will be demonstrated, the major variables are the cost of scrap versus hot metal
and the cost of capital.
For the BOF and IRSID processes, which use primarily hot metal, the charge was assumed to be
80% hot metal and 20% scrap. The EOF is for 50% scrap and 50% hot metal. The EAF estimates
are for scrap plus 50% DRI/HBI or 50% hot metal. These charge balances are for the production
of medium to high quality flat rolled which would be a likely product for a new plant. Materials
consumption for the BOF was a typical state of the art process. For the EAF, the HBI/DRI was
assumed to be 92% total iron at 92% metallization, 2% gangue, and 2% carbon. The energy and
materials consumptions were calculated from a comprehensive energy and materials model for the
EAF. The EOF materials consumption were from an operating furnace which were verified with an
energy and materials balance. The desulfurization cost for the EOF are high because of the use of
240
EAF
230
EOF
220 BOF
$/ ton steel
IRSID
210
200
190
100 110 120 130 140 150 160 170
$/ton scrap
Fig.13.9 Cost of liquid steel for various processes as a function of scrap price.
coal or coke for energy to melt the scrap. The BOF refractory costs are lower due to slag splash-
ing which extends refractory life. Other costs include dust handling and electrodes which are
higher for an EAF. The BOF and IRSID processes have slightly higher labor costs due to the charg-
ing and tapping operations. The capital costs were computed assuming 8% depreciation and 8%
return on investment or a total of 16% of the figures given in Table 13.2.
The most sensitive input is the cost of scrap. Therefore, a sensitivity analysis is done for the cost
and amount of scrap. In the past decade the price of scrap has varied by more than 50%. The price
of scrap given in Table 13.3 is the weighted average price of a typical charge mix for higher value
products. A sensitivity analysis was performed in which the price of scrap varied from $100 to
$160 per ton. The results are given in Fig. 13.9. As expected, the scrap based processes (EAF and
EOF) costs decrease with decreasing scrap prices and at $100 per ton of scrap their costs are equal
or less than those using less scrap.
The other major uncertainty is the cost of hot metal. Obviously an increase in hot metal costs will
adversely affect the economics of those processes using hot metal and its effect can be easily esti-
mated. Therefore, comparisons of processes which use differing amounts of hot metal and scrap
should be made only if precise information on the cost of the charge materials are known.
The IRSID process has lower capital cost. A BOF is actually only making steel about half the time
due to the time required for charging and tapping whereas the IRSID process is continuous and,
therefore, has a higher productivity to volume ratio. Also, the IRSID process requires less capital
because it has a permanent oxygen lance and does not require vessel rotation equipment. In the late
1970s, The United Nations25 examined continuous steelmaking processes studied up to that time.
They also concluded that the IRSID process was the most attractive alternative and potentially had
significantly lower capital cost than a conventional BOF.
The EOF has lower capital cost than the EAF because it does not require a transformer, electrode
mechanism, etc. It has a lower operating cost because of lower energy costs. However, it does
require a source of hot metal which is currently unavailable economically from a blast furnace. It
would require a COREX or new smelting ironmaking process which can produce iron economi-
cally at low production rates. The EAF could use hot metal reducing its energy cost. However, hot
metal is more costly and not usually available for an EAF. In comparing an EOF and EAF with
50% hot metal, the EOF cost may be lower because of lower capital but the EOF is less flexible.
Table 13.3 Operating and Total Cost of Producing Liquid Steel.
BOF IRSID EOF EAF EAF

Charge Mix 80% Hot Metal 80% Hot Metal 50% Scrap 50% Scrap 50% Scrap
20% Scrap 20% Scrap 50% Hot Metal 50% DRI 50% Hot Metal
Input (cost/unit) Units Cost Units Cost Units Cost Units Cost Units Cost
Hot Metal ($145/t) 0.85 123.30 0.85 123.30 0.55 79.70 0 0.54 64.80
Scrap ($120/t) 0.22 27.50 0.22 27.50 0.55 66.00 0.541 64.90 0.54 64.80
DRI/HBI ($130/t) 0.541 70.30
Flux ($55/t) 0.055 3.00 0.055 3.00 0.04 2.20 0.036 2.00 0.03 1.70
Oxygen ($40/t) 0.065 2.60 0.065 2.60 0.10 4.00 0.035 1.40 0.06 2.40
Carbon ($70/t) 0.025 1.80 0.01 0.70
13.5 Summary and Conclusions

Electricity ($0.04/kwh) 65 2.60 40 1.60 20 0.80 485 19.40 210 8.40
Refractories (kg/t) 1 1.00 2.5 2.50 2.5 2.50 1.5 1.50 1.5 1.50
Labor ($30/mh) 0.3 9.00 0.3 9.00 0.2 6.00 0.2 6.00 0.2 6.00
Desulfurization 7.00 7.00 12.00 7.00 7.00
Other(1) 20.00 20.00 20.00 25.00 25.00

Total Operating
Cost 196.00 196.50 195.00 198.20 195.10
Cost of Capital at
8% depreciation
and 8% ROI 13.30 8.00 9.00 15.20 16.70
Total Cost 209.30 204.50 204.00 213.40 211.80
(1)
Other includes electrodes, gas cleaning and dust disposal.
Copyright © 1998, The AISE Steel Foundation, Pittsburgh, PA. All rights reserved.
virgin iron and resemble a hybrid process between the EAF and the BOF as discussed in Chapter 10.
and will continue to change significantly. When one is considering steelmaking processes the EAF
and the BOF may not necessarily be the only options. In particular the EAF will use more oxygen and
Economic and strategic considerations with regards to steelmaking have changed in the past decade
757
In this chapter a technical evaluation of proposed and operating alternatives to a purely oxygen steel-
making process was conducted. The evaluation indicated that the IRSID continuous steelmaking
process may be an attractive alternative to a traditional BOF and the EOF could be an attractive option
when high scrap charge levels (50%) are required. The other processes considered, including the AISI
continuous process and the various trough processes, have significant technical problems.
An economic analysis was performed to compare the costs of the BOF, EAF, IRSID and EOF
processes. Whereas, any economic analysis is site specific and depends on the cost of the various
inputs, in particular scrap, the present analysis indicates that the IRSID and EOF processes are
competitive and may have an advantage over existing processes, the primary advantage being in
lower capital cost. The continuous nature of the IRSID process allows for greater productivity and
the fixed vessel and lance reduces its capital cost. The EOF is capable of melting scrap without the
cost of the electrical energy delivery system of the EAF and results in lower operating energy costs.
However, because of its high degree of development, optimization and capital investment it is
doubtful these will replace existing OSM plants.
In addition to the IRSID and EOF processes other processes under development should also be
considered. In particular, the IFCON process holds much promise as it purports to produce steel
directly from fine raw materials, thereby significantly reducing capital cost. More development
work may be required on the new processes but there is the potential of significant cost savings.
References
1. R. J. Fruehan and C. L. Nassaralla, “A Critical Review of Alternative Steelmaking
Processes”, Trans ISS, Iron & Steelmaker, Vol. 25, No.8, (1998): 59–68.
2. J. Szekely and N. J. Themelis, Rate Phenomena in Process Metallurgy (New York: Wiley
Interscience, 1971).
3. J. Szekely, “Proc. of Heat and Mass Transfer in Process Metallurgy,” I.M.M., London 1966,
p. 115.
4. T. W. Jenkins, N. B. Grey, and H. K. Worner, Metall. Trans. 2 (1971): 1258.
5. G. A. Brooks, N. G. Ross, and H. K. Worner, 80th Steelmaking Conference proceedings,
ISS-AIME, Chicago IL, 1997.
6. R. J. Fruehan: Iron and Steelmaker 1(1) (1976): 33.
7. C. A. Abel, R. J. Fruehan, and A. Vassilicos, Trans. ISS, I&SM (August 1995): 49.
8. G. Taylor, Proces. Ray Soc. 223A (1954): 446.
9. W. A. Tony, Iron and Steelmaker 15(9) (September 1988): 41.
10. K. Brotzman, 70th Steelmaking Conference Proceedings, ISS-AIME, Pittsburgh 1987, p. 3.
11. H. Ozawa et al., Proceedings of the Sixth International Iron and Steel Congress, 1990,
Nagoya Japan, ISIJ, pp. 33.
12. Y. Kishimoto et al., Proceedings of the Sixth International Iron and Steel Congress, 1990,
Nagoya Japan, ISIJ, pp. 41.
13. G. Golperine et al., 77th Steelmaking Conference Proceedings of ISS-AIME, Chicago IL,
1994, p. 73.
14. X. Zhang, R. J. Fruehan, and A. Vassilicos, “Modeling of the Two-Zone Smelter—Part 2:
Physical Modeling and Plant Trials,” Process Technology Conference Proceedings, Vol. 13,
1995, pp. 445-454.
15. A. Berthet et al., Journal of Iron and Steel Institute (June 1969): 790.
16. E. M. Rudzki et al., J of Metals 21 (1969): 57.
17. R. Nakagawa: Trans. ISIJ 13 (1973): 333.
18. H. K. Warner et al., J of Metals 21 (1969): 50.
19. R. J. Zaranek (USX Technical Center), personal communication, Monroeville PA.
20. P. Tarasoff: Met. Trans. B 15B (1984): 411.
21. P. Arthur, A. Siegmund, and M. Schmitt, Savard/Lee International Symposium, TMS, Mon-
treal, 1992, TMS, Warrendale PA, p. 668.
22. G. Gieger: U.S. Patent No. 5,139,568.

23. G. Gieger: “Carbide to Steel”: to be published by TMS-AIME, Warrendale, PA.
24. L. J. Fourie, Steelmaking Process, U.S. Patent 5,411,570, May 2, 1995.
25. The United Nations. The Increasing Use of Continuous Processes in the Iron and Steel
Industry. 1979. Sales No. E79IIE-7.

Chapter 13

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 13

Uploaded by

Copyright:

Available Formats

Chapter 13

Alternative Oxygen Steelmaking

13.2 General Principles and Process Types

Batch CSTR PFR

Mixing } complete complete only in dV

Concentration none none longitudinal

Concentration with time constant with distance

Fig. 13.1 Classification of steelmaking process types and their characteristics.

13.3 Specific Alternative Steelmaking Processes

Metal Feed Off

Fig. 13.2 Schematic diagram of a trough process with a barrier.

Table 13.1 Classification and Status of Steelmaking Processes.

Process Type Typical Charge Current Status

BOF batch 80% hot metal 65% of world production

* CSTR = continuous stirred tank reactor; ** PFR = plug flow reactor.

13.3.1 Energy Optimizing Furnace (EOF)

13.3.2 AISI Continuous Refining

13.3.3 IRSID Continuous Steelmaking

Decanting Slagging off

Hot metal Raw

Fig. 13.5 Schematic diagram of the IRSID continuous steelmaking process.

13.3.4 Trough Process

Refining Slag dam

13.3.5 Other Steelmaking Alternatives

Fig. 13.7 Conceptual flow sheet for carbide to steel.

O2 for post O2 for post

13.4 Economic Evaluation

BOF IRSID EOF EAF EAF

13.5 Summary and Conclusions

Other(1) 20.00 20.00 20.00 25.00 25.00

22. G. Gieger: U.S. Patent No. 5,139,568.

You might also like