United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-453/R-94-065
Final Report
September 1994
&EPA ALTERNATIVE CONTROL
TECHNIQUES DOCUMENT--
NO,, EMISSIONS FROM
IRON AND STEEL MILLS
-------
-------
EPA-453/R-94-065
ALTERNATIVE CONTROL
TECHNIQUES DOCUMENT-
NOX EMISSIONS FROM IRON AND STEEL MILLS
Emission Standards Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1994
-------
-------
ALTERNATIVE CONTROL TECHNIQUES DOCUMENTS
This report is issued by the Emission Standards Division,
Office of Air Quality Planning and Standards, U. S. Environmental
Protection Agency, to provide information to State and local air
pollution control agencies. Mention of trade names and commercial
products is not intended to constitute endorsement or
recommendation for use. Copies of this report are available --as
supplies last-- from the Library Services Office (MD-35),' U. S.
Environmental Protection Agency, Research Triangle Park, North
Carolina 27711 ([919] 541-5514) or, for a nominal fee, from the
National Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161 ([800] 553-NTIS).
-------
-------
TABLE OF CONTENTS
Chapter
Page
LIST OF FIGURES iv
LIST OF TABLES • . ' v
1 INTRODUCTION 1-1
2 SUMMARY ............ 2-1
2.1 SOURCES OF NOX EMISSIONS . . . . . . .. . . 2-1
2.2 UNCONTROLLED NOX EMISSIONS 2-3
2.3 NOX EMISSIONS REDUCTION TECHNIQUES . . . . .2-4
2.4 COSTS AND COST EFFECTIVENESS OF NOX CONTROL
TECHNIQUES ....2-8
2.5 IMPACTS OF NOX CONTROLS . 2-8
2.5.1 Air Impacts of Combustion
Modification Techniques . . . . . . 2-12
2.5.2 Air Impacts of SCR and SNCR ..... 2-13
2.5.3 Solid Waste Impacts . 2-13
2.5.4 Energy Impacts 2-14
3 INDUSTRY DESCRIPTION . . 3-1
3.1 BACKGROUND 3-1
3.1.1 A Historical Perspective . . . . . . 3-2
3.1.2 An Overview of the U.S. Iron
and Steel Industry . .3-2
3.1.2.1 Integrated Producers . . . .3-5
3.1.2.2 Mini-Mills 3-5
3.1.2.3 Specialty Producers . . . . 3-5
3.2 INDUSTRY CHARACTERIZATION ......... 3 - 8
3.2.1 An Overview of Iron and Steel
Manufacturing 3-8
3.2.1.1 Raw Materials and Preparatory
Processes 3-8
3.2.1.2 Ironmaking . . . . . . . . . 3-10
3.2.1.3 Steelmaking . 3-11
3.2.1.4 Finishing 3-14
3.3 PRODUCTION PROCESSES WITH NOX EMISSIONS . . 3-15
3.3.1 Sintering ..... .3-15
3.3.2 Cokemaking ..... 3-17
3.3.3 Blast Furnaces and Blast Furnace
Stoves . • . . 3-18
3.3.4 Basic Oxygen Furnaces . ... . . '. . . .3-19
3.3.5 Soaking-Pit and Reheat Furnaces . . 3-20
3.3.6 Annealing Furnaces . . . 3-23
3.4 REFERENCES 3-24
4 UNCONTROLLED NOX EMISSIONS .4-1
4.1 MECHANISMS OF NOX FORMATION 4-1
-------
TABLE OF CONTENTS (con.)
4.1.1 Thermal NOX Formation 4-1
4.1.2 Fuel NOX Formation ." . . . . . . . . . . 4-4
4.1.3 Prompt NOX Formation .......... 4-6
4.2 EMISSIONS OF NOX FROM IRON AND STEEL MILLS . 4-7
4.2.1 Emission Factors . . . . . . . . . .4-7
4.2.2 Coke-oven Underfiring . . . . . . . .4-7
4.2.3 Sintering . . . . . . . . . . . . . .4-9
4.2.4 Blast Furnaces and Blast
Furnace Stoves . . . . .4-10
4.2.5 Basic Oxygen Furnace . . . . . . . .4-10
4.2.6 Electric-arc Furnace . . . . . . . .4-11
4.2.7 Soaking Pits 4-12
4.2.8 Reheat Furnaces 4-i3
4.2.9 Annealing Furnaces . ... 4-14
4.2.10 Galvanizing Furnaces ... 4-15
4.2.11 Summary . . . 4-15
4.3 REFERENCES 4-18
5 NITROGEN OXIDES CONTROL TECHNIQUES 5-1
5.1 BACKGROUND . 5 -1
5.2 CONTROL TECHNIQUES APPLICABLE TO IRON AND STEEL
FACILITIES 5-1
5.2.1 Control Techniques Applied . . ; . .5-1
5.2.1.1 Low Excess Air . . . > . . .5-3
5.2.1.2 Low NOX Burners . . . . .. . 5-3
/ 5.2.1.3 Low NOX Burner Plus Flue Gas
Recirculation .5-8
5.2.1.4 Selective Catalytic
Reduction . , 5-8
5.2.2 Other Control Techniques 5-11
5.2.2.1 Off-Stoichiometric or Staged
Combustion . .5-11
5.2.2.2 Fuel Switching 5-16
5.2.2.3 Selective Noncatalytic
Reduction . 5-16
5.3 NITROGEN OXIDES CONTROLS FOR SPECIFIC
FACILITIES . 5-18
5.3.1 Coke Ovens 5-18
5.3.2 Sinter Plants . 5-19
5.3.3 Blast Furnaces and Blast
Furnace Stoves . 5-21
5..3.4 Basic Oxygen Furnace ........ 5-22
5.3.5 Electric Arc Furnaces . . . . . . . . 5-23
5.3.6 Reheat Furnaces . . . •'. .5-23
5.3.7 Soaking Pits ... 5-2-5
5.3.8 Annealing and Galvanizing 5-25
ii
-------
TABLE OF CONTENTS (con.)
5.3.8.1 Annealing Furnaces . . . . . 5-25
5.3.8.2 Galvanizing Furnaces . . . . 5-27
5.4 SUMMARY OF CONTROLLED NOX EMISSIONS . . . . 5-27
5.5 REFERENCES 5-29
6 COSTS OF NOX CONTROLS 6-1
6.1 STEEL REHEAT FURNACES 6-2
6.1.1 Model Furnaces . . . 6-2
6.1.2 Costs and Cost Effectiveness of LEA
Control for Reheat Furnaces . . 6-3
6.1.3 Costs and Cost Effectiveness of
LNB Control for Reheat Furnaces . . . 6-3
6.1.4 Costs and Cost Effectiveness of
LNB Plus FOR Control 6-5
6.2 ANNEALING FURNACES ... . 6-8
6.2.1 Model Furnaces .6-8
6.2.2 Costs and Cost Effectiveness of LNB
Control 6-11
6.2.3 Costs and Cost Effectiveness of LNB
plus FGR Control . . . .. .. . .. 6-9
6.2.4 Costs and Cost Effectiveness of
SNCR Control 6-9
6.2.5 Costs and Cost Effectiveness of SCR
Control of Annealing Furnaces .... 6-14
6.2.6 Costs and Cost Effectiveness of LNB plus
SNCR for Annealing Furnaces ... 6-14
6.2.7 Costs and Cost Effectiveness of LNB
plus SCR Controls for Annealing
Furnaces ...... 6-17
6.3 GALVANIZING FURNACES ..... 6-17
6.3.1 Model Furnaces 6-17
6.3.2 Costs and Cost Effectiveness of LNB
plus FGR Controls for Galvanizing
Furnaces .6-17
6.3.3 Costs and Cost Effectiveness of LNB
for Galvanizing Furnaces 6-19
6.4 REFERENCES ........... 6-22
7 ENVIRONMENTAL AND ENERGY IMPACTS OF NOX CONTROLS 7-1
7.1 INTRODUCTION . . 7 -1
7.2 -AIR IMPACTS . . . 7-1
7.2.1 NOX Emission Reductions ........ 7-1
7.2.2 Air Impacts of Combustion Modifications . 7-3
7.2.3 Air Impacts of SCR and SNCR 7-4
7.3 SOLID WASTE IMPACTS 7-6 .
7.4 ENERGY IMPACTS ......... 7-6
7.5 REFERENCES . 7-8
APPENDIX A TABULATION OF UNCONTROLLED NOX EMISSIONS DATA. .A-l
ill
-------
-------
LIST OF FIGURES
Number
2-1
3-1
3-2
5-1
5-2
5-3
5-4
5-5
5-6
5-7
Simplified steelmaking flow chart
Simplified steelmaking flow chart . . .
into major steel product forms ....
Schematic representation of.progress of
refining in a top-blown basic-lined EOF
Staged-air low NOX burner ......
Staged-fuel low NO^ burners
Illustration of flue gas recirculation
Illustration of the SCR Principle ". .
Biased burner firing .........
Burners out of service (BOOS)
Overfire air ......
Page
2-2
3-9
3-9
3-21
5-5
5-6
5-9
5-10
5-13
5-14
5-15
IV
-------
-------
LIST OP TABLES
Number
Page
2-1 SUMMARY OF UNCONTROLLED NOX EMISSIONS 2-5
2-2 NOX CONTROLLED EMISSIONS AND .
PERCENT REDUCTIONS ... 2-7
2-3 COSTS AND COST EFFECTIVENESS - EXISTING REHEAT FURNACES
(300 MMBtu/hr) . 2-8
2-4 COSTS AND COST EFFECTIVENESS - EXISTING ANNEALING
FURNACES (200 MMBtu/hr) ............. 2-9
2-5 COSTS AND COST EFFECTIVENESS - EXISTING GALVANIZING
FURNACES (100 MMBtu/hr} . 2-11
3-1 SUMMARY OF CURRENT STEEL INDUSTRY FACTS .... 3-4
3-2 SUMMARY INFORMATION RELEVANT TO INTEGRATED
STEEL PRODUCERS ........ 3-6
4-1 CALCULATED EQUILIBRIUM CONCENTRATIONS (in ppm)
OF NO AND N02 IN AIR AND FLUE GAS 4-3
4-2 FACTORS CONTROLLING THE FORMATION OF THERMAL NOX 4-5
4-3 NITROGEN OXIDE EMISSION FACTORS FROM AP-42
AND NAPAP 4-8
4-4 SUMMARY OF UNCONTROLLED NOX EMISSIONS 4-12 -
5-1 OVERVIEW OF NOX CONTROL TECHNIQUES 5-2
5-2 NOX CONTROLLED EMISSIONS AND PERCENT REDUCTION . 5-28
6-1 COSTS AND COST EFFECTIVENESS OF LEA FOR
REHEAT FURNACES. . '..... 6-4
6-2 COSTS AND COST EFFECTIVENESS OF LNB FOR
REHEAT FURNACES 6-7
6-3 COSTS AND COST EFFECTIVENESS OF LNB/FGR FOR
REHEAT FURNACES 6-7
6-4 COSTS AND COST EFFECTIVENESS OF LNB FOR
ANNEALING FURNACES 6-10
6-5 COSTS AND COST EFFECTIVENESS OF LNB/FGR FOR
ANNEALING FURNACES '. 6-11
6-6 COSTS AND COST EFFECTIVENESS OF SNCR FOR
ANNEALING FURNACES 6-13
6-7 COSTS AND COST EFFECTIVENESS OF SCR FOR
ANNEALING FURNACES 6-15
6-8 COSTS AND COST EFFECTIVENESS OF LNB+SNCR
FOR ANNEALING FURNACES . 6-16
6-9 -COSTS AND COST EFFECTIVENESS OF LNB+SCR FOR
ANNEALING FURNACES .... 6-18
6-10 COSTS AND COST EFFECTIVENESS OF LNB+FGR FOR
GALVANIZING FURNACES 6-20
6-11 COSTS AND COST EFFECTIVENESS OF LNB FOR
GALVANIZING FURNACES 6-21
7-1 NOX EMISSION REDUCTIONS--IRON AND STEEL
MILL PROCESSES 7-2
-------
-------
CHAPTER 1
INTRODUCTION
Congress, in the Clean Air Act Amendments of 1990 (CAAA),
amended Title I of the Clean Air Act (CAA) to address ozone
nonattainment areas. A new Subpart 2 was added to Part D of
Section 103. Section 183 (c) of the new Subpart 2 provides that:
[w]ithin 3 years after the date of the enactment of the
[CAAA], the Administrator shall issue technical documents
which identify alternative controls for all categories of
stationary sources of ... oxides of nitrogen which emit, or
have the potential to emit 25 tons per year or more of such
air pollutant.
These documents are to be subsequently revised and updated as
determined by the Administrator.
Iron and steel mills have been identified as a stationary
source category with emission sources that emit more than 25 tons
of nitrogen oxides (NOJ per year. This alternative control
techniques (ACT) document provides technical information on
various NOX controls for various iron and steel mill processes
that State and local agencies may use to develop and implement •
regulatory programs to control NOX emissions from iron and steel
mills. Additional ACT documents have been developed for eight
other stationary source categories.
The information in this ACT document was generated from
previous EPA documents and literature searches and contacts with
iron and steel manufacturers, engineering firms, control
equipment vendors, and Federal, State, and local regulatory-
agencies . Chapter 2 presents a summary of the findings of this
study. Chapter 3 provides a process description and industry
characterization of iron and steel mills. A discussion of
uncontrolled NOX emission levels is presented in Chapter 4.
Alternative control techniques and achievable controlled emission
levels are discussed in Chapter 5. Chapter 6 presents control
costs and cost effectiveness for each control technique.
Environmental and energy impacts associated with the use of NO,
control techniques are discussed in Chapter 7.
1-1
-------
-------
CHAPTER 2
SUMMARY
2.1 SOURCES OF NOX EMISSIONS
Integrated iron and steel mills produce steel by reducing
iron ore to iron in a blast furnace and, subsequently, removing
excess carbon and other impurities from the iron in a basic
oxygen furnace. Other processes involve beneficiating iron ore
(e.g., pelletizing), recycling of iron-bearing materials (e.g.,
sintering), coke-making, and steel finishing processes such as
shaping, annealing, and galvanizing. All of these are high
temperature processes, usually involving the combustion of fossil
fuels, and all are potential sources of NOX emissions.
Mini mills and specialty producers process steel through
some subset of the full range of processes found in integrated
iron and steel mills. Typically, they enter the process by
melting scrap steel in an electric arc furnace, bypassing the
iron-making process and attendant support activities such as -
sintering and coke making.
The processing order is illustrated in the simplified flow
diagram in Figure 2-1. In coke making, coal is destructively
distilled in coke ovens that are fired with, typically, a mixture
of coke oven gas (COG) and natural gas (NG) and/or blst furnace
gas(BFG). Coke oven underfiring is a high-temperature process
and NOX emissions from coke making are appreciable.
In the sinter plant, iron ore fines, coke fines, other
iron-bearing materials, and (often) flux are well-mixed and
spread uniformly on a traveling grate and ignited, typically with
NG. As the grate travels, a forced draft causes the coke fines
and other combustibles in the bed to burn. The mixture is thus
heated to a fusion temperature, creating a sinter suitable for
use in the blast furnace.
In the blast furnace, iron ore is reduced to molten iron
(also called pig iron or hot metal). The blast furnace is a
closed system with no atmospheric emissions. The effluent from
the furnace, blast furnace gas (BFG) that is rich in carbon
monoxide (CO), is cleaned of particulates and used as a fuel in
the blast furnace stoves. Each blast furnace has three or four
associated stoves that preheat the air blast supplied to the
blast furnace. Because these stoves are heated primarily with
BFG, NOX emissions from the stoves have low concentrations.
2-1
-------
CD
O
1
M—
O)
CD
"55
"8
•o
o>
Q.
<75
-------
Molten iron from the blast furnace, along with scrap steel,
is charged to the basic oxygen furnace (EOF) where high-purity
oxygen is blown on the molten bath (or upward through the bath in
the case of a Q-BOP [basic oxygen process] type furnace). The
oxidation of undesired elements in the bath (carbon, silicon,
manganese, etc.) converts the hot metal into steel and brings the
bath to a suitable pouring temperature. Combustible off-gases
from the process are typically collected in an open hood above
the furnace that admits air and are burned. Some furnaces are
equipped with suppression-type hoods that prevent air from mixing
with the off-gases above the furnace, and the off-gas is
subsequently flared. During the combustion of the off-gas,
thermal NOX is generated.
Scrap steel is melted in electric arc furnaces (EAF's) where
radiant heat from an electric arc established between the
electrodes (usually powered from a three-phase utility-based
supply) and the scrap or molten steel bath is used to bring the
charge to pouring temperature. Heating due to the electric
current passing through the scrap or molten bath is a minor part
of the total heat input. The use of electricity for steel
melting transfers the generation of NOX from the iron and steel
mill to a utility generating plant. However, oxygen and NG are -
sometimes used to preheat the charge. Thus EAF's are NOX
emission sources.
Molten steel from steel-making furnaces is typically
processed through a continuous caster where it is shaped into
slabs, billets, or blooms. Alternatively, it may be cast into
ingots and stored for subsequent processing. Slabs, billets, and
blooms from the continuous caster are typically reheated to
suitable working temperatures in reheat furnaces prior to being
passed through mills for further shaping. Ingots are typically
reheated in soaking pits prior to subsequent processing. Reheat
furnaces and soaking pits are high-temperature, fossil fuel
(typically natural gas) burning furnaces and are sources of NOX
emissions.
Finishing processes such as annealing and galvanizing also
involve reheating steel products to suitable temperatures for
processing. Consequently, these finishing processes are also
sources of NOX emissions.
2.2 UNCONTROLLED NOX EMISSIONS
There are three fundamentally different mechanisms of NOX
formation. These mechanisms yield (1) thermal NOX, (2) fuel NOX,
and (3) prompt NOX. The thermal NOX mechanism arises from the
thermal dissociation and subsequent reaction of nitrogen (N2) and
oxygen (02) molecules in combustion air. The fuel NOX mechanism
arises from the evolution and reaction of fuel-bound nitrogen
2-3
-------
compounds with oxygen. The prompt NOX mechanism involves the
intermediate formation of hydrogen cyanide (HCN), followed by the
oxidation of HCN to NO. Natural gas and most distillate oils
have no chemically bound fuel nitrogen and essentially all NO,
formed from the combustion of these fuels is thermal NOX.
Residual oils and coals all have fuel-bound nitrogen and, when
these are combusted, NOX is formed by all three mechanisms.
Iron and steel mill processes tend to use gaseous fuels,
i.e., NG, COG, BFG, and oxygen, and the NOX generation tends to
be thermal NOX. Exceptions include sintering where coke fines
are burned as a fuel and coke ovens where coal is destructively
distilled in' the absence of air. Emissions from sintering and
fugitive emissions from coke ovens may be sources of fuel NOX
emissions. Prompt NOX formation is not a major factor. It forms
only in fuel-rich flames, which are inherently low NOX emitters.
Thermal NOX formation is the predominant mechanism of NOX
generation at iron and steel mills.
Very little NOX emissions data are available in the
literature relevant to iron and steel processes. Table 2-1
summaries uncontrolled NOX emissions from the major process
facilities found in iron and steel mills. This summary is based -
on the available emissions data obtained during the preparation
of this document. The data are presented in Appendix A of this
document.
The uncontrolled NOX emissions tabulated in Table 2-1
indicate that coke-oven underfiring, sintering, reheat furnaces,
annealing furnaces, and galvanizing furnaces are facilities with
emission factors that range from 120 to 940 ppm at 3 percent 02.
Data available for both preheated combustion air and cold
combustion air furnaces show that NOX emission factors are much
higher when the combustion air is preheated. For some
facilities, e.g., reheat furnaces, there is much scatter in the
data, and the averages presented in Table 2-1 may not be
representative of individual furnaces.
2.3 NOX EMISSIONS REDUCTION TECHNIQUES •
Control techniques for NOX emissions can be placed into one
of two basic categories: techniques designed to minimize NOX
generation and techniques to remove previously generated NOX from
the waste'effluent stream. Combustion modification techniques
such as low-NOx burners (LNB's) and flue gas recirculation fit
into the first category. Add-on flue gas treatment techniques
such as selective catalytic reduction (SCR) and selective
noncatalytic reduction (SNCR) are examples of the second.
Few facilities found at iron and steel mills have NOX
controls. For many facilities, a suitable control technique has
2-4
-------
CO
S3
O
CO
CO
H ^
W ^
••"•
Q §
S^
|g
o
CJ
D ^
o °<*°
S CO ©
g CO
|8f
w
!*
FIRING 1
PROCESS FACILITY
vo
vr>
o
o
o
00
1
. 1
)ke-oven underfiring
\^
o
r-
o
o
^j*
ro
'
Sintering
O5 ' '
o ; i i
• ii
o o o o
m VD co H
ex
1 1 1
1 ,
last furnace stoves
asic Oxygen Furnace
during 02 blow
during non blow
EAF
PQ m
«t^
H
O
i
I
i
4J
_J
to
>
-H -H
4J 4-1
Regenera
Recupera
Reheat Furnace
*3*
H
O
O
ex
H
•H
i
T)
rH
O
U
in o in
H <* H
H O O
o o o
oo m ex
c^ m H
Q) 0)
> >
-H -H ^1
4J 4-) -H
Regenera
Recupera
Cold-A
Annealing Furnace
ex
H
O
<*<
CTl
<1)
>
-H
1J
Regenera
Galvanizing Furnace
w
0
o
o
m
CO
0)
-H
4-)
Recupera
^*
H
O
O
ex
H
-H
i
•d
rH
o
o
-2-5
-------
not been demonstrated. These facilities include sinter plants,
coke ovens, blast furnace stoves, and steelmaking furnaces. The
Japanese have investigated using SCR for sinter plants and coke
ovens. These efforts appear to be experimental, and.SCR is not
used routinely on these facilities in Japan. Iron and steel
process facilities known to have NOX controls applied are reheat
furnaces, annealing furnaces, and galvanizing furnaces. Control
techniques known to have been applied to these facilities are as
follows:
Reheat furnaces
Annealing furnaces
Galvanizing furnaces
Low excess air (LEA)
LNB'S
LNB plus flue gas
recirculation (FGR)
LNB
LNB plus FGR
SCR
LNB plus SCR
LNB
LNB plus FGR.
Other control techniques are discussed in this document.
Table 2-2 summarizes controlled NOX emissions data and
estimates and percent reductions. Percent reductions range from
13 to 90%. For reheat furnces controlled emissions range from 25
ppm @ 3% 02 for cold air furnace controlled by LNB+ FGR to 560
ppm for a regenerative furnace controlled by LEA. Emissions
reductions for the reheat furnaces with preheated combustion air
range from 0.03 Ib NOx/MMBtu for an LEA-controlled recuperative-
fired furnace to 0.61 Ib NOx/MMBtu for regenerative-fired
furnace. Due to low uncontrolled NOX emissions, the emissions
reductions for cold air reheat furnace are much lower.
For annealing furnaces, controlled emissions range from 10
ppm for a cold-air fired furnace controlled by LNB+SCR to 390 ppm
for a regenerative-fired furnace controlled by LNB. Annealing
furnace emissions reductions range from 0.07 Ib NOx/MMBtu for a
cold-fired furnace with LNB to 0.85 Ib NOx/MMBtu for a
regenerative-fired furnace with LNB plus SCR controls.
For galvanizing furnaces, controlled emissions range from 50
ppm for a'cold- air fired furnace controlled by LNB+FGR to 470
ppm for a regenerative-fired furnace controlled by LNB.
Emissions reductions for galvanizing furnaces range from 0.07
Ib/MMBtu for a cold-air fired furnace with LNB to 0.69 Ib/MMBtu
to a regenerative-fired furnace controlled by LNB+FGR.
2-6
-------
CO
6
H
H
1
r«i
8
u
rt
H
i
CO
6;
O
H
CO
CO
1
3
0
o
u
i
CM
1
CM
W
€
i-5
Hy
fc O
^4 | j
IB
s §
Pi
Cft ^""*
l^j 3 Ot
^^ **"r '"J
<*> fj p
S m °
w
l|i
CO ^ (14
tj pQ (3
A h^ O
M~ CQ
Pi
CQ->
co e S
03 -^.s
U m w
BBS
j
o
p^
o
o
w
^
"
w
r)
1
rou>c^ oooinoo o
t~H VC r*^ LO ^D ^0 GO 00 Q\ if}
™S?n r-c-r-oim^ r-
rngg ooooog o
0 CD 0 00000^ 0
_ CO VD
^^ i ^^ .. ^J \O vO VO CO ^^ ^3
HOO CNHHOOO CN
O Q . Q OOOOOO O
o*\ r^ co oo co co ^ c^ c*^
VO-l ^ i-J
O
O S
Mr i i
^ H N
W ^ S
ffi ^ ^
W § ' >
3
-------
2.4 COSTS AND COST EFFECTIVENESS OF NOX CONTROL TECHNIQUES
Tables 2-3 thru 2-5 present costs and cost- effectiveness
estimates for NOX controls for reheat, annealing and galvanizing
furnaces, respectively. These estimates are based on a limited
data base. Controlled NOX missions (and, consequently, emissions
reductions) are often based on test data from a single furnace.
Costs are also based on limited data, often a single example
furnace, that do not account for site-specific factors. All
costs are in April 1994$.
Costs and cost effectiveness vary with furnace firing
capacity. In discussing cost effectiveness in Tables 2-3 thru 2-
5, values used refer to the mid-size capacity furnace unless
stated otherwise and are for existing furnaces. For1 combustion
modification controls such as LNB and LNB/FGR, costs and cost
effectiveness for new furnaces are approximately 1/3 - 1/4 the
corresponding number for new sources. For add-on controls
only(SNCR and SCR), costs and cost effectiveness are, the same for
both new and existing sources.
As shown in Table 2-3, the cost effectiveness of existing
mid-size reheat furnaces of all control techniques range from
$90/ton NOX removed (LNB on a regenerative-fired furnace) to
$2,400/ton NOX removed (LEA on a cold-air fired furnace). Only
LEA has cost effectiveness greater than $700/ton. Capital costs
and annual costs for the various controls range from $190,000 to
340,000 and $51,000 to 83,000 respectively.
For existing mid-size annealing furnaces (Table 2-4) cost
effectiveness of all control techniques range from $200/ton NOX
removed (for LNB at a regenerative fired annealing furnace) to
$7,900/ton NOX removed (for LNB+SCR at a cold-air fired furnace).
Capital costs range from $540,000 to 4,500,000 and annual costs
from $77,000 to 870,000.
For existing mid-size galvanizing furnaces (Table 2-5), cost
effectiveness of all control techniques range from $110/ton for
LNB on a regenerative-fired furnace to $1,200/ton for LNB+FGR on
a cold-air fired furnace. Capital costs for the two controls are
$250,000 and $ 380,000. Annual costs are $58,000 and $70,000.
2.5 IMPACTS OF NOX CONTROLS
All of the NOX control techniques listed in Section 2.3 have
the potential to impact other air emissions in addition to NOX,
and all may have energy impacts. SCR units may have a solid
waste impact in the disposal of spent SCR catalyst. None of the
listed techniques have a wastewater impact.
2-8
-------
W5
o tj ^
w
c
Cj j^**v
(D ^
o
*c3 *^*
3 •*-» ^^
*2 C^ j?^
C r\ CP
ts.
•f !§„
u UE
1
I
2
1
u
"* — r fvf
°, 0 0
;-. en cs •
O 0 O
4) O
£ EH i_
3 -a
&0 0 —
O O O
fti & U
a
0 0-0
O\ 10 O
o o o
vo vo VO
0 O O
en en en
fN «S CS
W rt
fl} 0} •**
CO, «?
Si "°
§
o o o
• »— i ^- VO
o • o o
en oo cvj
en en en
00 00 00
o o o
*T ^T Tt
en en en
•s -s
g g .B
1) i* 2
o o o
c^ rt o
o
1
2-9
-------
CO
e / — ^
|ll
(*-
c
.2
"8 "g
o
0*^
^
ctf co i~
P|
111
o t£
a
^c
8
•+-*
o
U
o o S
i S s
t-» t> r~
t-> r- r-
O 0 0
IO >O 1/1
0 »
2 2 .fc
| .§• ?
M o 2
o o o
(Vj py] rj
§
O O S
O 0 (-3
10 O S
-* «N ^
o o o
0 O 0
oo oo oo
.1 3
22-
o ai -a
C B. f
& s 1
0
S
o § S
OO fo 6^
10 ^^ en
S I-P
\O \o vo
CN es ts
o o o
^ ON ON
NO NO VO
5 ^
•a *±
TO Cv
c .0, *?
wi 5 2
U U O
(Vj fV| ^_)
u
C/3
o o o
o o o
i— CN r-
O : Q O
O ' O o
o : o o
.55 • .55
1 i -
c : a '?
p . a TJ
w o -S
O d3 O
C£i (Vj tj
0
00
2-10
-------
s
-H
J-)
0
O
CJ)
I
CNJ
PQ
<
H
C ' — •>
111
u S tS-
§
e
= 'E
^ c
2
0*"^
Cti W3 ^
g II
— o*
-S to g
•a o 0
8-u -
u t^
1
H
e
;c
,
o
U
888
— m r~
o o o
00 00 r-.
^^ (^5 ^^
r~ r- r-
oo oo oo
o o o
o o o
10 >o vo
£ g
2 2 u
S SL '?
ass
«;> S o
p/j ^5 O
U
w
4-
^
P i 1
— « >
s e
U O '3
CO.?
-------
JE
12
CO
CO ,
« 1?
a5 C'
fc
_o
^ c
" O
O
Ctf C/3 ^^
I5!
III
O £t
H
c
;c
2
•*->
o
u
2 .8-8
— . f* ON
^D O ^5
r(- CS •<*
CO —
0 O O
CM CM CN
O ID
£ .£
4-< -4-J
e e w
g a •?.
a s s
D S O
« Ot$ 0
§
o o o
2 5 r4-
0 0 0
VO VO VO
0 0 O
oo oo oo
co co co
m o
^ •§ v.
S & '?
(D S "O
g> 8' -o
os « o
o
^!
§
2-12
-------
2.5.1 Air Impacts of Combustion Modification Techniques
All of the listed combustion modification NOX control
techniques, LEA, LNB, and LNB plus FGR, may increase CO and
unburned hydrocarbon (HC) emissions. The NOX reduction
mechanisms inherent in these modification techniques are the
reduction of peak flame temperatures, which is exponentially
related to the formation of thermal NOX and the reduced
availability of excess oxygen needed to form NOX. Reducing the
availability of oxygen to the combustion process increases the
likelihood that some HC will not be burned and that some CO will
not be oxidized to C02. However, if the control is properly
designed and applied, NOX control can be achieved without
increasing CO or HC emissions.
Data contained in the process heaters and glass ACT
documents indicate that decreases in excess oxygen levels begin
to impact CO emissions below 3 percent excess oxygen. Data in
the utility boilers ACT document show that CO emissions increase
for some boilers and decreace for others when implementing
combustion modifications. In a steel reheat furnace with LEA for
NOX control, CO emissions of 39 ppm at 3 percent O2 and 26 ppm at
3 percent 02 are reported for the uncontrolled and controlled
tests, respectively. These results indicate that CO emissions do
not necessarily increase when LEA is implemented for NOX control.
Other available information reports CO emissions from an
LNB-controlled reheat furnace to be 20 to 30 ppm at 3 percent O2.
Further, laboratory and field tests on a galvanizing furnace
yielded NOX emissions of 550 to 1,200 ppm at 3 percent 02 and,
concurrently, negligible CO emissions. Modifications to the
burners reduced NOX emissions to 350 to 430 ppm at 3 percent 02
and, concurrently, increased CO emissions to 30 to 60 ppm at 3
percent 02. The purpose of the modifications was to reduce NOX.
No explicit data are available relevant to the impact of NO,
control techniques on HC emissions from iron and steel process
facilities. In general, controls that reduce NOX by reducing the
availability of excess oxygen in the high-temperature regions of
a furnace would not increase either CO or HC emissions unless
oxygen availability was reduced excessively. Limite d data in
the Industrial-Commercial-Institutional Boilers ACT document
indicate that HC emissions do not change due to the
implementation of combustion modifications.
Reheat, annealing, and galvanizing furnaces predominantly
use natural gas as a fuel. Natural gas does not contain sulfur
and, consequently, SO2 is not emitted. Gaseous fuels, including
natural gas, can produce soot and carbon black when burned if
insufficient oxygen is present. However, no evidence indicates
that NO. controls increase particulate emissions in these iron
2-13
-------
and steel facilities.
2.5.2 Air Impacts of SCR and SNCR
SCR units are add-on, flue gas treatment facilities that
reduce NOX by injecting ammonia (NH3) upstream of a catalyst
reactor. Within the catalyst, NOX reacts with the NH3 and is
reduced to N2 and water '(Hp) . There is a potential for unreacted
NH3 to escape with the flue gas from the SCR unit. Any such
emissions are referred to as ammonia slip.
Two examples of SCR controls on annealing furnaces at iron
and steel mills are reported. One of these units is operational
with more than 3 years' operating history, and one is still under
construction. In the case of the former, the NH3 slip was
guaranteed to be less than 10 ppm initially and less than 12 ppm
after 1 year. The observed NH3 was initially less than 10 ppm.
Subsequent observations are not reported. The typical NH3/NOX
molar ratio for this unit is 0.9.
Other reports indicate that NOX removal rates of 70 to 90 -
percent can be achieved with SCR using NH3/NOX molar ratios
between 0.9 and 1.0, and that the NH3 slip will be between 5 and
10 ppm. These levels are considered to be well below health and
odor thresholds.
SNCR has not been applied to iron and steel mill process
facilities. In other SNCR applications, ammonia slip is
controlled to acceptable levels by controlling the ammonia to NOX
molar ratio. These levels are similar to ammonia emissions from
SCR applications, e.g., 10 ppm.
Pilot-scale testing and chemical kinetic modeling of SNCR
processes have shown that nitrous oxide (N20) emissions are a by-
product of both ammonia (NH3). and urea injection. The N20
formation resulting from these processes has been shown to be
dependent on the reagent used, the amount of reagent injected,
and the injection temperature.
Full-scale tests on fossil-fuel-fired boilers have shown
that direct emissions of N20 are less than 15 ppm and do not
generally correlate with NOX emissions. N20 production is higher
for urea injection than it is for NH3 injection.
V
2.5.3 Solid Waste Impacts
The only NOX control with solid waste impacts is SCR due to
2-14
-------
the disposal of spent catalyst materials. Titanium dioxide and
vanadium/titanium have been identified as the catalysts in the
two SCR annealing furnace applications cited. Other commonly
used materials are vanadium pentoxide, tungsten trioxide,
platinum, zeolites, and ceramics. Of these, vanadium pentoxide
is a toxic compound and a cause for concern. However, worker?
safety precautions adequately prevent any increased risk to
workers handling the catalyst, and stack emissions of vanadium
pentoxide are 1 million times less than industrial worker
exposure.
Most catalyst manufacturers arrange to recycle and
reactivate the catalyst. When that is not practical, the spent
catalyst can"be disposed of in an approved landfill in accordance
with the Land Disposal Restrictions in 40 CFR Part 268,
Subpart D.
2.5.4 Energy Impacts
All of the combustion modification control techniques have
the potential to impact energy requirements by affecting the
thermal efficiency of the process. No data are available to
quantify the impact of these controls on iron and steel mill
process facilities.
SCR results in a pressure drop across the catalyst that
requires additional electrical energy for the flue gas fan. One
estimates a cost of $537/yr assuming 8,000 hours operation per
year and electricity at 8 cents/kWh.
2-15
-------
-------
CHAPTER 3
INDUSTRY DESCRIPTION
Iron and steel are essential commodities in a modern,
industrialized society. Steel is a widely used industrial
material affecting every facet of society. Steel is a major
component of every transportation system, present in every
motorized vehicle and in the network of highways, bridges, and
traffic controls. It is a widely used construction material,
present in every home, skyscraper, and dam. Steel is essential
for most industrial facilities, prevalent in defense hardware,
and found in most durable goods. The uses of steel are too
prevalent and too diverse to be listed completely or to serve as
a basis for classification.
Section 3.1 of this chapter presents a brief background/
historical discussion of the evolution of iron and steel
processing. Section 3.2 is a brief industry characterization
including an overview description of the major production
processes and an overview of the industry. Section 3.3 presents
a description of those processes with a potential for large NOX
emissions.
3 .1 BACKGROUND
Iron (Fe) is one of the more abundant and widely distributed
elements in the earth's crust, constituting not less than 4
percent of the total crust. It is the fourth most abundant
element in the earth's crust, outranked only by aluminum,
silicon, and oxygen. Pure iron is a silvery white, relatively
soft metal that readily combines with other elements, e.g.,
oxygen and sulfur. Consequently, native metallic iron is rarely
found in nature.1'2 Iron oxides are the most prevalent natural
form of iron. That portion of the iron oxides that is of
commercial significance, i.e., economically and spatially
available for industry use, is referred to as iron ore, and iron
ore deposits are widely distributed. These deposits vary widely
in mineralogy, chemical composition, and physical
characteristics. The United States has abundant reserves of iron
ore. These reserves are grouped into five areas: Lake Superior,
Northeastern, Missouri, Southeastern, and Western. Of these, the
Lake Superior reserves are the most important.3
Steel is the generic name for a group of ferrous metals
composed principally of iron which, because of their abundance,
durability, versatility, and low cost, are among the most useful
metallic materials known. Most steels contain more than 98
percent iron. Steel also contains carbon, up to about 2 percent,
and may contain other elements. By controlling the carbon
3-1
-------
content and alloying elements, and by proper selection of heat
treatment in the finishing processes, steel can be produced with
a wide range of mechanical and physical properties.
3.1.1 A Historical Perspective
Facilities for reducing iron ore (iron oxide) to metallic
iron have evolved to the modern blast furnace that produces iron
from ore and other iron-bearing materials (e.g., sinter, pellets,
steelmaking slag, and scrap), coke, and flux (limestone and
dolomite). In the blast furnace, a blast of heated air and, in
most instances, a gaseous or liquid fuel, are injected near the
bottom of the furnace. The heated air burns the fuel and the
coke to produce the heat required by the process and the reducing
gas that removes oxygen from the ore. The reduced iron melts and
pools in the bottom of the furnace where it is periodically
drained through tapping holes as molten pig iron or hot metal.4
Modern steel making dates from the introduction of the
pneumatic or Bessemer process in 1856. The Bessemer process
involved forcing air through molten pig iron to oxidize the major
impurities (silicon, manganese, and carbon). The Bessemer
process was quickly followed by development of the open-hearth
furnace that evolved into the principal means of producing steel
throughout the world.5 Open-hearth furnaces were used in the
United States to produce steel into this decade, but all have now
been replaced by modern basic oxygen furnaces(BOF).
In the modern basic oxygen furnace, high-purity oxygen is
blown onto the surface of the hot metal (top-blown or basic
oxygen process) or up through the hot metal (bottom-blown or Q-
BOP process), oxidizing or removing excess carbon, silicon,
manganese, and other impurities from the hot metal to produce
steel of a desired composition.6
3.1.2 An Overview of the U.S. Iron and Steel Industry
Steel production is an international industry serving a
highly competitive, international market. The United States is
one of about 85 steel-producing nations and accounts ;for about 12
percent of the 1991 world production. In 1990, the United States
produced about 99 million tons of raw steel. Of this total, 37
percent was produced in EAF's and the balance predominantly by
BOF's. In 1991, the United States produced about 87 million tons
of raw steel. In 1991, 76.percent of the steel produced was-
processed through continuous casting machines rather than through
ingot casting. Continuous casting is projected to account for
about 84 percent in 1995. It should be noted that 1991 was not a
good year for steel production. World steel production was down
about 5 percent from 1990, and U.S. production was down 12
percent. Shipments to both the automobile and construction
3-2
-------
industries were down significantly in 1991. The automobile
industry is the largest single consumer of steel in the United
States, and the construction industry is.also a large consumer.7
Emphasizing the international character of the steel
industry, the import share of the United States steel market,
during the first half of 1991, was 8.24 million tons or 18.9
percent. Exports for the same period were 3.2 million tons.
Annual steel production capacity in the United States_in
1991 was about 120 million tons. High and low capacities in
recent times were 160 million tons in 1977 and 112 million tons
in 1988. In 1990, the steel industry in the United States was
composed of approximately 300 companies, of which 83 produced raw
steel at 127 locations.8
Table 3-1 summarizes data pertinent to the iron and steel
industry to characterize the current status of the industry.
Basic oxygen converters (BOP's and Q-BOP's) remain the major
source of steel production, but electric arc furnaces now account
for about 37 percent. Open hearth furnaces are no longer in use
in the United States.
A 1991 directory of iron and steel plants in North America
defines three categories of steel producers as follows:9
• Integrated Steel Producers are defined as those
companies having blast furnace or direct reduction
facilities and whose principal commercial activity is
the production and sale of carbon steel.
• Specialty Producers are defined as those companies _
whose principal commercial activity is the production
and sale of stainless steels, alloy steels, tool
steels, bars, wires, pipe, etc.
• Mini-Plants or mini-mills are defined as those
companies whose production is based on electric
furnace-continuous caster-rod/bar mill operations,
generally rolling carbon steel products--rebar, rounds,
flats and small shapes.
The differences between these categories are not always clear.
Some companies in all three categories use EAF's to melt steel
and have continuous casters. The integrated producers are the
3-3
-------
TABLE 3-1. SUMMARY OF CURRENT STEEL INDUSTRY PACTS7'1
U.S. percentage of
world production (%)
Total U.S. production
(106 tons)
Total U.S. shipments
(106 tons)
Total production by (%) :
Basic oxygen converter
EAFs
Open hearths
U.S. raw steel capacity
(106 tons)
Number of companies
Raw steel producers
Imports (105 tons)
Percent by continuous
casting (%)
1991
11.6
87
79
37
120
83
75.8
1992
92.9
82.2
113
1993
96.1
88.5
110
3-4
-------
most distinct group because they alone produce iron from iron ore
in blast furnaces or direct reduction iron (DRI) facilities.
Integrated producers also use coke in blast furnaces and either
produce coke on site or purchase it from coke producers. As of
May ,1991, there was only one DRI facility in the United States.10 ..
Mini-mills use only electric arc furnaces and use scrap steel as
feed stock.
3.1.2.1 Integrated Producers. The integrated producers,
although much fewer in number than specialty producers and mini-
mi 11s, produce the greater volume of steel. In 1990, integrated
producers accounted for about 70 percent of raw steel production
in the United States.7'11 The referenced 1991 directory lists 20
integrated steel producers in the United States. However, it is
not clear that all of those listed conform to the preceding
definition of an integrated producer, i.e., use blast furnaces or
DRI facilities and produce carbon steel. Those known to be
integrated producers are listed in Table 3-2 along with other
available information.9
3.1.2.2 Mini-Mills. The mini-mills are a growing segment of
the steel-producing companies that, in 1991, accounted for more -
than 20 percent of production in the United States. It has been
suggested that the term mini-mill is a misnomer. The steel-
melting facility used by mini-mills (i.e., EAF's) can produce up
to 130 tons of steel per hour, not far from the 200- to-300-
tons/hr typical of a EOF.12 Once viewed as suppliers of
unsophisticated, low-quality products requiring minimal
technology, mini-mills are now recognized as playing a growing
technological role and for having made permanent inroads into the
traditional domain of integrated producers.13 In 1990, Nucor
Corporation and North Star Steel Company, two mini-mill
producers, ranked seventh and eighth, respectively, among United
States steel producers. Nucor, for example, has an annual
production capacity of 3 million tons, and North Star an annual
capacity of about 2.5 million tons. By contrast, USX Corporation
(formerly U.S. Steel Corporation), the nation's leading steel
producer, shipped about 12 million tons in 1990.9>12 The 1991
directory of iron and steel plants lists 42 mini-mill companies.
3.1.2.3 Specialty Producers. It is more difficult to
characterize the specialty producers than integrated or mini-mill
producers. There are more companies, a total of about 120; they
produce a wide variety of specialty products; and it is difficult
to ascertain the starting point in their production process from
the literature. 'Company sizes vary widely (e.g., the number of
employees number from as few as 8 to as many as 5,000) and annual
capacities range from 62,000 to 1,500,000 tons of product. As a
group, they operate EAF's and a host of other furnaces including
annealing and reheat furnaces. By contrast, some weld or do
3-5 ,
-------
ffv
to
PJ
H
0
P
Q
o
Pi
Pi
j
W
to
_
p
uQ
M
3
5
W
c
H
O
IH
E?
5jj
1
§
H
NPORMAT
H
1
to
•
CS*
1
f)
W
t"3
1
•|
4J .P
a A
So1
ft i-i
1
ft vo
tt rH
00
CJ eo
i-l
BETHLEHEM STEEL
Bethlehem, PA
(215) 694-2424
-O
0)
•» N
0)-H
«a
* rH ?
S C 01
-H -H
4J a
Ol -H
-rl • i-l
ft CD 4J
rH 01
blooms ,
chemica
sheet ,
_
b
o
o
r-
tN
o
o
0
o
o
k. '
H
^
4J
4)
(1) •
J3 o
to *o
Q) Q
o •
0)
c a
O-H
w a
•H
• 0) m
Q) JJ rH
.y *-H
O rH O
cj au
m
GENEVA STEEL
Provo, UT 8460
(801) 227-9090
_
4
eri
H
c»
IN
•d
o
o JH o
o ao
o o
• i-i •
o tN
• u •
rH 05 H
••0
0) 0)
4-* N
m-H
• rH
m (6
rH Ol
u a
-rl -H
0) -U
x; to
u
rH <1>
(I) 0)
O J3
cj tn
• in
cj n
2i CM
H H
I
• i-l
i-3 0
H a\
GULF STATES STE
Gads den, AL 35
(205) 543-6100
^
M
•*
* s
u
(D fl5
S-°
rH •
am
i-H
Sheets,
chemica
billet
CJ
25
H
k.
co
Pa
H
Oi
CO ro
D o
P vo
INLAND STEEL IN
Chicago, IL 60
(219) 399-1200
M
SC
5
•
Id m
rH a
"a
4J •
0) Ol
rH C
rH -rl
0 <0
a o
H U
in
• H
£3
M ^
LTV STEEL CO.,
Cleveland, OH
(216) 622-5000
- ~-
..
S
a
•H
4J
m
4J"
0)
0) -a
J3 0>
m H
•H
i«J
4J O !>
«} rH rH
i-l rH «5-
ft «J 01
m
co
H
McLOUTH STEEL
Trenton, MI 48
(313) 285-1200
,_
tu
p
JIj
4J
<—•
3-6
-------
^^
•0
Q)
§
•H
§
O
^•*>
(N
TABLE 3
"§
•H
ij ii
SB)
0
H 0
O) rH
co
U 4J
Coal , chemi
sheet , ingo
galvanized,
Ingot, slab
• in
o< «* H
e! in CN
O U3 i-H
U •«* CO
• ^f
J o O
MHO HO
(ft (*) W *• rH
§ (C CS CO H H
K & O •#
o rt -— *• w _n f^
H J3 en O tn i
EH CQ H E3 rt O
2 5EJ ^^ Pi Q CO
$
0
r~
o
o
o
o
o
fc
^
0)
t 6^
rH
V H
8 1
Strapping,
tubing
gavlvanized
• rH
Da CJ
te H
^5 ^D
u H in
t-
W < rt
U CU i
EH H
rH CTl
!5rH
BJ 1-1 CS
CO fc —
«.
Ol H
S'S
»
o
Cn
C -
"H iJ O
C"H rH
O ft rt
Coke, pig i
pipe, rods,
galvanized,
\o
r*
T
ON
rH
CS
in
rH
grH
«rt
• Dim
i!5
U tn —
4J (N
X -U H
D ft -"
§
o
0
0
o
o
in
H
•O
i-H r-H
rH S
CO t-
a c"?
O 0
EH 4-J -^^>
Pi M *J*
H -H O'
» S ~-
§
•0
V
4J N
O-H
Cn C co
C rt 4J
-H > V
rH 3
* flj 'o
Coke , sheet
tinpiate, g
roofing pro
8
w.
B
CO
K ro
O o
0J O
»D VD
CQ f^
co o
EH o
H ^ M
1 .T}l
2J C CN
H -H
r-3 rH •—*.
PQ O ^
ta o) o
J5 5 -"
m
Q)
OJ
CO
>1
^)
m
C
o
•rH
OJ
rt
Cu
(1)
4J
2
tn
0)
4J
C
H
3-7
-------
other single-function operations that do not include the onsite
combustion of fuels. Products may range from a single item such
as cold-drawn wire to highly specialized steels.9 In terms of
their contribution to domestic capacity and production, they are
grouped with mini-mills.
3.2 INDUSTRY CHARACTERIZATION
3.2.1 An Overview of Iron and Steel Manufacturing
Figure 3-1 is a simplified flow diagram illustrating the
principal steps involved in the production of steel. All of
these processes, and more, are found in modern integrated steel
mills. An overview of these processes is presented here.
Individual processes that produce NOX are discussed in more
detail in Section 3.3 of this chapter.
3.2.1.1 Raw Materials and Preparatory Processes. With
reference to Figure 3-1, the basic raw material input for the
production of iron and steel is iron ore, which is reduced to
metallic iron in the blast furnace. The principal iron ore
beneficiating processes are pelletizing and sintering.
Pelletizing is usually accomplished in specially designed
furnaces located at or near iron ore mines rather than at iron
and steel mills. Its purpose is to pelletize fine, low-grade
ores prior to shipping to an iron and steel mill. Pelletizing is
rarely practiced in an iron and steel mill.
In contrast to pelletizing, sintering is often practiced at
integrated iron and steel mills. Its primary purpose is to
agglomerate dusts and fines from other process, e.g., ore fines,
coke fines, and flue dust, into particles with suitable mass,
size, porosity, and strength to charge into the blast furnace.
Fines charged to a sintering furnace differ in size from those
charged to a pelletizing furnace, and they originate at the iron
and steel mill rather than at the mine. Alternatively, fine
particles charged into the blast furnace will be blown out by the
rapid countercurrent flow of the furnace gases.
Another basic material used in the conversion of iron ore
into metallic iron is coke. Coke is the primary residue that
remains when a blend of pulverized coking coals is heated
gradually to high temperatures, about 900 to 1,100 °C (1,650 to
2,010 °F) .in the absence of air for approximately 18 hours.
About 90 percent of the coke produced in the United Stated is
charged to blast furnaces for the production of pig iron.14
The conversion of coal to coke is performed in long, narrow,
slot ovens, i.e., by-product coke ovens, which are usually
designed and operated to permit the separation and recovery of
the volatile materials evolved from the coal during the coking
3-8
-------
pl§
*- re
IS u
£g
0) h;
i
CD
C -
3-J2
go.
to
1
(A
o
»
-8
CO
is
0) !=
«5
en
o
t o
43 (0
CD
CO
2
re
p
t^
oc
o
Q)
U
II
«!«
c-^: o
(75 £
o
g
i
tsl
t
re
u
I
«P
O)
c.
(0
"33
v
o.
(75
CO
I
O)
J2 o
'
|
-
CO
15
8
o
§
TO
.5
•s
Q.
3-9
-------
process. In addition to coke, the primary product, a number of
by-products may also be recovered including breeze, crude tar,
crude light oil, ammonia, and coke oven gas. These may be used -,
on site or marketed. Coke is likely to be used on site but may
be marketed, coke oven gas is likely to be used on site, and
breeze is likely to be recycled to the sintering process.15
A third raw material input to the iron and steel making
process is flux. In the iron smelting process, i.e., when iron
is separated from the ore by fusion in a blast furnace, a flux is
used to enhance the process by combining with ash in the coke and
gangue in the ore to make a fluid slag that can be readily
separated from the molten iron. Selection of the proper flux for
a given process is a well-established process requiring knowledge
of the composition and properties of the materials involved.
Limestone and/or dolomite are the fluxes used in blast furnaces.
The proportions of each depend principally on the constituents of
the slag and the amount of sulfur that the slag must remove.16
In the basic oxygen steelmaking process, lime normally is
added as calcined or burnt lime or burnt dolomite. The calcium -
oxide in either the burnt lime or-burnt dolomite fluxes the
silica formed upon the oxidation of silicon in the hot metal.17
Often, limestone is also charged into the sinter plant to
produce a precalcined or self-fluxing sinter. The use of self-
fluxing sinter has both economic and performance advantages.18
3.2.1.2 Ironmaking. Iron ore, coke, selected fluxes (e.g.,
prepared limestone), and sinter are charged to the blast furnace
where iron is reduced from its ore. The product of the blast
furnace contains more than 90 percent iron and is referred to as
pig iron or, if in a molten form, hot metal. Most of the blast
furnace product is transported directly to onsite steelmaking
furnaces as hot metal.
The blast furnace charge of iron-bearing materials
(including iron ore, sinter, mill scale, steelmaking slag, and
scrap), coke, and flux (limestone and/or dolomite) is placed in
the furnace, and a blast of heated air and, in most instances, a
gaseous, liquid, or powdered fuel is introduced through openings
near the bottom of the furnace just above the furnace's hearth
crucible. The blast of, heated air burns the injected fuel and
most of the coke to produce the heat required by the> process and
the reducing gas that removes oxygen from the ore. The reduced
iron melts and pools in the bottom of the hearth. The flux
combines with the- impurities in the ore to produce a slag, which
also melts and accumulates on top of the liquid iron.
Periodically, the iron and the slag are drained from the furnace
through tapping holes.4
3-10
-------
The blast air delivered to the process through the furnace
tuyeres is preheated by passing it through regenerative blast-
furnace stoves that are heated primarily by combustion of the
blast furnace off-gas. Blast furnace gas is often enriched with
fuel of a higher calorific value such as natural gas to achieve
higher temperatures. Thus, some off-gas energy is returned to
the blast furnace in the form of sensible heat. This procedure
enhances the efficiency of the process by lowering fuel
requirements.
A modern blast furnace will typically have three or four
blast furnace stoves as auxiliaries. These are alternately fired
with blast furnace gas to raise the temperature of the stoves'
brick lining and then, by reversing the gas flow, preheat the
supply of blast air to the blast furnace to temperatures of 760
to 1,150°C (1,400 to 2,100 °F).19
To produce a metric ton of pig iron requires about 1.7
metric tons of ore or other iron-bearing material, 450 to 650 kg
of coke and other fuel, about 250 kg of limestone or dolomite,
and 1.6 to 2.0 metric tons of air.4
Although the blast furnace remains the dominant source of
iron for steelmaking, there are processes that produce iron by
the reduction of iron ore below the melting point of iron. These
are classified as direct reduction processes, and the products
are referred to as direct-reduced iron(DRI). DRI produces
several percent of the total iron produced worldwide. The major
part of DRI production is used as a substitute for scrap in
electric-arc steelmaking furnaces.20 There are many DRI
processes. The objective of these processes is to improve iron
ore until it is sufficiently iron-rich to be charged to an
electric arc furnace. A typical process uses heat to drive
oxygen from the ore, leaving an iron-rich residue called sponge
iron.6
3.2.1.3 Steelmaking. Steel is made from molten iron (hot
metal), delivered directly from a blast furnace, and scrap in
steelmaking furnaces. Two basic types of steelmaking furnaces
dominate this process: (1) ; the EOF and (2) the EAF. The open-
hearth steelmaking furnace, developed in the 19th century and a
major source of steel well into the latter half of the 20th
century, is no longer used in the United States.7 There are
three basic types of the EOF. All use oxygen of high purity
(>99.5 percent) to oxidize excess carbon, silicon, and other
impurities in the hot metal, thereby producing steel. One type
of EOF blows the oxygen on the top of the hot metal pool. In the
United States, this process is called the basic oxygen process or
BOP. The second type blows the oxygen through tuyeres in the
bottom of the furnace and is referred to as the bottom-blown or .
Q-BOP process. The BOP process is the most widely used form in
3-11
-------
the United States.
other two.
The third type of EOF is a combination of the
To make steel in a EOF, molten pig iron (hot metal) and
usually scrap steel are charged to the furnace. The scrap
consists of the by-products .of steel fabrication and wornout,
broken, or discarded articles containing iron and steel. The
furnace is mounted in a trunnion ring to facilitate tilting
during processing. At the beginning of a heat, i.e., the
processing of a batch of materials to make steel, the furnace is
tilted to receive a charge of materials through the open top.
Scrap steel, if any, is charged first. Scrap can form up to 30
percent of the total charge (up to 45 percent if preheated).1
The furnace is then tilted in the opposite direction to receive a
charge of hot metal (molten iron) from a transfer ladle. The
furnace is then returned to the vertical position and a water-
cooled retractable lance is inserted through the open top of the
furnace and positioned above the bath level. A water-cooled hood
is positioned over the open top. A jet of gaseous oxygen is
blown at high velocity onto the surface of the hot metal bath.
No external heat source is required during this process. -
Slag-forming fluxes, e.g., burnt lime, dolomitic lime, and
fluorspar, are added in controlled amounts through a chute built
into the side of the hood. The flux is added shortly before or
after the oxygen jet is started. '''.-..
The oxygen striking the surface of the molten bath
immediately forms iron oxide, part of which disperses rapidly
through the bath. The iron oxide reacts with carbon in the
molten bath to form carbon monoxide, which gives rise to a
violent circulation that accelerates the refining process.
Impurities such as carbon, manganese, silicon, sulfur, and
phosphorus are oxidized and transferred to the slag. These
oxidizing reactions take place very rapidly. A 300-metric-ton
heat (or charge), for example, can be processed in about 30
minutes.6
During the oxygen blowing, gases emitted by the process are
collected by the water-cooled hood and conducted to a cleaning
system where solids are removed from the effluent gas before it
is discharged to the atmosphere. The solids may be returned to
the sintering process or discarded, depending on whether or not
they contain contaminants. When the oxygen blowing is completed,
the lance is withdrawn and the molten steel is checked for
temperature and composition. If the steel is too hot, it is
cooled by the addition of scrap or limestone. If the steel is
too cold, the oxygen lance is reinserted and additional oxygen is
blown. When the temperature and composition are satisfactory,
the steel is tapped into a waiting ladle where alloying materials
are added. Subsequently, the remaining slag is dumped from the
3-12
-------
furnace, and the furnace is returned to a charging position to
begin the next heat.1'6'22
The bottom-blown or Q-BOP process is also used in the United"'
States. The Q-BOP furnace has tuyeres or double pipes in the
bottom of the furnace. Oxygen is blown into the furnace through
the center pipe and natural gas or some other hydrocarbon is
blown into the furnace through the annular space between the two
pipes. Considerable heat is generated when the oxygen oxidizes
the carbon, silicon, and iron in the molten bath. The
hydrocarbon stream provides essential cooling by thermal
decomposition. Without this cooling, the refractories
surrounding the oxygen jet would be destroyed.1'6
Although the BOP and Q-BOP processes have many similarities,
there are also significant differences. The Q-BOP process
operates much closer to equilibrium conditions between the metal
and the slag and is therefore much lower in oxidation potential
than the BOP. The manifestations of this characteristic of the
Q-BOP are lower iron losses as ferric oxide (FeO) to the slag,
higher manganese recoveries, faster blow times, improved
phosphorus and sulfur control, and lower dissolved oxygen and
nitrogen contents. Because of these improvements, yield of the
Q-BOP is 1.5 to 2.0 percent higher than the BOP, i.e., there is
less iron loss as FeO to the slag. The lower FeO generation,
however, results in lower scrap melting in the Q-BOP as compared
to the BOP. Process control of the Q-BOP is much easier than the
BOP because of the highly consistent metallurgical behavior of
the Q-BOP reactions and the absence of the variability caused by
oxygen lance practices as encountered in the BOP.21
The second major classification of steelmaking furnaces is
the EAF. In 1991, EAF production accounted for about 37 percent
of the total steel produced in the United States and Canada.7
Almost all of the balance was produced in basic oxygen furnaces.
Electric arc furnaces produce liquid steel primarily by
melting steel scrap. However, metallic iron, including pig iron
and direct-reduced iron, are sometimes added to the charge.
Generally, steel is produced in an EAF from a metallic charge of
about the same composition as the steel to be made.1 Reactions
taking place in the EAF are similar to those in the. basic oxygen
furnace. The charge is melted, impurities are oxidized, fluxes
are added to aid in the formation of slag, and alloying elements
are added to achieve the desired composition.
There are several different variations of the EAF. However,
the direct-arc, three-phase electric furnace is the most common.23
In operation, the three-phase electric current flows from one
electrode, through an arc between the electrode and charge,
through the charge, then through an arc between the charge and a
3-13
-------
second electrode. The charge is heated primarily by radiation
from the arc although heat developed by current through the
electrical resistance of the charge makes a slight contribution.
Three electrodes are used in the EAF, and the current flow to and"
from the electrodes is consistent with the ,features of the
utility-based, three-phase power supply.
Electric arc furnaces offer several advantages including low
construction cost, flexibility in the use of raw materials (steel
can be made directly from scrap without having a source of molten
iron), the ability to produce steels over a wide range of
compositions, improved process control, and the ability to
operate below full capacity. The improved process control makes
the EAF well suited for stainless steel and other high-alloy
steels. Economic constraints tend to favor the use of EAF's for
low to medium tonnage steelmaking facilities.1'24
Molten steel, whether from a EOF or an EAF, is tapped from
the furnace to a ladle and transported to a finishing process.
Specialty steels, i.e., "clean" steels which meet extraordinarily
stringent requirements of certain critical applications, may be .
subjected to additional processing in the ladle. These processes
are generally referred to as ladle metallurgy and as secondary
steelmaking. There are numerous ladle metallurgy processes
including ladle temperature control, composition control,
deoxidation, degassing, cleanliness control, and others.25
3.2.1.4 Finishing. Although there are many variations, most
steel follows one of two major routes to final processing. Both
routes produce solid, semifinished products: ingots or cast
slabs. The more traditional route is the ingot route. Molten
steel is poured from the ladle into an ingot mold where it cools
and begins to solidify from the outside toward the center. When
the ingot is solid enough, the mold is stripped away and the
ingots are transported to a soaking pit or reheat furnace where
they are soaked in heat until they reach a suitable, uniform
temperature throughout. The reheated ingots are then transported
to a roughing mill where they are shaped into semifinished steel
products, usually blooms, billets, or slabs. Blooms are large
and mostly square in cross section, and they are frequently used
in the manufacture of building beams and columns. Billets, which
are made from blooms, are also mostly square, but are smaller in
cross section and longer than blooms. Billets are processed
further to produce bars, pipes, and wire. Slabs are the wide,
semifinished product from which sheets, strip, and other flat
rolled products are made.26 The roughing mills may be the first
in a series of continuous mills that continue to shape the steel
into more useful, more finished products.
More recently, molten steel is increasingly routed to a
continuous casting process that bypasses the soaking pit or
3-14
-------
reheat furnace process associated with the ingot route. In this
process, the molten steel is lifted in the ladle to the top of
the continuous caster where it is allowed to flow into a .
reservoir called the tundish and from the tundish into the molds
of the continuous casting machine. These molds are cooled with
water so that a thin skin forms on the outside of the liquid
metal. Emerging from the mold, the continuous column of steel is
further cooled by a water spray, causing the skin to thicken.
The steel is further shaped by various designs of casters as it
continues to flow. As in the ingot process, the steel may be
shaped into blooms, billets, or slabs and, subsequently, into
more useful products.26 The continuous casting process has a
significant advantage in that the soaking-reheating step is
eliminated. Thus, it offers a fuel savings and the potential to
reduce undesirable emissions including NOX.
Downstream of the series of mills (roughing stands and
finishing stands) that shape the steel blooms, billets, slabs,
and other semifinished products are a multitude of processing
steps that may be used to produce finished steel products. The
actual processes followed may include a combination of many, few,-
or perhaps none of this multitude of options. The processes that
may be encountered include pickling, annealing, galvanizing,
coating, painting, and additional mechanical finishing steps.
Some of these steps require additional heating or perhaps
reheating and, thus, pose a potential for further combustion of
fuels with the attendant potential for generating additional NO,
emissions. These include annealing and all reheating furnaces.
3.3 PRODUCTION PROCESSES WITH NOX EMISSIONS
In this section, iron and steel processes with a potential
to produce NOX emissions are discussed in detail. Sintering,
coke ovens, soaking pits, and reheat furnaces are identified as
some of the more important iron and steel combustion processes
and, thus, some of the largest sources of NOX emissions.
3.3.1 Sintering
The sintering and pelletizing processes are the two most
important processes used to improve iron ore prior to making
iron. Of-these, pelletizing is usually accomplished near iron-
ore mines where ore is ground to fine particles, separated from
the gangue minerals either magnetically or by flotation, and then
made into pellets. This process is usually carried out at or
near the mines to reduce the cost of shipping the beneficiated
ore to iron and steel mills. At iron and steel mills, a parallel
process, sintering, takes fine iron-bearing materials recovered
from ore handling and other iron and steel operations and fuses
these fine particles into materials suitable for charging to
blast furnaces or direct reduction facilities to make iron.26
3-15
-------
In sintering, the mixture of iron ore fines and other iron-
bearing materials (iron-bearing dusts and slags) and fluxes (lime
or dolomite) are thoroughly mixed with about 5 percent of a
finely divided fuel such as coke breeze or anthracite and
deposited on a traveling grate. The traveling grate is shaped
like an endless loop of conveyor belt forming a shallow trough
with small holes in the bottom. The bed of materials on the
grate is ignited by passing under an ignition burner that is
fired with natural gas and air. Subsequently, as the grate moves
slowly toward the discharge end, air is pulled down through the
bed. As the coke fines in the bed burn, the generated heat
sinters, or fuses, the fine particles. The temperature of the
bed reaches about 1,300 to 1,480 °C (2,370 to
2,700 °F). At the discharge end of the sintering machine, the
combustion will have progressed through the thickness of the bed.
The sinter is then crushed to eliminate lumps, is then cooled,
and finally screened.1'18
Good sintering requires thorough mixing of the feed
materials and careful deposition on the traveling grate so as to
achieve a uniform bed. Proper ignition is also important, and _
some processes replace part of the solid fuel with gaseous fuel
to improve ignition and improve the sinter. Plants using this
process have approximately 25 percent of the length of the sinter
bed covered with a gas-fired ignition-type hood. Temperature in
the hood ranges from about 1,150 °C (2,100 °F) where ignition
begins to about 800 °C (1,500 °F) at the exit of the hood.
Depending upon characteristics of the ore materials and sintering
conditions, average production rates of 22 to 43 metric
tons/m2/day (2.3 to 4.4 tons/ft2/day) of grate area are expected,
and rates in excess of 49 metric tons/m2/day
(5 net tons/ft2/day) have been attained.27
The major source of energy used in the production of sinter
is the carbon content of coke breeze and flue dust. The amount
of ignition fuel required is about 140 J/g (0.12 MMBtu/ton) of
sinter produced. The total fuel requirement, including coke
breeze, is about 1.74 kJ/g (1.5 MMBtu/ton) of sinter produced.27
Beyond the sintering grate, the sinter is cooled so that it
can be transported by conveyor belts. The exhaust air from these
coolers is normally at too low a temperature to permit economical
recovery of heat. Recent developments in sinter cooling have
been directed toward on-strand cooling, which could improve heat
recuperation, sinter quality, and dust collection. The use of
sinter in the blast furnace charge, and especially fluxed sinter,
also improves blast furnace performance. One improvement is that
less coke is required in the blast furnace. Available data
indicate that for each net ton of limestone removed from the
blast-furnace burden and charged to the sinter plant to make
fluxed sinter, approximately 182 kg (400 Ib) of metallurgical
3-16
-------
coke is saved.18 This also has a potential for reducing NOX
emissions through fuel conservation.
3.3.2 Cokemaking
Coke is an essential component in the production of pig iron
(hot metal) in a blast furnace. Consequently, coke ovens are
found in most integrated steel plants. Coke ovens are
constructed in batteries that contain as few as 10 to more than
100 ovens. Coking chambers in a battery alternate with heating
chambers so that there is a heating chamber on each side of a
coking chamber. Underneath the coking and heating chambers are
regenerative chambers, a brick checkerwork that preheats
combustion air to conserve fuel, increase thermal efficiency, and
give a higher flame temperature. The ovens are constructed of
refractory brick.
In operation, selected blends of coal are charged to the
ovens through openings in the tops. Subsequently, the coal
undergoes destructive distillation during a heating period of
about 16 to 18 hours. At the end of this portion of the coking -
cycle, doors at both ends of the oven are opened and the
incandescent coke is pushed from the oven into a railroad-type
car called a quench car. The coke is then transported to the end
of the battery to a quench tower where it is deluged with water
to end the burning process. The operation of each oven in the
battery is cyclic, but the batteries usually contain a large
number of ovens so that the yield of by-products is essentially
continuous. The individual ovens are charged and discharged at
approximately equal time intervals. Practically all of the coke
produced in the United States is made by this process.13
The product of a coke battery is either furnace coke
(requiring 15 to 18 hours of distillation) or foundry coke
(requiring 25 to 30 hours of distillation).M The by-products of
the process are driven off during distillation and are recovered
in the by-product coke-making process through condensation. These
by-products include tar, light oils, and heavy hydrocarbons. The
noncondensable gaseous product remaining is known as coke oven
gas, which, on a dry basis, has a heating value of about 22
MJ/Nm3 (570 Btu/ft3) . Approximately 35 percent of the coke oven
gas produced is used in heating the ovens.24'27
Excess coke-produced in a cokemaking operation that is sold
to others is referred to as merchant coke, and merchant coke can
be either furnace coke or foundry coke.
In 1983, about one-sixth of the total bituminous coal
produced in the United States was charged to coke ovens. On the
average, 1.4 kg of coal is required for each kilogram of coke
produced. Each kilogram of coal carbonized requires 480 to 550
3-17
-------
kJ (450 to 520 Btu) of energy.
1,480 °C (2,700 °F) ."
Flue temperatures are as high as
Current annual coke production in the United States is about'
29 million tons. Of this amount, about 72 percent or 21 million
tons comes from steel plant coke ovens (furnace coke) with the
balance from independent coke producers (merchant coke). All of
this coke is produced in 12 States, and approximately 70 percent
is produced in five States: Pennsylvania, West Virginia, Ohio,
Indiana, and Illinois. The only nonrecovery coke-oven facility
producing metallurgical coke in the United States is located in
Virginia.28
The underfire combustion of fuel in coke ovens produces NOX,
among other emissions. Because of the many different emissions
from coke ovens, coke ovens will be subj ected to more stringent
environmental regulations in the 1990's and beyond. Apart from
these environmental concerns, most coke ovens in the United
States have reached obsolescence. The International Iron and
Steel Institute projects that approximately 40 percent of the
coke ovens in North America must be replaced within the next 10
years unless technology to extend their life is developed. The
costs of replacing these facilities, already considerable, are
significantly increased by the costs of meeting increasingly
stringent environmental regulations. These issues are of much
concern to the steel industry because coke is an essential
material for the integrated steel producers.8'28
3.3.3 Blast Furnaces and Blast Furnace Stoves
Iron ore is converted to molten iron in a blast furnace. A
charge consisting of iron ore, sinter, limestone, and coke is
charged to the top of the vertical-shaft blast furnace. The coke
provides thermal energy for the process and a reductant gas for
the iron-ore conversion, and it serves as the pathway passage.for
gases that pass through the furnace burden. The limestone
becomes calcined, melts, reacts with, and partially removes
sulfur from the molten iron.
Heated air is injected through tuyeres near the bottom of
the furnace and moves upward through the burden, consuming the
coke carbon and thereby providing energy for the process. Blast
furnace gas leaves the furnace through offtakes at the top of the
furnace, is cleaned of particulates, and subsequently used as a
fuel. The BFG contains about 1 percent hydrogen and about 27
percent carbon monoxide, and it has a heating value that ranges
from about 2,540 kJ/Nm3 (65 Btu/ft3) to 3,600 kJ/Nm3 (92
Btu/ft3) ,27'32 Between 2.2 and 3.5 kg of BFG is generated for each
kilogram of pig iron produced in the blast furnace.27
Molten iron and slag are tapped periodically from the bottom
3-18
-------
of the furnace." In operation, the blast furnace is kept under
pressure by an arrangement of pressure locks that function to
seal the top of the furnace. These locks are in the form of two ..
inverted cones, called bells, which operate sequentially to
maintain a seal in the furnace. Thus, the blast furnace is a
closed unit with no atmospheric emissions.27
A blast furnace typically has about three blast furnace
stoves associated with it. These stoves primarily burn BFG in a
process designed to preheat the air used to combust the fuel in
the blast furnace they serve. In an installation with three
associated blast furnace stoves, for example, two are generally
being preheated while the third is supplying preheated air to the
furnace. While being heated, hot gases from the blast furnace .
are cleaned, cooled, and some portion is then routed to the
stoves that are being heated where it is burned, often with other
gases, to heat a thermal storage, regenerative checkerwork
(refractory material) in the stove. The heat thus stored is then
used to preheat combustion air to the blast furnace. The BFG may
also be burned elsewhere in a steel plant. The blast-furnace
stoves require very large quantities of fuel for heating.
However, because the stoves are heated primarily with BFG, the
flame temperature is reduced, which reduces NOX generation.27
The BFG burned in the blast furnace is often enriched by the
addition of fuels with a. higher calorific value such as coke-oven
gas or natural gas 'to obtain higher hot-blast temperatures, often
in excess of 1,100 °C (2,010 °F) . This enhances the efficiency
and productivity of the process.19
3.3.4 Basic Oxygen Furnaces :
Most of the steel produced in the United States is made
using an oxygen steelmaking process. There are three oxygen
steelmaking processes: a top blown process called the BOP
process, a bottom blown process called the Q-BOP process, and a
third process, which is a combination of the other two. In all
three processes, scrap and molten iron (hot metal) are charged to
the furnace and high-purity (>99.5 percent) oxygen is used to
oxidize excess carbon and silicon in the hot metal to produce
steel. The major differences between, the three processes are in
the design of the furnace and the equipment for introducing
oxygen and fluxes.21 These differences do not have any
significant impact on the potential for NOX generation and are
not discussed at length here. The BOP process is the most
commonly used process in the United States.
In the BOP process, oxygen is blown downward through a
water-cooled lance onto a bath of scrap and hot metal. Heat
produced by the oxidation of carbon, silicon, manganese, and
phosphorus is sufficient to bring the metal to pouring
3-19
-------
temperature, and auxiliary fuel is not required.1
Oxygen striking the surface of the molten bath immediately -.
forms iron oxide. Carbon monoxide generated by the reaction of
iron oxide with carbon is evolved, giving rise to a violent
circulation that accelerates refining. The oxidizing reactions
take place so rapidly that a 300-metric-ton heat, for example,
can be processed in about 30 minutes. The intimate mixing of
oxygen with the hot metal permits this rapid refining.1 Figure
3-2 is a schematic representation of the refining in a top-blown
BOP. It illustrates the decreasing content of several
undesirable elements in the molten metal bath during the period
of an oxygen blow.21
Scrap use in a BOP furnace is limited to about 30 percent of
the charge. If the scrap is preheated, the scrap content can be
increased to about 45 percent.1 In addition to scrap preheating,
other fuels may be burned to dry out refractories and to keep the
BOF furnace from cooling between heats. The use of fuel for the
latter two purposes amounts to about 230 kJ/kg (200,000 Btu/ton)
of steel produced. _
About 470 kJ of carbon monoxide are produced in the BOF per
kilogram of steel produced (400,000 Btu/ton). Typical practice
is to burn combustible gases in water-cooled hoods mounted above
the BOF vessel. In most cases, the BOF vessels are equipped with
open hoods that admit air for combustion of carbon monoxide on a
relatively uncontrolled basis. Some new plants use suppressed
combustion hoods, which do not inspire air and burn off-gases.
New BOF capacity is expected to continue this trend, which may
cause a decrease in total NOX emissions. During the combustion
of the waste gas, the potential for NOX production exists.
3.3.5 Soaking-Pit and Reheat Furnaces
Soaking-pit and reheat furnaces are large furnaces used to
raise the temperature of steel in the course of processing to a
temperature suitable for hot working or shaping. They are
designed to accommodate the steel being processed at a suitable
rate, heat it uniformly, and hold it at a desired temperature for
a specified length of time. There are numerous design.
variations. Two major variations are batch-type and continuous -
type furnaces. Within the latter category are variations on how
the charge is moved through the furnace. These variations
include roller hearth furnaces in which the material moves as the
series of rollers that constitute the hearth rotate, walking-beam
furnaces in which material is moved in a controlled step-wise
manner, pusher-type furnaces in which a continuous line of
material is pushed over skids, and rotary-hearth furnaces with
circular hearths that rotate in a horizontal plane. Other
variations involve the introduction and removal of the charge and
3-20
-------
8 40
2
O
d 30
ut
20
o
o
ffi
e
10
0
!\
SV1
ooe
006 a
oc
o
z
A.
v>
004 O
&
ec
002 z
a.
0 2 4 6 ' 8 10 12 14 16 18 20 22
TIME (MINUTES)
Figure 3-2. Schematic representation of progress
of refining in a top-blown basic-lined
BOF,21
3-21
-------
the circulation of heat. The circulation of heat in fuel-fired
soaking pit and reheat furnaces is accomplished by natural
convection and stack draft.29
The sizes of soaking-pit and reheat furnaces are usually
described by their hearth area. Soaking-pit furnaces range from
about 100 ft2 to over 300 ft2, and reheat furnaces range up to
4,000 ft2 of hearth area.29 The capacity of these furnaces is
determined primarily by the area of the surface of the material
to be heated; the shape, thickness, and composition of the
material; and the temperature of the material and the furnace.
The thermal efficiency (defined in this case as the amount of
heat required to raise the temperature of the charge from its
initial to its desired temperature as a percentage of the gross
heat input to the furnace) also varies widely due to differences
in the temperatures of the heated stock and of the charged
material, provisions for heat recovery, furnace insulation,
operating schedules, and heating requirements. Large production-
line furnaces, such as continuous furnaces with recuperators and
good insulation, generally give over 30 to 40 percent thermal
efficiency over an average month's operation.29
Soaking-pit furnaces provide uniform heating of ingots to
the desired temperature with a minimum of surface overheating.
The normal range for heating ingots is between 1,180 °C and
1,340 °C (2,150°F and 2,450 °F) ,29 Soaking pits also function as a
reservoir to correct irregularities in the flow of ingots between
the steelmelting facilities and the primary rolling mills.
The use of continuous casters is affecting the operations of
both soaking pits and reheat furnaces. There is less need for
soaking pits as more steel is routed to reheat furnaces and, in
general, the steel is entering reheat furnaces at higher
temperatures. The net effect is that, in the aggregate, less
fuel is being consumed in soaking pits and reheat furnaces. In
1991, about 76 percent of the steel produced was processed by
continuous casting rather than forming ingots that subsequently
needed to be processed in a soaking-pit furnace.7 Although the
trend to continuous casting will continue, some soaking-pit
capacity will likely be retained to cope with upset or irregular
conditions.
Reheat furnaces function to adjust or to maintain steel at a
suitable temperature as the steel is processed. They are similar
in many respects to soaking-pit furnaces. Reheat furnaces are
being integrated into continuous casting facilities for slabs,
blooms, and billets. When these are charged hot into a reheat
furnace, a moderate fuel savings results. Fuel inputs to large
soaking-pit and reheat furnaces range from 1.2 to 5.4 MJ/kg (1.0
to 4.7 MMBtu/ton) heated.27
3-22
-------
3.3.6 Annealing Furnaces
Flat-rolled steel products such as coils of sheet steel are
sometimes annealed to enhance some physical properties of the
product. Annealing involves subjecting the product to a
supplemental heat treatment under controlled conditions. Only a
portion of the steel produced is annealed. It is usually cold-
reduced coils that are annealed because hot-rolled coils are
"self-annealed.n The great bulk of annealing is done at
temperatures of about 675 °C (1,250°F). Some products are
annealed at much higher temperatures[760 °C to 1,210 °C,
(1,400 °F to 2,200 °F) ] .30 Annealing is done in a protective,
deoxidized atmosphere that is predominately nitrogen.
Annealing can be a batch or a continuous process although
the trend is strongly toward continuous annealing. Batch
annealing is done in a box-type furnace that consists of a
stationary base, several stools on which coils of steel are
stacked, individual cylindrical covers for each coil stack (to
provide for the protective atmosphere), and the furnace, which is
lowered by crane over the base with its load, stools, and
cylindrical covers. Subsequently, the charge is heated slowly
but uniformly to a specified temperature, soaked for a period of
time, and then allowed to cool. After a period of cooling, the
furnace is removed to begin a cycle on another base. However,
the inner covers are left in place to preserve the protective
atmosphere. After the charge has cooled to about 150 °C
(300 °F), the charge can be exposed to air without oxidizing. In
this cycle, the cooling period takes at least as long as heating
and soaking combined.
Continuous annealing is done in large furnaces in which the
steel coil is threaded vertically around rollers located at the
top and bottom of the furnace. Thus, the residence time of the
steel in the furnace is dramatically increased as it passes
continuously through the furnace. A typical, continuous
annealing furnace will have several zones including a gas-fired
heating zone, an electrically heated holding zone, an
electrically heated slow cooling zone, and a fast cooling zone.
Steel coil will thread through these zones at a rate of about
1,500 ft/min (7.6 m/s). Threading back and forth, the steel will
make 10 passes through the heating zone in about 20 seconds.
Subsequently, it will be cooled to about 540 °C (1,000 °F) in the
slow cooling zone and then to 115 °C (240 °F) in the fast cooling
zone. The entire process takes about 2 minutes and is carried
out in an atmosphere of nitrogen (95 percent) and hydrogen
(5 percent) .31
3-23
-------
-------
3.4 References
1. Kirk Othmer Encyclopedia of Chemical Technology. New York,
John Wiley. 3rd ed. 1981. Vol. 13. p. 552.
2. Lankford, W. T. , N. L. Samways, R. F. Craven, and H. E.
McGannon. The Making, Shaping and Treating of Steel, Tenth
Edition. Pittsburgh, Herbich & Held. 1985. p. 256.
3.
4.
.5.
6.
7.
Ref .
Ref .
Ref.
Ref.
Labee
2
2
2
2
/
/ P
/ P
/ P
/ P
C.
. 257.
. 541.
. 24.
. 599.
J. , a
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Labee, C. J., and N. L. Samways. Developments in the Iron
And Steel Industry, U.S. and Canada-1991. Iron and Steel
Engineer. February 1992.
Labee, C. J., and N. L. Samways. Developments In The Iron
And Steel Industry, U.S. and Canada-1990. Iron and Steel
Engineer. February 1991.
Directory, Iron and Steel Plants. Association of Iron and
Steel Engineers. Pittsburgh, PA. 1991.
American Iron and Steel Institute.
Report 1990. Washington, DC.
Annual Statistical
Telecon, Parker C., Research Triangle Institute, with
Steiner, B., American Iron and Steel Institute. February
12, 1992. Iron and steel industry
Parkinson, G. Steelmaking Renaissance. Chemical Engineer.
May 1991.
Usher, Thomas J. Steel Industry in the Nineties. Iron and
Steel Engineer. February 1991.
U.S..Environmental Protection Agency. Benzene Emissions
from Coke By-Product Recovery Plants - Background
Information for Proposed Standards. Research Triangle Park,
NC. Publication No. EPA-450/3-83-0162. May 1984.
Ref. 2, p. 156.
Ref. 2, p. 325.
Ref. 2, p. 326.
3-24
-------
18,
19.
20.
21.
22,
23.
24.
25
26.
27
28
29
30
31
32
Ref. 2,
Ref. 2,
Ref. 2,
Ref. 2,
Ref. 2,
Ref. 2,
P-
P-
P-
P-
P<
P-
307.
557.
507.
612.
611.
627.
Stallings, R. L. (Research Triangle Institute). Organic
Emissions from Ferrous Metallurgical Industries:
Compilation of Emission Factors and control Technologies.
Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/2-84-003.
September 1983. 108 pp.
Ref. 2, p. 671.
American Iron and Steel Institute.
Washington, DC. February 1989.
Steelmaking Flowlines.
U.S. Environmental Protection Agency. Control Techniques
for Nitrogen Oxides from Stationary Sources -Revised Second
Edition. EPA-450/3-83-002 . January 1983. pp 5-75, 76.
Prabhu, D. U. , and P. F. Cilione. 1990 Clean Air Act
Amendments: Technical/Economic Impacts. U.S. Coke and
Steelmaking Operations. Iron and Steel Engineer. January
1992.
Ref 2., pp. 841-843.
Ref. 2, p. 1111.
Ref. 2, p. 1116
Letter and attachments from Finke, H. P., Bloom Engineering
Company, Inc. to Neuffer, W. J., EPA/OAQPS/ISB. May 14,
1993.
3-25
-------
CHAPTER 4
UNCONTROLLED NOX EMISSIONS
This chapter presents available information on uncontrolled
NOX emissions from process facilities at iron and steel mills.
Section 4.1 discusses the mechanisms of NOX formation. Section
4.2 discusses emission factors and emissions from specific
sources, and Table 4-4 presents a summary of emissions from these
sources. Appendix A is a tabulation of the available emissions
data.
4.1 MECHANISMS OF NOX FORMATION
Nitrogen oxides refer to the combination of nitric oxide and
nitrogen dioxide. However, flue gas resulting from the
combustion of fossil fuel consists primarily of NO with NO
representing 90 to 95 percent of the total NOX due to kinetic
limitations in the oxidation of NO to NOj.1 There are three
fundamentally different mechanisms of NOX formation. These
mechanisms are (1) thermal NOX, (2) fuel NOX, and (3) prompt NOX.
The thermal NOX mechanism arises from the thermal dissociation
and subsequent reaction of nitrogen and oxygen molecules in
combustion air. The fuel NOX mechanism is the evolution and
reaction of fuel-bound nitrogen compounds with oxygen. The prompt
NOX mechanism involves the intermediate formation of hydrogen
cyanide, followed by the oxidation of HCN to NO. Natural gas and
most distillate oils have no chemically bound fuel nitrogen and
essentially all NOX formed from the combustion of these fuels is
thermal NOX. Residual oils and coals all have fuel-bound nitrogen
and, when these are combusted, NOX is formed by all three
mechanisms. The formation of prompt NOX is only significant in
very fuel-rich flames. These three mechanisms are discussed in
more detail in the following subsections.
4.1.1 Thermal NOX Formation
At the temperatures encountered in combustion air, both N2
and O2 molecules are dissociated into their respective atomic
states, N-and O. The subsequent reaction of these atoms to create
thermal NOX is thought to proceed through mechanisms first
formulated by Zeldovich:1'2
N2 + 0 <* NO + N
N + O2 5^ NO + 0
2 x 1014 exp (-76500/RT) (4-1)
kf= 6.3 x 109 exp (-6300/RT) (4-2)
In each reaction, kf is the forward rate constant for that
reaction. The high activation energy of reaction (4-1), 76.5
kcal/mol, shows that this reaction is the most temperature
sensitive. These relationships assume that the combustion has
-------
reached equilibrium. Because it has a large activation energy,
reaction (4-1) is generally believed to be rate limiting. Oxygen
atom concentrations are assumed to have reached equilibrium
according to
(4-3)
where M denotes any third substance (usually N2) ,
Experiments at atmospheric pressure have indicated that,
under certain conditions, the amount of NO formed in heated N2/
O2/ and argon (Ar) mixtures can be expressed as3
[NO] =
[N2] [02]1/2t
(4-4)
where
[NO] = mole fraction
kk = constants
T
t
temperature, and
time.
Although equation (4-4) cannot adequately describe NO formation
in a turbulent flame, it does point out several features of
thermal NOX formation. Like reactions (4-1) and (4-2) , it
reflects a strong, exponential relationship with NO emissions and
temperature. It also suggests that NO formation is directly
proportional to N2 concentration, residence time,, and the square
root of 02 concentration. .Moreover, equation (4-4) also suggests
several control measures for reducing thermal NOX formation:
• Reduce local nitrogen concentrations at peak temperatures,
• Reduce local oxygen concentrations at peak temperatures,
• Reduce the residence times at peak temperatures, and
• Reduce peak temperatures.
These control measures are discussed further in Chapter 5 .
For the purposes of the computations presented in Table 4-1,
the N2 and 02 concentrations in flue gas are defined to be 76
percent and 3.3 percent, respectively. The strong, exponential
dependency on temperature is evident in these tabulations, and a
dependency on 02 concentration can be inferred. Note, for
example, that the equilibrium NO, is much higher in air where the
02 concentration is higher than it is in flue gas.
4-2
-------
TABLE 4-1. CALCULATED EQUILIBRIUM CONCENTRATIONS OF NO AND N02 IN
AIR AND FLUE GAS4(ppm)
Temperature
K
Air
NO
Flue gas
NO,
NO
NO,
300
800
1,400
1,873
80
980
2,060
2,912
3.4{10)-10
2.3
800.
6,100.
2.1(10)"
0.7
5.6
12.
I.KIO)'10
0.8
250.
2,000.
3.3(10)-3
0.1
0.9
1.8
4-3
-------
In practice, flue gas NOX concentrations tend to be much
higher than suggested by the computations in Table 4-1. After
NOX is formed in the high temperatures of a flame, the rate of
its decomposition (the reverse of reactions (4-1) and (4-2)) is
kinetically limited at the lower temperatures and lower N and 0.
atomic concentrations characteristic of the postcombustion zone
of the flame. Thus, although NO, is thermodynamically unstable
even at high temperatures, its decomposition is kinetically
limited. The result is that the NOX concentration in flue gas is
higher than predicted by equilibrium and depends, to a large
extent, on the mixing of fuel and combustion air in the flame.
The factors discussed above affect thermal NOX formation on
a macroscopic scale. However, local microscopic conditions
ultimately determine the amount of thermal NO, formed. These
conditions are intimately related to a host of variables such as
local combustion intensity, heat removal rates, and internal
mixing effects.
Studies on the formation of thermal NO, in gaseous flames,
for example, have confirmed that internal mixing can have large
effects on the total amount of NOX formed.3 Burner swirl,
combustion air velocity, fuel injection angle and velocity, quarl
angle, and confinement ratio all affect the mixing between fuel,
combustion air, and recirculated products. Mixing, in turn,
alters the local temperatures and specie concentrations, which
control the rate of NOX formation.
Generalizing these effects is difficult because the
interactions are complex. Increasing swirl, for example, may
both increase entrainment of cooled combustion products (lowering
peak temperatures) and increase fuel/air mixing (raising local
combustion intensity). Thus, the net effect of increasing awirl
can be to either raise or lower NOX formation. In summary, a
hierarchy of effects, depicted in Table 4-2, produces local
combustion conditions that promote thermal NOX formation.3
4.1.2 Fuel N0r Formation
Nitrogen oxides can also be produced by oxidation of
nitrogen compounds contained in fossil fuels. The mechanisms are
more complex than the Zeldovich model for thermal NOX formation.
Several studies indicate that the fuel-bound nitrogen compounds
react to form NOX in two separate mechanisms, one a solid-phase
char reaction (with solid fuels) and the other a homogeneous gas-
phase reaction resulting from evolution and cracking of volatile
compounds (solid and liquid fuels).1
4-4
-------
m K
O
s
>Jj
H 0
03 r^
E M-4
-H
M 73
Pi pj
nJ
0 JJ 0
V-i nj SH
4J 0 4J
rH nj g nj
0 >H -H >H
> 0 4J 0
d) pj p[
• >~i € d) £s
dj j-i Q}
a 4J 3 4J
0 CQ
CD ^ O 4«i
>i rt ft rt
X 0 X0
O ft H ft
>i a o a
4J O 0 -H O
•rl -H g 4J -H
m 0 jj 3 m jj
C JJ CO rH Vi SH
0 rt 3 H i (H JJ
fi g O -H CO 0
•H rH O JJ ft) -H fl
ns u e -v. 73 o
« > -H rH N
O O H C SH co
3 O 0 M 4J
4J C! -H JJ -H 0
(C U) JJ JJ (t rH
>H -H -H (C ft CQ rH
0 co CQ ft73 O0 JH
ft 0O C r( rH -H
g 73 ft £3 (d 73 U S
0 >i gO -H CO
JJJJ X O -H i-H M-l JJ
-H O O 4J0 O5-I 5-1
4J U X3 U3 rt 0
0O 0 rH 01W 0ft Cl
rH rH >-| 0 -n N >H
C0-H 3 C! «H -H M 3
H>[X( Cu HO COO «
CO
CO
rH G
nj O
C -H
M 4-J
0 -H
4J 73
X73
cq nj
4-5
-------
The char nitrogen reaction is not well understood although
data show that the char nitrogen conversion to NOX is weakly
dependent on the flame temperature and strongly dependent on
stoichiometric ratio, fuel/air mixing, and on the char
characteristics.1"3 The precise relationships, however, are not
known. Conversion rates to NOX of 15 to 35 percent have been
documented.1
The gas-phase reaction is postulated to include a number of
intermediate species (e.g., HCN, HOCN, NHj) which are produced at
rapid reaction rates. The decay rate of the intermediate species
into N2 (fuel-rich) and NOX (fuel-lean) is slower by at least an
order of magnitude. These reaction rates are strongly dependent
upon the stoichiometric ratio and the gas phase fuel nitrogen
concentration and weakly dependent upon the flame temperature and
the nature of the organic nitrogen compound. It is the weak
influence of temperature on gas-phase NOX conversion that reduces
the effectiveness of NOX controls that rely on temperature effects
in the combustion of nitrogen-bearing fuels.1
The relative contribution of fuel NOX and thermal NOX to total
NOX emissions from sources firing nitrogen-containing fuels has net
been definitively established. Estimates indicate that fuel NOX is
significant and may even predominate.3 In one laboratory study,
residual oil and pulverized coal were burned in an argon/oxygen
mixture to eliminate thermal NOX effects. Results show that fuel
NOX can account for over 50 percent of total NOX production from
residual oil firing and approximately
coal firing.5
80 percent of total
NOX from
The nitrogen content of most residual oils varies from 0.1 to
0.5 percent. Nitrogen content of most U.S. coals lies in the 0.5
to 2 percent range. Fortunately, only a fraction of the fuel
nitrogen is converted to NOX for both oil and coal firing.3
Furthermore, the percent of fuel nitrogen converted to NOX ,
decreases as nitrogen content increases. Thus, although fuel NOX
emissions undoubtedly increase with increasing fuel nitrogen
content, the emissions increase is not proportional. In fact,
data indicate only a small increase in NOX emissions as fuel
nitrogen increases.3
4.1.3 Prompt NOX Formation
Prompt NOX is produced by the formation first of intermediate
HCN via the reaction of nitrogen radicals and hydrocarbons,
NO + HC + H,
HCN + H20
(4-5)
followed by the oxidation of the HCN to NO. The formation of
prompt NOX has a weak temperature dependence and a short lifetime
of several microseconds. It is only significant in very fuel-rich
flames, which are inherently low-NOx emitters.1
4-6
-------
4.2 NOX EMISSIONS FROM IRON AND STEEL MILLS
Integrated iron and steel mills import three basic raw
materials and process these to produce steel in a variety of
compositions and forms. The conversion processes, described in
Chapter 3 of this document, tend to be high-temperature processes
in which large quantities of fuels are consumed. As a
consequence, most of these processes are sources of NOX emissions.
Mini-mills and specialty producers do not have the full range
of processing facilities characteristically found in integrated
plants. They do not, for example, produce coke or sinter for use
in blast furnaces or operate blast furnaces to produce pig iron or
hot metal. Instead, they may produce steel by melting scrap in an
electric-arc furnace, and they may use some subset of the
processes found in integrated mills to produce a variety of steel
products. They, too, may use a number of high-temperature, fuel-
consuming processes and produce NO, emissions. In this section,
the major sources of NOX emissions inherent in steelmaking plants
are identified along with the fuels and process temperatures
characteristically used in these processes. Further, estimates of
uncontrolled NOX emissions are also given.
The following subsections discuss uncontrolled NOX emission
factors and NOX emissions from specific iron and steel process
facilities. These data appear in the literature or were provided
by Section 114 responses for a variety of units, many of which
cannot be easily interpreted for other conditions. These data are
summarized in tabular form in Section 4.2.11 and are tabulated in
Appendix A.
4.2.1 Emission Factors
NOX emissions data in the literature relevant to iron and
steel mill processes are limited. This lack of a good emissions
database is reflected in the major compilations of emission
factors.6-8 Table 4-3 is a compilation of emissions factors from
AP-426-8 and from the National Acid Precipitation Assessment Program
(NAPAP) .7>s Note that the units used for these factors are unusual
and not easily interpreted or used. Also, note that their quality
ratings are poor (D for AP-42 and E for NAPAP), i.e., they are
based on a single observation of questionable quality or an
extrapolation from another factor for a similar process. If these
emission factors are to be used, it has been recommended that
AP-42 factors be used whenever possible and that the NAPAP
emission factors be used when AP-42 factors are unavailable.6
These emission factors are not used in this document.
4.2.2 Coke-oven Underfiring
As described in Chapter 3, coke is made by the destructive
distillation of coal in the absence of air. During distillation,
4-7
-------
1
(Sj
a
Z
S
N
I
S
u
s
0!
b
2
<
b
2
0
H
CO
CO
H
W
a
H
g
53
w
8
g
H
2
1
^*
H
r3
I2C
pj
"^
m
u
•rl
c
J?
4J
•H *-i
C PJ
cT^ls
rH £5
;.
-rl C-3
!z *^» 0j
35
m
tion/proces
S
o
o
T^ T3 'O *O T5 *O
O Q) 0) 0) 0) Q)
0> 01 0) 01 01 01
*O (fl (tJ flj (0 fC (0
(U ,cj pC »C 1^1 .C J2
c u o o u u o
3 pH iH rH i-l i-l iH
»Q rt (fl rt nj nj nJ
o o o o o o
•Jj U U O O U O
10 CO OQ CQ CQ 02
v C C C! C C C
o O O O O O O
rH EH EH EH EH EH H
O O PO ^l H
• • O VO O O
CO O ... .
04 00 O O O 0
ro H '
O O
0 0
Cn
14-1
e
m
m j-i
^ (fl
O Cn Q)
U C OJ rH
ffl cn-ri M
(fl 4J Cn C Vi (fl a)
Cn uc cn-H -H Q) T3
3 -H a J3 >t-i I-H -H
to n3 Cn-H u v< to
CO O l-l J3 C (I) W i
d)OO VinStaaiTJOft
U to O &J5 3 3 C 0 0
O CQ tj ' t^ Oi ^y D O ^
r) >,
D< ffl
•o
0)
U
3
•5
o
a
^4
0)
JJ
c
•H
m
m
1
f>
o
ffg
TJ>8
a) c
JJ-H
C S
-H
CO
•a TI -a TJ -a T) -a
o) a) a) d) a)
u u u u u u u
3333333
*O *O *O *o *o *o *t3
o o o o o o o
Di Qi Qi Oi Oi d ft
Illiiil
H EH EH EH EH EH EH
00 C4 in
rH O)O O O H O\
o Ss o o o o H
I
rH
«1
u c
2 g_
o T
» o
10 4J
" AJ tt) 0)
0) U U
U (fl (fl -
m u en c u
c a a C c
fi I -H 3 -H
3 t) 4J «4H N
C 44 O tD
O O b (fl D) •>
•H 0 m J3 O 0 C C
4J rl 4-> (I) -rl -H
U (fl -rl C O 4J 1J
3 ft d) 3 (fl —
•o u a •• o d)
OTHOiOtn3lHCn
rl W C C C 4J C
ft jj -H -H -iH i -H
d)rHOO<80d)O
0) b3 CO CO E^ CJ tC CJ
4J
CO
MH
tt t Q
o
SO
-H
-H JJ
4J Rl
r m 5} co
m ra S °>
m W n d)
•§ 8 ° 8
d) O,8,&&?
^SsSt
rH LJ TO Tj 01
2 fl) S« Q) "jl
O1 ^ jj
o o S
r* i
K,^ H S^ *^
«w jj «3 .j
"H -rl "
1 •«&««
•C Qfr£3 Qi O
S «S «j-S
s3s«H-
0 g 0 g £
o o
L^ U1 L| & S
« .O 0
4-8
-------
a noncondensable gaseous product known as coke-oven gas (COG),
with a, heating value of about 22 MJ/Nm3 (570 Btu/f t3) , is evolved.
The major combustible component of COG is methane; thus, it is
very similar to natural gas. Some 30 to 40 percent of_the COG is
used to heat the coke oven, and the remainder is used in other
heating processes in the plant. Coking is accomplished at
temperatures that range from about 900 to 1,100 °C (1,650 to 2,010
°F) , and flue-gas temperatures can reach about 1,480°C (2,700 °F) .
A regenerative, reversible checker-work is used to preheat the
combustion air used to burn the COG, yielding fuel conservation
and a higher flame temperature. Nitrogen oxides emissions from
coke underfiring result from the high temperatures and the large
quantity of fuel consumed.9 Uncontrolled NOX emissions reported
for coke ovens are tabulated in Appendix A.
The wide range of NOXemissions from heating of coke ovens is
due to wide variations in the geometry of the combustion chambers
and the combustion variables of fuel and air mixtures,
temperature, humidity, and other factors. While the principle of
all coke ovens is the same, each coke oven and coke-oven operation
is unique.
Much of the coke oven data included in Appendix A is
incomplete such that it cannot be compared on a common basis with
other data. The data that can be compared on a common basis (ppm
@ 3% 02, Ib/MMBtu, and Ib/ton of product) are summarized below.
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
ppm @ 3% 0;
8
254
1452
802
448
Ib/MMBtu
11
0.10
2.06
0.66
0.72
Ib/ton
12
0.15
2.15
0.88
0.64
4.2.3 Sintering
Sintering is a continuous operation in which feed materials,
including coke breeze or fines and flue dust, are deposited in a
uniform bed on a traveling grate and ignited by passing under
burners fired with natural gas. Beyond the ignition region, air
is pulled,down through the bed. The coke fines in the bed
continue to burn, generating the heat required for the sintering
to proceed. Ignition temperatures range from about 1,150 °C
(2,100 °F) , where' ignition begins, to about 800°C (1,500°F) at the
ignition hood exit. As the coke fines in the bed burn, bed
temperatures reach 1,300 to 1,480 °C (2,370 to 2,700 °F) .
Available uncontrolled NOX emissions data for sinter plants are
tabulated in Appendix A. These data vary widely, and much of
these data are incomplete, i.e., test conditions are seldom
specified. A summary of these data is tabulated below. The data
in terms of Ib/ton show a wide range from 0.007 to 0.70.
4-9
-------
ppm @ 3%
Ib/MMBtu
Ib/ton
n(No. Samples) 2
Min. Value 284
Max. Value 395
Avg. Value 339
Std. Dev.
4
0.22
0.59
0.47
0.17
16
0.007
0.70
0.34
0.28
4.2.4 Blast Furnaces and Blast Furnace Stoves
The blast furnace is fueled by the combustion of coke that,
along with iron ore and flux, is a component of its charge, and
the off-gas is recovered and burned for its fuel value. It is a
closed unit that does not have atmospheric emissions. However,
the associated blast furnace stoves do generate NOX emissions.
These stoves, typically three or four for each blast furnace, burn
the blast furnace gas (BFG) to preheat the combustion air supplied
to the blast furnace. The BFG has a low heating value ranging
from 3.0 to 3.5 MJ/Nm3 (80 to 95 Btu/ft3) .9 More recently, values
ranging from a low of 2.5 MJ/Nm3 (65 Btu/ft3) have been reported
with changes in blast furnace burdens cited as the cause of the
deterioration.16 Blast furnace gas burns with a low flame
temperature. Its combustible component is carbon monoxide, which
burns clean.6 However, the BFG is often enriched with other fuels
(e.g., natural gas or COG) to obtain blast-air temperatures as
high as 1,100 °C (2,010 °F) .3
The available uncontrolled NOX emissions data for blast
furnace stoves are tabulated in Appendix A. These data vary
widely in terms oflb/MMBtu and Ib/ton, and much of these data are
incomplete, i.e., test conditions are seldom specified. A summary
of these data is shown below.
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
ppm @ 3% 0;
1
28,
Ib/MMBtu
11
0.00.2
0.057
0.021
0.019
Ib/ton
10
0.003
0.072
0.037
0.022
4.2.5 Basic Oxygen Furnace
The basic oxygen process of steelmaking is an exothermic
process that does not require the burning of fossil fuels. The
process does produce off-gases, principally carbon monoxide, which
are usually burned prior to discharge into the atmosphere. Some
newer BOF facilities use suppressed combustion hoods, which
suppress the influx of air, and the off-gases are subsequently
flared. A potential for NOX generation does exist during the
combustion of waste gases. Nitrogen oxides production from this
process has been estimated to range from 30 to 80 ppm, or 180 to
500 ng NOx/kg (0.36 to 1 Ib N0x/ton) of steel produced.10
-------
Appendix A presents NO,emissions data for BOFs, including
Q-BOPs, and these data are summarized below. Data for which the
status of the furnace is unknown (i.e., whether or not the furnace
is in an 02blow period) are not included in the summary. A wide ._
range in terms of ppm@ 3%02 is reported (18-180) but in terms of
Ib/ton the range is much narrower (0.042-0.22) .
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
During 0*, Blow Period
ppm (aver)
12
12.3
84.0
24.0
19.6
ppm @ 3% 0?
7
18
180
58
56
Ib/ton
7
0.042
0.222
0.119
0.059
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Durincr Non 0, Blow Period
ppm (aver)
2
14.3
14.5
14.4
ppm @ 3% 0?
2
200.
366.
283.
Ib/ton
4.2.6 Electric-Arc Furnace
The electric arc furnace largely transfers the generation of
NOX emissions from the steelmelting facility to a utility plant
where it is easier to control. The only use of fossil fuels in
the electric arc facility is for scrap preheating, which may or
may not be practiced. However, some EAFs also fire oxy-fuel
burners in addition to electric arcs during meltdown.17' The
available NOX emissions data, presented in Appendix A and
summarized below, suggest that concurrent oxy-fuel firing during
meltdown does increase NOX emissions above the emissions from
electric-arc melting alone as NOX increased from 12 to 98 ppm. The
range of NOX emissions is narrow from 0.5 - 0.6 Ib/ton and 83 - 100
Ib/heat.
EAF With Concurrent Oxy-fuel Firing
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
ppm (avg.)
6
80.
110.
98.
10.
Ib/ton
6
0.50
0.60
0.54
0.05
Ib/heat
6
83
100
89
8.2
-------
EAF Without Concurrent Oxy-fuel Firing
n(No. Samples)
Min. Value
Max. Value
Avg. Value
ppm (avg.)
2
7
17 '
12
4.2.7 Soaking Pits
Soaking pits reheat or hold slabs, ingots, and other forms of
steel to temperatures suitable for further shaping and processing.
They predominantly use natural gas for fuel although some COG and
fuel^oil are used. In recent years, the use of soaking pits has
declined as the use of continuous casters increased.
The available NOX emissions data for soaking pits are included
in Appendix A. These data correspond to a range of conditions
that are not always specified. A summary of these data are
tabulated below:
Summary of all Appendix A Soaking Pit Data
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
Ib/MMBtu
5
0.064
0.148
0.108
0.039
Ib/ton
9
0.091
0.361
0.175
0.078
ppm @ 3% 0,
3
49
689
307
337
A sub-set of the Appendix A data for soaking pits includes
more^test-condition information than is available for the
remainder of the data. It is clear that these data correspond to
a complete cycle of the soaking pits. Hot and cold charges are
specified and the temperature of the combustion air is given. A
summary tabulation corresponding to this sub-set is presented
below, and a distinction is made between soaking pits with
preheated combustion air and those with cold or ambient combustion
air. NOX for the preheat air is approximately doubled that for the
cold air soaking pit. '
Summary of a Sub-set of Appendix A Soaking Pit Data
Ib/MMBtu
Cold Air
n(No. Samples 2
Min. Value 0.064
Max. Value 0.069
Avg. Value 0.066
Std. Dev.
Preheat Air
3
0.121
0.148
0.135
0.016
Ib/ton
Cold Air
2
0.091
0.104
0.098
Preheat Air
3
0.164
0.194
0.164
0.016
-------
4.2.8 Reheat Furnaces
Most reheat furnaces are either recuperative- or
regenerative-fired, i.e., they preheat the combustion air in order
to increase fuel efficiency. Some use cold combustion air. The '•
temperature of the combustion air has a large impact on
uncontrolled NOX emissions. Increasing the combustion air
temperature from 38 °C (100 °F) to 540° C (1,000 °F), for example,
will increase uncontrolled NOX emissions by a factor of about 6.
Combustion air preheated in regenerators has a much higher
temperature than air preheated in recuperators. While the higher
combustion air temperature increases furnace fuel efficiency, it
also increases NOX generation and NOX emissions. Consequently,
regenerative-firing is not usually practiced without combustion
modifications for NOX control, but there may be exceptions.
Appendix A contains a tabulation of available uncontrolled
emissions data for reheat furnaces. These data correspond to a
variety of condition, some known and some unknown. Summaries of
the Appendix A data follow. The first summary represents all of
the Appendix A data except three values that are more than an
order of magnitude lower than any other value and two values from
an ejector stack that may not be representative. Included are
data for both regenerative- and recuperative-fired furnaces and
others for which the firing configurations are not specified.
This tabulation shows a wide range (0.023 - 0.91 Ib/MMBtu).
Summarv of Appendix A Reheat Furnace Data
n(No. Samples)
Min. Value
Max. Value
Avg. Value
Std. Dev.
Ib/MMBtu
28
0.023
0.909
0.226
0.198
ppm @ 3% 0,
14
65.
740.
292.
166.
Ib/ton
11
0.054
0.327
0.198
0.084
The following is a summary of the Appendix A data for
recuperative-fired reheat furnaces.
Summary of •• Recuperative -fired Reheat Furnace Data From Appendix A
Ib/MMBtu
n(No. Samples) 16
Min. Value 0.080
Max. Value 0.40
Avg. Value 0.200
Std. Dev. 0.101
ppm @ 3%
8
65.
326.
220.
81.5
Ib/ton
8
0.157
0.327
0.237
0.058
The following is a summary of the Appendix A data for
regenerative-fired .reheat furnaces. Data from the recuperative-
fired furnace is much lower than regenerative reheat furnaces (0.2
vs. 0.91 Ib/MMBtu and 220 vs. 740 ppm at 3% 02. ,
4-13
-------
Summarv of Regenerative-fired Reheat Furnace Data From Appendix A
n(No. Samples)
Min. Value
Max. Value
Ave. Value
Ib/MMBtu
2
0.675 ,
0.909
0.792
ppm @ 3%
2
550.
740.
645.
None of the reheat furnace emissions data in Appendix A are
known to correspond to cold combustion air firing. For the
purposes of this Chapter, the uncontrolled emissions data for
cold-air-fired reheat furnaces are estimated using controlled
emissions data from an LNB-controlled reheat furnace (80 ppm. at 3%
02) assuming a control efficiency of 27 percent.18 Thus,
uncontrolled emissions from cold-air-fired reheat furnaces are
estimated to be 110 ppm at 3% 02 (0.135 Ib/MMBtu) .
4.2.9 Annealing Furnaces
Annealing processes vary greatly depending on specific
objectives and starting materials. In general, annealing furnaces
subject steel products to a planned time-temperature profile in a
reducing atmosphere with heating supplied from gas-fired, radiantf-
tube burners. There are two basic approaches. Batch annealing
typically includes slow heating of the steel to a desired
temperature, soaking at temperature, and controlled slow cooling,
a process that may take several days. In continuous annealing,
steel sheet is passed continuously through an annealing furnace
where it is subjected to a desired time-temperature profile.
Temperatures typically peak at about 675 °C (1,250 °F). However,
some processes may require temperatures as high as 760 to 1,200 °C
(1,400 to 2,200 °F).
Uncontrolled NOX emissions data from operating annealing
furnaces are not available. The emissions data included in
Appendix A are based on estimates and on laboratory measurements
of emissions from furnaces without NOX controls. These data were
provided for new furnaces or, in one case, for a furnace still
under construction. In each case, the furnace is regenerative
fired. For the purposes of this Chapter, the average of the two
values is used as the uncontrolled emissions from a regenerative-
fired annealing furnace, i.e., 775 ppm at 3% 02 (0.952 Ib/MMBtu).
The actual value would be affected by the temperature of the
preheated air. The 1,000 ppm value is possible for regenerative-
fired furnaces without Nox controls (because the air preheat is
very high with regenerator technology), but this configuration is
not used in practice (there may be exceptions) .16
In lieu of uncontrolled emissions data for recuperative- and
cold-air-fired annealing furnaces, estimated values are used for
the purposes of this Chapter. Values of 204 and 450 ppm at 3% 02
(0.25 and 0.55 Ib/MMBtu) have been suggested as reasonable for
recuperative-fired furnaces.16'19 The average of, these is 327 ppm at
3% 02 (0.402 Ib/MMBtu). A value of 120 ppm at 3% 02 (0.147
Ib/MMBtu) has been suggested as reasonable for cold-air-fired
furnaces.19
-------
4.2.10 Galvanizing Furnaces
Appendix A lists uncontrolled NOX emissions from two
regenerative-fired galvanizing furnaces.14 (These furnaces are
described as galvanizing/galvannealing and aluminizing furnaces.)-.
One estimate is based on measurements at laboratory facilities
with NOX control features disabled, and the second is based on
field measurements. These values are 1,000 and 880 ppm at 3% 02,
respectively, and the average is 940 ppm at 3% 02 (1.15 lb/MMBtu) .
No uncontrolled NOX emissions data are available for
recuperative- or cold-air-fired galvanizing furnaces. In lieu of
any data, values for annealing furnaces are used as estimates for
the purposes of this Chapter, i.e., 327 ppm at 3% 02 (0.402
lb/MMBtu) and 120 ppm at 3% 02 (0.147 lb/MMBtu) for recuperative-
fired and cold-air-fired galvanizing furnaces, respectively.16'19 In
practice, the actual value would depend on many variables
including furnace and combustion air temperatures.
4.2.11 Summary
Table 4-4 is a summary tabulation of uncontrolled NOX
emissions from the major NOX-emitting sources at iron and steel
mills. These values are based on the available emissions data
presented in Appendix A and, in instances where no data are
available, on estimates. There is not an abundance of data, and
test conditions are often not specified. For example, NOX
emissions will vary with furnace operating temperature and
combustion air temperature, and these conditions are usually not
specified. In each case, the text of this Chapter and Table 4-4
includes information indicative of the quantity and quality of the
data. The uncontrolled NOX emissions data tabulated in Table 4-4
are used to calculate NOX emissions reductions and cost
effectiveness of NOX controls in Chapters 5 and 6, respectively.
4-15
-------
I
H
CO
CO
H
$
Q
I
§
O
CJ
m
C
o
•H
m
m
"g
W
M
o
$5
CD
en
i
m
c
o
•H
to
m
•H
K
O
^
0)
?
Q)
S
g
4J
^N^
ja
3
rQ
5
S
\
rH
O
tV*
g (Y)
Q<
a©
a
g
4J
^s^
A
rH
C
*J
g
yj
^o
rH 3
C
o"
,<,,
g fO
Pi
ft ©
^i
4J
•H
rH
•H
U
n)
^
03
m
0)
u
O
H
o<
in
H
•
CN
in
H
o
vo
0
CN
O
H
O
I CN
T* in
in TJI
CN H
«N
H
CN
CO
CO
0
H
H
CO
in
vo
0
CO
o
o
CO
H1
C-H
0) )H
^ -rl
O M-l
i rl
0) CD
O C
U 3
i-l
^
CO
•
o
1
r-
0
o
o
1
ro eri
H CO
CN in
O 0
in
Cn
ro
i
CO
^*
CN
VO
H
H
*3*
ro
0
^
in
vo
<"
o
CN
O
^*
ro
Cn
C
•H
CD
4J
C
•H
CO
CN
r-
o
•
o
i
ro
o
o
o
i
CN t--
o in
o o
O 0
i
o
H
t-
m
0
0
H
H
H
CN
O
0
H
CO
CN
CD
U
m o
i-H U
CQ CO
CN
CN
CN
•
O
I
^4 1.
O
O
1 1
vo
o vo
co ro
H <
t O
CO 0
rH CN
f- I
crt
H i
H
o
1 1
1 1
t-CN
co ro
in co
CN
Cn
C
-H S
•O * O
3 f) l~l
O ^
c c c
<1)-H „ 0
0)"-^ O C
X1 to Oi Gi
O 0) C C
U -H -H
U 16 -~ )-l M
-H (3 0< 3 3
CO H O "O "O
«! 3 CQ
03 4-1 O
O
vo
«
0
1
O i
in
o
i i
i
f^
H B
I Q
t- a
VO i
^ 1
in
0
i i
i i
i
CN
i
CN
__,
~ 0)
Cn 3
-H i
• 4J-H X
O C"H O
'h 0)
(d >-l rH JJ
i 1-1 Q) 3
U 3 3 O Cn
-H d) U iw J3 C
JH U C i 4J -H
4J (8 O ><-H ^1
U C U X S -H
CD }-i — • o ^*m
rH 3
H M-l
crt o
rH rH
• •
O 0
1 1
VD CM
rH O
p o
I i
rH CO *f CA
CN Tjl VO VO
H H O O
O O P O
i i
ro CN
vo en
rH O
o o
fi CN
in vo
ro vo
rH O
o o
1 1
1 . 1
^
•rl
n)
•H
*i a
03 0 •
S" 1
(d u
Cn 0)
C X! -O
•HO) rH
X iH O
it! ft U
0
co
CN
ro
•
p
• in i
H
0
o
*4*
•
0
in en o
f^ 0 CO I
VO CD P
P P 0
p
*J( VO
t^ CN
i ro i
p i
in in
in vo
l CO i
i ro i
CN
0
rH
(U
4J
VO O
CN H S5
CN o in
en o ro
t-- N rH
o o o
rH
0)
^
O
CN CO 2
in P o
I1 CN H
VO CN H
•0
CO i I 0)
CD 0) CD tH
U > > -H
id -H -H IH
54J 4J I
(d (d rl
3 ^i v< -H
b cu cu (d
c a •
4J CD 3 T3
(d Cn-O u T) r-i
Q) CD CP CD (D O
,C Pi rl Pi rl U
CD -H -H
K > -H
3 -rl -H MH
M-l 4J 4-> i
(d cd M
(71 ^| ^ -H
Co) o id
-rl C pt i.
rH 0) 3 T!
(d Cn-O u -O IH
CD Q) CD 0) 0) O
C Pi rl PJ rl U
S 'H "H
4-16
-------
1
H
CO
CO
H
X
g,
Q
W
§
O
U
o
>H
CO
I
m
g
-H
m
a
-H
w
M
o
55
0)
Ol
c
»!
a
C
0
-H
(Q
m
w
if
O
tn
($
0)
£
c
o
4J
A
i-l
^
4J
CD
g
•x^
.Q
r-H
es
O
<*>
C f*l
a
a®
c
8
4J
~~x,
rH
C
4J
1
-x.
-1 3
C
o
<*>
& M
a©
>•
-H
r-H
-H
U
ft)
1^4
*-x^
to
m
> -H
-H -H >H
OlW 4J i
C IS IB >H
NO) 0) (1) 0) Q) 0) 0
rH o; n a M u
It! -H -H
C5 "4-1 <4-l
4J
C
s
m
4J 1C
m o
0) >H -
3 t-i M ft!
I-H «H g m
itf -O -HO
!> m -O w
4J CL> m p
>H It) 4J 0) tt)
0) g 3 -H
£ -H a e r-i
e -u E o
3 to o ^ C
E! W U fc H
C i-i CN m ^
0) 0) 0) 0)
4J 4J 4J 4J
O O O O
2 a a a
4-17
-------
-------
4.3 REFERENCES
1.
5.
6
7.
8.
9.
10,
11.
12
Campbell, Lisa M., D. K. Stone, and G. S. Shareef (Radian
Corporation). Sourcebook: NOX Control Technology Data.
Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/2-91-029, July
1991.
Zeldovich, J. The Oxidation of Nitrogen in Combustion and
Explosions. Acta. Physiochem. 21(4). 1946.
U.S. Environmental Protection Agency. Control Techniques for
Nitrogen Oxides Emissions from Stationary Sources-Revised
Second Edition. EPA-450/3-83-002. January 1983.
U.S. Environmental Protection Agency. Control Techniques for
Nitrogen Oxide Emissions from Stationary Sources. Pub. AP-
47. National Air Pollution Control Administration,
Washington, DC. 1970.
Ref. 3, p. 4-4. . -
Barnard, W.R. (E.H. Pecham and Association, Inc.). Emission
Factors for Iron and Steel Sources-Criteria and Toxic
Pollutants. Prepared for U.S. Environmental Protection
Agency. Research Triangle Park, NC. Publication No. EPA-
600/2-90-024. June 1990. 7 p.
Stockton, M.B., and J.H.E. Stalling (Radian Corporation).
Criteria Pollutant Emission Factors for the 1985 NAPAP
Emissions Inventory. Prepared for U.S. Environmental
Protection Agency. Research Triangle Park, NC. Publication
No. EPA-600/7-87-015. May 1987. 212 p.
U.S. Environmental Protection Agency. Compilation of Air
Pollutant Emission Factors, Vol. 1: Stationary Point and
Area Sources. Research Triangle Park, NC. Publication No.
AP-42 (GPA 055-000-00251-7), Fourth Edition (including
supplements A, B, C, and D). September 1991.
Ref. 3, pp. 5-72, 5-73.
Ref.-3, p. 5-75.
Ref. 3, p. 5-77.
Hovis, J. E. Low NOX Burners for Metallurgical Furnaces.
and Steel Engineer. pp. 43-46. June, 1986.
Iron
4-18
-------
13. Letter and attachments from Sulc, D. A., Nucor Steel,
Crawfordsville Division, to Jordan, B. C., EPA/OAQPS.
October 6, 1992. Response to Section 114 letter on iron and
steel mills. ;
14. Letter and attachments from Gilbert, F.C., North American
Manufacturing Company, to Jordan, B.C., EPA/OAQPS. November
11, 1992. Response to Section 114 letter on low-NOx burners.
15. Letter and attachments from Harmon, M. L., LTV Steel Company,
to Jordan, B. C., EPA/OAQPS. July 2, .1992. Response to
Section 114 letter on iron and steel mills.
16. Letter and attachments from Finke, H. P., Bloom Engineering
Company, Incorporated to Neuffer, W. J., EPA/OAQPS/ISB.
May 14, 1993.
17. Letter and attachments from Felton, S. S., Armco Steel
Company, L. P., to Jordan, B. C., EPA/OAQPS. June 26, 1992.
Response to Section 114 letter on iron and steel mills.
18. Letter and attachments from Gilbert, F. C., North American - .
Manufacturing Company, to Parker, C. D., Research Triangle -
Institute. May 14, 1993. Response to a request for
information.
19. Letter from Dickinson, G., Engineered Combustion Systems,
Inc., to Parker, C. D., Research Triangle Institute. May 19,
1993. Response to a request for information.
4-19
-------
CHAPTER 5
NITROGEN OXIDES CONTROL TECHNIQUES
5.1 BACKGROUND
Because the types of process facilities involved in the
making of iron and steel are numerous and different, they are
discussed separately in this chapter. Control techniques
applicable to each process facility are identified in the
appropriate subsection.
Control techniques for NOX emissions can be placed into one
of two basic categories: techniques designed to minimize NOX
generation and techniques designed to remove previously generated
NOX from the waste effluent stream. Combustion modification
techniques, including low NOX burners(LNB's) and flue gas
recirculation(FGR), fit into the first category. Add-on flue gas
treatment techniques such as selective catalytic reduction(SCR)
and selective noncatalytic reduction(SNCR) are examples of the
second. Table 5-1 summarizes these techniques and indicates that
combustion modifications tend to be less effective than flue gas -
treatment as a NOX control. Of these techniques, low excess
air(LEA), LNB's, FGR, SCR, and combinations of these controls are
known to be used to control NOX emissions from iron and steel
process facilities. It should be noted that LNB's and often FGR
incorporate staged combustion in their burner design.
5.2 CONTROL TECHNIQUES APPLICABLE TO IRON AND STEEL FACILITIES
Historically, few facilities found at iron and steel mills
have NOX controls. Requests for information from the industry
and control equipment vendors on NOX controls yielded information
on only a few facilities with NOX controls. Those facilities are
either new or are still under construction, and little
performance data are available. The control techniques known to
be used (LEA, LNB's, FGR, and SCR) are discussed in Section
5.2.1. These techniques are applied to reheat furnaces,
annealing furnaces, and galvanizing furnaces. The applications
are discussed in Section 5.3. Other control techniques that may
be applicable to iron and steel processes are discussed in
Section 5.2.2.
5.2.1 Control Techniques Applied
Control techniques known to have been used on iron, and steel
process facilities are LEA (reheat furnaces), LNB (reheat
furnaces and galvanizing furnaces), LNB plus FGR (reheat
furnaces, annealing furnaces, and galvanizing furnaces), SCR
(annealing furnaces), and LNB plus SCR (annealing furnaces).
5-1
-------
TABLE 5-1 OVERVIEW OF NOX CONTROL TECHNIQUES1
Technique
Emission
reduction
Comments
Combustion
modification
Low excess air
Staged
combustion
Flue gas
recirculation
Low-N0x
burners
10 to 50+
percent
One or more of
these has the
potential for
application to any
combustion unit.
Performance and
costs are highly
site specific.,
Flue gas treatment
Selective noncatalytic 30 to 80
reduction percent
Ammonia and urea
inj ection are used
on fossil-fuel-
and refuse-fired
units in the U.S.,
Japan, and Europe.
Selective catalytic
reduction
50 to 90+
percent
Used in U.S. on
gas turbines,
interal combustion
engines, process
heaters and some
boilers. Used on
gas-, oil-, coal-,
and refuse-fired
units in Japan and
Europe.
5-2
-------
These control techniques are discussed in the following
subsections.
5.2.1.1 Low Excess Air. Low excess air(LEA) is a
combustion modification technique in which NOX formation is
inhibited by reducing the excess air to less than normal ratios.2..
It reduces the local flame concentration of oxygen, thus reducing
both thermal and fuel NO,, formation. It is easily implemented
and is used extensively in both new and retrofit applications,
either singly or in combination with other control measures. It
can be used with all fuels and all firing methods.2'3
The potential of LEA as a NOX control technique is limited
by the onset of smoke or CO emissions. Tests on utility boilers
have indicated that LEA firing can reduce NOX emissions between
16 and 21 percent as compared to baseline levels.2 Other sources
have suggested reductions of up to 15 percent.4 In the case of.
utility boiler applications, LEA also increases thermal
efficiency. It decreases the volume of combustion air to be
heated, allowing more heat of combustion to be transferred, thus
lowering fuel requirements for a given output.3 To maintain
proper control of the furnace pressure, positive pressure must be
maintained in the furnace to prevent the influx of tramp air.s
A number of other factors affect the excess air levels that
can be implemented. These include the type of fuel fired,
uniformity of the air/fuel ratio, air and fuel control lags
during load swings, and other combustion control features such as
staging of fuel or air.2
For utility boilers, LEA firing is considered a routine
operating procedure and is incorporated in all new units.
Because it is efficient and easy to implement, expectations are
that LEA will be used increasingly in other applications as well.
Although it is a feasible technique for furnaces, specifically
commercial furnaces, the trend in NOX control for these sources
has been in improved burner design.2
In a reheat furnace application at a steel mini-mill, LEA
as a NOX control technique yielded a NOX emission reduction of
about 13 percent.6
5.3.6.
This application is described in Section
5.2.1.2 Low NOr Burners. Low NOX burners(LNB) have been
used since the early 1970's for thermal NOX control. These
specially designed burners employ a variety of principles
including LEA, off-stoichiometric or staged combustion (OSC), and
FGR. The objective in the application of LNB's is to minimize
NOX formation while maintaining acceptable combustion of carbon
and hydrogen in the fuel.2
5-3
-------
Literature references to LNB applications discuss industrial
and utility boiler and process heater applications primarily.
However, they are applicable to other combustion processes
including reheat furnaces and continuous annealing and
galvanizing line furnaces.7'8 Annealing and galvanizing line
furnaces are more like boilers and process heaters in that their
operating temperatures are in the same range. Reheat furnaces
and soaking pits, in general, operate at much.higher
temperatures.
The differences between a low NO, burner and a burner
featuring LEA or FGR, for example, are not always clear. In
general, LNB's implement LEA,- OSC, FGR, or a combination of these
techniques. In a stricter sense, LNB's have been defined as
burners that control NO^ formation by carrying out the combustion
in stages (OSC) and, further, by controlling the staging at and
within the burner rather than in the firebox.3 Consistent with
this definition, there are two distinct types of designs for
LNB's: staged air burners and staged fuel burners. Staged air
burners are designed to reduce flame turbulence, delay fuel/air
mixing, and establish fuel-rich zones for initial combustion.
The reduced availability of oxygen in the initial combustion zone
inhibits fuel NO, conversion. Radiation of heat from the primary-
combustion zone results in reduced temperature as the final
unburned fuel gases mix with excess air to complete the
combustion process. The longer, less intense flames resulting
from the staged stoichiometry lower peak flame temperatures and
reduce thermal NOX formation.3
Figure 5-1 is an illustration of a staged-air LNB. All of
the fuel and less-than-stoichiometric air (primary air) are
initially supplied to the primary combustion zone. Staged or
secondary air is supplied beyond the primary zone where
combustion of the fuel is completed, again under off-
stoichiometric conditions.3 Staged air burners generally
lengthen the flame configuration so their applicability is
limited to installations large enough to avoid flame impingement
on internal surfaces. The installation of replacement burners
may require substantial changes in burner hardware, including air
registers, air baffles and vanes, fuel injectors, and throat
design. Existing burners can incorporate staged air burner
features by modifying fuel injection patterns, installing air
flow baffles or vanes, or reshaping the burner throat. Staged
air burners are effective with all fuel types.3
Figure 5-2 is an illustration of a staged-fuel LNB.3 Staged
fuel burners mix a portion of the fuel and all of the air in the
primary combustion zone. The high level of excess air greatly
lowers the peak flame temperature achieved in the primary
combustion zone, thereby reducing formation of thermal NO,. The
secondary fuel is injected at high pressure into the combustion
5-4
-------
Staged air is mixed
with the combustion
products from the
primary zone. This
lowers the peak flame
temperature which
limits the formation
of NO.
Staged air
Secondary air
Sub-stoichiometric
conditions in primary zone
increase the amount of
reducing agents (H2 & CO).
Primary Air
Figure 5-1.. Staged-air Low NOX burner.
5-5
-------
Secondary Combustion
High Air-
Ratio in
Secondary Fuel
Primary Fuel
Connection
to-Fuel
Primary Zone
Combustion Air
Secondary Fuel Connection
Figure 5-2. Stagsd-fusl Low NOX burners.
5-6
-------
zone through a series of nozzles that are positioned around the
perimeter of the burner. Because of its high velocity, the fuel
gas entrains furnace gases and promotes rapid mixing with first-
stage combustion products. The entrained furnace gases simulate
flue gas recirculation. Heat is transferred from the first-stage.
combustion products prior to the second-stage combustion. As a
result, second-stage combustion is achieved with lower partial
pressures of oxygen and temperatures than would normally be
encountered.3
Unlike the staged air burner, staged fuel burners are
designed only for gas firing. The staged fuel burner is able to
operate with lower excess air levels than the staged air burner
due to the increased mixing capability resulting from the high-
pressure second-stage fuel injection. An additional advantage of
the staged fuel burner is a compact flame. Cooling of the
combustion products from the first-stage zone in the staged air
burner is accomplished primarily by radiation to the process.
However, in a staged fuel burner the entrained furnace products
give additional cooling to the flame. This particular
characteristic permits more intense combustion with reduced NOX
levels.3
In addition to the common viewpoint of LNB as a burner that
stages combustion within the burner rather than in the firebox
other designs also include FGR within the burner. The FGR
feature may be a more effective control than OSC. For example, a
radiant tube burner that uses a vitiated air stream (i.e., a
burner featuring FGR) is being developed. The combustion takes
place within the burner, and reduced NOX generation has been
demonstrated.9
Full-scale tests on boilers in Japan have shown NOX
emissions reductions from 40 to 60 percent with gas- fired LNB's.
Other subscale test results with LNB's include 55 percent NOX
reduction with specially designed nozzles and burner blocks, and
55 percent NOX reduction with designs that create fuel-rich and
fuel-lean combustion zones (staged combustion).2
Other estimates on NOX reductions for LNB's include 20 to 50
percent NOX reductions in oil- and gas-fired package boilers
using shaped fuel injection ports and controlled air-fuel mixing,
and 55 percent (typical) NOX reductions in process furnaces using
self-recirculating (an FGR feature) and staged combustion type
burners. In summary, LNB's appear to offer potential NOX
reductions around 50 percent in boiler and process heater
applications.2
There are three known applications of LNB's used by the iron
and steel industry in the United States. LNB's are being used at
a reheat furnace, an annealing furnace, and a galvanizing
5-7
-------
furnace. The use of LNB's may change the heat transfer
characteristics and the oxidizing/reducing atmosphere of a
furnace. This is not a problem for new furnaces as the design
and construction of new furnaces can accommodate these changes.
For existing furnaces, in particular annealing furnaces, which
operate at a very specific flame and furnace geometries to
achieve a specific "set point" at which steel processing is most
efficient, major problems may occur for a specific furnace
without a large amount of equipment reconstruction.
5.2.1.3 Low NOT Burner Plus Flue Gas Recirculation. As
discussed earlier, LNB's are often designed to incorporate OSC
and FGR within the burners. However, additional, external FGR is
added to enhance NOX control in some furnace installations. Flue
gas recirculation is implemented by recycling a portion of the
flue gas to the primary combustion zone. This principle is
illustrated for a boiler in Figure 5-3. It reduces NOX formation
by two mechanisms. The recycled flue gas contains combustion
products that act as inerts during combustion and lower the peak
flame temperature, reducing thermal NOX formation. To a lesser
extent, FGR also reduces thermal NOX formation by lowering the
oxygen concentration in the primary flame zone. These factors
lower flame temperature, altering heat distribution and reducing
fuel efficiency.
Flue gas recirculation has been applied principally to
boilers and to a few process heaters. Emissions tests on utility
boilers with flue gas recirculation control have achieved NOX
reductions of 13 to 60 percent.2
Controlled NOX emissions data are available for two reheat
furnaces, an annealing furnace, and a galvanizing furnace with
LNB plus flue gas recirculation control.9'10 These data are
discussed in Section 5.3.
5.2.1.4 Selective Catalytic Reduction. SCR is the most
advanced of the flue-gas treatment methods for reducing NOX
emissions and the one on which the great majority of flue gas
treatment units is based.2 The principle of SCR is illustrated
in Figure 5-4. SCR units use NH3 to selectively reduce NOX. The
ammonia, usually diluted with air or steam, is injected through a
grid system into the flue/exhaust gas stream upstream of a
catalyst bed. On the catalyst surface, the ammonia reacts with
NOX to form molecular nitrogen and water. Depending on system •
design, NOX removal rates of 80 to 90 percent and higher are
achievable.3
5-8
-------
Secondary
air wind-
box
Suck
breechlnf
Figure 5-3. Illustration of flue gas recirculation
11
5-9
-------
The major reactions that occur in SCR are the following:
4NO
2N02
4NH3
4NH3
4N2 + 6H20 , ' and
3N
(5-1)
(5-2)
Of these, reaction (5-1) predominates since flue gas NOX consists
primarily of NO. The optimum temperature range for these
reactions is generally 260 to 540 °C (500 to 1,000 °F) . At
higher temperatures, the NH3 oxidizes to NOX or ammonium nitrate
and ammonium nitrite. Lower temperatures do not provide
sufficient energy to initiate the reaction.1 The catalysts used
in SCR units are predominately oxides of titanium and vanadium.2
However, platinum, zeolites, and ceramics are also used.1 In-
gas-fired applications, catalyst pellets in a fixed bed are
commonly used. For oil- or coal-fired applications where the
flue gas contains particulate matter, catalyst designs usually
include honeycomb, pipe, or parallel shapes, which allow the flue
gas to pass along the catalyst"surface.2
NH,
Gts
+
•—NO,
•— "P.
NH.
O—NO,
X-NH,
O—NO,
X-NH,
0-HjO
Figure 5-4 Illustration of the SCR principle.12
The effectiveness of the SCR in removing NOX is also
dependent on the NH3:NOX mole ratio. As the NH3:NOX mole ratio
increases, the NOX reduction increases and ammonia slip (i.e.,
unreacted ammonia slipping through the catalyst bed) increases.
In practice, removal efficiencies of up to 90 percent are
commonly achieved with ammonia slip values below 5 ppm. On-line
NOX analyzers and feed-back control of NH3 injection are required
to maintain control of the SCR process.
Site-specific factors other than temperatures. and NH3:NOX
ratios that affect NOX emission rates achievable with SCR include
the ratio of flue gas flow rate to catalyst volume (the inverse
of residence time), catalyst poisoning (by metals, acid gases, or
particulate entrainment), and pressure drop across the catalyst
which reduces fuel efficiency. Most of these potential problems
have been addressed successfully in commercial operations.3
5-10
-------
There has been limited experience with SCR at iron and steel
mill. This is primarily due to the lack of stringent NOX
regulations. In the U.S., SCR has been applied to oil and gas
fired combustion units such as gas turbines, internal combustion .
engines, and process heaters and several coal-fired boilers. At
iron and steel mills, there are 2 SCR units in the U.S. An SCR
unit is being used to control NOX emissions from a gas-fired,
radiant tube, continuous annealing furnace at a steel mini-mill
in the United States. This furnace also has LNB's. Controlled
emissions from this unit are about 33 ppm at 3 percent O2.14 A
second SCR unit, currently under construction, will be used to
control NOX emissions from an annealing furnace at an integrated
steel plant in the United States. This furnace does not have
LNB's. The unit has a guaranteed NO, reduction of 90 percent.15
Techniques for NOX control adopted by the Japanese iron and
steel industry are, for the most part, based on combustion
modifications. However, the Japanese have investigated "flue gas
denitrification" using the SCR process for sintering plants, coke
ovens, and reheating furnaces.7 They concluded that the SCR
process was impractical for use on sintering plants due to the
many technical problems that remain to be solved.16 These
problems include the limited service life of the catalysts,
energy requirements to raise the exhaust gas temperature, and the
large installation space required. However, in some
environmentally critical areas, flue gas denitrification units
(presumably SCR) are used on sinter plants. SCR applications to
coke ovens and reheat furnaces appear to be experimental. They
are not routinely used on these facilities in Japan.8
5.2.2 Other Control Techniques
LEA, LNB's, LNB plus FGR, SCR, and SCR plus LNB are NOX
control techniques that have been used to reduce NOX emissions
from iron and steel process facilities in the United States.
Other techniques that may be applicable, some of which have been
used in Japan, are discussed in this section.
5.2.2.1 Off-Stoichiometric or Staged Combustion(OSC). OSC
reduces NOX generation by carrying out initial combustion in a
primary, fuel-rich combustion zone and subsequently completing
combustion at lower temperatures in a second, fuel-lean zone.
This technique is essentially the control technique implemented
through the use of LNB's; however, it can be implemented using
conventional burners by staging the combustion in the furnace or
firebox rather than within the burner itself as is done with
LNB'S.
In practice, OSC is implemented through biased burner firing
(BBF), burners out of service (BOOS), or overfire air (OFA). BBF
and BOOS are very similar and are generally applicable to
5-11
-------
furnaces with multiple burners arranged in rows. In BBE, fuel
distribution among the burners is redistributed so that the lower
rows of burners are fired more fuel-rich than the upper rows, as
illustrated in Figure 5-5. The additional air needed to complete
combustion is provided through the upper rows of burners, which ..
are fired fuel-lean. In BOOS firing, selected burners or rows of
burners are made inactive (taken out of service) by turning off
their fuel supply and using them to admit air only to the
furnace. The total fuel demand is supplied by the remaining or
active burners. Thus, the active burners are fired fuel-rich,
and the additional air required to complete combustion is
supplied through the BOOS or air-only burners. This principle is
illustrated in Figure 5-6. Both BBF and BOOS implement staged
combustion by creating two combustion zones within the firebox:
one fired fuel-rich or with low excess air and the second fired
with excess air at a reduced temperature. These methods are
applicable to all fuels, and both fuel- and thermal-NOx formation
are reduced.2
These techniques, BBF and BOOS, are attractive options for
existing facilities with suitably configured burners since few
equipment modifications are required. In some cases, process
facilities may need to be derated if the active burners (BOOS) or-
fuel-rich burners (BBF) have limited firing capacities.
Implementing BBF or BOOS requires careful, accurate
monitoring of the combustion process and the flue gas in order to
control the process. Alternatively, furnace efficiency and
safety may be compromised.2
Emissions tests of BOOS firing on utility boilers have
indicated average NOX reductions of 31 to 37 percent for coal,
oil, and natural gas firing compared to baseline values.2 It
seems reasonable that BBF would yield similar results due to the
similarity of the two techniques.
A third type of staged combustion, illustrated in Figure 5-
7, is OFA. With OFA, the burners are fired more fuel-rich than
normal, and the additional air needed to complete combustion is
admitted through overfire air ports or an idle top row of
burners. This technique is applicable with all fuels. In
emissions tests on utility boilers,. OFA has achieved average NO^
reductions of 24 to 59 percent with oil, coal, and natural gas
firing compared with baseline levels.2
These staged combustion techniques, especially OFA, are more
easily implemented on large, utility-scale boilers than on .
smaller units and furnaces.2 As furnace size decreases, furnace
volume decreases more quickly than furnace wall area. Thus,
residence times for fuel combustion may become a problem in
smaller units. Moreover, space for additional ductwork, furnace
5-12
-------
Fud-rich
Zoo*
Figure 5-5. Biased Burner Firing!!
5-13
-------
Figure 5-6. Burners out of service.
11
5-14
-------
Ovcrfire Air
Ports
Figure 5-7. Overfire air.
5-15
-------
penetrations, and fans may be a problem if these additions are
needed. These factors are potential impediments to implementing
these OSC techniques in iron and steel process facility furnaces,
especially on a retrofit basis.
There is no evidence that FGR has been used to control NOX
emissions from iron and steel process facilities. However, a
radiant tube LNB featuring a vitiated air stream concept (i.e.,
FGR) is being developed.7 As discussed in Section 5.2.1.3, FGR
has been used with LNB's to reduce NOX emissions from a reheat
furnace.
5.2.2.2 Fuel Switching. Fuel switching and fuel
denitrification are possible NOX control techniques. In Japan,
these techniques are currently used or have been used to reduce
NOX emissions from reheat furnaces, sintering furnaces, and coke
ovens (denitrification of fuel only).8
Natural gas is an attractive fuel from a NOX control
perspective because of the absence of fuel-bound nitrogen (fuel
NOX) and the flexibility it provides for implementing other
combustion modification control techniques. Because natural gas -
is already the principal fuel used in reheat furnaces in the
United States, fuel switching is not a reasonable NOX
control option for reheat furnaces (or soaking pits, for the same
reason) in the United States. Moreover, for large fuel-burning
processes, fuel choice decisions tend to be based on fuel costs
and on other regulatory constraints.
Coke ovens reduce coal to coke in a reasonably closed
reducing environment. There are fugitive emissions of NOX from
leaks and during charging and pushing operations. Coke oven
underfiring is done predominantly with coke oven gas, blast
furnace gas, natural gas ,or some combination of these three. It
is not clear that fuel switching is a reasonable option for NOX
reduction in the case of coke ovens in the United States, and it
is not clear from the literature to what extent the Japanese use
fuel switching as a coke oven NOX control strategy. Two
strategies are cited. These are (l) switching to lighter fuel
oil (not an option if COG, BFG, and NG are baseline fuels), and
(2) taking credit for the nitrogen removal that is incidental to
the desulfurization of COGV(an oxides of sulfur control
measure).&
5.2.2.3 Selective Noncatalytic Reduction. SNCR is
generally an applicable NOX.control technique for applications
that are suitable for SCR. SNCR is very similar to SCR: i.e.,
NH3 or urea is injected into the flue gas stream where the
temperature is suitable for the reaction with NOX to proceed.
However, SCR and SNCR differ in that their ideal reaction
5-16
-------
temperatures differ. Depending on catalyst type, the optimum
temperature for SCR reactions is 260 to 450 °C [500 to 840 °F],
and the optimum temperature for SNCR is 870 to 1,250 °C (1,600 to
2 300 °F) . Generally, 40 to 70 percent NOX reduction is achieved
with NH3:NOX molar ratios of 1:1 to 2:1. This control effiency ,
generally results from the difficulty of achieving uniform mixing
of NH3 with the flue gas and from the variations of flue gas
temperature and composition usually present.
There are two commercially available SNCR processes that use
either ammonia or urea. The SNCR process that uses ammonia is
the Thermal DeNOx (TDN) system developed by EXXON. In this
process, gaseous ammonia is injected into the air-rich flue gas
stream where it reacts with NOX in accordance with two competing
reactions:3
2NO
4NH
4NH3 + 502
20
4NO
3N2 + 6H20, and
(5-3)
(5-4)
Thus, the TDN process has the potential to either reduce NOX in
accordance with reaction (5-3) or increase NOX in accordance with
reaction (5-4). Temperature is the primary variable for
controlling the rates of these reactions. In the temperature
range of 870 to 1,200 °C (1,600 to 2,200 °F), reaction (5-3)
dominates, resulting in a reduction of NOX. Above 1,200 °C
(2,200 °F), reaction (5-4) dominates, resulting in an increase of
NO. Below 870 °C (1,600 °F), neither reaction is of sufficient
activity to either increase or decrease NOX. The reduction of NOX
is maximized in the relatively narrow range of 870 to
1 040 °C (1,600 to 1,900 °F) with an optimum temperature of
about 950 °C (1,750 °F) . The favorable, 170 °C (300 °F) reaction
window for reaction (5-3) can be lowered to about 700 to 820 °C
(1,300 to 1,500 °F) by introducing hydrogen, a readily oxidizable
gas.3-17
Without the use of a catalyst to increase the reaction
rates, adequate time at optimum temperatures must be available
for the NOX reduction reaction to occur. Design considerations
should allow ample residence time and good mixing in the required
temperature range. Long residence times (>1 second) at optimum
temperatures tend to promote relatively high performance even
with less than optimum initial mixing or temperature/velocity
gradients: However, when the NH, injection zone is characterized
by low temperature arid/or steep temperature declines, a loss of
process efficiency results.
The initial ratio of ammonia injected to NOX concentrations
is another parameter to control in the process. Maximum NOX
reductions (50 to 70 percent) require 1.5 to 2.0 NH3:NOX injection
ratios. At these ratios, significant concentrations of NH3 can
5-17
-------
exit the convective zone, creating corrosive ammonium sulfates
and/or a visible NH3 stack plume. Unreacted ammonia emissions
from a TDN system are usually higher than from SCR systems.3
The second of the two commercially available SNCR processes
is the Nalco/Fuel Tech NOXOUT process. In this process, a urea-
type compound is injected at a favorable location to reduce NOX
according to the following reaction: •
NH2 + NO -» N2 + HjO
(5-5)
The exact chemical mechanism is not fully understood, but it
involves the decomposition of urea [CCNH^O]. The likely
decomposition products include the NHj groups. The reaction
takes place at temperatures of 870 to 1,200- °C (1,600 to
2,300°F) .3
Originally developed by the Electric Power Research
Institute (EPRI) in the early 1980's, the process is currently
licensed by Nalco/Fuel Tech where patented process modifications
and control techniques, combustion unit computer modelling and
proprietary additives that allow NOX reduction capability over a
temperature range of 760 to 1,200 °C (1,400 - 2,300 °F) have been,
developed. As with the other post-combustion NOX control
systems, temperature is the primary control variable for the
selective reactions. NOX reductions up to 80 percent are
achievable with this technology. The performance of the urea
injection process is limited by the time/temperature/flow
characteristics of the flue gas. Residence time in the
temperature window and the urea-to-NOx ratio impact the
performance in a manner similar to that for Thermal DeNOx. The
NO, reduction capability is limited by mixing because the
reaction times are relatively quick.3
The SNCR technique is discussed here because of its
successful application to facilities that are similar to iron and
steel facilities, e.g., hot-side, coal-fired utility boilers,
heavy oil fired industrial boilers, and wood-waste fired co-gen
plants. In general, SNCR is applicable in applications where a
temperature suitable for a favorable reaction is acheivable.
SNCR has not been used for any iron and steel mill processes.
5.3 NITROGEN OXIDES CONTROLS FOR SPECIFIC FACILITIES
In this section, the NOX emission sources at iron and steel
mills are discussed with a view toward identifying and evaluating
applicable control techniques.
5.3.1 Coke Ovens
Coke ovens produce metallurgical coke from coal by the
5-18
-------
distillation of volatile matter. A by-product of the process is
coke oven gas, a fuel commonly used (along with blast furnace
gas) to fire the ovens. Although NOX emissions tend to be
minimized by slow mixing in the combustion chamber, they are
nonetheless substantial because of the large quantity of fuel
consumed. Coke ovens are among the major NOX emission sources at
iron and steel mills.2 Coke ovens with NOX controls in the United
States have not been found.
The Japan Iron and Steel Federation has reported the
installation of flue gas denitrification (SCR) units on coke
ovens. They also cite low-air-ratio combustion and
denitrification of fuel as NOX control techniques applicable to
coke ovens. Denitrification of fuel may refer to the practice of
desulfurizing coke gas, which also has the effect of removing
nitrogen and, thus, reduces NOX emissions. Flue gas
recirculation is also cited. The Japanese acknowledge that there
are many limitations and difficulties associated with applying
these controls and some may be applicable only to new ovens. No
details are provided about the effectiveness or levels of
performance of these controls.
5.3.2 Sinter Plants
Sintering is also a major source of NOX emissions from iron
and steel mills. In general, only integrated steel plants
operate sinter plants where the sinter is charged to a blast
furnace.
Sinter is produced from iron ore fines, limestone fines,
coke fines or breeze, flue dust, and possibly other additives._
These materials are mixed, uniformly distributed onto a traveling
grate, ignited, and heated to a fusion temperature that causes
the materials to agglomerate into sinter. Subsequently, the
sinter is cooled, crushed, and screened for size prior to
charging to a blast furnace.
The major source of energy used in the production of sinter
is the carbon content of the coke breeze and flue dust. An
auxiliary fuel such as natural gas or coke oven gas is used to
ignite the bed. This ignition fuel typically represents about 8
to 10 percent of the thermal energy required by the process.
Following ignition, combustion is continued by forcing air
through the bed as it travels with the grate. The combustion of
the coke breeze and flue dust (about 5 percent of the bed
material) creates sufficient heat and temperature [about 1,300 to
1,480 °C (2,370 to 2,700 °F)] to sinter the fine ore particles
into porous clinkers that, following cooling and screening, are a
suitable burden for a blast furnace.17 No information was
obtained on NOX emission control techniques for sinter plants in
the United States.
Japanese sources cite sinter plants as a major source of NOX
5-19
-------
emissions at steel plants and report on research on SCR
applications, an electron beam radiation process, and technology
for reducing nitrogen in the coke breeze as they apply to sinter
plants. The results of the research on reducing N2 in the coke
breeze are not readily available; however, it is noted that many .
problems remained to be solved.16
More recently, the Japanese have reported that some steel
mills have controlled NOX generation at sinter plants by
pretreating the sinter materials, i.e., lowering the oxidation
level of the sintering atmosphere through adjustment of the coke
blending ratio, basicity, etc.7 Most of the NOX in sinter plant
flue gas appears to come from coke in the sinter bed, which
usually contains about 1 percent nitrogen. Japanese research in
this area progressed from batch tests to a continuous test plant
with a capacity for treating 20 tons of coke per day. The
process was to preheat pulverized coke in a two-stage, fluidized-
bed heater and then use an electric furnace to heat the coke to
1,700 °C (3,090 °F) using an electric current through the coke
itself. The retention time is about 2 hours. The coke is
subsequently cooled.19
About 70 percent of the N is removed from the coke by
heating to 1,700 °C (3,090 °F). It is claimed that by removing
70 percent of the N from coke, the NOX in the flue gas was also
reduced by about 70 percent. This indicates that, in the sinter
plant, most of the NOX is generated from N in the coke.19
It was estimated that the coke denitrification process
requires about 30 kWh of electricity per ton of sintered iron
ore. Concern was also expressed about the flue gas from the
treatment process, i.e., its composition and treatment. The
Japanese were also uncertain as to whether the process could be
commercialized.
The Japanese also report that some sintering plants, located
in areas where national emission standards are not met
sufficiently to ensure environmental quality, are equipped with
flue gas denitrification units. These are SCR units, and there
are still technical problems with this application, e.g., the
catalyst life is limited and large amounts of additional heating
energy are required.8 ;
In one sinter plant application, the commercial SCR plant
was constructed as a permitting compromise when a large, new
blast furnace was constructed. The capacity of the SCR unit was
750,000 Nm3/hr. The sinter flue gas contained considerable
amounts of particulates and S03 and, thus, was difficult to
treat. The flue gas treatment consisted of a flue gas
desulfurization (FGD) process followed by a wet electrostatic
precipitator (ESP). The cleaned gas was then heated to 400 °C
(750 °F) for SCR treatment. Sinter flue gas typically has a
temperature of about 150 °C (300 °F), 150 to 300 ppm of NOX
(mostly NO), 150 to 400 ppm of SOX, and 200 to 300 mg/Nm3 of
5-20
-------
particulates containing ferric oxide ;and potassium chloride
vapor. It is not an easy gas to treat with SCR because the
effluent gas has to be heated and because particulates and
potassium chloride vapor in the effluent tend to contaminate the
catalyst.19
The following information was obtained about this major SCR
unit:19
Completed
Treated facility
Capacity
Load factor
Pretreatment of gas
SCR inlet
Reaction temperature
NOX:NH3 ratio
Catalyst life
Nox removal
1976
Iron ore sintering plant
762,000 Nm3/hr
70 to 100 percent
Electrostatic precipitator
Flue gas desulfurization
Wet ESP
Heating
200 to 300 ppm NOX
5 to 20 ppm SOX
3 to 10 mg/Nm3 particulates
11.2 percent 02
380-390 °C (716-734 °F)
1:1
1 year
95 percent
In 1979, the Japanese assessment of this control application was
that the investment and operating costs were so high that the
process could not be widely used.19
5.3.3 Blast Furnaces and Blast Furnace Stoves
The blast furnace is a tall, shaft-type furnace where
iron-bearing materials (iron ore, sinter, slag, scrap, etc.) are
reduced to iron (pig iron or hot metal). The burden or charge to
the blast furnace also includes flux (limestone or dolomite) and
coke. A preheated air blast is charged to the blast furnace
through tuyeres, nozzle-like openings near the bottom of the
furnace. The coke in the burden produces the heat required for
smelting the charge, and it provides carbon and carbon monoxide
required for reducing the iron ore. Moreover, because the coke
retains its strength at temperatures above the melting
temperature of pig iron and slag, it provides the structural
support that keeps the unmelted burden materials from falling
into the hot, molten metal that collects in the hearth. A
description of the blast furnace and its operation can be found
in Chapter 3 of this document, and a very detailed discussion can
be found in Reference 18.
The blast furnace itself is a closed unit with no
atmospheric emissions. The hot air blast reacts with the coke in
the furnace to produce more carbon monoxide than is needed to
reduce the iron.ore. The excess,CO leaves the top of the blast
5-21
-------
furnace with other gaseous products and particulates. This blast
furnace effluent, called BFG, is captured, cleaned, cooled, and
subsequently used as a fuel in other processes. BFG contains
about 1 percent hydrogen and 27 percent CO and has a heating
value of about 2,540 - 3,600 kJ/Nm3 (65 - 92 Btu/ft3) .5
A blast furnace typically has three or four blast furnace
stoves as auxiliaries. These stoves are used to heat the blast
air required by the blast furnace. They function somewhat like a'
regenerator in that one or more of the stoves are on blast (i.e.,
they are being used to heat the blast air being used by the blast
furnace} while the others are being heated. They differ from a
regenerator in that the source of heat is the combustion of a
fuel rather than the extraction of heat from an effluent stream.
Between 2.2 and 3.5 kg of BFG are generated for each
kilogram of pig iron produced in a blast furnace. About 18 to 24
percent of this gas is used to heat the checker brick in the
associated stoves.2 The remainder is used elsewhere in the
plant. Because of its low heating value, the BFG used in the
stoves is often enriched by the addition of fuels with a much
higher calorific value such as natural gas or coke-oven gas to
obtain higher blast air temperature, often in excess of 1,100 °C -
(2,010
Blast furnace stoves are a NOX emissions concern because
they consume large quantities of fuel. However, the primary fuel
is BFG, which is largely CO, has a low heating value, and
contains inerts, factors that reduce flame temperature. Thus,
the NOX concentration in blast furnace stove flue gas tends to be
low and the potential for NOX
small.2
reduction is considered to be
Apparently there are no blast furnace stoves with NOX
emissions controls in the United States. The Japanese also
indicate that NOX emissions from blast furnace stoves are not a
concern for the reasons noted above. They also suggest that LEA
combustion can be implemented, but they do not explicitly claim
to have implemented LEA.8
5.3.4 Basic Oxygen Furnace
The EOF is the principal steelmaking process used in the
United States. In operation, the EOF is tilted to receive a
charge of scrap metal and, subsequently, molten pig iron from a
blast furnace. After charging, it is returned to a vertical
position and blown with high-purity oxygen, which converts the
charge to steel. It is finally tapped and the molten steel is
poured into a transfer ladle. Following tapping, the EOF is
again tilted to discharge the remaining slag. A more .detailed
description of the EOF facility and the oxygen steel -making
5-22
-------
process can be found in Chapter 3 of this document and in
Reference 18.
In the EOF, oxygen is blown onto the bath at supersonic
velocity (to penetrate slag and metal emulsions) where it
oxidizes impurities in the bath to create steel. The oxidation
of carbon produces about 467 kJ of CO per kilogram of steel
(400,000 Btu/ton) .2 Typically, the off-gas, which is mainly CO,
is burned in an open, water-cooled hood mounted above the EOF.
These hoods admit air for combustion of the CO on an uncontrolled
basis. Some newer BOF's use suppressed combustion hoods that do
not admit air or burn the off-gas. In these instances, the
effluent gas is flared.
The combustion of the waste gas results in the generation of
NOX. There is no information available to indicate that existing
BOF's have any kind of NOX emission controls or that any control
techniques are suitable.2
A variation on the EOF is the Q-BOP furnace. It differs
from the BOF principally in that oxygen is blown through the
molten metal bath through tuyeres in the bottom of the vessel
rather than onto the top surface of the bath. From a NOX
emissions and controls perspective, the two processes appear to
be essentially the same, i.e., there is no indication that Q-BOP
furnaces have NOX emissions controls or that any control
techniques are suitable. In each case, NOX is generated during
the oxygen blow when CO in the off-gas is combusted either at the
mouth of an open, combustion-type hood or at the flare of a
suppressed hood system.20
5.3.5 Electric Arc Furnaces
The use of electricity to melt steel scrap in the EAF
transfers NOX generation from the steel mill to a utility power
plant. There is no information that NOX emissions controls have
been installed on EAF's or that suitable controls are available.
5.3.6 Reheat Furnaces
Reheat furnaces bring steel to a uniform temperature of
about 1,180 to 1,340 °C (2,150 to 2,450 °F), a temperature
suitable for hot working.18 They are major fuel - consuming
facilities and major sources of NOX emissions.
LEA, LNB's, and LNB plus FGR are being used to control NOX
emissions from reheat furnaces in the United States.5-10'11'21 Major
modifications to furnace structure and refractories to install
alternate burners may be required in some existing reheat
furnaces. In an LEA application, the control was retrofitted to
5-23
-------
a 98-ton/hr (112-MMBtu/hr) capacity reheat furnace. LEA was
selected as the most cost-effective option because the burner
system did not need replacement. Alternatively, another control
option would likely have been chosen. A test report shows the
LEA control efficiency to be 14 percent.5 Assuming uncontrolled ,
emissions of 220 ppm at 3 percent 02 (0.200 Ib NOx/MMBtu) (from
Table 4-4) for a recuperative furnace and a control efficiency of
13 percent, controlled NOX emissions are estimated to be 190 ppm
at 3 percent 02 (0.174 Ib NOx/MMBtu) . The corresponding NOX
reduction is 0.026 Ib NOx/MMBtu. For regenerative furnaces,
controlled emissions are 560 ppm at 3% 02 and 0.689 Ib/MMBtu.
For cold-air fired reheat furnaces, LEA lowers emissions to 96
ppm at 3%02 and 0.117 Ib/MMBtu.
LNB or FOR may require major modifications to furnace
structures and refractories to install alternate burners. These
combustion modifications do not necessarily -reduce efficiency and
capacity. The addition of LNB to a recuperative-fired furnace is
not likely to affect either capacity or fuel requirements. The
addition of FGR to a recuperative-fired furnace with or without
LNBs will likely involve derating the furnace. The addition of
LNB plus FGR to a regenerative fired furnace will not likely
reduce capacity.22
Average controlled NOX emissions from two reheat
furnaces(regenerative-firing) with LNB plus FGR controls are 150
ppm at 3 percent O2 (0.18 Ib NOx/MMBtu) .to>" Thus, the control
efficiency is 77 percent and the NOX removed is 0.61 Ib NOx/MMBtu.
This efficiency is also used for recuperative and cold air
firing.
Two reheat furnaces with LNB NOX control, one using
regenerative preheated combustion air and one using cold
combustion air, have controlled NOX emissions of 220 ppm at 3
percent O2 (0.27 Ib NOx/MMBtu) and 80 ppm at 3 percent 02 (0.10 Ib
NOx/MMBtu) , respectively.11'24 There are no uncontrolled NOX
emissions data available for either of these furnaces.
Uncontrolled NOX emissions from reheat furnaces using preheated
regenerative combustion air are estimated to be 645 ppm at 3
percent 02 (0.792 Ib NOx/MMBtu) as shown in Table 4-4. Thus, for
the reheat furnace using regenerative preheated combustion air,
the control efficiency is estimated to be 66 percent, and the NOX
removed is 0.52 Ib NOx/MMBtu.
From Table 4-4, the uncontrolled NOX emissions for cold-air
fired reheat furnaces are 110 ppm @ 3% 02 (0.135 Ib/MMBtu). The
control efficiency is therefore 27% or 0.035 Ib/MMBtu. For
recuperative-fired reheat furnaces, the controlled efficiency for
LNB is estimated to be 45% or controlled emissions are 120 ppm@
3% 02 (0.11 Ib NOx/MMBtu) . The NOX removal is 0.09 Ib/MMBtu.
In Japan, fuel switching, denitrification of fuel, LEA, and
LNB's have been used to reduce NOX emissions from reheat
5-24
-------
furnaces.8'19 No details about the performances of these controls
are available.
5.3.7 Soaking Pits
Soaking pits have many similarities to reheat furnaces.
Both are used to raise or maintain the temperature of steel in
preparation for hot working or shaping. They burn the same fuels
(natural gas predominantly, but also COB, BFG, and oil) and
operate at approximately the same temperatures. Soaking pits
tend to be smaller than reheat furnaces and, in the aggregate,
use much less fuel. In 1979, for example, soaking pits used only
about one-fifth the fuel used in reheat furnaces. Subsequently,
the use of soaking pits has continued to decline as the practice
of continuous casting has increased.
No examples of soaking pits with NOX emissions controls were
found. However, the similarity of reheat furnaces and soaking
pits indicate that controls suitable for reheat furnaces would
also be suitable for soaking pits.
5.3.8 Annealing and Galvanizing
Annealing and galvanizing are-two finishing processes often
practiced in iron and steel mills. They are separate,
independent processes; however, both can be accomplished in a
single continuous-type facility that includes, for example, a
continuous annealing furnace followed immediately by a
galvanizing facility in which the continuously moving steel sheet
is immersed in a molten zinc bath. As compared with most iron
and steel process, which take place at temperatures of 1,090 to
1,650 °C (2,000 to 3,000 °F) , annealing and galvanizing are
accomplished at moderate temperatures usually below 540 °C
(l,000°F); however, some annealing processes may require higher
temperatures. Because of'these much lower temperatures, NOX
emissions from these processes should be lower- These two
finishing processes are discussed independently in the following
subsections.
5.3.8.1 Annealing Furnaces. Annealing may be accomplished
in a large, batch-type furnace processing tons of coiled steel in
a single, multiday cycle or in a large continuous-type furnace
processing a continuous sheet of steel passing through the
furnace. "Annealing is a highly specialized technology that can
alter the properties of steel in useful ways. In general,
annealing relieves cooling stresses induced by hot-or-cold-
working and softens the steel to improve its machinability or
formability. This is accomplished by subjecting the steel to a
controlled thermal profile or cycle with moderate peak
temperatures.
SCR plus LNB, LNB plus FGR, and SCR are being used in the
United States to control NO, emissions from annealing furnaces.
5-25
-------
Information is available on four annealing furnaces, one that is
operational and three that are under const rue t ion.u>14>15 All are
radiant tube, continuous annealing furnaces firing natural gas.
Due to the many exhaust collection points at annealing furnaces,
there may be problems in installing SCR at existing furnaces.
The control efficiency of the SCR unit alone on the LNB-
plus-SCR controlled furnace is consistently 90 percent or more,
and controlled emissions at or near capacity firing rates average
33 ppm at 3 percent 02 (0.04 Ib NOx/MMBtu) .M There are no
uncontrolled emissions data available for this furnace, only NOX
concentration data referenced to the inlet of the SCR.
Uncontrolled NOX emissions from two annealing furnaces are
reported to be 1,000 ppm at 3 percent O2 (1.23 Ib NOx/MMBtu) (from
Table 4-4). Thus, the control efficiency of the LNB plus SCR
control is estimated to be 97 percent. Individual control
efficiencies of 53 percent for the LNB's and 93 percent for the
SCR, for example, would yield this result. The NOX removed by
this control combination is 1.19 Ib NOx/MMBtu.
Laboratory data are available for one of the two LNB-plus-
FGR-controlled annealing furnaces under construction. Controlled-
NOX emissions from this furnace are reported to average 180 ppm
at 3 percent 02(0.22 Ib NOx/MMBtu)." Uncontrolled emissions,
from Chapter 4, are 1,000. ppm at 3 percent 02 (1.23 Ib NOx/MMBtu) .
Thus the control efficiency is estimated to be 82 percent, and
the NOX removed is estimated to be 1.01 Ib NOx/MMBtu.
Controlled NOX emissions for the SCR-controlled annealing
furnace under construction are estimated using 1,000 ppm at 3
percent O2 as uncontrolled emissions (from Chapter 4) and
assuming a control efficiency of 85 percent. This yields
estimates of controlled emissions and NOX removed of 150 ppm at 3
percent 02 (0.18 Ib NOx/MMBtu) and 1.1 Ib NOx/MMBtu, respectively.
Although no data or examples are available, controlled NOX
emissions and NOX removed for annealing furnaces with only LNB's
for control are estimated by assuming a control efficiency of
50 percent. The resulting values are controlled NOX emissions of
500 ppm at 3 percent 02 (0.61 Ib NOx/MMBtu) and 0.61 Ib NOx/MMBtu
removed.
No evidence of the use of SNCR for NOX control on annealing
furnaces (or other iron and steel facilities) has been found.
However, because of SNCR's similarity to SCR, NOX emissions from
annealing furnaces with SNCR controls are estimated. As
discussed in Section 5.2.2.3, favorable temperature windows for
the desired reactions differ between SCR and SNCR, and the point
of injection for the ammonia or urea reactant will differ. These
are design issues that will have to be considered if SNCR is
adapted to annealing furnaces. Assuming a control efficiency of
60 percent for the SNCR control unit yields controlled emissions
and NOX removed estimates of 400 ppm at 3 percent 02 (0.49 Ib •
5-26
-------
NOx/MMBtu) and 0.74 Ib NOx/MMBtu, respectively.2
Similarly, controlled emissions and NOX removed estimates
are made for LNB-plus-SNCR-controlled annealing furnaces by
assuming a control efficiency of 80 percent (50 percent for LNB -.
control and 60.percent for SNCR control). The resulting
estimates are controlled NOX emissions of 200 ppm at 3 percent 02
(0.25 Ib NOx/MMBtu) and 0.98 Ib NOx/MMBtu removed.
5.3.8.2 Galvanizing Furnaces. Steel products are often
coated with a protective layer of zinc, aluminum, terne (a lead-
tin alloy) or a zinc-aluminum alloy to provide protection against
corrosion. This process is called galvanizing. At iron and
steel mills, steel sheets, strips, and other products with a
light cross section are typically galvanized. A typical molten-
zinc bath temperature, for example, is 450 °C (840 °F).l
Controlled emissions data are available for two, preheated
combustion air galvanizing furnaces with LNB plus FOR NOX
controls and one cold combustion air galvanizing furnace with LNB
NOX controls.11'21 Controlled NOX emissions from the preheated
combustion air furnaces average 220 ppm at 3 percent 02 (0.27 Ib -
NO /MMBtu). Given uncontrolled emissions (from Chapter 4) of 940-
ppm at 3 percent 02 (1.15 Ib NOx/MMBtu) , the LNB plus FGR control
efficiency is 77 percent and the NOX removed is 0.88 Ib NOx/MMBtu.
Controlled NOX emissions data for the cold combustion air,
relatively low-temperature, LNB-controlled galvanizing furnace
are given as 21 ppm at 3 percent O2 (0.026 Ib NOx/MMBtu) .21 In
Chapter 4, uncontrolled NOX emissions for this furnace were
estimated from the controlled emissions assuming a 34 percent
_ _ _ i a _«* ^f**. * _J '._ ._. ._ ^ J «•« ^ >^ *9 4— ^* V1^/-^ f*\ ^ "1 ^
control efficiency.
NO./MMBtu.
The NO, removed is estimated to be 0.014 Ib
5.4 SUMMARY OF CONTROLLED NOX EMISSIONS
Table 5-2 summarizes controlled NOX emissions data,and
estimates for controlled facilities taken from preceding sections
of this chapter. The values tabulated in Table 5-2 are used in
Chapter 6 to estimate cost effectiveness of the control
techniques. No controls are reported for coke ovens, sinter
plants, blast furnaces or blast furnace stoves, basic oxygen
furnaces, electric arc furnaces, or soaking pits in the United
States. the Japan Iron and Steel Federation reports experimental
work with coke oven controls (LEA, fuel denitrification, and flue
gas denitrification), but no definitive data are provided.8 They
also report experimental work with sinter plant controls
(pretreating sinter materials and SCR).19 Again very little
information is provided. This work appears to be experimental/
very selectively applied, and not appropriate for application in
the United States.
5-27
-------
CO
O
H
I
p<
I
H
CO
CO
H
I
8
N
i
in
H
I
o
H
W
P4
Q
W
CO
OJ
8
m vo
H VO
o o o
in us ko
in o o
oo oo a\
H
S
c>-
o
r-
o
O
o
o
o
n
o
<;
W
1
w
oo
o
•
o
o
•
o
VO
O
5-28
-------
5.5 REFERENCES
1. Mclnnes, R., and M. B. Van Wormer. Cleaning Up NOX
Emissions. Chemical Engineering. September 1990.
2.
3.
4.
5.
6.
7.
8.
9.
10
U.S. Environmental Protection Agency. Control Techniques
for Nitrogen Oxides Emissions from Stationary Sources-
Revised Second Edition. EPA-450/3-83-002. January 1983.
Campbell, Lisa M., D.K. Stone, and G.S. Shareef (Radian
Corporation). Sourcebook: NOX Control Technology Data.
Prepared for U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/2-91-029.
July 1991.
Plant Engineering. March 1991.
Letter and attachments from Finke, H.P., Bloom Engineering
Company, Inc., to Neuffer, W.J., EPA/ESD. May 14, 1993.
Comments on draft ACT document..
Letter and attachments from Lee, D., Cascade Steel Rolling, .
Mills, Inc., to Jordan, B.C., EPA/ESD. July 6, 1992.
Response to Section 114 letter on iron and steel mills.
Finke, H.P. Low-N0x Recuperative Burner Technology. AISI
Conference Paper. May 1988.
The Japan Iron and Steel Federation. NOX Regulations and
Control Technology in the Japanese Iron and Steel Industry.
AISI Conference Paper, May. 1988.
Finke, H.P. Low NOX Recuperative Burner Technology. AISI
Conference Paper., May 1988.
Letter and attachments from Postlethwait, W., Nucor Steel,
Darlington Division, to Jordan, B.C., EPA/OAQPS.
December 2, 1992. Response to Section 114 letter on iron
and steel mills.
11. Letter and attachments from Gilbert, F.C., North American
Manufacturing Company, to Jordan, B.C., EPA/OAQPS.
November 11, 1992. Response to. Section 114 letter on low-
NOX burners.
12. Joseph, G.T., and D.S. Beachler (Northrop Services, Inc.).
APTI Course 415, Control of Gaseous Emissions, Student
Manual. Prepared for U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-450/2-81-
085. December 1981.
13. SCR Committee, NOX Control Division, Institute of Clean Air
Companies, Inc. White Paper - Selective Catalytic Reduction
5-29
-------
(SCR) Controls to Abate NOX Emissions. September 1992.
14. Letter and attachments from McMahon, K.R., USS-POSCO
Industries, to Neuffer, W., EPA/ISB. August 17, 1992.
Response to Section 114 letter on iron and steel mills.
15. Letter and attachments from Anderson, D.M., Bethlehem Steel
Corporation, to Jordan, B.C., EPA/ESD. July 13, 1992.
Response to Section 114 letter on iron and steel mills.
16. lijima, K. Environmental Control Measures in the Japanese
Steel Industry. Industry and Environment, April/May/June.
1980.
17. Ref. 1, p. 3-42.
18. Lankford, W.T., N.L. Samways, R.F. Craven, and H.E.
McGannon. The Making, Shaping, and Treating of Steel, Tenth
Edition. 1985.
19. U.S. Environmental Protection Agency. NOX Abatement for
Stationary Sources In Japan. Research Triangle Park, NC. . _
Publication No. EPA-600/7-79-205. August 1979. .
20. Barnard, W.R. (E.H. Pechan and Associates, Inc.). Emission
Factors for Iron and Steel Sources-Criteria and Toxic
Pollutants. Prepared for U.S. Environmental Protection
Agency. Washington, DC. Publication No. EPA-600/2-90-024,
June 1990.
21. Letter and attachments from Harmon, M.L., LTV Steel Company,
to Jordan, B.C., EPA/ESD. July 2, 1992. Response to
Section 114 letter on iron and steel mills.
22. Letter and attachment from Steiner, B.A., American Iron and
Steel Institute, to Neuffer, W.J., EPA/ESD. May 25, 1993.
Comments on draft ACT document.
23. Kobayashi, H. Oxygen Enriched Combustion System Performance
Study, Volume II - Market Assessment, Phase I - Final
Report. DOE/ID/12597-3, DE89 004061. September 1988.
24. Letter and attachments from Sulc, D.A., Nucor Steel,
Crawfordsville Division, to Jordan, B.C., EPA/OAQPS.
October 6, 1992. Response to Section 114 letter on iron and
steel mills.
5-30
-------
CHAPTER 6
COSTS OF NOX CONTROLS
in Chapter 5, reheat furnaces, annealing furnaces, and
galvanizing furnaces were identified as iron and steel mill
processes to which NOX emissions control techniques have been
applied. In this chapter, costs and cost-effectiveness^estimates
are developed for controls applied to these process facilities.
The applicable control techniques for these furnaces are as
follows:
Facility
Reheat furnace
Annealing furnace
Galvanizing furnace
Control technique
LEA
LNB
LNB plus FGR
LNB
SCR
SNCR
LNB plus FGR
LNB plus SCR
LNB plus SNCR
LNB
LNB plus FGR
Controls applicable to reheat furnaces are also considered
to be applicable to soaking pits because of the similarity of
soaking pits and reheat furnaces. They have similar purposes
(i.e., they keep or reheat steel to temperatures suitable for
further processing), burn the same fuels (usually natural gas),
and operate at about the same temperature (e.g., 1,320 °C [2,400
°F]) .
Costs and cost-effectiveness estimates are developed for
controls for models of each of the three process facilities
described above. For each facility, the models correspond,
approximately, to small, middle, and large sizes selected from
information obtained from iron and steel industry personnel and
control apparatus vendors.
All costs are in April 1994 dollars. Costs are based on the
Chemical Engineering Plant Cost Index.1
6-1
-------
No differentiation is made in this chapter between new and
retrofit installations of SCR and SNCR NOX control techniques.
These control technologies are add-on units, and cost differences
would be site-specific. In the case of LNB and LNB plus FGR
controls, new burners and associated peripherals (ducts,
plumbing, and controls) may not be required in existing burners.
For new units the costs associated with combustion modifications
are based on the incremental increase in costs for these burners
over the cost of an ordinary burner. In the case of LEA, the
total capital investment (TCI) costs are for monitors and
combustion controls that would be required for either new or
retrofit installations.
In lieu of definitive cost data, the following generally
accepted relationship for scaling costs between facilities of
different size or capacity is used as needed in this chapter:2
Costs for size 1
Costs for size 2
[ Capacity of size 1] °'6
[ Capacity of size 2 ]
(6-1)
Capital costs are annualized as needed in this chapter using
the following capital recovery factor:3 -
CRF •
(6-2)
where
CRF
i
n
capital recovery factor,
annual interest rate, and
system economic life in years,
An interest rate of 7 percent is used in this chapter. The
system economic life used, n, was taken from equipment life
predictions provided by Section 114 respondents. Cost
effectiveness is determined by dividing the total annual cost of
the NOX controls by the tons of NOX removed per year by the
control.
6.1 STEEL REHEAT FURNACES
6.1.1 Model Furnaces .
Steel reheat furnace capacities of 75, 150, and 250 tons of
steel processed per hour are reasonably representative of small,
mid-size, and large furnaces, respectively. Therefore, these ..
capacities are selected as model facilities. Further, a linear
regression of available data from 10 examples that define both
production capacities (in tons per hour) and fuel firing
capacities (in million Btu's/hr) indicates that the corresponding
fuel-firing capacities for these three sizes are 140, 300, and
6-2
-------
520 MMBtu/hr, respectively.4"9 The regression equation is
MMBtu/hr = 2.17 (ton/hr) - 21.1 (6-3)
The 10 reheat furnaces used to generate Equation (6-3) use
preheated combustion air, i.e., they are either recuperative or
regenerative furnaces. However to cover the entire population,
reheat furnaces are separated into regenerative, recuperative and
cold-air furnaces. The uncontrolled emission factors are shown
in Table 4-2 and are listed below. These emission rates_are
assumed to be relatively constant over a wide range of firing
rates and furnace capacities.
Firing Type
Regenerative
Recuperative
Cold-air
Uncontrolled Emissions
ppm @ 3% O? Ib/MMBtu
645 0.792
220 0.200
110 0.135
6.1.2 Costs and Cost Effectiveness of LEA Control for Reheat
Furnaces , .
Costs and cost-effectiveness estimates for LEA controls for
reheat furnaces are based on a single example for which costs and
NOX emissions estimates are available.4 NO, emissions for this
furnace are estimated to be reduced by 13% by low excess air.
The TCI costs and annual operating costs for this 112-MMBtu/hr
furnace are $101,000 and $12,500/yr, respectively.4 In Table 6-
1, these costs are scaled to the model reheat furnaces using
Equation (6-1). Capital costs range from $120,000 to 260,000.
The TCI costs are annualized using Equation (6-2), assuming an
economic life of 10 years and 7 percent interest. For each
furnace, the annual costs are the sum of the annualized TCI and
operating costs. Annual costs range from $32,000 to 70,000. The
NOX reductions are computed as the difference between
uncontrolled and controlled emissions assuming 8,000 operating
hours annually at capacity rate. The cost effectiveness ranges
from $330/ton for the 520-MMBtu/hr regenerative reheat furnace to
$3,200/ton for the 140-MMBtu/hr cold-air reheat furnace. For the
same size furnace, cost effectiveness for recuperative and cold-
air fired furnaces are approximately four and six times that of
regenerative furnaces.
6.1.3 Costs and Cost Effectiveness of LNB Control for
Furnaces
Reheat
The costs and cost effectiveness of LNB controls for
regenerative and recuperative-fired reheat furnaces presented
here are based on NCL emission reduction of 66%. LNB costs were
6-3
-------
Cost
Effectiveness
($/ton)
§
11
°1 J.
O
"^ w g"
60 ^
Ijf
5 U %
Firing Type
ctf t) _p
M!
s a 2
C*i C*} fl
3 2 S
«N CN CS
g g
1 1 *.
1 §• 3
I 1 a
i
M!
§ ' s a
!/•> \r> 10
a s s
III
g g
11 *.
S 8. '?
a g 2
O 4J O
8
CO
o 8 8
co co 2K
.CO * »
ro Q O
cs * ^
00 0
c^ c^ c**
cs cs cs
CO CO CO
o o o
\o \o vo
CS CN
-------
determined by using the costs for LNB/FGR for reheat furnaces
shown in Section 6.1.4 and multiplying these costs by 0.67.
The ratio in TCI's for LNB versus LNB+FGR capital costs is 0.67
for 265 MMBtu/hr natural gas fired boiler in the
Industrial/Commercial/Institutional Boiler ACT Document. 10 This,
data is used as data was limited for the iron and steel industry.
The only data for reheat furnace with LNB were for a 440 MMBtu/hr
furnace where the PEC costs were $300,000 and the total burner
system PEC plus direct and indirect installation costs for a 233
TPH(555 MMBtu/hr) furnace under construction are $ 3,110,000.7
As shown in Table 6-2, TCI for LNB controls for existing model
furnaces range from $ 150,000 to 310,000 and for new furnaces
from $40,000 to 90,000. These costs are annualized using
Equation (6-2), assuming a 5-year economic life and 7 percent
interest (CRF « 0.244). As shown on Table 6-2, annual costs
range from $ 38,000 to 74,000 for existing furnaces and from
$10,000 to $21,000 for new furnaces. A control efficiency of 66%
was used. Cost effectiveness for new reheat furnaces range from
$20/ton for the 520-MMBtu/hr regenerative-fired reheat furnace to
$ 200/ton for the 140-MMBtu/hr cold-air-fired reheat furnace.
Cost effectiveness for existing furnaces range from $70/ton for
the 520-MMBtu/hr regenerative-fired furnace to $760/ton for the -
140-MMBtu/hr cold-air-fired reheat furnace.
6.1.4 Costs and Cost Effectiveness of LNB Plus FGR Control
Purchased equipment costs (PEC) were provided for three
examples of LNB plus FGR controls for reheat furnaces.6'7
Estimates of direct and indirect installation costs were also
provided for one of these examples, and the sum of these costs is
106 percent of the PEC.6 Therefore, to convert PEC into total
capital investment(TCI) a factor of 2.8 is used which is
consistent with the factors used in other NOX ACT documents.
The example furnace capacity is 78 tons/hr (assuming a hot
charge) and 154 MMBtu/hr. Thus, it is reasonably consistent with
Equation (6-3). Its burner system PEC is $735,000 and thus, the
TCI for this example furnace is $ 2,060,000 for retrofit
installations (78 TPH) . For a new 78 tons/hr reheat furnace the
capital cost is 0.275($2.0 million) or $ 550,000. The portion of
the TCI attributable to the LNB plus FGR NOX control system over
the TCI of a conventional burner system is 15% of the total.11
Therefore, the TCI is $ 230,000.
In Table 6-3, the TCI for the NOX control system for this
example furnace is scaled to the model furnaces using the
relationship defined in Equation (6-1). It is recognized that
costs for a given facility, new or retrofit, are site-specific
and may differ greatly from the values used in this chapter.
Capital costs vary from $ 62,000 to $130,000 for new sources and
from $230,000 to 460,000 for existing sources.
6-5
-------
Cost
Effectiveness
($/ton)
J-e
1 *i
iil
«3 u S
Firing Type
Furnace
Capacity
I (MMBtu/hr)
|V§1
-.3 O O Q
(fl {T) ^M \O
X
SW1 O
^H *
| . . .
! « « «
| , , .
•a g g g
1 1
88,'?
& 1 1
Oi OJ O
•-
Vi O O
<«
-------
Cost
Effectiveness
($/ton)
*Ji
1 n|
x^-
3 -a §~
Firing Type
u £>=5
M -| S
^ o o o
2^r ^* ^O
^ ^M CS
•g O O O
S 8 8
1 a .» a-
f
,tS "^
| S S '8
a*
•5 o o o
IS en en en
•g ts «s cs
II
'I ! •*
ft, s -o
g1 8 "o
C^ P^ O
s
en en °
a I I.
en oo «N
en en en
es es «N
CO CO CO
00 OO 00
. . .
CO CO CO
1 1
1 1' I
43 45 O
06 06 U
'
O Q O
« 5 "2
8 1 1
^ en «s
O O O
en en en
O O O
CO CO CO
1 I.I
| |
§ 8 -a
If 1
06 06 U
e^
IT)
6-7
-------
The annual costs in Table 6-3 are the annualized costs of
the TCI. In lieu of definitive cost data, it is assumed that
other cost components, e.g., annual operating costs, are
approximately the same as comparable costs for conventional
burners and should not be charged to NOX control. In annualizing,
these costs, a 5-year life is assumed for the LNB system
(estimates from 3 to 7 years are provided), and a 7 percent
interest rate is used. These costs are annualized using the
capital recovery factor (CRF) defined in Equation (6-2).3 The
CRF is 0.244. Annual costs range from $15,000 to 32,000 for new
reheat furnaces and $ 56,000 to 110,000 for existing furnaces.
The NOX reductions in Table 6-3 are computed from the
controlled and uncontrolled emissions assuming 8,000 hours of
operation per year. A control efficiency of 77% is used to
determine emission reductions. For new reheat furnaces, the cost
effectiveness ranges from $ 30/ton for the 520-MMBtu/hr
regenerative-fired reheat furnace to $ 930/ton for the 140-
MMBtu/hr cold-air fired reheat furnace. The cost effectiveness
for retrofit reheat furnaces are 3-4 times the cost effectiveness
of new reheat furnaces.
6.2 ANNEALING FURNACES
6.2.1 Model Furnaces
Information was obtained on four annealing furnaces, one
that is operational with an operating history and three that are
under construction. All are strip-anneal rather than batch-type
furnaces. The operating furnace has LNB plus SCR for NOX
emissions control.12 Of the furnaces under construction, two will
have LNB plus FGR and the third will have SCR for NOX control.5>7>13
One of the LNB-plus-FGR-controlled furnaces is nearing completion
and some emissions data are available. The furnace capacities
are given or estimated from data as 127 ton/hr and 95 MMBtu/hr,
126 ton/hr and 180 MMBtu/hr, and 140 ton/hr and 95 MMBtu/hr. No
explanation for the sharp difference in product capacity to
firing capacity ratios is provided. For the purposes of this
section, furnaces with firing capacities of 100, 200, and 300
MMBtu/hr are selected as model annealing furnaces.
The uncontrolled emission factors used in this section to
calculate-emission reductions are listed below:
Firing Type
Regenerat ive
Recuperative
Cold-air
PPM@ 3% P.,
645
220
110
LB/MMBTU
0.792
0.200
0.135
6-8
-------
6.2.2 Costs and Cost Effectiveness of LNB Control of Annealing
Furnaces
The control efficiency.used for LNB control is 50%. _ Costs
and cost effectiveness are shown in Table 6-4. Cost estimates
for LNB controls are calculated by multiplying the cost of
LNB/FGR as given in Section 6.2.4 by 0.67 as was done for reheat
furnaces. Only cost data for LNB/FGR were available for
annealing furnaces. The results of these estimates are presented
in Table 6-4. As was done earlier for reheat furnaces, the
capital costs for new furnaces was determined by multiplying the
costs for existing furnaces by 0.275. Capital costs for retrofit
applications vary from $ 360,000 to 690,000 and for new sources
from $100,000 to 190,000. To determine annual costs a life of 10
years and an interest rate of 7% were used. Annual costs for
existing furnaces vary from $50,000 to 98,000 and for new
annealing furnaces from $14,000 to 27,000. For existing
annealing furnaces, cost effectiveness ranges from $170/ton for a
300 MMBtu/hr regenerative furnace to $l,680/ton for a 100
MMBtu/hr regenerative furnace. For new annealing furnaces, the
cost effectiveness range from $50/ton for the 300-MMBtu/hr
regenerative annealing furnace to $470/ton for the 100-MMBtu/hr _
annealing furnace fired with cold-air.
6.2.3 Costs and Cost Effectiveness of LNB plus FGR
The control efficency for this control was estimated to be
60%. This reflects the efficiency achieved by this control for
boilers. Costs for an LNB+FGR controlled, 180-MMBtu/hr annealing
furnace are estimated as $2,400,000 for the burner system PEC.5'
Using the 106 % (for installation costs costs) and 15 % (LNB_plus
FGR cost premium over a conventional burner system). TCI estimate
of $760,000 for LNB+FGR control for 180 MMBtu/hr is obtained.
Equation 6-1 is used to calculate costs for the model furnace
sizes.
In Table 6-5, the costs and cost effectiveness of the LNB
plus FGR control for model annealing furnaces are given. For new
annealing furnaces, capital costs range from $150,000 to
$280,000; annual costs from $21,000 to 40,000 and cost
effectiveness from $60/ton to 530/ton. Only .the cold-air
furnaces had cost effectiveness greater than $220/ton. For
existing annealing furnaces, capital costs range from $ 530,000
to 1.03 million; annual costs from $75,000 to 146,000 and cost
effectiveness from $210/ton to 1,900/ton.
6.2.4 Costs and Cost Effectiveness of SNCR for Annealing
Furnaces
SNCR has not been applied to any iron and steel mill
facilities. However, the similarity of SNCR to SCR suggests that
6-9
-------
3
1
3
§
u
i
u
f2
I
u
w
w
s
1.11
ur
C2
o 'w
§1!
PS ~
E
J ^ -
<£
•1 » i
u ° S
^
eo
•c
£
Furnace
Capacity
(MMBtu/hr)
S o o o
jjr f^ 8O f-
fill
i
5? o o
5) oo cr>
1 2 2.2
.S3 o - t~- t-
O 0 O
>o
>o in n <— > O
*— ' CO
g 1 i
00 ^ Os
R B S
OO 00 OO
ON ON ON
ON ON ON
ON ON ON
4) O
•5 '-IS
1 1 I
O S O
•
-------
*-> § c
ur
i'lt
^ £
1 "I
•^^
H!
Firing Type
o £» ^
>_ 3C YMI £f}
^* CS V)
S"
S 8 g 8
iS
8 ••
|/8 s a •
! s s P
^H « »-l
1 8 § -8 -
u
43 1>
1 1 ^ •
^11
« a; u
g
...
8 S i
O O O
o o o
en en _ en
§ a a.
° (M CS
2 2 2
00 00 00
•111
60 W *5
O 45 Q
eS «! O
•-
• a,
i ij
ti a
§ § g
a a. i
o o o
Sen en
O O
ti U
! i
CCu
1 fr 1
II -3
g
en
-11
-------
SNCR is potentially applicable to facilities that are controlled
with SCR provided the flue gas stream to be controlled has a
suitable temperature for the SNCR reaction to proceed. As
discussed in Chapter 5, a temperature suitable for the SNCR
control reaction to proceed is much higher than the temperature
required for the SCR control reaction. SNCR is considered
feasible for annealing furnaces and the potential control costs
and cost effectiveness of SNCR were developed.
Cost estimates have been developed for SNCR applications to
process heaters. In the process heater ACT, TCI for an SNCR unit
for an 80-MMBtu/hr heater is $392,000.14 In lieu of better data,
it is assumed that this TCI is the same as for an SNCR unit for
an annealing furnace of the same firing capacity. In Table 6-6,
TCI for the model annealing furnaces are scaled from the process
heater example using Equation (6-1). Capital costs range from
$460,000 to 880,000.
Operating costs are modeled in Reference 14 as follows:
1. NH3 Costs - (Q)*(lb NOx/MMBtu)* [0.370 (lb per mole
NH3/lb per mole NOX) ] * (mole NH3/mole
N0x)*($/lb NH3)*(8,760 hr/yr) * (CF) , (6-5)
2. Electricity Costs = (0.3 kWh/ton NH3) * (tons NHj/yr) *
($/kWh), and (6-6)
3. Fuel Penalty Costs
(0.015)*(Q)*(8,760 hr/yr)*($ fuel
costs/MMBtu)*(CF), (6-7)
where Q is the furnace-firing capacity in MMBtu/hr and CF is the
capacity factor, i.e., operating hours per year/8,760. The tons
of NH3/yr value in Equation (6-6) is computed as (Q)*(controlled
emissions-uncontrolled emissions) * [0.370 (lb per mole NH3/lb per
mole NOX)]* (moles NH3/moles NOX) * (CF) * (8760/2000) . The
0.3 kWh/ton NH3 value is taken from Reference 14, The fuel
penalty cost estimate (1.5 percent or the 0.015 factor in
Equation 6-7) represents a loss in thermal efficiency in a
process heater and is used here to account for possible thermal
losses from using SNCR on an annealing furnace.4
For the purposes of this section, the parameter values used
in the operating cost models are as follows:
Natural gas costs
NH3 cost
Moles NH3/moles NOX
$/kWh
CF
$ 3/103 ft3
$ 164/ton,
1.5,
0.07, and
8000/8760.
6-12
-------
S
u
i
<
ttl
ctf
1
I
w
&
s
ca
ea
•4
•JU
O JjHj ^p;
o C-
X *"^ «•?'*
0 3 H
-3i iil/
PS
ilsl
60 ,>>
1 i § 1
Pi
Su
60
"S
Furnace
Capacity
(MMBtu/hr)
V§ 1
o o o
a. 2 *
o o o
ir>
O O O
CO 00 CO
o o o
1 1
^ 2 «-
1 1 1
pg 05
-------
The annual operating costs are estimated using the preceding
models and parameters. The total annual costs are the sum of the
annualized TCI plus the annual operating costs. Costs are the
same for both existing and new furnaces although there may be a ,
wide variation in costs from one furnace to another. As shown on
Table 6-6, the total annual costs range from $150,00 to 370,000.
The NOX reductions in Table 6-6 are estimated with a NO^
reduction efficiency of 60%. Eight thousand hours of operation
per year at furnace capacities is assumed. The cost
effectiveness ranges from $540/ton for the regenerative-fired
300-MMBtu/hr annealing furnace to $3,800/ton for the 100-
MMBtu/hr cold-air furnace.,
6.2.5 Costs and Cost Effectiveness of SCR Control
The control efficiency assumed for SCR is 85%. The PEC for
an SCR unit in a construction permit for an 95 MMBtu/hr
annealing furnace is $1.25 million.15 Converting this to April
1994 $(multiplying by 1.018), converting PEC to TCI by
multiplying by 2.06 and using Equation 6-1, the capital costs as -
shown in Table 6-7 range from $2.7 million to $5.1 million. The-
annual operating and maintenance costs were determined by using
the costs in the Process Heaters ACT Document for the 69 MMBtu/hr,
distillate oil-fired heater and using Equation 6-I.14 The total
annual costs are the sum of the annual o&m costs and the
annualization of the total capital costs. An interest rate of 7%
and a life of 10 years were used to calculate the annualization
of capital costs. As shown on Table 6-7, annual costs range from
$550,000 to $1,040,000. Also on this Table, cost effectiveness
varies from $l,100/ton to $ll,000/ton. Only the regenerative
fired annealing furnaces have cost effectiveness .less than
$2,000/ton.
6.2.6 Costs and Cost Effectiveness of LNB plus SNCR for
Annealing Furnaces
The control efficiency for LNB plus SNCR was estimated to be
80%. Capital and annual cost estimates for these controls are
taken by adding the respective numbers in Tables 6-4 (for LNB)
and 6-6 (for SNCR). The resulting estimates of costs, NOX
reduction, and cost effectiveness are tabulated in Table 6-8.
For existing annealing furnaces, capital costs range from
$820,000 to $1.6 million; annual costs from $200,000 to 470,000
and cost effectiveness from $520/ton to 5,000/ton. For hew
annealing furnaces, capital costs range from $560,000 to $1.1
million; annual costs from $160,000 to 400,000 and cost
6-14
-------
Cost
Effectiveness
($/ton)
§1!
tr
pi
Firing Type
Furnace
Capacity
(MMBtu/hr)
** ** v^*
t"M '^f ^^
S 2 • o '
CN CO V}
CO "^
v>
Vi _^
1 I
S 8, '?
0 3 •«
00 O --^
o C O
ft (A U
8
US
0 O O
Cn Ch Cn
q q q
1 1
88.'?
8) S S
5y S o
CA & U
'
!- ! !
82°
IS!
ui vi >/S
Q> 4)
_> £
I \ 1
Bi f* O
8
6-15
-------
s
CO
3
a
Cost
Effectiveness
($/ton)
*!§
§11
pi **~*
*T3 •>?*
PI
Firing Type
I ^1
1,11
••= g s 8
•2 5 g 8
x « C3 Vi
H
8^D f^
rj ^
S O O O
2 o
o 6 6
•I g S S
•g d e> e
« g
's '2
1 f 1 :
& 1 a
8
g 8 I
•* a S
"« — cr,
IS •
OO 00 OO
s • s §
CO CO CO
00 OO 00
odd
«, 2" 2
4) O
III
1
111
•8' 1 1
I.I!
1 1 1
o o o
r~ r- r-
^ ^ ^
vo vq vo
1 1
& S «-
S 2, '?
1 1 3
a; as u
8
en
6-16
-------
effectiveness from $440/ton to $4,000/ton. Only the cold-air
annealing furnace has cost effectiveness greater than $2,000/ton.
6.2.7 Costs and Cost Effectiveness of LNB plus SCR for Annealing.
Furnaces
The control efficiency for LNB plus SCR is 90%. Capital and
annual costs for the LNB-plus-SCR are estimated by adding the
respective numbers in Tables 6-4 (for LNB) and 6-7 (for
SCR) . NOX reductions used in these computations are based on an
assumed 8,000 hours of operation per year at furnace-firing
capacities. As shown in Table 6-9, for existing annealing
furnaces, capital costs range from $ 3.1 million to 5.8 million;
annual costs from $600,000 to $1.14 million and cost
effectiveness from $l,lOO/ton to $12,000/ton. For new annealing
furnaces, capital costs range from $2.8 million to 5.3 million;
annual costs from $560,000 to 1.07 million and cost effectiveness
from $1,000 to 11,000/ton. Only the regenerative furnace had
cost effectiveness less than $2,000/ton.
6.3 GALVANIZING FURNACES
6.3.1 Model Furnaces
Information was obtained on three galvanizing furnaces. One
is a 79-ton/hr furnace, and the firing capacity of this furnace
is estimated from a test report to be 150-MMBtu/hr.5 Firing
capacities of the other two are given as 38 MMBtu/hr and
179 MMBtu/hr.7 Using these furnaces as representative of
galvanizing furnaces, firing capacities of 50, 150, and 200
MMBtu/hr are selected as model furnaces. The 150-MMBtu/hr
furnace has LNB's for NOX control, is a cold combustion air
furnace, and its operating temperature is several hundred
Fahrenheit degrees below many other galvanizing furnaces.16 The
other two furnaces have LNB plus FGR for NOX control and are
preheated combustion air furnaces.
The uncontrolled emission factors used to calculate emission
reductions are shown below:
Firing Type ppm @ 3%O? Ib/MMBtu
Regenerative 940 1.15
Recuperative 330 0.40
Cold-Air - 120 0.14
6.3.2 Costs and Cost Effectiveness of LNB Plus FGR Controls for
Galvanizing Furnaces
Burner system purchased equipment costs for the 38-MMBtu/hr
and 179-MMBtu/hr furnaces are given as $511,000 and $1,430,000,
respectively.7 Assuming the 10.6 percent (for installation costs)
6-17
-------
%
I
C6
?
M
2
§
en
Ed
I
I
i
I
H
Cost
Effectiveness
($/ton) |
o "5*
K *** .?"*
ill
erf s~'
lA
3 In c5
I
oo
E
Furnace
Capacity
(MMBtu/hr)
I!!!
1 § 8 1
fl - * a
CO ~"
1 R 1 1
1
>5J 00 00 00
F— 1 ...
a1
• •< *• 4 t ^u
W
O O
1 I -a-
& 1 I
erf erf o
8
51$
I '•§' 1
— CO P>
g g 2
\O CN «-i
222
OO OO 00
OO OO OO
•*•*•«*
in v> \r>
o «
erf erf" C5
•
>^ pf vo
§ 8 '§
-« P< r-
R o 9
I- 5 S
§ § §
f«M V"4 V««
CO CO CO
00 CO 00
fl>
-------
and 15 percent (cost premium for LNB plus FGR as compared with
conventional burner systems) factors, the TCI costs for NOX
control for the two burner systems are $158,000 and $442,000,
respectively. These costs were converted to April 1994 $ by
multiplying by 1.018. A cost was obtained per MMBTU ($4,240 for -.
38 MMBtu/hr and $2,510/MMBtu) for the 179 MMBtu/hr. For the 50
MMBtu/hr model furnace, the capital cost for existing galvanizing
furnaces was determined by multiplying 50 by $4,240. For the 150
and 200 MMBtu/hr furnaces, $2,510/MMBtu was used. The capital_
costs for new galvanizing furnaces were determined by multiplying
the capital costs for existing furnaces by 0.275. As shown on
Table 6-10, capital costs for existing furnaces range from
$210,000 to 500,000 and for new furnaces from $60,000 to 140,000.
Operating costs are assumed to be approximately the same as
operating costs for a conventional burner system and, thus, not
chargeable to NOX control. Capital costs were annualized using
an useful life of 9 years and an interest rate of 7% giving a CRF
of 0.153. Annual costs for existing furnaces range from $32,000
to 77,000 and for new furnaces $9,000 to 21,000.
The control efficiency estimated for LNB + FGR is 60%.
The emission reductions shown on Table 6-10 are also based on an-
assumed 8,000 operating hours per year at furnace-firing
capacity. Cost effectiveness for existing furnaces range from
$140/ton to 2,000/ton. For new furnaces, cost effectivenes range
from $40/ton to 560/ton. '
6.3.3 Costs and Cost Effectiveness of LNB for Galvanizing
Furnaces
In lieu of any cost data for the LNB-controlled galvanizing
furnace, the cost estimates used in Table 6-10 for the LNB-plus-
FGR-controlled model furnaces are multiplied by 0-.67 to obtain
the costs for LNB only controls. This is the same factor used
earlier to calculate costs for these combustion controls for
reheat and annealing furnaces.
LNB is assumed to reduce NOX by 50%. This estimate is used
to compute the NOX reductions included in Table 6-11. These
reductions are based on an assumed 8,000 operating hours per year
at furnace- firing capacity. The capital costs range from
$140,000 to 340,000 for existing furnaces and 40,000 to 90,000
for new furnaces. Annual costs range from $21,000 to 52,000 for
existing furnaces and from 6,000 to 14,000 for new furnaces.
Cost effectiveness range from $llO/ton to $l,500/ton for
existing furnaces and from $30/ton to 430/ton.
6-19
-------
1
o
g
•<
o
I
ft
-J
fe
o
H
u
Cd
fe
8
en
I
U
Ed
J
CO
ts If
ur
di
OS
- ^
|0|
1 « §
<3°S
Firing Type
111
^ o g o
S1
1 '8 1 I
Ed
i-H .
| . . ,
*
| 8
^^ ^5 ^3 vg^
.5 m 0 O
\3 O ^_ ^-
| H " N-
.1 I •
!}!
1 1 a
o
g O O
? 2 8
1 3 °
rf —<
a a a'
oo oo oo
8 ,8 8
OO 00 00
CO CO CO
> 5
'' 1 '-a
1 1 1
8
o ° 2
* "-1 O
— tn
III
8 S .P
cv c< ts
§ i 5
•:> •
1 i
1 ! 1
1— C
6-20
-------
o
<
i
I
J
(El
O
8
H
O
w
I
on
S
8
v£
w
1
§ 'e'
.£ 2
S tO
S o
U
W)
III
S &§
22
-t
M
I «
| 2
a
8
«s
K
S
O
0
8
oo o
en ON
§
CS
ON
CO CO
bo
o
ss
O
o
6-21
-------
-------
6.4 REFERENCES
1. Economic Indicators: Chemical Engineering Plant Cost Index.
Chemical Engineering. July 1994. p.174.
2. Cooper, C. D.f and F. C. Alley. Air Pollution Control: A
Design Approach. Boston. 1986. 73 pp.
3. Vatavuk, W. M. Cost Estimating Methodology. In: OAQPS
Control Cost Manual, Fourth Edition, Vatavuk, W. M. (ed.).
Research Triangle Park, NC, U.S. Environmental Protection
Agency. Publication No. EPA 450/3-90-006. January, 1990.
p 2-13. . . .
4 Letter and attachments from Lee, D., Cascade Steel Rolling
Mills, Inc., to Jordan, B. C., EPA/OAQPS. July 6^1992.
Response to Section 114 letter on iron and steel mills.
5. Letter and attachments from Harmon, M. L., LTV Steel
Company, to Jordan, B. C., EPA/OAQPS. July 2, 1992.
Response to Section 114 letter on iron and steel mills.
6. Letter and attachments from Postlethwait, W., Nucor Steel,
Darlington Division, to Jordan, B. C., EPA/OAQPS.
December 2, 1992. Response to Section 114 letter on iron
and steel mills.
7. Letter and attachments from Gilbert, F. C., North American
Manufacturing Company, to Jordan, B. C., EPA/OAQPS.
November 11, 1992. Response to Section 114 letter on low-
NOX burners.
8. Letter and attachments from Finke, H. P., Bloom Engineering
Company, Inc., to Jordan, B. C., EPA/OAQPS. October 14,
1992. Response to Section 114 letter on low-NOX burners.
9. Letter and attachments from Felton, S. S., Armco Steel
Company, L. P., to Jordan, B. C., EPA/OAQPS. June 26, 1992.
Response to Section 114 letter on iron and steel mills.
10. Alternative Control Techniques Document-- NO,. Emissions from
Utility Boilers. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-453/R-94-
023. March 1994.
11. Telecon. Gilbert, F., North American Manufacturing Company,
with Parker, C., Research Triangle Institute. Section 114
letter response.
6-22
-------
12. Letter and attachments from McMahon, K. R., USS POSCO, to
Jordan, B. C., EPA/OAQPS. August 17, 1992. Response to
Section 114 letter on iron and steel mills.
13. Letter and attachments from Sapia, J. P., Bethlehem Steel
Corporation, to Method, T. J., Indiana Department of
Environmental Management. January 18, 1991. Construction
permit application.
14. Alternative Control Techniques Document--NOX Emissions from
Process Heaters. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-453/R-93-
015. February 1993. pp 6-6 to 6-7.
15. Letter and attachments from Sapia, J. P., Bethlehem Steel
Corporation, to Method, T. J., Indiana Department of
Environmental Management. January 18, 1991. Construction
permit application.
16. Telecon. Sistek, B., LTV Steel Company, with Parker, C.,
Research Triangle Institute. February 24, 1993. Response
to Section 114 letter.
6-23
-------
CHAPTER 7
ENVIRONMENTAL AND ENERGY IMPACTS OF NOX CONTROLS
7.1 INTRODUCTION
This chapter discusses the environmental and energy impacts
of the NOX emissions control techniques identified in Chapter 5.
The control techniques that have been applied or are potentially
applicable to iron and steel mill process facilities are the
following:
Facility
Reheat furnace
Annealing furnace
Galvanizing furnace
Control technique
LNB
LEA
LNB plus FGR
LNB
SCR
SNCR
LNB plus FGR
LNB plus SCR
LNB plus SNCR
LNB
LNB plus FGR
Of these control techniques, LNB's, FGR, and LEA are combustion
modification techniques, and SCR and SNCR are postcombustion
control techniques. Potentially, all techniques can impact other
air emissions in addition to NOX, and all may have energy
impacts. None have water pollution impacts. However, SCR may
have a solid waste impact.
7.2 AIR IMPACTS
7.2.1 NOX Emission Reductions
NOX emission reductions for the emission control techniques
are summarized in Table 7-1. Control efficiencies range from 13
percent for LEA on a reheat furnace to 90 percent for LNB plus
SCR on an annealing furnace. Emission reductions range from 40
tons/yr for LNB control on a 150-MMBtu/hr, cold-air galvanizing
furnace to 730 tons/yr for LNB plus FGR controls on a 300-
MMBtu/hr regenerative reheat furnace. The cold-air galvanizing
furnace reduction is low because the uncontrolled NOX emissions
are low for this relatively low-temperature furnace.
7-1
-------
S3 oj
o G
CQ tjc<
ivj ^ t"i ^^ •
rn O CJ CM <—
CO !S D E-" Q
1 r8
i4 O H
|— I I I U1
£ X E"* ^ j3
8 ° °* H
H W ^CD
w rt g
§ w
Sf ° S
O H — S
H §g^ji
i Q *— * S
CO § §
§ rt
H
B g^S
g CO O S
g H rf g
0 SS|
CQ
CQ
H
8 #
o
8
H
C-
W
Q H
P &
B
w
o
I
o 3 § g ° g 3 § S
i— 1 rH *"• ""I t"
_OO OOOOOO
*»o oo VD o o c^ in oo
^HH rHOJCNMCNCS
OOO OOOOOO
dCNn ooinin'^fooo
tH VO f*» CO ^J* ^^ \& ^O **&
OOO OOOOOO
OOO OOOOOO
mmro oaojojcNcsc^
(V (V P, ct<
CD CD fe CJ
pTi p^| ^. ^~* 1^/3
"^g ^ 1' ™ " B 1
§ § a J
CD
§ H
w d
W H -
H g
o o
«* in
o o
CM •*
H H
0 O
«* CN
rn ^*
o o
in in
H rH
«
CD
fn
14 03
J
1
N
y
3
>
CD
7-2
-------
7.2.2 Air Impacts of Combustion Modifications
Combustion modification techniques applied to process
facilities at iron and steel mills are LNB's, LNB's with external
FGR, and LEA. LNB's incorporate one or more of the following
combustion modification techniques: LEA, off-stoichiometric or
staged combustion (either staged air or staged fuel), and FGR.
Any of these combustion modification techniques have the
potential to increase carbon monoxide and unburned hydrocarbon
emissions. The NOX reduction mechanisms inherent in these
modification techniques are the reduction of peak flame
temperatures that are exponentially related to the formation of
thermal NOX and the reduced availability of excess oxygen needed
to form NOX. Reducing the availability of oxygen to the
combustion process increases the likelihood that some HC will not
be burned and that some CO will not be oxidized to C02. However,
it is not necessarily true that CO or HC emissions will increase
if combustion modification techniques are used to limit NOX
emissions. If the control is properly designed and applied, NOX
control can be achieved without increasing CO or HC emissions.
Experimental evidence from tests with process heaters, for
example, indicate that decreases in excess oxygen levels begin ta
impact CO emissions below about 3 percent excess oxygen.1 Also,_ -
data from glass melting furnaces indicate that CO emissions begin
to increase at excess oxygen levels of 3.2 percent.2 Limited
data for ICI boilers indicate that HC emissions do not change due
to the implementation of combustion modifications.3 Data in the
utility boiler ACT show that CO emissions increase for some
boilers and decrease for some boilers when implementing
combustion modifications.4
For the steel reheat furnace with LEA for NOX control, CO
emissions of 39 ppm at 3 percent O2 and 26 ppm at 3 percent O2 are
reported for the uncontrolled and controlled tests,
respectively.5 These results indicate that CO emissions do not
necessarily increase when LEA is implemented for NOX control. It
should be noted that, in this example, LEA is actually
implemented using a PID control loop that functions to satisfy a
CO ppm setpoint input. As the furnace operates independently to
achieve individual zone temperature setpoints, the CO control
loop continuously adjusts zone oxygen setpoints to minimize
excess oxygen via the air-to-fuel ratio in each zone without
violating the CO controller setpoint.
Other available sources report CO emissions from
LNB-controlled reheat furnaces to be in the range of 20 to 30 ppm
at 3 percent 02.6 Further, laboratory and field tests on a
galvanizing furnace yielded NOX emissions of 550 to 1,200 ppm at
3 percent 02 and, concurrently, negligible CO emissions.
Modifications to the burners reduced NOX emissions to 350 to 430
ppm at 3 percent O2 and, concurrently, increased CO emissions to
7-3
-------
30 to 60 ppm at 3 percent 02.7 The purpose of the modifications
was to reduce NOX, and further modifications were planned.
No data are available relevant to the impact of NO, control
techniques on HC emissions from these iron and steel facilities. •
In general, controls that reduce NOX by reducing the availability
of excess oxygen in the high-temperature regions of a furnace
would tend to increase both CO and HC emissions if the control
were not properly designed and applied. In each case, given the
availability of oxygen, available CO and HC molecules could be
further oxidized according to the following reactions:
2CH4
30
2 CO + O2
2 CO + 4^0
-* 2CO,
(7-1)
(7-2)
The reaction rates for HC to HjO + CO and CO to C02 differ; thus,
they are competing reactions. Nevertheless, reducing the
availability of oxygen to reduce NOX can also inhibit these
reactions, resulting in an increase in both CO and HC emissions
if the oxygen availability is reduced excessively.
Reheat, annealing, and galvanizing furnaces predominantly
use natural gas as a fuel. Natural gas does not contain sulfur,
and thus, SO2 is not a pollutant of concern.8 Gaseous fuels,
including natural gas, can produce soot and carbon black when
burned if insufficient oxygen is present.9 However, no
information on the effect of NOX controls on particulate
emissions in iron and steel facilities is available.
7.2.3 Air Impacts of SCR and SNCR
As described in Chapter 5, SCR units are add-on, flue gas
treatment facilities that reduce NO, by injecting ammonia
upstream of a catalyst reactor. Within the catalyst, NOX reacts
with the NH3 and is reduced to N2 and H20. There is a potential
for unreacted NH3 to escape with the flue gas from the SCR unit.
Any such emissions are referred to as ammonia slip.
Section 114 respondents cite two instances of SCR controls
on annealing furnaces at iron and steel mills. One of these
units is operational with more than 3 years' operating history,
and one is still under construction.10'" In the case of the
former, the NH3 slip was guaranteed to be less than 10 ppm
initially and less than 12 ppm after 1 year. The observed NH3
was initially less than 10 ppm. Subsequent observations are not
reported. The typical NH3/NOX molar ratio for this unit is O.9.11
For the unit under construction, a NOX reduction of 90
percent is guaranteed.10 The anticipated NH3/NOX molar ratio is
1.0. There is no NH3 slip guarantee. The construction permit
7-4
-------
requires 80 percent NOX reduction and a minimum NH3/NOX molar
ratio of 0.8. There are no requirements relevant to NH3 slip.
Other reports indicate that NOX removal rates of 70 to 90
percent can be achieved with SCR using NH3/NOX molar ratios
between 0.9 and 1.0, and that the NH3 slip will be between 5 and
10 ppm.12 These levels are considered to be well below health and
odor thresholds.13 The Occupational Safety and Health
Administration (OSHA) standard for NH3 in the workplace is 25 to
30 ppm, and in the atmosphere it tends to form nonhazardous_
compounds and disappear within 2 weeks.14 In other applications
in the United States, permitted NH3 slip levels are often on the
order of 5 to 10 percent. In Europe, 5 ppm is typically
specified and achieved, and levels as low as 0.5 to 2.0 ppm have
been achieved; and in Japan, levels as low as 4 ppm are
specified.13
No examples of SNCR applications to iron and steel mill
process facilities have been found. In other SNCR applications,
ammonia slip is controlled to acceptable levels by controlling
the ammonia or urea to NOX molar ratio. These levels are similar
to ammonia emissions from SCR applications, e.g., 10 ppm. - -
Pilot-scale testing and chemical kinetic modeling of SNCR
processes have shown that nitrous oxide emissions are a by-
product of both ammonia and urea injection. The N20 formation
resulting from these processes has been shown to be dependent on
the reagent used, the amount of reagent injected, and the
injection temperature.15
Full-scale tests on fossil-fuel-fired boilers have shown
that direct emissions of N2O are less than 15 ppm and do not
generally correlate with NOX emissions. Results from tests on a
small pilot-scale combustor using NH3 injection and encompassing
a range of test conditions show a maximum N2O production of about
45 ppm corresponding to an initial NO level of 700 ppm, an NH3
injection rate corresponding to a NH3/NO molar ratio of 2, and a
temperature of 880 °C (1,620 °F). The corresponding NOX
reduction was 75 percent. The peak NOX reduction corresponding
to an initial NO level of 700 ppm and an N/NO ratio of 2 was
about 95 percent at 980 °C (1,800 °F). The corresponding N2O
production was about 21 ppm or 3 percent of the NOX reduction.
N2O production is higher for urea injection than for NH3
injection. Given the same initial NO level (700 ppm) and_
reactant-to-NOx ratio (corresponding to a NH3/NO molar ratio of
2), the peak N20 production was about 160 ppm at 980 °C (1,800
°F) and the corresponding NOX reduction was at a peak value of
about 82 percent. Thus the N2O production was about 28 percent
of the NOT reduction.15 . ;
7-5
-------
Other pilot-scale tests have shown that, with urea
injection, the level of CO present affects the NOX removal
efficiency and the NOX removal temperature dependency. The
presence of CO also affects N20 by-product emissions. In the
temperature range of 820 to 870 °C (1,500 to 1,600 °F), for
example, increasing CO from 0 to 1,000 ppm caused an increase in
N20 emissions from 10 to 35 ppm. The initial NOX level was 125
ppm and the NH3/NO ratio was 2.16
7.3 SOLID WASTE IMPACTS
Combustion modifications do not have solid waste
environmental impacts. There is a potential for SCR to have
solid waste impacts in the disposal of spent catalyst materials.
Titanium dioxide and "vanadia/titania" have been identified
as the catalysts in the two SCR annealing furnace applications
cited.10'11 Other commonly used materials are vanadium pentoxide,
tungsten trioxide, platinum, zeolites, and ceramics.12'" Of
these, vanadium pentoxide is a toxic compound and a cause for
concern. However, worker safety precautions adequately prevent
any increased risk to workers handling the catalyst, and stack
emissions of vanadium pentoxide are 1 million times less than
industrial worker exposure.13
Most catalyst manufacturers arrange to recycle and
reactivate the catalyst. When that is not practical, the spent
catalyst can be disposed in an approved landfill in accordance
with the Land Disposal Restrictions in 40 CFR Part 268,
Subpart D. The volume of catalyst materials has been low and
their useful lives are 4-5 years for coal-fired boilers, 7-10
years for oil-fired boilers, and more than 10 years for gas-fired.
boilers.13
7.4 ENERGY IMPACTS
All of the combustion modification control techniques have
the potential to impact energy requirements by affecting the
thermal efficiency of the process. No data are available to
quantify the impact of these controls on iron and steel mill
process facilities. A lack of discussion of these issues in theft
literature suggests that these impacts, if any, are not a major
concern in installations using combustion modifications.
In the case of the SCR, the pressure drop across the
catalyst requires additional electrical energy for the flue gas
fan. Typical flue gas velocities over the catalysts are about 20
ft/s, and the pressure drop is about 3 inches of water. The
volume flow of one typical unit is 1,550 ft3/m. Given these
parameters and assuming an efficiency of 0.65 for the motor and
fan, the additional power demand is 0.84 kW, which corresponds to
7-6
-------
$540/yr assuming 8,000 hours of operation per year and
electricity at 8*/kWh.17
7-7
-------
-------
7.5 REFERENCES
1.
U.S. Environmental Protection Agency. Research and
Development, Emissions from Refinery Process Heaters
Equipped with Low-N0x Burners. Industrial Environmental
Research Laboratory. Research Triangle Park, NC.
EPA/7-81-169. October 1981.
Barklage-Hilgefort, H., and W. Sieger. Primary Measures for
the NOX Reduction on Glass Melting Furnaces. Glastech.
62(5):151. 1989.
Alternative Control Techniques Document --NO,. Emissions from
Industrial/Commercial/InstitutidnaKICI) Boilers. U.S.
Environmental Protection Agency. Research Triangle Park,
NC. Publication No. EPA-453/R-94-022. March 1994.
4. Alternative Control Techniques Document--NOX Emissions from
Utility Boilers. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-453/R-94-
023. March 1994. , -
2.
3.
5.
6.
7.
8.
9.
10.
11.
Letter and attachments from Lee, D., Cascade Steel Rolling -
Mills, Inc., to Jordan, B.C., EPA/OAQPS. July 6, 1992.
Response to Section 114 letter on iron and steel mills.
Letter and attachments from Dickinson, G., Engineered
Combustion Systems, Inc., to Jordan, B.C., EPA/OAQPS.
October 12, 1992. Response to Section 114 letter on low NOX
burners.
Letter and attachments from Gilbert, F.C., North American
Manufacturing Company, to Jordan, B.C., EPA/OAQPS.
November 11, 1992. Response to Section 114 letter on low
NOX burners.
Perry, R.H., C.E. Chilton, and S.D. Kirkpatrick. Perry's
Chemical Engineers' Handbook, Fourth Edition. New York,
McGraw-Hill Book Company. 1963. p. 9-8.
Devitt, T.W. Fossil Fuel Combustion. In: Handbook of Air
Pollution Technology, Calvert, S., and H. England (eds.).
New York, John Wiley and Sons. January 1984. pp. 375-417.
Letter and attachments from Anderson, D.M., Bethlehem Steel
Corporation, to Jordan, B.C., EPA/OAQPS. July 13, 1992.
Response to Section 114 letter on iron and steel mills.
Letter and attachments from McMahon, K.R., USS POSCO, to
Jordan, B.C., EPA/OAQPS. August 17, 1992. Response to
Section 114 letter on iron and steel mills.
7-8
-------
12. Garg, Ashutosh. Trimming NOX From Furnaces. Chemical
Engineering, p. 122-128. November 1992.
13. Institute of Clean Air Companies, Inc. Selective Catalytic
Reduction (SCR) Controls to Abate NOX Emissions, A White
Paper. Washington, DC. September 1992. 25 pp.
14. Makansi, Jason. Ammonia: It's Coming to a Plaint Near You.
Power. p. 16-22. May 1992.
15. Muzio, L., and T. Montgomery. N20 Formation in Selective
Non-Catalytic NOX Reduction Processes. .Paper presented at
1991 Joint Symposium on Stationary Combustion NOX Control.
Washington, DC. November 1991.
16. Teixeira, D. Widening the Urea Temperature Window. Paper
presented at 1991 Joint Symposium on Stationary Combustion
NOX Control. Washington, DC. November 1991.
17. Vatavuk, W.M. Cost Estimating Methodology. Ins OAQPS
Control Cost Manual, Fourth Edition. U.S. Environmental
Protection Agency. Research Triangle Park, NC. Publication
No. EPA 450/3-90-006. January 1990. p. 2-13.
7-9
-------
APPENDIX A.
TABULATION OF UNCONTROLLED NOX EMISSIONS DATA
A-l Coke Ovens (underfiring)
A-2 Sinter Plants
A-3 Blast Furnace Stoves
A-4 Basic Oxygen Furnaces
A-5 Electric-Arc Furnaces
A-6 Soaking Pits
A-7 Reheat Furnaces
A-8 Annealing Furnaces
A-9 Galvanizing Furnaces
-------
-------
s?
•H
t!
0)
!§
CQ
C
I
1
a
1
o
4J
C
Comme
*
IW
Cd
4J (1)
m jj
0) (8
E-T3
Sir
S*
3 m
T3 a
0 0
O, «"
01
4J
Emissions uni
m
c
o
•H
m
m
X-H
o e
55 4)
dl
-5 3
JJ
4J CQ
c8 2
Oi 2
a —
"
t-H
s
EL,
rH
0)
4J
0
55
n
CO
00
n
^_^
M
0
55 rH
(8
m o
(8 3 O
H CQ C
T3 rt 2 C
CTl O
m in 01
fO fl O rH
n
. X
s e^
CU pi .0
pt P[^-
CO CO r-
•* CS 0
r-
O
o
s
CTl
*
CO
O
O
rH
M
O
<*>
n
e E~-
in o
n cs
o cs
t-H
O
U
cs
0)
4J
O
55
CD
w .
cs
o
\
rH
.-*
O
<*= O
CS
C- M Z
CQ
® ® 2
6S|
t—
H CN M
rH H
^
ro
vo
O
EL,
CQ
C3
O
U
cs
0)
4J
o
s
0
CTl
M
rH
CTl
m
O
04
<*> 0
• <*> 3
vo ro -U
CQ
® ® 2
2
p) f^[ ^£
VO
0
CTl O CS
rH rH
in
CO
CO
O
EL,
CQ
U
O
U
-------
•o
to
I
4J
O
u
-S,
M
PH
P-i
fl
•H
•H
4-4
0)
rj
g
CO
rl
0)
^
CD
0)
*p?
u
tn
a
§
a
E
o
u
~
""•
0)
Pi
4J 4)
1 to 4J
4) m
H TJ
c
0 —
•H M
1 1 Jf*
0 ~-s
3 tn
•a c
0 O
~~"
tn
4j
•H
C
3
tn
c
0
-rl
tn
tn
•rj
£
H
ta
c
0
•rl
ta
ta
X-rl
o E
•g 4)
n'l
•rl 3
U
U £0
m 2
4) 2
K —
rH
4)
3
O
en
^
N
0
Cn
^,
H
H
M
O
c>P W
ro o
o #> 3
H ro 4J
a
© © 2
2
e e •>-.
Qj flj .Q
m
in
rH •* O
co p-
eg *f
O
O
ro
0
«
O
O
U
&
o\
^
M
CO
CO
•*s^
r-
N
0
M
tX> 0
• #> 3
r- n 4J
CD
© © 2 ^l
2 fi
ft ftrH rH
(N
ro
•*
ro co o CA
eg 01 ^*
rO ^ rH
in
^J4
ro
O
CO
o"
o
0
z
en
*
3
H ro 4J
.a
© © 2 M
2 . i
g £ ^^^^x
Pt Pi.n .0
O4 OirH r-H
CO
in
cs j
rH CN
ID
0
rH
O
O
U
^
^•^
cn
o
o-v
^>^
CN
in
0
. —
01 (tf 3
>
Q4 iir-l rH rH
oV
ro vo
CO O (N ^< ^
p- ui o eg o
co cn ro
U)
•
in
ro
H
O
O
O
^
~~-
cn
0
cn
*-^
cs
ro
p-
r>
»—
Oi rij 3
> 4) 4J
m a,m c
*— •* ^— ' ^S* Li A
A M u
2 S *"*•* ***NI ^>*
Q. O, O ^Q f>
Qj CUrH rH rH
ro rH
p* o eg cn ^
P- in 0 rH 0
co o*\ ro
VD
•
Ul
ro
H
O
O
u
2;
^.^
cn
o
a\
^*^
og
in
u>
—
01 ffl 3
> 4) JJ
r8 CUCQ C
^^ ^-^ 5?" Ll A
•£< rl w
S £ "^ ^x*^^-
P, p, Q Q Q
(i O
a a
— k 0
x: 4J
£ ***» ^^
O4X) XI
CO
Oft?
o ^ o
Cn rH
o
o
u
A-2
-------
t
•H
-H
4-1
r-l
CD
"S
CQ
fl
(U
0)
0
CQ
g
g
o
U
I
«j
41
JJ 4)
CO JJ
4) Hi
EH T3
C
0 ~
•H M
JJ XI
U ^
3 CQ
T3 C
0 O
rl JJ
2
•rH
C
3
CQ
s
0
•H
§
H
a
CQ
C'
C
a
Cf
O £
S 4)
S
c -^
•H 3
JJ
JJ CQ
(S S
4) S
EC —
ffl
Be.
t
0
•«— '
i CD
ro
C?
**-"
0V
0
G\
*^»
CO
CO
rH
ro
__
IS C
XI JJ
ftXl X!
ftr-l rH
C\
O U> O
O ^* r-
0 CO
C*J
t
O
b
CQ
CU
0
c
0
XI
CQ
tn
•H
rl
73
m
ca
0)
CQ
O
O
O
T3
IS
XI
•H
CQ 4)
0) >
rH O
ft CQ
g T3 QJ
IS C g
CQ 0 -H
I CJ JJ
tN 4)
CQ JJ
• • IS 4)
4) M
r-i a co
ft 0 M-l
IS 4) -H
CQ WO
M
4) Xl IS JJ
r-l IS XI IS
Xi r) O
is tn >•«
tn M
0)
o
•H
c
•H JJ
. 3 IS
tJ M m g
O i O
CJ IW
0
JJ >
c • c
is tn CD
^ 5 S 5
o. «; o Q
4-1
' is
• 4)
JJ
0
4)
JJ
o
A-3
-------
s
0
M
tn
CO
CQ H
55 H
5H
H H
CO t3
CO H
H CJ
§ r*
XCO
O CO
& to
§1
^ &
>4
Q J
Qj |_*|
pj H
5^5 £
O
P
6g
CO
•
C4 P
1 jgj
^i ^
a
X O
M p{
S H
U
CM
PH
<
«— — »
CQ
^^
o
rt
4J
CQ
X
0
f
-H
gj
*— •
C/}
4J
C
5
O4
S-i
0)
4J
rt
-H
CO
o
JJ
d
E
E
0
U
*
*— *
,
U-l
a)
erf
4j Q>
01 JJ
1) nt
EH TJ
O -^
•H ^
JJ 45
3 a
•0 d
o o
M JJ
ft —
a
jj
-r)
d
3
01
o
•H
m
CO
•H
E
H
a
d
o
•r)
m
CQ
X-H
O E
-H 3
JJ
JJ CQ
CS E
0) S
K —
i-H
(U
3
Cu
a
•^^
CTI
^
rH
r-
00
•v^
in
i1
r-
in
H
^^
01
(S
"Si
>3
to JJ
— CQ d
son
> E JJ 43
S3 3 3
m oo
rH 0
o N cf> in
in o o oo
*j* ^*
p-
•
ro
^
cs
O
aT
>^<
o
0
rH
0)
u
0
hH
^-*
CTl
*
rH
0
•
O
00
H
.M.
01
^ rt 'oi
nf O) nj
0) > 3 —
ft 18 JJ
— — ea d
E 0 M
> > 3S AJ.C
Qj OiiH i-H rH
in CTI
CTI ro
O O *}" «a* d
CN in o o UD
\O lf> 00
^4
•
^
t-
^
O
0)
l>
0
u
0
-»-•
CTl
*
tN
CTl O
' CO CTl ^-^
"*^ n ***» &•
CTl 0 CTl -"
1
(B ^
— O1
M (8
O *^ ^^
01
(M >
m o n!
» -^^
in *>
rH m 3
jj
© ® CQ
E E ^^--
pt pi n Q
O. Oli-H rH
CTl
CO
o in cs
CTl rf O n
CO CO CO
CJ M
in
•
01
(T)
"*
o
o
CD
oT
^*
o
o
£5
—
CTl
».
3
n JJ
® CQ .
® S in
IE-^
Cu CU o _Q
Cb CXrH rH
^t*
vo
o in M
in in o m
CTI CTI n
m cs
in
•
m
rH
«*
o
o
u
0)
r^
O
u
^^^
<:
^-^
CTl
0
CTl
^^
rH
0
in
H
^x
•^^> pirf
01 0)
^"^ *— • J^
M 0
Pi Q{) ^Q VQ
Ol GirH rH
p»
CTl
in in in o
in u> ^< o
rH
i
i
^-^
0)
-53*
>H O
E E^.^.
Pi Pi _f^ c^
CU O-K
^
01 rt
> 0)
<8 O4
.— — • d
>>|S
a ax) xi
ft fti-H rH
P-
H
CN P- ^< rH
P- P- CTl O
rH
t
O
O
u
a)
^*
o
• u
A-4
-------
3
8
S
•*" -H
CO
4J
s
rH
CM
rH
m 4J
o -.
~4J 43
U^
3 m
•a c
o o
Emissions
S H
CN ID o
rH CN O
M 0
> A U
~-
> X!
OUr-H rH
m
m
H
R1
i' c
r4 0
> 43 W
ma xi
CUrH 1-H
O 0 rH
M in O
o
0
0
u
O
U
en
in ^*
o cs in
<4< vo o
^ co
0
U
rH
O
M vo in
o co o
^< P-
co
en
o
-------
•r-~»»
ca
o
to
CO
X
O
=9
a
•H
&
W
4J
9
rH
04
t 1
o
•H
CO
1 *""
1
L
MJ
a)
-•
l| CQ 4J
11 fl) nJ
EH TJ
1 C
1 vr^ ^
II O^S-
3 a
•a c
o o
11 M xJ
* ""*
i
i
i
j CO
4J
"c
3
tn
c
o
•rl
a
m
'E
m
c
0
•H
a
X-H
0 6
S 0)
17
'J3
•rl 3
4J
ij ca
nJ •£•
m E
rH
0)
3
£J
01
N
O}
•^
in
n
in
H
?
fd
-- c
rl O
O.J3 JQ
i-H
(Tl
O 00 U)
in in o
in o
H
i
i
P?
a\
at
^^
in
n
in
rH
?
rt
^^ C
>H O
e\\
pi Q f>
O* i^ i~H
CO
o^
o t*» vu
r** vo o
w o
H
1
1
ca
u
•H
(U
13
o
o
Pk
rt
u
0.
a.
rt
0
o
4J
a
a
A-6
-------
en
OJ
JJ
CO
g
§
en
tC
iH
PQ
ta
JJ
0)
g
0
u
H-l
0)
JJ 0)
tn JJ
0) (8
H
3 tn
0 0
H JJ
p. •—
ta
jj
•H
d
3
to
o
-H
ta
01
•H
U
to
c
o
-H
tn
m
X-rl
O g
a «>
•rl 3
Jj
JJ • CQ
m 2
?E *—*•
CO
in
o
u>
tH
?
(8
tn 3
> -W
i-
— -"— S
2
g g
>
Cil Qj
CO
CO
• *> 3
d JJ
ca
r- 01
Ul t-
O O VD
o o in o o
4) 3
(8 OiJJ
— ^- CO
S M
> > 2 .C
g
&, O.rH rH
CO -*
iH-cn-fM
r> o o co o
in co o co o
0
r-
C3 C3
fc O
CQ U
ijl (S
> 4) 3
ro Q.JJ
~^~- CQ
2
> > 2
g
O. &rH r-l iH
in
(N
r> o o o
co in o o o
H CN CN
CM
r-
U O
fa O
m u
tn rt
> a)
rt
2 i-i C
>
g
Ot ftXl Xl -Q
Q, ft.r-1 i-t
IM
O 0 •
o n o vo o
CJ M H
O
co
r-
ua
ft. O
CQ U
ppmv (avg
Ib/MMBtu
Ib/hr
Ibt
VD,
rH
0
0 CM 0
u o
fo O
ca u
A-7
-------
in
SP
>
S^
4-<
CO
CD
0
(U
e
rj-
ra
rH
CQ
I
0
u
1 v
erf
4J 4)
(0 JJ
1 0) a
1 Hr°
§~
•rl JH
0^
3 0
-0 n
0 O
1 M JJ
o. ~-
4->
•H
3
§
°
•H
tfl
a
I •r|
1
0
•H
X-H
O E
g a)
•H>
u
4J CO
(8 2
0) 2
a —
«
3
W
en
CO
CO
"^
o
rH
r-
O
M
?3
18 4J
— CQ C
2 in 0
ttrH rH rH
VD CO
H *f ro
o en o
c\i o vo o
rH O O O
VO
•
VO
ro
<*
o
fa
CQ
CB
m
00
00
^.
o
rH
[^
O
C*3
(8 4-1
-~CQ C
2 >H 0
> 2 43 4J
E ^"v.-^
G 3
(8 4J
*^ CQ 13
2 SH 0
> 2 43 4J
&iXl XI A
QjrH rH rH
rH O
O.rH ^<
in o vo o
co o in o
in
.
0
o
in
0*0
fa O
CQ U
3
en
n
*x^
ro
o
H
1?
> 3
(8 4J
2 M 0
> 2 43 4J
O)i-H rH rH
«* O'COV
H en ^*
CO O [*• O
O O VO O
rH
in
o
o
in
O O
ta O
m u
3
en
ro
(Tv
ro
o
rH
"Si
> 3
(8 4J
— CQ 13
2 rJ 0
> 2 43 4J
OuXI X! XI
CUrH rH rH
^ ^
rH in ^^
O O (N O
rH O t~ O
rH
in
0
O
in
fa O
CQ U
(U
rH
O4
a
18
vw
•H
a)
•o
0
o
4J
(3
A-8
-------
o
CO
0)
o
fl
n
fc
_<
0)
en
1
o
•H
to
rt
CQ
,j
C
01 4J
41 n!
H Tl
C
0 '—
'jJ X!
O •-•»
3 n
•o c
0 0
M 4J
o, ^
m
•H
a
3
Emissions
01
a
o
•H
CQ
01
X-H
o e
S 0)
X
•H -
4-
jj m
rt 2
4) 2
K •—
c-i
4)
h
3
C*
^
rH
c-
CO
*-^
ui
0
rH
VO
"w «
> 41
(S ft C
•^^ M 0
E e\\
ft ftXt XI
04
O4
1* O4
^ O LO O
CO VO CO
H rH
O
O
0\
«
og
O 0
u\ o\
*^^^~^
0 H
rH H
O
rH
04
H
® 0
4J
2
o
Vi> O
0
^
o
C
4)
01
•H
01
C
0
41 U
E C
0 -H
W
•3
0^
^
04
00
00
^-^
^
CJ
o
m
CO
® 0
jJ
E-^
OS
ro o
rg o
o
O
4)
i-i
O
01 !>i
C 0
0
4) -H vo
Ol ffl
(8 01 >H
k -H 4)
4) E >
> 41 0
<
og
H
"\
O
rH
O
*
rH
C
® ^ 0
A J->
fti-H i-l
fO
C- rH
O] «* 0
rH O4
0
01
4)
r-i
V
01 >i
C 0
o
41 -H 00
01 m
m m M
M -H 0)
4) S >
> 4) 0
<
04
H
a\
*^
rH
H
ft
ft
•*
rH
og
o
O
4)
rH
U
m S
C CJ
o
4) -H in
O! 01
nj 01 JH
r4 -H 4)
41 £ >
> 4) 0
rt
og
04
CPl
^•^
VO
o
co
co
rH
ft
ft
04
rH
04
O
01
01
i-H
O
to >-.
C 0
o
41 -H VO
Ol ffl
SH -H 4)
4) E >
> 4) 0
"*•
05
04
C\
^^v
vo
M
0
5
vo
rH
ft
ft
<*1
M
rH
(N
O
H
- "*-'
4)
a ft 5
>-•«— 0 in
4J X!
E E-^^
rtj pj j^ j2
ft fti-H r-i
0
in
C\ O CO r-
r~ m 04 o
rH VO VO
O
H
**-'
|