United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
Technology Transfer
EPA/625/5-86/020
Nitrogen Oxide Control
for Stationary
Combustion Sources
-------
EPA/625/5-86/020
July 1986
Nitrogen Oxide Control for
Stationary Combustion Sources
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
This document has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
Contents
Chapter Page
1 Introduction 1
1.1 Background 1
1.2 NOX Emission Regulations 1
1.3 Mechanisms of NOX Formation 3
1.4 General Schemes for NOX Reduction 4
1.5 NOX Emission Sources Considered in This Document 5
1.6 Organization of This Document 8
2 NOX Control Alternatives 11
2.1 Introduction 11
2.2 Precombustion Control Technologies 11
2.3 Combustion Modification Technologies 11
2.4 Postcombustion Technologies 20
3 Performance and Cost Data: Utility and Large Industrial Boilers 23
3.1 Introduction 23
3.2 Coal-Fired Boilers 23
3.3 Oil- and Gas-Fired Boilers 34
4 Performance and Cost Data: Packaged Boilers Firing Oil or Gas 39
4.1 Introduction 39
4.2 Reduced Combustion Air Preheat 39
4.3 Flue Gas Recirculation 39
4.4 Low NO* Burners 39
4.5 Other Technologies 40
5 Performance and Cost Data: Gas Turbines and Stationary
Reciprocating Engines 41
5.1 Introduction 41
5.2 Gas Turbines 41
5.3 Stationary Engines 42
6 References 47
-------
Figures
Number Page
1-1 Basic Mechanism of NOX Formation 3
1-2 General Schemes for NOX Reduction 4
1-3 Classification of Coal-Fired Boilers 6
1-4 Utility Boiler Firing Configurations for Pulverized Coal 6
1-5 Main Types of Stokers Used in Industrial Boilers 7
1-6 Front View of Circular Burner Used for Oil Combustion 7
1-7 Front View of Cell Burner Used for Gas Combustion 7
1-8 Pulverized Coal-Fired Boiler Employing Horizontally
Opposed Burners 8
1-9 Simple-Cycle Turbine 9
2-1 Typical Staged Combustion: Overfire Air 13
2-2 Arch-Fired Boilers 14
2-3 Three Temperature Reduction Methods for Boilers 14
2-4 Low NOX Burner: Staged-Air Design 15
2-5 Low NOX Burner: Staged-Fuel Design 15
2-6 Low NOX Burner: Foster Wheeler Controlled Flow/Split
Flame Burner 16
2-7 Low NOX Burner: Babcock & Wilcox Dual Register
Burner / Compartmented-Windbox System 17
2-8 Low NOX Burner: Riley Stoker Controlled Combustion
Venturi Burner 17
2-9 Low NOX Burner: Riley Stoker Directional Flame Burner 18
2-10 Low NOX Burner: EPA Distributed Mixing Burner 18
2-11 Low NOX Burner: Combustion Engineering Low NOX
Concentric Firing System 19
2-12 In-Furnace Destruction by Mitsubishi Advanced Combustion
Technology (MACT) Process 19
2-13 NOX Reduction Options for Stationary Engines 19
2-14 Classification of NOX Flue Gas Treatment Processes 20
2-15 Typical Flow Diagram for Selective (Ammonia) Catalytic
Reduction Process 21
2-16 Thermal DeNOx System — Process Flow Diagram 21
3-1 NOX Reduction by BOOS for Single Wall-Fired Boilers
Burning Coal at 120 Percent Excess Air 24
3-2 NOX Reduction by BOOS for Horizontally Opposed Wall-
Fired Boilers Burning Coal 24
3-3 NOX Reduction by BOOS for Tangential-Fired Boilers
Burning Coal 25
3-4 Typical Retrofit Arrangement for Overfire Air 25
3-5 NOX Reduction by Overfire Air for Tangential-Fired
Boilers Burning Coal 26
3-6 Costs of Retrofitting Coal-Fired Boilers for Overfire Air 26
3-7 Pilot-Scale Test Results for the CCV Burner 27
3-8 Theoretical IMOX Emissions Versus Burner Stoichiometry
for the Controlled Flow/ Split Flame Burner 27
-------
Figures (continued)
Number Page
3-9 NOX Emissions for Dual Register Versus Circular
Burners in Coal-Fired Boilers 28
3-10 Pilot-Scale Results for NOX Reduction by Low NOX
Cell Retrofit Burners 29
3-11 NOX Emissions for the Low NOX Concentric Firing
System in Coal-Fired Tangential Boilers: No Overfire Air 29
3-12 NOX Emissions for the Low IMOX Concentric Firing
System in Coal-Fired Tangential Boilers: Full Overfire Air 29
3-13 Offsetting Retrofit Costs of Controlled Flow/Split Flame Burners
with Oil Savings 30
3-14 NOX Reduction by MACT for a 600-MW Coal/Oil Boiler 31
3-15 NOX Emissions from a Slagging Combustion Pilot Plant
Firing Coal 31
3-16 NOX Emissions from a Slagging Combustion Pilot Plant
Firing Coal, Oil, and Mixtures 32
3-17 Test Results for SCR on a Coal-Fired Boiler 32
3-18 Incremental Capital Cost of SCR and Other NOX Control
Technologies for a New Tangential-Fired Boiler 33
3-19 NOX Reduction Due to Flue Gas Recirculation in a Stoker-
Fired Boiler 34
3-20 NOX Reduction by Lowering Excess Air in a Gas-Fired
Utility Boiler 34
3-21 NOX Reduction by BOOS for a Gas-Fired Utility Boiler 35
3-22 NOX Reduction by SNR for Oil-Fired Utility Boiler 36
5-1 NOX Reduction by Water or Steam Injection: Gas Turbine
Firing Natural Gas 41
5-2 NOX Reduction by Water or Steam Injection: Gas Turbine
Firing Distillate Oil 41
5-3 Performance of Selective Catalytic Reduction on a Gas
Turbine 42
5-4 NOX Reduction by Water/ Fuel Emulsion for Diesel Engine 43
5-5 NOX Reduction by Water/ Fuel Emulsion for Diesel Engine 43
5-6 NOx Reduction by Exhaust Gas Recirculation for Diesel
Engine 44
5-7 Effect of Lean-Burning Torch Ignition on Spark-Ignition
Gas-Fired Engine 44
5-8 NOX Reduction by Charge Cooling for a Spark-Ignition
Gas-Fired Engine 45
5-9 NOX Reduction by Selective Catalytic Reduction for
Spark-Ignition Engine 46
5-10 NOX Reduction by Selective Catalytic Reduction: Long-
Term Test for Spark-Ignition Engine 46
-------
Tables
Number Page
1-1 Stationary Sources of NOX 1
1-2 New Source Performance Standards for NOX Emissions from
Utility Boilers 1
1-3 Proposed New Source Performance Standards for NOX Emissions
from Industrial Boilers 2
1-4 New Source Performance Standards for Stationary Gas
Turbines 2
1-5 Typical Nitrogen Content of Selected Fuels 3
1-6 Characteristics of Stationary Reciprocating Engines 9
2-1 Combustion Modification Technologies 12
3-1 Average Reported NOX Reduction with Low Excess Air Firing
in Coal-Fired Utility Boilers 23
3-2 Average Reported NOX Reduction with Overfire Air Firing in
Tangential Coal-Fired Utility Boilers 26
3-3 Cost of Retrofitting Coal-Fired Boilers for Overfire Air 27
3-4 NOX Performance for PM Burner System in Tangential-Fired
Boilers 30
3-5 Cost of Retrofitting with PM Burners 30
3-6 Capital Cost and Cost-Effectiveness of Thermal DeNOx Process 33
3-7 Economic Evaluation of SCR for Coal-Fired Utility Power
Plants: 80 Percent NOX Removal 33
3-8 NOX Removal in Oil-Fired Industrial Boilers with Selective
Catalytic Reduction 36
3-9 Annual Operating Cost Estimates for Thermal DeNOx on a
200,000 Ib/h Oil-Fired Industrial Boiler 37
3-10 Capital and Operating Cost Data for Selective Catalytic
Reduction Systems for Oil-Fired and Gas-Fired Boilers 37
4-1 NOX Emission Performance for Low NOX Burners in Single-
Burner Oil- and Gas-Fired Boilers 39
4-2 NOX Emission Reduction for Gas-Fired Fire-Tube Boiler with
a "Fiber" Low NOX Burner 40
5-1 NOX Reduction by Exhaust Gas Recirculation: Natural
Gas Spark-Ignition Engines 44
5-2 NOX Reduction by Ignition Retard: Natural Gas-Fired
Engines 45
5-3 NOX Reduction by Selective Catalytic Reduction for
Diesel Engine 45
5-4 Capital Costs for SCR Systems for Lean-Burn Spark-
Ignition Engines 46
5-5 NOX Reduction by NCR: Results for Tests of 13 Rich-
Burn Spark-Ignition Engines 46
5-6 Cost of NCR Systems for Rich-Burn, Spark-Ignition Engines 46
-------
Acknowledgments
Several individuals contributed to the preparation and review of this document. It
was written by Thomas Beggs, JACA Corporation, Fort Washington, Pennsylvania.
Robert E. Hall, J. David Mobley, and James A. Eddinger were the U.S.
Environmental Protection Agency reviewers. Other reviewers were Gary Bisonett,
Pacific Gas & Electric Co., San Francisco; John Maulbetsch, Electric Power
Research Institute, Palo Alto, California; F. Richard Kurzynske, Gas Research
Institute, Chicago; and Howard Mason, Acurex Corporation, Mountain View,
California. The Contract Project Officer was Norman Kulujian, Center for
Environmental Research Information, U.S. Environmental Protection Agency,
Cincinnati, Ohio.
For additional information on nitrogen oxides control alternatives, contact:
Combustion Research Branch
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Chapter 1
Introduction
1.1 Background
Nitrogen dioxide is a criteria pollutant under the Clean
Air Act. Accordingly, nitrogen oxide emissions (notably
nitrogen dioxide and nitric oxide, usually referred to as
NOX) are considered a major environmental concern.
Most NOX emissions result from fossil fuel combustion.
Mobile sources of combustion, mainly motor vehicles
and aircraft, contributed about 45 percent of total 1980
NOx emissions nationwide.(1) Stationary sources, con-
tributing about 55 percent, are covered in detail in this
document.(1) Table 1-1 characterizes NOX emissions
from stationary sources.
Table 1-1. Stationary Sources of NOX (1)
Source
Utility boilers
Industrial boilers
Gas turbines
Stationary engines
Miscellaneous
Total
Approximate
No. of
Sources
2,000
6,000
Appropriate
Size Range
MW
40-1,000
10-200
10-1,000
0-0.1
Total
NOX, %
53
14
2
20
11
100
This document covers the first four sources listed in
Table 1-1, which together represent about 90 percent of
all stationary source emissions of NOX. This provides in
one publication the basic information needed by
managers and others who are involved with the major
stationary sources of IMOX emissions to make prudent
decisions for controlling these emissions to meet ap-
plicable regulations. The document provides a
technology overview for managers of power plants and
other stationary sources; state and local air pollution
agency personnel charged with monitoring the com-
pliance status of sources; and vendors and consultants
actively engaged in developing equipment, systems,
and approaches for reducing stationary source emis-
sions of NOX. The goal is not necessarily to provide all
the information necessary to make a final decision on a
means of NO* control, but rather to present the
available options with a brief assessment of the
achievable results and, where it exists, the cost of the
options.
1.2 NOX Emission Regulations
The development of control technologies for stationary
sources of NOX has, in large measure, resulted in emis-
sion regulations for new sources that are based on the
reduction achievable by these demonstrated tech-
nologies. In order to fully appreciate the later discus-
sions of these technologies it is appropriate to review
the regulations that must be met by the various source
categories.
Utility boilers, which represent the largest stationary
source sector, have been regulated at the Federal level
since 1971. The New Source Performance Standards
(NSPS) have since been revised as shown in Table 1-2,
which lists the NOX requirements of the 1978 NSPS.
Note that the highest emission limits are for coal,
reflecting the relative difficulty of NOX removal from
coal burning as discussed in detail later in this docu-
ment. Note also that emission limitations for synthetic
fuels are in general higher than for their fossil fuel
counterparts due to the higher fuel nitrogen content.
Table 1-2. New Source Performance Standards for NOv
Emissions from Utility Boilers'
,a,b
Fuel
Emission Limit,
ng/J (lb/106 Btu)
Bituminous and anthracite coal, certain 260 (0.60)
lignites, and solid fuels not elsewhere
classified'1'''
Subbituminous coal and coal-derived fuel 210 (0.50)
Distillate and residual oil 130 (0.30)
Coal-derived oil and shale oil 210 (0.50)
Natural gas 86 (0.20)
Coal-derived gas 210 (0.50)
"Source: 40 CFR, Part 60, Subpart Da.
hApplies to units for which construction commenced after September
18, 1978, and which are capable of combusting more than 73 MW
(250 x 10- Btu/h).
'Lignites meeting certain conditions of source and type of combustion
have an emission limit of 340 (0.80).
rfSolid fuels containing more than 25 weight percent coal refuse are
exempt from the NOX standard.
The relative stringency of the 1978 NSPS is largely
responsible for the recent efforts to develop so-called
low NOX burners, which constitute a major control
technology, as discussed later, and which are designed
to significantly reduce NOX emissions. Also because of
these requirements, most of the results presented in
later sections are in terms of percent reduction of
NOX. An NSPS for paniculate matter and N0xwas pro-
posed for industrial boilers in June 1984 (Table 1-3).
1
-------
Under a court order, EPA is required to promulgate
regulations by November 1986. Industrial boilers that
were placed in operation after June 1984 are subject to
the standards. Again the limitations vary with fuel type
and, in this case, within fuel type by method of combus-
tion or fuel nitrogen content. The final regulation may
differ from that summarized in Table 1-3 but is expected
to be applied retroactively to June 1984.
Table 1-3.
Proposed New Source Performance Standards
for NO, Emissions from Industrial Boilers"'6 c
Fuel
Emission Limit,
ng/J (lb/106 Btu)
Coal, pulverized 301 (0.70)
Coal, mass-feed stoker 215 (0.50)
Coal, spreader stoker and fluidized bed 258 (0.60)
and certain lignites'*
Distillate oil 43 (0.10)
Residual oil > 0.35 weight percent nitrogen 172 (0.40)
Residual oil < 0.35 weight percent nitrogen 129 (0.30)
Natural gas 43 (0.10)
Mixtures of natural gas or distillate oil 129 (0.30)
with wood or solid waste
"Source: 40 CFR, Part 60, Subpart Db.
'Applies to units for which construction commenced after June 19,
1984, and which are capable of combusting more than 29 MW (100 x
10" Btu/h).
cThe standard includes a formula for the emission limit for mixtures of
coal, oil, or natural gas with any other fuel except the special category
of lignite.
dLignites meeting certain conditions of source and type of combus-
tion have an emission limit of 340 (0.80).
The NSPS for stationary gas turbines, as shown in
Table 1-4, is more complicated than the new source
standards for boilers. NOX emission limitations under
this standard are determined by one of two formulas
depending on the size of the unit, with a stricter limit for
larger turbines. Each formula includes an allowance for
heat rate (the lower the heat rate, the greater the
allowable emissions) and a term (F) for nitrogen con-
tent. Note that F applies for fuel nitrogen contents
greater than 0.015 percent and varies as a gradually in-
creasing function of nitrogen content (N) up to N =
0.25, above which F is constant at 0.005.
The control of NOX emissions from stationary sources
may also be regulated at the Federal level by the Preven-
tion of Significant Deterioration (PSD) and Emission
Offset (EO) programs. Under PSD, siting a new source
may be contingent upon meeting specified air quality
impact limits. Therefore, if the ambient NOX concentra-
tion attributable to a source is limited, an effective limit
is placed on the actual emissions from that source.
Under EO, the NOX emissions from a source may be
voluntarily controlled to a stringent level in order to
qualify for a less stringent emission limitation for
another source in the same area. In either case, NOX
control technologies as efficient as those required by
NSPS regulations will be required.
Furthermore, state and local air pollution control agen-
cies may regulate NOX emissions in areas which are in
nonattainment of the National Ambient Air Quality
Standard (NAAQS) or to assure that the standards are
met in the future. For example, some of the air quality
management districts in the State of California are cur-
rently enforcing utility and industrial boiler standards
which are more stringent than the Federal NSPS. In ad-
dition, several states — including Texas, Florida, and
New York — are employing an industrial boiler regula-
tion for existing sources that is equivalent to the propos-
ed industrial boiler NSPS.
Table 1-4. New Source Performance Standards for Station-
ary Gas Turbines"
Type Turbine
Allowable Emissions
Vol. % NOX @ 15% O2, Dry6
Electric utility units with a heat
input at peak load of > 107.2 GJ (100
x 10* Btu) per hour
Units with a heat input at peak
load of > 10.7 GJ (10 x 10" Btu) per
hour but < 107.2 GJ (100 x 10' Btu)
per hour, and units with a base load
at ISO conditions of 30 MW or lessc
0.0075 14.4 + F (1)
0.0150 14.4 + F (2)
"Source: 40 CFR, Part 60, Subpart GG. Refer to the source for several
exceptions to the standard.
6Y = manufacturer's rated heat rate at manufacturer's rated load
(equation 1) or rated peak load (equation 2) in KJ per watt hour, or ac-
tual measured heat rate based on lower heating value of fuel as
measured at actual peak load for the facility. Y must be no greater
than 14.4. F varies with fuel nitrogen content (by weight) as follows:
for N< 0.015, F = 0; 0.015 < N < 0.1, F = 0.04 (N); 0.1 0.25, F = 0.005.
CISO conditions: 288°K, 60% relative humidity, and 101.3 kPa
pressure.
With this brief introduction as a background, the re-
mainder of Chapter 1 presents several general topics
designed to establish a background for later sections.
First, the mechanisms of NOX formation are briefly
reviewed. Understanding these mechanisms is essential
to gaining a full appreciation for the development of
NOX control technologies, most of which are aimed at
preventing NOX formation during combustion. Many of
the technologies thus developed are designed to
primarily reduce NOX formation from the specific
mechanism associated with a particular combination of
fuel and combustion conditions.
Next, the general schemes for NOX control are discuss-
ed. An overview is presented of the three classes of
control; that is, control before, during, and after com-
bustion. As explained later, control for major sources,
such as coal-fired utility boilers, is limited by practical
considerations and often confined to during-
combustion technologies. This is because before-
combustion techniques may be unavailable or limited,
and after-combustion techniques may be cumbersome
and cost-ineffective unless also required for control of
other pollutants.
-------
Finally, a subsection is devoted to describing the
sources covered in this manual, namely, utility boilers,
industrial boilers, gas turbines, and stationary engines.
This information may be helpful to at least three groups
of readers: (1) field personnel or managers interested in
following NOX control developments for other sources
to ascertain if technologies can eventually be transfer-
red to their source; (2) state and local agency personnel
who may not be totally familiar with the sources that are
being regulated; and (3) consultants and equipment
vendors who supply engineering solutions to the prob-
lems of NOX control. Therefore, the brief discussion of
sources is provided for those readers who may feel that
this background will be necessary to their understand-
ing of control technology applications and results
presented in later sections.
1.3 Mechanisms of NOX Formation
In stationary source combustion approximately 95 per-
cent of the NOX formed is nitric oxide (NO), which can
oxidize in the atmosphere to form nitrogen dioxide
(N02), a criteria pollutant. The formation of NOX during
combustion of fossil fuels occurs by two mechanisms:
oxidation of atmospheric nitrogen present in the com-
bustion air at elevated temperatures (usually called ther-
mal NOX), and oxidation of a portion of the bound
nitrogen in the fuel (fuel NOX). The latter is less depen-
dent on temperature than on fuel nitrogen content, fuel
properties, and the stoichiometric conditions present at
combustion. The two mechanisms are shown
schematically in Figure 1-1. In general, fuel NOX is
responsible for the bulk of NOX formation in the com-
bustion of coal and certain fuel oils with high nitrogen
content. For combustion of gas or low nitrogen fuel
oils, thermal NOX is the dominant mechanism.
Two important reactions in thermal NOX formation are:
Figure 1-1. Basic Mechanism of NOX Formation.
N2 + 0 = NO + N
N + 02 = NO + 0
(1-1)
(1-2)
Note that Reaction 1-1, which is highly temperature
dependent, provides the atomic nitrogen (N) necessary
for Reaction 1-2. Both reactions, however, are equally
important in terms of the amount of NO formed. Note
further that the reverse reactions are not favored by the
presence of molecular oxygen; therefore, in the oxidiz-
ing environment that normally prevails downstream
from the actual combustion (due to excess air for com-
bustion), the NO that has been formed is essentially fix-
ed. Finally, it has been noted that in regions of the com-
bustion chamber in which the temperature is less than
1,200C (2,200F), formation of NO is not significant.
The kinetics of fuel NOX formation are not as well
understood as thermal NOX. The most significant fac-
tors in fuel NOX formation are nitrogen content of the
High Temperature
Combustion
Process
fuel and the degree to which the fuel is mixed with air
during the early stages of combustion when bound
nitrogen is liberated from the fuel. Table 1-5 shows the
nitrogen content of selected fuels, including several
nontraditional fuels, some of which exhibit relatively
high nitrogen contents. The molecular nitrogen in
natural gas responds in the same way as nitrogen in the
combustion air. Because it is not bound it does not
behave as fuel nitrogen.
Table 1-5. Typical Nitrogen Content of Selected Fuels
Fuel
Nitrogen
Content,
Weight %
Coal, anthracite, Pennsylvania 0.6-1.4
Coal, high-volatile "B," Ohio 1.4
Coal, subbituminous "B," Wyoming 1.0
Lignite, North Dakota 1.7
Fuel oil, No. 1 0.003
Fuel oil, No. 2 0.006
Fuel oil. No. 4 0.24
Fuel oil. No. 6, low sulfur 0.28-0.5
Tar sands oil 0.07
Shale oil 0.01
Coal-derived synthetic oil
SRC-II heavy distillate 1.03
H-Coal 0.57
Natural gas, mid-continent 3.2*
Natural gas, Pennsylvania 1.1*
Coke oven gas 3.4
Crude oil
Kern Co., California 0.5-0.83
Saudi Arabia, light 0.098
*Molecular nitrogen, N2.
-------
As the fuel enters the flame zone, it is pyrolyzed into
small, reactive, nitrogen-containing molecules which
react with oxygen to form NO. If little oxygen is present,
as in the fuel-rich zone of staged combustion (a com-
bustion modification NOX reduction technology), the
nitrogen-containing fuel fragments encounter and react
with each other, and convert the nitrogen to molecular
nitrogen (N2).
This theory then serves as the basis for many of the
combustion modification technologies for NOX reduc-
tion that are discussed in this document.
1.4 General Schemes for NOX Reduction
Because NOX formation results primarily from combus-
tion, three general schemes for NOX reduction suggest
themselves immediately:
• Reduction before combustion by reducing potential
for formation
• Reduction during combustion by modification of the
combustion process itself
• Reduction after combustion by some means of flue
gas treatment.
These are shown schematically in Figure 1-2; the con-
trol technologies shown are described in detail later.
Precombustion schemes for reducing NOX center on
switching to fuels with a lower nitrogen content or a
lower flame temperature. The choices are severely
limited, however. The fuel choice is usually dictated by
economic factors that transcend the economics of NOX
control. Furthermore, because nitrogen in solid and li-
quid fuels is chemically bound to organic chemical con-
stituents, it is not efficiently removed. Therefore, no
technology similar to physical cleaning of coal for sulfur
removal is practical.
Emulsifying diesel oil with water is essentially a combus-
tion modification technique in that the emulsification
itself lowers the flame temperature and aids atomiza-
tion. However, because it involves pretreatment of fuel,
this technology is discussed further in Chapter 2 under
precombustion control technologies.
NOX reduction during combustion, usually referred to
simply as combustion modification, has been employed
since the early 1970s to effect moderate NOX emission
reductions. It is currently the principal NOX reduction
scheme for moderate control; a growing list of
technologies in this category is being studied or applied.
These technologies, discussed in detail in Chapter 2,
suppress thermal or fuel NOX by modifying the condi-
tions for combustion, namely, stoichiornetry,
temperature, and residence time. These modifications
can be achieved by modification of the burner itself or
externally through modification of the air or fuel flow to
the combustion chamber. Research, as well as pilot-
scale and full-scale development of these technologies,
is being carried out primarily by burner and boiler manu-
facturers, the Environmental Protection Agency (EPA),
the Electric Power Research Institute (EPRI), and the
Gas Research Institute (GRI).
Postcombustion schemes all involve some type of treat-
ment of the flue gas and are normally classified as dry or
wet processes. Some — mostly wet processes — are
also designed for simultaneous removal of sulfur diox-
ide, primarily from coal-fired boilers. Many of the flue
gas treatment processes have been developed in Japan,
where NOX emission limits are generally stricter than in
the United States and where sulfur oxides (SOX)
removal processes have been widely applied.
Dry flue gas treatment processes usually consist of
either reduction of NOX to nitrogen by reducing agents
or adsorption onto solids. The reduction processes may
be' catalytic or noncatalytic. Noncatalytic reduction is
typified by the Exxon Thermal DeNOx process in which
ammonia is used to selectively reduce NOX at
temperatures of 925 to 980C (1,700 to 1,800F) and by
the EPRI urea injection process. These reduction pro-
cesses may emit residual reducing agents (principally
ammonia) and their by-products such as ammonium
sulfate. Adsorption processes add expense due to the
solid waste generated.
Wet flue gas treatment processes offer perhaps only
one advantage over dry processes — simultaneous
removal of sulfur oxides. These processes may be at-
tractive for applications requiring stringent control of
both pollutants. These processes are currently in the
development stage and have not been commercially ap-
plied in the United States.
Figure 1-2. General Schemes for NOX Reduction.
PRECOMBUSTION TECHNIQUES
• Switch Fuel
• Emulsify Fuel with Water
• Fuel Denitrification
COMBUSTION MODIFICATION TECHNIQUES
• Stoichiornetry
• Temperature
• Residence Time
FUEL
Combustion
Chamber
EXHAUST
OR FLUE GAS
POSTCOMBUSTION TECHNIQUES
• Flue Gas Treatment, NOX Only
• Simultaneous SOx/NOx Treatment
4
-------
To summarize, although there are a myriad of theoreti-
cal options currently available for reducing NOX emis-
sions from combustion sources, the actual selection for
a specific case will come down to a few very practical
choices. Until a few years ago nearly all NOX reduction
for stationary sources was brought about by simple
combustion modification. Recently, however, catalytic
postcombustion technologies have been employed in
increasing numbers, particularly for stationary engines.
The normal sequence has usually been to reduce NOX
emissions from stationary sources with the simplest and
most economical combustion modification that will
achieve the necessary emission reduction. As further
reduction is required, progressively more efficient (and
more difficult to implement and hence more costly)
combustion modifications are employed. In the case
where the most efficient combustion modification still
does not bring about the required reduction, postcom-
bustion technologies are considered, often in addition
to one or more combustion modifications.
1.5 NOX Emission Sources Considered in
This Document
This document considers the following stationary
sources of NOX emissions:
• Utility boilers, which account for approximately 53
percent of all stationary source emissions
• Industrial boilers, accounting for about 14 percent
• Gas turbines and stationary engines (gas and diesel),
which together represent approximately 22 percent
of stationary source emissions.
Therefore, this document covers technologies that ap-
ply to about 90 percent of the stationary sources of NOX
emissions. Not covered in this document are commer-
cial boilers, residential heaters, enhanced oil recovery
steam generators, industrial process heaters, and
miscellaneous combustion and noncombustion sources
which account for most of the remaining 10 percent of
emissions.
Each major source will now be described in detail suffi-
cient to provide the necessary background for
understanding the later discussions of control
alternatives.
7.5.7 Utility Boilers
Approximately 80 percent of fossil fuel steam genera-
tion is from coal firing.(2) The remainder is nearly evenly
divided between natural gas and oil (predominantly
residual fuel oil such as No. 6). In addition, the uncon-
trolled NOX emission factors for coal-fired boilers are
approximately twice those for natural gas-fired and oil-
fired units (on a Btu basis).
All fossil fuel-fired utility boilers generate steam by
transferring the heat from combustion of the fuel to
water. The steam, in turn, produces electricity by ex-
panding through a turbine. The steam generated may
be saturated or superheated and is essentially always
confined inside the tubes with the outside of the tubes
exposed in part to combustion gases. This is the water-
tube arrangement as opposed to the fire-tube arrange-
ment found in low capacity industrial boilers.
Coal is fired either in a bed, or stoker, or in a state of
suspension. Stoker-fired units are rarely found in large
utility boilers and in terms of nationwide NOX emissions
are of much less importance than suspension-fired
boilers.
Suspension-fired boilers are categorized (Figure 1-3) as
either cyclone-fired or pulverized coal- (PC) fired
boilers. Cyclone boilers fire coal ground to about a
4-mesh size and carried into a cylindrical combustion
chamber by primary air in a cyclonic flow pattern. Small
particles burn in suspension while larger particles
adhere to the molten slag on the furnace walls where
they are burned with the addition of secondary combus-
tion air.
In contrast, pulverized coal-fired boilers fire coal of par-
ticle sizes on the order of 70 percent passing 200-mesh.
The coal particles burn in a state of suspension in the
combustion chamber.
For our purposes, these boilers are categorized in accor-
dance with the position of the burners in the combus-
tion chamber (Figure 1-4). Wall-fired boilers have
burners mounted horizontally either in a single wall
(front or rear) or in two opposite walls horizontally op-
posed. Turbo-fired boilers also employ burners in
horizontally opposed pairs but the burners are inclined
downward to induce a turbulent flow. Tangential- or
corner-fired boilers employ burners in groups of four,
each firing horizontally from a corner of the furnace. At
each elevation of four burners, each burner is aimed at
the tangent to an imaginary circle in the center of the
furnace, which sets up a vertically oriented vortex
throughout the combustion zone. Arch- or vertical-fired
boilers employ burners that fire vertically downward in-
to the combustion chamber and may exhibit a turbulent
flow pattern if the horizontally entering secondary air
flows are so designed. This type of boiler is effectively
obsolete but may find application again in the future for
combustion of difficult-to-ignite fuels, such as chars
from coal conversion processes, which are low in
volatile matter.
The method of ash removal is also important to NOX
emissions in suspension-fired boilers and is another way
to classify such boilers. Dry-bottom boilers burn coal
with a high ash fusion temperature and therefore are
designed for a dry ash removal system. On the other
hand, wet-bottom, or slag-tap, boilers remove molten
ash (slag) resulting from combustion of coal with lower
ash fusion temperatures. Wet-bottom boilers exhibit
higher uncontrolled NOX emission rates than do dry-
bottom boilers due to higher combustion temperatures.
Combustion in these boilers is very intense; thermal
NOX formation is increased by the higher temperature
while fuel NOX formation is enhanced by the more tur-
bulent mixing.
-------
Figure 1-3. Classification of Coal-Fired Boilers.
Figure 1-4.
Utility Boiler Firing Configurations for
Pulverized Coal .
Wall-Fired,
Elevation View
Horizontally Opposed,
Wall-Fired, Elevation View
Turbo-Fired,
Elevation View
Arch-Fired,
Elevation View
Tangential Firing,
Plan View
Stoker-type boilers are classified according to the
method of feeding coal shown in Figure 1-5.
The four main burner configurations in coal firing —
wall-fired (single), horizontally opposed-fired, turbo-
fired, and tangential-fired — also predominate for oil-
and gas-fired utility boilers, although the burners
themselves are different.
The most frequently used burners for oil or gas combus-
tion are the circular burner and the cell burner. These
burners are also used for pulverized coal or for firing any
of these fuels in combination. The maximum firing
capacity of the individual circular burner is on the order
of 165 million Btu/h; cell burners can fire up to about
495 million Btu/h.
The chief difference between the two types of burners
is the fuel injection location. In the circular burner, fuel
is introduced at one location — the center of the circle.
In the cell burner, fuel is introduced through multiple
spuds arranged annularly around the center of the
burner. Each spud is a pipe with a pattern of holes at the
end to discharge the fuel. Figure 1-6 depicts a circular
burner and Figure 1-7 shows a cell burner. (4)
Figure 1-8 shows a "typical" utility boiler — in this case
a pulverized coal-fired boiler employing horizontally op-
posed burners.
-------
Figure 1-5. Main Types of Stokers Used in Industrial
Boilers (3).
Figure 1-6. Front View of Circular Burner Used for Oil
Combustion.
(a) Underfeed
.vy» •.••'•'.• Fuej .';•;•:'r:V
i _ i _ :: _ — -/,/
i ________ *
Air
Ash Pit
(b) Crossfeed Spreader
crc*
*V I « i_ i-t'^
Air
Ash Pit
Impeller
Oil Atomizer
Residue Figure 1-7. Front View of Cell Burner Used for Gas
Combustion.
Gas Spud
Impeller
(c) Overfeed Spreader
7.5,2 Industrial Boilers
Industrial boilers, used primarily to generate steam for
heating or process uses, also fire coal, oil, gas, or waste
fuels. Most coal-fired industrial boilers are water-tube
boilers; fire-tube boilers are subject to ash plugs and
other operational problems. Large coal-fired units
(steam flow rate greater than 350,000 Ib/h) generally
fire pulverized coal and are very similar to comparably
sized utility boilers. Smaller coal-fired industrial boilers
(less than 100,000 Ib/h of steam) are normally stoker
fired, and a mixture is found in the middle size range.
Because of the vast number of smaller boilers, stoker
firing accounts for most of the coal consumption by in-
dustrial boilers.
The main stoker types in use are illustrated in Figure 1-5.
Spreader stokers propel coal into the combustion
chamber; a portion of the coal actually burns in suspen-
sion while the remainder burns on a grate. Grates may
be of the stationary, dumping, or travelling variety.
Because they do not allow for continuous removal of
ash, stationary grates must be zoned to allow for
periodic removal. That is, not all zones are firing at any
one time; this allows ash removal from one zone at a
time. The underfeed stoker employs a ram to feed coal
upward through the burning bed in the same direction
as the flow of combustion air to the tuyeres. Multiple
retort underfeed stokers are used for coking coals with
high ash fusion temperatures. In contrast, the overfeed
stoker supplies coal to the bed from above. The latter is
normally used with travelling or pulsating grates and is
amenable to burning nearly all types of coal.
Oil- and gas-fired industrial boilers are either water-tube
or fire-tube boilers. The former employ burner flames on
the outside of the tubes with boiler water flowing
through the tubes, while the latter reverse this design.
Fire-tube boilers tend most often to be small, packaged
units, whereas water-tube boilers can be either packag-
ed units or large, field-erected units. Packaged units are
available up to a capacity of 350,000 Ib/h of steam.
-------
Figure 1-8. Pulverized Coal-Fired Boiler Employing
Horizontally Opposed Burners (5) •
To
Precipitator
Burners
Underfire
Air Ports
Burners
Underfire
Air Ports
1.5.3 Gas Turbines
Gas turbines are used in industry for compressing gas or
pumping liquids, and in utilities for generating electric
power. Gas turbines fire natural gas or fuel oil; the oil is
normally a distillate such as No. 2 fuel oil. The principle
of operation is simply to expand the products of com-
bustion through a turbine to generate power.
The combustion chamber of a gas turbine typically con-
sists of two zones: the primary zone, where essentially
all of the combustion takes place under low excess air
conditions; and the secondary zone, where secondary
air is introduced in quantity sufficient to cool the pro-
ducts of combustion to a temperature compatible with
the turbine materials of construction.
There are basically three configurations for gas tur-
bines. The so-called simple-cycle turbine is the basic
type of gas turbine (Figure 1-9). Compressed air and
fuel are directed to the combustion chamber, where
combustion and dilution of exhaust gas take place. The
exhaust gas is then expanded through the turbine which
provides energy not only for the load (pump, generator,
etc.) but also for the combustion air compressor.
The first refinement of the simple-cycle turbine is the
regenerative-cycle turbine. This engine uses a simple-
cycle turbine but also includes a recuperative heat ex-
changer to preheat combustion air with the turbine ex-
haust gas. This engine is more energy-efficient than the
simple-cycle gas turbine.
The second refinement is the combined-cycle turbine.
The heart of this type of turbine is a simple-cycle tur-
bine, but it also includes a waste heat boiler to produce
steam from heat exchange with the turbine exhaust
gas.
1.5.4 Stationary Reciprocating Engines
This category of NOX emission sources includes
compression-ignition engines and spark-ignition
engines (see Table 1-6). Compression-ignition engines
are normally fired with diesel oil or with a combination
of oil and natural gas (generally the oil is injected into
the cylinder only to initiate combustion). The latter are
referred to as dual-fuel engines. Spark-ignition engines
are typically fired with natural gas. Stationary
reciprocating engines have wide-ranging uses; perhaps
their most important application, involving units of
several thousand horsepower, is to drive large com-
pressors in distribution of natural gas.
Compression-ignition engines are normally four-cycle
engines of large bore that operate on this familiar cycle:
admission of air and fuel; compression and ignition;
expansion; and exhaust. Thermodynamically, com-
pression-ignition engines follow a constant pressure, or
diesel, cycle. Compression ratios are relatively high (on
the order of 20 to 1) and compression pressures range
up to several hundred psi. At such high pressures, com-
bustion is spontaneous; therefore, no ignition sources
(such as sparkplugs) are required.
Spark-ignition engines follow the spark-ignition, or
Otto, cycle of constant volume combustion. These
engines feature either four-cycle or two-cycle opera-
tion. In two-cycle operation, the air-fuel mixture is com-
pressed outside the cylinder and expels the exhaust
upon entering the cylinder. In two-cycle engines, a
scavenging arrangement is normally used in order to
hasten the exhaust of combustion products and
minimize the escape of the air-fuel mixture with the ex-
haust. Immediately after the bulk of the exhaust gas
leaves the cylinder, a jet of scavenger air enters the
cylinder. The scavenger air is deflected usually into
either a helical or cyclic pattern in order to force the re-
maining exhaust gas out the exhaust port. The cylinder
is then relatively exhaust-free when the air-fuel mixture
is introduced, which occurs nearly simultaneously with
scavenging. The pressure for scavenging can be sup-
plied by the crankcase (common in smaller engines) or
by an external blower that is driven by the engine (com-
mon in larger engines).
1.6 Organization of This Document
The remaining chapters are devoted to a more in-depth
examination of NOX control technologies, specifically as
they apply to, and have been refined for, the sources of
NOX discussed earlier in this chapter.
8
-------
figure 1-9. Simple-Cycle Turbine.
AIR
Combustion
Chamber
A
Shaft
Turbine
External
Load
Table 1-6. Characteristics of Stationary Reciprocating Engines
Spark-Ignition
Compression-Ignition
Typical fuel
Thermodynamic cycle
Compression ratio
Maximum cylinder pressure
Operation
Scavenging
Natural gas
Otto (constant volume)
6:1 to 12:1
Over 2,000 kPa
2-cycle or 4-cycle
Usually required for 2-cycle;
crankcase pressure or
external blower used
Diesel oil or dual fuel (oil and
natural gas)
Diesel (constant pressure)
11.5:1 to 22:1
Over 4,800 kPa
Usually 4-cycle
Not required for 4-cycle; 2-cycle
engines often employ blower-
scavenger
Chapter 2 presents a detailed discussion of all the NOX
control alternatives that are generally commercially
available and for which performance or cost data are
available. Most pilot-plant technologies and untested
conceptual designs are not included here; rather, those
technologies offering a reasonable promise for success
are presented. Chapters 3 through 5 present actual NOX
reduction data and system cost data for application of
these technologies to utility and large industrial boilers,
packaged boilers firing oil or gas, gas turbines, and sta-
tionary reciprocating engines. Both retrofit and new ap-
plications are considered where quality data on each are
available. Note that no new cost data were developed
for this document; rather, cost data appearing in the
literature are merely reported as they were found in the
references. Therefore, this report makes no representa-
tion as to the accuracy of such cost data.
-------
-------
Chapter 2
NOX Control Alternatives
2.1 Introduction
Three conceptual schemes for reducing NOX emissions
from stationary sources of combustion were introduced
earlier: precombustion, or the reduction of potential
formation; combustion modification; and postcombus-
tion removal of NOX from the flue gas. The numerous
technologies that are commercially available for NOX
reduction are now described in some detail, grouped in-
to the above three categories. A more complete discus-
sion of these technologies including results achievable,
associated costs, and deleterious impacts, is presented
in Chapters 3 through 5.
2.2 Precombustion Control
Technologies
Switching to a lower nitrogen fuel is often economically
unattractive, and removal of fuel nitrogen is usually not
cost-effective solely for NOX control. Furthermore, with
many fuels such as natural gas and low-nitrogen fuel oil,
the majority of NOX emissions are from thermal NOX
formation. Therefore, the choices in this category are
quite limited. The process of removing sulfur from fuel
oil by hydrodesulfurization has the side benefit of
nitrogen removal but is very expensive in comparison
with combustion modification techniques that produce
a comparable NOX reduction.
2.2.1 Change of Fuel
Clearly, this is the trivial case. Where it is economically
attractive to change to a lower nitrogen fuel, changes
for economic reasons, not for NOX reduction incentives,
have been or are made. Cost considerations are para-
mount and merely switching to a lower nitrogen fuel
may not effect a large enough reduction to meet NOX
emission regulations, which in any case are fuel-
specific. Therefore, combustion or postcombustion
NOX reduction techniques may be necessary.
2.2.2 Fuel Oil/ Water Emulsions
This control technology could arguably be classified as
a combustion modification technique because it alters
the conditions for combustion; however, it is discussed
here because it involves mixing of fuel prior to combus-
tion. The technique has been tested on stationary
engines firing diesel fuel. Tests of this technique have
been typically performed on diesel engines in the 100 to
300 hp size range and have employed fuel mixtures from
7 to 45 percent water. Emulsions have been delivered to
the fuel injection system in a number of ways: a simple
mixer and pump arrangement; a low-energy shear fluid
mixer into which the oil-water mixture is pumped; or a
high-energy emulsor in which the mixture is pumped
through an orifice at a high (on the order of 2,000 psi)
pressure drop.
The primary mechanism for NOX reduction provided by
fuel oil/water emulsions is lowering of the flame
temperature. Heat transfer from the flame to the flue
gas is enhanced by the presence of water vapor which
serves to increase the overall specific heat capacity of
the products of combustion, thus lowering the flame
temperature. Another probable mechanism taking place
is the limiting of the NO formation reaction resulting
from the water vapor's dilution of the oxygen in the
flame zone. Finally, the rapid vaporization of water in an
oil droplet is thought to increase the atomization of fuel,
thus enhancing combustion efficiency.
2.3 Combustion Modification
Technologies
The majority of NOX control technologies for combus-
tion sources involve modifying the parameters of com-
bustion. All of the techniques are aimed at achieving
one or more of the following goals: reducing the
available oxygen at critical stages of combustion; lower-
ing the peak flame temperature; and reducing the
residence time during which oxidation of nitrogen takes
place. Table 2-1 summarizes the technologies covered
in this subsection and the mechanisms these
technologies rely on.
Where possible, the technologies to be considered
under this heading have been grouped by similarity and
ordered by degree of complexity. Retrofit technologies
are presented first, in order of increasing complexity,
followed by technologies for new combustion sources.
In terms of meeting compliance NOX emissions levels
both on a source-by-source basis and in a general sense
for fossil fuel-fired boilers as a whole, a logical progres-
sion of control technologies can be employed in order of
increasing removal efficiency and cost. In this manner
the appropriate and most cost-effective technology can
be found most expediently. For example, low excess air
would be the first technology employed; if the NOX
emission reduction achieved at the limit of low excess
air is not sufficient, any of various types of staged com-
11
-------
Table 2-1. Combustion Modification Technologies
Technology
Low excess air
Staged combustion
Lowers Flame
Temperature
Yes
Reduces
Available
Oxygen
Yes
Yes
Shortens
Residence
Time
-
Other
-
BOOS
Biased firing
Overfire air
Arch firing
Reducing combustion air
preheat
Water injection
Exhaust gas recirculation
Low NOX burners
Staged air
Staged fuel
Variations
In-furnace destruction
(reburning)
Lean-burn, torch ignition
Turbocharging
Charge cooling
(refrigeration)
Retardation (ignition
or injection)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Reduction
of NO to N2
Yes
bustion typically would be applied. The logical progres-
sion would proceed in this manner until the desired
reduction is achieved. As mentioned earlier, this pro-
gression applies for individual sources and also parallels
the development of these technologies to meet ever-
increasingly stringent NOX emission levels. For retrofit
applications, the feasibility and cost effectiveness are
dependent on space requirements and operational side
effects resulting from modified combustion conditions.
Some control technologies may not be practiced
because of cost or effect on boiler efficiency. In any
event, the actual order in which technologies are im-
plemented depends on the required emission reduction.
2.3.1 Low Excess Air
Operating burners with low excess air is perhaps the
simplest NOX reduction technology to implement on a
retrofit basis; no capital equipment is required and com-
bustion chamber modifications are normally un-
necessary. However, the NOX reductions achievable are
modest and may not, in specific retrofit situations, be
sufficient to comply with NOX emission regulations. The
degree of control is constrained by the onset of carbon
monoxide (CO) emissions (generally held to 50 to 100
ppm) or plume opacity at low excess air. Other factors
such as safety considerations (e.g., minimum air flow
requirements) may also be considered. The technology
is applicable to coal, oil, and gas firing in boilers of all
sizes.
Most new coal-fired boilers are designed for low excess
air firing. In order to meet New Source Performance
Standards, however, they also usually are equipped
with low NOX burners or other combustion modification
NO* reduction technique. Oil-fired and gas-fired boilers
in many instances can be operated at excess air levels of
5 percent and lower. However, coal-fired boilers nor-
mally require a greater excess air level in order to ensure
essentially complete carbon burnout and to minimize
emissions of carbon monoxide.
The low excess air method accomplishes NOX reduction
of both thermal and fuel NOX. Reduced availability of
oxygen suppresses the formation of NO from nitrogen
in the fuel and in the combustion air.
2.3.2 Staged Combust/on
Staged combustion is actually a whole family of
technologies. The most frequently employed for retrofit
of large industrial and utility boilers are burners out of
service (BOOS) and biased firing. For new units, over-
fire air is most frequently applied. Arch firing, a boiler
configuration inherently low in NOX formation, has also
received attention lately as a form of staged combus-
tion. Staged combustion can also be achieved internally
by burner modification. A key characteristic of this
technology is stretching of the combustion zone, which
may cause flame impingement on side walls or
superheater tubes and affect steam balance and conse-
quently unit efficiency. When staging lowers
stoichiometric air ratios below 1.0, there is the potential
for increased tubewall corrosion rates.
2.3.2.1 Burners Out of Service (BOOS)
This technology is generally applicable to wall-fired utili-
ty and large industrial boilers. BOOS is normally im-
12
-------
plemented on existing units by employing an upper row
of burners — on one wall or on both walls of opposed-
fired units — as sources of secondary combustion air
only, without introduction of fuel through these
burners. Also, selected burners can be removed from
service other than the top row if the BOOS pattern is
balanced. The effect is to provide so-called overfire air
above the active burners which comprise the primary
combustion zone. This zone is maintained under fuel-
rich conditions by introducing air to these burners at
lower than stoichiometric rates.
Operation of the lower furnace zone with a substoichio-
metric amount of air (fuel-rich conditions) lowers the
conversion of fuel nitrogen to nitrogen oxide. Subse-
quent cooling and dilution of the combustion gas with
the secondary air from the upper (out of service)
burners reduces peak flame temperature and thus
minimizes the formation of NO by thermal fixation.
BOOS could presumably be employed in new installa-
tions but finds the vast majority of its applications as a
retrofit technology. It is applicable to suspension-fired
coal boilers as well as boilers firing oil or gas.
2.3.2.2 Biased Firing
Biased firing is a retrofit technology generally only ap-
plicable to oil-fired and gas-fired utility and large in-
dustrial boilers. It is roughly equal to BOOS in ease of
implementation; no new equipment is usually
necessary, nor are any major boiler modifications.
In biased firing, the overall combustion chamber
stoichiometry is preserved (which may in all probability
involve low excess air) while air and fuel flows to in-
dividual burners are varied. There is no set or generally
accepted pattern to variation of the burner conditions,
as many combinations may yield the desired effect. The
goal is to create fuel-rich and fuel-lean regions in the
combustion chamber with an effect similar to that of
overfire air or BOOS: The fuel-rich regions generate
relatively low thermal and fuel NOX. The best combina-
tion is usually found by experimenting with various
patterns.
2.3.2.3 Overfire Air
The third widely used type of staged combustion is
overfire air, which is applicable to coal-, oil-, and gas-
fired utility boilers and large industrial boilers. Overfire
air is usually used in new boiler designs. Unlike BOOS
and biased firing, overfire air requires modification of
the combustion chamber. Specifically, air ports above
most or all of the rows of burners must be added to pro-
vide the secondary combustion air above the burners.
The result is similar to BOOS operation: Fuel-rich
burners reduce fuel and thermal ISIOX formation, and in-
terstage cooling by boiler tubes reduces peak flame
temperatures which also suppresses thermal NOX. Over-
fire air operation is shown in Figure 2-1.
Figure 2-1. Typical Staged Combustion: Overfire Air (6).
Furnace Outlet
Overfire Air •
Main Burner
/ COMBUSTION \
! COMPLETION 1
v ZONE /
/ \
' MAIN \
/ BURNER \
. COMBUSTION '
\ ZONE I
\ /
2.3.2.4 Arch Firing
This boiler configuration is one of the earliest designs
for burning pulverized coal in utility boilers and is receiv-
ing some attention lately because it is an inherently low
IMOx formation combustion technology. Burners are
mounted such that they fire vertically downward in the
combustion chamber; secondary air is injected farther
down the vertical walls of the boiler. The result is a
J-shaped flame in which the combustion is staged and
therefore NOX emissions are low. Figure 2-2 shows
schematic views of three different arch-fired boilers.
Because this technology involves an entire combustion
chamber configuration, its applicability on a retrofit
basis is limited, especially when compared to the
relative ease of retrofitting the furnace with low NOX
burners. Its best application may be for new units,
although there are two basic drawbacks: The cost of an
arch-fired boiler is generally significantly higher than a
comparable wall-fired boiler; and the technology has
not been applied to new boiler designs for years and
there may be a natural reluctance on the part of utilities
to employ it.
2.3.3 Temperature Reduction Technologies
Several NO* reduction technologies employ some
method of reducing peak flame temperature to reduce
thermal NOX formation. These include reducing the
combustion air preheat, injecting water, and recir-
culating exhaust gas. These technologies are presented
schematically in Figure 2-3. Temperature reduction
technologies are generally applicable to oil- and gas-
fired boilers and engines; they are not effective for
sources firing coal or high nitrogen oil because thermal
NOX formation is usually the less important mechanism
in such units.
13
-------
Figure 2-2. Arch-Fired Boilers (7).
125 MW
2.3.3.1 Reducing Combustion Air Preheat
This technology is applicable primarily to utility and in-
dustrial boilers burning oil or gas. The technique is
merely to lower the temperature of the incoming com-
bustion air, usually by providing a bypass around the air
preheater for a portion of the combustion air. The result
is that the peak flame temperature is correspondingly
reduced, thus reducing thermal NOX formation. In utility
and some large industrial boilers, economizers can be
used to recover some of the thermal efficiency that is
sacrificed in reducing the combustion air preheat.
However, the loss of thermal efficiency can represent a
severe economic penalty. This technology, therefore, is
often considered for interim control.
2.3.3.2 Water Injection
Water injection technology involves a modest modifica-
tion to allow water to be injected into the combustion
air stream. Normally, nozzles are mounted in the wind-
box in a manner that permits vaporization of the water
before it enters the combustion chamber. The vaporiza-
tion removes some of the heat from the combustion
chamber, thus lowering peak flame temperature. It is
important to design the system so that vaporization of
the water occurs before the combustion air reaches the
combustion chamber to avoid corrosion in this area. As
it is a temperature reduction technology, water injection
may also carry a significant energy penalty and thus
should perhaps be viewed as an interim measure for
utility boilers.
2.3.3.3 Exhaust Gas Recirculation
Exhaust gas recirculation (EGR), or flue gas recircula-
tion (FGR), is a technique similar to that used for
275 MW
Figure 2-3. Three Temperature Reduction Methods for Boilers
To Stack
Combustion
Air
Preheater
FLUE GAS
flue gas
PREHEATED COMBUSTION AIR
Bypass a portion
of eombtisttop air around heater
Boiler
WATER
Inject water
14
-------
automobile engines. In fact its greatest applicability is
with stationary engines and turbines. It is applicable on
a new or retrofit basis for boilers, process heaters, and
engines. A portion of exhaust gas or flue gas is recycled
to a point where it joins, and therefore dilutes, the inlet
combustion air flow. This dilution serves to lower peak
flame temperature, thus reducing thermal NOX forma-
tion. The local oxygen level in the flame is also reduced,
which helps to suppress NOX formation. This technique
is normally not applicable for coal-fired boilers (where it
is normally called "flue gas recirculation") because the
greater portion of NOX emissions are from fuel NOX in
coal combustion.
2.3.4 Low NOX Burners
Low NOX burners have been and continue to be
developed for coal-, oil-, and gas-fired boilers. These
burners are applicable on a new or retrofit basis and can
be used for utility boilers, large industrial boilers, or
small packaged boilers. Low NOX burners are normally
developed by boiler and burner vendors and therefore
exhibit a wide variety of designs. However, the principle
for all NOX burners is the same: They inherently
generate lower NOX emissions due to internal staging of
fuel combustion.
For retrofit of low NOX burners, the number, type and
arrangement of original burners, the structural con-
figuration of the firing walls, and the nature of the low
NOX burners to be retrofitted all need to be considered.
Some retrofits are relatively straightforward; however,
others are complex and may prove to be infeasible due
to the complexity.
2.3.4.1 Staged-Air Burners
One of the first low NOX burners to be developed
employs staged air within the burner itself to effect NOX
reductions (Figure 2-4). The effect is similar to staged
combustion except that the detailed design of the
burner — as opposed to the arrangement of burners in
the combustion chamber or the design of the combus-
tion chamber itself — is responsible for the staged com-
bustion. Low NOX burners of this and other types are
equally applicable in new or retrofitted boilers. With
some designs, the firebox geometry may not permit a
burner retrofit without derate or flame impingement.
The staged-air burner employs primary and secondary
air for combustion in such a way that conditions in the
primary flame zone are substoichiometric (fuel-rich).
The remaining air (tertiary or staged air) is injected after
a brief delay so that the flame experiences a predeter-
mined residence time under reducing conditions. As
with staged combustion, the peak flame temperature is
thus lowered, resulting in lower rates of NOX formation.
2.3.4.2 Staged-Fuel Burners
In staged-fuel burners, all of the air required for com-
bustion is introduced into the primary combustion zone
Figure 2-4. Low NOX Burner: Staged-Air Design (8).
Staged
Air
Secondary
Air
Secondary
Air
Primary Air
and Coal
(Figure 2-5). The fuel is staged, however, so that the
amount of fuel which the primary zone receives is less
than stoichiometric. Primary combustion is under fuel-
lean conditions and therefore NOX formation is repress-
ed by the lower flame temperature brought about by the
excess air. The remaining fuel is injected into the flame
through a peripheral pattern of high-velocity nozzles.
Figure 2-5. Low NOX Burner: Staged-Fuel Design (9).
Secondary
Nozzle
Primary
Nozzle
Secondary
Nozzle
15
-------
The design ensures rapid mixing by the entraining ac-
tion of the injected fuel, which brings about results
similar to exhaust gas recirculation: NOX formation is
reduced by a lower flame temperature in the secondary
zone and by the decreased availability of oxygen in the
staged zone.
2.3.4.3 Variations
A myriad of commercial variations of internally staged
low NOX burners are offered by vendors for both new
and retrofit applications. For example, Foster Wheeler
offers the controlled flow/split flame burner for coal-
fired boilers (Figure 2-6). Adjustable inner and outer air
registers provide staged air while a tapered annular split-
coal nozzle separates the coal into several streams and
thus generates several flames. The result is that primary
combustion is at substoichiometric (fuel-rich) condi-
tions and mixing is delayed until the secondary zone.
Babcock & Wilcox has offered for several years a dual-
register burner, which is essentially a staged-air burner
for coal-fired boilers (Figure 2-7). In fact this burner has
been the standard for all new B&W units sold since
1972. Normally, a compartmented windbox is also
employed. Therefore, the company refers to their low
NOX burner system as the dual-register burner/
compartmented-windbox system.
Riley Stoker offers two major types of low NOX burners
for coal-fired boilers: the controlled combustion Venturi
burner for wall-fired boilers, and the directional flame
burner for turbo-fired boilers. The former (Figure 2-8)
provides a fuel-rich zone along the burner axis sur-
rounded by layers of progressively leaner mixtures. The
Figure 2-6.
Low NOX Burner: Foster Wheeler Controlled
Flow/Split Flame Burner (10) •
Outer Register
for Secondary Air
Inner Register
for Secondary Air
Ignitor
Perforated Plate Air Hood
Movable Sleeve
Flame
Scanner
Split Flame
Coal Nozzle
Tangential
Coal Inlet
latter (Figure 2-9) is actually a burner arrangement
wherein the combustion zone is designed to produce
turbofiring and overfire air is used to produce a staged-
combustion effect.
EPA has pioneered the development of the distributed
mixing burner. It consists (Figure 2-10) of a circular
burner that operates under reducing (fuel-rich) condi-
tions in a recirculation zone where the inner secondary
air combines with the primary air and fuel. Tertiary air is
supplied through outer ports. This arrangement allows
sufficient residence time in the burner zone to reduce
bound nitrogen compounds to molecular nitrogen and
also allows radiation heat transfer to reduce peak flame
temperatures. The tertiary air provides an overall oxidiz-
ing atmosphere to efficiently complete the combustion.
These burner concepts can also be applied with other
commercial burners. The use of tertiary air ports,
however, does complicate the retrofit application by re-
quiring modifications to pressure parts.
A type of low NO* burner — which is actually a burner
arrangement — specifically designed for tangential-fired
boilers is the low NOX concentric firing system
developed by Combustion Engineering in conjunction
with EPA. In the conventional tangential-firing arrange-
ment the primary air/fuel and secondary air streams are
aimed at the same imaginary circle in the center of the
combustion chamber. In the concentric firing system
(Figure 2-11), the secondary air is directed at a larger
concentric circle and therefore the initial combustion
takes place in an atmosphere of reduced oxygen
because the primary air/fuel stream does not entrain
the secondary air stream as rapidly as in the conven-
tional arrangement.
2.3.5 In-Furnace Destruction
This technology, which may be referred to as reburning
or fuel staging, is here classified as a combustion
modification technique, although it involves reduction
of NOX after it has been formed in the combustion zone.
The technology is being investigated for both new and
retrofit applications and for utility and large industrial
boilers and for stationary engines, although it has not
yet achieved full commercialization in the United
States. The technology is applicable for pulverized coal,
gas, or oil.
The principle of in-furnace NOX destruction by reburn-
ing has been the subject of numerous investiga-
tions.(13,14,15,16) Although the basic reactions are
similar to those that occur in staged combustion pro-
cesses, the actual pathways are rearranged. The
hydrocarbon fragments from the decomposing reburn-
ing fuel react directly with the nitrogen oxides from the
primary zone to form fixed nitrogen intermediates (e.g.,
HCN), which can subsequently react under fuel-rich
conditions to form molecular N2. The actual extent of
NOX reduction depends on the time and temperature in
the fuel-rich reburning zone. Above the reburning zone,
overfire or burnout air is added to complete the com-
16
-------
Figure 2-7. Low NOX Burner: Babcock ft Wilcox Dual Register Burner/Compartmented-Windbox System (11).
Compartmented
Windbox
Typical
Burner
Secondary Air
Nozzle
Coal and
Primary Air
Secondary Air
Section X
Section X
Burner Secondary Air
Control Dampers
Burner Secondary
Air Foils
bustion of unburned material. In this zone, some of the
nitrogen intermediaries can be reconverted to NO.
Because of this reconversion there is an optimum fuel-
rich stoichiometry that exists for the reburn zone.
Using the reburning concept originated by Wendt and
Sternling, Mitsubishi Heavy Industries (Japan)
developed the Mitsubishi Advanced Combustion
Technology, or MACT, process (Figure 2-12). The pro-
cess has been tested on various fuels and under condi-
tions where the hydrocarbons for upper injection are the
Figure 2-8. Low NOX Burner: Riley Stoker Controlled
Combustion Venturi Burner (5).
same as or different from the main boiler fuel. Since
1980 further advances have been made through EPA-
funded extramural and in-house projects.(17,18,19,20)
The process can be optimized by using low NOX burners
in the primary combustion zone (e.g., the low NOX
pollution minimum [PM] burners shown in Figure 2-12'
2.3.6 Other Types of Combustion Modifications
A significant number of additional combustion
modifications are employed for reducing NOX emissions
from specific sources. Most are employed for stationary
engines or for turbines. Several of these are considered
in this section and are shown schematically in Figure
2-13.
4-Bladed
Conical
Coal
Spreader
Coal and
Primary Air
Burner
Front
Plate
Venturi
Nozzle Tip
Burner
Throat
2.3.6.1 Lean Burning, Torch Ignition
This technology is applicable to stationary spark-
ignition engines burning natural gas. In this technology,
the air-to-fuel ratio is increased to a level not normally
attainable by the use of spark ignition. An ignition
chamber, rather than sparkplugs, is used in which a
burning jet is created by igniting a fuel-rich mixture.
This torch provides the ignition for the engine and ex-
tends the lean limit that otherwise would be limited by
misfiring, incomplete combustion, or overworking the
turbocharger. A possible drawback to this technology
appears to be relatively high levels of hydrocarbon emis-
sions; however, development work in this area is ongo-
ing. The technology can be applied for new engines or
on a retrofit basis; however, the latter could be quite in-
volved and would also require extensive engineering to
design and construct the most effective ignition
chamber.
17
-------
Figure 2-9. Low NOX Burner: Riley Stoker Directional Flame Burner (5).
Coal and
Primary Air
Directional
Vanes
Coal
Spreader"
s
A
\l
A
i
o
1
V
^
^
I/
Section A-A'
2.3.6.2 Turbocharging
Turbocharging, often employed primarily to increase an
engine's available power, is a technology that also can
be used to reduce NOX emissions by changing the
fuel/air ratio to a leaner mixture for stationary engines.
Turbocharging involves passing the engine exhaust
through a turbine coupled to a compressor. The com-
pressor provides higher pressure combustion air and
therefore increases the mass of air in the cylinder per
unit fuel.
Turbocharging is normally employed on a retrofit basis.
In cases where air temperature rises significantly due to
Figure 2-10. Low NOX Burner: EPA Distributed Mixing Burner (12).
Cast
Refractory
Exit
Tertiary Air Port
Adjustable Primary Air
Swirl Vanes and Coal Inlet
Fixed Vane
Primary Air Swirl
Oil Gun for
Ignition
Secondary
Air Inlets
Adjustable Outer
Tertiary Air Port Secondary Register
for Swirl Generation Adjustable Inner
Secondary Register
for Swirl Generation
Oil Gun
Primary Air
and Coal Inlet
Outer Inner
Secondary Air Secondary Air
Inlet Inlet
A. Development Distributed Mixing Burners (12.5, 50 and 100
x 10" Btu/h)
B. Full Scale Distributed Mixing Burner (12.5 x 10" Btu/h capacity)
18
-------
Figure 2-11. Low NOX Burner: Combustion Engineering Low
NOX Concentric Firing System (6)-
Direction of
Axis for
Coal/Primary
Air Flow
\
Burner
Assembly
Direction of
Axis for
Secondary
Air Flow
Burner
Assembly
Burner
Assembly
Burner
Assembly
compression, an intercooler is also required. Tur-
bocharging is applicable to compression-ignition
(diesel) and spark-ignition (gas) stationary engines.
2.3.6.3 Charge Cooling
This technology, often referred to as charge refrigera-
tion, is employed in diesel as well as spark-ignition sta-
tionary engines. Charge cooling can be applied on a
new or retrofit basis. The NOX reduction principle is
simply to decrease the peak flame temperature by cool-
ing the air/fuel mixture prior to combustion. Normally,
a refrigeration system is arranged as a heat exchanger
upstream of the intake manifold to cool the combustion
air. The air temperature reduction is usually on the order
of 50 to 75F.
Charge cooling is often employed in conjunction with
other NOX reduction techniques, particularly multiple
Figure 2-12.
Additional Air
Upper Burner
Overfire Air
In-Furnace Destruction by Mitsubishi Advanced
Combustion Technology (MACT) Process (21).
Combustion
Completion
Zone
NOX
Decomposing
Zone
Main
Combustion
Zone
Figure 2-13. NOX Reduction Options for Stationary Engines.
AIR
Turbocharging provides
high pressure (and there-
fore greater mass) air,
permitting lean burning
Charge cooling lowers
peak flame temperature by
cooling the air/fuel mixture
IGNITION
Torch ignition permits lean
burning (spark-ignition
engines only)
Delayed ignition reduces
the mean residence time of
fuel in the cylinder, thus
shortening the high NO*
emission period (spark
ignition engines only)
FUEL
Delayed fuel injection
reduces the mean
residence time of fuel in
the cylinder, thus shorten-
ing the high NOX emission
period (diesel engines only)
Charge cooling lowers
peak flame temperature by
cooling the air/fuel mixture
sparkplug firing, in order to rectify the problem of misfir-
ing that is often experienced when combustion air is
cooled to low temperatures.
2.3.6.4 Retardation
NOX emissions from all types of stationary engines can
be reduced by adjustments that retard either the ignition
timing (spark-ignition engines) or the fuel injection
(diesel engines). This is strictly a retrofit technology and
is normally applied with greater frequency on high-
compression (diesel) engines and engines that run lean
(such as turbocharged engines). Retarding both ignition
and injection delay combustion for an instant. The
theory is that the first portion of fuel that burns pro-
duces an inordinate amount of NOX because it is expos-
ed to high temperatures the longest, and because it is
heated by compression, while the balance of the fuel is
burned. Therefore, this technology brings about a
reduction in NOX formation by reducing the dwell time
of the fuel in the cylinder.
An added benefit of NOX reduction by ignition or injec-
tion retard is an apparent reduction in emission of un-
burned hydrocarbons. However, there is often an
increase in smoke emissions in diesel engines. Further-
more, cylinder exhaust valves are exposed to higher
temperatures with this technology and may experience
reduced life.
19
-------
2.4 Postcombustion Technologies
There are many technologies for reducing NOX once it
has been formed by the combustion process. These do
not compete directly with combustion modifications but
rather are considered only after the easier to implement
and less costly technologies have been exhausted and
even more stringent control is required. Usually referred
to as flue gas treatment (FGT) processes, these
technologies are seldom cost-effective for moderate
NOX reduction when compared to combustion
modifications for several reasons: relatively high initial
cost, high operating cost, and possible waste disposal
costs. Accordingly, these technologies are used when
stringent regulations require high NOX reductions. In the
future, flue gas treatment may be the best approach for
removal of NOX and sulfur oxides, and a number of pro-
cesses are under development to achieve these goals
simultaneously.
FGT processes can be categorized as either dry or wet
and can be further divided by the chemical reaction
principles involved in the conversion. Figure 2-14 shows
these processes, organized by type.
2.4,1 Dry Processes
A wide variety of dry processes are either commercially
available or well along in research and development.
They range from catalytic and noncatalytic reduction to
adsorption processes and irradiation with electron
beams. Dry processes usually involve less equipment
and therefore are less costly than wet processes, and
generally also produce less waste to dispose of.
2.4.1.1 Selective Reduction
Selective reduction technologies are either catalytic or
noncatalytic. In the United States, both methods have
been employed for oil-fired and gas-fired utility and in-
dustrial boilers and process heaters and are being
researched for use in coal-fired utility boilers. In addi-
tion, catalytic processes have been used for spark-
ignition stationary engines and for gas turbines.
In selective catalytic reduction (SCR), which is the most
popular FGT process in international utility use today,
ammonia is employed as the reducing agent. In the SCR
process, NOX is reduced to N2 and H20 by ammonia
(NH3) at 300-450°C in the presence of a catalyst. NH3 is
an acceptable reducing agent for NOX in combustion
gases because it selectively reacts with NOX while other
reducing agents, such as H2, CO, and CH4, readily react
with 02 in the gases. Figure 2-15 shows a typical
flowsheet for a selective catalytic process with am-
monia. Flue gas from the boiler is passed through a
reactor column which contains the catalyst bed. So-
called parallel flow catalyst beds may be used in which
the gas flows through channels rather than pores to
minimize blinding of the catalyst by particulates. Am-
monia vapor is injected into the flue gas upstream of the
reactor. The treated flue gas then passes through the
combustion air preheater and then to paniculate and
perhaps sulfur dioxide removal equipment before ex-
iting up the stack. The major items of process equip-
ment are the reactor and the equipment to store,
vaporize, and inject ammonia.
In the early stages of its development, SCR had the
following problems: catalyst poisoning by SOX in the
gas; plugging of the catalyst by dust; ammonium
bisulfate deposition on the catalyst below about 300°C;
ammonium bisulfate deposition in the air preheater
below about 250°C; catalytic promotion of oxidation of
SO2 to S03; and erosion of the catalyst by fly ash from
coal. (22)
Those problems have been addressed by the following
countermeasures:
• Use of base metal catalysts with Ti02 instead of
AI203 or Fe203 substrates
• Use of parallel-flow type catalysts such as
honeycomb, plate, and tube catalysts
• Maintaining the gas temperature above 330°C by us-
ing an economizer by-pass system
Figure 2-14. Classification of NOX Flue Gas Treatment Processes .
Blocks may represent several available processes,
and many processes simultaneously remove SOX.
20
-------
Figure 2-15. Typical Flow Diagram for Selective (Ammonia)
Catalytic Reduction Process (8).
Figure 2-16. Thermal DeNOx System — Process Flow
Diagram (8).
Boiler
Coal
i Flue Gas
Flue Gas
Desulfurization
Unit
H\
Stack
Unloading
Compressor
Rail or
Truck Hook-Up
Ammonia A Steam
Storage Tank/
Air
NH3 Vaporizer
• Keeping unreacted NH3 (IMH3 at the reactor outlet)
below about 3 ppm
• Using a low-oxidation catalyst
• Using a moderate gas velocity, a hard catalyst, and a
device for erosion prevention such as dummy spacer.
There are two noncatalytic selective reduction pro-
cesses. The Thermal DeNOx process was developed by
Exxon and is shown schematically in Figure 2-16. It
employs ammonia as the reductant but the reaction is
carried out at high temperature rather than under the in-
fluence of a catalyst. A similar process using urea as the
reducing agent was developed by EPRI.I23) Both pro-
cesses require that the reducing agent be injected
directly into the superheater section of the boiler; it is
therefore less complicated and expensive to retrofit
than the catalytic process. However, temperature con-
trol is critical to avoid producing more NO or releasing
unreacted reductant to the stack; therefore, ensuring
that the injection is at the right place even though
temperature patterns fluctuate is difficult and may limit
the reduction efficiency. This process is being used for
several boiler heaters, primarily on the West Coast.
2.4.1.2 Nonselective Reduction
This technology encompasses catalytic reduction
without a reductant that is selective to NO. This
1 1
1 Superheater 1 —
F
C
£3
NH3
ressure
ontrolle
( NH3 Storage Tank J
NH3Su
»
pply System
-J
Nh
C
r S
Vaporizer
Air Supply System
1 Compressor J ^
Heat
Exchanger
3s-,
a Flow
ontrol
ation i
j
t
Boiler
Flue Gas
r
Injection Grid
System in
Superheater
Section of Boiler
Hiiiif
1 23456
1 Zones
technology involves application, on a new or retrofit
basis, of a catalytic converter of the type employed in
most newer automobiles in the United States. The
technology in terms of stationary sources is limited to
stationary reciprocating engines, particularly rich-burn,
spark-ignition engines.
This technology is applied by installing the catalytic
converter in the exhaust line from the engine. The con-
verter consists of a catalyst bed supported within a
pressure-tight housing. Exhaust gas passes straight
through with a large percentage of the NO reduced to
molecular nitrogen. Catalysts employed are normally
noble metals such as platinum.
2.4.1.3 Simultaneous NOX/SOX Removal
Dry processes for simultaneous NOX/SOX removal that
are under development include reaction of sulfur
dioxide (S02) with copper oxide simultaneous with
selective catalytic reduction of NOX with ammonia; ad-
sorption onto alkalized alumina; and irradiation with
electron beams. Although many process concepts are
promising, they have not been demonstrated commer-
cially in the United States.
2.4.2 Wet Processes
Most wet flue gas treatment processes for NOX removal
represent natural extensions of processes originally
developed for SOX removal only. To date, wet pro-
cesses are still in the developmental stage. Thus they
cannot be considered as alternatives until they ex-
perience technical breakthroughs.
21
-------
-------
Chapter 3
Performance and Cost Data:
Utility and Large Industrial Boilers
3.1 Introduction
In this chapter data on NOX removal efficiency and
associated cost are presented for actual operating
technologies employed for utility and large industrial
boilers. Systems are described in detail, including re-
quirements for retrofitting the technologies to existing
boilers. In general, performance and costs are more
favorable for new unit installation since space and
operational requirements can be designed into the unit.
Retrofit costs and performance are usually quite site-
specific due both to unique boiler characteristics and
space constraints.
3.2 Coal-Fired Boilers
Coal-fired boilers are either pulverized coal-fired or
stoker-fired, the latter usually involving small units.
Most of the NOX control technologies developed for
coal-fired boilers — and for boilers in general — have
been developed for pulverized coal units. Data for these
technologies are presented below, followed by discus-
sions of technologies developed for stokers.
3.2.1 Pulverized Coal Boilers
The primary NOX reduction technologies for pulverized
coal boilers have been low NOX burners, various types
of staged combustion, and selective reduction techni-
ques. Low excess air is easy to implement but may not
bring about the required reductions. These technologies
are presented below.
3.2.1.1 Low Excess Air
The effectiveness of low excess air firing for coal-fired
utility boilers has been summarized for tests on boilers
of the three most prevalent configurations.(14,24) The
data are shown here in Table 3-1.
Note from Table 3-1 that for relatively modest reduc-
tions in excess air (for example, from 24 percent to 16
percent excess air, or a reduction of 8 percent for
tangential firing), significant reductions in NOX emis-
sions can be achieved. In most cases higher NOX reduc-
tions are precluded by operational problems that arise at
lower levels of excess air. At very low excess air rates
smoke and excessive carbon monoxide emissions result
from incomplete combustion. Furthermore, depending
on the type of coal fired, slagging and/or corrosion may
increase to unacceptable levels and boiler tubes may
suffer premature failure. Although it is currently a com-
plex matter to predict the minimum level of excess air
that will permit safe operation, further research in this
area may both define the lower limit and perhaps extend
it further with newer combustion chamber designs.
There is no significant additional cost to implement low
excess air firing in new boilers. (3) In fact, decreasing ex-
cess air can in many cases actually reduce the cost of
operating a boiler by increasing boiler efficiency as
much as 5 percent. The estimated capital cost to retrofit
a boiler for low excess air firing would be about $400 per
Table 3-1. Average Reported NOX Reduction with Low Excess Air Firing in Coal-Fired Utility Boilers (14).
Baseline
Low Excess Air (LEA)
Equipment Type
Tangential"
Opposed wall6
Single wallc
Average (mean)
Number
of
Tests
21
11
23
55
Stoichiometry
to Active
Burners, %
124
126
123
124
NOX Emissions
ppm dry
<5> 3% O2
459
746
624
609
Stoichiometry
to Active
Burners, %
116
118
114
116
NOx Emissions
ppm dry
@ 3% O2
373
660
525
522
Average NOX
Reduction, %
19
12
16
(1)"
16
Maximum NO,
Reduction
Reported, %
42
23
25
(3)"
30
"Burners firing from the furnace corners on a tangent to an imaginary circle in the center of the boiler.
^Burners firing from two opposed walls in the furnace.
cBurners firing from only one wall in the furnace.
^Numbers in parentheses refer to boilers originally designed for coal firing with wet-bottom furnaces.
Note: To convert values in ppm (3% 02) to lb/10' Btu, multiply by 0.0014 for coal and oil, 0.0012 for natural gas. (These factors are approx-
imate and highly dependent on fuel characteristics.)
23
-------
MW (1984 dollars), which is mostly for modifications to
the combustion air handling system and windboxes.(25)
A range of $640 to $740 per MW (1983 dollars) for
capital cost, with a negligible annual cost, has also been
reported.(7) Many operators require an oxygen trim
system for low excess air firing to closely control the ex-
cess air rate; this will entail additional cost.
3.2.1.2 Burners Out of Service
One method of staging combustion in existing boilers is
to employ burners out of service (BOOS). Figure 3-1
presents results for BOOS operation for several single
wall-fired boilers in terms of NOX reduction versus per-
cent BOOS. For these tests the total number of burners
ranges from 15 to 24, with those out of service from 2 to
8. For example, 2 of 16 burners out of service means
12.5 percent BOOS. The percentage reductions in
Figure 3-1 represent between 100 and 600 ppm (dry, 3
percent oxygen basis), from baseline NOX emissions of
400 to 700 ppm.
Figure 3-2 shows results for BOOS operation of horizon-
tally opposed wall-fired boilers. Note that with two ex-
ceptions BOOS operation resulted in NOX emission
reductions of 15 to 30 percent from the baseline level.
The Gaston Unit 1 boiler was equipped with specially
designed (low NOX) burners and therefore did not ex-
perience a dramatic NOX reduction from BOOS opera-
tion.
Substantial NOX reduction by BOOS operation has also
been shown for tangential-fired boilers, as apparent
from Figure 3-3. Again, there is a reasonably good cor-
relation between degree of BOOS (percentage of
Figure 3-1. IMOX Reduction by BOOS for Single Wall-Fired
Boilers Burning Coal at 120 Percent Excess Air
(15,16).
Figure 3-2. NOX Reduction by BOOS for Horizontally Op-
posed Wall-Fired Boilers Burning Coal (15,16).
60
Station/Unit
• Crist 6
• Edwards 2
A Mercer 1
50 I— •Johnston 2 ~~?
*Widows Creek 5 >£
^Shawnee 10 > *
- 40
T3
0)
cc
X
o
30
20
/ /
-/~~,'—
*/
10
15
20 25
BOOS, %
30
35
1000
900
800
ON
* 700
co
2
^. 600
Q
I
0~ 500
Z
400
300
200
Open symbols = baseline operation
Closed symbols = BOOS operation
Four Corners 4
Gaston 1
123 456
Boiler Excess 02, %
burners out of service) and NOX reduction achieved.
The units tested range in size from 16 burners (95 MW)
to 56 burners (800 MW). For this boiler configuration
the most effective BOOS pattern was observed to be
operation with the top row of burners out of service, a
pattern that simulates overfire air.
Retrofitting a coal-fired boiler for BOOS operation nor-
mally does not involve a significant capital expense.
However, there are subtle costs involved in such a
retrofit that are difficult to predict. Considerable ex-
perimentation may be required to find the optimum
number and arrangement of burners to be taken out of
service, although most BOOS patterns consist of taking
burners out of service in the upper regions of the boiler.
In any case, a BOOS design creates fuel-rich and oxy-
gen pockets that essentially stage the combustion. This
not only takes operator and management time but also
requires testing the flue gas for NOX as well as for
smoke and carbon monoxide. Tests are necessary to
verify the arrangement that will minimize NOX while
holding smoke and carbon monoxide emissions to ac-
ceptable levels. Each case will be different in that there
are no "normal" or "average" combinations of coal pro-
24
-------
perties, boiler layouts, and combustion conditions.
BOOS may also require a boiler derating if the pulverizer
mills cannot supply the additional fuel to the lower (in-
service) burners. If so, there may be a substantial addi-
tional cost for replacement power.
3.2.1.3 Biased Firing
Little performance or cost information is available for
biased firing in coal-fired boilers. The meager data that
are available indicate that NOX reductions on the order
of 7 to 8 percent were found in limited testing.
3.2.1.4 OverfireAir
Overfire air is a new and retrofit technology for staging
combustion in coal-fired boilers (including stoker-fired
boilers). It has been used extensively, especially for
tangential-fired boilers where it is offered for new
boilers. A typical overfire air retrofit is shown
schematically in Figure 3-4.
Limited data are available on the effectiveness of over-
fire air for single and horizontally opposed wall-fired
boilers. A 22 percent NOX reduction was achieved in a
test of simulated overfire air in which the top row of
burners in a single wall unit was taken out of service.
This mode of operation provided about 16 percent of
the total air as overfire air to the boiler. Tests of a
horizontally opposed 350 MW unit showed NOX reduc-
tions of 50 percent and higher (from a baseline of 650
ppm) with a significant (greater than 50 percent)
amount of air delivered as overfire air. Although im-
pressive, this testing was limited and further verification
in the field is required.
Figure 3-3. NOX Reduction by BOOS for Tangential-Fired
Boilers Burning Coal (15,16).
Figure 3-4. Typical Retrofit Arrangement for Overfire Air
(26).
50
40
30
20
7
10
/ /
I I
15 20
BOOS, %
Station /Unit
• Barry 2
• Barry 4
ANaughton 3
ANavajo 2
^Widows Creek 7
i l
25 30
F - Fuel and air
A-Air
0 - Overfire air
Substantial data showing the effectiveness of overfire
air for NOX control are available for tangential-fired
boilers. Figure 3-5 shows the reductions achieved in
tests at three boilers ranging in capacity from 130 to 800
MW, each operating at 120 percent excess air. Note
that significant (10 to 30 percent) reductions were
achieved at relatively high overfire air rates. A 1980
Acurex Corporation study reported on 46 tests of
tangential-fired boilers wherein NOX emissions were
measured for the baseline condition and the condition
of overfire air. The results are shown in Table 3-2. Note
from Table 3-2 that diverting about 20 percent of the
total combustion air to the overfire air ports resulted in
an average reduction of 31 percent in NOX emissions.
Considerable cost data are available for retrofitting over-
fire air to existing boilers. Figure 3-6 presents capital
cost in terms of dollars per kilowatt versus boiler size in
megawatts. Note that data are from three different
sources and that the actual boiler modifications were
likely to be different for each. Therefore a cost range
has been designated, as the shaded area indicates. Note
also that the cost data plotted are from three different
time periods.
Table 3-3 summarizes capital and annualized costs
(1983 dollars) for retrofitting three model-sized boilers
with overfire air. For retrofits, overfire air is in general
relatively inexpensive; it mainly involves the installation
of several ports above the burner rows and the
associated air piping. Low excess air is the least costly
option and in larger sizes may actually provide a net
credit due to increasing the efficiency of the boiler.
25
-------
Figure 3-5. NOx Reduction by Overfire Air For Tangential-
Fired Boilers Burning Coal (7).
Figure 3-6. Costs of Retrofitting Coal-Fired Boilers for
Overfire Air (7).
T3
CD
tr
40
30
20
10
Station/Unit
• Comanche 1
• Barry 2
A Navajo 2
50 75
Overfire Air, %
100
125
3.2.1.5 Arch Firing
Because this technology involves elemental changes in
the design of a boiler, it cannot be conceived of as a
retrofit technology but rather must be viewed as a
potential "new" technology. Arch firing is a technology
primarily limited to boilers firing difficult-to-burn an-
thracite coal and coke and a small number of
bituminous coal-fired units of Wisconsin Electric Power
Co.(27) There are few studies comparing the effective-
ness of arch firing and wall firing from which a mean-
ingful estimate of NOX reduction could be calculated.
Tests of existing arch-fired boilers (27) have shown that
NOX emissions are on the order of 200 to 350 ppm NOX
(at 3 percent oxygen), which is relatively low for "un-
controlled" NOX emissions. However, the units tested
were relatively small, ranging in capacity from 80 to 265
MW; most of the units constructed in the future would
be expected to be in a larger size range.
Available data (27) indicate that the total installed cost
for a base arch-fired boiler (500 MW) is estimated to be
5.4 percent greater than the cost of a comparable wall-
fired unit; an alternate arch-fired boiler was estimated to
cost 26 percent more than the base wall-fired unit. The
alternate unit selected corresponded to a larger boiler
due to different geometry, although it too was rated at
500 MW.
3.2.1.6 Low NOX Burners
Burners have been designed, primarily by equipment
and boiler vendors, that are inherently low in NOX pro-
o
4.0
3.0
2.0
1.0
A Jain, etal. (1972)
• Bartok, etal. (1969)
• Shimizu (1975) LADWP
200 400 600 800
Unit Size, MW
1000
duction, usually because of internal combustion
staging. Many of these burners have been tested at
laboratory, demonstration, and full scale, with cor-
responding data for NOX reductions achieved and cost.
The data available on cost are usually for retrofitting an
existing boiler. The reader is cautioned, however, that
the performance and cost data presented below are for
specific applications and are given as examples of
achievable results. Particularly for retrofits, every case
will be different and more than one NOX reduction
technology may be required in some retrofit
applications.
Riley Stoker Corporation has developed and tested the
Controlled Combustion Venturi (CCV) burner for retro-
fitted wall-fired boilers. Figure 3-7 shows results of pilot-
scale tests for the traditional Riley flare burner and the
CCV burner. The tests were conducted at the EER Cor-
poration test facility at firing rates up to 50 x 106 Btu/h.
The NOX reductions achieved by the CCV burner were
about 55 percent compared to the baseline of the flare
burner.
The effectiveness of the CCV burner as a retrofit to
replace the traditional flare burner was also tested for
full-scale operating units. A single-wall boiler rated at
400 MW was tested before and after replacing the 24
flare burners with CCV burners.(5) The NOX emission
Table 3-2. Average Reported NOX Reduction with Overfire Air Firing in Tangential Coal-Fired Utility Boilers (14).
Baseline Overfire Air (OFA)
Number
of
Tests
Stoichiometry
to Active
Burners, %
NOX Emissions
ppm dry
@ 3% O2
Stoichiometry
to Active
Burners, %
Furnace
Stoichiometry, %
NOX Emissions
ppm dry
@ 3% 02
Average NOX
Reduction, %
Maximum
NOX Reduction
Reported, %
46
129
454
105
122
311
31
41
26
-------
Table 3-3. Cost of Retrofitting Coal-Fired Boilers for Overfire Air (7)
(1983 dollars)
Overfire
Air
New
Retro
MW
250
500
750
250
500
750
Capital
Invest.
$/kW
0.30
0.20
0.15
1.17
0.75
0.60
Annual
Capital
$/kW
0.054
0.036
0.027
0.21
0.14
0.11
Oper.
$/kW
0.005
0.003
0.003
0.02
0.01
0.01
Annual
Fuel Total
$/kW $/kW
0.059
0.039
0.030
0.23
0.15
- 0.12
Electrical
mills /kWh
0.011
0.007
0.005
0.043
0.027
0.022
rate was reduced approximately 50 percent to levels of
320 to 440 ppm (3 percent oxygen) while burning a high
volatile "C" bituminous coal at 22 percent excess air.
Loss on ignition tests of combustible materials in the
ash established a decrease in boiler efficiency of only
0.25 percent.
Also tested was a 360-MW, horizontally opposed boiler
with a total of 24 burners.(5) After retrofitting with CCV
burners the NOX emission level fell from 810 ppm to bet-
ween 353 and 397 ppm (all on 3 percent oxygen basis).
Again, no significant adverse effects on boiler perfor-
mance were noted.
The Controlled Flow/Split-Flame (CF/SF) burner has
been developed by Foster Wheeler Energy Corporation
and has been tested on a variety of pilot and full-scale
units. Figure 3-8 presents idealized NOX emissions ver-
sus burner stoichiometry for the CF/SF burner as com-
pared to the intervane burner which represents the
baseline case for retrofitting a Foster Wheeler boiler.
Foster Wheeler has also tested NOX emissions for a
360-MW front wall-fired boiler and a 525-MW opposed-
fired boiler as well as for a 50 x 106 Btu / h test burner, all
equipped with CF/SF burners.CIO) Although none was
tested before retrofit, the company claims NOX reduc-
tions of about 60 percent by retrofitting with CF/SF
burners. Currently, two domestic utility boilers are
operating with retrofitted CF/SF burners: a 350-MW
Figure 3-8. Theoretical NOX Emissions Versus Burner
Stoichiometry for the Controlled Flow/Split
Flame Burner (10).
Figure 3-7. Pilot-Scale Test Results for the CCV Burner (5).
1000
800
£ 600
^ 400
O
200
Riley Flare Burner
with 20% Flue
Gas Recirculation
i Distributed Mixing Burner Staged
I
100 200 300
Burning Area Heat Release Rate, 103 Btu/h-ft2
400
Controlled Flow Split-Flame Burner
0.3 -
0.2
90 100 110
Burner Stoichiometry, %
120
27
-------
single-wall-fired unit and a 525-MW opposed-fired unit.
Tests indicate that both units operate at well below the
NSPS limit of 0.5 lb/106 Btu for subbituminous
coal. (10)
Foster Wheeler, in conjunction with EER Corporation
and EPA, has also developed the Distributed Mixing
Burner (DMB). Full-scale test results may be available
presently but to date the bulk of data available are for
research burners and in some cases for research
furnaces.
Babcock & Wilcox Company developed the Dual
Register burner early in the history of low NOX burners
and as a result, several thousand of these burners have
been sold domestically, mostly for new wall-fired
boilers. This burner has also been retrofitted and in fact,
the company's NOX emission guarantees are based on
extensive testing in a retrofitted boiler. Considerable
data are available on the emissions and emission reduc-
tions achievable with this burner simply because there
are so many currently in use. Figure 3-9 shows the per-
formance of the dual register burner compared to the
B&W circular burner it has replaced. Note that the
reductions achieved are on the order of 50 to 60 percent
throughout the range of boiler sizes and are nearly all
comfortably below the NSPS level of 0.5 Ib/106 Btu for
subbituminous coal.
Babcock & Wilcox is also, in conjunction with EPRI,
developing a retrofit low NOX burner to replace B&W
cell burners. (28) Cell burners, inherently high in NOX
emissions, represent a significant segment of the pre-
NSPS pulverized coal-fired boiler population. Each cell
consists of two circular burners mounted together to
produce a high-velocity turbulent flame in horizontally
opposed wall-fired boilers.
The company has developed two low NOX replacement
burners, able to be retrofitted into the existing cell. One
is a cell in which the upper burner supplies overfire air
only; the other is a pair of distributed mixing burners. In
pilot-scale testing the former arrangement provided 65
percent NOX reduction (Figure 3-10). In addition to
reducing NOX emissions to below the NSPS level, the
pilot tests of the ash revealed that for both coals tested,
operation with low NOX burners resulted in slightly
lower levels of unburned hydrocarbons, suggesting that
thermal efficiency is preserved after retrofit.(28) This
burner arrangement is currently being evaluated in an
actual operating boiler of 610 MW capacity, and will
undergo further subscale tests in an EPA test facility.
Low NOX burners for tangential-fired boilers have been
developed primarily by Combustion Engineering, Inc.,
and their licensee, Mitsubishi Heavy Industries. The
Low NOX Concentric Firing System (LNCFS) was
developed by C-E primarily as a retrofit technology for
existing coal-fired boilers.
In conventional tangential firing, the burners are corner
mounted with their axes tangent to an imaginary circle
in the center of the boiler. In LNCFS, the auxiliary air is
directed at a larger concentric circle so that the flame
front is stabilized and the devolatilization of the coal oc-
curs in a fuel-rich atmosphere. The retrofitting involves
modifications to the boiler windboxes alone; therefore,
strictly speaking the LNCFS is not really a low NOX
burner. However, modifications of this type are often
classified as low NOX burners.
Results of full-scale (400 MW) tests of the LNCFS ver-
sus the conventional burner arrangement are shown in
Figure 3-11. Note that NOX reductions on the order of 20
to 30 percent from baseline levels of 460 to 480 ppm are
Figure 3-9. NOX Emissions for Dual Register Versus Circular Burners in Coal-Fired Boilers (14).
0.8
m 0.6
&
X
o
Z 0.4
0.2
0
i
//
I
"*.*.
• Circular
Burner
EPA NOX
Emission Limit
1!
Ii
i
!
1
!
I
II
II
V-
1.
Dual Register
Burner
90 330 470 470 550 550- 550 575 580 600 650 675 700 700 700
Unit Capacity, MW
28
-------
Figure 3-10. Pilot-Scale Results for NOX Reduction by Low
NOX Cell Retrofit Burners (28).
2.5
1.6
df
d
0.5
D Standard cell burner with Ohio No. 6
• Standard cell burner with lower Kittaning
A Low NOX cell burner with Ohio No. 6
A Low NOX cell burner with lower Kittaning
3.4
4.4
Excess 02,
possible. Even greater reductions can be achieved
where overfire air is employed in addition to retrofitting
with LNCFS (Figure 3-12). In these tests, potential im-
pacts on boiler efficiency were gauged by unburned car-
bon in flyash. Levels of unburned carbon in flyash were
actually lower after retrofit (on an equivalent NOX emis-
sion basis), indicating no adverse impact of the LNCFS
on efficiency. (6)
MHI has also developed the Low NOX Pollution
Minimum (PM) Burner System to meet stringent NOX
Figure 3-11. NOX Emissions for the Low NOX Concentric Fir-
ing System in Coal-Fired Tangential Boilers: No
Overfire Air (6).
regulations in Japan. The PM burner basically divides
the coal/air mixture into fuel-rich and fuel-lean streams
and supplies auxiliary air in an intricately designed wind-
box. The design is a refinement of the SGR burner that
itself was a wall-fired version of the LNCFS concept.
Projections of NOX reductions for a 600-MW boiler and
a 575-MW divided boiler are shown in Table 3-4. Note
the significant NOX reductions estimated over conven-
tional tangential firing. Case I refers to a retrofit design-
ed for minimum boiler modifications, while Case II
refers to a retrofit where maximum NOX reduction was
the goal. The "new unit" was actually Plant A retrofit-
ted in such a manner as to reflect how the boiler would
be constructed if new.
The cost of retrofitting with the PM burner system is
given later in this chapter.
Figure 3-12. NOX Emissions for the Low NOX Concentric Fir-
ing System in Coal-Fired Tangential Boilers:
Full Overfire Air (6).
m
o
n.
o.
500
400
300
200
Full load (390-420 MW)
Full overfire air
• Baseline
• Postmodification
2.0
3.0
4.0
Excess 02, %
0.5
6.0
600
550
500
d
£ 450
o
I" 400
35°
z
300
250
200
Full load (390-420 MW)
No overfire air
• Baseline
• Postmodification
2.0
3.0
4.0
Excess 02, %
5.0
6.0
3.2.1.6.1 Side Effects of Low NOX Burner and
Staged Combustion
Commercial methods of reducing NOX in coal-fired
boilers are often characterized by two conditions poten-
tially detrimental to the longevity of boiler tubes: a
reducing atmosphere that may promote slagging and
remove the oxidized coating of tubes and expose them
to accelerated corrosion; and flame impingement
resulting from longer flames which can overheat tubes
and cause premature failure by thermal stress.
Air staging (BOOS or overfire air) can potentially result
in increased tube wall corrosion. Proper application and
design can minimize or even eliminate corrosion.(17)
Extensive testing of the corrosion potential of low NOX
operation has been conducted with major conclusions
as follows:
• It is critical that the burner and boiler design and
operating conditions avoid sulfidation of tubes by
29
-------
Table 3-4.
IMOX Performance for PM Burner System in
Tangential-Fired Boilers (6)
NOX Guarantee"' ppm
% NOX Reduction
Plant A6
Case 1
Case II
New Unit
Plant Bc
190
150
150
160
46"
57"
57d
50*
"Corrected to 3% 02.
*600 MW boiler.
C575 MW divided boiler.
''Based on Plant A emissions of 350 ppm corrected to 3% O2.
"Based on Plant B emissions of 320 ppm corrected to 3% 02.
flame impingement of unburned sulfur pyrites and
corrosion by liquid pyrosulfate.
• Proper selection of materials of construction can
minimize stress corrosion from alternating oxidizing
and reducing environments.
• Corrosion can be minimized by selection of certain
stainless steels and other alloys over carbon steel.
While corrosion is the side effect of most concern
resulting from low NOX combustion, there are several
others, which become constraints on how the unit is
controlled:
• Increased carbon monoxide emissions which may be
generated by incomplete combustion at very low ex-
cess air rates
• Smoke (carbon) emissions which may likewise in-
crease as a result of low excess air
• Problems associated with flame instability
• Limits on flexibility of the combustion unit (changes
in load)
• The presence of carbon in the fly ash.
These problems may impose limits on fuel switching or
even require a derating of the unit.
3.2.1.6.2 Cost of Low NOX Burners
Data on the cost of retrofitting coal-fired burners are not
plentiful and the data that are available are naturally
highly site-specific. However, a few examples are
presented here in order to show the magnitude of the
cost of such retrofits. In terms of new boilers, the in-
cremental cost of the low NOX burners themselves is im-
possible to separate out. Low NOX burners or controls
such as low excess air, biased firing, and so on have
become state-of-the-art and thus are offered as stan-
dard equipment for new boilers at relatively low invest-
ment cost.
Figure 3-13 shows the cost (1983 dollars) in terms of
dollars per ton of NOX removed for retrofitting boilers of
two different ages with the Foster Wheeler controlled
flow/split flame burner. Note that because of increased
capability to fire at low loads without significant aux-
iliary fuel, the costs are presented as a function of "oil
reduction" in hours per year, which means the hours
that the original Class 1 ignitors (which fire auxiliary oil
Figure 3-13. Offsetting Retrofit Costs of Controlled
Flow/Split Flame Burners with Oil Savings*(10).
20 Years of
Boiler Life Remaining 10Yearsof
~ Boiler Life Remaining
Breakeven
0 50 100 150 200
Reduction in Oil Usage by Auxiliary Oil Burners, h/yr
*This offset is a function of remaining boiler life, reductions in
annual oil usage by auxiliary oil burners (expressed in hours per
year), and the retrofit cost of the controlled flow/split flame
burners (expressed as dollars per ton of NOX removed).
to sustain combustion at low boiler loads) are not fired.
Further note that for the newer boiler, oil reduction in
excess of about 70 hours per year will offset the cost of
retrofitting and bring about a net savings at higher
values.
Estimated capital and annual power costs for retrofitting
with PM burners are shown in Table 3-5 in 1981 dollars.
Note the high capital cost for the extensive modification
(Plant A, Case II) and for the retrofit in a divided boiler
(Plant B). These costs are from a feasibility study (29)
and are only presented as an example. Every retrofit is a
special case with a corresponding cost that may not ap-
ply to any other boiler.
Table 3-5. Cost of Retrofitting with PM Burners°l20)
(1981 dollars)
Total Capital
$/kW
Plant A
Case I
Case II
New Unit
Plant B
7.91
15.90
5.44
14.29
Power Costs
mills /kWh
0.22
0.45
0.15
0.40
"Confidence levels ±20%.
3.2.1.7 Reburning
There have been a number of development efforts, in-
itially in Japan, to apply reburning for the control of NOX
from large industrial and utility boilers. The activity in
the United States has extended this technology to other
combustion sources and to the consideration of
simultaneous S02 control by dry sorbent injection.
30
-------
These have been supported by EPA, Department of
Energy (DOE), EPRI, and GRI. Several specific
technologies have arisen such as MACT (for Mitsubishi
Advanced Combustion Technology), an application to
cyclone combustors (Japanese Ministry of International
Trade and Industry) (30), an application to furnaces
(Hitachi Zosen) (31), and the In-Furnace NOX Reduction
(IFNRU32)
The bench- and pilot-scale results have generally
demonstrated greater than 50 percent reduction of NOX
using between 10 and 20 percent of the fuel for reburn-
ing.(30,32,33,34) A relatively complete data base is
available on the impact of the major process variables
from these studies. The data indicate that less reduction
can be achieved at lower initial NOX levels and that the
parameters within the reburn zone control the
achievable NOX reduction. In particular, the reburn fuel
type can have a significant impact on the level of con-
trol. The nitrogen in the reburn fuel is an impediment to
the process. Natural gas is an effective reburn fuel
because it has no fuel nitrogen, reacts rapidly so that
short reburn zones are achievable, does not slag or have
burnout problems as with other fuels, can achieve more
reduction in NOX especially at lower initial values, and
finally, can work at lower temperatures. Pilot-scale tests
with practical times, temperatures, and primary fuels
have generally demonstrated over 60 percent
reduction.(35,36) EPA in-house tests have shown that
distillate oil is slightly more effective than natural gas as
a reburn fuel.
Full-scale test results are available for a 600-MW
tangentially fired boiler (37) firing coal and oil and using
oil as the reburn fuel. There were five burner elevations
for oil and two for coal. Figure 3-14 shows the results
which are for a 90:10 ratio of oil to coal. Note that NOX
reductions on the order of 40 to 50 percent were found
in these tests. There are no known instances of full-
scale testing of reburning for coal-fired boilers in the
United States, although a number of studies are cur-
rently underway.
3.2.1.8 Slagging Combustors
Recently there has been a significant amount of
research at the laboratory and pilot scale concerning
slagging or external combustion. The technology has
emerged out of the field of magnetohydrodynamic
(MHD) power generation where such combustors are
used to provide heat to ceramic heat exchangers.(38) In
certain applications, especially in converting gas- or oil-
fired boilers to coal firing, it is advantageous or
necessary to initiate combustion external to the boiler
and to remove slag from the combustion gas prior to en-
try into the boiler. For boilers designed for oil or gas fir-
ing the use of coal would result in fouling of heat
transfer surfaces and plugging of the boiler gas
passages, rendering the boiler inoperable in a short time
period. The benefits which accrue from the retrofit of
coal-fired boilers are greatly reduced boiler tube
wastage, NOX and sulfur dioxide emission reduction.
Figure 3-14. NOX Reduction by MACT for a 600-MW Coal/
Oil Boiler (37).
100
m
o
CD
O
80
60
40
20
Coal consumption fixed over load range.
Oil/Coal = 90/10
Non-MACT Operation
150 200 300 450
Boiler Load, MW
600
and greater efficiency that would result from reduced
coal flyash deposits on boiler heat transfer surfaces.
In slagging combustion the initial combustion process is
carried out quickly and under fuel-rich conditions so
that formation of fuel NOX is minimized.
Sulfur oxides can be absorbed in this stage usually by
addition of a calcium-based sorbent. The sorbent reacts
with the S02 to form solids which are removed with slag
or collected in the paniculate emission control device.
Provided a brief cooling period is allowed before the gas
enters the boiler, overfire air can be added to complete
combustion without significant formation of thermal
NOX.
Several configurations of slagging combustors have
been tested at a pilot scale where the radiant heat
transfer of a boiler has been simulated. Results of some
tests are shown in Figures 3-15 and 3-16 on different
configurations. Note in Figure 3-15 that the low
Figure 3-15. NOX Emissions from a Slagging Combustion
Pilot Plant Firing Coal (39).
0 500
S 400
|.300
Q
| 200
a
0 100
0
U.S. NSPS-Bituminous Coal
" U.S. NSPS — Subbituminous Coal
Average Data:
• Burner exit
• Boiler exit (plotted
at burner exit
• stoichiometry)
A • LNSB Goal
200-
3
CL
_C
O)
iooc-
A
i i i i
Q.
-0.4 |
m
-0.2 |
-0.1
.4 0.5 0.6 0.7 0.8
Burner Stoichiometric Ratio, Air/Fuel
31
-------
Figure 3-16.
350
INOx Emissions from a Slagging Combustion
Pilot Plant Firing Coal, Oil, and Mixtures (38).
• Coal
OOil/Coal (1/3 Weight Ratio)
DON/Coal (1/1 Weight Ratio)
• Oil
0.8 1.0 1.2
First Stage Equivalence Ratio
N0x/S0x Burner (LNSB) program goal of 100 ppm was
not realized in all tests, but results were well below the
NSPS levels. In Figure 3-16, the minimum NOX of about
100 ppm was realized at a stoichiometric ratio of about
0.8 for all fuels (in the primary combustion stage). Fur-
ther pilot-scale research is underway to test the ability
of slagging combustors to simultaneously reduce S02
emissions by sorbent injection. This technology is a
long way from commercialization as full-scale tests will
not be completed for several years. TRW is presently
performing tests on an industrial boiler.
3.2.1.9 Dry Flue Gas Treatment Technologies
Two major dry FGT processes — the selective non-
catalytic reduction process and the selective catalytic
reduction process — are still in the development stage
in terms of their application to coal-fired boilers in the
United States. Selective catalytic reduction has been
applied extensively in Japan and is being planned for
several European facilities. Figure 3-17 shows results for
a Japanese utility boiler (175 MW). NOX removal has
been a steady 80 percent and higher. Note that max-
imum NOX reductions were found at an NH3/NOX mole
ratio of 1.0 with little ammonia slip. Also, NOX reduc-
tions averaging about 90 percent were found in EPA
pilot-scale tests at a Georgia Power Company
boiler. (40) The pilot plant treated a 1,500 scfm
slipstream of flue gas from a 60-MW boiler. It was noted
that when the NH3/NOX ratio was below 1.0, the NOX
reduction achieved dropped to well below 90 percent.
A number of major concerns need to be resolved before
this process can be considered totally commercial for
coal-fired boilers in the United States. These are:
• Ammonia slip
• Sulfate and bisulfate formation and deposition
• Catalyst durability, cost, and reliability
• Ammonia and NOX control systems.
Ammonia slip refers to that portion of the ammonia that
passes through the boiler without reacting with nitrogen
oxide. Once excess ammonia has "slipped" through the
NOX control system, it can react with sulfur trioxide to
form ammonium bisulfate. This compound is a par-
ticulate that can foul the combustion air preheater and
could conceivably, in extreme cases, jeopardize par-
ticulate compliance. Therefore, ammonia slip must be
controlled; accordingly, Japanese utilities employing
SCR on their boilers usually specify strict limits for their
vendors on the amount of ammonia permitted to slip.
Some Japanese boilers continuously monitor for am-
monia in the flue gas.(14)
There are cost data and calculations available for SCR
as applied to coal-fired boilers. One example, shown in
Figure 3-17. Test Results for SCR on a Coal-Fired Boiler (2).
>; 100
.1
o
c 80
.2
I 60
350
300
250
^ 200
§
I 150
O
n
z
O 100
50
Unit load = full load
Flue gas temperature = 340C
3% 02 concentrations assumed
-Inlet
Outlet NOX
Outlet NH,,
0.7 0.8 0.9 1.0
NH3/NOxMole Ratio
1.1
32
-------
Figure 3-18, is the projected cost of an SCR retrofit for a
300-MW tangential-fired boiler relative to other NOX
control technologies. Note the distinct cost disadvan-
tage, but nevertheless, the NOX removal efficiency ad-
vantage represented by SCR compares to other
technologies employed more frequently for tangential-
fired boilers.
Selective noncatalytic reduction, as represented by the
Exxon Thermal DeNOx process, reportedly costs less
than SCR but NOX removal is significantly lower. Table
3-6 presents vendor-supplied data on capital cost and
cost-effectiveness for the Thermal DeNOx process. The
costs, which are in 1983 dollars, were developed by the
vendor for a hypothetical, 500-MW boiler. The design
NH3/NOX mole ratio was 1.5.
Table 3-6. Capital Cost and Cost-Effectiveness of Thermal
DeNox Process (48) (1983 dollars)
Table 3-7. Economic Evaluation of SCR for Coal-Fired
Utility Power Plants: 80% NOX Removal (42)
NOx"
300 ppm
600 ppm
Investment
$/kW
19
25
Cost Effectiveness
mills /kWh $/lb NOX Removed
2.03 0.57
2.94 0.41
"Initial (before SCR) NOX concentration at 3% 02.
An EPA-sponsored economic evaluation for coal-fired
utility power plants included SCR for NOX control.(42)
The SCR systems consist of two trains of insulated
reactors. An ammonia storage and handling system in-
jects an ammonia/air mixture in the inlet duct. The
catalyst life is assumed to be 1 year. Table 3-7 sum-
marizes the capital investment and annual revenue re-
quirements.
The most important capital and annual cost is for the
catalyst. Other than the catalyst, the main factor affec-
ting NOX control costs is the flue gas volume which
determines the fan and ductwork costs and the catalyst
volume.
Coal MW
East 200
East 500
West 200
West 500
West 200
West 500
S02 Control
Limestone FGD
Limestone FGD
Lime Spray Dryer
Lime Spray Dryer
Limestone FGD
Limestone FGD
Capital
Investment
$/kW"
103.1
83.7
121.0
100.2
121.6
96.1
Annual
Revenue*
mills /kWh
8.8
8.0
10.6
9.6
10.1
9.0
"1982 dollars.
b1984 dollars.
3.2.1.10 Wet Flue Gas Treatment
Many processes have been proposed for simultaneous
removal of SO* and NOX. However, there are no data on
their effectiveness for full-scale coal firing. Limited
testing by the Pittsburgh Energy Technology Center has
shown that NOX removals of 60 to 70 percent are
achievable along with 90 percent removal of SOX.(43)
No cost data on wet flue gas treatment processes for
actual coal-firing installations are available. Model plant
cost estimates are available for several processes.
3.2.2 Stoker-Fired Boilers
The primary NOX reduction technologies that have been
applied to stoker-fired boilers are overfire air control and
flue gas recirculation. The former has traditionally been
employed to reduce smoke emissions from stoker-fired
boilers while the latter has recently been proposed for
the same purpose but has the added benefit of reducing
NOX emissions. Exhaust gas recirculation reduces fuel
NOX by lowering oxygen concentrations in the bed
where fuel nitrogen is evolved.
Figure 3-18. Incremental Capital Cost of SCR and Other NOX Control Technologies for a New Tangential-Fired Boiler (26).
x
o
Q.
Q.
20
15
10
(80-90)
(40-60)
(45-55)
(15-20)
(20-30)
T-
Firing
Base
T-Firing
w/OFA
LNCFS
w/OFA
PM
w/OFA
MACT
w/PM
&OFA
SCR
Baseline
NOX removal in parentheses.
T-Firing Tangential firing
OFA Overfire air
LNCFS Low NOX concentric firing system
*PM Pollution minimum burner
*MACT Mitsubishi advanced combustion technology
*SCR Selective catalytic reduction system
*C-E/MHI licensed technology
Based on material and construction costs for a new
300 MW coal-fired unit.
SCR system costs are based on 80% NOX removal.
SCR system costs do not include NH3 tank farm
equipment or external structural support steel.
These cost approximations may change depending
on specific unit design requirements.
33
-------
3.2.2.1 Overfire Air
NOx reductions of between 10 and 25 percent have
been reported for employing additional overfire air ports
(44); existing overfire air ports are designed primarily for
smoke control and may not be optimally located for
control of NOX emissions. This method is severely
limited in that redirecting a large portion of air from the
bed to the overfire ports reduces the cooling effect of
the air on the grate and disturbs the natural staged com-
bustion of the stoker.
3.2.2.2 Flue Gas Recirculation
Flue gas recirculation, also called stoker gas recircuta-
tion in this instance, has been suggested for application
to stoker-fired boilers. Full-scale testing of stoker-fired
boilers is reported (45) in which slight NOX reductions
were achieved for flue gas recirculation in spreader
stokers. The boiler, rated at 100,000 Ib/h of steam, fired
Western Kentucky coal at excess oxygen levels of 8 to
10 percent. It was found that in addition to paniculate
emission reduction (about 40 percent) and efficiency in-
creases (6 to 7 percent), a side benefit was reduction in
NOX emissions. Since the recirculated gas is injected
along with the primary combustion air, the technique
results in lower excess air which in turn reduces both
fuel NOX and thermal NOX. Figure 3-19 shows test
results indicating that NOX emission levels approaching
100 ppm (3 percent oxygen basis) are achievable. Note
that NOX emissions for a stoker-fired boiler are relatively
low to begin with; note further that the modified (flue
gas recirculation) system results in lower excess oxygen
rates and hence reduced NOX emissions.
No cost data for either overfire air or flue gas recircula-
tion for stoker-fired boilers are available.
3.3 Oil- and Gas-Fired Boilers
Many of the techniques discussed above for reducing
NOX emissions from coal-fired boilers are also applicable
to oil-fired and gas-fired utility and industrial boilers of
comparable size. Most of the techniques that have been
used are those that reduce peak flame temperatures. A
discussion of available data on the effectiveness and
cost of these techniques follows.
3.3.1 Low Excess Air
Modest (average 11 percent) NOX reductions can be
achieved by lowering excess oxygen to the vicinity of
2.5 percent for combustion of residual and distillate
oils.(44) A side benefit is an increase in boiler efficiency;
however, a possible detriment is increased emission of
carbon monoxide, hydrocarbons, particulates, and visi-
ble plume.
The reductions achievable for gas-fired, water-tube
boilers are also modest, in the range of 3 to 15
percent. (44) However, the technology is desirable
because it is easy to implement, can be combined with
another technology, and again has the added benefit of
increasing boiler efficiency. The excess oxygen can be
safely reduced to about 2 percent in most boilers
without jeopardizing carbon monoxide and paniculate
compliance. Figure 3-20 shows the results of tests of a
110-MW gas-fired utility boiler, where NOX emissions
were reduced by over 30 percent by lowering oxygen in
Figure 3-20. IMOX Reduction by Lowering Excess Air in a
Gas-Fired Utility Boiler (46).
11,000
10,500
10,000
Figure 3-19. NOX Reduction Due to Flue Gas Recirculation
in a Stoker-Fired Boiler (45).
500
Q
Q.
Q.
400 -
300
O 200 -
100
6 8
Excess Oj, % Dry
Q
Q.
Q.
8
v>
c
x
O
260
240
220
200
180
160
140
0.5 1.0 1.5
02 in Flue Gas, 9
2.0
2.5
34
-------
the flue gas from 2.2 to 0.6 percent (corresponding to
about 2.8 percent excess air).(46) At oxygen levels in
flue gas of about 0.6 percent and lower, flame instability
and combustibles in the flue gas were found, making
this a practical lower limit for reducing excess air on that
boiler.(46) Any reduction in excess oxygen should be
made only after evaluating a specific application and
determining that a potential safety hazard will not arise
at the lower oxygen level.
3.3.2 Burners Out of Service
This technology is only available as a retrofit and ob-
viously only applies to boilers with more than one
burner. Some utility and industrial boilers burning
residual fuel oil or natural gas have used this
technology. NOX reductions range from 10 to over 40
percent (44) and of course are highly dependent on
whether or not the BOOS pattern tested is the optimum
pattern. Figure 3-21 shows results of testing for a
110-MW gas-fired utility boiler in which BOOS was
tested in combination with opening of an overfire air
port to further enhance the BOOS effect. A potential
drawback to this technology is possible derating of the
boiler unless modifications to the fuel piping are made
so that the design heat input can be maintained after
retrofit. Heat rate may also be increased with BOOS.
3.3.3 Reduced Combustion Air Preheat
This technique requires extensive modifications to the
air and flue gas handling systems of a boiler and is ob-
viously applicable as a retrofit only for boilers employing
Figure 3-21. NOX Reduction by BOOS for a Gas-Fired Utility
Boiler (46).
100
90
fc-
Q 80
£ 70
50
Air only to one bottom burner
1.2% 02 in flue gas
100 MW
Air only to one top burner
1.1 % O2 in flue gas
95 MW
20 40
NOX Port Opening, %
60
80
combustion air preheaters. The best application for this
technology would probably be a new boiler where it is
necessary to include heat recovery equipment to
preserve the thermal efficiency of the boiler. The
technology has been tested on a limited basis for oil-
fired and gas-fired boilers.(44) The NOX reductions
demonstrated for firing residual oil ranged from 5 to 16
percent; results of only two tests for gas-fired boilers in-
dicate NOX reductions of perhaps 30 percent are possi-
ble with a combustion air temperature decrease of
about 50C OOF). Unless alternate heat recovery equip-
ment is used to minimize stack losses, this technique
usually produces unacceptable efficiency penalties.
3.3.4 Flue Gas Recirculation
As with reduced combustion air preheat, flue gas recir-
culation is difficult to retrofit because of extensive
modifications to air and flue gas handling systems. The
amount of flue gas that is typically recycled is in the
range of 15 to 20 percent, although levels as high as 45
percent have been used for gas firing.(44) The propor-
tion is limited in oil-fired boilers by the onset of flame in-
stability; this is less of a problem with gas-fired boilers if
other than ring burners are employed.
Testing of this technology for oil-fired and gas-fired
boilers has not been extensive. However, preliminary in-
dications are that substantial NOX reductions can be
achieved. For example, tests for residual oil firing show-
ed NOX reductions of 15 to 30 percent while results for
firing distillate oil (one boiler only) ranged from 58 to 73
percent. Tests of three gas-fired boilers showed NOX
reduction ranging from 48 to 86 percent.
3.3.5 Low NOX Burners
Several burner manufacturers have for some time been
engaged in developing and providing low NOX burners
for oil- and gas-fired boilers. In some instances the
burners thus developed are modifications of pulverized
coal burners in which the basic principles of air or fuel
staging have been employed. Many of these low NOX
burner designs have been licensed for sale in Japan
where operators of oil- and gas-fired utility and in-
dustrial boilers in general face more stringent NOX emis-
sion regulations than their U.S. counterparts.
Babcock & Wilcox has developed the Primary Gas-Dual
Register Burner (PG-DRB) for oil-fired and gas-fired
boilers. In this burner—which is a modification of the
coal-firing DRB—the primary air zone is surrounded by
a primary gas recirculation zone. Here, recirculated flue
gas shields the flame and reduces oxygen availability.
The company claims NOX emission reductions of 65 per-
cent for oil and 75 percent for gas firing, as compared to
conventional circular burners.(41) The field experience
with this type of burner includes utility as well as in-
dustrial boilers, primarily in Japan. There are 114 PG-
DRB's firing oil in four generating stations in Japan
ranging in size from 175 to 600 MW; 176 PG-DRB's fire
gas in six stations ranging in size from 175 to 1,000 MW.
The total capacity covered by these burners is 4,700
MW.
35
-------
Coen Company, Inc. supplies low NOX burners for oil or
gas firing in industrial boilers.(47) Two basic types of
staged air burners are offered: the front wall arrange-
ment that is especially suited for residual oil (high fuel
nitrogen); and the side wall arrangement that is well
suited for gas firing. The company claims NOX reduc-
tions of 25 to 30 percent for oil firing and 35 to 40 per-
cent for gas firing, employing these burners. The com-
pany literature reveals a total of 11 installations ranging
in size from one burner and 20,000 Ib / h of steam to four
burners and 400,000 Ib / h of steam that employ this type
of burner. The guaranteed or actual NOX emissions for
these units are 0.1 lb/106 Btu and less for natural gas,
and 0.2 to 0.3 lb/106 Btu for oil.(47)
3.3.6 Dry Flue Gas Treatment
Figure 3-22 shows results of tests of the Exxon Thermal
DeNOx process at full load on a 235-MW oil-fired utility
boiler in California. The figure shows that NOX removals
of over 50 percent were achieved at the full load condi-
tion. At partial load, the change in convective section
temperature reduced the NOX reduction efficiency to
below 50 percent. There was also significant ammonia
slip during these tests. Exxon reports that the ammonia
compounds formed from slippage do not cause unac-
ceptable corrosion or preheater fouling. Deposits form-
ed can be removed from the system by periodic
washing with water.(48) Exxon is continuing to develop
the process to improve efficiency.
Catalytic technologies have been tested and used ex-
tensively in Japan for coal-, oil-, and gas-fired industrial
Figure 3-22. NOX Reduction by SNR for Oil-Fired Utility
Boiler (48).
80
c
o
1
•o
I
o
60
40
20
1.0 1.2 1.4 1.6 1.8
NH3/Initial NOX Concentration
2.0
boilers. Processes with fixed beds and moving beds
have been tested. Performance data for actual
operating systems are available; some representative
results for oil-fired boilers are shown in Table 3-8. Per-
formance data for gas-fired boilers are also available in
which NOX removal efficiencies of 90 percent have been
realized over operating periods of several thousand
hours.(43)
An SCR system was tested that handles half of the flue
gas from a 215-MW gas-fired utility boiler of Southern
California Edison Company.(49) The system attained
the design goal of 90 percent NOX reduction in 18,000
hours of operation; however, it is not clear that the
Table 3-8. IMOX Removal in Oil-Fired Industrial Boilers with Selective Catalytic Reduction (43)
Operation Parameters of Major Plants
Operation Data of a Commercial SCR
Plant for Dirty Gas
SCR Plant by Mitsui
Engineering and jhipbuilding Co.
Completed
Plant site
Gas source
Capacity (NnWhr)
Load factor (%)
Pretreatment of gas
Reactor inlet
NOX (ppm)
SOX (ppm)
Dust (mg/Nm3)
02 (%)
Reactor type
Reaction temp.
NOx/NHj ratio
Catalyst No.
SV (hr1)
NOX removal (%)
Pressure drop by SCR
reactor (mm H20)
Catalyst life
11/75
Yokkaichi
Oil-fired boiler
440,000
50-100
EP", FGD,
heating
150
80-130
30-100
3.2
Fixed bed
420
1.0
304
10,000
80*
160
1 year
Gas for SCR (NmVhr)
Fuel
Load fluctuation (%)
Stack height (ml
Inlet gas composition
02 (%)
SOX (ppm)
NOX (ppm)
Particulates after EP
(mg/Nm3)
FGD unit
SV (rr'1
Temperature (C)
NOX removal (%)
NH3/NOxmol ratio
Leak ammonia (ppm)
Type of reactor
Plant completed
300,000
Oil IS = 0.7%)
60-100
140
6
400
200
10-20
Scheduled
5,000
320
Over 90
1.0
10-20
Moving bed
10/76
Capacity (NnWhr)
Gas composition
NOX (ppm)
SOX (ppm)
Dust (mg/Nm3)
Catalyst and reactor
Catalyst carrier
Catalyst shape
SV (hr1)
Temperature (C)
NrVNOxmol ratio
NOX removal (%)
Total pressure drop
(mm H20)
Leak NH3 (ppm)
Operation start
Plant cost (106 yen)
220,000
150
300
100-150
TiO
PP
4,000
350-400
1.0
Above 90
180
Below 10
7/77
260
"Electrostatic precipitator.
6lncluding leakage in heat exchanger.
Note: At 8/9/85 exchange rate of 237.70 yen per U.S. dollar.
36
-------
system can achieve the design of 10 ppm maximum am-
monia slip. During the test period the boiler was fired
with oil for periods of time. During these periods,
however, "heavy" deposits of ammonium bisulfate and
iron sulfate were found in the air preheater. At its worst,
this condition resulted in an increase in pressure drop
through the air preheater of 50 percent.(49)
The estimated cost for a Thermal DeNOx process for a
200,000 Ib/h oil- or gas-fired industrial boiler is
available.(49) Table 3-9 shows the estimated annual
operating costs for this application. The total invest-
ment cost is estimated to be $389,000. If the system is
assumed to have a useful life of 20 years, the annual
Table 3-9. Annual Operating Cost Estimates for Thermal
DeNOx on a 200,000 Ib/h Oil-Fired Industrial
Boiler (30) (1982 dollars)
Annual Unit Annual
Cost Item Consumption" Cost, $ Cost, $
Ammonia6
Electric powerc
Steamd
Maintenance
(Material and Labor)"
Total
164 t
55 MWh
3,890 t
220/t
50/MWh
13.50/t
36,080
2,750
52,515
7,560
98,905
"Assumes 65% load factor.
^Includes ammonia for direct injection.
""Includes power requirement for ammonia vaporizer.
dLow pressure steam (0.1 MP ag minimum) for carrier.
"Assumed to be 7% of direct investment cost.
capital-related cost would be $45,690 at a 10 percent in-
terest rate, and thus the total annualized cost of the
DeNOx system would be about $160,000, including a
small percentage of capital cost for insurance, taxes,
and administration.
Table 3-10 shows Japanese data (43) on capital and
operating cost of selective catalytic reduction systems
for oil-fired and gas-fired boilers. The smaller flowrate
would correspond to an industrial boiler while the large
flowrate is more typical of a utility boiler. Note the
economy of scale evident in the capital cost figures
while the operating cost is much closer to linear with
flowrate.
Cost data are also available (49) for an SCR system ap-
plied to gas firing. The data are in the form of 20-year
levelized costs and are for combustion of oil. The cost is
a strong function of the unit's capacity factor ranging
from greater than $0.0607 kWh at a capacity factor of 30
percent, down to $0.010/kWh at a capacity factor of 80
percent. (49)
Table 3-10.
Gas Flow
Rate, Nm'/h
Capital and Operating Cost Data for Selective
Catalytic Reduction Systems for Oil-Fired and
Gas-Fired Boilers (43) (1981 dollars)
Capital Cost, $
Operating
Cost, $
50,000
1,200,000
0.5 x 106
5.0 x 106
0.2 x 106
3.5 x 106
37
-------
-------
Chapter 4
Performance and Cost Data:
Packaged Boilers Firing Oil or Gas
4.1 Introduction
This chapter presents available data on the NOX removal
efficiency and associated cost of those technologies
employed for smaller, packaged boilers that fire oil, gas,
or both. Much of the material in the latter parts of
Chapter 3 pertaining to large industrial oil-fired and gas-
fired boilers also applies to packaged boilers. However,
many of those technologies have limited application to
smaller boilers. These limitations will be discussed in
this chapter.
Packaged boilers are for the most part industrial rather
than utility and are small. There has not been a great
deal of research and development of NOX control
technologies for these boilers. NSPS were proposed in
June 1984, but have not yet been promulgated. State
and local regulations that may apply usually only limit
emissions of particulates and sulfur dioxide.
Where low NOX technology has been required for
packaged boilers, the normal procedure has been to go
directly to a low NOX burner. On a new or retrofit basis,
low NOX burners are usually straightforward to imple-
ment (packaged boilers normally employ only one
burner) and relatively cost-effective compared to more
complicated technologies of lesser NOX reduction effi-
ciency. A few other technologies have received limited
attention, however, and these are discussed first.
4.2 Reduced Combustion Air Preheat
This technology is available for use with packaged
boilers of the water-tube design but, as with large
boilers, requires a combustion air preheater. Fewer than
20 percent of packaged boilers have preheaters; most
use economizers for heat recovery. Limited data show
NOX reductions in the range of 5 to 16 percent, which is
the same for industrial boilers in general.(44) To recover
lost thermal efficiency boilers equipped with this
technology (new or retrofit) require heat recovery
systems. Due to this severe energy penalty, this
technology is not expected to proliferate, especially in
view of the ease of implementing low NOX burners and
the much greater NOX reduction achievable.
4.3 Flue Gas Recirculation
Flue gas recirculation is a technology available for both
fire-tube and water-tube packaged boilers. It is a
technology that is being increasingly implemented on
packaged units. Tests have reflected a NOX reduction
ranging from 15 to 30 percent for residual oil, 58 to 73
percent for distillate oil, and 48 to 86 percent for gas-
fired boilers. (44) For gas-fired units the technique is
more effective with water-tube boilers than fire-tube
boilers.
4.4 Low NOX Burners
In Chapter 3 a discussion of low NOX burners for oil and
gas firing was presented as part of the material on utility
and large industrial boilers. In large part, that discussion
also applies here for packaged boilers.
Several burner manufacturers have designed low NOX
burners for oil and gas firing that may be used in
packaged boilers. Particularly applicable are the data
from Coen Company, Inc.(47) in which two basic types
of staged air low NOX burners are discussed. To
reiterate, these burners are claimed to be capable of
NOX reductions of 25 to 30 percent (oil) and 35 to 40 per-
cent (gas). Of the eleven installations mentioned, seven
are relatively small, one-burner, packaged units. This in-
formation is summarized in Table 4-1.
As pointed out earlier there are a number of burner
manufacturers offering low NOX burners for small oil-
and gas-fired boilers. Although performance data are
scarce, many manufacturers are agreeing to guaranteed
emission levels comparable to the rather stringent levels
of the proposed industrial NSPS.
Table 4-1. IMOX Emission Performance for Low NOX Burners
in Single-Burner Oil- and Gas-Fired Boilers (47)
Boiler Capacity, Ib/h
Fuel
Guaranteed or Actual
NOX Emission
20,000
68,200
80,000
55,000
30,000
100,000
50,800
Refinery or
natural gas
Natural gas or
No. 2 fuel oil
Refinery or
natural gas
Natural gas or
No. 2 fuel oil
Natural gas or
No. 2 fuel oil
Natural gas or
No. 2 fuel oil
Natural gas
70 ppm (@3% 02)
75 ppm (@3% 02)
0.15lb/10' Btu
0.09lb/106Btu
0.12lb/10" Btu
0.12 lb/106 Btu
0.08 Ib/ 10' Btu
Note: To convert emissions in lb/10' Btu to ppm for firing natural gas,
multiply by 833.
39
-------
EPA has developed a low NOX heavy oil burner that
generates no more than 75 ppm NOX regardless of the
fuel nitrogen content. This precombustion type burner
has been applied only to an enhanced oil recovery
steam generator rated at 60 x 106 Btu/hr. The precom-
bustor is operated fuel rich at very high temperatures to
take advantage of very low NOX equilibrium levels.
Burnout air is added at the precombustor outlet to com-
plete the combustion process at lower temperatures.
In addition, Alzeta Corporation, under contract to EPA,
has developed a "fiber" low NOX burner for gas-fired
fire-tube boilers. Flame temperatures are kept low
because a substantial portion of the heat is provided by
radiant transfer.
The fiber burner has been demonstrated on gas-fired
fire-tube boilers ranging in rating from 245 to 980 kW.
Results are shown in Table 4-2. Note that substantial
reductions in NOX emissions (on the order of 80 percent)
are possible at varying boiler loads. Earlier tests (51)
disclosed that 10 percent excess air was the optimum
operating point. Below this level, carbon monoxide
emissions were excessive and above this level, efficien-
cy was compromised. The carbon monoxide emissions
at 10 percent excess air were comparable to those for a
conventional burner.
4.5 Other Technologies
Several other technologies are available for packaged
boilers but have not been demonstrated. Reasons may
include doubts about cost-effectiveness, lack of interest
due to the availability of alternatives, or technical limita-
tions. Low excess air is available and readily im-
plemented. As with overfire air for stoker-fired boilers,
however, low excess air for packaged boilers has tradi-
tionally been employed for a reason other than NOX
control (in this case for fuel savings) with reduction of
NOX emissions as a side benefit.
Typically, an oxygen trim system has been used to
balance oxygen against carbon monoxide and smoke
emissions. Either oxygen or carbon monoxide in the flue
gas is monitored to approach the lowest practical ex-
cess air level while still complying with any applicable
regulations on particulates (smoke) and carbon monox-
ide. At this low excess air level (generally about 2.5 per-
cent), NOX emissions are reduced approximately 11
percent. (44)
Tests performed on a package boiler simulator and a
commercial fire-tube boiler, both rated at 0.73 MW or
2.5 x 106 Btu/hr, have shown that reburning can effec-
tively reduce NOX by 50 percent or greater with minimal
facility modifications.(19) However, for very low
primary flame NOX levels (less than 200 ppm), reburning
fuel nitrogen content is a limiting factor. Reburning with
a low nitrogen content fuel such as natural gas or
distillate oil may be necessary. Selective noncatalytic
reduction is offered commercially and has been applied
to several packaged boilers and enhanced oil recovery
steam generators in California.
Table 4-2. NO, Emission Reduction for Gas-Fired Fire-Tube Boiler with a "Fiber" Low NOx Burner" (50)
Site
Alzeta Lab
Santa Clara, CA
York-Shipley
Boiler Test Bay
York, PA
Vandenberg AFB
CA
Peter Paul Cadbury
York, PA
Hall Chemical Co.
Wickliffe, OH
Boiler
Size
kW
245
588
392
588
980
Burner Type
Conventional
Fiber
Conventional
Fiber
Conventional
Fiber
Conventional
Fiber
Conventional
Fiber
Excess
Air
%
16
10
15
10
55
10
20
7
7
13
CO
ppm
11
10
10
10
240
0
400
35
1,000
35
NO
ppm
57
10
NT
NT
NT
NT
NT
20
80
18
"All results at high fire conditions, emissions corrected to 0% 02-
60n segmented burner.
'Three different burners employed.
Note: NT = not tested
NA = not applicable.
HC
ppm
0
10
NT
NT
NT
NT
NT
NT
NT
NT
Boiler
Eff.
%
82.8
82.3
85.0
85.8
81.0
82.3
83.2
84.0
82.7
83.7
Hours of
Operation
NA
500*
NA
2,000
NA
3,800
NA
3,450
2,500
NA
3,450'
2,500C
1,500C
40
-------
Chapter 5
Performance and Cost Data:
Gas Turbines and Stationary Reciprocating Engines
5.1 Introduction
Available data on the effectiveness and cost of NOX
reduction technologies for turbines and engines are
presented in this chapter. With the exception of selec-
tive catalytic reduction technologies, the techniques
presented are unique to these sources and have not
been discussed in Chapters 3 and 4.
5.2 Gas Turbines
The two primary methods of NOX reduction that have
been employed for gas turbines are water/steam injec-
tion and selective catalytic reduction. The former
technology is the more advanced in the United States;
the latter has been used extensively in Japan for several
types of sources, including gas turbines, and it has been
used or tested at a few U.S. installations.
5.2.7 Water/Steam Injection
Nearly 50 percent of the gas turbines in California
employ water/steam injection for NO* reduction.(52) A
water injection system consists of nozzles mounted in-
side the combustor, a proportional controller to ensure
the correct injection rate, a pump, and the associated
piping. Typical water- or steam-to-fuel ratios are 0.21:1
to 1:1 for water and 1:1 to 2:1 for steam. Water for in-
jection must meet the purity requirements of boiler
feedwater in order to avoid corrosion in the turbine. (24)
The reference also reports that for water-to-fuel ratios in
the range of 1:1, NOX reductions up to 75 percent are
realized. This is basically confirmed by data shown in
Figures 5-1 and 5-2. (53) Increased (less than 5 percent)
fuel consumption is a drawback.
Water/steam injection is only effective in reducing ther-
mal IMOX. Therefore, it is not recommended for turbines
combusting fuels with significant fuel-bound nitrogen
such as coal-derived liquid, shale oil, and high-nitrogen
residual oil.
A by-product of water/steam injection is a modest (1 to
11 percent) increase in power output but a slight
decrease in turbine efficiency.(53) It is difficult to
generalize on turbine wear and maintenance problems
because the studies (53) do not indicate any discernible
trend. Suffice it to say that increased turbine wear and
more frequently required maintenance may be conse-
quences of water/steam injection in rare instances. The
formation of ice fog could be a deterrent to the use of
this technology in populated areas located in cold
climates, such as Alaska.(53)
Figure 5-1. NOX Reduction by Water or Steam Injection:
Gas Turbine Firing Natural Gas (53).
3 50 -
cc
O
0.4
0.6 0.8 1.0
Water-to-Fuel Ratio, Ib/lb
1.2
1.4
Figure 5-2. NO* Reduction by Water or Steam Injection:
Gas Turbine Firing Distillate Oil (53).
0.2 0.4 0.6 0.8 1.0
Water-to-Fuel Ratio, Ib/lb
1.2
1.4
41
-------
Another possible disadvantage is an increase in emis-
sion of carbon monoxide. For turbines operating at base
load with water-to-fuel ratios of 1:1, CO emissions are
approximately 300 ppm, which is about five times the
baseline level.(52) However, for these units operating in
California, the carbon monoxide emissions are well
below the limit of 2,000 ppm. (These concentration
levels assume a 3 percent oxygen basis.) In any case it is
difficult to generalize on this point because tests are not
consistent with turbine size and fuel type.(53)
5.2.2 Selective Catalytic Reduction
Extensive testing of selective catalytic reduction has
been carried out in Japan for actual operating gas tur-
bines. Figure 5-3 presents data for the Kawasaki Power
Station Number 1 combined cycle turbine which is rated
at a total of 141 MW (97 MW prime turbine plus 44 MW
steam turbine). Note that at all turbine loads greater
than 25 percent the NOX reduction efficiency is well over
80 percent and ammonia emissions—a potential
drawback to use of SCR—are low. At the time of the
site visit, the SCR catalyst had performed through
11,000 hours of operation including daily start-ups.(54)
Seven larger (370 MW to 1,000 MW) gas turbines that
employ selective catalytic reduction for NOX removal are
currently in construction or start-up phase in Japan with
no results available to date. Most, if not all, of these
units are designed for an outlet NOX concentration of 15
ppm at 15 percent oxygen.
Figure 5-3.
Performance of Selective Catalytic Reduction
on a Gas Turbine (54).
. o
u
» o £
C O 0.
350
250
100
;
80
80
3
o
*
o
O
0*
z
Q.
D.
20
0
20
z 25
•Actual 02
NH3/NOX = 1.03
Ambient temp = 15C
50 75
Peak Load, %
100
Data are also available on the variation of NOX reduction
with space velocity, gas temperature, and NH3/N03
mole ratios. These data are also for Japanese installa-
tions.
Cost data are available (54) for selective catalytic reduc-
tion units serving four large (General Electric "Frame
6") gas turbines (total rating approximately 300 MW). In
1982 dollars the total capital cost would be about
$2,650,000. The operating cost, including all direct and
indirect costs as well as the incremental fuel cost, is
estimated to be a maximum of $0.00125 per kWh. In
order to employ SCR, a means of cooling the exhaust
gas must be provided. The temperature range required
is from 300 to 400C (570 to 750F), depending on
catalyst.(53)
Efforts are underway in Japan and the United States to
develop new combustor designs for turbines that will
reduce NOX emissions without water/steam injection or
use of SCR. Both catalytic combustion and staged
combustion techniques have been developed. EPA has
developed the rich-burn quick quench (RBQQ) concept
combustor system that uses staging to reduce both
thermal and fuel-bound NOX emissions.(55) Testing so
far has been limited and performance data for actual
operating units are not available.
5.3 Stationary Engines
Several methods for reducing NOX emissions from sta-
tionary engines have been employed or at least tested.
They include: water/fuel emulsions (diesel engines
primarily), exhaust gas recirculation (diesel), lean
burning-torch ignition (spark-ignition engines only), tur-
bocharging, charge cooling (spark), ignition retard
(spark), injection retard (diesel), selective catalytic
reduction (spark), and nonselective catalytic reduction
(spark).
5.3.1 Water/Fuel Emulsion
EPA has conducted demonstration tests on the use of
water/fuel emulsions to reduce NOX emissions from
diesel engines. Results for a four-stroke, six-cylinder
turbocharged diesel engine with a generator output
rated at 165 kW are shown in Figure 5-4.(56) NOX reduc-
tions of about 60 percent were consistently achieved
over a wide range of loads. The load range also
represents a range of water-to-fuel ratios of 0.27 to
0.60.
An important part of emulsion research involves in-
vestigation of the side effects of NOX reduction through
use of water/fuel emulsions. Of concern in this
research were emissions of carbon monoxide, hydrocar-
bons, particulates, and sulfur oxides. Carbon monoxide
emissions were found to increase dramatically at low- to
mid-range loads and then to decrease to a level of about
20 percent above baseline at full load.(56) A similar ef-
fect was noticed for hydrocarbons except that at about
85 percent of load the hydrocarbon emissions dropped
below the baseline to a value of about 60 percent of the
baseline at full load. Particulate emissions were also
higher with NOX control dropping rapidly from over 2.5
42
-------
Figure 5-4. NOX Reduction by Water/Fuel Emulsion for
Diesel Engine.
1.4
1.2
1.0
§ 0.8
O
z
0.6
0.4
— Baseline
"••• Control effect
• Catalyst
• Emulsion
A Combined
recirculation. Tests reported for the same engine
described above were carried out by water scrubbing
the exhaust gas from the turbocharger (to remove par-
ticulates) and returning a portion to mix with the incom-
ing combustion air.(57) The optimum value was found
to be about 15 percent exhaust gas by volume in the in-
let air, at which NOX reductions were approximately 50
percent. Higher reductions were obtained but the fuel
penalty was excessive and smoke emissions also
became a problem at higher proportions of exhaust gas.
Fuel penalty was reported in BSFC, or brake-specific
fuel consumption, which has units of Btu per brake
horsepower-hour. Test results are shown in Figure 5-6.
Data have also been collected for exhaust gas recircula-
tion for large-bore spark engines firing natural gas. (58)
Table 5-1 shows data correlating NOX emission reduc-
tion with the amount of exhaust gas recirculated. The
BSFC penalty in these tests ranged from 1 to 3 percent.
5.3.3 Lean-Burning Torch Ignition
Another technique that has been used to reduce NOX
emissions from spark-ignition engines is lean-burning
torch ignition. Use of torch ignition extends the lean-
burning limit of the engine. Figure 5-7 shows test results
(57) for a single-cylinder, two-stroke, gas-fired engine
rated at 681 hp at 330 rpm. Note that the use of torch ig-
nition extended the lean-burn limit to a fuel/air
equivalence ratio of 0.55 to 0.60. (This ratio is the
percentage of stoichiometric.) As a result, NOX emis-
sions can be reduced by 50 percent or more by this
0.2'
I
40
80
Load, kW
120
160
times the baseline at zero load to less than baseline at
about 25 percent of load. At higher loads particulate
emissions stayed below the baseline level. Sulfur diox-
ide emissions were found to be essentially constant in
that they are fixed by the sulfur content of the fuel (0.24
percent in this case). Conversion of sulfur dioxide to
sulfate is consistently low except at zero load.
Performance data for water/fuel emulsions with diesel
engines are also available.(57) The test engine was a
single-cylinder, four-stroke turbocharged diesel engine,
rated at 350 hp at 1,000 rpm. Water content in fuel rang-
ed from 7 to 45 percent by volume. Both high energy
and low energy emulsifiers were tested with similar
results. Figure 5-5 shows test results which indicate
NOX reductions of about 35 to 40 percent from a
baseline emission level of approximately 10.4 g/bhp-h
(Btu per brake horsepower-hour, units used for state
and local regulations for engines). Losses in efficiency
were 4 percent or less in these tests and increased with
water/fuel ratio.
5,3.2 Exhaust Gas Recirculation
Another method for NOX reduction that has been tested
primarily for large-bore diesel engines is exhaust gas
Figure 5-5. NOX Reduction by Water/Fuel Emulsion for
Diesel Engine (57).
1.4
1.2
1.0
Z 0.8
ro
•35
ir
0.6
0.4
0.2
Emulsor
Bore, Power, Plunger,
mm
216
270
400
O Gear pump
$Gear pump
AWesthalea^
DHomogenizer 400
•Gaulin
AGaulin
*Gaulin
Hydroshear
280
280
280
280
hp
350
234
234
350
mm
24
24
20
24
0.2 0.4 0.6
Water/Fuel Ratio by Volume
0.8
1.0
43
-------
Table 5-1. NOX Reduction by Exhaust Gas Recirculation:
Natural Gas, Spark Ignition Engines (58)
% NOX Reduction
Figure 5^7. Effect of Lean-Burning Torch Ignition on Spark-
Ignition Gas-Fired Engine (57).
Approximate
% EGR
12
17
20
Pump-Scavenged
Engines
20
35
40
Blower-Scavenged
Engines
20
30
40
method. Note in Figure 5-7 that slight increases in
hydrocarbon emission also result from this modifica-
tion. Factory test data are available that show an NOX
reduction of 80 percent and higher for new engines
modified to use lean-burning torch ignition in the form
of "jet-cell igniters."(59) Operation at fuel/air
equivalence ratios of less than 0.6 may affect tur-
bocharger performance somewhat because the exhaust
gas now has comparatively less energy content;
however, reducing emissions to 2 g/hp-h only involves
a fuel penalty of approximately 2 percent.(58) Note in
Figure 5-6. NOX Reduction by Exhaust Gas Recirculation
for Diesel Engine (57).
Timing: 21,5°btdcf 1 24°btdc
1.0
.55
0.60 0.65 0.70
Fuel/Air Equivalence Ratio, $
0.76
8 12 16 20
Exhaust Gas Recirculation, %
24
Figure 5-7 that NOX emissions are given in grams per
brake horsepower-hour.
5.3.4 Charge Coo/ing
Charge cooling, sometimes referred to as charge
refrigeration, is another technique that has been applied
to spark-ignition engines, usually on a retrofit basis.
Test results are available for this technique as applied to
the earlier described engine.(57) Again, the NOX emis-
sions were measured at various fuel/air equivalence
ratios but also at incoming fuel/air temperatures rang-
ing from a baseline of 43C (110F) down to 2C (35F).
Figure 5-8 shows the results. Note that for a given
fuel/air ratio, NOX reductions of about 40 percent were
observed for charge cooling to 2C (35F). The side
effects of charge cooling are the same as those for
operation at leaner fuel mixtures, namely, the ignition is
further delayed and the duration of combustion is ex-
tended. In order to combat these conditions which
could lead to misfiring, it may be advisable to combine
charge cooling with torch ignition or use of high energy
sparking. Temperature drops to less than 5.5C (10F)
above ambient would require extensive and expensive
retrofitting including refrigeration.(60) Thus, if such a
close temperature approach is required to meet a NOX
emission level, this technology would not be cost-
effective.
44
-------
Figure 5-8. NOX Reduction by Charge Cooling for a Spark-
Ignition Gas-Fired Engine (57).
0.60 0.65 0.70 0.75 0.80
Fuel/Air Equivalence Ratio, 4
0.85
5.3.5 Ignition Retard
Data are available on NOX reductions in spark-ignition
engines resulting from retarding ignition timing, which
is used almost exclusively on a retrofit basis. Table 5-2
summarizes results for several types of natural gas-fired
engines.(58) Typical NOX reductions achievable range
from 15 to 30 percent. However, the fuel consumption
penalty is not negligible (up to 3 percent).
5.3.5 Turbocharg/ng
Limited data are available on the NOX reduction effect of
turbocharging a natural gas-fired engine. (58) For tur-
bocharging to 15 in Hg, NOX reductions were 55 percent
for pump-scavenged engines and 45 percent for blower-
scavenged engines, each with aftercooling to 38C
(100F). Also, fuel consumption decreases between 10
and 13 percent were reported with a corresponding
power increase of between 25 and 33 percent.
Table 5-2. NOX Reduction by Ignition Retard: Natural Gas-
Fired Engines (58)
Type of Engine
Retard, ° NOX Reduction, %
Pump-scavenged, atmospheric*
Blower-scavenged, atmospheric*
2-stroke, turbocharged
4-stroke, turbocharged,
medium pressure
4-stroke, turbocharged,
high pressure
4
4
4
5
5
15
25
30
17
25
*Not turbocharged.
5.3.7 Postcombustion Controls
In addition to during-combustion technologies for con-
trolling NOX emissions, both selective and nonselective
catalytic reduction technologies have been employed
for stationary engines. Most of the applications of these
after-combustion technologies have been in Japan and
in California.
Long-term (2,300 h) testing of SCR on a diesel engine
has been accomplished. The engine tested was rated at
165 kW with a displacement of 10.5 I (638 in3) and fired
No. 2 diesel fuel. The results, shown in Table 5-3, are
for extensive testing at an 80 kW load and at various
NH3/NO injection rates. Note that 90 percent reduction
(from a baseline of about 400 ppm) was achievable at
reasonable NH3 injection rates until sometime before
2,000 hours of operation by which time catalyst perfor-
mance had declined significantly. There was, however,
significantly improved NOX reduction after catalyst
cleaning.(61) Fouling of catalyst by diesel particulate
was indicated as one possible factor in catalyst deac-
tivation.
SCR has also been tested for spark-ignition engines.(57)
The engine was a six-cylinder model rated at 1,350 hp at
330 rpm and burning natural gas. Results are shown in
Figure 5-9. Note that reductions of 95 percent were
found over a temperature range 38C (100F) wide. With
the proper choice of catalyst, 95 percent reduction
Table 5-3.
Cat. Hours
NOX Reduction by Selective Catalytic Reduction
for Diesel Engine (61)
NO/NO Base
NH3/NO Base NO, ppm
10-500 1.0
0.582
0.213
0.089
0.043
0.0173
0.0173
0.0173
1,000 1.0
0.586
0.471
0.371
0.214
0.143
0.077
2,000 1.0
0.635
0.541
0.486
0.446
0.378
0.292
0.243
2,300 1 .0
0.372
0.268
0.216
0.169
0.138
0.117
0.099
0
0.41
0.81
1.04
1.09
1.37
1.65
1.98
0
0.45
0.60
0.73
0.81
0.99
1.18
0
0.432
0.570
0.708
0.778
0.949
1.114
1.592
0
0.70
0.88
1.05
1.22
1.40
1.57
1.75
433
252
92
38
18.5
7.5
7.5
7.5
350
205
165
130
75
50
27
370
235
200
180
165
140
108
90
384
143
103
83
65
53
45
38
45
-------
Figure 5-9.
110
100 -
NOX Reduction by Selective Catalytic Reduc-
tion for Spark-Ignition Engine (57).
Table 5-4. Capital Costs for SCR Systems for Lean-Burn
Spark-Ignition Engines'* (24) (1984 dollars)
I
I
400 500 600 700 800
Catalyst Temperature, F
900 1000
could probably be achieved at any gas temperature
from 288C (550F) to 425C (800F). Figure 5-10 shows
results of extended tests for two catalyst-temperature
combinations. Note that if greater than 80 percent
reduction is required, catalyst life may be quite limited.
The California Air Resources Board has investigated the
cost of SCR for spark ignition (primarily lean-burn)
engines.(42) Table 5-4 shows some relatively recent
capital costs for several engines ranging in size from 660
to 4,000 hp. For units with manual controls for ammonia
rejection, the major cost item is catalyst; where
automatic ammonia controls are used, the control cost
may outweigh the catalyst cost, especially for engines
of less than 1,000hp.
Figure 5-10. NOX Reduction by Selective Catalytic Reduction:
Long-Term Test for Spark-Ignition Engine (57).
20,000 h-1
12.5% 02
NrVNO = 1.1
Engine Size, bhp
460
820
1,280*
2,500''
5,150*
SCR Catalyst
Cost, $
29,000
47,000
71,000*
130,000*
263,000*
Cost, $
NH3 Addition Controls
Manual
6,500
7,000
7,500
8,500
11,000
Automated
50,000
51,000
52,000
54,000
59,000
"Engelhard 2-stroke engines.
*For 4-stroke engines, use 65% of this cost.
Spark-ignition engines are also controlled by nonselec-
tive catalytic reduction (NCR) units, especially rich-
burning engines in California which are required to
reduce NOX emissions by 90 percent. Table 5-5 shows
data for NCR units applied to 13 rich-burn engines rang-
ing from 50 to 1,100 hp and firing natural gas or, in two
cases, digester gas.(62) NOX reduction efficiency after
4,000 h of operation was still in excess of 80 percent on
average.
Costs for NCR (in 1984 dollars), including catalyst and
air/fuel controls (manual or automatic), are shown in
Table 5-6.(24)
Table 5-5. NO* Reduction by NCR: Results for Tests of 13
Rich-Burn Spark-Ignition Engines (62)
No. of
Units
3
1
1
2
1
1
1
1
1
1"
Uncontrolled
Mean
404
Engine
Manufacturer
Waukesha
Waukesha
Waukesha
Waukesha
IHC
Climax
Climax
Caterpillar
Superior
Ingersoll-Rand
NOXLppmb Controlled
Std. Dev. Mean
221 36.4
Engine Characteristics
Rating, bho
83
71
51
818
50
500
300
130
500
1,100
NOX, ppm*
Std. Dev.
33.0
Loading
Cyclic
Steady
Steady
Variable
Cyclic
Variable
Variable
Steady
Steady
Steady
Average NOX
Reduction, %
91
"Fueled by digester gas; all others by natural gas.
bCorrected to 15% 02.
Table 5-6. Cost of NCR Systems for Rich-Burn, Spark-
Ignition Engines" (62) (1984 dollars)
Air/Fuel Controls
Engine Size NSCR Catalyst Cost, $ Dual or
40 60
Elapsed Time, h
100
bhp
70
280
585
1,170
Cost, $
2,000
6,200
11,400
19,700
Manual
1,300
1,300
2,000
2,000
Automatic
7,900
7,900
10,900
10,900
Single
Single
Single
Dual
Dual
"Engelhard Systems.
46
-------
Chapter 6
References
1. Mason, H.B. and M.A. Neither. Emission Inventory
for Stationary NOX Sources (USA). In Air Pollution
by Nitrogen Oxides, Elsevier Publishing Company,
Amsterdam, 1982.
2. Aoki, H., et al. Introduction of IHI-Denitrification
System for Coal-Fired Steam Generator. Presented
at the 1985 Joint Symposium on Stationary Com-
bustion NOX Control, Boston, MA, May 1985.
3. Perry, R.H. and D. Green (editors). Chemical
Engineers' Handbook. 6th Ed., Section 9, McGraw-
Hill, New York, NY, 1984.
4. Babcock and Wilcox. Steam/Its Generation and
Use. 38th Ed., Chapter 7, 1972.
5. Lisauskas, P.A. and A.H. Rawdon. Status of NOX
Controls for Riley Stoker Wall-Fired and Turbo-
Fired Boilers. Proceedings of the 1982 Joint Sym-
posium on Stationary Combustion NOX Control,
Volume 1: Utility Boiler Applications. EPA Report
no. 600/9-85/022 (NTIS no. PB85-235604/AS),
July 1985.
6. Kokkinos, A., et al. Low NOX Firing System for
Tangentially Coal-Fired Utility Boilers — Preliminary
Testing. Proceedings of the 1982 Joint Symposium
on Stationary Combustion NOX Control, Volume I:
Utility Boiler Applications. EPA Report no.
600/9-85/022 (NTIS no. PB85-235604/AS), July
1985.
7. Muzio, L.J., et al. Control of Nitrogen Oxides:
Assessment of Needs and Options. Technical Sup-
port Document, Volume 5: Emissions Control
Technology for Combustion Sources, Section 2,
EPRI EA-2048, July 1983.
8. Siddiqi, A.A. and J.W. Tenini. Coping with Air
Pollution. Hydrocarbon Processing, Volume 63,
No. 11, November 1984, pp. 123-129.
9. Yoshitoshi, Sekiguchi, et al. Development of New
NOX Combustion Control Method — Low NOX
Burner Based on a New Two-Stage Fuel Supply
and Combustion Methods. U.D.C. 661.5:662.9.
10. Vatsky, J. Foster Wheeler's Low NOX Combustion
Program Status and Developments. Proceedings of
the 1982 Joint Symposium on Stationary Combus-
tion NOX Control, Volume 1: Utility Boiler Applica-
tions. EPA Report no. 600/9-85/022 (NTIS no.
PB85-235604/AS), July 1985.
11. Barsin, J.S. Fossil Steam Generator NOX Control
Update. Proceedings of the 1982 Joint Symposium
on Stationary Combustion NOX Control, Volume 1:
Utility Boiler Applications. EPA Report no.
600/9-85/022 (NTIS no. PB85-235604/AS), July
1985.
12. Folsom, B., et al. NOX Emissions Control with the
Distributed Mixing Burner. Proceedings of the 1982
Joint Symposium on Stationary Combustion NO*
Control, Volume 1: Utility Boiler Applications. EPA
Report no. 600/9-85/022 (NTIS no. PB85-
235604/AS), July 1985.
13. Proceedings of the 1982 Joint Symposium on Sta-
tionary Combustion NO* Control. EPA Report no.
600/9-85/022 (NTIS no. PB85-235604/AS), July
1985.
14. Proposed Guidelines for the Control of Emissions
from Coal-Fired Power Plants. California Air
Resources Board, State of California, 1981.
15. Crawford, A.R., E.H. Manny, and W. Bartok. Field
Testing: Application of Combustion Modifications
to Control NOX Emissions from Utility Boilers. EPA
Report no. 650/2-74-066 (NTIS no. PB237-
344/7BA), June 1974.
16. Crawford, A.R., E.H. Manny, and W. Bartok. Con-
trol of Utility Boiler and Gas Turbine Pollutant Emis-
sions by Combustion Modification — Phase I, EPA
Report no. 600/7-78-036a (NTIS no.
PB-281-078/6), March 1978.
17. Natanson, P.S., E.H. Manny, and A.R. Crawford.
Evaluation of Tubewail Corrosion Rates on a Coal-
Fired Utility Boiler Using Staged Combustion for
NOX Reduction. EPA Report no. 600/7-83-054
(NTIS no. PB84-118-231/REB), October 1983.
18. Greene, S.B., S.L. Chen, W.D. Clark, M.P. Heap,
D.W. Pershing, and W.R. Seeker. Bench-Scale
Process Evaluation of Reburning and Sorbent Injec-
tion for In-Furnace NOX/SOX Reduction. EPA
Report no. 600/7-85-012 (NTIS no. PB85-185-
890/REB), March 1985.
47
-------
19. Mulholland, J.A. and R.E. Hall. Fuel Oil Reburning
Application for NO* Control to Firetube Package
Boilers. Presented at the Joint ASME/IEEE Power
Generation Conference, Milwaukee, Wl, October
20-24, 1985.
20. Mulholland, J.A. and W.S. Lanier. "Application of
Reburning for NOX Control to a Firetube Package
Boiler." Journal of Engineering for Gas Turbines
and Power, Volume 107, July 1985. pp. 739-743.
21. Breen, B.P. Nitric Oxide Control in Practical Appli-
cation / Temperature, Time, Heat Transfer, and
Kinetics. Central State Section — The Combustion
Institute, Spring Technical Meeting, March 1982.
22. Ando, J. Recent Developments in S02 and NO*
Abatement Technology for Stationary Sources in
Japan. EPA Report no. 600/7-85-040 (NTIS no.
PB86-110-186/AS), September 1985.
23. Arand, J.K., et al. Urea Reduction of NO x in Com-
bustion Effluents. U.S. Patent 4,208,306, June 17,
1980.
24. Proposed Rule 1110.0 — Emissions from Stationary
Internal Combustion Engines. South Coast Air
Quality Management District, El Monte, CA,
September 1984.
25. Eskinazi, D. An Overview of Retrofit NOX Controls
for Coal-Fired Utility Boilers. American Power Con-
ference, Chicago, IL, 1984.
26. McCartney, M.S. and M.B. Cohen. Techniques for
Reducing NOX Emissions from Coal-Fired Steam
Generators. Presented at the First International
Conference on Acid Rain — Regulatory Aspects
and Engineering Solutions, Washington, DC,
March 27-28, 1984.
27. Vatsky, J. and J.E. Cichanowicz. Economic and
Engineering Analysis of Arch-Fired Furnaces. Pro-
ceedings of the 1982 Joint Symposium on Sta-
tionary Combustion NOX Control, Volume 1: Utility
Boiler Applications. EPA Report no. 600/9-85/022
(NTIS no. PB85-235604/AS), July 1985.
28. LaRue, A.D. and L.W. Rodgers. Development of
Low NOX Cell Burners for Retrofit Applications.
Presented at the 1985 Joint Symposium on Sta-
tionary Combustion NOX Control, Boston, MA,
May 1985.
29. Kokkinos, A., et al. Feasibility Study of a Low-NOx
Retrofittable Firing System with U.S. Coals. Pro-
ceedings of the 1982 Joint Symposium on Sta-
tionary Combustion NOX Control, Volume 1: Utility
Boiler Applications. EPA Report no. 600/9-85/022
(NTIS no. PB85-235604/AS), July 1985.
30. Mulholland, J.A. and R.E. Hall. The Effect of Fuel
Nitrogen in a Reburning Application to a Fire Tube
Package Boiler. Presented at the 1985 Joint Sym-
posium on Stationary Combustion NOX Control,
Boston, MA, May 1985.
31. Kawamura, Y. Combustion Apparatus. Agency for
Industrial Science and Technology, Ministry of In-
ternational Trade and Industry, Tokyo, Japan, U.S.
Patent 4,389,186, June 21, 1983.
32. Miyamae, S., et al. Evaluation of In-Furnace NOK
Reduction. Presented at the 1985 Joint Symposium
on Stationary Combustion NOX Control, Boston,
MA, May 1985.
33. Greene, S.B.; S.L. Chen; D.W. Pershing; M.P.
Heap; and W.R. Seeker. Bench-Scale Process
Evaluation of Reburning for In-Furnace NOX Re-
duction. ASME Paper No. 84-HPGC-APC-9, 1984.
34. Kelly, J.T., et al. Fuel Staging for Pulverized Coal
Combustion NOX Control. Proceedings of the 1982
Joint Symposium on Stationary Combustion NOX
Control. EPA Report no. 600/9-85/022 (NTIS no.
PB85-235604/AS), July 1985.
35. Okigami, N., et al. Three-Stage Pulverized Coal
Combustion System for In-Furnace NOX Reduction.
Presented at the 1985 Joint Symposium on Sta-
tionary Combustion NOX Control, Boston, MA,
May 1985.
36. Seeker, W.R., et al. Controlling Pollutant Emissions
from Coal and Oil Combustors Through the Sup-
plemental Use of Natural Gas. Final Report, GRI
5083-251-0905, 1985.
37. Murakami, N. Application of the MACTIn-Furnace
NOX Removal Process Coupled with a Low-N0x
SGR Burner. Presented at the 1985 Joint Sym-
posium on Stationary Combustion NOX Control,
Boston, MA, May 1985.
38. Zanderer, B. and K.S. Fujimura. NOX Control in an
Air-Cooled Cyclone Coal Combustor. Presented at
the 1985 Joint Symposium on Stationary Combus-
tion NOX Control, Boston, MA, May 1985.
39. Dykema, O.W. Development of a Low NOX/SOX
Burner. Presented at the 1985 Joint Symposium on
Stationary Combustion NOX Control, Boston, MA,
May 1985.
40. Mobley, D.J. and J.M. Burke. Evaluation of NOX
and NOX/SOX Flue Gas Treatment Technology for
Coal-Fired Sources. Proceedings of the 1982 Joint
Symposium on Stationary Combustion NOX Con-
trol, Volume 1: Utility Boiler Applications. EPA
Report no. 600/9-85/022 (NTIS no. PB85-
235604/AS), July 1985.
41. Babcock and Wilcox. Primary Gas-Dual Register
Burner. Manufacturer's literature, undated.
42. Maxwell, J.D. and L.R. Humphries. Economics of
Nitrogen Oxides, Sulfur Oxides, and Ash Control
Systems for Coal-Fired Utility Power Plants. EPA
48
-------
Report no. 600/7-85-006 (PB 85-243 103/AS),
September 1985.
43. NO* Flue Gas Treatment. In Emission Control
Technology for Industrial Boilers, ed. by A.E. Mar-
tin, pp. 405, Noyes-Data Corporation, 1981.
44. NOX Combustion Modification. In Emission Control
Technology for Industrial Boilers, ed. by A.E. Mar-
tin, pp. 405, Noyes Data Corporation, 1981.
45. Maloney, K.L. Combustion Modification for Coal-
Fired Stoker Boilers. Proceedings of the 1982 Joint
Symposium on Stationary Combustion NOX Con-
trol, Volume 2: Flue Gas Treatment. Fundamental
Studies and Industrial, Commercial, and Residential
Applications. EPA Report no. 600/9-85-022b (NTIS
no. PB85-235612/AS), July 1985.
46. Bullin, J.A., and D. Wilkerson. NOX Reduction in a
Gas-Fired Utility Boiler by Combustion Modifica-
tions. Control Technology News, June 1982.
47. Coen Company, Inc. Introduction to COEN Low
NOX Burner Techniques. Manufacturer's literature,
undated.
48. Hurst, B.E. Technology Advances in Selective
Non-Catalytic NOX Reduction (Thermal DeNOx>.
Presented at the Sixth Annual Energy-Sources
Technology Conference and Exhibition, Houston,
TX, January 31, 1983.
49. Kerry, H.A. and A. Weir, Jr. Operating Experience
on Southern California Edison's 107.5 MW Selec-
tive Catalytic Reduction DeNOx System. Presented
at the 1985 Joint Symposium on Stationary Com-
bustion NOX Control, Boston, MA, May 1985.
50. Kesselring, J.P. and W.V. Krill. A Low-N0x Burner
for Gas-Fired Fire-Tube Boilers. Presented at the
1985 Joint Symposium on Stationary Combustion
NOX Control, Boston, MA, May 1985.
51. Kesselring, J.P. and W.V. Krill. Low-N0x Fiber
Burner Application to Gas-Fired Fire- Tube Boilers.
Proceedings of the 1982 Joint Symposium on Sta-
tionary Combustion I\IOX Control, Volume 2: Flue
Gas Treatment. Fundamental Studies and In-
dustrial, Commercial, and Residential Applications.
EPA Report no. 600/9-85-022b (NTIS no. PB85-
235612/AS), July 1985.
52. Yee, G., et al. Suggested Control Measure for the
Control of Oxides of Nitrogen Emissions from Elec-
tric Utility Gas Turbines. California Air Resources
Board, Stationary Source Control Division, October
1980.
53. Emissions from Stationary Gas Turbines New
Source Performance Standard. Radian Corporation
Draft Report prepared for U.S. Environmental Pro-
tection Agency, Research Triangle Park, NC,
November 1984.
54. Pease, R.P. Status Report on Selective Catalytic
Reduction for Gas Turbines. South Coast Air Quali-
ty Management District Engineering Division
Report, July 1984.
55. Mosier, S.A. and R. Pierce. Advanced Combustion
Systems for Stationary Gas Turbine Engines. EPA
Report no. 60077-80-017a, b, c, d (NTIS nos. Vol.
1, PB80-175599; Vol. 2, PB80-175607; Vol. 3,
PB80-177108; Vol. 4, PB80-179849, January 1980.
56. Wasser, J.H. Emulsion Fuel and Oxidation Catalyst
Technology for Stationary Diesel Engines. Pro-
ceedings of the 1982 Joint Symposium on
Stationary Combustion NOX Control, Volume 2:
Flue Gas Treatment. Fundamental Studies and In-
dustrial, Commercial, and Residential Applications.
EPA Report no. 600/9-85-022b (NTIS no. PB85-
235612/AS), July 1985.
57. Wilson, R.P., Jr. Emission Control Methods for
Large Stationary Engines: Part 1 — Combustion
Modifications for Spark Gas Engines; Part 2 —
Combustion Modifications for Diesel Engines; Part
3 — Catalytic NOX Reduction by Ammonia. Pro-
ceedings of the 1982 Joint Symposium on
Stationary Combustion NOX Control, Volume 2:
Flue Gas Treatment. Fundamental Studies and In-
dustrial, Commercial, and Residential Applications.
EPA Report no. 600/9-85-022b (NTIS no. PB85-
235612/AS), July 1985.
58. Retrofit Controls for NOX Reduction from Large-
Bore Natural Gas Engines. Acurex Corporation,
1985 (unpublished).
59. Danyluk, P.P. and F.S. Schaub. Emission Reduc-
tion by Combustion Modification in Two-Stroke,
Spark-Ignited Gas Engines and by Catalytic Conver-
sion. American Society of Mechanical Engineers,
81-DGP-7, January 1981.
60. Hibbard, J.W. Field Retrofits to Improve Efficiency
and Lower Undesirable Emissions. Presented at the
American Gas Association Transmission Con-
ference, Chicago, IL, May 17-19, 1982.
61. Wasser, J.H. and R.B. Perry. Diesel Engine NOX
Control with SCR. Presented at the 1985 Joint
Symposium on Stationary Combustion NOX Con-
trol, Boston, MA, May 1985.
62. Rawuka, A. NOX Reduction Program for Gas Fueled
Stationary Engines. Presented at the 77th Annual
Meeting of the Air Pollution Control Association,
APCA 84-42.8, June 24-29, 1984, San Francisco,
CA.
49
------- |