May 1980 EPA-600/8-80-027
GUIDELINES FOR NOX CONTROL BY COMBUSTION MODIFICATION
FOR COAL-FIRED UTILITY BOILERS
Procedures for Reduction of NOX Emissions and Maximization of Boiler Efficiency
Guidelines Intended
— As a Reference for Combustion Modification
Techniques and Procedures for IMOX Emission
Control
— For Use by Utility Personnel in Designing
Individual NOX Reduction and Boiler Efficiency
Optimization Programs
— For use by Boiler Manufacturers and Operators as
a Guide and Supplement to Operating Procedures
oEPA
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
rial Environmental Research Laboratory
Research Triangle Park, NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the SPECIAL REPORTS series. This series is
reserved for reports which are intended to meet the technical information needs
of specifically targeted user groups. Reports in this series include Problem Orient-
ed Reports, Research Application Reports, and Executive Summary Documents.
Typical of these reports include state-of-the-art analyses, technology assess-
ments, reports on the results of major research and development efforts, design
manuals, and user manuals.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
TABLE OF CONTENTS
Page
1.0 INTRODUCTION 1
2.0 BOILER INSPECTION AND MAINTENANCE 3
2.1 Furnace 3
2.2 Air Heaters 4
2.3 Pulverizers 4
2.4 Combustion Controls 5
2.5 Pulverized Coal Burners 5
2.6 Control Panel Instruments 6
2.7 Emission Measuring Instruments 7
3.0 OPERATING VARIABLES AFFECTING NO 8
x
3.1 Load 8
3.2 Excess Air 9
3.3 Staged Firing 9
3.4 Burner Tilt 10
3.5 Burner Registers 10
3.6 Flue Gas Recirculation 11
3.7 Pulverized Coal Fineness 11
4.0 STEAM TEMPERATURE CONTROL 12
4.1 Attemperation 12
4.2 Burner Tilt 13
4.3 Biasing Dampers 13
4.4 Flue Gas Recirculation 13
4.5 Exesss Air 13
4.6 Furnace Slag Blowers 14
5.0 OPERATING TEST DATA 15
6.0 EMISSION MEASUREMENTS 17
6.1 Comparing NO Emissions 17
a
7.0 CARRYING OUT THE NO REDUCTION PROGRAM 19
X
7.1 Test Program Plan 19
7.2 The ABC' s of Reducing NOX Emissions 21
7.3 Basic Steps in Reducing NO Emissions -
Step by Step Procedures 24
7.4 Analyzing the Results 30
7.5 Translating Optimized NOX Emission Into
Operating Procedures 31
iii
-------
TABLE OF CONTENTS (CONT'D)
Appendices Page
A. COMPREHENSIVE FIELD TEST PROGRAMS FOR NOX
EMISSION REDUCTION ON LARGE BOILERS 35
B. COAL CHARACTERISTICS ". . 49
C. FORMATION OF COMBUSTION-GENERATED POLLUTANTS 55
D. METHODS OF MEASURING POLLUTANTS AND OTHER
COMBUSTION PRODUCT EMISSIONS 64
E. BOILER EFFICIENCY 87
F. CONVERSION FACTORS 92
iv
-------
TABLES
Number Page
1. Major Operating Variables Affecting NO 8
X
2. Superheat and Reheat Steam Temperature
Control Features 12
3. Boiler Test Data 16
4. Experimental Test Plan 20
5. Operating Parameters for Low NO Operation 34
A-l. Test Program Plan (300 MW Boiler-Front Wall Fired) 36
A-2. Test Program Plan (800 MW Boiler - Tangential Firing
with Overfire Air) 41
B-l. Classification of Coals by Rank 50
B-2. Progressive Stages of Transformation of
Vegetal Matter into Coal 51
D-l. Location of Traverse Points in Circular Stacks 68
D-2. Techniques for Analyzing Gases and Vapors
in Flue Gases ^ 81
E-l ASME Test Form for Abbreviated Efficiency Test
- Summary Sheet ........... i 89
E-2 ASME Test Form for Abbreviated Efficiency Test
- Calculation Sheet ... 90
-------
FIGURES
Number Page
1. Fundamental Steps in Reducing NOX Emissions 23
2. Basic Steps in Reducing NOX Emissions 25
A-l. 300 MW Boiler-Front Wall Fired 40
A-2. Effect of Burner Tilt 45
A-3. 800 MW Boiler-Tangential Firing Staged Firing -
Overfire Air Dampers 100% Open 46
A-4. 800 MW Boiler-Tangential Firing Full Load-Staged Firing 47
A-5. 800 MW Boiler-Tangential Firing Effect of Overfire Air
Dampers on NOX Emissions Full Load-Staged Firing 48
C-l. Three Regimes of Coal Particle Combustion 56
D-l. Determination of the Minimum Number of
Traverse Points 66
D-2. Duct Cross Section 67
D-3. Circular Cross Section of Stack Divided
into 12 Equal Parts 67
D-4. Sampling Velocity and Potential Sampling Errors 70
D-5. EPA Method 5 Sampling Train Schematic 73
D-6. EPA Method 5 Sampling Train Illustration 74
D-7. S02 Profile 78
vi
-------
ACKNOWLEDGMENTS
This guideline was prepared by Mr. E. H. Manny of Exxon Research
and Engineering Company under EPA Contract Number 68-02-1415.
The author wishes to acknowledge the assistance and cooperation
of Mr. R. E. Hall, the EPA Project Officer, for his comments and guidance
during the preparation of the guideline. Special thanks are also extended
to members of the industry review committee for their many valued contri-
butions, recommendations and suggestions. Members of this committee were:
J. E. Chichanowicz, Electric Power Research Institute; D. J. Coppock, Public
Service Company of Colorado; B. F. Kee, Tennessee Valley Authority; G. 0. Lyman,
Gulf Power Company; J. A. Panacek, Public Service Electric and Gas Company (NJ);
M. A. Trykoski, Edison Electric Institute; M. L. Zwillenberg, Public Service
Electric and Gas Company (NJ); and Dr. W. Bartok and A. R. Crawford (retired),
Exxon Research and Engineering Company. The author also wishes to acknowledge
and express his appreciation for the helpful comments, assistance and other
contributions received from J. A. Barsin, the Babcock and Wilcox Company;
W. F. Edmunds, the Louisville Gas and Electric Company; T. T. Frankenberg,
American Electric Power Service Corporation; Dr. G. A. Hollinden, the Tennessee
Valley Authority; Dr. C. R. Guena, the Public Service Electric and Gas Company
(NJ); R. W. Robinson, Combustion Engineering, Inc. and A. P. Selker, Combustion
Engineer ing, Inc.
vii
-------
c5350
SECTION 1.0
INTRODUCTION
Combustion modifications for NOX emission control were investigated
and applied to utility boilers in California in the early 1960's. The techniques
developed in these investigations concerned oil and gas fired boilers only,
presumably because of the local need to control NOX emissions in the Los
Angeles area. Little if any further work was done until 1970 when field studies
conducted under the sponsorship of the U.S. Environmental Protection Agency (EPA)
commenced. Major emphasis of these studies concentrated on controlling emissions
in pulverized coal-fired boilers.
More than 35 large, pulverized coal-fired utility boilers have been
tested during the past 8 years in the EPA-sponsored field tests and the informa-
tion developed can be combined with combustion modification test results
obtained by other sources, such as, boiler operators and manufacturers, The
Electric Power Research Institute (EPRI) and their contractors. One of the
goals of the EPA-sponsored field tests was to establish whether combustion
modifications can be used effectively for the control of NOX emissions from
coal fired utility boilers without incurring undesirable side effects.
Short term emission tests established a range of 20% to 60% reduction
in NOX emissions from pulverized coal fired utility boilers, with the average
reduction being on the order of 35-40%. Other emissions did not increase as a
result of the combustion modifications applied, nor were boiler efficiency,
operability and safety adversely affected. The only remaining question is the
long term effect on fireside tubewall slagging and corrosion when firing high
sulfur and iron content bituminous coals. (EPA-sponsored field studies are in
progress to provide an answer to this question.)
The purpose of this guideline is to compile the experience gained in
the EPA-sponsored NOX reduction tests and make it available to utility personnel
who are required to reduce NOX emissions and maximize efficiency in large
utility boilers, especially those fired by pulverized coal. The objective is
to present the combustion modification techniques and procedures, found to be
successful in these investigations, in a readily understood form easily translat-
able into individual programs. The approach taken is that knowledge or under-
standing of NOX combustion modification techniques and experience in planning
test programs is lacking since utility personnel would not ordinarily be
expected to plan test programs, stage firing patterns, bias mill output or
operate routinely at low excess air levels. The assumption is made, however,
that the responsible individual conducting the program possesses a practical
understanding of combustion phenomena as related to utility boiler operation
and a working knowledge of power plant combustion control systems. Although
combustion techniques are different for low NOX operation, the same general
principles of boiler operation apply without violating normal good practice.
- 1 -
-------
c5350
Operating variables and parameters which affect NOX emissions and their interre-
lationship with steam temperature and automatic combustion controls are discussed
to provide an insight into how to plan a test program, lower NOX emissions, and
translate these achievements into standard operating practices.
Test program planning is illustrated in the guidelines by examples
encompassing investigation of major operating variables to determine the range
of NOX emission reduction possible for each combustion modification operating
mode considered. Procedures for following the test plan are detailed in
step-by-step progression leading to the determination of achievable NOj emissions
and operating configuration. Once this goal has been achieved, methods are
suggested for translating the results into normal operating practice and for
determining optimum boiler efficiency at reduced NOX emitting conditions.
It is important that this entire guide be read and understood before
attempting any boiler adjustments or modifications. In many cases it may be
desirable to acquire the assistance of outside combustion control specialists,
or to consult with the boiler manufacturer before proceeding with the
emission reduction tests.
- 2 -
-------
c5350
SECTION 2.0
BOILER INSPECTION AND MAINTENANCE
Financial requirements dictate that utility power plants operate as
much as possible (normally 50 weeks per year) with a short annual outage
(usually 2 weeks) for major repair requiring a shutdown and entry into the unit
for cleaning, proper inspection, and maintenance to restore the boiler and
turbine to original operating condition. Other opportunities do occur for
minor repairs during forced outages but these are of short duration with little
time for attention to other needed repairs. Consequently, boiler operating
conditions are probably best immediately following the annual maintenance
outage and deteriorate with continued operation to some level of less than
optimum performance. Obviously it would be unrealistic to expect a boiler to
be in prime condition throughout its yearly operating cycle. In pursuing
a NOX emission reduction program it should be recognized that ideal operating
conditions may prevail for only a short period and that optimum NOX emission
levels may not always be achieved.
Before attempts are made to reduce NOX emissions or to optimize
efficiency on a boiler, it is wise to ascertain the operating condition of all
components and, ideally, make corrections where necessary to assure maximum
working capabilities* Where outages are not possible in order to effect the
desired adjustments, recognition of these operating limitations should be noted
for future correction and possible achievement of lower NOX emissions. In the
interim, critical boiler components should be inspected including automatic
combustion controls, emission measurement instrumentation, fans, dampers,
furnace and convection section surfaces, etc., with special attention given to
the fuel burning equipment, i.e., pulverizers and burner registers, vanes and
impellers. Experience has shown that one or two bad burners can severely limit
the effective operation of a boiler, so a check of these critical components
is important in fine tuning. A suggested inspection checklist follows. For a
more complete list, boiler manufacturers recommendations should be consulted
for specific type equipment in your plant. Ideally, these inspections should
be made before initiating and periodically during the NOX emission reduction
program.
2.1 Furnace
• Inspect the burner throats for eyebrows or whisker accumulations of ash
which might interfere with the complete and uniform combustion of the
pulverized coal. Establish that burner throat surfaces are smooth
providing no rough projections on which the slag may cling and build
up.
• Ascertain that superheater and reheater tubes are clean with no fouling
of flue gas passages to block and cause maldistribution of the gases
resulting in higher stack temperatures.
- 3 -
-------
c5350
• Make sure that furnace surfaces have no excessive deposits which might
cause higher flame and gas temperatures resulting In high NO^ emissions
and potential decreases In boiler efficiency.
• Inspect all soot and slag blowers making sure of their capability to
remove slag and ash deposits so that higher gas temperatures can be
avoided and efficiencies and NOZ emissions may be maintained at optimum
values*
• Observation and furnace Inspection doors should all be in operable
condition. Clear visibility of burner throats, combustion conditions
and furnace surfaces Is Important in assessing the effect of combustion
modifications.
• On balanced draft boilers, it should be established that furnace casings
are tight with minimal air infiltration especially in the burner area.
Excessive hopper leakage on such units is also undesirable since
it could Interfere with and limit NO^ reduction techniques.
2.2 Air Heaters
• Air heater surfaces should be inspected for excessive fouling which
could bottle up the furnace and interfere with the capability of the
forced draft and induced draft fans in maintaining air and gas flow
through the unit.
• Air heater blowers and water washing devices should also be
in operable condition to keep the air heaters in the best condition
possible over the operating cycle of the unit.
• Air heater seals should be inspected and repaired or readjusted if there
is excessive leakage in order to mnlntnln full air flow delivery to the
furnace at i««-«-t«i™« load conditions.
2.3 Pulverizers
• Pulverizers should be inspected and adjusted for rated coal delivery to
the furnace at the proper fineness (usually 70Z through a 200 mesh
screen). Excessively fine coal could prohibit rated fuel delivery while
coarse coal would tend to have the opposite effect but result in more
uneven distribution of fuel particles which could have an adverse effect
on combustion. Also, coarse coal,. as discussed in Section 3.7, could
lower NO emissions but this effect would be minor.
Experience shows that coal distribution between individual burner lines
is notoriously uneven. Where maldistribution is excessive (which would
be exhibited by rich combustion in some flames), steps should be taken
to even out the distribution of coal between pipes. Remember that one
bad burner can upset the entire combustion process.
- 4 -
-------
d5350
• Check pulverizer feeder controls to minimize cycling of the coal feed.
Otherwise, stack opacity limitations may be exceeded while operating
under low NOX conditions.
2.4 Combustion Controls
• Inspect and adjust all automatic combustion control instrumentation to
respond smoothly to steam flow, steam pressure, and feedwater flow
demands. Adjust master controllers and corresponding regulators
freeing their operation so that hang-ups are minimized and jerky opera-
tion or "jumps" in combustion regulation do not occur. Reason: smooth
regulation and safe control of combustion process is essential for
optimum efficiency and NOX emission control.
• Windbox dampers and forced draft and induced draft fan control dampers
should be inspected for free operation. Play in damper linkage controls
should be eliminated so that response is positive upon demand.
• Control pressure regulators should be inspected for proper operation to
assure that output control pressures are sufficient to produce instant
response at all firing rates.
• Normally automatic combustion controls are set up to anticipate load
changes which are transmitted to the air controllers on load increases
and to the fuel controllers on a decrease in load. Thus, more than
sufficient air is supplied for combustion during load swings which can
have an adverse effect on efficiency and NOX emissions. On many control
installations a wide safety margin is given to air "lead" over fuel
and the automatic combustion control system on these installations
should be readjusted to provide only the amount of "lead" necessary to
follow load swings.
• Controls should also be adjusted to minimize effects of cycling or
"hunting" above or below the control setpoint which makes it difficult
to control NOX emissions and can be wasteful of fuel.
2.5 Pulverized Coal Burners
Burner condition, without a doubt, is the single most important
element in achieving NOX emission reductions and optimizing combustion.
Ideally coal and air flow should be uniform to each burner but in reality, this
is seldom possible, especially in a pulverized coal installation. Coal
distribution between burner pipes on a pulverizer is never uniform and the
distribution can change with changes in load. Air distribution between burners,
especially on utility boilers having a common windbox, is similarly non-uniform.
This problem is compounded by inoperable or "sticking" burner registers or by
burner register indicators which do not reflect the true percent opening of the
registers or vanes. Thus, resistance to air flow through the burners can vary
appreciably among burners and should be minimized for best results. There-
fore, before a NOX emission reduction program is begun, the following is
recommended:
- 5 -
-------
b5350
• Burner registers be inspected to ensure free operation, especially when the
boiler is in operation and the register operators are hot and have the
greatest tendency to bind.
• Make sure that burner register indicators reflect the true opening of the
register. Goal: to be able to set all registers to the same percent open
so that air flow may be as uniform as possible through the burners.
• Repeat the above for burner vanes even though the purpose of burner vanes
normally is to direct the flow or rotation of air through the boiler (not
regulation).
• Inspect all burner impellers (where they exist) for condition and coke
deposits. Install new impellers where needed. Reason: non-symmetrical
impellers or coke deposits can cause lopsided flames.
• Inspect the alloy tips or burner parts on the ends of pulverized coal
burner pipes to make sure they have not burned off and are in good shape.
Replace bad tips so that flame shape and symmetry will not interfere with
uniform coal and air distribution at the burner.
2.6 Control Panel Instruments
Utility boilers are operated remotely from a central control room
through means of the automatic combustion control system which selectively, at
the operator's desire, provides for either automatic or hand control operation.
Critical control instruments are strategically located in the control panel
within sight and easy reach of the operator. Other instruments essential to
defining complete operating conditions such as, temperature, pressure, and flow
indicators and recorders, ammeters, 02 meters, draft gauges, damper position
indicators, etc., are also grouped in the control panel. Modern utility
control rooms also have dataloggers or computer systems for scanning and hourly
printout of essential operating data which can be programmed to log desired
test information.
Preventative maintenance of the control panel instruments, automatic
combustion control systems, and plant instrumentation is generally the responsi-
bility of the plant instrument department. Instruments are normally maintained
and calibrated on a periodic basis on a frequency dictated by the type of
instrument and the manufacturers recommendations. Critical instruments and
controls are also given special attention and recalibration whenever required,
generally in compliance with operator requests.
Control room, panel board instrument data, as discussed in Section
5.0 and indicated in Table 3, will provide essential information characterizing
the mode of operation for individual tests. Before undertaking a NOX emissions
reduction program, a decision should be made concerning which instrument data
are important to the program. The instruments selected should then be cleaned,
adjusted and recalibrated so that reliable data may be recorded during test
runs.
- 6 -
-------
c5350
2.7 Emission Measuring Instruments
Control room instrumentation in most power plants, except for 02
meters, does not include instrumentation for the measurement of CO or NOX. In
carrying out a NOX emission reduction campaign, reliable 62, CO, and NO (or
NOX) instruments are a necessity. Existing plant 62 meters could be used
provided it has been ascertained that their readings accurately reflect the
average values of the bulk of the gas flowing in the duct. Otherwise, it is
recommended that separate instrumentation be obtained and a special sampling
system be installed to measure, as a minimum, 02, CO, and NOX which are
essential to the conductance of a successful NOX emission reduction program.
Ideally a C02 analyzer should also be included since it could not only serve
as an alternate to the 02 meter in calculating the diluent effect but the
values obtained could be used directly in calculations of boiler efficiency
for key test runs. Refer to Section 6.0 for further discussion of the importance
of emission measurements.
- 7 -
-------
b5350
SECTION 3.0
OPERATING VARIABLES AFFECTING NO,
Once the boiler, furnace, firing equipment, dampers, controls,
instrumentation, etc., have been inspected, and the necessary adjustments,
calibration and repairs have been made wherever possible in accordance with the
suggested check list, attention can be turned to an operating plan for combustion
modifications for the reduction of NOX emissions and the optimizatic/n of boiler
efficiency. In formulating the plan, a knowledge of major operating variables
and their potential effect on emissions is very helpful. Major operating
variables which affect emissions on a pulverized coal-fired utility boiler are
listed in Table 1 and are discussed briefly below.
TABLE 1
MAJOR OPERATING VARIABLES AFFECTING N0y
Major Variables Settings Affecting N0y
Load High, Medium, Low
Excess Air Normal - Low
Staged Firing Various Patterns Including OFA
Burner Tilt (Tangential Units) Up - Horizontal - Down
Burner Registers Open - Partially Closed
Flue Gas Recirculation 0 to 30% of Gas Flow
Pulverized Coal Fineness Coarse - Fine
Investigations have shown that NOX emissions are derived from two
sources, i.e., (1) from the nitrogen contained in the combustion air and (2)
from the nitrogen chemically combined in the fuel. High temperature fixation
of molecular nitrogen in the combustion air occurs in the flame to yield
"thermal NOX." Simultaneoulsy, conversion of the chemically bound nitrogen in
the fuel during the combustion process results in "fuel NOX." Combustion
modification variables which affect flame temperature are most effective in
controlling thermal NOX such as, load, burner tilt, burner register settings,
flue gas recirculation and pulverized coal fineness. Other combustion variables
like excess air and staged firing which limit oxygen availability and affect
temperatures in the furnace, reduce both thermal and fuel NOX emissions. The
effect on the latter, however, is limited. Fuel NOX emissions are strongly
dependent on oxygen availability and relatively insensitive to temperature.
3.1 Load
Boiler output, or load, is one of the major operating variables
affecting NOX emissions. On a given unit very little control can be exercised
over this variable which is dictated by system demand requirements. NOX
- 8 -
-------
b5350
emissions in general, but not always, increase with load and are highest at
maximum load. In extreme cases boiler capacity could be derated to control
emissions. Obviously, this is undesirable economically and should be avoided,
when possible.
3.2 Excess Air
An excess of combustion air is always supplied to a boiler to ensure
complete burnout of the fuel and to provide a "cushion" or safety margin for
the operation of the boiler. Excess air is one of the most important operating
variables in the control of NOX, especially when used in combination with
staged firing which is discussed later. NOX emissions are lowest and boiler
efficiency highest at lowest excess air levels. Typically, excess air levels
on coal-fired utility boilers vary from 18 to 25 percent (3.5 to 5% Q£ in the
flue gas). Only recently has it been found possible to reduce oxygen levels on
pulverized coal-fired boilers down to 1 to 2 percent to minimize NOX emissions
(and possibly increase efficiency) in short term tests. Normally, levels of 2
to 3% 02 can be achieved on a long term basis (without exceeding the arbitrary
200 ppm CO level discussed below).
Practical, minimum excess air levels will vary widely from boiler to
boiler depending upon furnace design factors, fuel characteristics, boiler and
burner condition and operating control flexibility. Minimum excess air level
is defined as the level in the boiler flue gases where CO is a maximum of 200
ppm with normal stack plume opacity. Experience has shown that unburned carbon
loss at this level is low and that boiler efficiency, in most cases, is not
affected. A boiler designed and maintained so that equal, or nearly equal, air
to fuel ratios can be established on all operating burners will be able to
achieve minimum excess air levels. Thus, burner design, combustion air controls
and operating condition of air and fuel controls have an important effect on
the level of excess air for proper combustion. Minimum excess air operation is
obtained under steady, full load conditions, with all burners operating and
with operating time available for "fine tuning" adjustments. Operating at
reduced or varying loads will require higher excess air levels. Fuel character-
istics such as low volatile, high ash content coal and the degree of coal
fineness may also influence air requirements.
Low excess air operation reduces both thermal and fuel NOX generation
and increases thermal efficiency. Establishing practical minimum excess air
levels requires measurement of flue gas composition to be assured that combustion
is complete (CO should be less than 200 ppm and Q£ probably should not be
lower than 1%). Flame stability, shape, and possible impingement should be
checked and any affect on steam temperature control (discussed later) should
be noted. For long term operation, tendencies toward increased slagging
requiring more frequent slag and soot blowing operation should be observed.
3.3 Staged Firing
In the past the pattern of burner usage was dictated largely by load
and steam temperature control demands with all burners firing at full load and
top row burners kept in service during low load operation to maintain superheat
_ 9 -
-------
b5350
temperature. In the last 10 years staging the firing pattern in combination
with low excess air operation has been found to be the most effective technique
for limiting 1TOX emissions. Various combinations of burners are employed
(discussed later) resulting, in general, in an initial combustion zone in the
lower part of the furnace where combustion takes place under sub-stoichiometric
conditions with burn-out being completed in the upper stage with the admis-
sion of overfire air (OFA). In some new boilers designed to meet New Source
Performance Standards (NSPS) for NOX, overfire air ports have been built into
the boiler. In existing boilers staging has been simulated using top rows of
burners as overfire air ports. The latter technique, while effective, may be
limited if full load capability is not possible with top row burners out of
service. Staging, however, renains strictly a NOX emission control technique
and would not be practiced where emission controls are not required.
3.4 Burner Tilt
Tilting burners are employed on laost Combustion Engineering, Inc.
tangetially fired boilers for the express purpose of controlling superheat
and reheat steam temperatures. Burners can be tilted within a range of -30°
(dovmwards) or +30° (upwards) from the horizontal position. At low loads,
burners are tilted upwards to provide more heat in the superheater and reheater
areas. As load is increased the tilt is decreased to keep steam temperatures
at the desired control point. Burner tilt has also been found to be an important
operating variable in the control of NOX emissions. Burner angles at or near
the horizontal position usually produce less NOX. This should be checked on
each individual unit. The angle of burner tilt for NOX control, however, may
be in conflict with steam temperature control demands. This variable must be
investigated to determine optimum tilt angles to satisfy both temperature
demands and NOX emission limits. Usually there is sufficient flexibility in
attemperation for steam temperature control to arrive at a compromise between
both, although the range of superheat and reheat control nay be narrowed in
some cases*
3.5 Burner Registers
Fine tuning of the combustion in individual burners is accomplished
in part, by adjustments in burner register position. These adjustments are
intended to improve the combustion process, usually by changing direction (but
not the flow) of the incoming secondary air. Register adjustment is not for
the purpose of combustion air regulation rather adjustments normally increase
or decrease burner turbulence. Although burner registers are capable of
closing off, they normally are only shut when a burner is out of service.
Changes in burner register adjustment have also been found useful as
a variable in reducing KOX emissions, especially in staged combustion operation.
Partial closing of the registers on first stage burners to approximately 40%
open while opening the registers on second stage burners (the ones used as
overfire air ports) to 80 or 100% open has the effect of reducing the air to
the lower burners and increasing the overfire air flow. In most installations
where this technique has been used, substantial reductions in NOX emissions
have been achieved. The variable of burner register adjustment as an assist to
staged firing for NOX control, therefore, should not be overlooked.
- 10 -
-------
c5350
3.6 Flue Gas Reclrculation
Recirculation of up to as much as 30 percent of the flue gas flow is
practiced on some boilers for superheat and reheat temperature control. The
recirculated gas usually is injected through ports in the furnace walls in the
hopper area, having the effect of "blanketing" the furnace walls and increasing
the mass flow of gas through the superheater and reheater surfaces holding the
steam temperature at the desired control point over a wider load range.
Recirculation of flue gas has also been found to be a relatively effective
variable for NOX emission control on gas and oil fired boilers, especially
where the flue gas is mixed with the combustion air* In the latter case the
oxygen available for combustion is diluted, resulting in lower flame temperatures
and reduced NOX emissions. The effectiveness of flue gas recirculation for NOX
control is not as great when the gas is admitted directly to the furnace. Flue
gas recirculation is not as effective on coal-fired boilers because it is
relatively ineffective on NOX formation from fuel bound nitrogen. A conflict
also exists with the use of recirculated gas for steam temperature control versus
NOX control, so investigation of this variable is important to establish optimum
conditions for both, over the load control range. For these reasons, FGR is not
a cost effective means for NOX control on coal-fired boilers.
3.7 Pulverized Coal Fineness
The fineness of coal pulverization has been noted as a variable which
can affect NOX emissions, at least to a minor degree. Theoretically the
coarser the coal, the longer it takes the coarser particles to burn out. The
slower diffusion rate of the air into the coarse fuel particle results in
longer flames, lower temperature and reduced NOX emissions. While the
fineness of the coal is usually not a variable which is changed but, rather, is
a result of mill wear gradually over a period of time, it is nevertheless a
factor which can affect NOX emissions. On most mills coal fineness can be
adjusted so this variable, too, bears investigation in the achievement of lower
NOX emissions.
- 11 -
-------
c5350
SECTION 4.0
STEAM TEMPERATURE CONTROL
Changes in boiler operating variables, as discussed previously, have
a corresponding effect and cause changes in superheat and reheat steam tempera-
tures especially in utility boilers which are of more complex design than
industrial type boilers. In order to control superheat, and reheat steam
temperatures at constant levels over as wide a load range as possible,, special
design features are incorporated into utility boiler designs. Some of these
have already been discussed above as operating variables but for clarification
purposes they will be discussed briefly again relative to steam temperature
control. Table 2 lists the arrangements incorporated in utility boiler designs
for the specific purpose of controlling superheat and reheat steam temperatures.
Note, however, that not all design features will be found on any one boiler.
TABLE 2
SUPERHEAT AND REHEAT STEAM TEMPERATURE CONTROL FEATURES
Attemperation
Burner Tilt (Tangential Boilers)
Biasing Dampers
Flue Gas Recirculation
Excess Air
Furnace Slag Blowers
4.1 Attemperation
Most modern boiler designs incorporate an attemperator (or desuper-
heater) between the primary and secondary superheater sections for the purpose
of decreasing the temperature developed in the primary section and controlling
the final superheat steam temperature over a wider range of boiler load. The
superheater is designed for full temperature at some partial load point.
The excessive superheater surface results in correspondingly high temperatures
at higher loads and the excess temperature is "trimmed" off by the attemperator
by spraying water into the steam in sufficient quantity to maintain final steam
temperature emitting from the secondary superheater at the desired control
set point. This arrangement is a relatively delicate type of control and is
sensitive to changes in gas temperature and mass flow of the gas across the
superheater and reheater. Variation in excess air and flue gas recirculation
for NOX control affects both the mass flow and temperature of the gases
passing through the unit which will have a corresponding affect on attemper-
ation. A trade off in NOX emission control with steam temperature control
therefore, may be necessary in some boilers. The effect could also merely'be a
narrowing of the superheat and reheat temperature control range.
- 12 -
-------
d5350
4.2 Burner Tilt
As indicated above, most Combustion Engineering, Inc. tangentially-
fired utility boilers incorporate devices so that the burners may be tilted
through an arc range of -30° to +30° from horizontal* Tilting the burners
upward so they fire more directly into the superheater puts more of the heat
near the superheater entrance, effectively bypassing a portion of the furnace,
making it possible to maintain steam temperatures at lower loads. As the load
increases and less heat is needed in the superheater or reheater, the burner
angle is depressed to maintain temperature. As previously indicated, the
burner tilt angle can have an appreciable effect on NOX emissions with
lowest emissions normally being obtained with burners firing close to the
horizontal position. Optimum points may be worked out so that the desired
steam temperature can be maintained with minimum NOX emissions.
4.3 Biasing Dampers
Some boilers are designed with the superheater and rsheaters side-by-
side, rather than in tandem in the direction of gas flow which is more common
in modern designed boilers. On boilers with superheaters and reheaters side-by-
side one method of steam temperature control employed consists of an arrangement
of biasing dampers installed downstream of the convection section which can be
varied to divert more of the flue gas flow from the superheater side to the
reheat side and vice versa. This type of steam temperature control should be
independent of NOX control techniques but, again, a change in gas temperature
or mass flow as a result of NOX combustion modifications will affect steam
temperature control, thus requiring readjustment.
4.4 Flue Gas Recirculation
In past years flue gas recirculation (FGR) was a popular form of steam
temperature control. In recent years this type of control seems to have lost
favor especially on pulverized coal fired boilers due to fan problems and
difficulties in handling dust laden gases. On most units where FGR is employed,
only a small amount (or zero) of gas is recirculated as load drops off to keep
superheat and reheat temperatures at the control set point. Gas recirculation
is most effective for NOX control when mixed with the combustion air before
the burners. However, when FGR is used solely for steam temperature control,
it is usually injected into the hopper zone. In any event, the amount of gas
recirculated can affect NOX emissions and optimum conditions must be established
to satisfy both steam temperature control and NOX emissions. As mentioned
before, flue gas recirculation is of limited effectiveness for NOX control
with coal-fired boilers because it does not retard conversion of fuel nitrogen.
4.5 Excess Air
Varying the amount of excess air at a given load affects both the
flue gas temperature and the gas mass flow which has a corresponding effect on
superheat and reheat steam temperatures. Some boilers have been designed
employing this principal for steam temperature control. Fortunately, not too
many units of this type are in existence since one of the most effective NOX
control techniques employs low excess air in combination with staged combustion
to achieve low NOX emission values.
- 13 -
-------
b5350
4.6 Furnace Slag Blowers
On pulverized coal fired boilers, some of the ash in the coal deposits
on the furnace walls. Eventually, with time, these deposits may form sizeable
accumulations of slag which can effectively blanket the furnace walls, decrease
the heat absorption in the furnace and result in increased gas temperatures
entering the superheater and reheater sections. If the process is allowed to
continue, a point would be reached where superheat and reheat temperatures
could not be kept in control by the means discussed above and the unit would
have to be taken out of service to clean the furnace. However, slag blowers
are strategically placed in the furnace and are operated on a regular basis
(at least once per shift) to remove slag accumulations and keep the furnace in
a "clean" condition. Thus, the slag blower operating cycle is an important
factor in keeping steam temperatures under control. Under low NOX emission
conditions there may be a tendency for slag accumulations to build up faster
requiring more frequent furnace slag blower operation.
It is also important, from a steam temperature control standpoint, to
maintain superheater and reheater surfaces free of ash deposits which might
interfere with heat transfer or the flow of gas through these passages. Soot
blowers are provided for this purpose in the superheater and reheater sec-
tions and their operating schedule is similarly influenced by the same factors
discussed above.
- 14 -
-------
c5350
SECTION 5.0
OPERATING TEST DATA
In conducting a program to reduce NOX emissions from a boiler, it is
essential that pertinent control room board data, emission measurements and
operating parameters be thoroughly documented to completely characterize each
test. This would include such information as burner register settings, burner
tilt angle, opacity, observations on flame and slagging, notes on staging
patterns, etc. The recorded test data will become a permanent record of NOX
emission capabilities which may be used to plan subsequent tests. More important,
the data obtained will permit exact repetition of operating conditions after
determining which mode of firing resulted in optimum NOX reduction capability.
A sample data sheet, illustrative of the type of test data which
should be obtained, is shown in Table 3. Instrumentation will vary, of course,
with individual boilers and the data sheet should be expanded to include other
pertinent operating data wherever possible.
It should be noted, with reference to Table 3, that large utility
boilers usually have split flue gas streams downstream of the economizer.
Thus, there may be two airheaters, two ID and FD fans, etc. Also each boiler
will have from 4 to 10 pulverizers depending on size of the unit. Test data,
therefore, should be recorded for each piece of equipment in the installation.
- 15 -
-------
TABLE 3. BOILER TEST DATA
Utility
Station
Boiler No.
Fuel
Date
Test No.
Time
Load, MW
Sim. Klow, Lb./Hr.
Air Flow, Lb./Hr.
Food Water Flow. Lb./Ur.
Coal Scale. Start
Coal Scale, Finish
Steam Press. , PSIG
Superheat Temp., F
Re he «t Temp., F
SH Attempt. H,0 Flow, Lb./Hr.
RH Attempt. H,0 Flow, Lb./Hr.
Burner Tilt. Def>.
Over fire Air Damper Position
Overfire Tilt, Deg.
Burner Register Position
FD Fan Amps
ID Fan Amps
Furn. Press., in H»0
Uindbox/Furn. Dlif.. in H.O
Airheater Dlff.. in H,0
Pulverizer Temp. . °F
Airheater Air in Temp.. °F
Airheater Air Out Temp..
Airheater Air In Temp.. °F
Airheater Gas Out Temp., °F
Opacity! Z
0. Percent
CO. PPM
IW . PPH
Staging Pattern
SlaftKlng Obierv.
Remarks
- 16 -
-------
c5350
SECTION 6.0
EMISSION MEASUREMENTS
In carrying out a NOX emission reduction program, it is essential
that accurate and reliable emission measuring equipment be used. This equipment
must be located such that a representative sample of the boiler gas stream is
measured. As a minimum, equipment for measurement of 02, CO and either NO or
NOX is required. C0£ may be measured as an alternative to 02 for the purpose
of computing the diluent effect.
Older, pre-NSPS sources are usually equipped with only an 02 measure-
ment system, with the sensor/probe located in the ductwork between the economizer
and airheater. If this system is to be used for the NOX reduction program, it
must be evaluated in terms of accuracy and reliability. If it is used, the NOX
and CO measurement equipment must be located at the same place. U.S. EPA regula-
tions concerning the installation, operation and maintenance of continuous
monitoring systems provided in "Performance Specification 2 (PS2), 40 CFR part
60-as amended" are recommended for guidance. If the source already has a certi-
fied continuous monitoring system, it should be used for the NOX reduction
program.
If temporary emission measurement equipment must be installed for the
proposed program, consideration should be given to equipment which can operate
unattended for extended periods of time. Provisions should also be made for
some form of automatic data recording, such as stripcharts. If the NOX reduction
program will be used for a compliance demonstration, this type of installation
is normally not acceptable unless approved in advance by the appropriate
agency. Where compliance tests are required, Federal Reference Methods (as
mentioned above and in Appendix D), must be used for confirmation.
The installation and operation of any emission measurement equipment
should be under the supervision of appropriate, qualified personnel. Appendix
D covers the essential details and methods for measuring combustion product
pollutants.
6.1 Comparing NOV Emissions
Nitrogen oxide (NOX) emissions for any given test are measured on a
volumetric basis at the specific excess air conditions under which the test is
conducted. Since excess air is varied to reduce NOX emissions, the data
developed are not readily comparable, and the results are obscured by the dilution
effect of the excess air in the flue gases when reported in parts per million,
ppm. For example, when excess air is decreased, NOX emissions also decrease
but the apparent measured value may seem higher than expected due to the
simultaneous effect of concentration of the flue gases (less 02). Conversely,
- 17 -
-------
c5350
experience shows that NOX levels increase as excess air increases, yet measured
values may not appear too high since N0y emissions are being diluted as
excess air increases. It is important to understand the diluting effect on NO
(and other) emissions emitted from the boiler and the need for a common compar-
ison basis when results are reported as parts per million.
Early investigators resolved this matter by converting and reporting
all emission results on a common basis of 3% 62 in the flue gases. This has
carried over and been accepted by the industry and has been adopted by the EPA
as standard practice for evaluating and comparing emissions. After average
oxygen and nitrogen oxide emissions are established for a particular test, the
NOX emission value is corrected to the 3 percent oxygen basis by multiplying by
a correction factor determined as follows:
correction factor = 21-3% 02 (standard basis)
21-%02 (average for test)
18
21-%02 (average for test)
For example, emissions for a given operating mode average 485 ppm NOX
with an average of 1.6 percent 02 in the flue gases:
Therefore:
NOX (at 3% 02) = NOX (at test conditions) x correction
19.4
- 485 x .9278
» 450 ppm
NOX emissions measured on a volumetric basis in ppm, may be converted
to a weight and unit of heat input basis through the use of F factors. The use
of the F factor in calculating particulate emission levels from new stationary
sources was promulgated in the October 6, 1975 Federal Register. Other
publications of the F factor methods of conversion are (1) EPA report EPA-340/
1-77-015 "Standards of Performance for New Stationary Sources," November 1977
and (2) EPA's Air Pollution Training Institute, Manual 450, "Source Sampling
for Particulate Pollutants," January 1978. The purpose of the F factor is to
reduce the amount of data necessary to calculate emissions in terms of the
standard expressed as lbs/10^ Btu, heat input. Procedures for the
conversion of emissions from a dry (usual) to a wet basis are also detailed
in the F factor publications mentioned above.
Emissions given as mass per unit of heat input or mass of fuel burned
do not need to be corrected for dilution. However, most continuous monitors
provide results in units of ppm.
- 18 -
-------
d5350
SECTION 7.0
CARRYING OUT THE NO,, REDUCTION PROGRAM
A. i
At this point, it is assumed that the boiler on which it is desired
to reduce NOX emissions has been inspected in accordance with the suggested
check list, and items needing repair, adjustment or calibration have been
attended to insofar as may be feasible. It is also assumed that control,
recording and emission measuring instrumentation has been inspected and test
data sheets have been prepared to record the important test data.
The foregoing reviewed the major operating variables, and the changes
which NOX combustion modifications would be- expected to have on emissions and
superheat and reheat temperatures. Before proceeding further, consideration
should be given to whether the objective of the NOX emission reduction plan is
for compliance or merely to establish the emission capabilities of the boiler.
Establishment of these goals will help to determine how deep a cut may be
necessary. The suggested plans and procedures which follow, probe the maximum
NOX emission reduction capabilities of the boiler which may be relaxed if wide
latitudes in complying with regulations are found to exist.
It is not within the scope of this guideline to provide the necessary,
detailed combustion control knowledge required to carry out a NOX emission
program; rather, it is assumed that competent boiler operators well versed in
boiler operation are in charge. Combustion modifications and operation for NOX
emission control are different than for normal, routine operation. For instance,
excess air levels are lower than for normal operation and with normal firing no
attempts are made to stage the combustion process or to make adjustments in
burner registers other than those recommended by boiler manufacturer's operating
instructions. Therefore, a test program plan is necessary, even for experienced
operators, where a step-by-step procedure is followed for each test, leading in
a progressive fashion to the establishment of optimum efficiency and "low NOX"
emission operating conditions. A word of caution is necessary here, how-
ever. Changes should be made in small, incremental steps so as not to lose
control of the operation at any time.
7.1 Test Program Plan'
One of the first tasks in the approach to a successful NOX emission
reduction program is to develop a test program. Operating variables (which are
specific for each boiler) should be carefully reviewed for inclusion in the
experimental program. This is best illustrated by the example in Table 4
showing a simplified experimental test program plan for a 125 MW front wall
fired boiler, firing pulverized coal in 16 burners arranged in a matrix of four
burners wide by four high as shown in Table 4. Operating variables included in
the plan are:
- 19 -
-------
TABLE 4. EXPERIMENTAL TEST PLAN
i
NJ
o
Secondary
Pattern Air
Firing
S 16 Coal
1 0 Air Only
S 14 Coal
Dj D4 Air
S, 14 Coal
A., A4 Air
S 12 Coal
A! AZ A3 A^ Air
S 12 Coal
Al \ B2 B3 Alr
L. - Full Load (125 MW)
A. - Normal Air
201
Open
R2
(3) 2.8%
610
(5) 3.8%
632
(11) 4.57.
532
607.
Open
Rl
(1) 3.2%
* 577
(7) 4.0%
558
(9) 4.1%
518
A2 - Low Air
20%
Open
R2
(4) 1.9%
505
(6) 2.0%
372
(12) 1.7%
345
60%
Open
Rl
(2) 2.0%
491
(8) 1.5%
406
(10) 2.7%
368
L - Reduced Load (110 MW)
A^ - Normal
207.
Open
R2
(15) 4.9%
681
(17) 4.5%
399
(23) 4.9%
496
607.
Open
Rl
(13) 4.87.
629
(19) 4.5%
460
(21) 4.47.
480
A- - Low Air
207.
Open
R2
(16) 2.87,
464
(18) 2.77,
297
(24) 3.47.
306
607.
Open
RI
(14) 2.77.
450
(20) 3.07.
345
(22) 2.67.
342
* Numbers In parentheses In boxes are test run numbers (in sequence of
testing) . Figures In boxes are average percent 02 during test and ppm
NOX emissions corrected to 3% 02 dry.
NOTE: Experimental test plan Is also used as a "Score Sheet."
Teat results shown In boxes only become available as
each test Is completed.
Pulverizer-Burner
Configuration
Mill
A -Top
B-2nd
C-3rd
Row
Row
ftow
D-Bottom Row
Burner No.
1.
0
0
0
0
2
0
0
0
0
3
0
0
0
0
4
0
0
0
0
-------
d5350
Load - High (L^), intermediate (L2)
Excess Air - High or normal (Aj) and low (A2)
Burner Registers - Normal setting, 60% open (R^)
- Closed down, 20% open (R2)
Staging Patterns - S^, normal firing, all burner in
service
- 82, staged, two burners, Dj and
D^ on air only
- 83, staged, two burners, A^ and
A^ on air only
- 84, staged, four burners, A^, A2,
A3 and A4 on air only
- 85, staged, four burners, Aj_, A^,
62 and 63 on air only
The test plan in the example shown in Table 4 calls for testing at
two loads only, i.e., maximum rated load and intermediate load. In many cases
testing at a low load or additional intermediate ratings may be desirable. In
such cases, the test program plan should be expanded accordingly. Also, as
explained previously, testing at each stage is normally conducted at two
excess air levels only; i.e., normal (relatively high) and minimum (low excess
air) as defined by the 200 ppm maximum CO level. Furthermore, it should be
pointed out that the minimum excess air level most likely will vary on a given
boiler depending on the particular combustion modification operating mode, coal,
condition of the boiler and combustion system, etc.
Testing at two different burner register settings only are included
in the simplified test plan example. With more complicated burner arrangements,
testing at additional burner register settings may be required to adequately
optimize the effect of burner register adjustments on NOX emissions, resulting
in the necessity to further expand the test plan.
Several different staging patterns are included in the test plan
example shown in Table 4. Since, initially, the effect of different burner
arrangements on NOX emissions is unknown, various combinations of burners must
be tried until the optimum configuration becomes apparent. It is known that
wider separation of burners (improved cooling between burners) and overfire air
has a beneficial effect on NOX emissions. Various burner staging arrangements
to accomplish these objectives should be included in the test plan in order to
develop the optimum combustion mode. Larger more complicated boilers with
greater numbers of burners may require more extensive staging pattern investi-
gation than those shown in the simplified plan in Table 4. However, sufficient
information on staging the combustion process should be available from Table 4
to develop a more extensive test program. Examples of more complete experimen- .
tal test programs are given in Appendix A. Actual test plans and results can
be found in References 1 & 2.
7.2 The ABC's of Reducing NOV Emissions
In any NOX emission reduction program three fundamental, basic steps
are taken at each incremental phase of the program. These steps are illustrated
in the flow plan in Figure 1 as the ABC's of NOX reduction. Essentially the
approach is the same at any given point with the three steps being repeated
- 21 -
-------
c5350
over and over again, but under different operating conditions, whether they
occur at the beginning, at the end, or at some intermediate stage of the program.
An understanding of the objectives of the ABC approach, therefore, is most
important.
Referring to Figure 1, the three basic steps which are repeated at
each stage of the program for each combustion configuration are:
(A) Establishment and characterization of the operating mode.
(B) Measurement of NOX emission at normal (or high) excess air
operation for the operating mode established in A.
(C) Reduction of excess air to minimum levels and measurement of NOX
emissions.
Considerable thought must obviously be given to the establishment of
the operating mode regarding potential effects on NOX emissions. Step A,
therefore, is of utmost importance in planning each phase of the NOX reduction
program. For instance, the effect on NOX emissions of closing burner registers
20 to 40% from normal may not be readily predictable. What may be assumed to
have a beneficial effect may in actuality have an adverse effect on NOX emissions,
A series of burner register positions (different combustion modes) must conse-
quently be incorporated into the overall NOX investigation program in order to
fully assess and optimize NOX emission reductions. Similarly other operating
variables must also be investigated separately.
Once the operating mode has been established and characterized, the
second step (step B) is to measure NOX emissions for the operating mode selected
at "normal" excess air operation. The immediate question which arises is "what
is normal excess air"? Most utility boilers are designed to operate with about
17 to 20 percent excess air, i.e., 3.5 to 4.0 percent 02 in the flue gases. In
the field these boilers have been observed to be operating in excess of design
conditions, generally at about 4 to 5 percent Q£ and sometimes higher. In any
event, measurement of NOX emissions in step B are intended to be made at the
excess air level which the boiler normally operates (usually relatively high
excess air conditions).
In step C the potential for reducing NOX emissions for the operating
mode established in step A is explored by reducing excess air to a minimum and
again measuring NOX emissions (Step C). This is done by reducing the combustion
air in small incremental steps, keeping careful watch of 02 and CO levels.
Experience shows that the baseline CO level for normal operation on utility
boilers ranges between 30 and 40 ppm. As the air rate is decreased, oxygen
levels will be observed to drop with very little change in CO readings. A
point will be reached where CO levels commence to increase. Further decreases
in the combustion air should then be made until CO values (from the lowest 0%
probe if multiple gas sampling ports are used) reach a level of about 200 ppm.
This point is defined as the minimum excess air level. Reducing excess air
beyond this point normally will increase stack opacity conditions to unacceptable
levels. On some boilers under some operating conditions, especially at low
- 22 -
-------
Establish
Operating
Mode
Determine
NO Emissions
x
at Normal
(Baseline) Excess Air
Determine
NO Emissions at
x
Minimum Excess Air
Figure 1. Fundamental Steps in Reducing NOX Emissions.
- 23 -
-------
d5350
loads, it may not be possible to reach the 200 ppm CO level. Stack emissions
and flane appearance observations should be nade in these cases for instability
indications as the cotnbustion air is decreased. These observations should then
take precedence over the 200 ppm CO limitation. A good criteria for the latter
is the differential pressure between the windbox and the furnace. Care should
be exercised not to reduce excess air very nuch below the point where the
windbox to furnace pressure differential drops below 1 inch of
This completes the three (ABC) steps for the first operating node. A
new operating mode would then be established and the three steps repeated as
described next.
7.3 Basic Steps in Reducing NOV Emissions - Step by Step
Procedures
Since combustion conditions are most stable and secure at or near
full load it is advisable to start the test program under these conditions.
Besides, NOX emissions generally are greatest at maximum boiler output and
one of the objectives is to determine the maximum baseline emission level on
the unit.
Operating at full load, normal conditions, the boiler/turbine
should be disconnected from the "Automatic Dispatch System" so that load nay be
held constant during the test. The automatic combustion control system may be
kept on automatic as long as load docs not vary appreciably and thus obscure test
results. Later, after firing conditions for optimum low NOX emissions have
been defined, the boiler can again be operated in the automatic position once
the necessary adjustments have been nade in the automatic combustion controls.
It is assumed that some logical sequence of test runs has been
established in the development of a test plan similar to that suggested in
Table 4. This sequence should be established in the planning stage considering
the ease of changing from one mode of operation to another. Note that the
suggested test program plan (Table 4) may be used as a scoresheet to record
test results, i.e., average percent oxygen during the test and average NOX
emissions converted to the standard base of 3 percent C>2 (explained in Section
6.1), As the test progresses, a glance at the test plan sheet will show which
mode of operation is most successful in reducing NOX emissions.
The flow plan shown in Figure 2 provides a basic step-by-step procedure
for conducting the experimental test program and is applicable to any type of
test plan. Although the flow plan in Figure 2 follows the test plan example in
Table 4, other operating variables such as flue gas recirculation, air preheat
temperature, burner tilt, biasing fuel between pulverisers, etc., may be added
to the plan and the sane procedures followed as described below in the examples
for burner register settings or staged combustion patterns.
• Baseline Operation - Full Load
At the start (step A) with the boiler operating normally at full
load, the first task is to completely characterize the conditions under which
the boiler is operating regarding load, burner register setting, number of
- 24 -
-------
Start
I
Normal Excess Air
Baseline Operation
B
- Measure NOX (1}
Yes
Low Excess Air
Normal Firing
Operation
— Measure NO,
Can
Excess Air
Be Reduced?
No
Normal Excess Air
Adjusted Burner
Register Operation
Measure NOX
Yes
No
Low Excess Air
Adjusted During
Registration Operation
— Measure NO,
1
Normal Excess Air
1st Staged Firing
Operation
— Measure NOX
Yes
No
Low Excess Air
1st Staged Firing
Operation
— Measure NOx
1
Normal Excess Air
2nd Staged Firing
Operation, etc.
— Measure NO,
Can
Excess Air
Be Reduced?
Yes
Low Excess Air
2nd Staged Firing
Operation, etc.
Measure NO
Finish
Figure 2. Basic Steps in Reducing NOX Emissions.
-------
d535G
burners in service, etc. For instance, following the example in Table 4, test
Ho. 1, the boiler would be operating at full load (125 ITU) with all 26 burners
firing at normal air and with burner register settings 60% open. At this point the
level of excess air probably will not be known but it cii^ht be assumed that it
nay be in excess of design conditions. An indication may, of course, be
obtained from the boiler Oo meters. These pertinent operating parameters
would be recorded to characterize the operating mode and step A would be
complete.
The second task (step B) is to measure NOX emissions (test No. 1) and
record operating data for the first test run at baseline (normal) conditions.
The data obtained will establish baseline NOX emissions and the excess air
level for normal operation of the boiler. The first and second objectives
(steps A and B) are now complete, i.e., the operating mode has been established
and characterized and emissions have been measured for normal operation of the
boiler.
The next step will be the first attempt at reducing NOX emissions for
the node of operation established in step A. At this point there may be a
question of whether excess air can be reduced. A quick glance at the D£ and CO
values obtained in the first teat run will give a good indication of the
possibilities. If Oo measurements are in excess of 3.5 percent and CO values
are no higher than 40 ppm, then there should be adequate leeway to reduce
combustion air levels. Generally, as mentioned previously, excess air levels
on most boilers are high (in the 20-25" region) and there usually is no question
that reductions can be made. Assuming that this is the case, combustion air
should now be decreased to establish firing conditions for step C (Test No. 2).
As described, in Section 7.1, decreases in combustion air should be made slowly
in small controllable steps until the minimum excess air level is reached
(defined as the point where maximum CO concentration reaches 200 ppm). After
operation has stabilized, a set of readings should then be taken to record
operating parameters for Test No. 2. This completes steps A, B, and C and
Test Nos. 1 and 2 for the mode of operation determined in A for those cases
where excess air can be reduced.
Returning briefly to step B, in certain isolated cases it may be
found that the boiler is already at or near minimum, excess air levels and
combustion air cannot be reduced further without increasing CO emissions beyond
the 200 ppm limit. NOX emissions, therefore, are already at a minimum and,
referring to the flow plan in Figure 2, step C has been accomplished before
step B. The alternative in those instances where a boiler may be operating
optimumly (low excess air and low HOX) is to arbitrarily increase the combustion
air to typical levels (A to 5% 03) and measure NOX emissions at these psuedo
baseline conditions (step B) to show the spread in NOX emissions as a function
of excess air level.
At this point the ABC's of baseline emission characterization is
complete for the initial phase of the test program and combustion modification
techniques are waiting to be explored. Before changing the combustion pattern
however, it is recommended (after step C) that excess air should be increased
to the normal range approximating baseline operation determined in step B.
- 26 -
-------
d5350
Sufficient margin of safety will thereby be provided to make the changes for
the next operating rede to be investigated and to ncconnodate unexpected or
abrupt changes in conhustion without losing control of the operation.
Folloving the flow plan in Figure ?., changes nay now be made in the
operating mode, such as closing burner registers down fron 60 to 20 percent
open, as suggested in the test plan in Table 4. Upon completion of the
adjustments a new operating node will have been established and characterized
(repeat of step A). NOX emissions and other operating data (Test Ho. 3) are
then recorded (repeat of Step B) for the new operating node. Afterwards,
excess air nay again be slowly reduced to minimum levels (200 ppn CO max) and
pertinent operating data recorded (Test ITo. 4) in a repeat of step C. Test
llos. 3 and 4 of the suggested test plan in Table 4 will be complete at this
stage and the boiler will be ready for further testing after again increasing
excess air to baseline conditions.
An appraisal of the results obtained in Test No. 4 may not give a
clear-cut indication of an improvement in NOX emissions as indicated in the
flow plan in Figure 2. A decision nust be made at this point whether data on
additional burner register settings is warranted to optimize WOX emission
reductions. The full effect of burner register settings may not be apparent
until staged firing runs are completed. Better judgment of these effects,
therefore, nay be delayed until staged firing runs are completed. However, if
further data are desirable, other burner register settings can be attempted
at this tine and the process described above should be repeated for each
of the new settings. If sufficient information does appear to be available,
the test program may proceed to the next phase, "investigation of the effects
of staging the combustion process on NOX emissions." At this point baseline
operating conditions (S^ in Table 4) at full load and two different burner
register settings have been characterized. Also, HOX emissions have been deter-
mined for normal and minimum excess air operation.
• Staged Combustion - Full Load
Staging the combustion process to reduce !IOX emissions on boilers
not equipped with overfire air ports, as discussed below, requires taking
burners out of service in various combinations. These burners are used as air
ports by stopping the coal flow to the burner but permitting air flow to the
furnace to continue, generally (but not always) with the burner dampers in
the wide open position to obtain the optimum effect. Individual burners on
sotae boilers equipped with burner line valves can be removed from service by
shutting the burner line valve. Other burners supplied by the same pulverizer
can be kept in service firing coal in the usual manner. The choice of firing
patterns on such units can be varied widely as desired. Recent trends in fuel
burning equipment design have eliminated burner line valves due to problems
experienced with this type of equipment. Staging the firing pattern in units
without burner line valves is limited to removing all burners from service
which are served by an individual pulverizer. This, of course, restricts
the choice of firing patterns and, in nany instances, dictates testing for 1JOX
emission reduction at lower loads. Modern design practice, however, permits
full load operation in most units with one pulverizer out of service. Greater
- 27 -
-------
d5350
flexibility in the choice of staged firing patterns is thereby accorded
in these units. Appendix A covers examples of staged firing where entire rows
of burners served by an individual pulverizer must be removed from service.
The following example discusses staged firing patterns permitted by individual
burner removal.
With combustion air restored to normal levels after test Ho. 4,
adjustments may now be made in boiler operating conditions to stage the combus-
tion process. Following the test plan in Table 4, coal supply is shut off to
the outside burners Dj and D^ in the bottom row of burners as shown on the burner
configuration diagram in Table 4. Air supply to these burners is opened wide
so that combustion air only is introduced through the Dj and D^ burners underneath
the combustion zone. Burner registers remain at 20% open on the other burners;
the same setting prevailing after test No. 4. After completion of these
adjustments, the first staged firing combustion modification pattern (82,
Table 4) has been characterized (Step A) and the boiler is ready for determination
of NOX emissions.
NOX emission data are then recorded for staged firing pattern 82
with the boiler operating at full load under normal air conditions and
burner registers set at 20% open on the burners firing coal. After these
recordings are complete, excess air level is again cautiously reduced to
minimum levels following previous guidelines and NOX emission data is again
recorded. Test Nos. 5 and 6 (Table 4) will then be complete for staged
firing pattern 82 and 20% burner register settings.
Following the flow plan in Figure 2, excess air level is again
increased to normal before making any adjustments in the combustion pattern
after which burner registers (except on DI and 04) are reset to the 60%
open position. 1IOX data are then recorded for test No. 7 (normal air, 60%
register setting) followed by a reduction in excess air to minimum levels
and recording of data (test No. 8) as previously described. 110 emission
reduction capabilities for the 82 staged firing pattern, at this point,
have been determined for 20% and 60% burner register positions. It nay be
desirable to investigate further the optimization of burner register positions.
If so, additional tests must obviously be run in accordance with the procedures
detailed. However, at this point, it may be advisable to proceed with the
exploration of alternate staging patterns when more data will be available to
judge the best route for further exploration.
After test Ho. 8, excess air is restored to normal in preparation for
a change to staged firing pattern 83 (Table 4). Burners Dj and D^ are placed
back in service, readjusting the burner registers to 60% open, and top burners
AI and A4 are taken out of service. Secondary air supply on burners Aj and
A^ are opened wide resulting in a supply of air over the top cf the combustion
zone. This produces a two-staged effect with combustion at fuel rich conditions
in the bottom of the furnace with an excess of air at the top of the combustion
zone for burn out. As mentioned previously, in this example, the boiler
does not have the capability of maintaining full load with the four top burners
, A£, A3 and A^) out of service. A compromise with the two "wing" burners
and A4) used as air ports is, therefore, investigated.
- 26 -
-------
c5350
After combustion conditions have stabilized, NOX emission data is
recorded for test No. 9 at full load, staged firing pattern 83, burner registers
set at 60" open (except on burners A^ and A^ which are wide open). This is
followed by test No. 10 at nininun excess air in the same manner as previously
described. Excess air is then increased to nomal, burner registers are closed
down to 20°' open (except on A^ and A^) and emission data recorded for tests
11 and 12 following the sane procedures. This completes definition of the
83 staged firing pattern for 20 and 60% open register positions as indicated
in the test plan example in Table 4. It will be desirable to investigate other
staged firing patterns in order to optimize N0:c emissions. For the sake of
clarity, however, these will not be discussed here except to indicate a fev
possibilities, such as, using burners A2 and A3, B^ and B/+, 62 and 83 etc. as
air ports on air only. Investigative procedures remain the same for these
patterns, as previously described and as shown on the flow plan in Figure 2.
Information in Appendix A provides further details on more extensive testing.
• Baseline Operation - Reduced Load
Attention is now directed to investigations at reduced load as shown
in Table 4. The choice of boiler load at which the tests will be conducted
should be determined by the load carrying ability of the boiler with one
pulverizer out of service. In the example this means four burners out of
service, since each pulverizer supplies four burners. The boiler operators,
who are well aware of these capabilities, should be consulted to establish
these conditions. In the test plan shown in Table 4, maximum boiler load with
four burners out of service was established at 110 MW.
With all pulverizers and burners in service but load on the boiler
reduced to the 110 MI.7 level and excess air and burner register settings reset
to normal and 60% open positions, respectively, NOX emission data are recorded
for test No. 13. After reducing excess air to the minimum level, data for test
No. 14 is recorded. Following described procedures, burner registers (except
those on air only) are again closed down to 20% open and test Nos. 15 and 16
are run in sequence at high air and low air, respectively. This characterizes
and establishes baseline conditions and emissions at reduced boiler load at 20
and 60/» open register settings with all burners in operation.
• Staged Combustion - Reduced Load
After test No. 16, excess air is increased to normal and the pulverizer
supplying burner row A is removed from service. Secondary air dampers and
registers on burners A^, A2, A3 and A^, however, are opened wide to supply a
sheet of combustion air over the top of the burner zone to effect complete
two-stage combustion. Burner registers in the bottom three burner rows remain
set at 20% open as they were for test No. 16. Cranping down on the combustion
air flow in the bottom burners increases the flow of air to the top burner row
enhancing the two-stage effect. Test Nos. 17 and 18 are then run sequentially,
as in the previous examples, and NOX emission data are recorded for staged
combustion pattern S/t (Table 4). Excess air is then again increased to
normal and burner registers on the active burners are adjusted to the 60% open
position. Tests 19 and 20 are then recorded completing data on the 84 staged
firing pattern.
- 29 -
-------
e5350
Referring to the test plan in Table 4, the boiler is again in readi-
ness to proceed with the second reduced load, staged firing pattern (85) after
excess air is restored to normal levels. Pulverizer A is placed in service
firing through burners A2 and A3 with registers adjusted to 60% open. Coal
flow is shut off on burners AI and A^ and registers remain 100% open to conbus-
tion air flow. Coal flow on burners B£ and 83 is shut off and registers are
opened wide. Overfire air is thus introduced through burners A^, A4, Bo and BT
with registers on these burners 100% open. All other burner registers are
maintained at 60% open for test Nos. 21 and 22. In the manner described above
test ITos. 21 and 22 are run followed by a change in burner register settings
to 20% open on active burners and the recording of NOX emission data for test
Nos. 23 and 24. At this point, the second staged firing pattern investi-
gation at reduced load is finished and the experimental test plan in Table 4
has been completed.
As mentioned previously the test plan in Table 4 was abbreviated for
the sake of clarity to demonstrate the procedures and techniques involved in
conducting a 1IOX emission reduction program. The principles employed are
applicable to and may be used to conduct an expanded more detailed program.
Appendix A provides examples of more extensive test programs which may be
used in planning specific IIOX emission reduction test programs.
7.4 Analyzing the K.esults
As indicated in the previous sections, Table 4 not only outlines a
suggested experimental test plan to be conducted sequentially in the order of
the test plan numbers, but the plan can also be used to serve the dual purpose
of recording the results of each test as the data becomes available. A brief
study of the test plan after completion of the test program, or as the tests
progress, givers an indication of which combustion mode is most effective in
suppressing NOX emissions. Also an insight may be gained from the data on
other operating modes which might bear further investigation.
For example, referring to the first four tests in Table 4 conducted
under baseline operation with all burners in service, it can be seen that test
No. 2, conducted at low excess air with burner registers set at 60% open
results in lowest 1TOX emissions of 491 ppm. This is approximately 15% lower
than baseline operation at normal air of 557 ppm in test Ho. 1 under the same
operating conditions. Results of these four tests also show that closing the
burner registers to 20% open (test Nos. 3 and 4) is less effective than with
register settings of 60% open. With only these data available, one would be
inclined to believe that closing the burner registers is the wrong way to go.
Also one could speculate that register settings greater than 60% open might
result in even lower emissions. Additional tests with different burner register
settings could, obviously, be run to optimize NOX emissions at the stated
conditions but these would not provide the complete answer.
If staged firing combustion modifications are then investigated as in
test NOG. 5 through 12, it can be seen that MOX emissions may be reduced to
345 ppn (test No. 12), a reduction from baseline operation of 40 percent, it
may also be noted that these lower emissions can be achieved with burner
register settings at 20°: open in the active burners. Reducing the air supplied
- 30 -
-------
d5350
to the active burners decreases the stoichionetric air ratio in these burners
and forces nore air to burners Aj and A4 (registers 100°' open) used as air
ports. Thus, under staged firing conditions, closed down register settings on
the active burners gives better emission reduction performance, which is contrary
to what one night expect judging fron the data obtained in test Uos. 1 through 4.
Tests at reduced load (test tlos. 15 and 24) provide a similar
pattern of results as those conducted at full load except that emissions are
generally lower due to the lower load and staged firing pattern S^. With
four burners (A]_, A2, A3 and A^) used as overfire air ports, greater
NO,, reduction is achieved. For instance, test No. 13 at baseline conditions
with registers at 60% open and all burners in service produces NOX emissions
of 629 ppm at normal air levels. Staged firing test No. 18 at low excess
air with burners Aj, A2> A3 and A.^ used as overfire air ports and active
burner registers set at 20S open, reduces NOX down to 297 ppra. This is a
reduction of almost 53 percent over baseline conditions and represents
the lowest emissions obtained of the firing patterns investigated in the
example.
To better understand the results achieved it is also helpful to
analyze the data statistically and plot the least squares regression lines
fitted to the data points corresponding to the various firing patterns.
Examples of this type of analysis are given in Appendix A.
The step-by-step procedures outlined and discussed in the above
examples should enable the experienced boiler operator to plan, set up and
execute his own NOX emission and efficiency optimization test progran. Once
the test program has been completed, the data tabulated and analyzed, and
the optimum combustion modes established for 1IOX reduction, tests may be
repeated at these operating modes and fly-ash samples obtained in accordance
with established procedures so that carbon loss data (% carbon on fly-ash)
may be obtained. TJith these data in hand, boiler efficiency calculations
can be made. It is possible that excess air levels that are too low may
result in lower boiler efficiency due to increased carbon loss with some
operating modes. If this happens, excess air levels should be increased and
further data taken at the same operating mode to bracket or optimize
efficiency, perhaps at the expense of a compromise in NOX emissions.
However, if the 200 ppm maximum CO level is generally observed for the test
involved, experience shows that boiler efficiency should not be impaired
when operating under "low NOX" emission conditions.
7.5 Translating Optimized NOX Emission Into
Operating Procedures
At this point all of the tests have been run, the data have been
tabulated and analyzed, and any additional tests required to further define
optimum combustion modes for NOX reduction have been completed. The problem
now is to translate the test data into meaningful operating procedures for the
boiler operators so the boiler nay be operated over its normal range under low
NOX emission conditions. Normally, existing boilers are exempt from NOX
- 31 -
-------
c5350
New Source Performance Standards (NSPS); these standards applying only to new
boilers designed and built to meet NSPS after December 23, 1971, or subsequent
NSPS revisions. On new NSPS boilers, operating procedures would be established
by the boiler manufacturer's engineers in cooperation with utility personnel in
order to ensure compliance with air pollution regulations and the boiler
manufacturer's guarantee. In other cases, where utilities may desire to
operate existing boilers under low NOX conditions, operating procedures must
be established by the utility personnel after the low NOX emission test
program has been completed.
For example, using the data developed in the simplified test plan
shown in Table 4, one of the major operating parameters of concern to the
operator would be load and load range. In the test program it was demonstrated
that full load could be developed either with all burners in service or with
a maximum of two burners out of service (using the burners as air ports)
but tests were only run at the 125 MW, full load rating. It was further
determined that a maximum of 110 MW (reduced load) could be produced with
four burners out of service. Between 110 and 125 MW, therefore, a choice
can be made between using all burners or operating with two burners out of
service (staging). Referring to Table 4, lowest NOX emissions firing
all burners (491 ppm) were achieved in test No. 2 with secondary air registers
60% open. Staging the combustion process at full load with burners Aj and
A^ on air only, optimized NOX emissions at 345 ppm at the 1.7 percent 02
level (test No. 12). Consequently, from a NOX emission standpoint, loads
between 110 and 125 MW should be carried with the wing burners A± and A^
used as air ports. However, the lower range of 110 MW may require a somewhat
higher level of air in order not to exceed the arbitrary 200 ppm maximum CO
level previously discussed. In order to properly define these operating
parameters, therefore, another test should be run at 110 MW with coal shut
off in the wing burners Aj and A4 and air 100% wide open in these two
burners (20% open in active burners). Again, operating mode adjustments
would be made at normal excess air levels, after which excess air would be
decreased to the point where CO (in the lowest 02 probe) would approach
the 200 ppm maximum. Emission and operating parameter data would then be
recorded, as before.
With these data in hand, operating parameters recorded during the
tests may be used to establish procedures for low NOX operation at loads
between 110 and 125 MW. Within this load range, staged firing with burners
AI and A4 on air only (registers wide open) and active burner registers
cramped down to 20% open (as in test No. 12) would be employed. Minimum
02 levels of 1.7% would prevail at 125 MW possibly ranging upward to
somewhat higher levels as established for operation at 110 MW. Note that
these are minimum levels determined for "optimum" NOX emissions. Realis-
tically, some higher level of air will be necessary, effecting a compromise
with NOX emissions, in order to provide some degree of flexibility in
operation. This necessary "cushion" must be established for each individual
job and will be dictated by the control capabilities and the required degree
of NOX reduction. In the example, perhaps 2.0 percent 02 would be
targeted for operation at 125 MW and 2.5 percent for 110 MW to provide the
required flexibility.
- 32 -
-------
c5350
Other operating indicators would also be observed and established
as parameters for the definition of operating conditions for the different
loads. For example, in addition to 02, windbox to furnace differential
pressures recorded during the tests, may serve as additional operating
guideposts. If windbox to furnace pressure differentials of 1.5 and 1.3
inches of H20 were recorded for full load (125 MW) and 110 MW loads,
respectively, operating procedures would use these minimum levels as guideposts
to operation with corresponding interim levels at loads between 110 and 125
MW. Similarly, steam flow/air flow levels and other pertinent operating
parameters such as, forced draft fan discharge pressure, etc., which have
been recorded for individual tests, can be incorporated in the operating
procedures in a like manner. In the example, operating parameters and
procedures for staged firing (A^ and A^ burners used as air ports) for
boiler loads of 110 MW through 125 MW, respectively, would consist of air
levels ranging between 2.5 and 2.0 percent 02; windbox to furnace differ-
ential pressures ranging from 1.5 to 1.3 inches of t^O minimum; forced
draft fan discharge pressure, e.g., varying from 25 to 24 inches of 1^0
maximum and corresponding levels of steam flow and air flow recorded at
these loads during the tests. Thus the operator is provided with a simplified
number of important guidepost parameters delineating the type of operation
required for low NOX emissions.
For loads below 110 MW, a similar approach would be used in
establishing operating procedures. In the example in Table 4, it may be
seen that staged firing pattern with the top row of burners (Aj_, A2, A3,
and A/i) used as air ports (air wide open) and registers on active burners
adjusted to 20% open, gave lowest NOX emissions, i.e., 297 ppm at 2.7%
Oo (test No. 18). These emissions were achieved at 110 MW boiler load and
no data exist for lower load operation. Accordingly, another test or tests
should be run to establish operating conditions at minimum load for this
combustion mode with the top row of burners out of service. In a fashion
similar to that described before with the boiler operating at 110 MW and
normal air levels, all burners in service, shut off the coal flow in burners
Ai» A2» ^3 and A4 and open the registers to the wide open position.
Re-adjust registers on the 12 active burners to 20% open. Decrease boiler
load to a comfortable minimum capability with 12 burners in service. Decrease
excess air to minimum values (200 ppm CO maximum) and record emission and
operating data. Note where a wide load range spread is encountered, inter-
mediate loads should be run to define operating parameters more closely.
Using these data, pick out the important operating parameters, such as, load,
combustion mode, 02 level, minimum windbox to furnace pressure differential,
forced draft fan discharge pressure, steam flow, air flow, etc., and incor-
porate these data into the operating procedures for these loads.
Pertinent operating parameters for the simplified experimental
test plan example (Table 4), therefore, might appear as shown in Table 5
below, which contains the fundamental basic ideas. Table 5 may be expanded
as required for more extensive test programs.
The description of the translation of NOX emission test data
into operating procedures, detailed above, is intended only as an example
for those not completely familiar with boiler operation. Experienced boiler
operators, such as those who undoubtedly would be intimately involved in any
NO reduction program, should have little difficulty setting up operating
parameters and procedures once the basic data is on hand.
- 33 -
-------
d5350
TABLE 5
OPERATING PARAMETERS FOR LOW NOX OPERATION
Burners Used as A^, A4 Aj, A2, A3 & A4
Overfire Air Ports*
Active Burner
Register Position, 20% Open 20% Open
Load, MW 125 110 110 75
02 Level, % Min. 2.0 2.5 2.7 3.2
Windbox/Furnace
Press. Diff., in. H20 Min. 1.5 1.3 1.3 1.4
Forced Draft Fan
Discharge Press., in H20 25 24 21 22
Min.
The appendices which follow provide more detailed and general
background information helpful to anyone conducting a NOX emission reduction
program. For example, two comprehensive test plans are included in Appendix
A for larger, more complicated boilers. One of these is for a 300 MW front
wall fired unit (also applicable to horizontally opposed firing) and the
other for an 800 MW tangentially fired boiler equipped with overfire air
ports for NOX control. Coal quality is discussed in Appendix B from the
viewpoint of the boiler operator and potential effects on the NOX emission
reduction program. An understanding of the parameters affecting NOX forma-
tion is offered in Appendix C which discusses the fundamentals of the
formation of combustion generated pollutants with special emphasis on
pulverized coal firing. Similarly, Appendix D provides helpful background
information on the significance and fundamentals of pollutant measurements.
Finally, Appendix E discusses the need for optimization of boiler efficiency
from the relationship of the NOX reduction program.
- 34 -
-------
d5350
APPENDIX A
COMPREHENSIVE FIELD TEST PROGRAMS FOR
NOX EMISSION REDUCTION ON LARGE BOILERS
Section 7 covered the essential, underlying procedures employed in
planning and carrying out a NOX emission reduction program on a relatively
small, pulverized coal fired, utility boiler. Basic, fundamental principles
were illustrated using a simplified, abbreviated test plan example deliberately
kept simple in order to avoid confusion. The basic procedures covered in
detail in Section 7 are the same as would be employed in more extensive NOX
emission reduction test programs for larger boilers of more complicated designs.
The fundamental ABC procedures discussed in Section 7, therefore, must be
thoroughly understood, otherwise undertaking a more comprehensive NOX reduc-
tion program would be futile.
In this appendix, examples of more comprehensive field test programs
are given for two larger boilers of more complicated design. The first is a
300 MW, front wall fired, twin furnace boiler and the second an 800 MW tangentially
fired boiler with tilting burners for steam temperature control and overfire
air ports for the control of NOX emissions. The salient features of each
boiler design are considered and incorporated in the test programs so that the
effects on NOX emissions may be fully explored. These examples are representa-
tive of the types of boilers and design features encountered in modern utility
power plants and the experimental test programs may be used as a guideline in
formulating specific NOX emission reduction programs tailored to individual
needs.
Test Variables - Front Wall Fired Boiler
One of the first steps in the approach to a successful NOX emission
reduction program is to develop a test program plan. Operating variables
(which are specific for each boiler) should be reviewed for inclusion in the
experimental program. This is best illustrated by the example in Table A-l
showing a plan for a 300 MW, twin furnace, front wall fired boiler, firing
pulverized coal, in 24 burners; twelve burners in each furnace, three rows high
and four across. Operating variables included in the plan are:
Load - High (Lj), medium (L2>, and low (13).
Excess Air - High or normal (A^) and low (A2>.
Staging - Sj, baseline (normal fitting).
-82, biased fired (firing minimum coal in the top
burners and maximum in the lower burners).
- 83, staged, top burners used as overfire air ports
(air only).
- 84, baseline (normal firing for low load) top
burners out of service (air and coal off).
- 35 -
-------
TABLE A-l. TEST PROGRAM PLAN
300 MW BOILER-FRONT WALL FIRED
Burner
Staging Patterns
St Base
Firing)
S Biased
Firing
(Top Burner a
Fuel Lean)
83 Staged
Firing
(Top Burner!
Air Only
S. B*»e
(Top Burner*
Air and Coal
Off
Air Register
Settings
R, Normal
701 Open
R2 Closed Down
401 Open
Rl Normal
701 Open
R2 Closed Down
401 Open
R! Normal
701 Open (2)
R2 Closed Down
401 Open
R3 Closed Down
Ignite Position
R! Normal
701 Open
Lt - (300 Ml)
AI Normal
Excess Air
(1) 3.91
900
(5) 3.9*
936
(3) 3.2*
SOS
(7) 3.91
758
A2 Low
Excess Air
(2) 1.97.
746
(6) 1.951
675
(4) 1.351
600
(8) 1.81
570
L? - (200 »')
(Hax.T.oad-1 Mill Off)
AI Normal
Excess Air
(11) 5.91
739
(17) 5.71
740
(13)
(15)
(9) 6.81
500
(27) 6. 61
557
(19) 63%
661
(21)
A2 Low
Excess Air
(12) 3.151
663
(18) 3.51
735
(14)
(16)
(10) 3.31
213
(28) 3.11
237
(20) 2.31
398
(22) 3.21
609
-
L3 - (150 MM)
AI Normal
Excess Air
(25) 7.01
448
(29) 8.01
547
(24) 6.61
590
A2 Low
Excess Air
(26) 4.31
222
(30) 4.57.
314
(23) 5.07.
506
4.07.
465
(1) Numbers in parenthesis are test numbers. Figures are 02 values In percent and NOX In ppm (3% 02, Dry Basis).
(2) Registers on all burners set at 701.
(3) Secondary Air Registers: Top row of burners set at 401 open, bottom and middle registers set at 701 open.
(4) Top row of burner registers s'et at (Ignition position), bottom and middle row burners at 701 open.
-------
c5350
Burner Registers - Rj, normal (open) secondary air register
setting 70% open.
- R2 (partially closed down) secondary air
registers setting 40% open.
- R3 (closed down) secondary air registers set
at the ignition position for burner start
up.
Test Program Plan - Front Wall Fired Boiler
For example, following the plan in Table A-l, the operating mode for
Test No. 1 is baseline (S^) or normal operation at full load in the manner the
boiler is normally fired. A set of control room and emission instrumentation
readings are recorded for Test No. 1. In the example, boiler emissions average
900 pptt NOX (at 3% 62) with the boiler operating at an average of 3.9 percent
oxygen. The first and second objectives (Steps A and B) are complete at this
point, i.e., the operating mode has been established and emissions have been
measured •
step by Step Procedures - Front Wall Fired Boiler
Combustion air is now decreased to establish firing conditions for
Test No. 2. This is done in small controllable steps keeping watch of 02 and CO
levels (in cases where multiple probes are used, sampling is from the probe
with the lowest 62 reading). Experience shows that background CO levels for
normal operation on utility boilers average between 30 and 40 ppm. Oxygen
levels will be observed to drop as the air is decreased, with very little
change in CO. A point will be reached where CO levels commence to increase.
Continue to slowly reduce combustion air until the CO reading (from the lowest
0? probe) reaches a level of about 200 ppm. This point is defined as the
minimum acceptable excess air level. Reducing excess air beyond this point
normally will increase stack opacity to unacceptable levels. On some boilers
under some operating modes, especially at low loads, it may not be possible
to reach the 200 ppm CO level. Stack opacity and flame appearance observations
should be made for instability as the combustion air is decreased. A good
criteria for the latter is the differential pressure between the windbox and
the furnace. Care should be exercised in reducing excess air below the point
where the windbox/ furnace pressure differential drops below 1"
Once the minimum excess air level has been established no further
changes should be made and firing conditions should be allowed to stabilize (30
minutes). A set of readings should then be taken to record operating parameters
for Test No. 2. In the example (Table A-l) the average 62 measured was 1.9%
with NOX emissions of 746 ppm and the three objectives (ABC's - Section 7.2)
of the first step of the test plan for baseline operation with normal register
settings have been achieved.
At this point, following the test plan, excess air is increased to
normal levels prior to changing to the biased firing mode (82). Coal rate to
the top mill or pulverizer is then decreased gradually while increasing the
coal rate to the bottom mills. This is accomplished while maintaining load at
a constant value. This combustion mode is an attempt at a quasi-staging
pattern without taking a pulverizer out of service. Full load capability could
not be maintained on this boiler with one mill off. Under these conditions,
- 37 -
-------
d5350
combustion will be lean in the top burner rows and rich in the botton two
burner rows thus effecting a partially staged pattern. Once conditions have
stabilized, Test No. 3 data are recorded as in the previous procedure described
for Test No. 1. Excess air is then decreased as described above until maximum
CO levels reach 200 pptn. Data for Test No. 4 is then recorded after operating
equilibrium has been established (30 min.). Referring to Table A-l, it nay be
seen that emissions in the example were 805 ppm NOX at high air levels of 3.2%
02 for Test No. 3 and 600 ppm NOX for Test tlo. 4 at low excess air (1.35% 02).
Again, the ABC steps for the second phase of the test phase (Tests 3 and 4)
have been achieved at this point.
In the example in Table A-l, after Test No. 4, excess air is again
increased to normal levels, fuel rate is evened out on all three pulverizers
and burner registers are partially closed to the R2 position, returning to the
baseline node of operation but with the burner registers on all burners partially
closed. Test data for Test No. 5 is then recorded after which combustion air
is decreased to minimum values and data for Test No. 6 is recorded. Emissions
for baseline operation at full load (Sj) have now been characterized in Test
Nos. 1, 2, 5, and 6. It may be noted that NOX emissions of 675 ppm are lowest
for the operating mode of Test No. 6, a reduction of 25% or 225 ppm NOX from
baseline values of 900 ppm in Test No. 1.
After Test No. 6, with the secondary air registers still at position
R2 (40% open), excess air is raised to normal levels, the mills are biased as
before, and data is recorded for Test No. 7. Afterwards, excess air is reduced
to minimum levels and data for Test No. 8 is recorded, completing emission
characterization for biased firing conditions at full load. Note that emissions
in Test No. 8 have been reduced by 37% to 570 ppm NOX. This completes the
optimization of NCX emissions at full load since this particular boiler is
incapable of maintaining full load capacity with one pulverizer out of ser-
vice as required for firing patterns 83 and 84.
The same procedure is repeated at 200 MU load for Test Nos. 11, 12,
17, and 18. Staging the firing pattern, as indicated for 83, is now possible
at this load with one mill off. With the top burners used as overfire air
ports, the effect of varying the amount of overfire air is investigated by
operating with maximum (Test Nos. 9 and 10), intermediate (27 and 28) and
minimum (19 and 20) overfire air, respectively, by making adjustments in the
burner secondary air registers as indicated on Table A-l to vary the amount of
overfire air flow. From the example in Table A-l, note that minimum NOX emissions
of 213 ppm at 3.3% 03 are achieved in Test No. 10 with maximum overfire air
flow. Note, also, that emissions increase to 398 ppm NOX in Test Ho. 20 for
the 83 operating mode when the use of the top row burners as overfire air ports
is practically discontinued with the top burner registers set at the ignition
position.
Low load tests at 150 IIW are then conducted in a similar manner as
indicated in the test plan (Table A-l) for firing patterns 83 and SA (tes^ Nos
23, 24, 25, 26, 29 and 30, respectively). At this load, staging the firin«
pattern (83) usins the top row of burners as overfire air ports with burner
registers open as wide as possible, Pq (70% open), again produces lowest einissio
(Test No. 26) similar to Test No. 1C at 200 11W load. Note, however, that it S
- 38 -
-------
d5350
is not possible to reduce excess air as much (4.3% ©2) in Test No. 26 at the
150 KW load as compared to Test Ho. 10 (3.3% 03) at 200 11W and NOX emissions
are correspondingly lower at the higher load, i.e., 213 ppm vs. 222 ppm.
Analyzing the Results
After the field test work has been completed, the emission data
tabulated on the test program plan (Table A-l) may be plotted so as to better
understand what has been achieved. Figure A-l is a plot of ppm NOX (3% 02, dry
basis) vs. average percent Oo in the flue gas for each test run condition.
Lines have been drawn through the data for the normal (or high) and low excess
air firing points for each operating condition (boiler load, firing pattern and
secondary air register setting). Referring to Figure A-l, it may be seen
that all of the operating variables had a significant effect on NOX emission
levels. Baseline operation (Test No. 1) resulted in an emission level of 900
ppm NOX. Test No. 8, under low excess air biased firing conditions, is clearly
shown to produce the lowest NOX emissions (570 ppm) at full load. The boiler
should be operated in this manner if NOX emissions are to be limited. However,
if faced with stricter emission standards, consideration should be given to
installing separate overfire air ports so that more effective staging patterns
could be employed at full load to potentially achieve further reductions in
emissions.
Figure A-l clearly shows that staged firing (83) results in lowest
emission levels at intermediate load (200 MVJ) and low load (150 »W) . Test
Nos. 10 and 26, respectively, under low excess air firing product 110X emissions
of 213 and 222 ppm. With separate overfire air ports it might even be possible
to effect further emission reductions.
Test Program Plan-Tangentially Fired Boiler
Table A-2 shows another test program plan developed for full load
firing of an 800 MW, twin furnace boiler with pulverized coal. Burners fire
in a tangential pattern and can be tilted for steam temperature control. On
this unit separate overfire air ports were built into the top of the windbox
above the top row of burners. Sufficient pulverizer capacity was included in
the design to operate the boiler at full load with one pulverizer out of
service.
Operating variables to be investigated include:
\
Load - Full
Excess Air - Normal (A^) and Low (A2)
Staging - Si, Normal firing (no overfire air)
82, Staged firing (top row of burners used as
overfire air ports)
84, Staged firing (25% OFA Ports)
S5, Staged firing (50% OFA Ports)
S6, Staged firing (75% OFA Ports)
87, Staged firing (100% OFA Ports)
- 39 -
-------
900 —
Full Load (283-296 MW)
4 5
Average % Oxygen in Flue Gas
Figure A-l. 300 MW Boiler-Front Wall Fired.
(PPM NO (3% 0 , Dry) vs Average % 0^ .
-------
TABLE A-2. TEST PROGRAM PLAN
800 MW BOILER - TANGENTIAL FIRING WITH OVERFIRE AIR
(TEST RUN NO., % OXYGEN AND PPM NOX [3% 02, DRY BASIs]
FULL LOAD-800 MW)
Burner
•P- s Tilt
Firing
Pattern
S-£ Normal
Firing
S Top Row
Air Only
S3* 25% OFA
S^* 50% OFA
S5* 75% OFA
S,* 100% OFA
b
A- - Normal Excess Air
11 -10°
to -15°
(3) 3.9%
332
(19) 5.2%
346
T2
Horizontal
(1) 4.9%
343
(21) 5.4%
329
T~ + 10°
to 15°
(5) 5.2%
378
(9) 5.0%
318
(11) 4.0%
332
(13) 3.9%
310
(15) 3.8%
292
(17) 5.2%
354
(17A) 4.6%
346
T> + 25°
(7) 4.8%
492
A» - Low Excess Air
TI -10°
to -15°
(4) 3.8%
314
(20) 3.5%
271
T0
2
Horizontal
(2) 4.2%
309
(22) 3.8%
289
T3 +10°
+15°
(6) 3.6%
298
(10) 3.6%
259
(12) 3.6%
328
(14) 3.6%
326
(16) 4.0%
289
(18) 3.6%
282
(ISA) 3.6%
288
T4 +25°
(8) 3.8%
411
I
-p-
\->
I
*0verfire air registers set at 25% open for 83, 50% open for 84, 75% open for 85 and 100% for 85.
-------
d5350
Burner Tilt - Tlf -10 to -15° below horizontal
To, Horizontal
T3, +10 to 15° above horizontal
?4, +25° above horizontal
Step-By-Step Procedures - Tangentially Fired Boiler
The test plan in Table A-2 was devised to first investigate emissions
from the boiler fired as a normal tangentially fired boiler without using
the overfire air ports. Following the test plan, set the burner tilt at
horizontal (0°) with the boiler operating at full load. After operation
has stabilized, record the required test data for Test No. 1 with horizontal
burner tilt, normal firing pattern S^. When this complete, reduce the
combustion air to the boiler in snail increments until CO levels reach 200
ppn maximum on the sampling probe with the lowest Oo. Record the data for
Test No. 2. As illustrated in the example in Table A-2, NOX emissions are
reduced by about 10%, fron 343 to 309 ppra between Test Hos. 1 and 2, respec-
tively. The three objectives (ABC) for this phase of the program (T2, Sj)
are now complete.
At this point excess air is increased to normal levels and burner
tilt is reset to -10 to -15 degrees (downward) from horizontal for the
second operating mode to be investigated (T^, 8^). After operation
stabilizes, data are recorded for Test No. 3. Excess air is then reduced as
previously discussed, until CO reaches a maximum (200 ppn) in the lowest
Q£ sampling probe. Data for Test No. 4 is then recorded after operation
has reached equilibrium. In the example (Table A-2) 1IOX emissions decrease
from 332 to 314 ppn for Test IIos. 3 and 4, respectively, and the three
objectives for the second phase of the program (Tj, S^) are complete.
Preparing for the third operating mode (T3» Sj), excess
air is again increased to normal and burner tilt is reset slowly in careful
increments to the +10° to +15° position (above horizontal). Data for
Test No. 5 is now taken after which excess air is again reduced to the
maximum CO limitation (200 ppn) and data for Test No. 6 is recorded.
Emissions of 378 and 298 are recorded for Test Nos. 5 and 6, respectively,
and testing of the third operating mode is finished.
The same procedure is repeated for the fourth operating mode
("4» Sj) burners tilted to +25° (above horizontal) and data are
developed for Test IIos. 7 and 8. This completes the investigations for the
four operating modes devised to show the effect of burner tilt in reducing
NOX emissions.
Attention is now given to the investigation of staging the combus-
tion patterns in the furnace using overfire air 83, 83, 84, 85, and Sg.
Tests begin with the 82 operating node where staging consists of using the
top rows of burners as overfire air ports as might be necessary if overfire
air ports were not built into the boiler. Following Test No. 8, the burner
tilt is reset slowly to the +10° to +15° above horizontal fron the
+25° position. After superheat and reheat steam temperatures are stabilized
excess air is once again increased to normal. The top row pulverizer is
slowly taken out of service shutting coal off to the top burners while
- 42 -
-------
d5350
increasing coaL flow in the remaining pulverizers and burners to maintain
constant load. Registers on the top burners are not closed but left open
as if the burners were in service. Overfire air ports, as in all previous
tests, are left closed. Once boiler operation has equilibrated after making
these changes, the test data are recorded for Test No. 9, operating mode
TS, S2- Excess air is then slowly decreased until the 200 ppm max. CO
level (in the probe with the lowest 02) is reached and the test data are
recorded for Test No. 10. Referring to the example in Table A-l, it may be
noted that NOX emissions are reduced from 318 to 259 ppm for Test Nos. 9
and 10, respectively. This completes operating mode T3, 82 characterization.
The next phase of testing concerns the investigation of the use of
the built-in overfire air ports in suppressing NOX emissions. After Test
No. 10 is complete, excess air is increased to normal and the top row
pulverizer and burners are put back into service in the reverse manner to
that described above. The overfire air dampers are then carefully opened
one by one to the 25% open position while carefully observing combustion
conditions and panel board instrumentation. It is recommended that one or
two dampers be opened at a time allowing sufficient time between opening
the next set of dampers for operation to stabilize so that abrupt upsets do
not occur. After all overfire ports are opened 25%, boiler operation should
be allowed to stabilize and a set of readings should then be recorded for
Test No. 11, operating mode 1$, 83. Afterwards, excess air should be
slowly decreased to minimum (200 ppm maximum CO) and another set of readings
taken for Test No. 12. In the Table A-l example, NOX emissions are reduced
from 332 to 328 ppm for Test Nos. 11 and 12, respectively, and conditions
for operating mode T3, 83 are complete.
This procedure is repeated for operating modes T3, 84; T3,
S5; and T3, 8$ respectively. However, since overfire air ports would
normally be used in the wide open position, the effect of burner tilt with
100% OFA would again be investigated as in modes Tj, 85; T£, 85; T3, 85;
and T^, 85. The same procedures as at the start of the test would be
repeated but, this time, the overfire air ports would be used (100% open)
rather than using the top row of burners as overfire air ports.
The test program plan discussed concentrated solely on investiga-
tions of NOX emission reduction techniques under full load or rated
capacity conditions. A similar program plan could be developed for inter-
mediate and low load operation. The techniques and procedures, however,
are the same as described above and will not be repeated here.
Analyzing the Results
Once the test program plan has been completed, it is helpful to
plot the average emission data for each test to better understand the
results. The emission data tabulated for the test plan program example in
Table A-l are plotted in Figures A-2, A-3, A-4 and A-5. The effect of
burner tilt on NOX emissions for full load, normal firing (no staging,
overfire air ports closed) is shown in Figure A-2 for the 82 operating
mode. Clearly it may be seen that the horizontal position (0° tilt)
produces minimum NOX emissions. With tilt at -10 to -15° downward
- 43 -
-------
d5350
missions are somewhat higher than the horizontal position due to the
concentration of heat in the bottom of the furnace. Burner tilt of +25°
upwards has no beneficial effect on NOX as there is minimum separation of
overfire, second stage air with first stage combustion.
The effect of burner tilt with staged firing and overfire air
ports 100% open is shown in Figure A-3, 87 operating mode. Horizontal
burner tilt is again shown to result in lower NOX emissions at higher
excess air levels. At low excess air, however, there is little to choose
from with respect to the effect of burner tilt on NOX emissions. Any
burner tilt appears to produce about the same 1IOX emissions at these
levels of excess air.
Figure A-4 is a plot of ppm NOX vs. percent oxygen for the 82,
83, 84, and 85 staged firing operating modes. The line labeled Sg
from Figure A-3 (+10 to +15° tilt) was drawn on Figure A-4 for comparison
purposes. Short lines parallel to the 85 line were drawn through the
averaged data for 83, 84, and 85 test runs. Lowest MOX emissions
resulted fron test Nos. 9 and 10 (operating mode 82) operating with the
top row of burners as overfire air ports with the top pulverizer shutoff.
The beneficial effect of increasing overfire air register openings from 25%
(S3) to 50% (84) and then to 75% (85) is apparent in Figure A-4. The
effect of changing overfire air register settings is more dramatically shown
in the plot in Figure A-5. Only test runs conducted with approximately equal
excess air levels (3.6 to 4.0% 02> and burners tilted at T3 (+10° to +15°)
are shown on Figure A-5 so the effect of overfire air damper settings on NOX
emission levels can be seen directly.
In summary, lowest NOX emissions occur on this boiler at normal
excess air levels when operating with the burners firing horizontally (0°
tilt). At low excess air levels, burner tilt position has little or no effect
on HOX emissions. When staging the combustion process, increasing the air
flow through the overfire air ports, as might be expected, has a corresponding
decreasing effect on 1IOX emissions reaching a minimum with maximum air
flow* This may be observed in Figures A-4 and A-5. Lowest NOX emissions,
however, resulted when staging the conbustion pattern using the top row of
burners as overfire air ports (Figure A-4).
- 44 -
-------
500
$%)
01
-H
CO
400
350
I 30°
-10 to -15°
Burner Tilt
+25° Burner Tilt
+10 to +15°
Burner Tilt
0° Burner Tilt
250
I
3.0
3.5 4.0 4.5 5.0 5m5
Average % Oxygen Measured in Flue Gas
EFFECT OF BURNER TILT
Figure A-2. 800 MW Boiler-Tangential Firing
Full Load, Normal Firingi
6.0
- 45 -
-------
500
400
CO
•H
CO
5
tw
Q
CM
O
350
300
+10 to +15 Burner Tilt
(-10 to -15°
Burner Tilt)
0 Burner Tilt
250
1
3.0
3.5 4,0 4.5 5.0 5.5
Average % Oxygen Measured in Flue Gas
EFFECT OF BURNER TILT
Figure A-3. 800 MW Boiler-Tangential Firing
Staged Firing-Overfire Air Dampers 100% Open.
6.0
- 46 -
-------
400
T
T
Note: Burner Tilt Constant
at +10° + 15° (T3)
350
Cfl
Cfl
-------
350
03
•H
CO
300
CM
O
250
Note: Burner Tilt Constant
at 4-10° to +15° (T3)
0
20
40
60
80
3.6% Oxyg«n
3.6% Oxygen
100
Overfire Air Dampers - % Open
Figure A-5. 800 MW Boiler-Tangential Firing
Effect of Overfire Air Dampers on NOX Emissions
Full Load-Staged Firing
- 48 -
-------
c5350
APPENDIX B
COAL CHARACTERISTICS
Although coals have been classified or ranked in a number of
different ways, from the standpoint of the power plant operator it is
generally sufficient to know that coals are ranked in the following broad
categories:
Anthracite
Bituminous
Guhbituninous
Lignite
It is the operator's job to cope with the burning of the coal
which is supplied to the plant and it is only when operating problems
develop, such as slagging, fouling, handling difficulties, etc., does he
become concerned with coal supply matters. Even then the decision made
generally involves a switch to a different quality coal within the above
broad classification ranks. Only when new plants are proposed does the
potential exist for the operator to be consulted concerning major changes in
coal supplies. The above classification then fails to closely define rank
and the need for further definition of boundary lines becomes important.
The American Society for Testing Materials has developed a classification
system which is probably referred to most frequently because it provides a
clear definition of coal ranking. The ASTM information is provided in Table
B-l and classifies coals by rank according to the degree of metamorphisn or
progressive alteration in the natural series of coals from lignite to
anthracite. Criteria are fixed carbon, volatile matter and calorific value
on a mineral-matter-free, fixed carbon and volatile matter basis. Note that
higher rank coals are classified on a dry, mineral-matter-free basis while
lower rank coals are classified on a moist, mineral-matter-free calorific
value basis.
Although the ASTM classification of coal provides more information
on the broad ranking of coals, the power plant operator needs more detail
regarding coal quality. Table B-2, which lists the analyses of various coals
classified according to rank, as defined by ASTM Standards, together with
wood and peat as representing the earliest stages of transformation of
vegetal natter through wood and peat to lignite, provides additional coal
quality information helpful to the boiler operator.
Coal Analysis
Coal quality may be defined or expressed in several ways through
various tests and methods of analyses. The proximate analysis provides
useful information on principal characteristics while the ultimate analysis
gives details of the exact chemical analysis of the coal without reference
to the physical forn in which the compounds appear. Broadly speaking, the
proximate analysis provides the power plant operator with information to
judge the combustion of the coal in his furnace. Details of the ultimate
analysis are required for combustion calculations.
- 49 -
-------
TABLE B-l. CLASSIFICATION OF COALS BY RANK3
Class Group
1 Meta-anthraclte
1 Anthracitic 2 Anthracite
3 SemianthraciteC
1 Low volatile bituminous coal
2 Medium volatile bituminous coal
II Bituminous 3 High volatile A bituminous coal
4 High volatile B bituminous coal
5 High volatile C bituminous coal
1 Subbituminous A coal
III Subbituminous 2 Subbi Luminous B coal
3 Subbituminous C coal
...... 1 Lignite A
IV Lignite 2 uj|lte B
Fixed carbon
limits, Z
(Dry, mineral-
matter-free
basis)
Equal or
greater
than
98
92
86
78
69
—
—
™*—
—
—
~
Less
than
98
92
86
78
69
—
"
—
—
--
Volatile matter
limits, Z
(Dry, mineral-
matter-free
basis)
Greater
than
2
8
14
22
31
—
""
—
—
—
Equal or
less
than
2
8
14
22
31
—
—
~~
—
—
—
Calorific value
limits, Btu/lb
(Moistb
mineral-matter
free basis)
Equal or
greater Less
than than
—
"
__
14,000d
13,000d 14,000
/11.500 13,000
\10,500 11,500
10,500 11,500
9,500 10,500
8,300 9,500
6,300 8,300
6,300
Agglomerating
Character
> Nonagglomerating
)
Commonly
Agglomerating
Agglomerating
^1
I
> Nonagglomerating
J
o
I
a - This classification does not include a few coals, principally noni-anded varieties, which have unusual physical and chemical properties and whii:h
come within the limits of fixed carbon or calorific value of the high-volatile bituminous and Subbituminous ranks. All of these coals either con-
tain less than 48 percent dry, mineral-matter-free fixed carbon or have more than 15,500 moist, mineral-matter free British thermal units per pound.
b - Moist refers to coal containing its natural inherent moisture but not including visible water on the surface of the coal.
c - If agglomerating, classify in low-volatile group of the bituminous class.
d - Coals having 69 percent or more fixed carbon on the dry, mineral-matter-free basis shall be classified according to fixed carbon, regardless of
calorific value.
e - It is recognized that there may be nonagglomeratlng varieties in these groups of the bituminous class and there are notable exceptions in high
volatile C bituminous group.
Reprinted with permission from the Annual Book of ASTM Standard*, Copyright American Society for Testing and Materials, 1916 Race St.
Pennsylvania 19103.
Philadelphia,
-------
IABLE B-2. PROGRESSIVE STAGES OF TRANSFORMATION OF VEGETAL MATTER INTO COAL
Fuel CUnlflcation
Wood
Put
Llgnlta
Lignlta
Subbituatlnoua C
Subbltanlnoua B
Subbltunlnoui A
Bltunlnoua High Volatlla C
BlCininoua High Volatile B
lltuainoua High Volatile A
BitUBinoua Medium Vol«tll«
BitUBinoua Volatile
Seolanthracite
Anthr«clt«
Mata-anthracita
Locality
North Dakota
Texa*
Wyoming
Wyoming
Wyoolng
Colorado
IllinoU
Pennaylvanla
Vmt Virginia
W*at Virginia
Arkaoaai
Pennsylvania
Rhode liland
*••*
.1
z •
3S
3 t
*•
46.9
64.3
36.0
33.7
22.3
15.3
12.8
12.0
8.6
1.4
3.4
3.6
5.2
5.4
4.5
Analysts oo dry baa IB
Proxlaate
V.M.
78.1
67.3
40.8
44.1
40.4
39.7
39.0
38. 9
35.4
34.3
22.2
16.0
11.0
7.4
3.2
F.C.
20.4
22.7
3$.l
44.9
44.7
53.6
55.2
53.9
56.2
59.2
74.9
79.1
74.2
75.9
82.4
Aah
1.5
10.0
12.1
11. 0
14.9
6.7
S.8
7.2
8.4
6.5
2.9
4.9
14.8
16.7
14.4
Ultimate
5
0.4
1.8
0.8
3.4
2.7
0.4
0.6
1.8
1.3
0.6
0.8
2.2
0.8
0.9
H2
6.0
5.3
4.0
4.6
4.1
5.2
5.2
5.0
4.8
5.2
4.9
4.8
3.4
2,6
0.5
C
51.4
52.2
64.7
64.1
61.7
67.3
73.1
73.1
74.6
79.5
86.4
85.4
76.4
76.6
82.4
N2
0.1
1.8
1.9
1.2
1.3
1.9
0.9
1.5
1.5
1.4
1.6
1.5
0.5
0.8
0.1
°2
41.0
30.0
15.5
18.3
14.6
16.2
14.6
12.6
8.9
6.1
3.6
2.6
2.7
2.3
1.7
«
_3
5
»fr
e *o >—*
ft *-* *
4J *r*
J3 «
£2
8835
9057
11038
11084
10596
12096
12902
13063
13388
14396
15178
15000
13142
12737
11624
V.M.
F.C.
S
"2
C
N,
Volatila Mattar
Flxad Carbon
Sulfur
Hydrogen
Carbon
Nltrogan
Raprtnted with panaiaaion from Co*bu«elon Boglnatrlag Inc., Windsor, Connecticut
0609S.
-------
c5350
For power plant work, as-received air dried or noisture-free
analyses are generally used. Classification of coals normally are nade on a
noisture and ash-free and moisture and nineral-free analyses basis. The
as-received analysis cones closest to the condition of the coal delivered to
the power plant and compares more closely with the as-shipped or as-fired
values. Loss or gain of noisture between the tine of sampling and analysis
depends greatly on the method of handling the sample, the type of coal, its
size and weather conditions.
Coal Quality
The following discusses briefly coal quality characteristics
included in the proxinate and ultimate analyses of Table B-2 and thej-r
relationship and effect on combustion problens in the power plant.
lioisturG
All coal contains sone natural moisture - fron 1% to 5% in most
eastern coals, and up to 45S in some lignites. This moisture lies in the
pores and forms a true part of the coal, being retained when the coal is air
dried. Surface moisture, on the other hand, depends on conditions in the
mine, and the. weather during transit. Moisture increases shipping costs and
decreases boiler efficiency through moisture losses from the boiler.
tfoisture generally is determined quantitatively in two steps: air
drying and oven drying. The air-dried component of the total noisture value
should be reported separately, bec'ause this Infornation is required in the
design and selection of coal-handling and coal-preparation equipment. Keep
in mind that it is the surface moisture that must be evaporated from anthracite
and bituminous coal before pulverization to maintain high grinding efficiencies.
Volatile Matter
Volatile natter is that portion of the coal which is driven off in
gaseous forn when the fuel is subjected to a standardized temperature test*
It consists of combustible gases, such as methane and other hydrocarbons,
hydrogen and carbon nonoxide, and nonconbustible gases. v Since the quantity
of volatile natter indicates the amount of gaseous fuel present, it affects
firing mechanics. It also influences furnace volume and the arrangement of
heating surfaces*
Fixed Carbon
Fixed carbon consists mainly of carbon but nay contain small
amounts of oxygen, nitrogen, sulfur and hydrogen not driven off with the
volatile matter. Essentially, it is the combustible residue renaining after
the volatile natter distills off. The hardness of the fixed carbon is an
indication of the caking properties of the coal which can be important in
the selection of fuel handling equipment.
Ash
Ash is an impurity that increases shipping and fuel handling costs.
It is that portion of the coal left over after conbustion. It deposits on
- 52 -
-------
c5350
furnace surfaces causing slagging and carries over into superheater, reheater,
economizer, and air heaters contributing to fouling problems. Portions
escape with the flue gases which must be collected in precipitators or dust
collectors. Part of the ash and slag drops to the bottom of the furnace
into the ash hopper, requiring complicated handling systems for removal. It
is important to know the amount and character of ash before coal is purchased
to avoid or to minimize some of the problems mentioned above. Ash also
tends to have an effect on the combustion process, resulting in increased
carbon carried to the ash pit or to the precipitator.
Sulfur
Sulfur is an undesirable element contained in raw coal in amounts
as high as 8 percent. The effects of burning sulfur result in many nuisance
side effects in the boiler not the least of which is the emission of sulfur
oxides to the atmosphere which, under present regulations, requires instal-
lation of expensive sulfur removal systems. Corrosion of furnace tubes,
stacks, and cold end sections of air heaters, has also been directly
attributable to sulfur. Reliable data on fuel sulfur content is important
in negotiating equitable fuel contracts.
Organic sulfur (combined with the coal substance), pyritic sulfur
(combined in the form of narcasite or with iron in the form of mineral
pyrite) and sulfate sulfur (calcium or iron sulfate) are the three forms of
sulfur contained in raw coal. Organic sulfur and the finely divided pyrites
are the forms of sulfur considered to be non-removable economically under
present day technology. Technology exists, however, for the removal of
pyritic sulfur by washing the coal. Sulfate sulfur is not an important
problem, being present in coal in minute amounts. Aside from its nuisance
value, the calorific heating value of sulfur is low, making it undesirable
from an efficiency standpoint.
Nitrogen
Until recently, when emphasis was placed on pollution control,
little attention was given to the nitrogen content of raw coal. Experience
has shown that NOX emissions occur from two sources (1) fixation of the
nitrogen in the combustion air (termed "thermal NOX") and (2) from nitrogen
in the fuel (termed "fuel NOX"). Thermal NOX emissions can be greatly
reduced as discussed in this guideline but control of "fuel NOX" formation
while affected by certain combustion modifications is not as easy to control.
Therefore, fuel nitrogen content is of importance to the control of NOX
emissions. Unfortunately, economic processes do not exist in today's
technology for the removal or reduction of fuel nitrogen. Nitrogen content
of coals generally used in coal fired power plants, as indicated in Table B-2,
ranges between 0.8 and 1.9 percent for bituminous, subbituninous, and
lignite fuels. Coal selection solely on the basis of nitrogen content does
not appear to be feasible.
Heating Value
The calorific, or heating value, of a fuel is of utmost importance
in the purchase of coal since BTU's or energy is what is actually being
- 53 -
-------
d5350
purchased. From Table B-2 it nay be seen that the higher ash and oxygen
contents of the lignites and some subbituninous coals adversely affect the
heating value of the coal. The adverse effects of high ash content have
already been mentioned and it should be recognized that high oxygen content
has an equal deleterious effect on heating value. The price of the coal
should reflect these undesirable qualities.
It should also be recognized that when a coal sanple is burned in
a bomb-type calorimeter filled with oxygen under pressure, the fuel's higher-
heating value is measured. The latent heat of water vapor contained in the
combustion products is lost in the stack gas since the water vapor in the
flue gas is not cooled below its dewpoint during normal boiler operation.
Thus the latent heat is not available for making steam and it is often
subtracted from the high-heating value to give the net or lower-heating value,
- 54 -
-------
c5350
APPENDIX C
FORMATION OF COMBUSTION-GENERATED POLLUTANTS
This appendix section has been designed to provide fundamental
background information on the formation of combustion-generated pollutants.
It emphasizes those pollutants most closely identified with the combustion
of pulverized coal. This section also provides some background on why
pollutants form and an indication of how to minimize pollutant fornation
through combustion modifications.
This section begins by discussion nechanisns of coal-fired combustion.
Discussed subsequently are the fornation routes for pollutant emissions
of:
Nitrogen oxides
Sulfur oxides
Particulates
Hydrocarbons
Carbon monoxide
Coal Combustion liechanisns
The combustion of coal has been considered to occur in three
regimes: totally detached flame, attached diffusion flane and char burn-out.
These will be discussed with reference to a pulverized coal particle fired
in suspension. The three cases are illustrated in Figure C-l.
The first regime occurs during periods of rapid volatilization of
the coal particle and large relative velocities between the particle and the
surrounding gases. Combustion occurs in the wake trailing the particle
where the fuel and oxidizer have the opportunity to mix, at least partially,
before combustion starts. This regime will be found in coal firing only
when the rate of heating is high causing the particle to release large
quantities of. volatiles explosively, or when the velocity of volatiles
escaping from the particle is high. The volatilized species will be com-
busted far from the fast travelling particle.
The second regime, the attached diffusion flame, occurs when the
rate of vaporization is slow enough to allow a flane to attach itself to
the particle. Here most of the combustion occurs in a thin flame sheet
surrounding the particle and its wake. The reactions occurring in this
diffusion flame sheet are quite different from those occurring in the case
of the totally detached flame where some premixing of air and volatiles
occurs. In the case of the attached diffusion flame, there are hot, fuel-
rich regions where precombustion pyrolysis can occur. In such a flame front
the "fuel" can be considered to be the products of pyrolysis.
In the third regime, char burn-out, the volatile constituents
initially present in the coal particle are depleted and the remaining heavy
ends are pyrolyzed and convereted into a char. As the volatiles are deleted,
the flame front approaches the particle until oxygen molecules can attack
- 55 -
-------
A. Totally Detached Flame. Rapid mixing, rapid volatilization
Fuel-rich regions
regions for pyrolysis, cracking
HCN, N, pyrolysis compounds formed
Flame Front
Attached, Diffusion Flame. Pyrolysis occurs in fuel-rich region
char is being formed
C. Char Burn-Out. Surface combustion occurs when volatiles have been
driven out. CO formed as a result of surface reaction of carbon and
02- This CO burns externally to the solid particle in a diffusion
fl
ame.
Figure C-l. Three regimes of coal particle combustion.
- 56 -
-------
c5350
the char directly. In burn-out, carbon Is converted at the surface to CO
and nitrogen probably winds up forming NO. In the flame surrounding the
char particle, the CO will burn to completion and the NO may possibly be
further converted to other products.
Nitrogen Oxides
Flue gas from fossil fuel fired utility boilers usually contains
concentrations of nitric oxide (NO) and nitrogen dioxide (N02> of 1500 ppm
or less. NO and N02 are collectively referred to as NOX. At high temperatures
equilibrium favors the formation of NO, so essentially all of the nitrogen
oxides in the firebox of a combustion unit are in that form. Lower temperatures
favor the oxidation of NO to N02, but the residence time of flue gas in the
unit at low temperature is too short for this reaction to occur to any great
extent. So most of the nitrogen oxides (>90%) in the flue gas are in the form
of NO. However, after the NO is emitted to the atmosphere there will be a much
longer residence time in which the oxidation can occur. The rate and extent of
conversion of NO to NOX depend on the amount of dilution with ambient air,
temperature, sunlight, and the presence of other materials such as hydrocarbons
and oxidants with which it can interact.
NO is a colorless, odorless gas which is not generally regarded as
a health hazard at concentrations found in the atmosphere. Because NO is
converted to N0£ in the atmosphere, most exposure is, in fact, exposure to
N0o» Nitrogen dioxide is a yellow-brown colored gas with a pungent odor.
The effects of exposure to low levels of N0£ (e.g., <50 ppm) may take
several days to develop. At higher levels such as 60-150 ppm, exposure can
result in immediate respiratory system reaction such as nose and throat
irritation, coughing, etc. Chronic exposure to even low concentrations can
produce chronic respiratory tract irritation.
NOX which is formed in combustion processes arises from two
sources:
1. "Fixation" of atmospheric nitrogen, i.e., the reaction of N2
from the atmosphere with 02 under the intense conditions of
combustion. Because of the high temperatures involved, NOX
formed via this mechanism is termed "thermal NOX". The rate of
formation of thermal NOX is very sensitive to temperature.
2. Conversion of "fuel" nitrogen, (i.e., nitrogen atoms bound into
fuel molecules) to NOX. This is termed "fuel NOx". Fuel NC^
formation is not significantly temperature dependent.
The NOX formed from both sources is chemically identical and its origin
can be determined only through controlled experiments.
The relative importance of the two sources depends on fuel type and nitrogen
content as well as the type, size and operating conditions of the combustion
unit in which the fuel is burned. In some cases thermal NOX will constitute
most or all of the NOX emission. In other cases, fuel NOX will be responsible
for most or all of the NOX emission.
- 57 -
-------
d5350
The formation of thermal NOX is known to proceed via a free
radical mechanism. The mechanism is initiated by the rupturing of an oxygen
molecule, 02, into two oxygen radicals due to the high flame temperatures.
Thus, the variables governing the amount of NOX formed include:
• peak flame temperature (the higher it is, the greater the fraction of
oxygen molecules coverted to radicals).
• residence time at elevated temperatures (favors radical formation and
reaction).
• availability of excess oxygen (the more oxygen molecules present, the
more oxygen radicals can be formed).
On a practical basis, NOX formation tends to be high in combustion
units in which the flame zone temperature is high. Factors which influence
flame zone temperature include:
Size - flame temperatures in larger boilers are frequently higher
than in smaller boilers.
Type of firing - wall-fired boilers have a hotter, more intense
flame than tangentially fired boilers, where there are burners at
all four corners.
Refractory - covering water tubes in the firebox with belts of
refractory increases flame zone temperature.
Typically, 15 to 100% of the nitrogen contained in fuel is converted
to NOX, with the greater conversions occurring when the fuel nitrogen
level is low. The fuel nitrogen compounds responsible for fuel NOX are
considered to be derivatives of such components as pyridene and quinoline
which are volatilized in the high temperatures near the combustion zone, and
are subsequently burned. It is believed that a substantial fraction of the
NOX resulting from coal combustion is fuel NOX. Combustion modifications
affect formation of both thermal NOX and fuel NOX but are less effective on
the fuel NOX portion which is not as readily controllable.
As noted above, the primary variables governing the amount of
NOX formed are the availability of excess oxygen, peak flame temperature,
and the residence time available at elevated temperatures. NOX formation
can be controlled by modifying combustion conditions to minimize the values
of any or all of these variables. The most common approaches include staged
combustion low excess air. In extreme cases, load reduction may be legislated
but this normally is not viable economically.
Combustion modifications such as flue gas recirculation and staged
combustion have been used effectively on gas and oil fired utility boilers
and the latter has been demonstrated to be effective on coal-fired utility
boilers. Flue gas recirculation utilizes cooled flue gas recirculated into
the combustion air supply to reduce flame temperature. With staged firing,
combustion is initiated with less than stoichiometric air in the first
stage and is completed, possibly after some interstage cooling, with the
introduction of additional air in the second stage. This method both
reduces peak flame temperatures and the local availability of oxygen.
- 58 -
-------
e5350
In utility boilers staged combustion is often accomplished
by admitting 70-90% of stoichiometric air through the burners, with the
remaining air then brought in through auxiliary ports or through burners
which have been taken out of service. Staged conbustion has occasionally
been used in combination with flue gas recirculation to provide further
NO,, reductions than can be provided by either technique individually.
X
Operating at low excess air levels usually serves to minimize
NOX formation by reducing the availability of oxygen, whereas running at
reduced throughput and firing rates often reduces NOX enissions by lowering
peak conbustion temperatures. Of these two approaches, the former has the
advantage of improving combustion efficiency, while the latter may not be
practical from an operating standpoint.
The formation of fuel NOX is only slightly temperature dependent, so
that staged combustion and low excess air are the only combustion approaches
effective in its control. On the other hand, NOX formed by the thermal
fixation of nitrogen and oxygen in the combustion air is extremely temperature
dependent and can be controlled by any or all of the combustion modification
techniques. Other approaches to controlling thermal NOX which have been
investigated include lowering combustion air preheat temperatures or injecting
steam or water into the combustion zone. These techniques, however, suffer
from efficiency losses.
gulfur Oxides
When coal is combusted, most of the sulfur in the coal is converted to
sulfur oxides. The oxides formed are sulfur dioxide, SC>2, and sulfur trioxide,
503. Sulfur dioxide is a colorless, nonflammable gas which is generally
regarded as being highly irritating. It can be detected by an individual by
taste and smell in concentrations of 3-5 ppin. Sulfur trioxide is a vapor in
the hottest zones of most combustion equipment but can condense to a mist in
the cooler boiler regions or can condense after being emitted from a stack as a
blue-tinted plume. 803 is a strongly acidic pollutant which is responsible
for back end corrosion in boilers. The sum of S02 plus 803 is frequently
termed SOX. The total quantity of SOX emitted is generally a direct func-
tion of the sulfur content of the fuel.
Typically, in the hottest sections of a boiler about 1 to 3% of
SOX will be present in the fora of 803, although S(>2 may be oxidized
to 803 in the ambient air and downstream of the air preheater.
tf a fuel such as coal contains large quantities of ash, or if the
ash constituents are basic, some SOX will be adsorbed on the ash particles
or will react with the ash constituents to form sulfates. The quantity of
SOX which is absorbed by coal ash depends upon the nature of the coal, the
quantity and type of basic constituents in the coal and the firing method.
In one test in which bituminous coal was fired, sulfur retention was less
than 5%. In another test in which lignite was fired, sulfur retention
ranged from 10 to 40%. Lignite ash is basic due to the high concentration
of alkali and alkaline earth compounds it contains.
Combustion modification cannot reduce the total quantity of SOX
formed in coal firing. This is a function of the coal sulfur level and the
- 59 -
-------
e5350
nature of the coal which can change the quantity of SOX converted to ash.
However, the 503/802 ratio can be varied through combustion nodifications.
Thus, the factors effecting SOj fornation will be considered briefly.
Sulfur trioxide fornation can occur via either "homogeneous" or
"heterogeneous" pathways. The former is known to occur both in the flame
and in the high temperature reaction zone in the flue gas just beyond the
flarie. It is this formation route which is most directly influenced by
combustion modification and which will be discussed here in greater detail.
The latter nechanisn involves suspended fly ash and furnace deposits.
The relative importance of these depend on the furnace, fuel and operating
conditions.
v
The factors which influence homogeneous 803 formation are:
• excess air level
• flame temperature
• flame quench rate
Let us see why these factors are important.
High temperature 803 fornation occurs when free radical oxygen
atoms react with S(>2 converting it to 803. The free radical oxygen atoms
are formed because of the high temperature of the flame. Up to a point, the
greater the excess air level, the greater the number of free radical oxygen
atoms that can be formed and consequently the greater the quantity of 803
formed. Thus, 803 formation can be minimized by minimizing the level of
excess air used. This is also good for NOX control. However, increasing
excess air above a certain point (estimated to be about 30-40%) will reduce
flarie temperature and will thereby reduce the quantity of free radical oxygen
atoms available for reaction thus lowering the quantity of 803 formed.
However, this may result in higher levels of tKX- and will have a negative
effect on efficiency. 803 is kinetically more stable at low than at high
temperatures. Consequently, after 863 has formed at high temperatures, it
will begin to decompose if kept at that temperature because of kinetic consider-
ations. If it is cooled just after being formed, most will remain as SOo.
Thus, flame quench should be gradual to permit the 863 to decay as it will at
high temperatures.
Particulates
Particulate matter and smoke can be emitted when coal is combusted,
especially if the equipment is operated under maladjusted conditions. The
form of particulate matter emitted from stationary combustion sources
includes:
Smoke - submicrometer-size particles of carbon formed in the
vapor phase during combustion.
Cenospheres (or chars) - unburned carbonaceous residues of fuel
particles. Ilay be 5 to 50-micrometer particles.
Ash - non-cor.bustible residues of fuel. Kay be 5 to 10-micrometer
particles, or submicrometer if the ash vaporizes and recondenses.
- 60 -
-------
c5350
Size, composition and quantity are significant factors which must
be considered in evaluating the importance of particulates. Small particles
(< 1 micrometer) settle slowly in ambient air and reduce visibility more
per unit weight than larger particles. Of particular concern are particles
<15 micrometers in diameter which are termed inhalable particulates because
they can be breathed.into the lungs where they may produce adverse health
effects.
The particulate mass produced by coal-fired power plants is
significant. Electric power generation by utilities accounts for 13% of the
total man-made emissions of particles in the U.S. Coal combustion accounts
for 98% of particulate emissions from utilities. It is important to realize
that these emissions occur after treatment by emission control devices such
as electrostatic precipitators, cyclones, scrubbers and filters. Particulate
control devices have an overall particle collection efficiency ranging from
80 to 99%, the last figure being for boilers firing pulverized coal with
electrostatic precipitators, and the first for stoker fired boilers using
mechanical collectors. Particulate control devices are generally relatively
effective in the control of large particulates. The control devices are
substantially less efficient for the collection of respirable or inhalable
particulate. For example, efficiency of control devices in boilers firing
pulverized coal amounted to 75% in one survey. Significantly, in these
units, 35% of all particulates by mass were of the fine, respirable variety.
Before discussing the forms of particulate matter emitted, let us
review the processes which occur to a coal as it is combusted.
The coal particle is swept into the furnace where it receives heat
from the surroundings. Organic volatiles in the particle begin to evaporate.
The volatiles burn instantaneously and completely. At the same time, the
particle swells and forms a hollow, porous sphere. Once volatilization has
ended, oxygen has access to the sphere and the carbon in it begins to burn,
forming CX>2 (or CO which is oxidized to C02>. The residue consists of
inorganic ash and any uncombusted carbon.
Smoke is composed of submicrometer-size carbon particles formed in
the vapor phase which surround the coal particle during the first stage of
combustion. The mechanism of smoke formation is generally considered to
include cracking of volatilized molecules in the zone where there is insuffi-
cient oxygen, followed by polymerization of the cracked fragments to form
hydrogen deficient condensed ring structures. These collect into particles
on the order of 0.1 micrometer in size. Most of the smoke particles burn out in
the flame, imparting a characteristic yellow-orange color. Even small
amounts of smoke by weight in the flue gas are highly visible. If 0.1
to 0.2% of the fuel forms smoke, it will appear as a dense, black cloud on
leaving the stack. Mixing, combustion zone temperature, air/fuel ratio and
particle size are significant determinants of the quantity of smoke emitted.
Smaller particles and better mixing decrease the possibility of forming
fuel-rich pockets in which smoke forms. High excess air levels permit
the flame to be close to the particle minimizing the space in which cracking
can occur. High combustion zone temperature favors rapid volatilization so
that combustion is completed early and there is ample opportunity for smoke
to be consumed.
- 61 -
-------
b5350
Most of the factors which minimize smoke formation tend to favor
NOX formation since they involve short, intense flames with high, localized
temperatures and good oxygen distribution in the hot zone. Thus, the
problem of minimizing both smoke and NOX emissions from fuel combustion
is basically that of minimizing time-temperature history for NOX control
without increasing smoke emissions.
Cenospheres form during combustion when the volatile portion of a
coal particle vaporizes, leaving a residue of carbon and ash. The carbon
which remains tends to burn slowly. Thus, long residence times at high
temperature with sufficient excess air reduces the emission of cenospheres.
These additions are more likely to be met in a large boiler than in a small
one, and large boilers have been found to emit fewer cenospheres than small
boilers. Coal particle size is also a factor. The finer the particle, the
greater the resulting surface to volume ratio which favors diffusion of
oxygen to the surface so the cenosphere can burn out.
Ash remains after the carbonaceous matter has burned out of the
cenospheres. The quantity of ash formed is directly related to the ash
content of the coal. Certain elements present in coal are volatile at flame
temperature. These include sodium and vanadium. Thus, these ash constituents
may vaporize and recondense to form submicrometer particulates in the flue
gas.
Normally the cenospheres and ash residue account for a major part
of the weight of particulates emitted from coal firing. Thus they have a
major effect on emission factors and their elimination would cause a signi-
ficant reduction in the weight of particulate emitted. Yet, there is some
question about how effective this would be as a particulate control measure.
Because of their size, 10-50 micrometers, cenospheres settle out rapidly
near the point of emission. However, the smaller particulates especially
those below 1 micrometer probably constitute a much more important contri-
bution to the ambient air quality and require more stringent controls.
Carbon Monoxide
The formation of carbon monoxide in the combustion process is
generally associated with incomplete combustion. Carbon monoxide, CO, is a
colorless, odorless toxic gas. In well-controlled combustion processes, CO
emissions will generally be quite low. For example, pulverized coal combusted
in utility boilers produces CO levels of approximately 30 to 40 ppm.
Lowering the excess air level below a critical value will increase
CO emissions sharply, forming a "knee". Both the shape of the knee in the
CO curve and the excess air level at which it occurs generally vary from
boiler to boiler. These characteristics are a function of fuel type, combus-
tion unit type and burner characteristics. The sharp rise in CO level occurs
because of imperfect mixing of the fuel and air, a condition accentuated as
the ideal stoichiometric fuel/air ratio is approached. Thus, partially com-
busted portions of coal gases on coal char do not mix well with sufficient
air to complete their combustion before leaving the flame zone. In addition
quantities of fuel can bypass the normal combustion zone and mix with flue
gas which does not contain sufficient oxygen for complete combustion. Under
proper combustion conditions, virtually all CO formed is combusted to
- 62 -
-------
c5350
The use of CO as a combustion control parameter is increasing in
the U.S. It has been used for this purpose in Europe. To control combustion,
use is made of the knee. Holding CO emissions at a specified level such as
200 ppm or less should minimize excess air (low excess air firing), resulting
in increased boiler efficiency and lower emissions of NOX and 803.
CO emissions are generally controlled through proper combustion
operating practices and combustion equipment design. Unfortunately, some
combustion practices designed to reduce CO tend to raise NOX emissions. For
example, CO formation can be minimized by providing for rapid reaction
rates. This means assuring rapid contact through the mixing of air and fuel,
providing sufficient air for complete combustion and preheating fuel and
air, all of which tend to increase NOX. Clearly, this is a trade-off
situation in which CO and NOX emissions must be minimized.
A prime operating practice which minimizes CO formation is the
effective adjustment of the fuel/air ratio. The minimum amount of air which
produces acceptable CO and NOX levels is to be used. If air well over the
stoichiometric amount is used, the combustion temperature will be lowered,
efficiency will decrease, and increased levels of NOX will result.
Hydrocarbons
Hydrocarbons consist of uncombusted or partially combusted carbon-
containing vapors and gases of diverse molecular composition. The types of
hydrocarbons, HC, which can be formed may vary greatly, thereby precluding
broad generalizations regarding -their characteristics.
The presence of HC emissions are symptomatic of incomplete combustion.
Thus, hydrocarbons are produced due to poor mixing and the localized insuffi-
cient oxygen for combustion. The level of hydrocarbon emissions resulting
from the combustion of pulverized coal in power plants is low, ranging from
less than one ppm to ten ppm. Thus, when coal combustion is under good control,
HC emissions are generally not a problem.
The procedures to follow for reducing HC emissions involve increasing
the excess air level and/or improving air-fuel mixing. Generally, as excess
air is reduced, CO emissions increase before HC emissions increase. Therefore,
if the CO level is low, HC levels should also be low.
- 63 -
-------
c5350
APPENDIX D
METHODS OF MEASURING POLLUTANTS AND
OTHER COMBUSTION PRODUCT EMISSIONS
This appendix is a discussion of the fundamentals of pollutant
measurement, and is divided into two primary sections. These deal with
particulate emissions and gaseous emissions. This appendix is designed to
provide sufficient background information so that the reader can understand
more fully the significance of pollution measurements, and the underlying
principles. This information can be of value in dealing with air pollution
officials, outside contractor personnel, and vendors of monitoring equipment.
D.I Particulate Emissions
This section describes the procedure for obtaining a representative
sample of particulate emissions. Information is presented relating to particu-
late measurements including particulate mass, particulate size and opacity.
D.I.I Sampling Particulate Emissions
Obtaining representative samples is one of the most important
factors leading to valid emission measurements. In fact, it has been said
that more errors result from poor or incorrect sampling than from any other
part of the measurement process. To obtain valid emission samples requires
the identification of proper sampling locations. Often the best locations
can be hard to reach points which can be hazardous and high above ground.
The most critical sampling requirements are for the measurement of
particulates. Very specific requirements for particulates have been established
by EPA to assure that a representative sample is obtained. Bends in ducting
lead to the stratification of particulate in duct regions downstream of the
bend and a traverse will be required if a representative sample is to be
obtained. The requirements for gaseous measurements are only slightly less
rigorous, as gaseous stratification can also occur.
The selection of a sampling site and the number of sampling points
needed are based on attempts to get representative samples. Rules for accom-
plishing this are established in the EPA New Source Performance Standards,
NSPS (code of Federal Regulations, Title 40, Part 60, EPA, July 1, 1977 and
as amended). The EPA NSPS have been established for compliance purposes.
Because of their widespread use for compliance testing, the NStS methods are
often the standards of performance against which other methods are judged.
They are also used for non-compliance purposes. The sampling site should be
at least eight stack or duct diameters downstream and two diameters upstream
from any bend, expansion, contraction, valve, fitting or visible flame. For
rectangular ducts, the equivalent diameter can be calculated from the
expression: Equivalent diameter - 2 (length x width)/length + width).
After determining the sampling location(s), provision must be made to
- 64 -
-------
b5350
"traverse" the stack or duct. That is, the actual sampling must be performed
at a number of traverse points in the stack. These multiple samples are
necessary because of the extreme gradients of flow and concentration that
occur in some ducts and to a lesser extent in stacks.
When the 8 downstream and 2 upstream diameter criteria can-
not be met, EPA has specified a minimum number of sampling points for stack
diameters greater than 0.6 m and less than 0.6 m. The minimum number of
traverse points required is illustrated in Figure D-l. To use this figure,
it is necessary to first determine the distances from the chosen sampling
location to the nearest upstream and downstream disturbances. This distance
is then divided by the diameter or equivalent diameter to determine the
distance in terms of the number of duct diameters. Figure D-l is then used
to determine the minimum number of traverse points that corresponds (1) to
the number of duct diameters upstream and (2) to the number of diameters
downstream. If a different number of traverse points is required by distur-
bances upstream and downstream, the greater number is to be selected.
Samples are to be obtained from zones of equal areas. This
is easy to visualize for a rectangular duct. Such a duct is shown in cross
section in Figure D-2. Travese points would be located as the centroid of
each equal area as is shown in Figure D-2.
Obtaining samples from zones of equal areas may be less clear for
a circular stack because most people are not accustomed to determining areas
of circular cross sections. Figure D-3 illustrates the circular cross
section of a stack divided into 12 equal areas. Note that the division as
shown requires that samples be obtained from stack ports which are at right
angles. A chart for determining the location of traverse points on a
circular stack is shown as Table D-l. The chart presents the percent of
stack diameter from the inside wall to the traverse point. For example, if
the stack is 30 meters wide and 10 traverse points are to be obtained along
the stack diameter, the third point will be: 10 meters x 0.146 -1.46
meters. That is, 1.46 meters from the stack wall.
Once the traverse points have been established, and safe access to
the sampling location has been provided, velocity measurements are needed to
determine gas flow.
Stack-gas velocity is determined from a measurement of the velocity
pressure, made using a pitot tube. The velocity pressure is the difference
between the total pressure (measured against the gas flow) and the static
pressure (measured perpendicular to the gas flow).
The S-type pitot tube is specified by the EPA for measuring
velocity. The S-type pitot tube is designed for easy entry into small holes
in the stack wall, and because of its high relatively large openings does
not readily plug when in the presence of high concentrations of particulate
matter. However, it requires a separate calibration for the particular
velocity being measured, and thus does not directly read the velocity
pressure.
- 65 -
-------
50
40
o
OH
V
CO
(J
> 30
20
10
Duct Diameters Upstream From Flow Disturbance*(Distance A)
0.5 1.0 15 2.0
T
T
T
*From Point of any Type of
Disturbance (Bend, Expansion, Contraction, etc.)
1
1
2.5
^
T
A
i
I
—
J
1
p.
— -
t
4
Disturbance
Measurement
F-- Site
Disturbance
S
3456789
Duct Diameters Downstream From Flow Disturbance*(Distance B)
10
Figure D-l. Determination-of the Minimum Number of Traverse Points.
- 66 -
-------
o
o
— — —
0
1
1
1
-i-
1
1
T-
1
1
1
o
— — •
o
— -~"
o
1
1
1
t--
1
1
-t-
1
1
1
1
0
o
*_ .
o
1
1
1
-\
1
1
1
-1
1
1
1
0
— — — •
o
™ — — —
o
Example showing rectangular stack cross section divided into
12 equal areas with traverse points at centroid of each area
Figure D-2. Duct Cross Section.
Traverse
Point
1
2
3
4
5
6
Distance,
% of Diameter
4.4
14.7
29.5
70.5
85.3
95.6
Example showing circular stack cross section divided into 12 equal
areas, with location of traverse points, at centroid of each area.
Figure D-3. Circular Cross Section of Stack Divided into 12 Equal Parts.
- 67 -
-------
TABLE D-l
Location of Traverse Points in Circular Stacks
(Percent of stack diameter from inside wall to traverse point)
Traverse
point
number
on a
diameter 24 6 8 10 12 14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
14.6 6.7 4.4 3.3 2.5 2.1 1.
85.4 25.0 14.7 10.5 8.2 6.7 5.
75.0 29.5 19.4 14.6 11.8 9.
93.3 70.5 32.3 22.6 17.7 14.
85.3 67.7 34.2 25.0 20.
95.6 80.6 65.8 35.5 26.
89.5 77.4 64.5 36.
96.7 85.4 75.0 63.
91.8 82.3 73.
97.5 88.2 79.
93.3 85.
97.9 90.
94.
98.
8
7
9
6
1
9
6
4
1
9
4
1
3
2
16
1
4
8
12
16
22
28
37
62
71
78
83
87
91
95
98
.6
.9
.5
.5
.9
.0
.3
.5
.5
.7
.0
.1
.5
.5
.1
.4
18
%
1.
4.
7.
10.
14.
18.
23.
29.
38.
61.
70.
76.
81.
85.
89.
92.
95.
98.
4
4
5
9
6
8
6
6
2
8
4
4
2
4
1
5
6
6
20
1
3
6
9
12
16
20
25
30
38
61
69
75
79
83
87
90
92
96
98
.3
.9
.7
.7
.9
.5
.4
.0
.6
.8
.2
.4
.0
.6
.5
.1
.3
.3
.1
.7
22
1.
3.
6.
8.
11.
14.
18.
21.
26.
31.
39.
60.
68.
73.
78.
82.
85.
88.
91.
94.
96.
98.
1
5
0
7
6
6
0
8
1
5
3
7
5
9
2
0
4
4
3
0
5
9
24
1.1
3.2
5.5
7.9
10.5
13.2
16.1
19.4
23.0
27.2
32.3
39.8
60.2
67.7
72.8
77.0
80.6
83.9
86.8
89.5
92.1
94.5
96.8
98.9
Source: Code of Federal Regulations, Title 40, Part 60, EPA,
July 1, 1977.
- 68 -
-------
c5350
Isokinetic sampling is required if a representative sample
of particulate is to be obtained. To be isokinetic, the velocity of the
stack gas stream entering the probe nozzle must be the same as the velocity
of the stream passing the nozzle. If the sampling velocity is too high
(super-isokinetic sampling), there will be a disproportionately large
concentration of small particles collected (because the inertia of the
larger particles prevents them from following the stream lines into the
nozzle). Alternatively, in sub-isokinetic sampling, where the sampling
velocity is below that of the flowing gas stream, the gas samples would
contain a higher-than-actual concentration of large particles (because
heavier aerosol particles will enter the nozzle, but light particles will be
diverted). Thus, even though isokinetic sampling is preferable, if the
stack gas stream velocity fluctuates, it is better to sample slightly
super-isokinetic than sub-isokinetic because of an increase of small particles
has less effect on mass measurements than an increase of large particles.
Isokinetic, super-isokinetic and sub-isokinetic sampling are schematically
illustrated in Figure D-4.
It has been found that inertia effects become more significant
when particle size exceeds about 3 micrometers in diameter. Therefore, if a
reasonable proportion of the particles exceed this size, isokinetic sampling
is necessary. Because of the requirement for isokinetic sampling, the
sample volume extracted from each equal area zone will be proportional to
the velocity, assuming the velocity to be constant.
D.I.2 Particulate Measurements
This section discusses the types of particulate measurements which
generally can be undertaken at power generation facilities. Certain parti-
culate-related measurements at power plants are undertaken often and routinely,
others less frequently and still others only on rare occasions, if at
all-
One measurement which is made at many utility generating stations
On a routine basis is opacity. Monitors for accomplishing this measurement
a.re often permanently installed and operate continuously and on a real time
basis*
The measurement of particulate mass is generally made less fre-
quently than that of opacity. The sample for particulate mass measurements
is obtained on a grab sample basis. Sample collection can require up to
several hours. Particulate mass measurements are frequently obtained for
ompliance or to evaluate precipitator efficiency. These measurements may
l,e made by contractors who could require some form of governmental certification.
Measurements of particulate size are even less routine. There has,
however, been increasing emphasis on the importance of very fine particle
emissions «2.5 micrometers) into the atmosphere. The inhalable particulate
range, those less than 15 micrometers, is believed to have the greatest impact
On health. The measurement of inhalable particle emissions and particle size
distribution may some day be as commonplace as particulate mass measurements.
- 69 -
-------
Gas Stream
Gas Stream
Gas Stream
Isoki.net ic
Super-Isokinetic
V % V
V > V
Sub-Isokinetic
V < V
where V, = velocity in probe
V- = velocity in duct or stack
Figure D-4* Sampling Velocity and Potential Sampling Errors.
- 70 -
-------
b5350
The following sections discuss:
1. Opacity
2. Partlculate Mass
3. Partlculate Size
D.I.3 Opacity
The measurement of opacity is perhaps the most widespread of all
continuous source monitoring measurements made in the U.S. Opacity can be
defined as the ratio of light attenuated by particulate emissions relative
to the incident light. Monitors to accomplish this measurement are known as
opacity meters or monitors, transmissometers or smoke meters.
The most commonly used opacity monitors are based on a system
which measures the decrease in light transmission caused by particulates in
a stack or duct. In a typical instrument the light source will be on one
side of the duct and the photoelectric detector on the other. The measurement
obtained is a function of the particulate concentration, and the wavelength
of the radiation used, with appropriate compensation made for temperature.
The width of the stack or duct is a permanent value which is used to calibrate
the instrument.
It is necessary to measure the opacity of emissions in the visible
range of the radiation spectrum if there is to be a meaningful correlation
with the opacity value seen by an observer. Even if visible light is used,
certain light attenuation instruments are often found to be biased toward
large size particles and others toward small ones.
In order for an opacity monitor to remain in good operating
condition, it is necessary to reduce to a minimum the effects of temperature
and vibration. As a consequence, this type of instrumentation should be
constructed ruggedly, with provision to maintain the specified optical
alignment and clean optical surfaces. The latter is generally accomplished
by blowing streams of air over the light source and detector. Typically, a
split beam optical system is utilized. The source radiation is split into
reference and measurement beams which are detected and compared. This
referencing technique compensates for temperature effects in the stack
and aging of the radiation source.
l
Opacity monitors are designed to automate the measurements
of a trained observer. Such an Observer will have been trained to identify
the opacity of smoke plumes by observation at an EPA certified smoke school.
Their training consists of observing controlled smoke plumes until they can
identify the opacity settings at which the smoke generator was being operated.
Opacity read by the trained observer method is related to many
factors, including meteorological conditions and position of the observer
relative to the plume and the sun, as well as the makeup and concentration
of particulates in the plume. Plume composition, particle size distribution
and concentration, and stack diameter directly affect the optical character-
istics of plumes and therefore what will be seen by observers. Meteorological
- 71 -
-------
c5350
conditions and plume dispersion play a particularly important role in
plumes containing condensable emissions. It is therefore not surprising
that a reading obtained by an in-stack opacity meter can differ from that
obtained by an observer.
One advantage of using the opacity meter rather than the services
of an observer is that a correctly aligned and properly functioning instrument
produces a more consistent, more objective measurement which is easier to
define in terms of physical concepts. Opacity monitors can also be used on
a continual basis whereas observers can only function during the day under
good atmospheric conditions.
D.I.4 Particulate Mass
• EPA Method 5
The EPA Method 5 train has been specified by the U.S. Environmental
Protection Agency as the method to be used for particulate mass compliance
testing of stationary sources. It is a grab sample technique in which
particulate matter is withdrawn isokinetically from a stack or duct and
trapped in an out of stack particulate filter. The method 5 stack sampler
is shown in schematic in Figure D-5 and a typical version is illustrated in
Figure D-6. Versions of the EPA Method 5 sampling train are commercially
available from a number of manufacturers, the most common systems being
operated at 1 cfm. A large volume system (5 cfm) is available for sources
having lower particulate concentrations. The train and the procedure for
its use are described in detail in Code of Federal Regulations, Title 40,
Part 60, EPA, July 1, 1977 and as amended. The train will be discussed here
briefly.
The stack gas is extracted isokinetically through a nozzle
and is pulled through a heated probe and heated collection box. The box
contains a cyclone to separate out large particles (>7 mm) and a filter with
a collection efficiency greater than 99.5+%. The gas is then cooled and
dried in a series of chilled impingers. Isokinetic flow is achieved using a
vacuum pump and valve. The appropriate flow is determined from the flow
measurement using a pitot tube by the use of a nomograph or calculator. The
dry gas meter registers the gas flow.
In obtaining the particulate mass measurement, the filtering
medium is weighed before and after sampling. To that weight difference is
added the cyclone catch and particulate removed by washing from the probe
and sampling train, after drying. This represents the total mass. By
knowing the volume sampled, emissions expressed as mass per volume of stack
gas can be calculated.
As noted above, EPA Method 5 is the method to be used for all
compliance testing. Because of its widespread use, it is the method with
the greatest amount of resource data available and with the greatest number
of experienced practitioners of the art.
- 72 -
-------
Temperature
Sensor
Heated Area /
/ Absolute Filter Holder
Temperature
Sensor
Probe
Reverse-Type
Pitot Tube
Air-Tight
Pump
Check
Valve
Impingers
-Pass Valve
Vacuum
Line
Dry Test Meter
Figure D-5. EPA Method 5 Sampling Train Schematic,
-------
Sample
Tube
Sample Box
Meter Box
Electrical Cord
Power
Cord
Check Valve
Pitot arid
Sampling lubes
Figure D-6. EPA Method 5 Sampling Train Illustration.
-------
b5350
• Automtated Particle Mass
Over the years attempts have been made to obtain particle mass
measurements on a continuous basis. Although some have been sold commercially,
most are experimental in nature. In certain cases, the readings of opacity
meters can be correlated on a case by case basis with mass emissions using a
reference technique such as Method 5 as a calibration tool. One automated
measurement method which has been commercially available uses beta radiation
attenuation. This system uses extractive sampling. It has been found that
probe losses are so significant that the data cannot be considered valid.
Until an improved transport system or an in situ particulate mass monitor is
developed, manual methods will continue to be used.
D.I.5 Particulate Size
The measurement of particulate size is not undertaken routinely
at power plants. It is, however, performed with sufficient frequency and
for a sufficient number of purposes (such as to evaluate the efficiency of
particulate removal equipment) to warrant its inclusion here. The most
widely used device utilized to obtain particulate size distribution measure-
ments is the cascade impactor. This device can be used in the stack or
externally in association with an extractive sampling system. In-stack use
of a cascade impactor is preferable as it eliminates sample losses in the
probe which could influence the particulate size distribution obtained.
Use of the in-stack system is generally more difficult than using an external
system. Used externally, the cascade impactor is placed in the heated
sample box associated with Method 5 equipment. Typically a tap is taken off
at the primary Method 5 stream.
The cascade impactor permits simultaneous collection and sizing of
the particulate emissions. Cascade impactors such as the Andersen and the
Brink collectors are multi-stage devices depending upon particle momentum
and aerodynamic drag to achieve separation. At each stage, successively
smaller particles are separated from the gas stream and are deposited on
collection plates. A final filter removes those particles passing through
the impaction stages, with the extent of recovery being limited by the
collection efficiency of this filter. The particle size distribution is
obtained by measuring the mass of particulate collected by each stage.
Total particulate mass is the sum of the collected mass of each stage plus
the mass collected by the final filter.
Problems associated with the use of cascade impactors include:
• wall losses (some particles may adhere to impactor walls rather
than to collector stages).
• reentrainment (not all the particles striking a collector stage stick
to it).
• weighing accuracy (the mass of the particles accumulated on certain
stages may be so small as to be almost immeasurable unless an extremely
sensitive balance is used).
- 75 -
-------
C5350
• isokinetic sampling (the design of a cascade impactor requires that the
flow cannot vary during the run. If the stack gas flow varies, the
sample rate may not be variable unless only a slip stream is being used
to supply stack gas to the impactor).
Cascade impactors are generally used to divide particulates
into a number of fractions. Thus, 8, 10 or more stages are common with
impactors. When fewer cuts are desired, appropriately sized cyclones plus a
final filter in a Method 5-type heated sample box can be used to get a size
distribution. Another system using either 3 or 5 cyclones in situ has also
been tested.
Advanced instrument systems for obtaining particle size distribution
have been under development for a number of years. None, however, ate in
widespread use.
D.2 Gaseous Emissions
This section describes considerations important in the sampling of
gaseous emissions. Also included is information on frequently used reference
methods of gas analysis as well as on instrumental methods of gas analysis.
D.2.1 Gaseous Emissions Sampling
Sampling for gaseous emissions involves the same type of procedures
as in particulate sampling. Again, it is very important to obtain a repre-
sentative sample. Typically, in a power plant, most routine gaseous measure-
ments will be obtained using permanently installed continuous monitoring
instruments. These may be solid state systems or wet chemical. Alternatively,
grab samples may be obtained. These are usually analyzed using wet chemical
methodology. Within the last few years, in situ, i.e., in-stack, monitors
have been developed which do not require sampling systems. In situ monitors
are based on optical principles and sight completely across a duct or stack
thereby eliminating the need for multipoint meaurements. In situ monitors
of this type are available for some of the most common gaseous pollutants
and also opacity measurements. To differentiate between in situ and the
techniques in which a sample is withdrawn, the latter are termed extractive.
In addition to locating an appropriate sampling point, extractive sample
consists of the following operational elements.
• Sample extraction
• Sample transport
• Sample conditioning
• Sample analysis
As in particulate sampling, extractive sampling for gases requires
introducing a sampling probe into a stack source and withdrawing a sample.
Analysis may be performed using wet chemical methods or continuous monitoring
instrumentation. In the case of wet chemical analysis, the sample is drawn
through a sampling train into a collection medium. The collection medium
may be in the form of filters or a set of bubblers which are then taken to
- 76 -
-------
b5350
the laboratory for analysis. Continuous monitoring is much more efficient,
productive, and less time consuming than laborious grab sampling and wet
chemical analytical method, but requires a greater capital investment.
Regardless of the analytical method employed, the fact remains that after a
sample has been extracted from the source it requires sample transport and
sample treatment to ensure that the sample presented for analysis is compat-
ible with the analytical procedure that follows. Also, the method chosen
may depend on the nature of the tests. If measurements are taken for
compliance, it will be necessary to use required reference techniques.
Several potential sources of error can exist in the sampling
system prior to the instrument. Sample integrity can be destroyed by:
• Chemical reaction with surface materials
• Chemisorption on particulate matter
• Water condensation in the sampling line
• Leaks in the sampling line
Extractive source monitoring systems either condition the sample
and follow with an analyzer for measurement at source level concentrations
or condition the sample by a dilution network coupled with an ambient air
level instrument. Conventional source level sampling systems employ a probe
filter for particulate removal and a drying device (usually a refrigerated
dryer) before the measurement is made. Permeation drying tubes, which
operate on the permeation distillation principle, have also become popular
for moisture removal.
Recently a new concept has been introduced in source level extrac-
tive monitoring which is termed reflux filtration. In this system the
sample from the stack is drawn through the clean reflux sample stream where
dirt particles and aerosols (acid mist) are blown back by the action of high
velocity gas molecules. After the sample is scrubbed, it passes through a
guard filter and then is pressurized by a pump. A small portion of the
sample is sent to the dryer but most of the sample is sent back as the clean
reflux sample stream. After scrubbing the incoming sample, the reflux
stream, along with dirt particles, discharges into the process stream.
Dilution techniques can offer an advantage in sample conditioning
by eliminating heated sample lines and water vapor removal system, if the
stack gas sample can be quantitatively diluted as close to the source as
possible. Dilution systems also offer the advantage that both ambient air
and source monitoring could be accomplished by the same instrument.
If stratification of gaseous pollutant species exists, more than
one sampling inlet may be required to approach the equal areas concept noted
for particulates. Indeed, it has been found experimentally that stratification
does exist in the flue gas ducting of power plants and the single point
sampling is inappropriate for obtaining representative gas samples. A
typical SC>2 concentration stratification contour measured for a utility
boiler is shown in Figure D-7. To reduce the significance of gaseous
stratification multipoint sampling can be used to sample from the inner 50%
of the duct. Results of in-stack tests, as opposed to in-duct tests,
- 77 -
-------
.450 m
1.960 m h-
1.470 m h
0.979 m |-
0.489 m r~
0.000 m
0.000 m 2.500 m
5.000 m 7.500 m
Duct Width
10.000 m 12.500 m
Figure D-7. SO- Profile.
- 78 -
-------
d5350
indicated that stack conditions are extremely uniform and should be the
preferred extractive sampling location, provided that practical access to
such a sampling location is available. One drawback to stack sampling is that
air leakage into the air preheater can result in dilution of stack emissions.
Sampling between the economizer and air preheater will eliminate this problem.
Frequently Used Preference
Methods for Gas Analysis
The EPA new source performance standards specify methods to be
used in measuring certain pollutants. Some of these methods are wet chemical
and others are not. Wet chemical methods are generally not used for contin-
uous or routine monitoring, because instrumental methods generally can
accomplish the measurement far more efficiently. However, in certain cases
no satisfactory instrumental method has been developed and wet chemical
methods continue to find utility. This section describes briefly the
reference methods for 803/302, NOX, and CO. The first two are wet chemical
and the last is instrumental. The 803 measurement is one for which no
instrumental method is available and so it is used perhaps more frequently
than any other wet chemical source monitoring technique. The descriptions
which follow are not meant to be complete but they provide a feel for
general requirements. Additional details on the methods described can be
found in the Code of Federal Regulations, Title 40, Part 60, EPA, July 1,
1977 (and as amended).
• S02/S03 Reference Technique (EPA Method 8)
In this method, a gas sample is extracted at temperatures above
the acid mist dew point. As a consequence, 863 will be present as a
vapor, not a particulate. The flue gas is extracted from the exhaust stream
through a heated probe. The sample is then passed through two absorbing
solutions in a total of three impinger bottles. The first contains an 80%
isopropyl alcohol solution to absorb the sulfur trioxide. Two subsequent
impingers contain a hydrogen peroxide solution to absorb the S02« A filter
located between the isopropanol impinger and the first peroxide impinger
traps any 803 mist which is not absorbed by the isopropanol. The quantities
absorbed are determined by chemical titration using barium perchlorate as
the titrant and thorin as the indicator. The concentrations of the individual
sulfur oxides can then be used to calculate the 503/802 ratio.
A high level of care must be exercised if the determination is to
be performed properly. After sampling, the train must be purged with air
for at least 30 rainutes in order to release the S02 which might have been
retained by the isopropanol.
• NOX Reference Technique (EPA Method 7)
The gas sample is collected in an evacuated flask containing dilute
sulfuric acid-hydrogen peroxide absorbing solution. The nitrogen oxides,
except for nitrous oxide, are oxidized to nitric acid by the hydrogen
peroxide. After careful destruction of the peroxide with heat, the nitric
acid produced is measured of the peroxide with heat, the nitric acid produced
is measured colorimetrically as nitrophenol-disulfonic acid. The laboratory
procedure is lengthy, requires a number of steps, and analysis must be
performed by a qualified laboratory technician. Errors may arise due to
- 79 -
-------
c5350
improper procedures and deterioration of reagents. This method is suitable
for NOX concentrations between 15 and 1500 ppm by volume and has a sensitivity
of about 1.5 ppm.
• CO Reference Technique (EPA Method 10)
An integrated or grab sample is extracted and analyzed for CO
content using a nondispersive infrared analyzer. Any substance having a
strong absorption of infrared energy will interfere to some extent. Dis-
crimination ratios for 1^0 and C02 are 3.5% H20 per 7 ppm CO and 10% C02
per 10 ppm CO, respectively, for devices measuring in the 1500 to 3000
ppm range. For devices measuring in the 0 to 100 ppm range, interference
ratios can be as high as 3.5% 1^0 per 25 ppm CO and 10% C02 per 50 ppm CO.
D.2.3 Instrumental Methods of Gas Analysis
A large number of instrumental methods have been utilized to
monitor gaseous pollutants. The earliest instruments available were simply
automated versions of wet chemical procedures. Colorimetric and electrochem-
ical techniques predominated. Over the years, however, there has been a
pronounced movement towards all solid state instrumentation, which frequently
is based on optical techniques. Table D-2 provides an indication of the
large number of methods which have been used in pollution monitoring.
The outline of instruments used for monitoring gaseous pollutants
on a source basis presented here is divided into two sections: instruments
which require an extracted sample, and those which measure across a stack or
duct without extracting a sample, the in situ instruments.
D.2.3.1 Extractive Instrumentation
Most continuous source emission measurement methods utilize
extractive sampling in which a gas sample is withdrawn from a stack or duct.
With most instrument methods, the sample is sent through sample lines to a
sample conditioning system and from there to the measurement instrument.
The sample conditioning system prepares the gas for analysis, making it
possible for the analyzer to perform effectively and minimize interferences.
The sample conditioning required varies from method to method. It generally
involves filtration to remove particulates and frequently includes a water
removal system such as refrigeration coils and/or permeation driers.
Certain instruments may require only heated sampling lines to prevent the
condensation of water. To withstand the corrosive effects of flue gases,
the complete sampling system must be constructed of 316 stainless steel,
Teflon or other materials not easily affected by stack gases. Extractive
instrumentation is potentially cheaper than other systems such as in situ
in that more than one analyzer can monitor more than one emission source on
a sequential basis. Extractive instrumentation may also be more accessible
for servicing than an in situ monitor which could be located high on a
stack.
- 80 -
-------
d5350
TABLE D-2
TECHNIQUES FOR ANALYZING GASES AND VAPORS IN FLUE GASES
NOV S02 C0_ H£
Wet Chemical X
Colorimetric X X
Color titration X
Electrochemical
Coulometric transducer X X
Radiation Attenuation
Nondispersive infrared XX XX
absorption
Ultraviolet absorption XX X
Photometric
Chemiluminescence X
Flame photometer X
Pulsed fluorescense X
Flame lonization X
Gas Chromatography X XX
- 81 -
-------
d5350
The following paragraphs describe the nost widely used extractive
instrumentation for monitoring the criteria pollutants of NOX, SCb, CO, and
hydrocarbons. A brief description of the operating principles of these instru-
ments for each type of pollutant noted is given here for the benefit of the
reader. The nethods used include:
• NOx
- TTHemiluninescent
- UV absorption
- Electrochemical
- NDIR
• S02
- TJV absorption
- NDIR
- Pulsed UV fluorescence
• CO.
- 11DIR
• HC
- FID
- NDIR
D.2.3.1.1 NOx
- Chemilutainescent
Instruments desisned using the chemiluninescent principle utilize
the light given off from the reaction between NO and ozone, Oj. A photo-
nultiplier tube detects the light emitted by this reaction. Because of the
nature of the reaction and filters that screen out illumination resulting
from other possible reactions, the chemilurainescent technique is largely
interference free. The nethod is also highly sensitive with models bein^
available to measure ambient as well as source NOX. The technique does,
however, require extensive sample preparation including removal of particulates
and much water vapor. Also many instruments operate at negative pressures
to permit the reaction to 30 more smoothly. Although this can add to
mechanical complexity, most of the commercially available cheniluminescent
instrumentation is reported to work quite well.
As noted above, the reaction is between ozone and HO. To monitor
total NOX, i.e., to get a value for 1102, the NOo must first be reduced
to NO. This is typically accomplished using converters constructed of
stainless steel, molybdenum, carbon or catalysts such as gold and tantalum.
Thus, values can be obtained for 110, 1K>2, or total NOX.
- UV Absorption
This instrumental method uses the absorption of ultraviolet
radiation by N02 molecules over a heated, several foot long path length
optical cell. W has little absorbence in the visible and UV range and
- 82 -
-------
d5350
consequently it must be converted to NC>2 to be monitored. This is accom-
plished by reacting NO with oxygen at a pressure of 5 atmospheres. The use
of high pressure oxygen may be undesirable in certain applications and
environments but few problems have been reported. UV absorption instru-
ments utilize a heated sample line and an "all hot" system to prevent
condensation of water and SOX. This analyzer requires no further condi-
tioning because particulates, water vapor and other stack gas constituents
do not interfer because of a split beam arrangement. Some UV absorption
instruments for NOX also have the capability for monitoring SC>2 as well
by a change in wavelength. This is generally done on a sequential basis
with one pollutant monitored for a given period, then another.
- Electrochemical
An electrochemical transducer has found widespread use as a NOX
analyzer. Typically a gas sample is passed through a membrane and into an
electrochemical cell. A signal is generated within the cell which is
proportional to the NOX concentration in the gas sample. This signal is
amplified and displayed on a meter. Some versions of electrochemical NOX
analyzers can also be used to determine SC>2, and generally these instruments
are equipped with plug-in interchangeable detectors, one for each gas.
These detectors can be changed as the measurement need requires.
Electrochemical instruments are inexpensive, rugged and simple to
use. They can give reasonably reliable readings when properly calibrated.
There are, however, a number of potential interferences and the electrochemical
cells require frequent replacement.
- ND1R
Non-dispersive techniques (IR or UV) have been used for monitoring many
gases throughout the years. Typically, energy from a source is split into two
beams, one of which passes through a cell containing the gas to be monitored,
and the other passes through zero reference cell without this gas. The
intensity of the radiation passing through these two cells is compared yielding
a concentration measurement. Non-dispersive instrumentation has been used
extensively in the past to measure NO, although newer instrumental methods are
more selective, interference-free and require less maintenance. The maintenance
of NDIR systems is genearlly considered to be high relative to other techniques
and the use of NDIR instrumentaiton to measure NO has diminished. In addition,
water and C02 are known to interfere. Thus, NDIR requires extensive sample
conditioning such as the removal of water vapor and particulate.
D.2.3.1.2 S02
- UV Absorption
The UV absorption method for S02 is very similar to that used for
NOX. As noted under NOX monitors, some S02/N02 monitors are convertible.
The instrument can alternately monitor S02/N02 by automatically changing
the optical filter to select the appropriate light wavelength for each measure-
ment. The UV absorption instrument for S02 has the following characteristics:
- 83 -
-------
d5350
• all hot sampling and measurement system
• long path length
• minimal interference
- NDIR
Non-dispersive infrared spectroscopy was one of the original methods
used to monitor for S02« Non-dispersive techniques (IR or UV) have been used
for monitoring many gases throughout the years. Typically, energy from a
source is split into two beams, one which passes through a cell containing the
gas to be monitored, and the other passes through a zero reference cell without
this gas. The intensity of the radiation passing through these two cells is
compared yielding a concentration measurement of 862• Non-dispersive instru-
mentation can no longer be recommended for most applications because new
instrumental methods are more selective, interference-free and require less
troublesome maintenance. The maintenance of NDIR systems is generally considered
to be high relative to other techniques. In addition, water and C02 are known
to interfere. Thus, NDIR requires extensive sample conditioning such as the
removal of water vapor and particulate. Water removal via a condensation
system is especially tricky because of the solubility of S(>2 in water. NDIR
is no longer recommended for SC>2 measurements but it has been used extensively.
- Pulsed UV Fluorescence
The commercial availability of pulsed fluorescence instrumentation
for source monitoring is relatively recent. It is, however, a highly
regarded method for the measurement of 862. A sample is drawn into a
chamber where it is irradiated with a brief pulse of ultraviolet light.
This causes 802 molecules to emit characteristic radiation which is
directly proportional to their concentration. The judicious choice of
wavelength for excitation and emission radiation filtering is designed to
minimize interference from other potential interferents such as water vapor,
oxygen and nitrogen. The pulsed fluorescence technique does require some
sample conditioning, specifically the removal of particulate matter and
sufficient water vapor to prevent condensation.
D.2.3.1.3 CO
- NDIR
Non-dispersive infrared is a widely used method for monitoring CO
emissions. It is, in fact, specified in the EPA NSPS as Method 10. The
NDIR system for CO works exactly in the same way as for S02« Here, too,
interference can be a big problem. Any substance having a strong absorption
of infrared energy will interfere to some extent. Discrimination ratios for
H20 and C02 are 3.5% H20 per 7 ppm CO and 10% C02 per 10 ppm CO, respectively,
for devices measuring in the 1500 to 3000 ppm range. For measuring in the 0
to 100 ppm range, interference ratios can be as high as 3.5% 1^0 per 25
ppm CO and 10% C02 per 50 ppm CO. Interference from water vapor can be
minimized by reducing the humidity, by using refrigeration (water solubility
is a problem here, too, but it is less than with 802), drying agents, or by
the use of filters.
- 84 -
-------
d5350
D.2.3.1.A Hydrocarbons
- Flame lonization Detector
The flame iontzation detector, FID, is the most widely used analytical
method for total hydrocarbons. It is continuous and real time. Typically
with this system an emission gas sample is introduced into a hydrogen flame.
The combustion of hydrocarbons, HC, produces ionization which is proportional
to the number of carbon atoms present. A current collector in the vicinity
of the flame collects the ions generated and the current which results is
proportional to the number of carbon atoms in the gas stream. The presence
of nitrogen, halogen or oxygen atoms attached to HC molecules lowers response
somewhat. Other carbon-containing species such as CO and C02 do not inter-
fere, and water vapor interference is minimal. A number of different types
of HC can be present but the standard FID cannot distinguish among them.
Differentiation among HC is possible using a gas chromatograph, GC, to
initially separate HC before they are introduced into the FID for monitoring.
Use of a GC will reduce the FID instrument's real time capabilities but
should permit a true reading of individual hydrocarbons.
- NDIR
NDIR was one of the earliest techniques used for hydrocarbon
monitoring. In monitoring HC, this technique is subject to the interferences
of water vapor and C02 noted earlier for other NDIR applications. Thus, an
extensive sample conditioning system is required. NDIR also responds
differently to different categories of hydrocarbons, being sensitive to
paraffins and only about 5% as sensitive to aromatics.
D.2.3.2 In situ Instrumentation
In situ monitoring of pollutants is a relatively recent development.
A typical stack or duct-mounted system consists of a light source, detector-
analyzer and a mounting pipe. In situ monitor is in fact an absorption
spectrometer in which the cell path length is the width of the stack or
duct. The light source sends a polychromatic radiation beam through the gas
to be measured. The detector-analyzer receives the beam and separates
it into the wavelength(s) to be measured. The beam intensity at the measure-
ment wavelength is ratioed with a nearby nonabsorbing wavelength and the
resulting electronic signals are translated into pollutant mass. Knowing
the gas stream temperature and pressure, the total volume can be determined
and thus the emission concentration calculated.
The reading given by an in situ system is rapid in that the sample
is measured right in the stack or duct and need not be moved to an analyzer.
It is potentially unaffected by stratification in that it obtains an "optical
traverse" across the gas stream being measured. It is also unaffected by
the presence of particulates because of the ratioing procedure and of course
requires no sample conditioning. For a single installation it is potentially
cheaper than an extractive system in that it requires no sample conditioning
system and the ability to measure more than one species can be added in a
modular manner.
- 85 -
-------
c5350
Disadvantages most often noted are keeping the system optically
aligned, properly calibrated, and temperature compensated. A manufacturer
of this type of instrumentation has undertaken extensive studies to compare
the results obtained with the wet chemical PDS reference method. In all
tests shown, agreement between the methods was very good.*
In situ instrumentation is commercially available or has been
developed for a number of pollutants including NO, S02, CO, C02, HC, H2S, N02
and NH3.
*"Analytical Methods Applied to Air Pollution Measurements" by R. K. Stevens
and W. F. Herget, published by Ann Arbor Science Publishers, Inc., Chapter
11, by H. C. Lord "Adsorption Spectorscopy Applied to Stationary Source
Emissions Monitoring."
- 86 -
-------
d5350
APPENDIX E
BOILER EFFICIENCY
Experience has shown that carbon content on particulate tends to
increase under low NOX operating conditions, especially in front wall
fired and on some horizontally opposed fired boilers. While this does not
necessarily happen in all cases, it is obvious that when carbon loss does
increase the effect is to lower boiler efficiency. However, if the CO limit
of 200 ppm maximum in the flue gases is strictly observed as a criteria
for setting up the optimum NOX emission firing mode, then efficiency
should not be impaired or changed since any loss in efficiency due to
increased carbon loss will be offset by the increase in efficiency resulting
from operating at low excess air. Unfortunately, unless at least a brief
form of efficiency test is conducted it will not be known whether boiler
performance under low NOX conditions is equal to or less than efficiency
at baseline operation.
ASME EFFICIENCY TEST
Once the test program has been completed and the optimum low NOX
emission operation modes have been defined, it is recommended that baseline
and optimum low NOX tests be repeated but, this time, additional data
should be obtained to be able to calculate boiler efficiency. With the
required data in hand, boiler efficiency may be readily calculated using the
A.S.M.E. Steam Generating Units, Power Test Codes, Abbreviated Efficiency
Test, heat loss method. Examples of typical performance data required and
calculations made are shown in ASME test forms shown in Tables E-l and E-2.
Two pieces of information are required to make these calculations in addition
to some of the test data taken during the test program. These are: (1)
ultimate fuel analysis on an as received basis, and (2) percent carbon on
particulate. The ultimate analysis is not generally available in most power
plants and requires taking coal samples during the test and submitting them to
the laboratory for the required determinations. Likewise, carbon loss data
also are not readily available and must be obtained by extracting particulate
samples simultaneously as each test is run and analyzing for carbon content.
Particulate samples may be obtained using a variety of test techniques. The
prescribed method for obtaining representative samples under isokinetic condi-
tions for compliance with particulate emission regulations is EPA Method 5 as
printed in the Federal .Register, Vol. 36, No. 247, December 23, 1971 (and as
amended). This method, however, requires considerable effort to conduct the
test and, since the objective is not compliance, other more simple methods
could be used, such as, the alundum thimble apparatus which is more simple and
easier to operate. This apparatus is an "in stack" sampler with the thimble
located in the flue gas stream requiring no external heating complications.
Fly ash samples are collected in the thimble as the sample gases are drawn
through the apparatus isokinetically. The plant engineering department or
results department personnel should be consulted for assistance in obtaining
suitable particulate grab samples for this purpose.
- 87 -
-------
e5350
For ready reference, the following are the required data necessary
for the A.S.M.E. Abbreviated Efficiency Test:
• Coal as fired ultimate analysis
(1) Moisture
(2) BTU
(3) Carbon
(4) Hydrogen
(5) Sulfur
(6) Ash
• Percent carbon on particulate (fly ash)
• Temperature of air for combustion
• Gas temperature leaving boiler, economizer, air heater
• Flue gas analysis (percent by volume)
02
C02
CO
N2 (by difference)
Efficiency calculations may be made using the above data for each
test. The efficiency for "Low HOX" emission operation can then be compared
to baseline operation efficiency to make sure that there is no loss in
efficiency for low NOX operation. As indicated above, if the 200 maximum
ppm CO limitation is observed as the criteria for setting up the low NOX
combustion mode, efficiency should be equal to or greater than that at baseline
conditions. If not, excess air should be increased for the firing mode in
question and efficiency tests repeated until optimum conditions are established.
PERFOR1IAHCE CHECKS FOR EFFICIENCY
Once optimum conditions for the low NOX emission combustion modes
have been determined, test data should be recorded and filed for future
reference and comparison with similar data obtained in periodic checks of
boiler performance (efficiency). These spot checks, however, need not be
calculated by the A.S.M.E. method requiring ultimate analyses and carbon
loss data as discussed above but, instead, can be made simply by a spot
check of combustion conditions. The important combustion criteria affecting
boiler performance are 02 or C02, CO, exit gas temperature and stack opacity.
Operation of the boiler should be set up duplicating optimum low NOX condi-
tions with controls either on automatic but preferably on hand control and
the boiler locked out of the automatic dispatch control load control system.
This way the boiler will operate at constant load and the above measurements
and observations can then be made at steady state conditions and the results
compared to previous data recorded for the more complete efficiency tests.
Deviations in the values of 02 or COo, CO, exit gaa temperature and stack opacity
will pinpoint changes in efficiency. The reliability of the spot check
will, of course, be dependent on the accuracy of the values measured,
especially with reference to 02 and CO. Assuming no major or drastic
- 88 -
-------
TABLE E-l
SUMMARY SHEET
ASME TEST FORM
FOR ABBREVIATED EFFICIENCY TEST
PTC 4.1-a(1964)
TEST NO. 1A BOILER NO. 6
DATE^-13-72
O^NER OF PLANT TVA LOCATION widows Creek
TEST CONDUCTED BY Esso Research & Engineering Co .OBJECTIVE OF TEST Boiler Performanc®u9ATiOMi Hrs .
BOILER, MAKES. TYPE B&W Radiant RATED CAPACITY 125 MU
STOKER, TYPE &. SIZE
PULVERIZER, TYPE & SIZE TyPe E BURNER, TYPE S. SIZE
FUEL USED Bituminous Coal MINE COUNTY STATE
SIZE AS Fl = fD
PRESSURES 4 TEMPERATURES FUEL DATA
1
2
3
4
5
6
7
8
• 9
10
II
' * j
3
14
STEAM PRESSURE IN BOILER DRUM
STEAM PRESSURE AT S. H. OUTLET
STEAM PRESSURE AT R. H. INLET
STEAM PRESSURE AT R. H. OUTLET
STEAM TEMPERATURE AT S. H. OUTLET
STEAM TEMPERATURE AT R.H. INLET
STEAM TEMPERATURE AT R.H. OUTLET
WATER TEMP. ENTERING (ECON.) (BOI LER)
STEAMQUALITY7. MOISTURE OR P. P.M.
AIR TEMP. AROUND BOILER (AMBIENT)
TEMP. AIR FOR COMBUSTION
TEMPERATURE OF FUEL
GAS TEMP. LEAVING (Boiler) (Eton.) (Air Htr.)
corrected to auorontee)
PSIO
psia
psio
psia
F
F
F
F
F
F
F
F
F
%%
?7 ZL*
UNIT QUANTITIES
15
16
17
18
19
20
21
22
23
24
25
ENTHALPY OF SAT. LIQUID (TOTAL HEAT)
ENTHALPY OF (SATURATED) (SUPERHEATS D)
STM.
ENTHALPY OF SAT. FEED TO (BOILER)
(ECON.)
ENTHALPY OF REHEATED STEAM R.H. INLET
ENTHALPY OF REHEATED STEAM R. H.
OUTLET
HEAT ABS/LBOF STEAM (ITEM 16 - ITEM 17)
HEAT ABS/LB R.H. STEAMOTEM 19-ITEM 18)
DRY REFUSE (ASH PIT » FLY ASH) PER LB
AS FIRED FUEL
Btu PER LB IN REFUSE (WEIGHTED AVERAGE)
CARBON BURNED PER LB AS FIRED FUEL
DRY GAS PER LB AS FIRED FUEL BURNED
Btu/lb
Btu/lb
Btu/lb
Btu/lb
Btu/lb
Btu/lb
Btu/lb
Ib/lb
Btu/lb
Ib/lb
Ib/lb
HOURLY QUANTITIES
26
27
28
29
?.o
.11
ACTUAL WATER EVAPORATED
REHEAT STEAM,xFLOW
RATE OF FUEL FIRING (AS FIRED ~t)
TOTAL HEAT IMPIJT (Item 28 X Item 41)
1000
HEAT OUTPUT IN BLOW-DOWN WATER
HEAT*" "'"" 2o'""n JOHherr. 27-lt.m 21) i(|._ 7n
OUTPUT 1000
Ib/hr
Ib/hr
Ib/hr
kB/hr
kB/hr
kB/hr
7^12
iNi-J
d. 64
II. I,
FLUE GAS ANAL. (BOILER)IECON) (AIR HTR) OUTLET
32
33
34
35
36
CO,
0,
CO
N, (BY DIFFERENCE)
EXCESS AIR
% VOL
\ VOL
% VOL
T. VOL
r3
/Y-ity
iJ !_ -.,
t)i (>*
too . »
COAL AS FIREO
PROX. ANALYSIS
7
33
39
40
MOISTURE
VOL MATTER
FIXED CARBON
ASH
TOTAL
41
42
Btu per Ib AS FIREO
ASH SOFT TEMP.'
ASTM METHOD
% wt
-^4-
Ul4-$1-
COAL OR OIL AS FIREO
ULTIMATE ANALYSIS
43
44
45
46
47
40
37
CARBON
HYDROGEN
OXYGEN
NITROGEN
SULPHUR
ASH
MOISTURE
TOTAL
67, Z7
^.2-?
0.77
/J^, "2&
£,C32.
o.f
I2»H>
Not Required for Efficiency Testing
t For Point of Measurement See Par. 7.2.8.1-PTC 4.1-1764
89
-------
CALCULATION SHEET
TABLE E-2
ASME TEST FORM
FOR ABBREVIATED EFFICIENCY TEST
PTC 4.1-b (1964)
Revised September, 1965
OWNER Or PLANT TVA TEST NO. ]^
30
24
25
36
65
66
67
6*
6»
70
71
72
HEAT OUTPUT IN BOILER BLOW-DOWN WATER »L8 OF WATER BLOW-DOWN PEf
// impractical to weigh refust, this
item con be estimated at follows
% ASH IN AS FIRED COAL
100 — Jfc COMB. IN REFUSE SAMPLE
ITEM 43 fir EM 22 ITEM2J~
CARBON BURNED &7«2-7 01/^27 X f /^ / Q. 66
PER LB AS FIRED - - • »
FUEL L I4,SUB ^
DRY CAS PER LB 11 CO, » 80, + 7{N, * CO)
BURNED «CO« * C0) /
ITEM 32 ITEM 33 I ITEM 35 ITEM 34
11 x /4,y.* 8x 3,3 * 7y*TE 4-18-72
ITEM 15 ITEM 17
1 HP x —
1000
MOTE; IF F
PIT REFUSE
IN COMBUST
SHOULD BE
SEPARATEL
COMPUTATI
kB/hr
LUE OUST A ASH
DIFFER MATERIALLY
IBLE CONTENT, THEY
ESTIMATED
Y. SEE SECTION 7.
ON*.
LB AS FIRED FUEL f ' S)
I I ITEM 24 ITEM 47
L 267 J
n C0 ITCU — ITEM 33 -
.26I2N, - in _ CO .
2 .2682 (ITEM 35) - (ITEM
,t _ ITEM 34 }
2
HEAT LOSS EFFICIENCY
HEAT LOSS DUE LB DRY CAS ITEM 25 (ITEM 11) -(ITEM III ^, ^
TOORVGAS * PERLSAS xC x ('!,,_ '.,-,) * '/, * *»•" a /7 -j P 5 - 790
FIRED FUEL p Uni, //«b 3 / ^" 10 ' '
HEAT LOSS DUE TO _LBH,0 PER LB . f ,ENTHALPY OF VAPOJl ATI PSIA*
MOISTURE IN FUEL 'AS FIREO FUEL X l (ENTHALPY OF VAPJJ*^ I 1 r».A *
-(FHTHALPY OF UQUIDAT T AIR)] * ITEM.V »[(ENTH
- ^g «oo
AT 1 PSIA & T ITEM 13) -IENTHALPY OF LIQUID AT T IT!
T CAS LVG1
/XZ7/7
ALPY OF VAPOR
M 11)1 ««•?/. 3
HEAT LOSS DUE TO H,0 FROM COMB. OF H, « 9H, X [(ENTHALPY OF VAPOR AT 1 PSIA 8. T CAS
Lf-,"ifl /2.T/7 9 LVC) -(ENTHALPY OF LIOUID^T T AIR)]
« » X ' 'EM U X {(ENTHALPY OF VAPQR ATJ.PSIA fc T ITEM 13) - (ENTHALPT OF LIQUID AT
100 T ITEM 11)] • ty-SZftS.. .
HEAT LOSS DUE TO ITEM 72 ITEM 23 n
COMBUSTIBLE IN REFUSE « £ /^~J?7 * *jl2'l * 1 fif'i/"'O
HEAT LOSS DUE TO TOTAL BTU RADIATION LOSS PER HR
RADIATION* LB AS FIRED FUEL - ITEM M
UNMEASURED LOSSES ••
TOTAL
EFFICIENCY * (100 - ITEM 71)
Biu/l.
AS FIREO
FUEL
"7^)0
£33
V&t*
w.*
^
LOSS w
•RHv-*
100 «
— x 100 >
41
~ X 100 >
41
_X100 «
41
68
41
" X 100 .
41
41
•^^•"•M^M
LOSS
*
Mo
in
— — — — .
I,U
.
-' ' ' .
-
t For rig«raw* ^«'^rmir>arf«n «f CMCVS* *ir ••• App«nd«« ?.? — PTC 4.1-1964
* II loit«i or* n-.r n.oiur.d. ui. ABM A Stondo>d Rxliation L»it CK*n, Fij. I, PTC 4.1-1944
•* Unm«*tvr>d laxtl li»*d in PTC 4.1 but not tobuloKd ooa» r»oy ky BfO'idcd l«r ky •>li|nin| • mu>u«llr
•t'**d vpon >alv> lor (Km 70.
90
-------
c5350
changes in coal quality and accurate $2 an^ CO measurements duplicating the
initial test run, boiler efficiency should also be about the same, provided
that there is no change in exit gas temperature. An increase in exit gas
temperature (+10 to 20°F) with no change in 02, CO or stack opacity would
signal slagging of the furnace walls, or fouling in the superheater, reheater,
economizer or air heater sections. Steps should then be taken to clean up the
boiler surfaces by thorough and perhaps prolonged operation of the slag and
sootblowers, by dropping load severely to shock and deslag furnace surfaces, or
by taking the boiler out of service for clean up.
An increase in stack opacity at the same 02 level could result
from a change in coal quality. Under these circumstances, however, CO
values for equivalent 02 readings, should also be higher. This occurrence
would require an increase in excess air so that maximum CO levels (200 ppm)
are re-established to clean up the stack. Obviously, then it would not be
possible to repeat the optimum low NO combustion mode firing this type
of coal.
Most utility boilers maintain hourly logs of pertinent boiler data
which include 62 and boiler exit gas temperature measurements. Periodically,
say on a frequency of once per week or bi-weekly, CO measurements and stack
opacity observations could be made at optimum low NOX combustion conditions,
as discussed previously, to document and ascertain that boiler efficiency
standards are being maintained. Boiler log data over a period of time then
will show if there is any drift in operating conditions leading to lower
efficiencies and pinpoint the need for preventative maintenance to restore
the boiler to full capabilities.
- 91 -
-------
APPENDIX F
To Obtain
g/Mcal
106 Btu
MBH/ft2
MBH/ft3
Btu
10 3 Ib/hr* or MBH
103 Ib/hr* or MBH
Ib/MBtu
i ft
to
1 ±n
2
ft
ft3
Ib
Fahrenheit
psig
psia
iwg (39.2°F)
From
ng/J
GJ
GJ-hr «m
GJ.hr-l.m-3
gtn cal
GJ/hr
MW
ng/J
m
cm
2
^
m
3
m
Kg
Celsius
Kelvin
Pa
Pa
PA
CONVERSION FACTORS
SI Units to Metric or English
Multiply By
0.004186
0.948
0.08806
0.02684
3.9685 x 10~3
0.948
3.413
0.00233
3.281
0.3937
10.764
35.314
2.205
tp = 9/5 (tc)+32
t_. = 1.8K - 460
r
psig pa
PPsia= CPp.) (1.45*10-*)
Piwg -
-------
APPENDIX F
CONVERSION FACTORS
English and Metric Units to SI Units
To Obtain
ng/J
ng/J
GJ'hr^-m"2
GJ-hr^-m"3
GJ/hr
MW
m
cm
2
m
m
kg
Celsius
Kelvin
Pa
Pa
Pa
From
Ib/MBtu
g/MCal
MBH/ft2
MBH/ft3
103 Ib/hr*
or 106 Btu/hr
103 Ib/hr*
or MBH to MW
ft
in
ft2
ft3
Ib
Fahrenheit
psig
psig
iwg (39.2°F)
Multiply By
430
239
11.356
37.257
1.055
0.293
0.3048
2.54
0.0929
0.02832
0.4536
tc - 5/9 (tp-32)
tR - 5/9 (ty-32) + 273
P - (P . + 14.7) (6.89x10
pa % psig
Ppa ' (Ppsia) <6-895xl03)
PPa * ^iwg^249'"
Multiply Concentration
To Obtain ne/J of in ppm at 3Z 00 by
Natural Gas Fuel
CO 0.310
HC 0.177
NO or NO (as equivalent NO.) 0.510
S02 or S0x 0.709
Oil Fuel
CO 0.341
HC 0.195
NO or NOX (as equivalent N02) 0.561
S02 or SOX 0.780
Coal Fuel
CO 0.372
HC 0.213
NO or NOX (as equivalent N02) 0.584
S02 or SOX 0.850
*lb/hr of equivalent saturated steam
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/8-80-027
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Guidelines for NOx Control by Combustion
Modification for Coal-fired Utility Boilers
5. REPORT DATE
May 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
E.H. Manny
8. PERFORMING ORGANIZATION REPORT NO.
EE.116E.79
9. PERFORMING OROANIZATK N NAME AND ADDRESS
Exxon Research and Engineering Company
P.O. Box 101
Florham Park, New Jersey 07932
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-1415
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Special; 6/74-12/79
14. SPONSORING AGENCY CODE
EPA/600/13
15.SUPPLEMENTARY NOTES BERL-RTP project officer is Robert E. Hall, Mail Drop 65 919A
541-2477.
is.ABSTRACT The reporj-? wnjch has been reviewed by industry experts, reflects the
experience developed in successfully applying combustion modifications to reduce
NOx emissions from coal-fired utility boilers. Although the report emphasizes coal-
fired equipment, the same principles can be applied to gas- and oil-fired systems.
Techniques , methods, and step-by-step procedures are detailed by example to guide
utility personnel who may desire to conduct their own NOx emission reduction pro-
grams. Background information on operating parameters affecting NOx, necessary
to understanding NOx emission control, is also included. Field studies were conduc-
ted from 1971 to 1979 to assess the feasibility of combustion modification to control
NOx and other pollutants from large utility boilers. During these investigations,
significant NOx reductions were demonstrated. For example, using a combination of
staged combustion, low excess air firing, and other techniques reduced NOx by an
average of 38%, over a range of 12 to 62%, in more than 35 utility boilers.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Air Pollution
Coal
Utilities
Boilers
Combustion
Nitrogen Oxides
Sulfur Oxides
Dust
Aerosols
Slagging
Air Pollution Control
Stationary Sources
Combustion Modification
Staged Firing
Low Excess Air
Flue Gas Recirculation
Particulate
13 B
2 ID
13A
21B
07B
11G
07D
13H
B. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report/
Unclassified
20. SECURITY CLASS (This page)
Unclassified
21. NO. OF PAGES
101
22. PRICE
EPA Form 2220-1 (9-73)
94
------- |