U.S. Environmental Protection Agency Industrial Environmental Research EPA"600/
Office of Research and Development Laboratory
Research Triangle Park, North Carolina 27711 Jllly 1977
EPA~600/7~77~073b
PROCEEDINGS OF THE SECOND
STATIONARY SOURCE
COMBUSTION SYMPOSIUM
Volume II. Utility and Large
Industrial Boilers
Interagency
Energy-Environment
Research and Development
Program Report
LIB?
U. S. EIIV;• '
EDISCif, K»J>,
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research-and Development, U.S.
Environmental Protection Agency, have been grouped into seven series.
These seven broad categories 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 seven series
arc:
' <.
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
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series" result from
the effort funded under the 17-agency Federal Energy/Environment
Research and Development Program. These studies relate to EPA's
mission to protect the public health and welfare from adverse effects
of pollutants associated with -energy systems. The goal of the Program
is to assure the rapid development of domestic energy supplies in an
environmentally—compatible manner by providing the necessary
environmental data and control technology. Investigations include
analyses of the transport of energy-related pollutants and their health
and ecological effects; assessments of, and development of, control
technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal
Agencies, and approved for publication. Approval does riot
signify that the contents necessarily reflect the views and
policies of the Government, nor does mention of trade names
or commercial products constitute endorsement or recommen-
dation for use.
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161-.
-------
EPA-600/7-77-073b
July 1977
PROCEEDINGS OF THE SECOND
STATIONARY SOURCE
COMBUSTION SYMPOSIUM
Volume II. Utility and Large
Industrial Boilers
(N
Symposium Chairman Joshua S. Bowen.
Vice-Chairman Robert E. Hall
Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Program Element No. EHE624
LIB? .'••.?.
;^_ii.J.,_ 03317
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington' D.C. 20460
-------
-------
PREFACE
These proceedings document the more than 50 presentations and discus-
sions of the Second Symposium on Stationary Source Combustion held August
29 - September 1, 1977, at the Marriott Hotel in New Orleans, Louisiana.
Sponsored by the Combustion Research Branch of the EPA's Industrial
Environmental Research Laboratory-Research Triangle Park, the symposium
presented the results of recent research in the areas of combustion
processes, fuel properties, burner and furnace design, combustion
modification, and emission control technology.
Dr. Joshua S. Bowen, Chief, Combustion Research Branch, was Symposium
Chairman; Robert E. Hall, Combustion Research Branch, was Symposium Vice-
Chairman and Project Officer. The Welcoming Address was delivered by Dr.
John K. Burchard, Director of IERL-RTP; the Opening Address was delivered
by Robert P. Hangebrauck, Director, Energy Assessment and Control Division,
IERL-RTP; and Dr. Howard B. Mason, Program Manager NOX Environmental Assessment
Program, Acurex Corporation, delivered the Keynote Paper.
The symposium consisted of six sessions:
Session I:
Session II:
Session III:
Session IV:
Session V:
Session VI:
Small Industrial, Commercial and Residential Systems
Robert E. Hall, Session Chairman
Utility and Large Industrial Boilers
David G. Lachapelle, Session Chairman
Special Topics
David G. Lachapelle, Session Chairman
Stationary Engine and Industrial Process Combustion
Systems
John H. Wasser, Session Chairman
Advanced Processes
G. Blair Martin, Session Chairman
Fundamental Combustion Research
W. Steven Lanier, Session Chairman
m
-------
-------
VOLUME II
TABLE OF CONTENTS
- SESSION II: UTILITY AND LARGE INDUSTRIAL BOILERS -
Page
"Field Testing: Application of Combustion Modification to Power
Generating Combustion Sources," A. R. Crawford, E. H. Manny,
W. Bartok 3
"Analysis of NOX Control in Stationary Sources," 0. W. Dykema .... 41
"Overfire Air Technology for Tangentially Fired Utility Boilers
Burning Western U.S. Coal," A. P. Selker, R. L. Burrington 67
"The EPRI Program on NOX Control Using Combustion Modification
Techniques," K. E. Yeager, D. P. Teixeira 103
"Design and Scale-Up of Low Emission Burners for Industrial and
Utility Boilers," R. Gershman 121
"Cyclone Boilers — Their NOX Emissions and Population," T. E.
Ctvrtnicek, S. J. Rusek 123
"Statistical Aspects of Corrosion From Staging Combustion in a Wall
Coal-Fired Utility Boiler," 0. W. Tukey 143
"Noncatalytic Reduction of NOX with NH3," W. Bartok 145
"Western Coal Use in Industrial Boilers," K. L. Maloney, P. L.
Langsjoen 163
- SESSION III: SPECIAL TOPICS -
"A Survey of Sulfate, Nitrate, and Acid Aerosol Emissions and Their
Control," J. F. Kircher, A. Levy, J.O.L. Wendt 211
"Inventory of Atmospheric Emissions from Stationary Point Sources,"
V. E. Kemp, 0. W. Dykema 257
-------
TABLE OF CONTENTS (Concluded)
"Emissions Assessment of Conventional Combustion Systems," B. J.
Matthews
Panel: Combustion Source/Air Pollution Regulations ~ Present and
Projected
279
301
-------
SESSION II:
UTILITY AND LARGE INDUSTRIAL
BOILERS
DAVID G. LACHAPELLE
CHAIRMAN
-------
-------
FIELD TESTING: APPLICATION OF COMBUSTION MODIFICATION
TO POWER GENERATING COMBUSTION SOURCES
By:
A. R. Crawford, E. H. Manny, and W. Bartok
Exxon Research and Engineering Company
Linden, NO 07036
-------
-------
ABSTRACT
A field study program was conducted to assess the applicability of com-
bustion modification techniques to the control of NOX and other pollutant
emissions from power generating combustion sources. This work is part of a
continuing series of field test programs performed by ER&E on utility boilers
and other power generating combustion sources. The goal of this research is
to determine whether known modifications can be applied to the combustion
process for NOX control, without causing deleterious side effects.
The studies reported in this paper include field tests on utility
boilers and gas turbines employed for power generation. Comprehensive,
statistically designed test programs were conducted to evaluate the effect of
combustion modifications on NOX and other gaseous emissions. In addition,
particulate mass and size distribution and boiler efficiency were determined
under baseline and low NOX operating conditions.
The most extensively studied combustion modification for utility boiler
applications was staged firing at low excess air. This approach can achieve
reductions in NOX emissions up to about 50% based on the results of short term
tests. With the focus of the program on NOX emission control for coal fired
utility boilers, special attention was paid to the determination of potentially
adverse side effects—increased combustible emissions, unwanted changes in
particulate mass loading and size distribution, reduced boiler efficiency,
increased waterwall slagging and external corrosion, and flame problems.
Short term tests indicate that staged combustion may be applied to coal
fired utility boilers. The extent of waterwall corrosion and slagging could
not be determined unequivocally based on the results of 300-hour corrosion
probing runs under low NOx and baseline operating conditions. For this reason,
a long term waterwall corrosion test of at least six month duration was initi-
ated at Gulf Power Company's Crist station on a 500 MWe front wall fired
boiler. This program is conducted jointly with Foster Wheeler, the boiler
manufacturer. In addition to corrosion probing, ultrasonic mapping of the
waterwall tube thicknesses is carried out at the beginning and end of the
baseline and low NOX operating periods, and measurements are made on specially
installed waterwall test panels to determine the rate of corrosion under staged
firing with portions of the furnace operated under fuel rich, reducing con-
ditions .
-------
ACKNOWLEDGMENTS
The research described in *:his paper was performed under the sponsorship
of the U. S. Environmental Protection Agency, pursuant to Contract No.
68-02-1415. The field testing studies on the Mercer Station Boiler No. 1 of
Public Service Electric and Gat. Company was funded in part by the Electric
Power Research Institute (RP 2UO). The authors wish to acknowledge the con-
structive involvement and comments of Mr. R. E. Hall, the EPA Project Officer.
Thanks are due to Professor John Tukey of Princeton University for his advice
on statistical test design. The cooperation of equipment manufacturers,
General Electric Company, Foster Wheeler, Babcock and Wilcox, and Combustion
Engineering in selecting representative units is highly appreciated. Special
thanks are due to Mr. J. Vatsky of Foster Wheeler for their effective parti-
cipation in the long term corrosion test program on the Crist Boiler No. 7 of
Gulf Power Company, which is being carried out under a subcontract to Foster
Wheeler. The voluntary participation of the power generation combustion
equipment operators (Public Service Electric and Gas Company of New Jersey,
Houston Lighting and Power Company, Gulf Power Company, East Kentucky Power
Cooperative, and Colorado Public Service Company) in making available their
boilers and gas turbines for testing is gratefully acknowledged. Finally,
thanks are due to Messrs. L. W. Blanken, J. E. Bond, J. J. Eggert, W.
Petuchovas, R. W. Schroeder, and Mrs. M. V. Thompson for their assistance in
performing these studies.
-------
SECTION 1
INTRODUCTION
Exxon Research and Engineering Company has been conducting field studies
under EPA sponsorship (and in part supported by the Electric Power Research
Institute) on the application of combustion modification techniques to the con-
trol of pollutant emissions from utility boilers. The emphasis in these studies
has been on controlling NOX emissions without adverse side effects, such as
increases in other pollutant emissions, and equipment safety and operability
problems.
Because of the difficulty of controlling NOX emissions from coal fired
boilers, the emphasis in the Exxon Research field studies has been on coal fired
units. Using a mobile sampling-analytical system designed and built for the
first Exxon Research field studies conducted under EPA sponsorship (i) on gas,
oil, and coal fired utility boilers, gaseous emission measurements have been
obtained on a large number of coal fired utility boilers (2_».3.). This compre-
hensive program includes the measurement of not only gaseous species under
baseline and modified combustion conditions, but also particulate mass loading
and size distribution measurements. This characterization work is being ex-
tended to potentially hazardous inorganic and organic trace constituents of
combustion gases.
Reflecting the major remaining question that interferes with the application
of combustion modifications (staged combustion) to coal fired utility boilers
for NOX emission control, an extensive effort has been mounted to resolve whether
external furnace tube corrosion is accelerated by staged firing of high sulfur
coals.
Earlier short term measurements in the Exxon field studies using corrosion
probes could not produce an unequivocal answer to this problem. Therefore, a
long term corrosion test program was initiated in 1976 on Crist boiler No. 7 of
Gulf Power Company in cooperation with that utility and the manufacturer of the
unit, Foster Wheeler Energy Corporation, as Exxon Research's subcontractor.
A status report will be presented on the emission tests performed during
the current phase of this program and on 'the long term corrosion tests at Gulf
Power's Crist Station.
-------
SECTION 2
GASEOUS EMISSIONS
2.1 Mercer Station, Boiler No. 1
Public Service Electric and Gas Company (New Jersey)
Mercer Unit No. 1 is a twin furnace, front-wall fired, wet-bottom Foster
Wheeler boiler. This unit was selected for testing because of its flexibility.
for combustion modification, and to determine the side-effects of such modifi-
cations on a wet-bottom unit. The boiler has 3579 m2 (38,526 ft2) of furnace
heating surface, a furnace volume of 5333 m3 (188,332 cubic feet), with each
furnace measuring 11.95 m C39 feet 2-1/2 inches) in width and 7,99 m (26 feet
2-1/2 inches) in depth. Maximum continuous rated steam flow is 934.4 t/hr.
(2,060,000 Ib/hr) at 16.7 MPa (2400 psig) pressure and 867 K (1100'F) super-
heat steam temperature and 839 K (1050°F) reheat steam temperature. Three
ball-type pulverizers feed the 24 burners arranged in three rows of four
burners in each of the two twin furnace front walls. The pressurized furnaces
are equipped with flat floors and slag-taps.
Analysis of these test results indicate that all of the operating vari-
ables included in the experimental program had a significant effect on NOX
emission levels. As shown in Figure 1, at full load (290 MWe), baseline
operation resulted in 739 ng/J (1383 ppm) NOx emissions. This high NO level
is caused by the unusual furnace design of this boiler in which pulverized
coal can be burned at low loads with a wet bottom furnace. The flat furnace
floor is relatively close to the bottom row of burners so that unusually high
gas temperatures are maintained in the bottom of the furnace in order to
maintain the slag in a molten state. Within the limited operating flexibility
under full load operation, firing with low excess air was the most important
variable, reducing NOg emissions by an average of 24%. Biased firing (top row
burners fuel-lean; bottom and middle row burners fuel-rich) reduced NO*
emissions by an average of 16%. Reducing the secondary air register setting
from F-2 (maximum opening) to F-l (partially closed down) on the reheat fur-
nace increased NOx emission levels by about 4% under normal firing operation
(Si) and reduced NOg emissions by an average of 8% under biased firing
operation (82). At full load, the lowest NOx emission levels were obtained
under test run No. 8 operating conditions of biased firing, low excess air,
and closed down secondary air registers. This test condition produced 876
ppm NOX, a reduction of 36% from 793 ng/J (1383 ppm) produced under baseline
operation.
-------
Twelve test runs were conducted at approximately 220 MWe, the maximum
load achievable under staged firing (1 mill on air only) operation. Under
normal firing operation (S^) , NOx emissions were reduced by an average of 12%
due to the load reduction from 290 to 220 MWe (24% reduction) . Low excess air
operation reduced NOx emissions by an average of 5% under normal firing (S^)
and by 50% under staged firing (83). Staged firing (top row of burners on air
only) was carried out with the secondary air registers of the top row of
burners set at the maximum opening F-2, partially closed down F-l, or almost
completely closed down position. As expected, the greatest reduction of NOx
emissions occurred when the top row secondary air registers were set at the
maximum opening, F-2. Thus, the average reduction from the NOx level for
normal firing of 618 ng/J (1078 ppm NOX) at 200 MWe were 24%, 43%, and 48%
under staged firing (83) conditions when the secondary registers were set at
closed, partially closed, and open positions, respectively. Test run No. 10
operating conditions of staged firing, low excess air and normal secondary air
register settings (F-2) produced an average level of 240 ng/J (356 ppm) NOX,
or 69% below the 651 ng/J (1136 ppm) level experienced under baseline operation
at about the same load.
Nine test runs were conducted at approximately 155 MWe which is the
normal night-time low load conditions for this boiler. Under baseline opera-
tions at this load (top burners with air and coal off), NOx emissions were
34% less than under full load operation. Low excess air and staged firing
operation at this load again resulted in large NOX emission reductions. Thus,
low excess air combined with staged firing (run No. 26 conditions) lowered
NOX emissions by 61% from baseline operation at low load.
In summary, this wet bottom furnace boiler demonstrated significant
emission reduction capabilities through combustion modification from the high
baseline operating level of 793 ng/J (1383 ppm) at full load. Excess air
level, firing pattern, and secondary air register setting were all important
combustion control variables. The optimum operations at full load (290 MWe),
intermediate load (220 MWe), and low load (155 MWe) produced NC^ emission
levels and percent reductions of 502 ng/J (876 ppm) or 27%, 204 ng/J (356
ppm) or 74%, and 201 ng/J (351 ppm) or 75%, respectively.
2.2 Sewaren Station - Boiler No. 5
Public Service Electric and Gas Company (New Jersey)
A field test program was conducted to determine the NOX emission reduction
capabilities of the No. 5 330 MWe oil fired boiler at the Sewaren generating
station, Sewaren, N.J. of the PSE&G Co. of New Jersey. Sewaren Boiler No. 5
is a Babcock and Wilcox, horizontally opposed fired unit with three rows of
four burners in both front and rear walls. The high pressure feedwater heaters
on this unit were out of service during the tests, limiting the maximum load to
approximately 285 MWe (gross) output. A total of 24 runs were made on this
unit. The major variables included in the experimental program were load
(201-275 MWe net), excess air, flue gas fecirculation (not into the windbox) ,
various staged firing patterns, and the effect of a combustion improver additive
supplied by Appollo Chemical Co. to the utility company for this boiler.
-------
Figure 2 is a plot of NO emissions vs. oxygen in the flue gas for the
24 test runs. Referring to Figure 2 it may be noted that emissions on this
unit are relatively low even at the higher excess air levels. At full load,
uncontrolled NOX emissions were estimated to be at the 229 ng/J (440 ppm)
level. Because of stringent plume opacity limitations imposed by local
authorities, operation at excess air levels below 2% 02 was not possible. As
shown by Figure 2, excess air level had a significant effect on NOX emission
levels for both normal and staged firing; 9 to 14 ng/J (16 to 24 ppm) NOX
reduction per 1% reduction in 02 level. Staged firing reduced NOX levels by
22% at 285 MWe. Staged firing patterns II, III and IV gave about equal results
with 5 burners on air only giving slightly better results than 4 burners on air
only. Lowering load by 26% reduced NOX levels by 19%. Injecting flue gas into
the lower level of the furnace had only a small effect (a reduction of 7%) on
NOX emissions. The combustion improver additive had no noticeable effect on
NOX emissions (Data points No. 20 and 21).
2.3 T. H, Wharton Station, Gas Turbine No. 42
Houston Lighting and Power Company
T. H. Wharton1a gas turbine No. 42 was tested in our current field test
program to obtain data on NOX emissions under baseline and wet control
operations. This General Electric model MS 7001B gas turbine is of modern
design, has a rated output of 50 MWe with a peak load of 54 MWe. It is
equipped to fire gas or oil and has water injection capabilities with either
fuel. Operating variables were gross load (0 to 55.5 MWe) and amount of
water injected (0 to 2.3% of combustion air, or about 0 to 1.2 g H20/g fuel).
Flue gas samples were taken from the centers of 12 equal areas in the
duct work leading to the stack. Gas samples were taken from each of the 12
sample points and analyzed separately in the first test run on this unit
(Run No. 10). As previously found in tests conducted by General Electric,
there was an insignificant degree of gaseous stratification within the duct.
For example, the 02 level varied from 16.4% to 16.5%, and NOX emissions varied
from 44 to 48 ng/J (77 to 83 ppm) on an as-measured basis. Consequently, on
the remaining test runs the gaseous samples from short, medium, and large
tubes were composited into the 4 probes.
Figure 3 is a plot of NOX emissions vs. gross load (MWe). Least squares
regression lines have been drawn through the data points obtained with 0%,
0.75-0.80% and 1.5% water injection rates. The baseline NQx emission level
at full load (51 MWe) was 73 ng/J (127 ppm) compared to 76 ng/J (133 ppm)
previously measured on Mbrgantown Station Unit No. 3 when fired at 54 MWe (2) __
Reducing load by 61% to 20 MWe resulted in reducing NO* emission levels by
43% to 41 ng/J (72 ppm).
Water injection was extremely effective in reducing NOx emission levels
on this oil fired gas turbine. An 83% reduction was attained at full load
when operating with 2.3% water injection. Lower water injection rates were
less effective, but still reduced NO-^. emission levels by substantial amounts
as may be seen in Figure 3.
-------
2.4 T. H. Wharton Station, Gas Turbine No. 43
T. H. Wharton's Gas Turbine No. 43, a General Electric model MS 7001B
gas turbine is a duplicate of T. H. Wharton's GT No. 42 in every respect.
Tests were conducted on this turbine while firing natural gas, the normal
fuel used at this plant. Major operating variables included in the experi-
mental program were gross load (20 to 56 MWe) and amount of water injected
(0 to 1% of the combustion air, or about 0 to 0.5 g H20/g fuel). Since inlet
air relative humidity had been shown to have a significant effect on NOg
emission levels, measurements of this variable were also included. Flue gas
samples were taken from the centers of four equal areas from four probes.
Figure 4 is a plot of the NC^ emissions vs. gross load for the 16 test
runs conducted on Unit No. 43. Linear, least squares regression lines have
been drawn through the data for 0%, 0.5% and 0.75% water injection rates. As
expected, both gross load and level of water Injection has a significant
effect on NOx emission levels. From a baseline NOX emission level of 42 ng/J
(73 ppm, 15% 02 basis) at 54 MWe, a load reduction of 53% (to 20 MWe) reduced
the NOx emission level by 45% to 23 ng/J (40 ppm). Thus, the percent re-
duction in NOX level is less than proportional to the percent reduction in
load.
Water injection had a larger influence on NOX emissions than load
reduction. Injecting water at rate of 1% of the air mass flow decreased NOX
emission levels from 74 ppm to 25 ppm, or by 66% at a gross load of 56 MWe.
Combining both load and water injection into a single linear multiple
regression analysis resulted in the following equation:
PPM NOX (15% 02 basis) - 21 + 0.92 gross load (MWe) - 52.2 (% water
injected)
Ninety-seven percent of the variation in NOx emission levels are related to
these two operating variables.
Figure 5 is a plot of the 9 test runs conducted at about full load (52
to 57.5 MWe). The vertical axis is PPM NOX (15% 02, dry basis) and the
horizontal axis is % water injected. A least-squares curve could be fitted
based on the correlation of Shaw (4_) to these data to produce the following
relationship:
PPM NOX (15% 02, dry basis) - 7o73e-l-00? tt H20)
As shown by the smoothness of the curve in Figure 5, a very good fit was
obtained with r2 = 0.996.
2.5 Gulf Power Company. Crist Station,'Boiler No. T
The No. 7 boiler at the Crist Station of the Gulf Power Company in
Pensacola, Florida is a horizontally opposed fired, pressurized unit built by
the Foster Wheeler Energy Corporation. This boiler, as will be discussed later in
this paper, was selected for long-term testing of the external corrosion of
furnace tubes that may result from staged firing of high sulfur coal. The
unit has a rated capacity of 500 MWe with superheat and reheat temperatures
11
-------
of 811/811 K (100Q°/1000eF)» respectively. Six pulverizers supply pulverized
coal to 12 burners each in the front and rear walls. Burners are arranged in
3 vertical rows in each wall with 4 burners in each row for a total of 24
burners. Operating variables included in the experimental field test design
were: gross load (full load and medium load), excess air level (normal and
high), staged firing patterns (including firing top burners lean and lower
burners rich), with lower burner registers closed down to 40% open, and top
row burner registers set at 80% open.
Figure 6 is a plot of NOX emissions vs. flue gas Q£ for the tests con-
ducted on this boiler. Linear, least squares lines were calculated and drawn
through the data representing Si (normal firing at 460 to 520 MWe), 82 (top
burners lean at 430-450 MWe) and 83 (top burners on air only at 410-440 MWe)
operations. Baseline NOX emissions under full load, normal excess air (3.6%
flue gas 02) were calculated to be 588 ng/J (1025 ppm). Biased firing, 83
operation at 10% excess air reduced NO^ emissions to an average of 281 ng/J
(490 ppm) at 10% excess air. As expected, low excess air and staging of the
combustion process has a considerable effect on reducing NOg emissions from
this boiler, in agreement with results obtained on other pulverized coal
fired units tested in this program.
2.6 East Kentucky Power Cooperative, Inc.
John Sherman Cooper Station, Boiler No. 2
The Cooper Station, No. 2 boiler is a front-wall fired Babcock and Wilcox
designed boiler rated at 200 MWe (gross load) and 703 t/h (703,700 Kg of steam
per hour). Six Babcock and Wilcox Type EL pulverizers (using the ball bearing
principle for grinding) feed 18 burners arranged in five rows. The furnace
has a width of 12.8 m (42 ft.) and measures 7.3 m (24 ft.) from front wall to
rear wall. Heating surface measurements are: boiler 385.8 m2 (4,153 ft.2),
furnace 1983.3 m2 (21,350 ft.2), primary superheater 6370.2 (68,575 ft.2),
secondary superheater 664.8 m2 (7,157 ft.2), reheat superheater 2708.6 m2
(29,158 ft.) and economizer 2717.9 m2 (29,258 ft.2), respectively. Superheat
steam outlet temperature 813 K is (1005°F) at 12.9 MPa (1890 PSI) pressure.
Reheat outlet steam temperature is 813 K (1005*F) at 3.1 MPa (445 PSI)
pressure.
This unit was selected for field testing as a candidate for a second
long term corrosion test on a front wall, high sulfur coal fired boiler repre-
sentative of the design practices of B&W. The proximate analysis of the coal
used in designing the boiler was 8% moisture, 33% volative matter and 46.5%
fixed carbon. Ultimate analysis components were 12.5% ash, 4.5% S, 4.2% H,
63.5% C and 1.2% N. Because of local regulations, this unit will be forced to
fire low sulfur coal, a factor that tends to disqualify this unit as a candidate
for long term corrosion testing.
A statistically designed experimental test program was conducted on this
boiler to characterize emission levels using key operating variables of gross
load (MWe), excess air, secondary air register settings, and staged firing
patterns including firing top row burners lean, and bottom burners rich.
12
-------
Figures 7 and 8 are plots of ppm N0xvs. % oxygen for the data obtained
in 32 gaseous emission test runs. Figure 7 presents data points collected at
high loads of 166 to 227 MWe while Figure 8 presents data for lower boiler
loads of 123 to 155 MWe. Least squares regression lines have been drawn
through the data representing each firing pattern tested. There is somewhat
more scatter of the data points around these lines than usually encountered.
This is likely to be due to the following two reasons. First, some of the
test runs were conducted when the boiler was operating under Automatic Dis-
patching System CADS), i.e. fluctuating loads. Boiler loads were largely
determined by East Kentucky Power Cooperative, Inc. system demand. Since
Cooper No. 2 unit is the most flexible unit in the system, its electric
output is controlled by ADS. Second, data have been combined for a wider
range of loads than usual, accounting for the wider scatter.
Referring to Figure 7, it can be seen that, as expected, excess air has
a highly significant effect on NO emission levels. At high loads (160-227
MWe) , NOx emissions were reduced by an average of 38 ng/J (66 ppm) and 36 ng/J
(62 ppm) for each 1% drop in average flue gas oxygen content under S^, normal
firing operation and S-, top burners fired lean operation, respectively.
Similar results are shown in Figure 8 for the lower load test results. Base-
line operation at high loads resulted in 330 ng/J (576 ppm NC^) at 20% excess
air, while low excess air, staged firing reduced NOx emission levels to a
low of 246 ng/J (429 ppm), a 26% reduction.
Under low load operation (150 to 155 MWe), as shown in Figure 8, NOX
emissions increased when the heat release per active burner was increased.
For example, test run 45 (15 active burners and 150 MWe load) produced NOX
emissions of 362 ng/J (631 ppm), while at about the same load (155 MWe) test
run 43 (18 active burners) produced only 330 ng/J (567 ppm) NOX. There are
several firing patterns for low load operation that are capable of reducing NO^
emission levels to below 229 ng/J (400 ppm) . Operating with the top row
burners on air only, however, is more effective in reducing NO emissions than
operating with top row burners firing lean. x
2.7
Colorado Public Service Company, Pueblo, Colorado
Comanche Station, No. 2 Boiler
Comanche No. 2 unit is a horizontally opposed fired, balanced draft
Babcock and Wilcox designed boiler equipped with overfire air ports for
controlling NOX emission levels. Each of the four pulverizers feed eight
conventional, circular register burners arranged in four rows of four burners
each in the front and rear walls of the furnace. There is a single row of
four overfire air ports above the top row of burners in both the front and
rear walls of the furnace. Secondary air is fed to the burners through three
compartmentalized wind boxes arranged so that dampers control the volume of
air fed to the lower two rows of burners, upper two rows of burners, and
overfire air ports of both front and rear walls. This boiler has a maximum
continuous rating of 1150 t/h (1,150,436 Kg/hr of steam at 350 MWe). Steam
temperatures and pressures at the superheater and reheater outlets are 814 K
(1005°F) at 17.3 MPa (2500 psig) and 814 K (1005°F) at 3.9 MPa (555 psig)
respectively. The furnace volume is 7590 m3 (268,000 ft3). Total heating
13
-------
surface is 31,244 m* (336,350 ft2) of which 25,904 m* (287,860 ft2) is
convection surface. '
Predicted boiler efficiency is 84.57%, based upon the following coal
specifications. Proximate analysis - 29.0% moisture, 33.4% volatile natter,
32.4% fixed carbon and 5,2% ash; Ultimate analysis - 5.2% ash, 0.6% sulfur,
3.8% hydrogen, 47.5% carbon, 29,0% water* 0.9% nitrogen and 13% oxygen;
HHV = 4586 cal/g (8250 BTU per pound); grindability index 55, and ASTM ash
softening temperature of 1450 K (21509F) at reducing conditions.
This boiler was selected for field testing because it represents an
example of a NSPS unit manufactured with built-in second stage air ports, a
NOX control approach deemed suitable for low sulfur Western coal firing by
its manufacturer. Operating variables included in the experimental design
were: gross load (full load and minimum load while using all 4 mills),
excess air level (normal and high) and overfire air dampers (0 to 100% open).
Secondary air registers were maintained at normal opening (about 65%) and
mill fineness was maintained at the normal level. Four probes, each con-
taining three sampling tubes were placed in the ducts downstream of the
economizer, and positioned to sample gases from the centers of 12 equal duct
areas.
Figure 9 is a plot of NOX emissions vs. the degree of opening of the
overfire air dampers. Lines have been drawn through the data representing
three different operating conditions; upper line - full load at high excess
air, middle line - full load at normal excess air, and lower line - reduced
load at normal excess air. Figure 10 is a plot of NOx emissions vs. oxygen in
the flue gas for data obtained in full load test runs.
Baseline operation (0% open overfire air ports and normal excess air) at
full load produced NOX emissions of 416 ng/J (726 ppm). This relatively low
level for a 350 MWe unit can be partially accounted for by the low nitrogen
(nominally 0.9%) and high moisture contents (29%) of the coal. Full use
of the overfire air ports (100% open) reduced NOX emissions by 62% to
159 ng/J (278 ppm) from baseline operation at full load. Partial use of
overfire air ports gave intermediate levels of NOX reductions: 56% re-
duction at 80% OFA, 47% reduction at 60% OFA, 47% reduction at 40% OFA, and
21% reduction at 20% OFA. Increasing excess air from normal (4.5% excess 02)
to high (5.2 to 5.7% excess 02) increased NOX emissions by 25 to 34% under
full load operation. Thus, excess air, as expected, has a large influence
on NOX emission levels. Reducing boiler gross load from 355 MWe to 275 MWe
(23%) reduced NCXg emissions by an average of 8%. This result is in line
with results obtained on other pulverized coal fired boilers tested in this
study.
H
-------
SECTION 3
POTENTIALLY ADVERSE SIDE EFFECTS OF "LOW NO '' OPERATIONS
X
An Inevitable consequence of operating a coal fired boiler with staged
firing and reduced excess air for NOX emission control is the change from an
oxidizing to a net reducing atmosphere in the region of the lower burner rows
of the furnace. Under these conditions flames lengthen out, become dark and
smokey, and there is a tendency toward incomplete combustion. Potentially,
there may be a change in particulate emissions due to burnout problems; particle
size distributions could change adversely, resulting in enhanced fine particulate
emissions and affecting the efficiency of collector devicesj boiler efficiency
may decrease as a result of increased amounts of unburned combustibles; flame
patterns could change causing problems due to instability or Impingement;
increased slagging of the furnace may be experienced as a result of lower ash
softening temperatures under reducing conditions; and, perhaps foremost,
particularly when firing high sulfur content coals, slagging and external cor-
rosion of the furnace tubes may increase as a result of the reducing conditions
of the furnace atmosphere. Studies on these potential side effects have been
conducted by Exxon Research under EPA Contracts No. 68-02-0277 and 68-02-1415.
The results of these studies C?_,.3) indicated that only external corrosion of
the furnace tubes appeared to be a significant problem for the coal fired
boilers tested. In the present paper, the extension of these studies to addi-
tional utility boiler field studies is summarized.
3.1 Particulate Emissions
Staging the combustion process and decreasing excess air levels to reduce
NOX emissions produces longer, "lazier" flames with a tendency toward Incomplete
combustion. Any Increase in unburned carbon as a consequence of "low NOX"
operation would have a corresponding adverse effect on boiler efficiency and
might tend to Increase particulate emissions from the boiler. Particle size
distribution might also be affected by this type operation which could have an
adverse effect on precipltator performance. These problems have been the
subject of past Exxon Research field studies (2) (3) which concluded that no
significant changes in particulate mass loading or particle size distribution
could be observed resulting from staged firing of coal in the utility boilers
tested.
Table I presents particulate mass loading data obtained in recent studies
which are in typical agreement with the results obtained in the prior programs.
These data were obtained using EPA Method 5 sampling trains for particulates.
15
-------
The data summarized in Table I was obtained on the pulverized coal fired, No. 1
boiler at the Mercer Station, and on the oil fired No. 5 boiler at the Sewaren
Station of PSEG Co. of Nev Jersey. Comparing the low NOX data with baseline
results it can be seen that there are no significant differences in particulate
mass loading for Mercer Boiler No. 1. Particulate mass loading results for
oil firing of Sewaren Boiler No. 5 are, as expected, considerably lower than
the levels measured in coal firing. This is due to the low ash content of the
fuel oil fired. Interestingly, for the oil fired boiler particulate emissions
measured under baseline and low NOX firing conditions are essentially identical.
Particulate size distribution data are presented in Table II. The results
for the coal fired Mercer Boiler No. 1 are typical of the particle size distri-
bution information obtained on other coal fired units tested in this program.
Note that there appears to be no significant difference in the size distribution
of particulates produced under baseline and low NOX operations, respectively.
Consequently, the previous conclusion is further supported that "low NOX" firing
of coal does not affect particle size distribution adversely. This observation
±B especially noteworthy in regard to the submicron, resplrable particulate size
range.
The particle size distribution data obtained on the oil fired Sewaren
Boiler No. 5 Is interesting in two respects. First, it agrees directionally
with the results for coal firing, i.e., low NOX firing appears to result in
only minor changes, if any, compared with baseline conditions. Second, the
fraction of particulates in the submicron (especially in the < 0.5 um) size
range is larger than for coal firing. This result may be indicative of the
mechanism of particulate formation through cenospheres in oil firing.
3,2 Boiler Efficiency
The effect of combustion process modification on boiler performance was
investigated and evaluated in prior Exxon Research field studies where particu-
late data were obtained under baseline and optimum "low NOX" firing conditions.
A small tendency for particulate carbon content (unburned combustibles) to
increase with "low NOX" firing was noted in these earlier studies (I), especially
in front wall and horizontally opposed fired boilers. An increase in unburned
carbon should result in lower boiler efficiency but this adverse side effect
did not materialize in the previous studies due to the offsetting effect of low
excess air operation under low NOX operating conditions which increases boiler
efficiency because of reduced stack losses. These results were confirmed in
subsequent field studies (3).
Table III summarizes boiler efficiency results that are typical of those
obtained in prior field test programs. Table III shows that the Mercer No. 1
boiler under low NOX conditions operated at exit gas 02 levels nominally below
2%, whereas under baseline operation 02 levels were in excess of 3%. The
partlculate carbon content under low NOX conditions was somewhat higher than
under the baseline conditions. The calculated values of the boiler efficiency
were, however, essentially the same for both baseline and "low NOx" operations,
i,e. about 90%. These results again confirm the offsetting influence of low
excess air operation (with low NOX firing) on boiler efficiency, to compensate
for the loss in efficiency due to increased carbon losses.
16
-------
3.3 Slagffing
As mentioned previously, low NOX combustion modifications for NOX emission
control result in a change In the furnace atmosphere adjacent to the lower
burners from oxidizing to net reducing conditions due to staging of the combus-
tion process and a reduction In overall excess air levels in the boiler. A
potential consequence of "low NOX" operation could be an increased tendency
toward slagging of the furnace walls in the vicinity of the lower burners and
in the critical hopper slope areas. Ash fusion temperatures in most coals
decrease by approximately 111 K C200*F) when measured under reducing conditions.
Thus, if a reducing atmosphere prevails in the bottom of the furnace under low
NOX conditions, the tendency would be for the ash to melt quicker and be more
fluid and sticky, fouling the furnace surfaces more readily.
Most coal fired boilers are designed for operation at 15 to 25 percent
excess air with oxidizing atmospheres prevailing in the furnace. In most
boilers, furnace surfaces approximately 3.0 m (10 ft.) above the burners are
adequately covered with slag blowers to remove any ash accumulating on, and
blanketing the furnace heat transfer surfaces. Under normal design conditions
operation of these blowers two or three times per day will keep the furnace
tubes clean. If lower ash fusion temperature coal is burned, as happens on
occasion, these areas will slag more rapidly requiring more frequent operation
of the slag blowers. Operating a boiler under "low NOx" conditions could have
a similar effect. However, low NOX operation results in changes in the furnace
atmosphere in the lower regions of the furnace where slag blowers normally are
not installed. Potentially, therefore, an adverse side effect of "low NOX"
firing could be an increased tendency towards slagging of the furnace surfaces
where no means exist for removal of such ash accumulations. As a consequence,
boiler availability could be impaired by forced shutdowns to remove the accumu-
lated slag. In instances where this occurs, a solution to this problem would
be the installation of additional slag blowers in the lower areas of the furnace
to remove accumulations before they have a chance to build to critical pro-
portions.
In past Exxon Research field test programs — — utility boilers have
been operated for periods up to 1000 hours under low NOX firing conditions
without noticeably increasing slagging conditions or causing boiler shutdowns.
More frequent use of existing slag blowers may have been necessitated on
occasions, but, if so, increased slagging problems were not apparent. The
lower furnace surfaces, where no removal facilities exist, also presented no
unusual Increased slagging problems during these tests. Accordingly, under most
conditions studied, it appears that "low NOX" operation did not increase
slagging conditions to a point where normal removal facilities could not handle
the deposits.
3.4 Flame Problems
Pulverized coal combustion systems Inherently are much more complicated
than those used for oil or gas firing. Liquid or gaseous fuels can be handled,
distributed and controlled uniformly to each burner with relative ease. On the
other hand, pulverized coal systems require pulverizers, first, to grind the
coal to the required fineness and then distribute the fuel to the burners
17
-------
pnematically. On utility Boilers each pulverizer normally serves 4 to 6
burners through huge pipes. The fuel flow to these burners is virtually always
non-uniform. In addition, Because of problems of erosion, there are no
regulating valves in the burner lines which might be used to balance fuel flow
between burners, as is possible with oil for gas fuels. Furthermore, constant
change is gradually, but continually taking place in the combustion parameters
due, in most part, to pulverizer wear. Combustion control in a pulverized coal
system, because of these factors, is considerably more difficult than with other
types of firing systems.
In the early days of the pulverized coal combustion system development,
flames were unstable and sensitive to changes in firing rate. Flameouts occurred
frequently and difficulties were experienced with flame impingement on the side
and rear walls of the furnace. Unburned combustibles (carbon loss) problems,
associated with mill grinding capabilities, were also of serious concern.
Therefore,^ there was legitimate concern on the part of the utility industry that
low excess'air and/or staged firing of pulverized coal In utility boilers could
result in flame control problems. However, the accumulated experience of the
Exxon Research field studies of combustion modifications for NOX emission control
on over 30 pulverized coal fired utility boilers shows that relatively few flame
problems had been encountered. Flame instability or impingement problems have
not manifested themselves in these studies, and increases in unburned combustibles,
as discussed previously, have been found to be minor, resulting in negligible
p-ffects on boiler efficiency.
3.5 Corrosion Probe Measurements —
Nitrogen oxide emissions are limited most effectively in utility boilers by
staging the combustion process and maintaining the overall excess air level at
minimum values consistent with safe, efficient operation. Staging the firing
pattern consists of first, burning the fuel In the lower regions of the furnace
under fuel-rich or (substolchlometric air supply) conditions, followed by
second stage air addition above the primary combustion zone to complete burnout
of the remaining combustibles, Furnace tube wall corrosion potentially could
occur or be aggravated by operating a utility boiler at "low NOX" firing condl~
tions. Even though overall excess air is maintained at a reasonable level under
"low NOX" operation, conditions in the first stage combustion zone may approach
levels as low as 80% of stolchlometric. Atmospheres at the furnace sidewalls
where corrosion could take place under these conditions, might well be even more
reducing. For normal firing the atmosphere-in the furnace is oxidizing.
However, under "low NOX" staged firing conditions using upper burner rows as
overflre air ports, only the top area of the furnace is under oxidizing
conditions while the lower region (at the middle and bottom burners) is now in
a reducing atmosphere, and it is this condition which may lead to potentially
aggravated furnace tube external corrosion problems.
To gain insight into this potential problem area, corrosion probes have been
used by Exxon Research in two EPA sponsored programs, Contracts No. 68-02-0227,
(2) and 68-02-1415 (3), respectively. As discussed elswehere (2) (3), the
design of the corrosion probes was based on information supplied by Combustion
18
-------
Engineering, with appropriate modifications for this work. Essentially, the
design consists of a "pipe within a pipe", where the cooling air from the plant
air supply is admitted to the ring-shaped coupons exposed to furnace atmospheres
at one end of the probe, through a 19 mm (3/4 inch) stainless steel tube roughly
centered inside of the coupons, The amount of cooling air is automatically
controlled to maintain the desired set-point temperature of the coupons. The
cooling air supply tube is axially adjustable with respect to the corrosion
coupons, so that the temperatures of coupons can be balanced. Because the
cooling air returns along the 63.5 mm (2-1/2 inch) extension pipe and discharges
outside of the furnace, the cooling air and the furnace atmosphere do not mix at
the coupon location.
The approach used for measuring corrosion rates in the initial program was
to expose corrosion coupons Installed on the end of probes inserted into available
openings located near "vulnerable" areas of the furnace under both baseline and
low NOX firing conditions. Coupons were fabricated of SA 192 carbon steel, the
same material as that used for furnace wall tubes. Exposure of the coupons for
300 hours at elevated temperatures of 742 K (875*F) (higher than normal furnace
tube wall temperature of about 489 K (600°F) was chosen in order to deliberately
accelerate corrosion so that measurable values could be obtained. Coupons were
also mildly acid pickled to remove the existing oxide coating prior to exposure to
eliminate differences potentially caused by surface conditions. The conclusion
of these earlier corrosion probing tests was that no major differences in
corrosion rates could be found between coupons exposed to low NOX firing condi-
tions, compared to coupons exposed under normal boiler operating conditions.
Coupon corrosion rates were, however, considerably higher under baseline and
low NOX conditions than those corresponding to normal furnace tube wastage rates,
because of the accelerated nature of these corrosion probing tests.
Significant changes were made in the conditions for measuring corrosion rates
in subsequent field tests to better relate measurements obtained with corrosion
probes to actual furnace waterwall corrosion. The approach was similar to the
earlier work, but with several important differences. First, corrosion coupons,
which were all fabricated and machined in the same manner, were no longer mildly
acid pickled but Instead, were dipped in acetone, and air dried prior to weighing
to removal any oil deposited during machining. Second, coupon temperatures were
controlled at temperatures approximating those of the furnace waterwall tubes,
603-658 K (625-725°F) to more closely approximate actual furnace conditions.
Third, three coupons were installed on each probe, to increase the amount of
data obtained compared with only two coupons per probe in the prior program.
Time of exposure (300 hours) was held the same so that the results of the cor-
rosion probing runs could be compared to the earlier work. Other test conditions
were also kept the same, I.e., probes were inserted through openings in the
furnace wall as close as possible to vulnerable furnace areas, the analytical
procedures used were also the same in both programs, etc. Thus, each coupon was
visually Inspected after exposure and was photographed to record its appearance.
Scale was then removed from the outside diameter surfaces by dry honing of the
inside diameter surface scale after which the coupon was reweighed to determine
the weight loss from the Inside surfaces. Corrosion rates were then calculated
as the loss in mils per year (m/yr) using the weight loss data, the combined
exposure coupon areas, the metal and scale and densities, and the exposure time.
19
-------
Using this approach, the coupon corrosion rates are considerably lower
than those measured under accelerated conditions in the initial program. The
lower and more consistent coupon corrosion rates- measured in the latter program
reflect the changes made in test procedures to more closely approximate actual
furnace wall tube conditions. However, the rates are still an order of magnitude
greater than the 1 to 3 mils per year corrosion rates that are expected for the
wastage of actual furnace tubes under normal firing conditions. Therefore,
the corrosion probing results are viewed as only a relative measure of corrosion
tendency under baseline and low NOX firing conditions. It was concluded from
these studies that only long term corrosion measurements of actual furnace tube
wastage could answer the question of the magnitude of corrosion rate increase
caused by staged firing of utility boilers with pulverized coal. However,
corrosion probe data and techniques may still be useful in the future if reliable
correlations with actual tube wastage can be established.
20
-------
SECTION 4
LONG TESM CORROSION TESTS
During the course of the corrosion probe investigations conducted by
Exxon Research tinder EPA sponsorship, it became apparent that data obtained
with probes could not provide an unequivocal resolution to the question
whether external furnace tube corrosion might increase with "low NOX" operation
on utility boilers fired with pulverized coal. Even though "low ROX" combus-
tion conditions did not produce any major increases in corrosion rates as
measured by corrosion probes, it was not only difficult to relate these data
to actual tube wastage rates, but one could not rely on such information for
boiler design and operation. From the power generation industry standpoint
it is imperative that this question be resolved and, if corrosion is indeed
a problem, engineering solutions are required for the application of "low
NOX" operation to coal fired boilers. A comprehensive corrosion investigation
program, therefore, was undertaken in an attempt to settle this issue conclu-
sively. Long term tests sponsored by the EPA have been undertaken in a joint
cooperative venture by Exxon Research and Engineering Company, Foster Wheeler
Energy Corporation, and the Gulf Power Company. These tests have been under-
way since June of 1976 on the No. 7 boiler at the Crist Station of Gulf Power
Company in Pensacola, Florida. Three types of corrosion rate determinations
are being employed:
1. Measurements on specially installed furnace tube panel test
specimens.
2. Ultrasonic mapping of the thickness of the furnace tubes and
test panels.
3. Corrosion probes exposed for varying times.
(5)
Details of these methods are discussed elsewhere v—'. The highlights of
the long term corrosion test program are summarized below.
4.1 Test Panel Design. Installation and Measurements
Boiler manufacturers' experience indicates that external corrosion of
furnace tubes occurs in areas in a largely unpredictable and random pattern,
although the general problem areas are at the burner elevations of the furnace
sidewalls. This presents a major problem in determining where to place the
furnace corrosion panels to ensure that they are in areas where corrosion
occurs. There is no ideal solution to this problemt and from an economic
standpoint the number of panels used must, of necessity, be limited.
21
-------
Eight corrosion test panels were installed In the No. 7 boiler at the
Crist power station during May, 1976 at strategic locations believed to be
most helpful in defining potential corrosion effects. Windows were cut out of
the furnace walls the size of the corrosion panels and the sections removed were
retained for future laboratory inspection. Panel and corrosion probe locations
are shown in Figure 11. It can be seen that seven of the panels were installed
in the left furnace side wall and one in the right wall. The reason for
installing most of the panels in the one wall is to provide maximum areas
of exposure to corrosion. Since the corrosion probe areas of the sidewalls
are normally at the midpoint at the burner elevations, four of the panels
(No. 3, 4, 5 and 6) were installed in this area. Three of these panels
(No. 3, 4, and 5) are at the middle burner row elevation, and panel No. 6 is at
the top burner elevation. Note that panel Nos. 4 and 6 are located in the
middle of the sidewall where the most severe corrosion can be anticipated.
Panel Nos. 1 and 2 are located in the hopper area where corrosion on other
boilers occasionally has been experienced. Panels 7 and 8 in the left hand
and right hand sidewalls, respectively, installed in the upper reaches of the
furnace (above the burners), are expected to experience lower corrosion rates
(oxidizing atmosphere) and will serve as "control" panels for comparison pur-
poses. This scheme of panel arrangement was conceived to provide the maximum
amount of data within the constraints of reasonable level of effort.
Each test panel is five (5) tubes wide by 1.5 meters (5 ft.) in length.
Tubes 1, 3 and 5 were made of the same low carbon steel material as the furnace
tubes; SA-210 grade A-l. Tubes 2 and 4 are SA-213 grade 1-2, a higher grade
alloy carbon steel expected to have greater resistance to corrosion than normal
furnace tube -material. The use of two materials will provide additional use-
ful information on rates of corrosion which would not be available if only one
material were used.
Prior to installation, the panels were characterized in the laboratory
after fabrication. Thickness measurements were made ultrasonically at 3 inch
intervals on the side of the panel exposed to the furnace and at 6 inch intervals
on the opposite side, for control purposes. In addition, points near the end
of the tubes were measured independently by an accurate micrometer on both sides
of the panel, so that an independent measure of precision could be developed from
paired mechanical vs. electronic measurements.
Examination of the test panels while still in place in the furnace is al-
most identical to that for wall tubes. Samples of corrosion products will be
removed periodically from the tube surfaces by chipping, and the extent of
metal loss is being determined by ultrasonic measurement. The major advantage
of test panels is that they can be removed from the furnace and sectioned to
give precise indications of metal loss and of the composition of corrodents.
22
-------
Metallographic examination of the panel tube metal can provide important
information. Such methods, by examining the grain structure after exposure,
give a good indication of the mode by which corrosion is occurring and can
detect incipient metal attack along grain boundaries not detectable otherwise.
Sulfldation, in particular, can be shown by metallographic examination. Also,
if the metal should have been overheated at any stage, the extent of spheroi-
dization will provide a rough measure of this change in metal structure.
Two steps have been taken to obtain information characterizing corrosion
panel exposure conditions. These include (1) measurement of corrosion panel
tube metal temperatures and (2) analysis of the gaseous atmospheres in the
location of the panels. Three thermocouples were installed in each panel (in
the 1st, 3rd, and 5th tubes). The thermocouples are designed to measure the
temperature of the tube as closely as possible to the surface exposed to the
flame in the furnace. The objective is to determine whether any unusual tem-
perature conditions may exist which would contribute to corrosion. Thirty gas
taps were installed throughout the furnace area and in the front, rear and
side walls including taps in the corrosion panels. Periodic surveys and analy-
ses of the gases taken from these taps at different loads and combustion
conditions will provide information to characterize the atmosphere in the
furnace, especially under "low NO " operation.
A
4.2 Ultrasonic Tube Thickness Measurements
In order to determine the .extent of potential furnace tube corrosion the
tubes must be measured before and after prescribed periods of time. In the
recent past highly accurate ultrasonic thickness measuring equipment was not
available. Accuracy within 5 mils only was possible (5). Today, Instruments
are available capable of accuracy to tenths of a mil. In the program being
conducted on the No. 7 boiler at the Crist Station, two Krautkramer-Bronson
CL 204 ultrasonic instrument gages, the most accurate currently available, are
being used in making the measurements. Accuracy of these instruments is 0.01%
of an inch, thus assuring that measurements to the nearest tenth of a mil are
possible. This degree of accuracy is an essential feature for determining
corrosion rates in a reliable fashion. Procedures and methods have been up-
graded to achieve the highest degree of accuracy possible as experience with
the Instrument has been gained.
The condition of the tubes is a complicating factor requiring sand blast-
ing, or some other suitable form of slag, deposit, or scale removal before
reliable measurements can be made. Then the question remains after the tubes
have been cleaned bare, are they now more susceptible to corrosion, and there-
fore no longer representative since oxide scales, coatings or protection have
been removed? It is also difficult to relocate a spot previously measured for
rechecking, and measurements at other places on the tube may be unreliable for
comparison due to variations in tube manufacturing tolerances. Untortunateiy,
sandblasting, wire brushing or other means of cleaning to remove ash or slag
coatings from the tubes down to the base metal, necessary for accurate thickness
measurement, undoubtedly will remove any protective coating making the tube
more vulnerable to corrosion at this spot.
23
-------
To statistically eliminate bias in the results due to the greater corro-
sion vulnerability of tube areas previously cleaned and measured, a large
number of measurements are taken at a given elevation at the beginning of the
test in a random pattern, on some tubes exactly at the elevation, and on others
slightly above and below the exact elevation. At the conclusion of the test,
many measurements are repeated and center measurements are made exactly at
the elevation on those tubes previously measured above and below the elevation.
1 Ultrasonic tube thickness measurements were made at six different furnace
elevations in the No. 7 boiler at the Crist station in May 1976, after the cor-
rosion test panels were installed. A second set of measurements were taken at
the same elevations in October of 1976, after the boiler had been operating for
5 months under baseline operating conditions. These data will provide corrosion
rate information for future comparison with similar data presently being obtained
under "low NOX" firing conditions. Figure 11, showing corrosion panel locations,
also shows the six furnace elevations where ultrasonic tube thickness measure-
ments were made. The only elevation on Figure 11 at which measurements were not
made is elevation 30.4 m (99'-8"), the junction of the furnace with the hopper.
Elevations 28.4 m (93'), 34.4 m (112'-8"), 36.9 m (121'-2n), and 39.5 m
'(129'-8") include all elevations in the furnace area where corrosion may be
anticipated. Measurements at elevations 42.1 m (138'-2") and 48.1 m (157f-8")
are in the oxidizing zone where corrosion is least likely to occur. These
measurements will provide "control" information.
4.3 Corrosion Probe Measurements
Corrosion probes (as described in Section 3) are again being employed on
the No. 7 boiler at Crist Station to obtain corrosion rate information concur-
rently with corrosion rate data taken by corrosion panels and ultrasonic tube
measurements. The objective of these measurements is to establish correlation of
actual tube wastage experience with corrosion probe data, so that confidence in
the reliability of the less expensive corrosion probe methods may be achieved.
Several changes, however, have been incorporated in the corrosion probe proce-
dures to improve the program and to provide a greater amount of useful Infor-
mation. First, probes are being exposed for 30, 300, and 1000 hours, both-under
baseline and "low NOX" firing conditions. The 30 hour exposure will provide
data on initial corrosion and the 300 and 1000 hour information will show the
effect of corrosion with time. Also, correlation with previous investigations
on other boilers at 300 hour exposure should be possible. Second, special
openings were incorporated in the furnace corrosion panels to accommodate the
corrosion probes so that, for the first time, corrosion probes could be located
in areas of greatest anticipated corrosion and in "control" areas.
4.4 Status of the Long Term Corrosion Tests
, As indicated above, eight corrosion panels were installed in the Mo. 7
pulverized coal fired boiler at the Gulf Power Company's Crist power generating
station in Pensacola, Florida, during May 1976. New openings were also provided
in the furnace sidawallsduring this outage for the more advantageous location
of corrosion probes in areas vulnerable to corrosion, to better define the cor-
rosion problem using this method of approach. At the beginning'of the long
24
-------
term corrosion test (May 1976), over 1000 ultrasonic tube thickness measure-
ments were made on the No. 7 unit (including measurements of corrosion panel
tubes) at the six elevations mentioned earlier. The unit was then run over
the summer during the peak load demand period under baseline operating con-
ditions for five months, until October 1976, when the boiler was taken out
of service for a scheduled maintenance outage. Ultrasonic tube measurements
were again made during the latter outage in accordance with the statistical
plan developed for these tests. These results obtained will be subjected to
detailed statistical analysis and compared to those that will be forthcoming
from the low NOX operation.
The No. 7 Crist boiler is now being operated under low NOX conditions,
and will continue this type of operation until the next scheduled maintenance
outage in September/October 1977. At that time, the corrosion panels will
be removed from the boiler (after about 1 year of low NOX operation) and
returned to the laboratory for re-measurement of the tubes and metallographic
examination of the specimens. Ultrasonic tube thickness measurements will also
be made at the previously prescribed elevations. These data, when compared
to the measurements made in October 1976, should provide definitive information
on external furnace tube corrosion experienced under "low NOx" operation. A
comparison can then be made to actual wastage experienced under baseline con-
ditions which should prove whether or not "low NOx" firing leads to increased
corrosion of the furnace tubes in this pulverized coal fired boiler.
Corrosion rate data have been taken under the baseline operating period and
are currently being obtained under "low NOx" firing conditions at exposures of
30, 300 and 1000 hours for correlation with actual tube wastage rates determined
on corrosion panels and by actual furnace tube measurements.
25
-------
REFERENCES
1. W. Bartok, A. R. Crawford and G. J. Piegari, "Systematic Field Study of
NO Control Methods for Utility Boilers," Esso Research and Engineering
Company Final Report No. GRU.4G.N0.71, Contract No. CPA 70-90, December
1971 (NTIS No. PB 210 739).
2. A. R. Crawford, E. H. Manny and W. Bartok, "Field Testing: Application of
Combustion Modifications to Control NOX Emissions from Utility Boilers",
EPA-650/2-74-066, June 1974.
3. A. R. Crawford, E. H. Manny, M. W. Gregory and W. Bartok, "The Effect of
Combustion Modification on Pollutants and Equipment Performance on Power
Generation Equipment", Proceedings of the Stationary Source Combustion
Symposium, Volume III, p. IV-3, EPA-600/2/76-152c, June 1976.
4. H. Shaw, "The Effect of Water on Nitric Oxide Production in Gas Turbine
Combustors," ASME Paper No. 75-GT-70, Houston, March 2-6, 1975.
5. G. A. Hollinden, J. R. Crooks, N. D. Moore, R. L. Zielke and C. Gottschalk,
"Control of NOX Formation in Wall Coal-Fired Boilers", Proceedings of the
Stationary Source Combustion Symposium, Volume II, p. 111-31, EPA-600/2-76-
152b, June 1976.
6. E. H. Manny, et al., "Studies of Waterwall Corrosion with Staged Combustion
of Coal", Presented at the International Conference on Corrosion and
Deposits from Impurities in .Combustion Gases; ASME/Engineering Foundation
Conferences, New England College; Henniker, New Hampshire, June 26-
July 1, 1977.
26
-------
*
H
CO OS
ss
M
H O
co pa
Ba
§s
M fc
C/5
W rJ
(•£{ O
Q
W S5
EH
U 5
M
H
_^
4-1 rH
•x "5
.1
i«
rH
0
B>5
r-(
tj < 0)
01
C W
O flj
rH
M e 3
O O
4J £ -H
5 ,J 1J
0 (2
e >, •-.
O O w • p
"H C Cl X H
-d *J cy o> i-l BO
0-0 -rt K
rHU-l -U O rH
s
•H i-l
h •»
•rl C
[£< Q
U
o
55
4J
n
e-i
^,
^J
•H
(H
•H
4J
O 00 r-4 O 00 rH
CM O 00 rH
in CNJ i/> r-* CM
in co o m oo o-.
vO vD vO vO \O vO
CM CM CM CM CM CM
^H "r^ ^H O O O
?H rH ^H SE£ JZ3 rZ
(U 3
S (S rt ,3 ^ »3
£M -^ ^O rH rO 1^
v ^
He
«
* rH
* •
n- &
01 01
s 8 3
OT 01 O
1 1
1 1
1 1
1 1
1 1
1 1
1 1
J 1
1 1
fr^ f?}
o o
CO CO
ro en
l-l rH
VO l~-
CM CM
O O
VO VD
0 0
O O
00 O
OO 00
CM CM
d)
$ o"
-i g
41
CO 3
M S>
Q U
v_.
^
4e
^f
^ l/l
• *
1 5
ss c
\M> (U ^4
IJ 4)
o s -5
W S -H
to cu n
O> «n (3 '
CO
0)
u
•H
oi
•o
J^
o
4J
U
rH
rH
O
CJ
O
g
01
4J
CL.
3
w
111
iH
O
H
01
4J
s
•H
eg
4-1 00
eg C
00 h
C i-l
•H (L.
iH
tlH
ed
VI U
*
* *
tn>
•5
^1
(b
•H
O
*
*
•K
27
-------
en
H
6-8 M
** r^t
55 ^2
O p>4
l_l ^3
C_| £)
& CJ
M OT
tf oS
H 0
H p
n gj
w
N O
H M
W 0
^JJ p£J
u u
H 0
53 o
0U i™4
-1
M W
H
H M
M ffi
H £
^— •
•
0
*
rl
4)
•rl
O
PQ
i
13
0
•H
4J
CD
•U
tfl
ff
0)
(1)
w
*
„. J
f™"l
*
o
^
(U
rH
•H
T
a
O
•H
4J
Q
•*
l/^
cn
rH
CM
0)
pjj
01
N
iH
CO
O O O if) r~~ co
• •••••
O CM CM CM CN O
00 r4
^ c^ r*^ co co CM
• •*•»•
<}- O H i-l H 00
00
o r*^ r^- *3" \o ^o
• •••••
O O^ CM CM O «*
00
f- o% in CM o r~.
O CM O CM H CM
0\
VO CM CM CO CM IO
00 CO CM H CM CM
00
o> o« m <* H r»
rH CM rH O O CM
00 rH
CO O^ *d" O ^n ON
00 sr CM CM rH O
00
oo r*1* o^ oo ON ^ft
• •«•••
^- CM CM f> rH CO
00
e S S S 9 g
• •••••
'CM CM rH rH O O
A V
CM
rH ^N
in
0
0
0)
o
-------
' 1
1
1
CO
H
CO
s
r*
U
i
i
@
M
O
rn
1
•
I
;
PM
W
O
M
1
£
•
l_4
H
V
s
3
H
1
—
^
01
•H
O
pa
c
o
C
o
•Ct
u
fr£
r-l 4J
O £ St
" < ^
Ctt
10
C
O IN
i-i O
01
m s-s
•H co
i
i
-
I
4->
HJ
H
ao
•H
0)
rH
•H
O
CP
•
6
ig
0.
a
S
N
U
G
•H
U
•H
SM
w
01
I-I
3
•H
VJ
O
P-.
O St «™^ i— * fs. QO
O O O O O Os
*~* PO csj o rx oo
cs — H CN m ^H m
^NJ r** co o r«- fo o
oil in M o m CM o
3
O O 4-1
PM X w
29
-------
1400
1200
to
£ 1000
CD
>
D;
Q
CM
O
-, c
0s**
en
X
o
CL.
£L
800
600
400
200
/
FULL LOAD (283-296 MWe)
Ijj)-} (232-237 MWe)
(210 MWe)
(150-165 H
&
(150-163
MWe) 1
(151-155 MWe)
S"S
I
Firing
Pattern
S^-Normal
S2~Biased
S3-Staged
$4-Top Row
Off
Sec. Air Registers
Open
O
A
<
O
Partly
.Closed
V
>
Closed
ID>
3456
AVERAGE % OXYGEN IN FLUE GAS
Figure 1. NOX emissions measured for
Mercer No. 1 boiler.
30
-------
300
250
200
—
<7i
<
m
>
tz.
o
CM
O
§ 15°
X
100
50
NORMAL FIRING - 285 MWe
^@
STAGED FIRING - II - 286 MWe
NORMAL FIRING -
210 MWe
STAGED FIRING - II, III, IV -
285 MWe
STAGED FIRING - V, VI
145 MWe
I
NORMAL FIRING
STAGED FIRING II - FGR HIGH
STAGED FIRING II - FGR LOV
\/ STAGED FIRING III - FGR LOW
L>_ STAGED FIRING IV - FGR LOW
[J STAGED FIRING V - FGR LOW
STAGED FIRING VI - FGR LOW
_ I J
4 5
% OXYGEN IN FLUE GAS
Figure 2. Effect of operating variables on
NOX emissions for oil fired boilers
(Sewaren boiler No. 5).
31
-------
120-
100-
l/)
CD
CM
O
*•£
o"-.
IT)
i—1
X
CL
a.
20
0 .5%
0% H20 Injected
0.75-0.80% H20
1.5%
2.0%
2.3% H20
I
10
20 30 40
GROSS LOAD - MWe
50
60
Figure 3. Effect of water injection on NOX
emissions from oil fired gas turbine
(T. H. Wharton gas turbine No. 42).
32
-------
120
100
CO
CQ
C\J
O
X
O
80
&> 60
40
20
I
0% H00
1.9% H20
0
10
20 30 40
GROSS LOAD - MWe
50
60
Figure A. Effect of water injection on NOX
emissions from gas fired turbine
(T. H. Wharton unit No. 43, gas fired)
33
-------
PPM NO =70.3e-1-007*H2°
A
I
0.4
0.8 1.2 1.6 2.0
% WATER INJECTED (% OF COMBUSTION AIR)
2.4
Figure 5. Correlation of NOX emissions with water injection
rate for natural gas fired gas turbine
(Houston L&P Wharton No. 43 unit).
34
-------
1200
1100 -
1000
(
<
CD
or
900
800
X
O
Q_
DL
700
600
500
1 yI I
1l)'Sx-(460-520 MW)
82) S2-(430-450 MW)
S3-(410-440 MW)
400
I
I
345
% OXYGEN IN FLUE GAS
Figure 6. NOX emissions measured for
Crist No. 7 boiler (A duct)
35
-------
700i
600
500
IT)
<
CO
CM
O
,0
O*"*
CO
X
o
D-
CL
S, -NORMAL FIRING
1 (178-227 MWe}
, - D-MILL LEAN
i (166-210 MWeTl
400
300
200
' NORMAL FIRING
- D MILL LEAN
100|
I
2345
AVERAGE % OXYGEN IN FLUE GAS
Figure 7. Effect of excess air and staging
on NOX emissions
(Cooper No. 2 boiler - 166 to 227 MWe).
36
-------
700
600
500
to
DO
CM
O
a***-
en
400
x 300
o_
Q.
200
100
0
_ S15 - 150 MWe
- 130 TO 147 MWe
,Si - 155 MWe
Oh
VS
>
12
13
- NORMAL FIRING
- A MILL OFF
- D MILL LEAN (B MILL OFF)
- D MILL LEAN (F MILL OFF)
- D MILL ON AIR ONLY
- D MILL LEAN, (MILL AIR ONLY)
, - 125
D MWe
14 - 125-
TO 132 MWe
MILL/BURNER CONFIGURATION
Q &
^-~-
! 3 4 5 6
AVERAGE % OXYGEN IN FLUE GAS
Figure 8. Effect of excess air and staging
on NO emissions
(Cooper No. 2 boiler - 123 to li>5 MWe) .
8
37
-------
800 -
5.7% 0
Q.
Q.
300
200
100
5.7% 02
.2% 02 /ULL LOAD, HIGH EXCESS AIR
REDUCED LOA..I, 5.0% 02
NORMAL EXCESS
AIR
FULL LOAD, NORMAL EXCESS AIR
.5% 02
I
J_
20 40 60 80 100
OVERFIRE AIR DAMPERS - % OPEN
Figure 9. Effect of secondary air addition on NOX emissions
(Comanche No. 2 boiler - 275 to 355 MWe).
38
-------
800-
700-
600
CO
to
00
> 500
a:
Q
CM
o
400
X
O
a.
a.
300
200
100
Of—
3.0
50%
I I
OVERFIRE AIR
DAMPERS -
20% OPEN
3.5 4.0 4.5 5.0 5.5 6.0
AVERAGE % OXYGEN IN FLUE GAS
6.5
7.0
Figure 10. Effect of excess air in full
load NOX emissions
(Comanche No. 2 boiler - at
full load of operation)„
-------
157'8'
138'2'
129'8'
121'2"-
112'8"
99'8'
93'0'
®
Side Wall
Left (Right)
7(8)
So
It Corrosion Test Panels
-------
ANALYSIS OF NOX CONTROL IN STATIONARY SOURCES
By:
0. W. Dykema
The Aerospace Corporation
El Segundo, CA 90009
41
-------
-------
ABSTRACT
The subject program is a three-year effort to analyze NO control in utility
A
boilers by combustion modifications. Results of a previous Aerospace Corporation study
on this subject and those of the first year of the current study were presented two years
ago at the first EPA Stationary Source Combustion Symposium. Those studies concluded
that there appeared to be no inherent limit to reduction in NO by combustion
A
modification in utility boilers except those that may result from other undesirable side
effects. Among those discussed were: (a) excessive water-wall erosion/corrosion, (b)
excessive losses in plant efficiency, and (c) combustion instability. The first of these is
being investigated experimentally by several agencies. The1 latter two were addressed
during the past year in the Aerospace program and are the subject of this paper.
In general, the analysis of a large sample of data from tests on natural gas-and
oil-fired utility boilers showed no significant effects on plant efficiency due to staged
combustion (or burners-out-of-service) or the use of NO ports. The data available to
this study were not adequate to evaluate this conclusion with respect to coal-fired
boilers or for any other combustion modification technique (specifically, water injection
or combustion air temperature control).
The analysis of combustion instability in utility boilers, however, did show that
operating the burners fuel-rich (as in staged combustion) does tend to create more
unstable combustion, specifically in air-side feed system-coupled modes of combustion
instability. The fuel-rich burner operating conditions in staged combustion increase the
dynamic response, or gain, of combustion in the furnace and the air flow rate through the
burner. A method of analysis was developed which shows that, with proper design, these
modes can be stabilized even with very fuel-rich burner operation.
43
-------
NOMENCLATURE
C
Fa(i) (i=
H
F(r)
L
Lb
M
P
R
- Cross-sectional flow area, for air, in a burner.
- Capacitance within a burner.
- Arbitrary functions used to designate the furnace pseudo-
acoustics, in five of the six directions.
- A function describing changes in local furnace pressures
resulting from changes in the local combustion air/fuel ratio
(through changes in the combustion temperature and molecular
weight of the combustion products). *
- Constant describing damping during acoustic wave travel and
the efficiency of wave reflection at solid boundaries.
- Inertance of the air within a burner.
- Length of a burner, in the flow direction.
- Molecular weight.
- Pressure; furnace (Pf); furnace response (PfQ); furnace input, or
driving pressure (P*.); constant windbox pressure (P .).
- Flow resistance; inlet to a burner (R.); burner exit region (due
to the presence of flame) (R^); linearized (R-A R^); linearized
resistance to the flue gas flow from one burner in leaving the
radiant section (exit) of the boiler (R ); steady state resistance
C
in the burner exit region due to the presence of a flame (R,).
O
- The LaPlace Operator.
44
-------
T T T T
ll> 12> 3' 4'
T T
la' I3a
T
a
e
S
\
j
r
*
Time constants in the expression for the dynamic response of a
burner (equation (1)). "l-^ , ,-.
Temperature i'v£
The acoustic velocity urine air within a burner.
Designates an exponential?term (natural).
The acceleration of gravity.
An index. ;
SUPERSCRIPT
- The weight air/fuel ratio.
- Weight flow rate; air at the burner inlet (w j); air leaving the
burner (w.); fuel leaving the burner (constant) (w ^b); delayed
burner flow rate (*bd); returning acoustic flow rates (war).
- Weight of gases stored in the control volume in the furnace.
- Time delay; from the burner exit to the region of concentrated
combustion (combustion~time delay) at the steady-state burner
flow velocity ( r ); for acoustic wave travel from a burner exit
to and from the exit from the radiant section of the furnace (T
_); for acoustic wave travel from a burner exit to and from a
6
solid boundary, in the (i) direction ( T.).
- Frequency, in radians per second.
- Time-invariant quantities.
45
-------
i<
/•
S
-------
SECTION i
INTRODUCTION
A paper was presented by the author two years ago at the first EPA Stationary
Source Combustion Symposium that described Aerospace efforts to develop useful
guidelines for the control of NO in natural gas-, oil-, and coal-fired utility boilers by
X
combustion modifications (1). Fundamental to that effort was the use of large samples
of NO and operational data from full-scale operating utility boilers. In general, results
X
indicated no limits on NO reduction that could be considered inherent in the
X
combustion modification techniques per se. Such limits would very likely result from the
appearance of other undesirable side effects such as significant losses in plant efficiency,
combustion instability, or water-wall erosion/corrosion. A number of activities/agencies
have been examining the latter possibility. In the second year of an EPA grant, then,
Aerospace undertook a study of the effects of combustion modifications made for the
purpose of NO reduction on plant efficiency and combustion stability (2).
Overall plant efficiency data were available from the natural gas- and oil-fired
utility boiler samples of previous NO control studies (3). These data showed efficiency
X
losses of as much as 7 to 10 percent that, at least superficially, appeared to be related to
lower NO levels. These would certainly be significant limits to NO reduction by
A X
combustion modification if such losses could be attributed to the NO control technique.
A.
Briefly, the study of plant efficiency showed that, at least within the
limitations of the combustion modifications represented in available data sample, the
NO control techniques caused no discernible changes in overall plant efficiency. The
task to evaluate combustion instability, however, did show that, without proper design
consideration, this could represent a significant limit on NO reduction by combustion
modification. A method of analysis of air-side feed system coupled modes of combustion
instability in utility boilers was developed showing that fuel-rich burner operation, as in
47
-------
the staged combustion technique, can lead to violent combustion instability. The analysis
was verified by comparison with one (the only available) case of instability in a natural
gas-fired utility boiler. The writer is aware of at least one case where combustion
instability is currently limiting NO reduction in an oil-fired utility boiler to levels higher
X
than current regulation.
; Because the plant efficiency study results are rather straightforward whereas
the combustion instability study shows a potential problem, and offers a solution, plant
efficiency is discussed only briefly here and further disussion largely concerns the stability
study.
Work has continued, at Aerospace, under EPA funding, on NO control methods
beyond the efficiency and stability studies. This effort has concentrated largely on
simplifying and improving the NO control model and in demonstrating application of
A
results to utility boiler design for minimum NO within the bounds of high plant
efficiency, combustion stability and low emissions of other air pollutants. Because this
work is as yet incomplete, discussion of those results is also limited.
SECTION 2
CURRENT WORK
Studies just completed on NO control in utility boilers, under EPA funding,
A
largely concerned simplification and improvements in the NO control analysis technique
A
arid demonstration of the use of the resulting guidelines. Major simplifications result from
the observation, in previous Aerospace and other studies, that NO reduction by the
A
staged combustion technique is nearly always greatest when the air-only burners and NO
ports are all located above all of the active burners. For this reason, all burner
configurations involving air-only burners below the top level of active burners (and the
resulting enormous complexity of multiple and varied mixing zones throughout the
furnace) were eliminated from further consideration. With this limitation the combustion
product composition could be considered constant throughout the active burner region.
This limitation so greatly simplifies the active burner region that some of the
simplifications and limitations of the previous NO control analyses could then be
improved. It is now possible to directly integrate the NO formation rate equation along
A
48
-------
flow paths through the active burner region and, with a small number of steps, through the
/.one where the remaining air is mixed in with the fuel-rich products from the active
burner region. That limitation also allows better descriptions of combustion product
cooling by radiation (proportional to the fourth power of the product temperatures) and
finite gasification and initial mixing rates. While regression analysis of large samples of
data from full-scale utility boilers is still used to quantify the resulting analysis,
preliminary results indicate that NO levels calculated directly from the analysis are now
A.
much closer to measured levels than was previously possible. As of this writing, work in
this area is continuing. Results will be published as an EPA report.
SECTION 3
PLANT EFFICIENCY
Briefly, the study of the effects of combustion modifications made for the
purpose of NO reduction showed no distinguishable efficiency losses that could be
A
attributed to the use of staged combustion (burners-out-of-service and/or NO ports). All
J\
of the efficiency losses that could be correlated in the data (up to 8 percent) appeared to
result from load reduction. Although it is clear that NO does decrease with load (at
X
least in the boilers in the data sample), load reduction is not considered a combustion
modification for the purpose of NO control.
X
It is also clear, from the available data, that the combustion air temperature
decreases with load. The observed NO reduction with load is very likely caused by the
A
reduction in the combustion air temperatures. The loss in plant efficiency with load,
however, appears to be primarily related to off-design operation of the steam turbines at
reduced load. Thus the observed plant efficiency losses may not be related at all to the
phenomena that cause the NO reduction. Unfortunately, there was insufficient data
A.
available to this study to directly evaluate the effects of combustion air temperature (or
other combustion modification techniques such as water spray or flue gas dilution in the
combustion air) on efficiency. Since these are considered potential NO control
A
techniques, their effects on plant efficiency still need to be evaluated. Also, the data
available to the previous study of coal-fired utility boilers (4) was not adequate to develop
plant efficiency data.
49
-------
• SECTION 4
COMBUSTION STABILITY
Models and analysis techniques of feed system coupled modes of instability
developed in the rocket industry were used as a basis for the development of a method of
analysis for such feedback systems coupled to the air flow system in utility boilers. A
major modification of those models and techniques was necessary to adequately describe
the coupling between burner flow rate perturbations and resonances in the three
i
coordinates of the furnace cavity. This modification complicated the analysis but it was
!
shown that in the limiting case (primarily, in the case where the boiler cavity dimensions
are very small) the analysis developed here becomes identical with that long used in the
rocket industry.
j
Basically, the model can be described as follows. The boiler windbox is taken as
a large, constant pressure plenum from which combustion air enters the burners. The air
in the burners has compressibility and inertia. Resistance to air flow through the burner
is in two parts, a constant resistance at the burner inlet (through the air registers) and a
variable resistance near the burner exit that is a function of the degree of initial
combustion within the burner.
In conventional utility boilers the pressure drop across the burners, from the
windbox to the furnace, is very small, measured in inches of water. As a result, this
pressure drop, and the resulting air flow rates, are quite sensitive to variations in furnace
pressures at the burner exit. Small perturbations in furnace pressure at the burner exit
cause large perturbations in the air flow rate through the burner. However, the pressure
drop across the fuel injectors (or orifices) is usually quite large, measured in pounds per
square inch (psi). Perturbations in furnace pressure, therefore, have small effects on fuel
flow rates. As a result of constant fuel flow rates mixing with varying air flow rates, the
air/fuel ratio leaving the burner is also varying. Figure 1 shows a model of the dynamics
of air flow through a utility boiler.
The effect, in turn, of varying air flow rates and air/fuel ratios entering the
furnace is in two parts. The total volume flow rate perturbations begin immediately
upon leaving the burner to mix with the gases in the furnace, decelerating and generating
50
-------
acoustic waves which propagate away in all six directions. The majority of the
combustion takes place at some later time (the combustion time delay), further out in
the furnace. Although the flow velocity perturbations have been damped out, the
air/fuel ratio variations have not. Heat release rates, then, largely vary as a function
only of the varying air/fuel ratio. When the burner is operating very fuel rich (a
combustion modification for the purpose of NO control) the heat release rate varies
A
strongly with air/fuel ratio. The varying heat release rates also create acoustic waves
that propagate away in all six directions. Figure 2 shows a schematic of the dynamic
coupling between the active burner flow and furnace pressure.
After appropriate time delays for acoustic wave travel to the limits of the
furnace cavity, and reflection off of solid boundaries (at some efficiency), the waves
return (at different times) to the burner exit and add together to create the furnace
pressure variations that, in turn, cause further variations in air flow rates and air/fuel
ratios coming out of the burner. Acoustic waves that travel to the exit of the radiant
section of the boiler are partially dissipated as they cause varying rates of flow of gases
out of the radiant section into the back pass.
Figure 3 shows a block diagram of a simple feed system coupled mode of
combustion instability (called "chug") in a rocket engine (5). To apply such a model to a
utility boiler using staged combustion for NO control , necessary major modifications
A
include the effects of varying air/fuel ratio on the heat release and of acoustic
phenomena in the large furnace cavity. Direct solution of the three-dimensional acoustic
wave equation, coupled to the burner flows and including damping during wave travel and
imperfect reflections from solid boundaries, introduces unnecessary complications in the
analysis. Instead, a relatively simple set of "pseudo-acoustics" were developed that
incorporate these phenomena in an analytically manageable form. The resulting block
diagram is shown in Figure 4. The functions F (i) account for acoustic wave travel in the
fl
six Cartesian coordinate directions, reflection off solid boundaries and return to their
origin, with damping during wave travel. The model shown in Figure 4, when applied to
the conditions of a chug mode of instability in a rocket engine, reduces to that
schematically described in Figure 3.
The burner air flow response, developed from the model shown schematically in
51
-------
Figure 1, is described analytically by the expression:
b =
where:
C + L
Ril + Rtt
T,
/I
I2
LC\
)
1/2
1 + T4S + T S
1 + ?1S -•• T^ S2 + TJJ S3
1/3
(1)
(inertance)
AbLb
(capacitance)
2R. *. (inlet resistance)
dR
2R3 *b + *b | dwb (exit, flame, resistance)
Similarly, the response of the furnace pressure at the burner exit is
described by:
52
-------
w.
= H
(2)
-rS
e ,
e +
w R
s e
where:
F(r) =
dT
dr
1
M
dM
-3F
The overall, open looped response of an air-side feed system coupled mode of
combustion instability in a utility boiler, as shown in the block diagram of Figure 4, is the
product of Eqs. (1) and (2). To evaluate the frequency of this open loop the substitution
S = jw (3)
can be made. The system will be unstable, then, at those frequencies where the magni-
tude of the loop response is greater than one and the phase shift around the loop is 180
degrees (or(2m-l)7r).
The only obvious observation which can be made directly from Eq. (1) is that the
magnitude of the burner response can be reduced, and the overall loop made more stable,
by increasing the steady-state resistance to air flow through the active burners,
primarily by increasing the burner air inlet resistance, R.< . {This is called "gain
stabilization" (6).) It will be shown later that increased resistance due to the presence of
partial combustion within the burner, R.J , can have a strong destabilizing effect under
some circumstances. This is particularly true if the flame is not firmly anchored in the
burner exit and can move in and out of the burner as a result of (decreasing and
increasing) air flow rate variations. The worst case of this, of course, is periodic flame
lift-off.
53
-------
Equation (2) shows a destabilizing effect of staged combustion, through
the function F(r). This function represents the effect of air/fuel ratio variations on
furnace pressure through the combustion heat release. With active burners operating at
air/fuel ratios of the overall boiler, or near to stoichiometrie, air/fuel ratio perturbations
result in little or no pressure perturbations. The only destabilizing effect is that due to
total flow rate variations. When an active burner is operated very fuel-rich (for purposes
of NO control), however, small variations in the air/fuel ratio of the mixture leaving the
A.
burner cause large variations in the heat release in combustion (F(r) is large) and
subsequent large furnace pressure variations. In the extreme case, where the steady-
state burner air/fuel ratio is near the fuel-rich flammable limit, the flame could
alternately lift-off of, and flash back to the burner exit. In this case the function F(r) is
essentially infinite and violent instabilities should result.
Some of the more significant results of more detailed calculations of the
magnitude of the open loop response (stability) as a function of frequency are shown in
the remaining figures. Figure 5 shows a comparison of this calculation (for a full-scale,
natural gas-fired utility boiler), using Eqs. (1) and (2) of this analysis and developed from
the model shown in Figure 4, with the simple rocket engine analysis described by the
model shown in Figure 3. The agreement is good except that this analysis introduces the
effects of the furnace cavity resonances and the low frequency stabilizing effect of the
steady-state resistance of partial combustion within the burner.
t
Figure 5 also shows that the intermediate peaks in response are not at the
resonant frequencies of the furnace cavity, as might be expected, but are between these
resonances. This is because the burner air flow rate response is 180 degrees out-of-phase
with the furnace pressure at the burner exit (the minus sign in Eq. (1)) and a strong
resonance tends to damp burner flow rate oscillations.
Figure 6 shows a comparison of results of this analysis with experimental
observations from the single case of instability in a full-scale (natural gas-fired) utility
boiler available to this study. The instability frequencies prediced by this analysis are
shown by the circles drawn on the response curves (the frequencies where the open loop
phase shift is 180 degrees). Unstable operation is predicted where the magnitude of the
open loop response is greater than one at the frequencies noted by the circles. The
54
-------
response curves shown are for an active burner located at the bottom of the burner
array. Because the appropriate value of the combustion time delay is not well-known,
Figure 6 shows the response curves for each of three values of this time delay.
The degree of agreement is considered reasonably good. A strong instability is
predicted at 10 to 11 hertz and a strong instability was observed at about 12.5 hertz.
The analysis shown in Figure 6 implies a possible instability in the 27 to 35 hertz
range, but none was observed. Although not shown in this paper, response curves similar
to those shown in Figure 6 for burners higher up in the active burner array show that the
overall response of the total burner array would be soundly stable in this frequency
range.
The response curves shown in Figure 6 indicate a potentially unstable mode in
the 43 to 45 hertz range. It seems clear that slightly different modeling asumptions or
input constants could show a possible weak instability in this frequency range. Mild
oscillations were observed in the 40 to 50 hertz range which appeared and disappeared as
operating conditions changed.
Figures 7 and 8 show the two effects discussed earlier relative to Eqs. (1) and
(2). Figure 7 shows the effects of a poorly anchored flame within an active burner. The
numbers, n, with which the curves are labeled essentially represent the sensitivity of the
fraction of combustion completed within a burner to variations in the air flow rate
through the burner. A value of (n=0) represents a solidly anchored flame, with the
fraction of combustion completed within the burner independent of air flow rate
variations. In this case, the presence of partial combustion within the burner has a
stabilizing effect, resulting from the increased steady-state resistance to air flow
through the burner. Not shown in Figure 7 is another possible curve, for (n = infinity).
This represents the case where the flame is so poorly anchored that, in response to air
flow rate variations, it alternates between positions deep within the burner and
completely lifted off the burner exit. Such a case could be violently unstable.
Figure (8) shows the effect of staged combustion on stability. As the fuel is cut
off to more burners (more air-only burners in the total burner array), the open loop
response at the (lowest) unstable frequency becomes larger. Near the fuel-rich
flammable limit in the burner the case of alternate lift-off and flash-back can be
55
-------
encountered and again the system can become violently unstable.
The violent instabilities which can accompany both of these cases can be
avoided by careful attention to burner design to provide: (a) soundly steady flame
anchoring within or just downstream of the burner exit; and (b) local air/gaseous-fuel
ratios in the flame issuing from the burner, which are maintained well above the fuel-
rich flammable limit until combustion is nearly complete. Combustion instability can
represent a significant limit to NOX reduction by the staged combustion techique unless
this attention is paid to burner design.
56
-------
REFERENCES
O. W. Dykema and R. E. Hall, "Analysis of Gas-, Oil- and Coal-Fired Utility
Boiler Test Data, "/proceedings of the EPA Symposium on Stationary Source
Combustion, EPA-/2-76-152C, The Aerospace Corporation, El Segundo,
California, (June 1976). '
O. W. Dykema, Effects of Combustion Modifications for NO Control on
A^^^^~
Efficiency and Combustion Stability in Utility Boilers, The Aerospace
Corporation, El Segundo, California (to be published as an EPA report).
O. W. Dykema, Analysis of Test Data for NOx Control in Gas- and Oil-Fired
Utility Boilers, EPA-650/2-75-012 (NTIS PB 2~41918/AS), U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina, The Aerospace
Corporation, El Segundo, California (January 1975).
O. W. Dykema, Analysis of Test Data for N0x Control in Coal-Fired Utility
Boilers, EPA-600/2-76-274 (NTIS PB 261066/AS), U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina, The Aerospace
Corporation, El Segundo, California (October 1976).
O. W. Dykema, "Feed System Coupled Instability in Gas/Gas Combustors,"
proceedings llth JANNAF Combustion Meeting - VII, CPIA Pub. 261, p. 51,
The Aerospace Corporation, El Segundo, California (September 1974).
JANNAF Working Group on Combustion, Design and Development Procedures
for Combustion Stability in Liquid Rocket Engines, O. W. Dykema,
Committee Chairman, CPIA Pub.256 (September 1974).
57
-------
t-,
"o
>>
3
CO
C
.a
a
e
03
P>
•D
-------
CD LU
CD r—
CD
F=5So^
OOCC^SE?
CD ^ CC CD ^
CD LU
-------
ID LJU
CDGC
03
3 6
>>£
0
*D co
O C
2 =
if
If
I
60
-------
co_
i~
a?_
II
£
>>
3
OB
a
*->
w
O
O
O
•o
O
e
O
o
£
a>
•O
0)
c
05
•o
O
a?
ca
61
-------
GO
o
Q_
oo
LLJ
cc
Q_
CD
o
LU
O
§
Z
CD
3
T
T
NOMINAL OPERATING CONDT10NS
'0
10
20 30
FREQUENCY, HERTZ
40
50
Figure 5. Comparison of results of the current analysis with those of
conventional rocket engine analysis.
62
-------
CO
•ZL
CD
Q_
CO
LU
oc
Q_
CD
CD
Q_
CD
5
4
CD
LU
CD
3
0
NOMINAL OPERATING CONDITIONS
MEASURED FREQUENCIES ON TEST WITH
PARTICULARLY VIOLENT VIBRATIONS
180° PHASE SHIFT
AT THIS FREQUENCY
COMBUSTION TIME
DELAY, milliseconds
- 40
---- 50
k- ESTIMATED RANGE OF
SEVERE VIBRATIONS
ESTIMATED RANGE OF
. MILD VIBRATIONS
20 30
FREQUENCY, HERTZ
50
Figure 6. Comparison of analytical predictions with experimental observations.
63
-------
10
8
GO
CD
Q_
GO
DC
Q_
R 6
Q_
CD
CD
ID
CD
-------
2.0
1.8
-------
-------
OVERFIRE AIR TECHNOLOGY FOR TANGENTIALLY FIRED
UTILITY BOILERS BURNING WESTERN U.S. COAL
By:
A. P. Selker and R. L. Burrington
Combustion Engineering, Incorporated
Windsor, CT 06095
67
-------
-------
ABSTRACT
The paper reviews the results of a program designed to study methods of
reducing NOX formation in tangentially fired steam generating units firing
western U.S. coal types. High (western bituminous) and low (sub-bituminous)
rank coals were studied. This program was performed under the sponsorship of
the Office of Research and Development of the Environmental Protection Agency
(Contract 68-02-1486).
69
-------
CONVERSION FACTORS
SI METRIC UNITS TO ENGLISH UNITS
To Convert From
kg/s
ng/J
MJ/S
ug/J
J/G
MN/m2
KW/m2
To
103 Ib/hr
lb/106 Btu
106 Btu/hr
lb/106 Btu
Btu/lb
PSIA
106 Btu/hr-ft2
Multiply By
7.936640
2.326E-3
3.412141
2.326
4.299226E-1
1.450377E+2
3.16998E-1
ENGLISH UNITS TO SI METRIC UNITS
To Convert From
To
Multiply By
10J Ib/hr
PSIA
lb/106 Btu
lb/106 Btu
106 Btu/hr
Btu/lb
106 Btu/hr-ft2
kg/s
MN/m2
ng/J
ug/J
MJ/S
J/G
KW/m2
1.259979E-01
6.894757E-3
4.29922E+2
4.29922E-1
2.930711E-1
2.326
3.154594
1.8
32
70
-------
ABBREVIATIONS AND SYMBOLS
Abbreviations
Definitions
NO
X
THC
NA
X~S
WW
MCR
TA
EA
FFZ
NSPS
Oxides of Nitrogen
Total Hydrocarbons
Not Available
Excess
Waterwall
Maximum Continuous Rating
Theoretical Air to Fuel
Firing Zone
Excess Air
Fuel Firing Zone
New Source Performance Standard
Symbols
N02
CO
°2
S02
C02
Nitrogen Dioxide
Carbon Monoxide
Oxygen
Sulfur Dioxide
Carbon Dioxide
Note: % TA = Percent theoretical air to the active fuel firing
zone. 100% TA equals stoichiometric air required
for combustion.
EA = Percent excess air measured at the economizer outlet
(with no overfire air 115% TA = 15% EA, with overfire
air 105% TA + 10% OFA - 15% EA).
71
-------
-------
INTRODUCTION
The emphasis on improved quality of the environment has led to the design
of coal-fired steam generators with overfire air equipment to reduce and con-
trol NOX emission levels. In tangentially-fired steam generators, the overfire
air is admitted through registers in an extended windbox.
Tests conducted on Combustion Engineering coal-fired steam generators have
demonstrated that overfire air with tangential firing has been effective in
reducing NOX emission levels by as much as 50 percent of uncontrolled values.
Some of the field tests were performed by Combustion Engineering, Inc.
under an Environmental Protection Agency sponsored two-phase program to iden-
tify, develop, and recommend the most promising combustion modification tech-
niques for the reduction of NOX emissions from tangentially coal-fired utility
boilers with a minimum impact on unit performance.
The previous two-phase program is briefly described as follows:
Phase I (EPA Contract 68-02-0264) consisted of the selection of a
suitable utility boiler to be modified for experimental studies to
evaluate NOX emission control, and a preliminary application economic
study indicating the cost range of a variety of combustion modification
techniques applicable to existing and new boilers (1).
Phase II (EPA Contract 68-02-1367) consisted of modifying and testing
the utility boiler selected in Phase I to evaluate overfire air and
biased firing as methods for NOX control. This phase also included;
1. The completion of detailed fabrication and erection drawings.
2. Installation of analytical test equipment.
3. Updating of the preliminary test program.
73
-------
4. A baseline operation study.
5. Analysis and reporting of test results.
6. The development of control technology application guidelines
for existing and new tangentially coal-fired utility boilers.
This program was conducted at the Barry Steam Station, Unit No. 2 of
the Alabama Power Company (2,3).
The major portion of the early C-E test programs, and those completed for
EPA were conducted on units firing Eastern or Midwestern bituminous coals.
In recent years, the utilization of low sulfur Western U.S. coals as an
energy source has increased significantly. The incentive for their use is the
capability of meeting SOX emission levels without the use of flue gas scrubbers,
These fuels are abundant and may be used in lieu of oil and natural gas, which
are in short supply.
Following the Phase II tests for EPA, a recommendation was made to inves-
tigate Western coal types for NOX emissions. A contract (EPA Contract 68-02-
1486) was awarded to Combustion Engineering, Inc. to field test a Western
bituminous coal and a Western subbituminous coal.
The objective of this program was to investigate the effectiveness of
employing overfire air as a method of reducing NOX emission levels from tangen-
tially fired steam generators firing Western U.S. coals. The effect of redu-
cing NOX emission levels was evaluated with respect to unit performance, unit
efficiency, waterwall corrosion rates, and related gaseous emission levels.
Specifically, the factors considered in realizing this objective were
as follows:
1. The program was conducted on two steam generating units designed
with overfire air registers, the first unit firing a Western U.S.
subbituminous coal and the second unit firing a Western U.S.
bituminous coal.
2. The test program evaluated baseline, biased firing, and overfire
air operation and consisted of approximately 60 steady-state
tests per unit averaging two to three tests per day and two
months of waterwall corrosion rate studies per unit.
3. The effect of NOX control methods on all gaseous constituents
74
-------
was evaluated during all tests. The following constituents were
measured: NOX, SOX, CO, THC, 02 and particulate samples for un-
burned combustible analysis.
4. The effects of NOX control methods on steam generator performance
were evaluated during all tests by obtaining necessary temperatures,
pressures, flows, etc., with calibrated equipment.
5. Based on the results of this program, conclusions and recommen-
dations were made pertaining to the acceptable application of
staged firing with respect to NOX emission levels, corrosion
rates, and unit operation for each type of coal tested.
6. The results of this program were compared with the results obtained
under Contract 68-02-1367 for a unit equipped with an overfire air
system not included in the original design.
This paper will report major results of the Western coal test program
conducted at Utah Power and Light Company's Huntington Canyon No. 2 unit and
Wisconsin Power and Light Company's Columbia No. 1 unit. Additionally, major
results from Alabama Power Company's Barry No. 2 unit are reported, so that
results from all three units can be readily compared.
Side elevations of the three test units are shown in Figures 1, 2, and 3.
Major design features of the test units are presented in Table I, and average
coal properties are given in Table II.
75
-------
OVERFIRE AIR SYSTEM DESIGN
The overfire air (OFA) system (Fig. 4) retrofitted to the Barry Unit No.
2 provided for the introduction of up to 20 percent of the total combustion air
above the fuel admission nozzles at full unit loading. The overfire air was
introduced into the furnace tangentially through two separate compartments
near each furnace corner located approximately 2.4 meters above the fuel
admission zone.
While the Barry Unit No. 2 required separate OFA compartments due to unit
structural considerations, the Huntington Canyon No. 2 and Columbia No. 1 OFA
systems (Fig. 5) were designed as vertical extensions of the corner windboxes.
All three systems provided for fuel/air and OFA nozzle tilting (+ 30
degrees from vertical plane) and separate compartment flow control dampers to
permit a study of the effects of various flow rates, introduction angles, and
compartment airflow distributions.
76
-------
TEST INSTRUMENTATION
The effect of using overfire air as a combustion process modification
technique for emissions control was evaluated using the following instrumen-
tation (Figure 6) and methods.
1. A Scott chemiluminescence NO-NOX analyzer (0-2000 PPM).
2. An L&N paramagnetic 02 analyzer (0-25 percent).
3. A Beckman non-dispersive infrared CO analyzer (0-1000 PPM).
4. A Scott flame ionization total hydrocarbon (THC) analyzer (0-1000
PPM).
5. ASM! particulate train and ASTM carbon analysis.
Unit performance was monitored using the instrumentation and analytical
procedures shown in Table III.
77
-------
RESULTS
BASELINE OPERATION STUDY
It has been well documented that the formation of NOX is dependent upon
excess air and the oxygen concentration in the combustion zone, the oxygen
concentration in the combustion zone being directly related to excess air and
also to the theoretical air (TA) to the fuel firing zone. Theoretical air to
the fuel firing zone is a computational tool used by Combustion Engineering,
Inc. that accounts for variations in position and leakage in all windbox
compartment dampers. This method allows for the accounting of leakage in the
compartments above the top active fuel compartment and, therefore, is a
better approximation of the actual air (i.e., oxygen) available for combustion
in the fuel firing zone than is total excess air (EA). Therefore, all param-
eters are plotted versus theoretical air to the fuel firing zone, rather than
the total excess air. For the baseline operation study, the TA is essentially
the same as the total excess air measured at the economizer outlet.
Figure 7 is a plot of N02* versus TA for the full load baseline tests at
Alabama Power Company's Barry Station Unit No. 2, Utah Power and Light Compa-
ny's Huntington Canyon Station Unit No. 2 and Wisconsin Power and Light
Company's Columbia Energy Center Unit No. 1. As shown by this figure, N02 is
proportional to TA and, therefore, to oxygen concentration in the fuel firing
zone and excess air.
Figure 8 is a plot of NOo versus TA for the half load tests for all three
units. As with the full load tests, the half load tests also show increasing
* Throughout this paper, oxides of nitrogen (NOX) are expressed as nitrogen
dioxide (N02> to be consistent with the reporting requirement of the
Standards of Performance for New Stationary Sources (4).
78
-------
N02 emission levels with increasing TA. Comparison of the full and half load
tests show that at similar theoretical air levels, the N02 emission levels for
the half load tests are lower or equal to the NC>2 levels for the full load
tests. The effect of load is better shown in Figure 9, where NC>2 emission
levels are plotted against theoretical air level for full, three quarter, and
one half load baseline tests. This plot shows that in some, but not all cases,
N02 levels tend to increase with unit loading. It can also be shown that
occasionally the opposite trend was observed. While N02 levels correlated
well with TA, attempts to find what effect fuel nozzle tilt and furnace condi-
tion had on NOX formation were not as successful. The effect of fuel nozzle
tilt was found to have a wide and inconsistent variation with NC>2 emission
levels.
Other investigators have found that increased slagging of the furnace
walls tend to increase NO,., by increasing the furnace outlet temperature and,
A,
therefore, the bulk flame temperature (5,6). Bulk flame temperature increases
due to the reduced heat transfer from the hot combustion gases to the water-
cooled furnace walls. The amount of reduction in heat transfer may depend
greatly upon the type of slag on the furnace walls. The furnace conditions
for the full and half load tests are indicated on Figures 7 and 8. Furnace
condition showed no discernable effect on N02 emission levels. Furnace condi-
tion was measured by visual observation of the furnace waterwalls. Since
waterwall absorption is closely related to furnace condition, an attempt was
made to correlate NOj emission levels with furnace waterwall absorption and
therefore with furnace condition. This attempt produced no meaningful results.
The lack of correlation between N02 emission levels and furnace condition may
be partially attributed to the fact that the visual observation of furnace
waterwall deposits is very subjective.
The effect of reducing TA on CO emission levels and carbon heat loss is
shown on Figures 10 and 11 for the full load tests. Both CO emission levels
and carbon heat loss increase with decreasing TA. This trend is a result of
the reduced oxygen available for complete combustion. CO emission levels show
no effect due to furnace condition. However, carbon heat loss appears to
decrease with increasing furnace waterwall deposits. This may be related to
the higher bulk flame temperatures encountered in a heavily slagged furnace.
79
-------
accomplished by changing the overfire air register damper opening. The maxi-
mum overfire air rate corresponds to the overfire air register dampers being
100 percent open. With the exception of Barry No. 2, the overfire air systems
were designed to introduce up to 15 percent of the total combustion air above
the top level of fuel nozzles at MCR. Barry No. 2 was designed to introduce
20 percent of the total air as overfire air. During normal boiler operation,
the overfire air dampers are opened just enough to cool the overfire air
registers.
As the overfire air dampers are opened, the NC^ emission levels drop for
a constant excess air level. This trend is shown in Figure 14. Six excess
air levels have been shown, with the trend being similar for all excess air
levels.
Theoretical air to the fuel firing zone and overfire air damper opening
are closely related, with TA decreasing as the overfire air damper opening
increases. Figure 15 is a plot of N02 versus TA for the damper variation
tests for all three units. For these tests, as in the baseline and biased
firing studies, the N02 emission levels are found to increase with increasing
TA. The evidence shown in Figures 14 and 15 indicates that NO is more depen-
dent upon TA rather than EA.
Once the optimum excess air level and overfire air rate had been deter-
mined for each unit, the second test series was run. This test series involved
a variation in tilt of the overfire air registers and fuel nozzles. The
variation in tilt refers to how many degrees toward or away from each other
the fuel nozzles and overfire air registers are moved. This variation is
calculated by taking the difference in degrees that the overfire air registers
are angled toward or away from the fuel nozzles, i.e., overfire air register
tilt minus fuel nozzle tilt.
Tilt variation of the fuel nozzles and overfire air registers is designed
to move the fuel firing zone both in the furnace and in its position relative
to the overfire air registers. Movement of the fuel nozzles and overfire air
registers away from each other accentuates the effect of staged combustion.
Movement of the fuel nozzles and overfire air registers toward each other mini-
mizes the effect of staged combustion because the air is being forced down
into the firing zone.
80
-------
BIASED FIRING OPERATION STUDY
Biased firing involves the removal of a fuel firing elevation from service
with the dampers left open to admit air through the idle fuel nozzle eleva-
tions. The effect on N02 emission levels when taking various fuel elevations
out of service is shown in Figure 12. The lowest N02 levels for each unit
were obtained when the top fuel firing elevations were removed from service
and the respective compartment air dampers were 100 percent open. Overfire
air operation is simulated by this method of unit operation. The trend is for
increased N02 levels as the elevation being removed is lower in the windbox.
The increase in N0£ levels can be attributed to the increased oxygen available
in the fuel firing zone.
Examination of the units on an individual basis showed a slight reduction
in N02 levels when the bottom fuel firing elevation was removed from service.
This reduction in IK^ might be caused by a cooling of the hot combustion gases
by the cooler combustion air being admitted through the bottom fuel firing
elevation.
In Figure 13, N02 is plotted versus TA for the full load biased firing
tests. The correlation found for the baseline tests is also evident for the
biased firing tests, N02 being directly proportional to TA.
CO emission level and carbon heat loss plots for the biased firing tests
have not been included. Preliminary plots of these variables against TA
revealed wide and inconsistent variations. This inconsistency is most prob-
ably due to firing with different fuel elevations out of service.
OVERFIRE AIR OPERATION STUDY
The overfire air operation studies were divided into three separate test
series, each designed to determine an optimum operating condition. The three
test series were:
1. Excess air and overfire air rate variation
2. Overfire air register tilt variation
3. Load and furnace waterwall deposit variation at optimum conditions
The first of these test series involved the variation of the overfire air
rate at various excess air levels. Variation of the overfire air rate is
81
-------
Figure 16 is a plot of NC^ versus the difference in tilt of the fuel
nozzles and overfire air registers. N02 emission levels are found to be high-
est when the overfire air registers and fuel nozzles are angled toward each
other and lowest when they are angled away from each other. From the stand-
point of NOX reduction, the optimum tilt variation would be with the overfire
air registers and fuel nozzles angled away from each other. However for ease
of boiler operation, parallel operation of the overfire air registers and fuel
nozzles would be best.
In Figure 17, NC>2 is plotted versus TA for the second series of tests in
the overfire air study. Again, NC^ emission levels are found to be directly
proportional to TA.
In the final series of tests for each unit, the effects of load and fur-
nace waterwall deposits on NOX formation are examined. Boiler operation was
at the optimum conditions determined in the previous test series for each unit.
Half, three-quarter, and full load tests were conducted on each unit at clean
and dirty furnace conditions. Figure 18 is a plot of the N0? emission levels
versus TA for each test in this series. This figure attempts to minimize the
effect of TA and show the effect of load and furnace condition on N0?
emission levels. Both Huntington No. 2 and Columbia No. 1 show increases in
N02 levels as unit load rises from half load to full load. The effect of
furnace condition on these units shows inconsistent variation in the results.
Except for one half load test, Barry No. 2 results also indicate an increase
in NC>2 levels with increasing unit load.
For the overfire air studies, plots of CO emission levels and carbon heat
loss versus TA produced the same trend that was established in the baseline
operation studies. The CO levels and carbon heat losses were found to increase
with decreasing theoretical air levels.
BOILER PERFORMANCE
Figure 19 is a plot of unit efficiency versus excess air for the full
load tests performed on the subject units. As can be seen in Figure 19, biased
firing and overfire air boiler operation did not affect unit efficiency. In a
previous section, it was shown that NOo emission levels can be reduced through
the use of overfire air. Therefore, these results indicate that it may be
82
-------
possible to reduce NO 2 emission levels without adversely affecting boiler per-
formance or operation.
In general, unit efficiency is found to decrease with increasing excess
air. The decrease in unit efficiency with increasing excess air levels can be
attributed to the increasing economizer outlet gas flows and temperatures and
therefore to increased dry gas losses.
The two to three percent difference in unit efficiency between Columbia
Energy Center, Unit No. 1 and Barry No. 2 or Huntington No. 2 can be attributed
to higher dry gas losses and moisture in the fuel losses for Columbia No. 1.
These higher losses are due to the type of coal being fired at Columbia No. 1.
WATERWALL CORROSION COUPON EVALUATION
Thirty-day waterwall corrosion coupon evaluations were performed at the
baseline and optimum overfire air conditions for each unit. The purpose of
these evaluations was to determine what effect low excess air or staged com-
bustion would have on waterwall tube wastage.
The method used to evaluate corrosive potential, waterwall tube wastage,
in a boiler is by exposing samples of tube material to furnace conditions for
finite periods of time and then measuring the weight losses. This is accom-
plished by inserting test probes each consisting of five coupons into the
furnace fuel firing zone and maintaining them at typical waterwall metal temp-
eratures. Figure 20 depicts the type of probe and coupons used to obtain such
information. This particular probe utilized air to keep the coupon at the
desired temperature.
Typical instrumentation to automatically maintain the desired temperature
consists of an electronic controller, and a. pneumatic controller. The pneu-
matic controller operates as a switching device, using solenoid valves, to
regulate the amount of cooling air to the probe. The amount of air is based
on a signal from the electronic controller, which is tied into the sensing
thermocouple at the probe coupon.
At the end of the exposure period, the coupons are evaluated for weight
loss and visual evidence of attack. The average weight losses for the base-
line and overfire air modes of boiler operation are shown in the following
83
-------
table. The results indicate that waterwall tube wastage is unaffected by mode
of boiler operation.
Unit
AVERAGE CORROSION COUPON WEIGHT LOSSES
Baseline
Operation
Alabama Power Company
Barry Station No. 2
Wisconsin Power & Light Co.
Columbia Energy Center No. 1
Utah Power & Light Co.
Huntington Station No. 2
2.6381 mg/cm2
8.0770 mg/cm2
3.4266 mg/cm2
Overfire Air
Operation
4.4419 mg/cm2
8.0933 mg/cm2
2.6357 mg/cm2
The weight losses for Barry No. 2 and Huntington No. 2 are within the
range of losses that would be expected for the oxidation of carbon steel for a
thirty-day period. This premise was verified by control studies conducted in
C-E Power Systems' Kreisinger Development Laboratory.
The weight losses measured at Columbia No. 1 are slightly higher than
expected. One possible reason for the higher losses is that some of the probes
overheated during the thirty-day tests. Another possible reason for the
higher weight losses is that the coal being burned at Columbia No. 1 is sub-
bituminous, while the Barry and Huntington units both burn bituminous type
coals. The results for the Columbia tests, however, show the weight losses
are equivalent, regardless of the mode of boiler operation.
84
-------
SUMMARY
Percent excess air, bulk flame temperature, and residence time of the com-
bustion gases, all directly affect the formation of oxides of nitrogen (NOX).
The two oxides of nitrogen that are significant are nitric oxide (NO) and ni-
trogen dioxide (NC^). NO is more predominant and accounts for 90 to 95 per-
cent of the total NO generated in a utility boiler. Once it enters the atmos-
X
phere, NO is converted to N02» which is more hazardous to human health. Most
references in this report to N02 are actually referring to total nitrogen
oxides. This method of expressing NOX as N02 is in agreement with EPA practice.
While it is not the subject of this report, it should be noted that NOX
generated by the combustion of coal can occur by two mechanisms. One mechanism
is by the oxidation of atmospheric nitrogen (thermal NOX), while the other
mechanism involves the conversion of fuel-bound nitrogen (fuel NO ). The for-
X
mation of thermal NOX is known to be dependent on flame temperature, oxygen
concentration in the combustion zone, and residence time at temperature.
Several investigators have observed that the formation of fuel NOX is
responsible for a significant portion of the total NO emitted from the combus-
X
tion process (5-8). The reaction can take place at a much lower flame tempera-
ture and has also been shown to be dependent on the oxygen concentration in
the combustion zone. The coals being fired at Barry No. 2 and Huntington
Canyon No. 2 had analyses ranging from 1.1 to 1.3 percent nitrogen by weight.
Columbia No. 1 had an analysis ranging from 0.6 to 0.8 percent nitrogen by
weight. Preliminary plots of N02 versus the coal nitrogen content did not
show any correlation between N02 and coal nitrogen content. Any correlation
would probably have been masked by the limited range of the nitrogen content
of the coals being fired and by the variation in excess air levels.
The test programs conducted on the subject units showed that overfire
air operation is effective in reducing NOX emission levels. Reductions in NOX
emission levels can be accomplished without adversely affecting unit perfor-
mance . Unit loading was found to have a minimal effect on NOX formation,
while waterwall slag conditions showed wide and inconsistent effects on NO .
XV
emission levels.
85
-------
REFERENCES
1. Blakeslee, C. E. and A. P. Selker, "Program for Reduction of NOX from
Tangential Coal-Fired Boilers - Phase I," EPA-650/2-73-005 (NTIS PB 226-
547/AS), U.S. EPA, Research Triangle Park, NC, 1973, p. 195.
2. Selker, A.. P., "Program for Reduction of >IOX from Tangential Coal-Fired
Boilers - Phase II," EPA-650/2-73-005-a (;>ITIS PB 245-162/AS), U.S. EPA,
Research Triangle Park, NC, 1977, p. 144.
3. Selker, A. P., "Program for Reduction of NOX from Tangential Coal-Fired
Boilers - Phase Ha," EPA-650/2-73-005b (NTIS PB 246-889/AS), U.S. EPA,
Research Triangle Park, NC, 1975, p. 37.
4. Standards of Performance for New Stationary Sources," Federal Register,
Vol. 36, No. 247, Part II, Dec. 23, 1971.
5. Winship, R. D. and P. W. Brodeur, "Controlling NOX Emissions in Pulverized
Coal-Fired Units," Engineering Digest, September, 1973, pp. 31-34.
6. Vatsky, J. and R. P. Welden, "NOX, A Progress Report," Heat Engineering,
July/September, 1976, pp. 125-129.
7. Haynes, B. S. and N. Y. Kirov, "Nitric Oxide Formation During the
Combustion of Coal," Combustion and Flame, Vol. 23, 1974, pp. 277-278.
8. Graham, J., "Combustion Optimization," Electrical World, June 15, 1976,
pp. 43-58.
86
-------
TABLE I. MAJOR DESIGN FEATURES OF TEST UNITS
Barry No
Generator rating, Mw
Main steam flow @ MCR (Ib/hr)
(kg/s)
Reheat steam flow @ MCR (Ib/hr)
(kg/s)
Superheat outlet temp. (F)
(C)
Superheat outlet press. (PSIG)
(MN/m2)
Reheat outlet temp. (F)
(C)
Reheat outlet press. (PSIG)
(MN/m2)
Mills (number)
Fuel elevations
125
900,000
113
800,000
101
1000
538
1875
12.9
1000
538
404
2.8
4
4
TABLE II. AVERAGE PROPERTIES
Barry No
Carbon 65.4
Hydrogen 4.3
Nitrogen 1.3
- Oxygen 7 . 4
Sulfur 2.3
Moisture 8.8
High heating value 11,701
. 2
Huntington Columbia
. 2 Canyon No. 2 No. 1
400
3,036,000
382
2,707,000
341
1005
541
2645
18.2
1005
541
559
3.8
4
5
OF COALS TESTED
Huntington
Canyon No. 2
66.8
5.1
1.3
10.6
0.5
8.0
12,110
520
3,800,000
479
3,392,000
427
1005
541
2620
18.1
1005
541
556
3.8
4
6
Columbia
No. 1
48.8
3.4
0.7
12.2
0.8
24.9
8,485
(Btu/lb)
87
-------
TABLE III. INSTRUMENTS AND ANALYTICAL
PROCEDURES FOR MONITORING UNIT PERFORMANCE
Flow rates
Temperatures
Pressures
Parameter
Steam and Water
Feedwater flow
Reheat and superheat
desuperheat spray
Reheat flow
Laboratory
analysis
Air and Gas
Total Air and Gas
Weight
Overfire air
Air heater leakage
Steam and Water
Deg. F
Unit absorption rates
Waterwall absorption
Air and Gas Deg. F
Steam and Water
PSIG
Unit absorption rates
Unit draft loss
Temperature and
Pressure
Fuel and Ash
Instrument/
Analytical Procedure
Flow orifice
Heat balance (deg. F &
PSIG) around desu-
perheater
Heat balance around
reheat extraction
and estimated turbine
gland seal losses
Calculated
Pitot traverse
Paramagnetic 02
analyzer
Calibrated stainless
sheathed CR-C well &
button TC's
Calibrated stainless
steel sheathed CR-C
chordal WW TC's
CR-C TC's
Water cooled probes
Pt/Pt-10% Rh TC's
Pressure gauges and/or
transducers
Water manometers
C-E data logger capac-
ity: 400 temperatures,
50 pressures
ASTM procedures
-------
Figure 1. Unit side elevation, Alabama Power Company, Barry Station No. 2
89
-------
Figure 2. Unit side elevation, Utah Power and Light Company,
Huntington Station No. 2
90
-------
Figure 3. Unit side elevation, Wisconsin Power and Light Company,
Columbia Energy Center No. 1
91
-------
V
F- FUEL AND AIR
A- AIR
0-OVERFIRE AIR
Figure 4. Test unit overfire air system schematic
92
-------
WIND8OX
SECONDARY AIR DAMPERS
SECONDARY AIR
DAMPER DRIVE UNIT
OVERFIRE AIR
NOZZLES
SIDE ICNITOR
NOZZLE
SECONDARY
AIR NOZZLES
— COAL NOZZLES
OIL GUN
Figure 5. Corner windbox showing overfire air system
93
-------
Figure 6. Gaseous emissions test system
94
-------
110 115 120 125 130 135
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
UO
LEGEND
OAlabama Power Co.
Barry #2
Awi scons in Power I
Light Co.
Columbia II
QUtan Power i Light Co.
Huntington #2
Furnace Condition
13 Clean
(•Moderately Dirty
•Dirty
* NSPS New source performance standard
Figure 7. N02 vs. theoretical air, baseline study, maximum load
360
360
340
320
300
280
Z60
240
220
200
180
160
NSPS
115
150
120 125 130 135 140 145
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
USEND
OAlabama Power Co., Barry n d Clean
QWisconsin Power & Light Co., Columbia #1 (^Moderately Dirty
QUtah Power & Light Co., Huntington #2
155
Figure
8. N02 vs. theoretical air, baseline study, 1/2 load
95
-------
380
360
?4n
320
300
260
240
200
NSPS
9
/
J©
/*
A
AG
*rt
C7J
^W
/
0
(
&7
/H
•
H
'gl
•"
^
,«
•
A
Lrf
pa
p
A
A
Ld
f^
<
^
1
if-
LEGEND
0 Alabama Power Co.
Barry #2
Light Co.
Columbia #1
El Utah Power & Light Co
Huntington Canyon 12
Unit Loading
Q Three Quarter
9 One Half
100 110 120 130 140 150
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
160
Figure 9. NOo vs. unit loading, baseline study
40
36
32
28
24
2°
,6
12
8
4
0
0
105 110 115 120 125 130 135
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
140
LEGEND
©Alabama Power Co.
Barry *2
^Wisconsin Power &
Light Co.
Columbia fl
Dutah Power & Light Co.
Huntington *2
Furnace Condition
Ddean
LlModerately Dirty
• Dirty
Figure 10. CO vs. theoretical air, baseline study, maximum load
96
-------
A
105 110 115 120 125 130 135
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
LtGENn
OAlabama Power Co.
Barry *2
^Wisconsin Power &
Light Co.
Columbia tl
Hlltah Power & Light Co.
Huntington »2
Furnace Condition
Qciean
(iModerately Dirty
• Dirty
140
Figure 11. Carbon heat loss vs. theoretical air,
baseline study, maximum load
A A
B E
8 5
5 x
c
LEGEND
OAlabama Power Company
Barry «2
AWiscons in Power &
Light Co.
Columbia #1
Qutah Power & Light Co.
Huntington #2
120 140 160 180 200 220 240 260 280 300 320
NO,,, ng/J
Figure 12. Fuel elevation out of service vs. N0?, biased firing study
97
-------
320
300
280
260
P 240
d>
T 220
° 200
180
160
140
120
NSPS
•-2T
90 95 100 105 110 115 120
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
125
LEGEND
©Alabama Power Co.
Barry »2
^Wisconsin Power &
Light Co.
Columbia *t
GJutah Power & Light Cc.
Huntington *2
130
Figure
13. NO- vs. theoretical air, biased firing study, maximum load
350
300
250
200
150
100
0 20 40 60 80 100
OVERFIRE AIR REGISTER DAMPER OPENING. % OPEN
LEGEND
0Alabama Power Co,
Barry *2
^Wisconsin Power S Light Co.
Columbia *I
Qutah Power j Light Co.
Huntington #2
Figure 14. N02 vs. OFA damper opening, overfire air study
98
-------
350
300
250
200
150
100 '-
80 90
I
LEGEND
©Alabama Power Co.
Barry 12
3/4 Load
Awisconsin Power & Light Co.
Columbia «!
Full Load
ID Utah Power & Light Co.
Huntington 12
Full Load
100 no 120
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
130
Figure 15. N09 vs. theoretical air, overfire air study
NSPS
300
280
260
240
i 220
;vi 200
180
160
140
120
60
0
20
40
40 20
TOWARD AWAY
OFA REGISTER AND FUEL NOZZLE TILT DIFFERENTIAL, DEGREES
LEGEND
©Alabama Power Co.
Barry «2
^Wisconsin Pu.-,'er f,
Light Co.
Columbia «1
Qlltah Power « Light to.
Huntington »2
60
Figure 16. NO,., vs. tilt differential, overfire air study
99
-------
NSP5
300
280
260
240
220
j 200
180
160
140
120
_,
j
r t
i
85
&
;© ;
0 jS<§>
_.f
s
/
90 95
E
B
Q
-• \
A
1
\ /^ !
. .. 0. x
>*
E
B-
S[*\
—&~
a.... *.... , —
1
&
S~ ' ;
A . |
i 1
1
i i
i
' 1
•
L£GEND_
©Alabama Power Co.
Barry K
^Wisconsin Power S
Light Co.
Columbia «1
0Utah Power & Light Co.
Huntington *2
100 105 110
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT
Figure 17. NO.,
240
230
220
210
200
?, 190
c
i
eg 180
^
170
160
150
140
130
n
i
:
,
j
«;
V
*
•'
on
vs. theoretical
A
&
.
t>
G
0
Ci
€>
^
a
QE 1 {HI
air, overfire air study
LEGEND
1
1 ;
• j
<> i
T
&
1 rtt
i
.
t>
1 1 f\ 11
Alabama Power Co.
Barry #2
© Full Load
E3 3/4 Load
^1/2 Load
Utah Power & Light Co.
Huntington #2
&Full Load
&3/4 Load
Gil/2 Load
Wisconsin Power & Light Co.
Columbia #1
QFull Load
3/4 Load
(2)1/2 Load
Furnace Condition
OLight
115 OModerately Dirty
THEORETICAL AIR TO FUEL FIRING ZONE, PERCENT •Dirty
Figure 18. N02 vs. theoretical air, overfire air study
100
-------
90
89
88
O
10 20 30 40
a. Alabama Power Co., Barry 12
10 20 30 40
b. Utah Power & Light Co., Huntingtort 12
88
87
86
A* *\^l
* .fc^'A
10 20 30 40
c. Wisconsin Power & Light Co., Columbia *1
EXCESS AIR AT ECONOMIZER OUTLET, PERCENT
O
LEGEND
Baseline
Study
Biased Firing
kOverfire Air
Figure 19. Unit efficiency vs. excess air
101
-------
u
AIR
OUTLET
ADAPTER
PLATE
AIR
INLET
OXYGEN
SAMPLING
INSERT^ --
\ TT
CORROSION
COUPONS
THERMO-
COUPLE
WELLS
LOCK NUT
Figure 20. Corrosion probe assembly drawing
102
-------
THE EPRI PROGRAM ON NOX CONTROL USING
COMBUSTION MODIFICATION TECHNIQUES
By:
K. E. Yeager and D. P. Teixeira
Electric Power Research Institute
Palo Alto, CA 94303
103
-------
-------
THE EPRI PROGRAM ON NOX
CONTROL USING COMBUSTION MODIFICATION TECHNIQUES
Kurt E. Yeager
Assistant Department Director
Fossil Fuel Power Plants Department
Donald P. Teixeira
Program Manager, Air Quality Control
Fossil Fuel Power Plants Department
ABSTRACT
The utility industry, through the Electric Power Research
Institute (EPRI), is actively pursuing a significant NOX control
technology program to assure that the NOX emissions do not become
a factor limiting the construction or operation of fossil fueled
power plants. The program is directed primarily at understanding
the reliability, performance and cost implications of alternative
control technologies. This paper specifically reviews the status
and results to date from three major combustion control projects
in the EPRI program: (1) low-NOx coal combustion development,
(2) selective non-catalytic ammonia reduction of NOX, (3) low-NOx
turbine combustor development.
The low-NOx combustion process development is a cofunded project
with Babcock & Wilcox. The configuration and design parameters
of a two-stage combustion system for coal will be described.
The selective, non-catalytic ammonia reduction of NOX being
performed in cosponsorship with Exxon Research and Engineering
Co. is directed at verifying the behavior of ammonia with NOX in
the presence of coal fly ash. The laboratory-scale (3 x 106
Btu/hr) examination of four bituminous and subbituminous coals
will be described.
The objective of the low-NOx turbine combustor development
program is to evaluate the potential for a gas turbine combustor
capable of meeting 75 ppm NOX at 15% 02 without water injection.
The performance of dry control configurations incorporating pre-
mixing coincidental with prevaporization will be described.
105
-------
-------
THE EPRI PROGRAM ON NOX
CONTROL USING COMBUSTION MODIFICATION TECHNIQUES
INTRODUCTION
Of the three major pollutants (SOX, participates and NOX) emitted
by power plants, only NOX is limited to a current control effi-
ciency of 30 to 50%. The utility industry, through the Electric
Power Research Institute (EPRI), is therefore actively pursuing a
significant NOX control technology program to assure that NOX
emissions do not become a factor limiting the construction or
operation of fossil fueled power plants. The program is directed
at understanding the reliability, performance and cost impli-
cations of alternative control technologies. The objective is to
provide the utility industry an informed basis for not only
responding to, but also shaping, the future course of NOX control
requirements.
The EPRI NOX control program, defined in Figure 1, consists of 15
projects and a six-year budget of $12 million. This program is
directed to four objective areas: (1) analysis and control of
fuel nitrogen conversion, (2) combustion control for steam
generators, (3) combustion control for gas turbine/combined cycle
power plants, and (4) post-combustion control of NOX.
Relatively minor changes in burner designs and boiler operating
procedures (i.e., staged combustion and low excess air) have been
applied to satisfy the 0.7 lb/106 Btu NOX standards for coal-
fired steam generators recently proposed by EPA, although
questions regarding furnace tube corrosion and potentially toxic
byproducts, such as polycyclic organic matter, are as yet un-
answered. However, EPA is considering lowering NO stndards even
further: EPA research goals of 200 ppm NOX in 1980 and 100 ppm
in 1985 have been identified for pulverized-coal steam gener-
ators. Because of the relative simplicity and economic savings,
the emphasis on NOX control at EPRI has been on combustion
modification. A lower priority has been placed on post-
combustion alternatives because of the substantially higher costs
involved and the likelihood of major reliability impacts.
This paper reviews the status and results to date from three
major combustion control projects: (1) the low-NOx boiler
combustion development, (2) selective non-catalytic ammonia
reduction of NOV, and (3) low-NOv turbine combustor development.
107
-------
LOW NOX BOILER COMBUSTION PROCESS DEVELOPMENT (RP899)
The major problem in trying to achieve low NOV levels from coal
A
is similar to that encountered with sulfur — the nitrogen
organically bound within the coal molecules is a major source of
the emissions. Were this fuel-bound nitrogen not present, such
control technologies as wind-box flue gas recirculation and
staged combustion would be adequate to control NOX to very low
levels, as evidenced by considerable data from natural gas and
residual oil-fired boilers. The fuel-bound nitrogen, however, is
unlike its sulfur counterpart in that it does not necessarily
result in a solid byproduct scrubber sludge, sulfuric acid, or
even elemental sulfur that somehow must be disposed of. There is
a considerable amount of data indicating that the fuel, nitrogen
may be reduced to molecular nitrogen by properly manipulating the
combustion chemistry.
The fundamental requirement to accomplish the conversion of fuel
nitrogen to N2 is through fuel-rich combustion. However, this
probably cannot be accomplished by an extension of current
staged-combustion techniques with conventional burners. More
likely, it will require a completely new burner technology that
can provide the proper temperature, time and stoichiometry
specifically for low NOX. The system must physically isolate the
fuel-rich combustion process from the secondary air injection
zone, which is required to maintain an overall oxidizing condi-
tion in the boiler. One such approach is the primary combustion
furnace concept (Figure 2) proposed by Babcock & Wilcox (B&W)
under EPRI project RP899. Pulverized coal is introduced into a
conventional B&w dual-register burner with less air than is
required for complete combustion. Any resemblance to existing
burners ends at this point. The extended length of the combustor
provides the necessary residence time to partially oxidize the
coal and permit the desirable ^-producing reactions to occur.
Heat removal will also occur along the combustion chamber to
avoid slagging. Secondary air is added at the exit of the
primary combustion furnace to bring the combustion products to
oxidizing conditions for the balance of their passage through a
conventional convective section.
The development of the low-NOx combustion process will be
performed on two scales. The first tests will be at 4 x 10
Btu/hr. These tests will evaluate the process variables
necessary to accommodate low NOX while maintaining acceptable
combustion characteristics. Heat removal, residence time and
quantity of air in the primary combustion furnace are major para-
meters to be defined. Due to the small scale, only gross aspects
of reliability can be evaluated in this research. Following
successful completion of testing at the 4 x 10 Btu/hr scale,
108
-------
research will then move to a 50 x 10" Btu/hr facility. This
research will confirm the NOX and combustion process variables
determined in the earlier work and evaluate material require-
ments, mechanical design, and longevity. Results of the 50 x
10" Btu/hr tests can be extrapolated to typical full-scale
utility burner ratings (150-200 x 106 Btu/hr).
Cost estimates for this technology have been provided by B&W.
New unit costs are estimated at under $5/kW and retrofits are
projected to be under $20/kW. While these figures must be
regarded as preliminary, the attractiveness of the combustion
control approach is obvious when one considers that post-
combustion control techniques for new units are currently being
estimated at $30/kW and up.
Four coal types will be evaluated on this low-NOx combustor
development effort: two Eastern bituminous, one subbituminous
and one lignite. Support efforts include: {1} state-of-the-art
evaluation of nitrogen conversion chemistry with laboratory study
of fuel nitrogen chemistry in both well stirred and plug flow
reactors, and (2) analysis of commercial utility steam generator
systems incorporating the low-NOv combustion.
X
SELECTIVE NON-CATALYTIC NO,, REDUCTION
An alternative to limiting the formation of NOX through
combustion modifications is processes which lead to destruction
of combustion-generated NOX. The gas phase decomposition of
nitric oxide is one attractive control approach for NOX in boiler
flue gases. The primary advantage of gas phase decomposition
over a catalytic system is the avoidance of the catalyst costs
and maintenance problems associated with the more complex
catalyst systems. Further, if a selective reduction in nitric
oxide occurs, then the amount of reducing agent can be minimized.
The first use of ammonia or reducing NO in a combustion system
was reported by Wendt et al. (1) . They injected ammonia down-
stream of the reaction zone of a premixed flat flame operating
with 2% excess oxygen. Wendt et al. found that significantly
larger reductions in nitric oxide were obtained with ammonia
injection than with methane. They attributed this to a pyrolysis
of the ammonia to hydrogen in the injector with the hydrogen then
reacting with the nitric oxide. Recently, Lyon (2) and Lyon and
Longwell (2] have shown that selective gas phase decomposition of
nitric oxide can occur in the presence of oxygen by the addition
of ammonia to the combustion products. These gas-phase reactions
occur over the temperature range of 1200°F to 2000°F. A
patent (2) has been issued to Exxon Research and Engineering
covering the use of ammonia for the selective homogeneous
109
-------
reduction of NOX.
LABORATORY EVALUATION OF THE HOMOGENEOUS GAS-PHASE
DECOMPOSITION OF OXIDES OF NITROGEN (RP461)
EPRI, under a contract with KVB, Inc., has experimentally
evaluated the potential for selective gas phase NOX reduction
under utility boiler operating conditions (4J . The study deter-
mined the conditions of reducing agent type, concentration,
temperature and time which would lead to the selective reduction
of NOX in the presence of varying amounts of oxygen and nitric
oxide. In addition, the study also investigated the potential
for NOX reductions through the injection of reducing agents into
fuel-rich combustion products laden with NOV.
A
By far the most sensitive parameter in the selective non-
catalytic reduction of NOX is temperature. With ammonia as the
reducing agent, a selective homogeneous gas phase reduction of
nitric oxide is achieved over the temperature range of 1200°F to
2000°F. This effect of temperature is shown in Figure 2 for
typical boiler conditions: an excess oxygen level of 4% and
initial NO level of 300 ppm and various amounts of injected
ammonia. Significant NO decomposition begins at a temperature of
about 1500°F with peak reduction of NO occurring near 1750°F. As
the temperature at the point of ammonia injection is further
increased above 1750°F, the decomposition of NO by ammonia
becomes less effective. At temperatures of about 2000°F, no
reduction is observed. In fact, as can be seen in Figure 2, the
injected ammonia oxidizes to nitric oxide at temperatures greater
than 2300°F. Thus, a temperature window of less than 250°F
exists in which substantial reductions in nitric oxide can be
achieved with the injection of ammonia into the combustion
products.
The ratio of injected ammonia to initial nitric oxide is the
second critical parameter in the selective reduction process. As
seen in Figure 3, NO reductions vary from 30% to 92% as the ratio
of (NH^J/fNO) varies from 0.3 to 1.6. From the data presented,
it may be concluded that a selective reduction in NO is occurring
with ammonia injection. For the conditions of Figure 2, a ratio
of injected NH3 to initial NO level of about 106 (based on NH3 +
3/402 1/2H2 + 3/2H20) would be required to consume all of the
excess oxygen and render the combustion products stoichiometric.
However, the data show that only trace amounts of ammonia are
required, relative to the stoichiometric condition to effect
large NO reductions. This is further confirmed since no measure-
able change in the excess oxygen level occurred over the range of
ammonia injection rates studied.
110
-------
The small-scale laboratory experiments conducted during this
study indicated that the selective homogeneous gas phase decompo-
sition of nitric oxide with ammonia or other reducing agents may
be a viable emission control technique for conventional utility
boilers. Further study is necessary to establish the viability
of this approach either as a technique for specialized appli-
cation, or indeed as a universally applicable approach to NOX
control. Specifically, additional research is needed to confirm
the observed NOX reduction levels on a larger scale. The impact
of fly ash and possible reactions between sulfur compounds and
the reducing agent on reliability and byproduct emissions also
need to be determined. Specific concerns include the potential
for NH^HSO^ fouling of low temperature heating surfaces and
sulfate formation. In addition, evaluation of the feasibility of
maintaining the narrow range of effective temperatures necessary
for low NOX emissions in existing and new large utility boilers
is required. This information is necessary before determination
of the commercial applicability of this technique can be made.
From this study the following specific conclusions may be drawn
with regard to the laboratory scale investigation:
1. A selective reduction in nitric oxide occurs when controlled
quantities of ammonia are injected into the combustion
products.
2. The temperature region in which the selective reduction of
NOX occurs is between approximately 1300°F and 2000°F, with
the peak reductions taking place at about 1750°F with
ammonia injection.
3. At the peak reduction temperature, a ratio of injected
ammonia to initial NO of unity yielded an 80% reduction.
4. The NO is primarily reduced to molecular nitrogen (N2).
5. Ammonia emissions were maintained at less than 10 ppm if the
reducing agent is injected at a temperature slightly higher
than the peak effectiveness temperature. This results in
the excess reducing agent being consumed by the excess
oxygen.
6. The tests further indicate that the presence of sulfur
oxides in the combustion products had no measurable effect
on the NO reductions obtained.
7. Selective removal of NO can also be achieved by the
injection of other amine species (e.g., CH3NH2, (CH3)2NH,
and (CH3)3N); however, these are of little practical
interest.
-------
8. When ethane, methane, carbon monoxide, or hydrogen are used
as reducing agents,1 a nonselective reduction occurs with
significant nitric oxide reductions occurring only when the
combustion products become fuel rich.
9. Ammonia, ethane, and methane were also effective in reducing
nitric oxide in fuel rich combustion products. For this
case, the maximum reductions occur at higher temperatures
(typically greater than 2300°F) with ammonia again being the
most effective reducing agent.
LABORATORY EVALUATION OF COAL-FIRED N0y REDUCTION WITH
AMMONIA REJECTION (RP835)
Under project RP835, EPRI is continuing to investigate the direct
reduction of NOX with ammonia rejection, specifically as it
appies to coal combustion. This effort is being conducted in co-
sponsorship with Exxon Research and Engineering Company at KVB,
I;nc. The research is directed at determining the levels of NO
removal possible for four different coals (New Mexico subbitumi-
nous, Pittsburgh Seam 8 bituminous, Illinois No. 6 bituminous and
Utah low-sulfur bituminous) using selective, noncatalytic ammonia
injection. Tests are also being performed to (1) obtain data on
the type and concentration of potential byproduct emissions, and
{2} determine the extent to which hydrogen can lower the effec-
tive process temperature range.
The tests will be conducted in the KVB 3 x 106 Btu/hr facility.
A test matrix, as shown in Figure 4, will be undertaken to
determine the effect of temperature, ammonia injection rate and
coal type.
EVALUATION OF A PREMIXED, PREVAPORIZED GAS TURBINE COMBUSTOR
(RP359)
In addition to the regulations being considered for pulverized-
coal power plants, EPA has issued proposed emissions standards
for industrial gas turbines.
75 ppm at 15% 0? for both liquid and gaseous fuels.
NOX emissions will be limited to
The only means currently available for meeting these standards
involve water (or steam) injection into the combustor. Unfortu-
nately, this technique has a capital cost of at least S10-$15/kW
and a fuel consumption increase of 2-3%. Increased maintenance
costs are also probable, so a system that avoids water injection
is desirable. In conjunction with Solar Division, International
Harvester Company, EPRI has undertaken a project to evaluate the
feasibility of a low-emission combustor that does not use water
or steam. This is commonly referred to as the dry approach.
112
-------
There are several ways of controlling NOX without water
injection. All these, however, require that the fuel and air be
completely mixed prior to combustion. The most difficult conven-
tional fuel to accommodate is No. 2 distillate because it must be
vaporized as well as mixed before combustion occurs.
Accordingly, most research has centered on this fuel. The most
common premixed combustion method uses high-pressure { 10 atm)
and high-temperature { 650°F) combustor inlet air to provide the
heat of vaporization of the distillate oil. Figure 5 is a
diagram of the combustor tested. The main fuel injection is into
the premixing ports. During its passage through the ports, the
fuel is evaporated and mixed to a uniform stoichiometry with the
airstream. At this point, the fuel-air charge enters the primary
zone, where combustion occurs. Subsequent secondary and dilution
zones are designed to use essentially conventional combustor
design principles. Fuel can also be introduced through the pre-
combustor. This provides the capability to independently vary
inlet air temperature and also permits added turndown
flexibility.
Preliminary results of the emissions performance of the dry
combustor have been obtained (5} . Up to about 7 atm combustor
pressure, emissions are within proposed EPA standards. However,
at the design operating pressure of 10 atm, emissions several
times higher than required were observed. The probable cause of
the high emissions is incomplete evaporation and mixing of the
fuel in the vaporization tubes. One solution to improved fuel
vaporization is increased combustor inlet air temperatures.
However, increasing the temperature produced autoignition of the
fuel-air mixture in the vaporization tubes. Autoignition
resulted in failure of the fuel preparation ports.
The autoignition results can also be interpreted in terms of
mixing rates. Slow mixing between vaporizing fuel and air favor
locally high equivalence ratios. These higher-than-average
ratios can then produce autoignition at conditions other than
those governed by chemical processes at the overall stoichio-
metry. These considerations form the basis for future tests
involving the effect of increased initial fuel dispersion (more
fuel injection sites) and droplet size.
Although the evaporation/autoignition problem may be resolved
through one or more of the techniques described, the solution may
still be of only academic interest if practical problems of
safety, reliability and availability are considered. For
example, a perfectly acceptable solution to autoignition may be
to use a fuel injector with more than four fuel injection sites
and larger vaporizing ports at compressor exit air temperature
entering the vaporization tubes. However, the system may be
113
-------
marginally stable, and only a minor perturbation in flow or
temperature could lead to catastrophic results.
Other problem situations are conceivable. For example, if 10-
point fuel injectors were found to be an acceptable solution,
this would mean, for example, in a typical gas turbine having
eight cans, with six premixing tubes per can and each premixer
tube having the requisite 10-point fuel injection sites, that 480
fuel injectors would be required. The reliability problems
associated with this complexity are clearly high. Finally, if
significant blockage of some of the fuel lines were to occur, the
balance between a stable vaporization/preignition situation would
be disrupted. This could shift the vaporizing/premixing post-
equivalence ratio to a value in the autoignition range, with
subsequent mechanical failure of the combustor.
Optimistically, a combustor design solution avoiding these poten-
tial problems can be developed. It is also clear, however, that
this achievement of a dry control method for gas turbine NOX will
require considerable additional research before a commercially
acceptable solution can be found.
114
-------
REFERENCES
1. Wendt, J.O.L., Sternling, C. V., and Matovich, M. A.,
"Reduction of Sulfur Trioxide and Nitrogen Oxides by
Secondary Fuel Injection," 14th Symposium (International) on
Combustion, The Combustion Institute, 1973.
2. Lyon, R. K., "Method for the Reduction of the Concentration
of NO in Combustion Effluents Using Ammonia," U.S. Patent
No. 3,900,554, assigned to Exxon Research and Engineering
Company, Linden, New Jersey, August 1975.
3. Lyon, R. K. and Longwell, J. P., "Selective, Non-Catalytic
Reduction of NOX with NH3," EPRI NOX Control Technology
Seminar, San Francisco, California, February 5 and 6, 1976.
4. Muzio, L. J. and Arand, J. K., "Homogeneous Gas Phase
Decomposition of Oxides of Nitrogen," Electric Power
Research Institute, RP253, August 1976.
5. Teixeira, D. P., White, D. J. and Ward, M. E., "Evaluation
of a Premixed, Prevaporized Gas Turbine Combustor for No. 2
Distillate," American Society of Mechanical Engineers,
77-6T-69, March 1977.
115
-------
CO
3
O
(B
in
a
o
c
o
o
cc
I
X
o
\
i-i
cn
-H
PH
116
-------
Figure,2.Primary combustion furnace concept as proposed by Babcock & Wilcox Co. The
extended combustion permits N2-producing reactions to occur.
117
-------
1.0
0,8
0,1
0.2
0
120.
I
=(NH3)/(NO)
I
I
1600 1800
TEMPERATURE, °F
2000 2200
Figure 3. Effect of Temperature on NO Reductions with Ammonia
Injection. (Excess Oxygen 4%, Initial NO 300 ppm).
118
-------
FIGURE 4.
EXPERIMENTAL MATRIX
NH3 Injection Tests
Excess Oxygen
NOX Level
Natural
Approx. 6%
Burner Produced
400-700 ppm
NHo Flow Rate
(d)
Number of tests with each fuel
Total NH3 tests: 56
(c)
Temperature at 1400°F to 1900°F 4
Injection Point
(NH3)/(NO) = 0-5 4
molar
16
(b)
Gas Coal Av ' Coal Bv ' Coal C
(b)
.(b)
1
1
16
1
1
1
1
Coal D
1
1
4
(b)
(a) Numbers indicate the number of variations.
{b} Coal A - New Mexico sub-bituminous;
Coal B - Pittsburgh Seam 8 bituminious;
Coal c - Illinois bituminous;
Coal D - Colorado bituminous;
(c) NOX will be added to the burner air.
(d) Where there are two NH3 runs, the ratios will be 1.5:1 and 2.5.1,
Variable
Excess Oxygen
NOX Level
Temperature at Injection
Point
NH3 Flow Rate
NH3 Flow Rate
Total NH3 + H2 Tests:
Fuel Type
Coal A
NH3+ H? Injection Tests
Range
Approx. 6% 1
Burner Produced 400-700 ppm 1
1000°F to 1500°F 2
(NH3)/(NO) = 0-3 2
(NH3)/(NO) = 0-to be determined 4
16
SUMMARY:
Total number of Gas Tests:
Total Number of Coal Tests:
Total Number of Tests:
119
16
56^
72
-------
O
-P
if)
o
o
X
C
(t
>-,
tr-
c
•H
•H
fa
120
-------
DESIGN AND SCALE-UP OF LOW EMISSION BURNERS FOR
INDUSTRIAL AND UTILITY BOILERS
By:
R. Gershman
Energy and Environmental Research Corporation
Santa Ana, CA 92705
This paper was not received in time for publication, and therefore will be
inlcuded in Volume V.
121
-------
-------
CYCLONE BOILERS - THEIR NOX EMISSIONS AND POPULATION
By:
T. E. Ctvrtnicek
Monsanto Research Corporation
Dayton, OH 45407
and
S. J. Rusek
Owens-Corning Fiberglass Corporation
Granville, OH 43023
123
-------
-------
ABSTRACT
There are 149 boiler units in the United States that are fired by a total
of 736 cyclone furnaces. All of the cyclone furnaces are located in the East
and Midwest. Three states, Illinois, Missouri, and Indiana, account for nearly
half of the total-fired steaming capacity and one-third of the boilers.
Since cyclonic combustion takes place at high combustion temperatures,
cyclone furnaces firing utility and industrial boilers are high NO emitters.
Based on available data, their measured full-load emissions were 576 ng/J to
718 ng/J for bituminous coal firing, 546 ng/J for subbituminous coal firing,
291 ng/J to 355 ng/J for lignite firing, 217 ng/J to 318 ng/J for residual oil
firing, and 208 ng/J to 325 ng/J for natural gas firing.
The authors estimate that 0.76 x 106 metric tons of NO were emitted from
X
all cyclone coal-fired utility boilers in 1973. This represents from 19% to
22% of the NO produced by all coal-fired utility boilers in the U.S. Corres-
"X
pondingly, between 10% and 13% of coal consumed by all coal-fired utilities was
used by cyclone fired units. Similar estimates for industrial boilers could
not be obtained due to insufficient data.
Several combustion modification techniques have been applied to cyclone
boilers/furnaces in an attempt to lower their NO emissions. These include
X
boiler load reduction, low excess air firing, two-stage firing, and switching
fuels. Even though significant reductions in NO were achieved, none of the
X
techniques was shown to reduce NO emissions to the level meeting the EPA's
X
New Source Performance Standards for NO .
125
-------
ACKNOWLEDGEMENT
This work was supported by the U.S. Environmental Protection Agency under
Contract 68-02-1320, Task 20. Much of the data and information presented here-
in were kindly provided by the Babcock & Wilcox Company and by the Commonwealth
Edison Company.
126
-------
SECTION 1
INTRODUCTION
The first cyclone-fired boiler began operation 33 years ago. At that time,
cyclone firing represented a major breakthrough in the art of firing trouble-
some coals high in ash content and having a low ash fusion temperature. Both
of these factors, frequently an annoyance to both boiler operators and design-
ers, were an advantage in cyclone furnace operation.
Successful operation of the cyclone furnace depends on maintaining a
liquid or wet slag within the cylindrical furnace. Crushed coal (95% will pass
through a 4-mesh screen) is introduced tangentially through a primary burner at
the front of the cyclone furnace, thrown to the walls of the cyclone, and
caught in the running slag, Figure 1. Tangentially supplied secondary air at
velocities of 91.4 m/s sweeps past the embedded coal particles, quickly oxi-
dizing them. The cyclone is typically 1.8 ra to 3.0 m in diameter and about
3.4 m long and is water-cooled. In order to maintain proper slagging, low
furnace heat absorption rates and low ash fusion temperatures are maintained
and the cyclone is operated at temperatures as high as 1920 K. Large quanti-
ties of fuel are combusted within a relatively small volume, resulting in
furnace high heat release rates (4.7 MW/m3 to 7.8 MW/m3; a pulverized-coal-fired
unit typically has a heat release rate of 0.2 MW/m3).
Since cyclonic combustion intrinsically requires high combustion tempera-
tures, the cyclone-fired boilers are high NO emitters. This paper briefly
summarizes information that was compiled on cyclone boiler population and NO
emissions, both with and without modifications made to decrease such emissions.
127
-------
SECTION 2
POPULATION
The first full-scale cyclone-furnace-fired boiler was placed on stream
in 1944 at the Calumet Station of the Commonwealth Edison Company, Calumet,
Illinois. Since then, a total of 84 cyclone-fired installations have been
built in the U.S. These installations are located in 26 states, contain a
total of 149 boiler units fired by a total of 736 cyclone furnaces, and have
a primary steam-generating capability of approximately 26,000 kg/s (about 9%
of the total U.S. steam-generating capacity). Illinois, Missouri, and Indiana
account for nearly half of the total cyclone steaming capacity and one-third
of the boilers. Table I gives a further breakdown of the cyclone-fired boiler
population. It shows that over 94% of the total primary steaming capacity is
held by the electric utility sector (24,253 kg/s) which operates 116 of the
149 boilers. These boilers are fired by 677 furnaces. The remaining 33
boiler units are owned by private industry and institutions. Table I also
indicates that primary steam-generating capacities of individual boiler units
built range from 16 kg/s to 70 kg/s for industrial and commercial units and
from 23 kg/s to 1,160 kg/s for the electric utility units. All these units were
built by the sole manufacturer of cyclone furnace boilers, Babcock & Wilcox
Company, who estimates that the majority of the boilers listed in Table I are
still in use even though some may have been derated because of their age.
Since about 1973, the Babcock & Wilcox Company has not sold a single
cyclone unit. The decline of sales started with the strict Federal SO regu-
lations imposed on new stationary combustion sources. The low ash fusion
temperature coals burned in the cyclone boiler normally have a high sulfur con-
tent. Switching to low-sulfur coals normally results in ash with a high fusion
128
-------
temperature. The final event which restricted the sale of bituminous-coal-
firecl cyclones was the limitation of NO emissions for stationary'combustion
x
sources.
129
-------
SECTION 3
BASELINE EMISSIONS FROM CYCLONE FURNACE INSTALLATIONS
Although all existing cyclone furnaces were originally designed to burn
coal, many other types of fuels have been and are still being successfully
fired in them. These fuels include residual and distillate oils, solid wastes
(wood bark, coke), and natural gas. Baseline emissions from these units are
defined to be those NO emissions reflecting normal or near normal boiler
X
operation at various loads. Full-load (91% to 100%) baseline emissions are
summarized in Table II by type of fuel along with the New Source Performance
Standards (NSPS) for NO . Data in Table II were obtained from 14 boiler
X
units field tested by the Exxon Corporation, the Tennessee Valley Authority,
the Babcock & Wilcox Company, KVB, Inc, and NAPCA (EPA).
130
-------
SECTION 4
CYCLONE COMBUSTION MODIFICATIONS
Only a relatively small number of cyclone boilers were found to have been
examined and tested in some way to determine the effects of combustion modifi-
cations on NO emissions. One reason for the lack of field data on this com-
X
bustion equipment class is the relative inflexibility of the cyclone furnaces
and boilers in respect to combustion process modification. Altering the cyclone
operation sufficiently to control NO can result in a furnace that is no
X
longer a cyclone.
Four types of combustion modifications have been applied, however, either
singly or in combination to reduce NO emissions from cyclone furnaces. These
X
are: low excess air firing, load reduction, staging, and switched fuel firing.
Twelve boilers were tested under modified combustion conditions. The
modification techniques applied most often have been load reduction and low
excess air firing because they require no physical modification or changes of
existing cyclone units.
LOW EXCESS AIR FIRING
In one test performed on a lignite-fired boiler at full load, reduction
in the oxygen content of the flue gas by 75% (from 6.4% 02 to 1.6% O2)
reduced NO emissions by 47%. However, CO concentrations increased from
X
0 vppm to 17 vppm (see Table III). Firing the cyclone with 1.1
flue gas required supplemental oil fuel to maintain ignition.
in the
131
-------
Likewise, for oil-fired boilers at full load decreasing the excess oxygen
lowered NO emissions. These results are summarized in Table IV. In no case
x
were NO emissions reduced enough to meet the NSPS. The lower NO levels
x 6 x
achieved by reducing excess air again could not be justified because of the
increased levels of CO.
LOAD REDUCTION
Load reduction in cyclone-fired boilers results in consistently lower NO
emissions compared to the same boilers at full or normal loads. This is
usually considered an economically unattractive method for reducing NO ends-
X
sions, however, because of the penalties incurred and because of reduced ther-
mal efficiency and reduced boiler flexibility at reduced loads. Maximum loads
reduction for a cyclone boiler appears to be limited to about 40% of the
maximum continuous steam rating. Below this point load reduction causes flame
instability with possible loss of ignition, lack of adequate steam temperature
control, and excessive slagging on the cyclone walls when coal is fired. Five
of six units tested showed an overall reduction of NO emissions as load was
x
reduced. Two of these boilers were oil-fired. With minor extrapolation it
was possible to compare the results from the five boilers, as shown in Figure 2.
With a 20% reduction in load, three bituminous-coal-fired units showed a
reduction in NO emissions ranging between 24% and 32%. For the same load
reduction two oil-fired boilers showed NO emission reductions ranging between
X
8% and 30%. With the exception of one oil-fired boiler operating at the
reduced load of 57%, the reduction in NO emissions was nevertheless inadequate
X
to meet NSPS.
STAGING
Staging requires physical modification of the cyclone boiler. There are
definite limits to the extent to which existing units can be modified, however.
Consequently, the field test data obtained under staged firing conditions are
132
-------
limited in nature and representative only of the specific conditions and
arrangements existing during the tests.
Staged firing is based on sustaining the initial part of the combustion in
a reducing atmosphere zone. Several forms of staged firing have been applied
to cyclone boilers, mainly two-staging and pattern firing.
Applying two-staging to the cyclone boilers involves reduction of the
amount of combustion air fed into the cyclone. The remainder of the air is
sent into the boiler at a point near the exit of the cyclone furnace proper.
Two-staging has been applied to an eastern-coal-fired boiler and a gas-
fired boiler. The first unit showed a 28% to 36% reduction in NO emissions.
X
However, when firing coal, two-staging may require an oil supplement to main-
tain ignition and flame stability. The gas-fired cyclone boiler showed a 48%
reduction in NO emissions.
x
Pattern firing can only be applied to a boiler with multiple cyclones in
stacked configuration. The technique is based on use of different air-to-fuel
ratios for the stacked cyclones in such a way as to produce a staged effect.
Using this technique 21% to 24% reduction in NO emissions was achieved in a
X
residual-oil-fired unit. At the same time the level of
creased from 1.3% to 1.6%, a 23% increase.
in the stack in-
Staging has been shown to reduce NO emissions to a limited extent, but
X
long-term test data are not available. Sustained operation of cyclones in a
reducing atmosphere can cause catastrophic failure due to tube corrosion. For
these reasons the cyclone boiler developer (B&W) does not recommend staging
as a viable method of NO reduction.
x
OTHER METHODS
Additional efforts have been made to investigate cyclone boiler opera-
tions in which more than one combustion modification technique was applied.
133
-------
None of these efforts reduced the NO level enough to meet NO emission
X A
standards.
The same is true for switching fuels. As indicated by baseline emis-
sion data none of the fuels fired in the cyclone furnace can meet existing
NO standards. Consequently, fuel switching in itself does not appear to be
X
a promising solution.
The final report on this study was published by EPA in January 1977.
This report, entitled "Applicability of NO Combustion Modifications to Cyclon*
X
Boilers (Furnaces)," EPA-600/7-77-006, is available from NTIS as PB 263960/AS.
134
-------
SECTION 5
CONCLUSIONS
All attempts to date to modify cyclone furnace combustion by methods
including low excess air firing, load reduction, staging, and switched fuel
firing have failed to reduce NO emission level sufficiently to meet the EPA's
X
NO New Source Performance Standards for any type of fuel.
.X.
135
-------
REFERENCES
Ctvrtnicek, T. E., and S. J. Rusek. Application r*f NO Combustion
Modifications to Cyclone Boilers (Furnaces). EPA-600/7-77-006, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina,
January 1977. 121 pp.
Additional 29 references are cited in Reference 1 above.
136
-------
H •
II
w
g -3
y iu w
> T-4 CN rj **t fi
I %Q O i
CM H
rig:
i O ^ M M M 1
137
-------
TABLE II. FULL-LOAD BASELINE NO EMISSIONS
A
Fuel
Q
Bituminous coal
o
Subbituminous coal
Lignite3
Residual fuel oil
Natural gasc
NOX emissions,
dry 3% 02
basis, ng/J
576 to 718
546
291 to 355
217 to 318
208 to 325
NOX NSPS,
dry 3% 02
basis, ng/J
301b
**•
258
129
86.0
Reduction
needed to meet
the NSPS, %
48 to, 58
b
11 to 27
41 to 59
59 to 74
1 vppm % 0.6 ng/J was assumed.
Not available.
'1 vppm £ 0.5 ng/J was assumed.
TABLE III.
NOX EMISSIONS FROM A
LIGNITE FIRED BOILER
UNDER LOW EXCESS AIR
CONDITIONS3
NOX9 dry 3%
in flue 02 basis, CO,
gas, % ng/J vppm
6.4
5.6
5.1
4.9
4.6
4.3
4.0
2-9,
H
1.6a
411
384
345
337
302
360
384
324
216
_c
c
c
c
_c
10
12
12
17
Data courtesy of the Babcock &
Wilcox Company.
Data reported in vppm, conver-
sion factor of 1 vppm = 0.6 ng/J
was assumed.
^
Not available.
Low excess air required supple-
mental oil to maintain ignition.
138
-------
TABLE IV.
NO EMISSIONS FROM OIL-FIRED BOILERS
x
Average flue gas measurements8
Excess air
level
Boiler A
Normal
Intermediate
Lowc
Boiler B
Normal
Low
NOX, dry
3% 02 basis
0?, % ng/J
1.5 254
1.1 228
0.5 181
2.2 206
1.6 181
CO, 3%
02 basis,
vppm
57
74
1,523
85
231
Flue gas measurements made on composite gas samples
from three individual sampling tubes. Measurements
shown are averages of three analyses from three
sampling tubes (short, medium, and long) for each
of four probes.
Boiler 02 meter.
"Excessively high CO emissions at this condition.
139
-------
CRUSHED COAL
Figure 1. Cyclone furnace side view.
140
-------
750
700
650
600
550
* 500
CO
450
CJl
c
350
300
250
200
150
100
SYMBOL
38
25
23
6
61
41
EXTRAPOLATION
BITUMINOUS
COAL FUEL
DATA
.NEW SOURCE PERFORMANCE
STANDARD APPLICABLE TO COAL/IRING^
RESIDUAL
OIL FUEL
DATA
4-2/NEW SOURCE PERFORMANCE STANDARD APPLICABLE TO OIL FIRING
4.6
I
40 50 60 70 80 90 100
% OF MAXIMUM CONTINUOUS RATED BOILER LOAD
* 1 vppm = 0.6 ng / J
Figure 2. Overall reduction of NOX emissions for four coal- and
two oil-fired cyclone furnace boilers using load reduction
(stack % 02 levels indicated adjacent to data points).
141
-------
-------
STATISTICAL ASPECTS OF CORROSION FROM STAGING COMBUSTION
IN A WALL COAL-FIRED UTILITY BOILER
By:
J. W. Tukey
Bell Laboratories
Princeton University
Murray Hill, NO 07974
This paper was not received in time for publication, and therefore will be
included in Volume V.
143
-------
-------
NONCATALYTIC REDUCTION OF NOX WITH'NHs
By:
W. Bartok
Exxon Research and Engineering Company
Linden, NJ 07036
145
-------
-------
ABSTRACT
Exxon Research and Engineering Company has developed a post-flame injec-
tion process for the reduction of NOX to nitrogen with ammonia (U.S. Patent
No. 3,900,554). The feasibility of this technique has been demonstrated by
ER&E and Exxon's Japanese affiliate for an oil fired boiler. The technology
has been commercialized for oil and gas fired boilers. The ammonia injection
technology (Thermal DeNOx process) "is~viewed as potentially a useful supple-
ment to available combustion modification techniques for attaining low NOX
levels for installations that require such degree of control.
An analysis of utility boiler types will be made to determine what coal
fired boiler types by design, size, or manufacturer are most likely to be
amenable to the Thermal DeNOx process. Budget type cost analysis for the
application of the Thermal DeNOx .process will be made as a function of utility
boiler size, fuel, appropriate boiler characteristids and degree of,NOx reduc-
tion. The costs of the Thermal DeNOx process will be compared with those for
extreme combustion modifications that would be required to achieve very low NOX
levels.
147
-------
-------
SECTION 1
INTRODUCTION
Exxon Research and Engineering Company has developed a new process called
Thermal DeNOx for reducing emissions of oxides of nitrogen from large stationery
combustion sources. This non-catalytic process is based on the selective reduc-
tion of NOX with NH3 in the homogeneous gas phase (i»D • The Thermal DeNOx proc-
ess has been commercially demonstrated in gas and oil-fired steam boilers and
process furnaces, and tests have also been conducted on a muncipal incinerator.
Exxon Research and Engineering Company has granted licenses on this process in
Japan where NOX emission regulations are very stringent.
The Thermal DeNOx process involves injection of ammonia into the hot flue
gas within a narrow and critical temperature range. In the first full-scale
commercial demonstration, conducted in 1974 at Tonen Sekiyu Kagaku K.K.'s
Kawasaki Plant, a reduction in NOX emissions of up to 70 percent was achieved
on a 70-ton-per-hour oil-fired steam boiler. Although the temperature sensi-
tivity will cause the reaction's effectiveness to vary from one installation
to another, the NOX reduction is essentially independent of the concentration
of oxides of sulfur or particulate matter in the flue gas. The specific level
achievable is dependent upon a number of factors, including the heater design,
operating mode, and initial NOX level.
Thermal DeNOx may be applied to boilers for additional NOX reduction after
combustion modifications such as staged firing or flue gas recirculation have
been implemented. As Thermal DeNOx is a post-flame injection process, it is
not affected by certain limitations imposed on combustion modifications, e.g.,
by reduced boiler load capability in retrofit applications; Thus, the Thermal
DeNOx process is viewed as an effective supplement to available combustion
modification techniques for attaining low NOX levels for combustion instal-
lations that require such high degree of emission control.
In the present paper, the technical background' of Thermal DeNOx is reviewed
and the objectives of a pending contract between Exxon Research and Engineering
Company and the U.S. Environmental Protection Agency on the application of this
process to coal fired utility boilers are discussed.
149
-------
SECTION 2
CHEMISTRY OF THE PROCESS
The process chemistry relies on the selective reaction between NH3 and ,'
NOX to produce nitrogen and water. This reaction proceeds in the presence of
excess oxygen within a critical temperature range. The overall NO reduction
and production reactions are summarized in equations (1) and (2), respectively:
NO
NH3 + 5/4
1/4
3/2
NO + 3/2
(1)
(2)
In typical flue gas environments, the NOx reduction shown as equation (1)
dominates at temperatures around 950°C (1740°F). At higher temperatures, the
NOX production reaction shown as equation (2) becomes significant, and it
dominates at temperatures over about 1000°C (1830°F). As temperatures are
reduced below about 900*>C (1650°F), the. rates of both reactions slow, and the
ammonia flows through unreacted. These effects of temperature on NOX and NHj
concentrations are shown in Figure 1.
The following chain reaction cycle was proposed by Lyon
NH-NO-02 reaction system:
for the
NH2 + NO
NH + NO
+ H + OH
(3)
(4)
H + 0- -> OH + 0
0 + NH,
OH + NH
H
OH + NH2
H.,0 + NH,
(5)
(6)
(7)
(8)
This chain reaction mechanism is sufficient to explain qualitatively the obser-
ved reduction of NO by NH in the presence of 0 .
Exxon's technology also includes means of altering the utilizable temper-
ature range. The addition of H2 shifts the temperature window to lower levels,
as shown in Figure 2. It should be noted that hydrogen does not widen the
150
-------
temperature range, but merely changes it. The magnitude of this shift is
mainly a function of the amount of H2 injected relative to the NH-. At H
ratios on the order of 2:1, the NO reduction reaction can be forced to proceed
rapidly at 700°C (1290°F) . By judiciously selecting the H2/NH3 injection
ratio, flue gas treatment can be accomplished over the range of 700-1000°C.
The same effect produced by H- can also be provided by other combustible
gases, such as hydrocarbons or carbon monoxide. Use of these additives for
temperature control, however, is not generally recommended because of the pos-
sible formation of undesirable by-products such as small amounts of hydrogen
cyanide .
In addition to temperature, the process is also sensitive to initial NOX
and NH3 concentrations. The NH3 injection rate is generally expressed as a
mole ratio relative to the initial NOX concentration. The reductions obtained
with various initial NOX levels are shown as functions of this parameter in
Figure 3.
Other variables affecting performance are excess oxygen and available resi
dence time at the reaction temperature. Minimizing excess air tends to enhance
the NOV reduction, as does maximizing residence time.
The issue of possible pollutant by-products (HCN, N20, CO, 803 and
was addressed by Exxon Research studies \£/ . As mentioned before, HCN can be
produced only if hydrocarbons are present in the Thermal DeNOx reaction zone.
Under normal conditions, hydrocarbons are absent from this zone. As regards
N20 production, it represents only one to two per cent of the NOX reduced Q,»A) .
The Thermal DeNOx process does not generate CO by reducing C02- However, CO
oxidation is inhibited by NH3, so that if CO is present, it would be emitted
unreacted into the atmosphere. This effect is of no consequence under normal
operating conditions for oil and gas fired boilers, as CO oxidation is complete
before the NH3 injection point.
Detailed laboratory experiments have shown no interaction between the
Thermal DeNOx process and sulfur compounds in the high temperature flue gas
regions. That is, sulfur or its oxides do not interfere with the NH3-NOX-02~H2
chemistry. Additionally, ammonia injection has been shown to cause neither
additional homogeneous nor additional heterogeneous oxidation of S02 to SO^ .
To the extent that the thermal reduction of NO leaves some NH^ unreacted,
and as the combustion gases cool, NH3 reacts with SOg and 1^0 to form ammonium
sulfates. (Ammonium bisulfate is a corrosive liquid at air heater temperatures.)
Based on laboratory and commercial tests, these sulfates do not create either
severe corrosion or unacceptable air heater fouling problems when Thermal
DeNOx is used in accordance with its design specifications. Long term tests
conducted in two oil-fired boilers by Tonen Sekiyu Kagaku K.K. in Kawasaki,
Japan, revealed these deposits could easily be removed by waterwashing the air
heaters at reasonable intervals.
151
-------
SECTION 3
ENGINEERING CONSIDERATIONS
When applying the Thermal DeNOx process to commercial equipment, perform-
ance is generally limited by the extreme temperature sensitivity of the reac-
tion and its dependence on the local concentrations of reactants, NHj, NOX,
02, and H2- The Exxon technology provides a means of adapting the chemistry
requirements to industrial equipment environments, and NOX reductions up to
about 70% can often be achieved by the use of Thermal DeNOx technology in
existing boilers. Application to new, grass-roots designs is usually easier
because the internal configuration of the high temperature zone can be adjusted
to complement the process demands.
The Thermal DeNOx process utilizes proprietary Exxon gas phase mixing tech-
nology to rapidly and efficiently mix the small volume of reagents with the hot
flue gas. Correct distribution of reactants is required because of non-line-
arities in the reaction rates. Locally high concentrations of NHj will decrease
the maximum attainable NOX reduction and will also result in the breakthrough
of unreacted ammonia.
Accommodating flue gas temperature variations is important if high
rates are to be achieved. Not only does the system have to accommodate flue
gas temperature changes caused by normal load and operating variations, but it
also must allow for fluctuations across the reaction zone caused by non-uniform-
ities in flow and heat transfer. It follows, therefore, that a case-by-case
evaluation of flue gas temperatures and local conditions is required for the
application of Thermal DeNO for each installation considered.
152
-------
SECTION 4
PROCESS COSTS
The costs associated with the Thermal DeNO process are sensitive to the
particular circumstances of the application. Factors influencing cost include
initial NOX concentration, reduction target, compatibility of the heater design
and operation, and local price and availability of chemicals and utilities.
An an example, consider applying the process to a 300 MWe oil-fired utility
plant with an initial NOX level of 225 ppm (about 0.3 Ib. NOX/M Btu fired).
Assume the boiler geometry and operating conditions provide a temperature in
the reaction zone which does not require H2, and that for a 50% NOX reduction
target, an approximate NH3/NOX injection ratio of 1.0/1 is feasible. Thus,
Thermal DeNOx will have the following estimated operating costs:
(a) NH. (§1.0 mole per mole NO (assume 170 $/ton) = 0.9 C/M Btu
«s X
(b) Utility air @ 210 SCF per M Btu fired (assume 0.005 C/SCF, including
compressor cost) =1.0 C/M Btu
The total operating cost is estimated at 1.9 C/M Btu.
Note that 20 psig utility air is used as a diluent in the injection system. An
alternative approach would be to use a similar quantity of low pressure steam,
resulting in a different operating cost.
The availability of chemicals and utilities usually represents the greatest
variable in the installation cost. In situations where such facilities are
already available on site, the equivalent cost for capital investment for a
large utility boiler can be as low as about 1 C/M Btu (assuming annual charges
for finance, maintenance, and depreciation total 20% of investment).
The equivalent cost for the above example totals about 2.9 C/M Btu fired.
With the assumed reduction of NOX emissions from 225 to 112 ppm, the cost-effec-
tiveness is about 390 $/ton of NOX removed (expressed as KO^). As previously
stated, total cost and effectiveness will vary for other cases depending on both
technical and economic factors. Where higher DeNOx severities are required, or
where optimum flue gas temperatures are not available, H2 and higher N
ratios would be required, thus significantly increasing overall cost.
153
-------
SECTION 5
COMMERCIAL SCALE EXPERIENCE
Thermal DeNOx has been demonstrated in six commercial boilers and furnaces
to date. Actual capability often represents a compromise between the technical
limits of the process chemistry and cost-effectiveness. In many situations,
performance is maximized at full load operation, and smaller NOX reductions
accepted at reduced loads resulting in lower reaction zone temperatures. In such
installations, total NOX emissions are generally at target levels over the
full range of operating conditions because of the reduced NOX production
at lower loads. Results from all six demonstrations are shown over their
range of operating conditions as a function of flue gas temperature in Figure 4.
154
-------
SECTION 6
APPLICATION OF THERMAL DeNOx
TO COAL FIRED UTILITY BOILERS
At the time of writing this paper (June 29, 1977), a contract is being
negotiated between Exxon Research and Engineering Company and the U.S. Environ-
mental Protection Agency on the Thermal DeNOx process. The purpose of this
pending contract will be to evaluate and document the feasibility of Exxon
Research's proprietary Thermal DeNOx process to coal fired utility boilers for
reducing NOX to nitrogen. A program consisting of four task elements is envi-
sioned for this contract. These tasks are as follows:
Task 1. The Thermal DeNOx process technology will be documented with rele-
vant background information and data. The data and information generated by
Exxon Research in the development of the Thermal DeNOx process will include
those resulting from bench scale laboratory experimentation; development test
data obtained with an oil fired boiler and with a gas fired test furnace; and
commercial scale test data obtained at Exxon's affiliate's Kawasaki plant on
an oil fired steam boiler.
Task 2. A broad analysis will be made of utility boiler types to determine
if certain boiler types by design, size, or manufacturer, are more amenable to
Thermal DeNOx than others. This analysis will be made based on information
available on the current population of coal fired utility boilers in the U.S.
Task 3. Budget type cost analysis will be made of the Thermal DeKOx proc-
ess as a function of utility boiler size, fuel, appropriate boiler character-
istics and degree of NOX reduction desired. This analysis will include cost
components of engineering, capital investment, and operating costs. Key cases
to be considered for coal fired utility boilers are (a) trimming NOX emissions
to meet the current NSPS of 0.7 lb./106 Btu on coal fired boilers unable to
meet the standard with state-of-the-art combustion modification techniques; and
(b) further reduction in NOX emissions to the 0.3-0.4 Ib./lO^ Btu range.
\
As part of the budget type cost analysis effort, a comparison will be made
of the costs of the Thermal DeNOx process with that of extreme combustion modifi-
cations that would be required to achieve very low NO,, levels for coal fired units.
155
-------
Task 4. Without reimbursement by EPA, ER&E will conduct a laboratory
scale test program to test the effectiveness of the Thermal DeNQg process for
coal firing. The effects of Thermal DeNOx on NOX, S02» S03, HC, CO, HCN, NH3,
and particulates (including sulfate and nitrate particulates) will be deter-
mined. Relevant results of this laboratory scale test program will be made
available to EPA as part of the pending contract for their evaluation of the
results.
156
-------
REFERENCES
1. Lyon, R, K., "Method for the Reduction of the Concentration of NO in
Combustion Effluents Using Ammonia," U.S. Patent 3,900,554, August 19, 1975,
2. Lyon, R. K., "Communication to the Editor: the NHo-HO-02 Reaction,"
International Journal of Chemical Kinetics, g, '315-318 (1976).
3. Lyon, R. K. and Longwell, J. P., "Selective, Non-Catalytic Reduction of
NOX by NH3," Paper presented at EPRI NOX Seminar, San Francisco, February
1976.
157
-------
200
150 -
Q.
Q.
UJ
LL
UJ
100 -
800 900
FLUE GAS TEMPERATURE, °C
1000
Figure 1. Thermal DeNOx reaction products as
functions of temperature without hydrogen.
158
-------
200 s,
150 -
E
n.
CL
g 100
LL
U-
UJ
50 -
700
800 900
FLUE GAS TEMPERATURE, °C
1000
Figure 2. Thermal DeNOx reaction products as
functions of temperature with hydrogen added.
159
-------
r 40
h-
o
It)
o:
Excess Oxygen: 2°/
Temperature: 960°C
Initial NOX, ppm
200
100
NH3/NOX, MOLE RATIO
Figure 3. NOX reduction as a function of
NH-j injection ratio.
160
-------
70
60
50
O
Hi
ce
.ox
40
30
20
10
i
..•.$$£$&
frf8H8
•0^^0^^
^f^0f^^0^
\
; !
•
•••••
i
'.'• : £#&&&#XX£ •&•&£••''' ^
$ i mmi ^ \
^mniP^ \
P" \
UNIT DESCRIPTION \
• 70 t/hr STEAM BOILER \
* 150 kbbl/d CRUDE HEATER »
A 430 t/hr STEAM BOILER ,
v 430 t/hr STEAM BOILER \
o 120 t/hr STEAM BOILER \
0 150 kbbl/d CRUDE HEATER \
1
700 800 900 1000
INJECTION ZONE FLUE GAS TEMPERATURE, °C
Figure 4. Performance of Thermal DeNOx
systems in commercial applications.
161
-------
-------
I
WESTERN COAL USE IN INDUSTRIAL BOILERS
By:
K. L. Maloney, Ph.D.
KVB, Incorporated
Tustin, CA 92680
and
Peter L. Langsjoen
KVB, Incorporated
Minneapolis, MN 55426
163
-------
-------
WESTERN COAL USE IN INDUSTRIAL BOILERS
ABSTRACT
Ten small and intermediate-sized [4.5 Mg/hr to 113 Mg/hr (10,000
Ib/hr steam to 250,000 Ib/hr steam)] coal-fired boilers in the upper Midwest
have each been tested on both a bituminous eastern coal and a subbituminous
western coal.
The purpose of this study was to determine the feasibility of sub-
stituting western subbituminous coal for eastern bituminous coal as a means
of reducing the SO emissions from this class of boiler and to demonstrate ,
X.
the feasibility of greatly expanded western coal utilization as a means of
reducing the use of oil and gas.
The scope of this study was such that the representative boiler
types were tested on both eastern and western coal for a period of time
sufficient to completely characterize their individual emission and opera-
tional characteristics.
This research was supported under Environmental Protection Agency
Contract No. 68-02-1863.
The authors would like to express their appreciation to Mr. Dave
Lachapelle, EPA, for his continued interest in the use of western coal*
DISCLAIMER
This report has been reviewed by the National Environmental Research
Center, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views
and policies of the U.S. Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endoresement or recommen-
dation for use.
165
-------
-------
SECTION 1
INTRODUCTION
Faced with the problem of complying with sulfur dioxide control
regulations/ electric utilities and industries in the Midwest have been
increasing their use of low-sulfur western coal. The extent to which Mid-
western demand for western coal will continue to increase depends on a
number of factors. Foremost among these are: (1) the evolution of federal,
state, and local sulfur dioxide control regulations, (2) the growth of coal
as a boiler fuel, and (3) the cost of western coal relative to the costs
of alternate fuels and control technologies.
The upper Midwest region (Minnesota, Wisconsin, Iowa, Nebraska, and
Illinois) is presently an area where low-sulfur western subbituminous coal
is cost competitive with midwestern and eastern coals. Within this region,
there is considerable variation with regard to western coal use versus the
traditional eastern supply. This variability is due in part to equipment
limitations which dictate that a certain coal be burned.
For this reason it is necessary to determine the operational compa-
tibility of western coal with existing industrial coal-fired equipment,
if fuel substitution is to be considered a viable alternative sulfur oxides
control strategy.
The purpose of this program, the test results of which are detailed
in this paper, was to assess the effectiveness of the use of lower sulfur
western coals as a means of reducing sulfur oxides emissions from industrial-
sized boilers in the size range 4.5 to 113 Mg/hr (10,000 to 250,000 Ib/hr
steam). The impact on SO , NO , CO, particulates, and unburned hydrocarbon
•*( X
emissions has been assessed as a consequence of this fuel conversion.
167
-------
The scope of the testing program included testing ten representative
types of coal-fired industrial boilers for a period of one month each on
eastern and western coal. During this testing period, the pollutant emis-
sions listed above were measured both in a baseline configuration and in
an optimized firing mode. Operational problems of the unit were character-
ized for each coal. Potential reductions of pollutant emissions have been
estimated for each unit type and each coal tested.
168
-------
SECTION 2
PROPERTIES OF WESTERN SUBBITUMINOUS COALS
A large supply of low sulfur, subbituminous coal exists in the
Powder River region of Wyoming and the Fort Union region of southeast
Montana. This coal is being mined at a rapidly increasing rate. One mine
in Wyoming, for example, increased production from 0.89 million tons per
year in 1973 to 3.3 million tons per year in 1974, a factor of 3.7 in only
one year. However, the most impressive statistics are the reserve capacity
of these western coal fields. That same mine in Wyoming whose production
increased so dramatically in 1974 has a reserve capacity of 18.5 billion
tons. This translates to a lifetime of 50 years at current production
rates. The large reserves, coupled with the relative ease of strip mining,
point to a ready supply of coal for fuel if other constraints are met.
One of these constraints is the subject of this paper.
The compatibility of these western subbituminous coals with exist-
ing industrial boilers could be a hinderance to their wide acceptance as
a boiler fuel. The compatibility of coal and boiler are determined both
by coal properties and boiler design. Since the boiler designs are fixed
in existing units, the coal properties are the variables of interest.
Western coal characteristics are those of a typical subbituminous
coal: an ash-free higher heating value of 19 to 24 MJ/kg (8,200 to 10,500
moist Btu/lb), and a high moisture content of 20% to 30%. The ash content
of most of these coals is less than 10% by weight. The western subbitumi-
nous coals exhibit high volatile to fixed carbon ratios, typically approach-
ing a value of one. The coals tested during this study are presented in
Table I. The mineral analysis is also included for selected coals. The
sequence number keys the ultimate analysis with the mineral analysis.
169
-------
The western subbituminous coals are also classed as "free-burning"
coals. In the free-burning coals, the pieces do not fuse together, but
burn separately or, after fusion, the mass breaks up quickly into fragments.
This characteristic causes problems in certain types of stokers where there
is inadequate control of undergrate air distribution.
The high moisture content of the western coals causes the greatest
combustion difficulty in industrial-sized equipment. In most units with
superheaters, it leads to high steam superheat temperatures. It also causes
flame stability problems in pulverized coal combustion and ignition problems
in stoker-fired units. In order to recover the lost steam capacity, some
pre-drying of western coals will be necessary for firing in units designed
for eastern coal.
The second major problem with western coal is the size distribution
of the delivered coal. Most western coals do not travel or weather well.
The coal has a tendency to break into fine sizes while in transit. There-
fore, even if the coal has been sized before shipment, the as-received coal
will exhibit a change in size distribution toward the smaller sizes. This
shift becomes more severe with longer transit and/or storage periods. The
effect of this coal property on stoker unit performance is discussed below.
This paper is divided into a discussion of pulverized coal firing
and stoker firing of both eastern and western coal. A general overview of
boiler performance is presented in Table II. Here, the units tested are
rated in terms of emissions, efficiency, and overall ease of operation.
The type and source of the coals tested are also given for each boiler.
170
-------
SECTION 3
PULVERIZED COAL COMBUSTION
Unit No. 3, a pulverized coal-fired boiler, was tested at Dairyland
Power Cooperative at the Alma, Wisconsin generating station. This four-burner
face-fired unit manufactured by Riley Stoker Corporation is rated at 104.3
MG/hr (230,000 Ib/hr) steam flow. The coal is pulverized with two ball
tube mills, one mill for the upper two burners and one for the lower two
burners. The unit is equippped with a spray steam attemperator. Fly ash
collection is accomplished with a UOP-designed cold-side electrostatic pre-
cipitator (ESP).
The two fuels used during the testing were:
Eastern Base Coal
o 4% sulfur
o 16% ash
o 24 MJ/kg (10,500 Btu/lb)
o 18% volatiles
and
Montana Coal
o 0.77% sulfur
o 12% ash
o 19.5 MJ/kg (8,400 Btu/lb)
o 37% volatiles
BOILER PERFORMANCE - ALMA UNIT NO. 3
The boiler performed well on both coals, although the unit was limi-
ted in maximum load due to excessive superheat steam temperature on the
Montana coal. The steam attemperation system was not adequate to reduce
the temperature to the desired 755 °K (900°F) level at loads above 78.9 MG/hr
(174,000 Ib/hr) steam on western coal. This compares to a maximum load of
92.5Mg/hr (204,000 Ib/hr) steam on eastern coal. The boiler is design rated
at 104.3Mg/hr(230,000 Ib/hr) steam, however, this load is no longer achieved.
171
-------
The primary factor causing the excessive steam temperature is the
high moisture content of the coal. The water reduces the flame temperature
which in turn reduces the radiant heat flux to the water walls/ resulting
in lower steam generation. This lower heat transfer (a function of temper-
ature to the fourth power) removes less heat in the radiant section; how-
ever, the gas still contains a large enthalpy which then acts on a decreased
amount of steam in the convective section, resulting in increased steam
temperatures. The water in the fuel also results in greater gas flows which
increase heat transfer rates in the convective pass.
The excessive steam temperature problem is a function of boiler
design. For example, a boiler designed for western coal might not be able
to make design steam temperature on eastern coal.
increased steam attemperation would result in full capacity opera-
tion on western coal.
PULVERIZING MILL PERFORMANCE
Eastern coking coals, when exposed to furnace temperatures, will
swell and form lightweight, porous coke particles. These may float out
of the furnace before they are completely burned. As a result, carbon loss
will be high unless pulverization is very fine. Free-burning (western)
coals, on the other hand, do not require the same degree of fineness because
the swelling characteristic is absent.
High volatile (western) coals ignite more readily than those with
a low volatile content. Therefore, they do not require the same degree of
fine pulverization. With the exception of anthracite, however, the low-
volatile coals are softer, and therefore have a higher grindability. As
a result, mill capacity is greater at increased fineness than with high
volatile coals (Ref. 1).
Table III shows the screen analyses and the loads of the coal burned
in tests 9, 16, 57, 63, 75, and 78. Tests 9 and 16 were on eastern coal.
Test 16 was with one mill out of two operating so the load in the mill was
172
-------
the same as it would have been with both mills operating at 47.2 Mg/hr (104,000
Ib/hr) steam. The screen analyses of tests 16 and 78 may then be compared.
It is seen that the western coal did not grind quite as well as the eastern
coal. However in the opinion of de Lorenzi {Ref. 1), free-burning coals
need not be ground as finely as coking coals, and this was not thought to
be a severe problem. An equally important factor in mill grinding capacity
is moisture. From Reference 1, frequently too much emphasis is placed on
grindability, while other factors such as moisture, which also affect mill
capacity, are almost entirely overlooked. The capacity of a pulverizer is
not directly proportional to the grindability of a coal. Correction must
be made for variation in fineness and moisture content.
Without quantitative analysis, it can be seen in Table III that the
moisture content of test 78 is a factor of eight times higher than test
16. The grindability of the other tests tend to follow the moisture content.
Test 57 with the highest moisture content exhibited the poorest grindabil-
ity, followed by tests 75, 73, 78, and 63 in order of increasing grindability.
The poorly pulverized coal burns more slowly resulting in lowered
heat transfer in the near-flame region (radiant section) and increased heat
transfer to the convective section. At high loads (tests 57 and 75), the
poor grind probably contributed to the excessive superheat steam temperature
problem.
EMISSIONS FROM ALMA UNIT NO. 3
A coal performance comparison for Alma Unit No. 3 is presented in
Table IV. In this table, western coal (test 66)., is compared to the nearly
identical eastern coal (test 9).
SO emissions
x
Significant differences in coal performance are noted for:
o
o NO emissions
o Carbon carryover
o Uncontrolled particulate emissions
o Unit efficiency
173
-------
For the first four items, the western coal performed better than the eastern
coal. Sulfur oxides emissions were reduced by a factor of 3 by substituting
western coal. At the same time nitric oxide emissions were reduced 24% and
carbon carryover was virtually eliminated. Uncontrolled particulate loadings
were reduced 33%. The performance of the ESP was not affected by the fuel
switch. It continued to operate at 99+% efficiency. Carbon monoxide and
unburned hydrocarbon emissions were generally less than 100 ppm each. In
the optimum furnace configuration, these emissions are controlled by excess
air. Below 3% excess O in the flue gas, these emissions became signifi-
cant. Soot formation, resulting in a "black stack" was also a problem
below 3% excess O . However, boiler efficiency was lower on western coal
due to high moisture losses resulting from fuel-contained water.
DISCUSSION
Figure 1 is a plot of nitric oxide as a function of excess O in
the flue for western coal at four loads. Figure 2 contains the same type
data for the base eastern coal. Both figures show increasing NO with
increasing O at a constant load; however, the absolute magnitude of NO
emissions from western coal is less at any given load and 0 level. Most
of the NO data on Figure 1 fall below the EPA limit for new coal-fired
units of 0.7 lb of NO as NO per million Btu (about 500 ppm). Attempts
X •"
to reduce the NO emissions of the eastern coal to these same (less than
500 ppm) levels resulted in high CO emissions.
Included in the factors that influence NO emissions are:
o Flame temperature
o Fuel nitrogen
o Excess oxygen
All three of these influence NO emissions when switching to western coal. The
temperature of the western coal flame is lower than the eastern due to the high
moisture content of the coal resulting in lower NO emissions from atmospheric
nitrogen fixation and to a lesser extent fuel nitrogen conversion. In general
the western coal could be fired at lower excess air before combustible losses
became a problem. This is due to the higher ratio of the volatile matter
174
-------
to fixed carbon content of the western coal which results in less solid
carbon to be burned out in the post-flame gases. The lower excess air
requirements result in lower NO emissions.
Table V contains data for all coals tested which show that the fuel
nitrogen content of the western coal is generally lower than eastern coal.
In any case, not all of the fuel nitrogen present in coal is converted to
nitric oxide. Typically, only 40% to 60% is converted with the amount
dependent on coal type, fuel nitrogen content, firing conditions, and the
structure of the nitrogen-containing molecule within the coal.
The substitution of western coal for eastern coal results in an
11% reduction in NO emissions on the average. The western coals used in
this comparison had 18% less fuel-bound nitrogen than the eastern coals.
The emission comparisons were based on western and eastern coal tests at
comparable loads and excess 0_ levels. Since NO arises from both conver-
2 x
sion of fuel-bound nitrogen and fixation of atmospheric nitrogen, it is
difficult to draw any correlation between fuel nitrogen content and NO
emissions. This is further influenced by the fact that different coals
have different tvpes of nitrogen-containing molecules which, depending on
their structure, are more or less easily oxidized to NO in the flame.
The conversion of fuel nitrogen to NO is a function of the structure
and distribution of the nitrogen-containing molecules within the coal. For
example, under certain conditions it could be important if the nitrogen
containing molecules are associated with the volatile fraction of the coal
rather than the fixed carbon portion. The chemical oxidation state of the
nitrogen species in coal is important since nitrogen that is partially oxi-
dized will be more easily converted to NO. For example, azide groups (N =
N) will more readily be reduced to N than -NH groups which will be more
easily oxidized to NO under flame conditions.
Figure 3 compares the NO emission behavior as a function of excess
O for three western coals on two pulverized coal units. The Fremont data
were taken on a 73 Mg/hr (160,000 Ib/hr) four-burner boiler while the Alma
data came from a 104 Mg/hr (230,000 Ib/hr) four-burner unit. These data
175
-------
indicate that the NO emissions are unit-dependent as well as coal-dependent.
Furnace volume, burner heat release rate, burner spacing, and fuel/air
mixing characteristics all have been found to affect NO emissions.
In order to control CO emissions from the eastern coal, it was
necessary to operate at higher 0 levels; this led to higher NO emissions.
For western coal firing, it has been shown that the furnace can operate at
lower excess 0 , thus lower NO. Western coal typically contains less bound
fuel nitrogen than eastern coal. This fuel nitrogen can be as little as
half the amount found in typical eastern coals.
The third factor affecting NO emissions is flame temperature. The high
moisture content of western coal causes the temperature of the western coal flame
to be lower than the eastern coal flame. This lower flame temperature lowers the
fixation of molecular nitrogen in the combustion air. The effect of flame temp-
erature on the conversion of fuel nitrogen to NO is not well understood.
Sulfur oxides emissions are largely a function of sulfur in the
fuel. There has been some work that indicates that coal ash composition
may affect the amount of sulfur oxides emitted (Ref. 2). The comparison
of the eastern and western coals at Alma (see Table VI) show the benefit of
fuel substitution in the control of SO emissions.
x
PARTICULATE EMISSIONS
Table VII contains particulate emissions data from the eastern and
western coals tested. The ash content of the coal and the combustible con-
tent of the fly ash emissions are given for a comparison of the maximum
potential emissions from each coal. For a 4-burner, 104 Mg/hr (230,000
Ib/hr) pulverized coal boiler at Alma firing on eastern coal, approximately
60% of the coal ash reported to the flue gas stream; whereas only 40% of
the western coal ash was found in the flue gas under identical firing con-
ditions and for coals with the same ash content. Electrostatic precipitator
efficiencies were unimpaired by the fuel switch. Combustible losses were
higher on eastern coal than on western coal. A 34% reduction in uncontrolled
particulate emissions was realized by switching to western coal.
176
-------
A 4-burner, 73 Mg/hr (160,000 lb/hr) pulverized coal unit was
tested at Fremont, Nebraska, on a Wyoming subbituminous coal and a Colorado
subbituminous coal. Again the Wyoming coal with a higher ash content had
less fly ash in the flue gas than the Colorado coal. However, the cyclone
dust collector efficiency was reduced to 72% on the Wyoming coal from 80.5%
on the Colorado which resulted in 110 ng/J (0.25 Ib/MBtu) greater controlled
particulate emissions.
177
-------
SECTION 4
STOKER-FIRED BOILERS
Coal firing of industrial boilers can be separated into two broad
classes — suspension firing and grate firing.
Suspension firing is normally applied in larger sized units, how-
ever, units as small as 16 Mg/hr (35,000 Ib/hr) steam have been built for
pulverized coal firing. Current economics indicate a break-even point in
the 91 to 113 Mg/hr (200,000 to 250,000 Ib/hr) steam flow range. Suspen-
sion firing includes both pulverized coal firing (70% through a 200 mesh
screen) and cyclone firing [crushed to 6.35 mm (1/4") with about 10% through
a 200 mesh screen].
Grate firing comprises three general stoker types:
o Underfed
o Overfed
o Spreader
Within these three types, there are a number of variations in feed methods
and grate design. Stoker-fired boilers have been built covering the entire
capacity range of this study: 4.5 to 113 Mg/hr (10,000 to 250,000 Ib/hr)
steam. The present stoker-fired boiler population represents a highly indi-
vidualized array of equipment.
Table II presented previously lists stoker types tested in this study.
From this assortment of units, the emissions and operating characteristics
of western coal firing have been determined.
OPERATIONAL CHARACTERISTICS OF WESTERN COALS IN STOKERS
i
Two properties of western subbituminous coals result in operational
problems for stoker-fired units. They are:
178
-------
o Coal weathering - resulting in size reduction
o "Free burning" characteristic - resulting in an
uncovered grate
Many older underfed and traveling grate stokers were manufactured with insuf-
ficient control of the undergrate air to use western coal as a fuel. The
problem is manifested when a dark spot of unburned coal develops on the grate.
This patch of coal can grow into a large clinker if special measures are
not taken to remove it. The "black patches" occur because there is insuffi-
cient local air pressure under the patch to maintain vigorous burning. The
loss of local air pressure occurs because some other portion of the grate,
in the same plenum control area, has become thin or bare and allows the com-
bustion air to pass through easily. These units were designed for an eastern
coal that formed some coke while burning and in turn maintained even coverage
of the grate. The free-burning western coals, on the other hand, tend to
form a fine powdery ash which either blows off or falls through the grate,
leaving it bare. This problem is compounded by the serious size reduction
that occurs while the western coal is in transit. The small coal particles
burn more rapidly when there is available air, however, when there is insuf-
ficient air they tend to plug the grate and fuel bed openings and form dark
patches which turn into clinkers.
The older underfed and overfed stokers designed for eastern coal
will have to have modifications to the undergrate air chamber to allow better
control of the air distribution, if western coals are to be used.
Spreader stokers are affected by the same coal properties but to
a lesser extent since approximately half of the combustion takes place in
suspension. This suspension burning reduces the number of "fines" that
reach the grate. However, the fines in the coal tend to burn close to the
spreader, sometimes flashing back into the feeder opening. This flash back
mode can be dangerous since there is the possibility of a fire in the coal
feeder. Coke and slag also have a tendency to build up on the spill plates
and rotor blades if the flash back is allowed to persist. This problem can
be alleviated somewhat by rotor speed and spill plate adjustments.
179
-------
The western coal performed well in the spreader stoker units. In
some units designed for eastern coal, the maximum attainable load was about
80% while on western coal. This was due to insufficient induced draft fan
capacity and as in pulverized units, high superheat steam temperatures.
Removing the major part of the moisture from the coal prior to combustion
would alleviate both of these problems .
EMISSIONS FROM STOKER-FIRED UNITS
Sulfur Oxides
The emissions of sulfur oxides from stokers is to a large degree,
governed by the sulfur in the fuel. These emissions are independent of load
and excess air in the flue gas. Table VI, presented earlier, contains the
results of a SO emission comparison for all the coals tested in this study.
The overall average SO emissions from actual operating industrial type
boilers decreases from 1827 ng/J (4.25 Ib/MBtu) on eastern coal to 619 ng/J
{1.44 Ib/MBtu) on western coal, or 66%. This is to be compared to the reduc-
tion as calculated from the fuel analysis from 2025 ng/J (4.71 Ib/MBtu) on
eastern to 778 ng/J (1.81 Ib/MBtu) on western coal, or 62%. The sulfur
content of the fuel was calculated from analyses of actual fuels burned.
The mineral analyses of the coals tested, given in Table 1, show
that the western coal contains a high percentage of lime, CaO; and magnesia,
MgO. The amount of sulfur trioxide retained in the ash closely approximates
the lime content of the ash in all cases. This suggests that the CaO may
tie up some of the sulfur as a sulfate. The western coal with its greater
lime content retains more sulfur than the eastern coal as indicated by the
data in Table VI.
Gronhovd, et al. (Ref. 2) have published a study of sulfur oxides
emissions from lignite-fired power plants. They found significant amounts
of sulfur retained by the ash. By using the following relationship, they
could satisfactorily correlate their data.
Sulfur emitted, as % of sulfur in coal =
*°
-12.7
Na O
- 48.1 —- + 110.1
180
-------
This correlation could not predict the amount of sulfur emitted in
the flue gas as SO when the subbituminous coals used in this study were
compared to the actual emissions. The Gronhovd correlation gave consis-
tently higher emission factors than actually measured.
Table VI contains the results of fuel sulfur analyses and SO
emission analyses for six western and eastern coals. For each coal, the
fuel analysis and the SO emissions analysis for the same tests were com-
pared. Where there was more than one test, the results were averaged.
The results show that on the average, the western subbituminous coal emitted
only 80% of the fuel sulfur available, whereas the eastern base coal emitted
90.4% of the available fuel sulfur under comparable boiler operating condi-
tions .
It can be concluded from this data that naturally reduced sulfur
emissions are influenced by coal type and are of a magnitude such that the
reductions should be considered when choosing a coal for reasons of SO
X
compliance.
Nitric Oxide
Nitric oxide emissions from stokers exhibited a similar dependence
on excess O in the flue gas as the pulverized coal firing. At constant
load, nitric oxide emissions increased with increasing excess O , as well
as increasing slightly with increasing load.
However, the slope of the NO vs. 0 curve is less for stoker-fired
units than for the higher intensity combustion devices. Figure 4 shows an
interesting NO vs. 0 result for a water-cooled vibrograte stoker. The
western coal {Wyoming Bighorn) has a slope of 12 (ppm NO/% O ) compared to
the eastern coal (Kentucky Vogue) which has a slope of 35 (ppm NO/% O ).
Figure 5 gives the NO vs. O plot for the same two coals on an overfed tra-
veling grate stoker without a water-cooled grate. On this unit, both coals
exhibit the same NO vs. O dependence. In fact, of the boilers tested, the
water-cooled grate was the only unit having different NO vs. O slopes for
the two coals tested. It is speculative as to whether the additional cool-
ing of the grate affects the conversion of fuel-bound nitrogen to NO.
181
-------
Stokers have overall lower NO emissions than pulverized coal units
since the stokers operate in a "staged combustion" configuration. The stoke -s
that have little or no suspension burning such as underfed and overfed stokers
have a greater degree of staging than do the spreader stokers. In the stoker,
the fuel devolatizes in the fuel bed under reducing conditions, then is mixed
with the combustion air above the bed. Mixing is provided by overfire air
jets or by front or rear arches in the furnace. Clinkering in the fuel bed
establishes a limit to the degree of staging that car be achieved on stokers.
Figure 6 shows these limits for a 72.6 Mg/hr (160,000 Ib/hr) steam spreader
stoker.
Carbon Monoxide and Unburned Hydrocarbons
Carbon monoxide (CO) and unburned hydrocarbons (UHC) emissions from
stokers as with all combustion systems, can be controlled by providing ade-
quate excess air and proper mixing to insure complete combustion. High
excess air conditions can cause CO and UHC emissions as well as too low
excess air. Figure 7 gives the results of CO emission measurements on a
45.4 Mg/hr (100,000 Ib/hr) steam spreader stoker as a function of excess
air for both eastern and western coal. At high load [40.8 Mg/hr (90,000 Ib/hr) J.
steam, CO emissions increase with decreasing excess air; however, at low
and intermediate loads, a point is reached where increasing excess air
results in rapidly increasing CO emissions. This behavior was observed
for both coals. At low excess air, CO results from inadequate mixing of
fuel and air. At high excess air settings, the fuel bed is thin even to
the extent of some uncovered grate area which is thought to lead to local
quenching of the flame by the combustion air and incomplete oxidation of
CO to CO . The western coal can be fired at 2% lower excess o at high
load while producing comparable levels of CO emissions. This translates
to higher unit efficiency because of lower dry gas and combustible losses.
Table VIII contains unburned hydrocarbon (UHC) emission data for an
eastern and a western coal on the same spreader stoker unit described above.
Unburned hydrocarbon emissions were higher at low load and high excess air
than at high load and normal excess air, thus following the same trends as
the CO emissions. No appreciable differences in UHC emissions were noted
between eastern and western coals.
182
-------
Carbon monoxide emissions are a much more sensitive measure of incom-
plete combustion than are unburned hydrocarbons. A comparison of CO emissions
and carbon carryover can be made. Figure 8 is a plot of percent carbon in
the outlet flyash of a 72.6Mg/hr (160,000 Ib/hr) steam spreader stoker firing
western (Montana) coal. This unit exhibited rather high carbon losses which
increased with unit load. The carbon losses on eastern coal were even larger
than for the western coal. However, the point to be made here is that by
measuring the carbon monoxide emissions, an indication of the other combust-
ible losses can be gained. The CO emissions for the same tests are shown
in Figure 9.
Particulate Emissions
Three types of stoker-fired boilers were tested. They were
o spreader stoker
o vibrating grate stoker
o traveling grate stoker
Within the spreader stoker category, four unit sizes and two grate
configurations were tested. The stoker at St. Johns was fitted with a dump-
ing grate while the boilers at Madison, Willmar, and Fairmont were all equipped
with traveling grates. The spreader stokers with their greater degree of
suspension burning and thin fuel bed have higher particulate emissions than
the mass fed vibrating grate and traveling grate stokers. The spreaders
are intermediate between the pulverized coal units and the mass fed stokers.
Uncontrolled particulate emissions from spreader stokers average about 858
ng/J (2 Ib/MBtu)-
In three of the four spreader stokers, western coal produced mark-
edly lower particulate emissions. In the case of Madison/ both eastern
and western coal produced the same particulate loadings although the western
coal had 16% more ash. The combustible content of the western coal fly ash
was half that of the eastern.
Dramatic reductions in particulate emissions were obtained on both
a vibrating grate stoker (65%) and a traveling grate stoker (31%) by switch-
ing to western coal. These units both have inherently low particulate emissions
because the combustion takes place in thick fuel beds with little or no sus-
pension burning.
183
-------
For a given ash content in the coal, the quantity of particulate
matter in the flue gas from stoker-fired boilers depends primarily upon
the amount of burning that takes place in suspension or on the grate.
Table VII gives an average flue gas particulate loading from both eastern
and western coals as measured before the control device for the stoker
types tested as well as pulverized coal-fired boilers.
Spreader stokers with the greater suspensiou burning have from two
to three times the particulate loading of the traveling rate and the vibro-
grate stokers. On the average, the western coal test results showed a 32%
lower particulate loading than the eastern coal.
In summary it can be concluded that there is a distinct advantage
from a particulate emissions standpoint, for switching to western coals.
BoilerEfficiency
Figure 10 presents boiler efficiency data as a function of boiler
load at two different coal-fired boilers each burning an eastern bituminous
low moisture coal and a western subbituminous high moisture coal. Data are
shown for a 45.4Mg/hr (100,000 lb/hr) spreader stoker with a traveling grate
and a 4-burner wall-fired, 104.3 Mg/hr(230,000 Ib/hr) pulverized coal unit."
The latter boiler was equipped with a tubular air preheater while the stoker
was equipped with a feedwater economizer.
An examination of the curves in Figure 10 suggests both boiler-to-
boiler efficiency differences and the importance of coal properties on the
efficiency characteristics of an individual boiler. The efficiency of the
pulverized coal boiler is greater than the stoker efficiencies over the load
range primarily due to higher excess oxygen and combustible loss character-
istics of the stoker boiler. Although dissimilar heat recovery devices are
used at each boiler {air preheater versus economizer) this has little impact
on the efficiency comparisons since stack temperatures were roughly equiva-
lent at both boilers.
Firing with western coal reduced the efficiency of the pulverized
coal boiler by approximately 5% while very similar efficiencies were exhibited
by both coals on the stoker unit. In the first case the shift in efficiency
184
-------
is attributed to the dissimilar moisture content of the two coals which
resulted in different "moisture" heat losses. Slight variations in excess
O levels, stack temperatures, and combustibles had no appreciable effect
on the other heat losses (dry gas loss and solid and gaseous combustible
losses).
At the stoker unit, a similar impact on efficiency would be expected
if coal properties were the only variable. However, in this case, the com-
bination of higher combustible losses experienced with the eastern coal
resulted in similar efficiencies. As a point of interest, the combustible
losses were less than 1% for all the pulverized coal tests shown, whereas
combustible losses on the stoker unit were 2% to 4% and 7% to 11% for the
western and eastern coals respectively. The absence of cinder reinjection
and combustion air preheat on the stoker boiler contributed to the rather
high combustible losses.
185
-------
SECTION 5
CONCLUSIONS
This study has shown that western subbiturrinous coals can be sub-
stituted for eastern bituminous coals as an industrii."1 boiler fuel. The
western coals are compatible with industrial coal-fired units of current
design. Two unit types of older design (underfed and traveling grate stokers)
were found to experience difficulty burning western coal. Some cases have
been noted where the maximum load capacity of the boiler had to be limited.
This problem can be eliminated by predrying the coal or by increased super-
heat steam attemporation capacity.
Western subbituminous coals were found to be superior to eastern
coals in terms of SO , NO , particulate, and unbumed hydrocarbon emissions.
A. X
The western coals could be fired at lower excess air and exhibited substan-
tially lower combustible losses than eastern coals.
The size of delivered western coal proved to be a problem in most
of the stoker-fired units tested. The coal generally had too large a per-
centage of fine coal which resulted from the poor weathering characteristics
of western coals.
Stoker performance on western coal could be improved if the coal
were sized locally at the point of use so that delivery distances could be
reduced to about 200 miles.
Boiler efficiencies on western coal were lower due to the high mois-
ture content of the western coal. The reduced efficiency due to the moisture
losses were somewhat offset by the lower combustible losses and lower excess
O required on western coal combustion.
This study has defined the operational parameters that must be fol-
lowed in order to successfully burn western coal in industrial-sized stokers
and pulverized coal units. Excess O and carbon monoxide monitors for
186
-------
combustion control would improve overall industrial boiler performance on
both eastern or western coal. These controls are necessary since many times
the margin of success can be as small as + 0.5% excess O in the flue.
~~" 2
For the most part, present instrumentation does not provide sufficient pre-
cision in combustion control. Operator training and education must go hand-
in-hand with improved controls for successful western coal firing.
187
-------
SECTION 6
REFERENCES
1. de Lorenzi, 0. (ed.)r Combustion Engineering. Combustion Engineering
Company, Inc., p. 7-8, 1947.
2, Gronhovd, G. H., Tufte, P. H., and Selle, S. J., "Some Studies on
Stack Emissions from Lignite-Fired Power Plants," Presented at 1973
Lignite Symposium, Grand Forks, ND, May 9-10, 1973.
188
-------
a
H
EH
g
U
Di
O
H
W
3
H
S
Q
2
1
•>si%)
,-ed
Orcanic
W -r* V
0 a> *>
U, o «
a* »«
5" a
-i c
^ *j Cl
(fl «CJ C
-r4 V Q
tn > ,0
.-* QJ 10
So u
< GC 1
.. 2 c
*» *S "^J 4J
g >• t7
^ £
U]
1
S. *>
x:
a " **• o
•-4 IA
u "5 fl S
O* C .-+ r-«
< O "*
Location
t*
o
u
(A o
**
f\! o* d -i d
o o o o o
o o o o d
•H O O • H O
»*>ir ^o i-« **i o Cft cno r-4 o~» *» tro CNM
fN^H IM'~' »>*»-< *N-«- (N ^H f^rHj fn'-'
^ ^f m rt r-«
ru m »H O O
^ »n ^ m v
tt>^a>cocotoco^)r*co
f*irtOOO(NMOOO
^r\5oia>trv^^«o«-4r^
C^gr-tor^oirt^^rN
•r O \D KD (T> in
i o »D *** t*- o a>
*N o^ eft in ^H o
l » • » * . .
i ™ «* ^T *•» r* ^
I «i
W t i 1 U
2 -. -- s ?
3 ^ $
a "
2 -
s
5 g f t! . g1
0) IB -H 4) ^
« s I 1 ' * Ist
3 S £ i |
a>l*r^p5r-ri"r"°
r* *-l |M rt ^H
•-* o* »o *-H t*t m *o ODC^ o
ft ** f4 ft fi ft n
a) v 4 O »-< O O
o £ o o o
CT* d tA CO O
d d ft o ft
lAf"*1 k0 co inr^ ^o*f IA i— t tO*c t*> co t*> to
SfM 00 itr Mr* r»cn cno ^^
p^ewotneof-irtf1*
^OfMrnr-^^^f1*
ipcocoajco^o.^
n tA O
^ •to o*
1 S
51 5"
y s
** < *J 4 -*
C i C C * C O
«t 4 U 4 C
2 g « x « -}
* £ s S «
O fnmm-H^cD^
fl N «-i r-i «*l
s s s s s s- a s
189
-------
6.2
o o o o o o o
r-o CD -* o*o>
CO r>* COIN
CN IN f*i tfi r» cn < , \o
Q *** O CO fS
-------
•p
c
0
H .-I
. .
CQ a
< -H
S H
u 9
« 04
o -^ o o o o o
tD^DOCOvOCD»Hr»-f>li-fChOCft™in'»iO
\Otninr^irvrJ4-»rMO^Hni
OOOOOOOOOO O O O --«
oooooooooc
191
-------
TABLE II. DESIGN TYPE OF UNITS TESTED AND
OVERALL PERFORMANCE ON EASTERN AND WESTERN COALS
nrt OF WIT TKSTED
I-ULVKXIZEU CUM, (Riley)
o 24) CJ/hr (210 Xlb/hri Stem
o Four Burner Face-Fired
9 TWO. Ball TlttM Kill P0lvericer>
a UOP ESP
VIBRATING CRATE ITOTXH. {Detroit)
e Hater-Cooled Grate
0 47 CJ/hr Hi KU>/hr> Steaa
o n> ran/Maturaj ID
o CiivJer Trap Partic. Renaval
TRAVELING CUTE STOKER ILeClcde)
„ *3 cj/hr (to r.lb/nr) suw
o rp Fan/Natural ID
a No Paniculate Controls
UNDERFED STOKE* (Meetinghouse)
0 Multiple Retort
0 10i 'CJ/hr (100 Ub/hr) £tee>
o cyclone Dost Collector
SPREADER STOKER (Detroit)
0 10' UJ/or (liO iUi>/hr> St«u
o Traveling GTftte
o Hulcicione Cyclone
o TD end ID rent
0 Superheat, CcononiKer,
Mid Air Heeler
SPREADER STOKER IWeeti.iahOuae)
0 105 CJ/hr (100 Klb/hrl Steon
o Traveling C»t*
0 ro end ID rune
o Superheet. Econonixer
SPKEMKR STOHER [Erie City)
o 84 CJ/hr (BO KU>/hi) St*»
o Traveling Crete
« n> end ID Feu
o Superheet
5PKI»DER STOKES (Kreler)
0 14 CJ/hr (I3.S Klb/hr) Steea
0 Dumping Crete
0 FD F*rl and HAturel ID *
o NO Perticulate Control*
SPRTADtR STOKER (tticke*)
o 11 CJ/hr (JO Klb/hr) SteeK
o Traveling Crate
o FD and ID Fane
o Sujicrhcab
o Cyclone Duct Collector
PULVERIZED COAL (94W)
v H« CJ/hr 1160 Klb/hrl Stea»
o 4-&urner tront-li'ircd
o fl.ill and Harv I'^lvrriket*
e Cyelono IXi-Jt Collector
OA.RALI, rcnt>H-:i: PATIIR
PF.R CnAL
CooJ
lantern
and
Weatara .
Eattern
Eesten
Vcittrn
and
Caatem
C««lcrn
Weetem
and
Cavtcm
VcBtem
Ejntera
l<uiin.)
and
Kosi.m
Stttblldkln.
Fair
Mcitem
Eactem
Me>tern
Blend
.
Kx,r
veetem
ta»tern
VnAcc«rr>t4blr:
Heatem
CfJviritTS
Reducitd pWHunii* capacity
Isprorved coal aixing
vould ia^rove perforvance
Severely affected by coa!
• ile
Specially tiled western
coal wea used far the
teat, however, the unit
would not respond to load
deiaand. Mortifications
are necessary to under-
arate air systen in order
to bum western coal.
Ka*i:«ai load reduced to
JOO CJ/hr {130 Klb/hrl
• teaa on western cdel due
to lilgn aupemeat tesfi^
eratures^ Large catbon
losses on eastern coal--
anoking.
Able to itaintain full
load on western coal
Blending problem due to
line* in western coal*
fcstem coal used ail
the tlaia with no
prob learn.
Ho prdblevuj with
western coal.
m>th the hit««inoue
ami llu1 MuUMtumlnout
CkVll* |X-»-|\>tTV*J WOll
on Hit* unit.
EASTCUl ^*~~~^~"
^-f^*~***~^ VKSTCRN
Neitern Kentucky ^r
IVogue) jS
.S Mycaiing
f Blend
j^V* So. 111.
SM6 1/3 Kontana
jf (Col atrip)
Dntuiown jr
jS Hyomlng
s/S (Big Morn)
jf Kuntana
Colorado jS
(Kaidenl ^r
S^ Kyo»lng
jT (Ikuui)
192
-------
TABLE III. SCREEN ANALYSES OF PULVERIZED COAL
Test No.
Load Mg/hr
Klb/hr
- 80 mesh, %
- 80 +100 mesh, %
-100 +140 mesh, %
-140 +200 mesh, %
-200 mesh, %
Moisture, %
EASTERN
9
56.7
125
0.65
0.75
2.75
6.95
88.90
5.15
16
23.6
52
1.00
1.00
3.35
8-20
86.45
2.20
WESTERN
57
77.1
170
2.90
2.30
20.65
34.60
39.55
22.05
63
59.4
131
1.40
1.30
4.05
8.30
84.95
12.75
73
42.6
94
0.65
2.32
8.91
32.26
55.66
17.94
75
77.6
160
8.74
7.70
21.65
16.75
45.15
19.33
78
49.9
110
1.53
2.23
6.60
17.46
72.18
17.75
193
-------
TABLE IV. COAL PERFORMANCE COMPARISON
ALMA UNIT NO. 3
Test No-
Western Eastern
66 ESP Inlet 9 ESP Inlet
Load, Mg/hr (Klb/hr)
59 (130)
3.4
996
372*
31
Participate! ng/J (Ib/MBtu) 2266(5.28)
ESP Efficiency, % 99.6
Carbon Carryover, % by wt 0.55
Unburned HC, at 3% O_, ppm 25
'if
Boiler Efficiency, % 85
Excess O , %
SO at 3% 0 , ppm
NO, dry at 3% O_, ppm
CO, at 3% O , ppm
<£*
59 (130)
3.4
3283
490 +
21
3411(7.947)
99.6
4.13
31
88.5
*223 ng/J (0.52 Ib/MBtu)
f296 ng/J (0.69 Ib/MBtu)
sUncontrolled
194
-------
s
o
a;
t,
o
2 w
H 1-3
x <
U O
H O
H
5 U)
W 3
W H
D S
fn 5
0^
H CQ
ca
w P
D W
Q
u> S
2 M
O fr>
H 10
W W
W 5
H
2 O
W H
O W
Z D
*i
O M
s
H D
O H
< M
IX (Q
w
E-i
w
§
EH
JK
* U
O "*
v in
tr
2 3
J (M
Kit
o
•9
•0
g
C
o
tr
o
55
3
U.
O
r>
±J
IB
•D
§
C
Cl
R
1,
3:
01
£
fe
*-7
"^ a
~
3
"I
1
3
t£?
J3
r-|
00
4
*-H *- (">
SU> S* i?S SS SS ™« 0(N
£ S S S S S £
irt vn ff* *n vc •*> M
5 £ ~ «
a* m o r* v m ui
--« -4 M •-< ^H o <^r
x t7* >, , n ^ p> n
X C X Jf v44JJC "-<
3KDW;f cos c«
C ^ c d> cm *HOI C 3 --« u
SO fij 3 W D -1 C «* O i-lig
> xtr «tr MO M c MX
•~t O O • -W X *4
a 3 s a s- .-
s S S ? s £ S
SS S3 PS SS SIN §!-» o'
0 0 O 0 0 0 O
T ^ ^» *n M Cl M
„„„„„„ 5
M^> to CD *r-- g^J JJ^ ^^ p; *
ci i d ^i CJ c £
O O O O ^H O O
'a
V W ** t^
2 ,— . "S "ol ^- ^^.^.HM
L> C -H •-* R C O
M O^i< QL| fflLr I7>M O* W t O *
c>. co e -w C4J co co "^ §
wir So* JJ^H JJ-H eo> eir c**
en O-H co co o "^ o -^ tt c **>
g2 f& % H £B ££. £•£ 3 z r5
o »H irt r*- o i^ »H
§ « —
— — -" O) (ft
a in — in
."• 0> V> « *—
— 0 — — U «
Up •* c V
0. — C W * 3 K
3 0 « ^ 9 .5
•^ *l flj v^ fQ 4J iQ
< M X * U « £
I
P
O
1
0
£ n
o
0
CO •
p
rj S
0
p
r* «
g.
2
<
195
-------
U
U
E->
U)
<*
U
Q
§
U
H
W
Qi
O
b
I
H
S
o
o
§
H
W
W
M
O
w
3
oq
in
c
0)
"e
u
o
(0
i-l *J
0} *H
3 S
(n U
tf> -
H W
1-3 X *
•-X O U<
CU13
n 4^
rH CO
Coal Source (Mine)
V
-H
in
to
r-tfcKUjcomoir-r-
SSoPSSoS?
>HH(Nr-lr--l
•SSSSPSSSS
r-ipoentOinOfM
rH tH
oSaSSSSSco'
fMHeNCNf<.-HfNO.;H
isfssfssl:
SSS.2P32S
OOOHOOrHO
0) 43 0)
o — - «— o) tn
u- "c -H -H 1: ^ o e it
i,HHH^Ka
>, O -U -U O O 1 -O >
{XA tn n^^ h
OF Id -H IB
5 n H o c c
*j u> 6 u ^j 5 5
id 3 -rt ^H Eg
E O V fH 9 0> U
^ iJ S -H 10 JJ H . H
o in o o o r» v
CO 00 CTH 01 C^ CD Cft
•-I f^ r^ CM 1/1 r«| 1T»
co o v in in f-il (N
_ — — ^- „ ^-| _
CTt ^f 00 f*^ lA **1 (s
to on in n co ca
<*> w *-i \o r> rg
O i-l CO <*> fO tO
rn C4 ^ r>< 01 >H
^ to in f* in n ^
to r- \o r- o t t-
to *v **> ^* in n ^
O ^ to o to r» «> >M IE 01
C Ol 41 —~ — -HO>
•rt W C 41 4) 18 «(
M 0 3 3 £ M
* 4J CP O* Itf W
^4 Q) W O O W >
JJ 3 - > > - 5
> tjl — —
-rj o ra i» H
K > -H >, >, -H ^H
— • •— o ji x o n
c o o £ ii
>i >i -H 3 p -H 41
J< Ji .H i> *• ^< >
0 OH C C r-t O
3 3 M Cl 01 i-i
-P 4J ii ^
c ' ct • •
qj ffi o • • O
>4 tZ u> » y ta
g
V, "3 C S
« ^ 1 3 2 §
Id 3 r-l .^ ^
E O »-l 3 -O -H
i-l -U — 1 nj ifl (0
•< «0 S H X b
8
3
xi
m
CD —
3
CM
l->
to
o
i
«
i
CJ
3
ID
03
£
O
O
in
o
en
S1
10 C
en 10
Oi --I
3 S
o
•o
01
§
•o
u
Q
A
T!
U
X X
S S
0>
S
196
-------
Q
U
E-i
g
O
§
ELI
03
H
S,
I
u
>
CQ
8
til
H
W
g
!-
, .«
. in <
m 3 _,
«i a -i <*
•H e o)
E O 3
U L> fc,
• U
ss*
•o
4>
1 i-f
C 0
tn —, O 14
C 3 y **
H X. fft
at & 2 -a
6 £ O ^
u — o o
c n
3 *J
^1
g A
8
A
5
» o
Emiss.-
Combus.
Fuel Ash
%
• *J
w c
=i n
w u
2
^
tn — , o to
C 3 0 4J
o *~r •*-*
-H x. m
w en E -D
W C X, | 4
m 5
«H O" X
«J C V.
t, .H
!
1
jj .
V "*
*j a
-H S
to K
« ^
< . I + 1 ill
in i M i i i-t v
trt
T l/> ^
CO M) 0
CD -H i i i o* r* i i£> ^
o i i i ™ . t r- •
mo -i O
o « — •
^^oSco r-S i ^,5^1
^""^r'|;*3 «' S U IS
H£ l£ si ^s ss 5S ss
O l> U O U 0
C C C > -* C ^-*
41 « 41 O «-t O -H
X N! K C H X M
C O ' O
sc * * D w at w
O^ CD ^ ^4 ^r <*i 0>
«.,.,.,.,
m
in i III ||
O CO O
N N
c* fN i i i o r** fi fN »*) o (N6^
o i i i (Nfcn'-«r~-or»i HV
s ^s^s^s' MS
SS O^r 3ri £?*r ^?— O? mSr 2*n
«B 2 d ,S -2 -i ' 2 S
,s ^ is |l 5i ,i s| si si
*"* '*-
*
* O C *3
4 V1 id O^ D**O H) O> 4J JJ
c c - , c c c ~ m c c MM
u>^ n ^H
g g
•H opouc 3 ^ 9P
Bo 3 o "O w >vi o y *-*< w o o -^ t/j
3ft, fl H 3 u* *J OT t4& — ( rt *J > «tf w
— U" X— W*- fc.— 3:*- W— (M —
r<*
1
'
h-
o ***
r- •
^ S
v —
O ^r
ss
r*
in O
O
sF
— 4
CD O
r« ^
'.«
AVERAGE
197
-------
TABLE VIII. COMPARISON OF HC EMISSIONS
FROM EASTERN AND WESTERN COALS
for 45.4 Mg/hr (100 Klb/hr) Steam
Spreader Stoker
Mg/hr
18
27
41
41
{Klb/hr}
40
60
90
90
Eastern Coal
02(%)
15.
12.7
9.7
8.7
HC
{cor ppm)
114
54
48
44
Western Coal
02(%)
13.8
11.3
—
8.8
HC
(cor ppnQ
125
18
—
44
198
-------
800,
700
600
500
rst
O
<#>
-U
10 400
o
z
300
200
(High CO)
O 40.8 Mg/hr (90
Q 27.2 Mg/hr (60 Klb)
A 59.0 Mg/hr (130 Klb)
O 77.1 Mg/hr (170 Klbf
10
12
14
°
Figure 1. Nitric oxide vs. oxygen - Alma Unit No. 3, western
coal. Fuel N = 0.79%.
199
-------
700
600
! 500
04
(N
400
Q
O
300
200
O 1
50
/
47Q
54 /BOOS 51
BOOS 40
(High COlQ
S 52
(High CO)
-J3OOS 41
Msmoke)
nl
III!
1 1
20 Q
O 90.7 Mg/hr (200 Klb)
A 59.0 Mg/hr (130 Klb)
Q 27.2 Mg/hr (60 Klb)
1 I
10
12
°
Figure 2. Nitric oxide vs. oxygen - Alma Unit 3, eastern coal.
Fuel N = 1.09%.
14
200
-------
700
600
Fremont (Wyoming)
at 59 Mg/hr
(130 Klb/hr)
500
O
of
Q
O
400
300
200
Alma (Montana)
at 59 Mg/hr
(130 Klb/hr)
Fremont (Colorado)
at 59 Mg/hr
(130 Klb/hr)
I
0
10
o2(%)
Figure 3. Nitric oxide vs. oxygen - comparison of 104 Mg/hr
(230 Klb/hr) four-burner PC boiler and 73 Mg/hr
(160 Klb/hr} four-burner PC boiler.
201
-------
250
E
O.
4J
nj
>,
)-i
Q
O
2
225
200
175
150
125
6.8 Mg/hr
.3 Mg/hr
Klb/hr)
Eastern Coal
(15 Klb/hr)
(25 Klb/hr)
(40 Klb/hr)
Coal
Mg/hr (15 Klb/hr)
Mg/hr (16 Klb/hr)
1 J
Excess 0,
Figure 4. Comparison of western and eastern coal nitric oxide emissions
(University of Wisconsin-Stout). Water-cooled vibrograte
stoker, 20.4 Mg/hr (45 Klb/hr) steam.
202
-------
350
300
R4
a,
. 250
-U
rd
200
150
8
12
/
/
/
Western Coal (Wyoming
Bighorn)
/
Load
/
14-16 Mg/hr (30-35 Klb)
7-8 Mg/hr (15-18 Klb)
/
*
Base Eastern Coal
Load
A 10-12 Mg/hr (23-27 Klb)
7-8 Mg/hr (15-18 Klb)
10
12
14
16
18
Excess O , %
Figure 5. Overfed traveling gra.te stoker, 27.2 Mg/hr (60 Klb/hr)
:. steam (University of Wisconsin-Eau Claire).
203
-------
12
11
0)
c
o
o
r~{
c
o
to
w
10
45
ID Fan Limited
Normal Operation
CO, Clinker, or Excessive
Smoke Limitation
1
51 57 63 69
% of Rated Load
75
81
Figure 6. Stoker firing staging limits, 72.6 Mg/hr (160 Klb/hr)
steam, western coal (Willmar Unit 3).
204
-------
2000
a,
CL,
O™ 1500
-u
its
O
u
0)
TD
•H
X
o
I
1000
I I I
40.8 Mg/hr (90 Klb/hr)
36-3 Mg/hr (80 Klb/hr)
O 27.2 Mg/hr (60 Klb/hr}
D 13.6 Mg/hr (30 Klb/hr)
Eastern Coal
. Western Coal
500
6 8 10
Excess Oxygen {%)
o
i
i
i
d -
14
16
Figure 7. Carbon monoxide as a function of excess oxygen (University
of Wisconsin-Madison) - 45 Mg/hr (100 Klb/hr) steam spreader
stoker.
205
-------
45
51
57 63
% of Rated Load
69
75
81
Figure 8. Percent carbon in outlet flyash, 72.'6 Mg/hr {160 Klb/hr)
steam spreader stoker (Willmar Unit 3), western coal.
206
-------
2200
1800
1600 —
45
51
57 63
% of Rated Load
69
75
Figure 9. Carbon monoxide emissions versus load (Willmar Unit 3),
western coal, 72.6 Mg/hr (160 Klb/hr) steam spreader
stoker.
207
-------
100
95
90
85
df
>,8°
U
c
-------
SESSION III:
SPECIAL TOPICS
DAVID 6. LACHAPELLE
CHAIRMAN
209
-------
-------
A SURVEY OF SULFATE, NITRATE, AND ACID
AEROSOL EMISSIONS AND THEIR CONTROL
By:
J. F. Kircher and A. Levy
Battelle's Columbus Laboratories
Columbus, OH 43201
and
0. 0. L. Wendt
University of Arizona
211
-------
-------
ABSTRACT
The objective of this analytical study was to evaluate the effects of
fuel and combustion modifications on the formation of primary acid aerosols
and their significance as combustion generated pollutants from large station-
ary sources. The term acid aerosol is used here in its broadest sense to
include all sulfates, nitrates, chlorides, and fluorides in all their forms.
Primary acid aerosols are those aerosols which are emitted directly from a
source or formed, primarily by condensation reactants, in the immediate vi-
cinity (0.5 mi); secondary aerosols, formed by reactions downstream in the
plume, are not considered. Available field data were collected and
interpreted in view of current knowledge of mechanisms of formation of
potential acid aerosols and their precursors.
Sulfates, nitrates, chlorides, and fluorides were considered initially;
however, based on the data available, only sulfates appear to make a,signifi-
cant contribution to primary acid aerosols. Thermodynamic calculation's bring
out significant differences to be expected in the emission of sulfates of
minor constituents from coals and oils.
The various combustion modifications for NOX control, including low
excess air, staged combustion, flue gas recirculation, and reburning, are
expected to have little effect on emissions of primary acid aerosols. The
exception to this conclusion may be firing with low excess air which has the
potential to abate both NOx and acid aerosol emissions. Combustion
modifications and fuel changes may lead to an increased formation of small
particles which could increase the formation of acid aerosols through various
heterogeneous reactions. Most of the effects are rather speculative due to
the meager data available. An important technological gap brought out in
this study is the lack of information on the specific sulfates being emitted
from stationary sources today.
This report was submitted in fulfillment of Contract No. 68-02-1323,
Task 49 by Battelle's Columbus Laboratories under the sponsorship of the U.S.
Environmental Protection Agency. The work covers the period from February to
September, 1976.
213
-------
-------
INTRODUCTION
In recent years there has been increasing evidence that sulfates in the
atmosphere may be of more concern as a health and environmental hazard than
sulfur dioxide. Part of this concern is reflected in the fact that S02 levels
in the atmosphere have been on the decline, while sulfate levels remain un-
changed^1'2'. Historically, it has generally been stated that only about 1
to 3 percent of the sulfur in a fuel is emitted from the combustion system
as 803 or acid. However, since such acid can lead to various sulfates which
might be a part of the particulate emissions, it is important to consider
these particulates as well as SOj as part of the primary acid aerosol. Further
as various combustion modifications (CM) become more widely applied to control
NOX emissions one must be concerned that these previously-held postulations
regarding SO-j and sulfate emissions are valid.
The overall purpose of this study was to survey, compile, and evaluate
data on the generation and emission of acid aerosols formed within stationary
combustion devices and in particular to determine what effect, if any, CM
might have on the formation and emission of primary acid aerosols and what the
ju
potential might be for CM to control primary acid aerosol emission.
Generally it is considered that the acidic sulfates, such as sulfuric acid
and ferric sulfate, are of greater concern as health hazards than the neutral
or basic sulfates. When speaking of "acid aerosols" in this study, however,
the term is used in the broadest sense. Acid aerosols refer here to any and
all liquid and solid particles containing sulfates, nitrates, chlorides, and
fluorides, as well as sulfates and nitrites of sodium, calcium, ammonia, etc.,
and all are of equal concern to this study and are included under the term
"acid aerosols". The study is concerned not only with total acid aerosol
* EPA Contract No. 68-02-1323 Task 49.
215
-------
emission but also their speciation and size distribution. It is important
to be so all inclusive in terminology at this time because of limitations
to specific relationships between health effects and individual sulfates,
nitrates, etc., and because of a current lack of specificity in chemical
characterization of particulate emissions.
Although an attempt was made to examine the production and emission of
sulfates, nitrates, chlorides, and fluorides in combustion, by far the great-
i
est emphasis was on the sulfates. This came about quite naturally due to the
instability of nitrates under combustion temperatures and the dearth of infor-
mation on chlorides and fluorides in combustion processes. As regards the
formation and emission of sulfates as primary aerosols, we have arbitrarily
defined as primary aerosol any particulate emitted from the stack and/or
produced within the first half-mile in the plume. The need for this definition
of primary aerosol is apparent since the chemistry of SC>2 oxidation in plumes
is specifically avoided in this study. The slow, secondary oxidation of S(>2
in plumes is covered by numerous other studies(3), Similarly we have not
stressed homogeneous gas phase combustion reactions in this review since they
have been summarized well by others (4,5). The secondary reactions are well
documented , and it is important in the context of the present study to re-
cognize at the outset that the emission of primary sulfates, nitrates, etc.,
is small relative to other pollutants or relative to second generation acid
aerosols.
One final comment by way of introduction: throughout this study there
was considerable uncertainty in correlating and interpreting such data as
was available because of the difficulties in sampling and analyzing for
acid aerosol components, especially as regards SC>2, 803, and sulfates.
This has led to a great deal of uncertainty in much of the data and has
made quantitative interpretation extremely difficult. A deep-rooted conclusion
that underlies this entire study is that to objectively evaluate the sulfate
issue additional work to enhance sampling and analysis capabilities is
urgently needed.
216
-------
ACID AEROSOLS FROM STATIONARY SOURCES
The great majority of emissions which may lead to acid aerosols are
sulfur compounds, sulfuric acid, SO-j, and sulfates, although it is recognized
that not all sulfates are acidic. Nitrates have not been observed nor are
they expected in stack particles, but a small amount of nitrate may be formed
in the near plume. The sparse information available on HC1 or chlorides is
in general agreement with basic thermodynamic considerations that the chlorine
in fuel will be emitted primarily as gaseous HC1 from the stack. Evidence
indicates that total primary sulfates (i.e., those observed within the first
half-mile) can be as high as 20 percent of total sulfur emissions or as low
as 2 percent.
A significant fraction of the primary sulfates consists of H2S04, based
on field measurements. Equilibrium considerations for coal-firing indicate
that the remaining sulfates are distributed among CaS04, MgS04, and ZnS04.
Bisulfates are not observed. Specific sulfates are not indentified but
field data suggest Na2S04, ^804 and FeS04 are also formed. For oil-firing,
NiS04 and Na2S04 are major components of the sulfated fly ash.
The composition and characteristics of particulate matter generated in
the combustion process depend on a wide variety of variables including fuel
composition, firing method, flame temperature, and amount of oxidant or excess
air. Coal and No. 5 and No. 6 fuel oil are the two general types of fuels of
most interest in generation of particulate emissions because of their
significant ash content and also potential for containing significant amounts
of sulfur. They also contain significant bound nitrogen, but nitrates are
relatively unstable and it is not surprising that we have uncovered no nitrate
particulate information in this study.
Coal Combustion
Most of the mineral matter in coal is made up of kaolinite, calcite, and
pyrites. The major elements found in the fly ash are from these minerals and
consist of mostly Si, Al, Na, Mg, K, Fe, and Ca. Table I illustrates concentra-
tion of the major ash constitutents for a variety of coals taken from the
general literature and also illustrates some typical measured concentrations
217
-------
of fly ash sulfur compounds, reported as 803, which are of major interest
in analyzing the impact of fly ash as a primary acid aerosol.
Generally, the sulfate level (reported as SO^) is on the order of less
than 1 percent up to about 2 percent. Two cases studied by Walker, however,
yielding sulfate contents of over 20 percent with the higher case being 24.2
percent
(6)
Bolton also recorded an unusually high sulfur content (reported
in this case as sulfur instead of 803) in one particular coal fly ash sample' '.
It is difficult to say why these high fly ash sulfur contents were observed.
They apparently had no relation to coal sulfur content as the two high readings
obtained by Walker were on coal of less than 1 percent sulfur while coal used
in the Bolton test had a sulfur level of over 5 percent. Also, of all the 13
coal/boiler combinations analyzed by Walker only 3 had over 6 percent sulfate
in the fly ash, as SO^, despite the fact that in all cases coal sulfur content
was less than 1 percent. Table II shows variations in the ranges of concentra-
tions of Na, Ca, Mg, K, and Fe in Walker's data for the 3 cases with high
sulfate compared to average high and low ranges for 7 cases with low sulfate.
Na, Ca, and Mg were consistently higher in the high sulfate cases whereas K
and Fe exhibited no clear trend.
Additional data on sulfate content in fly ash is shown in Table III.
Three of these cases show sulfate contents for two different locations in
the boiler. Where 863 was added ahead of the electrostatic precipitator^1^'
to enhance ESP performance, a clear increase in sulfate content was shown at
the ESP outlet. Three out of four readings taken at the air heater and stack
show a clear increase in sulfate concentration in the stack^ . A fifth
reading taken with soot blowing illustrates a clear increase in sulfate
concentration at the air heater but a low reading in the stack.
In addition to the major ash constitutents, fly ash also contains a long
list of minor or trace constitutents. Certain of these trace elements are of
importance in determining the potential for particulates formed in combustion
to affect the emission of acidic aerosols. The six elements identified as
having the most potential for catalytic effects for converting S(>2 to 803 in
the flue gas are V, Pe, Ni, Pt, Na, Cr, and Cu. The approximate concentrations
218
-------
of these elements as reported in 3 different studies are given in Table II.
The recent results of both Sheibely^14' and Abel and Rancitelli^15^ were
related to an NBS/EPA Standard fly ash sample. The results of Bolton were
based on results of fly ash samples taken from the Thomas A. Allen Steam
Plant
(7)
Two numbers are given for the Bolton results indicating measure-
ment by neutron activation analysis (NAA) and by spark source mass spectro-
metry (SSMS) .
The size range of particles resulting from coal combustion is an important
factor in estimating their pollutant potential in that it determines the
particle surface area available for contact with flue gases, hence affecting
adsorption rates with various gas components, and also the relative ease and
efficiency with which the particles can be collected.
The results of particle size measurements made at the boiler exit on fly
ashes from 69 pulverized coal fired boilers (IGCI/ABMA) indicate that for the
most probable distribution the mass median diameter of the particles is about
10 micrometers. This agrees well with data by Walker which indicates that
for 30 tests on pulverized coal fired boilers plus 2 cyclone fired units the
mass median diameter of the particles was from 5 to 15 micrometers '"' .
Stoker fired boilers tend to produce fewer small particles than pulverized
or cyclone fired units resulting in an overall larger size distribution. This
is due to the lower combustion intensity or heat release rate per unit volume
and also to the fact that the coal is burned in larger lumps or pieces. Data
on a particular traveling grate stoker fired unit indicates that in 21 tests
the mass median diameter ranged from 12.5 to 37 micrometers^ .
Oil Combustion
Particulate emissions from oil fired units result in much the same manner
as those from coal combustion. Ash or non-combustible constituents in the oil
form particulates both by ashing of refractory components and by condensation
of more volatile constituents. Also, there is the possibility of f ermine;
condensed carbon particles or soot depending on the combustion conditions and
degree of fuel atomization.
219
-------
Whereas coal ash typically can be at least 5 percent or more by weight
of fuel, ash in oil seldom exceeds 0.1 percent. Sulfur contents of oil are
also usually lower than those in coal but in the case of residual oil can
range up to 5 percent by weight of fuel. Based on analysis of up to 150
samples of residual oil over the 3 year period reported by Orr' , sulfur
ranged from a low of 0.29 to a high of 5.25 percent and ash ranged in content
from 0.004 to 1.9 percent (only one-tenth of the samples contained more than
0.1 percent).
Typical ranges of compositions of oil ashes for oils from different
regions of the United States and from overseas are shown in Table IV. Also
shown is the wide range of concentration of ash constituent in residual fuel
oil. Sulfates occur in oil ash in much higher concentrations than in coal
ash (values of over 40 percent shown in Table IV). It is noteworthy that oils
can contain significant amounts of V and Ni, elements that have been identified
as having a potential catalytic effect on oxidizing S02 to SQ$ thus having a
potential effect on sulfate emissions. This differs from coal ash where V
and Ni are present in ash at only a few hundred parts per million. Also the
nature of vanadium compounds in crude oil is such that they are stable up to
800 F. Hence they are not destroyed by refinery operations and as a result
they concentrate in the residuals' ^.
Carbon is also seen as a significant constituent in many samples, account-
ing for over half of the ash content in some cases. According to Novakov
carbon could be an important constituent in allowing fly ash particles to adsorb
S0.j from the flue gas forming acid constituents^ .
However, it is important to keep speculation about mechanisms in
perspective with regard to the formation of primary acid aerosols. In
general, N0£ and 803, precursors to acid, are found to be only several
percent of the NO and S02 in the combustion system effluent. Sulfate is
found, occasionally at high levels, in fly ash and when gaseous and parti-
culate sulfates are combined the total may be of the order of 10 percent
(23)
Nitrate is almost never observed. Chlorine in the fuels would be expected to
produce some HC1 in the effluent but there is almost no data on this point.
220
-------
However, HC1 has been observed'24)an^ ^n one case, for a power plant
burning pulverized coal, it was determined that virtually all the fuel
/251
chlorine appeared as HC1 in the flue gasv but the fuel chlorine is
generally low for most US coals. Thus, based on data presently available
to us, to the extent that there is a primary acid aerosol problem, it
is a problem of SOo and sulfate. The situation may be summarized by
considering the results of Cato'^'and Sommer' ' shown in Figure 1.
Where the ratio of SOj to (802 + 803) is shown for a variety of fuels with
a range of sulfur contents in a number of different boiler systems. The
apparent increase in conversion to 863 at low sulfur concentration may be
real or it may reflect the uncertainty of current analyses at low concentra-
tions.
Thermodynamic Considerations
Generally the data on particulates does not show the distribution among
species that might lead to acidic aerosols. Metals are generally reported as
oxides and sulfur as 803 or sulfate without identifying the molecular forms.
However, equilibrium thermodynamic considerations can be used to provide some
insight even though it is recognized that we are not dealing with an equilib-
rium system. That is, the observed species concentrations are characteristic
of the system at equilibrium at a temperature different from that measured.
In industrial combustion systems this generally means that the product distri-
bution of pollutants is characteristic of the systems at relatively high
tempertures, e.g., at the fire box exit, even though the measurements were
made at relatively low temperatures, e.g., in the stack. Thus, the composition
is kinetically controlled. This is most striking in the case of SO^/SO,
where S03 is generally observed to be only a few percent of SO in the flue
gas. At high temperatures most of the sulfur would be in the form of SO
at equilibrium. But as the equilibrium system cools, SO becomes the
predominant gaseous sulfur compound. That the predicted result is not
observed means simply that as the temperature drops the rate of SO oxidation
slows and the gases pass through the system before they reach equilibrium,
i.e., the chemistry is kinetically controlled. This is a gross over-
simplification of the many factors involved, of course.
221
-------
Consideration of the thermodynamic equilibrium chemistry can be instru-
tive since It describes the direction the reactions are moving. Figure 2
is the result of calculations based on combustion of coal of an average
composition (2 , containing 3.2% sulfur. The formation of both gaseous and
condensed species are included in the calculations and they are based on
burning with 10% excess air. Calculations were done for lower excess air but
are not shown since the results are almost identical. The figure shows the
distribution of sulfur among a variety of compounds. The various compounds
considered in this analysis are given elsewhere^ . At high temperatures
the sulfur appears almost entirely as 502- As the temperature decreases, the
system at equilibrium, S02 is converted to 863 and the two become almost equal
at about 800 K (980 F). However, the formation of solid sulfates also
increases and as the temperature continues to decrease the solids become the
predominant sulfur species and sulfur in the gas phase falls to very low
values.
Various gaseous species may contribute to. acid aerosol formation, parti-
cularly sulfates, and these are indicated in Figure 3, again for the same
coal composition with 10% excess combustion air. These are indicative of the
expected equilibrium flue gas compositions. It will be noted that sulfur
compounds dominate the situation from the standpoint of acid or acid precursors
until the temperature drops below about 700 K (801 F). Below this HC1 be-
comes the predominant species because at equilibrium the formation of solid
sulfates removes almost all sulfur from the gas stream. That is, at low
temperature equilibrium, one would expect most of the sulfur in the system in
the solid phase (ash) and most of the chlorine in the gas phase.
In the solid at equilibrium at about 800 K (980 F) the distribution of
sulfates, nitrates, chlorides, etc., is dominated by CaSO^. On the basis of
mole fraction in the solid the distribution of SO^ is
1.4 x 10~1 ZnSOA 2.6 x 10~3
CaS04
MgS04
1.4 x 10'
6.9 x 10
,-2
PbSOA
6.2 x 10~5
-3
222
-------
All other sulfates and sulfites are at the level of 10 . The sum of all
such species would be about 2 x 10"-* and include FeSO^, Na2S03, NaHS03,
and BeS04 in order of decreasing concentration.
A variety of chlorides also would be expected and together they would
total about A x 10 mole fraction with individual compounds in the range of
10~6. It is interesting to note that very little nitrate or nitrite appear
as products in these calculations with a total of only about 2 x 10~^ mole
fraction.
Fluorides also are present in the same concentration range and the total
of various compounds is about 2 x 10"-* mole fraction. Two phosphorous acids
also are expected, phosphoric (H3P04) and metaphosphoric (HP03), and their
total concentration would be about 9 x 10~" mole fraction.
There was essentially no difference between the 10% and 2% excess
combustion air cases calculated for all of these various compounds except
some slight general decrease with decreasing excess air, but the change is
hardly significant. At still lower temperatures than considered above
( <800 K), there is little change among the species except for Al2(20^)3
which at equilibrium would be the major species. This would occur at the
expense of gas phase sulfur (S02 atid 803), but may not happen in practice
because the kinetics become too slow. The rest of the species mentioned
above are essentially unchanged.
Based on many observations in a wide variety of combustion systems,
the composition observed at the stack is more characteristic of equilibrium
at higher temperature than the exit temperature. That is, the composition
appears to be "frozen" at the higher temperature distribution. If it is
assumed that above 1300 K (1800 F) the reaction system is at equilibrium
and below it the composition is unchanging, then flue gas compositions can
be estimated. On a molar basis for coal with 3.27% S and 10% excess combustion
air, the distribution of potential acid aerosol components would be:
SO? ^ 2000 ppm NO-TN
2 J
S03 % 200 ppm H2S02 £
-------
0.3 ppm
If the same coal burned with only 2% excess combustion air, then the expect-
ed distribution would show an increase in 802/803 ratio and a decrease in
both NO and N02- The expected distribution would be:
S02 * 2400 ppm H2SOM
803 ^ 70 ppm HFj
HC1 *v» 100 ppm N02 ^ <0.1 ppm
NO *> 10 ppm
If we made the same assumptions about rates for the solid components
and further, that the fly ash would contain about the same distribution of
compounds as the total solids, then we would expect in the fly ash, on a
molar basis: CaS04 i> 1%; MgSO^ ^ 0.1%; ZnS04 * 100 ppm.
The various other sulfates, chlorides, etc., mentioned previously would
appear at about the ppm level.
In a similar manner, the equilibrium predictions can be compared for
fuel oil combustion. Since we are still burning a hydrocarbon fuel, the
basic combustion parameters, particularly the temperature range, remain the
same. The major difference results from the changed distribution of im-
purities which can lead to acid aerosols. A distillate fuel will generally
have relatively low levels of sulfur, for instance, and very low levels of
ash, i.e., the metallic constituents which can lead to a wide variety of
inorganic sulfates as in the coal case. A residual oil, on the other hand,
will have comparable levels of sulfur, fuel nitrogen, and possibly ash.
Therefore the results of equilibrium calculations are similar to those for
coal.
Figures 4 and 5 summarize the results for a typical No. 6 fuel oil
/OQ 26)
composition * 'and similar calculations were also done for a No. 2 fuel.
The products included in the computations are the same as those used in the
coal computation except, of course, that those compounds involving elements
not included in the fuel are not included in the calculation.
224
-------
It should be pointed out that chlorine is not included although it is
recognized it may be present in some oils. However, the oil analyses
available did not include chlorine. The No. 6 oil contained about an
order of magnitude more sulfur than the No. 2 fuel and relatively large
amounts of vanadium, iron, and aluminum, as well as smaller amounts of a
number of other metals. Although calculations were performed for both
10% and 2% excess air, only the 2% cases are shown since the results were
nearly the same.
The distribution of potential acid aerosol species in the flue gas is
shown in Figures 4 for the No 6 oil. The major difference between the fuel
oil cases and that of coal, for equilibrium conditions, is that more of the
initial sulfur in the fuel oils remains in the flue gas at low temperature.
This results from the lesser ash and hence less formation of solid sulfates.
As a result of more sulfur in the flue gas at low temperatures oxidation
proceeds further and at the lowest temperature sulfuric acid becomes the major
gas phase sulfur species. For the No. 6 oil this could amount to about 0.1
percent of the flue gas at equilibrium, Figure 4. As mentioned previously,
however, the system does not come to equilibrium and the calculation represents
the potential for sulfuric acid formation, not a prediction of the amount
actually formed.
As noted in the coal case, there is little effect due to differences in
excess air. From the standpoint of combustion modifications, the maximum
difference would be expected from the situation in various types of staged
combustion wherein less than stoichiometric air is initially utilized. A
summary of effects on potential contributors to acid aerosols is shown in
Table V where 2% less than stoichiometric combustion air is compared to 10%
excess. It will be seen that at equilibrium at maximum temperature almost
all of the sulfur is in the form of S02- There is a large increase in 803
in all cases when the air is Increased but even so the SO^/SOg ratio remains
approximately 1000. The main effect in going from less than to more than
stoichiometric air is the increase in NO and N02« However, the ratio of NO/
N02 is again about 1000. In staged combustion, additional air is added in
some manner to complete combustion at a reduced temperature. For equilibrium
225
-------
calculations, the results at the reduced temperature and Increased air
are the same regardless of the distribution of products at the higher
temperature so that the calculated combustion system effluent Is unaffect-
ed by considerations of staged combustion.
The distribution of sulfur in the solid phase at low temperature is
what one would expect from the combustion of oil. For residual fuel oils
where the total metal content and the number of different metals are both
relatively high, the solids will contain a variety of sulfates. For the
No. 6 fuel oil used in this example, at low temperature equilibrium, one
would find aluminum, iron, magnesium, and nickel sulfates and sodium bi-
sulfate. However, in contrast to the coal case where total ash was much
higher, the solids in this example contain only a couple of percent of the
total sulfur available in the fuel as shown in Figure 5. Almost all of the
sulfur is present as uncondensed sulfuric acid-
Making the same 'assumptions of frozen compositions below 1300 K (1800 F)
as in the coal case, the expected concentration of acid aerosol precursors
in the flue gas for the No. 6 oil with 2 percent air and containing 2.8
weight percent sulfur are: S02 ^ 1500 ppm; SOj 'v 100 ppm; NO 'v 20 ppm.
The fly ash would be expected to contain about equal amounts of magnesium,
nickel, sodium, and calcium sulfates amounting to a total of a few mole
percent .
The No. 2 oil, containing 0.2 weight percent sulfur would be expected
to produce in the flue gas, based on the same assumptions: S02 ^ 90 ppm;
SO-j ^ 4 ppm; NO ^ 20 ppm. In this case there would be much less fly ash but
what there was. could contain as much as 10 to 20 mole percent N32SO. plus
A comparison of equilibrium considerations as they relate to observed
SO-$ production in practical boiler systems has been given by Hedley^and
is shown in Figure 6. The solid lines are theoretical equilibrium calcula-
tions for the conversion of S02 to 803 with either 10% or 0.1% excess air.
The dotted line indicates typical actual values at temperatures characteristic
of various parts of the boiler system. The horizontal lines indicate expected
temperature ranges for various parts of the system. This is consistent with
the overall views of mechanisms and kinetics. At high temperatures, point Y,
the formation of 803, is controlled by superequilibrium oxygen atom concentration.
226
-------
As the gas moves away from the flame and begins to cool the oxygen atoms
recombine or react and 863 concentration falls toward true equilibrium.
When the temperature reaches about 1200 K the kinetics become too slow for
SO-j concentration to f611ow equilibrium as the gases pass rapidly through
the boiler system. Though the concentration of 803 may rise somewhat it
will not reach true equilibrium.
EFFECT OF FUEL COMPOSITION
That the nature of the fuel obviously will have an effect on acid aerosols
and the increased propensity for sulfates with increasing fuel sulfur hardly
needs to be stated. However, there are essentially no field data that
indicate low concentration of trace species have a major effect on primary
acid aerosols. Although there are apparent correlations between sulfates
and certain metals in fly ash and deposits, it is difficult to state
unequivocally that the sulfate resulted from the presence of the metals.
Based on fundamental data and experiments to show catalytic activity
of fly ash constituents, it is to be expected that trace metals will have
a large effect on sulfate emissions, but existing data from practical
combustion systems do not allow trace metal effects to be quantified.
In the case of coal particulates, the major species are silicon and
aluminum, oxides which probably are inert. However, Fe, Na, K, and Ca
oxides are also generally present in lesser quantities and these have been
shown to have some catalytic activity for S02 oxidation under at least
some conditions. Different metal distributions among coals in conjunction
with different combustion modifications probably lead to variations in the
distribution and speciation of these metals in the particulate, therefore
variation in oxidation of S02 is to be expected.
Catalytic Effects of Fly Ash and Deposits
In the temperature range characteristic of the boiler convective heating
surfaces, heterogeneous reactions leading to increased oxidation of S02 can
occur. Several investigations have been made of the catalytic activity of
fly ash components in this regard. Fletcher and Gibson ^^^showed that for
227
-------
temperatures above 600 C (1112 F), Fe2(>3 greatly Increased the formation of
sodium sulfate from sodium chloride and 802, while at temperatures below
600 C, the sulfate was formed only from 803 already present. Thus Fe203 is
a strong catalyst in the oxidation of S02- These findings were confirmed
in the work of Vogel, et al, who found that high relative humidity helped
promote this fly ash catalysis
(31)
That materials other than Fe000 in a boiler system can be effective
(32)
catalysts for the oxidation of S02 to S0_ was demonstrated by Wickert
He found that although Fe-0 was the most active of the materials that he
tried, a sample of fly ash brought about_a-fflaximum Of 36 percent_
conversion of SC^; to 803 at about 760 C (1400 F). On the other hand, Si02
and A1203 were only weak catalysts in this system. What is most significant,
however, was the observation that these catalytic reactions were highly
temperature dependent as shown in Figure 7. The catalytic effect of the
^C2^3 *s Sreatest at superheater metal temperatures while that of the fly
ash goes through a maximum at a slightly higher temperature. The broader
peak with Fe203 indicates the greater importance of this compound as a catalyst
for the formation of $03, as it operates through a wider temperature range.
Obviously the surface area available and length of time for fly ash to spend
inside the catalytically active temperature window are of paramount importance
in determining the importance of fly ash catalysis.
Manganese dioxide is a powerful converter of S02 to sulfate over a wide
range of temperatures, from room temperature(33)to at least 340 C (644 F) '
and so may play a role throughout the entire convection and stack zone of a
power plant. High relative humidities (without condensation) are necessary
for the catalysis to proceed.
One of the best known catalysts for the conversion of 803 to 803 is V205.
Residual fuel oils from the Middle East and from Venezuela contain significant
amounts of vanadium and in the combustion process this is converted to V20c.
As a consequence, there is great potential for 803 formation by heterogeneous
catalytic reaction with V^O^-.ini the combustion of these oils. Wickert also
investigated the effect of 7205 and mixtures containing 7205 and other boiler
deposit components on the oxidation of 802^ . His results are shown in
Figure 8 in which the temperature dependence of the catalytic reaction is again
228
-------
apparent. In this case V20$ was a better catalyst for the reaction than
Fe^. A mixt:ure containing 90 percent 7205 and 10 percent Na2SO^ was just
as good a catalyst as the FejiOj. However, the boiler deposit from the burn-
ing of residual oil containing vanadium was the best catalyst of all, causing
a 90 percent conversion of S02 to S03. Catalysis by V205 also was examined
by Napier and Stone using short contact times with typical flue gas composi-
tions'-"'. The catalyst consisted of V^O- and I^SO^ on a silica support.
With 1000 ppm S02 in the gas stream and the catalyst at 430 C (806 F), from
94 to 98 percent conversiton of the S02 to 803 was achieved with contact times
ranging from 90 to 430 milliseconds. When the contact time was held constant
at 170 milliseconds, S02 conversions ranging from 92 to 98 percent were
obtained when the SC^ content in the gas stream was varied from -340 to 2700 ppm.
It was concluded that the required contact time for catalytic oxidation at
low S02 concentrations is much lower than that used in the contact process for
sulfuric acid.
The potential catalytic effect of deposits has been demonstrated further
by the work of Gleboy, et al^36'. They demonstrated that S02 could be oxidiz-
ed to 863 by molecular oxygen in the presence of boiler deposits over the
temperature range of 900 to 400 C (1670 to 752 F). The most active catalyst
powder was found to be a deposit from the convective bundles.of oil-fired
biolers. The deposit showed catalytic activity comparable to a vanadium pen-
toxide powder in the experiments. Maximum conversion occurred at a catalyst
temperature of 560 C (1040 F). Further, effective catalysts were prepared
from mixtures of V205 + Fe203 which had a maximum conversion at 640 C (1184 F).
Using empirical coefficients in conjunction with a simple model of a
boiler system, Glebov predicted 803 in flue gas as a function of deposit thick-
ness, convective bundle surface area and temperature, and excess air. The
predicted values were compared to measured values from a boiler burning a high
sulfur oil and good agreement was found. These results indicate that when
deposits which may contain vanadium pentoxide are allowed to build up, it is
possible that catalysis by the deposits could control the 803 effluent. Hence,
•»
heterogeneous catalysis by V205 in flue gases can well be an important source
of 803. However, V20s is effective .only in a high temperature "window", and
229
-------
this again poses the question of how combustion conditions affect the avail-
i
able surface area and time that fly ash particles•spend in this catalytically
effective high temperature window. Furthermore, V2dc is ineffective at room
temperatures, and so probably does not contribute to secondary sulfates.
As the simple sulfates remain in the deposits for extended periods of
time, they are gradually converted to complex sulfates by the action of the
sulfur oxides in the flue gas stream. The formation of alkali-iron tri-
sulfates such as Na^Fe(804)3'from the reactions of sulfur oxides with
and Fe203 was studied using radioactive tracer techniques by Krause, et
at 600 C (1112 F) and with 2500 ppm S02 and 30 ppm S03 in a gas stream contain-
in 3 percent 0(2. It was demonstrated that the reaction rate of SOo to form
the trisulfate was 970 times that with S02- A similar compound is formed
with potassium and in this case the reaction rate of SO^.exceeded that of 802
plus oxygen by a factor of 1260. Experience with trisulfate formation in an
i
operating boiler was reported by Anderson and Diehel, who placed a probe in
front of the superheater tubes^^. Gas temperatures in the region of the
probe were 982 to 1094 C (1800 to 2000 F) and the metal surface temperature
of the specimens was maintained at 566 C (1050 F). In this case the fly ash
collected from the boiler was found to contain 10.3 percent of sulfate expressed
as 863. The initial deposit collected on the probe after a week's exposure
contained 15.7 percent 803, and after several weeks time, the 803 concentration
reached 35 percent.
The significance of the formation of sulfates in'deposits and the build-
up of high concentrations of sulfates stems from the fact that portions of
these deposits are removed periodically from the boiler tubes by soot blowing.
This operation is carried out at least once a shift and by its nature creates
a large amount of particulate in the boiler in a short period of time. As a
consequence the capacity of the electrostatic precipitators is taxed during this
period and it is quite likely that a significant portion of ..particulate sulfate
passes through the precipitator and is emitted from the stack. Unfortunately,
virtually no data are available yet on these overload conditions which are
*r
potential sources of particulate sulfate in the atmosphere.
230
-------
Catalytic Oxidation of SOn by Soot and/or Carbon
The role of carbon as a catalyst to oxidize SC^ to SO/ has been reported,
primarily by Novokov^l^). He found that freshly generated soot or graphite
convert SC>2 to SO^ which is bound to the particles. Evidence that soot was
also catalytically active, especially in the present of water vapor was also
found. Futhermore, soot was active at only a specified distance down
stream from a flame. This distance may imply the existence of a tempera-
ture "window" or limitations of capacity to adsorb and react S0?.
The data shows that the decrease of SCL across a sooty filter is indepen-
dent of S0? inlet concentration, a fact which Novokov explains by
hypothesizing that the number of active sites on soot particles is controlling
rather than inlet species concentrations. Novokov also states that increases
in SC>2 oxidation occur at higher 02/C3H8 ratios, although none of these data
lie in the fuel lean regime. He attributes this to an increase in the number
of "ultrafine, high surface area particles" although it may be due simply to
increased 02 availability. In all his experiments there was a pronounced
saturation effect, implying a finite capacity for this process.
However, it cannot be determined from Novokov's work how much conversion
is possible through this mechanism. It is not clear whether particle age or
temperature is the determining factor for catalytic activity. There is at
present no demonstration that this carbon mechanism is not important, and so it
seems reasonable to conclude that the mechanism might account for significant
primary SO^ formation, especially when fresh carbon particles or soot are
formed. Clearly further work is required to quantify this effect and to deter-
mine its practical significant.
In summary, it is generally found that the sulfate in coal fly ash is less
than 2 percent, although occasional higher values are observed, and the sulfur
is largely on the surface and present as SOT rather than adsorbed 803. Studies
of the particulate surfaces suggest the sulfate may be present largely as iron
and/or calcium sulfates. Iron is a major constituent of fly ash along with
sodium and lesser amounts of calcium. Furthermore, since studies of deposit
chemistry have shown that ferric oxide can be an effective catalyst for S02
oxidation and sulfate formation, the fragmentary evidence available suggests
iron and its eventual distribution and speciation may be an important factor in
231
-------
the effects of particulates on primary acid aerosol emissions.
Particulates from oil combustion, particularly high ash residuals,
are in many respects similar to those from coal except that they are much
more likely to be high in carbonaceous materials and the total amount of
particulate will be smaller. The major difference is the presence of
vanadium and nickel oxides, which do not generally appear in coal particu-
lates, and are known active catalytic materials for SC>2 oxidation. This
occurence correlates with the generally higher sulfate levels found in
particulate from oil combustion.
Fuel composition in conjuction with combustion conditions is expected
to have a large effect on particle-size distribution. In general, the
evidence suggests that the more volatile metals will be concentrated in the
smaller particles and maximum flame temperture will effect the amount and
species volatilized. These small particles with their relatively large
surface area can be particularly effective catalysts for S02 oxidation and
sulfate formation.
In addition to catalytic effects, differences in size distribution and
speciation are also expected to effect the sorptive properties of fly ash
and deposits for S02 and 803 which can effect the eventual formation of
sulfates by noncatalytic mechanisms. However, at present it is virtually
impossible to quantify these effects.
SPECIFIC EFFECTS OF COMBUSTION MODIFICATION ON ACID AEROSOL
There are very little pilot or field test data which directly demonstrate
that a particular combustion modification employed to reduce NO and M>2 will
have an effect, good or bad, on primary acid aerosol. The weight of the
evidence is that anything which tends to reduce super-equilibrium oxygen atom
concentration in the flame zone will tend to reduce SO-j. On the other hand
if the production of particulate, especially very small particles, is increased
then the production of acid and sulfate solids might be expected to increase
through heterogeneous processes. At this time, conclusions regarding the effect
of a particular combustion modification on specific equipment must be highly
speculative.
232
-------
Staged Combustion
There is little experimental information from practical-sized equip-
ment on the effect of staged combustion on SO->. The pilot scale work of
(39)
Archer, et al., investigated two-stage combustion of a high vanadium
residual oil with 2.4% sulfur. Their results demonstrated that SCh can be
reduced essentially to zero when the first stage is slightly fuel rich.
They explain their results by noting from previous work that carbonaceous
particles inhibit S03 formation, react with S03, and physically adsorb it.
Such changes do not mean S03 is completely eliminated from the boiler,
however. When air is added at the second stage to complete combustion SO-j
may well be formed, as observed by Hedley^' , in excess of that which would
have been formed in single-step combustion with the same total air. Also,
heterogeneous reactions in the boiler section of the system may produce as
much S03 in spite of staged combustion. This position was summed up by
Schwieger (*°) •• t , .catalytic oxidation of S02 to 803 in the superheater
and reheater section generally is considered to contribute most of the 803.
Thus, there might be an unacceptably high level of SO^ at the air-heater in-
let despite an acceptable S03 level at the furnace outlet." Pilot scale and
more basic studies tend to confirm this expectation, for instance, the
previously discussed work of Glebov, et alP6'. However, there is not data
from practical systems which substantiate these heterogeneous effects when
staged combustion is used,
Flue Gas Recirculation
The situation is quite similar when flue recirculation is used. It is
'
well known that thermal NO and N02 are reduced by this CM ' , but there
is little direct evidence on S02 and S03. In one investigation Koizumi, et
ali ', in studying the combustion of a 2-1/2 percent sulfur heavy fuel oil
7 1
in a compact combustor (about 10 W/nr') , noted that the variable flame length,
for the excess air conditions used, decreased as recirculation increased up to
20 percent, then increased slightly up to 40 percent recirculation, before
starting a final decrease. Furthermore, the acid dewpoint (mearsured just
beyond the combustor) showed a parallel trend and correlated quite well with
233
-------
the flame length. The authors ascribe this effect to improved mixing.
Whether this acid decrease would be maintained in view of possible hetero-
geneous reactions in other parts of the system is questionable. However,
at this time there is essentially no data regarding SO-j from practical systems
employing this CM.
Low Excess Air
It is well known that low excess air is effective for reducing NO and
NC>2 and limiting acid in boilers. Basic studies indicate that as excess
air approaches zero the ratio 803/802 also approaches zero. Csaba and
Macfarland^ ^compute theoretically the values for various mixture ratios, for
specific fuel compositions and a range of product temperatures. They
demonstrate the expected increase in ratio of 803 to S02 as the excess air
increases. It should be noted that the effect of mixture ratio on the
803/502 ratio persists throughout the furnace in their calculations. These
results are consistent with "normal" conversion of SO, to 803 at this point1
which Gills reported as 0.2 percent to 2.5 percent^ .
Experience with oil-fired systems, where low excess air operation is
most practical at the present time, has demonstrated that this mode of
operation minimizes the formation of sulfates in deposits in the high tempera-
ture portion of the boiler, reduces the amount of sulfuric acid formed, and
eliminates the emission of acid smuts. Successful operation with low excess
air requires that the oxygen in the flue gas be maintained at levels below
0.2 percent. Such operation requires precise control of the fuel-air ratio in
all parts of the combustion system to prevent thermal cracking of hydrocarbons
i
and the emission of smoke. Consequently, low excess air operation has been
limited to oil-fired systems, because the technology for burning pulverized
coal with such little oxygen does not exist. Normal operation with 12 to 20
percent excess air results in the formation of 25 to 30 ppm 803 in the flame
with fuels containing 2 to 3 percent sulfur. The excess air must be less than
2 percent to decrease the 803 by about half. Further lowering of the excess
air results in a rapid drop of the 803 level, and at about 0.1 percent oxygen
in the flue gas the 803 concentration will be reduced essentially to zero.
234
-------
By redesigning the oil burners and exercising very close control on
the fuel-air ratio, Glaubitz in Germany was able to lower excess oxygen to
0.2 percent for routine operations. Under these conditions, the surfuric
acid was reduced to such an extent that the dewpoint approached that of
water. Glaubitz stated that after 12,000 hours of operation, the boiler
still did not have to be shut down for cleaning, indicating that the strong-
ly bonded deposits which build up as a result of the formation of large
amounts of sulfates had not developed in this boiler
(52)
Low Air Preheat
Lower air preheat is another change of imput conditions which lowers
the formation of NO and NCL. There is considerable information regarding
the lower preheat effect on the 803/802 ratio. However, as Glevov(36)
points out, "data on the influence of flame temperature on process of
formation of SO-j is very inconsistent. It has been firmly established that
in pulverized-fuel-fired boilers, the content of SO- in the gases decreases
—with increasing temperature in the furnace. However, Crumley, et al.'"'
on the basis of experimental data they obtained. . ." using kerosene and
distillate show an increase in S03 to a flame temperature of 1750 C (3182 F)
followed by a leveling off. The difference in the results from the two fuels
is considerably less than the difference in 2 percent sulfur in the kerosene
and 3 percent in the distillate. At 70 percent excess air with kerosene, about
7 percent of the sulfur was in the form of 803; at 28 percent excess air, about
5 percent.
, in agreement with Macfarland^50^, Csaba^9^, and Gudzyuk, et
shows with thermodynamic calculations that the S0,/S0? ratio decreases
with increasing perheat temperature. But Glebov's data on the combustion of
high sulfur fuel oil show a constant value of SO from 2100 to 2500 C (3800-
4530 F) for two values of excess air. Gudzyuk, et al., indicate possible
effects on SO- of high excess air regions near cool walls which might explain
some of the contradictory results. As discussed previously, SO- can be removed
from the flue gases by reaction with metal oxides to form solid sulfates, thus
reducing the SO- concentration in some regions. Or heterogeneous reactions
might increase SO- under some conditions. Therefore, it is virtually impossible
235
-------
without additional data to predict what the effect of lower air preheat on
S0» might be in a given system.
Load Reduction
Based on very meager data, it appears that load reduction has no
( 36}
significant effect on S0_ emissions. Glebov found no effect of load
on SO, over a range of 20 percent to 80 percent design load in his study
of high sulfur, heavy oil in an experimental furnace. In his theoretical
computations he found no change in going from 100 percent to 70 percent
load, assuming a catalytic activity of deposits equivalent to that produced
by Fe_0^, but some increase in SCL with decreasing load, assuming catalysis
£ J -J
by V205.
SUMMARY AND CONCLUSIONS
The great majority of emissions which may lead to acid aerosols are
sulfur compounds, sulfuric acid, SO., and sulfates; although it is recognized
that not all sulfates are acidic. Nitrates have not been observed nor are
they expected in stack particles, but a small amount of nitrate may be formed
in the near plume. The sparse information available on HC1 or chlorides is
in general agreement with basic thermodynamic considerations that the chlorine
in fuel will be emitted primarily as gaseous HCl from the stack. Evidence
indicates that total primary sulfates (i.e., those observed within the first
half-mile) can be as high as 20 percent of total sulfur emissions or as low
as 2 percent.
A significant fraction of the primary sulfates consist of H-SO,, based
on field measurements. Equilibrium considerations for coal-firing indicate
that the remaining sulfates are distributed among CaSO, MgSO., and ZnSO,.
Bisulfates are not formed. Specific sulfates are not identified but field
data suggest Na2SO,, K2S04 and FeSOA are aiso formed. For oil-firing, NiSO,
and Na.SO, are major components of the sulfated fly ash.
A significant effect of fuel composition on sulfate emissions should be
expected. Laboratory and field data indicate that as fuel sulfur level is
decreased, the fractional conversion to SO, is increased although total
236
-------
emission decreases. Moreover, based on fundamental data on catalytic activity
of various metals, it is expected that there will be large effects.on total
primary sulfate emissions.
In addition to affecting total SO," formation, fuel composition is
expected to have a large effect on emitted particle-size distribution. Vola-
tile metals condense to form very small nuclei, giving a high surface to
volume ratio and therefore accelerating catalysis. It should be noted that
volatile elements jnay not always form the most volatile combustion products.
Furthermore, small particles may increase the H_SO,/SO ~ ratio, especially
if large particles are removed.
Field data to support evidence for V, Ni, etc. , catalysis of SO,, oxida-
tion in fly ash are lacking. There is a scarcity of field data on speciation
of sulfate in fly ash, although a correlation exists between the Na, Ca, Fe,
and Mg content of fly ash and sulfate content.
Primary acid aerosols are formed by at least two general mechanisms:
(1) High-temperature homogeneous S0_ oxidation
(2) Dry gas-solid reactions converting SO- to SO, .
The basic mechanisms of the homogeneous reactions are reasonably well under-
(4 5)
stood and have been extensively reviewed by others ' . However, the effects
of heterogeneous reactions are much less clearly defined. Such reactions
have the potential to control primary acid aerosol formation but the actual
extent of their participation is not certain.
In general, the acid species are not destroyed once formed. Some may
be removed from the gas stream by adsorption on particles where they may be
in part neutralized and some particles are removed by precipitators, for
instance, but such processes are not completely effective. There is positive
evidence for the adsorption of sulfur oxides, metal oxides, and chlorides by
deposits and subsequent conversion of these compounds to sulfates. Soot
blowing removes some such material and probably overloads the precipitator
and results in emission of sulfates, vanadates, and other species. Although
field data are lacking in this regard, it is probable that such deposits
catalyze SO™ oxidation to some extent but carbon deposits probably do not
contribute to the catalysis at deposit temperatures.
237
-------
Where in the combustion system, and to what extent, primary acid
aerosol is formed is highly speculative at the present time. However, the
investigators best estimate is:
Flame zone .10 percent of total S converted
Convective pass 10 percent of total S converted
Stack 0 percent of total S converted
Near plume 1.percent of total S converted
Atmosphere 80 percent of total S converted
These estimates of conversion refer to that part of the sulfur in the
fuel contributing to aerosol emissions and do not include the sulfur retained
in the ash, slag, etc. That is, it is estimated that of the sulfur in the
stack effluent, up to 10 percent might be converted to acid aerosol consti-
tuents in the combustion zone. Similarly, another 10 percent may have been
converted in the convective passes so that up to 20 percent of the sulfur in
the effluent may contribute to primary acid aerosol. Probably 80 percent or
more of the sulfur emitted in the stack gases will be SO™, which will be
further oxidized in the atmosphere at some later time.
Finally, there is no evidence to indicate that CM (combustion modifica-
tion) will, in general, be an effective procedure for acid aerosol abatement
although low excess air firing, where practical, may be an exception. It
might be expected, however, that standard CM techniques for NO abatement may
X
adversefly affect the quantity, speciation , and particle size of acid aerosol
exhaust emissions through increases in the formation of fine particulates and
carbonaceous materials.
ACKNOWLEDGEMENT
The authors are pleased to acknowledge the assistance of A. A. Putnam,
D. A. Ball, H. H. Krause, and R. W. Coutant of BCL and J. M. Genco of the
University of Maine during the course of this study and also the continuous
support and encouragement of W. S. Lanier, Environmental Protection Agency.
238
-------
REFERENCES
1. Altshuller, A. P., "Regional Transport and Transformation of Sulfur
I Dioxide to Sulfates in the U.S.", J. Air Pollution Control Assoc.,
26, 318-324 (1976).
2. Squires, A. M., "Control of Emissions of Sulfuric Acid Vapor and Mist,
in Air Quality and Stationary Source Emission Control", National
Academy of Sciences. Government Printing Office, Serial No. 94-4
(1975), pp 458-473.
3. Levy, A., Drewes, D. R., and Hales, J. M., "SO- Oxidation in Plumes,
A Review and Assessment of Relevant Mechanistic and Rate Studies",
Final Report (Draft) from Battelle's Columbus and Pacific Northwest
Laboratories to EPA, Contract No. 68-02-1982 (February, 1976).
4. Cullis, C. F. and Mulcahy, M.F.R., Comb, and Flame, 18, 225 (1972).
5. Hedley, A. B., "Factors Affecting the Formation of Sulphur Trioxide in
Flame Gases", Journal of Institute of Fuel, 40. 142-151 (1967).
6. Walker, A. B., "Characteristics and Electrostatic Collection of Parti-
culate Emissions from Combustion of Low Sulfur Western Coals", 65th
Annual Meeting of the Air Pollution Control Association, Denver,
Colorado (June, 1974).
7. Bolton, N. E., et al., "Trace Element Mass Balance Around a Coal Fired
Steam Plant", Advances in Chemistry Series 141, Trace Elements in Fuel,
Edited by Suresh P. Babu, American Chemical Society (1975).
8. Davis, R. E., "Properties of Cements and Concretes Containing Fly Ash",
report to the Research Corporation, University of California at Berkeley
(June, 1938).
9. Bickelhaupt, R. E., "Electrical Volume Conduction in Fly Ash", Journal
of the Air Pollution Control Association, ^4 (2) (March, 1974).
10. Burton, J. S., et al., "Baseline Measurement Test Results for the Cat-Ox
Demonstration Program", Mitre Corporation Report to EPA, M-73-42.
Contract No. 68-02-0650 (April, 1973).
239
-------
11. Cuffee, S. T., et al., "Air Pollutant Emissions from Coal Fired Power
Plants: Report No. 1", Journal of the Air Pollution Control, 14, (9)
(September, 1964).
12. Cowherd, et al., "Hazardous Emission Characterization of Utility Boilers",
Midwest Research Institute, EPA 650/2075-066 (July, 1975), PB 245 017.
13. Oglesby, S., "A Survey of Technical Information Related to Fine-
Particulate Control", EPRI RP-259 (SORI-EAS-74-351) (April, 1975) PB
242 383.
14. Sheibley, D. W., "Trace Elements by Instrumental Neutron Activation
Analysis for Pollution Monitoring", Advances in Chemistry Series 141
Trace Elements in Fuel, edited by Suresh P. Babu, American Chemical
Society (1975).
15. Abel, K. H. and Rancitelli, L. A., "Major, Minor, and Trace Element
Composition of Coal and Fly Ash as Determined by Instrumental Neutron
Activation Analysis", Advances in Chemistry Series 141, Trace Elements
in Fuel, edited by Suresh P. Babu, American Chemical Society (1975).
Bradway, R. M., et al., "Fractional Efficiency of a Utility Boiler
Baghouse Nucla Generations Plant", EPA Report 600/2-75-013a (August,
1975) NTIS PB 246 641.
Orr, A. A. and Van Sant, R. W., Petroleum Products Handbook, McGraw Hill,
New York (1960).
Reid, W. T., External Corrosions and Deposits, Boilers and Gas Turbines,
American Elsevier, New York (1971).
Novakov, T., "Sulfates in Pollution"Particulates", U.S. Atomic Energy
Commission, Contract W-7405-ENG-48 (May, 1974).
Novakov, T., Chang, S. G., and Barker, A. B., "Sulfates as Pollution
Particulates: Catalytic Formation on Carbon (Soot) Particles", Science,
186. 259-261 (October 18, 1974).
20. Ellis, C., The Chemistry of Petroleum Derivatives, Reinhold, New York
(1937).
21. Thomas, W. H., Science of Petroleum, Volume II, Oxford University
Press (1938).
22. Huge, E. C. and Plotter, E. C., "The Use of Additives for the Prevention
of Low-Temperature Corrosion in Oil-Fired Steam Generating Units",
Transactions of ASME, 7£, 267 (1955).
23. Homolya, J. B., Barnes, H. M., Fortune, C. R., "Characterization of
Gaseous Sulfur Emissions from Coal and Oil-Fired Boilers", Fourth
Conference on Energy and the Environment, Cincinnati, October, 1976.
16.
17.
18.
19.
and
240
-------
24. Piper, J. D. and Van Vliet, H., Trans. ASME, 80, 1251 (1958).
25. lapalucci, T. L., Demski, R. J., Bienstock, Di, "Chlorine in Coal
Combustion", Bie. Mines RI-7260 (1969).
26. Cato, G. A., Muzio, L. J., and Shore, D. E., "Application of Combustion
Modifications to Control of Pullutant Emissions From Industrial Boilers
—Phase I", EPA-650/2-74-078a (October, 1974).
27. Sommerer, D. K., "Report of Information on Sulfur Emissions from Power
Plant Boilers", FEA Report No. Y-8833, September, 1975.
28. Ruch, R. R., Cluskoter, A. J., and Shimp, N. F., "Distribution of Trace
Elements in Coal", EPA Symposium on Environmental Aspects of Fuel
Conversion Technology, St. Louis, Missouri (May, 1974).
29. Levy, A., et al., "A Field Investigation of Emissions from Fuel Oil
Combustion for Space Heating", API Report 4099 (November, 1971).
30. Fletcher, A. W. and Gibson, E. J., "The Use of Carbon-14 and Sulphur-35
in Chemical Problems of Fuel Research", Proceedings of the 2nd Radio-
isotopes Conference, Oxford, England, Volume II (July, 1954), pp 40-48.
31. Vogel, R. F., Mitchel, B. R., Massoth, F. E., Environmental Sci. Tech.,
8, 432 (1974).
32. Wickert, K., "Chemical Reactions in the Combustion Chamber of a Slag-Tap
Boiler", Brennstoff-Warme-Kraft, 9_, 104-118 (March, 1957).
33. Corn, M. and Cheng, R. T. , JAPCA, 22_, 870 (1972).
34. Wickert, K., "The Behavior of Inorganic Constituents of Fuel Oils in
Steam Generator Furnaces and Gas Turbines", Brennstoff-Warme-Kraft, 11,
266-279 (1959).
35. Napier, D. H. and Stone, M. H., "Catalytic Oxidation of Sulphur Dioxide
at Low Concentrations", Journal of Applied Chemistry, JJ, 781-786 (1958).
36. Glebov, V. P., "Investigation of the Formation of SO with Combustion
of Liquid High-Sulfur Content Fuel in a Cyclone Chamber", Thermal Eng.
20, 51-54 (1973).
Glebov, V. P., et al., "The Formation of SO,, in the Combustion Products
of High Sulfur Oil Under the Action of Catalytic Active Boiler Heating
Surface", Thermal Engineering, _20, 23-28 (1973).
37. Krause, H. H., Levy, A., and Reid, W. T., "Sulfur Oxide Reactions:
Radioactive Sulfur and Microprobe Studies of Corrosion and Deposits",
Journal of Engineering for Power, Trans. ASME, Series A, 90. 38-44
(1968).
241
-------
38.
39.
40.
41.
42.
43.
44.
and
45.
46.
47.
48.
and
Anderson, C. H. and Diehl, E. K.> "Bonded Fireside Deposits in Coal-
Fired Boilers—A Progress Report on the Manner of Formation", ASME
Winter Annual Meeting, Chicago, Paper No. 55-A-200 (November, 1955).
Archer, J. S., Grout, P. D., and Eisenklam, F., "Multistage Combustion
of Residual Fuel Oil", J. Inst. Fuel, 43, 397-404, 451-460 (1970).
Schwieger, R. G., "Select and Evaluate Additives for Oil- and Coal-
Fired Steam Generators", Power, 118 (10), 54-59, 88-89; (12), 51-53
(1974).
Tomany, J. P., Koppang, R. R., and Surge, H. L., "A Survey of NO
Control Technology and the Development of a Low NO Emissions ,
Combustor", J. Eng. Power, Trans. ASME, 293-299 (19*71).
Turner, D. W., Andrews, R. L., and Siegmund, C. W., "Influence of
Combustion Modification and Fuel Nitrogen Content on NOX Emissions
from Fuel Oil Combustion", AIChE Symposium, Series 68, 126, 55-65
(1972). [See Combustion, 21-30 (August, 1972)].
Rawdon, A. H. and Sadowski, R. S., "An Experimental Correlation of
Oxides of Nitrogen Emissions from Power Boilers Based on Field Data",
J. Eng. Power, Trans. ASME, 165-170 (1973).
Muzio, L. J., Wilson, R. P., Jr., and McComis, C., "Package Boiler
Flame Modifications for Reducing NO Emissions", EPA Report R2-73-2926
(June, 1974).
Muzio, L. J., Wilson, R. P., Jr., and McComis, C., Package Boiler
Flame Modifications for Reducing NO Emissions, Phase II of III", EPA-
R2-73-292-b (June, 1974).
Armento, W. J., "Effects of Design and Operating Variables on NO from
Coal-Fired Furnaces, Phase 1", EPA-650/2-74-002a (January, 1974)?
Koizumi, M., Mizutani, H., Takamura, Y., and Nagata, K., "High Space
Heat Release and Low Excess Air Combustion of Heavy Fuel Oil Using
Exhaust Gas Recirculation Methods", Bull. J.S.M.E., 12, 530-538 (1969).
Barrett, R. E., Miller, S. E., and Locklin, D. W., "Field Investigation
of Emissions from Combustion Equipment for Space Heating", Data Supple-
ment, EPA-R2-73-084b (June, 1973).
Brown, T. D., Mitchell, E. R., and Lee, G. K., "Low NOX Combustion:
The Effect of External Flue Gas Recirculation on Emissions from Liquid
Fuel Combustion", Combustion Institute European Symposium, 487-492 (1973),
Brown, R. A., Mason, H. B., and Schreiber, R. J., "Systems Analysis
Requirements for NO Control of Stationary Sources", EPA-650/2-74-091
(September, 1974) PB 237 367..
242
-------
49. Csaba, J., "A Method for the Determination of Adiabatic Temperature and
Gas Composition Resulting from the Burning of Sulphur Containing Fuels",
J. Inst. Fuel, A7, 262-267 (1974).
50. MacFarland, J. J., "The Formation of Sulphur Trioxide in the Combustion
Products from Petroleum Fuel Oils", J. Inst. Fuel, j(5, 502-505 (1962).
51. Gills, B. G., "Production and Emission of Solids, SO , and NO from
Liquid Fuel Flames", J. Inst. Fuel, 4£, 71-76 (1973)? X
52. Glaubitz, F., "The Economic Combustion of Sulfur-Containing Heating Oil,
A Means of Avoiding Dewpoint Difficulties in Boiler Operations", Corabus
Combustion, 34, 31-35 (January, 1963).
and
Glaubitz, F., "Operating Experiences with Oil-Fired Boilers in the
Combustion of Sulfur-Containing Heating Oil with Lowest Possible Air
Excess", Combustion, ^4, 25-32 (March, 1963).
53. Crumley, P. H. and Fletcher, A. W., "The Formation of Sulphur Trioxide
in Flue Gas", J. Inst. Fuel, 29, 322-327 (1956).
54. Gudzyuk, V. L., et al., "The Effect of Operational Factors in the
Formation of Sulfur Trioxide When Burning Oil of High Sulfur Content",
Thermal Eng., 17, 39-42 (1972).
55. Kircher, J. F., et al, "A Survey of Sulfate, Nitrate, and Acid Aerosol
Emissions and Their Control", EPA-600/7-77-041, April, 1977.
243
-------
&
sa 33
M * «4
u ^i ^: -
5a s5
« -4 •*
• • • II
O U O I I
(MOO NO O
mm o -i'
<->_<
00 ^
i
Walker
•« a;
n n "
5 2 £Ji
55 s*
§^ £ j
•"SS
A > U
fNOM 4O O iHO
-» ON
•no
o
o
s:
__ ^
r» o.
^ I
O .•>
U X
r-l U
0 •**
A «
244
-------
H
| «
fcH
O to
U
fc, ,J
O Pj
w <
M CO
O
525 W
< W
S5
SE ^
0 J
r4 5
f/J
525 &
3
gn
W t22
jJHj <^[
M 35
o G
W
< O
P4
[^
O co
*£
0 W
CO JD
M l-l
3S
£g
O O
0 0
a
M
M
u
rJ
to
<
H
y*s
4-1
C
cu
o
M
0)
P.
V-'
01
00
c
fl
&
c
0
•H
4J
cd
r-l
4-)
C
0)
o
c
o
u
0)
00
2
u
<
o
,v
MM
0
>r
o
00
as
o -
«0
u
o
CM
. sr
I i
O ON
• *
\D CM"
00 ON
• *
CO ON
CM iH
1 1
«* CM
• •
rH m
CM fH
CM O
00 ON
• *
CM O
1 1
•» CM
oo in
• •
o o
CO
• \o
o
CM CM
I I
\O CO
* •
l-» rH
iH
£
00 >
•H 01 O 0)
X *J rH 4-1
cd cd
9) U-l M (4-1
01 rH 01 rH
CO 3 CO 3
co co cd co
u u
c c
CO >H !>. >H
o
z
M
PS
M
fo
^
o
o
1
w
PM
»
(A
<
{H
r4
Ih
rZ
M
CO
H
z
w
u
0
u
til
H
*
£
w
•
M
M
M
U
J
(O
<
H
u
C
01
5
j t
*j
c
O)
0
M
01
a
M
J3
W
-. U M
33 33
CO CO W CO
cd ctj CB ctj
£ S S S
^J ^* ^4 ^<
/~V
x>
XwS
1^ OO
co co
tH
CM vO
• •
CM rH
r- oo
* *
o m
rH
/•"v
a
«-v CM CO x-v
rH rH rH O
rH N— ' ^ rH
**S V*"
*t *
•O >N
01 M A C
01 01 CO O
u-i ,e o) u
14-1 > rH U
3 O 00 3
u u o m
•
d,
CO
M
UM
O
•O
(0
*N
CO -Q
V-X >>^
245
-------
ri
o
a
S3
s
M
W
O
s
u
&
tn
3
f».
If
O
w
w
i
K
•
f>
5
M
S
c:
«
o
tj
41
n
4J
c
5
i.
•H
4
1 C
1 *•
1 U
°
rH
O
o
vt
o
CM
tt
0
it
5^
O
'r*
K
m
O
CN
!>
O
l!
X
O
g
O
«
u
c*
o
r*
3
f*
o
41
tv,
n
o o
: s:
/-s
r-l
CM
^>
*
r^
,0
^x
H
.C
u d
4 0
•rl
-1 rH
° it
a |J
«
U Ck
- C r-l
IP U Tl
U ii O
S -•
v to
e, e. -H
It S !
CT) 10
0^» 1 1
*» ex
CM rH
m m
«»i
*M CA **l
• • * •
CO O m rH
0 *
in rH o •*
r*.
rH
1 * II
1 O II
^» cj
»-i i
*"\ r*V
N »
04 rH
w w
»
^^ *
U »-«
-O
t) o
So M
3 C
M n
>
•
***
g
w
41
U
g
U
vl
Vi
*J
8
S
•
rH CO
i-l Ou
° f
4 . H
H to
fl 4) TJ
Ct»C
o cd to
IM u
vl V) X".
ri °-
« T) O
O 4) H
U vx
g id u
O & 11) <*1
S O
^3 tt O * ^"
g rg H. « r-}
* " « « •*
SCB -O X T»
«> Vl rH C
%H O "M fd ^
rl C 0 0
Pi U rH « t*
f<9 tK O
rH (S
B a »• «
o o o o.
•j tj
8«w »4 H T>
9 «
n « 3 C
U «l V T> vl
M *O 3 vl rO
a 3 rH 0) B
1 e 5 s 5
**N y-1* r-X r*» r"S
« ,0 0 TJ «t
^r" W \^ %^ ^r
•
O
-------
u.
o
00
< f-
CM
H '-'
OS
M 0
< 0
oo
w
U3 t*j
U O
X
u w
erf
(•Jt* t—i^
O H
U W
Cz3 pu
Cn X
b (U
W H
>*
U
,J
EQ
<
H
O
as
CM
o
55
o
2
n
O
(A
CM
O
tfl
U
IM Wl
O *J
u d
cow.
01 -H -H
U X <
u u
(1) -rl
CLl O
IO
1 1
0 0
xx ! ! S !
0 •*»
r^ Q\
1 1 1
* S S S
! x ! x i x
00 CM so
• • •
in oo r-
IO f^ lO CO \f\ CO
II II II
o o o o o o
_J ^J -rl —J 1 ^J
r—i ti F-I (^ r^ r"t
KXw" X Xtn «X
^Nsroxa\cMMoocM
WJ »«O • «O • •
%*{ * *
!*• CM O
*
OOOvD r-400 NO tH CO f-
s-' 1 1 -H | | -H 1 1
oooooooo
i— 1 «H r-l t— 1 rH t-l i-H
Oj \O CM
O X X XX XX
o
C5 00 O CO ih-t Q |*^« C*^
r^ CM r*« CM sr r-<
c*> co f) f> •* in
II II II
O O O 0 O 0
XX XX XX
r-. >» r-. m r-t oo
« • * • • •
CM CM t-t r-i r-t O\
O O O
CM rH CM iH CM i-l
1 + 1 4- 14-
g
t^
Ui
V
•-I
•»(
-------
O
4)
a
£
c
tomizatio
<
T3
a
1
L. a> T3
g> .
2. ° * to o
CTCJ & -g o
— . . . Q. 3 C >»
poo O in Q. 3 o
•"• *^ "^ ^^
-}Z2 'I
<5>
r e
X
0°
DO e
.
X
O
^
d
^a
x ^
O
~
4
^V
*s
4^
>
X
D®<^c
^
e
^^
^^
3
«.
9
©
K^
<^
S/t
^
<
q q " o q <~J~" q o q o <
CM
1O
8
CO
CM
I1
„£
K>
O -J;
o o
8 ^
t3
C
.2
o 2
S e
o
>
L
!22si2®«>*cMC>
3 3
-------
439 710 900
2780
10'
AI2(S04)>
FeS04U)
I i I
JiOO
700
900
1100
1300
1500
1700
1900
Figure 2. Kquil ibrliim sulfur products for coal combusLlou with 10 pc-rccnL
excess air, 3.27 percent sulfur
249
-------
439 710 900
2780
500
700
900
1500
1700
1100 1300
°K
Figure 3. Equilibrium flue gas components for coal combustion with 10
percent excess air, 3.27 percent sulfur
1900
250
-------
439 710 900
2780
10
500
FIgu re
700
900
1100
1300
1500
1700
1900
°K
!••')u (1 1 hr i urn flue pns components for fl(> oil cntnlxiRt Ion witlt
2 pert-cut CXCUHS air, 2.80 percent HiiLTur
251
-------
2780
10
500
Figure 5.
700
900
1500
1700
1100 1300
°K
Equilibrium sulfur products for //6 oil combustion with 2
percent excess air, 2.80 percent sulfur
1900
252
-------
-~ INCREASED TIME
1600 KOO 1200 1000 800
TEMPERATURE. °K.
600
400
Figure 6. The variation of the theoretical equilibrium yield and possible
actual yield of 803 with time in a boiler.
253
-------
c
V
o
5
a
*
x>
a>
M
•a
K
O
400 600 600 .000 .200 1400 1600 .800 2000 2200 2400
Temperature, F
Figure 7. Catalytic oxidation on S0? to SO. by
various materials
254
-------
255
-------
-------
INVENTORY OF ATMOSPHERIC EMISSIONS FROM
STATIONARY POINT SOURCES
By:
V. E. Kemp and 0. U. Dykema
The Aerospace Corporation
El Segundo, CA 90009
257
-------
-------
ABSTRACT
This paper describes the first and second years of a stationary source
combustion-related atmospheric emissions inventory being conducted by The Aerospace
Corporation for the EPA. This is a 3-year study aimed at assisting in the establishment
of priorities for detailed studies of techniques for the control of combustion-related
emissions from stationary sources. The inventory includes emissions of oxides of
nitrogen, unburned hydrocarbons, carbon monoxide, and particulate matter, not only from
recognized major stationary combustion sources, but also from other stationary source
categories in which combustion plays a secondary role. During the first year of the
study, emissions were established for 1975 and estimated for 1980 from boilers, internal
combustion engines, chemical manufacturing, and petroleum refining. During the second
year, emissions were calculated for 1976 and 1981 for primary metals and hydrocarbon
evaporation, as well as for the four industries studied the first year. This paper
identifies approximately 68 percent of particulate matter and 93 to 97 percent of
nitrogen oxide, hydrocarbon, and carbon monoxide emissions from stationary point
sources. The third year of the study will raise the inventoried emissions to greater than
90 percent for all four pollutants. The emission rates were calculated based on process
usage rates, emission factors, and time rate of change of these variables for the
respective categories of stationary source of emissions. The usage rates and their slopes
were obtained from agencies such as the U. S. Department of Commerce and the EPA
National Emissions Data System. The Emission Factor values and slopes were based on
data extracted from various reports reflecting either empirically or theoretically derived
emissions.
In addition to estimating the annual nationwide emission rates of the four
referenced pollutants, the uncertainty of those rates was established. Statistical
engineering estimates, current and potential legislative environmental controls, and
several independent sources of data were considered in calculating these uncertainties.
259
-------
-------
SECTION 1
INTRODUCTION
For several years Aerospace has been studying control of combustion-related air
pollution from stationary point sources; specifically those of oxides of nitrogen (NO ),
A_
hydrocarbons (HC), carbon monoxides (CO), and particulates. Since current methods of
control of oxides of sulfur (SOJ are not combustion-related, air pollution by SOV was
A A
not studied. In order to put combustion-related air pollution in its proper perspective, it
was also necessary to study the major non-combustion point sources of the four air
pollutants. A point source of air pollution is defined by the National Emissions Data
System (NEDS) as a single point of discharge of more than 100 tons of a given air
pollutant per year. The aggregate of large numbers of air pollution sources which are too
small to qualify as point sources are called area sources. Although area sources were
specifically ruled out of this study, in some cases the possible nationwide pollution
contributed by certain area sources was very large compared with the point sources.
Some effort was made, in such cases, to at least point out this potential.
The basic source of data, and the model for the data cataloging and reference
system was the NEDS. The NEDS system, on the average, tends to represent data from
the 1970 to 1974 period and does not provide a means of updating or projecting into the
future. The NEDS, however, does represent the largest single nationwide emissions data
base available. In order to provide data on which decisions can be based regarding
allocation of resources for research in control of combustion-related air pollution, it was
considered necessary that the emissions be projected at least 5 years into the future, and
that some estimate be made of the uncertainty of the resulting projections. As a result,
the NEDS data were used as one source of relatively current data, particularly for the
distribution of emissions between various detailed sources within an industry, but other
surveys and analyses were studied as well to develop means of predicting changes in the
fuel or process usage rates and emission factors in the near future.
261
-------
The primary data stored in the Aerospace computer data base are: (a) the best
estimates of current annual charge rates and emissions factors; and (b) the probable
i
linear slopes, or changes with time, of these two parameters into the near future.
Engineering estimates of the range of uncertainties in the charge rates, emission factors,
and the slopes of each are also included. The computer program, then, can project usage
rates and emission factors to any date in the future, along with the uncertainties in those
projections. Total nationwide emissions at those future dates are calculated from the
product of the projected charge rates and the projected emission factors. Because of the
uncertainties in the slopes, of course, the uncertainties in emissions projected far into
the future become so large that the projection becomes useless. Figure 1 shows an
example of projection of the best estimates and uncertainties of the annual charge rate
and the emission factor and the resulting projected emissions of, in this example, NO .
A
The first year of this study included the categories of industrial and utility
steam boilers, stationary internal combustion engines, chemical manufacturing, and
petroleum refining(l). The second year added the categories of primary metals and
evaporation (2). The third year, nearly complete, adds mineral products, secondary
metals, and wood products, as well as including a short study to update the rapidly
changing category of steam boilers (3). These nine categories of air pollution sources are
divided into more than 400 sub-categories, in five levels of primary and summary data.
Each primary data sub-category requires 40 separate data entries to describe charge
rates, emission factors, slopes, and uncertainties for the four air pollutants.
Table I shows the scope of the inventory. All data in Table I except the
distribution of the four air pollutants among the stationary point source categories were
obtained from a summary of NEDS data run in 1976. Although the NEDS data contain
updated emission factors, they are based on process usage rates of the 1970 to 1974
period. The distribution of emissions between the stationary point and area sources and
the mobile sources, therefore, generally reflects that time period. Those data are shown
here only to orient the scope of the subject inventory with respect to all other sources.
This inventory, then, restricted to stationary point sources, addressed the emissions from
(approximately) as little as 30 to as much as 80 percent of all of the artificial sources of
the four air pollutants.
262
-------
SECTION 2
RESULTS
The data shown in Table II for the category of stationary point sources were
developed in this inventory for the year 1976. Data are shown both for the total of all
stationary point sources and for those sources which are combustion-related. Within the
category of stationary point sources, steam boilers (utility and industrial) clearly
dominate the total of NO and particulates emissions. This is not necessarily because
A
these sources are particularly dirty but because the magnitudes of the fuel usage rates
are so high. This category includes the generation of electricity and all other industrial
process uses of steam but excludes steam heating of commercial and residential
buildings.
Similarly, evaporation of petroleum products, surface coatings, and cleaning
solvents (non-combustion sources) dominates the emission of hydrocarbons from
stationary point sources. Major emissions of carbon monoxide are shared primarily by
the petroleum and the primary metals industries.
Table II also shows that, perhaps as expected, NO and CO emissions nearly all
A
result from combustion. About half of the particulate emissions result from combustion,
largely from coal combustion in steam boilers. Since the combustion processes, to be
efficient, should oxidize all of the hydrocarbons in the fuels, little of the HC emissions
results from combustion.
Figure 2 shows the trends in the emissions of the four air pollutants over the
period 1976 to 1981, as projected in this inventory. All are shown to be decreasing,
nominally from as little as 13 to as much as 34 percent. Figure 2 also shows the
estimated ranges of uncertainty in the data and projections.
Table III shows a more detailed breakdown of the major stationary combustion-
related source of NO and particulate emissions (i.e., steam boilers). NO from
* X
stationary point sources is clearly dominated by utility and industrial boiler combustion.
263
-------
Bituminous coal combustion currently contributes about three-quarters of the NO from
steam boilers, both because nearly two-thirds of the heat input to steam boilers comes
from bituminous coal combustion and because the EPA NO regulation for coal-fired
A
utility boilers is 2 to 3 times those for oil- or natural gas-fired boilers (0.7 versus 0.3 and
C
0.2 Ib of NO™ per 10 Btu heat input for coal, oil, and gas, respectively). By 1981
bituminous coal is expected to contribute more than 92 percent of the NO emissions
A
from steam boilers and nearly three-quarters of the NO from all stationary point
A
sources.
Table HI also shows that nearly half of all of the particulate emissions from
stationary point sources currently result from steam boilers and 94 percent of this results
from bituminous coal combustion in these boilers. The small increases in some of the
particulate source percentages are largely due to a small decrease in the total of
particulate emissions from all stationary sources over that time period.
Figure 3 shows a further breakdown of the NO emissions from utility boilers
A
firing bituminous coal projected for the year 1981. As expected, conventional pulverized
coal boilers dominate the NO emissions, but, perhaps unexpectedly, NO emissions
X X
from cyclone-type boilers remain high. This is largely because the NO emission factor
A
for a cyclone furnace is more than twice that of the average for bituminous coal-fired
utility boilers. Thus, although approximately a 20 percent improvement in the emission
factor is projected and fuel usage in cyclone furnaces is expected to remain low (14
percent of total bituminous coal combustion in utility boilers), the NO emissions from
this type of boiler are expected to remain significant.
The next largest sub-category of NO emissions from bituminous coal-fired
utility boilers is the tangential configuration. This results from the large fraction of this
coal which is burned in utility boilers of this configuration (39 percent of the bituminous
coal) rather than from a high emission factor (53 percent of the average).
Figure 4 shows a similar breakdown of particulate emissions projected for 1981
from the predominant combustion source, utility boilers burning bituminous coal. Again,
conventional pulverized coal-fired boilers represent the largest fraction of particulate
emissions from this source category (91 percent). Particulate emission factors, however,
are much more a function of the control equipment used on the coal-fired boilers than of
the firing type. As a result, the fractions of the total particulate emissions represented
264
-------
by each sub-category tend to represent the fraction of coal burned in boilers of each of
the firing types.
The major contributions to carbon monoxide (CO) emissions from combustion-
related sources are shown in Figure 5. More than half of this CO results from processes
involved in the catalytic cracking of petroleum during the refining process. The related
combustion process is the periodic regeneration of the catalyst by burning off the coke
(with air) which becomes deposited on the catalyst. There is a great deal of uncertainty
in both the current and projected levels of CO emissions from this source (±38 percent in
1976 levels and ±68 percent in 1981 levels). This results largely from the great disparity
in current emission factors obtained from various sources and from the lack of data from
which to project the rate at which refineries will be modernized in the future and
brought into compliance with new standards of performance. The best estimate of this
inventory for CO emissions from petroleum refineries for the year 1976 was higher than
those reported in the NEDS summary by more than a factor of five. This discrepancy,
and the large uncertainties, have not been adequately resolved.
The CO emissions from blast furnace and basic oxygen furnace operations,
shown in Figure 5, are included as combustion-related sources simply because they
represent the high temperature oxidation of carbon. Carbon black manufacture involves
fuel-rich combustion of natural gas and oil, to form the carbon black. As a result, the
off-gases are rich in unburned hydrocarbons and CO. While a great deal of effort is
made, primarily for reasons of process efficiency, to capture the hydrocarbons, CO
emissions are essentially uncontrolled.
Figure 6 shows a breakdown of the major stationary point sources of total
hydrocarbon emissions. The figure shows that nearly two-thirds of these emissions result
from non-combustion-related sources, principally from evaporation of various
hydrocarbon fluids. This relationship is due to the fact that complete combustion of
hydrocarbons is necessary to achieve high combustion efficiency. Three of the major
combustion-related sources, then, involve processes wherein energy conversion is not the
primary objective (carbon black production, ammonia production, and fluid catalytic
cracking). The other two major combustion-related sources of hydrocarbon emissions
(stationary internal combustion engines and steam boilers) do represent processes where
energy conversion is the prime objective and they are relatively efficient combustion
265
-------
processes (i.e., low hydrocarbon emission factors). They appear as major contributors
only because of the massive quantities of hydrocarbon fuels that are burned in these
systems.
SECTION 3
OTHER COMMENTS
Two other comments are of interest here, both involving uncertainties in the
date presented in the published reports (1-3). One is related to the magnitude of, and the
proportions between, the consumption of coal, oil, and natural gas in utility and industrial
boilers. The other concerns the large uncertainties in the emissions from the large
numbers of stationary internal combustion engines that are individually too small to be
classified as point sources.
The major environmental, availability, and cost perturbations which have been
affecting fuels for utility and industrial boilers are well-known. Initial trends toward
lower cost, easily handled natural gas prior to 1970 were accelerated in the early 1970s
as a result of environmental considerations. Then in the mid-1970s the lack of
availability of natural gas sharply reversed this trend. Today, strong efforts are being
made to decrease total fuel consumption, to increase the use of coal, and to eliminate
natural gas in these applications. As a result, predictions of the magnitudes and relative
proportions of the use of fossil-fuels in utility and industrial boilers are very uncertain.
Figure 7 graphically shows the problem, as exemplified by three attempts to
predict trends in natural gas consumption in industrial and utility boilers. Past history, in
this case, is no guide because gas consumption in these boilers has been increasing
continuously for many decades and very strongly in the recent past. Neither are data on
new boilers sales or boilers in fabrication good guides because boilers are being converted
to pil- or coal-firing and, where conversion is not possible, some natural gas-fired boilers
are being shut down entirely.
In the first attempt to project natural gas usage rates in utility and industrial
boilers, in 1974, the trend to reduce natural gas usage in utility boilers was just becoming
apparent. As a result, the best estimate trend was established to show some reduction,
with an increased uncertainty range about that projection. Before that estimate was
published, further public statements were made which appeared to indicate forthcoming
266
-------
regulations that would even more severely reduce natural gas usage, and the original
projected decreasing trend was steepened even further. Recent review of those data and
projections, particularly in the light of new data and analyses published in 1976 by the
Federal Energy Administration (4), resulted in the first data update, shown in Figure 7.
That study has indicated that the reversal of the usage trend expected in the 1974-1975
period has not been as sharp as expected and the trend projected for the near future
appears to be between those developed in the first two attempts. The negative slope in
this updated trend is entirely due to projected decreases in natural gas usage in utility
boilers while a small increase (2.1 percent per year) in usage in industrial boilers is
projected.
Clearly, as a result of rapidly changing compromises between energy,
environmental, economic, and political constraints, projection of fossil-fuel usage in
utility and industrial boilers, in total magnitude of the heat input or in the split between
the three major fuels, is very uncertain. As a result, emissions of the four air pollutants
are similarly uncertain. For example, NO emissions from steam boilers projected to
1980 are considered uncertain within ±15 percent.
Another area of significant uncertainty involves the huge numbers of stationary
internal combustion engines individually too small to qualify as point sources. Previous
studies (5-6) have shown, for example, that well over one million gasoline-powered
internal combustion engines were shipped from manufacturers every year for at least the
last 10 years for uses ranging from small power tools to 1000 horsepower and greater
compressors, pumps, and electrical power installations. Little data are available,
however, on the actual applications of these engines, their average useful life, or their
usage rates. Under one set of assumptions, NO and CO annual emissions from these
A
small stationary internal combustion engines were estimated at about 3 and 14 million
tons, respectively, in the year 1980; but these estimates could easily be low by factors of
two or three. Thus, this category could be the largest single stationary source of both
NO and CO. The NEDS data identify less than 2 and 6 million tons of NO and CO,
X X
respectively, in the total area source category, but the NEDS system of data collection
could also have missed these large numbers of small engines. Since the subject
inventory was limited to stationary point sources, no further effort was made to
investigate this category. Efforts to trace at least the larger of these engines to the
user and to estimate numbers of operating engines and their duty cycles certainly appear
warranted.
267
-------
REFERENCES
(1) O. W. Dykema and V. E. Kemp, Inventory of Combustion-Related Emissions
from Stationary Sources, EPA-600/7-76-012, The Aerospace Corporation, El
Segundo, California, (September 1976).
(2) O. W. Dykema and V. E. Kemp, Inventory of Combustion-Related Emissions
from Stationary Sources (First Update), EPA-600/2-77-066a, The Aerospace
Corporation, El Segundo, California (March 1977).
(3) V. E. Kemp and O. W. Dykema, Inventory of Combustion-Related Emissions
from Stationary Sources (Second Update) , The Aerospace Corporation, El
Segundo, California, (to be published as an EPA Report).
(4) Federal Energy Administration, National Energy Outlook, FEA-N-75/713 (1976).
(5) W. V. Roessler, et al, Assessment of the Applicability of Automotive Emission
Control Technology to Stationary Engines. EPA-650/2-74-051, The Aerospace
Corporation, El Segundo, California (July 1974).
(6) C. R. McGowin, Stationary Internal Combustion Engines in the United States,
EPA-R2-73-210, The Shell Development Company, Houston, Texas, (April 1973).
268
-------
03
0.
oo
•
o
in
CO
00
iH
j^J
INVENTOf
fa
O
W
(X
O
Q
O>
w
_5
05
<
E"
s-.
gj
CD
"of
8
*••
o
^H
• i—<
E
Emissions,
i-H
cO
o «o 3
?< -t 0
u t_
^H ^
rH ^
" 1
M
O
<4-.
0
&
w^
o> -
o o. S
ffi « o
M i
i
W
O> «N> OS O
• » » •
^* in o> o
CO (NO
i— 1
CO CXI O O
• * « •
c c
o •*
.2
E
u
i—t
Oj
§
CO <--
o i
.22
o
f
en
CM
269
-------
Q
w
w
>
z
^H
CO
M
5
o
o
w
H
<
Q
w
o
tz;
D
O
CO
H
£
o
OH
55
O
*HH
E-
<
CO
w
J
CQ
<
6-
CO
4-1
CO
Q
to
t-
en
M
0)
o
o
CO
^J
c
'o
ex,
g>
CO
c
o
at
CO
o
1
"i
U3
en
4_)
OS
3
«
a.
o
o
en
o
t.
E-.
•o
X
X
O
*£<
JO
c
h*
O
O
^
•*-•
H
JD
6
6
i-H
05
•*->
O
c™*
•
JD
1
r*-l
4->
^H
(N
*
X>
£
o
O
^^
i-H
p-4
cO
O
E-
1 a
^•c
t/J O
bo
• £
° "OB
Source
Category
01
*
CO
co
en
CO
CO
CO
^
eo
1-1
l£>
«
in
N
^f
Tf
00
^
•*
oo
C4
t«
Steam
Boilers
,
,
i
i
^*
at
in
i
-•N
CO
"""i
tj^
GO
Evaporation
t*
»
fn*i
t-
*
i-H
00
^
oo
CM
co
O^
in
T-H
Chemical
Manufact-
uring
i
270
i
,
1
i-H
*
in
TH
*
in
o>
in
o>
m
^*
i-H
Stationary
I.C. Engines
,
o
1— 1
t-
CM
1
1
1
1
at
^•J*
Secondary
Metals
,
„
at
i
i
i
c-
CO
t-
co
^*
t-
Mineral
Products
o
CM
O
C-3
00
oo
CM
1
i
1
1
CO
Wood
Products
,
_
0
c-
CM
|
0
OS
1
OO
o
1
V
•*->
O
CO
i-H
o
o
o
1-1
eg
cn
O
o
o
T- 1
^
00
I-H
o
*
o
o
04
O5
o>
o
C5
o
I— I
CO
»— 1
^
3
o
E-
9)
3
'Ss
CO
O
CO
CO
o
'J
CO
*->
CO
CO
J3
E
a
CM
I
O
CO
as
O
CO
•<->
CO
-------
w
o
H
OS
<
cu,
Q
55
0*8
* 8
P §
oo
en
o
»
OO
c- co
**" eo
CO rH
O
n
oo
to 1-4 » CN
(O
D
a
«
E-|
fc
0
O
CO
O
cu
>H
tf
«*<
^N
~i
CO
t-
05
1-1 C*>
t- rH
00
^
*
^f
OO
Q
«
O
a
i—i
M
-3
ffl
CO
c_
0)
-5
'S
JD
E
«
Q)
*->
CO
(A
CIA
• *H
t-,
-*-•
"O
£
cn
tn
r2
1
>>
+J
3
'j
D
•
cd
-------
EXAMPLE: MODIFIED SOURCE CLASSIFICATION
CODE *= 101002021
100000000 = EXTERNAL COMBUSTION, BOILERS
101000000 = ELECTRIC GENERATION
101002000 = BITUMINOUS COAL
101002020 = >100mmBtu/hr PULVERIZED, DRY
101002021 = TANGENTIAL FIRING
co
o
2
o£ 1
0
20
I 10
ja
0
200
GO
J2100
ANNUAL NOX EMISSIONS
NOy EMISSION FACTOR
ANNUAL CHARGE RATE (fuel usage)
I I
J1970
1975
1980
1985
Figure 1. Example of the projections of best estimates and uncertainties.
272
-------
120
100
80
GO
LU
GO
LU
00
en
60
£40
20
0
NOMINAL (typical)
V
CO
V
oo
'OD
NOX
T:
co
r-
oo
HC
I
CD
'00
,03
CO
I
'l.
co
'i—
03
7".
CO
CJD
PART.
CD
CC
>-
<
CC
LU
CJ
Figure 2. Projected trends in the four air pollutants from Stationary Point Sources.
273
-------
CONVENTIONAL
PULVERIZED,
WET BOTTOM
CYCLONE
130%)
VERTICAL-
FIRED
STOKER
OTHER
SINGLE-WALL-
FIRED
(17%)
CONVENTIONAL LARGE
PULVERIZED, DRY BOTTOM
(58%)
TANGENTIAL-FIRED
(21%)
OPPOSED-FIRED
(17%)
Figure 3. Projected NO emissions (4.6 million tons) from utility boilers firing
bituminous co&l in the year 1981.
274
-------
VERTICAL-
FIRED
CONVENTIONAL
PULVERIZED,
WET BOTTOM
(16%)
SINGLE
WALL-
FIRED
(16%)
CYCLONE
(5%)
STOKER
OTHER
OPPOSED-
FIRED
(16%)
CONVENTIONAL
PULVERIZED, DRY
BOTTOM
(75%)
TANGENTIAL-FIRED
(40%)
Figure 4. Projected particulate emissions (3.9 million tons) from utility boilers
firing bituminous coal in the year 1981.
275
-------
SINTERING
BASIC
OXYGEN
FURNACE
(9%)
BLAST
FURNACE
CHARGING
(20%)
CARBON
BLACK
MANUFACTURE
(12%)
PRIMARY
METALS
(34%)
CHEM
MFG
PETROLEUM
INDUSTRY
(52%)
FLUID CATALYTIC CRACKING
(51%)
Figure 5. Projected CO emissions (19.8 million tons) from combustion-related
stationary point sources in the year 1981.
276
-------
CLEANING
SOLVENTS
CARBON BLACK
PRODUCTION
STATIONARY
I.C. ENGINES
(11%)
AMMONIA
PRODUCTION
STEAM
BOILERS
MAJOR
COMBUSTION
SOURCES
(36%)
PET MKG
& TRANS
18%)
FLUID CAT
CRACKING
MAJOR
NON-COMBUSTION
SOURCES, EVAPORATION
(64%)
PETROLEUM
STORAGE
(13%)
SURFACE COATING
(39%)
Figure 6. Projected hydrocarbon emissions (2.8 million tons) from major
stationary point sources in the year 1981.
277
-------
10
CJ
CV1
CD
-
1ST PROJECTION-
(1974)
(preliminary)
CORRECTION
11975)
IRef. 1)
I
1965 1970 1975 1980
1985
Figure 7. Natural gas usage rate projections for utility and industrial boilers.
278
-------
EMISSIONS ASSESSMENT OF CONVENTIONAL
COMBUSTION SYSTEMS
By:
B. J. Matthews
TRW, Incorporated
Redondo Beach, CA 90278
279
-------
-------
ABSTRACT
The Industrial Environmental Research Laboratory (IERL) of the
Environmental Protection Agency, in association with TRW Environmental
Engineering Division, is conducting an extensive multimedia assessment
of emissions from conventional stationary combustion systems. The
program's purpose is the assessment of air, water, and solid waste
emissions from approximately 51 categories of Residential, Commercial,
Industrial and Utility combustion sources, burning gas, oil, coal and
refuse. This involves the collection of existing test data plus an
extensive sampling and analysis program. The forty-four (44) month
program is scheduled for completion in the spring of 1980. Reports on
specific types of combustion sources will be issued periodically starting
in late 1977,
This study, Contract No. 68-02-2197, is being conducted by TRW
Environmental Engineering Division under sponsorship of the United States
Environmental Protection Agency.
281
-------
-------
EMISSIONS ASSESSMENT OF CONVENTIONAL
COMBUSTION SYSTEMS
INTRODUCTION
The Industrial Environmental Research Laboratory (IERL) of the Environ-
mental Protection Agency, in association with TRW Environmental Engineering
Division, is conducting an extensive emission assessment of stationary combus-
tion systems. The program's purpose is the assessment of air, water and solid
waste emissions from approximately 51 categories of Residential, Commercial,
Industrial and Utility combustion sources. This involves the collection of
existing test data plus an extensive sampling and analysis program. The
forty-four (44) month program, directed by Dr. Ronald Venezia of IERL, is
scheduled for completion in the spring of 1980. Reports on specific types
of combustion sources will be issued periodically starting late 1977.
PROGRAM DESCRIPTION
TRW Environmental Engineering Division of Redondo Beach, California and
its subcontractors, GCA/Technology Division of Bedford, Massachusetts and
Engineering Sciences Incorporated of Arcadia, California are conducting an
extensive multimedia assessment of emissions from stationary combustion systems.
The combustion device categories being considered are shown in Table I.
They are classified in terms of application and fuel type. The four (4)
application categories are Residential, Commercial/Institutional, Industrial,
and Electricity Generation. The seven (7) fuel categories are gas, distillate
283
-------
oil, residual oil, lignite coal, bituminous coal, anthracite coal, and refuse.
The major program goal is to provide an accurate and complete emissions
data base for conventional combustion sources. To achieve this, existing
data are being collected and analyzed for accuracy and adequacy. Using
as a basis the adequacy of existing data, a test plan, which is subject to
periodic revision, was developed. The purpose of the test plan is to identify
those tests which will supply information currently missing from the existing
data base and to collect additional information in areas where the existing
data are questionable.
The sampling and analysis phase of this program is being conducted
jointly by TRW and GCA. Tests are being conducted throughout the country at
facilities provided on a voluntary basis by their owner/operator. Some
analysis will be conducted in the field but the bulk of the analytical work
will be done at TRW's and GCA's laboratory facilities at Redondp Beach,
California and Bedford, Massachusetts, respectively.
EXISTING DATA
An extensive search for existing emission data was conducted. Data
quality was evaluated in terms of criteria developed as part of this program.
To be acceptable, the test data had to meet the following six (6) criteria:
0 Only actual test data were acceptable. Emission estimates based on
emission factors or engineering estimates were not acceptable.
{This criteria eliminated data bases such as the National Emissions
Data System (NEDS).)
0 The combustion device had to be described adequately (i.e., design
heat rate, type of burners, type of draft, etc.)
0 The operating mode had to be defined adequately (i.e. load during
the test).
0 The design and operation of emission control devices had to be specified.
284
-------
0 A fuel analysis was required, with a minimum of trace elements,
sulphur, and ash specified.
0 The sampling and analysis methods had to be approved EPA, ASTM, or
API methods.
The resulting data base was examined and those areas with insufficient
data were identified. For most combustion source categories and pollutants,
the existing data are not adequate.
The data gethering task is continuing throughout the program. As
data become available from other sources, they will be included in the
data base. TRW is working in cooperation with the following companies to
collect additional test data for this program:
0 Radian Corporation
0 Hittman Associates
0 Tennessee Valley Authority (TVA)
0 Arthur D. Little
0 Monsanto Research Corporation (MRC)
0 Aerotherm/Accurex
0 Battelle Corporation
TEST PLAN
A test plan was developed which calls for data collection in those areas
where existing data are currently inadequate. In addition, projected changes
in fuel use patterns and expected changes in the types of combustion equipment
were considered. For example, because the use of anthracite coal is declining,
fewer tests were scheduled on anthracite fired boilers than those burning
other types of coal. Similarly, since the use of stoker fired units in the
electric utility industry is declining, proportionally fewer tests were
scheduled on this type of unit.
285
-------
The test plan will be revised periodically. Currently about one-third
of the available test dates have been left unscheduled. As data from this
and other programs become available, the remaining tests will be assigned
as appropriate. The purposes of the test plan are to maximize the amount
of data collected with the resources available and to collect those data
that will be most useful in future years.
SAMPLING AND ANALYSIS
A two tier approach to sampling and analysis is being used. At approxi-
mately 170 test sites a series of methodologies designated Level I procedures
are being employed. Level I results are, when compared to most existing
test data, quite detailed and sophisticated. The procedures are, however,
not the most sophisticated available. Level I tests are designed to provide
large amounts of data in a cost-effective manner. For example, stack gas
measurements are being made at a single representative sampling point under
semi-isokinetic conditions instead of using multipoint sampling under true
isokinetic conditions. Similarly, the analysis of trace elements is being
done with spark source mass spectrometry (SSMS) which is generally considered
to be semiquantitative.
Data from the Level I tests and other test programs will be used to
determine which sites are to be tested using the more sophisticated Level II
procedures. State-of-the-Art Level II procedures provide more precise and
detailed information on the composition and quantity of emissions. They
consist of the most sophisticated sampling and analysis procedures available.
Both the sampling and analysis, however, are more time consuming and expensive.
For this reason, Level II sampling and analysis will be applied to
approximately 21 sites.
The sampling and analysis procedures for flue gas are centered around
the Source Assessment Sampling System (SASS). The SASS train, which was
developed by Aerotherm/Accurex, is shown schematically in Figure 1. Cyclones
collect particulates in three (3) size ranges: 10+M, 3 to 10/x, and 1 to 3M.
A filter downstream of the cyclones collects any material that passes through
286
-------
the cyclones. A condenser containing a bed of molecular sieve material
follows, Water is condensed out and organic and inorganic material are
trapped. A series of impingers complete the train. The impingers collect
the most volatile organic and inorganic (such as mercury) compounds.
In addition to the six (6) SASS trains that are available for sampling
stack emissions, a variety of other sampling equipment is available for taking
samples of liquid effluents and solid wastes. Each will be used as circum-
stances dictate.
Three smaller vans and a trailer are available to support the two (2)
29-foot Mobile Environmental Assessment Laboratories that TRW designed and
outfitted for this program (Figures 2, 3 and 4). Each mobile lab is equipped
to provide lab and field facilities for the test crews. Each is equipped
with 13 kilowatts of onboard power; potable and high purity water systems;
a laminar flow hood; gas chromatographs; a refrigerator; an ice making machine;
and a broad range of analytical instrumentation and supplies. There are also
storage facilities for the SASS train and other sampling equipment.
On-board equipment will be used to analyze C-, through Cg organics, CO,
SOw and NO,,. Samples for the other analyses will be prepared and shipped
to the laboratory facilities in Redondo Beach and Bedford.
Figures 5, 6 and 7 show the basic analysis schemes for air emissions,
water effluents and solid waste, respectively. Table II summarizes the analyses
that will be conducted.
Samples will be analyzed using a variety of instrumental and wet-chemical
analysis methods including spark source mass spectrometry (SSMS), gas
chromatography, gas chromatography/mass spectrometry (GC/MS), atomic
absorption (AA), liquid chromatography, and infrared spectroscopy (IR).
After each particulate size range has been weighed separately, they are
combined to give two samples — particles that are smaller than 3^ and those
that are larger than 3». Organics are removed by extraction and, sample
size permitting, separated further into eight (8) classes of compounds.
287
-------
Each class is analyzed by infrared (IR) and mass spectrometric (MS) techniques.
The remaining particulate is subjected to elemental analysis and analysis
for sulfate (S04~). Portions of the particulate samples will be stored as
received for bioassay and particulate morphology studies. Currently these studies
are under consideration.
The substances that pass through the front part of the train as gases are
trapped either in the condenser condensate, the molecular sieve adsorbant,
or the impingers. The samples collected at each of these locations are
analyzed for organics and inorganics. The adsorbant is analyzed specifically
for PCB's and ROM's. The organic fractions are separated into eight (8)
classes and analyzed in the manner described above.
The solid and liquid portions of liquid and slurry effluents are analyzed
separately. The solids are subjected to the solid waste protocol; the
liquids are analyzed in a manner similar to the SASS train condensate. The
organic portions of each are separated into eight (8) classes for IR and MS
analysis. In addition, a variety of water quality parameters such as pH,
conductivity, chemical oxygen demand (COD), and biochemical oxygen demand.
(BOD), are measured.
Solid wastes are separated into water soluble and insoluble components.
The organics are separated into volatile and non-volatile components and
analyzed separately. The non-volatile components are separated into eight
(8) classes and analyzed separately.
The general plan is: all solid, liquid and gas samples, regardless
of the form in which they were originally emitted, will be analyzed for
volatile and non-volatile organic and inorganic components. It is recognized,
however, that it will not be feasible nor reasonable to conduct all tests
on all samples. For example, particulate emissions from gas fired units are
very low. It is impractical to collect a large enough particulate sample
to conduct all of the organic and inorganic analyses. Furthermore, since
gas fuel contains essentially no trace elements, it seems unreasonable to
search for them in the flue gas.
288
-------
QUALITY CONTROL
A comprehensive quality control function for both sampling and analysis
is being conducted. The objectives of this part of the program are to
establish and supervise procedures that assure reliable data. More specifically
the objectives are:
0 To establish acceptable limits on data quality;
0 To establish procedures that ensure the quality of data from various
sites and laboratories;
0 To establish guidelines for the selection and use of site-specific
measurement methods;
0 To develop and implement quality control programs on each specific
sampling technique and/or analysis;
0 To identify areas requiring new or improved measurement methods.
SUMMARY
The combustion emissions assessment program will provide detailed and
accurate data on the air, water, and solid waste emissions from stationary
combustion systems. The data being collected include particulate size
distributions,.-'trace element composition {including volatile elements),
organic emissions and composition, and standard water quality parameters.
These data will provide a sound basis for assessing the environmental
impact of stationary combustion sources as well as a data base for other
research programs.
289
-------
TABLE I. COMBUSTION SOURCE CATEGORIES
Residential Ext Comb Anthracite
Residential Ext Comb Bituminous
Residential Ext Comb Dist Oil
Residential Ext Comb Gas
Residential Ext Comb Wood
Residential Ext Comb Lignite
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Comniercial/Institutional
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Commercial/Institutional
Bottom
Commerci a1/Ins ti tuti onal
Bottom
Ext Comb Resid Oil Other
Ext Comb Resid Oil Tang Fire
Int Comb Uist Oil
Ext Comb Dist Oil Tang Fire
Ext Comb Dist Oil Other
Ext Comb Gas Other
Int Comb Gas
Ext Comb Gas Tang Fire
Ext Comb Anthracite Stoker
Ext Comb Bituminous Stoker
Ext Comb Bituminous Pulv Dry
Ext Comb Bituminous Pulv Wet
Industrial Ext Comb Resid Oil Other
Industrial Ext Comb Resid Oil Tang Fire
Industrial Ext Comb Bituminous Stoker
Industrial Ext Comb Bituminous Pulv Wet Bottom
Industrial Ext Comb Bituminous Cyclone
Industrial Ext Comb Gas Other
Industrial Int Comb Gas Recip Eng
Industrial Int Comb Gas Turbine
Industrial Ext Comb Gas Tang Fire
List of Abbreviations:
Ext Comb - external combustion
Int Comb - internal combustion
Pulv - pulverized
Tang - tangential
Recip Eng - reciprocating engine
290
-------
TABLE I. COMBUSTION SOURCE-CATEGORIES (CONT.)
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Electricity
Industrial
Industrial
Industrial
Industrial
Industrial
Industrial
Industrial
Generation Ext Comb Bituminous Pulv Wet Bottom
Generation Ext Comb .Bituminous Cyclone
Generation Ext Comb Bituminous Stoker
Generation Ext Comb Res id Oil Other
Generation Int Comb Dist Oil Turbine
Generation Int Comb Dist Oil Recip Eng
Generation Int Comb Gas Turbine
Generation Int Comb Gas Recip Eng
Generation Ext Comb Gas Other
Generation Ext Comb Gas Tang Fire
Generation Ext Comb Dist Oil Other
Generation Ext Comb Dist Oil Tang Fire
Generation Ext Comb Anthracite Stoker
Generation Ext Comb,Anthracite Pulv Dry Bottom
Generation Ext Comb Lignite Pulv Dry Bottom
Generation Ext Comb Lignite Pulv Wet Bottom
Generation Ext Comb Lignite Cyclone
Generation Ext Comb Lignite Stoker
Generation Ext Comb Refuse
Int Comb Dist Oil Recip Eng
Ext Comb Oil Other
Int Comb Dist Oil Turbine
Ext Comb Dist Oil Tang Fire
Ext Comb Refuse
Ext Comb Anthracite Stoker
Ext Comb Lignite Stoker
291
-------
TABLE II. SUMMARY OF ANALYSES TO BE PERFORMED
AIR
Particulates, by
size fractions
NO.,
S°
C1~C12
Greater than C,2
organics by functional
group
PCB
POM
Trace Elements
(seventy metals)
Selected Anlons
(such as fluorides,
chlorides, and
nit.rates)
Other Selected
Parameters
WATER
C,-C-|2 Organics
Greater than C,2
organics by functional
group
pcb
POM
Trace Elements
(seventy metals)
Selected Anions
(such as fluorides,
chlorides, sulfates,
nitrates, cyanide, and
phosphates)
N i trogen-Ammoni a
PH
Conduct!vi ty
Total suspended
particulates
Oil and Grease
Other Selected
Parameters
SOLIDS
Organics
Trace Elements
(seventy metals)
Selected Anions
(such as sulfates,
nitrates, chlorides
and fluorides)
Other Selected
Parameters
292
-------
u
•f—
•M
R}
-------
O)
•»—
O
4->
<0
O
t/1
(D
V)
c
O
-O
O
CvJ
OJ
cn
294
-------
OJ
*>
-------
-------
u*
1/»
Z>
_j
u.
«_4_J
I »
-+
1
r
t/»
u>
in
0
1
1
1
1
1
1
-z.
<
«c
*-
3
•Si
\
\
\
3
ul
1
1
1
1
1
1
^ -HBH
>
»•
— 4
— ^
— »•
*
-- fr
k
K
O
» ?
2*
UJ <£
u
o
. e
* R
§
O
sS
«A
oc
bJ
o
1
tt
tLl
O
sa
i
0
Ik
3-
O
3.
n
O
a
V
— 1
- — *
>,
' — >•
—
'
-- fc
— »•
X
» o
v\
y
H
Ou
ux
li
Ofc
Of (^
z^
— ^ UJ
if
H
z
o
R
u
z
o
s
— >•
— ^
— fr
—
— »
.
Sy
"•Z
o ^
fi
— H
— >
>
* .
*A
^>
»• Q^
^ IU
»-£j
>-
u,O
50
*8
^y
y£
K- tKf
??
tal
fr-
<
V*
>
S
O
r
<
g
c
»—
X
tu
•~^^ «"^^«« ^^^^ ^^^^H. ^_ «_^
i
V)
c
o
•^
(/)
•^-
E
iij
i-
•P-
s_
o
u.
^
,
«c
m
t
u t.
Z 3
£ en
ki *' -i—
i 1 ^
"* m1
— > s i
i 1 2 I
C 6 ,
z 5:
k. < U I
* s> ••• j
S 1
1
1
-* " 1
° i
§ 1
297
-------
o
I
o
o
Sz
£2
O
82?
3
o
11
o £
3s
G
>>
So
OS
rT u<*
I/)
c
(U
S-
V
-u
S-
o
o
(U
O1
Jfs_
^JOO
|oo
z*»
298
-------
SOLID SOURCE
COAL PIU
ASH PILE
SLUDGE AND
SEDIMENTS
WATER SOLUBLE
WATER
INSOLUBLE
I INORGANIC
ELEMENTS
f ORGANIC 1
GROUPS
1
I VOLATILES ! I NON-
VOLATILES
Figure 7. Analysts Scheme For Solid Waste
299
-------
-------
PANEL: COMBUSTION SOURCE/AIR POLLUTION REGULATIONS — PRESENT
AND PROJECTED
1 -- Federal Regulations
Jack R. Farmer
2 — Regional Regulations
Robert Dupree
Panel discussion abstracts will be included in Volume V.
301
-------
TECHNICAL REPORT DATA
(Please read instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/7~77-073b
4. TITLE AND SUBTITLE PROCEEDINGS OF THE SECOND
STATIONARY SOURCE COMBUSTION SYMPOSIUM
Volume D. Utility and Large Industrial Boilers
7 AUTMoms> Symposium Chairman J.S. Bowen, Vice-
Chairman R.E. Hall
9. PERFORMING ORGANIZATION NAME AND ADDRESS
NA
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO,
S. REPORT DATE
July 1977
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
EHE624
11, CONTRACT/GRANT NO.
NA (Inhouse)
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings: 8/29-10/1/77
14. SPONSORING AGENCY COOS '
EPA/600/13
is. suppLEMiNTARY NOTES T£RL-RTP project officer for these proceedings is R.E. Hall,
Mail Drop 65, 919/541-2477.
is. ABSTRACT
proceedings document the 50 presentations made during the Second
Stationary Source Combustion Symposium held in New Orleans, LA, August 29-
September 1, 1977. Sponsored by the Combustion Research Branch of EPA's Indus-
trial Environmental Research Laboratory — RTP, the symposium dealt with subjects
relating both to developing improved combustion technology for the reduction of air
pollutant emissions from stationary sources, and to improving equipment efficiency.
The symposium was divided into six parts, and the proceedings were issued in five
volumes: Volume I--Small Industrial, Commercial, and Residential Systems; Volume
II-~Utility and Large Industrial Boilers; Volume HI— Stationary Engine, Industrial
Process Combustion Systems , and Advanced Processes; Volume IV- -Fundamental
Combustion Research; and Volume V — Addendum. The symposium was intended to
provide contractor, industrial, and Government representatives with the latest infor-
mation on EPA inhouse and contract combustion research projects related to
pollution control, with emphasis on reducing nitrogen oxides while controlling other
emissions and improving efficiency.
t7. ' KEY WORDS AND DOCUMENT ANALYSIS '
S. DESCRIPTORS
Air Pollution, Combustion, Field Tests
Combustion Control, Coal, Oils
Natural Gas , Nitrogen Oxides , Carbon
Carbon Monoxide , Hydrocarbons , Boilers
Pulverized Fuels , Fossil Fuels , Utilities
Gas Turbines, Efficiency
18. DISTRIBUTION STATEMENT
Unlimited
b.lDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Combustion Modification
Unburned Hydrocarbons
Fundamental Research
Fuel Nitrogen
Burner Tests
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
c. COSATI Field/Group
13B 21B 14B
21D 11H
07B
07C 13A
13G 14A
21. NO. OF PAGES
308
22. PRICE
EPA Form 2220-1 (9-73)
302
-------
|EP 600/7
/-073b
EPA
Ind.
Res. Lab.
Proc. of the second stationai
TITLE source combustion symposium.
y.2;Utility & large industri-
OAVLORD *>
BORROWER'S NAME
-------
DATE DUE
BORROWER'S NAME
-------
DATE DUE
-------
------- |