c/EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/7-79-050a
February 1979
Proceedings of the Third
Stationary Source
Combustion Symposium;
Volume I.
Utility, Industrial,
Commercial, and Residential
Systems
Interagency
Energy/Environment
R&D Program Report
-------
-------
EPA-600/7-79-050a
February 1979
Proceedings of the Third
Stationary Source Combustion
Volume I. Utility, Industrial,
Joshua S. Bowen, Symposium Chairman,
and
Robert E. Hall, Symposium Vice-chairman
Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Program Element No. EHE624
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
PREFACE
These proceedings document more than 50 presentations and discussions
presented at the Third Symposium on Stationary Source Combustion held March
5-8, 1979 at the Sheraton Palace Hotel, San Francisco, California. Sponsored
by the Combustion Research Branch of the EPA's Industrial Environmental
Research Laboratory - Research Triangle Park, the symposium papers emphasized
recent results in the area of combustion modification for NOX control. In
addition, selected papers were also solicited on alternative methods for
NOX control, on environmental assessment, and on the impact of NOX control
on other pollutants.
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 Clyde
B. Eller, Director, Enforcement Division, U.S. EPA, Region IX and the opening
Address was delivered by Dr, Norbert A. Jaworski, Deputy Director of IERL-RTP.
The symposium consisted of seven sessions:
Session I:
Session II:
Session III:
Session IV:
Session V:
Session VI:
Session VII:
Small Industrial, Commercial and Residential Systems
Robert E. Hall, Session Chairman
Utilities and Large Industrial Boilers
David G. Lachapelle, Session Chairman
Advanced Processes
G. Blair Martin, Session Chairman
Special Topics
Joshua S. Bowen, Session Chairman
Stationary Engines and Industrial Process Combustion
Systems
John H. Wasser, Session Chairman
Fundamental Combustion Research
W. Steven Lanier, Session Chairman
Environmental Assessment
Wade H. Ponder, Session Chairman
n
-------
VOLUME I
Table of Contents
Session I: Small Industrial, Commercial, Residential Systems
Page
"Evaluation of Emissions and Control Technology for
Industrial Stoker Boilers," R. D. Giammar, D. R. Hopper,
P. R. Webb, E. Radhakrishan 3
"Field Tests of Industrial Stoker Fired Boilers for
Emission Control," J. E. Gabrielson, P. L. Langsjoen 35
"Guidelines for Adjustment of Residential Gas Burners
for Low Emissions and Good Efficiency," D. W. Locklin,
R. L. Himmel, D. W. DeWerth . 59
"Field Verification of Low-Emission Integrated
Residential Furnaces," A. S. Okuda and L. P. Combs 81
Session II: Utility and Large Industrial Boilers
"Status of NO Control Implementation for Utility
Boilers," L. W. Waterland, K. J. Lim and R. J.
Schreiber 119
"Field Testing of Utility Boilers and Gas Turbines
for Emission Reduction," E. H. Manny, A. R. Crawford,
and W. Bartok 157
"Corrosion Testing of Utility Boiler Combustion
Modifications," P. S. Natanson, A. R. Crawford,
E. H. Manny and W. Bartok 199
"Field Evaluation of Low NO Coal Burners on
X
Industrial and Utility Boilers," G. B. Martin ..... 213
"Applicability of the Thermal DeNO Process to
X
Coal-Fired Utility Boilers," G. M. Varga and W. Bartok ........ 233
"Combustion Modification Concepts for Stoker Boiler
Applications," J. H. Wasser (Abstract) (See Volume V) • • 253
ill
-------
SESSION I
SMALL INDUSTRIAL, COMMERCIAL, RESIDENTIAL SYSTEMS
ROBERT E. HALL
SESSION CHARIMAN
-------
EVALUATION OF EMISSIONS AND CONTROL TECHNOLOGY
FOR INDUSTRIAL STOKER BOILERS
by
Robert D. Giammar, David R. Hopper, Paul R. Webb,
and E. Radhakrishnan
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
ABSTRACT
In support of our nation's continual commitment to utilization of
coal in an environmentally acceptable manner, EPA has funded a research and
development program to identify and demonstrate improvements in stoker-coal
firing that can provide an incentive for greater industrial use of coal.
The overall objectives of this program are (1) to characterize the spectrum
of emissions from industrial coal-fired stoker boilers firing several types
of coal under various stoker firing conditions, (2) to investigate control
methods to reduce these emissions, (3) to determine the effect of these
control methods and variations in stoker-boiler operation on the overall
performance of the stoker boiler, and (4) to assess the environmental impact
of new technology on the future acceptability of stoker boilers.
This program is divided into two phases. In Phase I, emission
characteristics were determined for a variety of coals fired in a 20-bhp
stoker. Emphasis was focused on identifying coals with low pollutant
potential. These included both physically and chemically treated coals.
Phase II is currently being conducted to identify and evaluate potential
control concepts for control of emissions from full-scale industrial stoker
boilers. The Battelle-Columbus steam plant 25,000-lb steam/hr spreader-
stoker boiler is being utilized and modified to evaluate potential control
concepts.
In Phase I, it was observed that significant amounts of sulfur
were retained in the lignite and western subbituminous coals. Fuel
nitrogen conversion to NO was found to be between 10 and 20 percent. In
addition, a limestone/high sulfur coal pellet was developed and found
effective in capturing 80 percent of the fuel sulfur.
In Phase II, emissions were characterized for several coals under
a variety of operating conditions. Firing of the limestone/coal pellet
resulted in 75 parcent sulfur retention in the ash, thus verifying the
results of the small-scale Phase I study. Although there were some minor
operational problems in firing the pellet, it can be concluded that the
limestone/coal pellet has the potential to be a viable SO control
technique for industrial boilers.
-------
ACKNOWLEDGMENT
The research covered in this report was pursuant to Contract
No. 68-02-2627 with the U.S. Environmental Protection Agency, Combustion
Research Section. The authors acknowledge the assistance of the EPA
Project Officer, J. H. Wasser, who participated in planning this program
and provided helpful comments. The authors thank other Battelle-Columbus
staff who have contributed to this study—R. Coleman, J. H Faught,
H. G. Leonard, T. C. Lyons, and the plant facility and operation personnel.
In addition, we extend our appreciation to Harold Johnson of Detroit Stoker
and several plant engineers who provided us practical guidance.
-------
SECTION I
INTRODUCTION
The stoker-fired boiler provides an option for industry to meet
their energy needs. This option has not been exercised by a significant
number of industries primarily because oil- and gas-fired equipment have
been, and still are, more environmentally and economically attractive.
However, with the dwindling supplies of oil and gas, the rising costs of
these fuels, and increased attention given to coal utilization, industry
once again is considering the stoker coal-fired boiler.
In support of our nation's commitment to cleaning the environment
and to utilizing coal, EPA has funded a research and development program
to identify and demonstrate improvements in stoker-coal firing that can
provide an incentive for greater industrial use of coal. The overall
objectives of this program are (1) to characterize the spectrum of emissions
from industrial coal-fired stoker boilers firing several types of coal
under various stoker firing conditions, (2) to investigate control methods
to reduce these emissions, (3) to determine the effect of these control
methods and variations in stoker-boiler operation on the overall performance
of the stoker boiler, and (4) to assess the environmental impact of new
technology on the future acceptability of stoker boilers.
This program is divided into two phases. In Phase I, emission
characteristics were determined for a variety of coals fired in a 20-bhp
stoker. Emphasis was focused on identifying coals with low pollutant
potential. These included both physically and chemically treated coals.
Phase I experiments were conducted to provide some guidance for Phase II
* EPA Contract No. 68-02-2627, "Evaluation of Emissions and Control
Technology for Industrial Stoker Boilers".
-------
especially in the selection of a "design" fuel, a coal that was physically
or chemically treated to make it more environmentally acceptable for stoker
firing. Phase II is currently being conducted to identify and evaluate
potential concepts for control of emissions from full-scale industrial stoker
boilers. The Battelle-Columbus steam plant, 25,000-lb steam/hr spreader-
stoker boiler, is being utilized and modified to evaluate potential control
concepts.
This paper presents the preliminary results of Phase II experi-
ments focused at assessing emission levels for a variety of stoker-boiler
system operating variables. Combustion system design modifications as
possible control technologies are not addressed here; this aspect of Phase II
will be conducted in 1979. A brief summary of Phase I is also included in
this paper.
-------
SECTION 2
PHASE I STUDY
The overall performance of a stoker-boiler system and the
emissions that it generates are dependent upon a number of factors that
include stoker-boiler system design, coal properties, and combustion
operating parameters. As a consequence, when comparing emission and
performance data from several systems, each of these factors must be taken
into consideration. Phase I experiments were conducted in a 20-bhp stoker-
boiler facility (Figure 1) described in an earlier EPA report (Reference 1).
An underfeed stoker and a custom-built "model" spreader, schematically
shown in Figure 2, were both used.
The Phase I experiments were designed to generate data to provide
a relative measure of emissions from a variety of coals, including coals
that could not be conveniently or economically evaluated in larger indus-
trial systems. For example, in Phase I, chemically and physically treated
coals not commercially available as well as coals representative of major
reserves in the U.S. were fired. Additionally, it was felt that much of
the data would be representative of a variety of units which have been
verified by the Phase I experiments in both the underfeed and "model"
spreader stokers. Gaseous emissions correlate with coal properties and
appear to be somewhat independent of system design. Gaseous emission levels
obtained from firing the same coal in both the "model" spreader and an
industrial spreader stoker were similar. On the other hand, particulate
emission levels appear to be dependent on system design and may not be
representative of industrial-size units.
EXPERIMENTAL PROCEDURES
Experimental procedures are similar to those described in
Reference 1. Briefly, the stack gas was sampled at the boiler outlet and
levels were determined by: paramagnetic analysis for oxygen; nondispersive
-------
infrared for carbon monoxide, carbon dioxide, and nitrogen oxide; and a dry
electrochemical analyzer for sulfur dioxide. Particulate emissions were
measured in a straight section of stack, ten stack diameters downstream of
the boiler outlet. Particulate levels were determined by the EPA Method V
procedure.
Summary of Significant Findings
Fuel NO—
Table I indicates that a small fraction of the fuel nitrogen is
converted to NO (assuming no thermal NO). In comparison to pulverized coal
combustion, where as much as 30 to 40 percent of fuel nitrogen is converted
to NO, the fuel nitrogen conversion is relatively low for stoker firing.
This low conversion may be attributed to the inherent staged combustion of
stoker firing accomplished by supplying the total combustion air as primary air
through the grates and secondary air through overfire air jets located above
the bed. Additionally, as suggested in Reference 2, the relatively high
levels of CO that are present in the fuel bed may serve to reduce the NO
to N2-
Fuel Sulfur—
Table II shows the percent fuel sulfur emitted as S0« for selected
coals. It is well recognized (References 3 and 4) that the alkaline content
in the coal can be effective in retaining a portion of the fuel sulfur as
sulfate in the ash. Thus, when firing coals with naturally high calcium and
sodium content and those treated with these elements, only a portion of the
fuel sulfur is emitted as S02 as evidenced in Table II.
The measured values of SO,-, were exceptionally low for the lignite
and the western subbituminous coals. Maloney in his work (Reference 5)
suggests that lime was found more effective in capturing fuel S as sulfide
in fuel-rich regions, A representative sample of fuel ash could not be
obtained in the 20-bhp stoker facility to verify this observation. However,
it was observed that when firing lignite and subbituminous coals, combustion
was nonuniform in the fuel bed—a condition that often leads to unburned
carbon in the bottom ash.
-------
Limestone was mixed with Illinois No. 6 by two methods. One was
simply to introduce limestone (-50 mesh) into the feed system along with
stoker-size coal. In the other method, pulverized limestone and coal were
mixed together with a suitable binder and then extruded in a mill to form
a pellet of about 1/2 in. diameter and 3/4 in. length. The data indicate
that the limestone/coal pellets were significantly more effective in
capturing sulfur than simply mixing crushed limestone with stoker-size coal.
This increased sulfur capture may be attributed to the intimate contact of
the limestone with the coal particle in the pellets and possibly the more
reactive surface of the pellet.
The hydrothermally treated (HTT) coal is a coal that has been
chemically treated to remove sulfur (Reference 6). The process results
in some alkali residuals that are effective in capturing some of the
remaining fuel sulfur. The HTT coal fired in this study contained about
7 weight percent Ca and 0.5 weight percent Na.
In summary, the Phase I experiments provided gaseous emission
data from a variety of coals that appear to agree with those of Phase II.
In addition, during Phase I, a limestone/coal pellet was developed that
offers potential S0? control for industrial boilers.
10
-------
SECTION 3
PHASE II STUDY
Preliminary results of the Phase II experiments focused at
assessing emission levels for a variety of stoker-boiler system operating
variables are presented below. These experiments were conducted using the
No. 1 spreader-stoker boiler in the central heating plant at Battelle's
Columbus Laboratories. A variety of coals that could be conveniently
supplied to the test site were fired in addition to the limestone/coal
pellets.
FACILITY
The heating plant facility includes a coal-fired stoker utilizing
balanced induced draft and forced draft fans for combustion air and a boiler
for generating steam. The combustion particulate control system comprises
two cyclone dust collectors. The ash from the grate discharge, grate
sittings, and cyclone dust collectors are discharged into a hopper which
is periodically emptied via dump trucks. Figure 3 is a schematic showing
the relative locations of the system components. Sample ports for measuring
particulate emission concentrations are located in the exhaust breeching
downstream of the induced draft fan while a port for gaseous emission samp-
ling is located between the boiler outlet and induced draft fan. The same
type of instrumentation used in Phase I was also used in Phase II.
Stoker
The stoker is a 2A-Hoffman "Firite" Pulsating Ash Discharge (PAD)
type with a nominal feed-rate of about 1 ton/hr. The principle of operation
of the Hoffman mechanical spreader stoker is to spread coal uniformly on a
level, specially designed high resistance grate. The coal is burned in a
thin layer on top of the ashes and in suspension.
11
-------
Coal feed Is regulated by changing the length of stroke of the
reciprocating feed plates. Length of throw is regulated by the speed of
the distributor blades. Uniformity of distribution is a natural consequence
of the underthrow principle used exclusively by Hoffman. Distribution is
modified by changing the pitch angle of the circular tray.
The grate surface area is 77 sq ft; 11 ft by 7 ft. Incorporated
in this area are 18 individual grates. The six grates positioned in the
center are 9 in. by 39 in., while the remaining twelve grates located on
the sides of the boiler are 12 in. by 39 in. in dimension. The 12-in. wide
grates have 4 rows of tubes (16 in each row) and each tube has five 1/4-in.
flower-like openings on the grate surface to distribute the air. The 9 in.
grates have 3 rows of tubes (16 in each row) with similar holes.
The grate is activated periodically to discharge ashes according
to a predetermined time cycle. The time interval between discharge periods
is varied manually or automatically. The amount discharged per cycle is
adjusted manually to suit the particular type of coal and usually requires
no further adjustment for load change.
The grate assembly is energized by a revolving eccentric weight
in such a manner that the resulting reciprocating motion or pulsations
move the ash bed toward the discharge end. The rotating eccentric weight
shaft is driven through speed changing devices either from the line shaft
or by a separate motor. The purpose of the speed changer is to adjust for
the optimum speed after which the speed setting remains fixed. This adjust-
ment determines the distance the grate moves in each pulsation.
Fly Ash Reinjection System—
Fly ash is collected in hoppers under the boiler passes and is
reinjected to the furnace to complete its burning. Nozzles of the Venturi
type (Hoffman design) entrain the fly ash using air discharged from the
overfire air blower. The fly ash nozzles do not need any special attention
except to keep the nozzle piping free and clear. During operation, the
minimum air pressure that will insure continuous trouble-free operation is
used.
12
-------
Overfire Air Jets—
There are 4 overfire air jets in the rear and 5 in the front of
the furnace. These are located approximately 10 in. above the grate. The
maximum overfire air pressure attainable is around 16 in. W.G. The various
overfire air jets have individual and branch dampers adjustable by hand.
The minimum air pressure is selected that will provide turbulence in the
furnace.
Boiler—
The water tube gas/oil fired boiler is a Keeler Type MKB manufac-
tured by E. Keeler Company, Williamsport, Pennsylvania. The boiler was
installed at Battelle-Columbus in 1964 and altered for coal firing in 1976.
It has an operating pressure of 125 psig with an ASME Power Test Code
heating surface of 4119 sq ft utilizing a furnace volume of 1400 cu ft.
The boiler can operate at a capacity of 25,000 Ib steam per hour, continuous,
with a heat release rate of 22,700 Btu/cii ft/hr.
COAL ANALYSES
A variety of coals were fired in Boiler No. 1. The properties of
these coals comprising proximate and ultimate analyses, ash fusion tempera-
tures, and size consist are presented in Table III.
RESULTS
Table IV contains a summary of the measured emissions during the
Phase II experiments for test firing with the various coals.
Table V outlines the operating variables investigated during the
program. These operating variables are discussed with reference to the
emission data summary presented in Table IV.
Effect of Operating Variables on
Emissions and Boiler Performance
Excess Air—
The effect of different excess air rates was investigated at high
load for test firing with the low-S Ohio coal. This was accomplished by
13
-------
varying primarily the underfire combustion air for a constant coal
feed rate.
It was noted that carbon monoxide (CO) and smoke levels could be contro
by providing adequate excess air to insure complete combustion. However,
the limiting factor in the minimum excess air rate was not smoke or CO
levels, but clinker formation. Clinkers formed at flue gas 02 levels around
5.5 percent—CO was around 120 ppm and smoke opacity around 25 percent at
this relatively low 02 condition. Local hot spots which result from uneven
distribution of the coal are primarily responsible for clinker formation at
the reduced excess air rates.
No appreciable influence on S02 emissions was observed for varia-
tions in excess air at the full-load condition, while a reduction in NO
levels was observed at lower excess air rates. Optimum excess air levels
were found to be around 8.0 percent 0^ for the coal fired. Test No. SP-1
(Table IV) represents this optimum condition. Optimum excess air rates
were also determined for the other coals fired and these are discussed later.
Overfire Air Rate—
A variety of overfire air/total air flow rate ratios (0 to 0.25)
were investigated at full-load operation for test firing of the low-S Ohio
coal. Changes in this ratio affected smoke and CO levels. The reduction
in smoke levels was much more significant than that in CO at increased
overfire air rates.
During the test period, both the rear and front overfire air jets
were deactivated. Smoke levels increased to unacceptable levels (about 30
percent opacity) while CO increased from 30 ppm to 50 ppm. NO levels
decreased slightly upon deactivating the jets—this can be attributed to
several factors including lower above-bed temperatures, reduced oxygen
levels, and heterogeneous reduction reactions with smoke particles.
The optimum overfire air rate was about 18.0 percent of the total
combustion air, corresponding to 7.5 to 7.7 percent 0? in the flue gas. For
a constant total air flow rate, an increase in the overfire air/total air
flow rate ratio above 18 percent would further reduce smoke, but the reduc-
tion of underfire air through the bed would create severe clinkerine
Test No. SP-4 (Table IV) was conducted at the optimum overfir
air condition. A comparison with Test No. SP-1 shows lower CO a-nA ^
emu smoke
14
-------
levels at the higher o'verfire air ratio. Particulate loadings are also
lower at the optimum overfire air ratio, revealing that overfire air can
reduce particulate emissions by promoting fuel burnout.
Fuel Bed Depth—
Experiments to explore the effects of different bed depths at
full load were conducted for test firing with the medium-S Kentucky coal.
This coal has a low ash content and a high heating value, properties which
were found more suitable for these tests. For higher ash coals, satisfactory
combustion could only be achieved over a narrow range of bed depths.
By operating without grate vibration for 3 hours, the bed built
up to about 4-1/2 inches. Clinker formation would occur at greater
bed depths. With a grate vibrating cycle of 6 minutes, the bed depth could
be reduced to about an inch- which is the minimum limit. This limit is
imposed by the danger of blowing the insulating ash layer off the grate
and exposing che grate metal to the heat of the furnace.
A bed depth of about 2-1/2 inches was found to be optimum and
resulted by using a vibrating cycle of 35 minutes for the grate. Test SP-3
(Table IV) which was made at this condition shows that a lower excess air
rate (around 7.0 percent 0?) is required for this coal. Sulfur retention—
about 10 percent—is lower than that obtained with the low-S Ohio coal
(20 to 30 percent), This may be attributed to the lower ash content of the
Kentucky coal. Particulate loadings are, however, comparable to that with
the Ohio coal—this may be due to factors including the thinner fuel bed,
the friable nature of the Kentucky coal, the greater amount of fines
present in it, and its higher S content (leading to condensable sulfate
formation in the fly ash).
Boiler Loads—
The effect of load on boiler operation was studied for test firings
of the low-S Ohio coal. The different loads studied and the corresponding
Test Run Numbers (from Table IV) are given below:
1. Full Load — around 21,000 to 22,000 Ib per hr steam
(Test No.s SP-1 and SP-4)
2. Partial Load — around 15,000 Ib per hr steam
(Test No. SP-2)
15
-------
3. Low Load — around 10,000 Ib per hf steam
(Test No. SP-5)
The excess 0 required at partial load (7.5 percent) was comparable
to that at full load, while excess 02 at low load (11.8 percent) was sub-
stantially higher. A greater percentage of overfire air was required at
lower loads to maintain a clean stack and reduce excessive CO emissions.
Bed temperatures were around 2575 F at full load, 2450 F at partial
load, and about 2200 F at low load.
NO levels were not measured during all four tests—the difference
in NO emissions for Runs SP-4 and SP-5 is not significant. S02 emissions
are comparable at partial and full load conditions, but about 15 to 20 per-
cent lower at the low-load condition. This result appears to be verified
by the thennodynamic prediction (References 5 and 7) that S retention in
bed ash increases with decreasing temperatures below 2500 F and increasing
excess air, conditions associated with low-load operation. A comparison of
dust loadings in Tests SP-1, SP-2, and SP-5 shows a trend of increasing
particulates with increasing boiler load. This is probably due to an
increased rate of fly ash carryover at higher underfire air rates required
for higher loads.
Boiler efficiencies were not significantly different at the
various loads—these were 78 percent, 76 percent, and 75 percent, respec-
tively, at full, partial, and low load.
A significant achievement of the low load experiments was the
alleviation of the low-load smoke problem. This was achieved by supplying
a controlled amount of underfire air to the bed. At high underfire air
settings, smoke formation occurs due to incomplete combustion resulting
from local quenching of the flame by the combustion air. However, there
is a minimum level of excess air that is required to avoid clinker formation.
Smoke levels below 5 percent opacity were achieved for low-load operation
at excess air levels around 12.0 percent oxygen (Test No. SP-5). It must
be pointed out that the attention and skill of the boiler operator are
important factors in controlling low-load smoke, as smoke levels can
increase suddenly and drastically for any changes in boiler load or
operation.
16
-------
Fly Ash Reinjection—
The effect of fly ash reinjection on particulate emissions and
combustion efficiency was investigated for test firing with the following
coals:
II Low-Ss washed Kentucky coal (Test Nos. SP-11, SP-12)
2. High-S, washed run-of-the-mine Ohio Coal (Test Nos.
SP-13, SP-14)
3. High-S, washed stoker-grade Ohio coal (Test Nos.
SP-15, SP-16).
The runs during which fly ash reinjection was practiced corres-
pond to Test Nos. SP-11, SP-14, and SP-16, while no fly ash was reinjected
during Tests SP-12, SP-13, and SP-15. Furthermore, only about 65 percent of
the total fly ash was reinjected during the fly ash reinjection runs for
the high-S Ohio coals. Each coal was fired with and without fly ash
reinjection under approximately the same boiler operating conditions--full
load (21,000 to 22,000 Ib per hr steam) was maintained during these runs.
No significant trend in gaseous emissions was observed, while a
reduction in smoke was observed for runs made without fly ash reinjection.
The effect on smoke iss however, difficult to determine as smoke levels
under all conditions were below 5 percent opacity.
The most significant effect was the reduction in particulate
loadings without fly ash reinjection, as shown in Table VI. These results
suggest that particulate loadings may be reduced by 10 to 25 percent by
operating the spreader-stoker boiler without fly ash reinjection. Table VI
also shows that the carbon loss by not reintroducing the fly ash is less
than 1-1/2 percent of the total carbon input. With fly ash reinjection,
the carbon loss is about 3 percent of the total carbon input with certain
coals.
These results, plus the fact that portions of the fly ash reinjec-
tion system have to replaced periodically due to erosion, make fly ash
reinjection unattractive for the Battelle boiler.
Coal Types—
With the exception of the unwashed stoker Ohio coal, and to a
certain extent, the Illinois No. 6 coal, the other coals were fired
17
-------
satisfactorily. The high ash content of the unwashed stoker coal and the
high moisture content of the Illinois No. 6 coal are partly responsible
for the problems encountered in firing them. The high amount of fines
present in these coals added to the difficulty in firing them.
The Kentucky and the low-S Ohio coals were superior to the high-S
stoker-grade and run-of-the-mine coals in terms of S02, NO, and particulate
emissions. Moreover, as Table IV shows, they could be fired at lower excess
air rates.
Sulfur oxide emissions are largely a function of the sulfur
in the fuel. It appears that Ohio coals with less than 0.7 percent S and
Kentucky coals with less than 1.25 percent S would meet the Federal S02
emission standard of 1.2 Ib/mBtu. Again, Table IV reveals that some of the
eastern Bituminous coals have a sulfur capture efficiency around 25 percent
at full-load operation.
Nitric oxide emissions were low and below 0.6 Ib/mBtu for all
fuels tested. Conversion of fuel nitrogen to NO (assuming no thermal NO)
was around 12 to 20 percent (Table II), similar to that seen in the Phase I
study.
Particulate emissions were close to or below the Ohio EPA regula-
tion of 0.24 Ib/mBtu during test firing of the low-S Ohio and the Kentucky
coals. However, as can be seen from Table IV, they were much higher for the
high-S Ohio coals. Analyses of the 'fly ash emitted from firing the high-S
coals show a higher percentage of sulfate, signifying that an appreciable
quantity of condensable sulfates may be present in the fly ash from these
coals.
Table IV shows that CO emissions were low for all fuels tested
(with the exception of the limestone/coal pellet) and less than 120 ppm at
3 percent 0«.
Boiler efficiencies were around 76 percent for the low-S Ohio
coal, around 84 percent for the medium-S Kentucky coal, 71 to 74 percent
with the high-S washed coals, and 68 percent for the unwashed stoker coal.
This shows that boiler efficiencies are appreciably higher for the low-
moisture, low-ash Kentucky coal. Combustion losses also appear to be lower
for this fuel.
18
-------
Limestone/high-S coal pellet—The limestone/high-S coal pellets,
which showed potential for sulfur capture in the Phase I experiments, were
fired in the 25,000 Ib steam/hr .stoker. Test Nos. SP-18 and SP-19 represent
the pellet runs. The limestone/coal pellets were found to possess a sulfur
capture efficiency of about 75 percent. Some operational problems were
encountered in maintaining a stable bed due to the high percentage of ash
(=50 percent) in the pellet. Also, a fairly high excess air rate (around
11.0 percent 0,-,) had to be supplied to maintain clinkerless operation.
The limestone/coal pellet developed for this study offers a
potentially interesting and exciting new means of sulfur control for indus-
trial boilers. The use of limestone for capturing sulfur is by no means
new. Exactly how and why limestone should be effective in capturing sulfur
on a stoker bed is not well established at this time. Thermodynamically,
one would guess that the coal bed is much too hot to retain sulfur as
calcium sulfate. On the other hand, it is possible that the stoker bed
is a very nonuniform combustion mixture. In such a state, one might
anticipate significant lean air regions or areas. As such, it is conceivable
that reduced conditions may exist in parts of the bed which ultimately lead
to sulfur capture via sulfide formation. This is also an explanation
offered by Maloney, et al, in their recent study. At this stage one must
only consider this as a hypothetical explanation since the overall bed
analyses tend to show little sulfide present, These speculations notwith-
standing the use of limestone/coal pellets merit serious attention as
viable alternative means for controlling SO-.
From an industrial point of view, the possibilities of using
limestone/coal pellets for removing S0« in situ via a dry process is
considered more inviting than the use of scrubbers. The additional costs
for pelletizing the coal mixture and for the removal of possibly twice as
much ash is far more attractive than the high cost of operating and maintain-
ing wet scrubbers.
19
-------
SECTION 4
SUMMARY OF PHASE II RESULTS
The Phase II experiments have demonstrated that emission levels
can be reduced by proper control of the stoker operating variables. In
addition, the limestone/coal pellets have been demonstrated to offer potential
control for S0». In summary, the major findings are:
• The limestone/high-sulfur coal pellet showed a sulfur
capture of about 75 percent.
• Sulfur capture efficiencies of around 25 percent were
noticed with some eastern bituminous coals.
• High excess air rates at low loads result in increased
sulfur retention in the bed ash.
• CO and smoke levels can be controlled by providing
adequate excess air. CO levels were low for all
fuels tested.
• Clinker formation may be a limiting factor in
determining the minimum excess air rate.
• NO levels increase slightly with increase in excess
air.
• Conversion of fuel nitrogen to NO was between 12 to
20 percent, assuming no thermal NO.
• An increase in overfire air/total air flow rate
ratio reduces CO and smoke, the latter more signi-
ficantly. Particulate loadings are also lower with
increased overfire air.
• NO is lower for operation with inactive jets.
• Clinker formation occurs readily if bed depths
become excessive, while the danger of burning the
grates exists for operation with very thin beds.
Bed depths around 2.5 to 3.0 inches appear to be
optimum for lower ash coals.
20
-------
• A higher excess air rate is required for low-load than
for partial- or full-load operation. A greater percen-
tage of overfire air is required at low load. Low load
smoke can be reduced by a reduction in underfire air,
coupled with attentive boiler operation.
• Particulate loadings may be reduced by 10 to 25 percent
by operating at full load without fly ash reinjection.
Fuel savings of less than 1.5 percent were observed
for operation with fly ash reinjection.
• The high-S Ohio coals had to be fired at higher excess
air rates than with the low-S Ohio and Kentucky coals.
The high-ash unwashed stoker coal and high-moisture
Illinois No. 6 coal could not be fired satisfactorily.
SECTION 5
CONVERSION FACTORS
This report has been prepared using English engineering units,
Table VTI provides factors for conversion to SI units, if desired.
21
-------
REFERENCES
1. Giammar, R. D., R. B. Engdahl, and R. E. Barrett. Emissions from
Residential and Small Commercial Stoker-Coal-Fired Boilers Under
Smokeless Operation. EPA-600/7-76-029, U.S. Environmental Protection
Agency, Washington, D. C. 20460, October, 1976.
2. Merryman, E. L., S. E. Miller, and A. Levy. Reduction of NO in the
Presence of Fly Ash. Submitted for publication in the Combustion and
Science Technology Journal.
3. Gronhovd, G. H., P. H. Tufte, and S. J. Selle. Some Studies on Stack
Emissions from Lignite-Fired Power Plants. Presented at 1973 Lignite
Symposium, Grand Forks, ND, May 9-10, 1973.
4. Maloney, K. L., G. L. Moilanen, and P. L. Langsjoen. Low Sulfur Western
Coal Use in Existing Small and Intermediate Size Boilers. EPA-600/7-78-
153a, U.S. Environmental Protection Agency, Washington, D. C. 20460,
July, 1978.
5. Maloney, K. L. , P. K. Engel, and S. S. Cherry. Sulfur Retention in Coal
Ash. KVB 8810-482-b, EPA Contract No. 68-02-1863, Industrial Environ-
mental Research Laboratory, EPA, Research Triangle Park, NC, November,
1978.
6. Stambaugh, E. P., et al. Combustion of Hydrothermally Treated Coals.
EPA-600/7-78-068, U. S. Environmental Protection Agency, Washington, B.C.,
20460, April, 1978.
7. Giammar, R. D., and R. W. Coutant. Experimental Studies on the
Feasibility of In-Furnace Control of S02 and NO Emissions from
Industrial Stokers. A Battelle Energy Program,XBattelle, Columbus
Laboratories, December, 1975.
22
-------
Figure 1. 20-bhp Stoker-Boiler Facility
23
-------
HOPPER
MODEL SPREADER STOKER
VARIABLE SPEED
TRANSMISSION
VIBRATORY
FEED
ROTOR
GRATE
Figure 2. Model Spreader Stoker
24
-------
WATER
TREATING
AGENTS
WATER
TREATMENT
STACK -^-
n
SAMPLE PORTS
PRIMARY
CYCLONE
SECONDARY
CYCLONE
SPENT
TREATING
AGENTS
\
INDUCED
DRAFT
FAN
UNDER FIRE
FORCED DRAFT
GRATE
DISCHARGE
GRATE
SIFTINGS
ASH
DISCHARGE
Figure 3. Schematic of 25,000 Ib steam/hr
Stoker-Boiler Facility
25
-------
TABLE I. CONVERSION OF FUEL NITROGEN TO
NO FOR SELECTED COALS (assuming
no thermal NO)
Coal
Nitrogen
Content,
%
Measured
NO,
ppm
Theoretical
NO,
ppm
Fuel N
Conversion,
%
Lignite 0.61
Western Subbituminous 0.60
Western Bituminous 1.40
Illinois No. 6 0.93
Eastern Bituminous 1.30
Hydro thermally Treated 1.10
Coal
Illinois No. 6/
Limestone Pellet
120
160
200
200
270
180
1500
1800
2400
1450
1800
1850
8
8
14
14
10
TABLE II. AVAILABLE FUEL SULFUR EMITTED AS SO,
FOR SELECTED COALS
Coal
Lignite
Western Subbituminous
Western Bituminous
Eastern Bituminous
Illinois No. 6
Illinois No. 6 w/Limestone
(Ca/S=7)
Illinois No. 6/Limestone
Pellet (Ca/S=7)
Hydro thermally Treated
Coal Pellet
Sulfur
Content,
1.10
0.67
0.55
1.20
4.40
—
—
1.10
Measured
so2,
ppm
350
560
400
720
3800
3200
560
220
Theoretical
S02,
PPm
1100
940
450
740
4100
4100
2910
820
Fuel S
Emitted as
en '°
l-l
32
60
89
97
93
78
19
27
26
-------
TABLE III.
Ultimate Analysis, %
(as received)
Proximate Analysis, %
(as received)
Fuel Type
Low-S Ohio Coal
Medium-S Kentucky
Vola tiles
33.12
38.20
Fixed
Carbon
47.59
53.15
Ash
9.60
4.95
Mois-
ture
9.69
3.70
Carbon
64.81
76.96
Hydro-
gen
4.26
5.15
Coal
Stoker-Grade,
Washed Ohio Coal
37.84
42.58
10.34 9.24 63.40 4.54
Stoker-Grade,
Unwashed Ohio Coal
Run-of-the-Mine ,
Washed Ohio Coal
Low-S Kentucky Coal
Illinois No. 6
Limestone/High-
Sulfur Coal Pellet
36.34
38.89
36.87
37.30
39.20
40.95
43.47
52.83
39.97
NA*
18.50
8.60
7.71
8.63
45.60
4.21
9.04
2.59
14.10
12.60
59.98
65.02
75.42
60.52
29.70
4.23
4.56
5.03
4.23
1.54
* NA - Not Available.
27
-------
SUMMARY OF COAL ANALYSES
Ultimate Analysis, %
(as received)
Oxy-
gen
Nitro- Chlo- Sul- (differ-
gen rine fur ence)
Heating Ash-Fusion Fuel Size
Value Temperatures, F Consist
(as reed.) (Initial Deformation) (% less
Btu/lb Reducing Oxidizing than 1/8")
1.26
1.26
0.05
0.14
0.70
1.38
9.63
6.46
11,473
13,786
NA
2125
2700+
2450
7
12
1.21
0.06
3.00
8.21
11,452
2045
2470
1.09 0.06 3.94 7.99
0.98 0.04 3.19 8-57
11,009
11,732
2080
2005
2485
2400
17
16
1.53 0.11 0.89 6.72
1.10 0.14 3.46 7.82
0.48 — 1.96 8.64
13,480
10,903
4,257
2480
2020
NA*
2700
2310
NA*
12
6
28
-------
TABLE IV. EMISSION DATA SUMMARY FROM PHASE II EXPERIMENTS—
Test
No.
SP-1
SP-2
SF-3
SP~4
Sf-5
SP-6
SP-7
SP-8
SP-9
SP-10
SP-11
SP-1 2
SP-13
SP-14
SP-15
SP-16
SP-17
SP-18
Coal Type
Low-S Ohio
Low-S Ohio
Medium-S Kentucky
Low-S Ohio
Low-S Ohio
Stoker-grade,
washed Ohio
Stoker-grade,
unwashed Ohio
Run-cf-the-mine,
washed Ohio
Stoker-grade,
washed Ohio
Run-cf-the-mine,
washed Ohio
Low-S Kentucky
Low-S Kentucky
Run-of-the-mlne ,
washed Ohio
Run-o f- the-mine ,
washed Ohio
Stoker-grade,
washed Ohio
Stoker- grade,
washed Ohio
Illinois No. 6
Limes tone/ hi gh-S
Load,
103 Ib
/hr
21.5
14.5
21.0
21.0
9.5
21.5
21.0
21.0
22.5
22.0
22.5
22.0
20.0
20.5
20.0
21.0
17.0
16.0
Overfire
Air
Total
Air
0.14
0.23
0.16
0.18
0.32
0.17
0.15
0.17
0.16
0.18
0.18
0.16
0.17
0.19
0.18
0.19
0.19
0.13
Flue Gas Composition
02,
%
8.0
7.5
7.0
7.7
11.8
8.0
10.8
9.0
9.4
9.0
8.5
8.8
9.3
9.8
10.0
9.2
9.5
11.2
co2,
%
11.0
11.4
12.4
13.0
8.4
11.2
9.0
10.5
10.1
10.8
10.4
10.4
9.8
9.4
9.8
10.2
10.2
10.0
CO,
ppm
50
60
34
25-30
70
28-44
32-72
40-72
40-68
35-50
35-45
25-35
30-45
40-68
20-35
20
36-40
420-600
so2,
ppm
360
350
700
320
190
1800
2200
1800
1400
1900
340
330
1500
1300
1350
1700
1900
560
NO,
ppm
__
__
—
230
185
280
250
230
240
245
250
235
250
240
230
260
230
145
Smoke
Opacity,
%
10
4-5
7
4
5
6
7-11
6-9
6
6
4
2
3
4
2
3
5
22
coal pellet
SP-19 Limestone/high-E
coal pellet
15.0
0.16
11.8 9.5 1000
515 140 25
29
-------
BATTELLE'S 600-bhp SPREADER STOKER
CO at 3%
02, ppm
70
80
44
34-41
140
38-60
52-120
60-80
63-107
52-75
50-65
37-52
60-70
65-110
33-58
31
57-63
792-1132
S02 at 3% C
Computed
into System
625
625
1010
625
625
2635
3650
2750
2635
2750
665
665
2750
2750
2635
2635
3190
4035
>2, ppm
Measured
Emissions
505
470
910
440
375
2475
3635
2740
2192
2825
490
488
2310
2092
2230
2602
2974
1057
Fuel S
Emitted,
80.8
75.2
90.1
70.4
60.0
93.9
99.6
99.6
83.2
(102.7)
73.7
73.4
84.0
76.1
84.6
98.7
93.2
26.2
NO at 3%
Computed
into System
2565
2565
2110
2565
2565
2430
2310
1930
2430
1930
2615
2615
1930
1930
2430
2430
2320
2255
02, ppm
Measured
Emissions
—
—
—
320
365
385
413
350
375
365
360
350
385
386
380
398
360
274
% Conver-
sion of
.Fuel N
to NO
—
—
—
12.5
14.2
15.8
17.9
18.1
15.4
18.9
13.8
13.4
19.9
20.0
15.6
16.4
15.5
12.2
Particu-
lates,
Ib/mBtu
0.26
0.18
0.25
0.17
0.15
0.45
—
0.55
0.51
0.67
0.17
0.13
0.36
0.40
0.37
0.41
—
1.43
2010
4035
1035
25.6
2255
281
12.5
2.24
30
-------
TABLE V. OPERATING VARIABLES STUDIES FOR SYSTEM
CHARACTERIZATION IN PHASE II
EXCESS AIR
- Minimum levels
- Optimum operating levels from both an emissions and
efficiency viewpoint
OVERFIRE AIR RATE
- Jets inactive
- Optimum rate from an emissions viewpoint
FUEL BED DEPTH
- Maximum
- Minimum
- Optimum
BOILER LOADS
- Low load (30 to 40 percent of full load)
- Partial load
- Full load
FLY ASH REINJECTION
- Maximum
- Partial
- None
COAL TYPES
- Size distribution
- Ash content
- Ash-fusion properties
- Sulfur levels
- Treated coal: limestone/coal pellet
31
-------
TABLE VI. EFFECT OF FLY ASH REINJECTION
Type of Coal
Particulate Loading,
Ib/mBtu
w/ Fly Ash w/o Fly Ash
Reinjection Reinjection
Percent
Reduction Percent
in Particulate Increase
Loadings in C Loss
Kentucky
0.17
0.13
23.5
1.0
Run-of-the-Mine Ohio
0.40
0.36
10.
1.3
Stoker-Grade Ohio
0.41
0.37
9.8"
0.5
* Around 65 percent of total fly ash was reinjected during the fly ash
reinjection runs.
32
-------
TABLE VII. ENGLISH AND METRIC UNITS TO SI UNITS
To Obtain
ng/J
ng/J
m
cm
2
m
3
m
kg
Celsius
Kelvin
Pa
Pa
MW
From
lb/106 Btu
g/Mcal
ft
in.
2
ftZ
ft3
Ib
Fahrenheit
Fahrenheit
psig
iwg (39.2 F)
106 Btu/hr
Multiply By
430
239
0.3048
2.54
0.0929
0.02832
0.4536
tc = 5/9 (tF - 32)
tr = 5/9 (tw - 32) + 273
K F
P = (P . + 14.7)(6.895 x 103)
pa psig
P =• (P. )(249.1)
pa iwg
0.293
33
-------
FIELD TESTS OF INDUSTRIAL STOKER FIRED BOILERS
FOR EMISSION CONTROL
BY
J. E. GABRIELSON
P. L. LANGSJOEN
KVB, INC.
6176 OLSON MEMORIAL HIGHWAY
MINNEAPOLIS, MINNESOTA 55422
35
-------
ABSTRACT
A program funded by the Department of Energy and the U, S. Environmental
Protection Agency is underway to produce information which will increase manu-
facturers ability to design and apply stokers and boilers which are an economic
cal and environmentally satisfactory alternative to importation and combustion
of fuel oil which may not be available in the event of a national emergency.
Initially, the program was to test six spreader stokers, but it has been
expanded to also include five mass fed stokers.
To date, three boilers with spreader stokers and one with a Vibra-Grate
Stoker have been tested under this ongoing program. Some of the preliminary
findings are presented in this paper. Emissions and efficiency are reported
as functions of amount and geometry of overfire air admission, degree of
flyash reinjection, excess air level, coal properties, air velocity through
the grate and the two types of stokers are compared. Measurements include
particulate mass loading at both the boiler outlet and the mechanical collector
outlet, oxides of nitrogen, excess oxygen and carbon monoxide. Combustible
and dry gas heat losses are also discussed.
36
-------
INTRODUCTION
In recent years the vast majority of industrial boiler installations have
been shop assembled gas and oil fired units which could be purchased and
installed at substantially lower cost than conventional coal burning boiler-
stoker equipment. Because of the decline in the coal firing market areas for
equipment manufacturers, little or no work has been done in recent years to
develop cleaner burning stoker fired boilers. Just as important, updated
emission specification data have not been made available to consulting engineers
and purchasers of coal burning equipment. The implementation of more rigid air
pollution regulations coupled with the lack of good boiler emission data has had
a further negative influence on the installation of coal burning installations.
It is difficult for many coal burning installations to comply in a cost effective
manner with required stack emission limits.
The national interest in coal utilization has increased dramatically within
the last few years, As a result, the Department of Energy (DOE) and the Research
Branch of the Environmental Protection Agency (EPA) are co-funding a tesL program
whose oblective is to develop information which will increase manufacturers
ability to design and apply stoker fired boilers which are an economical and
environmentally satisfactory alternative to importation and combustion of fuel
oil which may be in short supply in the event of a national emergency. The
American Boiler Manufactuers Association (ABMA) is managing the test program
through its Stoker Technical Committee. KVB, Inc.,, a combustion consulting firm,
is performing the testing.
The scope of this program covers testing of eleven stoker fired boilers in
the size range of 20 to 90 MW (75,000 to 300,000 Ib/hr steam flow). Only new
"state-of-the-art1' stoker fired boilers are being tested. By comparison of test
data between various units, design specifications will be updated. Primary
37
-------
emphasis is on particulate loading and sizing at the boiler outlet and after
any ash reinjection device. Emissions and efficiency are being determined as
functions of overfire air admission, degree of flyash reinjection, excess air,
coal sizing and analysis, grate air velocity, etc., for various boiler-stoker
designs. Technical reports on test results for each individual boiler tested
and the final technical report at the conclusion of this program will be made
available to the public through the National Technical Information Service
(NTIS) . The reports will also be available through EPA, DOE, ABMA and KVB.
This paper will present some of the initial results of testing on the
first four boilers tested under the program entitled, "A Testing Program to
Update Equipment Specifications and Design Criteria for Stoker Fired Boilers,"
DOE Contract No. EF-77-C-01-2609,
38
-------
TEST RESULTS
OVERFIRE AIR
Overfire air jets were originally installed above the fuel bed on stoker
boilers to promote mixing of the product gases in the flame zone, increase
combustion and reduce smoke, Overfire air systems have undergone numerous
modifications over the years. Very little hard test data which detail the
impact of overfire air on particulate and ntiric oxide emissions and boiler
efficiency are available to the public.
With the goal of establishing the effectiveness of overfire air, overfire
air tests have been run on the first four boilers tested under the program
described above. In each case, particulate loading at the boiler outlet was
reduced by increasing the quantity of overfire air. Figure 1 illustrates the
reduction in particulate loading resulting from an increase in overfire air
pressure from 750 Pa (3 iwg) to 2500 Pa (10 iwg) at three boiler loads. The
unit tested was an 87 MW (300,000 Ib/hr) spreader stoker having two rows of
overfire air jets on both the front and rear wa.ter walls. In these tests,
particulate loading reductions of 25 to 50% were experienced. This reduction
was found to be due primarily to burnout of combustibles in the flyash and resulted
in a 0.8 to 2,0% increase in combustion efficiency.
In similar tests run on a 62 MW (200,000 Ib/hr) spreader stoker, 25%
reductions in particulate loading were measured. On a 66 MW (132,500 Ib/hr)
spreader stoker the particulate reduction was ten percent.
Tests on a 26 MW (90,000 Ib/hr) Vibra-Grate Stoker showed at full load
that the particulate loading was nearly independent of overfire air. The data
at full load are shown in Figure 2.
The effect on nitric oxide (NO) emissions of increasing overfire air
varied with the unit being tested. On the 87 MW spreader stoker (300,000 Ib/hr)
39
-------
nitric oxide emissions were reduced by an average five percent or 16 ppnt when
overfire air was increased. However, this reduction may not be statistically
significant. On the 26 MW boiler the nitric oxide showed scattered results
with changes in overfire air. On the 62 MW spreader stoker (200,000 lb/hr)
the nitric oxide concentrations again decreased but were not statistically
significant. By contrast, on several occasions nitric oxide concentrations
increased with increasing overfire air on the 66 MW C182,500 lb/hr) spreader
stoker by twenty percent or 65 ppm. Rationalization of this result will be
attempted when more data are available.
Tests such as those described above demonstrate the usefulness of overfire
air in controlling emissions and improving boiler efficiency. However, overfire
air use has limitations. Extensive use of overfire air adds to the dry gas
heat loss thereby reducing boiler efficiency; also, if the total excess air is
held constant and the undergrate air is reduced proportional to the increase
in overfire air, clinkering on the grate or grate overheating may result. Current
efforts are being directed toward defining how much overfire air is optimum
and where it can best be utilized in the furnace.
The first step toward this goal is to define quantitatively how much air
is actually (_in comparison with design estimates) entering the furnace through
the overfire air and the pneumatic flyash reinjection lines and relate it to
the total combustion air. Air flow measurements were made in the various
branches of the overfire air and reinjection air ductwork using a thermocouple
and standard pitot tube. The results of one such set of measurements on the
66 MW spreader stoker (.182,500 lb/hr) are presented in Figure 3. With this
figure the fractional contribution of overfire air and flyash reinjection air
to the total combustion air can be determined at any load and excess air.
Additional charts were made to represent the air flow from the individual air
headers as a function of header pressure so that a complete breakdown of air
infiltration to the furnace could be made for each test.
The overfire air was biased in different directions to determine where it
was contributing most to improved combustion. On the 66 MW (182,500 lb/hr)
spreader stoker, the front and rear rows of overfire air jets had an equal in-
fluence on particulate loading whereas increasing air flow through the lower
40
-------
rows of overfire air jets resulted in a particulate loading at the boiler outlet
which was twenty percent lower than when the flow through the upper rows was
increased.
The investigation into overfire air effectiveness will continue at each
site tested. The end results will be translated into design guidelines for
new units and operating guidelines for existing units equipped with overfire
air systems.
FLYASH REINJECTION
Today, the majority of spreader stoker installations reinject into the
furnace a portion of the flyash collected by the mechanical dust collector.
Selective type dust collectors are often specified which have the ability to
separate out the finest 30% of the collected flyash and allow reinjection of
only the larger particles which contain the highest carbon content. Total re-
injection from a high draft loss collector has been and is discouraged in the
vast majority of cases. The mechanically collected flyash (primarily above
ten micron size) contains 50 to 80% combustibles and may represent five percent
or more in potential fuel savings. One design criteria to be considered with
flyash reinjection is the question of how much of the reinjected ash is re-
entrained into the flue gas stream without settling on the grate and burning.
It may recirculate several times through the boiler until it is small enough,
either by successive heating and cooling, mechanical abrasion, or combustion,
to pass through the mechanical collector. This results in increased particulate
loading at both the boiler outlet and at the mechanical collector outlet.
Consequently, boiler tube erosion very likely increases when flyash is reinjected,
The importance of this effect has not been substantiated.
An example of f.lyash re-entrainment is given in Table 1. In this case,
flyash reinjection increased particulate loading at the boiler outlet by 315%
and at the mechanical collector outlet by 70%. The increase in collection
efficiency at the higher inlet loading is most likely due to a change in particle
size distribution resulting from the reinjected and re-entrained ash. Size
distribution data are not available at this time.
On the 87 MW (300,000 Ib/hr) spreader stoker, flyash reinjected from the
mechanical collector hopper increased combustion efficiency by 1.5 to 2.5% and
41
-------
increased particulate loading at the boiler outlet by 22 to 39%. Figure 4
illustrates this increase in particulate loading with reinjection for two test
sets at different boiler loads. On the 62 MW (200,000 Ib/hr) spreader stoker
the particulate loading increased 65% at the boiler outlet and 23% at the
mechanical collector outlet. At this time, it is believed that the differences
in particulate increase are due more to varying rates of flyash reinjection than
to design factors in the reinjection system.
The investigation into the impact of flyash reinjection on pollutant
emissions and efficiency of coal burning will continue at future test sites.
Computer modeling will be attempted to determine what fraction of the reinjected
flyash is being re-entrained r and what fraction of the re^-entrained ash is
subsequently collected and reinjected. It is a goal of this program that these
studies will lead to improved methods of reinjection,
EXCESS AIR
The impact of excess air on boiler efficiency is widely understood.
Reducing the excess air reduces the dry gas heat loss. Careful combustion air
control is also effective in reducing emissions. On the 87 MW (300,000 Ib/hr)
spreader stoker, a reduction in the level of excess O2 from six percent to five
percent resulted in particulate loading reductions of five to twenty percent.
Boiler efficiency was improved. Combustibles losses were reduced by an average
of 0.5% of the energy input. Dry gas losses were reduced by an average of 0.8%
of the total energy input. Nitric oxide emissions were reduced an average of
55 ppm.
Nitric oxide trend lines are being determined for each boiler tested.
Figure 5 presents the trend lines for the 87 MW (300,000 Ib/hr) spreader stoker.
It is well documented that nitric oxide concentrations increase with increasing
boiler load and excess 02. In stoker fired boilers the factors affecting fuel
nitrogen conversion to nitric oxide and the effects of various design changes
on nitric oxide emissions are not well understood. It is anticipated that one
of the outcomes of this program will be a better understanding of the formation
of nitrogen oxides in stoker fired boilers.
Carbon monoxide (CO) concentrations are being examined because they
are a very sensitive indicator of the onset of incomplete or unsatisfactory
combustion. In general, carbon monoxide has been four.d to remain at levels
42
-------
below 400 ppm under normal operating conditions. As the excess air is reduced,
a point is reached at which carbon monoxide begins to rise rapidly. This "CO
limit" is used to tune boilers for efficient and satisfactory operation and is
dependent on boiler load as shown in Figure 6. While studying the 87 MW unit,
it was found at 80% of full load CO increased rapidly as the excess oxygen was
reduced from 5.5% to 4.5% while at lower loads the limit was encountered at
lower excess oxygen levels. Other factors such as uniform fuel and air distri-
bution on and through the grate are important in reducing the "CO or malcombustion
limit," Unlike gas and oil fired equipment, the operating point in terms of
excess air cannot be set strictly by the smoke or CO limit without taking into
consideration grate temperature and clinkering. Because coal properties are
more variable than those of gas and oil, generally larger operating margins
must be established above the minimum excess air operating point.
COAL PROPERTIES
Three or four coals have been fired at each of the first three test
sites. In the 87 MW (300,000 Ib/hr) spreader stoker, a three percent
ash Wyoming coal produced one-third less particulates at the boiler outlet
than an eight percent ash Colorado coal. This is not to imply a direct relation-
ship between coal ash and particulate loading because other coal properties are
involved. However, this variable will be carefully studied throughout the
test program.
Coal size consist is being measured for each test, The ABMA has published
graphs stating recommended coal size consist for various stoker fired equipment.
According to this graph the percent coal passing a 1/4" square mesh screen
should be in the range of 20 to 50% for spreader stokers. Coal having 60%
passing 1/4" Csee Figure 7) was successfully fired on the 87 MW (300,000 Ib/hr)
spreader stoker on a day-to-day basis. Revision of these recommended limits
may be one of the outcomes of this program.
In the future, tests will be conducted in which coals will be screened at
the site to remove their fines thereby altering the size distribution. The
effect of coal fines on particulate emissions and operating problems such as
piling in front of the feeders will then be examined.
43
-------
GRATE AIR VELOCITY
Grate air velocity is a design and operating parameter which has been
used to reduce particulate emissions. Certain limitations prevent this para-
meter from being the total answer to particulate emissions. Among the limitations
are increased cost, reduced turndown ratio, and clinkering caused by low grate
air velocity. Reductions in grate air velocity usually reduce particulate
emissions. The velocity can be reduced by fabricating larger grates, reducing
load, reducing excess air, or increasing overfire air while holding total excess
air constant.
On the 62 MW C200,000 Ib/hr) spreader, all of the test data for particulate
loading at the boiler outlet correlated fairly well with grate air velocity as
shown in Figure 8, The particulate loading is in mass per unit energy input.
While such data is often described as showing particulates as a function of
load, the grate air velocity can often be reduced as much as twenty percent
either by increased use of overfire air, reduced excess air or by a twenty
percent increase in the design of the grate area.
Figure 8 includes data obtained on four different coals, various excess
air levels, different loads and changes in flyash reinjection rates. It will
be important to compare this figure with similar ones from other boilers.
COLLECTION EFFICIENCY
On the spreaders the multiclones were remarkably efficient. On the Vibra-
Grate stoker the opposite was true. Table II shows the comparisons. The
difference is primarily due to the fact that the particle size consist of the
particulates leaving the vibra-Grate Stoker is much smaller. Figure 9 shows
a comparison of the particle size of the ash leaving this boiler and one of the
spreaders. From the Vibra-Grate Stoker, the particles are much smaller.
In addition, it should be remembered that the particulate loading leaving
the Vibra-Grate stoker boiler is much less and, therefore, with the reduced
collection efficiency the emissions are similar.
44
-------
CONCLUSION
Much has been learned about stoker boiler operation and efficiency in
the short time this program has been in operation. Much more will be learned
in the coming months. When the final report is written, the knowledge
gained will be distributed through the ABMA Stoker Technical Committee to the
manufacturers of stoker/boiler equipment, the consulting engineers and the
boiler purchasers and operators. Then, when its use makes coal combustion
in industrial sized stoker fired boilers financially more competitive with
the combustion of fuel oil, and environmentally acceptable, then the program
will have been a success.
45
-------
ACKNOWLEDGMENTS
The authors wish to express their appreciation for the assistance and
direction given the program by project monitors, W. T. (Bill) Harvey of the
United States. Department of Energy (DOE) and R. E. (Bob) Hall of the United
States Environmental Protection Agency (EPA). Thanks are due to their agencies,
DOE and EPA, for co-funding the program.
We would also like to thank the American Boiler Manufacturers Association
(ABMA) staff memebers W. M. (Bill) Marx, Executive Director, W. H. (Bill)
Axtman, Assistant Executive Director, and B. C. (Ben) Severs, Project Manager,
and the members of their Stoker Technical Committee chaired by W. B. (Willard)
McBurney of the McBurney Corporation for providing support through their time
and travel to manage and review the program. The participating committee
members listed alphabetically are as follows:
F. C, Belsak Island Creek Coal
R. D. Bessette Island Creek Coal
T. Davis Combustion Engineering
J, Dragos Consolidation Coal
T. G. Healey Peabody Coal
W. B. Hoffmann Hoffmann Combustion Engineering
N. H. Johnson Detroit Stoker
K. Luuri Riley Stoker
J. Mullan National Coal Association
E. A. Nelson Zurn Industries
E. Poitrass The McBurney Corporation
P- E. Ralston Babcock and Wilcox
D, C. Reschley Detroit Stoker
R. A. Santos Zurn Industries
W. Sisken U.S. Department of Energy
We would also like to recognize the KVB engineers and technicians who spend
most of their time in the field, often under adverse conditions, testing the
boilers and gathering data for this program. Those involved to date are
46
-------
Jim Burlingame, Jon Cook, Jim Demont, Mike Jackson, George Moilanen, Russ
Parker, and John Rech.
Finally, our gratitude goes to the host boiler facilities who invited
us to test their boilers. At their request, these facilities will remain
anonymous to protect their own interests. Without their cooperation and
assistance this program would not have been possible.
47
-------
10000
en
c.
Q
O
_J
UJ
H
O
IT
Q.
8000
6000
4000
2000
MINIMUM •
OVERFIRE AIR Q.
-O
MAXIMUM
OVERFIRE AIR
1.2 1.4 16 1.8 2.0
GRATE HEAT RELEASE, W/m2
FIGURE 1
PARTICULATE LOADING VS OVERFIRE AIR ON AN 87 MW
(300,000 LB/HR) SPREADER STOKER AT CONSTANT EXCESS AIR
48
-------
600
o
H
o
a
O)
400
CO Q
I™
H
§
05
CD
•*-«
O
O
(-3
O
H
o
QL
8
en
en
G
O
200
1000
2000
3000
Overfire Air Pressure, Pascals
-------
80 60 40 20 0 '80
EXCESS AIR, % |
100 120 140 160 180
BOILER LOAD, Ib/hr STEAM x 103
FIGURE 3
NOMOGRAPH FOR DETERMINING CONTRIBUTION OF COMBINED OVERFIRE AIR
AND REINJECTION AIR TO TOTAL AIR ON A 66 MW (182,500 LB/HR SPREADER
STOKER. EXAMPLE: AT 140,000 LB/HR STEAM FLOW, 15 IWG OVERFIRE AIR
PRESSURE, AND 60% EXCESS AIR, IT IS DETERMINED THAT 30% OF THE COM-
BUSTION AIR IS INTRODUCED ABOVE THE GRATE AND CONVERSELY THAT 70%
OF THE AIR IS INTRODUCED THROUGH THE GRATE. (103 LB/HR STEAM = 0.364
MW; 1 IWG = 249 PA)
50
-------
10000
en
c
Q
O
_l
HI
<
cc
<
a.
8000
6000
4000
2000
FULL FLYASH
REINJECTION
REDUCED FLYASH
REINJECTION
1.2 1.4 1.6 1.8 2.0
GRATE HEAT RELEASE. W/m2
FIGURE 4
PARTICULATE LOADING VS FLYASH REINJECTION ON
AN 87 MW (300,000 LB/HR) SPREADER STOKER
51
-------
CM
O
ro
O 600
Q
UJ
£ 500
CE
O
O
oE 400
Q
CL
Q_
- 300
UJ
Q
X
O
O 200
tr
T
BOILER LOAD
80% 70% 60% 50% 40%
567
EXCESS OXYGEN, %
FIGURE 5
NITRIC OXIDE (NO) TRENDS VS EXCESS OXYGEN AND BOILER
LOAD ON AN 87 MW (300,000 LB/HR) SPREADER STOKER
52
-------
80% LOAD
70% LOAD
D 60% LOAD
O 50% LOAD
V 40% LOAD
567
EXCESS OXYGEN, %
FIGURE 6
CARBON MONOXIDE (CO) VS EXCESS OXYGEN AND BOILER LOAD ON AN
87 MW (300,000 LB/HR) SPREADER STOKER: A COMPOSITE OF 62
DATA POINTS GATHERED OVER A THREE-MONTH TEST PERIOD UNDER
VARYING CONDITIONS.
53
-------
AVG. AND STD. DEV.
OF-.COAL FIRED
ABMA RECOMMENDED
LIMITS
50 20
US STANDARD
SQ. MESH SCREEN,
inches
FIGURE 7
AVERAGE AND STANDARD DEVIATION OF SIX COAL SIEVE ANALYSES VS
ABMA RECOMMENDED LIMITS FOR SPREADER STOKERS. THIS COAL WAS
SUCCESSFULLY FIRED ON A DAY TO DAY BASIS ON AN 87 MW
(300,000 LB/HR) SPREADER STOKER (1 inch = 2.54 CM)
54
-------
7000 -
O)
c
6000
Q
O
_!
HI
I—
O
I-
QC
5000 -
4000 -
3000 -
© REDUCED REINJECTION
O ALL OTHER TESTS
0.2 0.3 0.4 0.5 0.6 0.7 0.8
GRATE AIR VELOCITY, m/s
FIGURE 8
RELATIONSHIP BETWEEN GRATE AIR VELOCITY AND PARTICULATE
LOADING ON A 62 MW (200,000 LB/HR) SPREADER STOKER
55
-------
FIGURE 9
PARTICULATE SIZING FOR A SPREADER STOKER (SITE B)
AND A VIBRA-GRATE STOKER (SITE D)
CD
N
C/)
T3
0)
•4—1
03
*-•
C/)
C.
03
Q)
15
E
C/)
•*—•
c
0
O
0)
Q.
99.9
Equivalent Particle Diameter, micrometers
56
-------
TABLE I
EFFECT OF FLYASH REINJECTION ON PARTICULATE LOADING ON A
66 MW (182,500 LB/HR) SPREADER STOKER AT MAXIMUM LOAD
FLYASH REINJECTION
PARTICULATE LOADING, NG/J MECHANICAL
(LB/106BTU) COLLECTOR
BOILER MECHANICAL EFFICIENCY
OUTLET OUTLET PERCENT
None
Boiler Hopper Only
2590
(6.0)
3010
(7.0)
Boiler Hopper and
Mechanical Collector Hopper 10750
(25.0)
207
(0.48)
213
(0.50)
350
(0.81)
92.0
92.9
96.7
57
-------
TABLE II
MECHANICAL COLLECTION EFFICIENCY
BOILER
TYPE
COLLECTION
EFFICIENCIES
PERCENT
87 MW
62 MW
66 MW
26 MW
Spreader
Spreader
Spreader
Vibra-Grate
70-94
93-97
92-97
25-75
58
-------
GUIDELINES FOR ADJUSTMENT OF RESIDENTIAL GAS BURNERS
FOR LOW EMISSIONS AND GOOD EFFICIENCY
By:
David W. Locklin
Battelle, Columbus Laboratories
Columbus, Ohio 43201
Robert L. Himmel
Douglas W. DeWerth
American Gas Association Laboratories
Cleveland, Ohio 44131
59
-------
GUIDELINES FOR ADJUSTMENT OF RESIDENTIAL GAS BURNERS
FOR LOW EMISSIONS AND GOOD EFFICIENCY
ABSTRACT
This paper reviews the basis for development of Guidelines for the
adjustment of residential gas burners to minimize air pollution and to
achieve efficient use of fuel. The Guidelines were developed jointly by
Battelle-Columbus and the American Gas Association Laboratories under con-
tract with the U. S. Environmental Protection Agency. They are intended for
use by service managers in training of technicians and by skilled service
technicians in maintenance and adjustment of gas burners used in residential
furnaces, boilers, and water heaters.
Emission characteristics of typical gas-fired equipment are presented in
the paper,and the Guidelines are summarized. Fourteen steps in the recom-
mended adjustment procedure are outlined with condensed descriptions. They
comprise preparation steps, adjustment steps, combustion diagnosis steps, and
final checks.
The Guidelines include input and reviewer comments from representatives
of the gas-heating industry, especially from committees of the American Gas
Association and the Gas Appliance Manufacturers Association. The document
is scheduled for distribution by EPA early in 1979 and is expected to be
reprinted for distribution by the gas industry.
60
-------
GUIDELINES FOR ADJUSTMENT OF RESIDENTIAL GAS BURNERS
FOR LOW EMISSIONS AND GOOD EFFICIENCY
INTRODUCTION
Air pollutant emissions from gas-fired residential heating equipment
are usually relatively low, compared to equipment firing other fuels.
Nonetheless, proper burner adjustment will contribute to maintaining low
emissions plus safe and efficient operation. Adjustment Guidelines for
atmospheric gas burners have been developed by the authors for distribution
by the U. S, Environmental Protection Agency* as "Guidelines for Adjustment
of Atmospheric Gas Burners for Residential and Commercial Space Heating and
Water Heating" (Reference 1). This paper is intended to provide an overview
of these Guidelines and the basis for their development.
The Guidelines were prepared for use by service managers in training of
technicians and by skilled technicians in gas burner servicing. By follow-
ing the step-by-step procedure outlined in the Guidelines, the skilled gas-
burner service technician will be able to adjust residential and commercial
heating equipment and water heaters to minimize air pollution, attain opti-
mum efficiency, and provide safe operation.
Similar Guidelines for oil burners were published by EPA (References 2
and 3) arid their development was described in APCA papers (References 4 and
5). Air-pollution control agencies may wish to promote the use of the
Guidelines in this series as a means of updating good practice by burner
servicing organizations to achieve both low emissions and efficient
operation.
EPA Contract No. 68-02-2653- Project Officer, R. E. Hall.
61
-------
TYPICAL EMISSION LEVELS
While burner design has an effect on the emission of NOX or CO, the
appliance to which the burner is applied is also important. Forced-air
furnaces, water heaters, space heaters and boilers all have different heat
transfer rates, combustion intensities, recirculation patterns, and flame
aerations. Each of these factors can influence the formation of NOX as much
as burner design parameters.
Emissions for Various Gas Appliances
Table I presents emissions measured by the A.G.A. Laboratories (Refer-
ences 6, 7, and 8) for several types of gas heating appliances operating in
the laboratory. These data were obtained for units operating with well-
adjusted flames after an equilibrium operating condition was achieved. All
samples were taken downstream from the draft hood, because that point is
representative of what actually is emitted into the atmosphere. (Note: For
determining combustion and appliance efficiency of a field-installed unit,
the recommended sampling point for measuring C02, CO, and temperature is
before the draft hood.)
All values of NO, N0~ and CO presented in Table I are on an air-free
basis, based on an ultimate C02 of the natural gas fuel used of 12.0 per-
cent. An air-free basis, or an emission factor basis (pounds per million
Btu of natural gas burned) is necessary in order to make comparisons.
To obtain the actual sample concentration in the flue gases emitted to the
atmosphere, the air-free concentration is divided by the ratio of the ulti-
mate C02/sample C02-
These data on emissions are for conventional gas burning equipment with
typical burner adjustments and are in general agreement with those reported
by others (References 9 and 10). The data are also representative of
emissions from current burner designs; however, R&D efforts have shown that;
lower NOX emissions are feasible with some burner designs.
62
-------
Effects of Burner Adjustment
Table II presents average emissions for 38 forced-air furnaces
equipped with atmospheric-injection burners. The overall average is sub-
divided into averages for single-port burners and for multiport burners. In
all cases, the yellow-flame condition (which represents a misadjusted flame)
showed higher CO and aliphatic aldehyde levels when compared to the well-
adjusted blue-flame condition. With the yellow-flame condition, average
emissions of NO were lower, but N02 emissions were higher; the combined
result, expressed as NOX, averaged slightly lower for the yellow-flame
condition.
Tables I and II also show that NOX levels for single-port burners were
slightly less than for multiport burners. This small difference is statisti-
cally significant, but may not be practically significant. We believe the
difference is due to the following:
• Multiport burners provide more sites for near-
stoichiometric flame fronts, with attendant maximum
flame temperatures which cause higher NOX production,
• Single-port burners generally have larger flame volumes
with less regions of high intensity which lowers the
potential for optimum NOX production, and
• The longer flames of single-port burners allow more
time for NOX decay.
TYPICAL CHARACTERISTICS OF NATURAL GAS-FIRED
BURNERS IN HEATING APPLIANCES
Figure 1 shows flame regions of stable blue flames and yellow-tip
flames controlled by gas input rate and primary air adjustment for typical
atmospheric gas burners; the region of undesirable yellow flames occurs at
high input rates and low primary-air aeration. (Note in Table II that
emissions of CO and aldehyde were increased when operating burners in the
yellow-flame combustion region.)
63
-------
Figure 2 illustrates the effects of combustion air on CO and CC>2 con-
centrations for typical atmospheric gas burner installations. It demon-
strates that the CO is very low at the proper C02 adjustment region. The
figure also shows that the C02 reading does not always predict the CO level.
With a deficiency of combustion air, it is still possible to obtain a high
C02 level, but with an undesirable CO level. Therefore, C02 level alone
cannot predict the CO concentration. For this reason, it is necessary to
measure the CO concentration to be certain it is at an acceptable level.
It should be recognized that gas burners, as applied to matched furnace-
burner or boiler-burner units, do not have as wide a range of air adjustment
available to the service technician as do residential oil burners. The pri-
mary air setting adjustment has a realtively narrow range of adjustment, and
the design of the appliance mainly determines the secondary air supply. For
conversion burners, the service technician usually has more range of adjust-
ment (Reference 11).
Figure 3 demonstrates the expected effect of excess air on CO, NOX, and
hydrocarbon (HC) emission of a gas burner installation by showing the rela-
tionship of these emissions to excess air supply (or C02 value). This
figure shows the normal operating excess-air region for natural and LP gas
appliances in terms of excess air.
CO emission levels are at unacceptable levels under a deficiency of
combustion air. HC emissions also are high with a deficiency of combustion
air and may increase under very high excess-air conditions because of lift-
ing flames. The NOX (NO + N02) emission reaches a peak very near a stoichi-
ometric air-gas mixture (zero percent excess air) because the maximum flame
temperature occurs at this condition.
Figure 3 shows that it is desirable to operate under an excess air con-
dition of from about 40 to 90 percent in order to minimize the pollutant
emission levels and maintain good efficiency. Most atmospheric gas furnaces
will perform as shown. For some burners, the normal region will be wider or
narrower; a few burners will operate satisfactorily outside the region
64
-------
shown." The general trend of the pollutant emission, however, will be similar
to that shown by the figure.
The service technician should visualize the interaction of emissions
and burner adjustment to develop an appreciation that the relationship
between CO and C02 is basic to setting a burner for minimum emissions by use
of field measurements. Note that evaluation by using only CC^ measurements
could result in high CO and HC. This can be avoided by measuring both the
C02 and CO levels. The heating industry has traditionally used C02 readings
as the basis for burner adjustment, mainly because of the availability of
effective field-type instruments. However, 02 measurements offer the advan-
tage of positively identifying whether operation is in the fuel-rich or
fuel-lean region. While instruments for field measurement of 02 are being
introduced (Reference 12), and may ultimately gain acceptance, the Guidelines
that follow are based on C02 and CO measurements.
RECOMMENDED ADJUSTMENT PROCEDURES
Important criteria in making gas burner adjustments are:
• To provide safe and reliable operation,
® To provide efficient fuel utilization, and
® To minimize emissions of air pollutants.
The Guidelines summarized below are intended for adjustment of atmospheric-
type gas burners for residential and commercial space heating and for water
heating systems, The procedures are supplemental to manufacturers' instal-
lation instructions, the National Fuel Gas Code (Reference 13), other
installation codes or other handbooks (Reference 14), and apply to burners
fired with either Natural Gas or Liquified Petroleum (LP) gases.
The adjustment procedures apply to atmospheric-type gas burners of the
single-port and multiported types such as ribbon, slotted-port and drilled-
port as used in manufacturer-designed burner/furnace units, burner/boiler
units, or complete water-heater units. Although the principles apply to
65
-------
conversion-type gas burners, these burners require special adjustment proce-
dures, and the burner manufacturer's instructions should be followed.
The following steps are emphasized in the Guidelines from the viewpoint
of minimizing air-pollutant emissions, plus maintaining good efficiency and
safe operation. A total of fourteen steps are outlined in four different
categories: preparation steps, combustion adjustment steps, combustion
diagnosis steps, and final checks. A condensed summary follows.
PREPARATION STEPS
1. Clean Burner
and Pilot
2. Adjust
Manifold
Pressure
3. Adjust Pilot
4. Adjust Main
Burner Input
The most common malfunction of main burners and pilots
is caused by excess dirt, lint, or other debris. Block-
age of primary-air openings causes the burner to oper-
ate with incomplete combustion and can lead to soot and
excess CO formation. Remove burners and clean the bur-
ners, burner ports, burner and pilot orifices and pri-
mary air openings. Make preliminary adjustment of the
primary air shutters, if any, by following step 6.
Adjust pressure regulators to deliver gas
to burners at the proper pressure. If the appliance
rating plate specifies a manifold pressure, adjust to
that pressure; if not specified, set pressure as
follows:
Natural Gas:
3.5 inches water column
(Appliances with regulators)
LP Gases: 11 inches water column
(Regulator may be at tank)
The pilot flame should be soft blue in color. Propane
pilots may have slightly yellow tips. The flames
should surround the tip of the thermocouple or flame
sensor and, if there is a separate ignition port, flame
should reach the main burners. A pilot adjusting screw
is usually located on the main gas valve or pilot shut-
off valve.
After the unit has been operating at least 10 minutes,
check the gas input rate by timing one revolution of
the gas meter test dial, making sure that all other
gas-burning appliances are not operating. Note the
number of seconds for one full revolution, and
calculate input as follows:
66
-------
5. Check Ignition
-r /i cubic ft per rev ^rnn ^,
Input, Btu/hr = -, £- x 3600 x HV
seconds per rev
where HV = gross heating value of gas, Btu/cubic foot.
Adjust the input to within ±5 percent of the value
noted on the rating plage by adjusting the pressure
regulator or by changing burner orifice.
Allow the burner to cool and check operation of igni-
tion system by cycling several times while the burner
is cold to ensure prompt ignition on cold start. Then
operate the burner at least 10 minutes, or until thor-
oughly hot, and repeat several ignition cycle checks to
ensure prompt ignition on hot start after about 5 sec-
onds OFF time.
COMBUSTION ADJUSTMENT STEPS
6, Adjust Burner
7. Sampling
8.
Check C02
To adjust primary air, allow at least 10 minutes for
the burners to heat up, close primary air shutter until
yellow tips appear, then open air shutter until all
yellow completely disappears. The flame should be a
clear blue color. Too much primary air causes noisy,
hard flames which could result in flame lifting, flash-
back and emission of unburned hydrocarbons. Too little
primary air causes yellow flames which could produce
high CO and soot emissions. (See Figure 1)
Secondary air is controlled by the burner and appliance
design and should not need modification. However, con-
version burners may need secondary-air adjustment and
the manufacturer's instructions should be followed.
(Reference 11)
Flue gas samples for analysis should be taken from in-
side the appliance approximately 1 inch upstream of
the draft hood inlet. C02 and CO are usually measured
with instruments using color-sensitive indicators. If
several flue outlets are provided, individual samples
should be taken from each, or a partial sample taken
from each outlet to make up the full sample.
If it is desired to assess the burner adjustment in
terms of the best practical efficiency, follow this
procedure. After the burner has been in operation for
at least 10 minutes, measure and record the C02 in the
flue outlet (following the instructions provided by
the C0£ instrument manufacturer).
For safe and most efficient operation, the C02 concen-
tration should be near 8.5 percent for Natural gas and
67
-------
9. Check CO
10.0 percent for LP gases. See Figures 2 and 3 for
typical operating regions.
This step is important for safety, whether or not step
8 is followed. After the burner has been in operation
for at least 10 minutes, make a CO measurement (follow-
ing the CO instrument manufacturer's instructions).
Typical CO values for properly adjusted appliances,
operating at rated input, range from 25 to 100 ppm
(0.0025 to 0.010 percent); however, values as high as
300 ppm may be encountered in the flue sample,since
this is equivalent to the maximum allowable value of
400 ppm (air-free) at a C0£ of 8.5 percent as allowed
by the ANSI standards (References 11, 13, and 14).
COMBUSTION DIAGNOSIS
10. Check
Performance
11. Efficiency
Checks
A well-adjusted gas burner should be capable of opera-
ting with a CO level not greater than 100 ppm in the
flue gases with a C02 level between 6.0 and 8.5 percent
for Natural gas and 7.0 to 10.0 percent for LP gases.
If these values cannot be reached, check the following:
a. Too low a C02 level could mean circulating air
leakage into the combustion air, due to a faulty
seal around the secondary air opening of the unit.
Too low a C02 level could be caused by a defect in
the furnace heat exchanger. If the C0£ levels ob-
tained with and without the blower operating differ
by more than 0.4 percent, a check of the combustion
chamber and heat exchanger should be made.
b. If the C02 and/or CO are too high, this could mean
an improper combustion air supply. Spillage from
the appliance draft hood due to negative pressure
in the room also indicates inadequate air supply.
Spillage will also result from a blocked chimney or
vent. If inadequate air supply is suspected, refer
to Section 1.3.4 of the National Fuel Gas Code
(Reference 13) or comparable provisions in local
codes for specific detailed recommendations for
providing air for combustion and ventilation.
To determine the steady-state efficiency, follow this
step. Measure the flue-gas temperature after at least
10 minutes of operation at the same point that the C02
sample was taken. Determine the net flue-gas tempera-
ture by subtracting room temperature from the flue-gas
temperature reading.
68
-------
Using this net temperature value and the percent of
C02, determine the .appliance flue loss from Figure 4.
The appliance thermal efficiency is equal to (100 -
flue loss). Use the values shown in Table III to deter-
mine whether the unit is operating at a satisfactory
efficient level. Note that this gives steady-state
efficiency. Seasonal efficiency will be less, due to
flue losses occuring during OFF cycles. These losses
are minimized by avoiding overfiring that occurs when
gas input rate exceeds that needed to satisfy the
actual steady-state heat losses.
FINAL CHECKS
12. Ignition
13. Controls
14. Regular
Cleanup
Check operation over several repeated cycles at about
5-second intervals, to ensure prompt ignition.
Check setting of all operating and limit controls be-
fore leaving installation. For modern forced-warm-air
systems, usual good practice is to set the fan control
at 115 to 125 F ON and 90 to 100 F OFF for practical
performance and good efficiency.
Annual burner checkup is recommended, with cleaning as
necessary.
Experienced service technicians will observe that these procedures are
essentially the same as they have followed in normal good practice. Adjust-
ment of gas burners by these procedures will help to provide safe and reli-
able operation, to minimize emissions, and to provide efficient fuel
utilization.
ACKNOWLEDGMENTS
Significant inputs to the development of these Guidelines were made by
a number of reviewers from the gas industry, the equipment servicing indus-
try, and others. Special acknowledgment is made to reviewers representing
the following:
69
-------
American Gas Association (A.G.A.)
- by members of the A.G.A. Customer Service and
Utilization Committee
Gas Appliance Manufacturers Association (G.A.M.A.)
- by members of the Joint A.G.A./G.A.M.A. Customer
Service Task Group
Air Pollution Control Association (APCA)
- by members of the APCA Residential Fuel Combustion
Committee.
REFERENCES
1. Himmel, R. L. , D. W. DeWerth, and D. W. Locklin. Guidelines for Adjust-
ment of Atmospheric Gas Burners for Residential and Commercial Heating.
U. S. Environmental Protection Agency, Document prepared under EPA
Contract 68-02-2653, 1979.
2. Locklin, D. W. and R. E. Barrett. Guidelines for Residential Oil-
Burner Adjustments. EPA Report No. EPA-600/2-75-069a, U. S. Environmen-
tal Protection Agency, NTIS Report PB-248292, October 1975.
3. Locklin D. W. and R. E. Barrett. Guidelines for Burner Adjustments of
Commercial Oil-Fired Boilers. EPA Report No. EPA-600/2-76-088, U. S.
Environmental Protection Agency, NTIS Report PB-251919, March 1976.
4. Locklin, D. W., R. E. Barrett, and R. E. Hall. Guidelines for Residen-
tial Oil-Burner Adjustment to Minimize Pollutant Emissions. APCA Paper
No. 47-24.3, presented to APCA Annual Meeting, Toronto, Ontario, June 20-
24, 1977.
5. Barrett, R. E., D. W. Locklin, and R. E. Hall. Field Investigation of
Emissions from Commercial Boilers. APCA Paper No. 76-27.7, presented to
APCA Annual Meeting, Portland, Oregon, June 27-July 1, 1976.
6. DeWerth, D. W. and R. L. Himmel. An Investigation of Emissions from
Domestic Natural Gas-Fired Appliances. SP-8 Proceedings, 67th Annual
APCA Meeting, Denver, Colorado, June 1974.
7. Belles, F. E., R. L. Himmel, and D. W. DeWerth. Measurement and Reduction
of NOX Emissions from Natural Gas-Fired Appliances. 68th Annual APCA
Meeting, Boston, Massachusetts, June 1975.
8. Himmel, R. L., E. H. Tausch, and D. W. DeWerth. Further Measurements of
Emissions from Gas-Fired Appliances. 70th Annual APCA Meeting, Toronto,
Ontario, June 1977.
70
-------
9. Levy, A., S. E. Miller, R. E. Barrett, E. J. Schulz, R. H. Melvin,
W. H. Axtrnan, and D. W. Locklin. A Field Investigation of Emissions
from Fuel Oil Combustion for Space Heating. API Publication No. 4909,
American Petroleum Institute, November 1, 1971.
10. U. S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors, AP-42, 2nd Edition, April 1973. Supplement No. 3,
July 1974.
11. American National Standards Institute. Installation of Domestic Gas
Conversion Burners. American National Standard Z21.8-1971, American
Gas Association, Arlington, Virginia, 1971.
12. Torborg, R. H. and U. Bonne. Instrumentation for Adjusting Burners for
Improved Efficiency. APCA Paper No. 78-49.6, APCA Annual Meeting,
Houston, Texas, June 1978.
13. American National Standards Institute. National Fuel Gas Code. Ameri-
can National Standard, Z223.1-1974, American Gas Association, Arlington,
Virginia, 1974.
14. American National Standards Institute. Gas-Fired Gravity and Forced-
Air Central Furnaces. American National Standard Z21,47-1973, American
Gas Association, Arlington, Virginia, 1973.
15. Griffiths, J. C. Method of Calculating the Flue Loss of Gas-Fired
Equipment. American Gas Association Laboratories Research Report No.
1509, Catalog No. U07176, Cleveland, Ohio, 1976.
71
-------
Lifting Flames
-------
CO
"o
o
k.
ex
03
O
&
Q.
e
_g
"o
c.
2 concentration for a
typical natural gas and LP gas burner showing proper
adjustment region.
73
-------
g
'in
en
"e
o
o
-------
FOR NATURAL GAS
Example: 8.2 percent C02
22 percent flue Joss
(78 percent efficiency)
F net
Figure 4. Nomograph for determining flue loss and steady-state
efficiency from CC>2 and flue-gas temperature for
natural gas firing (Reference 15).
Limited to use with natural gas as follows:
970 - 1100 Btu/standard cubic foot
0.57 - 0.70
Heating value, gross:
Specific gravity:
Ultimate CO
2:
11.7 - 12.2 percent
75
-------
TABLE I. EMISSIONS FROM TYPICAL HEATING APPLIANCES
WITH NATURAL-GAS FIRED ATMOSPHERIC BURNERS
Appliance and Burner Type
Forced Air Furnaces
Water Heater
Domestic
Commercial, Multiport
Commercial, Single Port
Hot Water Boilers
Multiport
Single Port
Steam Boilers
Multiport
Number
of
Units
38
17
3
1
11
3
4
Sample
C02,
Percent
5.8
5.4
5.4
5.1
4.3
3.4
4.6
Average Flue Gas
Concentration,
(Air Free) PPM
CO NO N02
8.1 88.8 4.7
7.0 110.1 4.8
6.6 179.6 5.3
3.8 74.1 3.7
61.0 146.6 12.0
11.0 89.4 6.9
69.0 143.9 10.9
NOX
Emission
Factor, . .
lbs/106 Btu^
0.098
0.120
0.194
0.084
0.168
0.102
0.161
(a) Sampled downstream of the draft hood during steady-state operation with 5 feet of
vertical stack
(b) Sum of the NO and N02 calculated as N02
-------
TABLE II. EFFECT OF BURNER ADJUSTMENT MADE ON AVERAGE EMISSIONS
FROM NATURAL-GAS-FIRED FORCED-AIR FURNACES^a^
Average
Overall
Multiport
Burner
Single Port
Burner
Sample
C02,
Flame'"' Percent
Blue
Yellow
Blue
Yellow
Blue
Yellow
5.
6.
5.
6.
5.
5.
8 ±1.0
0 ±1.0
9 ±0.9
0 ±0.9
5 ±1.1
8 ±1.1
Average Flue Gas Concentration, ppm, Air Free
8.
208.
8.
201.
7.
255.
CO
1 ±2.6
0 ±4.0
2 ±2.5
0 ±3.0
8 ±2.1
0 ±6.0
88.
73.
94.
80.
75.
56.
NO
8 ±14.0
6 ±16.0
0 +11.7
0 ±10.0
0 ± 9.6
9 ±17.6
4.
9.
4.
8.
4.
12.
N02
7 ±2.4
7 ±3.9
8 ±2.5
8 ±2.6
4 ±2.2
1 ±5.5
0.
0.
0.
0.
0.
0.
HCHO(C)
18 ±0.14
60 ±0.39
20 ±0.14
62 ±0.40
14 ±0.15
60 ±0.40
NOX Emission
Factor,
lbs/106 Btu(d
0.098
0.088
0.104
0.093
0.084
0.073
±0
±0
±0
±0
±0
±0
.015
.014
.012
.010
.010
.014
(a) Sampled downstream of the draft hood during steady-state operation with 5 feet of vertical stack.
The ± value is the standard deviation of the average shown.
(b) Yellow flames are undesirable and result from a misadjusted burner.
(c) Total aliphatic aldehydes expressed as formaldehyde (HCHO).
(d) Sum of NO and N02 calculated as N02-
-------
TABLE III. TYPICAL STEADY-STATE THERMAL EFFICIENCIES
OF CONVENTIONAL GAS-FIRED RESIDENTIAL AND
COMMERCIAL HEATING EQUIPMENT
(a)
Thermal Efficiency, Percent
System
Forced Air Heating
Gravity Heating
Hot Water or Steam Boiler
Domestic or Service
Water Heater
Good
Practice
75 or more
70 or more
75 or more
72 or more
Acceptable
72 -
67 -
72 -
68 -
75
70
75
72
(a) Overall Thermal Efficiency = 100 - flue loss, percent,
where flue loss is expressed as a percent of the gas
input (gross heating value) that is lost up the flue in
steady-state operation. This assumes that any jacket
loss is useful heat.
(b) A field measurement program by four gas utilities showed
that the "as found" thermal efficiency of about 200
furnace installations had a mean of 74.5 percent thermal
efficiency with a standard deviation of 3.77 percent.
78
-------
TABLE IV., CONVERSIONS TO SI UNITS
The Guidelines have been prepared using English engineering units,
as commonly used in the gas heating industry and as most familiar
to the service technician. xhis provides factors for conversion
to SI units, if desired.
Multiply
Btu
cubic feet
Btu/cubic foot
cubic feet/min
degrees F
inches
pounds /million Btu
by
1.056
2.832 x 10~2
37.26
4.720 x 10~4
| (F - 32)
2.54
0.430
To Obtain SI Units
kilo joules
cubic meters
kilojoules/cubic meter
cubic meters/sec
degrees C
centimeters
grams /Mega j oule
kJ
3
m
kJ/m3
m /sec
cm
g/MJ
79
-------
FIELD VERIFICATION OF
LOW-EMISSION INTEGRATED RESIDENTIAL FURNACES
By:
A. S. Okuda and L. P. Combs
Rockwell International
Environmental and Energy Systems Division
Energy Systems Group
Canoga Park, California 91304
81
-------
ABSTRACT
A prototype low-emission, improved-efficiency, oil-fueled, warm-air
residential furnace was previously developed under EPA sponsorship. In
laboratory tests, the prototype furnace produced approximately 65% lower
emissions of oxides of nitrogen (NO), and, at the same time, operated at high
A
thermal efficiency (Reference 1).
Six units of the prototype low-emission furnace were field tested during
the 1977-78 winter heating season in homes in the Boston, Massachusetts, and
Albany, New York, areas. Their emissions behavior was monitored by making
monthly measurements of flue gas composition with a mobile laboratory. Their
performance was monitored with a set of continuously recording time, tempera-
ture, and flow rate instrumentation.
All six field-tested prototype furnaces functioned satisfactorily through-
out the heating season with only a few minor shakedown problems. The emission
goals, set very close to what had been achieved in the laboratory, were met.
The NO emissions goal was 0.65 g/kg fuel burned - a reduction of 65% from the
average of existing furnaces. Emission levels of carbonaceous pollutants -
smoke, carbon monoxide, and unburned hydrocarbons —were also acceptably low.
None of the furnaces exhibited significant drift of any emission level during
the season.
Steady-state thermal efficiencies were approximately 83 to 84%, while cycle-
average thermal efficiencies averaged approximately 74%. Fuel consumptions were
compared with those of the predecessor furnaces during the 1975-76 and 1976-77
heating seasons by normalizing to a per degree-day basis. All six units con-
sumed significantly less fuel, averaging 18.5% less than the prior two seasons.
The data base is currently being expanded by continuing the field testing
for a second full heating season.
82
-------
ACKNOWLEDGMENTS
Work on which this paper is based was performed pursuant to Contract
68-02-2174 for the U.S. Environmental Protection Agency, Industrial Environ-
mental Research Laboratory, Research Triangle Park, North Carolina. The
support of the EPA and the advice and helpful discussions with the EPA project
officer, Mr. G. Blair Martin, are gratefully acknowledged. Views expressed in
the paper are the authors', however, and do not necessarily reflect the views
or policies of the EPA, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
Acknowledged with appreciation are Mr. Clifford Hauenstein, Program
Manager at Rockwell International, Messrs. Clark Zeh and Ronald VonRonne of
Main-Care Heating Service, Delmar, New York, and Mr. Joseph lorio, Sr., of
Atlantic Heating and Air Conditioning Company, Brook!ine, Massachusetts, for
their cooperation and contributions leading to the successful execution of the
1977-78 field test.
83
-------
INTRODUCTION
Under a previous EPA contract, design criteria were determined whereby
gun-type pressure-atomizing oil burners, such as are commonly used in resi-
dential and commercial space-heating systems, may be modified so that they
produce substantially lower emissions of oxides of nitrogen (NOX) and burn
smoke free at more efficient operating conditions. Those design criteria were
used to modify existing burners of two sizes - a 1.05-ml/s (1.0-gph) residen-
tial burner, and a 9.47-ml/s (9.0-gph) commercial burner - and were shown to
be valid in laboratory testing (Reference 2).
Further laboratory research with the residential-size, low-emission
burner provided additional design criteria for fireboxes matched to the burner
to achieve even lower NO emissions (Reference 3). Thereafter, proof-of-
A
concept experiments were carried out in the laboratory using a prototype
residential warm-air furnace embodying several design modifications (Refer-
ences 4 and 5). The resultant NO emissions were reduced to about 35% of the
/\
estimated average established by a field survey (Reference 6) of comparable,
existing, installed units. Further, the laboratory performance data showed
that the prototype furnace's cyclical efficiency should be 10 or more percent-
age points higher than the estimated average of the existing residential oil
furnace population.
The present investigation is a logical continuation from those encouraging
laboratory results. Potential benefits of commercializing the derived tech-
nology are being demonstrated by conducting field tests of several low-emission
furnaces in actual residences. First, however, it was appropriate to effect
further refinements of the burner and furnace designs, partially to further
optimize emissions and efficiency performance and partially to improve
commercial producibility. Thus, the two objectives were: (1) beginning with
84
-------
the prototype furnace tested earlier, to optimize further the design of an
integrated, low-emission, high-performance, oil-fueled, residential warm-air
furnace, and (2) to verify its pollutant emissions and thermal efficiency
performance by operating units over an entire winter heating season in actual
residential installations.
The task of further design optimization was addressed in the first phase
of this study. The approach taken was to make modifications in the prototype
low-emission furnace described in Reference 4 and to retest the unit in the
Rockwell furnace laboratory. Eventually, this led to the construction and
verification testing of a second, all-new integrated furnace unit. The
results of this effort are summarized in a paper presented at EPA's Second
Stationary Source Combustion Symposium (Reference 1).
The above effort was carefully scheduled so that the second phase, field
verification, would be initiated early summer of 1977 to allow time for con-
struction and installation of six integrated furnaces for testing in the
1977-78 winter heating season. The effort and the results of the 1977-78
heating season field testing of the integrated low-emission furnaces are the
primary subjects of this paper.
85
-------
FIELD TEST PREPARATION
DESCRIPTION OF THE INTEGRATED FURNACES
The effort in the first phase of this contract (Reference 1) was to
finalize a configuration that incorporated the low-emissions design criteria
and applied fuel-conserving techniques into a viable, commercially producible
furnace unit. Six furnaces of this configuration would then be constructed
for field testing.
The most expedient way of constructing only six furnaces is to purchase
commercially available units and institute the desired modifications. The
Lennox Industries Model 011-140 furnace was selected as the foundation for the
modifications because of its compact heat exchanger and outside dimensions.
Figure 1 is a line drawing of the integrated furnace configuration surrounded
by photographs of the modified components. Most components are minor refine-
ments of laboratory prototypes. The finned firebox, however, received major
refinements for the field test configuration, changing from welded-steel
fabrication to cast-iron formed for mass productbility. The fin heights and
placements were changed for better temperature distribution to reduce internal
stresses, thereby increasing potential service life. A carbon-steel ring
integrally attached into the top of the cast firebox facilitated mating of the
stock heat exchanger by conventional welding.
Baffles were added along the side walls (see Figure 2) to assure air flow
to the finned firebox. These baffles continue up alongside the heat exchanger
panels to reduce bypass flow. The firebox and heat exchanger compartment is
insulated on all sidewalls with 1-in. "Cerafelt" insulation, reducing cabinet
losses to near zero.
The burner, also shown in Figure 2, incorporates an optimized head, quiet
stator, oversized static disc, and a standby draft control assembly with
86
-------
microswitch safety circuitry. The burner head is a nonretention type, using
the plug-flow, low-emissions design criteria and fabricated from 22-gauge,
Type 310 stainless steel sheet stock. To reduce feed-system induced pulsa-
tions, a quiet stator on the blower and an oversized static disc in the blast
tube are incorporated in the optimum burner. The draft control assembly
serves a twofold purpose of reducing standby heat losses by convection, and
thereby maintaining the temperature of the firebox for a cleaner start on the
next firing.
The integrated furnace system is completed with a "sealed-air" subsystem,
which transports outside air to the burner and the flue damper. This reduces
dwelling heat losses through the burner and damper. The subsystem consists of
7-in. diameter insulated ducting from an opening to the outside, to an enclo-
sure around the flue damper, and continuing on with 4-in. diameter ducting to
the top of the burner vestibule. Immediately under the entry to the vestibule
is a filter box to assure consistent burner blower characteristics. The
louvers in the vestibule cover are closed over by sheet metal, thereby
"sealing" this air system from the dwelling.
The complete integrated system is approximately 75 Ib heavier than the
original Lennox 011-140, primarily because of the intentionally massive heat-
retaining firebox.
TEST LOCALE SELECTION
Three field test units were located in each of two test locales where oil
heat is used in a substantial fraction of single-family homes and which have
distinctly different winter climates. Winter climatic data averaged over
several years were reviewed for approximately 35 cities, most of which are
distributed throughout the New England, Great Lakes, and Northern Midwestern
States. All cities considered have normal accumulative degree-days (based on
65°F) exceeding 4500 for the 9-month period of September through May. Atten-
tion was also directed to nonclimatic factors such as oil usage patterns,
predominant types of residential construction, state and local code require-
ments, availability of local support, and logistical requirements. The cities
selected for the field tests were Boston, Massachusetts, and Albany, New York.
87
-------
Boston's Atlantic shoreline location provides a maritime climate, and the
Appalachian mountain range isolates Albany, New York, to a continental
climate.
With the completion of the selection of the two cities, effort was
immediately started on establishing subcontractor support. Both cities were
visited by Rockwell personnel, the field test plan was presented to several
contractors, and, through standard bidding procedures, commitments were estab-
lished with two of the contractors, one in each city.
SELECTION OF HOST SITES
A list of potential host dwellings was supplied by each of the two
service contractors from their existing customer lists. The selection of the
six host sites (single-family dwellings) was based on criteria that included
consideration of (1) dwelling construction, (2) dwelling location, (3) exist-
ing dwelling furnace installation, (4) fuel usage history, and (5) host family.
The final selection resulted in six dwellings of a variety of characteristics
that are listed in Table I. The dwellings range from 14 to 80 years of age,
one to five occupants, and a poor fuel utilization of 3.52 to the best at
11.48 degree-day per gallon of fuel. This fuel utilization factor is based on
the fuel consumption and climatologica'l histories of the prior two heating
seasons. All of the dwellings, with the exception of Woodbine, have insula-
tion or insulative design incorporated in the wall surfaces. The Woodbine
dwelling has no insulation and the greatest exposed wall surface, and the
effect on fuel utilization is very obvious in Table I. All host dwellings
have been firing No. 2 fuel oil at 0.79 ml/s (0.75 gph), which is the same as
the installed integrated furnace, so as to alleviate any changes in efficiency
induced by changes in rate of heat delivery.
The furnaces being replaced represent a wide spectrum, from a 20-year
old unit representing furnaces near the end of their service life, to a 3-1/2
and a 5-year old Lennox 011-140 representing the most modern-design, commer-
cially available furnace. Coincidentally, the Lennox 011-140 is the foundation
of the integrated furnace body and was selected because of its compact design.
-------
FIELD INSTRUMENTATION
Monitoring of the pollutant emissions is the primary scientific objective
of the field test; however, the effect of pollution reduction devices on fuel
utilization is also of great concern in these times of energy shortages.
Therefore, the field test instrumentation consisted of two sets of equipment
with different monitoring techniques, each set dedicated to one of the two
areas of concern. The pollutant emissions measurement equipment consisted of
a number of gas analyzers, assembled into a mobile unit and transported to
each site on a periodic basis. The thermal efficiency measurement equipment
consisted of one programmable data logger system and five supplementary
recorders, all operating on a continuous monitoring basis.
Emissions Monitoring Mobile Laboratory
The pollutants of concern are oxides of nitrogen (NO ), carbon monoxide
)\
(CO), unburned hydrocarbons (UHC), and smoke. Carbon dioxide (C02) and oxygen
(02) are components also measured to accurately establish the burner operating
conditions. The instruments to measure these gas components at flue gas
concentrations were available in the Rockwell furnace laboratory. The instru-
ments were removed from the stationary consoles and remounted, along with the
required supplementary systems, into two rack assemblies designed to fit into
the freight bay of a 3/4-ton Ford van. The two assembled instrumentation
modules are shown in Figure 3 alongside the carrier vehicle. Visible on the
modules are the primary instruments, strip chart recorders, hydrogen generator,
calibration gas cylinders, sample pumps, gas valves, and flowmeters. Figure 4
shows Instrument Module 1 installed in the vehicle on rubber vibration isola-
tion mounts, six of which are visible in the photograph along the bottom of
the module. The enclosed bay of the vehicle is completely insulated with fire-
retardant "ethafoam" for the winter duty. The white surface of the foam slabs
appears as the floor of the van in this photo. Table II lists the capabili-
ties of the instruments in the emissions-monitoring mobile laboratory along
with the methods used for analyses. Although the chemiluminescent NO analyzer
is equipped with a NOp to NO converter allowing it to measure NO (NO + NOgK
no attempt has been made thus far to measure the NO- component in the field,
89
-------
because of the requirement for a condensation-free (heated) sample transport
line. Assembly of a heated line system is planned, and the N02 concentration
will be checked in the field to confirm previous laboratory readings.
The CO and C02 analyzers use the method of absorption of infrared energy,
and they require well-stabilized cell operating conditions. To minimize de-
stabilizing interruptions in electrical power to these instruments, the mobile
laboratory has two 120 V ac power systems. The primary system consists of two
passive, external power supply circuits with provisions for connection to the
dwelling's 120 V ac power circuit. This, of course, provides only for the
stationary mode of operation. To keep the infrared cells operating contin-
uously, a 12 V dc to 120 V ac power inverter provides a secondary source of
electrical power for the in-transit mode. The result of having this primary-
secondary electrical system has been reliable performance from the CO and C02
analyzers through 10 months of varying ambient conditions and 8000 miles of
travel duty (i.e., vibration).
The flue gas sampling train consists of several elements that condition
and transport the gas from the basement furnace installation to the instru-
ments in the van, which is located typically about 25 m (80 ft) away. The
sample is drawn by diaphragm-type suction pumps located within each instru-
mentation module; the total flow rate required is approximately 1.0 x 10
2
m /s (13.0 scfh). The sample is extracted from the flue by a stainless steel
probe and drawn through a short length of T.F.E. (Teflon) tubing to an ice-
cooled condensibles and particulates trap, where nearly all of the condensi-
bles and particulates are removed by the cooling effect and glass-wool filter
element. A coil of stainless steel tubing is available for attachment to the
probe for sampling of gases from the combustion zone. This acts as an air-
cooled heat exchanger to lower the combustion gas temperature to a tolerable
level for the Teflon tubing. The removal of condensibles at this point is
important to reduce the likelihood of plugging by freezing of the water when
the sample is transported outdoors to the van through a 30-m (100-ft) length
of Teflon tubing. The tubing connects to an external fitting on the vehicle,
and thereafter the gas sample is carried internally by stainless steel tubing
to a second glass-wool particulates filter. The sample path then splits and
90
-------
the portion destined for the infrared analyzers is dried further by passage
through a desiccant (CaSCL) column before being admitted to the instruments.
The second gas sample path leads directly to the NO/NO and the UHC analyzers,
J\
avoiding any chemical desiccants that may influence the analyses. Calibration
gases are available to each instrument through 4-way gas valves located just
ahead of sample flowmeters that lead into each instrument.
Thermal Efficiency Data Loggers
The commercial thermal efficiency rating method generally used in the
industry today is the steady-state operation, flue-gas loss method. This
method uses only the measurement of flue gas COp concentration and net tem-
perature to determine air/fuel ratio (i.e., relative heat input), and unused
heat energy output, from which the ratio of heat-out to heat-in is derived.
This method is a useful indicator but does not account for many of the losses
associated with furnaces in actual operation. The combination of the cyclical
nature of operation of fired-furnace heating systems and variables contributed
by the loose control tolerances of the subsystems produces complex heat loss
characteristics. It was concluded, however, that evaluation of the actual
operating thermal efficiency would provide valuable new data, not only to this
study of balancing emissions with efficiency, but also to attract the solely
efficiency-minded sector into areas of environmental concern.
Because of the large number of variables affecting the operating cycle
(e.g., weather and control inconsistencies), a large number of cycles would
have to be measured to provide a statistically valid data base. To accomplish
this task of evaluating installed operating efficiency, a programmable data
logger system was assembled to continuously record firing cycle efficiency and
timing parameters. Supplementing this system were time recording data loggers
attached to the five other furnaces to accurately monitor furnace utilization.
The ultimate objective of this dual path scheme was to establish overall
heating season thermal efficiency values for all six furnaces by correlating
efficiency to cycle timing and applying the correlation to cycle timing data
acquired over the entire season.
91
-------
Programmable Data Logger--
The programmable data logger system was assembled around a Hewlett-
Packard System 3051A9 which consists of a 10-channel scanner unit, a digital
multimeter, and a programmable controller. Figure 5 shows that attached to
this system are two channels of ac voltage sensing, five channels of tempera-
ture measurement (Chrome!-Alumel thermocouples and a thermistor),, and one
channel of differential pressure information. Channel 1 senses burner power
to determine the boundary of the firing cycle and the duration of heat input.
Channel 2 senses warm-air fan power, which is only used as a signal for mea-
surement of heat output parameters. Channel 3 is return air temperature
required for both temperature gain and air flow rate calculations. Channel 4
is a series matrix of nine thermocouples to measure output air temperature.
The sum of the nine thermocouples is read in and divided by nine for an aver-
age temperature to reduce any bias if a nonuniform temperature profile exists.
Channel 5 monitors maximum flue-gas temperature during the burner firing and
is used in conjunction with Channel 89 burner inlet air temperature for cal-
culation of a net flue-gas temperature. The outputs from the above four
thermocouple circuits are connected to "floating" reference junctions, which
are allowed to vary with ambient temperature but are stabilized in a thermally
insulated enclosure. The temperature in the reference junction enclosure is
monitored by Channel 73 a thermistor device, and the appropriate temperature
correction is applied to all the thermocouple readings. Channel 6 is con-
nected to a very low differential pressure transducer, 0 to 138 Pa (0 to 0.5
in. WG), which measures the pressure differential across a laminar flow
element in the return air ducting. This, in conjunction with the return air
temperature (Channel 3), enables the weight flow rate of heated air to be
calculated.
The monitoring program continuously scans all eight channels and counts
the number of scans to establish a time base. Upon sensing burner power-on,
an on-time count is initiated (heat-in), and inlet air and flue-gas tempera-
tures are monitored. Whenever warm-air fan power is sensed, the "heat-out"
parameters of inlet temperature and pressure drop, and outlet temperatures are
measured. Summations of these measurements are accumulated in temporary
storage until the end of the cycle, which is marked by the start of the next
92
-------
firing. At that time, they are normalized by the number of readings to estab-
lish cycle-averaged values, which are then used in the calculation of total
heat extracted from the furnace by the warm-air flow (heat-out). The values
of on-time, off-time, maximum flue-gas temperature, and heat-out are stored in
interim storage. When 20 firings are accumulated, the data are then trans-
ferred to magnetic tape storage. Also, if the on-time is greater than 2.8
hours or the off-time is greater than 27.8 hours, the transfer instruction is
executed to secure a significant time value into storage.
A total of 4300 firings can be stored on one tape cartridge. This
allows well over a month of typical furnace operation to be recorded. A data
reduction program and the data on tape can then be loaded and processed by the
data logger's computerized controller. The controller also has an automatic
start capability that enables it to reload itself and restart the system
should a power outage occur.
Cycle Time Data Loggers--
To supplement the data acquired by the above system, furnace utilization
data were recorded on the remaining five furnaces. To do this, specifications
were drawn up for a recording clock system that would: (1) have digital
output with 1-second time resolution, (2) monitor a 120 V ac circuit, record-
ing the moment of voltage-on and voltage-off, and distinguishing between the
two, (3) provide 48-hour standby power to maintain correct time through power
outages, (4) record a minimum of 8000 data points, preferably on magnetic
cassette tapes, and (5) be compatible with a tape reader/computer input de-
vice. The system procured is an Instrumentation Technology Corporation Mod-
el %76i a derivative of an earlier model, modified to the above specifica-
tions.
93
-------
RESULTS OF THE 1977-78 HEATING SEASON FIELD TEST
TEST OBJECTIVES
The scientific objectives of the field test were to maintain close-to-
laboratory levels of pollutant reduction and demonstrate an ability to achieve
improvement in fuel consumption by careful application of the low-emissions
design criteria. Consistency of performance and maintainability of the fur-
naces were also of interest.
The goals set on the pollutant emission concentrations were derived from
consideration of both laboratory results (References 4 and 5) and a field
survey (Reference 6). The emission concentration limits are expressed on a
cycle-averaged basis for a fixed operating cycle of 4 minutes on, 8 minutes
off. The NO emissions goal was set at 0.65 g/kg of fuel burned., a reduction
of approximately 65% below the average established by the field survey. The
limits for CO and UHC emissions were set at 1.00 and 0.10 g/kg, respectively,
allowing no compromise of these pollutants for achieving low NO. Smoke emis-
sion's at or below No. 1 smoke spot reading on the Bacharach smoke scale were
deemed acceptable.
Greater than 10 percentage points improvement in gross thermal efficiency
over the average of the furnaces encountered in the field was another goal set
for the integrated furnace system. Therefore, limitation for near unity
stoichiometric ratio (S.R.) operating condition, S.R. = 1.20, was also imposed
to aid in achieving an improvement in fuel utilization.
CLIMATIC CONDITIONS
Heather records from the Albany County Airport and the Blue Hills Obser-
vatory (Milton, Massachusetts) show totals of 7055 degree-day and 6725 degree-
day, respectively, for the field test period. The test period was the second
94
-------
of two unusually cold winter seasons in the Northeastern and Midwestern
United States. The test area encountered two severe winter storms in the
months of January and February 1978. The second storm paralyzed the Boston
area for 7 days, with more snowfall recorded in that area in a 24-hour period
than that of the entire winter of 1972-73. Temperature extremes in the
Boston area are buffered somewhat by the ocean mass (maritime climate effect),
however, resulting in winter seasons warmer than those experienced in the
Albany area (continental climate).
The above-reported degree-day values are based on the generally accepted
65°F baseline, from which any daily mean temperature below this value is
subtracted and presented as degree-days. The heating service subcontractor in
the Albany area, however, has found that use of a 70°F baseline value applied
to the mean temperatures measured at the airport results in a better repre-
sentation of heating oil consumption of their clientele. This is primarily
because of the relatively low location of the Albany County airport in the
Hudson River Valley (i.e., warmer temperatures) compared with the location of
the bulk of their customers. We have found this to hold true for the three
Albany test locations and have adhered to the use of this method for the
Albany data. All degree-day values reported hereafter in this paper for the
Albany sites are based on the 70°F reference.
PERIODIC MONITORING
The field test monitoring schedule was set up on a monthly basis. The
initial furnace checkout visit occurred early in October 1977, with eight
subsequent monthly visits ending in June 1978. These visits primarily in-
volved the measurement of pollutant emissions. Many of the visits required
the movement of the programmable data logger from one furnace system to an-
other; but, generally, the efficiency-measuring systems required only the
retrieval of magnetic tape data cassettes.
Pollutant Emissions Results
The results of the monthly pollutant emissions monitoring visits are
presented in Tables III and IV for the Boston and Albany areas, respectively.
95
-------
The data tables list the pollutant emissions goals, monthly measurements, and
season averages. Although reliable heating service was maintained through the
test period, difficulties in measurement of stoichiometric ratio were encoun-
tered. Inspiration of air into the negative pressure flue-pipe system upstream
of the gas sampling point resulted in false readings of excess air levels.
Although the flue-pipe systems were initially installed properly, thermal
cycling of the systems led to changes in fit tolerances that affected the pipe
joints. In one case, the galvanized piping warped significantly on the back-
side of the joint to the exhaust manifold (i.e., highest pipe temperature),
causing approximately a 10% error in the determination of burner excess air
level. Since the pollutant emission concentrations are normalized to the
amount of fuel burned and not the amount of air, however, the effect of the
air leaks were of only minor consequence to the emissions data. Subsequently,
to monitor the integrity of the flue-pipe system, gas samples were occasion-
ally taken from the firebox area to establish the operating stoichiometric
ratio of the burner. Before this condition was recognized as a false stoi-
chiometric ratio reading, however, burners were on occasion adjusted incor-
rectly to conditions below S.R = 1.20. This resulted in high carbonaceous
pollutant levels or noisy (combustion roughness) operation.
Boston Area Installations--
Woodbine Site—Most of the unanticipated variables in the field test were
encountered by the three Boston area installations, but these were not asso-
ciated with the maritime climate effect. The Woodbine site encountered:
(1) carbonization of the nozzle (see Table III), (2) flue-pipe air leaks,
(3) dirty fuel supply system, and (4) unusually high fuel-bound nitrogen. The
suspected carbonization of the nozzle in January led to the replacement of the
nozzle and retuning of the burner because of variations in nozzle flow char-
acteristics. Although not recognized at that time, the flue-pipe air leakage
may have already been significant, causing the burner to be incorrectly ad-
justed below its design stoichiometric ratio operating point. This possi-
bility is reinforced by the high smoke readings during that period. There
were also complaints from the hosts of noisy operation, requiring readjustment
of the burner in February to a supposedly "above design specification" air
setting. Fuel samples taken during this period showed significant amounts of
96
-------
bottom contaminants, which were thought to be the primary cause of the high
smoke emissions. 'Further evaluation of the unusual characteristics, however,
led to a strong suspicion of air. leaking into the flue. The internal gas
sampling technique was tried on the subsequent visit, and it revealed a sig-
nificant discrepancy between the reading taken at the flue pipe (S.R. = 1.31)
and that taken above the firebox (S.R. = 1.24). Another internal sample was
taken at a point higher above the firebox, at the entrance to the heat ex-
changer, and this also showed a higher excess air condition. This led to the
belief that the leak was internal, probably at the union of the cast-iron
combustor and the carbon-steel heat exchanger. Post-test disassembly of the
unit, however, revealed that the leak was caused by warping of the flue pipe
at a very inaccessible area, on the backside of the connection to the furnace.
The negative-pressure, low-velocity system apparently allowed the cooler
inspirated air to propagate down into the heat exchanger to bias the internal
reading at the entrance to the heat exchanger. The season-averaged stoichi-
ometric ratio for the Woodbine unit, even including the suspect January and
February readings, is 1.23, very close to design specification. The CO and
UHC averages also satisfied the respective goals.
The season-averaged NO emissions for the Woodbine site is very close to
the test goal, biased slightly high by three unusually high readings taken in
Marchs April, and May. This was attributed to a fuel-bound nitrogen level of
200 ppm, more than twice that of other oil samples analyzed. The source of
this high nitrogen fuel oil stock was not traced, since there are no refinery
guidelines on nitrogen composition and locating the source would only be
academic. The smoke emissions field test criterion is almost fulfilled with
the season average, biased by the two months of "mistuned" operation. The
flue-gas temperature reflects the effect of the soot buildup with a rise in
net temperature starting at the time of the initial mistuning.
Richard Site—The Richard site also experienced air leaks in the flue-pipe
system. By January, a trend of increasing stoichiometric ratio conditions was
noted, and at that time was assumed to be caused by degradation of oil nozzle
performance. The oil nozzle was changed; however, smooth operation could not
be attained at design stoichiometric ratio conditions, and the unit set at an
97
-------
apparently high S.R. = 1.27 (flue-gas sample) condition. The stoichiometric
ratio varied on the two subsequent visits, and on the third subsequent visit
(April 1978), the internal sampling technique was used, confirming the exist-
ence of air leaks biasing the stoichiometric ratio readings. It was found
that the actual operating conditions in the firebox were at design specifica-
tions, in spite of the high excess air levels measured in the flue pipe. The
season average is probably more on the order of S.R. ~ 1.22 than the S.R. =
1.26 reported in Table III.
The CO emissions measured at the Richard site were acceptable, with the
exception of two measurements in the early months of the heating season. The
CO profiles through the 4-minute firing cycle started with a significant
"spike," followed by a slow rate of decay, resulting in high cycle-averaged
values. During the first month of operation, the oil bypass return line
plugged, resulting in a back pressure that forced oil past the shaft seals
into the burner housing. In investigating the problem, the serviceman par-
tially disassembled the burner, which required readjustment of the air setting
on the November visit. With the flue-pipe air leak developing, the burner may
have been mistuned, resulting in the high CO emissions and traces of smoke.
In January, the oil nozzle was changed, calling for another readjustment of
the air setting, which later was shown to be very close to design specifica-
tions. The CO emissions were then under control and remained so through the
remainder of the test period. The NO emissions field test goal was satisfied
throughout the heating season. The goals for UHC and smoke emissions were
also achieved, with consistent performance below the stated limits.
Pond Site—The Pond installation encountered only one problem that affected
emissions performance - the accumulation of carbon on the nozzle. The problem
was first noted at this site with a trend of increasing stoichiometric ratio
(i.e., decreasing oil flow rate) requiring a nozzle change after only 2 months
of service. It was first believed that the nozzle may have been "off specifi-
cations," but when another nozzle change was required on the following visit,
a closer examination was made of the nozzle and burner assemblies. It was
then determined that the electrodes were positioned too c'ose to the nozzle,
and repositioning was effected on this and all other test units. Thereafter,
98
-------
with the exception of one anomalously high reading in March, the stoichi-
ometric ratio remained near design conditions. The limits on CO, NO, and UHC
emissions were never exceeded at any time during the field test. The smoke
emissions exceeded the test limit only once during the period of degraded
nozzle performance, but the average for the season is well within specifi-
cations. All pollutant emissions goals were achieved at the Pond installa-
tion.
Albany Area Installations—
Graven Site—The three integrated furnace field test installations in the
Albany, New York, area performed relatively trouble free through the entire
test period. The Graven site showed a slight rise in stoichiometric ratio
(see Table IV); and in February, when No. 1 Bacharach smoke reading was
measured, carbonization of the nozzle was suspected. Removal of the nozzle
showed only minor surface blackening, however, and the smoke reading persisted
with a new nozzle installed. Finding the nozzle not to be the cause, the
burner was adjusted to its as-found condition of S.R. = 1.28, in keeping with
the minimum tampering field-test guideline. The isolated smoke trace recorded
in February was probably induced by a fuel delivery the previous day, agitat-
ing bottom contaminants into suspension in the fuel oil. The cause of the
drift in stoichiometric ratio was not identified, but the condition recovered
to near design point at the end of the heating season. Pollutant emissions
remained stable and controlled throughout the heating season at this installa-
tion.
Stovepipe and Fissette Sites—The Stovepipe installation did encounter a
nozzle carbonization condition, which was corrected by nozzle replacement and
readjustment of the spark electrode. The Fissette site showed no indication
of nozzle problems and completed the heating season without adjustment. The
integrated furnace test units at both these installations showed a peculiar CO
emissions characteristic, the cause of which has not yet been indentified.
During the warmer periods of early fall and late spring, the units show an
unusually high start spike on the CO emissions profile, resulting in the
cycle-averaged value being above specifications. Some effect on UHC emissions
are also noted during this period. This characteristic is suppressed during
99
-------
the colder midwinter period, and the CO emissions goal is satisfied. Since
the emissions measurements are taken on a fixed firing cycle, the temperature
of the furnace components at start are relatively equal and not greatly influ-
enced by changes in ambient temperature. The temperature of the oil trans-
ported into the furnace varies with basement temperature, however, and it is
suspected that this may have an effect on the pump delivery characteristics at
startup. The duration of the phenomenon is very short, making it difficult to
isolate its cause. Because of the much improved performance during the colder
months, the season-averaged CO emission levels for both installations are
right at the limit of 1.0 g/kg. The NO, UHC, and smoke emissions for both
sites are all well under the season-averaged emissions limits, thus achieving
all the field-test emissions goals.
Thermal Efficiency Performance
Thermal efficiency of residential oil-fired furnaces is often rated by a
steady-state operation, flue-gas heat loss method (Reference 7). This method
is simple and a useful indicator, but it does not include many significant
operational losses (e.g., cabinet or standby). For purposes of general com-
parisons, the steady-state efficiency of the integrated furnaces was measured
by the flue-gas loss method and found to be between 83 and 84%, very close to
the 85% maximum achievable efficiency for noncondensing flue furnace systems.
These values are gross efficiency values based on the higher heating value of
No. 2 fuel oil, and all efficiency values reported in this paper are computed
on this basis.
Actual installed efficiencies for oil-fired, residential furnaces are
substantially lower than the steady-state efficiencies, primarily because of
the cyclical nature of operation. For meaningful comparisons of performance,
the installed efficiency was determined in the field test for each integrated
furnace. Very little comprehensive data exists on evaluation of the installed
efficiency of residential heating systems, so the field test comparisons were
made against the fuel utilization histories of the host dwellings for the
previous two seasons. Detailed cycle-averaged efficiency versus cycle timing
data were recorded for each test furnace for part of the test period, while
100
-------
cycle timing data were recorded through the entire heating season. From
these, the relationship of overall fuel consumption can be related to cycle-
averaged efficiency, and estimates of efficiency of the original furnace
installations can be made. This will be accomplished before presentation of
this paper by correlating cycle-averaged efficiency to timing parameters,
applying the correlation to the cycle timing data for the entire season to
arrive at an average efficiency for the season, and adjusting this value by
the ratio of present and past fuel utilization to arrive at an estimate of
prior operating efficiency.
The programmable data logger system was used to record cycle-averaged
thermal efficiency and cycle timing parameters. To characterize all six test
furnaces, the system monitored each furnace for a minimum of 1 month, record-
ing more than 500 firing cycles on each furnace. The results of the data
recorded are summarized in Table V, showing the average efficiency of the
furnace during the period monitored. The thermal efficiencies reported here
are based on actual heat delivered to the distribution system, which accounts
for all furnace losses (e.g., flue gas, standby, and cabinet) exclusive of
dwelling-related losses. The burner-on time-weighted averaging method is
applied here, instead of a simple summation of cycle efficiencies, in order to
eliminate erroneous upward biasing of the average by abnormally terminated
firing cycles. The time-weighted method reduces to simple summations of heat-
out and heat-in over the monitering period.
The average efficiency values presented in Table V for all six integrated
furnace units are above 70%, with a mean of 74.4% and a standard deviation of
±3.8%. Typically, laboratory experimental data will produce a "scatter" on
the order of ±5%, so the ±3.8% deviation is acceptable. The 80.6% performance
achieved at the Woodbine site degraded approximately 2% by soot buildup during
the subsequent mistuned operation period.
Although identical in construction, differences in efficiency character-
istics are noted between field test units. Distinct differences are noted in
the efficiency trend below 20% furnace utilization. "Dropoff" in performance
varies for each furnace at low utilization. Part of the function of the inlet
101
-------
air draft flap is to inhibit cooldown of the furnace by restricting free
convection through the furnace. The path from the firebox to the flue is not
mechanically closed, however, and restricted circulation could exist in some
installations and not others, resulting in the differences in standby loss
characteristics. These different characteristics have impeded the resolution
of an overall correlation of integrated furnace efficiency versus utilization.
Although not fully defined at the time of writing, simpler evaluations of the
data indicate an overall performance of the integrated furnaces to be approx-
imately 73.5% thermal efficiency over the heating season.
The only direct comparison to the existing furnace systems is in the
comparison of previous fuel utilization histories to those recorded with the
integrated furnaces during the field test. The occupants were instructed to
maintain thermostat temperature settings and dwelling use patterns as close as
possible to the prior two seasons, leaving weather as the only variable. To
correct for the differences in weather, degree-day information generally used
in the heating industry to estimate fuel consumption was recorded along with
the amount of fuel consumed. These data are presented in Table VI for both
the original furnace and the integrated furnace service periods. The amounts
of fuel consumed in each period are normalized by the respective weather
factors, and these are the values used for comparisons. The changes in fuel
utilization brought about by the integrated furnaces are significant, showing
reductions in all cases, ranging from 10 to 30% and averaging 18.5%. The wide
band of net change values is caused primarily by the wide variety of original
furnaces being compared. Considering the 73.5% season-averaged thermal effi-
ciency of the integrated furnaces and the resultant 18.5% improvement in fuel
utilization, the mean season-averaged thermal efficiency of the original
furnaces is calculated to be approximately 60% using a method described in
Reference 5. This fits within the 55 to 75% range calculated by Janssen and
Bonne (Reference 8) and the 60 to 65% range suggested by the data obtained by
Turner (Reference 9). The net improvement in season-averaged gross thermal
efficiency of the integrated furnaces is 13.5 percentage points, exceeding the
field test efficiency goal.
102
-------
CONCLUSIONS
The six integrated furnace field test units successfully completed the
1977-78 heating season with no system failures. No major component needed to
be replaced, and only minor shakedown problems were encountered. The season-
averaged goal of a 65% reduction in NO emissions was achieved with CO, UHC,
and smoke controlled to the established limits.
The compatibility of emissions reduction with efficient fuel utilization
was demonstrated. Overall improvements in fuel consumption per degree-day of
10 to 30% (18.5% average) by the six integrated furnace systems were calcu-
lated at the end of the 1977-78 field test, indicating high payback potential
in careful application of low-emissions technology.
103
-------
REFERENCES
1. Combs, L. P., and A. S. Okuda. Design Optimization and Field Verifica-
tion of an Integrated Residential Furnace. In: EPA-600/7-77-073a,
Proceedings of Second Stationary Source Combustion Symposium, New Orleans,
Louisiana, 1977- pp. 85-121.
2. Dickerson, R. A., and A. S. Okuda. Design of an Optimum Distillate Oil
Burner for Control of Pollutant Emissions" EPA-650/2-74-047, Final Report
on Contract 68-02-0017, Rockwell International, Rocketdyne Division,
Canoga Park, California, 1974. 251 pp.
3. Combs, L. P., and A. S. Okuda. Residential Oil Furnace System Optimiza-
tion - Phase I. EPA-600/2-76-038, Phase Report on Contract 68-02-1819,
Rocketdyne, Canoga Park, California, 1976. 210 pp.
4. Combs, L. P., and A. S. Okuda. Residential Oil Furnace System Optimiza-
tion - Phase II. EPA-600/2-77-028, Phase Report on Contract 68-02-1819,
Rocketdyne, Canoga Park, California, 1977. 122 pp.
5. Combs, L. P., and A. S. Okuda. Design Criteria for Reducing Pollutant
Emissions and Fuel Consumption by Residential Oil-Fueled Combustors.
Paper 76-WA/Fu-10, presented at the 97th ASME Winter Annual Meeting, New
York, New York, 1976.
6. Barrett, R. E., Miller, S. E., and D. W. Locklin. Field Investigation of
Emissions from Combustion Equipment for Space Heating, EPA-R2-73-084a
(also API Publ. 4180), Environmental Protection Agency, Research Trinagle
Park, North Carolina, 1973. 206 pp.
7. American National Standard Performance Requirements for Oil-Powered
Central Furnaces, ANSI Z91.1 1972, American National Standards Institute,
Inc., New York, New York, 1972. 24 pp.
8. Janssen, J. E., and U. Bonne. Improvement of Seasonal Efficiency of
Residential Heating Systems. Paper 76-WA/Fu-7, presented at the 97th
ASME Winter Annual Meeting, New York, New York, 1976. 5 pp.
9. Turner, D. W., et al. Efficiency Factors for Domestic Oil Heating Units.
In: Addendum to the Proceedings, Conference on Improving Efficiency in
HVAC Equipment and Components for Residential and Small Commercial Build-
ings, Purdue University, LaFayette, Indiana, 1974. pp. A18-A22:
104
-------
o
en
SEALED COMBUSTION AIR SYSTEM
STANDBY DRAFT CONTROL
COMBUSTION AIR FIL
QUIET PULSE FREE STATOR
OPTIMUM BURNER HEAD
AIR COOLED FINNED FIREBOX
Figure 1. Low-Emission Integrated Furnace Components
-------
Figure 2. Integrated Furnace Finned Firebox and
Heat Exchanger Assembly and Optimum Burner
106
-------
8486-3
Figure 3. Rockwell International Pollutant Emissions Monitoring Laboratory
-------
Figure 4. Flue Gas Analysis Instrument Module No. 1 Installed in the Mobile Laboratory
-------
FLUE
GAS
DATA LOGGER
HEWLETT-PACKARD
SYSTEM 3051A
THERMISTOR
(CONAXTH13-SS12)
CHROMEL-ALUMEL
THERMOCOUPLES
SENSING ONLY
FORtBURN ANDtFAN J
TRANSDUCER
(SETRA MODEL 239)
OUTLET DUCT
REFERENCE
JUNCTION
BOX
INTEGRATED
FURNACE
SYSTEM
OIL FLOWMETER
(CUMU-
LATIVE)
LAMINAR
FLOW
ELEMENT
8486-5
Figure 5. Schematic of Automatic Field Test Furnace
Efficiency Data Acquisition System
109
-------
TABLE I
DESCRIPTION OF THE HOST DWELLINGS SELECTED FOR FIELD TEST
OF THE INTEGRATED FURNACE
Test Site
Woodbine Street
Newton, MA
Richard Road
Needham, MA
Pond Avenue
Brook! ine, MA
Graven Drive
Delmar, NY
Stovepipe Road
Clarksville, NY
Fissette Drive
Feura Bush, NY
Type of
Archi-
tecture
Wood Frame
Raised Ranch
Colonial
Duplex
Ranch
Wood Frame
Age
(Years)
80
21
32
20
14
25
Levels
of
Living
Space
3
1-1/2
2
1
1
1
Type of Wall
Construction
Wood
Wood
Wood
Brick/Wood
Sheet Rock
Sheet Rock
Age of
Furnace
(Years)
5
21
3-1/2
20
14
5
Number of
Occupants
2
2
3
1
4
5
Average
K-Factor
(D-D/gal.)*
3.52
7.84
8.04
10.92
10.49
11.48
*Degree-day per gallon
-------
TABLE II
FLUE GAS EMISSIONS MOBILE LABORATORY INSTRUMENTATION
Species
Range
(Resolution)
Instrument
Manufacturer and Model
Measurement Method
CO
C02
NO/NOx*
02
UHC
Smoke
Temperature
0 to 1500 ppm (2 ppm)
0 to 15% (0.1%)
0 to 1000 ppm (1 ppm)
0 to 25% (0.2%)
0 to 3000 ppm (2 ppm)
0 to 9 Bacharach Scale
0 to 500°F (2°F)
MSA LIRA 300
MSA LIRA 300
TECO 10A
Beckman
MSA
Bacharach
Infrared
Infrared
Chemiluminescent
Polarographic
\\2 Flame lonization
Hand Pump
Mercury Thermometer
*Special heated sampling train required for N0£ measurement
-------
TABLE III
1977-78 INTEGRATED FURNACE FIELD TEST POLLUTANT EMISSIONS SUMMARY
FOR THE BOSTON AREA INSTALLATIONS
Field Test
Location
Test Goals
Woodbine St.
Newton, MA
Unit 5
Richard Road
Needham, MA
Unit 4
Pond Avenue
Brook! ine, MA
Unit 1
Stoich.
Ratio
1.20
1.23
1.23
1.21
1.22*
1.27+
(1.24)
(1.22)
(1.22)
(1.24)
1.23
1.23
1.22t
1.26+
1.27*+
1.45+
1.24+
1.19**
1.23
1.22
1.26
1.22
1.30
1.21*
1.20*
1.24
1.31
1.18
1.23
1.19
1.23
CO
(g/kg)
£1.00
0.65
0.39
0.80
0.40
0.59
0.88
0.87
0.86
0.84
0.70
0.45
5.60
2.97
0.68
0.54
0.68
0.72
0.81
0.81
1.47
0.32
0.78
0.59
0.55
0.47
0.46
0.60
0.59
0.60
0.55
NO
(g/kg)
10.65
0.558
0.679
0.498
0.589
0.579
0.845
0.876
0.952
0.619
0.688
0.690
0.350
0.447
0.507
0.517
0,470
0.585
0.610
0.647
0.536
0.583
0.570
0.532
0.565
0.580
0.590
0.601
0.630
0.606
0.584
UHC
(g/kg)
5D.100
0.065
0.028
0.027
0.060
0.072
0.090
0.103
0.070
0.065
0.064
0.065
0.050
0.048
0.019
0.033
0.036
0.060
0.028
0.043
0.042
0.065
0.069
0.041
0.020
0.038
0.043
0.038
0.019
0.018
0.039
Smoke
Bach-
arach
11. 0
0.5
0.7
0.8
2.5
4.5
0.2
0.4
0
0
1.1
0.3
0.6
0.4
0.5
0
0
0
0
0
0.2
0.5
2.0
0
0
0.5
0.1
0
0.1
0
0.4
Net TFG
(°cj
—
134
132
135
167
173
178
192
174
163
161
144
—
140
147
156
142
153
138
146-
158
174
169
169
169
186
170
164
140
167
Date
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
* New 0.75-7QO-A nozzle installed
** Flue pipe joints sealed
+ Air leak suspected
( ) Combustion gas analysis
112
-------
TABLE IV
1977-78 INTEGRATED FURNACE FIELD TEST POLLUTANT EMISSIONS SUMMARY
FOR THE ALBANY AREA INSTALLATIONS
Field Test
Location
Test Goals
Graven Road
Del mar, NY
Unit 2
Stovepipe Road
Clarksville
Unit 6
Fissette Drive
Feura Bush, NY
Unit 3
Stoich.
Ratio
1.20
1.22
1.22
1.24
1.31
1.28*
1.25
1.29
1.27
1.24
1.26
1.21
1.21
1.31
1.23*
1.20
1.23
1.21
1.21
1.18
1.22
1.22
1.19
1.20
1.18
1.19
1.24
1.20
1.19
1.19
1.20
CO
(g/kg)
<1.00
0.41
0.52
0.49
0.52
0.68
0.56
0.78
0.67
0.53
0.57
1.12
2.23
1.22
0.49
0.52
0.65
0.72
0.75
1.25
0.99
2.24
1.27
0.83
0.55
0.76
0.65
0.76
0.91
1.46
1.05
NO
(g/kg)
<0.650
0.590
0.586
0.494
0.522
0.527
0.529
0.585
0.641
0.676
0.572
0.466
0.310
0.420
0.540
0.476
0.613
0.583
0.535
0.636
0.509
0.552
0.525
0.493
0.521
0.582
0.493
0.529
0.485
0.637
0.535
UHC
(g/kg)
<0.100
0.028
0.082
0.042
0.050
0.039
0.057
0.036
0.034
0.037
0.045
0.018
0.110
0.040
0.019
0.027
0.046
0.024
0.018
0.018
0.036
0.064
0.400
0.063
0.036
0.050
0.043
0.050
0.045
0.061
0.090
Smoke
Bach-
arach
<1.0
0
0
0
0
1.0
0
0
0
0
0.1
0
0
0.4
0
0.1
0
0
0
0
0.1
0
0
0
0
0
0
0
0.1
0
0
Net TFG
(°C)
—
159
160
153
160
181
148
166
152
147
158
137
150
146
168
176
161
158
—
161
157
126
152
132
138
148
149
166
147
134
144
Date
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
Oct 1977
Nov 1977
Dec 1977
Jan 1978
Feb 1978
Mar 1978
Apr 1978
May 1978
June 1978
Average
*New 0.75-70°-A nozzle installed
113
-------
TABLE V
TABULATION OF INTEGRATED FURNACE CYCLE-AVERAGED THERMAL EFFICIENCY,
UTILIZATION, AND TEST PERIOD WEATHER DATA
Furnace
Test Site
Woodbine Street
MA (Unit 5)
Richard Road
MA (Unit 4)
Pond Avenue
MA (Unit 1)
Graven Road
NY (Unit 2)
Stovepipe Road
NY (Unit 6)
Fissette Drive
NY (Unit 3)
Furnace
Efficiency
(nt,%)
80.6
77.1
73.0
70.5
73.5
73.2
70.4
Furnace
Usage
(tRj%)
24
48
32
34
29
14
13
5
Average
Degree-
Day/Day
14
34
37
43
52*
39*
24*
10*
Test Period
Oct-Nov 1977
Dec 1977
Dec 1977 to
Jan 1978
Jan 1978
Feb-Mar 1978
Mar-Apr 1978
Apr-May 1978
May-June 1978
*Based on 70 F reference
where:
= - 10° • 4on
st
100
tn =
on
St
"ft
on
off
100
woil
2t
on
SC-
out 100
-------
TABLE VI
EFFECT OF INTEGRATED FURNACE ON FUEL OIL CONSUMPTION
Test Site
Woodbine Street
Newton, MA
Unit 5
Richard Road
Needham, MA
Unit 4
Pond Avenue
Brookline, MA
Unit 1
Gravon Drive
Delmar, NY
Unit 2
Stovepipe Road
Clarksville, NY
Unit 6
Fissette Drive
Feura Bush, NY
Unit 3
Previous 2 Winters
Fuel
(gal.)
3543
1618
1578
1532
1593
1457
Weather
(D-D)*
12489
12679
12679
16722
16722
16722
Gal. per
D-D
0.2837
0.1276
0.1244
0.0916
0.0953
0.0871
Integrated Furnace
1977-78 Field Test
Fuel
(gal.)
1455
710
722
512
563
614
Weather
(D-D)*
6108
6693
6478
8171
8058
7760
Gal. per
D-D
0.2383
0.1061
0.1114
0.0626
0.0699
0.0791
Overall Average
Net
Change
(%)
-16.0
-16.9
-10.4
-31.7
-26.6
-9.5
-18.5
*Degree-Day, °F-Day (Reference temperature: MA = 65°F, NY = 70°F)
-------
SESSION II
UTILITY AND LARGE INDUSTRIAL BOILERS
DAVID G. LACHAPELLE
SESSION CHAIRMAN
117
-------
STATUS OF NO CONTROL IMPLEMENTATION
A
FOR UTILITY BOILERS
By:
L. R. Waterland, K. J. Lira, and
R. J. Schreiber
Acurex Corporation
Energy and Environmental Division
Mountain View, California 94042
119
-------
ABSTRACT
Utility boilers currently represent the largest contributors to total sta-
tionary source NO emissions in the U.S. and their contribution is projected to
X
remain significant through the end of the century. Thus control of NO emis-
X
sions from these sources has and will continue to receive significant attention
in past and future NO air quality regulation strategies. Currently, modifying
X
the combustion process conditions is the most effective and widely used tech-
nique for achieving moderate (20 to 60 percent) reduction in utility boiler
emissions. Favored techniques have included staged combustion firing, flue gas
recirculation, and new burner and furnace designs. This paper discusses past
experience with these techniques in terms of individual and combined control
effectiveness and operational problems encountered, and presents costs of
implementing these controls in both retrofit application and new unit designs.
Application to coal, oil, and gas fired utility boilers are treated. Finally,
the status of application of these methods in existing and planned field instal-
lations is reviewed.
120
-------
ACKNOWLEDGEMENT
The work presented in this paper was performed as part of the NO Control
X
Technology Environmental Assessment program under Contract 68-02-2160 to the
U.S. Environmental Protection Agency, Industrial Environmental Research Labora-
tory, Combustion Research Branch. The support and assistance of Dr. J. S. Bowen
and Messrs. R. E. Hall, and D. G. Lachapelle are most gratefully acknowledged.
The authors would also like to thank the following individuals for gra-
ciously supplying background and support information: J. Barsin and E. Campo-
benedetto of the Babcock and Wilcox Company; J. Vatsky of the Foster Wheeler
Energy Corporation; R. Sadowski of the Riley Stoker Corporation, W. Barr,
F. Strehlitz, and E. Marble of the Pacific Gas and Electric Company; D. Norton,
formerly of Southern California Edison, R. Meinzer of the San Diego Gas and
Electric Company, and W. Pepper of the Los Angeles Department of Water and Power,
121
-------
SECTION 1
INTRODUCTION
Utility boilers represent the largest contributors to total stationary
source NO emissions in the U.S. In fact, as Figure 1 shows, utility boilers
x
were the origin of 48.3% of all stationary NO emissions for the year 1974 (Ref-
X
erence 1). Moreover, their contribution to total NO emissions is projected to
X
remain significant through the end of this century. As shown in Table I, under
a high energy growth projection, and assuming no additional utility control
application beyond that required to meet 1971 New Service Performance Standards
(NSPS), total NO emissions more than double by the year 2000, while the utility
X
contribution more than quadruples. Even under a low growth scenario, and assum-
ing aggressive promulgation of additional, more stringent, NSPS for coal fired
utility boilers (258 and 215 ng/J in 1978, 172 ng/J in 1981, 120 ng/J in 1985,
and 86 ng/J in 1988), total nationwide NO emissions increase by 25% in the year
x
2000 and the utility contribution increases by about 10%. Thus, it should be
clear why utility boilers have received the greatest attention in past NO
X
regulatory strategies, and are expected to be emphasized in future plans, both
nationwide and regional, to attain and maintain NO ambient air quality standards.
X
Currently, modifying the combustion process conditions is the most effec-
tive and widely used technique for achieving moderate (20 to 60 percent) re-
ductions in utility boiler NO emissions. Techniques presently in the later
X
stages of development and application include combustion staging by means of
overfire air addition (OFA) or removing burners from service (BOOS), flue gas
recirculation (FGR) (for oil and gas fired boilers), and new burner and furnace
designs (NBD). Coupled with all these control approachet is the use of low
excess air firing (LEA). However, LEA is not considered a control method here,
but instead a widely used operational adjustment.
122
-------
This paper discusses past experience with each of these techniques in
terms of control effectiveness and operational problems encountered, presents
costs of implementing these controls in both retrofit and new unit applications,
and reviews the status of application of these methods in existing and planned
field installations.
123
-------
SECTION 2
PROCESS IMPACTS OF NOV CONTROL
A
This section summarizes the major impacts of combustion modification con-
trols on boiler operation. The discussion is organized by fuel type.
COAL FIRED BOILERS
The effects of low NO operation on coal fired boilers are summarized in
x
Table II. The table describes experience obtained through field test programs
on the several units listed. The most commonly applied low NO technique for
X
coal fired boilers is staged combustion through either overfire air (OFA), or
burners out of service (BOOS). Average NO reductions of 30 to 50 percent
X
(controlled emissions of 215 to 300 ng/J) can be expected with either technique.
Flue gas recirculation has been tested,'but found to be a relatively ineffective
control giving only about 15 percent NO reduction (Reference 6). More recently
X
new, low NO burner designs (NBD) have been installed on some units and found to
X
be at least as effective as OFA. The combination of OFA with NBD has resulted
in 40 to 60 percent NO reduction (controlled emissions of 170 to 260 ng/J).
X
The major concerns regarding low NO operation on coal fired boilers have
X
been the effects on boiler efficiency, load capacity, water wall tube corrosion
and slagging, carbon loss, heat absorption profile, and convective section tube
and steam temperatures.
In most past experience with staged combustion, optimal excess air levels
have been comparable to those used under baseline conditions. In these cases
the efficiency of the boiler would remain unaffected if unburned carbon losses
do not increase appreciably. However, in some cases when, due to nonuniform
fuel/air distribution or other causes, the excess air requirement increases
under staged firing, a significant decrease in efficiency may occur. From
124
-------
Table II, it is seen that efficiency decreases up to 1 percent have been ex-
perienced. It is also seen that the same boiler (Widows Creek No. 5) tested at
a different time with BOOS showed an average increase in efficiency of 1 percent.
Many new boilers now come factory equipped with OFA ports. Older boilers
can be retrofitted with OFA ports, or can operate with minimal hardware changes
under BOOS firing. BOOS firing is normally accomplished by shutting off one or
more pulverizers supplying the upper burner levels. If the other pulverizers
cannot handle the extra fuel to maintain the total fuel flow constant, boiler
derating will be required. From Table II, it is seen that boiler derating of
10 to 25 percent is not uncommon with BOOS firing.
The possibility of increased corrosion has been a major cause for concern
with staged operation. Furnaces fired with certain Eastern U.S. bituminous
coals with high sulfur contents may be especially susceptible to corrosion
attack under reducing atmospheres. Local reducing atmosphere pockets may exist
under staged combustion operation even when burner stoichiometry is slightly over
100 percent. The problem may be further aggravated by slagging as slag gener-
ally fuses at lower temperatures under reducing conditions. The sulfur in the
molten slag may then readily attack tube walls. Still, as noted in Table II,
experience has generally been that no significant acceleration in corrosion rates
occurs under staged firing. More recent experience has substantiated this con-
clusion (Reference 9). Nevertheless, the issue cannot be considered resolved
until definitive results from long term tests with measurements on actual water
wall tubes are available. Insofar as slagging is concerned, short term tests
performed to date indicate no significant increase in slagging or fouling of
tubes under staged combustion.
Increased carbon loss in flyash may occur with staged firing if complete
burnout of the carbon particles does not occur in the furnace. High carbon loss
will result in decreased boiler efficiency and may also cause electrostatic pre-
cipitator (ESP) operating problems. From Table II, it is seen that increases
in carbon loss vary over a wide range and can be as high as 70 to 130 percent
in some cases. However, increased carbon loss is not perceived as one of the
major problems associated with staged combustion. If the carbon content in fly-
ash increases to levels where it threatens to impair the operation of dust col-
lection systems, the unburned carbon can usually be easily controlled by
125
-------
increasing the overall excess air level in the furnace. Although this will tend
to increase stack heat losses, the decrease in boiler efficiency will be par-
tially compensated for by reduced unburned carbon losses.
Extension of the combustion region to higher elevations in the furnace may
result in potential problems with excessive steam and tube temperatures. How-
ever, among the numerous short term combustion staging tests conducted, no such
problems have been reported. In some tests where furnace and convective section
tube temperatures were measured directly, no significant increase was found.
Changes in heat absorption profiles were also found to be minor, thus indicating
no need for addition or removal of heat transfer surfaces. Superheater attemper-
ator spray flowrates tripled in one case under OFA operation, but in all cases
were well within spray flow capacities of the units. Reheater attemperator
spray flowrates did not show any increase due to staged operation, thus cycle
efficiencies were not affected.
Many new wall fired coal boilers are being fitted with low NO burners (NBD).
X
These burners are designed to reduce NO levels either alone or in some cases in
combination with OFA ports. Using the new burner designs has the advantage of
eliminating or decreasing the need for reducing or near reducing conditions near
furnace walls. Corrosion problems associated with reducing atmospheres should
thus not arise with this system. Although low NO burner flames can be expected
X
to be less turbulent and hence longer than flames from normal burners, the com-
bustion zone will probably not extend any farther up the furnace than with staged
combustion. Potential changes in heat absorption profile and excessive steam
and tube temperatures are, therefore, less likely to occur.
As fuel and airflows are controlled more closely in NBD equipped systems,
nonuniform distribution of fuel/air ratios leading to excessive CO generation, or
high excess air requirements should be eliminated. Boiler efficiencies should,
therefore, not be affected. However, Table II shows that the efficiency of one
boiler decreased slightly when retrofitted with low NO burners. The decrease
x
in efficiency was mainly due to the large increase in unburned carbon loss. How-
ever, such problems noted in retrofit applications can be avoided in units
specifically designed with the low NO burners included. Corrosion rates inferred
X
from tests with corrosion coupons showed no significant increase with the new
burners. Some BOOS tests were also carried out on the NBD equipped boiler. A
126
-------
substantial decrease in NO emissions resulted, below those already achieved
X
with the new burners alone. However, the boiler was derated by up to 30 per-
cent. Other, potential problems noted above as being associated with staged
combustion could also arise with this type of firing.
It should be emphasized that the effects of NO control, in many cases,
X
will be critically dependent on boiler operating conditions. Still, with proper
design of retrofit systems and adequate maintenance programs, low NO operation
should not result in a substantial increase in operational problems over normal
boiler operation. Moreover, when NO controls are designed into new units,
X
potential problems can be anticipated and largely corrected.
OIL FIRED BOILERS
The effects of low NO operation on oil fired boilers are summarized in
Table III. As the table shows the most commonly used low NO techniques for
oil fired boilers are staged combustion and flue gas recirculation (FGR), both
employed with low excess air firing. Other techniques which have been tested
are water injection (WI) and reduced air preheat (RAP). However, these have
found little application due to attendant efficiency losses.
Staged combustion has been applied through the use of overfire air ports
(OFA) and by removing burners from service (BOOS). Typical NO reductions using
X
OFA are 20 to 30 percent (controlled emissions of 150 to 172 ng/J), while BOOS
has been slightly more effective giving 20 to 40 percent reductions (controlled
levels of 129 to 172 ng/J). Flue gas recirculation also typically gives 20 to
30 percent NO reductions. The combination of staged combustion with FGR has
X
been most effective, resulting in 30 to 60 percent reductions (controlled emis-
sions of 100 to 172 ng/J).
The major concerns regarding low NO operation on oil fired boilers are
X
effects on boiler efficiency, load capacity, vibration and flame instability,
and steam and tube temperatures.
Staged combustion operation generally increases the minimum excess air re-
quirements of the boiler, which may result in a loss in boiler efficiency. In
extreme cases when the boiler is operating close to the limits of its fan ca-
pacity, boiler derating may be required. Derates of as much as 15 percent have
been required in some cases due to the lack of capability to meet the increased
127
-------
airflow requirements at full load. In addition, under BOOS firing the fuel flow
to the active burners must be increased if load is to remain constant. In many
cases, it has been necessary to enlarge the burner tips in order to accommodate
these increased flows.
Other potential problems attendant with applying staged combustion in oil
fired boilers have concerned flame instabilities, boiler vibrations, and ex-
cessive convective section tube temperatures. However, in past experience, none
of these problems has been significant. Staged operation does usually result
in hazy flames and obscure flame zones. Thus new flame scanners and detectors
are often required in retrofit applications. In additions because staged combus-
tion produces an extended flame zone, flame, carryover to the convective section
may occasionally occur. However, in one case where intermittant flame carryover
occurred, no excessive tube temperatures were recorded.
Similarly, there are a number of potential problems which can occur in
retrofit FGR applications. The most common problems, such as FGR fan and duct
vibrations, can usually be avoided by good design. Other problems such as flame
instability, which can lead to furnace vibrations, are caused by the increased
gas velocity at the burner throats. Modifications to the burner geometry and
design such as enlarging the throat, altering the burner tips, or adding diffuser
plates or flame retainers may then be required.
Another potential problem associated with FGR is high tube and steam tem-
peratures in the convective section. The increased mass velocities which occur
with FGR cause the convective heat transfer coefficient to rise. This, coupled
with reduced furnace heat absorption, can give rise to high convective section
temperatures leading to tube failures, exceeding attemperator spray flow limits,
or loss in cycle efficiency due to excessive reheat steam attemperation. In-
creased mass flowrates in the furnace may also cause furnace pressures to in-
crease beyond safe limits.
The combination of staged combustion and FGR is very effective in reducing
N0x emissions. However, the problems associated with each technique are also
combined. Tube and steam temperature problems in the upper furnace are particu-
larly exacerbated, as both combustion staging and FGR tend to increase upper
furnace temperatures and convective section heat transfer rates. In addition,
128
-------
boiler efficiencies usually decline slightly with combined staged combustion
and FGR firing due to higher EA requirements and greater fan power consumption.
As with coal fired boilers, before low NO techniques are instituted on an
oil fired boiler, it is important to assure that it is in good operating con-
dition. Uniform burner air and fuel flows are essential for optimal NO control.
X
Retrofit NO control systems must be designed and installed properly to minimize
X
potential adverse effects. Many of the problems experienced in the past can
now be avoided because of hindsight and experience. Thus, retrofit systems can
now be designed and installed with care to avoid any potential adverse effects.
GAS FIRED BOILERS
The effects of low NO operation on gas fired boilers are summarized in
X
Table IV. The most commonly applied techniques, as with oil fired boilers, are
staged combustion through the use of OFA or BOOS with FGR. Typical NO reduc-
X
tion under either OFA, BOOS, or FGR is 30 to 60 percent (controlled emissions
of 86 to 172 ng/J). The combination of staged combustion and FGR has resulted
in 50 to 75 percent reductions (controlled levels of 65 to 129 ng/J).
The effects on low NO firing on gas fired boilers are very similar to those
X
for oil fired boilers. Usually, there is no distinction between oil and gas
fired boilers as they are designed to switch from one fuel to the other according
to availability. Since boiler design details, NO control methods, and the
X
effects of low NO operation are similar for gas and oil fired units, most of
X
the above discussion of applicable NO control measures to oil fired boilers and
X
potential problems resulting applies. Some effects specific to gas fired boilers
alone are treated briefly in the following.
NO emissions oftentimes are difficult to control after switching from oil
X
to gas firing. Residual oil firing tends to foul the furnace due to the oil ash
content. Thus, NO control measures which have been tested on a clean furnace
x
with gas may be found inadequate after oil firing due to the changed furnace
conditions.
Boilers fired with gas usually have higher gas temperatures at the furnace
outlet than when fired with oil. The upper furnace and convective section inlet
surfaces are thus subject to higher temperatures with gas firing. These tem-
peratures may increase further under staged firing or FGR. Upper furnace and
129
-------
convective section tube failures and excessive steam temperatures are therefore
more likely to occur with staged firing and FGR applied to gas fired boilers.
The situation may be aggravated further if switching from gas fuel occurs after
an oil burn, as fouling will further reduce furnace absorption and, hence, in-
crease gas temperatures. Excessive tube temperatures will usually require de-
rating of the system. However, problems with gas firing are not commonly en-
countered at present due to the scarcity of natural gas fuel.
130
-------
SECTION 3
COSTS OF NOV CONTROL
A
Estimated costs of applying the NO controls discussed in Section 2 are
X
summarized in the following. Retrofit control costs are presented first, fol-
lowed by a discussion of the incremental costs for incorporating NO controls in
new units.
RETROFIT CONTROL COSTS
Detailed retrofit control costs were calculated using an annualized revenue
requirement formulation similar to that described in References 17 and 18. Repre-
sentative control costs were prepared for the boiler/control combinations shown
in Table V. For each combination, preliminary engineering designs of the NO
X
controls treated were prepared. This design work provided an estimate of the
hardware and installation requirements for applying the retrofit controls. Up
to date vendor quotes were then obtained to serve as input to the costing algo-
rithm. All cost input data and assumptions employed are discussed in more detail
in Reference 19.
TABLE V. BOILER/CONTROL COMBINATIONS COSTED
Boiler Fuel
Tangential/Coal
Opposed Wall /Coal
Opposed Wall /Coal
Opposed Wall /Coal
Single Wall /Oil , Gas
Single Wall/Oil, Gas
MCRa
(MW)
225
540
540
540
90
90
NOX
Control
OFA
OFA
NBD
BOOS
BOOS
OFA & FGR
aMaximum continuous rating in MW of electrical output
131
-------
It was assumed that the units being retrofitted were relatively new, say
5 to 10 years old, with at least 25 years of service remaining. As Table
V shows, overfire air and new burner designs were selected as the retrofit
control methods for coal firing. Burners out of service was not necessarily
recommended for coal fired units, but was included to demonstrate the high cost
of derating a unit, as is often the case for pulverized coal units. Burners
out of service, and flue gas recirculation through the windbox combined with
overfire air were selected as the retrofit control methods for the single wall
oil and gas fired unit.
Estimated costs for applying the treated NO controls, in 1977 dollars, are
X.
summarized in Table VI. The table shows initial capital investment, annualized
capital investment with other indirect costs, annualized direct costs, and total
annualized cost to control. The table indicates that the preferred combustion
modification generally costs between $0.50 to 0.70/kW-yr to install and operate.
One major exception to this is the use of BOOS firing on coal fired units if
derating is required due to insufficient mill capacity. In this instance the
high cost of BOOS implementation reflects the need to purchase makeup power,
and to account for lost capacity through a lost capital charge.
CONTROL COSTS FOR NEW UNITS
Estimating the incremental costs of NO controls for NSPS boilers is in
X
some respects an even more difficult task than costing retrofits. Certain modi-
fications on new units, though effective in reducing NO emissions, were orig-
X
inally incorporated due to operational considerations rather than from a control
viewpoint. For example, the furnace of a typical unit designed to meet 1971
NSPS has been enlarged to reduce slagging potential. But this also reduces NO
X
due to the lowered heat release rate. Thus, since the design change would have
been implemented even without the anticipated NO reduction, the cost of that
x '
design modification should not be attributed to NO control.
x
Babcock & Wilcox has estimated the incremental costs of NO controls on an
x
NSPS coal fired boiler (Reference 20). The two units used in the comparison
were identical except for NO controls on the NSPS unit which included:
X
132
-------
• Replacing the high turbulence, rapid-mixing cell burner with
the limited turbulence dual register (low NO ) burner
X
• Increasing the burner zone by spreading the burners vertically
to include 22 percent more furnace surface
• Metering and controlling the airflow to each row of burners
using a compartmented windbox
To provide these changes for NO control, the price increase was about $1.75 to
X
$2.50/kW (1977 dollars). If these costs are annualized they translate to 0.28
to 0.40 $kW-yr.
In addition, Foster Wheeler has performed a detailed design study aimed at
identifying the incremental costs of NO control inclusive in NSPS units (Refer-
X
ence 21). Foster Wheeler looked at these unit designs with the following results:
Boiler Design Relative Cost
Unit 1: Pre-NSPS base design 100
Unit 2: Enlarged furnace, no 114
active NO control
J\.
Unit 3: NSPS design; enlarged 115.5
furnace, new burner
design, perforated hood,
overfire air, boundary air
Assuming the cost of a pre-NSPS coal fired boiler to be about $100/kW in
1969, or $180/kW in 1977 construction costs (References, 22, 23, 24). the
incremental cost of active NO controls (NBD plus OFA) is $2.78/kW, or about
X
$0.44/kW-yr annualized. The Foster Wheeler estimate which includes both NBD
and OFA, thus agrees quite well with the Babcock & Wilcox estimate, which in-
cludes only NBD and associated equipment.
133
-------
SECTION 4
FIELD APPLICATION OF NO CONTROL
J\.
Given the preceding discussion of the operational effects and costs of NO
control for both retrofit and new unit applications. Tables VII and VIII inventory
the extent of existing and planned implementation of these controls. Table VII
documents the application of retrofit control; Table VIII lists new units, in
operation or planned, with designed-in control capability. Neither table should
be considered complete, however. Table VII was compiled through direct contact
with utility companies known to have participated in various EPA and privately
sponsored field application tests. Table VIII was combiled largely through direct
contact with utility boiler manufacturers, supplemented by published trade publi-
cation surveys. Both tables undoubtedly suffer some omissions. For example,
information on planned control applications for tangential units was incomplete.
However, hopefully the data represented in the tables are sufficiently complete
for illustrative purposes.
Table VII illustrates that the most extensive use of retrofit controls has
occurred on gas and oil fired units in California. Staged combustion through
OFA or BOOS, and/or FGR have been most commonly employed. Over 14,500 MW of gas
and oil fired capacity have come under control through retrofit application;
representing about 6.6 percent of installed gas/oil fired capacity. Retrofit
control application to coal fired units has been insignificant.
Table VIII shows that new unit controls have focused on coal fired instal-
lations, paralleling new unit sales trends (very few new gas or oil fired units
have been sold since 1970). The use of new burner and furiace designs, OFA, and
the combination of NBD and OFA have been the favored techniques. By the end of
1979 over 34,600 MW of coal fired capacity will have come under control for NO ,
i X
almost all through controls incorporated into the units original design. This
134
-------
will represent about 14 percent of the nation's coal fired utility boiler
capacity. In 1980 and beyond this fraction will, of course, increase as all
new units coming on line will be built to conform to the 1971, or the proposed
1978, New Source Performance Standards.
In addition, Table VIII shows that by the end of 1979, about 12,500 MW of
gas and oil fired capacity will have been installed with new unit controls.
Thus a total of 17,000 MW of gas/oil fired capacity will be controlled, repre-
senting about 7.7 percent of installed capacity. No new gas or oil fired units
are currently forecast to come on line after 1981.
135
-------
SECTION 5
SUMMARY AND CONCLUSIONS
Modifying the combustion process conditions is currently the most effective
and widely used method to effect 20 to 60 percent reductions in NO emissions
from utility boilers. Staged combustion through overfire air addition is the
favored technique for retrofit application to coal fired units, with the use of
new burner designs, or new burners in combination with OFA favored for new
units. Combustion staging through OFA or removing burners from service, flue
gas recirculation, or the combination of FGR with staged combustion has been most
extensively used in retrofit applications to oil and gas fired boilers.
Potential problems with the use of these techniques have concerned possible
adverse effects on boiler efficiency, load capacity, water wall tube corrosion
and slagging, fouling, carbon loss, steam temperature, and flame stability and
vibration. However, recent field experience has shown that adverse effects can
be minimized to acceptable levels with proper care in design for retrofit appli-
cation, and largely eliminated in new unit designs.
Annualized NO control costs for retrofit implementation generally fall in
X
the $0.50 to 0.70 per kW-yr. These include costs for installation of new burners
or for OFA in coal fired boilers and for using BOOS in oil and gas fired boilers.
Annualized incremental costs for constructing new coal fired units with prov-
vision for NO^ control fall in the $0.25 to 0.45 per kW-yr range.
By the end of 1979 approximately 7.7 percent of the oil and gas fired
utility boiler capacity, and 14 percent of the coal fired capacity will be con-
trolled using combustion modification. These percentages are projected to in-
crease with time as:
136
-------
• New, controlled, coal fired units begin operation
• Older, uncontrolled, gas/oil and coal fired units
are retired
• More existing units install retrofit control in
response to local regulatory demands.
137
-------
REFERENCES
1. Salvesen, K. G. NO Emission Inventory for Stationary Sources. Report
X
TM-78-210, Acurex Corporation, Mountain View, California, June 1978.
2. Selker, A. P. Program for Reduction of NO from Tangential Coal-Fired
X
Boilers, Phase II and Ha. EPA-650/2-73-005a and 5b, June 1975 and
August 1975.
3. Burrington, R. L., et_ al. Overfire Air Technology for Tangentially Fired
Utility Boilers Burning Western U.S. Coal. EPA-600/7-77-117, October 1977.
4. Crawford, A. R. , &t_ al_. Field Testing: Application of Combustion Modi-
fications to Control NO Emissions from Utility Boilers. EPA-650/2-74-066,
X
June 1974.
5. Crawford, A. R., at auL. The Effect of Combustion Modification on Pollu-
tants and Equipment Performance of Power Generation Equipment. In:
Proceedings of the Stationary Source Combustion Symposium, Volume III.
EPA-600/2-76-152c, June 1976.
6. Thompson, R. E., et_ aJ_. Effectiveness of Gas Recirculation and Staged
Combustion in Reducing NO on a 560 MW Coal-Fired Boiler. EPRI Report
X
No. FP-257, September 1976.
7. Unpublished data supplied by Hollinden, G. A. Tennessee Valley Authority,
Chattanooga, Tennessee, August 1977.
8. Crawford, A. R. , et_ _al_. Field Testing: Application of Combustion Modi-
fication to Power Generating Combustion Sources. In: Proceedings of the
Second Stationary Source Combustion Symposium, Volume II.
EPA-600/7-77-073b, July 1977.
138
-------
9. Bartok, W. , at al^. Combustion Modification for the Control of Air
Pollutant Emissions from Coal-Fired Utility Boilers. ASME Winter Annual
Meeting, ASME 78-WA/APC-7, December 1978.
10. Unpublished data supplied by Meinzer, R. P., Jr. San Diego Gas & Electric
Company, San Diego, December 1977.
11. Barr, W. H., et^ _al. Modifying Large Boilers to Reduce Nitric Oxide Emis-
sions. Chem. Eng. Prog. J., 73(7):59-68, July 1977.
12. Norton, D. M. , jit. _al. Status of Oil-Fired NO Control Technology. In:
X
Proceedings of the NO Control Technology Seminar. EPRI SR-39, February
X
1976.
13. Norton, D. M., et_ a^. Modifications to Ormond Beach Steam Generators for
NO Compliance. ASME Winter Annual Meeting, ASME 75-WA/PWR-9, November
X.
1975.
14. Unpublished data supplied by Campobenedetto, E. J. Babcock and Wilcox
Co., Barberton, Ohio, and Barr, W. H. and Marble, E., Pacific Gas &
Electric Co., San Francisco, February 1978.
15. Unpublished data supplied by Meinzer, R. P., Jr., San Diego Gas & Electric,
San Diego, California, October 1977-
16. Barr, W. H., and James, D. E. Nitric Oxide Control — A Program of Signi-
ficant Accomplishments. ASME Winter Annual Meeting, ASME 72-WA/PWR-13,
November 1972.
17. McGlamery, G. G., et^ al. Detailed Cost Estimates for Advanced Effluent
Desulfurization Processes. EPA-600/2-75-006, January 1975.
18. Waitzman, D. A., et^ _al. Evaluation of Fixed-Bed, Low-Btu Coal Gasifica-
tion Systems for Retrofitting Power Plants. EPRI Report No. 203-1,
February 1975.
19. Lim, K. J., et al. Environmental Assessment of Utility Boiler Combustion
Modification NO Controls. Report TR-78-105, Acurex Corporation,
X
Mountain View, California, April 1978.
20. Personal communication from Campobenedetto, E. J. Babcock & Wilcox
Company, Barberton, Ohio, November 1978.
139
-------
21. Vatsky, J. Effectiveness of NO Emission Controls on Utility Steam
X
Generators. Foster Wheeler Energy Corporation, Livingston, New Jersey,
Report to Acurex Corporation, November 1978.
22. Electrical World., 188(9);39, November 1973.
23. Chemical Engineering., 85(11):189, May 1978.
24. Electrical World., 184(10):43, November 1975.
25. NO Control Review., 3(3):6, Summary 1978.
X
140
-------
Noncombustion 1.
Warm Air Furnaces 2.7%
Industrial Process Combustion3.6
Gas Turbines 3.7%
Fugitive 4.1%
Incineration 0.3°n
Reciprocati ng
1C Engines
15.455
Utility Boilers
Packaged Boilers
Warm Air Furnaces
Gas Turbines
Reciprocating 1C Engines
Industrial Process Combustion
Noncombustion
Incineration
Fugitive
Total
6g
5,810
2,446
320
440
1,856
426
193
40
498
12,029
1,000 Tons
6,391
2,691
352
484
2,042
469
212
44
548
13,233
Percent
Total
48.3
20.3
2.7
3.7
15.4
3.6
1.6
0.3
4.1
100
Figure 1. Distribution of stationary anthropogenic NOX
emissions for the year 1974 (Reference 1).
141
-------
TABLE I. NATIONWIDE NOX EMISSIONS PROJECTIONS (Gg)
ro
Total Nationwide
NOX Emissions
Utility Sector
NOV Emissions
A
High Growth
No Post-1971 NSPS
1974
20,931
5,813
1985
25,857
10,833
2000
43,852
24,066
Low Growth
(National Energy Plan)
Stringent Utility NSPS
1974
20,931
5,813
1985
22,251
7,191
2000
26,084
6,298
-------
TABLE II. EFFECT OF LOW NO OPERATION ON COAL-FIRED BOILERS
Boiler
Tangential
Barry No. 2
Co 1 iimb i a
No. 1
Huntington
Canyon No. 2
Barry No. 4
Navajo No. 2
Comanche No. 1
Opposed Wall
Harllee Branch
No. 3
Four Corners
No. 4
Hatfield No. 3
Low NOX
Technique
BOOS
OFA
OFA
OFA
BOOS
BOOS, OFA
OFA
BOOS
BOOS
BOOS
Efficiency
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
0.6% average
decrease
0.6% increase
0.3% decrease
Corrosion3
Measured 75%
increase, but
within normal
range
Measured 70%
increase, but
within normal
range
No change
Measured 25%
decrease, but
within normal
range
No significant
change
No significant
change
No significant
change
Slight Increase
No significant
change
—
Load
Capacity
20% derate
Unaffected
Unaffected
Unaffected
20% or more
derate
Unaffected
Unaffected
Up to 17%
derate
Up to 25%
derate
10% derate
Carbon Loss
in Flyash
Slight increase
Slight increase
Slight increase
Slight increase
~50% average
decrease
No change
~30% average
decrease
~ 130% average
increase
-E05S average
decrease
~ 30% average
increase
Other Effects,
Comments
Minor changes in heat
absorption profile
SH attemperation
increased by 70%
Minor changes in heat
absorption profile
SH attemperation
increased over 200%
Minor changes in heat
absorption profile
SH attemperation
increased by 70%
Minor changes in heat
absorption profile
No SH attemperation
required
No slagging or fouling.
No significant
increase in tube
temperatures.
Reference
2
3
3
4
5
5
4
4
6
i
T-833"
a—Denotes not investigated
-------
TABLE II. Concluded
Bo i 1 er
E.G. Gaston
No. 1
Single Wall
Widows Creek
No. 5 (TVA
test)
Widows Creek
No. 5 (Exxon
test)
Widows Creek
No. 6
Mercer Station
No. 1 (wet
bottom)
Crist Station
No. 6
Low NOX
Technique
FGR
NBD, BOOS
BOOS
BOOS
BOOS
Biased
f ~ing
BOOS
Efficiency
0.4* decrease in
boiler effi-
ciency. Some
decrease in cycle
efficiency due to
RH attemperation
0.3% decrease
on average (NBO
baseline)
1% decrease
1% average
increase
Unaffected
Unaffected
0.4* decrease
Corrosion3
No significant
increase
Results of tests
inconclusive
No significant
increase
—
No significant
increase
—
Load
Capacity
Unaffected
Up to 30*
derate
(NBD with
BOOS)
Unaffected
Unaffected
Unaffected
Unaffected
Up to 15*
derate
Carbon Loss
in flyash
-120% average
increase
-130* average
increase (NBO
baseline)
30* increase
30* average
decrease
70* average
increase
80* average
increase
60* increase
Other Effects,
Comments
Stable flames and
uniform combustion.
Increase in RH
attemperation.
No significant
increase in
tube temperatures.
Unit retrofitted
with low NOX
burners. Baseline
and BOOS tests with NBD
compared to baseline
tests on sister
boiler with no NBD.
Reference
6
5
7
8
4,7
8
4
a — Denotes not investigated T-833
-------
TABLE III. EFFECT OF LOW NO OPERATION ON OIL-FIRED BOILERS
Bo i 1 er
Tangential
South Bay No. 4
Pittsburg No. 1
SCE tangential
boilers
Opposed
Ual.l
Moss Landing
Nos. 6 and 7
Ormond Beach
Nos. 1 and 2
SCE BAW Units
Sewaren Station
No. 5
Low NOX
Technique
BOOS
RAP
OFA and
FGR
BOOS and
FGR
OFA and
FGR
BOOS and
FGR
Water
injection
BOOS and
FGR
BOOS
Efficiency3
Decrease in efficiency
compared to LEA due to
increased excess air
requirements
Unaffected due to
special preheater
design
—
Increased excess air
requirements resulting
in decreased efficiency
Increased excess air
requirements resulting
in decreased efficiency
Increased sensible and
latent stack losses
FGR reduced minimum
excess air require-
ments increasing
unit efficiency
Load
Capacity3
—
Slower startups
and load changes
—
10 to 155S derate
due to maxed FD
fan capacity
—
—
—
Vibration and
Flame Instability2
—
FGR fan vibration
problems
—
FGR fan and duct
vibration, furnace
vibration problems.
Associated flame
instability
Flame instability
and associated
furnace vibration
Boiler vibration
problems
—
Steam and Tube
Temperatures3
—
High water wall
tube
temperatures
—
—
—
Other Effects,
Comments
No other adverse effects
reported
Limited tests. NOX
control effectiveness not
demonstrated.
No adverse effects
reported
High furnace pressures.
Increased FGR and forced
draft fan power
consumption
Flame detection problems
due to change in flame
characteristics
Limited tests carried out
with WI at partial loads.
Excess air requirements
increased
Flame detection problems
due to change in flame
characteristics
Tests carried out at
partial loads. No adverse
effects reported.
Particulate loading and
size distribution
unaffected.
Reference
10
-
11
12,13
11,14
12,13
12
8
a— Denotes not investigated T-831
-------
TABLE III. Concluded
Boiler
Single Wall
Encina Nos. 1,
2 and 3
Turbo
South Bay No. 3
Potrero No. 3-1
Low NOX
Technique
BOOS
(2 burners
on air
only)
BOOS
(3 burners
on air
only)
Water
injection
RAP
OFA and
FGR
Efficiency9
Increased unit effi-
ciency. Some adverse
effect on cycle effi-
ciency due to lower
steam temperatures
Increased excess air
requirements resulting
in reduced efficiency
6% decrease at full
load
Reduction in effi-
ciency greater than
that with water
injection
Higher excess air re-
quirements, but addi-
tion of economizer
surface expected to
improve efficiency
Load
Capacity3
5% derate due to
maxed ID fan
capacity
„
__
5% derate due to
excessive tube
temperatures
Vibration and
Flame Instability3
„
In most tests no
flame instability
or blowoff noted
No flame instability
noted even at high
rates of WI
__
Side to side
windbox oxygen
cycling
Steam and Tube
Temperatures3
Decrease in SH & RH
steam temperature
Intermittent flame
carryover to SH
inlet but tube
temperature limits
not exceeded
„
Tube and steam
temperature limits
approached.
Increased SH
tube failures
Other Effects,
Comments
No other adverse effects
reported
No abnormal tube fouling,
corrosion or erosion
noted. Increased
tendency to smoke and
obscure flame zone.
No other adverse
effects reported
Limited tests
Increased tendency to
smoke required higher
minimum excess 03
levels. RH surface
removed to avoid
excessive RH steam
attemperation. Larger
economizer installed
to compensate for RH
surface removal.
Reference
15
10
11
CT>
a—Denotes not investigated
T-831
-------
TABLE IV.
EFFECT OF LOW NO OPERATION ON GAS-FIRED BOILERS
/\
Boiler
Tangential
South Bay No. 1
?
Pittsburg No. 7
Horizontally
Opposed
Moss Landing
Nos. 6 and 7
Pittsburg
Nos 5 and 6
Contra Costa
Nos. 9 and 10
Sinqle Wall
Encina Nos. 1,
2 and 3
Low NOX
Technique
BOOS
OFA and FGR
OFA and FGR
OFA and FGR
OFA and FGR
BOOS
(2 and 3
burners out
of service)
Efficiency3
Slight decrease in
efficiency due to
increased excess air
requirements
0.83! decrease in cycle
efficiency due to RH
steam attemperation
"
Low EA levels were
possible even with
BOOS, resulting in
increased efficiency
Loada
Capacity
253£ derate due to
excessive steam
temperatures
slower load
change response
Load curtailment
to 50« after oil
burns due to SH
tube temperature
limits being
exceeded
"
No derate. Load
pickup response
not affected
Vibration and3
Flame Instability
Fan and duct
vibration problems
Furnace and duct
vibration problems.
Flame instability.
FGR fan and duct
vibrations. Flame
instability problems.
FGR duct vibrations
Some pressure
pulsing at
corners of
firebox
Steam and Tube3
Temperatures
High tube and RH
steam temperatures
RH spray and SH tube
temperature limits
approached after oil
burns upper wall tube
failures
Upper water wall
tube failures
High SH and RH steam
temperatures. SH
tube temperature
limits being
approached.
Some flame carryover
to SH but no
problems with high
tube temperature or
tube wastage
Other Effects, Comments
No other adverse effects
reported
Furnace pressure limit
approached. FGR fan power
requirements increased by
as much as 665!. Problems
associated with switching
to gas after oil burning
could he eliminated only
with complete water washing
of furnace.
Boiler initially restricted
to manual operation due to
problems with flame insta-
bility nn automatic control
Furnace pressure limits
approached after oil firing.
FGR fan preheating required
to r°riuce vibrations on cold
boiler startups.
No other adverse effects
reported
Reference
10
11
11,14,16
11
11
15
T-819
3—Denotes not investigated
-------
TABLE IV. Concluded
Boiler
Turbo
South.Bay No. 3
Potrpro No. 3-1
Low NOX
Techn ique
Water injection
OFA and FGR
Efficiency6
105! decrease at full
load
Installation of larger
economizer expected to
improve efficiency
Load3
Capacity
__
5% derate due to
problems with high
temperatures
Vibration and3
Flame Instability
No flame instability
noted even at high
rates of WI
Side to side
windbox oxygen
eye ring
Steam and Tube3
Temperatures
_.
Tube metal and steam
temperature limits
reached at high
loads
Other Effects, Comments
No other adverse effects
reported
Hardware modifications
included partial RH surface
removal to avoid excessive
RH steam attemperation.
Larger economizer t^en
installed to compensate for
smaller RH surface.
Reference
10
11
-pi
oo
3—Denotes not investigated
T-819
-------
TABLE VI. SUMMARY OF RETROFIT CONTROL COSTS'
Boiler /Fuel Type
Tangential /Coal -Fired
OFA
Opposed Wall/Coal-Fired
OFA
NBD
BOOS
Single Wall/Oil- and Gas-Fired
BOOS
FGR/OFA
Initial
Investment
($/kW)
0.90
0.62
2.03
0.08
0.30
5.71
Annualized Indirect
Operating Cost
($/kW-yr)
0.21
0.16
0.34
5.34
0.05
1.14
Annualized Direct
Operating Cost
($/kW-yr)a
0.32
0.52
0.06
24.78
0.44
1.91
Total to Cost
Control
($/kW-yr)a
0.53
0.69
0.40
30.12
0.49
3.05
a--Based on 7000-hour operating year. Typical costs only.
T-870
-------
TABLE VII. RETROFIT UTILITY BOILER NOX CONTROL INVENTORY'
Utility
Alabama Power Co.
Los Angeles Dept.
of Water & Power
Pacific Gas &
Electri c
San Diego Gas &
Electric
Southern California
Edison
Plant Name
Barry 2
Harbor 1
Harbor 2
Harbor 3
Harbor 4
Harbor 5
Haynes 1
Haynes 2
Haynes 3
Haynes 4
Haynes 5
Haynes 6
Scattergood 1
Scattergood 2
Moss Landing 6
Moss Landing 7
Pittsburg 5
Pittsburg 6
Pittsburg 7
Contra Costa 9
Contra Costa 10
Potrero 3
South Bay 1
South Bay 2
South Bay 3
South Bay 4
Silvergate 1
Silvergate 2
Silvergate 3
Silvergate 4
Silvergate 5
Silvergate 6
Encina 1
Encina 2
Encina 3
Encina 4
Alamitos 1
Alamitos 2
Alamitos 3
Alamitos 4
Unit
Type
T
W
W
W
W
W
W
W
HO
HO
HO
HO
W
W
HO
HO
HO
HO
T
HO
HO
Tu
W
W
Tu
T
W
W
W
W
W
W
W
W
W
W
W
W
T
T
Manu-
facturer
CE
NAv
NAv
NAv
NAv
NAv
CE
CE
BW
BW
BW
BW
CE
CE
BW
BW
BW
BW
CE
BW
BW
RS
BW
BW
RS
CE
BW
BW
BW
BW
BW
BW
BW
BW
BW
BW
NAv
NAv
CE
CE
Rated
Output
MW
130
75
75
90
90
95
240
240
245
245
350
350
180
180
750
750
330
330
735
330
330
300
140
142
198
220
19
19
32
32
64
64
100
102
102
287
175
175
320
320
Fuel
C
G,0
R,0
G,0
G,0
G,0
G,0
G,0
6,0
G,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
6,0
NOX
Control
OFA
BOOS
BOOS
BOOS
BOOS
•BOOS
BOOS
BOOS
BOOS,
OFA
BOOS,
OFA
BOOS,
OFA
BOOS,
OFA
BOOS
BOOS
BOOS,
FGR
BOOS,
F6R
FGR,
OFA
FGR,
OFA
FGR,
OFA
FGR,
OFA
FGR,
OFA
FGR,
OFA
RAP
OFA
OFA,
RAP
OFA
BOOS
BOOS
BOOS
BOOS
BOOS
BOOS
BOOS
BOOS
BOOS
FGR
OFA
OFA
FGR
FGR
Year
Onl ine
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
65
65
66
67
NAv
NAv
71
71
64
64
72
65
65
72
60
62
64
71
43
43
48
48
50
52
54
56
58
73
NAv
NAv
NAv
NAv
Reference 25
150
-------
TABLE VII. (Concluded)
Utility
Southern California
Edison
Southern Electric
Generating Company
Tennessee Valley
Authority
Western Fanners
Electric Coop.
Plant Name
Alamitos 5
Alamitos
El Segundo 1
El Segundo 2
El Segundo 3
El Segundo 4
Etiwanda 1
Etiwanda 2
Etiwanda 3
Etiwanda 4
Huntington Beach 1
Huntington Beach 2
Huntington Beach 3
Huntington Beach 4
Ormond Beach 1
Ortnond Beach 2
Mandalay 1
Mandalay 2
Redondo 5
Redondo 6
Redondo 7
Redondo 8
E. C. Gaston 1
Widows Creek 5
Moorland 3
Unit
Type
HO
HO-
W
W
T
T
T
T
T
T
W
W
HO
HO
HO
HO
W
W
W
W
HO
HO
HO
W
Tu
Manu-
facturer
BW
BW
NAv
CE
CE
CE
CE
CE
CE
CE
NAv
NAv
NAv
NAv
FW
FW
NAv
NAv
NAv
NAv
BW
BW
BW
BW
RS
Rated
Output
•MM
480
480
175
175
335
335
135
135
320
320
215
215
215
225
785
785
215
215
175
175
480
480
270
125
135
G,0
G,0
G,0
G,0
G,0
6,0
G,0
G,0
6,0
G,0
G,0
6,0
6,0
6,0
G,0
6,0
6,0
6,0
6,0
G,0
G,0
6,0
C
C
6
MOX
Control
BOOS,
FGR
B'OOS,
FGR
BOOS,
OFA
BOOS,
OFA
FGR
FGR
BOOS
BOOS
FGR
FGR
BOOS,
OFA
BOOS,
OFA
OFA
OFA
BOOS,
FGR,
OFA
BOOS,
FGR,
OFA
BOOS
BOOS
OFA
OFA
BOOS,
FGR
BOOS,
FGR
NBD
BOOS
BOOS
Year
Onl ine
65
66
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
NAv
71
73
NAv
NAv
NAv
NAv
66
67
60
50
75
Unit Type
HO Horizontally opposed
T Tangentially fired
Tu Turbo-fired
W Single wal 1 fired
NOX Controls
BOOS Burners out of service
FGR Flue gas recirculation
NBD New burner or furnace design
OFA - Overfire air
RAP Reduced air preheat
Manufacturer
CE Combustion Engineering, Inc.
BW Babcock & Wilcox Co.
FW Foster Wheeler Energy Corp.
RS Riley Stoker Corp.
Fuel
C Coal
G Gas
0 Oil
Others
NAv Not available
151
-------
TABLE VIII. NEW UTILITY BOILER NOX CONTROL INVENTORY*
Utility
Alabama Electric
Cooperative
Alabama Power Co.
Allegheny Power
System
American Electric
Power
Arizona Electric
Power Coop.
Arizona Public
Service Company
Arkansas Power &
Light Company
Austin, City of
(Texas)
Baltimore Gas
& Electric
Basin Electric
Big Rivers Elec-
tric Corporation
Black Hills Power &
Light, Pacific Power
& Light
Brazos Electric
Power Corporation
Bryan, City of
Cajun Electric
Power Coop.
Carol i na Power &
Light Company
Central Illinois
Light Company
Central Illinois
Public Service Co.
Central Louisiana
Electric Company
Central Maine
Power Company
Central Power &
Light Company
Cincinnati Gas
& Electric
Plant Name
Tombigbee 2
Tombigbee 3
Miller 1
Miller 2
Miller 3
Miller 4
Pleasants 1
Pleasants 2
Breed 2
Apache 2
Apache 3
Cholla 4
White Bluff 2
Decker Creek 2
Brandon Shores 1
Brandon Shores 2
Laramie River 1
Laramie River 2
Reid 2
Wyodak
San Miguel 1
Dansby 1
Big Cajun 1
Big Cajun 2
Mayo 1
Mayo 2
Mayo 3
Mayo 4
Roxboro 4A
Roxboro 4B
Duck Creek Site 1
Meredosia 4
Rodemacher 2
Cousins 4
Coleto Creek 1
Miami Fort 3
East Bend 2
Unit
Type
Tu
Tu
HO
HO
HO
HO
HO
HO
HO
Tu
Tu
T
T
HO
HO
HO
HO
HO
NAv
HO
HO
W
Tu
Tu
HO
HO
HO
HO
Tu
Tu
W
W
HO
W
T
HO
HO
Manu-
facturer
RS
RS
BW
BW
PW
BW
'-W
FW
BW
RS
RS
CE
CE
BW
BW
BW
BW
BW
BW
BW
BW
BW
RS
RS
FW
FW
FW
FW
RS
RS
RS
FW
FW
FW
CE
BW
BW
Rated
Output
MW
255
255
650
660
660
660
600
600
1300
203
203
130
800
425
600
600
580
580
240
350
448
105
540
540
375
375
375
375
360
360
378
200
550
600
550
552
650
Fuel
C
C
C
C
C
C
C
C
C
C
C
C
C
G,0
0
0
C
C
C
C
C
G,0
C
C
C
C
C
C
C
C
C
0
C,0,G
0
C
C
C
NOX
Control
NBD
NBD
NBD
NBD
NBD
NBD
OFA
6FA
NBD
NBD, OFA
NBD, OFA
NBD
OFA
OFA.FGR
NBD, OFA
NBD, OFA
NBD
NBD
NBD
NBD
NBD
OFA
NBD
NBD
OFA
OFA
OFA
OFA
NBD
NBD
FOR
FGR,OFA
FGR.OFA
OFA
OFA
NBD, OFA
NBD
Year
Onl ine
78
79
78
79
80
80
78
79
79
77
78
80
81
77
78
78
79
80
79
78
79
78
79
80
81
81
84
84
75
76
76
76
82
78
79
77
79
Reference 25
152
-------
TABLE VIII. (Continued)
Utility
Cliffs Electric
Service Company
Colorado Public
Service Company
Colorado Springs
D.P.U.
Colorado Ute Elec-
tric Assn. Int.
Columbus & Southern
Ohio Electric Co.
Commonwealth Edison
Consolidated Edison
Consumers Power
Dairyland Power
Company
Dallas Power &
Light Company
Dayton Power &
Light
Pelmarva Power
& Light
Detroit Edison
Detroit Public
Lighting Comm.
East Kentucky
Rural Coop.
Florida Power &
Light Company
Fremont, City of
(Nebraska)
Gainesville, City
of (Florida)
Georgia Power
Company
Gulf Power
Gulf States Util-
ity Company
Plant Name
Presque Isle 7
Presque Isle 8
Presque Isle 9
Comanche 1
Comanche 2
NAv
New Fountain 1
Craig 1
Craig 2
Poston 6
Collins 1
Collins 2
Collins 3
Collins 4
Collins 5
Astoria 6
Campbell 3
Alma 6
Martin Lake 2
Kill en 1
Kill en 2
Indian River 4
Belle River 1
Belle River 2
Greenwood 1
Mistersky 3
Charleston Bottoms
Manatee 1
Manatee 2
Martin 1
Martin 2
Wright 8
NAv
Scherer 2
Ellis 1
Ellis 2
Nelson 5
Sabine 5
Unit
Type
Tu
Tu
Tu
T
HO
HO
W
HO
HO
T
HO
HO
HO
HO
HO
W
HO
Tu
T
HO
HO
Tu
HO
HO
W
Tu
HO
W
W
W
W
W
Tu
T
HO
HO
T
T
Manu-
facturer
RS
RS
RS
CE
BW
FW
BW
BW
BW
CE
Bl<>
BW
BW
BW
BW
FW
FW
RS
CE
BW
BW
RS
BW
BW
FW
RS
BW
FW
FW
FW
FW
BW
RS
CE
FW
FW
CE
CF
Rated
Output
MW
85
85
85
350
350
500
220
470
470
413
550
550
550
550
550
800
840
280
750
650
650
400
650
650
800
60
333
800
800
800
800
85
81
818
575
575
520
480
Fuel
C
C
C
C
C
C
C
C
C
C
0
0
0
C,0
C,0
0,6
C
C
L
C
C
C
C
C
0
0
C
0
0
0
0
C
C
C
C
C
C
0
NOX
Control
NBD
NBD
NBD
OFA
OFA
OFA
NBD
' NBD
NBD
OFA
OFA.FGR
OFA,FGR
OFA.FGR
NBD
NBD
FGR,OFA,
NBD
OFA
NBD, OFA
OFA
NBD
NBD
NBD
NBD
NBD
FGR.OFA
NBD
NBD, OFA
FGR,OFA
F6R.OFA
F6R.OFA
FGR.OFA
NBD
NBD
CFA
OFA
OFA
OFA
OFA
Year
Online
78
78
79
NAv
75
80
79
79
78
83
77
76
76
77
78
77
80
77
78
82
80
79
NAv
NAv
79
78
76
76
77
80
81
76
75
82
82
84
86
79
153
-------
TABLE VIII. (Continued)
Utility
Grand River Dam
Authority
Hoosier Energy
Division
Houston Lighting
& Power (Texas)
Illinois Power
Indiana Pub! it
Service
Interstate Power
Iowa Power & Light
Iowa Public Service
Iowa Southern
Utilities Company
Jersey Central
Power & Light
Company
Kansas City Board
of Public Utilities
Kansas City Power
& Light Company
Kentucky Utilities
Lafayette, City of
Lakeland, Florida
Louisville Gas &
Electric
Lower Colorado
River Authority
Mississippi Power
Company
Mississippi Power
& Light
Montana Power
Company
Nebraska Public
Power
New Mexico Public
Service Company
New York State
Electric & Gas
New York State
Power Authority
Plant Name
Unit 1
Unit 1
Unit 2
Parish 5
Parish 6
Cedar Bayou 3
Havana 6
Gibson 3
Gibson 4
Gibson 5
Lansing 4
Council Bluff 3
George Neal 4
Ottumwa 1
Coho 1
Seward 7
Nearman Creek 1
latan 1
La Cygne 2
Ghent
Bon in 3
Plant 3, Unit 2
Mill Creek 3
Mill Creek 4
Fayette 1
Jackson County 2
Andrus 1
Colstrip 3
Col strip 4
NAv
San Juan 1
San Juan 2
San Juan 3
San Juan 4
Lake Cayuga
NAv
Unit
Type
HO
Tu
Tu
HO
HO
HO
HO
HO
HO
HO
Tu
HO
HO
T
T
T
Tu
HO
HO
HO
W
W
HO
HO
T
T
HO
T
T
HO
W
W
HO
HO
HO
HO
Manu-
facturer
FW
RS
RS
BW
BW
FW
BW
FW
FW
FW
RS
BW
FW
CE
CE
CE
RS
BW
BW
FW
BW
BW
BW
BW
CE
CE
BW
CE
CE
BW
FW
FW
BW
BW
BW
FW
Rated
Output
MW
490
490
490
675
675
750
472
660
660
660
260
700
600
674
800
800
235
680
685
511
187
125
450
534
550
500
750
700
700
930
375
350
600
600
930
750
Fuel
C
C
C
6,0,0
6,0,0
0,G
C
C
C
C
C
C
C
C
C
C
C
C
C
C
G,0
0
C
C
C
0
0
C
C
C
C
C
C
C
C
C
NOX
Control
OFA,NBD
NBD.OFA
NBD,OFA
NBD
"NBD
FGR,OFA
NBD
OFA, NBD
OFA,NBD
OFA,NBD
NBD
NBD
OFA
OFA
OFA
OFA
NBD
NBD
NBD
OFA
OFA
FGR
NBD
NBD
OFA
FGR
FGR
OFA
OFA
NBD
OFA, NBD
OFA
NBD
NBD
NBD
OFA
Year
Online
81
81
80
76
76
75
78
82
78
79
77
79
78
81
87
84
78
80
77
83
76
75
78
79
79
79
74
80
81
80
77
74
81
81
83
NAv
154
-------
TABLE VIII. (Continued)
Utility
Niagara Mohawk
Northern Indiana
Public Service
Northern States
Power
Oklahoma Gas- &
Electric Company
Oklahoma PuHic
Service Company
Omaha Public Power
Pennsylvania
Electric
Pennsylvania
Power
Portland General
Electric
Potomac Electric
Power Company
Salt River Project
San Diego Gas &
Electric
South Carolina
Public Service
Authority
South Mississippi
Electric Power
Association
Southern Indiana
Gas & Electric
Southwestern
Electric Power
Springfield, City of
Tallahassee, City
of (Florida)
Tampa Electric
Company (Florida)
Texas Power &
Light Company
Texas Utilities
Public Service
Utah Power and
Light
Plant Name
Oswego 5
Oswego 6
Schaefer 15
Sherburne Co. 3
Sherburne Co. 4
Sooner 1
Northeastern 1
Northeastern 2
Nebraska City 1
Homer City 3
Mansfield 3
Boardman 1
Morgantown 1
Navajo 2
Unit 1
Unit 2
Encina 5
Georgetown 2
Morrow 1
Morrow 2
A. B. Brown
Welsh 1
Flint Creek
Southwest
Hopkins 2
Big Bend 3
Twin Oak 1
Monti cello 3
Forest Grove 1
Naughton 4
Naughton 5
Unit
Type
W
W
HO
HO
HO
T
T
T
HO
HO
HO
HO
T
T
Tu
Tu
W
Tu
Tu
Tu
HO
HO
HO
Tu
W
Tu
T
HO
HO
HO
HO
Manu-
facturer
FW
FW
FW
BW
BW
CE
CE
CE
FW
BW
FW
FW
' CE
CE
RS
RS
NAv
RS
RS
RS
BW
BW
BW
RS
BW
RS
CE
BW
BW
BW
BW
Rated
Output
MW
850
850
500
800
800
510
450
450
575
680
880
500
575
800
380
380
300
280
180
180
250
550
550
202
238
450
750
775
775
450
450
Fuel
0
0
C
c
C
c
c
c
c
c
c
c
c
c
c
c
0
c
c
c
c
c
c
c
G,0
c
L
c
c
c
c
NOX
Control
FGR.OFA
FGR.OFA
OFA
NBD
NBD
' OFA
OFA
OFA
OFA
NBD
OFA
OFA
OFA
OFA
NBD, OFA
NBD, OFA
OFA
NBD, OFA
NBD.OFA
NBD, OFA
NBD
NBD
NBD
NBD
OFA,FGR
NBD,FGR
OFA
NBD
NBD
NBD
NBD
Year
Online
76
79
79
80
81
79
79
79
79
77
80
80
NAv
NAv
78
79
79
77
78
78
NAv
77
78
76
77
76
82
77
80
81
83
155
-------
TABLE VIII. (Concluded)
Utility
Utah Power and
Light
Wisconsin Electric
Power Coop.
Wisconsin Power
and Light
Wisconsin Public
Service Corporation
Plant Name
Huntington
Canyon 2
Pleasant Prairie 1
Columbia 1
Weston 3
Unit
Type
T
Tu
T
T
Manu-
facturer
CE
RS
CE
CE
Rated
Output
MW
400
616
520
322
Fuel
C
C
C
C
NOX
Control
OFA
NBD
OFA
OFA
Year
Online
77
79
77
82
Unit Type
HO Horizontally opposed
T Tangentially fired
Tu - Turbo-fired
W Single wall fired
NOX Controls
BOOS Burners out of service
FGR Flue gas recirculation
NBD New burner or furnace design
OFA Overfire air
RAP Reduced air preheat
Manufacturer
CE - Combustion Engineering, Inc.
BW - Babcock & Wilcox Co.
FW Foster Wheeler Energy Corp.
RS Riley Stoker Corp.
Fuel
C Coal
G Gas
L Lignite
0 - Oil
Others
NAv - Not available
156
-------
FIELD TESTING OF UTILITY BOILERS AND GAS TURBINES FOR EMISSION REDUCTION
By:
E. H. Manny, A. R. Crawford and W. Bartok
Exxon Research and Engineering Company
Government Research Laboratories
Linden, New Jersey 07036
157
-------
ABSTRACT
The U. S. Environmental Protection Agenc3? has sponsored a series of field
test programs by Exxon Research and Engineering Company over the past eight.
years to determine the effect of combustion modification techniques on
combustion related emissions, with emphasis on oxides of nitrogen (NO ). The
2C
combustion modifications found to be effective in earlier programs in gas and
oil fired utility plants were tried ivad alsc found to be effective in the
more complicated pulverized coal fired utility boilers in later programs.
These tests have shown that a combination of staged combustion and low excess
air firing is a cost-effective means of controlling NO emissions in gas, oil,
X
and coal-fired utility boilers. NOX reductions averaging 40 percent have
been achieved in pulverized coal-fired utility boilers in short term tests
without significant adverse side effects on other emissions, boiler
efficiency, boiler slagging and fouling. The long term effect of staged
combustion on furnace tubewall corrosion rates is unresolved, but current
and planned EPA sponsored tests should resolve this question and determine
where corrective measures are needed.
158
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge the constructive participation of
Mr. R. E. Hall, EPA Project Officer and Mr. R. C. Carr of the Electric Power
Research Institute, co-sponsors of this program, in planning the field test
programs and providing coordination with boiler operators and manufacturers.
The assistance and cooperation of the General Electric Company personnel in
helping in selecting gas turbines for testing is also gratefully acknowledged.
The helpful cooperation, participation and advice of Babcock and Wilcox, Com-
bustion Engineering and Foster Wheeler were essential in selecting represen-
tative boilers for field testing and conducting the program. The voluntary
participation of electric utility boiler operators in making their boilers
available is gratefully acknowledged. These boiler operators include the
Southern Electric Generating Company, the Alabama Power Company, the Tennessee
Valley Authority, the Potomac Electric Company, the Salt River-Project,
the Public Service Company of Colorado, the Public Service Electric and Gas
Company (N.J.), the East Kentucky Power Cooperative, the Gulf Power Company,
and the Houston Lighting and Power Company. The authors also express their
appreciation for the extensive coal analyses services provided by Exxon
Research's Coal Analysis Laboratory at Baytown, Texas and to Messrs. A. A.
Ubbens and E. C. Winegartner for their contributions and advice on coal
related matters. The valuable assistance of Messrs. P- S. Natanson, L. W.
Blanken, R. W. Schroeder, W. Petuchovas, J. J. Eggert, M. Tyszkicwz, and Mrs.
M. V. Thompson in performing these field studies is also acknowledged.
159
-------
INTRODUCTION
Since 197Q,. Exxon Research and Engineering Company (ER&E) has been
conducting field studies under EPA sponsorship on the application of
combustion modification techniques to control pollutant emissions from
utility boilers. The emphasis in these studies has been on controlling
NOX emissions without causing adverse side-effects.
Under EPA Contract No-, CPA 70-90 (1), significant reductions of NOX
vere achieved for gas and oil-fired boilers using combustion modification
techniques in field testing of limited duration, without attempting to
optimize the technology, The principal modifications investigated consisted
of minimizing excess air, staged introduction of the combustion airs flue
gas recirculation, varying boiler load, and varying air preheat temperature.
Also, as part of this study, it was possible to achieve significant re-
ductions in NOX emissions for two of the seven coal-fired boilers tested,
through the combination of low excess air with staged firing.
Because of the difficulty of controlling NOX emissions from coal-fired
boilers, in the subsequent EPA-sponsofed ER&E study (Contract No. 68-02-0227)
(2) the emphasis shifted to a more detailed investigation of emission control
for coal-fired utility boilers, again in cooperation with boiler .owner-
operators and manufacturers. These field studies, on twelve coal-fired
units representative of the current design practices of the major U.S.
boiler manufacturers (Babcock and Wilcox, Combustion Engineering, Foster
Wheeler, and Riley Stoker), have produced very promising results. It was
possible to achieve reductions in NOX emissions ranging between about 30% and
50%, without apparent adverse side-effects. In addition to gaseous emissions
measurements, the studies included particulated mass loading and unburned
combustible measurements, accelerated furnace corrosion rate probing, deter-
mination of boiler efficiency, and observations on changes in boiler opera-
bility which particular attention to slagging, fouling, and flame problems.
160
-------
Based on the encouraging results of the above work, the Environmental
Protection Agency and the Electric Power Research Institute decided to
jointly fund the present ER&E field study. The scope of the work was
broadened to determine the effects of combustion modification techniques
on the control of pollutant emissions and on the performance of fossil fuel
fired power generation equipment. In this field program, the combustion
equipment selected for study included coal-fired, mixed-fuel fired and
waste fuel-fired boilers to be studied, in addition to short-term tests
on stationary gas turbine and I.C. engine equipment. The principal
continuing emphasis in this work remained on coal-fired utility boilers.
All but one of the coal-fired boilers tested in this study had dry-bottom
furnaces.
In the field studies detailed in this report, the combustion equip-
ment selected was tested in cooperation with the owner-operators and the
manufacturers. The effectiveness of equipment modifications designed to
provide NOX emission control, such as coal-fired boilers constructed with
overfire air ports, and the use of low NOX emitting improved burner
design were explored. Potential adverse side-effects of combustion
modifications were studied in more detail than previously. Thus in
addition to the gaseous emission measurements (including pollutants,
stable combustion products and unburned combustibles), particulate mass
and size distribution measurements were made under both normal and low
NOX modes of boiler operation.
As in the earlier field studies, the effect of combustion modifi-
cations on boiler efficiency and operability was determined. Special
attention was paid to the demonstration of potential furnace fireside
tube wall corrosion problems that may result from staged firing with
coal, i.e., with parts of the boiler operated under fuel rich, reducing
conditions. For this purpose, 300-hour corrosion probe runs with air-
cooled coupons were made under both low NOX and baseline firing conditions
in the boilers studied. corrosion probe runs of 30S 300 and up to 1000
hours were also made during a special long term corrosion test on one
boiler included in this program to obtain information on the effect of
corrosion with time.
161
-------
FIELD STUDY PLANNING AND PROCEDURES
This section summarizes the boiler selection criteria, cooperative
arrangements made with boiler manufacturers and operators, the test
program strategies used, and the specific test procedures employed.
Test Program Design
The criteria recommended for the selection of coal fired utility
boilers were classified into five groups: (1) boiler design factors,
(2) boiler operating flexibility, (3) boiler measurement and control
capability, (4) utility management's willingness to participate in
research programs, and (5) logistic and scheduling considerations.
Boilers representing the current design practices of the U.S. utility
boiler manufacturers (Babcock and Wilcox, Combustion Engineering, Foster
Wheeler, and Riley Stoker) were selected. Boiler size (150 MW or larger),
type of firing (wall, tangential, turbofurnace and cyclone) furnace
loading, burner configuration (size, number and spacing), draft system
(balanced and pressurized), and furnace bottom design (wet or dry bottom)
were all taken into account.
Earlier ER&E field test programs with coal (1,3) on one tangentially
fired and one wall fired utility boiler have established that the combi-
nation of staged combustion with low excess air could lead potentially
to the largest reductions in NOX emissions. Although well researched
for gas and oil fired installations, there was no report in the literature
of the application of such combustion modifications to coal fired utility
boilers for NOX control. Because of this early indication of significant
NOX reduction potential, subsequent ER&E field programs have emphasized
the above approaches (2,4).
162
-------
As discussed before, the boiler units to be tested were selected on
the basis of criteria established for this purpose. Testing the applica-
bility of combustion modifications, two principal goals were set: (1)
establishing sufficient information by boiler and fuel characteristics so
that the knowledge gained could be applied to other existing installations,
and (2) generalizing the information adequately to help guide the design
of new installations for meeting the New Source Performance Standard (NSPS)
of 301 ng/J (0.7 lb/106 Btu) heat input for NOX emissions.
The test program evolved through several stages — initially, short
term gaseous emission measurements were made to assess the applicability
of combustion modifications for NOX control. This was followed by field
test programs tailored to individual boilers characteristic of current
design practices, where a series of short term (Phase I) gaseous emission
measurements were made. These were followed by a 1-2 day (Phase II)
period under low NOX combustion modification conditions to check for
potential operating problems and to make' any required adjustments. Finally,
300-hour sustained runs were conducted under both low NOX and baseline
boiler operating conditions to allow short-term fireside tube corrosion
tests to be made with corrosion coupons, and also to determine the effect
of combustion modifications on boiler efficiency, operability, particulate
mass loading and size distribution, and S02/S03 concentration in the flue
gas. (For the purposes of this study, low NOX operation denotes modified
firing of the boiler resulting in the lowest NOX levels without apparent
short-term side effects. Baseline operation is at or near full load
conditions corresponding to low NOX operation, but with conventional
firing.)
With a sufficient data base in hand on combustion modifications for
coal fired utility boilers of different design and size characteristics,
additional units were selected for testing based on special features. The
latter included units designed and built by their manufacturers to meet the
NSPS for NOX, and specially modified boilers that provided the flexibility
for testing combustion modifications and new low NOX burner designs,
Statistical principles provided practical guidance in planning the
Phase I test programs', e.g., how many, and which test runs to conduct, as
163
-------
well as the proper sequence in which they should be run. These procedures
allow valid conclusions to be drawn from analysis of data on only a small
fraction of the total possible number of different test runs that could
be made in principle, resulting in increased productivity of the research
work.
Statistical principles were also used extensively in the analysis
of emission data recorded from each test run. For example, least-squares
regression analysis was used to obtain valid estimates of the relationship
between NOX emission levels and operating variables, such as excess air
level. These techniques provide objective, uniform methods of analysis
that can be easily verified, provide the basis for efficient assessment of
outlier data, and limit the risk of drawing false conclusions.
Gaseous Sampling and Analysis
The mobile sampling and analytical system used in this study to
obtain reliable gaseous emission data from field tests was described
elsewhere in detail (1,4). The gaseous species analyzed are NO, N02,
02, C02> CO, S02 and hydrocarbons. The instruments contained in the ER&E
sampling and analytical van were also described in detail (1,4) according
to manufacturer, operating technique employed, and measurement range
capabilities.
A major consideration in obtaining reliable gaseous emission data
is to insure that the sample gas is virtually moisture free. Moisture
in the sample gas can influence the readings obtained by some of the
analytical monitoring insturmentation. For example, moisture is an
unwanted interference with NO, C0£ and CO nondispersive infrared ana-
lyzers. This problem is avoided effectively in the sampling system used
by passing the sample gases through a refrigerated water trap coil.
Recently, permeation-type drying tubes (Permatubes supplied by Perma
Pure, Inc., Oceanport, N.J.) have become available for removing moisture
in gas sampling systems without changing the concentration of other
gaseous species. For this reason, Permatube driers were adapted to the
sampling train as an addition to the refrigerated traps to insure a
maximum drying capability for the gaseous emission sampling system. This
164
-------
system has proven to be effective in obtaining dry sample gases.
In the ER&E sampling-analytical system, gaseous samples are taken
from zones of "equal areas" in the flue gas ducts. At least two probes
are installed in each flue gas duct, or a minimum of four are used when
there is only one large flue gas duct on the boiler. Each of the probes
consists of three stainless steel sampling tubes (short, medium and long)
reaching to the mid-point of zones of equal area through the depth of the
duct. Thus, a minimum of six sampling points per duct} or 12 per boiler>
are provided, assuring that representative gas samples are obtained. The
soundness of this approach to gaseous emission sampling was verified by
the results of another EPA supported ER&E study on the stratification of
gaseous species in boiler ducts (5).
A complete range of calibration gas cylinders in appropriate con-
centrations with N£ purge and zero gas for each analyzer are installed
in the system. The instruments are calibrated daily before each test,
and also in-between tests, if necessary, to assure reliable and accurate
results.
Particulate and SOx Sampling
To control NOX emissions, combustion modification techniques are
used at less intense combustion conditions than those corresponding to
conventional firing methods. Adjusting secondary air dampers, lowering
excess air, or staging the combustion pattern may increase unburned
combustibles. Beyond the issue of combustibles (soot particulates and
unbumed carbon) , modifications to the combustion process for NOX emission
control may also affect the quantity, size distribution and composition
of the particulate matter emitted from the boiler.
To obtain information on the effect of combustion modifications
on particulate emissions, the field studies included the measurement of
particulate mass loading and fly ash particle size distribution upstream
of the particulate collector device of pulverized coal fired utility
boilers. Measurements of total mass loading and particle size distri-
bution were made under baseline and optimized low NOX operating conditions
to determine the extent of adverse effects of staged firing and other NOX
165
-------
emission reduction techniques. Flyash was analyzed for unburned carbon
content for use in these comparisons and in boiler efficiency calculations
discussed later.
Two Joy Manufacturing Company Method 5 trains and two Acurex High
Volume EPA type particulate sampling trains designed according to EPA
Method No. 5 (6) were used to obtain particulate mass loading data. These
trains were modified to incorporate Brink multi-stage cascade impactors
in the heated sampling box for the determination of particle size
distribution. This arrangement permits particle size distribution
determinations on the outside of the boiler duct under isokinetic sampling
conditions.
Wet chemical analyses for the determination of sulfur dioxide and
sulfur trioxide content of boiler flue gases were made using an Exxon
Research and Engineering Company modification of EPA Method No. 6. Flue
gas samples were extracted from the boiler ducts through a heated probe
and passed through a series of two absorbers, the first one containing
isopropanol to absorb the sulfur trioxide and the second one a hydrogen
peroxide solution in isopropanol to absorb the sulfur dioxide. The amounts
absorbed were determined titrimetrically using barium perchlorate as the
titrant and thorin as the indicator.
Furnace Fireside Tube Wall Corrosion Probe Measurements
Under certain conditions, pulverized coal fired boilers are subject
to corrosion of the furnace tubes. Normal corrosion occurs in oxidizing
atmospheres due to the corrosive effect of iron alkali sulfate attack on
the tube metal surface. However, under fuel-rich reducing conditions,
corrosion of the furnace tubes may be accelerated (particularly with high
sulfur, high iron content coals) due to increased slagging on the tube
surface and penetration of iron sulfide into metal surfaces. Under normal
firing conditions this type of corrosion is most likely to occur in areas
where a localized reducing invironment might exist adjacent to the midpoint
of furnace sidewalls near burner elevations where flame impingement might
occur. To counteract such effects normal boiler operating practice is to
increase the excess air level so that an oxidizing atmosphere prevails at
166
-------
these locations, and to increase the fineness of pulverization, so that the
oxidation of the pyrites in the coal is completed before these species
can come into contact with the furnace wall tube surface. For new boilers,
an improved design feature consists of increasing the separation between
the burners and the sidewalls, thus minimizing potential flame impingement
problems.
The most effective combustion modifications used at present for
controlling NOX emissions from coal fired boilers are staged combustion
and low excess air firing; i.e., conditions that are potentially conducive
to furnace tube wall corrosion, In the early stages of these field
programs, the approach used for measuring corrosion rates was to expose
corrosion coupons installed on the end of probes. The probes were inserted
into available openings located near "vulnerable" areas of the furnace
under both baseline.and low NOX firing conditions. Coupons were fabri-
cated of SA 192 carbon steel, the same material as that currently 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 surface
temperature of about 589 K (600°F)] was chosen in order to deliberately
accelerate corrosion so that measurable values could be obtained. Coupons
were also mildly pickled to remove the existing oxide coating prior to
exposure to eliminate potential differences 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 conditions and coupons exposed under normal boiler operating
conditions. Coupon corrosion rates were, however, considerably higher under
both baseline and low NOX conditions than those corresponding to normal
furnace tube corrosion rates because of the accelerated nature of these
corrosion probing tests.
In subsequent corrosion coupon tests, the acid pickling feature was
eliminated. Instead, the coupons were dipped in acetone and air dried
prior to weighing tc remove any oil deposited during machining. Also,
coupon temperatures were reduced to simulate actual furnace tube wall
conditions, at about 589 K (600°F), and three coupons per probe were
used, instead of two, to obtain more information per test. The design of
167
-------
the corrosion probes was based on information supplied by Combustion
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 603-672 K (625-750°F) for the coupons. The cooling
air supply tube is axially adjustable with respect to the corrosion
coupons, so that temperatures of coupons may be balanced. The cooling air
travels backwards and discharges outside the furnace.
Boiler Efficiency
The application of staged combustion for NOX emission control is
based on fuel rich operation of a number of burners. This results in
less intense combustion conditions which may increase carbon burnout
problems. Thus, because this mode of boiler operation may increase the
amount of unburned combustibles, there may be an adverse impact on boiler
efficiency. However, potential efficiency losses may be compensated by
the increase in efficiency resulting from reduced stack losses if low
excess air firing is employed. To determine the effect on boiler efficiency
of combustion modification techniques applied for NOX control, particulate
emission measurements were run for each boiler under baseline and low
NOX operating conditions and the fly ash samples were analyzed for unburned
carbon. Boiler efficiency was then calculated using the ASME Abbreviated
Efficiency Test based on the heat loss method, and the results were
compared to evaluate the effects of low NOX firing on efficiency.
NOV Tests in Gas Turbines
NO emissions were measured in three gas turbines during the course
X
of the current program to establish emission levels with gas and oil
fuels and to investigate the potential for NO reduction. All turbines
X
tested were General Electric MS7001B models having a capacity rating of
50 MW. The first turbine tested was Potomac Electric Power Company's
Morgantown GT No. 3 turbine firing distillate fuel oil with no provisions
168
-------
for water injection. Houston Lighting and Power Company's LH Wharton GI
No. 42 and No. 43 were the second and third turbines tested. Both turbines
have water injection capabilities for NO emission control. Turbine No.
X
42 was tested while firing distillate fuel oil and No. 43 with natural gas.
Gas turbines are much less flexible than boilers to combustion modi-
fication. Combustion air fuel ratios can be changed within limits but
making the required changes would have been a time consuming operation
beyond the scope of this program. As a consequence, investigations of NOX
emission levels from gas turbines in this program were confined to the
effects of gross load and variations in water injection rates with the
different fuels tested.
169
-------
RESULTS AND DISCUSSION
Results of the initial EPA sponsored program covering field test in-
vestigations on twelve coal fired utility boilers employing combustion modi-
fications to reduce NO emissions (2) were reported in Report No. EPA-650/
X
2-74-066 dated June 1974. Tests on nine additional coal fired boilers were
conducted under the current, ongoing contract sponsored by EPA. Testing of
a tenth boiler, also included in the current contract, covered extensive
long and short term testing of corrosion and gaseous emissions on the No. 7
unit at the Crist station of the Gulf Power Company, a 500 MW, horizontally
opposed fired, Foster Wheeler boiler. Results of the long term corrosion
tests are not available at the time of writing this paper.
Table 1 summarizes results of the tests on the nine boilers, mentioned
aboFe, including design and operating features, numbers of short-term tests
conducted, and emission levels determined for baseline and optimum low NO
X
firing conditions for each boiler tested. It may be noted in Table 1 that
uncontrolled NO emissions in the nine boilers ranged from 341 to a maximum
X
of 1383 ppm with only three boilers meeting the New Source Performance
Standard (NSPS) of 0.7 Ibs NO /106 BTU (525 ppm) and two of these, Navajo
X
No. 2 and Comanche No. 1, were designed with overfire airports for NO con-
X
trol. Also note that optimum NO emissions, achieved through combustion
X
modification, successfully reduced emissions below the present NSPS
in all but one case (Mercer No. 1). Average NO reductions were 38%,
X
ranging from 12 to 62%, almost precisely the same as achieved in testing
12 boilers in the prior contract.
Table 2 contains a summary of the five wall-fired boilers tested. It
may be seen that all of the wall-fired boilers under baseline operating
conditions produced NO emissions exceeding the New Source Performance
X
Standards for coal-fired boilers. However, the Gaston No. 1 unit, equipped
with B & W's low N0x burners, had emissions of 389 ppm NO under baseline
conditions, well below the NSPS NOX level. With modified combustion
170
-------
operation, four of these boilers could be operated within the NSPS limits.
The exception, Mercer No. 1, a wet-bottom furnace with high heat release
and limited operating flexibility, required a 20% load reduction in ad-
dition to combustion modifications to meet the standard. Also note that the
average percent reduction in NO emissions for these boilers was 38 percent.
3C
Tangentially fired boilers are summarized in Table 3. It can be seen
that the three boilers equipped with overfire airports operating at full
load, baseline operation with cooling air only through the airports,
emitted NOX below the NSPS level. Morgantown No. 1 (without airports)
emitted NOX slightly above this level when firing 100% coal. Modified
firing using the overfire airports for NOX control, however, reduced NOX
emissions to levels below NSPS requirements. NOX emission reductions
of 27 to 45% were achieved in these boilers for an overall average of 37%,
essentially the same as for the wall-fired boilers.
Nitrogen Oxides Emissions
Details of the results obtained in these field studies are available
through reports prepared for the EPA (2,3). To illustrate the type of
emission reductions achieved in these investigations one tangentially fired
boiler (Salt River Project, Navajo Boiler No. 2), and one wall (horizontally
opposed) fired boiler (Southern Electric Generating Company, E. Co Gaston
Boiler No. 1) retrofitted with low NO burners have been chosen for discus-
2C
sion.
Navajo Unit No. 2 is a twin furnace, 800 MWe, tangentially fired, Com-
3
bustion Engineering boiler. The furnace has a volume of 13,182 m (465,000
cu. ft.), width of 25.5 m (83 ft. 6-3/4 in.) and front to rear length of
12.45 m (40 ft, 10-1/4 in.). Maximum continuous steam flow is 2,450,000
o
kg/hr (5,400,000 Ib/hr), at 814 K (1005 F), with a reheat steam flow of
o
2,200,000 kg/hr (4,850,000 Ib/hr) at 813 K (1003 F) . The normal fuel fired
is Black Mesa sub-bituminous coal with a higher heating value of 24,950
kJ/kg (10,725 Btu/lb), containing 10.4% ash, 10.3% moisture, 38% volatile
material and 41.4% carbon. At maximum continuous rating, 296,000 kg/hr
(652,000 Ib/hr) of coal is fired and design efficiency is 88.77%. Seven
pulverizers feed 56 burners arranged to fire at seven different levels.
Overfire tilting airports are located above the top row of burners.
171
-------
Operating variables included in the experimental program were gross
load, excess air level, burner tilt, and firing pattern.
Fig. 1 is a plot of NO emissions vs. percent oxygen for various
X
staging patterns, including operation of the top level burners on air only.
The least-squares line labelled S7, corresponding to 100% overfire air (OFA),
was drawn from other data for purposes of comparison. Short lines parallel
to the 87 line were drawn through the averaged data for 84, 85 and Sg test
runs. The lowest NO emissions resulted .with the top tier of burners on
X
air only (S ) and cooling air through the overfire air registers. The
beneficial effect of increasing overfire air register openings from 25%
(8.) to 50% (8 ) and then to 75% (S,) is apparent from Fig. 1.
Fig. 2 illustrates the effect of changing overfire air register settings
more directly. Only test runs conducted with approximately equal excess air
levels (3.6 to 4.0% 0 ) and burner tilts (+ 10° to + 15°) are shown on
Fig. 2 so that the effect of overfire air damper settings on NO emission
X
levels can be seen directly. Average NO levels decreased from 193 ng/J
X
(330 ppm) to 186 ng/J (318 ppm), 170 ng/J (290 ppm), and 165 ng/J (283 ppm)
as the overfire air damper openings were increased from 25% to 50%, 75% and
100%, respectively.
The effect of burner tilt on NO emissions with the overfire air
x
dampers 100% open may be seen in Fig. 3. At the higher excess air levels
at which boilers normally operate (4.5 to 5.0% 0 ), burner tilt can have
an appreciable effect on NO emissions as shown in Fig. 3. With combustion
X
modifications and low excess air operation for NO control, however, it may
be seen that emissions are relatively insensitive to burner tilt.
Because Boiler No. 1 at the Ernest C. Gaston Station of the Southern
Electric Generating Company had been retrofitted with newly designed, low
N0x burners by its manufacturer, Babcock and Wilcox, this boiler was
of special interest for testing in these studies. These dual register,
compartmentalized windbox burners produce a limited turbulence, controlled
diffusion flame designed to minimize the amount of fuel and air mixed at
the burner to that necessary to obtain ignition and to sustain combustion
of the fuel (7).
172
-------
Gaston Boiler No. 1 and its sisten unit, Gaston No. 2, are 270 MWe,
18-burner, horizontally opposed, pulverized coal fired Babcock and Wilcox
boilers. Two division walls divide each furnace into three equal compart-
ments, each having six burners arranged three high in both the front and
rear walls of the furnace. The furnace is 18.3 m (60 feet) wide, with a
3 9
volume of 3950 m (139,500 cubic feet) and a total wall area of 2391 m
(25,732 square feet). Six pulverizers, each with 16,300 kg (36,000 pounds)
capacity, feed coal to three burners each, as shown in Fig. 4. The maximum
continuous rating of each boiler is 771,000 kg (1,700,000 pounds) of steam
per hour. Steam temperature controls include flue gas recirculation into
the hopper area of the furnace and spray attemperators for primary super-
heat and reheat.
Fig. 5 represents a plot of NO emissions vs. average percent stoichio-
X
metric air to the active burners. It may be noted that NO emissions in
x
Boilar No, 2 (equipped with conventional burners) at full load of 270 MWe?
averaged 346 ng/J (591 ppm) at 24% excess air. In comparison Boiler No. 1
when operated at full load of 270 MWes averaged only 220 ng/J (377 ppm) at
24% excess air. Therefore, the new dual register burners used in Boiler
No. 1 reduced NO emissions significantly; i.e., by 36% compared to con-
X.
ventional Babcock and Wilcox burners.
In Fig* 5, least-squares, linear regression lines were calculated from
the data points representing normal firing (S ) at 270 MWe, normal firing at
205 MWe, staged firing (S9 - top mill burners on air only on front or rear
wall) at 250 MWe and staged firing operation (S - top mill burners of front
and rear walls on air only) at 190 MWe.
From a full load, base level NO, emission level of 220 ng/J (377 ppm)
•fv
(124% stoichioinetric air), lovr excess air operation reduced NO emissions
A
from Boiler No, 1 by about 20% to 178 ng/J (305 ppm). Two staged firing
patterns on air only (S ). Low excess air, staged firing at 250 MWe reduced
NO emissions to as low as 140 ng/J (240 ppm) with B mill burners on air
X
only (SJ. Low excess air, staged firing at lower loads of 190 and 148 MWe
reduced NO emissions to as low as 106 and 87 ng/J (182 and 148 ppm), res-
x
spectively, with both E and B mill burners on air only (S ). Analysis of
secondary air register setting vs. NOX shows that the lowest NOX level was
173
-------
generally reached when the register setting was approximately 70% open, with
both lower and higher settings producing higher NO levels.
X
PARTICULATE EMISSIONS
As mentioned previously, low NO combustion modifications may tend to
X
increase burnout problems. Particulate mass loading could increase as a
consequence of the increased carbon in the fly ash and precipitator collec-
tion efficiency could be adversely affected. Changes in particle size dis-
tribution as a result of low NO firing could also have a similar adverse
X
effect on the efficiency of the gas cleaning device. In these studies low
NO combustion modification effects on dust loading were investigated using
x
an EPA Method 5 type sampling train incorporating a Brink cascade impactor
for particle size determination. In the latter phase of the contract dust
loading measurements were made with EPA's SASS train sampling system. Re-
sults of the analyses of the latter tests, however, are not available at
this writing and will be presented later.
Comparison of typical particulate mass loading data obtained in this
program using the EPA Method 5 sampling train may be made from the informa-
tion tabulated in Table 4. Comparison of baseline with low NO operation
in the table shows, in some cases, that particulate mass loading tends to
increase with low NO firing, i.e., Gaston No. 1 and No. 2, and Navajo No.
X
2. The effect, however, appears to be minor. In the other boilers,
particulate emissions are either roughly the same or lower with low NO
X
firing. It was concluded from these and other data obtained in prior
program that changes in particulate mass loading due to combustion modifi-
cations for low NO firing are not significant.
The potential effect of low NO operation on changes in flue gas
X
particle size distribution was investigated in this program using a Brink
Cascade impactor. As indicated above, precipitator collection efficiency
could be impaired if low NO operation resulted in a larger proportion of
X
fine material being produced. Table 5 tabulates typical particle size
distribution data obtained in this program and indicates that particle
174
-------
distribution under low NOX conditions is essentially the same as for base-
line operation. It is concluded, therefore, that there are no significant
differences in particle size distribution under low NO firing.
X
BOILER PERFORMANCE
The potential side effect of low NO operation on boiler efficiency
X
was also investigated in these field studies. Typical results of boiler
efficiencies calculated from the data obtained during these tests, using
the ASME Abbreviated Efficiency Test Method, are tabulated in Table 6. The
data in Table 6 shows that in only two cases (E. C. Gaston No. 1 and Mercer
No. 2) are the carbon losses (unburned combustibles) higher for low NO
X
firing. Boiler efficiency consequently would be expected to be adversely
affected. However, calculated boiler efficiencies for low NO operation
X
show no appreciable or significant changes. Debits in efficiency due to
incomplete combustion are offset by increases in efficiency resulting from
low excess air operation. It is concluded- that no significant changes in
efficiency occur when operating at optimum NO emission levels.
X
SULFUR OXIDES EMISSIONS
Flue gas samples taken from the ducts at the air heater inlet were
analyzed wet chemically for S09/S0_ content using an Exxon Research modi-
fication of EPA Method 6. Table 7 provides illustrative data on typical
sulfur oxide emissions measured. A comparison of calculated SO values,
based on 100 percent conversion, with measured values shows that calculated
values are higher. This is as expected since it is well known that a portion
of the sulfur remains in the ash in pulverized coal fired boilers. In a
typical dry bottom boiler, 15 to 25% of the ash remains in the furnace
and is extracted as bottom ash. The remaining 80 to 85% is carried out
with the flue gas where better than 98 percent in most units is collected
in the electrostatic precipitator. A portion of the total sulfur remains
in these accumulations to decrease the amount emitted as S0_ or S0q in the
flue gases.
Emissions of 803 are essentially the same for baseline and low NOX
175
-------
firing, ranging from 5 to 16 ppm. These values are within the anticipated
range of 5 percent of total flue gas SO .
X
FURNACE TUBE CORROSION
Staging the firing pattern to control NO emissions results in com-
X
bustion at substoichiometric conditions in the first stage, creating a
reducing atmosphere in the lower furnace areas. Potentially this type of
operation may result in corrosion of the furnace waterwall tubes due to
iron sulfide attack. Earlier attempts to quantify this potential side
effect were limited to exposure of corrosion probes for 300 hours under
accelerated conditions to obtain "measurable wear". Examples of the
typical high corrosion rates obtained are shown in Table 8 for the Harllee
Branch No. 4 boiler of Georgia Power Company. The wide range of corrosion
rates measured is typical of the accelerated tests.
In the current study, pickling of the corrosion coupons was eliminated
and exposure temperatures were reduced to furnace tube metal temperatures
in order to approximate actual conditions. This resulted in lower corrosion
rates as shown in Table 8 for the Widows Creek and E. C. Gaston boilers,
ranging from 7 to 15 mils per year. However, these results are still
significantly higher than the 2 to 3 mils per year anticipated for actual
furnace tubes under normal firing conditions. Even though somewhat higher
corrosion rates are indicated for low NOX vs. baseline firing, the scatter
in the data was such that these differences could not be established as
statistically significant.
The effects of corrosion rates with time were explored in tests on the
No. 7 boiler at the Crist Station of Gulf Power Company. Probes were
exposed under baseline and low NO firing conditions for approximately
X
30, 300, 600, 740 and 1000 hours. Data obtained are plotted in Fig. 6 with
curves drawn through points of average corrosion rates determined for the
test. As seen in Figure 6 corrosion decreases with time>which is typical
of most high temperature corrosion mechanisms. Initial corrosion rates
are very high with considerable scatter.
176
-------
Exposure times above 600 hours yield low corrosion rates very closely equal
to the steady state corrosion rate observed at 1000 hour exposure time.
Fig. 6 shows that corrosion rates obtained with probes under low NOX condi-
tions in these tests are no higher than for baseline operation.
Long term, steady state corrosion rates measured with corrosion probes
of 9 to 10 mils per year are still significantly higher than the 2-3 mils/
year experienced with actual furnace tubes under normal firing conditions.
It is difficult to correlate the corrosion probe measurements to that
experienced in furnace tubes. A long term corrosion test conducted during
the current program sponsored by EPA, was completed recently in cooperation
with Gulf Power Company and Foster Wheeler. Corrosion rates were measured
by three means: corrosion probes (as described above), ultrasonic mapping
of tubewall thickness at six furnace elevations, and exposure of 8 furnace
tube test panels in corrosion prone areas of the furnace. The results of
the long term corrosion test at Crist No. 7 boiler will be compared to those
of the corrosion probe tests.
Currently, Exxon Research has initiated under contract to EPA
long-term corrosion tests on four coal fired boilers. These tests will be
discussed in further detail in the next presentation.
Gas Turbine Emissions
As mentioned previously, three gas turbines were tested in the current
program for emission levels and potential NOX reduction. Emission levels
as a function of gross load were investigated in all three turbines and the
effects of variation in water injection rate were fully explored in the
gas fired and oil fired turbines equipped with water injection facilities
Table 9 summarize the pertinent data obtained during these tests.
Baseline (uncontrolled) NOX emissions in the two oil fired turbines
(Morgantown No, 3 and Wharton No. 42) appear to be essentially the same at
all loads. As expected, NOX emissions from the natural gas fired turbine
(Wharton No. A3) are only 55 to 60% of those In the oil fired units. As
shown in Table 9, both gross load and water injection rate have, a signif-
icant effect on NOX, CO and HC emissions.
177
-------
The effect of gross load and water Injection rate may be observed
more clearly in Figure 7. At full to peak load, baseline NOX emission
for the Wharton No. 42 oil fired turbine is 214 ng/J (127 ppm-15% 02, dry
basis). As the top line on Figure 7 illustrates NOX emissions decreased
linearly with decreasing load. Reducing load from 51 MW to 20 MW reduced
NOX emissions by 43% (127 to 72 ppm). However, water injection had a larger
influence on NOX emissions than load reduction did. At 55.5 MW, 2.3% water
injection (based on combustion air) reduced NOX emissions to 22 ng/J (13 ppm-
15% 02» dry basis) or about a 90% reduction over baseline emissions. Similar
effects may be observed in Figure 8 for the natural gas fired Wharton No. 43
gas turbine.
178
-------
CONCLUSIONS
The results of this field studies of combustion modifications on coal
fired utility boilers lead to the following conclusions:
1. Tangentially fired utility boilers operate under baseline conditions
with NOX emissions at or near the NSPS level of 301 ng/kj (0.7 lb/106
Btu) . Wall fired units are higher NO emitters under baseline con-
A
ditions.
2. Combustion modifications can significantly reduce NOX emissions for
all types of firing by 40%, albeit with potential loss in boiler
capacity if no special features are incorporated into the boiler system.
3. With special features such as overfire air ports or low NOX burners as
an integral part of the boiler, all types of firing systems appear to
be capable of meeting (or operating below) the NSPS NOX level.
4. Low excess air, staged firing mode of operation is at present the most
effective combination of furnace combustion modifications for
controlling NOX emissions from coal fired utility boilers. The optimum
firing pattern generally established is the operation of the top burners
on air only. This is consistent with the favored use of built-in over-
fire air ports.
5. Low NOX burners appear to produce significant reductions in NOX
emissions from coal fired utility boilers. The effectiveness of
operating low NOX burners with furnace combustion modifications appears
promising, but requires further investigation.
6. The side effects of combustion modifications used for NOX emission con-
trol are not significant, with the possible exception of furnace tube-
wall fireside corrosion. Long term tests are required to assess the
effect of staged firing with coal on fireside corrosion. The effect of
combustion modification on slagging varies from coal to coal and may
179
-------
limit the extent to which combustion modification can be used.
NOX emissions in large gas turbines firing either natural gas or light
fuel oil can be reduced by 80 to 90 percent with water injection.
Natural gas firing produces NOX emissions about 40 percent lower than
with, distillate (light) fuel oil fired in the same size turbine.
180
-------
REFERENCES
1. Bartok, W., Crawford, A. R., and Piegari, G. J., "Systematic Field
Study of NOX Control Methods for Utility Boilers," Esso Research and
Engineering Company Final Report No. GRU.4G.NO.71; (EPA No. APTD 1163,
NTIS No, PB 210739), Dec. 1971.
2. Crawford, A. R., Manny, E. H., and Bartok, W., "Field Testing: Appli-
cation of Combustion Modifications to Control NOX Emissions from Utility
Boilers," EPA-650/2-74-066, NTIS No. PB-237344, Exxon Research and
Engineering Company Final Report, June 1974.
3. Bartok, W., Crawford, A. R., and Piegari, G. J., "Systematic Investi-
gation of Nitrogen Oxide Emissions and Combustion Control Methods for
Utility Boilers," in A1 r Pollution and Its Control, AIChE Symposium
Series, Vol. 68, No. 126, 1972, pp. 30-74.
4. Crawford, A. R., Manny, E. H., and Bartok, W., "Control of Utility
Boiler and Gas Turbine Pollutant Emissions by Combustion Modification -
Phase I," EPA~600/7-78-036a (NTIS No. PB 281078), Exxon Research and
Engineering Company report, EPRI Project No. 200, March 1978.
5. Gregory, M. W., et al., "Determination of the Magnitude of S02, NO,
C02 and 02 Stratification in the Ducting of Fossil Fuel Fired Power
Plants," Paper No. 76-35.b, Air Pollution Control Association Annual
Meeting, Portland, Oregon, June 27-July 1, 1976.
6. Environmental Protection Agency, "Standards of Performance for New
Stationary Sources," Method 5, Federal Register, Vol. 36, No. 247,
Dec. 23, 1971, p. 24888.
7. Brackett, C. E., and Barsin, J. A., "The Dual Register Pulverized Coal
Burner, 'A NOX Control Device'," The Proceedings of the NOX Control
Seminar 5 EPRI Special Report SR-39, Feb. 1976.
181
-------
400
05
03
PQ
M
Q
CM
o
S-S
O
a
350
300
250
87 (From
" igure 3-7)
- Top Row Air Only
- 25% OFA
- 50% OFA
- 75% OFA
- 100% OFA
JL
3.0 3.5 4.0 4.5
Average % Oxygen Measured in Flue Gas
5.0
5.5
Figure 1. PPM NO Emissions Vs. % Oxygen for Navajo No. 2 Unit
Under Staged Firing Operation
182
-------
350
to
•H
co
H)
M
300
250
8
g
P.
1
20 40 60 80
Overfire Air Dampers - % Open
17)3.6% Oxygeti
18)3.6% Oxy
•^A
18E)3.7% Oxyken
100
Figure 2. Effect of Opening Overfire Air Dampers on W^ Emissions From Navajo
No. 2 Unit at Full Load
183
-------
500
400-
350
CO
rt
P3
o
&-S
CO
o* 300
250
+10 to +15 Burner Tilt
AlO to -15
^Burner Tilt
y
0 Burner Tilt
1
3.5 4.0 4.5 5.0 5.5
Average % Oxygen Measured in Flue Gas
6.0
Figure 3. PPM NO Emissions vs. % Oxygen for Navajo No. 2 Unit Operating
With Overfire Air Dampers 100% Open
184
-------
Division Walls
CO
tn
Mill
B
D
Front-Wall Burners
Figure 4. E. C. Gaston - Boiler No. 1
-------
600 -
500 -
a
PC
400
300
200 -
100
70
80
S3-190 MWe
MWe
Sji-270 M
(Results from Boiler
No. 2 equipped with
conventional burners)
JL
S,-205 KV
90 100 110 120 130 140
Average 7, Stoichiometric Air to Active Burners
ISO
160
Figure 5. E. C. Gaston - Boilers No. 1 & No. 2
-------
09
CRIST STATION - NO. 7 BOILER
CORROSION PROBE MEASUREMENTS
EFFECT OF CORROSION WITH TIME
Tine (Ura.)
Figure 6. Effect of corrosion with time.
-------
120 h-
Q.
I
co
3
CM
O
O
C/5
CO
X
100 h
80 h-
Figure 7. PPM NOX emissions vs
gross load on Wharton No. 42, firing oil.
188
-------
120
100
o.
i
2 80
I
60
o
to
2 40
20
0,75% H20
2 H20
1,9% H20
20
30 AO
GROSS LOAD - W
50
60
Figure 8. PPM NOX vs gross
load - Wharton No. 43 unit firing gas.
189
-------
TABLE 1. SUMMARY OF COAL AND MIXED FUEL FIRED BOILERS TESTED
NOX Emissions
Boiler Operator
Tennessee Valley Authority
Southern Electric
Generating Company
Alabama Power Company
Potomac Electric
Power Company
Public Service Company
of Colorado
, Public Service Electric
UD
O East Kentucky
Power Cooperative, Inc.
Public Service Company
of Colorado
Average of Coal Fired Boilers
Station and Boiler
Boiler No. Mfr.
(a)
Widows Creek-5 B&W
E. C. Gaston-1 B&W
Barry- 2 CE
Morgantown-1 CE
Navajo-2 CE
Comanche-1 CE
Mercer-1 F-W
Cooper-2 B&W
Comanche-2 B&W
i
Type of Fuel
Firing Burned
(b) (c)
RW C
H0(f) C
T(g) CG
T CO
T(g) C
FW(h) C
FW C
HO C
MCR
(MHe)
125
270
130
575
800
350
270
220
350
No. of
Burners
16
18
16
40
56
20
24
18
32
Test Test
Variables Runs
4
5
6
5
4
4
4
4
4
31(d)(e)
37(d)(e)
38
27
36(d)(e)
30(d)(e)
33(d)(e)
101
-------
TABLE II. SUMMARY OF NO EMISSIONS FOR PULVERIZED COAL, WALL-FIRED BOILERS
FULL LOAD
NOX Emissions
Boiler- Type of Firing (a)
1. Widows Creek No. 5-RW
2. Ernest C. Gaston No. 1-HO
(Low NO-, Burners)
A
3. Mercer No. 1-FW
A, Cooper No. 2-FW
5, Comanche No, 2-HO
(Overfire Air Ports)
Operating Mode
(Gross Load - MWe) % 0^
Base
"Low NOX - I"
Base (No. 2 Unit)-
"Low NOX - I"
Base -
"Low NOX - I"
Base -
"Low NOX - I"
Base
"Low NOX - I"
Average f
125
125
270
270
290
284
208
205
358
355
• Base
Low
4.0
3.2
4.3
2.4
3.9
1.8
3.1
2.6
5.2
4,5
NO
•v
ppm
(3% 02)
597(567)(b)
468
591(491) (b)
278
1383(1354) (b)
876
557(575) (b)
483
726(488) (b)
278
771
477
Lb.
106 Btu
0.81
0.64
0.80
0.38
1.88
1.19
0.76
0.66
0.99
0.38
ng/J
349
274
348
163
809
512
325
282
425
163
ppm CO
(3% 0 )
29
88
36
65
22
—
30
61
19
63
% Reduction 38
(a) Types of Firing: RW, rear wall; HO, horizontally opposed; FW, front wall.
(b) Baseline NO emissions calculated for 20% excess air operation.
-------
to
TABLE III. SUMMARY OF NO^ EMISSIONS FOR PULVERIZED COAL, TANGENTIALLY FIRED BOILERS
FULL LOAD
NOY Emissions
1.
2.
3.
4.
Boiler
Navajo No. 2
(Overfire Air Ports)
Comanche No . 1
(Overfire Air Ports)
Barry No. 2
(Overfire Air Ports)
Morgantown No . 1
Operating
(Gross Load
Base
"Low NOY -
A
Base
"Low NOX -
Base
"Low NOX -
Base (b)
"Low NOX -
Average
Mode
- MWe)
- 800
I" - 802
- 323
I" - 333
- 130
I" - 129
- 570
I" - 572
f Base
% 02
4.8
3.6
4.0
3.7
4.9
3.3
4.9
4.7
< Low NOX
v % Reductions
ppm
(3% 02 )
492(400) (a)
282
417(370)(a)
261
341(250) (a)
189
552(534) (a)
403 (c)
451
284
37
Lb.
106 BTU
0.67
0.38
0.57
0.35
0.46
0.26
0.75
0.55
ng/J
288
165
244
153
199
111
323
236
ppm CO
(3% 02)
64
330
27
35
22
49
21
20
(a) Baseline NOX emissions calculated for 20% excess air operation.
(b) Fuel burned = 93% coal, 7% oil.
(c) Fuel burned = 74% coal, 26% oil.
-------
TABLE IV. PARTICULATE MASS LOADING RESULTS
(a)
UD
Boiler
Widows Creek No. 5
E, C. Gaston No. 2
E, C. Gaston No. 1
Nava j o No . 2
Comanche No . 1
Mercer No . 1
Firing
Condition
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Load
MWe
124
125
272
268
792
805
324
321
267
267
Mans Loading
Gr/SCP
4.45
3.55
4.74
5.05
2.86
4.22
2.58
2.13
1,95
2,32
np;/J
2290
1830
2550
2670
1580
2290
1350
1120
1100
1230
Ib/MBtu
5.33
4.26
5.94
6.20
3.7
5.3
3.1
2.6
2.6
2.9
Carbon on '
Flyash, Wt %
9.1
8.1
1.9
4.4
1.6
1.2.
0.6
0.3
1.9
3.5
^Ash in
Coal, Wt %
18.3-
12.3
15.0
10.6
7.8
8.1
5.4
5.4
9.9
9.9
(a) Flue gas stream measured at air heater outlet upstream of particulate collector.
-------
TABLE V. TYPICAL PARTICLE SIZE DISTRIBUTION, WT, %
BRINK CASCADE IMPACTOR
Navaj o No . 2
Size Range
>2.5y
2.5y
1.5y
l.Oy
0.5y
<0.5y
Base
92.7
3.0
0.7
1.0
1.2
1.4
Low NO
X
93.9
1.9
0.6
0.9
1.1
1.6
Comanche No . 1
Base
81.8
8.9
2.0
2.6
2.9
1.6
Low NO
X
80.7
8.9
2.3
2.9
3.3
1.9
Mer.cer No . 1
Base
85.0
3.5
2.3
2.2
1.2
5.8
Low NO
X
86.4
5.3
1.8
2.0
1.2
3.3
-------
TABLE VI. SUMMARY OF BOILER PERFORMANCE CALCULATIONS
10
en
Boiler
Widows Creek No. 5
E. C. Gaston No. 1
Navaj o No . 2
Comanche No . 1
.Mereer No . 1
Firing
Mode
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
Load,
MWe
124
125
270
270
792
798
325
330
267
267
NOx Emissions
Corrected to 3%
% 0
2
2.5
2.5
4.2
3,9
3.9
3.7
3.7
3.4
3.4
1.9
PPM
441
412
606
349
332
280
380
205
1113
897
Coal"
Ash, %
02 (Wet
Basis)
18.27
12.30
13.28
12.12
5.94
6.04
5.36
5.62
9.93
9.90
% Carbon on
Particulate
9.1
8.1
2.1
5.7
1.8
1.5
0.66
0.59
1.9
3,5
Boiler
Efficiency, '
86.2
87.3
89.2
89.2
89.6
89.7
80.8
80.3
90.2
90.1
-------
TABLE VII. TYPICAL WET CHEMICAL S02/S03 FLUE GAS ANALYSES
CT>
Generating
Station/Boiler
Coraanche No . 1
Salt River No. 2
Mercer No . 1
Firing
Mode
Baseline
Low NOX
Baseline
Low NOX
Baseline
Low NOX
02
3.0
" 3.0
3.0
3.0
4.4
2.4
Calculated
ppm
376
(b)
393
(b)
1039
990
Measured SOX
S02
ppm
224
238
330
207
980
722
S03
ppm
9
10
6
10
10
10
S03/S02
Mole Ratio
0.034
0.035
0.018
0.047
0.010
0.014
(a) Based on 100% conversion of fuel sulfur to SO .
x
(b) Fuel analysis not available.
-------
TABLE VIII. FIRESIDE CORROSION RATE MEASUREMENTS
C300 Hour Corrosion Coupon Exposure)
Boiler
Harllee Branch No, 4
Widows Creek No. 6
No. 5
E. C. Gaston No. 2
E, C. Gaston No. 1
Firing Condition
(Baseline
Baseline
f Low NO
V Low NOX
(Baseline
Baseline
/ Low NOX
V Low NOX
f Baseline
V Baseline
f Low NOX
VLow NO.,
Coupon Corrosion Rate
Mlls/Yr.
73(a)
37 (a)
75 (a)
115 (a)
12
11
15
13
9
7
14
9
(a) Accelerated corrosion tests,
-------
TABLE IX. SUMMARY OF NO , CO AND HC EMISSIONS FOR GAS AND OIL-FIRED GAS TURBINES
x
CO
Emissions
Operating Mode
Gas Turbine Fuel Burned -Gross Load, MWe
1. Morgantown Station
Gas Turbine No. 3 No. 2 Fuel Oil Base
Base
Base
Base
2. Wharton Station
Gas Turbine No. 42 No. 2 Fuel Oil Base
0 Q TT (-,
^ . J ri^U
Base
1.5% H20
Base
0.8% H20
3. Wharton Station
Gas Turbine No. 43 Natural Gas Base
1.9% H20
Base
0.5% H20
Base
0.75% H20
Base
- 54
- 48
- 25
- 10
- 51
- 55.5
- 40.0
- 37.5
- 20.0
- 20.0
- 54
- 56
- 43
- 43
- 30
- 30
- 20
°2
ppm
% (15% 0?)
14.7
15.2
17.3
17.9
15.3
14.6
16.5
16.4
17.3
17.4
15.5
15.1
16.0
16.0
17.5
17.6
18.1
133
125
101
82
127
13
106
17
72
19
73
11
62
34
64
10
42
NO
Lbs.
106 BTU
0.58
0.55
0'.47
0.36
0.50
0.05
0.41
0.07
0.03
0.07
0.26
0.04
0.22
0.12
0.23
0.04
0.15
ng/J
224
214
182
138
214
22
178
29
121
31
112
17
95
52
97
15
64
CO
HC
ppm ppm
(15% 0?)(15% 02)
81
30
4
4
4
53
3
142
19
'114
4
60
4
7
18
336
42
5
6
10
24
3
2
3
20
3
19
2
51
4
4
5
131
10
-------
CORROSION TESTING OF UTILITY BOILER COMBUSTION MODIFICATIONS
By:
P. S. Natanson, A. R. Crawford, E. H. Manny and W. Bartok
Exxon Research and Engineering Company
Government Research Laboratories
Linden, New Jersey 07036
199
-------
ABSTRACT
Previous Exxon Research and Engineering Company field studies have
shown that NOX emissions from coal fired utility boilers can be reduced by
about 30 to 50 percent by combustion modifications. Simultaneous studies of
side effects of combustion modifications showed no significant problems in
short term tests. However, only limited and inconclusive data are available
on the effect of combustion modification on fireside corrosion rates of
boiler tube walls. Exxon is presently under contract to the United States
Environmental Protection Agency to determine the long term effect of com-
bustion modifications on the corrosion rate of furnace tube walls of four new
coal fired utility boilers designed to meet the New Source Performance
Standard of 0.7 lb/lC)6 Btu (301 mg/J) for NOX emissions. The respective
manufacturers of these boilers will be involved as subcontractors in this
long term corrosion test program. Comprehensive emission measurement
programs will be conducted in parallel with the corrosion tests on the
four boilers selected for this study. Corrosion rates will be determined in
two-year test programs on each boiler by three methods—installation of
corrosion panels in the furnace tube walls, ultrasonic mapping of furnace
tube wall thicknesses, and exposure of corrosion probe coupons. This paper
reports the work planned and the present status of this contract.
200
-------
ACKNOWLEDGEMENTS
Exxon Research and Engineering Company wishes to thank Mr. David
Lachapelle, the EPA Project Officer on this contract for his guidance. We
also acknowledge the cooperation of boiler manufacturers and electric uti-
lities without which this program would not be possible.
201
-------
INTRODUCTION
Exxon Research and Engineering (ER&E) Company under contracts funded
by the United States Environmental Protection Agency (EPA) has shown in
short term tests typical NOX reductions of 30 to 50% from coal fired utility
boilers by modification of the combustion process (1). Subsequent short
term studies of the side effects of combustion modification did not reveal
any significant problems (2) with the possible exception of aggravated
furnace tube wall corrosion and slagging that may result from the fuel rich
conditions associated with staged combustion of coal, particularly with high
sulfur content. Short term corrosion rate measurements have since demon-
strated that low NOX firing of coal does not result in catastrophic corrosion
tendencies, however, actual wastage rates could not be determined from these
measurements. Currently, U.S. boiler manufacturers rely on combustion modi-
fications such as staged combustion and/or low NOX burner design to meet the
New Source Performance Standard (NSPS) of Q.7 NOX/1Q6 Btu for new coal fired
boilers. Additional boiler flexibility for combustion modification on NSPS
boilers is provided by additional combustion control features such as
tiltable burners, variable flue gas recirculation and overfire air. A key
question is whether combustion modifications, particularly for high sulfur
coals will result in accelerated tube wall corrosion rates, and if they do,
what engineering solutions need to be employed.
The long term (2-year) corrosion studies to be done under this program
(3) will be performed on four new coal fired utility boilers which have
been designed by their manufacturers to meet the new source performance
standards (NSPS) for NOX emissions (0.7/m Btu). U.S. boiler manufacturers
(Babcock and Wilcox, Combustion Engineering and Foster Wheeler) will function
as Exxon's subcontractors in performing the long term corrosion tests on
boilers of their design.
202
-------
Because of the potential corrosion problems associated with low NO
X.
operation, a special long term corrosion study was made under a previous
contract (4) at the #7 unit of Gulf Power Company's CRIST Generating Station
in Pensicola, Florida. CRIST #7 is a 500 MW, horizontally opposed, coal
fired boiler containing 24 burners and manufactured by Foster Wheeler. It
is a pre-NSPS unit.
Sections of boiler wall tubes were replaced with specially measured
and characterized sections (corrosion panels). Approximately one year later,
these special sections were removed and corrosion rates were determined.
Gulf Power Company agreed to operate under low NOX conditions to the extent
practical during the one year test, in accord with guidelines based upon
extensive tests at that boiler and others. Tube wall thicknesses were
ultrasonically measured at hundreds of points within the boiler to provide a
thorough map of tube wall thickness both before and after the one year
corrosion panel exposure period. Corrosion rates were measured by probes,
replaceable wall tube panels, and ultrasonic thickness mapping. Furnace
gas composition was mapped by gas taps, and boiler operating data were
recorded from the control room instruments. Data were gathered to check the
correlation between these different variables to determine if simpler
techniques could provide the corrosion information desired.
The new program involving NSPS boilers will include all of the corrosion
tests performed at CRIST #7, in addition to other comprehensive testing
described in the next section of this paper.
203
-------
PROGRAM
OBJECTIVES
During the test program, extensive corrosion data will be gathered by
probes, panels (removable tube wall sections), and periodic ultrasonic
thickness (UT) mapping. An attempt will be made to show correlation among
these three sets of data and with data from older programs. Boiler per-
formance and flexibility will be evaluated. Corrosion rates will be
correlated with operating conditions (base, low NOX, etc.). Other side
effects of NSPS low NOX operation will also be evaluated, including complete
pollutant assessment by a Level 1 test, and a 30-day continuous flue gas
monitoring period with extensive reference method and calibration checks
at each boiler.
BOILER SELECTION
Each of the four boilers in this test program represents designs of the
four major manufacturers, and as such, the manufacturers have worked closely
with us recommending candidate host sites. Each boiler burns high sulfur
(>2.25 wt %) bituminous coal and produces more than 125 MW of electricity.
These boilers were designed to meet the New Source Performance Standards
(NSPS) for NO emissions of 0.7 lb/106 Btu.
X
TEST PROGRAM DESIGN
A specific test program is designed for each boiler selected so
that the effects of each operating variable on a particular boiler can be
statistically evaluated. Independent variables include fuel feed rate,
burner tilt, overfire air flow rate and total air/fuel ratio. Dependent
variables include boiler efficiency, slagging rate, gaseous and particulate
emissions, and corrosion rates. Time and cost considerations will not allow
a separate test of each combination of operating variables, thus emphasizing
the importance of statistically designed test programs.
204
-------
BOILER CHARACTERIZATION
Before corrosion testing can begin, the safe limits pf boiler
flexibility are determined. This involves a statistically designed program
lasting several weeks during which (determine) the effect of the various
combustion controls on emissions and boiler efficiency will be determined.
During this period detailed records of operating conditions and flue gas
composition are kept. Two conditions of special interest are established
and quantified:
Baseline: Uncontrolled, operation at full load.
Low NOX: Full load with combustion controls safely set to yield
lowest practical NOX emissions.
Having established the optimum operating conditions for low NOX
emissions, the boiler is operated at these conditions for 2 to 3 days.
During this period problems not found during the shorter tests, such as
slagging and fouling, can be detected and corrected by adjustment of
operating conditions, thus redefining the optimum low NOx conditions.
Corrosion rates are then measured at both baseline and low NOX conditions.
CORROSION RATF. MEASUREMENTS
Tube wall corrosion rates are measured by three methods: probes, ultra-
sonics mappingj and pantos. Furthermore, furnace gas composition is
monitored in areas of high anticipated corrosion, as well as in certain
control areas where corrosion is not expected to occur.
Furnace Gas Composition
At the start of the test program, furnace gas taps are installed through
the furnace walls in 30 statistically selected locations. Gas samples are
withdrawn and analyzed from each of the taps at various typical operating
conditions so that highly corrosive local furnace environments can be
identified.
Corrosion Probes
After analysis of the gas tap data and identification of corrosive
regions of the furnace, 5 corrosion probes are installed through
205
-------
selected points of the furnace walls where more detailed corrosion data are
desired. Each corrosion probe consists of several specimens (called coupons)
of boiler tube material and other alloys affixed to an air cooled support
such that the coupons are maintained at normal tube wall temperature (^650-
700°F). The probes are inserted through the tube walls into the fur-
nace where they remain for 30, 300, or 1,000 hours. Corrosion rate is
determined by weight loss analysis. Because of the simplicity and relatively
short duration of corrosion probe tests it is possible to use this method to
determine relative corrosion tendencies under a variety of conditions
(including base and low NOX) within a few weeks. This method can also be
used to "spot check" corrosion rates.
Data from the corrosion probe test will:
1. Aid in locating corrosion panels,
2. Establish possible corrosion problems early in the program so that
less corrosive operating conditions can be selected if necessary,
3. Correlate furnace gas composition with probe corrosion rates.
4. Allow correlation of corrosion data between this boiler and other
boilers tested during previous programs.
5. Allow comparison between corrosion rates at low NO and baseline
X
conditions-.
6. Determine applicability of short term corrosion tests to the
prediction of long term corrosion rates.
Ultrasonic Thickness (UT) Measurements and Corrosion Panels:
Long term (1 to 2 year) low NOX operation will begin only after the
shorter tests, just described, establish problem-free operation under low NOX
conditions. Prior to the start of the two year low NOx operating period, the
boiler is entered during a scheduled outage and selected sections of wall
tubes are removed from the boiler circuit and replaced with approximately
6 carefully characterized sections called corrosion panels, each about 5 to
10 tubes wide and 10 feet long. The location of the panels will have been
determined from gas tap and corrosion probe information. In this way, more
206
-------
data.can be collected in certain locations of interest. Each panel contains
three thermocouples so that a temperature history of the boiler walls can
be recorded. While the panels are being installed, ultrasonic thickness
measurements are performed at hundreds of points encompassing 6 elevations of
each wall of the interior of the boiler furnace. This information provides a
thickness map of the boiler walls prior to low NOX operation.
The ultrasonic thickness measurements have been carefully planned so
that subsequent cleaning and measurement of the same vicinities at a later
time will yield statistically meaningful results. The cleaning procedure
used to prepare the surface is designed not to interfere with the measurement.
Periodic visits will be made to the boiler during the 2 year period of low
NOX operation to measure emissions, corrosion rates by probes, and verify the
effectiveness of low NOX operation. Ultrasonic thickness measurements will
be performed twice more during the 2-year test, and a section of each
corrosion panel will be removed for analysis during each UT measurement
period. This analysis includes physical'measurements as well as chemical
analysis of slag, deposits, and the metallurgical characterization of tube
wall material.
Since periodic corrosion probe tests will have been made during the 2
year corrosion panel exposure period, it should be possible to correlate
data from probes, panels, ultrasonic thickness measurements, and furnace
gas taps.
EMISSION TEST PROGRAM
During the period of boiler characterization, flue gas monitors are
used to record NO and other emissions while combustion control parameters
are varied to explore optimum low NO conditions. Flue gas composition will
X
also be monitored periodically during the 2 year period of low NOX operation,
especially while corrosion probe tests are in progress. More extensive
emission measurements done during low NOX operation include a Level 1 test
of effluent streams, and a period of 30 days during which emissions are
monitored continuously.
Flue Gas Analysis
Flue gas is monitored at 12 points between the economizer and air heater
207
-------
according to procedures that have been shown to yield meaningful results (5).
The monitors are contained in a mobile emissions sampling van and include 02,
.CO, C02, light hydrocarbons, NO, NOX, and S02 analyzers. S02 and S03 are
also measured by wet chemical methods.
Level 1 Tes^t
A Level 1 test is performed once on each boiler during its low N0x
operating period. This is an extensive pollutant assessment test involving
all pertinent streams solid, liquid, and gas entering or leaving the
boiler. Gas and particulate samples are taken upstream and downstream of the
electrostatic precipitator (ESP) so that the effect of low NOX operation on
particulate loading and composition and ESP efficiency can be determined by
comparison with similar tests done at base conditions earlier in the program.
Level 1 tasks also include fuel analysis, and biological toxicity tests of
flue gas particles, bottom ash, and precipitator ash.
The source assessment sampling system (SASS), a modified Method 5 train,
is used for flue gas and particulate sampling. It gives information on size
distribution, and trace species concentrations as well as data on polychlori-
nated biphenyls (PCB) and polycyclic organic matter (POM).
Continuous Monitors
The low NOX operating period will include 30 days of continuous
monitoring for N0/N02 and S02 according to EPA guidelines and with
frequent reference method checks. An attempt will be made to run this
test concurrently with the Level 1 test, corrosion probe tests, and flue
gas analyses.
PRESENT STATUS AND PLANS
As of this writing, agreement has been reached with two utilities for
testing of their units with the restective boiler manufacturers.
• Columbus & Southern Ohio Electric Company
Conesville Unit 5 (410 MWe)
Manufacturer: Combustion Engineering, Inc.
Start of Testing: Fall, 1978
208
-------
• Louisville Gas & Electric Company
Mill Creek Unit 3 (450 MWe)
Manufacturer: Babcock & Wilcox
Start of Testing: Spring, 1979
Foster Wheeler is presently screening a number of candidate units
suitable for long term testing. This boiler should be selected by early
1979.
A fourth unit has yet to be selected. Manufacturers are screening their
customers for a candidate boiler. It is expected that this unit will be selected
early in 1979.
209
-------
IN SUMMARY
The data gathered during this program will determine the effects of
various combustion controls on flue gas emissions and'boiler efficiency.
These, together with careful observations of side effects such as slagging
and fouling will help determine optimum operating conditions for low NO
X
emission for large, high-sulfur coal fired, NSPS utility boilers. Corrosion
rates will be determined at this low NOX condition and a basis will be
established for comparison with corrosion rates at base conditions. The
effects of low NOX operation will be extensively quantified and compared with
baseline emissions. Toxicity of specific effluents will also be determined
under low NOX conditions by biological testing.
The data are expected to yield valuable information to utility boiler
operations and manufacturers on corrosion rates under low NOX operating
conditions of coal fired NSPS boilers. This information will be used to
determine whether corrective measures are required to reduce corrosion rates
resulting from low NOX firing of coal.
210
-------
BIBLIOGRAPHY
1. Bartok, W. , Crawford, A. R. , Manny, E. H., and Piegari, G., "Reduction
of Nitrogen Oxide Emissions from Electric Utility Boilers by Modified
Combustion Operation." Presented at the "American Flame Days"
Symposium sponsored by the American Flame Research Committee, Sept. 6-7,
1972, by Exxon Research and Engineering Company, Linden, N.J.
2. Crawford, A. R., Manny, E. H., and Bartok, W., "Control of Utility Boiler
and Gas Turbine Pollutant Emissions by Combustion Modification - Phase I.
EPA-600/7-78-036a (NTIS No. PB 281078), Exxon Research and Engineering
Company Report, EPRI Project No. 200, March 1978.
3. United States Environmental Protection Agency Contract No. 68-02-2696.
4. United States Environmental Protection Agency Contract No. 68-02-1415.
5. Crawford, A. R. , Gregory, M. W., Manny, E. H. and Bartok, W., "Magnitude
of S02, NO, C02, and 02 Stratification in Power Plant Ducts."
EPA 600/2-75-053. Exxon Research and Engineering Company Report,
Sept. 1975.
211
-------
FIELD EVALUATION OF LOW NOV
/\
COAL BURNERS ON INDUSTRIAL AND UTILITY BOILERS
G. B. Martin
Combustion Research Branch
Energy Assessment and Control Division
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Presented at
Third Stationary Source Combustion Symposium
San Francisco, California
March 5-8, 1979
213
-------
INTRODUCTION
The trend in the U. S. is to the increased use of coal for industrial
and utility heat and steam generation with the potential for an attendant
increase in pollutant emissions. For example, a recent report (1) has esti-
mated that N0x emissions might increase by up to 80 percent by the year
2000 for a scenario of high coal usage with current control technology.
Therefore, the EPA program for development of advanced combustion concepts
is directed toward optimum control of NO and other pollutants from a
.A
range of conventional coal utilization technologies. The program, which
addresses not only direct combustion of coal, but also solid, liquid, and
gaseous fuels that may be derived from coal and/or shale, has been
described in a recent paper (2). A major element of that program is the
development and field evaluation of pulverized coal combustion techniques
directly applicable to conventional steam generators.
projects for field evaluation of a low NO burner concept on industrial and
X
The purpose of this paper is to briefly describe the recently initiated
icts for field evalua-
utility steam generators.
BACKGROUND
The development of the low emission coal burner was initiated in 1970
under a contract with the International Flame Research Foundation for a
study of the effects of burner design on emissions from pulverized coal
combustion. The results led to the conclusion that NO formation resulted
A
predominantly from the conversion of nitrogen compounds released with the
volatile matter and that a burner design which reacted the volatiles in a
fuel-rich atmosphere could result in very low conversions to NO (3).
X
214
-------
A triple concentric burner design was developed which created the optimum
conditions for low NO emissions while maintaining the flame shape required
/\
for boiler application. This development work was done at 2 to 3 thermal
MW (6-9 X 10 Btu/hr) in a refractory tunnel. Both burner scale-up to
nearer practical size and performance evaluation in a system having more
realistic geometry and heat transfer characteristics were required.
The development of scale-up criteria was initiated under a contract
with Energy and Environmental Research Corporation. As a part of that
contract a versatile experimental facility capable of simulating the
geometry and heat transfer characteristics of a field-erected watertube
boiler was constructed. The configuration of the firebox is shown in
Figure 1. The most significant features are: 1) up to 40 thermal MW
rated capacity for either single or multiple burner arrays; 2) variable
geometry for constant volumetric heat release over the range of firing
rates; 3) water-spray cooled sheet steel walls for control of temperature;
and 4) flexibility of operation over a wide range of well controlled
conditions. A smaller system simulating a D-type package watertube and
capable of firing up to 4 thermal MW was constructed for burner design
screening. Using these two facilities, a distributed mixing burner was
developed. One experimental burner configuration, shown schematically
in Figure 2, has three points for addition of combustion air: 1) primary
air used to transport the pulverized coal; 2) secondary air admitted
through two annular channels around the fuel pipe; and 3) tertiary air
admitted through four ports located around the burner perimeter. The
amount of air admitted through each can be controlled independently to
determine optimum conditions. The secondary air is divided into two
streams with independent swirl to allow not only flame shape tailoring but
also increased turndown flexibility. The burner variables that influence
NO emissions include: 1) primary air percentage and velocity; 2) fuel
injector design; 3) secondary air percentage, velocity, and swirl distribu-
tion; 3) quarl design; and 4) tertiary air percentage, velocity, direction,
and displacement from burner center!ine. A representative graph of results
obtained at a nominal 15 thermal MW is shown in Figure 3, where NOX emis-
sions are plotted versus excess oxygen in the flue gases. The major
215
-------
variable is the amount of secondary air. With the primary air maintained
at 25 percent of theoretical, secondary air was controlled to give total
air through the burner throat over the range of 50 to 110 percent of
theoretical air. The balance of the air was admitted through the tertiary
ports to bring the total air to the stated excess oxygen conditions.
For this particular case NO emissions range from approximately 500 ppm
A
down to less than 100 ppm. To examine the effects of scale, three
geometrically scaled burners were constructed and tested. The 4 thermal
MW burner was tested in the small system watertube simulator while the
'16 and 35 MW burners were tested in the large watertube simulator at
comparable volumetric heat release rates. Data for all three burners,
run at an "optimum" condition for the 16 MW burner, are shown in Figure
4. The NO level for the 16 MW burner is significantly lower than for
/\
the 4 MW burner while the 35 MW results fall in between. Further work
on the effects of scale is in progress to explain these observations.
Having established the potential of the low NO burner concept for a
/\
limited number of coal types, the development program was expanded to
extend its applicability to the full range of U. S. coals, to further
optimize single burner concepts, and to explore additional multiple burner
configurations. The approach is to use a small scale furnace for screening
of fuels and concepts followed by large burner experiments on fuels and
concepts of specific interest. The fuels selected for the screening cover
the full range of U. S. coals from anthracite to bituminous plus selected
foreign coals. In addition, chemically and/or physically cleaned coals
will be included as they become available. Based on the results of the
small scale screening of about 36 coals, 12 will be selected for burner
testing at a nominal 4 thermal MW and 6 for testing at the larger scale
(15 to 40 MW). The results of these tests will establish criteria for
optimizing burner performance as a function of fuel characteristics. The
small scale furnace will also be used to screen other combustor concepts
capable of achieving lower emission. A more extensive discussion is
presented elsewhere in this Symposium for both the small scale fuel experi-
ments (4) and the pilot scale development work (5).
216
-------
To provide a basis for widespread application of the technology to
pulverized coal fired boilers it is necessary to evaluate the long term
performance of prototype burners in field boilers. The purpose of this
paper is to describe the general structure of two programs to perform
this field evaluation.
DISCUSSION
A significant fraction of industrial and utility steam generators are
coal fired and projects indicate that the fraction will increase. While
stoker firing shares the market in the smaller sizes, pulverized firing
dominates the large industrial and new utility boiler market. For both
sectors* multiple burner watertube boilers of a variety of designs are
used. The principal distinctions between the sectors are boiler capacity,
duty cycle, and steam quality. Industrial boilers are generally below
293 thermal MW (a nominal 10 Ib of steam/hr) and supply process steam;
whereas, new utility boilers are substantially larger (e.g., 600 elec-
trical MW) and supply high pressure superheated steam for electric
generation. While the similarities are sufficient to make the low NO
A
burner technology equally applicable to each, the differences require
a separate evaluation of the technology for each class of boilers. For
purposes of the program, the target size range for the field evaluation
boilers was 29 to 145 thermal MW (100 to 500 thousand Ib of steam/hr)
for the industrial boilers and 100 to 300 electrical MW for the utility
boilers. Two boilers of each type will be evaluated.
The field evaluation of the low NO coal burner must address a variety
X
of technical issues on the translation of an experimental technology into
a system that can be used in practical stearn generators. Since the experi-
mental burners have been tested at heat inputs comparable to burners cur-
rently used in many industrial and utility boilers, it appears that the
classical problems of scale up should be minimized. The remaining problems
are related to mechanical reliability and operational flexibility of the
prototype burners. The specific requirement imposed OR the burner design
217
-------
will depend on the characteristics of the host boiler. Other aspects of
the program, such as schedule., will also be impacted by the boiler selection.
Therefore, the field evaluation program requires the definition of host
boiler requirements and the selection of specific sites as the first step.
This is followed by the design of a prototype burner to meet the specific
requirements for each host boiler. Based on the design phase for each
boiler, a prototype burner can be fabricated and mechanical operability
and performance optimization can be carried out in the experimental system.
This prototype testing will allow the specification of the final burner
configuration for each boiler. The field evaluation will entail the fabri-
cation and installation of the required number of burners, the reoptimiza-
tion of performance in the boiler, and the long term operation under prac-
tical conditions. Finally, the results must be used to generalize the
technology for application to the full range of U. S. boiler designs and
coal types.
PROGRAM STRUCTURE
The program of the field evaluation is virtually the same for indus-
trial and utility boiler contracts, with some differences in the detailed
requirements. The scope of work consists of nine tasks which are discussed
below.
Program Definition
The program definition task provides for the initial planning and
continuing review and management of the contract effort. Specific elements
include: host boiler selection, engineering design of the prototype
burner, definition of the measurement plan, and preparation of an overall
program plan.
218
-------
Boiler Selection--
The results of this subtask will establish the requirements to be
satisfied in all other tasks of the contract. Existing information will
be screened to establish design and operating characteristics of the
specific boiler population of interest. This study will deal with
historical data on existing boilers and with projections of trends for
future sales. Based on this, a set of criteria will be derived for
selection of two boilers of distinctly different characteristics required
to show the generalization of the technology. Following these criteria,
a list of existing boilers will be assembled and used to select potential
host sites. This list will be screened for additional desirable charac-
teristics, such as: 1) redundant capacity at the host site to allow the
greatest flexibility in the test program, 2) continuous steam require-
ment to maximize data collection over the evaluation period, 3) past
characterization of emissions, and 4) good history of reliable operation
with minimum problems. The potential hosts will be contacted to deter-
mine their interest in participating in this program and detailed discus-
sions will be held with all potential hosts. The contractor and the EPA
project officer will make a final selection of host sites based on the
results of this process. Following final negotiations with the hosts,
the contractor will prepare a site specific plan for each boiler, including:
1) schedule of work, 2) provision of required clearances, 3) definition
of equipment lead time, 4) schedule of boiler outage, and 5) coordination
of all involved parties.
Burner Engineering Design--
The experimental data on the low NO coal burner will be translated
/\
into a prototype burner design engineered to meet the specific requirements
of each host boiler. This design package will include specifications not
only of the burner but also of all support and auxiliary systems required
to interface the burner to the boiler. It will also identify compliance
with all codes governing fabrication, installation, and operation of the
burners. The burner design package will be subjected to detailed review
by the EPAS the representatives of the host site, and an independent commit-
tee of experts, prior to final approval.
219
-------
Analytical Measurements—
The approach to the measurement of pollutant emissions and boiler
thermal performance must be defined in a measurement protocol to ensure that
the data from the study are complete and accurate. The frequency with which
any given measurement is made will depend on both the complexity and the
costs involved. The protocol will specify the measurement frequency for
input and output parameters. For monitoring performance of the burners,
certain gas phase species (CO, 02, and N0x) will be monitored continuously
over the duration of the long term tests. Other gas phase species (NO, S02>
and HC) will be measured periodically. Under conditions that warrant
further investigation, more extensive measurement will be made including:
1) particulate mass, 2) ash carbon content, 3) S03, and 4) sulfur reten-
tion in the ash. Finally, for a very limited number of conditions, a
comprehensive analysis of the combustion side effluent streams will be
made. The procedures to be used for this comprehensive analysis will
satisfy the requirements of the EPA's Environmental Assessment proce-
dures (6).
The periodic gas phase measurements and the more comprehensive
analysis of the effluent streams will be made during three to five 30-
day detailed measurement periods during boiler baseline characterization
and long term performance evaluation. For the 18 month evaluation, two
of the 30-day periods will occur in the first and last months, with the
balance equally spaced in between. Detailed evaluation of the thermal
performance will also be conducted during these periods.
Overall Program Plan--
The results of the three subtasks discussed above will be combined
into a program plan covering all aspects of the performance evaluation.
This plan will be reviewed periodically as part of overall contract
management.
220
-------
Prototype_Construc.t1on and Testing
The first step in the evaluation of the prototype burner design dis-
cussed above is the comparison of its performance to that of the experimen-
tal burner on which it was based. For each host site one prototype burner
will be constructed and tested in the large watertube simulator (LWS)
(see Figure 1). This testing will be carried out with both the baseline
coal from the burner development study and the coal being used at the host
site. The prototype burner will be operated over a range of conditions
simulating the requirements of the host site duty cycle. The performance
of the burner will be optimized and any mechanical problems will be recti-
fied. Once the burner design has been verified, comprehensive measurements
will be made for selected operating conditions. Any modifications of the
burner will be incorporated in the design package. A direct comparison
of the prototype burner to the original equipment manufacturer's (OEM)
burner as installed at the host site will be made. The OEM burner will
be installed in the LWS and operated over a range comparable to the
prototype burner. The results of this comparison will be used to esti-
mate the performance of the prototype burner in the host boiler and will
provide additional data for generalization of the technology.
Boiler Baseline Evaluation
The detailed measurement of the host site boiler operating at the
normal duty cycle will provide a benchmark for evaluation of the perform-
ance of the low NO burner. The baseline evaluation will be directly
/\
comparable to a 30-day detailed evaluation period to be run with the low
NO burner and will include both continuous gas phase measurements and
A
comprehensive analysis of samples at selected conditions. The boiler will
also be operated under modified conditions (e.g., low excess air, staged
combustion) to determine the degree of NOX reduction achievable by tuning
the existing firing system. These tests will be of short duration and
detailed measurements will be limited.
221
-------
Burner Installation
Following the baseline characterizations a scheduled outage will allow
installation of the required number of prototype low NOX coal burners and
all support systems in each host boiler. Other equipment for complete
evaluation of the technology will be installed, including: 1) air and fuel
flow measurement and/or control, 2) boiler thermal performance instrumen-
tation, 3) wallports for sampling and/or observation, 4) a color TV camera
and videotape for remote flame observation, and 5) removable tube panels for
corrosion evaluation. In addition, there is an option for a 30 day evalua-
tion of dry SO control on one industrial boiler if pilot scale results and
/\
short term tests on the boiler warrant. Ultrasonic measurements will also
be made to determine boiler wall thickness at a number of locations not
covered by the corrosion panels. Finally, the boiler will be checked out
and started up using normal procedures and the operability of all systems
will be verified. The boiler will be operated over the normal load range to
establish that the low NO burner performs in a manner fully comparable to
A
the unmodified boiler.
Performance Evaluation
The performance evaluation task is the heart of the low NO burner
J\
program. The first step is the optimization of the burner performance for
the specific host boiler. The initial burner settings for low NO operation
X
will correspond to those identified during testing in the LWS. The burner
conditions will also be varied to determine if the optimum condition in
the multiple burner ooiler is significantly different than those chosen
based on the experimental single burner system. In addition, the use of
overfire air and/or biased firing may be explored to determine if further
NO reduction is achievable.
A
The final burner settings for long term operation will be selected
based on minimum N0x achievable with low carbonaceous emissions, good flame
stability, and acceptable load-following characteristics. Then the 18
month performance evaluation will commence. Detailed measurement of thermal
and emission performance will be conducted during the initial 30 day period
222
-------
of operation under normal duty cycle. Detailed 30 day duration measure-
ments will be made periodically during the next 16 months (one to three
periods depending on need) and again during the 18th month. The contractor
will provide for continuous monitoring of NO, CO, and 09 during the 18
X C.
months. An onsite representative will monitor the overall burner perfor-
mance, provide for any service required on the burners, investigate the
causes of any shift in emission performance, and reduce data on a continu-
ing basis.
Industry Coordination
The purpose of this task is to secure the advice of manufacturers and
users of pulverized coal burners on practical aspects of the technology
and to ensure the timely dissemination of the program results to potential
users. In addition, both the industrial and the utility boiler evaluation
contracts must be coordinated with the continuing development activity
to generalize the applicability of the technology to the full range of U. S.
coals and to other system configurations. This will be accomplished through
the existing Technical Review and Technology Transfer Panels under the
development contract (68-02-2667). The Technical Review Panel, which already
includes representatives of the field evaluation contractors, will be
expanded to include representatives of each host site, and review of all
three contracts will be accomplished during the semiannual meetings.
The plans for and results of the field evaluation will be communicated to
the Technology Transfer Panel, which consists of representatives of Govern-
ment agencies and industry associations, at the annual meetings.
Restoration
At the completion of the 18 month period, a scheduled outage will be
arranged and the boiler will be restored to the configuration when the
program began. The corrosion panels will be removed and used to assess
the extent of tube wastage. Ultrasonic measurements will be made at the
locations measured at the time of low NO burner installation. The
X
223
-------
restored boiler will be started up and checked out to verify that its
function is comparable to that at the start of the program.
Data Analysis
Data will be analyzed continuously over the period of the contract.
Specific analyses include:
1) Compare experimental burners and prototype burner to identify
design changes and establish optimum conditions.
2) Compare prototype burner and OEM burner performance in the
LWS to estimate the performance of the low NOX burner in a multiple
burner watertube boiler.
3} Analyze the baseline data to serve as a benchmark for com-
parison of low NO burner performance.
A
4) Optimize the performance of the low NO burner in each host
boiler and select the condition having the lowest emission
potential.
5) Analyze long term data continuously to detect performance
changes and/or define alterations required in burner operation or
testing procedures.
6) Verify the procedures for extrapolating the experimental
system results into practical burner systems.
Guideline Manual
The ultimate goal of the field evaluation contracts is to provide
the basis for widespread application of low NO burner technology to indus-
/\
trial and utility pulverized-coal-fired boilers. To assist in accomplishing
this goal, a guideline manual will be prepared for each class of boilers.
These manuals will summarize the results of the contract and provide informa-
tion on the methodology of translating the experimental burner data into
practical burner design. Specific areas to be covered include the follow-
ing:
224
-------
1) A summary of experimental data for single and multiple burners
that served as the basis for the prototype burner design.
2) Experimental data on effects of fuel type on burner perfor-
mance.
3) The design approach for the prototype burner and the method of
applying the technology to specific boiler situations. Approach to
multiburner effects shall be emphasized.
4) Comparison of prototype and original equipment burners in the
LWS to the experimental burner on the same facility.
5) Comparison of performance of the low emission burner in the
field operating boilers to projected performance from the LWS
experiments,
6) Identification of problem areas and solutions employed during
this study.
7) Effects of burner technology on non-criteria pollutants and
methods of estimating prior to field test application of a tech-
nology.
8) Specific guidelines for application of the burner technology
to field operating boilers.
SUMMARY
Low NO burner technology has been developed which is capable of
x /•
achieving NO emissions below 86 ng/J (0.2 lb/10 Btu) in an experimental
single burner system operating at 35 thermal MW (105 X 10 Btu/hr). Although
this approaches the size of practical burners for large industrial and util-
ity boilers, reliable operation in multiple burner watertube boilers under
a normal duty cycle must be achieved before the full potential of the
technology can be assessed. Therefore, two programs have been initiated
for long term field evaluation of the low NOV burner technology for indus-
X
trial and utility boilers. Significant features of each program are as
follows:
1) Design and performance verification of prototype burners
based on experimental burner results.
2) Installation of the required burners and support equipment
on host boilers.
225
-------
3) Operation of the low NO burner under normal duty cycle on
each boiler for 18 months.
4) Continuous monitoring of NOX, CO, and 0,,.
5) Three to five detailed 30 day evaluations of emission and
thermal performance, with continuous gas phase analysis and more
comprehensive analysis on selected conditions.
6) Provision for assessment of total environmental impacts of the
technology.
7) Potential for evaluation of dry S02 control technology on one
industrial boiler.
8) Corrosion evaluation including both removable tube panels and
ultrasonic measurements.
9) Preparation of a guideline manual to assist in generalization
of the technology.
The industrial boiler contract was awarded to Energy and Environ-
mental Research Corporation (EERC) with a proposed subcontract to a
major boiler manufacturer. The utility boiler contract was awarded to
Babcock and Hi 1 cox with a proposed subcontract to EERC. Both contracts
are scheduled to be completed in late 1982.
226
-------
REFERENCES
1. Salvesen, K. G., et al. Emission Characterization of Stationary NO
Sources: Volume I. Results. Acurex Corp. EPA Report No. EPA-
600/7-78-120a, NTIS No. PB 284-520/AS, June 1978.
2. Martin, G. B. and D. G. Lachapelle. "The EPA Program to Develop
Advanced Conventional Coal Combustion Methods." Presented at
the Second EPRI NO Control Technology Seminar, Denver, Colorado,
November 8-9, 1978.
3. Heap, M. P., et al. Burner Criteria for NO Control, Vol. I.
Influence of Burner Variables on NO in Pulverized Coal Flames.
International Flame Research Foundation. EPA-600/2-76-0"6Ta., NTIS
No. PB 259-911/AS, March 1976.
4. Heap, M. P., et al. "The Influence of Fuel Characteristics on NO
Formation - Bench Scale Studies," to be presented at the
Third Stationary Source Combustion Symposium^ San Francisco,
California, March 5-8, 1979.
5. Zallen. D. M., et al. "The Generalization of Low Emission Coal
Burner Technology," to be presented at the Third Stationary Source
Combustion Symposium, San Francisco, California, March 5-8, 1979.
6. Hamersma, J. W., S. L. Reynolds and R. F. Maddalone. "IERL-RTP
Procedures Manual: Level 1 Environmental Assessment." TRW Systems
Group. EPA-600/2-76-160a, NTIS No. PB 257-850/AS, June 1976.
227
-------
8.4m
ro
rsj
oo
-n
— i.
to
-$
ro
1C)
n>
QJ
r*
-5
rt-
C.
cr
3
01
<-«•
o
-$
\
\
\
\
7.2m
MAXIMUM I
VOLUME J
•— 3.6 m -i
MINIMUM
VOLUME
12,7m
GROUND LEVEL
FUELS: - PULVERIZED COAL
- HEAVY OIL
FIRING CAPACITY:
- 15-40 THERMAL MW (50-120 X 106 Btu/hr)
HEAT RELEASE RATE:
- 1.6to2.4kW/m3
VARIABLE GEOMETRY
SHEET STEEL WALLS:
- WATER SPRAY COOLED
PULVERIZER-RAYMOND CE (MOD 473A):
6000 kg/hr
AIR PREHEAT TEMPERATURE:
- 700 K
INDEPENDENT CONTROL OF ALL AIR STREAMS
-------
U3
c
-s
n
y
(K
Oi
ft-
^*.
o
a
m
x
m
-5
(T)
3
ft-
©
SWIRL VANES
O
o
g
COAL &
PRIMARY
ASR
era
RETRACTABLE
OILMOZZLE
CERAMIC QUARL
SECONDARY
AiR
TERTIARY
AIR
•COMBUSTOR WALL
-------
700
c
-$
fD
o
o
-a
CO
ro
oo
o
at
cr
fB
i/i
o
3
E
a,
a.
~CNJ
O
h-
O
DC
LLJ
a.
o
X
O
600
500
LARGE WATERTUBE SIMULATOR
LOAD-16 THERMAL MW
PRIMARY AIR - 25 percent OF THEORETICAL
SECONDARY AIR - VARIABLE
400
300
200
100
PRIMARY + SECONDARY,
PERCENT OF THEORETICAL
O> 110
D 110
A 90
O 70
SOLID SYMBOLS - FLAME STABILIZED WITH OIL
0
456
EXCESS OXYGEN, percent
8
-------
400
COAL - UTAH
PRIMARY AIR - 25 percent THEORETICAL
OPTIMUM BURNER CONDITIONS
O4 THERMAL MW
Q16 THERMAL MW
A35 THERMAL MW
40 50 60 70
PRIMARY PLUS SECONDARY AIR (THEORETICAL), percent
Figure 4. Effect of Burner Scale on NOX Emissions.
231
-------
APPLICABILITY OF THE THERMAL DeNO PROCESS
TO COAL FIRED UTILITY BOILERS
By:
G. M. Varga, Jr. and W. Bartok
Government Research Laboratories
Exxon Research and Engineering Company
Linden, N.J. 07036
233
-------
ABSTRACT
Under an EPA contract, the performance and cost of the Exxon Thermal
DeNO Process were projected for coal fired utility boilers. In this pro-
X
cess, ammonia is injected into boiler convective flue gas passages to reduce
NO selectively without a catalyst within a critical temperature range.
x
Eight coal fired utility boilers were selected to determine process appli-
cability to boilers of different manufacturers, sizes, firing methods, and
by coal types.
Thermal DeNO performance was projected both with and without combustion
X
modifications (CM) . Assumed NO reduction targets were (i) the proposed
X
NSPS of 0.5-0.6 Ib. NO /MBtu, and (ii) a deep reduction level of 0.3 to 0.4
X
Ibs. NO /MBtu. The maximum practical level of NOX reduction which could be
achieved was also projected. A total of six cases were established. Per-
formance was projected to be equally applicable to all boilers studied at
full and one or more reduced loads down to 50%, despite significant differ-
ences in flue gas temperature profiles and flow path configurations. Maximum
projected Thermal DeNO performance ranged between 50 and 59% for the coal
X
fired boilers selected for study. For actual applications, design limi-
tations and operating problems that might limit the level of DeNO attainable
x
will have to be assessed in each case.
Budget type cost estimates were prepared for full load. Costs ranged
from 0.25 to 1.23 mills/KW-Hr, excluding preliminary engineering costs and
licensing roya .es. Based on the projected capability of Thermal DeNO to
x
supplement CM for reaching low levels of NO emissions from coal fired
x
utility boilers, a full scale demonstration program is recommended. The
demonstration of Thermal DeNOx on a coal fired utility boiler would include
the potential downstream effects identified from experience with oil fired
boilers.
234
-------
ACKNOWLEDGMENTS
The authors acknowledge the contributions of Messrs. M. E. Tomsho,
B. H. Ruterbories and G. J. Smith of Exxon Engineering - Technology Depart-
ment of Exxon Research and Engineering Co. for their contributions to the
engineering assessment reported here. We also acknowledge with thanks the
boiler manufacturers, Babcock and Wilcox, Combustion Engineering, Inc.,
Foster Wheeler Corp., and Riley Stoker Corp., for providing technical infor-
mation required on their boilers for the assessment of the applicability of
Thermal DeNO to different boiler types.
x
235
-------
INTRODUCTION
This paper presents a summary of an engineering study sponsored by the
U.S. Environmental Protection Agency to project the performance and estimate
the cost of the Exxon Thermal DeNO Process on selected, representative coal
X
fired utility boilers (1). The non-catalytic Thermal DeNO Process is based
X
on the selective reduction of NO with NEL in the eas phase. Thermal DeNO
x 3 x
has been commercially demonstrated on gas-and oil-fired boilers and process
furnaces. A pilot scale test on a coal fired combustor produced NO reduc-
X
tions similar to those obtained with oil and gas firing (1,2).
The following sections of this paper provide information on the chem-
istry and engineering considerations which impact on the applicability of
the Thermal DeNO Process, the methodology used in undertaking the assess-
ment, the projections and results of the study, as well as recommendations
for future work.
236
-------
CHEMISTRY AND ENGINEERING CONSIDERATIONS
The Thermal DeNO Process relies on the selective reaction between
x
NH« and NO to produce nitrogen and water (3,4). The reaction requires the
j X
presence of oxygen and proceeds within a critical temperature range. The
overall NO reduction and production reactions are summarized in equations
1) and 2), respectively;
NO + NH3 + 1/4 02 -> N2 + 3/2 H20 1)
NH3 + 5/4 02 -> NO + 3/2 H20 2)
In typical flue gas environments, the NO __ reduction shown as equation
A
1) dominates at temperatures around 950*0 (174Q°C). At Mgher tempera-
tures, the NO production reaction shown as equation 2) becomes significant,
3C
and it dominates at temperatures over about 1000°C (1830°F). As tempera-
tures are reduced below about 900°C (1650°F), the rates of both reactions
slow down, and the ammonia flows through unreacted*
In addition to temperature, the process is also sensitive to initial
NO and NH concentrations. The NH, injection rate is generally expressed
X ~J 3
as a mole ratio relative to the initial NO concentration. Other variables
X
affecting performance are excess oxygen and available residence time at the
reaction temperature.
Exxon Research and Engineering Company (ER&E) has found that hydrogen
can be used to shift the DeNOx temperature window to lower levels. The
magnitude of this shift is mainly a function of the amount of Ho injected
relative to the NHo. At H2/NH3 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.
237
-------
Thermal DeNO may be applied to boilers for additional NO reduction
X X
after combustion modifications such as low excess air firing, the use of low
NO burners, or overfire air ports have been implemented. As Thermal DeNO
x x
is a post-flame injection process, it is not affected by certain limitations
such as derating that may affect the usefulness of CM in retrofit applica-
tions. The Thermal DeNO process is viewed as an effective supplement to
x
available combustion modification techniques for attaining low NO levels for
X
combustion installations that require a high degree of emission control.
Although the temperature sensitivity will cause the reaction's effectiveness
to vary from one installation to another, the NO reduction is essentially
X
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 boiler design, operating mode, and initial NO level.
X
Thus, each boiler to which Thermal DeNO is applied must be considered on an
X
individual basis.
When applying the Thermal DeNO process to commercial equipment, per-
X
formance is generally limited by the extreme temperature sensitivity of the
reaction and its dependence on the local concentrations of reactants, NH ,
NO , 0 , and H . The Exxon technology provides a means of adapting the
chemistry requirements to industrial equipment environments, and maximum
practical reductions ranging from 35 to 65% have been achieved by the use of
Thermal DeNO in retrofit boiler installations.
x
The Thermal DeNO process utilizes proprietary ER&E gas phase mixing
X
technology to rapidly and efficiently mix the small volume of reagents with
the hot flue gas. Correct distribution of reactants is required in order to
minimize the breakthrough of unreacted ammonia.
Accommodating flue gas temperature variations is important if high DeNO
X
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-
uniformities in flow and heat transfer. It follows, therefore, that case-by-
case evaluations of flue gas temperatures and local conditions are required
for the application of Thermal DeNO to each installation considered.
X
238
-------
In initial applications, ammonia was injected only into boiler cavities,
(boiler regions between tube banks) which can be considered to be isothermal
to a first approximation. Subsequent experimentation by ER&E has shown the
feasibility of injecting ammonia into boiler tube bank regions as well.
Thus, satisfactory NO reduction performance could be obtained by locating
X
the injector grid either in the boiler cavity or in tube banks. The ability
to inject ammonia at post-combustion boiler locations with temperatures
ranging from 760°C to 1000°C has substantially increased the flexibility of
the Exxon Thermal DeNO process.
x
The temperature in the post-combustion convective zone of a boiler is
shifted by changes in boiler load. For example, as load is reduced from full
to 50%, the temperature for optimum Thermal DeNO will shift toward the fire-
X
box. Depending on the magnitude of this shift, more than one ammonia injec-
tion grid may be required in order to obtain DeNO coverage over the range of
X
practical boiler loads. Thus, one grid may be adequate for boiler loads be-
tween 100% and 70%, while another would cover the 70% to 50% load range. It
should be noted, however, that in some cases the use of hydrogen with its
ability to lower the DeNO temperature window could permit effective DeNO
X X
over practical boiler loads with only one grid. In other cases, both hydro-
gen addition and the use of multiple grids may be required to accommodate
load changes. A specific case was considered in which the higher reagent
costs of operating a single grid with ammonia and hydrogen versus those of
ammonia only with two grids (and higher onsite costs) were compared.
The issue of possible by-products associated with the Thermal DeNO
X
reaction has been the subject of extensive study (1, 2, 4) and should be
further considered in undertaking a full scale test of the Thermal DeNO
X
process on a coal-fired utility boiler. For example, the formation, depo-
sition and emissions of ammonium bisulfate should be determined. Also, a
comprehensive emissions measurement program should be performed. The
presence of ammonium bisulfate and/or ammonia can influence the performance
of electrostatic precipitators. Both beneficial and detrimental effects have
been reported. Only through a full scale test of Thermal DeNO on a coal-
X
fired power plant can its effect on ESP performance be quantified.
239
-------
Another factor which should be considered is slagging and fouling
deposits associated with the firing of coal. These deposits can change the
heat transfer characteristics of the boiler and thereby influence the temper-
ature at any given moment in the Thermal DeNO reaction zone. This is an
X
important consideration affecting the design of a Thermal DeNO system to be
X
installed in a coal-fired boiler. The Thermal DeNO system elements must be
x
made compatible with the coal as well as with soot blowing equipment and
procedures.
240
-------
PROGRAM METHODOLOGY
In undertaking this engineering study, eight typical coal-
fired utility boilers representative of current coal fired utility
boiler systems were selected. The boilers were chosen to permit an
evaluation of the Thermal DeNO Process on different utility boiler sizes,
X
firing methods and with different coal types (see Table I). Thermal DeNOx
performance and process costs were determined for two NOX reduction targets:
a. Trimming NO emissions to meet the proposed New Source
X
Performance Standards (NSPS) of 0.6 Ib. NO /MBtu (450
X
ppm NO ) for bituminous coal and lignite fired boilers
A
and 0.5 Ib. NO /MBtu (375 ppm NO ) for boilers fired
X X
with subbituminous coal.
b. Deep reduction in NO levels to 0.4 Ib. /MBtu (300 ppm
X
NO ) for boilers fired with bituminous coal and lignite
Jt
and 0.3 Ib. NO /MBtu (225 ppm) for subbituminous coal
Jt
fired boilers.
Also considered was the:
c. Maximum practical reduction in NO levels which could
X
be realized by the application of the Exxon Thermal DeNO
X
Process. Ratio of NH,. to initial NO level 1.5 to 1 was assumed.
•J X
Two initial NO levels were considered for each of the above NO
x x
targets: (i) uncontrolled and (ii) reduced by combustion modifications.
A total of six cases were thus established from the combination of
the two initial NOX levels and three target NOX levels. These cases
are:
Case 1. Uncontrolled baseline NO ; NSPS target
X
Case 2. Initial NO level reduced by CM ; NSPS target
X
Case 3. Uncontrolled baseline NO ; deep reduction target
X
241
-------
Case 4. Initial NO level reduced by CM; deep reduction target
X
Case 5. Uncontrolled baseline NO ; maximum NO reduction
X X
Case 6. Initial NO level reduced by CM, maximum NO reduction
X •"•
In addition to the six cases noted above, special sensitivity analyses
were performed for flue gas temperature nonuniformity and the use of
hydrogen with ammonia to permit load following. Also, the Thermal DeNOx
costs for reaching NO levels of 0.3 to 0.4 Ib/MBtu were compared with
X
the costs of combustion modifications required to reach these levels.
The selection of the locations where ammonia is to be injected is
based on a number of factors which include: flue gas temperature and
conditions, flow path geometry, the reaction time available and the
suitability of the reaction zone with respect to its dimensions and
configuration of the injector grid. A Performance Prediction Procedure
developed by Exxon Research and Engineering Company was used to determine
the locations of the ammonia injection grids.
The Exxon Performance Prediction Procedure is a multistep calculation
procedure which utilizes and/or determines the above noted factors. The
calculation procedure can forecast the percent reduction in initial NO
X
level which would result from the location of an ammonia injection grid at
any number of locations along the flue gas path. For making these projec-^-
tions of Thermal DeNOx performance for coal fired utility boilers, the
required temperature profile, dimensional and other informaiton were supplied
by the four U.S. utility boiler manufacturers, Babcock and Wilcox, Foster
Wheeler, Combustion Engineering and Riley Stoker. In the case of an actual
installation, after the tentative selection of the location(s) of one or
more grids using the Exxon Performance Prediction Procedure, an experimental
program would be conducted to measure temperature, flow and concentration
distributions in the reaction zone. This information would then be used to
confirm or adjust as required the injector location selected and would be
utilized as input for the final injector design. For this study, two
grid locations were selected for each boiler. The grid locations were
optimized for 100%, 75% and 50% load.
242
-------
The cost estimates prepared in this study were designed to illustrate
the costs associated with the Thermal DeNOx process itself. The techniques
and procedures used in producing these estimates were the same as would be
applied to a more completely defined project. In the cases considered here
the projects were not completely specified with respect to a number of factors
which could have an impact on costs. In addition, certain costs which are
associated with the commercial application of the process such as licensing
fees and preliminary engineering and testing costs were not included.
The cost estimates were formulated for the second quarter of 1977 and
were assigned a U.S. Gulf Coast location. Where practical, costs were
determined for material and labor, and were adjusted to include field
labor, overhead and burden.
DeNOx performance was projected for all boilers at full load and one
or more reduced loads. Costs were projected for full load only.
243
-------
PROJECTIONS AND CONCLUSIONS
The performance of the Exxon Thermal DeNO Process was projected to be
X
essentially equivalent for all eight boiler types evaluated, even though
significant differences existed in flue gas temperature profiles and flow
path configurations among boilers. These differences resulted in the selec-
tion of significantly different injection grid locations among the boilers
of different manufacturers. The analysis determined that the proposed NSPS
of 0.5 Ib./MBtu for subbituminous coal and 0.6 Ib./MBtu for lignite and
bituminous coal could be met by all boilers considered using the Thermal
DeNOT Process. All boilers studied, except the cyclone boiler fired with
5i.
Lignite could meet the deep reduction targets of 0.3 and 0.4 lh/MBTU using
Thermal DeNOx coupled with available combustion modifications. Table II
summarizes the initial and target NOX levels and projected costs for Cases 1
through 4, while Table III presents predicted NOX levels and costs for
Cases 5 and 6.
It was projected that the ammonia injection grid location would not be
effected significantly by assuming a 50°C larger temperature range in the
injection plane than that used for baseline calculations. However, a
temperature range increase of this size would reduce DeNOx performance by
5 to 10 percentage points (e.g., from 50% to 40-45%). It was also found
that overall NOX removal costs increased when hydrogen (rather than multiple
grids) was used with only one grid to achieve effective DeNOx performance at
less than full boiler loads.
Other specific projections and conlcusion were as follows:
• All units could reach the proposed NOX NSPS using Thermal DeNOx alone.
Five units could also reach this level using combusiton modifications
alone.
244
-------
• All units except one could meet the deep NO reduction target
X
when Thermal DeNO was used in combination with combustion
X
modifications. The one exception was the cyclone boiler
fired with lignite.
• Projected costs to reach the proposed NSPS from an uncontrolled
base level ranged from 0.25 mills/KW-Hr for the 250 MW CE boiler
to a high of 1.17 mills/KW-Hr for the lignite fired cyclone
boiler. The average cost for all boilers considered was
0.57 mills/KW-Hr, or 0.49 mills/KW-Hr not including the cyclone
boiler.
• Four of the eight boilers could reach the deep reduction
target using Thermal DeNO alone. Costs ranged from 0.38
X
mills/KW-Hr to 0=83 mills/KW-Hr for these boilers.
• Projected costs to reach the deep reduction target using Thermal
DeNO with combustion modifications ranged from 0.38 mills/KW-hr
x
to 0.51 mills/KW-Hr, with the average being 0.44 mills/KW-Hr for
the seven boilers reaching the target level.
a NO reductions ranging from 62% to 76% and averaging 70%
X
relative to an uncontrolled base case could be achieved using
Thermal DeNO at a maximum practical level in combination with
combustion modifications. NO levels in the 0.20 to 0.23 Ib.
x
NO /MBtu range could be realized for most of the boilers. Costs
X
ranged from 0.55 to 1.14 mills/KW-Hr and averaged 0.68 mills/KW-Hr
for all boilers studied. With the lignite boiler excluded» the
range was 0.55 to 0.67 mills/KW-Hr and the average was 0.61
mills/KW-Hr.
e The costs for onsites and the carrier were found to be
proportional to boiler size.
• The total ammonia reagent costs for all cases? normalized for
the amount of NOX removed expressed as N02 (ANOX)» were nearly
equal for all eight units studied at $0.09/lb. ANOX* This
245
-------
parameter was considered to be a good overall judgment criterion
of the chemical efficiency and economic efficacy of the Thermal
DeNO process.
The Exxon Thermal DeNO Process was considered to be equally
X
amenable to all units studied.
The costs for reaching NO levels in the 0.3 to 0.4 Ib./MBtu
X
range were compared for Thermal DeNO and combustion modifications.
The costs of most conventional combustion modifications and combi-
nations thereof were lower than that of Thermal DeNO . Extreme NO
x x
reduction methods such as derating or staged combustion that
incurred derating would be more expensive.
246
-------
RECOMMENDATIONS
The primary recommendation resulting from this engineering assessment
is that a demonstration of the Exxon Thermal DeNOx Process on a coal fired
utility boiler is required to test the process effectiveness and determine
potential side-effects under realistic, full scale conditions. After the
selection of a suitable candidate boiler, a detailed performance and cost
analysis must be prepared. Temperature and velocity profile measurements
in the boiler heat transfer region will be required to verify grid placement
and performance estimates. After installation and startup of Thermal DeNOx
system, a careful measurement and evaluation program will be needed to assess
DeNO performance, costs, and potential side effects that may result from the
X
use of the Thermal DeNOx process.
In undertaking this demonstration, a high level of attention should be
accorded to those factors which could reduce the overall effectiveness of the
Thermal DeNOx Process on coal fired utility boilers, or could have adverse
side effects on boiler operation or the environment. These factors include:
• Ammonia and by-product emissions
• Effect of slag formation and fouling on DeNOx reaction zone temper-
atures, and on resulting DeNOx performance.
$ Effective DeNOx performance under differing boiler load conditions.
» Effect of deposition of ammonium sulfates on metal surfaces and on
electrostatic precipitator performance.
• Compatibility of Thermal DeNOx system elements with coal ash levels
and with soot blowing equipment and procedures.
247
-------
REFERENCES
1. Varga, G. M., Jr., M. E. Tomsho, B. H. Ruterbories, G. J. Smith and
W. Bartok, Assessment of Exxon Thermal DeNOx Process Applied to Coal
Fired Utility Biolers, U.S. Environmental Protection Agency, Research
Triangle Park, N.C.
2. Muzio, L. J., J. K. Arand.and K. L. Maloney, Noncatalytic NOX
Removal with Ammonia. EPRI Report FP-735, Electric Power Research
Institute, Palo Alto, California, April, 1978.
3. Lyon, R. K., Communication to the Editor: The NH3-NO-C-2 Reaction,
International Journal of Chemical Kinetics, J3, 315-318 (1976).
4. Lyon, R. K. and J. P. Longwell, Selective, Non-Catalytic Reduction of
NOX by NH3, Paper presented at EPRI NOX Seminar, San Francisco,
February 1976.
248
-------
TABLE I. BOILERS SELECTED FOR STUDY
Boiler
'Manufacturer
Firing Method
Boiler
Size, MW
Coal Type
Babcock and
Wilcox
Babcock and
Wilcox
Babcock and
Wilcox
Combustion
Engineering
Combustion
Engineering
Foster Wheeler
Foster Wheeler
Riley Stoker
Front Wall
Horizontally
Opposed
Cyclone
Tangential
Single Furnace
Tangential
Front Wall
Horizontally
Opposed
Turbo Furnace
130
333
400
350
800
330
670
350
Subbi tuminous
Bituminous
Lignite
Bituminous
Subbituminous
Bituminous
Subbituminous
Bituminous
249
-------
TABLE II. TARGET CASES FOR REDUCING NO EMISSIONS OF COAL FIRED UTILITY BOILERS USING THERMAL DeNO
x
Case 1
U/0 Comb. Mod.
ro
on
O
Size,
Manuf. MW
B&W 130
*
333
400
CE 350
800
F-W 330
670
Firing
Type
FW
HO
Cyclone
Tan
Tan
FW
HO
Coal
Subbit.
Bit.
Lig.
Bit.
Subbit.
Bit.
Subbit.
Initial
NO^. ppm
500
700
1000
500
530
850
700
Cost
Mills /KW-Hr
0.49
0.45
1.17
0.25
0.34
0.71
0.63
Case 2
With Comb. Hod.
Initial
NO^. ppm
300
420
900
450
375
510
420
Cost
Mills /KW-Hr
0.06*
0.06*
0.99
0.08*
0.08*
0.33
0.31
Final
NOs. ppm
375
450
450
450
375
450
375
Case 3
U/0 Comb. Mod.
Initial
NO . ppm
500
700
1000
500
530
850
700
Cost
Mills/KW-Hr
**
0.63
**
0.38
0.62
**
**
Case 4
With Comb. Mod.
Initial
NO ppm
300
420
900
450
375
510
420
Cost
Mills/KW-Hr
0.47
0.38
**
0.42
0.42
0.50
0.51
Final
NO^ ppm
225
300
300
300
225
300
225
Percent
Reduction
From
Uncontrolld
Case
55
57
—
40
58
65
61
RS 350 Turbo. Bit. 700 0.54
420
0.10* 450 700 0.83 420 0.45
300
57
Thermal DeNO not required. Final NO level attainable using combustion modifications.
Target NO level cannot be reached.
-------
en
TABLE III.
MAXIMUM PRACTICAL NOV REDUCTION PREDICTED USING THERMAL DeNOv
A _ . A
Case 5
Without Combustion Modifications
Case 6
Vttth Combustion Modifications
Boiler
Size,
Manuf. MW
B&W 130
333
400
CB 350
800
F-U 330
670
Firing
Type
FW
HO
Cyclone
Tan
Tan
FW
HO
Coal
Subbit.
Bit.
Lignite
Bit.
Subbit.
Bit.
Subbit.
Initial
Level,
PPm
500
700
1000
500
530
850
700
NO
Final"
Level,
250
290
430
210
230
390
290
Percent
NO
Reduction
50
59
57
58
57
54
59
Total
Cost,
Mills /KW-Hr
0.71
0.70
1.23
0.57
0.62
0.82
0.87
Initial
Level,
300
420
900
450
375
510
420
NO
Final*
Level,
PPm
150
175
385
190
160
230
170
Percent
NO
Reduction
50
58
57
58
57
55
59
Total
Cost,
Mllls/KW-Hr
0.60
0.55
1.14
0.61
0.59
0.62
0.65
Percent
Total NO Reduction
PosslSle From
Uncontrolled Case
70
75
62
62
70
73
76
RS
350 Turbo Bit.
700 295
58
0.84
420 175
58
0.67
75
-------
COMBUSTION MODIFICATION CONCEPTS
FOR STOKER BOILER APPLICATIONS
By:
John H. Wasser
Combustion Research Branch
Energy Assessment and Control Division
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
253
-------
ABSTRACT
This contract program will apply combustion modification concepts inves-
tigated in prior government and industry programs to full-scale coal-fired
stoker boilers. In Phase I, based on reported field test results and R&D
information on combustion modification integrated with a perspective of the
stoker boiler spectrum in the United States, a contractor will select and
negotiate for the utilization of two representative large industrial spreader
stoker systems. Representative fuels (including cleaned or processed fuels)
will be selected and used in the study. A comprehensive Operation Plan will
be prepared to bring together all elements of the program in a study that
will address the problems involved in applying new technology to stoker
boilers and include design of the hardware for applying the combustion modi-
fication concepts to the specific units selected. While the plan is reviewed
for EPA approval, the contractor will conduct the baseline study of the
boiler systems to establish the unmodified performance characteristics.
After approval of the Operation Plan, the contractor will construct the
modifications to the system and proceed with the comprehensive study of
the modified system.
The program will continue with analysis of the data and preparation of
a thorough evaluation of the environmental aspects of the applied modification
concepts. Subsequent to the analysis and evaluation, a document will be
prepared in cooperation with boiler owners and manufacturers that will
promote application of the results of this program to new design and retrofit
installations.
In Phase II, the contractor will apply combustion modifications to
other types of stoker systems. Boilers representative of commercial moving
grate and underfeed stokers (about three units) will be selected for the
program. This phase will consist of the same series of tasks as the large
spreader stoker phase, progressing from selection of the boilers and fuels,
through planning, design, baseline determination, modification construction,
modification testing, comprehensive analysis and evaluation, to guideline
document preparation.
An industrial review panel will be established to provide practical
guidance for the program and insure that the results would have the maximum
benefit in the shortest possible time for stoker boiler manufacturers and
owners.
254
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-050a
2.
3. RECIPIENT'S ACCESSION-NO.
proceedings of the Third Stationary
Source Combustion Symposium; Volume I. Utility, In-
dustrial, Commercial, and Residential Systems
5. REPORT DATE
February 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Joshua S. Bowen, Symposium Chairman, and
Robert E. Hall, Symposium Vice-chairman
B. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12.
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings: 3/79
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES JERL-RTP project officer is Robert E. Hall. MD-65, 919/541-
2477. EPA-600/7-77-073a thru -073e and EPA-600/2-76-152a thru -152c are pro-
ceedings of earlier symposiums on the same theme.
. ABSTRACT
proceedings document the approximately 50 presentations made during
the symposium, March 5-8, 1979, in San Francisco. Sponsored by the Combustion
Research Branch of EPA's Industrial Environmental Research Labor atory-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 in seven parts , and
the proceedings are in five volumes: I. Utility, industrial, Commercial, and Resi-
dential Systems; TJ. Advanced Processes and Special Topics; HI. Stationary Engine
and Industrial Process Combustion Systems; TV. Fundamental Combustion Research
and Environmental Assessment; and V. Addendum. The symposium provided contra-
ctor, industrial, and government representatives with the latest information on EPA
inhouse and contract combustion research projects relating to pollution control,
with emphasis on reducing NOx while controlling other emissions and improving
efficiency.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Air Pollution
Combustion
Field Tests
Assessments
Combustion Control
Fossil Fuels
Boilers
Gas Turbines
Nitrogen Oxides
Efficiency
Utilities
Industrial Pro-
cesses
Hydrocarbons
Air Pollution Control
Stationary Sources
Environmental Assess-
ment
Combustion Modification
Trace Species
Fuel Nitrogen
2 IB
14B
2 ID
ISA
isir
07B
07C
8. DISTRIBUTION STATEMEN1
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
258
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
255
------- |