EPA-600/7-81-090
Evaluation of Emissions and
Control Technology for Industrial
Stoker Boilers - ^
Battelle
Columbus Laboratories
-------
EVALUATION OF EMISSIONS AND CONTROL
TECHNOLOGY FOR INDUSTRIAL STOKER BOILERS
by
Robert D. Giammar, Russell H. Barnes, David R. Hopper,
Paul R. Webb, and Albert E. Waller
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
Contract No. 68-02-2627
March, 1981
EPA Project Officer: John H. Wasser
U.S. ENVIRONMENTAL PROTECTION AGENCY
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
-------
FOREWORD
The coal-fired stoker boiler provides an option for industry
to meet its 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 coal-fired stoker
boiler.
In support of our nation's commitments to maintain a clean
environment and to utilize coal, EPA 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 were to
a Characterize the spectrum of emissions
from industrial coal-fired stoker boilers
using several types of coal under various
stoker-firing conditions
Investigate control methods to reduce these
emissions
Determine the effect of these control methods
and variations in stoker-boiler operation on
the overall performance of the stoker boiler,
and,
0 Assess the environmental impact of new
technology on the future acceptability of
stoker boilers.
This program was divided into three phases. In Phase I,
Alternative Fuels Evaluation, emission characteristics were determined
for a variety of coals fired in a 200-kW stoker boiler. Emphasis was
focused on identifying coals with low pollutant potential, including
both physically and chemically treated coals. In Phase II, Control
-------
Technology Evaluation, potential concepts for control of emissions for
full-scale industrial stokers were evaluated. In Phase III, Limestone/
Coal Pellet Development, a limestone/coal fuel pellet was developed and
evaluated as a viable SO control for industrial stoker boilers.
This report presents the results of the three phases of work.
The report is organized into the following three parts corresponding to
the three phases of work:
Phase I. Alternative Fuels Evaluation
Phase II. Control Technology Evaluation
Phase III. Limestone/Coal Pellet Development
These parts actually represent separate reports but are included under one
cover.
-------
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 wish to express
their appreciation for the assistance and direction given the
program by project monitor John H. Wasser.
We would also like to acknowledge Harold Johnson of
Detroit Stoker, William Engelleitner of Mars-Mineral, Sam Spector
of Banner Industries, and Donald Hansen of Alley-Cassetty Coal
Company for providing advice and assistance to the program.
Finally, we would like to recognize Battelle-Columbus
staff membersJohn Faught, Tom Lyons, Paul Strup, Don Hupp,
Luis Kahn, and Andrew Skidmore, and acknowledge the cooperation of
John Clayton and his Facilities staff who allowed us to use the
Battelle steam plant boiler during the program.
-------
PHASE I. ALTERNATIVE FUELS EVALUATION
-------
PHASE I
CONTENTS
FIGURES ill
TABLES . » . iv
EXECUTIVE SUMMARY v
I. BACKGROUND 1-1
II. OBJECTIVE AND SCOPE 1-2
III. EXPERIMENTAL PLAN 1-3
Underfeed Stoker 1-6
Model Spreader 1-6
Coal Selection and Preparation 1-7
Coal Selection 1-7
Coal Preparation 1-8
Coal Analyses 1-9
Experimental Procedures 1-9
Model Spreader Simulation 1-9
Sampling and Analytical Procedures 1-10
Coal Sampling 1-10
IV. EXPERIMENTAL RESULTS 1-13
Underfeed Stoker Results 1-13
Gaseous Emissions 1-15
Particulate Loadings 1-16
POM Loadings 1-16
Model Spreader Stoker Results 1-18
Gaseous Emissions 1-18
Particulate Loadings 1-20
POM Loadings 1-20
Particle Size Distribution 1-22
Gas Probing Observations 1-22
Bed Temperature Profiles 1-24
Observations of Suspension and Fixed Bed Combustion 1-28
-------
PHASE I
CONTENTS (Continued)
Page
V. ALTERNATIVE FUELS FOR STOKERS--PELLETIZATION APPLICATIONS .. 1-33
VI. CONCLUSIONS 1-34
REFERENCES 1-35
APPENDIX A - BINDER IDENTIFICATION AND RESTRUCTURING TECHNIQUE
REVIEW
ii
-------
PHASE I
LIST OF FIGURES
Page
Figure 1-1. 200 kW Stoker-Boiler Facility 1-4
Figure 1-2. Spreader Feed Mechanism 1-5
Figure 1-3. Particle Size Distrubution of Fly Ash for Runs a) MS-
72, b) MS-73, c) MS-5 and d) MS-103 1-23
Figure 1-4. Bed Temperature Profiles for a High Volatile Coal . . 1-26
Figure 1-5. Bed Temperature Profiles for a Low Volatile Coal. . . 1-27
Figure 1-6. Fuel Bed Isotherms for a 27 Percent Volatile Austra-
lian Coal 1-29
iii
-------
PHASE I
LIST OF TABLES
Page
Table 1-1. Analysis of Selected Untreated Coals 1-10
Table 1-2. Analysis of Selected Treated Coals I-ll
Table 1-3. Summary of Emissions for Selected Coals Fired in The
Underfeed and Modal Spreader Stokers 1-14
Table 1-4. POM Quantification for Underfeed Stoker Experiments. . 1-17
Table 1-5. POM Quantification for Model Stoker Experiments. . . . 1-21
Table 1-6. Bed Temperatures 1-25
Table 1-7. Proximate and Ultimate Analyses (In Weight Percent) of
4 Coal Samples Obtained During Combustion. ...... 1-31
iv
-------
EXECUTIVE SUMMARY
A 200-kW stoker-boiler facility was used to evaluate
characteristics of emissions from combustion of a variety of coals,
including coals that could not be conveniently or economically
evaluated in larger industrial systems. The stoker was initially
operated in an underfeed mode to expand the data base developed in
an earlier EPA program . This facility was modified to accommodate
a model spreader stoker more typical of an industrial boiler.
Raw coals with low pollution potential and treated coals were
evaluated. Because there was only one treated coal available during the
time framework of the program, Battelle developed, as part of this pro-
gram, a limestone/high sulfur coal fuel pellet.
Results of the Phase I emission characterization were as
follows .
NO
For the underfeed stoker, less than 10 percent of the fuel
nitrogen was converted to NO, assuming no thermal NO. For the model
spreader-stoker, between 10 and 20 percent of the fuel nitrogen was con-
verted to NO.
Coals naturally high in calcium and sodium and those treated
with these elements retained significant percentages of the sulfur
in the ash. For the eastern bituminous coals, with relatively small
amounts of calcium and sodium but significant amounts of iron, sulfur
retention in the ash was as high as 20 percent. Note that bed tempera
tures in these laboratory stokers are significantly lower than these
measured in an industrial stoker.
v
-------
CO
CO levels can be controlled by the use of overfire air and
were generally less than 100 ppm.
Particulate Loading
Particulate loadings did not correlate consistently with either
the ash content of the coal nor its size consist prior to feeding. It
appears that the friability and inherent moisture content of the coal may
affect particulate loading since these properties influence the amount of
fines generated.
POM Loadings
POM loadings for continuous operation of the underfeed stoker
were significantly less than those reported earlier ~ for intermittent
operation.
Particle-Size Distribution
For the model spreader, the average stack particle size ranged
between 15 and 30 micrometers.
Treated Coals
No commercially available, chemically treated coals were identified.
Treated coals required pelletization for firing in stokers.
The Battelle Hydrothermally Treated (HTT) coal was available for
laboratory evaluation. The treatment reduced the fuel sulfur from 2.6
percent to 1.1 percent. Because of the relatively high calcium and sodium
residual from the treatment, only 28 percent of the remaining sulfur was
emitted as SO..
Also, the limestone/coal fuel pellet, with a Ca/S molar ratio of
7, reduced SO,, emissions by over 70 percent. Even at the elevated fuel-bed
temperatures (> 1100 C), the calcium reacts with the coal sulfur and retains
it as a sulfide/sulfate as part of the fuel ash.
vi
-------
PHASE I ALTERNATIVE FUELS EVALUATION
SECTION I
BACKGROUND
The overall performance of a stoker-boiler system and the
emissions that it generates depend on several factors, including stoker-
boiler system design, combustion operating parameters, and coal properties.
Generally, for a given stoker-boiler system, the overall boiler performance
and combustion-generated emissions can be controlled to a limited extent by
optimizing stoker-boiler operating parameters. In contrast, the type of
coal fired in system has the most dominant effect on both emission levels
and system performance.
Unfortunately, coals in major U.S. reserves located near the
industrial sector contain relatively high levels of sulfur. Firing these
coals in an environmentally acceptable manner, requires either pretreatment
to remove the sulfur or burning in systems with flue-gas desulfurization
(FGD) equipment. Industrial boiler owners/operators have resisted FGD
because of the high cost and low reliability. As a result, there has been
increased interest in coal treatment processes. In addition, as an alter-
native to high-sulfur eastern bituminous coals, consideration has been
given to increased industrial utilization of low-sulfur coals such as lignite,
western subbituminous, and western bituminous coals. Because the treated
coals have not been available and the western coals have not been used
extensively, there is relatively little combustion generated emission data
on these fuels to assess their impact on the acceptability of stoker
boilers.
1-1
-------
SECTION II
OBJECTIVE AND SCOPE
The overall objective of the Phase I research program was to
evaluate the emission characteristics resulting from the combustion of
a representative number of raw coals and available treated coals.
Because of the limited availability of treated coals and the logistics
problems and high costs of procuring large quantities of western coals,
the combustion tests were conducted in a 200-kW (20-bhp) stoker-boiler
facility used on an earlier EPA stoker program . This facility was
initially designed and operated with an underfeed stoker. A limited
number of coals were fired in this operational mode primarily to extend
the data base developed from the earlier program. However, because the
spreader stoker is the type preferred by industry, a "model spreader"
was designed to simulate that mode of operation. The majority of experi-
ments were conducted with the model spreader. In addition, as part of
Phase I activities, limited emission measurements were made in the
Battelle (8 MW) (25,000 Ib steam/hr) steam plant spreader stoker for
comparison with those of the model spreader.
As first conceived, this program was to focus primarily on
evaluating the emission characteristics of treated coals. However,
this scope was altered because the various processes to treat coals
did not develop sufficiently within the time framework of the program.
As a consequence, the development of a limestone/high sulfur coal
pellet, conceived by the Battelle staff, was pursued as part of Phase I
activities.
1-2
-------
SECTION III
EXPERIMENTAL PLAN
Phase I experiments focused on assessing emission characteristics
for a variety of coals for both the underfeed and model spreader operations.
The effects of system operating variables for relatively small systems are
difficult to scale to larger systems because wall effects, heat losses, and
particle dynamics are more pronounced. Accordingly, these variables were
of secondary interest. Nine raw coals and two treated coals were evaluated
using particulate loading, CO, S0?, and NO emission levels as the primary
basis for evaluation; a limited number of POM emission measurements were
also made.
FACILITY
Figure 1-1 is a photograph of the stoker research facility used
in the Phase I experiments. The basic element of this facility is a
Kewanee 3R-5 200 kw (20 bhp), fire-tube, hot water boiler capable of firing
coal at rates up to 34 kg/hr. Coal was fed to the boiler by either a
Will Burt 34 kg/hr underfeed stoker or a "model" spreader stoker. The
"model" spreader, shown in Figure 1-2, was designed and constructed by
the Battelle-Columbus staff. This spreader can distribute coal uniformly
over a 0.6-m wide by 1-m deep grate at rates ranging from 13.6 to 68 kg/hr.
The "model" spreader was installed through the firebox door of the boiler,
the normal location of a burner for oil or gas firing. These two types of
stokers provide radically different burning methods suspension plus bed
burner for the spreader and entirely bed burning for the underfeed. They
even have opposite combustion air flow patterns counter flow for the
spreader and parallel flow for the underfeed.
1-3
-------
FIGURE 1-1. 200 KW STOKER-BOILER
FACILITY
1-4
-------
MODEL SPREADER STOKER
VARIABLE SPEED
TRANSMISSION
H
HOPPER
VIBRATORY
FEED
r
GRATE
FIGURE 1-2. SPREADER FEED MECHANISM
-------
These units have fixed beds with no provision for automatic ash
removal. The combustion zone is surrounded by the water wall of the
boiler.
Combustion gases were vented through a 0.35-m diameter, insulated
stack that included provision for sampling at a location 10-stack diameters
downstream of the boiler outlet. A damper (approximately 1.5 m above the
sampling port) provided a control at the boiler outlet. Ports for smoke
and gaseous-emission sampling and for temperature and pressure measurements
were provided at the base of the stck.
Underfeed Stoker
The Will-Burt stoker is a conventional bituminous underfeed stoker.
The majority of coal is burned in the retort with the remaining coal being
burned on a ceramic hearth surrounding the retort. Normally, the primary
combustion air would be supplied by a fan; however, to provide for a greater
range of flow rates, laboratory compressed air was used for these experi-
ments. The flow rate was measured with a standard ASME orifice.
Four overfire air jets, 150 mm on centers and 7 mm in diameter,
were installed approximately 300 mm above the retort. The overfire or
secondary air was measured with a laminar flow element.
Mod el Spreader
This spreader mechanism introduced coal into the boiler through
the fire door. The spreader mechanism was installed on the boiler in
much the same way as a gas or oil burner. Seals were used to minimize
leakage. Pitch of the four rotor baldes on this spreader mechanism was
varied by cold flow tests to provide uniform fuel distribution. However,
there was still some tendency for the coal to build up slightly in the
center of the grate. Coal was fed to the rotors from the coal hopper by
a vibrating feeder. The flow rate was controlled by the vibration of the
feeder.
The stationary grate was designed with a free flow area of 4
percent to provide for a uniform distribution of the air through 6-mm holes.
1-6
-------
To compensate for the slight build up of coal in the center of the grate,
the 6-mm holes along the outer edge were plugged.
The water-cooled walls of the boiler firebox were lined with
high-temperature (1400 C) blanket-type insulation. This insulation reduced
the heat loss the from the burning zone and improved combustion in the suspension
zone above the bed.
Sampling ports located 25 mm, 100 mm, and 200 mm above the grate
were used to measured bed temperature. Bed temperature was extremely
sensitive to the bed structure. Accordingly, procedures were developed
and coal feeding apparatus was modified to establish and maintain a
reasonably uniform fuel bed. This was accomplished with continuous scrutiny
and adjustment of the feed mechanism.
COAL SELECTION AND PREPARATION
Coals considered for this program were categorized as
Untreated naturally occurring raw coals
Reconstituted physically treated/clean coals
that require restructuring
Processed chemically treated coals that may
require restructuring.
Selection was based primarily on availability and the low pollution potential
of the coal. Coal preparation ranged from crushing and screening the
untreated coals to pelletizing the treated coals.
Coal Selection
The untreated coals were to be representative stoker fuels and/or
have low pollution potential based upon chemical analyses. The reconsti-
tuted and processed coals were to be coals treated to obtain the desired
properties of low emission fuels. Additionally, the availability and
cost of procuring and preparing the coal was considered.
The untreated coals were commercially available. The costs of
these coals at the mine were nominal and the transportation costs were
1-7
-------
reasonable for the 5 ton lots. In the case of the processed coals, only
one (Battelle Hydrothermally Treated) was available for this program.
This coal also required pelletization for stoker firing. Likewise, the
reconstituted coals or coals that were physically treated to "deep clean"
were not available. Because of the unavailability of treated coals,
Battelle developed a reconstituted fuel by physically mixing finely ground
limestone with pulverized high-sulfur coal. This limes tone/coal mixture
was restructured by pelletization producing a pellet approximately 12 mm
in diameter by 18-mm long that was suitable for stoker firing.
Coal Preparation
For reasons discussed later, the untreated coals were crushed
and screened to a top size of 15 mm, with about 20 percent fines (less
than 5 mm but greater than 1-mm diameter). The reconstituted fuels and
processed fuels required restructuring.
Each of the initial fuels evaluation experiments required 100
kg to 150 kg of fuel pellets. To prepare this quantity of pellets,
Battelle used a California Laboratory Pellet Mill capable of producing
about 5 kg of pellets per hour. For the limestone/coal pellets, the base
coal, Illinois No. 6, was pulverized to 100 percent through -20 mesh and
50 percent through -100 mesh with a hammer mill. A -50 mesh limestone
(>90 percent CaCCO was used. This limestone, the Illinois No. 6 coal,
and appropriate binders were then mixed in either (1) a Hobart mixer for
small batch quantities of pellets for mechanical strength evaluation or
(2) a portable cement mixer for larger quantities for combustion tests.
For the Battelle Hydrothermally Treated Coal, similar procedures were
used except that limestone was not added.
Restructuring of these fuels required a binder that provided
enough mechanical strength that the resultant fuel pellet could be
handled without serious physical deterioration. Thus, an effort was
undertaken to identify and evaluate binders. In addition, different
restructuring techniques were reviewed. These efforts are described in
Appendix I-A.
1-8
-------
Coal Analyses
Tables 1-1 and 1-2 present the ultimate and proximate analyses
and the free-swelling index for the untreated and treated coals fired in
Phase I. The free-swelling index is a. measure of caking tendency of coal
and is applicable to the underfeed stokers rather than the model spreader.
The nitrogen content of the washed Western Kentucky coal appears low at
0.75 percent in comparison to the unwashed western coal at 1.38. Washing
coal does not reduce the nitrogen content (as is the case for the ash and
sulfur content) but rather increases the fuel N percentage because of the
removal of ash.
EXPERIMENTAL PROCEDURES
For the underfeed stoker experiments, procedures previously
(1-1)
developed on the earlier EPA stoker program were used. For the
model spreader experiments, procedures were developed to simulate the
operating characteristics of an industrial spreader stoker. Proven
analytical techniques were used to determine the criteria pollutants.
Model Spreader Simulation
Spreader stokers burn up to 50 percent of the coal in suspension
in the combustion zone above the fuel bed. The remaining coal is burned
in the fuel bed. The amount of coal burned in suspension depends on the
percentage of coal fines and the stoker boiler design (residence time
above the bed). The model spreader was designed to provide for both
suspension and bed burning. To compensate for the decreased residence time
in the freeboard of the model spreader (residence time could not be "scaled"
in the model spreader), the as-received stoker coal was crushed to reduce
the overall particle size distribution to a top size of about 15 mm. The
bottom size could not be reduced substantially as particles less than 0.5 mm
to 1 mm would be transported by the combustion gases to the stack without
reaching the burning regime. Also, it was observed that if the coal feed
contained more than 20 percent fines, the fuel bed tended to mat. Evidently,
these coal fines did not burn to completion in suspension and accumulated
on the fuel bed. Accordingly, the percentage of coal fines was controlled
to about 20 percent of the total feed.
* References on page 1-35 1-9
-------
TABLE 1-1. ANALYSIS OF SELECTED UNTREATED COALS
Ultimate Analysis, %
(dry)
Proximate Analysis, %
(as run)
Fuel Type
SE Kentucky
Illinois No. 6
H Western Ky.
M (unwashed)
o
E . Kentucky
W. Virginia
(Bishop mine)
Western Sub-
bituminous
Lignite
Western Ky.
(washed)
Volatiles
36.
39.
39.
40.
21.
28.
29.
41.
94
31
7
0
4
4
1
76
Fixed
Carbon
54.01
46.90
49.8
53.5
70.8
35.6
28.3
52.6
Moisture
3.89
11.07
6.88
2.35
0.9
26.1
32.6
7.48
Ash C
5.37 79.9
13.79 66.9
10.46 70.4
4.7 78.2
6.9 84.7
9.86 50.6
9.97 62.2
5.62 74.64
H N
5.5 1.53
4.7 1.04
4.6 1.38
5.1 1.3
4.7 (1.5)
3.2 0.8
3,8 0.9
5.41 .75
1.
4.
4.
1.
0.
0.
1.
3.
S
27
94
57
22
61
67
57
53
Oxygen
(differ-
ence)
6.3
8.5
8.5
9.5
1.6
34.8
21.6
10.01
High
Heating
Value
KJ/g
33.3
28.1
29.1
32.8
34.1
24.1
21.9
31.6
Free
Swelling
Index
6-1/2
3-1/2
2
2
7-1/2
Non-
Agglom.
Non-
Agglom.
2-1/2
-------
TABLE 1-2. ANALYSIS OF SELECTED TREATED COALS
Fuel Type
Battelle (Hydro-
thermal ly
Treated
L/C 50/50 (a)
Latex Bound
L/C 50/50
Cement Bound
L/C 50/100
Cement Bound
Proximate Analysis, %
(as run)
Fixed
Volatilea Carbon Moisture
34.17 46.42 7.15
46.01 9.72 7.06
44.9 2.96 12.60
43.9 19.82 6.99
Ultimate Analysis,
(dry)
Ash C HNS
19.41 65.3 4.3 1.17 1.16
44.3 37.6 2.2 0.63 2.21
52.1 34.0 1.8 0.55 2.24
36.3 46.5 2.9 0.78 2.87
%
Oxygen
(differ-
ence)
8.55
13.0
9.3
10.5
High
Heating
Value,
KJ/g
26.5
12.8
11.3
17.4
(a) L/C = limestone/coal, fuel pellets
-------
Spreader stokers have relatively thin, fast-burning beds. In
the model spreader there was no provision for continuous ash disposal
and the fuel bed became progressively deeper. High ash fuels, such as
the limes tone/coal pellet, accentuated this problem. Deep beds generally
resulted in a nonuniform burning of the bed. Accordingly, to provide
some measure of control of the fuel bed when making emission measurements
before emission sampling was began, the stoker was operated long enough
to establish a stable fuel bed, but not so long that ash build-up would
significantly interfere with the stoker performance.
Sampling and Analytical Procedures
Particulate and POM levels were determined by a modified EPA
Method 5 procedure with the probe wash and filter catch being used to
determine the filterable particulate loadings and an absorbent column
(1-2)
being used to determine POM loadings. The extract from the column
was analyzed by GC/MS techniques.
Gaseous emissions were determined by: paramagnetic analysis
for oxygen; nondispersive infrared (NDIR) analysis for carbon monoxide,
carbon dioxide, and nitrogen oxide; and a dry electrochemical analyzer
for sulfur dioxide. Gas samples from the stack first passed through a
heated filter that was coupled close to the stack, and then through an ice
trap to reduce the dewpoint. After this trap, the sample gas was conveyed
through teflon tubing to a manifold and then distributed to each instru-
ment. For the NO measurement, a dry ice trap was located between the
manifold and the NDIR instrument to eliminate all moisture in the stack
gas.
Coal Sampling
Because coal is not a homogeneous fuel, its composition can
vary substantially. An accurate determination of the coal composition
is needed for combustion calculations. In the model spreader experiments,
coal samples were collected from the feed system through the run. These
aggregate samples were then ground and analyzed. In addition to this
procedure, coal was sampled from each drum in which it was stored and
then ground for analysis. Generally, there were some minor discrepancies
that were attributed to the difficulty of obtaining representative samples,
For the combustion calculations, coal analyses were based on those samples
collected during the experiment.
1-12
-------
SECTION 4
EXPERIMENTAL RESULTS
The emissions of a stoker-boiler system depend on its design as
well as its operation and the coal being burned. Accordingly, these
factors must be considered when comparing emission and performance data
from several systems.
Phase I experiments were designed to provide a relative measure
of emissions from a variety of coals, including coals that could not be
conveniently evaluated in larger industrial systems. The majority of the
gaseous emission data correlate primarily with coal properties and to a
lesser extent with system design parameters. As a consequence, these data
would be representative of a variety of industrial units, including under-
feed and spreader stokers. In fact, gaseous emission levels obtained from
firing the same coal in both the model and industrial spreader stoker
were similar. On the other hand, particulate emission levels depend on
system design. Thus, particulate emission data from the laboratory-size
stoker-boiler may not be representative of data for industrial-size units.
UNDERFEED STOKER RESULTS
The underfeed stoker experiments were designed to extend the
data base of the earlier EPA stoker program to cover additional
coal types and operating conditions. In the earlier program, the under-
feed stoker was most often operated under cyclic conditions (on/off mode),
while in this program, the stoker was operated continuously.
Table 1-3 summarizes the results of these underfeed stoker
experiments for the coals fired and the operating conditions listed.
1-13
-------
TABLE 1-3. SUMMARY OF EMISSIONS FOR SELECTED COALS FIRED IN THE
UNDERFEED AND MODEL SPREADER STOKERS
M
I
Run No.
a. llndei
UF-1
UF-2
OF- 3
UF-4
UF-5
b. Model
HS-1
MS-2
MS-3
MS-4
HS-5
HS-6
MS-7
HS-8
MS-9
MS- 10
MS-ll
MS-12
HS-11
HS-1 4
HS-15
MS- 16
MS-17
HS-18
HS-19
MS-20
HS-21
HS-22
MS-23
HS-24
MS- 2 5
MS-26
MS-2 7
HS-28
HS-29
HS-30
HS-11
Coal Type
feed
E.Kentucky Bit.
Lignite
Western Sufabit.
Illinois No. 6
Illinois No. 6 n/llaeutone
Spreader
Lo-vol W. Va.
E Kentucky Blt.,TP Mine
SE Kentucky Bit.
SE Kentucky Bit.
Illlnoli No. 6
Illinois No. 6
Lignite
Western Bit. Col
Western Bit. Col
Pellets, Ca/S - 7, Latex
binder
Pellets, Ca/S - 7, Latex
binder
Pellets, Ca/S - 1.5, Latex
Illinois No. 6
Lo-vol W. Va.
SE Kentucky, Bit.
W. Kentucky, unwashed
W Kentucky, washed
E Kentucky, Bit.
Lignite
Pellets, Ca/S-3.5 Lutex
Illinois No 6
Pellets, Ca/S - 7, Ceuent
binder
Pelletn, Ca/S - 7, Cement
binder
HTT Pellets
Lo-vol W. Vn.
Lignite
SE Kentucky
C Kentucky
Illinois No. 6
Uefltern Bituminous
Pellets, Ca/S - 7.0
Avera
°2
X
7.5
8.5
7.4
6.0
6.0
7.2
6.8
7.5
8.2
8.8
8.6
10.2
9.1
8.8
9.5
10.2
8.8
7.9
6.8
6.8
7.7
8.2
7.6
9.6
10.7
9.4
9.3
9.7
7.5
7.5
7.0
7.0
5.0
8.3
7.5
11.0
ge Flue
Z
12.4
11.2
12.4
13.0
13.2
11.8
12.0
11.0
11.1
10.2
10.2
10.4
11.0
10.5
11.5
11.0
10.5
10.4
12.9
11.5
11.1
10.6
13.4
9.7
10.4
10.9
11.0
11.1
12.0
12.8
17.0
14.8
11.0
12.8
10
Gas C
CO
ppm
40
70
60
80
240
20
24
50
40
30
20
10
10
15
70
50
250
30
20
40
10
50
30
15
24
50
20
11
20
20
20
20
28
15
30
so2
ppm
720
350
560
3800
3200
.-
-
1200
1300
2200
-
~
790
3100
560
700
220
210
380
640
760
2800
400
M>0
Ion
HO
ppm
160
120
160
160
190
120
250
160
225
220
300
170
170
200
180
200
150
270
190
240
150
230
170
310
320
180
160
180
220
320
300
300
280
290
210
CO IXOj
PP»
54
100
81
9/
290
26
11
67
56
44
30
17
15
23
110
84
370
42
26
51
U
70
40
20
42
78
31
.
15
27
26
26
23
41
18
55
SO at 3X0, ppn
Calculated
940
1570
1270
4700
4700
390
940
940
940
4700
4700
1570
590
590
3830
3830
3840
4700
390
940
3500
2760
940
1570
3840
4700
4020
407.0
1090
390
1570
940
940
4700
590
4020
Measured
970
510
760
4630
1880
i
-
-
1370
4852
880
1160
100
280
490
830
860
4090
540
1100
Fuel
Sulfur
enlttcd
X
(103)
32
60
98
83
~
-
-
36
(103)
22
29
28
72
31
88
91
87
92
27
Nil at 1Z02- PP"
Calculated
2280
2030
2510
2220
2220
2400
2280
2710
2710
2220
2220
2030
3400
3400
2560
2560
1690
2220
2400
2710
2490
2280
2030
2020
2200
2250
2250
2510
2400
2070
2710
2280
2220
3400
2250
HeHStired
2 2O
160
220
190
230
160
320
480
320
330
300
290
260
310
280
340
220
380
240
310
210
320
230
540
500
280
2CO
240
300
410
390
340
410
ISO
420
Fuel N
Converted
to NO
X
9.6
8.0
8.7
8.6
10
6.7
7.1
18
12
15
14
14
7.6
9.1
11
7.5
13.
17.
10.
11.
8.4
10.
--
27
23
12
12
9.6
13
20
14
15
18
11
H
Particulate
Loading
Bg/SNl
64
338
160
120
-
290
400
1360
700
360
330
500
380
170
780
430
1020
390
170
340
200
160
180
110
POM
Load 1 ng
.5
.9
.7
.3
17
9
60
17
-------
Gaseous Emissions
NO. Table 1-3 indicates that a small fraction, less than 10
percent, 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 attri-
buted to the inherent "staged combustion" of stoker firing and/or low excess
air combustion occurring within the stoker bed. Additionally, as suggested
in Reference 1-3, the relatively high levels of CO that are present in
the fuel bed may serve to reduce the NO to N .
SQg. Table 1-3 shows the percent fuel sulfur emitted as S02 for
selected coals. It is well recognized (References 1-4 and 1-5) 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 tiring 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 1-3.
The measured values of S0_ were exceptionally low for the lignite
and the western subbituminous coals coals with relative high sodium and
calcium contents. Maloney suggests that lime is more effective in
capturing fuel sulfur as sulfide in fuel-rich regions. It was observed
in the underfeed stoker 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.
In Run UF-5, -50 mesh limestone was fed at a rate equivalent
to a Ca/S molar ratio of 7 along with the stoker-sized coal. S02 emissions
were reduced by 17 percent, a considerably lower reduction than for coals
with naturally occurring alkaline content and lower Ca/S ratios. This
result suggests that the manner in which the alkaline material contacts
the sulfur and the conditions (oxygen level, temperature) at which this
contact occurs are important.
1-15
-------
CO. CO levels can generally be controlled with the overfire
or secondary air. For the coals fired, overfire air rates of 10 to 20
percent of the total air were required to minimize CO emissions as well
as to achieve near smokeless operation. CO levels varied between 50 and
100 ppm, except when limestone was added to the Illinois No. 6 coal
(Run UF-5). The addition of limestone appeared to flux the ash resulting
in clinkering. The clinkering caused nonuniform bed burning, increasing
the CO levels to 290 ppm.
Particulate Loadings
In comparison to the high volatile Eastern coal, and to a lesser
extent the Illinois No. 6, the lignite and western subbituminous coals
generated higher particulate loadings. Generally in underfeed firing, there
is very little fly-ash carryover. However, because of the friability of
the lignite and western subbituminous coals, an excessive amount of fines
were produced while conveying the coal to the retort. A rough measure
indicated about 50 percent fines were introduced into the retort. Also,
because of the high sodium content of the lignite, it appears that some
portion of the particulate loading may be attributed to condensed sodium
vapor.
POM Loadings
Previously reported POM data indicated that loadings varied by
an order of magnitude for essentially identical runs firing fuel oils
under controlled laboratory conditions. The mechanisms involved in the
formation of POM are not fully understood and thus it is difficult to
explain these apparent discrepancies. Thus, in reviewing the POM loadings
presented in Table 1-3, there appears to be no significant difference in
the levels that could be attributed to coal type or the combustion
conditions. Of greater significance is the fact that the POM levels
generated during continuous firing of this unit were at least 1000 less
than those measured from the same unit operating under cyclic conditions.
Table 1-4 presents POM species quantification.
1-16
-------
TABLE 1-4. POM QUANTIFICATION FOR UNDERFEED STOKER EXPERIMENTS
Component Notation
Anthracene/Phenanthrene
Methyl anthracenes
Fluoran there
Pyrene
Methyl Pyrene/Fluoranthene
Benzo(c)phenanthrene ***
Chrysene/Ben2(a)anthracene *
Methyl chrysenes *
7, 12-Dimethylbenz (a) anthracene ****
Benzo fluoranthenes **
Benz(a)pyrene ***
Benz(e)pyrene
Perylene
Methylbenzopyrenes
3-Methylcholanthrene ****
Indeno(l,2,3,-cd)pyrene *
Benzo ( ghi) pery lene
Dibenzo(a,h)anthracene ***
Diebenzo(c,g)carbazole ***
Dibenz(ai and ah)pyrenes ***
Coronene
Total
UF-1,
ng
384
98.4
380
30.2
0.1
5.2
10.4
0.1
ND(a)
ND
910ng
UF-2, UF-3,
ng ng
863 527
257 123
385 405
88.8 0.1
0.1 0.1
4.0 2.4
12.7/ 2.4 7.4
0.1 0.1
0.1
5.5
2.7
2.2
0.1
1620ng 1070ng
UF-4 ,
ng
169
22.7
281
4.8
0.1
3.7
11.8
0.1
490ng
(a) None detected.
1-17
-------
The POM loadings in Runs UF-1 through UF-4 are very low. These
levels are about an order of magnitude lower than those measured from
firing a 500-kw package boiler on oil and gas. This is a somewhat
surprising result when coal is believed to generate higher levels than
oil or gas for similar combustion condition. A review of the sampling
and analysis procedures indicated identifiable changes in technique
occurred that could have resulted in the POM levels being altered. Because
the levels are extremely low, the data are considered suspect.
MODEL SPREADER STOKER RESULTS
Table 1-3 summarizes the results of the model spreader stoker
experiments for the coals fired and the operating conditions listed.
Gaseous Emissions
NO. For the most part, 10 to 20 percent of the fuel nitrogen is
converted to NO (assuming no thermal NO). This percent conversion is
somewhat higher than that observed in the underfeed runs and may be attri-
buted to two factors. First, in the model spreader stoker some portion
of the coal is burned in suspension. (In industrial spreader stokers as
much as 50 percent of the coal is burned in suspension.) In suspension
burning, the coal particles are surrounded by an oxidizing atmosphere,
which can contribute to a higher percentage of fuel nitrogen conversion
to NO than in bed burning alone. The other factor is that in spreader
stoker firing, the fuel bed is thinner and there is more surface burning
than in underfeed firing. This condition results in a greater portion of
the combustion occurring in an oxidizing atmosphere.
SO,,. The percent fuel sulfur emitted as S0« for selected
conventional coals fired in the model spreader stoker was similar to that
observed during the underfeed stoker runs. In Run MS-24, because of the
relatively high alkaline content in the lignite, only about 31 percent
of the fuel sulfur was emitted as SO-. Likewise for the bituminous
1-18
-------
coals that have lower alkaline contents, between 72 and 100 percent of
the fuel sulfur was emitted as S0_. Unfortunately, due to degradation
of an electrolytic cell, the Faristor analyzer used to measure SO was
found to give inconsistent results during Runs MS-1 through MS-19. As a
consequence, the data are not reported. The problem was corrected after
Run MS-19.
Hydrothermally Treated Coal. The hydrothermal treatment of coal
removes approximately 50 percent of the fuel sulfur depending on whether
the sulfur is organic or inorganic. The process results in some
alkali residuals (7 percent Ca and 0.5 percent Na by weight) that are
effective in capturing a portion of the remaining fuel sulfur. As
indicated by Run MS-24 in Table 1-3, only 28 percent of the fuel sulfur
contained in the HTT coal was emitted as S0_.
Limestone/Coal Pellets. The most significant finding of the
model spreader stoker experiments was that the limestone/high-sulfur coal
pellet was effective in capturing as much as 78 percent of the fuel
sulfur. For a Ca/S molar ratio of 7, as indicated in Runs MS-22, MS-23,
and MS-31, only 22, 29, and 27 percent, respectively, of the fuel sulfur
was emitted as SCL. Similarly, for a Ca/S molar ratio of 3-1/2 in Run
MS-20, only 36 percent of the fuel sulfur was emitted as S02. Illinois
No. 6 coal was used as the base coal (Runs MS-21 and MS-29).
In comparison to Run UF-5, where limestone was simply introduced
into the feed system along with stoker-size coal, the data indicate that
the limestone/coal pellets were significantly more effective in capturing
sulfur. This increased sulfur capture may be attributed to the intimate
contact of the limestone with the coal particle in the pellets and possibly
to the more reactive surface of the pellet.
CO. As in the underfeed stoker experiments, during the model spreader
runs CO levels-were controlled with overfire air. CO levels were generally less than
100 ppm. CO levels were noticeably high for two of the pellet runs (MS-10
and MS-12), which may be attributed to several factors. Observation of
1-19
-------
the fuel bed during these runs indicated nonuniform burning and a general
appearance of low heat release rate. Temperatures in and above the bed
for these runs, as well as for other pellet runs, were less than 1000 C
while those for the conventional coal runs were greater than 1000 C as
indicated in Table 1-5.
Particulate Loading
The model spreader stoker, as expected, generates significantly
higher particulate loadings than does the underfeed stoker firing the
same coals. The higher particulate loadings are attributed to the suspen-
sion burning that occurs in spreader stoker firing that results in a
greater amount of fly-ash (carbon) carryover than is observed in either
underfeed or overfeed stokers. Another significant, but expected, finding
was that particulate loadings from the limes tone/coal pellet runs were
significantly higher than those from the conventional coal runs. These
higher loadings of the limestone/coal pellet are attributed to the high
ash content of the pellet (about 50 percent for Ca/S = 7) and the fines
fed into boiler.
The particulate loadings do not correlate with either the ash
content of the coal nor its size prior to feeding. It was observed
that the feed system would often crush the coal and reduce its size.
The amount of crushing and size reduction appears to be dependent on the
initial coal size and, also, the friability of the coal. Unlike a large
industrial stoker, the model spreader does not provide sufficiently
long residence times at elevated temperatures to burn coal fines in
suspension. As a consequence, an unusually high amount of fly carbon,
in addition to the fly ash, is emitted to the stack. Thus, the particulate
loadings of the model spreader experiments were affected by factors other
than ash content of the coal that are not easily quantified.
POM Loadings
POM loadings were measured for Runs MS-3, MS-4, MS-5, and MS-6.
There does not appear to be any relationship between coal type and POM
1-20
-------
TABLE 1-5. POM QUANTIFICATION FOR MODEL STOKER EXPERIMENTS
NAS MS-2 MS-3, MS-5 ,
Component Notation ng ng ng
Anthracene/Phenanthrene
Methyl anthracenes
Fluoranthene
Pyrene
Methyl Pyrene/Fluoranthene
Benzo ( c) phenanthrene
Chrysene/Benz (a) anthracene
Methyl chrysenes
7, 12-Dimethylbenz (a) anthracene
Benxo fluoranthenes
Benz(a)pyrene
Benz(e)pyrene
Perylene
Methylbenzopyrenes
3-Methylcholanthrene
Indeno(l,2, 3,-cd)pyrene
Benzo ( ghi) perylene
Dibenzo (a , h) anthracene
Diebenzo(c,g)carbazole
Dibenz(ai and ah)pyrenes
Coronene
TOTAL
15,013 5,313 20,955
2,648 1,553 3,210
10,927 2,516 42,340
2,591 1,409 7,142
1,528 ND ND
*** 921 911 1,034
* 1,242 972 2,027
* 1,041
****
** 1,329
*** 1,137
****
* 1,062 1,076
1,035 1,047
***
***
***
36,000 12,700 82,000
MS-6,
ng
17,417
4,125
13,405
4,542
2,677
1,099
2,716
1,677
2,207
1,286
1,161
1,024
53,300
ND = None Detected
1-21
-------
levels nor do the POM levels correlate with CO levels or particulate
loadings. The POM levels are, as expected, somewhat higher than those
measured in the oil/gas 500 kW packaged boiler. ~6' Table 1-5
gives the POM specie quantification.
Particle Size Distribution
Particle size distribution, as determined by the Coulter Counter
method using the filter catch, was similar for Runs MS-2, MS-3, MS-5,
and MS-10. The range was from about 5 to 100 micrometers with an average
particle size between 15 and 30 micrometers, about the same as that
observed for pulverized coal firing. The upper cutoff limit could be
attributed to the probe collecting particles larger than 100 micrometers
while the lower limit is attributed to the particular Coulter Counter
procedure used. Figure 1-3 shows these distribution curves.
Gas Probing Observations
In an attempt to determine the combustion phenomena occuring
in the stoker boiler, the furnace environment was sampled for gaseous
constituents at various locations within the firebox in a plane 0.3 m
above the bed using a watercooled stainless steel probe. Evidently,
the overfire air jets and the combustion generated sufficient turbulence
that consistent data could not be collected. Gaseous constituents varied
significantly. For example, 0? varied from 3 percent to 20 percent and
CO from 40 ppm to 1000 ppm. The sample ports were located in approxi-
mately the same plane as the overfire air jets and, as a result,
measurements were significantly affected. Gas probing 0.6 m above the
bed showed that the levels of gaseous constituents in this region were
similar to those in the stack.
1-22
-------
100
100 3
Particle Size, urn
10
d.
too
FIGURE 1-3. PARTICLE SIZE DISTRIBUTION OF FLY ASH
FOR RUNS a) MS-72, b) MS-73, c) MS-5
and d) MS-103
1-23
-------
Bed Temperature Profiles
Temperatures were measured at three points within the furnace
at 25 mm above the grate, well within the bed; at 100 mm above the grate,
close to the surface of a well-established bed; and at 200 mm above the
grate definitely in the freeboard (suspension burning zone) for most of
the run. Because the fuel bed is fixed and there are no provisions for
removing ash, the bed increases in depth and the flame front continually
changes relative to the location of the thermocouples. These temperature
measurements were made over a period of time and were used to estimate
relative burning rates for several coals. Stack temperatures at the outlet
of the furnace were also recorded. The average values for these tempera-
tures during each run are reported in Table 1-6, along with the feed rates
and the final bed depth.
Figures 1-4 and 1-5 show bed temperature profiles for high (40
percent) and low (21 percent) volatile bituminous coals. Since both the
combustion conditions and stack temperatures of these runs (MS-1 and MS-4)
were essentially the same, overall or total heat releases from the combus-
tion of these coals can be assumed to be essentially equivalent. Thus, a
comparison of bed temperature (25 mm location) and freeboard temperature
(200 mm location) provides an indication of the relative differences in
heat release between these two-regions. For example, the low volatile
coal produces a higher bed temperature than freeboard temperature
indicating a higher heat release within the bed than in the combustion
zone above the bed. On the other hand, the high volatile coal produces
a higher freeboard temperature than bed temperature indicating a rela-
tively higher heat release in the freeboard region than in the fuel bed.
Toward the end of a run, as the bed becomes deeper, the temperatures at
the 200 mm point for each approach each other. Temperature at the 25 mm
point drops 'as this zone becomes buried in the ash and is cooled by
underfire air. With high ash coals and deep beds, for example with the
limes tone/coal pellets, this dropoff becomes quite dramatic as the
combustion zone moves above this point. Profiles of this type are a
1-24
-------
TABLE 1-6. BED TEMPERATURES
Average Temperatures, C
Run No.
~
MS-1
MS-2
MS- 3
MS-4
MS-5
MS- 6
MS- 7
MS- 8
MS-9
MS-10
~
MS-11
MS-12
MS-1 3
MS- 14
MS-15
MS-16
MS- 17
MS- 18
MS- 19
~
MS-20
MS-21
MS-22
MS-23
MS-24
Coal
SE Kentucky
SE Kentucky
SE Kentucky
SE Kentucky
SE Kentucky
E Kentucky
Lo-Vol West Virginia
E Kentucky
SE Kentucky
SE Kentucky
Illinois No. 6
Illinois No. 6
Lignite
Lignite
Lo-Vol West Virginia
West Bituminous
West Bituminous
L/C Pellets Ca/S=7
SE Kentucky
SE Kentucky
L/C Pellets Ca/S=7
L/C Pellets Ca/S=3-l/2
Illinois No. 6
Lo-Vol West Virginia
SE Kentucky
W Kentucky, unwashed
W Kentucky, washed
E Kentucky
Lignite
Western Bituminous
L/C Pellets Ca/S=3.5
Illinois No. 6
L/C Pellets Ca/S=7
L/C Pellet Ca/S=3-l/2
HIT Pellets
25mm
1200
950
1430
1450
1180
1430
~
1290
1180
1200
1150
1200
1400
420
800
800
1220
1220
610
520
890
1400
1000
980
1310
680
820
200
1190
100mm
1180
1230
~
1320
1260
1320
1480
1230
1260
1040
1260
1260
820
1120
1290
830
1270
890
1210
1370
660
760
1300
1420
1090
960
1200
1220
1040
1100
1200
870
290
1250
200mm
1450
1150
1320
1230
1200
1340
1230
1200
1150
1150
820
870
1320
1100
1220
970
1210
1280
1130
1120
1120
1210
1000
800
1180
1130
1050
1140
720
1240
860
1050
1080
Stack
350
360
~
370
400
340
340
350
350
360
330
320
270
280
360
330
350
310
380
360
300
290
330
350
340
300
340
360
300
320
330
370
330
300
300
Final
Bed
Dept. ,
mm
100
100
250
550
150
330
100
180
90
100
150
130
510
150
200
460
<100
180
280
280
130
90
50
80
50
50
150
100
150
250
330
250
Coal
Feed,
kg/hr
27
22
22
22
32
27
27
25
28
27
28
27
27
50
25
28
30
61
32
25
43
46
31
23
24
26
25
25
42
27
37
29
43
36
45
Total
Air,
kg/hr
300
300
250
340
300
300
300
300
300
300
300
300
220
220
310
300
300
220
360
300
250
250
300
300
300
300
300
300
.200
300
270
270
260
250
250
1-25
-------
1750
1500 -
1250
o 1000
-------
1750
1500
1250
O 1000
of
750
500
250
200mm over grid
100mm over grid
25 mm over grid
Low Volatile (21%) W.Va Bituminous Coal
345
Duration of Run, hr.
FIGURE 1-5. BED TEMPERATURE PROFILES FOR A LOW VOLATILE COAL
1-27
-------
graphical indication of the relative amounts of suspension burning. These
results relating bed temperatures to coal volatility confirm those of
Marskell, et al/Z~8^
Marskell conducted pilot scale testing of a variety of coals
under spreader firing conditions. His studies are relevant in that
the fuel-bed isotherms that were generated provide some insight as to
regions in which the calcium-sulfur reactions may occur. His studies
indicate that fuel-bed temperatures can be as high as 1480 C to 1590 C
depending on the type of coal. Moreover, his data indicate that there are
significant regions within the fuel bed where temperatures are below 1370 C.
It is in these regions that we suspect Ca/S reactions will occur. Figure
(18)
1-6 is an isotherm for a 27 percent volatile Australian coal.
Observations of Suspension and
Fixed-Bed Combustion
Suspension Burning. In addition to the bed temperature profiles,
several observations support a conclusion that suspension burning is
significant in the model spreader stoker. After a uniform bed was established
in these experiments, the coal feed was stopped. CO levels then increased
from 50 ppm to greater than 1250 ppm. (The same phenomenon was observed in
the 8-MW spreader stoker.) Evidently, the combustion of the coal in
flight provided a sufficiently high temperature to burn out the CO
evolving from the bed. Also, the combustion zone above the bed provided
a thermal barrier between the fuel bed and the heat transfer surface.
Without this combustion zone, the fuel-bed temperature dropped quickly
resulting in a reduced burning rate. Thus, when the coal feed was stopped,
while maintaining the combustion air flow rate, 0- concentration in the
stack increased significantly, suggesting that significant suspension
burning occurred and that bed burning was reduced. Apparently volatile
matter was driven off above and immediately at the surface of the bed and
provided a region of intense burning. The coal remaining in the bed
appeared to be essentially all char. To support this observation, four
samples were collected during various phases of stoker combustion. Table
1-28
-------
* I
I Burning- '
1 I
| out zone i
I
Non-equilibrium zone
0.9 0.8
0.7 0.6 0.5 0.4 0.3
Fraction of combustion time
0.2.
FIGURE 1-6, FUEL BED ISOTHERMS FOR A 27 PERCENT
VOLATILE AUSTRALIAN COAL(I~8)
1-29
-------
1-7 lists the ultimate and proximate analyses of these samples. Sample
No. 1 is the raw coal, Sample No. 2 is a lump of coal that was extracted
from the stoker after about 4 minutes of burning, Sample No. 3 was extracted
from the burning bed, and Sample No. 4 was removed from the bed after
combustion was completed. These data indicate that the coal devolatilizes
readily while carbon burnout proceeds more slowly. In addition, analysis
of Sample Nos. 2 and 3 on a mass basis suggests that sulfur is burning at
about the same rate as the carbon.
Fixed Bed Combustion. In another set of experiments with the
southeastern Kentucky coal, bed depth was varied (1) to determine the
optimum depth and (2) to optimize excess air. Thermocouples were located
in the fuel bed and a gas sample probe and thermocouple were located above
the bed. Measurements from these instruments along with visual observations
were used in characterizing the stoker combustion. Maximum bed temperatures
of 1370 C were observed for shallow beds (50 to 100 mm). Temperatures
decreased with increasing bed depth. For bed depth of 200 mm or more, the
average bed temperature was about 820 C with the higher temperatures occur-
ring near the surface of the bed. Surface temperatures of 1260 C to
1370 C were measured for shallow beds and dropped to about 1100 C for the
deeper beds.
These results were somewhat confusing in that classical fixed
bed combustion indicates that the maximum flame temperature should occur
at the flame front near the incoming air. However, this classical
approach deals with a homogeneous bed. This is not the case for our
model spreader as the coal continually devolatilizes, the rate decreasing
for increasing bed depth. The maximum bed temperature occurs in the
region of maximum reactivity. In the case of our model spreader stoker,
this appears at or near the surface of the bed.
These results indicate that if limestone were to be effective
in capturing SO-, the stoker should be operated with a deep bed. (This
would not be the case if most of the SO,, comes off in suspension burning.)
Temperatures are lower and there is more reactive surface. This supports
1-30
-------
TABLE 7. PROXIMATE AND ULTIMATE ANALYSES (IN WEIGHT PERCENT) OF
4 COAL SAMPLES OBTAINED DURING COMBUSTION
Moisture
C
H
N
Cl
s
Ash
°2
Vol.
Fixed
carbon
No. 1
Raw Coal
3.89
76.74
5.24
1.53
0.12
1.22
5.16
6.10
36.94
54.01
No. 2
Devolatized
Coal Luat>
1.27
73.59
1.45
1.17
0.04
0.55
21.48
0.45
5.62
71.63
No. 3
"Burning Bed"
0.48
87.28
0.38
0.84
0.08
1.06
11.56
-1.68
2.68
85.70
No. 4
"Dead" Bed
0.18
7.59
0.11
0.06
0.02
0.10
92.54
-0.60
1.41
5.87
1-31
-------
our observations on the underfeed effort in which the limestone captured
up to 70 percent of the sulfur in a deep bed. The effectiveness of the
limestone in capturing sulfur that comes off with the volatiles remains
questionable, however.
These results also suggest that the ratio of the overfire to
underfire air should be a function of bed depth. If the coal (containing
up to 40 percent volatiles and a somewhat lower percentage of the heating
value) devolatilizes at or above the bed then, perhaps, an amount of air
sufficient to burn these volatiles should be introduced in this region
rather than through the underfire grate. An excess amount of air passing
through the grate will cool the coal and retard the burnout of the carbon
in the bed.
1-32
-------
SECTION V
ALTERNATIVE FUELS FOR STOKERS
PELLETIZATION APPLICATIONS
During Phase I, no candidate "treated" industrial stoker-
boiler fuels were identified that would be commercially available in
the near future. However, some alternative fuels that have potential
to become stoker fuels include:
Most chemically cleaned coals
« "Deep" physically cleaned coals
Limestone/coal mixtures
Waste coal fines
Refuse-derived fuels
Wood wastes.
These fuels may require restructuring before they can be handled and fed
properly into a stoker-boiler. Restructuring consists of a technique to
upgrade fuel particle size. Such a technique must provide a restructured
product that has:
Variable particle size
Mechanical strength, durability and
weatherability similar to raw coals
without the addition of large amounts
of binder
Controlled density
No undesirable combustion characteristics
Acceptable production concepts.
Some of the pelletization processes that were briefly explored
show promise of meeting these requirements. Pelletization is not a new
technology but the application of it to restructuring alternative
stoker boiler fuels is. Consequently, some research effort will be
1-33
-------
required before techniques are fully developed. Prior to that, some
potential market area must be created. The development of the limestone/
coal pellet as a viable industrial boiler fuel may help create that
market.
SECTION VI
CONCLUSIONS
Phase I experiments indicated that
The conversion of fuel No to NO was less than
20 percent for stokers, somewhat less than
pulverized coal firing
POM levels for continuous combustion were
significantly lower than for intermittment
combustion
Coals containing alkaline material retain
significant amounts of sulfur in the ash.
Intimate contact of the sulfur and alkaline
material is essential to achieve substantial
sulfur retention.
Additionally, no treated or modified coals were identified that could be
evaluated during Phase I. As a result, a limestone/coal fuel pellet was
developed. This pellet offers a potentially new means of sulfur control
for industrial boilers. Further refinement of the pellet and evaluation
in larger-scale equipment are needed to assess its full potential.
1-34
-------
REFERENCES
1-1. Giammarj, 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.
1-2. Jones, P. W., et al. Improved Measurement Techniques for Polycyclic
Aromatic Hydrocarbons from Combustion Effluents. In: Carcinogenesis
A Comprehensive Survey, Vol 1, Polynuclear Aromatic Hydrocarbons,
Chemistry, Metabolism, and Carcinogenesis, Raven Press, NY, 1976.
1-3. Merryman, E. L., S. E. Miller, and A. Levy. Reduction of NO in the
Presence of Fly Ash. Combustion and Science Technology, 20, 160-163,
1973.
1-4., Gronhovd, G. H., P. H. Tufte, and S. J. Selle. Some Studies on Stack
Emissions from Lignite-Fired Power Plants. Presented at the 1973
Lignite Symposium, Grand Forks, ND, May 9-10, 1973.
1-5. 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
Environmental Research Laboratory, EPA, Research Triangle Park, NC,
Nov, 1978.
1-6. Giammar, R. D., Weller, A. E., Locklin, D. W., and Krause, H. H.,
Experimental Evaluation of Fuel Oil Additives for Reducing Emissions
and Increasing Efficiency of Boilers, U.S. Environmental Protection
Agency Report No. 600/2-77-008b, Jan, 1977.
1-7. Stambaugh, E. P., et al., Combustion of Hydrothermally Treated Coals,
EPA-600/7-78-068, U.S. Environmental Protection Agency, Washington, D.C.,
20460, April, 1978.
1-8. Marskell, W. G., and C. W. Pratt, J. Inst. Fuel, 28, 212-21, 1952.
1-35
-------
APPENDIX I-A
BINDER IDENTIFICATION AND
RESTRUCTURING TECHNIQUE REVIEW
-------
APPENDIX I-A
BINDER IDENTIFICATION AND
RESTRUCTURING TECHNIQUE REVIEW
Binder Identification
A cursory review of the literature indicated that a number of
binders could be used to palletize the finely ground limestone and coal.
Table I-A-1 lists those binders that were evaluated during this phase
to produce pellets. Of these, cement and a latex emulsion were found to
be the most desirable binders for the initial combustion tests because
they were readily available, relatively inexpensive, easy to handle, and
produced a. pellet that remained intact during feeding and handling. No
sophisticated test procedures were used to evaluate the mechanical strength
properties of these pellets in this phase of the program. The pellet was
dropped from 6 m, and if it remained intact, the pellet was considered
satisfactory for the combustion evaluation experiments.
A review of the literature indicated that binders improve the
strength of compacts and generally are classified as matrix type, film
*
type, or chemical . They usually take advantage of combinations of the
forces due to solid bridges and/or immobile liquid bridges. Table I-A-1
lists some of the important characteristics of the binders tested. Highly
viscous materials such as asphalt and pitch form immobile liquid bridges.
These bridges between the compacted particles more fully exploit available
adhesion and/or cohesion forces producing greater binding ability than the
mobile bridges. This results in significantly stronger pellets. These
immobile bridges fail by tearing apart the weakest bond, leaving the
remainder intact. Such pellets tend to be maleable and deform rather
than powder. However, many of these binders can be difficult to handle
due to their viscous nature. A tempering process is frequently required
during pellet production. Pitch for example is normally kept within a few
degrees of its softening point during the entire operation.
* Komarek, Chem. Eng., 74 (25), 154 (1967).
-------
As the initial thrust was to pelletize coals which have the
potential to generate less SO , some potentially good binders containing
high amounts of sulfur were not included. The tests of HIT coal and the
initial limestone/coal (L/C) pellet tests in the model spreader were all
performed with pellets utilizing a latex emulsion (Du Pont Hycal-S83) as
a binder. This binder is a room temperature film former that utilizes both
immobile liquid bridges and solid bridges to increase pellet strength.
Pellets produced with this binder had sufficient mechanical strength for
the model spreader, but were crushed severely when encountering the severe
mechanical stresses of the underfeed stoker. Additionally, the pellets
deteriorated when fed through the Battelle steam plant boiler coal-handling
system.
Common cement as a binder produced a pellet with sufficient strength
to survive the steam plant boiler system. Cement is a matrix-type binder
which relies mostly on solid bridges formed when this "gel" sets to produce
additional strength. Cement normally contains approximately 4 percent sulfur
which would not effect SO levels if used in small quantities. Pellets
utilizing this binder exhibit excellent weathering characteristics as
might be expected.
Though many potential binders were not tested, this investigation
was terminated at this point as the primary objective of mechanical strength
had been met. Many interesting aspects remain to be explored. What role
could the binder or other additions play during combustion was virtually
ignored in this work. Cement is an inert that increases the ash content,
while polyvinyl alcohols are ash free and contain appreciable heating
value. Substances as or in addition to the binder which volatilize readily
in the combustion zone could allev iate this potential problem. For example
fuel oils and motor oils do not produce strong pellets, but their addition
with cement binder would increase the volatility and weatherability of the
fuel. Disposal of used crankcase oil is a problem area due to its heavy
metal content. When added to the L/C pellets these metals might be trapped
in the ash, while significantly increasing the volatility of the fuel.
I-A-2
-------
TABLE I-A-1. BINDERS TESTED FOR USE IN RECONSTITUTING FUELS
Binder
Type
Bonding Mechanisms
Comments
Water
Pitch-asphaltum
Starches
(corn, potatoe)
Sugar
Sterotex (hydro-
genated cotton-
seed oil)
Fuel oil
Motor oil
X 83
Bentonite
Cement
Film
Film
Lignin sulfonates Film
Matrix
Matrix
Matrix
Film
Film
Hycar 2800 (latex Film
resin) X 138
Film
Elvanol (polyvinyl Film
alcohol) Matrix
Methyl Cellulose Film
Matrix
Matrix
Chemical
Matrix
Chemical
Mobile liquid bridges
Forms immobile liquid bridges
hard pitch can form solid bridges
Immobile liquid bridges
Solid bridges formed during
crystallization
Solid bridges formed during
crystallization
Solid bridges formed as oil
recrystallizes
Mobile liquid bridges
Mobile liquid bridges
Immobile liquid and solid bridges
Immobile liquid and solid bridges
Solid bridges during
crystallization
Solid bridges during
crystallization
Solid bridges
Solid bridges formed when
"gel" sets
Present in all pellets. Some fuels such as
BTC produce very hard pellets with water alone.
A high percentage is usually required. Also
needs tempering but exhibits good weathering.
Rejected due to handling problems and the
amount of sulfur contained.
Readily available from papermills. Tends to
produce maleable pellets. Rejected due to
high sulfur content.
Inadequate pellet strength - brittle
Inadequate pellet strength - brittle
Relies on heat of pelleting process to melt
the oil. Exhibits good weatherproofing as well
as lubrication of the die. Combustible. Re-
jected - brittle pellets.
Add heating valve and volatility . Good weather-
proofing. Pellets too soft.
Add heating valve and volatility. Good weather
proofing. Soft pellets.
Required curing. Combustible, ash free strong
pellet.
Room temperature. Film former. Easier to
handle and somewhat stronger than X 138. The
binder used in the majority of model tests.
Binder difficult to handle-moisture becomes
critical. Ash free with significant heating
value. Tests not conclusive.
Similar to elvanol
No heating value - 100 percent ash. Produces
a strong pellet but softens badly when wet.
For final strength and weatherproofing it must
be fired.
No heating value - 100 percent ash. Produces
a very strong pellet with excellent weatherability.
Binder selected for Phase II experiments. Crushes
when excess force applied.
I-A-3
-------
PHASE II. CONTROL TECHNOLOGY EVALUATION
-------
PHASE II
CONTENTS
LIST OF FIGURES
Page
iii
LIST OF TABLES 1V
EXECUTIVE SUMMARY v
I. BACKGROUND II-l
II. OBJECTIVE AND SCOPE II-2
III. PLAN OF EXPERIMENTAL INVESTIGATION II-3
Facility II-3
Stoker II-3
Fly Ash Reinjection System II-5
Overfire Air Jets II-6
Boiler II-6
Sampling Ports II-6
Coal Handling System II-7
Coal Properties II-7
Pellet Production II-9
Material Preparation II-9
Processing Operations II-9
Sampling Equipment and Procedures 11-11
Mass-Rate Determinations 11-12
IV. EXPERIMENTAL RESULTS 11-14
Stoker-Boiler Emissions 11-14
Nitrogen Oxides 11-14
Sulfur Oxides 11-16
CO 11-22
Particulate Loading 11-22
POM 11-22
Effect of Operating Variables on Emissions and Boiler
Performance 11-25
Excess Air . . .' 11-25
Overfire Air 11-34
Fuel-Bed Depth 11-35
Boiler Load 11-36
-------
PHASE II
CONTENTS (Continued)
Fly Ash Reinj action 11-41
Coal Types 11-42
Combustion System Design Modifications 11-44
Feed System 11-45
Grate Design 11-45
Overfire Air 11-46
Fly Ash Reinjection 11-47
SUMMARY 11-48
REFERENCES 11-49
APPENDIX A - OVERFIRE AIR JETS
-------
PHASE II
List of Figures
Page
Figure II-l. Schematic of the 8-MW Stoker-Boiler Facility. . . . II-4
Figure II-2. Pellet Procuction Process Flow Diagram 11-10
Figure II-3. NO Emission Levels as a Function of Excess Air For
Washed, Run-Of-The-Mine Ohio Coal 11-30
Figure II-4. NO Emission Levels as a Function of Excess Air For
Unwashed Ohio Coal 11-31
Figure II-5. NO Emission Levels as a Function of Excess Air
For Washed, Stoker Ohio Coal 11-32
Figure II-6. NO Emission Levels as a Function of Excess Air For
LOW-Sulfur Kentucky Coal 11-33
Figure II-7. Fuel-Bed Temperature Distribution 11-40
111
-------
PHASE II
Table II-l.
Table II-2.
Table II-3.
Table II-4.
Table II-5.
Table II-6.
Table II-7.
Table II-8.
Table II-9.
Table 11-10.
Table 11-11.
Table 11-12.
Table 11-13.
Table 11-14.
Table 11-15.
Table 11-16.
Table 11-17.
List of Tables
Properties of Coals Fired in Phase II II-8
Operating Variables Studied for System Characteri-
zation in Phase II 11-13
Emission Summary Data of 8 MW Boiler Experiments II-15
Conversion of Fuel Nitrogen to NO for Coals Fired
In Phases I and II 11-17
Sulfur Balances for Steamplang Stoker Runs 11-19
Distribution of Coal Ash Within The Stoker-Boiler
System 11-20
CHNS and Ash Analyses of Bes Ash, Reinjected Fly Ash,
Cyclone Catch, and Filler Catch, Wt Percent 11-21
Coal Properties Affecting Particulate Loading. . . . 11-23
Listing of POM Loadings, Particulate Loadings Carbon
Particulate, CO Emission Level, and Smoke Opacity
For Selected Runs 11-24
POM Quantification 11-26
PAH Quantification pg Total Sample 11-27
Comparison of Particulate Loading Between Optimal And
Normal Stoker Operation Firing the Medium Sulfur
Kentucky Coal 11-29
Smoke, CO, and Particulate Emissions for Two Overfire/
Total Air Ratios 11-35
The Effect of Boiler Load on Performance and Emission
Levels 11-37
SO,, Levels As A Function of Boiler Load and Fuel Bed
Temperature 11-39
Effect of Fly Ash Reinjection 11-42
Boiler Efficiencies for Selected Coals 11-43
IV
-------
EXECUTIVE SUMMARY
Potential control concepts were identified and evaluated in
the Battelle 8 MW , (25,000 Ib steam/hr) spreader stoker boiler. Con-
th
trol strategies were limited to:
Use of compliance coals
Combustion-system operational modifications
Minor combustion-system design modification
Use of treated coal (limestone/coal fuel pellet).
Flue-gas clean-up techniques were not considered. Criteria pollutants
were used as the basis for evaluation.
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 for S02 control. In summary, the major findings are:
The limestone/high-sulfur coal pellet showed a
sulfur capture of about 75 percent for a Ca/S
molar ratio of 7.
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 except the limestone/coal pellet.
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
significantly. Particulate loadings are also
reduced with increased overfire air.
« NO is lower for inactive overfire air jets.
* Clinker formation occurs readily if bed
depths become excessive, while the danger
of burning the grates exists for operation
with very shallow beds. Bed depths around
6.3 to 7.6 cm appear to be optimum for
low ash coals.
3
a POM levels ranged from 13 to 24 yg/Nm .
They were somewhat lower than those of the
model spreader and only slightly higher than
those from a 500 kW packaged boiler firing
natural gas and fuel oil.
« A higher excess air rate is required for low-
load than for partial- or full-load operation.
A greater percentage 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.
0 At full load, fly-ash reinjection increase
boiler efficiency by 1.5 percent. However,
particulate loadings were reduced by 10 to 25
percent by operating without fly-ash reinjection.
The high-sulfur Ohio coals had to be fired
at higher excess air rates than did the low-
sulfur Ohio and Kentucky coals. The high-
ash unwashed stoker coal and high moisture
Illinois No. 6 coal could not be fired
satisfactorily.
VI
-------
PHASE II. CONTROL TECHNOLOGY EVALUATION
SECTION 1
BACKGROUND
Industrial utilization of coal could be enhanced provided tech-
nology is made available to control air pollution emissions in an economi-
cally and environmentally acceptable manner. Control technologies to
consider for industrial boilers include:
Use of compliance coals
Flue-gas cleanup
Combustion-system operational modifications
Combustion-system design modifications
Use of treated or modified coals.
Depending on the specific stoker-boiler design and local air pollution
regulation, one or more of these control technologies may be required to
operate a boiler within the standards.
Data collected from industrial stoker boilers to evaluate the
effect of these control technologies on their performance and emissions
are limited. Before improvements in stoker firing can be achieved,
baseline performance and emission level data must be generated so that
pollution control potential can be accurately assessed.
II-l
-------
SECTION 2
OBJECTIVE AND SCOPE
The overall objective of the Phase II program was to identify
and evaluate potential concepts for control of emissions for full-scale
industrial stoker boilers. For this phase of the program, control
strategies were limited to operational and minor design modification
of the Battelle 8 M!«Tth steamplant (25,000 Ib steam/hr) spreader
stoker boiler and to firing of low pollution coals (either naturally
occuring or those that had been chemically or physically treated).
Research on flue gas clean-up equipment was not within the scope of
this program.
The operating variables that were investigated for system
characterization studies included:
excess air levels
overfire air rate
fuel-bed depth
boiler load
fly ash reinjaction
0 coal types (including treated coals).
Combustion design modifications were considered where the system charac-
terization studies identified a need and extensive modification of the
boiler was not necessary.
II-2
-------
SECTION 3
PLAN OF EXPERIMENTAL INVESTIGATION
Phase II experiments focused on assessing emissions for a
variety of stoker-boiler system operating variables. The experiments
were conducted in a Battelle steam plant 8 MW _ spreader-stoker,
th
boiler. Seven coals were fired in addition to the limestone/coal
fuel pellets. Particulate loading, S0~, NO, CO, particulate, and
smoke emission levels, plus a limited number of POM measurements,
were obtained.
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. This boiler was designed for
a traveling-grate stoker and initially fired on gas and oil. During
conversion to coal, the boiler was modified for spreader-stoker firing.
The combustion particulate control system consists of two cyclone dust
collectors. The ash from the grate discharge, grate siftings, and
cyclone dust collectors is discharged into a hopper, which is periodi-
cally emptied into dump trucks. Figure II-l is a schematic showing the
relative locations of the system components.
Stoker
The stoker is a 2A-Hoffman "Firite" Pulsating Ash Discharge
(PAD) type with a nominal feed-rate of about 910 kg/hr. This spreader
stoker spreads 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.
II-3
-------
TREATING
AGENTS
WATER
WATER
TREATMENT
MAKEUP
WATER
SPENT
TREATING
AGENTS
COAL
AIR
OVERFIRE AIR JETS-
J I
SLOWDOWN
STEAM
LOOP
GAS SAMPLING
PORT
REINJECTION
STOKER
PARTICULATE
SAMPLE PORTS
PRIMARY SECONDARY
CYCLONE CYCLONE
A_
FLUE
GAS
w
SETTLINGS
FLY ASH
OVERFIRE AIR
JETS
UNDER FIRE
FORCED DRAFT
GRATE
DISCHARGE
GRATE
SIFTINGS
INDUCED
FAN
ASH
DISCHARGE
FIGURE II-l. SCHEMATIC OF THE 8-MW STOKER-BOILER FACILITY
II-4
-------
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 conse-
quence of the underthrow principle used exclusively by Hoffman. Distri-
bution is modified by changing the pitch angle of the circular tray.
The grate surface area is 5 m2 (2.4 m by 2.1 m). Incorporated
in this area are 18 individual grates. The six grated positioned in
the center are 23 cm by 99 cm, while the remaining twelve grates located
on the sides of the boiler are 30 cm by 99 cm in dimension. The 30 cm-
wide grates have four rows of tubes (16 in each row) and each tube has
five 0.635 cm flower-like openings on the grate surface to distribute
the air. The 23 cm grates have three 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 of
ash 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 speed changer adjusts for the
optimum speed, after which the speed setting remains fixed. This
adjustment 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. Venturi type nozzles
(Hoffman design) entrain the fly ash using air discharged from the over-
fire 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 opera-
tion is used.
II-5
-------
Fly ash is reinjected through two nozzles located at the rear
of the boiler. These nozzles are approximately 30 cm from the sidewalls
and 25 cm above the grate.
Overfire Air Jets
There are four overfire air jets in the rear and five in the
front of the furnace. These are located approximately 25 cm above the
grate. The maximum overfire air pressure attainable is approximately
40 cm. W. G. The various overfire air jets have individual and branch
dampers adjustable by hand. The minimum air pressure that will provide
turbulence in the furnace is selected.
Boiler
The water tube gas/oil fired boiler is a Keeler Type MKB manu-
factured 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 860 kPa (125 psig) with an ASME
n
Power Test Code heating surface of 383 m utilizing a furnace volume of
3
40 m . The boiler can operate at a capacity of 8 MW , continuous, with
<3 tO.
a heat release rate of 845,615 kJ/m /hr (22,700 Btu/ft3/hr).
Sampling Ports
Figure II-l shows the relative location of the particulate and
gas emissions sampling ports. Particulate sampling ports were located
in the breaching between the induced draft fan and the stack. This
specific location of the breaching is common to the three Battelle boilers,
the other two of which fire gas only. Most tests were conducted when
only the stoker-boiler was on-line. On occasion, one of the gas-fired
boilers was in operation, but the particulate loading contribution
from this boiler is negligible in comparison to that of the stoker-boiler.
The gas sampling port was located at the boiler outlet immediately
upstream of the dust collector and induced-draft fan.
II-6
-------
Coal Handling System
Coal is initially fed into a concrete pit from a front-end
loader or dumped by truck. A vibrating feeder inside the pit delivers
coal to the bottom of a bucket elevator. A flop-gate at the top of
the 15-m bucket elevator diverts the coal into a silo (for storage) or
onto a screw conveyor. The screw conveyor empties into a day hopper
located directly above the stoker. Coal is gravity fed through a conical
chute to the two stokers.
Throughout the coal-handling system, the coal can encounter
significant mechanical stresses that may create an excessive amount of
fines.
COAL PROPERTIES
Eight coals were fired in the steamplant boiler during Phase II
experiments, including the limestone/coal fuel pellet (Ca/S molar ratio
of 7). Table II-l lists ultimate and proximate analyses, heating values,
.ash-fusion temperature, and size consist for these coals. The size
consist was determined before the coal was introduced into the handling
system. These coals were selected to provide a range of fuels suitable
for industrial stoker-boiler firing. Based on Phase I, the fuel
pellet was selected as the treated f.uel with the greatest potential to
become a viable SO., control in the near future.
The Ca/S molar ratio of 1 was selected to assess whether the
concept of a limestone/pellet had any potential as an S0? control in an
industrial system without optimizing the pellet. Phase I development
of the fuel pellet concept indicated that optimization of the pellet in
terms of
binder selection
coal and limestone size distribution
size and shape
may allow for the reduction of Ca/S ratio without significant reduction in
S02 retention. Program plans were to optimize the pellet as part of Phase
III, provided the firing of the Ca/S = 7 pellet showed promise.
II-7
-------
TABLE II-l. PROPERTIES OF COALS FIRED IN PHASE II
Proximate Analysis
(As Received), %
M
M
1
CO
Coal Type
Low-S Ohio
Mediurn-S Kentucky
Stoker-grade,
washed Ohio
Stoker-grade,
unwashed Ohio
Rim-of-the Mine,
washed Ohio
Low-S Kentucky
Illinois No. 6
Volatiles
33
38
37
36
38
36.
37,
.12
.20
.84
.34
.89
.87
.30
Fixed
Carbon
47.59
53.15
42.58
40.95
43.47
52.83
39.97
Ultimate
Analysis
(As-Received), !£
Oxygen
Ash
9.60
4.95
10.34
18.50
8.60
7.71
8.63
Moisture
9
3
9
4
9
2
14
.69
.70
.24
.21
.04
.59
.10
Carbon
64
76
63
59
65
. 75
60
.81
.96
.40
.98
.02
.42
.52
Hydrogen
4.26
5.15
4.54
4.23
4.56
5.03
4.23
Nitrogen
1
1
1
1
0
1
1
.26
.26
.21
.09
.98
.53
.10
Chlorine
0.05
0.14
0.06
0.06
0.04
0.11
0.14
Sulfur
0
1
3
3
3.
0,
3.
.70
.38
.00
.94
.19
.89
.46
(Difference)
9
6
8
7
8
6
7
.63
.46
.21
.99
.57
.72
.82
Heating
Value
K.I/8
26
32,
26
25
27
31
25
.7
.1
.6
.6
.3
.3
.3
Ash-Fus Ion
Temperature, F
(Initial
Reducing
NA
1160
1120
1140
1100
1360
1100
Deformation)
Oxidizing
1480*
1340
1350
1360
1320
1480
1270
Fuel Size
Consist,
% less
than 3mm
7
12
6
17
16
12
6
Limes tone/high-
Sulfur coal pellet
39.20 NA 45.60 12.60 29.70 1.54
0.48
1.96
8.64
9.9
* NA - Not Available.
-------
PELLET PRODUCTION
Approximately 100 tons of limestone/coal pellets with a Ca/S
molar ratio of 7 were produced for evaluation in the steamplant spreader
stoker boiler. Figure II-2 is a flow diagram for the pellet production
process. There are many components in this process with the major element
being the pellet mill.
Pellet formulation was the same as in Phase I, namely
Illinois No. 6 coal (48 percent)
Piqua limestone (47 percent)
Cement binder (5 percent)
Material Preparation
As received, the limestone and cement required no further pre-
paration. The coal required size reduction from -2.5 cm to -20 mesh
which was accomplished in a Model 2W Mikropulverizer hammermill. (This
mill processed about 400 kg/hr of raw coal and was operated about 12
hrs/day.)
Processing Operations
In addition to coal grinding and materials handling, the pellet
process consisted of the following four separate operations:
(1) Blending
(2) Pelleting
(3) Size Classification
(4) Drying
Blending. The coal, limestone, cement, and water were thoroughly
3
blended prior to introduction into the pellet mill. A 0.2 m Voeller
batch mixer, typically used for batch mixing in the glass industry,
was used. The capacity of this mixer, 800 kg/hr, was the limiting
factor in the pellet production process.
II-9
-------
I
M
O
RECYCLE FINES
1/2 HP FEED
AUGER
30 HP
PELLET MILL
PRODUCT STREAM
2000 Ibi/hi
DUMPS TER
r-r
1
vie
s
VIBRAIINa
SCREEN
FIGURE II-2. PELLET PRODUCTION PROCESS FLOW DIAGRAM
-------
Pelleting. Pellets, nominally 1.3 cm in diameter and 1.9 cm
long, were produced with a California Pellet Mill (22 kW Master Model).
There were a number of operational problems with this mill, primarily
associated with roller damage due to coal influx into the bearing
lubricant. Once these were resolved, the mill produced pellets at
a rate of about 1200 Kg/hr.
Size Classification. The mill output contained about 10 percent
fines along with the pellets. These fines were removed with a model ME
38 Midwestern Screen Separator and recycled back to the mill.
Drying. After separation, the pellets required drying, which,
because of the relatively small batch quantities could be done in air.
However, on a larger scale production or during inclement weather, air
drying might be inadequate and another method of drying would be required.
Materials Handling. The pellet production operation was
highly labor intensive requiring six men for operation. Most of this
labor was associated with materials handling, especially in the blending
operation. Larger scale operations with automation would significantly
reduce the labor intensity.
Summary. The 100 tons of pellets were produced for about $600/
ton. A large-scale, mine-mouth operation would reduce these costs signi-
ficantly to about $15/ton. However, we experienced significant mainten-
ance and "downtime" that must be minimized for this operation to be
economically attractive.
Sampling Equipment and Procedures
Sampling equipment and procedures were similar to those of the
Phase I experiments with one exception. Stack gases were transported
through a heated sampling line, about 6-m long, before filtering and
moisture removal rather than filtering the particulate and condensing
the moisture in a trap adjacent to the stack. Analysis procedures were
also similar.
11-11
-------
Mass-Rate Determinations
Bell mouth orifices were installed on inlets of both force-
draft fans, one supplying the undergrate air and the other supplying
overfire air. These orifices provided a measure of the total air-flow
rate into the boiler. Steamflow rates were also determined from an
orifice. The coal feed rate was determined by measuring the time
required to displaced a fixed volume of coal (in the day hopper). Fly
ash reinjection rates were determined by collecting the reinjection
fly ash in drums and weighing the contents for a specific collection
time.
11-12
-------
SECTION 4
EXPERIMENTAL RESULTS
The performance of the stoker boiler and the emissions that
it generates depend on a number of factors, including the combustion
operating parameters and the properties of the coal. Table II-2 out-
lines the operating variables investigated during this Phase II effort.
Table II-3 summarizes the gaseous and particulate emissions generated
during the firing of the eight different coals for the conditions
listed. Because there are few measurements to characterize the fuel
bed other than visual observations, qualitative observations are
necessary to describe the fuel-bed combustion. Accordingly, such
observations are included with the experimental measurements to interpret
the data.
In the subsections below, a general discussion of the emissions
generated from the steam plant stoker is given first. Then, the effect
of stoker boiler operating variables on emissions and boiler performance
is discussed. Finally, based upon observations of the stoker-boiler
system and analysis of the data, combustion modifications that have the
potential to reduce emissions and improve boiler performance are identified.
STOKER-BOILER EMISSIONS
The criteria emissions are discussed in terms of coal properties.
Nitrogen Oxides
Nitrogen oxide emissions are largely a function of the nitrogen
content of the coal. Unfortunately, coal cleaning or treatment techniques
do not remove nitrogen; in fact as a result of reducing the ash content, the
percentage of nitrogen in the coal appears to be increased. This
emission can be controlled by operating a low excess air levels.
11-13
-------
TABLE II-2. OPERATING VARIABLES STUDIED FOR SYSTEM
CHARACTERIZATION IN PHASE II
a 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 RE INJECTION
- Maximum
- Partial
None
COAL TYPES
- Size distribution
Ash content
- Ash-fusion properties
- Sulfur levels
- Treated coal: limestone/coal pellet
11-14
-------
TABLE II-3.
EMISSION SUMMARY DATA OF 8 MW , BOILER EXPERIMENTS
tn
n
H
1
1 l
t_n
No.
SP-1
SP-2
SP-3
SP-4
SP-5
SP-6
SP-7
SP-8
SP-9
SP-10
SP-11
SP-12
SP-13
SP-14
SP-15
SP-16
SP-17
SP-18
SP-19
Overflre
Air /
Total
Coal Type I FullrLoad Air Rat:
Low-S Ohio
Low-S Ohio
Medlum-S Kentucky
Low-S Ohio
Low-S Ohio
Stoker-grade,
washed Ohio
Stoker-grade,
unwashed Ohio
Run-of-the-mine,
washed Ohio
Stoker-grade
washed Ohio
Run-of-the-mlne,
washed Ohio
Low-S Kentucky
Low-S Kentucky
Run-of-the-mine,
washed Ohio
Run-of-the-mine,
washed Ohio
Stoker-grade,
washed Ohio
Stoker-grade,
washed Ohio
Illinois No. 6
Limestone/hlgh-S
coal pellet
Limes tone/high-S
coal pellet
86
58
84
84
38
86
84
84
90
88
90
88
80
82
80
84
68
64
60
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
0.16
i° °x'
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
11.8
Flue Gas Composition
C02,
X
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
9.5
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
1000
S02,
ppm
360
350
700
320
190
1800
2200
1800
1400
1900
340
330
1500
1300
1350
1700
1900
560
515
Smoke
NO, Opacity,
ppm X
230
185
280
250
230
240
245
250
235
250
240
230
260
230
145
140
10
4-5
7
4
5
6
7-11
6-9
6
6
4
2
3
4
2
3
5
22
25
S02 at 3Z
02, ppm
Fuel S
CO at 3X Computed Measured Emitted,
Oj, ppm into System Emissions X
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
2010
625
625
1010
625
625
2635
3650
2750
2635
2750
665
665
2750
2750
2635
2635
3190
4035
4035
505
470
910
440
375
2475
3635
2740
2192
2825
490
488
2310
2092
2230
2602
2974
1057
1035
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
25.6
NO at 3X
Computed
into System
2565
2565
2110
2565
2565
2430
2310
1930
2430
1930
2615
2615
1930.
1930
2430
2430
2320
2255
2255
Oj, ppm
Measured
Emissions
320
365
385
1
413
350
375
365
360
350
385
386
380
398
360
274
281
X 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
12.5
Partlcu-
lates
ng/J
110
77
110
73
65
190
240
220
290
73
56
150
170
160
180
610
960
POM
g/N»3
_
^ _
14
13
22
24
22
22
12
21
174
-------
Table II-4 gives the conversion of fuel nitrogen to NO for
coals fired in both Phases I and II. The conversion was based on the
assumption that no thermal NO was present. For the steamplant boiler,
the fuel nitrogen conversion ranged from 12 to 20 percent with an average
of 15 percent. The rate of fuel N conversion for the model spreader
runs ranged from 8 to 20 percent with an average of about 12 percent.
This is a. somewhat lower rate of fuel N conversion than observed in
the steamplant stoker. This difference may be attributable to the
higher percentage of bed burning in the model spreader than in steam-
plant stoker and also to the higher bed temperatures in the steamplant
stoker.
The rate of fuel nitrogen conversion for the underfeed stoker
is even lower, averaging about 9 percent. This extremely low rate of
fuel N conversion may be attributed to the inherent staged combustion in
the operation of an underfeed stoker and the low excess air combustion
occurring in the bed.
Sulfur Oxides
Sulfur oxide emissions are largely a function of the sulfur
content of the fuel. Sulfur is captured in the ash as a sulfate or
emitted from the stack as S0?, plus some small fraction (generally
only a few percent) as SO,,. Because the fate of sulfur within a stoker
boiler system is fairly well bounded, one might expect reasonable
complete sulfur balances. However, review of data in Table II-5,
and earlier work by BCL , Gronhovd , and Maloney ,
indicates that good sulfur balances from coal combustion are difficult
to achieve. This is because accurate mass balances must be made through-
out the entire system. From a practical sense, only estimates of mass
balances can be reasonably achieved. Furthermore, sulfur balances
are determined from analysis of relatively small samples that may not
be representative of the total mass in question.
The scatter of the stoker data, as shown in Table II-5, is
similar to that obtained by other researchers and reflects a general pro-
blem of achieving sulfur balances. Of primary interest is determining
the amount of sulfur that can be retained with the ash, thus reducing
the amount of SO,., emitted. Phase I results, and those of Gronhovd and
II-16
-------
TABLE II-4. CONVERSION OF FUEL NITROGEN
TO NO FOR COALS FIRED IN
PHASES I AND II
Nitrogen Content
Coal MAF, %
(a) Steamplant Stoker
Low-S Kentucky
Low-S Kentucky
Low-S Ohio
Low-S Ohio
Stoker Grade, Washed, Ohio
Stoker Grade, Washed, Ohio
Stoker Grade, Washed, Ohio
Stoker Grade, Washed, Ohio
Stoker Grade, Unwashed, Ohio
Medium- S Kentucky
Illinois No. 6
ROM washed, Ohio
ROM washed, Ohio
ROM washed, Ohio
L/C Pellet, Ca/S=7
L/C Pellet, Ca/S=7
1.7
1.7
1.6
1.6
1.5
1.4
1.4
1.4
1.2
1.2
1.2
1.1
1.1
Fuel N
Conversion, %
13.4
13.8
12.5
14.2
15.8
15.4
15.6
16.4
17.9
10
15.5
18.1
18.1
20.0
12.5
12.2
11-17
-------
TABLE II-4. (continued)
(b) Model Spreader
Western Bituminous
Western Bituminous
Western Bituminous
Low-S Kentucky
Low-S Kentucky
Low-S Kentucky
Low Volatile W. Virginia
Low Volatile W. Virginia
Western Kentucky, Unwashed
Hydro thermal
Eastern Kentucky
Eastern Kentucky
Eastern Kentucky
Illinois No. 6
Illinois No. 6
Illinois No. 6
L/C Pellets, Ca/S=4
L/C Pellets, Ca/S=7
L/C Pellets, Ca/S=7
L/C Pellets, Ca/S=7
L/C Pellets, Ca/S=7
Lignite
Lignite
(c) Underfeed Stoker
E. Kentucky
Illinois No. 6
Lignite
Western Subbituminous
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.6
1.5
1.4
1.4
1.4
1.4
1.2
1.2
1.2
1.2
1.1
1.1
1.1
1.1
1.0
1.0
1.4
1.2
1.0
.9
11
9
8
18
12
11
13
10
8
10
15
10
7
15
14
17
13
12
12
8
11
14
20
10
9
8
9
-------
Maloney, indicate that significant amounts of sulfur can be retained in
the ash because of the presence of alkaline mineral compounds in the
coal. However, Maloney's data, as well as data generated in Phases I
and II, suggest that lesser amounts of sulfur may be retained in the ash,
even without the presence of alkaline material.
As indicated in Table II-5, for Runs SP-1 through SP-17, all
eastern coals, the fuel sulfur emitted as SO ranged from 60 to 103
percent, with an average of 85 percent. This compares with Maloney's
data, which indicated that for stoker firing eastern coals, the average
fuel sulfur emitted was 90 percent. Furthermore, the data in
Table II-5 indicate the lowest value of fuel sulfur emitted as SO-
occurred at low load operation while higher values occurred at higher
boiler loads, confirming Maloney's observation.
However, for these same runs, the average amount of fuel
sulfur retained in the ash was about 2 percent. The sulfur retention
in the ash was based on the sulfur content of the bottom ash. The
bottom ash provides a fairly accurate indication of the residual ash
constituents as it accounts for about 80 percent of the coal ash input,
as shown in Table II-6. Table II-6 indicates that the other major
deposit of ash is from the fly ash reinjection hopper. A comparison
of the CHNS and ash analyses from the ash discharge (bottom ash) and
fly ash reinjection hoppers, shown in Table II-7, indicates a reasonably
close agreement; sufficiently close, that using the bottom ash as
representative of all the coal is reasonably accurate. Intuitively,
it would appear that the sulfur analysis of the bed ash would be
reasonably accurate making the S02 measurement either from the monitor
or the procedure suspect. Calibration of the S02 measurement against
Method 6 wet chemistry techniques, indicate that the S02 monitor
did provide reasonably accurate results.
The data in Table II-5 indicate sulfur balances for runs with
higher sulfur coals were superior to those of runs with the lower sulfur
coals. This observation suggests that there may be some fixed amount
of sulfur that cannot be accounted for; for example, fuel sulfur in the
coal analyses or some fixed amount of SO that is removed in the sampling
system.
11-18
-------
TABLE II- 5. SULFUR BALANCES FOR STEAM-
PLANT STOKER RUNS
Run No.
SP-1
SP-2
SP-3
SP-4
SP-5
SP-6
SP-7
SP-8
SP-9
SP-10
SP-L1
SP-12
SP-13
SP-14
SP-15
SP-16
SP-17
SP-18
SP-1 9
Computed As
Fuel S
524.6
524.6
860
524.6
524.6
2253.2
3078.8
2339.2
2253.2
2339.2
567.6
567.6
2339.2
2339.2
2253.2
2253.2
2730.5
3956
3956
S02 (ng/J)
Measured As
so2
425.7.
395.6
774
369.8
318.2
2115.6
3065.9
2330.6
1874.8
2403.7
421.4
417.1
1965.1
1780.2
1909.2
2223.1
2008.1
946
1032
Retained
in Bedash
8.6
4.3
5.59
4.3
25.8
-._
21.5
8.6
73.1
4.3
4.3
90.3
12.9
86
430
2365
2021
Unaccounted, %
17
23
9
29
39
5
I
0
17
-6
25
25
12
23
12
0
10
16
23
11-19
-------
TABLE II-6. DISTRIBUTION OF COAL ASH WITHIN
THE STOKER-BOILER SYSTEM
Ash Sample Location
Ash Hopper
Grate Sittings
Flyash Reinjection
Cyclone
( I Method 5 Catch
jjj Total Measured
O
Calc. From Coal Feed
Rate and Analysis
Ash Sample Location
Ash Hopper
Grate Sittings
Flyash Reinjection
Cyclone
Method 5 Catch
Total Measured
Calc. From Coal Feed
Snmp Le
Col Section
Rate, Kg/hr
52.
0.
12.
1.
1.
68.
16
32
7
27
81
5
Sample
Collection
Rate, Kg/hr
56.
0.
11.
0.
4.
73.
25
41
8
45
99
94
RUN
Analysi
Sample
C
7.1
7.0
59
44.4
12.1
RUN
SP-12
s.of
, ^
Ash
93
93
25
55
84
SP-15
Analysis of
Sample, %
C
3.2
2.1
30.2
18.8
8.8
Ash
98
99
71
76
23
RUN SP-13
Ash
Rate,
Kg/hr
48.08
0.32
3.2
0.68
1.54
53.5
77.1
Distribu-
tion of
Ash, %
89
1
6
1
3
100
Sample
Collection
Rate, Kg/hr
109.8
0.45
31.75
0.18
4.63
146.97
Analysis of
Sample, 7.
C Ash
17.1 80
6.4 95
29 71
24 74
10 24
Ash
Rate
Kg/hr
87.54
0.45
22.68
0.136
1.13
112.04
95.26
RUN SP-17
Ash
Rate
Kg/hr
55.34
0.41
8.16
0.36
1.09
65.32
108.86
Distribu-
tion of
Ash, %
83
1
13
1
2
Sample
Collection
Rate, Kg/hr
94.35
0.91
14.52
0.68
NA
110.68
Analysis of
Sample, %
C Ash
70
95
70
70
Ash
Rate
Kg/hr
66.23
0.91
9.53
0.45
78.02
85.73
Distribu-
tion of
Ash, %
79
0
20
0
1
RUN SP-18
Distribu- Sample Analysis of Ash Distribu-
tion of Collection Sample, % Rate tion of
Ash, % Rate, Kg/hr C Ash Kg/hr Ash, %
85 884.52 6.4 86 762.5 78
1 1.36 5.1 93 1.36
1* 228.16 6.1 88 201.85 21
1 0.59 9.2 83 0.45
13.61 6.8 85 11.34 1
1128.1 972.97
1011.53
Rate and Analysis
-------
I
t-0
TABLE 11-7. CHNS AND ASH ANALYSES OF BED ASH,
REINJECTED FLY ASH, CYCLONE CATCH,
AND FILLER CATCH, WT PERCENT
Bed Ash
Run No.
SP-1
SP-2
SP-3
SP-4
SP-5
SP-6
SP-7
SP-8
SP-9
SP-10
SP-11
SP-12
SP-13
SP-14
SP-15
SP-16
SP-17
SP-1 8
SP-19
C
?2. 7
10.1
1.7
7.7
10.5
3.9
1.7
9.2
10.6
7.1
17.1
3.7
19.2
1.8
27.6
6.1
7.6
H
0.2
0.1
0.1
0.1
0.2
0.1
0.1 .
0.4
0.2
0.1
0.2
0.4
0.2
0.2
0.2
0.3
0.1
N
0.1
<0.1
<0.1
0.1
0.1
<0.1
<0.1
0.1
0.1
<0.1
0.1
0.1
0.1
0.2
0.1
0.2
S
0.1
0.04
0.17
0.03
0.37.
0.37
0.10
1.14
0.07
0.1
1.40
0.17
1.09
0.18
6.34
2.33
2.16
Ash
75.3
88.8
98.6
~
90.9
90.3
96.3
98.6
90.6
89.9
92.6
79.8
96.8
82.4
98.5
70.0
86.4
93.5
Relnjected Flyash
C
31.6
48.5
41.3
65.2
46.3
51.7
48.8
50.7
50.9
59.0
29.4
30.2
6.1
13.7
H
0.2
0.4
0.6
0.3
0.5
0.7
0.5
0.6
0.7
__
0.3
0.3
0.2
0.1
0.5
N
<0.1
0.1
0.2
0.2
0.3
0.3
0.2
0.2
0.2
0.1
0.2
0.1
<0.1
0.1
s
0.21
0.19
5.32
0.88
0.53
1.29
1.45
1.33
1.26
0.4
0.72
0.68
2.42
2.45
Ash
69.5
49.2
46.6
21.5
39.0
37.9
42.9
37.6
41.6
--
24.9
71.4
71.4
88.4
77.1
C
40.2
30.9
33.7
--
26.9
25.9
23.3
31.5
34.2
47.3
44.4
20.9
18.8
9.2
11.7
Cyclone Catch
H
0.4
0.3
0.3
0.6
0.3
0.3
0.3
0,7
0.4
0.3
0.2
0.2
0.2
0.3
N
0.2
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.3
0.1
0.1
0.1
0.1
0.1
s
0.23
0.28
0.53
0.28
0.61
0.95
0.84
1.07
0.41
0.5
0.95
0.77
2.81
2.82
Ash
61.4
64.1
59.9
62.2
71.9
70.9
62.8
56.6
52.4
54.8
70.9
76.2
82.5
78.3
C
11.3
18.7
16.1
16.0
37.6
4.83
10.3
8.73
5.1
14.6
12.1
2.26
3.23
3.06
3.08
6.8
9.4
Filter Catch
11 N
2.2 0.28
2.9 0.43
2.3 0.37
2.3 0.30
4.05 0.87
1.7 0.12
4.32 0.31
6.07 0.21
3.4 <0.1
1.6 <0.1
1.9 <0.1
2.44<0.1
3.4 <0.1
2.2 <0.1
1.6 <0.1
0.3 0.1
0.4 0.2
S
3.1
3.5
7.4
4.11
3.76
12.1
12.2
12.33
15.0
7.2
11.9
17.4
7.7
8.1
6.1
7.7
15.9
Ash
66
12
49
54
35
28
24
23
22.4
81.6
83.7
*
A
*
*
85.4
81.8
* Deposit could not be removed from filter.
-------
CO
CO levels were less than 70 ppm for all the conventional coals
fired. The use of overfire air provided an adequate control.
Particulate Loading
The data in Table II-8 suggest that for conventional coals,
the ash content of the coal is not the only, or even the dominant,
factor contributing to the particulate load for industrial stoker
boilers, as was suggested by the EPA emission factor publication.
This observation was also made in an earlier EPA study on underfeed
stokers. It appears that the amount of coal fines introduced
into the boiler system would be a superior indicator of particulate
loading than ash content. At least it should be considered along
with the ash content as an indicator of particulate loading.
As seen in Table II-8, the limestone/coal fuel pellet generated
the highest particulate loading. This fuel contained a relatively
high amount of ash. Although the pellets contained small amounts of
fines as produced, they degraded when transported through the fuel-
handling system. As a result, 50 percent fines were introduced into
the boiler. This, coupled with the high ash content, resulted in the
unusually high particulate loading.
POM
POM were collected for those selected runs listed in Table II-l.
POM are related to unburned carbonaceous materials, particulate
loadings, percent of carbon deposited on Method 5 filters, CO levels,
and smoke opacity; these data are included in Table II-9. For the
conventional coals (Runs SP 9-16), POM levels were similar even though
carbon and particulate loadings were not. POM levels ranged from
o
13 to 24 yg/Nm and were somewhat lower than levels generated by the
model spreader. Also, the levels of POM generaged by the steamplant
stoker were only somewhat higher than those observed from firing a 500 kW
packaged boiler on natural gas, distillate oil, and residual oil.
This was a rather surprising result in that coal combustion generates
higher POM levels than oil or gas combuation. Other factors must
11-22
-------
TABLE II-8. COAL PROPERTIES AFFECTING PARTICIPATE LOADING
Coal Type
Optimum
02> %
Particulate
Loading
ng/J
Ash
Ash-Fusion Temperature
Initial Deformation,
Oxidizing
Fuel Size Consist
(percent less
than 3 mm)
M
I
Low-Sulfur, Ohio
Medium-Sulfur, Kentucky
Stoker-Grade Washed,
Ohio
Stoker-Grade Unwashed,
Ohio
Run of the Mine,
Washed, Ohio
Low-Sulfur, Kentucky
Illinois No. 6
Limestone/Coal Pellet
7.7
7.0
8.0
10.8
9.0
73
110
190
240
9.6
5.0
10.3
18.5
8.6
1480
1315
1370
1370
1315
7
12
6
17
16
8.5
9.5
11.2
73
610
7.7
8.6
45.6
1480
1260
1480
12
6
50
(entering boiler)
-------
TABLE II-9. LISTING OF POM LOADINGS, PARTICULATE LOADINGS
CARBON PARTICULATE, CO EMISSION LEVEL, AND
SMOKE OPACITY FOR SELECTED RUNS
Run No.
SP-9
SP-10
SP-11
SP-12
SP-13
SP-15
SP-16
SP-18
SP-19
POM. Particulate
Loading, Loading,
Coal yg/Nm ng/J
Stoker- grade,
washed Ohio
Run- o f - the-Mine
washed Ohio
Low-S Kentucky
Low-S Kentucky
Run-o f- the-Mine ,
washed Ohio
Stoker-grade,
washed Ohio
Stoker-grade,
washed Ohio
L imes t one / hi gh- S
coal pellet
Limestone/high-S
coal pellet
14
13
22
24
22
22
12
21
174
220
290
73
56
150
160
180
610
960
Percent C C0>
on filter ppm Smoke Opacity, %
10.3
9.2
14.6
12.1
3.6
4.8
6.8
9.4
80
60
60
40
60
50
30
950
2000
6
6
4
2
3
2
3
22
25
11-24
-------
be considered, however, such as size, design, and operation of the equip-
ment. The POM loading of Run SP-19 was significantly higher than those
of the other runs. With the exception of Run SP-18, the CO level, smoke
opacity and carbon and particulate loading of Run SP-19 were significantly
higher than the other runs, as expected. In Run SP-18, the CO level,
smoke opacity, carbon and particulate loading were not appreciably
different than those in Run SP-19, yet the POM levels are significantly
less. These results suggest that perhaps another factor may have a
dominant effect on POM levels. It is difficult to generalize becuase
the formation of POM is not fully understood nor are the sampling and
analytical techniques completely developed. This creates some diffi-
culties in interpreting the data.
Tables 11-10 and 11-11 give POM species quantification. The
differences in species identification in the two quantification tables
represent changes in analytical procedures. The total POM levels are
not affected by these changes. The procedure used to generate data
in Table 11-11 improves the resolution of high-molecular weight species«,
EFFECT OF OPERATING VARIABLES ON
EMISSIONS AND BOILER PERFORMANCE
In general, those modifications in the operation of the stoker-
boiler system that reduce emissions also improve boiler efficiency.
Those modifications that improve the overall combustion, tend to decrease
emissions by improved control of the air and fuel distribution through-
out the stoker-boiler system. The operating variables that were inves-
tigated were listed in Table II-2 and are discussed below.
Excess Air
Operation. The limiting factor in determining the minimum
excess air level at which the stoker boiler could be fired was not high
emission levels but clinker formation on the fuel bed. This observation
/TT_O TT_Q\
is verified by other researchers. ' Infiltration into the
boilers and leakage around stoker grates may be significant, hence the
excess air level (generally measured in the stack) at which clinkers
are formed depends on the specific design and condition of the stoker-
boiler system, as well as the coal type and size.
11-25
-------
TABLE 11-10. POM QUANTIFICATION
NAS
Component Notation
Anthracene/Phenanthrene
Methyl anthracenes
Fluoranthene
Pyrene
Methyl Pyrene/Fluoranthene
Benzo (c ) phenanthrene ***
Chrysene/Benz (a) anthracene *
Methyl chrysenes *
7 , 12-Dimethylbenz (a) anthracene ****
Benzo Fluoranthenes **
Benz(a)pyrene ***
Benz(e)pyrene
Perylene
Methylbenzopyrenes
3-Methylcholanthrene ****
Indeno(l,2,3,-cd)pyrene *
Benzo (ghi)perylene
Dibenzo (a, h) anthracene ***
Diebenzo (c,g)carbazole ***
Dibenz(ai and ah)pyrenes ***
Coronene
TOTAL
SP-9
23.17
2.83
5.92
1.79
0.427
ND
0.247
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
34
SP-10
19.77
1.84
4.10
1.11
0.291
ND
0.313
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
27
SP-11
24.51
3.20
12.66
2.35
0.747
0.079
0.371
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
44
SP-12
24.30
3.56
15.12
4.45
1.58
0.258
1.04
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
50
11-26
-------
TABLE 11-11. PAH QUANTIFICATION
pg Total Sample
PAH
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Anthracene
Phenanthrene
Methyl Anthracenes
Fluoranthene
Pyrene
Methyl Pyrene /Fluor anthene
Ben'zo (c) phenanthrene
Chrysene
Benz (a) anthr acne
Benzofluoranthene's
Benz(a)pyrene
Benz(e)pyrene
Perylene
Indeno-pyrene
Benzo (ghi) perylene
Total
SP-13
12
0
2
1
0
18
3
6
3
0
0
0
0
*
*
»
.
.
B
.
»
6
946
51
31
717
1
75
04
24
135
261
399
154
SP-15
16
0
1
1
0
21
3
4
2
0
0
0
0
ND
ND
ND
ND
ND
ND
50
54
.3
.811
.86
.28
.785
.7
.40
.92
.43
.102
.112
.098
.074
ND
ND
ND
ND
ND
ND
SP-16
5
0
0
0
0
5
1
3
1
0
0
0
0
18
.27
.088
.647
.342
.165
.90
.45
.04
.36
.072
.106
.164
.078
ND
ND
ND
ND
ND
ND
SP-18
14
0
3
0
0
13
1
5
1
0
0
0
0
42
.1
.236
.71
.888
.566
.4
.93
.04
.36
.069
.125
.254
.070
ND
ND
ND
ND
ND
ND
SP-19
77
3
37
14
5
113
15
8
3
0
0
0
0
0
0
0
0
280
.1
.61
.3
.8
.79
.5
.00
.76
.223
.298
.669
.218
.165
.077
.133
.224
ND
ND
11-27
-------
Table II-8 indicates the optimum levels of excess air that were
determined for the various coals fired in the Battelle stoker boiler at
the same load. Optimum is defined as the lowest excess air that could
be continually maintained while providing an adequate margin of safety
before clinkers formed. For example, the low sulfur Ohio coal could be
fired as low as 6 percent 0~ before clinkers formed. From Table II-8,
it appears that the minimum excess air level that can be achieved depends
on more than one of the coal properties given and/or others not included
in the Table. For example, mineral analysis of the ash is used to pre-
dict the slagging potential of ash for pulverized coal firing but not
for stoker firing. It appears that this analysis would be useful in
determining the clinkering tendency. Table II-8 does suggest a weak
relationship between minimum excess air and ash content. This may be
attributed to the fact that higher ash coals form deeper beds than lower
ash coals. Bed ash is a thermal insulator between the burning bed and
the significantly cooler grate. Increasing the bed depth reduces
heat losses from the bed, allowing an increase in bed temperature and
clinker formation. In addition, because of the higher resistance of
deeper beds, more air is bypassed around the grate. Though the indi-
cated level of excess air in the stack may be the same, the amount
of excess air in the bed may be significantly less for deeper beds.
The Illinois No. 6 coal could not be fired at full load with-
out clinker formation even with the forced draft fan operating wide
open. The high excess air of the limestone/coal pellet firing can be
attributed to the abnormally deep and compact bed causing a significant
amount of air to leak around the grate bypassing the fuel bed. The
compact bed was caused both by the excessive amount of fines (50 percent)
and the vibratory action of the grate tending to pack the bed.
Emissions. Data from the Battelle steamplant boiler indicate
that low excess air operation can effectively control particulate
loading and NO. Particulate loadings were measured at 1) a baseline
condition in which the Battelle stoker operator established the normal
operating condition, and 2) the optimum excess air condition established
by the Battelle laboratory team. Both runs were made at approximately
the same boiler load firing the medium-sulfur Kentucky coal. The
results of the runs are tabulated below in Table 11-12.
11-28
-------
TABLE 11-12. COMPARISON OF PARTICIPATE LOADING
BETWEEN OPTIMAL AND NORMAL STOKER
OPERATION FIRING THE MEDIUM SULFUR
KENTUCKY COAL
Particulate
Excess Loading CO, (@3% 0 ),
Conditions 0 , % Air, % ng/J ppm
Baseline
Optimum
9
7
78
49
150
110
60
44
Reduction of excess air from 78 to 49 percent resulted in about a 30
percent reduction in particulate loading and about the same reduction
in CO level. The lower particulate loading can be attributed to the
higher fly-ash carryover. Also, the reduction of excess air increased
boiler efficiency by about 3 percent.
NO. Figures II-3 through II-6 show that reduction in excess
air from 100 percent to 60 percent can result in as much as a 20 to 25
percent reduction in NO levels. These observations are in general
agreement with data reported by KVB researchers. In pulverized
coal firing, low excess air is also used as an NO control, but the
X
excess air levels are considerably less than those from stoker firing.
However, the air availability to the stoker fuel bed may be similar
to that of the pulverized coal particle.
Two features stand out in examining the NO emissions: 1) NO
levels in stoker firing are substantially less than those in pulverized
coal firing and 2) the NO emissions are directly related to excess air
levels in and above the bed. Both of these observations are corrobated
by other investigators.
Figures II-3 through II-6 show a range of NO levels (3 percent 0_,
dry) of 340-460 ppm at excess air levels of 55-110 percent for coals
of 1 to 1.5 percent nitrogen. These NO levels compare with levels of
300-1000 ppm NO (3 percent 0_, dry) at excess air levels of 5-25 percent
X ^
for firing of pulverized coals ranging from 0.8 to 1.4 percent nitrogen.
The marked difference in the NO emissions from the two systems reflects
11-29
-------
460
CVJ
O
ro
o
5
420
8 380
e
ex
ex
340
300
Washed ROM Ohio
Fuel N = I %
"
50
60
70 80 90
Excess Air, percent
100
110
FIGURE II-3. NO EMISSION LEVELS AS A FUNCTION OF EXCESS AIR FOR
WASHED, RUN-OF-THE-MINE OHIO COAL
11-30
-------
460
CVJ
O
ro 420
_o
o
-------
460
CO
O
ro
o
o>
"o
0)
i_
o
o
E"
Q.
Q.
"55
420
380
340
O
WASHED, STOKER OHIO
Fuel N = 1.21 %
O
O
O
300
I
50
60
70 80 90
Excess Air, percent
100
10
FIGURE II-5. NO EMISSION LEVELS AS A FUNCTION OF EXCESS
AIR FOR WASHED, STOKER OHIO COAL
11-32
-------
460
55
ro
T3
(O
O
£
o
O
E"
Q.
Q.
0)
*
o
420
380
340
Low-S Kentucky
Fuel N= 1.53%
V V
V
V V
V
300
50
60
70 80 90
Excess Air, percent
100
IK
FIGURE II-6.
NO EMISSION LEVELS AS A FUNCTION OF EXCESS AIR FOR
LOW-SULFUR KENTUCKY COAL
11-33
-------
the nautre of stoker combustion. Stoker combustion is primarily bed com-
bustion and as such is similar (to a degree) to fluidized bed combustion.
NO levels are lower in bed combustion (stoker or FBC) because of the
significant reduction that occurs in the bed in spite of the overall
lean stoichiometry.
The relationships between NO and excess air are generally
X
similar for all coal firingstoker, FBC, pulverized coal, i.e., NO
X
increases with increasing fuel-N and with increasing excess air. Our
results corroborate the trend for increasing excess air but are not
as obvious relative to fuel-N content. This is in part due to the fact
that bed heights, bed temperatures, and overfire/bed air levels were
not controlled as closely as desired.
S0?. No appreciable influence on SO- emission levels was
observed for variations in excess air at full load operation. Maloney's
data indicated that there was a tendency for S0_ levels to be reduced
for higher excess air levels. This reduction was not significant
and there was sufficient scatter in the data to suggest a very weak
relationship between SO,, and excess air levels.
Overfire Air
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-sulfur Ohio coal. The total air flow rate was held constant. 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 (less than 20 ppm) upon deactivating the
jets. This can be attributed to several factors including lower
above bed temperatures, reduced oxygen levels, and perhaps heterogeneous
reduction reactions with smoke particles.
11-34
-------
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 0.18 would further reduce smoke,
but the reduction of underfire air through the bed would create severe
clinkering.
For constant total combustion air, increased overfire air has
two effects that can be illustrated by comparison of data of Runs SP-1
and SP-4 in Table 11-13.
TABLE 11-13. SMOKE, CO, AND PARTICULATE EMISSIONS
FOR TWO OVERFIRE/TOTAL AIR RATIOS
Total Carbon
Smoke Particulate Particulate
Overfire Air Opacity, CO, Loading, Loading,
Run Total Air Ratio percent ppm ng/J ng/J
SP-1
SP-4
0.14
0.18
10
4
70
40
110
73
12.1
11.7
The increased overfire air improved aerodynamic mixing above the bed
resulting in decreased emissions and smoke opacity. Furthermore
increased overfire air allows for decreased underrate air. Decreasing
undergrate air reduces the amout of fly ash carryover, as evidenced
by the 30 percent reduction in particulate loading. It is interesting
to note carbon loading remained virtually unchanged.
Fuel-Bed Depth
Experiments to explore the effects of different bed depths at
full load were conducted for test firing with several coals. It was
observed that for most coals, the stoker boiler could be fired only
over a narrow range of fuel bed depths and still maintain satisfactory
combustion. The stoker boiler was designed for about 8 to 10 cm deep
ash bed at the front end of the grate for a 5 to 10 percent ash coal.
11-35
-------
For ash bed depths less than 8 cm, full-load operation with complete
carbon burnout could not be achieved. Additionally, if the bed becomes
to shallow, there is danger of blowing the insulating ash layer off the
grate and exposing grate metal to the heat of the furnace. In ash beds
deeper than 15 to 20 cm, clinkers would form. Clinker formation in
deeper beds is attributed to 1) increased bed resistance allowing more
air to bypass the fuel bed and effectively lowering the air/fuel ratio
in the bed, and 2) the insulating effect of the ash resulting in higher
bed surface temperatures.
The medium sulfur Kentucky coal, because of its low ash content
and high heating value, could be more satisfactorily fired over a range
fuel bed depths than the other coals investigated. A shallow bed of
about 6 cm was found to be optimum for this coal in terms of responsive-
ness to load swings and low excess air firing. Run SP-3 was made at this
optimum condition, indicating that the medium sulfur coal could be fired
at 7 percent 0«the lowest CL level of any of the coals for continuous
satisfactory operation. Even though this coal had lower ash content,
the particulate loading was comparable to that of the low sulfur Ohio
coal in Run SP-1. This may be attributed to factors that include a
shallower more active bed, the friable nature of the Kentucky coal, the
greater amount of fines, and the higher sulfur content, leading to con-
densible sulfate formation in the fly ash.
As discussed earlier, increased bed depths require increased
excess air levels.
Boiler Load
The effect of boiler load on stoker performance and emission
levels was investigated in Runs SP-1, and SP-5, firing low sulfur Ohio
coal. The data are summarized in Table 11-14.
11-36
-------
TABLE 11-14. THE EFFECT OF BOILER LOAD ON
PERFORMANCE AND EMISSION LEVELS
Overfire Air Smoke Particulate
Load, Total Air Q^i CO, SO-, Opacity, Loading, Boiler
Run % Ratio % ppm ppm ppm ng/J Efficiency, %
SP-1
SP-2
SP-5
86
58
38
14
23
32
8.
7.
11.
0
5
8
50
60
70
505
470
375
10
5
5
110
77
65
78
76
75
11-37
-------
Performance. The excess 09 required at 58 and 86 percent loads
was about the same, but was significantly higher at 38 percent load.
The higher ) level at low load was required for two reasons: (1) to
avoid clinker formation and (2) to reduce CO and smoke. Observations
indicated that the amount of undergrate air required for satisfactory
operation was somewhat independent of boiler load. A certain level of
air flow through the grate had to be maintained to avoid clinker forma-
tion. Fow low loads, this resulted in reduced temperatures in the
combustion zone above the bed. To compensate for this effect, the
overfire air/total air flow ratio was increased to improve mixing to
control CO and smoke levels. For example, the overfire/total air ratio
of 0.15, CO levels were 150 ppm while smoke opacity was about 15 percent.
Increasing this ratio to 0.32 decreased CO levels to about 70 ppm and
smoke opacity to less than 10.
Boiler Efficiency. As indicated in Table 11-14 boiler effi-
ciencies were not significantly different for the various boiler loads.
Particulate Loading. A comparison of the particulate loadings
in Runs SP-1, SP-2, and SP-5 shows a trend of increasing particulates
with increasing boiler loads. This increase was attributed to the
increased rate of fly ash carryover because of increased velocities
(higher total flow rate and higher furnace temperature) at higher loads.
Smoke. 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 combus-
tion 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 (Run
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.
11-38
-------
SO.,. Bed temperatures, measured with an optical pyrometer about
5 cm above the bed in the central region of the grate, were around 1410
C at full load, 1340 C at partial load, and 1200 C at low load. The
reduction in bed temperature would seem to account for the decrease in
SO, emissions, that is greater sulfur capture for decreasing load, as
shown in Table 11-15.
TABLE 11-15.
so2 LEVELS AS A FUNCTION OF BOILER
LOAD AND FUEL BED TEMPERATURE
Run
SP-1
SP-2
SP-5
Load, %
86
58
38
SO (@3% 02),
ppm
505
470
375
Fuel S
Retained
in Bed,
percent
19
25
40
Bed
Temperature, C
1410
1340
1200
Analysis of the bed ash, however, does not verify the reduction in SO
observed in the stack gases.
In order to examine the relationsihp of bed temperature to
sulfur retention in the ash, the fuel bed was probed with a thermo-
couple, 2.5 cm below the surface. Figure II- 7 is a plan view of the
grate, indicating the approximate location of the temperature measure-
ments and the recorded temperatures for washed Ohio coal fired at 9.3
percent excess 0 and 80 percent load. (There was not sufficient room
in front of the boiler to insert a probe the full length of the grate.)
Furnace temperatures about 5 cm above the bed in the central region of
the grate, measured with an optical pyrometer, were around 1250 C to
1300 C and were in reasonably good agreement with the thermocouple mea-
surements. The region of lower temperature (1040 C) about 1.2 m from the
front wall was attributed to a cracked section of the grate that
admitted an additional amount of underfire air. It is expected that
bed temperatures would increase if the excess 0~ were reduced, the boiler
load increased, or if a higher heating value coal was fired.
11-39
-------
Boiler Backwall
o>
o
CO
2.4
J
i2'1
§ '-5
E
o 1.2
o '9
"w
b
.6
X
o
Q.
*^ .5m
- I290C I320C
- I260C 1270 C
I300C I040C
I290C I230C
1270 C 1260 C
- 1060 C 1100 C
a
Dnilar FV rtn + \*j/i i 1
O
a>
o
en
o
CD
FIGURE II-7. FUEL-BED TEMPERATURE DISTRIBUTION
11-40
-------
Fly Ash Reinjaction
The effect of fly ash reinjection on particulate emissions and
combustion efficiency was investigated for test firing with the following
coals:
0 Low-sulfur, washed Kentucky coal (Runs
SP-11 and SP-12)
High sulfur, washed run-of-the-mine Ohio
Coal (Runs SP-13 and SP-14)
High-sulfur, washed stoker-grade Ohio coal
(Runs SP-15 and SP-16)
Fly ash was reinjected during Runs SP-11, SP-14, and SP-16,
while no fly ash was reinjected during Runs 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-sulfur Ohio coals. Each
coal was fired with and without fly ash reinjection under approximately
the same boiler operating conditionsfull load (9525.6 to 9979.2 Kg/hr)
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 is, 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 11-16. These
results suggest that particulate loadings may be reduced by 10 to 25
percent by operating the spreader-stoker boiler without fly ash reinjec-
tion. This table also shows that 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 for those coals investigated.
The fly ash carbon loss measured as high as 5 percent for the
Battelle stoker-boiler facility. Although the potential existed to
increase boiler efficiency by reinjecting fly ash, the fly ash reinjection
system was not sufficiently effective to justify the increased maintenance.
Portions of the system have to be replaced periodically due to erosion
and higher particulate loading. Consequently, it was removed.
11-41
-------
TABLE 11-16. EFFECT OF FLY ASH REINJECTION
Particulate Loading,
Percent
Reduction in
Percent
Type of Coal
w/ Fly Ash w/o Fly Ash Particulate Increase
Reinjection Reinjection Loadings in C Loss
Kentucky
Run-of-the-Mine Ohio
Stoker-Grade Ohio
73.1
172
176.3
55.9
154.8
159.1
23.5
10.0*
9.8*
1.0
1.3
0.5
* About 65 percent of total fly ash was reinjected during the fly ash
reinjection runs.
Particulate emissions were significantly higher for the higher
sulfur coals. Analysis of the filter catch for these high-sulfur runs
indicates that as much as 25 percent of the particulate loading can
be attributed to condensible sulfates compared to less than 10 percent
for the low-sulfur coals. Dust collectors are ineffective in cap-
turing condensible sulfates, which pass through the collector as SO., and
then condense in the presence of water below 180 C.
The unwashed, stoker-grade, Ohio coal could not be fired satis-
factorily in the Battelle steam plant boiler. When firing this coal,
the boiler could not follow load swings. Additionally, clinkers were
formed at loads above 50 percent. This unsatisfactory performance is
attributed to the high ash content and high amount of coal fines.
Similarly, the Illinois No. 6 coal could not be fired above 60 percent
load without severe clinkering. This severe clinkering was attributed
to the excessive amount of fines, which caused the bed to mat, as well
as to the lower ash-fusion temperature.
Coal Types
The type of coal fired has a significant effect on the stoker-
boiler emissions and performance. The characterization of emissions
generated by the combustion of various coals provides a basis for selec-
ting coals to meet emission standards. With the exception of the unwashed,
11-42
-------
stoker grade, Ohio coal, and to a certain extent the Illinois coal, all
coals listed in Table II-3, including the limestone/coal fuel pellet,
could be fired satisfactorily in the Battelle steamplant boiler.
Conventional Coals. The Kentucky and low-sulfur Ohio coals
were superior to the high-sulfur, stoker-grade and run-of-the mine coals
in terms of S0«, NO, and particulate emissions, as previously discussed.
Furthermore, these coals could be fired under a wider range of operating
conditions. As shown in Table 11-17, the highest boiler efficiencies
were achieved with these coals.
TABLE 11-17. BOILER EFFICIENCIES FOR SELECTED COALS
Boiler
Coal Efficiency
Medium-S Kentucky 84
Low-S Ohio 76
Stoker-grade Ohio washed 74
ROM washed Ohio 71
Stoker-grade Ohio unwashed 68
Limestone/High Sulfur Coal Fuel Pellet. The results of the
Phase I study indicated that the limestone/high-sulfur coal fuel pellet
showed excellent potential for sulfur capture. To test the concept
further and to identify any operational problems, 90 Mg of pellets
were fired in the Battelle steamplant boiler in Runs SP-18 and SP-19.
In these runs, sulfur captures of better than 70 percent
were achieved, based on SO., emission levels in the stack. However, as
indicated in Table II-5, only 84 and 77 percent of the fuel sulfur was
accounted for in Runs SP-18 and 19, respectively, based upon a mass
balance.
11-43
-------
Several operational problems were encountered. These were
attributed to the low-heating value, high ash content, and excessive
amount of fines of the fuel product. The deep bed of compact fuel
and ash significantly increased the resistance through the bed. As
a result, there was noticeable air leakage around the grate. Conse-
quently, it would appear that the air/fuel ratio in or just above the
grate was substantially less than that indicated in the stack. Addi-
tionally, because of the excessive amount of fines and lack of a consis-
tent size distribution, the fuel could not be fed uniformly to the bed.
This may have been a factor in the relatively high CO emissions.
The limestone/coal pellet developed for this study offers a
potentially interesting and exciting new means of sulfur control for
industrial boilers. The use of limestone for capturing sulfur is by
no means new. Exactly how and why limestone should be effective in
capturing sulfur in 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. It
is conceivable that reduced conditions (by insufficient combustion air
diffusing into the pellet) exist in the fuel bed, ultimately leading
to sulfur capture via sulfide formation. This is also an explanation
offered by Maloney.
COMBUSTION SYSTEM DESIGN MODIFICATIONS
As part of Phase II, combustion modifications with the potential
to reduce emissions and improve overall boiler performance were to be
identified. Experiments with the Battelle boiler indicated that those
modifications that would improve air/fuel ratio control in the fuel
bed and the combustion zone above the bed would provide optimum stoker-
boiler performance from an emissions and performance viewpoint. This
air/fuel ratio is dependent on the design of a number of system design
components that include:
feed system
grate design
9 overfire
fly-ash reinjection.
11-44
-------
Feed System
Although not a primary combustion modification technique,
improvements in the coal feed system will increase the control of fuel,
distribution within the stoker boiler. This includes the rotor-feed
mechanism, as well as the transport system to the feeders. The feeder
must spread the coal uniformly on the grate and avoid regions of coal
buildup. Many of the feed problems encountered are caused by the
coal transport system segregating the coal size by distributing a dis-
proportionate amount of fines to certain feeders. For example, in the
Battelle stoker, one feeder received a disproportionate amount of fines
while the other received a disproportionate amount of top sizes; the
result was nonuniform combustion and excessive fly ash carryover.
Grate Design
The grate not only supports the fuel bed but also serves as a
distributor plate providing a uniform flow of air through the bed. The
pressure drop across the grate is only about 25 mm to 50 mm H.O at full
load. As a result, any maldistribution of the coal on the bed will
result in a nonuniform flow of air. The shallower portions of the bed
receive more air than the deeper portions. Increasing the pressure
drop across the grate would tend to compensate for maldistribution of
the coal. However, increased pressure drop would require improved
air seals around the stoker grate.
Air leakage around the grate, or infiltration of air into the
boiler, creates problems in controlling the air/fuel ratio. The excess
air level of a stoker-boiler system is generally measured in the stack.
With significant amounts of air infiltration, it is difficult, if not
impossible, to determine the air/fuel ratio in the combustion zone.
In the Battelle stoker, during cold flow tests on a bare grate, 14 percent
of the air bypassed the grate through leakage around the grate. This
percentage would increase during stoker operation. As a result, the
stoker boiler is often operated with significantly more excess air than
is required to ensure that there is sufficient undergrate air to burn the
coal and keep the bed from clinkering. This additional amount of air acts
only as a diluent and does directly enter into the combustion of the fuel.
It does, however, reduce boiler efficiency, increase fly ash carryover,
and perhaps increase NO levels.
11-45
-------
Overfire Air
The overfire air jet system is designed to provide effective
aerodynamic mixing above the bed. To accomplish this, the overfire air
jet system must effectively mix fuel fragments and air above the bed
without disrupting the fuel bed.
The overfire air systems is one of the more promising combus-
tion modification techniques in terms of reduced emissions and improved
performance. For this reason, and because modifying the overfire air
jet system on a large system is expensive and difficult, it was decided
to analytically investigate overfire air jet systems.
Analysis of the operation of such jet systems shows that the
individual jets, which can be round or rectangular with a low aspect
ratio, can be treated as simple round jets for the initial part of their
trajectory. However, consideration must be given to the difference in
density of the jet and its surrounding medium, which both affects the
aspiration rate of the surrounding gases and imposes a negative buoyancy
effect on the jet. Furthermore, the jet is in a cross-wind of the
rising gases from the fuel bed.
As the jet approaches the opposite wall (or, for directly
opposed jets, the mid-plane), two further effects occur. First, the
jet is affected by the aspiration of adjacent jets. Second, the pre-
sence of a barrier wall forces the jet to turn upward away from the bed.
These complications, plus the fact that the system is three-dimensional,
make physical studies, physical modeling studies, and mathematical
modeling studies quite difficult. This situation is discussed in more
detail in Appendix II-A. As a result of these difficulties, we fine
both a paucity of data in the literature on all but single jets in the
open, and an overabundance of empirical rules of design. In fact,
using all the empirical results that seem logical, one arrives at an
overspecified system with no indication of which rules to drop. On the
experimental side, even the most recent treatments can refer only
to very early work for experimental results on overfire air jets.
11-46
-------
The review in Appendix II-A shows that a. water-salt water
model can be used satisfactorily to develop either general rules for
the optimization of overfire air jets or to optimize conditions for
overfire air jets in any specific application. For most ranges of
operation, it is expected that buoyancy effects can be ignored. In
this case, air models can be used. Finally, techniques are now avail-
able that make three-dimensional mathematical modeling of the overfire
air system economically feasible.
Fly-Ash Reinjection
Fly ash reinjection systems are used to return the carryover
back into the high-temperature zone above the grate for burning. The
manner and location in which the fly ash is reinjected can have a signi-
ficant effect on particulate loadings. Ideally, the fly ash should
be reinjected into the furnace in a region in which the carbon is
burned out and the reinjected fly ash remains on the grate for disposal
rather than being recirculated. Gravity flow systems, in which the fly
ash is fed by gravity from the collecting hoppers to the rear end of
a forward moving grate, have been used. In systems in which the fly
ash is pneumatically conveyed back to furnace, the reinjection nozzle
should be located between the overfire air jets to insure complete burn-
out of the fly carbon. In the Battelle steamplant, the fly ash from
collecting hoppers was reinjected through nozzles along the boiler
sidewalls. Because of the high amount of air leakage around the grate,
much of this reinjected fly ash was carried to cooler regions of the
boiler before the fly carbon could be burned.
11-47
-------
SUMMARY
Of the criteria pollutants, the most troublesome emission to
control in an industrial stoker-boiler system is S02. The level of S02
emission is related to the quantity of sulfur in the coal and is essen-
tially independent of the combustion conditions. Thus, a modified or
treated fuel, such as the limestone/coal fuel pellet, offers an option
to flue-gas scrubbing. Treated or modified coals to control SO are
still in the development stage and must be shown to be economically
viable before commercialization. NO emissions from stokers are lower
X
than those from conventional pulverized-coal burners and until standards
become more stringent, this emission does not appear to be a major
deterrent to the use of stoker coal. Furthermore, NO emissions can be
x
effectively controlled by combustion design or operating modifications.
CO and unburned hydrocarbon emissions are related to the design and
operation of the stoker-boiler system and can be controlled by careful
regulation of the air and fuel distribution. Similarly, particulate
loading is dependent on the design and operation of the stoker boiler
system and can be controlled accordingly. In addition, particulate
control devices, such as mechanical collectors or baghouses, have been
found to be effective control devices.
The results of the Phase II program indicate that environ-
mentally the lack of a viable S02 control is a deterrent to increased
coal utilization. As a result, and because limestone/coal fuel pellet
data are encouraging, it is recommended that the fuel pellet be refined
further.
11-48
-------
REFERENCES
II-l. 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.
II-2. Gronhovd, G. H., P. H. Tufte, and S. J. Selle. Some Studies on
Stack Emissions from Lignite-Fired Power Plants. Presented at
the 1973 Lignite Symposium, Grand Forks, ND, May 9-10, 1973.
II-3. Maloney, F. L., P. K. Engel, and S. S. Cherry. Sulfur Retention
in Coal Ash. KVB 8810-482-b, EPA Contract No. 68-02-1863,
Industrial Environmental Research Laboratory, EPA, Research
Triangle Park, NC, Nov, 1978.
II-4. Compilation of Air Pollution Emission Factors. AP-42. U.S.
Environmental Protection Agency, 1973.
II-5. Giammar, R. D., et al., Emissions from Residential and Small
Commercial Stoker-Coal-Fired Boilers under Smokeless Operation.
EPA-600/7-76-029, U.S. Environmental Protection Agency,
IERL, Office of Energy, Minerals, and Industry, Research
Triangle Park, NC 27711, Oct, 1976.
II-6. Giammer, R. D., et al., Experimental Evaluation of Fuel Oil
Additives for Reducing Emissions and Increasing Efficiency
of Boilers. EPA-600/2/77-008b. US Environmental Protection
Agency, Research Triangle Park, NC. Jan, 1977.
II-7. Jones, P. W., et al., "Efficient Collection of Polycyclic Organic
Compounds from Combustion Effluents", presented at the 68th Annual
Meeting of the Air Pollution Control Association, Paper No.
75-33.3, Boston, June 15-20, 1975.
II-8. Gabrielson, J. E. and P. L. Langsjoen. Field Tests of Industrial
Stoker Fired Boilers for Emission Control. EPA-600/7-79-050a,
Proceedings of the Third Stationary Source Combustion Symposium;
Vol 1. Utility, Industrial, Commercial, and Residential Systems, Feb, 197!
II-9. Heap, M. P. and R. Gershman. Pollutant Formation During Coal
Combustion. EPA-600/7-78, Proceedings of the Engineering
Foundation Conference on Clean Combustion of Coal. April, 1978.
11-49
-------
REFERENCES
(continued)
11-10. Pereira, F. J., J. M. Beer, B. Gibbs, and A. B. Hedley,
NOX Emissions from Fluidized-Bed Coal Combustors.
Fifteenth Symposium (Int'l) on Combustion, Tokyo, Aug, 1974.
11-11. Niessen, W. R. Combustion and Incineration Processes.
Marcel Dekker, Inc., NYC, p 129-170, 1977.
11-12. Davis, R. F., The Mechanics of Flame and Air Jets. Proc. Inst.
Mech. Eng., 137 (11-72) 1937.
11-50
-------
APPENDIX II-A
OVERFIRE AIR JETS
-------
APPENDIX II-A
OVERFIRE AIR JETS
INTRODUCTION
Overfire air jets are used in a stoker-fired furnace to produce
a uniform composition of gases a short distance above the burning fuel
surface. This not only prevents cold streams of air from passing com-
pletely through the combustion region above the fuel surface, but produces
a uniform flux of heat back to the surface. This heat flux tends to level
out variations in the surface combustion rate.
In this discussion, following certain general comments, the
performance of the simple turbulent jet and then the turning effects on
such jets are discussed. This is followed by comments on physical and
mathematical modeling problems, and suggestions for their use in specific
research programs.
GENERAL COMMENTS
Simple turbulent jets are characterized by a jet core of about
6 diameters in length, with a uniform velocity on the axis. This is
followed by a decay of the velocity on the axis. Axially directed
velocities off the axis are related to the maximum velocity on the axis
by a distribution curve, that is approximately Gaussion in nature. For
round jets, the decay rate of velocity (and the jet component of the
mixture) is inversely proportional to the distance, starting at the end of
the core region. The aspiration of the surrounding gas is directly pro-
portional to the distance. Slot jets follow a similar, but more complex
relationship.
Recent emphasis in jet turbulence and mixing studies has been
related to the large vortices that are formed and move along the inter-
action surface between the jet and the surrounding environment. This
results in an unmixedness of a considerably different character than
is usually ascribed to phenomena associated with Gaussion-type curves.
-------
Jets used in overfire air systems are subjected to transverse
buoyancy effects and to cross winds. In a cross wind or with transverse
buoyancy effects, the jet tends to maintain its original character but is
bent. For cross winds, the path shape in terms of jet characteristics
becomes a function of the momentum ratio of the cross wind fluid to the
jet fluid. The higher the ratio, the more the jet is bent. With severe
bending, the jet cross section distorts from a circle to a bean shape.
The buoyancy effects enter in the form of a Froude number,
2
V p/gD Ap. The larger the inverse value of this dimensionless group, the
more the jet is deflected.
Since an overfire jet, even using recirculated products rather
than air, is more dense than the surrounding medium composed of the hot
gases coming off the fuel bed, the upward flow of gases from the burning
fuel surface and the buoyancy act in opposite directions on the jet.
The multiple jets of an overfire system will interact, but for
many situations, the effects can be predicted by a simple addition of the
effects of the individual jets.
SIMPLE JET PERFORMANCE
Jet Size
The parameters involved in jet phenomena are taken from
Schlichting . Other sources are equally acceptable, with little
difference in the characteristic values. The equation assumes a point
source, ignoring the potential core. The radius of the jet at the
distance where the velocity is half the velocity on the axis, b, as a
function of distance from the source x, is
b = .0848 x.
The total volume flow rate, Q, is given by
^"^i
Q = 0.404 Y K x,
where
f 2
K = 2 IT / u rdr.
0
References are listed on p II-A-13.
II-A-2
-------
The value of the axial velocity, u, on the axis, U, is given by
1.59 Ub =
If the jet gas is assumed to be at the furnace gas temperature, so the
performance can be compared with that of a jet in an environment of the
same properties, the value of K based on the true jet entrance conditions
must be multiplied by (p /Prr) to preserve the momentum. Thus,
K = (po/pf)(M/dpQ)2 (4/31) .
where M is the mass flow rate, and subscripts o and f refer to jet and
furnace gas conditions.
Aspiration of Furnace Gas by Jet
If we assume an active length of jet, L, equal to the distance
across the furnace, it would seem reasonable that one would want to
aspirate into the jet a volume produced by that portion of the burning bed
assigned to that particular jet. For more intensive mixing, one might want
to aspirate in n times the volume of gas produced. If the mass supply rate
of gas per unit of the bed area is m, then
/'
nWLm/pf = .404 y K L,
f
where W is the width of the area served by one jet. Since m and p are known,
this specifies the jet momentum as a function of the width, W, of the area
served by the jet. For a given air temperature relative to the furnace gas
temperature, the product of jet velocity and diameter (or ratio of jet mass
flow rate to diameter) is thus proportional to nWm. For a constant ratio
of the jet to the fuel bed mass velocity, n is therefore proportional to
d/W.
Ratio of Overfire Air to Total Air
If the ratio of the overfire air to total air, A, is specified,
another design relation can be established. Similarly, if recirculated
products are used for overfiring, a design relationship can be established.
II-A-3
-------
If M is the mass flow rate of air per jet serving an area of WL,
then
M/(M + WLm ) = A
a
where m is the gas flow rate per unit area. If m is the fuel consumption
a c
rate per unit area,
m = m + m .
c a
If recirculated products are used, M in the denomination is set equal to zero.
O *
For design purposes, m is 20 to 40 Ib/hr ft , m /m is 14 to 20,
C 3 C
and A is 0.1 to 0.2.
Rewriting the equation for A,
M = A m WL/(1 - A).
a.
» »
Since m and A are specified, M/WL is a constant. Furthermore, L/d becomes
3
a simple function of n. Specifically,
1/7
L/d = (2.194 n)(m/ma)(po/pf)±// [(1 - A)/A)]
For A = .15, m /m =17, T = 600 R and T = 2860 R, L/D = 28.8 n.
3. C O L
If L = 10 ft and n = 1, d = 4.18 in. For n = 2, d = 2.09 in.
JET SPACING
The simple jet was shown to expand with a radius to the half
axial velocity of
b = .0848 x.
One might assume that
W ' 2mb(L)
where b, , is evaluated at the distance, L.
Thus,
W = .1696 mL.
The sample computation then gives
W = 5.73 mnd;
for m = n = 1, this would result in 24-in. spacing of jets
II-A-4
-------
TURNING EFFECTS ON JETS
Cross Flow and Buoyancy
Abramovich suggests, in terms of the parameters defined
above, the relationship
/P \ 2 2'55
y [ f\ / m/pf \ /_x
d = I o~ / I i 9 \ d
vox \4 M/p irdz / \
for the deflection of a jet by cross flow. This simplifies to
y/d = (P0/Pf) Ud2 m/4M)2 (x/d)2'55
Gray and Robertson suggest for the effect of buoyancy,
/T - T \
y/d = .0460 ° _
2 22
V(4M/ud p ) d
(Note that the coefficient in the reference is in error.) Thring
suggests a more complex relation, but it is not worth the trouble to use
at this point. As a result the total deflection is
,, / / \
y/d - (po/p£)
Pf
- .046 (1 --^-) r-^ g<
Mo \4 M /
Now considering the condition where y/d comes back to 0 after the original
deflection,
(xo/d)'45 = 21.7 [(Pt/(PQ- Pf)I
P z gd,
f
II-A-5
-------
As a sample computation, using the same values as before, and assuming
that m = 30 Ib/ft hr, and d = 4.18 inches,
(K /d)*45 = 66.17
o
Thus, x = 3846 ft. This seems ridiculous. Calculating the cross wind
effect separately, leads to a lift of 8.65 ft in 10 ft. However, the
gravity fall will only be 7.1 in. at 10 ft for these particular conditions.
If for some condition, the jet did fall to the surface rapidly or impact
the surface, the aspiration effect would change as the jet spread out
over the hot surface.
Wall Effect
We note that in the above computation, the impinging wall was
neglected. If jets are directed out from one wall, or are directly opposite
each other so there is a virtual wall in the center, then the aspirated
gas plus the jet gas must turn upward as they approach the impinging wall.
This phenomenon has nothing to do with either cross-wind effects or buoyancy
effects. Some estimates can be made from impingement studies. One might
assume that the j ets are tipped down slightly and impinge on the floor and
are evenly distributed. Then treating the problem as a two-dimensional
potential flow problem, or using available experimental data, one can
produce the velocity profiles. The upward velocity of the hot gases would
then be superimposed.
When jets are staggered on opposite walls, they interlace. It
would appear that the gross recirculation pattern described above for jets
on only one wall would disappear.
Length of Throw
One may assume that the jet is ineffective when the velocity on
the jet axis is equal to some multiple, p, of the upward velocity at the
fuel surface. Thus, combining previous equations,
1/2
pucLp = 6.57 (po/pf) u.
II-A-6
-------
where L is the penetration distance. Stephens and Mohr assume
a value of 4 for p. Erigdahl assumed a value of -the order of 2.
An alternate form is
i ij
L =6.57 (p_/p ) / (4 M/TTdmp).
p to
For the assumed values of the various terms
L /L = 0.2925 (W/pd) .
P
Removing W/d by use of previously derived relations,
L /L = 1.676 mn/p.
P
Ifm=n=l,L =Latp= 1.676, close to the value used by Engdahl.
P
COMMENTS ON MODEL STUDIES
Mathematical Modeling
The above discussion indicates that the jets should probably be
fairly close together. They might be from one side, opposed from two
sides, or staggered. The plan view of possibilities resolves into two
cases, as follows:
Jets from one
side, or
opposed
-------
In Case a, there is a surface of symmetry through the jet axis and between
each pair of jets. In Case b, the same situation occurs, but the conditions
on the surface between jets have to have a reflective character.
The sketch indicates that the region for which a solution is
required is long and thin; in the vertical direction, the plane extends
from somewhat below the jet to far above the jet. The jet may be directed
in other than a horizontal manner. In any case, boundary layer type
assumptions (uniform pressures through the narrow dimension) can be used.
Since the jets are controlling the mixing, another simplification
may be made. Round jet flow equations can be based on a constant turbulent
kinematic viscosity assumption. In fact, the character of turbulence
can probably be considered uniform throughout the system.
It therefore appears that primitive 3-dimensional equations can
be used in a simplified form to solve the flow problem. Small cold regions
in the reaction surface can be considered, comparing the uniformity of
temperature at higher levels as a function of the different variables.
In the long run, more sophisticated mixing equations and reactions
above the bed can be included, but the simple approach outlined above should
answer major questions as to optimum design, and the variations obtained by
using air, steam, steam injected air, recirculated products, and steam
injected recirculation products, with various jet sizes, velocities, and
angles, for different fuel consumption rates, air supply rates, and overfire
air rates.
Dimensional Analyses for Physical Modeling
If one were to investigate only the flow pattern with no temperature
variation and no cross flow, and in a highly turbulent situation, the vector
velocity at any point would be given by
u = f(x, y, z, W, L, H, a, U , d )
where H is the heigth of the jet above the grate, and a is the direction
of the jet, zero when horizontal and positive upward. All measurements
x, y, z are from the orifice center. This obviously can be simplified to
II-A-8
-------
, ,. ,x y z W L H
u/u = f (p £, ₯, z , d ' d.' a
3 J 3-
If C is the concentration of the jet material,
r/r /£ Z .5. W L H
C/Cj " (L' H' N' d.' d.' d.' a)
where
0 < f < 1, -1 < § < 1, -1 <
Note that f indicates function, but not necessarily the same function,
For
i- f t - 1/2-
the data should correlate in the form
Wu., C/Cj-f (i.i.i.
J J J
This can be modified to
/ n/^ f I /x cos a + y sin d >
u/u., C/C. = f I ( - - * - ),
(y cos a - x sin a) 2]
If the flow velocity from the grate, U , is now added as a variable,
&
the additional parameter, U/U., must be added to the independent variables,
C? -J
and a must be reinserted as a variable.
We next consider the case where there is a difference in temperature
between the jet and the upward flow. Not considering gravity effects, the
relation p./p must be added, or T /T.. Since momentum effects are the key
Jo & J
II-A-9
-------
to the flow pattern, one would expect to change the ratio U /U to
1/2 s J
(U /U.)(T./T ) . For the regions some distance from the jet exit, one
O J Jo
can follow through with the Thring criteria based on constancy of momentum
22 ~ 2- 2
in modeling, p.d. U. = p d u. and constancy of mass flow
J J J o J
2 - 2- _ o-2 1/9
P.d U - P d U.. Thus, U. = U., and p.d. = p d , or d = d.(p./p ) '
J J J o J J J J j g J 3 §
1/2
= d.(T /T.) , This indicates the substitution of d for d. in the
JJ § J J
correlation relations.
The last factor to be considered is the effect of gravity. It is
expected that the Froude number will enter at this point, in the form
U 2 T /(gd )(T - T ).
J J Jo J
The final correlation relation, therefore, is
u/u., C/C., and (T -T)/(T -T.) =
J J 8 S J
U \ /U \ / T- X1/2
, H \ / 2 1 \
A refinement which can be made later is to alter the terms on the LHS so
they all go from 1 at the jet to zero at £ * °°.
H
II-A-10
-------
ADDITIONAL RESEARCH CONSIDERATIONS
Mixing data are required for the two basic overfire jet configura-
tions as a function of jet size, jet angle, spacing, and heigth, relative
to the furnace width. Three possibilities suggest themselves, namely
water modeling, air modeling, and mathematical modeling. Each is
discussed in turn.
Water Model
A water model can be used to simulate the entire system, including
the buoyance effect. Salt water is used to simulate the overfire jets and
pure water the gas flowing from the bed. A movable salt water source in
the bed can be used to simulate a cold spot and determine the mixing. Con-
centration is determined by a measure of the local conductivity. Coloring
matter can be used to trace flow paths and determine dispersion effects.
While the density difference is not as great as that for the combustion
process, 1.2:1 being about the most that can be obtained, there is a large
gain in Reynolds number, of the order of 20. A size reduction of about 4
is possible with no change in Reynolds number of the jet,-while simulating
buoyance effects.
One visualized model would consist of a square stack, open at
the top and with a perforated plate that could be moved up from the bottom.
Part way up on two opposite sides would be slots where nozzle bars could
be sealed in. These would contain the desired number and size of nozzles,
at the desired angle, and with a feed supply. Salt water flow would be
balanced to each nozzle bar and measured. Flow of fresh water through the
perforated plate would also be measured.
Air Model
If the effects of gravity can be ignored, air modeling can be used.
The nozzle size is increased to compensate for the higher density of the
jet fluid relative to the surroundings and the modeling proceeds in the
manner used in many investigations.
II-A-11
-------
The effect of gravity, which involves the Froude number
2
pU /dgAp, may pose a problem. Even using a small model furnace, so that
2
Ap/p is the same as in the prototype, U /d must be held constant. Now
considering the Reynolds number of the model jet compared to the full-scale
jet, Re /Re - (d /d )3/2. If d /d = 1/5, then Re /Re = 1/11.2. For
mpmp- mp mp
large diameter jets, at low overfire air ratios and low firing rates, this
puts the model Re in a questionably low range. However, it should be
possible to ignore this effect when the full-scale Re is sufficiently large.
The next item to consider, though, is that the jet velocity in the model
would have to be reduced by 2.25. Since there is a definite relationship
between jet velocity and bed velocity, the bed velocity would have to be
reduced accordingly. The velocity would no longer correspond to that
from a hot bed, and thus the hot gas would have to be generated separately
and supplied through a simulated bed.
In an air model, a cold jet or tracer gas up through various spots
in the simulated bed can be used to determine the degree of mixing. Both
colored traces and temperature measurements could be used.
The model visualized would be similar to that described above
for water.
Mathematical Model
The third possibility is to set up a mathematical model of the flow
system. This would have to be three-dimensional, which complicates the
problem considerably. However, programs are available in primitive variables
(pressure and velocity) that can be used for this purpose. Since no chemical
reactions are considered, the kinetic aspects are not involved; this simplifies
the problem immensely. Depending on the size of the modules, the round jets
can be simulated by square jets, or multiple square jets. At a reasonable
distance from orifice, this would have no effect on the results.
From such a program, once could easily check the importance of
certain factors in the analysis, such as the inclusion or exclusion of
buoyance effects, the substitution of a larger diameter jet with a uniform
gas temperature (momentum and volume flow rate matched) and the exclusion
of flow from the bed in the computation. It is important to note, however,
that certain parameters entering into the mathematical analyses would require
experimental determination.
II-A-12
-------
REFERENCES
II-A-1. Schlichting, H.. Boundary Layer Theory. McGraw Hill,
607-609, 1960.
II-A-2. Afaramovich, G. N. The Theory of Turbulent Jets.
M.I.T. Press, 1963. Equation 12.127.
II-A-3. Gray, F. A., and Robertson. The Deflection of Hot Jets Due
to Buoyancy. J. Inst. Fuel, 29, 424-427, 1956.
II-A-4. Thring, M. W. The Science of Flames and Furnaces,
Win Clowes & Sons, 1962.
II-A-5. Stephens, R. H. and C. M. Mohr. Reduction of Combustible
Emissions in Municipal Incinerators Using Overfire Jets.
AIChE Symposium Series No. 147, 71, Air I. Pollution
Control and Clean Energy, 134-140, 1975.
II-A-6. Engdahl, R. B. Design Data for Overfire Jets. Combustion,
March, 1944.
II-A-13
-------
PHASE III. LIMESTONE/COAL PELLET DEVELOPMENT
-------
EXECUTIVE SUMMARY
The Phase III program focused on refinement of the limestone/
coal fuel pellet and evaluation of its suitability as an industrial
stoker-boiler fuel. This program consisted of four major tasks.
1. Pellet Development aimed at developing a fuel pellet
with mechanical strength characteristics that can
withstand weathering and the severe stresses of an
industrial stoker coal-handling and feeding system,
burns at reasonable rates, and captures sufficient
sulfur to be competitive with other control
strategies. Mechanical strength characteristics
were evaluated with standard laboratory tests.
Burning characteristics and sulfur capture were
determined in a fixed-bed reactor simulating the
fuel bed of a spreader stoker.
2. Process Variables Selection combining a mathematical
model analysis with a series of experimental studies
to develop a more comprehensive understanding of
the processes that influence the combustion of the
fuel pellet and control the capture of sulfur.
3. Laboratory Evaluations conducted in both the 200 kWth
model-spreader stoker and the.8 MWt^ Battelle steam
plant boiler to evaluate the most promising candidate
pellets.
4. Economic Analysis aimed at developing pellet process
costs.
The major results and conclusions of the four tasks are:
Pellet Development
A fuel pellet was produced that, according to
laboratory tests, has mechanical strength and
durability characteristics similar to those of
conventional coals.
Pellets produced by auger extrusion or pellet mill
processes had better mechanical strength than those
produced by disc pelleting or briquetting.
VI
-------
Binders that provide some resistance to the weather
were identified. However, no binder was identified
that provided complete weather proofing.
9 The fixed-bed reactor experiments indicated a weak
dependency between Ca/S ratio and sulfur capture
for Ca/S ratios above 2.
e Calcium oxide is a superior absorbent to limestone,
but is not economically competitive with limestone.
Additives do not appear to enhance sulfur capture.
Process Variables Selection
The mathematical model predicts an optimum coal size
(35-40 mm diameter) for maximum sulfur retention.
The model indicates a weak dependency on the calcium/
sulfur ratio.
Scanning electron microscopy and x-ray diffusion are
powerful tools for the study of solid-state reactions
in the pellets. Results indicate that sulfur is
retained predominantly as CaS04.
Sulfur may react directly with limestone by solid-
state processes without involving the formation, of
so2.
Laboratory Evaluations
« Auger-extruded and milled pellets burned better than
briquets and disc-agglomerated pellets.
Sulfur capture of about 65 percent was achieved at
Ca/S molar ratios of 3.5.
Sulfur capture of about 50 percent was achieved in
the steam-plant stoker. In comparison to the model
spreader, this lower S02 capture was attributed to
higher temperatures (in excess of 1300 C).
Sulfur capture appeared to be weakly dependent on
fuel-bed temperature.
In the Battelle steam power plant, the fuel pellets
burned as well as low-sulfur coal.
Economic Analysis
It is estimated that limestone/coal fuel pellets
can be produced for about $15.40/Mg ($14/ton) of
pellets above the costs of the high-sulfur coal.
vii
-------
PHASE III
CONTENTS
List of Figures i
List of Tables iii
Executive Summary vi
I. BACKGROUND III-l
II. OBJECTIVE AND SCOPE III-2
III. PELLET DEVELOPMENT III-3
Binder Selection III-4
Mechanical Properties III-7
Pellet Durability Index (PDI) III-7
Fuel Pellet Strength Test III-7
Stoker Fuel Pellet Weatherability Index (WI). III-9
Postweathering Test III-9
Laboratory Test Results III-9
Baseline Data III-ll
Effect of Production Technique III-ll
Effect of Binder 111-12
Discussion of Experimental Results 111-20
Binder 111-23
Conclusions from Fixed Bed Reactor Study. . . 111-27
IV. PROCESS VARIABLES SELECTION Ill-28
Pellet Details 111-29
Experimental Results 111-33
Pellet Characterization Studies 111-42
Scanning Electron Microscopy 111-49
X-Ray Diffraction 111-54
Description of Mathematical Model 111-54
Modeling Analysis of Pellet Combustion Data. . . . 111-59
Determination of Rate Constants 111-64
Transient Heat-Transfer Analysis 111-74
-------
CONTENTS (Continued)
Page
Detailed Modeling Calculations 111-77
Modeling and Characterization Summary and Conclusions 111-84
V. LABORATORY EVALUATION 111-85
Model Spreader Experiment 111-85
Sampling System 111-87
Ca/S Ratio 111-87
Production Technique 111-87
Binder Type 111-89
Pellet Preparation and Properties 111-90
Experimental Procedures 111-94
Checkout Runs 111-96
Demonstration Test 111-99
Summary III-103
VI. LIMESTONE/COAL FUEL PELLET PROCESS COST SUMMARY. . . . III-104
Basic Assumptions Ill-104
Process Flowsheet III-106
Sources of Information III-106
Capital Cost Estimates IH-110
References Ill-114
-------
PHASE III
List of Figures
Page
Figure III-l. Apparatus for Testing Fuel Pellet Strength III-8
Figure III-2. Fixed-Bed Reactor Ill-15
Figure III-3. CC>2 Concentration Measured in the Gas Stream from
the Fixed-Bed Reactor for the Limestone Pellets
in Run 110 111-37
Figure III-4. Concentrations of CC>2, SC>2, 0£, and CO Measured
in Gas Stream from the Fixed Bed Reactor for the
50 Weight Percent Coal/50 Weight Percent Limestone
Pellets in Run 104 111-38
Figure III-5. Concentrations of C02, S02, 02, and CO Measured
in Gas Stream from the Fixed Bed Reactor for the
70 Weight Percent Coal/30 Weight Percent Limestone
Pellets in Run 101 111-39
Figure III-6. Release Rates for C02 and S02 from 50 Weight
Percent Coal/50 Weight Percent Limestone Pellets
During Run 104 111-43
Figure III-7. Release Rates for C02 and S02 from 70 Weight Percent
Coal/30 Weight Percent Limestone Pellets During
Run 101 111-44
Figure III-8. Thermocouple Measurements for Fixed Bed Reactor
Run 101 with 70 Weight Percent Coal/30 Weight Percent
Limestone Pellet 111-45
Figure III-9. Metallographic Cross Sections of 50 Weight Percent
Coal/50 Weight Percent Limestone Pellets as a
Function of Burning Time III-46A
Figure 111-10. Scanning Electron Micrographs and Elemental X-Ray
Maps for Unburned 50 Weight Percent Coal/50 Weight
Percent Limestone Pellet (Pellet 72) III-50a
Figure III-11. Scanning Electron Micrographs and Elemental X-Ray
Maps for Outer Ash Zone in 50 Weight Percent Coal/
50 Weight Percent Limestone Pellet (Pellet 65)
Burned for 12 Minutes in the Fixed Bed Reactor. . III-51a
Figure 111-12. Scanning Electron Micrographs and Elemental X-Ray
Maps for White Reaction Zone in 50 Weight Percent
Coal/50 Weight Percent Limestone Pellet (Pellet 65)
Burned for 12 Minutes in the Fixed Bed Reactor. . . III-52a
-------
List of Figures (continued)
Page
Figure 111-13. Scanning Electron Micrographs and Elemental X-Ray
Maps for Inner Core in 50 Weight Percent Coal/50
Weight Percent Limestone Pellet (Pellet 65) Burned
for 12 Minutes in the Fixed-Bed Reactor III-53a
Figure 111-14. CuKa Radiation Scatter of Unburned Coal 111-55
Figure 111-15. CuKa Radiation Scatter of Unheated Limestone 111-55
Figure III-16. CuKa Radiation Scatter of Unburned 70/30 Mixture. . . 111-55
Figure 111-17. CuKa Radiation Scatter of 70/30 Limestone/Coal
Burned Pellet. Major Identified Constituents
CaO (Lime), CaS04 (Anhydrite) 111-55
Figure 111-18. Summary of Results of the Chemical Analyses on
50 Weight Percent Coal/60 Weight Percent Limestone
Pellets from Fixed Bed Reactor Experiments (Data
from Table 8) 111-62
Figure 111-19 Summary of Results of the Chemical Analyses on 70
Weight Percent Coal/30 Weight Percent Limestone
Pellets from Fixed Bed Reactor Experiments (Data
from Table 9) 111-63
Figure III-20. Unburned Core Radius as a Function of Burning Time
Calculated Using Simplified Model (Data Points
Represent Measurements on 50/50 Pellets) 111-65
Figure 111-21. S02 Release Rate and Calculated ECO Analyzers
Based on the Simplified Model (Equation 7) for
50 Weight Percent Coal/50 Weight Percent
Limestone Pellet 111-70
Figure 111-22. Oxygen Consumption Rates Calculated Using
Simplified Model 111-71
Figure 111-23 Sulfur Retention and Burning Time for 50 Weight
Percent Coal/50 Weight Percent Limestone Pellet
Calculated as a Function of Pellet Diameter Using
the Simplified Model 111-72
Figure 111-24. Burning Time for 50 Weight Percent Coal/50 Weight
Percent Limestone Pellet Calculated as a Function
of Temperature Using the Simplified Model 111-73
Figure 111-25. Transient Heat Transfer Calculations for a 70
Weight Percent Coal/30 Weight Percent Limestone
13.7 mm Diameter Spherical Pellet Ill-75
ii
-------
List of Figures (continued)
Figure 111-26. Transient Heat Transfer Calculations for a 50
Weight Percent Coal/50 Weight Percent 13.7 mm
Diameter Limestone Spherical Pellet 111-76
Figure 111-27. Calculated Combustion Rate for 70 Weight Percent
Coal/30 Weight Percent Limestone Pellet 111-78
Figure 111-28. Calculated S0£ Release Rate for 70 Weight Percent
Coal/30 Weight Percent Limestone Pellet 111-79
Figure 111-29. Calculated 0£ Consumption Rate for 70 Weight
Percent Coal/30 Weight Percent Limestone Pellet. . 111-80
Figure 111-30. Calculated CaO Distribution at 35.10 sec in
70 Weight Percent Coal/30 Weight Percent
Limestone Pellet 111-81
Figure III-31. Calculated S02 Distribution at 35.10 sec in 70
Weight Percent Coal/30 Weight Percent Limestone
Pellet 111-82
Figure 111-32. Calculated 02 Distribution at 35.10 sec for 70
Weight Percent Coal/30 Weight Percent Limestone
Pellet 111-83
Figure 111-33. Model Spreader Sampling System 111-88
Figure III-34. Location and Orientation of Gas Sample Probe . . . 111-95
Figure III-35. Coal/Limestone/Cement Pelletizing Process
Flowsheet III-107
iii
-------
PHASE III
List Of Tables
Page
Table III-l. Agglomeration Methods III-5
Table III-2. Binders Evaluated During the Phase III Program . III-6
Table III-3. Pellet Optimization Data Ill-lOa
Table III-4. Comparison of Test Results By Production Technique 111-12
Table III-5. Reproducibility Series with 50/50 Pellets, Fur-
nace Temperature 1040 C, with 33% Overfire Air . 111-18
Table III-6. Fixed-Bed Experiment Runs 111-19
Table III-7. The Effect of Ca/S Ratio on Sulfur Capture . . . 111-21
Table III-8. The Effect of Sorbent on Sulfur Capture 111-22
Table,-III-9. Comparison of the Effect of Limestone and Lime
on Sulfur Capture 111-23
Table 111-10. The Effect of Binder Type on Sulfur Capture. . . Ill-24
Table III-11. The Effect of Production Method on Sulfur Capture 111-25
Table 111-12. The Effect of Production Technique on Reactivity 111-26
Table III-13. The Effect of Additives on Sulfur Capture. . . . 111-26
Table 111-14. Typical Pellet Densities and Porosities 111-31
Table 111-15. Summary of Chemical Analyses on Raw Coal and
Pellets 111-32
Table 111-16. Summary of Fixed-Bed Reactor Experiments Performed
in Support of the Modeling Studies 111-34
Table 111-17. Results from Chemical Analyses of Pellets From
Fixed Bed Reactor Experiments 111-36
Table 111-18. Response Parameters for Gas Analyzers 111-41
Table 111-19. Results from Gas Analyzer Measurements 111-47
iv
-------
List of Tables (Continued)
Page
Table 111-20. Comparison of Gas Analyzer Measurements for
Coal/Limestone Pellets 111-48
Table 111-21. Summary of Results of the Chemical Analyses
on 50 Weight Percent Coal/50 Weight Percent
Limestone Pellets from Fixed Bed Reactor
Experiments 111-60
Table 111-22. Summary of Results of the Chemical Analyses
on 70 Weight Percent Coal/30 Weight Percent
Limestone Pellets from Fixed Bed Reactor
Experiments 111-61
Table 111-23. Summary of Chemical Analyses and Physical
Measurements on the Limestone Pellets from the
Fixed Bed Reactor Experiments 111-66
Table 111-24. Summary of Burning Rate Measurements for 50
Weight Percent Coal/50 Weight Percent
Limestone Pellets Based on Metallographic Studies. 111-67
Table 111-25. Summary of Calculations Using Simplified Model 111-68
Table 111-26. Model-Spreader Studies 111-86
Table 111-27. Potential Tolling Companies for Large Quantities
of Pellets 111-91
Table 111-28. Ultimate, Proximate, and Ash-Fusion Temperature
Analyses for Limestone/Coal Fuel Pellet Ca/S=3.5 111-92
Table 111-29. Mineral Analysis of Ash 111-93
Table 111-30. Comparison of Emissions From Combustion of a Low
Sulfur Coal and Limestone/Coal Pellet 111-98
Table 111-31. Emission Data Summary for Fuel Pellet Demonstra-
tion III-100
Table 111-32. Analysis of Method 5 Filter Catch (Weight Percent) III-101
Table 111-33. Analysis of Grate Discharge (Weight Percent) III-101
Table 111-34. Sulfur Balance III-101
Table 111-35. Summary of Limestone/coal Pelletizing Process
Costs III-105
Table 111-36. Coal/Limestone/Cement Pelletizing Process. . . . Ill-
Table 111-37. Capital Cost Estimates III-
v
-------
PHASE III. LIMESTONE/COAL PELLET DEVELOPMENT
SECTION III-l
BACKGROUND
During Phase II of this program, Control Technology Evaluation,
S02 was considered as the most troublesome emission to control in an
industrial stoker-boiler system. The other criteria emissions could be
controlled with existing technology or were within the current emissions
requirements. As a result, and because the limestone/coal fuel pellet
offers a means for environmentally acceptable burning of high-sulfur coal
in existing boilers, EPA continued the development of the fuel pellet as
part of this program.
From an industrial point of view, the possibilities of using
limes tone/coal pellets for removing SO- in situ via a dry process is
more acceptable than the use of scrubbers. The additional costs for
pelletizing the coal/limestone mixture and for the removal of 3 to 4 times
as much ash is far more attractive than the high cost of operating and
maintaining wet scrubbers.
III-l
-------
"SECTION III-2
OBJECTIVE AND SCOPE
The overall objective of the Phase III program was to refine
the limestone/coal fuel pellet and evaluate its suitability as an
industrial stoker boiler fuel. Phase III focused on the following
major tasks:
o Pellet Development aimed at producing a limestone/
coal fuel pellet that is economically competitive
with other control technologies, has sufficient
mechanical strength, durability, and weatherability
characteristics, and burns like ordinary coal.
Process Variables Selection aimed at developing a
coal pellet combustion model to be used as: (1) a
predictive tool to support pellet development and
(2) an interpretive tool to provide information on
the burning time and sulfur capture behavior of coal/
limestone pellets.
Laboratory Evaluations focused on evaluating the most
promising candidate pellets in the model-spreader
stoker boiler. In addition, a limited number of
pellet tests were conducted in the steam plant
stoker.
Economic Analysis aimed at estimating costs for
the manufacture of limes tone/coal fuel pellets.
III-2
-------
SECTION III-3
PELLET DEVELOPMENT
The results of the Phase II short-term steam plant demonstration
indicated that the physical properties of pellets had to be improved if
they were to be a viable industrial-stoker boiler fuel. Cement-bound
pellets used in Phase II did not have adequate strength or durability.
They broke on handling so that up to 50 percent fines was introduced
into the boiler. Furthermore, both from an economical and operational
standpoint, the amount of limestone required to capture the target goal
of 70 percent sulfur had to be limited to stay within a Ca/S molal ratio
of 3 to 4. Accordingly, an experimental research and development effort
was undertaken to provide an improved pellet over that evaluated in
Phase II.
PELLET PRODUCTION TECHNIQUES
The concept of agglomerating coal particles is not new. The
Institute for Briquetting and Agglomeration (IBA) was formed originally as
the International Briquetting Association in 1949 and has held biannual
conferences since. Initially, these conferences focused on coal briquetting
but have more recently expanded to include all types of agglomeration.
Agglomeration, in the general sense, is any process where smaller particles
are brought together and formed into larger masses of compacted material.
Agglomeration processes include pelleting, briquetting, and pelletizing.
Pelleting is described as an extrusion of small, particle-sized
material through a die, the material being forced by an auger or a ring mill.
Briquetting involves high-pressure compaction of material, often between rolls
that contain pockets to form the desired size and shape. Pelletizing is a
particle agitation process that involves rolling of fine, moist powders
in a disc or drum causing the particles to coalesce and form spheres
III-3
-------
All of these methods have advantages and disadvantages that affect their
suitability for coal agglomeration. They have all been used on at
least a laboratory scale for coal agglomeration, but no clearcut advan-
tage has been cited in the literature for any of them. Relative processing
characteristics for different agglomeration methods are summarized in
Table III-l, based on a review of IBA proceedings.
BINDER SELECTION
The binder is an essential element in fuel-pellet formulations.
With the right binder, a pellet will have the mechanical strength, durability,
and weatherability needed for a stoker-boiler fuel. Accordingly, an exten-
sive search was conducted to identify candidate binders. Table III-2 lists
the more promising binders evaluated during this phase of the work.
A review of the literature indicated that while extensive work
has been done on coal briquets, there has been little work on disc pelletizing,
mill pelleting, or auger extrusion of coal fines. Further, most of the coal
briquets were produced using a hard pitch binder, which requires an extensive
tempering process. This prior research, however, was useful in producing
a list of candidate binders to be investigated further.
(III-D*
Komarek provides a very useful discussion on binders used
during pressure compaction of various materials. From this source, as well
as data from Pennwalt Chemicals Corporation reproduced in the Chemical
Engineer's Handbook , candidate binders were identified. This
list was expanded during discussions with members of the ceramic and polymer
chemistry groups at Battelle, as well as technical representatives of
companies manufacturing agglomerating equipment. Most of the major chemical
companies were also consulted as were companies in the pulp-paper industry.
Late in the program, unpublished data from a study on pelletizing fines
were made available to this program. This study, performed by Babcock
Contractors for the U.S. Department of Energy, Pittsburgh Mining Technology
Center, concentrated only on disc pelletizing but the data should still
extrapolate well to other processes. All binders that Babcock Contractors
found promising were already included in the candidate list.
* References are listed on page III-114.
III-4
-------
TABLE III-l. AGGLOMERATION METHODS
H
M
Ui
Comments
Moisture Required (w/o)
Lubricant Required
Binder Required
Maintenance
Power Costs
Die Wear Costs
Labor Required
Operational Problems
Agglomerate Size
Relative
Size Distribution of
Feed Material
Relative
Agglomerate Density
Relative
Output
Drying Required
Relative Strength
Other
Mill
Pelleting
10-20
Desirable
Yes
High
Moderate
High
Moderate
Possible
Small Cylinders
Moderate
Moderate
Small
Yes
Moderate
Generates Heat
Briquetting
4-10
Desirable
Not Necessarily
High
High
High
Low
Few
Assorted Shapes
and Sizes
Large
High
High
Possibly
High
Feed Screw
Wear Possible
Disc
Palletizing
10-20
No
Yes
Low
Low
N/A
Moderate
Possible
Spheres 1/4-1 in. dia.
Small
Low
Moderate
Yes
Low
Longer Start Up
and Waste
Auger
Extrusion
10-20
Yes
Yes
. Moderate
High
High
Moderate
Possible
Continuous
Ribbons or Cylinders
Moderate
Moderate
High
Yes
Moderate
Mixing May
not be Required
-------
TABLE III-2. BINDERS EVALUATED DURING THE
PHASE III PROGRAM
Inorganics
Cement Bentonite clay
Epsom salt Calcium borate
Magnesium sand
Magnesium hydroxide
Phosphoric acid
Sulfuric acid
Natural Organics
Potato starch Creosol
Corn starch Anthracene oil
Parafin Coal tars
Asphalt emulsion (5) Pitches
Lignin Fuel oil
Ligno sulfonate Sawdust
Waxes Wheat flour
Dextrose
Processed Organics
Dowell M-107.01 latex resin
Dowell M-166 latex resin
Goodrich X-183 latex resin
Goodrich X-83 latex resin
Dow methocel F (methyl cellulose sum)
Dow methocel A (methyl cellulose gum)
Borden polyco (3) vinylacetate copolymers
Dow P-4000 polyethylene glyco copolymer
Dow E-8000 polyethylene glyco copolymer
Union Carbide PEG-100, polyethylene glyco copolymer
Union Carbide PEG-200, polyethylene glyco copolymer
Epon G-15 epoxy resin
Sterotex* hydrogenated cottonseed oil
Allbond 200, pregeletanized corn starch
III-6
-------
MECHANICAL PROPERTIES
Limestone/coal fuel pellets must have sufficient mechanical
strength, durability, and weatherability to survive the severe tumbling,
crushing, and weathering actions associated with typical fuel handling
and storage for industrial stoker firing. The following laboratory test
procedures were developed to evaluate different formulations for mechanical
properties.
Pellet Durability Index (PDI)
The PDI is determined using the equipment and procedure specified
by ASAE (Reference S-269.1). This test, similar to the ASTM tumbling index
test, simulates the tumbling action and impact forces encountered in handling
operations. The PDI indicates the percentage of pellets that remain intact
after such handling. In this test, a 500-gram sample is placed in a 0.6 m
square, 127 mm wide test bin, which is rotated at 50 rpm for 10 minutes.
The contents are then screened (4 mesh) and the +4 mesh fraction weighed.
The PDI is defined as:
weight of +4 mesh fraction (grams) .,nn
PDI - X J.UU.
FuelPellet Strength Test
This test and the associated equipment (Figure III-l) were designed
to simulate the crushing forces a pellet might experience during storage in
a pile, transportation by truck or railroad car, and feeding and handling.
The pellet is placed between two 25-mm diameter circular plates so that
radial forces are applied. (Cylindrical pellets are usually capable of
withstanding greater axial forces than radial forces.) The top plate is
2
attached to an air-driven plunger. The pressure on the 775-mm plunger is
increased until the pellet fractures. The compressive force required to
crush the pellet is obtained by multiplying the air pressure by the plunger
area, i.e.,
F Ibs = psi x 1.2
III-7
-------
2.54 cm diameter
circular plate
Pressure gage
Effective area
6.8cm2
I
Air supply
Vent
Test specimen
FIGURE III-l. APPARATUS FOR TESTING FUEL PELLET STRENGTH
III-8
-------
Stoker Fuel Pellet Weatherability Index (WI)
This test is designed to simulate the severe weathering conditions
a fuel pellet may encounter during outdoor winter storage. It has been
derived from ANSI/ASTM Test No. C666, "Resistance of Concrete to Rapid
Freezing and Thawing". The test consists of four freeze/thaw cycles of
4 hours each.
A 0.45 Kg sample of pellets is alternately placed in 4 C water
for 2 hours, then frozen in -18 C air for 2 hours. After four such cycles,
the pellets are drained and exposed to ambient room conditions overnight.
After the fuel pellets reach equilibrium with the room conditions, a screen
analysis (4 mesh) is run and the weatherability index is reported as the
percentage of whole pellets:
WI = lbs mesh) x 100
Postweathering Test
Since coal that is exposed to severe weathering conditions during
storage must retain its mechanical strength and durability characteristics,
the whole pellet samples were retested for durability and strength after
the weatherability tests. These retests are referred to as postweathering
tests.
Laboratory Test Results
During the pellet development work, the four production techniques
referred to earlier were used to produce 166 test samples. These samples
were evaluated for durability, strength, weatherability, and postweathering
durability and strength. The data from these tests are presented in
Table III-3. To provide a baseline for comparison and a target goal, four
different coal types and the 50/50 pellet from Phase II were evaluated using
the same test procedures. These are listed as the first five tests in
Table III-3.
III-9
-------
TABLE III-3. PELLET OPTIMIZATION DATA
o
03
Test
Ito
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Production
Method
Raw Coal
Raw Coal
Raw Coal
Raw Coal
20 Hp CPH Mill
Mars-Disk
Mars Disk
CPM Lab Hill
CPM Lab Hill
CPU Lab Hill
CPM Lab Mill
CPM Lab Hill
CPH lab Kill
CPM Lab Mill
CPM Lab Mill
CPH Lab Hill
CPH Lab Mill
CPM Lab Mill
CPM Lab Hill
CPM Lab Mill
CPH Lab Mill
CPM Lab Hill
CPM Lab Hill
CPM Lab Mill
CPM Lab Mill
CPH Lab Mill
CPU Lab Kill
CPM Lab Hill
CPH Lab Hill
CPM Lab Mill
CPH Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPH Lab Hill
CPH Lab Mill
CPM Lab Mil]
CPM Lab Hill
CPM Lab Mill
Disk by Coalmet
Disk
Disk
Disk
Disk
Disk
Coal
Type
Illinois it,
I. Kentucky
Lignite
Rosebud
111 mo
Illlno
II lino
lllino
Illino
111 mo
Illino
lllino
Illino
Illtno
Illino
lllino
111 ino
Lignlt
Illino
Illino
Illlno
Illino
Illino
Illino
I) lino
Illlno
Illino
Illino
Illino
Illino
Illino
Illino
lllino
Illino
Illino
Illino
Ulino
Illlno
Illino
Illino
Illlno
Illino
Illino
Hlim>
Illino
s iff,
s tt
s #6
s 16
s K
s K
s 16
s rt.
s 16
s if)
s 16
s »6
s 16
s It
s «6
s 16
s If,
s 16
s It
s it
s 16
s 16
s 16
s 16
s 16
s 16
s 16
s 16
s if,
s 16
s 16
s 16
s (6
s #6
s 16
s It,
s K
s 16
s 16
s 16
Pellet Formulation*8'
i
100
100
100
100
50
50
50
70
70
70
70
70
70
70
70
70
70
100
70
70
70
70
70
70
70
70
70
70
50
60
70
80
70
80
70
70
70
70
70
70
83
71
71
70
70
Limeston
Type
--
__
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Pi qua
Piqua
Piqua
Piqua
Lime (CaO)
Piqua
Piqua
Fedonia
Basic Lime
Howard
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
e
X
-
._
50
50
50
30
30
30
30
30
30
30
30
30
30
0
30
30
30
30
30
30
30
30
30
30
50
40
30
20
30
20
30
30
30
30
30
30
17
29
29
30
30
Binder
Type *
-I
.-
Cement 5
"Super" 10
"Super" 10
Epon 81 SI 4
16 Fuel Oil 5
16 fuel Oil 2
Sawdust 5
Sawdust + Oil 7
Epan 8151 5
Coal Tar 5
Cresol 5
Anthracene Oil 5
Indolimat Oil 5
None
M-167.01 5
M-166 5
Coal Tar 5
Cresol 5
Anthracene 5
Koppers Solvent 5
Asphalt - £65 5
Asphalt - [4868 5
Asphalt - 13GR 5
Anthracene Oil 5
Cement 5
Cement 5
Cement 5
Cement 5
Cement 5
Cement 5
H-167.01 1
M-166 1
Cement 5
Cement 5
Cement 5
Coalmet ?
Asphalt 4868 8.3
Asphalt 4868 8.5
Asphalt 4868 8 5
M-167.01 2.5
M-167.01 1 2
Remarks
Baseline Data
Baseline Data
Baseline Data
Baseline Data
Baseline Data
Air Dried
Oven Dried
Epoxy Resin
(5 Sawdust -2 16 Oil)
Oven Cured
Hlxed & Run
Hixed a Run
Mixed & Run
Lignin
Pelletlzed Lignite Fines
Dowel! Latex
Dowel 1 Latex
Hix to Heat Treat - 1 hour 9 160 C
Mix to Heat Treat - 1 hour 9 160 C
Hix to Heat Treat - 1 hour 9 160 C
Hix to Heat Treat - 1 hour (» 160 C
Emulsion from Armak
Emulsion from Armak
Emulsion from Arniak
Test #16 Used - 19 days
Ca/S Ratio Series
Ca/S Ratio Series
Ca/S Ratio Series
Ca/S Ratio Series
Production Run
Reagent Grade CaO
Dowel 1 Latex
Dowel 1 Latex
Limestone Type Series
Lime tone Type Series
Lime tone Type Series
Prop ietary Organic
Coar e Coal - 8 mesh
Coar e Coal - 8 mesh
Fine Coal - 100 mesh
Dowel 1 Latex
Dowel 1 Latex
Durability
Index(b)
85 +- 2
85 v 2
77 + 4
84 + 2
37 + 10
--
15 + 2
17 + 2
8+2
1 + 1
7 + 2
0
25 + 5
2 + 1
15 + 1
3 + 2
18 + 2
79 t 1
61 + 1
23 + 7
7 + 2
21+2
25 + 5
13 + 2
9+3
4 + 2
28
49 + 2
43+3
45 + 2
26 + 0
44 + 0.5
49 + 4
9+3
4 + 1
25 + 1
32 + 3
19 t 1
96
86
28
15
2
3
Compression Po" Ueattarins
Strength, Weather Durability(b) Strength,*
Ib * IndextM Index Ib
74+12 89+1 75 58 34
33 + 22 94 + 1 83 94 16
92+22 80+4 34 45 12
50 + 15 79 + 2 20 68 29
42+1 59+2 40 2
24+8 --
19
10+1
8 + 2 --
17+5 --
6+1 --
4 t 1 __ -
3 + 2 --
24+3 --
21+7 --
38 +10 --
17+5 --
12+4 --
88 +15 --
49 +15 -
33 + 5 -
29 +17 --
5+1 ~
29+1 --
19 t 5 -- -
25 t 4 -- -
14 + 2
i34 -- -
38 + 3 -
39+5 --
36+4 --
32 + 3 - -
49+9 -
46 +11 ...
22+3 --
19+5 --
26+3 -- -
32+7 ...
36 + 2 - -
92 + 19 lOO(c)- 53 + 15
22+1 a -
8+3 a -
-,, r o -
.>, 1 (,
^1 .
(a) Water added as needed
(b) Percent survival = 100 - percent fines.
(ci Some evidence of fracturing
-------
1-1
c
o
a
01
I
w
if.
' ss
Hi
i
r- in O O O r-
; £ s I
: I I«i^ =
V VI >i o >i >l^:
- x s< a "3 x -^ *= i t
SfUOlSoi- >>>>-o~o
. ---.r
a o oaa!ii£EO
-------
TABLE III-3. (continued)
Pellet Formulation*9'
Test
No.
107
108
109
110
111
112
113
114
115
116
H H7
l I
1 118
1 '
r? 119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
Production
Method
CPM Lab Hill
CPM Lab Mill
CPM Lab Mill
Briquetts
Briquetts
Briquetts
Briquetts
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
Banner Extrusion
Banner Extrusion
Banner Extrusion
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPH Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab dill
CPM Lab Mill
CPM Lab Hill
CPU Lab Mill
CPM Lab Mill
Banner Extrusion
CPM Lab Mill
Coal
Type
Illinois #6
Illinois #6
Illinois 16
Ohio
Ohio
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Unknown
Illinois #6
Illinois #6
Illinois 16
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois $6
Illinois #6
Illinois #6
Illinois #6
Illinois 16
Illinois 16
Illinois #6
Illinois 16
Limestone
%
70
70
70
70
70
70
70
70
70
70
70
70
70
100
70
70
70
70
70
70
70
70
89
94
71
73
70
70
70
Type
Piqua
Piqua
Piqua
(?)
(?)
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
CaO
CaO
Piqua
Piqua
Piqua
Piqua
Piqua
%
30
30
30
30
10
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
11
6
27
23
30
30
30
Binder
Type
P-4000
E-8000
PEG- 100
Unknown
Unknown
Methocel f
Cement
PEG-200
Polyco 11755
Polyco 2140
Polyco 2136
Cement
Methocel F
All bond 200
Polyco 2136
Polyco 2136
Polyco 11755
Polyco 11755
All bond 200
All bond 200
Cement
Cement
Cement
Cement
Cement
Cement
Blend M-6
Methocel
Blend F-l
%
5
5
5
1
5
5
5
5
5
5
1
1
1
0.5
1
0.5
5
1
5
5
5
5
5
5
1.5
1
3
Compression
Durability Strength, Weather
Remarks Index"" Ib * Indexlb)
Dow polyethylene
glycol
Dow polyethylene
glycol
Union Carbide poly-
ethylene glycol
Supplied by
Evergreen
Supplied by
Evergreen
Methyl cellulose
gum
Bentonite added
Union Carbide poly-
ethylene glycol
Bordon-vinyl acetate
latex
Bordon-vinyl acetate
latex
Bordon-vinyl acetate
latex
M-167.01 added for
plasticity
Methyl cellulose gum
Pregellatinized corn
starch
Bordon-vinyl acetate
latex
Borden-vinyl acetate
latex
Borden-vinyl acetate
latex
Borden-vinyl acetate
latex
Pregellatinized corn
starch
Pregellatinized corn
starch
0.1% Nad added as
catalyst
0.1% Fe-, 0., added as
catalyst
Would not pelletize
1.2* CaO added
3.1* CaO added
0.5% Methocol +1.0%
M-167.01
Methycellulose gum
2% All bond 200 +
W M-167.01
14
1
6
25
96
60
9
7
71
51
78
--
59
45
46
46
78
92
-
46
62
57
91
92
95
26
5
18
30 100
36 100
28 0
8 0
10
73 94
71
72 89
43 .0
> 113
80
40
33
20
23
96 98
85
-
35
41 62
47 64
109 40
101 0
111 87
Post Weathering
Durability^) Strength,
Index Ib*
24 43
80 53
--
66 90
53 90
..
--
--
89 111
32
38
36
112
-------
TABLE III-3. (continued)
H
H
H
O
o.
Pellet Formulation(a)
Test
No.
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
Production
Method
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
Banner Extrusion
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab Mill
CPM Lab tlill
CPM Lab Mill
Coal
Type
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Muskingum
Muskingum
Muskingum
Muskingum
Illinois 16
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Illinois #6
Limestone
%
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
62
98
90
87
89
89
70
70
70
100
Type
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
Piqua
CaO
CaO
Piqua
Piqua
Piqua
%
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
31
7.1
11
11
11
30
30
30
--
Binder
Type
Blend F-2
Blend M-7
Blend M-8
Blend F-3
Asphalt
47808
Asphalt
47808
Asphalt
Asphalt
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend M-9
Blend F-4
Blend F-5
Unknown
Unknown
Unknown
Unknown
Allbond 200
Allbond 200
Blend F-6
Blend A-l
Blend F-l
Allbond 200
%
3
1.5
1.5
3
5
10
5
10
2
2
2
2
2
2
2
2
2
6
2.5
1.5
1.9
1.7
1.6
5
10
10
9
3
2
Compression
DurabiJJJby Strength,
Remarks Index10' Ib *
2% Allbond 200 +
n Polyco 11755
0,5% methocel + U
Polyco 11755
0.5% methocel + 1%
Polyco 2136
2% Allbond + 1%
Polyco 2136
Bituminous Mat. Co.-
slow setting
emulsion
Bituminous Mat. Co.-
slow setting
emulsion
Bituminous Mat. Co.
med. setting
emulsion
Bituminous Mat. Co.
med. setting
emulsion
Low density
Medium density
High density
High density
Low density
Low density
Low density
Medium density
High density
5% Allbond 200 +
1% M-167.01
1.5% Allbond 200 +
1% M-167.01
Banner sample 10M
Banner sample 11M
Banner sample 12M
Banner sample 13M
5% Allbond 200 +
5% asphalt 47808
5% asphalt 47808 +
4% 11-167.01
ReproducibiHty
check on Test 137
92
55
52
87
64
67
2
36
75
86
80
81
82
85
85
89
87
99
94
38
0
92
78
96
92
82
74
98
112
54
59
36
48
40-110
70
72
81
66
70
70
90
62
110
84
110
88
73
102
24
0
64
76
> 112
97
Post Weathering^
Weather. Durability'15) Strength
Index(b) Index Ib *
93
62
62
100
75
76
0
63
100
100
--
91
90
--
--
84
30
27
85
30
10
2
98
62
88
42
--
> 112
> 112
43
30
30
> 112
60
--
> 112
48
--
-------
Baseline Data
The data in Table III-3 indicate that the eastern bituminous
coal was superior to the others in mechanical properties. This coal
retained its mechanical strength characteristics after the weatherability
tests. Those values listed in Table III-3 for the eastern bituminous
coal (Run No. 2) were used as the target goals.
Lignite and western subbituminous coals are susceptible to
weathering as evidenced by the postweathering indexes. The cement-bound
pellet from Phase II is clearly inferior to the stoker coals.
Effect of Production Technique
The characteristics of pellets produced by each of the four
methods are summarized in Table III-4. These characteristics based on
the laboratory standard tests are similar to those of eastern bituminous
coals and superior to those of lignite and the western subbituminous
coals.
The data in Table III-4 indicate that pellets produced by mill
pelleting and auger extrusion have comparable mechanical strength
characteristics. The mechanical strength characteristics of the disc
pellets and briquets are clearly inferior. Disc pellets, however,
exhibit a good PDI compared to their compression strength. This
characteristic is attributed to the spherical shape, which means there
are no corners to be broken. The disc pellets, however, have relatively
low weatherability. The high porosity of these pellets allows water to
penetrate them readily, dissolving any water soluble binder and creating
fractures during freezing. The briquets could not be readily procured
and thus extensive evaluation could not be made.
III-ll
-------
TABLE III-4. COMPARISON OF TEST RESULTS BY
PRODUCTION TECHNIQUE
Test
No.
82
136
112
71
(a)
Production Technique
Pelleting - CPM
Auger extrusion - Banner
Briquetting - Evergreen
Disc palletizing - Mars
Binder
Loading,
percent
1
1
1
5
Compression ^
PDI
96
92
60
99
Strength,
498
449
124.5
267
N
(a) All tests have been formulated as follows:
70 percent coal
30 percent limestone
Methylcellulose binder
* 1 Ib = 4.448 N
Effect of Binder
The most desirable binders are those that can produce pellets
with physical properties similar to conventional eastern bituminous coals
and yet be relatively inexpensive. Since the binder can contribute to as
much as 20 percent of the cost of producing a limestone/coal fuel pellet
with a Ca/S ratio of 3.5, economics were considered in binder selection.
Inorganics. Six inorganic binders were tested. With these
binders, the amount of limestone affected the mechanical strength, with
higher limestone additives producing stronger pellets. The calcium appears
to present an active attachment site for the binders. None of the pellets
produced with these binders were as strong as the cement-bound pellet.
Because of this, and a feeling that inorganics add to the ash content and
would inhibit combustion, no additional inorganic binders were tested.
111-12
-------
Naturally Occurring Organics. The distinction between naturally
occurring and processed organics is somewhat arbitrary. Naturally
occurring substances there were limited to materials like starches and
asphalts that need minimal additional processing and/or waste materials
such as lignin or ligno sulfonates that are readily available. Due to
the availability and low cost of these materials and the potential for
enhanced combustion, the Phase II effort had focused heavily in this area.
The best binder investigated in this category was a slow setting asphalt
emulsion supplied by Bituminous Material Co. Pellets with this binder
showed a PDI of 64, a compression strength of 240 N, and a WI of 75.
This represents an improvement over the cement bound pellets but does
not meet the larger goals. Relatively high (5-10 percent) binder loadings
are required to produce these pellets. No other binder in this category
produced pellets worthy of further consideration.
Processed Organics. Included in this category are synthetics
such as methycellulose and further processed natural organics such as
Allbond 200, a gelatanized cornstarch. These products offer the optimum
potential for strength, weatherability, and combustion enhancement.
Many of these materials do in fact produce pellets with the desired pro-
perties. The major drawbacks to this category for a commercial applica-
tion are cost and availability. For example, methyl cellulose, at today's
prices, might add as much as $40 to the cost of 910 Kg of pellets.
The Allbond 200 binder, with a strength of 30 nt, a PDI of 78, and a WI
of 98, while adding only $5.50 Mg ($5/ton) to the cost of pellets, was
the most promising. This binder is 20 percent water soluble so there is
some concern about long-term weatherability.
Combustion Characterization. The standardized laboratory tests
provided relatively quick comparison of the mechanical strength charac-
teristics of a wide variety of pellet formulations produced using different
techniques. To complement this effort, small-scale tests were conducted
to determine the combustion characteristics (sulfur capture and reactivity)
of various formulations of limestone/coal pellets. Variables that were
considered include:
111-13
-------
Ca/S ratio
« Sorbent type
9 Production technique
9 Binder type and concentration
Catalysts or other additives.
A tubular, fixed-bed reactor that could be operated to 1300 C was selected
for these tests.
Fixed-Bed Reactor'Experiments. The tubular fixed-bed reactor pro-
vides a relatively fast method of evaluating a matrix of variables while
minimizing fuel preparation costs. The more promising formulations can then
be evaluated on a larger scale.
Fixed-Bed Reactor. Figure III-2 is a drawing of the fixed-bed
reactor test facility. This facility has the following features to simulate
fixed-bed burning and provide for collection of data of interest;
High-temperature (1300 C) operation
Controlled temperature set point
Controlled air input
Small charge-batch operation
Analysis of off-gases
« Easily collectible ash
* Rapid quenching of the combustion reactions
No wall effects.
The fixed bed reactor consists of a tubular electrical resistance
furnace capable of heating the quartz reactor tube to 1300 C. Furnace
temperatures can be controlled by varying the electric power input.
Primary combustion air is introduced at the bottom of the bed. A porous,
stainless steel distributor plate, located 50 mm inside the furnace hot
zone, supports the fuel bed and provides for air distribution. Test-bed
depths of 50 to 300 mm are possible, and the small quantity (50 to 300
grams) of fuel required minimizes fuel preparation costs. Secondary air
is preheated and introduced above the bed by a helical coil that is heated
by the furnace. Both air flows are metered and controlled between 0 and
.0005 m /sec (0 and 1 cfm) by rotameters. Rapid introduction of inert gas
111-14
-------
Heated
1/4" S/S
*
Sample Line
Particulate
Tube for ,!, E//.X Filter
Pre-Heat-*
of Over
Fire Air
0
g
{/
V
\
y
/*
K
+4^-^
\w^
^
^-^3^5
v_ A /s /
^
CL'
4
I
^
%
f
^'
£
l^D
1^
^
>*
W
/
Filter
Flex
Tef
--] ^~"
ater
Diaphraqt
line- Pump
'°n ^-r, ^>
^m^ j-
/ _!_.
n
/^s
/ u
Particulate LLk. '
Filter 1 Flnw
Meter
Trao
Ice Bath
^-Oven
1
Over
Fire
Air
Distribution
'/,
'/
y
^
-^
~1
-L
v
T
^
V
-*-
Tu
Manifold
TECO
S0? Faristor
S02
1 I
V V
0(
(,
Jl i
-Flow
Meter
)
NDIR
CO
I
V
f
{
i
NDIR
C02
V
ji
-it Vent/
Taylor
f
V
(
i
-* ^Flow
Met
)
_Q p Span/Zero Gases
nder Fire
Air
. I-
C°m^reSSed TUBULAR FIXED BED REACTOR
FLOW SCHEMATIC
FIGURE III-2. FIXED-BED REACTOR
111-15
-------
(N?, C0?, or argon) can quench the combustion reactions at any time.
Temperatures can be monitored at any chosen point in or above the bed
by thermocouples introduced through holes provided in the distributor
plate,
A probe inserted in the reactor tube exit withdraws samples of
the flue gas through a conventional stack sampling train for
analysis. The sample is continuously analyzed by the following equipment:
» Taylor paramagnetic analyzer for oxygen
e Beckman NDIR analyzer for CO and CO
Faristor analyzer for S0?
TECO pulse fluorescence analyzer for SO .
The output from the analytical instruments, as well as all thermo-
couples, is recorded on a data acquisition system for later processing by
computer.
The fixed-bed reactor also provides ash samples with a clearly
defined history for chemical and metallographic analysis. Wet chemical
analyses for S09 and H_S are made for selected runs using the EPA Method 6.
Experimental Procedure. After systematically investigating
different modes of operation for the fixed-bed, the following experimental
procedure was developed, based on the most consistent set of runs:
1. Preheat the furnace to the selected temperature.
2. Load the reactor tube with the selected fuel charge.
3. Connect the air lines, establish flow rates, and
place thermocouples.
4. Insert the charged reactor tube into the heated furnace.
5. Place the sample probe in the outlet.
The coal ignites in the preheated furnace. The run is continued until
the C02 and S02 instruments return to zero indicating the combustion
reactions are complete. This typically takes about 30 minutes. When
combustion is complete, the reactor tube is removed from the furnace
and allowed to cool. The residue is then collected, weighed, and analyzed
for carbon, sulfur, ash, and moisture content. From these data, the mass
of carbon and sulfur remaining in the residue is determined. The tape from
the data acquisition system containing the flue gas analysis is input to
111-16
-------
produces concentration-versus-time profiles of O^j C02, co> an
-------
TABLE III-5. REPRODUCIBILITY SERIES WITH 50/50
PELLETS, FUBNACE TEMPERATURE 1040 C,
WITH 33% OVERFIRE AIR
Run D 44 45 46 47 Average
Percent C Accounted
Flue gas 98.5 88.6 93.4 92.6 93.3 ± 4.1
Total 99.0 89.2 93.9 93.3 93.9 ± 4.0
Percent S accounted
Flue gas 17.5 12.3 17.8 18.1 16.4 ± 2.8
Total 92.1 86.9 97.3 93.5 92.5 ± 4.3
Run time 19.78 18.00 17.85 19.92 18.89 ± 1.11
Max SO,, 2646 1878 2213 2942 2420 ± 469
fuel sulfur was being accounted for. It was also noted that the Faristor
S02 analyzer was malfunctioning, possibly because of H-S poisoning. Since
these effects were not observed previously, a series of tests (59-69)
were made to check these results and to determine the effect of the
increased coal charge. These runs indicated that the increase in coal was
contributing in part to the sulfur-balance problems. During this period
(Run 65), the TECO S02 analyzer was added to the sampling system. Correc-
tions for interference by other stack gas constituents were found to be
minimal and easily handled by the computer. The S02 curves produced by
the Faristor and the TECO analyzers have some marked differences. The TECO
data frequently show a higher peak value. This is probably due to a faster
response time. Of more interest is a TECO drop near to zero during the
period of low oxygen. This dip, which is more severe at higher coal charges,
is felt to represent a shift from SO^ formation as less oxygen becomes
available. The Faristor, which responds to H2S, shows no dip. However,
sensitivity is severely degraded by prolonged exposure to H?S.
111-18
-------
TABLE III-6. FIXED-BED EXPERIMENT RUNS
Distribution
Run
No.
48
49
50
51
52
53
54
55
56
57
58
59
60
61
64
65
67
68
69
70
71
72
73
74
75
76
77
78
79
80
31
82
83
84
86
87
88
89
Fuel
No.
29
30
31
32
34
37
38
39
71
82
76
32
34
37
5
82
5
31
37
52
53
54
112
113
118
119
128
130
31
31
126
127
132
133
157
158
159
160
Fuel
Charge,
Fuel Type grams
50/50 Cement (CPM)
60/40 Cement (CPM)
70/30 Cement (CPM)
80/20 Cement (CPM)
CaO (at limestone 70/30 ratio)
70/30 Fredonia (CPM/a^
70/30 Baseline (CPM)
70/30 Howard Limestone (CFM)
70/30 5Z Methocel (Disc)
70/30 5Z Methocel (CPM)
70/30 5Z Dowell (CPM>
80/20 Cement (CPM)
CaO (at limestone 70/30) (CPM)
70/30 Fredonia (CPM)
50/50 Cement (CPM WJ)
70/30 53! Methocel (CPM)
50/50 Cement (CPM WJ)
70/30 Cement (CPM)
70/30 Fredonia (CPM)
70/30 Epsom Salt (CPM) (a)
70/30 Magnesium Sands (CPM)
70/30 Magnesium Hydroxide (CPM)
70/30 1Z Methocel (Briquet)
70/30 Cement + Bentonite (Briquet)
70/30 Cement + Latex (Banner)
70/30 1Z Methocel (Banner)
70/30 0.11 NaCl (CPM)
70/30 0.1Z Fea03 (CPM)
70/30 Cement (1500 F)
70/30 Cement (1900 F)
70/30 5Z Allbond
70/30 1Z Allbond
94/6 Coal/CaO
70/30 +1.27 CaO
Banner 10M-30Z Limestone
Banner HM-Muskingum Coal
Banner 12M-7Z Limestone
Banner 13M-11Z Limestone
77
64
55
48
48
55
55
55
55
55
55
48
48
55
11
55
55
55
55
41
41
41
41
41
41
41
41
41
41
41
41
41
41
49
30
33
35
Run
Time,
min
23
25
27
26
24
30
25
25
22
28
30
27
29
28
23
25
20
27
29
25
23
28
36
51
29
30
25
26
25
21
28
26
24
21
34
34
38
30
Flue Gas
Z
S
10.96
11.03
11.41
19.97
7.05
12.86
22.07
13.74
24.72
15.26
20.10
22.64
10.43
25.93
16.69
25.91
16.85
15.69
17.31
15.82
20.21
17.18
19.38
18.69
14.18
15.52
25.33
18.63
24.82
23.45(b)
17.23 (b)
26.92
29.82tW
24.9(b)
54.3(b)
55.9°»
53.8(b>
Z
C
88.72
88.72
91.4
88.22
89.76
87.80
83.22
80.47
86.40
74.23
89.87
87.40
91.19
87.76
23.15
71.17
86.20
89.55
87.28
93.31
84.04
93.05
102.93
105.68
96.70
104.74
84.05
83.75
89.83
89.29
82.67
91.79
80.22
89.39
97.5
96.0
88.3
93.4
Ash
Z
S
62.84
65.23
60.9
61.38
87.05
68.49
71.38
76.71
71.20
66.75
75.0
62.18
78.40
62.08
92.27
64.46
80.33
70.50
58,32
68.93
73.17
67.00
72.10
66.42
73.31
76.83
63.73
62.60
74.07
64.01
63.19
60.27
51.89
79.10
58.4
1.0
7.9
14.5
%
C
1.45
1.14
2.26
0.23
0.96
1.70
1.29
1.91
1.81
2.20
2.35
0.75
1.61
1.32
0.56
0.05
0.72
0.07
0.40
0.25
0.58
0.49
0.78
2.90
0.83
1.25
0.19
0.34
2.09
1.87
1.50
1.89
1.56
2.24
0.8
0.1
0.2
0.4
Total
Z
S
73.80
76.26
72.3
81.35
94.10
81.35
93.45
90.45
95.92
82.01
95.10
84.82
88.83
88.01
68.96
90.37
97.18
86.19
75.03
84.75
93.38
84.18
91.43
85.11
87.49
92.34
89.06
81.28
98.88
87.46
80.42
87.19
96.91
108.92
83.3
55.3
63.8
68.3
%
C
90.15
89.86
93.66
88.45
90.72
89.50
84.51
82.38
88.21
76.43
92.22
88.15
92.80
89.09
83.71
71.22
86.92
89.63
87.68
93.56
84.62
93.54
103.71
108.58
97.53
105.99
84.24
83.79
91.92
91.16
84.17
93.68
81.78
92.18
98.3
96.1
88.5
93.8
S03
Max,
PP»
2017
2347
2263
2148
728
1567
4013
2864
3844
2145
2947
3635
1170
4449
2628
4036
1808
2860
2170
2888
3630
3248
3234
2172
2669
1634
2398
1032
2500
2789
3190(b)
2989 (b)
2704 <">
3664
2056
(a) Changed Faristor analyzer cell.
(b) Taco data.
111-19
-------
In light of these findings, it was felt that the decision to run
at 1040 C to produce the "real world" conditions more closely using the higher
coal charge, may have been an error. Runs 79 and 80 were made with
41 grams (28 grains of coal) of 70/30 cement bound pellets at 820 and
1040 C, respectively. The data show excellent agreement of the TECO
and Paristor instruments at 820 C. Further, the profiles are essentially
identical. At 1040 C, the TECO indicates a severe drop in SO concen-
tration but nearly the same dosage as at 820 C. The Faristor value is
low at 1040 C, one half the value at 820 C. The differences in Method 6
data, which implies more sulfur capture at the higher temperature, are
not understood.
Except for Runs 48-51, the sulfur balance is not exceedingly
low. But it is always low, never high as might be expected with normal
random fluctuations. This suggests that sulfur capture as determined
by the fixed-bed reactor tests may not be accurate in absolute terms.
However, the generally constant overall sulfur balance is reason to
believe that these tests rank the various formulations correctly and
thus satisfy their objective.
The TECO and Faristor comparisons indicated that a 28-gram
coal charge and 820 C furnace temperature, or possibly more secondary
air, would be desirable to eliminate problems with H S formation. However,
to conform with the previous work, the remaining runs were made at 1040 C.
It is interesting to note that with different charges of the same fuel,
run times were essentially the same. Furthermore, variability in the -
sulfur balance comes more from the ash than the flue gas composition. For
these reasons, it is felt that comparisons of runs at these conditions
are valid.
Discussion of Experimental Results
Table III-6 summarizes the data of Runs 48-89 from the fixed-bed
reactor experiments. As previously discussed, Runs 1 through 43 were for
checkout purposes and provide little direct information on the evaluation
of pellets. The results of reproducibility runs (44-47) were shown
111-20
-------
in Table III-5. The effect of pellet formulation variables and system
operating temperature on sulfur capture and reactivity are discussed
below.
Ca/S Ratio. The effect of Ca/S ratio was examined by making
cement-bound pellets in the CPM lab mill with various levels of Piqua
limestone. Table III-7 gives the results of these experiments.
TABLE III-7. THE EFFECT OF Ca/S RATIO ON
SULFUR CAPTURE
Run
48
49
50
68
51
59
Fuel
29
30
31
31
32
32
Limestone/
Coal
1.0
0.67
0.43
0.43
0.25
0.25
Sulfur (a)
Capture,
percent
89
89
89
84
80
78
Reactivity
(burn time)
minutes
23
25
27
27
26
27
Approximate
Ca/S Ratio
8
6
3.5
3.5
2
2
(a) Based on sulfur found in flue gas.
These results indicate that the Ca/S ratio (as reflected by the
limestone/coal ratio) has only a small effect on sulfur capture at limestone/
coal ratios at or above 0.43 (Ca/S = 3-4). At a limestone/coal ratio of
0.25 (20 percent limestone, Ca/S = 1.7-2.3), the sulfur capture begins to
drop noticeably.
The limestone/coal ratio may have a slight effect on reactivity.
Pellets with the higher limestone content appear to burn slightly faster.
Because of the higher ash content at higher limestone levels, an opposite
result was anticipated. Though lime is known to catalyze some gasification
and combustion reactions, the levels present here (20 percent) should swamp
any concentration effect. Perhaps this effect is due to the higher C0_
values within the pellet from the calcining limestone and the surface reaction
C + CO -» 2 CO taking place.
111-21
-------
Type of Sorbent. Piqua, Fredonia, and Howard limestones, as well
as "basic lime", (a precipitated CaCO of very small particle size), were
compared at a limestone/coal ratio of 0.43 (Ca/S - 3-9) in cement-bound
CPM pellets. Table III-8 gives the results of these tests.
TABLE III-8.
THE EFFECT OF SORBENT ON
SULFUR CAPTURE
Run
50
53
61
69
54
55
Fuel
31
37
37
37
38
39
Type of
Sorbent
Piqua
Fredonia
Fredonia
Fredonia
Basic Lime
Howard
Sulfur
Capture,
percent
88.6
87.1
74.1
82.7
77.9
86.3
Reactivity
(burn time) ,
minutes
27
30
28
29
25
25
These results show that "basic lime" is the poorest sorbent. The
three limestones appear about equal, with some indication that Fredonia
limestone is slightly inferior. These results are also contrary to what
was anticipated. The "basic lime" was included since it was the smallest
particle size available and it was believed that more finely divided CaCO,,
would prove to be a better sorbent. Perhaps the precipitation process
reduces the reactivity of this material.
The "basic lime" and Howard limestone pellets show the highest
reactivity, and the Fredonia limestone the lowest reactivity. The reasons
for this are unexplained.
Calcium Oxide. Calcium oxide and Piqua limestone were compared
at a Ca/S ratio of 3 to 4 in cement bound CPM pellets, as shown in
Table III-9.
111-22
-------
TABLE III-9.
COMPARISON OF THE EFFECT OF LIMESTONE
AND LIME ON SULFUR CAPTURE
Run
50
68
52
60
83
84
Fuel
31
31
34
34
132
133
Sorbent
Limestone
Limestone
CaO
CaO
CaO
Limestone
and CaO
Sulfur
Capture,
percent
88.6
84.3
93.9
89.6
55.0
70.2
Reactivity,
Burn Time,
minutes
27
27
24
29
24
21
Approximate
Ca/S
3.5
3.5
3.5
3.5
2.0
4.0
As expected, CaO is the superior sorbent. The limestone must be
calcined before it can react with the sulfur while the CaO can react
immediately. Any sulfur released prior to this calcining cannot be
captured by limestone but could be captured with CaO. This leads to the
speculation that a small addition of CaO to the limestone might significantly
improve the sulfur capture. This is not supported by the one test conducted
with this combination, where capture was poorer that with limestone alone.
Additionally, however, the significant reduction in sulfur capture in
Run 83 is surprising. Further work is needed to investigate lower Ca/S
ratios using CaO and CaO-limestone mixtures.
Under the test conditions, the substitution of CaO for limestone
does not appear to alter reactivity, but this conclusion is weak in view
of the widely different burn times obtained with pellets containing CaO.
Binder
Binders were compared with CPM pellets at a limestone/coal ratio
of 0.43 using Piqua limestone. Table 111-10 gives the results of fixed-bed
reactor experiments.
111-23
-------
TABLE 111-10.
THE EFFECT OF BINDER TYPE
ON SULFUR CAPTURE
Run
50
68
57
65
58
70
71
72
81
Fuel
31
31
82
82
76
52
53
54
126
Binder
Cement
Cement
Methylcellulose
Methylcellulose
Dowell Latex
Epsom Salts
Mag Sands
Mg (OH)
Allbond 200
Sulfur
Capture,
percent
88.6
84.3
84.7
79.6
79.9
84.2
80.9
82.7
82.8
Reactivity
Burn Time,
minutes
27
27
28
25
30
25
23
28
28
Peak Reactor,
temperature, C
1184
1173
1347
1178
1152
1130
1170
1095
1120
The cement binder appears most favorable for sulfur capture and
the organic binders appear slightly inferior to the inorganic binders.
The poorer performance of organic binders might be caused by increased
pellet temperature caused by combustion of the binder, or the volatility
of these organic binders may drive off additional sulfur. Also, the
organic binders were used at 5 percent concentration, which may be higher
than necessary for pellet strength and durability. At lower concentrations,
the organic binders might capture sulfur better.
The Epsom Salts and "Magnesium Sand" binders, which were
believed to have some catalytic properties, produced the most reactive
pellets in terms of run time, and the Dowell Latex the least reactive
pellet. The Allbond 200 cornstarch pellets were similar to cement-bound
pellets. Because of two rather different results in replicate runs, the
position of the methylcellulose pellets is uncertain, but may be approxi-
mately the same as the cement-bound pellets.
The peak temperature reached in the reactor does not correlate
with the run time, which is unexpected. The peak temperatures were similar,
with the exception of one run with methylcellulose pellets, which reached
a peak temperature 182 C higher than the average of the other runs.
111-24
-------
Pellet Production Method. Pellets with a limestone/coal ratio
of 0.43 using Piqua limestone and methylcellulose binder were produced
by each of the four production methods. Table III-ll gives the results
of the fixed-bed reactor experiments using these pellets.
TABLE III-ll. THE EFFECT OF PRODUCTION METHOD ON SULFUR CAPTURE
Sulfur Reactivity , ,
Capture, (burn time), Density
Run Fuel Method percent minutes grams/cc
56(b)
57(fe)
65 0>)
73
76
71
82
82
112
119
Pelletizing
Mill pelletizing
Mill pelletizing
Briquet
Extrusion
75.3
84.7
79.6
80.0
84.5
22
28
25
36
30
0.79
1.39
1.39
1.26
0.85
(a) Determined by displacement.
(b) These runs were made with a 55-gram fuel charge, whereas
only 41 grams were used in the other two runs listed.
Except for the extruded pellets, sulfur capture increased as pellet density
increased. This is because the higher density produces a more tortuous path
and lower diffusion rate for the escaping sulfur. Since excess calcium is
present, the increase in contact should permit the kinetically controlled
reaction to go further towards completion. The high performance of the
extruded pellets, in these terms, is anomalous.
Reactivity, as measured by burn time, does not correlate well with
the pellet density either, as shown in Table 111-12.
111-25
-------
TABLE 111-12.
THE EFFECT OF PRODUCTION
TECHNIQUE ON REACTIVITY
Pellet Type
Mill pelleting
Briquetting
Extrusion
Disc palletizing
Density
1 (high)
2
3
4 (low)
Ranking
Sulfur
Capture
2
3
1 (high)
4 (low)
Reactivity
2
4 (low)
3
1 (high)
Lower-density pellets should be more receptive to diffusion of
air into and combustion products out of the pellet, thereby promoting combustion.
The low-density, disc-agglomerated pellets support this premise by being
the most reactive. However, the second most reactive pellet is the CPM
pellet which has the highest density. Evidently sulfur capture and
reactivity are not related to pellet density or porosity in any simple way.
Additives. The addition of 0.12 percent NaCl or Fe-0,, to cement-
bound pellets with a Piqua limestone/coal ratio of 0.43 was expected to
enhance sulfur capture or, in the case of Fe 0_, possibly enhance burning
rate. Table 111-13 shows the results of these tests.
TABLE 111-13. THE EFFECT OF ADDITIVES ON SULFUR CAPTURE
Run
77
78
80 N
50 ^
68^
Fuel
128
130
31
31
31
Additive
0.1% NaCl
0.1% Fe203
None
None
None
Sulfur
Capture,
percent
73.6
80.3
76.5
88.6
84.3
Reactivity
(Burn Time) ,
minutes
25
26
21
27
27
(a) These runs were made with a 55-gram fuel charge,
whereas only 41 grams were used with the other
three runs listed.
111-26
-------
Neither additive appears to be beneficial, although the comparisons
are clouded by the noted change in the fuel charged to the reactor.
Conclusions from Fixed Bed Reactor Study
The data from the fixed bed reactor tests showed few effects of
formulation on reactivity. Often, the effects that were observed were
opposite of those expected. This might indicate an insensitivity of the
test equipment and procedure to actual reactivity.
Bearing this in mind, the conclusions reached from this study are:
« Ca/S ratio is relatively unimportant at values
above 3-4. Sulfur capture is only slightly
lower at a Ca/S ratio of 2.
* Three limestones gave comparable sulfur capture
with only the precipitated "basic-lime" inferior.
9 CaO is a superior sorbent to limestone. A small
amount of CaO in conjunction with limestone might
be expected to improve sulfur capture, but this
effect was not observed.
* Use of organic binders instead of inorganic sub-
stances results in slightly lower sulfur capture.
This disadvantage might be reduced at lower binder
concentrations.
Sulfur capture increased with pellet density
when comparing production methods with the
exception of the anomalous extruded pellets.
The pellet mill produces highest density and
highest sulfur capture.
NaCl and Fe 0 additives (0.12 percent) did not
improve sulfur capture or reactivity.
The effect of the variables on pellet reactivity,
as evidenced by the burning time, were not clearly
distinguished. Some differences were observed.
111-27
-------
SECTION 4
PROCESS VARIABLES SELECTION
A mathematical modeling analysis, in combination with a series
of experimental studies, was performed to develop a more comprehensive
understanding of the processes that influence the burning of the coal/
limestone pellets and control sulfur capture. The purpose of the
experimental studies was to provide physical and chemical rate constants
for the model, and to provide a basis for assessing model reliability.
The main objectives of the modeling studies were to predict burning
rates and sulfur capture as a function of pellet properties such as
size, composition, and physical structure under different combustion
conditions. Knowledge of how these parameters affect pellet performance
is important for the development of optimum pellets for specific boiler
applications.
The mathematical model described here is a preliminary model
with a number of simplifying assumptions incorporated into it. An
important purpose of this effort was to demonstrate how modeling can be
practically and usefully employed to understand how coal/limestone
pellets behave and how to improve their performance. The basic structure
of this preliminary model can be expanded to include either additional
processes or refined mechanisms to increase its accuracy.
In the model, the pellet's cylindrical geometry is represented
by an equivalent sphere having the same surface-to-volume ratio. The
burning of the pellet is represented by the shrinking core model developed
, TT , , (III-3 to III-6) _ . , , ,
by Wen and co-workers. It is based on a burning reac-
tion zone that penetrates into the pellet and is supported by oxygen
that diffuses through the ash layer surrounding the burning core. SO ,
formed at the burning interface between the unburned core and ash, is
assumed to diffuse out through the ash layer where it is captured by the
CaCO., in the limestone. The calcination process is represented by the
reaction
111-28
-------
-> CaO + C02
and sulfur capture by the reaction
CaO + S02 + 1/2 02 -> CaSO, .
SO that does not react with the CaO while diffusing to the outer surface
of the ash layer is considered to be released from the pellet. The
model, in its present form, neglects factors such as heat transfer,
devolatilization of the coal, the reaction between CO,., and carbon forming
CO, and solid-state reactions. These mechanisms can be incorporated into
extended versions of the present model as appropriate.
The experimental studies associated with the modeling effort
were performed using the fixed-bed reactor described in detail in
section 4. These experiments involved heating pellets in the fixed-bed
reactor in a flowing air stream for various periods of time corresponding
to different levels of fuel consumption in the pellets. The combustion
gases exiting from the reactor were analyzed continuously for .S0?, CO,
and 0~ to follow the kinetic rates of the combustion and S0?-release
processes. Burned pellets were recovered, weighed, analyzed chemically,
and subjected to examination by metallographic techniques, scanning
electron microscopy, and X-ray diffraction. Some limitations, however,
exist in accurately applying some of these data to the model because of
uncertainties in the fixed-bed reactor experiments. These difficulties
can be eliminated in future work by using the model developed here as
a guide for designing the experiments.
Further details on both the fixed-bed reactor experiments and
the model are presented below.
PELLET DETAILS
Two different pellet compositions were used in these studies:
70 wt. percent coal/30 wt. percent limestone and 50 wt. percent limestone.
These pellets were produced at Battelle using a California pellet mill.
In addition to the coal and limestone, the pellets also contained corn
starch and latex at levels equivalent to 2 wt. percent and 1 wt. percent,
respectively, of the weight of coal plus limestone. The latex contained
50 wt. percent solids. The pellets were cylindrical in shape with a
12.5 mm diameter. The ends of the pellets had a fracture texture with
111-29
-------
a variable overall length of approximately 12 mm. Variability in
the end conditions and lengths of the pellets may have been responsible
for some of the variations noted in the experimental results. For
comparison purposes, some information is also included on the Banner
and Alley Cassetty pellets used in the steam-plant demonstra-
tion experiments performed at Battelle in November 1979.
Information on densities and porosities for typical pellets
is summarized in Table 111-14. The density values were calculated from
measured volumes and weights of typical pellets. Porosities were
calculated using densities for the solid coal, solid limestone, corn
3
starch, and latex of 1.21, 2.60, 1.53 and 0.92 g/cm , respectively. The
Alley Cassetty pellet ran about 8.5 percent less dense than the Battelle
70/30 pellet, while the Banner pellet was over 12 percent less dense.
Results of chemical analyses on the coal used to prepare the
Battelle pellets and on typical pellets are given in Table 111-15.
Calcium-to-sulfur ratios and the percentage of total sulfur occurring as organic
sulfur for the Battelle pellets were:
Pellet Calcium/Sulfur Ratio, Organic
Composition g-moles Ca/g-moles S Sulfur, percent
50/50 5.04 58.15
70/30 3.14 51.43
Total sulfur occurring as organic sulfur in the raw coal was about 51.74
percent corresponding closely to the result of the 70/30 analysis.
FIXED BED REACTOR EXPERIMENTS
The fixed-bed reactor experiments were conducted mainly with the
Battelle 50/50 and 70/30 pellets with the exception of several runs with
pellets of only coal and only limestone. Each experiment involved three
pellets that were rapidly introduced into the high temperature zone of the
furnace via the quartz tube using an open wire support designed for
minimum contact with the pellets. Thermocouples attached to the wire cage
were used to determine pellet temperatures. At the end of a preselected
burn time the air flow passing through the quartz tube was changed to argon
111-30
-------
TABLE 111-14. TYPICAL PELLET DENSITIES
AND POROSITIES
Pellet Pellet Pellet Density, Porosity,
No. Source Composition g/ciiH percent
70 Battelle 50 w/o coal/50 w/o limestone 1.50 9
76 Battelle 70 w/o coal/30 w/o limestone 1.41 2
30 Battelle 100% coal 1.18 2.5
16 Battelle 100% limestone 1.84 29
200 Alley-Cassetty 70 w/o coal/30 w/o limestone 1.29 10
202 Banner 70 w/o coal/30 w/o limestone 1.24 15
111-31
-------
TABLE 111-15. SUMMARY OF CHEMICAL ANALYSES ON
RAW COAL AND PELLETS
Chemical Composition, w/o
M
M
1
U)
M
Pellet
No.
79
80
81
205
206
18
Pellet
Source
Battelle
Battelle
Battelle
Banner
Alley-
Cassety
Battelle
Pellet
Composition
*
Raw coal
100 w/o coal
70 w/o coal/30 w/o
limestone
50 w/o coal/50 w/o
limestone
70 w/o coal/30 w/o
limestone
70 w/o coal/30 w/o
limestone
100 w/o limestone
C Ca C03
59.5 0.67
63.0 0.50 1.03
46.4 11.0 20.1
39.4 17.0 29.7
45.4 10.1 17.7
46.6 10.8 17.4
11.4 31.0 53.7
Fe Ash
3.7 15.6
1.99 12.7
1.43 29.0
1.34 37.8
2.53 32.4
1.22 26.5
51.0
H20
9.77
3.1
2.0
1.6
2.0
2.1
8.3
Total
Sulfur
4.02
4.5
2.8
2.7
2.8
2.7
Organic
Sulfur
2.08
2.36
1.44
1.57
1.45
1.63
Pyritic
Sulfur
1.79
1.97
1.25
1.04
1.22
1.10
* Analysis of Illinois No. 6 coal used to produce Battelle pellets.
-------
to quench the combustion, and the pellets were removed from the high-
temperature zone. All pellet experiments were performed with the fur-
nace set at 1040 C. Along with thermocouple temperature, the outputs
from the CO., 0~, CO, and SO^ analysers monitoring the combustion gases
from the fixed-bed reactor were recorded continuously, A total flow
3 3
rate of 0.00012 m /sec was used in the reactor with 0.00002 m /sec of
this total added above the pellets as over-fire air. The intention was
to maintain high excess air to convert any H S released from the pellets
to SO , which would be detected by both the Faristor and TECO analyzers
to give a complete accounting of sulfur release. H_S is not detected
by the TEGO analyzer but does produce a spurious response with the Faristor.
After each experiment, each pellet was weighed and selected samples
were analyzed chemically or sectioned for metallographic examination,
scanning electron microscopy, and X-ray diffraction.
Experimental Results
Weight of the pellets is given in Table 111-16 with the chemical
analyses for selected burned pellets presented in Table 111-17. For
several cases where pellets were not completely burned, values are given
in the two tables for both the ash and the unburned core. The postheating
weights for the limestone pellets missing from Table 111-16 could not
be determined because the pellets exploded when heated and could not
be completely recovered.
Analyses of the exhaust gas from the fixed-bed reactor are
summarized in Figures III-3, III-4, and III-5. CO release from the
limestone pellets, shown in Figure III-3, was somewhat erratic, which
may be associated with the explosion of one of the pellets, shown in
Figures III-4 and III-5. Most of the SO. release occurred within the
first 5 minutes, while oxygen consumption and CO,., production continued
for significantly longer times. Some of the sulfur released from the
pellets into the gas stream could be present as H«S. Earlier experiments
with a number of pellets in the fixed-bed reactor (Runs 86-89) showed H,,S
levels ranging from about 6 to 10 percent of the SO . Use of only three
2
pellets and of high excess air was intended to convert all H_S to S0»,
but this has not been confirmed experimentally. Since the TECO monitor detects
only S02, any sulfur present as H^S would not be detected and higher sulfur
retention by the pellets would be indicated.
111-33
-------
TABLE 111-16.
SUMMARY OF FIXED-BED REACTOR EXPERIMENTS
PERFORMED IN SUPPORT OF THE MODELING STUDIES
H
H
H
I
Pellet
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
19
20
21
22
23
24
25
26
27
37
38
39
40
Pellet Pellet
Composition Batch Run No.
Limestone 169 109
n n n
n n n
110
n n n
n n n
111
n n n
n n n
" " 112
n n n
n n n
113
n n n
n n n
Coal 172 94
n ii M
M n n
95
II II II
II M II
96
n n n
n M n
70/30 173 97
n n n
n ii n
98
Burn
Time,
Min.
(a)
43
II
/ \
( B. I
3V '
"
n
18
"
n
12
n
"
6
"
"
(a)
(a)
(a)
(a)
(a)
(a)
5
"
n
20
11
n
5
Original
Pellet
Weight, g
3.2630
3.2750
3.2650
2.4990
3.3905
2.5205
2.7619
3.1688
3.0329
3.1518
3.0487
3.2448
2.7385
2.9415
2.8381
1.7275
2.2102
1.7402
1.9568
2.1032
1.7939
1.8690
1.9036
2.0847
2.5113
2.2216
2.3316
2.3100
Pellet Weight
Heating,
Total Ash
1.7405
1.6178
1.2874
1.3747
1.5483
1.6205
1.5681
1.6400
1.7543
1.5212
1.5844
0.7600
0.6475
0.5752
0.6186
0.9694
After
g
Unburned
Core
-------
TABLE 111-16. (Continued)
i
U)
Ul
Pellet
No.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
73
74
75
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
Pellet Pellet
Composition Batch Run No.
70/30 173 98
II II It
99
ii n n
n n it
" " 100
n n n
n n it
101
II II II
II II II
11 " 102
n n n
it n n
(b)
103^ '
n n ii
n n
50/50 174 104
n n n
M n n
105
M n n
n n n
" " 106
n n n
n n n
107
n n n
n n n
108
n n n
Burn
Time,
Min.
5
"
11
II
II
15
II
11
20
"
"
10
"
n
10
II
II
22^'
n
11
4
"
"
6
"
"
12
"
"
14
"
Original
Pellet
Weight, g
2.2035
2.2364
2.0811
2.2428
2.3287
2.1567
2.2779
2.0071
2.2786
1.9107
2.1246
2.5570
2.0349
2.2760
2.3062
2.2079
1.9507
2.4193
2.6733
2.5346
2.0642
2.2296
2.4874
2.7737
2.2662
2.5910
2.3511
2.4552
2.4155
2.4564
2.6856
2.7084
Pellet Weight
Heating,
Total
0.8876
0.8750
0.5505
0.6603
0.6352
0.5746
0.6243
0.5470
0.5770
0.5163
0.5877
0.8559
0.6354
0.6709
0.7422
0.6549
0.5856
0.7814
0.5788
0.8129
0.9466
0.9544
1.0345
1.0768
0.8120
0.9820
0.8388
0.8590
0.8133
0.8515
0.8950
0.8763
Ash
0.1605
0.4436
0.3799
0.3603
0.1949
0.3929
0.4507
0.3771
0.2996
0.5280
0.7774
_^
After
g
Unburned
Core
0.7145
0.1916
0.0000
0.0000
0.4760
0.2751
0.4760
0.3493
0.2042
0.2085
0.7349
0.4540
0.0359
0.0000
(a) Denotes complete combustion or reaction as determined by gas analyzers.
CM Experiment No. 103 was nerformed with the three oellets located side-bv-side in
-------
TABLE 111-17.
RESULTS FROM CHEMICAL ANALYSES OF PELLETS
FROM FIXED-BED REACTOR EXPERIMENTS
I
OJ
Chemical Composition.
Pellet
No.
1
6
9
10
15
40
42
45
48
51
52
73
57
60
63
66
69
61
_
Pellet
Composition
Limestone
"
11
ii
ii
70/30
"
"
it
"
"
"
it
70/30
"
50/50
"
50/50
50/50
50/50
11
Type of
Sample
Total pellet
"
"
II
II
II
Ash
Unburned core
Ash
Unburned core
Total pellet
"
Ash
Unburned core
Ash
Unburned core
Total pellet
Ash
Unburned core
Ash
Unburned core
Ash
Unburned core
Total pellet
Ash
Unburned core
C
0.2
0.1
0.1
1.5
0.6
35.0
13.8
46.3
1.2
39.0
0.7
0.2
8.8
46.0
6.0
46.3
0.2
8.0
27.7
5.0
30.2
0.6
26.9
0.4
9.4
27.8
Ca
60.2
60.6
60.2
56.9
59.9
22.70
32.1
20.8
42.9
24.7
34.5
35.5
36.1
20.2
35.0
20.5
45.0
37.4
31.5
34.2
16.5
34.0
37.5
42.7
32.3
C03
1.89
0.81
0.54
7.33
2.71
3.02
7.43
7.82
Fe Ash
99.3
99.7
99.5
93.5
97.5
3.23 64.20
84.0
57.1
A
66.0
98.9
98.9
97.2
60.8
99.1
61.3
98.5
91.0
72.6
97.7
68.6
99.4
98.0
2.94 90.7
2.19 69.6
H20
(a)
(a)
(a)
(a)
(a)
0.1
<0.1
<0.1
<0. 1
<0.1
<0.1
<0. 1
<0.1
0.3
<0.1
0.1
<0.1
0.1
<0.1
0.2
0.2
, 2.C/2.C.
Total
Sulfur
__
4.30
4.6
3.9
5.5
4.4
6.2
6.0
6.4
4.2
6.7
4.8
3.9
4.6
3.1
4.6
3.9
4.6
3.4
4.5
3.9
2.8
Organic Pyritic
Sulfur Sulfur
3.73 0.13
2.67 0.07
1.87 0.48
(a) Indicates that sample gained weight during analysis and value could not be determined.
-------
FIGURE III-3.
CO CONCENTRATION MEASURED IN THE GAS STREAM
FROM THE FIXED-BED REACTOR FOR THE LIMESTONE
PELLETS IN RUN 110
111-37
-------
Time, mm
FIGURE III-4.
CONCENTRATIONS OF C02, S02, 02, AND CO
MEASURED IN GAS STREAM FROM THE FIXED-
BED REACTOR FOR THE 50 WEIGHT PERCENT
COAL/50 WEIGHT PERCENT LIMESTONE
PELLETS IN RUN 104
111-38
-------
C02 0 0 » 10"* )
SOj (I 5 x I0'«)
02 (I 0 x 10"')
CO (60x IO'4)
IS
Time, mm
20
25
30
FIGURE III-5.
CONCENTRATIONS OF C02, S02, 02, AND CO
MEASURED IN GAS STREAM FROM THE FIXED-
BED REACTOR FOR THE 70 WEIGHT PERCENT
COAL/30 WEIGHT PERCENT LIMESTONE PELLETS
IN RUN 101
111-39
-------
Measurements were made to determine the time response charac-
teristics of the CO and the S0? gas analyzers. Sampling lines
from the fixed-bed reactor to the gas analyzers introduce a response delay
and the finite volume of the detector cell in the instruments along with
other factors can retard instrument response. Response measurements were
made by separately introducing S0_ and CO- into the fixed-bed reactor and
recording the output of the respective instruments as they followed the
rise and decay of the concentrations when the gas flows were switched on
and off. The response data were analyzed using the following differential
equation to approximate response time:
_
dt V F V m
m tn
where
o
C = concentration measured by analyzer, moles /cm
o
V = effective volume of sensor in analyzer, cm
m
f = flow rate through analyzer, cm /sec
2
F = flow rate through fixed-bed reactor, cm /sec
R(t) = release rate, moles/sec
t = time, sec.
For the instrument response measurements, R(t) was either a finite constant,
RO, or 0, and the above equation could be integrated to give
when the detected gas was switched on or
_L ,
R ~ V
C = e m
m F e
when the gas was switched off after having achieved steady state. Response
parameters determined for the 0- and SO analyzers are summarized in Table 111-18.
The difference in the CO^ analyzer response parameters between increasing
and decreasing transients indicates the inadequacy of the response equation.
Release rates were calculated for Runs 101 and 104 using the
differential form of the response equation
111-40
-------
TABLE 111-18. RESPONSE PARAMETERS FOR GAS ANALYZERS
M
M
H
J>
M
Analyzer
Beckman C0»
* TECO SO
* Faristor S02
Analyzer
Flow Rate/3'*
cm /sec
1.52
8.10
8.44
Delay
Time , sec
36
19.6
17
Response Parameter ,
Increasing Release
0.0764
0.0367
0.0223
f /Vm,sec
Decreasing Release
0.0501
0.0367
0.0223
(a/Flow rate based on a pressure of 29.54 in. Hg.
-------
R(t)
with the analyzer concentration curves and their derivatives, and the
response parameters from Table 111-18. The results of the calculations,
3
which are based on a flow-rate, F, of 0.0006 m /sec are presented in
Figures III-6 and III-7. In general, the response of the analyzers
followed the release rates with only a slight lag. So for the experi-
ments reported here, it was acceptable to use the concentrations from
the analyzers without response corrections.
Results from the gas analyzer measurements are presented in
Table 111-19 in terms of total 0 consumption and CO , CO and SO
release. In the table, parentheses designate values corrected for
instrument responses that show no significant differences when compared
with the results obtained by integrating the concentration curves. The
total C07 observed in Run 110 seems to be twice as high as would be
anticipated from stoichiometric considerations. The data from Table
111-19 for the coal/limestone pellets are compared in Table 111-20.
In general, about 10 percent of the observed CO released would be
expected to come from calcination of the limestone. The amount of CO
observed is small, running only about 1 to 2 percent of the CO^. Figures
III-4 and III-5 both show that the major CO release, like the S02, occurs
during the first 3 minutes after heating begins and does not continue
at an appreciable rate for longer times as does C02-
Thermocouple measurements for Run 101, which involved 70/30
pellets, are given in Figure III-8. Here it is seen that the maximum
pellet surface temperature runs about 60 C higher than the furnace
temperature because of the heat released by combustion of the pellets.
PELLET CHARACTERIZATION STUDIES
Unburned and burned pellets were sectioned metallographically
and examined microscopically by scanning electron microscopy and by
X-ray diffraction. Typical longitudinal cross sections of 50/50 pellets
are presented in Figure III-9 which shows an unburned pellet and two
111-42
-------
so, (2 o « icr6)
Time, mm
FIGURE III-6.
RELEASE RATES FOR C02 AND S02 FROM
50 WEIGHT PERCENT COAL/50 WEIGHT
PERCENT LIMESTONE PELLETS DURING
RUN 104.
111-43
-------
2100 -
1800 -
0
10 15
Time, min
20
25
3O
FIGURE III-8.
THERMOCOUPLE MEASUREMENTS FOR FIXED
BED REACTOR RUN 101 WITH 70 WEIGHT
PERCENT COAL/30 WEIGHT PERCENT
LIMESTONE PELLETS
111-45
-------
(a) Pellet 72
Unburned Pellet
(9J665)
FIGURE III-9.
METALLOGRAPHIC CROSS SECTIONS OF 50 WEIGHT PERCENT
COAL/50 WEIGHT PERCENT LIMESTONE PELLETS AS A
FUNCTION OF BURNING TIME
III-46a
-------
(9J666)
(b) . Pellet 59
4-Minute Burning Time
FIGURE II1-9. (CONTINUED)
III-46b
-------
Outer Ash
Zone
White Reaction
Zone
(c) Pellet 65
12-Minute Burning Time
FIGURE III-9. (CONTINUED)
(9J668)
III-46c
-------
M
TABLE III-19. RESULTS FROM GAS-ANALYZER MEASUREMENTS
(a)
Run
No.
110
104
101
Pellet 02 Consumed, CO Released,
Composition Moles Moles
Limestone
50
50
w/o
w/o
70 w/o
30
w/o
-
Coal/ 2.49 x lo"1 2.79 x 10~3
Limestone
-
Coal/ 3.18 x 10"1 4.25 x 10"3
Limestone
- -
co2
1.
(1.
2.
(2.
2.
(2.
Released,
Moles
67
60
27
26
17
05
x 10 1
x 10 )
x 10"1
1
x 10 )
x 10""1
..
x 10 )
so2
8
(8
I
(1
Released, ^
Moles
.02
.83
.70
.73
_
-
x 10
x 10
x 10
-4
-/i
-3
x 10 ")
(a)Values in parentheses corrected for instrument response. All other values were obtained
directly from concentration curves.
(b)Based on measurements from TECO Pulaed Fluorescence Analyzer.
-------
TABLE III-20. COMPARISON OF GAS-ANALYZER MEASUREMENTS
FOR COAL/LIMESTONE PELLETS
Run 104 Run 101
(50 w/o Coal/ (70 w/o Coal/
50 w/o Limestone) 30 w/o Limestone)
* Percent Sulfur Released 13.8 31.3
as S02
Percent Carbon in Pellets 80 84
Measured by Gas Analysis(a)
Percent C02 Occurring 1.2 2.0
as CO
* Ratio of 0 Consumed to Coal 1.0 1.3
Carbon Present in Pellets,
Moles 0~/moles Coal Carbon
(a)Run 104 is based on 0.250 moles of carbon from coal and 0.038 moles of
carbon from limestone in the three pellets. Run 101 based on 0.244
moles of carbon from coal and 0.019 moles of carbon from the limestone
in the three pellets.
111-48
-------
partially burned pellets. The unburned central zone or core is surrounded
by a white halo, which is in turn surrounded by porous ash. The unburned
central core shows evidence of this porosity, presumably from devolatiliza-
tion of the coal.
Scanning Electron Microscopy
Photomicrographs and elemental X-ray maps of a typical area in
an unburned 50/50 pellet (Pellet 72 in Figure 7) are presented in Figure
111-10. The distribution of calcium, sulfur, iron, magnesium, aluminum,
and silicon in the unburned pellets is shown in the X-ray maps. These
maps show sulfur dispersed throughout the coal but strongly associated
with the iron, probably as pyrites. Iron is also associated with magnesium,
aluminum, and silicon. Calcium is seen surrounding the iron-containing
particle and interdispersed between the coal particles.
Figures III-ll, 111-12, and 111-13 show the scanning electron
micrograph and X-ray maps for the three different zones shown in Figure
III-9 for Pellet 65, which was burned in the fixed-bed reactor for 12
minutes. The results for the outer ash zone are presented in Figure III-ll.
The 50X photomicrographs show that sulfur is associated primarily with the regions
containing calcium and small levels of magnesium. The photomicrographs
in Figure 111-12 for the white region again indicate high porosity, with
the sulfur dispersed throughout the material that contains high levels of
calcium.
Results for the unburned inner core, depicted in Figure 111-13,
indicate that some of the coal is still intact. However, there is evidence
of porosity formation, which is probably associated with partial devolati-
zation of part of the coal. There is also evidence, as can be seen by
comparing the X-ray maps for calcium, sulfur, and iron, that sulfur may
pass from the pyrites into the surrounding calcium by a solid-state reaction
mechanism.
Further information on the chemical structure of the constituents
in the pellets was obtained by X-ray diffraction.
111-49
-------
t-i
Ul
o
03
SOX
Photomicrograph
(a)
500X
Pho tomlcrograph
FIGURE 111-10.
SCANNING ELECTRON MICROGRAPHS AND ELEMENTAL X-RAY MAPS FOR
UNBURNED 50 WEIGHT PERCENT COAL/50 WEIGHT PERCENT LIMESTONE
PELLET (PELLET 72)
-------
M
t_n
O
C7-
500X
(b)
Calcium X-Ray Map 500X Sulfur X-Ray Map
FIGURE 111-10. (CONTINUED)
-------
H
M
M
Ln
O
n
500X
(c)
Iron X-Ray Map 500X Magnesium X-Ray Map
FIGURE III-10. (CONTINUED)
-------
M
Ui
O
(d)
500X
Aluminum X-Ray Map
500X
Silicon X-Ray Map
FIGURE III-10. (CONTINUED)
-------
i
Ln
50X
(a)
Photomicrograph
500X
Photomicrograph
FIGURE III-ll. SCANNING ELECTRON MICROGRAPHS AND ELEMENTAL X-RAY MAPS FOR OUTER ASH ZONE
IN 50 WEIGHT PERCENT COAL/50 WEIGHT PERCENT LIMESTONE PELLET (PELLET 65)
BURNED FOR 12 MINUTES IN THE FIXED-BED REACTOR
-------
H
V
500X
Calcium X-Ray Map
FIGURE III-11.
500X
(CONTINUED)
Sulfur X-Ray Map
-------
I
M
O
500X
(c)
Iron X-Ray Map 500X Magnesium X-Ray Map
FIGURE III-ll. (CONTINUED)
-------
Ul
(d)
500X Aluminum X-Ray Map 500X Silicon X-Ray Map
FIGURE III-ll. (CONTINUED)
-------
I
Ul
(a)
500X
Chlorine X-Ray Map
500X
Photomicrograph
FIGURE 111-12.
SCANNING ELECTRON MICROGRAPHS AND ELEMENTAL X-RAY MAPS FOR
WHITE REACTION ZONE IN 50 WEIGHT PERCENT COAL/50 WEIGHT
PERCENT LIMESTONE PELLET (PELLET 65) BURNED FOR 12 MINUTES
IN THE FIXED-BED REACTOR
-------
I
Ul
*H1K-+ff?. »?"-
; '.^^?$m%
*-JL«*SI «JTJV , >,..A3fhy.
£gp?|^
-------
I
t_n
500X
(c)
Iron X-Ray Map 500X Magnesium X-Ray Map
FIGURE 111-12. (CONTINUED)
-------
1
Ul
P.
(d)
500X Aluminum X-Ray Map 500X Silicon X-Ray Map
FIGURE 111-12. (CONTINUED)
-------
FIGURE 111-13.
SCANNING ELECTRON MICROGRAPHS AND ELEMENTAL X-RAY MAPS FOR INNER CORE IN
50 WEIGHT PERCENT COAL/50 WEIGHT PERCENT LIMESTONE PELLET (PELLET 65)
BURNED FOR 12 MINUTES IN THE FIXED-BED REACTOR
-------
M
01
500X
(b)
Calcium X-Ray Map 500X Sulfur X-Ray Map
FIGURE 111-13. (CONTINUED)
-------
Ut
u>
n
500X
(c)
Iron X-Ray Map 500X Magnesium X-Ray Map
FIGURE III-13. (CONTINUED)
-------
H
un
u>
500X
(d)
Aluminum X-Ray Map 500X Silicon X-Ray Map
FIGURE 111-13. (CONTINUED)
-------
X-Ray Diffraction
The X-ray diffraction measurements are summarized in Figures III-14
through 111-17 for a coal pellet (Pellet 36), a limestone pellet (Pellet 17),
an unburned 70/30 pellet (Pellet 78), and a 70/30 pellet (Pellet 37) that
was burned for 20 minutes in the fixed-bed reactor. A peak analysis was
performed only for Pellet 37. The identified compounds in the burned
pellet were CaO, CaSO, (anhydrite), Ca(OH) , MgO, and CaFe02, indicating
that the predominant sulfur containing compound is CaSO,.
Further use of x-ray diffraction techniques should be quite
valuable for studying the kinetics of the solid-state reactions in the
coal/limestone pellets, although the relative insensitivity of x-ray
diffraction will be troublesome.
Description of Mathematical Model
The pellet burning and sulfur capture model is based on a spherical
approximation represented by the three coupled partial differential equations
given below. The 0? concentration distribution in the spherical pellet is
given by
3C1
- k^C3C4/C^ = e (1)
with the SO concentration given by
\
3C
3
and the CaO concentration distribution by
3C
TT = - ^sv75!
where
2
C1 = 0« concentration, moles/cm
3
C,. = S09 concentration, moles/cm
3
C, = CaO Concentration, moles/cm
D = Effective diffusion coefficient for Q~ in pore
structure, cm^
D- = Effective diffusion coefficient for S02 in pore
structure,
111-54
-------
liO 100 90 80 70 60 50 40 30 20
FIGURE 111-14. Cu K ,,_RADIATION SCATTER OF UNBURNED COAL
110 100 90
FUGIRE 111-15.
80 70 60 50 40 30
Cu IU. RADIATION SCATTER OF UNHEATED
LIMESTONE
20
no 100 90
FIGURE 111-16.
i
80
70
i ' i ' ; ' ] '
60 50 40 30
Cu KA RADIATION SCATTER OF UNBURNED
70/30 MIXTURE
20
o
o
MgO
t
V
o>
c
o
Ca(OH)2-j
M9°7 CcO I !
Ca(OH)2;
t
IIC 100
FIGURE 111-17.
90 80 70 60 50 40 30
Cu K-A,RADIATION SCATTER OF 70/30 COAL/
LIMESTONE BURNED PELLET. MAJOR IDENTIFIED
CONSTITUENTS CaO (LIME), CaSO, (ANHYDRITE)
111-55
-------
9/2 / 3/2
k' = S09 capture rate constant, cm /mole -sec
s ^
e = pellet porosity.
The mechanism for the SO^-CaO reaction is based on the work of Wen and
Assciates. ' ~ ' The initial conditions assumed are
1. C-j^Cr.o) = C° r=R
= o o< r
-------
The model contains several assumptions that may limit its
accuracy. These assumptions are:
Heat-transfer effects and internal heat generation
are negligible
The rather simple chemical mechanisms assumed in
the model are adequate
The calcination reaction is instananeous
Use of a spherical geometry is adequate to simulate
the pellets which are cylindrical in shape.
Of the four major assumptions, the one neglecting heat transfer is probably
the most critical. All these assumptions can be refined, however, by further
extensions of the present model.
To facilitate the preliminary analysis of data from the fixed-bed
reactor experiments, a simplified version of the above model was derived,
The simplified model is based on the assumptions that:
The steady-state approximation is valid
The change in the CaO content due to sulfur capture
is negligible
The oxygen dependency of the SO- capture reaction can
be represented by an average constant 0? concentration.
The burning rate constant, k, is expressed as
k, cm/ sec = etk
s
S02 capture and mass transport in the pellet are represented by
+ .
dr2 r dr * D3 C3
with the boundary conditions
C_ = C3 at r = R
and
dC,
D3
where
K = S02 capture rate constant, sec
(3 = stoichiometric coefficient,
°2
111-57
-------
CL consumption is represented by
d2C,
dr
JL^l
r dr
= 0
r
-------
with
C1D1
62 =-
ryTTiT
Cc / rc R J
kr
Lv
Co = C° at r = R for t^o.
A program based on the above equations was written for a Hewlett-
Packard 9825A calculator for use in analyzing data from the fixed-bed reactor
experiments. Details on the application of the simplified model to the
experiments are covered in the following section.
MODELING ANALYSIS OF PELLET
COMBUSTION DATA
Results of the chemical analyses on the burned pellets from
the fixed-bed reactor experiments are summarized in Tables 111-21 and 111-22.
These results are also presented in Figures 111-18 and 111-19 for better
visualization of the data. Data scatter is quite high and as a result the
curves in these figures were drawn subjectively. The burning rate for
the 50/50 pellets appears to be faster than for the 70/30 pellets, which
is consistent with the gas-analyzer data from Runs 101 and 104. However,
the chemical-analysis data indicate that sulfur retention is higher with
the 70/30 pellets, which would not be expected since the 50/50 pellets
contain more limestone. Nor is this in agreement with the gas-analyzer
data, which showed the 50/50 pellets retain more sulfur than the 70/30
pellets (see Table 111-20). The reason for this inconsistency is not
understood at this time.
111-59
-------
TABLE III-21. SUMMARY OF RESULTS OF THE CHEMICAL ANALYSES
ON 50 W/0 COAL/50 W/0 LIMESTONE PELLETS
FROM FIXED-BED REACTOR EXPERIMENTS
Burn
Time,
Min.
0
4
6
6
12
14
22
Pellet
No.
81
60
63
61
66
69
57
Pellet
Weight
Loss, %
0
58.4
62.1
61.2
66.3
67.7
67.9
Sulfur (a)
Retained ,
%
100
63.5
71.1
67.1
64.3
55.5
Carbon /a\
Burned ,
%
0
72.9
81.0
98.2
99.6
99.8
(a) Based on the use of the ash content as a constant
reference for calculations.
111-60
-------
TABLE 111-22. SUMMARY OF RESULTS OF THE CHEMICAL ANALYSES
ON 70 W/0 COAL/30 W/0 LIMESTONE PELLETS
FROM FIXED-BED REACTOR EXPERIMENTS
Burn
Time,
Min.
0
5
5
10
10
11
15
20
Pellet
No.
80
42
40
52
73
45
48
51
Pellet
Weight
Loss, %
0
60.9
58.0
66.5
67.8
72.7
72.7
72.3
Sulfur
Retained,
%
100
67.3
69.4
69.7
74.0
64.9
62.8
Carbon , -,
Burned,
%
0
59.4
65.9
76.1
80.8
99.6
99.9
(a) Based on the use of the ash content as a constant
reference for calculations.
111-61
-------
100
90
80
70
- 60
c
03
O
o5 en
Q_ OU
40
30
20
10
Carbon Burned
Sulfur Retention
10 15
Heating Time, min
20
25
FIGURE 111-18.
SUMMARY OF RESULTS OF THE CHEMICAL
ANALYSES ON 50 WEIGHT PERCENT COAL/
50 WEIGHT PERCENT LIMESTONE PELLETS
FROM FIXED BED REACTOR EXPERIMENTS
(Data from Table 8)
111-62
-------
10 15 ZO
Heating Timei min
25
30
FIGURE 111-19,
SUMMARY OF RESULTS OF THE CHEMICAL
ANALYSES ON 70 WEIGHT PERCENT COAL/
30 WEIGHT PERCENT LIMESTONE PELLETS
FROM FIXED BED REACTOR EXPERIMENTS
'(Data from Table 9)
111-63
-------
The chemical analysis data and dimensional measurements for
the limestone pellets heated in the fixed-bed reactor are given in Table
111-23. The data show a decrease in pellet diameter with heating and
indicate that calcination is nearly complete within 6 minutes. This is
not consistent with the gas-analyzer data, which show significant
release of C09 after 6 minutes. This difference and the fact that the
analyzer data gave a higher integrated CO- release value are also
unexplained at this time.
Determination of Rate Constants
Burning rate data from the metallographic measurements on 50/50
pellets are presented in Table 111-24. The 70/30 pellets disintegrated
to the extent that it was not possible to obtain similar data. The data
from Table 111-24 were correlated with the burning-rate expression given
by Equation (5) for the simplified model. A calculated burning velocity
curve based on the derived constants is shown in Figure 111-20 along
with a curve calculated for the 70/30 pellet based on a burning time of
20 minutes. The calculations are based on an effective spherical pellet
radius, R, equal to 0.635 cm, the radius of a sphere having the
same surface-to-volume ratio as a right circular cylinder with a diameter
and height of 12.7 mm. Temperature was assumed to be 1040 C. Assuming
appropriate physical constants, the values of the burning rate constant
and the diffusion coefficient for oxygen diffusion in the ash were
calculated. Using the same burning-rate constant, the analysis was also
repeated to determine the oxygen diffusion coefficient for the 70/30
pellet. The results were then incorporated into the simplified SC>2
release equation, Equation (6), with the value of K, the S07 capture
rate constant, selected to give the levels of sulfur capture determined
by the gas analyzer results. The values of the constants used in the
calculations and the results are given in Table 111-25. It was
impossible to obtain a valid SO capture vrate for the 70/30 pellet that
corresponded to 68.7 percent sulfur retention by the pellet. For a
range of reasonable variables, the model predicts a sulfur capture in
excess of 80 percent, just under that calculated for the 50/50 pellet.
The model is extremely sensitive to the values of the constants, and
small uncertainties in the data can produce large variations in the
111-64
-------
10 IS 20
Heating Time, mm
25
30
FIGURE III-20.
UNBURNED CORE RADIUS AS A FUNCTION
OF BURNING TIME CALCULATED USING
SIMPLIFIED MODEL (Data Points
Represent Measurements on 50/50
Pellets)
111-65
-------
TABLE 111-23. SUMMARY OF CHEMICAL ANALYSES AND
PHYSICAL MEASUREMENTS ON THE
LIMESTONE PELLETS FROM THE FIXED-
BED REACTOR EXPERIMENTS
Pellet
No.
17
14
11
7
4
Heating
Time, Min.
0
6
12
18
31
Pellet Diameter
After Heating,
In.
0.506
0.501
0.494
0.498
0.496
Weight
Change^
%
0
48.3
46.2
43.9
47.4
Total C02 Released,
%
0
95.7
88.4
99.1
98.7
(a) The weight change for complete calcination of
CaCO_ would be 44 percent.
111-66
-------
TABLE III-24. SUMMARY OF BURNING RATE MEASUREMENTS
FOR 50 W/0 COAL/50 W/0 LIMESTONE PELLETS
BASED ON METALLOGRAPHIC STUDIES
Pellet
No.
72
59
62
65
68
56
Burning
Time,
Min.
0
4
6
12
14
22
Relative Burning
Interface Radius,
rc/R
1
0.75
0.64
0.44
0.25
0
111-67
-------
TABLE 111-25. SUMMARY OF CALCULATIONS USING SIMPLIFIED MODEL
M
M
M
1
Oo
Constants
R,cm
Temperature, °F
C° moles 02/cm3
C", moles S02/cm3
<^, moles 02/cm3
B
e, (ash porosity)
Derived Rate Coefficients
K, cm/sec
K, sec
2 ,
D^, cm /sec
T>2> cm /sec
50 w/o Coal/50 w/o Limestone
Pellet
0.635
1900
1.84 x 10 6
° -2
6.11 x 10
2.07 x 10
0.66
3.5 x 102
2.6 x 10 (a)
1.77
1.17
70 w/o Coal/30 w/o Limestone
Pellet
0.635
1900
1.84 x 10
o
6.74 x 10 ;
1.83 x 10
0.73
3.5 x 102
(c)
2.17
1.43
(b)
(a) Based on a sulfur retention of 86.2%.
(b) Assumed value.
(c) Would not correlate with the measured sulfur retention of 68.7%.
-------
results. The relation of diffusion rates to porosity was determined
using the effective diffusion constants for the 50/50 and 70/30 pellets
and the relationship
where D is the effective diffusion coefficient, e is the porosity, and D
is the true diffusion coefficient. For oxygen diffusion, n was determined
2
as 2.28 with the true diffusion coefficient for oxygen equal to 4.5 cm /sec.
The value for n corresponds to those observed by Wen and Chuang
for coal ash. The true value of the diffusion coefficient for 0 seems
somewhat high but may indicate higher temperatures in the pellets,
possibly nearer 1470 C. The value of K also matches very closely with
values reported by Field, et al. The value of K for the 50/50 case
is considerably smaller than would be anticipated from the results of Wen,
- (III-ll, 111-12) .,_-,. ,. , ,-,
et al with pure limestone. This may not be unreasonable,
though because of the presence of the coal ash and other impurities in
combination with the limestone in the pellets.
Using the derived data for the 50/50 pellet in Equation (6) ,
the SO release rate and the concentration of S02 as it would be
measured by the TECO analyzer was calculated. These results (Figure 111-21)
show the same characteristics as the actual experimental curves, however,
experimental response times initially seem to lag behind the calculated
response. This may be due to heat-transfer effects, which are discussed
below. Oxygen consumption calculated from Equation (5) for the 50/50
and 70/30 pellet compositions is presented in Figure 111-22.
To illustrate the predictive capabilities of the simplified
model, calculations shown in Figures 111-23 and 111-24 were made to
estimate sulfur retention and burning time as a function of pellet
diameter, and burning time as a function of temperature for a 50/50
pellet. The burning rate constant used for the calculations in Figure
111-24 was
17,967
cm/sec
(7)
C?A T e cm/sec
where the temperature dependency was based on the work of Field, et al.
In the above equation, A was determined to be 1.3 x 10~ using the rate constant
111-69
-------
SOj Concentration
S02 Release Rale
(7 5 « I
20 40 60 80 IOO 120 140 160 ISO 200 220 240 26O 280 300
FIGURE 111-21.
S02 RELEASE RATE AND CALCULATED TECO
ANALYZERS BASED ON THE SIMPLIFIED
MODEL (EQUATION 7) FOR 50 WEIGHT
PERCENT COAL/50 WEIGHT PERCENT
LIMESTONE PELLET
111-70
-------
10
&
c.
_o
"5.
3
70/30 Pel let
IO'7
j_
50/50 Pellet
10 15 20
Heating Time, min
25
30
FIGURE III-22. OXYGEN CONSUMPTION RATES CALCULATED
USING SIMPLIFIED MOEDL
111-71
-------
98
96
94
0>
-I 92
I
en
190
o
88
86
84
I
I
525
450
375
c
'£
300 of
c
'c
225?
D
,0
150
0.5 1.00 (.50
Pellet Diameter, in
2.00
2.50
FIGURE 111-23.
SULFUR RETENTION AND BURNING
TIME FOR 50 WEIGHT PERCENT
COAL/50 WEIGHT PERCENT LIMESTONE
PELLET CALCULATED AS A FUNCTION
OF PELLET DIAMETER USING THE
SIMPLIFIED MODEL
111-72
-------
25
24
c
'£
'c
2!
20
19
1700
1800
1900
2000
Temperature, F
2100
2200
2300
FIGURE 111-24.
BURNING TIME FOR 50 WEIGHT PERCENT COAL/
50 WEIGHT PERCENT LIMESTONE PELLET
CALCULATED AS A FUNCTION OF TEMPERATURE
USING THE SIMPLIFIED MODEL
111-73
-------
determined for the 50/50 pellet assuming a temperature of 1040 C. T is
the temperature in °K and C° is the oxygen concentration, as previously
defined. The temperature dependency of the burning rate constant is
very approximate and should be accurately determined through more detailed
experiments than those reported here, if definitive conclusions are to
be drawn about the performance characteristics of the pellets. An
interesting conclusion from Figure 111-23 is that there is an optimum
diameter for sulfur retention. Sulfur retention reaches a maximum of
about 96 percent for a pellet diameter of about 38 mm. A more detailed
model or actual experiments may change this conclusion, but the effect
of size and geometry probably merits further attention.
Transient Heat-Transfer Analysis
A simple transient heat-transfer analysis based on a treatment
by Heisler was performed to assess the effects of thermal lag on
pellet combustion and S02 release. The results of the heat-transfer
analysis are presented in Figures 111-25 and 111-26 for the 50/50 and
70/30 pellets, respectively. The calculations assumed a spherical
pellet with a radius of 0.635 cm at an initial temperature of 21 C which
is instantaneously introduced into a furnace at 1040 C. The surface
temperatures of the two pellets respond at the same rate, with the
center temperature of the 70/30 pellet lagging slightly behind that
of the 50/50 pellet because of its lower thermal conductivity. The
surface temperature of both compositions reached the furnace temperature
in just over 1 minute, while the centers of the pellets required about
2-1/2 minutes to attain the furnace temperature of 1040 C. Heat generated
by pellet combustion could increase the temperatures more rapidly
and significantly increase the pellet temperatures over the furnace
temperature while endothermic reactions, such as calcination of the lime-
stone and devoletilization of the coal, will delay the temperature rise.
A more detailed heat-transfer analysis would be required to examine
these effects. The simplified analysis does indicate that initial
heating effects could play an important role in pellet burning and SO,.,
release. These effects are probably responsible for the deviation between
the model prediction of S0~ and the experimental gas-analyzer data. One
of the next refinements in the model should be to include heat-transfer
effects.
111-74
-------
2000
ISOO -
1900 F
^Center of
/ ^Pellet
/
f 1000
w
01
Q.
1
5OU
70
n
f /
/ Pellet Parameters
/ Density: 1.50 g/cms
Thermal Conductivity: 0.23 hr-ft-F
/ B+u
i neat Capacity- O.29O ^K _c
! I
Time, min
FIGURE 111-25
TRANSIENT HEAT TRANSFER CALCULATIONS FOR
A 70 WEIGHT PERCENT COAL/30 WEIGHT PERCENT
LIMESTONE 13.7 MM DIAMETER SPHERICAL PELLET
111-75
-------
2000r
1500 -
Li.
o.
4)
1000 -
1900° F
500
70
Pellet Parameters
Density'. 1.41 g/cm'
Therman Conductivity10.18
Heat Capacity:0.275 JtiL.
lbn
Time, min
FIGURE 111-26.
TRANSIENT HEAT TRANSFER CALCULATIONS
A 50 WEIGHT PERCENT COAL/50 WEIGHT
PERCENT 13.7 MM DIAMETER LIMESTONE
SPHERICAL PELLET
FOR
111-76
-------
Detailed Modeling Calculations
For economic reasons, computer calculations using the more
detailed model were limited. Improvements still need to be made in the
numerical analysis to decrease the computational time. An example of a
calculation using the program is shown in Figures 111-27 through 111-32
for a 70/30 pellet at 1040 C. The following constants were employed
in the calculations:
k - 3.5 x 104 -r
^ _»-/ A J-U / 1 \
s /moles \
\ cm /
9/2
k1 = 4.5 x 106 °m
s
sec
,
moles *sec
C£ = 4.2 x 10"3 moles Ca/cm3
C° = 1.84 x 10~6 moles 02/cm3
C° = 1.2 x 10"2 moles S/cm3
-1 3
C2 = 1.0 x 10 moles coal/cm
f3 = 1.19 x 10~2
2
DI = 2.17 cm /sec
2
D_ = 1.43 cm /sec
R = 0.635 cm
e = 1
The curves for both SO- and 0_ appear to correspond roughly with the
experimental measurements in Figures III-5 and III-7. Longer time
calculations are needed, however, to accurately determine the validity
of the model. The results obtained to date do indicate that extension
of the model to include heat transfer and the effects of coal devolati-
lization would be appropriate.
111-77
-------
in
w
o
o
o
o
0-000 5.850 11.700 17.550 23.400
Time of Combustion (sec)
29.250
35.100
FIGURE 111-27. CALCULATED COMBUSTION RATE FOR 70
WEIGHT PERCENT COAL/30 WEIGHT
PERCENT LIMESTONE PELLET
111-78
-------
CM
(O
O
X S
o
LU
CO ^_
CO o
LU
O
z:
10
(K S
LU
CO
g
CC
CO
CO
o °
LU
I
<
°^ oi
o
o
o
o
o
0.000
5-850
1
11.700
17.550
23.400
29.250
35.100
Time of Combustion (sec)
FIGURE 111-28. CALCULATED S02 RELEASE RATE FOR 70 WEIGHT
PERCENT COAL/30 WEIGHT PERCENT LIMESTONE
PELLET
111-79
-------
o
U3
O
CO
UJ
in
o
CM
I
~ IT)
X O
o
UJ
CO
N.
n
o
o
a: S
LU
u_
CO
g
o;
CO
CO
\f>
fe 5
UJ
i
<
o
o
o
o
o
0.000
T
5.850
11.700
I
17.550
23.400
29.250
1
35.100
Time of Consumption (sec)
FIGURE 111-29. CALCULATED Oo CONSUMPTION RATE FOR 70 WEIGHT
PERCENT COAL/30 WEIGHT PERCENT LIMESTONE PELLET
111-80
-------
f\J
*
o
t
O
ro
z:
o
\
CO
LU
n
o
cj
n
o
to
(M
o
. to
LU
O _
o
o
o
o
o
o
606 .611 .616 .620 .625
Radius of the Pellet (CM)
.630
.635
FIGURE III-30.
CALCULATED CaO DISTRIBUTION AT 35.10 SEC
IN 70 WEIGHT PERCENT COAL/30 WEIGHT
PERCENT LIMESTONE PELLET
111-81
-------
IT)
O)
O
LU
o
z
o
o
CO
00
O
I
O
X
O
\
CO
LU
o
11 <0
ce
-------
CONCENTRATION (MOLES/CM3) (XI 0 -4)
o.ooo
.002
.005
.007
.009
.012
.014
016
018
oo
CO
^
M
O
I
CO
N>
IIP
n o
M
a
o
S3
Hi O
H
CO
M
n
CL
H-
c
-J c!
or-1 o
> HI
MM rt
s o
W ffi O
CO HM
H
o TI e»
!Z! M M
M fa co
O H
M a i<
t-< a
W O H
H O M
Jt- O
ft)
rt
n
o>
o
0>
en
O)
O)
M
o
o>
ro
in
en
u
O
o>
C4
cn
-------
MODELING AND CHARACTERIZATION
SUMMARY AND CONCLUSIONS
The results of the preliminary modeling analysis and the pellet
characterization studies described suggest that further study and optimi-
zation of coal/limestone pellets for sulfur capture offers potential
benefits.
The simplified model developed for correlation of the experimental
data approximates some of the performance characteristics of the pellets.
For the range of calcium/sulfur ratios considered, the model indicates a
weak dependency on this ratio, which is in line with observations. Also,
the model predicts an optimum pellet size for maximum sulfur retention.
This possibility should be given further attention.
Scanning electron microscopy and X-ray diffraction analysis
are powerful tools for the study of the solid-state reactions
in pellets. Results show that the captured sulfur in the burned
pellets is tied up predominantly as CaSO,. There is evidence that the
sulfur may react directly with the limestone by solid-state processes that
do not involve formation of SO.
Recomendations for further studies are:
The present model should be extended to include heat-
transfer effects, calcination, coal devolatilization,
and solid-state reactions.
Detailed experimental measurements should be made
to determine SO^-capture and burning rate constants
as a function of temperature.
Further information on the pore and ash structure
of the pellets and its influence on the mass trans-
port processes should be identified.
* The pellet burning model should be coupled with
models of different boiler combustion systems to
predict the interaction between pellet parameters
and systems operating conditions.
111-84
-------
SECTION III-5
LABORATORY EVALUATION
The promising pellet formulations identified during the mech-
anical strength and fixed-bed reactor experiments were evaluated in the
model spreader-stoker boiler. This evaluation was based on gaseous
emissions (primarily SO-) and visual observations of the fuel bed. In
addition, 18 Mg of the most promising pellet formulation were fired in
the Battelle steamplant stoker. Criteria pollutants, visual observations,
and ash analyses were used in these evaluations. Both facilities have
been described in Parts I and II of this report.
MODEL SPREADER EXPERIMENTS
To supplement the fixed-bed reactor experiments, the model
spreader-stoker boiler was used to evaluate the more promising pellet
formulations. Compared to the fixed-bed reactor, the model spreader pro-
vides an improved simulation of the operation of an industrial stoker
boiler and thus evaluates the fuel pellet more realistically.
Table 111-26 presents the results of these experiments. In these
experiments, the effect of Ca/S ratio (3.5 and 7), the four pellet pro-
duction techniques, and binder type (cement and methylcellulose) were
investigated. Additionally, for comparison, experiments were conducted
with medium-S Kentucky coal, Illinois No. 6 coal, and the 50/50 pellets
produced during Phase II. Prior to experimentation, the sampling system
and procedures were modified to minimize any reactions that may occur in
the sampling system.
111-85
-------
TABLE 111-26. MODEL-SPREADER STUDIES
00
Run No.
MS-20(b)
MS- 31
MS-32
MS- 33
MS- 34
MS- 35
MS- 36
MS- 3 7
MS- 38
MS- 39
MS-40
MS- 41
MS- 4 2
Fuel
100/50 CPM pellets
(Latex)
50/50 CPM pellets
(cement)
50/50 CPM pellets
(cement)
70/30 CPM pellets
(cement)
Medium S Kentucky
Illinois #6 Coal
50/50 CPM pellets
(cement)
Illinois #6 Coal
70/30 CPM pellets
(methocel)
70/30 briquets
(methocel)
70/30 disc pellets
(methocel)
70/30 extrusion
(methocel)
70/30 CPM pellets
(methocel)
Fuel
Size,
mm
12.5 x 19
Ditto
Ditto
Ditto
~
12.5 x 19
12.5 x 19
12.5 x 25
12.5 dia
12.5 x 19
Ca/S Ratio
(Approx.)
4
7
7
3.5
0
0
7
0
3.5
3.5
3.5
3.5
3.5
Average
S02,(a)
ppm
1370
1100
1040
1220
1050
4120
1240
3700
1480
1780
1370
1260 , .
(1220) (C)
Predicted
S02,
ppm
3840
4020
3700
3700
900
3700
3700
3700
3700
3700
3700
3700
Average
Stack
Temp, C
295
340
350
360
340
300
325
335
365
375
345
Average
Excess
Air,
percent
100
100
140
110
120
95
120
100
90
80
60
75
Average
Sulfur
Retention,
percent
64
73
72
67
-15
-11
67
0
60
52
63
67
percent
10.7
11.0
11.5 - 15.0
10.0 - 13.8
9.5 - 19
8.0 - 12.7
13.5
10.6
9.5 x 11.5
8.2 - 9.8
7.4 - 9.2
8.4 -10.2
C02,
percent
9.7
10
5.9 - 8.0
NA
8.2-10
8.4 - 10.8
8.5
9.2
8.0 - 11.2
10.8 - 11.6
10.8 - 12.6
10.4 - 12
CO,
ppm
42
55
150
100
300
100
150
85
90
50
(a) Normalized to 3 percent 02.
(b) 1978 data.
(c) By Method 6.
-------
Sampling System
The sampling system was a modification of that used during
the Phase II experiments. An in-stack filter was used upstream of the
water trap and the water trap was coupled as close to the stack as
physically possible. Figure 111-33 is a schematic of the modified
sampling system.
Modifications were made to minimize the presence of water
(especially water with calcium-laden particulates) in the sampling system
that could remove SO-. Provisions were also made to span the instruments
by injecting calibration gases through the entire sampling system before,
during, and after the experiments. This procedure indicated no loss of
S0~ in the sampling train at any time during the experiment.
A comparison of the model-spreader pellet data from the Phase
III experiments with those from Phase I indicate that sulfur capture was
not as great (about 10 to 15 percent lower) for the Phase III experiments.
This small reduction may be attributed to the improved sampling system
where precautions were taken to minimize sulfur capture in the sampling
line. Because of the difficulty in obtaining a representative sample, sulfur
retention based on SCU emission levels could not be verified from a
sulfur analysis of the bed ash.
Ca/S Ratio
The fixed-bed reactor experiments indicated that the Ca/S ratio
had little or no effect on sulfur capture for Ca/S ratios greater than
3.5. The model spreader data presented in Table 111-26 confirm this
observation. Visual observations indicated, as expected, that the pellets
with less limestone (Ca/S = 3.5) burned more uniformly and rapidly
than those with more limestone (Ca/S = 7).
Production Technique
Pellets using the same formulation consisting of Illinois No. 6
coal, limestone (Ca/S = 3.5), and methylcellulose binder, were prepared
by the following production techniques:
111-87
-------
To Analyzer
Calibration
Gases
A
Flow
Meter
Manifold
n
Quartz wool
Lined U type
Particulate
Filter
[_Heated (I25-I50C)
\
Flow Valve
S S Bellows
Sample Pump
Cold Finger
Water Trap
FIGURE II1-33, MODEL-SPREADER SAMPLING SYSTEM
-------
Pellet mill (prepared by Battelle staff)
Auger extrusion (prepared by Banner Industries)
Disc agglomeration (prepared by Mars Mineral Corporation)
Briquets (prepared by Evergreen Company).
The pellet-mill and auger-extruded pellets burned satisfactorily,
having sulfur captures of 67 and 63 percent, respectively. The auger
extruded pellets were observed to burn more uniformly than the pellet mill
pellets perhaps because they were more porous (-1.0 g/cc compared to ~1.9
g/cc).
The briquetted formulations burned satisfactorily and showed relatively
low sulfur retention (52 percent) a surprising and unexplained result. The
disc agglomerated pellets were entirely unsatisfactory when fired in the
model spreader. These pellets disintegrated in the combustion zone producing
excessive amounts (greater than 50 percent) of fines. Such fines matted the
bed causing nonuniform air distribution. Fuel-bed conditions degraded so
rapidly that meaningful data could not be obtained.
Binder Type
Comparison of sulfur retention data of the auger-extruded and mill-
pellets made with organic (methylcellulose) and inorganic (cement) binders
indicated no significant difference. The binders are used in very small
quantities (less than 4 percent) and do not have any catalytic effects. As
a result, it appears that the type of binder does not significantly effect
the combustion behavior of the pellet providing the physical properties of
the pellet are retained. Cement-bound pellets with satisfactory physical
properties could not be made by the disc agglomeration and briquetting methods.
STEAMPLANT STOKER DEMONSTRATION
Eighteen Mg of the limestone coal fuel pellets with a Ca/S
molal ratio of approximately 3.5 were fired in the steamplant boiler.
Two types of pellets were useda lower density of (0.9 to 1.2 g/cc)
111-89
-------
pellet produced by Banner Industries using auger extrusion and a higher
density pellet ("1.4 g/cc) produced by Alley-Cassetty Coal Company
using a pellet mill. Both types of pellets were fired under a variety
of boiler conditions. Evaluations were based on visual observations,
criteria pollutants, and ash analyses.
PELLET PREPARATION AND PROPERTIES
Throughout Phase III, efforts were made to identify companies
that could provide tonnage quantities of pellets to our specifications at
no more than $1100/Mg, excluding grinding and raw material costs. Table
111-27 identifies candidate tolling firms. Of these, only Banner Industries
and Alley-Cassetty Coal Company could meet the criteria. Neither had on-
sight grinding capabilities. The Illinois No. 6 coal was pulverized by
Battelle and shipped with the necessary Piqua limestone and binders to
the respective preparation sites.
Allbond 200 cornstarch and M-167 latex emulsion were used as
binders, although this combination was not evaluated in the model spreader.
These binders were readily available, while the methylcellulose binder
previously used could not be procured in sufficient quantities in time for
the demonstration. In addition, the mechanical-strength characterization
tests and fixed-bed reactor experiments indicated that the Allbond and
methylcellulose bound pellets were comparable.
The resulting pellet formulation (dry basis) consisted of:
67 percent Illinois No. 6 coal
30 percent Piqua limestone
2 percent Allbond 200 binder
1 percent M-167 latex binder.
These pellets were cylindrical, about 13 mm in diameter and 25- to 75-tnm
long. Table 111-28 gives the proximate and ultimate analyses and also
the ash-fusion temperature (initial deformation) for these pellets.
Table 111-29 gives the mineral analysis of the ash.
111-90
-------
TABLE 111-27. POTENTIAL TOLLING COMPANIES FOR
LARGE QUANTITIES OF PELLETS
-------
M
I
VO
TABLE 111-28. ULTIMATE, PROXIMATE, AND ASH-FUSION TEMPERATURE ANALYSES FOR
LIMESTONE/COAL FUEL PELLET Ca/S =3.5
Proximate Analysis Ash-Fusion
(As received). % Heating Temperature. C
Fixed 0 Value, Initial Deformation
Volatiles Carbon Moisture Ash C HNS (difference) KJ/g Reducing Oxidizing
42.9 21.8 2.15 33.2 47.0 3.2 .9 2.9 10.5 18.6 1500+ 1500+
-------
TABLE 111-29. MINERAL ANALYSIS OF ASH
Compound
Silica, Si02
Alumina, Al^O*
Titania, Ti02
Ferric oxide, Fe20-
Lime, CaO
Magnesia, MgO
Potassium oxide, K20
Sodium oxide, NaJ?
Sulfur trioxide, SO-
Phos. pentoxide, P205
Strontium oxide, SrO
Barium oxide, BaO
Manganese oxide, M^O*
Undetermined
Percent Weight
14.20
5.42
0.24
7.10
51.88
6.94
0.55
0.42
10.77
0.09
0.00
0.01
0.06
2.32
100.00
111-93
-------
The high CaO in the ash of the treated coal preclude
the usual procedures for evaluating ash characteristics, which are
limited to-about 20 percent CaO. The excess CaO above that which
will react with the other ash constituents, principally with SiO-, to
form a low-viscosity slag, provides a matrix of solid CaO particles. Thus,
in the ASTM cone fusion determination, this matrix retains the original
shape of the cone, probably even at temperatures above 1650 C, which
explains the anomalous data on the "fusion temperature" of the ash.
The high CaO content physically interferes with flow of the fluid
slag that would result with a smaller addition of CaO. For example, if
the CaO content of typical Illinois No. 6 coal ash were increased to only
20 percent on a. normalized basis, the resulting fluxed coal ash would have
a viscosity of only 10 poise at 1430 C, comparable to that of castor oil
at room temperature. With 60.7 percent CaO, the ash will behave as a
solid rather than a liquid because of all the unreacted CaO.
The pellets remained sufficiently intact during storage and
handling that an acceptable pellet was fed into the boiler. However, it
was observed that some pellets softened during exposure to rain. Weather-
ability tests on these pellets were rerun showing approximately the same
characteristics previously observed, i.e., no softening. It appears that
the weatherability test used during pellet development has some limitations
and that pellets will require some undercover storage or further formulation
refinement for weatherproofing.
EXPERIMENTAL PROCEDURES
Experimental procedures were similar to those of Phase II
experiments with modification of on-line gaseous monitoring. Gases from
the combustion of pelletized coal were sampled as follows. First, a
representative gas sample was continuously removed from the boiler outlet
by placing a "rake" in the center of the first and second cyclone via
an access port. Figure 111-34 shows the relative location of the rake,
boiler outlet, and cyclones. The sample port holes along the length of the
111-94
-------
Boiler Outlet
M
I
Flow path of gas
and partlculate
Gas entry
Particulate build up
Cyclone
Gas Entry
Gas Sample
Probe
FIGURE 111-34. LOCATION AND ORIENTATION OF
GAS SAMPLE PROBE
Induced
Draft Fan
-------
rake were located and sized to draw an equal amount of gas through each
hole. The holes were also located to minimize interference by
particulates, as seen in Figure III-34.
A heated filter maintained at approximately 150 C to avoid water
condensation was located immediately downstream of the sample probe to
remove particulates. The gas was passed through a wet ice trap to remove
excessive moisture and then transported to the on-line monitors through a
6-mm diameter teflon line. An additional fiber glass filter was placed in
the sample line just before the sample pump to remove any extraneous
material that may have passed through the particulate/condensate trap.
Relatively high sample rates were maintained to insure fast
response and short residence times.
All instruments were checked for operational condition before
and after use fay using nitrogen for a zero response gas and an appropriate
certified span gas to set the gain for each of the individual monitors.
EXPERIMENTAL RESULTS
Checkout Runs
Prior to the demonstration test, the fuel pellets were fired for
10 hours to determine the necessary stoker adjustments and to establish
a range of operating conditions.
The limestone/coal pellets were fired without any adjustment to
the stoker mechanism, previously set for a low-sulfur Ohio stoker coal.
The stoker feed mechanism distributed the pellets uniformly over the
grate. This was unexpected since the pellets were all approximately
the same size. It was observed, however, that approximately 50 percent
of the pellets broke randomly into smaller pieces providing a reasonably
good size distribution.
a. Phase II/Phase III Pellet Comparison. Pellets fired in the
Phase III study were significantly superior to those fired in the Phase
II steamplant runs. They burned more readily at lower excess air rates,
111-96
-------
provided improved boiler response (shallower bed), ignited more readily,
and generated lower CO and smoke levels. These improvements are attri-
buted to the fact that the Phase III fuel had a higher heating value,
contained an organic (rather than inorganic binder), contained less ash,
and exhibited superior mechanical strength. However, sulfur retention
was not as high with the Phase III pellets.
b. Stoker Coal/Phase III Pellet Comparison. Phase III pellets
appeared to burn equally as well as the low-sulfur Ohio coal that is
normally fired in the Battelle steamplant boiler. The boiler appeared
to be as responsive to the load and could be operated at comparable excess
air levels. Table 111-30 compares these two fuels. Emissions are corrected
to 3 percent 0,,.
c. Effect of Operating Parameters on Sulfur Retention. Because
it was not the intent of the checkout runs to characterize the emissions
for a variety of boiler operating conditions nor was it possible with the
limited supply of fuel pellets, only limited amounts of data were collected
in the checkout runs.
Sulfur retention was observed to decrease for increasing load
as indicated below for relatively constant excess air (about 80 percent).
Boiler Load, Sulfur Retention, Bed Temperature,
percent full load percent C
0.64 50 1315
0.80 48 1405
0.85 47 1425
The bed temperatures were measured with an optical pyrometer sighted on
the combustion zone at the top surface of the bed. Sulfur retention
varies with bed temperature. However, this observation must be tempered
as the combustion conditions were not closely controlled throughout these
tests.
111-97
-------
TABLE 111-30. COMPARISON OF EMISSIONS FROM COMBUSTION OF A
LOW SULFUR COAL AND LIMESTONE/COAL PELLET
Smoke Fuel N Fuel S
Opacity, Converted, Emitted,
Coal Type percent CO NO percent S02 percent
Low-S coal 10 70 480 18 540 90
Fuel pellet 20 400 310 20 1800 45
111-98
-------
and the observed temperature measurement may not be a good indication of
the actual bed temperature.
At a low-load condition, the excess 02 was varied from 9.5
percent to 16 percent with no significant change in the SO. retention (46
to 50 percent). Bed depths were also varied from 76 to 152 mm. SO-
retention increased somewhat with deeper beds. The increased retention
was attributed to the lower bed temperatures measured for the deeper beds.
Demonstration Test
During the limestone/coal fuel pellet demonstration, the pellet
feed rate was maintained at approximately 1360 Kg/hr for a boiler load of
80 percent. Tables 111-31, 111-32, 111-33, and 111-34 summarize the
results of this test.
a. Sulfur Capture. As indicated in Table 111-31, sulfur
capture was 45 percent during the demonstration test. This sulfur retention
is less than that observed for the model spreader and fixed-bed reactor
experiments firing pellets of similar formulations. Additionally, as
previously discussed, a 75 percent sulfur retention was achieved when
firing a cement-bound pellet with a Ca/S ratio of 7 in the steamplant
during Phase II. The greater sulfur retention of these other experiments
is attributed to the lower bed temperatures, which seldom exceeded 1260 C.
The bed temperatures in the Phase III steamplant demonstration were seldom
less than 1370 C and ran as high as 1455 C. Additionally, with a pulsating
ash-discharge stoker, the fuel bed is violently disturbed. Ash can be
recirculated back into the hot zone. Thus, if sulfur is retained in the
ash at a lower bed temperature, it may be released when the ash is exposed
to a higher temperature.
The average S07 emission level of 1600 ppm during the Method 5
test was verified by the Method 6 wet-chemistry technique. (Wet chemistry
gave an SO. emission level of 1590 ppm.) In addition, as indicated in
Table 111-34, the sulfur balance based on the fuel pellet analysis, the
S02 emission and the sulfur content in the bottom ash (Table 111-33)
was complete.
111-99
-------
TABLE 111-31. EMISSION DATA SUMMARY FOR FUEL PELLET DEMONSTRATION
M
I
Smoke CO at _ ,. _._ Fuel N _ ,_ Fuel S
£ Load. 0 , C02. CO. NO. SO^ Opacity. 3X 02. WO at 32 02. ppro Converted. S02 at 3I._°2« PPm Emitted. Partlculates.
O X ZZZZZZ ppro Computed Measured X Computed Measured X ng/J
80 8.4 10.5 300 310 1600 20 420 2250 440 20 4100 2250 55 258
-------
TABLE 111-32. ANALYSIS OF METHOD 5 FILTER CATCH (Weight Percent)
Ash C Ca CO- Fe Total S
81 19 11 4 54
TABLE 111-33. ANALYSIS OF GRATE DISCHARGE (Weight Percent)
Ash C Ca C03 Fe Total S
97.7 1.8 36.5 0.7 5.8 3.9
TABLE 111-34. SULFUR BALANCE
Sulfur Retained in
Computed Fuel S In, Emitted,as S0?, Bed Ash as S02,
lb/10 Btu lb/10 Btu lb/106 Btu
7.4 (3182 ng/J) 4,1 (1763 ng/J) 3.3 (1419 ng/J)
III-101
-------
b. CO Levels. CO levels from pellet firing were relatively
high compared to those from the firing of conventional stoker coals which
are usually <100 ppm. These higher CO levels may be related to the
nature of the fuel bed and to the fact that the overfire air flow rate
was decreased during the pellet tests. Higher CO levels have been
observed in other pellet firings. Because of the compactness of the
pellet and the limited access of air into it, the capture process first
involves the formation of calcium sulfide via
2CaO + FeS2 £ FeO + CO,
which can account for part of this increase in CO.
Another possible explanation for the higher CO levels was that
the overfire air rate was significantly decreased during pellet firing.
In the Battelle boiler the overfire air jets are only 254 mm above the
grate. With the increased bed depth from pellet firing, the overfire
air jets would have impinged upon the fuel bed if the normal flow rate
were maintained. The impingement would increase ash carryover, increasing
particulate loadings.
c. Particulate Loading. The Battelle steamplant boiler
facility has a mechanical collector to control particulates. Depending
on the ash and sulfur content of the coal, the experiments in Phase II
showed that particulate loadings varied between 86 and 258 ng/J (0.2
and 0.6 lb/10 Btu). Generally, for low S, low ash coals, particulate
loadings were less than 129 ng/J (0.3 lb/10 Btu).
The particulate loading from the firing of the fuel pellet
was 258 ng/J (0.6 lb/10 Btu). This loading was not unusually high for
a spreader stoker firing a 33-percent-ash coal. This loading should be
significantly less for a chain-grate stoker. The smoke opacity was
only 20 percent, which would appear low for a. particulate loading of
258 ng/J if the fly ash collected was from conventional stoker coal.
However, the fly ash from pellet firing is about 50 percent more
III-102
-------
dense and considerably more coarse than from conventional coals. For
equivalent mass loadings, optical density varies inversely with particle
size and density. Thus, the apparent discrepancy between smoke opacity
and particulate loading is explained partially by laws of optics. As
indicated in Table 111-32, about 19 percent of the fly ash was carbon,
a negligible carbon loss.
d. Grate Discharge. Table III-3A shows that the unburned carbon
content in the grate discharge was less than 2 percent. This indicates that
the fuel pellets were burned essentially to completion. Analysis indicates
that Ca and SO, were present and could have combined with water to form a
solid mass. Some minor plugging problems were experienced in the ash-
disposal system when steam was used to control dusting during transport of
the ash.
SUMMARY
The steamplant demonstration indicated the limestone/coal fuel
pellet could be fired in an acceptable manner without modifying the
facility. During the demonstration, sulfur capture levels that would make
the fuel pellet a viable SO control were not achieved. The data suggest
that improved SO- retention could be realized if bed temperature could be
reduced to below 1315 C, perhaps with flue gas recirculation. In addition,
a quiescent fuel bed in a stoker boiler may increase the sulfur retention
in the bed and should reduce particulate emissions.
III-1Q3
-------
SECTION III-6
LIMESTONE/COAL FUEL PELLET PROCESS COST SUMMARY
Table 111-35 summarizes an economic analysis of the limestone/
coal pellet process. This analysis considers costs related to raw
materials, utilities, labor, and capital, including profit, interest,
and income tax. It indicates a process cost of approximately $15.40/Mg
($14/ton) of pellets in addition to the cost of the high sulfur coal.
Increased costs of firing the boiler are not considered. As an example
of such costs, because of the high ash content of the pellet, ash
handling and disposal costs would be higher than for the low-ash con-
ventional coals.
The estimated cost of $15.40/Mg of pellets above the price of
the raw coal is based on the best available data. The cost may vary
depending on the type of system used and whether the process may be inte-
grated into a physical coal cleaning preparation plant. This cost is for
a product with a heating value of 18.6 KJ/g (8000 Btu/lb) and thus adds
Q
about $0.95 per 10 joules ($1 per million Btu) for S02 control. It
indicates that the limestone/coal pellet is cost competitive with other
control strategies.
BASIC ASSUMPTIONS
The following assumptions were used in the analysis.
Mine-mouth operation
Limestone and coal ground to 60 to 100 mesh
Pellet composition:
65 percent high sulfur coal
32 percent limestone
2 percent pregelenized cornstarch
1 percent latex emulsion
Plant capacity of 54.4 Mg/hr (60 tons/hr).
III-104
-------
TABLE 111-35. SUMMARY OF LIMESTONE/COAL PELLETIZING PROCESS COSTS
M
O
Ui
Basts: 60 tons per hour product with 65 percent coal, 32 percent limestone, 3 percent binder
23 hours per day, 330 days per year
1380 tons per day, 455,400 tons per year of product
Fixed plant investment $2,790,000
Working capital 80,000
Interest during construction 250.000
$3,120,000
Item
Hew Materials
Limestone 18 tons/hr, 136,620 tons/year at $8/ton delivered
Pregelatin cormtarch, 9100 tona/year at $20/ton delivered
Latex emulsion, 1.2 ton/'hr, 9100 ton/year at $150/ton delivered
Utilities
Process water 12 tons/hr (48 gpn) 21.9 MM gallon/year at $0.2/M gal
Fuel oil 32 MMBtu/hr, 243 trillion Btu/yr of $3/MMBtu
Pouer 75 percent of 1917 KM or 1440 KW at 50.035/KU-hr
Diesel fuel 5 gph, 37,950 gallon/year et $0.80/gal
Labor Related
Direct labor 7 operators 9 $8/hr plus 25 percent payroll burden
($10/hr total); staffed 365 daya/yr
Supervision -- 15 percent of direct labor
Overhead 50 percent of direct labor and supervision
Capital Related
Maintenance 6 percent of fixed plant investment
Special pelletizer maintenance at $0.30/ton plus
$0.55/ton die and rollers
Front-end loader maintenance at $0.22/hr per machine
Taxes and Insurance 1.5 percent of fixed plant Investment
Depreciation 11 year, straight line on fixed plant Investment
Profit, Interest, income tax 30 percent of total employed capital
TOTAL
Annual Cost*,
Dollar
$1,092,960
1.138,500
4.400
728,600
382,500
22,800
613.200
91,980
306,500
167.400
387,100
3,300
41,850
250,000
936,000
$6,167,100
Per Ton Product,
Dollars
2.4
3.0
0.01
1.60
0.84
0.06
1.30
.20
.70
0.37
0.85
0.01
0.09
0.56
2.05
-$14.00
-------
The pellet composition was based on the results of the pellet development
effort.
PROCESS FLOWSHEET
The economic analysis was based on the process flowsheet presented
in Figure 111-35. In this process
Coal is taken from a pile instead of directly from
an existing mine operation conveyor
Limestone is delivered to a pile by truck
Portland cement is delivered directly to a bin from
a truck by pneumatic feeding system suggested by
Jeffrey Manufacturing Company
Relatively long inclined conveyors from the coal and
limestone piles are assumed. Costs would be about
35 percent less for horizontal conveyors combined
with bucket elevators.
A paddle-type mixer, as suggested by California
Pellet Mill, is used
California Pellet Mill pelletizers and dryers are
costed.
A California Pellet Mill was used in the analysis since cost information was
available. However, pellets can be produced by an extruder at perhaps a
lower cost. Specifications for processing equipment are given in Table
111-36.
SOURCES OF INFORMATION
Information on equipment included in the flowsheet was obtained
from the following sources:
Front-end loaders Caterpillar Tractor
Conveyors/elevators Jeffrey Manufacturing
Storage bins Butler Manufacturing
Feeders Jeffrey Manufacturing
Solids mixer Rapids Machinery
Pelletizers California Pellet Mill
Coolers California Pellet Mill
III-106
-------
Vent Filter
H
M
I
(-"
o
PBODUCT
FIGURE 111-35. COAL/LIMESTONE/CEMENT PELLETIZING
PROCESS FLOWSHEET
-------
TABLE 111-36. COAL/LIMESTONE/CEMENT PELLETIZING PROCESS
Capacity: 60 tons per hour
Feed Storage and Handling Section
1. Front-end loaders
2. Coal conveyer from pile to bin. 45.7-m long, 9.1-m elevation, 36.3
Mg/hr, with hopper, feeder, tresses, walkway, 3.7 KW motor
3. Coal storage bin, 47.7 m^ (2.74-m dia x 7.3 m), 1 hour
4. Limestone conveyer from pile to bin. 30.5-m long, 9.1-m elevation,
18.1 Mg/hr, with hopper, feeder, tresses, walkway, 2.2 KW motor
5. Limestone storage bin. 47.7 m-^ (2.74-m dia x 7.3 m)
6. Portland Cement storage bin. 60.9 nr* (3.7-m dia x 4.4 m) with bin
vent filter and auxiliaries
7. Cement screw auger from bin; 4.6 m, 228.6 m screw, with motor 1.1 KW,
316 SS, 5.4 Mg/hr
8. Cement elevator to feed hopper; 6.1-m elevation, 5.4 Mg/hr, 1.5 KW
9. Cement bin agitator with 1.5 KW motor
3
10. Cement feed hopper; 5,7 m
11. Cement feeder: 5.4 Mg/hr
Coal and Limestone Grinding Section
12. Limestone air-swept mill with drying from 12 to 1%, moisture, 18.1
Mg/hr (for Hardgrove 70), with hot air heater, cyclone, valve, bag-
house, motors, conveyers to crusher
13. Coal air-swept mill with drying from 12 to 1% moisture, 36.3 Mg/hr
(for Hardgrove 55) with hot air heater, cyclone, valve, baghouse,
motors, feeder conveyers to crusher
3
14. Ground limestone storage bin; 47.7 m (2.74-m dia x 7.3 m)
15. Ground limestone feeder to mixer, 18.1 Mg/hr
3
16. Ground coal storage bin; 47.7 m (2.74-m dia x 7.3 m)
17. Ground coal feeder to mixer, 36.3 Mg/hr
18. Fuel oil tank, 76,000 liters, 3-1/2 days
19. Fuel oil pump, 19 1pm
III-108
-------
Mixer Section
3
20. Recycle fines storage bin, 47.7 m (2.74-m dia x 7.3 m)
21. Recycle fines feeder
22. Feed conveyer to mixer from bins; 61.0-m belt, 12.2-m long, 59.9 Mg/hr
23. Feed elevator to mixer; 14 x 8, 9.1-m elevation, 59.9 Mg/hr
24. Mixer, continuous paddle; 1.06 x 3.65 x 1.37 m, 2.66 m max, with
44.7 KW motor drive
3
25. Mix surge hopper, 47.7 m (2.74-m dia x 7.3 m)
Pelletizer Section
26. California Pellet Mill, 13.6 Mg/hr, 23 hours/day, with motors,
starters, feeders, magnet separator, dies, mixer (15 sec), oil pump
27. Conveyers from mix surge hopper feeder to pelletizers, 3 m, 14.85 Mg/hr
28. Conveyers from mix surge hopper feeder to pelletizers, 6 m, 14.85 Mg/hr
29. Water pump, 190 1pm, 450 KPa
Cooling Section
30. Conveyers from pelletizers to coolers, 3 m, 14.85 Mg/hr
31. Elevators from pelletizers to coolers, 6-m elevation, 29.7 Mg/hr, .3 x
.15 m
32. Pellet coolers from 90 C to 40 C, 11.6 m double door with belts, with
980 m3/m
Screening and Product Discharge
33. Screens, 30 Mg/hr, 10% fines with suspension, motor 1.5 KW, starter
34. Recycle fines pickup conveyer, 7.6 m, 2.7 Mg/hr, .75 K motor
35. Recycle fines conveyer to bin, 18.2 m, 5.4 Mg/hr, 1.1 KW motor
36. Recycle fines elevator to bin, 9.1 elevation, 5.4 Mg/hr, 1.5 KW motor
37. Product pickup conveyer, 7.6 m, 27 Mg/hr, 1.1 KW motor
38. Product conveyer inclined with swivel spout, 12.8-m elevation, 45.8
long, 54 Mg/hr, 6 KW motor, walkway
III-109
-------
Screens California Pellet Mill (Rotex)
Other Guthrie
Factors for installation costs Guthrie
Pulverizers a paper "Estimated Cost, Beneficiation, and
Applications of Ultrafine Coal Pulverization" by Foo,
DeCarlo, Jamgochian, and Foster. Presented at 1978 ASME
Winter Annual Meeting,
CAPITAL COST ESTIMATES
Capital cost estimates were based on mine-mouth operation for a
plant of 60 ton/hr capacity and are summarized in Table 111-37.
COMPARISON TO OTHER CONTROL STRATEGIES
The limestone/coal fuel pellet is an attractive control for two
major reasons:
(1) No major modification of the stoker boiler
facility is required to fire the pellets
(2) The cost of $15.40/Mg is competitive with other
control strategies such as used flue gas scrubbers
or low sulfur coals.
The steamplant experiments indicate that neither the stoker
boiler facility nor its operation will require major modification to
fire fuel pellets. The pellets burn similarly to a lower heating value
coal. In contrast, the addition of a flue gas scrubber is a major facility
modification and increases system maintenance.
Cost comparisons of the various types of control strategies are
difficult to interpret, primarily because of different sets of basic
assumptions and different reference points. However, the pellet process
g
costs of $15.40/Mg or $0.95 per 10 joules ($1 per million Btu) are
competitive with flue gas scrubbers. Foley indicated costs of
between $22 and $33/Mg ($20 and $30/ton) of coal for the gas scrubber
for small to medium-sized industrial boilers based on 1973 figures.
III-110
-------
TABLE 111-37. CAPITAL COST ESTIMATES
Installed Cost,
Thousands of Dollars
220
1,200
90
710
220
180
Process Sections
Feed storage and handling
Coal and limestone grinding
Mixer
Pelletizer
Cooling
Screening and product discharge
2,620
Offsites
Site preparation: at 2% of $2,620,000 52
Industrial buildings: at 2% of $2,620,000 52
Miscellaneous utilities: (substations, wells, etc.)
at 1% of $2,620,000 26
Auxiliary facilities: (scales, etc.) at 0.5% of
$2,620,000 i3
Offsite piping: (water, air, oil) at 1.0% of $2,620,000 26
TOTAL (OFFSITE) 169
TOTAL FIXED PLANT INVESTMENT 2,790
Working Capital
Raw material inventories
Product inventories
Spare equipment and tools at 2% of
fixed plant investment
Miscellaneous cash
Thousand
Dollars
Negligible
,ditto
56
34
TOTAL
80
Interest During Construction
9% of fixed plant investment
(1-1/2-year construction period)
Thousand
Dollars
250
III-111
-------
Operating Requirements
Connected Power
Load, KW
Feed storage and handling
Coal and limestone grinding
Mixing
Pelletizing
Drying
Screening and product discharge
TOTAL
Direct Labor, Men
per Shift
10
898
52
S2
123
13
2
1
1
1-1/2
1/2
1
1917
III-112
-------
Mason, et al reported flue gas scrubber costs of about
$22/Mg ($20/ton) for larger industrial boilers based on 1978 figures.
Additionally, the cost differential between high- and low-sulfur coals is
approaching $22/Mg and should increase as the availability of low-sulfur
coals decreases. Another point to consider is that the limestone/coal
pellet concept can be applied to waste coal fines, a readily available
and inexpensive source of fuel.
III-113
-------
REFERENCES
III-l. Komarek. Chem. Eng., ]± (25): 154, 1967.
III-2. Perry, Robert H. and Cecil H. Chilton. Chemical Engineer's
Handbook. 5th Ed., McGraw-Hill.
III-3. Wen, C. Y. Noncatalytic Heterogeneous Solid Fluid Reaction
Models. Ind. Eng. Chem., 6j3 (9): 34, 1968.
III-4. Wen, C. Y. and S. C. Wang. Thermal and Diffusional Effects in
Noncatalytic Solid Reactions. Ind. Eng. Chem., 62_ (8): 31, 1970.
III-5. Ishida, M. and C. Y. Wen. Comparison of Zone Reaction Model and
Unreacted-Core Shrinking Model in Solid-Gas Reactions - I.
Isothermal Analysis. Chem. Eng. Sci., 26; 1031, 1971.
III-6. Ishida, M. and C. Y. Wen. Comparison of Zone Reaction Model and
Unreacted-Core Shrinking Model in Solid-Gas Reactions - II.
Non-Isothermal Analysis. Chem. Eng. Sci., 26; 1043, 1971.
III-7. Ishida, M. and C. Y. Wen. Effectiveness Factors and Instability
in Solid-Gas Reactions. Chem. Eng. Sci., 23; 125, 1968.
III-8. Wen, C. Y. and T. Z. Chuang. Entrained-Bed Coal Gasification
Modelling. Interim Report. U.S. Department of Energy Report
FE-2274-T1, 1978.
III-9. Field, M. A., D. W. Gill, B. B. Morgan, and P.G.W. Hawksley.
Combustion of Pulverized Coal. The British Coal Utilization
Research Association, Leatherhead, England, 1967.
111-10. Mulcahy, M.F.R. and I. W. Smith. Kinetics of Combustion of
Pulverized Fuel: A Review of Theory and Experiment. Rev. Pure
and Appl. Chem., 19_: 81, 1969.
III-ll. Wen, C. Y. and M. Ishida. Reaction Rate of Sulfur Dioxide with
Particles Containing Calcium Oxide. Environ. Sci. Technol. , !_'
703, 1973.
111-12. Kito, M. and C. Y. Wen. Analysis of S02~Limestone Reaction
Systems: Part II. Simulation. AIChE Symposium Series No. 147,
Vol. 71, 119, 1975.
III-114
-------
111-13. Heisler, M. P. Temperature Charts for Inducation and Constant-
Temperature Heating. Trans. ASME, 69; 227, 1947.
111-14. Foley, G. J., et al. Control of SOX Emissions from Industrial
Combustion. Proceedings of First Annual AIChE Southwestern
Ohio Conference on Energy and the Environment, Oxford, Ohio,
October 25-26, 1974.
111-15. Mason, et al. Operating History and Present Status of the
General Motors Double Alkali S02 Control System. Proceedings
Symposium on Flue Gas Desulfurization Volume II, Las Vegas,
Nevada, March, 1979.
III-115
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
4. TITLE AND SUBTITLE
EVALUATION OF EMISSIONS AND CONTROL TECHNOLOGY
FOR INDUSTRIAL STOKER BOILERS
7. AUTHOR(S)
Robert D. Giammar, Russell H. Barnes, David R. Hopper,
Paul R. Webb, and Albert E. Weller
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle, Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NW 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
March, 1981
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2627
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report presents the results of a 3-phase program to evaluate emissions
and control technology for industrial stoker boilers. In Phase I, emission
characteristics were determined for a variety of coals fired in a 200-kW
stoker boiler. 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/coal fuel
pellet was developed and found effective in capturing 80 percent of the fuel
sulfur. Phase II focused on identifying and evaluating potential control concepts.
An 8-MW spreader stoker boiler was used. It was found that improved control of
combustion air, that is underfire and overfire air, resulted in lower excess air
operation (improved efficiency), reduction in particulate loading, smoke, CO
and NO emissions, and had no effect on S02 levels. The limestone/coal pellet
(Ca/S= 7) was successfully fired achieving 75 percent S02 reduction. In Phase III,
the limestone/coal fuel pellet was refined. A pellet was produced that had
physical properties that could survive an industrial coal handling system. This
pellet with a Ca/S molar ratio of 3-1/2 was fired in the 8-MW boiler achieving
sulfur captures of 50 percent. The cost of this pellet would add approximately
one dollar per million Btu to the cost of the raw, high sulfur coal.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Boilers
Combustion
Efficiency
Flue Gases
Criteria Pollutants
POM
Limestone/Coal Pellet
Stokers
Coal
18. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources .
Treated Coals
Stoker Operating Variable
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
c. COSATI Field/Group
s
21. NO. OF PAGES
247
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in s;
type or otherwise subordinate it to mam title. When a report is prepared in more than one volume, repeat the primary title, add vo
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, dot
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing ors
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented'at conference
To be published in. Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 "words or less) factual summary of the most significant information contained in the report. If the report contaii
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
------- |