United States
Environmental Protection
Agency
EPA-600/7-83-042
August 1983
Research and
Development
EVALUATION OF
COMBUSTION MODIFICATION EFFECTS
ON EMISSIONS AND EFFICIENCY OF
WOOD-FIRED INDUSTRIAL BOILERS
Prepared for
Office of Air Quality Planning and Standards
Prepared by
Industrial Environmental Research
Laboratory
Research Triangle Park NIC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
-------
ABSTRACT
Results of full scale tests to evaluate combustion modifications for
emission control and efficiency enhancement on two wood-fired industrial
boilers are reported. These modifications consisted of lower excess air and
variations in the overfire air system operation.
The boiler at Location 3 is fueled with a combination of wood bark and
coal. The implementation of lower excess air reduced NOV emissions by
A
37.2 percent and improved thermal efficiency by 1.2 percent. Variations in
the overfire air system reduced NO by 20.7 percent and improved efficiency by
1 .6 percent. The combination of lower excess air and overfire air system
modification reduced NO by 18.5 percent and improved efficiency by
0.9 percent. A 51 percent load reduction produced only a 3.7 percent NO
reduction and a 4.0 percent loss in efficiency.
The boiler at Location 5 uses hogged wood as the primary fuel and oil
as the supplemental fuel. The effectiveness of lower excess air in reducing
NO., was 12.5 percent with a slight improvement in efficiency (0.6 percent).
X
Adjustment of the auxiliary air dampers produced a 17.2 percent NOX reduction
and a 1.7 percent improvement in efficiency. Polycyclic organic matter (POM)
sampling was performed at both baseline and optimum low-NO conditions. On a
yg/m basis the POM for low-NO conditions exceeded the baseline results by a
factor of two to three. The results obtained are compared to previous
sampling on industrial steam boilers.
ii KVB72-806015-1308
-------
CONTENTS
Section Page
ABSTRACT ii
1.0 INTRODUCTION AND SUMMARY 1-1
2.0 TEST EQUIPMENT DESCRIPTION 2-1
2.1 Emissions Sampling Equipment 2-1
3.0 WOOD BARK/COAL BOILER, LOCATION 3 3-1
3.1 Boiler Description 3-1
3.2 Emissions Sampling 3-4
3.3 Baseline Tests 3-4
3.4 Combustion Modifications 3-11
4.0 HOGGED FUEL BOILER, LOCATION 5 4-1
4.1 Boiler Description 4-1
4.2 Fuel Description 4-3
4.3 Modifications and Tests Performed 4-5
4.4 Polycyclic Organic Matter (POM) Sampling 4-22
5.0 REFERENCES 5-1
6.0 CONVERSION FACTORS 6-1
APPENDIX:
A. GASEOUS AND PARTICULATE EMISSIONS TEST METHODS
AND INSTRUMENTATION
ill
KVB72-806015-1308
-------
FIGURES
Figure Page
3-1 Two-drum Stirling boiler for suspension and thin-bed burning 3-2
of wood
3-2 Air-swept distributor spout for spreader stoker 3-3
3-3 NO emissions as a function of G>2 before overfire nozzle 3-9
modification
3-4 Aerodynamic particle diameter - baseline conditions 3-10
3-5 NO emissions as a function of load for a wood bark boiler 3-12
3-6 NO emissions as a function of stack 02 for a wood bark 3-13
boiler
3-7 Aerodynamic particle diameter - low-NO conditions 3-15
Jt
4-1 Flow diagram of combustion air and hogged fuel induction
systems 4-2
4-2 Boiler total particulate emissions as a function of NO 4-13
emissions for various test conditions on a hogged fuel boiler
4-3 Aerodynamic particle diameter as a function of cumulative 4-15
proportion of impactor catch at low-O2 conditions in a
hogged fuel boiler
4-4 Aerodynamic particle diameter as a function of cumulative 4-16
proportion of impactor catch at low-NO conditions in a
hogged fuel boiler
4-5 Aerodynamic particle diameter as a function of cumulative 4-17
proportion of impactor catch at low-NO conditions in a
hogged fuel boiler
4-6 Total particulate as a function of wood ash content for a 4-19
hogged fuel boiler
4-7 NO emission as a function of wood moisture content for a 4-20
hogged fuel boiler
4-8 NO emissions as a function of stack oxygen for three 4-22
loads in a hogged fuel boiler
4-9 NO emissions as a function of stack oxygen in the 4-24
baseline and low-NOv configurations
A.
4-10 Mark III adsorbent sampling system 4-26
iv KVB72-806015-1308
-------
TABLES
Page
Summary of Combustion Modifications at Location 3 1-2
Summary of Combustion Modifications at Location 5 1-3
Summary of Emissions from Location 3 Wood Bark Boiler 3-5
Location 3 Fuel Analyses 3-8
Location 5 Fuel Analyses 4-4
Location 5 Ash Analyses 4-6
Summary of Gaseous and Particulate Emissions, Location 5 - 4-9
Hogged Fuel Boiler
4-4 Additional Emissions Data, Location 5 4-11
4-5 Emissions of NO at Baseline and Optimum Low-NOx Conditions 4-12
4-6 Summary of POM Analyses for Location 5 - Wood-Fired Spreader 4-28
Stoker
4-7 POM Emission from Oil-, Coal-, and Wood-Fired Boilers: 4-30
Comparison with Present Data
KVB72-806015-1308
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of Robert E. Hall,
the EPA Project Officer, whose direction and evaluation were an important
contribution to the program.
Acknowledgment is also made of the contributions of the environmental,
engineering, operating and maintenance personnel of the plant at which the
boilers were located. Without their willing cooperation, the program could
not have been conducted.
Limitations on Application of Data Reported
The pollutant emission data cited in this report pertain to two
specific wood boilers. These data should not be used to estimate mean
emissions from other types of wood boilers or to predict emission reduction
potentials of combustion modifications until the modifications have actually
been carried out on a greater number of such combustion types.
VI
KVB72-80601 5-1 308
-------
SECTION 1.0
INTRODUCTION AND SUMMARY
This report is one of a series which describe tests of combustion
modifications to control NO emissions from industrial process equipment.
X
This work was performed under EPA Contract No. 68-02-2645 over the time period
from August 1977 to June 1981.
The activities reported herein include tests performed on a wood bark/
coal-fired boiler (Location 3) and a hogged wood fuel boiler (Location 5).
Oil was the supplemental fuel at Location 5.
Variations in load, excess air and overfire air were the combustion
modifications common to both boilers. In addition, lowering combustion air
preheat and positioning of the supplemental fuel oil air damper were performed
at Location 5. Polycyclic organic matter (POM) sampling was also conducted at
Location 5 at both baseline and optimum low-NO conditions.
A
Table 1-1 summarizes the reductions in NO and changes in efficiency
measured at Location 3 for each of the combustion modifications. The overfire
air system modification consisted of increasing each of the overfire air ports
from 1 inch (2.54 cm) to 1-1/2 inches (3.81 cm) in diameter. As shown in
Table 1-1, the lowest NO level obtained resulted from implementing lower
excess air before the modification of the overfire air ports. This arrange-
ment also produced an increase in boiler efficiency of 1.2 percent.
Table 1-2 summarizes the NO reductions achieved at Location 5 and the
change in efficiency for all modifications except reduced combustion air
preheat. This modification could not be fully implemented since the combus-
tion air temperature could only be reduced by 16 - 22 K. Also noted in this
table is the NO mass emission factor measured after each modification had been
implemented.
Of some interest for this particular boiler is the effect of load
changes on NO emissions. As noted, increasing load (18 percent) actually
1-1 KVB72-806015-1308
-------
TABLE 1-1. SUMMARY OF COMBUSTION MODIFICATIONS AT LOCATION 3
Control
Lower Excess Airt (146)tt
Lower Excess Air§ (184)
Overfire Air Dampers§ (174)
Load Reduction (51%)§,# (140)
NO Reduction,
%
37.2
18.5
20.7
7.9
Efficiency
Change, %
+1.2
+0.9
+1 .6
-4.0
NO After Control
ng/J*
92
150
138
129
*NO as N02.
tBefore overfire air system modification.
§After overfire air system modification.
ttLoad reduction referenced to nominal operation at 80% of rating.
ttValue in parenthesis is baseline NO (ng/J) before combustion modification.
1_2 KVB72-806015-1308
-------
TABLE 1-2. SUMMARY OF COMBUSTION MODIFICATIONS AT LOCATION 5
NO Reduction, Efficiency NO After Modification
Modification % Change, % ng/J*
Lower Excess Air (40 )§
Increase Overfire Air (46)
Auxiliary Air Damper (36)
Load Changet
+18% (40)
-30% (40)
12.5
21 .7
17.2
27.5
30.0
+0.6
-1 .3
+1.7
+0.9
+1 .8
35
36
30
29
28
*NO as NO2.
tLoad change referenced to nominal operation at 76.5% of rating.
§Value in parenthesis is baseline NO (ng/J) before combustion modification.
KVB72-806015-1308
-------
reduced the NO concentration and mass emission factor. This characteristic of
peak NO occurring in the mid-load range is somewhat unusual.
Polycyclic organic matter (POM) samples "were collected and analyzed at
both baseline and low-NOx (auxiliary air damper adjustment) conditions. The
significant finding was that the total POM at the low-NO,. condition could be
X
two to three times higher than that measured under baseline conditions. This
large difference could be due more to fuel property variations than to com-
bustion modification although the trend of higher POM with Tower NO ' has been
• A
observed previously on a coal-fired spreader stoker (Reference 1).
1-4 KVB72-806015-1308
-------
SECTION 2.0
TEST EQUIPMENT DESCRIPTION
2.1 EMISSIONS SAMPLING EQUIPMENT
Gaseous emission measurements were made using analytical instruments
and equipment contained in a government-furnished mobile instrumentation
laboratory. A plan view of the trailer is shown in Appendix A. Total partic-
ulate measurements were made using an EPA Method 5 sampling train produced by
Joy Manufacturing Company. Particulate size distribution was measured using a
Brink Cascade Impactor. Total oxides of sulfur were measured by wet chemistry
methods using the sampling train and analytical procedure of the Goksoyr-Ross
method. Smoke density was measured using an automated Bacharach smoke spot
pump. Stack opacity readings were made during particulate tests according to
EPA Method 9.
2.1.1 Gaseous Emissions Sampling System
The laboratory is equipped with analytical instruments to continuously
measure concentrations of NO, NO2, CO, C02, 02, SO2, and hydrocarbons. The
sample gas is delivered to the analyzers at the proper condition and flow rate
through the sampling and conditioning system described below. Appendix A
describes the analytical instrumentation and the details of the sampling
system.
A flow schematic and description of the flue gas sampling and analyz-
ing system is presented in detail in Appendix A. Briefly, the sampling system
used pumps to continuously draw flue gas from the heater into the labora-
tory. A high-capacity heated positive displacement diaphragm pump was used to
draw a high volume of flue gas into the analyzers to assure quick response to
source variations.
Special precautions were required to obtain a representative sample
for the analysis of N02, SO2/ and hydrocarbons. These precautions consisted
of insuring that the sample was kept above its dew point, to prevent loss of
2~1 KVB72-806015-1308
-------
sample components in condensed water. For this reason, an electrically heated
9.5 mm (3/8 in.) Teflon sample line was used to bring the sample into the
laboratory for analysis. The hot pump provided heated sample directly to the
hydrocarbon, SCU/ and NO analyzers.
A portion of the sample pump discharge was also sent through a refrig-
erated condenser [to reduce the dew point to 275 K (35°E)],' through, a large
rotameter with flow control valve, and then to the O_, NO, GO, and CO- instru-
mentation. Flow to each individual analyzer was measured and controlled with
smaller rotameters and flow control valves. Excess sample was vented outside
the laboratory.
2.1.2 Particulate Emissions
Particulate samples were taken at the same sample location as the gas
sample using a portable effluent sampler produced by Joy Manufacturing
Company. These tests were made at baseline and optimum low-NO operating
X
conditions. This system, which meets EPA design specifications for Test
Method 5, Determination of Particulate Emissions from Stationary Sources
(Federal Register, Volume 36, No. 27, page 24888, December 23, 1971) was used
to perform both the initial velocity traverse and the particulate sample
collection. Dry particulates were collected in a heated case that contained a
110 mm glass-fiber filter for retention of particles down to 0.3 microm-
eters. • Condensible materials were collected in a train of four impingers in a
chilled water bath.
Particle size was measured at baseline and optimum low-NO conditions
X
at the same sample location as the Method 5 tests. A Brink Cascade Impactor
was used for all of the sizing tests because of its high grain loading
capability. This impactor was capable of fractionating particles in-situ into
six aerodynamic size ranges (five collection stages and one backup filter).
The size range capability of this impactor was approximately 0.4 ym to 10 pm.
The Method 5 sampling train and procedures, and the impactor opera-
tional procedures, are discussed in Appendix A.
2-2 KVB72-806015-1308
-------
2.1.3 Wet Chemical SO Measurement
The Goksoyr-Ross technique was used to sample the stack gas for S03.
This method uses controlled condensation of the stack gases in a coil main-
tained at 333-344 K (140-160°F) by a water bath. This temperature is below
the sulfuric acid (H-SO.) dewpoint so that the SO3 and the water vapor in the
flue gas condense upon the coil walls to form H2SO4 droplets. The SO2 in the
flue gas passes through the coil and is collected in impingers containing a
weak hydrogen peroxide solution. Following the impingers, the flue gas flows
through a dry gas meter and is then discharged into the atmosphere.
The coil rinse and impinger liquid are each analyzed by means of an
acid-base titration with a sodium hydroxide solution. Both the SO3 and SO2
concentrations may be determined from this procedure.
2.1.4 Smoke Spot and Opacity Measurement
On combustion equipment where smoke numbers normally are taken, such
as oil-fired boilers, the smoke number is determined using test procedures
according to ASTM Designation: D 2156-65. The smoke number is determined at
each combustion modification setting of the unit. Examples are baseline,
minimum excess air, low load, etc., and whenever a particulate concentration
is measured.
Smoke spots are obtained by pulling a fixed volume of flue gas through
a fixed area of a standard filter paper. The color (or shade) of the spots
that are produced is visually matched with a standard scale. The result is a
"Smoke Number" which is used to characterize the density of smoke in the flue
gas.
Opacity readings were taken by a field crew member who is a certified
graduate of a U.S. Environmental Protection Agency approved "Smoke School."
Observations are made at the same time that particulate measurements are made
and as often in addition as deemed necessary to gather the maximum amount of
information. The procedures set forth in EPA Method 9, "Visual Determinations
of the Opacity of Emissions for Stationary Sources," were followed.
2-3 KVB72-806015-1308
-------
SECTION 3.0
WOOD BARK/COAL BOILER, LOCATION 3
3.1
BOILER DESCRIPTION
This boiler uses by-product wood bark from the plant's pulping opera-
tion and coal, or coal only. The unit was built in 1966 by the Wickes Boiler
Company and is rated at 100,000 Ib/hr (45.4 Mg/h) steam flow firing a combina-
tion of coal and bark or coal only. The boiler is equipped with a travelling
grate spreader stoker. The wood bark is injected pneumatically above the
three coal feeders through three ports located above the front overfire air
ports. The smaller wood particles burn in suspension while the rest of the
bark burns on the grate. Preheated combustion air is introduced under the
grate after passing through a tubular air preheater. The unit is equipped
with a multiclone dust collector. Figure 3-1 is a cross-section of a typical
boiler used for suspension and thin-bed burning of wood. Figure 3-2 is a bark
distributor which is similar to the three air-swept distributors used in the
Location 3 unit. Other boiler characteristics are:
Maximum Continuous High
Pressure Steam Output:
Steam Conditions at
Superheater Outlet:
Heating Surface:
Firing coal and bark or coal
only - 100,000 Ib/hr (45.4 Mg/h)
Temperature - 800 °F (700 K)
Pressure - 600 psig (4.14 MPa)
11,247 ft2 (1,045 m2;
6,060 ft2 (563 m2)
1,768 ft2 (164 m2)
Furnace Volume 6,150 ft3 (174 m3)
Boiler
Air Heater
Water Walls
Bark flow rate to the boiler was controlled as well as possible. Flow
fluctuations and interruptions occurred from time to time and are considered
normal operation. The percentage of bark heat input was estimated from the
steam chart when bark flow was interrupted.
The test unit was limited in the amount of wood bark being burned due
to high superheat metal temperature. Whenever superheat metal temperature
3-1
KVB 72-806015-1308
-------
Stack
Induced Draft
Fan
Figure 3-1. Two-drum Stirling boiler for suspension
and thin-bed burning of wood.
3-2
KVB72-806015-1308
-------
Bark Feed
Distributor
Spout Air
Rotating Damper
for Pulsating Air Flow
Figure 3-2. Air-swept distributor spout for spreader stoker.
3-3
KVB72-806015-1308
-------
reached 810°F (706 K), the bark flow was reduced until the tube temperature
dropped below 800°F (700 K). The high superheat metal temperature is caused
by the large amount of small bark particles which are injected above the front
overfire air ports, burning in suspension above the grate near the superheater
tubes. While burning coal only, the unit encountered no problem in maintain-
ing the design superheat metal temperature of 750°F (672 K).
3.2 EMISSIONS SAMPLING
Gaseous and particulate emissions measurements were made at a single
port in the duct work downstream of the multiclone dust collector and induced
draft fan. A heated sample line was used to sample all gaseous emissions. No
access was available for measurements upstream of the dust collector. During
the test series problems were encountered with the heated sample pump and the
problem could not be fixed at the test site. For test numbers 3-20 through
3-32 only cold line data could be taken. Therefore, neither N02/ SO2 nor
hydrocarbons could be sampled or measured during these tests.
At the dust collector outlet a Lear Siegler Optical Transmissometer
was installed by the boiler owner. The readings of this instrument were
recorded in the control room data sheets.
Appendix A describes the instrumentation employed.
3.3 BASELINE TESTS
Baseline emission measurements were made with the boiler in the "as-
found"* condition firing about 20 percent wood bark and 80 percent coal. All
test results are summarized in Table 3-1. The results of the wet chemical
measurements made during test 3-25 were:
S02—891 ppm (corrected to 3% 02), 794 ng/J
SO3—4 ppm (corrected to 3% O2), 4 ng/J
*As-found relates to the unit operation at the time of the KVB test crew
arrival. Baseline refers to the unit operating at its nominal conditions.
As-found and baseline conditions are usually the same, however, they can
differ.
3-4 KVB 72-806015-1308
-------
TABLE 3-1. SUMMARY OF EMISSIONS FROM LOCATION 3 WOOD BARK/COAL BOILER
Nominal Steam
Load
Test
No.
3-1
3-2
3-3
3-4
3-5
3-6
3-7
3-8
3-9
3-10
3-11
3-12
3-13
3-14
3-15
3-16
3-17
3-18
3-19
3-20
3-21
3-22
3-23
3-24
3*25
3-26
3-27
3-28
3-29
3-30
3-31
3-32
Date
1979
4/11
4/11
4/11
4/11
4/12
4/13
4/16
4/16
4/16
4/16
4/17
4/17
4/17
4/17
4/18
4/18
4/18
4/18
4/19
4/19
4/19
4/19
4/19
4/20
4/20
4/23
4/24
4/24
4/25
4/25
4/25
4/25
Mg/h
37.2
37.7
37.2
37.2
15.9
36.3
25.0
24.0
25.0
25.9
35.4
35.4
35.4
35.4
35.4
35.4
36.3
35.4
35.4
35.4
35.4
35.4
34.9
34.9
29.5
37.2
37.7.
38.1
18.2
27.2
36.8
46.7
103
Ib/hr
82
83
82
82
35
80
55
53
55
57
78
78
78
78
78
78
80
78
78
78
78
78
77
77
65
82
83
84
4°,
60
81
103
Heat
Input
Rate
MW
29.4
29.8
29.4
29.4
12.6
28.7
19.8
19.0
19.8
20.5
28.0
28.0
28.0
28.0
28.0
28.0
28.7
28.0
28.0
28.0
28.0
28.0
28.7
28.7
23.3
29.4
29.8
30.2
14.4
21.5
29.1
37.0
°2
9.3
10.8
8.2
7.8
11.9
8.7
10.1
8.7
11.0
9.8
9.6
10.4
10.6
9.8
9.7
10.4
9.2
8.5
9.8
10.0
9.6
10.6
9.9
9.9
8.8
8.2
9.6
9.4
11.9
10.8
9.3
9.1
co2
10.3
8.6
12.4
12.8
8.3
9.9
9.8
10.8
8.4
9.4
8.7
8.9
8.1
8.3
8.9
7.7
7.6
10.1
9.0
7.7
7.7
6.8
9.5
8.9
10.2
10.2
9.8
10.2
7.5
8.6
9.5
9.9
NOx
Ppm'
Converter
3ut of
X
i
205
256
202
170
311
204
230
306
438
321
301
310
231
251
275
i
i
i
ti
o
I
01
in
0
$
0
1
o.
o
X
i
ng/J
ervice
i i i
i i i
~
130
162
128
107
196
129
145
193
217
203
190
196
146
159
174
— —
—
—
__
—
—
—
--
NO
ppm'
231
394
184
145
205
247
198
170
306
201
216
297
415
312
292
306
227
238
275
221
229
243
218
222
168
192
212
178
205
214
221
268
ng/J
146
249
116
92
130
156
125
107
193
127
136
188
262
197
184
193
143
150
174
140
145
154
138
140
106
121
134
112
130
135
140
169
HC
ppm
0
0
0
0
0
108
206
0
0
924
1895
1995
1341
723
2506
6708
4653
1728
849
i
i
i
0)
0
-rj
e
o
4J
3
O
i
&,
u
o
X
ng/J
0
0
0
0
0
23
44
0
0
196
402
423
284
153
532
1423
987
367
180
— —
—
—
— -
—
—
--
—
CO
ppm'
362
574
387
498
613
292
1032
505
207
587
268
304
87
129
288
450
275
238
493
319
308
243
154
292
465
192
391
472
828
281
230
364
ng/J
139
220
149
191
235
112
396
194
80
225
103
117
33
50
111
173
106
91
189
123
118
93
59
112
179
74
150
181
318
108
88
140
so2
ppm'
677
724
647
584
268
1062
396
527
1728
225
1011
1596
2769
2507
1180
645
610
864
1489
i
i
s
I
11
10
o
4J
O
1
O.
ng/J
596
637
569
514
236
934
348
464
1520
198
889
1404
2436
2205
1038
567
537
760
1310
— —
—
—
__
—
__
—
—
—
3-5
KVB72-806015-1308
-------
TABLE 3-1.
(CONTINUED)
Test
No.
3-1
3-2
3-3
3-4
3-5
3-6
3-7
3-8
3-9
3-10
3-11
3-12
3-13
3-14
3-15
3-16
3-17
3-18
3-19
3-20
3-21
3-22
3-23
3-24
3-25
3-26
3-27
3-28
3-29
3-30
3-31
3-32
Total
Particulate
* Q lb/106
Bark Btu ng/J
31
31
31
31
50
28 0.244 105
23
23
0
24
15
15
0
0
16 0.504 217
14
6
6
15
15
15
15
15
21
26
16 0.320 138
24 0.400 172
29
10
10
10
10
Solid
Particulate Stack
lb/10S Temperature
Btu ng/J °F K
420
430
430
440
420
0.204 88 428
431
417
408
435
430
428
417
410
0.428 184 428
425
431
435
431
432
435
431
429
430
440
0.274 118 437
0.361 155 435
440
420
419
420
430
489
493
494
500
489
493
495
487
482
497
494
493
487
483
493
491
495
497
495
495
497
495
494
494
500
498
497
500
489
488
489
494
Eff.
%
81.56
79.77
82.77
82.55
79.01
80.99
80.31
81.78
79.16
79.88
79.35
79.70
78.76
79.46
79.68
79.06
77.22
80.39
79.19
77.72
77.61
76.86
80.48
79.34
80.47
79.31
79.56
80.02
76.42
78.65
79.61
77.82
Comments
As found boiler test - high barkflow
0 variation - high
0 variation
0 variation - low
Low load - approx 50% barkflow
Baseline particulate - high barkflow
Medium load - approx 23% barkflow
Medium load, low 0 - approx 23% bark
Medium load - coal only
Medium load - high barkflow^
Baseline - low barkflow
High air - low barkflow
High air - coal only
Normal air - coal only
Baseline particulate - approx 16%
0 variation
0 variation
O variation
Baseline - overfire air variation
Overfire air variation
Overfire air variation
Overfire air variation
bark
Overfire air variation - low excess air
Low NOx - Cascade impactor
SOx test
Low NOx particulate
Baseline particulate - approx 24%
Baseline - Cascade impactor
Load variation
Load variation
Load variation
Load variation
bark
*ppm corrected to 3%
3-6
KVB72-806015-1308
-------
Also shown in Table 3-1 is the bark contribution to the total boiler heat
input (with coal supplying the balance).
Fuel and ash analyses are presented in Table 3-2. Of note is the high
carbon content of the bottom and fly ash. The boiler efficiencies noted in
Table 3-1 include the loss associated with unburned carbon. Tests 3-1 through
3-6 were conducted with the original one inch (2.54 cm) overfire air
nozzles. Tests 3-7 through 3-32 were run after the 12 upper and five lower
overfire air ports in the back of the boiler were changed to 1-1/2" diameter
(3.81 cm) to enlarge overfire air capacity and increase turbulence.
Boiler load during the test series was approximately 80 percent of
rated load except for the load variation tests. NO emissions at the baseline
condition were 146 ng/J or 231 ppm at 3 percent Oj. NO emissions as a
function of O~ before the overfire air nozzle modifications are shown in
Figure 3-3. The smoke limit was found at 7.8 percent O2 at a load of 82,000
Ib (37.2 Mg/h) of steam per hour."
A baseline test (Test 3-11) immediately after the overfire air port
modification showed 136 ng/J of NO or 216 ppm at 3 percent O2- NO and S02
emissions increased as coal flow increased. Highest NO emissions occurred
when only coal was burned, i.e., 196-277 ng/J (Tests 3-9, 3-13, and 3-14).
Solid particulate emissions during baseline conditions were measured
with two different amounts of bark flow. The highest particulate emissions
(184 ng/J or 0.428 lb/10 Btu, solid) occurred when approximately 16 percent
of the fuel was made up by bark. At 25 percent bark flow solid particulate
emissions decreased by 16 percent to 155 ng/J or 0.361 lb/10 Btu. Particu-
late size distribution was also measured using a Brink cascade impactor. Data
on the size distribution are shown in Figure 3-4 for test number 3-28, where
particle diameter as a function of cumulative proportion of impactor catch is
plotted. Approximately 46 percent of the particulate is below 3 ym
aerodynamic diameter.
3-7 KVB 72-806015-1308
-------
TABLE 3-2.
LOCATION 3 FUEL ANALYSES
CD
Test No.
Date
Fuel Type
Ultimate Analysis
Moisture, % weight
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen (diff.), %
Proximate Analysis
Moisture, % weight
Ash, %
Volatile Matter, %
Fixed Carbon, %
Heat of Combustion
Gross Btu/lb
Net Btu/lb
Bottom Ash
Carbon, *
Gross Btu/lb
Fly Ash
Carbon, *
Gross Btu/lb
3-6
4-13-79
Coal
6.63
64.89
4.25
1.48
2.36
8.94
11.45
6.63
8.94
38.78
45.65
12,050
11,660
33.21
4,370
3-15
4-16-79
Coal
6.06
66.12
4.43
1.53
1.96
8.46
11.44
6.06
8.46
37.30
48.18
12,160
11,750
would
3-26
4-23-79
Coal
6.62
61.83
4.33
1.45
3.01
12.20
10.56
6.62
12.20
37.52
43.66
11,580
11,180
1.1.56
not ignite
25.74
3,590
3-27
4-24-79
Coal
6.77
62.60
4.33
1.50
2.30
11.07
11.43
6.77
11.07
37.49
44.67
11,470
11,070
3-6
4-13-79
Wood
51.51
25.58
2.84
0.16
0.030
1.51
18.37
51.51
1.51
37.21
9.77
8,950
NR*
3-15
4-18-79
Wood
53.87
23.93
2.77
0.17
0.029
1.42
17.81
53.87
1.42
37.23
7.48
8,890
NR
3-26
4-23-79
Wood
35.78
33.59
3.77
0.17
0.025
1.61
25.05
35.78
1.51
51.53
11.08
9,060
NR
3-27
4-24-79
Wood
41.23
31.01
3.55
0.20
0.026
1.56
22.42
41.23
1.56
46.51
10.70
8,890
NR
*NR = not reported by testing laboratory.
-------
400
i r
i—i—r
300
Fuel: s 70% Coal
30% Wood Bark
( ) Test Number
Boiler Load: 37.2-37.7 Mg/h
•a
o
<*>
g
a
o
2;
200
100
I
I
(1)
(4)
Smoke limit
@ 7.8% 0,
5 6
>, %, dry
10
11
Figure 3-3. NO emissions as a function of O before
overfire air nozzle modification.
3-9
KVB72-806015-1308
-------
10
H1
O
to
00
o
en
o
M
Ul
i
H
CO
O
CD
o
w
EH
H
Q
O
EH o 9
p, u.y
U
p
8
a
0.3
0-1
LI I I I I I I 1—I 1 I I I I I I
0.01
i—rr
Location 3 Wood Bark Boiler
Test No. 3-28
29% Wood Bark and 71% Coal
Load 38.1 Mg/h
O^ = 9.4%
Brink Impactor
Downstream of Multiclone
I I I i I i i 1 I I I I I 1 I I I I II
5 10 20 60 80 90 95 98 99
Cumulative proportion of impactor catch, % by mass
Figure 3-4. Aerodynamic Particle Diameter - Baseline Conditions
-------
3.4 COMBUSTION MODIFICATIONS
Combustion modification testing included load variation, variation of
excess air and overfire air adjustments with the modified nozzles described
previously. NO emissions as a function of load are presented in Figure 3-5.
A test series (Tests 3-15 through 3-18) was conducted to evaluate the
effect of stack oxygen on NO emissions. The overfire air was left constant -
100 percent open which is normal for these tests. Stack ©2 was controlled by
adjusting the damper on the forced draft fan which supplies the undergrate
air. The baseline condition for the 02 variation (Test 3-15) was 9.7 percent
02 with 184 ng/J of NO emissions. Stack O2 during these tests varied from a
high value of 10.4 percent to a low value of 8.5 percent. The effect of stack
©2 on NO emissions is shown in Figure 3-6 which includes all the test data
measured. Lowering the stack 02 resulted in a decrease in NO from 193 ng/J at
10.4 percent 02 to 150 ng/J at 8.5 percent 02. During the duration of the low
02 test no clinkering was observed and combustion conditions appeared to be
good.
After the O2 variation test series the boiler was returned to the
baseline condition to provide a baseline check point for the overfire air
tests. The dampers for the two overfire air headers in the back of the boiler
were set in their normal position which is 100 percent open. The fly ash
reinjection headers were also in their fully open position. The front over-
fire air dampers were 30 percent open for the upper header and 90 percent open
for the lower header. At baseline conditions (Test 3-19) NO emissions were
174 ng/J with 189 ng/J CO emissions.. The dampers for the lower row of over-
fire air jets and for both fly ash reinjection headers were reduced to
50 percent open. NO emissions at this test point (Test 3-23) dropped to
138 ng/J with 59 ng/J of CO.
During the test series, it was observed that the SO2 measurements were
varying. This was due to fluctuations in wood bark flow. At a stable load
condition whenever bark flow decreases more coal is automatically fed onto the
grate to keep the load steady. The higher coal flow then increases the SO2
emissions.
KVB 72-806015-1308
-------
300
200
(31)
(30)
(29)
o
<#>
0-
o
z;
100
Fuel: s 80% Coal
= 20% Wood Bark
( ) Test Number
J L
1
1
1
10
20
30
40
50 60 70 80
Load, 103 Ib steam/hr
90 100
110
Figure 3-5.
NO emissions as a function of load for a
wood bark boiler.
3-12
KVB72-806015-1308
-------
400
n
T3 300
(N
O
<*>
fl
4J
(0
a
Qj
200
i
100
c
1 1 1 1 1 1 1 1 1 1
^b Baseline before modification
m O Variation before modification
/ \ Overfire air (13)\^
~~ C"} Baseline after modification (2) ^V
1 J O variation after modification
\/ Coal only
(14)Oa(1€
"~ (15)O
O(19)
(6) i ^
d8)CT riS^^
- (26) A (ID ^727) _
(3)" O(28)
• (^
Fuel: = 80% Coal
= 20% Wood Bark
Boiler Load 34.9-38.1 Mg/h
( ) Test Number
1 1 1 1 1 1 1 1 1 1
)123456 789 10 11
Stack Oxygen, %, Dry
Figure 3-6. Location 3 — NO emissions as a function of stack
3-13
KVB72-806015-1308
-------
Total and solid particulate emissions were measured at the low NOX
condition described earlier. Solid particulate concentration was 118 ng/J
(0.274 lb/106 Btu) with the unit operating at 8.2 percent stack O2. NO
emissions were 121 ng/J at this condition.
The low-NO cascade impactor test (Test 3-24) is shown in
X
Figure 3-7- Particulate diameter as a function of cumulative proportion of
impactor catch is plotted. About 27 percent of the particles are below 3 um
aerodynamic diameter. The geometric mean particle size and geometric disper-
sion coefficient are 6 ym and 1.099 \m, respectively. A comparison of the
baseline (Figure 3-4) and low-NOx results (Figure 3-7) indicates that the
geometric mean particle size for baseline operation is approximately
50 percent of that measured during low-NOx operation (3.2 ym vs. 6 ym) .
Closing the dampers for the overfire air and fly ash reinjection (low-NOx
configuration) resulted in the production of larger particulates, but at a
reduced mass rate (118 ng/J vs. 155 ng/J).
A wet chemistry SOX test (Test 3-25) resulted in 784 ng/J of S02 and
4 ng/J of SO-,.
The conclusions from these tests are that closing the dampers on both
the overfire air and fly ash reinjection resulted in lower particulate emis-
sions. Little change in NO emissions was found when compared with data
obtained before the modification. Overfire air adjustment reduced CO emis-
sions (i.e., more complete combustion) even at low excess air conditions.
Reduced excess air firing reduced NO emissions and helped this unit's problem
with high superheater metal temperatures.
3.4.1 Efficiency
Efficiency of the wood bark boiler was calculated using the heat loss
method described in the ASME Power Test Code. The appropriate fuel analysis
in Table 3-2 was used in the calculations. Stack gas losses were calculated
from the flue gas analyses and radiation loss was estimated from the ABMA
standard Radiation Loss Chart. The efficiency data for each test condition is
3-14 KVB 72-806015-1308
-------
10
U)
I
g
o
LO
a:
u
EH
a
w
ij
o
M
H
«
-^
DJ
u
1
0.9
0.5
i r
O
1 I
i r
T
Location 3 Wood Bark Boiler
Test No. 3-24
21% Wood Bark, 79% Coal
Load: 34.9 Mg/h
O =9.9%
2
Brink Impactor
Downstream of Multiclone
Q
§
a
to
I
CO
o
CTl
o
u>
o
oo
0.1
I I I
I
I
I
J I I I I I I
I I
I I
0.01
2 5 10 20 50 95 98 "
Cumulative proportion of impactor catch, % by mass
Figure 3-7. Aerodynamic Particle Diameter - Low-NO Conditions
x
-------
presented in Table 3-1. Efficiency varied from 82.77 percent at 8.2 percent
02 and a load of 82,000 Ib/hr (37.2 Mg/h) steam flow to 76.86 percent at
10.6 percent 02 and 78,000 Ib/hr (35.4 Mg/h) steam flow. Low excess air
firing resulted in improved efficiency without adversely affecting other
operating conditions.
3-16 KVB 72-806015-1308
-------
SECTION 4.0
HOGGED FUEL BOILER, LOCATION 5
4.1 BOILER DESCRIPTION
Full-scale combustion modification tests were carried out on a hogged
fuel power boiler located at a large pulp and paper mill. The boiler is a
spreader stoker rated at 58.6 MW thermal input (200,000 Ib/hr steam flow —
25.3 kg/s) when firing wood. Wood is the primary fuel and oil is used as a
supplementary fuel only when the demand exceeds 58.6 MW (200,00 Ib/hr — 25.3
kg/s). The unit was placed into operation in September, 1977- The design
specifications of the Combustion Engineering boiler are as follows:
1. CAPACITY - 200,000 Ib/hr (58.6 MW)-wood only; 250,000 Ib/hr
(73.3 MW)-wood and oil
2. DESIGN PRESSURE - 700 psig (4,923 kPa)
3. OPERATING PRESSURE - 390 psig (2,785 kPa)
4. STEAM OUTLET TEMP - 600 deg. F (316 K)
5. STOKER - traveling grate (heating surface = 352 sq. ft. or
32.7 m2)
6. OIL BURNERS - (4) at one elevation, oriented tangentially
7. IGNITORS - High energy arc (1) per burner
8. BOILER HEATING SURFACE 15,242 sq. ft. (1,416 m2)
9. WATER WALL HEATING SURFACE 5,482 sq. ft. (509 m2)
A schematic of the combustion air, flue gas, and wood handling systems
is given in Figure 4-1. The hogged fuel was fed from the storage bin via
conveyor belt to the screw feeders (large pieces of wood are recycled through
the hogger where they are chopped and sent back through the system). The wood
flow rate was controlled by the screw feeder RPM. The wood flow was deter-
mined volumetrically by integrating over the screw feeder revolutions. A
strip chart in the control room displayed the mass flow rate of the wood fuel
4-1 KVB72-806015-1308
-------
I
to
-J
to
I
CO
o
en
o
i->
LTl
I
h-1
CO
O
CD
TO
SCRUBBER
AND
STACK
(Sample
Ports C)
NO. 10 POWER BOILER
STEAM COIL
AIR HEATER
I.D. FAN
HOG FUEL
STORAGE BIN
BELT CONVEYOR
FROM STORAGE BIN
TO FLIGHT CONVEYOR
AND SCREW FEEDERS
F. D. FAN
SPREADERS
UNDERGRATE
AIR
Figure 4-1. Flow diagram of combustion air and hogged fuel induction systems.
-------
which was determined using an average density for the wood. The wood mass
flow measurement was not very accurate because of varying densities. This
fact was verified after the tests by calculating wood flows based on steam
flow, actual fuel heating value (as determined by laboratory analysis), and
efficiency data.
The balanced draft combustion air system consisted of an air heater,
undergrate air system (four zones front to back), overfire air ports (four
elevations each with four ports located in the corners and oriented tangen-
tially), and an air system for the four oil guns (controlled by five dampers
located near each gun).
The flue gas, after leaving the boiler proper, passed through the air
heater making a right-angle bend at the air heater hoppers. From this point
the flue gas passed through the multiclone and on through a venturi-type
scrubber and out the stack. A bypass stack was used when the scrubber was
off-line.
Most of the emissions sampling was done at the boiler outlet above the
entrance to the air heaters (Ports A). Simultaneous particulate measurement
was made at the multiclone outlet during one test (Ports B) and at the stack
(Ports C) during another test to determine the particulate removal efficiency
of different sections of the flue gas system.
4.2 FUEL DESCRIPTION
The hogged fuel consists of sawmill wastes purchased from neighboring
mills. All hogged fuel arrives by barge. Bark and wood waste from fir and
hemlock logs constitutes approximately 90 percent of the hogged fuel. Mois-
ture content of this fuel can vary from 44 percent to 58 percent depending on
source and season. Salt content of the fuel varies from 0.7 percent to
1.6 percent. The salt content of the hogged fuel is the result of storing or
transporting logs in salt-containing waterways. Most of the hogged fuel is
unloaded directly into an inside storage bin.
Three wood fuel analyses and two No. 6 oil analyses are presented in
Table 4-1. Analyses of ash samples obtained during Test 5/2-1d are shown in
4-3 KVB72-806015-1308
-------
TABLE 4-1. LOCATION 5 FUEL ANALYSES
Sample Identification:
1. Wood fuel sample Loc. 5
2. Wood fuel sample Loc. 5
3. Wood fuel sample Loc. 5
4. No. 6 Fuel Oil Loc. 5, 9-13-79, Test 5/1-1
5. No. 6 Fuel Oil Loc. 5, 10-23-79, Test 5/7-2C
9-20-79, Test 5/2-lb
10-3-79, Test 5/2-4a
10-17-79, Test 5/7-2
Wood Fuel Samples:
Proximate Analysis:
Moisture, %
Volatile Matter, %
Ash, %
Fixed Carbon, %
Ultimate Analysis (Dry Basis):
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Heat of Combustion (Dry Basis)
Gross Btu/lb (kJ/kg)
No. 6 Fuel Oils
Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Heat of Combustion:
Gross Btu/lb (kJ/kg)
Net Btu/lb (kJ/kg)
1
51.76
37.59
2.55
8.10
i):
• S):
1
47.13
5.66
0.16
0.11
5.29
41.65
8207
(19090)
4
89.
8.
0.
1.
0.
0.
2
44.64
42.27
3.64
9.45
2
48.53
5.50
0.21
0.088
6.57
39.10
8438
(19626)
21
27
45
52
019
53
50
38
2
9
3
49.81
5.69
0.27
0.091
4.04
40.10
8674
(20175)
5
89.69
7.65
0.39
1.71
3
.05
.29
.02
.64
Avg.
48.49
5.62
0.21
0.10
5.30
40.28
8440
(19630)
0.030
0.53
17,550 (40703)
16,800 (39075)
17,220 (40052)
16,520 (38424)
Avg.
Ash-Free
51.20
5.93
0.22
0.11
42.53
8912
(20728)
Avg.
89.45
7.96
0.42
1.62
0.02
0.53
17,385 (40435)
16,660 (38749)
4-4
KVB72-806015-1308
-------
Table 4-2. The carbon content of both ash streams was included in the boiler
efficiency calculations.
4.3 MODIFICATIONS AND TESTS PERFORMED
The combustion modifications performed at Location 5 were the
following:
1. Excess air variation
2. Load variation
3. Overfire air variation
4. Auxiliary air damper adjustments
5. Combustion air preheat variation
Excess air variation was accomplished by adjusting the induced- and
forced-draft fan dampers to increase or decrease combustion air flow while
maintaining as constant a fuel flow as possible. Since the wood fuel composi-
tion was quite variable (it was not possible to control the fuel composition
during the tests), some difficulties were encountered in achieving steady
conditions. At times the fuel flow had to be adjusted to maintain the desired
steam flow, thus, oxygen levels tended to fluctuate by plus or minus approxi-
mately one percentage point. The average oxygen was varied from 5.0 percent
to 9.3 percent during these tests.
In the load variation series of tests, steam flow was varied from
baseline (approximately 70-80 percent of full capacity) to 90 percent of
capacity and then to approximately 50 percent of capacity. Two different
oxygen conditions were set up at both the high and the low steam flow rates.
In this test series and in other tests at Location 5, some load fluctuations
occurred. This is partly because of variability of the wood fuel, but also
because the boiler was designed as a load-following unit. The other boiler in
the powerhouse at Location 5 is a black liquor recovery boiler. In operation,
the recovery boiler is baseloaded and the hogged fuel boiler takes load
swings. Although the plant was quite cooperative in maintaining constant load
on the hogged fuel boiler during emissions tests, some unsteadiness was
4-5 KVB72-806015-1308
-------
TABLE 4-2. LOCATION 5 ASH ANALYSES
Test 5/2-1d
Bottom Ash
Moisture, %
Ash, %
Sulfur, %
Carbon, %
As Received
0.88
93.94
0.16
3.85
Dry Basis
0
94.77
0.16
3.88
Heat of Combustion (gross and net]
Btu/lb
kJ/kg
331
770
334
777
Fly Ash
As Received
Dry Basis
Moisture, %
Ash, %
Sulfur, %
Carbon, %
4.71
44.88
0.27
47.37
0
47.10
0.28
49.71
Heat of Combustion (gross and net]
Btu/lb
kJ/kg
7,047
16,390
7,395
17,200
KVB72-806015-1308
-------
unavoidable. In general, load could be held to within about ±5 percent of
some average value over the course of the day's tests, however, over several
days the load variations were on the order of +15 percent and more for a few
tests. In some cases, No. 6 oil had to be burned in order to keep the fire
steady and to maintain load.
Three values of overfire air (OFA) were evaluated: a baseline value
of 5.7 percent of total combustion air, a high of 9.7 percent, followed by
zero OFA. As shown in Figure 4-1, there were four elevations of overfire air,
each with four tangentially-oriented ports. At baseline conditions only the
fourth elevation (highest above grate) was used. At the high overfire air
condition the top two elevations were used.
The overfire air elevations were either on or off; no throttling of
the flow was possible. The lowest two elevations of overfire air are not used
by the plant because it was determined soon after the boiler installation that
stack opacity increased when they were used and efficiency was decreased. In
addition, when more than two elevations of overfire air were used, grate
temperatures became dangerously high due to the reduced undergrate air flow.
The auxiliary air dampers are part of the oil air system shown in
Figure 4-1. They are located near each of the four oil guns which are mounted
in the four corners above the overfire air ports. There are five air regis-
ters for each of the oil guns. In the baseline condition when no oil was
fired there was no air flow through these registers. In the modified condi-
tion the lowest of the five registers (auxiliary air damper "AA") was opened
10 percent and air was thus injected at approximately the level of the oil
guns without any oil flow. Thus, the adjustment provided an alternative to
overfire air as a means of staged combustion air. The other oil air registers
were also adjusted but did not appear to give NOX emissions as low as the
auxiliary air damper "AA" adjustment.
The combustion air temperature was varied from baseline level to a low
inlet air temperature by bypassing the steam coil portion of the air heater
for that particular test. The combustion air temperature dropped from 517°F
4-7 KVB72-806015-1308
-------
(543 K) to 477°F (521 K) at high load and from 495°F (531 K) to 467°F (515 K)
at baseline load. Thus, only modest changes in combustion air temperature
were achievable.
A complete set of emissions measurements including gaseous, total and
solid particulate and particulate size, wet chemical SOx, and polycyclic
organic matter (POM) was made for the baseline condition. The same series of
tests (except for wet chemical SO measurement) was run at the optimum low NOx
conditions which was a combination of low excess air and auxiliary air damper
adjustment. Nearly complete sets of gaseous O2, CO, C02, NO, NOX, S02, and HC
emissions were obtained on all tests. Appendix A provides a description of
the measurement equipment.
4.3.1 Results
The data from Location 5 is summarized in Tables 4-3 and 4-4. The
particulate measurements in Table 4-3 were obtained at the boiler outlet
(upstream of multiclone) while those in Table 4-4 were obtained downstream of
multiclone. The data yields two important observations:
1. NOX emissions were low with only a single reported concen-
tration over 100 ppm. >
2. Particulate emissions prior to any dust collecting device
were high and variable. The range of total particulate
concentrations measured at the boiler outlet was 1270 - 3780
ng/J (2.96 - 8.79 lb/106 Btu). The overall fly ash removal
efficiency based on one set of simultaneous particulate
measurements at the boiler outlet and at the stack was
84 percent.
The average NO emissions for the baseline and optimum low NO configurations
are shown in Table 4-5. The average reduction in NO concentration based on
those values is 17.2 percent.
Boiler efficiency increased at the low NO conditions to 72.3 percent
X
(average for Tests 5/5-2, 5/7-2b, and 5/7-2c) from the baseline efficiency of
71.1 percent (average for Tests 5/5-1, 5/7-1b, and 5/7-1c) for a gain of
1.7 percent. The auxiliary air damper "AA" adjustment thus gave reduced NO
emissions and the maximum efficiency condition at baseline load.
4-8 KVB72-806015-1308
-------
TABLE 4-3. SUMMARY OF GASEOUS AND PARTICULATE EMISSIONS, LOCATION 5 -
HOGGED FUEL BOILER
Steam Flow
Test No.
5/1-1
5/2-la
5/2-lb
5/2-lc
5/2-ld
5/2-le
5/2-lf
5/2-2
5/2-3
5/2-4
5/2-4a
5/3-1
5/3-2
5/3-2a
5/3-3
S/3-3a
5/4-1
5/4-2
5/4-3
5/4-4
5/4-2a
5/5-1
5/5-2
5/5-3
5/6-1
5/6-2
5/6-la
5/6-2a
5/7-1
5/7-2
5/7-la
5/7-2a
5/7-lb
5/7-2b
5/7-lc
5/7-2C
Date
1979
9-17
9-21
9-20
9-24
9-25
9-26
10-1
9-21
9-21
9-21
10-3
10-4
10-4
10-4
10-4
10-4
10-1
10-1
10-1
10-1
10-2
10-16
10-16
10-16
10-15
10-15
10-15
10-15
10-17
10-17
10-18
10-18
10-22
10-22
10-23
10-23
Kg/s
28.0
19.3
16.6
14.4
20.0
17.6
21.0
19.3
19.3
19.3
20.2
17.5
22.7
22.7
13.5
13.5
19.5
19.5
19.5
19.5
19.2
18.1
18.1
18.1
26.2
23.6
21.4
20.4
18.8
18.8
18.9
18.9
17.6
17.6
18.0
18.0
Flow
103
lb/hr
222
153
132
114
159
140
167
153
153
153
160
139
180
180
107
107
155
155
155
155
152
144
144
144
208
187
170
162
149
149
150
150
140
140
143
143
Heat Input
Rate
MW
78.0
53.9
46.3
40.2
56.0
49.2
58.6
53.9
53.9
53.9
56.3
48.9
63.3
63.3
37.5
37.5
54.5
54.5
54.5
54.5
53.3
50.7
50.7
50.7
73.3
65.6
65.6
56.9
52.5
52.5
52.8
52.8
49.2
49.2
50.4
50.4
1C6
Btu/hr
266
184
158
137
191
168
200
184
184
184
192
167
216
216
128
128
186
186
186
186
182
173
173
173
250
224
204
194
179
179
180
180
168
168
172
172
°2
*
8.3
7.2
8.7
9.6
6.9
7.1
5.4
6.5
9.3
5.7
5.0
7.4
4.8
7.7
10.3
7.0
7.3
8.1
7.4
8.1
7.7
8.3
5.8
7.4
6.5
7.1
7.4
8.0
6.9
5.1
8.1
5.1
8.7
5.6
7.6
6.2
%
12.7
12.0
10.0
9.0
12.4
13.6
14.5
13.6
10.2
12.9
15.0
12.6
15.9
12.4
8.9
12.9
12.8
10.3
11.7
11.0
12.3
11.8
14.8
12.8
10.9
9.9
12.1
12.9
13.5
15.8
10.6
16.2
10.6
14.4
12.3
13.9
N
ppm'
103
80
81
86
50
45
59
106
88
69
69
61
49
53
72
53
85
67
57
58
67
89
62
81
86
99
89
85
82
61
76
55
61
54
66
65
°x
ng/Jf
56
43
44
47
27
24
32
57
48
37
37
33
27
29
39
29
46
36
31
31
36
48
34
44
47
54
48
46
44
33
41
30
33
29
36
35
NO
ppm*
97
74
72
72
48
43
55
101
82
65
68
61
49
53
72
51
85
67
53
56
65
85
61
78
83
96
84
83
78
56
73
51
59
54
62
59
ng/J+
53
40
39
39
26
23
30
55
44
35
37
33
27
29
39
28
46
36
29
30
35
46
33
42
45
52
45
45
42
30
40
28
32
29
34
32
HC
ppm»
81
47
213
557
91
44
139
35
204
96
130
40
112
141
315
72
223
214
89
39
60
48
68
79
50
36
45
57
50
103
92
ng/J
15
9
40
105
17
a
26
7
38
18
24
8
21
27
59
14
42
40
17
17
11
9
13
15
9
7
8
11
9
19
17
CO
ppm*
929
1522
>2656
>3144
>2585
>2258
>2312
894
>3077
847
2172
2310
>2222
>2707
>3364
>2571
1130
865
1853
1298
1780
1312
>2153
2038
1460
700
1055
670
>1463
>1165
1988
2422
ng/J
306
502
>875
>1036
>852
>744
>762
295
>1014
279
716
761
>732
>892
>1109
>847
372
285
611
428
587
432
>710
672
481
231
348
221
>4S2
>384
655
798
so..
ppm*
44
13
30
20
42
97
3
0
0
3
39
8
0
0
0
0
0
0
0
0
47
ng/J
33
10
23
15
32
73
2
0
0
2
29
6
0
0
0
0
0
0
0
0
35
4-9
KVB72-806015-1308
-------
TABLE 4-3.
(CONTINUED)
Test No.
5/1-1
5/2-la
5/2-lb
5/2-lc
5/2-ld
5/2-le
5/2-lf
5/2-2
5/2-3
5/2-4
5/2-4a
5/3-1
5/3-2
5/3-2a
5/3-3
5/3- 3a
5/4-1
5/4-2
5/4-3
5/4-4
5/4-2a
5/5-1
5/5-2
5/5-3
5/6-1
5/6-2
5/6- la
5/6-2a
5/7-1
5/7-2
5/7-la
5/7-2a
5/7-Lb
5/7-2b
5/7-lc
5/7-2C
Total Solid
(Blr.Out.) (Blr.Out.)
Fuel Mix
Wood/Oil lb/106 lb/106
% of heat in Btu ng/J atu ng/J
75/25 8. 50 3660 9.37 3660
100/0
100/0 5.12 2200 5.07 2180
80/20 7.41 3190 7.27 3120
100/0
100/0 2.96 1270 2.87 1240
100/0
100/0
100/0
100/0
100/0 8.79 3780 8.77 3770
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0 3.82 1640 3.81 1640
100/0
100/0
100/0
85/15
85/15
100/0
100/0
100/0
100/0 5.64 2430 5.62 2420
100/0
100/0 3.50 1505 3.48 1500
100/0
100/0
91/9
91/9 5.83 2510 5.82 2500
Air Heater
Gas Out
Temperature
°F
463
465
455
441
471
460
453
465
470
470
462
465
480
460
420
440
465
440
455
445
454
428
433
430
454
405
394
440
450
440
440
440
430
453
440
443
K
513
514
508
500
517
511
507
514
516
516
512
514
522
511
489
500
514
500
508
503
508
493
496
494
508
480
474
500
505
500
500
500
494
507
500
501
Boiler
Eff.
»
79.2
68.6
67.5
75.6
68.8
69.8
70.5
69.6
67.1
69.0
70.4
69.1
70.3
69.2
67.7
69.8
69.1
68.2
68.8
68.6
69.2
69.8
71.2
70.4
75.6
73.0
67.6
70.0
68.9
70.4
74.7
75.4
Comments
As Found
Baseline
Baseline-Brink
Baseline-Multiclone efficiency-bad wood
Baseline-Brink and SO -bad wood
Ba^.Gline-POM-bad wood
Baseline-Brink
Low O -unsteady conditions
High 0,
Minimum O
Minimum 0 -Brink
Baseline
High Load-Low 0
High Load-High 0
Low Load-High O
Low Load-Low O.,
2
Baseline
High OFA
Zero OFA
Repeat Baseline
High OFA
Baseline
Aux.Air "AA" dpr. adj.
Oil Air "C" adj.
High Load
Low C.A. Temp. -High load
Low C.A. Temp.
Baseline
Baseline
Low NO -Brink
Baseline
Lew NO -Brink-POM
Basel ine
Low NO -Brink
Baseline
Low MO -Total part ic. removal efficiency
* dry, corrected to 3* 0 , dry
t as NO.
4-10
KVB72-806015-1308
-------
TABLE 4-4. ADDITIONAL EMISSIONS DATA, LOCATION 5
03
O
O1
O
Steam Flow
Test No. Date,
(Location) 1979 kg/a 103 Ib/hr
5/2 - Ic 9-24 14.4 114
(Multiclone
Outlet)
5/7 - 2c 10-23 18.0 143
(Stack)
5/2 - 1d 9-25 20.0 159
(Boiler
Outlet)
Wet Total Solid Particulate
Chemical SO, Particulate Particulate Oa „ ,
£ KG mO va J.
02 Efficiency
%, dry ppm* ng/J lb/106 Btu ng/J lb/106 Btu ng/J Percent Comments
9.6 — — 3.02 1300 2.85 1230 59.2 Multiclone efficiency only-
multiclone partially plugged
9.0 — -- 0.92 394 0.88 377 84^2 Total particulate collection
efficiency
7.0 18 13.5 — — — — -- Zero detectable SO3 -
Goksoyr-Ross method
*Dry, corrected to 3%
U)
O
oo
-------
TABLE 4-5. EMISSIONS OF NO AT BASELINE AND OPTIMUM LOW-NO CONDITIONS
Condition
Baseline
Low NO
Mean, Standard Deviation
No. of
ppm* ng/Jt Data Pts .
67.9, 13.5 36.8, 7.3 14
56.2, 4.0 30.4, 2.0 5
(aux. air dpr. "AA"
adjustment)
*dry, at 3 percent
tNO as NO0
The particulate emissions at the optimum low NO condition ranged from
1500 - 2420 ng/J (3.50 - 5.83 lb/106 Btu). It was assessed that this emission
level was not different from that found at baseline, which was 1270 - 3190
ng/J (2.96 - 7.41 lb/106 Btu).
However, when all of the total particulate data from all of the tests
conducted upstream of the multiclone are plotted as a function of NO emis-
sions, as is done in Figure 4-2, a trend of lower particulate emissions with
lower NO emissions is noted. The reason for this behavior is unclear at
present. Correlations were made of particulate emission with excess air level
and with the pressure drop across the grate, however, no meaningful result was
obtained with either variable. The wood fuel composition and the size of the
wood fuel may be the most important variables affecting the particulate
emissions from the hogged fuel boiler, and may influence NOV emissions as
X
well.
The Brink cascade impactor tests conducted at the boiler outlet were
analyzed to determine if the solid particulate size distribution was related
4-12
KVB72-806015-1 308
-------
-p
ffl
O
rH
\
£1
EH
O
E-i
3870
(9.0)
3440
(8.0)
3010
(7.0)
2580
(6.0)
2150
(5.0)
en
2
O
H
in
en
§ 1720
„ (4.0)
u
£ 1290
< (3.0)
860
(2.0)
430
ri.o)
50 60 70
NO, ppm, Dry at 3% 0,
80
90
100
Figure 4-2.
Boiler total particulate emissions as a function of NO
emissions for various test conditions on a hogged fuel
boiler.
4-13
KVB72-806015-1308
-------
to boiler operation. (Three of these distributions are shown in Figures 4-3
through 4-5.) These distributions, characterized by their mass median
particle size, are summarized below:
Test.
No.
5/2-4a
5/7-2b
5/2-1f
5/7-2
5/2-1b
5/7-2a
uperamng
Condition
Minimum ©2
Low NOv
X
Baseline
Low NO
X
Baseline
Low NO
iieuj-dii LJ
ym
200
250
600
38
250
22
lb/106 Btu
8.77
6.92
5.79
5.62
5.07
3.48
ng/J
3770
2980
2490
2420
2180
1500
These data indicate that there was no clear relationship between the
mass median diameter and either the particulate emission factor or the boiler
operating condition. (As will be discussed shortly, there is a relationship
between the particulate temperature factor and wood ash content.)
Although the size of the fuel was not characterized, it is anticipated
that for fuels of equal ash content, smaller pieces of fuel would be more
easily entrained in the flue gas after combustion and would lead to larger
particulate concentrations having smaller aerodynamic particle diameters.
Larger pieces of fuel would tend to burn on the grate and fall into the ash
pit after combustion and less fine material would be carried out of the boiler
in the flue gas. The boiler did not have the instrumentation necessary to
separately measure the amount, if any, of sander dust or wood shavings used
during a test.
While the total impactor catch weight was considerable, much of the
particulate material was large and, therefore, was caught in the precutter
cyclone before reaching the stages. Figures 4-3 through 4-5 show that only
20 - 40 percent of the total catch actually impacted on the stages. During
some of the tests the impactor jets became plugged and sampling had to be
discontinued prematurely. Thus, the sample weight on the stages was sometimes
less than ideal.
4-14 KVB72-806015-1308
-------
1
H
l_n
-J
NJ
1
co
o
0
I-J
Ln
1
f— '
U)
o
CD
-LU
8
6
4
6
3 3
o
Q
K 2
W
W
H
Q
a i
o
BO. 8
rf!
rl 0.6
o
H
1 0.4
Q
O
§0.3
0.2
0.1
J 1 1 1 1 1 1 1 1 1 1 1 | | | 1 I 1 1 | |l '
~ LOCATION #5 - HOGGED FUEL BOILER ~~
TEST #5/2-4A
FUEL - Hogged Wood
— LOAD - 160 x 10 Ibs/hr. _
O 0 - 5.0%
~~ 1 BRINK IMPACTOR ~~
/ BOILER OUTLET
— / TOTAL CATCH WEIGHT - 0.1560 g —
r~\ (includes cyclone catch)
1^
— / —
(i
W
/
/
_ / _
JL
_ O _
r _
/
— i —
1 -
o
— —
— —
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II II l
.01.05.1 .2.512 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 99.99
CUMULATIVE PROPORTION OF IMP ACTOR CATCH, PERCENT BY MASS
Figure 4-3. Aerodynamic particle diameter as a function of cumulative proportion
of impactor catch at low-O conditions in a hogged fuel boiler.
-------
10
i
h->
(Tv
to
I
30
o
rr>
o
h-1
01
i
H
U)
O
00
o
LT>
Q
v 2
K
Q
^0.8
o
H
P
1.4
JO.3
0.2
0.1
.JillI\T
1 T
1 I TT~I I \ I I TT
Tl L.
HOGGED FUEL BOILER
LOCATION #5 -
TEST #5/7-2
FUEL - Hogged Wood
LOAD - 149 x 10 Ibs/hr.
0 - 5.1%
BRINK IMPACTOR
BOILER OUTLET
TOTAL CATCH WEIGHT - 0.1177 g
(includes cyclone)
O
111 III
I
I I I I I I I
I I
I 1
I I
I
.01.05.1.2 .5
12 5 10 20 30 40 50 60 70 80 90 95 98 99
CUMULATIVE PROPORTION OF IMPACTOR CATCH, % BY MASS
99.8 99.9 99.99
Figure 4-4.
Aerodynamic particle diameter as a function of cumulative proportion
of impactor catch at low NO conditions in a hogged fuel boiler.
-------
i
{—>
-~j
tv>
I
oo
o
(T>
O
M
Ul
U)
o
00
10
8
6
o
LO
Q
. 2
K
£
H
Q
W 1
So.8
u
H
0.6
0.4
0.3
0.2
0.1
TT~I—TT~T \ 1 1—I I I I I—I I I TT
TT
LOCATION #5 - HOGGED FUEL BOILER
TEST #5/7-2A
FUEL-Hogged Wood
Load - 150 x 10 Ibs/hr.
BRINK IMPACTOR
BOILER OUTLET
TOTAL CATCH WEIGHT - 0.094 g
(includes cyclone)
1 1 1 1
1
1 1 1 1 1 1
1 1
11
.01.05.1 .2 .5
12 5 10 20 30 40 50 60 70 80 90 95 98
CUMULATIVE PROPORTION OF IMPACTOR CATCH, % BY MASS
99 99.8 99.9 99.99
Figure 4-5.
Aerodynamic particle diameter as a function of cumulative proportion
of impactor catch at low NO conditions in a hogged fuel boiler.
-------
The fuel composition data obtained show that there is some correlation
of total particulate emission with the ash content of the wood fuel. This
correlation is shown in Figure 4-6. Since a large portion of the fuel ash
(probably over 70 percent) is fly ash rather than bottom ash one would expect
such a correlation. Because of the spread in the data it was not possible to
separate the effects of different operating conditions on emissions. Thus,
the curve shown is an overall curve not specific to only one set of
conditions. The curve in Figure 4-6 was fit by li-near least square regression
analysis. A correlation coefficient approaching unity indicates a good fit
and a coefficient approaching zero indicates a poor fit. As noted, the corre-
lation in terms of wood ash content can explain 68.9 percent of the variation
in total particulate.
To determine whether or not there was a relationship between the
moisture content of the fuel and total particulate emission, the data were
plotted and a linear regression analysis performed. The result was a correla-
tion coefficient of only 10.1 percent indicating a weak relationship between
particulate emission and wood moisture content. The data indicates that,
under different boiler operating conditions, total particulate can vary by as
much as a factor of three for nearly the same fuel moisture content. Although
it is reasonable to suppose that the amount of moisture in the fuel affects
the burnout of the wood pieces, other factors appear to play a more signifi-
cant role in the combustion and carryover of the wood. More data would be
required at the various operating conditions to determine their true effects
on particualte emission.
Figure 4-7 was generated by linear regression of the NO emission
versus fuel moisture data. A somewhat better correlation was found (r2 =
26.3 percent), however, there is still a large amount of data scatter. Based
on previous tests of steam and water injection in boilers, it might be
expected that increased moisture content of the fuel would result in lowered
flame temperatures and reduced NO emission, however, this does not appear to
be the case for the hogged fuel boiler at Location 5.
In a similar manner, NO emission was evaluated as a function of fuel
nitrogen in an attempt to determine whether or not any correlation existed.
4-18 KVB72-806015-1308
-------
3
4-1
m
(U
4J
rO
r-l
3
0
•H
-P
r-l
03
4J
O
EH
10.0
(4300)
8.0
(3440)
6.0
(2580)
4.0
(1720)
2.0
(860)
5/2-4a
5/2-lc Q
QMin. 02
High OFA
Baseline
/\ Optimum Low NO
LOAD: 18.0 Kg
steam/s
(143,000 Ib
steam/hr)
y=l.983x+1.138 (Curve Fit-Linear Regression)
Correlation coefficient = r2 = 0.689
I
I
J_
0123456
Wood Ash Content (Wet Basis), %
Figure 4-6. Total particulate as a function of wood ash content for
a hogged fuel boiler.
4-19
KVB72-806015-1308
-------
80
70
60
CN
t*>
n
-P
fl3
50
I 40
ft
c
0
-H
m
to
•H
w
30
20
10
I I
High OFA
Baseline
/\ Optimum Low NOX
. 0
5/2-4a
5/4-2a
LOAD: 18.0 Kg steam/s
(143,000 Ib stearn/hr) 5/7-2c
o
10 20 30 40 50
Wood Moisture Content (Wet Basis), %
60
Figure 4-7. NO emission as a function of wood moisture content
for a hogged fuel boiler.
4-20
KVB72-806015-1308
-------
The data were too scattered to develop a meaningful curve. However, the data
indicate that there was no significant increase in NO emission as fuel nitro-
gen increased. (The weight percent of nitrogen in the fuel varied over a
narrow range from 0.08 percent to 0.14 percent on a wet basis.)
Very low SO2 emissions were measured either by wet chemical technique
or continuous gaseous analyzer. The SCU emission data were presented in
Table 4-2. There are two reasons for the low SG>2 emission:
1. Low fuel sulfur content (0.09 - 0.11 percent by weight on a
dry basis).
2. Ash sulfur retention. Analysis of the fly ash and bottom ash
yielded sulfur values (dry basis) of 0.28 percent and
0.16 percent, respectively. These values exceeded that for
the wood fuel indicating some degree of sulfur concentration
in these ash streams.
Changing fuel sulfur content and, no doubt, sulfur content of the fly ash and
bottom ash (sulfur in the fly ash and bottom ash was measured for only one
test) caused some variation in SO2 emission, however, the absolute level of
S02 was generally less than 50 ppm, dry at 3 percent 0,,. The error in the
continuous analyzer S0? data at this low level for these tests approached 20 -
30 percent due to problems with the analyzer.
Gaseous SO2 measurement was made for the first twenty-one tests with
the DuPont 400 Photometric Analyzer. A wet chemical SO test (Goksoyr-Ross
A
method) yielded 18 ppm (13.5 ng/J) of SO2 and no detectable SO-j. By compari-
son, the DuPont analyzer indicated 42 ppm of SO2 during the same test. The
DuPont analyzer was down for repair during the remainder of the test program
at Location 5.
The results of the excess air variation test series at three different
boiler loads are shown in Figure 4-8. The maximum NO reduction obtained by
reducing the excess air from baseline (7.2 percent oxygen) to 5.7 percent
oxygen was 12 percent. The efficiency at the reduced excess air condition was
slightly higher than that at baseline condition. CO emissions were also
reduced in two of the three lowered excess air tests.
4-21 KVB72-806015-1308
-------
*
m
a
a
o
2
140
120
100
80
60
40
20
1 I T
-L--L
1 I I I \ I
STEAM FLOW
kg/S (10 Ib/hr)
19'3
22.7 (180)
13.5 (107)
FUEL: Hogged Wood (100%)
O
5/2-2
5/3-3
5/3-2A
5/3-2
5/3-3A
56 7 8 9
STACK OXYGEN, % DRY
10
11 12 13 14
Figure 4-8. NO emissions as a function of stack oxygen for three
loads in a hogged fuel boiler.
4-22
KVB72-806015-1308
-------
Similar behavior of NO emission versus excess oxygen was found to
X
occur at full capacity and half capacity. NO emissions at both high and low
loads were less than emissions at baseline by approximately 30 percent. CO
emissions, however, were high for both full and half capacity tests. Unburned
hydrocarbons (as methane) are reported in Table 4-3; however, the data appear
to be quite variable.
NO emissions were reduced by increasing the overfire air in another
A.
series of modification tests. The reduction in emissions from baseline was
21 percent (from 85 ppm to 67 ppm) and CO emissions were reduced at the same
time from a baseline level of 1130 ppm to 865 ppm. However, the efficiency
was reduced by 1.3 percent at the high overfire air condition. These calcula-
tions are made by comparing the baseline value from test 5/4-1 to the high
overfire air test 5/4-2. (Test 5/4-2a was also conducted at high overfire air
conditions, but was done on a different day.)
Shutting off the overfire air produced no significant change in NO
emissions as is seen by comparing the emissions from test 5/4-3 to those of
5/4-4, however, CO emission increased considerably at zero overfire air. No
definite relationship of particulate emissions with overfire air variation can
be determined from the data.
Lowering the combustion air temperature by only a modest amount had
little impact on NO emissions, but increased CO emissions. The effect of the
A.
optimum low NO register adjustment (10 percent open auxiliary damper "AA",
X
others closed) on NO emissions is shown in Table 4-4. The effect of this
X
modification on efficiency has already been discussed. Other oil air regis-
ters were adjusted but failed to achieve the same NO reduction.
X
The effect of the auxiliary air damper "AA" adjustment coupled with
lowered excess air on NO emissions is shown in Figure 4-9.
4.4 POLYCYCLIC ORGANIC MATTER (POM) SAMPLING
POM sampling was conducted at Location 5 for baseline conditions and
for the optimum low NOx condition with the boiler operating at 70 - 75 percei
of capacity. One POM test, at the boiler outlet, was run at each condition.
4-23 KVB72-806015-1308
-------
160
140
120
cTioo
0\°
on
-P
rd
80
a
60
40
20
= BASELINE _
["I = LOW NO
FUEL: Hogged Wood (100%)
except for Tests
5/7-1C and 5/7-2C ~"
where 9% of heat
input came from
No. 6 Oil.
STEAM FLOW: 18.3 Kg/s
(145,000 Ib/hr)
5/5-1
5/5-2
5/7-2B
5/7-2
5/7-2A
10
STACK OXYGEN, % DRY
Figure 4-9.
NO emission as a function of stack oxygen in the
baseline and low- NO configurations.
4-24
KVB72-806015-1308
-------
The sampling system is a modified Method 5 sampling train developed by
Battelle Columbus Laboratories. A combination of conventional filtration with
collection of organic vapors by means of a high surface area polymeric
adsorbent (XAD-2) proved highly efficient for collection of all but the more
volatile organic species. The modified sampling system consists of the
standard EPA train with the adsorbent sampler (Figure 4-10) located between
the filter and the impingers. With this system filterable particulate can be
determined from the filter catch and the probe wash according to Method 5,
whereas the organic materials present can be determined from the analysis of
the filterable particulate and the adsorbent sampler catch. The impingers are
only used to protect the dry-gas meter, and their contents are discarded.
4.4.1 POM Emissions
Sample time was extended to two hours to provide a large enough sample
for analysis. Following the sampling period, the organic resin module was
sealed and returned to the laboratory for analysis. The sampling probe and
glassware were washed with a 50-50 mixture of methylene chloride and methanol
per laboratory instructions. The filter and wash were also sent to the
laboratory following weighing.
These samples were analyzed by capillary-El, GC-MS utilizing a 30M
SE-52 column with hydrogen as a carrier gas. All data were collected by
single ion monitoring (SIM) to improve selectivity and sensitivity.
The extractions from the samples received from KVB, reagent blanks,
calibration standards and extracts from spiked filters and XAD-2 cartridge
blanks were analyzed using GC/MS procedures.
After running all the samples, the areas of the peaks were determined
by use of a computer. For the calibration solutions, the peak area of the
standard compound is ratioed to the nearest internal standard peak. A least
square analysis is performed which provides a slope and intercept used to
quantitate the sample solutions. The intercept is taken to represent the
quantitation limit. A correlation coefficient is also calculated which gives
an indication of the linearity of the calibration. An acceptable curve has a
correlation coefficient of at least 0.990.
4-25 KVB72-806015-1308
-------
GLASS WATER
JACKET
8-MM GLASS
COOLING COIL
ADSORBENT
GLASS WOOL PLUG
RETAINING SPRING
28/12 BALL JOINT
FLOW DIRECTION
GLASS FRITTED
DISC
FRITTED STAINLESS STEEL DISC
15-WM SOLV-SEAL JOINT
Figure 4-10. Mark III adsorbent sampling system.
4-26 KVB72-806015-1308
-------
The correlations obtained were as follows:
Phenanthrene 0.997
Anthracene 0.997
Fluoroanthrene 0.998
Phyrene 0.998
Benzo(a)anthracene 0.996
Chrysene 0.997
Benzo(e)pyrene 0.997
Benzo(a)pyrene 0.997
Perylene 0.997
Indeno-pyrene 0.994
Benzo(g,h,i)perylene 0.994
Coronene 0.999
Using the computer determined areas for the sample peaks, these areas
were ratioed to that of the nearest internal standard. These ratios were then
inserted into the appropriate quantitation equation to determine the total
amount of the particular POM (expressed in yg) in the sample. Finally, the
values for the reagent blank were subtracted from the amount calculated to
yield the POM actually in the sample. The results are shown in Table 4-6.
•The results of the analyses are presented in yg per total sample. The
quantitative detection limit was 0.5 yg, thus samples with POM's present at
levels lower than this are reported as <0.5 yg (the standard deviation at
lower levels was prohibitively high for accurate quantitation). Samples
reporting POM values of ND (none detected) are at a level of less than 0.1 yg
(the approximate qualitative detection limit). The standard deviation on
points around 0.5 yg averaged around ±20 percent, at levels around 5 yg it
averaged around ±15 percent, and at levels above 12 yg the standard deviation
averaged around ±10 percent.
The POM analyses for the low NO condition are presented in the first
X
six columns of Table 4-6 for the XAD-2 module, the filter/probe wash/cyclone
and total. The corresponding data for the unmodified or baseline condition
are shown in the last six columns. The POM emissions in the low NO,, condition
A.
4-27 KVB72-806015-1308
-------
TABLE 4-6.
SUMMARY OF POM ANALYSES FOR LOCATION 5 - WOOD-FIRED SPREADER STOKER
POM
Phenanthrene
Anthracene
Methyl Anthracenes/
Phenanthrene a
Fluoranthene
Pyrene
Methyl Pyrene/Fluoranthene
Benzo (c)phenanthrene
Benz (a)anthracene
Chrysene
Methyl Chrysenes
Dimethylbenz anthracenes
Benzofluoranthenes
Benzo(a)pyrene
Benzo (e)pyrene
Perylene
Methylcholanthrenes
Indeno (1 ,2 ,d-cd)pyrene
Benzo (g,h,i)perylene
Dibenz anthrancenes
Dibenzpyrenes
Co rone ne
TOTAL **
Sample volume, m
Low NO Test
X
XAD-2
Pg
0.7
ND*
0.7
<0.5
ND
<0.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
<1.2
ND
ND
ND
ND
ND
pg/m3
.70
ND
.70
<.5
ND
<.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
<1.2
ND
ND
ND
ND
ND
1.4O.6
Probe Rinse,
Cyclone & Filter
Pg
2.5
<0.4
2.7
o.e
1.0
0.7
ND
ND
<0.5
ND
ND
ND
<0.6
0.6
ND
<1.2
ND
<1.2
ND
0.1
2.2
pg/m
2.5
<.4
2.7
0.8
1.0
.7
ND
ND
<.5
ND
ND
ND
<.6
.6
ND
<1.2
ND
<1.2
ND
.1
2.2
10.6<14.5
Total
pg
3.2
<.4
3.4
0.8<1.3
1.0
.7<1.2
ND
ND
<.5
ND
ND
ND
<.6
.6
ND
<2.4
ND
<1.2
ND
.1
2.2
pg/m
3.2
<.4
3.4
.8<1.3
1.0
.7<1.2
ND
ND
<.5
ND
ND
ND
<.6
.6
ND
<2.4
ND
<1.2
ND
.1
2.2
12.0<18.1
Baseline Test
XAD-2
P9
0.9
ND
1.0
ND
ND
<0.5
ND
ND
ND
ND
<0.5
<0.5
ND
ND
ND
<1.2
ND
ND
ND
ND
0.1
pg/m3
.7
ND
.a
ND
ND
<.4
ND
ND
ND
ND
<-4
<.4
ND
ND
ND
<-9
ND
ND
ND
ND
.1
1.6<3.7
Probe Rinse.
Cyclone & Filter
pg
1.6
ND
1.1
<0.5
<0.5
<0.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
<1.2
ND
ND
ND
ND
ND
pg/m
1.0
ND
.8
<.4
<.4
<.4
ND
ND
ND
ND
ND
ND
ND
HD
ND
<.9
ND
ND
ND
ND
ND
1.8<3.9
Total
pg
2.5
ND
2.1
<.5
<.5
<1.0
ND
ND
ND
ND
<.5
<.5
ND
ND
ND
<2.4
ND
ND
ND
ND
0.1
pg/»3
1.7
ND
1.6
<.4
<.4
<.8
ND
ND
ND
ND
<.4
<.4
ND
ND
ND
<1.8
ND
ND
ND
ND
.1
3.4<7.6
*ND
Not Detected, less than 0.1 yg
**Two totals are shown, e.g., 1.4<3.6 where 1.4 is total of all quantified amounts and
3.6 is total of quantified amounts plus all values indicated as <, which indicates
that a compound was observed but cannot be quantified at a value below the amount
shown.
-------
were two to three times higher than the emissions at baseline. This large
difference could be due more to the fuel change than to the combustion modifi-
cation although the trend of higher POM with lower NO has been observed
before (Ref. 1 ).
A comparison of previous POM measurements of coal- and oil-fired
boilers with the present data is provided in Table 4-7. The POM emissions
from the hogged fuel boiler in the low NO mode were higher than those from
X
the two coal-fired spreaker-stokers and the residual-oil fired unit tested on
an EPA thirty-day field test program (Refs. 1,2,3). The baseline POM
emissions, however, were approximately equal to those of one coal-fired
spreader-stoker (Ref. 1) and were slightly less than the oil-fired unit. In
both baseline and low NO conditions, the POM emissions from the hogged fuel
X
boiler at Location 5 were well below those of the pulverized coal unit tested
on the Thirty-Day program (Ref. 4).
The following conclusions were made based on the present study:
1 . POM was higher in the low NO mode than under baseline
conditions.
2. POM emission for this hogged fuel boiler was greater in the
low NO configuration than two coal-fired spreader stokers
and one residual-oil-fired boiler tested previously by KVB;
at baseline, the POM emission was on a par with that of one
coal-fired spreaker-stoker and slightly below that of the
oil-fired unit.
3. At baseline and in the low NO mode, POM emission was well
below that of a pulverized coal boiler previously tested.
4. The POM emissions are likely to be dependent on the type of
wood burned. The wood fuel composition could not be con-
trolled because of the nature of the wood sources.
Therefore, further study is required to determine the exact
extent of its influence on POM emissions.
4-29 KVB72-806015-1308
-------
TABLE 4-7. POM EMISSION FROM OIL-, COAL-, AND WOOD-FIRED
BOILERS: COMPARISON WITH PRESENT DATA
i
U)
O
30-Day
(Ref .
30-Day
(Ref.
30-Day
(Ref.
30-Day
(Ref.
Hogged
Location
Tests, Site 1
1)
Tests, Site 2
2)
Tests, Site 3
4)
Tests, Site 4
3)
Fuel Boiler
Boiler Capacity Control
and Type Fuel Technology
Spreader-Stoker Coal LEA
12.6Kg/s (100,000 Ib/h)
Residual-Oil-Fired No. 6 Oil SCA (BOOS)
lO.OKg/s (79,000 Ib/h)
Pulverized Coal-Fired Coal SCA + LEA
32.8Kg/s (260,000 Ib/h)
Spreader-Stoker Coal LEA
20.2Kg/s (160,000 Ib/h)
3
Total POM, yg/m
Low NO
Baseline x
1.6K10.8 4.17<10.9
7.59<11.4 4.57<9.68
66.9<70.8 64.57<68.21
1.24<2.52 1.035<2.76
Spreader-Stoker Wood LEA + Air Register 3.6<7.9 12.0<18.1
Location 5 (Present Data) 25.2Kg/s (200,000 Ib/h)
Adjustment
-------
SECTION 5.0
REFERENCES
1. Carter, W. A., and Buening, H. J., "Thirty-Day Field Tests of
Industrial Boilers, Final Report, Site 1 - Coal-Fired Spreader
Stoker," EPA-600/7-80-085a, April 1980.
2. Carter, W. A., and Tidona, R. J., "Thirty-Day Field Tests of
Industrial Boilers, Final Report, Site 2 - Residual-Oil-Fired Boiler,"
EPA-600/7-80-085b, April 1980.
3. Carter, W. A., and Hart, J. R., "Thirty-Day Field Tests of Industrial
Boilers, Final Report, Site 4 - Coal-Fired Spreader Stoker," EPA-
600/7-80-085d, April 1980.
4. Carter, W. A., and Buening, H. J., "Thirty-Day Field Tests of
Industrial Boilers, Final Report, Site 3 - Pulverized Coal-Fired
Boiler," EPA-600/7-80-085c, April 1980.
5-1 KVB72-806015-1308
-------
cr>
1
h- '
§
-j
to
I
-806015-:
To Obtain
g/Mcal
106 Btu
Btu
lb/106 Btu
ft
in.
ft2
ft3
Ib
Fahrenheit
Fahrenheit
psig
psia
iwg (39.2 °F)
106 Btu/hr
GJ/hr
*These conversions
The values given
From
ng/J
GJ
gm cal
ng/J
m
cm
2
m
3
m
kg
Celsius
Kelvin
Pa
Pa
Pa
MW
MW
depend
are for
SECTION 6.0
CONVERSION FACTORS
SI Units to Metric or English Units
Multiply By
0.004186
0.948
3.9685x10
0.00233
3.281
0.3937
10.764
35.314
2.205
-3
tp = 9/5 (tc) + 32
tp = 1.8tK - 460
P . = (P )(1.450xlO~
psig pa
P . = (P ) I
psia pa
P. = (P ) (4
iwg pa
3.413
3.60
To Obtain ppm
Multiply*
Concentration
at 3% O0 of
Wood
CO
HC
NO or NO
x
SO. or SO
2 x
Oil Fuel
CO
HC
NO or NO
x
SO,, or SO
2 x
Coal Fuel
CO
HC
NO or NO
x
SO or SO
in ng/J by
2.31
4.05
1.41
1.01
2.93
5.13
1.78
1.28
2.69
4.69
1.64
1.18
o
03
-------
English and Metric Units to SI Units
Multiply*
1
M
--J
NJ
I
CO
O
cn
o
M
Ln
I
M
OJ
O
00
To Obtain
ng/J
ng/J
GJ
m
cm
2
m
3
m
kg
Celsius
Kelvin
Pa
Pa
Pa
MW
MW
*mv.^<-'^ ^~~,.~
From
lb/106 Btu
g/Mcal
106 Btu
ft
in.
2
ft
3
ft
Ib
Fahrenheit
Fahrenheit
psig
psia
iwg (39.2 °F)
106 Btu/hr
GJ/hr
Multiply By
430
239
1.055
0.3048
2.54
0.0929
0.02832
0.4536
tc = 5/9 (tp - 32)
t = 5/9 (t - 32) + 273
K F
P = (P . + 14.7) (6.895x10 )
pa psig
P = (P . ) (6.895x10 )
pa psia
P = (P. ) (249.1)
pa iwg
0.293
0.278
i
To Obtain C
ng/J of in
Wood
CO
HC
NO or NO (as NO )
x 2
SO or SO
2 x
Oil Fuel
CO
HC
NO or NO (as NOJ
x 2
SO,, or SO
2 x
Coal Fuel
CO
HC
NO or NO (as NO0)
x 2
SO^ or SO
2 x
Toncentrat
ppm @ 3% i
0.432
0.247
0.710
0.988
0.341
0.195
0.561
0.780
0.372
0.213
0.611
0.850
The values given are for typical fuels.
-------
APPENDIX A-1.0
GASEOUS AND PARTICULATE EMISSIONS TEST METHODS
AND INSTRUMENTATION
All emission measurement instrumentation was carried in a 9.8 m x
2.4 m (8 x 42 ft) mobile laboratory trailer. A plan view of the trailer is
shown in Figure A-1. The gaseous species measurements were made with
analyzers located in the trailer.
The emission measurement instrumentation used was the following:
TABLE A-1. EMISSION MEASUREMENT INSTRUMENTATION
Species
Manufacturer
Measurement Method
Model
No.
Hydrocarbon
Carbon Monoxide
Oxygen
Carbon Dioxide
Nitrogen Oxides
Particulates
Sulfur Dioxide
Particle Sizing
Smoke Spot
Opacity
Sulfur Oxides
(sox)
Beckman Instruments
Beckman Instruments
Teledyne
Beckman Instruments
Thermo Electron Co.
Joy Manufacturing C
DuPont Instruments
Brink
Bacharach
Flame lonization
IR Spectrometer
Polarographic
IR Spectrometer
Chemiluminescent
EPA Method 5 Train
UV Spectrometer
Cascade Impactor
ASTM 2156-65
EPA Method 9
Goksoyr-Ross
402
865
326A
864
10A
EPA
400
BMS1 1
RCC
A-1 .1
GAS SAMPLING AND CONDITIONING SYSTEM
A flow schematic of the flue gas sampling and analyzing system is
shown in Figure A-2. The sampling system uses three positive-displacement
diaphragm pumps to continuously draw flue gas from the stack into the
A-1
KVB72-806015-1308
-------
Calibration Gas
Bottles
Door and Stairs
Spare Calibration
Gas Bottles
Air
Conditioning/
Heater
oooo
f~\ Sample Handling/
^ Conditioning
O Room
-— TT-
. . Counter Top/
|Sirik~[ Cabinets
Instrument
Console
Air Conditioning/Heating Duct and Ven
J J 11 J
Counter Top/Cabinets
ts
I
•Over
Lv ^» •
••» ••> «^ •
Fume
Hood
OOOO
. ErfiMfi. S-ta^Aifi.
/Storage
Room
2.4m x 12.8m Double Axle
Semi Trailer
to
i
oo
o
CTi
o
H
Ul
Figure A-l. Instrumentation trailer floor plan.
o
CO
-------
I
oo
o
cr>
o
M
ui
i
Hot
Sample Dry Sample Lines
Line (Typical Set-Up Six Lines):
o
Hot Pump
Pressure
Hot Pump
Vacuum
HOTBOX
Hi-Span
Ini-opan
(Dry Sample)
Filters (6)
(7 micrometers)
Condenser
16
Hot/Cold
Switch
| Manifold )
1
Refrigeration Condenser
Q-Osample Pressure
span
OJ
o
oo
Figure A-2. Flue gas sampling and analyzing system.
-------
laboratory. The sample pumps pull from six unheated sample lines. Selector
valves allow composites of up to six points to be sampled at one time. The
probes are connected to the sample pumps with 0.95 cm (3/8") or 0.64 cm (1/4")
nylon line. The positive displacement diaphragm sample pumps provide unheated
sample gas to the refrigerated condenser (to reduce the dew point to 35°F), a
rotameter with flow control valve, and to the G>2, NO, CO, and C02
instrumentation. Flow to the individual analyzers is measured and controlled
with rotameters and flow control valves. Excess sample is vented to the
atmosphere.
To obtain a representative sample for the analysis of NO2, S02 and
hydrocarbons, the sample must be kept above its dew point, since heavy hydro-
carbons may be condensible, and S02 and NO2 are quite soluble in water. For
this reason, a separate electrically-heated sample line is used to bring the
sample into the mobile laboratory for analysis. The sample line is 0.95 cm
(3/8-inch) Teflon line, electrically traced and thermally insulated to main-
tain a sample temperature of up to 478 K (400°F). A heated diaphragm pump
provides hot sample gas to the hydrocarbon, SO2 and NOX analyzers.
A-1.2 INSTRUMENTAL CONTINUOUS MEASUREMENTS
The laboratory trailer is equipped with analytical instruments to
continuously measure concentrations of NO, NO_, CO, CO_, 0~, S0_, and hydro-
carbons. All of the continuous monitoring instruments and sample handling
system are mounted in the self-contained mobile laboratory. The entire system
requires only connection to on-site water, power, and sampling lines to become
fully operational. The instruments themselves are shock mounted on a metal
console panel. The sample flow control measurement, and selection, together
with instrument calibration are all performed from the console face. Three-
pen recorders provide a continuous permanent record of the data taken. The
sample gas is delivered to the analyzers at the proper condition and flow rate
through the sampling and conditioning system described in the previous
section. The sections below describe the analytical instrumentation.
A-4 KVB72-806015-1308
-------
A-1.2.1 Nitric Oxide (NO) and Total Nitrogen Oxides (NO )
X
Both the total nitrogen oxides (NO ) and nitric oxide (NO) concen-
A.
trations are measured from a sample gas obtained using a heated sample line at
394 K (250°F). In addition, the nitric oxide concentrations are measured
sequentially from samples obtained using the unheated sample line that is
connected to the same analyzer in the laboratory trailer. In the latter case,
water is first removed from the sample gas by a refrigeration unit. The
analytical instrument that is used for all of these measurements is the Thermo
Electron Model 1OA chemiluminescent gas analyzer.
For NO analyses, the sample gas is passed directly into the reaction
chamber where a surplus of ozone is maintained. The reaction between the NO
and the ozone produces light energy proportional to the NO concentration which
is detected with a photomultiplier and converted to an electrical signal. Air
for the ozonator is drawn from ambient through an air dryer and a 10-micro-
meter filter element.
The chemiluminescent reaction with ozone is specific to NO. To
detect NO2, a thermal converter has been designed to dissociate the NO2 to NO
by the bi-molecular reaction: 2 NO2 ->• 2 NO + O-. A model 700 thermal
converter is used in conjunction with the chemiluminescent gas analyzer as
shown in Figure A-3. The converter is a coil of resistance-heated stainless
steel tubing whose purpose is to drive the N02/N0 ratio to its chemical
equilibrium value at the converter temperature and pressure. The unit is
designed to operate at a temperature of 923 K (1200 °F) and pressure of .3 kPa
(10 torr). For these conditions and typical stack gas 02 concentrations, the
equilibrium N02 concentration is 0.2 percent of the total NOx concentration.
Therefore, when a gas sample containing any N02 is passed through the con-
verter, essentially all the NO? would be converted to NO. The resulting total
NO is then measured using the chemiluminescent analyzer and the difference
between the actual NO and the "total NO" would be the sample N02
concentration. The "total NO" is interpreted as NO .
.A
A_5 KVB72-806015-1308
-------
Capillary
0.020cm x 3.8cm
Optical
.Filter
Photo-
multiplier
Oxygen
Regulator
Analyzer
-J
Reference I
Sample V. )
Regulator"-]—
Reaction
Chamber
Sample
Gage
I
Flow
Meter
Capillary
0.013cm
x
2.9cm
Inlet
Outlet
Capil-
-lary
3.051cm
x
3.8cm
Oxygen
or
Dry Air
unit
NO
1
Anplifier
r-
I Me
-. -«._ J_ i
i
HV
5uppl\
— o
Low p
>ter p
i 1—
r
— '
1
fciS
Plugged
si—6—6—1
Room
Air
Water
Trap
—i Trap
^T~
To
Bypass
Pump
0.013cm 0.051cm i
Capillaries i
Model 700 !
Heated
Sample
Line
Figure A-3.
Schematic of NO /NO chemiluminescent analysis system.
X
A-6
KVB72-806015-1308
-------
N0~ may react upon contact with H^O (liquid phase) to form HNO,
(nitric acid). Under field test conditions, the exhaust gas may contain
significant H20 (depending upon the process and the ambient meteorological
conditions), and it is necessary to convert the N02 to NO before the H20 is
allowed to condense in the sampling system. By using the heated sample line
and the Thermo Electron Model 700 heated NOX module, NO2 concentrations will
effectively be measured. In reference to Figure A-3, the sample is maintained
above the H20 dew point up to and through the 127 ym (0.005 in.) capillary in
the heated module. Downstream of this capillary, the flow network is main-
tained at 1.3 kPa (10 torr), where the partial pressure of the H_0 in the
sample is sufficiently low to prevent any condensation at ambient temperature.
When using the heated system, NO, NO0, and NO are measured on a wet
£ X
basis. When not using the heated system, a condenser is placed upstream of
the analyzer and NO is measured on a dry basis.
Specifications
Accuracy: 1% of full scale
Span stability: ± 1% of full scale in 24 hours
Zero stability: ± 1 ppm in 24 hours
Power requirements: 115 ± 10V, 60 Hz, 1000 watts
Response: 90% of full scale in 1 sec (NOX mode);
0.7 sec (NO mode)
Output: 4-20 ma
Sensitivity: 0.5 ppm
Linearity: ± 1% of full scale
Vacuum detector operation
Range: 2.5, 10, 25, 100, 250, 1000, 2500, 10,000 ppm full scale
A_7 KVB72-806015-1308
-------
A-1.2.2 Carbon Monoxide and Carbon Dioxide (CO and CO-)
Carbon monoxide and carbon dioxide concentrations are measured using
Beckman Model 864 and 865 short-path-length nondispersive infrared analyzers
(see Figure A-4). These instruments measure the differential in infrared
energy absorbed from energy beams passed through a reference cell (containing
a gas selected to have minimal absorption of infrared energy in the wavelength
absorbed by the gas component of interest) and a sample cell through which the
sample gas flows continuously. The differential absorption appears as a
reading on a scale of 0% to 100% and is then related to the concentration of
the species of interest by calibration curves supplied with the instrument. A
linearizer is supplied with each analyzer to provide a linear output over the
range of interest. The operating ranges for the CO analyzer are 0-100 and 0-
2000 ppm, while the ranges for the C02 analyzer are 0-5% and 0-20%.
Specifications
Span stability: ± 1% of full scale in 24 hours
Zero stability: ± 1 ppm in 24 hours
Ambient temperature range: 273 to 322 K (32°F to 120°F)
Line voltage: 115 ± 15V rms
Response: 90% of full scale in 0.5 or 2.5 sec
Linearity: Linearizer board installed for one range
Precision: ± 1% of full scale
Output: 4-20 ma
A-1.2.3 Oxygen (O2)
A Teledyne Model 326A oxygen analyzer is used to automatically and
continuously measure the oxygen content of the flue gas sample. The analyzer
utilizes a micro-fuel cell which is specific for oxygen, has an absolute zero,
and produces a linear output from zero through 25 percent oxygen. The micro-
fuel cell is a sealed electrochemical transducer with no electrolyte to change
or electrodes to clean. Oxygen in the flue gas diffuses through a Teflon
membrane and is reduced on the surface of the cathode. A corresponding
A-8 KVB72-806015-1308
-------
Reference
Cell
Detector
Infrared Source
- Sample in from source
Sample
Cell
mm
*• Sample out
Diaphragm
Distended
Absorbs infrared
energy in region of
interest
Other molecules
Control
Unit
Figure A-4. Schematic of NDIR analyzer.
A-9
KVB72-806015-1308
-------
oxidation occurs at the anode internally and an electric current is produced
that is proportional to the concentration of oxygen. This current is measured
and conditioned by the instrument's electronic circuitry to give an output in
percent CU by volume for operating ranges of 0% to 5%, 0% to 10%, and 0%
to 25%.
Specifications
Precision: ± 1% of full scale
Response: 90% in less than 40 sec
Sensitivity: 1% of low range
Linearity: ± 1% of full scale
Ambient temperature range: 273 K to 325 K (32°F to 125°F)
Fuel cell life expectancy: 40,000% +-hrs
Power requirement: 115 VAC, 50-60 Hz, 100 watts
Output: 4-20 ma
A-1.2.4 Total Hydrocarbons (HC)
Hydrocarbon emissions are measured using a Beckman Model 402 high-
temperature hydrocarbon analyzer. The analyzer utilizes the flame ionization
method -of detection which is a proven technique for a wide range of concen-
trations (0.1 to 120,000 ppm). A flow schematic of the analyzer is presented
in Figure A-5. The sensor is a burner where a regulated flow of sample gas
passes through a flame sustained by regulated flows of air and a premixed
hydrogen/nitrogen fuel gas. Within the flame the hydrocarbon components of
the sample stream undergo a complex ionization that produces electrons and
positive ions. Polarized electrodes collect these ions, causing current to
flow through electronic measuring circuitry. Current flow is proportional to
the rate at which carbon atoms enter the burner.
The analysis occurs in a temperature-controlled oven. The sample is
extracted from the stack with a stainless steel probe which has been thermally
treated and purged to eliminate any hydrocarbons existing in the probe
itself. An insulated heat-traced teflon line is used to transfer the sample
A-10 KVB72-806015-1308
-------
Figure A-5. Flow schematic of hydrocarbon analyzer (FID).
A-ll
KV372-806015-1308
-------
to the analyzer. The entire heated network is maintained at a temperature
sufficient to prevent condensation of heavier hydrocarbons.
The flame ionization detector is calibrated with methane, and the
total hydrocarbon concentration is reported as the methane equivalent. FID's
do not respond equally to all hydrocarbons but generally provide a measure of
the carbon-hydrogen bonds present in the molecule. The FID does not detect
pure carbon or hydrogen.
Specifications
Full-scale sensitivity: adjustable from 5 ppm CH^ to 10% CH^
Ranges: Range multiplier switch has 8 positions: X1, X5, X10, X50,
X100, X500, X1000, and X5000. In addition, span control
provides continuously variable adjustment within a dynamic
range of 10:1
Response time: 90% full scale in 0.5 sec
Precision: ± 1% of full scale
Electronic stability: ± 1% of full scale per 24 hours with ambient
temperature change of less than 5.6 K (10°F)
Reproducibility: ± 1% of full scale for successive identical samples
Analysis temperature: ambient
Ambient temperature: 273 K to 317 K (32°F to 100°F)
Output: 4-20 ma
Air requirements: 250 to 400 cc/min of clean, hydrocarbon-free
air, supplied at 2.07 x 105 to 1.38 x 106
n/m2 (30 to 200 psig)
Fuel gas requirements: 75 to 80 cc/min of fuel consisting of
100 percent hydrogen supplied at 2.07 x 1O5
to 1.38 x 106 n/m2 (30 to 200 psig)
Electric power requirements: 120 V, 60 Hz
Automatic flame indication and fuel shut-off valve
A-12 KVB72-806015-1308
-------
A-1.2.5 Sulfur Dioxide
A Dupont Model 400 photometric analyzer is used for measuring SO-.
This analyzer measures the difference in absorption of two distinct
wavelengths (ultraviolet) by the sample. The radiation from a selected light
source passes through the sample and then into the photometer unit where the
radiation is split by a semi-transparent mirror into two beams. One beam is
directed to a phototube through a filter which removes all wavelengths except
the "measuring" wavelength, which is strongly absorbed by the constituent in
the sample. A second beam falls on a reference phototube, after passing
through an optical filter which transmits only the "reference" wavelength.
The latter is absorbed only weakly, or not at all, by the constituent in the
sample cell. The phototubes translate these intensities to proportional
electric currents in the amplifier. In the amplifier, full correction is made
for the logarithmic relationships between the ratio of the intensities and
concentration or thickness (in accordance with Beer's Law). The output is
therefore linearly proportional, at all times, to the concentration and thick-
ness of the sample. The instrument has a lower detection limit of 2 ppm and
full scale ranges of 0-200 and 0-2000 ppm.
Specifications
Noise: Less than 1/4%
Drift: Less than 1% full scale in 24 hours
Accuracy: (± 1% of analyzer reading) + (+ 1/4% of full scale range)
Sample cell: 304 stainless steel, quartz windows
Flow rate: 0.05 dm3/s (6 cfh)
Light source: Either mercury vapor, tungsten, or "Osram"
discharge type lamps
Power rating: 500 watts maximum, 115V, 60 Hz
A-13 KVB72-806015-1308
-------
Reproducibility: 1/4% of scale
Electronic response: 90% in 1 sec
Sample temperature: 378 K (220°F)
Output: 4-20 ma d.c.
A-1.3 PARTICULATE MATTER TOTAL MASS CONCENTRATION
Particulate matter is collected by filtration and wet impingement in
accordance with US-EPA Method No. 5. Nomograph techniques are utilized to
select the proper nozzle size and to set the isokinetic flow rates.
Gas samples for particulate sampling can be taken from the same
sample port as those for gas analysis and passed through the Joy Manufacturing
Company Portable Effluent Sampler. This system, which meets the EPA design
specifications for Test Method 5, Determination of Particulate Emissions from
Stationary Sources (Federal Register, Volume 36, No. 27, page 24888,
December 24, 1971, and revisions thereof) is used to perform both the initial
velocity traverse and the particulate sample collection.
Dry particulates are collected in the heated case that may contain a
cyclone to separate particles larger than 5 ym and a 110-mm glass-fiber filter
to retain particles as small as 0.3 )am. Condensible particulates are col-
lected in four Greenburg-Smith impingers immersed in a chilled water bath.
The sampling probe is positioned through an exhaust port and attached
to the sampling box. The probe consists of a sampling nozzle, heated probe,
gaseous probe, thermocouple, and pitot tube. The ball joint from the heated
probe connects to the cyclone and glass filter holder assembly. These assemb-
lies are positioned in the heated sampling box which is maintained at 394 K
(250°F), in order to eliminate condensation. The sample then passes from the
heated section to four Greenburg-Smith impingers immersed in an ice bath.
Only the second impinger has the original tip, the other three have had the
tip removed to decrease the pressure drop through them. The first and second
impingers are filled with 150 milliliters of distilled/deionized water. The
third impinger is left dry. The fourth impinger is filled with approximately
200 grams of indicating silica gel to remove entrained water. The use of
A-14 KVB72-806015-1308
-------
silica gel assures that a dry sample is delivered to the meter box. After
sampling, the spent silica gel is discarded and not used for any further
analysis.
An umbilical cord connects the last impinger, the pitot tube, and the
heating elements to the meter box which is located in a convenient place
within 15 m of the sampling ports. The meter box contains a vacuum pump,
regulating valves, instantaneous and integrating flow meters, pitot tube
manometers, vacuum gauge, and electrical controls.
Particulate matter (solids and condensibles) is collected in three
discrete portions by the sampling train: the probe and glassware upstream of
the filter; the filter; and the wet impingers. The probe and glassware are
brushed and rinsed with acetone; the matter is captured for gravimetric analy-
sis. The probe and glassware are then rinsed with distilled water and the
rinsings transferred to a second container for analysis. The filter is desic-
cated and analyzed gravimetrically. The combined impinger liquid is heated to
drive off uncombined water and the residue retained for analysis. The parti-
culate matter analysis is illustrated schematically in Figure A-6.
US EPA Method 5 considers the particulate matter captured in con-
tainers (1) and (3); the filter, probe brushing, and probe acetone rinse. As
EPA source standards are based on solid particulates only, care is taken to
differentiate between solid and the total (including condensible) particu-
lates. The water wash is performed because KVB's test experience has shown
that a significant amount of water-soluble material may sometimes be captured
by the probe.
The dry sample volume is determined with a dry test meter at a mea-
sured temperature and pressure and then converted to standard conditions. The
volume of condensed water in the impingers is measured in miHiliters and the
corresponding volume of water vapor is then computed at standard conditions.
The dry sample volume and water vapor volume are then summed to give the total
sample volume. The dry sample volume is used in the data reduction
procedures.
A_15 KVB72-806015-1308
-------
PARTICUUTE MATTER MASS DETERMINATION
sampling
train
component
particulate
matter
transfer
procedure
container
processing
analysis
result
Probe Cyclone
Y y y
Brushing
Acetone
Rinse
Distilled
Water
Rinse
\_/ i
Filter
i
t
Impingers
Distilled
Water
Rinse
Bake at 215"F to drive off uncombined 1^0 and Acetone
K
V
Gravimetric to 0.1 milligrams
I
Dig
Samples stored for Compositional Analysis
Figure A-6. Processing and analyzing particulate matter.
A-16
KVB72-806015-1308
-------
A point of interest is the method chosen to calculate particulate
emissions in ng/J or lb/10^ Btu from the experimental data. The particulate
sampling train, properly operated, yields particulate mass per unit flue gas
volume. Having measured g/m , it is necessary to establish the flue gas
volume per unit heat input if emissions in ng/J are desired. The original
Method 5 involved determining a velocity traverse of the stack, the cross-
sectional area, the flue flow rate, and fuel heating value. A revised and
more accurate method has been promulgated by the Environmental Protection
Agency that utilizes a fuel analysis (carbon content, hydrogen content, high
heating value, etc.) and the measured 02 in the exhaust to calculate the gas
volume generated in liberating 1.055 GJ (a million Btu's). The velocity
traverse approach generally results in a 20 to 30 percent higher value and is
believed to be less accurate.
KVB72-806015-1308
\~ -L /
-------
A-1.4 SMOKE SPOT
On combustion equipment where smoke numbers normally are taken,
such as oil-fired boilers, KVB, Inc. determines the smoke number using
test procedures according to ASTM Designation: D 2156-65. The smoke
number is determined at each combustion modification setting of the
unit. Examples are baseline, minimum excess air, low load, etc., and
whenever a particulate concentration is measured.
Smoke spots are obtained by pulling a fixed volume of flue gas
through a fixed area of a standard filter paper. The color (or shade) of
the spots that are produced is visually matched with a standard scale.
The result is a "Smoke Number" which is used to characterize the density
of smoke in the flue gas.
The sampling device is a hand pump similar to the one shown
in Figure A-7 . It is a commercially available item that can pass 36,900
+_ 1650 cu cm of gas at 289K and 1 atmosphere pressure through an enclosed
filter paper for each 6.5 sq cm effective surface area of the filter
paper.
Sampling Tube
T-' ..
T '*—-.- . (
Handle'
Figure A-7. Field-service-type smoke tester.
The smoke spot sampler is provided with a motor-driven
actuator to ensure a constant sampling rate independent of variations
in stroke rate that can occur when the sampler is operated manually.
The smoke scale required consists of a series of ten spots numbered
consecutively from 0 to 9, and ranging in equal photometric steps from white
through neutral shades of gray to black. The spots are imprinted or other-
wise processed on white paper or plastic stock having an absolute surface
A-18 KVB72-806015-1308
-------
reflectance of between 82.5 and 87.5%, determined photometrically. The smoke
scale spot number is defined as the reduction (due to smoke) in the amount of
light reflected by a soiled spot on the filter divided by 10.
Thus the first spot, which is the color of the unimprinted scale, is
No. 0. In this case there is no reduction in reflected incident light directed
on the spot. The last spot, however, is very dark, reflecting only 10% of the
incident light directed thereon. The reduction in reflected incident light
is 90%, and this spot is identified as No. 9. Intermediate spot numbers are
similarly established. Limits of permissible reflectance variation of any
smoke scale spot will not exceed +_ 3% relative reflectance.
The test filter paper is made from white filter paper stock having
absolute surface reflectance of 82.5 to 87.5%, as determined by photometric
measurement. When making this reflectance measurement, the filter paper is
backed by a white surface having absolute surface reflectance of not less
than 75%.
When clean air at standard conditions is drawn through clean filter
paper at a flow rate of 47.6 cu cm per sec per sq cm effective surface area
of the filter paper, the pressure drop across the filter paper falls between
the limits of 1.7 and 8.5 kPa (1.3 and 6.4 cm of mercury).
The sampling procedure is exactly that specified in D 2156. A clean,
dry, sampling pump is used. It is warmed to room temperature to prevent
condensation on the filter paper. When taking smoke measurements in the
flue pipe, the intake end of the sampling probe is placed at the center line
of the flue. When drawing the sample, the pressure in the flue gas stream
and the sampler is allowed to equalize after each stroke.
The smoke density is reported on the Mobile Lab Data Sheet as the Smoke
Spot Number on the standard scale most closely corresponding to test spot.
Differences between two standard Smoke Spot Numbers are interpolated to
the nearest half number. Smoke Spot Numbers higher than 9 are reported
as "Greater than No. 9."
This procedure is deemed to be reproducible to within +_ 1/2 of a
Smoke Spot Number under normal conditions where no oily stain is deposited
on the disk.
A~19 KVB72-806015-1308
-------
KVB's field experience with industrial boilers has been that the
human factor involved in the interpretation of the smoke spot by an experi-
enced observer does not cause a significant lack of precision.
A-1.5 OPACITY
Opacity readings are taken by a field crew member who is a certificated
graduate of a U.S. Environmental Protection Agency approved "Smoke School".
Observations are made at the same time that particulate measurements are
made and as often in addition as deemed necessary to gather the maximum
amount of information. The procedures set forth in EPA Method 9, "Visual
Determinations of the Opacity of Emissions for Stationary Sources" are
followed.
Observations are Imade and recorded at 15 second intervals while
particulate concentration is being measured and after the unit has stabilized
at other times. Before beginning observations, the observer determines that
the feedstock or fuel is the same as that from which the sample was taken
for the fuel analysis.
Before beginning opacity observations, the observer makes arrangements
with the combustion unit operator to obtain the necessary process data for the
standard KVB Control Room Data Sheet. The control room data are recorded for
the entire period of observations, as is customarily done by KVB during an
emissions test. The process unit data that are obtained include:
a. Production rates
1. maximum rated capacity
2. actual operating rate during test
b. Control device data
1. recent maintenance history
2. cleaning mechanism and cycle information
The observer requests the appropriate plant personnel to
briefly review and comment on the opacity measurements and process
data and the observer comments on:
A-20 KVB72-806015-1308
-------
a. the basis for choosing the observation periods used.
b. why it is believed the periods chosen constitute periods
of greatest opacity.
c. why the observations span a time period sufficient to
characterize the opacity.
Consideration is given to postponing the EPA Method 5 particulate
tests during periods of cloudy or rainy weather because of the inability
of the observer to monitor the smoke.
A-l.5.1 Sulfur Oxides (SO )
x
Goksoyr-Ross Method—Wet Chemical Method
The Goksoyr-Ross Controlled Condensate (G/R) method is used for the
wet chemical SO /SO determination. It is a desirable method because of its
fc «J
simplicity and clean separation of particulate matter, S0_ and ¥. SO (SO ) .
This procedure is based on the separation of H_SO. (SO ) from S0_ by cooling
the gas stream below the dew point of H_SO but above the HO dew point.
£, f± £
Figure A-8 illustrates schematically the G/R test system.
Particulate matter is first removed from exhaust gas stream by means
of a quartz glass filter placed in the heated glass filter holder. Tissue-
quartz filters are recommended because of their proven inertness to H2S04.
The filter system is heated by a heating tape so that the gas out temperature
of 533 K (500°F) is maintained. This temperature is imperative to ensure that
none of the H^SO, will condense in the filter holder or on the filter.
The condensation coil where the H^SO. is collected is cooled by water
which is maintained at 333 K (140°F) by a heater/recirculator. This tempera-
ture is adequate to reduce the exhaust gas to below the dew point of H_SO..
A-21 KVB72-806015-1308
-------
Adapter for Connecting Hose
Asbestos Cloth
Insulation
Glass-Cloth Heating
Mantle ^~^-
Stack
Gas Flow
Rubber
Vacuum
Hose
Recirculator
'hermometer
Styrofoam Ice Chest
Vacuum
Gauge
Dry Test
-way
Valve
Drierite
Figure A-8. Schematic of Goksoyr-Ross controlled condensation system (CCS).
A-22
KVB72-806015-1308
-------
Three impingers are shown in Figure A-8. The first impinger is
filled with 3 percent H202 to absorb SO2. The second impinger is to remove
carryover moisture and the third contains Drierite and a thermometer to mea-
sure the exhaust gas temperature to the dry gas meter and pump. The sampling
rate is 0.04 dm3/s (0.08 cfm).
A-1.5.2 Analysis Procedure
For both SO2 and H2S04 determination, the analytical procedure is
identical. The H2SO, sample is washed from the back part of the filter holder
and the coil using a 5 percent isopropyl alcohol solution. The sample from
the first impinger which is assumed to be absorbed and reacted S02 in the form
of H2SO^ is recovered with distilled water washing. The amount of H2S04 in
the condensate from the coil and from the H202 impinger is measured by H+
titration. Bromphenol Blue is used with NaOH as the titrant.
A-1.6 PARTICLE SIZE DISTRIBUTION
Particle size distribution was determined using a Brink Model BMS-11
cascade impactor (Figure A-9).
A-1.6.1 Design
The Brink sampling probe is a modular cascade impactor apparatus
suitable for sampling dust in a wide range of flue gas conditions. The pri-
mary components are a cyclone separator, a cascade impactor, an absolute
filter, and a critical orifice. The cascade impactor comes with sampling tips
of various sizes and is constructed of 316 stainless steel. The impactor
consists of a number of stages arranged in series. From one to five stages
can be used. Each impactor stage has an orifice and a collection cup. These
are shown in Figure A-9. The orifice diameter and the distance between the
orifice and the cup determine the particle collection characteristics of the
stage. These dimensions are listed in the table in Figure A-9. An absolute
filter follows the final impactor stage. This back-up filter was a Gelman
Type A Glass Fiber Filter. Special collection substrates (e.g., glass fiber,
aluminum foil, etc.) placed on the collection plates can be used. The stack
A-23 KVB72-806015-1308
-------
DIMENSIONS OF CASCADE IMPACTOR JETS
Dimensions, Cm
Spacing of
Jet Opening*
0.747
0.533
0.419
0.282
0.220
*From collection cup surface
COLLECTION
CUP
SPRING
•JET SPINDLE
GASKET
-3 SLOTS
The in-line impactor has five stages. Particles
in the range of 0.3 to 3.0 microns are collected
by successive impingement.
Collection cups are positioned so that
the distance from the jet decreases
as the jet diameter becomes smaller.
Annular slots around cup minimize
turbulence.
Figure A-9. Design of a single stage from a Brink-type cascade impactor.
A-24
KVB72-806015-1308
-------
sampler is designed to operate at 2.8 l/min or less. An ideal flow rate is
around 2
A-1.6.2 Operation
The impactor will be carefully loaded with the stage cups and the
preweighed stage substrates. The Brink should be tightened with wrenches to
make certain the high temperature No. 116 asbestos gaskets are sealed. The
appropriate nozzle for isokinetic sampling is now added. The flue gas
temperature should be above the dew point but less than 450°F. After mounting
the impactor on the preheated sample probe, it will be inserted into the duct
to be preheated for at least 30 minutes. The inlet nozzle will be pointed
downstream of the flow during the heating phase. A predetermined flow rate is
established immediately and is maintained constant, since any attempt to
modulate flow to compensate for changes in the duct flow rate to provide
isokinetic sampling will destroy the utility of the data by changing the cut
points of the individual stages. Establishment of the correct flow rate
quickly is especially important for the short sampling times typical of high
dust loaded streams.
Sample times will normally vary from 5 to 45 minutes depending on the
dust loading. A total sample weight of <8 mg per stage should be collected.
The stages of the Brink impactor will yield cuts of 0.25, 0.50, 1.0, 1.5, and
2.5 ym. After the sampling cycle has been completed, the impactor is cooled
and disassembled. Proper disassembly is critical to make sure the collected
material stays where it originally impacted. After the desiccating of the
collection media, weighing is performed to determine the net particle accumu-
lations .
A-1.6.3 Data Presentation
To determine the concentration of particulates for any size range,
first determine the percentage of total particulates for each stage. Then the
cumulative percentage is determined beginning with the last stage of the
impactor. By plotting the effective cutoff diameter and the cumulative
percent on logarithmic probability graph paper, the particle concentration by
weight for any specific size can be determined.
KVB72-806015-1308
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
1. REPORT NO.
EPA-600/7-83-042
3. RECIPIENT'S ACCESSION"
4. TITLE AND SUBTITLE
Evaluation of Combustion Modification Effects on
Emissions and Efficiency of Wood-fired Industrial
Boilers
5. REPORT DATE
August 1983
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. J. Tidona, W. A. Carter, H. J. Buening, and
S. S. Cherry
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
KVB, Inc.
18006 Skypark Boulevard
Irvine, California 92714
11. CONTRACT/GRANT NO.
68-02-2645
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13 TYPE OF REPORT AND PERIOD COVERED
Final; 4/79 - 10/79
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES iERL-RTP project officer is Robert E. Hall, Mail Drop 65, 919 /
541-2477.
is. ABSTRACT
report gives results of full-scale tests to evaluate combustion modifi-
cations (lower excess air and variations in the overfire air system operation) for
emission control and efficiency enhancement on two wood-fired industrial boilers.
One boiler, rated at 100,000 Ib steam/hr, is fueled with a combination of wood bark
and coal. Implementation of lower excess air reduced NOx emissions by 18. 5 per-
cent and improved thermal efficiency by 0.89 percent. Variations in the overfire air
system reduced NOx by 20. 7 percent and improved efficiency by 1.63 percent. The
second boiler, rated at 200,000 Ib steam/hr when fired with hogged wood, can
achieve 250,000 Ib steam/hr when fired with hogged wood and an oil supplement.
The effectiveness of lower excess air in reducing NOx was about 14 percent with a
slight (0.6 percent) improvement in efficiency. Adjusting the auxiliary air dampers
reduced NOx by 17.2 percent and improved efficiency by 1. 7 percent. Polycyclic
organic matter (POM) was sampled at both baseline and optimum low-NOx conditions.
Under baseline conditions, POM emissions were similar to those of a coal-fired
spreader stoker and an oil-fired boiler, but were well below those of a pulverized-
coal-fired boiler tested previously. For the wood-fired boiler, POM emissions in the
low-NOx mode were higher than those at baseline.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b-IDEN-HFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Combustion
Wood
Fuels
Fossil Fuels
Nitrogen Oxides
Boilers
Dust
Industrial Processes
Particle Size
Stoichiometry
Pollution Control
Stationary Sources
Combustion Modification
Particulate
Overfire Air
13 B
2 IB
11L
2 ID
07B
13 A
11G
13H
14G
07D
3. DISTRIBUTION STATEMEN1
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
88
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
A-26
------- |