United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-83-038 Oct. 1983
v°/EPA Project Summary
Characterization of the NOx and
S02 Control Performances;
Southern Indiana Gas and Electric
Company, A. B. Brown Unit 1
Edward F. Peduto, Jr., Robert R. Hall, and Guy Tucker
A continuous emissions monitoring
program was conducted at Southern
Indiana Gas and Electric Company's
(SIGECO-s) A. a Brown Power Plant to
characterize the nitrogen oxide (NOJ
and sulfur dioxide (SO^) control per-
formances of Unit 1 in terms of process
variables. NOX results show that the
unit operated at significantly below70
percent of the existing NOX standard
(301 ng/J). Daily averages were 135-
219 ng/J, with a mean of 163 ng/J.
Thirty-day rolling averages were 160-
167 ng/J. SOj results indicate a mean
removal efficiency of 88.0 percent and
emissions of 344 ng/J for the north
tower, and 86.5 percent and 391 ng/J
(respectively) for the south tower.
Thirty-day rolling averages were 85.8-
90.5 percent and 85.3-88.0 percent for
the north and south towers, respec-
tively. Thirty-day rolling average SO^
outlet emission rates were 274-396
ng/J and 355-418 ng/J for the north
and south towers, respectively.
This Project Summary was developed
by EPA's Industrial Environmental Re-
search Laboratory. Research Triangle
Park, NC, to announce key findings of
the research project that is fully doc-
umented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
On June 11,1979, the EPA promulgated
the New Source Performance Standards
(IMSPS) for Utility Steam Generators, con-
tained in Subpart Da of 40 CFR 60 and
applying to generators on which construc-
tion started after September 18, 1978.
After promulgating a standard, the EPA
is required to assemble data and to review
the standard every 4 years. In preparing
for this review, EPA's Office of Air Quality
Planning and Standards (OAQPS) initiated
an overall program pertaining to the SC^
emission standard, aimed at documenting
the performance of high efficiency SO2
scrubber systems. This program subse-
quently involved the testing of two dual
alkali and two lime wet scrubber systems.
These system types were considered
"state-of-the-art." demonstrating consis-
tently high sulfur control performance.
GCA/Technokxjy Division was contracted
by EPA's Industrial Environmental Research
Laboratory at Research Triangle Park
(IERL-RTP) to conduct the test program at
the A. B. Brown Power Plant Primary
objectives of the program were to assess
the SO2 and NO, control of the FMC dual
alkali scrubber and the Babcock & Wilcox
(B & W) wall-fired boiler, respectively. Of
secondary importance was the evaluation
of these performances in conjunction with
various process data that were available
Sulfur emission control performance was
of primary concern to OAQPS. while the
correlation of process data to operational
performance was of more interest to
IERL-RTP.
The program lasted 10 months. Delays
in the original proposed schedule resulted
from a longer-than-anticipated monitor
system setup time and various problems
during data reduction
The final report for this program is in five
volumes, all of which are covered by this
single project summary.
-------
Facility Description
Southern Indiana Gas and Electric Com-
pany's (SIGECCTs) A. a Brown Unit 1 is a
modern pulverized-coal-fired boiler with a
generation capacity of 265 MWe and has
been operating since early 1979. The unit
is subject to the 1971 Federal NSPS
which limit SQ2 emissions to 1.2 lb/106
Btu (516 ng/J), NOX emissions to 0.7
lb/106 Btu (301 ng/J), and paniculate
emissions to 0.1 lb/106 Btu (43 ng/J).
Emissions of SO2 are controlled by an
FMC concentrated-mode dual alkali scrubber
system. The three-stage two-module
scrubber, shown schematically in Figure
1, was designed to meet the NSPS, when
4.5 percent sulfur coal is burned, by
treating all the flue gases at an efficiency of
about 85 percent SIGECO normally bums
3.5 percent sulfur coal and, reportedly,
has been able to meet the standard by
treating 90 percent of the flue gas at 90
percent efficiency while by-passing the
remaining 10 percent
The boiler, designed and built by B&W,
includes their dual register burners as
shown in Figure 2. Emission tests have
demonstrated that these burners limit NOX
emissions to less than 0.5 lb/106 Btu
Retractable Lighter
and Auxiliary Burner Assembly
Water Cooled
Outer Throat
Outer Register
Spin Vanes
Inner Register
Adjustable Air
Vanes and Registers
Figure 2. Dual register burner.
f
To Exhaust Stack
Lime Reactor
Main Chemical
Reaction:
CafOH)s + 2NaHSO3 - CaS03 + Na2S03 + 2HZO
CafOHh
Donut
pH6.S
Main Recirculation
Loop Components:
NaHSOs > Aqueous
Bleed
Stream
Main
Scrubber
Reaction:
SO* + A/2O + /Va2SO3 - 2NaHSO3
I
Regenerated
Aqueous
Na2S03
Solids
to
Disposal
Figure 1. FMC Corporation's concentrated double a/kali simplified process flow diagram.
2
-------
(less than 70 percent of the NSPS emis-
sions limitations). The burner limits NOX
formation to acceptable levels, by utilizing
a relatively long, narrow flame. The initial
burning in the center of the flame is in a
fuel rich mixture. Turbulence is kept low to
limit the degree of the fuel/air mixing to
that which is required to sustain combus-
tion and complete burning. The remaining
air needed to complete combustion is
admitted through a separate chamber to
totally surround the inner combustion zone
The resulting slow but efficient combustion
spreads heat evenly through the furnace,
lowering flame temperature and reducing
NOX formation.
Particulates are removed from the flue
gases by a Buell Envirotech coldside elec-
trostatic precipitator (ESP) prior to S02
removal. Tests with the scrubber offline
indicate that the ESP meets the standard
of 0.1 lb/106 Btu (43 ng/J). Opacity
monitors are at the ESP outlets, prior to the
scrubbers, for compliance monitoring.
Program Approach
The technical program emphasized
selected program objectives. The primary
objective was to determine the S02 collec-
tion efficiency of each of the two scrubber
modules and the NOX emissions from the
B&W boiler. A secondary, but also important
objective was to determine the influence
of process parameters on S02 performance.
Emissions of NOX were also measured,
and the effects of any variations in flue gas
oxygen or CO content were evaluated.
However, since NOX emissions were ap-
proximately 0.5 lb/106 Btu, no variations
in the operation of the dual register burners
were suggested.
The primary program data were acquired
using a mobile continuous emissions mon-
itoring laboratory maintained by IERL-
RTP. This system was used to acquire the
appropriate emissions and diluent data at
the inlets to and outlets from the parallel
scrubber modules. Concurrent with the
emissions data collection, applicable boiler
and scrubber process data were continu-
ously acquired. These data provided proc-
ess documentation for the emissions data
on a real time basis. The process and
emissions data files were subsequently
used to determine factors affecting the
emission control performance of the unit
Factors that may affect S02 collection
efficiency were an important focus of the
process evaluation. It was anticipated that
for example, gas flow to each module
might vary efficiency by 85 - 92 percent
Absorber pH, sulfite ion concentration,
and regenerator flow are other important
factors affecting S02 efficiency which were
considered in the initial tests.
Emissions parameters which applied to
the S02 control device included inlet and
outlet S02 and diluent levels. These mea-
surements were conducted at the specified
locations on both modules. In addition, the
mobile laboratory could measure the gas
flow rate through each module. Measure-
ments related to NOX included total NOX
prior to the scrubber, CO as a gauge for
combustion upset conditions, excess air
prior to the air preheater, and diluent at the
monitoring points for NOX.
In addition to the measurements/param-
eters mentioned above, other process
signals logged from the plant control panel
included various scrubber and boiler pa-
rameters and (initially) the plant's stack
emissions measurements for S02/02.
However, the stack emissions measure-
ments were discarded due to the erratic
behavior of this system. All data param-
eters were logged and processed by an
onboard minicomputer.
Utilizing data obtained during the data
acquisition phase of the program, GCA
evaluated the performance of the NOX and
S02 control equipment Throughout the
data collection phase, a field engineer
periodically observed and reviewed the
data At the end of the data collection
phase, all data were evaluated to fulfill the
program objectives.
Operational Profile
A B. Brown Unit 1 is the most expen-
sive plant to operate in the SIGECO system;
40
35:
30 \
,25i
D
I
15
10
5
Average
Boiler
Load
(129 MWe)
therefore, it is the last unit to be dispatched
and the first to reduce toad The actual load
profile during the test program depended
on the weather and the availability of other
units. In addition A. B. Brown Unit 1
experienced pulverizer problems that forced
load reductions. The average load during
the test program was 50 percent of capacity,
although both higher and lower loads were
encountered Figure 3 shows an hourly
frequency distribution for boiler load
About 70 percent of the hourly average
boiler loads were below 130 MWe; the
average was 129 MWa
Average excess air near the boiler (prior
to the air preheater) was 36 percent.
Oxygen concentrations at the same point
averaged 5.45 percent. The range of ob-
served oxygen concentrations is shown
in the frequency distribution of Figure 4.
Test Results
Data from the test program are sum-
marized in three categories: emissions
control performance, effects of process
variables on emission control performance,
and measurement system results.
Emissions Control Performance
NOX Emissions
The mean NOX emission rate for the full
test program was 163 ng/J (0.38 In/106
Btu). All the hourly readings were below
210 ng/J (0.49 lb/106 Btu) which is
equivalent to 70 percent of the 1971
NSPS. Daily average emission rates were
135-219 ng/J (0.3-0.51 lb/106 Btu).
90 100 110 120 130 140 150 160 170 180 190 ZOO 210 220 230 Z4O 250
Boiler Load Midpoint (MWe)
Figure 3. Hourly frequency distribution for boiler load.
3
-------
The 30-day rolling averages for NOX emis-
sions were all below 168 ng/J (0.4 lb/106
Btu), ranging from 160 to 167 ng/J
(0.370 to 0.388 lb/106 Btu) as shown in
Table 1. Table 2 is a statistical summary of
the NOX data
SQ2 Emissions
Emissions of S02 are controlled by two
parallel FMC dual-alkali scrubbers. For
these tests, SIGECO operated the FGD
system in its customary manner to meet
required regulations. The system was not
operated to optimize SQ2 collection effi-
ciency. Valid data were collected for the
performance of the north module for 68
days and for the south module for62 days.
This data collection spanned a total time
frame of 70 days. The boiler was shut
down for 2 days late in the program due to
pulverizer problems, and the south scrubber
module model was offline during the first
5 days of the test program as a result of
recirculation pump failure.
The mean SO2 removal efficiency for the
north module was 88.4 percent compared
to 86.6 percent for the south module for
the 30-day rolling average. The higher
average efficiency for the north module
was, in part, attributable to the first 6 days
of data collection when the north module
averaged 95.2 percent efficiency and the
south module was not operating. Tables 2
and 3 and Figure 5 show the 30-day rolling
averages beginning at the 30th day. The
decline in the rolling average SO2 collec-
tion efficiency coincides with the reduc-
tion in average boiler load which was lower
during the end of the test program. As
shown by the dotted line in Figure 5, the
scrubber system consistently operated
above the design guarantee
Emissions of SO2 for the north module
averaged 344 ng/J (0.80 lb/106 Btu),
while the south module averaged 391
ng/J (0.91 lb/106 Btu) on a daily average
basis. Thirty-day rolling averages for the
SO2 emissions are shown in Table 3 and
Figure 6. Thirty-day rolling averages for
the south tower were 355-418 ng/J
(0.81 -0.98 lb/106 Btu).
Effects of Process Variables on
Emission Control Performance
Various regression techniques were used
to investigate possible effects of process
variables on emission control performance
Hourly data consisting of up to 1400
hours were used to develop correlations
between process variables and emission
control performance as indicated by NOX
emission rates and S02 efficiencies from
each modula
Average O2
Concentration
12
10
i6
4
2
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
O2 Prior to Air Heater (%)
Figure 4. Hourly frequency distribution for oxygen concentration prior to the air preheater.
Table 1. NOX Emissions. 30-Day Rolling Averages
Day
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
Blr
load
(MWe)
99
141
137
118
132
113
142
115
105
104
160
121
134
161
178
101
102
112
173
173
178
166
131
143
209
177
129
106
109
141
142
139
111
104
102
Daily
average
NOX
(ng/J)
149
155
170
159
173
166
170
162
166
168
167
179
166
166
161
161
169
171
174
174
167
170
180
170
158
155
162
168
164
163
154
152
161
159
b
30-day
rolling
average
(ng/J)
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
160
160
160
161
161
161
Day
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262C
263C
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
Blr
load
(MWe)
102
112
99
123
158
153
167
191
117
109
157
114
122
100
97
-
.
98
106
106
119
109
107
106
108
120
143
126
105
105
101
156
134
151
191
Daily
average
NOX
(ng/J)
b
168
163
164
143
135
146
151
154
153
144
166
173
172
175
-
-
219
209
198
191
190
189
196
200
204
189
139
184
187
175
172
190
182
166
30-day
rolling
average
(ng/J)
162
162
163
162
162
162
162
161
161
161
160
161
160
161
160
-
-
160
161
162
163
164
164
164
164
165
166
167
167
167
167
166
166
166
167
'Not applicable.
bInsufficient data (<18 hrs).
c Boiler down.
-------
Table 2. Unit 1 - Performance Summary Statistics
Averaging
period
Hourly
Daily
30-day
Rolling
Table 3. SO2
Blr
load
Day (MWe)
212 99
213 141
214 137
215 118
216 132
217 113
218 142
219 115
220 105
221 104
222 160
223 121
224 134
225 181
226 178
227 101
228 102
229 1 12
230 173
231 173
232 178
233 166
234 131
235 143
236 209
237 177
238 129
239 106
240 109
241 141
242 142
243 139
244 111
245 104
246 102
247 102
248 112
249 99
250 123
251 158
252 153
253 167
254 191
255 117
256 109
Sulfur emissions (ng/J)
Statistical
parameter North South
Mean 341 391
StdDev 115 101
Min 29 102
Max 662 887
Mean 344 391
StdDev 92 62
Min 67 216
Max 544 600
Mean 341 394
StdDev 35 19
Min 274 355
Max 396 418
Tower efficiency (%)
North South
88.0 86.5
4. 1 3.4
75.9 71.2
99.0 96.7
88.0 86.5
3.3 2.2
81.1 82.5
97.7 92.2
88.4 86.6
1.6 1.0
85.8 85.3
90.5 88.0
NOX
Emissions
163
15.8
76
209
163
10.9
135
219
163
2.5
160
167
Performance: Daily and 30-Day Rolling Average
North tower
Outlet
emissions
(ng/J) Efficiencies^
30-day 30-day
Daily Rolling Daily Rolling
148 ' 94.5
129 ' 95.6
67 97.7
134 95.2
158 94.4
1 70 93.5
235 91.0
396 85.3
503 81.1
357 86.6
y?7 Q1 9
AO / j i .0
288 89.4
241 91.6
315 89.6
273 91.0
258 93.0
247 91.3
239 91.7
182 93.8
293 90.6
252 91.4
211 93.2
242 92.2
323 89.5
314 90.0
242 92. 1
398 86.9
372 87.5
322 89.0
306 274 89.7 90.5
373 278 87.3 90.4
464 283 83.9 90.2
465 286 83.8 90.1
434 297 83.9 89.9
445 304 c 89.8
506 313 c 89.6
524 320 84.5 89.5
410 319 86.7 89.7
383 321 87.4 89.5
332 319 87.9 89.7
274 320 89.3 89.6
244 318 91.7 89.7
277 319 89.7 89.6
283 318 89.1 89.6
345 320 87.7 89.5
South tower
Outlet
emissions
(ng/J) Efficiencies?*,)
30-day
Daily Rolling Daily
b b
b b
b b
b b
b b
b b
362 86.3
413 84.8
433 83.8
293 88.9
218 92.0
216 92.2
330 88.6
382 87.6
365 88.1
392 88.8
343 88. 1
328 88.7
299 90.1
423 86.3
274 90.6
245 92.0
325 89.5
404 86.8
358 88.4
331 89.2
379 87.6
461 84.6
364 87.6
440 355 85. 1
397 357 86.3
484 359 83.2
477 360 83.3
496 366 82.5
523 370 c
600 375 c
493 377 84.2
475 378 84.6
438 380 85.6
441 383 83.7
353 386 86.2
328 384 88.9
365 386 86.3
359 385 86.2
451 388 83.8
30-day
Rolling
88.0
87.9
87.8
87.7
87.7
87.7
87.7
87.8
87.9
87.8
87.7
87.5
87.6
87.5
87.4
87.3
Oxides of Nitrogen
Emissions of NOX appeared to be related
to the flue gas oxygen concentration as
measured at the inlet locations (the north
tower inlet value was used), boiler load.
and CO concentration. It was expected
that the oxygen concentration in the boiler.
as indicated by the concentration prior to
the air preheater, might affect NOX emis-
sions, but this proved to be a very weak
correlation.
The equation selected to predict NOX
emissions is:
NOX (ng/J) = 27.8 + 1 2.9 (02, %)
+ 0.2 1 6 ( load. MWe) - 0.0495 (CO, ppm)
(1)
This equation explained 52 percent of the
variation in NOX emissions and was highly
significant as indicated by the F value of
491. The predicted and measured NOX
emissions are compared in Figure 7. If the
linear regression explained all the variation
in NOX emission rates, all the points would
fall on the indicated 45 degree lina Be-
cause the selected equation explains 52
percent of the variation in emissions, the
data points are scattered around its 45
degree line
The linear regression equation developed
for the NOX data collected at A. & Brown
defines a real and statistically significant
relationship that was observed in the data
base. It does not prove a physical or
chemical cause and effect relationship nor
should it be applied to data from other
sites. Further investigation of the observed
relationship between process variables
and emissions may be appropriate.
Sulfur Dioxide
The development of correlations be-
tween SO2 control (efficiency) and proc-
ess parameters was approached using
venturi scrubber models as background
information. According to these models.
the main variable which affects collection
efficiency is liquid drop surface area The
surface area is directly proportional to the
flue gas velocity. The higher the gas velocity,
the smaller the drop size and, consequently.
the larger the surface area an atomized
liquid will exhibit
Similarly, FMCs tower design depends
on gas velocity for atomization. Conse-
quently, these towers are expected to
exhibit higher collection efficiencies at full
loads.
Collection efficiencies for both towers
were affected by gas flow only as shown
by the equations below:
North Module-
mm LI i fvi\>HJuic.
eff = 6.4x1 0-7 (flow)2 5 - 3.4x1 0« (flow)3-5
44.7x1 0-& (flow)45 (2)
R2 = 0.9814forN=530
-------
Table 3. (Continued)
North tower
Outlet
emissions
(ng/JI
Blr
load
Day (MWe)
257 157
258 114
259 122
260 1OO
261 97
262 rf
263 d
264 98
265 106
266 106
267 119
268 109
269 107
270 106
271 108
272 120
273 143
274 126
275 105
276 105
277 101
278 156
279 134
280 151
281 191
'Not applicable.
Daily
325
298
360
386
402
d
d
418
443
438
409
407
376
399
436
394
441
474
544
524
517
442
422
373
365
30-day
Rolling
322
324
326
333
336
d
347
352
357
363
364
364
367
372
373
375
378
382
389
393
392
395
395
396
Efficiencies (%)
Daily
88.3
89.1
86.6
85.5
85.2
d
84.7
83.3
84.2
85.1
85.4
86.3
85.6
84.3
85.7
84.2
83.6
81.6
82.2
82.1
85.2
86.3
87.1
87.4
30-day
Rolling
89.5
89.4
89.4
89.0
88.8
d
d
88.4
88.1
87.8
87.4
87.4
87.3
87.1
86.9
86.8
86.7
86.6
86.3
86.1
85.9
85.9
85.8
85.8
85.8
South tower
Outlet
emissions
(ng/J)
Daily
373
355
429
446
461
d
d
c
454
462
424
408
362
382
424
388
368
385
510
490
489
388
380
358
318
30-day
Rolling
387
387
389
394
395
d
d
407
409
413
417
418
414
415
414
414
412
412
413
414
413
412
410
407
405
South Module:
eff = 1 x1 0-3 (flow)2-5 - 6.3x1 0'6 (flow)3-5
+1 . 0x1 0'8 (flow)4-5 (3)
R2 = 0.9280 for N = 506
Efficiencies (%)
Daily
86.5
87.1
84.0
83.3
83.0
d
d
c
83.1
83.3
84.5
85.4
86.9
86.3
84.8
86.0
86.7
86.8
83.1
83.5
83.2
86.8
87.1
87.8
89.1
"Scrubber module offline.
'Insufficient data (<18 hrs).
^Boiler offline.
y *
90
?
g 89
-S
.5
ui
| 88
|
«
U) &7
86
85
V
where eff istheS02 removal efficiency (%),
flow is the gas flow rate ( 1 03 acfm).
R2 = goodness of fit constant and
Roiling N=numberofdatapointsanalyzed
872 These equations describe greater than
872 92 percent of the process variation for the
87.2 data segment analyzed. Note, however.
86-9 that these correlations may indicate only
d the data set obtained from this test site;
d they may not represent performances
86.3 which may be exhibited by other dual alkali
86. i systems.
856 ^s 'On9 as tne P*"* °^ tne system 's 'n
85,5 tolerance, the variation of control perform-
85.6 ance is only a function of gas flow rate. In
85-6 the FMC system, the scrubbants are highly
HJ buffered and automatically regenerated. In
85.6 addition, the liquid circulation rate is held
85.6 constant through the tower and regenerated
85-4 automatically as required.
||^ Figure 8 is a plot of S02 removal efficien-
85.3 CY as a function of gas flow rate. As shown.
85.4 the removal efficiency is only affected by
85.4 the gas flow rate through the tower as long
8S-4 as the pH of the system is within design
tolerances. According to the data set and
the extrapolation of the curve, one can
expect a removal efficiency approaching
90 percent at a gas flow rate of 400,000
acfm. As stated above, this plot depicts a
trend in the data and should not be applied
to other dual alkali scrubber systems.
Jv Boiler Offline on
\ v*^^*^"*^*^ DayS 262 and 263
•
>V
x
TT
>*-+-**
^^•^^™W
X^
^-*\i
>-*^
s*+
*%
\
\
^
South Module— r *\. '
m
«
\
"
V- North Module
\
V*.
••%
*
\
V^
\
\
V
•-•*_ __
Xc
X
^*Jt
A
-». M H
Desion Guarantee ^^ • a-*-*'*
.,.,.,,,
241
246
257
256
261
Day
266
271
276
Figure S. SOi efficiency. 30 day rolling average.
No equations were generated to describe
outlet emission performance, since the
equations describing efficiency can also
be rearranged to describe outlet emission
rates.
Measurement System Results
Operational Experience
The mobile emissions laboratory, used
to collect the primary emissions data.
initially was anticipated to require a setup
and checkout period of about 1 month.
Subsequently, 1 week was scheduled for
the performance specifications tests de-
signed to assess the precision and relative
accuracy of the measurement system.
Actual setup and commencement of mon-
itoring required an additional 6 weeks for
equipment troubleshooting, remedial modi-
fications and subsequent checkout
Problems most commonly encountered
28 1 within the extraction system and a spurious
voltage induction problem which caused
the baseline of the analytical instrumenta-
tion to shift. The latter problem was reme-
-------
I
c*
o
to
576
420
400
380
360
340
320
300
280
270
• Design Guarantee-
Boiler Offline on
Days 262 and 263
- South Module
North Module
1.2
0.98
0.95
0.88
I
0.83 €
w
c
.o
0.79 I
ki
0.74
0.70
0.65
241 246 251 256 261 266 271 276 281
Day
Figure 6. SOz emissions, 30 day rolling average.
d\ed prior to monitoring. Leakage through-
out the system was due to valve diaphragm
ruptures and the flowing of Teflon sample
lines at connection points throughout the
system. These problems were remedied
by system modifications.
After correcting the system problems,
approximately 68 days of data were col-
lected from July 31 through October 8,
1981. During this time, the plant was
offline 2 days.
After system start-up, very little data
loss resulted from hardware failure. Hard-
ware availability during the program ap-
proached 98 percent while the data avail-
ability rate (including primary and backup
data systems) approached 95 percent
Valid data capture was approximately 93
percent for the primary emission param-
eters, and 70 percent for associated proc-
ess measurements. The process data cap-
ture rate was significantly lower than the
emissions capture rate because these sig-
nals were not connected to a backup
logger and most were often not available
from control room log sheets.
During the tests, the mobile laboratory
was attended by a full-time operator. Daily,
the operator conducted calibration checks,
performed routine maintenance, accessed
I the previous day of data (printouts), and
filled out daily maintenance check lists and
site logs. Most of the day was spent
reducing data and performing program
related paperwork. All data were proc-
essed into report format by the onsite
computer.
Stratification Results
Before initiating the routine monitoring
phase, stratification tests were conducted
at each monitoring location. This test se-
quence ensured that the probe locations
would provide representative flue gas sam-
ples. The procedure involves traversing
the cross-sectional area of the flue gas
stream to define the spatial variability of
the analyte of interest This procedure is
normally conducted where the probability
of stratification is high (e.g., after wet
scrubbers). Periodically, the tests were
repeated to verify that the representative-
ness of the flue gas samples remained
unchanged
Table 4 summarizes the stratification
results. Data in the first two columns were
obtained during the initial stratification
tests. Based on these initial tests the
probes were placed in the geometric center
of each monitoring location. Subsequent
tests at the outlet locations indicated that
the average points of concentration at the
outlet locations were unchanged. Further
testing was not conducted at the inlets
due to the low probability of variable
stratification.
Performance Specifications Tests
Results of the performance specifications
tests are summarized in Tables 5 and 6.
Due to a shortage in span gases, the
optional 2 hour drift test was not conducted.
All analyzers conformed to the 24 hour
drift and calibration error criteria, except
the outlet S02 analyzer (which exceeded
the midscale calibration error criterion)
and the NOX analyzer (which exceeded the
calibration drift and the high scale calibra-
tion error criteria). Dueto these results, the
NOX span gases were analyzed to reverify
the "true" concentration; however, results
of the analyses did not alter the perform-
ance results.
Results of the relative accuracy testing
are shown in Table 6. Generally speaking,
the results were favorable and, except for
the NOX results, were less than the stated
compliance limit of 20.0 percent relative
accuracy in terms of emission rate. The
results for the NOX analyzer were 30
percent in terms of emission rate. This
high value for relative accuracy is the
result of a large confidence interval and
not an absolute bias. This denotes random
scatter when comparing the differences
between the reference method and the
analyzer.
Emission rates for S02 were calculated
using both the 02- and C02- based "F"
factors. As shown by the results, the
different methods for calculating the emis-
sion rate in some cases did not yield
comparable relative accuracy results. This
discrepancy may be a result of the high
C02 relative accuracy (eg., 0.8 percent
C02) and the opposite relative bias for the
02 and C02 analyzers.
Monitoring System Precision and
Accuracy
Individual instrument precision and ac-
curacy estimates were determined accord-
ing to the proposed Quality Assurance
Regulations, scheduled to be promulgated
as 40 CFR 60, Appendix F. These precision
estimates were determined by GCA, utiliz-
ing the daily zero and span data; the
accuracy of each instrument was assessed
by an independent auditor.
Results of instrument precision and ac-
curacy are listed in Table 7. Precision
estimates for the total data collection period
appear satisfactory: results of the perform-
ance audit were "acceptable" as defined
by the auditors.
-------
220
210
Legend:
A = 1 Point. B = 2 Points,
C = 3 Points, etc.
200
'1
190
780
170
160
750 <
140
130
720
A A
tet
A
A
A A
A A
A
A AA AAX*BUA A A A
AA A A A AAF-iByr IHAAA C AB A AA A
A 1CRBABC CCLOCLA4CB*'**£ OCC 0 A AAB
ASA BCAC OEHiGCrV&t'SOtSCD EftOCBA A AAAM
lOF?e?09MFAS»A A A AAAA A
CCJC V.,:!S?iFXruC3tGCl«OC»CAF BAA A A I
A A A
A A A
A A
770
BAA
A A A
AA A A
A CiD.
A a Aeac*:3iooHr:j:rbrooo3 OSODOCOCBA AA
a ecCDOCCS5CC::=^<»»CA8A£AFC5C A»tC
AAbeACCBbFCCCC^UP&ABASCCAA A « A
B C ?C£3«L0lLCiCa ABA A A
C AC CO
6 A A 369AA AA
a A AA
24 A* E A
BC AC A A
-------
100.
75-
g
IK
| 50-
-------
Table 6. Summary of Relative Accuracy Test Results for A R Brown Unit 1
North Tower (% RA or % CO2/02)
Inlet Outlet
Parameter Test 1 Test 2 Test 3 Test 1 Test 2
SO2 16.9 5.5 21.4 16.4 14.6
02 0.46 0.30 1.10 0.13 1.14
C02 — — — 0.81 2.29
NOX 23.9 29.0 26.5
Eso2,o2F>" 17-3 S.I 13.6 17.0 12.9
Eso2(co2Ff - - - 21.V 27.9*
ENO(02Fi 36.6" 31.0" 23.3d
"Test runs invalidated.
b Relative accuracy based on O2 "F" factor emission rate.
c Relative accuracy based on CO2 "F" factor emission rate.
d Exceeds 20 percent relative accuracy based on emission rate.
—Not applicable.
Table 7. AB. Brown Precision Estimate Results
Zero precision
estimate {%)
Location Parameter Month Lower Upper
Met SO2 August -0.9 0.5
(Horiba) September -0.3 0.8
October0 -0.2 0.8
O2 August -1.0 0.7
(MSA) September -0.3 0.0
October' -0.3 0.0
Outlet S02 August -7.2 6.1
(DuPont) September -8.2 12.5
October8 -2.3 2.9
O2 August 0.0 0.5
(MSA) September 0.0 0.6
October* -0. 1 0.3
CO2 August -0.5 1.2
(Horiba} September -0.5 1.3
October* 0.5 0.6
Test3
13.5
0.55
1.17
14.1
8.7
South Tower (% RA or % CO2
Inlet
Test 1 Test 2 Test 3 Test 1
28.7 12.5 14.7 14.8
0.53 0.14 1.09 0.40
1.98
a a a
28.6d 12.5 12.8 11.8
16.1
a a a
/OJ
Outlet
Test 2 Test 3
14.7 18.4
0.62 1.33
0.97 1.14
14.2 14.4
12.8 19.2
Span precision
estimate (%)
Lower
-1.6
-1.5
-0.4
-4.0
-1.8
-2.7
-16.2
-8.3
-6.6
-2.5
3.7
0.6
-4.0
-2.7
1.8
Upper
2.3
1.5
-3.9
1.8
1.4
1.3
9.1
4.4
5.0
1.7
4.6
0.8
0.8
2.5
4.0
•
"Less than 10 entries used for calculating these estimates.
10
-------
Edward F. Peduto, Jr., Robert R. Hall, and Guy Tucker are with GCA/Technology
Division, Bedford, MA 01730.
J. David Mobley is the EPA Project Officer (see below).
The complete report consists of five volumes, entitled "Characterization of the
NO* and SOi Control Performances; Southern Indiana Gas & Electric Company,
A. B. BrownUnit 1:" all five volumes are available as a set: Order No. PB83-240
663; Cost: $91.00 or individually as—
"Volume I. Program Results," (Order No. PB 83-240 671; Cost: $13.00)
"Volume II. Program Documentation," (Order No. PB 83-240 689; Cost:
$23.50)
"Volume III. North Module Sulfur Dioxide Data Reports," (Order No. PB
83-240 697; Cost: $23.50)
"Volume IV. South Module Sulfur Dioxide Data Reports," (Order No. PB
83-240 705; Cost: $23.50)
"Volume V. Oxides of Nitrogen Data Reports," (Order No. PB 83-240 713;
Cost: $23.50)
The above prices are subject to change and the reports are available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
11
-------
JO
3 5
Q) O
*g
Is
la
(6
00
8
C 0. 3
2<
« 8 3
a o
Jo
a X
Sm
_, O
™ m
5S
0
8
> m i
s.
-• o
O11
X
A
01
N>
o>
a>
------- |