United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EMB Report 84-IBR-24
August 1984
Air
Industrial Boilers
Emission Test
Report
General Electric Noryl
Products Division
Selkirk, New York
-------
EMISSION TEST REPORT
METHOD DEVELOPMENT AND TESTING FOR
INDUSTRIAL BOILERS, PM AND NO
General Electric Company
Selkirk, New York
ESED NO. 76/13
EMB NO. 84-IBR-24
by
PEI Associates, Inc.
11499 Chester Road
P.O. Box 46100
Cincinnati, Ohio 45246-0100
Contract No. 68-02-3849
Work Assignment Nos. 7 and 8
PN 3615-7 and 3615-8
Prepared for
Mr. Dennis Holzschuh
Task Manager
U.S. ENVIRONMENTAL PROTECTION AGENCY
EMISSION STANDARDS AND ENGINEERING DIVISION
EMISSION MEASUREMENT BRANCH
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
January 1985
-------
CONTENTS
Figures iii
Tables iv
Acknowledgment vi
Quality Assurance Element Finder vii
1. Introduction 1
2. Summary of Test Results 3
2.1 Test protocol 3
2.2 Continuous emission monitor data 6
2.3 Particulate and fuel analysis test results 14
3. Quality Assurance 30
3.1 Continuous emission monitors 31
3.2 Manual tests—particulate and NO 51
/\
4. Sampling Locations and Test Methods 64
4.1 Sampling locations 64
4.2 Continuous emission monitors—sample extraction,
analysis, and data reduction 64
4.3 Particulate test methods and analytical procedures 69
4.4 Manual test methods for NO 69
7\
5. Process Description and Operation 71
5.1 Boiler description 71
5.2 Burner description 73
5.3 Operating history 74
5.4 Control procedures 76
References 78
Appendices
A Computer Printouts and Example Calculations A-l
B Field Data Sheets B-l
C Laboratory Data Sheets C-l
D Sampling and Analytical Procedures D-l
E Equipment Calibration Procedures and Results E-l
F Quality Assurance Summary F-l
G Project Participants and Field Log G-l
ii
-------
FIGURES
Number Page
1 Example NO Calibration Curve 40
A
2 Example Op Calibration Curve 41
3 Example CO Calibration Curve 42
4 Example C02 Calibration Curve 43
5 On-Site Audit Data Sheet 53
6 Field Audit of Dry Gas Meter (Meter Box FB-4) 54
7 Field Audit of Dry Gas Meter (Meter Box FB-8) 55
8 Example of an Unacceptable Meter Box Audit 56
9 Example Calculation Form Used by PEI During Test Series 57
10 Audit Report NO Analysis Results on Samples Received
8/15/84 x 60
11 Audit Report NO Analysis Results on Samples Received
8/16/84 x 61
12 Audit Report NO Analysis Results on Samples Received
8/17/84 and 8/18/84 62
13 Audit Report NO Analysis Results on Samples Received
8/22/84 x 63
14 Boiler 4 Exit Stack (No Scale) 65
15 Boiler 5 Exit Stack (No Scale) 66
16 CEM System Layout 67
17 Unit 5 Layout 72
18 Coen Parallel Flow Type LEA Burner 75
-------
TABLES
Number Page
1 Summary of Boiler Operating Parameters and Emissions Tests 4
2 Summary of Continuous Emission Monitoring Data 7
3 Summary of Maximum and Minimum CEM Data by Test Block 10
4 Relationship Between NO Concentration and Excess Air
Levels—Residual Oil x 11
5 Relationship Between NO Concentration and Excess Air
Levels—Natural Gas 12
6 Summary of NO Emission Rates 15
X
7 Summary of Particulate Sampling Conditions, Boiler No. 5
Stack 17
8 Summary of Particulate Emission Results 18
9 Between-Run Statistical Data for Similar Sample Types
and Boiler Loads 21
10 Summary of Fuel Analytical Results 22
11 Method 5B Relative Percent Weight Loss at 160°C 26
12 Residual Sulfate (S0.=) Analysis of Methods 5 and 5B
Sample Fractions 27
13 Analytical Results From Ether/Chloroform Extraction of
Back-Half Solutions 29
14 Monitor Stratification Test—Boiler 4 (8/14/84) 32
15 Monitor Stratification Test—Boiler 5 (8/16/84) 32
16 Test Results for NO Monitor 24-Hour Zero and Calibration
Drift x 33
17 Test Results for CO Monitor 24-Hour Zero and Calibration
Drift 34
iv
-------
TABLES (continued)
Number Page
18 Test Results for 0~ Monitor 24-Hour Zero and Calibration
Drift c 35
19 Test Results for C09 Monitor 24-Hour Zero and Calibration
Drift c 36
20 Monitor Response Time 37
21 NO Monitor Response Time 37
A
22 0? Monitor Response Time 38
23 CO Monitor Response Time 38
24 C02 Monitor Response Time 38
25 NO Linear Regression Data 44
A
26 02 Linear Regression Data 45
27 CO Linear Regression Data 46
28 CO^ Linear Regression Data 47
29 Summary of NO CEM Audit Results 48
A
30 Comparison of Reference Method 7 and NO CEM Test Results 49
A
31 Comparison of Oxygen and Carbon Dioxide Results—CEM and
Reference Method 3 (Orsat) 50
32 Field Equipment Calibration 52
33 Particulate Filter and Reagent Blank Analysis 58
34 Boiler Process Data 77
-------
ACKNOWLEDGMENT
Mr. Dennis Holzschuh, EPA Task Manager, provided overall project coor-
dination and guidance and observed the test program. Messrs. Kevin Johnson
and John Martinez of Radian Corporation, an EPA Contractor, provided project
coordination relative to project scope and process operation. Mr. Len Keck
of General Electric Company provided assistance in scheduling and process
operation. Mr. Charles Bruffey was the PEI Project Manager. Principal
report authors were Messrs. Charles Bruffey, Paul Reinermann, and Daniel
Scheffel.
-------
QUALITY ASSURANCE ELEMENT FINDER
Location
Section Page
(1) Title page
(2) Table of contents
(3) Project description
(4) QA objective for measurement of data in terms
of precision, accuracy, completeness, repre-
sentativeness, and comparability
(5) Sampling procedures
(6) Sample custody
(7) Calibration procedures and frequency
(8) Analytical procedures
(9) Data reduction, validation, and reporting
(10) Internal quality control checks and frequency
(11) Performance and system audits and frequency
(12) Preventive maintenance procedures and schedules
(13) Specific routine procedures used to assess data
precision, accuracy, and completeness of specif-
ic measurement parameters involved
(14) Corrective action
(15) Quality assurance reports to management
1
Appendix F
Section 3
Appendix D
Section 4
Appendix C
Appendix E
Section 3
Appendix D
Section 4
Appendix F
Section 3
Appendix F
Section 3
Appendix F
Section 3
Appendix F
11
1
F-2
D-l
C-l
E-l
D-l
F-3
F-5
F-3
F-12
Appendix F F-4
Appendix F F-ll
Appendix F F-12
vn
-------
SECTION 1
INTRODUCTION
The United States Environmental Protection Agency (EPA) is developing
standards of performance for industrial boilers in accordance with Section
111 of the Clean Air Act as amended August 1977. The Act requires that the
standards be based on the "... best technological system of continuous emis-
sion reduction which the Administrator of EPA determines has been adequately
demonstrated." Accordingly, EPA is interested in the nitrogen oxide (NO )
/\
control capability of Low Excess Air (LEA) burners on industrial boilers.
To support the standards development process and provide data to charact-
erize emissions from industrial boilers, PEI performed a series of atmospheric
emission tests on two oil- and gas-fired boilers equipped with the latest
commercial LEA burners at the General Electric Noryl Products facility in
Selkirk, New York. These tests were conducted under contract to EPA's Emis-
sion Measurement Branch (EMB) from August 14 to 20, 1984. The primary objec-
tives of the test program were:
0 To characterize NO emissions as a function of fuel type, boiler
load, excess air [oxygen (02) level], and combustion air preheat
temperature.
0 To obtain particulate emissions data.
No major problems were encountered during the test program and project objec-
tives were met.
All testing was performed on Boilers 4 and 5. Both boilers were tested
while firing natural gas, and Boiler 5 only was tested while firing No. 6
residual and No. 2 distillate oil. Continuous emission monitor (CEM) systems
for NO , 02, CO, and C02 were used to characterize these pollutants as a
function of boiler load, excess air, and combustion air reheat temperature.
Manual testing was conducted concurrently with the CEM tests while Boiler 5
-------
was firing residual and distillate oil to determine the concentration and
mass emission rate of participate matter according to procedures described in
EPA Test Methods 5 and 5B.* Flue gas volumetric flow rates, temperature, and
moisture content were determined in conjunction with the particulate tests.
As a data quality assurance check for the NOV CEM system, EPA Method 7 (NOV)
/^ /\
tests were also conducted over the course of the test program. In addition,
fuel oil samples were collected during the test program and subjected to a
proximate and ultimate analysis.
Section 2 of this report summarizes the results of the test program.
Section 3 addresses quality assurance activities undertaken to assure repre-
sentative data collection. Section 4 summarizes the testing and analytical
procedures used and describes the sampling locations. Section 5 describes
the process and its operation during the test series. Appendices A through F
contain computer printouts and calculations, all field and laboratory data
sheets, detailed descriptions of the testing and analytical procedures used,
and equipment calibration procedures and results.
40 CFR 60, Appendix A, July 1984.
-------
SECTION 2
SUMMARY OF TEST RESULTS
This section details the results of the field test program. For the
convenience of the reader, emission data are presented in both metric and
English units where applicable. Also, subsections are used to present each
phase of the sampling program and the corresponding emission results.
2.1 TEST PROTOCOL
Table 1 presents a summary of the process operating parameters and the
types of tests performed during the program. For reporting purposes, the
sampling dates, test times, and types of tests are grouped by fuel type (oil
or natural gas). The actual sequence of events varied somewhat because of
boiler operating problems and plant production schedules that affected boiler
load.
Two identical boilers (Boilers 4 and 5) capable of firing both natural
gas and fuel oils were used for this study. Each has a rated capacity of
150,000 pounds of steam per hour, but normal operation is in the 50,000 to
60,000 Ib/h range. Both boilers possess Coen parallel-flow, low-excess air
burners. Flue gas exits the top of the boiler, turns, and then passes
downward through an economizer before it enters the exit stack via a
rectangular breeching. Testing was conducted in the exit stacks of the two
boilers. Boiler 4 was fired with natural gas, and Boiler 5 was fired with
both No. 6 residual oil and No. 2 distillate oil as well as natural gas. A
total of 19 individual test blocks were conducted during which NO , 0?, CO,
/\ C~
and C02 emissions were continuously monitored. Of this total, six test blocks
were conducted on Boiler 4 and the remainder on Boiler 5. Ten of the test
blocks were for natural gas (Boilers 4 and 5), and nine were for residual and
distillate oil (Boiler 5 only).
As shown in Table 1, particulate emission tests on Boiler 5 were con-
ducted according to EPA Methods 5* and 5B* sampling and analytical procedures
-------
TABLE 1. SUMMARY OF BOILER OPERATING PARAMETERS AND EMISSIONS TESTS
Test
block
1
2
3
4
5
6
7
8
9
NG-1
NG-2
Date
(1984) and
Time (24-h)
8/16
1215-1415
8/16
1707-1907
8/17
0027-0227
8/17
1112-1442
8/17
1645-1910
8/17
2201-2341
8/18
1725-1925
8/18
2049-2249
8/19
0014-0214
8/15
1310-1740
8/15
1823-2018
Boiler
ID
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 4
No. 4
Operating parameters
Fuel9 type
R.O.
R.O.
R.O.
D.O.
D.O.
Low viscosity
R.O.
R.O.
R.O.
R.O.
N.G.
N.G.
Loadb
1/2
1/2
1/2
1/2
1/2
1/2
Full
3/4
3/4
3/4
1/2
02 Level c
High
Low
Low
High
Low
Low
Low
Low
High
Low
Low
Air
preheat
tempera-
ture, °F
Ambient
Ambient
si 60
si 50
si 50
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Emission tests
CEM's
(NO , 0,,
C02, CO)
X
X
X
X
X
X
X
X
X
X
X
Particulate
(Methods 5-5B)
X
X
X
X
X
X
X
X
X
-
-
Reference
Method 7
(NOX)
X
X
-
-
-
X
X
X
-
X
-
(continued)
-------
TABLE 1 (continued)
Test
block
NG-3
NG-4
NG-5
NG-6
NG-7
NG-8
NGr9
NG-10
Date
(1984) and
Time (24-h)
8/19
1041-1241
8/19
1440-1540
8/19
1626-1726
8/19
1837-1937
8/19
2052-2152
8/19
2227-2257
8/19
2315-2345
8/19-20
2351-0021
Boiler
ID
No. 4
No. 4
No. 4
No. 4
No. 5
No. 5
No. 5
No. 5
Operating parameters
Fuel9 type
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
Loadb
3/4
Full
Full
1/2
1/2
3/4
3/4
3/4
02 Level0
High
High
Low
High
High
High
Low
Low
Air
preheat
tempera-
ture,0 °F
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
sfl60
Emission tests
CEM's
(NO. 0?,
C02, CO)
X
X
X
X
X
X
X
X
Particulate
(Methods 5-5B)
-
-
-
-
-
-
-
Reference
Method 7
(NOX)
—
-
-
X
-
-
-
aFuel type: R.O. = No. 6 Residual oil; D.O. = No. 2 Distillate oil; N.G. = Natural gas
bBoiler load: 1/2 (=70,000 Ib/h steam); 3/4 (=105,000 Ib/h steam); Full (=140,000 Ib/h steam)
C0xygen level for natural gas^and residual oil tests: Low (=0.5-1.5%), High (>3.0%); for distillate oil test:
Low (2.0%), High (>3.0%).
Ambient air temperature is approximately 70°-90°F.
-------
while the boiler was firing residual and distillate oil. No participate
measurements were made during any of the natural gas test blocks. Manual
emission tests for NO (Method 7*) were conducted during 7 of the 19 test
/\
blocks, as a quality assurance check of the NO CEM system.
A
Throughout the test program, boiler load and excess air (02) levels were
varied to characterize NO emissions as a function of these operational param-
/\
eters. Personnel from Radian Corporation, the EPA's NSPS Contractor, coordi-
nated changes in process operation with the boiler operators and monitored all
pertinent data during each test block. As indicated in Table 1, boiler loads
of one-half capacity (representing =70,000 Ib/h), three-fourths capacity
(representing =105,000 Ib/h), and full capacity (representing =140,000 Ib/h)
were evaluated.
Excess air (0^) levels were designated as high (ranging from 3 to 4
percent) and low (typically less than 1.5 percent). Combustion air temperatures
were generally at ambient levels between 21° and 32°C (70°-90°F); however, 2
of the 19 test blocks were conducted at a preheat temperature of 71°C (160°F).
The following subsections present the results of the test program.
2.2 CONTINUOUS EMISSION MONITOR DATA
An extractive monitoring system was assembled on site to characterize
emissions data for NO , 02, CO, and C02 as a function of boiler load, excess
air (0?) level, and combustion air preheat temperature. Table 2 summarizes
the boiler operating parameters and corresponding CEM data for each designated
test block. Table 3 presents the maximum and minimum pollutant concentrations
recorded for each test block.
A test block typically represented a 2-hour monitoring period during
which the boiler load and oxygen level remained relatively constant at the
specified conditions listed in Table 2. Several of the test blocks run on
Boilers 4 and 5 while these boilers were firing natural gas were 30 minutes to
1 hour in duration.
Regardless of the test time, the reduction of the CEM data was accom-
plished by taking an average chart reading for every 5-minute period and
determining the corresponding pollutant concentration by the use of linear
40 CFR 60, Appendix A, Reference Methods 5, 5B, and 7, July 1984.
-------
TABLE 2. SUMMARY OF CONTINUOUS EMISSION MONITORING DATA
Test
block
1
2
3
4
1 5
6
7
8
9
NG-1
NG-2
Date
(1984) and
Time (24-h)
8/16
1215-1415
8/16
1707-1907
8/17
0027-0227
8/17
1112-1442
8/17
1645-1910
8/17
2201-2341
8/18
1725-1925
8/18
2049-2249
8/19
0014-0214
8/15
1310-1740
8/15
1823-2018
Boiler
ID
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 4
No. 4
Operating parameters
Fuel3 type
R.O.
R.O.
R.O.
D.O.
D.O.
Low viscosity
R.O.
R.O.
R.O.
R.O.
N.G.
N.G.
Loadb
1/2
1/2
1/2
1/2
1/2
1/2
Full
3/4
3/4
3/4
1/2
02 Level0
High
Low
Low
High
Low
Low
Low
Low
High
Low
Low
Air
preheat
tempera-
ture, °F
Ambient
Ambient
=160
=150
=150
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Average emission data, dry basis
NO ,
ppm
354
376
385
119
130
420
429
400
380
196
169
N0y
at 3%
02, ppm
362
338
346
123
124
382
388
367
400
175
151
02, %
3.4
1.0
1.0
3.6
2.1
1.2
1.1
1.4
3.9
0.9
0.8
C02, %
13.3
15.0
15.3
12.9
13.7
15.6
14.7
15.4
13.1
11.4
11.6
CO, ppm
0
26
56
1
170
20
46
87
0
25
7
(continued)
-------
TABLE 2 (continued)
Test
block
NG-3
NG-4 -
NG-5
NG-6
NG-7
NG-8
NG-9
NG-10
Date
(1984) and
Time (24-h)
8/19
1041-1241
8/19
1440-1540
8/19
1626-1726
8/19
1837-1937
8/19
2052-2152
8/19
2227-2257
8/19
2315-2345
8/19-20
2351-0021
Boiler
ID
No. 4
No. 4
No. 4
No. 4
No. 5
No. 5
No. 5
No. 5
Operating parameters
Fuel9 type
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
N.G.
Loadb
3/4
Full
Full
1/2
1/2
3/4
3/4
3/4
02 Level0
High
High
Low
High
High
High
Low
Low
Air
preheat
tempera-
ture, °F
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
S160
Average emission data,6 dry basis
NO ,
ppm
151
146
211
122
100
118
176
209
N0y
at 3%
02, ppm
152
147
188
126
104
121
159
189
02, %
3.1
3.1
0.8
3.6
3.7
3.4
1.1
1.1
C02, %
10.0
10.1
11.6
9.9
10.0
10.5
11.7
11.5
CO, ppm
0
0
12
0
0
0
19
25
00
aFuel type: R.O. = No. 6 Residual oil; D.O. = No. 2 Distillate oil; N.G. = Natural gas
bBoiler load: 1/2 (=70,000 Ib/h steam); 3/4 (=105,000 Ib/h steam); Full (=140,000 Ib/h steam)
^Oxygen level for natural gas and residual oil tests: Low (S0.5-1.5%), High (>3.0%); for distillate oil tests:
Low (-2%), High (>3.0%).
Ambient air temperature, between 70°-90°F.
eAverage emission results for the indicated test block time interval, dry basis.
-------
regression equations established from the calibration data for each monitor.
For a 2-hour test block, this procedure yielded 24 data points; for a 1-hour
test block, 12 data points; and for a 30-minute test block, 6 data points.
Pollutant concentrations reported in Table 2 represent average values deter-
mined from the number of data points for each test block. The concentrations
reported in Table 3 represent the maximum and minimum values recorded during
the designated test periods.
Nitrogen oxide (NO ) concentrations are reported in parts per million by
A
volume on a dry basis. These concentrations have also been corrected to 3
percent CL as a standard by which emission trends can be evaluated. Oxygen
and carbon dioxide (CO-) concentrations are reported in percent by volume and
carbon monoxide (CO) concentrations are reported in parts per million by
volume, all on a dry basis.
In the evaluation of the data relative to the specific process param-
eters, NO concentrations corrected to 3 percent 09 were used. Analysis of
A £.
the test data showed several general trends:
1. Concentrations of NO were significantly higher at the same load
conditions when Boiler 5 was burning No. 6 residual oil rather than
No. 2 distillate oil. When corrected to 3 percent 02, NO concentra-
tions ranged from 338 to 362 ppm for the three residual oil tests
compared with values of 123 and 124 ppm for the two distillate oil
tests at half load.
2. Combustion air preheat temperature appeared to have no significant
impact on NO concentration during residual oil firing. For similar
boiler load and oxygen levels, the results from Test Blocks 2 and 3
showed an average NO concentration of 338 ppm at ambient preheat
temperature compared with 346 ppm at a preheat temperature of 71°C
(160°F), or less than a 2.5 percent difference. For natural gas,
however, the results of Test Blocks NG-9 and NG-10 indicate that NOX
concentrations increase by more than 10 percent with an increase in
preheat temperature. The average concentration at a preheat tempera-
ture of 71°C (160°F) was 189 ppm compared with 159 ppm at an ambient
preheat temperature, or a 16 percent difference.
3. When Boiler 5 was firing residual oil, NO concentrations were
generally greater at high excess-air leveTs. Under similar boiler
load conditions, the results from Test Blocks 1 (one-half load, high
excess air) and 2 (one-half load, low excess air) were 362 and 338
ppm, respectively. For Test Blocks 8 (three-fourths load, low 02)
and 9 (three-fourths load, high 02), NO concentrations were 367 and
400 ppm, respectively. For Test Blocks 4 and 5 (while Boiler 5 was
firing distillate oil), NO concentrations were 123 ppm for Test
Block 4 (one-half load, hitjh 02) and 124 ppm for Test Block 5 (one-
half load, low 02).
-------
TABLE 3. SUMMARY OF MAXIMUM AND MINIMUM CEM DATA BY TEST BLOCK
Test Block
RO-1
RO-2
RO-3
DO-4
DO-5
RO-6
RO-7
RO-8
RO-9
NG-1
NG-2
NG-3
NG-4
NG-5
NG-6
NG-7
NG-8
NG-9
NG-10
Date
(1984)
8/16
8/16
8/17
8/17
8/17
8/17
8/18
8/18
8/19
8/15
8/15
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
CEM Data3
NO , ppm
Max.
370
381
391
126
135
424
435
414
388
200
172
153
147
213
123
100
119
178
210
Min.
338
366
377
116
124
418
420
389
374
191
157
149
145
210
120
99
116
175
208
02, %
Max.
3.7
1.1
1.4
4.0
2.5
1.4
1.2
1.5
3.9
1.2
0.9
3.1
3.1
0.9
3.6
3.7
3.4
1.2
1.2
Min.
3.3
0.9
0.9
3.1
1.9
1.0
1.0
1.2
3.9
0.8
0.8
3.1
3.1
0.8
3.6
3.7
3.3
0.9
0.8
C02, %
Max.
13.7
15.2
16.1
13.1
14.0
15.8
14.9
15.6
13.2
11.7
12.0
10.4
10.2
11.8
10.1
10.0
10.5
11.8
11.7
Min.
12.7
14.4
14.4
12.2
13.4
15.4
14.4
15.3
13.0
11.1
11.0
9.7
9.9
11.1
9.6
9.9
10.5
11.7
11.3
CO, ppm
Max.
0
58
134
0
234
133
87
124
0
68
9
0
0
13
0
0
0
23
47
Min.
0
10
5
0
35
2
27
20
0
3
6
0
0
11
0
0
0
12
13
Dry basis.
10 .
-------
4. When firing natural gas, NO concentrations were generally higher at
low excess-air levels. For Test Blocks NG-4 (full load, high 02)
and NG-5 (full load, low 02), NO concentrations were 147 and 188
ppm, respectively. A similar pattern was observed for Test Blocks
NG-8 (three-fourths load, high 0?) and NG-9 (three-fourths load, low
02)9 during which NO concentrations were 121 and 159 ppm, respec-
tively.
5. When Boiler 5 was firing residual oil, the increase in NO concen-
tration was less than 10 percent between one-half boiler Toad and
three-fourths and full boiler loads. Test Blocks 1, 2, 3, and 6
(one-half loads) showed an average NO concentration of 357 ppm.
Test Blocks 7, 8, and 9 (three-fourths and full boiler loads) showed
an average NO concentration of 385 ppm. No full or three-fourths
load test data are available for distillate oil; therefore, compari-
sons based on boiler loads are not possible.
6. When natural gas was being fired, NO concentrations averaged 127
ppm for one-half boiler loads (Tests NG-2, 6, and 7), 159 ppm for
three-fourths boiler loads (Tests NG-1, 3, 8, 9, and 10), and 168
ppm for full boiler loads (Tests NG-4 and 5).
Additional data reduction for NO and 09 was performed for those periods
A £.
prior to actual test blocks when the boiler was in a changeover mode, either
from high 0^ to low 0^ or vice versa. The purpose of this was to establish a
clearer relationship between NO concentrations, excess-air levels, and fuel
/\
types. Tables 4 and 5 summarize these data.
TABLE 4. RELATIONSHIP BETWEEN NOY CONCENTRATION AND EXCESS
AIR LEVELS—RESlDUAL OIL
Test Block
Prior to Test 2
(one-half load,
R.O. , low 02)
Date
(1984)
8/16
8/16
8/16
8/16
8/16
8/16
Time
(24-h)
1615
1625
1638
1645
1655
1700
Average CEM concentrations
02, %
3.5
3.4
1.2
0.8
1.5
0.9
NO , ppm
As measured
340
340
360
374
337
366
At 3% 02
350
348
327
333
311
328
Only one set of data was available for residual oil firing, but these
data are generally comparable to the actual test block data, which show a
decrease in NO emissions at low excess-air levels. Seven sets of data were
A
11
-------
TABLE 5. RELATIONSHIP BETWEEN NOY CONCENTRATION AND EXCESS
AIR LEVELS—NATURAL GAS
Test Block
Prior to Test NG-1
(three-fourths
load, N.G. , low
02)
NG-1 (Process upset
from 13:35-16:15,
boiler parameters
readjusted)
Prior to Test NG-3
(three-fourths
load, N.G. , high
02)
Prior to Test NG-5
(full load, N.G.,
high 02)
Date
(1984)
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/15
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
Time
(24-h)
1230
1240
1250
1300
1310
1320
1330
1505
1515
1525
1535
1545
1555
1605
1615
0935
0945
0955
1005
1015
1025
1035
1041
1606
1610
1616
1626
Average CEM concentrations
02, %
3.2
2.9
2.6
1.8
1.1
1.2
1.2
3.7
3.1
2.8
2.4
1.9
1.7
1.6
0.9
1.1
1.4
1.8
2.1
2.2
2.6
2.9
3.0
3.2
2.7
1.8
0.9
NO , ppm
As measured
144
151
164
181
193
192
194
136
156
167
177
185
185
186
196
188
183
178
178
173
166
153
149
145
166
195
212
At 3% 02
147
150
160
170
174
174
176
142
157
165
171
174
172
172
175
170
168
167
169
166
162
152
149
147
163
183
190
(continued)
12
-------
TABLE 5 (continued)
Test Block
Prior to Test NG-6
(three-fourths
load, N.G. , high
02)
Prior to Test NG-7
(one-half load,
N.G., high 02)
During to Test NG-10
(three-fourths
load, N.G. , low
02)
Date
(1984)
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
Time
(24-h)
1821
1823
1825
1827
1831
1835
2020
2024
2026
2028
2035
2038
2040
2042
2046
2049
2345
2351
0001
0011
Average CEM concentrations
02, %
1.6
2.4
2.6
3.0
3.8
3.5
2.2
3.5
4.0
4.3
4.6
4.0
3.5
3.2
3.4
4.0
0.4
0.6
0.9
1.2
NO , ppm
As measured
146
143
142
140
117
122
127
105
95
90
83
93
101
110
105
92
199
210
210
208
At 3% 02
135
138
139
140
122
126
122
108
101
97
91
98
104
111
107
97
174
185 .
188
189
13
-------
available for the natural gas tests, and these data correlate with the actual
test block data, which show an increase in NO emissions at low excess-air
A
levels.
For data presention and informational purposes, the NO CEM data were
X
used to calculate mass emission rates in pounds per million Btu. This was
accomplished by converting the reported average parts per million of NO
/\
concentration (corrected to 3 percent 02) to pounds per dry standard cubic
foot and then multiplying this value by the appropriate F-factor based on fuel
type and an excess air correction factor. The F-factor relates the amount of
dry flue gas generated to the calorific value of the fuel combusted and is
expressed in dry standard cubic feet per million Btu heat input.
Subpart D of the Federal Register* lists an F-factor of 9220 dscf/106 Btu
for residual and distillate oil and an F-factor of 8740 dscf/10 Btu for
natural gas. These factors were used in all calculations. Table 6 summarizes
the NOV mass emission rate data for each test block. For the distillate oil
6
tests, NO emission rates were 0.16 lb/10 Btu. For the residual oil tests,
x /-
NO emission rates ranged from 0.43 to 0.51 lb/10 Btu and emission rates
x c
ranged from 0.13 to 0.23 lb/10 Btu for the natural gas tests.
2.3 PARTICULATE AND FUEL ANALYSIS TEST RESULTS
Particulate emission measurements were made on Boiler 5 while it was
firing residual and distillate oil. Two individual sampling trains were used
for simultaneous traversing of the cross-sectional area of the stack during
each of the nine test blocks. One sampling train was an EPA Reference Method
5** [sample collection temperature of 121°C (250°F)] and the other was an EPA
Reference Method 5B** [sample collection temperature of 160°C (320°F)]. The
analytical procedures for these methods are similar in all respects except
that the probe rinse and filter particulate catch for Method 5B are heated in
an oven for 6 hours at 160°C prior to gravimetric weighing.
Previous EPA studies at sources where the emission streams contained
sulfur oxides indicate significant positive biases can occur on particulate
40 CFR 60, Subpart D, New Source Performance Standards for Fossil-Fuel-
**Fired Steam Generators, July 1984.
40 CFR 60, Appendix A, Reference Methods 5 and 5B, J.uly 1984.
14
-------
TABLE 6. SUMMARY OF NOV EMISSION RATES
X
Test Block
RO-1
RO-2
RO-3
DO-4
DO-5
RO-6
RO-7
RO-8
RO-9
NG-1
NG-2
NG-3
NG-4
NG-5
NG-6
NG-7
NG-8
NG-9
NG-10
Date
(1984)
8/16
8/16
8/17
8/17
8/17
8/17
8/18
8/18
8/19
8/15
8/15
8/19
8/19
8/19
8/19
8/19
8/19
8/19
8/19
NOV concentration3
ppm at 3% 02
362
338
346
123
124
382
388
367
400
175
151
152
147
188
126
104
121
159
189
Ib/dscf x 10"b
4.32
4.04
4.13
1.47
1.48
4.56
4.62
4.37
4.78
2.09
1.80
1.82
1.76
2.25
1.51
1.24
1.45
1.90
2.26
NO mass .
emission rate,
lb/106 Btu
0.47
0.43
0.44
0.16
0.16
0.49
0.50
0,47
0.51
0.21
0.18
0.19
0.18
0.23
0.15
0.13
0.15
0.19
0.23
Conversion factor:
M
385.1 x 10"
where M = the molecular weight of N02(46).
bF-factor: 9220 dscf/10^ Btu for R.O. and D.O.
8740 dscf/10° Btu for N.G.
Ib NO /106 Btu = N0¥ cone. (Ib/dscf @ 3% 09)
/\ /\ £
x F-factor (dscf/106 Btu @ Stoichiometric 0~)
20.9
/Stoichiometric 02 \
20.9-3.0 ^dilution by excess air
15
-------
measurements as a result of the retention of condensible sulfate material in
the front half (probe and filter) of the standard Method 5 sampling train.
In most cases, sulfuric acid (f-LSO.) is the predominant sulfate species in
these emission streams. These same EPA studies have shown that increased
sampling temperatures coupled with a thermogravimetric analytical procedure
reduce, but do not eliminate, the effects (artifact formation) of H,,S04 col-
lected in the sampling train. Therefore, conducting both test methods makes a
direct comparison between methods possible with regard to potential biases
caused by condensible sulfate material.
Tables 7 and 8 summarize the particulate sampling conditions and emission
results, respectively. Sample volumes are reported in dry normal cubic meters
(dNm3). Volumetric flow rates are presented in actual cubic meters per minute
(m3/min), actual cubic feet per minute (acfm), dry normal cubic meters per
minute (dNm3/min), and dry standard cubic feet per minute (dscfm).
Particulate weights are reported in milligrams, concentrations are re-
ported in milligrams per dry normal cubic, and grains per dry standard cubic
feet. Mass emission rates are reported in kilograms per hour, pounds per
hour, and pounds per million Btu heat input. The mass emission rates in
kilograms/hour (pounds/hour) were calculated by multiplying individual par-
ticulate concentrations by the average flow rate measured during each specific
test block. Mass emission rates in pounds/10 Btu were determined by using
the particulate concentration, the 0~ level measured by the CEM system for the
specific test block, and the listed F-factor for residual and distillate oil,
9220 dscf/106 Btu.*
As reported in Table 7, sample volumes were generally consistent for each
train, and isokinetic sampling rates ranged between 92.1 and 104.7 percent,
which is within the acceptance range of 90 to 110 percent.
The probe and filter temperatures represent average values determined
from data recorded on the field data sheets. Method 5 filter temperatures
ranged from 121° to 141°C (250° to 284°F), and probe temperatures ranged
from 115° to 154°C (238° to 301°F), Method 5B filter temperatures ranged from
157° to 176°C (314° to 349°F), and probe temperatures ranged from 158° to
181°C (316° to 358°F).
*40 CFR 60, Subpart D, July 1984.
16
-------
TABLE 7. SUMMARY OF PARTICULATE SAMPLING CONDITIONS, BOILER NO. 5 STACK
Run No.
M5-1
M5B-1
MS -2
M5B-2
M5-3
MSB-3
MS -4
MSB -4
M5-5
HSB-S
M5-6
MSB-6
M5-7
M5B-7
M5-8
M5B-8
M5-9
M5B-9
Sample
type
MS
MSB
MS
MSB
MS
MSB
MS
MSB
MS
MSB
M5C
MSB
MS
MSB
MS
MSB
MS
MSB
Date
(1984)
8/16
8/16
8/16
8/16
8/16
8/16
8/17
8/17
8/17
8/17
8/17
8/18
8/18
8/18
8/18
8/19
8/19
Time
(24-h)
1219-1555
1217-1553
1732-2235
1733-2236
0035-0310
0036-0311
1128-1435
1129-1436
1649-1926
1650-1925
2210-0035
1728-1902
1729-1901
2050-2212
2051-2213
0025-0150
0026-0151
Samplea
volume,
dNm3
3.34
3.19
1.78
1.71
1.60
1.68
1.90
1.92
1.65
1.65
1.61
1.71
1.71
1.33
1.43
1.58
1.63
Sampl ing
time,
min
120
120
120
120
120
120
120
120
120
120
120
72
72
72
72
72
72
Isokinetic
sampl ing
rate, %
99.5
92.1
104.7
104.3
101.7
99.5
101.5
103.2
104.3
103.4
101.9
98.4
97.9
99.0
98.5
99.0
100.1
Sampling temperature
Probe
"C (°F)
130 (266)
172 (342)
121 (250)
158 (316)
120 (249)
163 (326)
115 (238)
176 (350)
160 (320)
168 (335)
161 (322)
154 (310)
181 (358)
149 (301)
164 (327)
137 (278)
162 (323)
Filter
°C( °F)
123 (253)
172 (341)
125 (257)
164 (328)
127 (261)
157 (314)
135 (275)
167 (333)
121 (250)
163 (325)
169 (336)
140 (284)
176 (349)
130 (266)
169 (337)
128 (263)
167 (332)
b
Flue gas conditions
Volumetri flow rate
mVinin (acfm)
971 (34,300)
869 (31.100)
833 (29,400)
960 (33.900)
823 (29,100)
810 (28,600)
1704 (60,200)
1281 (45,250)
1524 (53.800)
dNmVmin (dscfm)
579 (20.400)
544 (18,000)
492 (17.400)
572 (20,200)
486 (17,200)
481 (17.000)
889 (31,400)
713 (25,200)
823 (29.100)
Temperature,
°C (°F)
157 (315)
157 (314)
149 (300)
157 (313)
153 (307)
147 (296)
214 (417)
182 (361)
201 (393)
Moisture
content, %
12.3
14.5
14.7
12.3
13.6
14.3
12.7
12.8
12.0
aSample volume in dry normal cubic meters per minute. Standard conditions: 760 mraHg (29.92 in.Hg); 20°C (68°F); and 0 percent moisture.
bVolumetr1c flow rate, gas temperature, and moisture content data represent the average values measured by each Individual sampling train.
cRun MS-6 exhibited an excessive leak rate at the sample port change, and the run is considered void.
-------
TABLE 8. SUMMARY OF PARTICULATE EMISSION RESULTS
Test
block
1
2
3
4
5
6
7
8
9
Date
(1984)
8/16
8/16
8/16
8/17
8/17
8/17
8/18
8/18
9/19
Sample
ID
M5-1
M5B-1
M5-2
M5B-2
M5-3
M5B-3
M5-4
M5B-4
M5-5
M5B-5
M5-6a
M5B-6
M5-7
M5B-7
M5-8
M5B-8
M5-9
M5B-9
Analytical results
Particulate
weight, mg
Probe
30.7
9.1
18.2
4.0
16.9
6.0
13.8
6.3
15.8
6.0
6.8
58.0
40.3
34.8
13.6
32.0
12.1
Filter
228.9
169.6
127.7
103.2
100.9
97.8
9.4
13.6
4.1
4.0
98.4
282.7
211.9
111.3
100.4
118.0
101.1
Total
258.9
178.7
145.9
107.2
117.8
103.8
23.2
19.9
19.9
10.0
105.2
340.7
252.2
146.1
114.0
150.0
113.2
Concentration
mg/dNm3
77.5
56.1
82.1
62.8
75.8
62.0
12.2
10.4
12.1
6.1
65.5
198.9
148.0
109.7
79.9
94.9
69.5
gr/dscf
0.034
0.0245
0.036
0.027
0.033
0.027
0.005
0.0045
0.005
0.003
0.029
0.087
0.065
0.048
0.035
0.0415
0.030
Mass emission rate
kg/h
2.7
2.0
2.6
1.9
2.1
1.9
0.42
0.35
0.35
0.18
1.7
10.6
7.9
4.5
3.5
4.7
3.5
Ib/h
5.9
4.3
5.7
4.2
4.7
4.2
0.93
0.78
0.78
0.39
4.2
23.5
17.3
10.0
7.8
10.3
7.6
lb/10b Btu
0.05
0.04
0.05
0.04
0.05
0.04
0.01
0.01
0.01
0.01
0.04
0.12
0.07
0.07
0.05
0.07
0.05
-------
The reported flue gas volumetric flow rates, temperatures, and moisture
contents represent the average of the individual train measurements. In each
case, individual train measurements compared within ±10 percent.
Test Blocks 1, 2, 3, and 6 were conducted at approximately one-half
boiler load during the firing of residual oil. Average flue gas flow rates
ranged from 810 m2/min (28,600 acfm) during Test 6 to 971 m3/min (34,300 acfm)
during Test 1. ' Temperatures ranged from 147° to 157°C (296° to 315°F), and
moisture content ranged from 12.3 to 14.7 percent. Particulate concentrations
and mass emission rates as measured by the Method 5 sampling trains for these
test blocks ranged from 75.8 mg/dNm3 (0.033 gr/dscf) and 2.1 kg/h (4.7 lb/h;
0.05 lb/106 Btu) to 82.1 mg/dNm3 (0.036 gr/dscf) and 2.6 kg/h (5.7 Ib/h; 0.05
lb/10 Btu). The corresponding Method 5B test results showed particulate
concentrations and mass emission rates ranging from 56.1 mg/dNm3 (0.0245
gr/dscf) and 2.0 kg/h (4.3 lb/h; 0.04 lb/106 Btu) to 65.5 mg/dNm3 (0.029
gr/dscf) and 1.7 kg/h (4.2 lb/h; 0.04 lb/106 Btu). Run M5-6 was void because
of an excessive mid-test leakage rate; therefore, results are not reported.
Test Blocks 8 and 9 were conducted at approximately three-fourths boiler
load while Boiler 5 was firing residual oil. Average flue gas flow rates
ranged from 1281 m3/min (45,250 acfm) during Test 8 to 1524 m3/min (53,800
acfm) during Test 9. Gas temperature and moisture content were 182°C (361°F)
and 12.8 percent during Test 8 and 201°C (393°F) and 12.0 percent during Test
9. Method 5 results during Test 8 showed a particulate concentration of 109.7
mg/dNm3 (0.048 gr/dscf) and a mass emission rate of 4.5 kg/h (10.0 lb/h; 0.07
lb/106 Btu). The corresponding Method 5B results were 79.9 mg/dNm3 (0.035
gr/dscf) and 3.5 kg/h (7.8 Ib/h; 0.05 lb/106 Btu). During Test 9, Method 5
showed a concentration of 94.9 mg/dNm3 (0.0415 gr/dscf) and an emission rate
of 4.7 kg/h (10.3 lb/h; 0.07 lb/106 Btu); Method 5B results were 69.5 mg/dNm3
(0.030 gr/dscf) and 3.5 kg/h (7.6 Ib/h; 0.05 lb/106 Btu), respectively.
Test Block 7 was conducted at full boiler load while Boiler 5 was firing
residual oil. The average gas flow rate, temperature, and moisture content
during this run were 1704 m3/min (60,200 acfm), 214°C (417°F), and 12.7 per-
cent, respectively. Method 5 results showed a particulate concentration of
198.9 mg/dNm3 (0.087 gr/dscf) and a mass emission rate of 10.6 kg/h (23.5
lb/h; 0.12 lb/106 Btu). The Method 5B results were 148.0 mg/dNm3 (0.065
gr/dscf) and 7.9 kg/h (17.3 lb/h; 0.07 lb/106 Btu).
19
-------
Test Blocks 4 and 5 were conducted at approximately one-half boiler load
while Boiler 5 was firing distillate oil. For Test 4, the flue gas flow rate
averaged 960 m3/min (33,900 acfm), temperature averaged 157°C (313°F), and
moisture content averaged 12.3 percent. Method 5 results showed a particulate
concentration of 12.2 mg/dNm3 (0.005 gr/dscf) and a mass emission rate of 0.42
kg/h (0.93 Ib/h; 0.01 lb/106 Btu). The Method 5B results were 10.4 mg/dNm3
(0.0045 gr/dscf) and 0.35 kg/h (0.78 Ib/h; 0.01 lb/106 Btu). For Test 5, the
gas flow rate, temperature, and moisture content averaged 823 m3/min (29,100
acfm), 153°C (307°F), and 13.6 percent, respectively. Method 5 results showed
a particulate concentration of 12.1 mg/dNm3 (0.005 gr/dscf) and a mass emis-
sion rate of 0.35 kg/h (0.78 Ib/h; 0.01 lb/106 Btu). The Method 5B results
were 6.1 mg/dNm3 (0.003 gr/dscf) and 0.18 kg/h (0.39 Ib/h; 0.01 lb/106 Btu).
Table 9 presents a summary of the between-run statistical data for simi-
lar sample types and boiler loads. For each set of data, the table lists the
mean particulate concentration, the standard deviation with N-l weighting for
sampling data, and the percentage coefficient of variance (CV), which ex-
presses the standard deviation as a percent of the mean concentration. The
number of data points included in each calculation is shown for consideration
in data evaluation. The data show that regardless of the difference in con-
centrations as determined by Methods 5 and 5B, the between-run data for simi-
lar load conditions and sample types are reproducible.
Table 10 summarizes the fuel analytical results for residual and distil-
late oil samples collected during the particulate test program. Samples of
the oil were collected immediately prior its entering the boiler burner sys-
tem. For Test Blocks 1 through 3 and 7 through 9, samples of residual oil
were collected at the indicated times, and a single composite sample was
analyzed by applicable ASTM Methods. Individual samples were collected for
Test Blocks 4 through 6.
The residual oil samples showed an ash content ranging from 0.052 to 0.14
percent, a sulfur content ranging from 2.18 to 2.42 percent, and a nitrogen
content ranging from 0.34 to 0.43 percent. The distillate oil samples showed
an ash content of less than 0.02 percent, an average sulfur content of 0.55
percent, and a nitrogen content of less than 0.02 percent.
20
-------
TABLE 9. BETWEEN-RUN STATISTICAL DATA FOR SIMILAR SAMPLE
TYPES AND BOILER LOADS
Run No.
M5-1
M5-2
M5-3
M5B-1
M5B-2
M5B-3
M5B-6
M5-8
M5-9
M5B-8
M5B-9
M5-4
M5-5
M5B-4
M5B-5
Boiler load
1/2
1/2
1/2
1/2
1/2
1/2
1/2
3/4 .
3/4
3/4
3/4
1/2 - Distillate oil
1/2 - Distillate oil
1/2 - Distillate oil
1/2 - Distillate oil
Participate
concentra-
tion, mg/dNm3
77.5
82.1
75.8
56.1
62.8
62.0
65.5
109.7
94.9
79.9
69.5
12.2
12.1
10.4
6.1
No. of
data
points
3
4
2
2
2
2
Statistical data
x,a
mg/dNm3
78.5
61.6
102.3
74.7
12.15
8.25
b
o»
mg/dNm3
3.26
3.96
10.5
7.35
0.07
3.04
CV,C %
4.2
6.4
10.2
9.8
0.6
37.0
Mean participate concentration.
Between-run standard deviation with N-l weighting for sampling data.
Coefficient of variance is the standard deviation expressed as a percentage of
the mean.
21
-------
TABLE 10. SUMMARY OF FUEL ANALYTICAL RESULTS
Laboratory ID
Composite No. 1
No. 2 DS064
No. 3 DS065
Mo. 4 DS063
Composite No. 5
Date
(1984)
8/16
8/16
8/16
8/17
8/17
8/17
8/17
8/18
8/18
8/19
Time
(24-h)
1130
1908
2200
0330
1430
1930
2355
1940
2250
0225
Fuel type
No. 6 residual oil
No. 2 distillate oil
No. 2 distillate oil
No. 6 residual oil
No. 6 residual oil
Fuel analysis, % by weight
Ash
0.14
0.018
0.013
0.052
0.082
Carbon
85.75
85.24
85.23
85.57
83.64
Sulfur
2.18
0.53
0.57
2.18
2.42
Hydrogen
7.08
11.97
12.84
11.43
9.92
Oxygen
4.42
2.24
1.35
0.42
3.51
Nitrogen
0.43
<0.02
<0.02
0.34
0.42
Heating
value,
Btu/lb
18,690
19,280
18,660
18,660
18,610
ro
ro
-------
The reported analytical results and heating values generally correspond
to the range of values expected for these specific fuel types according to
reference material.*
Analysis of the particulate test results revealed the following readily
evident factors:
1. Regardless of the test type, particulate emissions increased signif-
icantly as boiler load increased.
2. Particulate emissions generated by the firing of residual oil were
substantially higher than those generated by the firing of distil-
late oil.
3. The particulate sample results (concentration basis) produced by EPA
Method 5B were consistently lower by an average of approximately 30
percent than the within-run Method 5 sample results.
Items 1 and 2 were as expected given previous test experience and the
fuel analysis results presented in Table 10. With regard to Item 3, past EPA
studies (as mentioned previously) have shown the effects of condensible
sulfate biases on particulate measurements from sources with emission streams
laden with sulfur oxides.
The sulfur data in the fuel analysis would lead one to expect sulfate
species in the gas emission stream from this boiler. Sulfuric acid (hLSO,)
and/or its metal and ammonium salts are the most probable forms of water-
soluble sulfates that exist in many complex chemical forms. Of these, HpSO^
is the most common.
The EPA studies cited in the report have shown that increased sample
temperature coupled with a thermogravimetric analysis (i.e., Method 5B) will
minimize the positive biases on particulate measurements caused by the reten-
tion of ^SO. and its associated water in the front half of the standard
Method 5 sampling train.
These same studies have shown that the greater the amount of ^SO. re-
tained in the probe and in the filter, the greater the positive bias in par-
ticulate measurements. Increased hLSO. retention also increases the potential
for artifact formation. The data suggested that the greater the amount of
*
The Babcock and Wilcox Company. Useful Tables for Engineers. 13th ed.
1979. p. 58.
23
-------
condensible sulfate (HpSCh) initially collected, the more that remained on the
samples, regardless of treatment temperature. The analytical data indicated
that HUSCL and its associated water were significantly reduced by heating
sample fractions to at least 160°C (320°F) prior to gravimetric analysis.
Observed weight losses at higher treatment temperatures were primarily attrib-
utable to the volatilization of residual HUSO, and other water-soluble sul-
fates not removed by heating at 160°C.
According to current understanding of the thermogravimetric principle,
only sulfate present as HpSO^ and its associated water is removed in the heat
procedure. Other water-soluble sulfates would not be removed at 160°C.
Therefore, if H^SO, is the predominant sulfate species, the Method 5B results
would be expected to be lower than the Method 5 results, and correcting for
total water-soluble sulfate should account for a significant portion of the
differences in weights.
In an effort to characterize this difference in test results, the follow-
ing additional analyses were performed.
1. A single weighing of each Method 5B sample fraction was performed
prior to heating to 160°C. In this way, the weight loss was evalu-
ated as a function of heat treatment temperature.
2. A single Method 5 test (M5-9) was subjected to heat treatment at
160°C to determine sample weight loss for comparison with the with-
in-run Method 5B data (M5B-9).
3. The probe rinse and filter samples from Runs 1 through 9 were ana-
lyzed for residual sulfate (SOi* ). Each filter was cut into small
pieces and placed in a 125-ml flask equipped with a stopper and air
condenser. About 50 ml of deionized, distilled water was added, and
the contents were gently refluxed for 6 to 8 hours. The solution
was then cooled and diluted with water to exactly 250 ml in a volu-
metric flask. A 15-ml aliquot of the settled solution was then
pipetted in a separate container. The sulfate ion (SOit") concentra-
tion in each aliquot was then determined by ion chromatography. The
results of this analysis yielded total milligrams of S04 , or total
water-soluble sulfate, on the filter samples. The probe rinse
residue was diluted to 150 ml with deionized, distilled water and
the contents were heated to approximately 90°C and gently stirred
for 2 to 3 hours. This solution was cooled and diluted with deion-
ized, distilled water to 250 ml in a volumetric flask. A 15-ml
aliquot of the solution was then pipetted into a separate container
and the sulfate ion (SOO concentration was determined by ion
chromatography.
24
-------
4. The back-half impinger solutions (H?0) from Runs 3 and 7 were ana-
lyzed by ether/chloroform extraction to determine condensible organ-
ic content for evaluation of the potential contribution of condens-
ible organics to the difference in reported weights.
As shown in Table 11, the Method 5B sample fractions exhibited weight
losses relative to their ambient weights ranging from 30 to 83 percent in the
probe fraction and 6 to 20 percent in the filter fraction. Also shown in this
table is a comparison between the ambient Method 5B sample weights, and con-
centrations and the within-run Method 5 results reported in Table 8. For Runs
1, 7, 8, and 9, the Method 5 concentrations were within 10 percent of the
within-run Method 5B data versus approximately a 30 percent difference after
the Method 5B fractions were heated. For Runs 2, 3, 4, and 5, ambient Method
5B concentrations were actually greater than the within-run Method 5 results.
The ambient Method 5B weights were obtained immediately prior to heating.
Only one weight was obtained for each fraction after a minimal desiccation
period. Some uncombined water may have been retained on these sample frac-
tions, which could cause a positive bias in ambient weights. Thus, exact
comparisons between the ambient Method 5B and Method 5 data are probably
distorted to some degree. The data do show, however, that condensible matter
is retained in the front half of the sampling train regardless of sample
temperature and that the positive effects of the condensible matter on partic-
ulate measurements are reduced by heating the sample fractions to 160°C.
As an additional check, sample fractions from Run M5-9 were heated at
160°C for 6 hours and then cooled and weighed to a constant weight. The probe
and filter fractions showed relative weight losses of 49 and 20 percent,
respectively. The Method 5 concentration after heating was 70.0 mg/dNm3
versus 69.5 mg/dNm3 for the within-run Method 5B sample.
Table 12 presents the results of the residual sulfate analyses performed
on the rinse and filter fractions for each run. As shown, both methods con-
tained a significant amount of water-soluble sulfate (S0.~), particularly the
filter fractions.
Because of limited distillate oil test data and the fact that the par-
ticulate concentrations measured were significantly less than the residual oil
test data, our discussion on the differences in weights between methods is
limited to the residual oil test data.
25
-------
TABLE 11. METHOD 5B RELATIVE PERCENT WEIGHT LOSS AT 160°C
ro
en
Test No.
M5B-1
M5-1
M5B-2
M5-2
M5B-3
M5-3
M5B-4
M5-4
M5B-5
M5-5
MSB-6
M5-6
M5B-7
M5-7
M5B-8
M5-8
MSB -9
M5-9
Ambient temperature3
Sample
weight, mq
Probe
20.4
30.7
23.4
18.2
20.6
16.9
17.4
13.8
22.1
15.8
19.5
57.4
58.0
28.9
34.8
28.7
32.0
Filter
202.8
228.2
123.5
127.7
118.3
100.9
14.4
9.4
4.3
4.1
123.4
261.1
282.7
119.1
111.3
119.3
118.0
Total sample
weight, mg
223.2
258.9
146.9
145.9
138.9
117.8
31.8
23.2
26.4
19.9
142.9
318.5
340.7
148.0
146.1
148.0
150.0
Concen-
tration,
mg/dNm3
70.0
77.5
85.9
82.0
82.7
73.6
16.6
12.2
16.0
12.1
88.8
186.3
199.2
103.5
109.8
90.8
94.9
160°Cb
Sample
weight, mg
Probe
9.1
4.0
6.0
6.3
6.0
6.8
40.3
13.6
12.1
16.2
Filter
169.6
103.2
97.8
13.6
4.0
98.4
211.9
100.4
101.1
94.4
Total sample
weight, mg
178.7
107.2
103.8
19.9
10.0
105.2
252.2
114.0
113.2
110.6
Concen-
tration,
mg/dNm3
56.0
62.7
61.8
10.4
6.1
65.3
147.5
79.7
69.5
70.0
Relative percent
weight loss, %
Probe
55
83
71
64
73
65
30
53
58
49
Filter
16
16
17
6
7
20
19
16
15
20
aThe Method 5B sample fractions were desiccated less than 2 hours prior to heating at 160°C. One weight
was obtained for each fraction. The Method 5 sample fractions were weighed following procedures
described in the Federal Register (40 CFR 60, Appendix A, July 1984).
The reported Method 5B results represented the sample weights obtained after heating each fraction to
160°C for 6 hours with a 2-hour desiccation and cooling period per Method 5B.
cupinht in« - Ambient weight (mq) - 160°C weight (mg) 1QO
weignt loss Ambient (mg) * luu>
-------
TABLE 12. RESIDUAL SULFATE (SO/) ANALYSIS OF
METHODS 5 AND 5B SAMPLE FRACTIONS
Sample
ID
M5-1
M5B-1
M5-2
M5B-2
M5-3
M5B-3
M5-4
M5B-4
M5-5
M5B-5
M5-6
M5B-6
M5-7
M5B-7
M5-8
M5B-8
M5-9
M5B-9
Sampling
type
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Method 5
Method 5B
Residual sulfate as S0.~,a
Probe
rinse, mg
12.3
7.2
3.2
2.6
2.5
2.0
2.6
2.8
2.5
2.1
2.3
11.4
5.2
5.1
2.8
9.6
6.3
Filter,
mg
105.8
100.9
65.0
55.1
47.8
51.6
3.5
2.3
4.1
2.2
54.7
92.9
67.5
50.8
48.0
57.5
47.3
Total ,
mg
118.1
108.1
68.2
57.7
50.3
53.6
6.1
5.1
6.6
4.3
57.0
104.3
72.7
55.9
50.8
67.1
53.6
Particulate concen-
tration, mg/dNm3
Uncorrected
77.5
56.1
82.1
62.8
75.8
62.0
12.2
10.4
12.1
6.1
65.5
198.9
148.0
109.7
79.9
94.9
69.5
Corrected
42.2
22.1
43.7
28.9
42.2
29.9
9.0
7.7
8,1
3.5
29.9
138.2
105.0
67.8
44.2
52.5
36.6
Analysis by ion chromatography. Represents total water-soluble sulfate as SO/
The uncorrected particulate concentrations are as reported in Table 8. The
corrected concentrations were calculated by subtracting the quantity of
residual sulfate from the original sample weights (Table 8) and dividing by
the appropriate sample volume.
27
-------
For the residual oil tests, the quantity of water-soluble sulfate found
on the Method 5 samples ranged from 50.3 to 118.1 mg, which represents approx-
imately 31 to 47 percent of the total particulate collected (Table 8). The
Method 5B samples exhibited similar characteristics with water-soluble sulfate
quantities ranging from 50.8 to 108.1 mg or approximately 29 to 60 percent of
the total particulate after heating to 160°C.
Two distinct differences between the previous studies cited " and this
project are as follows:
1. In the previous studies, all Method 5 samples were subjected to a
thermogravimetric analysis prior to extraction and determination of
residual sulfate. In this study, the Method 5 samples were ex-
tracted after desiccation at ambient temperature and pressure.
2. In the previous studies, the majority of the Method 5 particulate
catch at ambient conditions was in the probe rinse and this fraction
exhibited a significantly higher relative percent weight loss when
heated to 160°C. In contrast, the majority of the Method 5 particu-
late catch at ambient conditions was on the filter fraction.
As mentioned previously, if HUSO, is the predominant sulfate species in
the gas stream, the Method 5B particulate weights after heating to 160°C
(Table 8) would be expected to be lower than the Method 5 results (which they
are), and correcting for total water-soluble sulfate should account for a
significant portion of the differences in weights (which it does not, Table
12). Since the quantity of water-soluble sulfate is comparable for each
method, correcting the particulate concentrations from Table 8 for residual
sulfate does not significantly reduce the difference in results between the
two methods.
Due to the fact that the Method 5 samples were not heated prior to ex-
traction, the remaining difference in weights between the two methods (Table
12) may be attributable to water associated with the sulfate in the Method 5
sample.
The comparability of the Method 5 and 5B SO. results actually suggests
that hLS04 is not the predominate sulfate species. If this were the case, the
Method 5B sulfate results should be.considerably lower than the Method 5
results since HUSO* would be significantly reduced by the thermogravimetric
analytical procedure of Method 5B. The difference in SO." results for Runs 1,
2, 7, 8, and 9 indicate that some HUSO, is present on the sample fraction but
28
-------
the predominate sulfate species are probably metal sulfates. This could ex-
plain the difference in participate distribution between the probe and filter
fractions observed in this study compared to the previous studies cited where
the majority of the particulate and HpSO, was collected in the probe rinse
fraction. In this study, the particulate catch in the probe was small com-
pared to that collected on the filter.
This test data suggest that the difference in weights between the two
methods (Table 8) is attributable to H^SO. and its associated water and/or
water associated with metal sulfates in the particulate.
The comparability and quantity of the organic data for Runs 3 and 7
(Table 13) suggest no significant bias resulted from condensible organics.
Considering the sample collection temperature, one would expect to find a
greater organic content and difference between the two sample types than
demonstrated here if condensible organics contributed significantly to the
difference in measured concentrations.
TABLE 13. ANALYTICAL RESULTS FROM ETHER/CHLOROFORM EXTRACTION
OF BACK-HALF SOLUTIONS
Test No.
M5-3
M5B-3
M5-7
M5B-7
Back-half condensible
analysis, mg
Organic
5.9
3.1
4.0
2.6
Inorganic
87.2
73.8
87.5
47.3
Concentration, mg/dNm3
Organic
3.8
1.85
2.3
1.5
Inorganic
56.1
44.1
51.1
27.8
29
-------
SECTION 3
QUALITY ASSURANCE
The objective of testing is to produce representative emission results;
therefore, quality assurance is one of the main facets of stack sampling.
Quality assurance guidelines provide the detailed procedures and actions
necessary for defining and producing acceptable data. Four such documents
were used in this test program to ensure the collection of acceptable data and
to provide a definition of unacceptable data. The following documents com-
prised the source-specific test plan prepared by PEI and reviewed by the
Emission Measurement Branch of the EPA; the EPA Quality Assurance Handbook
Volume III, EPA-600/4-77-027; the PEI Environmental Emission Test Quality
Plan; and the PEI Environmental Laboratory Quality Assurance Plan. The last
two, which are PEI's general guideline manuals, define the company's standard
operating procedures and are followed by the emission testing and laboratory
groups.
Relative to this specific test program, the following steps were taken to
ensure that the testing and analytical procedures produced quality data.
0 Calibration of all field sampling equipment.
0 Checks of train configuration and calculations.
0 Onsite quality assurance checks, such as sampling train, pitot tube,
and Orsat line leak checks and quality assurance checks of all test
equipment prior to use.
0 Use of designated analytical equipment and sampling reagents.
0 Internal and external audits to ensure accuracy in sampling and
analysis.
Quality Assurance activities for each specific phase of this project are
summarized in the following subsections:
30
-------
3.1 CONTINUOUS EMISSION MONITORS
Each CEM system was set up and operated according to specifications
outlined in the monitor operating manuals. Performance specifications (zero
drift, span drift, and response time) outlined in 40 CFR 60, Appendix B,
Performance Specifications 2 and 3, were followed throughout this test pro-
gram. Prior to actual stack gas monitoring, a pollutant profile was estab-
lished by traversing the stack cross section and comparing individual sample
point values for NO and 0/> against a reference point (stack centroid); this
A £.
permitted a determination of possible gas stratification in the stack. A
difference of less than 10 percent between individual sampling points and the
reference data point indicated no significant stratification problem existed
at the sampling locations. Tables 14 and 15 show stratification results for
Boilers 4 and 5, respectively.
At the beginning of each test day, each monitoring system was leak-
checked and system checks for zero drift, span drift, and response time were
conducted. The performance specification tests followed were established for
"continuous on-line" analyzers in operation for long periods of time. The
tests applied to monitors in this test series were used as general checks to
ensure reasonable response times and minimal drifts from day-to-day testing.
Tables 16 through 24 summarize the results of the checks for zero drift,
span drift, and response time. The data in Tables 16 through 19 represent
summary data for 24-hour zero and span drift checks. All drift checks were
well within the expected operating ranges of the monitors and showed consist-
ent analyzer response from day-to-day operation. Response time checks are
shown in Tables 20 through 24. All monitors had response times of less than
three minutes for both high-level calibration gas and stack effluent readings.
Both response times and drift checks show consistent monitor operation
throughout this test program.
A three-point calibration was performed on each monitoring system to
cover the low, mid, and high values of the specific pollutant concentration
measured. This system check was conducted at the beginning and end of most
test blocks. Single-point calibration checks were performed between test
blocks when sufficient time was not available for a three-point calibration
check. Calibration gases were transported through the sample-conditioning
31
-------
TABLE 14. MONITOR STRATIFICATION TEST—BOILER 4 (8/14/84)
Traverse
Point No.
West Port
1
2
3
4
North Port
1
2
3
4
NO
concentra-
tion, ppm
_
133
132
138
142
_
140
142
141
143
Reference
NO con-
centra-
tion, ppm
134
132
132
136
142
140
142
141
141
146
NO Devia-
tion, %
_
+0.75
0
+1.5
0
_
-1.4
+0.71
0
-2.1
02 con-
centra-
tion, %
_
3.2
3.2
3.1
3.3
_
3.2
3.1
3.2
3.2
Reference
02 con-
centra-
tion, %
3.2
3.2
3.2
3.3
3.2
3.1
3.1
3.1
3.1
3.0
02 Devia-
tion,0 %
_ i
0
0
-6.1
+3.1
^
+3.2
0
+3.2
+6.7
Reference point is the sampling point located in the center of the sample matrix.
•Went deviation . Tr"erie^g.^gnce- PP" x 100.
TABLE 15. MONITOR STRATIFICATION TEST—BOILER 5 (8/16/84)
Traverse
Point No.
West Port
1
2
3
4
North Port
1
2
3
4
NO
concentra-
tion, ppm
_
352
347
342
329
.
342
342
341
327
Reference
NO con-
centra-
tion, ppm
352
352
352
345
331
337
347
347
339
327
NO Devia-
tion, %
_
0
-1.4
-0.9
-0.6
.
+1.5
-1.4
+0.6
0
02 con-
centra-
tion, %
_
1.5
1.3
1.3
1.6
^
3.6
3.5
3.5
3.5
Reference
02 con-
centra-
tion, %
1.3
1.3
1.3
1.3
1.2
3.5
3.6
3.5
3.5
3.5
02 Devia-
tion,0 %
_
+15.4
0
0
+25
^
0
0
0
0
eference point is the sampling point located in the center of the sample matrix,
ercent deviation .
""
x 100.
32
-------
TABLE 16. TEST RESULTS FOR NO MONITOR 24-HOUR ZERO AND CALIBRATION DRIFT
(ppm NO except as indicated)
X
Test No.
1
2
3
4
5
Date (1984)
Start
8/11
8/14
8/16
8/17
8/18
End
8/12
8/15
8/17
8/18
8/19
Test time
Start
1215
0800
1500
0830
0800
End
0930
0800
1500
0800
0900
Zero reading
Start End
(A)
-0.5
2.4
2.4
0.7
0.5
I(B)
0
0
0.7
0.5
0
Arithmetic mean (AM)
95% confidence interval (CIg5)
24-hour drift,3 %
Zero
drift
(C=B-A)
0.5
-2.4
-1.7
-0.2
-0.5
0.9
1.35
0.23
Span reading
Start End
(D)
222
227
426
441
470
IE)
225
223
444
470
427
Span
drift
(F=E-D)
3
-4
18
29
-43
Cali-
bration
drift
(F-C)
2.5
1.6
19.7
29.2
-42.5
19.1
34.2
5.3
co
co
24-hour drift =
CI
1QQO
x 100.
Zero drift
Calibration drift
CIQ, = 2'776 x| 5 (9.19) - (18.5) = 1.35
yb 5 xTT
95
= 2-776 J 5 (3056) - (110.3) = 34.2
~~
-------
TABLE 17. TEST RESULTS FOR CO MONITOR 24-HOUR ZERO AND CALIBRATION DRIFT
(ppm CO except as indicated)
Test No.
1
2
3
4
5
6
7
Date
Start
8/11
8/12
8/13
8/14
8/15
8/17
8/18
1984)
End
8/12
8/13
8/14
8/15
8/16
8/18
8/19
Test time
Start
1215
0930
0820
0800
0800
0830
0800
End
0930
0820
0800
0800
1100
0800
1940
Zero reading
Start End
(A)
0
-0.2
0
0.5
-0.2
0
0.5
(B)
-0.2
0
0.5
-0.2
0
0.5
0
Arithmetic mean (AM)
95% confidence interval (CIg5)
24-hour drift,9 %
Zero
drift
(C=B-A)
-0.2
0.2
0.5
-0.7
0.2
0.5
-0.5
0.4
0.44
0.40
Span reading
Start End
(D)
202
201
204
203
201
203
206
(E)
201
204
203
201
203
206
204
Span
drift
(F=E-D)
-1
3
-1
-2
2
3
-2
Cali-
bration
drift
(F-C)
-0.8
2.8
-1.5
-1.3
1.8
2.5
-1.5
1.74
1.77
1.14
co
-P.
a% 24-hour drift
Zero drift
CI95 =
|AM| + CI95
x 100
ZTtfc 1
• \l 7 ( 1 36) (0 04) ~ 0 44
7 xT6~
Calibration drift
rT _ 2.365 1 7
cm , N '
yo 7 c
/ \^ b
(24.16) - (4) = 1.77
-------
TABLE 18. TEST RESULTS FOR 02 MONITOR 24-HOUR ZERO AND CALIBRATION DRIFT
(% 02 except as indicated)
Test No.
1
2
3
4
5
6
7
Date
Start
8/11
8/12
8/13
8/14
8/16
8/17
8/18
(1984)
End
8/12
8/13
8/14
8/15
8/17
8/18
8/19
Test time
Start
1215
0930
0820
0800
1500
0830
0800
End
0930
0820
0800
0800
1500
0800
0900
Low-range
reading
Start End
(A)
0.99
0.96
0.94
0.96
1.02
1.00
0.97
(B)
0.96
0.94
0.96
0.97
0.99
0.97
0.97
Arithmetic mean (AM)
95% confidence interval (CIg5)
24-hour drift, b %
Low-
range
drift
(C=B-A)
-0.03
-0.02
0.02
0.01
-0.03
-0.03
0
0.02
0.02
0.04
Span reading
Start End
(D)
8.08
8.05
8.01
8.03
7.93
7.93
7.70
(E)
8.05
8.01
8.03
8.04
7.93
7.70
7.79
Span
drift
(F=E-D)
-0.03
-0.04
0.02
0.01
0
-0.23
0.09
Cali-
bration
drift
(F-C)
0
-0.2
0
0
0.3
-0.2
0.9
0.05
0.08
0.13
CO
in
aThe low-range calibration gas (1.003 % 02) data were substituted for the zero drift check. The data test
monitor is set up on a calibration gas basis (no actual zero); zero readings are dependent on the low-range
calibration gas used.
\ 24-hour drift as % 02 = |AM| + CIg5.
Zero drift
Calibration drift
CIQ, = 2'365 J 7 (0.0036) - (0.006) = 0.02
yb 7\| 6
CI
2.365
95
\ 7 (0.05) - (0.01) = 0.08
-------
TABLE 19. TEST RESULTS FOR C02 MONITOR 24-HOUR ZERO AND CALIBRATION DRIFT
(% C02 except as indicated)
Test No.
1
2
3
4
5
6
7
Date
Start
8/11
8/12
8/13
8/14
8/16
8/17
8/18
[1984)
End
8/12
8/13
8/14
8/15
8/17
8/18
8/19
Test time
Start
1215
0930
0820
0800
1500
1500
0800
End
0930
0830
0800
0800
1500
1800
0900
Zero reading
Start End
(A)
0
0.04
0
0
0
0.10
0
(B)
0.04
0
0
0.02
0.10
0.10
0
Arithmetic mean (AM)
95% confidence interval (CIg5)
24-hour drift,3 %
Zero
range
drift
(C=B-A)
0.04
-0.04
0
0.02
0.10
0
0
0.03
0.04
0.07
Span reading
Start End
(D)
15.4
15.8
16.3
15.9
14.7
15.1
15.6
(E)
15.8
16.3
15.9
15.5
15.1
15.6
15.7
Span
drift
(F=E-D)
0.40
0.50
-0.40
-0.40
0.40
0.50
0.10
Cali-
bration
drift
(F-C)
0.36
0.54
-0.40
-0.42
0.30
0.50
0.10
0.37
0.36
0.73
GO
CTl
a% 24-hour drift
Zero drift
qc -
as % C02 = |AM| + CI95.
2Oi;c i—
.JD3 y /Q Q136) (Q 01
7 r~r N '
Cal ibration drift
nn\ n 04 TT - ^•JDO 1 7
yb 7 J~l~
(1.11) - (0.96) = 0.36
-------
TABLE 20. MONITOR RESPONSE TIME
(seconds)
Test No.
NO
°2
CO
co2
Upscale9 at
1110 (8/17/84)
2:00
2:00
2:45
2:30
Downscale
2:10
2:20
2:50
2:15
Response time needed to record stable stack effluent
reading.
Response time needed to record stable high-level calv
bration gas reading.
TABLE 21. NO MONITOR RESPONSE TIME
(seconds)
Test No.
1
2
3
Date
(1984)
8/12
8/18
8/19
Downscale9
1:58.9
2:05
2:00
Response time needed to record stable high-level calv
bration gas reading.
37
-------
TABLE 22. 02 MONITOR RESPONSE TIME
(seconds)
Test No.
1
2
3
Date
(1984)
8/13
8/13
8/14
Downscale9
2:15.0
2:30.0
2:15
Response time needed to record stable high-level cali-
bration gas reading.
TABLE 23. CO MONITOR RESPONSE TIME
(seconds)
Test No.
1
2
3
Date
(1984)
8/12
8/13
8/14
Downscale
3:08.6
2:48.1
2:30
Response time needed to record stable high-level calv
bration gas reading.
TABLE 24. C02 MONITOR RESPONSE TIME
(seconds)
Test No.
1
2
3
Date
(1984)
8/12
8/13
8/14
Downscale3
2:17.5
1:54.3
2:00
Response time needed to record stable high-level cali-
bration gas reading.
38
-------
system and sample line as a system check and an indicator of possible sample
dilution or contamination. All calibration gases were Master Gas-Certified,
which means the gas values were within ±2 percent of indicated values. Along
with calibration gases, zero nitrogen was used to zero all monitors and to
purge sample lines to guarantee a clean sampling system.
Data generated by the CEM calibrations (three-point and single-point)
were used to define calibration curves for each monitoring system. Each
calibration response had a chart division reading and a corresponding cali-
bration gas concentration (ppm, %). A linear regression analysis of these
data was conducted to establish the relationship between response and concen-
tration, or, the degree of correlation (linearity). Figures 1 through 4
present example calibration curves. The linear regression equations estab-
lished for each monitor on a daily basis were then used to define pollutant
concentrations for each specific test block. The final reduction of data was
accomplished by taking an average chart reading for every 5-minute period and
determining the pollutant concentration by the linear regression equation.
Tables 25 through 28 summarize the CEM linear regression data for each test
block conducted.
The EPA supplied NO audit gases to check monitor response and accuracy.
A
These audit gases were analyzed daily throughout the test program. Table 29
summarizes the results of the NO CEM system audits. As shown, the NO CEM
A A
response compared favorably with the audit cylinder values.
As a final check of the NO CEM system, several stack samples were col-
A
lected and analyzed according to procedures described in EPA Reference Method
7.* Table 30 summarizes the comparative data. Results for the majority of
the Method 7 samples collected were within ±20 percent of the NO CEM values
A
recorded during the sample collection period.
Table 31 summarizes a data comparison between 02 and CO,, results from
those test blocks for which both CEM and EPA Reference Method 3 data are
available. The CEM data represent average values for each designated test
block. The Reference Method 3 data were obtained by collecting a gas sample
in a Tedlar bag during the particulate tests. The gas sample collection probe
was attached to a particulate train sample probe during each test so that an
40 CFR 60, Appendix A, Reference Method 7, July 1984.
39
-------
NOV CEM CONCENTRATION, 8/15/84,
TEST BLOCK 1-G
XPPM 0.425
CORRELATION COEFFICIENT = 0.9994
120 160 200
NO CONCENTRATION, PPM
Figure 1. Example NOX calibration curve.
40
-------
100
90
80
70
60
50
eC
U
40
30
20
10
09 CEM CONCENTRATION, 8/15/84,
* TEST BLOCK 1-G
u Y-4.41
10.18
CD-4.41
10.18
CORRELATION COEFFICIENT = 0.9998
10
12 14
Figure 2. Example 02 calibration curve.
41
-------
o
100
90
80
70
60
50
40 •
30 -
20 -
10 -
40
CO CEM CALIBRATION, 8/15/84,
TEST BLOCK 1-G
v.Y-5.51
* 0,404
CO
_CD-5.51
PPM 0.404
CORRELATION COEFFICIENT = 0.9998
80 1ZO 160 200
C02,PPM
240
280
Figure 3. Example CO calibration curve.
42
-------
100
Y
90
80
70
60
o
•—H
t—I
>•
S 50
i—
or
3C
40
30
20
10
C02 CEM CALIBRATION, 8/15/84,
TEST BLOCK 1-G
X=
Y-6.08
4.856
rn y-CD-6.08
V 4.856
CORRELATION COEFFICIENT = 0.9988
6 9 12 15 18
co2,%
Figure 4. Example C02 calibration curve.
43
21
-------
TABLE 25.
NO LINEAR REGRESSION DATA
/\
Test Block
l-Ga
2-G
lb
2
3
4
5
6
7
8
9
3-6
4-6
5-6
6-6
7-6
8-6, 9-6
and 10-6
No. of cali-
bration points
7
7
8
8
8
8
8
8
6
8
8
8
6
6
6
4
6
Y-intercept
6.80
7.68
5.90
5.84
5.55
5.41
5.40
5.13
6.07
5.84
5.60
5.01
5.02
5.01
4.98
5.40
5.24
Slope
0.4252
0.4125
0.0994
0.0975
0.0970
0.1027
0.1050
0.1070
0.0950
0.1013
0.1038
0.1009
0.1020
0.1032
0.1047
0.1057
0.1044
Correlation
coefficient
0.9994
0.9994
0.9996
1.0000
0.9998
0.9999
0.9996
0.9998
0.9995
0.9954
0.9978
0.9998
1.0000
0.9999
0.9999
0.9995
0.9996
JTest Blocks 1-6 and 2-G on 0-250 scale.
5A11 other test blocks on 0-1000 scale.
44
-------
TABLE 26. 0, LINEAR REGRESSION DATA
Test Block
1-6
2-6
1
2
3
4
5
6
7
8
9
3-6
4-6
5-6
6-6
7-6
8-6, 9-6
and 10-6
No. of cali-
bration points
5
5
6
6
6
6
6
6
4
6
6
6
4
4
5
4
5
Y-intercept
4.41
4.82
7.39
7.24
7.22
4.78
4.68
4.92
5.04
4.85
5.37
4.58
4.53
4.44
3.64
1.87
3.33
Slope
10.18
9.97
9.90
9.67
9.70
10.03
9.88
9.14
8.78
8.88
9.00
9.88
9.90
9.62
9.82
10.20
9.91
Correlation
coefficient
0.9998
0.9998
0.9994
0.9999
0.9998
1.0000
0.9994
0.9944
0.9991
0.9994
0.9999
0.9999
0.9999
0.9999
0.9997
1.0000
0.9995
45
-------
TABLE 27. CO LINEAR REGRESSION DATA
Test Block
1-G
2-6
1
2
3
4
5
5a
6
7
8
9
3-6
4-6
5-6
6-6
7-6
8-6, 9-6
and 10-G
No. of cali-
bration points
7
7
8
8
8
8
8
4
8
6
8
8
8
6
6
6
4
5
Y-intercept
5.51
5.95
5.46
5.60
5.49
5.45
5.65
6.34
5.56
6.84
6.76
6.36
5.32
5.08
5.09
5.17
4.95
4.97
Slope
0.4039
0.4028 .
0.4061
0.4063
0.4068
0.4064
0.4066
0.1857
0.4072
0.4068
0.4064
0.4084
0.4052
0.4065
0.4058
0.4075
0.4089
0.4082
Correlation
coefficient
0.9998
0.9999
0.9999
0.9999
0.9999
0.9999
1.0000
0.9963
0.9999
0.9998
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999
1.0000
1.0000
0-500 ppm scale; all other tests 0-250 ppm scale.
46
-------
TABLE 28.
C02 LINEAR REGRESSION DATA
Test Block
1-6
2-6
1
2
3
4
5
6
7
8
9
3-6
4-6
5-6
6-6
7-6
8-6, 9-6
and 10-6
No. of cali-
bration points
7
7
8
8
8
8
8
8
6
8
8
8
6
6
6
4
6
Y-intercept
6.11
6.10
5.55
5.85
6.15
6.54
6.94
6.65
6.14
6.99
6.79
6.30
5.98
6.28
5.89
5.45
5.52
Slope
4.85
4.79
4.70
4.73
4.88
4.88
4.82
4.76
4.88
4.83
4.94
4.93
4.77
4.76
4.97
5.14
4.99
Correlation
coefficient
0.9988
0.9994
0.9998
0.9996
0.9985
0.9993
0.9988
0.9982
0.9993
0.9991
0.9995
0.9987
0.9995
0.9992
0.9990
1.0000
0.9990
47
-------
TABLE 29. SUMMARY OF NO OEM AUDIT RESULTS
Date
(1984)
8/11
8/11
8/14
8/15
8/16
8/16
8/17
8/17
8/18
8/18
8/18
8/18
8/18
8/19
8/19
EPA audit
cylinder ID
LL-4348
AAL-6222
LL-4348
LL-4348
AAL-6222
AAL-6222
AAL-6222
LL-4348
AAL-6222
AAL-6222 '
AAL-6222
AAL-6222
AAL-6222
AAL-6222
LL-4348
EPA audit .
value NO, ppm
109
225.2
109
109
225.2
225.2
225.2
109
225.2
225.2
225.2
225.2
225.2
225.2
109
CEM NOC
value, ppm
113
225
110
113
223
226
222
109
215
225
220
238
223
223
110
Percent
difference
+3.5
-0.09
+0.90
+3.5
-0.99
+0.35
-1.4
0
-4.7
-0.09
-2.4
+5.4
-0.99
-0.99
+0.90
aGas cylinders provided by U.S. EPA.
Audit values of nitric oxide (NO) with the balance of gas being nitrogen.
cValues determined from PEDCo NOX CEM system. All audit gases were introduced
at the sample probe.
48
-------
TABLE 30. COMPARISON OF REFERENCE METHOD 7 AND NO
CEM TEST RESULTS >
Method 7
Test No.
G-1A
G-1B
G-1C
G-1D
1A
IB
1C
ID
7A
7B
7C
8B
8C
G-6A
G-6B
G-6C
G-6D
Date
(1984)
8/15
8/15
8/15
8/15
8/16
8/16
8/16
8/16
8/18
8/18
8/18
8/18
8/18
8/19
8/19
8/19
8/19
Time
(24-h)
1642
1650
1658
1704
1350
1400
1914
1920
1756
1805
1815
2125
2130
1826
1831
1835
.1837
Testing
location
No. 4
No. 4
No. 4
No. 4
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 5
No. 4
No. 4
No. 4
No. 4
Method 7
test
result, ppm
151
161
151
160
379
404
348
368
331
363
359
374
333
138
116
134
122
CEM NO
value, ppm
193
194
196
199
361
361
381
381
430
430
433
394
394
140
120
124
124
Percent
difference
+28
+20
+30
+24
-5
-11
+9
+4
+30
+18
+21
+5
+13
+1.4
+3.4
-7.5
+1.6
Difference =
CEM - Method 7
Method 7
x 100.
49
-------
TABLE 31. COMPARISON OF OXYGEN AND CARBON DIOXIDE
RESULTS—CEM AND REFERENCE METHOD 3 (ORSAT)
Test Block
1
2
5
8
Date
(1984)
8/16
8/16
8/17
8/18
CEM value9
02
3.5
1.0
1.2
1.4
C02
13.2
15.0
15.6
15.4
Reference Method 3
02
4.6
2.3
2.0
2.5
C02
12.6
13.7
13.5
13.5
Represents average monitor value calculated for the designated
test block.
'Represents data from Method 3 analysis (Orsat) of integrated bag
samples collected during the designated particulate tests.
50
-------
integrated (traverse) sample was collected. No attempt was made to collect
gas samples at a point near the CEM sample probe.
Several Method 3 samples (Test Blocks 3, 4, 6, and 7) were considered
void because of broken bags or leakage problems within the gas sampling sys-
tem. Accordingly, the CEM data from these test blocks were used in all par-
ticulate calculations. As presented in Table 31, the EPA Reference Method 3
data consistently exhibited higher CL and lower CCL values than the average
CEM results. Though no specific explanation was found for the discrepancy,
the Method 3 samples might have been diluted during sample collection or
immediately prior to analysis. Considering the low levels of CL encountered,
such a dilution would tend to distort the comparison between the two measure-
ment systems. In contrast, the difference between the Method 3 and CEM C02
values ranged between 5 and 15 percent, which is considered acceptable.
3.2 MANUAL TESTS--PARTICULATE AND NO
X
Table 32 lists the sampling equipment used to perform the particulate
tests and the calibration guidelines and limits. In addition to the pre- and
post-test calibrations, a field audit was performed on the metering systems
and thermocouple digital indicators used for sampling. Critical orifices
constructed by PEI were used in the dry gas meter audits. Figures 5 through 7
present results of the onsite audits. These data were used to assess the
operational status of the sampling equipment relative to EPA guidelines.
Figure 8 is an example of an unacceptable meter box audit. The audit
value for the dry gas meter Y-factor was greater than ±5 percent, which was
considered unacceptable by the PEI Project Manager; therefore, the meter box
was not used for this test program.
The sample data and isokinetic sample rates were calculated on site by
PEI personnel. The data were rechecked and validated at the end of the test
program by computer programming.
Figure 9 presents an example calculation form PEI used during the test
series. Computerized calculations are presented in Appendix A of this report.
As a check of the analytical methodology used, blank filter and reagent
(acetone) samples were analyzed in a similar fashion as the actual field
sample analyses. Table 33 presents the gravimetric blank analysis blank.
51
-------
TABLE 32. FIELD EQUIPMENT CALIBRATION
01
ro
Equipment
Meter box
Pilot tube
Digital indi-
cator
Thermocouple
and stack
thermometer
Orsat analyzer
Inpinger
thermometer
Trip balance
Barometer
Dry gas
thermometer
Probe nozzle
ID
Mo.
FB-4
FB-8
188
515
126
221
201
205
0,
c62
CO
385
291
M-l
229
FB-4
FB-8
8-107
8-110
6-109
6-110
Calibrated
against
Wet test meter
Standard pitot
tube
Millivolt signals
ASTM-2F or 3F
Standard gas
ASTM-2F or 3F
Type S weights
NBS traceable
barometer
ASTM-2F or 3F
Caliper
Allowable
error
Y ±0.02 Y
AH @ iO.15
(Y ±0.05 Y post-test)
Cp *0.01
0.5*
1.5%
(±2* saturated)
±0.5%
±2°F
±0.5 g
±0.10 in.Hg
(0.20 post-test)
±5°F
Dn ±0.004 in.
Actual
error
0.67
0.05
-0.68
0.02
0.05
-2.4
0.0
<0.5%
<0.5l
<0.50%
-------
ON-SITE AUDIT DATA SHEET
Audit Name:
Date: g///AV Auditor:
r*
TW>.
I*
+*•
Equipment
'-V Meter box
.^ inlet thermo.
-/ Meter box
-i. outlet thermo.
"^ Impinger
•6 thermometer
„ Stack
thermometer
or
Thermocouple
Orsat
analyzer
Trip
balance
Barometer
Reference
ASTM-3F at
ambient temp.
ASTM-3F at
ambient temp.
ASTM-3F at
ambient temp.
ASTM-3F at
ambient temp.
ASTM-3F at
stack temp.
% 02 in
ambient air
IOLM std.
weight
Corrected*
NWS value
Reference
Value
•^yf
W
77-
7>V^
I*'?
20.8%
ii
^
Value
Determined
7/V
73 V
e*°
Wf
irf
>fl.«
Deviation
Z>
-7£
^3*^
-*•/*
-/'?
0,1
Max. Allowable
Deviation
5°F
5°F
2°F
7°F
See table
0.7%
0.5 grams
0.20 in. Hg
Reference temp. °F
Max. deviation °F
32-140
7
141-273
9
274-406
11
407-540
13
541-673
15
674-760
17
* Correction factor:
NWS value (in. Hg) - [Altitude (ft)/1000(ft/in. Hg)] + 0.74 in. Hg**
** 0.74 1n. Hg is the nominal correction factor for the reference barometer
against which the field barometer was calibrated.
If It Is not feasible to perform the audit on any piece of equipment, record
"N/A" In the space provided for the data.
Iff
It'F
I
Iff
-3'f
Figure 5. On-site audit data sheet.
53
-------
FIELD AUDIT REPORT: DRY GAS METER
BY CRITICAL ORIFICE
DATE:
CLIENT:
BAROMETRIC PRESSURE (Pbjr):^T.79 in.Hg METER BOX NO.
ORIFICE NO. 3 PRETEST Y: ^
ORIFICE K FACTOR: 5^3">7X /*'v AUDITOR:
AHP
1n.H20
Orifice
nanometer
reading
AH,
in.HzO
/.^
Dry gas
meter
reading
w
ft3
1i.ee>
f*-lfa_i
Temperatures
Ambient
Tai/Taf
°F
1M-
73
Average
V
°F
13*
Dry gas meter
Inlet
VTif
°F
?>
°ll
Outlet
Toi/Tof
°F
n
VB
Average
Tm'
"F
g6.7^
Duration
of
run
0
min.
/jTosy
Dry gas
meter
V fts
/***
Vm
mstd'
ft3
/*.*1
Vm
macf
ft3
/^.^
Audit,
Y
/..»/
Y
devia-
tion, %
'•>•<•&
Audit
AH(?,
1n.H20
) /./^
AHP Devia-
tion, 1n.H20
& - . o^ (p4-s
"std
17.647(Vj(Pbar + AH/13.6)
460)
ft3
"act
1203( 0 )( K ){Pbar)
(Ta + 460)
112
ft3
Audit Y
"'act
"'std
Y deviation
Audit Y - Pre-test Y
Audit Y
x 100
Audit AH? = (0.0317)(AH)(P.aJ{T + 460)
DO r m
Audit Y must be in the range, pre-test Y ±0.05 Y.
Audit AHP must be in the range pre-test AH@ ±0.15 inches H20.
Figure 6. Field audit of dry gas meter (Meter Box FB-4)
54
-------
FIELD AUDIT REPORT: DRY GAS METER
BY CRITICAL ORIFICE
DATE:
CLIENT:
BAROMETRIC PRESSURE (P^):/??.?? 1n.Hg METER BOX NO.
ORIFICE NO.
PRETEST Y:
AHP 1.11 1n.H20
Orifice
manometer
reading
AH,
1n.H20
' '9 ""V
Dry gas
meter
reading
VV
ft3
700.*
7*t.±
Temperatures
Ambient
Tai/Taf
°F
77
75-
Average
V
°F
^
Dr
Inlet
VTif
°F
ev
&<£,
1 gas meter
Outlet
w
°F
77
7^
Average
Tm-
°F
e,r
Duration
of
run
0
min.
*:«*
Dry gas
meter
V ft3
/,?.7^
mstd-
ft3
/P.VV
macf
ft3
/p.sr
Audit,
Y
/.009
Y
devia-
tion, X
/-/.^^
Audit
AH@,
1n.H20
/.e
-------
FIELD AUDIT REPORT: DRY GAS METER
BY CRITICAL ORIFICE
DATE:
CLIENT:
BAROMETRIC PRESSURE (Pbar):.*?.S7 in.Hg METER BOX NO.
ORIFICE NO. 3 PRETEST Y: _^
ORIFICE K FACTOR: j'377*/l>~'f AUDITOR: C
AHP A7V 1n.H20
(\
Orifice
manometer
reading
AH,
in.HjO
s
» £"/**
Dry gas
meter
reading
VV
ft3
Jtf. 4o
3*x*>
Temperatures
Ambient
Tai/Taf
°F
16
11-
Average
Ta'
°F
^
7.i
Dry gas meter
Inlet
T11/T1f
°F
73
7/
Outlet
Toi/Tof
°F
73-
73
Average
Tm-
°F
72
Duration
of
run
min.
//"•<*
Dry gas
meter
Vm- ft3
/;.y
Vm
"std1
ft3
IM
Vm
macf
ft3
/3,-)l
Audit.
Y
,.»11
Y
devia-
tion, %
'-'ft
Audit
AHP,
in.H20
AH@ Devia-
tion, in.H20
"std
"act
1203( 0 )( K )(Pbar)
(Tfl 4 460)
\Ti
#*
ft3
fi.Tl
AuditY=
Y deviation *
Audit Y
std
lo° g ^'
Audit
(0.0317)(AH)(Pbar)(Tm + 460) (V(t, , AH/l3.6
in.H20
Audit Y must be in the range, pre-test Y ±0.05 Y.
Audit AHP must be in the range pre-test AH@ ±0.15 inches H20.
Figure 8. Example of an unacceptable meter box audit.
56
-------
14
ISOKINETIC CALCULATION
SITE.
fi/* $
-if
RUN 1
RUN 2
RUN 3
RUN 4
1.
VeliMi of dry fit tM^lrt cormttd to
tundtrd condition. HeU: »„ mitt kt
Cormtld for iMUgt 1f t*J )l*k«at
ntet fiCMd l().
• I7.M » »_ « T tar • TIT
V.. ft3
Mf A ' Z.
AH. 1n.H20
0.7
.7
V . dscf
"$td
2. Voluat of Mttr »«por tt tUndcrd con-
d1t1en». ft'.
Vlc. 9
Vd
0.04707V
..
*e
.ft
to,(ej
1-74
3. ItoiiUM content in lUek 9ts.
•«•
•ttd
V • V
•»td *
4. Dry •oltculir M * » "« *
Ms. Ib/lb-nole
.
x. a, . / >«|-
^- / • 7 U
(. Stack *t1oc1tjr at tuck condition*.
»P».
Pstat1c* 1n
-------
TABLE 33. PARTICULATE FILTER AND REAGENT BLANK ANALYSIS
Sample type and
identification
Method 5 - ambient
Filter No. 0002771
(D5079)
Acetone (D5088)
(Blank volume
= 170 ml)
Method 5B - 160°C
Filter No. 0003396
Acetone D5088
(Blank volume
= 170 ml)
Original tare
weight, mg
347.5
101,882.1
349.5
106814.0
Blank
weight, mg
347.7
101,888.8
349.6
106815.1
Net
weight, mg
+0.2
+6.8 mg
= 0.051 mg/g
+0.1
+1.1
58
-------
Figures 10 through 13 present the results of the NO (Method 7) audit
/\
results from both the field and laboratory analyses. Audit solutions supplied
by EPA were analyzed according to the procedures described in Method 7. The
results indicate good analytical technique.
59
-------
AUDIT REPORT NOV ANALYSIS
Plant
(?,£,
PN Number
Date sample
Samples ana
Reviewed by
Sample
Number
3SJ3
s received & -/£
lyzed by //• L ,
7/2-,
f
mg N02/dscm
Determined
^~ 3 % ?
'-?*/ Date a
Date o
Source of
Sample
r*-^
nalyzed
f Review
Accepted
Value
W.9
?-/*-??
S/3//9"/
Difference
-9.V
Figure 10. Audit report NO analysis results on samples
received 8/15/84.
60
-------
AUDIT REPORT NOX ANALYSIS
Plant
PN Number
7
Date samples received <^-/
Samples analyzed by tf> /.
Reviewed by
Date analyzed
Date of Review
Sample
Number
LOf OH';!*,
yj?3
mg N02/dscm
Determined
$J3
Source of
Sample
7" fay^^
Accepted
Value
93 7, A
%
Difference
+ /0*3
Figure 11. Audit report NO analysis results on samples
received*8/16/84.
61
-------
Plant
AUDIT REPORT NOX ANALYSIS
PN Number
Date samples received y-/9-fY. ?-/^/Date analyzed P'/f-P^
Samples analyzed by /^«//
Reviewed by
Sample
Number
* S°SA
'%/.%'. Date of Review %/?//f"/
// f f
mg NC>2/dscm
Determined
/£>/&
Source of
Sample
r.*r~
Accepted
Value
f«*
Difference
.*?
Figure 12. Audit report NO analysis results on samples
received 8/17/84 and 8/18/84.
62
-------
AUDIT REPORT NOV ANALYSIS
Plant £;£ 6<,'(/6.//l ~ (JtCPii
Date samples received
Samples analyzed by &
Reviewed by /T/. a---
PN Number
Date analyzed
Date of Review % '5,'/
Sample
Number
3 *n -~}
e>(j I t-1-
'R5o^iT
?£??
T^o^t*
mg N02/dscm
Determined
3/f.5"^
^7 U ° «r
/ 7^
-------
SECTION 4
SAMPLING LOCATIONS AND TEST METHODS
4.1 SAMPLING LOCATIONS
All tests were run in the exit stacks of Boilers 4 and 5, as depicted in
Figures 14 and 15. In the Boiler 4 exit stack, two sampling ports, 90 degrees
off center, were located approximately 3.2 duct diameters downstream and more
than 8 duct diameters upstream from the nearest flow disturbance in the 152-cm
(60-in.) i.d. round stack. A total of 24 traverse points (12 per port) were
used to measure gas velocity and temperature. As mentioned in Section 2, no
particulate measurements were made in this stack.
In the Boiler 5 exit stack, four sampling ports, 90 degrees off center,
were located approximately 2.7 duct diameters downstream and more than 8 duct
diameters upstream from the nearest flow disturbance in the 183-cm (72-in.)
i.d. round stack. A total of 24 sampling points (12 per port) were used to
conduct the particulate tests. Two individual sampling trains (a Method 5
with a sample collection temperature of 120°C and a Method 5B with a sample
collection temperature of 160°C) were used to isokinetically traverse the
cross-sectional area of the stack for each of nine test runs. During Tests 1
through 5, each point was sampled for 5 minutes, for a total test time of 120
minutes. During the remaining tests, the total test time was reduced to 72
minutes, or 3 minutes per sampling point. A brief description of the test and
analytical procedures used is presented in the following subsection.
4.2 CONTINUOUS EMISSION MONITORS—SAMPLE EXTRACTION, ANALYSIS, AND DATA
REDUCTION
An extractive monitoring system was assembled on site to provide a con-
tinuous emissions data base for NO , 02, CO, and C02- Figure 16 presents the
CEM system layout.
64
-------
TO ATMOSPHERE
SAMPLE PORTS
10-cm (4-in.)
I.D.
ROOF LINE
152 cm
(60-in.) I.D.
02 MONITOR
CROSS-SECTION
MONITOR LOCATION
3.7 m
(12 ft. 2 in.)
4.9 m
(16 ft.)
1.2 m
(»3 ft. 10 in.)
FLOW FROM
BOILER
Figure 14. Boiler 4 exit stack (no scale).
65
-------
/* r
t C
PORT
LOCATION
10-cm (4-in.)
l.D.
ROOF LINE
_ 183 cm „
(72 in.) l.D.
.^
^
02 MONITOR
1
FLOW FROM
BOILER
• •
(
3.
(12 ft
1
i
1.
(=3 ft
<
CROSS-!
/v
M
10-cm (4-
SAMPLE PORT
,
7 m
3 in.)
4.
(16
m
. 9 in.)
r
SECTION
\
}
in.) i.d.
S (90° O.C.)
9 m
ft.)
1
Figure 15. Boiler 5 exit stack (no scale).
66
-------
CTl
CALIBRATION
GAS
1/4-in. TEFLON
ANALYZERS
CALIBRATION GAS LINE
S.S. 1/4-in.
3-WAY S.S PROBE
VALVE ,
=31 m SAMPLE LINE - 1/4-in. TEFLON
(100 ft) ,
S.S. CONDENSER
(ICE BATH)
TRAILER
TRANSPORT SAMPLE
PUMP
STACK WALL
. PARTICIPATE
FILTER
FLOW
GLASS WOOL
TEFLON PUMP
BUBBLE METER
'.AMPLE MANIFOLD - 1/4-in. S.S.
O
EXHAUST
Figure 16. CEM system layout.
-------
A single Teflon sample line was used to transport the gas sample to the
NOX, CO, C02> and (^ monitors. The gas conditioning system consisted of an
in-stack glass-fiber filter (Reeve Angel 934AH) to remove particulate matter,
followed by an ice-bath condenser to remove moisture. The conditioned sample
gas then passed through a flow regulator, a pump, an additional moisture
knockout jar, and into a Teflon manifold with an individual tee connection for
each monitor. Flow at the outlet of the manifold was monitored to ensure that
the sample pump was supplying a constant excess of sample gas.
System leak checks and checks for zero drift, span drift, and response
time were performed daily on each monitor. Guidelines set forth in 40 CFR 60,
Appendix B, Performance Specification Tests 2 and 3, were followed during this
test series.
A three-point calibration check was performed on each monitor at the
beginning and end of each test day. This check covered the low, mid, and high
values of the specific pollutant concentrations measured. Single-point cali-
bration checks were conducted between test blocks to ensure proper monitor
response. Calibration gases were delivered through the gas sampling system
(condenser and sample line) as a check on total sample system integrity.
Upon completion of system checks and calibration of monitors, the sample
probe was inserted in the stack at the designated sample point. Stack gases
were purged through the sampling system for 10 minutes, or until stable read-
ings were achieved on the monitors. Data were then recorded for each desig-
nated test period. The particulate filters and condenser were cleaned as
necessary between test blocks. At the end of each test block, all monitors
were zeroed, calibrated (single or multi-point as time allowed), and prepared
for the next test block.
All CEM's used for this test series have linear response curves. The
three-point calibration conducted at the beginning and end of each day was
used to verify instrument linearity. Each calibration response had a chart
division reading and a corresponding calibration gas concentration (parts per
million, percentage). A linear regression analysis was conducted to determine
the relationship between response and concentration or the degree of correla-
tion or linearity.
68
-------
The final reduction of data was accomplished by taking an average chart
reading for every 5-minute period and determining concentration by the linear
regression equation established from the monitor calibrations.
4.3 PARTICIPATE TEST METHODS AND ANALYTICAL PROCEDURES
Particulate was measured concurrently with the CEM data acquisition while
Boiler 5 was firing both residual and distillate oil.
Two individual sample trains were used to traverse the cross-sectional
area of the stack. One train was a standard EPA Method 5* sampling system
with a sampling temperature of 121°C (250°F), and the other was an EPA Method
5B* sampling system with a sampling temperature of 160°C (320°F). Each train
consisted of a stainless steel sampling nozzle, a heated glass-lined probe, a
heated glass-fiber filter, and a series of Greenburg-Smith impingers followed
by a vacuum line, vacuum gauge, leak-free vacuum pump, dry gas meter, thermom-
eters, and a calibrated orifice. For each train, the probe and filter temper-
atures were set at a predetermined temperature and monitored by the use of
multiterminal digital indicators with thermocouple leads located in each probe
and immediately behind the Method 5 filter frits.
The nozzle, probe, and front filter holder portions of each sampling
train were acetone-rinsed at the end of each test. For the Method 5 samples,
the acetone rinse and particulate caught on the filter media were dried at
room temperature, desiccated to a constant weight, and weighed on an analyti-
cal balance. For the Method 5B samples, the acetone rinse was evaporated at
room temperature. The resulting rinse residue and the filter were then heated
in an oven for 6 hours at 160°C (320°F), cooled in a desiccator, and weighed
on an analytical balance. Total filterable particulate matter was determined
by adding the rinse and filter values.
4.4 MANUAL TEST METHOD FOR NOX
Flue gas samples were collected from each stack during the test program
and analyzed for NO according to procedures described in EPA Reference Method
A
40 CFR 60, Appendix A, Reference Methods 5 and 5B, July 1984.
69
-------
7.* These data were used to verify the NO CEM relative accuracy and to
/\
provide additional quality assurance data for the NO CEM system. Most of
/\
these samples were recovered and analyzed on site. Aliquots of the sample
solutions were also retained and analyzed in our Cincinnati laboratory as a
check of the field analysis data.
40 CFR 60, Appendix A, Reference Method 7, July 1984.
70
-------
SECTION 5
PROCESS DESCRIPTION AND OPERATION
This section presents a brief process description of the boiler test
units. Included are characterizations of the boiler and burner and discus-
sions of the plant operating history and control procedures.
5.1 BOILER DESCRIPTION
Boiler Units 4 and 5 were manufactured by Babcock and Wilcox (B&W) in
1974 and 1975, respectively. Both are packed watertube boilers, Model
FM117-97 with "D"-type furnace construction (the upper steam drum, lower drum,
and furnace wall water tubes are in the shape of a "D"). The boilers are
essentially identical, but they are right- and left-hand versions of the same
design. Each boiler is rated at 150,000 Ib/h steam generating capacity, or
170 x 10 to 180 x 10 Btu/h heat input (based on firing oil or gas, with an
economizer). The outer dimensions for this size boiler are 15 ft height x 12
ft width x 29 ft length (to the nearest foot). Figure 17 shows the layout of
Unit 5.
Combustion air passes through a forced draft fan located on the boiler
house roof. It flows down through the air preheater, which is a series of
four steam tube coils, where it can be heated to 150° to 200°F. Then the air
enters the windbox, which houses the Coen low excess air (LEA) burner (which
is described more thoroughly in the next subsection).
The flame exiting the burner travels down an open channel within the
furnace. The furnace sidewalls, roofs, and floor all contain water tubes to
cool the walls and adsorb radiant heat energy from the flame. The rear wall
of the furnace is a heavy duty bank of water tubes as opposed to a refractory-
lined surface.
The hot gases turn the corner at the back of the furnace and come forward
through the convection section. This section is where the two drums and
71
-------
FORCED DRAFT
FAN
ROOF LINE
SWUNG
PORTS
AIR
PREHEATER'
x C0/C02 MONITOR
• ROOF LINE -
STACK
ECONOMIZER
BOILER GAS OUTLET
LOWER
Figure 17. Unit 5 layout.
72
-------
additional water tubes are located. The flue gas exits through the side of
the boiler at the front, flows up through some ductwork, passes downward
through the economizer, and is ducted to the stack.
Boiler feed water under pressure is preheated from 245°F in the econo-
mizer to s300°F by the hot flue gases. It exits the economizer, enters the
upper drum, and circulates through the water tubes of the convection and radi-
ant sections. The steam produced is collected in the upper steam drum, exits
the top of the drum, and passes to the plant's header (pressure control)
system.
The current operating pressure of the boilers is 160 psig of saturated
steam (original design 185 psig) and the maximum pressure is 250 psig. There
are steam vents to act as a pressure relief system if the pressure gets too
high.
The boiler convection section has 7095 ft2 heat transfer surface area.
The water tube in the furnace contains 1072 ft2 of radiant heat transfer sur-
face area exposed to flame, giving a total commercial heating surface of 8167
ft2. The furnace volume is 1674 ft3—corresponding to approximate furnace
dimensions of 11 ft height x 6 ft width x 25 ft length (to the nearest foot).
At full load, the furnace heat release rate is approximately 165,000 Btu/h ft2
or about 100,000 Btu/h ft3.
Each boiler is equipped with an economizer. The economizer in Unit 4 has
2988 ft2 of bare tube surface area, but the steel tubes in the economizer in
Unit 5 have cast iron fins. These fins provide additional heat transfer
surface area and therefore Unit 5 operates cooler. As Unit 5 is more amenable
to residual oil firing, the plant prefers to fire oil only in Unit 5. This is
why there are only gas tests on Unit 4, plus the fact that all the controls
for firing oil are already set for oil firing on Unit 5.
5.2 BURNER DESCRIPTION
Each boiler contains a single burner manufactured by Coen Company, Inc.
The burners in Units 4 and 5 are identical. They are parallel flow type burn-
ers capable of operating at very low excess air levels in the 5 percent excess
combustion air range, corresponding to about 1 percent flue gas 0?. The
combustion air is accelerated to velocities sufficient to give good mixing and
73
-------
complete combustion even at these air/fuel ratios. This burner is shown in
Figure 18.
The burners have 30 inch diameter throats. These LEA burners were the
first ones of this size that Coen installed. They are designed to fire natu-
ral gas or residual oil (a combination of these two fuels could be fired in
other installations). For natural gas firing, the gas is introduced in the
burner through eight gas spuds on a ring header.
For No. 6 oil firing, the fuel is first preheated to lower the viscosity
to 250 SSU. It is then atomized using 160 psig saturated steam and is dis-
charged through the burner gun assembly running axially down the burner
throat. Combustion air also enters the burner throat axially (i.e., in paral-
lel flow) through two narrow bellmouth assemblies. By adjusting these inner
and outer concentric bellmouths, the relative air flows to three radial burn-
er zones (core, annulus, and sheath) can be varied. Using these adjustments,
the same burner can fit an appropriate flame pattern into similar sized fur-
naces of different shape.
5.3 OPERATING HISTORY
In the early 1970's the plant's process steam demand increased due to
various process expansions. Units 4 and 5 were the last two, of a total of
five, packaged boilers installed at this site. This brought the total capac-
ity of the steam plant to over 500,000 pounds of steam per hour.
Since that time with increasing fuel costs, the manufacturing plant has
invested in heat recovery equipment. This has lowered the total process steam
demand to approximately 100,000 Ib/h today. At this load, the steam plant
normally operates two boilers, Units 3 and 5. Unit 5 normally swings load to
follow the process steam demand. Unit 4 serves as a backup to Unit 5 and can
be fired with natural gas during times of high steam demand, such as shutdown
of major heat recovery equipment or a general plant startup.
Originally, Units 4 and 5 were equipped with oil-only burners supplied by
Babcock and Wilcox. As an additional fuel cost saving investment, the plant
installed a Coen LEA burner in Unit 5 in January 1981. During startup, this
burner had a resonant vibration problem firing natural gas; the burner was
74
-------
Figure 18. Coen parallel flow type LEA burner.
75
-------
removed and redesigned. The redesigned burner was installed and became opera-
tional in Unit 5 in September 1982. At the same time, an identical burner (of
the new design) was retrofitted in Unit 4.
5.4 CONTROL PROCEDURES
Normally the boiler plant operates in the automatic control mode. Steam
flow and pressure measurements for the total process steam line to the manu-
facturing plant are input signals to the master control block of the computer
control system. This total demand is compared to the sum of each operating
boiler's steam production. The computer control blocks for the fuel control
and air flow of the swing boiler are cascaded onto the master control block
and respond accordingly. The control blocks for the combustion air and fuel
valves automatically follow the action of the opposite control valve to keep
the air/fuel rates within limits. The boiler feed water flow controller
follows steam output and is also subject to a drum level control override, as
is the combustion air flow to flue gas 0^ with CO set point adjustment.
During the emissions testing of Units 4 and 5, steady-state conditions
with regard to load and excess air were required. Because of this, whichever
unit was being tested was operated manually and the other unit was kept on-
line to handle the swings in process steam demand. Fuel, air, and boiler feed
water flow rates were wet manually, with the drum level alarm set at a pre-
determined safe level.
All of the continuous tests for which emissions data are presented in
this report were performed at essentially steady-state conditions. When a
problem did occur, as in the No. 2 oil tests, testing was ceased and then
recommenced when the boiler upset had passed.
5.5 PROCESS CONDITIONS DURING TESTS
Table 34 summarizes the pertinent process data recorded by the plant
during the test runs. Process parameters included here are steam production,
economizer inlet temperature, stack .temperature, and stack 02 concentration.
Plant values for 0? concentration differ from the data presented in Tables 2
and 3 because of the differences in monitors and sampling locations. However,
these differences between the plant's and PEI's 02 data are consistent.
76
-------
TABLE 34. BOILER PROCESS DATA
Test
block
NG-5
NG-4
NG-1
NG-3
NG-2
NG-6
NG-9
NG-10
NG-8
NG-7
5
4
7
8
9
2
3
1
6
Boiler
unit
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
Fuel
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
No. 2 Oil
No. 2 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
Flue aas
02,a %
0.8
3.1
0.9
3.1
0.8
3.6
1.1
1.1
3.4
3.7
2.1
3.6
1.1
1.4
3.9
1.0
1.0
3.4
1.2
Special
conditions
_
_
-
-
-
-
_
Air preheat
-
-
Air preheat
Air preheat
_
-
-
_
Air preheat
-
Low viscosity
Steam flow,
1000 Ib/h
138
136
110
108
74
71
107
107
104
73
72
73
132
105
103
76
77
75
72
Econo-
mizer inlet
tempera-
ture, °F
683
695r
N/AC
640
N/A
560
686
687
706
608
637
654
N/A
N/A
769
641
611
N/A
629
Stack
tempera-
ture, °F
374
386
N/A
353
N/A
305
340
335
349
299
300
309
403
355
387
301
292
N/A
294
Stack
02,a %
0.5
3.1
0.8
3.0
0.5
3.5
0.7
0.5
2.7
2.9
1.5
3.1
0.8
0.9
3.1
0.8
0.8
3.2
0.9
Flue gas 02 denotes PEI monitor data and stack 02 denotes plant monitor data.
After the test program was completed, the plant calibrated the Unit 5 steam flow meter (orifice plate)
with a more accurate turbine flow meter. Based on this check and a review of some steam/header/feed
water flow rate data, we suspect the plant steam flow measurements are slightly low - ranging from
about 5000 Ib/h at i load up to 10-15,000 Ib/h at full load.
N/A means not available.
-------
REFERENCES
1. Mitchell, W. J., and M. R. Midgett. A Means to Evaluate the Performance
of Stationary Source Test Methods. Environmental Science and Technology,
10:85-88, 1976.
2. Oldaker, G. B. Condensible Particulate and Its Impacts on Particulate
Measurements. Draft Report. Prepared under EPA Contract No. 68-01-4148,
Task No. 69. May 1980.
3. Peters, E. T., and J. W. Adams. Sulfur Dioxide Interaction With Filters
Used for Method 5 Stack Sampling. In: Workshop Proceedings on Primary
Sulfate Emissions From Combustion Sources, Volume I - Measurement Tech-
nology. EPA-600/9/78-020a, 1978. pp. 199-202.
4. Gushing, K. W. Particulate Sampling in Process Streams in the Presence
of Sulfur Dioxide. In: Workshop Proceedings on Primary Sulfate Emis-
sions From Combustion Sources, Volume I - Measurement Technology. EPA-
600/9-78-020a, 1978. pp. 202-227.
5. PEDCo Environmental, Inc. Comparative Evaluation of EPA Methods 5 and
17. Draft Report. Prepared under EPA Contract No. 68-02-3431, Task Nos.
88, 103, and 163. February 1983.
6. PEDCo Environmental, Inc. Method Development and Testing for FCCU Regen-
erators. Final Reports. Prepared under EPA Contact No. 68-02-3546, Task
Nos. 14 and 20. February 1984.
7. PEDCo Environmental, Inc. Method Development and Testing for Boilers.
Draft Report. Prepared under EPA Contract No. 68-02-3546, Task No. 15,
June 1983.
78
------- |