vIEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EMB Report 86-SPB-2
March 1986
Air
Fluidized
Bed Boiler
Emission Test
Report
Canadian
Forces Base
Summerside,
Prince Edward
Island
Canada
-------
TRC-A87-192U)
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air Quality Planning and Standards
Emissions Measurement Branch
Research Triangle Park, NC
Contract No. 68-02-3851
Work Assignment No. 11
George Walsh - Project Officer
Dennis Holzschuh - Task Manager
EMISSION TEST REPORT
CANADIAN FORCES BASE
PRINCE EDWARD ISLAND, CANADA
VOLUME I
Final Report
EMB No. 86-SPB-2
May 1987
Prepared by
John M. Foley
Howard Schiff
ALLIANCE TECHNOLOGIES CORPORATION
213 Burlington Road
Bedford, Massachusetts 01730
(617) 275-9000
-------
DISCLAIMER
This Final Report was furnished to the Environmental Protection Agency by
the Alliance Technologies Corporation, Bedford, Massachusetts 01730, in
fulfillment of Contract No. 68-02-3851, Work Assignment No. 11. The opinions,
findings, and conclusions expressed are those of the authors and not
necessarily those of the Environmental Protection Agency or the cooperating
agencies. Mention of company or product names is not to be considered as an
endorsement by the Environmental Protection Agency.
-------
CONTENTS
Figures iv
Tables v
1.0 Introduction 1
2.0 Summary and Discussion of Results 2
3.0 Process Description and Operation 8
3.1 General Information 8
3.2 Design Information 8
3.3 Operating Information 10
3.4 Operating Procedures 15
3.5 Process Operation 16
3.6 Process Data 18
4.0 Sampling Location 25
5.0 Sampling and Analytical Procedures 28
5.1 Overview 28
5.2 Measurement of Flue Gas Emissions 28
6.0 Program Quality Assurance 36
6.1 Introduction 36
6.2 Precision, Accuracy and Completeness 36
Appendices
A. 30-Day Hourly CEM Emissions Summary A-l
B. 30-Day Coal Analyses Results B-l
C. 30-Day Limestone Analyses Results C-l
D. Process Generated Materials' Analytical Results D-l
E. Hourly Field Coal Analysis Lob E-l
F. Relative Accuracy Field and Reduced Data Sheets F-l
G. Zero and Calibration Drift Data Sheets G-l
H. Field Team's Daily Log H-l
I. Daily Calibration Log 1-1
J. Calibration Gas Certification Reports J-l
K. 30-Day Digistrip Data Log (Volume II) K-l
111
-------
FIGURES
Number Page
3-1 Foster Wheeler boiler at Prince Edward Island , . 9
3-2 Flow diagrams for coal, wood and limestone 12
3-3 FBC layout 14
3-4 Emission source test summary 21
4-1 Schematic of Sampling Locations 26
4-2 Stack sample traverse point locations 27
5-1 GEMS program schedule 29
5-2 EMR Continuous Emissions Monitor 30
-------
TABLES
Number Page
2-1A Results of CEM Relative Accuracy Tests 3
2-1B Summary of Measurement Values 4
2-2 Summary of Analysis Procedures Performed on Samples
Obtained During 30-Day CEM Program 6
2-3 Summary of Sieve Analyses 7
3-1 Baghouse Design Data 11
3-2 Operating Anomalies During FBC Emissions Tests at
Prince Edward Island 19
3-3 Daily Average Process Flow Rates 23
3-4 Daily Average Process Parameters and Emission Results .... 24
5-1 Analyzer Specifications 31
5-2 Calibration Gas Concentrations 31
5-3 CEMS System—Monitor Response Time Results (Seconds) 33
5-4 Analytical Procedures 35
6-1 Performance Specification Requirements 37
6-2 Results of CEM Relative Accuracy Tests 38
6-3 Results of EPA Cylinder Gas Audit 40
6-4 Results of EPA Reference Method 6 Audit 40
-------
SECTION 1.0
INTRODUCTION
The objective of this test program was to demonstrate the sulfur dioxide
(802) and nitrogen oxide (NOX) control capabilities of a fluidized bed
combustion (FBC) unit over an extended period of time (30 days). The data
presented in this report will be used to develop and support New Source
Performance Standards for small steam-generating units (100 million Btu/hr or
less heat input).
The test program was conducted at an FBC facility located at the Canadian
Forces Base (CFB), Summerside, Prince Edward Island, Canada. The facility has
two FBC boilers that generate a maximum of 40,000 Ib steam/hr each. These
boilers supply space heating to the base and are operated as peak load units
from September through May. One of the boilers was used for testing purposes.
Emissions from the unit were characterized utilizing a Continuous
Emissions Monitoring System (GEMS) provided by Energy Mines and Resources
(EMR)/Canada. Oxygen, carbon dioxide, carbon monoxide, oxides of nitrogen,
and sulfur dioxide were monitored at the stack location for 30 consecutive
days.
In addition to gaseous emissions monitoring, coal and limestone samples
were obtained from the unit's solids handling systems on a regular basis.
These samples were analyzed by Spotts, Stevens & McCoy Laboratories, Reading,
PA. These data, in conjunction with the CEMS data and process logs, will
allow a comprehensive evaluation of the operational characteristics of the FBC
unit.
This report summarizes the emissions test program conducted by Alliance
Technologies Corporation (formerly GCA Technology Division, Inc.) on the FBC
facility from February 17 through March 28, 1986. In addition to the data
base generated by 30 days of CEMS data, the results of coal and limestone
analyses, pertinent field logs, and data related to CEMS certification are
included in this report.
-------
SECTION 2.0
SUMMARY AND DISCUSSION OF RESULTS
The fluidized bed combustion unit (FBC) located at the Canadian Forces
Base at Summerside, Prince Edward Island, Canada, underwent a 30-day
continuous emission monitoring (CEM) period during February and March, 1986.
Sulfur dioxide, oxides of nitrogen, carbon monoxide, carbon dioxide, and
oxygen concentrations were monitored by the CEM system in the flue gas stream
at the final emission point downstream of the baghouse.
The 30-day CEM data summaries are found in Appendix A. These data result
from the raw data obtained and stored by the Kaye Digistrip III in the field.
Due to the volume of this raw data, only the summary tables are included in
this report. The raw data file, also containing operational data, is found in
Volume II, Appendix K of this report. Certification of the CEM system was
performed both prior to and following completion of the 30-day test program in
accordance with procedures outlined in 40 CFR Part 60 Appendices A, B, Da,
and F (proposed). Certification results can be found in Section 6 of this
report and are summarized in Table 2-1A.
Concurrent with the continuous monitoring, coal and limestone feed lines
to the FBC were also sampled during three 8-hour shifts. The coal samples
were taken every half-hour during the 24-hour period. These samples were
composited over an 8-hour period to form three composite samples for each
calendar day. The limestone samples taken during the 24-hour period were
composited to form a single, daily composite sample for each calendar day. In
addition to the CEM and coal and limestone sampling, ash samples were obtained
from the FBC unit at various points in the emission control device system
(i.e., bed drains, cyclone reinfection point, and baghouse). The ash samples
were collected at least once a week at selected intervals during the 30-day
period. Analytical results for the coal, limestone, and ash can be found in
Appendices B through D.
The data sets presented in Appendices A through E will be utilized by
others in computing pollutant emission rates, sulfur capture efficiency of the
unit, and a correlation of the emissions with fuel sulfur content and process
operating conditions. A summary of these average values will be presented in
this section, as well as, the ranges of sulfur content in the coal limestone
and ash, and the S02, NOX, CO, C02, and ©2 concentrations in the flue
gas are given in Table 2-1B.
-------
TABLE 2-1A. RESULTS OF GEM RELATIVE ACCURACY3 TESTS
Specification Initial set
(percent) (percent)
9 runs
02 0.5 0.45 02 c
C02 1.0 0.39 C02 c
CO 10.0 5.4a
NOX 20.0 7.8a
S02 20.0 10. 4a
Final setb
(percent)
3 runs
0.30 02 d
0.07 C02 d
13. 5b
8.1b
20. 4b
a |_
RA = 1 d. + cc x 100
RM
RA = | d 1 x 100
RM
AA =|dl-
cc
AA
Where:
RA = Relative accuracy.
AA = Absolute accuracy.
d = Algebraic average of differences between GEMS
value and Reference Method value.
cc = 95% confidence coefficient.
RM = Reference Method value.
-------
TABLE 2-IB. SUMMARY OF MEASUREMENT VALUES
Parameter Value Range . Mean Value
Coal, sulfur content, % dry basis 5.34 - 7.62
% wet basis 4.74 - 7.10 5.89
Limestone, purity 96.4 - 98.5 97.7
sulfur content, % <0.01 <0.01
Ash, sulfur content
Bed material 11.1 - 14.4 13.00
Bed drain material 7.08 - 18.34 7.70
Multicyclone ash for reinjection 7.02 - 9.46 8.38
Baghouse ash 5.73 - 7.31 6.90
Ash silo material 7.23 - 7.78 7.51
Flue gas
Sulfur dioxide, ppm S02 2 - 1,658 226
Oxides of nitrogen, ppm NOX 180 - 340 270
Carbon monoxide, ppm CO 330 - 1,060 481
Carbon dioxide, % C02 4.1 - 12.5 9.4
Oxygen, % 02 9.7 - 12.9 10.7
-------
A single limestone sample was tested for reactivity by Westinghouse
Electric Corporation, Research and Development Center, located in Pittsburgh,
PA. In the Westinghouse thermogravimetric reactivity te«t, the limestone was
found to be highly reactive with SC^- The Fluidized Bed Combustor sulfur
capture model estimates that 90 percent sulfur removal would be achieved with
a calcium to sul-iur ratio of 3:1. Information and correspondence relating to
the analysis, as well as results, are provided in Appendix C-3. The analyses
of the ash samples showed a relatively high percentage of sulfur present, as
shown in Table 2-1B. The overall results of the ash analyses can be found in
Appendix D.
In addition to the emissions data, results of coal and limestone analyses
are provided. Coal and limestone samples were composited over three 8-hour
intervals/day U = 0000-0800; t = 0800-1600; t = 1600-2400). Samples were
bagged, labelled, and prepared for shipment onsite. Analyses were conducted
by Spotts, Stevens & McCoy, Inc. of Reading, PA.
Daily coal samples obtained during the first 8-hour interval
(t = 1600-2400) were subjected to moisture, higher heatiog value, and sulfur
analyses (see Appendix B-3). Also, on a weekly basis (days 1, 7, 14, 24,
and 29 of the CEM program), a single coal sample for the t = 0000-0800
composite sample was subjected to proximate, ultimate, higher heating value,
sieve, and ash speciation analyses. These data are provided in Appendix B-l.
The three 8-hour limestone samples generated per day were composited to
create a single sample for each day of testing. These samples were subjected
to calcium carbonate and sulfur analyses; results of the analyses are provided
in Appendix C-2. One daily composite sample/week was selected to undergo
calcium carbonate, sulfur, and sieve analyses. These data are provided in
Appendix C-l.
Finally, samples of bed material, bed drain material, multicyclone ash
for reinjection, baghouse ash, and ash silo material were obtained at various
times during the test program. These samples provide additional information
regarding the combustion characteristics. The samples underwent ultimate
analysis, and the results are located in Appendix D. Table 2-2 provides a
breakdown of the analyses procedures performed on all samples obtained during
the 30-day test program. Samples of coal, limestone and ash material were
subjected to sieve analyses which are summarized in Table 2-3.
-------
TABLE 2-2. SUMMARY OF ANALYSIS PROCEDURES PERFORMED ON SAMPLES OBTAINED DURING 30 DAY GEM PROGRAM
Analysis Type
Daily
Coal
Daily
Limestone
Weekly
Coal
Sample Description
Bed
Weekly Bed Drain Cyclone
Limestone Material Material Ash
Multlcyclone
Ash for
Reinjection
Baghouse. Ash Silo
Ash Material
Proximate 1 ,
Ultimate 1
HHVa 1,2,3
Sieve
Ash Speciation
Moisture 2,3
Sulfur 2,3
CaC03
Reactivity 4e
I - Samples obtained over the t - 0000-0800 composite interval.
2 - Samples obtained over the t - 0800-1600 composite interval.
3 - Samples obtained over the t - 1600-2400 composite interval.
4 - Samples are a composite of the three daily composite samples; effective sample interval t =0000-2400.
5 - Miscellaneous grab samples.
aHigher heating value, Btu/lb
bNote exceptions: 3/21/86 t = 1600-2400 composite analyzed.
C0ne sample.
dNote exceptions: 3/07/86 t = 0000-0800 composite analyzed.
3/08/86 t = 0800-2400 composite analyzed.
3/18/86 t = 0000-2400 composite analyzed.
elndicates that a single 24-hour composite sample (3/5/86) underwent the reactivity test.
-------
TABLE 2-3. SUMMARY OF SIEVE ANALYSES
Coal
Retained on:
Passing:
Limestone
Retained on:
Passing :
Asha
Retained on:
Passing:
•-
3/4", %
1/2", %
3/8", %
1/4", %
4 mesh (0.185"), %
8 mesh (0.09"), %
8 mesh (0.09"), %
3.320 mm - 6 mesh, %
2.000 mm - 9 mesh, %
1.397 mm - 12 mesh, %
1.190 mm - 14 mesh, %
0.133 mm - 20 mesh, %
0.595 mm - 28 mesh, %
0.595 mm - 28 mesh, %
4 mesh (0.185") (4.7 mm), %
8 mesh (0.093") (2.362 mm), %
10 mesh (0.065") (1.65 mm), %
20 mesh (0.0328") (0.833 mm), %
50 mesh (0.0117") (0.297 mm), %
100 mesh (0.0058") (0.147 mm), %
100 mesh (0.0058") (0.147 mm), %
Mean value
21.2
26.7
14.6
12.7
5.0
8.5
11.4
7.3
44.2
32.9
7.3
3.6
7.1
3.6
Multicyclone ash
for reinjection
0.0
0.0
0,0
0,0
5.8
14. ,7
79,5
Value Range
13.8
8.9
8.9
9.2
2.0
1.5
2.8
2.3
40.0
16.3
2.7
2.4
0.4
1.0
- 31.0
- 45.0
- 18.4
- 17.3
- 8.8
- 23.6
- 2.4
- 23.3
- 49.8
- 38.4
- 10.1
- 5.3
- 3.1
- 10.5
Baghouse ash
0.0
0.0
0.0
0.0
0.8
13.9
85.3
aSingle sample of multiclone ash and baghouse ash underwent sieve analysis.
-------
SECTION 3.0
PROCESS DESCRIPTION AND OPERATION
This section presents a process description of the fluidized bed
combustion (FBC) test unit. Included are characterizations of the boiler/
auxiliary equipment design and operation, discussions of the boiler operating
history, control procedures, and process test conditions.
3.1 GENERAL INFORMATION
The two FBC boilers located at the Summerside, Prince Edward Island site
were built to provide 40,000 Ib/hr of steam each for space heating at the
Canadian Forces Base , and are operated as full-scale demonstration units by
the Canadian government. The two identical FBC boilers (Unit Nos. 1 and 2)
are field-erected, bubbling bed units manufactured by Foster Wheeler (FW)
Canada Ltd. The emission test data were gathered on Unit No. 2.
Two 25+ year old dump and grate stoker boilers are also located at the
Summerside site. The two stoker units have a combined steam production
capacity of 35,000 Ib/hr. The two FBC units were purchased to meet the
additional steam needed for Base heating and are operated as peak load units.
Fluidized bed combustion was chosen because the Canadian government wanted to
demonstrate on a commercial scale that FBC is a viable, clean, "off oil"
technology.
During early operation of the FBC units, attention was drawn to the great
amount of make-up water required by the boilers. This concern led to a steam
leak repair program in the steam tunnels of the Base, resulting in a
significant reduction in the steam demand. The reduction was so large that
typically, only one FBC unit is needed on-line to accommodate the additional
steam demand.
3.2 DESIGN INFORMATION
A side view of one of the FBC boiler units located at the site is shown
in Figure 3-1. Each of the atmospheric fluidized bed (ABF) boilers is
designed to produce 40,000 Ib/hr of saturated steam at 110 to 140 Ib/in^
(psi). The heat input capacity of each unit is rated at 50 million Btu/hr.
The fluidized beds are operated at temperatures near 1560°F (850°C). The
units were guaranteed by FW Canada Ltd. to achieve greater than or equal to
80 percent boiler efficiency. In actual testing, the boiler efficiency has
been measured at 83 percent. Each FBC boiler consists of two combustion beds,
A and B, each fed by its own overbed spreader stoker. In Figure 3-1,
8
-------
C MaiTlOB n tXi »»t
Xll.it.tl
/ ^
MiWf* ' (. M» W4» M«
VT
IL • I I
^
11 • «i
Figure 3-1. Foster Wheeler boiler at Prince Edward Island.
9
-------
combustion bed A is indicated as the preferential bed, while combustion bed B
is indicated as the secondary bed. The two beds are divided by a waterwall
which has a 1 ft^ opening located at the bottom center of the wall. This
opening allows bed material to circulate between the two beds when both are in
operation. This two-bed system allows for greater flexibility in turndown. A
turndown ratio of. 8:1 has been achieved during regular operation by taking
bed B off-line.
As shown in Figure 3-1, tubes are located both in and above the bed. The
total heating surface is 9,930 ft^. In-bed tubes are mounted vertically
along the walls and then diagonally across the unit. Above-bed tubes are
mounted vertically along the wall. When fluidized, the bed depth is
approximately 4.5 ft and the freeboard extends 22 ft above the bed. The
slumped (non-fluidized) bed height is 2 ft. Bed A is 4 ft wide and 9.5 ft
long, while bed B is 4.5 ft wide and 9.5 ft long.
Each FBC unit is equipped with a fly ash reinjection system. All of the
fly ash collected in the convective tube bank and by the mechanical dust
collector is recycled back to bed A. The mechanical dust collector is a
multi-tube cyclone, housing five small-diameter cyclone tubes which operate in
parallel. The multi-tube cyclone was manufactured by Enviro Systems and
Research, and was designed to capture particles greater than 40 microns in
diameter.
Each FBC unit is also equipped with a baghouse unit designed to capture
nearly all of the particulate matter remaining in the flue gas exiting the
cyclone. The baghouse was also manufactured by Enviro Systems and Research.
The design data for these units are shown in Table 3-1.
3.3 OPERATING INFORMATION
Annual operation is from early September to late May, and the typical
load over this heating period is 75 to 80 percent of maximum design load.
During operation of the boiler, coal and limestone are trucked to the facility
and dumped into separate underground hoppers as shown in Figure 3-2. The coal
and limestone are sifted by screens and the larger sizes are re-crushed before
being carried by bucket elevators to onsite storage bunkers. Bucket elevators
transport the coal and limestone from the storage bunkers to the respective
coal and limestone day bins located above the boiler unit. From the day bins,
the coal and limestone flow by gravity to weigh feeders. The coal feed system
is equipped with, one weigh feeder which delivers the coal to a pant leg, where
the coal flows by gravity and is distributed to separate stoker feeders, one
feeder for each bed. The limestone feed system is equipped with a separate
weigh feeder for each bed. From the weigh feeders, the limestone flows by
gravity through separate feed pipes to beds A and B.
The coal normally fired in the FBC units, as well as during the emission
testing period, is an unwashed eastern bituminous coal which has been sized to
1 in. x 0. The coal has an average heating value of 11,500 Btu/lb and an
average sulfur content of 5.5 percent. This corresponds to potential S02
emissions of 9.57/million Btu on an uncontrolled basis. Typical ash,
10
-------
TABLE 3-1. BAGHOUSE DESIGN DATA
Flue Gas
Inlet gas volume
Inlet gas moisture
17,278 ACFMd @ 350°F
8 Wt. Percent
Dust
Loading
Particle size
2.4 gr/SCFD
40 Percent minus 10 microns
Number of cells
Number of bags
Total cloth area
Air-to-cloth ratio
12/Unit
36/Cell ,,
4,898 Ft*
3.52 (All cells on-line)
Operating Information
Collector design pressure
Particulate matter emission
(Baghouse outlet)
Collection efficiency
±20 in. water gauge
0.02 gr/SCF .
99.2 Percent'
ACFM = actual cubic feet per minute.
gr/SCF = grains per standard cubic foot.
GCollection efficiency calculated on the basis of the inlet loading
and outlet emissions.
11
-------
v
1
Coa
V
I Bun
v
I
kers
t
V-
No. 1 Boiler ^ •
U
^
\r~
v
I
We
Coal
Bins
} (
sigh Feed
} /
/
i
3rs
t
1
^ —
V
i
Coa
V
Bun
V
*
Kers
V
L-r '
. J No. 2 Boiler
Flow Diagram • Coal and Wood
r^
Scree
?S
1
_ L
C
.t
_
-|
J
I
Bunker
\s.
L
t
-^
— _l
pr^
1 1
Limestone Bins
VY^ V
T ^T^I * j
O_L>j (cj cj
Weigh Feeders
ZI
Y
Jn
1
1
«—
1_—
n7
i
Bunker
V
1
Flow Diagram • Limestone
MATERIALS HANDLING - RATES AND CAPACITIES FOR EACH FBC UNIT
Transfer to receiving hoppers
to bunkers
Bunkers capacity
Transfer to bins
Bins capacity
Burn rate
COAL
25 tons/hour
4 x B2.5 tons
5 tons/hour
2 * 5 tons
2.5 tons/hour
LIMESTONE
25 tons/hour
2x7$ tons
2.5 tons/hour
2 x 2.5 tons
1.13 tons/hour
Figure 3-2. Flow diagrams for coal, wood and limestone.
12
-------
nitrogen, and moisture contents are 19.3, 1.0, and 0 to ].0 percent,
respectively. The fixed carbon is 47.2 percent and the volatile matter is
33.5 percent.
The limestone normally used at the site is Havelock limestone with a
calcium carbonate- (CaCC^) content of 94 to 95 percent. Currently, it is
sized to 0.0934 in. x 0.0331 in. (8 x 20 mesh) with certain percentages at 10,
12, and 15 mesh. The FBC unit can also utilize a limestone sized to
1/4 in. x 0, which contains a higher percentage of fines than does the 8 x 20
mesh limestone. Limestone is used as the sorbent for purposes of sulfur
dioxide (SO2) removal. As discussed in Section 3.4, limestone is added to
the bed as one method to control bed level. During normal operation, the
limestone-to-coal feed ratio is maintained near 1 to 5, which corresponds to a
calcium-to-sulfur (Ca/S) ratio of approximately 1.1:1 baaed on a 5.5 percent
sulfur coal. SC>2 reductions usually range from about 40 to 50 percent at
this operating condition. This unit was designed to achieve 83 percent SC>2
reduction at Ca/S ratios ranging from 2.4 and 4.1, depending on the fuel
burned.
The material handling system shown in Figure 3-2 can also handle wood
chips. The FBC units can combust wood, but only 50 percent load can be
achieved due to size limitations of the wood feed system. (The original
design called for 30 percent of the heat input to be from wood chips.)
Screw conveyors are operated under the beds to remove ash and sulfated
limestone (i.e., bed material). The desired bed level is maintained by
adjusting either the limestone feed rate or the screw conveyor discharge
rate. Each bed is equipped with a separate drain pipe which allows bed
material to drain by gravity when the screw conveyors are operated.
The FBC layout is shown in Figure 3-3. Combustion air passes through a
forced draft (FD) fan and enters the plenum. The air flow is distributed
across the two beds at fluidizing velocity by means of distributor nozzles
which are uniformly spaced across the base of each bed. Fluidized bed
material freely circulates through the door in the water wall which separates
beds A and B. This movement of material, as the two beds; seek a common level,
promotes fuel and sorbent mixing and improves the operational stability of the
boiler.
Hot flue gas leaves the bed at approximately 1560°F (850°C), passes
through the freeboard section, and flows across the conv€:ctive tube bank
before exiting the boiler. All of the fly ash collected in the convective
tube bank is gravity fed back into bed A.
The flue gas exiting the boiler enters the mechanical dust collector
where fly ash is captured and gravity fed through a straight leg reinjection
system back into bed A. Spargers inject air into the reinjection leg at the
reinjection point to assist fly ash recycle and prevent plugging of the
reinjection leg.
13
-------
NORTH
\ I
1. CONTROL ROOM
2. F.G. FAN
3. BOILER
4. MECHANICAL OUST COLLECTOR
5. ECONOMIZER
6. BAGHOUSE
7. I.D. FAN
8. STACK
9. COAL RECEIVING HOPPER
10. COAL BUNKER
11. COAL BIN
12. LIMESTONE RECEIVING HOPPER
13. LIMESTONE BUNKER
14. LIMESTONE BIN
15. WOOD CHIP BIN
16. DE-AERATOR
17. DIESEL GENERATOR
18. LIGHT OIL TANK
19. OFFICE
20. ASH SILO
Figure 3-3. FBC layout.
14
-------
The flue gas then passes through the finned-tube economizer where it is
cooled before entering the baghouse. In the economizer, the feed water
temperature is raised from 220°F (105°C) to 275°F (135°C). In the baghouse,
nearly all remaining particulate matter is removed. From the baghouse, the
flue gas is drawn through an induced draft (ID) fan and exhausted to the stack.
The ash which is collected from under the bed and in the baghouse is
transported in a vacuum-type pneumatic handling system to the ash silo. From
the ash silo, the waste is trucked to disposal trenches located on the Base.
3.4 OPERATING PROCEDURES
During normal operation, the boiler plant master controller operates in
the automatic mode and regulates the steam output from the operating boilers.
The controller modulates the total steam flow from the boiler plant to
maintain constant pressure in the common steam header to the Base. During
peak heating periods, when more than one boiler is operating, the control
strategy is to operate one unit at constant load conditions while the peak
load unit is allowed to track the excess steam demand. This control strategy
is accomplished by setting the boiler master controller for the constant load
unit in manual mode, and operating the boiler master controller for the peak
load unit in automatic mode.
As mentioned earlier, the FBC boilers are normally operated as peak load
units. Operating in the automatic mode, the FBC master controller adjusts the
coal feed rate to the unit to compensate for swings in the plant steam
demand. The combustion air controller, sensing the coal rate change, adjusts
the total air flow to the unit to maintain a constant air-to-coal ratio. The
total air flow is further adjusted by an oxygen trim controller, as necessary,
to maintain the bed at optimum sulfation temperature.
The FBC unit design control scheme calls for the limestone feed
controllers to automatically track the stack gas S0£ concentration and
adjust the sorbent feed rates to beds A and B based on an SO2 emission
set-point. However, this loop was never sufficiently tuned to allow fully
automatic control and, therefore, has been by-passed as part of the overall
FBC unit control scheme. During normal operation, the limestone feed
controllers are operated in manual mode, and the limestone feed rates to
beds A and B are adjusted by operating personnel to maintain desired bed
levels.
The boiler feed water flow responds to the steam output requirements.
The boiler feed water flow rate is also subject to control by the steam drum
level which is a slower control loop. The steam drum level is controlled at
the mid-level height of the drum.
15
-------
3.5 PROCESS OPERATION
The primary operating objective of the emission source test was to
demonstrate long-term FBC system performance of approximately 93 percent S02
reduction. Secondary operating objectives of the test were: 1) to obtain
system performance data demonstrating the effect of limestone particle size on
sulfation capacity, and 2) to determine the maximum S02 reduction achievable
during stable two-bed unit operation.
During the first 3 weeks of the testing period, the FBC unit was operated
at near 70 percent load, and the limestone feed rate was manually adjusted to
maintain the S02 removal efficiency at the targeted 93 percent level. As
discussed in Section 3.4, the control system between the S02 outlet flue gas
concentration and the limestone feeders was never sufficiently tuned to allow
fully-automatic control of S02 emissions using the limestone feed rate. As
a result, the limestone feeders to beds A and B were controlled manually by
plant operators.
A target SO2 emissions level was established consistent with the feed
coal sulfur and heating values and desired S02 reduction levels. While it
would have been physically possible for the plant operators to adjust the
limestone feed rates on a near-continuous basis to match the target 862
emissions level and load changes, this would have required an operator to be
dedicated to limestone feed adjustments. This was impractical and was
inconsistent with normal plant operation in which limestone feed rates were
set manually, and only adjusted infrequently.
Another approach was to hold the unit load at a steady level so that a
relatively constant Ca/S feed ratio could be maintained. This was the control
approach used for all but the last week of the emission source test. It was
desired that the FBC unit operate at a high a load as was practicable in order
to fully test the capabilities of the limestone feed system, solid waste
withdrawal system, and sulfur capture reactions in the bed. Unfortunately,
weather conditions during the emissions test period were unseasonably mild,
such that Base steam demand was below normal. To operate the FBC system at
loads in the range of 80 to 90 percent would have required the venting of
steam during much of the test period, at considerable cost to the plant. A
unit load at or above 70 percent was selected as a reasonable compromise
between the desire to demonstrate FBC system performance under the adverse
conditions of relatively high load, and the economic penalties of venting
large quantities, of steam. During short periods when the steam demand
exceeded that available from the FBC unit operating at or above 70 percent
load, the dump and grate stokers were quickly brought on-line to meet the
extra demand.
The actual steam output of the FBC boiler was not known because the steam
flow rate recorder was out of calibration and was considered unreliable for
measuring actual flows. However, the recorder was considered reliable for
indicating flow trends. The unit load target of 70 percent was maintained on
the basis of the coal feed rate, the heating value of the coal, and the known
maximum heat input capacity of the FBC boiler.
16
-------
For these reasons, the control strategy used during emissions testing was
to set the coal feed rate to the test unit to a level approximating a
70 percent load condition, and to set the limestone feeders to achieve a
target SC>2 emissions level. During the emission testing., the bed level was
maintained by adjusting the screw conveyor bed drain discharge rate. The dump
and grate stoker -load or the steam vent rate was then adjusted as necessary to
allow the FBC unit to operate at the steady load condition during the test
period. The boiler combustion air and feed water controls were operated in
the automatic mode during the test period to ensure smooth boiler operation at
the constant load condition.
Variations from the desired load condition during testing were due to
fluctuating coal characteristics such as moisture content: and bulk density.
At constant controller settings, the bulk density affected the coal delivery
rate to the unit while increased coal moisture content lowered the effective
heat input to the unit. Operating personnel monitored flue gas SC>2
concentration trends during the test period and adjusted the limestone feed
rates step-wise, as necessary to maintain desired S(>2 reduction performance.
During the last week of the source testing period, two short-term
parametric tests were conducted. During the first parametric test, system
performance data were gathered over an 8-hour period, while 1/4 in. x 0
limestone was fed to the unit. The 1/4 in. x 0 limestone: evaluated during
this period was obtained from the same mine as the 0.0934 in. x 0.0331 in.
(8 x 20 mesh) limestone used during the remainder of the source testing. The
data gathered during this period were used to compare the 862 reduction
efficiency of the two limestones at equivalent Ca/S ratios.
During the second parametric test, the unit was operated at high Ca/S
ratios (by increasing the limestone feed rate) to determine if mass transfer
restrictions would prevent the FBC unit from operating at: S02 flue gas
concentrations in the range of 20 to 40 ppm(v). For a unit firing low sulfur
coal ( 2 percent sulfur), the flue gas SO2 concentration must be maintained
in this range to achieve approximately 93 percent SC>2 reduction. The boiler
was operated at near 50 percent load during this 12-hour test to avoid
overloading the baghouse at the high Ca/S ratios. The dump and grate stokers
were brought on-line as necessary to handle any additional steam demand.
Due to inclement weather conditions, electrical power to the Base was
interrupted four times during the source testing period. When a power outage
occurred, a diesel-fired generating system at the plant was started and
brought on-line to provide electricity for the FBC boiler operation. The
switching process from main power to back-up power required a few minutes to
execute; as a result, all feed flows to the unit were interrupted during this
period. As the back-up power was brought on-line, controller settings were
restored to their pre-outage positions. However, the limestone feed was
discontinued until the coal and combustion air rates had been adjusted to
bring the bed temperatures back within desired operating ranges.
The system operating performance and SO2 reduction performance were
unsteady while the unit was operated on back-up power due; to the process
upsets (e.g., interruption of feed flows). Also, the electricity generated by
17
-------
the diesel system was at a slightly different frequency than that of the main
power system. This affected the performance and reliability of sensitive
electronic equipment such as the feed totalizers.
Once main power was restored to the facility, the same switching process
was repeated as the back-up power supply was taken off-line and main power to
the boiler was restored. The fuel and combustion air feed flows were again
interrupted for a short period, and the limestone feed was discontinued until
bed temperatures stabilized within operating ranges. After the return to main
power, a period of 8 to 15 hours was normally required for coal and limestone
feed rates to stabilize and for the FBC system to return to near steady
operation.
Due to the frequency of power outages at this facility during winter
storm periods, a more advanced back-up system will be installed in the future
to instantaneously provide electricity to the unit during power outage
periods. The new back-up system will allow the unit to operate smoothly
during the transition from main to back-up power supply.
For these reasons, the data gathered during periods of back-up power
generation were not considered to be representative test data. Once the unit
was restored to main power and the system performance steadied to set
conditions, testing was recommenced. However, the data gathered during
process upset periods caused by unavoidable euipment malfunctions and
operating anomalies other than power outages, were considered representative
test data. Data gathered during these periods were evaluated in order to
obtain an accurate estimate of achievable long-term SO2 reduction
performance during normal day-to-day operation. A summary of the operating
anomalies encountered during the emissions source test is included in
Table 3-2.
A summary of the FBC unit status for the source testing period is shown
in Figure 3-4. Power outages occurred on the following dates: March 6,
March 9, March 15, and March 18. On March 7, a severe winter storm forced the
Base to be closed to all non-essential personnel. The ash hauling trucks,
therefore, were not available to transfer waste ash from the ash silos to
disposal trenches located on the Base. To avoid overloading the ash silos
during this period, the boiler was operated at low limestone feed rates. The
unit was returned to test conditions on the morning of March 8. The March 18
power outage resulted in the formation of clinkers in bed B of the test unit.
The clinkers prevented fluidization of bed B, making this bed inoperable. The
unit was taken off-line and allowed to cool so that operating personnel could
remove the clinkers. The test unit was restarted on March 21 and was
operating at test conditions by approximately 12:30 p.m. During the week
after restart-up, the parametric tests associated with the secondary operating
objectives were conducted.
3.6 PROCESS DATA
During the source testing period, process data were logged automatically
and printed every 15 minutes. The recorded data reflect the instantaneous
process conditions at the time the data were logged. The recorded process
18
-------
TABLE 3-2. OPERATING ANOMALIES DURING FBC EMISSIONS TEST AT PRINCE EDWARD ISLAND
From:
To:
Date Time Date
lime
Duration
(Days)
Operating Anomaly
2/28 2300 2/29 0800 0.38
3/5 0900 3/6 0830 0.98
3/9 1800 3/10 0900 0.63
3/16 0800 3/16 1045 0.11
3/16 2200 3/16 2315 0.05
Operators bypassed flue gas around the baghousje. The
baghouse was operating at high differential pressures
(6 to 6.5 inches of HpO). Operators were manually
cleaning the ash from the baghouse.
General inspection. Coal rate to the boiler was
lowered such that no steam was vented; load was reduced
to about 50 percent. During the early morning of
March 3, load was at 66 percent.
Limestone feed rate did not respond to increased
limestone control settings. Consequently, S02
emissions were high (300 to 400 ppm).
One of the pins holding a coal bucket came loose in
primary bucket elevator. While switching coal flow to
the other bucket elevator, the coal supply was
exhausted. Began feeding coal from emergency
stockpile. Coal weigh feeder jammed and bed
temperature varied more than normal after switching to
emergency coal due to high moisture and fines content.
Very wet coal from emergency stockpile extinguished
furnace flames. Limestone feed was discontinued until
furnace fire was restarted and the bed temperature
returned to 1560°F (850°C).
-------
TABLE 3-2 (continued)
From:
To:
Date
Time Date Time
Duration
(Days)
Operating Anomaly
3/16 2315 3/18 0945 1.44
3/21 1330 3/22 0900 0.81
to
o
3/25 1345 - 3/25 1400 0.01
Operated with emergency coal until new coal shipment
arrived on Base. Operating problems occurred with coal
handling and feeding systems.
Discontinued limestone feed due to artifically high
baghouse differential pressures (between 7 and 7.5
inches of H20). The high differential pressure was
caused by plugging in the pneumatic line of the
controller.
Replaced faulty electrical board in coal weigh feeder
totalizer. Coal feed rate readings from March 21 @
0945 to March 25 @ 1400 were artifically low.
-------
Ul
0000 •
0200 -
0400 -
0600
0600
1000
1200
1400
1600 1
1800
2000
2200-
2400
DATE
2/27 2/28 3/1 3/23/3 3/43/53/63/7 3/83/9 3/10 3/11 3/12 3/13 3/14 3/15 3/16 3/17 3/18 3/19 3/20 3/21 3/22 3/23 3/24 3/25 3/26 3/27 3/28
Unit Operating at Test Conditions B Unit Operating but not at
S Test Conditions
Unit not Operating - down for Repairs
Figure 3-4. Emission source test summary.
-------
parameters included Ithe coal feed rate, limestone feed rate, total air flow,
bed temperatures, average freeboard temperature, boiler exit gas temperature,
stack gas temperature, steam flow and pressure, economizer inlet oxygen
content, and pressure drop across the beds. Also included were fan suction
and discharge pressures, flue gas operating pressures, boiler feed water
temperature and pressure, and steam drum level. A copy of the process data
recorded by the plant data logger during the emissions test period is included
in Appendix K.
The coal, limestone, and steam flow rates were also indicated by
electronic totalizers. Coal and limestone totalizer readings were recorded on
an hourly basis during day-shift operation in order to compute average hourly
feed rates during these periods. Coal, limestone, and steam totalizer
readings were also recorded at approximately midnight each day by operating
personnel in order to compute the total flow for the 24-hour period.
The average hourly coal, limestone, and steam flow rates during the
testing period are summarized in Table 3-3. The flow rates shown in the table
were calculated from the midnight totalizer readings recorded by the operating
personnel and, therefore, represent average hourly rates based on total daily
flows. On test days interrupted by power outages, the average coal and
limestone rates shown in Table 3-3 were computed from the midnight and latest
available day-shift readings prior to the power outage. The coal and
limestone rates, therefore, represent hourly flow rate averages indicative of
test operation. However, the average steam rates computed during test periods
interrupted by power outages also include steam output during off-test
conditions. These data, therefore, were not included in the Table 3-3 summary.
As mentioned in Section 3.5, the recorded steam rates, although
considered inaccurate on an exact rate basis, were useful in determining unit
operating trends. Table 3-4 presents the daily average process parameters
affecting emissions, as well as the daily average emission results.
22
-------
TABLE 3-3. DAILY AVERAGE PROCESS FLOW RATES
Daily Hours of
Date Test Operation
Average
Coal Feed
Rate (Ib/hr)
Average
Limestone Feed
Rate (Ib/hr)
Average
Steam Production
Rate (Ib/hr)
2/27
2/28
3/1
3/2
3/3
3/4
3/5
3/6
3/7
3/8
3/9
3/10
3/11
3/12
3/13
3/14
3/15
3/16
3/17
3/18
3/19
3/20
3/21
3/22
3/23
3/24
3/25
3/26
3/27
3/28
24
24
24
24
24
24
24
12.5
0
16
9
24
24
24
24
24
20
24
24
9a5
°a
oa
11.5
24
24
24
24
24
24
12
3010
3080
3180
3120
3080
3160
2960
3240
-
3080
3260
3010
3230
3150
3030
2670
3280
2810
3040
3120
-
-
2370
2380
2300
2230
2740
2800
2760
2460
2190
2120
2280
2380
2290
2290
1960
2260
-
2120
2040
2270
2270
2150
2100
2060
2010
1910
2010
2070
.
-
1520
2230
2470
2440
2680
2680
2900
3290
37,400
37,600
38,200
38,700
37,900
37,100
35,100
-
-
-
-
36,800
38,100
37,900
38,600
36,900
-
36,900
36,600
-
-
-
-
42,200
35,300
35,400
39,500
36,300
33,000
-
Boiler was shut down.
23
-------
TABLE 3-4. DAILY AVERAGE PROCESS PARAMETERS AND EMISSION RESULTS
Date
2/27
2/28
3/1
3/2
3/3
3/4
3/5
3/6
3/7
3/8
3/9
3/10
3/11
3/12
3/13
3/14
3/15
3/16
3/17
3/18
3/19
3/20
3/21
3/22
3/23
3/24
3/25
3/26
3/27
3/28
Load,
106 Btu/hr
35.6
36.4
37.4
36.5
36.0
37.1
34.0
36.1
-
36.1
35.9
35.2
37.9
37.5
36.2
31.4
36.7
32.7
35.2
36.7
-
-
27.6
27.8
27.7
26.5
32.2
33.3
32.3
28.1
Calcium- to-
Sulfur Ratio
3.9
3.4
3.7
4.1
3.9
3.8
3.3
3.6
-
2.7
3.4
3.9
3.6
3.7
3.6
4.4
3.3
4.0
3.6
3.3
-
-
3.7
5.1
5.8
5.6
5.1
4.8
6.0
7.7
S02 Reduction,
Percent
94.2
93.1
93.3
95.2
94.1
95.9
93.1
91.3
88.3
73.2
89.0
93.4
97.1
96.6
95.4
-
94.7
92.8
93.8
95.0
-
-
-
94.1
95.2
95.4
93.8
89.6
95.8
-
Outlet S02
Emissions,
lb/106 Btu
0.58
0.71
0.66
0.48
0.61
0.42
0.75
0.88
3.53
2.03
1.11
0.66
0.31
0.33
0.47
-
0.53
0.68
0.61
0.53
-
-
-
0.57
0.47
0.48
0.62
1.08
0.40
-
Outlet NO
Emissions,
lb/106 Btu
0.58
0.63
0.63
0.65
0.65
0.67
0.56
0.62
0.66
0.67
0.74
0.71
0.65
0.65
0.62
-
0.59
0.59
0.61
0.62
-
-
.
0.69
0.69
0.69
0.65
0.63
0.68
-
24
-------
SECTION 4.0
SAMPLING LOCATION
Continuous emissions monitoring and EPA Reference Methods 3, 6, 7E,
and 10 were conducted at the stack location. A schematic of the stack
location is shown in Figure 4-1.
Figure 4-2 shows the placement of traverse points for stratification
testing. The stack is 42 inches in diameter. Traverse points were positioned
according to EPA Reference Method 1.
25
-------
to
BAG HOUSE
r
A
T
7'
SAMPLING* PLATFORM CFB SUMMERSlDE
DNNBTABf
I CEM PORT
2 RM PORT
Figure 4-1. Schematic of Sampling Locations.
-------
1 -
Reference
Method
Probes
2 - GEM Probe
Point
1
2
3
4
5
6
7
8
9
10
11
12
Distance from wall (inches)
41.1
39.2
37.0
34.6
31.5
27.0
15.0
10.5
7.4
0.9
Figure 4-2. Stack sample traverse point locations,
27
-------
SECTION 5.0
SAMPLING AND ANALYTICAL PROCEDURES
5.1 OVERVIEW
Alliance conducted the continuous monitoring program over a period of
approximately 6 weeks (see Figure 5-1). The first phase of the study required
certification of the monitoring instruments according to EPA Performance
Specifications 2, 3, and 4. During the initial performance specification
tests, zero and calibration drift, instrument response time, and relative
accuracy were determined. EPA Reference Methods 3: (02/C02), 6: (802),
7E: (NOX), and 10: (CO) were used to determine the accuracy of the monitors
in use. Upon successful completion of the relative accuracy tests, the
monitoring program commenced. Prior to completion of the monitoring period,
the analyzers were again subjected to relative accuracy tests.
5.2 MEASUREMENT OF FLUE GAS EMISSIONS
Flue gas emissions were measured on a dry basis with an Energy Mines and
Resources/Canada Continuous Emissions Monitoring System (CEMS). The CEMS,
based upon extractive gas sampling principles, consists of three subsystems:
sample acquisition/conditioning, sample analysis, and data acquisition (see
Figure 5-2).
The Sample Acquisition/Conditioning unit was designed to deliver a sample
stream, representative of the stack gas stream, to the sample analysis
subsystem. Since the stream must be clean and dry for proper analyzer
operation, a filter and condenser were used for particulate and moisture
removal. The sample line, which carries the gas from the stack location to
the conditioning system, was heat traced to eliminate in-line condensation.
Sample was drawn into the system by a vacuum pump, which was also heated to
eliminate condensation.
Sample analyses for sulfur dioxide, nitrogen oxides, carbon monoxide,
oxygen, and carbon dioxide were achieved with the analyzers specified in
Table 5-1. In each case, accurate interpretation of analyzer response
required systematic calibration of the instrument against gases of known
concentrations (see Table 5-2). A calibration curve was determined from a
linear regression of known gas concentrations versus instrument response. The
equation used to convert instrument signal to concentration units is as
follows:
concentration = m(response) + b
28
-------
sD
00
I
I
CM
CO
00
I
CM
Z
i
vD
00
-*
CS
I
CM
z
o
•o
CO
CO
CO
Z
O
vO
00
I
CO
Z
§
vO
00
I
CO
Z
o
vO
00
I
-3-
CM
I
CO
oo
I
CO
CO
Z
i
vO
00
I
Z
o
S3
VO
TRAVEL AND
SET UP
GEM CALIBRATION
DRIFT
DETERMINATION
CEM
RELATIVE
ACCURACY
DETERMINATION
30 DAY CEM
MONITORING
BREAKDOWN AND
TRAVEL
I 02 CO N0>| S02
°
1
J
30 DAY
i
MONITORI
W PROG
?AM 1
cn
1 1
I I INDICATES ACTUAL PROGRAM TIMETABLE
I 1 PROGRAM AS SCHEDULED
Figure 5-1. GEMS program schedule,
-------
CALIBRATION
HEATED SAMPLE
BOX WITH F1LTERS
HEATED BOX
CO
o
CER PROBE VALVE y. HEATED
N? ZERO
2.97%0
3 03°'CO L0 MIX
2
67.9ppmCO
461ppmSO
11.76%02
99T°/rin __.._..
463ppmCO
872ppmSO
20.0%00
2 HI MIX
2
999ppmCO
4850ppmS02
LO NO
1 54pptnNO
„. Mn MID NOx
1 HFAT TRAPFft HF.AT |_J k .1 , _J_ _ ._
SAMPLE LINE IRA^E U pmp J | M
-------
TABLE 5-1. ANALYZER SPECIFICATIONS
Pollutant/diluent
Sulfur dioxide (802)
Oxides of nitrogen (NOX)
Carbon monoxide (CO)
Carbon dioxide (C02)
Oxygen (02)
Manufacturer
Western Research
TECO
Beckman
Beckman
Beckman
Model No.
721
10AR
865
864
755
Operating principle
NDUV
Chemi luminescence
NDIR
NDIR
Paramagnetic wind
TABLE 5-2. CALIBRATION GAS CONCENTRATIONS
S02
NOX
CO
02
C02
Zero
N2
N2
N2
N2
N2
Low
461 ppm
154 ppm
67.9 ppm
2.97%
3.03%
Mid
872 ppm
315 ppm
463 ppm
11.76%
9.83%
High
4,850 ppm
999 ppm
20.0%
19.0%
31
-------
where: m = slope of calibration curve
response = instrument signal (volts)
b = y - intercept of calibration curve
All calibration gases utilized during this program were obtained from
AIRCO with an analytical certification of concentrations (Protocol No. 1; see
Appendix J). The Data Acquisition System consisted of a Kaye Digistrip III
Data Logger programmed to accept daily calibration data, and to perform the
previously mentioned concentration calculation. Average concentration values
(derived from the concentration calculation) for each analyzer were logged at
15-minute intervals, on the quarter-hour.
Stratification Check
Before beginning relative accuracy tests, a stratification test was
conducted at the sampling location to verify a homogeneous gas stream. The
stratification test consisted of measuring the velocity, pollutant, and
diluent concentration at each traverse point (see Figure 4-2). Data collected
at each traverse point was utilized to define the spatial variation of the
flue gas stream. The results of the stratification test indicated a
homogeneous stack gas stream and Reference Method sample probes were placed in
close proximity to the GEMS probe.
Response Time
Monitor response time was used as a Quality Control check and is reported
as the slower of the average of three sets of up-scale and down-scale
determinations. The up-scale determination is the time it takes the monitor
to respond from a zero calibration gas reading to a stable stack effluent
reading. Conversely, the down-scale determination is the time it takes the
monitor to respond from a high-level calibration gas concentration to a stable
stack effluent reading. The mean up-scale and down-scale response times are
determined, with the slower value reported. Reporting the slower value
identifies an important operational limitation of the monitor. Results of the
monitor response time check are provided in Table 5-3.
Instrument Drift
Instrument drift was determined during the first and second week of the
test program according to procedures outlined in Performance Specification
Tests 2, 3, and 4. Drift determinations were based upon the routine daily
calibration checks conducted at 24-hour intervals. Instrument drift data are
included in Appendix G.
Calibration Gas Traceability
All calibration gases used during the test program were obtained from
AIRCO Industrial Gases with an analytical certification of concentration (NBS
Traceable Protocol No. 1). Appendix J provides copies of the Calibration Gas
Analytical Reports.
32
-------
TABLE 5-3. GEMS SYSTEM — MONITOR RESPONSE TIME RESULTS (SECONDS)
Upscale 1
2
3
X
Downscale 1
o
f-
3
X
02
88,
89
83
87
89
93
85
89
C02
66
63
64
64
65
66
65,
65
CO
64
64
65
64
63
68.
68 ,
68
NOX
63
62
59
61
67
63
69
66
S02
74
78
70
74
109
101
105
105
Response Time
89
65
68
66
105
33
-------
Coal and Limestone Sampling and Analysis
Coal and limestone grab samples were collected and composited (over
8-hour periods: 0000-0800; 0800-1600; 1600-2400, continually throughout the
GEMS program. An individual day's sample consisted of three 8-hour composites
(total approximate weight - 96 Ibs) for both the fuel and sorbent. Prior to
bagging, the limestone samples were composited into a single daily sample.
These samples were bagged and labelled in the field for shipment to Alliance.
Upon arrival at Alliance, the samples were again logged and tagged for
shipment to Spotts, Stevens and McCoy (SSM), Reading, PA. SSM was responsible
for the analyses summarized in Table 5-4, with the exception of the reactivity
test which was conducted by Westinghouse Research and Development Center.
Appendices B and C contain the data sheets as received from SSM and
Westinghouse.
In addition to the 8-hour composite coal sampling, Alliance personnel
analyzed hourly grab samples of coal for moisture and sulfur content at the
plant site. This information provided process engineers with the data
necessary to determine sulfur emissions on a day-to-day basis. This
information was used in conjunction with emissions data to adjust coal and
limestone feed rates and, concurrently, the calcium to sulfur ratio.
Coal samples were analyzed on a hourly basis for sulfur content and
heating value from February 27 through March 5. For the remainder of the
program, 3-hour intervals were composited such that eight 3-hour composites
were analyzed per day. The coal analyses involved weighing, drying for
24 hours, and again weighing each sample to determine moisture content. The
samples were then pulverized and analyzed in triplicate for sulfur content
with a LEGO MODEL 10 sulfur analyzer. The hourly coal sample and analysis
logs are provided in Appendix E.
Additional samples, obtained periodically throughout the program, were
taken to further aid in characterizing FBC operation. They included: Bed
Material, Bed Drain Material, Multi-cyclone Ash for Reinjection, Baghouse Ash,
and Ash Silo Samples. The analyses of these materials are provided in
Appendix D.
34
-------
TABLE 5-4. ANALYTICAL PROCEDURES
Material
Parameter
ASTM Method
Other Method
Coal
Limestone
Ash Bed,
Multiclone
Baghouse
Silo
Sieve Analysis
Preparation for Analysis
Ultimate Analysis
Carbon and Hydrogen
Sulfur
Nitrogen
Ash
Moisture
Oxygen
Proximate Analysis
Moisture
Ash
Volatile Matter
Fixed Carbon
Ash Analysis
Calcium
Magnesium
Sodium
Calcium
Sulfur
Sieve Analysis
Reactivity
Ultimate Analysis
Carbon and Hydrogen
Sulfur
Nitrogen
Ash
Moisture
Oxygen
Carbonate Carbon
Sieve
D410
D2013
D3176
D3178
D4239B
D3179
D3173
D3173
Calculated by Difference
D3172
D3173
D3174
D3175
Calculated by Difference
D3682
C25, E508
D423B
D410
Westinghouse Research &
Development Thermogravimetrie
Analysis
D3176
D3178
D4239B
D3179
D3174
D3173
Calculated by Difference
D1756
D410
-------
SECTION 6.0
PROGRAM QUALITY ASSURANCE
6.1 INTRODUCTION
A detailed Test and Quality Assurance Plan was written, submitted, and
approved for use under this project. This document was intended to serve as a
guide for use during the field, laboratory, and data handling segments of the
project.
6.2 PRECISION, ACCURACY AND COMPLETENESS
Alliance followed recommended procedures outlined in 40 CFR Part 60,
Subpart Da and Appendices A, B, and F (proposed) to assess the precision,
accuracy, and completeness of the continuous emissions monitors used during
this program. These checks include determination of relative accuracy,
precision estimates, tests for instrument drift and response time, and
completeness of the data base. Table 6-1 provides Performance Specifications
for Relative Accuracy (R.A.) and Calibration Drift as delineated in 40 CFR 60.
Relative Accuracy Tests
Relative accuracy (R.A.) tests were performed to assess the precision and
accuracy of the continuous monitors. Reference Methods 3: (02, C02),
6: (S02), 7E: (NOX), and 10: (CO), outlined in 40 CFR Part 60 Appendix A,
were performed to generate information for comparison with CEM data.
Reference Method field data sheets and R.A. calculations are presented in
Appendix F. All the analyzers surpassed the established criteria during the
initial relative accuracy evaluation conducted prior to beginning the 30-day
CEM period (February 25-27, 1986). A second abbreviated relative accuracy
test (three runs) was conducted during the final week of testing. The
analyzers did not fare as well during the second set of tests, as the CO and
SC>2 analyzers failed. Failure of the analyzers during the second set of
tests can be traced to the extreme ambient temperature fluctuations that
occurred as local weather patterns changed.
The analyzers were located within 10 feet of the roof of the boiler
building in an uncontrolled temperature environment. On the day of the final
relative accuracy and cylinder audit tests the outside ambient temperature was
between 60° and 70°F, with the temperature surrounding the analyzers within
the building between 100° and 110°F. Prior to this, the outside ambient
temperatures were much lower with the temperature surrounding the analyzers
between 70° to 80°F. Once a relative accuracy test series commences on a
36
-------
TABLE 6-1. PERFORMANCE SPECIFICATION REQUIREMENTS
Relative
accuracy
(RA)
Calibration drift
S02
NOX
C02*
02*
CO
+20%
+20%
+1% (co2)
+0.5% (02)
+ 10%
+2.5% for 6 of 7 test days
+2.5% for 6 of 7 test days
+_0.5% for 6 of 7 test days
+0.5% for 6 of 7 test days
+5% for 6 of 7 test days
*No criteria are designated for individual diluetnt analyzers,
GCA uses these criteria as a basis for judging proper
operation.
37
-------
TABLE 6-2. RESULTS OF CEM RELATIVE ACCURACY TESTS
Specification
(percent)
--
02 0.5
C02 1.0
CO 10.0
NOX 20.0
S02 20.0
Initial set
(percent)
9 runs
0.45 02 c
0.39 C02 c
5.4a
7.8a
10. 4a
Final setb
(percent)
3 runs
0.30 02 d
0.07 C02 d
13. 5b
8.1b
20. 4b
RA
llU
cc x 100
RM
RA = Id | x 100
RM
AA = d + cc
AA =| d.
Where: RA = Relative accuracy.
AA = Absolute accuracy.
d = Algebraic average of differences between CEMS
value and Reference Method value.
cc = 95% confidence coefficient.
RM = Reference Method value.
38
-------
given day, the analyzers are not recalibrated until the next calendar day.
Therefore, any drift due to temperature fluctuation, after initial calibration
early in the morning, is not corrected as the temperature fluctuates. In
reality, the location of the monitor system was not ideal and it should have
been in a controlled temperature environment.
Both the CO "and SC>2 analyzers were biased high relative to Reference
Methods values. The higher values for SC>2 would, therefore, affect the
SC>2 capture efficiency calculations; i.e., showing a lower percent capture
than actual. A summary of relative accuracy tests results is provided in
Table 6-2.
Calibration Drift
Zero and calibration drift tests were used to indicate an individual
analyzer's ability to remain accurately calibrated over an extended period of
time. Performance Specifications (Table 6-1) set a level for maximum
allowable drift per day. This criteria must be met for seven consecutive days
as part of the CEM certification routine. The Alliance Irield team used the
data from daily calibrations to determine individual analyzer drift. The
62, CO, NOX, and SO2 analyzers all eventually passed the tests; the
C02 did not. Appendix G contains zero and calibration drift field data
sheets.
Completeness
The completeness of the emissions monitoring data was determined based
upon the 60 minute data file (Appendix A). Excluding 71 hours of FBC unit
downtime (unit repair), the CEMS completeness percentage for the test program
was 97.5 percent. The analyzers were down only 2.5 percent of the time for
maintenance and repair.
Results of EPA Audit
Alliance conducted cylinder gas audits on the S02 and NOX analyzers.
The audit gases were introduced to the CEM system, per normal calibration
routine at the probe location, thereby eliminating the potential for system
bias. The results of the audit are presented in Table 6-3.
Due to relatively high temperatures surrounding the analyzers on the day
of the audit, the S02 analyzer had drifted positively from the original
settings, which "had been set earlier in the day, prior to conducting the
audit. Once a calibration has been set at the beginning of a 24-hour period,
the analyzers are not recalibrated until the start of the next 24-hour
period. The differences between the cylinder value and the analyzed values
were 18.3 ppm and 28.5 ppm for the 69.4 and 296.1 ppm cylinders,
respectively. The higher percent difference for the low cylinder audit is
due to its value, as the differences are apparently equal. As the S02
monitor was reading in the 200 to 350 ppm range during most of the test
period, the cylinder gas audit showed that the concentrations presented by the
monitor are fairly accurate.
39
-------
TABLE 6-3. RESULTS OF EPA CYLINDER GAS AUDIT
Concentration
A B
Pollutant Cylinder value Analyzed value
B-A
x 100
NOX
SO 2
70.4
402.6
69.4
296.1
69.8
397.6
87.7
324.6
-0.85
-1.24
26.37
8.78
Two EPA Audit Ampules for S0£ (Reference Method 6) were analyzed along
with the initial Reference Method Runs. Table 6-4 provides the results of the
Method 6 Audit.
TABLE 6-4. RESULTS OF EPA REFERENCE METHOD 6 AUDIT
Concentration
EPA audit
sample #
2478
3666
A
Actual
concentration
297.4 mg/dscm
282.2 mg/dscm
B
Reported value
298.2
294.4
B-A
A
0.
4.
x 100
2%
3%
Coal Analyses
Four U.S. EPA Quality Assurance coal samples were provided by the
Emissions Measurement Branch for auditing the analyses conducted by Spotts,
Stevens and McCoy. Each sample was analyzed twice on each of two different
days. The results of these analyses are presented in Appendix B-4.
An internal Quality Control sample was also analyzed by Spotts,
Stevens and McCoy on four separate days during the analytical program. These
results are presented in Appendix B-4.
40
------- |