EPA-600/R-92-115
June 1992
DEMONSTRATION OF SORBENT INJECTION TECHNOLOGY
ON A WALL-FIRED UTILITY BOILER
(EDGEWATER LIMB DEMONSTRATION)
by
Paul S. Nolan
Thomas W. Becker
Peter 0. Rodemeyer
Edward J. Prodesky
The Babcock & Wilcox Company
Barberton, Ohio 44203
EPA Contract 68-02-4000
Project Officer
David G. Lachapelle
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
Prepared for
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
-------
TECHNICAL REPORT DATA
(Please ren,l Instructions on the reverse bejore complet1
i. ncPoru no. 2.
EPA-600 / LI-9 2-115
3 PB92-
-201136
4. 1 I I LE AND yum ITLE
Demonstration of Sorbent Injection Technology on a
Wall-fired Utility Boiler (Edgewater .LIMB
Demonstration)
5. nEf'OHT DA T B
June 1992
6. PERPORMING ORGANIZATION CODE
7. AUTMOR(S)
P. S. Nolan, T. W. Becker, P. O. Rodemeyer, and
E.J. Prodesky
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The Babcock & Wilcox Company
Barberton, Ohio 44203
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-4000
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 9/84-6/89
14. SPONSORING AGENCY CODE
EPA/600/13
is.supplemeniauy notls AEERL project officer is David G. Lachapelle, Mail Drop 4, 919/
541-3444.
its. aoshiact 'I'lle report gives results of the full-scale demonstration of Limestone In-
jection Multistage Burner (LIMB) technology on the coal-fired, 105 MW, Unit 4 boi-
ler at Ohio Edison's Edgewater Station. Developed as a technology aimed at moder-
ate levels of sulfur dioxide (S02) and nitrogen oxides (NOx) emissions control for
relatively low-cost retrofit applications on older plants, LIMB operation at a calciun
to sulfur (Ca/S) molar ratio of 2.0 is shown to be capable of achieving about 55 to
72% S02 removal. Removal depends on the sorbent in use and the degree of flue gas
liumidification employed. Sorbents tested included a commercial calcitic hydrated
lime, both with and without a small amount of calcium lignosulfonate, a material
added to improve reactivity. Results are given for liumidification of the flue gas both
to an 11 C approach to the saturation temperature (where an increase of 10% absolute
in S02 removal efficiency is obtained), and for minimal humidification. (where water-
is added solely to maintain adequate electrostatic precipitator-~ESP--performance).
The performance of the XCL low-NOx burners, with an overall emission of 206 ng/
J is also presented. The report also discusses the impact of LIMB technology on
boiler and plant operations. The effects are related primarily to the increased quan-
tity of particulate that must flow through the boiler, ESP, and other ash equipment.
17. KEY WORDS AND DOCUMENT ANALYSIS
•l. DESCRIPTORS
b. IDENTIF 1E RS/OPEN ENDED TERMS
c. cosati Field/Group
Pollution Nitrogen Oxides
Utilities Emission
Boilers Humidification
Sorbents Electrostatic Pre-
Lime stone cipitators
Sulfur Dioxide Particles
Pollution Control
Stationary Sources
Utility Boilers
Limestone Injection
Multistage Burners
Particulate
13 B
14 G
13 A 13 H
11G
08G 131
07 B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
210
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
£ PA Form 2220-1 (9-73)
i
-------
ABSTRACT
This report presents the results of the full-scale demonstration of Limestone Injection
Multistage Burner (LIMB) technology conducted on the coal-fired, 105 MWe, Unit 4 boiler at Ohio
Edison's Edgewater Station. Developed as a technology aimed at moderate levels of sulfur dioxide
(S02) and nitrogen oxides (NOx) emissions control for relatively low-cost retrofit applications on
older plants, LIMB operation at a calcium to sulfur (Ca/S) molar ratio of 2.0 is shown to be capable
of achieving approximately 55 to 72% S02 removal. The removal depends upon the specific
sorbent in use and the degree of flue gas humidification employed. Sorbents tested included a
commercial calcific hydrated lime, both with and without a small amount of calcium lignosulfonate,
a material added to improve reactivity. Results are presented for humidification of the flue gas
both to a 20°F approach to the saturation temperature (where an increase of 10% [absolute] in
S02 removal efficiency is obtained), and for minimal humidification (where water is added solely to
maintain adequate electrostatic precipitator [ESP] performance). The performance of the
DRB-XCL™ low-NOx burners, with an average emission of 0.48 lb/106 Btu over all loads, is also
presented.
The report also discusses the impact of LIMB technology on boiler and plant operations.
The effects are related primarily to the increased quantity of particulate matter that must flow
through the boiler, ESP, and ash handling equipment. The need for effective sootblowing is seen
as the single most important requirement when considering application of the technology: without
it, the increased particulate loading can decrease heat transfer efficiency between sootblowing
cycles. Humidification can also be necessary, as sorbent injection tends to produce an ash whose
electrical resistivity may render it more difficult to collect. The other effects discussed in some
detail result from the quicklime component of the ash and the precautions that must be taken to
avoid the potential for deposits at wet/dry interfaces and for high pH conditions associated with
ash handling.
Finally, the application and economics of the technology are discussed in terms of a
hypothetical LIMB system on a 150 MWe boiler. This forms the basis for a comparison of the
capital and operating costs of the technology with those of a conventional limestone wet scrubbing
system.
ii
-------
CONTENTS
Abstract ii
Figures iv
Tables vii
Acknowledgments ix
1. Executive summary 1
2. Introduction 4
3. Project objectives and scope 7
4. Design application to Edgewater 8
5. Detailed system description 19
6. LIMB operational issues 41
7. LIMB performance evaluation 47
8. LIMB commercial design and economics 113
9. Summary and conclusions 141
References 144
Appendices
A. Metric conversion table 147
B. Quality assurance/quality control 148
C. DRB-XCL™ burner characterization test report 176
iii
-------
FIGURES
Number Page
1 The LIMB process at Edge water 10
2 Sorbent injection locations 14
3 Sorbent storage and transfer system 20
4 Sorbent feed system 22
5 Injection nozzle and furnace wall port 24
6 Original circular register burner 26
7 Retrofit low NOx DRB-XCL™ burner 27
8 Bypass system - plant orientation and schematic 29
9 Humidification chamber detail 31
10 Cross-section of atomizer with nacelle 34
11 Airfoil lance and atomizer array 35
12 Ash removal and ash pond pH control 37
13 Furnace gas flow pattern initially assumed 62
14 Differential mass concentration plot for test BL03 at the ESP outlet 67
15 Differential mass concentration plot for test BL05 at the ESP outlet 68
16 Differential mass concentration plot for test BL06 at the ESP outlet . 69
17 Circular and DRB-XCL™ burner NOx emissions 74
18 S02 removal vs. Ca/S stoichiometry for commercial lime with minimal humidification . . 76
19 S02 removal vs. Ca/S stoichiometry for commercial lime with 20 °F approach to
saturation 77
20 S02 removal vs. Ca/S stoichiometry for lignosulfonated lime with minimal
humidification 78
21 S02 removal vs. Ca/S stoichiometry for lignosulfonated lime with 20 °F approach to
saturation 79
22 LIMB operation: January 8 - June 9, 1989 85
23 LIMB operational time 86
24 S02 and 02 concentrations during Ca/S variation on June 7, 1989 88
25 S02 removal comparison at less than optimal conditions 90
26 Effect of humidification on S02 removal 93
iv
-------
FIGURES (continued)
Number Page
27 DRB-XCL™ burner NOx emissions vs. load, February 1989 94
28 10 min Average DRB-XCL™ burner N0X emissions, January 8 - June 9, 1989 95
29 Hourly baseline circular burner NOx Emissions, May 27 - June 4, 1986 97
30 Hourly average DRB-XCL1" burner NOx emissions, January 8 - February 28, 1989 .... 98
31 Daily average DRB-XCL™ burner NOx emissions, January 8 - April 20, 1989 99
32 Hourly average DRB-XCL™ burner NOx emissions, January 8 - 25, 1989 100
33 24 hr Rolling average DRB-XCL™ burner NOx emissions, January 8 - 25, 1989 101
34 30 day Rolling average DRB-XCL™ burner NOx emissions, February 28 - April 20, 1989 102
35 In situ resistivity measurements without humidification, September 1987 104
36 Typical sorbent injection effects on ESP electrical characteristics, September 1987 ... 105
37 Typical LIMB ash particle size distributions at ESP inlet and outlet 107
38 ESP electrical conditions during LIMB operation with humidification to 275°F 108
39 ESP electrical conditions during LIMB operation with humidification to 165°F 109
40 Resistivity of the LIMB ash 110
41 Fractional collection efficiencies of the ESP during LIMB operation 111
42 Capital cost sensitivity to unit size and coal sulfur (base case conditions) 131
43 S02 removal cost sensitivity to unit size and coal sulfur (base case conditions) 133
44 Total capital requirement and S02 removal cost sensitivities to unit size (base case
conditions) 136
45 Total capital requirement and S02 removal cost sensitivities to coal sulfur (base case
conditions) 137
46 Total capital requirement and S02 removal cost sensitivities to capacity factor (base
case conditions) 138
47 Total capital requirement and S02 removal cost sensitivities to removal efficiency (base
case conditions) 139
B-1 CEMS sample acquisition system 154
C-1 NOx emissions as a function of load 186
C-2 NOx emissions as a function of economizer outlet 02 187
C-3 NOx emissions as a function of air heater inlet 02 188
C-4 Circular and DRB-XCL™ burner NOx emissions 190
v
-------
FIGURES (continued)
Number Page
C-5 Carbon in ash as a function of load 191
C-6 Carbon in ash as a function of economizer outlet 02 192
C-7 Carbon in ash as a function of air heater inlet 02 193
C-8 Carbon in ash burner comparison 195
C-9 DRB-XCL™ burner pressure loss 197
C-10 CO emissions as a function of air heater inlet 02 198
vi
-------
TABLES
Number Page
1 Summary of high sulfur coal analyses - truck samples 53
2 Summary of high sulfur coal analyses - bunker samples 54
3 Summary of high sulfur coal analyses - pulverized (burner pipe) samples 55
4 Summary of lime analyses 56
5 Circular burner baseline and characterization tests summary - S02 and N0X emissions . 58
6 DRB-XCL™ burner characterization tests summary - S02 and NOx emissions 59
7 Circular burner baseline tests summary - other parameters 60
8 Summary of circular burner baseline tests ash analyses 70
9 Results with commercial calcitic hydrated lime at minimal humidification 80
10 Results with commercial calcitic hydrated lime at close approach to saturation 81
11 Results with lignosulfonated calcitic hydrated lime at minimal humidification 82
12 Results with lignosulfonated calcitic hydrated lime at close approach to saturation .... 84
13 .Results with commercial calcitic hydrated lime at less than optimal conditions 91
14 Average DRB-XCL™ burner NOx emissions 103
15 Effect of LIMB on boiler operational parameters 125
16 EPA economic premises for LIMB and FGD cost analyses 126
17 EPA technical premises for LIMB and FGD cost analyses 127
18 Summary of capital cost for the base case (B&W and EPA comparison) 128
19 Summary of operation and maintenance costs for the base case (B&W and EPA
comparison) 129
20 Capital sensitivity ($/Kw) to unit size and coal sulfur (B&W and EPA comparison) .... 130
21 Sensitivity of cost per ton of S02 removed to unit size and coal sulfur (B&W and EPA
comparison) 132
22 EPA comparison of LIMB and limestone forced oxidation FGD economics 135
B-1 Test history 149
B-2 Summary of estimated precision, accuracy, and completeness objectives 150
B-3 Summary of manual flue gas sampling and analysis 156
B-4 Summary of process sample collection/analysis 156
B-5 Triple beam balance audit 157
B-6 Dry gas meter performance audit results 158
B-7 Orsat analyzer performance audit results 158
vii
-------
TABLES (continued)
Number Page
B-8 Summary of EPA Method 5 flue gas parameters and isokinetic sampling rate 159
B-9 Results of performance audit - proximate analyses of coal 160
B-10 Results of performance audit - ultimate analyses of coal 160
B-11 Results of performance audit - mineral analyses of coal ash 161
B-12 Fly ash audit results - ICAPES metals analysis 161
B-13 Summary of the CEMS certification and baseline drift check results 163
B-14 Summary of the CEMS quality control sample results for baseline testing 164
B-15 Results of the CEMS performance audits 165
B-16 Summary of CEMS analytical quality control checks, frequencies, acceptance criteria,
and corrective actions 166
B-17 Summary of analytical operations quality control checks, frequencies, acceptance
criteria, and corrective actions 167
B-18 Analyzer performance during the demonstration period 168
B-19 Summary of internal performance audits showing the range (minimum/maximum) of
percent relative error values 171
B-20 Summary of performance audit results for manual sampling methods 171
B-21 Summary of CEMS relative accuracy results 172
B-22 Summary of CEMS calibration gas data 173
B-23 Summary of the CEMS drift check results 173
C-1 Edgewater Unit 4 baseline test results 177
C-2 Original randomized test sequence and conditions 179
C-3 Actual test sequence and conditions 180
C-4 Coal analyses 181
C-5 DRB-XCL™ burner characterization test data summary 182
C-6 Oxygen and excess air analysis 184
viii
-------
ACKNOWLEDGMENTS
The authors wish to thank the many individuals who contributed to the success of the
LIMB Demonstration Project. We especially wish to acknowledge the efforts of Mr. Thomas B.
Hurst, the first Project Manager for B&W, and Mr. Robert V. Hendriks, the first Project Officer for
the U.S. Environmental Protection Agency. We also wish to thank the personnel at Ohio Edison's
Edgewater Plant for their work and understanding throughout the operational phase of the
demonstration. For others too numerous to mention individually, we thank you for your advice,
support, and cooperation through mention of your organizations:
U. S. Environmental Protection Agency
Ohio Edison Company
Ohio Coal Development Office
Consolidation Coal Company
Radian Corporation
Stone & Webster Engineering Corporation
Energy and Environmental Research Corporation
Southern Research Institute
U. S. Department of Energy
ix
-------
SECTION 1
EXECUTIVE SUMMARY
The results of the Limestone Injection Multistage Burner (LIMB) Demonstration Project on a
full-scale, coal-fired utility boiler show that the technology has met or surpassed program goals
with respect to sulfur dioxide (S02) and nitrogen oxides (NOx) emissions control. Moreover, in spite
of early difficulty in controlling particulate emissions with the electrostatic precipitator (ESP), the
design and installation of a full-scale humidification system appears to have overcome the adverse
aspects of the increased quantity of fine, high resistivity ash. Specifically, the use of a
lignosulfonate-doped hydrated calcitic lime, produced in bulk by a commercial supplier, together
with operation of the humidifier at a 20 °F* approach to saturation resulted in S02 removal
efficiencies of up to 72% at a calcium to sulfur (Ca/S) molar ratio of 2.0, far in excess of the 50%
program goal. Even without the lignosulfonate additive, up to 65% capture was possible under the
same humidified conditions. Removal efficiencies of about 62% and 55% were obtained with only
modest humidification to approximately 275 °F for the modified and commercial sorbents,
respectively. The S02 data have also been reduced and presented in a form that characterizes the
removal over a range of stoichiometric ratios.
At the same time, an overall average NOx emission level of 0.48 lb/106 Btu was achieved
over months of operation with the DRB-XCL™ low NOx burners installed to meet a goal of 0.5
lb/10® Btu or less. Averages of the weighted 24-hour and 30-day rolling averages of 0.47 and
0.49 lb/10® Btu were calculated for the periods of January 8 to 26 and February 21 to April 22,
1989, respectively. The individual burner characterization tests performed before sorbent injection
commenced indicated some increase in the unburned carbon content of the ash as a result of the
longer flame length of the low-NOx burner. Ash samples, corrected for dilution by unreacted
* For those more familiar with metric units, see the conversion table in Appendix A.
/¦
-------
sorbent and reaction products, appeared to show that the effect was minor.
Particulate emission control was effectively restored by modest levels of humidification as
indicated by opacity measurements at the ESP outlet. Particulate mass emission levels of 0.01 to
0.02 lb/106 Btu were found to be far below the 0.1 lb/10® Btu State of Ohio operating limit, al-
though the tests had to be conducted at conditions representing approximately 75% of the full
boiler load. The reason for this was that long term loss of refractory material in the vicinity of the
nose arch exposed support beams to unacceptably high temperatures from flue gas leakage.
Unfortunately, load limitations that were to last until the scheduled maintenance outage had to be
imposed at the very time that the particulate tests were to be conducted. Nevertheless, the
conditions were such that the humidifier was operating at the full design flow since repairs to
various sources of air in-leakage in the boiler had not been fully successful.
The overall impact that LIMB technology had on boiler and plant operations at Edgewater is
particularly notable in a few areas, all of which are considered manageable when viewed from the
perspective of further commercialization. The single greatest impact comes simply from the
increased particulate loading through the boiler, ESP, and ash handling system. Unless restored by
effective sootblowing, heat transfer efficiency in the boiler decreases in proportion to the amount
of sorbent injected. At higher Ca/S ratios the limitations of the compressed air sootblowing system
at Edgewater soon pointed toward the need for future commercial systems to address sootblowing
capability more thoroughly than had originally been anticipated for the demonstration project. In
comparison with these effects, the impact of increased gas flow from air used to disperse the
sorbent and to atomize water in the humidifier is thought to be quite low, especially if only
moderate humidification to perhaps 275 °F is required to restore ESP efficiency. While the data
indicate an overall energy penalty of about 1.5% in boiler efficiency for the combined impacts, it is
strongly suspected that much of this can be recovered with more effective sootblowing.
Again because of the increased quantity of material, LIMB technology makes additional
demands upon the ash handling system that must be taken into account in the application for
future commercial systems. Beyond this, the chemical composition of the LIMB ash, and the
quicklime component in particular, should be reviewed carefully in light of possible deposit
formation at wet/dry interfaces where the pozzolanic properties of the ash can be important.
Likewise the lime component of the ash is important with regard to potentially high pH conditions
in any wetting operation that might be used as part of the ash disposal process.
2
-------
At Edgewater the existing ash handling system provided the capability of both wet and dry
unloading from the ash silo. Dry unloading was normally used when ash was sold for use as a
cement additive, and would be the likely choice for similar use of LIMB ash as a byproduct. More
commonly, wet unloading through a pug mill fitted with water sprays was used when the ash was
trucked to a landfill. The water was added to help control fugitive dust emissions. This wet
system, upgraded most significantly in its water feed capacity, was used for the treatment of the
LIMB ash prior to disposal in the landfill.
It was known when the project began that the reaction of water with quicklime in the LIMB
ash was highly exothermic and could result in significant steam generation, especially when low
levels of humidification were employed. Although a large blower was provided to exhaust the
steam generated, the quantity experienced at Edgewater still made it difficult for operators to see
the level of the wetted ash as discharged into the truck. The steaming subsides within about
15 min and presents no extraordinary handling problem thereafter. Though it was not practical at
Edgewater, an ideal system would have included wetting as part of a radial stacking operation
where the ash could be allowed to sit for at least a short period of time. Alternatively, a more
extensive positive exhaust system or perhaps a gravimetric loading device might be employed in
situations similar to those at Edgewater.
Analysis of LIMB technology in terms of a commercial application reveals it to be the
economic choice for S02 and N0X emission abatement in unit sizes under approximately 300 MWe
when compared to a wet limestone flue gas desulfurization (FGD) system. If the S02 and NOx
control levels of a LIMB system meet or exceed the required levels in a utility's overall strategy for
compliance with Clean Air Act legislation, a LIMB system offers the following advantages over a
wet FGD system:
• lower capital cost
• equal or lower overall operating cost
• significantly easier operation and maintenance
• reduced outage time for installation
• shorter lead time for material supply and installation
• smaller reduction of plant electrical output.
3
-------
SECTION 2
INTRODUCTION
Control of emissions of sulfur and nitrogen oxides from existing coal-fired boilers is a
significant concern of the U.S. Environmental Protection Agency (EPA), the utility industry, and
other private and public organizations. These two pollutants are widely believed to be major
precursors of acid rain. Coal-fired utility boilers account for about 65% of the S02 and 29% of the
NO„ emissions in the United States.1
Recent legislation calls for reduction of S02 and NO„ emissions by different phases with
separate implementation dates. The choice of control techniques for limiting S02 and N0X
emissions from utility boilers will most likely include a mix of technologies to achieve the desired
reduction at minimum cost. Of the existing commercial control options, flue gas desulfurization
(FGD) is costly, and switching to a lower sulfur content fuel may have severe economic and logistic
impacts. Only 10% of the boilers east of the Mississippi River are subject to New Source
Performance Standards (NSPS).2 Many units in that region have a remaining useful life of 10 to 30
years. There is a need, therefore, for lower cost alternatives to FGD which can be retrofit to this
existing boiler population. One such technology is Limestone Injection Multistage Burner (LIMB).
This technology is based on the injection of a dry sorbent into the boiler for direct capture of S02 in
the flue gas, combined with the use of low NOx burners for reduction of NO* emissions.
BACKGROUND
The LIMB Process has its roots in the widely observed fact that significant sulfur retention
frequently occurs in fly ash of coals with low sulfur content and high alkali content. This is
particularly true with U.S. lignites and European brown coals. This fact was only an interesting
observation until the late 1960s, when the U.S. Congress began to debate broad and sweeping
legislation to limit sulfur emissions. Following several bench- and pilot-scale experiments at the
National Air Pollution Control Administration (NAPCA, the EPA's precursor), and at various
industrial sites, the Babcock & Wilcox Company (B&W) and the Tennessee Valley Authority (TVA)
under sponsorship of NAPCA, performed a full-scale demonstration of furnace limestone injection
on a 150 MWe unit at TVA's Shawnee Station.3 This project encompassed a period from 1969 to
4
-------
1971. The principal outcome of that test program was that the level of S02 removal (typically less
than 30%) was judged to be unsatisfactory. Following that test program, the commercial interest
in sorbent furnace injection waned in the early 1970s. This technology simply could not meet the
requirements of the 1971 federal NSPS.
Two situations began to evolve in the late 1970s which gradually rekindled interest in
sorbent injection technology. First, a number of potential applications began to appear for which
S02 removal efficiencies of 50% or less could be beneficial. For example, one Eastern utility
determined that if sorbent injection were capable of only 25% S02 capture, then this technology in
combination with coal cleaning could permit it to meet the 1971 NSPS on one of its units.
The second situation consisted of improvements in the performance capabilities of this
technology. These advancements were initially made by several German companies who were
looking for moderate sulfur removal improvements for brown coal units in anticipation of pending
air pollution control regulations. One approach tested was to mix limestone with the fuel and then
burn the mixture in multistage, low NOx burners. Preliminary pilot-plant results of both German
tests and tests by EPA indicated S02 removals as high as 80% at a Ca/S ratio of 3:1. Thus, the
acronym LIMB (Limestone Injection Multistage Burner) was born.
Although subsequent experiments revealed that other sorbents were more effective than
limestone and that other means of sorbent injection could perform as well or better than injection
through low N0X burners, these initial LIMB experiments did stimulate new scientific inquiry into the
technology. Some of the more important results of that research are summarized as follows:
• The temperature window for the sulfation reaction was identified to be about 2300 °F
to about 1600°F. The upper bound is a thermodynamic limit and the lower bound is a
kinetic limit. Since the reaction rate diminishes exponentially, the flue gas temperature
quench rate has become a design/performance parameter.4 6
• The importance of reactant surface area was identified. Flash calcination of limestone and
Ca(OH)2 can produce CaO surface areas of over 60 m2/g. The surface area generated in this
way is over an order of magnitude greater than is typically created by conventional limestone
calcination in rotary kilns. However, it was discovered that this reactant surface area can
diminish very rapidly in the furnace environment. Sintering, caused by exposure of the
calcined lime to excessive furnace temperatures for too long, is one reason for the loss of
5
-------
reaction surface. The other major cause is pore pluggage resulting from S02 reactions with
the lime.®
• As a result of the above two factors, a new and increased emphasis has been placed
on sorbent/flue gas mixing. Mixing must be accomplished rapidly within the
temperature window. Attempts at injection at temperatures much above 2300°F have
invariably resulted in poorer performance, presumably because of deadburning of the
lime and/or CaS04 instability.7
Thus, EPA's concept of LIMB gradually moderated from emphasis on limestone injection
through the burners to "sorbent" injection within the reaction temperature window. However, for
continuity, EPA has continued to use the acronym LIMB even when the temperature window
dictates sorbent injection through separate injection ports.
By the spring of 1984, EPA was prepared to demonstrate the LIMB technology on a
commercial scale. The demonstration project at Ohio Edison's Edgewater Station was the result.
The LIMB Demonstration contract was awarded to Babcock & Wilcox on September 28, 1984,
based upon a competitive procurement process. The project team consisted of Babcock & Wilcox
(project management, injection system design), Ohio Edison (host utility), Stone & Webster
Engineering (facilities engineering), Radian Corporation (emissions testing), and Energy and
Environmental Research Corporation (preliminary LIMB technology review). When the contract was
modified to include humidification, Consolidation Coal Company (Consol) joined the team and
assisted B&W in the technical development of the humidifier design. The demonstration was
co-funded by EPA, the Ohio Coal Development Office, and the project team members.
By late 1985, the possibility of adding a Coolside Process demonstration to the LIMB
project began to be considered. Coolside involves in-duct sorbent injection and would require the
addition of a humidifier. During this time period, the U.S. Department of Energy (DOE) issued a
Program Opportunity Notice (PON) as part of the Clean Coal Technology Program. The agreements
reached resulted in the inclusion of an in-duct humidifier to be built as an extension of the EPA
contract and a program to test the Coolside Process as part of the DOE contract.
6
-------
SECTION 3
PROJECT OBJECTIVES AND SCOPE
The basic goal of the LIMB Demonstration was to extend LIMB technology to a full-scale
application on a representative wall-fired utility boiler. The specific objectives were to demonstrate
that:
• S02 reductions of 50% or more and NOx emissions of less than 0.5 lb/106 Btu can be
achieved at a fraction of the cost of add-on FGD systems.
• Boiler reliability, operability, and steam production can be maintained at levels existing prior
to the LIMB retrofit.
• Technical difficulties attributable to LIMB operation, such as additional slagging and fouling,
changes in ash disposal requirements, and an increased particulate load, can be resolved in a
cost-effective manner.
The demonstration project consisted of several distinct phases - a preliminary phase to
develop the LIMB process and design applicable to the host boiler, a construction and start-up
phase, and an operating and evaluation phase. The development of the Edgewater LIMB design
was completed in January 1986, and major boiler-related components were installed during a boiler
outage in the autumn of that year. Fine tuning and characterization of the DRB-XCL™ low-NOx
burners were carried out during the first half of 1987, followed by shakedown of the sorbent
injection system that summer. The first S02 removal tests were conducted in September 1987
when the need for humidification became apparent as ESP performance degraded beyond
expectations. These tests were conducted over about a three-week period during which only a few
hours of continuous operation at a time were possible before the opacity climbed to unacceptably
high levels. Discontinuing sorbent injection allowed the ESP to recover within a number of hours
depending on the Ca/S ratio used and the unit load during and after each test.
Extensive laboratory and pilot tests on humidification had been on-going throughout 1987
under the EPA contract modification. The humidifier design was finalized in late 1987, with
procurement and installation lasting until mid-1988. Testing at Edgewater resumed in September
1988 and continued until June 1989.
7
-------
SECTION 4
DESIGN APPLICATION TO EDGEWATER
The design criteria for the LIMB system were established early in the program to include
provision for either lime or limestone as the sorbent for use with a nominal 3 wt % sulfur coal. The
parameters used to develop the design are summarized as follows:
Coal Composition
wt %
wt %
c
67.5
0
7.0
H
4.8
H20
6.5
S
3.0
Ash
10.0
N
1.2
CI
0.0
Heating Value 12,850 Btu/lb (as received)
Sorbents
Hydrated Calcitic Lime - Ca(0H)2
a)
b).
Boiler Rating
Steam Flow, Ib/hr
Heat Input, Btu/hr
S02 Removal, %
Ca/S Stoichiometry
Sorbent Flow, Ib/hr
Sorbent Composition
Bulk Density, lb/ft3
Maximum Continuous Rating
700,000
963 x 106
50
2.5
13,978
93% Ca(OH)2, 7% Inerts
30
Particle Mass Median Diameter, //m 5
Limestone - CaC03 (Although lime was selected for use as the
sorbent, the equipment was designed for limestone use as well.)
Boiler Rating
Steam Flow, Ib/hr
Heat Input, Btu/hr
S02 Removal, %
Ca/S Stoichiometry
Sorbent Flow, Ib/hr
Sorbent Composition
Bulk Density, lb/ft3
Particle Size
Peak Rating
770,000
1060 x 10®
50
2.5
20,453
95% CaC03, 5% Inerts
75
95% < 44//m (325 mesh)
8
-------
DRB-XCL™ Burner Specifications
Number of Burners
Burner Arrangement
Number of Pulverizers
Burner/Pulverizer Arrangement
12
4 high x 3 wide
4 (Series E64)
3 Burners/Mill,
1 Mill/Burner Row
Design Criteria (per Burner)
Rated Heat Input, Btu/hr
Fuel Flow, Ib/hr
Primary Air Flow, Ib/hr
Secondary Air Flow, Ib/hr
Primary Air Temperature, °F
Secondary Air Temperature, °F
Windbox to Furnace Differential, in WC
80 x 108
6.8 x 103
11.9 x 103
57.3 x 103
150
500
5.0
LIMB DEMONSTRATION SITE
The Edgewater Station of Ohio Edison is located in Lorain, Ohio, and is situated on Lake
Erie. The B&W boiler used for the demonstration is designated as Boiler No. 13 and is part of Unit
No.4, generating steam for the 105 MWe turbine. First commissioned in June 1957, it has a
radiant secondary superheater, a reheater, a horizontal primary superheater, an economizer, and a
tubular air heater. Compressed air was the primary medium used for sootblowing, although four
new steam blowers were added for the LIMB project. (All but four air heater sootblowers have
since been converted to steam for the LIMB Extension Project.) Four B&W EL pulverizers supply
coal to an array of 12 burners arranged three across by four levels high on a single wall. In 1982,
a new Lodge-Cottrell electrostatic precipitator (ESP) with a design specific collection area (SCA) of
612 ft2/103 ACFM was added. This 6-field, retrofit ESP is located about 100 ft from the boiler and
was expected to operate at 99.38% particulate removal efficiency with the original design coal.
LIMB PROCESS APPLICATION
Figure 1 is a schematic representation of the LIMB/humidifier system installed at
Edgewater. The sorbent system consists mainly of sorbent storage facilities, a gravimetric feed
system, a sorbent distribution system, and a set of sorbent injectors. From an equipment
standpoint, the space requirements were relatively minor. Only the sorbent storage silo and
neutralization acid storage tank required the use of additional outside space. All of the other
9
-------
Figure 1.
The LIMB process at Edgewater.
-------
sorbent facilities were located around the boiler at various elevations. The humidifier depicted in
Figure 1 required a significant new installation. The humidifier was built in a bypass loop on the
boiler house roof in a space previously occupied by a decommissioned ESP. In fact, the humidifier
was located over one set of hoppers from this old ESP. This feature was included to provide a
means to collect ash in case ash buildup became a problem in the humidifier. However, these
hoppers remained covered throughout the project, leaving the humidifier resembling an in-duct
configuration. The decision to build a separate humidifier rather than to use the existing duct
between the air heater and ESP was a matter of prudence. Since the demonstration was intended
to be a long-term test and since questions existed concerning the operating reliability of in-duct
humidification, the decision was made to not use the existing duct as a humidifier at that point in
time. It is now clear, however, that the technology requires substantially less in the way of new
capital equipment or additional space than would be required for a wet FGD system on a
moderately sized boiler. The smallest impact case would entail a sorbent storage silo, minimal
humidification achieved in an existing duct, and ash treatment off-site. At the other end of the
spectrum, humidification to a close approach to saturation and on-site ash treatment, as was done
at Edge water, add to the cost and space requirements, though not to the extent that a wet FGD
system would impose.
DESIGN BASIS
As described earlier, laboratory and pilot experiments identified the LIMB sulfation reaction
window to be between 2300°F and 1600°F. This fact, coupled with the importance of reactant
surface area, dominated the process design for LIMB. The primary design issues are summarized
as follows:
• Determination of the physical location of the sulfation temperature window within the
Edgewater boiler. Since the location of this temperature window changes with load as
well as with other operating variables, it was necessary to determine its location under
different operating conditions.
• Delivery of sorbent from the feed silo to the individual injection ports in an accurate, evenly
distributed, non-slugging manner.
• Design of a sorbent injection system that will disperse the sorbent quickly and uniformly
within the boiler.
• Selection of burners that satisfy the N0X emission goals and are compatible with the physical
11
-------
dimensions of the Edgewater boiler.
• Evaluation and/or modification of the particulate control system to maintain particulate
emission limits in the face of a 2-to-3-fold increase in the amount of ash.
• Design of an environmentally acceptable waste handling and disposal system.
BOILER CHARACTERIZATION
Due to the critical importance of the boiler temperature and velocity profiles on sorbent
reactivity and sorbent mixing with combustion gases, a considerable effort was undertaken to
characterize the Edgewater boiler. The work included flow characterization in a 1/15th scale
acrylic model of the boiler, use of a computerized boiler model to estimate temperature profiles, and
direct measurement of temperature and velocity at selected points in the boiler.
The boiler characterization data revealed important implications for sorbent injection. The
temperature window is located in the upper furnace from a plane just above the nose to a plane in
the convective section. A reduction in boiler load moves this zone back toward the furnace.
Bubble streak flow visualization in the scale model showed that the flow is nearly uniform across
the width of the boiler. The residence time in the optimum temperature window is very short, on
the order of 1 second, indicating that this boiler provides a rigorous test for sorbent injection.
SORBENT FEED SYSTEM
The main design issues in the sorbent feed system were to obtain a known, controllable
flow of sorbent from the feed silo, to deliver that sorbent pneumatically to a distribution system,
and then split the sorbent uniformly to a number of injector nozzles operating in parallel.
Furthermore, this must be done without creating a slugging condition at the injection point. This is
a critical operational issue. Pressure and visual observations have shown that the sorbent feed
system meets all of these requirements. The injection system design is based upon a similar
system which B&W developed and employed on pulverized-coal injection into blast furnaces.
SORBENT INJECTION
Three levels of injector nozzles were installed on the front and side walls of the boiler. The
locations were selected based upon the results of the boiler characterization studies, as well as
-------
flow modeling of the Edgewater unit. These injector locations are depicted in Figure 2. The lowest
level of injectors is located at an elevation of 181 feet. Eight nozzles are spaced uniformly across
the front wall of the boiler. A second row of injection nozzles is located 6 feet above the first.
This row also has eight nozzles along the front wall, plus two nozzles along each of the side walls.
The third row of injector nozzles is located 4 feet above the second row in an arrangement identical
to the first row. All nozzles can be tilted through a 30° arc. Each injector consists of a center pipe
which carries the pneumatically-transported sorbent and an outer pipe which carries booster air.
From a design point, the injector is sized to achieve a momentum flux which will permit the sorbent
to penetrate the flue gas from the wall.
BURNER SELECTION
In order to meet or surpass the objectives of the program, an evaluation of commercially
available low NOx burners was undertaken early in the project. A literature survey of low NOx
burner designs was conducted to support the selection of a suitable burner for the LIMB
demonstration. This literature survey focused on the following types of burner designs as
applicable in the commercial scale:
• Babcock & Wilcox's Dual Register Burner (DRB)
• Foster Wheeler's controlled flow/split flame burner
• Riley Stoker's controlled combustion venturi burner
All three commercial low NOx burners were designed to modify the air/coal mixing at the
burners. The principal objective in the burner design was to regulate the mixing rate at the burner,
resulting in fuel-rich pockets in the near burner region, and gradual downstream mixing with the
secondary air. This serves to control NOx formation by reducing oxygen availability during
combustion. The flame patterns from these three burners are similar in that the flame volume
tends to increase relative to the original circular burners, due to more gradual air/fuel mixing and a
concentrated central fuel-rich zone exists in the near-flame region.
These commercial burners had been developed by the boiler manufacturers for burner
retrofits in pre-NSPS (New Source Performance Standards) units and direct installation in
post-NSPS units. B&W's DRB was designed to replace the circular burner, Foster-Wheeler's
controlled flow/split flame burner was to replace the intervane burner, and Riley Stoker's controlled
13
-------
Elev. 191'
Elev. 187'
Elev. 181'
-------
combustion venturi burner was to replace the flare burner. The circular, intervane, and flare
burners are single-register coal burners designed for high air swirl and rapid mixing at the burners to
obtain complete carbon utilization.
It was concluded at the end of the literature survey that all three commercial burners had
applied similar design principles. The B&W DRB had the advantage of being the most compatible
with the design of the B&W boiler.
At about the same time Babcock-Hitachi K.K. (BHK), the exclusive licensee of the DRB for
applications in Japan, succeeded in developing a new low N0X burner that provided even greater
N0X reduction. The BHK burner was designed to create a N0X "reburning" reaction in the region
immediately downstream of the burner's exit. This low N0X burner, denoted as the HT-NR burner,
had some similarities to the DRB: two annular air zones, the inner with adjustable spin vanes and
the outer with a circular, louver door register. In order to achieve the desired combustion
reactions, BHK varied the air zone splits and velocities and added enhancements to the burner to
create a unique flow pattern which increases flame stability and enhances a reburning reaction by
increasing hydrocarbon radical generation in the flame's fuel-rich core. Babcock-Hitachi's HT-NR
burner had demonstrated NOx reductions of over 50% on pulverized coal and was in commercial
use in Japan in several boilers.
B&W holds an exclusive license for the application of the HT-NR burner in North America.
In an effort to better meet the demand for a retrofit low NOx burner in the U.S., B&W had
undertaken a program to modify this burner to provide enhanced mechanical reliability, air
measurement capabilities, and a reduced burner pressure drop.
The resultant burner is referred to as the DRB-XCL™ burner. The major mechanical change
was to replace the circular register design with an air sleeve not unlike the inner air zone of the
DRB and HT-NR. This modification reduces the burner's overall pressure drop and allows for air
measurement instrumentation to be installed on the burner. The addition of air directional vanes on
the inlet to the burner also helped reduce pressure drop. Flexibility is increased by independent
controls for flow and air swirl in the inner and outer air zones. Additional full-scale burner tests
were conducted with all three designs (DRB, HT-NR, and DRB-XCL™) before the DRB-XCL™ was
eventually selected for the demonstration. Results of the tests are provided in Section 7.
15
-------
ELECTROSTATIC PRECIPITATOR CONSIDERATIONS
Early in the project, the effect of LIMB on electrostatic precipitator (ESP) performance was
evaluated. The concerns included the higher dust loading (by a factor of 2 to 3), finer particle size,
and higher ash resistivity - all of which can result in reduced performance. On the basis of
analogous experience and mathematical modeling, it was concluded that this "over-sized" ESP
could meet the required emissions and opacity requirements. Nevertheless, this evaluation gave
rise to strong consideration of gas conditioning as a possible additional requirement for the process.
The usual method of conditioning is to inject sulfur trioxide (S03) into the flue gas upstream of the
ESP. However, because the LIMB system contains a large excess of free lime which would react
with the sulfur trioxide (S03) in the flue gas, an exceptionally large amount of S03 would be
required to achieve sufficient reduction in resistivity.8
Humidification provided another option for conditioning flue gas upstream of an ESP. The
conditioning mechanism is similar to the S03 mechanism. Fly ash is a refractory material whose
electrical resistance increases sharply with decreasing temperature. The use of S03 can alter this
condition by condensing small quantities of sulfuric acid on the surface of the fly ash. This creates
an electrically conductive path through the layer of ash on a collection plate in the ESP. With S03,
this phenomenon typically becomes significant around 300°F. If S03 is absent from the flue gas as
is the case with LIMB, water vapor is thought to serve the same function, although it requires a
lower temperature before its effect becomes significant.
The combined factors of flue gas conditioning, the potential for additional S02 removal in
the humidifier, and the plans for demonstrating the Coolside Process all contributed to the decision
to proceed with design and installation of the humidifier at Edgewater.
HUMIDIFICATION
Development of the humidification process presented a number of design issues that
needed to be resolved:
• The cross-sectional area of the humidifier determines the gas velocity and hence the
residence time of the water droplets in the humidifier. If the area is too small, droplets exit
16
-------
the unit before complete evaporation occurs. These droplets may impact a wall or turning
vane at the first elbow downstream, causing surface wetting and a subsequent ash build-up
problem. An oversized cross-section, which would maximize residence time, reduces gas
velocities to the extent that the flue gas lacks enough momentum to deflect the spray jet
away from the walls. The expanding jet will impact and wet the walls early in the humidifier,
leading to a build-up problem. The optimum humidifier cross-section was therefore required.
• Additional information needed for humidifier design consisted of the type and number of
atomizers, the optimum distance between atomizers, and the minimum distance from the
array's outermost atomizers to the wall.
• Good distribution of the flue gas at the inlet of the humidifier is an absolute necessity to
eliminate localized areas of high water concentration and resulting wet areas in the
humidifier. The exact placement of flow straightening devices was necessary design
information.
• At Edgewater, the humidifier, the bypass duct in which it was installed, and the flue gas
reheater impact the gas side system pressure drop and the induced draft fan capacity.
Assurance that fan capacity, a site specific factor, would remain adequate was necessary.
The approach taken to develop the required design information was to evaluate the
performance of a number of commercially available atomizers, choose the superior performer(s) and
test it(them) in an array in a 6 foot square duct. Concurrently, a mathematical model was
developed to improve understanding of the humidification process and a 1 /12th scale cold flow
model was constructed and operated to investigate gas flow patterns and gas pressure drop.
WASTE HANDLING AND DISPOSAL
Application of LIMB technology creates additional waste and changes the fly ash properties.
The Edgewater waste handling system is a dry vacuum system in which ash is pneumatically
removed from the ESP hoppers to an ash silo, sprayed with a small amount of water in a pug mill
for dust control, and trucked to a nearby landfill. It was found that the lime in the ash could be
hydrated in the pug mill and a dust-free material could be produced for hauling to the landfill.
Control of the wetting process is critical to avoid pug mill pluggage, hardening of the material in the
truck, and for dust control (fugitive emissions).
Early in the project, studies were conducted on LIMB ash obtained during full-scale tests at
17
-------
Ontario Hydro's Lakeview Station, as well as from a pilot combustor at Southern Research
Institute." In addition to providing information that would be necessary for the safe handling,
transportation, and disposal of the ash, the samples were subjected to the Extraction Procedure
(EP) Toxicity Test10 as specified in the Resource Conservation and Recovery Act (RCRA). The
results showed that leachable toxic materials were far below levels considered to be hazardous.
Subsequently, the bulk of the LIMB-modified ash generated at Edgewater was disposed of in an
existing landfill adjacent to the pile of conventional ash. However, approximately 120 truckloads of
LIMB ash were used to construct two separate test cells for observation and testing over a period
of several years.11
18
-------
SECTION 5
DETAILED SYSTEM DESCRIPTION
SORBENT STORAGE AND TRANSFER SYSTEM
The sorbent storage and transfer system shown in Figure 3 holds up to a 43 hr supply of
lime (73 hr supply of limestone) and moves it to the feed silo on demand. Sorbent is delivered to
the plant by pneumatic, self-unloading trucks, which normally carry 20-25 ton per load. Two
trucks can unload simultaneously while a locally controlled bag filter vents the transport air. The
sorbent transfer system operates automatically to fill the feed silo upon a low sorbent supply
signal. An aerated bin bottom promotes sorbent flow from the silo, with low pressure, dry air
supplied by a dedicated lobe-type blower and an air dryer.
Flowing from the silo through the slide gate and rotary valves, sorbent enters the Fuller-
Kinyon powder pump. Compacted as the screw moves it through the barrel, the material provides a
seal against the backflow of transport air. The sorbent, discharged into the pump's mixing
chamber, is fluidized by compressed air introduced through a series of jets, and carried into the
transport line in the dilute phase. A rotary vane compressor supplies the transport air for the Fuller-
Kinyon pump.
The 8 in carbon steel sorbent transport line conveys the sorbent from the outdoor (300 ton
working capacity) storage silo to the indoor (15 ton working capacity) feed silo. At the feed silo,
an inlet alleviator and a pulse jet vent filter separate the sorbent from the conveying air. Control
cabinets for the sorbent transfer operation are located in the boiler house near the feed silo. In
addition, hand on/off switches are available at the equipment for maintenance purposes.
The transfer system is sized to deliver 50 ton/hr of either hydrated lime or limestone to the
feed silo. At this rate, the system operates for about 20 min every 2 hr and 20 min when using
hydrated calcific lime at a Ca/S stoichiometry of 2.5 under full load conditions.
19
-------
To Atmosphere
Figure 3. Sorbent storage and transfer system.
20
-------
FEED SYSTEM
The feed system shown in Figure 4 provides a controlled feed rate of sorbent to each of
two injection elevations. There are two independent feed systems, each including a differential-
weight-loss feeder manufactured by Acrison, Inc. The feeders have their own hoppers with an
auger discharge from the bottom. Mounted in the feed silo's two discharge chutes above each
feeder's hopper are DeZurik slide gate valves to isolate the hopper from the feed silo. Mounted
directly below each slide gate is a rotary valve to prevent the lime from flooding into the feeder's
hopper. This hopper is mounted on a balanced scale system which determines the feed rate of the
sorbent. The motor speed of the feeder discharge auger is automatically adjusted to match the
actual feed rate to the set feed rate. When the sorbent level in the hopper has dropped to a set
level, the feeder's control system signals the slide gate to open and the rotary valve to turn. At the
same time, the Vibranetic's bin vibrator mounted above the slide gate is signaled to start sorbent
flow from the feed silo. Once the feeder's hopper has been filled, the vibrator automatically stops
and the slide gate closes. The sorbent set feed rate can be controlled manually or automatically
from selector stations located in the main boiler control room. In the automatic mode, the feed rate
will be boiler load following. From each feeder, the sorbent passes through a rotary valve. The
rotary valves are designed to seal the feeders from the 20 psig conveying air.
Conveying air carries the sorbent, in a dilute phase flow, to the two distribution bottles
feeding two elevations of injection nozzles. A 4 in inside diameter (I.D.) hose connects the pick-up
point at the discharge of each rotary valve with the 4 in connection to its distribution bottle. The
distribution bottle is designed to provide a smooth and even distribution of sorbent to each injection
line. The injection lines were arranged such that equal resistance is created in each line minimizing
the possibility of uneven flow to each injection nozzle. One distribution bottle feeds the 12 ports at
elevation 187 ft, while the other supplies sorbent to the 8 ports at either elevation 181 or 191 ft.
Although all sorbent piping was originally 1.25 in I.D. opening up to 1.5 in I.D. within the injection
nozzle, the feed piping from the 8-port distribution bottle was changed over to 2 in I.D. all the way
through the nozzle to reduce back-pressure. This was done before testing began in 1988. The
feed system is designed to feed 2.0 to 10.25 ton/hr of sorbent per feeder. The dilute phase air
flow is determined from the dilute phase mixture density which lies below 3 lb/ft3 for stable flow.
21
-------
Distribution
Bottles
Figure 4. Sorbent feed system.
22
-------
CONVEYING AIR SYSTEM
The conveying air is provided by a rotary screw compressor. From the compressor, the air
passes through a dryer and into a receiver. From the receiver the air passes through a coalescing
filter and then to the control valve station, where the flow is automatically controlled for proper
dilute phase conveying.
BOOSTER AIR SYSTEM
One booster air fan, provided by Garden City Fan Company, supplies the required booster
air for all injection nozzles. The high velocity booster air accelerates and carries the sorbent into
the furnace. An impact tube/flow transmitter at the fan discharge sends a total booster air flow
signal to the distributed control system (DCS) which controls the inlet vortex damper. Each
individual booster air line to the nozzle has a butterfly valve for throttling. Though only used in
early tests when flow was split between two elevations, impact tubes with local flow indicators
were provided to measure the booster air flows to the individual injection nozzles. If the booster air
fan is turned off or tripped off, a vent in the booster air duct automatically opened to allow outside
air into the piping. Suction provided by the boiler pulled the air through the injection lines to serve
as a cooling medium for the injection nozzles.
INJECTION NOZZLES
The injection nozzles essentially consist of two concentric pipes with sorbent feed through
the inner pipe, while booster air is introduced through the annular area created by the outer 3 in,
schedule 40 pipe. The material of construction is 304 stainless steel. Pipe sizes were chosen to
allow variation of momentum flux ratio within the range expected to be required. As shown in
Figure 5, the nozzle assembly is hinged and provided with a locking bar in order that it can be tilted
over a range of 30°. The nozzles for elevation 181 can be varied ±15° from horizontal, while,
those for the upper elevations can vary from horizontal to 30° down.
SOOTBLOWERS
Four additional Diamond IK sootblowers were installed in anticipation of increased fouling.
One set of the new sootblowers is located upstream of the horizontal secondary superheater inlet
23
-------
Figure 5. Injection nozzle and furnace wall port.
24
-------
tube bank. The second set is located between the first and second bank of the primary
superheater tubes. The sootblowers utilize steam as the blowing medium. The steam is obtained
from the main steam header and reduced to 300 psig through a steam reducing station. Each
blower has a local start/stop switch for local operation and trouble-shooting. A new Diamond
Microselect 64 system was installed to control the sootblowers.
DRB-XCL™ BURNERS
The original B&W circular register burners shown in Figure 6 were replaced by B&W
DRB-XCL1" burners shown in Figure 7. This burner was developed to provide low NOx emissions,
high combustion efficiency, and mechanical reliability while firing a pulverized coal (PC) fuel. The
burner is derived from B&W's Dual Register PC burner and utilizes concepts from Babcock-Hitachi's
HT-NR burner.
The burner consists of a centrally located nozzle for the pulverized fuel/air mixture and two
secondary air zones. The DRB-XCL™ burner is designed for flexibility in retrofit applications and
can utilize different nozzle configurations for specific applications. For a new boiler or an opposed-
fired retrofit, the DRB-XCL™ burner would utilize a conical diffuser/deflector arrangement to
distribute the fuel/air mixture in the nozzle. This arrangement provides a low nozzle exit angle,
leading toward lower NOx generation and longer flames. For single wall-fired retrofits such as at
Edgewater where flame length is of concern, an impeller is utilized in the coal nozzle to disperse the
fuel into the secondary air very rapidly. The impeller has a beneficial effect on flame length,
though NOx emissions increase somewhat due to the increased turbulence at the flame's base.
In optimizing the application of DRB-XCL™ burners for Edgewater, a conical
diffuser/deflector arrangement was installed with flame stabilizing rings on the second and fourth
rows from the bottom, while 30° impellers and distribution cones were installed in the first and
third rows. The overall effect reduced NOx generation while utilizing the impellers to minimize
carbon loss. The flame stabilizing ring recycles air back toward the coal nozzles which holds the
base of the flame close to the burner.
The secondary air through each burner is introduced to the furnace through two axial flow
air zones. The majority of the air passes through the outer air zone. This zone has two sets of
25
-------
Adjustable
Figure 6. Original circular register burner.
26
-------
Figure 7. Retrofit low NOx DRB-XCL™ burner.
27
-------
vanes to add turbulence to the combustion air, fixed upstream vanes, and adjustable vanes that
provide the initial spin direction at a very low pressure drop. The downstream vanes are adjustable
and provide the balance of spin direction required for proper combustion.
Air flow to the outer zone is controlled by a sliding damper that rides on the inner air zone
barrel. To facilitate air flow balancing, each burner is equipped with a pitot-grid measurement
device at the outer air zone inlet. Utilizing the pitot differential, the windbox air distribution can be
properly balanced.
The inner air zone has one set of spin vanes which are adjustable from outside the
windbox. Air flow to this zone is controlled by the spin vanes. Lighter, scanner, and observation
port penetrations pass through the inner air zone. An air separation sleeve is attached to the outlet
of the inner zone. This sleeve redirects the air flow from the outer zone to enhance the delayed-
mix action of a dual zone burner. It also creates recirculation zones that support the flame's
stability.
By reducing the complexity of the burner's mechanical aspects, the DRB-XCL™ burner has a
high degree of reliability. Enhanced delayed-mix actions provide NOx emissions lower than first
generation delayed-mix burners. The high turbulence created by the DRB-XCL™ burner creates a
stable combustion zone that provides high burnout efficiency.
BYPASS FLUE SYSTEM FOR HUMIDIFICATION
The bypass system shown in Figure 8 became necessary when Ohio Edison could not risk
use of their existing flues for the humidification process since problems could cause the shut-down
of Unit 4. As a system, the bypass provides a separate run of flue including the humidification
chamber and functions in parallel with the existing flues running from the air heater to the ESP.
New flues connect the inlet to the chamber with the air heater outlet flue, and the outlet of the
chamber with the existing flue further downstream.
Modulating louver dampers are supplied to bring the bypass system into operation or to
take it out while the boiler is in operation. These may also be used for controlling the flow through
the bypass when only a partial flow is desired. Additionally, guillotine shut-off dampers are used to
28
-------
Stack
S3
CO
Existing
Air Heater
Outlet
New Bypass
Flue With
Humidification
Chamber
Outlet
Flue To
Precipitator
Outlet
Flue From
ID Fan
Guillotine Isolation Damper
Louver Control Damper
Steam
Coil Reheater
Figufe 8.
Bypass system - plant orientation and schematic.
-------
isolate the humidification chamber for internal maintenance with the boiler in operation. B&W's
standard turning vanes were supplied at all new flue bends for flow uniformity and to limit the
increase in draft loss. Safety interlocks are applied to this system to insure against furnace
overpressure, precipitator or flue implosion, or loss of gas path due to improper damper operation.
HUMIDIFICATION CHAMBER
The basic dimensions of the humidification chamber are shown in Figure 9. The 14.6 by
14.6 ft cross-section reduces the peak load inlet flue gas velocity to 31 ft/sec which provides a
chamber residence time of about 2 sec. These conditions were considered in the design of the
atomizer and its array in order to minimize the amount of unevaporated water droplets that might
impact on walls and downstream turning vanes.
As noted briefly in the preceding section, the humidification chamber was installed over the
hoppers of the retired ESP on the boiler house roof. Hopper baffles were modeled and included in
the design to optimize the flow characteristics in the chamber in case ash had to be removed
through open hoppers. These baffles were used to support removable steel plates that served as a
chamber floor in the oversized duct configuration employed. Although it was never necessary,
these plates could have been removed had there been unexpectedly large deposits or ash fall-out
that would have required usage of the hoppers.
Structural design parameters for the humidification chamber were based on supporting the
equivalent of one foot of wet ash on the chamber floor, with the accumulation of deposits being
monitored by a system of strain gauges mounted on the chamber support legs. The ash was
generally free-running with friable deposits when they occurred. Both were easily removed with
vacuum hoses when a totally clean chamber was desired for test purposes. The single longest
period of continuous operation at close (20 to 25 °F) approach to saturation was 11 days during
the LIMB Extension/Coolside test program.
ATOMIZATION WATER FEED SYSTEM
Water used for humidification is supplied from the existing Unit 4 service water system,
which consists of filtered Lake Erie water. Water is first fed in parallel through the atomizing air
compressor first and second stage intercoolers, and the auxiliary instrument air conditioner. Its
30
-------
Figure 9. Humidification chamber detail.
31
-------
pressure is boosted by an in-line pump to obtain the necessary flow through the intercoolers. The
water is then either returned to the raw water intake channel through a back-pressure control valve
or fed through a basket strainer into the water storage tank through a level control valve.
From the storage tank, the water is pumped through another duplex basket strainer to a
flow control valve that feeds water to the atomizers at a controlled rate to maintain a set approach
to the flue gas saturation temperature downstream of the humidification chamber. Inlet pressure to
the flow control valve is maintained by a back-pressure control valve, which returns water to the
storage tank.
At the spray deck the water feeds through the third and final duplex basket strainer before
splitting its flow to the north and south atomizer supply headers. Each header supplies eleven
lances through a flow element, a three-way air purge valve, and an air operated ball valve. This
last valve is operated by atomization air providing an interlock that stops water flow to each lance
if air has not already been supplied to that lance. This prevents unatomized water from entering
the chamber.
The atomizer supply headers are vertical. To compensate partially for effects of static head
on the atomization pressure, two control valves are located after the top four and before the
bottom three take-offs. They operate to maintain offsetting pressure drops among the three groups
of atomizers.
ATOMIZATION AIR SUPPLY SYSTEM
A centrifugal compressor pulls air through an air inlet filter mounted on the roof. In this
three-stage compressor, air passes through the first and second stage intercoolers and moisture
separators. Air is discharged from the third stage and goes directly to the receiver without
aftercooling.
From the receiver, the air passes through a Y-type strainer and a control valve which
maintains a set differential between the water and air header pressures. The air parallels the water
by splitting into two vertical headers supplying eleven lances each on the north and south sides of
the humidification chamber. The lines to each atomizer have a manual shut-off valve and a Y-type
filter.
32
-------
ATOMIZATION WATER INJECTION SYSTEM
Performance tests of both B&W proprietary and commercially available nozzles led to the
selection of the B&W Mark XII atomizer depicted in Figure 10. This design was based on the B&W
l-Jet atomizer used in dry scrubbing applications, though scaled down for this application.
The optimum size of the atomizers and their array in the humidification chamber was
developed by tests that evaluated residence time and wetting factors. Residence time had to be
sufficient to allow evaporation of the atomized droplets of water before leaving the chamber. The
distance from the wall was set to avoid impacting the wall. The distance between atomizers was
set to avoid droplets impacting one another and coalescing into large drops that would fall out of
the gas stream and wet the floor of the chamber.
The original atomizer array, 10 across on 12 in centers and 11 rows vertically on 12 in
centers, consisted of an airfoil assembly at the inlet to the humidification chamber. Eleven lances
are inserted from each side, each including five atomizers mounted within an airfoil assembly
containing water and compressed air piping and vent air passageways. The water and compressed
air are supplied from the headers through flexible hoses. The vent air is provided by sucking in
ambient air through the lance wall mounting plate. The compressed air and water are mixed within
the atomizer and blow out through orifices providing atomization. The vent air flows around the
outside of the atomizers and prevents eddy currents from recirculating droplets which could wet
the atomizer tips and cause a buildup of fly ash to occur. While the original vent air ports were
simply holes in the airfoil assembly, three of the lances were replaced with an improved design that
incorporated nacelles around the individual nozzles. These assemblies were even more effective in
minimizing ash deposits on the atomizers. These modifications were able to be made to only three
lances a few months before the end of the project. Ten rows of atomizers were used for most of
the testing since the jets of the lowest row would impact and wet the accumulated nozzle ash
deposits that fell from the overall assembly. The completely modified array used during the LIMB
Extension work is depicted in Figure 11.
FLUE GAS REHEAT SYSTEM
A condensing type steam coil reheater designed to increase the flue gas temperature from
150 to 190°F was installed in the flue just downstream of the ESP. It contained four removable
33
-------
Flue Gas Flow ¦
¦Compressed Air
Nacelle
Atomizer
Water
Shield Air
Airfoil
Figure 10. Cross-section of atomizer with nacelle.
34
-------
Water
Supply
\
/
Atomizer
Air Supply
Air Foil Lance
Support
Pin
Figure 11 Airfoil lance and atomizer array.
35
-------
tube sections built from 1 in outside diameter (O.D.) by 0.065 in average wall, bare, carbon steel
tubes. Saturated steam at 1589 psig from the boiler feeds through a pressure reducing station and
a flow control valve to enter the reheat sections at 300 psig and 423 °F. The flow of condensate
(water rate) to the condenser hotwell is controlled to maintain a constant level in the condensate
drain tank. Each section is separately valved to allow for the removal of individual sections for
repair or replacement. Maintenance platforms and trolley beams for hoists are supplied to augment
the section replacement operation.
ASH REMOVAL SYSTEM
The ash removal system, shown in Figure 12, incorporates the existing vacuum conveying
system with the following changes made to accommodate the LIMB test program:
• The hopper air intake valves for the four air heater/economizer hoppers were replaced with
header systems feeding heated air past the hopper ash discharge valves. A 144 Kw electric
duct heater raises ambient air temperature to 400°F for transporting the ash. This prevents
condensation which could cause pluggage or corrosion problems.
• A new hydraulic exhauster and dust pugger were installed to produce additional vacuum and
ash transport capacity.
• Prior to the demonstration project, the existing Unit 4 air heater/economizer ash conveying
line was routed to discharge into the Unit 3 ash storage silo as a matter of convenience. In
order to collect all the alkaline LIMB ash in one silo, this conveying line was rerouted to the
Unit 4 ash silo.
The hydraulic exhauster (eductor) provides the vacuum required to pull the heated air-ash
mixture through the insulated ash transport lines to the ash silo. The ash is separated from the air
by a cyclone separator and baghouse in series on top of the ash silo. The cleaned air is then pulled
through the eductor where it contacts the water used to produce the vacuum. The air/water
mixture flows into the air separator where the air is vented to the atmosphere. The water drains
back to the pond, past the acid injection point where its alkalinity is reduced to a range complying
with water quality standards.
In the ash silo, aerator pads fluidize the ash which exits the silo and enters the variable
speed airlock rotary valve. The ash is then fed into the ash conditioner where it is mixed with the
36
-------
Air
Figure 12. Ash removal and ash pond pH control.
37
-------
proper quantity of water to slake the unreacted lime and eliminate fugitive dust. The water-ash
mixture weight ratio, typically about 1:7, is tightly controlled. Adding too much water to the ash
causes the pozzolanic mixture material to set up and become hard. Adding too little water creates
fugitive dust control problems.* An excess of water is added to the ash since some of it is
evaporated during the exothermic reaction between lime and water. This reaction continues for ten
to fifteen minutes after the mixture is dumped into the disposal truck.
pH CONTROL SYSTEM
The pH control system, located as shown in Figure 12, was supplied to neutralize alkaline
water that would flow into the ash pond if left untreated. The condition results from LIMB
operation in that some alkaline ash mixes with recycled pond water in the hydraulic exhauster,
while additional lime collects in the yard sump as a result of spillage and surface runoff in the ash
truck loading and lime truck unloading areas. Neutralization is effected by addition of sulfuric acid
which is held in a 5000 gal storage tank and moved by a transfer pump to a day tank on the acid
skid. A controlled volume, diaphragm metering pump adds acid to the water being discharged to
the pond in order to maintain the pH within the 6 to 9 range.
SUMMARY OF MAJOR MODIFICATIONS
The following modifications of the original design were made to improve the operations of
various systems and to obtain the reliability required for acceptable test conditions. They are all
incorporated in the final design as described above.
Sorbent Feed System
The back-pressure generated at the discharge of the rotary valves caused excessive leakage
back through their seals. This aerated the lime entering the rotary valve, decreasing its density and
throughput. It also placed pressure on the outlet connection of the differential weight loss feeder,
upsetting its mechanical balance and accuracy. As a result, the following corrective measures
were taken:
• The dense phase portion of the lime conveying piping was eliminated and the 1.5 and
2.5 in pipe sizes were replaced with 4 in I.D. hose between the pick-up points and the
-------
distribution bottles. The elimination of dense phase conveying also removed the
complication of associated valves, instruments, and controls.
• The distribution bottle inlets were increased from 2.5 to 4 in and the distribution pipes
feeding elevation 181 ft injection nozzles were increased from 1.25 to 2 in diameter.
• The inner pipes of the injection nozzles at elevation 181 ft were increased in diameter from
1.5 to 2 in. The 3 in outer booster air pipes remained the same.
• The rotary valve inlet side and pocket vents were rerouted to the air heater inlet
instead of to the feed silo. Venting to this negative pressure area decreased the
pressure on the feeder outlet connection. It was observed that the amount of lime
vented with the air was negligible.
In order to prevent the lime from flooding into the differential weight loss feeder hoppers,
rotary valves were placed in the discharge pipe from the lime feed silo.
Sorbent Transfer System
The speed of the Fuller-Kinyon lime transfer pump was increased to meet required capacity
by resheaving the V-belt drive. This was required because of the fineness of the hydrated lime.
The felt bags in the lime feed silo vent filter were replaced with Gore-Tex™. This material
successfully handled the very fine particles which tended to blind the original bags.
Humidification
The airfoil lance assemblies were revised by adding nacelles for vent air to reduce deposits
on atomizer tips and improve humidifier performance and availability.
Pond dH Control
The pond pH control system was upgraded as follows during the demonstration to reflect
the system shown in Figure 12:
• The yard sump pump discharge was rerouted into the pond piping upstream of the
39
-------
primary pH probe. Previously it had gone untreated to the pond. This sump collects
LIMB ash silo unloading spills with its high unreacted lime content which could cause
high pond pH.
• A primary pH probe was added just upstream of the acid injection point to sense the
need to inject acid and provide a feed-forward signal to the acid metering system.
• Just downstream of the acid injection point a section of the 10 in ash water pipe was
replaced with a loop of about 40 ft of CPCV piping. This permitted thorough mixing of the
acid with the effluent to protect the balance of the carbon steel pipe running underground to
the pond.
• A second pH probe was installed after the mixing loop to measure the pH of the water going
to the pond and provide a feedback signal to the acid metering system.
40
-------
SECTION 6
LIMB OPERATIONAL ISSUES
1987 LIMB SYSTEM START-UP EXPERIENCE
Initial, continuous sorbent injection during July 1987 was characterized by intermittent
operation due to problems encountered with the sorbent feed system and the necessity of
incorporating design changes to attain full-flow capability. The initial maximum feed rate was
limited to approximately 4000 Ib/hr of hydrated lime, far below the design feed rate of 18,000
Ib/hr. This problem was traced to the inability of a rotary valve to seal against the high back-
pressure of the transport air into which the lime feeds, causing air to leak back through the feed
system and disrupt flow. At this time the problem was diminished by reducing the back-pressure
by switching to larger diameter lines.
At first, the feed rate was found to be somewhat difficult to control. This was primarily
due to the flow properties exhibited by the finely-sized hydrated lime and pressure changes when
the feed silo filled. Consistent feed control for the short-term tests was obtained by utilizing a
single feed silo full of sorbent for each test. Rotary valves were then installed above each feeder
before testing resumed in 1988.
Once sorbent feed rates corresponding to a Ca/S ratio of 2.0 (10,000 to 13,000 Ib/hr) were
achievable, operation was characterized by steady degradation of ESP performance as indicated by
an increase in stack opacity. The current versus voltage relationship indicated that the high
electrical resistivity of the ash was producing back corona in the ESP. The condition was so severe
that all attempts to find an effective electrical solution were unsuccessful. Measurements of ash
resistivity using an in situ point-plane probe showed that sorbent injection resulted in an increase
from approximately 3 x 1010 to 3 x 1012 ohm-cm8. As expected, mass train data showed that the
particulate loading almost doubled, while particle size measurements indicated a decrease in
particle size.
As a result of the decreased ESP performance, continuous sorbent injection was limited to
3 to 8 hours, depending on the feed rate. Typically, injection would be terminated when stack
41
-------
opacity began to rise rapidly into the 10 to 15% range in order to avoid exceeding the opacity limit
of 20%. Upon termination of sorbent injection, stack opacity would gradually drift downward, and
ESP electrical conditions would return to normal. Approximately 4 to 6 hours were required for
return to the original condition where another sorbent injection test could be conducted.
Fortunately the humidifier installation permitted continuous operations once it was put into regular
service during the summer of 1988. In fact, relatively high humidifier outlet temperatures of 250 -
275°F, i.e., a modest level of humidification, were found adequate to maintain the opacity at
normal 1 to 7% levels.
LIMB Effects on Boiler Operation
The impact of LIMB technology on boiler operation is related primarily to the increased
particulate loading and depends heavily on the adequacy of the sootblowing system. With sorbent
injection in the upper furnace, as was done at Edgewater, LIMB technology results in a higher rate
of ash accumulation on tubes and other available surfaces, roughly in proportion to the additional
particulate matter. Virtually all of the sorbent appeared to follow the gas stream through the
convective pass, rather than collecting as bottom ash. Thus injection at the elevation of the nose
and above avoids the potential for sintering and slagging conditions that would probably have
resulted had the sorbent been introduced at a higher temperature and/or a lower elevation. The
four new steam sootblowers added in the primary superheat area as part of the project were
effective within their range, but did not make up for limitations elsewhere.
The LIMB ash that did accumulate on the tubes was found to be about as easily dislodged
by sootblowing as is normal coal fly ash. The most notable difference was that, instead of
operating on a normal once per shift basis, the sootblowers were run about three times as
frequently, effectively in almost continuous cycles. As one might expect, this strained the capacity
of the compressed air system at Edgewater, to the point that all but the four air heater blowers
were converted to a steam-driven system for the LIMB extension work to follow. This is not to say
that an air system would not work, only that the capacity of the Edgewater system could not keep
up with the heavier duty cycle.
The overall effect of ash accumulation on the tubes was a decrease in heat transfer and
higher flue gas temperatures throughout the boiler. Heat transfer rates through the tubes appeared
to return to normal immediately after being blown, only to begin the cycle once more. Because of
42
-------
this and the relatively low capacity of the air-driven sootblowers, the net result was a flue gas
temperature of about 350°F at the air heater outlet, compared to a more normal temperature of
approximately 300°F under steady state conditions. However, effective sootblowing is expected
to reduce this temperature rise significantly.
The other source of energy loss necessarily associated with LIMB is the excess air used
both for sorbent dispersion in the furnace and water atomization in the humidifier. At Edgewater,
ambient air was used for both, although preheated air, recirculated flue gas, or even mechanically
assisted devices could be used for sorbent dispersion in another installation. The "booster air fan"
arrangement at Edgewater used approximately 30,000 Ib/hr of air to feed 12,000 Ib/hr of sorbent
into 1,000,000 Ib/hr of flue gas. Humidification to a 20°F approach to saturation required about
20,000 Ib/hr of air to atomize 44,000 Ib/hr of water. The atomization system was designed
conservatively with an air compressor motor rated at 1250 HP, to ensure the production of the fine
droplet size required. Close approach to saturation at Edgewater also required flue gas reheat to
restore plume buoyancy. A system coil reheater with the capability of raising the flue gas
temperature 40 °F was installed for this purpose at the outlet of the ESP, just before the induced
draft fan. In applications where only minimal humidification would be required for ESP operation,
the energy penalty associated with high levels of humidification and reheat would be significantly
lower.
Humidifier Operation
The humidification system installed at Edgewater proved to be one which was relatively
problem-free and achieved the purposes for which it was designed. Early mathematical and plastic
flow modelling and quarter-scale simulation tests, all conducted in the early design phase, provided
meaningful insight into the flow patterns and evaporation processes involved. For the most part,
the wet/dry interface problems were limited to relatively small, manageable areas. The most
significant problem was a build-up of material on the lances near some of the atomizers. The
material would build to several inches in length and then fall to the floor of the horizontally
configured chamber where material would accumulate. A modified airfoil lance assembly for the
atomizer array was designed and installed in 3 of the 22 lance locations for the last 3 months of
the program. These modified assemblies had a greatly reduced tendency for deposit formation, and
the complete atomizer array was replaced before continuing with the Coolside process tests.
43
-------
When the humidifier was operated at outlet temperatures greater than 40 °F above the
saturation temperature, essentially no difficulties were encountered. Moreover, the presence of
unreacted quicklime, CaO, in the ash undoubtedly contributed to favorable conditions by virtue of
its exothermic reaction with water. Although the quantity of water vapor present in the flue gas as
a result of combustion was more than sufficient for total rehydration of the CaO, the gas/solid
reaction kinetics appeared to be quite slow when the humidifier was operated considerably above
the saturation temperature. Under this operating condition, much steam evolved when mixing
LIMB ash with liquid water just prior to disposal (described in more detail later). Notably less
steaming occurred when the humidifier was operated at close approach to saturation, suggesting
that, whatever the mechanism(s), the excess CaO has a strong affinity for liquid water.
Deposits formed in the center area of the humidifier outlet turning vanes after extended
operation at a 20°F approach to saturation. The deposits tended to build to a thickness of about 4
in and then sloughed off onto the floor. These deposits are not considered to be a problem
because the chamber was being tested at the extreme limit of its design and several remedial
alternatives could be used in a commercial system. These include not only operation at a slightly
higher temperature, but also intentional capture of large, unevaporated droplets with periodic
deposit collection and removal through a hopper system (see, for example, the description of the
Droplet Impingement Device12). If site-specific spatial limitations permit, a vertical, down-flow
configuration with a hopper for solids collection could be considered for a commercial humidifier
required to operate at close approach temperatures.
Ash Handling and Disposal
LIMB technology also has significant impacts on the power plant's ash handling and
disposal practices. The most obvious of these is the sheer quantity of material that must be
processed. Operation at 2.0 Ca/S at Edgewater with a 10% ash coal almost triples the rate at
which ash must be collected by the ESP and then transferred to the ash silo. As long as lines were
free of obstructions and the transfer equipment was in good working order, the increased quantity
generally did not pose a problem since normal demand and the basic design of the ash handling
system could accommodate such amounts. By the same token, it was important to react to upsets
quickly since there was less time available to make required repairs.
Beyond the quantity of ash itself, the other major considerations regarding LIMB ash
44
-------
handling and disposal stemmed from the quicklime component and the pozzolanic properties of the
ash. At Edgewater, ash is pneumatically conveyed from the ESP hoppers into the ash storage silo.
Even during the relatively short tests in 1987, LIMB ash tended to bridge over at the wet/dry
interface near the aspirating water jets used to create the vacuum. A device designed to ram
through the deposit periodically was procured and installed to overcome this seemingly small, but
critical, problem.
Greater difficulty was encountered due to the amount of steam generated by the
water/quicklime reaction occurring in the beds of the trucks used to haul ash to the landfill. Water
was mixed with dry LIMB ash from the silo in a pug mill that discharged directly into the truck
waiting below. As the level rose in the bed, the steam that resulted soon made it impossible for
the operator to see how much room remained. As noted earlier, this was true especially when
minimal humidification was employed. Although the problem had been anticipated and a large fan
had been mounted to blow or suck the steam away, the severity of the steaming was far beyond
the fan's capability. After considering a number of alternatives including weighing and sonic
techniques, a system was devised that involved the operator's lowering a freely hanging
thermocouple to the desired fill level in the empty truck bed. The thermocouple then read ambient
temperature up until the time it would begin to sense the hot ash. While not regarded as a
permanent "commercial" solution to the problem, it proved to be an expedient, practical remedy for
the purposes of the demonstration. In time, a "reference" thermocouple was added in the
discharge chute of the pug mill, since some variation in temperature did occur, typically in the 150
to 250°F range, depending on the stoichiometry and degree of humidification.
For future commercial systems, a more sophisticated design including those based on
gravimetric, sonic, or even spectrophotometric techniques is envisioned, if treatment is restricted to
an ash unloading system such as exists at Edgewater. Assuming that rehydration of the quicklime
component is the preferred method of ash disposal, another approach is possible where there are
pre-existing facilities, such as a pug mill and radial stacker at the disposal site, which can readily
accommodate steam from the ash. The steam emanating from the surface of the ash presents no
sustained problem in that it subsides to faint wisps within about 15 min.
Before the steaming problems were brought under control, they gave rise to a further
complication that had been underestimated. While underfilling the truck was the norm, on a couple
of occasions overfills did occur that had to be picked up with a front end loader. The area then
45
-------
was washed down as a final cleanup measure. When this was done, the lime-rich ash would raise
the pH of water draining toward the plant's ash pond. While provisions had been made for sensing
and neutralizing high pH water that could result from a failure in the pneumatic ash conveying
system and from minor yard spills of lime or ash, the amount that spilled from overfilling the truck
had not been anticipated. Moreover, sumps that accumulate significant quantities of lime or LIMB
ash due to low flow or agitator failure could also produce undesirably high pH. The solution to all
of these related difficulties was rerouting all of the potential sources through the neutralization
system.
Most of the LIMB ash from Edgewater was placed in Ohio Edison's ash disposal site,
adjacent to the normal ash disposal area. Despite apprehensions that the cementitious properties
of the ash might cause some difficulty with the bulldozers, no significant problems were
encountered, probably because the pozzolanic reactions did not proceed to any appreciable extent
at the relatively low water/ash ratios used to produce a material that can be readily dumped from
the truck. In addition to the routine LIMB ash disposal, approximately 140 truckloads were placed
in two test cells in an isolated area of the disposal site for study over the next several years under
a related DOE-sponsored program.11
46
-------
SECTION 7
LIMB PERFORMANCE EVALUATION
EXPERIMENTAL PROCEDURES
The collection and reduction of large amounts of data to be obtained during the project was
an early concern that called for a state-of-the-art computer system that would permit on-line
analysis of system performance. This was accomplished with the system described in the next
paragraph, except for the initial burner and baseline tests conducted while the appropriate
programs were being written. The basic equations used in data reduction and analysis were the
same, however, regardless of whether the calculations were performed manually or automatically.
The overall performance of the total system - boiler, sorbent injection system, and Radian
Corporation's Continuous Emission Monitoring System (CEMS) - is monitored by the Babcock &
Wilcox Boiler Performance Diagnostics System 140™. Since the data presented in this report are
taken from the output of this device, the following few paragraphs are provided to help the reader
understand how the data were collected and handled by the System 140™. This computerized
data acquisition system was expanded beyond its normal boiler performance function to a
customized device capable of monitoring the additional equipment and analyzers associated with
the LIMB technology. It was also programmed to perform a variety of calculations specific to the
technology, such as Ca/S stoichiometry and momentum flux ratios, as well as to reduce the data
from the CEMS according to EPA-accepted procedures.
The input data for the System 140™ consisted of approximately 700 temperature, pressure,
flow, and gas concentration signals that were automatically read once a minute, together with
manually entered coal, sorbent, and ash analyses that are updated as necessary. The system was
later upgraded to accept humidifier and pulverizer data and ESP voltages and currents in order that
the System 140™ could serve as the primary data recording device for the project.
As each set of readings is taken, the System 140™ performs about 600 calculations that
may be grouped into categories related to boiler performance, boiler cleanliness factors, sorbent
47
-------
injection parameters, humidification parameters, and emission data. As time progresses, ten
minute rolling averages of the input and calculated data are calculated for display in data lists
and/or 2, 8, 24 or 168 hour trend charts as appropriate. In many cases the most recent individual
input data are similarly displayed. In addition, hard copies of the data lists and trend charts may be
obtained as needed.
The System 140™ also performs several other functions associated with on-line error
analysis, the redetermination of the LIMB system's performance based upon recalculation of the
stored ten minute averages with updated, ultimate coal and sorbent analyses received several days
after the fact, and x-y plotting of trended values. The recalculation feature was included such that
the ultimate coal analyses could be entered at a later date when the analysis was complete and to
provide an opportunity to calculate a thirty day rolling average of pertinent variables.
Since much of the emphasis of this report is directed primarily at the S02 removal efficiency
of the LIMB system as a function of the Ca/S stoichiometry, the calculations of particular
importance are those associated with the determination of coal firing rate and "inlet" S02, sorbent
injection rate, and S02 emissions. These are described individually in the following paragraphs.
Coal Firing Rate and "Inlet" SO-
It was decided early in the project that the coal feeders at the Edgewater Station were of a
type and vintage that would not readily provide the desired accuracy in feed rate. The
determination of coal firing rate by the System 140™ was considered to be more reliable in that it
calculates the rate based on the coal analysis and a series of heat and material balances derived
from the temperatures, pressures, and flows measured throughout the boiler. While the exact
equations within the System 140™ are derived from the original design of the Edgewater boiler and
are considered proprietary, the pertinent portions here are quite similar to ASME PTC4.1, ASME
Heat Loss Method, a standard calculation of boiler efficiency in which fuel rate is also determined.
The product of the coal firing rate and the sulfur content of the coal is the basis for the sulfur term
in the Ca/S stoichiometric ratio. The ratio of coal sulfur to its as-fired, high heating value (HHV)
allows calculation of the "inlet" S02 on a lb/108 Btu basis according to the equation:
48
-------
Inlet SO lb/106 Btu = 'Coal S twt %I. lb) (1x10e Btu/106 Btu) (64.062 lb SO.)
2' (100 lb Coal) (HHV, Btu/lb coal) (32.064 lb S)
Sorbent Injection Rate
Lime is fed from either of two Acrison differential weight loss feeders to any two of three
elevations in the upper furnace. The feed rates output from the two feeders are summed by the
System 140™ to give the total sorbent injection rate. The product of this value and the Ca(OH)2
content of the lime determined by analysis is the basis for the calcium term in the molar Ca/S
stoichiometric ratio calculated as:
Ca/S = (Sorbent. Ib/hr) (Sorbent Ca(OH). fwt %1. lb) (32.064 lb/lb mol S) (100 lb Coal)
(Coal, Ib/hr) (Coal S [wt %], lb) (74.1 lb/lb mol Ca(OH)2) (100 lb Sorbent)
It is noted that the feed silo from which the lime is fed to the Acrison feeders is equipped
with load cells which permit an independent measure of the feed rate. During the shakedown
period of July 7 to September 4, 1987, weights were separately recorded as a function of time to
verify the feed rates indicated by the System 140™. This data, concurrent "spikes" to lower S02
emissions, and other visual observations made by those operating the feeders, made it possible to
discriminate between a controlled lime feed rate and one where large excesses flooded past the
feed screw.
SO- and NO„ Emissions
Radian Corporation's CEMS continuously analyzes several gases in the duct between the
induced draft fan, following the ESP, and the stack. The sampling point was confirmed to be a
representative location according to EPA quality criteria as defined in 40 CFR, Part 60, Appendix A,
Method 1 of the Federal Register. An extractive system is used in which calibration gases have
been sent through the whole train except for the initial filter in the duct, especially to verify that
there is no appreciable S02 capture within the sampling/analytical system itself. The filter in the
duct is maintained at high (approximately 300°F) temperature by the flue gas where no significant
S02 capture occurs. The sampled flue gas is cooled to condense moisture, and split among the
various analyzers which include S02, NOx, 02 (oxygen), C02, (carbon dioxide), CO (carbon
49
-------
monoxide), and total hydrocarbons (analyzer output calculated as propane in air). The total
hydrocarbons were almost continuously below detection limits for several months and were
therefore no longer monitored later in the project.
Of interest here, S02 and N0X (each parts per million by volume [ppmv]) and 02 (vol %)
values are read every minute by the System 140™. The S02 and NOx concentrations are then
converted to a lb/106 Btu basis using the standard "F-factor" equation and the measured 02
concentration to correct for dilution. The difference between this outlet S02 value and the "inlet"
value described earlier represents the removal which, when divided by the "inlet" S02> gives rise to
the removal efficiency calculated by the System 140™ using the most recent ten minute rolling
average values at any point in time. The equations used in the calculations are:
Fd, Dry Standard Cubic Feet (dscf) Flue Gas/10® Btu =
.1x108 Btu, ,(3.64 x H) + (1.53 x C> +(0.57 x S) + (0.14 x N> - 0.46 O) dscf Flue Gas/lb Coal,
10® Btu 1 Coal HHV, Btu/lb '
where H, C, S, N, and 0 are the weight percentages of these elements and HHV is
the high heating value of the coal on an "as received" basis.
Outlet S02/ lb/10® Btu =
(S02, ppmv dry) 'Fd' dscf/108 Btu) (1.66x10'7 lb S02/(ppmv dscf))
' vj
where 02 is the volume percentage of oxygen in the flue gas measured at the same
location where the corresponding S02 concentration (ppmv) is determined. All
outlet gas concentrations were determined on a dry basis at the stack, and on a
wet basis elsewhere (individual 02 at the economizer outlet, air heater outlet, and
humidifier inlet and outlet; S02, C02, and H20 at the humidifier inlet/outlet where a
dilution probe sampling and analytical system alternated between inlet and outlet
sample gas. When appropriate for further calculations, gas concentrations
measured on a wet basis were corrected to a dry basis, as for example:
S02, ppmv dry = (S02, ppmv wet) ( ... ^ >
100 - H20, vol %
50
-------
where H20 is the volume percentage of water assumed or measured at the sample
location.
Similarly,
Outlet NOx as N02, lb/10® Btu =
(NOx, ppmv dry) (2° ^P' % 0?) (Fd, dscf/108 Btu) (1.194x10"7 lb N02/(ppmv dscf))
ZUiv " Uj
The S02 removal efficiency is then given by:
S02 Removal Efficiency, % = 100 % (lnlet SO, - Outlet SQ->, lb/108 Btu)
2 Inlet S02, lb/106 Btu
One final area that requires explanation before presentation of the results is the method of
establishing the "baseline" or zero removal efficiency for each test. Since the high temperature
precluded continuous measurement of S02 concentration upstream of the lime injection point(s) in
the furnace, the method of calculating an "inlet" value as described previously was selected. This
was considered the only reasonable, practical method for the long term demonstration, during
which pulverized coal is automatically sampled daily from four burner coal pipes just before entering
the furnace.
While it was reasonable to program the System 140™ for recalculation on a daily basis, it
would have been extraordinarily complicated to have made it possible to make recalculation
infinitely variable over short injection periods such as those characteristic of the tests here. What
was done instead was to adjust the coal sulfur such that the calculated S02 removal efficiency was
near zero (typically between -5 and 5% during the 1987 tests and between -2.5 and 2.5% during
the 1988-89 tests) at the beginning of the test. After lime injection was discontinued, during the
1987 tests the S02 removal efficiency would return to approximately the same zero after a few
hours and the average of the two values was subtracted from the removal efficiency of the test
period to yield the efficiency reported herein. Given the generally longer test periods of the
1988-89 tests, the baseline "zero" efficiency was determined from the "before" and "after" values
judged to be most reliable in light of on-site sulfur analyses and checks on analyzer drift. It was
likewise theoretically necessary to correct the sulfur term in the Ca/S stoichiometry for the nonzero
51
-------
baseline S02 removal efficiency, NZB (%), value according to:
Corrected Ca/S = Original C?/S
(1 - (NZB/100 %))
While this was done for the 1987 data, the correction was not considered worth the additional
effort, given the tighter ±2.5% criterion applied during the 1988-89 tests.
The validity of manipulating the data in this manner is most strongly dependent upon the
invariability of the coal sulfur and lime Ca(OH)2 contents. Several observations are made along
these lines. For the long term demonstration a Leco sulfur analyzer was purchased for frequent on-
site determination of the approximate sulfur content of the coal sampled from the burner pipes.
This instrument was used in the course of the tests and typically provided very good agreement
between the value analyzed and that selected to "zero" the baseline. Further reasonable agreement
was found with the sulfur content of burner coal pipe samples determined in the ultimate analysis
of the coal by Commercial Testing & Engineering Company (CTECo). Finally, the relatively
invariable trace of S02 concentration under steady boiler and combustion air (that is, 02
concentration) conditions, both during and outside of test periods, suggest that coal sulfur did not
change radically for significant lengths of time in the course of the tests.
The coal analyses associated with the project are summarized in Tables 1 to 3. These are
the "truck" and "bunker" samples collected and analyzed by Ohio Edison as part of their normal
analytical procedures. In accordance with a statistically designed sampling program, the truck
samples are collected with a sampling auger and the bunker samples with a cross-cut coal stream
sampler. The correspondence between the averages of the truck and bunker samples is excellent.
In order to have as invariable a composition as possible, Ohio Edison unloaded trucks at the rear of
the coal pile and then bulldozed the coal forward through the pile. This tended to blend the coals
together and lowered the variability of the coal loaded to the bunker. This lower variability is
manifested in the lower standard deviation of the bunker samples.
Table 3 contains the results of analyses performed on pulverized coal samples obtained
from the piping exiting each of the four pulverizers at Edgewater. An automatic pulverized coal
sampler was used to collect the coal from the pipes coming from each of the four pulverizers. The
samples were generally obtained daily through the 5-day work week. Composite samples were
52
-------
TABLE 1. SUMMARY OF HIGH SULFUR COAL ANALYSES - TRUCK SAMPLES (Ohio Edison Analyses)
Dry Basis
Mois-
ture
and
Aah
Free
Vola-
Heat-
Heat-
tile
Fixed
ing
ing
Sample
Time
Matter
Carbon
Ash
Sulfur
Value
Value
Ash Index
S02 Index
Typo
Period
wt %
wt %
wt %
wt %
Btu/lb
Btu/lb
lb/10" Btu
lb/10s Btu
Coal Y
06/27/86
Avar age
37.64
49.04
13.13
3.23
12818
14766
10.26
6.04
7,129
ton
to
Standard deviation x 2
2.34
2.42
2.28
0.63
412
320
2.04
1.12
06/04/86
No. of analysss
16
16
16
23
23
23
23
23
Coal Y
06/26/87
Average
NA"
NA
13.39
2.89
12707
14671
10.66
4.66
16,811
ton
to
Standard deviation x 2
NA
NA
2.37
0.66
417
176
2.20
1.16
08/11/87
No. of analysee
NA
NA
36
36
36
36
36
36
Coal A
08/08/88
Average
NA
NA
7.67
3.03
13389
14486
6.66
4.62
18,161
ton
to
Standard deviation x 2
NA
NA
1.87
0.64
338
216
1.63
0.81
06/09/89
No. of analyses
NA
NA
123
123
123
123
123
123
Coal B
08/08/88
Average
NA
NA
10.70
3.01
13081
14860
8.24
4.60
64,666
ton
to
Standard deviation x 2
NA
NA
6.02
0.84
729
303
4.62
1.38
06/09/89
No. of enelyses
NA
NA
220
220
220
220
220
220
Coal C
08/08/88
Average
NA
NA
11.76
3.19
12799
14603
9.20
4.99
42,418
ton
to
Standard deviation x 2
NA
NA
2.66
1.21
484
317
2.41
2.01
06/09/89
No. of analyse*
NA
NA
188
188
188
188
188
188
Coals
08/08/88
Average
NA
NA
10.36
3.08
13063
14660
7.98
4.72
A + B + C
116,244
to
Standard deviation x 2
NA
NA
4.89
0.96
726
328
4.27
1.69
ton
06/09/89
No. of analyses
NA
NA
631
631
631
631
631
631
" NA - Not snalyzed.
53
-------
TABLE 2. SUMMARY OF HIGH SULFUR COAL ANALYSES - BUNKER SAMPLES
Dry Beeie
Drv E
Semple
Type
Time
Period
Vole-
tile
Metter
wt %
Fixed
Cerbon
wt %
Aeh
wt %
Sulfur
wt %
Heet-
ing
Velue
Btu/lb
Moie-
ture
end
Aah
Free
Heet-
ing
Velue
Btu/lb
Aeh Index
lb/104 Btu
SO, Index
lb/104 Btu
Cerbon
wt %
Hydro-
wt %
Nitro-
0«n
wt %
Ohio- Oxy-
rine gen
wt % wt %
Bunker
(Ohio
Edieon)
cn
4*
06/27/86
to
06/06/B6
Averege
Stenderd devietion x 2
No. of enelyeee
NA
NA
NA
NA
NA
NA
13.13
2.87
12
3.12 12788 14734
1.21 316
12 12
210
12
10.27
2.44
12
4.87
1.96
12
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Bunker
(Ohio
Edieon)
08/06/87
to
08/26/87
Averege
Stenderd devietion x 2
No. of enelyeee
NA NA 13.08 2.76 12647 14437
NA NA 1.26 0.26 670 717
NA NA 17 17 17 17
10.43
1.00
17
4.38
0.40
17
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Bunker
(Ohio
Edieon)
06/08/88
Averege
Stenderd devietion x 2
No. of enelyeee
NA
NA
NA
NA
NA
NA
11.30
2.66
168
3.08 12800
0.68 380
168 168
14643
266
168
8.77
2.28
168
4.77
0.98
168
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
* NA « Not enelyzed.
-------
TABLE 3. SUMMARY OF HIGH SULFUR COAL ANALYSES - PULVERIZED (BURNER PIPE) SAMPLES
Dry Bee is
Dry Bee m
Sample
Type
Time
Period
Vole-
tile Fixed
Matter Carbon
wt % wt %
Aah
wt %
Sulfur
wt %
Value
Btu/lb
Mois-
ture
end
Aeh
Free
Heat-
ing
Value
Btu/lb
Aeh Index
lb/100 Btu
SO, Index
lb/10* Btu
Cerbon
wt %
Hydro-
wt %
Nitro-
oen
wt %
CMo- Oxy-
rine gen
wt % wt %
Pulverized
(CTECo)
06/27/86
to
06/06/86
Averege
Standard deviation x 2
No. of analyeee
37.06
0.91
6
60.26
1.84
6
12.68
1.10
6
2.99 12941
0.40 196
11 6
14820
69
6
9.80
1.00
6
4.76
0.79
6
72.06
1.20
6
4.63
0.09
6
1.39
0.11
11
0.24
0.06
6
6.91
0.66
6
CJ»
CJ1
Pulverized
(CTECo)
09/08/87
to
09/26/87
Averege
Standard deviation x 2
No. of analyeee
36.37 60.28 13.36 2.44 12692 14648
1.97 2.39 1.48 0.17 272 111
17 17 17 17 17 17
10.63
1.38
17
3.86
0.36
17
71.42
1.63
17
4.69
0.21
17
1.41
0.08
17
0.17
0.06
17
6.66
1.00
17
Pulverized
(CTECo)
09/21/88
to
06/09/89
Average
Standard devietion x 2
No. of analyeee
37.24
3.18
118
61.80
2.19
118
10.98
2.46
118
2.68 13027
0.36 327
118 118
14636
236
118
8.44
2.06
118
4.11
0.63
118
72.38
1.69
118
4.89
0.27
118
1.48 NA
0.16 NA
118
NA
7.60
1.21
118
Pulverized
(Radian
- Leco
analyzer)*
09/21/88
to
06/09/89
Average
Standard devietion x 2
No. of analyeee
NA
NA
NA
NA
NA
NA
NA
NA
NA
2.73
0.33
118
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
" NA - Not enelyzed.
* Repreeente the veluee obtained with convereion to a dry baeie ueing the moisture determined by CTECo in the correeponding ultimate analysis.
-------
sent to CTECo for the ultimate analyses, while 4 to 6 samples were typically analyzed with the
on-site Leco sulfur analyzer periodically throughout the day. The lower sulfur content in the
ultimate and Leco analyses, as compared to the bunker analyses, is believed to be due to the
removal of pyritic sulfur in the "pyrite traps" built into the design of the coal pulverizers. No credit
was taken for this sulfur removal, however, since the calculations were all performed on an "as-
fired" basis (recall that pulverized coal samples collected for the ultimate and Leco analyses were
collected downstream of the pyrite traps, and that "baseline" values were obtained from the S02
concentrations resulting from the coal actually burned).
As shown in Table 4, lime quality was very consistent with an average value of 92.1 %
Ca(OH)2 for both the commercial and lignosulfonated limes. Values of 92.4 and 91.3%,
respectively, were used in the System 140™ for the two sorbents during the course of the
program. In preparing this report, no corrections were made to the Ca/S stoichiometries for such
small differences in purity.
TABLE 4. SUMMARY OF LIME ANALYSES
Commercial Calcific
Hydrated Lime
Ugnosutfonate-Doped
Calcific
Hydrated Lime
Dates of
analyses
From To
Weight
Represented
Ton
Number
of
Analyse*
Average
Available
Ca(OH}3
wt %
Standard
Deviation
x 2
Average
Available
Ca(OH),
wt %
Standard
Deviation
x 2
00/10/87
09/18/87
602
7
92.66
0.21
NA*
NA
09/22/87
09/26/87
109
4
NA
NA
92.13
0.36
10/26/88
03/07/89
6260
28
92.07
1.16
NA
NA
03/14/89
03/23/89
731
14
NA
NA
92.01
0.39
03/27/89
03/30/89
608
4
92.02
0.37
NA
NA
04/14/89
06/19/89
942
11
NA
NA
92.13
0.62
06/23/89
06/09/89
790
12
91.96
0.38
NA
NA
Overall
09/10/87
06/09/89
7169
61
92.10
0.96
NA
NA
09/22/87
06/19/89
1783
29
NA
NA
92.07
0.46
* NA = Not applicable
56
-------
BASELINE TESTS WITHOUT SORBENT INJECTION
A number of tests were conducted prior to full operation of the sorbent injection
components of the LIMB system in order to establish baseline conditions for the burners in
particular, as well as for performance of the unit as a whole. Specific characterization tests with
the original circular burners were carried out during the period of May 13 to 15, 1986, and were
followed by more general baseline tests of May 27 to June 6, 1986. The plant's "normal"
compliance coal containing a nominal 1.6% sulfur and 1.4% nitrogen was used during the burner
characterization tests, while the baseline tests were run while burning a coal with about the same
nitrogen, but higher sulfur content (see Tables 1 to 3). The DRB-XCL1" burner characterization
tests were conducted during the period of July 13 to 16, 1987 following an extended start-
up/shakedown and optimization period for these new burners. It is also noted that the System
140™ was first put in service for the DRB-XCL™ burner characterization tests. The results obtained
during the three test periods, each grouped according to high (-105 MWe), medium (-80 MWe),
or low (-60 MWe) load, are summarized in Tables 5, 6, and 7. The load conditions represent peak
and full loads of-770,000 and 700,000 lb steam/hr, respectively, a mid-load of -610,000 Ib/hr,
and a low load of -500,000 Ib/hr, the so-called control load which represents the lowest practical
load at which one can operate with all burners in service. It is noted that the unit can and does
operate at loads down to as low as -300,000 Ib/hr (-31 MWe) by taking some burners out of
service. Such conditions were not included in these specific test plans since they would have
generated a relatively large variety of flow and temperature profiles that would have made
interpretation of results extremely complex. The conditions are represented in later results,
however, since the system was operated in a load-following mode outside of rigorous test periods
within the main demonstration program itself.
Each of the three test periods is discussed individually in the following sections. Data is
generally presented in summary form since the voluminous quantity collected throughout the
demonstration makes it impractical to include all in one report. For this reason the discussion
occasionally refers to measurements made in support of the primary goals of an individual set of
tests in order to provide the reader with an understanding of the scope of the work done, without
specifically providing each and every detail. This is particularly true for the baseline and burner
characterization tests where considerable effort was initially expended on 02 measurements to
assure proper, representative operation of the boiler.
57
-------
TABLES. CIRCULAR BURNER BASELINE AND CHARACTERIZATION TESTS SUMMARY - SOa AND NOx EMISSIONS
Economizer Outlet 02
vol % wot
Air
Mein
Furnace
Heater
Air
Grooo
Steam
Exit
Inlet
Heater
Stack
Unit
Flow
oa
oa
Infot
oa
Stack
Stack
Carbon in
Toot
Stort
Load
Ib/hr
vol %
vol %
NO,
vol %
NO.
SO,
Ash
No.
Timo
Date
MWo
x 10-®
dry
Eaot
West
Averege
dry
lb/10fl Btu
dry
fb/10" Btu
lb/10* Btu
wt %
Characterization Toots
BC02
18:30
6/13/86
106
770
NA'
2.3
8.6
6.6
4.8
0.89
NA
NA
NA
1.90
BC04
10:06
6/14/86
103
766
NA
1.6
6.2
3.9
3.6
0.69
NA
NA
NA
3.62
BC08
10:30
6/16/86
98
777
3.0
1.8
4.4
3.1
4.4
0.99
6.6
0.88
NA
2.48
BC01
16:16
6/13/86
84
610
NA
2.0
8.6
6.3
4.9
0.76
NA
NA
NA
2.63
BC03
08:16
6/14/86
84
610
NA
1.4
6.0
3.2
3.6
0.61
4.7
0.60
NA
3.60
BC09
14:00
6/16/86
79
608
3.6
2.6
6.7
4.6
6.0
0.94
6.1
0.84
NA
1.77
BC06
12:10
6/14/86
66
496
NA
2.0
6.8
3.9
3.6
0.69
NA
NA
NA
3.82
BC06
14:60
6/14/86
66
496
3.9
3.0
8.3
6.7
6.1
0.71
8.0
0.79
NA
1.36
BC07
07:46
6/16/86
62
600
NA
1.2
6.7
3.6
3.8
0.67
6.0
0.64
NA
4.60
Boooiino Toots
BL01
08:00
6/27/86
102
800
NA
2.0
7.7
4.9
4.60
0.81
6.32
0.79
4.40
NA
BL04
08:30
6/30/86
103
761
NA
2.4
6.6
4.6
4.46
0.88
6.41
0.94
4.80
NA
BL06
07:00
6/03/86
101
804
NA
2.7
6.8
4.2
4.46
0.82
6.24
0.82
4.84
NA
BL03
08:30
6/28/86
82
646
NA
2.9
6.2
4.6
4.71
0.73
6.33
0.78
6.09
NA
BLOB
07:00
6/06/86
82
636
NA
2.8
6.2
4.6
4.82
0.69
6.04
0.81
6.20
NA
BL09
07:00
6/06/86
81
636
NA
2.6
6.2
4.4
4.62
0.71
6.03
0.81
4.93
NA
BL02
07:00
6/28/86
64
493
NA
2.8
6.9
4.4
4.B6
0.60
6.60
0.64
4.86
NA
BL06
07:30
6/02/86
66
618
NA
3.4
6.8
6.1
4.61
0.69
6.73
0.73
6.00
NA
BL07
07:00
6/04/86
66
497
NA
2.6
6.3
4.4
4.49
0.66
6.61
0.66
4.76
NA
" NA - Not analyzed.
-------
TABLE 6. XCL BURNER CHARACTERIZATION TESTS SUMMARY - SO, AND NOx EMISSIONS
Economizer Outlet Oa
vol % wet
Air
Main
Furnace
Heater
Air
Groea
Steam
Exit
Inlet
Heater
Stack
Stack
Unit
Flow
oa
oa
Inlet
Oa
Stack
Stack
Carbon
CO
Teat
Start
Load
Ib/hr
vol %
vol %
NO,
vol %
NO„
SO,
in Aeh
ppmv
No.
Time
Date
MWe
x 10 3
dry
Eaat
Weet
Averaoe
dry
lb/10" Btu
dry
lb/10* Btu
lb/106 Btu
wt %
dry
Characterization Teata
BT02
16:00
7/13/87
91
680
NA*
0.4
3.0
1.7
4.60
0.42
7.1
0.42
NA
10.61
746
BT06
11:30
7/14/87
92
682
NA
2.0
3.0
2.6
6.16
0.44
7.7
0.47
NA
7.86
64
BT08
14:16
7/16/87
93
680
NA
2.6
6.0
4.3
6.06
0.66
8.7
0.67
NA
3.96
16
BT11
10:26
7/16/87
93
676
NA
3.0
6.0
4.0
6.10
0.69
8.6
0.61
NA
3.78
13
BT03
19:60
7/13/87
79
680
NA
2.0
4.0
3.0
6.30
0.42
7.7
0.47
NA
6.68
68
BT04
0B:00
7/14/87
80
680
NA
3.0
4.0
3.6
6.20
0.48
8.4
0.62
NA
4.08
13
BT09
17:40
7/16/87
80
670
NA
2.0
6.2
3.6
4.60
0.37
7.4
0.40
NA
12.96
386
BT01
11:30
7/13/87
64
460
NA
3.2
6.0
4.6
8.20
0.44
8.7
0.43
NA
3.96
18
BT06
16:64
7/14/87
64
468
NA
1.6
3.6
2.6
4.76
0.30
7.2
0.32
NA
12.97
638
BT07
10:20
7/16/87
66
462
NA
2.6
2.6
2.6
6.30
0.38
7.6
0.40
NA
8.04
83
BT10
07:30
7/16/87
66
466
NA
2.0
4.0
3.0
6.10
0.41
7.6
0.41
NA
10.96
81
* NA - Not analyzed.
-------
TABLE 7. CIRCULAR BURNER BASELINE TESTS SUMMARY • OTHER PARAMETERS
Teat
No.
Particulate Loadina
ESP
Removal
Efficiency
%
Fly Aah Maaa
Mean Diameter
ESP Inlet Aah Reeiativity
Maaaurementa and Condition
Other Stack
(ESP Outlet) Emiaeione
ESP
Inlet
gr/dacf
ESP ESP
Outlet Outlet
gr/decf lb/108 Btu
ESP
Inlet
//m
ESP
Outlet
fjm
Reeiativity
ohm cm
Temperature
°F
Rue
Oaa
HaO
vol %
CO,
vol %
dry
CO
ppmv
dry
THC1
ppmv
dry
Baeelirte Teata
BL01
2.84
0.029
0.040
99.0
na"
NA
NA
NA
NA
12.8
16
0
BL04
3.66
0.069
0.083
98.3
39
NA
NA
NA
NA
13.7
17
0
BL04*
3.49
0.040
0.066
99.8
NA
NA
NA
NA
NA
NA
NA
NA
BL06
4.16
0.067
0.093
98.3
64
6.6
1.9E+11
309
7.3
13.6
16
NA
BL03
3.79
0.017
0.024
99.6
91
2.8
NA
NA
NA
13.0
14
1
BL08
3.19
0.02B
0.039
99.1
48
NA
1.7E+11
306
8.4
12.3
16
NA
BL09
3.29
0.028
0.039
99.2
NA
NA
NA
NA
NA
12.7
18
NA
BL09*
3.73
0.017
0.024
99.6
NA
NA
NA
NA
NA
NA
NA
NA
BL02
3.66
0.017
0.023
99.6
64
NA
NA
NA
NA
13.3
16
0
BLOB
3.41
0.022
0.031
99.3
37
2.1
NA
NA
NA
12.8
16
NA
BLOB1
3.66
0.019
0.027
99.4
NA
NA
NA
NA
NA
NA
NA
NA
BL07
3.14
0.036
0.060
98.8
46
NA
6.6E+10
290
7.9
12.9
18
NA
" NA - Not enelyzod/meeaured.
f THC - Total hydrocarbon* ae propane.
* Replicate particulate loading meaeurementa war* mada during teata BL04, BL06, and BL09.
-------
Circular Burner Characterization Tests
The primary objective of the characterization tests performed on the existing B&W circular
burners was to confirm that the burners were operating according to normal commercial standards.
Specifically, the tests sought to confirm that the required quantity of combustion air was available
in the furnace, and that the air was properly distributed within the windbox. Beyond this, the tests
were also intended to provide some baseline NOx emission data for the circular burners.
Before the tests actually began, the circular burners were inspected to assess their
mechanical condition. No broken or bound register linkages were found and all of the registers
moved freely. Overall, the burners were in excellent working order, and only the packing seal
between the burner and the furnace wall had to be replaced before the start of testing.
B&W also adjusted the burners based on visual observations and 02 traverses. All impellers were
initially set at the zero position with the back of impeller flush with the end of the coal nozzle.
Registers were positioned to provide short, bright flames attached to the throat. The burner
settings resulted in a flame length of approximately 15-18 feet.
In order to ascertain the existing windbox air distribution, a point-by-point 02 traverse was
performed with the eight-point gas sampling grid at the inlet of the air heater. Based on
experience, it was assumed that the unit's gas flow was stratified throughout the boiler as shown
in Figure 13. The grid was arranged to draw gas samples at points 3/4 and 1/4 deep into the air
heater in order to provide 02 measurements relating to upper burners and lower burners,
respectively. Four probes were placed across the furnace width to provide side-to-side
measurements.
The first series of point-by-point 02 data, taken just before beginning test BC01, showed a
slight bias of air to the upper east corner of the windbox. This was further confirmed by the
observation of less stable and less bright flames in the lower burners relative to the upper burners.
Burner adjustments were made to balance the air by biasing burner register settings from nearly
closed on the upper east corner to near-full open on the lower west side. This resulted in better
distribution, and was confirmed by visual observation of brighter lower burner flames. Once the
general windbox air distribution was understood, the tests proceeded as planned. The point-by-
point 02 traverse using the sampling grid at the air heater inlet was repeated just before tests BC03
and BC07 to reconfirm the air distribution.
61
-------
Figure 13. Furnace gas flow pattern initially assumed.
62
-------
For tests BC06, BC08, and BC09, fourteen-point 02 traverses were performed at the
furnace exit (defined as the plane where the flue gas enters the pendant secondary superheater
bank) to document the operating 02 level of the furnace. The analysis of the data was performed
by averaging all of the traverse points for each of these three tests to determine an average
furnace outlet 02 level of 3.5 vol %. This compared with an average air heater inlet 02 level of 4.8
vol %, indicating air leakage into the convection pass not uncommon in older units such as this.
During the tests the boiler operator used the numeric average of the two economizer outlet
02 readings to establish the control set points. Despite the relatively large differences between the
east side and west side readings, the 4.3 vol % 02 average obtained over all nine tests was the
same as that obtained from the eight-point composite sampling grid at the air heater outlet. As can
be seen in Table 5, the NOx emissions calculated at the air heater inlet using the 02 data at the
same point were comparable to those measured at the stack where the data were subjected to the
more rigorous quality control criteria. The corresponding
combustion efficiency of the circular burners, as reflected in the weight percent carbon in the ash,
is also included in the table.
Overall, the results of the burner characterization tests show that the Unit 4 boiler operated
within normal commercial standards, and combustion performance was as would be expected from
any similar unit. The furnace combustion zone operated at approximately 1.3% 02 lower than the
economizer outlet, and produced a full load NOx emission level of about 0.85 lb/10® Btu with
approximately 2.2 wt % carbon in the ash.
Boiler Performance Baseline Tests
Baseline tests to determine the operating characteristics of the boiler prior to installation of
the LIMB system were conducted between May 27 and June 6, 1986. Data were obtained at 30
minute intervals over 10 hr periods on 9 test days ~ 3 days each at gross generation loads of
approximately 102, 82, and 65 MWe. All tests were performed with all 12 circular burners in
service. The data was then averaged in groups of 5 - 30 minute tests and subsequently loaded
into B&W's P140 boiler design program for analysis.
Test Procedures-
The majority of the boiler performance data was obtained from the panel board in the main
63
-------
control room. Additional data points were recorded using the following methods:
• The superheat and reheat flow was measured using a flow transmitter. A flow transmitter
uses a pitot tube to record a pressure differential across an orifice and converts it to a 4 to
20 mA signal which was recorded and converted to an equivalent flow.
• A multipoint oxygen sampling grid was installed at the air heater inlet to provide continuous
monitoring of the oxygen in the flue gas. A grid of thermocouples was inserted at the same
location as the gas probes to measure flue gas temperature. The sampling grid provided the
average oxygen and temperature measurements of the gas passing through the cross
sectional area of the flue.
• Radian Corporation's Continuous Emission Monitoring System (CEMS) was used for
measurement of the S02, NOx( 02, CO, total hydrocarbons (THC), and C02 at the stack.
Analysis of Boiler Performance--
The data obtained, together with pertinent data on the geometry of the unit and original
design information, were input into a Hewlett-Packard HP9845 computer where they were used to
calculate heating surface cleanliness factors (Kf's). Kf's have been and are used by B&W to assess
boiler performance and to account for deviations from expected heat transfer rate due to ash
deposits, differences in fuels, and actual versus theoretical operation. The original design assumed
Kf's of 1.0 for all heating surfaces. Kf's above 1.0 indicate absorption is greater than expected,
while values below 1.0 indicate less. The Kf's calculated for the various absorption banks from the
test data were:
Boiler
1.00
Secondary Superheater
1.11
Pendent Reheater
1.15
Primary Superheater
1.35
Economizer
1.01
The Kf's were compared for various tests and found to be in close agreement, which
indicates good overall precision of the test data. In addition, they all fall within the normal range of
values B&W has calculated from data on comparable operating units.
The model used to calculate the Kf's includes the air infiltration through the convection
64
-------
pass that was measured during the circular burner characterization tests. Air in-leakage
represented by the 1.3 vol % 02 increase determined from 02 traverses at the furnace exit and the
air heater inlet. This is equivalent to 6.2% excess air infiltration through the convection pass. In
the development of the computer model for the Kf study, it was assumed that 1 % excess air
leaked in before the economizer, and that the remaining 5.2% leakage occurred at the junction of
the convection steam cooled enclosure and the economizer enclosure. Although coal samples were
taken throughout the testing, the fuel analysis of June 3, 1986 was used in the Kf study. It was
most representative of the samples analyzed and reduced the computational time considerably.
S02 and NOx Emissions-
CEMS data for S02 and NOx emissions at the stack are summarized in Table 5. Average
NOx emissions ranged from 0.67 lb/108 Btu at the low boiler load tested (64 MWe) to 0.86 lb/106
Btu at a load of 102 MWe. The S02 emissions for the nine test runs ranged from 4.40 to 5.20
lb/10® Btu and averaged 4.87 lb/106 Btu for the baseline test period. Analyses of the pulverized
coal samples (Table 3) collected during the test period showed an average sulfur content of 2.99
wt % and an average heating value of 12,941 Btu/lb on a dry basis. These translate to a predicted
S02 emission rate of 4.76 lb/10® Btu, only slightly below the average of the CEMS-measured
values.
Other Combustion Parameters-
Results of continuous monitoring of 02, C02, CO, and total hydrocarbons (THC) during the
nine baseline test runs are presented in Table 7. As shown in this table, average concentrations of
02, C02, and CO were similar for all three boiler loads. Total hydrocarbons (THC), with the
measurement based on propane calibration standards, were found at concentrations of 1 ppmv or
less during tests BL01 through BL04.
Particulate Emissions/ESP Efficiency-
Results of EPA Method 5 testing at the ESP inlet and outlet are also included in Table 7. All
of these tests were conducted with all six fields in service. Averaged for each of the three loads,
ESP inlet particulate loadings ranged from 3.41 to 3.51 gr/dscf. Similarly averaged ESP outlet
particulate concentrations were 0.024 and 0.023 gr/dscf for the 65 MWe and 82 MWe tests,
respectively, but approximately twice as high (0.049 gr/dscf) during the tests at 102 MWe. This is
as would be expected as one begins approaching the design conditions of the unit.
65
-------
ESP removal efficiency was slightly lower for the high load tests (98.9 %) when compared
to the 82 and 65 MWe tests (99.4 and 99.3 %, respectively). Corresponding average outlet
particulate emissions on a heat input basis were 0.068, 0.032, and 0.033 lb/10a Btu, respectively.
Cascade impactors were used to determine particle size distributions (PSD) at the ESP inlet
and outlet. A cascade impactor classifies the flue gas particulate sample according to aerodynamic
diameter. Two types of impactors were used during the baseline testing. A high capacity
Anderson impactor was used at the inlet to the ESP, while an eight-stage Anderson MK III impactor
was used at the ESP outlet sample location. Because of the few number of runs, trends between
the boiler load and the PSD data could not be determined. The mass mean diameters of ESP inlet
and outlet particulate taken from the determinations of the particle size distribution data are also
presented in Table 7. The mass mean diameter of the inlet fly ash was on the order of 50 fjm
throughout the baseline test period. Outlet particle size distribution data, taken during tests BL03,
BL05, and BL06, indicate a mass mean diameter of approximately 2 to 7 fjm. Differential mass
concentration plots for these three tests are presented in Figures 14 through 16. These plots
exhibit a bimodal distribution of the particles. The peaks of the distribution are found at 1.2 fjm
and 6.5 fjm, with most of the mass concentrated in 2-10 fjm range.
Ash Resistivity-
Fly ash resistivities were also determined during baseline tests BL06, BL07, and BL08 and
are presented in Table 7 along with the temperatures and moisture concentrations of the flue gas
at the time of the measurements. Resistivities ranged from 6.6 x 101° to 1.9 x 1011 over the load
range and appeared to be on the high end of what might be expected for a 3% sulfur bituminous
coal ash.
Ash Analyses-
A summary of the analytical results for fly ash, economizer/air heater ash, and bottom ash
samples is provided in Table 8. Comparison of ash compositions for the three boiler loads shows
little change in the composition of each ash type with changes in load. The calcium and sulfate
contents of the fly ash averaged 1.28 and 0.93 wt %, respectively, indicating that there was no
reason to expect appreciable S02 capture by the ash.
66
-------
Figure 14. Differential mass concentration plot for test BL03 at the ESP outlet.
67
-------
5.0E-04
4.0E-04 _
3.0E-04 _
-j 2.0E-04 _
1.0E-04
O.OE + OO
0.1
1.0 . 10.0
Aerodynamic Particle Diameter, jum
100.0
Figure 15. Differential mass concentration plot for test BL05 at the ESP outlet.
68
-------
4.0E-04 _
3.0E-04 _
2.0E-04 _
1.0E-04 _
O.OE + 00 I I Ij
0.1 1.0 10.0 100.00
Aerodynamic Particle Diameter, ^m
Figure 16. Differential mass concentration plot for test BL06 at the ESP outlet.
69
-------
TABLE 8. SUMMARY OF CIRCULAR BURNER BASELINE TESTS ASH ANALYSES
Ash Type
Rvash'
Economizer/Mr Heater Hoooer'
Bottom Ash
Test No.
BL06
BL08
BL07
BL06
BL08
BL07
BL06
BL08
BL07
Load. MWe
101
82
66
101
82
66
101
82
66
Al, wt %
0.71
9.64
10.0
(10.1)*
4.68
(4.79)
6.27
4.34
3.92
3.18
2.03
Ca, wt %
1.40
1.37
1.18
(1.18)
1.04
(1.12)
1.21
0.84
0.69
0.62
0.24
Fe, wt %
13.6
14.1
12.6
(13.4)
33.6
(33.2)
31.2
29.9
8.66
7.19
21.9
K, wt %
13.3
13.2
12.8
(13.1)
0.66
(0.69)
0.78
0.63
0.62
0.64
0.46
M0, wt %
0.32
0.33
0.30
(0.31)
0.18
(0.19)
0.23
0.16
0.16
0.14
0.12
Na, wt %
0.17
0.07
0.12
(0.12)
<0.006
«0.006)
<0.006
<0.006
<0.003
<0.003
<0.006
Si, wt %
13.4
12.6
12.6
(12.7)
7.03
(8.89)
10.6
8.26
6.71
4.43
3.89
Ti, wt %
0.66
.66
0.67
(0.68)
0.26
(0.27)
0.29
0.24
0.23
0.19
0.09
S03", wt %
<0.002
<0.002
<0.002
K0.002)
<0.001
K0.001)
<0.001
<0.001
<0.002
<0.001
<0.002
SO/, wt %
1.00
0.96
0.88
(0.88)
0.19
(0.20)
0.20
0.21
0.24
0.07
0.23
H30, wt %
0.09
0.12
0.06
(0.01)
0.12
INA")
0.16
0.22
32.4
48.4
19.7
Fly ash from the EPA Method 6 particulate temple* were analyzed.
' The temples consisted of a 60/60 composite of ash collected from the sir heater and economizer hoppers.
* Parenthetical values represent duplicate analyses.
Not analyzed.
DRB-XCL™ Burner Characterization Tests
The DRB-XCL™ burner characterization tests, carried out during the week of July 13, 1987,
were performed to benchmark the burner performance prior to starting sorbent injection. These
tests provided data under conditions that closely approximated those used in the circular burner
characterization and baseline tests described in the two preceding sections. The results obtained
are summarized in Table 6.
The performance objectives set for the DRB-XCL™ burners included not only the primary
goal of 0.50 lb N0x/10e Btu, but also secondary goals of achieving this performance with a burner
pressure loss of <5 in WC, <5 wt % carbon in the ash, and with a flame length of <22 ft.
70
-------
DRB-XCL™ Burner Optimization Pretests-
The DRB-XCL™ burners used in the demonstration project were the result of development
that incorporated design features of both the B&W dual register burner and the Babcock Hitachi
HT-NR burner as described in Section 4. In conjunction with a separate project, the development
included the operation and testing of one full-scale burner prior to installation at Edgewater in
August 1986. Once all twelve burners were placed in operation, initial tests showed that the
burners were able to meet all the goals except that the NOx emission was slightly higher than
desired (approximately 0.53 lb/106 Btu). Project participants met to discuss the trade-offs in the
design between lower NOx and longer flame lengths and higher carbon in the ash. As a result, it
was agreed that further optimization of the hardware would be pursued prior to beginning the
demonstration. This occurred during the spring of 1987 and is briefly described here as
background leading to the DRB-XCL™ burner characterization tests that followed. During the
optimization tests several iterations of burner modifications identified various factors that had
detrimental effects on the performance of the DRB-XCL™ burners. Prominent among them were
operational constraints associated with the coal pulverizer control system, and uncontrolled air
infiltration into the convection pass and the air heater. In spite of the difficulties encountered, by
May 1987 a final burner configuration was found that the participants agreed would serve the
purposes of the demonstration project.
The resultant burner configuration utilized shallow angle impellers on the lowest burners (D
mill) and the upper middle burners (B mill). The lower middle burners (A mill) and the uppermost
burners (C mill) were equipped with a standard conical diffuser/deflector for coal distribution within
the nozzle, and with a flame stabilizing ring on the end of the coal nozzle. This arrangement
provided conditions that met all program goals except for the carbon in the ash, which averaged
6.2 wt % during the optimization tests. This concession was deemed to be acceptable since the
depth of the Edgewater furnace at the burner level was about as shallow as one might expect to
find among potential retrofit units. With more room in a deeper or an opposed-wall-fired furnace,
substantially higher carbon burn-out would be expected.
Prior to and during the DRB-XCL™ characterization tests, oxygen concentrations in the flue
gas were determined at four locations in the gas path, namely, at the economizer inlet, the air
heater inlet and outlet, and the ESP outlet (effectively at the stack). The measurements at all
points were crosschecked for consistency when changes in concentration occurred (for example.
71
-------
with load change) and led to the conclusion that they were a reasonably accurate reflection of the
differences in 02 concentration at the various locations. Comparison of the differences in 02
concentrations between the economizer outlet and air heater inlet was especially notable between
the time of the circular burner baseline and characterization tests and the time of the DRB-XCL™
characterization tests. Whereas there had been almost no difference when the circular burners
were tested, the air heater inlet 02 was significantly higher than the economizer average value at
the time of the DRB-XCL™ tests. This, coupled with the existing air infiltration within the air
heater, raised the overall excess air level to about 50 wt % as seen by the induced draft fan.
(Subsequent inspection of the economizer and air heater hopper areas revealed that the air in-
leakage was most likely due to gaps near the economizer inlet header connections where the
economizer tubes penetrate the casing.) The net effect, when combined with higher than normal
cooling water temperatures at the time of the tests, was a reduction in the high load condition to
about 680,000 lb steam/hr (approximately 92 MWe) attainable at the time of the DRB-XCL™
characterization tests. During the fall of 1987 repairs were made to reduce the air infiltration in the
economizer and air heater areas, though air intentionally introduced as part of the humidification
process brought the excess air level at the induced draft fan back up to the 50% level .
At the time of the tests a good portion of the load reduction was assumed to be due to the
DRB-XCL™ burner because of the higher pressure drop relative to the circular burners. However,
when data collection resumed in August 1988 following installation of the humidifier, operation at
gross generation loads of 100 to 105 MWe and main steam flows of 700,000 to 750,000 Ib/hr
(about those obtainable with the circular burners) was once again possible. Unfortunately, load
was again curtailed just as intensive testing began under EPA sponsorship due to unacceptably high
temperatures in the vicinity of the nose arch, a condition which was not considered to be the result
of LIMB technology. It did, however, force the project into operation at less than full load
conditions for most of the remaining time. Since then, however, operation at or near the full load
condition has been routinely possible during the DOE-sponsored LIMB Demonstration Project
Extension. It therefore appears that either the cooling water temperature had a much greater effect
than originally assumed, or that other unknown factors were responsible for the load limitations
during the DRB-XCL™ burner characterization tests.
Summary of DRB-XCL™ Burner Tests-
At the time they were conducted, the DRB-XCL™ burner characterization tests summarized
72
-------
in Table 6 were divided into tests at low, medium, and high excess air conditions at each of the
three loads. The original interim report, included as Appendix C, discussed the details of the
results. The report included specific comparisons between the circular and DRB-XCL™ burners and
appropriate interpretation of the high carbon in ash and stack carbon monoxide (CO) values
typically obtained under low excess air conditions. During the main portion of the demonstration,
however, normal boiler operation did not include the high degree of combustion air control
exercised during the characterization tests. As a result, the ten minute average NOx emissions as
recorded by the System 140™ usually varied over a range of 0.1 to 0.2 lb/10® Btu. For this reason,
the N0X emissions of the two burners are compared in Figure 17 for all excess air conditions
tested.
RESULTS AND DISCUSSION OF S02, NOx, AND PARTICULATE EMISSION CONTROL TESTS WITH
LIMB
Test runs of LIMB were initially conducted in September 1987 when S02 removals of
approximately 55-60% were achieved at a Ca/S ratio of 2.0. With start-up of the humidifier in
1988, extended LIMB run times could be realized and the system could be adjusted and optimized.
The humidifier added a potential complicating factor in the determination of LIMB S02 removal,
since S02 could also be captured in the humidifier itself. However, this removal occurred only at
close approach to the adiabatic saturation temperature of the flue gas (approximately 25 °F). This
was indicated by an immediate increase in S02 concentration whenever the water was turned off
while running at the close approach temperatures. Once above a 50 to 60° approach temperature
no significant change in S02 concentration was detectable when water flow was interrupted.
Data taken while the humidifier was operated at high temperatures, typically about 275°F
where no S02 removal occurred, were used to determine S02 removal in the furnace. Later tests
with humidifier outlet gas temperatures approaching the saturation point were used to define the
incremental removal in the humidifier. With humidification, rigorous test periods with no significant
changes in operating conditions extended to as long as 15 hours.
The September 1987 tests included some runs using commercial, hydrated calcific lime
treated with calcium lignosulfonate, an additive that appeared to improve S02 removal in earlier
bench scale studies.13 Additional tests with more extensive use (1673 tons over two periods of
73
-------
1.00
0.80
m 0.60
O
r—
£
x
o
0.40
0.20
0.00
I I
o
' I ' I
Circular Burners
1 1 1
O
—
—
o
0
0 o
o
—
¦ <§>
o
o
0 Dn
—
~
~
—
—
~ D
—
~
DRB-XCL™ Burners
—
—
Composite data from baseline and characterization tests
over the range of excels air conditions used
I I I I I I I I I
—
400
500 600 700
Main Steam Flow, Ib/hr x 10~3
800
900
Figure 17. Circular and DRB-XCL™ burner N0X emissions.
74
-------
10 and 25 days each) of the modified sorbent were included in the 1988-89 demonstration.
Figures 18 through 21 present the overall results of the demonstration project. The plotted data
are included in Tables 9 through 12, respectively, along with other operational information. The
data points on the graphs represent averages of the 10-minute averages collected by the System
140™ over continuous periods under steady operating conditions. Figures 18 and 19 depict the
performance of commercial hydrated calcific lime both at "minimal humidification" and at 20°F
approach to saturation, respectively. The former term covers results obtained both before the
humidifier was installed and when the humidifier was used just to maintain a suitable flue gas
temperature and humidity for effective electrostatic precipitation, normally about 275°F. Figures
20 and 21 show the corresponding curves for the lignosulfonate-doped sorbent. The curves
indicate that the modified sorbent increases S02 removal from 55 to 63% with minimal
humidification, and from 65 to 72% at close approach to saturation, when operating at a Ca/S of
2.0. These are relative increases of 15 and 11%, respectively. The increase is believed to be due
to a combination of the finer size and the nonagglomerating properties of the modified sorbent, as
well as the ability of the modified sorbent to resist sintering at furnace temperatures14.
All LIMB operation in 1989 is depicted in the middle band of Figure 22, showing how
closely the program approached continuous operation. In this figure, boiler outages are represented
in the lower band, together with two interruptions in the test program designated by an "I." The
upper band indicates LIMB outages due both to failures of the mechanical equipment, and to time
spent on LIMB-specific tasks such as repairing and modifying an acid addition system to neutralize
lime and LIMB ash spills, and humidifier inspection and cleaning. The outright mechanical failures
are designated by an "F" and are considered to be representative of the types of failures that might
be expected in a fully commercial system. The difficulties encountered and the solutions developed
are discussed more completely in Section 6. The more frequent interruptions, typically for several
hours, allowed purging the sorbent from the furnace. This provided a "clean" start for a given
day's test activities, as the value to be derived from the tests was considered more important than
simply running with continuous injection. Going beyond the rigorous test periods, the longest
continuous operating period with no interruption in lime flow, other than perhaps a momentary trip
of the feeder, was 5.9 days.
Figure 23 presents the cumulative time of actual operation in 1989. "Available" time also
depicted in this figure represents time when the LIMB system was considered to be available. This
75
-------
100
>
o
c
03
[o
LU
"ro
>
o
E
03
cc
CN
O
CO
0.50
1.00 1.50
Inlet Ca/S Molar Ratio
2.00
2.50
Figure 18. S02 removal vs. Ca/S stoichiometry for commercial lime with minimal humidification.
76
-------
100
1.00 1.50
Inlet Ca/S Molar Ratio
2.50
Figure 19. S02 removal vs. Ca/S stoichiometry for commercial lime with 20°F approach to saturation.
77
-------
100
1.00 1.50
Inlet Ca/S Molar Ratio
2.50
Figure 20. S02 removal vs. Ca/S stoichiometry for lignosulfonated lime with minimal humidification.
78
-------
100
80 -
60 -
CO
>
o
E 40 -
Q)
cc
IN
O
C/3
20 -
o u
0.00
0.50 1.00 1.50
Inlet Ca/S Molar Ratio
2.00
2.50
Figure 21. S02 removal vs. Ca/S stoichiometry for lignosulfonated lime with 20°F approach to saturation.
79
-------
181
jectic
ort T
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TABLE 9. RESULTS WITH COMMERCIAL CAICITIC HYORATEO LIME AT MINIMAL HUMIDIFICATION
Main Flue Elevation Elevation Momentum
Inlet Humidifier Steam Gas 181 187 Flux
Ca/S S02 Outlet Flow Temperature Pulverizera Injection Injection Ratio
Duration Molar Removal Temperature Ib/hr at Nose Out of Porta Porta at
min Ratio % °F x 103 °F Service Operating Operating Noea
140 2.20
60 2.04
100 0.77
70 0.98
30 1.22
430 1.87
100 2.03
80 2.06
230 2.06
160 2.04
120 1.47
460 1.66
680 2.02
400 1.97
30 2.48
30 2.42
67.3 322
67.8 316
23.7 320
30.6 320
37.4 290
64.0 276
62.8 276
66.7 276
69.6 291
49.3 300
40.4 260
61.1 276
60.2 276
64.6 276
63.6 260
60.6 266
676 2200
678 2200
676 2160
712 2230
720 2240
306 1868
688 2289
673 2214
669 2177
313 1771
342 1879
332 1846
602 2098
610 2081
416 2044
407 2043
None 8
None 8
None 8
None 8
None 8
NA 8
None 8
None 8
None 8
NA 8
NA 8
NA 8
NA 8
NA 8
NA 8
NA 8
0 NA"
0 NA
0 NA
0 NA
0 NA
0 773
0 118
0 294
0 290
0 628
0 NA
0 NA
0 176
0 126
0 NA
0 NA
-------
Elevation
181
Injection
Port Tilt
o
0
0
0
0
0
0
0
0
0
0
0
0
0
TABLE 10. RESULTS WITH COMMERCIAL CALCITIC HYDRATED LIME AT CLOSE APPROACH TO SATURATION
Main Flue Elevation Elevation Momentum
Inlet Humidifier Steam Gaa 181 187 Flux
Ca/S S02 Outlet Flow Temperature Pulverizers Injection Injection Ratio
Duration Molar Removal Temperature Ib/hr at Noee Out of Porta Porta at
min Ratio % °F x 10® °F Service Operating Operating Noee
170 1.62 49.6 146 301 1923 NA 8 0 NA
60 0.86 27.9 146 299 1876 NA 8 0 NA
70 0.84 36.2 146 418 1993 NA 8 0 NA
90 0.86 34.6 146 416 2014 NA 8 0 NA
70 1.24 46.2 146 419 2022 NA 8 0 NA
60 1.63 69.4 146 416 2029 NA 8 0 NA
60 2.46 73.3 146 416 2023 NA 8 0 NA
90 0.84 33.9 146 397 2100 NA 8 0 NA
60 1.21 42.6 146 379 2030 B 8 0 NA
60 1.67 61.4 146 420 2066 NA 8 0 NA
60 2.40 71.7 146 412 2034 NA 8 0 NA
130 0.83 39.2 146 411 2067 NA 8 0 NA
120 1.96 62.7 146 420 2066 NA 8 0 NA
-------
TABLE 11. RESULTS WITH L1GN0SULF0NATED CALCITIC HYDRATED LIME AT MINIMAL HUMIDIFICATION
Main Flue Elevation Elevation Momentum Elevation
Inlet Humidifier Steam Gas 181 187 Flux Booeter 181
Ca/S SO, Outlet Flow Temperature Pulverizera Injection Infection Ratio Air Injection
Start Duration Molar Removal Temperature Ib/hr at Noee Out of Porta Porta at Flow Port Tilt
Time Date min Ratio % °F x 10' °F Service Operating Operating Noee Ib/hr °
00
N>
22:39
13:11
17:11
12:01
16:21
12:48
14:16
13:61
14:38
16:44
16:16
19:37
22:67
04:07
14:47
23:07
04:17
16:07
20:18
03/13/89
03/16/89
03/16/89
03/17/89
03/20/89
03/21/89
03/21/89
03/22/89
03/23/89
06/02/89
06/03/89
06/06/89
06/06/89
06/06/89
06/06/89
06/06/89
06/07/89
06/07/89
06/09/89
440
200
110
220
90
40
90
30
30
40
40
130
180
160
460
230
220
760
270
2.17
1.86
1.83
1.94
1.92
1.30
1.97
1.96
1.96
1.98
2.06
1.09
1.08
1.07
1.06
1.06
1.06
1.06
1.64
88.7
60.8
63.3
68.9
66.4
39.4
69.0
67.7
60.6
66.6
66.6
36.4
31.4
32.1
30.0
30.0
30.6
34.4
48.0
276
276
276
276
276
276
276
276
263
264
270
273
276
276
276
276
276
276
276
317
674
666
647
668
662
627
431
674
461
367
372
301
299
301
290
292
292
373
1817
2117
2146
2173
2102
2120
2062
1966
2240
2262
2107
2083
1932
1928
1926
1907
1936
1924
2096
NA
NA
NA
NA
NA
NA
NA
C
NA
D
D
D
D
D
D
D
D
D
D
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NA
366
371
367
301
312
290
604
347
609
936
878
1196
1241
1186
1206
1176
1248
741
20000'
41287
42933
39939
38218
40168
37869
38067
36627
37246
39778
37862
38200
39639
39684
39120
38703
39867
34328
0
0
0
0
•16*
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NA - Not available.
* Estimated value*.
® Down from horizontal.
-------
TABLE 11 (continued).
Mein
Intet Humidifier Steam
Ca/S S02 Outlet Flow
Start Duration Molar Removal Temperature Ib/hr
Time Date min Ratio % °F x 10"®
Flue Elevation Elevation Momentum Elevation
Gaa 181 187 Flux Booeter 181
Temperature Pulverizers Injection injection Ratio Air Injection
at Noee Out of Porta Porta at Flow Port Tilt
°F Service Operating Operating Noee Ib/hr °
03:08
12:69
14:29
01:39
03:69
13:34
20:24
23:24
06:44
19:26
00:06
06/11/89
06/12/99
06/12/89
06/13/89
06/13/89
06/17/89
06/17/89
06/17/89
06/18/89
06/18/89
06/19/89
80
90
110
90
110
330
160
420
130
260
370
1.62
1.97
2.03
1.83
1.79
1.94
1.47
1.63
1.47
0.96
1.00
62.7
60.2
63.9
64.6
60.1
67.6
66.2
43.6
63.3
38.6
30.3
276
276
276
276
276
276
276
276
276
276
276
336
667
634
344
314
696
689
30.1
670
604
308
1896
2406
2376
1992
1912
2367
2427
1941
2299
2070
1829
A
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
838
223
219
697
772
316
340
963
342
246
663
34236
26260
26996
30376
30468
30602
32646
34486
31761
33124
33776
NA « Not available.
^ Estimated veluee.
-------
TABLE 12. RESULTS WITH LIQNOSULFONATED CALCITtC HYDRATED LIME AT CLOSE APPROACH TO SATURATION
Main Flue Elevation Elevation Momentum Elevation
Inlet Humidifier Steam Gee 181 187 Flux Booeter 181
Ca/S S02 Outlet Flow Temperature Pulverizer* Injection Injection Ratio Air Injection
Start Duration Molar Removal Temperature Ib/hr at Noee Out of Porta Porta at Flow Port Tilt
Time Date min Ratio % °F x 10'* °F Service Operating Operating Noee Ib/hr 0
00
4*
11:21
13:18
12:44
11:56
11:16
03/22/89
03/23/89
06/02/89
06/03/89
06/04/89
10:46 06/06/89
12:07 06/08/89
16:28
11:48
13:38
06/09/89
06/10/89
06/10/89
140
80
180
190
180
180
130
160
110
60
1.96
1.90
1.98
2.00
1.97
1.94
2.06
2.12
1.99
1.99
71.4
66.6
73.8
66.1
66.8
69.0
73.6
67.2
62.4
73.1
161
146
146
146
146
146
146
146
146
146
429
676
460
368
367
381
377
376
366
367
1976
2248
2227
2099
2067
2141
1967
2061
2376
2366
C
NA*
D
D
D
D
D
D
D
D
692
311
679
888
801
764
762
716
442
1148
37129
34621
36764
39696
37618
36777
37676
32923
20669
34662
NA - Not available.
-------
LIMB Mechanical
Failures (F)
and
Other LIMB Down Time
LIMB
Operating
Unit Down
and
Nontest Periods (I)
-------
January-June, 1989
Figure 23. LIMB operational time.
86
-------
includes the actual operating time, plus time when operation was intentionally interrupted (primarily
to obtain "zero" values) for reasons unrelated to the technology. The total time during which LIMB
would have operated had there been no such interruptions, or mechanical failures, is indicated as
"Test period" time in the figure.
SO- Capture in the Furnace
S02 removal is determined by measuring S02 at the stack and comparing this level with an
uncontrolled calculated S02 level based on the coal sulfur content as previously described.
Typically the S02 concentration falls rapidly when sorbent is first injected into the boiler, and then
continues to decrease gradually to a steady state value within about 30 minutes. This incremental
increase in S02 removal is thought to be due to supplemental reaction of sorbent that settles out
on boiler surfaces. After this initial acclimation of the boiler, the S02 levels remain relatively
unaffected by sootblowing and other boiler operations since normally there is a continual renewal
of the deposits throughout such activities.
S02 control is primarily a function of Ca/S molar ratio and the degree of humidification as
presented in Figure 24, a reproduction of the S02 and 02 traces from the continuous emission
monitoring (CEM) system. Here the initial gradual increase is masked by the simultaneous lowering
of the humidifier outlet temperature to 145°F (about 20 °F above the adiabatic saturation
temperature of the flue gas). However, the figure serves as a good example of the effect of
stoichiometry as it was increased stepwise throughout the day. In addition, the enhancement
resulting from operation at close approach to saturation is evident when it is lost at about 17:07.
At this time the humidification water flow was reduced to control the humidifier outlet temperature
at 275 °F (150°F approach to saturation) which in turn resulted in an increase in S02
concentration.
Considerable pilot scale tests by a number of investigators had indicated that S02 removal
was particularly dependent upon the temperature at the injection point. Consequently, limited
parametric optimization tests were conducted to quantify this effect at full scale. The variables,
all of which are related to the temperature, included injection at different elevations in the furnace,
momentum flux ratio (essentially injection velocity and furnace penetration at a given load), the
angle of injection (nozzle tilt), and boiler load. The results show that these parameters have little
87
-------
2500
2000
| 1500
Q.
Q.
CSI
o
« 1000
o
A3
cn
500
Start Ca/S to 0.8
1
ATas*= 20°F
ATas* = 150°F_
W1
• • — •••¦ 02
*ATas Approach to adiabatic saturation temperature
10:33
12:33
14:33
Time
16:33
25
20
15
10
>
*D
o
>
o
CO
4-"
C/)
Figure 24. S02 and 02 concentrations during Ca/S variation on June 7, 1989.
88
-------
effect on S02 removal in the Edgewater boiler over most of the ranges tested, indicating that the
system is very insensitive to minor changes, if the initial design parameters enable near-optimum
operation.
The only factors that influenced S02 removal significantly were those related to sorbent
injection at the low end of the 2300 to 1600°F sulfation temperature window or with a gross
reduction in injection velocity. Figure 25 compares data obtained under preferred conditions (Figure
18, Table 9) with the data obtained during periods when unfavorable conditions existed (Table
13). They include injection at the upper elevation in the furnace, periods when the boiler was at
minimum load and/or when the top row of burners was off. Under such conditions the estimated
temperature at the injection point was 1900°F or lower and the increased residence time did not
appear to be sufficient to compensate for the difference.
At one point outside of the test period, the booster air was turned off entirely, leaving the
lime transport air as the sole means of conveying the lime into the furnace. A dramatic loss in
efficiency resulted, indicating that there is a point at which the momentum flux ratio becomes
important. The condition was beyond the range of control on the booster air flow, however, and
could not be readily tested. Injection into higher temperature zones was attempted in the form of
tilting the injection nozzles at elevation 181 down 15° from horizontal. No appreciable change in
S02 capture was apparent. While this was not expected to lead to a dramatic difference, it leads
one to suspect that elevation 181, where the temperature is approximately 2300°F, was very
close to an optimum injection point for this boiler. Injection ports could have been located at a
lower elevation where temperatures are higher; however, recirculation in the lower furnace would
probably have resulted in sintering and decreased efficiency.
SO- Removal bv Humidification
The Edgewater project demonstrated that LIMB S02 removal could be enhanced by
humidification of the flue gas. Most of the tests conducted at close approach to saturation were
limited to a period of 8 to 12 hours so that the baseline removal could be certified. The longest
continuous run with no interruption was 29 hours. Subsequently, however, the humidifier has
been operated continuously for as long as 11 days at a 20 to 25°F approach temperature as part
of the Coolside Process tests under the U.S. Department of Energy-sponsored LIMB Demonstration
Project Extension.
89
-------
100
80 -
~ + Injection under optimum conditions, minimal humidification
.2 60 -
40 ~
20 -
x Injection at low temperature and/or with low velocity,
minimal humidification
I I I I I I I
0.50
1.00 1.50
Inlet Ca/S Molar Ratio
2.00
2.50
Figure 25. S02 removal comparison at less than optimal conditions.
90
-------
181
fectic
xt Ti
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TABLE 13. RESULTS WITH COMMERCIAL CALCIT1C HYDRATED LIME AT LESS THAN OPTIMAL CONDITIONS
Main Rue Elevation Elevation Momentum
Inlet Humidifier Steam Gae 181 187 Flux
Ca/S SOa Outlet Flow Tempereture Pulverizere Injection Infection Ratio
Duration Molar Removel Temperature Ib/hr at Noee Out of Porte Porte at
min Ratio % °F x 10® °F Service Operating Operating Noee
40
260
80
60
90
70
230
120
80
230
170
210
230
140
770
160
690
1.96
1.79
1.06
1.86
0.88
1.81
1.63
1.38
1.38
1.07
1.96
1.96
2.12
2.06
1.93
1.60
1.66
64.3
44.2
32.2
44.9
32.4
47.6
42.0
30.1
30.7
31.3
46.6
46.1
44.9
43.6
39.7
37.1
39.6
326
367
349
369
368
376
371
366
367
369
276
300
300
269
276
276
276
700
736
639
740
694
734
328
697
639
667
694
316
316
326
318
482
296
2120
2237
2011
2136
1994
2139
1702
2066
2020
2173
2238
1811
1818
1828
1880
2064
1826
None
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
76
664
NA
NA
NA
-------
The effect of humidification is readily apparent in Figure 24 where the trace of the S02
concentration suddenly increased as soon as the humidifier outlet temperature control was reset
from 145 to 275°F (150°F approach to saturation). Figure 26 summarizes the effect of
humidification to a 20°F approach temperature for the range of stoichiometries tested. As can be
seen, humidification increases S02 removal from 55 to 65% at a Ca/S ratio of 2.0. The tests run
at 40 and 60 °F approach temperatures suggest that the enhancement is lost somewhere slightly
above a 60°F approach. However, the few data points obtained are not considered sufficient to
define a truly representative curve. Pilot work and dry scrubber experience predict an exponential
increase in S02 removal as one approaches the saturation temperature.16
NO„ Emission Control
In order to reduce N0X emissions, the existing circular burners were replaced with B&W's
DRB-XCL™ burners as part of the demonstration project. As was shown in Figure 17, discrete
baseline tests with the older burners indicated average emission levels of approximately 0.7 to 0.8
lb/10s Btu over the mid- to peak main steam flow range of 500,000 to 800,000 Ib/hr. Discrete
tests with the DRB-XCL™ burners indicated a corresponding emission range of 0.4 to 0.5 lb/108
Btu. Figure 27 shows the NOx emission data collected by the CEMS (each point represents a
10-minute average value) over full load range used at the plant during February 1989. From this
perspective the DRB-XCL™ burners do not appear to be as sensitive to load conditions as the
discrete tests had shown.
Substantial portions of the NOx emission data, with an overall average value of 0.48 lb/106
Btu, are presented in Figure 28. The data cover the full range of operating conditions during the
January - June 1989 period. The DRB-XCL™ burners operate at about 4 to 5 in WC pressure drop,
in comparison to the 2 to 3 in WC required by the circular burners. The unburned carbon content
averaged 1.54 wt % for four LIMB ash samples collected isokinetically while sorbent was being
injected at a Ca/S ratio of 2.0 and NOx emissions were about 0.48 lb/10s Btu. This would equate
to approximately 4.6 wt % in an ash undiluted by reaction products and excess sorbent. Also note
that there appeared to be no interactive effects between sorbent injection and NOx reduction.
Since the System 140™ data collection from the CEMS was unavailable at the time of the
circular burner baseline tests, the variability of NOx emissions obtained with that burner cannot be
92
-------
100
80 -
x Commercial lime, 20°F AT.
0.50
+ Commercial lime, minimal humidification
1.00 1.50
Inlet Ca/S Molar Ratio
2.00
2.50
Figure 26. Effect of humidification on S02 removal.
93
-------
0
200 400 600
Main Steam Flow, Ib/hr x 10"3
800
Figure 27. DRB-XCL™ burner N0X emissions vs. load, February 1989.
94
-------
Reproduced from
best available copy.
1.0
0.8
£ 0.6
ID
0
£
1 0.4
0.2
0.0
Figure 28. 10 min Average DRB-XCL™ burner N0X emissions, January 8 - June 9, 1989.
1/22 2/11
3/3 3/23 4/12 5/2
January-June, 1989
5/22 6/11
95
-------
compared directly with those of the DRB-XCL™ burner (measured at the stack) as shown in Figure
28. However, some idea of the magnitude of the variation can be seen in Figure 29 where the
readings taken approximately hourly during the baseline tests are plotted. For unknown reasons,
the values recorded earlier in the day are frequently somewhat different than those recorded later.
Although operator changes were thought to be involved to some extent, the differences did not
occur every day, nor did they always coincide with the precise shift change. The NO„ emissions
determined at the air heater inlet (measured during the baseline tests only) are also shown in Figure
29. Though not subjected to as rigorous quality control as those at the stack, they generally agree
reasonably well with the stack values and exhibit the same sort of variability.
As noted earlier, Figure 28 contains the individual 10 minute averages recorded by the
System 140™. When emissions are reported to regulatory agencies, however, weighted (by the
coal firing rate) rolling average values over a specified period of time are generally used. While it
was beyond the scope of this project to delve into a full statistical analysis of all of the emission
data as has been done with data in another study,16 a portion of the NOx data were further reduced
to provide at least some indication of what might be expected when various averages are
calculated. Weighted hourly and daily averages were first determined for the periods shown in
Figures 30 and 31, respectively. The hourly averages (shown in expanded form in Figure 32 for
the January 8 to 26, 1989 period) were then used to develop the 24 hour rolling averages plotted
in Figure 33. The 30 day rolling averages presented in Figure 34 were similarly developed from the
daily averages in Figure 31. With twice the standard deviation as a measure of variability at the
95% confidence level, the data are summarized in Table 14.
Particulate Emission Control
An early concern of the LIMB demonstration was the ability of the Edgewater ESP to handle
the two- to threefold increase in particulate loading that accompanied sorbent injection. In
addition, the finer size of the particulate was expected to lower particulate collection efficiency.
However, the preliminary tests in 1987 dramatically indicated that the extremely high electrical
resistivity of unconditioned LIMB ash was by far the most immediate impediment to continuous
operation. As noted earlier, the high resistivity, in particular, precluded continuous operation
beyond several hours as the opacity climbed toward the 20% limit at that time. The overall
conclusion, discussed in much greater detail in a separate report8 on a complementary study, was
96
-------
0.8
D
m 0.6
CO
o
jo
O* 0.4
0.2
0
¦fcr
i r
mtti.
#
"tf
+t
+¦
¦(%
rr
•r
~ Air heater inlet N0X + Stack N0X
J L
J L
W".
J L
05/27 05/29
05/31 06/02
May-June, 1986
06/04 06/06
Figure 29. Hourly baseline circular burner N0X emissions, May 27 - June 4, 1986.
97
-------
0.2 -
1 I ! I I I I
01/02 01/22 02/11 03/03
January— February, 1989
Figure 30. Hourly average DRB-XCL™ burner NOx emissions, January 8 - February 28, 1989.
98
-------
0.8 -
i r
£ 0.6
tO
o
£
d 0.4
+H-
4+
+++ ++ s
+¦ _L+ 4-
+ +++J+
L++' ++ , 1++ ++ + ^ +
++ +
v+*
0.2 -
0
J I L
I I
01/0.2 01/22 02/11 03/03 03/23
January— April, 1989
04/12
Figure 31. Daily average DRB-XCL™ burner NOx emissions, January 8 - April 20, 1989.
99
-------
1
0.8
3
m 0.6
o
o 0.4
z
0.2
J_L
01/08
01/12
01/16 01/20
January, 1989
01/24
01/28
Figure 32. Hourly average DRB-XCL™ burner NO emissions, January 8 - 25, 1989.
100
-------
0.8
0.6
0.4
0.2
0
01/08 01/12 01/16 01/20 01/24 01/28
January, 1989
Figure 33. 24 hr Rolling average DRB-XCL™ burner N0X emissions, January 8 - 25, 1989.
-------
02/21
03/13 04/02
February— April, 1989
04/22
Figure 34. 30 day Rolling average DRB-XCL™ burner NOx emissions, February 28 - April 20,1989.
102
-------
TABLE 14. AVERAGE DRB-XCL" BURNER NO, EMISSIONS
NO. Emieeion
Circular Burner
Hourly recorded value
DRB-XCL* Burner
10 Minute average
10 Minute average
Hourly average
Daily average
24 Hour rolling average
30 Oay rolling average
Time period
6/27 - 6/04/86
1/08-6/09/89
1/08-4/20/89
1/08-4/20/89
1/08-4/20/89
1/08 - 3/02/89
1/08-4/20/89
Figure No.
Average Variability
lb/106 Btu lb/109 Btu
28
28
30, 32
31
33
34
0.48
0.48
0.48
0.49
0.47
0.49
0.10
0.11
0.10
0.08
0.08
0.02
* Over which NO, emieaion and variability were calculated.
that the high resistivity ash gave rise to a back corona condition that proceeded sequentially from
the ESP inlet field toward the outlet. In situ resistivity measurements obtained on the LIMB ash at
the time are shown in Figure 35. Figure 36 presents "V-l" (secondary voltage vs. current density)
curves measured both with and without sorbent injection. The former is typical of a back corona
condition, with little or no useful power being applied to the ESP fields.
The humidifier installed for the 1988-89 operations proved to be an effective means of
conditioning the ash and restoring the efficiency of the ESP to near normal levels. Although the
Edgewater humidifier was designed to accommodate close approach to saturation for S02 removal
purposes, adding even modest amounts of water (lowering the flue gas temperature to about
275°F) was sufficient to maintain opacity generally in the 1 to 7% range throughout the
demonstration. Specific ESP tests included measurement of in situ resistivity, operating voltage
and current, and inlet and outlet gas flow, mass loading, and particle size distribution. These
intensive tests were conducted during the week of May 22, 1989, and cover operation of the ESP
with both three and five of the six fields in service. Unfortunately, the unit was restricted during
the tests to the 75 MWe load limit by the long term loss of refractory material in the vicinity of the
nose arch mentioned in Section 1. Nevertheless, it did represent a condition under which the
humidifier was operating at the full design flue gas flow since repairs to various sources of air in-
leakage were not fully successful.
103
-------
1012 —
£
o
o
¦w
'¦w
w
'in
CD
QC
1011
1010 _
109
o
o
O °
c£°
o
§
o
o
o
o
0 -8 ~ °
O With sorbent injection
Baseline 2.5% S coal
290 300 310 320 330 340 350 360 370 380
Temperature, °F
Figure 35. In situ resistivity measurements without humidification, September 1987.
104
-------
15
10
| I I I I | I I I I | I | I I | I I I
- With LIMB
Without LIMB
"i i i i i l i i i i l i i i i l i i i i
20
30 40 50
Secondary Voltage, kV
60
Figure 36. Typical sorbent injection effects on ESP electrical characteristics, September 1987.
105
-------
Figures 37 to 41 summarize the particulate size distribution, "V-l" curves, resistivity, and
ESP collection efficiency data obtained. For the purposes of the tests, the ESP inlet temperature
was maintained at 275°F except for one test with an inlet temperature of 165°F (Figure 39). The
results of these tests, coupled with those from associated laboratory tests and mathematical
modelling, have recently been presented in detail.17 The data indicate that humidification to 275°F
lowered resistivity to 2.3 x 1011 ohm-cm, permitted continuous operation at acceptable opacity
levels, and resulted in particulate emissions on the order of 0.01 to 0.02 lb/106 Btu or less at the
conditions tested.
Comparison of Boiler Characteristics Before and After LIMB
Prior to the demonstration, it was anticipated that the LIMB ash would follow the gas path,
especially since early studies had not indicated the degree of increased sootblowing eventually
found necessary to maintain surface cleanliness at Edgewater. Fouling of the convective surface
occurred at a rate much greater than the sootblower system could handle. Upon starting sorbent
injection, gas temperatures throughout the boiler convective pass increased rapidly, though
steadily, with the air heater outlet gas temperature eventually rising to about 40 °F over preinjection
temperatures. The corresponding Kf's for the various convective surfaces showed the reduction in
heat transfer rates caused by the deposition of material on the tubes. Immediately after
sootblowing a bank of heating surface, however, the Kf's would return to their original values,
indicating surface cleanliness could be restored.
In retrospect and based on B&W's experience on high-ash subbituminous and lignitic coals,
it is felt that surface cleanliness could have been maintained at or near preinjection values had the
sootblowing system been upgraded even more than it was to handle the higher deposition rates.
However, meaningful, quantitative comparisons of boiler heat transfer rates with and without the
LIMB system are not possible in the absence of that capacity.
Conclusions from LIMB Tests
The results of the LIMB Demonstration Project on a full-scale, coal-fired utility boiler show
that the technology has met or surpassed program goals with respect to S02 and NOx emission
control. Moreover, in spite of early difficulty in controlling particulate emissions with the ESP, the
design and installation of a full-scale humidification system appears to have overcome the adverse
106
-------
hiiIr
i—i i i i 11
ESP Outlet
o —
o _
o
o
o
o
o
o
o
o
o
o
o
o
o
ESP Inlet
o
o
o
o
o
o
• o
• - O
o
o
1CH
I I I I I I I I I I I I I I I ll I I M I I I ll
10° 101
Particle Diameter,
102
Figure 37. Typical LIMB ash particle size distributions at ESP inlet and outlet.1
107
-------
15
o 10 —
0
I | I I I I
Fourth Field
Second Field
Third Field
Fifth Field
Inlet Field
20
30 40 50
Secondary Voltage, kV
60
Figure 38. ESP electrical conditions during LIMB operation with humidification to 275°F."
108
-------
15
| 10
<
C
-------
1014
1013
o
£
JZ
o
>
+->
[>
'•4-'
-------
99.999
99.99
99.9
99
Figure 41. Fractional collection efficiencies of the ESP during LIMB operation.'7
J
I I I I I I 111 I I I I I 1111 I I III
L
0
—
0
—
ooo0oo0oooo0
o
o°
o
—
—
—
—
0
—
-
o
—
—
O O
-=
O O
—
o
o
o
o
o
—
I I I I I I III I I li Mill I 111!
Ill-
1CT1 10° 101 102
Particle Diameter, um
111
-------
aspects of the increased quantity of fine, high resistivity ash. Specifically, the use of a
lignosulfonate-doped hydrated calcific lime, produced in bulk by a commercial supplier, and
operation of the humidifier at a 20° F approach to saturation resulted in S02 removal efficiencies of
up to 72% at 2.0 Ca/S, far in excess of the 50% program goal. Even without the lignosulfonate
additive, up to 65% capture was possible under the same conditions. Corresponding removal
efficiencies of about 62% and 55% were obtained in the absence of humidification to the close
approach temperature. The S02 data have also been reduced and presented in a form that
characterizes the removal over a range of stoichiometric ratios. At the same time, an average NOx
emission level of 0.48 lb/108 Btu was achieved over months of operation with the DRB-XCL1"
burners installed to meet a goal of 0.5 lb/106 Btu or less. Particulate emission control was
effectively restored by modest levels of humidification as indicated by opacity measurements at the
ESP outlet. Particulate mass emission levels of 0.01 to 0.02 lb/10® Btu were found to be far below
the 0.1 lb/106 Btu goal, although the tests had to be conducted at conditions representing
approximately 75% of the full boiler load.
112
-------
SECTION 8
LIMB COMMERCIAL DESIGN AND ECONOMICS
The commercialization of LIMB technology based on the Edgewater demonstration
experience requires the design of retrofit LIMB systems that include the most successful equipment
and methods found. The "commercial designs" are intended to provide recommendations and
guidelines for future utility applications. They also serve as a basis for developing cost
comparisons with wet scrubbers on similar size units.
Utility and industrial boiler owners will be faced with complex decisions on how they will
comply with emissions legislation. Various combinations of technologies (scrubbers, sorbent
injection, low sulfur or cleaned coal) in addition to fuel switching, boiler retirements, new capacity
addition, and power purchases all have to be considered singly or in combination. Various
strategies can be developed to comply with the legislation. The best implementation plan will be
dependent on what must be attained, the capital and operating costs of the particular technology
or combination of technologies selected, loss of net generation capacity, and the ability to
implement the technologies on specific boiler installations.
To cover the range of capital and operating costs and the S02 removal capabilities, two
LIMB systems were considered and are evaluated here. The scope for both begins with truck
delivery of dry hydrated calcitic lime to the site and ends with removal of a water-conditioned LIMB
ash in trucks as was done at Edgewater. A system with the humidification capability to restore
ESP performance only is the lowest in overall cost, and can be retrofitted to almost any wall-fired
unit. The other system has the ability to humidify to a 20° F ATgs in order to provide maximum
S02 removal capability. This system is more difficult to retrofit into an existing plant and has a
higher cost both to install and operate because it requires:
• A much larger atomizing air compressor and water pump capacity, with attendant electrical
power needs.
• A flue gas reheater after the ESP to eliminate condensation in the stack and increase plume
buoyancy.
113
-------
• The addition of heaters for ash transport air and other measures to protect against
condensation in the ash removal and collection system.
• Addition of a humidification chamber with 2.2 sec retention time in a straight section
downstream of the humidification atomizers.
In the commercial design, a number of typical constructions have been combined to form a
hypothetical average boiler installation to which the LIMB processes could be advantageously
applied. This boiler would have the range of temperatures, gas-flow patterns, and residence time
over the load range that is conducive to sulfur capture when dry calcific hydrated lime is injected
into the furnace. Although these characteristics may be found in other boiler designs, a 150 MWe
B&W radiant boiler of the "Carolina" type with design parameters typical of utility installations built
during the 1950-1970 period has been selected. It has a regenerative type air heater, a new or
recently installed ESP, and burns pulverized bituminous coal.
SYSTEM DESIGN - EQUIPMENT AND LOCATION
The equipment required for a LIMB/humidification retrofit can be most easily grouped by the
systems in which they operate. These are listed in the following sections with a brief description
of the equipment included in each group and their locations.
Lime Unloading. Storage, and Feed System
Lime is delivered to the site by pneumatic trucks, two of which can unload lime into the 40
ft diameter silo at the same time. The silo can hold a seven-day supply of lime when the unit is
operating at its maximum continuous rating (MCR).
Lime discharge at the conical bottom is aided by internal air slides on the sloping sides. On
opening a rotary valve, lime drops into a rotary feeder whose speed is controlled by boiler load.
The rotary feeder prevents flooding and feeds at a controlled rate into a powder pump. Sorbent is
compacted as it is pushed through the pump barrel by the screw sealing against the transport line
back-pressure. It is then fluidized by compressed air in the pump mixing chamber sufficiently to be
carried through the transport line to the distribution bottle.
114
-------
Furnace Injection System
Based on furnace gas temperature ranges, lime is injected into the upper furnace at a level
generally close to the screen tubes and entrance to the superheater. Twelve injection nozzles
similar to the Edgewater design penetrate the front wall carrying dry air-transported hydrated lime.
Additional "booster" air is injected through annular openings surrounding the lime nozzles to aid
penetration and mixing with furnace gases.
The flow of lime to the nozzles is equalized in the distribution bottle. Booster air is supplied
by a 150 HP fan located as close to the injection points as possible to reduce the distribution
ductwork.
Humidification System
Humidification takes place in the flue after the air heater and ahead of the ESP. To obtain
optimum close approach to saturation (20°F ATgs), the placement of the atomizers must allow
about 2.2 sec residence time in a straight section of duct after atomization. This allows the water
droplets to evaporate completely, thus preventing water from wetting the walls and causing
particulate build-up. The optimum velocity that will allow good atomization and the shortest length
of straight flue is about 27 ft/sec. This requires about 60 ft of straight flue downstream of the
atomization point, considerably less than that needed for the existing gas velocity of about 60
ft/sec.
To obtain this reduced velocity while maintaining gas flow, the cross-sectional area of the
flue must be increased to 304 ft2, and the 60 ft straight length, with transitions, must be designed
to fit into the existing flue arrangement with minimum cost and minimum increase in system
resistance. This section is called the humidification chamber.
Using humidification section dimensions of 15.25 ft wide x 20 ft high, there would be an
array of atomizers 12 wide x 17 high on 12 in centers, with 2 or more feet of clearance from the
flue walls. This arrangement requires 17 spray lance assemblies projecting from each side, with 6
atomizers in each lance.
An enclosed, heated, and insulated platform is required on each side of the chamber to
115
-------
monitor the operation and maintain the spray lances. These enclosures would contain atomization
air and water headers, distribution piping, and associated valves, filters, and instruments.
Compressed air and water for humidification would require new equipment which is located
in the "LIMB island" described below. Three 50% capacity compressors (1000 HP each), two
100% water pumps (80 HP each), three duplex water strainers, a 2000 gal air receiver, and a
7500 gal water tank (11 ft diameter x 12.5 ft high) would be supplied along with associated
piping, valves, instruments, and controls.
Ash Collection. Storage, and Removal
The application of LIMB to a boiler increases the total fly ash by a ratio of about 3:1 when
operating at a 2.0 Ca/S ratio. For this evaluation it has been assumed that, similar to the
Edgewater installation, the existing ash removal system can collect and transport the additional ash
by more frequently emptying of both the ESP hoppers and the hoppers under the air heater and
convection pass. It is noted, however, that some plants would not have this margin available and
would find it necessary to consider increasing the capacity of all or part of their ash handling
system.
The evaluation has been based on the assumption that the existing storage and disposal
systems do not have adequate capacity. As a result, the design includes a new ash silo with
primary and secondary collectors, fed by new ash lines which tie into the existing lines from the
ESP and convection pass hoppers. An additional mechanical exhauster provides the vacuum with
the transport air being pulled through an atmospheric air heater. An ash conditioning system is also
supplied at the ash silo discharge. More detailed descriptions of the equipment are provided in the
following paragraphs.
Collection System-
As at Edgewater, ash removal systems installed during this period are mostly of the vacuum
type included in this design. It includes a primary cyclone type collector and a secondary pulse jet
bag filter, both mounted on the top of the silo and emptying into it through double-dump discharge
gates. The vacuum is generated by two 100% mechanical exhausters (125 HP each) mounted at
ground level with an alternator valve for exhauster selection and a vacuum breaker and guard filter
assembly on the suction side.
116
-------
This system will be tied into the existing ash transport lines after they leave the ESP and
boiler pick-up points. Diverter valves will allow ash to be routed from either or both systems to the
new silo as required.
Because of the additional moisture in the flue gas from the low ATas operation, a system of
heaters and lead-in piping was added for the transport air taken from the atmosphere. When
combined with the flue gas from the hoppers, transport line temperatures will then stay above
saturation when properly insulated. This prevents condensation from corroding the equipment,
causing build-up and pluggage of the piping and hoppers, and blinding of the filter bags.
Storage System-
The new ash silo is 40 ft. in diameter and will hold seven days of LIMB ash at MCR. It will
have a relief valve and level sensors for monitoring the collecting and discharge operations. The
conical hopper will have a 60° slope and a 5 ft. diameter outlet connection.
Ash Removal System-
Ash will be discharged into trucks through a gravity system including a vibrating bin
bottom, an adjustable speed rotary feeder, and a paddle type ash conditioner (pug mill). Truck
loading will be under the control of a plant operator working in a local control room which
overlooks the operation. From this point he/she will be able to direct the truck positioning and
control the ash flow and conditioning process. The operation of the ash discharge slide gate,
rotary feeder, ash conditioner, and water supply will be automatically sequenced, with ash and
water rates manually adjustable from the panel. The loaded trucks normally wait about 15 min for
the steaming to subside, then proceed to the plant scales and on to the disposal site.
LIMB Island Concept
Those items and groups of equipment that were not necessarily located in the existing
boiler building and associated structures have been brought together and housed in a single
separate structure which has been named the "LIMB island". This structure will support and/or
enclose:
• The lime storage silo truck unloading, bin discharge, and feed equipment.
• The ash storage silo, collecting equipment, unloading equipment, and truck drive-through.
117
-------
• The atomization water tank, feed pumps, and two sets of strainers.
• The atomization air compressors, receiver, and instrument air conditioning equipment.
• Rooms for electrical switchgear, motor control centers, ash truck loading, and LIMB control
room.
• Noise abatement enclosures for the centrifugal air compressors and exhausters.
The advantages of this "island" concept include:
• Consolidation of design, engineering, and construction of major LIMB equipment into one
new building. This avoids a number of costly separate and unique efforts to locate most of
the equipment listed above in available spaces throughout the existing plant and makes
efficient use of the area under the two silos.
• Ease of operational monitoring and maintenance by having most new rotating equipment,
electrical equipment, and the central control point in one building.
• Reduced interference with the plant's existing space, equipment access and operations both
during and after the construction period.
• From the marketing point of view it will be easier to develop and present designs common to
a number of plants; reducing the need to consider a completely new approach for each.
Pipe Rack
The design includes a 150 ft long pipe rack for supporting the pipe runs between the LIMB
island and the boiler structure. The support elevation will be 20 ft above grade to clear roadways.
This is also the elevation of the main floor of the LIMB island providing easy access to the
maintenance catwalk. The rack will carry the interconnecting piping, electrical wiring in cable
trays, and a walkway with outdoor lighting fixtures.
Flue Gas Reheat System
A new condensing type steam coil is installed in the existing flue downstream of the ESP.
Assuming an available saturated steam source at 300 psig, a 40 °F increase in gas temperature at
MCR, and a gas-side pressure drop of 1.5 in WC, the heater would have a face area of 292 ft2, be
40 in deep and have 8 rows of 1 in O.D. bare carbon steel tubes. The complete installation would
require flue transition pieces, supply and condensate piping, insulation, control valves, and a
118
-------
condensate drain tank. Flows are generated by the condensing rate while maintaining a constant
level in the tank.
An average retrofit would require the addition of maintenance access platforms and
allowances for reheater section replacements. The latter would involve a monorail and hoist
system. The required support structure design would depend on the proximity of existing
structures, the possibility of framing in to them, and available ground space for column placement.
Instrumentation and Controls
Recognizing that each plant will have its own philosophy and that any final design would
have to make accommodations for its unique facilities and requirements, this design would attempt
to keep most LIMB operator functions out of the main control room except those necessary for
boiler operation. These would include initiation of furnace injection, humidification and reheat as
complete systems, full sootblowing control, and monitoring LIMB/humidification conditions critical
to unit performance and safety. Problems would be annunciated on a system basis only.
Subsystem and individual equipment start/stop operations would be from the LIMB island
control room. The design includes computerized process control and operator interface units so
that most process functions can be monitored and adjusted from one point, close to the operating
equipment.
Required instrumentation for local or remote status indication is included with recording and
alarms for critical conditions. Annunciation of trouble spots are highlighted on the computer
graphics.
The assumption is that the plant will have a continuous emission monitoring system in place
or will provide one.
Electrical Equipment
Assuming the plant can provide a power lead from its 4160 VAC bus, the design would
include the necessary transformer and switchgear to power the LIMB island and the booster air fan.
The electrical equipment room is located on the main floor of the LIMB island and will house the
119
-------
4160 and 480 VAC load centers, motor control centers and lighting panel. A grounding grid, 120
VAC utility outlet and control power system, electrical heat tracing, and indoor and outdoor lighting
is also included.
Sootblowers
The design allows for the addition of four new sootblowers in the convection pass. These,
plus more frequent blowing cycles, are expected to keep the pendant section and convection pass
surfaces clean when injecting lime. The existing controls will be upgraded to add the four new
blowers in the blowing sequence.
Heating. Ventilating, and Air Conditioning
The LIMB island enclosure and the humidification enclosures will be heated and ventilated.
The LIMB control and electrical equipment rooms will be air conditioned.
Miscellaneous
A sump is located in the LIMB island to collect ash and lime truck spill wash-down water,
atomization water tank overflow, and miscellaneous drip and floor drains, not including sanitary,
roof downspouts, or storm drains. Two 100% 15 HP vertical sump pumps will pump the ash/lime
waste water into a disposal system by others.
OPERATIONAL CONSIDERATIONS
When a LIMB/humidification system is added to a utility unit, its effect on existing
operations needs to be considered both from the standpoint of operating the equipment and
maintaining it. The LIMB addition is of a magnitude and complication that significantly increases
the quantity of machinery that has to be started, stopped, adjusted, controlled, monitored, and
maintained.
Effect on Equipment Operating Requirements
To minimize the impact of this addition in the existing boiler control room, the routine
120
-------
operator duties added have been assigned to the "LIMB operator" in the LIMB island control room.
The boiler operator will retain responsibility for overall control of operations by initiating the start or
stop of lime furnace injection, humidification, and reheat as total systems. This will relieve him/her
of concerns about individual machines or subsystems. Lime silo loading and LIMB ash collection
and unloading are totally within the control of the LIMB operator.
LIMB Operator Responsibilities
To bring the responsibilities into perspective, listed below by systems are the electrically
and pneumatically driven equipment under the LIMB operator's direct supervision.
Lime Silo Loading
Vent filter/fan (1)
Miscellaneous
Sump pumps (2)
Ash Disposal System
Vibrating bin bottom (1)
Rotary feeder (1)
Ash conditioner (1)
Ash Collection System
Mechanical exhauster (2)
Pulse jet bag filter with double dump discharge (1)
Cyclone separator with double dump discharge (1)
Diverter valves (2)
In addition, other operating elements requiring operator surveillance include filters, tanks/silo levels,
and control valves.
All of the above items would be checked during the LIMB operator's walk-down of
equipment which would be a normal scheduled activity when he/she was not required in the control
121
Lime Feed System
Aeration air blower (2)
Aeration air dryer (1)
Rotary shut-off valve (2)
Rotary feeder (2)
Lime pumps (2)
Transport air compressor (2)
Humidification System
Atomization water pump (2)
Atomization air compressor (2)
-------
room. The boiler operators would add check-outs of the booster air fan, humidifier and reheater to
their walk-down list.
Lime and Ash Hauling
Another consideration is the increase in the number of trucks, both delivering lime and
removing ash, from about 5 to 18/day for the base conditions described in the economic analysis
to follow. This analysis assumes the use of hydrated lime at a Ca/S ratio of 2.0 while burning a
2% sulfur, 8.7% ash coal at a capacity factor of 65%. The additional 13 trucks (5 for lime and 8
for LIMB reaction products) reflect 5 day/wk receiving and unloading using 25 ton trucks, with silo
storage capacity designed to accommodate weekends and short holiday periods.
MAINTENANCE CONSIDERATIONS
The addition of the LIMB equipment would increase maintenance. Whether additional
manpower is required depends on the workload of the existing crew. In the present climate of
minimizing costs, it is most probable that additional personnel will be required, but their activity
would not be limited to the LIMB building. The existing maintenance crew would have to be
increased by an estimated one or two persons.
Maintenance Access
Included in the design are monorails for the hoisting and removal/replacement of all
equipment or parts that could not be manhandled or reached by mobile lifters in the plant's
inventory. The atomizing air compressors will have removable panels in the sound enclosure for
maintenance and hoist access, with a monorail system carrying equipment outside for truck
loading. Under the lime silo a maintenance hatch is located in the main floor for lowering
equipment by a monorail/hoist system onto a truck. A truck can be driven into the building on the
ground floor level through "garage" doors in the enclosure sidewall.
Design parameters would ensure sufficient access to allow normal inspection, lubrication,
and upkeep (such as repacking pump glands) of the equipment. Access platforms will be located
accordingly.
122
-------
Maintenance Facilities
The plant machine or maintenance shops are assumed to be adequate to handle any
additional maintenance imposed by this design. There is, however, space on the LIMB island
ground floor, within the enclosure, for bench-mounted and limited other equipment should the plant
elect to provide local tooling or maintenance parts storage.
Replacement Parts
Normal replacement parts are not included in the scope of the material supply. It is
assumed the plant's normal spare parts inventory control system will procure what is necessary.
The design will include a comprehensive listing of spare parts recommended by the manufacturers
for holding in inventory. In addition, a complete listing of all parts will be included in the operating
instructions with each vendor's item.
ECONOMIC EVALUATION OF A 150 MWe COMMERCIAL SYSTEM
Evaluation of potential retrofit technologies by the boiler owner must cover capital cost and
operating cost. In addition, unit availability and loss of net generating capacity are equally
important considerations. The LIMB demonstration has shown the process to be highly reliable,
and the simplicity of the commercial design will increase its reliability. It also requires significantly
less electrical load for its operation as compared to wet scrubbers. Specifically the following
operational impacts were considered in the economic evaluation:
• Power required by process equipment such as fans, pumps, and compressors results in a
direct reduction of electrical generation capacity. For the minimal humidification case, this
was quite small. For the close approach to saturation, the size and power requirement of the
compressor were significant.
• Transport and booster air utilized for injection of sorbent results in a 3.4% increase in the
amount of flue gas through the convective heat transfer area of the boiler. This additional
flow decreases boiler efficiency by approximately 0.23%. This efficiency loss and the
additional steam used for sootblowing have been factored into the economic comparison.
• LIMB technology also has an indirect impact on boiler performance as a result of the
increased particulate flow through the furnace. This is reflected in the overall loss in heat
123
-------
rate that occurs as a result of increased sootblowing requirements and/or an effective
lowering of the overall heat transfer efficiency. The experience at Edgewater showed that
the injection of sorbent resulted in a substantially higher deposition rate on the boiler
convective surface. The capacity of the existing sootblowing system was insufficient to
maintain acceptable cleanliness. The resulting lower heat transfer rate increased the air
heater outlet gas temperature about 50 °F. However, the material was found to be easily
removed from the surfaces. Based on the Edgewater operation, it is felt that the addition of
sootblowers and/or increased blowing frequency can maintain surfaces at the same
cleanliness as that experienced prior to injection.
• Other indirect effects result from the use of low NOx burners. The largest effect here is
related to the increased combustion air pressure drop, on the order of 1 in WC across the
burner. In many units there is sufficient margin in the forced draft fans to accommodate this
increase, so only the increased power consumption is involved.
• The other indirect effect is related to the nature of the burner. With all else equal, the
unburned carbon content of the ash increases slightly due to the lower flame temperatures
inherent in the design. This effect is regarded to be disappearingly small, especially when
considered in the light of the potential for increased efficiency. This results from the
inclusion of slide disks within the DRB-XCL™ burner design, allowing significantly improved
control of the air flow distribution.
• Several aspects of LIMB and humidifier installation affect the induced draft system. The
booster and transport air add additional gas flow, as does the atomizing air, shield air, and
the atomizing water. The additional flue gas flow for the two LIMB systems is factored into
the economic analysis, in addition to the 1.5 in WC increased draft loss from the flue gas
reheater for the 20°F ATas system. This was 0.8 in WC in the boiler and air heater, and 0.3
in WC and 1.0 in WC for the minimum and maximum humidification systems, respectively.
• For the purposes of the economic comparison, it is assumed that the induced draft fans have
sufficient margin in their design. At specific installation sites the boiler may be I.D. or F.D.
fan limited, which would add to the capital cost of installation of a LIMB system. At these
installations the fans may have to have tips added and/or motors replaced.
The additional gas flow due to sorbent injection has a minimal impact on boiler and air
heater performance. The expected changes in boiler operating parameters are summarized in
Table 15.
124
-------
TABLE 16. EFFECT OF UMB ON BOILER OPERATIONAL PARAMETERS
Without UMB
Furnace exit qm temperature, °F 2192
Rim qm leaving economizer, Mj/hr x 10^ 1427
Qm temperature leaving economizer, °F 640
Boiler efficiency, % 88.17
Boiler draft Iom, in WC 4.44
Air heater draft Iom, in WC 6.38
Reheat spray flow, Ib/hr x 16.8
Superheat spray flow, IM* x 10*9 29.2
Maximum qm velocity, ft/sec 63.9
With LIMB
2190
1476
646
88.94
4.77
6.80
12.9
40.4
66.3
ECONOMIC COMPARISON OF LIMB WITH WET FLUE GAS DESULFURIZATION
Results of an economic comparison of LIMB and wet FGD are presented here.
Estimates of LIMB costs were each made by B&W and EPA. Estimates of FGD costs were
made by EPA. The framework for the economic analysis is a 150 MWe unit firing a 2.0%
sulfur coal. Specifics of the economic and technical premises used for the base case and for
alternate scenarios are provided in Tables 16 and 17, respectively. The computational
methodology used by B&W and EPA follows procedures suggested in the Electric Power
Research Institute's (EPRI) TAG™ Technical Assessment Guide,™ where those interested can
find the detailed definitions and equations used in this analysis.
The B&W cost estimates were derived from vendor quotations for components, based
on the system design described earlier in this section. A B&W computer model using these
values together with appropriate algorithms, allowed for estimating variations from the base
case. The EPA cost estimates were similarly made using its fourth version of a computer model
called Integrated Air Pollution Control Systems (IAPCS-4). This cost estimating model has been
developed and upgraded under EPA sponsorship over a period of about ten years. It provides a
modular approach for estimating the cost of various individual control devices and/or
combinations thereof. The base case LIMB design included a humidification system that
provides gas conditioning to a level sufficient to restore ESP performance. An alternate case
evaluated a separate humidification system capable of cooling the flue gas to within 20 °F of
the adiabatic saturation temperature with a stack reheat system to restore plume buoyancy.
The wet FGD case is based on a limestone forced oxidation (LSFO) system.
125
-------
TABLE 16. EPA ECONOMIC PREMISES FOR LIMB AND FGD COST ANALYSES
UMB
Wbt FGD
Reference date of coet estimate
July 1900
July 1990
Unit book life, yr
IB
16
Tax life, yr
IB
16
Levelling factor for O&M*
1.37
1.37
Levelling factor for 1st year carrying charges*
21.20
21.20
Leveltzing factor for 15 yr carrying charges*
19.39
19.39
Indirect coats, % of total direct capital
General facilities
20.0
10.0
Engineering
10.0
10.0
Protect contingency
17.B
17.3
Process contingency
2.B
2.6
Construction period, yr
1
2
Factor for allowance for funds used during construction (AFDC)
1.00
1.03
Baee hydrated calcific lime coat, $/ton
60
NA""
Alternate hydrated calcitic lime coet, $/ton
70
NA
Calcific limestone coet, $/ton
NA
16.37
Electricity, mill/Kwh
60
60
Solids disposal, $/ton
8.60
4.14
(dry)
(wet)
Water, */1000 gel
0.66
0.66
Steam, «/1000lb
3.60
3.60
Labor rate, $/hr
20.60
20.60
Land, $/acre
6660
6660
O&M « Operation end maintenance.
* Conetant dollar analysis.
* Current dollar analysis.
Not applicable.
Table 18 provides a summary of the major categories that comprise the estimates of
capital cost for the base case. The B&W and EPA estimates compare quite favorably, with the
$93/kW EPA estimate of the total capital requirement being about 9% lower than the $102/kW
B&W estimate. A couple of differences contribute to this. First, the EPA model appears to
underpredict the humidification system. EPA is aware of this, and plans to correct this cost
module as part of a general model upgrade. Second, differences appear in the category of
inventory costs. Here, a contributor to the difference is due to the extent of sorbent inventory,
30 days for B&W versus 90 days for EPA. However, B&W and EPA agreed that the estimates
126
-------
TABLE 17. EPA TECHNICAL PREMISES FOR UMB AND FGD COST ANALYSES
UMB
Wet FGD
Baee unit see, MWe
160
160
1st alternate unit size, MWe
100
100
2nd alternate unit etze, MWe
300
300
Net boiler heat rate, Btu/kW
9926.7
9926.7
Baee unit capacity factor, %
66
66
Alternate unit capacity factor, %
40
40
Heat irput 9 160 MWe, Btu/hr
1489.0
1489.0
Coal input 0 160 MWe, ton/hr
61.9
61.9
Baee coal sulfur, wt %
2.0
2.0
1 at alternate coal sulfur, wt %
1.6
1.6
2nd alternate coal eulfur, wt %
2.6
2.6
Coal heating value, Btu/lb (aa received)
12020
12020
Coal aah, wt %
8.70
8.70
Baee system SO, removal, %
64
64
Rue gas bypassing 90% removal FGD absorber, %
NA*
29
Baee case S02 removed, ton/yr
8803
8803
Alternate system S02 removel, %
72
72
Flue gaa bypeeeing 90% removel FGD abeorber, %
NA
20
Ca/S stoichiometry, mol Ca fed/mof S02 inlet
2.0
NA
Ca/S stoichiometry, mol Cs fed/mol SOa inlet
NA
1.1
FGD sparing, %
NA
100
FGD liquid-to-gas (UG) ratio, gel/1000 ft9
NA
96
FGD retrofit factor
NA
1.46
NA = Not applicable.
were sufficiently close so as to proceed with a sensitivity study to evaluate the effect of other
variables that would influence LIMB costs. These included: coal sulfur, unit size, capacity
factor, removal efficiency, sorbent cost, and 15 year levelized current dollar analysis. The
results of this effort are discussed later in this section.
Operation and maintenance (O&M) costs and first year and levelized annual revenue
requirements were also estimated by B&W and EPA for the base case. These are shown in
Table 19. Again, the overall agreement is quite good, with the EPA estimate of $925/ton of
S02 removed being about 2% higher than the $905/ton B&W estimate. Differences in the
127
-------
TABLE 18. SUMMARY OF CAPITAL COSTS FOR THE BASE CASE IB&W AND EPA COMPARISON)
B&W
EPA
Direct cost*, $
LIMB System: includes reagent receiving end transport,
injection system, eeotbiowers and ash handling, controls,
engineering, freight, and contingency for corrective actions
and design changes
7,349,700
7,167,400
Gas conditioning: indudee humidifier, controls, engineering,
freight, and contingency for corrective actions and design
changes
2,1»4,200
1,170,000
Subtotal direct capital
9,643,900
8,327,400
Indirect costs, $
General fecilitiee 9 20%
1,976,600
1,666,600
Engineering 9 10%
964,400
832,700
Project contingency Q 17.6%
1,670,200
1,466,600
Procees contingency 9 2.6%
236,600
208,700
Subtotal indirect coete
4,839,800
4,163,400
Total plant cost, $
14,383,700
12,490,800
Other capital costs, $
Royalties
0
0
Preproduction costs
707,700
707,400
Inventory costs
260,300
723,600
Catalyst
0
0
Land
0
21,700
Subtotal other capital costs
966,000
1,462,600
Total capital requirement, $
16,361,700
13,943,400
Equivalent unit capital requirement, $/kW
102
93
O&M estimates are primarily due to assumptions built into the respective cost estimating
models. For example, the EPA estimate assumes a slightly higher level of operating
manpower. Since maintenance and supervision costs are derived from the total plant cost,
differences are reflected here.
In the area of consumables, the EPA model does not include steam use for the
additional sootblowers needed to maintain surface cleanliness with sorbent injection. The EPA
128
-------
TABLE 19. SUMMARY OF OPERATION AND MAINTENANCE COSTS FOR THE BASE CASE (B&W AND EPA COMPARISON)
Fixed O&M costs for operating and supervisory labor, t/yr
B&W
EPA
Operators
384,000
464,800
Maintenance
Labor
162,700
199,800
Material
229,100
299,800
Supervision
161,100
196,400
Subtotal fixed O&M coats
928,900
1,160,800
Variable operating coats for consumables, $/yr
Steam use for sootblowers
96,600
na"
Hydrated calcitic Hme {B&W 0 34,310 and EPA 0 34,200 ton/yr)
2,068,600
2,062,000
Water (B&W © 6,238,000 and EPA Q 62,100,000 gal/yr)
3,400
33,900
Electricity (B&W 9 242 and EPA Q 196 kW)
68,800
66,600
Waste disposal (B&W © 76,218 and EPA © 79,100 ton/yr)
646,700
680,100
Fty ash disposal credit
(181,800)
NA
Boiler efficiency loss
22,600
NA
Subtotal variable operating coat
2,714,700
2,821,600
First year costs, $
Capital carrying charges © 19.39%
2,976,700
2,703,600
Fixed O&M
926,900
1,160,800
Variable O&M
2,714,700
2,821,600
Total first year cost
6,618,300
6,676,800
Levelized cost, $/yr
Capital carrying chargee 9 19.39%
2,976,700
2,703,600
(Fixed + variable O&M coat) x 1.37
4,988,900
6,442,100
Total 16 yr levelized annual coat
7,966,600
8,146,600
Levelized cost, mill/kWh
Capital carrying chargee © 19.39%
3.49
3.17
(Fixed + variable O&M cost) x 1.37
6.84
6.37
Total equivalent levelized cost, mill/kWh
9.33
9.64
Total equivalent levelized coat, $/ton S02 removed
906.
926.
NA «= None aeeumed.
model indicates a higher water usage for humidification and this is thought to be related to
assumptions regarding the approach to saturation temperature for the humidification module
129
-------
(which is planned for upgrading). Minor differences are shown for electricity usage to support
LIMB system requirements. The minor differences in waste disposal charges result from slightly
different estimates of the amount of material to be handled annually. Furthermore, B&W
provides a credit for disposing of the coal fly ash that is a component of the LIMB waste. This
approach correctly provides for the net increase in disposal charges between the conventional
fly ash and the LIMB waste. The EPA model assumes that there is no significant boiler
efficiency loss. First year and levelized annual revenue requirements are computed based on
the fixed O&M costs, variable operating costs, and the total capital requirement.
As mentioned earlier, B&W and EPA each used their cost estimating methodology to
perform a sensitivity analysis of key parameters that determine capital and operating costs,
both directly and in terms of the cost per ton of S02 removed. The effect of changing coal
sulfur and unit size on the total capital requirement (expressed as $/kW) are shown in Table 20
and Figure 42. All other model parameters were held constant for this analysis. The two
analyses agree well, within 10% of their average. The effect of increasing unit size is apparent
when going from 100 to 300 MWe with the 2.0% sulfur coal. This change accounts for
overall capital requirement decreases of 35 and 41 % in the EPA and B&W analyses,
respectively, and reflects the economy of scale. Similarly, a range of coal sulfur, where LIMB is
thought most likely to be used, was selected for the sensitivity analysis. These results show
that the effect of increasing coal sulfur on capital requirements is not quite as pronounced as is
that of unit size. For example, increases of 5 and 10 $/kW are shown for the 150 MWe unit in
TABLE 20. CAPITAL SENSITIVITY («/kW) TO UNIT SIZE AND COAL SULFUR (B&W AND EPA COMPARISON)
Unit Size, MWe
Coal S. wt % Source 100 150 300
1.6 EPA NA* 88 NA
B&W NA 100 NA
2.0 EPA 113 93 73
B&W 128 102 76
2.6 EPA NA 98 NA
B&W NA 106 NA
NA = None enelyzed.
130
-------
150
130
I 110
<13
'3
O"
a>
1 90
"q.
CD
o
75
i° 70
50
I I
I I I
I I
—
—
o
1.5% S
—
—
•
—
— \
A
~
2.0% S
—
— \ \
—
2.5% S
—
— \
\i ¦
—
—
d\
—
V.N. B&W
—
—
—
EPA
—
I I
I I I
I I
—
50 100 150 200 250 300 350 400
Unit Size, MWe
Figure 42. Capital cost sensitivity to unit size and coal sulfur (base case conditions).
131
-------
the B&W and EPA analyses, respectively, as coal sulfur is increased from 1.5 to 2.5%. This
reflects the additional equipment and indirect costs associated with handling the increased quantity
of solids and sorbent (at a constant Ca/S stoichiometric ratio of 2.0) as coal sulfur is increased.
A similar analysis was performed to evaluate the effect that changing unit size and coal
sulfur has on S02 removal cost. In Table 21 and Figure 43, the economy of scale is again evident,
with overall decreases of 25 and 29% shown for EPA and B&W, respectively, as unit size is
increased from 100 to 300 MWe for the 2.0% sulfur case. Similar levels of change in the cost per
ton of S02 removed are seen as sulfur content is varied for the 150 MWe case. The EPA and B&W
analyses indicate an overall 23 to 26% decrease, respectively, as coal sulfur is increased from 1.5
to 2.5%. Thus, the increased tonnage of sulfur dioxide removed more than offsets the incremental
capital and operating costs associated with the higher sulfur coals, and this results in lower cost
per ton removed.
TABLE 21. SENSITIVITY OF COST PER TON OF SO, REMOVED TO UNIT SIZE AND COAL SULFUR (B&W AND EPA COMPARISON)
Unit Size. MWe
Coal S. wt % Source 100 160 300
1.6 EPA NA* 1081 NA
B&W NA 1076 NA
2.0 EPA 1060 926 796
B&W 1049 904 760
2.6 EPA NA 833 NA
B&W NA 796 NA
' NA« None analyzed.
In addition to evaluating the effects of unit size and coal sulfur, EPA used IAPCS-4 to
evaluate further the effects of capacity factor, higher removal efficiency, and increased sorbent
cost on the capital requirement, the first year and levelized annual requirements, and first year and
levelized cost per ton of S02 removed. Furthermore, this analysis was expanded to include the
LSFO wet FGD process which could compete with LIMB as a candidate in an S02 control strategy.
132
-------
1200
o
cn
O
o
TJ
0)
1100
TJ
0)
>
o
E
CD
g 1000
c
o
•w
1.5% S
900
800
700
0
100 200
Unit Size, MWe
300
400
Figure 43. S02 removal cost sensitivity to unit size and coal sulfur (base case conditions).
133
-------
These types of analyses have been performed in the past by a variety of government and private
sector organizations. Typically, a 50 to 60% S02 removal LIMB system has been compared to a
70 to 90% S02 removal (i.e., New Source Performance Standard) FGD system. For these
scenarios, LIMB has been shown to compete favorably with FGO, especially with regard to capital
requirements and the cost per ton of S02 removed. LIMB'S economics, relative to both wet and
dry FGD systems, generally improve with decreasing:
• Boiler size
• Coal sulfur content
• Capacity factor
• Remaining unit life
• Ease of retrofit and space requirement
• Sorbent cost
For this current analysis, a slightly different approach was taken in that the removal
efficiency for both LIMB and LSFO was set at 64 and 72%. The rationale was that the cost
information developed might be useful for those applications or utilities requiring or considering
moderate to intermediate levels of S02 control, but at lowest cost for compliance. This approach,
therefore, provides another perspective of control costs, but at equivalent removal levels. As
shown earlier in Table 17, the LSFO system uses a 90% control device with flue gas bypass to
achieve the required net 64 and 72% removal. Other characteristics of the LSFO system are listed
in this table. The LIMB system achieves 72% removal with the use of the lignosulfonate-doped
hydrated calcific lime at a Ca/S ratio of 2.0. Additionally, this higher removal system incorporates
a humidification system capable of cooling the flue gas to within 20 °F of the adiabatic saturation
temperature upstream of the ESP. A stack reheater is used to restore plume buoyancy.
The results of the comparative analysis of LIMB and LSFO economics are provided in Table
22. As a reference point, Table 22 includes costs for a full 90% S02 removal NSPS LSFO system.
Graphical examples for base case conditions are depicted in Figures 44 to 47 which show the
influence of unit size, coal sulfur, capacity factor, and removal efficiency on the total capital
requirement and the cost per ton of S02 removed. Examination of these results indicates that the
capital requirement for LIMB is typically 30 to 34% of that for LSFO. The exception to this is the
72% removal LIMB system with its attendant higher capital cost resulting from the larger humidifier
needed to cool the flue gas to within 20°F of the adiabatic saturation temperature. This higher
performance system's capital requirement is about 43% of that of the LSFO FGD system's. LIMB
134
-------
TABLE 22. EPA COMPARISON OF LIMB AND LIMESTONE FORCED OXIDATION FGD ECONOMICS
16 year Levelized
Annual Coat
15 year Levelized
Annual Coat
LIMB aa % of LSFO FGD
16 yaar Level izad
Annual Coat
1.5 wt % auffur
2.6 wt % eutfur
100 M We
300 MWe
40% Capacity factor
72% S02 removal
00% SO] removal
$70rton a or bent
1at yaar coat
Capital
Coat
»/kW
93
88
98
113
73
93
136
NA"
94
93
mill/kWh
9.64
8.36
10.73
10.92
8.19
12.66
13.20
NA
10.12
8.11
$/ton SO,
removed
926
1081
833
1060
796
1229
1138
NA
982
787
Capital
Coat
$/kW
300
296
308
373
216
300
317
378
NA*
300
mill/kWh
16.26
16.87
16.78
20.10
11.86
24.91
17.16
20.92
NA
16.68
$/ton SO,
removed
1676
2047
1299
1947
1149
2413
1476
1443
NA
1609
Capital
Coat
*fkW
31
30
32
30
34
31
43
NA
NA
31
mill/kWh
69
63
64
64
69
61
77
NA
NA
62
t/ton SO,
removed
69
63
64
64
69
61
77
NA
NA
62
NA - Not applicable.
-------
2250
2000
1750
1500
1250
1000
750
500
250
- H
5 LIMB $/kW
7} LSFO $/kW
S3 LIMB levelized $/ton
E LSFO levelized $/ton
I3_Hi
pi ^
100
T
150
Unit Size, MWe
300
Figure 44. Total capital requirement arid S02 removal cost sensitivities to unit size
(base case conditions).
136
-------
"O
£
0
>
0
E
0
4->
u.
c
CN
0
O
E
CO
0
c
'3 ~a
0
4—1
D"
c
0
cu
CC
•>
1c
4->
to
O
'o.
U
CO
"D
0
0
N
75
'^Z
0
.O
>
1—
0
_l
2500
2250
2000
1750
1500
1250
1000
750
500
250
T
1.5
G LIMB $/kW
3 LSFO $/kW
E3 LIMB levelized $/ton
™ LSFO levelized $/ton
I1
JH
2.0
Coal Sulfur, %
2.5
Figure 45. Total capital requirement and S02 removal cost sensitivities to coal sulfur
(base case conditions).
137
-------
3000
£
0)
>
o
E
H
0
2000
1000
0
Si LIMB $/kW
2 LSFO $/kW
IS LIMB levelized $/ton
5 LSFO levelized $/ton
jn_
40
65
Capacity Factor, %
Figure 46. Total capital requirement and S02 removal cost sensitivities to capacity factor
(base case conditions).
138
-------
2000
1750
1500
1250
1000
750
500
250
0
M
~ LIMB $/kW
~ LSFO $/kW
£3 LIMB levelized $/ton
S LSFO levelized $/ton
22
i
64
72
S02 Removal, %
Figure 47. Total capital requirement and S02 removal cost sensitivities to removal efficiency
(base case conditions).
139
-------
is known to be somewhat more sensitive to reactant cost than are other FGD technologies. Under
base conditions, a 17% increase in sorbent cost (from $60/ton to $70/ton) translates into an
incremental S02 removal cost increase of approximately 6%, indicating that LIMB can continue to
show economic advantages over LSFO FGD technology even with a somewhat higher priced
sorbent.
In conclusion, this economic analysis shows that LIMB technology has favorable economics
over a fairly wide range of conditions that typically influence both capital and overall annual
revenue requirements. These advantages should be factors in considering LIMB technology in
utility acid gas emission control strategies. The results of this analysis continue to be consistent
with those reported in other studies where various S02 control technology economics are
compared. Similar conclusions can be made to those mentioned earlier regarding LIMB's economics
improving relative to FGD with decreasing boiler size, coal sulfur content, capacity factor,
remaining unit life, ease of retrofit, and sorbent cost.
140
-------
SECTION 9
SUMMARY AND CONCLUSIONS
The primary conclusion reached during the demonstration project is that significant
reduction of S02 and NOx emissions, either individually or in tandem, can be achieved by the
application of LIMB technology on full-scale, coal-fired utility boilers. In addition, the adverse
impact of sorbent injection on the control of particulate emissions by an ESP can be overcome by
humidification of the flue gas. Furthermore, humidification results in enhanced S02 removal when
control is maintained at a close approach to the adiabatic saturation temperature. The tests
conducted also show that the use of a lignosulfonate-doped hydrated calcific lime, produced in bulk
by a commercial supplier, improves the overall S02 capture when compared to that achieved with
the unmodified sorbent. Specifically, the use of the lignosulfonated lime combined with operation
of the humidifier at a 20 °F approach to saturation resulted in S02 removal efficiencies of up to
72% at 2.0 Ca/S, far in excess of the 50% program goal. Even without the lignosulfonate
additive, up to 65% capture was possible under the same conditions. Corresponding removal
efficiencies of about 62% and 55% were obtained when only minimal humidification was employed
to maintain ESP performance and all S02 capture occurred in the furnace.
The tests were conducted in such a way that the S02 removal capabilities have been
characterized over a range of stoichiometric ratios under the various conditions. At the same time,
an average NOx emission level of 0.48 lb/10s Btu was achieved over months of operation with the
DRB-XCL™ burners installed to meet a goal of 0.5 lb/106 Btu or less. There was no evidence of
interactive effects between the S02 and NOx control techniques, that is, sorbent injection and/or
humidification appeared to have no effect on NOx emissions, nor did the use of the DRB-XCL™
burners result in S02 emissions other than what would have been expected from combustion of the
coal used. Particulate emission control was effectively restored by modest levels of humidification
as indicated by opacity measurements at the ESP outlet. Particulate mass emission levels of 0.01
to 0.02 lb/10e Btu were found to be far below the 0.1 lb/106 Btu goal, although the tests had to be
conducted at conditions representing approximately 75% of the full boiler load.
The overall effects of LIMB technology on boiler and plant operations are regarded as
141
-------
manageable when viewed from the perspective of what a future commercial installation may
involve. Some were perhaps more pronounced at Edgewater than they might have been at another
plant, while others might present more difficulty in a different setting.
The single greatest impact comes simply from the increased particulate loading through the
boiler, ESP, and ash handling system. Unless restored by effective sootblowing, heat transfer
efficiency in the boiler decreases in proportion to the amount of sorbent injected. At Edgewater,
the use of the higher Ca/S ratios strained the capacity of the compressed air sootblowing system.
Since the capabilities of such systems vary from site to site, the application of the technology on
future commercial systems will require a thorough review of sootblowing capability to determine
what, if any, changes or upgrades may be needed.
In comparison with the effect of increased ash loading, the impact of increased gas flow
from air used to disperse the sorbent and to atomize water in the humidifier is thought to be quite
low, especially if only moderate humidification to perhaps 275°F is required to restore ESP
efficiency. While the data indicate an overall energy penalty of about 1.5% in boiler efficiency for
the combined impact on the Edgewater unit, it is strongly suspected that much of this can be
recovered with more effective sootblowing in future commercial installations.
The increased quantity of material associated with the LIMB technology makes additional
demands upon the ash handling system. These must also be taken into account in future
applications, although here the prime consideration is one of making sure that the material
transport system is adequately sized for the amount of ash generated.
The other significant effect of applying LIMB technology relates to the chemical
composition of the ash, and to the quicklime component in particular. The potential for deposit
formation at wet/dry interfaces must be considered since the pozzolanic properties of the ash can
result in the formation of buildups. These tend to be similar to those formed by ashes of some
lignite and western subbituminous coals which have a comparable chemical composition. The lime
component of the ash is also important with regard to potentially high pH conditions in any wetting
operation that might be used as part of the ash disposal process. Provision must be made for a
suitable neutralization system in accordance with site-specific considerations.
Finally, comparison with a limestone wet scrubbing system indicates that a LIMB system
142
-------
competes favorably on a $/ton of S02 removed basis when applied for S02 and NOx emission
control on units of approximately 250 MWe or less. Given the degree of emission control
attainable with LIMB fits within a utility's overall strategy for compliance, the relative advantages
of this technology over wet FGD include overall lower capital and operating cost, ease of operation
and maintenance, reduced outage time for installation, and a lower reduction of the plant's
generation capacity.
143
-------
REFERENCES
1. J.K. Wagner, R.A. Walters, L.J. Maiocco, and D.R. Neal, "Development of the 1980
NAPAP Emissions Inventory," EPA-600/7-86-057a (NTIS PB88-132121),
December 1986.
2. G.B. Martin and J.H. Abbott, "EPA's LIMB R&D Program - Evolution, Status, and Plans," in
Proceedings: First Joint Symposium on Dry S02 and Simultaneous SOz/NOx Control
Technologies, Volume 1, EPA-600/9-85-020a (NTIS PB85-232353), July 1985.
3. F.E. Gartrell, "Full-Scale Desulfurization of Stack Gas by Dry Limestone Injection? Volumes
I, II, and III, EPA-650/2-73-019a, b, and c (NTIS PB228447, PB230384, and PB230385),
August 1973.
4. S. Bortz and P. Flament, "Recent IFRF Fundamental and Pilot Scale Studies on the Direct
Sorbent Injection Process," in Proceedings: First Joint Symposium on Dry S02 and
Simultaneous S02/N0x Control Technologies, Volume 1, EPA-600/9-85-020a (NTIS
PB85-232353), July 1985.
5. B.J. Overmoe, S.L. Chen, L. Ho, W.R. Seeker, M.P. Heap, and D.W. Pershing, "Boiler
Simulator Studies on Sorbent Utilization for S02 Control," in Proceedings: First Joint
Symposium on Dry S02 and Simultaneous S02/N0x Control Technologies, Volume 1,
EPA-600/9-85-020a (NTIS PB85-232353), July 1985.
6. D.A. Kirchgessner, "Laboratory-scale Production and Characterization of High Surface Area
Sorbents," in Proceedings: First Joint Symposium on Dry S02 and Simultaneous S02/N0x
Control Technologies, Volume 1, EPA-600/9-85-020a (NTIS PB85-232353), July 1985.
7. B.M. Cetegen, T.R. Johnson, D. Moyeda, and R. Payne, "Influence of Sorbent Injection
Aerodynamics on S02 Control," in Proceedings: 1986 Joint Symposium on Dry S02 and
Simultaneous S02/N0X Control Technologies, Volume 2, EPA-600/9-86-029b (NTIS
PB87-120457), October 1986.
144
-------
REFERENCES (Continued)
8. J.P. Gooch and J.L. DuBard, "Evaluations of Electrostatic Precipitator Performance at
Edgewater Unit 4 LIMB Demonstration," EPA-600/7-88-020 (NTIS PB89-109177),
September 1988.
9. R.S. Dahlin, C.L. Lishawa, C.C. Clark, P.S. Nolan, and N. Kaplan, "Disposal, Recycle, and
Utilization of Modified Fly Ash from Hydrated Lime Injection into Coal-Fired Utility Boilers,"
presented at the 80th Annual Meeting of the Air Pollution Control Association, New York,
NY, June 21-26, 1987.
10. EP Toxicity Test Procedures. Code of Federal Regulations 40. Part 261, Appendix II, July
1984, p. 365.
11. L. Holcombe, A. Weinberg, R. Butler, and J. Harness, "Field Study of Wastes from a Lime
Injection Technology," in Proceedings: 1990 S02 Control Symposium, Volume 1, EPA-
600/9-91-015a (NTIS PB91-197210), May 1991.
12. L.S. Hovis, R.E. Valentine, B.J. Jankura, P. Chu, and J.C.S. Chang, "E-S0X Pilot
Evaluation," in Proceedings: First Combined FGD and Dry S02 Control Symposium, Volume
3, EPA-600/9-89-036c (NTIS PB89-172175), March 1989.
13. D.A. Kirchgessner and J.M. Lorrain, "Lignosulfonate-Modified Calcium Hydroxide for Sulfur
Dioxide Control," l&EC Research, 26: 2397 (1987).
14. D.A. Kirchgessner and W. Jozewicz, "Enhancement of Reactivity in Surfactant-Modified
Sorbents for Sulfur Dioxide Control," l&EC Research, 28: 413 (1989).
15. H. Yoon, M.R. Stouffer, W.A. Rosenhoover, and R.M. Statnick, "Laboratory and Field
Development of Coolside S02 Abatement Technology," presented at the Second Annual
Pittsburgh Coal Conference, Pittsburgh, PA, September 16-20, 1985.
145
-------
REFERENCES (Continued)
16. L.L. Smith, W.S. Pitts, R. Rush, and T. Flora, "Long-term Versus Short-term Data Analysis
Methodologies: Impact on the Prediction of NOx Emission Compliance," in Proceedings:
1987 Joint Symposium on Stationary Source Combustion NO„ Control, Volume 1, EPA-
600/9-88-026a (NTIS PB89-139695), December 1988.
17. R.F. Altman, R.L. Chang, E.C. Landham, Jr., E.B. Dismukes, M.G. Faulkner, R.P. Young,
and L.S. Hovis, "Electrostatic Precipitation of Particles Produced by Furnace Sorbent
Injection at Edgewater," presented at the EPA/EPRI Eighth Symposium on the Transfer and
Utilization of Particulate Control Technology, San Diego, CA, March 20-23, 1990.
18. TAG™ Technical Assessment Guide, EPRI Report P-6587-L, Electric Power Research
Institute, Palo Alto, CA, September 1989.
146
-------
APPENDIX A: METRIC CONVERSION TABLE
To convert from To Multiply bv
micron
m
1.000 x 10"4
in
m
2.540 x 10'2
ft
m
3.048 x 10"1
mile
km
1.609
kg
4.536 x 10'1
ton
kg
9.072 x 102
Btu
J
1.055 x 103
HP
W
7.460 x 102
acre
m2
4.047 x 103
gal
m3
3.785 x 10 3
ft3
m3
2.832 x 10"2
ft/sec
m/s
3.048 x 10'1
cfm (actual - acfm, standard - scfm)
m3/s
4.719 x 10"4
ft2/1000 cfm
m2/1000 m3/s
1.968 x 10'2
gr/dscf (68 °F)
kg/m3 (273 K)
2.288 x 10"3
lb/ft3
kg/m3
1.602 x 101
in WC (39.2°F)
Pa
2.491 x 102
lb/108 Btu
ng/J
4.299 x 102
°F C C = (5/9)(°F-32)
°F K K = 273.15 + (5/9)(°F-32)
psig Pa (absolute) Pa = 6895 (psig + 14.7)
147
-------
APPENDIX B: QUALITY ASSURANCE/QUALITY CONTROL
INTRODUCTION
The overall objective of the Limestone Injection Multistage Burner (LIMB) program was to
demonstrate the feasibility of retrofitting LIMB technology to an existing full-scale utility boiler.
Radian Corporation, under subcontract to Babcock & Wilcox (B&W), was responsible for
characterizing boiler emissions and ESP performance during the LIMB program. Greatest emphasis
was placed on performance data concerning sulfur dioxide (S02), nitrogen oxides (N0X), coal sulfur
content, and particulate matter emissions.
The test program comprised four major activities: baseline, low-NOx burner tests,
optimization, and LIMB demonstration. The baseline testing was conducted prior to any boiler
modifications. Low-NOx burner tests, to assess NOx removal with LIMB, were performed after the
new burners were installed and adjusted. Immediately following the low-NOx burner tests, the lime
injection system was optimized and the demonstration test period started.
The LIMB test program was designed and executed with emphasis on completeness and
data quality. A comprehensive quality assurance (OA) and quality control (QC) program was an
integral part of the overall test program. Specifications for the QA/QC program were detailed in
the Quality Assurance Project Plan and approved by EPA managers prior to testing. The goal of the
QA/QC effort was to ensure that the data collected were of known accuracy and precision and that
they were complete, representative, and comparable.
The EPA portion of the LIMB Demonstration Project ran from September 1984 to
June 1989. A test history of the parts of the program which involved Radian Corporation is shown
in Table B-1. This report first describes the OA objectives of data collection for the LIMB program.
This is followed by an outline of the various sampling and analytical procedures used to obtain the
data. The results of the QA/QC procedures that were carried out throughout the baseline and
demonstration phases are then presented, together with details on the annual certifications of the
continuous emissions monitoring system (CEMS). The certifications were carried out to ensure that
the instruments met EPA standards of precision, accuracy, and repeatability. In order to maintain
148
-------
TABLE B-1. TEST HISTORY
Time Period
Teat
May - June 1986
April 1987
May 1987
July 1987
May • September 1987
November 1987 - May 1988
June 1988
June 1989
Baaafine taats: Manual tempting
Initial CEMS certification
CEMS certification (f«M)
CEMS certification repeated (paaeed)
DRB-XCL" burner taata
LIMB optimization and initial teats
Humidifier installation and ahakedown
CEMS certification (paaaad)
End of EPA teat period
this standard of operation throughout the rest of the project, daily QC checks were also made on
the instruments throughout the work week.
QUALITY ASSURANCE OBJECTIVES FOR DATA PRECISION, ACCURACY, AND COMPLETENESS
The objective of the OA effort for this program was to assess and document the precision,
accuracy, and adequacy of the data collection systems including sampling and laboratory analysis.
Table B-2 summarizes the QA objectives for each major measurement parameter. Data
comparability was achieved by using standard units of measure as specified in the methods
indicated in Table B-2.
•
Continuous Emission Monitoring Quality Assurance and Quality Control Objectives
The CEMS was certified for N0X, S02, oxygen (02), and carbon dioxide (C02) measurements
three times during the test period (June 1986, May 1987, and June 1988). Certification of the
CEMS was performed according to EPA's Performance Specifications 2 and 3 in the Standards of
Performance for New Stationary Sources (40 CFR Part 60). The CEMS certifications were
conducted while the boiler was burning the 1.6% sulfur coal normally used to meet compliance.
This coal has a nitrogen content (~ 1.2%), similar to that of the 3.0% sulfur coal used as a test
coal during the baseline tests and LIMB demonstration. Concentrations of NOx in the flue gas were
expected, therefore, to be similar for both coals. Concentrations of S02 from the low sulfur coal
without control and from the higher sulfur with LIMB (at 50% sulfur reduction) were also expected
to be similar.
149
-------
TABLE B-2. SUMMARY OF ESTIMATED PRECISION, ACCURACY, AND COMPLETENESS OBJECTIVES
Perameter (Method)
Data
Precieion, % Accuracy, % CompJetenees, %
S03 (CEM)
6"
±20*
90
NO, (CEM)
6*
±20T
90
02 (CEM)
10*
±20f
90
C03 (CEM)
10*
±201
90
CO (CEM)
10"
±20*
90
THC (CEM)
10"
±20*
90
SO, (EPA Methods 6 and 6)
4*
±7*
90
NO, (EPA Method 7)
16*
± 20*
90
Particulate maae (EPA Method 6)
11*
±10*
90
SOJHjSOt (controlled condensation, IC analysie)
6"
±7*
90
Velocity/volumetric flow rete (EPA Methods 1 and 2)
6*
±10*
90
Moieture content (EPA Method 4)
20*
±10*
90
Molecular weight (07 and C02 by EPA Method 3)
4*
±10*
90
Particle size distribution (caacede impactor)
11
±10
90
Coal S content (ASTM D3177-84}
7
±10
90
Coal N content (ASTM D3170-84)
8
±10 •
90
Coefficient of variation for daily analysts of control sample, where:
fit^dard Deviation
CV
x 100%
Relative error (%) for eudit analyeee or reference method tests, where:
„ Maattirad Value - Actual Valua x
Actual Vafcie
* Precision and accuracy baaed on EPA collaborative tests.
Baeed on analytical precision.
The certification procedures consisted of the following major components:
• CEMS installation and measurement location
• Performance specifications
• Calibration drift tests
• CEMS relative accuracy tests
• Calculations
The site-specific aspects of each of these components, are discussed in the following
sections.
150
-------
ESP Outlet Sampling Location-
The ESP outlet sampling location is situated on the side of the horizontal 11.5 ft square
duct (~ 10 ft equivalent diameter) leading from the induced draft fan to the stack. The sampling
location consists of 5 vertical ports (4 in diameter) placed approximately 2 ft apart, and each
having 9 in diameter flanges. The sampling location is 80 ft (8 equivalent diameters) downstream
of the induced draft fan and 24 ft (2 equivalent diameters) upstream of a 90° bend. In accordance
with EPA Method 1, sampling was conducted with 15 traverse points in a 5 x 3 grid. The
temperature in the duct was about 350°F, and the gas static pressure was slightly positive.
CEMS Location and Installation-
The CEMS probe was located in the 11.5 x 11.5 ft ESP outlet duct approximately
8.5 equivalent diameters downstream of the ID fan and 1.5 diameters upstream of a 90° bend. No
significant S02/02 or N0„/02 stratification was experienced at this location because good gas
mixing occurred in the ID fan. To verify this condition, S02/02 and N0x/02 stratification profiles
were determined for the ESP outlet duct during the baseline period, prior to installation of the
CEMS probe. No significant stratification was found; therefore, the CEMS probe was located in the
centroidal area of the duct. Stratification was considered to be a condition identified by a
difference in excess of 10% between the average concentration in the duct and the concentration
at any point more than 1.0 ft from the side of the duct.
Performance Specifications and Calibration Drift Tests--
The calibration drift (CD) checks were conducted over a 168 hr (7 day) certification test
period. The magnitude of the CD was determined at 24 hr intervals for the 7 consecutive days.
The CD was determined at the low- (or zero) and high-level calibration gas concentrations by
comparing the CEMS response with the cylinder values. The requirement was that the calibration
must not drift or deviate from the calibration values by more than 2.5% of the span value for NOx
and S02 and by more than 0.5% (absolute) for C02 and 02.
Relative Accuracy Tests-
The CEMS were certified by conducting a series of 9 to 12 relative accuracy tests for NOx,
S02, 02, and C02 during the 7 day calibration drift test period. The relative accuracy tests were
performed using EPA Methods 3, 4, 6, and 7. The relative accuracy testing was conducted at the
ESP outlet sampling location using one of the five existing vertically oriented ports in the outlet
duct. A minimum of three points were traversed for each flue gas sample.
151
-------
Method 6 tests were conducted in conjunction with at least half of the Method 5 tests
during the baseline testing to provide reference method data for comparison with the DuPont S02
analyzer used as part of the CEMS. For the 10-month demonstration test period, relative accuracy
tests for both NOx and S02 were performed prior to the start of testing and once each quarter
through the demonstration test period. This was done to comply with the recently promulgated
40 CFR Part 60 Appendix F - "Quality Assurance Procedures for Gas Continuous Emission
Monitoring Systems Used for Compliance Determination."
Data Validation-
All measurement data were validated based upon representative process conditions during
sampling or testing, acceptable sample collection/testing procedures, consistency with expected
and/or other results, adherence to prescribed QC procedures, and specific acceptance criteria. Any
suspect data were flagged and identified with respect to the nature of the problem with validity.
Suspected outliers were tested using the Dixon Criteria at the 5% level.
Acceptance criteria for instrument response linearity checks were based upon the
correlation coefficient, r, of the best fit line for the calibration data points. The correlation
coefficiency reflects the linearity of response to the calibration gas mixtures and is calculated as:
r - n ffxy) • £*) ffy)
VlnfEx2) - (Ex)2] InfEy2) - (Ey)2]
where:
x = calibration concentrations
y = instrument response (peak area)
n = number of calibration points (x,y data pairs)
SAMPLING AND ANALYTICAL METHODS
This section describes the test procedures used by Radian Corporation. These procedures
include CEMS operation, manual flue gas sampling, and process sampling and analysis.
152
-------
Continuous Emissions Monitoring System Analyzers
A schematic of the CEMS sample acquisition system is shown in Figure B-1. The sample
for the CEMS was acquired using a heat-traced sample line which was maintained at a temperature
of at least 100°C to prevent condensation. A gas conditioner was used to remove moisture, thus
providing a dry gas stream for analysis.
Data from the CEMS instruments were collected and recorded using a microprocessor data
acquisition/reduction system. A hard copy of the reduced data was printed continuously and raw
data was stored on disc.
The CEMS instruments were used to sample carbon monoxide (CO), C02, 02, THC, N0X,
and S02. The instruments used to analyze these gases are described below.
Carbon Monoxide and Carbon Dioxide Analyses-
A Beckman Model 865-11 analyzer was used to measure CO concentrations in the flue gas.
The Beckman 856-11 is a non-dispersive infrared (NDIR) analyzer. The instrument measures the
concentration of CO by infrared absorption over a broad spectrum. A Beckman Model 865-23 was
used to determine C02 concentration. The Beckman 865-23 is also a NDIR analyzer. The typical
instrument ranges used were 0 to 1000 parts per million by volume (ppmv) for CO and 0 to 20%
(vol %) for C02.
Oxygen Analysis-
A Thermox WDG-III 02 analyzer was used to continuously measure flue gas 02
concentrations. The Thermox 02 analyzer utilizes an electrochemical cell to produce a linearized
voltage signal that is proportional to the ratio of 02 concentrations of a reference gas (usually
ambient air) and the 02 concentration of the sample. The typical range used was 0 to 25 vol %.
Nitrogen Oxides Analysis-
A Teco Model 10AR analyzer was used for NOx measurement. This instrument determines
N0X concentrations by converting all nitrogen oxides present in the sample to nitric oxide and then
reacting the nitric oxide with ozone. The reaction produces a chemiluminescence proportional to
the NO concentration in the sample. The chemiluminescence is measured using a high-sensitivity
photomultiplier. The typical range used was 0 to 1000 ppmv.
153
-------
Exhaust
Gas
Conditioner
with Pump
Slipstream Sample Gas Manifold
02
NO*
Monitor
Monitor
II
II
Recorder
Recorder
J
CO
Monitor
C02
Monitor
Recorder
Recorder
Conditioner/Converter
Microcomputer
Heated Sample Line
>>>>>>>>> >>>>>}>>}»^
S02
Monitor
nr
Recorder
I I
I I
Printer
Th
Mor
C
itor
Reco
rder
Figure B-1. CEMS sample acquisition system.
154
-------
Sulfur Dioxide Analysis-
A DuPont Model 400 analyzer was used to measure S02 concentrations on a dry basis.
The DuPont analyzer determines S02 concentration of the sample based on the absorption of
ultraviolet (UV) light in the 280 to 313 nanometer (nm) range. The typical range used was 0 to
2500 ppmv.
Total Hydrocarbon Analysis-
A Beckman Model 400 hydrocarbon analyzer was used to measure THC in the flue gas.
This instrument is a continuous flame ionization analyzer (FIA). The detector is a burner where a
regulated flow of sample gas passes through a flame sustained by regulated amounts of air and
hydrogen. Hydrocarbons passing through the flame undergo a complex ionization that produce
electrons that are detected by polarized electrodes. The THC analyzer was calibrated using
propane standards, and sample concentrations were reported in ppmv as propane. Early in the test
program it was evident that THC levels of less than 1 ppmv were typical. As a result, EPA agreed
with the recommendation to discontinue its use.
Manual Flue Gas Sampling and Analysis
Manual flue gas sampling methods were used to characterize flue gas emissions as well as
to certify the CEMS. These methods were used to determine particulate emissions, particle size
distribution, S02, NO„, 02, CO, and COz concentrations, flue gas molecular weight, and volumetric
flow rates. The particular method used to determine each of these parameters is given in Table B-
3. To indicate the frequency of the monitoring, sampling analysis frequencies are given in
Table B-3 for both the baseline and demonstration phase test series.
Process Sample Collection and Analysis
Process samples collected by Radian included coal, ESP ash, economizer/air heater ash and
bottom ash. A summary of the process samples, analyses, and sampling frequencies is shown in
Table B-4.
OA RESULTS - BASELINE TEST
The baseline tests were started in May 1986. The purpose of these tests was to
155
-------
TABLE B-3. SUMMARY OF MANUAL RUE GAS SAMPLING AND ANALYSIS
Sample or
Location
Parameter
Sampling
Method
Analytical
Method
Sampling Analysis
Frequency
(Demonstration Phase)
Sampling Analysis
Frequency
(Baselins)
ESP outlet
NO,. S03, CO,
CO„ 0„ THC
H«st>trec«d Una/
gas conditioner
Continuous monitor
24 hr/day
24 hours/day
ESP outlet/
Particulate
EPA Method 6*
Analytical balanos
2 per day per
location
4 test rune per
location
ESP outlet/
Particle size
distribution
Caacade
impact or*
Analytical balance
1 par day per
location
6 per location
ESP outlet/
Mat
Manual SO,
EPA Method $'
Titrimetrie or ion
chromatography
(IC)
2 per day (inlet) or
1 per day (outlet)
4 per location
ESP outlet
Manual NOx
EPA Method 7
Spectrophotometer
During relative
accuracy test
9-12 test runs
ESP outlet
so„ h3so4
Controlled
condensation
IC
3 per week
1 teet run
This sampling included volumetric flow rate, moisture content, and molecular weight determinations by EPA Methods 2, 3. and 4.
* This sampling included moisture content and molecular weight determinations by EPA Methods 3 snd 4.
TABLE B-4. SUMMARY OF PROCESS SAMPLE COLLECTION/ANALYSIS
Sample or
Locetion
Parameter
Sampling
Method
Analytical
Method
Sampling Analysis
Frequency
(Demonstration Phase)
Sampling Analysis
Frequency
(Bsseline)
Coal
Sampling
Semi-continuous
automatic samplers
NA*
Daily 24 hr
composite
Daily 10 hr
composite
Ash elemental
Same
ICAPES*
1 per week
2 samples
Ultimate/proximate
Same
ASTM Methods
3 per week
4 samples
Sulfur/nitrogen
Same
ASTM Methods
1 per dsy
9 samples
Aah
Sampling
Grab
NA
1 per dsy
Daily compoeite
ESP,
economizer/
air heater,
and bottom ash
Metals, anions,
sulfate, sulfite
Grab
ICAPES, IC,
Titrimetrie
3 per week
3 samples
(1 per load)
ESP ash
Resistivity
NA
In situ probe
1 per dey
1 per loed
Urrve
Metals, snions
Grab
ICAPES, IC
1 per dsy
3 per week
1 per load
Pulverized coal samples were collected from each of the four mills and composited to provide a single daily sample.
' NA - Not applicable.
* ICAPES = Inductively coupled argon plaama emission spectroscopy.
characterize the emissions from the boiler, prior to the retrofitting of the LIMB system, thus
providing a set of baseline reference data which could be used to assess the effectiveness of the
LIMB system. It was necessary, therefore, that the data were reliable, accurate, and met all OA
156
-------
objectives (see Table B-2). The results of the various OA procedures are presented in the following
sections.
Manual Gas Samples
The performance audit of the manual gas sampling procedures included two EPA methods
and two integral parts of other methods. Results for each of the systems audited are presented in
the following sections.
Analytical Balance-
The triple beam balance used for weighing impingers to determine the percent moisture in
the flue gas according to EPA Method 4 was audited with calibrated weights. The results of the
audit are presented in Table B-5. Greater than half the weights audited showed no discrepancy
between the measured and calibrated values. The measured weights which did deviate from the
calibrated values were within ±0.8 g.
TABLE B-6. TRIPLE BEAM BALANCE AUDIT
Audit Weight, g
Measured Weight, g
Absolute Error, g
0.0
0.0
0
6.0
6.0
0
10.0
10.0
0
60.0
60.0
0
100.0
100.8
0.8
160.0
160.0
0
166.0
166.4
-0.6
441.0
441.0
0
Dry Gas Meters-
A reference dry gas meter (DGM) was used to check the flow correction factor of the four
DGMs used during the LIMB project. The reference DGM is a meter dedicated specifically for
calibrations for which the calibration and handling history is well known. The results of the DGM
performance audit are summarized in Table B-6. All of the calibration flow correction factors were
within 1 % of those determined using the reference DGM.
157
-------
TABLE B-6. DRV GAS METER PERFORMANCE AUDIT RESULTS
Calibration
Dry 6m Last Row Audit How
Meter Calibration Correction Correction Relative
Number Data Factor Factor Error, %
RAC #9 2/04/86 1.0000 1.0002 0.0
RAC #2 3/19/86 1.0013 1.0064 0.4
RAC #12 0/13/86 1.0024 0.9987 0.4
RAC #10 2/18/88 1.0062 1.0147 0.9
Method 3 (Orsat Analyzer)--
The Orsat analyzer was used to determine the wet molecular weight of the flue gas
according to EPA Method 3. The performance audit was conducted by analyzing three test
atmospheres of certified C02 and C02 composition with the orsat analyzer. The result of the
performance audit is shown in Table B-7. In most cases the relative error of the certified and
measured values was less than 1.5%. All of the results were within 6% relative error which is less
than the accuracy objective of ±10%.
TABLE B-7. ORSAT ANALYZER PERFORMANCE AUDIT RESULTS
Teat
Teat
Atmosphere
Meaaured
Relative
Atmoephere
Meaaured
Relative
C03( vol %
C03. vol %
Error, %
03, vol %
Oa, vol %
Error, %
2.01
2.0
-0.6
6.01
6.1
1.6
6.99
6.6
-6.6
9.96
10.1
1.4
16.0
16.2
1.3
21.0
19.8
-6.7
Method 7 (Nitrogen Oxides)--
Method 7 was used to determine the N0X concentration in the flue gas during certification
of the NOx CEMS analyzer. The performance audit consisted of supplying a standard test
atmosphere of nitric oxide (NO) to a vented tee apparatus from which a flask sample could be
taken. Two flasks were selected at random from the batch prepared for the day's sampling and
used to collect the audit samples. The samples were recovered with the other flasks collected that
day, then express shipped to the Radian laboratory for analysis. The results for duplicate samples
indicated 99 and 102% agreement with the 719.8 ppm NO audit atmosphere.
158
-------
Method 5 Sampling Data-
Comparison of the gas flow rate, temperature, and moisture content data presented in
Table B-8 shows excellent consistency between the ESP inlet and outlet measurements and among
measurements taken at each boiler load. The comparatively high moisture content observed for
Method 5 test runs 1-10 and 0-11 may be due to analytical errors. The results of isokinetic
sampling rate calculations presented in Table B-8 indicate that all of the Method 5 tests met the
OA objective of 100 ±10%, with the exception of Method 5 test run 0-11.
TABLE B-8. SUMMARY OF EPA METHOD E FLUE GAS PARAMETERS AND ISOKINETIC SAMPLING RATE
Boiler
ESP
Rue Qaa
Rut Gee
Rue Qm
Moleture
Isokinetic
Baaeiine
Teet
Load,
Teet
Method 6
Temperature,
Row,
Row,
Content,
Sampling
Teat No.
Date
MWe
Location
Test Run
°F
ACFM
•cfm (dry)
vol %
Rate, %
BL01
6/27/86
102
Inlet
1-01
310
408,600
244,200
8.16
99.6
Outlet
0-01
347
420,600
246,200
7.82
94.4
BL04
6/30/86
103
Inlet
I-04
320
428,100
261,700
7.89
101.9
Outlet
0-04
327
440,000
260,800
B.68
98.9
Inlet
I-06
326
436,600
262,600
8.46
99.6
Outlet
0-06
326
441,300
264,900
7.67
99.1
BL06
6/03/86
101
Inlet
1-08
309
367,800
221,860
7.29
109.3
Outlet
O-08
311
383,200
236,880
7.24
103.6
BL03
6/29/86
82
Inlet
1-03
311
337,700
200,200
6.80
97.0
Outlet
0-03
296
326,700
204,400
7.91
96.6
BL08
6/06/86
82
Inlet
1-10
306
336,900
192,900
12.04
102.7
Outlet
0-10
307
331,400
202,600
8.40
106.4
BL09
6/06/86
81
Inlet
1-11
302
326,700
194,200
8.88
98.7
Outlet
0-11
312
320,900
187,600
11.76
110.4
Inlet
1-12
303
328,800
200,100
8.89
100.7
Outlet
0-12
306
336,700
201,000
7.92
102.6
BL02
6/28/86
64
Inlet
1-02
286
249,200
164,300
8.78
101.2
Outlet
0-02
281
262,100
160,100
8.66
100.3
BLOB
6/02/66
66
Inlet
1-06
276
267,100
169,600
7.68
99.3
Outlet
0-06
273
283,200
186,600
6.88
101.0
Inlet
1-07
292
264,400
164,800
7.62
97.2
Outlet
0-07
282
276,400
179,300
6.37
101.6
BL07
6/04/86
66
Inlet
l-OS
290
264,000
167,400
7.92
100.0
Outlet
0-09
283
269,600
166,600
7.71
101.4
Coal and Flv Ash Analyses
Performance audit samples of pulverized coal and fly ash were submitted to Commercial
Testing & Engineering Company (CTECo) as though they were normal process samples. The
pulverized coal matrix has a certified proximate/ultimate and mineral analysis. The audit sample for
159
-------
fly ash was a National Bureau of Standards "Standard Reference Method" sample with certified
concentrations of trace metals.
The results of the performance audit on the coal sample are presented in Tables B-9, B-10,
and B-11. All measurements for the ultimate and proximate analyses (Tables B-9 and B-10) were
within the target range of ±10% stated in the LIMB Quality Assurance Project Plan (QAPP).
Results for the mineral constituents (Table B-11) were all within the 10% range except titania
(Ti02), magnesia (MgO), sodium oxide (Na20), and phosphorus pentoxide (P206). These exceptions
all comprised less than 0.9% of the audit sample. A higher analytical variability for the lesser
constituents is not unusual.
TABLE B-9. RESULTS OF PERFORMANCE AUDIT - PROXIMATE ANALYSES OF COAL
Certified
Meaaured
Relative
Parameter
Value
Value"
Error, %
Ash, wt %
0.03
6.06
0.6
Volatile matter, wt %
42.39
44.92
6.0
Fixed carbon, wt %
61.68
49.02
-6.0
Sulfur, wt %
3.14
3.31
6.4
Heating value, Btu/lb
13,661
16,683
0.2
Dry baat*.
TABLE B-10. RESULTS OF PERFORMANCE AUDIT •
• ULTIMATE ANALYSES OF COAL
Certified
Meaaured
Relative
Parameter
Value*
Value"
Error, %
Carbon, wt %
74.18
76.38
0.8
Hydrogen, wt %
6.64
6.34
•6.3
Nitrogen, wt %
1.62
1.48
-2.6
Chlorine, wt %
0.01
0.01
0.0
Sulfur, wt %
3.14
3.31
6.4
Aah, wt %
6.03
6.03
0.0
Dry baait.
The results of the fly ash analysis performance audit are presented in Table B-12. These
samples were analyzed by Radian Analytical Services (RAS) using ICAPES for the metals. The
160
-------
TABLE B-11. RESULTS OF PERFORMANCE AUDIT - MINERAL ANALYSIS OF COAL ASH
Certified
Meaaured
Relative
Parameter
Value"
Value"
Error, *
SHica (SiOj)
36.33
34.67
-2.2
Alumina (A^O,)
20.92
20.67
*1.3
THania (TtOa)
0.86
0.08
•20.
Ferric Oxide (Fe,03)
36.20
30.80
1.7
Lime (CaO)
2.36
2.24
-4.7
Magneeia (MgO)
0.06
0.40
-29.
Potaeeium Oxide (K20)
0.80
0.89
-7.3
Sodium Oxide (Na,0)
0.30
0.34
13.
Sulfur Trioxide {SO,)
2.09
2.08
-0.6
Phoaphorua Pentoxide (Pa06)
0.11
0.10
46.
Strontium Oxide (SrO)
t
0.03
t
Barium Oxide (BeO)
t
0.00
t
Manganeee Oxide (Mrv,OJ
t
1.06
t
Weight percent ignited beaie.
* Value net certified in temple.
* Relative error cannot be calculated.
TABLE B-12. FLY ASH AUDIT RESULTS • (CAPES METALS ANALYSIS
Parameter
Measured
Value, wt %
Certified
Value, wt %
Recovery", %
Aluminum lAI+a)
10.8
14.t
77
Calcium (Ca+?)
0.864
1.11 ± 0.01
77
Iron (Fe+a)
7.28
9.4 ± 0.1
77
Potaeeium (K+)
1.43
1.68 ± 0.06
76
Megneeium (Mg+3)
0.367
0.466 ± 0.01
78
Sodium (Na+)
0.213
0.17 ± 0.01
126
Silicon (Si+4)
14.4
22.8
03
Titanium ffi+4>
0.671
0.8*
71
i« . rlftMsured Vakrt - Certified V«k«\ M
|1 ~ ( C^m«d VtkM 11 x10°
* Value not certified; reported value baaed on nonreference method or not determined
by two or more independent method*.
161
-------
measured values were reasonably near the certified values. Recovery was greater than 70% for
aluminum, calcium, iron, potassium, magnesium, and titanium. The recovery of silicon was only
slightly lower at 63%. The recovery of sodium was somewhat high at 125%.
Continuous Emission Monitoring System
The CEMS was operated throughout the certification and baseline testing period to
characterize boiler combustion conditions and emissions. Internal QC activities for the CEMS
included a certification period where it was established that the drift and relative accuracy of the
NOx, C02, 02, and S02 monitors were within prescribed limits. Additionally, daily drift checks and
QC checks were conducted throughout the certification and baseline testing periods. The results of
these internal QC efforts are presented in the following sections.
Relative Accuracy Tests-
Prior to initiation of the baseline testing, the continuous NOx, 02, and C02 monitors were
certified according to procedures outlined in EPA's Performance Specifications 2 and 3. The S02
analyzer was certified by collecting Method 6 samples during the baseline testing. Certification of
these instruments requires that the drift and relative accuracy are within prescribed limits. The
results of the relative accuracy tests were as follows.
The relative accuracy of the TECO NOx analyzer was determined using EPA Method 7 to be
14.3% (compared to an acceptable limit of 20%). Similarly, the relative accuracies of the 02, C02,
and S02 analyzers were determined by EPA Methods 3 and 6 to be 10.6%, 2.7%, and 17.9%,
respectively. All these values were well within the 20% limit.
Drift Checks-
Calibration drift checks were performed daily on the CEMS to ensure that data collected
was both precise and accurate. During the certification period both low and high concentration
calibration gases were introduced into the system to determine the drift since the previous
calibration. In order to conserve calibration gases, only a single point drift check based on the QC
gas was performed during the baseline test period. The instrument drift is defined as the difference
between the monitor value from the last calibration and the current monitor reading, expressed as a
percentage of the monitor span value. The acceptable drift for the NO,, S02, CO, and THC
monitors was 2.5% of the span value. For the 02 and C02 monitors, the acceptable limit was
162
-------
0.5 vol % 02 or C02 (absolute).
Table B-13 summarizes the average CEMS drift check data for each of the monitors. In
general, the CEMS monitors exhibited drifts which were within acceptable limits. Excessive drift
over a finite period could usually be traced to failure of an instrument component or the sample
acquisition system. The DuPont S02 and Beckman THC analyzers exhibited the largest drifts
throughout the testing period.
TABLE B-13. SUMMARY OF THE CEMS CERTIFICATION AND BASELINE DRIFT CHECK RESULTS
IBS=S=5BSa
Drift
Limits
Maan
Standard
Deviation
Maximum
Positive
Drift
Maximum
Negative
Drift
NO,, % of span
2.6
-0.1
1.0
1.7
-2.4
SOj, % of span
2.6
0.1
1.8
3.1
-3.9
CO, % of span
2.6
-0.2
0.6
0.4
-2.4
THC, % of span
2.6
-1.1
2.7
3.0
-6.0
02, vol % (absolute)
0.6
0.0
0.2
0.6
-0.3
CO,, vol % (absolute)
-0.1
-0.1
0.3
0.4
-0
Daily Quality Control Results--
As an additional QC check for the CEMS data, each monitor was challenged with a QC gas
mixture on a daily basis. The QC gas concentrations were representative of the typical flue gas
concentration for each monitor. This QC check was done following the daily monitor calibration.
The acceptance criteria for these control sample analyses was agreement between the CEMS
response and the certified value within 5% of the reference value for NOx and S02; within 10% for
CO and THC; and within 1 vol % (absolute) for 02 and C02. The average QC sample results for all
monitors during the baseline test period are summarized in Table B-14. The mean CEMS value for
each monitor is within the specified limits of the certified QC value and all monitors had a
coefficient of variation of less than 5%.
Performance Audit-
The performance audit of the CEMS was conducted by introducing test atmospheres, which
covered the range of the individual monitors, through as much of the sampling systems as possible.
This included the heated sample line, the sample conditioning system, and the glass manifold from
which each monitor receives its sample. The accuracy criteria for the performance audit was a
163
-------
TABLE B-14. SUMMARY OF THE CEMS QUALITY CONTROL SAMPLE RESULTS FOR BASELINE TESTING
Coefficient
Certified of Maximum Minimum
Specie* Mean QC Value Variation, %* Value Velue
NO, 603 ppmv 697 ppmv 1.4 623 ppmv 690 ppmv
SO, 2160 ppmv 2000 ppmv 2.2 2270 ppmv 2088 ppmv
CO 106 ppmv 100 ppmv 4.4 111 ppmv 80 ppmv
Oa 6.1 vol % 6.94 vol % 4.9 6.6 vol % 6.7 vol %
COj 11.6 vol % 11.2 vol % 3.0 12.1 vol % 10.7 vol %
CmMm* at VtoMon - SMwtotd DrMton x 100
monitor response within ±20% of the certified gas composition. The results of the performance
audit on the S02, N0X, 02/ C02, CO, and THC monitors are presented in Table B-15. All of the
emission monitors passed the accuracy criteria of the performance audit. Most of the monitors
were within at least ±10% of the certified concentrations.
QUALITY ASSURANCE RESULTS - DEMONSTRATION PHASE
The term "Demonstration Phase" is used here to describe all the testing carried out
subsequent to the baseline test. This encompasses the period of May 1987 to June 1989.
Throughout the testing period various QC checks were made. In addition to daily controls, internal
audits were carried out seven times during the 1987 to 1989 period and the CEMS was subjected
to two annual certification tests. Four external EPA audits were conducted during the
demonstration. Summaries of these QA checks and audits are presented in the following sections.
Quality Control Checks
Quality control checks were carried out by the Radian operating personnel on a daily or per
sample basis as appropriate. The particular checks, their frequencies, acceptance criteria, and
suggested corrective actions are listed in Tables B-16 and B-17 for the CEMS and analytical
operations, respectively.
164
-------
TABLE B-16. RESULTS OF THE CEMS PERFORMANCE AUDITS
Parameter
Audited
Cylinder
Number
Auditor
Actual
Concentration
Meaaured
Concentration
Relative
Error, %
SOa
AAL11738
K. Rozacky
0
-12
NA'
AAL6436
K. Rozacky
61
66
7.8
AAL8949
K. Rozacky
218
240
10.0
AAL8448
K. Rozacky
720
766
8.2
AAL15064
K. Rozacky
2470
2466
-0.6
NO,
AAL11738
K. Rozacky
0
6
NA
AAL6436
K. Rozacky
48
64
12.0
AAL8949
K. Rozacky
237
261
6.9
AAL9448
K. Rozacky
732
782
6.8
CO
AAL11738
D. Ranum
0
3
NA
AAL11609
D. Ranum
100
116
16.0
AAL8002
D. Ranum
302
303
0.3
THC
C014798
D. Ranum
0
1
NA
AAL11609
D. Ranum
21.6
21
-2.3
AAL8002
D. Ranum
69.4
69
-0.7
o,
AAL6436
K. Rozacky
0
-0.3
NA
AAL11641
K. Rozacky
3.06
2.7
-10.3
AAL6616
K. Rozacky
0.10
6.7
-6.6
AAL6616
K. Rozacky
9.96
9.4
-6.6
AAL11738
K. Rozacky
21.0
20.0
-4.8
COa
C016798
D. Ranum
0
0
NA
AAL11609
D. Ranum
6.99
7.4
6.9
AAL8002
D. Ranum
16.09
16.6
4.0
Concentrations are expreeeed in ppmv for SO,, N0„ CO, and THC, and in vol % for Oa and COa.
' NA = Not applicable.
The results of the daily calibration and drift checks carried out for each analyzer are
summarized in Table B-18. This table shows the percentage of CEMS calibration and drift checks
which met the OA criteria listed in Table B-16. It can be seen that the NOx, 02, and C02 analyzers
proved to be very reliable instruments and, therefore, almost all the data recorded by these
monitors could be interpreted as useful information.
The S02 analyzer just met the OA criteria due to difficulties that appeared to be associated
with the auto-zeroing feature of the instrument. Several attempts to correct the problem failed to
make a permanent improvement. Since calibrations were being performed daily, and sometimes
165
-------
TABLE B-10. SUMMARY OF CEMS ANALYTICAL QUALITY CONTROL CHECKS, FREQUENCIES, ACCEPTANCE CRITERIA, AND CORRECTIVE ACTIONS
Quality Control Chock
Froquoncy
Acceptance Criteria
C orractive Action
SO., NO,
Low level/high (aval
calibration drift ehacka
Daily
Difference between CEMS
raeponat and reference
valua £2.6% of apan valua
£6 conaaeutiva day* of
axceeaive drift
Adjuet V.9., aat) lew
level/high laval raaponat
Sarviea inatrument
Control aampla analyaaa
(following daily calibration)
Comparison of monitor
raaulf to raferanca method
teat reeuits
Daily
Drift £10% of apan valua
Agraamant within 6% of
certified calibration
Relative accuracy within
20% of mean valua for
reference method teat data
Sarviea inatrument
Repeat calibration
Check inatrument
calibration
Repeat reference method
02, CO,
Low level/high level
celibration drift checks
Daily
Diffarence between CEMS
raaponee and reference
value £0.5 vol % 02
(abeolute)
Service inatrument if still
unacceptable
Adjust eet) low
level/high level reaponee
£6 coneecutive daya of
exceeeive drift
Service inatrument
Control sample analyses
(following daily celibration)
Daily
Drift £10% of apan value
Agreement within 1 vol %
(abeolute) of certified
calibration
Service inatrument
Repeat calibration
Comparison of monitor
reeuits to reference method
test reeuits
Relative eccurecy within
20% of meen value for
reference method teet data
Check instrument
celibration
Repeat reference method
teats
CO, THC
Low level/high level
calibration drift checks
Control sample analyses
(following daily calibration)
Daily
Daily
Difference between CEMS
and reference value
£2.6% of the epen velue
£6 coneecutive days of
days of exceeeive drift
Drift £ 10% of apan value
Agreement within 10% of
certified concentration
Service inatrument if still
unacceptable
Adjust (/,*,, eet) low
level/high level response
Service instrument
Service instrument
Repeat calibration
more frequently, the feature was not deemed essential and was disconnected. Unfortunately, drift
problems occasionally resulted with one instance occurring during an external EPA audit by
Research Triangle Institute (RTI) in 1989. Data recorded on a day on which an instrument did not
meet OA criteria were not used unless a fair estimate of correction factors could be made. In spite
166
-------
TABLE B-17. SUMMARY OP ANALYTICAL OPERATIONS QUALITY CONTROL CHECKS, FREQUENCIES, ACCEPTANCE CRITERIA. AND CORRECTIVE ACTIONS
Parameter
Quality Control Check
Frequency
Acceptance Criteria
Corrective Action
NO, (Method 7,
spectrophotometry)
Multipoint calibration
Control sample analyeee
Daily
In triplicate with each a
of eamplae
Agreement of daily
calibration factor within
16% of factor from
previous day
Agreement of each QC
sample within 16% of
Repeat calibration
Prepare new stenderde
Service instrument
Repeat control sempte
analyeee
SOa (Method 6, titrimetric) Control sample analyses
Duplicate sample analyses
S0j/H2S04 (controlled
condensstion, ion
chromatography)
Peroxide blank analysis
Linsarity check (multipoint
calibration)
Daily, or with each set of
sample analyses
100 % of samples
Daily, or with each a
sample analyses
Twice for each set of
sample analyees
Agreement within 1 % for
duplicate analyeee
96 to 106% recovery,
beeed on the mean value
for duplicate analyeee
Agreement within 1 % or
S0.006 mmol/l, whichever
is larger
None; used in blenk
correction
Correletion coefficient
20.996
Slope within 10% of
running mean for previous
6 determinatione
Recalibrate
Repeat control sample
analyses
Root end ardize titrant
Repeat sample analyees
Not applicable
Repeet linearity check
Prepere new etanderde
Calcium by atomic
abeorption
epectrophotometry; eulfate
by ion chromatography
Control sample analyses
Duplicate sample analyses
Blank analyses (reagent)
Lineerity check (multipoint
calibration)
in duplicate with each «
of sample analyses
20% of samplee
Twice for each set of
temple analyses
Twice for eech set of
samplee analyses
Control sample snelyees In duplicate with each eet
of sample analyeee
Duplicate sample analyses 10 of samples
Agreement between
duplicates within 6% of the
mean
96 to 106% recovery
Agreement between
duplicetes within 6% of the
mean
None
Correlation coefficient
*0.996
Slope within 10% of
running mean for previoue
6 determinations
Agreement between
duplicatee within 6% of the
mean
96 to 106% recovery
Agreement between
duplicatee within 6 % of the
Repeat control sample
analysea
Recalibrate
Repeet analyeee
Not applicable
Repeat lineerity check
Prepere new standards
Service instrument
Repeat control sample
analyeee
Recalibrate
Repeet analyeee
Blank analyeea (reagent)
Twice for each eet of
sample analyses
Not applicable
167
-------
TABLE B-17 (continued).
Parameter Quality Control Chock Frequency Acceptance Criteria Corrective Action
Sulfite (titrimetric)
Duplicate eemplae
100% of samples
Agreement within 6% of
the mean
Repeat anelyeee or flag
data
Control sample anal yea*
With eeeh set of sample
analyeee
06 to 106% recovery
Repeat analyeee
Root end ardize iodine
eokition
Blank analyeee
In duplies
of sampt*
te with eeeh set
i analyeee
None
Not applicable
Met ale, anione by ICAPES
anaiyeie
Linearity cheek (multipoint
calibration)
Daily
Correlation coefficient
10.996
Repeet calibration
Adjust inetrunent
Control sample analyeee
Weekly
Meeoured value within 6%
of true vekie for aU
elements
Prepare new calibration
etenderde
Repeat calibration
Duplicate samples
10% minimum or 2 per
batch of samples
Analyeee agreement within
20 % for species of interest
Repeat sempta enelyees
Solid* moieture content
Duplicate sample*
10% of sampiee
Agreement between
duplicetee within 10% of
the mean
Repeet analyse* or fleg
deta
TABLE B-1B. ANALYZER PERFORMANCE DURING THE DEMONSTRATION PERIOD
Period
Analyzer
% Accept eble
Performence Days
06/B7-
06/89
NO,
88
06/87-
¦06/89
o3
100
06/87
¦06/89
SO,
91
06/87-
¦06/89
C02
94
06/87-
-06/89
CO
86
07/87-
-10/88
THC
66
Defined as thoee days upon which drift and celibrstion OA criteria were met.
of the problems, Table B-18 shows that enough of the daily drift checks met the required QA
standard for the S02 data collected to meet the completeness criteria given in Table B-2.
Furthermore, relative accuracy tests carried out in 1987 and 1988 as part of the CEMS certification
tests, showed that readings from the S02 analyzer were in good agreement (i.e., within QA limits)
with manual method measurements. Thus, confidence can be maintained in the data collected.
(Although time constraints precluded acquisition of a replacement monitor during the EPA-
sponsored project, a new analyzer was obtained for the DOE-sponsored LIMB Demonstration
168
-------
Project Extension that followed. Results obtained with the new analyzer under essentially
equivalent conditions appear to confirm the data reported herein.)
Table B-18 shows that QC checks on the CO analyzer did not meet the OA criteria for
completeness when averaged over the whole demonstration period. However, there were periods
{e.g., June 1987 through July 1987) when there was a run of consistently poor QC results which
heavily influence this average. Overall the monitor was found to be working within QC limits.
During one series of poor results, the monitor did not respond to normal adjustments. The problem
appeared to be related to the calibrant or contaminants in the gas bottle, as the QC checks
improved following a change of the bottle.
Table B-18 also indicates that the THC monitor failed to meet the OA criteria. This was
due to the fact that, although the actual errors were small, the amounts of hydrocarbon detected
were so small that they resulted in high percentage errors. As noted earlier, the concentrations
were so low that the THC analysis was discontinued in October 1988.
Coal quality was also the subject of QC checks, especially with regard to its sulfur content.
Pulverized coal ash samples were taken from the boiler on a regular basis (often hourly) and were
analyzed on site for sulfur content using a Leco Corporation Model SC-32 analyzer . Each
weekday, one of the samples was divided into two parts, with one half being analyzed on site, and
the other being sent to CTECo for independent analysis. Agreement between the two values was
found to be always within the 10% QA window. The agreement between the two was later
verified by the RTI external auditors in 1989.
Internal Audits
Performance and technical audits were conducted by Radian personnel as part of the QA
support effort for the LIMB test program.
Performance Audits-
The performance audit consisted of challenging the CEMS with representative reference
standards to provide a point-in-time assessment of the accuracy of each component of the system.
The CEMS is operated for the measurement of NOx, S02, C02, 02, and CO. Some audits also
covered other instruments and the coal analysis. The results of the performance audits are
169
-------
summarized in Tables B-19 and B-20. From these tables it can be seen that in the majority of
cases, the OA objectives given in Table B-2 were satisfied. Where this was not so, the data are
marked and explained as footnotes to the tables.
Technical Systems Audit-
A technical systems audit was conducted concurrently with the performance audit. The
systems audit focused mainly on analyst technique, compliance with established QC protocols
identified in the QAPP, and the documentation associated with the on-site sampling and analytical
activities. No major problems were identified which significantly impacted the quality of the data
generated by this project.
Results of Continuous Emissions Monitorino System Certification
The results of the CEMS certification testing are presented in this section for both 1987
and 1988. In both years, the CEMS was shown to comply with the certification requirements
before further emissions data were collected. The following two sections present the relative
accuracy test and drift check results, respectively.
Relative Accuracy Tests-
The continuous NOx, 02, and C02 monitors were certified according to procedures outlined
in Performance Specifications 2 and 3 in Appendix B of 40 CFR Part 60. Certification of these
instruments requires that the drift and relative accuracy are within prescribed limits. The manual
sampling for the reference methods was conducted at the ESP outlet location about 20 ft upstream
of the CEMS sample probe. The relative accuracy was determined from 9 of the 12 individual
runs.
EPA Methods 3, 6, and 7 were used to determine 02 and C02, S02, and NOx, respectively.
Table B-21 presents the results of the 1987 and 1988 certifications. All instruments met the
required QA objectives given in Table B-2.
The relative accuracy of the DuPont S02 analyzer was determined using EPA Method 6
samples collected on June 5, 8, and 10, 1987. The recovery and analysis was performed in
Radian's on-site laboratory using a barium-thorin titration method. The relative accuracy for the
DuPont S02 analyzer was 4.3%, well within the 20% prescribed limit.
170
-------
TABLE B-19. SUMMARY OF INTERNAL PERFORMANCE AUDITS SHOWING THE RANGE (MINIMUM/MAXIMUM) OF PERCENT RELATIVE ERROR VALUES
Date of Audit
Analyzer/Method 6/87 9/67 7/88 1/89 4/89 8/89
OEMS:
SO,
•11.0/-0.4
-3.7/13.8
7.9/9.2
•19.0/ -2
0.3/3.9
2.6/ 8.7
NO.
0.8/ 4.2
2.0/11.4
6.2/12.2
6.4/14
2.2/2.3
-1.9/-27.2
COa
-8.4/ 0.3
-27.3/-2.71
0.6/1.7
-6.6/ 1.7
0.0/9.1
2.9/ 4.1
CO
na"
-11.6/-2.3
-4.6/-0.3
-17.0/-14
-20.0/6.3*"
-10.6*
0,
-0.3/ 2.1
2.4/18.7
1.9/4.9
1.0/ 11
2.3/3.1
-4.7/ -0.6
THC
NA
•8.4/ 6.6
•12.0/16
NA
NA
NA
Dry Gas Meter
0.1*
NA
NA
NA
NA
NA
Analytical Balance
-0.6/ 0.1
NA
0.0/1.0
NA
NA
NA
Leco
NA
0.0/0.001
NA
NA
NA
NA
Orsat
0.0/ 3.2
NA
NA
NA
NA
NA
Sulfur Content
NA
-2.8 T
NA
NA
NA
NA
NA = Not audited.
' Single determination.
* Instrument ia not linear in response at levels below 11 % (44% of full scale). This is also ths lowest level QC gas available to the analyst.
Quality control objective met for all but one CO measurement.
ff NOj portion of sample was not being measured due to analyzer malfunction.
TABLE B-20. SUMMARY OF PERFORMANCE AUDIT RESULTS FOR MANUAL SAMPLING METHODS
Date Parameter
EPA
Method
Expected
Range of
Recovery, %
Sample
Recovery, %*
4/87 SO, as S04
(Titrimetric)
6
93-107
96
6/87 NO, as NO
7
80-120
101
NO, as NO
7
80-120
CO
7/88 Oa
3
90-110
98
COa
3
90-110
100
S03
6
93-107
100
SO,
6
93-107
99
NO,
7
80-120
64"
NO,
7
80-120
93"
" Accuracy obtsctivee specified for the method in the QAPP.
* Sample recovery, % ¦ (Reported value/Audit value) x 100%.
* Percent recovery is outside expected renge of recovery.
Initial results produced a recovery value of 64%. Although a second determination yielded a value of 234%
at first, follow-up revealed improper sample preparation, for which correction was possible to the final 93% valus.
171
-------
TABLE B-21.
SUMMARY OF CEMS RELATIVE ACCURACY RESULTS
Number
96% Confidence
Relative
Spetioe
of TMt*
Interval
Accuracy, %
June 1987
NO,
12
6.9 ppmv
13.6
so,
9
29.7 ppmv
4.3
oa
12
0.17 vol %
4.2
co3
12
0.20 vol %
1.9
Julv 1988
no.
13
8.6 ppmv
10.8
SO,
11
14.4 ppmv
11.9
0,
9
0.32 vol %
7.6
C03
9
0.22 vol %
4.6
Calibration Drift Checks-
The calibration drift (CD) checks were conducted over a 168-hr (7 day) certification test
period. The magnitude of the CD was determined once each day at 24 hr intervals. When the
boiler was taken out of service during the 7 day test period (e.g., for normal weekend outages), the
CEMS was maintained in a standby mode. When the boiler was returned to service, the CD checks
were continued. The first CEMS calibration responses after the restart were treated as day 0 data
so that the 7 day test count resumed with the next 24 hr calibration cycle. Thus, a CD test may
have required more than seven days to accumulate the required data. If the test period was
interrupted due to a failure in the CEMS, then the CD test was restarted at a new day 0 and the
previous data were discarded.
The CD was determined at the low- (or zero) and high-level calibration gas concentrations
by comparing the CEMS response with the calibrated or adjusted response recorded at the previous
24 hr calibration cycle. The difference is expressed as percent of instrument span for the N0X and
S02 monitors and absolute percent difference for the 02 and C02 monitors. A summary of the
CEMS calibration gas values is presented in Table B-22. Table B-23 contains the results of the drift
checks for both the 1987 and 1988 tests. During the 1987 certification period, all monitors
exhibited drifts which were within the prescribed limits.
172
-------
TABLE B-22. SUMMARY OF CEMS CALIBRATION GAS OATA
Calibration Gaa
Concentration
N0„ ppmv
SO,, ppmv
Oa, vol %
C03, vol %
CO, ppmv
THC, ppmv"
June 1987
Low Laval Value
0
0
0
0
0
0
High Level Value
976.3
1822
20.98'
18.62
486.3
91.1
QC Value'
403.8
1007
8.01
12.02
99.81
48.1
July 1988
Low Leva! Value
0
0
0
0
na""
NA
High Level Value
764.8
1814
21.2
18.9
NA
NA
QC Value
264.8
1009
8.08
11.0
NA
NA
Total hydrocarbons were meaeured using propane aa the standard.
* An alternate high laval eoncantration gas (20.91 %) waa uaad for days 6 and 7 of the calibration drift teat.
* QC standard* war« selected to represent approximate flue gaa coneantrationa for each parameter.
" NA - Not audited.
TABLE B-23. SUMMARY OF THE CEMS DRIFT CHECK RESULTS
Drift
Limits
Mean
Standard
Deviation
Maximum
Positive
Drift
Maximum
Negative
Drift
June 1987
NO,, % of spen
2.6
•0.3
0.9
1.0
•2.3
SOj, % of span
2.6
0.4
1.1
2.4
-1.1
0], vol % (absolute)
0.6
0.0
0.1
0.4
-0.1
C03, vol % (absolute)
0.6
0.0
0.2
0.4
•0.6
Julv 1988
NO,, % of span
2.6
-0.06
0.73
2.2
-1.0
S02, % of span
2.6
2.1
1.6
4.9
-0.72
0,, vol % (absolute)
0.6
0.02
0.3
0.7
-0.64
COj, vol % (absolute)
0.6
0.11
0.23
0.7
-0.1
During the 1988 certification period, only the N0X analyzer met the OA specification. Both
the S02, 02, and C02 monitors exceeded the acceptable drift limits. However, it has already been
shown that all analyzers met OA specifications for relative accuracy which is a more stringent test
of an analyzer performance. As was noted earlier, the drift problems were, for the most part,
overcome by daily monitoring, and when necessary, correction of the analyzers in order to ensure
that the problems did not significantly affect the collection of data.
173
-------
Apart from the CEMS certification period, daily drift checks were carried out on the CEMS
throughout the whole test period. This ensured that the instruments were maintained at the
required high standard of operation. The certification drift check data shown in Table B-23 confirm
that the instruments were maintained in good operating condition throughout the project.
External EPA Audits
Four external EPA audits were performed by Research Triangle Institute (RTI) under the
direction of the EPA/AEERL OA Manager during the LIMB Demonstration Project. A brief summary
of each of these audits follows.
• On June 2, 1986, a technical systems audit (TSA) and a performance evaluation audit (PEA)
were conducted. The results of this audit concluded that the tests were being conducted in
accordance with the QAPP. Minor recommendations for improving operations were made.
• On September 16-17, 1987, a second TSA was conducted. Recommendations included
improved tagging of equipment, more frequent manual checking of some flow rates
calculated by the data acquisition system, and the use of improved stripchart recorders for
some of the CEMS data.
• On May 8-9, 1989, a PEA and an Audit of Data Quality (ADQ) were conducted. The PEA
indicated satisfactory results were being obtained, although improved procedures for QC
checks were recommended. The ADQ indicated an apparent sulfur mass balance
discrepancy between the inlet and outlet values. This was not unexpected since the data
audited had not been reduced to its final form. For this reason and because the work was to
continue under the DOE-sponsored LIMB Demonstration Extension Project, a follow-up audit
was recommended. In addition, B&W and Radian decided to replace the S02 analyzer with a
new, more stable unit for the extended project.
• On October 11, 1989, the follow-up ADQ was conducted. This audit found improved sulfur
mass balances, the reasons for which the auditors thought may have been related to lower
coal sulfur variability and/or the improved QC checking procedures. The installation of the
new S02 analyzer had improved the quality of the data being generated due to improved
stability.
174
-------
[Project personnel shared the auditors' concerns regarding the data as it was being
generated, with the realization that the apparent inconsistencies could only be resolved after
full analysis and reduction of the data. The programming logic originally used in the System
140™ for much longer term trending was not readily adaptable to data reduction over what
came to be relatively short test periods during the project. For this reason, transient,
"erroneous" data {e.g., a low gas concentration associated with water blowdown from the
CEMS sample gas conditioner [chiller]) was represented disproportionately and had to be
removed manually from each test period. While this was feasible for the auditors at the time,
it was not so for the project personnel on a daily basis. The apparent inconsistencies were
eventually resolved as data reduction proceeded.]
QUALITY ASSURANCE/QUALITY CONTROL SUMMARY
QA/QC activities were an integral part of the LIMB Demonstration Project and were
undertaken at a level commensurate with the overall scope of work involved. Both internal and
external audits were routinely performed in order to assure the continued generation of complete
and reliable information. While questions arose regarding sulfur mass balance during the later
audits, these were subsequently resolved. Other audit findings generally revealed adherence to the
QAPP and resultant recommendations for improvements in equipment and procedures were
implemented in a timely manner.
175
-------
APPENDIX C: DRB-XCL™ BURNER CHARACTERIZATION TEST REPORT
INTRODUCTION
During the week of July 13, 1987, burner characterization tests were performed on Unit 4
of Ohio Edison's Edgewater Station (B&W reference RB-231) to determine actual operating
characteristics of the DRB-XCL™ burners which were retrofitted to this unit as part of the
Limestone Injection Multistage Burner (LIMB) program.
The primary objectives of this portion of the LIMB program was to confirm that the DRB-
XCL™ burners were meeting the program's goals for the burners. This testing was performed
without limestone injection in order to characterize the burners' performance as a baseline prior to
limestone injection.
Ohio Edison's Edgewater Unit 4 that is utilized as the LIMB host site is a Babcock & Wilcox
Radiant Boiler, vintage 1957. The boiler was originally designed for 690,000 Ib/hr steam flow at
1480 psig and 1000°F, providing 104 MWe. Ohio Edison typically operates this unit at 770,000 to
830,000 Ib/hr at 1500 psig. Originally, this unit was equipped with four (4) EL-64 pulverizers and
twelve (12) PC circular burners. Each mill feeds three burners, all located on the same elevation.
The unit is balanced draft. For the LIMB program, the circular burners were replaced with DRB-
XCL™ burners. No pressure part changes or other modifications were required for the burner
retrofit.
PROGRAM GOALS FOR THE DRB-XCL™ BURNER
The following goals were set as objectives for the burner performance in the LIMB program:
NOx emissions:
Burner pressure loss:
Percent carbon in fly ash:
Flame length:
0.50 lb/106 Btu
less than 5.0 in WC
less than 5.0%
less than 22 ft
176
-------
The N0X emission goal was established by the EPA for the LIMB program and the other goals were
set based on industry-acceptable levels and existing host boiler constraints.
BASELINE TEST RESULTS WITH ORIGINAL CIRCULAR BURNERS
Tests were performed during March 1986 to determine the operating characteristics of
Edgewater Unit 4 with the original equipment. This information was developed as a baseline to
evaluate the impact of a LIMB retrofit. The baseline tests provided conclusive data that the boiler
was performing as would be expected from any unit of similar design. Table C-1 shows typical
emissions/operations data from the baseline tests.
TABLE C-1. EDGEWATER UNIT 4 BASELINE TEST RESULTS
Main ateem flow 770,000 Ib/hr
O, entering air heeter 4.8 vol %
N0X emieeione 660 ppmv (dry) Q 3 vol % 0,
0.89 lb/10'Btu
Carbon in a*h 1.9-2.5%
Burner preesure loea 1.0 in WC
PRE-TEST OPTIMIZATION WORK
Prior to the characterization tests without sorbent injection, optimization tests were
performed to tune the DRB-XCL™ burner's operation. These tests were quite extensive and are
briefly discussed here. A complete description of the optimization tests will be covered in a
separate report (to be prepared at a later date). [Ed. note: This report was originally issued August
7,1987.]
The DRB-XCL™ burners were installed in Ohio Edison's Edgewater Unit 4 in August 1986
during a five (5) week outage. The burner configuration, as installed, had a shallow angle impeller
installed in each burner for the purpose of flame length control. This configuration was agreed
upon by the EPA and B&W as being the most suitable configuration for the retrofit as determined
from DRB-XCL™ burner development tests.
177
-------
Initial testing of the burner provided information that showed the burner capable of meeting
all program goals except NOx with the original configuration. NOx emissions with the impellers in all
burners was approximately 0.53 lb/106 Btu (exceeding program goals by 6%). It was agreed to by
the involved parties (EPA, Ohio Edison, and B&W) to optimize the performance of the DRB-XCL™
burners with regard to NOx with minimum impact to the other program goals. Several iterations of
burner modifications were made toward optimization with the following items being identified as
problem areas with regard to combustion performance:
• Tubular air heater leakage was approximately 40% excess air
• Convection pass and economizer hopper infiltration was very high
• Pulverizer control system did not provide consistent coal feed per mill during operation
A final burner configuration was arrived upon that all parties involved agreed was
acceptable to utilize for the purposes of the LIMB demonstration program. This configuration
utilized the shallow angle impellers on the bottom burner elevation (D mill) and the upper middle
burner elevation (B mill). The lower middle burner elevation (A mill) and the top burner elevation (C
mill) were equipped with a standard conical diffuser/deflector for coal distribution within the nozzle
and a flame stabilizing ring on the end of the coal nozzle. This arrangement provided conditions
that met all program goals (during optimization tests) except for the carbon in ash, which averaged
6.2%.
During the period of optimization tests, the optimum burner register settings were
determined. These settings were based on even air distribution to all burners in service at all loads.
Each DRB-XCL™ burner is equipped with a pitot tube grid in the outer air zone that was calibrated
prior to installation at Edgewater. Using the indicated pitot differential, it is possible to balance air
flow to all burners in service. The assumption is made that total fuel flow is evenly distributed to
all burners in service. For this installation, it was determined that proper air distribution can be
maintained at all load conditions using two positions for each burner: an "open" position for an in-
service burner, and a "cooling" or "closed" position for an out-of-service burner. Even though the
assumption of even fuel flow distribution is not consistently valid for this particular application, the
two position settings provides the optimum burner settings short of constant adjustment to
compensate for uneven fuel distribution.
178
-------
TEST PLAN
The fully randomized test conditions originally selected for the burner characterization test
series is shown in Table C-2. These test conditions represent a repeat of the baseline test
conditions.
TABLE C-2. ORIGINAL RANDOMIZED TEST SEQUENCE AND CONDITIONS
StMm Row, Ib/hr Exewa Air, %
770,000
16
610,000
20
460,000
26
610,000
26
460,000
16
770,000
20
460,000
20
760,000
26
610,000
16
To confirm the DRB-XCL™ burner operating data, tests BTOI to BT06 were scheduled to be
repeated [as BT07 to BT12] in the same order.
For the actual test series, the test matrix was rearranged to meet the daily load demands of
Ohio Edison's system. Tests were scheduled for three (3) per day and the daily tests were
rearranged to fit the load requirements. Table C-3 shows the actual test matrix utilized and
includes the tests that were repeated. The test series was stopped at BT11 due to a superheater
tube leak that took the unit off-line.
As can be noted from Table C-3, the high load condition was reduced from the original
770,000 Ib/hr steam flow to 680,000 Ib/hr steam flow. This was required during the test series
due to higher than normal cooling water temperatures that adversely affected the performance of
the condenser, limiting boiler load.
179
-------
TABLE C-3. ACTUAL TEST SEQUENCE AND CONDITIONS
Tact Number Steam How, Ib/hr Exceee Air, %
BT01
460,000
26
BT02
080,000
16
BT03
680,000
20
BT04
680,000
26
BT06
880,000
20
BT06
480,000
16
BT07
480,000
20
BT08
880,000
26
BTO0
680,000
16
BT10
460,000
20
BT11
680,000
26
TEST PROCEDURE
For each test, the following items were performed:
• Setup: Boiler set to test conditions by Ohio Edison. Excess air, mill conditions, flames and
burner settings checked by B&W.
• Data: Control room panel board data - manual
Burner settings - manual
Precipitator data - manual
Air heater inlet emissions • manual
System 140™ acquisition system:
All data points
Calculated data
Trends
Radian's emissions data (ESP out)
Dust loading for carbon in ash
Fuel sample (daily)
Precipitator and bottom ash sample (daily)
180
-------
FUEL ANALYSIS
Table C-4 shows the average analysis of four samples for the fuel utilized during the DRB-
XCL™ burner characterization test program. This is the typical fuel fired by Ohio Edison in
Edgewater Unit 4.
TABLE C-4. COAL ANALYSIS
wt % (aa received)
Proximate:
Fixed carbon 61.33
Volatile matter 36.32
Aah 9.68
Moieture 3.77
Uttimate:
Carbon 71.61
Hydrogen 4.68
Sulfur 1.42
Nitrogen 1.40
Aah 9.68
Moisture 3.77
Chlorine 0.16
Oxygen 7.38
Higher heating value, Btu/lb 12,490
TEST RESULTS
Table C-5 presents a results summary for the DRB-XCL™ characterization test series.
Oxvaen Analysis
During the test program, excess oxygen data was taken at four (4) locations:
• Economizer outlet - 6 Westinghouse in situ probes:
2 - six ft (control room and System 140™)
4 - twelve ft (System 140™ only)
181
-------
TABLE C-6. XCL BURNER CHARACTERIZATION TEST DATA SUMMARY
Typet
BT07
BT08
BT10
00
fO
Unit load, MWe
LCR panel
64
91
79
80
92
64
66
93
80
66
93
Main eteem flow, Ib/hr x 10"®
CR panel
460
680
680
680
682
468
462
680
670
466
676
Burner pressure toes, in WC
CR panel
1.8
3.4
2.6
3.1
3.7
1.0
1.4
4.6
2.1
1.3
4.7
Air Data
Total airflow, Ib/hr x 10*
S140-C24
610
784
692
N/A*
809
617
673
893
641
682
866
Excess air leaving economizer, %
S140-C22
18
8
10
N/A
13
6
11
22
7
14
21
Oxvoen Data, vol %
Economizer outlet east - OE*
CR panel
3.20
0.40
2.00
3.00
2.00
1.60
2.60
2.60
2.00
2.00
3.00
Economizer outlet west • OE
CR panel
6.00
3.00
4.00
4.00
3.00
3.60
2.60
6.00
6.20
4.00
6.00
Economizer outlet average • OE
Man-calc
4.60
1.70
3.00
3.60
2.60
2.60
2.60
4.26
3.60
3.00
4.00
Economizer outlet average • B&W
S14O-C60
4.26
1.44
2.66
3.74
2.96
1.94
2.24
4.69
2.97
2.30
4.26
Economizer outlet average - B&W
S140-C67
2.92
1.39
1.64
3.33
2.30
N/A
1.96
3.47
1.36
2.28
3.38
Economizer outlet average • All
S140-C63
3.37
1.43
1.98
3.47
2.62
1.94
2.07
3.86
2.43
2.28
3.67
Air heater Inlet average - B&W
Manuel
6.20
4.60
6.30
6.20
6.16
4.76
6.30
6.06
4.60
6.10
6.10
Air heater outlet average - B&W
S140-C68
8.62
6.96
7.77
8.60
7.74
7.66
7.84
8.91
7.68
7.64
8.62
ESP outlet • Radian
S140
8.70
7.10
7.70
8.40
7.70
7.20
7.60
8.70
7.40
7.60
8.60
Emissions Data
Air heater inlet NO* ppmv
Manual
260
280
267
290
287
200
246
340
260
260
336
Air heater inlet NO* ppmv 9 3% Oa
Man-calc
316
306
306
363
326
222
281
409
274
263
406
Air heater inlet NO* lb/10* Btu
Man-calc
0.44
0.42
0.42
0.46
0.44
0.30
0.38
0.66
0.37
0.41
0.69
Air heater inlet CO, ppmv
Manual
26
1000
120
46
100
876
160
26
600
120
22
Air heater inlet CO, ppmv 9 3% 0,
Man-calc
30
1091
138
66
114
969
172
30
669
136
27
ESP outlet NO,, ppmv
S140
216
233
249
267
268
182
220
286
221
211
287
ESP outlet NO„ ppmv 0 3% 0,
Man-calc
316
302
337
381
349
238
296
419
293
281
417
ESP outlet NO„ lb/10* Btu
Man-calc
0.43
0.42
0.47
0.62
0.47
0.32
0.40
0.67
0.40
0.41
0.61
ESP outlet CO, ppmv
S140
18
746
68
13
64
638
83
16
386
81
13
ESP outlet CO, ppmv 9 3% 02
Man-calc
26
966
92
19
73
701
111
22
611
108
19
Carbon in aah, wt %
Man-calc
3.96
10.61
6.68
4.08
7.86
12.97
8.04
3.96
12.96
10.96
3.78
OE - Ohio Edieon.
* LCR panel - LIMB control room panel; CR panel
Men-celc - Manually calculated data.
1 Not available.
Plant control room panel; S140 - Syatem 140w input data; S140-Cxx - System 140™ calculated data on channel #xx; Manual - Manual data;
-------
• Airheater inlet - 8 gas sampling probes to composite bubbler (manual data)
• Airheater outlet - 4 Westinghouse in situ probes (System 140™)
• Precipitator outlet - 1 gas sampling probe (Radian CEMS)
The measurements from all points were cross-checked for accuracy/consistency by various
methods prior to and during the test program. It was determined that each oxygen analyzer being
used was accurately measuring the excess 02 at the point of measurement. Based on this
conclusion, differences in 02 at different points in the boiler were determined. These differences
are shown in Table C-6. A point of significant infiltration (2.75% 02/20% excess air) is between
the economizer outlet and the air heater inlet. The economizer and air heater hopper area have
been inspected to determine infiltration points. Significant infiltration occurs around the
economizer inlet header connections where the economizer tubes penetrate the west side casing.
During the baseline tests with the circular burners, the infiltration between the economizer outlet
and the air heater inlet was measured and found to be near zero (0.02% 02), identifying a
significant degradation in that area of the boiler during the time period from the baseline tests to
the characterization tests.
Table C-6 also shows significant leakage of combustion air to the gas side in the tubular air
heater. An average leakage of 49% excess air during the test program is partially the reason for
reduced loads (680,000 Ib/hr steam flow) being used for the high load test condition. At higher
steam flows, air quantities were not available to perform a 25% excess air test. (Also, cooling
water temperatures during the test program were higher than normal, affecting the performance of
the condenser and therefore limiting load capabilities).
Currently, Ohio Edison uses two multipointers in the control room panel for oxygen
indication. The multipointers correspond to the east and west side six foot 02 probes at the
economizer outlet. To provide adequate combustion air to the burners during four mill operation,
the control room operator should not allow the numerical average of these two multipointers to fall
below 3% 02. For mill out-of-service conditions, the control room 02 should be increased by a
minimum of 1.5% 02 per mill out of service [i.e., 4.5% for 3 mills operating, 6.0% for 2 mills
operating, etc.).
The System 140™ data acquisition system installed as part of the LIMB program has a
183
-------
TABLE C-6. OXYGEN AND EXCESS AIR ANALYSIS
Typo*
BT01
BT02
BT03
BT04
BT06
BT06
BT07
BT08
BT09
BT10
BT11
Oxvoen Data, vol %
Economizer outlet eaat - OE*
Economizer outlet weet • OE
Economizer outlet average - OE
Economizer outlet average - B&W
CR panel
CR panel
Man-calc
S140-C66
3.20
6.00
4.60
4.26
0.40
3.00
1.70
1.44
2.00
4.00
3.00
2.66
3.00
4.00
3.60
3.74
2.00
3.00
2.60
2.96
1.60
3.60
2.60
1.94
2.60
2.60
2.60
2.24
2.60
6.00
4.26
4.69
2.00
6.20
3.60
2.97
2.00
4.00
3.00
2.30
3.00
6.00
4.00
4.26
Economizer outlet averege - All
S140-C63
3.37
1.43
1.98
3.47
2.62
1.94
2.07
3.86
2.43
2.28
3.67
Air heater inlet averege - B&W
Manual
6.20
4.60
6.30
6.20
6.16
4.76
6.30
6.06
4.60
6.10
6.10
Mr heater outlet averege - B&W
S140-CES
8.62
6.96
7.77
8.60
7.74
7.66
7.84
8.91
7.68
7.64
8.62
ESP outlet - Radian
S140
8.70
7.10
7.70
8.40
7.70
7.20
7.60
8.70
7.40
7.60
8.60
Differences
Averege for
all tests
Economizer outlet average:
IMervcalc - S140-C63), vol % 0,
0.66
1.23
0.27
1.02
0.03
•0.02
0.66
0.43
0.40
1.17
0.72
0.33
(Economizer outlet - eir heater inlet) averege
(S140-C63 - Menuel], vol % 02
Approximate exceee eir equivalent, wt %'
-2.76
-20
-2.83
-23
-3.07
-20
•3.32
-23
-2.73
-22
-2.63
•19
-2.81
-19
-3.23
-22
-2.20
•18
-2.17
-16
•2.82
-20
•2.43
-20
(Economizer outlet • eir heater outlet) average
(SI 40-C63 - S140-C681, vol % 0,
Approximate exceee air equivalent, wt %
•6.34
-49
-6.26
•63
-6.62
-43
-6.79
-60
•6.03
-60
-6.22
-47
-6.61
-48
-6.77
-60
-6.06
-66
•6.16
-46
-6.36
-47
-4.96
•62
(Air heater inlet * outlet) everege
(Manual - S140-C68J, vol % 02
•2.69
•2.42
-2.46
-2.47
-2.30
•2.69
-2.80
-2.64
-2.86
-2.98
•2.64
•2.62
(Air heater outlet everege • ESP outlet)
[S140-C68 - S1401, vol % 03
0.10
•0.08
-0.16
0.07
0.10
0.04
0.36
0.24
0.21
0.18
0.14
0.02
* OE ¦ Ohio Edison.
* Approximated from • B&W standard chart.
* CR panel - Plant control room panel; S140 - System 140" input data; S140-Cxx - System 140" calculated data on channel #xx; Msnual - Manual data; Man-caic - Manually calculated data.
-------
series of oxygen outputs using the economizer outlet and air heater outlet 02 probes. If Ohio
Edison has the opportunity to use the System 140™ terminal, the operator should utilize the
economizer outlet 02 average for control purposes. The six probe configuration will provide a
better operating parameter than two probes over an extended period of time. When using the
System 140™ data, the control room operator should not allow the economizer outlet 02 average to
fall below 2.5% 02 during four mill operation. For mill out-of-service conditions, the 1.5% 02 per
mill rule (as stated above) would apply. This operating condition will provide minimum required
oxygen. The optimum operating point would be 3.0 to 3.5% 02 as measured by the six probes at
the economizer outlet.
N0„ Emissions Analysis
Figures C-1 and C-2 show the N0X performance of the DRB-XCL™ burner during
characterization tests. Figure C-1 shows the NOx performance versus boiler steam flow. The data
falls into typical patterns, with the NOx emissions increasing with increasing load. The high load
test condition (680,000 Ib/hr steam flow) with normal (mid) 02 provided NOx emissions of
approximately 0.48 lb/106 Btu. The NO„ emissions of all normal condition tests were below the
program goal of 0.50 lb/106 Btu. The higher load, high 02 conditions generated NOx in excess of
0.50 lb/106 Btu, but these conditions are not representative of typical operating conditions for the
boiler.
Utilizing the data from the test series, the NOx emissions may be extrapolated to what is
considered a normal full [peak] load condition of 770,000 Ib/hr steam flow. Assuming adequate
combustion air was available to provide 3.0% 02 at the economizer outlet with 770,000 Ib/hr
steam flow, the expected NOx emissions would be approximately
0.48 lb/106 Btu.
Figure C-2 shows NOx performance versus oxygen leaving the economizer. The high load
data is as expected, with NOx steadily decreasing versus decreasing 02 levels. The mid load and
low load data appear erratic when plotted versus economizer outlet 02. It was noted that a plot of
NOx versus air heater inlet 02 (shown in Figure C-3) shows trends that resemble typical NO„ versus
02 plots. The points on Figure C-2 that appear out of line correspond to the tests that show a high
differential between economizer outlet and air heater inlet 02 in Table C-6. Higher than average
185
-------
Main Steam Flow, Ib/hr x 10-3
Figure C-1. NOx emissions as a function of load.
186
-------
1 2 3 4
Excess Oxygen at Economizer Outlet, vol%
Figure C-2. NO„ emissions as a function of economizer outlet 02.
187
-------
Excess Oxygen at Air Heater Inlet, vol%
Figure C-3. NOx emissions as a function of air heater inlet 02.
188
-------
differences in the oxygen readings may account for the nonuniformity of the NOx emissions plotted
against the economizer 02.
Figure C-4 provides a comparison of the DRB-XCL™ burner to the original circular burners in
terms of NOx performance. This data is for an operating condition of 3.0 - 3.5% 02 leaving the
economizer. Across the load range, the DRB-XCL™ burner provides 40 - 50% NOx reduction versus
the uncontrolled circular burner. Although this NOx reduction does not appear astounding for a
state-of-the-art low NOx burner, it must be noted that only half of the burners on Edgewater Unit 4
operate in a truly low NOx mode. The lower and upper middle elevations of burners are equipped
with rapid mix impellers which substantially decrease the low NOx potential for these six burners.
Although this arrangement was required to provide overall acceptable performance in this boiler, it
does not truly characterize the NOx reduction potential of the DRB-XCL™ burner. For a boiler of
similar size as Edgewater Unit 4 without the constraint of inconsistent fuel distribution, the upper
level burners would not require rapid mix impellers to provide adequate combustion conditions.
Although the lower burners would still require impellers for flame length control, an additional 15 to
20% NOx reduction would be expected with consistent fuel distribution and optimized burner
arrangement.
Unburned Carbon Analysis
For each test condition, a fly ash sample was obtained using an isokinetic sampling probe
(B&W design SLM probe). The sample was obtained from a six-point grid in the duct between the
air heater outlet and the precipitator inlet. These samples were analyzed for carbon content by
ASTM method D2178-73 to determine the percent carbon in the ash. The goal for the program is
less than 5% carbon in ash for all operating conditions.
Figure C-5 shows the carbon in ash versus boiler load for the test series. Carbon in ash is
shown as relatively flat versus load, with lower loads showing slightly higher carbon in ash than
higher loads. Figure C-6 shows the effect of varying oxygen levels on carbon in ash. The same
inconsistencies noted for NOx versus economizer outlet 02 show up in the carbon in ash data.
Figure C-7 more clearly presents the effect of oxygen on carbon in ash as it is plotted against the
air heater inlet 02. As can be seen from Figure C-7, the carbon in ash increases very sharply for
oxygen levels less than 5.5 to 5.6% measured entering the air heater. From Figure C-6, this same
189
-------
Main Steam Flow, Ib/hr x 103
Figure C-4. Circular and DRB-XCL™ burner NOx emissions.
190
-------
400
500 600 700
Main Steam Flow, Ib/hr x 10"3
800
Figure C-5. Carbon in ash as a function of load.
191
-------
Higher Loads
~ High Load
A Mid Load
¦ Low Load
1 2 3
Excess Oxygen at Economizer Outlet, vol%
Figure C-6. Carbon in ash as a function of economizer outlet 02.
192
-------
15
14
13
12
11
Vp
10
O""-
§
9
j£
8
CO
<
c
7
c
o
6
CD
5
u
4
3
2
1
0
i—i—i—r
i—i—i—rr
Higher Loads
~ High Load
A Mid Load
¦ Low Load
4.4 4.8 5.2 5.6 6 6.4
Excess Oxygen at Air Heater Inlet, vol%
6.8
Figure C-7. Carbon in ash as a function of air heater inlet 02.
193
-------
critical point falls at approximately 2.7% oxygen as measured leaving the economizer. As
discussed in the oxygen analysis section above, the minimum operating 02 should be greater than
2.5% leaving the economizer. This value is based on the need for minimum air quantities required
for complete combustion and is reflected in the carbon in ash data.
For the normal full load condition of 770,000 Ib/hr steam flow, an extrapolation of the data
would predict carbon in ash values of approximately 5.5 to 6.0% with 3.0 to 3.5% 02 leaving the
economizer.
Figure C-8 shows a comparison of carbon in ash values for the DRB-XCL™ burner to those
obtained during the baseline tests with the circular burner. In order to compare the burners on an
equal basis, the carbon in ash data is plotted versus the oxygen leaving the economizer as
measured by Ohio Edison's in situ probes. These probes represent the only consistent 02 data
between the two test series due to the infiltration around the economizer outlet/air heater inlet area
having increased by approximately 2.5% 02 since the baseline tests. Figure C-8 shows
approximately 1.0 to 1.5% difference in carbon in ash values between the circular burner and the
DRB-XCL™ at normal 02 levels. The DRB-XCL™ burner becomes much more sensitive to oxygen
than the circular burner at low 02 levels.
Flame Length Analysis
During the optimization test series, much work was done to optimize flame lengths,
resulting in the use of impellers on the bottom row of burners. For the characterization tests, there
was no incidence of flame impingement on the rear furnace wall at normal or high 02 conditions.
For the lower 02 tests, the flames from the lower middle row of burners ("A" mill) were somewhat
longer and slightly impinged on the rear wall. In no case was the impingement severe to the point
that slag began to build on the wall, but the flames from "A" mill burners did turn up the rear wall
and reach into the upper furnace cavity. These observations were consistent with observations
from the optimization tests, where low 02 conditions were noted to create longer flame lengths
than normal operating conditions.
194
-------
3 DRB-XCL™ Burner
A Circular Burner
3 5
02 Leaving Economizer (Ohio Edison probes), vol %
Figure C-8. Carbon in ash burner comparison.
195
-------
Burner Pressure Loss Analysis
Figure C-9 shows the DRB-XCL™ burner pressure loss as a function of boiler load. For all
conditions studied during the characterization tests the burner pressure drop remained under the
program goal of 5.0 in WC. The highest pressure drop condition was high load, high 02, which
provided 4.6 to 4.7 in WC pressure drop from windbox to furnace. For a normal full load condition
of 770,000 Ib/hr steam flow and normal oxygen, the DRB-XCL™ burner pressure drop is expected
to be approximately 4.8 to 5.0 in WC.
CO Emissions Analysis
Even though there is not a stated program goal for CO emissions, a commercially accepted
value for a coal-fired utility boiler is less than approximately 200 ppmv with emissions typically less
than 100 ppmv. Figure C-10 shows CO emissions versus air heater inlet 02. As noted with carbon
in ash, a breakpoint in CO emissions occurs between 5.0 and 5.5% 02 measured at the air heater
inlet. For the test series, CO emissions were acceptable for all normal operating conditions. During
low oxygen tests, the CO emissions were extremely high and quite erratic, an indication of poor
combustion conditions.
CONCLUSIONS
Based on the data generated from the characterization tests series, the following
conclusions can be made concerning the DRB-XCL1" burners installed on Unit 4 of Ohio Edison's
Edgewater Station with respect to performance for the LIMB demonstration program:
• Air heater leakage is approximately 49% excess air at full load conditions. This excessive
leakage condition presents difficulties in providing adequate combustion air to the burners at
full load conditions. A load restriction has been realized due to this condition. [Following this
report, attempts were made to reduce this excess air infiltration. The effort included
replacement of 1,250 tubes in the air heater and major repairs in the vicinity of the
economizer header. Unfortunately, the reductions realized, while substantial, were not as
great as had been hoped due to the cumulative inleakage throughout the unit.]
196
-------
10
9
O
£
c
CO
03
o
—I
(U
3
U5
W
03
3
CO
8
7
£ 3
1
0
400
500 600 700
Main Steam Flow, Ib/hr x 10 3
800
Figure C-9. DRB-XCL burner pressure loss.
197
-------
1200
1000
800
>
a 600
o.
400
200
I I I I
i i i i i i i i
i i
= High Load
A Mid Load
—
¦ Low Load
—
Higher Loads
Lower Loads
—
"mm
i
i r
4.4 4.8 5.2 5.6 6 6.4
Excess Oxygen at Air Heater Inlet, vol%
6.8
Figure C-10. CO emissions as a function of air heater inlet 02.
198
-------
In order to provide adequate combustion air to the burners, Ohio Edison should utilize the
measured oxygen values for the economizer outlet oxygen - average from the System 140™
data acquisition system as a control factor. For four (4) mill operation, this value should not
be less than 2.5% 02, with an optimum value of 3.0 to 3.5% 02. For each mill out of
service, this value should increase by 1.5% 02:
Economizer Outlet 0- - Average, vol %
Mills in Service Minimum Optimum
4 2.5 3.0-3.5
3 4.0 4.5 - 5.0
2 5.5 6.0-6.5
NOx emissions during the LIMB program should be in the range of 0.48 to 0.50 lb/106 Btu for
full load operating conditions. NOx emissions should decrease with decreasing load. This
represents a reduction in NOx emissions of approximately 40 to 50% compared to the circular
burners but does not represent the DRB-XCL™ burners full NOx reduction potential due to
circumstances specific to this application.
Carbon in ash values for normal full load conditions are in the range of 5.0 to 8.0%.
Differences can be accounted for by deviations in operating oxygen levels or attributed to on-
line fluctuations in fuel distribution. Carbon in ash values will increase slightly with reduced
load and will increase greatly for a drop in oxygen levels below those specified. With proper
oxygen levels, carbon in ash will typically meet the program goal of less than 5% at the full
load condition.
The circular burner and the DRB-XCL™ burner provide near the same combustion
performance, with respect to carbon in ash, at normal oxygen levels. The DRB-XCL™ burner
is more sensitive at low oxygen levels and therefore has poorer combustion performance at
low oxygen levels.
DRB-XCL™ burner pressure drop is below the program goal of 5.0 in WC and should range
from 4.8 to 5.0 in WC for normal full load conditions.
199
-------
• CO emissions, which are a function of operating 02, are at an acceptable level well below
100 ppmv for all normal operating conditions. CO emissions increase greatly for reduced
oxygen conditions.
• Flame lengths of the DRB-XCL™ burner are well short of the 22 ft furnace depth at all normal
operating conditions.
RECOMMENDATIONS
The following recommendations are made pertaining to the use and operation of the DRB-
XCL™ burners in this application to the LIMB program:
• Ohio Edison should be made aware as soon as possible of the oxygen levels necessary to
insure complete combustion (as outlined above). In order to utilize the System 140™ oxygen
data, a terminal should be made available to the control room operator.
• Measures should be taken to correct the problems associated with inconsistent fuel
distribution on this boiler. An upgrade to the existing control system along with gravimetric
feeders would insure correction of the problem.
200
------- |