United States
Environmental Protection
Agency
&ERA Research and
Development
DEMONSTRATION OF NATURAL GAS
RE BURN FOR NOX EMISSIONS REDU CIION
AT OHIO EDISON COMPANY'S
CYCLONE-FIRED NILES PLANT UNIT NO. 1
EPA-600/R- 99-063
July 1999
Prepared for
Office of Air and Radiation
Prepared by
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
-------
FOREWORD
The U. S. Environmental Protection Agency is charged by Congress with pro-
tecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions lead-
ing to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, EPA1 s research
program is providing data and technical support for solving environmental pro-
blems today and building a science knowledge base necessary to manage our eco-
logical resources wisely, understand how pollutants affect our health, and pre-
vent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks
from threats to human health and the environment. The focus of the Laboratory's
research program is on methods for the prevention and control of pollution to air,
land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites and groundwater; and prevention and
control of indoor air pollution. The goal of this research effort is to catalyze
development and implementation of innovative, cost-effective environmental
technologies; develop scientific and engineering information needed by EPA to
support regulatory and policy decisions; and provide technical support and infor-
mation transfer to ensure effective implementation of environmental regulations
and strategies.
This publication has been produced as part of the Laboratory's strategic long-
term research plan. It is published and made available by EPA's Office of Re-
search and Development to assist the user community and to,link researchers
with their clients.
E. Timothy Oppclt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/R-99-063
July 1999
DEMONSTRATION OF NATURAL GAS R1BURN
FOR NOx EMISSIONS REDUCTION
AT OHIO EDISON COMPANY'S
CYCLONE-FIRED NILES PLANT UNIT NO. 1
Richard W. Borio, Robert D. Lewis, and Robert W. Koucky
ABB Power Plant Laboratories
Windsor, Connecticut 06095-0500
EPA Contract 68-02-4280
EPA Project Officer: Robert E. Hall
National Risk Management Research Laboratory
Research Triangle Park, North Carolina 27711
Prepared for:
Environmental Protection Agency (EPA)
(Office of Research and Development, Washington, DC 20460)
Gas Research Institute (GRI)
Electric Power Research Institute (EPRI)
Department of Energy-Pittsburgh Energy Technology Center (DOE-PETC)
Ohio Coal Development Office (OCDO)
East Ohio Gas Co. (EOG)
-------
A Disclaimer on the Economics of Reburning Presented in This Report
This report describes the first large-scale demonstration of reburning technology on a U.S. coal-
fired boiler. Technical information developed in this demonstration has been used in subsequent
commercial applications and demonstrations of this technology. Recent NOx regulations (i.e., the
acid rain regulations and the ozone transport regulations) have created a market for NOx control
technology applications in the U.S. The competitive nature of this market coupled with a
growing technology experience base has had a favorable impact on costs of NOx control
technologies. Since the development of this report, reburning technology has been successfully
applied on several coal-fired boilers. Taking into account these factors, the costs of reburning
presented in this report are considered to be higher than those expected to be applicable in the
current competitive environment.
ii
-------
ABSTRACT
The report describes a demonstration of rebuming on a cyclone-fired boiler. The project included
a review of reburn technology, aerodynamic flow model testing of reburn system design
concepts, design and construction of the reburn system, parametric performance testing, long-
term load dispatch testing, and boiler tube wall thickness monitoring, The report also contains a
description of Ohio Edison's Niles No. 1 host unit, a discussion of conclusions and
recommendations derived from the program, and tabulations of data from parametric and long-
term tests. A primary focus of the report was to document performance of the Niles Unit No, 1
when employing natural gas rebuming, but it was equally as important to be able to use the
information to make technical and economic judgements on the application of natural gas
rebuming to the entire family of cyclone boilers. Performance of the reburn system at Niles
might not represent the highest nitrogen oxide (NOx) reduction possible on cyclone boilers
because of differences in design and operating parameters, including boiler size and mode of
operation (e.g., cyclic vs base-loaded). Larger units are expected to have greater NOx reduction,
both percent reduction and total NO* removed, than those found at Niles.
iii
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge the individuals and organizations that contributed to this
successful demonstration project. The support provided by Ohio Edison throughout the project
is gratefully acknowledged. Particular appreciation is expressed to Sher Durrani and Dan Steen
of the Akron headquarters and Steve Brown of the Niles plant. The efforts of Niles Plant
Superintendents and all of the Niles plant personnel involved in the project were highly
instrumental in successfully completing the test program.
The following organizations provided significant support during the project. Although many
people from these organizations contributed to the success of the project, key individuals are
cited, where appropriate: U. S. Environmental Protection Agency (Robert Hall); Gas Research
Institute (Robert Lott, John Pratapas); Electric Power Research Institute (Angelos Kokkinos
[currently ABB C-E], George Offen); U. S. Department of Energy-Pittsburgh Energy
Technology Center (Doug Gyorke); Ohio Coal Development Office (Jackie Bird, Howard
Johnson); East Ohio Gas (Joe Kienle); Energy Systems Associates (Ben Breen, George Dusatko
and Richard Booth); Spectrum Diagnostix, Incorporated (Steve Johnson); and Stone and Webster
Engineering Corp. (Brett Krippene). Aziz Lookman of Energy Systems Associates (currently at
Carnegie-Mellon University) wrote portions of the section which describes the test
measurements and the sections which present results of parametric testing and long term testing.
He also provided technical support for data acquisition during the testing. Milt Manos
coordinated Ohio Edison's written contributions for the report. These contributions were by Dale
Corfman in the section which describes Niles Unit No. 1 and by A1 Waddingham in the section
which discusses boiler tube thickness monitoring. Steve Johnson provided contributions initially
as an employee of Spectrum Diagnostix, Incorporated and later as President of Quinapoxet
Engineering Solutions, Inc.; he provided major technical inputs and writing for the reburn system
economics section.
Special thanks are extended to Prof. Janos Beer of the Massachusetts Institute of Technology,
Prof. Jon McGowan of the University of Massachusetts, Mark Keough and Richard LaFlesh of
ABB C-E Services, Inc., and David Anderson, Andrew Kwasnik, Art Plumley, and
Winfred Roczniak of ABB Power Plant Laboratories.
iv
-------
CONVERSION TABLE
To Convert From
la
Multiply bv
in
m
2.540 x 10"2
ft
m
3.048 x 10"'
ft2
m2
9.290 x 10"2
ft3
m3
2.832 x 10"2
mile
km
1.609
ib
kg
4.536 x 10 1
ton
kg
9.072 x 102
ft/see
m/'s
3.048 x 101
lb/hr
kg/sec
1.260 x 10"4
tons/hr
kg/sec
2.520 x 10"1
gal
m3
3.785 x 10"3
'j
lb/in
kPa
6.895
HP
W
7.460 x 102
Btu
J
1.055 x 103
Btu/lb
kJ/kg
2.326
Btu/hr
W
2.931 x 10"1
cfm
m3/s
4.719 x 10"4
ft2/1000 cfm
m2/1000 m3/s
1.968 x 10"2
gr/dscf
kg/m3
2.288 x 10"3
in WG
Pa
2.491 x 102
Ib/mmBtu
ng/J
4.299 x 102
°F
°c
°C = (5/9) (°F-32)
psig
Pa (absolute)
6895 (psig + 14.7)
lb NOx/mmBtu
ppm NQx at 3% O2
714.
-------
This page is intentionally left blank.
vi
-------
EXECUTIVE SUMMARY
Passage of the 1990 Clean Air Act Amendments underscored the need for establishing
commercially acceptable technologies for reducing power plant emissions, especially oxides of
nitrogen (NOx) and sulfur dioxide (SOi). NOx and SO2 lead to formation of acid rain by
combining with moisture in the atmosphere to produce nitric and sulfuric acids. NOx also
contributes to the formation of "ground level" ozone. Ozone is a factor in the creation of smog,
leads to forest damage, and contributes to poor visibility. Currently, electric utility power plants
account for about one-third of the NOx and two-thirds of the SO2 emissions in the U.S. Cyclone-
fired boilers, while representing about 9% of the U.S. coal-fired generating capacity, emit about
14% of the NOx produced by coal-fired utility boilers.
Given this background, the Environmental Protection Agency (EPA), the Gas Research Institute
(GRI), the Electric Power Research Institute (EPRI), the Department of Energy-Pittsburgh
Energy Technology Center (DOE-PETC), and the Ohio Coal Development Office (OCDO)
sponsored a program led by ABB Combustion Engineering (ABB-CE) to demonstrate reburning
in a cyclone-fired boiler. Ohio Edison provided Unit No. 1 at its Niles Station for the rcbum
demonstration along with financial assistance. The Consolidated Natural Gas Company (CNG),
specifically East Ohio Gas, also provided technical and financial contributions. Working as
subcontractors with ABB-CE on the program were Energy System Associates (ESA) and
Spectrum Diagnostix, Incorporated.
Reburn technology reduces NOx emissions by creating a second combustion or "reburn" zone
downstream from the primary combustion zone. The injection of this reburn fuel creates a fuel-
rich zone in which the NOx formed in the main combustion zone is converted to molecular
nitrogen, carbon dioxide, and water vapor by the reaction of NOx with carbon-hydrogen
intermediates from the second, or reburn, fuel feed. Any unburned fuel leaving the reburn zone
is subsequently burned to completion in a downstream burnout zone where burnout air is
injected. Reburning is especially attractive for cyclone-fired furnaces since conventional low
NOx burners, or low NOx burners in concert with over fire air, cannot be used since low NOx
burners typically operate at lower temperatures, a condition which would prevent slag flow, a
necessary requirement for cyclone furnaces. Additionally, most cyclone boiler operators do not
want to alter fuel/air stoichiometrics in the cyclone because of the potential negative effects on
tube wastage.
A natural gas rebum system was installed on Ohio Edison's Niles Unit No. 1, a 115 MW (gross)
cyclone-fired boiler. The objective was to demonstrate that 50% NOx reduction could be
achieved at full load and that the reburn system could be operated without adversely affecting
boiler thermal performance and component life.
vii
-------
The project at the Niles plant represented the first commercial demonstration of a natural gas
reburn system. Although the effectiveness of reburning as aNOx reduction technique was shown
in many laboratory and pilot scale experimental tests, the subject demonstration was the first to
look at the total impact of a rebum system in a commercial boiler. Though NOx reduction was
the focus of the demonstration, it was even more important that the rebum system not cause any
unacceptable side effects on boiler operation and component life. Indeed, execution of this
project turned up a few unexpected results illustrating just why R&D demonstrations are
conducted.
It was the intent of the project to generate specific information on the performance of Unit No. 1
at the Niles station and to use this information to make technical and economic judgements on
the application of natural gas reburning to the entire family of cyclone boilers. It was recognized
that performance of the reburn system at Niles might not represent the highest NOx reduction
possible on cyclone boilers because of many design and operating factors, including boiler size
and mode of operation, i.e. cyclic vs. base-loaded. In fact, it was judged that many of the newer,
larger units will have NOx reductions exceeding those found at Niles, both the percentage NOx
reduction and the total NOx removed.
The original reburn system was designed to employ flue gas recirculation (FGR) as a carrier gas
for mixing of the natural gas with the bulk flue gas in the rebum zone. The original system met
the NOx reduction and boiler thermal performance objectives. However, much thicker slag
deposits formed on the back wall of the furnace. The deposits, which were as much as 12 inches
thick, had little or no effect on boiler performance and did not prevent completion of the original
system test program. However, long-term operation of the original rebum system was
unacceptable for several reasons. Slag falls during boiler operation could have a damaging effect
on screen tubes at the bottom of the furnace; the possibility of slag falls during slag removal
operation was a risk to personnel; and slag accumulation could cause blockage and misdirection
of the rebum fuel jets and reduced durability of the rebum nozzles. For these reasons there was a
need to identify the cause of the problem and to resolve it.
Importantly, resolution of the deposition problem with the original rebum system led to a
simpler, less expensive rebum system. The original rebum system employed flue gas
recirculation (FGR) as a means of better mixing the natural gas (rebum fuel) with the bulk flue
gas. Proof-of-concept (POC) testing showed that the thicker ash deposit was returned to its
normal thickness when the FGR was eliminated. The relatively cool FGR had caused the
normally thin, molten, running slag layer on the back wall of the secondary furnace to become
cooler and therefore thicker than the basecase condition.
A modified rebum system was designed and installed in which the use of FGR was eliminated.
As indicated in the POC testing, deposits on the back wall returned to normal thickness. NOx
reduction was initially lower than with the original system; but with continued operation and
increased operator attention to cyclone air/fiiel ratio control, NOx reduction improved, and during
the last period of long-term testing, full load NOx reduction was better than that achieved with
the original rebum system. Importantly, there were a number of other advantages with the
modified system, both operational and economic: the modified system showed heat transfer
-------
distribution within the boiler to be much closer to the base case conditions, and the cost of the
reburn system was lower due to the elimination of the FGR and associated equipment. The plant
net heat rate was also improved by eliminating the power requirement for the gas recirculation
fan. The elimination of FGR was considered sufficiently important from both an operational and
economic standpoint that a rebum system employing direct injection of natural gas was used as
the preferred design when conducting the economic analysis of natural gas reburn systems for the
entire family of cyclone furnaces.
The modified reburn system design and installation at Niles Unit No. 1 were relatively simple.
The key components for the reburn zone were the rebum fuel injectors, modifications to the
furnace water walls to permit penetration of the reburn fuel injectors in the rebum zone, and the
natural gas piping, controls, valves, and connections between the natural gas pipeline and the
furnace. Key components for the burnout zone were the ductwork, associated control dampers,
and the windboxes and nozzle assemblies where combustion air was injected into the gas mixture
from the reburn zone. Placement and configuration of reburn fuel and burnout air injectors were
important to achieve sufficient residence time and mixing.
The reburn test program included parametric testing of the original reburn system in which
natural gas was injected into the reburn zone mixed with recirculated flue gas (FGR), as well as
parametric testing of the modified reburn system followed by long-term dispatch testing to
measure system performance and operability during normal boiler operation. The most
significant period of long term operation with the modified reburn system occurred in June '92.
At this time operators had gained familiarity with the system and fuel air ratios for the individual
cyclones were more uniformly maintained. Out of a total of sixty-one (61) tests during the June
time frame, twenty-one (21) one-hour tests were conducted wherein the rebum zone
stoichiometry (RZS) was between 0.9 and 1.0, the CO was less than 200 ppm and the boiler was
operating at 90 MWe, or higher. Performance under these conditions, during June *92,
represented a culmination of the efforts in the project to meet the stated goal of 50% NOx
reduction at full load without adversely affecting boiler operation or component life. The
following key conclusions are drawn regarding emissions reduction, performance and operability
of the natural gas reburn process as applied to cyclone-fired furnaces:
* Natural gas reburn significantly reduced NOx emissions from the Niles Unit No. 1 cyclone
fired furnace. Reburn also affected CO emissions. Specific NOx and CO emissions behavior
was observed as follows:
- With the modified reburn system an average NOx reduction of 52.1 % was achieved at
full load with acceptable boiler operation and CO emissions lower than 200 ppm
during the final series of long-term dispatch tests when the reburn zone stoichiometry
(RZS) was between 0.90 and 1.00.
- Reburn zone stoichiometry (RZS ) was the most significant operating variable
affecting NOx reduction by the reburn process.
- NOx emissions decreased linearly as RZS was decreased.
- CO emissions increased exponentially as RZS was decreased.
ix
-------
- For long-term operation of a commercial reburn system RZS should be maintained
slightly above 0.9 to simultaneously minimize both NOx and CO. Because of the
inability to maintain precise coal/air ratios in each of the cyclones at Niks No. 1
during long-term testing, simultaneous NOx and CO emissions were minimized at
RZS of 0.94.
- NOx reductions of 30 to 70% were measured during parametric testing of the original
system at full load.
Natural gas reburn had a minimal effect upon boiler performance and electrostatic
precipitator (ESP) performance.
- During 18% natural gas rebum testing with the original system, waterwall heat
absorption decreased by approximately 5% and convective pass heat absorption
increased by 5%; attemperator spray flows, operating in a normal range, were able to
control steam temperatures at the design levels.
- Boiler efficiency decreased by 0.6% with 18% natural gas reburning in the original
system due principally to higher latent heat of vaporization losses caused by greater
moisture formation from combustion of natural gas.
- ESP collection efficiency was lowered slightly during reburn system operation due to
lower ESP inlet loading and a non-optimized flue gas conditioning system.
Operation of the original reburn system led to the buildup of much thicker ash deposits on the
rear wall of the furnace at Niles No. 1.
- Long term operation of the reburn system could not be sustained with the original
reburn system due to abnormally heavy slag buildup on the back wall and over the
reburn fuel injectors.
- The primary cause of thicker ash deposits was the cooling effect of FGR on the rear
wall.
The cooler FGR caused the normally thin, molten deposits to become thicker, sintered
deposits as they equilibrated to the change in the thermal environment.
The original rebum system was replaced by a modified reburn system in which the FGR
system was eliminated. Eliminating FGR eliminated the ash buildup deposition problem.
The modified reburn system also provided several cost and operations advantages over the
original rebum system.
- Lower capital cost.
- Smaller space requirement.
- Elimination of the high maintenance, energy intensive FGR fan.
- More favorable furnace heat absorption distribution. Radiant section heat absorption
increased and convective section heat absorption decreased resulting in lower
attemperator water flow requirement. Boiler efficiency was essentially the same as
that of the original system.
-------
The modified reburn system initially showed a NOx removal efficiency about 8% lower than
the original reburn system. Possible causes for the lower NOx reduction were initially
thought to be soot formation by the natural gas in the absence of the recirculated flue gas and
decreased mixing of the natural gas due to elimination of the recirculated flue gas. However,
NOx reduction improved as long-term testing continued; during the last period of long-term
testing, NOx reduction was greater than that achieved with the original reburn system.
Operator familiarity with the system and closer control of individual cyclone fuel/air ratios
was thought to be the reason for improvement.
Water injection into the reburn zone was initially thought to improve NOx reduction during
testing with the modified reburn system. A water leak in one of the water-cooled reburn fuel
injector guidepipes seemed to correspond directly with increased NOx reduction. However
controlled water injection tests conducted after completion of the long-term tests provided no
improvement in NOx reduction compared to the NOx reduction achieved during the final
series of long-term tests. Controlled water injection did however accomplish the following:
- Lower CO levels; CO emission of 46 ppm and NOx emissions of 325 ppm, corrected
to 3% 02, were achieved with water injection compared to CO emission of 110 ppm
at the same NOx emission level without water injection.
- The ability to operate the reburn zone at lower stoichiometries (lower NOx), while
maintaining the CO at acceptable levels.
Reburn systems installed on pressurized furnaces, such as Niles Unit No. 1, can result in a
hazardous situation if a casing leak occurs in the vicinity of the reburn zone because of the
presence of combustible gases. Possible commercial solutions were suggested:
- Convert pressurized units to balanced draft by adding an induced draft fan and
associated equipment.
- Convert tangent tube pressurized units such as Niles No. 1 to fusion welded walls by
adding fusion welds between the tubes.
- Erect an enclosure around the reburn zone which would operate at a slightly higher
positive pressure than the furnace pressure to assure that any leakage would be into
the furnace.
- Erect a "hood-like" structure around the upper part of the furnace so that gas
composition could be constantly monitored for possible changes.
It is unlikely that the first two could be economically justified. However, the third and
fourth options would be much less capital-intensive and could be configured to ensure
safe reburn system operation.
Operational constraints place a limitation on the rebum fuel feed rate and corresponding NOx
reduction during reduced load conditions.
-------
In order to assure effective tapping of slag from cyclone-fired units, it is necessary to
maintain a minimum heat release rate to the primary furnace and the corresponding
coal feed rate to the cyclone combustors.
- The minimum heat release requirement in the slag tap region of the primary furnace is
a function of the furnace size, cyclone design, and coal ash fusibility.
- Since the fuel fed to the reburn zone does not contribute to heat release in the slag tap
region, reburn fuel must be reduced and finally discontinued as boiler load is reduced
beyond a certain level.
- The need to decrease and ultimately discontinue reburn fuel is most severe in older,
smaller cyclone furnaces such as Niles No. 1 since less energy is available to maintain
effective tapping of liquid slag in these units. Reburn fuel feed was discontinued at
Niles during long-term testing when the boiler load was reduced below 80 MW gross.
- Because the proportion of reburn fuel used at reduced boiler loads is decreased and
ultimately turned off below a certain load, overall NOx reduction is less for reburn
systems installed on cyclone-fired furnaces which operate at reduced load for
substantial periods. The NOx concentration in the stack with reduced load, however,
tends to remain nearly constant because the "baseline" NOx also decreases with
reduced load.
The possibility of tube wastage during operation of the reburn system existed because the
reburn process generated a substoichiometric (reducing) gas mixture in the reburn zone. A
boiler tube monitoring program was conducted during the reburn system testing to address
this possibility. The findings of tube monitoring program were as follows:
The ultrasonic thickness testing in the waterwall sections was inconclusive since
changes in tube thickness were below the sensitivity of the U.T. measurement.
However, visual inspection of the waterwalls revealed that the tube surface appeared
to be unaffected by reducing atmosphere corrosion.
- Ultrasonic thickness measurements of the superheater and reheatcr sections, following
operation of the original reburn system, showed areas with an approximate 10% wall
loss, with wastage in areas of the fifth stage superheater as high as 0.100" over a 20
month timeframe. Indicated tube loss is thought to be from a combination of erosion
and corrosion.
- Tube wall thickness changes in the superheater and reheater sections during testing of
the modified reburn system were significantly less. The reduced tube wastage during
operation of the modified system, without FGR, is explained by the fact that the flue
gas mass flows/velocities during modified reburn system operation were returned to
basecase levels, thereby minimizing wastage due to erosion. Because tube wastage
was not uniform, it is believed that erosion was the larger contributing factor between
erosion and corrosion.
The remaining superheater/reheater tube life analyses performed before and after the
rebum project were inconclusive concerning any degradation due to high temperature
oxidation. Final inspection values gave higher remaining tube life values than did
initially obtained values.
-------
• The cost effectiveness of natural gas reburn retrofit for reducing NOx emissions from
cyclone-fired furnaces depends upon several factors including the following: (1) the baseline
NOx and the expected NOx removal efficiency of the process over the load range of the
boiler, (2) the load profile of the boiler, (3) whether or not it is necessary to terminate reburn
operation at some boiler load due to slag tapping requirements and if so at what load this
requirement is imposed, and (4) the difference in fuel costs between natural gas and coal. A
study of natural gas reburn economics indicated that natural gas reburning is most attractive
for newer large units, particularly, base-loaded units and least attractive for small, older units
used for cycling such as the Niles units.
• The economics study provided these data for Niles Unit No. 1: capital cost for installing a
rebum system = $34/kW; reburn cost-effectiveness = S5835/ton of NO* removed if the
natural gas/coal cost differential is $1.50/mmBtu; and reburn cost-effectiveness = $1592/ton
of NOx removed if there is no cost differential between natural gas and coal.
• Natural gas reburning is less attractive economically at Niles Unit No. 1 than at larger units
both because the baseline (uncontrolled) NOx emissions at Niles are low relative to larger
units and because the utilization of the NOx system at Niles is lower than at larger units.
Niles has four low heat-release cyclones which discharge gas into long, narrow passages
(high surface/volume ratio) resulting in relatively low gas temperatures and less NO*
formation than the NOx formation at larger more modern cyclone units. The utilization of the
reburn system at Niles is less than would exist at larger units because the Niles unit spends a
greater fraction of the operating time at part load and because the unit, since it operates at
lower gas temperatures, must cease reburn operation completely at a higher fraction of design
load.
Parametric testing and long-term testing during the Ohio Edison Reburn Demonstration project
provided several recommendations for reducing NOx and CO emissions by improvements to the
reburn system design and operation. These are:
• Improve the control system for feed of coal and air to the cyclones in order to have better and
more uniform control of RZS. In this way the reburn system will be better able to operate
nearer to the optimum RZS which will provide higher NOx reduction without aggravating CO
levels.
• CO levels turned out to be a limiting factor for NOx reduction. Decreases in RZS could
clearly produce lower NOx, but at the expense of unacceptably high CO. Better mixing of air
in the burnout zone and biasing residence times toward the burnout zone, rather than the
reburn zone, may result in lower NOx because of the ability to employ lower RZS while
maintaining acceptable CO levels,
• Introduce a small controlled amount of H2O with the natural gas in the reburn zone to reduce
CO formation; this would allow lower RZS, higher NOx reduction and acceptably low CO.
• Use stainless steel for water-cooled guide tubes and other components which are subjected to
high temperatures in order to reduce the possibility of failure of reburn zone components.
-------
1
TABLE OF CONTENTS
Section
Page
1
2
3
4
5
6
7
Disclaimer on Economics....
Abstract
Acknowledgements
Conversion Table
Executive Summary
Table of Contents ....
List of Figures
List of Tables ................
Introduction ...
Review of Reburn Technology
General Description of the Process
Fundamentals of the Reburn Process-
Experimental Investigations
Description of the Host Unit.
.,H
......iii
...... iv
v
.....vii
....1-1
....2-1
...3-1
...4-1
...5-1
...5-1
...5-3
...5-9
...6-1
Reburn System Design
Flow Modeling and Reburn System Conceptual Design.
Reburn System Design Requirements..
Reburn System Components and Installation
Boiler Thermal Performance
Control System ...
Test Planning/Measurements
Program Scope
Flue Gas Sampling and Analysis
Boiler Performance and Operations Data
Coal Composition
,..7-1
,..7-1
,7-10
,7-13
,7-14
7-19
...8-1
,..8-1
..8-2
,..8-6
,..8-7
1-1
-------
Table of Contents
Section Page
ESP Performance Data,..,..,....,..... .....8-7
Carbon in Ash .............8-7
Flue Gas Temperature and Flow Field 8-7
Reburn Zone Inlet Conditions 8-8
Data Analysis ...8-8
QA/QC Procedures for Gaseous Measurements 8-10
GA/QC Procedures for Boiler Operation Data 8-11
QA/QC Procedures for ESP Performance Measurements 8-12
9 Reburn System Installation and Startup ..9-1
Installation.... 9-1
Startup ....9-2
10 Parametric Testing (Original Reburn System) 10-1
Introduction 10-1
Ohio Edison Niles Plant Coal Analyses 10-3
Baseline NOx Emissions 10-3
NOx Emissions as a Function of Key Variables 10-5
Other Gaseous Emissions 10-9
Carbon in Ash 10-11
Furnace Gas Temperatures 10-11
Electrostatic Precipitator Performance 10-13
Boiler Thermal Performance 10-14
Ash Slagging Condition.......... 10-19
11 Design of Reburn System Without FGR 11-1
Analysis of the Slag Buildup Problem 11-1
Resolution of the Problem 11-3
12 Parametric Testing (Modified Reburn System}........ 12-1
Introduction 12-1
Modified System Emissions Performance and
Operating Characteristics 12-1
Modified System Optimization Tests......................... 12-5
Modified System Thermal Performance............... 12-8
Evaluation of the Modified Reburn System
Design and Performance 12-13
Project Planning 12-14
Water Injection 12-14
Conclusions 12-18
1-2
-------
Table of Contents
Section Page
13 Long-Term Load Dispatch Testing 13-1
Purpose of Long-Term Testing 13-1
Reburn System Operation 13-1
Long-term NOx and CO Emissions 13-2
Difference Between Parametric and Long-Term
Testing Conditions 13-9
Commercial Potential of Gas Reburn for NOx Control 13-11
Utility Operator's Assessment of NOx Reduction by the
Reburn Process 13-13
14 Boiler Tube Thickness Monitoring Program 14-1
Description of the Program 14-1
Testing Sequence 14-1
Testing Locations..................... 14-1
Instrumentation 14-2
Ultrasonic Tube Thickness Test Results 14-2
Remaining Tube Life Analysis Using Oxide Scale Measurements 14-6
Corrosion Probe Tests 14-7
Conclusions 14-7
15 Application of Reburnfng to Pressurized Furnaces 15-1
Background.... 15-1
Resolution of the Problem .........15-1
16 Reburn System Economics... .....16-1
Introduction. 16-1
Basis for Study... 16-2
Reburn System Design and Economics 16-8
Conclusions 16-38
17 Conclusions and Recommendations 17-1
18 References 18-1
Appendices
A "Ultrasonic Thickness (UT) Measurements at Ohio Edison
Company's Niles Plant Unit No. 1," W. R. Roczniak, ABB
Plant Laboratories, August 1992 A-l
B Ohio Edison memorandum, A. L. Waddingham to Sher Durrani
dated January 9, 1991, "Niles No. 1 Boiler Waterwall Survey" B-l
C "Interim Field Test Report," W. R. Roczniak, ABB Power Plant
Laboratories, April 23, 1992 C-l
1-3
-------
2
LIST OF FIGURES
Figures Page
5-1 Schematic of Reburn Process ...5-2
5-2 Schematic of the Formation and Destruction Mechanisms of NO 5-6
5-3 Generalized Conceptual Mechanism for N0X Formation
and Reduction 5-8
6-1 Ohio Edison Niles Unit No. 1 .........6-2
7-1 Schematic of the Ohio Edison Reburn Process 7-2
7-2 One-ninth Scale Flow Model of Niles Unit No. 1 7-3
7-3 Test Plane and Injector Locations ...................7-5
7-4 Baseline Axial Velocity Contours, Plane 1 7-6
7-5 Baseline Axial Velocity Contours, Plane 4............ .7-7
7-6 Recommended Reburn Fuel Injector Configuration ..7-9
7-7 Recommended Burnout Air Injector Configuration 7-11
7-8 Reburn Fuel Injector Windboxes 7-14
7-9 Furnace Heat Absorption Rate -108 MW 7-17
8-1 Boiler Exit Gaseous Emissions Sample Matrix 8-4
2-1
-------
List of Figures
Figures Page
8-2 Schematic of Flue Gas Sampling and Analysis System 8-5
10-1 Baseline N0X Versus Cyclone 02,108 MWe and 86 MWe Net 10-4
10-2 NOx versus Reburn Zone Stoichiometry at Various Gas
Flow Rates, 108 MWe, 10% FGR 10-5
10-3 NOx versus Cyclone 02 at Various Natural Gas Flows and Reburn
Zone Stoichiometries, 108 MWe, 10% FGR 10-6
10-4 NOx versus Percent Flue Gas Recirculation at Constant Reburn
Zone Stoichiometry 10-7
10-5 NOx versus Reburn Zone Stoichiometry at Various Gas
Flow Rates, 108 MWe and 86 MWe 10-8
10-6 CO versus Burnout Air Tilt at Several Reburn Fuel Injector
Tilts, 108 MWe Net, 5% FGR, 17.5% Natural Gas, 2.5% Cyclone 02. 10-9
10-7 Oz and CO versus Boiler Duct Sample Location,
Non-Optimized Operation 10-10
10-8 02 and CO versus Boiler Exit Duct Sample Location,
Optimized Operation 10-10
10-9 Flue Gas Temperatures, Reburn Zone Inlet, 108 MWe Net 10-12
10-10 Flue Gas Temperatures, Reburn Zone Inlet, 86 MWe Net.... 10-12
10-11 Flue Gas Temperature, Furnace Outlet 10-13
10-12 Reburn Nozzles and Rear Wall after Completion of Parametric
Testing 10-21
10-13 Original Reburn System Nozzles Prior to Parametric Testing 10-22
11-1 Slagging Mechanism at Niles Unit No. 1 11-2
«¦#
-------
List of Figures
Figures Page
12-1 Original and Modified Reburn System NOx
Emissions at Full Load 12-4
12-2 NOx - CO Emissions Comparison for the Original System
and Modified System at Full Load 12-4
12-3 NOx Emissions for Variations in Aspirating Air Flow
Rate and Gas Nozzle Tip Arrangements 12-6
12-4 NOx Emissions for Variations in Gas Injector Position..... 12-6
12-5 NOx Emissions for the Original Reburn System and
Modified Reburn System at 86 MWe .....................12-7
12-6 N0X Emissions for the Modified Reburn System and
Tests with Water Leaks... 12-15
12-7 Water Injection for the Modified System 12-16
12-8 Modified Reburn System NOx Emissions at Full Load for Parametric
Tests, Long-Term Tests in June 1992 and Parametric Tests with
Water Injection.......... 12-17
12-9 NOx-CO Emissions Comparison for the Modified System Without and
With Water Injection. 12-17
13-1 Variation of NOx Emissions with RZS at 110+ MWe 13-3
13-2 Variation of NOx Emissions with RZS at 100-110 MWe 13-3
13-3 Variation of NOx Emissions with RZS at 90-100 MWe 13-4
13-4 Variation of NOx Emissions with RZS at 80-90 MWe 13-4
13-5 Variation of NOx Emissions with RZS at 90-110+ MWe 13-8
13-6 NOx Emissions for the Original System Parametric Tests and
the Modified System June 1992 Long-Term Tests 13-8
13-7 Comparison of NOx Emissions at Different Loads 13-9
2-3
-------
List of Figures
Figures .......Page
13-8 Variation of CO Emissions with RZS at Full Load 13-10
13-9 Reburn Effectiveness at Niles Unit No. 1 for Different Loads 13-12
14-1 Vertical Corrosion Test Probe..... 14-3
14-2 Horizontal Corrosion Test Probe 14-4
14-3 Corrosion Probe Locations 14-5
16-1 Firing Arrangements Used for Cyclone Furnaces 16-6
16-2 Niles Load Profile 16-11
16-3 Load Profiles for High Load Factor (Base-Loaded) Units, Typical
Coal-Fired Units, and Intermediate Load Factor Units.. 16-12
16-4 Niles Unit No. 1 Ohio Edison Company ..16-15
16-5 Relative Effect of Cyclone Arrangement on NOx Emissions................. 16-19
16-6 Unit A - 75 MW Gross, 1969 16-21
16-7 Cost of Reburning on a 75 MW Cyclone-fired Boiler 16-24
16-8 Unit B -125 MW Gross, 1957 16-25
16-9 Cost of Reburning on a 125 MW Cyclone-fired Boiler 16-27
16-10 Unit C - 225 MW Gross, 1958 16-28
16-11 Cost of Reburning on a 225 MW Cyclone-fired Boiler 16-29
16-12 Unit D - 420 MW Gross, 1968 16-31
16-13 Cost of Reburning on a 420 MW Cyclone-fired Boiler 16-32
16-14 Unit E - 605 MW Gross, 1970 16-34
16-15 Cost of Reburning on a 605 MW Cyclone-fired Boiler 16-35
2-4
-------
List of Figures
Figures Page
16-16 Capital Cost of Natural Gas Reburning 16-37
16-17 NOx Removal Cost Effectiveness for a Range of Boiler Sizes Based
2-5
-------
3
LIST OF TABLES
Tables Page
5-1 Names used for the Reburn Process. 5-3
5-2 Reburn Experimental Parameters ..........5-10
5-3, Reburn Experimental Data Summary 5-11
Sheet 1 thru
thru 6 5-16
6-1 Description of Host Boiler: Niles Unit No. 1 ...........6-3
6-2 Description of Electrostatic Precipitator at Niles Unit No. 1 ..6-4
7-1 Calculated Boiler Efficiency -108 MW 7-16
7-2 Coal Analysis - % by Weight 7-16
8-1 Specifications of Gas Analysis Instruments. ....8-6
10-1 Ohio Edison Niles Unit No. 1 Coal Analyses (As-Received Basis) 10-3
10-2 Niles Unit No. 1 Operating Data 10-15
10-3 Summary of Boiler Thermal Performance for the Original
Reburn System 10-18
12-1 Full Load Parametric Test Emissions Measurements for the
Original Reburn System, Modified System, and System with
Water Injection........... 12-2
12-2 Operating Data 12-9
12-3 Summary of Results 12-12
3-1
-------
List of Tables
Tables Page
13-1 Distribution of Load and Reburn Conditions During
Long-Term Data Acquisition 13-2
13-2 Full-Load Long-Term Emissions Data for Reburn System
Operation with 16% or Greater Natural Gas Reburn Fuel...... 13-5
16-1 89 Cyclone Boilers with 1957 and Later Start-up Dates 16-2
16-2 Study Boilers 16-7
16-3 Reburn Design Criteria for Niles....... 16-8
16-4 Preliminary Design Summary 16-14
16-5 Summary of Returning Cost Effectiveness 16-17
16-6 Factors Affecting Baseline NOx 16-19
16-7 Other Boilers Retrofitted with Natural Gas Reburning... 16-36
16-8 Long-term Reburning Performance in Other Boilers 16-36
3-2
-------
4
INTRODUCTION
Passage of the 1990 Clean Air Act Amendments has underscored the need for establishing
commercially acceptable technologies for reducing power plant emissions, especially nitrogen
oxides (NOx) and sulfur dioxide (SO2). NOx and SO2 lead to formation of acid rain by
combining with moisture in the atmosphere to produce nitric and sulfuric acids (Brack (1987),
Hakkarinen (1987), and Johnson and Siccama (1983)). NOx also contributes to the formation of
"ground level" ozone. Ozone is a factor in the creation of smog, leads to forest damage, and
contributes to poor visibility.
Electric utility power plants account for about one-third of the NOx and two-thirds of the S02
emissions in the U.S. Cyclone-fired boilers, while representing about 9% of the U.S. coal-fired
generating capacity, emit about 14% of the NOx produced by coal-fired utility boilers.
Given this background, the Environmental Protection Agency (EPA), the Gas Research Institute
(GRI), the Electric Power Research Institute (EPRI), the Department of Energy - Pittsburgh
Energy Technology Center (DOE-PETC), and the Ohio Coal Development Office (OCDO)
sponsored a program led by ABB Combustion Engineering, Inc. (ABB-CE) to demonstrate
rebuming on a cyclone-fired boiler. Ohio Edison provided Unit No. 1 at their Niles Station for
the rebum demonstration along with financial assistance. The Consolidated Natural Gas
Company (CNG), specifically East Ohio Gas, provided technical guidance as well as financially
sharing in the program. Ohio Edison and East Ohio Gas both shared a portion of the differential
between the cost of natural gas and coal. Working as subcontractors to ABB-CE on the program
were Energy System Associates (ESA) and Spectrum Diagnostix, Incorporated. The Niles Unit
No. 1 reburn system was started up in September 1990. This reburn program was the first full-
scale rebum system demonstration in the U.S.
This report describes work performed during the program. The work included a review of reburn
technology, aerodynamic flow model testing of reburn system design concepts, design and
construction of the reburn system, parametric performance testing, long-term load dispatch
testing, and boiler tube wall thickness monitoring. The report also contains a description of the
Niles No. 1 host unit, a discussion of conclusions and recommendations derived from the
program, a diskette containing tabulation of data from parametric and long-term tests, and
appendices which contain additional tabulated test results.
Though a primary focus of the report was to document performance of the Niles, Unit No. 1
when employing natural gas reburning it was equally important to be able to use the information
to make technical and economic judgements on the application of natural gas reburning to the
4-1
-------
Introduction
entire family of cyclone boilers. It was recognized that performance of the rebum system at Niles
might not represent the highest NOx reduction possible on cyclone boilers because of differences
in design and operating parameters, including boiler size and mode of operation, i.e. cyclic vs
base-loaded. As will be seen later in the report, notably Chapter 16, Rebum System Economics,
larger units are expected to have greater NOx reduction, both the percent reduction and the total
NOx removed, than those found at Niles.
4-2
-------
5
REVIEW OF REBURN TECHNOLOGY
The process of rebuming, a fuel staging process that provides in-fornaee reduction of nitrogen
oxides (NQX) emissions, has been demonstrated in laboratory, pilot scale, and full scale
combustor test trials. In many cases, reduction of NOx emissions of 50% and greater were
demonstrated. These trials provided direction for practical application of reburning to foil scale
boilers. However, the results depend on the apparatus and operating conditions and must be
adequately interpreted to provide the criteria necessary for the design of full scale utility
reburning systems.
This review was conducted during the planning stage and equipment design for the Ohio Edison
Reburn Project. The objectives of the review were to study previous experiments and to interpret
the results in order to:
1) Establish reburn system design and operating parameters
2) Assess the relative importance of the parameters and establish appropriate values or value
ranges for these parameters in terms of system performance and NOx reduction efficiency.
3) Establish an overall set of design criteria for a full scale utility reburn system design.
General Description of the Process
The reburning process is an in-furnace NOx control technology that diverts some of the fuel and
combustion air flows from the main burners and injects them above or downstream of the main
flame. Rebuming can be employed with any fossil fuel, or combination of fossil fuels, typically
coal, oil, or natural gas. Natural gas is technically ideal because it contains no fuel nitrogen and
it can be burned with relatively lower residence times. The reburning process involves the three
zones shown in Figure 5-1:
1) Primary Zone: This is the main heat release zone where the majority of thermal energy is
released to the boiler. This zone operates under overall fuel lean conditions although the
burners can be of a low NOx producing design with low levels of excess air; this, however, is
not the case with cyclone-fired combustors which do not employ burners in the usual sense.
The level of NOx exiting from this zone is the level to be reduced in the reburning process.
5-1
-------
Review of Return Technology
2) Rcbum Zone: This is the zone into which rcburn fuel is injected (downstream of the primary
flame zone) wherein NOx reduction occurs. The nitrogen species entering this zone come
from two primary sources: (1) the "thermal NOx" from fixation of nitrogen in the primary
zone combustion air, and (2) "fuel NOx" from the nitrogen contained in the primary fuel
(coal in this case). Depending on the choice of reburn fuel (coal, oil, or natural gas) there
could also be some nitrogen produced by the reburn fuel, if coal or oil was chosen. The
reduction of NOx is the result of hydrocarbon species from the reburn fuel reacting with NO
and NOi from the primary zone to form N2. Other products of this reduction zone are
reactive nitrogen species (cyanogens) and partially reduced hydrocarbons. To optimize the
NOx reduction through rebuming it is necessary to minimize the total reactive nitrogen
leaving the rebuming zone.
3) Burnout Zone: In the burnout zone, air is added to produce overall fuel lean conditions in
order to oxidize the unreacted fuel from the rcburn zone. I11 the burnout zone the remaining
reactive nitrogen species (cyanogens) may be converted to either NO or N2.
MECHANISTIC MODEL FOR
NOX DESTRUCTION
NOx FORMATION INHIBITED DUE
TO FUEL RICH CONDITIONS IN
REBURN ZONE; NOx DESTRUCTION
IS PROMOTED DUE TO SECONDARY
FLAME RADICAL ATTACK ON NO
PRODUCED IN PRIMARY ZONE TO
FORM MOLECULAR NITROGEN
BURNOUT AIR
REBURN FUEL <
PRIMARY
FUEL-AIR
HYPOTHESIZEO NOx DESTRUCTION MECHANISM:
OH,H
NO
CHi
OH,H w
HCN — ~ NH
BURNOUT
ZONE
REBURN
ZONE
PRIMARY
ZONE
ZONE 3
ZONE 2
ZONE t
NO
N2
(DESIRED PATH)
Figure 5-1
Schematic of Rebum Process
The use of separate fuel combustion stages to control NO emissions is not a totally new concept.
The first practical system using this approach was commercialized by the John Zink Company in
1975 (U.S. Patent No. 3873671 to Reed et al.). The Zink system was given the trade name
NOxIDIZER and was sold to reduce NO emissions from nitric acid plants. The first
investigation for applying the process to reduce emissions from combustion processes was
performed by Wendt et al. (1973), who injected CO and CH4 downstream of the flame zone of a
laboratory scale burner and measured significant reduction of NO emissions, Myerson (1974)
carried out similar experiments using a second combustion stage to reduce NO emissions from
5-2
-------
Review of Return Technology
automotive engines. The first commercial application of the process to utility furnaces was by
Mitsubishi Heavy Industries.
The reburning process is known under different names, the names depending on the researchers
or manufacturers applying the process. Several names are listed in Table 5-1. For simplification
purposes, all of these processes can be called rebum. All of the processes involve fuel staging to
provide in-furnace reduction of NOx.
Table 5-1
Names used for the Reburn Process
ORGANIZATION/INVESTIGATOR
DATES
NAME
John Zink Company
1970s
NOxIDIZF.R
Wendt, et al ./Shell Development
1970s
Reburning
Mitsubishi Heavy Industries (MHI)
1978
MACT (Mitsubishi Advanced Coal
Technology)
EER
1980s
Reburning; Fuel Staging
KVB
1980s
In-Furnace Control of NO Formation
(ICNF)
Babcock Hitachi
1980s
IFNR (In-Furnace NOx Reduction)
Hitachi Zosen
1980s
Three-Stage Combustion System
Ishikawajima Heavy Ind. (IHI)
1980s
IFNR (In-Furnace NOx Reduction)
Acurex
1980s
Fuel Staging
Fundamentals of the Reburn Process
Overview
The technology for reducing NOx emission by the downstream addition of fuel has been under
investigation for several years. Myerson (1974) and Wendt et al. (1973) conducted fundamental
studies of the destruction of NOx by injection of secondary fuel (hydrocarbons) and named the
process "reburning". Since that time, research on reburning was conducted in Japan, and more
recently in the United States. For example, Takahashi et al. (1982) of Mitsubishi Heavy
Industries (MHI) documented MHI's research on the NOx reduction process through
hydrocarbon injection by reporting on the development of the Mitsubishi Advanced Combustion
Technology (MACT), the first Japanese reburning concept. The MACT system diverts a small
percentage of fuel from the main burner combustion zone and injects the fuel through upper
injection ports with an inert fluid, usually flue gas. The balance of the combustion air is
provided via overfire air ports. This research by MHI showed that nitrogen oxide (NO) formed
during the initial stages of combustion could undergo significant conversion to molecular
nitrogen by the injection of hydrocarbons. In the 1980's, research investigations and commercial
demonstrations of this technology were conducted both in Japan and the U.S. Several of the
projects were limited to the use of natural gas and/or oil as the reburn fuel; however a few
investigated reburning with coal as primary or reburn fuel. To illustrate the scope of issues
involved, a brief review of several recent investigations is given.
5-3
-------
Review of Reburn Technology
The work of Takahashi et al. at MHI (1982) led to further research by others including Miyamae
et al, from Ishikawaj ima-Harma (1985), Mulholland and Hall (1985) and Mulholland and Lanier
(1985) of Acurex/EPA, Greene, McCarthy, and Overmoe from EER (1985), and Mulholland and
Hall of Acurex/EPA (1987), This research yielded slightly different results as to the
applicability of different fossil fuels as a reburn fuel. For example, from the results obtained
from a small test facility, Takahashi concluded that the NOx reduction efficiency of the reburn
process was independent of the reburn fuel type and of the reburn zone inlet NOx level.
Subsequent results from other researchers (such as Mulholland et al. (1985)) indicated that
reburning efficiency is influenced by both of these parameters. Most of the early researchers
concluded that the effectiveness of the NOx reduction is a strong function of the residence time in
the reburn zone. Furthermore, they generally found that an optimal reburn zone stoichiometry
(defined as the actual oxygen in the region divided by the oxygen required for complete
combustion of the fuel to carbon dioxide and water vapor) of about 0.8 to 0.95 was desirable,
with MHI promoting a value of 0.95 as an "optimal compromise", taking boiler performance into
consideration.
The results reported by Mulholland et al. (1985) indicate that substoichiometric reburn zone
conditions are optimal and that the reburn zone inlet NOx level is a key parameter in determining
the rebuming NOx reduction efficiency. They also stated that the nitrogen content of the reburn
fuel was important in determining the NOx reduction efficiency that can be achieved. These
researchers also recommended an optimal reburn zone residence time of approximately 0.5
seconds,
Miyamae et al. (1985) worked with natural gas and gas phase volatile matter evolved from
pulverized coal. In their work it was concluded that a main burner stoichiometry of 1.1 was
optimal. Also, they indicated that while a 50% NOx reduction efficiency was possible in a
laboratory test facility, only 15 to 20% NOx reduction efficiency would be possible with a
multibumer full scale demonstration. They also concluded that a reburn residence time of about
0.5 seconds was optimal.
Okigami et al. (1985) utilized a South African coal and an Australian coal as the rebum fuel.
Although they did not present quantitative design and operating characteristics of the reburn
zone, residence time or stoichiometry, they did demonstrate that significant NOx reduction (up to
60%) could be achieved using coal as a reburn fuel.
Pilot scale work in the U.S., McCarthy et al. (1985), confirmed that the optimum rebum zone
stoichiometry is close to 0.9. The results also showed that there was an increase in rebuming
efficiency when the rebum zone temperature was increased from 2600°F to 2840°F.
Thermochemistry
The degree of conversion of fuel-bound nitrogen (FBN) to molecular nitrogen (N2) is determined
by the thermodynamics of the system (temperature, pressure, and chemical composition of the
gas mixture), and the rates of reaction in the flame zone. In practical combustion systems the
reaction of FBN and NO to N2 may be kinetically constrained and provision of sufficient
residence time in a fuel-rich zone is therefore essential for high conversion to N2. The partially
5-4
-------
Review of Reburn Technology
reacted nitrogenous or cyanotic (CN) compounds in a fuel-rich flame zone may have different
origins: they may be pyrolysis products of nitrogen-bearing fuels, products of the fixation of
atmospheric nitrogen by hydrocarbon fragments early in the flame, or the result of reactions due
to the secondary injection of hydrocarbons into the NO-containing burned gas, sometimes
referred to as "prompt NOx". The various pathways for formation and destruction of NO are
shown schematically in Figure 5-2. In the combustion air staging technique, conversion of FBN
to N2 rather than to NO is mainly due to fuel-rich conditions prevailing in the primary flame
zone near the burner.
In fuel staging, the destruction routes of NO shows that NO can be destroyed in two ways. It can
be reduced either by reacting with amines (NHj) to form molecular nitrogen or by reacting with
hydrocarbon radicals such as CH and CII2 to produce hydrogen cyanide which in turn is
converted to NH3. The ammonia thus formed can subsequently reduce NO to N2 or be directly
converted to N2. In the case of pulverized coal, volatile matter evolves from the coal upon
injection into the hot furnace environment (approximately 2500 °F). The volatiles thus formed
crack into compounds which contain nitrogen such as HCN and NH3 and non-nitrogen
containing species such as CH4 and C2H2. These species can then start the NOx reduction
process. Under reburning conditions, it is felt that the CHj radicals play an important part in
reducing the NOx to N?. A critical step in the reburning reaction sequence is the conversion of
HCN to NHj via interaction with the free radical pool (O, OH, H). Once fomied, the NHj species
further reacts with NO to form N2.
Kinetics
The fundamental reactions leading to the formation and destruction of NOx are too numerous and
complex to be described in detail here. However, the kinetics of the NOx production/destruction
reactions as applied to reburning can be summarized by the following discussion.
The formation of NOx during fossil fuel combustion is a process involving contributions from
both the fixation of atmospheric nitrogen (thermal NOx) and the oxidization of nitrogen bound
chemically in the fuel (fuel NO*). NOx generation via the thermal fixation of atmospheric
nitrogen can be approximated in large boilers by the use of a highly temperature dependent
chemical reaction rate determined by Zeldovich (1974), The rate of formation is exponentially
dependent on temperature and is proportional to the square root of the oxygen concentration.
Reducing both the amount of oxygen available to the fuel and reducing the combustion
temperature are effective methods of controlling NOx formation via the thermal mechanism.
Although only a fraction of the nitrogen in fuel is converted to NOx, fuel NOx can represent a
major fraction of the total NOx, Fuel nitrogen conversion is a particularly important source for
NOx formation in coal-fired furnaces. For example, when firing a high-nitrogen fuel in a
conventional steam generation unit, fuel nitrogen accounts for 50 to 80% of the NO* emitted.
5-5
-------
Review of Reburn Technology
Figure 5-2
Schematic of the Formation and Destruction Mechanisms of NO
5-6
-------
Review of Reburn Technology
A generalized conceptual mechanism for NOx formation and reduction is illustrated in Figure
5-3. Recent experimental evidence of NOx kinetics supports this sequence of events, Toqan et al.
(1987):
• Fuel nitrogen rapidly and irreversibly breaks down to HCN
• HCN is irreversibly converted to NCO through a rate-controlling step
• HOCN and NCO rapidly interchange with one another and react to form NH, species
• The NH; species are equilibrated among themselves and, except for possibly N, provide the
branching point for the production of NO or N2
• The oxidizer and NHj species which provide the branching point are uncertain.
Based on this model, when fuel containing nitrogen is fed into a furnace or combustor, a large
fraction of the nitrogen compounds evolve into the gas phase as the fuel volatilizes. The volatile
nitrogen is predominately in the cyanogen form during the combustion of oil or bituminous coal,
although a substantial fraction (originating as amino-bonded species) may evolve as NH; from
subbituminous or lignite coals. The cyano species are thought to react with the flame-generated
free radicals to form the amine species that can further react with oxygenated species to form
NO, or with NO to form N2. The fractional conversion to NO is therefore highly sensitive to the
amount of available oxygen (excess air) and the mixing conditions that determine the contact
time between reactive nitrogenous and oxygenated species. The amino and cyano subsystems
are often idealized for modeling purposes as being in partial equilibrium.
This model indicates that there are two pathways of NOx destruction. For the first process, NO
can react with hydrocarbon radicals (symbolized by CH) to reform HCN/CN. This reaction
takes place under foel-rich conditions in any staged combustion process, but is maximized during
the rebuming process. This recycle of NO to HCN provides a second opportunity to produce N2,
and can result in substantial reductions in NOx emissions. The other pathway to NOx destruction
takes place by reaction of NO with NHj radicals, and is known to occur at low temperatures
(1600-2000°F) and where local oxygen concentrations are low.
5-7
-------
Review of Return Technology
HOMOGENOUS GAS PHASE REACTIONS HETEROGENEOUS REACTIONS
Figure 5-3
Generalized Conceptual Mechanism for NOx Formation and Reduction
Detailed chemical kinetic modeling for the reduction of NO has been performed for simple
flames, shock tubes, and full-scale combustors. For gas-phase NOx modeling, typical kinetic
models contain mechanisms consisting of from 30 to over 100 elementary reactions. Each
reaction requires specification of reactants, products, thermochemistry, and rate coefficients as a
function of temperature. The rate laws for such mechanisms consist of sets of coupled ordinary
differential equations which must be integrated numerically. In general, the computations are
difficult due to large, rapid excursions in the rates of many of the free radical reactions
("stiffness"). An excellent example of numerical analysis for NOx formation/destruction is given
by Toqan, et al. (1987). The numerical calculations were carried out using CHEMKIN, a
chemical kinetic code developed at Sandia National Laboratory, Kee et al. (1980), coupled with
a differential equation solver developed at Lawrence Livcrmorc Laboratory.
5-8
-------
Review of Return Technology
Experimental Investigations
Overview of Experimental Parameters
This section provides a review of experimental work conducted to develop reburn technology.
Data for thirty-five (35) experimental studies are compiled to address experimental parameters
listed in Table 5-2. The thirty-five experimental studies are summarized in Table 5-3, Sheet 1
through Sheet 6, The tables list data for the Overall System, Primary Combustor, Reburn
Section, and Burnout Zone. More detailed descriptions of experience gained for each of the
sections of the reburn system are given in the following subsections.
Primary Zone
Three primary zone variables have bearings on the effectiveness of the reburn process: primary
zone fuel type, primary zone stoichiometry, and NOx level of the gas leaving the primary zone.
Table 5-2 indicates that reburn systems have been used with all three fuels: gas, oil, or coal, as
the primary zone fuel. The selection of primary zone stoichiometry is influenced by boiler
efficiency and corrosion considerations as well as NOx emissions. Three investigations, Maringo
et al. (1987), Maringo and McElroy (1987), and Farazan et al. (1989) indicate that for cyclone-
fired coal furnaces the primary zone stoichiometry should not be less than 1.1 in order to avoid
corrosion in the cyclone and to assure complete combustion of the primary fuel. Another
investigation, McCarthy et al. (1987), states that it is important to minimize excess air in the
main burner zone in order to minimize NOx leaving the primary zone. The NOx concentration of
the gas leaving the primary zone has an effect on the performance of the reburn zone. The NOx
level leaving the reburn zone is lowest when the NOx level entering the rebum zone is the lowest.
However, the percentage reduction of NOx increases as the NOx level entering the reburn zone
increases.
Return Zone
Reburn Zone Fuel. Table 5-3 shows that all three fuels have been used for the reburn fuel.
However gas was found to have advantages because it contains no fuel nitrogen, retrofits are
generally easier, and burnout can be more successfully accomplished in the limited furnace
volume available when gas is the reburn fuel.
Fraction of Reburn Fuel. For tests listed in Table 5-3 the fraction of rebum fuel varied from
a very small fraction up to 70%. Maringo and McElroy (1987) and Farzan et al. (1989) showed
that NOx levels decreased with increasing amounts of reburning fuel fraction (up to about 35%).
This can be generally attributed to higher values of reburning fuel fraction corresponding to
lower reburning zone stoichiometrics. Also, for the case of using gas as a reburning fuel,
substitution of higher percentages of gas for coal reduced the total nitrogen fuel input to the
furnace. Additionally, as the percentage of fuel in the primary zone is decreased, peak
temperatures are decreased and thermal NOx is thereby decreased. Mulholland and Srivastava
(1987) found that NOx emissions decreased with increased fuel staging. They attributed this to
NOx destruction by the rebum process and the reduced overall fuel nitrogen content. McCarthy
5-9
-------
Review of Reburn Technology
Table 5-2
Reburn Experimental Parameters
GENERAL
Organization
SPECIFIC NAME OF ORGANIZATION THAT CARRIED OUT EXPERIMENTAL MEASUREMENTS
Investigators)
AUTHORS OF SPECIFIC REFERENCE
Date (s)
DATES OF REFERENCES ON EXPERIMENTS
OVERALL SYSTEM
Furnace Type
SCALE OF EXPERIMENTAL COMBUSTION STUDY (LAB., PILOT, FULL SCALE)
Furnace Size
GEOMETRIC DIMENSIONS OF COMBUSTION CHAMBER
(MBTU/hr)
FUEL FIRiNQ CAPACITY IN MILLIONS OF BTU/HR.
Carbon Loss Calc. (YIN)
WERE MEASUREMENTS/CALCULATIONS CARRIED OUT TO CALCULATE CARBON LOSS?
Heat Flux Change Calc. (Y/N)
WERE EXPERIMENTAL MEASUREMENTS MADE TO DETERMINE THE COMBUSTOR WALL HEAT FLUX
PRIMARY COMBUSTOR
Fuel Type
SPECIFIC FUEL TYPE (GAS. OIL, OR COAL) IN PRIMARY COMBUSTOR
Varied Fue! (Y/N)
WAS FUEL TYPE VARIED
SR Primary
STOICHIOMETRIC RATIO AT PRIMARY COMBUSTOR EXIT
NOx Prtmaiy Exit (3 % 02)
RANGE OF NOX EMISSIONS AT PRIMARY COMBUSTOR EXTT
In-Fumace Species Data (Y/N)
WERE EXPERIMENTAL MEASUREMENTS MADE OF SPECIES CONCENTRATION IN PRIMARY ZONE?
REBURN SECTION
Reburn Fuel
SPECIFIC FUEL TYPE (GAS, OIL, OR COAL) IN REBURN COMBUSTOR SECTION
% Reburn Fuel
PERCENTAGE OF TOTAL COMBUSTOR ENERGY INPUT IN REBURN COMBUSTOR SECTION
Reburn Fuel Nitrogen Content
RANGE OF FUEL NITROGEN CONTENT OF REBURN FUEL
Fuel Transport Gas
SPECIFIC GAS USED FOR REBURN FUEL TRANSPORT
SR Rebum
STOICHIOMETRIC RATIO AT EXIT OF REBURN ZONE
No. of Rebum Stages
NUMBER OF COMBUSTOR JET STAGES IN REBURN COMBUSTOR SECTION
Rabum Temperature
RANGE OF MEASURED TEMPERATURES IN REBURN COMBUSTOR SECTION
Rebum Residence Time (sec)
RANGE OF GAS RESIDENCE TIME IN REBURN COMBUSTOR SECTION
Jet Mixing Studied (Y/N)
WAS JET MIXING OF REBURN JETS STUDIED IN REBURN ZONE?
In-Fumace Species Data (Y/N)
WERE EXPERIMENTAL MEASUREMENTS MADE OF SPECIES CONCENTRATION IN REBURN ZONE?
BURNOUT ZONE
SR Exit
STOICHIOMETRIC RATIO AT END OF BURNOUT ZONE OR COMBUSTOR EOT
NOx Exit (3%02)
RANGE OF NOX EMISSIONS MEASURED AT COMBUSTOR EXIT
In-Fumace Species Data (Y/N)
WERE EXPERIMENTAL MEASUREMENTS MADE OF SPECIES CONCENTRATION AT COMBUSTOR EXIT?
Rebum Efficiency (%)
PERCENTAGE DECREASE IN NOX LEVEL FROM PRIMARY COMBUSTOR SECTION
NOx Reduction Efficiency (%)
PERCENTAGE DECREASE IN NOX LEVEL FROM BASELINE COMBUSTOR VALUE
5-10
-------
Review of Rebum Technology
Table 5-3, Sheet 1
Reburn Experimental Data Summary
GENERAL
Organfzation
Babcock & Wilcox
Istiikawajimi-Harima
ishikavvajimi-Harima
Ishikaw^i'imi-Harima
IshiKawajimi-Harima
investigators)
Farzan and Rogers,
Miyamae, et. al.
Miyamae, et. al.
Miyamae, et al.
Miyamae, et, al,
Eckhart, et al.
Date (s)
1989
1985
1985
1985
1985
OVERALL SYSTEM
Furnace Type
Pilot Scale
Pilot Scale
Pilot Scale
Pilot Scale
600 MW Steam Gen,
Wall- fired
Furnace Size
Cyclone Burner
3 x 3 x 10m
3 x 3 x 10m
3 x 4,5 x 11m
-
(M BTU/hr)
up to 6
-40
- 80
- 60
-
Carbon Loss Calc, (Y/N)
Y
N
N
N
N
Heat Flux Change Calc. (Y/N)
N
N
N
N
N
PRIMARY COMBUSTOR
Fuel Type
Goal, CWF
Butane
Fuel Oil
Pulverized Coal
Oil(70%)/Coai(30%)
Varied Fuel (Y/N)
Y
N
N
N
N
SR Primary
1,0-1,2
0,7-1.3
0.8- 1.2
0.85 - 0,95
0.84- 0,91
NOx Primary Exit (3 % 02)
900-1200
-
-
,
-
In-Furnace Species Data (Y/N)
N
N
N
N
N
REBURN SECTION
Rebum Fuel
gas, oil, coal
Butane
Fuel Oil
Pulverized Coal
OH
% Reburn'Fuel
15- 35
5-18
5-18
0-16
0-10
Rebum Fuel Nitrogen Content
0-1.5
-
0*02%
1.50%
0.24%
Fuel Transport Gas
air
-
-
Air/ Flue gas
-
SR Rebum
0,85- 0.95
-
-
-
No. of Rebum Stages
1
1
1
1
1
Rebum Temperature
-
-
-
-
-
Rebum Residence Time (sec)
0.5- 0.0
-
-
-
-
Jet Mixing Studied (Y/N)
N
N
N
H
N
In-Fumace Species Data (Y/N)
N
N
N
H
N
BURNOUT ZONE
SR Exit
1.05- 1,2
-
-
NOx Exit (3%02}
250-550
35- 180
30- 95
85- 100
80-110
In-Fumace Species Data (Y/N)
N
N
N
N
N
Rebum Efficiency (%}
-
-
-
-
NOx Reduction Efficiency (%)
up to 57
-
-
-
-
5-11
-------
Review of Reburn Technology
Table 5-3, Sheet 2
Reburn Experimental Data Summary
GENERAL
Organization
Ishtkawajimr-Harima
Hitachl-Babcock
MHI
MHI
MHI
Mm
Investigators)
Miyamae, et. al.
Narita, et. al.
Murakami
Takahashi, et al
Takahashi, et al
Takahashi, et al
Date (s)
1985
1987
1985
1982/1981
1982/1981
1982/1981
OVERALL SYSTEM
Furnace Type
55 MW Steam Gen.
200 MWe Furnace
600 MW
Small Scale
Pilot Scale
Pilot Scale
Front- fined
Waif- fired
T- fired
Furnace
T-fired
T-fired
Furnace Size
-
-
-
0.8m dlam x 2.2m
•
-
(MBTU/hr)
-
-
-
-2.5
up to - 60
up to ~40
Carbon Loss Calc. (Y/N)
Y
Y
Y
H
N
N
Heat Flux Change Calc. (Y/N)
N
N
N
N
N
N
PRIMARY COMBUSTOR
Fuel Type
Pulwsrized Coal
Puherized Coal
Oii/Coal(to 30%)
Propane/Coal
Gas/ Fuel Oil
Pulverized Coal
Vaned Fuel (Y/N)
N
Y (5 types)
Y (%coal)
H
Y (6 oils)
N
SR Primary
0.8-1.05
.
.
.
NOx Primary Exit £3 % 02)
.
-
.
.
-
In-Furnace Species Data (Y/N)
N
N
N
Y
N
N
REBURN SECTION
Reburn Fuel
Pulverized Coaf
Pulverized Coal
Oil
Propane/Coal
Gas/Fuel Oil
Pulverized Coal
% Reburn Fuel
0- 1D
-
0-10
Q--20
0- -25
Reburn Fuel Nitrogen Content
2.00%
0.76-1.71%
-
0/1.0
0-0.9
-
Fuel Transport Gas
Air
Air
Flue gas
Air
Air
Air
SR Reburn
-
_
-
0.1-1.4
.
.
No. of Rebum Stages
1
1
1
1
1
1
Return Temperature
-
-
-
1850- 2350F
-
-
Reburn Residence Time (sec)
-
-
-
0.1- 0.8
-
-
Jet Mixing Studied (Y/N)
N
N
N
N
N
N
tn-Fumace Species Data (Y/N)
N
N
N
Y
N
N
BURNOUT ZONE
SR Exit
_
-
-
-
1.05-1.17
1.1-1-23
NOx Exit (3%02)
140-240
180- 260
72- 180
-
.
In-Fumace Species Data (Y/N)
N
N
N
Y
Y
Y
Reburn Efficiency (%)
.
.
.
10-95
.
.
NOx Reduction Efficiency (%)
-
-
-
.
-
5-12
-------
Review ofReburn Technology
Table 5-3, Sheet 3
Reburn Experimental Data Summary
GENERAL
Organization
Riley Stoker
Riley Stoker
Riley Sioker
Acurex/EPA
Acurex/EPA
Acurex/EPA
investigator(s)
Lisauskas, et a!.
Penterson, et al.
Penterson, et al.
Mulholland &
Mulholland, et. al.
Mulholland, et. a\.
Srivastava
Dale (s)
1985
1989
1939
1987
1987/1985
1987/1985
OVERALL SYSTEM
Furnace Type
Pilot Scale
Pilot Scale- 1GT
Pilot Scale- MSW
Package Boiler
Package Boiler
Package Boiler
End- fired
Simulator
Simulator
Simulator
Furnace Size
18 x 18 x 60ft
4.5 x 3 X 14ft
3x11.75x17ft
0.6m diam x 3 m
0,6m diam x 2.3m
0.6m diam x 2.3m
(MSTU/hr)
100
1.7
3
- 2
-2-27
- 2
Carbon Loss Calc. (Y/N)
N
N
N
N (CO data)
N
N
Heat Flux Change Calc. (Y/N)
M
N
N
H
N
N
PRIMARY COMBUSTOR
Fuel Type
Pulverized Coal
Simulated MSW
MSW
Natural Gas
Natural Gas
Fuel Oil
Varied Fuel (Y/N)
Y (4 types)
N
N
N
N
Y (light, heavy)
SR Primary
*
-
0,95-1.31
0.78
0.9-1,2
0.95-1.16
NOx Primary Exit (3% 02)
-
100- 300
120- 16S
.
43-430
43-430
In-Fumace Species Data (Y/N)
N
N
N
N
Y
Y
REBURN SECTION
Reburn Fuel
Puherized Coal
Natural Gas
Natural Gas
Natural Gas
Natural Gas
Fuel Oil
% Reburn Fuel
10-28
7- >30
7-15
0-35
0-3?
0-29
Reburn Fuel Nitrogen Content
1,03-1.44%
-
-
0
0-1.0%
0-0.5%
Fuel Transport Gas
Air
-
-
-
-
-
SR Reburn
0,75-10
0.6- 1.22
0.6- 1.25
0.85- 1,1
0-7- 1.1
0.8- 1.1
No. ofRebum Stages
1
1
1
1
1
1
Reburn Temperature
1950-2400
-
.
2000- 2400 F
2000- 2400F
Rebum Residence Time (sec)
.
1- 5.2
1.2- 5.2
-
0- 0,4
0.1- 0.4
Jet Mixing Studied (Y/N)
N
N
N
N
Y (injection pts)
N
In-Fumace Species Data (Y/N)
N
H
N
N
Y
Y
BURNOUT ZONE
SR Exit
1,2
-
-
1.15
1.0- 1.3
1,0- 1.3
NOx Exit (3% 02)
175- 400
70- 180
71- 142
140- 275
-
-
In-Fumace Species Data (Y/N)
N
N
N
N
Y
Y
Rebum Efficiency (%)
.
.
-
.
-0- 75%
-0- 40%
NOx Reduction Efficiency (%)
up to 75
up (o 70
up to 50
up to 50
-
-
5-13
-------
Review of Reburn Technology
Table 5-3, Sheet 4
Reburn Experimental Data Summary
GENERAL
Organization
Acurex/EPA
Acurex
Acurex
EER
EER
EER
Investigators)
Mulholland, et, at.
Kelly, et al
Brown & Kuby
McCarthy, et. at.
Green, et- al.
Clark.et. al.
Qvermoe, et al
Chen, et. al.
Date (s)
1987/1985
1805/1982
1885
1987/1S85
1985/1984
1982
OVERALL SYSTEM
Furnace Type
North American
Lab Subscale
Subscale Combustar-
Pilot Scale Comb
Lab Scale Comb
Lab Scale Comb.
Boiler
Combust or
Engine Exh. Simulator
(DownSred)
(Downfired)
(Downfired)
Furnace Size
-
8 in cflam x 72iri
4 to 6 in duct
4 x 4 x 26 ft
6ln diam x 4ft
6 in diam x 4 ft
(MBTU/hr)
- 2,5
- 0,055
-0.1
- 10
up to 0.083
- 0.07
Carbon Loss Calc. (Y/N)
N
Y
Y
Y
N
N
Heat Flux Change Calc. (Y/N)
N
N
N
H
N
N
PRIMARY CQM8USTOR
Fuel Type
Fuel Oil
Pulverized Coal
Gas
Gas/Oil/Coal
Gas/Coal
Gas/Puherized Coal
Varied Fuel (Y/N)
Y (No. 2, 5)
Y (5 types )
N
Y
Y
Y (2 Coals)
SR Primary
.
0.9-1.2
0.9-1.2
1.5-1.3
1,05-1.4
NOx Primary Exit (3 % 02)
.
.
430* 2800
~ 200-1000
-
In-Fumace Species Data (Y/N)
H
Y
Y
Y
N
N
REBURN SECTION
Rebum Fuel
Natural Gas
Natural Gas/ Coal
Gas
Gas/Oii/Coal
Gas/Coal
Propane/Coal
% Rebum Fuel
0-20
25- 50
10-38
0-30
0-36
-
Rebum Fuel Nitrogen Content
0-1.3%
0.7-1,5%
.
1.17-1.94
0.88-1.88
0-1,67
Fuel Transport Gas
-
-
-
Air/Flue Gas
Aif/N2
Air
SR Rebum
0,88-105
0.6-1,2
0.9- 1.05
0,7-1.0
0.7-1 25
0.8- 1.1
No. of Reburn Stages
1
1
1
1
1
1
Rebum Temperature
- 230QF
~ 2300-2600P
-
-2550F
"2200-2800F
-
Rebum Residence Time (sec)
- 0,2
0.5- 1-6
0.25- 0.3
0.8- 1.0
0.14- 0.75
-
Jet Mixing Studied (Y/N)
Y (injection pts)
N
N
Y
Y
N
In-Fumace Species Data (Y/N)
H
Y
Y
Y
Y
N
BURNOUT ZONE
SR Exit
-
1.2
-
1.1-1.35
- 1,25
1.05-1.3
NOx Exit (3%02)
-
up to 860
-
-200- 500
-
-
In-Fumace Species Data (Y/N)
U
Y
Y
Y
N
Y
Rebum Efficiency {%)
" 0- 50%
-
-
up to 50
up to 70
up to 60
NOx Reduction Efficiency (%)
up to 70
up to 70
•
-
5-14
-------
Review of Reburn Technology
Table 5-3, Sheet 5
Reburn Experimental Data Summary
GENERAL
Organization
EER
ENEL
KVB
KVB
KVB
kvs
lnvestfgator{s)
Clark,et ai.
Baldacci, et, al
Bortz &
Yang, et. al.
Yang, et, al.
Yang, et. al.
Chen, et. al.
OfTen (EPRI)
Date (s)
1082
1988
1987
nm
1964
1984
OVERALL SYSTEM
Furnace Type
Lab Scale
Lab Scale
Lab Scale Comb
Lab Scale Comb
Lab Scale Comb
Lab Scale Comb
Reactor
Reactor
(Downs red)
(Upfired)
(Cement Kiln Sim.)
Furnace Size
-
-
18 In dlam
14 In efiam x 7 ft
8 In dlam x 7 ft
Sin diam x 12ft
(MBTU/ftr)
-
0,2
0.5
up lo 0.25
0.13
0.17
Carbon Loss Calc. (Y/N)
N
N
N
N
N
N
Heat Flux Change Calc. (Y/N)
N
N
N
N
N
N
PRIMARY COMBUSTOR
Fuel Type
Gas
Heavy Fuel Oil
Pulverized Coal
Natural Gas
Natural Gas
Natural Gas/ Coal
Varied Fyet (Y/N)
N
N
N
N
H
H
SR Primary
1,1
1,0-1,115
0.S5-1.2
0,4-1,2
0.9-1.7
1-1,23
NOx Primary Exit (3 % 02)
up to ~ 500
950
90- 500
-
-
IrvFiimace Species Data (Y/N)
N
N
N
N
N
N
REBURN SECTION
Reburn Fuel
Methane
GPL
Gas (3 Types)
Natural Gas
Natural Gas
Coal
% Rebum Fuel
-
5- 30
0-20
up to 70
0-70
10- 60
Return Fuel Nitrogen Content
0- 0.5
-
.
-
0,84
Fuel Transport Gas
N2
-
-
-
-
Air
SR Rebum
0.75-1.25
0,8- 1.0
0.8-1,05
0.5- 1.15
0.5-1,15
0.7-1.2
No, of Rebum Stages
1
1
1
1
1
1
Rebum Temperature
-2100-3100F
> 2400F
2200- 2400F
-
-
-
Rebum Residence Time (sec)
-
0.1-0.4
0.25- 0.5
-
-
-
Jet Mixing Studied (Y/N)
N
N
N
Y (varied Inj. pt.)
Y(varied inj. pt)
N
in-Fumace Species Data (Y/N)
Y
N
N
N
N
N
BURNOUT ZONE
SR Exit
- 1,2
1,02-1,15
1.2
-
~ 115
- 1.22
NOx Exit (3% 02)
-
-
- 150-500
-
•
-
ln-Fumace Species Data (Y/N)
Y
N
Y
Y
N
N
Rebum Efficiency (%)
up to 50
_
-
.
.
.
NOx Reduction Efficiency (%)
-
up to 56
-
up to 90
up to 90
up to 90
5-15
-------
Review of Return Technology
Table 5-3, Sheet 6
Reburn F.xperimenta
Data Summary
Organization
KVB
Hitachi Zosen
Hitachi Zosen
IFRF
MIT
MIT
Investigators)
RadaK, et, al.
Okigami, et, al.
Okigami, et. al.
Knni, et at.
Toqan, et. al.
Farmayan,
-------
Review of Return Technology
et al. (1987) stated that if flue gas recirculation is used, reburn fuel fraction should be set at 20%.
If flue gas recirculation is not used, the rebum fuel fraction should be set at approximately 30%.
Maringo et al. (1987) stated that for large cyclone combustor designs, the rebum fuel fraction
should range between 15 and 25%. From this range of experience, the authors concluded that
25% reburn fuel would represent a good initial choice.
Reburn Fuel Nitrogen Content As one might expect, this parameter has a significant effect
on NOx emissions. The following are specific comments:
• McCarthy et al. (1987) stated that for highest rebuming efficiency, a nitrogen-free reburn
fuel should be used.
• Mulholland et al. (1987) noted that NO* reduction via rebuming is strongly dependent on
fuel nitrogen. They concluded that NOx reduction is adversely influenced by the presence of
bound nitrogen in the reburn fuel, especially in cases where the primary zone NOx is low.
• Green et al. (1984,1985) found that nitrogen-free rebum fuels were the most effective for
NOx reduction.
Reburn Fuel Transport Gas/Mixing Medium. Experimental work by several investigators
led to the following results:
• Green et al. (1984, 1985) stated that an inert rebuming fuel transport medium (oxygen free)
is desirable since less rebuming fuel is required to attain optimum stoiehiometry. Overmoe
et al (1985) stated that one should minimize the available transport oxygen and in particular
flue gas recirculation should be used as a transport medium, if possible. McCarthy et al.
(1987) concluded that coal fired systems, if feasible, should use flue gas recirculation as the
rebuming coal transport medium (with approximately 20% rebum fuel). However, if air
must be used for the rebuming fuel transport, they stated that the rebuming fuel ratio should
be increased to 30%.
• Takahashi et al. (1981) noted that the use of flue gas recirculation in the rebum zone lowers
the NOx emissions for the following reasons:
1) It improves the mixing of the reburn fuel into the main combustion gas stream.
2) It causes the production of radicals (CnHm) that improve the NOx removal process in the
rebum zone.
3) The water contained in the recirculated flue gas has the effect of suppressing the
production of soot through a water gas reaction, resulting in a decreased amount of
smoke, dust, etc.
• Maringo and McElroy (1987) reported that adding flue gas recirculation flow to the rebum
zone burners improved the NOx reduction capability of their pilot scale facility. Specifically,
with an addition of just 10% flue gas recirculation, they were able to show a NOx reduction
improvement of about 13% across a wide range of natural gas rebum fuel inputs.
5-17
-------
Review of Return Technology
Stoichiometric Ratio in the Reburn Zone. Most investigators agree that the stoichiometric
ratio in the reburn zone is one of the most important parameters for the reburn efficiency of a
combustion system. This value was varied over a wide range for the tests listed in the data
summary. Some of the comments regarding this variable are as follows:
• Toqan ct al. (1987) found an optimum stoichiometric ratio of 0.77 based on a theoretical
model; however, based on their experiments 0.91 was optimum.
• Farmayan et al. (1985) found that optimum NOx reduction was achieved with a reburn zone
stoichiometric ratio between 0.77 and 0.83.
• Mulholland et al. (1987) stated that NOx reduction via rebuming is strongly dependent on
reburn stoichiometry.
• Maringo et al. (1987) postulated that, for large-scale systems, the reburn zone in a cyclone
boiler should operate at a stoichiometric ratio of 0.85-0.95. This conclusion was based on
maximum NOx reduction from laboratory scale studies.
• Miyamae et al. (1985) noted that one of the dominant variables controlling NOx reduction by
reburning is the stoichiometry in the reburn zone.
• Lisauskas et al. (1985) showed a general decrease in exit NOx as reburn zone stoichiometry
was decreased from approximately 1,1 to 0.65.
• Green et al. (1984,1985) concluded that the reburning zone stoichiometry was optimized at
0.9.
• Takahashi et al. (1981) stated that NOx decomposition rate falls off rapidly after the reburn
zone stoichiometric ratio becomes greater than 0.9. Again, it should be noted that this
conclusion is based on the experimental results from small scale experiments.
• Eckhart et al. (1989) achieved maximum NOx emissions reduction at reburn zone
stoichiometrics of about 0.85 in their pilot scale experiments.
Reburn Zone Temperature. The general consensus of the experimental investigators on the
subject of temperature in the reburn zone is that the effectiveness of the reburn process increases
with increasing temperature. As shown in the experimental data summary, a wide range of
temperatures were reported for the reburn zone. The following comments apply to this
parameter:
• Takahashi et al. (1981) stated that the reburn temperature should be at least 1650 °F and that
temperatures higher than 2350°F are preferred.
• Green et al. (1984) concluded that the reduction of NOx increases with increasing
temperature in the range from 2400 to 2900 °F.
• McCarthy et al. (1987) and Overmoe et al. (1985) stated that the reburning fuel should be
injected into as hot a furnace environment as possible.
5-18
-------
Review of Reburn Technology
From this limited data, it can be concluded that the gas temperatures in the reburn zone should be
as high as possible, without creating thermal NOx.
Reburn Zone Residence Time. The reburn experimental results indicate that the residence
time in the reburn zone varied over a wide range, from 0.1 to 1 second. Despite this variation in
residence times, there is general agreement that this parameter is important for reburn zone
design. Some comments regarding this variable are as follows;
• Maringo and McElroy (1987) varied the reburn zone residence time from 0.5 to 0.8 seconds
and found that longer residence time provided the greatest NOx reduction.
• Mulholland et al, (1987) concluded that NOx reduction increased with rebum zone residence
time, with more rapid changes after 50 ms, and leveling off at 300 to 400 ms. Mulholland
and Hall (1985) noted that there was a practical design constraint of 500 ms or less residence
time in the reburn zone.
• McCarthy et al. (1987) and Overmoe et al. (1985) concluded that one should maximize the
rebum zone residence time. Green et al. (1984, 1985) stated that this variable has a strong
impact on NOx reduction efficiency, increasing with time (from a range of 100-750 ms).
• Maringo et al. (1987) postulated that, based on pilot and field scale tests, a 50-60% reduction
in NOx could be achieved at residence times greater than 450 ms.
• Lisauskas et al. (1985) found that NOx emissions decreased as residence time increased.
Thus, from the general consensus, it appears that a reburn zone residence time of at least 500 ms
is desirable, with longer times desirable, if practical. The difficulty of obtaining an accurate
value, or interpretation, of residence time in practical sized combustors where the flow is
generally not of the one dimensional plug flow type should be pointed out. This point was made
in the pilot plant study of Maringo and McElroy (1987), where they presented some serious
questions regarding the accuracy of their residence time calculations. In practice, detailed
measurements of furnace gas velocity and direction or other types of residence time
determination are highly desirable.
For the design of a full-scale reburn system, Borio et al. (1989) stated that the key design criteria
for the reburn system include;
• Inject reburn fuel into as high a temperature zone as possible, commensurate with releasing
all fuel-bound nitrogen upstream of the reburn zone.
• Maintain average stoichiometry between 0.90 and 0.95.
• Permit a small amount of O2 to promote the formation of OH and H radicals.
• Maintain the residence time between 0.5 and 0.7 seconds.
• Maximize entrainment, mixing, and dispersion of reburn fuel.
• Avoid direct fuel impingement on boiler walls.
5-19
-------
Review of Return Technology
• Minimize the number of required boiler penetrations,
• Locate fuel injection nozzles to minimize boiler/structural steel modifications.
• Provide for maximum flexibility of reburn fuel jet direction and flow rates.
• Provide a fuel flow rate control system with automatic load following capability
• Provide safeguards for fail-safe operation.
Reburn Zone Mixing. Since many of the experimental studies listed in Table 5-3 were
conducted in small scale facilities, reburn zone mixing was not studied as a separate design entity
in a majority of the experiments. However, mixing was recognized as an important reburn
design variable during several investigations. The following comments emphasize this point:
• Maringo and McElroy (1987) varied the mixing and residence time in the reburn zone by
moving overfire air ports and varying spin vanes in the reburn zone burners. They found that
lower swirl from the reburn burner enhanced the NOx reduction efficiency of the system.
• Mulholland et al. (1987) stated that uniformity of the reburn zone stoichiometry should be
important for optimizing NOx reduction. They varied injection geometry and location via a
reburn fuel boom inserted in the combustor reburn section. However, due to the existence of
large scale turbulent structures in their experimental apparatus, the reburn fuel injection
design did not influence reburning effectiveness.
• McCarthy et al. (1987) found that when recirculated flue gas is the transport medium, the
reburn fuel should be injected with jet penetration greater than 70% of the furnace depth and
coverage of the furnace cross section should be thorough. They also noted that if air must be
used for the reburning fuel transport, one should decrease the mixing rate of the reburning
jets. Overmoe et al. (1985) and Green et al. (1984,1985) stated that rapid mixing of the
reburn fuel led to more effective NOx reduction.
• Farmayan et al. (1985) stated that the mode of reburn fuel injection was a principal variable
in their experimental reburn systems.
It was concluded that the mixing process had great importance in the design of commercial
reburn system. For this reason isothermal flow modeling studies reported by Borio et al. (1989)
were conducted for the design of the reburn system for this project.
5-20
-------
Review of Return Technology
Burnout Zone
Key parameters in this zone are:
• Stoichiometric ratio
• Temperature
• Residence time
• Jet mixing
Although this zone is an important part of a practical reburn combustion system, very little
general information was found in the literature reviewed (Table 5-3). One exception was Knill
(1987) who noted that the residence time in the burnout zone must be sufficient for ensuring
complete fuel burnout. He concluded that this should not be a problem in gas flames, but in coal
flames char burnout may be affected by the size of this zone.
For the design of a full-scale reburn system Borio et al. (1989) stated that the key design criteria
for the burnout zone include:
• The injection of burnout air in as low a temperature zone as possible commensurate with
obtaining fuel burnout before entering the first convective surface.
• Provision for rapid mixing of air to minimize pockets of unburned fuel
• The avoidance of direct air impingement on furnace walls
• The minimization of final excess oxygen commensurate with obtaining good fuel burnout
• Provision for a residence time of 0.6 to 0.8 seconds.
• Minimization of boiler penetrations while providing maximum flexibility for air jet direction
and velocity.
5-21
-------
-------
6
DESCRIPTION OF THE HOST UNIT
The Ohio Edison System's Niles Plant is located in Northeastern Ohio on the southwest border
of the city of Niles, Weathersfield Township, Trumbull County. The plant occupies 130 acres
along the southern bank of the Mahoning River. The main power plant structure covers an area
of approximately 166 feet by 200 feet and consists of two cyclone coal-fired boilers and steam
turbine generating units. Approximate site elevation is 870 feet above sea level.
Both units were placed in commercial operation in 1954. Steam turbine conditions are 1450
psig, 1000°F main steam and 384 psig, 1000°F reheat steam. The original design rated gross
capacity for each steam turbine was 125 megawatts (MW). Effective January 1985, the
demonstrated capacity for each unit was decreased to 108 MW net, which is equivalent to
approximately 115 MW gross. The unit capacity rating decrease from the original design of 125
MW to 115 MW was necessitated by continual combustion problems associated with operation
of the cyclone-fired boilers.
At low loads, proper slag tap flow is a concern on these cyclone boilers. Any operating
conditions that cause reduced lower furnace temperatures can result in poor slag tap flow. Loads
below 55 MW net are possible under normal operating conditions. However, operation is
reduced to three cyclones in service below approximately 75 MW net. In the event only three
cyclones are operating and a second cyclone is forced out-of service, unacceptable steam
temperature swings can occur.
The boilers burn primarily high sulfur bituminous coal. They are pressurized radiant furnaces,
natural circulation, reheat type boilers with four (4) 9 feet by 12 feet cyclone burners on the front
wall and a primary and secondary furnace. The back wall of the secondary furnace has studded
waterwall tubes which are coated with refractory to provide sufficient flue gas temperatures in
the back passes of the boiler to maintain steam temperatures. Boiler design steam conditions at
the turbine governor valves wide-open design point are 885,000 lb/hr, 1650 psig and 1000°F. A
side elevation of the boiler is shown in Figure 6-1. Boiler data and operations data are given in
Table 6-1.
* For those more familiar with metric units, see the conversion table on Page v.
6-1
-------
Description of the Host Unit
Elev. 876-0"_
jSECONDARY
SUPERHEATER
OUTLET
-REHEAT
OUTLETS
-INLET HE4DER
REHEATER SECTION
SECONDARY
SUPERHEATER
INLET
BABCOCK & WILCOX RADIANT BOILER
WITH CYCLONE FURNACE
Mo*, cent. it«am cap-, lb/fir, #oeh, .885,000
Otiign pressure, p«i 1,450
PfflMWf* or lune'heoter ol/H«V pi\ . . 1,485
firsal lot. «teom remp., F., 1000
Reheat steam temp., f 1000
Fe«dwcri*r temp,, f 485
CYCLONES
RB-139-500-1
Figure 6-1
Ohio Edison Niles Unit No. 1
6-2
-------
Description of the Host Unit
Table 6-1
Description of Host Boiler: Niles Unit No. 1
Utility
Ohio Edison Company
Unit Identification
Niles Unit No. 1
Boiler Type
Cyclone Fired, Natural Circulation, Reheat
Manufacturer
Babcock & Wilcox
Date in Service
January 1, 1954
Boiler Nameplate Rating
125 MW Design Capacity (Turbine Generator)
Boiler Steam Conditions
1000°F, 1650 PSIG
Main Steam Flow
885,000 Ib/hr
Main Steam Conditions
1000°F, 1450 PSIG
Reheat Steam Conditions
1000°F, 384 PSIG
Net Heat Rate
9,465 Btu/'NKWH (1993)
Number of Cyclones
4
Coal Crusher Manufacturer
Pennsylvania Crusher
Total Heat Input @ Rated Capacity
1,199 mmBtu/Hr (Design Coal)
Heat Release
86,000 Btu/Sq. Ft./Hr.
Furnace Width
36'-0"
Gas Temperature Leaving Air Heaters
270°F
Soot Blowers
18 Copes-Vulcan, Service Air
Ash Removal
Pneumatic Transport From Air Heater and ESP
Hoppers
Air Heater
Babcock & Wilcox Tubular
Equivalent Availability
84.86% (1993)
Unit Capacity Factor
68.61% (1993)
Boiler Design Efficiency
90.3%
Boiler Actual Efficiency
87.81% (1993)
6-3
-------
Description of the Host Unit
Four (4) ceramic-extractive oxygen analyzers were originally installed across the back wall of
the secondary furnace at the turbine floor level. Each extractive sample probe was located in-
line with one of the cyclones. Because the combustion gases remain stratified as they flow
through the furnace, each oxygen analyzer provided an indication of the oxygen level resulting
from combustion from one cyclone. Therefore, the analyzers could be used to indicate relative
balance of combustion air between cyclones and were used to roughly tune the fiiel-to-air ratios.
The fuel flow rate is determined by volumetric coal feeders.
The original oxygen analyzers plugged with slag frequently and, as a result, provided erroneous
readings. The analyzers were never integrated into the boiler control system. Adjustments could
provide short-term optimized conditions, but balance of the combustion air between cyclones
was not maintained. Due to these problems, the analyzers were removed in October 1992 and
new zirconium oxide fuel cell type oxygen analyzers were installed at the air heater inlet.
Each boiler has an electrostatic precipitator (ESP). These were installed in 1981. The ESP's
have sufficient capacity to normally operate with only three of their five fields activated. ESP
design features are given in Table 6-2. The boilers originally had multi-cone type mechanical
dust collectors. Flue gases from each boiler are exhausted to a 393 foot chimney which contains
two 11 foot diameter steel-lined flues, each with a design capacity of 330,189 cubic feet per
minute. Two 300 foot exhaust stacks connected to the main plant structure were
decommissioned in 1981 after the ESP's were installed.
Table 6-2
Description of Electrostatic Precipitator at Niles Unit No. 1
Manufacturer
Wheelabrator-Frye
Installation Date
1981
Number of Fields
5
Collection Surface, ft2
278,168
Specific Collection Area, ft2/1000 acfm
520
Design Gas Temperature, °F
270
Velocity through Precipitator, ft/sec
<4.5
Efficiency, percent
99.0
Method of Ash Removal
Dry Pneumatic
Ash Collection and Storage system
Pneumatic Transport to Wet System and
Pumped to Pond
6-4
-------
7
REBURN SYSTEM DESIGN
Flow Modeling and Reburn System Conceptual Design
Rebum accomplishes in-furnace reduction of NOx by creating a reducing zone downstream of
the primary combustor by a second introduction of fuel as shown schematically in Figure 7-1.
The reducing zone creates intermediate chemical compounds composed of carbon, hydrogen,
oxygen, and nitrogen which subsequently react with NOx formed in the primary combustion zone
to convert NOx into the desired final product, molecular nitrogen. Unburned fuel leaving the
rebum zone is burned to completion by air introduced in the burnout zone.
This section describes the flow modeling studies and the design of the original reburn system.
Following parametric testing of the original system, a modified reburn system was developed.
The design basis and equipment design for the modified reburn system are discussed in Section
11.
The effectiveness of reburn for performing the reburn chemical reactions and burnout of the
reburn fuel depends upon good mixing of the reburn fuel with the NOx-containmg gases and
good mixing of the burnout air with the unbumed combustibles leaving the reburn zone. To
assist in the design of the rebum system, an isothermal flow model study of the unit was
performed. The specific tasks of the modeling effort were:
1. Construct a 1/9 scale isothermal flow model of Ohio Edison's Niles Station Unit No. 1.
2. Using the flow model, map the aerodynamic flow fields within the furnace in its existing
baseline configuration.
3. Develop and evaluate potential reburn fuel and burnout air injection system
configurations and operating parameters, based on the results of the baseline aerodynamic
characterizations, ideal reburn system operating conditions, and the geometric/physical
constraints imposed by the unit.
4. Recommend reburn fuel and bumout air injection system designs and operating
condition.
Experimental Test Program
The flow modeling was performed in the one-ninth scale model of the Niles Unit shown in
Figure 7-2. The model, constructed primarily of clear plastic, encompassed the entire furnace
from the cyclone combustors to the vertical furnace outlet plane. The cyclone combustors were
7-1
-------
Return System Design
SUPERHEAT/REHEAT
CONVECTIVE PASSES
GAS
RECIRCULATION
FAN
burnout air
CYCLONES
Figure 7-1
Schematic of the Ohio Edison Rebum Process
7-2
-------
Reburn System Design
Figure 7-2
One-ninth Scale Flow Model of Niles Unit No. 1
7-3
-------
Reburn System Design
designed to produce the correct swirl and momentum (axial and tangential) entering the primary
furnace. Upper furnace radiant and convective heat transfer surfaces were also modeled. A
header system fed by a high pressure blower controlled the introduction of smoke or tracer gas to
any one or a combination of reburn fuel or burnout air injection nozzles. Flue gas flow through
the model was simulated by drawing air through it with a large induced draft fan.
An initial series of isothermal flow modeling tests characterized the baseline gas flow
characteristics of the boiler. Following the establishment of the baseline reference data, flow
modeling of the reburn system consisted of two screening level tests and a final configuration
characterization test for both the fuel and burnout air:
• Screening Level 1 - Flow visualization (with smoke) of a large number of reburn
fuel/burnout air injector configurations.
• Screening Level 2 - Mixing study tests on best configuration candidates from Level 1.
• Level 3 - Final injection configurations based on three-dimensional analysis.
In Screening Level 1, smoke flow visualization tests were performed for each candidate injection
system at simulated full and 70% load furnace operating conditions. Each injection
configuration was evaluated at three injection velocities, three tilts, and a number of yaws. After
initial selection of the best injection configurations, Screening Level 2 consisted of methane
tracer gas injection tests with concentrations measured by a laser absorption spectrophotometer.
Final injection configurations were determined from results of detailed velocity profile
measurements using three dimensional (five-hole pitot tube) analysis techniques. Details of the
instrumentation used to carry out these measurements are given by Anderson et al. (1986),
Test Results. Baseline furnace velocity fields were measured at Test Planes TP1 through TPS
shown on Figure 7-3. Data were obtained under flow conditions simulating boiler operation at
100% and 70% MGR. Of particular interest for the design of the injectors is the bulk flue gas
flow field at the entrance to the reburn zone and burnout zone. Profiles for these planes, TP1 and
TP4, are shown in Figures 7-4 and 7-5. Test Plane 1 is characterized by two high velocity areas
along the rear wall of the boiler, corresponding to the flow originating from the two lower
cyclones. The outlet from the lower cyclones is partially below the dividing wall. As a
consequence, a large portion of the gases exiting these cyclones passes unimpeded under the
division wall into the secondary furnace.
At Test Plane 4 slightly higher velocities (than average) were found along the rear wall. The
side-to-side velocity distribution at this plane shows more flow along the right side of the unit
than the left. The cause is uncertain, but it is speculated that it is a function of swirl induced by
the cyclones. The side-to-side velocity distribution is more uniform at Test Plane 5.
Fourteen reburn fuel injector configurations were evaluated at three different yaws in the first
reburn. fuel injector screening test series. Simulated injection velocities ranging from 100 ft/sec
to 300 ft/sec were evaluated. Reburn fuel injection nozzle diameters for these velocities and
7-4
-------
Reburn System Design
Figure 7-3
Test Plane and Injector Locations
7-5
-------
Reburn System Design
TEST NO: REBURN-BASELINE-0002 TEST PLANE: 1
Normalized Velocity Profile
Figure 7-4
Baseline Axial Velocity Contours, Plane 1
7-6
-------
Reburn System Design
m
r-
«
n
T
Norma I Ized Velocity Profile
Figure 7-5
Baseline Axial Velocity Contours, Plane 4
7-7
-------
Rehurn System Design
intermediate velocities were sized for flow rates of natural gas mixed with flue gas recirculation
(FGR) flowing at a rate equal to 10% of the total flue gas flow rate. It was found that, at all
velocities tested, three injectors along the rear wall were insufficient to cover the plane. Five
injectors along the rear wall were found to be the best configuration. However, the two
outermost injectors suffered from jet wall attachment when injecting straight into the furnace.
Yawing these injectors toward the center of the unit eliminated this problem. Injection from the
side wall also provided generally good distributions, but no better than with the more economical
five rear wall injector configurations.
The addition of "pant-legs" dividers to the ends of the injection nozzle tips was found to be an
effective means of enhancing the dispersion of the reburn fuel jets. Pant-legs were found to
significantly improve the dispersion of the jet near the rear wall.
The most effective use of yawing was obtained by providing two fuel injection levels. Each of
the three inboard upper nozzles were split and yawed, using pant-legs, while the lower nozzles
were not split and allowed to penetrate to near the division wall. Since it would not make sense
to put pant-legs on the outermost sets of injection nozzles (being located next to the side walls),
these nozzles, both upper and lower, were yawed toward the center of the furnace.
The Screening Level 2 test matrix for the reburn fuel injectors was developed from the Screening
Level 1 results discussed above. Screening Level 2 mixing studies supported most of the
conclusions from the initial smoke flow visualization studies and permitted the selection of an
optimum reburn fuel jet configuration. The penetration and dispersion performance of the
injectors was a function of the flow field into which they were injected. Injectors that were firing
into the lower velocity segments along the rear wall of the furnace at a simulated velocity of 100
ft/sec were capable of penetrating all the way to the division wall, while those that injected into
the higher velocity zones, associated with the two lower cyclones, could not. It was found that
an injection velocity of 300 ft'sec was too high, resulting in jet impaction on the division wall
almost directly across from the point of rejection. Reducing the recirculated flue gas flow rate
below 10% generally resulted in reduced levels of dispersion. It was found that tilting the
nozzles down improved the overall dispersion of the jet at the outlet of the reburn zone, while
tilting upward reduced the dispersion. The configuration shown in Figure 7-6 was chosen as the
recommended reburn fuel injector configuration.
Locations for air injectors for the bumout zone were limited to the side walls and one central
location on the front wall because of interferences on the front wall of the unit. The choice of
candidate burnout air injection configurations/locations was also guided by the need to inject air
into a zone that was partially obstructed by cyclone burner hanger tubes. Figure 7-3 shows the
locations that were evaluated for burnout air injections.
Each burnout air configuration shown on Figure 7-3 was evaluated at 150 and 300 ft/sec, three
tilts (-20°F, 0°, +20°), and configuration specific yaws ranging between plus and minus 20°.
During all burnout air injection tests the recommended reburn fuel injection configuration was
installed and was in service.
7-8
-------
Reburn System Design
Division Wal]
100-150 fps
Bottom Nozzle
Plan View
;15'/
' :
100-150 fps
Top Nozzle
Plan View
Figure 7-6
Recommended Reburn Fuel Injector Configuration
7-9
-------
Reburn System Design
Smoke flow visualization tests of burnout air injectors showed trends consistent with the reburn
fuel injection system tests; i.e., the jet penetration and dispersion increased as the burnout air jets
were tilted into the flow and yawed for maximum dispersion.
The testing indicated that to effectively mix bumout air in the upper secondary furnace, a
burnout air injection velocity of 225 ft/sec should be used. The recommended burnout air
injector configuration is shown schematically in Figure 7-7. The configuration is represented by
locations "A" and "B" on each side wall with two nozzles, one above the other, at each location.
The injectors at "A" were directed straight in. Those at "B" were tilted down 10 degrees and
yawed 10 degrees toward the rear of the unit. To aid in field-tuning of the system each nozzle
was given a tilt and yaw capability of plus and minus 20 degrees.
Reburn System Design Requirements
The objectives of the reburn system design were:
(1) To meet performance criteria for effective NOx reduction while minimizing any impact
on boiler performance or boiler normal operation.
(2) To incorporate operational flexibility within the design to permit optimization of
performance in the field.
Injection of reburn fuel into the high temperature zone enhances N0X reduction by favoring
higher chemical reaction rates; however, reburn fuel should not be injected before the bulk of the
primary fuel has burned to completion. If injected too early in a coal fired boiler, natural gas, as
the reburn fuel, would preferentially bum before the coal char particles have burned to
completion. This could increase the possibility for unburned carbon while, additionally, not
permitting all the char bound nitrogen to be released prior to the reburn zone. A stoichiometry in
the range of 0.90 to 0.95 has been found to represent a reasonable balance between achieving a
desirable stoichiometry from the standpoint of NOx reduction chemistry and a stoichiometry that
will not exacerbate ash deposition and/or boiler tube wastage. Though the reaction kinetics for
NOx reduction in the reburn zone are quite fast, requiring on the order of 0.1 second, the
remainder of the residence time in the reburn zone is required to achieve good mixing of the
reburn fuel with the bulk flue gas. The naturally occurring small amount of oxygen in the flue
gas entering the reburn zone was found to be sufficient to promote the desired formation of OH
and H radicals. Effective and rapid mixing of reburn fuel ensures that all NOx entering the
reburn zone will contact the intermediate nitrogen-containing species so that maximum NOx
reduction is possible. Effective mixing must be achieved in such a way that there is no direct
fuel impingement on boiler walls. This impingement could exacerbate tube wastage or iron-
related ash deposition by creating low local stoichiometrics. Effective mixing would eliminate
extremes between highly oxidizing and highly reducing atmospheres which could cause
corrosion. Other practical considerations involve minimizing the number of boiler penetrations
and the avoidance of unnecessarily costly boiler modifications relative to the number and
placement of reburn fuel injectors. The number and placement of reburn fuel injectors must not
create thermal or structural boiler problems. The reburn fuel injection system should have
sufficient flexibility to permit on-line adjustment to maintain optimum mixing as a function of
boiler operational variables, such as load changes, that could alter gas flow patterns within the
7-10
-------
Return System Design
Test No. 412 J
M
= Straight in Injection
1
B
£
e
I*
U-
Elevation View
U,
B
Plan View
Figure 7-7
Recommended Burnout Air Injection Configuration
7-11
-------
Reburn System Design
reburn zone. Since the amount of reburn fuel required will likely change as a function of boiler
load, a control system should be provided which will provide automatic load following
capability. The reburn fuel control system should also have permissives which must be satisfied
to ensure safe operation.
The key design criteria for the reburn zone are summarized as follows;
• Inject reburn fuel into as high a temperature zone as possible, commensurate with releasing
all fuel-bound nitrogen upstream of the rebum zone
• Maintain average stoichiometry between 0.90 and 0.95
• Permit a small amount of Oz to promote formation of OH and H radicals
• Maintain a residence time between 0.5 and 0,7 seconds
• Maximize entrainment, mixing, and dispersion of reburn fuel
• Avoid direct fuel impingement on boiler walls
• Minimize the number of required boiler penetrations
• Locate fuel injection nozzles to minimize boiler/structural steel modifications
• Provide for maximum flexibility of reburn fuel jet direction and flow rates
• Provide a fuel flow rate control system with automatic load following capability
• Provide safeguards for fail-safe operation
For the burnout zone, unlike the rebum zone, air should be injected in as low a temperature gas
as possible to prevent the reformation of NOx. However, lower temperatures could prevent
complete burnout of combustibles leaving the reburn zone, so a balance must be struck between
the dual objectives of minimizing NOx reformation and complete combustible burnout. Rapid
and thorough mixing in the burnout zone is necessary. Although the reaction between fuel and
oxygen is quite rapid, the remainder of the recommended 0.6 - 0.8 second residence time is
needed to achieve effective mixing rather than for combustion reaction time per se. Direct
impingement of air on furnace walls should be avoided, more for reasons of preventing local
temperature increases than for any concern about the presence of an oxidizing atmosphere. The
amount of air should be just sufficient to achieve desired fuel burnout; an overabundance of
excess air contributes to dry gas losses and the potential for NOx reformation in the burnout
zone.
7-12
-------
Reburn System Design
Key design criteria for the burnout zone were;
• Inject burnout air in as low a temperature zone as possible, commensurate with obtaining fuel
burnout before entering the first convective surface.
• Provide for rapid mixing of air to minimize pockets of unburned fuel.
• Avoid direct air impingement on furnace walls.
• Minimize final excess oxygen, commensurate with obtaining good fuel burnout.
• Provide for a residence time in the range of 0.6 to 0.8 second.
• Minimize the number of required boiler penetrations, commensurate with obtaining good
mixing.
• Locate burnout air injectors to minimize boiler structural modifications while providing good
mixing.
• Provide for maximum flexibility of air jet direction and flow.
• Provide an air flow rate control system with automatic load following capability.
• Provide safeguards for fail-safe operation.
Reburn System Components and Installation
Five separate windboxes were installed at the rear wall of the unit for the reburn fuel nozzles.
Modified waterwall panels were installed at these locations for installation of the reburn fuel
nozzles. To supply recirculated furnace gas used as carrier for the natural gas in the original
system, ductwork from the gas recirculation fan and a windbox header were also installed. The
reburn fuel equipment was located at Elevation 882'. From inside the furnace, facing the rear
wall, the windboxes were designated 1A through IE from left to right. The three center
windboxes (IB, 1C, ID) face directly into the furnace and are square to the rear wall. The
centerline of the left and right windboxes (1A and IE) are yawed 20 degrees away from the
respective side wall (toward the center of the furnace) to avoid jet wall attachment.. Each
windbox was divided into three horizontal compartments. The upper and lower compartments
were 10" high and the middle ones were 8" high. All of the compartments were 16" wide. The
reburn injector installation is shown in Figure 7-8.
Four separate windboxes (two on each side) were installed at the side walls of the unit for the
burnout air nozzles. Modified waterwall panels were installed at these four locations. To supply
hot combustion air for the windboxes, connecting ductwork from the secondary air ducts was
installed. The burnout air equipment consisted of four tilting windbox assemblies located in the
left and right side walls of the furnace at elevation 905' 8". There were two windboxes in each
side wall. The nozzle tip arrangement was the same in all four windboxes. Each windbox was
divided into two horizontal compartments. The upper compartments were 11 1/2" high and the
lower ones were 10" high. All compartments were IE" wide.
7-13
-------
Rehurn System Design
Boiler Thermal Performance
Since operation of the reburn system required reduction of the main fuel by up to 20% and an
equivalent injection of reburn fuel into the lower portion of the secondary furnace, changes in the
boiler gas and steam side thermal performance were expected. A series of proprietary ABB/C-E
mathematical models, in conjunction with baseline data, furnace dimensions, and operating data
supplied by Ohio Edison, were used to verify that satisfactory thermal performance of the unit
would be achieved when operating the reburn system and that no adverse effects on the boiler
would occur. The calculations for boiler performance with reburn were for the original reburn
system design in which the natural gas was injected in a mixture which included recirculated flue
gas flowing at a rate equal to 10% of the exit gas flow rate. Specific items investigated in the
performance study were the following:
• Furnace heat absorption profile
• Convection pass performance
• Boiler efficiency
• Boiler circulation; departure from nucleate boiling.
Figure 7-8
Reburn Fuel Injection Windboxes
7-14
-------
Reburn System Design
The general approach used to evaluate boiler thermal performance was:
• Calculate or obtain physical data for the boiler components (e.g., heating surfaces, tube
diameters, tube arrangement, tube material, free gas areas).
• Set up computer programs to calculate boiler efficiency, cyclone/furnace performance,
convection pass performance, and air heater performance.
• Calibrate programs with baseline data. Determine required calibration factors to match
baseline data.
• Calculate baseline boiler performance.
• Calculate boiler performance with reburn.
• Compare boiler performance with reburn to baseline performance.
Baseline Boiler Performance
The first step in calculating the impact of the reburn system on boiler performance was to
establish the baseline performance for reference purposes. Baseline boiler performance was
calculated using the heat loss method. The calculated losses and resultant efficiency are shown
in Table 7-1. Knowing the boiler efficiency and the output of the unit, the energy input of the
coal was calculated. Based on the coal analysis shown in Table 7-2, combustion calculations
were performed to establish the gas and air weights. That data provided the necessary inputs for
the convective pass program. The convection pass program was run backwards to determine:
(1) furnace exit gas temperature, (2) surface effectiveness factors, and (3) intermediate steam and
gas temperatures.
Furnace/cyclone performance calculations were performed next using C-E's lower furnace
program. Program inputs were varied until conditions were met relative to cyclone combustion
efficiency, gas temperatures measured in the unit, and the furnace outlet temperature back-
calculated by the convection pass program.
A heat absorption baseline profile was then generated using C-E's lower furnace program. This
is shown by the solid line in Figure 7-9. Conditions for this calculation were 108 MW and 12%
excess air. The heat absorption rates shown are perimeter average rates. Where heat transfer
surfaces are more or less uniformly covered with refractory or ash deposits, the local rates should
be reasonably close to the average rates. Where tube sections are not covered with refractory or
ash deposits, local rates could be much higher than the average rates. The calculated average
rate for the cyclone is approximately 83,000 Btu/hr-ft2, and for the primary furnace and screen
tubes the rates are approximately 54,000 and 44,000 Btu/hr-ft2, respectively.
The total lower furnace heat absorption can be calculated by multiplying the heat absorption
rates from the profile by the EPRS (Effective Projected Radiant Surface) and by correcting for
casing heat loss. If the heat absorbed by the evaporative surface in the convection pass is added
in, the sum should equal the heat absorbed by the fluid from the boiler inlet to the steam drum
outlet; this was checked and was found to be in agreement within 2%.
7-15
-------
Reburn System Design
Table 7-1
Calculated Boiler Efficiency -108 MW
Dry Gas Loss
Moisture from Fuel Loss
Moisture from Air Loss
Radiation Loss
Ash Pit Loss
Miscellaneous
Total Losses
Boiler Efficiency
Stack Temp °F
Baseline Reburn
Coal 80% Coal/20% N.G.
2.84 2.63
4.47 5.32
0.07 0.06
0.24 0.25
0.74 0.62
0.50 0.50
8.86 9.38
91.14 90.62
267 269
Table 7-2
Coal Analyses - % by Weight
Ultimate
Moisture
Hydrogen
Carbon
Sulfur
Nitrogen
Oxygen
Ash
7.45
4.48
63.00
3.26
1.12
7.33
13.36
Total
100.00
Proximate
Moisture 7.45
Volatile Matter 35.05
Fixed Carbon 44.14
Ash 13.36
Total
100.00
HHV (Btu/lb)
11559
746
-------
"->4
>-rj |rj
£ ere'
CO ^
5 »
co -j
as
o
oo
100,000
90,000
80,000
70,000
60,000
tr
| 50,000
<
S
40,000
30,000
20,000
10,000
Cyclones
BASELINE
REBURN
K
0
-------
Return System Design
Boiler Performance with Natural Gas Reburn and Recirculated Flue Gas
Boiler performance with gas reburn was calculated in the normal forward design mode; i.e., the
programs were run in the following order: (1) cyclone/lower furnace, (2) convection pass, (3) air
heater, and (4) boiler efficiency. Process flow quantities were determined from assumptions
regarding reburn natural gas flow rates, recirculated flue gas flow rates, burnout air flow rates,
and the locations of the reburn fuel and burnout air injectors.
For predicting performance with reburn, the lower furnace program was run with the firing rate
in the cyclones reduced by 20%. Excess air was maintained at the same level as that of the base
case, 12%. Boundary surface conditions (waterwall deposits) were varied in the secondary
furnace: (1) in one case they were kept at the same condition as the back-calculated value for the
base case, and (2) in the other case it was assumed that there would be about 30% less thermal
resistance because of the decreased amount of coal being fired, the expected lower gas
temperatures, and changes in ash deposit characteristics.
The calculated heat absorption profile for the reburn case is shown with a dotted line in Figure
7-9. The profile indicates a 10% reduction in overall waterwall heat absorption with reburn for
the assumed case where the thermal resistance of ash deposits remained the same as for the base
case. For the assumed case where the thermal resistance dropped by 30% in the secondary
furnace, the overall waterwall heat absorption would be about 5% less with reburn than for the
base case.
Utilizing the output from the lower furnace program, the convection pass program was then run
to calculate superheater and reheater performance. The effective heating surfaces calculated
from the base line data was input to the program. A weighted fuel analysis (80% coal + 20%
natural gas) was used to calculate changes in gas properties. In the initial boiler performance
calculations which included recirculated flue gas and a resulting increased gas flow rate, more
heat was picked up in the convection pass with reburn than for the base case. Slightly more heat
is also picked up by the air heater with reburn.
One consequence of picking up more heat in the convection pass was that increased superheater
spray water flow was required. However, the calculated increase in superheater spray was within
the capability of the unit even under a worst case scenario. As shown earlier in Table 7-1 the
calculated boiler efficiency with natural gas reburn was about 0.5% less than the base case
primarily due to greater moisture from fuel losses; i.e., the higher hydrogen content of the natural
gas resulting in more water vapor being formed than when firing coal.
Two other boiler thermal performance related questions were addressed, namely the effect of
reburn on boiler circulation and the effect of reburn on departure from nucleate boiling (DNB).
DNB is defined as the occurrence of film boiling under which the tube inside (water side) heat
transfer coefficient drastically deteriorates and tube overheating/failure can occur.
A computer program was used to perform boiler circulation calculations. The program balances
the pressure drops of the multiple parallel circuits based on available thermal heads between the
downcomers and risers. Both baseline and reburn cases were investigated. The results of this
7-18
-------
Reburn System Design
study showed that tube offsetting, for purposes of making openings for fuel and air injectors,
would have an insignificant effect on circuit flow and exit circulation ratio.
Relative to the question of DNB, the division wall between the primary and secondary furnaces
was evaluated since this surface has the higher heat transfer duty. The criterion for evaluation of
DNB was specification of the maximum allowable steam quality, which depends on pressure,
heat flux, and mass flow of water/steam. To avoid DNB the actual circuit steam quality must be
kept less than the maximum allowable steam quality with adequate safety margin. Calculations
determined that 56% steam quality (or less) ensures a DNB free condition. The actual steam
quality in the highest duty location, the division wall, is calculated to be well under 10% with
reburn. The occurrence of DNB was therefore not seen as a problem.
Control System
The reburn control system used an Allen Bradley programmable controller to operate the reburn
system in an automatic, load-following mode. Natural gas flow, at a predetermined percentage
of unit heat input, and recirculated flue gas flow were based on coal flow demand input. The
burnout air flow was based on natural gas flow with the final excess oxygen designed to be
slightly lower than the normal cyclone excess oxygen level.
The reburn system was tied into the main boiler control system for safety and control purposes.
The natural gas reburn fuel controls were set up in a last-in-service/first-out-of-service logic.
The FGR system remained in service independent of the reburn natural gas, except for loss of
control power. All system dampers/valves fail shut except for the natural gas vent valves which
fail open.
7-19
-------
8
TEST PLANNING / MEASUREMENTS
Data for reburn system performance and boiler performance was measured during parametric
tests, maxi tests, and long-term load dispatch testing. Long-term corrosion monitoring was
conducted concurrently with the reburn system performance tests. Although the focus of the
project was NOx reduction, it was important to also measure boiler performance and component
life to assure that these requirements were not compromised. The scope of the tests is described
in the following subsections. Also described are the emissions and boiler performance
measurement equipment, the method used for data analysis, and the quality assurance/quality
control procedures.
Program Scope
Parametric Testing
A comprehensive set of parametric tests was performed on the unit under baseline and reburn
operation for the original reburn system during October 1990 through June 1991. Another set of
parametric tests was performed during October and in late 1991 on the modified reburn system.
The objective of these tests was to characterize the reburn system and document the effects of
varying operating parameters and equipment settings on NO* and CO emissions as well as boiler
performance and the steam temperatures. To expedite data collection, several parametric tests
were performed on a test day, and data collection was limited to flue gas composition and boiler
operating data. For selected tests, carbon in ash was also measured.
Maxi Testing
During the parametric testing of the original reburn system, four comprehensive tests (referred to
as "maxi" tests) were conducted. Maxi tests were run at generator loads of 108 and 86 net
megawatts for both baseline (100% coal firing) and 18% natural gas reburn conditions. An
additional maxi test at full load was conducted at reburn conditions for the modified reburn
system. The reburn configuration found to represent an optimum during the parametric
investigations was utilized during the maxi reburn tests.
The purposes of the maxi tests were to:
• provide fiill sets of test data for boiler performance calculations.
• assess the effect of reburning on the flue gas conditions entering the electrostatic precipitator
(ESP).
8-1
-------
Test Planning/Measurements
• measure the size distribution and mass loading of particulates entering the ESP.
• evaluate the effect of reburn on the collection efficiency of the ESP.
Long-Term Dispatch Testing
Long-Term Dispatch Testing was conducted between March 2, 1992 and June 19, 1992 after
optimum boiler and reburn system parameters for long-term testing had been identified. Gaseous
emissions data and boiler operating data were logged on a continuous basis at five-minute
intervals. Data logging was halted only for periods of equipment servicing, inspection, and
QA/QC procedures, and during periods when the reburn system or the boiler were off-line.
Corrosion Monitoring
Corrosion monitoring was performed to evaluate the effect of reburn operation on tube life, if
any. Corrosion measurement methods and monitoring results are described in Section 14,
"Boiler Tube Thickness Monitoring Program."
Flue Gas Sampling and Analysis
Sampling During Parametric Tests and Maxi Tests
Flue gas sampling and analysis were performed in general accordance with the EPA Methods.
Flue gas was sampled point-by-point using a ten-point sampling matrix located at the air heater
inlet as shown in Figure 8-1. Flue gas was extracted using stainless-steel probes which had
sintered metal filters. Impact shields were installed at the probe inlets to keep them from
plugging. The flue gas then passed through a second particulate filter and then to a solenoid
valve box. Individual probes were selected by switching on the solenoid valve for that particular
probe. The sample was drawn down to an instrument and gas analysis trailer at a flow rate of
20 - 25 cfm. This high rate of sample delivery minimized the residence time of wet, dirty flue
gas in the sampling system so that the removal and/or destruction of NOx in the sampling system
was minimized. The gas analysis train is shown in Figure 8-2. The sample was chilled to dry
the flue gas. Part of the sample was re-filtered and fed to the instruments to be analyzed for
NOx, O2, CO, CO2 and SO2; the remainder was discharged. Specifications of the instruments are
given in Table 8-1. Initially, a separate sample was drawn from the ESP breeching and analyzed
for THC (total hydrocarbons) but as its concentration was found to be negligible, THC analysis
was discontinued.
Sampling During Long-Term Dispatch Tests
Flue gas was extracted at the precipitator inlet breeching using an averaging probe made up of
three individual sampling probes of different lengths that were manifolded together. Probe
sampling lengths were chosen according to the equal area procedure described in EPA Method 1.
The probes were constructed in a similar manner as those used at the air heater inlet. After
leaving the averaging probe, the sample was filtered and drawn down to the instrument trailer.
8-2
-------
Test Planning/Measurements
Sample tubing exposed to the environment was insulated and heat-traced to keep the moisture in
the flue gas from freezing and blocking the sampling line. The flue gas sample was conditioned
and analyzed in a manner similar to that of the parametric testing sample.
8-3
-------
Test Planning/Measurements
Figure 8-1
Boiler Exit Gaseous Emissions Sample Matrix
8-4
-------
Test Planning/Measurements
Figure 8-2
Schematic of Flue Gas Sampling and Analysis System
8-5
-------
Test Planning/Measurements
Table 8-1
Specifications of Gas Analysis Instruments
Gas Specie
Made by:
Principle
Range
EPA Reference
Method
NO
TECO
chemiluminescence
of NO oxidized to
N02 by ozone
0 - 1000 ppm
7E
02
Thermox
fuel cell - difference
in potential between
flue gas and ambient
air
0 - 25%
3A
CO
Horiba
non-dispersive
infrared
0-1000 ppm
10
C02
Horiba
non-dispersive
infrared
0 - 20%
3A
S02
Western
Research
UV absorption
0 - 4000 ppm
6C
THC
Beckman
flame ionization
25A
Boiler Performance and Operations Data
A set of 62 boiler and reburn system operation parameters was logged on an IBM-PC compatible
computer directly from the plant's Bailey System 90 control system. During parametric and
maxi tests, additional data were recorded from control room instrumentation. The logged data
included:
• individual cyclone air and coal flows
• total gas flow
• primary and secondary cyclone air flow and the reburn system burnout air flow
• superheat and reheat steam temperatures and pressures
• unit load
• superheat and reheat attemperator spray flow.
A complete list of parameters logged is included in Appendix A.
8-6
-------
Test Planning/Measurements
Coal Composition
Coal samples were collected twice weekly as part of Ohio Edison's fuel specification check.
These samples were sent to ABB-CE for proximate and ultimate analyses. Coal samples were
also taken during the maxi tests from each of the active feeders.
ESP Performance Data
Fly Ash Loadings
To compute ESP inlet loading and ESP efficiency, fly ash samples were collected at the ESP
inlet breeching and at the stack. Both samples were collected in general accordance with EPA
Method 5, in which the flue gas sample is extracted isokinetically and the particulate matter is
collected on a filter placed in a box heated to 250° F.
Fly Ash Particle Size Distribution
Particle size distribution samples were collected from the stack using an Andersen Mark III
cascade impactor, with an Andersen pre-separator with a cut size of 10 microns used upstream of
the impactor.
Fly Ash Resistivity
In-situ fly ash resistivity was measured at the ESP inlet using a Wahleo resistivity probe. S03
concentration in the flue gas was computed by measuring the acid dew point using a probe
manufactured by Land Corporation.
Carbon in Ash
Fly ash samples were taken at the ESP inlet using a high volume sampler. Boiler bottom ash was
sampled from the slag tanks below the wet bottom slag taps. Both of these samples were
analyzed for carbon in ash to document imburnt fuel losses.
Flue Gas Temperature and Flow Field
Gas temperatures and velocities were measured periodically at three locations on the rear wall at
an elevation 2 ft. 6 in. below the reburn fuel injection elevation. Velocities were measured using
a five-hole pitot probe to obtain all three components of velocity. Gas temperature was also
measured at the reburn zone outlet immediately below the superheater.
8-7
-------
Test Planning/Measurements
Reburn Zone Inlet Conditions
To better understand the chemical and physical processes that take place in the reburn zone, a
series of measurements were made upstream and downstream of the reburn zone. Measurements
within the reburn zone could not be made because of physical access difficulties.
Cyclone Exit 02
Flue gas was extracted from the cyclone exit, actually downstream of the screen tubes and
immediately upstream of the reburn fuel injection location, using four water-cooled probes
mounted along the furnace rear wall. Each probe was connected to the plant's 02 instrument, and
its output was transmitted to the control system. During parametric and maxi testing, the 02
levels were also measured directly from the probes using a portable Teledyne electrochemical
cell 02 meter.
Cyclone Exit 02 for Long-Term Testing
Because the cyclone exit 02 probes were located in a very high temperature area, they plugged
frequently. Once the probes plugged, the reported cyclone exit 02 was unreliable because of high
air in-leakage into the 02 instrument. This made the plant-reported cyclone exit 02 value logged
off the plant's control system unsuitable for use in data analysis. A calculated value based on the
plant-reported cyclone air and fuel flows was used for cyclone exit 02 during long-term testing.
The equation is as follows:
Cyclone exit 02 (mole %, dry) = 20.9 x ^1-11.52 x
where: Mcoal = mass flow rate of coal
Mair = mass flow rate of air to the cyclones
The constant in the equation, 11.52, was determined from measurements of coal, air, and cyclone
exit 02 measured when the cyclone exit probes were not plugged.
Data Analysis
Reburn Zone Stoichiometry
One of the key variables affecting NOx emissions is reburn zone stoichiometry (RZS) defined as
the stoichiometry of the flue gas after the reburn fuel is injected but before the burn out air is
added to the flue gas. The equation relating RZS to cyclone exit 02 and measured gas and coal
feed rates, derived in the following paragraphs, depends on the chemical properties of coal and
gas and stoichiometric ratios for coal and gas.
-------
Test Planning/Measurements
RZS is related to mass flow rates of coal, air and gas, and stoichiometric ratios for coal and gas
as follows:
RZS = ^
^coal x ^coal ^gas x ^gas
where: Mair = mass flow rate of air
Mcoai = mass flow rate of coal
Mgas = mass flow rate of natural gas
Zcoal = air/coal mass flow ratio for complete combustion of coal
Zg2S = air/coal mass flow ratio for complete combustion of natural gas
Dividing the numerator and denominator in the above equation by the mass of air required for
stoichiometric coal combustion (Zcoa) x Mcoal), expressing the mass flows of coal and natural gas
in terms of their heat inputs, and simplifying yields:
RZS = — -:zs
Zgas R
1 + X-
Zcoal HI IV.
gas
HHVcoal
x(l-R)
where: CZS = Cyclone zone exit stoichiometry
R - natural gas energy input/(natural gas energy input + coal energy input)
HHVgas = high heating value of natural gas
HHVcoal = high heating value of coal
The cyclone zone exit stoichiometry is related to the cyclone exit 02 and additional
stoichiometric parameters as follows:
CZS = l-
Y * Mdfg
Msx(l~f)
where: y = normalized cyclone exit 02 = cyclone exit 02 (mole %, dry) /20.9
Mdfg = moles of dry flue gas product per pound of coal combusted
Ms = moles of air required per pound of coal for stoichiometric combustion
Values of some of the relevant ratios, based on average fuel compositions are as follows:
Ms
; 0,97
HHV„as
8a--~ 1.9 to 2.1
HHVcoal
2.0
2Coal
8-9
-------
Test Planning/Measurements
Approximating M
-------
Test Planning/Measurements
Instrument Drift
The instrument dri ft of the flue gas analyzers was checked to ensure that it was within allowable
limits (3% of span). If the drift was greater than 1.5% of span the instrument was re-calibrated.
Since measured drifts were always well within the allowable limits, it was believed that
inspections performed twice weekly were adequate to ensure that test equipment was within
specification.
Gas Sampling System Bias
The integrity of the gas sampling system was checked for system bias. Calibration gases were
introduced at the averaging sampling probe manifold exit and the instrument response from this
sample was compared to the response by directly injecting the calibration gases into the
instruments. The sampling system bias was well within the 5% allowed.
Sampling Location Bias
Parametric test results were based on an arithmetic mean of emission measurements sampled at a
ten-point grid located upstream of the air heater. Long-term results were based on emission
measurements made from a three-point composite sample drawn from the ESP inlet breeching.
To ensure that there were no significant biases between these two sets of emission data, flue gas
composition was simultaneously measured and compared at these two sampling locations under
identical steady-state boiler operation. This check was made several times during the long-term
testing. The NO emissions were within 5% of each other.
Comparison of NO and NOx Emissions
During the initial stages of the test program, the fraction of the total NOx that was in the form of
NO2 was documented by measuring the difference between the NO and NOx emissions from the
boiler. The NO2 fraction was found to be less than 5% of the total NOx. Therefore, in
accordance with EPA Method 7E, para. 5.1.2, only NO emissions were measured for the
remainder of the test program. In this report, the terms NO and NOx have been used
synonymously to refer to NO emissions from the boiler.
QA/QC for Boiler Operation Data
Boiler operation data was checked in a variety of ways. During parametric and maxi tests, boiler
operating data was simultaneously recorded manually from the instrumentation in the control
room to compare with the logged data. During long-term testing, coal mass flow based on feeder
instrumentation output was checked against bunker loading over a two-week period to calculate
a correction factor for the coal flow. Further, to check that the feeder calibration had not
excessively drifted during the test period, the average net plant heat rate was calculated for each
hour of operation. A sudden change in the plant heat rate would have indicated a change in
feeder calibration.
8-11
-------
Test Planning/Measurements
QA/QC Procedures for ESP Performance Measurements
Sampling procedures, selection of sampling location, sampling equipment, and calibrations were
performed according to the relevant EPA Methods, Reagents used for sample recovery were
reagent-grade.
8-12
-------
9
REBURN SYSTEM INSTALLATION AND STARTUP
Installation
The rebum system was installed with minimal disruption to normal power plant operation. The
four key phases of rebum system installation were: (1) procurement of material and delivery to
the site, (2) pre-outage activities, (3) outage activities, and (4) post-outage activities. A key
consideration was the installation of all direct boiler-related equipment/materials during the
utility's normal four week boiler outage.
Major items obtained during the procurement period included the reburn fuel and burnout air
windboxes, their waterwall tube panel inserts, FGR fan, recirculated flue gas and burnout air
ductwork, and the control system. The FGR fan and the control system were the items requiring
the longest lead time, about 26 weeks.
Pre-outage work included removal of the old FGR fan and associated ductwork along with
asbestos abatement. The natural gas pipeline was installed up to the point where it connected to
the windboxes. Structural steel was reinforced in those areas where the recirculated flue gas
ductwork was installed and minor revamping of access stairs and platforms was performed to
accommodate installation of the new ductwork.
At the commencement of the boiler outage on May 21, 1990, boiler casing and refractory at the
locations for the reburn fuel and burnout air windboxes were removed exposing the straight
sections of waterwall tubes to be cut out. Waterwall sections removed to accommodate the
prefabricated reburn fuel and burnout air tube panels were about 3 feet by 15 feet. After welding
in the tube panels, the windboxes were welded to flanges provided as part of the tube panel
structure, and seal boxes were built around each windbox and tube panel to prevent any furnace
leakage. Windboxes were tied into the previously installed ductwork by the installation of
expansion joints which allowed for growth of the boiler versus the stationary ductwork. The
boiler was hydrostatically tested, followed by the installation of refractory in the seal boxes and
seal welding of all outer casing. Following an air pressure test to locate and seal weld any
remaining furnace casing leaks, the boiler was fired up (to allow for chemical cleaning and
curing the refractory) and returned to service on June 25,1990.
9-1
-------
Re.hurn System Installation and Startup
Startup
A key activity during the post-outage timeframe was checkout and start-up of the rcburn system,
the objective being to verify that all components worked as designed. During the outage all
mechanical and electrical subsystems were verified to be operational. During system start-up,
the various subsystem interactions and sequencing were verified. Minor changes to the control
system programming and adjusting of the time delays based on actual device responses were also
completed. The gas reburn system was designed to operate in either a reburn mode (natural gas
being injected) or a non-reburn mode (no natural gas being injected). In the non-rebum mode
some minimal amount of cooling FGR or air was needed to maintain the integrity of the rebum
fuel and bumout air nozzles; minimum requirements for cooling FGR or air were determined
during the post-outage timeframe.
Rebum system operation was initially simulated without the use of natural gas to verify
operation of the comprehensive control system safety related permissives. Natural gas was
injected in small quantities for the first time on August 29,1990. Full-load automatic operation
with 19% natural gas was achieved on September 21,1990.
9-2
-------
10
PARAMETRIC TESTING (ORIGINAL REBURN
SYSTEM)
Introduction
This section provides a more detailed description of test results presented by Borio et al. f 1991)
for the original rebum system, the system which employed flue gas recirculation (FGR) for
transport of the natural gas reburn fuel. Borio et al. (1991) and the following subsections show
that the original system met the program objectives for NOx reduction and boiler performance.
However, a thick ash deposit formed on the back wall of the furnace during operation of the
original reburn system. This ash buildup did not prevent completion of the testing of the original
system but was unacceptable for sustained long term operation. The ash buildup is discussed in
further detail below in the subsection entitled "Ash Slagging Condition". Action taken to
circumvent the ash buildup is discussed in Section 11.
The reburn system incorporated a high degree of operational flexibility for examination and
optimization of reburn and boiler operating variables. The primary objective of the parametric
testing program was to determine the operational mode which would result in low NOx (not
necessarily lowest NOx) while minimizing other potentially detrimental effects on boiler
performance. These other effects included:
1. Minimizing other gaseous combustible and particulate emissions.
2. Minimizing fuel and auxiliary power costs.
3. Minimizing degradations in boiler performance (e.g., decreases in boiler efficiency, use
of reheat attemperator spray, or excessive superheater or reheat steam or tube metal
temperatures).
A secondary objective of the parametric testing was to establish a reburn database which could
be used to evaluate the process for application to other boilers.
During the initial parametric testing, approximately 150 test points were completed to examine
13 existing boiler and rebum system operational variables. The operational variables examined
included:
10-1
-------
Parametric Testing (Original Reburn System)
Baseline Test Variables
• Cyclone Excess Air
• Number of Cyclones in Service
• Boiler Load
Rebum Test Variables
• Rebum Zone
Natural Gas Flow Rate
Flue Gas Recirculation Flow Rate/Compartment Bias
Reburn Fuel Injector Tilt/Yaw Angle
Rebum Fuel Injector Horizontal Flow Bias
• Bumout Zone
Air Flow Rate
Bumout Air Tilt/Yaw Angle
Bumout Air and Rebum Fuel Injector Tilt/Yaw Angle Combination
Because of the large number of independent test variables, it was not possible to examine every
permutation and combination. The parametric testing was set up and conducted to step-through
the variables in a decreasing priority sequence for each of the three key boiler zones (cyclones,
rebum zone, and burnout zone). Initially, nominal operating conditions were selected for each
variable; then, once a variable had been examined, it was reset to a "near optimum" condition for
subsequent tests. Near optimum conditions were selected based on the above testing strategy.
The duration of each parametric test was one hour. Several parametric tests were conducted each
test day. Typically the rebum system operated continuously during testing times of ten hours per
test day. Tests were conducted five days per week over a time period of nine weeks.
To ensure comparability of the results, many tests were repeated. This was necessary because,
even though significant effort was expended, it was difficult to replicate cyclone operating
conditions on a day-to-day basis. During the parametric testing, a limited number of more
comprehensive tests were completed. These were referred to as "maxi" tests. The duration of
the maxi tests was eight hours. These tests were run at generator loads of 108 and 86 MWe (net)
at baseline (100% coal firing) and 18% natural gas rebum conditions utilizing the rebum
configuration found to represent an optimum during the parametric investigations. Purposes of
the maxi tests were to:
• Determine the effect of rebum system operation on the furnace gas temperatures entering the
rebum zone and the convective pass.
• Assess the effect of reburning on the flue gas conditions entering the electrostatic precipitator
(ESP).
• Measure the size distribution and mass loading of the particulates entering the ESP.
10-2
-------
Parametric Testing (Original Return System)
• Evaluate the effect of reburn on the collection efficiency of the ESP.
* Carbon in fly ash and bottom ash.
Ohio Edison Niles Plant Coal Analyses
The Eastern bituminous coal fired at the Niles plant arrived by track from approximately 15
mines located in the Ohio, Pennsylvania, and West Virginia area. No one mine supplied more
than 10% of the total coal supply used. Initially there was some concern that coal variability at
the Niles plant could add uncertainty to the results and conclusions drawn from those results.
However, frequent samples and subsequent analysis of the coal showed the fuel composition to
be very consistent. Table 10-1 presents a composite coal analysis based on 21 samples obtained
during the first series of parametric tests. Statistical data showing the good consistency of the
analyses is also shown. Based upon the consistency of the coal compositions, it was concluded
that coal variability had a negligible effect on results during this test series.
Table 10-1
Ohio Edison Niles Unit No. 1 Coal Analyses (As-Received Baste)
Maximum
Minimum
Standard
Proximate Analysis
Average
Value
Value
Deviation
% Moisture (Total)
7.8
9.3
6.6
0.76
% Volatile Matter
32.2
33.7
31.1
0.62
% Fixed Carbon
(By Difference)
47.3
49.3
45.3
0.94
% Ash
12.6
13.6
11.4
0.62
HHV Btu/lb
11576
11870
11277
176
lb Ash/106Btu
10.9
12.0
9.6
0.63
Ultimate Analysis
% Moisture
7.8
9.3
6.6
0.76
% Hydrogen
4.4
4.5
4.3
0.06
% Carbon
63.4
65.3
61.8
1.11
% Sulfur
3.3
4.1
3.0
0.31
% Nitrogen
1.4
1.5
1.3
0.05
% Oxygen (By Difference)
7.1
8.4
5.5
0.73
% Ash
12.6
13.6
11.4
0.62
Total
100.0
Baseline NOx Emissions
NO, emissions for the subject cyclone-fired boiler at 108 MWe (net) averaged approximately
705 ppm (all NOx emissions reported have been corrected to a 3% excess O, basis). This
emissions level was representative of normal operation with a mean cyclone excess oxygen level
10-3
-------
Parametric Testing (Original Reburn System)
of 2.0-2.5% 02 (10.6-13.6% excess air). Slight variations in individual cyclone operation
resulted in day-to-day data scatter of approximately ±25 ppm.
Changing the cyclone excess oxygen level changed the NOx emissions slightly. For example, a
1% decrease in cyclone excess oxygen, from 3 to 2% 02 decreased NOx emissions by
approximately 15 ppm.
Reducing the cyclone-firing rate also reduced NOx emissions. At 86 MWe (net), a 20% decrease
in boiler load, NOx emissions under normal operating conditions were approximately 630 ppm, a
75 ppm or 10% decrease in emissions from normal full-load operation. At a reduced boiler load,
a similar trend of decreasing NOx for decreasing cyclone excess oxygen was seen. Baseline NOx
emissions results showing the effects of boiler load and 02 are shown in Figure 10-1.
£
13
E
Q.
Q.
eg
O
S?
o
5
800
700
600
500
400
300
200
too
& 86 MWe Net
D 108 MWe Net
1
0.9
0.8
0.7
3
0.6 m
E
0.5 I
o
«s
_l
0.3
0.2
CYCLONE 02,%
Figure 10-1
Baseline NO, versus Cyclone 02,108 MWe and 86 MWe Net
Carbon monoxide (CO) emissions in the baseline mode of operation were typically very low,
under 30 ppm. Baseline SOx emissions varied between 2400 and 2700 ppm due to slight
variations in coal sulfur content. Negligible THC gaseous emissions were observed during
baseline and reburn testing.
10-4
-------
Parametric Testing (Original Reburn System)
N0X Emissions as a Function of Key Variables
Reburn Zone Stoichiometry
Some test variables were found to have a pronounced effect on NOx emissions, and other
variables had little or no effect on NOx. Rebum zone stoichiometry was found to be the key
parameter affecting NOx emissions. The equations used to calculate reburn zone stoichiometry
are discussed in Section 8. Figure 10-2 shows the effect of reburn zone stoichiometry on NOx
emissions. The reburn zone stoichiometry was varied either by adjusting the rebum natural gas
flow rate or the cyclone excess air level. For the full-load tests the reburn zone stoichiometry
was varied from 0.88 to 1.06.
NOx emissions are seen to be linearly related to reburn zone stoichiometry (for the test range)
and decreased by approximately 180 ppm per 0.10 (or 10%) decrease in reburn zone
stoichiometry. For a constant cyclone excess oxygen level an approximate 10% decrease in
reburn zone
t
¦o
£
a
a.
OJ
O
g
S
x
O
BOO
700
600
500
400
300
ZOO
100
0.80
0 „—-—"O""
&
4/
+/
0
No Natural Gas
X ~
+ >
&
10% Natural Gas
a
14% Natural Gas
+
18% Natural Gas
1
0.8
0.8
0.7
0.6
0.S
0.4
0.3
0.2
3
tn
£
E
"S
o
z
ca
0.85 0.80 0.9S 1.00 1.0S 1.10 1.15
REBURN ZONE STOICHIOMETRY
1.20
Figure 10-2
NOx versus Rebum Zone Stoichiometry at Various Gas Flow Rates, 108 MWe, 10% FGR
stoichiometry resulted from a 9% increase in rebum natural gas fuel fraction. For example, with
the normal cyclone excess oxygen level of 2.5% O2 (13.6% excess air), increasing the reburn
natural gas fuel fraction from 9 to 18% resulted in a decrease to the rebum zone stoichiometry
from approximately 1.03 to 0.93 and a decrease in the NOx emissions from approximately 480 to
300 ppm (±25 ppm).
10-5
-------
Parametric Testing (Original Reburn System)
Rebum natural gas flow (Figure 10-3) presents the NOx emissions data versus the amount of
reburn natural gas fired. Two significant results shown are: (1) the linearity of the NOx
reduction with increasing natural gas flow for a given cyclone excess oxygen level; and (2) for a
given reburn zone stoichiometry (RZS), the NOx emissions results were similar regardless of
whether the stoichiometry was achieved by changing the reburn natural gas flow rate or by
changing the cyclone excess oxygen level.
Recirculated Flue Gas Flow
The purpose of flue gas recirculation (FGR) in the reburn system is to assist in the penetration of
the reburn fuel and promote mixing of the reburn fuel with the bulk furnace gases without
significantly increasing the oxygen content or stoichiometry in the rebum zone as would happen
if air were used instead of FGR. Pilot scale research, Farzan, et. al., (1989), has also shown a
small incremental NOx reduction with increasing levels of FGR. Figure 10-4 presents the results
of tests where the FGR flow rate was varied from approximately 3 to 11% of the total flue gas
flow with constant natural gas flow and reburn zone stoichiometry. Both baseline (no natural
gas) and 18% natural gas reburn test series are shown. FGR had no appreciable effect on NOx
emissions with or without rebuming.
E
a.
o.
cf
O
8
55
<5
800
700
600
500
400
300
200
100
Q 1,5-2.0 Cyclone 02
O 2.0-2.6 Cyclon* 02
A 2.5-3.0 Cyclone 02
2.5-3.0 % CYCLONE 02
1
0.9
0.8
0.7
0.6
O.S
0.4
0.3
0.2
E
E
"S
o
z
m
5 10 15
PERCENT NATURAL GAS,%
Figure 10-3
NOx versus Cyclone O2 at Various Natural Gas Flows and Reburn Zone Stoichiometrics,
108 MWe, 10% FGR
10-6
-------
Parametric Testing (Original Reburn System)
&
"O
a
a
gj
o
m
5
£
z
700-
600"
800"
400"
300-
200-
100"
A Baseline
~ 18% Natural*
10
—-r-
12
1
O.S
O.B
0.7
2
0.6 OQ
E
0.5 |
0.4
0.3
0.2
14
APPROXIMATE FLUE GAS RECIRCULATIONS
Figure 10-4
NOx versus Percent Flue Gas Recirculation at Constant Reburn Zone Stoichiometry
o
z
3
The lack of any effect of FGR on NOx during the baseline (non-rebuming) tests was likely due
to: (1) coal combustion being essentially completed (no further fuel nitrogen release); and
(2) changes in thermal NOx being not measurably affected because of the relatively low thermal
dilution created by introducing FGR (previously measured temperatures showed approximately
2300-2400°F for the reburn zone inlet).
For the reburn tests, varying FGR between 3% and 11%, also had no effect on NOx emissions.
This was likely due to the eventual good mixing that occurred regardless of the FGR flow rate.
When FGR was reduced, rapid mixing was likely reduced; but, because of the ample residence
time, thorough mixing eventually still occurred and the net result was no change in NOx
emissions. Earlier flow modeling, Borio et al. (1989), had shown that cyclone effluent gases
tend to hug the rear wall where the reburn jets were placed. The importance of FGR flow is
likely to be very unit specific; e.g., in a large open furnace, if access to the reburn zone is limited,
FGR may be required for reburn fuel penetration and thorough mixing.
After determining the sensitivity of NOx reduction to FGR flow rate it was decided to operate at
a reduced level (about 5%) with the FGR fan inlet dampers nearly closed for the remainder of the
parametric testing. This was advantageous since lower levels of FGR minimized changes in
boiler steam side performance (discussed later) and also decreased auxiliary power consumption.
Later in the program, the reburn system was modified to eliminate the use of FGR altogether.
10-7
-------
Parametric Testing (Original Reburn System)
Other Reburn System Variables
N0X emissions were not directly affected by other reburn system operating variables, including
rebum fuel injector tilt, yaw, or flow bias or by bumout air tilt, yaw, or flow bias. However,
these variables had a significant effect on CO emissions and the O2 profile at the air heater inlet.
These effects are discussed below.
Reduced Boiler Load Testing
Reduced load testing was conducted at 86 MWe for the following reasons:
• this represents the approximate operating load where the fourth cyclone would be placed into
or taken out of service depending on whether boiler load was being increased or decreased.
• 86 MWe baseline coal only tests would have nearly equivalent cyclone loading to the 108
MWe full load tests when lull reburn (18% gas heat input) was employed.
Figure 10-5 shows NOx emissions plotted against reburn zone stoichiometry for both 86 and
108 MWe, net. At reduced load the NOx values were lower for all conditions than at full load.
Reburn effectiveness was also lower. The decrease in NOx for a ten percent (10%) change in
rebum stoichiometry at 86 MWe was approximately 130 ppm compared to the 180 ppm at full
load. This decrease in reburn effectiveness is due to lower initial NOx values and lower gas
temperatures which led to slower reactions in the reburn zone.
¦n
I
a
04
O
a?
m
£
0.8
0,9 1.0
REBURN ZONE STOICHIOMETRY
108 MWe Net
O No Natural Gas
A 10% Natural Gas
+ 14% Natural Gas
~ 18% Natural Gas
86 MWe Net
• No Natural Gas
A S.4% Natural Gas
¦ ta,3% Natural Qai
Figure 10-5
NOx versus Reburn Zone Stoichiometry at Various Gas Flow Rates, 108 MWe and 86 MWe
10-8
-------
Parametric Testing (Original Reburn System)
Other Gaseous Emissions
Baseline emission of CO ranged between 25 and 50 ppm. During shakedown of the reburn
system, high levels of CO emissions were observed, especially during high reburn fuel flow
rates. The high CO measurements were attributed to insufficient penetration and mixing of the
burnout air and occasional maldistribution of air to the cyclones. The airflow distribution to the
cyclones was corrected by monitoring oxygen at the sampling ports on the rear wall at an
elevation 2'-6" below the reburn fuel nozzles and making required adjustments to the airflow to
the cyclones. CO was minimized by down-tilting both the reburn fuel nozzles and burnout air
nozzles and optimizing the yaw of the burnout air nozzles. A 17 depee downward tilt of the
reburn fuel nozzles and a 10 degree downward tilt of the burnout air nozzles was selected (Figure
10-6). The burnout air nozzles were set to impart a clockwise swirl (viewed from above). With
these adjustments CO emissions were decreased to typically below 100 ppm. (Please note that
the CO data are presented on a logarithmic scale.) In addition, more uniform CO and O2 profiles
were generated across the boiler exit duct as shown by comparing Figures 10-7 and 10-8,
Emission of SO2 decreased with increasing natural gas flow as expected. On average the SO2
decrease was inversely proportional to the reburn fuel flow; however, there was a significant
amount of scatter (±10%) due to coal sulfur variations. Gaseous THC emissions were negligible
for all tests.
10000-
1000 •
£
a.
Q.
o~
o
too:
10"
•30
—(—
-20
-10
o Reburn Fuel Injector Tilt = -IT degrees
o Reburn Fuel Injector Tilt - 0 degrees
a Reburn Fuel Injector Tiit = -25 degrees
_1_
O
10
20
30
BURNOUT AIR VERTICAL TILT, degrees from horizontal
Figure 10-6
CO versus Bumout Air Tilt at Several Rebum Fuel Injector Tilts, 108 MWe Net, 5% FOR,
17.5% Natural Gas, 2.5% Cyclone 02
10-9
-------
Parametric Testing (Original Rebum System)
10000
1000 -
E
o.
Q.
o
o
to
18.4% Natural Gas
2.5% Cyclone 02
•17 degree Rebum Fuel Injector Tilt
Q degree Burnout Air Tift
0 degree Burnout Air Yaw
10
9
-7
6
A
5
-------
Parametric Testing (Original Reburn System)
Carbon in Ash
Carbon loss in fly ash was not significantly affected by rebuming. Fly ash samples were taken
and analyzed for approximately two-thirds of the reburn tests. Bottom ash samples were taken
once per day. Carbon levels in the fly ash during full-load tests ranged from 25% to 45%, with
carbon levels between 30% to 35% being most typical. Attempts were made to relate fly ash
carbon level to reburn natural gas flow and cyclone excess air, variables which might be
expected to have correlations with fly ash carbon levels. No relationship was found. Carbon in
the bottom ash was typically less than 1% of the bottom ash, by weight. Thus for a coal with
12.6% ash and a baseline fly ash/bottom ash ratio of 30:70, the baseline carbon heat loss was
approximately 1.2 to 1.4%; and for rebuming with a reduced coal flow and hence fly ash loading,
the carbon heat loss was approximately 1.0 to 1.2%.
Reasons for the relatively high unburned carbon values under baseline and reburn conditions are
unclear. Possible causes include coal properties, coal particle size distribution and cyclone
aerodynamics (greater expulsion of coal fines). During reduced load operation the average fly
ash carbon content decreased to about 20%. This would be expected with more residence time,
decreased cyclone loading, and decreased expulsion of particulate from the cyclones.
Furnace Gas Temperatures
Reburn Zone Inlet Gas Temperatures
Figures 10-9 and 10-10 present results of flue gas temperature traverses made at the inlet to the
reburn zone. The furnace depth at the traverse locations was 13 feet. The maximum traverse
depth was physically limited to 10 feet. At 108 MWe (net) the baseline average gas temperature
was 120°F higher than with 18% reburn. The tests at 86 MWe (net) showed a similar trend: the
baseline gas temperature averaged approximately 100°F higher than with rebuming. For both the
baseline and reburn tests, there was a 200 to 300°F decrease in flue gas temperature from the rear
wall to the division wall. The temperature profiles for baseline and reburn at 86 MWe paralleled
one another. The baseline and reburn gas inlet temperatures at 108 MWe showed considerable
difference near the back wall but approached the same value near the maximum probe insertion
measurement depth. Comparison of the average temperatures and profiles measured during the
108 MWe reburn test with the 86 MWe baseline test show very similar results. This is
reasonable because the coal loading to the cyclones for reburn with 18% natural gas at 108 MWe
is only slightly higher than at 86 MWe with 100% coal.
Furnace Outlet Gas Temperatures
Figure 10-11 shows the results of the temperature traverses from the left (west) side wall at the
furnace outlet plane. The traverse depth represents approximately one third of the boiler width.
The furnace outlet temperature with reburn averaged 130°F higher at 108 MWe than the base
case; i.e., 100% coal. At 86 MWe the average temperature with 18% rebum was about 65°F
10-11
-------
Parametric Testing (Original Reburn System)
3000
it
03
«*-»
to
-------
Parametric Testing (Original Reburn System)
3000
£ 2500
ffl
3
c5
3k.
m
t
-------
Parametric Testing (Original Return System)
Ammonia was injected into the exhaust duct at a location about ten (10) duct diameters upstream
of the ESP inlet port to control acid smut emissions. However, in addition to affecting acid smut
emissions, ammonia alters fly ash resistivity and particulate size, which in turn can affect ESP
performance and particulate emissions, Lookman and Glickert (1992). Ammonia was normally
fed at a constant rate, optimized for Ml base-load operation on 100% coal. At stoichiometric
ratios less than unity, ammonia reacts with the SO3 and water vapor in the flue gas to form
ammonium bisulfate (NH4HSO4), which is a sticky substance that is believed to be deposited
onto fly ash particulates at the flue gas temperatures prevailing in the ESP. During rebum
system operation no attempt was made to optimize the ammonia injection system to account for
the decrease in the amount of SO3 entering the ESP. Consequently, the ammonia injection rate
was excessive, and the ammonia to SO3 stoichiometric ratio was above unity. The ammonia then
reacted with SO3 to form ammonium sulfate (NH4SO4), which is a crystalline powder at the ESP
temperatures. This substance has a high resistivity, making it difficult to collect, unlike
ammonium bisulfate that is weakly ionic and actually lowers the fly ash resistivity. Also, unlike
ammonium bisulfate, which is sticky and promotes agglomeration of the particulates into larger,
easier to collect particulates, ammonium sulfate is formed as a fine powder, that is itself very
hard to collect. Therefore, to duplicate the 100% coal firing particulate loading levels leaving
the stack it is necessary to optimize the ammonia injection rate of the flue gas conditioning
system.
Boiler Thermal Performance
Boiler operating data for four parametric tests of the original reburn system was analyzed to
evaluate the effects of reburn on boiler performance. The four tests selected are the following;
Test No.
56A
57B
58A
59C
Date
12/11/90
12/12/90
12/13/90
12/13/91
Load, %
100
100
80
80
Reburn Fuel %
0
17.2
0
18.5
Excess Air %
15.0
14.2
19.6
17.3
Gas Recirculation, %
1.31
4.49
2.08
5.99
Main Steam Temperature,°F
997
1000
1000
1000
Final Reheat Temperature, °F
988
1000
975
982
Reburn Stoichiometry
-
0.94
-
0.99
Two of the tests are with the reburn system shut off (test 56A and 58A) and two are with the
reburn system in operation (tests 57B and 59C). Two loads were selected, 100% and 80%. The
operating data for these four tests are shown in Table 10-2.
10-14
-------
Parametric Testing (Original Return System)
Table 10-2
Niles Unit No. 1 Operating Data
Test No.
56A
57B
58A
59C
Baseline
Reference Data
DATE
12/11/90
12/12/90
12/13/90
12/13/90
2/3/88
TIME
822
1811
828
1857
MATRIX PT
9
58
67
92
MW (GROSS)
114.5
114.5
92.4
91.7
115
MAIN STM. FLOW K Ibs/HR
854.2
843.7
672.2
659.8
840
SH DES. FLOW E lbs/HR
8560
14420
1100
900
SH DES. FLOW W lbs/HR
1700
23340
10110
11523
SH DES, FLOW TOT lbs/HR
10260
37760
11210
12430
39000
RH DES. FLOW lbs/HR
0
2320
0
0
0
RH FLOW (EST.) (.887 x MS + SPRAY)
757.7
750.7
596.2
585.2
FW PRESS PSIG
2051
2061
2234
2247
DRUM PRESS PSIG
1533
1530
1508
1506
MAIN STM. PRES. PSIG
1470
1470
1470
1470
1470
RH IN PSIG
336
336
265
261
335
RH OUT PSIG
308
308
241
237
308
STEAM/WATER TEMP. (°F FOR ALL
TEMPERATURES)
FW TEMP.
483.4
483.9
462
460.6
PRIM, SH OUT. E
748.7
764.9
735.9
730.2
PRIM. SH OUT. W
722.3
750.5
718.5
716.9
PRIM. SH OUT. AVG
735.5
757.7
727.2
723.6
750
SEC. SH IN E
706.6
683.1
680.2
673.8
SEC. SH IN W
728.4
722.3
735.4
731.2
SEC. SH IN AVG
717.5
702.7
707.8
702.5
691
MAIN STM. E
993.5
1000.1
1000.9
1000.1
MAIN STM. W
1000.2
999.9
999.6
999.9
MAIN STM AVG
996.9
1000
1000.3
1000
1000
COLD RH E
672
677
644
642
COLD RH W
681.5
685,3
659.8
658.1
COLD RH AVG
681.5
685.3
659.8
658.1
692
COLD RH AFTER DESUPERHEATER
669
668
641
640
692
HTRHOE
990
1002.7
977.7
984.7
HT RHO W
984
996.6
971.7
978.5
HT RHO AVG
987.5
999.7
974.7
981.6
990
(Continued)
10-15
-------
Parametric Testing (Original Reburn System)
Table 10-2 (Continued)
Niles Unit No. 1 Operating Data
Test No.
56A
57B
58A
59C
Baseline
Reference Data
DATE
12/11/90
12/12/90
12/13/90
12/13/90
2/3/88
GAS/AIR TEMP. (°F)
GAS LV, PRI. SH E
655
700
600
605
GAS LV. PRI. SH W
630
660
610
618
GAS LV. PRI. SH AVG
643
680
605
612
GAS LV. LT. RH
670
650
645
658
GAS AHIE
680
688
652
652
GAS AHI W
680
682
668
672
GAS AHI AVG
680
685
660
662
747
GAS AHO E
252
248
230
240
GAS AHO AVG
251
250
233
241
267
AIR AHI
120
118
105
112
130
AIR AHO E
585
590
555
562
AIR AHO AVG
575
581
554
562
637
PRI. SH. DAMP POS. E
2.12
84.6
-5.28
-5.28
PRI. SH. DAMP POS. W
0.75
98.8
-0.2
-.23
RH DAMPER POS.
98.7
-0.78
101.7
101.7
SPRAY WATER TEMP.
235
236
221
221
02 AHI %
2.8
2.7
3.5
3.2
FGR (REBURN) #/HR
13050
44850
17390
48930
GAS WEIGHT - NO FGR
999800
998100
836000
816200
10-16
-------
Parametric Testing (Original Reburn System)
With the rebum system off there was still a small amount of gas (1.31% and 2.08% in tests 56A
and 58A, respectively) recirculated through the rebum nozzles for cooling. In addition, a small
quantity of cooling air was supplied to the burnout air ports. With the reburn system on (tests
57B and 59C), 17.2 and 18.5 percent of the heat supplied to the furnace was from the reburn fuel,
respectively. FGR was 4.5% and 6% respectively. Approximate reburn zone stoichiometry was
0.94 and 0.99, respectively.
Using proprietary ABB codes in conjunction with the data in Table 10-2 the following items
were calculated:
• Attemperator spray water flows, reheat flow and component heat absorption
• Boiler efficiency and heat supplied to furnace
• Gas and air weights
• Furnace exit gas temperature and gas temperature profile through convection pass
• Secondary superheater surface effectiveness
The results of the thermal performance calculations are summarized in Table 10-3. At full load
(114.5MWe, gross) the main impact of rebuming on boiler thermal performance was a shift in
the heat absorption from the waterwalls to the convective sections. The Niles unit does not have
an economizer; therefore, the increase in convective pass absorption was observed in the
superheater and reheater. Superheat attemperator spray water flow increased from 1.4% to 3.9%
with reburn. A small amount of reheat attemperator spray water was measured (0.26%) during
the reburn test, primarily due to leakage past the control valve. Reheater outlet steam
temperature was 12°F below design during the baseline test; therefore, reheater performance
improved with reburn.
At full load, rebuming decreased waterwall heat absorption by approximately 5% and increased
the convective section heat absorption by approximately 5%. The decrease in waterwall
absorption was due to the decrease in cyclone loading. The increase in convective pass
absorption was due to increased gas temperatures (calculated to be 30°F at the furnace outlet
plane) and increased flue gas weight (due to FGR) with rebuming. Reheater absorption
increased by only 4% while superheater absorption increased by 6% due to an adjustment of
backpass flow control dampers.
Steam temperature profiles were also monitored during this program. Thermocouples were
installed on approximately every fourth tube element at the primary and secondary superheater
outlet headers. Negligible changes were observed in primary or secondary superheat profiles
between baseline and rebum tests.
Boiler thermal performance for the four tests is summarized in Table 10-3. The boiler
efficiency with natural gas rebuming decreased by 0.62%. The largest change was a 1% higher
loss due to a higher moisture in the flue gas in the reburn cases. The higher moisture in the flue
gas is due to the higher hydrogen content in the natural gas versus the hydrogen content in the
coal. This loss was somewhat offset by a lower ash pit loss and a lower carbon heat loss due to
less coal being fired when rebuming was employed.
10-17
-------
Parametric Testing (Original Reburn System)
Table 10-3
Summary Of Boiler Thermal Performance for the Original Reburn System
TEST NO,
56 A
57B
58A
59C
TYPE OF TEST
BASELINE
REBURN
BASELINE
REBURN
GROSS MW
114.5
114.5
92.4
91.7
HEAT FROM COAL %
100
82,8
100
81.5
HEAT FROM GAS %
0
17.2
0
18,5
EXCESS AIR %
15.0
14.2
19.6
17.3
GAS RECIRC, %
1.31
4.49
2.08
5.99
STEAM TEMP. SHO-°F
997
1000
10W
1000
STEAM TEMP, RHO-°F
988
1000
975
982
MAIN STM FLOW LBS/HR
854200
843700
672200
659800
REHEAT STM FLOW LBS/HR
757700
750700
596200
585200
SH SPRAY FLOW LBS/HR
11575
32696
9637
10671
RH SPRAY FLOW LBS/HR
0
1959
0
0
GAS RECIRC FLOW LBS/HR
13050
44850
17390
48930
GAS FLOW THRU CONV PASS LBS/HR
1012850
1042950
853390
865130
AIR FLOW THRU AIR HTR LBS/HR
917100
921200
769900
755300
COMPONENT HEAT ABSORPTIONS - MBTU/HR
PRIMARY SUPERHEATER
124.5
132.6
93.0
89.3
SECONDARY SUPERHEATER
160.7
170.7
132.5
132.8
REHEATER SUPERHEATER
123.2
127.9
98.6
99.4
WATERWALLS
590.3
563.0
478.4
469.7
TOTAL
998.7
994.2
802.5
791.2
HEAT LOSSES - %
DRY GAS LOSS
2.70
2.66
2.73
2.67
MOIST FROM FUEL LOSS
4.35
5.34
4.38
5.46
MOIST FROM AIR LOSS
0.06
0.06
0.06
0.06
RADIATION LOSS
0.24
0.24
0.29
0.30
ASH PIT LOSS
0.27
0.22
0.27
0.22
CARBON LOSS
1.39
1.15
1.39
1.14
TOTAL
9.28
9.90
9,39
10.06
BOILER EFFICIENCY
90.72
90.10
90.61
89.94
BTU FIRED MBTU/HR
1099.1
1107.7
887.0
881.8
LBS FUEL FIRED
94791
87697
76498
69357
SURFACE EFFECTIVENESS
SECONDARY SUPERHEATER
0.903
0.907
0.935
0.935
GAS TEMPERATURES - °F
SECONDARY FURN. OUTLET
2112
2139
2011
1974
REHEATER INLET
1618
1634
1529
1495
REAR CAV. OUTLET
1373
1395
1283
1246
PRIMARY SUPERHEATER INLET
1359
1381
1270
1234
AIR HEATER INLET
680
685
660
662
AIR HEATER OUTLET
251
250
233
241
10-18
-------
Parametric Testing (Original Return System)
Comparing data for Tests 58A and 59C shows that the change in thermal performance due to
reburn was less noticeable at 80% load. Waterwall heat absorption decreased by 1.8%. Part of
this can be attributed to the slightly lower load of Test 59C. Gas temperatures entering the
convection pass were lower with reburn, offsetting the effect of increased gas flow. Reheater
outlet steam temperature was higher with the reburn system in operation although it was below
design. Boiler efficiency was lower with the reburn system in operation as was the case at full
load. The boiler efficiency decreased by 0.67%, nearly the same as at foil load. Overall, the
boiler performance did not change appreciably with natural gas rebuming. The minimal changes
in boiler efficiency measured during the parametric testing confirm the predicted minimal
performance changes as discussed in Section 7.
Ash Slagging Condition
During the planned year-end outage in late 1990 after completion of the parametric testing of the
original reburn system, a heavy buildup of slag was found on the rear wall of the furnace at
elevations from below the reburn fuel nozzles to above the beginning of the sloping rear wall.
The cause of this buildup and resolution of the difficulties caused by the buildup are related to
the studded, refractory coated rear wall of the secondary furnace of Niles Unit No. 1. This
furnace design, shown in Figure 6-1, includes a primary furnace and a secondary, or main
furnace. Hot combustion gases exit from the cyclones, flow downward through the primary
furnace, through slag screen tubes, and upward through the secondary furnace. The primary
furnace and slag screen are refractory-lined to keep slag discharging from the cyclones in a
molten state, permitting the slag to discharge through the slag tap at the bottom of the furnace.
In the normal design and operation of screened cyclone furnaces, the tube walls of the secondary
furnace are not covered with refractory and there is no running slag above the slag screen. (See
Farzan et al. (1993) for a description of the typical screen tube furnace design). However, as
discussed in Section 6, Description of the Host Unit, Niles Unit No. 1 (and the sister Unit No. 2)
has studded, refractory coated waterwall tubes on the rear wall of the secondary furnace to
provide higher gas temperatures in the back pass of the boiler to maintain steam temperatures.
During normal operation at Niles a layer of slag builds up on the rear wall due to particles
passing through the screen and impacting on molten slag on the rear wall. An equilibrium slag
layer thickness of two to four inches is reached with the accumulation of particles impacting and
remaining on the wall balancing the flow of slag running down the rear wall. As indicated in
Section 6, satisfactory slag tapping and steam temperatures are achieved at Niles with this
arrangement over the normal operating range of the unit.
A photograph of the rear wall of the unit and one set of reburn nozzles after completion of the
parametric tests is shown in Figure 10-12. The condition of the wall and nozzles, with deposits
as much as 12 inches thick at some places near the nozzles, is a sharp contrast to the clean
condition of the wall and nozzles before parametric testing, shown in Figure 10-13 prior to
accumulating the normal two to four inch equilibrium slag layer buildup that occurs on the rear
wall during normal operation. After completion of parametric testing there was speculation that
the buildup may have been the result of the natural gas used during the reburn parametric tests.
The slag was removed manually and the unit was restarted for a time period without rebum.
During this time period flue gas was recirculated to the reburn fuel nozzles at a flow rate equal
10-19
-------
Parametric Testing (Original Return System)
to approximately 1% of the total flue gas flow rate in order to protect the nozzles from
overheating. This flow rate was approximately 20 to 25 % of the flow rate used during reburn
tests. After operation in this mode for two (2) weeks, the furnace was again taken out of service
and inspected. Slag deposits of about the same size and appearance as those seen after
completion of the initial series of reburn parametric tests were again present on the rear wall.
As stated in the Introduction, the Niles No. 1 reburn program was the first full-scale
demonstration of reburn technology in the U.S. The condition of the reburn nozzle and rear wall
of the furnace seen in Figure 10-12 was not anticipated by any of the small-scale, short-term
reburn investigations discussed in Section 5. The experience at Niles clearly shows the
importance of long-tern demonstration programs as a necessary part in the development of new
emissions control technologies. The deposits, which were as much as 12 inches thick, as
discussed above, had little or no effect on boiler performance and did not prevent completion of
the original system test program. However, long-term operation of the original reburn system
was unacceptable for several reasons. Slag falls during boiler operation could have a damaging
effect on screen tubes at the bottom of the furnace; the possibility of slag falls during slag
removal operation was a risk to personnel; and slag accumulation could cause blockage and
misdirection of the reburn fuel jets as well as shorten nozzle life due to overheating. For these
reasons there was a need to identify the cause of the problem and to resolve it. These subjects
and the redesign of the reburn system are discussed in the next section.
10-20
-------
Parametric Testing (Original Return System)
Figure 10-12
Reburn Nozzles and Rear Wall after Completion of Parametric Testing
10-21
-------
Parametric Testing (Original Return System)
Figure 10-13
Original Rebum System Nozzles prior to Parametric Testing
10-22
-------
11
DESIGN OF REBURN SYSTEM WITHOUT FGR
Analysis of the Slag Buildup Problem
The steps for resolution of the slag buildup problem during the Ohio Edison Reburn Project
involved summarization of information that had a bearing on the problem, development of a
hypothesis to explain the phenomenon, and resolution of the problem by evaluating the
hypothesis. These actions and the modified reburn system design, which evolved by resolution
of the problem, are discussed in this section.
The important information and observations are summarized as follows:
— Deposit removal from the secondary furnace back wall occurs due to molten ash run-
off through the screen tubes onto the furnace floor where it is tapped with cyclone
slag.
— The deposit on the back wall reaches a steady state thickness when the deposition rate
equals the molten slag runoff rate; normal thickness is 2 to 4 inches.
— Following furnace deslagging during the 1990 year-end outage, ash deposition
reached pre-outage condition (up to 12 inches thick) in about two weeks time with
only reburn nozzle cooling flue gas in operation.
— The sister unit (No. 2) burning the same coal at the same load and excess air had
normal ash deposit thickness.
— It is arguable whether the ash deposition was greater with the reburn system in
operation or just the FGR system at the flow rate used for cooling the reburn nozzles.
However, in either case the buildup was significantly greater than operation without
FGR.
~ Ash deposition in and around some of the reburn fuel nozzles affected nozzle life.
These observations led to a hypothesis that the heavier than normal ash deposition on the back
wall of the secondary furnace was caused by a combination of cooler recirculated flue gas
flowing along the back wall, entrained fly ash in the recirculated flue gas, and new studs which
were installed on the five new panels (each 3 feet wide by 15 feet high) installed for the reburn
fuel nozzles.
The mechanism for reentrainment and redeposition of molten slag droplets is shown
diagrammatically in Figure 11-1. The mechanism and slag buildup hypothesis are supported by
the following rationale:
-- A boundary layer of relatively cool recirculated flue gas flowing along the rear wall
caused the deposit temperature on the back wall to decrease.
11-1
-------
Design of Reburn System Without FGR
Figure 11-1
Slagging Mechanism at Niles Unit No. 1
11-2
-------
Design of Reburn System Without FGR
— The decreased deposit temperature forced deposits to grow thicker in order for the
surface to reach a sufficiently high temperature for the run-off rate of the slag to
equilibrate with the deposition rate.
— Increased deposit thickness decreased heat transfer in the lower part (vertical wall) of
the secondary furnace, thus increasing bulk gas temperature at the upper elevations
(sloped wall).
— Higher bulk gas temperatures coupled with furnace aerodynamics drove more molten,
entrained slag droplets into the sloped section of the back wall causing deposition,
whereas previously the droplets were frozen or crystallized before impacting the
sloping wall. As a result, the sloping wall had thick deposits where previously it only
had small islands of deposits one to two inches thick.
-- Fly ash in the recirculated flue gas experienced a different time/temperature history
than ash coming directly into the secondary furnace from the cyclones. It was
speculated that this might have been a contributing factor in altering the morphology
of the deposit from a thin molten deposit to a thicker sintered deposit.
Resolution of the Problem
Several approaches were proposed for resolving the problem. The most attractive approach was
to completely eliminate the use of FGR for injection of the reburn fuel. One concern with this
approach was possible loss in NO* removal efficiency due to poorer mixing. In order to address
this concern, brief proof of concept (POC) tests were conducted to evaluate the impact of
eliminating FGR on NO* removal. For the POC tests, natural gas injectors were installed
temporarily through the four gas sampling ports located on the rear wall at elevation 879'-7"
(which is 2'-6" below the center-line of the original reburn fuel nozzles). Tests were run with
natural gas flows equal to 6%, 9%, and 12.9% of the total energy input. Results of these tests,
given in Figure 11-2, showed little or no difference in NOx removal compared to reburn tests
using 10% FGR at comparable natural gas flow rates. Note that FGR flow rate in the POC tests
was less than 1%, the minimum required for cooling the original reburn registers.
Based on these results a decision was made to redesign the reburn system with five water-cooled
natural gas injectors installed at the same locations as the original reburn fuel injectors and to use
no FGR. The sampling probes were reinstalled at the existing locations. The design of the
modified reburn fuel injectors for Niles Unit No. 1 is shown in Figure 11-3. It should be
emphasized, however, that the natural gas injectors of the type used at Niles No. 1 may not be
optimum for all furnaces because other furnace configurations may require FGR flow to achieve
proper mixing of the reburn fuel.
In addition to effective NOx removal, the modified reburn system had several economic
advantages. Changes to pressure parts (water walls) and construction of injectors for the
modified system were less costly. The elimination of the gas recirculation fan, controls, and
ducts represented a reduction in capital, maintenance, and operating costs. The space and time
requirements for the reburn system were also reduced.
11-3
-------
Design of Reburn System Without FGR
2.5-3.0 %
/
Cyclone
32
/
i/
/
A
t
1.
/
5-2.0 % C
fclone Oi
&
o 1 5 2 0 % Cyclone 02; 10 % FGR
O 2.0-2 5 % Cyclone 02; 10 % FGR
a 2,5-3.0 % Cyclone 02; 10 % FGR
# POC NOx w/o Nat. Gas; <1 % FGR
¦ POC NOx w/6 % Nat. Gas; <1 % FGR
# POC NOx w/9 % Nat. Gas; <1 % FGR
2
* POC NOx w/12.9 % Nal. Gas; <1 % FGR
8 10 12
Natural Gas Flow, (%)
14
16
18
1
0.9
0.8
• 0.7
3
0,6 10
E
0.5 |
O
04 j?
m
0.3
0.2
0.1
0
20
Figure 11-2
NOx Emissions at 108 MWc for POC Tests Compared to Tests with 10% FGR
11-4
-------
Design of Rehurn System Without FGR
Figure 11-3
Natural Gas Injector for the Modified Reburn System
11-5
-------
12
PARAMETRIC TESTING (MODIFIED REBURN
SYSTEM)
Introduction
After installation of the modified reburn system, parametric tests were run to measure emissions
and thermal performance of the system, to establish reburn system operating characteristics, and
to identify optimized reburn system operating parameters for long-term load dispatch testing.
These parametric tests were conducted between October 29,1991 and November 20,1991. A
limited number of measurements of water feed to the reburn zone were obtained in January 1992
when two water-cooled guide-tubes inadvertently developed water leaks. Controlled parametric
tests of water injection to the reburn zone for enhancing system performance were performed in
July 1992. The results of parametric tests and conclusions concerning modified system
emissions control performance are discussed in this section.
Modified System Emissions Performance and Operating Characteristics
Full load N(\ emissions for the original and modified reburn system are shown in Figure 12-1.
A comparison tabulation of NOx and CO emissions for full-load parametric testing of the
original reburn system, the modified reburn system, and the modified reburn system with water
injection (discussed later in this section) is given in Table 12-1. The modified reburn system
NOx emissions were 50 to 75 ppm higher than the original system at a given reburn system
stoichiometry. The NOx emissions reduction of the modified system did, however, satisfy the
program emission control goal of 50% NOx emissions reduction at lull load during the
parametric testing. Reasons for the lower NOx reduction with the modified system during
parametric testing were unclear, and indeed, later long-term results with the modified reburn
system (June 1992) were essentially the same as those with the original system. Possible reasons
for the different performance noted during parametric testing are discussed in further detail
below.
The modified reburn system was more sensitive to CO formation than the original reburn system.
Figure 12-2 gives a plot of CO versus NOx emissions for the original reburn system and the
modified system. CO levels for the modified system increased from nominal levels of 50 to 100
ppm to 700 to 800 ppm as NOx levels decreased to 325-350 ppm. The "knee" of the CO curve
was lower for the original reburn system where NO, levels were reduced to 300 ppm before
significant CO levels were reached.
12-1
-------
Parametric Testing (Modified Reburn System)
Table 12-1
Full Load Parametric Test Emissions Measurements for the Original Reburn System,
Modified System, and System with Water Injection
Original System
Modified System
Gross
Avg AHin
Avg AHin
NOx Red,
Reburn
Gross
Avg AHin
Ave AHin
NOx Red.
Reburn
Load
NOx
CO
(707 Base)
Zone
Load
NOx
CO
(707 Base)
Zone
Test No,
(MW)
(ppmQ3%)
(ppm)
<%)
Stoieh.
Test No,
(MW!
(ppm@3%)
ppm
(%)
Stoieh.
N1-27C
116.0
345
255
51.2
0.945
N1-102C
114.93
402
40
43.1
0.946
N1-27D
118.0
347
60
50.9
0.955
N1-102D
115.15
399
54
43.6
0.925
N1-27E
116.2
417
22
41.0
1.010
N1-103C
115.32
421
58
40.5
0.899
N1-27F
114,9
517
37
26.9
1.064
N1-104B
114,99
562
20
20.5
1.032
N1-30B
113.5
320
783
54.7
0.969
N1-104C
115,22
371
139
47.5
0.919
N1-31B
110,7
44S
121
36,6
1.002
N1-104D
114.99
347
292
50.9
0,903
N1-31C
115,2
282
550
60.1
0.906
N1-104E
112.18
308
759
56.4
0,909
N1-32C
116.0
508
28
28.1
1.042
N1-105B
115,72
473
45
33.1
0,990
N1-32D
114.0
426
40
39.7
1.002
N1-105C
114.54
319
725
54.9
0,911
N1-32E
113.6
365
22
48.4
1.014
N1-105D
115.45
323
681
54.3
0,913
N1-32F
114.2
295
114
58.3
1.025
N1-105E
115.89
373
145
47.2
0,951
N1-34C
118.4
303
686
57.1
0.974
N1-106B
115.88
340
276
51.9
0,932
N1-34D
115.3
301
773
57.4
0.938
N1-106C
114.95
353
359
50.1
0,922
N1-34E
114.5
301
380
57.4
0,948
N1-108D
115.35
556
20
21.4
1.053
N1-35B
114,9
305
1068
56.9
0.931
N1-108A
116.10
351
347
50.4
0.923
N1-35C
114.7
298
1095
57.9
0.921
N1-108B
116.28
476
38
32.7
0.996
N1-35D
115.4
295
790
58.3
0,942
N1-108C
116.67
581
32
17.8
1.076
N1-35E
115,0
272
1136
81.5
0.925
N1-35F
115,8
275
1501
61.1
0.917
N1-36A
114,5
304
692
57.0
0.935
N1-36B
114.9
288
1324
59.3
0.918
N1-36C
113.9
283
354
60,0
0.914
N1-36D
113,9
301
180
57.4
0.913
N1-37B
114.5
405
24
42.7
0,977
N1-37C
115.1
356
85
49.6
0.959
N1-38C
118.1
368
61
47,9
0,959
N1-39B
113.5
382
33
46.0
0.988
N1-39C
115.4
337
408
52.3
0.947
N1-41B
115.4
355
42
49.8
0,946
N1-41C
115.6
337
72
52.3
0.938
N1-42B
112.7
302
94
57.3
0.946
N1-43B
114.6
325
196
54.0
0.941
N1-43C
115.8
328
154
53.6
0.928
N1-43D
115.1
337
62
52.3
0.940
N1-43E
115.4
334
108
52.8
0.947
N1-43F
115.9
339
122
52.1
0.954
N1-45B
114.6
292
364
58.7
0.949
N1-46B
114.9
303
507
57.1
0.937
N1-46C
114.9
268
468
62.1
0.928
N1-46D
115.6
291
477
58.8
0.932
N1-46E
116,2
301
325
57.4
0.948
N1-47A
115.7
288
157
59.3
0.058
N1-47B
115.6
302
218
57.3
0,949
N1-47C
114.1
288
649
59.3
0,043
N1-47D
116.7
306
205
56.7
0.99S
N1-47E
114.1
306
133
56.7
0.949
N1-47F
116.2
291
246
58.8
0.944
N1-48B
115.0
310
59
56.2
0.049
N1-40C
115.6
311
126
56.0
0.942
N1-48D
115.1
333
35
52.9
0.046
N1-48E
115.1
321
101
54.6
0.937
N1-48F
114.8
295
53
58.3
0.930
N1-48G
114.5
292
85
58.7
0,924
N1-49A
115.2
288
89
59.3
0.045
N1-49B
116.0
295
239
58.3
1.001
N1-50B
116.5
276
483
61.0
0,951
N1-50C
115.0
283
488
60.0
0.S79
N1-67B
114.5
382
26
48.0
0.980
N1-B0B
114.5
281
356
60.3
0.959
N1-60C
114.6
453
29
35.9
1.038
N1-61B
114.5
303
14
52.0
0.954
N1-81C
114.7
324
15
48.6
0.957
N1-81D
113.5
330
14
47.9
0.941
(Continued)
12-2
-------
Parametric Testing (Modified Reburn System)
Table 12-1 (Continued)
Full Load Emissions Measurements for the Original Reburn System, Modified System, and
System with Water Injection
Modified System with Water Leak
Modified System with Water Injection
Gross
Ayg AHin
Avg AHin
NOx Red,
Reburn
Gross
Avg AHin
Awj AHin
NOx Red,
Reburn
Load
NOx
CO
(707 Base)
Zone
Load
NOx
CO
(734 Base)
Zone
Test No.
(MW)
(ppm@3%)
ppm
<%)
Stoich.
Test No,
(MW)
(ppm@3%)
ppm
(%)
Stoich.
110A
113
308
907
56.4
0.966
126H
114
273
157
62.8
0.908
111C
113
269
508
62.0
0.96
126H2
114
312
57
57.5
0.916
111D
114
233
1771
67.0
0.937
12612
113
314
57
57.2
0.918
111E
115
298
1243
57.9
0.946
126K
114
322
46
56.1
0.917
127C
114
334
56
54.5
0.935
127D
115
318
75
56.7
0.936
127E
114
324
44
55.9
0.938
127F
114
323
35
56.0
0.943
129B
114
325
48.2
55.7
0.964
129C
115
304
38
58.6
0,924
129D
114
335
28
54.4
0.945
129E
115
322
45
56.1
0,930
129F
114
332
40
54.8
0.917
130AR
114
307
71
58.2
0.961
1308
114
345
43
53.0
0.954
12-3
-------
Parametric Testing (Modified Reburn System)
800
ci
o
*
X
600
S00
400
300
200
0.8
Modified Rfthum Syttam
Orlgirtai Reburn System
1.0 1.1
Reburn Zone Stolchlometry
1.2
1
0.9
...1
E
0.7 >?
O
Z
0,6 03
_1
0.5
0.4
0.3
1.3
Figure 12-1
Original and Modified Reburn System NOx Emissions at Full Load
1600
1400
1200
1000
E
o.
800
o
o
600
400
200
0
100
0
» Original System
—
0
<
0
~ Mod
fied System
\ *
a
£ a
Original Syste
n
A
Modilie
Systen
i
\ 1
[Sly
4
0
200 300 400
NOx. ppm at 3% 02
500
600
Figure 12-2
NOx-CO Emissions Comparison for the Original System and Modified System at Full Load
12-4
-------
Parametric Testing (Modified Reburn System)
Another difference between the two reburn systems was the presence of greater luminosity in the
reburn zone during operation of the modified system. The luminosity was seen through
observation ports located on the side walls on the operating floor (Elevation 870'-0") and
primarily by a video camera located on a side wall on the operating floor. Although there was no
video camera installed during tests of the original reburn system, visual observation did not
appear to show the same degree of luminosity. The luminosity observed during operation of the
modified system suggested that the chemical environment was somewhat different in the reburn
zone between the two systems. This difference is discussed in further detail below.
Differences in operating characteristics were observed. The heavy build-up of slag on the rear
wall was eliminated by elimination of the FGD system. Elimination of FGD simplified
operation of the rebum fuel feed system since a flue gas recirculation fan can be a relatively high
maintenance piece of equipment. Also, capital costs for commercial installations of reburn
systems will be reduced by elimination of the flue gas fan and associated ductwork and controls.
Elimination of FGD and the thick coating of slag on the rear wall returned the boiler operation
and thermal performance back more closely to the mode that existed under baseline conditions.
The thermal performance of the modified reburn system is compared to baseline performance for
both full load and part load in a subsection below.
Modified System Optimization Tests
Although the modified system achieved the objectives of eliminating the excessive buildup of
slag on the rear wall while maintaining acceptable NOx reduction performance and good system
operability, the interest in bringing the NOx reduction performance up to the level of the original
reburn system remained. Therefore a series of tests was conducted to optimize the reburn system
fuel injector configuration and operating parameters and to evaluate NOx emissions reduction at
reduced loads. The results of these tests are discussed in this subsection.
Variation in Aspirating Air Flow Rate and Gas Nozzle Tip Arrangement
The natural gas injector design (Figure 11-3) provided for use of aspirating air as a precaution to
protect against slag build-up on the gas injector tips. Tests were conducted to evaluate the effect
of aspirating air flow on tip slagging and NOx emissions. In addition, tests were conducted with
the gas nozzle tips removed to evaluate the effect of nozzle area and hence natural gas injection
velocity on reburn effectiveness. Results of the testing are shown in Figure 12-3. Turning off
the aspirating air had a minimal impact on tip slagover since slagover was not found to be a
problem with or without aspirating air. NOx emissions were lower at a given RZS with the
aspirating air off. Therefore, this air was left off in the optimized modified reburn system
configuration. Figure 12-3 also shows that NOx emissions performance was about the same with
the nozzle tips in place or removed.
12-5
-------
Parametric Testing (Modified Return System)
Position of Gas Injectors
The injector nozzle design provided capability for varying the insertion depth of the nozzle tips.
Figure 12-4 shows no consistent effect of nozzle retraction on NOx reduction.
1000
900
aoo
s
700
i
ODD
a.
N
o
600
25
«*>
400
IS
X
300
o
z
200
100
0
0.8
-01— Modified System with Aspirator
• Modified System w/o Aspirator
¦ z5i- * * Modified System w/o Aspirator and w/o Tips
0.7 g
o.e |
O
04 z
m
«< _i
a,i
0.2
0.»
0
0.9 1.0 1.1
Reburn Zone Stoichiometry
1.2
1.3
Figure 12-3
NOx Emissions for Variations in Aspirating Air Flow Rate and Gas Nozzle Tip Arrangements
soo
fr
t3
E
S
cm"
O
s«
w
m
400
300 "
200
08 0.9 1 0 1.1
Reburn Zone Stoichiometry
Figure 12-4
NOx Emissions for Variations in Gas Injector Position
1.3
12-6
-------
Parametric Testing (Modified Reburn System)
Natural Gas Feed Rate
Tests of the modified system with 20 percent reburn gas consistently resulted in higher CO than
tests with 18 percent reburn gas. No difference in NOx emissions was found between these
conditions. In order to maintain CO within acceptable limits during long-term tests, the natural
gas flow and corresponding RZS were limited to the flow corresponding to 16% of the total
energy input.
Reduced Load Parametric Tests
Figure 12-5 shows NOx emissions for the modified reburn system at 86 MWe. The highest
emissions reduction is 35.5%. As with the original reburn system, a decrease in reburn
effectiveness was found at part load. The reduced reburn effectiveness was due to lower initial
NOx values and lower gas temperatures in the rebum zone, leading to slower reburn chemical
kinetics.
Reburn System Configuration and Operation for Long-Term Dispatch Testing
Based on the results of the parametric tests of the modified system, the following reburn system
parameters were selected for the long-term dispatch testing:
• No aspirating air
• No tips for the natural gas injectors
• Natural gas injectors inserted four inches from the furnace walls
• Maximum natural gas flow rate equal to 16% of the total fuel input.
800
700
-O 600
E
CL
a.
o
ss
««» 400
to
X
O -,nn
500
200
100
Original Rebum System
1
• Baseline (0% Gas)
A 9.4% Natural Gas
~ 18.3% Natural Gas
X
•
•
•
•
0.9
Modified Reburn System
0.8
% 16% Natural Gas
X
0.7
X
X ^
~
- 0.6
- 0.5
~
- 0.4
...4. I ...
'
1
0.3
~ 0.2
0,8
0,9 1,0 1,1 1.2
Reburn Zone Stoichiometry
1.3
1.4
09
Figure 12-5
NOx Emissions for the Original Reburn System and Modified Reburn System at 86 MWe
12-7
-------
Parametric Testing (Modified Return System)
Modified System Thermal Performance
Boiler operating data from four tests with the revised rebum system was analyzed to determine
the effect of the revised rebum system on boiler thermal performance with and without the
rebum system in operation. The four representative tests selected for study are the following:
inn t "*t
Test No.
106 A
106B
109E
109 A
Time
1020
1155
2340
2005
Date
11-6-91
11-6-91
11-20-91
11-20-91
Type of Test
Baseline
Reburn
Baseline
Rebum
Load - %
100
100
79
79
Reburn Fuel %
0
18
0
18
Excess Air %
9.2
11.9
24.5
21.4
Main Steam Temp °F
952.6
1000.4
966.8
957.4
Final Reheat Temp °F
925.0
981.8
924.2
920.0
Reburn Stoichiometry
-
0.93
-
0.98
As was the case with the original system, two of the tests were with the reburn system on and two
were with the system shut off (see table above). One pair of teste was at Ml load and the other pair
at 79% of MGR. The operating data for these four tests is shown in Table 12-1. Approximate
reburn stoichiometry was 0.93 and 0.98 for the two reburn tests.
12-8
-------
Parametric Testing (Modified Return System)
Table No. 12-2
Operating Data
Test No.
Baseline
106A
18% Kelttirit
106B
18% Helium
109 A
Baseline
109E
Baseline Reference Data
DATE
11-6-91
11-6-91
11-20-91
11-20-91
2/3/88
TIME
1020
1155
2005
2340
MW (GROSS)
114.4
115.9
91.0
90,3
115
MAIN SIM. FLOW K Ibs/hr
901.3
877.4
684.5
676.9
840
SH DES. FLOW E Ibs/hr
0
0
0
0
SH DES. FLOW W lbs/hi
0
2000
0
0
SH DES. FLOW TOT Ibs/hr
0
2000
0
0
39000
RH DES. FLOW Ibs/hr
0
0
0
0
0
RH FLOW (EST.) (.887 X MS + SPRAY)
799453
77S254
607152
600410
PW PRESS psig
1854
1883
2127
2136
DRUM PRESS psig
1538
1538
1508
1508
MAIN STM. PRES. psig
1470
1470
1470
1470
1470
RH IN. psig
348.4
346.5
266.8
264.5
335
RH OUT psig
NA
NA
NA
NA
308
STEAM/WATER TEMP. (°F)
PW TEMP.
486.8
488.2
464.3
463.9
PRIM. SH OUT. E
713.5
736.6
694.5
705.9
PRIM. SH OUT. W
681.5
709.4
696.3
693.0
PRIM. SH OUT, AVG
697.5
723.0
695.4
699,5
750
SFXL SH IN E
670.6
697.8
680.4
676.8
SEC. SH IN W
718,6
737.3
698.4
710,7
SEC. SH IN AVG
694.6
717.5
689.4
693.7
691
MAIN STM. E
961.0
996.81
953.99
973.81
MAIN STM. W
944.32
1004.0
961.0
959,8
MAIN STM. AVG
952.6
1000,4
957.43
966.8
1000
COLD RH E
NA
NA
NA
NA
COLD RH W
641.7
684.8
621.7
629.0
COLD RH AVG
641.7
684.8
621.7
629.0
692
COLD RH AFTER DES.
641.7
684.8
621,7
629.0
692
HTRHOE
928.1
984.87
922.5
927.1
HTRHOW
922.0
978.9
917,4
921.3
HR RHO AVG
925.0
981.8
920.0
924.2
990
(Continued)
12-9
-------
Parametric Testing (.Modified Reburn System)
Table No, 12-2 - Cont'd.
Operating Data
Test No.
Baseline
106 A
18% REBURN
166B
18% Reburn
109A
Baseline
109E
Baseline Reference Data
GAS/MR TEMPERATURES (°F)
GAS LV. PR] SH E
612
627
581
600
GAS LV, PEL SH W
609
616
599
614
GAS LV. PRI. SH AVG
610,5
621.5
590,0
607
GAS LV. LT. RH
63 5
645
612
633
GASAH1E
653
667
618
636
GAS A HI W
663
681
634
652
GAS AH) AVG
658
674
626
644
747
GAS AIIO E
246
249
230
236
GASAHO AVG
246
249
230
236
26?
AIR AHI
12?
127
125
134
130
AIR AHOE
554
5%
534
544
AIR AHO AVG
554
556
534
544
637
PRI. SH DAMP POS. E
21.00
33.02
9,108
9.108
PRI. SH DAMP POS, W
0.72
1,22
868
8 64
RH DAMPER POS.
11.20
11.20
ll.20
11.20
SPRAY WATER TEMP (°F)
295!
295 J
284 5
283.9
02 A IS % {FROM ESA)
1.8
23
3.8
4.2
FOR (REBURN) lbs/lu
0
0
0
0
GAS WEIGHT - NO PGR Ibs/hr
964700
995100
835100
843000
Based on test data the following items were calculated:
• Component heat absorptions
• Boiler efficiency
• Heat supplied to the furnace
• Gas and air weights
• Furnace exit gas temperature and gas temperature profile through the convection pass
• Secondary superheater surface effectiveness factor
The performance of the components in the rear pass of the unit could not be analyzed in detail
because of insufficient test measurements (gas flow through each of the three lanes was not
measured). However, the heat absorbed by the entire low temperature superheater could be
calculated. On the other hand, the heat absorbed by the low temperature reheater could not be
calculated because there was no outlet steam temperature measurement.
12-10
-------
Parametric Testing (Modified Rebum System)
Thermal Performance Results
The results of the thermal performance calculations are summarized in Table 12-3. At full load
(about 115 MW) the main impact of thermal performance was a shift in the heat absorption from
the waterwalls to the convective pass sections as was the case with the original rebum system.
Reburning decreased waterwall heat absorption by approximately 2.1% while the convective
section heat absorption increased by approximately 2.1%. The decrease in waterwall heat
absorption is due to the decrease in cyclone loading. The increase in convective pass absorption
is attributed to the higher furnace outlet gas temperatures (calculated to be 42°F at the furnace
outlet plane) and increased flue gas weight with reburning.
Comparing the full load tests 106A and 106B it can be seen that the firing of 18% natural gas
resulted in a substantial improvement in boiler operation. Steam temperature at the superheater
outlet increased from 952.6°F to 1000.4°F while at the reheater outlet the steam temperature
increased from 925.0°F to 981.8°F; steam temperature design targets in both cases being 1000°F.
Boiler efficiency, however, dropped 0.74% due to the higher hydrogen content of the natural gas
which results in a higher moisture from fuel loss.
At 79% load the change in thermal performance created by the reburn system was less noticeable
and in the opposite direction. Contrary to the full load tests, waterwall heat absorption increased
0,3% while the convection pass heat absorption decreased by 0.7%. The combination of the
lower gas weight, lower furnace outlet gas temperature and higher steam flow resulted in the
final steam temperatures going down instead of up with reburn. The reason for the increase in
waterwall heat absorption is not clear but may be due to an overall cleaner furnace or more rapid
fuel burnout. The 79% load tests were run at much higher excess air than the Ml load tests.
This could be why the same trends were not observed at both loads.
Boiler efficiency was lower at part load with the reburn system in operation as was the case at
Ml load. Boiler efficiency decreased by 0.59% compared to 0.74% at full load. At full load the
effectiveness of the secondary superheater was about 6% less with reburn. At 79% load the
effectiveness was up 3%. These changes are probably related to changes in overall furnace
cleanliness.
As discussed above, boiler efficiency was decreased during reburn operation because the higher
hydrogen content of the natural gas resulted in higher loss from moisture in the fuel. However,
boiler performance was improved at full load with natural gas due to more nearly achieving
design superheat and reheat steam temperatures. At reduced load the boiler performance was
about the same for the baseline and reburn cases.
12-11
-------
Parametric Testing (Modified Reburn System)
Table 12-3
Summary of Results
TEST NO,: 106A
TYPE OF TEST BASELINE
GROSS MW 114.4
HEAT FROM COAL % 100
HEAT FROM GAS% 0
EXCESS AIR % 9.2
STEAM TEMP. SHO-°F 952.6
STEAM TEMP. RHO-°F 925.0
MAIN STM FLOW LBS/HR 901300
REHEAT STM FLOW LBS/HR 799453
SH SPRAY FLOW LBS/HR 0
RH SPRAY FLOW LBS/HR 0
GAS RECTRC FLOW LBS/HR 0
GAS FLOW THRU CONY PASS 962300
LBS/HR
AIR FLOW THRU AIR HTR LBS/HR 879500
COMPONENT HEAT ABSORPTIONS -
MBTU/HR
PRIMARY SUPERHEATER 104.5
SECONDARY SUPERHEATER 161.3
REHEATER 121.4
WATERWALLS 627.5
TOTAL 1014.7
HEAT LOSSES - %
DRY GAS LOSS 2.32
MOIST FROM FUEL LOSS 4.32
MOIST FROM AIR LOSS 0.06
RADIATION LOSS 0.23
ASH PIT LOSS 0.52
CARBON LOSS 1,39
TOTAL 8.84
BOILER EFFICIENCY % 91.16
BTU FIRED MBTU/HR 1113.1
LBS FUEL FIRED 95990
SURFACE EFFECTIVENESS FACTORS
SECONDARY SUPERHEATER 0.975
GAS TEMPERATURES - °F
SECONDARY FURN. OUTLET 2078
REHEATER INLET 1555
REAR CAV. OUTLET 1283
PRIMARY SH & LTRH INLET 1269
AIR HEATER INLET 658
AIR HEATER OUTLET 246
106B 109E 109 A
18% REBURN BASELINE 18% REBURN
115.9 90.3 91.0
82 100 82
18 0 18
11.9 24.5 21.4
1000.4 966.8 957.4
981.8 924.2 920.0
877400 676900 684500
778254 600410 607152
2000 0 0
0 0 0
0 0 0
992600 841000 833100
914300 776700 771800
120.5 79.7 78.2
163.1 124.5 124.5
126.4 93.4 95.4
608.2 488.6 494.1
1018.2 786.2 792.2
2.41 2.26 2.24
5.29 4.28 5.17
0.06 0.05 0.05
0.23 0.30 0.30
0.43 0.52 0.44
1.16 1.39 1.19
9.58 8.80 9.39
90.42 91.20 90.61
1123.6 862.2 874.3
89359 74353 70056
0.912 1.045 1.074
2120 1913 1892
1615 1442 1417
1348 1195 1114
1334 1183 1104
674 644 626
249 236 230
12-12
-------
Parametric Testing (Modified Reburn System)
Evaluation of the Modified Reburn System Design and Performance
Reviews of parametric test results and discussions between project sponsors, contractors, and
consultants were held during December 1991 and early January 1992 to identify the reasons why
the modified reburn system gave lower NOx reduction than the original system and to develop
recommendations for improving NOx reduction. These discussions and the recommendations are
summarized in this subsection.
The observation of luminosity in the reburn zone during testing of the modified system was an
important input in identifying the cause of the lower NOx reduction for the modified system.
Another observation was the reduced thickness of the slag layer and resulting increased heat
absorption in the waterwalls although the elimination of FGR may have compensated all or in
part for this effect. The key reasons for the lower NOx reduction were thought to be:
• Pyrolysis of the natural gas during operation of the modified system, evidenced by the
greater luminosity rather than the desired chemical reaction, which is hydroxylation.
Pyrolysis converts natural gas into carbon (soot formation) and hydrogen rather than the
more reactive intermediate chemical compounds such as CII, Clla, CHQ, and OH which are
generated during hydroxylation and which are essential for the reburn process.
• Reduced temperatures in the reburn zone, which slowed the rate of the NOx-destroying
reburn chemical reactions. Possible lower temperatures in the reburn zone may have resulted
from a combination of these factors:
Greater heat transfer to the waterwalls in the reburn zone due to the reduction of the slag
layer thickness on the back wall.
- Greater heat transfer to the waterwalls in the reburn zone due to higher flame emissivity
of the more luminous gases.
Offsetting these effects to some extent but not sufficiently to counteract them completely
was the elimination of the recirculated flue gas and the attendant thermal dilution.
It was believed that C02, H2O, and the O2 in the recirculated flue gas used in the original design
caused hydroxylation of methane to occur rather than straight pyrolysis. A question which
remained unanswered, however, was how much flue gas is needed to prevent pyrolysis. The
quantity may be very small because the performance of the original system was essentially
unchanged when the FGR was reduced from 16% to 3%. Mixing effectiveness with the modified
reburn system did not seem to be a factor, at least not at Niles No. 1. This conclusion was based
on the fact that N0X removal performance was essentially unchanged when the natural gas tips
were removed or when other changes were made to the natural gas injector configurations. One
final observation about the importance of reburn zone temperature was provided by the fact that
the reduction in NOx was significantly lower at part load for the modified system than for the
original system. Under part load rebum operation with the modified system the reburn zone was
especially cool because of the combined effects of the reduced slag layer thickness, the greater
12-13
-------
Parametric Testing (Modified Reburn System)
gas luminosity, and the inherently cooler temperatures during part load operation. It should also
be noted that NOx reduction decreases as the inlet NOx concentration decreases.
The following recommendations were made for improving NOx reduction:
• Premix steam and/or water with the natural gas before injection into the reburn zone,
« Premix a small amount of flue gas with the natural gas before injection into the reburn zone.
Project Planning
A meeting of project sponsors, contractors, and consultants was held on January 15,1992. At
this meeting the modified system parametric test results were reviewed, reasons for the lower
NOx were discussed, recommendations for improving NOx reduction were presented, comments
from Ohio Edison regarding operation of the modified reburn system were presented, boiler tube
wastage measurements were reviewed, recommendations for effective continuous NOx
measurements were presented, the project budget status was discussed, and project plans were
formulated. Because of time and budgetary constraints, a decision was made to proceed with
long-term load dispatch testing using the modified reburn system and to postpone parametric
testing to improve NOx emissions reduction until completion of the long-term tests. The long-
term tests are discussed in Section 13.
Water Injection
Adding water to the reburn zone was determined to be the most economic and technically
feasible method for improving the performance of the modified reburn system. Design and
fabrication of combined natural gas and water injectors proceeded during the long-term dispatch
testing. However, an opportunity for a brief evaluation of water introduction to the reburn zone
was provided in January 1992 during the initial start-up of the long-term testing. This occurred
when water leaks developed in the water-cooled guide-tubes of two of the five natural gas
injectors. NOx and CO emissions data for the brief water-leak and the more detailed, controlled
water injection tests conducted in July 1992 are listed in Table 12-1. Data for the full-load
parametric tests of the original and modified reburn systems are also listed in Table 12-1. The
water injection tests are discussed in further detail in the following paragraphs.
Parametric Testing with Water Leaks through Guide Tubes
During the initiation of long-term load dispatch testing, water leaks were detected on reburn
Nozzle A and Nozzle E water-cooled guide pipes. Nozzle A was on the far left (west) side of the
furnace, and Nozzle E was on the far right (east) side of the furnace. These leaks added water
into the reburn zone in an uncontrolled manner since there was no measurement of the flow rates
and no positive indication of where on the guide pipes the leaks were located. However, since
the NOx monitoring indicated a reduction in NOx, NOx and CO emissions data were recorded.
NOx emissions comparisons at full load are shown in Figure 12-6. With natural gas injected only
through Nozzles B, C, and D (the center three nozzles) the NOx emissions were reduced by 50 to
75 ppm to the level comparable to NOx emissions measured with the original reburn system.
12-14
-------
Parametric Testing (Modified Reburn System)
With natural gas injected through all five nozzles, there was further reduction in NOx—a
reduction of approximately 100 ppm compared to previous data for the modified system.
0.8
Modified Reburn System
Water Leaks A & E, Gas Through B. C, D
Water Leaks at A & E, Gas Through All 5 Nozzles
1
0.1
0.i
0.1
0.1
OJ
(M
04
0.J
0.1
0
3
CQ
E
JE
CQ
09 1.0 1.1
Reburn Zone Stolchiometry
1.2
1.3
Figure 12-6
NOx Emissions for the Modified Reburn System and Tests with Water Leaks
Parametric Testing with Water Injection
As discussed above, the parametric reburn tests of the modified rebum system with water
injection were conducted in July 1992, after completion of the long-term dispatch tests to be
discussed in Section 13.
Figure 12-7 is a sketch of a modified system natural gas injector with a water injection atomizer
added. The natural gas tip was removed. Water entered through a pipe in the center of the
natural gas passage and was injected through a pressure atomizing spray nozzle. In a
modification of the design, the pressure atomizing water spray nozzle was removed and the
water entered the natural gas passage from the end of the water pipe, A third water injector,
called the "doughnut injector," was designed to simulate the leak in the guide tube. For this
design, water flowed through a separate annulus inside the water cooling passage and entered the
natural gas stream flowing radially inward through holes at the tip of the water annulus.
Doughnut injectors were installed in nozzle locations A and B for the final scries of water
injection tests.
12-15
-------
Parametric Testing (Modified Reburn System)
Cooling
Water
Outlet
Injection
Water
Inlet
J~L
#1
wd
Natural
Gas
Inlet
Cooling
Water
Inlet
Figure 12-7
Water Injection for the Modified System
NOx emissions for the full load parametric testing of the modified system both without and with
water injection are shown in Figure 12-8. Some of the NQX levels achieved during the water leak
tests (Figure 12-8) are lower than what was achievable during the controlled water injection
testing. In an effort to explain the differences between NGX levels during water leak tests vs.
controlled water injection testing the oxygen levels for each cyclone were reviewed. It was
found that two of the four cyclones were operating at very low O, during the water leak tests; this
is verified by the high CO levels as shown in Table 12-1 (Test No. 110A, 111C, 111D, and
11 IE) which ranged from 508 to 1771. It can be reasonably speculated that the NO, levels for
those cyclones with low 02 would have been lower than had the cyclones all been operating at
the target 2.5 to 3.0 percent 02. It can be further speculated that the inlet NOx to the reburn zone
would not have been as high in the water leak tests and that might be the real explanation for
why the water leak tests gave lower NOx which could not be repeated during the controlled water
injection testing.
Also shown in Figure 12-8 are data for the long-term tests conducted in June 1992, reburn tests
with water off conducted in July 1992, and the water leak tests discussed above. The data show
an improvement in NO, removal for the water injection tests relative to the parametric tests
conducted in October and November 1991. However, the water injection tests gave NOx
removal performance a bit lower than the performance for the long-term tests conducted in June
1992 and the three baseline tests with gas on and water off conducted in July 1992. The small
drop-off in NOx removal during the water injection tests may have been due to cooling of the
gases and resulting slowing of the reburn chemical kinetics. The excellent NOx removal
efficiency achieved during the June 1992 long-term tests is discussed further in Section 13.
12-16
-------
Parametric Testing (Modified Reburn System)
600
MO
a 400
o
s?
n
£ 300
0.
a
K
O
Z 200
100
Loru
Parametr
Term Tests \
b Tests o
<
Water Irtjecti
\,
in TestV 2^*
C
n
> \ °
cu | „,p-i
1
~ o o
p 8 ®
1° a
*
O
a a
a ¦
B
» Modified System Panraetrto Tests
o Modified System June 1992 L. T. Tests
a Modified System Wafer Injection Tests
¦ Modified System Wafer Leak Tests
4 July 1692 Tests wtlh Gas on and Water of)
-
—1 1
0
0.850
0.8
0,7
0.8 3
CO
E
o.s E
0,900 G.950 1.000
Reburn Zona Stoichiomatry
0.4
0.3
0.2
0.1
0
1.0S0
1.100
*
o
z
m
Figure 12-8
Modified Reburn System NOx Emissions at Full Load for Parametric Tests, Long-Term Tests in
June 1992 and Parametric Tests with Water Injection
NOx, ppm at 3% 02
Figure 12-9
NOx-CO Emissions Comparison for the Modified System Without and With Water Injection
12-17
-------
Parametric Testing (Modified Reburn System)
Figure 12-9 compares the CO emission data for the water injection reburn tests and the
uncontrolled water leak tests to the CO emission for the modified reburn system without water
injection. All of the tests showed an exponential rise in CO as reburn zone stoiehiometry, and
corresponding NOx emissions, were decreased. However, water injection provided a substantial
reduction in CO emission relative to rebum operation without water injection. A wide range of
water flows and a great variety of water injector configurations were tested including water
injection through open pipes with the water atomizers removed. The CO concentration was
nearly unchanged for this range of water flows and injector configurations. This suggests that
the CO reduction was a chemical phenomenon rather than a mixing - limited phenomenon and
that only a small amount of water may be needed to reduce CO to low levels. Inspection of
Figure 12-8 also shows that the water injection tests were conducted over a small range of reburn
zone stoichiometries. The water injection tests were conducted at the very end to the test
program under tight time constraints which did not permit testing over an adequate range of
water flows or rebum zone stoichiometries. It is interesting to speculate on what the results may
have been if testing had been conducted at reduced total water flow rates, such as 2 or 3 gallons
per minute, and lower reburn zone stoiehiometry. If reduced water flows had given NOx
measurements similar to the long-term tests and CO emission similar to the other water injection
tests, the operations would simultaneously have the minimum emissions of both NOx and CO.
Extrapolating the minimum emissions lines in Figures 12-8 and 12-9 would give estimates of
NOx and CO emissions of NOx = 249 ppm at 3% 02 and CO =180 ppm at a reburn zone
stoiehiometry of 0.90 and NOx = 230 ppm at 3% 02 and CO = 259 ppm at a reburn zone
stoiehiometry of 0.88. The corresponding NO, reductions, relative to a baseline NOs of 670 ppm
are 62.8% and 65.7% at reburn zone stoichiometries of 0.90 and 0.88, respectively.
Conclusions
Testing of the modified reburn system with water injection led to the following conclusions:
1. The NOx removal performance for the modified reburn system with water
injection was better than the performance achieved during the parametric testing
of the modified reburn system. However, the performance with water injection
was no better, and perhaps a bit poorer, than the NOx removal performance
achieved during long-term reburn testing in June 1992 and during rebum tests
with gas on and water off conducted in July 1992.
2. For all reburn systems tested, CO emission increased exponentially as rebum zone
stoiehiometry and corresponding NOx emissions were reduced.
3. The modified system with water injection provided a reduction in CO emissions
compared to all tests with the modified system without water injection.
4. The effectiveness of water for reducing CO emission was independent of the
quantity of water used over the full range of water flows tested.
5. The method of mixing water within the reburn zone appeared to be of minimal
importance, at least at Niles Unit No. 1, because NOx and CO emissions were
about the same for a wide range of water injection configurations.
6. Controlled water injection during natural gas reburning has the potential for
concurrently providing minimum emissions ofboth NOx and CO.
12-18
-------
13
LONG-TERM LOAD DISPATCH TESTING
Purpose of Long-Term Testing
Long-term testing was initiated after parametric testing had established the effect of major
reburn system variables on system performance. Long-term testing was conducted to:
• document the reliability of the system,
• compare system performance under fluctuating boiler load and excess oxygen operating
conditions as compared to the closely controlled operating conditions present under
parametric testing,
• evaluate the potential for changes in tube wastage caused by the reducing atmosphere created
in the reburn zone,
• document the effects of reburning on boiler equipment, operations and performance under
long-term commercial power-plant operation.
Data was logged for 1196 hours of operation between March 2,1992 to June 19,1992. This
section discusses long-term operation of the reburn system, discusses the NOx emission reduction
performance of the system in long-term service, including a comparison with performance
measured during parametric testing, discusses the commercial potential of gas reburn for NOx
emissions control, and presents a utility perspective of the reburn process for retrofit of cyclone-
fired furnaces based on 3 1/2 month's experience at Niles Unit No. 1. This information expands
on information given by Brown and Borio (1992) and Borio et al. (1993).
Reburn System Operation
The reburn system was designed to operate at the design reburn fuel heat input of approximately
16% at loads greater than 80 MW. Below 80 MW, the reburn fuel heat input was to be
proportionally ramped down to 0% at 65 MW. The reburn fuel flow was restricted in this
manner to maintain sufficiently high temperatures in the primary combustion zone to keep the
slag molten and permit tapping without undue difficulty. The reduction in reburn fuel flow led
to decreased NOx reduction during part-load operation. Moreover, during long-term testing, the
unit was never operated in a rebum mode below 80 MW, the primary reason being operator
judgment relative to slag tapping concerns.
Based on parametric testing conducted immediately after the modified reburn system was
installed, it was decided to operate the rebum zone at a target stoichiometry of 0.94. This
13-1
-------
Long-Term Load Dispatch Testing
stoichiometry was higher than that used as the target during parametric testing; reasons for the
higher stoichiometry will be discussed later.
Niles Unit No. 1 operated 2439 hours during the time period between March 2, 1992 and
June 19, 1992. Data acquisition was in operation for 1196 hours during this time period. Data
acquisition times included times when the reburn system operated at design conditions (gas
energy input 16% or greater fraction of total energy input), times of off-design reburn operation
(gas input 3% to 16% of total energy input), and times when baseline data were obtained by ,
operating the data acquisition system with the reburn system not in operation. The distribution
of data acquisition times by load range and reburn system operating range is listed in
Table 13-1. As noted above, the reburn system was never employed when the load dropped
below 80 MW. Additionally there were periods when reburn fuel was not injected at loads above
80 MW because of cooling water leaks in the reburn fuel guide pipes, a furnace casing leak, and
malfunctions of the natural gas control valve and instrument air compressor. Corrective actions
for these mechanical malfunctions, which can be considered typical for the first commercial
installation of a gas reburn system, are available.
Table 13-1
Distribution of Load and Reburn Conditions during Long-Term Data Acquisition
Load Range
(MW gross)
Hours of Operation
Design Reburn
Operation
(16+ % gas)
Off-Design
Reburn
Operation
(0 -16% gas)
Baseline
Operation
(0% gas)
Total
Data
Acquisition
Hours
110+
154
16
30
200
110-100
213
63
57
333
100 - 90
187
150
129
466
90-80
9
34
65
108
below 80 MW
0
0
89
89
Total
563
263
370
1196
Long-Term NOx and CO Emissions
Long-term NOx emissions data for tests between March 2 and April 29,1992 are presented for
four boiler load ranges in Figures 13-1,13-2, 13-3, and 13-4 where NO, is plotted against Reburn
Zone Stoichiometry (RZS). Each point is the arithmetic average of twelve measurements logged
at five minute intervals. One-hour average test data for full-load operation with reburn fuel
fraction of 16% or greater are also presented in Table 13-2. NOx emissions during the March
through April time frame averaged about 370 ppm at full boiler load rather than the original
system parametric test values in the 300 to 330 ppm range principally because of the higher
13-2
-------
Long-Term Load Dispatch Testing
800
£" 700
¦o
£
CL
CL
CM*
O
2 500
X
O
Z 400
600
300
200
March 2 - Aprll2fl, 1092 Data
A
.O
3-
°o
o Coflre (3 to 18% gas)
a Baseline (0% gas)
0.8
0.0 1.0 1.1
Rsburn Zone Stolchlometry
1.2
T 1
0.9
0.8 3
ffl
0.7 |
E
1?
0.6 O
0.5
0.4
0.3
m
1.3
Figure 13-1
Variation of NOx Emissions with RZS at 110+ MWe
BOO
700
T3
J™
o
°? 500
400
300
200
0.8
a ** March 2 - Aprlf28,1982 Data
m. Tt*
"i. _
4 * A \ A *¦ A
o C of ire (3 to 18% gas)
a. Baseline (0% gas)
1.0 1.1
Reburn Zone Stolchlometry
1.2
1.3
Figure 13-2
Variation of NOx Emissions with RZS at 100 -110 MWe
13-3
-------
Long-Term Load Dispatch Testing
800
& 700
n
£
!-
o
? 500
400
300
200
March 2 - Apf1129,1892 Data
A
0.8
o Cofire (3 to 18% gas)
a Baseline (0% gas)
1.0 1.1
Reburn Zone Stolehlometry
1.2
1.3
Figure 13-3
Variation of NOx Emissions with RZS at 90 - 100 MWe
BOO
700
600
E
CL
CL
e*
O
85
f 500
x
o
z 400
300
200
0.8
Msrch 2 - Apf II2S, 1982 Data
o
.A ^
Ass4*
O Q}0"" O
.'"So _
o°
*
A A
oo
o Cofire (3 to 18% gas)
jk Baseline (0% gas)
0.9 1.0 1.1
Reburn Zone Stolehlometry
1.2
Figure 13-4
Variation of NOx Emissions with RZS at 80 - 90 MWe
1.3
13-4
-------
Long-Term Load Dispatch Testing
Table 13-2
Full-Load Long-Term Emissions Data for Reburn System Operation with 16% or Greater
Natural Gas Reburn Fuel
Long Term Tests March 2-April 28, 1992 with
Long Tarm Tests June 12-June 19,1992 with
Load >=110 MW and Gas >= 18%
Load >= 80 MW and Gas >= 16%
Qross
ESP in
ESP in
Reburn
NOx Red,
Gross
ESP in
ESP ill
Reburn
NOx Rid.
Load
NOx
CO
Zone
(672 Base)
Load
NOx
CO
Zone
(670 Base)
Date
Hour
(MW)
;ppm®3%
ppm
Stole h.
(%)
Date
Hour
(MW)
(ppm® 3%
ppm
Stoic h.
(%)
304
19
115.1
291
468.0
0,9802
56.78
613
1
94
325.6
79.4
0,9536
51.41
304
20
114.2
292
437.5
0,9905
56.65
613
10
111.5
309.4
155.8
0.9474
53.83
318
7
110.7
391
66.7
0.9118
41.85
613
11
95.4
364.9
176.1
1.0227
45.54
316
8
113.5
385
24.5
0.9323
42.76
613
12
108.2
222.6
1514.3
0.8920
66.78
318
9
113.4
390
266.4
0.9417
42.04
613
13
110,1
234 1
1032.8
0.8909
65.06
318
12
111.8
421
147.9
0.9454
37.43
613
14
96.2
275
268.2
0.9602
58.06
318
13
114.8
409
148.4
0.9398
39.12
613
15
103,8
328
131.4
0.9644
51.05
318
14
115.3
408
275.6
0.9423
39,31
613
16
113.6
399.2
23.7
1.0167
40.42
318
15
113.5
426
113.9
0.9514
36.66
613
17
103,6
372.5
25
1.0167
44.41
318
15
114.8
405
248.8
0.9456
39,75
613
18
107.5
382.7
26.2
0.9979
42.89
318
17
113.6
406
198.8
0,9548
39.57
613
19
98.1
381.7
14,9
1.0155
43.04
318
19
114.6
396
301.3
0.9375
41.10
613
20
92.9
368.5
20.1
1.0337
45.01
318
20
113.3
397
232.2
0,9417
41.03
613
21
111.3
360.7
281.4
0.9856
46.17
318
21
113.0
387
363.8
0.9553
42.50
613
22
113.7
303.5
132.9
0,9462
54,71
319
8
114.3
386
212.0
0.9508
42.57
614
12
102.4
318,9
139.3
0.9889
52,41
319
9
113.5
405
67.4
0.9451
39.72
614
13
98.3
345.8
61.1
0.9942
48.39
319
10
113.1
402
76.1
0.9527
40.17
615
10
115.7
279.5
292
0.9343
58.29
319
13
112.7
394
80.2
0.9555
41.41
615
11
115,5
203
131.1
0.9434
56.27
319
21
111.6
389
109.7
0.9390
42.14
615
12
116.8
384.6
30.3
1.0227
42.60
320
19
115.3
410
206.2
0.9341
38.97
615
13
118,9
400.1
2.3
1.0143
40,29
320
20
110.6
417
191.3
0.9581
37.96
616
14
115,5
343.7
60.2
0.9718
48,71
320
22
111.3
409
86.1
0,9554
39.14
615
15
113,7
401.3
S.2
1.0251
40,11
321
19
112.5
412
337.7
0,9117
38.77
615
16
103.2
417.2
1.3
1.0412
37.74
321
20
111.4
401
216.6
0,9325
40.30
615
17
97.6
428.8
-3.9
1.0475
35,89
322
9
113.7
393
206.8
0.9229
41.59
615
20
101.7
387.9
32.2
1.0425
42.11
323
13
112.0
415
404.2
0.8948
38.23
615
21
113,2
365.6
86,8
1.0025
45.44
323
21
111.1
403
268.4
0.9356
40,14
615
22
115,8
365.5
2.1
1.0239
45.45
324
17
114.2
371
562.0
0.8932
44,80
616
9
100,2
257.5
657.8
0.9332
61.57
324
18
115.3
358
730,7
0.8920
46.70
616
10
101.5
309.8
136.2
0.9544
53.77
324
19
115.2
359
396.4
0.8872
46.55
616
11
106.3
328.1
73
0,9249
51,04
324
20
114,4
357
466.2
0.8924
46.94
616
6
100,1
362.3
50.4
1.0276
45,93
324
21
114.8
352
629.1
0.8998
47.67
616
7
101.2
372.1
67.9
1.0180
44,47
324
22
114,3
349
706.0
0.9094
48.07
616
8
100.4
358.2
53.1
1.0096
46.54
325
2
114.6
356
468.0
0.9212
47.00
616
9
106.4
382.5
102.4
1.0215
42.92
325
3
115.2
374
141.4
0.9226
44.34
616
10
100.1
327.3
332.8
0.9730
51.15
325
4
113.7
376
141.8
0.9354
44.03
616
11
94.9
388.2
104.1
1.0126
42.07
325
6
113.2
364
380,5
0,9348
45,88
616
19
95
325.9
346.4
1.0240
51.36
325
14
111.0
399
55.0
0.9263
40.60
616
20
106.4
341.8
267.8
1.0130
48 99
325
15
112.2
391
86.5
0.9389
41.81
616
21
105.1
310.9
231.5
0.9785
53.60
325
16
112.9
387
95,7
0.9375
42.39
616
22
109.3
318.4
189.2
1.0083
52,48
325
17
111.4
386
349.7
0.9400
42,63
616
23
114.8
309.2
262.5
0.9633
53,86
325
18
114.8
380
222.5
0.9101
43.51
617
0
115.8
293.2
178.8
0.9526
56.24
325
19
110,0
392
156.2
0.9310
41.75
617
1
115,7
300.4
132.6
0.9609
55.17
325
21
112.2
395
72.9
0.9187
41.29
617
2
114.5
322.8
54.4
0.9817
51.83
32S
a
111.2
408
197.2
0.9333
39.39
817
3
101.6
351
25.4
1.0437
47.62
326
10
111.8
403
323.3
0.9174
40.10
617
4
108.3
319.9
107.7
0.8991
52.26
320
13
111.2
382
130.2
0.9226
43.20
617
5
115.2
322.2
89.2
0.9865
51.92
327
a
111.9
395
71.8
0.9120
41.31
617
6
107.7
338.8
369.5
1,0015
49.44
327
13
112.2
429
106.0
0.9130
36.25
617
7
99.7
327.5
69.4
1.0083
' 51,12
328
8
111.6
405
591,3
0.8736
39.75
617
8
98,8
364.4
58
1.0399
45,62
323
8
115.6
382
1023.0
0.8721
43.25
617
9
112.9
337.4
160.7
0.9746
49,65
328
10
115.5
378
1063.9
0.8716
43.80
617
10
114.7
288.2
179.9
0.9497
56,99
328
11
110.3
384
586.1
0.8823
42.89
617
11
112.2
299
68.7
0.9675
55,38
328
12
112.2
401
208,2
0.9142
40.31
619
1
99,5
338.4
54.8
1,0000
49,50
329
9
110.8
390
862.1
0.9267
42.01
619
5
95.1
351.5
58
1.0450
47,54
330
2
110.3
413
176.0
0.9490
38.53
619
6
99.2
331
28.6
1.0130
50.60
330
6
112,3
383
416.3
0.9175
43,00
619
7
100.5
251,9
1124.6
0.9259
62.41
330
7
111.4
373
437.4
0.9259
44.50
619
8
102.6
227.8
1167
0.8064
68,00
330
8
113.5
377
149.8
0.9262
43,91
619
9
107.4
289.5
316.5
0.9248
58.80
330
S
111.9
381
80.1
0.9275
43.29
618
10
111
329.6
162.3
0.8515
50.81
330
16
110.4
399
227.8
0.9377
40.66
331
0
111.3
383
433.8
0.9114
42.89
331
1
113.2
385
613.1
0.8982
42.70
331
2
111.5
383
262.5
0.9065
43.10
331
3
111.3
383
286.1
0.9088
43.08
331
4
111.4
395
339,3
0.9168
41.29
(Continued)
13-5
-------
Long-Term Load Dispatch Testing
Table 13-2 (Continued)
Full-Load Long-Term Emissions Data for Reburn System Operation with 16% or Greater
Natural Gas Reburn Fuel
Long Term Tests March 2-April 28, 1B92 with
Long Term Tests June 12-June 19, 1992 with
Load >= 110 MW and Gas >= 16%
Load >= 80 MW and Gas >= 16%
Gross
ESP in
ESP In
Rebum
NOx Red.
Gross
ESP in
ESP ir.
Rebum
NOx Red,
Load
NO*
CO
Zone
[672 Base
Load
NOx
CO
Zone
[670 Base)
Date
Hour
(MW)
;ppm©3%
ppm
Stolch.
(%)
Date
Hour
(MW)
Sppm®3%
ppm
Stolch.
(%S
331
21
110.4
373
220.1
0.8997
44.53
401
6
113.4
382
701.7
0.8805
43.28
401
16
110.5
358
58,2
0.8935
46.69
405
20
115.1
362
80.1
0.9080
46.19
408
a
110.8
398
39.9
0.0653
40.81
400
21
111.7
446
20.2
0.9980
33.68
406
20
114.1
441
21.8
0.8650
34.36
406
21
111.7
#46
20.2
0.8960
33.68
407
16
111.3
126
2457.7
0.7777
81.30
407
17
112.7
314
825,8
0.9278
53.37
407
21
112.7
400
21,2
0.9957
40.50
408
7
111.7
155
2458.4
0.8061
76.96
400
18
110.8
201
1795.9
0.8432
70.11
408
21
113.6
185
2523,5
0.8281
71.04
411
10
112.4
354
46,1
0.8512
47.42
411
11
110.5
354
54.4
0.9389
47.38
413
22
115.1
280
1168,7
0.8740
58.44
413
14
115 9
422
16.9
0,8548
37.24
413
15
115.9
418
18.7
0.8587
37.77
413
16
112.2
421
18.1
0,8754
37.48
413
20
113.1
338
498.5
0.8995
49.68
413
22
115.1
280
1168.7
0,8740
58.44
413
23
115.4
335
91,4
0.9092
50.18
414
8
115.5
366
84.3
0.9202
45.61
414
0
116.3
363
36,5
0.9255
46.05
414
1
112.9
384
20.9
0.9425
42.92
414
2
112.8
380
23.0
0.9529
43.52
414
3
111.0
380
19.6
0.9525
43.53
414
4
110.6
380
28.8
0.9366
43.44
414
5
113.8
383
26.4
0.9428
42,96
414
6
113.8
390
20.1
0.9411
41.98
414
8
115,5
366
84.3
0.9202
45,61
414
11
110.7
365
119.6
0.9303
45.69
414
12
112.7
369
33.5
0.9520
45.19
414
13
111.1
365
42.8
0.9581
45.69
414
15
115.8
342
240.9
0.9285
40.12
414
16
114.0
347
202.6
0.9313
48.35
414
17
114.4
334
307.6
0.9212
50.31
414
18
115,5
334
176.6
0.9262
50,31
414
19
115.5
347
63.1
0.9332
48.36
414
20
115.8
348
59.1
0.9342
48.28
414
21
115.4
380
45.4
0.9329
48.46
414
22
114.8
340
85.6
0.9253
49,46
415
8
115,0
342
180.7
0.9106
49.10
415
6
110.0
351
228.0
0.9244
47.74
415
7
111.4
355
173.6
0.9188
47.15
415
8
115.0
342
180.7
0.9106
49.10
416
9
115.8
352
94.3
0.9250
47.65
415
10
115.0
370
36.6
0.9477
45.02
415
11
113.4
377
23.4
0.9510
43.99
415
12
112.1
381
25.7
0.9666
43.32
415
15
116.2
394
28.2
0,9417
41.37
415
16
115.3
391
23.1
0,9413
41.80
415
17
115.2
375
36.1
0,9335
44.23
415
18
114.5
364
41.1
0.9296
45.93
415
19
113.5
360
73.8
0.9313
46.38
415
20
115.9
363
56.0
0.9235
48.03
415
21
115.1
367
30.1
0.9690
45.47
415
22
114.7
330
79.1
0.9326
50.00
415
23
115,1
338
109.5
0.9143
49.58
416
0
115.8
335
161.0
0.9020
50.14
423
9
114,6
325
872.9
0.8838
51.68
423
10
111,4
263
2183.2
0.8674
60,91
423
11
114,9
285
2271.7
0.8818
57,65
423
12
115,4
343
853.2
0.9134
48,95
423
13
115,8
354
605.5
0.9168
47.41
13-6
-------
Long-Term Load Dispatch Testing
target RZS arid other factors which may differentiate performance between the original and
modified reburn systems as discussed in preceding sections. It is not uncommon for long-term
NOx emissions to run somewhat higher than values achieved during parametric testing (Wilson et
al. (1991)). The relatively steady state conditions that exist during parametric testing minimize
the fluctuations in air/fuel ratios that are bound to occur with load swings during normal boiler
operation.
NOx emissions during June 1992 reburn tests are presented in Figure 13-5 and Table 13-2. The
June long-term data show significantly lower NOx emissions than the NOx emissions data
measured between March 2 and April 29,1992. During the June 12 to June 19,1992 time periods
there were sixty one-hour tests when the unit operated with 90 MWe or larger gross load and 16
or greater percent reburn fuel. There were twenty-one (21) one-hour tests with operation in the
RZS range between 0.90 and 1.00. The average NOx reduction for tests at foil load with
acceptable boiler operation and CO emissions lower than 200 ppm was 52.1% when the RZS was
between 0.90 and 1.00. This NOx emissions reduction performance is superior to the NOx
reductions measured during parametric performance measurements for the original reburn system
as shown in Figure 13-6. The long-term measurements verify that the performance achieved
with the original reburn system can be duplicated with the modified reburn system. The most
reasonable explanation for why the excellent NOx emissions reduction were achieved in June
1992 is that the boiler operators became more experienced at maintaining target air/fuel ratios, in
particular more uniform air/fuel ratios among the cyclones. Significantly, these results were
obtained near the end of the long-term testing which provided increased operator familiarity and
the ability to control key operational parameters within a tighter margin. It is believed that the
June data are, indeed, representative of the performance that can be expected from the modified
reburn system.
It was generally observed that the slope of the curves relating NGX to RZS increased with
increasing load. Figure 13-7 depicts the variation of NOx with RZS for various load ranges. At
high loads, 100-110 MWe for example, the curve for the March-April time period has a steeper
slope than the 80-90 MWe load range curve. This higher percentage of NOx reductions at higher
loads is probably due to higher gas temperatures in the reburn zone at high loads and the higher
inlet NOx concentrations to the reburn zone. For comparison, the June data are also shown on
this plot. As previously, noted the NOx values measured in June were lower than the March-
April values.
13-7
-------
Long-Term Load Dispatch Testing
800
>. 700
E
&
. BOO
CM
O
aft
CT soo
&
X
Q
z 400
300
200
0.8
-i r
Ju0612- June 19,1982
* *
/
A
A X
V
o
o ,
go
L.
o
o o° 9*
a s
a §r v'
° dao
xf'
O O
offo
o C of ire (3 to 18% gas)
4 Baseline (0% gas)
0 J 1.0 1.1
Reburn Zone Stolchiomatry
12
1
0.9
3
o,8 «?
0.7 •»?
o
0.6
0.5
0.4
0.3
m
1.3
Figure 13-5
Variation of NOx Emissions with RZS at 90-110+ MWe
600
500
e 400
a.
a.
of
O
S* 300
200
100
0
0.8S0
Original Systei
n Parametric Tes
\
ts
*
Mc
dified System L.
T. Tests
\
• \>>
* » 3--'"°^
X ~ ~ J3
~ \ .
1
A
a a a
fH D
aH °
o
a
~ »
~ *
» Original System Parametric Tests
a Modified System June 1992 L. T.
Tests
-
1 1
0.8
0.7
0.6 3
m
£
0.5 E
¦»
o
0.4 Z
CQ
_J
0,3
0.2
0.1
O
0.S00 0.950 1.000
Rebum Zone Stoichiometry
1,050
1.100
Figure 13-6
NOx Emissions for the Original System Parametric Tests and the Modified System June 1992
Long-Term Tests
13-8
-------
Long-Term Load Dispatch Testing
Rebum Zone Stolchiometry
Figure 13-7
Comparison of NOx Emissions at Different Loads
Difference Between Parametric and Long-Term Testing Conditions
Increased Target RZS
During parametric testing, the target RZS was 0.90. For long-term testing, it was raised to 0.94,
primarily to maintain the CO level under 200 ppm (the baseline CO was about 35 ppm.) The
variation of CO emissions with RZS is shown in Figure 13-8. It is apparent that CO increases
exponentially with decreasing RZS.
During parametric testing when operating conditions could be controlled so that the primary
air/fuel ratios were relatively constant, the RZS could be held close to 0.90 without incurring
excessive CO because the cyclone O2 varied less than it did under normal boiler operating
conditions. A RZS of 0.90 with only a small deviation from the target value was achievable
during parametric testing. However during long-term testing the primary air/fuel ratio varied
more widely. If the RZS for long-term testing had been targeted at 0.90, the minimum RZS
during load swings would likely have been 0.85, which would have caused excessive CO
emissions (1,000+ ppm). Raising the target RZS to 0.94 provided a margin of safety. With this
target RZS the probability of the RZS falling below 0.90 and consequently excessive CO
emissions was minimized, but at the expense of higher NOx emissions.
13-9
-------
Long-Term Load Dispatch Testing
RZS
Figure 13-8
Variation of CO Emissions with RZS at Full Load
Effect of Unsteady Operation on NO and CO Emissions
As shown in Figures 13-7 and 13-8, NOx emissions vary essentially linearly with RZS while CO
emissions increase exponentially with decreasing RZS. Though NOx emissions would be nearly
identical under steady vs. unsteady operation (assuming the average RZS stayed the same) the
CO emissions would be substantially higher. For example, at a steady 2.0% cyclone 02 for one
hour, the NOx and CO emissions would be 330 and 50 ppm, respectively. If the unit were to then
operate with the cyclone O2 swinging between 1.4 and 2.6%, the average cyclone O2 for the hour
would still be 2.0%, and the average NOx would also be unchanged at 330 ppm, but the average
CO would be much higher at 215 ppm, a four-fold increase.
Variations in air/fuel ratio between cyclones can also lead to radical variations in CO
concentrations. Since the available measurements only showed the overall reburn zone
stoichiometry, there was no way to identify whether the air/fuel ratios of the individual cyclones
were uniform or highly divergent. Figure 13-8 shows variations in CO concentrations during
long-term tests in March and April 1992 from less than 25 to more than 925 ppm for rebum zone
stoichiometry between .90 and .95. This range of CO concentrations indicates that wide
variations in air/fuel ratios to the cyclones could have existed during the long-term testing.
13-10
-------
Long-Term Load Dispatch Testing
Increased Variance of RZS About Target Value
During long-term testing the rebum system controller modulated reburn fuel flow according to
coal flow. If primary air/fuel ratios had been held constant, then the RZS would have been
relatively constant. However, under normal boiler operating conditions, considerable variation
in primary air/fuel ratio did occur. Older boilers, which typically have older control systems,
will generally show more variation than a newer boiler which is equipped with a more modem
control system. At Niles the control system was replaced and upgraded in 1987; however, the
sensors and drivers for some of the components, damper drives for example, are still original
equipment. Most of the variation in RZS occurred because of changes in cyclone exit O2.
Rebum fuel input, by comparison, was held relatively constant between 16 and 18 percent of the
heat input basis.
Commercial Potential of Gas Reburn for NOx Control
Effect of Boiler Load
As noted earlier (Figure 13-7), the percent NOx reduction decreased with decreasing load even
with the same percentage of reburn fuel. However, the baseline NOx also decreased with
decreasing load; the result being that the quantity of NOx produced on an absolute basis stayed
relatively constant throughout the range where rebum could be employed, see Figure 13-9. The
rebum system was designed to operate at a constant 16% rebum fuel input down to 80 MW;
from 80 MW to 65 MW the reburn fuel was designed to be ramped from 16% to 0%. Because of
potential adverse effects on slag tapping, it was the operator's judgment to not operate the rebum
system below 80 MW; Figure 13-9 reflects this by showing no NOx reduction at loads of 75 MW
and lower. More favorable coal properties (lower ash fusibility temperatures) could have
facilitated rebum system operation down to the 65 MW design point. Long-term, cumulative
NOx emissions with rebuming would be a weighted average of the NOx produced at the actual
loads experienced during normal boiler operation. As inferred from Figure 13-9, NOx emissions
are a function of the boiler duty and therefore base loaded units can realize lower NOx emissions
than peaking units. However, even if rebuming cannot be used below some critical load, the
overall effect is a levelization of NOx emissions on an absolute (tons/hr) basis, throughout the
entire load range.
13-11
-------
Long-Term Load Dispatch Testing
0.6
Load (MW gross)
Figure 13-9
Reburn Effectiveness at Niles Unit No. 1 for Different Loads
Suitability of Gas Reburn for Seasonal NOx Control
Gas reburning may have good potential to a utility company on a seasonal basis. The price of
natural gas is typically lowest in the summer. During summer, ambient ozone concentration
(which is regulated under Title I of the 1990 Clean Air Act Amendments) tends to peak because
of the long duration of sunlight available to promote ozone formation reactions. NOx is a
precursor to ozone formation, and controlling NOx emissions has been demonstrated to be a
necessary part of reducing ozone concentration for most areas that are in non-attainment. During
the summer months, most units in a utility system are operated at close to maximum capacity, the
load at which the reburn process has been demonstrated to be most effective. Since most of the
cost of reburn system operation is the fuel cost differential between natural gas and coal, the
operating cost differential is at its minimum during the summer months. The combination of
maximum effectiveness and minimum operating cost for reburn system operation during summer
suggests that natural gas reburning is an ideal candidate technology for seasonal NOx control.
Also, the creation of emission allowances, through the substitution of gas (which contains little
or no sulfur) for coal, which almost always contains sulfur, is likely to add further justification to
reburning if the gas price is low enough.
13-12
-------
Long-Term Load Dispatch Testing
Utility Operator's Assessment of N0X Reduction by the Reburn Process
Operation of the system was fairly simple for the plant operators. The automated control system
and interface with the main boiler controls allowed for a nearly invisible system for the
operators. The reburn system operation put more heat into the superheat and reheat sections.
This increased attemperator flow rates by about 1 to 5%. The resulting effects on boiler
efficiency have been discussed in Section 12.
The overall XOx reductions which gas reburning can achieve on a long-term basis depends on
how the unit is loaded. The greatest NOx reductions observed at Niles occurred only when the
unit was operating at, or near, maximum load. At low load conditions (< 80% MCR), no NOx
reductions were achieved because the operators turned off the gas in order to keep the slag
running in the cyclones and primary furnace.
The results presented elsewhere in this report demonstrate the need for additional research of gas
rebuming at other locations before any federal or state regulations are developed based on
rebuming as a long-term NOx control technology for cyclone boilers. The Niles long-term
testing began March 2, 1992, and ended June 19,1992, a duration of only 3 1/2 months. During
long-term testing, the Niles unit operated within a load range of 114 MW maximum down to 65
MW minimum. When the generating unit was operating above 80 MW during the 3 1/2 month
test, the reburn system was not operated for roughly 50 percent of the time for a variety of
reasons, as follows:
Operator Judgment - Though the reburn system was programmed to operated down to 65 MW,
because of potential adverse effects on slag tapping, operators turned off the reburn system
below 80 MW.
Furnace Casing Leak - The occurrence of a casing leak in the reburn zone required the reburn
system to be shut off until the leak was repaired.
Reburn Fuel Guidepipe Leak - Water leaks occurred in several of the guide pipes necessitating
shutting off the reburn fuel to the affected injectors.
Instrument Air Compressor Failure - Loss of compressed air for instruments and controls caused
the reburn system to be turned off until the compressor was repaired.
Reburn Fuel Gas Control Valve Failure - Inability to accurately control natural gas flow rates
caused the reburn system to be turned off until the valve was repaired.
The last three problems listed above should be resolvable with engineering changes. Potential
solutions to the casing leak problem have been suggested. These will be discussed in Section 15;
however, no solution has yet to be demonstrated. Finally, as discussed above, the minimum load
at which the reburn system can be operated without incurring a slag tapping problem depends on
cyclone exit temperatures and stoichiometry, and on the ash fusibility characteristics of the coal.
13-13
-------
14
BOILER TUBE THICKNESS MONITORING PROGRAM
Description of the Program
During the planning for the reburn test program, Ohio Edison and Combustion Engineering
recognized the need to monitor degradation of boiler tubing during the testing. The possibility of
tube degradation existed because the reburn process altered the heat flux pattern within the
furnace and produced substoichiomctric (reducing) fuel/air gas mixtures downstream of the
cyclones. To address the possibility of boiler tube degradation, a comprehensive boiler tube
monitoring program was developed. The program included both non-destructive and destructive
testing techniques to assess the possibility of corrosion on the waterwall tubes caused by the
reducing atmosphere and long-term overheating and coal-ash corrosion damage on the
superheater and rcheater sections. The waterwall fireside corrosion was evaluated by ultrasonic
thickness testing and corrosion probe monitoring. The assessment of superheater and reheater
tube damage was performed by corrosion probe monitoring, remaining life evaluation by
ultrasonic testing, and tube sample removal.
Testing Sequence
A baseline inspection of the unit with ultrasonic tube thickness measurements was performed in
June 1990 during the installation of the reburn system. The first injection of natural gas into the
unit took place on August 29, 1990. The parametric testing with the original reburn system took
place during September through December 1990. A short outage at the end of December 1990
provided opportunity to obtain ultrasonic thickness data and a visual inspection for waterwall
tube wastage. In October 1991, installation of the modified reburn system was completed. At
that time another set of ultrasonic testing data was obtained and corrosion probes were removed.
In the months following October 1991, several tests, including parametric, long-term, and water
injection tests, were performed. In August of 1992 during the Unit 1 outage, an additional set of
ultrasonic measurements was made and data was obtained for remaining tube life analysis.
Testing Locations
Measurements were made at several elevations of the lower furnace waterwalls, superheater
sections, and reheat superheat section. Waterwall ultrasonic measurements were made on every
third waterwall tube and target wall tube in the furnace at elevations of 914', 902', 896', 890', and
14-1
-------
Boiler Tube Thickness Monitoring Program
880'. Three readings per tube (left, center and right) were obtained. An additional strip was
measured along the back wall upper bend at elevation 909'.
Because of the higher gas temperatures in the convection pass, ultrasonic thickness readings
were also obtained on the horizontal reheater and superheater sections. Three readings per
element were obtained for the horizontal reheater and each of the five stages of the secondary
superheater. Internal oxide scale measurements were also obtained during the baseline testing
and during the August 1992 outage.
In addition to the ultrasonic non-destructive testing, eight vertical temperature-controlled
corrosion probes were installed through waterwall openings to measure the corrosion rates. Four
probes were located at elevation 891'—two on the rear wall and one on each of the side walls.
Two probes were located on side walls at elevation 904', and two probes were located on side
walls at elevation 928'. The design of the vertical waterwall corrosion test probes is shown in
Figure 14-1. A horizontal corrosion probe was installed between the fourth and fifth stages of
the secondary superheater. The design of the horizontal corrosion test probe is shown in Figure
14-2. The waterwall probes were constructed of three 2" diameter test specimens of the
following materials: SA192 carbon steel, SA213-T22 and SA213-TP304 stainless steel. In
addition to these same materials, SA213-T11, T-91, and TP310 stainless steel were used in the
corrosion probes for the superheater and reheater sections. The corrosion probe locations are
shown in Figure 14.3.
Instrumentation
Ultrasonic thickness readings were obtained by Combustion Engineering's subsidiary, ABB AM
Data, Inc. Thickness readings were taken by using a Kraut-Kramer Branson USK7 flaw detector
with a contoured, dual-element 5MHz probe. Calibration was performed on a machined tube
with known wall thickness. The calibration was checked after each set of readings. The tube
surface was prepared by sandblasting to white metal. The couplant was a cellulose-gel type.
The remaining life analysis of the superheater and reheater sections was performed by using an
oscilloscope and pulser receiver. The pulser receiver was a Panametries Model No. TRX5052
(75 megahertz), and the oscilloscope was a Textronics Model No. 2246. The transducer was a
single element delay line with a frequency range between 15 and 30 megahertz.
Ultrasonic Tube Thickness Test Results
Ultrasonic thickness (UT) readings were obtained on four different occasions over a 20-month
period. The UT data are presented in Appendix A. The UT test results of the waterwall tubes
are inconclusive and could not be used to determine a corrosion rate. In fact, close examination
of plotted data revealed that many of the tubes gained wall thickness. The error in data may be
explained by several factors; the equipment, the technique, and the variation in location and
14-2
-------
Boiler Tube Thickness Monitoring Program
OUTER TUBE
0,200* WALL
SA-192
BRASS ROD
TO SPIRAL
COOLING
AIR
SIDE VIEW
Figure 14-1
Vertical Corrosion Test Probe
14-3
-------
Boiler Tube Thickness Monitoring Program
THERMOCOUPLE
PARTS
AIR
|l || 1
FURN.
WALL
T?C
TC
I
0
U
M
M
Y
TC
€
TC
3
2
TC
D
U
M
M
Y
~
CYLINDRICAL TEST SPECIMEN
Figure 14-2
Horizontal Corrosion Test Probe
14-4
-------
Boiler Tube Thickness Monitoring Program
SBmsUT/XEMUT
CON VICT] VIPASSES
GAS
JtlClSCOLA TING
fAJt
CYCLONES
Figure 14-3
Corrosion Probe Locations
Eight vertical corrosion test probe locations are identified by Items 1 through 8. The horizontal
corrosion test probe location is identified by Item 9.
14-5
-------
Boiler Tube Thickness Monitoring Program
calibration. More definitive insight concerning the effect of reburn on waterwall tubes was
provided by visual inspection. During the December 1990 outage an inspection was performed
on the waterwall tube surfaces. The inspection report is given in Appendix B. The examination
revealed that the tube surface appeared to be unaffected by reducing atmosphere corrosion.
UT measurements of the superheater and reheater sections showed areas of erosion/corrosion.
Although subject to the same errors mentioned above, the wall loss was significantly more
pronounced. Ultrasonic thickness measurements of the superheater and reheater sections,
following operation of the original reburn system, showed areas with an approximate 10% wall
loss, with wastage in areas of the fifth stage superheater as high as 0.100" between the June 1990
measurements before the initiation of reburn operation and October 1991 which was before the
initiation of testing of the modified system. Indicated tube loss is thought to be from a
combination of erosion and corrosion. Tube wall thickness changes during testing of the
modified reburn system (shown by measurements in October 1991 and August 1992) was
significantly less and in several instances measurements of the superheater and reheater sections
showed an inconsistent data pattern with smaller wall thicknesses in October 1991 than August
1992.
Because tube wastage was not uniform, it is believed that erosion was the larger contributing
factor between erosion and corrosion. The reduced tube wastage during operation of the
modified reburn system (without FGR) is explained by the fact that flue gas mass
flows/velocities during modified reburn system operation were returned to base-case levels and
in this way wastage due to erosion was minimized.
Remaining Tube Life Analysis Using Oxide Scale Measurements
The superheater and reheater sections were inspected for oxide scale and wall thickness before
and after the reburn project. The oxide scale thickness is representative of the operating
temperature which when combined with the time of operation can be correlated to a Larson-
Miller parameter. The dimensions of the tubing along with wall thickness are used to calculate
the mean diameter stress. Remaining life is predicted based on a linear oxide scale growth and
wall loss rate.
After reviewing the oxide scale data it was found that the results of the initial, baseline inspection
gave remaining life values lower than the final inspection values. The oxide scale readings
obtained during the initial inspection were always assumed to be at least 0.006", thus producing a
lower remaining life value. When the measurements were made during August of 1992, a new
technology allowed the technician to measure scale thicknesses below 0.006". Since a valid
comparison of remaining life before the initiation of reburn testing and after the completion of
reburn testing was not possible, the oxide scale tube life analysis could not be used to evaluate
the effect of gas reburn on the superheater or reheater tube life.
14-6
-------
Boiler Tube Thickness Monitoring Program
Corrosion Probe Tests
The corrosion probes were installed for the parametric testing conducted between January and
October 1991. The results of the corrosion evaluation are attached in Appendix C. The
corrosion probe analysis revealed that virtually no corrosion had occurred on the materials on
most of the probes at all locations. Two of the probes however did indicate severe corrosion
rates which were attributable to loss of probe cooling air.
Conclusions
A rebum system was installed in Niles Unit 1 at the end of June, 1990. Prior to the installation
Ohio Edison and Combustion Engineering developed a series of tests to evaluate corrosion
damage to the waterwall, superheater, and reheat superheater tubing. The key findings are as
follows:
• The ultrasonic thickness testing in the waterwall sections was inconclusive. Changes in tube
wall thickness were below the threshold of sensitivity of the UT measurement technique.
However, visual inspection of the waterwalls during the December 1990 outage revealed that
the tube surface appeared to be unaffected by reducing atmosphere corrosion.
• Ultrasonic thickness measurements of the superheater and reheater sections following
operation of the original rebum system showed areas with an approximate 10% wall loss,
with wastage in areas of the fifth stage superheater as high as 0.100". Tube wall thickness
changes were significantly less during testing of the modified reburn system. The reduced
tube wastage during operation of the modified reburn system (without FGR) is explained by
the return of flue gas mass flows/velocities to baseline levels during modified reburn system
operation, thereby minimizing wastage due to erosion. Because tube wastage was not
uniform, it is believed that erosion was the larger contributing factor between erosion and
corrosion.
• The remaining superheater/reheater tube life analyses performed before and after the reburn
project were inconclusive concerning any degradation due to high temperature oxidation.
Final inspection values gave higher remaining tube life values than did initially obtained
values.
• Corrosion probe tests showed very low corrosion rates. In two instances when corrosion
rates were high, the wastage was attributed to loss of cooling air to the probe.
14-7
-------
-------
15
APPLICATION OF REBURNING TO PRESSURIZED
FURNACES
Background
The application of reburn technology to pressurized furnaces such as Niles Unit No. 1 can create
unfavorable situations if a leak develops in the casing surrounding the furnace in the vicinity of
the reburn zone because the reburn process generates a fuel-rich gas mixture for converting NOx
into N2. Furnace gases usually are a mixture of the normal combustion products: carbon
dioxide, water vapor, nitrogen, and a slight concentration of oxygen. However, during the reburn
process the fuel-rich combustion products in the rebum zone contain carbon monoxide, a toxic
gas. In addition, some mixtures of fuel-rich combustion products can be in the flammabiiity
range, depending on the proximity of the leak to the reburn fuel injectors, and therefore can
create a hazard. During long- term testing on April 16,1992, five "flameletts" about two feet
long were observed attached to a comer of the furnace at the elevation of the rebum zone.
Apparently, a furnace gas leak occurred in the reburn zone and gases made their way through the
casing to the atmosphere where, with sufficient oxygen and being combustible gases, they
proceeded to burn. There was no indication that reburning had caused the casing leak; as a
matter of fact, leaks can occur during normal furnace operation of pressurized units. However
leakage of combustible gas creates a different situation than leakage of normal products of
combustion.
Resolution of the Problem
Discussions were held with project personnel, sponsors, and consultants to identify and resolve
the leakage problem. The issue was addressed in two categories: (1) what to do to assure safety
during the long- term reburn tests, and (2) what to do for an acceptable solution for commercial
application of reburn technology to pressurized furnaces.
Regarding category (1), it was decided that the most reasonable approach was to find and repair
the leak and institute a monitoring plan that would allow early detection of any new gas leaks
that might occur. The casing leak was found and repaired. The monitoring plan incorporated the
use of a portable hand-held gas analyzer which the boiler operators carried and used throughout
the plant during normal once-per-shift walkdowns of the unit. Long term reburn testing was
continued to completion.
Regarding category (2), several possible commercial solutions were suggested. A number of
options were considered: (1) convert pressurized units to balanced draft by adding an induced
draft fan and associated equipment, (2) convert tangent tube pressurized units such as Niles No. 1
to fusion welded walls by adding fusion welds between the tubes, (3) erect an enclosure around
15-1
-------
Application of Returning to Pressurized Furnaces
the rcburn zone which would operate at a slightly higher positive pressure than the furnace
pressure to assure that any leakage would be into the furnace and (4) erect a "hood-like"
structure around the upper part of the furnace so that gas composition could be constantly
monitored for possible changes. It is unlikely that options (1) and (2) could be economically
justified, given the remaining life of most cyclone units and the existence of competing
technologies. However, options (3) and (4) would be much less capital-intensive and could be
configured to ensure safe rebum system operation,
It should be noted that "commercial" resolution of this unanticipated problem was beyond the
program workscope. Indeed, discovery of this problem is an excellent example of why R&D
demonstration programs are conducted. The preferred selection between these alternatives
depends upon site-specific technical as well as economic considerations and can therefore only
be decided by a detailed technical and economic analysis.
15-2
-------
16
REBURN SYSTEM ECONOMICS
Introduction
There are currently about 100 cyclone-fired boilers operating in the United States. These units
range in size from about 15 to 1150 MW and were commissioned between the late 1940's
through 1981. Baseline NO* emissions from these cyclone units range from 600 to 2000 ppm
corrected to 3% O2 (0.85 to 2.7 lb NOx/mmBtu). These emissions can be reduced by 50 to 70%
using natural gas reburning. Since cyclone boilers do not employ burners in the conventional
sense, reburning is the only viable in-fumace NOx reduction technology that has been proposed
for NOx reduction for cyclone units. Other technology options include staged combustion and
post-combustion NOx control systems (selective non-catalytic reduction, selective catalytic
reduction, SNOX). Staged combustion is unacceptable for retrofit of cyclone furnaces due to the
potential for unburned carbon, increased cyclone watertube corrosion, and slag tapping problems.
The choice between reburning and the post-combustion technologies is driven by cost (dollars
per ton of NOx removed) as well as the impact of the technologies on boiler availability,
reliability and performance.
The natural gas reburning demonstration at Niles resulted in practical design and operating
experience that can be applied to other cyclone-fired boilers. In addition to the Niles
demonstration, there have been two other reburning demonstrations on cyclone boilers; one of
them employing natural gas and the other coal as the reburning fuel (Farzen. et al. (1993);
Folsam et al. (1995)). However, to apply reburn technology commercially, the process must not
only be technically feasible but also economically viable to be chosen over post-combustion
processes for NOx control. A study was conducted to evaluate reburning from an economic
perspective as a NOx reduction technology for the entire cyclone boiler population using the
Niles experience as the basis. The Niles results were applied to five other cyclone boilers
covering a range of sizes, ages, furnace configurations, cyclone arrangements, and megawatt
ratings. This section summarizes the findings of the study and reaches conclusions for the
technical and economic viability of natural gas reburning for cyclone boilers.
16-1
-------
Reburn System Economics
Basis for Study
Cyclone Boiler Population Application Criteria
The first criterion for applying natural gas reburning to cyelone-flred boilers is age of the unit.
The project team determined that a boiler should have at least 10 years of operation remaining in
its expected lifetime to justify the capital investment for any combustion modification or post-
combustion equipment. If it is assumed that 50 years is a reasonable boiler lifet ime from start-up
date and that the retrofit of reburning equipment will be completed in 1997, then the boilers that
are "eligible" for reburning retrofits are listed in Table 16-1.
Table 16-1
89 Cyclone Boilers with 1957 and Later Start-up Dates
Utility/Station
B&W
Contract No.
MW
Rating
Start-up
Date
Fuel
Type
Eastman Kodak, Rochester, NY
RB-230
-60
1957
bit
AEP/Ohio Power, Muskingum river #3
RB-248
225
1958
bit
Tampa Electric, Gannon #1
RB-254
105
1957
bit
International Paper, Mobil #1, #2
RB-255
-70 each
1957
bit
Jersey Central P&L, Sayreville
RB-256
133
1960
bit
AEP/Columbus & Southern, Conesville #1
RB-265
136
1958
bit
AEP/Ohio Power, Muskingum River #4
RB-268
225
1958
bit
Consolidated Water & Power, Biron
RB-274
16
1957
bit
AEP/Ohio Power, Kammer #1, #2
RB-280
225 each
1961
bit
TVA, Allen #1, 2, 3
RB-289
330 each
1964
bit
Tampa Electric, Gannon #2
RB-290
115
1959
bit
International Paper, Pine Bluff #1,2
RB-291
-70 each
1958
bit
Detroit Edison, St. Clair #5
RB-292
325
1960
bit
Rhinelander Paper, St. Regis
RB-296
-30
1959
bit
AEP/Ohio Power, Kammer #3
RB-297
225
1961
bit
Atlantic City Electric, Deepwater #1
RB-299
79
1960
bit
AEP/Columbus & Southern OH, Conesville #2
RB-303
136
1959
bit
Arkansas P&L, Ritchie #1
RB-305
356
1961
bit
Commonwealth Edison, Joliet #6
RB-311
360
1960
bit
(Continued)
16-2
-------
Return System Economics
Table 16-1 (Continued)
89 Cyclone Boilers with 1957 and Later
Start-up Dates
Utility/Station
B&W
Contract No.
MW
Rating
Start-up
Date
Fuel
Type
Wisconsin P&L. Nelson Dewey #1
RB 312
100
1960
bit
United Illuminating, Bridgeport Harbor #1
RB-320
75
1962
bit
Missouri Public Service, Sibley #1
RB-327
50
1960
bit
W. VA Pulp & Paper, Luke MD
RB-331
-60
1960
bit
Public Service of NH, Merrimac #1
RB-337
114
1961
bit
Nebraska Public Power, Sheldon #1
RB-338
105
1961
bit
Monongahela Power, Willow Island #2
RB-342
165
1961
bit
Tampa Electric, Gannon #3
RB-346
160
1960
bit
Missouri Public Service, Sibley #2
RB-347
50
1963
bit
Central Electric Power, Chamois #2
RB-348
48
1961
bit
Iowa Electric, Sutherland
RB-353
75
1962
bit
Kansas City BPU, Kaw #3
RB-359
66
1963
bit
Tampa Electric, Gannon #3 and 4
RB-361
180
1964
bit
Commonwealth Edison, State Line #4
RB-365
389
1963
bit
Baltimore G&E, Crane #2
RB-366
191
1963
bit
Atlantic City Electric, B.L, England #1
RB-368
125
1963
bit
Wisconsin P&L, Nelson Dewey #2
RB-369
100
1962
bit
NIPSCO, Bailly #7
RB-372
194
1964
bit
Iowa Public Service, Neal #1
RB-377
147
1964
bit
Owensboro Municipal Utility, Smith #1
RB-386
150
1965
bit
Northern States Power, Riverside #8
RB-390
228
1964
bit
Atlantic City Electric, B.L. England #2
RB-409
150
1965
bit
Kansas City BPU, Quindaro #3
RB-421
75
1968
bit
Associated Electric Coop., Hill #1
RB-427
175
1970
lig
St, Joseph P&L, Lake Road #1
RB-430
75
1969
bit
(Continued)
16-3
-------
Keburn System Economics
Table 16-1 (Continued)
89 Cyclone Boilers with 1957 and Later
Start-up Dates
Utility/Station
B&W
Contract No.
MW
Rating
Start-up
Date
Fuel
Type
Associated Electric Coop,, Hill #2
RB-434
270
1969
lig
Nebraska Public Power, Sheldon #2
RB-438
120
1968
bit
Wisconsin P&L, Edgewat-er #4
RB-442
330
1969
bit
Empire District Electric, Asbury #1
RB-447
200
1970
bit
Minnkota Power, Young #1
RB-457
235
1970
lig
Associated Electric Coop., New Madrid #1
RB-466
580
1973
lig
Associated Electric Coop., New Madrid #2
RB-483
600
1977
lig
Basin Electric, Leland Olds #2
RB-489
400
1974
lig
Otter Tail Power, et al, Big Stone #1
RB-490
400
1974
%
Southern Illinois Power Coop., Unit 4
RB-560
175
1978
%
Otter Tail Power, Coyote #1
RB-563
456
1981
lig
Southern Illinois Power Coop., Unit 5
RB-589
350
1980
lig
AEP/Ohio Power, Philo #6
UP-1
125
1957
bit
AEP/Ind. & Mich. Elec., Breed #1
UP-2
450
1960
bit
Baltimore G&E, Crane #1
UP-6
190
1961
bit
AEP/Ind & Mich. Electric, Tanners Creek #4
UP-9
580
1964
bit
TV A, Paradise #1
UP-10
704
1963
bit
TV A, Paradise #2
UP-11
704
1963
bit
Public Service E&G, Hudson #1
UP-12
420
1964
bit
Hartford Electric, Middletown #3
UP-16
240
1964
bit
CIPS, Coffeen #1
UP-18
365
1965
bit
Union Electric, Sioux #1
UP-19
489
1967
bit
Union Electric, Sioux #2
UP-20
489
1968
bit
NIPSCO, Bailly #8
UP-29
422
1968
bit
Commonwealth Edison, Kincaid #1,2
UP-30
660 each
1967
bit
Northern States Power, King #1
UP-36
574
1968
sub
(Continued)
16-4
-------
Rehurn System Economics
Table 16-1 (Continued)
89 Cyclone Boilers with 1957 and Later
Start-up Dates
Utility/Station
B&W
Contract No.
MW
Rating
Start-up
Date
Fuel
Type
Public Service of NH, Merrimac #2
UP-42
350
1968
bit
Missouri Public Service, Sibley #3
UP-45
419
1968
bit
TV A, Paradise #3
UP-49
1150
1969
bit
Illinois Power, Baldwin #1
UP-61
605
1970
bit
NIPSCO, Michigan City #12
UP-76
500
1974
bit
CIPS, Coffeen #2
UP-82
600
1972
bit
Illinois Power, Baldwin #2
UP-8 3
600
1973
bit
Commonwealth Edison, Powerton 35-1, 35-2
UP-89
430 each
1972
sub
Kansas City P&L/KG&E, La Cygne #1
UP-90
844
1973
sub
Commonwealth Edison, Powerton 6-1, 6-2
UP-103
430 each
1975
sub
NIPSCO, Schahfer #1
UP-112
520
1976
bit
Total Number 89
About 160 cyclone boilers have been built in the United States. Table 16-1 lists 89 units with
1957 or later start-up dates; this represents the number of retrofittable units according to the
earlier guidelines assuming a 50-year life and having 10 or more years of useful life remaining.
Out of this list, one unit has already been retired, one unit already has a reburning system, and
four have been targeted for SNCR retrofits. Therefore, the reburning retrofit candidates are
reduced to 83.
These boilers can be classified further by megawatt rating, main furnace configuration, and
cyclone configuration. Older units like Niles often fire the cyclones into a primary furnace to
maximize slag rejection. Slag droplets entrained in the cyclone exit gas are impinged against the
target wall of the primary furnace. The gases must pass below the target wall, through a bank of
screen tubes, then upward through the main furnace to the furnace exit. Such units are often
short and wide, and present challenges to the reburning system designer for placement of fuel
and air injectors where adequate mixing rates and reaction times are available.
Newer units were usually designed with open furnaces and with cyclones mounted on single or
opposed walls. Boilers with open furnaces are usually tall and thin, and their width is
determined by the number of cyclones that must be accommodated. Figure 16-1 from Steam. Its
Generation and Use (1992) illustrates the different cyclone furnace arrangements.
16-5
-------
Return System Economics
(a)
Screened Furnace
Arrangement
Single Wall
(b)
Open Furnace
Arrangement
Single Wall
(c)
Open Furnace
Arrangement
Double Wall
Figure 16-1
Firing Arrangements Used for Cyclone Furnaces (from Steam. Its Generation and Use.® Babcock
and Wilcox, 1992, Reproduced with permission)
The boilers chosen for further study allowed the following comparisons:
primary furnace versus open furnace (one-wall-fired)
one-wall versus opposed wall firing
one versus two cyclone furnace elevations
pressurized versus balanced draft operation
designs from the 1950's versus 1960's.
Table 16-2 lists the boilers (anonymously). Additional description of each unit is also provided
in this chapter.
Reburn System Design Criteria
The criteria used for the design of the Niles rebuming system were confirmed during the
demonstration tests. Those criteria are listed in Table 16-3. Changes to the commercial reburn
system design resulting from the test program are discussed below.
One of the significant findings of the testing at Niles was that effective penetration and mixing of
natural gas reburn fuel was achieved without the use of flue gas recirculation (PGR). The
elimination of FOR was considered sufficiently important both from an operational and an
economic standpoint that reburn systems employing direct injection of natural gas were used as
16-6
-------
Reburn System Economics
Table 16-2
Study Boilers
Unit
Start-up
Year
Gross MW
Rating
No. of
Cyclones
Cyclone
Arrangement
Draft
Furnace
Type
Furnace
Width, Ft.
Furnace
Depth, Ft.
*Furnaee
Height, Ft.
Normal
Residence Time, s
Niles
Unit #1
1953
115
4
2 over 2
Front Wall
Press.
Primary
36
13
~43
1.6**
Unit A
1969
75
2
Front Wall
Press.
Open
24
14
38.7
0.9*
UnitB
(Decom.)
1957
125
3
Front Wall
Press.
Primary
34
10.5
19.2
0.5**
UnitC
1958
225
5
2 over 3
Front Wall
Press.
Primary
48
16
20.5
0.8**
Unit D
1968
420
8
2 over 2
Opposed Wall
Bal
Open
36
27
90.0
1.1*
UnitE
1970
605
14
3 over 4
Opposed Wall
Bal
Open
60
33
116.8
2.0*
* ^ of top cyclones to furnace arch
** Main furnace only, screen tubes to furnace arch
-------
Return System Economics
Table 16-3
Reburn Design Criteria for Niles
•
Cyclones operate at SR of 1.1; 50% cyclone turndown
•
Inject reburn gas as close as possible to cyclones (T ~ 2700 F)
•
No FGR for furnace depth less than 17 ft (34 ft for opposed-fired)
•
Reburn zone S.R. = 0.90 at full load
•
Reburn zone residence time (nominal) = 0.6 s at full load
- minimum of 0.3 s
- maximum of 0.8 s
•
Burnout zone residence time (nominal) - 0.7 s at full load
- minimum of 0.5 s
the basis for the economic evaluation. At Niles, the furnace depth at the point of natural gas
injection was 13 ft. The gas jets, injected at sonic velocity, were observed to reach the opposite
wall. It was estimated that the jets could have penetrated several feet further had the furnace been
deeper; it was estimated that no FGR would be required as long as the boiler depth is less than
17 ft for one-wall-fired boilers or 34 ft for opposed-fired boilers with reburn nozzles installed on
both the front and rear walls. Since the largest opposed-fired cyclone boiler ever built (TVA,
Paradise Unit #3 - 1150 MW) is only 33-ft deep (by 96-ft wide), it is concluded that FGR would
not be required on any cyclone-fired boiler using natural gas as the rebum fuel.
Location of reburn fuel and burnout air injectors will be based on residence time available for
mixing, and affected by structural interferences that could prevent the ideal location from being
chosen. Reburn fuel injectors should be placed as close to the cyclone outlets as possible since
higher temperatures drive the NOx reduction reactions. Side spacing between rebum fuel
injectors should be close (4 to 8 ft) to assure that natural gas rapidly contacts the products of
combustion from the cyclones. Vertical distance between fuel injectors and burnout air injectors
should provide enough time for mixing and NOx reduction to take place. Theoretically, burnout
air injector locations should be chosen to assure burnout of gaseous hydrocarbon fragments
remaining after partial combustion of the reburning gas. Reaction should be rapid, so mixing
rates will dominate. However, in commercial reburn systems, the additional air must also burn
any carbon carryover from the cyclones that would normally have burned where the reburning
zone has been located.
Reburn System Design and Economics
Five boilers were chosen from the cyclone boiler population to be subjects of a technical and
economic assessment. For each unit, a general arrangement of reburning equipment was
prepared, NO* emissions before and after reburning were estimated, and rough capital and
operating costs were scaled from the Niles experience. The EPRI Technical Assessment Guide,
EPRI (1989), was used for the economic estimates, except a detailed breakdown of process
capital was beyond the limited scope of this assessment.
Specifically, the capital cost of reburning at Niles was taken from Bono et al. (1991) with
adjustments in capital costs for the change from the original reburn system to the modified
system, adjustments due to productivity gains, and adjustments for price escalation between 1991
16-8
-------
Reburn System Economics
and 1995. The changes to the modified reburn system included elimination of the flue gas
recirculation fan, motor, controls, and ductwork, simpler reburn fuel injectors, and simpler
modifications to the furnace water walls. It was estimated that these changes would reduce the
capital cost by S1.0M relative to the S4.22M capital cost presented by Borio et al. (1991) for the
original reburn system. The capital costs for this study are presented in 1995 dollars. It was
estimated that capital cost escalation between 1991 and 1995 would be balanced by cost savings
due to productivity gains between 1991 and 1995. Therefore the estimated process capital cost
forNiles in 1995 dollars is $3.22M.
To extrapolate Niles cost experience to different sized cyclone furnaces, the "factor" method was
used. In this method the Niles costs were scaled by the square root of the megawatt rating of
each unit studied as shown below:
, , f UnitMwV/2
Unit process capital = ($3.22 M)
\ Niles MW;
The factor method was appropriate for this study since the design of the reburn system
equipment for the other units would be similar to the Niles equipment and costs would be
expected to be analogous to unit sizes. The unit process capital calculated by the above equation
was applied in the EPRI TAG to derive capital and O&M costs.
The economic viability of reburning for Niles and each of the five study boilers is characterized
by the NOx removal cost effectiveness which is the cost, in dollars per ton of NOx removed. The
NOx removal cost effectiveness of natural gas reburning depends on these factors:
1. NOx removal efficiency over the operating range of the boiler
2. The load profile of the boiler
3. Minimum boiler load at which reburning can be applied
4. The differential in cost between coal and natural gas
The load profile of each individual boiler is an important economic variable in the calculation of
NO>: removal cost effectiveness because investment and fixed operating costs for NOx control
equipment are constant over the time period for amortization of the equipment, but the quantity
of NOx removed over this time period is dependent upon the load profile of the boiler. Load
profile is especially important for reburning in cyclone boilers for three reasons:
1. Turndown of individual cyclones is limited by the ability of each cyclone to maintain
molten slag from the spout to the slag tap at the bottom of the unit. Minimum load for
each cyclone is usually about 50% of full load heat input. When 18% of the heat input is
provided by the reburning fuel, the turndown range of each cyclone becomes 82% of its
former value.
16-9
-------
Return System Economics
2. Many cyclone boilers do not have the flexibility to remove cyclones from service to
achieve low load operation. Therefore, minimum cyclone load is often minimum boiler
load especially when the number of cyclones is small.
3. At decreased load, the NOx reduction usually decreases because the NOx entering the
reburn zone decreases, because the temperature in the rebum zone is lower (NOx
destruction slows down), and because cyclone outlet O2 increases (especially when
cyclones are taken out of service). An increase in cyclone outlet O2 at a constant
percentage of reburning gas results in a higher stoichiometric air-fuel ratio in the reburn
zone and less N(\ destruction.
The sensitivity of NOx reduction to load profile was explored during this study. The EPR1 TAG
method applies a 65% load factor to the full load NOx emission potential to determine the
amount of NOx removed per year. This simplifying assumption does not take into consideration
reduced NOx control effectiveness at intermediate loads or the need to turn off the reburning fuel
at very low load. Therefore, in addition to the EPRI TAG method, four load profiles were used
to evaluate the cost impact on each case study boiler:
1. A load profile derived from the Niles long-term demonstration data (65% load factor).
2. A high load profile representative of today's most competitive base loaded units (82%
load factor),
3. A typical base load profile for pulverized coal-fired boilers and larger cyclone-fired
boilers where more load flexibility is available (65% load factor).
4. An intermediate load profile for units with relatively high operating costs (49% load
factor).
Figure 16-2 shows the Niles load profile. Load data originally organized into 10 MW increments
(i.e. time at 100 to 110 MW, 90 to 100 MW, etc.) were merged into three categories: 100% load
(90 to 115 MW), 70% load (70 to 90 MW), and 50% load (40 to 70 MW). Time of day was
arbitrarily selected but does not factor into the calculation. Figure 16-3 shows the other load
profiles used in this study. Note that the Niles and the "typical" load profiles both produce the
same load factor (65%), but differ in the amount of off-peak time spent at very low loads.
16-10
-------
Reburn System Economics
100
10 12 14
Time of Day (Hr)
Figure 16-2
Niles Load Profile
16-11
-------
Reburn System Economics
100
80
60
40
20
High Load Factor
82.5% Load Factor
100
80
60
40
20
0
Typical Load Factor
= 65% Load Factor
100
80
60
40
20
0
0
—
— Intermediate Load Factor
= 49.2% Load Factor
i l 1 1
I
1
1 1 1 1 1
8 10 12 14 16 18 20 22 24
Time of Day (Hr) d-isto
Figure 16-3
Load Profiles for High Load Factor (Base-Loaded) Units, Typical Coal-Fired Units, and
Intermediate Load Factor Units
16-12
-------
Reburn System Economics
A summary of the rebuming system designs for Niles and the five boilers is given in Table 16-4,
The NOx removal cost effectiveness for Niles Unit No. 1 and each of the study units are
presented in the following subsections for a range of natural gas cost differentials. The costs are
calculated using the EPRI TAG method and for one or more of the four load profiles discussed
above, where applicable.
Niles Unit No. 1
The Niles reburning system design has been discussed elsewhere in this report, but is
summarized here for completeness. Niles is a pressurized unit rated at 115 gross MW. Four
cyclone furnaces arranged two over two (staggered) in the front wall fire into the primary
furnace. Gases are forced downward by the target wall through the screen tubes to the main
furnace, and then upward to the convective section. Figure 16-4 shows a sectional side view of
Niles No. 1 prior to rebuming retrofit.
To convert this unit to rebuming, furnace penetrations were added to allow natural gas and
bumout air injection. Five natural gas injectors spaced about 6 ft apart horizontally were added
to the rear wall of the main furnace at elevation 880 ft. Two burnout air ports were arranged
about 5 ft apart on both boiler sidewalls (total of four burnout air ports) at elevation 912 ft. The
air port arrangement was not optimal since each air jet had to quickly penetrate at least half way
across the 36-ft boiler width, but operation proved adequate as long as the cyclone combustors
were operating with low CO and unburned carbon.
16-13
-------
Re/rum System Economics
Table 16-4
Preliminary Design Summary
Gross
No. of
Pgr
No. of Reburn
Unit
MW
Cyclones
Required
Fuel Injectors
Niles #1
115
4
No
5
Unit A
75
2
No
5
Unit B
125
3
No
5
Unit C
225
5
No
7
Unit D
420
8
No
12
Unit E
605
14
No
14
Unit
Reburn Zone
Normalized
Residence Time
No. of Burnout Air
Injectors
Burnout Zone
Normalized
Residence Time
Estimated
Min. Load, %
Estimated Min,
Load w/Reburn. %
Niles #1
1.0
4
1.0
50
70
Unit A
0.77
5
0.88
50
67
Unit B
0,46
4
0.39
34
44
UnitC
0.46
4
0.55
30
40
Unit D
0.95
10
0.74
25
38
UnitE
1.18
14
1.72
29
49
Note: Normalized residence times are based on 1,0 for Niles Unit No. 1
16-14
-------
Reburn System Economics
REHEAT
OUTLETS
SECONDARY
SUPERHEATER
P3 OUTLET
SECONDARY
SUPERHEATER
INLET
CYCLONES
C-9138
Figure 16-4
Niles Unit No. 1 Ohio Edison Company
1645
-------
Return System Economics
The Niles boiler was a challenge for implementing natural gas reburning technology, but by no
means the worst case. The main furnace contained adequate residence time to meet design
criteria for NOx reduction in the reburn zone and carbon conversion in the burnout zone. As the
reader will see in the subsections that follow, other cyclone-fired boilers with primary furnace
designs are not so generous.
The economics of rebuming at Niles were summarized in previous ABB technical papers (e.g.
Borio et al. (1991)). Certain capital costs specific to Niles (FGR fan replacement, asbestos
abatement, an on-site natural gas pipeline, test ports for sampling and corrosion measurements)
added approximately $1,000,000 to the process capital for this project. Elimination of these
costs would reduce the total capital cost for a commercial reburning retrofit at Niles. Key
economic factors for Niles calculated using the EPRI TAG are the following:
process capital $3.22M
total plant investment $3.70 M
total capital required $3.91M
capital cost per kW $34/kW
10-yr. levelized busbar power charge 6.35 mills/kWh
cost effectiveness $2.200/ton NOx removed
10-yr. levelized cost (w/o fuel cost diff.) 1.73 mills/kWh
cost effectiveness (w/o fuel cost diff.) $600/ton NOx removed
Baseline NOx emissions measured at Niles were lower than those found at most cyclone-fired
boilers. The primary reason for low baseline NOx is a low heat input per cyclone compared to
most cyclone furnaces; this value is 30% less than the nominal design value for 10-ft cyclones
and 22% less than the nominal design heat input for 9-ft cyclones such as Niles.
The long-term test program showed that NOx could be reduced 55% at full load and 40% at 70%
load. Reburning was not used at loads below 70% due to concerns about slagtap freezing. This
unit, however, is operated between 50 and 70% load for a significant part of the day (especially
in the springtime when the long-term NOx data were obtained). Measured NOx results used to
calculate the cost effectiveness of reburning at Niles are as follows:
Load, % Baseline NOx, lb/mmBtu NOx Removed, %
100 0.93 55
70 0.87 40
50 N/A 0
For various load factors, the amount of NOx removed was calculated as follows:
NOx removed = (full load NOx) (% NOx reduction) (full load heat input)
(hours of full load operation) (0.9 availability)
+ (70% load NO,) (% NOx reduction) (70% load heat input)
(hours of 70% load operation) (0.9 availability)
16-16
-------
Reburn System Economics
+ (50% load NOx) (% NOx reduction) (50% load heat input)
(hours of 50% load operation) (0.9 availability)
Finally, cost effectiveness is the yearly levelized cost of the retrofit divided by the tons of NOx
removed per year.
Rebuming cost-effectiveness was much poorer using the Niles load profile. The resulting cost
was $5835 per ton of NOx removed at a natural gas cost $1.50/mmBtu higher than coal, and
$1592 per ton of NOx removed for equal coal and natural gas costs. The high NOx reduction cost
at Niles could be anticipated because it makes no economic sense to install a NOx control system
and then shut it off for more than half of the unit operating time.
Table 16-5 lists estimates of NOx removal cost effectiveness for Niles and the five study boilers.
Different load profiles were assumed for the study boilers, while the actual load profile was used
for Niles. The sections that follow detail the methodology used in deriving this table, and
discuss each case and the important factors that influence NOx removal cost effectiveness.
Table 16-5
Summary of Reburning Cost Effectiveness
Unit
Rebuming Cost-Effectiveness, S/T NOx Removed
EPRI TAG
High Load
Profile
Niles Load
Profile
Typical Load
Profile
Intermediate
Load Profile
Incl.
FCD
w/o
FCD
Incl.
FCD
w/o
FCD
Incl.
FCD
w/o
FCD
Incl.
FCD
w/o
FCD
Incl.
FCD
w/o
FCD
Niles
2200
600
N/A
N/A
5835
1592
N/A
N/A
N/A
N/A
Unit A
2069
686
2260
749
4596
1524
N/A
N/A
N/A
N/A
UnitB
2212
567
2011
515
2707
694
2994
767
5229
1340
UnitC
2057
409
1872
373
2520
502
2787
555
4868
970
Unit D
1242
182
1224
179
1889
276
1857
272
3526
516
Unit E
994
121
970
119
1470
180
1471
180
2768
338
Note: FCD = fuel cost differential
16-17
-------
Rebiirn System Economics
NOx Prediction Methodology
The cost-effectiveness of reburning at Niles No. 1 is based upon the results of long-term testing.
Baseline NOx emissions at various loads as well as controlled NOx levels were measured
directly. Baseline NOx emissions at Niles, however, are somewhat lower than those reported at
many cyclone-fired plants (Maringo, et al. (1987)). The following methodology was used to
estimate NOx emissions for Study Units A through E.
The high NOx emissions from cyclone boilers result from the high-temperature turbulent
combustion process in these units. Peak flame temperatures inside the cyclones and immediately
downstream depend on the amount of heat released in the cyclone (under near-adiabatic
conditions) and the rate of heat removal downstream of the cyclones. Recall that NOx formed
from nitrogen in the air (thermal NOx) increases exponentially with increasing temperature and
linearly with increasing time at that temperature.
Two boiler design factors and one fuel factor affect thermal NOx production in cyclone boilers.
The boiler design factors are:
1. Heat release in the cyclone
2. Heat removal rate downstream of the cyclones.
Cyclone heat release can be approximated by the cyclone capacity, MW/cyclone. Heat removal
rate is qualitatively related to the fraction of cyclones whose exit gases radiate directly to a
sidewall. For example, Figure 16-5 shows hypothetical examples of cyclone furnaces where the
cyclones are arranged in rows across the firing walls. The gases exiting the outer cyclones
radiate heat rapidly to the adjacent sidewall. The gases from the inner cyclones in the third
example are, however, somewhat shielded from the sidewalis, thus causing higher temperatures
to persist longer in the center of the furnace. In this example, 50% of the cyclones radiate to a
sidewall. If the same four cyclones were arranged two over two, as in the first example, all
cyclone exit gases would radiate to a sidewall and NOx emissions would be lower since the gas
temperature would be quenched more rapidly. Similarly, opposed-fired boilers, example two,
produce slightly higher NOx than single-wall fired boilers because the hot gases meet and create
a hot zone in the middle of the furnace. If the cyclones are offset, NOx is about the same as one-
wall furnace arrangements.
The fuel factor affecting NOx emissions is fuelbound moisture. Wet fuel depresses flame
temperatures and less NOx is formed. Lignites and most subbituminous coals contain more
bound moisture than bituminous coals. Lower NOx emissions have been measured in cyclone
fired boilers that burn low-rank coals.
Table 16-6 shows a comparison of the study units relative to the factors that affect baseline NOx
emissions. How these factors were evaluated in each case study is explained in the sections that
follow.
16-18
-------
Reburn System Economics
Top
Views
Front
Views
o
o
o
o
D-1932
Figure 16-5
Relative Effect of Cyclone Arrangement on NOx Emissions
The other quantity critical to calculating the NOx removal cost-effectiveness of reburning is the
percent NOx reduction to be expected. The following assumptions were made regarding NOx
reduction based on reburn process fundamentals:
1. The higher the temperature at the point of reburn fuel injection, the more rapid the NOx
reduction. Thus, the units with higher baseline NOx should achieve larger percentage
NOx reductions.
Table 16-6
Factors Affecting Baseline NO*
Unit
No. of
Firing Walls
MW/Cyclone
Cyclones
Adjacent to a
Sidewall, %
Fuel Rank
Est. Full Load
NOx,
lb/mmBtu
Niles
1
28.8
100
bituminous
0.93
A
1
37.5
100
subbituminous
1.2
B
1
41.7
67
bituminous
1.4
C
1
45.0
80
bituminous
1.4
D
2 (offset)
52.5
100
subbituminous
1.4
E
2(opposed)
43.2
57
bituminous
1.7
16-19
-------
Return System Economics
2. Increased residence time in either the rebum zone or the burnout zone would favorably
affect percentage NOx reduction if temperatures remained constant. At reduced load,
residence time increases but gas temperature in the reburn zone decreases. Therefore
NOx removal at reduced load will vary from unit to unit depending upon which factor
predominates.
3. The reburn zone stoichiometric ratio (RZS) is 0.90 at full load and 1,0 at minimum reburn
load. In between, the RZS varies linearly with load. Further, it is assumed that the
percent heat input from natural gas stays constant, and the variation in RZS is caused by a
gradual increase in combustion zone excess air.
Assumption 1 indicates that most boilers will achieve higher NOx reductions than Niles except
when residence times are limited. Units A and B had limited residence time available, so their
NOx reduction performance is not expected to be as good as Niles. The long-term tests at Niles
were run at a RZS of about 0.94 ± 0.02 over a load range of 70 to 100%. Better NOx reductions
would be expected in newer cyclone boilers where more accurate control of fuel and air flows to
each cyclone would allow operation at lower cyclone excess air levels, thus reducing foil load
RZS to 0.9.
Unit A
Unit A is located in the central part of the United States. It was designed to burn Illinois
bituminous coal when it started up in 1969, but recently converted to Powder River Basin
subbituminous. The unit produces about 75 MW at full load on either coal. Steam flow is
approximately 575,000 lb/h at 1950 psig at the superheater outlet. Steam temperature is 1005 F
leaving both the superheater and reheater.
Unit A represents the latest one-wall-fired cyclone design for bituminous coals. Only a few more
units were built between 1969 and 1971 when NOx emission requirements effectively made
cyclone boilers obsolete. This unit contains just two cyclones mounted on the front wall. The
cyclones are located 12 ft 3 in. apart (centerline to centerlinc), with a 5 ft lOVz in. clearance
between cyclones and the adjacent sidewalls. The main furnace is compact, having a mean bulk
gas residence time of only 0.9 s from the cyclone outlet plane to the horizontal plane at the
furnace arch.
Figure 16-6 shows Unit A configured for reburning. Like Niles, the reburning fuel injectors are
located on the rear wall opposite the cyclones. This arrangement not only maximizes reburning
zone residence time, but also should result in good dispersion of reburn fuel since the reburn fuel
jets directly oppose the cyclone jets. Five reburn fuel jets located 4 ft apart were chosen to
maximize natural gas contact with cyclone exit gas throughout the boiler cross section.
16-20
-------
Return System Economics
-14 ft-
-E-
N.
CO
CO
3-
Side View
17.5 ft
4 ft
~
-24 ft-
' T
.J
~
16.5 ft
4 ft 4 ft
O > 0 O O i o
""T" "T . i—
/ I \ 12ft3In. / |Sft 101/2In,
i-t-/
\ < /
-¦
r
i
i 7~
\ i
¦+
s
Rear View
C-8760
Figure 16-6
Unit A - 75 MW Gross, 1969
16-21
-------
Rebum System Economics
The burnout air ports are located on the front wall, 16 ft 6 in. above the reburn fuel injector
elevation. This location provides a normalized residence time of 0.77 in the reducing zone and a
normalized residence time of 0.88 above the burnout air ports to complete bumout. These
residence times are shorter than they were for Niles, but the narrow furnace at Unit A and the
closer side spacing of fuel and air injectors should result in faster mixing and similar NOx
performance.
Unit A is expected to produce about 1.2 lb NOx/mmBtu at full load prior to rebuming. The
baseline NOx is higher than Niles because the heat input per cyclone is higher, but not as high as
some of the other case boilers due to reduced flame temperature when burning western coals.
After rebuming is operational, the NOx emission should be reduced to about 0.45 lb/mmBtu
(62.5%). This NOx reduction is brought about by these factors:
1. The heat input to each cyclone is reduced by 20%, thus reducing peak combustion
temperatures and thermal NOx. The magnitude of this cyclone load reduction on NOx
emission is about 15%.
2. About 20% of the nitrogen-bearing fuel (coal) is replaced with a non-nitrogen-bearing
fuel (natural gas).
3. Natural gas reacts with NO to form reactive nitrogen species (like HCN, NH3) and N2.
The amount of NO destroyed in the reducing zone at a stoichiometric ratio of 0.9 is about
40 to 60% depending on residence time.
4. NOx can be reformed in the burnout zone by oxidation of reactive nitrogen species
escaping the rebum zone. Slow mixing of bumout air and low combustion temperatures
minimize NOx reformation. The NOx increase in the bumout stage ranges torn 0 to 20%.
This NOx reduction is higher than Niles because the baseline NOx is higher and because more
favorable furnace mixing conditions should result in lower CO emissions at Unit A than at Niles,
thus allowing operation at lower rebum zone stoichiometrics.
The other question regarding NOx emission predictions is, can the maximum NOx reduction be
maintained? The full load NOx reduction during short term tests at Niles ranged from 50% to
nearly 70%, depending mostly on the air/fuel balance among cyclones. Since Unit A only has
two cyclones and they are equipped with gravimetric coal feeders, NOx emissions should be
more stable and controllable at lower values than they were at Niles. However, NOx reductions
will probably decrease at low load.
Economics of a rebuming retrofit at Unit A are summarized below as estimated using the EPRI
TAG:
Process Capital
Total Plant Investment
Total Capital Required
S2.6M
$3.0M
$3.1M
16-22
-------
Reburn System Economics
Capital Cost per kW
10-Yr. Levelized Costs
Cost Effectiveness
Levelized Fuel cost Differential
10-Yr. Levelized Cost (w/o fuel Diff.)
Cost Effectiveness (w/o fuel cost Diff.)
$42 /kW
6.9 mills/kWh
$2069 /ton NOx removed
4.6 mills/kWh
2.3 mills/kWh
$686 /ton NC) removed
It can be seen that the 10-yr. levelized busbar power cost is dominated by the difference between
the cost of coal and the cost of natural gas. The first estimate assumes that natural gas escalates
to $1.50/mmBtu higher than the cost of coal, certainly a worst-case assumption and a drastic
change from today's market. During the time of the Niles demonstration, however, the fuel cost
differential paid by the project sponsors was $1.50/mmBtu. The second estimate assumes that
coal and natural gas have the same cost, a circumstance enjoyed over the last few years during
the summer months in some regions of the country but unlikely to continue in the future. Based
on these bracketing assumptions on natural gas price, the cost of reburning on Unit A calculated
using EPRI TAG methodology will be $686 to $2069/ton of NO* removed. The impact of fuel
cost differential is plotted in Figure 16-7.
Figure 16-7 also shows NOx removal cost effectiveness under the load profile scenarios
described above. The assumptions used in calculating NOx removed as a function of load are
tabulated below for Unit A:
This plot can be used to estimate the cost of implementing gas reburning for various assumptions
of load profile and natural gas cost. Like Niles, Unit A is limited to reburn operation at loads
above 67% because slag tapping will be a problem. Therefore, the unit is probably only a viable
candidate for reburning if it can be operated at high loads most of the time; and for this reason no
cost data are shown on Table 16-5 and Figure 16-7 for Unit A for the typical load profile and the
intermediate load profile.
Load, %
100
70
below 67
Baseline NOx, lb/mmBtu
1.2
1.0
N/A
NOx Removal, %
62.5
50.0
0
16-23
-------
Reburn System Economics
%
0)
E
a)
"a
I
3=
UJ
ffl
cc
x
o d
ffl z
c
p
ffl
cc
X
o
z
5000
4000
3000
2000 "
1000
EPRI TAG
0.50 1.00
Natural Gas Cost Differential ($/mmBtu above Coal)
1.50
Figure 16-7
Cost of Rebuming on a 75 MW Cyclone-fired Boiler
Unit B
Unit B is a 125 MW pressurized boiler located near the Appalachian coal fields. The furnace
arrangement is similar to Niles, except more compact. Three cyclone combustors located on the
same elevation fire into a primary furnace. Cyclone exit gases are forced downward by the
target wall, pass through a bank of screen tubes, and then upward through the main furnace.
Even though Unit B is rated about the same as Niles, the main furnace depth is 2-ft smaller, the
width is 1-fi shorter, and the furnace height (cyclone centerline to furnace arch) is almost 20 ft
shorter than the Niles unit. As a result, Unit B has much shorter residence times for reburning
and burnout compared to Niles or most other cyclone units.
Unit B was started up in 1957, four years later than Niles. It produced 675,000 lb/h of steam at a
superheater outlet pressure of 4550 psig and temperature of 1150 F. This unit was
decommissioned in the mid-1980s due to higher operating and maintenance costs compared to
other units in its system.
Certainly Unit B represents the most difficult case for cyclone-fired boiler reburning.
Figure 16-8 shows a general arrangement sketch of the unit, including potential fuel and air
injector locations. Five reburning fuel injectors are located on the rear wall opposite the cyclone
16-24
-------
Reburn System Economics
combustors. The reburn fuel injectors are spaced 5-ft 8-in. apart horizontally and tilted
downward to maximize residence time in the reducing zone. Burnout air ports are located on the
sidewalls 13-ft above the reburn fuel injectors. Sidewall air ports were chosen because, like
Niles, Unit B is likely to have no access for air duct work at this elevation on the front or rear
walls.
Figure 16-8
Unit B -125 MW Gross, 1957
These reburn fuel injector locations provide normalized reburning zone mean bulk-gas residence
time of 0.46. Normalized burnout zone time is 0.39. Both these residence times are less than
optimal for NOx reduction and carbon burnout. NOx reductions of only 40 to 50% are expected
at full load from Unit B.
Unit B baseline NOx at Ml load is about 1.4 lb/mmBtu. Full load NOx reductions are assumed to
be limited to 45% due to residence time constraints in this boiler. At partial loads (50 to 70%),
NOx reductions could increase slightly due to additional residence time available for mixing.
Both NOx destruction and carbon burnout times will increase at low loads. Longer burnout time
may further reduce NOx emissions by allowing operation at lower cyclone excess air levels
without worrying about increased unburned carbon in the flyash.
16-25
-------
Return System Economics
Assumptions for NOx reduction versus load are tabulated below for Unit B.
Load, % Baseline NOx, lb/mmBtu NOx Removal, %
100 1.4 45
70 1.2 50
50 1.0 50
below 44 N/A 0
Economics of a rebuming retrofit at Unit B are summarized below:
Process Capital
S3.22M
Total Plant Investment
S3.70M
Total Capital Required
S3.91M
Capital Cost per kW
$31.3 /kW
10-Yr, Levelized Costs
6.20 mills/kWh
Cost Effectiveness
$2212 /ton NOx
10-Yr. Levelized Cost (w/o fuel cost Diff.)
1.59 mills/kWh
Cost Effectiveness (w/o fuel cost Diff.)
$567 /ton NOx removed
Figure 16-9 shows the NOx removal cost effectiveness for this unit over a range of natural gas -
coal price differentials and load profile scenarios. Even with less effective rebuming, the cost
effectiveness of Unit B is comparable to that of Unit A at high load, and more flexibility exists
for achieving NOx reductions at low load.
16-26
-------
Reburn System Economics
Figure 16-9
Cost of Reburning on a 125 MW Cyclone-fired Boiler
Unit C
Unit C is a one-wall-fired cyclone boiler rated at 225 gross MW. It is located in the midwest and
first started up in 1958. Two elevations of cyclone combustors, arranged two over three, fire
bituminous coal. Slag is captured in a primary furnace, and the products of combustion pass
through a slag screen and into the main furnace where reburning takes place. Figure 16-10
shows a general arrangement sketch of Unit C with reburning applied.
The reburning fuel injectors are located on the rear wall of the main furnace and tilted downward
to maximize reburning foel mixing rate and residence time for NOx reduction reactions. There
are seven injectors spaced 6 ft apart to cover the entire boiler width. Burnout air ports are located
on the furnace sidewalls 13.5 ft above the reburn fuel injectors. Two air ports on each sidewall
are spaced about 5-1/2 fl apart. A potential problem with this layout is that burnout air must
penetrate and mix within a 48-ft boiler width. This distance is 14 ft larger than the boiler width
at Niles, thus increasing the risk of unburned combustibles in the flue gases. Unit C is another
very difficult retrofit candidate.
16-27
-------
Reburn System Economics
The costs of reburning calculated from the EPRI TAG and applied to Unit C arc summarized
below:
Process Capital
S4.50M
Total Plant Investment
S5.18M
Total Capital Required
I5.45M
Capital Cost per kW
$24.2 /kW
10-Yr, Levelized Costs
5.76 mills/kWh
Cost Effectiveness
$2057 /ton NOx removed
10-Yr. Levelized Cost (w/o fuel Diff.)
1.14 mills/kWh
Cost Effectiveness (w/o fuel cost Diff.)
$409 /ton NOx
Figure 16-10
Unit C - 225 MW Gross, 1958
Full load NOx emissions are estimated at 1,4 lb/mmBtu for Unit C. The heat input per cyclone is
within design limits for 10-ft cyclones and the exit gas from four out of five cyclones is cooled
by boiler sidewalls. Both these factors are similar to Unit B, indicating that full load NOx should
also be similar.
16-28
-------
Rehurn System Economics
Unit C is also expected to achieve better NOx reductions with reburning at partial load. Given
the cyclone configuration, minimum load of about 30% is achievable for short periods of time
with the lower three cyclones operating at about half their design heat input. Assumptions used
for calculating NOx removal as a function of load are tabulated below:
Load, % Baseline NOx , lb/mmBtu NOx Removal, %
100 1.4 45
70 1.2 50
50 1.0 50
below 40 N/A 0
Some economies of scale are realized since Unit C has double the megawatt production of Niles,
but inferior NOx reduction potential makes Unit C a less attractive candidate for gas reburning.
Figure 16-11 shows how reburning cost effectiveness for Unit C would be affected by natural
gas prices. The range of $409 to $2057/ton of NOx removed is comparable to Niles and Units A
and B using the EPRI TAG. Much better cost effectiveness is achieved for Unit C when
applying any of the load-following scenarios due to its ability to use reburning down as low as
40 percent load.
Figure 16-11
Cost of Reburning on a 225 MW Cyclone-fired Boiler
16-29
-------
Return System Economics
Unit D
Unit D is a modern cyclone boiler, having started up in 1968. It is located in the midwest and
designed to fire Illinois bituminous coals. It has been switched to low-sulfur Powder
River Basin subbituminous coal recently to comply with Title IV of the Clean Air Act
Amendments.
Unit D is a supercritical unit rated at 420 gross MW. Design steam temperatures are 1005 F for
both superheated and reheated steam at a superheater outlet pressure of 3810 psig. This unit is
fired by eight cyclones arranged two over two on opposed walls as shown in Figure 16-12. The
cyclones are staggered so that no two combustors are directly opposed. Cyclone horizontal
spacing is 15 ft (centerline to centerlinc) while vertical spacing is 17 ft.
The products of combustion pass from the cyclones into an open furnace where most of the
steam production occurs. Most open furnaces are wider than they are deep (in this case 36 ft
wide by 27 ft deep), and have ample volume available within which reburning can take place.
To apply reburning, natural gas injectors would be located on the front and rear walls at an
elevation 11 ft above the upper cyclone centerline. Each wall would contain six injectors spaced
about 5 ft apart to assure rapid mixing of reburning fuel with cyclone exhaust gas. The burnout
air ports would be located 40 ft above the reburn fuel injectors. The normalized mean bulk gas
residence time for NOx reduction would be 0.95. The 39 ft. between the burnout air ports and
the furnace arch would provide adequate residence time for burnout of the reburn fuel and any
unburned coal. The air ports would be located 6 ft apart horizontally to assure complete mixing
well before the furnace exit plane.
Reburning should be effective in reducing NOx by 70% at full load in Unit D (from a baseline of
1.4 lb/mmBtu to 0.4 lb/mmBTU) because all the design criteria listed on Table 16-3 are met, and
because the unit has gravimetric coal feeders that can help balance cyclone air/fuel ratios at
minimum excess air levels. The economics of the retrofit are summarized below as calculated
using the EPRI TAG:
Process Capital
$6.15M
Total Plant Investment
$7.08M
Total Capital Required
S7.46M
Capital Cost per kW
$17.8 /kW
10-Yr. Levelized Costs
5.41 mills/kWh
Cost Effectiveness
$1242 /ton NOx removed
10-Yr. Levelized Cost (w/o fuel Diff.)
0.79 mills/kWh
Cost Effectiveness (w/o fuel cost Diff.)
$182 /ton NOx
16-30
-------
Reburn System Economics
[
-27 ft-
39 ft
90 ft
40 ft
6 It , 6 ft
S.1 ft5.1
ft|5.1 ftj
o o o o o o
14ft
15 ft
7 ft
y
(
N
LA
\
)
y
\^
I
J \ ( \
\ , J \ , 7
7ft
IS ft
14ft
36 ft -
C-8768
Figure 16-12
Unit D - 420 MW Gross, 1968
Unit D has significantly higher energy input per cyclone than units B, C, and E. Higher cyclone
heat input usually means higher baseline NOx emissions. However, the staggered cyclone
arrangement, where the gases leaving each cyclone radiate to the main furnace sidewalls, will
16-31
-------
Reburn System Economics
result in relatively low peak temperatures in the main furnace and lower NOx formation
downstream of the cyclone furnaces. Therefore, baseline NOx similar to Units B and C is
predicted. Assumptions used to calculate NOx removal effectiveness for different load profiles
are listed below:
Minimum load for Unit D is about 105 MW for short term operation. This load is achievable
with only the lower four cyclones in service firing at about half load. The higher-than-normal
full load heat input is another advantage for this unit since the potential cyclone turndown could
be even more than 50% without freezing the slag layer in the combustor. Therefore, minimum
load with rebuming may be as low as 160 MW with a reasonable safety margin for slag tapping.
The dependency of cost effectiveness on natural gas price differential and load profile for Unit D
is shown in Figure 16-13. As expected, rebuming is much more cost-effective for larger boilers.
In addition, load profile is not quite as important as the cost effectiveness lines merge into 3
categories: high, typical, and intermediate.
Load, %
100
70
50
below 38
Baseline NOx, lb/mmBtu
1.4
1.2
1.0
N/A
NOx Removal, %
70
60
50
0
4000
0
0
0.50
1.00
1.50
Natural Gas Cost Differential ($/mmBtu above Coal)
Figure 16-13
Cost of Rebuming on a 420 MW Cyclone-fired Boiler
16-32
-------
Reburn System Economics
UnitE
UnitE is a 605 MW, once-through boiler located in the midwest. Started up in 1970, Unit E
burns Illinois bituminous coals as well as small amounts of waste fuels. Steam conditions are
1005 F superheat and reheat at a superheater outlet pressure of 2620 psig.
Figure 16-14 shows a sketch of Unit E retrofitted with natural gas reburning. The unit contains
14 cyclone furnaces, 7 mounted on the front and rear wall in a three-over-four anray, The main
furnace is 60-ft wide by 33-ft deep. Cyclone spacing is about 15-ft horizontally (see Figure
16-14 for detailed dimensions) and 17-ft vertically.
The reburn fuel injectors would be located on the front and rear walls about 15 ft above the
centerline of the top row of cyclones. Each wall would contain seven injectors spaced 8 ft apart.
The burnout air ports would be located 36 ft above the reburn fuel injectors, providing a mean
bulk-gas normalized residence time in the reburning zone of 1.18. Burnout air port side spacings
would be identical to those for the fuel injectors. There would still be 66 vertical ft from the air
ports to the furnace arch, providing a normalized mean, bulk-gas residence time for burnout of
1.72.
Baseline NOx for Unit E is expected to be around 1.7 lb/mmBtu at full load. The main factor
contributing to higher baseline NOx for this unit is the close packing of cyclone combustors in
the main furnace leading to high peak temperatures. These same high temperatures, however,
should help reburning effectiveness by increasing the rate of NOx destruction. Natural gas
reburning should be able to achieve a 70% NOx reduction down to 0.5 lb/mmBtu. The
economics as calculated by the EPRI TAG method for this retrofit are summarized below:
The assumptions used to calculate the NOx removal effectiveness for Unit E at different load
profiles are listed below:
Process Capital
Total Plant Investment
Total Capital Required
Capital Cost per kW
10-Yr. Levelized Costs
Cost Effectiveness
10-Yr. Levelized Cost (w/o fuel Diff.)
Cost Effectiveness (w/o fuel cost Diff.)
S7.39M
S8.49M
S8.97M
$14.83 /kW
5.26 mills/kWh
$994 /ton NOx
0.64 mills/kWh
$ 121 /ton NOx removed
Load, %
100
70
50
below 49
Baseline NOx, lb/mmBtu
1.7
1.5
1.3
N/A
NOx Removal, %
70
60
50
0
16-33
-------
Reburn System Economics
£
-33 ft-
66 ft
-3-
116.8 ft
36 ft
e*
17ft
6 ft , 8 ft
~ ~~~~on
6 It . 8 ft
~
-60 ft-
g q n ~ n
«^15 ft -^14.5 ft
16ft
15 ft ' 7 ft
C-8763
Figure 16-14
Unit E - 605 MW Gross, 1970
16-34
-------
Reburn System Economics
Minimum load of 175 MW for this unit is achievable with the eight lower cyclones operating at
half load. When reburning is added, each cyclone is derated by about 20% as heat input is
transferred to the rebum gas injectors. Leaving a margin of safety, rebuming is limited to loads
above 300 MW (49% of rated capacity). NOx removal efficiency will gradually decrease with
load as main furnace temperature decreases and cyclone outlet O2 increases.
Figure 16-15 summarizes the NOx removal cost effectiveness for Unit E. This unit appears to be
a good candidate for rebuming regardless of how coal and gas prices fluctuate over the next
decade.
3000
0.50 1.00
Natural Gas Cost Differential ($/mmBtu above Coal)
1.50
Figure 16-15
Cost of Rebuming on a 605-MW Cyclone-fired Boiler
Other Reburning Systems
Natural gas reburning has been applied by others to three other boilers in the United States
(Hong et al. (1993) and May et al. (1994)). These boilers are listed in Table 16-7. All three
projects are demonstrations funded partially by the U.S. Department of Energy under the Clean
Coal Technologies Program and by the Gas Research Institute.
16-35
-------
Reburn System Economics
Table 16-7
Other Boilers Retrofitted with Natural Gas Reburning
Illinois Power
Hennepin Station
Unit #1
Tangential
71 MW (g)
Public Service Company of Colorado
Cherokee Station
Unit #3
One-wall fired
172 MW (g)
City of Springfield, IL
Lakeside Plant
Unit #7
Cyclone
33 MW (g)
Recently capital cost data have been published for two of these units, Hennepin and Lakeside,
Swanekamp (1995). The installed capital cost at Hennepin was $38/kW while Lakeside came in
at $60/kW. EPRI TAG methodology was used in both cases,
Reburning performance has also been documented during long-term operation at both these
plants, May et al. (1994). Performance data are summarized in Table 16-8.
Table 16-8
Long-term Reburning Performance in Other Boilers
Performance Parameter
Hennepin
Lakeside
Load range
25 to 72 MW
23 to 34 MW
Average baseline NOx
0.75 Ib/mmBtu
1.0 Ib/mmBtu
Average controlled NOx
0.25 Ib/mmBtu
0.34 Ib/mmBtu
Range of daily NO average
0.18 to 0.32 Ib/mmBtu
0.21 to 0.47 Ib/mmBtu
Percent reburn fuel (Btu basis)
10 to 18%
20 to 26%
It can be seen that performance of these units was comparable to Niles.
Cost Summary
Figure 16-16 shows a comparison of the capital costs of reburning for the five study boilers.
Two other boilers (Hennepin and Lakeside) are also included in the comparison based on
published cost information. This figure shows that gas reburning is best justified on larger
baseloaded boilers. Further, the Hennepin and Lakeside data points lend credibility to the cost
escalation methodology employed in this study.
16-36
-------
Reburn System Economics
Figure 16-16
Capital Cost of Natural Gas Rebuming
Figure 16-17 shows the cost-effectiveness of rebuming for Niles and the study boilers using the
Niles load profile. Although boiler size has a significant effect on rebuming cost effectiveness
(due to both larger NOx reductions and economics of scale), the driver for implementation of
natural gas rebuming will be the cost of natural gas. As long as natural gas prices stay close to
the prices of coal, natural gas rebuming will be an attractive option for cyclone boiler NOx
control.
Smaller cyclone boilers (those having four or fewer combustors) can also be limited to rebuming
operation at 70 percent load or above. Many of these units like Niles are dispatched according to
system demands and subsequently operate at minimum load for much of the offpcak demand
periods. As Figure 16-17 shows, NOx removal costs increase sharply for Niles and Unit A at the
load profile typical of Niles. The larger the unit, the more flexibility the unit may have for
low-load operation with rebuming so that the technology becomes less dependent on load profile.
16-37
-------
Reburn System Economics
6000
Q
V
Niles
5000 -
z
3
LU
a>
DC
X 3000
O
S 4000-
E
2000
3000
1000-
0
0 100 200 300 400 500 600 700 800
Boiler Gross Megawatt Rating
Figure 16-17
NOx Removal Cost Effectivenesses for a Range of Boiler Sizes Based on the Niles Load Profile
Conclusions
A representative sampling of the cyclone boiler population has been evaluated for the feasibility
of retrofitting natural gas rebuming. Using design criteria from the Niles demonstration,
virtually all cyclone units can be retrofitted. Reburning NOx reduction, however, will range from
40 to about 70%, depending on the residence time available in the main furnace and the load
profile of the unit. Larger open-furnace designs will provide more residence time and thus
achieve greater NOx removal than smaller, primary-furnace designs. Base loaded plants will
achieve greater NOx removal since cyclone furnace turndown may limit how much natural gas
can be used at low load.
It was also found that cyclone boilers have shallower furnaces (less depth front-to-back) than
other types of boilers. The largest cyclone-fired units (TVA, Paradise #3 et al.) are only 33 ft
deep. The implication of this observation is that flue gas recirculation is not required to help
disperse the reburning fuel on any cyclone-fired boiler.
In summary, nearly all cyclone boilers can be retrofitted with gas reburning, but small furnaces
may limit the effectiveness of reburning in early boiler designs (1950's). The age and expec ted
lifetime of some cyclone units may make selective non-catalytic reduction processes (less
capital-intensive processes) better choices for NO* control in these older units. Reburning seems
most attractive for larger cyclone units built in the 1960's and 1970's (lignite-fired). The cost of
16-38
-------
Reburn System Economics
natural gas will be the most important single factor in determining the number of cyclone boilers
that can economically use rebuming to meet future NOx regulations.
Table 16-5, shown above, summarizes rebuming cost-effectiveness for all boilers and load
profiles, with and without a S1.50/mmBtu fuel cost differential between natural gas and coal.
For smaller boilers such as Niles or Units A and B, low load operating limits make rebuming
economically unattractive. Even for larger cyclone units, intermediate load operation will
increase the NOx removal cost by a factor of three. However, since most cyclone boilers were
designed for and are best suited for base loaded operation, rebuming is a good choice. Moreover,
for utility systems subjected to environmental dispatch, rebuming may in fact allow higher load
factors to be utilized in cyclone units. Each utility must weigh the technical and economic merits
of the technology for their own unique situation. It is the intent of Section 16 to provide enough
information to allow an intelligent first cut at the natural gas rebuming choice for all cyclone
boiler owners.
16-39
-------
17
CONCLUSIONS AND RECOMMENDATIONS
A natural gas reburn system was installed on Ohio Edison's Niles Unit No. 1, a 115 MW (gross)
cyclone-fired boiler. The objective was to demonstrate that 50% NQX reduction could be
achieved at full load and that the reburn system could be operated without adversely affecting
boiler thermal performance and component life.
The project at the Niles plant represented the first commercial demonstration of a natural gas
reburn system. Although the effectiveness of reburning as a NOx reduction technique has been
shown in many laboratory and pilot scale experimental tests, the subject demonstration was the
first to look at the total impact of a reburn system in a commercial boiler. Though NOx reduction
was the focus of the demonstration, it was even more important that the reburn system not cause
any unacceptable side effects on boiler operation and component life. Indeed, execution of this
project turned up a few unexpected results illustrating just why R&D demonstrations are
conducted. It is believed that results from this project were valuable in their own right and,
furthermore, that lessons learned here provided very useful input and direction to those who
would conduct follow-on demonstrations of reburn systems.
The original rebum system was designed to employ flue gas recirculation (FGR) as a carrier gas
for better mixing of the natural gas with the bulk flue gas in the reburn zone. Project objectives
were met with the original system relative to NOx reduction and boiler thermal performance.
However, much thicker slag deposits formed on the back wall, the one in which the reburn fuel
injectors were installed, compared to the base case deposits. The thicker deposits were found to
be caused by the relatively cooler FGR near the affected wall. The deposits, which were as much
as 12 inches thick (compared to the normal 2 to 4 inches), had little or no effect on boiler
performance and did not prevent completion of the original system test program. However, long-
term operation of the original reburn system was unacceptable for several reasons. Slag falls
during boiler operation could have a damaging effect on screen tubes at the bottom of the
furnace; the possibility of slag falls during slag removal operation was a risk to personnel; and
slag accumulation could cause blockage and misdirection of the reburn fuel jets as well as
shortened life of the nozzles due to overheating. For these reasons there was a need to identify
the cause of the problem and to resolve it.
Resolution of the slag buildup problem led to the development of the modified reburn system. In
the modified reburn system FGR was eliminated. Deposits on the back wall returned to normal
thickness. NOx reduction was initially lower than with the original system; but with continued
operation and increased operator familiarity, NOx reduction improved and during the last period
of long-term testing full load NOx reduction was greater than that achieved with the original
system. Importantly, there were a number of other advantages with the modified system both
17-1
-------
Conclusions and Recommendations
operational and economic: the modified system showed heat transfer distribution within the
boiler to be much closer to the base case conditions, and the cost of the reburn system was lower
due to the elimination of the FGR and associated equipment. The plant net heat rate was also
improved by eliminating the power requirement for the gas recirculation fan.
Long-term testing was carried out with the modified system under normal economic dispatch
conditions during a period of three and one-half months.
The following provides a summary of the most important observations and conclusions reached
during this demonstration project:
• Natural gas reburn significantly reduced NOx emissions from the Niles Unit No. 1 cyclone
fired furnace. Reburn also affected CO emissions. Specific NOx and CO emissions behavior
was observed as follows:
- An average NOx reduction of 52.1% was achieved at Ml load with acceptable boiler
operation and CO emissions lower than 200 ppm during the final series of long-term
dispatch tests when the reburn zone stoichiometry (RZS) was between 0.90 and 1.00.
- Reburn zone stoichiometry (RZS) was the most significant operating variable affecting
NOx reduction by the reburn process.
- NOx emissions decreased linearly as RZS was decreased.
- CO emissions increased exponentially when RZS was decreased.
- For long-term operation of a commercial reburn system RZS should be maintained
slightly above 0.9 to simultaneously minimize both NOx and CO. Because of the
inability to maintain precise coal/air ratios in each of the cyclones at Niles No. 1 during
long-term testing, simultaneous NOx and CO emissions were minimized at RZS of 0.94.
- NOx reductions of 30 to 70% were measured during parametric testing of the original
system at Ml load.
» Natural gas reburn had a minimal effect upon boiler performance and electrostatic
precipitator (ESP) performance.
- During 18% natural gas reburn testing with the original system, waterwall heat absorption
decreased by approximately 5%; attemperator spray flows, operating in a normal range,
were able to control steam temperatures at the design levels.
- Boiler efficiency decreased by 0.6% with 18% natural gas rebuming in the original
system due principally to higher latent heat of vaporization losses caused by greater
moisture formation from natural gas.
- ESP collection efficiency was lowered slightly during reburn system operation due to
lower ESP inlet loading and a non-optimized flue gas conditioning system.
• Operation of the original reburn system led to the buildup of much thicker ash deposits on the
rear wall of the furnace at Niles No. 1.
17-2
-------
Conclusions and Recommendations
- Long term operation of the reburn system could not be sustained with the original rebum
system due to abnormally heavy slag buildup on the back wall and over the rebum fuel
injectors.
- The primary cause of thicker ash deposits was the cooling effect of FGR on the rear wall,
- The cooler FGR caused the normally thin, molten deposits to become thicker, sintered
deposits as they equilibrated to the change in the thermal environment.
• The original reburn system was replaced by a modified rebum system in which the FGR
system was eliminated. Eliminating FGR eliminated the ash buildup problem. The modified
reburn system also provided several cost and operations advantages over the original rebum
system,
- Lower capital cost.
- Smaller space requirement.
- Elimination of the high maintenance, energy intensive FGR fan.
- More favorable furnace heat absorption distribution. Radiant section heat absorption
increased and convective section heat absorption decreased resulting in lower
attemperator water flow requirement. Boiler efficiency was essentially the same as that
of the original system.
• The modified reburn system, initially showed a NOx removal efficiency about 8% lower than
the original rebum system. Possible causes for the lower NOx reduction were initially
thought to be soot formation by the natural gas in the absence of the recirculated flue gas and
decreased mixing of the natural gas due to elimination of the recirculated flue gas. However,
NOx reduction improved as long-term testing continued; during the last period of long-term
testing, NOx reduction was greater than that achieved with the original rebum system.
Operator familiarity with the system and closer control of individual cyclone fuel/air ratios
was thought to be the reason for improvement.
• Water injection into the rebum zone was initially thought to improve NOx reduction during
testing with the modified rebum system. A water leak in one of the water-cooled rebum fuel
injector guide pipes seemed to correspond directly with increased NOx reduction. However,
controlled water injection tests conducted after completion of the long-term tests provided no
improvement in NOx reduction compared to NO* reduction achieved during the final series
of long-term tests. Controlled water injection did however accomplish the following:
- Lower CO levels; CO emission of 46 ppm and NOx emissions of 325 ppm (corrected to
3% O2) were achieved with water injection compared to CO emission ofllO ppm at the
same NOx emission level without water injection.
- The ability to operate the rebum zone at lower stoichiometrics (lower NOx), while
maintaining the CO at acceptable levels,
• Reburn systems installed on pressurized furnaces, such as Niles Unit No. 1, can result in a
hazardous situation if a casing leak occurs in the vicinity of the rebum zone because of the
presence of combustible gases.
17-3
-------
Conclusions and Recommendations
Possible commercial solutions were suggested:
- Convert pressurized units to balanced draft by adding an induced draft fan and associated
equipment,
- Convert tangent tube pressurized units such as Niles No. 1 to fusion welded walls by
adding fusion welds between the tubes.
- Erect an enclosure around the reburn zone which would operate at a slightly higher
positive pressure than the furnace to assure that any leakage would be into the furnace.
- Erect a "hood-like" structure around the upper part of the furnace so that gas composition
could be constantly monitored for possible changes.
It is unlikely that the first two could be economically justified. However, the third and fourth
options would be much less capital-intensive and could be configured to ensure safe reburn
system operation.
• Operational constraints place a limitation on the reburn fuel feed rate and corresponding NOx
reduction during reduced load conditions.
- In order to assure effective tapping of slag from cyclone-fired units, it is necessary to
maintain a minimum heat release rate and corresponding coal feed rate to the slag tap
region.
- The minimum heat release in the slag tap region is a function of the furnace size, cyclone
design, and coal ash fusibility.
- Since the fuel fed to the reburn zone does not contribute to heat release in the slag tap
zone, rebum fuel must be reduced and finally discontinued as boiler load is reduced.
- The need to decrease and ultimately discontinue reburn fuel is most severe in older,
smaller cyclone furnaces such as Niles No. 1 since less energy is available to maintain
effective tapping of liquid slag in these units. Reburn fuel feed was discontinued at Niles
during long-term testing when the boiler load was reduced below 80 MW gross.
- Because the proportion of reburn fuel used at reduced boiler loads is decreased and
ultimately turned off below a certain load, overall NOx reduction is less for reburn
systems installed on cyclone-fired furnaces which operate at reduced load for substantial
periods. The NOx concentration in the stack with reduced load however tends to remain
nearly constant because the "baseline" NOx also decreases with reduced load.
• The possibility of tube wastage during operation of the reburn system existed because the
reburn process generated a substoichiometric (reducing) gas mixture in the rebum zone. A
boiler tube monitoring program was conducted during the reburn system testing to address
this possibility. The findings of tube monitoring program were as follows:
- The ultrasonic thickness testing in the waterwall sections was inconclusive since changes
in tube thickness were below the sensitivity of the U.T. measurement. However, visual
inspection of the waterwalls revealed that the tube surface appeared to be unaffected by
reducing atmosphere corrosion.
17-4
-------
Conclusions and Recommendations
- Ultrasonic thickness measurements of the superheater and reheater sections, following
operation of the original reburn system, showed areas with an approximate 10% wall
loss, with wastage in areas of the fifth stage superheater as high as 0.100" over a 20 month
timeframe. Indicated tube loss is thought to be from a combination of erosion and
corrosion.
The modified reburn system, without FGR, maintained flue gas mass flows/velocities at
basecase levels, thereby minimizing wastage due to erosion. Because tube wastage was
not uniform, it is believed that erosion was the larger contributing factor between erosion
and corrosion.
- The remaining superheater/reheater tube life analyses performed before and after the
reburn project were inconclusive concerning any degradation due to high temperature
oxidation. Final inspection values gave higher remaining tube life values than did
initially obtained values.
• The cost effectiveness of natural gas reburn retrofit for reducing NOx emissions from
cyclone-fired furnaces depends upon several factors including the following: (1) the baseline
NOx and the expected NOx removal efficiency of the process over the load range of the
boiler, (2) the load profile of the boiler, (3) whether or not it is necessary to terminate reburn
operation at some boiler load due to slag tapping requirements and if so at what load this
requirement is imposed, and (4) the difference in fuel costs between natural gas and coal. A
study of natural gas reburn economics indicated that natural gas rebuming is most attractive
for newer large units, particularly, base-loaded units and least attractive for small, older units
used for cycling such as the Niles units.
• The economics study provided these data for Niles Unit No. 1; capital cost for installing a
reburn system = $34/kW; reburn cost-effectiveness = $5835/ton of NOx removed if the
natural gas/coal cost differential is $1.50/mmBtu; and reburn cost-effectiveness = $1592/ton
of NOx removed if there is no cost differential between natural gas and coal.
• Natural gas reburning is less attractive economically at Niles Unit No. 1 than at larger units
both because the baseline (uncontrolled) NOx emissions at Niles are low relative to larger
units and because the utilization of the NOx system at Niles is lower than at larger units.
Niles has four low heat-release cyclones which discharge gas into long, narrow passages
(high surface/volume ratio) resulting in relatively low gas temperatures and less NOx
formation than the NOx formation at larger more modern cyclone units. The utilization of the
reburn system at Niles is less than would exist at larger units because the Niles unit spends a
greater fraction of its operating time at part load and because the unit, since it operates at
lower gas temperatures, must cease reburn operation completely at a higher fraction of design
load.
Parametric testing and long-term testing during the Ohio Edison Rebum Demonstration project
provided several recommendations for reducing NOx and CO emissions by improvements to the
rebum system design and operation. These are:
17-5
-------
Conclusions and Recommendations
• Improve the control system for feed of coal and air to the cyclones in order to have better and
more uniform control of RZS. In this way the reburn system will be better able to operate
nearer to the optimum RZS which will provide higher NOx reduction without aggravating
CO levels.
® CO levels turned out to be a limiting factor for MOx reduction. Decreases in could
clearly produce lower NOx> but at the expense of unacceptably high CO. Better mixing of air
in the burnout zone and biasing residence times toward the burnout zone, rather than the
reburn zone, may result in lower NO* because of the ability to employ lower RZS while
maintaining acceptable CO levels.
• Introduce a small, controlled amount of H2O with the natural gas in the reburn zone to reduce
CO formation; this would allow lower RZS, higher NOx reduction and acceptably low CO.
• Use stainless steel for water-cooled guide tubes and other components which are subjected to
high temperatures in order to reduce the possibility of failure of reburn zone components.
17-6
-------
18
REFERENCES
Anderson, D.K., Bianca, J.D., and McGowan, J.G., (1986), "Recent Developments in Physical
Flow Modeling of Utility Scale Furnaces," Industrial Combustion Technologies. American
Society for Metals, Materials Park, OH.
Baldacci, L,, Bertacchi, S., and Michele, G. De. (1988), "Reduction of NOx Emissions by
Reburning Technology," Proceedings: Chemistry of Flames/Heat Transfer. Denmark, October
25-29.
Borio, R.W. et al, (1989), "Application of Reburning to a Cyclone Fired Boiler," Proceedings:
1989 Joint Symposium on Stationary Combustion NOx Control, San Francisco, CA, Volume 1,
EPA-600/9-89-062a (NTIS PB89-220529).
Borio, R.W. et al, (1991), "Rebum Technology for NOx Control on a Cyclone-Fired Boiler,"
Proceedings: 1991 Joint Symposium on Stationary Combustion NOx Control, Washington, D.C.,
Volume 1, EPA-600/R-92-093a (NTIS PB93-212843).
Borio, R.W., Lewis, R.D., Steen, D., and Lookman, A., (1993), "Long Term NOx Emissions
Results with Natural Gas Reburning on a Coal-Fired Cyclone Boiler," 1993 Joint Symposium on
Stationary Combustion NOx Control, Miami Beach, FL.
Bortz, S.J. and Often. G.R., (1987), "Reburning with Low and Medium BTU Gases,"
Proceedings: 1987 Joint Symposium on Stationary Combustion NOx Control, New Orleans,
March, Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).
Brown, S.W. and Borio, R.W., (1992), "Gas Rebum System Operating Experience on a Cyclone
Boiler," Proceedings: 1992 EPRI Symposium on NQX Control for Utility Boilers, Cambridge,
MA.
Brown, R.A. and Kuby, W.C., (1985), "Application of Reburning for NO„ Control in
Cogeneration," Proceedings: 1985 Symposium on Stationary Combustion NOx Control, Boston,
May, Volume 1, EPA-600/9-86-021a (NTIS PB86-225042).
Bruck. R.I., (1987), "Decline of Boreal Montane Forest Ecosystems in Central Europe and the
Eastern North America - Links to Air Pollution and the Deposition of Nitrogen Compounds,"
Proceedings: 1987 Joint Symposium on Stationary Source Combustion NOx Control, Volume 1,
EPA-600/9-88-026a (NTIS PB89-139695).
18-1
-------
References
Chen, S.L. et al, (1982), "N0X Reduction by Rcbiirning with Gas and Coal—Bench Scale
Studies," Proceedings of the 1982 Joint Symposium on Stationary Combustion NOx Control,
Volume 1, EP A-600/9-85-022a (NTIS PBS5-235604).
Chen, S.L. et al., (1983), "Influence of Coal Composition on the Effectiveness of NO, Control
Technology," Paper CSS/CT 83-03, Central States Section of the Combustion Institute Spring
Meeting.
Chen, S.L. et al, (1986), "Controlling Pollutant Emissions from Coal and Oil Combustors
Through the Supplemental Use of Natural Gas," Proceedings: 1986 Joint Symposium on Dry
S02 and Simultaneous NOx Control Technologies, Volume 1, EPA-600/9-86-029a (NTIS PB87-
120465).
Clark, W.D. et al, (1982), "Reburning with Pulverized Coal for NOx Control," Western States
Section of the Combustion Institute Spring Meeting, Paper No. WSS/CI82-12.
Eckhart, C.F. et al., (1989), "Conversion of Industrial Boilers to Burn Coal Water Fuel: Phase 1-
Pilot-Scale Development and Testing," U.S. Department of Energy Report: DOE/PC/90276-1,
Farmayan, W.F. et al, (1985), "Reduction of NOx Emission from Natural Gas Flames by Staged
Fuel Injection," Proceedings: 1985 Symposium on Stationary Combustion NOs Control, Boston,
May, Volume 2, EPA-600/9~86-021b (NTIS PB86-225059).
Farzan, H. et al, (1989), "Pilot Evaluation of Reburning for Cyclone Boiler NOx Control,"
Proceedings: 1989 Joint Symposium on Stationary Combustion NOs Control, San Francisco, CA,
Volume 1, EPA-600/9-89-062a (NTIS PB89-220529).
Farzan, II. et al., (1993), "Evaluation of Reburning NOx Control and Coal Switching in a
Cyclone Boiler," Engineering Foundation Conference on Coal Blending and Switching with
Western Low-Sulfur Coals, Snowbird, UT.
Folsom, B. et al., (1995), "Three Gas Reburning Field Evaluations: Final Results and Long-
Term Performance," EPRI/EPA 1995 Joint Symposium on Stationary Source Combustion NO,
Control, Kansas City, MO.
Greene, S.B. et al., (1985), "Bench-Scale Process Evaluation of Reburning and Sorbent Injection
for In-Fumace NOs/SOx Reduction," EPA-600/7-85-012 (NTIS PB85-185890).
Ilakkarinen, C. (1987), "An Overview of Environmental Issues Related to Nitrogen Oxides in the
Atmosphere," Proceedings: 1987 Joint Symposium on Stationary Source Combustion NOx
Control, Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).
Hong, C.C, et al., (1993), "Gas Reburning and Low NOx Burners on a Wall-Fired Boiler,"
Second Annual Clean Coal Technology Conference, Atlanta, GA.
Johnson, A.H. and Siccama, T.G., (1983), "Acid Deposition and Forest Decline," Environmental
Science and Technology, 17:294a - 305a.
18-2
-------
References
Kce. R.J. et al., (1980), "CHEMKIN: A General-Purpose Problem-Independent, Transportable,
Fortran Chemical Kinetics Code Package," Sandia National Laboratories Report: NTIS SAND
80-8003.
Knill, K.J., (1987), "A Review of Fuel Staging in Pulverized Coal Combustion Systems " IFRF
Doc. No. G13/V3.
Knill, K.J. and Morgan, M.E., (1989), "The Effect of Process Variables on NOx and Nitrogen
Species Reduction in Coal Fuel Staging," IFRF Doc. No. K70/a/l 1.
Lisaukas, R.A. et al., (1985), "Experimental Investigation of Retrofit Low-NOx Combustion
Systems," Proceedings: 1985 Symposium on Stationary Combustion NOx Control, Boston, May,
Volume 1, EPA-600/9-86-021 a (NTIS PB86-225042).
Lookman, A. and Glickert, R., (1992), "Effective Opacity Control Using Anhydrous Ammonia
Injection," Proceedings: 1992 Pittsburgh Coal Conference.
Maringo, G.J. and McElroy, M„ (1987), "Results from B&W Pilot Scale Reburning Tests
(EPRI/GRI)", September Progress Report.
Maringo, G.J. et al., (1987), "Feasibility of Reburning for Cyclone Boiler NOx Control,"
Proceedings: 1987 Joint Symposium on Stationary Source Combustion NOx Control, New
Orleans, March, Volume 2, EPA-600/9-88-026b (NTIS PB89-139703).
May, J.T. et al., (1994), "Gas Reburning in Tangential, Wall, and Cyclone-Fired Boilers,"
ASME-IEEE Joint International Power Generation Conference, Phoenix, AZ.
McCarthy, J.M. et al., (1986), "Pilot Scale Process Evaluation of Reburning for In-Furnace NOx
Reduction," EPA-600/7-86-048 (NTIS PB87-140323).
McCarthy, J.M. et al., (1987), "Pilot Scale Studies on the Application of Reburning for NOx
Control," Joint Symposium on Stationary Source Combustion NO, Control, New Orleans,
March, Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).
Miyamae, S. et al., (1985), "Evaluation of In-Furnace NOs Reduction," Proceedings: 1985
Symposium on Stationary Combustion NOx Control, Boston, May, Volume 1, EP A-600/9-86-
02 la (NTIS PB86-225042).
Mulholland, J. A. and Hall, R.E., (1985), "The Effect of Fuel Nitrogen in Reburning Application
to a Firetube Package Boiler," Proceedings: 1985 Symposium on Stationary Combustion NOx
Control, Boston, May, Volume 1, EPA-600/9-86-021a (NTIS PB86-225042).
Mulholland, J.A. and Lanier, W.S., (1985), "Application of Reburning for NOx Control to a
Firetube Package Boiler, "J. of Engineering for Gas Turbines and Power. 107. July,
18-3
-------
References
Mulholland, J.A. and Hall, R.E., (1987) "Fuel Oil Rcburning Application for NOx Control to
Firetube Package Boilers," I, of Engineering for Gas Turbines and Power. 109. April
Mulholland, J.A. et al., (1987), "Reburning Application to Firetube Package Boilers," EPA-
600/7-87-011 (NTIS PB87-177515).
Mulholland, J.A. and Srivastava, R.K., (1987), "Pilot-Scale Tests of a Multi-Staged Burner
Designed for Low NO, Emission and High Combustion Efficiency," Proceedings: 1987 Joint
Symposium on Stationary Combustion NOx Control, New Orleans, March, Volume 2, EPA-
600/9-88-026b (NTIS PB89-139703).
Murakami, N., (1985), "Application of the MACT In-Furnace NOx Removal Process Coupled
with a Low-NOx SGR Burner," Proceedings: 1985 Symposium on Stationary Combustion NOx
Control, Boston, May, Volume 1, EPA-600/9-86-021 a (NTIS PB86-225042).
Myerson, A.L., (1974), "The Reduction of Nitric Oxide in Simulated Combustion Effluents by
Hydrocarbon-Oxygen Mixtures," 15th Symposium (International) on Combustion. The
Combustion Institute.
Narita, T. et al., (1987), "Operating Experiences of Coal Fired Utility Boilers Using Hitachi NO,
Reduction Boilers," Proceedings: 1987 Joint Symposium on Stationary Combustion NO,.
Control, New Orleans, March, Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).
Okamoto, T. et al., (1983), "Fuel NOx Reduction by the Combination of Air Staging and
Secondary Fuel Injection," ASME-JSME Thermal Engineering Joint Conference Proceedings.
Honolulu.
Overmoe, B.J. et al., (1985), "Pilot Scale Evaluation of NOx Control from Pulverized Coal
Combustion by Reburning," Proceedings: 1985 Symposium on Stationary Combustion NOx
Control, Boston, May, Volume 1, EPA-600/9-86-02 la (NTIS PB86-225042).
Penterson, C.A. et al., (1989), "Reduction of NOx Emissions From MSW Combustion Using Gas
Reburning," Proceedings: 1989 Joint Symposium on Stationary Combustion NOs Control, San
Francisco, CA, Volume 1, EPA-600/9-89-062a (NTIS PB89-220529).
Radak, L.J. et al., (1982), "In-Furnace Control of NO Formation in Gas- and Oil-Fired Utility
Boilers," Proceedings of the 1982 Joint Symposium on Stationary Combustion NOx Control,
Volume 1, EPA-600/9-85-022a (NTIS PB85-235604).
Reed, R.D. et al., (1975), "Process for Disposal of Oxides of Nitrogen," U.S. Patent No.
3,873,671 (John Zink Co.).
Singer, J.G. (Editor), (1991), Combustion Fossil Power. Fourth Edition, Combustion
Engineering, Inc., Windsor, CT.
Steam. Its Generation and Use. (1992), 40th Edition, Babcock & Wilcox, a McDermott
Company, Barberton, OH.
18-4
-------
References
Swanckamp, R., (1995), "Natural Gas: Poised to Penetrate Deeper into Electric Generation,"
Power Magazine, Vol. 139. No. 1.
TAG - Technical Assessment Guide. (1989), EPRI Report P-6587-L, Vol. 1.
Takahashi, Y. et al., (1982), "Development of'MACT' In-Furnace NOs Removal Process for
Steam Generators," Proceedings of the 1982 Joint Symposium on Stationary Combustion NOx
Control, Volume 1, EPA-600/9-85-022a (NTIS PB85-235604).
Toqan, M. et al., (1987), "Reduction of NOx by Fuel Staging," Proceedings: 1987 Joint
Symposium on Stationary Source Combustion NOx Control, Volume 2. EPA-600/9-88-026b
(NTIS PB89-139703).
Wendt, J.O.L. et al., (1973), "Reduction of Sulfur Trioxide and Nitrogen Oxides by Secondary
Fuel Injection," 14th Symposium (International) on Combustion. The Combustion Institute.
Wilson, S.M. et al., (1991), "Demonstration of Low NOx Combustion Control Technologies on a
500 MWe Coal-Fired Utility Boiler," Proceedings: 1991 Joint Symposium on Stationary
Combustion NOx Control, Volume 1, EPA-600/R-92-093a (NTIS PB93-212843).
Yang, R J. et al., (1985), "Screening and Optimization of In-Furnace NOx-Reduction Processes
for Refinery Process Heater Applications," Proceedings: 1985 Symposium on Stationary
Combustion NO_ Control. Boston, May, Volume 1, EPA-600/9-86-021a (NTIS PB86-225042).
Yang, R.J., Arand, J.K., and Garcia, F.J., (1984), "Laboratory Evaluation of In-Fumace-NQ,-
Reduction for Industrial Combustion Applications," ASME Paper 84-JPGC-FU-12.
Zeldovich, Y. B., Sadovinkov, P. Y., and Frank-Kamenetskii, D. A,, (1974), "Oxidation of
Nitrogen in Combustion," Academy of Science of USSR, Moscow.
18-5
-------
APPENDIX A
ULTRASONIC THICKNESS (UT) MEASUREMENTS AT
OHIO EDISON COMPANY'S NILES PLANT UNIT NO, 1
W. R. Roezniak
ABB Power Plant Laboratories
Research and Technology
A Division of Combustion Engineering, Inc.
Windsor, Connecticut 06095-0500
August 1992
-------
Synopsis of Appendix A
This appendix is twenty (20) plots of waterwall, secondary superheater, and reheater tube
wall thickness measured during June 1990, December 1990, October 1991, and August
1992. The plots are labeled Figure 3A through Figure 12B. A discussion of the plots is
given in Section 14.
A-2
-------
A-3
-------
to
* » . .
: w
» » 1 •
: oj
*•4 i t
5—
Wi-
: Z3
• U)
tt..
• Ll
A-4
-------
A-5
-------
A-6
-------
A-7
-------
A-8
-------
A-9
-------
10
-------
A-ll
-------
A-12
-------
A-13
-------
A-14
-------
A-l 5
-------
-------
A-17
-------
A-18
-------
A-19
-------
-------
A-21
-------
A-22
-------
APPENDIX B
Memorandum, A. L. Waddingham to Sher Durrani,
"Niles No. 1 Boiler Waterwall Survey"
Ohio Edison
Chemical and Material Applications Center
1501 Commerce Drive
Stow, Ohio 44224
January 9,1991
-------
Synopsis of Appendix B
This appendix is a memorandum by A. L. Waddingham of Ohio Edison Co. describing the
condition of the boiler tubes at Mies No. 1 during an inspection of December 30, 1990. The
appendix also includes twelve (12) photographs taken during the inspection.
B-2
-------
f II F fi/5 -~
cj£- rvl
pr ' MEMORANDUM
TO: SMDurrani January 9, 1991
14th Floor
FROM; ALWaddinghara
Central Chemical Lab
SUBJECT: Niles No. 1 Boiler Waterwall Survey
On December 30, 1990, I monitored the subject inspection
which was performed by Combustion Engineering's ultrasonic
testing company. All readings were obtained by using a
Krautkramer Branson USK7 Flaw detector with a contoured,
dual-element 5 MHz probe. Calibration was performed^on a
machined tube and was checked after each set of readings.
The surface of the tubes was cleaned by sand blasting to
white metal; the couplant was a cellulose-gel type.
The following is a list of test strips that were sand-
blasted:
Elevation
914' All four walls
909' Rearwall only
902' All four walls
896' " " "
890' " " "
880' " " "
It should be noted that all areas were tested except the
front wall strips (target wall) at elevations 914' and 902',
due to inaccessibility, and every 3rd tube was tested at each
elevation except for elevation 909' in which they did every
tube. Furthermore, readings were obtained on the left,
center and right of each tube unless access was not available
or studding obstructed the transducer.
Although a copy of the data was not obtained, I did take
several photographs of the "surface" conditions (see
attachments). Based on the visual examination, I feel that
very little external "reducing" corrosion has occurred. Some
low values were recorded at elevation 909'. The low numbers
were not due to external corrosion, but due to internal
gouging at the bond.
If you have any questions, or require additional
information, please advise.
A.T. Waddin
... T~
ALW/lrp ' " |
cc: HCCouch RECEIVED
JMMurray
DLTackett y JAN 14 7Q"i
KHWorkman ^
„ Cth. FIT.
n«:t , papers
-------
ELEVATION 914'
B-4
-------
ELEVATION 909'
B-5
-------
ELEVATION 902'
B-6
-------
ELEVATION 896'
m
y|
B-7
-------
ELEVATION 890'
B-8
-------
ELEVATION 880'
9
jut * 5f \
; ^ -
v ... .rl ns. •?- w
^ ISr ft#1.*;/ iiK 1
r » "S^t4**!
H>
W'
jS^H^F -v i# - S® . a ^1
^jg^ v^jf jr j
«k,.
F OT
K1
l^jLp ,6 ail
fa
Wv \ 1
mm
^ * i rrm
' / Cv -
i i
IS
B-9
-------
APPENDIX C
INTERIM FIELD TEST REPORT
W. R. Roczniak
ABB Power Plant Laboratories
Research and Technology
A Division of Combustion Engineering, Inc.
Windsor, Connecticut 06095-0500
April 23,1992
-------
Synopsis of Appendix C
This appendix summarizes the results of boiler tube corrosion investigations conducted for the
Ohio Edison reburn project through April 1992.
C-2
-------
ff+Mm
April 23, 1992
Sher M. Durrani
Gen. Proj. Engr.
Ohio Edison
76 South Main St.
Akron, OH 44308
Attached is the interim report presenting the data associated with corrosion potential of surfaces
obtained at pre-start of testing, after parametric testing and after base line testing.
The corrosion probes show no significant wastage during the first two phases of testing. The
U.T. measurements taken after completion of the parametric testing phase and the base line test
phase are inconsistent The original readings obtained prior to testing, of the secondary
superheater stages suggest that the tubing walls were either much higher than the specifications
or that an adherent scale remained. Thus, they appear to be inconsistent with the readings
obtained during parametric and base line data or may have been obtained with a different
instrument.
The waterwaU measurements are similar for the pre-start data with the base line data but for the
parametric test data the readings are lower, of course, the tubing could not have increased in
thickness during operation.
The measurements to be taken after completion of the rebum phase will be incorporated in these
plots and should assist in resolving the U.T. measurement inconsistencies reported thus far.
WRR/haw
ce: R. Borio
A. L. Plumley
R. Lewis
ABB Combustion Engineering Systems
Combustion Engineering, Inc.
1000 Prosped MM Road
Posi Office Box 500
Windsor, Connecticut 06095-0500
Telephone (203) 688-1911
Fax (203! 285-9512
Telex 99297 COMflEN WSOR
C-3
-------
INTERIM FIELD TEST REPORT
INTRODUCTION
This report presents the data associated with corrosion potential of surfaces obtained during the
parametric phase and the base line phase of the test program in preparation for the reburn phase
operation. These data were utilized to determine any accelerated wastage during parametric
testing and to establish a wastage rate during normal operating parameters and will be used for
comparison to similar data generated during the reburn phase of the program.
BACKGROUND
Several studies have been conducted over the last twenty years to evaluate the reduction of NOx
in utility steam generators. The prime goal of these studies was the assessment of the
effectiveness of the processes and/or modifications employed to achieve NOx reduction. A
secondary goal was to determine the effects of the low NOx operating conditions on wastage
rates in these test vehicles. The determination of these wastage rates was obtained by the direct
measurement of heat surfaces by ultra sonic techniques (UT), and by the use of temperature
controlled corrosion probes or by the utilization of integral test sections. The integral test
sections were documented metal lographically prior to installation and removed after completion
of the test for metallographic evaluation. The U.T. measurements were obtained prior to
initiation of the test and at fixed intervals during the test program. The temperature controlled
probes were installed and removed from the unit without requiring unit outages.
Operating these units at minimal excess oxygen could potentially increase the wastage rates by
interrupting the formation of protective oxide coatings, thus influencing the rate of wastage.
In the previous test programs, accelerated wastage was not reported under any of the test
conditions studied.
FIELD TEST PROGRAM
The test program in the "Reburn" study conducted in Ohio Edison's Niles Unit No. 1, was
scheduled to be completed in a two year period. This program employs the use of temperature
WKR/05-92.wp
C-4
-------
controlled corrosion probes and extensive U.T. measurements between several phases of the
program. The large number and frequency of U.T. measurements was requested by Ohio
Edison. The measurements were obtained prior to the start of parametric testing, after the short
term parametric tests (approximately three weeks of demonstrations in a six month period) and
after the base line data were established. This program consists of three phases of testing, i.e.,
1) The parametric testing to establish operating parameters, 2) the base-line phase (approximately
one year of operation) and 3) the long term reburn phase.
The U.T. measurements were taken at five elevations in the furnace cavity and just above the
rear wall bend in the furnace cavity. Also, the leading tubes of each element of the secondary
superheater stages and the reheater were measured at three distances (1/4, 1/2, 3/4 the tube
length) of each of the stages. Reference marks were inscribed at the test elevations to facilitate
measuring the same locations at each outage.
Eight temperature controlled corrosion probes were exposed in the waterwall openings installed
for this purpose. Two probes were located in the rear wall along with one probe on each side
wall at the lowest elevation, which was just above the gas injection nozzles in the unit. Two
probes were installed at mid-point of the side walls, one on each side wall, and two probes just
below the entrance into the secondary superheater portion of the furnace. In addition to these
waterwall probes, a superheater probe was installed between the fourth and fifth stages of the
secondary superheater.
The complete set of probes were utilized during the base-line phase of the program and were
installed for the reburn phase of the program. A limited number of waterwall probes was
requested for the parametric testing, which was not within the original scope of the program.
SELECTION OF ALLOYS
Materials to be evaluated during these tests were the materials of construction utilized in the
furnace. These included carbon steel, T-l 1 material (11A % Chromium - Vi % Molybdenum),
T-22 material (21A % Chromium - 1 % Molybdenum) and T-9 material (9 % Chromium - 1 %
WRR/05-92.wp
C-5
-------
Molybdenum). Materials not utilized in fabrication of the unit included T-91 material (a
modified T-9 material), 3G4ss material (18% Chromium - 8% Nickel) and 310ss material (25%
Chromium - 20% Nickel). These additional materials were selected for evaluation to determine
their effectiveness in this environment should accelerated wastage occur to the materials of
fabrication.
TEST PROBES
Each of the waterwall probes was composed of five (5) threaded test rings machined to fixed
dimensions for insertion in the test locations. The superheater probe consisted of fifteen (15)
test rings which were held on the assembly by spring tension.
The hardware to control temperature, log, average and store temperature profile data consists
of a diskless industrial computer, thermocouple input cards, and an analog output card with
digital I/O lines. The analog output is used to control the proportioning air valves, and the
digital I/O for sensing limit switch status for enunciating alarms. All control equipment is
mounted in a dust-tight air conditioned enclosure. Data retrieval capability utilizing a telephone
modem provides a mechanism to monitor the operational status of the probes.
RESULTS
The corrosion probes exposed during the parametric testing phase were removed from the unit
in December of 1990. The test results of the ring specimens are shown in Table 1. The table
shows the weight loss of the materials from the probes exposed at the four elevations of the
furnace. Also included in the table are the maximum wall penetrations measured by micrometer.
Minimal weight loss differences were noted between the carbon steel rings and the T-22
material. The 304ss material was the least affected of these probe rings.
The wastage rates of the corrosion probes exposed during the base line test phase of the program
are found in Table 2. The waterwall probes in the rear wall at the lowest elevation experienced
exposure at high temperature above the control limits established for the test and consequently
greater wastage. All of the other probes operated within the preset temperature limits. The top
WRR/05-92.wp
C-6
-------
elevation experienced the highest wall penetrations of the remaining probes. These probes were
exposed from January 1991 into October 1991.
During the October outage, modifications were made to the rebum system. Several random
tubes located in the gas injector nozzle openings were removed for physical evaluation in the
laboratory. These were from the outer extremes of the assemblies near the straight tubes in the
furnace cavity. A schematic of the tubes is shown in Figure 1. The micrometer measurements
of the rings after cleaning are located in Table 3. These measurements assume the bench mark
as position A. The benchmark reading is the lowest in all of the test rings. As a general rule
the backside of the tube is usually the highest reading of exposed tubing but the measurements
show no appreciable wastage. The benchmark was established by the studs and the position of
the external deposits located on the surface of the tubes. Also noted is stud burn back both in
diameter and length.
The superheater corrosion probe was also removed in December of 1990, The test results of
the specimens are shown in Table 4. The table shows the weight loss of the materials as well
as the maximum penetration for each test ring. The weight loss is shown as a function of time
of exposure in Figure 2. Weight loss is shown increasing with exposure at increasing metal
temperature. Associated with this is the resistance to corrosion of the metals containing
increased chromium concentration. The effect of temperature is shown in the data shown at the
highest exposure time. These test rings were also exposed, at lower temperatures than the other
two exposure times shown in the Figure.
ULTRASONIC WALL THICKNESS MEASUREMENTS
U.T. measurements were obtained during each outage and prior to initiating parametric testing.
All the measurement were plotted for each elevation (see Figures 3-12).*
WATERWAIX MEASUREMENTS
The front wall (Figures 3) shows that no accelerated wastage has occurred over the entire length
of the wall. These measurements show 5 to 10 mils from June 1990 to October 1991. Both the
*The original Figures 3-12 were updated in August 1992.
WRR/054)2.wp
C-7
-------
right and left wall measurements (Figures 4 & 5) show no accelerated wastage. Isolated tubes
were measured and resulted in readings higher than the earlier outage measurement. This
apparently is erroneous since the tubes can not gain thickness during exposure. The rear wall
generated the same type of conditions (Figure 6). Accelerated external wastage was not detected
on any wall.
The rear wall just above the wall bend at elevation 909, had several specific tubes which were
considered below specification. The outward appearance displayed no wear flattening,
corrosion, or erosion. These tubes were reported to have internal wastage and Ohio Edison is
aware of the conditions of these tubes.
SUPERHEATER MEASUREMENTS
Ultrasonic thickness measurements were obtained at 1/4, 1/2 and 3/4 points of every tube in the
lower tube bank of each stage of the convection section (superheater & reheater bundles). See
Figures 7-12.
Wastage in general was slight except for the 15 tubes in the center section of the 4th and 5 th
stages. Apparent metal losses of up to 80 mils were observed on these few tubes (Figures 9 &
10) in a pattern suggesting that soot blower erosion or poor gas distribution coupled with fly ash
carryover were accelerating the wastage. Note that the thinning continued and even accelerated
during the period January to October 1991 when no reburn activity was occurring.
The UT measuring team had noted that the tubes were polished in a pattern more typical of
erosion than corrosion. The locations of these areas were reported to Ohio Edison personnel at
the time.
INTERIM TEST CONCLUSIONS
The data generated by the corrosion probes indicate increased penetration of the materials with
elevation in the boiler. Review of the three sets of ultrasonic thickness readings obtained at pre-
selected elevations on the waterways and in the convection section does not appear to indicate
WRJR/05-92.wp
C-8
-------
that there was any excessive general wastage occurring during the past 11/2 years of operation
with or without rebum.
Watcrwall measurements showed variations of ±5 to 10 mils between June 1990 and October
1991. The superheater and reheater measurements were also consistent at all three outages with
one exception. The center 15 tubes in the 4th and 5th stages of the superheater bundles were
found to have measurable thinning in a pattern which suggests sootblower erosion or excessive
carryover of ash due to mal distribution. The phenomenon was observed even at the time of the
original baseline tube measurements prior to any reburn activity.
WRJR/05-92.wp
C-9
-------
TABLE 1
CORROSION PROBES
PARAMETRIC TESTING
JUNE 1990 - DECEMBER 1990
LOWEST ELEVATION - RIGHT SIDE
LOWEST ELEVATION - RIGHT SIDE
MAXIMUM WEIGHT LOSS
CARBON STEEL 0.13 GRAMS/IN2
T-22 0.12 GRAMS/IN2
304ss 0.04 GRAMS/IN2
MAXIMUM WALL PENETRATION
0.001"
0.001"
NONE MEASURED
LOWEST ELEVATION
CARBON STEEL
T-22
304ss
0.09 GRAMS/IN2
0.07 GRAMS/IN2
<0.01 GRAMS/IN2
0.001"
0.001"
NONE MEASURED
MIDDLE ELEVATION*
CARBON STEEL
T-22
304ss
0.12 GRAMS/IN2
0.12 GRAMS/IN2
0.01 GRAMS/IN2
0.001"
0.001"
0.001"
TOP ELEVATION
CARBON STEEL
T-22
304ss
0.04 GRAMS/IN2
0,04 GRAMS/IN2
<0.01 GRAMS/IN2
0.001"
0.001"
NONE MEASURED
* PROBE EXPERIENCED HIGH TEMPERATURE
WRR/05-92.wp
C-10
-------
TABLE 2
CORROSION PROBES
BASE LINE TESTING
JANUARY 1991 - OCTOBER 1991
LOWEST ELEVATION - RIGHT SIDE
LOWEST ELEVATION - RIGHT SIDE
MAXIMUM WEIGHT LOSS
CARBON STEEL
T-22
304ss
0.78 GRAMS/IN2
0.66 GRAMS/IN2
1.0 GRAMS/IN2
MAXIMUM WALL PENETRATION
4 MILS
6 MILS
1 MIL
*REAR RIGHT SIDE
CARBON STEEL
T-22
304ss
2.12 GRAMS/IN2
3.42 GRAMS/IN2
2.60 GRAMS/IN2
14 MILS
23 MILS
21 MILS
*REAR LEFT SIDE
CARBON STEEL
T-22
304
2.90 GRAMS/IN2
1.42 GRAMS/IN2
3.14 GRAMS/IN2
16 MILS
8 MILS
11 MILS
LEFT SIDE
CARBON STEEL
T-22
304
0.54 GRAMS/IN2
0.30 GRAMS/IN2
0.14 GRAMS/IN2
2 MILS
1 MIL
<1 MIL
WRM)5-92.wp
C-ll
-------
MIDDLE ELEVATION - RIGHT SIDE
MAXIMUM WEIGHT LOSS
CARBON STEEL
T-22
304
0.18 GRAMS/IN2
0.24 GRAMS/IN2
0.04 GRAMS/IN2
MAXIMUM WALL PENETRATION
1 MIL
1 MIL
1 MIL
LEFT SIDE
CARBON STEEL
T-22
304
0.92 GRAMS/IN2
0,92 GRAMS/IN2
0.12 GRAMS/IN2
4 MILS
5 MILS
1 MIL
TOP ELEVATION - RIGHT SIDE
CARBON STEEL 2.92 GRAMS/IN2
T-22 1.32 GRAMS/IN2
304 0.18 GRAMS/IN2
7 MILS
4 MILS
<1 MIL
LEFT SIDE
CARBON STEEL
T-22
304
2.08 GRAMS/IN2
1.28 GRAMS/IN2
0.08 GRAMS/IN2
10 MILS
5 MILS
<1 MIL
* OVERHEATED
WRR/05-92.wp
C-12
-------
TABLE 3
Wall Thickness Measurements of Waterwall Tubes
Removed in October 1991
Position
Bench Mark
B
C
D
A
Tube 1
0.257
0.262
0.262
0.254
Tube 2
0.259
0.258
0.256
0.255
Tube 3
0.257
0.257
0.257
0.256
Tube 4
0.259
0.261
0.257
0.253
STUDS
Diameter
Length
Tube 1 (1 stud)
0.430
0.794
Tube 2 (2 studs)
0.438
0.810
0.444
0.798
Tube 3 (2 studs)
0.477
0.850
0.460
0.865
Tube 4 (2 studs)
0.442
0.857
0.450
0.820
WRR/05-92.wp
C-13
-------
TABLE 4
Superheater Corrosion Probes Weight Loss Data
Includes Time of Exposure and Temperature
Metal Temperature Test Rings
Ring No.
Material
Weight Loss
in Grams
Weight Loss
Gram/in2
Temperature
°F
Length of
Exoosure
1
T-22
3.236
0.26
910
5800 hours
2
T-22
4.812
0.39
930
1!
3
T-91
2.439
0.19
950
f!
4
T-91
3.432
0.27
970
?f
5
304ss
1.105
0.09
990
H
6
T-ll
12.225
0.98
1010
II
7
T-22
11.746
0.94
1030
5800 hours
8
T-22
14.706
1.18
1050
11
9-1
T-91
1.638
0.13
1070
1500 hours
10-1
T-91
1.766
0.14
1090
M
11-1
304ss
0.165
0.01
1110
1!
12-1
T-22
4.746
0.38
1130
II
13-1
T-91
2.019
0.16
1150
!!
14-1
304ss
0.123
0.01
1170
!!
15-1
31 Oss
0.061
0.01
1190
1500 hours
9-2
T-91
4.950
0.39
1070
4300 hours
10-2
T-91
6.389
0.51
1090
n
11-2
304ss
1.336
0.11
1110
It
12-2
T-22
25.607
2.05
1130
it
13-2
T-91
4.942
0.39
1150
1!
14-2
304ss
0.968
0.08
1170
It
15-2
31 Oss
0.349
0.03
1190
11
WRR/05-92.wp
C-14
-------
FIGURE 1
SCHEMATIC OF MEASUREMENT LOCATIONS
OF WATERWALL TUBES REMOVED DURING
OCTOBER 1991 OUTAGE
D
C
B
A
-------
FIGURE 2
SUPERHEATER MATERIAL PERFORMANCE
2.2
2
-
O
1.8
-
1.6
-
1.4
-
1.2
-
~
1
8
0,8
-
0.6
-
+
0.4
-
~
+
~
0.2
0
_
*
i a i i i
8
A |
a
+
0
1
0 2 4
(Thousands)
EXPOSURE TIME - HOURS
6
~ T22 + T91 0 TP304 A TP310
X T1 1
-------
In-situ Tubes
0.38
0.37
0.36
0.35
0.34
0.33
0.32 -
0.31
0.3
4th Bank - #19 5th Bank - #13 Reheater - #13
4th Bank - #34 5th Bank - #25 Reheater - #24
n
i
*<1
12-90
10-91
8-92
E3
8-92R
-------
GAS SLOW
.355
KL
f
.366
.322
.363
4TH BANK
TUBE #19
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.320
0.320
0.325
12-90
0.365
0.350
0.300
10-91
0.305
0.330
0.290
8-92
0.295
0.345
0.290
-------
GAS ELOW
.355
356
.350 H
If
k
B .361
1 1 ^
.370 p
0 .371
:
.364
.378
4TH BANK
TUBE #34
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.340
0.350
0.350
12-90
0.380
0.365
0.315
10-91
0.345
0.340
0.320
8-92
0.340
0.340
0.320
-------
GAS FLOW
? .356
5TH BANK
Tl IRC 4*4 O
I WJ *r K O
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.370
0.350
0.385
12-90
0.375
0.360
0.350
10-91
0.350
0.370
0.350
8-92
0.380
n
0.390
-------
5TH BANK
TUBE #25
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.405
0.385
0.400
12-90
0.385
0.360
0.360
10-91
0.360
0.340
0.330
8-92
0.360
0.360
0.360
C-21
-------
GAS FLOW
~ .356
REHEATER TUBE #13
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.350
0.355
0.350
12-90
0.320
0.330
0.340
10-91
0.330
0.325
8-92
0.335
0.360
0.345
-------
REHEATER TUBE #24
UT MEASUREMENTS
DATE
REAR
MIDDLE
FRONT
6-90
0.340
0.330
0.340
12-90
0.340
0.350
0.315
10-91
0.300
0.325
0.315
8-92
0.300
0.340
0.325
C-23
-------
SUPERHEATER CORROSION PROBES WEIGHT LOSS DATA
INCLUDES TIME OF EXPOSURE AND TEMPERATURE
METAL TEMPERATURE OF TEST RINGS
Ring Ncx
Material
Weight Loss
in Grams
Weight Loss
Gram/in2
Temp. °F
Length of
Exposure
202
T-22
12.179
0.98
910
2000
203
T-22
2.774
0.22
930
2000
204
T-9
1.540
0.12
950
2000
205
T-9
2.011
0.16
970
2000
206
304ss
0.414
0.03
990
2000
207
T-l 1
4.441
0.36
1010
2000
208
T-22
5.082
0.41
1030
2000
209
T-22
6.743
0.54
1050
2000
210
T-91
2.644
0.21
1070
1400
211
T-9
3.116
0.25
1090
1400
212
304ss
1.020
0.08
1110
1400
213
T-22
4.417
0.35
1130
1400
214
T-9
1.325
0.11
1150
1400
215
304ss
0.364
0.03
1170
1400
216
310ss
0.191
0.02
1190
1400
145-2
T-91
1.859
0.15
1070
600
146-2
T-9 2
1.986
0.16
1090
600
147-2
304ss
0.380
0.03
1110
600
148-2
T-22
3.566
0.28
1130
600
149-2
T-91
0.672
0.05
1150
600
150-2
304ss
0.110
0.01
1170
600
151-2
310ss
0.076
0.01
1190
600
C-24
-------
SUPERHEATER MATERIAL PERFORMANCE
o
o
X o
+ D
O +
+ +
o
+
i i i t t ° i i i i i i ° t i i £ i ft
0.9 0.94 0.98 1.02 1.06 1.1 1.14 1,18
(Thousands)
TEMPERATURE - F
a T22 + T91 O TP304 A TP310 X T11
-------
KTOMBT OTP i c7 TECHNICAL REPORT DATA
MKHKL. Kit" 13/ (Please read Instructions on the reverse before compU III IIINI lll!ll;|||| || |||||| 11
1. REPORT NO. 2.
E PA-600/R-99-063
3-f. in iiiiiimilium iimum
H PB99-15640A iAaO rv. 0
4. title ANDsueTiTLE]3emorlst]fation of Natural Gas Reburn
for NOx Emissions Reduction at Ohio Edison Com-
pany's Cyclone-fired Niles Plant Unit No. 1
5. REPORT DATfc i ft
July 1999 v
6. PERFORMING ORGANIZATION COOE
7. author(S! w, Borio, R. D. Lewis, and R. W. Koucky
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
ABB Power Plant Laboratories
Windsor, Connecticut 06095-0500
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 9/90-12/96
14. SPONSORING AGENCY CODE
EPA/600/13
16. supplementary notes i\ppcD project officer is Robert E. Hall, Mail Drop 65, 919/541-
2477.
te.abstractThe report describes a demonstration of reburning on a cyclone-fired boi-
ler. The project included a review of reburn technology, aerodynamic flow model
testing of reburn system design concepts, design and construction of the reburn sys-
tem, parametric performance testing, long-term load dispatch testing, and boiler
tube wall thickness monitoring. The report also contains a description of Ohio Edi-
son's Niles No. 1 host unit, a discussion of conclusions and recommendations deri-
ved from the program, and tabulations of data from parametric and long-term tests.
A primary focus of the report was to document performance of the Niles Unit No. 1
when employing natural gas reburning, but it was equally as important to be able to
use the information to make technical and economic judgements on the application of
natural gas reburning to the entire family of cyclone boilers. Performance of the re-
burn system at Niles might not represent the highest nitrogen oxide (NOx) reduction
possible on cyclone boilers because of differences in design and operating parame-
ters, including boiler size and mode of operation (e.g., cyclic vs base-loaded). Lar-
ger units are expected to have greater NOx reduction, both percent reduction and
total NOx removed, than those found at Niles.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSAT1 Field/Group
Pollution Nitrogen Oxides
Coal Emission
Combustion
Afterburning
Natural Gas
Boilers
Pollution Control
Stationary Sources
Reburning
Cyclone Boilers
13B 07B
21D 14 G
2 IB
13A
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Reportf
Unclassified
21. NO, OF PAGES
255
20. SECURITY CLASS (This page}
Unclassified
22. PRICE
EPA Form 2220-1 (9-731
C-26
------- |