United States EPA~600/7-81-l22a
Environmental Protection
July 19 81
Research and
Development
COMBUSTION MODIFICATION CONTROLS
FOR STATIONARY GAS TURBINE
Volume I. Environmental Assessment
Prepared for
Office of Air Quality Planning and Standards
Prepared by
industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-81-122a
July 1981
COMBUSTION MODIFICATION CONTROLS FOR STATIONARY GAS TURBINE
VOLUME I: ENVIRONMENTAL ASSESSMENT
By
R. Larkin, H. I. Lips, R. S. Merrill, and K, J. Lim
Acurex Corporation
Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
Acurex Final Report FR-79-18/EE
Prepared for
EPA Project Officer -- J. S. Bowen
Combustion Research Branch
Energy Assessment and Control Division
Industrial Environmental Research Laboratory
Research Triangle Park
North Carolina 27711
Contract 68-02-2160
-------
ABSTRACT
The report provides an environmental assessment of combustion modifica-
tion techniques for stationary gas turbines with respect to NOX control
effectiveness, operational impact, thermal efficiency impact, control
costs, and effect on emissions of pollutants other than NOx. Wet controls,
which inject steam or water directly into the combustion chamber, are the
only currently available methods sufficiently developed to reduce NOX
emissions below the recently recommended New Source Performance Stan-
dard of 75 ppm NO2 at 15 percent C>2 for clean fuels (greater than 50 per-
cent reduction). However, the effectiveness of wet controls decreases
significantly as the percentage of fuel-bound nitrogen increases. Emissions
of unburned hydrocarbons and CO can increase with wet controls; however,
a detailed level 1 test on a 60-MW utility gas turbine indicated that incre-
mental emissions other than NOx remained relatively unchanged. Wet con-
trols increase the cost of electricity by 2 to 5 percent due, in large part,
to the associated fuel penalty. Dry NOX controls are under development,
based on combustor modifications that do not involve water or steam injec-
tion. They hold much promise because of their NOX control effectiveness
for both clean and dirty fuels, and their expected lower operational and
cost impacts.
ii
-------
PREFACE
This is the third in a series of five process engineering reports
to be documented in the "Environmental Assessment of Stationary Source
NO Combustion Modification Technologies" (NO EA). Specifically,
/\ /\
this report documents the environmental assessment of stationary gas
turbines, with primary emphasis on NO combustion controls. The NO
X X
EA, a 36 month program which began in July 1976, is sponsored by the
Combustion Research Branch of the Industrial and Environmental Research
Laboratory of EPA (IERL-RTP). The program has two main objectives: (1)
to identify the multimedia environmental impact of stationary combustion
sources and NO combustion modification controls applied to these
y\
sources, and (2) to identify the most cost-effective, environmentally
sound NO combustion modification controls for attaining and maintaining
/\
current and projected NO^ air quality standards to the year 2000.
The NO EA will assess the following combination of process
/\
parameters and environmental impacts:
t Major fuel combustion stationary NO sources: utility
rt
boilers, industrial boilers, gas turbines, internal combustion
(1C) engines, and commercial and residential warm air
furnaces. Other sources (including mobile and noncombustion)
will be considered only to the extent that they are needed to
determine the NO contribution from stationary combustion
^
sources.
• Conventional and alternate gaseous, liquid and solid fuels
• Combustion modification NO controls with potential for
A
implementation to the year 2000; other controls (tail gas
cleaning, mobile controls) will be considered only to estimate
the future need for combustion modifications
m
-------
® Source effluent streams potentially affected by NO controls
« Primary and secondary gaseous, liquid and solid pollutants
potentially affected by NO controls
X
• Pollutant impacts on human health and terrestrial or aquatic
ecology
To achieve the objectives discussed above, the NO EA program
/\
approach is structured as shown schematically in Figure P-l. The two
major tasks are: Environmental Assessment and Process Engineering
(Task B5), and Systems Analysis (Task C). Each of these tasks is designed
to achieve one of the overall objectives of the NO EA program cited
/\
earlier. In Task B5, of which this report is a part, the environmental,
economic, and operational impacts of specific source/control combinations
will be evaluated. On the basis of this assessment, the incremental
multimedia impacts from the use of combustion modification NO controls
/v
will be identified and ranked. Task C will in turn use the results of
Task 85 to identify and rank the most effective source/control
combinations to comply, on a local basis, with the current NO- air
quality standards and projected N02 related standards.
As shown in Figure P-l, the key tasks supporting Tasks 85 and C are
Baseline Emissions Characterization (Task 81), Evaluation of Emission
Impacts and Standards (Task B2), Experimental Testing (Task B3), and
Source Analysis Modeling (Task D). The arrows in Figure P-l show the
sequence of subtasks and the major interactions among the tasks. The oval
symbols identify the major outputs of each task. The subtasks under each
main task are shown on the figure from the top to the bottom of the page
in roughly the same order in which they will be carried out.
As indicated above, this report is a part of the Process
Engineering and Environmental Assessment Task. The goal of this task is
to generate process evaluations and environmental assessments for specific
source/control combinations. These studies will be done in order of
descending priority. In the first year of the NO EA, all the sources
A
and controls involved in current and planned NO control implementation
programs were investigated. The "Preliminary Environmental Assessment of
Combustion Modification Techniques" (Reference P-l) documented this effort
IV
-------
tmmam mncnuta
CMAHACTIMmiHM ml (TAHOAW1 («l
UIHHIU COMMMJTKM
(OtMCf MOCIUtMMIOM
•ACKOmuNO
1 * 1
Of MHATt HUUHMOIA
CWlSfOM MVfHTOm
HCOHOAMT MM.fl«f«A
AUtKMI MH.THWOIA
•NVMOMMHTAt OOAl*
IMIG1I
1 ' 1 *
1 1
nojccnom COHHU
•lOIONAL VAMATION*
1
f unworn to MO*
* ^
f MMlMt M*ACT A
1 AMfUHCHTIIUMIMa 1
•
(mure •"•ttiom.
>"">j«cnotn
1
•H>ACT AlMSWCMt
MKMfCTHOf
t
f MMCT cnrrim*: ]
i (TANOAUDi raoNcnom 1
1 4
JUtTWT AHO
im>AT( MO*
1
un»Tt HOI
OOAl*
r
(IPIMMHTAl MMHICI AHAtOIS
nstmo !•« wnnrima mi
wrim swrtiic/
AHALIOIt
•EOUHWMmt
i
COHOUCT
rmo
maam
. 1 CMfMCAU 1
_L
ocmor utirnoi 10
coHPAflf rrriucNf
CONCEHTIIATHim TO UtO*
mil IMTACT rvAIUATIOM
t
AtHMBLC rOttUtANT
r AC ton MODFIS
OU1LMC mOCEDUdtS FOH
fCMFCHINn.
sift evAiu*rtoM
t
KMtlUM ANO^\ f MOHt rODMAI ^\
COHTIIOIUO | I oturiowif m »• IA |_^
ntwiioHt 1 1 W"'" *«"' y,
OATA i \««MU J
*
1
noctn tMamctiima AMD iwtciM
NVMONMFNfAt ASACASMtNf |»9| AMAttSIB «C|
COUPKt NOi COHfftOL
Miner as ftACNflHOUNO
t
IVAtUAIf WCflfmNIAL
fWISSIOHS OATA WITH
HOm CON T«fn 9
i PHtlMIMAIIV ftOtmcl/ 1
"*) CONIHOifPOLtUtAHT 1
•>
i
otnior MOCII*
t
COmf MIAHtO (OUHCC'
CONINOi mOCI9$ OATA
| j 1
CONDUCT OCTAHtD
MOCC99
•11HIV FOfl
ImouyraiAt •o^tris
UAS TUflBmtft
nrSIOf NtlAl HtATHHI
mnu5iniAi moctss
ADMANI rn COMBUST ion
1,
•OKI roil
immoNttrHlAl.
AltfHHAIIVf
t
•CKtlH CONIKOl
moimf MTHII ron
ATTAmiHOIWAIHTAINMa
AH OUAlItT
J
^^" . ^^-v
t AtwVtMFMI OF 1
-*\ COMH«KHOH SOUMCtt r*
V AHII NO* CONI«m • /
1
•fit CT AHD AOAVT
WACim AM OUAIITT
•OOtL
t
roojtCT tou»ct omtrra
AHO AHOIf HI lt*ND»D>
*
ran VAUious KFOUIATOIIT
HtOUMCWHIl
r m»T mtCTrn COHT«OI\
OPTIOMl COM»nt AHD 1
v01""""*" J
A 7n?in
i»f HAH
1 Kraut
\ JTOHM
* I from
Figure P-l. NOX EA approach.
-------
and established a priority rankings based on source emission impact and
potential for effective N0x control, to be used in the current ongoing
detailed evaluation.
This report presents the assessment of combustion modification
NO controls for the third source category to be treated, gas turbines.
Other environmental assessment reports documented are:
t Environmental Assessment of Utility Boiler Combustion
Modification NO Controls (Reference P-2)
/\
• Environmental Assessment of Industrial Boiler Combustion
Modification NO Controls (Reference P-3)
X
• Environmental Assessment of Combustion Modification Controls
for Stationary Internal Combustion Engines (Reference P-4)
• Environmental Assessment of Combustion Modification Controls
for Residential and Commercial Heating Systems (Reference P-5)
REFERENCES FOR PREFACE
P-l. Mason, H. B., et al., "Preliminary Environmental Assessment of
Combustion Modification Techniques: Volume II: Technical Results,"
EPA-600/7-77-119b, NTIS-PB 276 581/AS, October 1977.
P-2. Lim, K. J., j2t £l_., "Environmental Assessment of Utility Boiler
Combustion Modification NO^. Controls. Volume 1: Technical
Results. Volume 2: Appendices," EPA-600/7-80-075a and b, April
1980.
P-3. Lim, K. J., et jiJL, "Industrial Boiler Combustion Modification NOX
Controls: Volume I: Environmental Assessment," EPA-600/7-81-126a
July 1981.
P-4. Lips, H. I. et_£l_., "Environmental Assessment of Industrial Boiler
Combustion Modification Controls for Stationary Internal
Engines," EPA-600/7-81-127, July 1981.
P-5. Castaldini, C., jrta/L, "Combustion Modification Controls for
Residential and Commercial Heating Systems: Volume I-
Environmental Assessment," EPA-600/7-81-123a, July 1981
\M
-------
TABLE OF CONTENTS
Section Page
PREFACE iii
LIST OF ILLUSTRATIONS x
LIST OF TABLES xiii
1 EXECUTIVE SUMMARY 1-1
1.1 Overview of Stationary Gas Turbines 1-2
1.2 Emissions and Fuels 1-4
1.3 Status of Environmental Protection Alternatives . 1-5
1.4 Data Needs and Recommendations 1-7
2 INTRODUCTION 2-1
2.1 Background 2-1
2.2 Role of Gas Turbines 2-3
2.3 Objective of this Report 2-5
2.4 Organization of this Report 2-6
3 SOURCE CHARACTERIZATION
3.1 Technical Overview of Gas Turbines 3-2
3.1.1 Status of Development 3-2
3.1.2 The Future for Gas Turbines 3-3
3.1.3 Load Use 3-8
3.1.4 Fuels and Emissions 3-8
3.1.5 Capital and Operating Costs 3-11
3.2 Process Description 3-11
4 CHARACTERIZATION OF INPUT MATERIALS, PRODUCTS, AND
WASTE STREAMS 4-1
4.1 Input Materials 4-1
4.1.1 Gas Fuels 4-1
4.1.2 Liquid Fuels 4-5
4.1.2.1 Physical Properties 4-7
4.1.2.2 Chemical Properties 4-7
4.1.2.3 Fuel Treatment 4-8
4.2 NOX Formation 4-11
4.2.1 Thermal NOX Formation 4-11
4.2.2 Fuel NOX Formation 4-11
4.2.3 Fuel Potential for NOX Production 4-13
4.2.4 Products Characterization 4-23
vii
-------
TABLE OF CONTENTS (Continued)
Section Page.
4.3 Emissions Characterization ............ 4-23
4.3.1 NOX Emissions .................. 4-23
4.3.2 S02 Emissions .................. 4-25
4.3.3 Participate and Visible Emissions ........ 4-25
4.3o4 Hydrocarbons and Carbon Monoxide Emissions. . . . 4-25
5 PERFORMANCE AND COST OF CONTROL ALTERNATIVES ...... 5-1
5.1 Procedures for Evaluating Control
Alternatives .................. 5-1
5.2 Performance and Cost of NOX Control
Alternatives ................... 5-4
5.2.1 Control Techniques amd Fuels .......... 5-4
5.2.1.1 Wet Control Techniques and Clean Fuels .... 5-4
5.2.1.2 Wet Control Techniques and Dirty Fuels .... 5-12
5.2.1.3 Dry Control Techniques and Clean Fuels .... 5-12
5.2.1.4 Dry Control Techniques and Dirty Fuels .... 5-27
5.2.1.5 Summary .................... 5-34
5.2.2 Incremental Emissions of Pollutants Other
than NOX ..................... 5-35
5.2.2.1 Carbon Monoxide and Unburned Hydrocarbons . . . 5-35
5.2,2.2 Particulate Emissions ............. 5-40
5.2.2.3 Sulfates .................... 5_41
5.2.2.4 Organics. . .................. 5-41
5.2.2.5 Trace Elements ................. 5.43
5.2.2.6 Summary of Incremental Emissions ...... '. 5-43
5.2.3 Incremental Costs of NOX ............. 5.43
5.2.3.1 Wet Controls .............. 5..44
5.2.3.2 Dry Controls . ........... . . . . . 5-51
5.3 Regional Considerations Affecting Control
Selection
5.3.1 Fuel Considerations ............. c;_52
5.3.2 Equipment Considerations ......... '.'.'" 5~5?
5.4 Summary of Performance and Costs of NOX
Controls, .............. .... 5 54
5.4.1 Wet Controls
5.4.2 Dry Controls
b-bb
viii
-------
TABLE OF CONTENTS (Concluded)
Section Page
5.5 Performance and Cost for Pollutant Controls
Other Than NOX 5-58
6 ENVIRONMENTAL ASSESSMENT 6-1
6.1 Environmental Impact Analysis 6-3
6.2 Environmental Impacts on Air 6-13
6.2.1 Summary of NOX Control Regulations 6-3
6.2.2 Comparison of Emissions to Standards 6-10
6.2.3 Ambient Air Impact 6-11
6.2.4 Bioassay Results 6-12
6.3 Additional Impacts 6-13
6.4 Assessment of Impacts of NOX Controls on
Operations and Maintenance 6-13
6.4.1 Wet Controls for NOX Reduction 6-14
6.4.2 Dry Controls for NOX Reduction 6-16
6.5 Economic Impact of NOX Controls 6-18
6.5.1 Impact on Manufacturers 6-18
6.5.2 Impact on Consumer 6-20
6.6 Effectiveness of NOX Controls 6-22
7 DATA BASE EVALUATION AND NEEDS 7-1
7.1 Data Base Review 7-1
7.2 Data Needs 7-4
Ix
-------
LIST OF ILLUSTRATIONS
Figure
P-l NOX EA Approach .................... v
1-1 Projected Gas Turbine Generating Additions ...... 1-
2-1 Distribution of Stationary Anthropogenic NOX
Emissions for the Year 1974 (Stationary Fuel
Combustion: Controlled NOX Levels) .......... 2"
3-1 Gas Turbine Generating Additions ........... 3~4
3-2 Projected Gas Turbine Generating Additions ...... 3-5
3-3 Basic Simple Cycle Gas Turbine ............ 3-12
3-4 Typical Regenerative Cycle Gas Turbine ........ 3-14
3-5 Typical Combined Cycle Gas Turbine .......... 3-15
3-6 Comparison of Efficiency vs. Pressure Ratio for Simple
and Regenerative Cycle Engines ..... ....... 3-17
4-1 Ash Bearing Fuels Treatment .............. 4-10
4-2 Conversion of Fuel Bound Nitrogen to NOX in a
Combustor ........ . . 1 ............ 4-12
4-3 NOX Emissions as a Function of Combustor Exit
Temperature ......... . ............ 4-14
4-4 NO Formed in or very near the Primary Reaction Zone of
Ethyl ene Flames .................... 4-15
4-5 Temperature Distribution Maps of a Laboratory Scale Gas
Combustor ....................... 4-16
4-6 Nitric Oxide Concentration Maps of a Laboratory Scale
Gas Combustor ..................... 4-17
4-7 Effect of Residence Time on Nitrogen Oxides Emissions
for a Lean Primary Combustor Propane Fuel Inlet
Mixture Temperature, 800K; Inlet Pressure, 5.5 atm;
Reference Velocity, 25 and 30 m/s . . . . ....... 4-19
4-8 Predicted NOX Emission Levels of Various Fuels
Burning in Gas Turbine Combustors ........... 4-20
-------
LIST OF ILLUSTRATIONS (Continued)
Figure Page
4-9 NOX Emission Trends of a Hybrid Combustor Versus a
Conventional Combustor 4-22
4-10 Relationship Between Nitric Oxide Emissions and Total
NOX for Propane Combustion. Inlet Pressure, 5.5 atm;
Reference Velocity, 25 m/s; Equivalence Ratio, 0.40 to
1.0 4-22
4-11 CO Emissions vs. Turbine Size for Large Gas Turbines
Without NOX Controls when Operated at or near Full
Load 4-26
4-12 HC Emissions vs. Turbine Size for Small Gas Turbines
Without NOX Controls when Operated at or near Full
Load 4-27
5-1 NOX Emission Index Versus Gas Turbine Combustor
Inlet Temperature 5-6
5-2 Summary of NOX Emission Data from Gas Turbines Using
Wet Control Techniques 5-7
5-3 Effectiveness of Water/Steam Injection in Reducing NOX
Emissions 5-8
5-4 Mechanical Components Schematic 5-9
5-5 Electrical Control Schematic 5-9
5-6 Water Treatment System 5-11
5-7 Predicted Decrease in NOX Emissions through Water
Injection with Increasing Amounts of Bound Nitrogen
in Fuel Oil (Reference 5-9) 5-13
5-8 VAB Combustor Details 5-17
5-9 JIC Combustor Details 5-17
5-10 VAB Combustor Test Results -- Initial Configuration . . 5-19
5-11 VAB Combustor Test Results — Modified Fuel
Injection 5-19
XI
-------
LIST OF ILLUSTRATIONS (Concluded)
Figure Page
5-12 JIC Combustor NOX Test Results 5-20
5-13 Staged Premixed Combustor .... 5-22
5-14 NOX Emissions in Baseline Premix and Conventional
Combustors 5-23
5-15 NOX Emissions in Premix Combustor Configurations . . 5-23
5-16 NOX Emissions Comparison Corrected to 0 Percent
Excess Air 5-25
5-17 Rich Burner Arrangement 5-26
/
5-18 Rich Burner Characteristics, 345 kPa (50 psia),
590K (600°F), 0.5% Nitrogen 5-30
5-19 Rich Burner Simulated Engine Cycle Characteristics,
1030 kPa (150 psia), 0.5 percent Nitrogen 5-32
5-20 Variation of Minimum NOX Emissions with Primary
Zone Residence Time 5-33
5-21 Lean Burner High Pressure Characteristics,
689 kPa (100 psia) 5-38
5-22 Effect of Staged Combustion and Fuel Rich Combustion on
NGX and CO Emissions 5-39
5-23 Gas Turbine Particulate Emissions as a Function of
Load „ 5-42
xii
-------
LIST OF TABLES
Number
3-1
3-2
4-1
4-2
4-3
4-4
4-5
5-1
5-2
5-3
5-4
5-5
5-6
5-7
5-8
5-9
5-10
5-11
Forecasted Generating Additions For All
Generator Types Through 1986 (Reference 3-3)
1974 Gas Turbine Fuel Consumption (EJ)
Typical Properties of Common Gaseous Fuels
Typical Analysis and Properties for Low BTU Fuels. . .
Detailed Specifications for Fuel Oils3
Selected Fuel Constituents
Gas Turbine Criteria Pollutant Emission
Factors (ng/J)
Typical Water Quality Specifications
VAB Design Modifications
JIC Combustor Configurations
Model Gas Turbine Data Summary:
Catalytic Combustor
Effect of Water Injection NOX Control on
CO Emissions From a Gas Turbine
Effect of Water Injection NOX Control on UHC
Emissions From a Gas Turbine
1973 Water Injection Investment Cost (San Diego Gas)
and Electric
1975 Water Injection Cost (City of Pasadena)
Water and Stream Injection Costs
(City of Glendale)
Estimated Operating Cost of Uncontrolled Utility
Gas Turbine (50 MW)
Estimated Cost of Water Injection for Utiity
Gas Turbines
Page
3-6
3-10
4-3
4-4
4-6
4-9
4-24
5-12
5-20
5-21
5-28
5-37
5-37
5-46
5-46
5-47
5-49
5-49
xiii
-------
LIST OF TABLES (Concluded)
Number
5-12
5-13
6-1
6-2
6-3
6-4
6-5
6-6
6-7
Estimated Costs for Uncontrolled Industrial
Gas Turbine
Estimated Costs of Water Injection for Industrial
Gas Turbine
Emissions From a 60MW Utility Oil-Fired Gas Turbine . .
Emissions with Concentrations Greater Than DMEG
Values
Summary of State and Local NOX Emissions Standards
For Stationary Sources6
Degree of Wet Control Required to Meet Regional
NOX Standards
Projected Control Requirements for Alternate
NOX Emission Levels
Predicted Economic Impact of Wet NOX Controls
to Meet Proposed NSPS for Utility Gas Turbines:
1978-1983a
NOX Controls: Best Available Control Technology
(BACT) and Advanced Technology
5-50
5-51
6-4
6-5
6-6
6-10
6-19
6-21
6-23
XIV
-------
SECTION 1
EXECUTIVE SUMMARY
This is the third in a series of five special reports to be
documented in the "Environmental Assessment of Stationary Source NOV
A
Combustion Modification Technologies" (NO EA). Specifically, this
J\
report documents the environmental assessment of stationary gas turbines,
with primary emphasis on NO combustion controls. The program has two
A
main objectives: (1) to identify the multimedia environmental impact of
stationary combustion sources and NO combustion modification controls
^
applied to these sources, and (2) to identify the most cost-effective,
environmentally sound NO combustion modification controls for attaining
A
and maintaining current and projected N02 air quality standards to the
year 2000.
With more NO controls being implemented in the field and
n
expanded control development anticipated for the future, there is
currently a need to: (1) ensure that the current and emerging control
techniques are technically and environmentally sound and compatible with
efficient and economical operation of systems to which they are applied,
and (2) ensure that the scope and timing of new control development
programs are adequate to allow stationary sources of NOV to comply with
A
potential air quality standards. The stationary gas turbine EA helps to
address these needs by evaluating the operational, economic and
environmental impacts from applying combustion modification NO controls.
y\
Gas turbines are the fifth largest contributors of NO emissions
J\
from stationary anthropogenic sources in the U.S. — constituting a
2.0 percent share (Reference 1-1). A variety of factors including fuels,
electricity demand and increasing thermal efficiencies, will tend to
intensify the NOX problem from stationary gas turbines. Given this
1-1
-------
background and their potential for NOX control, stationary gas turbines
have been selected as one of the major source categories to be treated
under the NOX EA program.
1.1 OVERVIEW OF STATIONARY GAS TURBINES
Gas turbines are rotary internal combustion engines commonly,
although not universally, fired with natural gas or "clean" liquid fuels
such as diesel or distillate oils. The basic gas turbine consists of a
compressor, combustion chamber(s) and a turbine. Pressurized combustion
air, supplied by the compressor, and fuel are burned in the combustion
chambers. The hot combustion gases are rapidly quenched in the combustor
by secondary dilution air and then expanded through turbines which drive
the compressor and provide shaft power to, e.g., a generator, compressor,
or pump.
Three different thermodynamic cycles are typically used in
stationary gas turbine engines — simple, regenerative and combined
cycles. The simple cycle is the basic gas turbine engine while the
regenerative and combined cycles employ some form of exhaust waste heat
recovery. Throughout this report, stationary gas turbines have been
divided for analysis purposes into three capacity ranges (power output):
• Large capacity, including combined cycle, >15 MW (20,000 hp)
• Medium capacity, 4 MW to 15 MW (5,000 to 20,000 hp)
• Small capacity, <4 MW (5,000 hp)
Gas turbines have enjoyed spectacular sales growth through 1970 due
primarily to their inherent low cost and operational and maintenance
advantages over other prime movers and electrical generation types. A
growing economy combined with delays in nuclear plant licensing also
contributed to their popularity. With the 1970's came decreased oil
availability and a growing uncertainty among users concerning the
reliability of gas turbines, causing a subsequent steady decline in
sales. In addition, forecasts of new generating requirements by the
National Electrical Manufacturers Association (NEMA) have shown
substantial reductions over previous forecasts of gas turbine equipment
Figure 1-1 shows the Sixth Biennial Survey of Power Equipment Requirements
(Reference 1-2). The gas turbine generating additions predicted in 1978
have decreased 78 percent from NEMA's 1973 predictions. Although the
1-2
-------
8
SPER-71
SPER — survey of power equipment
requirements by NEMA-biannual
projections
to
en
•r—
CD
SPER-73
SPER-69
SPER-75
SPER-77
68
72
76
84
Figure 1-1. Projected gas turbine generating additions.
(Reference 1-2)(Reproduced by permission of the National
Electrical Manufacturers Association from the Sixth Biennial
Survey of Power Equipment Requirements of the U.S. Electric
Utility Industry 1977-1986, NEMA, PE-S-6-1978.)
-------
immediate future does not look bright, manufacturers are optimistic about
en upswing in the market, particularly for combined cycle plants.
1,2 EMISSIONS AND FUELS
Air emissions in the form of exhaust gases are essentially the only
effluent stream from stationary gas turbines. Stream composition depends
highly on the fuel burned, combustor geometry, combustion and operating
characteristics. NO emissions are highest and CO and unburned
/\
hydrocarbons (UHC) are lowest when the engine operates at design
conditions (i.e., rated power output). Off-design firing, while limiting
NOX, enhances the production of unburned species through incomplete
oxidation. Virtually all fuel sulfur is converted to sulfur dioxide
(S02) in a turbine engine, the concentration being purely a function of
the fuel sulfur content. Using low sulfur fuel is the only viable means
to control S02 emissions.
The only liquid and solid wastes from gas turbines are from the
water treatment facilities associated with water injection for NOV
A
control. These effluent streams are relatively small, generally
innocuous, and easily disposed of in landfill areas or to rivers or
municipal sewers. As dry NO controls gradually replace wet controls,
/\
these liquid and solid discharges will no longer exist.
Natural gas and distillate oils are the preferred fuels for gas
turbines because they are relatively clean burning. Those oils containing
significant ash and trace element concentrations, such as crude oil,
residual oil, and synthetic fuels may require some treatment before they
can be used.
NOX is one of the primary pollutants of concern with stationary
gas turbines. It is emitted in relatively large quantities, is
deleterious to human respiratory functions and acts as a key precursor to
photochemical smog. NOX in gas turbines, as in all combustion sources,
is formed primarily by two mechanisms ~ thermal fixation and fuel NO
formation. Thermal NOX results from the thermal fixation of molecular
nitrogen and oxygen in the combustion air and the rate of formation
increases exponentially with local flame temperature. Fuel NO
from the oxidation of organically bound nitrogen found in certain fuels
such as residual oil, and primarily depends on the nitrogen content of the
fuel. In general, liquid fuels yield a higher N0x emission level than
1-4
-------
gaseous fuels. This is due primarily to higher localized flame
temperatures resulting from droplet burning and, to some extent, on the
amount of fuel nitrogen.
Emissions of carbon monoxide (CO) and (UHC) are related to the
operating cycle and generally inversely follow load. For example,
decreasing load reduces NO , but increases CO and UHC concentrations.
A
Thus in all control development efforts, a balance must be maintained
between NOX and CO and UHC.
Particulate emissions from gas turbines are a function of the ash
content of the fuel, and the levels of carbon and unburned hydrocarbons
due to incomplete combustion. Thus, particulate emissions will increase
with the use of "dirty" fuels and as combustion becomes less efficient.
1.3 STATUS OF ENVIRONMENTAL PROTECTION ALTERNATIVES
Wet controls, which inject steam or water directly into the
combustion chamber, are the only currently available methods sufficiently
developed to reduce gas turbine NOX emissions below the recently
promulgated (1980) NO New Source Performance Standard (NSPS — 75 ppm
/\
at 15 percent oxygen for clean fuels) and more stringent local
regulations. Wet controls work on the principle of thermal quenching —
effective lowering of peak flame temperatures thereby reducing NO
/\
generation. The required degree of control is obtained by altering the
water-to-fuel ratio. However, depending on the water-to-fuel ratio and
operating conditions, emissions of unburned hydrocarbons and CO may
increase. Furthermore, the effectiveness of wet controls decreases
significantly as the percentage of fuel-bound nitrogen in the fuel
increases.
Dry NOX controls generally refer to combustor modifications and
do not involve water or steam injection. A number of general concepts are
currently being examined by dry control developers, but one in particular
is emerging as the most promising. The Pratt and Whitney Aircraft Group
under EPA sponsorship has recently completed demonstration of a new
combustor concept for stationary gas turbines known as rich burn/quick
quench (R8QQ), in a full-scale engine. NOV emissions were well below
/\
NSPS for both clean and dirty fuels, and CO was simultaneously held to low
levels.
Other organizations are developing different dry control concepts
with various degrees of success. Many of these programs are being
1-5
-------
conducted with meeting the aircraft jet engine emissions standards in
mind, thus the designs and dry control concepts are being evaluated for
aircraft engines which burn clean jet fuels. There is some disagreement
as to whether this approach to dry NOY controls will be adaptable to
A
stationary gas turbines designed to fire a variety of fuels.
Implementing wet NOX controls can significantly impact the total
operating cost of a stationary gas turbine, which varies greatly among
users. Various utilities have reported capital costs ranging from $5/kW
to almost $23/kW in 1978 dollars. A typical utility gas turbine will
cost, in total, approximately $150/kW in 1978 dollars. Actual costs are
very site specific and depend to a great extent on the required water
purification equipment and to a lesser extent on turbine modifications.
Also, a fuel penalty resulting from an increased heat rate with water
injection is a significant portion of this increased expense. The total
annualized cost of wet controls, including capital and operating costs,
raises the cost of electricity by about 2 to 5 percent.
At this stage of development of dry controls, it is difficult to
accurately predict their associated costs. Whether development costs are
passed on to the user and the market size over which these costs can be
spread, will be major factors in determining the magnitude of dry control
costs. It appears that dry NO controls which have development
/\
expenditures included in their total cost will cost somewhat less than wet
NOX controls for a comparably sized unit. If development costs are not
passed on, dry control combustors are expected to be only nominally more
costly than existing combustor models.
In summary, wet controls are currently the preferred, indeed the
only commercially demonstrated, NO control techniques which meet the
NOY NSPS for stationary gas turbines. Although the capital costs of
J\
water injection can be significant, approximately 10 percent of total unit
cost on a per kW basis, annualizing that cost along with operating costs
over the lifetime of the unit results in an incremental cost of
electricity of only 2 to 5 percent.
Dry controls, as they become fully developed for full size engines,
will begin to replace wet controls in new sources. They appear to be more
attractive from virtually all standpoints, including economics. Catalytic
1-6
-------
combustion, while yielding extremely high pollutant reductions in subscale
models, is only feasible for production engines in the very long term.
1.4 DATA NEEDS AND RECOMMENDATIONS
A substantial amount of good quality .data regarding wet control
emissions reduction was collected by EPA for use in the Standard Support
and Environmental Impact Statement (SSEIS — Reference 1-3). Results from
additional emissions testing programs since the completion of the SSEIS
have been added to the data base. These data support the general trends
reported in the SSEIS regarding NOX emission reductions and the
relationships between percent NO reduction and water-to-fuel ratio.
A
A significant amount of disparity was found in both capital and
operating cost estimates for wet controls from users, manufacturers, and
the EPA. Other than for large scale gas turbines where a few users could
supply actual costs, data in this report are primarily manufacturer and
EPA cost estimates. Even for large scale units, data were sparse since
most users are not yet required to control NO
/\
Given the existing state of development of dry NOY controls, it
/\
is not possible to perform a comprehensive environmental analysis. This
report relies heavily on experimental combustor program literature and
some industry contacts to qualitatively assess the postulated effects of
the most promising dry control concepts. Since much of the developmental
work is proprietary, the information presented is incomplete.
Furthermore, the work conducted so far has generally been performed on a
single combustor rig. Significant research and development will be
required before these dry control concepts can be reliably applied to
actual engines.
The available data for all dry NOX control techniques only
provide qualitative judgements and estimates when predicting costs,
incremental emissions, and operations and maintenance impacts. In
assessing this report's evaluation, it should be noted that dry controls
are an emerging technology and are at least two to four years from the
first full scale application. Numerous changes in the preferred design,
which could have a major impact on the predicted environmental effects,
may occur.
As users and manufacturers gain experience, additional emissions
data should be gathered and periodically evaluated during the life cycle
1-7
-------
of existing, modified and new -stationary gas turbines. This is necessary
because dry NO control techniques are just evolving as a technology and
/\
there is a considerable lack of long term operating experience with wet
controls.
Specific areas where opinions differ regarding impacts due to wet
controls include:
• Water injection cost data for capital equipment, operating and
maintenance expenses
• The cost/benefit ratio of wet controls for small gas turbines
(<4 MW) (Note: small engines have a 5 year exemption from
NSPS to develop dry controls)
• Quantification of the fuel penalty due to increased heat rate
and, additional power output resulting from more mass
throughput.
At present, it seems clear that dry NO controls will be the
A
preferred control technology option for new gas turbines within 5 years.
However, because they are still in the development stage, very little
actual data exist regarding emission levels, control costs and operations
and maintenance impacts for the application of dry controls to full scale
engines. All of these data are required to perform a meaningful
environmental assessment of dry NO control technology. As the
y\
direction of dry controls research becomes evident, additional testing
programs can be suitably designed to provide the proper data base. Then,
as dry controls become commercially feasible and users gain operating
experience, any additional data gaps may be filled accordingly. The types
of gaps will primarily be additional operating and maintenance costs that
can only be accurately predicted through long term expense accounting.
Only with long term experience, as well as careful front-end tracking of
dry control developments, can a comprehensive environmental assessment be
performed. The task should not be extremely complicated for gas
turbines. Dry controls do not require as much ancillary equipment as do
wet controls, so in this regard, they can be considered less complex.
Also, most manufacturers' expect no additional costs over wet controls, nor
do they expect any significant operational or maintenance impacts.
1-8
-------
REFERENCES FOR SECTION 1
1-1. Waterland, L. R., et a/L, "Environmental Assessment of Stationary
Source NOX Control Technologies — Final Report," Acurex Report
FR-80-57/EE, EPA Contract 68-02-2160, Acurex Corporation, Mountain
View, CA, April 1980.
1-2. National Electrical Manufacturers Association, "Sixth Biennial
Survey of Power Equipment Requirements of the U.S. Electric Utility
Industry," NEMA, Washington, D.C., March 1978.
1-3. Goodwin, D. R., et al., "Standard Support and Environmental Impact
Statement. Volume I: Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
1-9
-------
SECTION 2
INTRODUCTION
This report assesses the operational, economic, and environmental
impacts from applying combustion modification NO controls to stationary
A
gas turbines. With more NO controls being implemented in the field and
A
expanded control development anticipated for the future, there is
currently a need to: (1) ensure that the current and emerging control
techniques are technically and environmentally sound, and compatible with
efficient and economical operation of systems to which they are applied,
and (2) ensure that the scope and timing of new control development
programs are adequate to allow stationary sources of NO to comply with
A
potential air quality standards. The NO EA program addresses these
A
needs by (1) identifying the incremental multimedia environmental impact
of combustion modification controls, and (2) identifying the most
cost-effective source/control combinations to achieve ambient NO.
standards.
2.1 BACKGROUND
The 1970 Clean Air Act Amendments designated oxides of nitrogen
(NO ) as one of the criteria pollutants requiring regulatory controls to
A
prevent potential widespread adverse health and welfare effects.
Accordingly, in 1971, EPA set a primary and secondary National Ambient Air
Quality Standard (NAAQS) for N02 of 100//g/m3 (annual average). To
attain and maintain the standard, the Clean Air Act mandated control of
new mobile and stationary NO sources, each of which currently emits
A
approximately half of the manmade NO nationwide. Emissions from light
A
duty vehicles (the most significant mobile source) were to be reduced by
90 percent to a level of 0.25 g N02/km (0.4 g/mile) by 1976. Stationary
sources were to be regulated by EPA's New Source Performance Standards
2-1
-------
(NSPS), which are set as contr-ol technology becomes available. Additional
standards required to attain air quality in the Air Quality Control
Regions (AQCRs) could be set for new or existing sources through the State
Implementation Plans (SIPs).
Since the Clean Air Act, techniques have been developed and
implemented that reduce NO emissions by a moderate amount (30 to
A
60 percent) for a variety of source/fuel combinations. In 1971, EPA set
NSPS for large steam generators burning gas, oil, and coal (except
lignite). Recently, more stringent standards for utility boilers burning
all gaseous liquid and solid fuels have been promulgated. In addition,
NSPS have been promulgated for stationary gas turbines and are currently
being considered for stationary internal combustion engines and
intermediate size (industrial) steam generators. Local standards also
have been set, primarily for new and existing large steam generators and
gas turbines, as parts of State Implementation Plans in several areas with
NO problems. This regulatory activity has resulted in reducing NO
A A
emissions from individual stationary sources by 30 to 60 percent. The
number of controlled sources is increasing as new units are installed with
factory equipped NO controls.
A
Emissions have been reduced comparably for light duty vehicles.
Although the goal of 90 percent reduction (0.25 g NOp/km) by 1976 has
not been achieved, emissions were reduced by about 25 percent (1.9 g/km)
for the 1974 to 1976 model years and in 1979 were reduced to 50 percent to
1.25 g/km. Achieving the 0.25 g/km goal has been deferred indefinitely
because of technical difficulties and fuel penalties. Initially, the 1974
Energy Supply and Environmental Coordination Act deferred compliance to
1978. Recently, the Clean Air Act Amendments of 1977 abolished the 0.25
g/km goal and replaced it with an emission level of 0.62 g/km (1 g/mile)
for the 1981 model year and beyond. However, the EPA Administrator is
required to review the 0.25 g/km goal, considering the cost and technical
capabilities as well as the need of such a standard to protect public
health or welfare. A report to the Congress is due July 1980.
Because the mobile source emission regulations have been relaxed,
stationary source NO control has become more important for maintaining
A
air quality. Several air quality planning studies have evaluated the need
for stationary source NO control in the 1980's and 1990's in view of
A
2-2
-------
recent developments (References 2-1 through 2-9). These studies all
conclude that relaxing mobile standards, coupled with the continuing
growth rate of stationary sources, will require more stringent stationary
source controls than current and impending NSPS provide. This conclusion
has been reinforced by projected increases in the use of coal in
stationary sources. The studies also conclude that the most
cost-effective way to achieve these reductions is by using combustion
modification N0x controls in new sources.
It is also possible that separate NOX control requirements will
be needed to attain and/or maintain additional NCL related standards.
Recent data on the health effects of N02 suggest that the current NAAQS
should be supplemented by limiting short term exposure (References 2-4 and
2-10 through 2-12). In fact, the Clean Air Act Amendments of 1977 require
EPA to set a short term N02 standard for a period not to exceed 3 hours
unless it can be shown that such a standard is not needed. EPA will
probably propose a short term standard in 1980 when update of the NOp
air quality criteria document (Reference 2-13) is completed
(References 2-14 and 2-15).
EPA is continuing to evaluate the long range need for additional
NOX regulation as part of strategies to control oxidants or pollutants
for which NO is a precursor, e.g., nitrates and nitrosamines
A
(References 2-4, 2-10, and 2-14 through 2-17). These regulations could be
source emission controls or additional ambient air quality standards. In
either case, additional stationary source control technology could be
required to assure compliance.
In summary, since the Clean Air Act, near term trends in NOY
y\
control are toward reducing stationary source emissions by a moderate
amount, hardware modifications in existing units or new units of
conventional design will be stressed. For the far term, air quality
projections show that more stringent controls than originally anticipated
will be needed. To meet these standards, the preferred approach is to
control new sources by using low NOX redesigns.
2.2 ROLE OF GAS TURBINES
Stationary gas turbines are the fifth largest contributor of NO
/\
emissions in the U.S. Figure 2-1 shows that 2.0 percent (on a weight
2-3
-------
Noncombustion 1.9%
Warm air furnaces 2.0%
Gas turbines 2.0%
Others 4.1%
Industrial process
heaters 4.1%
Industrial
Boilers
14.4%
Reciprocating
1C Engines
18.9%
— Incineration 0.
Utility Boilers
52.0%
Total: 10.5 Tg/yr (11.6 x 106 tons/yr)
Figure 2-1. Distribution of stationary ,_,,,
for the year 1977 (controlled
(Reference 2-18).
'x emissions
2-4
-------
.basis) of all NO emissions from stationary sources in 1977 were from
A
stationary gas turbines (Reference 2-18).
In that year, gas turbines consumed approximately 1.1 EJ or
approximately 2 percent of the total U.S. energy use of about 48 EJ
(Reference 2-18). With electricity demand continuing to increase and
manufacturers constantly striving to improve thermal efficiency in their
units (which generally favors increased NO production), NOY emissions
/\ /\
from stationary gas turbines will continue to be a problem unless adequate
controls are developed. The problem may be aggravated when the current
trend to burn "clean" fuels is reversed and fuels such as residual oils
and synthetic fuels, which show a propensity for higher NO formation,
/\
become the fuels of choice. In addition, the proposed National Energy
Plan will encourage research and development aimed at substituting coal
and coal-derived fuels for natural gas and petroleum products (Reference
2-19). The implications for gas turbines involve increased pollution
potential from use of synthetic coal liquids and gases.
Given this background and their potential NO control, stationary
X
gas turbines were selected as the third source category to be treated
under the NO EA program. The "Preliminary Environmental Assessment of
/\
Combustion Modification Techniques" (Reference 2-8) concluded that
modifying combustion process conditions is the most effective and widely
used technique for achieving 20 to 70 percent reduction in oxides of
nitrogen. Nearly all current NOX control applications use combustion
modifications. Other approaches, such as treating postcombustion flue
gas, are being evaluated in depth in other programs (Reference 2-20) for
potential future use.
2.3 OBJECTIVE OF THIS REPORT
This report provides comprehensive, objective, and realistic
evaluations and comparisons of the important aspects of the available
combustion N0x control techniques, using a common and uniform basis for
comparison. The objective is to perform an environmental assessment of
NOX combustion modification techniques for stationary gas turbines to:
• Determine their impact on the achievement of selected
environmental goals, based on a comprehensive analysis from a
multimedia consideration
2-5
-------
• Ascertain the effect of their application on turbine
performance and identify potential problem areas
• Estimate the economics of their operation
• Estimate the limits of control achievable by combustion
modification
• Identify further research and development and/or testing
required to optimize combustion modification techniques and to
upgrade their assessments
2.4 ORGANIZATION OF THIS REPORT
Evaluating the effectiveness and impacts of N0x combustion
controls applied to stationary gas turbines requires assessing their
effects on both controlled source performance, especially as translated
into changes in operating costs and energy consumption, and on incremental
emissions of other pollutants as well as NO . To perform such an
A
evaluation, it is necessary to:
• Characterize the source category with regard to equipment,
fuels, emissions, and incremental costs (Section 3)
• Identify NO formation mechanisms and relate fuels to their
emissions potential (Section 4)
• Evaluate the performance of current and developing NO
A
control techniques through examination of NO and incremental
A
emissions under controlled conditions (Section 5)
• Estimate the capital and operating costs, including energy
impacts of implementing N0x control (Section 5)
• Evaluate the environmental impact of NO controls through the
analysis of incremental emissions (Section 6)
t Assess the total impact of NO controls on ambient air
A "5
economics, energy and operations and maintenance of the
turbine, thereby evaluating the effectiveness of current and
emerging control technology (Section 6)
2-6
-------
REFERENCES FOR SECTION 2
2-1. Crenshaw, J. and A. Basala, "Analysis of Control Strategies to
Attain the National Ambient Air Quality Standard for Nitrogen
Dioxide," presented at the Washington Operation Research Council's
Third Cost Effectiveness Seminar, Gaithersburg, MO, March 1974.
2-2. "Air Quality, Noise and Health — Report of a Panel of the
Interagency Task Force on Motor Vehicle Goals Beyond 1980,"
Department of Transportation, March 1976.
2-3. McCutchen, G. D., "NOX Emission Trends and Federal Regulation,"
presented at AIChE 69th Annual Meeting, Chicago, December 1976.
2-4. "Air Program Strategy for Attainment and Maintenance of Ambient Air
Quality Standards and Control of Other Pollutants," Draft Report,
U.S. EPA, Washington, D.C., October 1976.
2-5. "Annual Environmental Analysis Report, Volume 1 Technical Summary,"
The MITRE Corporation, MTR-7626, September 1977.
2-6. Personal communication with R. Bauman, Strategies and Air Standards
Division, Office of Air Quality Planning and Standards, U.S. EPA,
October 1977.
2-7. "An Analysis of Alternative Motor Vehicle Emission Standards," U.S.
Department of Transportation, U.S. EPA, U.S. FEA, May 1977.
2-8. Mason, H. 8., et al., "Preliminary Environmental Assessment of
Combustion Modification Techniques: Volume II, Technical Results,"
EPA-600/7-77-119b, NTIS-PB 276 68I/AS, October 1977.
2-9. Greenfield, S. M., et al., "A Preliminary Evaluation of Potential
NOX Control Strategies for the Electric Power Industry," Electric
Power Research Institute, EPRI TR-13300, April 1977.
2-10. French, J. G., "Health Effects from Exposure to Oxides of
Nitrogen," presented at the 69th Annual Meeting, AIChE, Chicago,
November 1976.
2-11. "Scientific and Technical Data Base for Criteria and Hazardous
Pollutants -- 1975 EPA/RTP Review," EPA-600/1-76-023, NTIS-PB 253
942/AS, January 1976.
2-12. Shy, C. M., "The Health Implications of an Non-Attainment Policy,
Mandated Auto Emission Standards, and a Non-Significant
Deterioration Policy," presented to Committee on Environment and
Public Works, Serial 95-H7, February 1977.
2-7
-------
2-13. "Report on Air Quality Criteria for Nitrogen Oxides " AP-84,
Science Advisory Board, U.S. EPA, June 1976.
2-14. "Report on Air Quality Criteria: General Comments and
Recommendations," Report to the U.S. EPA by the National Air
Quality Advisory Committee of the Science Advisory Board, June 1976.
2-15. "Air Quality Criteria Document for Oxides of Nitrogen; Availability
of External Review Draft," Federal Register, Vol. 43, pp. 58,
117-8, December 12, 1978.
2-16. Personal communication with M. Jones, Strategies and Air Standards
Division, Office of Air Quality Planning and Standards, U.S. EPA,
September 1976.
2-17. "Control of Photochemical Oxidants — Technical Basis and
Implications of Recent Findings," EPA-450/2-75-005, July 1975.
2-18. Waterland, L. R., et al., "Environmental Assessment of Stationary
Source NOX Control Technologies — Final Report," Acurex Report
FR-80-57/EE, EPA Contract 68-02-2160, Acurex Corp., Mountain View,
CA, April 1980.
2-19. "The National Energy Plan," Executive Office of the President,
Energy Policy and Planning, U.S. Superintendent of Documents, Stock
No. 040-000-00380-1, April 1977.
2-20. Faucett, H. L., et ^1_._, "Technical Assessment of NOX Removal
Processes for Utility Application," EPA 600/7-77-127 or EPRI
AF-568, November 1977.
2-8
-------
SECTION 3
SOURCE CHARACTERIZATION
This section presents a general characterization of gas turbines as
a stationary air pollutant source to aid in understanding subsequent
sections of this report. Unlike some sources, such as utility boilers,
which are categorized by many equipment types, each requiring individual
analysis, gas turbines are distinguished more by size categories.
Consequently much of the distinction in analysis within the following
sections is by categories defined according to rated power output. This
approach is consistent with the Preliminary Environmental Assessment of
Combustion Modification Techniques (Reference 3-1) and the Standard
Support and Environmental Impact Statement for stationary gas turbines
(Reference 3-2).
Simple and combined-cycle turbines offer the greatest potential for
NOX control with dry combustion controls (the preferred NOX control
technique for the 1980s), especially considering that regenerative cycle
turbines are decreasing in popularity due to their limited efficiency and
slowness in getting up to power. Large capacity gas turbines are treated
as a separate group. The classification includes both large industrial
engines and the large aircraft derivative engines. This is a valid
grouping with respect to NOX control technology because both types
exhibit similar combustor volume constraints with retrofit low-NO
A
combustors. Due to a relatively large population, medium capacity gas
turbines are also treated separately. Since the application and
distribution of small capacity gas turbines may present localized NOV
A
problems, they too will be treated separately. These divisions also
facilitate the economic impact assessment of NOV controls.
A
3-1
-------
Fuel requirements and use played a key role in defining the focus
of this report. The primary fuels for gas turbines are the clean
fuels ~ natural gas and distillate oils. The dirtier fuels, such as
residual oils, can be fired in suitably equipped gas turbines. But their
use can cause additional operational and maintenance problems, as well as
increased emissions of NO and SOY. Their use, however, is slowly
/\ /\
becoming more attractive as clean fuels escalate in cost and become less
available. For similar reasons, synthetic coal-derived fuels are also
approaching the point where they are becoming a viable gas turbine fuel.
The remainder of this section presents an overview of the gas
turbine industry and discusses technical aspects such as load use, cycle
types and efficiencies, and capital and operating cost considerations.
3.1 TECHNICAL OVERVIEW OF GAS TURBINES
This subsection presents a broad overview of gas turbines, leaving
specific details on equipment and cycle types to the next subsection.
Detailed process descriptions can be found in Section 3.2. Topics
included here are an overview of developments and trends in the industry,
brief discussions of applications, fuels, emissions, energy efficiencies,
and costs.
Throughout this report, stationary gas turbines are typically
divided into three capacity ranges:
• Large capacity, including combined cycle, 215 MW (20,000 hp)
• Medium capacity, 4 MW to 15 MW (5,000 to 20,000 hp)
• Small capacity, <4 MW (5,000 hp)
3.1.1 Status of Development
The growth of gas turbines as prime movers since the early 1960s to
early 1970s has been exceptional. There are numerous reasons for this
acceptance, but among the more important are the following gas turbine
characteristics:
0 Ability to operate on a variety of fuels
• Short delivery times
• Fast response to load changes
0 Quiet, reliable operation
0 Capability for remote operation
0 High power to size ratio
3-2
-------
Also, much of the spectacular rise in gas turbine generating additions
shown in Figure 3-1 has been due to a shortage of generating capacity
caused by delays in nuclear licensing and by the growth in the economy
during the late 60's and early 70's (Reference 3-3). On a unit basis
growth has not been as spectacular, even decreasing in 1968 and 1971.
This reflects the fact that most sales have been to electric utilities,
who primarily use large capacity units and that gas transmission sales
have steadily declined in the United States (Reference 3-2).
Figure 3-1 shows a steady decline in gas turbine generating
additions throughout the early and mid-1970s, primarily because the fuel
availability situation changed dramatically in this time period. Fuel
price increases (including natural gas) had a particularly detrimental
effect on the appeal of gas turbines, which tend to be less efficient
engines than diesels, for example. The capital cost advantage of gas
turbines was offset by the fuel penalty. Also, availability of clean
fuels to the utility industry was questionable, further contributing to
the sales decline.
The domestic gas turbine market is dominated by a few large
companies. In fact, the three largest, General Electric Company, Turbo
Power and Marine, and Westinghouse, control 80 percent of the market.
Also within capacity groups, certain companies dominate the market. For
example, Solar Turbines International of International Harvester,
Airesearch Manufacturing Company, and Detroit-Diesel Allison Division of
General Motors control a major share of the small capacity turbine market
while G.E., Westinghouse and Turbo Power and Marine control much of the
medium and large capacity sectors. Much of the new information in this
report was derived from these manufacturers, as well as users of their
models.
3.1.2 The Future for Gas Turbines
The Sixth Biennial Survey of Power Equipment Requirements (SPER)
sponsored by the National Electrical Manufacturers Association (NEMA)
shows substantial reductions in forecasts of new generating additions over
1973 projections (Reference 3-3). Gas turbine installations have
decreased 78 percent from SPER-1973 projections as shown in Figure 3-2.
The forecasted generating additions for all types of generation are shown
in Table 3-1. Gas turbines hold an average of 2.3 percent of the total
3-3
-------
7000
6000
5000
4000
3000
2000
1000
Note: Docs not Include
combined cycle units
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
Year
Figure 3-1. Gas turbine generating additions (Reference 3-3)
(Reproduced by permission of the National Electrical Manufac-
turers Association from the Sixth Biennial Survey of Power
Equipment Requirements of the U.S. Electric Utility Industry
1977-1986, NEMA, PE-S-6-1978.)
3-4
-------
8
6 -
(0 .
I 4
O>
•r™
CO
CO
I
en
o »
SPER - Survey of Power Equipment
Requirements by NEMA,
biannual projections.
1968
Figure 3-2.
1972
1976
1980
1984
Year
Projected gas turbine generating additions
(Reference 3-3). (Reproduced by permission of the
National Electrical Manufacturers Association from the
Sixth Biennial Survey of Power Equipment Requirements
of the U.S. Electric Utility Industry 1977-1986, NEMA,
PE-S-6-1978.)
-------
TABLE 3-1. FORECASTED GENERATING ADDITIONS FOR ALL GENERATOR TYPES
THROUGH 1986 (Reference 3-3)
Fossil Steam
Nuclear
Conventional Hydro
Pumped Storage Hydro
Gas Turbine
Combined Cycle
Total Additions
1977
18,278
7,165
392
550
404
1,158
27,947
1978
15,389
7,705
1,783
1,813
1,087
613
28,390
1979
14,155
9,169
1,609
234
621
0
25,788
1980
19,670
10,284
668
345
700
1,095
32,762
1981
13,015
12,154
309
1,254
1,452
349
28,533
1982
12,633
15,031
456
503
850
430
29,903
1983
15,856
15,599
808
2,776
217
766
36,022
1984
14,895
21,848
130
800
438
439
38,550
1985
16,731
11,940
660
425
593
1,948
32,297
1986
16,982
16,911
140
672
866
736
36,307
co
i
en
Generation additions in megawatts. EE-T-005
Gas Turbine category Includes a small number of dlesel units, statistically insignificant.
Reproduced by permission of the National Electrical Manufacturers Association from the Sixth
Biennial Survey of Power Equipment Requirements of the U.S. Electric Utility Industry 1977-
1986, NEMA, PE-S-6-1978.
-------
generating additions through 1986. This value will most certainly
increase for certain locations which favor gas turbines. Utilities are
currently planning only limited generating additions through the use of
gas turbines. However, under certain circumstances utilities may again
accelerate their orders of gas turbines. For example, if nuclear and
fossil fuel plants are experiencing long delays in a time of high peak
energy needs, whether for licensing or construction reasons, a shortfall
in electrical generating capacity will develop. Gas turbines would then
be in the best position to provide that capacity.
The outlook for the gas turbine industry in the immediate future (1
to 2 years) does not look bright according to some industry sources
(References 3-4 and 3-5). Domestic sales in particular are in a depressed
state for a variety of reasons. Fears of an economic recession are one
cause while fuels availability and pricing situation are also significant
factors. Prices currently tend to favor diesel 1C engines rather than the
less fuel efficient gas turbine. The future outlook, however, is
optimistic for a reversal in this trend. Apparently utilities have put to
use nearly all the excess capacity available and most of the easy and
convenient energy conservation measures have been implemented.
Consequently utilities may soon be looking to expand their capacity. Gas
turbine sales may also increase once combustors become commercially
available that can burn high nitrogen fuels and still meet NSPS.
The outlook for combined cycle plants is particularly bright for
the industry. A 1975 Survey of U.S. Electric Utilities (Reference 3-6)
shows that almost 60 percent of the respondents favored combined cycle for
future generation. A recent study by General Electric Corporate Research
and Development also looks toward combined cycle plants for long range
capacity (Reference 3-7). Of 21 advanced energy conversion cycles studied
for their technical and economic promise for use from 1985 to 2000,
various combined cycles realized the lowest busbar energy costs. Whether
it was an air cooled engine firing low Btu gas, a water cooled engine
firing low Btu gas, or a water cooled engine firing semiclean fuel, gas
turbine combined cycles were consistently among the two most economic
cycles for base, midrange and peak power modes. In addition, the high
thermal efficiencies of combined cycle gas turbine plants makes the future
look brighter, particularly with regards to the use of synthetic fuels.
3-7
-------
3.1.3 Load Use
The reasons cited in Section 3.1.1 for the popularity of gas
turbines also reflect their flexibility. The primary use of large gas
turbines is in peak load electric-generating use. Indeed, over 90 percent
of gas turbine capacity sold in the U.S. now goes to electrical utilities,
and this percentage is continuing to increase (Reference 3-1). Under
certain circumstances, large simple cycle machines are favored for base
load uses. Combined cycles are favored by utilities for intermediate and
base load applications due to their high efficiencies and fuel
flexibility. The second largest user of large stationary gas turbines is
the gas and oil industry, where gas turbines are ideal due to their
variety of sizes and ability to run remotely and unattended for long
periods. Additional uses include stand-by power, private electric
generation and additional industrial purposes requiring shaft power.
Taken on a size category basis, the primary load uses are as
follows:
Category Load Use
Large capacity t Primarily base, midrange and peak utility
electrical generation
Medium capacity t Electrical standby generation
• Pipeline compression and pumping
• Industrial electrical generation
• Variety of industrial uses requiring
shaft power
Small capacity • Standby electric generation for the oil
and gas industry
t Gas compression
3.1.4 Fuels and Emissions
Stationary gas turbines primarily utilize clean fuels, typically
natural gas and distillate oils. The oil shortage of the early 1970s
however, had a profound effect on the fuel situation. The whole question
of fuels availability has become very volatile and difficult to predict.
The industry is now being forced to examine options regarding what fuels
will be available when current supplies of clean fuels are expended or
3-8
-------
become unavailable due to regulation. Already some utilities are firing
residual oils, although, due to the high cost of pretreating residual
oils, the economics generally favor gas or distillate oils. Still,
Westinghouse reports considerable success with residual oil, with one of
their units running 2000 hours.
Some of the reported problems when firing dirtier fuels include
coking on fuel nozzles and clogging fuel filters. Typically some degree
of preliminary treatment is required, such as reducing sodium by
50 percent and using an additive to inhibit vanadium corrosion. In these
cases manufacturers report no problems with "hot" components (i.e., those
equipment parts in contact with combustion gases) (Reference 3-8). In
some cases the turbine inlet temperature would have to be lowered to
prevent sulfidation if sulfur is present in the fuel. This, however, may
have a detrimental effect on cycle efficiency.
Some of the most promising new clean fuels are low and high Btu
gases and process gases such as coke oven and blast furnace gases
(Reference 3-9). Improved thermodynamic cycle efficiencies and low NOV
A
emissions make these clean fuels attractive alternatives in broadening
basic energy sources. There are, however, a number of redesign
considerations with the use of certain low Btu fuels in conventional
engines. Modifications to the combustion and fuel systems are all that is
required with some fuels. But with others, significant problems arise
from a compressor-turbine mismatch due to inordinately high pressure
ratios caused by excessive turbine weight flow. A more comprehensive
discussion of gas turbine fuel alternatives is given in Section 4.1.
According to the Emission Characterization of Stationary NO
A
Sources (Reference 3-10) which cites fuel consumption data for stationary
gas turbines, the 1974 estimates are of high quality. The total energy
consumed by gas turbines was about 3.5 percent of the total stationary
source fuel consumption in 1974. As Table 3-2 shows, medium-capacity
units consumed more fuel than the large units. The bulk of the fuel
consumption of these medium-capacity turbines was either in the oil and
gas industry, where equipment operates almost constantly, or in private
sector electricity generation, where equipment operates about
three-quarters of the time. .
3-9
-------
TABLE 3-2. 1974 GAS TURBINE FUEL CONSUMPTION (EJ)
Gas Turbines
Gas Turbines
115 MW
Gas Turbines
4 MW to 15 MW
Gas Turbines
<4 MW
Natural
Gas
0.212
0.468
0.001
Oil a
0.264
0.579
0.001
Total
0.476
1.047
0.002
alncludes distillate, diesel, residual oils
Similarly, regional fuel consumption data are of high quality,
particularly that from the utility sector. Additional data were traced
through manufacturers. In general, distillate oil shows a strong
domination in the New England and Middle Atlantic regions while the
West-North-Central, East-North-Central, Mountain and Pacific regions are
primarily serviced by natural gas.
Air emissions are essentially the only effluent stream from
stationary gas turbines. The stream composition is highly dependent on
the fuel burned, combustor geometry and combustion characteristics. NO
/\
emissions are highest and CO and UHC are lowest when the engine is
operating at design conditions (i.e., rated power output). Off-design
firing, while limiting N0x, enhances the production of unburned species
through incomplete oxidation. One method to enhance combustion of these
unburned species is to increase residence time through enlargement of the
reaction zone of the combustor. This is particularly applicable with
today's combustors and lean primary zone burning. Virtually all the fuel
sulfur is converted to sulfur dioxide in a turbine engine, the
concentration being purely a function of the fuel sulfur content.
Selecting fuels low in sulfur is the only means to control SO
emissions. Flue gas desulfurization, when applied to gas turbines is
overwhelmingly expensive (Reference 3-2).
3-10
-------
The only liquid waste associated with gas turbines is from the
water treatment equipment in a NO control water injection system. The
/\
waste water flowrate, when compared to that from a comparably sized
utility steam boiler, is small and after pH treatment, is relatively
innocuous. Typical disposal techniques involve evaporation, return to
rivers or municipal sewers. Once dry N0x controls become commercially
available (1982 by most estimates), water wastes will no longer exist.
Wastes from water treatment evaporation ponds are the only solid
effluent from gas turbines. Again, these wastes are relatively innocuous
and easily disposed of as landfill. With the advent of dry controls, this
waste problem will also be obviated.
3.1.5 Capital and Operating Costs
This section describes the types of considerations involved in
evaluating the total costs associated with stationary gas turbines
equipped with some form of NOX control. Capital costs include
purchasing and installing the required equipment. For wet controls, the
cost includes the water purification system, water injection system, and
any additional turbine modifications. For dry controls the cost would be
for engine modifications. Fixed costs per year, taken as 20 percent of
installed costs, would include such factors as depreciation, taxes,
insurance, etc. Operating and maintenance costs are taken to be 3 to
5 percent of installed costs, depending on size and use of the turbine.
We consider operating and maintenance costs to include labor to operate
the control device and normal maintenance. Any maintenance cost to the
gas turbine resulting from the control device is in addition to the normal
operating and maintenance costs. The cost of any additional fuel usage
caused by the NOX control device and, in the case of wet controls, water
costs, must also be considered in assessing total costs of owning and
operating a stationary gas turbine with NOX controls.
3.2 PROCESS DESCRIPTION
Gas turbines are rotary internal combustion engines commonly,
although not universally, fired with natural gas or "clean" liquid fuels
such as diesel or distillate oils. The simple cycle gas turbine shown in
Figure 3-3 is a Brayton cycle engine consisting of adiabatic compression,
constant-pressure heating and adiabatic expansion processes. While this
cycle type predominates the stationary gas turbine population, there are
3-11
-------
OJ
I
r
Air
Air
L
Fuel
Combustion
chamber
Compressed
.air
Compressor
Stationary qas turbine
Rotary V
enerqy
Load
Figure 3-3. Basic simple cycle gas turbine.
-------
two other types in common use: the regenerative cycle and the combined
cycle shown in Figures 3-4 and 3-5. Both cycles recover waste exhaust
heat to improve thermal efficiency. The regenerative cycle uses turbine
exhaust to heat incoming combustion air while the combined cycle uses
turbine exhaust gases to produce steam in a waste heat boiler. The
product steam then can be used to power a steam turbine or to provide
process steam.
Gas turbines can range in size from 40 hp to over 75 MW
(100,000 hp) and are commonly installed in groups by utilities to obtain
more power output. Such an arrangement may involve single turbines with
their own generators or multiple turbines connected to one generator.
The basic gas turbine consists of a compressor, combustion chambers
and a turbine. The compressor delivers pressurized air to the combustors
at compression ratios of up to 20 to 1. Injectors introduce fuel into the
combustors. The hot combustion gases are rapidly quenched by secondary
dilution air, reducing gas temperatures to 1,370 K (2,000°F), before
being expanded through the turbine. The turbine drives the compressor and
provides shaft power to a generator, compressor, pump, etc.
These basic components can be arranged in a variety of ways. On
some engines the combustor is placed axially between the compressor and
the turbine. In this design, the combustor may be made up of a series of
individual "cans" encircling the drive shaft or of two concentric
cylinders mounted to produce a single annular combustion chamber (hence
the terms cannular and annular combustor designs). On other gas turbines,
the combustor is a single large volume chamber, connected to the
compressor and turbine by ducting, but not necessarily physically located
between the compressor and the turbine. A typical simple cycle gas
turbine may be a dual spool engine with can-annular combustors and a free
power turbine. The engine is dual spool because the high and low pressure
compressors each have independent rotor systems driven by independent
compressor turbines. The power turbine is free because it is not
connected to either compressor rotor. Other engines may simply have a
single shaft connecting the compressors, compressor turbine and power
turbine.
Manufacturers are continually seeking ways to improve cycle
efficiency, usually through cycle variations or through increased
3-13
-------
CO
I
EXHAUST
STATIONARY GAS TURBINES
ROTARY
ENERGY
LOAD
Figure 3-4. Typical regenerative cycle gas turbine.
-------
STEAM-
co
i—*
en
I STATIONARY GAS TURBINE
Figure 3-5. Typical combined cycle gas turbine.
-------
compression ratios and turbine inlet temperatures. Very significant gains
in system thermal efficiency can be achieved by recovering thermal energy
in the turbine exhaust. The regenerative air heater (recuperator) employs
this "waste" thermal energy to preheat incoming combustion air (already
compressed). In effect, this requires less heat (in the form of fuel) to
be added to meet the design gas conditions entering the turbine.
Figure 3-6 shows typical efficiency curves for different pressure ratios
for regenerative and simple cycle gas turbines at a specific turbine inlet
temperature. The figure shows that regenerators are beneficial only over
a limited pressure ratio. In high compression ratio machines, the
compressed air is already substantially heated. Consequently it is not-
feasible to transfer heat from exhaust gases which are only slightly
warmer than the compressed air.
Combined cycle plants recover "waste" thermal energy from the gas
turbine exhaust by using this heat to generate steam in a waste heat
boiler. Occasionally additional steam is generated by supplementary fuel
firing in the waste heat boiler. Due to their superior overall system
efficiencies, 44 to 47 percent versus 36 to 38 percent for a conventional
steam boiler, and their potential for even higher efficiencies, combined
cycle systems are one of the most promising alternative powerplant designs
for base and intermediate load service. This is particularly true now
that coal-derived synthetic fuels are becoming more economically viable.
3-16
-------
40
30
u
c
OJ
01
- 20
O)
u
10
Regenerative
10
Pressure ratio
15
20
Figure 3-6. Comparison of efficiency vs. pressure ratio
for simple and regenerative cycle engines-
3-17
-------
REFERENCES FOR SECTION 3
3-1. Mason, H. B., et al., "Preliminary Environmental Assessment of
Combustion Modification Techniques. Volume II, Technical Results,"
EPA-600/7-77-119b, NTIS-PB 276 681/AS, October 1977.
3-2. Goodwin, D. R., "Standards Support and Environmental Impact
Statement. Volume 1: Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
3-3. National Electrical Manufacturers Association, "Sixth Biennial
Survey of Power Equipment Requirements of the U.S. Electric Utility
Industry," NEMA, Washington, D.C., March 1978.
3-4. Personal communication with S. M. DeCorso, Westinghouse Electric
Corporation, Philadelphia, PA, July 1978.
3-5. Hallberg, K., ed., "Turbo Machinery International", Vol 19, No. 2,
March 1978.
3-6. Sawyer, J. W., ed., Sawyers Gas Turbine Catalog, Gas Turbine
Publications, Inc., Stamford, CT, 1975.
3-7. General Electric Company, "Comparative Study and Evaluation of
Advanced Cycle Systems," Phase I Report, May 1976.
3-8. Personal communication with P. Dinenno, Westinghouse Electric
Corporation, Philadelphia, PA, July 1978.
3-9. Schiefer, R. B., et_ aj_., "Low Btu Fuels for Gas Turbines," ASME
74-GT-21, November 1973.
3-10. Salvesen, K. G., et al., "Emission Characterization of Stationary
NOX Sources, Volume I," EPA-600/7-78-120a, NTIS-PB 284 520, May
1978.
3-18
-------
SECTION 4
CHARACTERIZATION OF INPUT MATERIALS* PRODUCTS, AND WASTE STREAMS
This section summarizes the physical and chemical characteristics
of input materials (fuels), products (shaft power and waste heat), and
waste streams (primarily gaseous emissions). The focus of this section
will, however, be on fuel properties since they are a dominant factor in
determining the type and quantity of emissions.
4.1 INPUT MATERIALS
The basic input materials to a gas turbine are fuel and combustion
air. In addition, if the gas turbine has wet NO controls, steam or
}\
water may also be input. The emphasis in this section is on fuels, gas
and liquid. Discussion of associated contaminants that may be entrained
with the combustion air or steam/water is incorporated within the
appropriate fuel subsections.
The basic design of gas turbines tends to limit the fuels of
combustion to the "clean" fuels, natural gas and distillate oils. Those
oils which contain significant ash and trace element concentrations, such
as crude and residual oils, may require some treatment before they can be
used. Coal and some liquids are used to synthesize low Btu gases. These
gases will find increased use due to the reduced availability of the
standard gas turbine fuels (References 4-1 and 4-2). Also, coal and shale
derived liquids are showing promise as gas turbine fuels.
4.1.1 Gas Fuels
Natural gas is one of the primary fuels used in gas turbines. It
can be used directly from the distribution center if the solid contents
from such contaminants as water, sand, rust, naphthalene and others are
less than 30 ppm (Reference 4-3). Greater impurity contents may require
some fuel pretreatment. The lessened availability of natural gas will
4-1
-------
lead to increased use of other fuels (References 4-1 and 4-2). Table 4-1
summarizes the typical properties of some common gaseous fuels
(Reference 4-4).
One of the most promising new fuels is low Btu gas, produced by air
injected coal or oil gasifiers, or other chemical processes. This fuel,
although promising, has emissions and operating problems of significantly
greater magnitude than those associated with natural gas.
The low Btu gas composition from a gasifier depends on its
configuration and operating conditions. Typically, as shown in Table 4-2,
the combustible constituents of the gas are hydrogen, carbon monoxide, and
some methane. Natural gas is presented as a reference (Reference 4-1).
Ammonia (NhU), sometimes present in product gases, may be converted to
NO in lean flames, possibly with yields of greater than 50 percent
(Reference 4-1). Future NO emission standards for gas turbines may
A
require engine manufacturers who wish to use low Btu fuels to establish
specific limits on the amount of MM, allowable in the fuel gas or to use
combustors designed to fire high NH3 fuels and still meet NSPS.
Similarly, hUS limits will be dictated by environmental concerns rather
than manufacturer requirements. Any H2S present will form S02 when
burned.
Two parameters of the combustion of low Btu fuels which affect
turbine design (modification) and emissions are the stoichiometric fuel-
to-air ratio and the stoichiometric temperature rise. The fuel-to-air
ratios for low Btu fuels are considerably greater than those of natural
gas (Reference 4-1). The stoichiometric ratio for natural gas is
typically 0.062 Fuel/Air (F/A) whereas with low Btu fuel, it may approach
0.712 (F/A) on a volume basis. This is approximately an order of
magnitude increase in the gaseous flow into the primary combustion chamber
(Reference 4-2). The increase in fuel flow requires larger control
valves, nozzles, and possibly primary combustion zones. (Reference 4-2).
But, primarily, a switch to low Btu fuels would cause a significant
redesign of engine controls due to a combustor/turbine mismatch. It
follows, then, that changes in input fuels to a machine must be carefully
considered.
The maximum flame temperature (i.e., stoichiometric temperature
rise) is a function of the fuel composition. For many of the low Btu
4-2
-------
TABLE 4-1. TYPICAL PROPERTIES OF COMMON GASEOUS FUELS (Reference 4-4)
Property
Heating Value, Btu/ft3
Heating Value, Kcal/m3
Specific Gravity
Composition, Volume %
Methane, CH4
Hydrogen, H-
Carbon Monoxide, CO
Nitrogen, N-
Carbon Dioxide, C02
Natural Gas
(Dry)
950/1150
8400/10,200
0.58/0.72
75/97
2/20
1/16
0.1
Coal Gas
(Low Btu)
110/165
1000/1500
0.80/0.92
0.5/4.5
2/32
30/55
0.5/10
Coal Gas
(High Btu)
500/700
4500/6200
0.41/0.48
20/35
2/4
5/15
4/11
2/4
Coke Oven
Gas
525/650
4500/6200
0.40/0.45
28/32
2/4
5/7
1/6
2/3
Blast Furnace
Gas
90/100
700/800
0.95/1.05
25/30
55/60
8/16
EE-T-010
Reproduced by permission of the General Electric Gas Turbine Division,
GER-2222J, 1974, by A. D. Foster.
4-3
-------
TABLE 4-2. TYPICAL ANALYSIS AND PROPERTIES FOR LOW Btu GASEOUS FUELS
(Reference 4-1)
Analysis
(% by volume)
°2
4
C02
CO
CH4
C2"6
C3H8
C4H10
CgHl2
H?0
H2S
Specific
Gravity
(f/a)sto1c
(by wt)
(by vol)
LHV
(Btu/scf)
(Btu/lbro)
HHV
(Btu/scf)
(Btu/lbm)
Natural
Gas Lurg1b
—
.5 30.2
1.8 10.7
10.7
15.4
93.3 4.4
3.49
0.68
0.18
0.04
27.8
0.5
0.600 0.772
0.062 0.712
0.103 0.923
930 121
20,255 2041
1030 150
22,441 2488
Fluidized Bedb
Coal Gas
—
50.4
.5
31.8
15.6
.5
—
—
—
.5
.7
0.827
0.674
0.815
154
2429
163
2569
BOH. 7644b
—
54.5
7.2
2.0
15.5
2.8
—
—
—
—
—
0.856
0.770
0.899
132
2021
143
2182
Winkler
11"
—
55.3
10
22
12
0.7
—
—
—
—
—
0.912
1.040
1.141
110
1578
117
1674
Winkler
#2b
—
1
19
38
40
2
—
—
—
—
—
0.705
0.344
0.488
250
4641
272
5049
Producer Gas
From Coal
—
52.4
4
29
12
2.6
—
—
—
—
—
0.871
0.711
0.817
150
2250
158
2379
Blast
Furnace
Gas
—
60
11
27
2
—
—
—
—
—
1.010
1.462
1.448
92.3
1194
93.3
1207
Blue Water
Gas
—
4.5
4.7
41
49
0.8
—
—
—
—
—
0.550
.248
0.450
274
6504
299
7105
Coal Gas
(Vertical Retort
w/Steamlng)
0.4
6.2
4
18
49.4
20
2
—
—
—
—
—
0.475
0.120
0.252
422
11.626
471
12,965
*For reference purposes only, not a low Btu gas.
"Gas from various types of coal gasification schemes (see Reference 4-3).
Reproduced by permission of Power Engineering
-------
fuels, the lower heating value results in lower maximum flame temperatures
(Reference 4-2). Since the rate of NO formation increases
/\
exponentially with local flame temperature when firing clean fuels, lower
N0x emissions will result (References 4-1, 4-2, 4-5).
Another priority pollutant emitted from the combustion of low Btu
fuel is carbon monoxide (CO). CO has the slowest combustion rate of any
species found in the combustion flames of typical hydrocarbon fuels
(Reference 4-1). Therefore, for new engines burning low Btu fuels the
residence time must be increased in the primary combustion zone (over that
of an existing engine).
Other contaminants which could be emitted in the flue gas are trace
elements which can enter the system independent of the fuel. These are
primarily sodium and potassium, conceivably entering through the compressor
inlet air, particularly in ocean environments; injected water or steam; or
evaporative cooler carryover. Some trace elements such as those in salt
water may also be entrained in the fuel. These elements can react with
sulfur, forming sodium sulfate (for example), which is a corrosive compound
at high temperatures. These corrosive compounds can be controlled by
limiting the amounts of trace elements entering the combustion zone in the
air/water/steam/fuel.
4.1.2 Liquid Fuels
As previously mentioned, existing engines have been designed to be
fired with clean fuels. In the case of oils, distillates are preferred.
Recent indications are that future trends, dictated by fuel availability
and economics, may be towards heavier residual oils.
Table 4-3 summarizes the specifications of the various grades of
oils. Those physical and chemical properties that will affect turbine
operation and emissions are discussed below. In general, true distillates
may be fired as received and at normal climatic conditions, while ash-
bearing fuels may require some pretreatment. So called synthetic fuels,
such as the middle and heavy distillates obtained from pyrolysis of coal,
are also becoming a potential gas turbine fuel. Indeed, synthetic fuels
may be the future fuel for gas turbines due to the changing market for more
conventional fuels, Federal fuel use regulations, and other considerations.
4-5
-------
TABLE 4-3. DETAILED SPECIFICATIONS FOR FUEL OILS (Reference 4-6)
Grade
No. 1
No. 2
No. 4
No. 5
(light)
No. 5
(heavy)
No. 6
Flash
Point
°C
(OF)
Min
38
(100)
38
(100)
55
(130)
55
(130)
55
65
Pour
Point
°C
(OF)
Max
0
(20)
(20)
Water
and
Sediment
Volume %
Max
trace
0.10
0.50
1.00
1.00
2.00
Carbon
Residue
on 10%
Bottoms,
%
Max
0.15
0.35
Ash,
Weight
%
Max
0.10
0.10
0.10
Distillation
Temperature
°C (°F)
10* 90%
Max
215
(420)
Min
286
(540)
Max
288
(550)
338
(640)
Saybolt Viscosity
Universal at
38°C (10QOF)
Min
(32.6)a
45
150
350
(900)a
Max
(37.4)3
125
300
750
(9000)3
Sulfur
%
Max
0.5
0.7
no limit
no limit
no limit
no limit
aNumber given for information only; not necessarily limiting.
(Reproduced by permission of Industrial Press, Inc. in Fuel Oil Manual, 1969.)
4-6
-------
4.1.2.1 Physical Properties
Contaminant concentrations and viscosity are among the primary
physical properties that affect operation (Reference 4-4). The fuel must
be free of foreign material that would foul accessories and decrease the.
useful life of components. This includes water, sediment and filterable
dirt. Complete precombustion particulate removal for particles larger
than 5 um is recommended by at least one equipment vendor (Reference
4-4). Sediment and other particulates can cause plugging and erosion
problems with the equipment.
The quantity of water found in oil varies significantly. In light
oils, such as No. 2, water is present in negligible quantities, unless it
is introduced by outside contamination such as tank leakage, condensation,
etc. Water can cause sparking and spitting of the flame, as well as
flash- back of the flame. Loss of flame is also possible if excess water
is present.
Viscosity is directly related to the atomization potential of the
oil. If the fuel is not properly atomized, combustion may not be
complete. The completeness of combustion, in turn, directly affects the
emission levels of hydrocarbons, carbon monoxide and NOX. Typically, as
combustion efficiency goes down, unburned hydrocarbon emissions increase
while N0x decreases. Turbines firing light distillate oils generally
use direct pressure or low-pressure air atomizing fuel nozzles and require
preheating only during abnormally low ambient temperatures. Most heavy
distillates and blends, and virtually all crude and residual oils require
preheating between 320 K and 400 K (120 and 260°F), use high-pressure
air atomizing fuel nozzles (Reference 4-4), and may cause a significant
cost impact due to a larger compressor requirement.
4.1.2.2 Chemical Properties
Certain constituents (impurities') present in most fuel oils are
noncombustible and form residual ash. Most ash found in fuel oil comes
from contaminants present in crude oil, with a minor fraction due to
contamination occurring during handling or refining. Heavier fuels
typically contain the most ash.
The composition of fuels, particularly residual oils, may include
trace elements. The five elements of prime concern are vanadium, sodium,
potassium, lead, and calcium. The first three can lead to accelerated
4-7
-------
corrosion of turbine blades in elevated temperature gas turbines
especially when sulfur is present. In addition, all five may cause hard
deposits on the blades which can lead to reduced machine output.
Table 4-4 presents ranges of the elemental analysis (excluding
combustibles) of the fuel oils of interest.
Trace metals can also be introduced into the combustion system
through the compressor and steam/water injection system. Sulfur from the
fuel, in combination with these trace elements, particularly sodium and
potassium, can combine to form compounds such as sodium sulfate and other
compounds which are extremely corrosive at elevated temperatures. Generally
it is best to prevent this type of corrosion, called sulfidation, by
removing the contaminant in the incoming stream by giving special
attention to inlet air filtration, steam/water quality, and fuel handling
and treatment processes. Also turbine inlet temperatures can be kept low
to prevent sulfidation. Improved turbine inlet pattern factors will also
help to prevent high temperature corrosion by eliminating hot spots.
4.1.2.3 Fuel Treatment
Depending on the physical properties and the contamination level of
the fuel, most ash-bearing fuels will require some degree of treatment.
Fuel treating systems are generally divided into two stages: oil-
desalting and additive treatment.
The oil-desalting stage is a water washing stage where the water
soluble salts of sodium, potassium, and calcium are removed. The water
and soluble trace elements are separated from the fuel by centrifuging or
electrostatic precipitation processes.
The second stage is designed to alter the form of oil-soluble trace
elements, such as vanadium. The additive found most effective in reducing
vanadium corrosion is magnesium. When added in concentrations of 3:1,
magnesium to vanadium, the effect'is a corresponding increase in the
melting point temperature of the ash and a coating effect of the thin ash
deposits that form on turbine blades (References 4-4, 4-8). The deposits
which collect on the mechanisms are significantly less corrosive and
easily removed during turbine outages. Figure 4-1 shows the treatment
process.
Sulfur removal processes such as hydrodesulfurization are also
available for the heavier oils. In hydrodesulfurization, sulfur compounds
4-8
-------
TABLE 4-4. SELECTED FUEL CONSTITUENTS (References 4-4 and 4-7)
^•vFuel Type
Fuel ^"^-^
Constituent ^v.
Sulfur, %
Nitrogen, X
Hydrogen, %
Ash (fuel as delivered), ppm
Ash (inhibited), ppm
Trace metal contaminants,
untreated
Sodium plus potassium, ppm
Vanadium, ppm
Lead, ppm
Calcium, ppm
True Distillates
Kerosine
0.01/0.1
0.002/0.01
12.8/14.5
1/5
~
0/0.5
0/0.1
0/0.5
0/1
No. 2
Distillate
0.1/0.8
0.005/0.06
12.2/13.2
2/50
--
0/1
0/0.1
0/1
0/2
Ash-Bearing Fuels
Blended
Residuals
and Crudes
0.2/3
0.06/0.2
12.0/13.2
25/200
25/250
1/100
0.1/80
0/1
0/10
Heavy
Residuals
0.5/4
0.05/0.9
10/12.5
100/1000
100/7000
1/350
5/400
0/25
0/50
Trace
Element
Limitations3
150 ppm
Legalb
1 ppm
10 ppm
aLimits recommended by manufacturer (Reference 4-8)
bVanadium levels less than 0.5 ppm do not require inhibition. Maximum vanadium levels
are usually dictated by local codes regarding resulting stack emissions and the user's
acceptable costs to Inhibit. Maximum values for gas turbines are not significant;
however, 400 ppm is generally slated as an upper limit.
4-9
-------
Emulsion breaker
additive system
Duplex
strainer
Self-cleanin
centrifuges
Raw fuel oil
storage tank
Water supply
Hot wash water
supply system
r~
->
_/ s
! k
!
U
Jater drain
I
Recycling
water pump
Sludge pump
Recycling
tank
storage tank
/>.
•*. /
ft
Suction
heater
To fuel fwd. pump
Figure 4-1. Ash bearing fuels treatment (Reference
4-3). (Reproduced witn permission of Power Engineering.)
4-10
-------
are reduced with hydrogen over a catalyst into hydrogen sulfide and
hydrocarbon remnants of the original sulfur compounds. This process is
performed at relatively high temperature and pressure (Reference 4-10).
Hydrodesulfurization is carried out by the oil refiner and not the user.
Other processes are available and are discussed in subsequent sections.
It should be noted that the treatment systems generate waste streams:
liquid streams containing dissolved solids which must be disposed of in an
acceptable manner, as well as gaseous emissions such as hydrogen sulfide.
Furthermore, increased fuel handling and processing also increases the
potential for fugitive air emissions, particularly unburned hydrocarbons.
Therefore the treatment systems themselves require careful control.
4.2 NCL Formation
"^ Y\ '•' ™n-"--M-r-n^™«^«™«"
Two mechanisms have been identified as responsible for NO
A
formation during the combustion process: thermal NO formation and fuel
A
NO formation.
A
4.2.1 Thermal NO,, Formation
^ ""^™""*"™y\ "'" •'•'••™ ••!• MI H TTIJ
Thermal NO results from the thermal fixation of molecular
A
nitrogen and oxygen in the combustion air. Its formation is extremely
sensitive to local flame temperature, residence time at this local flame
temperature, and somewhat less so to local concentration of oxygen.
Virtually all thermal NO is formed at the region of the flame which is
A
at the highest temperature. This kinetically controlled behavior means
that thermal NO emissions are dominated by local combustion conditions.
A
The great majority of fuels used in existing engine designs are very low
in fuel nitrogen. Consequently virtually all NO emissions are due to
A
thermal NO formation.
A
4.2.2 Fuel NO.. Formation
^^/^ ^™«^™^-^^^»—
Fuel NO derives from the oxidation of the organically bound
A
nitrogen present in certain fuels such as residual and distillate oils.
Fuel NO emissions are dependent on the nitrogen content of the fuel as
A
well as on combustion conditions. Fuel NO formation is strongly
A
affected by the local oxygen concentration under which combustion takes
place.
Figure 4-2 shows experimental results from a bench-scale combustor
indicating that fractional conversion of fuel nitrogen decreases with
increasing fuel nitrogen content (up to one percent fuel nitrogen)
4-11
-------
X
o
1.0
0.8
§ 0.6
c
o
0.4
0.2
VARIATION DUE TO
MEASUREMENT TOLERANCES
LABORATORY TEST DATA AT FUEL/AIR RATIO = .021
DRY COMBUSTION
0.01
0.02
0.04 0.06 0.08 0.1 0.2
FUEL-BOUND NITROGEN, percent BY WEIGHT
0.4
0.6 0.8 1.0
Figure 4-2. Conversion of fuel bound nitrogen to NOX in a combustor
(Reference 4-10).
4-12
-------
(Reference 4-10). For distillate fuels, with nitrogen contents generally
less than 0.015 weight percent, nearly all the fuel nitrogen is
converted. But fuel NO is not a problem (compared to thermal NO )
/\ "
with distillate oils because of their low nitrogen content. With residual
oils, however, fuel nitrogen concentrations are generally above 0.2
percent and can go as high as 2 percent (Reference 4-10). Fortunately,
only a fraction of that fuel nitrogen is ultimately converted to NO .
/\
Figure 4-2 predicts a 40 to 50 percent conversion and data by Dilmore
(Reference 4-11) indicate only a 20 to 30 percent conversion. However,
actual total NO concentrations may increase significantly due to the
/\
high fuel nitrogen content. More recent data from EPRI and Westinghouse
(Reference 4-12) indicate that the percentage of total NO conversion
/\
actually increases linearly, within a 20 to 50 percent range, with the
weight percent of fuel bound nitrogen.
4.2.3 Fuel Potential for NO., Production
^—^*^.^_M^_^H_p^ , | •,„• _______
The amount of NO produced is a function not only of the fuel,
J\
but also of the mechanical configuration and operation of the gas
turbine. In general, for a given fuel, time, temperature and mixing
considerations will dictate the quantity,of NO in the effluent gas.
/\
The rate of NO formation increases markedly with local flame
J\
temperature. Figure 4-3 shows the exponential increase in NO emissions
/\
with combustor inlet temperature. The same temperature dependence can be
seen in Figure 4-4 (Reference 4-14), which also shows the effect of
stoichiometry on NOV formation in or near the primary reaction zone of a
A
flame from a burner. Figures 4-5 and 4-6 show temperature and NO
X
concentrations respectively as a function of position within a laboratory
scale gas combustor burning liquid normal heptane. Note that as the
equivalence ratios, , are increased, (where 4> = (fuel/air)actual/
(fuel/air)stoichiometric), so does the area of the hot zone. The maximum
temperature does not increase but the size of the region at these highest
temperatures does increase. This is partially due to there being less
diluent air. The formation of NOX in the hottest regions of the
combustion zone is known as thermal NO and has been represented by the
/\
Zeldovich reactions combining nitrogen and oxygen in the combustion air
(Reference 4-16). Note in comparing Figures 4-5 and 4-6 the
correspondence of higher NOV at higher temperatures.
/\
4-13
-------
40
35
30
OJ
O)
25
C\J
o
20
X
OJ
•o
c
15
ox 10
0
o
O
a
X
ENGINE
MODEL
JT9D
JT3D
HUMIDITY
MASS RATIO
0.01
0.0123
0.0083
0.0002
0.01
0.0118
0.0130
0.0170
0.0046
0.0037
GTC8S-90-2
TSCP-700-4
JT8D Smokeless 0.0153
Calculated 0.01
/<
400 500 600 700 800
Combustor Inlet Temperature in °K
Figure 4-3. NOX emissions as a function of combustor
inlet temperature (Reference 4-13).
(Reproduced by permission of the ASME.)
4-14
-------
en
a.
a.
0.8
Flame T = 2300 K
A
-2100 K
I
Figure 4-4.
1.2 1.6
Stoichiometry
(Fuel/Air) primary reaction zone/(Fuel/Air) Stoichiometry
NO formed in or very near the primary reaction zone of ethylene flames
(Reference 4-14).
-------
Too*
* 0.119
AIIAk OIITAMCC . inCMCS
I
0.1 0.2
Meters
0.3
Figure 4-5. Temperature distribution maps of a laboratory
scale gas combustor (.Reference 4-15).
(Reproduced by permission of the ASME )
4-16
-------
$
0
ttttj.
'V
1
>
V
-a— _
• •i i
^^-
-a-^
^
^
ff
j-
4
1 I
1
^
* -C
1
I
.1:9
(
»
Alll
4
4> * 0.281
I I
• r • t
IU MtTMCC. Mutt
ii itai«
0
0.1
0.2
Meters
0.3
Figure 4-6.
Nitric oxide concentration maps of a laboratory
scale gas combustor (Reference 4-15).
(Reproduced by permission of the ASME.)
4-17
-------
A second phenomenon which promotes NO production is dwell time,
A
or residence time, of the combustion components in the hot regions of the
combustion chamber. Figure 4-7 illustrates this concept for combustion of
premixed, prevaporized propane fuel in a lean primary zone combustor
typical of today's engines. NO is plotted as a function of residence
A
time with equivalence ratio as a parameter (Reference 4-5). The
intersection of the combustion efficiency curves with the NO curves
A
gives the residence time required to achieve a given efficiency at a given
equivalence ratio. Note that residence time is a strong factor in NO
A
formation only at high equivalence ratio. It follows therefore, that NO
A
emissions increase with increasing equivalence ratio. The figure further
suggests that the way to achieve the lowest possible NO emissions while
X
maintaining high combustion efficiency is to burn as lean as possible but
to allow long residence times (Reference 4-5). Figure 4-4 also shows how
less NO is produced in the primary reaction zone of a flame as the
A
mixture becomes more lean. Recent data from Pratt & Whitney has shown
that in a newly developed, full-scale combustor with rich burning in the
primary zone and firing fuels containing bound nitrogen, NO emissions
A
decrease with increasing residence time.
Other mechanisms related to time, temperature, and mixing, impact
the NO emission levels. Turbulence (or the degree of mixing) can
A
affect NO emissions and is dependent on the type of fuel. As seen in
A
Figure 4-8, liquid fuels generally result in higher NO emissions than
X
gaseous fuels (Reference 4-17). However, this is partly due to the fact
that liquid fuels tend to have fuel bound nitrogen (which can contribute
to the total NO emissions) as well as to such factors as degree of
A
mixing. Therefore the comparison between liquid and gaseous fuels should
only be made in cases of relatively nitrogen free fuels such as distillate
oil and natural gas (methane). Gas fuels can be effectively mixed with
combustion air with high turbulence to produce a uniform fuel/air mixture
throughout the combustion zone and hence, uniform temperatures. By
reducing localized combustion temperature, NO emission levels can also
be reduced (Reference 4-5). Tests have also shown that better mixing can
reduce NO formation when firing liquid fuels (Reference 4-5).
A
Combustion of liquid fuels which are not prevaporized can be
thought of as droplet burning. The fuel is atomized and mixed with air in
4-18
-------
to,—
o
«r
*/•»
1
X
O
e .5
.2
EOUIVALfNCL
RATIO
COMBUSTION
EFFICIENCY.
.5 1.0 1.5 2.0
RESIDENCE TIME. MSEC
2.5
3.0
Figure 4-7. Effect of residence time on nitrogen oxides emissions for
a lean primary combus tor. Propane fuel inlet mixture
temperature, 800K; inlet pressure, 5.5 atm; reference
velocity 25 and 30 m/s (Reference 4-5).
(Reproduced by permission of the ASME.)
-------
NO emission level
2.0
2.5
i
r^s
o
Carbon monoxide (CO)
Crude oils
Distillate oils
Ethanol (C2«s OH)
Hydrogen (H2)
Methane (CH4)
Methanol (CHjOH)
Natural gases
Propane (CjHg)
Residual oils
Simulated coal gases
Reference fuel: No
2 distillate containing
0-100 ppm N by wt.
Figure 4-8. Predicted NOX emission levels of various fuels
burning in gas turbine combustors (N0>, levels are
normalized to emissions from burning #2 fuel oil)
(Reference 4-17).
(Reproduced by permission of the ASME.)
-------
the combustion chamber. Droplet burning is one mechanism by which burning
can result in local flame temperatures in excess of the mean combustion
zone temperature due to non-uniformities in the fuel/air ratio
(Reference 4-5). The more energy used for atomization and fuel/air
mixing, the more uniform is the combustion and resulting temperature, thus
producing less thermal NO at least in a fuel lean primary zone. Note
A
that the proper fuel/air mixing for lowering NO emissions with more
A
uniform combustion temperatures should be balanced against the opposing
effect of greater NO formation due to better oxygen availability, as
A
discussed earlier. Thus proper NO control requires controlled fuel/air
A ^™*^^ ^^-^^^^-^
mixing.
The above effects have led to several control mechanisms which
promote good mixing of fuel and air. One such mechanism is to promote
turbulent mixing within the combustion chamber. Another mechanism has to
do with increasing the atomization of liquid fuel. This can be
accomplished through preheating or prevaporization of the fuel. The
efficiency of fuel atomization is a function of the fuel viscosity
(Reference 4-18).
Still another concept is that of premixed fuel/air charges. This
is also known as hybrid combustion (Reference 4-19). The fuel and primary
combustion air are allowed to mix thoroughly prior to ignition in the
primary combustion zone (Reference 4-20). Figure 4-9 schematically shows
the postulated effect of premixed combustion mixture as opposed to
standard combustion techniques. Emissions at various equivalence ratios
using premixed systems are also shown.
It should be stressed, however, that these techniques to enhance
fuel/air mixing generally work only with lean primary burning of fuels low
in bound nitrogen. These principles may not be true with rich primary
burning of high bound nitrogen fuels.
Nitric Oxide/Nitrogen Dioxide
The relationship between nitric oxide emissions and total NO for
A
propane combustion is shown in Figure 4-10. Nitric oxide accounts for
about 90 percent of the total NO for most conditions (Reference 4-5).
A
This is typical of stationary source combustion of conventional fuels
(Reference 4-16). The remaining NO is emitted as N09.
A £
4-21
-------
i
STANDARD C
f PREHIX EM) I VALENCE
I0l RATIO
-t
DESIGN
TEMPERATURE
COHBUSTOR EXIT TEMPERATURE
Figure 4-9. NOX emission trends of a hybrid combustor versus
a conventional combustor (Reference 4-19).
(Reproduced by permission of the ASME.)
4aoor
INIET MIXTURE PROBE
TEMPERATURE, POSITION
K CM
600 800
10 »50 100 5*
TOTAL NITROGEN OXIDES (NOXI CONCENTRATION, PPM
4ooo
Figure 4-10.
Relationship between nitric oxide emissions and total
NOv for propane combustion. Inlet pressure, 5.5 atm;
reference velocity, 25 m/s; equivalence ratio, 0.40 to
1.0 (Reference 4-5).
(Reproduced by permission of the ASTM.)
4-22
-------
Other criteria pollutants include carbon monoxide (CO) and unburned
hydrocarbons (UHC). The quantity of each depends on the completeness of
combustion. The concentration of each is expected to be affected by
changes in operations for the reduction of NO For example, reductions
A
in combustion temperatures may reduce NO emissions but increase the CO
A
and UHC concentrations. A balance between NO and the other criteria
A
pollutants must be maintained. The effect of NO controls on other
A
pollutants is discussed in greater detail in Section 6.
4.2.4 Products Characterization
The principle product of a gas turbine is rotary shaft power. The
exhaust gas from a gas turbine combustor can also be considered a
"product." The hot exhaust gas can be used to supply waste heat via a
heat recovery device or, by virtue of its high excess oxygen content and
waste heat content, be used in a supplementary fuel fired waste heat
boiler. The product exhaust gas from a gas turbine is discussed in
further detail in the next Section, Emissions Characterization.
4.3 EMISSIONS CHARACTERIZATION
The emissions from a gas turbine without NO controls will
A
primarily be a function of the fuel composition, combustor geometry and
operating conditions of the system. Emissions of concern include NOV,
A
S02> CO, parti oilates, and unburned hydrocarbons.
Emission factors for uncontrolled emissions for the criteria
pollutants from gas turbines are presented in Table 4-5. The emission
factors come from field studies (References 4-10, 4-21, and 4-22) and an
EPA Document, AP-42 with supplements (Reference 4-23). Emission factors
for organic matter and sulfates from gas turbines cannot be determined at
present since extensive field test results have not yet been reported.
There are no liquid or solid effluents resulting from combustion related
gas turbine operation (with no NO controls).
A
4.3.1 NO Emissions
^^^™A ~••••••"^"•^•••^^™^^™
The amount of fuel bound nitrogen converted to NO is a function
A
of the supply of oxygen available for reaction. The degree of conversion
from Figure 4-2, also depends on the degree of fuel/air mixing. Thermal
NOY is a function of many variables, as described in the previous
A
section. Again, efforts to control NO must be carefully applied to
/\
assure acceptable emission rates of the other criteria pollutants.
4-23
-------
TABLE 4-5. GAS TURBINE CRITERIA POLLUTANT EMISSION FACTORS (ng/J)
Equipment Types
Gas Turbines
>15 MW (output)
Natural Gas
Diesel oil
Gas Turbines
4 MW to 15 MW
(output)
Natural Gas
Diesel oil
Gas Turbines
<4 MW (output)
Natural Gas
Diesel oil
N0x
194
365
•.
194
365
194
365
S°x
2.2
10.7
2.2
10.7
2.2
10.7
Part.
6.0
16.0
6.0
15.5
6.0
15.5
CO
49.0
47.0
49.4
47.3
49.4
47.3
HC
. 8.6
8.6
8.2
9.9
8.2
9.9
4-24
-------
4.3.2 jQg Emissions
Sulfur oxide emissions are almost exclusively a function of the
sulfur content of the fuel (Reference 4-10). 502 em''ssions are
controlled by burning low sulfur fuels. Some sulfur removal from fuels is
practiced, but only to a limited degree.
4.3.3 Particulate and Visible Emissions
Particulate emissions from gas turbines are a function of the ash
content of the fuel and combustion efficiency. Emissions from natural gas
firing are typically low, approximately 0.005 g/son (0.002 gr/scf) of gas,
with residual oil emissions approximately 0.24 g/scm (0.10 gr/scf).
Magnesium and other inhibitors added to liquid fuels to reduce vanadium
corrosion can also contribute to particulate emissions (Reference 4-10).
Reductions in visible and particulate emissions can be obtained by burning
clean fuels such as gas, and through combustor modifications for
improvement of the combustion efficiency.
4.3.4 Hydrocarbons and Carbon Monoxide Emissions
Hydrocarbon and CO emissions are functions of the combustion
efficiency of the unit. Since most units are designed for high efficiency
at maximum load, reduced load tends to cause the occurrence of increases
in CO and HC concentrations. Carbon monoxide reacts slowest of all
components formed during combustion, therefore it is emitted in the
largest concentrations. Figures 4-11 and 4-12 show the wide variations
that can occur as a function of unit size, operating conditions, fuel
type, and probably most significantly, design changes as sizes get
larger. Note that the values reported in the figures have been diluted to
15 percent oxygen.
The complete combustion of HC and CO emissions is a function of the
method of fuel injection, including atomization method and pressure,
degree of fuel/air mixing, and the residence time at combustion
temperature. It should be noted that improved atomization and fuel/air
mixing can reduce thermal NOX as CO and HC are reduced. However,
increased residence time and combustion temperature for more complete
combustion may increase NOX, at least for fuel-lean primary zone
combustors. This is not the case with fuel-rich primary zone combustors.
4-25
-------
200
02
o.
in
too
(Xi
en
0.
0.
I 50
CO
Ul
o
o
D NATURAL GAS
O LIQUID FUEL
D 0
o
^•^
O
B
°
O
°-bo
10 15 20 25 30 35
RATED OUTPUT,Mw
40
45
50
55
60
Figure 4-11.
CO emissions vs. turbine size for large gas turbines without
NO controls when operated at or near full load
(Reference 4-10).
-------
"p ? 1 1 r
16'D NATURAL GAS
O UOUIDFUEl
O O FUEL UNKNOWN
izl— O
— Q
- ° O
O
O °
•» ll—
O
~~ O
O
i r
O
O
1
0 0.5 1 2
RATED POWER OUTPUT, MM
Figure 4-12. HC emissions vs. turbine size for small gas
turbines without NO controls when operated
at or near full loaa [Reference 4-10).
4-27
-------
REFERENCES FOR SECTION 4
4-1. Schiefer, R. B., and D. A. Sullivan, "Low Btu Fuels for Gas
Turbines," ASME 74-GT-21, November 1973.
4-2. Hefner, W. J., "Alternate Fuels Capability of Gas Turbines in the
Process Industry," ASME 76-GT-119, January 1976.
4-3. Carpenter, R. J., "Fuel Systems for Gas Turbines," Power
Engineering, Vol. 74, No. 4, pp. 34-36, April 1970.
4-4. Foster, A. D., _et al., "Fuel Flexibility in Heavy-Duty Gas
Turbines," GeneralTTectric Report No. GER-2222J, General Electric
Company, Schenectady, NY, 1974.
4-5. Anderson, D., "Effects of Equivalence Ratio and Dwell Time in
Exhaust Emissions from an Experimental Premixing Prevaporizing
Burner," ASME 75-GT-69, December 1974.
4-6. Schmidt, P. F., "Fuel Oil Manual," Industrial Press Inc., Edison,
NJ, 1969.
4-7. Byam, J. W., et al., "Residual Fuel Treating and Handling for Base
Load Gas Turbines," ASME 75-GT-72, December 1974.
4-8. Krulls, G. E., "Gas Turbine Liquid Fuel Treatment and Analysis,"
Journal of Engineering for Power, Vol. 97, No. 1, pp. 55-63,
January 1975.
4-9. Amero, R. C., et al., "Hydrodesulfurized Residual Oils as Gas
Turbine Fuels^AW 75-WA/GT-8, July 1975.
4-10. Goodwin, D. R., et al., "Standard Support and Environmental Impact
Statement. Volume TT~ Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
4-11. Dilmore, J. A., e£ a_U_, "Nitric Oxide Formation in the Combustion
of Fuels Containing Nitrogen in a Gas Turbine Combustor,"
ASME 74-GT-37, November 1973.
4-12. Bauserman, G. W., et aK, "Combustion Effects of Coal Liquid and
Other Synthetic FueTs in Gas Turbine Combustors -- Part II: Full
Scale Combustor and Corrosion Tests," ASME 80-GT-68, March 1980.
4-13. Shaw, H., "The Effects of Water, Pressure, and Equivalence Ratio on
Nitric Oxide Production in Gas Turbines," Journal of Engineering
for Power, Vol. 96, No. 3, pp. 240-246, July 1974.
4-28
-------
4-14. Fenimore, C. P., "Formation of Nitric Oxide in Premixed Hydrocarbon
Flames," Proceedings of the 13th Symposium on Combustion, The
Combustion Institute, August 1970.
4-15. Starkman, E. S., et al., "The Role of Chemistry in Gas Turbine
Emissions," ASME 7U-TTF31, February 1970.
4-16. Mason, H. B., et al., "Preliminary Environmental Assessment of
Combustion ModTFi"caTi on Techniques: Volume II, Technical Results,"
EPA-600/7-77-119b, NTIS-PB 276 681/AS, October 1977.
4-17. Hung, W. S. Y., "The NOX Emission Levels of Unconventional Fuels
for Gas Turbines," ASME 77-GT-16, December 1976.
4-18. "Fuel Treatment Flexibility," Gas Turbine World, Vol. 7, No. 6,
pp. 34, January 1978.
4-19. Mumford, S. E., et a]^, "A Potential Low NOX Emission Combustor
for Gas Turbines Using the Concept of Hybrid Combustion,"
ASME 77-GT-15, December 1976.
4-20. Teixeira, D. P., et al., "Evaluation of a Premixed Prevaporized Gas
Turbine Combustor for No. 2 Distillate," ASME 77-GT-69, December
1976.
4-21. Hare, C. T., et al., "Exhaust Emissions from Uncontrolled Vehicles
and Related Equipment Using Internal Combustion Engines, Part 6:
Gas Turbine Electric Utility Power Plants," Southwest Research
Institute, San Antonio, Texas, Report to EPA, No. APTD-1495,
February 1974.
4-22. Dietzmann, H. E., and K. J. Springer, "Exhaust Emissions from
Piston and Gas Turbine Engines Used in Natural Gas Transmission,"
Southwest Research Institute, San Antonio, Texas, AR-923, January
1974.
4-23. "Supplement No. 4 for Compilation of Air Pollutant Emission Factors
(Second Edition)," U.S. EPA, Office of Air and Waste Management,
Office of Air Quality Planning and Standards, January 1975.
4-29
-------
SECTION 5
PERFORMANCE AND COST OF CONTROL ALTERNATIVES
Stationary gas turbines are characterized by a wide variety of
equipment types, cycle types, fuel requirements, and applications. All of
these factors contribute to a high degree of variability on the pollution
control potential from a specific turbine. Indeed, even gas turbines of
the same model and production run can have significantly different NO
A
emissions (Reference 5-1). For example, small machining differences in
fuel nozzles and air dilution holes can significantly affect localized
fuel/air ratios and residence times and ultimately NO generation.
A
This section evaluates the performance of existing and planned
pollution controls for stationary gas turbines. Since NO is the
A
principle and major pollutant of concern from gas turbines, the great
emphasis here is on NO controls (Reference 5-2). Control options for
/\
other pollutants are reviewed briefly in subsection 5.5. To make the
control evaluation meaningful and useful to researchers, regulatory
agencies and potential control users, pollutant reduction and/or
prevention must be considered, and the associated costs and impacts on
operations and maintenance must be analyzed.
5.1 PROCEDURES FOR EVALUATING CONTROL ALTERNATIVES
Ideally, consistent analysis and comparison procedures would be
used to judge fairly the effectiveness of various control alternatives.
These procedures would be applied to data obtained from tests and analyses
of actual full scale stationary gas turbines. However, there is a paucity
of data from long term control experience which can be applied to such
analysis procedures. The reason for this lack of data is that only a few
regions in this country have promulgated emissions standards for
stationary gas turbines (see subsection 6.3.1). And those that have, for
5-1
-------
the most part, have been promulgated for only 1 to 3 years. Only certain
counties in the South Coast California Air Quality Management District in
Southern California have had standards regulating NO emissions from
A
stationary gas turbines for more than 3 years (almost 7 years for Los
Angeles County). Of course, many stationary gas turbines that are not in
regulated regions have NOV controls (i.e., water injection). However,
J\
they are seldom used unless required by local NO regulations.
/\
Consequently this assessment relied on three primary sources of
data: users, manufacturers, and EPA, notably through the "Standards
Support and Environmental Impact Statement Volume 1: Proposed Standards
of Performance for Stationary Gas Turbines" (Reference 5-3).
The approach taken here is to compare baseline gas turbine
operation with that under controlled and low NO conditions. In
A
addition to pollutant reduction, other factors considered in comparing the
effectiveness of NO controls are fuels, incremental emissions of
A
pollutants other than NO , control economics, operating reliability,
A
additional maintenance requirements, current stage of control development,
and expected timescale of commercialization for new control technologies.
As mentioned earlier, data based on long term user experience with NO
A
controls are limited. Consequently, to compare baseline and controlled
firing, this investigation had to rely to a great extent on informed
conjecture and qualitative engineering judgement, particularly in
assessing operations and maintenance impacts. Furthermore, the available
user experience is extremely variable, if not contradictory. In general,
some problem areas have been identified; their frequency and degree depend
on equipment types, operating parameters and quality of maintenance.
The assessment of the economic impact due to NO control
A
techniques has had to rely to a great extent on real-time data supplied by
a few users and to a lesser extent, on costs quoted by manufacturers. The
manufacturer's figures are primarily for capital equipment costs and do
not include incremental operating costs. Much of the cost data was used
to update and confirm costs cited in the SSEIS. In performing a
comprehensive cost analysis, a significant hinderance was that many users
had never attempted to separate and specifically account for costs of wet
NO controls. Rather they simply looked at operating and maintenance
A
costs on a total turbine basis, not on a component by component basis.
5-2
-------
An additional concern when evaluating NO control alternatives
A
was the current state of development and the expected timescale of
commercialization of developing control concepts. For example, while
catalytic combustion appears to perform quite well in reducing all
pollutants from clean fuels in particular ( 98 percent reduction —
Reference 5-4), it is far from developed for use in full scale combustors
and consequently many years from fruition. Indeed, some manufacturers
have expressed skepticism whether it will ever become a reality for use
with stationary gas turbines (Reference 5-5).
There are a number of other factors that may cause delays in the
development of control techniques and that make evaluation difficult.
First, is the depressed domestic market for gas turbines. Although a slow
reversal of the trend is expected, the current sales picture is very weak
(Reference 5-6). Also there is a major hesitancy to invest in control
technique development because NSPS has only recently been promulgated.
Furthermore, development of new combustors is very expensive, approaching
$20 million to get a new combustor into a full-scale engine. Also the
fuels situation is unclear. Will clean fuels (low in fuel-bound nitrogen)
continue to be available, or will developers and users be forced to design
for dirty fuels? And finally, most development work is aimed at improving
thermal efficiency through higher turbine inlet temperatures, higher
pressure ratios, and combined cycles. This situation is forced by gas
turbines' competitive position with 1C engines which generally have higher
cycle efficiencies.
Ultimately, after review of current and future control
technologies, assessment of process and cost impacts, and emissions
reviews, this study will arrive at conclusions regarding the cost
effectiveness of particular NO control alternatives. Those controls
A
that show the best balance between emissions reduction performance and
capital and operating costs will be identified.
For all the control options studied, the analyses have been
conducted based on the currently available data. It must be remembered
that the figures are general and can vary widely for different
installations. In addition, there are still many unanswered questions
regarding long term cost, operations and maintenance data. Due to the
changeable nature of the gas turbine market and the current state of flux
5-3
-------
in control developments, these questions may become moot by the time they
can be reliably answered.
5.2 PERFORMANCE AND COST OF NOV CONTROL ALTERNATIVES
X
For all intents and purposes, the only gas turbine waste streams
resulting from combustion are exhaust emissions. The most effective and
economical way of reducing these emissions is by modifying combustion
process conditions. This subsection reviews the currently used and
developmental combustion modification techniques for control of exhaust
emissions. Also, because NO emissions, the primary gas turbine
X
effluent, are highly dependent on fuels as well as combustor design and
control technique, special emphasis is put on controls and the fuels
designed for.
Since stationary gas turbines typically operate at rated capacity,
combustion efficiencies are high. Complete combustion, while minimizing
production of CO and UHC, tends to maximize production of nitrogen
oxides. Sulfur dioxide emissions are solely a function of the fuel sulfur
content and are controlled by burning fuels with low sulfur content.
Particulate emissions are low because clean fuels (natural gas and
distillate oil with little or no ash) are generally used by gas turbines.
5.2.1 Control Techniques and Fuels
NO controls for gas turbines are usually classified into two
A
categories: wet techniques which inject water or steam into the
combustion zone, and dry techniques which involve some process
modification other than the addition of water. This typically takes the
form of combustor redesign. The following discussion reviews the
performance of wet and dry techniques while they are burning clean and
dirty fuels. Dirty fuels are those which are high in fuel-bound
nitrogen. Each will be treated separately in the following discussion.
The distinction is made due to the performance of certain control concepts
with certain fuel types. Subsequent discussions then focus on the impact
of these NO controls on other pollutants (Subsection 5.2.2). Finally,
A
the cost impact of these controls will be analyzed (Subsection 5.2.3).
5.2.1.1 Wet Control Techniques and Clean Fuels
The formation of thermal NO is highly dependent on flame
A
temperature. In fact, virtually all thermal NO is formed in the region
5-4
-------
of highest flame temperature and amounts formed increase exponentially
with increasing temperature as shown in Figure 5-1. With the injection of
atomized water or steam directly into the primary combustion zone, peak
flame temperatures are lowered which effectively lower NO . The heat of
J\
vaporization of the injected water effectively removes some of the heat
from the primary combustion zone. NO emission reductions as great as
/\
80 percent have been achieved with water injection on gas turbines.
Figure 5-2, updated from Reference 5-3, summarizes NO emissions data
A
from a variety of gas turbines using wet controls and firing clean fuels.
Figure 5-3, also updated from Reference 5-3, indicates the effectiveness
of water or steam injection in reducing NO emissions.
A
Typically, water injection equipment is skid-mounted while the
operating controls are mounted in the turbine control compartment.
Typical components of the water injection system include the water
injection pump, pump inlet strainers, a filter for outlet water, flow
meter, valves, motor starters, heaters, piping and instrumentation. A
mechanical components schematic from one gas turbine manufacturer is shown
in Figure 5-4. The electrical control schematic from the same
manufacturer is shown in Figure 5-5. This particular system is designed
to inject water over the complete load range to maintain the proper water-
to-fuel ratio predetermined to limit NO emissions to the federal
A
standard (Reference 5-7). Upon promulgation of the Standards of
Performance for New Stationary Gas Turbines, this capability is be
required for all gas turbines subject to the regulations. It is important
to note, however, that these engine controls are limited to existing
engines and current combustors. These controls and water injection would
not be suitable for new concepts such as premixing/prevaporization and
superclean combustors.
Water injection is commonly accepted as a valid technology to
control nitrogen oxides emissions from current combustor design. In fact,
General Electric currently has 16 Model MS 5001 and 43 Model MS 7001 gas
turbines equipped with water injection equipment (Reference 5-7). Some of
these are used to meet local air pollution regulations while others are
used to increase power output by increasing mass flow rates through the
turbine. One manufacturer guarantees their gas turbine NO emissions to
A
5-5
-------
X
UJ
a
o
on
40
30
20
10
9
1 i • i ' 1 '
4
ENGINE HUMIDITY '
MODEL MASS RATIO QQ
a . °-01 ¥
D ) OC123 /
o JT9D «» o'
«. 1 0.(WG2 r\
0 / ow
- 0 ) JT3D °01 ° -
fr 1 0.0113 •
0 ( 0013D 0°
o ) °-0"0 /
^ GTCSS-90-2 0.0046 D
Q TSCP.)OC-( 00037 '
x JT80 Smokeless 0.0)53 A
• Wculalei U.01 X/A
/
*P
/o
'
6.-^
O ^^
"° o , i
200 400 600 800 1000
367 478 589 700 811 K
COMBUSTOR INLET TEMPERATURE IN °F and °K
Figure 5-1. NOX emission index versus gas turbine combustor
inlet temperature (Reference 5-3).
5-6
-------
315
200
180
160
140
CSJ
o
£ 120
(J
U
Q-
LTJ
~ 100
«J
o.
P
i 80
o
-£
9i
^
-
^
10.56
j).43
0.6
C
5
<
9
0.7
<
1.1(
1.0
>
P
x. Leaend
[ ] Conibustor ria test
Amount of reduction
Q 0.5, etc. Water/fuel
9 Notes
<
0|5<
6
,
<
<
°"2
|>0.5
)0.7
S0.9 ^
9
i
9
'
' 9 (,
t.4.T, °'
(Si.o
-
9
•
9
m
L
\
'i
G.T. Size
(MW)
Fuel type
[2.2] 17.2 17.2 17.5 21.3 33
5260
13 13 17.2 17.5 61.5
0.5 2.5
60.4
• Liquid fuel (distillate)•
Natural aa=
Figure 5-2. Summary of NOX emission data, from gas turbines
using wet control techniques (Reference 5-3).
5-7
-------
60
o
UJ
cc
50
.
90
D NATURAL GAS
O LIQUID FUEL
10
D
70
/ O
/DO o
o „ /
I U0
"
00 /
/
O /
U S LJ /
40
,
I /
I /
I /
If
/
0.2 0.4 0.6 0.8 1.0 1.2
WATER/FUEL RATIO
Figure 5-3. Effectiveness of water/steam injection in reducing
NOX emissions (updated from Reference 5-3).
5-8
-------
MECHANICAL COMPONENTS SCHEMATIC
DUAL
MANIFOLD
-------
75 ppm at 15 percent oxygen in the flue gas, while another will supply wet
controls on an "as needed" basis.
Strict specifications for water quality are required to prevent
excessive corrosion in hot gas turbine parts. Control of impurities is
essential to maximize turbine life and keep the system economically
viable. Gas turbine water injection equipment does not require as
stringent water quality standards as does a steam boiler. General
specifications for turbines and boiler feedwater are shown in Table 5-1
(Reference 5-3). Water treatment requirements do vary somewhat between
turbines, depending mostly on the type of fuel burned. Turbine
manufacturers often specify total contaminants allowed in the turbine,
including those contributed by the fuel, water and combustion air.
Although water quality requirements vary to a small degree, water use
requirements can vary greatly among turbines. The water-to-fuel ratio for
optimum NO reduction can vary from 0.5 to 1.0 (Reference 5-8). It
/\
depends on such factors as combustor design, plant location, heat rate,
turbine inlet temperature, fuel characteristics (primarily fuel nitrogen
content), and operating mode. On a total volume basis, the water
requirements for NO control are small. Consequently, siting of a gas
/\
turbine generating station should not become a difficult problem due to
water availability, except possibly in arid or arctic regions. However,
there is a cost impact, which is discussed in subsection 5.2.3.
Water purification systems typically employ a series of filters
followed by a reverse osmosis unit and a demineralizer or deionization
unit. Figure 5-6 shows a water treatment system sized to provide water
injection for operating five 28 MW gas turbines 10 hours per day. The
high degree of cleanup that the reverse osmosis unit attains minimizes the
size requirements for the demineralizer and deionization unit with an
attendant cost savings.
Wet controls for thermal NO reduction have thus proven to be
X
very effective in stationary gas turbines of current des.ign and with
relatively clean fuels. The effect that water or steam injection has on
incremental emissions (i.e., changes in emissions of pollutants other than
NOX) and on the gas turbine operations and maintenance requirements is
discussed in subsections 5.2.2 and 6.5, respectively.
5-10
-------
influent
474,000 liters/day
(125,000 gal/day)
I Pump
Reverse Osmosis
Module (R. 0.)
Reject water equal to
20-30% of influent
TOS = 3 to 4 times
influent concentration
75,800 liters/day
(20,000 gal/day)
effluent from R. 0.
(90-95'< reduction in IDS,
70-85% of influent)
Sewer,
River or
Evaporization Ponds
Mixed Bed
Demineralizer
Daily back wash &
rinse to regenerate
ion exchange units
19,000 liters/day
(5,000 gal/day)
Storage
360,000 liters/day
(95,000 gal/day) turbines
IDS 5 PPM
NA + K + Pb + Va 0.5 "M
Ph 6.5 - 7.5
Figure 5-6. Water treatment system (Reference 5-3)
5-11
-------
TABLE 5-1. TYPICAL WATER QUALITY SPECIFICATIONS
(Reference 5-3)
Total dissolved solids )
+ >
Non dissolved solids (ppm) )
Sodium + potassium (ppm)
Silica (ppm)
Particle size (urn)
PH
Turbine
1.0 - 5.0
0.5
0.02
10.0
7.0 - 8.5
Boiler Feedwater
0.25
0.25
0.0
—
6.5 - 7.0
5.2.1.2 Wet Control Techniques and Dirty Fuels
Water and steam injection have been demonstrated in a number of
full-scale installations to be effective in controlling NO emissions.
A
However, recent studies have shown that the effectiveness of wet controls
decreases significantly as the percentage of fuel-bound nitrogen^(FBN) in
a fuel increases. Wet controls control primarily thermal NO production,
but as fuel nitrogen increases, fuel N0x can predominate. Wet controls
can actually be counter-productive. Conn, et al. (Reference 5-9) have
shown in tests in a subscale combustor version of a commercial
Westinghouse unit, that the performance of water injection decreases
significantly with high nitrogen fuels such as solvent refined coal (SRC)
fuels. Indeed, Figure 5-7 shows that with a high water-to-fuel mass ratio
and a high-nitrogen fuel, water injection is actually detrimental to NO
reduction.
5.2.1.3 Dry Control Techniques and Clean Fuels
Controlling emissions from gas turbines can be a very difficult
task, primarily because of the nature of their operating cycles. As
engines attain maximum load, emissions of NOX are maximized. From the
operator standpoint, this is a preferred condition since overall thermal
efficiencies are also maximized. Under partial load conditions or in a
5-12
-------
-20
5 20
o
3
&-
~* 40
OJ
O)
D-
60
80
100
SRC-II Fuel
(0.94XN)
I
I
I
I
0.0 0.20
0.40 0.60 0.80 1.0
Water/fuel mass flow ratio
1.2
1.4
Figure 5-7. Predicted decrease in NOX emissions through water
injection with increasing amounts of bound nitrogen
in fuel oil (Reference 5-9).
5-13
-------
start-up/shut-down mode, NO emissions decrease, but due to poorer
A
combustion efficiency, emissions of CO and unburned hydrocarbons
increase. Consequently an emissions tradeoff situation exists in existing
combustor designs. While it is desirable to decrease NOX emissions, it
is also undesirable to increase CO emissions. This can become a
particularly important consideration in local nonattainment areas with a
cluster of gas turbines operating in the spinning reverse mode (i.e., low
load in an electric utility system). Modeling studies performed by EPA
have shown this particular arrangement to cause sufficient CO emissions to
cause local violations of the National Ambient Air Quality Standards
(NAAQS) (Reference 5-3). It must be kept in mind that this is but one
example. The factors affecting gas turbine emissions and their effect on
ambient air quality are numerous and certainly not limited to the
combustor zone reactions.
However, the combustor itself is where emissions are most
effectively reduced. The amount of NO formed in any combustion process
A
is a function of:
• Flame temperature
• Dwell time of the combustion gases at peak flame temperatures
t Fuel/air mixing
• Quantity of oxygen and nitrogen present
• Fuel atomization and vaporization
• Combustor pressure
Flame temperature increases when combustion air temperature is
raised and as the fuel-to-air ratio in the primary combustion zone
approaches stoichiometric values. Gas velocity and combustor zone
dimensions generally define the dwell times. Thus, as compression ratios
and turbine inlet temperatures increase to improve thermal efficiency and
net plant heat rate, NOX generation in the combustors of stationary gas
turbines will increase significantly. By 1986, uncontrolled thermal NO
A
generation is expected to double over today's values while by 1996, they
are expected to triple (Reference 5-10). This estimate does not take into
account N0x generated from fuel bound nitrogen. In the future, as
economics and availability reverse the present trend and make the dirtier
fuels with higher fuel bound nitrogen more attractive, then these NO
emissions estimates will increase even more (without NO controls)
5-14
-------
In chronological order, the technological advances that are expected to
increase compression ratios and turbine inlet temperatures are:
1) convectively cooled turbine airfoils, 2) precooled turbine cooling air,
3) water cooled vanes and 4) water cooled airfoils.
All of this points to the fact that a combustion modification NOV
A
control technique is required that is more acceptable to stationary gas
turbine users and manufacturers while, at the same time, meets NO
A
emissions regulations. Dry controls (any control technique that does not
involve water injection and requires design or operational modifications)
are expected to provide this capability, by some estimates in 1982
(Reference 5-11).
Among the most promising dry NO control strategies are:
A
0 Improved fuel premixing and prevaporization
0 Fuel rich combustion in the primary zone
• Primary zone heat removal
• Controlled fuel/air ratio distribution
• Extended flammability limits for ultralean combustion
• Catalytic combustion
These are general strategies and many methods exist to attain one
of these general results. For example, the general strategy of controlled
fuel/air ratio distribution can include the specific techniques of fuel
staging, air staging, fuel/air premixing and virtual staging.
Furthermore, each of these techniques responds differently to different
fuels. The discussion in this section concerns those programs which have
developed dry NO control techniques for clean fuels (i.e. those fuels
/\
low in bound nitrogen).
Much of the past and current developmental work with clean fuels is
directed towards superlean primary zone combustors. One study by the
Pratt and Whitney Aircraft Group looked at 26 potential dry NO control
A
concepts, analytically screened them, and selected the most promising for
extended experimental evaluation on bench scale hardware (Reference 5-10).
The first concept tested in bench-scale equipment was the staged
centertube — a super-lean configuration which proved effective only on
clean fuels. CO and NOX were found to be controllable in the range of
25 to 50 ppmv over a range of operating conditions. A major limiting
factor to further development was a need for variable geometry in the
5-15
-------
combustor. Additional results from this program with a rich primary zone,
lean secondary zone and clean fuels are discussed later in this section.
The results of additional work while burning dirty fuels is discussed in
Section 5.2.1.4.
As part of NASA's experimental clean combustor program, General
Electric Company and Pratt and Whitney Aircraft (References 5-12 and 5-13)
have each performed tests using advanced combustor designs to minimize
emissions of NO , CO, HC and smoke under all operating conditions, idle
A
to full power. The important results of those programs are briefly
discussed here. A more extensive summary of each program can be found in
Reference 5-3. General Electric's program involved four different
combustor designs and included such emission control concepts as lean
burning, fuel/air premixing and fuel and air staging. Pratt and Whitney
tested three different combustor designs. The emission control concepts
investigated were fundamentally similar to General Electric's, although
specific approaches varied. Both programs resulted in a maximum of
approximately 60 percent NO reduction; General Electric with a
A
radial/axial staged combustor with premix and lean primary combustion and
Pratt and Whitney with their Swirl Vorbix configuration. This particular
combustor provided for long residence time at idle to maximize CO and HC
oxidation and rapid burning and quenching at full power to minimize NO .
A
The concept of lean burning and prevaporization and premixing fuel
has been found to be very effective in reducing NO emissions when
A
burning fuels containing negligible amounts of fuel bound nitrogen. Two
aircraft combustor concepts developed by Solar proved very successful in
substantially reducing N0x emission levels while also keeping CO and HC
emissions low (Reference 5-14). The program objective was to demonstrate
that two combustor designs, the Vortex Air Blast (VAB) and the Jet Induced
Circulation (JIC), were capable of low emissions in aircraft turbines at
simulated cruise conditions. Jet-Al fuel was used in both combustors.
Ultimately the VAB combustor had lower N0x emissions, approximately
14 ppmv at 15 percent oxygen under cruise conditions of 5 atmospheres
inlet pressure and 833 K (1040°F). The JIC combustor, while giving
higher emissions, held promise for further emissions reductions with more
developmental work on the fuel preparation and mixing system. Figures 5-8
and 5-9 show the details of the basic VAB and JIC combustors.
5-16
-------
SWIRL VANES (24)
CONVECTIVE
^COOLING AIR
0.133M
(5.25 IN.)
OIA.
L_ AIR BLAST FUEL
INJECTOR TUBES (24)
(10 IN.)
Figure 5-8. VAB combustor details (Reference 5-7).
(Reproduced by permission of the General Electric
Gas Turbine Division, GER-2506, 1978, by W.M.
Knox.)
AIR BLAST
FUEL INJECTION
TUBE (4)
CONVECTIVE COOLING
<4> AIR/FUEL
MIXING PORTS
X,
TORCH IGNITER
ENTRY POINT
.0.533 U
(21 IN.)
Figure 5-9. JIC combustor details (Reference 5-7).
(Reproduced by permission of the General Electric
Gas Turbine Division, GER-2506, 1978, by W.M.
Knox.)
-------
The baseline (before design modification) emissions versus
combustor temperature rise for the VAB combustor are shown in
Figure 5-10. Subsequent design changes modified the basic VAB combustor
so that ultimately the emissions shown in Figure 5-11 were obtained.
Table 5-2 shows the design changes in chronological order and the purpose
of each change. The emission values obtained for NOX, CO and UHC were
substantially lower. NO was reduced approximately 94 percent from
A
emissions in present day aircraft combustors.
Figure 5-12 presents a review of the various JIC combustor
modifications and the effect each had on NO emissions. Table 5-3 shows
X
the basic design differences between the baseline combustor and the final
design. Further design changes have been recommended for the JIC
combustor, particularly with the intent of improving initial fuel
distribution.
It should be noted that for both the VAB and JIC combustors, smoke
emissions were not detected under all conditions.
In summary, the important program conclusions were:
t Lean burning, prevaporization and premixing concepts have been
shown to considerably reduce NO emissions from those of
A
conventional aircraft combustors
• Further development work is required to enable the combustors
to operate stably over a full range of inlet and loading
conditions. Constant geometry premixed combustors are not
capable of stable operation over a wide range of inlet and load
conditions. Concepts which may improve these problems, such as
fuel staging or variable geometry, may significantly complicate
aircraft engi nes
• The VAB combustor obtained N0x levels of approximately
14 ppmv at 15 percent oxygen
• Although JIC N0x levels were substantially lowered from
conventional combustors, they did not reach the levels
attainable with the VAB combustor.
• VAB NOX emissions, contrary to most designs, did not appear
to be pressure dependent.
• Operation of both combustors at low inlet temperatures resulted
in rising CO and increased problems with flame stability.
5-18
-------
24.0
J 22.0
Ul
= 20.0
o 18.°
| 16.0
S 14.0
*
10.0
8.0
6.0
4.0
2.0
0
VAB COMBUSTOR
FUEL: JET-A1
P= 196.5 KPA (28.5 PSIA)
T 3 831 DEC. KU036 DEC. F)
AlRFLOW = 0.326 KG/S (0.719 PPS)
CO
NOX
1800 1900 2000 2100
TEMPERATURE RISE -(F. DEC.)
2200
1000
1050 1100
TEMPERATURE RISE
1150
•
-------
TABLE 5-2. VAB DESIGN MODIFICATIONS
Design Change
Purpose
t Modification of fuel injectors
• Increased outside diameter of
reaction zone
• Increased swirler axial throat
length
• Increased combustor length
• Fuel injection modification
t Eliminate fuel streaking on
swirler walls
• Improve stabilization character-
istics of the combustor
• Increase premixing through
greater residence times
• Provide more reaction time for
unburned species (CO + HC)
t Increase the degree of premixing
at the swirler exit by improving
the initial fuel distribution
5.0
a 3.0
o 2.0
JIC COMBUSTOR
FUEL: JET-A1
_ NOMINAL COMBUSTOR INLET CONDITIONS
1.0
833
172.4 KPAI25 PSIA)
INITIAL CONFIGURATION
MULTI-POINT TUEL INJECTION
INCREASED VOLUME
REACTION ZONE
J_
J_
JL
MIXING TUBES FLUSH
WITH REACTION ZONE
J I i
« 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700
COM8USTOR TEMPERATURE RISE-(DEC F)
I I I I i ' i
700
750 800 850 900
COMBUSTOR TEMPERATURE RISE -(DEC. K)
950
Figure 5-12 JIC combustor NO, test results (Reference 5-7)
(Reproduced by permission of/the General Electric Gas
Turbine Division, GER-2506, 1978, by W.M. Knox.)
5-20
-------
TABLE 5-3. JIC COMBUSTOR CONFIGURATIONS
Item
Initial
Final
Fuel Injection
Reaction Zone
Liner Cooling
Mixing Tube
• Air blast
• Mixing tube entry center
poi nt
• Injection with airflow
direction
Outside diameter = 0.133 m
Overall length = 0.133 m
Inter-port cooling strips
Penetrating into reaction
zone
t Air blast
• 13 points on two diameters
at each mixing tube entry
t Injection against airflow
direction
Outside diameter = 0.191 m
Overall length = 0.235 m
Inter-port cooling strips
deleted
End of mixing tube flush with
inside of reaction zone
Problems with off-design operation (i.e., under conditions other
than full power), in particular with combustion efficiency, and problems
with autoignition and flashback in premixing chambers have arisen in
different combustors employing the same concept of lean burning,
prevaporization and premixing to obtain a homogeneous firing fuel/air
mixture.
Pratt and Whitney Aircraft has conducted a study involving a
premixed combustor installed in a high pressure ratio aircraft gas turbine
engine (Reference 5-15). This particular design involved a two stage
approach whereby at low power, burning at stoichiometric equivalence
ratios was done to minimize UHC and CO emissions. At high power, lean
equivalence ratios would minimize NO production. Figure 5-13 shows a
A
schematic of the staged premixed combustor with the two annular premixing
chambers. It also contains 40 primary and 40 secondary fuel injectors.
Three different design airflow distributions were tested, one being
a baseline distribution. Four different fuel injection schemes were used
as well. A conventional production type combustor was tested for
comparison purposes. N0x emissions versus combustor exit fuel/air
5-21
-------
PRESSURE
ATOMIZING NOZZLE
40 PRIMARY
FUEL INJECTORS
4O SECONDARY
FUtL INJECTORS
PREMIXING
PASSAGE
Figure 5-13. Staged premixed combustor (Reference 5-15).
(Reproduced by permission of the AIAA.)
ratios for the conventional and baseline premix airflow distribution are
shown in Figure 5-14. Only at high power conditions (high fuel/air
ratios), with all injectors firing and with a 1:2 primary to secondary
fuel split, are N0x emissions significantly below emissions from a
conventional combustor. NO emissions indices for all three airflow
A
distributions are shown in Figure 5-15. Although NO emissions for all
/\
cases are on the order of one-half of those from conventional combustors,
the difference between the three premix configurations is small.
In spite of significant NO reduction over conventional
A
combustors (approximately 50 percent), the researchers have recommended no
further developmental work with this particular approach. Many difficult
problems needed further work, including autoignition in premix chambers,
exit temperature distributions, carbon deposits and ignition. However the
basic concept of premix appears very promising and should be pursued.
Among the newer techniques being investigated for control of all
emissions, as well as more efficient combustion, is catalytic combustion.
Gas turbine combustors appear well suited for adaptation to catalytic
combustion because of high volumes of excess air and the use of clean
fuels such as natural gas and distillate fuels. A number of investigators
5-22
-------
50
40
S 30
o* '° -
Primary
Secondary
FUEL SPLIT
100/0
100/0
* 50/50
* 33/67
Fuel Injectors
Primary
Secondary
40
20
40 20
40 40
Premix
Baseline
Combustor
11 )
D
cr
a
•
Conventional
Combustor
0
00
1
1
008
IDLE
.010
APPROACH
.012 .014
CLIMB
SLTO
.016 .018
COMBUSTOR EXIT FUEL AIR RATIO
.020
.022
(Comcttd to Nominal JT9O-7A Cycle Condition*)
Figure 5-14.
i
§.
20
'6
12
i2 *
5
PHI;SEC
FUEL SPLIT
tOO'O
* SO 50
NOX emissions in baseline premix and conventional
combustors (Reference 5-15).
(Reproduced by permission of the AIAA.)
PREMIX "CONFIGURATION
FUEL
INJECTORS
PRI SEC
III
40
40
40
20
40
BASELINE QUICK QUENCH
O
a
131
MODIFIED
PREMIX ZONES
O
9
1 1
008
.010
1
IDLt
012
APPR
1 1 1
014
3ACH
.016
018
CLI
I |
.020
MB
SLT
.022
O
COMBUSTOR EXIT FUEL AIR RATIO
(Corrected to Nominal JT90-7A Cycle Conditions)
Figure 5-15. jNOx emissions in premix combustor configurations
(Reference 5-15).
(Reproduced by permission of the
5-23
-------
have studied the application of catalytic combustion to gas turbines and
the results are reported in References 5-3 and 5-4. Varying degrees of
success in reducing NO emissions while simultaneously controlling
X
emissions of unburnt species have been achieved. In one of the more
recent programs, Acurex Corporation, in conjunction with General Electric
and United Technologies, has made significant advances in development work
and is now focusing on solving specific problems before demonstration in a
full scale gas turbine combustor. Thermal NOY produced from catalytic
A
combustion in a number of different catalyst cells versus thermal NO
A
produced via thermal combustion is shown in Figure 5-16.
In surrmary, aside from catalytic combustion, which is still in a
relatively early stage of development, super!ean primary zone combustion
has been one of the main NO control concept investigated for use with
A
clean fuels in new combustor designs for existing engines. While some
results have been promising, difficulties remain with specific
approaches. Problems exist with control over the engine's load range, a
need for varriable geometry which causes significant pressure drops,
flashback problems in fuel/air premixing tube configurations, and a NO
A
pressure dependency. All these problems would be exacerbated by the
higher pressure ratios expected with engines of the future.
The Pratt and Whitney Aircraft Group, which screened 26 potential
dry NO control concepts in an EPA-sponsored study, eventually selected
A
a concept known as the rich burn/quick quench (RBQQ) combustor as the most
promising for further development. Figure 5-17 indicates the key
components in the experimental ccmbustor employing the rich burning
concept. The premix tube premixes and prevaporizes the fuel/air mixture
so that a homogeneous rich mixture is burned in the primary zone. No
additional air for combustion is provided at this stage. The quick quench
slots provide sufficient quantities of air to terminate the fuel rich
burning and transform the mixture so it has an overall lean equivalence
ratio as it enters the dilution zone. The underlying concept of fuel rich
primary burning and quick quency to lean the mixture is not new. However,
the particular methods used to test the concept in this experiment have
proven extremely effective.
5-24
-------
350
Test Model
300
250
200
u
X
-------
-Primary
Zone
Dilution Zone
Quick Quench
Slots
Tertiary
Air Ports
Figure 5-17. Rich burner arrangement (Reference 5-10).
The RBQQ combustor was first tested in a bench-scale
configuration. While burning clean #2 fuel; NO emissions were
X
extremely low -- as low as 20 ppm. CO was simultaneously low. In
addition, the NO versus fuel/air ratio curve stays low and flat once the
primary zone becomes rich, and stays flat over the entire fuel-to-air
ratio range. Significantly, this indicates that there is no need for
variable geometry. Additional results indicate that the RBQQ on clean
fuels (and dirty fuels) exhibits no dependency on pressure and, in
opposition to conventional thinking, residence time in the primary zone
must increase to reduce NO .
A
The RBQQ combustor was then scaled up to a full-size combustor
suitable for a 25 MW machine. N0x and CO emissions continued to be very
low, on the order of 40 to 45 ppm over the load range. Emissions of NO
A
were not as low as in the bench-scale hardware, however, because of
combustor length limitations with the full-scale hardware. NO could
A
not be minimized because residence time could not be maximized. Other
difficulties with the RBQQ full-scale combustor were flashback in the
premix tubes, which were replaced with air-boost nozzles, and some wall
temperature problems in the primary combustor zone.
5-26
-------
With respect to engines of the future, where pressure ratios and
turbine inlet temperatures will rise, the RBQQ combustor seems promising.
The RBQQ NO emissions are not pressure dependent nor are they dependent
A
on combustor inlet temperature.
5.2.1.4 Dry Control Techniques and Dirty Fuels
If present trends continue, availability and economics may well
cause users and manufacturers to consider the use of the less traditional
gas turbine fuels. These may include #6 residual, SRC-II, shale oil, low
Btu gas, and others. These fuels are higher in impurities such as ash,
sulfur, trace metals, and bound nitrogen than are distillate oils. How do
the evolving dry NO control combustors respond to these fuels?
X
While super-lean primary zone combustors, such as the Pratt and.
Whitney staged centertube concept, showed some success with clean fuels,
they exhibit significant drawbacks when fired on fuels containing high
bound nitrogen. Results from the Pratt and Whitney staged centertube
bench-scale testing showed a 90 percent conversion of fuel bound nitrogen
when fired on number 2 fuel oil doped to 0.5 percent nitrogen as
pyridine. Indications are that this is due to the superlean primary zone
maximizing the availability of oxygen.
Catalytic combustion has also been demonstrated in bench-scale
hardware to be somewhat effective in controlling fuel NO . Acurex
A
Corporation (Reference 5-4) in a two stage catalytic combustor has
demonstrated a low 30 percent nominal fuel nitrogen conversion rate to
NO precursors. The advantages of the two stage arrangement were that
A
the catalytic bed temperatures could be controlled for the purpose of long
life and the first stage could be operated fuel-rich to minimize reactions
to fuel NO . Also, as part of this program, a graded cell catalyst was
A
applied to a model gas turbine combustor. A variety of fuels and
pressures were tested and the results are shown in Table 5-4.
Probably the most significant progress with dirty fuels has been
made with the Pratt and Whitney-EPA RBQQ concept. The functioning of this
concept was discussed in the previous section, so it will not be repeated
here.
Figure 5-18 shows the initial results obtained with this control
concept. Note that the point of low NO (equivalence ratio 0.18) and
A
5-27
-------
TABLE 5-4. MODEL GAS TURBINE DATA SUMMARY: CATALYTIC COMBUSTOR (Reference 5-4)
Test
Point
Total
Air
(%)
Acurex Tests
0112-05
0112-06
0112-03
0112-09
-250
Space
Velocity
(1/hr)
91,500
96,600
87,300
92,400
Pratt and Whitney Tests
1976
1977
1978
1981
1982
1983
312
350
283
397
504
326
162,900
165,800
167,100
185,000
428,400
141,300
m fuel
kg/h
(Ibm/h)
5.13
(11.3)
5.68
(12.5)
5.13
(11.3)
5.40
(11.9)
7.49
(16.5)
6.86
(15.1)
8.44
(18.6)
6.75
(14.87)
12.4
(27.24)
6.24
(13.75)
m air
kg/h
(Ibm/h)
208.9
(460.3)
193.7
(426.7)
200.5
(441.7)
211.2
(465.3)
366.1
(806.4)
373.9
(823.7)
374.3
(824.4)
418.6
(922.0)
974.7
(2147.0)
317.9
(700.2)
Tcombustor
inlet
K (OF)
651
(713)
657
(724)
700
(801)
703
(806)
646
(703)
656
(722)
650
(711)
657
(723)
716
(829)
632
(679)
Tcatalyst
bea
K (OF)
1480
(2200)
1480
(2200)
1480
(2200)
1480
(2200)
—
—
—
—
—
—
F
(atm)
1.16
2.06
3.13
3.42
3.06
4.97
4.97
6.77
10.04
3.10
CO
(ppm)
0
0
0
0
10
9
10
110
23
9
NO
(ppm)
2
1
3
2
2
1
2
1
1
0
UHC
(ppm)
--
--
--
0.6
0.3
0
0.3
0
0.6
Fuel
Propane
• Propane
en
i
CO
T-747
-------
TABLE 5-4. Concluded
Test
Point
Total '
Air
W
Space
Velocity
(1/hr)
Pratt and Whitney Tests
(continued)
1985
1986
1987
1988
1989
1991
1992
1993
860
752
829
819
1277
541
583
573
288,400
291,900
644,000
901,100
276,600
291,900
564,000
699,500
m fuel
kg/h
(Ibm/h)
5.36
(11.8)
6.22
(13.7)
12.5
(27.5)
17.5
(38.5)
3 50
(7.7)
8.63
(19)
15.4
(34)
19.5
(43)
m air
kg/h
(Ibm/h)
669.7
(1475)
677.8
(1493)
1496
(3294)
2093
(4609)
642.4
(1415)
677.8
(1493)
1310
(2885)
1624
(3578)
Tcombustor
inlet
K (°F)
752
(894)
756
(901)
710
(819)
634
(681)
745
(881)
687
(778)
674
(754)
670
(746)
Tcatalyst
bed
K («>F)
1200a
(1700)
1280a
(1850)
12803
(1850)
1140a
(1600)
11409
(1600)
—
—
—
P
(atm)
2.99
3.03
5.21
7.01
2.96
3.06
5.07
6.77
CO
(ppm)
1195b
710
1592
2202
1860
82
1285
1362
NO
(ppm)
3
5
3
3
44
145
74
68
UHC
(ppm)
79.5
34.7
23.9
222.6
High
80.1
24.8
15.6
Fuel
No. 2 oil
Propane
• No. 2 oil
+ pyridine
ro
10
aBed Temperature estimates due to bed nonuniformities.
t>High CO and UHC resulted from operating at low bed temperatures to avoid flameholding.
T-747
-------
1000
c_n
I
CO
O
0.1
0.2 0.3
Overall equivalence ratio
0.4
Figure 5-18. Rich burner characteristics, 345 kPa (50 psia), 590K (600°F),
0.5% nitrogen (Reference 5-8).
(Reproduced by permission of EPRI.)
-------
the point of high CO (equivalence ratio 0.12) do not coincide.
Furthermore, at an overall equivalence ratio of 0.18, indicating fuel rich
burning in the primary zone, emissions of NOV and CO measured only
A
50 ppmv corrected to 15 percent 02» substantially below the proposed
standard of 75 ppmv at 15 percent 02 and the program goal of 100 ppm.
This was accomplished when burning No. 2 distillate oil containing
0.5 percent fuel bound nitrogen. Although these tests were performed at
air pressures of 345° kPa (50 psia), and temperatures of 590 K (600°F)
substantially below those encountered in full scale stationary gas
turbines, the researchers have obtained similar results at more realistic
values of 130° kPa (150 psia) and 670 K (750°F). Apparently the fuel
rich/quick quench concept is not subject to the same kinetic constraints
that greatly increase NO production with lean burning and increasing
air temperatures and pressures. Figure 5-19 indicates the results
obtained at increased temperatures and pressures which more closely
simulate the conditions found in a stationary gas turbine.
Subsequent tests of fuels with bound nitrogen, which involved the
use of a variable damper to control primary airflow, were successful in
isolating NO control from CO control. While NO is controlled by
A A
primary zone stoichiometry and residence time, secondary zone temperatures
appear to control CO. This point is extremely significant because it says
that NO control does not have to be traded off for CO control.
A
Another significant characteristic of the RBQQ is that there is a
major dependence of NOY on residence time in the primary zone as shown
A
in Figure 5-20. Surprisingly, the NO decreases with the increasing
A
residence time. This relationship, however, proved to be a major limiter
to the RBQQ concept in realizing its full potential -- combustor length
could not be extended enough to provide the maximum residence time to
minimize NO . Low emission levels were nevertheless attained.
A
Tests were subsequently performed in a RBQQ combustor scaled up to
fit a 25 MW machine. Operating pressure was 689 kPa (100 psi) and
combustor inlet temperatures were up to 589 K (600°F). A range of
turbine inlet temperatures were tested up to a maximum of 1700 K
(2600°F). The fuels tested were SRC II (0.96 percent N), residual shale
oil (0.46 percent N), Indonesian/Malaysi an residual oil (0.24 percent N),
and distillate oil. Tests were run with and without a premix section.
5-31
-------
en
i
OJ
ro
CM
o
O
O
Q-
Q_
1000
600
400
200
100
60
40
20
10
7% Primary Air
at 670K (750 F)
NO.
J
CO
14% Primary Air
at 620K (650 F)
CO
0.1 0.2
Overall equivalence ratio
0.3
0.4
Figure 5-19. Rich burner simulated engine cycle characteristics, 1030 kPa
(150 psia), 0.5 percent nitrogen (Reference 5-8).
(Reproduced by permission of EPRI.)
-------
en
i
co
co
100
80
C\J
o
60
a.
CL
CO
c
o
to 40
20
590 K, 340 kPa
(600°F, 50 PSIA)
4% Pressure Drop
Fuel: Number 2 Oil with 0.5% N
Neat Number 2
O
'V
0.04
0.08
0.12
0.16
0.20
0.24
0.28
0.32
Primary Zone Residence Time-Sec (Base on Inlet Density)
Figure 5-20. Variation of Minimum NOX Emissions with Primary Zone Residence Time (Reference 5-39).
-------
NOX emission results (at 15 percent 02) with the premix xection
were 92 ppm for SRC II, 65 ppm for residual shale oil, 75 ppm for
Indonesian/Malaysian residual, and 42 ppm for distillate. The same
combustor with an air-boost nozzle and a straight swirler on the outside
gave NO emissions of 80 ppm for SRC II, 65 ppm for Indonesian/Malaysian
J\
residual, and 42 ppm for distillate. Residual shale was not burned in
this configuration.
Combustor tests without a premix section and turbine inlet
temperatures up to 1644-1700 K (2500-2600°F) gave NOX emissions
slightly higher than with premix. However, normal design limits for the
primary zone metal temperature were exceeded.
5.2.1.5 Summary
Wet controls for reduction of NOX from stationary gas turbines
are currently the only commercially available techniques. Emissions can
easily be reduced to levels below the NSPS by controlling the amount of
water or steam injected into the combustor primary zone. Wet controls
work well with clean fuels. However, as the amount of fuel nitrogen in
the fuel increases, wet controls become increasingly less effective, to
the point where they are actually detrimental to NOV reduction if the
A
percent nitrogen is high enough. Clean fuels, such as distillate oils,
are still relatively available, economical, and preferred for gas
turbines. But as economics and availability change, users may be put in a
position where they must consider dirtier fuels.
While wet controls may be sufficiently effective to reduce NO
A
below NSPS levels on fuels containing minimal bound nitrogen, some form of
dry controls will be required for dirtier fiels. Furthermore, users would
much prefer not to have to deal with the added expense, maintenance, and
operation problems attendant with water and steam injection. Thus, it
seems clear that dry NOX controls will replace wet controls within five
years. Dry control development seems to be heading in two directions,
superlean primary zone combustors for clean fuels and the RBQQ concept for
dirty fuels. RBQQ, in particular, appears to show a great deal of
promise. NOX -js kept well below the NSPS level despite the presence of
fuel nitrogen, while CO is simultaneously kept at a low level. RBQQ has
been successfully demonstrated in full-scale hardware on clean and dirty
fuels, such as SRC-II and residual shale oil. But further developments
5-34
-------
are needed in heat transfer problems and the combustor outlet pattern
factor.
Superlean combustors are making significant progress with clean
fuels, although some problems exist with controls over the load range,
premix tube flashbacks, and a pressure dependency which would intensify as
pressure ratios increase with future engines.
5.2.2 Incremental Emissions of Pollutants Other than NO.,
"••^™M^H«^HH.«MmKM•_^^vj^
The wet and dry NO control techniques were discussed in the
A
previous section in terms of their effectiveness in reducing NO
A
emissions. However, it must be noted that the same changes that affect
NO also can affect the other pollutants discharged from the system. If
A
the magnitude of change is ignored, the pollutants may have an adverse
effect on the environment. The changes which occur are called incremental
emissions. Control of incremental emissions can affect the overall system
thermal efficiency and performance. Control of CO and unburned
hydrocarbons are an integral part of NO control in new combustor
A
designs. Manufacturers are not likely to compromise combustion efficiency
for the sake of NOV control.
A
The pollutants of concern are the criteria pollutants, CO, UHC,
particulate (mass emission rates and particle size distribution), SOo,
and the noncriteria pollutants, sulfates, organics and trace metals. The
following subsections present the postulated formation mechanisms of these
pollutants and the available data supporting the postulations.
5.2.2.1 Carbon Monoxide and Unburned Hydrocarbons
CO and UHC in combustion product gases results from incomplete fuel
combustion. Generally, the conditions which lead to incomplete combustion
fall into the following categories:
t Insufficient oxygen available for reaction
• Incomplete fuel/air mixing
• Reduced overall flame temperature
t Decreased combustion gas residence time
t Cold wall impingement (resulting in reduced temperatures)
Any combustion modification which leads to any of the above can result in
increased CO and UHC emissions.
5-35
-------
Thus, some increase of CO and UHC with water injection may happen.
The heat required to vaporize the water, as well as dilution effects, tend
to impede combustion and result in lower combustor temperatures. This can
lead to reduced combustion efficiency and increased CO and UHC emission
levels. It is postulated that steam injection will also increase CO and
UHC emission levels, but to a slightly lesser degree due to the smaller
heat load required for vaporization.
Documented effects of water injection on CO emissions from gas
turbines are shown in Table 5-5. There appears to be only a slight trend
toward increased CO emissions with water injection for distillate and
diesel oil firing. However, the significant CO increase for natural gas
firing may be of concern, but it should be noted that the data are very
limited.
Unburned vapor phase hydrocarbon emissions from combustion sources
are of environmental concern because of their role in the atmospheric
reactions leading to photochemical smog. These hydrocarbons include
aliphatic, oxygenated, and low molecular weight aromatic organic compounds
which exist in the vapor phase of flue gas at noncondensing temperatures.
Thus these hydrocarbons include such organic compounds as alkanes,
alkenes, aldehydes, carboxylic acids, and substituted benzenes.
Like CO emissions, unburned hydrocarbon emissions are a function of
combustion completeness. Therefore, since water injection NO control
A
tends to impede complete combustion, hydrocarbon emissions can be expected
to increase with increasing water injection. Table 5-6 presents the
limited data available. Though not conclusive, a general trend towards
slightly increased unburned hydrocarbons is noted.
The documented effects of dry NOX controls on CO and UHC
emissions appear to indicate that both super!ean combustors and the RBQQ
combustor are capable of minimizing unburned pollutant species to
acceptable levels. There are, however, qualifications. For superlean
configurations, CO and UHC emissions can be minimized if the engine
control system is properly executed so that in response to high
temperatures, the gas stream is not overquenched, causing a rise in CO and
UHC. In addition, any deterioration in components could affect combustion
efficiency, thereby inducing CO and UHC formation. Figure 5-21 shows the
characteristic "scissors" curve of CO and NO in a superlean bench-scale
A
5-36
-------
TABLE 5-5. EFFECT OF WATER INJECTION NOX CONTROL ON CO EMISSIONS
FROM A GAS TURBINE (References 5-3)
Fuel
Natural gas
Distillate oil
Diesel
CO Emissions (ppm)a
Baseline
147
252
17
1,174
99
135
93
Water Injection
1,134
1,512
22
1,286
144
162
30
aAt three percent 02, dry basis.
TABLE 5-6. EFFECT OF WATER INJECTION NOX CONTROL ON UHC
EMISSIONS FROM A GAS TURBINE
(References 5-3 and 5-23)
Fuel
Natural gas
No. 2 Distillate
oil
Diesel
UHC Emissions (ppm)a
Basel i ne
234
141
36
8
24
Water Injection
372
246
27
10
12
aAt three percent Og, dry basis.
5-37
-------
200
a.
O.
o>
4J
O
O)
150
100
50
590 K (600°F)
0.1
Overall Equivalence Ratio
0.2
CL
Q.
ai
+->
o
a>
i.
L.
o
200
150
100
50
700 K (800°F)
NO
0.1
Overall Equivalence Ratio
0.2
Figure 5-21. Lean burner high pressure characteristics, 689 kPa
(100 psia) (Reference 5-16)
5-38
-------
combustor (Reference 5-10) at two inlet temperatures while burning clean
fuel. Due to the steep slope of the CO curves, any change in fuel/air
ratio, caused by such things as an improper control system response or
component wear, can cause significant increases in CO. UHC would respond
in a similar manner.
The Pratt and Whitney-EPA RBQQ combustor is able to limit CO and
UHC while minimizing NO by careful control of airflow in the primary
A
zone and temperature in the secondary zone. Figure 5-22 shows results
from bench-scale results where low NO concentrations have been obtained
A
over a range of overall equivalence ratios while CO remains constant (the
dashed line are values obtained at a lower overall air flow rate). These
results are extremely significant because NO control and CO control are
/\
isolated from each other. Results from full-scale tests were not quite as
successful. While NO remained low, CO increased due to a considerably
A
shortened combustor and a revamped secondary mixing pattern. If combustor
length, and thus residence time, could be increased and secondary
temperatures could be held high enough, minimizing CO should be no problem.
>
o.
o"
Ol
O
-------
5.2.2.2 Particulate Emissions
Although gas-fired units produce negligible amounts of particulate,
distillate fuel oil and "dirty" oils such as residual oil can emit
significant concentrations of particulate (Reference 5-17). Therefore the
impact of NO control techniques on particulate emissions can be of
A
concern.
The formation of particulates in a combustion source is intimately
related to combustion aerodynamics, the mechanisms of fuel/air mixing, and
the effects of these factors on combustion gas temperature-time history.
The optimum conditions for reducing particulate formation (intense, high
temperature flames as produced by high turbulence and rapid fuel/air mixing)
are not the conditions for suppressing NO formation. Therefore, most
A
attempts to produce low NO combustion designs have been comprised by the
A
need to limit formation of particulates or smoke (Reference 5-18).
Particulate emissions from oil-fired gas turbines can be composed
of soot (condensed organic matter) and ash (incombustible mineral
matter). Thus, particulate mass emissions are generally a function of the
completeness of combustion given that the ash content of emitted
particulate solely depends on the ash content of the fuel burned. When
controls for NO that reduce combustion efficiency are applied, an
A
increase in condensible organic matter (soot) can be expected.
The data on particulate emissions from gas turbines using wet
controls are very limited and contradictory. The effect of water
injection appears to be related to the specific injection method used and
on load. The manner in which particulates vary as wet controls are
applied, is, in general, similar to the effect on CO and HC emissions.
The same conditions which promote oxidation of CO and HC also promote
oxidation of filterable (carbon soot) and condensable (heavy hydrocarbons)
particulate matter. One may therefore expect a decline in combustion
efficiency to increase amounts of particulate emissions. The available
data, however, do not support or refute this conclusion.
The data on particulate emissions from gas turbines resulting from
applied dry NOX controls are also very limited. However, the available
data indicate that incremental particulate emissions from NO controls
A
follow trends similar to those for incremental CO and hydrocarbon
emissions.
5-40
-------
Figure 5-23 shows the effect of turbine load on particulate
emissions. Like CO and HC emissions, particle emissions increase as
turbine load is reduced. As the figure shows, particulate emission
increases average 40 percent when turbine load is decreased to 30 percent
of rated capacity.
5.2.2.3 Sulfates
Ambient sulfate levels are a matter of increasing concern in
regions with large numbers of combustion sources firing sulfur-bearing
coal and oil (notably, the northeast region of the U.S.). Although the
direct health effects of high ambient sulfate levels are currently
unclear, high sulfate aerosol concentrations are known to decrease
visibility and aggravate acid precipitation phenomena.
Ambient sulfates are comprised of directly emitted sulfates
(primary sulfates) and those derived from the atmospheric oxidation of
SO* (secondary sulfates). SOp arises from the sulfur contained in the
fuel. Essentially all sulfur entering the combustor is discharged as
SOp- The primary concern for incremental emissions, then, is to control
the ratio of primary sulfate to S02 (S04/S02)- The sulfate present
may exist as either sulfuric acid (H2S04) or as metal or ammonium
sulfates. The potential for internal corrosion in gas turbines from
sulfates is great, and thus unacceptable. Consequently the sulfur content
of the fuel is controlled.
The low concentrations of sulfur in the primary gas turbine fuels
and the normally high operating efficiency of gas turbines have resulted
in little concern for the level of these emissions. Data showing the
effects of wet controls on sulfate levels are extremely limited. However,
results from a recently completed test program performed on a 60 MW
utility gas turbine equipped with water injection indicated that the NO
n
control technique had no significant effect on SOo or SO* emissions
(see Section 6.1).
5.2.2.4 Organics
Organics are defined as those organic species not included in the
criteria pollutant class of unburned vapor phase hydrocarbons. These
remaining organic emissions are composed largely of compounds emitted from
combustion sources in a condensed phase. These compounds can generally be
classed into a group known either as polycyclic organic matter (POM) or
polynuclear aromatic hydrocarbons (PNA or PAH) (References 5-18 and 5-19).
5-41
-------
30
r
25
20
CO
c
O
•r—
to
to
E 15
01
-------
POM emissions have significant environmental impact because several
species are highly carcinogenic (Reference 5-19). The fact that they
generally exist as fine participate makes them an even more serious health
hazard.
Although polycyclic organic matter can conceivably be formed in the
combustion of any hydrocarbon fuel, it is considered more of a problem
when associated with soot (carbonaceous particulate) emissions from coal-
and oil-fired combustion equipment. Thus, POM production is of only minor
concern in gas-fired gas turbine systems and of some concern in oil-fired
systems.
5.2.2.5 Trace Elements
Emissions of trace metals are generally a concern only from
combustion sources firing coal and residual oil. They are a lesser
problem in oil-fired gas turbines since these sources tend to fire
distillate fuels, although present trends indicate an eventual shift to
coal derived synthetic fuels. Trace metal concentrations in distillate
oils are generally much lower than those in residual oils and synthetic
fuels. This, coupled with the fact that fuel specifications limit the
level of certain trace elements in the fuel fired, suggest that trace
element emissions from gas turbines should present no problems under
either controlled or uncontrolled operation.
5.2.2.6 Summary of Incremental Emissions
CO and UHC are products of incomplete combustion and result from
dropping temperatures too rapidly. An engine at idle and low power will
produce high CO and UHC because combustion efficiency is low. Full load
produces high combustion efficiencies and therefore low CO and high UHC.
While data demonstrating the effect of NO controls on CO and UHC
/\
emissions is limited, trends seem to indicate that water and steam
injection tend to increase these emissions. Dry controls, such as
superlean and RBQQ both appear to be capable of minimizing CO and UHC, but
each type of combustor has its limitations that need to be corrected
before they become commercially available.
5.2.3 Incremental Costs of NO., Controls
^^.^^^-^M^«««V^
The pollutants of concern with stationary gas turbines are
primarily N0¥, CO and UHC. NO is the major pollutant from gas
^ A
turbines; CO and UHC are already controlled for efficiency reasons.
5-43
-------
CO and UHC are emitted in small quantities and S02 emissions, emitted in
quantities directly proportional to fuel sulfur content, is best
controlled by regulating the amount of fuel sulfur. Gas turbines put out
a large amount of gas due to the very large excess air requirement. This
causes the cost of SCU flue gas scrubbers to be prohibitively high
(Reference 5-3). The proposed S02 emission standard for gas turbines
limits either SCL exhaust emission levels or sulfur in the fuel. Since
the sulfur content of most distillate oil is below the standard of 0.8
percent, there should be no problem as long as distillate oil is burned.
Burning residual oil could present more of a problem, but it has been
estimated that 85 percent of all residual oil is below 0.8 percent sulfur
(Reference 5-3). There will be a cost if users are forced to analyze
their fuel for sulfur on a daily basis, especially when the turbines are
used in remote areas. However, except for remote areas, this cost should
be small. In any event, it appears that the sampling requirement may be
modified where it would cause undue hardship (Reference 5-20).
The following sections detail estimates of NO control costs. In
A
estimating the equipment and operating cost of NO controls, there has
A
been some limited operating experience with water and steam injection on
large utility size turbines. For the smaller size turbines, there has
been very little operating experience with wet controls. Therefore the
costs given for the smaller units may not be as accurate as the costs for
the utility size units.
This report uses the same costs and procedures as in the SSEIS
report on gas turbines. However, the costs reported here have been
updated from user experience collected since the SSEIS was published. Wet
controls are discussed first and then dry controls. Since costs are site
dependent, the costs given can only be considered typical, not the exact
cost estimate for a particular site. As discussed above for sulfur, it is
assumed that costs required for fuel analysis will be minimal.
5.2.3.1 Wet Controls
This section describes the costs associated with wet methods for
controlling NO emissions from gas turbines. It will be assumed that
A
water will always be available. Since the proposed standard will probably
5-44
-------
be modified where there is a limited water supply (Reference 5-20), this
seems a reasonable assumption. As in the SSEIS report, it will be assumed
that there are no additional costs due to water pollution problems.
Wastes are produced when the water purification systems that are used for
the water injection systems are cleaned or recharged, but there should be
few problems in disposing of this waste. This is especially true for
utility installations which already have to purify large amounts of boiler
feedwater as well as dispose of the wastes. In many cases the existing
sewer system should be able to handle the wastes produced. One user
reported trouble with the waste from recharging the purifiers
(Reference 5-21) but that was due to lack of experience with the equipment
rather than inherent system problems.
In the following subsections, costs are first given for utility
size and then smaller size units. Since information on wet controls for
the smaller size turbines is limited, small and medium size turbines are
combined.
Utility Turbines
For wet controls on utility size turbines, the most expensive cost
is for the water purification system. The water injected into the turbine
must be very clean to prevent damage to the turbine. Utility plants have
systems for supplying water to their boilers. If there is enough excess
capacity, then the cost for the water purification system can be
lessened. This was not assumed in the cost analysis, so it might be
possible to obtain a lower cost than listed below. No costs are given for
steam generating equipment because steam injection would only be used with
a combined cycle plant where there is an existing source of steam.
The following three tables list costs for water and steam injection
systems as quoted by users. Table 5-7 lists 1973 costs reported by San
Diego Gas and Electric to control their gas turbines (Reference 5-22).
The turbines themselves cost about $100/kW in 1973. Also, since these
costs are for utility use, the turbine cost also includes the cost of the
electric generator. Tables 5-8 and 5-9 list costs quoted by the City of
Pasadena, California Power and Water Department and the City of Glendale,
California, Public Service Department, respectively (References 5-23 and
5-24). The cost for the water injection equipment averaged 5 to 10 percent
of the cost of the turbine itself. Based on Tables 5-7 through 5-9, a
5-45
-------
TABLE 5-7. 1973 WATER INJECTION INVESTMENT COST (SAN DIEGO GAS
AND ELECTRIC) (Reference 5-22)
Control System
Combustor modifications
including water injection
nozzles
Water injection pumps and
water control system
Associated piping and water
storage facilities
Water treatment equipment
General expenses including
engineering, administration,
testing, taxes
TOTAL
Gas Turbine Size
20 MW
$1.00/kW
$3.54/kW
$1.72/kW
$0.90/kW
$1.15/kW
$8.31/kW
49 MW
$0.86/kW
$2.88/kW
$1.05/kW
$0.47/kW
$0.82/kW
$6.07/kW
81 MW
$1.04/kW
$3.10/kW
$0.87/kW
$0.47/kW
$0.57/kW
$6.05/kW
TABLE 5-8. 1975 WATER INJECTION COST (CITY OF PASADENA)
(REFERENCE 5-23)
Cost Item
Cost for Two
26 MW Units
Water purification system
Other costs
Total per turbine
Cost/kW
$100,000 ($50,000/turbine)
$50,000
$100,000
$3.85
5-46
-------
TABLE 5-9. WATER AND STEAM INJECTION COSTS
(CITY OF GLENDALE, Reference 5-24)
Water injection costs (equipment only) (1973)
$300,000 out of $3,500,000 for 31 MW = $9.68/kW
Steam injection costs (equipment only) (1976)
$1,000,000 out of $20,000,000 for 120 MW
combined cycle (gas turbine produces 90 MW) = $8.30/kW
1978 cost of $10/kW appears appropriate for the water injection system.
The SSEIS (Reference 5-3) used a figure of $2.58/kW in 1975 dollars.
However, since the water injection systems listed in the SSEIS were
installed around 1975, this figure appears too low. Energy and
Environmental Analysis (EEA) has reevaluated water injection costs for the
EPA and has given an estimate of $7/kW for a 60 MW gas turbine
(Reference 5-25). Thus, it has been decided to use $10/kW, in 1978
dollars, in the present analysis. Even so, for some installations, even
this figure might be too low. San Diego Gas and Electric is now
estimating that it will cost $23/kW to convert an existing 32 MW gas
turbine to water injection (Reference 5-26).
Assigning typical incremental operating and maintenance costs as a
result of wet NOX controls is a more difficult task. Section 6.5 covers
the changes in operations and maintenance due to the wet controls. As
described there, it is unclear if there is any significant increased
maintenance due to the water injection other than the maintenance for the
water system itself. Therefore it will be assumed that there is no
increased maintenance cost for the turbine due to the water injection
system. From data compiled on operating and maintenance costs for
turbines (References 5-27, 5-29, 5-30), a figure of 3 percent of installed
costs per year appears more reasonable than the one percent used in the
SSEIS (Reference 5-3). A fixed charge of 20 percent of installed costs
per year will be used, the same value used in the SSEIS report.
5-47
-------
The greatest operating cost impact due to wet controls results from
the fuel penalty caused by lowered thermal efficiency. Estimates of the
drop in thermal efficiency range from zero to five percent (References 5-3,
5-23, 5-28, 5-31, 5-32). A figure of two percent, an approximate average,
is used in the calculations. There is also an increase in power output of
the turbine due to water injection. It will be assumed that the utility
company can use this increased capacity and a capital charge credit of two
percent will be used. A figure of $0.79/m^ is used for the cost of the
water, which agrees with the SSEIS report (Reference 5-3) and one user's
experience (Reference 5-23). A figure of 10.9 MJ/kWh is used for the heat
rate of a simple cycle gas turbine. This value was also used in the SSEIS
report. In actual practice, most existing turbines have heat rates of
12.7 to 16.9 MJ/kWh (References 5-27, 5-29), although new turbines are
expected to be operating at the lower heat rates (Reference 5-32).
The turbine is assumed to be operating on No. 2 oil at a cost of
$2.84/GJ (Reference 5-33). Table 5-10 estimates the cost of electricity
from an uncontrolled turbine. Table 5-11 estimates the increased costs
due to the water injection system. Based on the costs cited in
Table 5-12, it is estimated that water injection will increase the cost of
electricity by 4 percent from a peaking unit, 2.7 percent for a midrange
unit and 2.6 percent for a base load unit.
Smaller Gas Turbines
For turbines smaller than utility size, there has not been any
known long term use of water injection. Therefore, there are no operating
data to base the costs on. The same assumptions that were made for
utility size turbines when calculating the costs will be made for the
smaller units. Any significant changes are noted. Gas turbines in
pipeline service operate unattended (Reference 5-34), thus costs
associated with these applications will be estimated assuming unattended
operation. One utility operator was able to let his turbine run
unattended except for recharging the water purification system
(Reference 5-28). It is reasonable to expect that when the pipeline gets
its weekly check, the water system could also be checked.
In calculating the operating costs, the following changes in the
assumption used for utility turbines were made. A heat rate of
12.5 MJ/kWh (13,200 Btu/kWh) and a maintenance and operating charge of
5-48
-------
TABLE 5-10. ESTIMATED OPERATING COST OF UNCONTROLLED UTILITY GAS
TURBINE (50 MW), 1978 DOLLARS
Use
Installed cost ($/kW)
Heat rate (MJ/kWh)
Fixed costs (% of installed cost)
Operating and maintenance (% of
installed cost)
Fuel cost ($/GJ)
Costs
Fixed, operating, and maintenance
mills/kWh)
Fuel (mills/kWh)
Total (mills/kWh)
Peaking
150
10.9
20
3
2.84
69
30.9
100
Midrange
150
10.9
20
3
2.84
8.6
30.9
40
Basel oad
150
10.9
20
3
2.84
4.3
30.9
35
TABLE 5-11. ESTIMATED COST OF WATER INJECTION FOR UTILITY
GAS TURBINES, 1978 DOLLARS
Use
Size (MW)
Hours of operation per year
Water costs ($/m3)
Water feed rate (m3/h)
Installed cost ($/kW)
Annualized costs (mills/kWh)
Fixed, operating, and maintenance
Water cost
Fuel penalty
Capacity enhancement
Total (mills/kWh)
% change over uncontrolled
Peaking
50
500
0.8
7.6
10.0
4.6
0.1
0.6
-1.3
4.0
4.0
Midrange
50
4000
0.8
7.6
10.0
0.6
0.1
0.6
-0.2
1.1
2.7
Basel oad
50
8000
0.8
7.6
10.0
0.3
0.1
0.6
-0.1
0.9
2.6
5-49
-------
TABLE 5-12. ESTIMATED COSTS FOR UNCONTROLLED INDUSTRIAL GAS TURBINE,
1978 DOLLARS
Use
Size (kW)
Installed cost ($l/kW)
Heat rate (MJ/kWh)
Fixed costs (% of installed cost)
Operating and maintenance
(% of installed cost)
Fuel cost ($/GJ)
Hours per year operation
Annual ized costs (mills/kWh)
Fixed, operating, and maintenance
Fuel
Total
Midrange
2775
300
14
20
5
2.84
4000
18.75
39.8
59
Basel oad
2775
300
14
20
5
2.84
8000
9.38
39.8
49
5 percent were used, in agreement with the SSEIS. There is a wide
difference between costs quoted for the water injection system between
EPA's estimates in the SSEIS report, $7.50/kW (Reference 5-3) and two
manufacturers, $30/kW (Reference 2-35) and $35/kW (References 2-36 and
2-37). The SSEIS prices are based on estimates made in 1974 by
manufacturers while the process was still in the development stage for
large turbine applications. For the present report, a cost of $30/kW will
be used as the cost of wet controls for a typical 2.8 MW unit. Costs are
only given for a 2.8 MW unit because, except for the very small units,
they all have about the same cost per kW. Table 5-12 shows the costs for
4000 and 8000 hour operation while Table 5-13 shows the increased cost due
to water injection. The total cost of the turbine depends on how it is
used (i.e., for generation, pumping, compressing, etc.). A cost of
$300/kW is used but depending on application, this cost can be
significantly higher. Water injection causes an increased charge in kWh
5-50
-------
TABLE 5-13. ESTIMATED COSTS OF WATER INJECTION FOR INDUSTRIAL GAS TURBINE,
1978 DOLLARS
Use
Size (kW)
Installed costs ($/kW)
Hours per year operation
Water cost ($/m3)
Flowrate (m3/h)
Annuali zed costs (mills/kWh)
Fixed, operating, and maintenance
Fuel penalty
Water costs
Total (mills/kWh)
% of uncontrolled
Midrange
2775
30
4000
0.8
0.4
1.87
0.80
0.10
2.8
4.7
Basel oad
2775
30
8000
0.8
0.4
0.94
0.80
0.10
1.8
3.7
delivered of 4.7 percent for a 3 MW unit and 3.7 percent in a 6 MW unit.
No credit is taken for increased capacity.
5.2.3.2 Dry Controls
Since dry controls are still in the developmental stages, the costs
cannot be accurately estimated and only rough estimates are possible. One
manufacturer of utility size turbines (Reference 5-32) hopes to design a
new type of combustor liner to replace existing liners. They hope that no
new air or fuel controls would be needed and no additional maintenance
would be required. If they are successful, the only additional costs
would be the development costs. Since combustion liners are routinely
changed every several years, a large base is provided to spread
development costs over. Thus, dry controls for utility boilers will
probably be cheaper than wet controls, though this statement is based on
unproven results.
For smaller turbines, one manufacturer (Reference 5-37) estimated
equipment costs for dry controls to be as high as those for wet controls.
This is partly due to the small market for 3000 kW turbines, consequently
5-51
-------
there is a small base for spreading development costs over. They
estimated a cost of $95,000 for dry controls for turbines in the 3 MW
range. Another manufacturer (Reference 5-38) has estimated that the cost
of dry controls to the customer will be small. Large development costs
could, however, force this manufacturer out of the American market. A
full scale turbine with dry controls to meet the proposed NOX NSPS has
not yet been built so these costs are only rough estimates. Since there
should be no fuel penalty with dry controls, operating costs and cost per
kWh generated should be less than for wet controls. Using the same
figures given in Table 5-13 but subtracting out the fuel penalty, for
4000 hours per year operations, there would be a 3.4 percent increase in
the cost of power delivered. For 8000 hour/year generation, there would
be a 2.1 percent increase in the cost of power delivered.
5.3 REGIONAL CONSIDERATIONS AFFECTING CONTROL SELECTION
Regional considerations refer to situations or conditions external
to turbine operation which could adversely affect the operation of NOV
/\
control equipment, limit the degree of control or have a significant
impact on control costs. The following section discusses the types of
problems that one must consider when evaluating various control
alternatives. Fuel and equipment specific considerations are treated
separately.
5.3.1 Fuel Considerations
As discussed in Section 4, the composition of the fuel burned in
the gas turbine will have a large impact on the emission rate of NO .
A
The fuel composition also determines to a great extent the emission rate
of SOV, particulate, and trace metals.
/\
A potential partial solution for meeting the New Source Performance
Standards could be fuel switching. Generally fuel changes would be from a
"dirty" high pollutant potential fuel, to a "clean" low nitrogen, low
sulfur fuel. Fuel switching generally is an economic consideration;
substituting a more highly processed, higher cost fuel for a "dirty" low
cost fuel. The primary regional consideration would be the availability
of distillate oils and natural gas. Certain regions, such as the
northeastern United States, are faced with significant clean fuel
availability problems during the winter months. If circumstances occur
5-52
-------
which would reduce the availability of clean fuels, fuel switching may not
be a viable solution to NO , SO and particulate emissions.
An alternative to fuel switching is fuel treatment. As discussed
in Section 4, ash-bearing fuels and fuels with high levels of other
contaminants can be treated to reduce the pollution potential. Treatment
methods are available to reduce particulate, sulfur, and trace element
contaminants. However, these processes generate liquid waste streams
which are high in dissolved and suspended solids which must be disposed of
in an acceptable manner. Situations may occur in certain regions where
regulations governing liquid waste disposal may prohibit discharge of the
waste to municipal waste treatment facilities. The added burden of fuel
treatment and additional waste disposal systems may make this option
economically unattractive.
5.3.2 Equipment Considerations
Certain regions of the United States, such as the South Coast Air
Quality Management District (SCAQMD) have substantially more stringent
control requirements than other regions. Though NO control technology
n
is available to meet these control levels, there will be an increased
burden on existing equipment. For example, a higher water/fuel ratio
would be required for wet control of NO . This may result in increased
A
water volume requirements, larger water treatment facilities, potential
increases in equipment corrosion, and an increased fuel penalty. Existing
water treatment facilities must be able to meet the new water volume
requirements. Operation and maintenance requirements may be increased,
and water and fuel usage will be increased. If the turbine is located in
a region where water and fuel supplies are limited, or existing water
treatment facilities are straining to handle the increased burden, costs
may rise substantially.
Water availability alone could be a significant factor in
determining N0x control strategy. In regions with limited water
supplies, such as arid or arctic regions, it may not be reasonable to
require the use of water or steam injection. A second water related
problem is the increased volume of treatment waste from the water
purification system. If facilities are limited to handle this waste
material, water or steam injection may not be viable control options.
5-53
-------
Reductions in NO emissions are generally associated with slight
A
reductions in combustion efficiency. This generally results in increases
of CO and UHC emissions as discussed in Section 5.2.2. Certain regions
could be adversely affected by increases in incremental emissions such as
unburned vapor phase hydrocarbons because of their contribution to
photochemical smog and other health related situations. Those areas with
pollution problems of this type may limit the emissions of these other
criteria pollutants and thus substantially limit the effectiveness of
NO control devices.
A
It is thus quite important when considering NO control
A
technology to examine not only the basic system requirements, but also the
impact the control may have on other plant facilities and the surrounding
environment.
5.4 SUMMARY OF PERFORMANCE AND COSTS OF NOV CONTROLS
A
Performance and costs associated with wet and dry NO controls
A
for stationary gas turbines are summarized in this section. Details of
performance, cost and incremental impacts of specific control techniques
have been discussed earlier. There are certain aspects of both wet and
dry NOV controls where there is a serious lack of data necessary to form
A
even qualitative conclusions. Also, some aspects of this evaluation have
relied on data which shows considerable variation among users,
manufacturers and EPA. There does not seem to be much argument that wet
controls are now fully capable of reducing NO emissions to the standard
A
of 75 ppm at 15 percent 02- Also, manufacturers expect dry NO
controls (exclusive of catalytic combustion) to be fully developed and
operational in large scale turbines within 5 years. Assigning costs to
the various control options and evaluating their effect on operations and
maintenance is where the most inconsistency in estimates occur.
5.4.1 Wet Controls
Industry has long accepted water and steam injection as a suitable
method for NOX control. Depending on the water/fuel ratio, emission
reductions can be as great as 80 percent. Existing utility engines in the
60 MW range will be capable of reducing NOX emissions approximately
45 percent using a 0.5 water/fuel ratio. This is sufficient reduction to
lower NOX emissions below the proposed standard of 75 ppm at 15 percent
02- Recent studies have shown that wet controls can actually aggravate
5-54
-------
the NO problem when fuels containing bound nitrogen are used.
A
Depending on the water/fuel ratio and operating conditions, emissions of
unburned hydrocarbons and CO may increase. Water or steam may have a
localized detrimental effect on combustion efficiency within the primary
combustion zone, thereby inhibiting complete oxidation of hydrocarbons and
CO.
There is considerable disagreement about the impact that wet
controls have on the daily operations and maintenance of gas turbines. An
increase in engine heat rate, manifested as a maximum of 5 percent
increase in fuel usage, is the most significant impact on operations.
This may be offset somewhat by increased power output caused by the
increase in mass throughput. Periodic recharging of the water
purification system will most certainly be required. Indeed, a full time
operator/maintenance person may even be warranted for some installations.
Some users have reported significant maintenance problems with the water
treatment system itself and internal turbine problems due to water use.
The nature of the latter type generally fall into one of the following two
categories: hot parts embrittlement or particle deposition and
contamination. It appears these items are not only affected by water
quality, water/fuel ratio, and equipment types, but by day-to-day
operations and maintenance procedures. At least two utilities have
accumulated over 50,000 hours of wet NOX control experience and have
experienced no significant problems or outages directly attributable to
the control technique.
A rather wide range of user-reported costs for water injection
equipment on a per kilowatt basis has been found. A number of utilities
have cited costs ranging from approximately $5/kW to almost $23/kW in 1978
dollars. EPA originally used a value of $2.58/kW in 1975 dollars to
estimate the economic impact of NO controls. They have subsequently
n
revised their estimates to $7/kW inM978 dollars for a 60 MW gas turbine
(Reference 5-24). The analysis here has used $10/kW as a representative
value for large scale turbines. It appears that actual capital costs will
be very site specific and depend to a great extent on the required water
purification equipment and to a lesser extent on turbine modifications.
Additional operating and maintenance costs have been found to be
approximately 3 percent of installed costs. The fuel penalty due to an
5-55
-------
increased heat rate with water injection is a significant portion of this
additional expense.
The cost impact of wet NO controls on medium and small size gas
A
turbines is more severe than for large engines which can show savings
through economies of scale. Water injection for a 3 MW unit has been
found to cost $30/kW, considerably higher than EPA's estimate reported in
the SSEIS. Nonetheless, the total annualized cost to control using water
injection should only increase the cost of electricity produced by 2 to
5 percent.
5.4.2 Dry Controls
Dry NO controls refer to unconventional combustors which are so
/\
designed that NO formation is reduced by combustion modification rather
than by injection of water or steam. General concepts currently being
developed by manufacturers to meet the NOV NSPS involve modifications to
A
the combustor design, fuel premixing and prevaporization, and improved
airflow patterns. Catalytic combustion, although many years from full
development in gas turbines, has shown dramatic reductions in all
emissions approaching 98 percent in some cases. Catalytic technology has
received only minor treatment in this report since manufacturers speculate
that it has only a far term potential for gas turbines.
For detailed summaries of completed and some ongoing dry control
research and development programs, see Section 5.2.1.3 and 5.2.1.4. The
technical details of these programs will not be further summarized here.
However, the NO and incremental emissions will be summarized. Lean
A
burning in the primary zone and staged combustion in conjunction with some
form of fuel/air premixing and prevaporization appears to be among the
most promising dry control concepts for fuels containing minimal bound
nitrogen. Researchers have obtained NO reductions approaching 80
A
percent in laboratory combustor rigs. Other techniques such as rich
burning in the primary zone with quick quench in the secondary zone have
shown dramatic NO reductions with both clean and dirty fuels. Most of
A
these controls are being developed with the intent of minimizing emissions
of CO and UHC as well as NO . Manufacturers have had varying degrees of
success. While many designs face a NO /CO tradeoff, the EPA sponsored
A
Pratt and Whitney rich-burn, quick-quench combustor appears to have
isolated NOX control from CO control. The problem becomes particularly
5-56
-------
difficult for lean combustors during off-design conditions when combustion
is not very efficient. While formation of NOV is somewhat inhibited
/\
under these conditions, incomplete oxidation of CO and UHC is favored. CO
and NO emissions in lean combustors are very sensitive to changes in
A-
equivalence ratio. While a carefully implemented engine control system
could control equivalence ratio, other problems, such as component
deterioration, may tend to increase CO and UHC.
Since dry controls are essentially modified, although more complex,
conventional combustors, there probably will not be any additional impact
on operations and maintenance. Still, all of the problems manufacturers
are experiencing in the developmental stage must be solved before these
concepts are commercially employed on full scale engines. Moreover, new
problems to be solved will undoubtedly surface during the scale-up
process. Currently, it is not expected that dry controls will
significantly affect heat rate. Combustor liners may have to be replaced
more frequently than with conventional combustors. Other than periodic
replacement of the catalyst cell, it is not known what additional
maintenance procedures will be required with catalytic combustion
technology.
At the present stage of development, it is very difficult to
accurately estimate the economics associated with the implementation of
dry controls. For utility size turbines, manufacturers speculate that dry
controls will cost somewhat less than wet controls for a comparably sized
unit. The size of the base over which developmental costs can be
distributed seems to be a dominant factor in the actual cost. Maintenance
costs should not be great since dry control combustor liners can be
replaced on a periodic basis as are current combustor liners, although
possibly more frequently. The cost impact on smaller turbines is expected
to be more substantial since there is a smaller base to spread development
costs over.
In summary, wet controls are the preferred, indeed the only NO
A
control technique which are currently available to meet the recently
promulgated NO performance standard for stationary gas turbines.
/\
Although the capital costs of water injection can be significant,
approaching 20 percent of total unit cost on a per kW basis, total
5-57
-------
annualized control costs, including operating and maintenance, should
increase the cost of power by only 2 to 5 percent.
Dry controls, as they become fully developed for full size engines,
will begin to replace wet controls in new sources. They appear to be more
attractive from virtually all standpoints, the most important of which is
an emerging ability to control NOX from high-nitrogen fuels such as
residual oils and synthetic fuels. Catalytic combustion, while yielding
extremely high pollutant reductions in subscale models, may only be
feasible in the very long term.
5.5 PERFORMANCE AND COST FOR POLLUTANT CONTROLS OTHER THAN NOX
This section must be prefaced by a statement regarding the relative
importance of controlling the various pollutants produced by stationary
gas turbines. NO is the primary pollutant of concern with gas
turbines. Gas turbines also produce quantities of sulfur oxides, CO and
UHC. But for various reasons control of these pollutants is either
relatively easy, prohibitively expensive or obviated by certain operating
conditions of gas turbines.
Sulfur dioxide emissions from gas turbines are strictly a function
of the fuel sulfur content. Virtually all fuel sulfur is converted to
S02 in the combustion process. The easiest way to control S02
emissions is to regulate the sulfur content of the fuel. Gas turbines
have traditionally burned clean fuels low in sulfur so this method of
control should not be a great hardship. However, if trends in fuel use
move towards the use of heavier fuels with higher sulfur content, some
form of fuel desulfurization may be required. Economics and availability
of fuels will dictate when desulfurization becomes viable. Flue gas
desulfurization is at present extremely costly for gas turbines due to the
high volumes of gas that must be treated. Costs of flue gas desulfurization
are estimated to be from two to three times the cost of the gas turbine
itself, thus making it unattractive.
Typical operating characteristics of gas turbines and the
importance of controlling NO emissions considerably diminishes the
A
problem of CO and UHC emissions. Most stationary gas turbines operate at
full load and thus under conditions of high combustion efficiency. This
greatly reduces CO and UHC emissions. When load is decreased, combustion
efficiency decreases and emissions of unburned species increases. Control
5-58
-------
of these emissions is accomplished (secondarily to NO ) by combustion
A
modifications that enhance fuel/air mixing, fuel atomization, and control
of equivalence ratios, and increased residence times to promote better
combustion of UHC in the primary zone and CO in the dilution zone.
In summary, costs of S02 control are reflected directly in the
costs and sulfur content of the fuel. Performance and costs of CO and UHC
control follow control of N0x. That is to say, while it is desirable to
limit CO and UHC emissions, gas turbine pollution control research and
development efforts have as their primary goal, control of NO . CO and
A
UHC control are of secondary importance, and with the development of new
dry control combustors, manufacturers will control CO and UHC in order to
maximize combustion efficiency.
5-59
-------
REFERENCES FOR SECTION 5
5-1. Personal cormunication with T. Hensel, Turbo Power and Marine
Systems, Farmington, CT, August 10, 1978.
5-2. Mason, H. B., et aj_._, "Preliminary Environmental Assessment of
Combustion Modification Techniques: Volume II, Technical Results,"
EPA-600/7-77-119b, NTIS-PB 276 681/AS, October 1977.
5-3. Goodwin, D. R., et _a_L, "Standard Support and Environmental Impact
Statement, Volume 1: Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
5-4. Kesselring, J. P., et a!., "Design Criteria for Stationary Source
Catalytic Combustion SysTems," Acurex Final Report 78-278, EPA
Contract 68-02-2116, March 1978.
5-5. Personal conmunication with S. M. DeCorso, Westinghouse Electric
Corp., Philadelphia, PA, July 25, 1978.
5-6. "New Equipment Orders and Installations," Gas Turbine World,
Vol. 9, No. 3, pp. 33-38, July, 1979.
5-7. Knox, W. M., et al., "Water and Steam Injection for Emission
Control," General Electric Gas Turbine Reference Library
publication GER-2506, 1978.
5-8. Teixeira, D. P., "Overview of Water Injection for NOX Control,"
EPRI SR-39, February 1976.
5-9. Singh, P. P., e£ aj_., "Combustion Effects of Coal Liquid and Other
Synthetic Fuels in Gas Turbine Combustors - Part 1: Fuels Used and
Subscale Combustion Results," ASME 80-GT-67, December 1979.
5-10. Mosier, S. A., and R. M. Pierce, "Advanced Combustion Systems for
Stationary Gas Turbine Engines," Pratt and Whitney Aircraft Group,
presented at the Second Stationary Source Combustion Symposium New
Orleans, August 1977.
5-11. Personal communication with E. Zeltmann, General Electric,
Schenectady, NY, July 21, 1978.
5-12. Bahr, D. W., et^ a/K_, "Experimental Clean Combustor Program, Phase I
Final Report," NASACR-134737, June 1975.
5-13. Roberts, R., et a/L_, "Experimental Clean Combustor Program, Phase I
Final Report," NASACR-134736, October 1975.
5-60
-------
5-14. Roberts, P. B., et a\_._, "Advanced Low NOX Combustors for
Supersonic High Altitude Aircraft Gas Turbines," ASME Publication
76-GT-12, March 21-25, 1976.
5-15. Goldberg, P., et £K_, "Potential and Problems of Premixed
Ccmbustors for Application to Modern Aircraft Gas Turbine Engines,"
AIAA Paper No. 76-727, presented at the AIAA/SAE 12th Propulsion
Conference, Palo Alto, CA, July 1976.
5-16. Pierce, R. M. and S. A. Mosier, "Advanced Combustion Systems for
Stationary Gas Turbine Engines," Special Interim Report,
PWA FR-10846, Pratt & Whitney Aircraft Group, West Palm Beach, FL,
November 1978.
5-17. Particulate Polycyclic Organic Matter, National Academy of
Sciences, Washington, 1972.
5-18. Bittner, J. D., et al., "The Formation of Soot and Polycyclic
Aromatic Hydrocarbons in Combustion Systems," in Proceedings of the
Stationary Source Combustion Symposium, Volume I,
EPA-600/2-76-152a, NTIS-PB256 320/AS, June 1976.
5-19. Vapor Phase Organic Pollutants - Volatile Hydrocarbons and
Oxidation Products, National Academy of Sciences, Washington, 1976.
5-20. Personal communication with D. Blann, Acurex Corp., Research
Triangle Park, NC, August 1978.
5-21. Personal communication with W. B. Little, Houston Light and Power,
Houston, TX, August 3, 1978.
5-22. Shimizu, A. B., et_ £]_._, "NOX Combustion Control Methods and Costs
for Stationary Sources, Surmtary Study," EPA-600/2-75-046, NTIS-PB
246 750, September 1975.
5-23. Personal communication with R. Carlisle, City of Pasadena Power and
Water, Pasadena, CA, August 8, 1978.
5-24. Letter from W. R., Hall, City of Glendale Public Service
Department, Glendale, CA, August 11, 1978.
5-25. Personal communication with R. Jenkins, Office of Air Quality
Planning and Standards, EPA, Research Triangle Park, NC,
September 1978.
5-26. Personal communications with A. Dusi and B. Gilman, San Diego Gas
and Electric, San Diego, CA, September 1978.
5-27. "Gas Turbine Electric Plant Construction Cost and Annual Production
Expenses," F.P.C. S-240, 1972.
5-28. Personal communication with W. Price, New Mexico Electric, Hobbs,
NM, August 8, 1978.
5-61
-------
5-29. "Utility Cost Study - Part 1," Gas Turbine International, March -
April 1977.
5-30. Edwards, T. W., "Utility Experience in the Southwest,"
Turbomachinery International, March, 1978.
5-31. Personal communication with H. Dvorak, Turbo Power and Marine,
Farmington, CT, August 8, 1978.
5-32. Personal communication with R. Walker, General Electric,
San Francisco, CA, August 21, 1978.
5-33. "Task 7 Summary Report - Technical and Economic Bases for
Evaluation of Emission Reduction Technology," prepared by PEDCo in
support of OAQPS work on NSPS for industrial boilers, EPA Contract
68-02-2603, PEDCo Environmental, Inc., Cincinnati, OH, June 2, 1978.
5-34. Sawyer, 0. W., Editor, Sawyer's Gas Turbine Engineering Handbook,
Gas Turbine Publications, Inc., Stamford, CT, 1976.
5-35. Personal communication with D. Stenson, Solar, San Francisco, CA,
August 16, 1978.
5-36. Personal communication with G. P- Hanle, General Motors, Wamen, MI,
August 25, 1978.
5-37. Fisher, T. M., "General Motors Response to Stationary Gas Turbines
Standards of Performance for New Stationary Sources," submitted to
D. R. Goodwin, Emission Standards and Engineering Division, EPA,
Research Triangle Park, NC, January 31, 1978.
5-38. De Biasi, V. "FT50 Design Shortcut to 1980 Technology," Gas Turbine
World, November 1975.
5-39 Personal communication with S. Lanier, U.S. Environmental
Protection Agency, Research Triangle Park, NC, November 1978.
5-62
-------
SECTION 6
ENVIRONMENTAL ASSESSMENT
Unlike most stationary sources which have potential multimedia
pollution control problems, the impact of stationary gas turbines is
almost exclusively on ambient air. Solid or liquid waste impacts are
minimal. Although gas turbines represented only a 2 percent share of the
total anthropogenic NO emissions for the year 1977, this does amount to
/\
210,000 Mg (232,000 tons) (Reference 6-1). Due to the nature of most gas
turbine installations (i.e., clustered in groups of multiple units)
certain circumstances of operations, ambient background pollutant levels
and meteorological conditions can cause ambient standards to be exceeded
locally (Reference 6-2). Consequently the main focus of this section will
be on assessing the environmental impact of pollutant controls for gas
turbines on ambient air levels.
The environmental assessment discussion begins in Section 6.1 with
a quantitative multimedia impact analysis of a typical gas turbine,
comparing uncontrolled and controlled operation. Then the discussion
focuses in Section 6.2 on the environmental impact of gas turbines on air,
since the principle effluent is in the form of flue gas emissions.
Planned and existing emissions standards on the Federal, state, and local
levels are reviewed and compared with emission levels from uncontrolled
and controlled firing. In addition, potentially hazardous, but as yet
unregulated pollutants (including those with no proposed regulations), are
discussed. These include sulfates and polycyclic organic matter (POM).
The effect of NOY controls on incremental emissions of other pollutants
A
has been discussed at great length in Section 5.2.2. It was found that
these incremental emissions can vary greatly depending on the type of
control, the degree of care in which the control was implemented
(particularly with dry controls), the fuel type, the load, and other
operating conditions.
6-1
-------
Although primary focus is on air emissions, gas turbines do cause
additional environmental impacts. However, in comparison with the
relative magnitude of the air impacts, other multimedia "wastes" are
easily controlled and relatively innocuous. These include noise impacts,
solid wastes such as water treatment system residue and liquid wastes such
as waste water from the water treatment system. Evaluation of these
impacts is presented in Section 6.3.
Evaluating the effect that emission controls have on gas turbine
operation and maintenance is an integral part of assessing the performance
of those controls. This is discussed in Section 6.4. N0x controls can
affect operating parameters as well as maintenance requirements. Although
maintenance requirements are highly variable among different users, wet
controls appear to have a greater impact than do dry controls. For
example, some users and manufacturers report a 2 to 5 percent heat rate
increase with wet controls, while dry controls are not expected to affect
heat rate. Data for this evaluation have been collected primarily from
users and are, for the most part, only qualitative in nature. The
limitation is the lack of long term user experience.
In assessing the economic impact of NO controls, this assessment
A
took into account such factors as the economic environment of the
industry, incremental capital and operating costs of controls, and cost in
terms of efficiency losses, increased fuel use requirements, fuel
availability and net effects on operations and maintenance requirements
and schedules. Section 6.5 reviews the cost analysis results presented in
Section 5.2.3. Through a formalized costing procedure based on user
supplied data, through figures reported in the literature and supplied by
manufacturers and, in some case (for dry controls), through
semi-quantitative estimates, a ranking of controls based on cost
effectiveness is derived (Sections 6.5 and 6.6). The present stage of
control technology development was also an important consideration.
6.1 ENVIRONMENTAL IMPACT ANALYSIS
If the environmental assessment of stationary gas turbines is to be
compared on a consistent basis with other environmental assessments,
standardized methodologies must be employed. Using Level 1 analyses from
an Acurex test (sponsored by EPA under this NO EA program) on a 60 MW
A
oil-fired utility gas turbine, potential problem pollutants and pollutant
6-2
-------
control priorities have been determined. Through the use of an IERL
Source Analysis Model, specifically SAM IA, problem areas and priorities
can be identified by establishing maximum discharge severities and
weighted discharge severities (References 6-3 through 6-6).
The SAM IA model defines two indices of potential hazard. The
first, termed Discharge Severity (DS), is the ratio of a pollutant species
discharge concentration to that species' Discharge Multimedia Environmental
Goal (DMEG). The DMEG values represent maximum concentrations acceptable
in effluent streams and are defined to preclude adverse effects from acute
exposure. The DMEG's are generally defined to be pollutant levels safe for
occupational exposure.
The second SAM IA hazard index, termed Weighted Discharge Severity
(WDS), is defined to be the product of the DS with the effluent stream
mass flowrate.
A summary of the data from the 60 MW oil-fired utility gas turbine
tested is presented in Table 6-1. The concentration in the flue gas of
the species from each MEG category analyzed for the baseline and water
injection case, are presented. Table 6-2 presents the compounds for which
the DS exceeded unity for either of the two tests. The emissions of
C02, N02, S02, and H2S04 exceeded the DMEG values while those of
As, Cr, and Cl may have also. However, for these compounds there is a net
potential environmental benefit (i.e. decrease in emissions) when using
water injection. The organic compounds terphenyl, naphthalene, pyrene,
and fluoranthrene were found during the water injection test but not
during the baseline test. However, their DS values were less than 1.
6.2 ENVIRONMENTAL IMPACTS ON AIR
This section analyzes the ambient air impact of emissions from gas
turbines, with the primary emphasis on NOX. The discussion begins with
a review of existing or planned NO control standards and then compares
A
waste stream emissions, both controlled and uncontrolled, with these
standards.
6.2.1 Summary of NO,, Control Regulations
^•—•—^•••"^^"^•^•^.^•^•^••^—^—^^^^^•jfc™™^*™™.^™™^™^™™
The Federal Standards of Performance for New Stationary Sources
(NSPS) provide the limitations for emission of NO pollutants from
J\
stationary sources on a national basis. The NSPS are intended largely to
6-3
-------
TABLE 6-1. EMISSIONS FROM A 60MW UTILITY OIL-FIRED GAS TURBINE
MEG
Category
01A
05C
08A
08D
ISA
15B
ISA
21A
21A
21B
22A
32
36
37
41
42
45
46
47
49
50
51
53
54
55
62
65
68
69
71
72
74
76
78
81
82
83
Compound
Assumed
Methane
t-Pentanol
Maleic acid
Phthalate ester
Biphenyl
Terphenyl
Phenol
Phenanthrene
Naphthalene
Pyrene
Fluoranthrene
Be
Ba
B
Tl
CO
C02
Sn
Pb
N02
As
Sb
Bl
SO 2
H2S04
Se
Te
T1
V
Cr
Mo
Mn
Fe
Cc
Ni
Cu
Zn
Cd
Hg
Concentration (yug/m3)
Baseline Water Injection
1.6 x 103
620
620
2.0
60
—
1.0
0.5
—
—
—
<0.92
<3.5
<2.2 x 102
<11
7.0 x 103
8.0 x 107
<16
82
3.5 x 105
<14
<4.6
<1.8
3.1 x 104
8.1 x 103
<11
<3.6
<29
<22
<17
<5.8
<0.48
71
<0.55
<0.24
42
<760
<13
<2.8
2.4 x 103
520
520
2.0
60
5.0
1.0
1.0
1.0
0.5
0.5
<0.14
<3.6
<2.0 x 103
<10 .
1.0 x 104
8.4 x 107
<44.1
23
1.5 x 105
<14
<4.7
<2.3
3.4 x 104
6.0 x 103
<10
<3.4
<33
<49
<7.5
<4.8
<0.05
89
<0.13
<0.61
60
800
<0.55
<21
Total Flue Gas Flowrate:
Basel i
Water
ne 250 kg/s
injection 242 kg/s
6-4
-------
TABLE 6-2. EMISSIONS WITH CONCENTRATIONS GREATER THAN DMEG VALUES
MEG
Category
42
47
49
53
68
82
Compound
Assumed
C02
N02
As
S02
H2S04
Cr
Cl
Discharge
Baseline
8.89
38.9
<7
2.38
8.1
<17
<1.3
Severity (DS)
Water Injection
9.33
16.7
<7
2.62
6
<7.5
<0.055
assist in air quality maintenance by offsetting increases due to source
growth through the application of control technology. The development and
demonstration of N0x control technology allows EPA to prepare for the
setting of future NSPS based on the best system for emission reductions.
The NSPS NOX limit for stationary gas turbines with heat input more than
2.2 MW (7.5 MBtu/hr) is 75 ppm at 15 percent 0,,. Specific corrections
are allowed for efficiency and fuel nitrogen.
The primary responsibility for air quality attainment and
maintenance, however, rests with the individual states and the emission
standards established under the State Implementation Plans (SIP), not at
the Federal level. The state and local standards for new and existing
stationary fuel combustion equipment, including gas turbines, are
presented in Table 6-3. As in the case for the NSPS, the basis for these
standards is by application of combustion process modification as
demonstrated through retrofit technology. With the exception of the South
Coast Air Quality Management District (SCAQMD), the standards are largely
directed at future air quality maintenance rather than attainment. Some
areas have exercised the option to set more stringent standards than
required for maintenance by the SIP. The SCAQMD has a serious attainment
problem and has accordingly instituted the most comprehensive and
stringent emission regulations. In fact, the control development in the
SCAQMD has been so intense that it is useful as a guide to the limits of
current technology for existing gas- and oil-fired equipment. The trend
in the SCAQMD and elsewhere has been toward regulating smaller equipment
6-5
-------
TABLE 6-3. SUMMARY OF STATE AND LOCAL NOX EMISSIONS STANDARDS FOR
STATIONARY SOURCESe (Reference 6-1)
CALIFORNIA*1
Bay Area QM
Monterey Bay
San Diego
San Joaquin
Kern Co.
Santa Barbara APCD
Counties In SCAQMD:
LA Co.c
Orange Co.c
Equipment*'
New or
Existing
New
All
New
All
All
All
New
New
New
New
All
All
All
New
Existing
Existing
Existing
Type
Heat transfer
Heat transfer
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Equipment
Equipment
Equipment
Equipment
Equipment
Equipment
Heat Input
Capacity3
73 (250)
513 (1750)
440 (1,500)
440 (1.500)
15 (50)
520 (1,775)
520 (1,775)
520 (1,775)
520 (1,775)
73 (250)
147-630
500-2,150)
630 (2,150)
Standardise
Gas
Fired
125 ppm
175 ppm
64 kg/hr9
125 ppm
225 ppm
125 ppm
64 kg/hr9
125 ppm
64 kg/hrg
64 kg/hr9
125 ppm
225 ppm
125 ppm
125 ppm
225 ppm
225 ppm
125 ppm
Oil
Fired
225 ppm
300 ppm
64 kg/hr9
225 ppm/hr
225 ppm
225 ppm
64 kg/hr9
225 ppm
64 kg/hr9
64 kg/hr9
225 ppm
325 ppm
225 ppm
225 ppm
325 ppm
325 ppm
225 ppm
Coal
Fired
—
-
64 kg/hr9
—
--
225 ppm
64 kg/hr9
225 ppm
64 kg/hr9
64 kg/hr9
-
—
--
--
—
-
-
Effective
Date
4/19/75
9/16/76
1/1/75
1/1/75
12/31/71
12/31/74
12/31/72
12/31/72
12/31/75
Comments'^
cn
cr>
-------
TABLE 6-3. Continued
• San Bernardino Co.c
Riverside Co.c
SCAQMD
Ventura Co. APCD
CONNECTICUT
Equipment1*
New or
Existing
Existing
Existing
Existing
Existing
New
All
All
All
New
New
All
All
Existing
New
Existing
All
Type
Equipment
Equipment
Equipment
Equipment
Equipment
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
Fuel burning
GT
Heat Input
Capacity3
520 (1,775)
205 (700)
205 (700)
520 (1,775)
523-628
(1,786-2,143)
628 (2,143)
161-523
(550-1,786)
73 (250)
73-630
(250-2,150)
630 (2,150)
73 (250)
1.5-73
(5-250)
1.5-73
(5-250)
Standardb.e
Gas
Fired
125 ppm
225 ppm
125 ppm
125 ppm
64 kg/hr9
125 ppm
125 ppm
300 ppm
64 kg/hrQ
125 ppm
250 ppm
125 ppm
86.1 (0.2)
86.1 (0.2)
86.1 (0.2)
387.3 (0.9)
Oil
Fired
225 ppm
325 ppm
225 ppm
225 ppm
64 kg/hr9
225 ppm
225 ppm
400 ppm
64 kg/hr3
225 ppm
250 ppm
225 ppm
129 (0.3)
129 (0.3)
129 (0.3)
387.3 (0.9)
Coal
Fired
—
64 kg/hr9
325 ppm
225 ppm
400 ppm
64 kg/hr9
—
250 ppm
—
387.3 (0.9)
301 (0.7)
387.3 (0.9)
387.3 (0.9)
Effective
Date
1/1/75
12/31/71
12/31/74
1/1/75
1/1/77
Comments'!
Applies to West
Central Area Only
Applies to West
Central Area Only
Applies to
Balance of County
Except peaking
units at Hand a lay
Except GT, 1C
Except GT, 1C
Except GT, 1C;
variances
permitted
CTl
-------
TABLE 6-3. Continued
cr>
i
00
ILLINOIS
Lake, Mill,
Du Page, HcHenry,
Kara, Grundy,
Kendall, Kankakee,
Macon, St. Clair,
Madison Cos.
Cook Co.
INDIANA
MARYLAND
NEW MEXICO
NEM YORK
Equipment''
New or
Existing
Existing
lew
Existing
Existing
New
ew
xisting
11
11
Type
Fuel combustion
Fuel combustion
Fuel combustion
Fuel burning
Fuel burning
Gas burning
Gas burning
Oil burning
GT, 1C
Heat Input
Capacity9
73 (250)
59 (200)
59 (200)
73 (250)
73 (250)
1,055,000 GJ/
yr (1,000,000
MBtu/yr)
1,055,000 GJ/
yr (1,000,000
MBtu/yr)
1,055,000 GJ/
yr (1,000,000
MBtu/yr)
73 (250)
Standard13. e
Gas
Fired
129 (0.3)
86.1 (0.2)
129 (0.3)
86.1 (0.2)
86.1 (0.2)
86.1 (0.2)
129 (0.3)
86.1 (0.2)
Oil Coal
Fired Fired
129 (0.3)
129 (0.3)
129 (0.3)
129 (0.3)
129 (0.3)
129 (0.3)
129 (0.3)
387.3 (0.9)
301 (0.7)
387.3 (0.9)
301 (0.7)
215.2 (0.5)
301 (0.7)
Effective
Date
12/31/74
12/31/74
Comments''
Except cyclone &
horizontally
opposed fired
boilers burning
solid fuel
Applies to
"Priority Basin A"
only — none at
present
Applies to
"Priority I"
AQCR ' s
only
-------
TABLE 6-3. Concluded
CTl
10
OKLAHOMA
SOUTH DAKOTA
VERMONT
WYOMING
Equipment^
New or
Existing
New
All
NGW
New
New
Existing
New
New
Existing
Existing
Type
Fuel burning
Fuel burning
Combustion
GT
Gas burning
Gas burning
Oil burning
Oil burning
011 burning
011 burning
Heat Input
Capacity*
15 (50)
73 (250)
73 (250)
0.29 (1)
0.29 (1)
73 (250)
73 (250)
Standardb« e
Gas
Fired
86.1 (0.2)
86.1 (0.2)
—
—
86.1 (0.2)
99 (0.23)
—
—
—
—
011
Fired
129 (0.3)
129 (0.3)
129 (0.3)
129 (0.3)
—
129 (0.3)
258.2 (0.6)
197.9 (0.45)
258.2 (0.6)
Coal
Fired
301 (0.7)
-
—
—
~
~
—
Effective
Date
7/1/71
7/1/71
Comments'*
Except GT
Except 1C 59
(200)
Except 1C 59
(200)
Except 1C 59
(200)
Except 1C 59
200)
Except 1C 59
(200)
^Unless stated otherwise, units are MW (U)6 Btu/hr)
bUnless stated otherwise, units are ng/J (lb/106 Btu)
cRu1es put Into effect before SCAQHD was formed and replaced by SCAQMD rules
dGT refers to gas turbines; 1C refers to reciprocating internal combustion engines
eNOv emission standards In chronological order in so far as possible, standards which are similar to federal standards have been omitted
fAll ppm standards are at 3 percent 02
9140 Ibs/hr
-------
categories and tightening the regulations on larger equipment. The SCAQMD
regulations are currently being evaluated as part of the SIP revision to
determine if further emission reductions are practical.
Maintenance of air quality in the 1980's and 1990's will require
NO regulations in addition to those existing or planned. New source
A
controls will be emphasized since experience has shown them to be more
effective, less costly and less disruptive than retrofit control of
existing equipment.
6.2.2 Comparison of Emissions to Standards
The NOV reduction capability of wet N0¥ control techniques was
X "
presented in Figure 5-2. Table 6-4 compares the potential levels of NO
/\
reduction with wet controls to the Federal standard of 75 ppm at
15 percent 0? and to the more stringent regional standards.
In principle (although not always in practice), water or steam
injection is fully capable of controlling NO emissions to proposed and
/\
existing standards; it is simply a matter of the correct water/fuel
ratio. For example, a distillate oil-fired simple cycle gas turbine with
baseline (uncontrolled) NO emissions of 170 ppm (all emissions at
X
TABLE 6-4. DEGREE OF WET CONTROL REQUIRED TO MEET REGIONAL NOX STANDARDS
Fuel
Oil
Gas
Baseline
Emissions —
ppm @ 15% 02
170 ppm
110 ppm
Existing NOX Standard
— ppm @ 15% 02 and
Equipment Category3
75 ppm for new and
existing sources
42 ppm for new and
existing sources
Required Water/Fuel
Ratio to Meet
Standard
0.5
0.6
aRegulations from counties in the Southern California Air Pollution
Control District.
6-10
-------
15 percent CL) can meet the NSPS of 75 ppm with a 0.5 water to fuel
ratio. The same turbine firing gas would emit 110 ppm uncontrolled. In
this case, a water/fuel ratio of 0.6 would be sufficient to meet the
Southern California standard of 42 ppm.
Dry NO controls are being developed as a replacement for wet
A
controls, which many users claim are costly, complicated, and difficult to
maintain. Of course, they are being designed to meet the NSPS and for
some fuels, such as natural gas, are even expected to meet the most
stringent regional regulations. As an example, Pratt and Whitney's Low
N0¥ Combustor Program has met their program goal of 75 ppm NOV. In
« X
fact, even with the firing of fuel containing 0.5 percent nitrogen, NO
/\
emissions were as low as 50 ppm (Reference 6-7).
It appears then that while wet controls are fully capable of
meeting all NO regulations by varying the water/fuel ratio, dry
A
controls will probably evolve into the preferred technique once they are
commercially demonstrated to meet existing and proposed emission
regulations.
6.2.3 Ambient Air Impact
The ambient air impacts of S09, CO and NO emissions from gas
L. A
turbines have been analyzed by EPA (Reference 6-2) using the basic
Gaussian plume model with estimations and assumptions appropriate to the
special plume characteristics of stationary gas turbines. The modeling
studies have accounted for variations in fuels, degree of control,
equipment configurations (i.e., individual versus cluster arrangements),
and meteorological conditions. The concentration and ambient air impact
of emissions can be estimated over distinct averaging times through the
use of these dispersion models. The estimates can then be compared to the
National Ambient Air Quality Standards (NAAQS).
The environmental impact of nine gas turbine units was analyzed for
individual and clustered arrangements (Reference 6-2). CO emission
estimates exceeded the NAAQS in only one instance, with eight clustered
units operating in the spinning reserve mode. This mode of operating
represents idle (no load) conditions where CO emissions can approach
100 times the emission rate at design (full) load (Reference 6-2).
6-11
-------
The potential for SCL emission rates to exceed the short term
NAAQS occurred often. However, the SC^ average was found not to exceed
air quality standards on an annual basis. Obviously if point sources are
controlled to meet short term standards, the annual standard would never
be exceeded.
Annual average NO concentration estimates for individual gas
/\
turbines did not exceed the NAAQS. However, the short term average for
multiple (16-clustered) turbine units was found to be approximately five
times the impact of an individual unit. If the same factor is applied to
annual average estimates, uncontrolled units could potentially exceed the
NAAQS for N02<
Gas turbine units are generally characterized by short, rectangular
stacks. Results reported in Reference 6-2 show that higher concentrations
can occur in regions with stronger, steadier winds. Since the effluent
stacks from gas turbines typically reach the heights of nearby buildings,
the wind turbulence in the vicinity of the buildings can cause high ground
level concentrations. As might be expected, it was found that taller
stacks will result in reduced ambient concentrations.
Additional studies have been conducted which predict ambient air
impacts of NOY controlled gas turbines through the use of dispersion
/\
models (Reference 6-8). These studies have also included individual and
cluster arrangements as well as other variables. The results consistently
show that the annual ground level NO concentrations from the given gas
turbine units are in most cases, three orders of magnitude less than the
NAAQS for N0? of 100 g/m3 -- with the incremental contribution to
"- o
uncontrolled ambient concentrations ranging from 0.07 to 1.2 g/m .
Although small, this increase can have a significant impact on baseline
ground level concentrations for nonattainment regions, where ambient N02
concentrations already exceed the NAAQS.
6.2.4 Bioassay Results
The organic extracts from the XAD-2 resin portion of the SASS train
test on the 50 MW oil-fired gas turbine tested as part of this NO EA
/\
program (Reference 6-9) was subjected to bioassays. The Ames Salmonella
Microsome plate test did not show any mutagenicity. While the WI-38 human
cell cytotoxicity assay only showed low toxicity. These tests were
performed on samples collected during the water injection test. Thus, it
6-12
-------
appears that the exhaust from this gas turbine operating under water
injection has only low toxicity.
6.3 ADDITIONAL IMPACTS
A stationary gas turbine can represent a significant point source
of noise pollution. However, muffler systems are available that reduce
noise levels to 85 db, a safe level. The application of wet or dry NO
/\
controls is not expected to significantly alter these noise levels.
There are no radiation impacts associated with stationary gas
turbines.
6.4 ASSESSMENT OF IMPACTS OF NOX CONTROLS ON OPERATIONS AND
MAINTENANCE
A pollutant control technique is not considered acceptable if it
causes excessive' technical problems and is economically unfavorable.
Therefore as part of the assessment of NO control techniques for gas
A
turbines, this section appraises the potential operations and maintenance
effects as well as those reported by users who are already employing some
form of NO control. This evaluation, in conjunction with the analysis
A
of control technique performance, effect on incremental emissions, and
cost (Sections 5.2.1, 5.2.2, and 5.2.3 respectively), will provide a
realistic and comprehensive evaluation of gas turbine NO control
X
techniques.
Initially it was hoped that sufficient data would be available to
compare process variables under baseline or normal operating conditions to
those under controlled or low NO modes. However, unlike some
/\
stationary sources such as utility boilers where there are a large number
of parameters which must be continuously monitored, stationary gas
turbines do not have this requirement. Indeed, many can and do run
unattended for long periods. Furthermore, there is a large variation as
to which operating parameters are measured between users. Another added
hinderance is that some important variables, turbine inlet temperature for
example, are proprietary with most manufacturers. While there are
numerous large gas turbines currently with water injection capabilities,
most are not in use since few regions of the country currently regulate
N0x emissions. The equipment is usually purchased as a hedge against
future regulations. In light of this, much of the information for the
analysis in this section has been taken from user experience in the case
6-13
-------
of wet controls and from experimental test results in the case of dry
controls.
Significant operational and maintenance problems have been
experienced by many users, and these are discussed in detail. Some
companies have accumulated long hours of operation (in excess of
90,000 hours) using water injection and obtained favorable maintenance
records. Since for all intents and purposes dry NO controls have not
/\
been implemented, actual long term user experience and data do not exist.
Potential problems with dry controls noted here are based on research and
development experience. Of course most, if not all, of these problems
must be solved before dry controls become commercially viable. The
following discussion points out the types of operational and maintenance
problems inherent with new and complex combustor technology.
6.4.1 Wet Controls for NOW Reduction
.^.^-»i^fc«^.^^^^^^^^B P^^"—^ M^_^M_^H~
Wet controls for NO reduction in gas turbines refer to injecting
A
steam or more commonly, water, directly into the combustion chamber.
Since some of the thermal energy is used to heat the water, the peak flame
temperature decreases, resulting in lower NO emissions. One negative
/\
aspect of having to heat additional material in the combustor is decreased
overall system efficiency manifested as increased fuel consumption.
Reported values have ranged from zero to five percent increase in heat
rate with two percent as the average. The economic effect on overall gas
turbine ownership due to increased heat rate is discussed in Section 5.2.3.
Almost all the other reported operational and maintenance impacts due to
water injection can be classified as particle deposition and contamination
problems or difficulties related to the water treatment system. There are
some positive consequences of using water injection, aside from
substantially reducing NOX. While thermal efficiency can decrease,
maximum power output is simultaneously increased from two to three percent
depending on the water/fuel weight ratio. In fact, water injection has
been used for load augmentation in gas turbines since the early 1960's
(Reference 6-2). One major manufacturer commented that water injection
may also help to cool combustor walls, thereby prolonging service life.
This is speculation, however, and is not supported or refuted by actual
user data.
6-14
-------
Gas turbine manufacturers typically specify a total contaminant
level allowed in the combustor, based on impurities in combustion air,
fuel and water. The degree of water purification then depends on
contaminants in the other fluids in addition to the make-up water
quality. Some users have reported that water injection has decreased
component life by as much as 15 percent (Reference 6-10), although they
are not certain whether the increased maintenance is due to the water
injection or to the increasing age of the units. These same units,
General Electric Frame 5's burning natural gas and distillate oil, have
experienced embrittlement of hot combustion parts with water but not with
steam injection. The Public Service Department of the City of Glendale,
California has considerable experience with wet NO controls. One unit,
/\
a 31 MW simple cycle machine, has been in use since 1973 with a five
percent capacity factor. Thus far, no additional operational and
maintenance problems have been noted due to the water injection system. A
second unit is a combined cycle plant utilizing waste heat recovery
boilers to repower two old 20 MW steam turbines. The unit generates a
total of 120 MW, 30 MW of which is due to boilers, and is used for base
load power. Steam injection is used for NO control. A requirement for
A
more frequent cleaning of fuel nozzles is the only additional maintenance
item directly attributed to steam injection. One manufacturer has
reported coking in fuel nozzles with the use of water injection in a
particular installation (Reference 6-11).
It appears that a considerable number of turbine installations have
had particle deposition and other problems related to contamination due to
wet NO controls. There are many users, however, that report no
A
increased maintenance problems associated with water or steam injection.
A Southwest utility reports 2 year's, use of water injection on a
Westinghouse 66 MW turbine and no problems with the water injection system
or turbine problems caused by the wet NOY controls. In fact, no
X
significant deposits have yet been found on the turbine components
(Reference 6-12).
One utility used water injection (for NO emission control) on
/\
two combined cycle plants for a total of 54,180 hours, without making any
major changes to normal maintenance and operating procedures. They
followed procedures essentially identical to those required for a similar
6-15
-------
machine not using water injection and the plant experienced no outages
attributable to the water injection system during these operating
periods. Another company accumulated approximately 92,000 hours of
operating time with water injection on 17 turbines (1972 to 1978) with
only about 116 hours of outage due to their water injection system.
Maintenance records seem to indicate that, if run properly, water
injection does not substantially increase maintenance requirements. The
experience of Westinghouse in Japan seems to support this. They have four
501AA machines which use water injection continuously for N0x control
and Westinghouse indicates that they have received no negative reports
regarding the effect of water injection on operations and maintenance
(Reference 6-13).
Of the users contacted, approximately half were experiencing some
sort of problem with the water purification system. It appears that most,
if not all, of these problems can be obviated by good preliminary design
of the treatment system, and through careful installation and maintenance
procedures. In other words, water treatment systems employ conventional
water purification technology and there is nothing inherent in them that
would cause excessive problems which could not be avoided by the
appropriate design and operation.
6.4.2 Dry Controls for NO^ Reduction
"™"•^""••^••^•^^™y\i'""""™""-g^^^™^^^™^^™
Dry controls for NO emissions are classified as those not
J\
involving water or steam injection but some form of combustor modification
that favors conditions for limiting NO production. As previously
A
mentioned, dry NO controls have not generally been developed and
A
implemented to the degree where they can reduce NO emissions to 75 ppm
A
at 15 percent 0^. Most manufacturers agree that at least 3 to 5 years
is required before dry controls will be able to meet this emission goal
and become commercially viable. Thus, essentially no user experience on
the operational and maintenance effects of using dry controls is
available. This section simply points out the types of problems that
researchers have been experiencing during the dry control R&D programs.
Again, it must be kept in mind that most of these problems will be solved
before dry controls are implemented. This section discusses the generic
types of problems that future users potentially may experience if dry
6-16
-------
controls are not adequately developed or if they are not properly operated
and maintained.
It appears that most problems with dry control techniques are
contained in five general categories:
• Combustor exit gas temperature profile
• Combustor structural integrity
• Premixed combustor flashback and autoignition
t Combustor airflow distribution and control
• Carbon deposits.
Pratt and Whitney and Solar have done experimental combustor rig
work with premixed combustors (References 6-14 and 6-15). Both programs
were discussed in detail in Section 5.2.1.3. The Pratt and Whitney
premixed combustor for aircraft turbines experienced substantial problems
in virtually all of the above categories . Combustor exit gas
temperatures were found to be excessively high and characterized by a
distorted radial temperature profile. Certain combustor configurations
suffered burning, buckling and cracking as well as carbon deposits.
Solar's lean reaction premix type combustor experienced flashback within
the premix chambers under certain loads. This made it difficult to obtain
consistent levels of NO emissions.
n
United Aircraft studied a prevaporized premixed combustor concept
which employed an external heat exchanger to vaporize the fuel prior to
injection into the combustor. Significant problems were encountered with
coking and pyrolysis product deposition on heat exchanger tubes. It is
imperative that tubes be kept clean for long periods to maximize the heat
exchanger effectiveness which ultimately affects NOX emissions.
Another Pratt and Whitney program studied the concepts of vortex
mixing and burning and prevaporization-premixing. This program suffered
most of the generic problems mentioned earlier including:
• Airflow distribution problems causing high pressure losses
• Poor combustor exit temperature profiles
• Coking in heat exchangers
• Combustor durability
• Carbon deposits on combustor liners and in fuel injectors.
The researchers report that most of these problems can be eliminated
through proper design aimed at improving airflows and distribution.
6-17
-------
In summary, it appears that most difficulties related to dry NOX
controls, while a number of years away from being solved, will be resolved
before this control technology becomes accepted. However, these
combustors, in some cases, especially those requiring variable geometry,
appear to be more complicated, with an increased number of parts and more
complex geometry than today's conventional combustors. This may
complicate operations and increase the chance of failure. Combustor
liners may have to be replaced more often (Reference 6-16). On the
positive side, it appears that thermal efficiency will not be affected by
conversion to dry controls (Reference 6-16).
6.5 ECONOMIC IMPACT OF NOV CONTROLS
A
Wet controls will increase the price of gas turbines and increase
the fuel usage per kilowatt generated. Dry controls may increase gas
turbine prices but are not expected to increase fuel usage. Dry controls
are not presently available but should be by 1983. For utility size
turbines, wet controls are now being used but they are not being used on
the smaller size turbines. All of the above were discussed in
Section 5.2.3 along with estimated costs for typical turbines. This
section covers the total impact on the industry and consumer. The
economic impact will only be discussed through 1983 and mainly wet
controls will be discussed. Wet controls will have the largest effect
since dry controls, for all intents and purposes, will not be available in
the near term.
Table 6-5 summarizes projected installed costs and annualized cost
for various control techniques and NO emission levels. The two main
/\
points that these data indicate are: (1) that all costs increase as the
degree of control is increased and (2) dry control costs are expected to
be less than costs of wet controls. Considerd a far term technology,
costs of catalytic combustion are very difficult to estimate.
6.5.1 Impact on Manufacturers
The impact on the manufacturer comes from how these controls will
affect his sales. Can he pass the increased costs to the consumer and
still remain competitive? Also, will the N0x controls change the
operating characteristics so that other types of power generators will be
bought instead? Will one manufacturer have an advantage over another?
6-18
-------
TABLE 6-5. PROJECTED CONTROL REQUIREMENTS FOR ALTERNATE
NOX EMISSION LEVELS
NOX Emission
Level,
ppm @ 15 % 02
130
75
40
75
5
Control Technique
Water Injection
Water Injection
Water Injection
Dry controls (1983)
Catalytic combustion
(1985-1990)
Install Cost
$/kW
8
10
12
NA
NA
Annual i zed Cost
mills/kWh
3.5
4.0
4.7
NA
NA
Notes:
Peaking utility turbine, baseline NOX ~ 175 ppm @ 15% 02
Based on Table 5-11 assumptions and data
Fuel penalty and capacity enhancement assumed linear with
water/fuel ratio
Dry control costs expected to be somewhat less than wet controls
For utility turbines, all manufacturers will be adding the same
type of wet controls so no one should have an advantage over another. For
dry controls, if one manufacturer were to develop a dry control that was
easier to use and much lower in capital cost, he might have an advantage,
but as of now, this has not happened. For peaking units, gas turbines
have a large advantage over boilers because of their quick startup time.
NO controls should not change this. For wet controls, an isolated unit
A
might require a little more attention because of the water purification
system but at least one user did not find this to be a problem
(Reference 6-17). For base load operation, new turbines are becoming more
efficient and because of their lower capital cost, they may eventually
become competitive with boilers (Reference 6-18). Wet controls do
increase costs and lower efficiency but this should not change the
advantages turbines have as peaking units. Since dry controls are not
6-19
-------
expected to lower the efficiency nor are they expected to raise the cost
by a significant amount, they too should not diminish the advantages of
gas turbines as peaking units.
For the smaller size gas turbines, there is stiff competition from
diesel engines. Gas turbines are cheaper to manufacture but use more
fuel. Wet controls will increase the price of turbines and lower their
fuel efficiency, therefore if diesel engines are unregulated, gas turbines
will lose ground. However, in all probability there will also be
regulations promulgated for diesel engines. Also, even with wet or dry
NOY controls, gas turbines will remain cheaper to produce than diesels.
J\
Dry controls will have less impact than wet controls because they should
not increase the fuel usage. Also with dry controls, the turbines should
be easier to operate, which would lessen the impact on isolated operation
and minimize operating costs.
6.5.2 Impact on Consumer
The largest consumer, in terms of capacity, of stationary gas
turbines is the electric utility industry. Wet NO controls will increase
A
their cost of producing electricity through higher capital and operating
costs. In predicting the effect of wet controls, it was assumed that only
new installations will be affected and all new installations will be using
wet controls. Once dry controls become proven and available, wet controls
will be used less often. As a means of estimating the number of new gas
turbine installations, we have used estimates provided by the National
Electrical Manufacturers Association (Reference 6-19). Since smaller size
turbines (less than 2.2 MW heat input) will probably not require NO
A
controls for 5 years, they are not considered in this analysis.
Table 6-6 summarizes the predicted NO economic impact of wet
A
NOX controls to meet proposed NSPS for utility gas turbines from 1978 to
1983. To obtain a reduction of 175 ppm NO at 15 percent 0? to
75 ppm, that would be a reduction of 45.8 x 106 kg of NO in 1983 at a
A
cost of 0.8 mills/g NO for gas turbines added since 1977. The
A
electricity produced from these turbines would cost two to three percent
more because of these wet controls. Since less than 2 percent of the net
power generation of utilities is by gas turbines, these increased costs
are minor (Reference 6-2).
6-20
-------
TABLE 6-6. PREDICTED ECONOMIC IMPACT OF WET NOX CONTROLS TO MEET PROPOSED
NSPS FOR UTILITY GAS TURBINES: 1978-19833
Type cycle
Cost of 1983 increased fuel
Additions wet controls use due to NOX
(MM) (103 $) control (10° GJ)
1983 increased fuel
cost due to NOX Annualized control cost mills
control (103 $) (103 $) g~NOx
Simple
Combined
Total
4927
3253
8180
49,270
32,530
81,800
0.5
4.9
5.4
2,000
19,600
21,600
14,300
27,700
42,000
3.25
0.7
0.9
ro
Assumptions
Fuel cost ($/GJ) $4.00
Heat rate (GJ/kWh)
Simple 0.0109
Combined 0.0095
Cost for wet controls ($/kW) 10.00
NOX reduction (ng/J) required
to meet NSPS of 75 ppm
0 15% 02 167
Fuel penalty (X) 2
Hours/year used
Simple cycle 500
Combine cycle 8000
Annualized cost = 25% of installation cost + increased fuel cost, where 25X
+ 5% operation and maintenance
20% for fixed costs
aSince regenerative cycle machines constitute a low percentage of the installed turbines and the regenerative
market has been very slow, they were not considered in this analysis.
-------
6.6 EFFECTIVENESS OF N0¥ CONTROLS
A
Table 6-5 sunmarizes the percent of NO reduction achievable with
/\
the Best Available Control Technology (BACT) for stationary gas turbines,
i.e., water injection. The amount of NO reduction can be varied by
A
changing the water/fuel ratio. However, certain tradeoffs, such as greater
expense and increased maintenance and operation problems, tend to limit the
percent reduction achievable. Also shown in Table 6-7 are control
potentials for advanced technologies. While dry controls (employing various
combustion modification techniques) are expected to be commercially
available by 1983, catalytic combustion's future is less certain, although
the NO control payoffs are potentially much greater.
A
Wet controls are a proven NOV control technology. Section 5.2.1.1
A
has shown that water and steam injection are capable of reducing NOX
emissions on clean fuels to below the NSPS emission limit of 75 ppm at 15
percent 07. The degree of NO reduction achieved can be controlled by
£ /\
adjustment of the water/fuel ratio. Increasing this value increases the
percent of NO reduction. Typically a water/fuel ratio of 0.5 would be
/\
sufficient to reduce uncontrolled NOV emissions in a 60 MW oil-fired
A
simple cycle gas turbine to a value below 75 ppm. To obtain this same
level of NO reduction when firing natural gas, a ratio of only 0.25
A
would be required since the uncontrolled NO emission level when firing
A
gas is less than when firing oil. Wet controls are not likely to be
applied to engines firing dirty fuels since studies have shown them to be
counter-productive to NO reduction in the presence of fuel bound
A
nitrogen.
The costs associated with the degree of control through the
application of wet controls are difficult to accurately quantify. Utility
users have reported installed costs of water injection ranging from $5 to
$23/kW in 1978 dollars. In addition to capital costs, there are annual
operating and maintenance costs, including water treatment costs and a
fuel penalth due to a two to five percent increase in heat rate. These
cost effects are even more severe for small size turbines where economies
of scale cannot be realized.
Dry NOX controls are currently being developed with the intent of
fully replacing wet controls in new stationary turbines by the early
1980's. Indeed, users would welcome the day when they could replace wet
6-22
-------
TABLE 8-7.
NOX CONTROLS: BEST AVAILABLE CONTROL TECHNOLOGY (BACT)
AND ADVANCED TECHNOLOGY
Control Technique
NOX Emission Level,3
ppm at 15% 02
BACT
Water Injection
.25 water/fuel ratio
.50 water/fuel ratio
.75 water/fuel ratio
130
75
40
Advanced
Technology
Dry Controls (1983)
Catalytic Combustion (1985-1990)
40
10
aFor clean liquid and gaseous fuel, typical peaking utility gas turbine,
values are + 20 percent, baseline NOX emission - 175 ppm at 15% 03
controls with the more simplified and less costly dry controls. One of
the most promising designs comes from an EPA program being conducted by
Pratt and Whitney Aircraft. The experimental portion of this program has
been completed with the most promising dry control concept, rich
burn/quick quench, having been tested in a scale-up to the 25 MW size
range. All program goals were met and exceeded. NO levels were
/\
approximately 40 ppm with clean fuels and as low as 75 ppm with
0.5 percent fuel nitrogen. CO emissions were well below 100 ppm,
indicating excellent combustion efficiency. In addition, the dry control
combustor, when commercially available, is expected to cost only
15 percent more than a conventional combustor. This is just a fraction of
what wet controls would cost. Wet controls, both water and steam
injection, have been shown in numerous installations to be fully capable
of meeting NSPS. However, this does not come at negligible expense and
headache to the user. These control systems can be costly to install and
operate and some users have reported deleterious effects on the day to day
operations and maintenance of their engines. Considerable operating and
6-23
-------
maintenance problems and costs associated with wet controls will be all
but eliminated with dry controls. However, existing operational
complexities will have to be solved before dry controls become
commercially demonstrated.
NO emissions are probably the primary pollutant of concern with
A
stationary gas turbines. CO and HC emissions are somewhat smaller in
magnitude and not easily controlled without severely upsetting NOX
emissions. SO emissions are purely a function of the fuel sulfur
A
content and can only be economically controlled by selecting low sulfur
fuels. Flue gas desulfurization is, at present, prohibitively expensive
for gas turbines. Since by most estimates the earliest that dry NO
X
controls will be fully developed and demonstrated in full scale gas
turbines is by 1982 or 1983, wet controls are certainly the preferred,
indeed the only, means of controlling NO emissions in gas turbines.
J\
Dry NO controls, when they become commercially available, will
A
undoubtedly start to replace wet controls as the preferred NO control
A
technique in new sources. Assuming dry controls will eventually be
commercially available and can reduce NO emissions to meet the NSPS
A
without adversely affecting CO and HC emissions, they will then be the
much preferred control technique. Dry controls will be much simpler than
wet controls, involving simply a different kind of combustor can. None of
the ancillary equipment associated with wet controls will be required.
Cost estimates, while difficult to accurately predict at this stage of
development, are expected to be, at a maximum, the same as wet
controls. One developer expects them to cost approximately 15 percent
more than a conventional combustor can.
Ultimately, if catalytic combustion can be demonstrated in full
scale engines, then it will probably become the preferred option.
Emission reductions with this technique have been significant, approaching
98 percent. However, this technology is in its infancy and should not be
considered a real possibility until the late 1980's.
A number of organizations, including EPA, have done air quality
modeling studies to determine the effect that stationary gas turbines have
on ambient air quality. Emissions from units burning various fuels,
having different degrees of control, equipment configurations and
meteorological conditions, were compared to the NAAQS. The details of
6-24
-------
these studies were previously reported in Section 6.2.3. In sumnary, it
appears that most permutations of the above variables result in a very
small impact on the NAAQS. Under certain unusual circumstances the NAAQS
for CO and SO? was surpassed. While the N0¥ NAAQS were never exceeded
Cf A
in these modeling studies (assuming zero baseline ambient NO ), the
A
NAAQS could easily be exceeded by certain turbine arrangements in regions
that were approaching nonattainment.
EPA in Reference 6-2 has estimated that wet controls would reduce
national NOX emissions by 190,000 tons per year by 1982 and by
400,000 tons per year by 1987. It is estimated that electricity produced
from these turbines would cost an additional 2 to 3 percent.
6-25
-------
REFERENCES FOR SECTION 6
6-1. Water land, L. R., et al., "Environmental Assessment of Stationary
Source NOX Control Technologies — Final Report," Acurex Report
FR-80-57/EE, EPA Contract 68-02-2160, Acurex Corp., Mountain View,
CA, April 1980.
6-2. Goodwin, D. R., et_ aj_._, "Standard Support and Environmental Impact
Statement. Volume 1: Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
6-3. Schalit, L. M., and K. J. Wolfe, "SAM/IA: A Rapid Screening Method
for Environmental Assessment of Fossil Energy Process Effluents,"
EPA-600/7-78-051, NTIS-PB 277 088/AS, February 1978.
6-4. Hangebrauck, R. P., e_t ^K_, "Nomenclature for Environmental
Assessment Projects: Part 1 — Terminology for Environmental
Impact Analyses," U.S. Environmental Protection Agency, Industrial
Environmental Research Laboratory, Research Triangle Park, NC,
August 1979.
6-5. Waterland, L. R., and L. B. Anderson, "Source Analysis Modeling for
Environmental Assessment," Presented to Fourth Symposium on
Environmental Aspects of Fuel Conversion Technology, Hollywood, FL,
April 17, 1979.
6-6. Cleland, J. G. and G. L. Kingsburg, "Multimedia Environmental Goals
for Environmental Assessment. Volumes I and II," EPA-600/7-77-136a
and b, NTIS PB-276 919 and 920, November 1977.
6-7. Personal communication with S. Lanier, Industrial Environmental
Research Laboratory, U.S. EPA, Research Triangle Park, NC,
October 1978.
6-8. Zeltmann, E. W., "U.S. Environmental Standards," Mechanical
Engineering, 100 (2), 32 (1978).
6-9. Larkin, R. and E. B. Higginbotham, "Combustion Modification Control
for Stationary Utility Gas Turbines, Volume II: Utility Unit Field
Test," EPA 600/7-81-122b, July 1981.
6-10. Personal communication with D. Felsienger, San Diego Gas and
Electric, San Diego, CA, August 1978.
6-11. Personal communication with H. Dvorak, Turbo Power and Marine
Systems, Farmington, CT, August 1978.
6-12. Personal communication with W. Price, New Mexico Electric
Hobbs, MM, August 1978.
6-26
-------
6-13. Personal communication with P. Dinenno, Westinghouse Electric
Corporation, Philadelphia, PA, August 1978.
6-14. Goldberg, P., et al., "Potential and Problems of Premixed Combustors
for Application to Modern Aircraft Gas Turbine Engines," presented at
the AAIA/SAE 12th Propulsion Conference, Palo Alto, CA, July 1976.
6-15. Roberts, P. B., et al., "Advanced Low NOX Combustors for Supersonic
High Altitude Aircraft Gas Turbines," ASME 76-GT-12.
6-16. Personal communication with L. Ezzell, Pacific Gas and Electric,
San Francisco, CA, August 1978.
6-17. Personal communiction with W. Price, New Mexico Electric, Hobbs, NM,
August 1978.
6-18. de Bias, V., "FT50 Design Shortcut to 1980 Technology," Gas Turbine
World, November 1975.
6-19. "Sixth Biennial Survey of Power Equipment Requirements of the U.S.
Electric Utility Industry 1977 - 1986."
6-27
-------
SECTION 7
DATA BASE EVALUATION AND NEEDS
This section presents an evaluation of the data base which was used
in assessing the environmental impacts of stationary gas turbines.
Consideration was given to the sources of data as well as to the quantity
and quality of the data applied to each element of this report. The value
of completed and ongoing control technology research and development
programs are also discussed, with much of the emphasis put on dry
controls. It seems that as dry controls are developed and become
commercially feasible, little additional effort will be expended on
refining wet controls. Of course, wet controls are already capable, in
principle and in many actual operations, of reducing NO emissions to
X
below the proposed regulation of 75 ppm at 15 percent 0^.
In addition to discussing the environmental assessment data base,
this section outlines various data requirements to more fully and
accurately evaluate the total impacts due to NO controls on gas
A
turbines. In defining these data needs, primary focus has been placed on
development of standards support, control technology R&D, and on further
testing and studies to upgrade the environmental assessment.
7.1 DATA BASE REVIEW
Water and steam injection are widely accepted by the gas turbine
industry as valid means of controlling NO emissions. Indeed, water
X
injection has been used for load augmentation since 1961 by industrial
users. Numerous gas turbine units equipped with water injection have been
sold, particularly to utility customers. However, since there are
relatively few regional regulations controlling NOV emissions, most of
X
these users with water injection simply have not used it for appreciable
lengths of time. Consequently we have had to rely to a great extent on
7-1
-------
information and data supplied by manufacturers, EPA and those few users
who are required to meet a NO emission limit.
A
A substantial amount of good quality data regarding wet control
emissions reduction was collected by EPA for use in the SSEIS
(Reference 7-1). Results from certain additional emissions testing
programs since the completion of the SSEIS have been added to this data
base. These data support the general trends reported in the SSEIS
regarding NO emission reduction and the relationships between percent
A
NO reduction and the water/fuel ratio. Thus, these data in conjunction
A
with other supporting data in the literature indicate that water and steam
injection are viable control techniques for NO .
/\
A significant amount of disparity was found in users, manufacturers
and EPA estimates for costs of wet controls, both capital and operating
cost estimates. Other than for large scale gas turbines where a few users
could supply actual cost data, we have had to rely primarily on
manufacturer's and EPA's cost estimates. Even for large scale units, data
are sparse since most users are not required to control NO . Again, due
X
to the lack of NO regulations and due to differences in accounting
procedures, it is difficult to accurately predict any additional operating
and maintenance costs. Day-to-day operations were found to have a major
effect on the costs incurred by users. The most serious cost data gaps
are with wet control costs for small and medium size turbines and with all
aspects of cost estimates for dry controls on all size units.
As in the case with estimating NO control costs, it is also
difficult to accurately predict the effect that wet controls have on gas
turbine operations and maintenance. The best source for this type of
information is, of course, the long term user. Of the users we have
contacted, some of which have accumulated many thousands of hours of
operation with wet controls, the majority seem to have few additional
problems attributable to wet controls. On the other hand, some users have
reported significant problems, usually related to the water injection
equipment itself or to materials problems with hot components (i.e.,
deposits, corrosion and embrittlement). All sources agree that water
injection imparts a fuel penalty on normal operations, thereby increasing
the heat rate. A portion of this penalty may be offset by increases in
power output. The variability in the reported data seems to be primarily
7-2
-------
a function of day-to-day operations and maintenance procedures.
Furthermore, the data indicate that there is nothing inherent with water
injection systems that would cause excessive problems. It also appears
that proper operational and maintenance procedures can obviate most
difficulties encountered with these systems.
Dry NOX controls are in the development stages. Numerous
specific designs are being applied to a few general NOX control concepts
with a common goal to design a gas turbine combustor that will control
NOX to the regulated limit, without sacrificing other pollutant emissions
or combustion efficiency. Given the existing state of development, it is
not possible to perform a comprehensive and accurate environmental
analysis. We have had to rely almost totally on experimental combustor
programs reported in the literature and some industry contacts to
qualitatively assess the postulated effects of the most promising dry
control concepts. Unfortunately, the available literature does not give
the whole picture regarding evolving dry controls. Much of the design
information is proprietary to the manufacturers and is not available.
Further, the work conducted so far is generally on a single combustor rig
although some full-scale demonstrations have been performed. Significant
R&D effort will be required before these dry control concepts can be
reliably applied to actual engines. The degree of R&D required will be
determined by the sophistication of combustor rig experiments, how well
actual combustor inlet conditions are simulated and how accurately actual
engine conditions are duplicated. Moreover, once combustor rigs are
scaled up to full size engines, long term reliability and life tests will
have to be performed.
While the data indicate that the pollution reduction potential of
catalytic combustion is extremely promising, this technology must only be
considered as a long term control technique. As far as application to
full scale engines is concerned, this technology is many years from
fruition. Some manufacturers even voice skepticism that it will ever
prove commercially viable.
The available data for all dry NO control techniques make only
A
qualitative judgments and estimates possible when predicting costs,
jncremental emissions and operations and maintenance impacts. When
assessing this report's evaluation, one must keep in mind that dry
7-3
-------
controls are an emerging technology and are at least two to four years
from the first full scale application. Numerous changes in the preferred
design, which could have a major impact on the predicted environmental
effects, may occur.
7.2 DATA NEEDS
As users and manufacturers gain experience, additional data ought
to be evaluated on a periodic basis during the life cycle of existing,
modified and new stationary gas turbines. This is necessary due to the
nature of the evolutionary process that dry NO control techniques are
A
undergoing and the considerable lack of long term operating experience
with wet controls. In this manner maximum benefit will be derived through
use of emerging data to:
• Support standards development
• Support effects and control technology R&D
• Define needs for further testing to expand the data base
0 Define needs for further studies to upgrade the environmental
assessment.
There are specific areas where there appears to be a general lack
of consensus regarding certain impacts from wet controls. These include:
1) water injection cost data for capital equipment, operating and
maintenance expenses, 2) the cost/benefit ratio of wet controls for small
gas turbines (<4 MW electrical output), 3) quantification of the fuel
penalty due to increased heat rate and additional power output resulting
from more mass throughout, and 4) quantification of the effect of NO
A
controls on incremental emissions of pollutant species other than NO .
/\
Wet control technology and its associated NO reduction potential
A
are well developed and understood. In addition, wet controls will
essentially become obsolete for new sources with the expected advent of
dry controls by 1982. Dry controls, however, are an emerging technology,
and there are many unanswered questions regarding incremental effects and
associated costs. Manufacturers appear to be focusing in on the most
effective dry control concept in reducing NOX while minimizing
incremental emissions and maintaining acceptable system efficiencies. The
next critical step is scaling up to full size engines, assessing the
various environmental impacts and developing long term operating
experience.
7-4
-------
At present, it seems clear that dry NOX controls will be the
preferred control technology option for new gas turbines within five
ars« Due to their present state of development, essentially no data
regarding emission levels, control costs and operations and maintenance
impacts exist for the application of dry controls to full scale engines.
of these data are required to perform a meaningful environmental
dry NOX control technology. As the direction of dry
controls research becomes evident, additional testing programs can be
suitably designed to provide the proper data base. Then, as dry controls
become commercially feasible and users gain operating experience, any
additional data gaps may be filled accordingly. The types of gaps will
primarily be additional operating and maintenance costs that can only be
accurately predicted through long term accounting of such expenditures.
Only by such careful front-end tracking of dry control developments can a
comprehensive environmental assessment be performed.
REFERENCE FOR SECTION 7
7-1. Goodwin, D. R., et al., "Standard Support and Environmental Impact
Statement. Volume 1: Proposed Standards of Performance for
Stationary Gas Turbines," EPA-450/2-77-017a, NTIS-PB 272 422/7BE,
September 1977.
7-5
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-81-122a
2.
3. RECIPIENT'S ACCESSION- NO.
4. TITLE AND SUBTITLE Combustion Modification Controls for
Stationary Gas Turbine: Volume I. Environmental
Assessment
5. REPORT DATE
July 1981
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
i8. PERFORMING ORGANIZATION REPORT NO.
R. Larkin, H.I. Lips, R.S. Merrill, and
K.J. Lim
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex/Energy and Environmental Division
485 Clyde Avenue
Mountain View, California 94042
10. PROGRAM ELEMENT NO.
EHE624A
11. CONTRACT/GRANT NO.
68-02-2160
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERI.OD COVERED
Task Final; 2/78-6/80
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES IERL-RTP project officer is Joshua S. Bowen, Mail Drop 65,
919/541-2470.
is. ABSTRACT
report gives an environmental assessment of combustion modification
techniques for stationary gas turbines, with respect to NOx control effectiveness,
operational impact, thermal efficiency impact, control costs, and effect on emis-
sions of pollutants other than NOx. Wet controls, which inject steam or water direc-
tly into the combustion chamber, are the only available methods sufficiently devel-
oped to reduce NOx below the recently recommended New Source Performance
Standard of 75 ppm NO2 at 15% O2 for clean fuels (greater than 50% reduction). How-
ever, the effectiveness of wet controls decreases significantly as the percentage of
fuel-bound nitrogen increases. Emissions of unburned hydrocarbons and CO can
increase with wet controls; however, a detailed Level 1 test on a 60-MW utility gas
turbine indicated that incremental emissions other than NOx remained relatively
unchanged. Wet controls increase the cost of electricity by 2-5% due, in large part,
to the associated fuel penalty. Dry NOx controls are being developed based on
combustor modifications that do not involve water or steam injection. They are
promising because of their NOx control effectiveness for both clean and dirty fuels ,
and their expected lower operational and cost impacts.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Pollution
Gas Turbines
Combustion Control
Assessments
Pollution Control
Stationary Sources
Combustion Modification
Environmental Assess-
ment
13B
13G
21B
14B
13. DISTRIBUTION STATEMEN1
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
173
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
7-6
------- |