Professional Documents
Culture Documents
org/9780521764056
ht
tp
s:
//b
oi
le
rs
in
fo
.c
om
Gas Turbine Emissions
The development of clean, sustainable energy systems is one of the grand challenges of
our time. Most projections indicate that combustion-based energy conversion systems
will remain the predominant approach for the majority of our energy usage. Moreover,
gas turbines will remain a very significant technology for many decades to come, whether
for aircraft propulsion, power generation, or mechanical drive applications. This book
compiles the key scientific and technological knowledge associated with gas turbine emis-
sions into a single authoritative source. The book has three parts: the first part reviews
major issues with gas turbine combustion, including design approaches and constraints,
within the context of emissions. The second part addresses fundamental issues associated
with pollutant formation, modeling, and prediction. The third part features case studies
from manufacturers and technology developers, emphasizing the system-level and prac-
tical issues that must be addressed in developing different types of gas turbines that emit
pollutants at acceptable levels.
om
Timothy C. Lieuwen is professor of aerospace engineering and executive director of the
Strategic Energy Institute at the Georgia Institute of Technology. Lieuwen has authored
one textbook, edited two books, written seven book chapters and more than 200 papers,
.c
and received three patents. He chaired the Combustion and Fuels Committee of the
fo
International Gas Turbine Institute of the American Society of Mechanical Engineers
(ASME). He is also on the Propellants and Combustion Technical Committee of the
in
American Institute of Aeronautics and Astronautics (AIAA), and he previously served
rs
on the AIAA Air Breathing Propulsion Technical Committee. He has served on a variety
of major panels and committees through the National Research Council, Department
le
is the editor in chief of the AIAA Progress in Astronautics and Aeronautics series and
is serving or has served as an associate editor of the Journal of Propulsion and Power,
//b
Combustion Science and Technology, and the Proceedings of the Combustion Institute.
Lieuwen is a Fellow of the ASME and received the AIAA Lawrence Sperry Award and
s:
the ASME Westinghouse Silver Medal. Other recognitions include ASME best paper
awards, the Sigma Xi Young Faculty Award, and the NSF CAREER award.
tp
ht
Vigor Yang is the William R. T. Oakes Professor and chair of the School of Aerospace
Engineering at the Georgia Institute of Technology. Prior to joining the faculty at
Georgia Tech, he was the John L. and Genevieve H. McCain Chair in Engineering at the
Pennsylvania State University. His research interests include combustion instabilities in
propulsion systems, chemically reacting flows in air-breathing and rocket engines, com-
bustion of energetic materials, and high-pressure thermodynamics and transport. Yang
has supervised more than forty PhD and fifteen MS theses. He is the author or coauthor
of more than 300 technical papers in the areas of propulsion and combustion and has pub-
lished ten comprehensive volumes on rocket and air-breathing propulsion. He received
the Penn State Engineering Society Premier Research Award and several publication
and technical awards from AIAA, including the Air-Breathing Propulsion Award (2005),
the Pendray Aerospace Literature Award (2008), and the Propellants and Combustion
Award (2009). Yang was the editor in chief of the AIAA Journal of Propulsion and
Power (20019) and is currently the editor in chief of the JANNAF Journal of Propulsion
and Energetics (since 2009) and coeditor of the Cambridge Aerospace Series. He is a
Fellow of the American Institute of Aeronautics and Astronautics, American Society of
Mechanical Engineers, and Royal Aeronautical Society.
ht
tp
s:
//b
oi
le
rs
in
fo
.c
om
Cambridge Aerospace Series
Editors:
Wei Shyy
and
Vigor Yang
om
10. G. A. Khoury (ed.): Airship Technology, Second Edition
11. J. P. Fielding: Introduction to Aircraft Design
.c
12. J. G. Leishman: Principles of Helicopter Aerodynamics, Second Edition
13. J. Katz and A. Plotkin: Low-Speed Aerodynamics, Second Edition
fo
14. M. J. Abzug and E. E. Larrabee: Airplane Stability and Control: A History of the Technologies
in
that Made Aviation Possible, Second Edition
15. D. H. Hodges and G. A. Pierce: Introduction to Structural Dynamics and Aeroelasticity,
rs
Second Edition
16. W. Fehse: Automatic Rendezvous and Docking of Spacecraft
le
20. C. A. Wagner, T. Httl, and P. Sagaut (eds.): Large-Eddy Simulation for Acoustics
21. D. D. Joseph, T. Funada, and J. Wang: Potential Flows of Viscous and Viscoelastic Fluids
s:
22. W. Shyy, Y. Lian, H. Liu, J. Tang, and D. Viieru: Aerodynamics of Low Reynolds Number
tp
Flyers
23. J. H. Saleh: Analyses for Durability and System Design Lifetime
ht
Edited by
Timothy C. Lieuwen
Georgia Institute of Technology
Vigor Yang
om
Georgia Institute of Technology
.c
fo
in
rs
le
oi
//b
s:
tp
ht
cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town,
Singapore, So Paulo, Delhi, Mexico City
om
no reproduction of any part may take place without the written
permission of Cambridge University Press.
.c
First published 2013 fo
Printed in the United States of America
in
A catalog record for this publication is available from the British Library.
rs
Lieuwen, Timothy C.
Gas turbine emissions / Timothy C. Lieuwen, Vigor Yang.
oi
TJ778.L524 2013
tp
621.433dc23 2012051616
ht
Cambridge University Press has no responsibility for the persistence or accuracy of URLs
for external or third-party Internet Web sites referred to in this publication and does not
guarantee that any content on such Web sites is, or will remain, accurate or appropriate.
Contents
om
Preface xv
.c
Part 1 Overview and Key Issues fo
1 Aero Gas Turbine Combustion: Metrics, Constraints, and
in
System Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
rs
Willard Dodds
tp
vii
viii Contents
Index 363
om
.c
fo
in
rs
le
oi
//b
s:
tp
ht
Contributors
ix
Foreword
Alan H. Epstein
When I first became interested in jet engines, smoke trails from the then ultramodern
Boeing 707s were an arresting feature of that modern world. Ten years later, smoke
was regulated and the U.S. Federal Aviation Administration had canceled the
Boeing 2707 supersonic airliner program in the midst of growing environmental
concerns. Back in the early 1960s, ground-based gas turbines were a very small
business and concern for the environment was only minor. Over the five decades
since the 707, the role of gas turbines in our society has greatly expanded, and con-
cern regarding their emissions has grown even faster. Now, the electric power gen-
eration gas turbine business has outgrown that of aircraft engines and emissions
have become a market discriminator. Indeed, large fortunes have been won and
lost on the basis of the emissions performance of land-based gas turbine engines.
On the aero engine side, emissions performance is now featured in engine market-
ing campaigns.
Combustion emissions might be thought an arcane topic. It is certainly complex.
It is also of great importance to our society given the dominance of gas turbines
for aircraft propulsion and power generation. There are three, basically indepen-
dent, complicated problems associated with gas turbine emissions the design of
low-emissions combustors, the prediction of the effects of emissions on human
health and the global environment, and the formulation of balanced and effec-
tive policy and regulation. These challenges are important to three very different
groups technical folk, businesspeople, and policy makers and regulators. This book
will be of interest to them all.
For the technical community, the science of how emissions are generated in a
gas turbine combustor and their interactions with the atmosphere has always been a
fascinating but challenging subject. The relatively recent concern for climate change
has increased the complexity of the atmospheric science problem, especially for air-
craft engines, from one mainly concerned with local air quality at low altitude to
more complex interactions at the tropopause and in the stratosphere. During the last
fifty years, design engineers have risen to the environmental challenge by realizing
combustors with much lower emissions while at the same time significantly increas-
ing reliability and life. One important aspect of combustor engineering, however, has
xi
xii Foreword
not changed over this time we still do not have the technology needed to predict
gas turbine emissions from first principles. The lack of first principles capabilities
drives up product development costs and business risk.
Policy makers and regulators, who are not necessarily technical experts in the
fields they regulate, face interesting challenges as well. These can be grouped into
three general categories technical, political, and diplomatic. Technical questions
include, for example, consideration of currently unregulated emissions such as very
small particulates and CO2, as well as the role uncertainty plays in resolving con-
flicting requirements such as NOx and CO2. Political challenges abound and include
issues such as how to best balance environmental protection with economic growth
and how to balance local air quality with global climate change. Gas turbine emis-
sions have also become a major diplomatic challenge. Aviation is the most interna-
tional of endeavors, both in manufacture and operation. Most engines have parts
and major subsections designed and manufactured in several countries. Aircraft take
off and land in different countries thousands of times a day and so fall under the pur-
view of more than one regulator. It is critical to the efficient operation of the worlds
air transportation system that regulations be harmonized across the globe. This is the
job of the International Civil Aviation Organization (ICAO), a branch of the United
Nations with 189 member states. Getting 189 countries to agree on anything has
never been easily or quickly achieved. The rise of climate change as a major world-
wide issue with its attendant political and economic implications has only increased
the complications of international rule making.
From the point of view of technical and policy folks, gas turbine combustor
emissions bring fascinating challenges. For the business community, the fascination
turns to dread. Why the dichotomy? The confluence of regulation and technical chal-
lenge generates business uncertainty and risk, with financial penalties large enough
to destroy a business. Manufacturers of ground-based engines are often contrac-
tually responsible for the price of the electric power not produced if an engine is
deficient. An engine that does not meet local air quality standards cannot be oper-
ated, and may incur liabilities that dwarf the price of the engine. Manufacturers
of aircraft engines face similar challenges; that is, until an engine meets emissions
requirements, it will not be certified by regulatory authorities. Such engines cannot
be legally shipped, and so the airplanes, which cost ten times more than the engine,
cannot be delivered. Gas turbine development can cost up to two billion U.S. dollars,
so long production runs are needed to amortize the cost. The business risk asso-
ciated with emissions regulations is further amplified by the long-lived nature of
the products. Engines typically have service lives of thirty years or more. Over this
time span, emissions regulations usually change. Increased stringency can reduce
the residual value of an engine, hinder sales, and even prohibit operation of engines
in the field. Additional uncertainty is introduced by the degree to which regulations
are not harmonized across political boundaries since niche markets cannot support
high development costs. Thus, business planning for gas turbine emissions is a chal-
lenge and a concern.
Foreword xiii
These are hard problems. These are interesting problems. These are important
problems at the confluence of engineering, regulation, and business. This book is the
first to cover both the technical and regulatory aspects of gas turbine emissions. With
chapters authored by some of the worlds experts in their respective fields, it has the
breadth and depth to be of interest to all the stakeholders. It is valuable for experts
in the field and informative for those just getting involved.
Preface
The development of clean, sustainable energy systems is one of the grand challenges
of our time. Environmental and energy security concerns, coupled with growing
energy demand, require us to increase, diversify, and optimize the use of energy
sources while reducing the adverse environmental impacts of energy production,
transmission, and use. In particular, we are confronted with four interacting issues:
climate change, local air and water quality, energy supply, and energy security.
Global warming has led to significant discussions about reductions of carbon dioxide
emissions. Meanwhile, concerns about energy security and supplies for a growing uti-
lization base are driving us to consider broader and more reliable energy resources.
Finally, local air quality concerns are driving interest in other pollutants that lead to,
for example, acid rain or photochemical smog, and that have additional implications
for the management of power plant operations and emissions.
Gas turbines will continue to be an important combustion-based energy con-
version device for many decades to come, for aircraft propulsion, ground-based
power generation, and mechanical-drive applications. At present, gas turbines are
a principal source of new power-generating capacity throughout the world, and
the dominant source for air-breathing flight vehicles as well. Over the last decade,
power generation from alternative sources, such as solar and wind, has significantly
increased. Nevertheless, most projections indicate that the relative fraction of
energy supplied by these sources will remain small, even several decades from now.
These projections also indicate that gas-turbine-based combined cycle plants will
continue to represent the majority of new power generation capacity. Moreover, as
the supply of intermittent renewables grows, gas turbines will play an increasingly
important role in stabilizing the electrical grid, where the supply and demand of
electric power must match at every instant in time. The topic of gas turbine emis-
sions, both traditional pollutants (NOx, CO, UHC, particulates) and CO2, is clearly
of significant interest.
In the aviation sector, emissions regulations continue to tighten. Climate change
may lead the worldwide community to begin taxing carbon emissions for aircraft,
and cloud formation associated with water vapor emissions continues to be an area
of research. Particulate and NOx emissions can significantly influence local air quality
xv
xvi Preface
1.1Introduction
The aircraft gas turbine engine is a complex machine using advanced technology
from many engineering disciplines such as aerodynamics, materials science, combus-
tion, mechanical design, and manufacturing engineering. In the very early days of
gas turbines, the combustor section was frequently the most challenging (Golley,
Whittle, and Gunston, 1987). Although the industrys capability to design combus-
tors has greatly improved, they remain an important design challenge.
This chapter will describe how the combustor interacts with the rest of the
engine and flight vehicle by describing the relationship between attributes of
the engine and the resulting requirements for the combustor. Emissions, a major
engine performance characteristic that relies heavily on combustor design, will be
introduced here with more detail found in following chapters. The wide range of
operating conditions a combustor must meet as engine thrust varies, which is a
major challenge for combustor design, will also be described. Last, the relationship
between combustor exit temperature distribution and turbine section durability
will be discussed.
3
4 Aero GT Combustion
Optimize fuel consumption High combustion efficiency and low combustor pressure loss
Meet emissions requirements Minimize emissions and smoke
Wide range of thrust Good combustion stability over entire operating range
Ground and altitude starting Easy to ignite and propagate flame
Turbine durability Good combustor exit temperature distribution
Overhaul and repair cost Meet required combustor life by managing metal temperatures
and stresses
Stability Durability
Emissions
Altitude relight
Exit temperature
and starting
Core flow
Compressor
Temperature
Combustor Pressure
Turbine
Pressure Temperature
mechanisms, such as NOx formed in the flame zone itself, are also described in
Chapter 7, but are negligible for aircraft engines.
When fuel-rich regions of the combustor flow exist at high pressures and tem-
peratures, the formation of small particles of carbon can occur. These carbon par-
ticles result from complex chemical processes and undergo multiple processes within
the combustor such as surface growth, agglomeration, and oxidation prior to leaving
the combustor, as detailed in Chapter 5. These particles pass through the turbine and
exit the engine in the exhaust. When the concentration of the particles in the exhaust
is high enough to be visible, as was often the case in early gas turbines, it is referred
to as smoke or soot. Recently, the more general term particulate matter (PM) has
been used to describe this emission. Modern engine smoke levels are invisible but
still possess large quantities of very small soot particles and aerosol soot precursors
(see Chapter 5) at the exhaust. Emerging research on the effect of PM on health
and climate focuses more attention on measuring, modeling, and understanding the
processes governing PM production.
These relationships between engine power conditions and emissions production
lead to the behavior shown in Figure 1.3. As shown in the figure, levels of UHC and
CO are highest at low power and drop quickly with increasing thrust. Conversely,
NOx and PM increase with engine power and are typically maximized at maximum
power. Chapters 5 and 7 discuss these emissions formation processes in moredetail.
A typical commercial engine mission consists of ground starting, taxi, takeoff, climb
to altitude, cruise, deceleration to flight idle and descent, approach, touchdown,
1.5 Effect of Range of Thrust and Starting Conditions on the Combustor 7
60
HC
50 CO
NOx
Smoke
EI or Smoke Number
40
30
20
10
0
0 20000 40000 60000 80000 100000
Thrust
Figure 1.3. Emissions versus power level for the PW4084.
thrust reverse, and taxi in. The extremes in combustor operating conditions drive the
overall design approach. The combustor must meet performance, operability, and
emissions metrics over the full range of operation. In order to do so, it must operate
at the following extremes:
1. Minimum fuel-air ratio This occurs during decelerations from high power to
low power. Flight decelerations normally occur when descending from high alti-
tude cruise and during approach throttle movements. They can also occur in
emergencies. Minimum fuel-air ratio typically depends on the thrust decay rate,
as the time response of the engine turbomachinery that governs the airflow is
much longer than that of the fuel flow. Risk of weak extinction (flameout) is
highest during decelerations.
2. Minimum operating temperatures and pressures These occur at flight and
ground idle conditions. Low pressure and temperature challenges combustion
efficiency due to slower fuel vaporization and chemical kinetics.
3. High operating temperatures and pressures These occur at takeoff, climb,
thrust reverse, and cruise conditions. These conditions result in the bulk of NOx
formation and the most severe liner metal temperatures.
4. Ignition conditions Ignition normally occurs on the ground but also occasion-
ally in flight. Ignition is required at near surrounding ambient pressure and
temperature. High altitude and extremely cold conditions are typically the most
challenging to achieve ignition, flame propagation, and flame stabilization. These
conditions lead to low temperature (40F) and pressure (4 psia at 35,000 ft.)
combustor inlet conditions.
Thus, the combustor design must meet the performance, emissions, and durability
requirements at low- and high-power operations without compromising stability
8 Aero GT Combustion
(a)
(b)
Figure 1.4. (a) Can-annular combustor (Pratt & Whitney JT8D-200); (b) RQL annular com-
bustor (IAE V2500).
Fixed-geometry combustors have been used in the gas turbine industry since its
inception. Early designs used multiple cans in a circumferential array. The cans
transitioned through an annular duct to the turbine (Figure 1.4a). Later designs
used an annular duct geometry that allowed for reduced overall length and weight
(Figure1.4b). Annular combustors also have reduced liner surface area relative to
can-annular combustors and therefore use less cooling. All designs use multiple fuel
injectors to provide spray atomization and fuel-air mixing. Achieving good atomiza-
tion and fuel-air mixing is critical for efficient combustion, low emissions, and good
temperature uniformity into the turbine. Normally, the fuel is injected in the front end
of the combustor and flow recirculation is created to provide a stabilization region
for the combustion process. This is typically accomplished with air swirlers, which
leads to vortex breakdown and flow recirculation. The stabilization zone promotes
recirculation of hot product gases forward to the incoming fuel spray, thereby pro-
viding a continuous ignition source and faster fuel droplet evaporation. Accelerated
droplet evaporation is critical to high-efficiency combustion at low-power conditions,
when low air inlet temperatures are insufficient to provide fast enough evaporation.
1.5 Effect of Range of Thrust and Starting Conditions on the Combustor 9
Steady state
Combustor inlet condition
fuel-air ratio
Temperature
Pressure
Transient decel
fuel-air ratio
Idle Thrust Take-off Idle Thrust Take-off
Figure 1.5. Combustor operating conditions.
If continuous ignition is not provided at low power, the vaporization and reaction
times can exceed the combustor residence time and flameout occurs.
The airflow distribution in a fixed-geometry combustor must be selected to
achieve both low- and high-power performance requirements. Conditions at the
combustor inlet vary significantly between low-power idle and high-power takeoff
conditions. At idle, inlet temperature, pressure, and global fuel-air ratio are relatively
low. At takeoff, the opposite is true (Figure 1.5). The operating temperatures and
pressures are largely a function of the engine thermodynamic cycle; therefore the
most significant parameter for the combustor designer to consider is the fuel-air
ratio. Because air is introduced in stages along the length, the designer can tailor
the airflow distribution to achieve key performance metrics. This creates a distri-
bution in fuel-air ratio along the length of the combustor, leading to variations in
local temperature as power level is adjusted. The difference in fuel-air ratio between
high-power takeoff and low-power deceleration and idle conditions is critical because
it determines the range of local fuel-air ratio in the front end of the combustor. For
most modern gas turbines, the difference is large enough that the front end operates
fuel rich (f/a > 0.068 for jet fuel) at takeoff conditions. Consequently, fixed-geometry
combustors are referred to as rich-burning or rich-quench-lean (RQL) designs. This
refers to the rich front-end fuel-air ratio that is diluted (quenched) by additional
airflow in the downstream section of the combustor to reach the fuel-lean conditions
at the combustor exit. The RQL-type design has several advantages and challenges,
which are discussed later in this chapter.
As previously described, the challenges at low power are combustion efficiency
and stability. The local fuel-air ratio in the RQL combustor front end at idle is
designed to generate high recirculating gas temperatures (Figure 1.6). Therefore, the
local fuel-air ratio should be near the stoichiometric (f/a ~.068 for jet fuel) fuel-air
ratio to achieve high combustion efficiency. High combustion efficiency minimizes
unburned hydrocarbon and carbon monoxide emissions that predominate at idle.
Some increase in NOx emissions is generated by the hot front end, but emissions at
idle are not significant when compared to high power. By designing for near stoichio-
metric conditions at idle, stability can be ensured at deceleration conditions, where
minimum fuel-air ratio occurs. If the minimum fuel-air ratio during deceleration is
10 Aero GT Combustion
air CO
ing
Mix HC
n hes
bust
io enc
Com ear Qu ction
at n rea
1
=
CO consumed
Gas temperature
Threshold
temperature
Turbine
Compressor inlet
exit
not more than one-third below idle fuel-air ratio, the local fuel-air ratio in the front
end is maintained above the weak extinction limit and flameout is avoided. Limiting
of minimum deceleration fuel-air ratio is accomplished by the engine control and
controls the maximum thrust decay rate for the engine transient.
At high-power conditions, the principal emissions challenges are NOx and
smoke. The RQL combustor axial temperature distribution at high power is depicted
in Figure 1.7. The front end is fuel rich and consequently has lower flame tempera-
tures. The dilution or quench region is characterized by peak gas temperatures as
the fuel-rich mixture transitions through stoichiometric fuel-air ratio to the fuel-lean
conditions at the combustor exit. In the front end, smoke is formed due to the com-
bustion at fuel-rich conditions. Some of the smoke formed in the front end is oxidized
in the high-temperature, oxygen-rich quench region. Thus, the front-end airflow level
must be set with understanding of the formation and oxidation processes. The NOx
emissions are formed in both the front end and quench regions at high power. NOx
formation is exponentially a function of gas temperature, but also depends on the
residence time at the local temperature. The highest rate of formation occurs in the
quench region because it is the region where peak temperatures occur. However,
time at peak temperature in the quench region is relatively short due to high mixing
rates. In contrast, the formation of NOx in the front end is not negligible because it
has relatively longer residence time due to the flow recirculation. The presence of
cooling flow in the front end also leads to NOx formation when it interacts with the
fuel-rich gas mixture.
1.5 Effect of Range of Thrust and Starting Conditions on the Combustor 11
ion NOx
bust
Com = 1 Smoke
at gen
xy
Rich n as o ded
bust
io is ad
c o m
2
~
Rapid NOx
formation
Gas temperature
Threshold
temperature
Compressor Turbine
exit inlet
Recent advances have shown that substantial reductions in residence time and
NOx can be achieved without compromising combustor stability and low-power
performance. Use of fuel injectors that produce small droplets uniformly dispersed
within the airflow and rapid air jet mixing has enabled the residence time reduction.
These advanced RQL combustor designs (Figure 1.8) have demonstrated NOx reduc-
tion of over 50 percent when compared to early annular combustors. They are also
shorter and have lower volumes to reduce residence times. Reduced-length combus-
tors are lighter and also have reduced surface area requiring film cooling. Advanced
cooling schemes have been deployed to minimize NOx emissions and temperature
streaks into the turbines.
Overall, the RQL combustor has demonstrated excellent service history. Because
it does not require complex controls to modulate fuel between injectors, it has dem-
onstrated very good reliability. It also has inherently favorable stoichiometry for
stability because the front-end airflow is minimized for NOx control purposes. The
front-end airflow is established as the minimum amount required for smoke control.
If the fuel-air ratio range between high power and low power is large, the airflow
required to control smoke can be larger than desirable for flame stability during
decelerations. In these instances, the selected minimum transient fuel-air ratio must
be raised to protect flight safety and reliability. In turn, raising the minimum fuel-air
ratio limit increases the time required to decelerate the engine and can result in
a safety risk during emergencies. If the deceleration time cannot be met with the
revised minimum fuel-air ratio, then stability must be addressed by other means, such
as by clustering fuel injectors provided with either more fuel or reduced airflow. This
12 Aero GT Combustion
Figure 1.8. Advanced RQL combustor (Pratt & Whitney PW1500 TALON X).
zone remains above the weak extinction level locally and protects against flameout
at worst-case deceleration conditions.
The critical challenges for the RQL design approach are smoke and liner dura-
bility. As previously discussed, uniform mixing of fuel and airflow in the injectors
can result in reduced smoke levels. When the fuel injector stoichiometry is fuel rich
overall, the uniformity of the fuel-air distribution within the injector becomes crit-
ical. A poorly mixed injector with a wide distribution will have regions that range
from fuel lean to very fuel rich. The latter can produce the bulk of the smoke in
the combustor. This occurs because the highest smoke generation often takes place
in the most fuel-rich regions where there is sufficient residence time. Because the
front end is designed with gas recirculation to achieve stability, these zones can
produce smoke. Thus, the mixing and recirculation patterns are critical to smoke
control.
The presence of fuel-rich and stoichiometric gases also introduces a liner dura-
bility challenge. Because modern gas turbines operate at high temperatures and
pressures, peak gas temperatures can exceed 4200F. Metallic liners have a practical
temperature limit of <2000F for designs that meet typical durability life require-
ments. Therefore, the liner must be cooled to prevent failure. Virtually all aero engine
combustors feature hot side film cooling. Film cooling provides a protective layer of
airflow on the liner surface that prevents convective heat transfer from high temper-
ature gas. However, when fuel-rich gases in the front end interact with cooling, the
film air provides oxidant for high-temperature combustion. Therefore, the presence
of cooling air increases NOx formation in the forward portion of the combustor. In
the aft section of the combustor, cooling does not readily mix radially and there-
fore decreases gas temperatures near the walls. The result is higher temperatures in
1.5 Effect of Range of Thrust and Starting Conditions on the Combustor 13
the midstream. The midstream peaked temperature profile also increases the hot-
test streak temperature exiting the combustor. Therefore, aft cooling airflow affects
the temperature profile and uniformity entering the turbines. Consequently, it is
desirable to reduce cooling throughout the combustor. Improved liner designs have
enhanced heat transfer efficiency, enabled emissions reductions, and strengthened
turbine durability. The evolution of liner cooling designs will be discussed in a later
section.
cruise, climb, and takeoff. Therefore, the local fuel-air ratio in the main stage may
be unfavorable for efficient combustion at the lower temperatures and pressures.
Consequently, more complex staging systems may be required where the main stage
fuel injectors are turned on at different overall fuel-air ratios so that high efficiency
is maintained. These fuel-air ratios are referred to as staging points. Initial designs
were applied to engines with relatively low fuel-air lapse levels. These designs were
operated with two fuel stages and a single staging point. More recent designs applied
to engines with higher fuel-air ratio lapse may require more than a single fuel staging
point to maintain staging efficiency.
Staging can also affect engine acceleration time from idle to higher power con-
ditions. This is because of two factors. The first is the aforementioned combustion
efficiency near the staging point. Lower efficiency results in reduced heat release
and slower acceleration. The second is potential delay time to deliver fuel to the
main fuel injectors. If some of the fuel flow is needed to fill fuel manifolds and fuel
injectors, a delay occurs in the time to achieve combustion heat release and engine
acceleration. Therefore, it is desirable to keep the main stage fuel system as filled as
possible to achieve prompt acceleration when the throttle is moved. However, a full
main stage fuel system is vulnerable to fuel coking. Fuel coking refers to the hard
carbonaceous compounds formed in the internal passages of the fuel system when
the fuel undergoes pyrolysis reactions when it is heated in the absence of air. Such
compounds can block or reduce the flow of fuel through the main stage hardware.
Coking is most common inside the fuel injectors because they are exposed to the
high temperatures inside the diffuser casing. In the extreme, coking can limit thrust
by limiting fuel flow. Most modern engines have idle air temperatures near or above
the level at which significant coking occurs (400F). This air is in contact with the
main stage fuel injectors containing the stagnant fuel. To prevent fuel coking, cool-
ing and insulation features must be incorporated to prevent fuel from contacting
passage walls over the critical temperature for coking. Some designs use the pilot
fuel flow to cool the stagnant main fuel injectors. Other possibilities include using air
pressure to purge the fuel from the most vulnerable areas.
A final challenge to the fuel-staged combustor designer is combustion instability.
Combustion instability refers to temporal fluctuations in the heat release. Such fluc-
tuations can be attributed to several mechanisms, typically involving excitation of
natural fluid mechanic instabilities in the flow or fuel-air ratio oscillations. In the
extreme, instabilities can damage hardware and result in engine damage and failure.
All combustors have risk of instability, but staged lean combustors have been more
prone to them. It is unclear if this tendency is related to differences in acoustic driving
resulting from heat release distribution differences or to changes in acousticdamping
as the combustor is modified for lean-staged operation (Lieuwen and Yang, 2005).
Gas turbine combustors are required to ignite on the ground and in flight. Ignition
in flight is rare because it occurs after unplanned engine shutdown. The combustor
1.5 Effect of Range of Thrust and Starting Conditions on the Combustor 15
should ignite promptly after the fuel is turned on and provide efficient combus-
tion to accelerate the engine to idle power. Delayed ignition can cause excess fuel
accumulation in the combustor and increased pressure pulses at light off. Increased
pressure pulses can result in compressor stall that prevents engine acceleration to
idle. On the ground and at low-speed flight conditions, the engine rotors are turned
with a starter to provide airflow to the combustor for ignition and combustion. At
higher-speed flight conditions, the ram airflow turns the rotor in a process referred
to as windmilling. Ignition energy is typically delivered with a spark igniter. At least
two igniters are placed in the typical annular combustion chamber to provide redun-
dancy in the event of a failure. The spark produces plasma sufficient to initiate the
combustion reaction. The ignited reactants must then be transported to an area
where the reaction can stabilize and propagate to the other fuel injectors in the
combustor. The same features that provide flame stability at idle and deceleration
conditions are relied upon at sub-idle starting operations. The pressure at light off is
usually near the outside ambient pressure because the rotors are not producing sig-
nificant work. However, at higher flight speeds, the total pressure is typically slightly
higher than ambient due to the stagnation effect. Temperatures at ignition are highly
dependent on the thermal state of the engine. For the first start of the day on the
ground, temperatures are usually only slightly higher than ambient. Altitude relight
temperatures are highly dependent on the amount of time the engine has been shut
down. For quick relight attempts less than a minute after shutdown, temperature
at the combustor inlet can be greater than 200F. If the engine is shut down and
windmilling for thirty minutes or longer, the air temperature is closer to the outside
ambient.
Most commercial aircraft must meet requirements for both ground and altitude
starting. The ground starting requirements include a range of ambient temperatures
and airport altitudes. Typical ground starting ambient temperature requirements vary
between 40 and 120F. Airport altitude requirements typically range between sea
level and eight thousand feet. Altitude relight requirements are typically expressed
on a flight envelope (Figure 1.9). There is a high-speed windmilling envelope and a
lower-speed starter assisted envelope. The maximum altitude required for air start-
ing depends on the aircraft. Commercial airliners normally require altitude relight
capability of at least twenty-five thousand to thirty thousand feet. Business jets
often require capability at thirty-five thousand feet because of their higher cruising
altitude.
At the highest altitudes and in extreme cold, combustor ignition conditions can
be very challenging. Pressures of less than five psia and temperatures below 0F are
typical for an engine that windmills until cold. These conditions inhibit the atomiza-
tion of fuel and vaporization of droplets. Low temperature and pressure also slow
the reaction kinetics that promotes stabilization and propagation of flame. Therefore,
design of the combustor should provide for three key features that enable ignition: a
good fuel spray, a favorable airflow velocity, and the proper spark igniter location.
Small fuel droplets are critical to the formation of vapor necessary for igni-
tion. Two types of fuel injectors are typically used: pressure atomizing and airblast
16 Aero GT Combustion
x103ft
40
Successful ignition
No ignition
30
Altitude
20
10
0
0 0.2 0.4 0.6 0.8 1.0 1.2
Mach number
Figure 1.9. Altitude relight envelope (B777 with PW4084 engine).
atomizing (Figure 1.10). The former uses high pressure to push fuel through a small
orifice to generate the spray. The fuel can also be swirled prior to passing through
the orifice to provide angular momentum that produces a spray cone. Airblast atom-
ization uses the energy of the airflow to produce the spray. The fuel is typically deliv-
ered to a cylindrical surface between two swirling airstreams. The cylindrical surface
develops a thin film of fuel as a result of the action of the swirling inner airstream.
As the thin film reaches the tip of the cylinder, the shear between the two airstreams
atomizes the film into a spray. Airblast atomizer performance degrades as the air
pressure drop decreases and should not be used if insufficient airflow is available
to atomize the fuel. This occurs when windmill ignition is attempted at very low
airspeed and when insufficient starter torque limits rotor speed in assisted starts.
Often, airblast fuel injection systems will be supplemented with pressure atomizers
in the locations where the igniters are placed. Such injectors that incorporate both
pressure atomizing and airblast features are referred to as hybrid or duplex injectors.
Increased fuel flow at the igniter locations can help achieve ignition. This additional
nonuniform fuel flow is provided by upstream valves and is usually only present at
low-power settings. Manifolds that deliver the fuel must be designed in such a way
as to achieve the desired distribution of fuel.
Successful ignition also requires a favorable velocity in the region near the spark
plug and the stabilization zone. Because most combustors are swirl stabilized, the
recirculation of flow can be used to transport the ignited spark kernel to the sta-
bilization zone. However, even a properly designed stabilization zone can result
in poor ignition characteristics. This problem can stem from two factors. The first
is local velocities that are too high to sustain the reaction surrounding the spark.
This results in the convective heat loss from the reaction kernel exceeding the heat
1.6 Turbine and Combustor Durability Considerations 17
Air Fuel
Secondary
fuel
released by the reaction, quenching the reaction. This situation is created when there
is insufficient volume and cross-sectional area in the stabilization zone for the quan-
tity of airflow present. Therefore, care must be taken to ensure that local velocity
does not exceed the flame propagation speed at light-off conditions. The other cause
is improper igniter placement. If the igniter is placed in an area with flow direc-
tion away from the recirculation zone, the reaction kernel can be carried out of the
back end of the combustor. Igniters also must be placed in an area where the fuel
spray provides sufficient local fuel-air ratio to achieve ignition. Conditions at igni-
tion are often relatively high in overall fuel-air ratio because of low airflow levels,
but wide variations exist in local fuel-air ratio. As a result, spark igniters are often
placed at the downstream edge of the flow recirculation so that it receives robust
fuel-air mixture from the conical fuel spray, but also provides reverse flow direction
for stabilization.
combustor cooling does reduce the amount of airflow available to control emissions
and mix out temperature streaks. Therefore, it is desirable to minimize the amount
of cooling flow used.
As previously mentioned, the combustor exit temperature distribution has a
large impact on the amount of turbine cooling required and, thus, the engine per-
formance. Combustor exit temperature quality is normally described in terms of
the radial average temperature profile and the hottest streak intensity. These are
referred to as radial profile factor and pattern factor, respectively. They are typically
described as nondimensional parameters:
where Tra is the average temperature at a given radial position, Ti is inlet temperature,
and Te is the mass averaged exit temperature. The pattern factor is given by:
where Tstr is the maximum temperature anywhere in the combustor exit annulus,
commonly called the streak temperature. The radial profile factor is determined as a
function of radial position at the entrance to the turbine. The maximum pattern fac-
tor occurs at only one spatial position in the combustor exhaust (Figure 1.11). Often
this location is the result of random hardware variation and cannot be assumed
repeatable from engine to engine. However, the radial distribution of pattern factor
is also of interest to the turbine designer. Turbine static hardware (vanes and outer air
seals) are impacted by the local gas temperatures while rotating blades are impacted
predominantly by the radial average temperature profile because they rotate too fast
for metal temperatures to respond to local effects. Thus, the static hardware cooling
level is often set to protect against the hottest pattern factor streak, even though it
occurs in only one place. It is useful to know the radial distribution of pattern factor
so that static hardware cooling can be distributed more in the core region, where the
hottest streak is likely to occur, and less near the walls, where the average tempera-
tures are lower because of the effects of combustor cooling.
To achieve the target radial temperature profile and low pattern factor, the
designer must control the mixing of fuel and air in the combustor. To achieve the low-
est possible pattern factor, the designer would premix all of the airflow with the fuel
at the combustor front end. This would produce a flat, uniform temperature profile at
the combustor exit. However, considering liner cooling requirements, operability, and
radial profile requirements, this approach is not practical for aero engines. In practical
aero engine designs, cooling air is injected into the combustor in a way such that it
provides a protective film near the combustor walls. As such, it generally does not mix
readily with the other airstreams and the fuel in the combustor. Liner cooling is there-
fore not effective at controlling pattern factor, but can be effective at providing cooler
radial average profile near the inner and outer walls. The airflow not used for liner
cooling is used to control radial profile shape and pattern factor. In RQL designs, the
bulk of the non-cooling airflow enters through air jets downstream of the front end
in the liner walls. In fuel-staged designs, most of the airflow is incorporated into the
1.6 Turbine and Combustor Durability Considerations 19
0 10 20 30 40 50 60 70 80 90 100
Radial span (%)
Figure 1.11. Combustor exit profile and pattern factor.
fuel injection swirlers at the front end of the combustor. Therefore, the mixing pro-
cesses to achieve uniform exhaust temperatures are quite different. In swirling flow
mixers, multiple airstreams are often used to create shear layers that promote mix-
ing. Fuel is injected into the airstreams so that it is dispersed and mixed with the air.
Fuel injection is usually accomplished with jets, thin films, or pressure sprays. Swirling
airstreams may be co-swirling or counter-swirling. Designers have successfully used
both approaches. Counter-swirling airstreams produce the highest mixing rates, but
result in low net swirl if not designed with unequal flow quantities. These principles
are applied to both RQL and lean-staged designs because good fuel injector and
swirler mixing is required for both design approaches.
Jet mixing is dependent on the arrangement, size, and upstream conditions influ-
encing the jets. The penetration depth of a jet, Y, is proportional to the jet diameter,
dj, and the square root of the momentum ratio, J:
Y ~ dj J
J = (j U j2 )/(g Ug 2 )
Uj and Ug denote the jet and cross flow velocities, respectively. Specific correlations
depend on the geometry of the duct and the arrangement of the jets (Lefebvre and
Ballal, 2010).
The penetration of the jet can be controlled by the sizing, pressure loss, and
upstream flow quantity. Efficient mixing of upstream gases also requires jet spacing
dense enough to mix within the combustor length allocated. Therefore, the arrange-
ment and size of the jets are critical to the spatial delivery of airflow to the critical
regions where it is needed to mix temperature streaks and provide target radial aver-
age temperatures.
Staged lean combustors may only use air jets to control radial profile shape,
because the lean well-mixed front end delivers good pattern factor at high-power
20 Aero GT Combustion
conditions. For example, a row of smaller holes in the aft end of the combustor can
effectively cool the inner portion of the radial distribution because of their limited
penetration. At low-power conditions, staged lean combustors often have worse uni-
formity because of the reduced number of fuel injectors operating. RQL combustors
are more dependent on the jet mixing to deliver both the radial profile and pattern
factor targets. Researchers have conducted numerous experimental studies to deter-
mine the optimum jet arrangement for mixing flow in a duct. Holdeman found that
the optimum arrangement was given by:
H = ( p / h) J
where p is the hole pitch, h is the duct height, and H is the characteristic parameter
(Holdeman, 1993). For unopposed holes in a duct, the optimum value for H is 5
(Figure 1.12). This finding was for a uniform axial cross flow, but is often a good
arrangement from which to begin optimization. For real combustors where the
upstream flow is swirling, computational fluid dynamics analysis is useful to refine
the distribution. Rig testing is required to determine the maximum pattern factor
because CFD calculations typically cover a single fuel injector sector and thus do
not provide random effects.
Combustion imposes two different types of heat loads on the liner. The first is
radiation from the flame to the surfaces. The second is the convective effect of hot
gases contacting the liner cooling films. The convective load can result in film tem-
perature above metal temperature in some areas of the combustor. The radiation
flux is given by:
q r = 0.5(1+ w )( g Tg 4 g Tw 4 )
where w is the wall emissivity, g is the gaseous emissivity, g is the gas absorptiv-
ity at the wall temperature Tw, and Tg is the radiating gas temperature. The gaseous
emissivity is dependent on the flame luminosity and gas temperature, which in turn
depends on the combustion stoichiometry. Rich combustion tends to produce highly
luminous soot. Therefore, the forward section of a RQL combustor tends to have
more radiant heat load than a staged lean combustor, which produces very little
soot. The midsection of a RQL combustor produces peak gas temperatures as the
stoichiometry transitions from fuel rich to fuel lean. This region of a RQL also has
higher gaseous emissivity than a lean-staged combustor does.
1.6 Turbine and Combustor Durability Considerations 21
Film
Louver lips
Floatwall liner
Film
Convective heat load on the liner is dependent on the local gas temperature and
velocity and its interaction with the cooling film. The effectiveness of the film is crit-
ical to maintenance of acceptable metal temperatures because the film temperature
is a key driver in the heat flux:
q c = h(Tfilm Tmetal )
where qc is the convective flux, h is the convective heat transfer coefficient, Tfilm is
the film temperature, and Tmetal is the liner surface metal temperature. Film tempera-
ture is dependent on the local gas temperature and film effectiveness. Film effec-
tiveness depends on the nature of the film (slot flow, discrete holes, etc.) and the
ratio of the cooling momentum to the mainstream flow momentum. The momentum
ratio is referred to as the blowing parameter. In zones where the mainstream flow
momentum is low and the blowing parameter is high, the cooling film effectiveness
is reduced. This occurs most commonly in the front end of the combustor. Note that
for equivalent film effectiveness, the RQL design will have higher film temperature
because it reaches higher peak gas temperature than the staged lean combustor.
Cooling strategies for the outside of the liner wall vary widely with design
(Figure1.13). Outside cooling is important because it balances the hot side heat flux.
With high backside cooling effectiveness, higher hot side heat flux can be tolerated.
Initial liners made use of simple louvers that created slot films on the liner hot side
and had minimal heat transfer on the outside of the liner wall. The louver length was
determined by the distance that film effectiveness could be maintained. Louvered
designs later evolved to incorporate more effective backside cooling strategies for
the louver lips. All continuous ring louver liners fail because of thermal fatigue
cracks. The cracks result from high thermal stresses on the full ring hoop. As oper-
ating temperatures increased, more effective liner designs were required because
22 Aero GT Combustion
cracking was an even more severe problem. This led to tiled liner (floatwall) combus-
tors that have a cold structure that carries mechanically attached panel tiles. The cold
carrying structure virtually eliminates the hoop stresses that caused fatigue cracks
in prior designs. The tiles have multiple fin structures to augment convective heat
transfer incorporated on the backside. Film cooling air is first passed through these
fins to provide high effective heat transfer levels. More recent designs have relied on
film cooling holes to increase liner cold side heat transfer rates.
RQL design typically requires a higher cooling flux than a staged lean-type
design. However, this difference is only significant if it prevents the achievement of
other objectives. Cooling air generally stays near the liner walls of the combustor. It
causes the exit temperature profile to be cooler near the walls and more peaked at the
mid-span of the exit plane. This is generally desirable for turbine durability because
it reduces heat loading on the turbine platforms and seals. Using modern liner cool-
ing technologies, the cooling flow allocation has not limited the achievement of pro-
file and pattern factor targets in either type of combustor design. However, cooling
can have a significant impact on emissions and combustion stability.
Cooling affects emissions most significantly at low power when the inlet air
temperature is lowest. At low temperatures and fuel-air ratio, liner film cooling can
quench the near-wall combustion process, resulting in the generation of unburned
hydrocarbons and carbon monoxide. This occurs predominantly in the front end
of the combustor, where swirling and recirculating flow contacts liner film cooling.
These effects occur in both RQL and lean-staged designs. Such quenching can result
in disruption of flame stability in the extreme if the heat released is not sufficient
to sustain continuous ignition. At high power, cooling can result in formation of
NOx emissions in the front end of RQL combustors. The NOx forms when fuel-rich
front-end gases contact film cooling. The region where the contact occurs produces
stoichiometric combustion temperature and the highest NOx formation rates. In
lean-staged combustors, the cooling air does not increase NOx because the fuel-air
mixture is already leaner than stoichiometric and cooling causes a reduction in com-
bustion temperature.
Future aircraft engine cycles will require improved thermal and propulsive effi-
ciency to meet aggressive fuel burn goals and address CO2 emission concerns. Current
cycles have sufficient differential between the coolant and target metal tempera-
tures to allow effective cooling within allowable flow budget. As cycle temperatures
increase, improved liner cooling technology or increased temperature materials will
be required to maintain low enough cooling fluxes to meet all combustor metrics.
1.7Summary
Gas turbine combustors remain an interesting and complex design challenge.
Balancing the many requirements in an environment demanding low cost, low
weight, low emissions, and excellent safety and reliability is often difficult. Many of
the subjects introduced in this chapter will be discussed in more depth in subsequent
chapters of this book to provide a better understanding of these issues.
References 23
References
Golley, J., Whittle, F., and Gunston, B. (1987). Whittle: The True Story, Smithsonian Institution
Press, Washington, DC.
Holdeman, J. D. (1993). Mixing of Multiple Jets with a Confined Subsonic Crossflow. Progress
in Energy and Combustion Science 19 no. 1: 3170.
Lefebvre, A. H., and Ballal, D. R. (2010). Gas Turbine Combustion: Alternative Fuels and
Emissions, CRC Press, Boca Raton FL.
Lieuwen,T., and Yang,V. (2005).Combustion Instabilities in Gas Turbine Engines: Operational
Experience, Fundamental Mechanisms and Modeling. Progress in Astronautics and
Aeronautics 210 American Institute of Aeronautics and Astronautics.
2 Ground-Based Gas Turbine Combustion:
Metrics, Constraints, and System
Interactions
Vincent McDonell and Manfred Klein
2.1 Introduction
A serious need for future energy resources worldwide is apparent, driven by
high-population countries such as India and China that are rapidly developing infra-
structure for energy, as well as growth or repowering in developed countries. Gas
turbines play a preeminent role in the stationary power generation marketplace and
should remain a critical part of the market mix for the foreseeable future, despite
competition from reciprocating engines and newer technologies such as fuel cells.
Alternative technologies compete with gas turbines in certain size classes, but at
power generation levels above 5 MW, gas turbines offer the most attractive option
because of their relatively low capital, operating, and maintenance costs. Hence,
these engines are increasingly relied upon for clean power production from a variety
of fuels. The configurations for these systems involve high efficiencies as well. As a
result, the market will continue to demand gas turbines.
Chapter 1 discusses the drivers and consideration for aero gas turbines. While
much of that discussion applies to gas turbines in general, the use of gas turbines
for ground-based applications gives rise to additional and/or different metrics, con-
straints, and much wider possible overall system interactions relative to the com-
bustion system. These turbines vary in size from 10 s of kW to hundreds of MW.
The applications vary from power generation to mechanical work (e.g., Soares,
2008). In power generation, the gas turbine shaft is coupled to a generator either
directly or via a gearbox (direct drive). For mechanical work, the gas turbine
provides power to a mechanical device such as a compressor or pump (mechan-
ical drive). In power generation, the gas turbine may often be combined with
other equipment to form combined cycle systems (e.g., combining a gas turbine
generator with infrastructure to collect exhaust heat to produce steam to drive a
steam turbine).
This chapter introduces key aspects associated with ground-based gas turbine
systems and starts by summarizing the key differentiators between ground-based
and aero engines. Then, it discusses interactions between different systems and com-
ponents, starting with the overall electric grid, then moving to the power plant, then
24
2.2 Key Differentiators between Aero and Ground-based Gas Turbines 25
to the engine and combustor itself. Specifically, combustor and engine interaction
with the electrical grid is a unique aspect of ground-based power plants and has
substantial impacts on the operation of the gas turbine.
Next, a section on plant-level requirements is provided that illustrates the
trade-offs that come into consideration when including the gas turbine in different
types of cycles and applications as one part of an overall plant. For example, devel-
opments for advanced central plant designs couple the gas turbine with an integrated
gasifier combined cycle (IGCC) in configurations that may also be designed to iso-
late CO2. In these cases, the performance of the gas turbine is just one of a number
of considerations in terms of the overall operation of the plant.
The basic trade-offs associated with the engine operation and the combustion
system are then discussed. In some power generation applications, ability to follow
load is important. In this case, the gas turbine may be required to operate over a
range of output power, necessitating consideration for optimum operation at inter-
mediate loads. In many smaller power generation applications, local regulations may
prohibit the export of power back to the grid. In this case, if the electricity con-
sumption on the customer side of the meter is less than what the gas turbine pro-
duces, the output of the turbine must be reduced. In other applications, the primary
operating strategy requires as close to continuous 24/7 operation as possible. This
would be the case for central station power generation where full-power operation
generally results in the highest overall efficiency. In other applications, gas turbines
may be used strategically for tuning the power available to meet local peak use.
Such peaking plants rely on gas turbines because of their ability to rapidly start
up. In these cases, it might be common for the gas turbine to undergo hundreds of
start/stop cycles in a given year. In this case, parallels to aero engines are obvious
as the duty cycle appears more like a propulsion gas turbine. Not surprising, many
of the successful peaking-type turbines are derived from aero engine gas turbines
(aeroderivatives) because they are designed to quickly start and stop and yet still
achieve relatively high thermodynamic efficiency. As a result, consideration must be
given to these different operational requirements in the development and optimiza-
tion of the gas turbine system.
Finally, general combustion system constraints and design architectures are
introduced along with consideration for various types of fuels.
2.2.1 Emissions
Soot
HC (VOC, ROG, NMHC)
CO
NOx (NO, NO2)
SOx (SO2, SO3, sulfates)
CO2
H2O
An issue
High Priority
these cycles relative to combustion are discussed further in Section 2.5.3. As a result,
the heat transfer, reaction kinetics, pollutant chemistry, and materials issues are
dramatically altered and must be carefully accounted for. Regardless of these chal-
lenges, commercial developments involving all of these cycle enhancements have
been successfully implemented (e.g., Alstoms GT-24 and GT-26 reheat sequential
combustionsystem, GE LMS-100 intercooled industrial engine; Solar Mercury 50
recuperated engine).
Ground-based systems also allow inclusion of water addition (e.g., inlet fog-
ging and/or humid air turbine (HAT) cycles) as a means to achieve higher overall
efficiencies or power augmentation (e.g., Kharchenko, 1998; Kavanagh and Parks,
2009a, 2009b). In these systems, high water content in the airstream leads to addi-
tional considerations in terms of the combustion system as the added diluents will
play a role in the reaction kinetics and the subsequent pollutant chemistry. Hitachi
and Pratt & Whitney have developed examples of implementation of cycles involv-
ing water. Pratts effort involved the adaptation of an FT4000 turbine (EPRI, 1993
and ongoing work with the U.S. Department of Energy). Hitachi has carried out sev-
eral demonstration projects, including a 4 MW pilot plant (e.g., Higuchi et al., 2008;
Araki et al., 2012). The role of water in the NOx chemistry was particularly noted in
these examples. More discussion on these cycles and how they impact the combus-
tion conditions is provided in Section 2.5.3 along with discussion relative to systems
consideration involving water in Section 2.4.9.
Further, while aero engines have benefited from operation on essentially a single
fuel, ground-based turbines have to handle a multitude of fuel types and composi-
tions. While developments since the 1970s have focused on natural gas operation and
have resulted in tremendous growth for gas turbine power plants in the following
twenty to thirty years, attention continues to be directed at fuel flexibility to accom-
modate market and political forces. The desire to take advantage of carbon mitiga-
tion through use of renewable fuels, such as those from landfills and waste water
treatment, has provided an interesting opportunity for gas turbines. On the larger
scale, the development of gas turbines for operation on high-hydrogen-content fuels
28 Ground-Based Gas Turbine Combustion
has been driven by the desire to operate on gases produced from gasification of
a myriad of feedstocks. In the 2000s, development of turbine combustion technol-
ogy for high-hydrogen-content fuels associated with gasification of coal resulted in
considerable progress in terms of reliable, low-emissions gas turbine operation on
hydrogen. The implications of operation on these fuels are discussed throughout and
a summary of the fuels of interest is provided in Section 2.8.
with momentary torque on the generator that is several times larger than the normal
full-load torque. This load drop situation is also challenging for combustor stability.
In light of increasing amounts of renewable energy systems, namely solar and
wind, the need for dispatchable power is increasing substantially. Because of the
intermittent nature of solar and wind-derived power, operators need the ability to
back up this resource and stabilize the grid. As a result of this requirement, the role
of gas turbines and other high-efficiency dispatchable power generation devices may
increase substantially in the near future.
In some circumstances, small gas turbines may give a region the option to
island, allowing the region to eliminate power draw from the grid altogether. In
the case of remote power applications or with the concept of microgrids, the role
of power generation becomes more local in nature. However, it is imperative that
considerations for the larger-scale grid be given to prevent safety issues. It is impor-
tant to retain the capability to isolate loads fed from the generator from those fed
from the grid. This has led to extensive discussions and debate with utility providers
raising concerns about safety and loss of grid control and small generator adopters
faced with potential feed in or standby charges to allow the grid to back up the
local generation system. This debate is highly complex and regionally dependent.
Generally speaking, end users are not anxious to take on this complexity, resulting
in opportunities for third-party businesses to take this burden from the end users
through mechanisms such as a power purchase agreement (PPA). In this case, the
third party essentially contracts with the end users to pay for power used and then
deals with securing fuel prices, equipment procurement, interconnection agreements,
permits, rate negotiation, and the like. Often, incentives such as tax credits, rebates,
or renewable or GHG credits can play a major role in defining the economic viabil-
ity of such a project.
The notion of smarter grids that can adapt to increased intermittent power due
to increased use of renewables or local power generation, whether utility owned,
customer owned and operated, or third party owned and operated, is of tremendous
current interest. It is evident that utilities want to engage in this new paradigm of
energy and are working with regulators and agencies to evolve the nature of the grid
to mitigate potential issues associated with these scenarios.
water and energy security considerations. Pollution prevention and energy con-
servation with system efficiency are key to arriving at solutions that address both
economic opportunities and environmental sustainability for international clean
energy implementation. The greenhouse gas and climate change debates are creat-
ing the need for a fresh look at the issues. Establishing environmental best available
technology (BAT) management practices and technologies for a sector also requires
a balanced evaluation of several interrelated environmental issues. Gas turbine
energy facilities, as well as other systems, must address a combination of air, noise,
water, and land issues to be appropriately permitted. Key decision-making consid-
erations include:
What are the overall environmental objectives for the main stakeholders?
What does the word best refer to one issue or several?
What are the metrics against which performance can be judged?
Is the assessment or permitting aimed at individual equipment or at the entire
facility?
Are requirements based on performance standards or on prescriptive technol-
ogy choices?
How are overall system efficiency, safety, and reliability factored into
decisions?
For thermal energy systems, the objective may be to address as many of the
critical environmental impacts as feasible in a balanced and comprehensive man-
ner. Air emissions of criteria pollutants, CO2, and air toxics almost always happen
simultaneously, within the same system and from the same fuel choices. When GHG
emissions are prevented, one generally finds that other air pollution factors concur-
rently improve as well, thus adding value to the reduction in GHGs. Moreover, NOx
and CO2 emissions often increase in opposite directions, with the high pressures and
temperatures needed for efficiency and power creating more thermal NOx. For cer-
tain types of NOx control, important collateral impacts may arise, resulting in a need
to balance various constraints depending on the applications in which gas turbines
are used.
The rapid response of aeroderivative engines makes them good choices for peaking
applications. However, such compact high-pressure combustors may have difficult
challenges in lean premix dry low NOx DLN design, as space is limited for imple-
menting air and fuel control strategies. Large gas turbines have lower NOx emis-
sion rates in large DLN combustors for combined cycle plants versus slightly higher
NOx rates from annular combustors in small engines. However, these aeroderivative
gas turbines can often be allowed to have a higher NOx level, as they are usually
more efficient in cogeneration applications, being able to use most of their exhaust
heat output with a high heat-power ratio and with a lower GHG profile. Among
the many specific trade-offs in DLN design are the difficulties with more efficient
2.4 Plant-Level Requirements, Metrics, and Trade-offs 31
Natural
Air gas
Heat
recovery
Generator steam
Hot generator
exhaust
Gas turbine
Gas turbine combined cycles (GTCC) are a combination of gas turbine prime
mover (making about two-thirds of total power) with an associated steam turbine
(making one-third of total power), together producing electrical and/or mechanical
power more efficiently than if they were independent. Figure 2.1 illustrates a typical
plant layout and components. However, large condensing GTCC plants in green-
field locations may not be as environmentally sound as balanced CHP because
they tend to reject large amounts of otherwise usable low-grade heat to water or air
condensers. The operation of many of these plants is impacted by economics associ-
ated with electricity rate structures combined with often volatile fuel pricing, under-
scoring the complex relation between technology, policy, and market forces. Low
operating hours often lead to higher electricity prices as utilities strive to recover the
costs of these otherwise stranded assets.
As alluded to previously, for large combined cycle systems, the required steam
condensers are an environmental problem for several reasons, including large energy
losses, thermal pollution of local water bodies, vapor plumes, and noise impacts.
About half of all thermal energy is rejected to the environment through stacks, con-
densers, and off-gases. Large gas turbine systems can be built very quickly, and con-
sume large amounts of fuel from the natural gas delivery infrastructure. The remote
siting of a large number of combined cycles will lead to more power transmission
lines, condenser energy losses, and gas fuel supply and pricing uncertainty.
32 Ground-Based Gas Turbine Combustion
Environmental policy that promotes ultralow NOx limits often tends to favor
large combined cycles owing to economy of scale, and this has resulted in impres-
sive combustion technology advancements by the major frame engine OEMs. At the
same time, the relatively smaller near-term economic benefit for low-emission sys-
tem development for smaller systems makes implementation of more efficient, dis-
tributed systems more difficult. Moreover, large GTCC units, despite their advanced
low-emissions combustion systems, must sometimes also have backend selective cat-
alytic reduction (SCR) systems to meet ultralow NOx levels, resulting in an increase
in the systems fine particulate emissions, N2O greenhouse gases, ammonia slip,
and an efficiency drop in the heat recovery system, as discussed in Section 2.4.8.2.
These potential effects, coupled with plant cycling and marginal emission reductions,
requires care be taken when applying such technology.
Utility repowering of aging coal facilities and solid fuel gasification are the most
important opportunities for large gas turbine systems. Much of the worldwide
installed utility boiler and steam turbine capacity was constructed before 1975, and
many of these units are in need of overhaul, upgrading, or retirement over the next
decades. The majority of the electricity industrys emissions of GHGs and other pol-
lutants is sourced from these coal- and oil-fired plants, using boilers and condens-
ing steam turbines. This opportunity invites a significant energy choice in reducing
worldwide air pollution, toxics, and greenhouse emissions from existing thermal
power infrastructures (Figure 2.2).
Advances in emissions reduction and efficiency increases for large gas turbine
combined cycles make a strong case for using natural gas for some utility generation
by adding gas turbines with heat recovery steam generators (HRSGs) to replace old
boilers at the site, keeping the existing steam turbines. A variety of repowering tech-
niques are available that could partially or fully integrate gas fuel on sites, keeping
most of the existing steam system and auxiliaries. These intensive modifications may
involve the conversion of the plant to a high-efficiency GT combined cycle to retire
an older unit and greatly decrease overall emissions. Coal gasification offers another
potential form of combined cycle repowering, discussed in Section 2.4.6, in which
case other combustion system considerations are required.
Steam turbine
Steam
LP IP
Existing
steam
turbine
Cond.
Gas turbine
HRSG
Gas fuel
Figure 2.2 Repowering utility boiler with a GT/HRSG system.
Fuel
Power
Heat Heating
MWe generation
recovery load
technology
Absorption Cooling
chiller load
E 35 Energy output
(E) E (45)
65 Energy output (E,H)
Auxiliary losses
15 Auxiliary losses
48 24 10
E (33)
14 Auxiliary losses
Stack loss
Figure 2.4. Cycle efficiency trade-offs.
District energy systems (DES) use captured heat from industrial processes to distrib-
ute steam and hot or cold water to individual buildings through a network of pipes.
District energy systems are good candidates for CHP because they are installed near
electrical loads, they provide demand for low-grade thermal energy, and they usually
burn clean fuels that can be located close to the demand. Low-temperature thermal
2.4 Plant-Level Requirements, Metrics, and Trade-offs 35
output increases the efficiency of CHP systems, and district energy system load peaks
can coincide with peak electricity demand.
Here, CHP systems are encouraged to be a centerpiece of a utility island, to
use various types of feedstocks and waste products to supply energy services such as
electricity, steam, hot water, and/or cold water in an eco-industrial park concept.
An improved choice range of reliable and cost-effective small gas turbine units (with
moderate DLN NOx levels) would be valuable in creating additional distributed
CHP and DES applications for buildings. On larger scales, paper mill operations
and oilsands facilities can take advantage of these concepts, turning waste streams
(petcoke, bitumen, woodwastes) into energy services (Figure 2.9).
Hence, like the CHP applications, many opportunities for alternative fuels
are found with the subsequent combustion system impacts associated with fuel
interchangeability.
The term integrated gasification combined cycle (IGCC) is used to describe a type
of clean coal technology facility that reduces most of the air emissions produced
from coal-fired electricity generation. Gasification technology reforms solid fuels
such as coal, bitumen, or petroleum coke into a synthetic fuel gas for much cleaner
burning. The solid fuel is fed into the gasifier then subjected to high temperatures
and pressure and to low levels of oxygen to create syngas by partially burning the
coal. The syngas is a mixture of hydrogen and carbon monoxide that has about
one-quarter of the energy value of natural gas. The syngas is then burned in an
advanced efficient gas and steam turbine combined cycle. A water-shift chemical
reaction will allow the CO2 separation at a reasonable pressure, with carbon going
to delivery, and hydrogen to process or combustion. Figure 2.5 illustrates the
process.
As in the cases examined before, IGCC can impact combustion systems because
of the need for operation on high-hydrogen-content fuel streams. Hydrogen fuel
combustion presents many challenges due to flashback (Richards et al., 2001;
Lieuwen et al., 2008; Lieuwen et al., 2010). Additional combustion research is
needed to verify the technical potential and reliability of these systems for various
input feedstocks. Gasification could be very important as a comprehensive solu-
tion for solid fuels and in the power, oilsands, and refinery sectors. Acceptable NOx
emission levels must be set high enough to allow safe operation on hydrogen-rich
fuels for an effective carbon capture system. Gas turbine NOx emissions reduction
for hydrogen or syngas combustion will be a combination of nitrogen injection and
dilution, possibly steam injection. System reliability is critical if great expense will
be provided for high-pressure CO2 capture through water shift of syngas to hydro-
gen. Specific challenges for syngas combustion include (Lieuwen et al., 2010):
Air
Gas
Water
CO Gas turbine
Gasifier shift MW
H2 cleanup combined
reaction H2 cycle
Coal
Petcoke
Biomass Slag Sulphur N2 injection, dilution
When combined with coproduction of chemicals and CO2 capture, delivery, and
storage, gasification systems will improve energy efficiency to reduce air pollution
and greenhouse gas emissions. Pipelines for delivery of CO2 can be developed in
conjunction with a regional energy plan in areas where coal and oilsands activities
are prevalent.
The gas pipeline industry (and offshore platform applications) is also a large user
of gas turbines, especially on large-diameter, high-flow systems that use centrifugal
gas compressors. A balanced mixture of aero-derived and industrial frame engines
has been employed, although the former type is gaining popularity because of ease
of removal and maintenance. DLN combustion developments have comprised most
of the air pollution solutions, with water injection for NOx prevention sometimes
employed. Station or unit upsets can result from unreliable combustion in gas tur-
bines, causing problematic shutdowns with stops and starts, increased methane emis-
sions, station blowdowns, and venting noise. Safety on the offshore platforms cannot
be compromised. Many best operating practices are needed to minimize the overall
GHG profile of these energy systems and to maintain a high system reliability to
prevent mishaps.
For the high-pressure gas pipeline industry, methane leakage has been an
important full fuel cycle consideration. Methane emission, with a global warming
potential (GWP) at twenty-one times that of CO2, is an important consideration
for the pipeline industry. Although quantities emitted are small, averaging about
one-half percent for gas production with a 1,000-kilometer-long transmission sys-
tem, the high GWP poses issues. A common conclusion is that most fugitive CH4
2.4 Plant-Level Requirements, Metrics, and Trade-offs 37
6 MW
23 MW OTSG
HP
steam 20 MW
steam
7 MW turbine
26 MW OTSG
7 MW
26 MW OTSG
Air cooled
condenser
Pipeline
Compressor
Waste heat recovery
Figure 2.6. Pipeline compression with heat recovery.
emissions can be reduced by finding the top two to three leakage sources in a sta-
tion, often associated with unseated relief valve vents or unnecessary station blow-
downs. As a result, reliability of compressor stations is very important, implying a
need for robust combustion systems (Moore et al., 2009). Methane measurement,
quantification, component counts, and inventories are also now becoming impor-
tant tasks.
Another challenge for these applications is limited need for heat. As a result,
gas turbine electrical efficiency is of paramount importance. Thus, combustion sys-
tems may be impacted by need for higher firing temperatures or impacts associ-
ated with cycle enhancements. In addition, bottoming cycles such as steam turbines
or organic Rankine cycle (ORC) technologies can be used for waste heat recovery
(Figure 2.6).
Air pollutants affect the health of local populations, wildlife, and ecosystems through
contributions to smog, acid rain, and some toxicity residues. In severely impacted
regions, pollutant mitigation is often a very strong driver of combustion technology
advancement.
(a) 2.5
SO2 NOx PM
kg/MWhr 2
1.5
0.5
0
Coal Oil Gas GTCC GTCHP Bio IGCC
(b) 1000
Carbon Dioxide
750
kg/MWhr
500
250
0
Coal Oil Gas GTCC GTCHP Bio IGCC
Figure 2.7. Comparison of emissions from various power plant configurations.
System characteristics and technical choices for various cleaner energy sectors
will determine how the balancing of low-criteria air pollutants, greenhouse gases,
and air toxics can be optimized for compression, combustion, turbine output, and
heat recovery in gas turbine energy systems. A cost-effectiveness evaluation can be
important to evaluate alternative choices and trade-offs when several variables are
considered in the same analysis. When this is done, BACT concepts such as cogen-
eration, district energy, gasification, methane leakage prevention, absorption chill-
ing, and waste fuels will each become as important as the various NOx reduction
technologies for gas turbines. To avoid confusion, it is useful to ensure that when
emissions are discussed, it is always made clear as to which type GHGs, air pol-
lution, CFCs, or trace toxics is being considered in the situation. Care should also
be taken to consider interrelated effects on local noise, vapor plumes, water, and
land use impacts.
Ammonia slip from unreacted NH3 (less reaction during plant cycling
operation);
40 Ground-Based Gas Turbine Combustion
Aero-Derived Systems
- high pressure ratios (1535)
- small physical space, more complexity
- improving reliability, but challenges remain
Water impacts can be more important to local stakeholders than air emissions.
Condenser losses from large Rankine steam systems affect water quality in lakes
or rivers, and sometimes associated vapor plumes are a visibility issue. Condensing
steam turbines have less of an impact because of their smaller size relative to the
primary Brayton Cycle gas turbine units. These condensers may be water-cooled
surface units, air-cooled fan condensers, or a hybrid system with both moisture and
air. Water surface condensers using river or lake water can be once-through units
discharging into local water bodies, often limited to a 812C temperature rise at
the return outlet to protect the aquatic ecosystem. High temperatures are lethal to
fish and other aquatic life, and even modest temperature increases can harm growth
42 Ground-Based Gas Turbine Combustion
Sewage, Waste
Plastics, Rubber
Treatment
Waste Heat
L.P. Steam Woodwaste
Hot Water
Flyash, Gypsum
MSW, Tires Sulphur, Pet.Coke
5Rs
Cogeneration &
Waste Heat District Energy By-Product
Recovery Synergies
Sustainability
Natural Gas GHGs, CACs, Toxics, Water Gas
& Hydrogen Energy Security & Conservation Turbines
Energy Supply and Pricing
and reproduction of sensitive organisms, stimulate the growth of algae, and decrease
levels of dissolved oxygen, which are also harmful to aquatic life forms.
Sometimes evaporative cooling towers are used for the heat sink (rather than
river water) for very large boiler steam plants. These transfer process waste heat
to the ambient air, often with a large vapor plume. An air-cooled fan condenser is
another form of condensing system used frequently in gas turbine plants and where
water is in short supply. While less water impact results, the cycle efficiency is reduced
because of power consumption from the fans. Wet surface air condensers are also
common when surface area is restricted or when summer fan operation is not suffi-
cient to cool the steam. Nuisance impacts are related to low-speed fan noise (can be
cancelled with acoustics) or thermal plumes in winter from the water spray.
The major impact of condensing, however, is extensive energy losses of low-grade
heat. Large combined cycles depend upon capturing all of the energy value as
high-quality electricity, but, in doing so, the lower-grade heat is rejected to the envi-
ronment. The steam turbine requires this condenser to make steam back into boiler
2.5 Engine-Level Metrics and Trade-offs 43
feedwater. Thus, a 500 MWe GTCC plant (at 50 percent efficiency, using 1000 MWth
of fuel input) can produce 450 MW thermal of rejected heat, of which 300400 MW
thermal goes through the condenser. This energy could be used in a district energy
application to provide winter heat (or summer cooling) to many large buildings in a
nearby downtown city area.
As mentioned in Section 2.4.8.2, water can also be injected for NOx control in
simple cycle applications such as peaking duty, or, more recently, pipeline compres-
sion. Clean water is essential in this application to prevent any impurities or solids
formation inside the engine. Significant operating costs are associated with water
transportation, treatment, and disposal; modified combustor/turbine and control sys-
tem components; increased engine maintenance; and fuel penalty. High costs can be
incurred in isolated areas where the water acquisition and treatment costs are signif-
icant. Steam injection also requires extensive water treatment. During certain ambi-
ent air temperature conditions, the gas turbine engine can be fooled into thinking
that the air is at an ISO condition, such that more mass flow or colder/warmer air is
fed into the gas turbine inlet. Inlet fogging with vaporized water droplets for cooler
air can be effective in enhancing performance. A water droplet stream can also be
injected in the air filter and compressor inlet to provide both the cooling effect and
a wet compression cycle with more mass flow and power. Some of these system
requirements are discussed in Section 2.5.3.
2.5.1 Turndown
Turndown refers to the ability of the engine to operate over a range of conditions
while maintaining reasonable performance levels. Generally speaking, the overall
fuel-air ratio drops as load is reduced. To illustrate this, Figure 2.10 presents the rela-
tive fuel and airflows and the associated fuel-air ratio for a small gas turbine engine.
As shown, the overall fuel-air ratio drops as load decreases. As a result, if the system
is optimized to operate at minimum fuel-air ratio at full load (e.g., to minimize reac-
tion temperatures), it would not be possible to reduce the load of the system, as it
would blow off due to reaching the lean blowoff limit. In response to this problem,
lean premixed combustion systems are often staged, with multiple fuel injection
points that can be operated sequentially in parallel to allow tailoring of the local
fuel-air ratio for each point, while allowing the overall fuel-air ratio for the engine to
vary as needed to accomplish the turndown desired.
The key is to ensure local fuel-air ratios are established to allow the engine
to maintain operability. To illustrate this in the context of emissions, Figure 2.11
44 Ground-Based Gas Turbine Combustion
1400 140
Air
1200 Fuel 120
1000 100
800 80
600 60
400 40
200 20
0 0
0 20 40 60
Load (kW)
Figure 2.10. Example of total air and fuel flows for a 3.5:1 pressure ratio gas turbine engine.
Diffusion combustion CO
SOOT
HC
T
NO
[concentration]
Staged:
Pilot + Main
Staged:
Pilot
1.0
Idle Full
Figure 2.11. Comparison of staged and conventional combustion strategies idle
and full-power points shown for conventional strategy.
compares how staging allows the combustion system to stay within a lean com-
bustion regime (locally for each fuel injection point), whereas the conventional
non-staged engine must operate with combustion taking place over a wider range
of equivalence ratios and inevitably having to operate at conditions favorable for
high NOx emissions for at least some part of the load range.
This stated, how the staging is accomplished differs substantially. This is dis-
cussed in more detail in Section 2.7.3.
2.5 Engine-Level Metrics and Trade-offs 45
Gas turbine engine efficiency considerations remain a significant driver. With the
adoption of output-based emissions standards as well as the need to consider emis-
sions of greenhouse gases, overall cycle efficiency is becoming increasingly impor-
tant. Generally speaking, the heat engine basis of the gas turbine requires increasing
the temperature at which heat is added and decreasing the temperature at which it is
rejected. This efficiency metric has driven a steady increase in turbine firing tempera-
tures. The firing temperature is the defining criteria for cycle evolution in industrial
gas turbines evolving from D and E class to G, H, and J class. The firing temperature
for J-class engines is 1600C. To achieve this temperature at the turbine inlet, little or
no cooling is available for the combustor itself. Also, the 1600C temperature itself is
46 Ground-Based Gas Turbine Combustion
Base load:
nuclear hydro running
water,coal,steam plants
0 2 4 6 8 10 12 14 16 18 20 22 24
Day time
Figure 2.12. Representation of intermittent renewable power and role of gas turbines (both aeroderiv-
ative and combined cycle plants) (from Balling, 2010).
Exhaust
Fuel
Air
Combustor
Power
Compressor HP turbine turbine Generator
Figure 2.13. Simple cycle.
sufficient for thermal NOx production. This illustrates a trade-off between achieving
high efficiency while maintaining low NOx emissions. In this case, the cycle used is
the so-called simple cycle in that it is essentially a straight Brayton cycle. An illustra-
tion of this cycle is shown in Figure 2.13. In this cycle, the compression ratio is essen-
tially directly correlated to the combustor inlet temperature and pressure. Hence,
higher thermal efficiency is directly tied to potential for higher NOx emissions.
Alternatively, other cycles can be considered to gain efficiency without extreme
compression ratios. One relatively common cycle for smaller engines is the recuper-
ated cycle. Notable examples include the Solar Turbines Mercury 50 and the majority
of microturbines such as those manufactured by Capstone Turbine Corporation.
The recuperated cycle is illustrated in Figure 2.14. In this case, some of the otherwise
wasted heat from the Brayton cycle is partially recovered through the recuperator (a
gas-to-gas heat exchanger) and used to preheat the air entering the combustor. This
requires less fuel burn to achieve the same turbine inlet temperature. It also results
2.5 Engine-Level Metrics and Trade-offs 47
Exhaust
Recuperator
Fuel
Air
Combustor
Power
Compressor HP turbine turbine Generator
Figure 2.14. Recuperated cycle.
Inter Exhaust
Water Recuperator
cooler
in
Air
Fuel
Combustor
LP HP Power
compressors HP turbine turbine Generator
Figure 2.15. Intercooled/recuperated cycle.
0.50
ICR
0.45
0.40
Overall efficiency
0.35 Recuperated
Intercooled
0.30
0.25 Simple
cycle
0.20
1.5x 2x 2.5x 3x 3.5x
Specific power
Figure 2.16. Comparison of idealize efficiency versus specific power for fixed turbine inlet
temperature.
Economizer
Exhaust
Fuel
Air
Combustor
LP HP Power
compressors HP turbine turbine Generator
Figure 2.17. Humid air turbine cycle.
Exhaust
Recuperator
H2O DC power
Fuel SOFC
Air Fuel
Combustor
Power
Compressor HP turbine turbine Generator
Figure 2.18. Fuel cell/gas turbine hybrid cycle.
0.03
Constant T3, P3
Constant fuel composition
oscillations
Flash back
Stable
burning
0.01
Static
stability limit
0
0 0.25 0.50 0.75 1.00
Air mass flow (kg/s)
Figure 2.19. Illustration of combustion operability issues for gas turbine combustor at fixed
inlet temperature and pressure (adapted from Lieuwen et al., 2008).
change. Figure 2.19 also illustrates the presence of flashback, which can be an issue
as the velocities in the fuel injector/premixer become relatively low.
Finally, discrete points, often along the static stability limits, are shown that cor-
respond to operating conditions where combustion oscillations are problematic.
Each of these operability issues is discussed in the next sections.
2.6.1.1 Blowoff
Blowoff refers to the dynamic process of flame detachment and extinction. In gen-
eral, blowoff has been represented as a competition between time scales associated
with physical processes and kinetic processes. The ratio of the time scales is the
Dmkohler number. When the physical time scale (e.g., residence time) falls below
the kinetic time scale (e.g., reaction time), the combustor will be prone to blowoff.
Generally, the physical time scale is dictated by the design of the combustion system.
Parameters such as swirl strength and flame holder size dictate a time scale within
the combustion zone. On the other hand, the kinetic time scale is strongly affected
by temperature, pressure, and fuel composition. As a result, changing loads and/or
fuels significantly impacts the reaction times and, thus, the Dmkohler number.
This intuitively appealing description of the mechanism triggering blowoff has
resulted in decades of work essentially validating the concept (e.g., DeZubay, 1950;
Zukoski and Marble, 1955; Wright, 1959; Ballal and Lefebvre, 1979; Leonard and
2.6 Combustor-specific Metrics and Trade-offs 51
Mellor, 1983; Rizk and Lefebvre, 1986; Chaudhuri et al., 2010). This work has been
largely summarized by Shanbhogue and colleagues (2010) and essentially concludes
that this description effectively captures the global blowoff behavior. However, this
largely one-dimensional conceptual model may not fully capture the details of the
processes occurring. Indeed, local behaviors may indicate blowoff ahead of any time
ratio that may be responsible for the scatter in the data available. As fuel composi-
tion changes dramatically or inlet conditions reach levels outside of the conditions
for which the numerous experimental studies have been carried out, refinements to
the conceptualization of blowoff may be required. The impact of further transient
response requirements may also require new thinking in this area.
2.6.1.2 Flashback
Flashback is a phenomenon associated mainly with combustion systems relying
upon premixing of fuel and air prior to entry into the combustor. As summarized
by Lieuwen and colleagues (2008), flashback has been classified into at least four
different types of behavior: flashback into the core flow, flashback along a bound-
ary, flashback associated with combustion-induced vortex breakdown, and flashback
associated with combustion dynamics.
Regarding core flashback, the simplest design rule requires that the flow field
must not have strong local velocity deficits and that the flow velocity must be
substantially above the turbulent flame speed. In utilizing turbulent flame speed
data, it is important to recognize that multiple definitions of turbulent flame speed
exist, each applicable to different issues (Cheng, 2010). For flashback, the local
displacement speed is probably most relevant. Among the first equations used for
the calculation of turbulent flame speed were those developed theoretically by
Dmkholer (1947):
ST = SL + u
where
u = fluctuating velocity component.
Other theoretical expressions for turbulent flame speed were developed by Liu and
Lenze (1988) and Bradley (1992):
Additional results for ST,LD are reported by Littlejohn and colleagues (2008) using
measured velocities in a low-swirl burner operated on hydrogen:
ST u
= 1 + 3.15
SL SL
100
90 Damkohler, 1947
Liu and Lenze, 1988
80
Bradley, 1992
70
Littlejohn, 2008
60
ST,LD (m/s)
50
40
30
20
10
0
0 10 20 30 40
u' (m/s)
Figure 2.20. Summary of correlations from the literature for a constant SL= 0.34 m/s.
exist between these equations and experimental data. Figure 2.20 illustrates the sen-
sitivity of the estimated turbulent flame speed using these different expressions.
To illustrate the importance of the definition of turbulent flame speed used,
Figure 2.21 presents measured turbulent flame speeds based on global consump-
tion rates versus the local displacement rate as shown in Figure 2.20. Kido and
colleagues (2002) is shown as representative of ST, GD measurements because it
has been referenced often in turbulent flame speed discussions. Both experiments
shown were performed at atmospheric conditions (T = 298 K, P = 1 atm). As shown,
major differences in the turbulent flame speed will result depending on which defini-
tion is used.
In summary, flashback in the core flow depends on turbulent flame speed, which
in turn is fraught with challenges because of differences in definition. For gas tur-
bine application, those based on local displacement speed seem most appropriate,
but even then discrepancy is evident and the role of pressure, temperature, and
fuel mixtures is unclear. Some general trends are: leaner conditions results in a
slower ST, higher hydrogen content causes an increase in ST, increases in inlet bulk
velocitylead to increased ST, and the addition of diluents does not have a marked
effecton ST.
In gas turbine combustion, flame propagation along the burner walls is another
key flashback mechanism. Near the wall, the low velocities and turbulence in the
boundary layer promote flame propagation upstream. These effects compete with
flame quenching because of the heat loss of the burner wall. As flashback limits in
laminar flows clearly correlate with the velocity gradient at the wall, the concept of
the critical velocity gradient (Lewis and Von Elbe, 1943) has been widely adopted,
and is a function of the laminar flame speed and the thermal diffusivity, :
SL
gf
2.6 Combustor-specific Metrics and Trade-offs 53
30
CH4, Eq.Ratio = 0.7
CH4, Eq. Ratio = 0.9
25 CH4, Eq. Ratio = 0.98
Littlejohn, 2008
Linear (Littlejohn, 2008)
20
Based on Local Displacement Speed
ST/SL
15
10
Kido (2002). Based on Global Consumption Rate
0
0 5 10 15 20 25 30
u/SL
Figure 2.21. Kido et al. (2002) data collected using spark-ignited flame kernel shown here with
symbols (ST,GD). Correlations obtained using low-swirl burner (ST) (Littlejohn et al., 2008).
This equation can assess influence of the fuel on flashback in the boundary layer and
shows that the laminar flame speed has a substantial influence on the critical velocity
gradient required for flashback prevention. The equation implies that the velocity
gradient for hydrogen is approximately a factor ten times that of natural gas. This
equation is based on a decoupling between the approaching flow and the retreating
flame. However, as shown by Eichler and colleagues (2012), the flow field is actually
strongly influenced by the expansion of gases as a result of the flame. This different
view of the process could lead to an improved scaling method for predicting the role
of pressure, temperature, and fuel composition on flashback propensity in boundary
layers.
Whether the critical wall gradient for the corresponding turbulent boundary
layer is higher than that for the laminar case depends on the thickness of the quench-
ing distance with respect to the laminar sublayer (Wohl, 1952; Schfer et al., 2001).
Although no generalizations regarding flame propagation in turbulent boundary
layers are readily available, indications are that proper aerodynamic burner designs
produce substantially larger velocity gradients than required to avoid flashback
for natural gas. Unfortunately, the same conclusion cannot be made for fuels with
high-hydrogen content. Based on arguments for the laminar situation, gradients ten
times greater may be required. As a result, boundary layer flashback is a major area
of concern. In present systems, air can be judiciously utilized in the boundary layer
to mitigate conditions that can give rise to flashback. However, a difficulty remains
in that the addition of small amounts of air along the wall does not lead to the
desired diluted mixtures beyond the lean flammability limit in the critical near wall
zones for hydrogen containing syngas. Even with dilution, the laminar flame speed
54 Ground-Based Gas Turbine Combustion
near the wall may be substantially higher than for natural gas without dilution. As a
result, keeping the boundary layers as thin as possible is an essential design criterion
for syngas burners and, even more important, local separation zones in the mixing
zone must be avoided. Particularly critical are diffuser sections near the burner exit,
which lead to a rapid increase of the wall boundary layers.
For swirling flames, the presence of the flame can alter the vortex breakdown
behavior, which can lead to flashback. This behavior has been summarized in a num-
ber of works (e.g., Krner et al., 2003; Fritz et al., 2004; Kiesewetter et al., 2007;
Konle and Sattelmayer, 2010), which propose several correlations that can be used
to estimate the onset of vortex-induced breakdown. In this case, the relative mass
and thermal diffusivity, which are strong properties of the fuel, can play a role in
whether combustion-induced breakdown flashback occurs. This causes additional
design considerations to be required when bearing in mind fuel flexibility.
In summary, a major design criterion for the nozzle aerodynamics is that the
axial velocity in the nozzle must be as high and as uniform as possible and free of
strong wakes. Designs with constant or with slightly conical and accelerating airflow
paths downstream of the swirler provide the preferred solution. Strong acceleration
of the flow bears the danger of flame stabilization upstream near the fuel injector in
stoichiometric zones near the fuel jets, once the flame has passed the high-velocity
area downstream during unexpected events in gas turbine operations like compres-
sor surge and sudden breakdown of the burner mass flow. Thin boundary layers and
careful use of air in the boundary layer are necessary to avoid flashback along the
boundary layer.
2.6.1.3 Autoignition
Like flashback, autoignition is primarily a concern for systems using some degree of
premixing. It is particularly a concern for liquid fuels that might be used as either
primary or backup fuel, especially if a lean premixed strategy is utilized. Often, con-
densed fuel mist or lubrication oils can be an initiator of ignition. Because distillate
liquid fuels are so complex, it is difficult to identify a suitable kinetic mechanism that
can be used for estimating ignition delay. Some correlations have been developed,
such as the following one for Jet-A (Guin et al., 1998).
( 3377 / T ) 0.9
= 0.508e P (2.1)
As indicated, all a designer must know temperature and pressure, which are func-
tions of the cycle. This expression suggests that, at thirty bar and 700 K inlet tem-
perature, ignition delay times are around three ms. This is on the order of typical
premixer residence times and is consistent with other work done on ignition delay
for liquid fuels (Lefebvre et al., 1986).
For gaseous fuels, strong fuel compositional dependencies exist. Relative to
applications of autoignition in gaseous-fueled gas turbines, practical experience sug-
gests that this should not be a major concern. This has been summarized by Beerer
and McDonell (2008) and shown in Table 2.2. Of interest is the significant difference
2.6 Combustor-specific Metrics and Trade-offs 55
Table 2.2. List of current commercial engines with their approximate combustor inlet pressure
and temperature; along with estimations of ignition delay times for pure hydrogen, methane,
ethane, and propane at specific inlet conditions
between the full kinetic estimate of ignition delay time with that from correlations
observed in other measurements (e.g., Petersen et al., 2007). This observation has
led to considerable effort to explain the reasons (e.g., Chaos et al., 2010) and also to
illustrate that hydrogen, in particular, can develop inhomogeneities in the early igni-
tion time. Hence the influence of hydrodynamic effects may be important in these
systems.
Figure 2.22. Laser cut through a swirling flame showing flame distortion by vortical structures
(flow bottom to top). Image courtesy of T. Lieuwen.
Lieuwen, 2005). Fuel-air ratio oscillations arise because of the sensitivity of both the
fuel and airflow rates to perturbations. For example, pressure fluctuations at the fuel
injection point cause an oscillatory pressure drop across the fuel orifices, modulating
the fuel supply rate. Likewise, velocity oscillations cause the airflow rate the fuel is
mixing with to oscillate.
Similarly, the separating shear layers and other hydrodynamically unstable
flow features in the combustor are sensitive to perturbations. For example, shear
layers are naturally unstable and roll up into tightly concentrated regions of vor-
ticity in the absence of acoustic forcing. In the presence of acoustic forcing, these
concentrated vortices can pair and form a large-scale vortical structure whose pas-
sage frequency matches the acoustic excitation frequency. This large-scale vortex
then distorts the flame and causes its heat release to oscillate. An OH PLIF image
showing the rollup of the flame by an acoustically excited vortex is illustrated in
Figure 2.22.
While the reader is referred elsewhere for details (e.g., Lieuwen and Yang, 2005),
it is useful to summarize some of the key dependencies influencing the conditions
under which instabilities occur. In brief, these conditions are influenced by the natu-
ral acoustics of the combustor (controlled, in turn, by its size and the average sound
speed), and the distribution of heat release. Thus, the length of the flame and its sta-
bilization location strongly influence instability boundaries. In turn, these items are
influenced by fuel-air ratio (because of its effect on flame speed and, therefore, flame
length), ambient temperature, fuel composition, and flame stabilization approach. To
illustrate, consider Figure 2.23, which demonstrates these points. Figure 2.23a shows
the change in flame position associated with a change in, for example, fuel-air ratio
or combustor pressure and temperature. Figure 2.23b shows the effect of a different
geometry (in this case, a different center body), leading to a change in flame position
by changing the flame stabilization point from the shear layer to the forward stagna-
tion point of the vortex breakdown region.
2.7 Overview of Combustion Design Architectures 57
(a) (b)
Figure 2.23. Schematic showing (a) flames with two different lengths associated with change
in fuel-air ratio or air temperature (b) different center body geometries leading to different
flame stabilization point as in (a). Image courtesy of T. Lieuwen.
2.6.2 Emissions
Details regarding NOx and CO formation and emission are provided in Chapter 7.
For the majority of advanced gas turbine systems, NOx and CO tend to trade off
against each other. NOx is generally formed at high temperatures, which are favor-
able for promotion of CO oxidation. Of increasing importance is particulate emis-
sions, which is discussed in detail in Chapter 5 and Chapter 6. While particulates
have not been a focus for natural gas-fired advanced turbines, regional regulations
are looking again at this source. For example, sulfur compounds used to odorize
natural gas for safety is potentially a source of the fine particulate generated by gas
turbines operating on natural gas. Hence, particulates may be an issue associated
more with fuel specification than any sort of combustion behavior. Combustion may
be a contributor with operation on liquid fuels, where carbonaceous particles may
be generated. Evidence also suggests that combustion of higher hydrocarbons may
give rise to other species such as aldehydes.
For ground-based power, engines are classified to some extent based on their power
output. At the smallest scale (e.g., microturbines 25400 kW), the packaging is
optimized to be plug and play. In this case, all components can be integrated into
a single package that requires connection of fuel to the unit (and possibly water in
the case of an integrated combined heat and power unit) and power output from the
unit to an appropriate electrical feed-in. In this sense, if the site is prepped suitably,
the time from arrival to power generation can be a matter of one to two hours.
Larger industrial engines generally fall into heavy duty and aeroderivative
packages. The heavy-duty packages are purpose built for power generation.
58 Ground-Based Gas Turbine Combustion
To intercooler
Can Configuration
Can Annular
Annular
2.7 Overview of Combustion Design Architectures 59
Figure 2.26. GE-10 layout and scroll (Gas Turbine Short Course, 2002).
The first major category is can combustors. As suggested, these systems are essen-
tially quasi axisymmetric layouts with fuel injection at the centerline. An example is
shown in Figure 2.26. This silo approach is very convenient for maintenance as the
combustor is readily accessible. It also allows for convenient adaptation of the com-
bustion system to accommodate different fuel types and to change sizes and so forth.
The axisymmetric nature of the combustion system in this case can be appealing rel-
ative to overcoming combustion issues as the domain that needs to be addressed is
relatively compact. Any interaction of the combustion system with primary and dilu-
tion jets will have a circumferential nature allowing good penetration and mixing
with all the gases. However, a challenge with this particular approach is the need to
transform the hot gas path from the combustor into an annular feed to the turbine.
To accomplish this, a scroll is required as shown in Figure 2.26b. Addressing the heat
transfer issues with the scroll can be challenging.
A can-annular example is shown in Figure 2.27a. As shown, a series of cans is
positioned around the shaft. In this case, cross-fire tubes between the cans facilitate
the ignition process. As shown in Figure 2.27b, the outflow of each can is transitioned
into an annular flow that then hits the turbine/stator. In this configuration, the sym-
metry of the individual cans remains along with the potentially beneficial mixing
considerations, but the complexity of the scroll assembly is eliminated.
Other examples of can-annular designs can be found in many of the large frame
engines made by GE, Siemens, and others. Some of these offer the best of the main-
tainability offered by the silo approach with the other attributes outlined for can
systems. By essentially canting the cans out of parallel from the shaft, it allows
convenient access. Note the canted design shown in the GE 9H configuration
60 Ground-Based Gas Turbine Combustion
Figure 2.29. Rolls-Royce Allison 501 annular combustor (Courtesy of Rolls Royce).
presented in Figure 2.28 with the ring of combustor modules angled away from
the engine centerline.
The full annular arrangement is illustrated in Figures 2.29 and 2.30. In this case, fuel
injectors remain positioned at the head of the liner dome, but feed fuel and air into a
common combustion chamber. This removes considerable material from the combus-
tor system, which is appealing. This is one reason annular systems are found commonly
in aeroderivative engines. However, the annular system creates a situation where the
2.7 Overview of Combustion Design Architectures 61
Figure 2.31. Siemens Frame Engine Architecture. Courtesy by GE Energy and Siemens.
flames from the individual injectors can interact, as well as introduce transverse mode
acoustic interactions. It also leads to possible challenges with testing. With a can sys-
tem, one can reasonably expect a rig test with a single can to be representative of what
occurs in the engine. In the annular design, a sector would need to be considered,
which may still not capture the overall behavior, particularly with respect to acoustics.
Hence additional risk may exist in evolving a design or adapting a combustor to a
change in fuel type or cycle modification. In light of this, new frame engine designs fea-
ture can-type combustors. Examples include Siemens, with an early advanced annular
frame engine shown in Figure 2.31 along with its updated can counterpart.
which a small amount of fuel from the pilot circuit (~5 percent of the total fuel) can
be used to extend the blowoff limit of the reaction substantially.
In addition to staging on the individual injector, it is also possible to stage entire
burners. This approach is ubiquitous in current advanced lean premixed combustion
systems. As shown in Figure 2.30, the GE LM6000 can allow full burners to be staged
on and off in an essentially unlimited number of possible patterns. This can be used
to help overcome changes due to load, fuel composition, ambient conditions, and the
like. Staging adds additional complexity to the overall combustion system, in that the
interaction between the fired zones and cold unfired zones can lead to quenching of
the reaction, which may impact lean stability and emissions. Hence additional care is
required to achieve the combustion goals without impacting overall operability.
Other staging schemes involving multiple injectors include axial staging as used
in the GE DLN 1 (Figure 2.33) scheme. In this configuration, a combination of six
primary nozzles and one secondary nozzle can be operated over a range of loads to
minimize reaction temperatures and hot spots through careful mixing.
Another example is the 60 kW Capstone microturbine combustor shown in
Figure 2.34, which fuels two to six injectors depending on the load. More details on
staging strategies used in practice are discussed in Chapter 10 (Section 10.4).
2.8 Fuels
The fuels of interest for ground-based gas turbines are widely variable. Traditionally,
gas turbines have been omnivorous in terms of fuels. With the advent of
low-emissions requirements, however, gas turbines are increasingly sensitive to fuel
type. For example, numerous examples of fielded gas turbines operating on nearly
pure hydrogen can be found (e.g., oil operations), yet with relatively high NOx levels
as a result of burning the fuel in a diffusion flame mode.
2.8 Fuels 63
Primary
fuel nozzles
Centerbody
Lean and
premixing Secondary zone Dilution zone
primary zone
Secondary
fuel nozzle
A B
Outer liner
Dilution air
Primary
zone Dilution zone
A B Inner liner
Recuperator wall
Outer liner
Injector
Inner liner
Igniter
Figure 2.34. Capstone C-60 combustor.
In general, the cost of refining crude oils (either fossil or renewable oils) tends to
favor their use in premium applications such as aviation. However, for stationary
power generation, liquid fuels were widely used until the 1970s. With the onset of
increasingly stringent emission regulations, as well as the adoption of natural gas as
64 Ground-Based Gas Turbine Combustion
an attractive clean fuel and the recognition that liquid fuels are inherently attract-
ive for transportation applications because of high specific energy, the use of liquid
fuels in stationary gas turbines has dropped markedly.
However, depending on the region, the cost of liquid fuels may offer advantages
for use in stationary gas turbines. In addition, even for gas-fueled systems, cost sav-
ings can be available for operators willing to curtain operation to save gaseous fuels
for other applications. In these particular cases, if the gas turbine can operate on a
backup fuel, the operator can continue to produce power and still garner the benefit
of the interruptible gas rate. As a result, while generally not a first fuel choice, liq-
uid fuels are used in stationary systems.
For stationary power, the principal liquid fuel used is distillate fuel from refining
fossil fuels such as oil. The production of fuels from refined fossil fuels is a complex
process. The result is a fuel that may have a wide range of individual components.
These fuels also contain trace amounts of important species such as sulfur and con-
taminants such as water, gums, and metallic compounds. These trace species can have
a significant impact on the materials used in the combustor and fuel systems as well
as on pollutant emissions, particularly particulate (Rocca et al., 2003).
As discussed earlier, the more highly refined liquids are essentially reserved for
transport applications. These liquids would include gasolines. However, light distil-
lates and diesel fuels are common classes of fuels used in ground-based gas turbines.
Beyond that, various crudes, heavy distillates, and blended residual fuels may be
considered as well. These may otherwise be wasted, yet can be used in gas tur-
bines. The heavier crudes may require heating to evenly flow the liquid, yet the gas
turbine can handle it, albeit perhaps with elevated emissions levels. In regions of the
world with copious amounts of crude oil available, the accessibility of very low-cost
oil makes it attractive for use in gas turbines. Examples are especially prevalent in
regions of the Middle East with high growth. For example, in Riyadh, Saudi Arabia,
gigawatts of power are generated by combined cycle power plants operated on crude
oil (Bunz et al., 1984).
The most prevalent fuel is diesel fuel #2 (DF2), a relatively low-volatility liq-
uid also used in the transportation sector. Because of the latter, it is widely avail-
able. It is commonly used for remote power generation (i.e., in regions that lack
natural gas infrastructure) and also for backup power in the event natural gas is
curtailed. Diesel, as a distillate, is a relatively complex fuel. Figure 2.35 illustrates
a Mass-Spectroscopy Gas Chromatography analysis of a representative DF2. Each
peak represents a specific molecule. As shown, combinations of different hydrocar-
bon classes (e.g., aromatics and alkanes) are present. As a result, the combustion of
this fuel can be quite complex to describe.
During the 1970s and 2000s, concerns regarding sustainable oil supplies led to
considerable research in the area of alternative fuels. Early efforts studied oil from
tar sands and shale oil. In addition, the idea of using coal-water slurries to replace oil
was examined and progress was made demonstrating gas turbines on such fuels. In
some cases, the heavy oil from sources like tar sands led to development of method-
ologies to improve the utility of the basis fuel by altering it through emulsification.
By emulsifying the heavy oil with water, flowability improves, and the water can also
2.8 Fuels 65
7.61
Benzene
100
90 Aromatics
6.54
9.93
80 23+hydrocarbons > 2%
of total composition
70
Relative abundance
60 Alkanes
14.15
16.31
50
4.53
4.17
11.04
32.89
16.82
40
36.59
29.02
40.19
24.96
18.63
20.71
43.45
3.04
30
48.66
49.71
20 Volatility
21.31
52.66
25.42
10
T
0
0 10 20 30 40 50 60 70 80
Time (min)
Figure 2.35. MS-GC analysis of typical diesel fuel #2.
49.88
100
50.12
90 Primarily 4
fatty acid, methyl esters
80
70
Relative abundance
60
50
40
30
Volatility
44.35
20
10 T
0
0 10 20 30 40 50 60 70 80
Time (min)
Figure 2.36. MS-GC analysis of typical biodiesel.
pipeline infrastructure. Extensive work by the Gas Research Institute to define the
variability in the standard pipeline gas found throughout the United States has been
relied upon as an indicator of the typical variability (Liss et al., 1992). This variation
is summarized in Table 2.4. Because higher molecular weight hydrocarbons have a
substantially lower autoignition temperature, they can lead to autoignition in the
fuel-mixing region of the lean premixed combustion systems or catalytic combustion
systems (e.g., Richards et al., 2001; Lieuwen et al., 2008). In addition, the liquid
droplets can create locally high concentrations of higher hydrocarbons.
A second implication of varying concentrations of higher hydrocarbons is the
effect of the components on the flame position and stability. In this case, the kinetic
reaction rate of the higher hydrocarbons is substantially different (higher) than that
of methane, resulting in changes of the flame speeds and, therefore, flame location.
It is not clear to what extent the composition of natural gas throughout the
United States and the world will vary. Clearly it will depend upon regions of gas
2.8 Fuels 67
Table 2.4. Concentration of fuel constituents for natural gas in the United States
* Without peakshaving
Sources: Liss et al., 1992
combustion system (CO2 can also affect the behavior of a catalyst in a catalytic
c ombustion system). If CO2 is present, it can affect the combustion processes
through its high heat capacity and by its competition for H radical, through the
CO+OH > CO2+H reactions.
Natural gas will also contain some moisture, although usually by the time it is
delivered it is almost completely dry. At the levels likely to be present, water prob-
ably will have little effect; however, scientists can explore if a satisfactory means of
introducing moisture can be established.
Another factor associated with natural gas is the sulfur content. The United
States boasts large reserves of natural gas; however, the sweet gases are rapidly
being depleted, leaving increasingly sour gases for future use. Today the sour nat-
ural gases, that is, those containing hydrogen sulfide or other organo-sulfur com-
pounds, are extracted and diluted with sweet or sulfur-free gases to maintain the
desired low levels of sulfur in the gas. In the near future, sulfur removal will become
necessary and this will mean that all natural gases will tend toward having the maxi-
mum allowable sulfur levels. It is not clear what these maximum levels may be in the
future, but today, based on gas transfer contracts, levels are between ten to twenty
grains per 100 ft3 (0.220.45 grams/m3). Of course, sulfurous compounds are added
to natural gas as an odorant for safety purposes.
NGLs nat
ura
l ga
sp Liquefied
ipe
line petroleum gas
LPG
LPG storage
Underground
NG storage Liquefied LNG
natural gas
LNG storage
High pressure NG
transmission to load
centers
LPG/LPG-air
system base load
Future / remote
natural gas
consumers
LPG-air systems
Firm natural gas standby or base load
consumers
Interruptible
natural gas
consumers
Figure 2.37. North American Gas Energy Grid: Natural Gas and LPG (Standby Systems, Inc.).
reserves (George and Bowles, 2011). A still not fully understood potential downside
to the use of shale gas is that the new extraction methods (hydraulic fracturing)
require considerable water use and involve injection of several chemicals, including
benzene, to fracture the shale to release the gas. The environmental impact of this
water/chemical use may well impact the overall feasibility of shale gas.
Table 2.5. Fuel composition from various sources around the world
have gained attention for regions with limited natural gas reserves. This led
to significant efforts to assess the impact of such scenarios on a wide range of
combustion devices (NGC+, 2005). The variation in composition expected with
high LNG import scenarios would feature more C3+ content than expected for any
natural gas as summarized by the GRI report (Liss et al., 1992; Liss and Rue, 2005).
The exact range of values depends on the origin of the LNG. As shown in Table 2.5,
the amount of the C3+ values anticipated could exceed the maximum levels in the
legacy United States pipeline gas (Table 2.4) by more than 50 percent.
2.8.2.2.2 Refinery Gas. The refining of fuels leads to the production of off-gases that
can potentially be used to produce power. The composition of these gases is different
from LNG-type fuels in that they contain high quantities of hydrogen. An example
of a typical composition of these gases is shown in Table 2.6 (Rao et al., 1996). In
regions with refining operations, such gases and strategies for using them for energy
production are already being pursued, though generally motivated by regulatory
2.8 Fuels 71
H2 25.7
CO 1.5
CH4 37.4
N2 2.9
C2 27.5
C3 2.9
C4 1.7
C5 0.4
LHV (Btu/ft3) 984
1
Adapted from California Energy Commission http://www.energy.ca.gov/research/renewable/
biomass/landfill/index.html.
72 Ground-Based Gas Turbine Combustion
an important issues with landfills. Sulfur compounds can range from negligible to
1,700 ppm, and other compounds like siloxanes can be significant. In recent years,
attempts to convert landfill gas to energy have required varying degrees of care
relative to gas cleanup to prevent damage or coating of critical power generation
device parts.
When landfill gas is vented, the GHG implications are serious as methane is a
much stronger greenhouse-effect-forcing species than CO2. However, conversion of
the LFG to power generates CO2 and also potentially criteria pollutants. While the
benefits of generating power from this otherwise wasted fuel are apparent, regula-
tory pressures require a significant reduction in criteria emissions.
slow decay of matter, which can be accelerated with the use of digestion strategies.
Essentially, these processes can accelerate the breakdown of organic material and
generation of methane gases. Most common is the use of anaerobic (i.e., without
air but really without oxygen to the extent possible) digestion (AD), a biological
process in which biodegradable organic matters are broken down by bacteria into
biogas, which consists of methane (CH4), carbon dioxide (CO2), and other trace
amounts of gases. The biogas can generate heat and electricity. Other important
factors, such as temperature, moisture and nutrient contents, and pH are also critical
for the success of AD. In terms of temperature, either mesophilic AD (3040C) or
thermophilic AD (5060C) can be used. In general, AD at lower temperature is
more common, but thermophilic temperature has the advantage of reducing reaction
time, which corresponds to reduction of digester volume. Moisture contents greater
than 85 percent are suitable for AD.
Because of the nature of the digestion processes, the composition of fuel gas
from these systems varies far less than from landfills and tends to have overall higher
methane content.
AD technology is well developed worldwide with an estimated 5.36.3 GW
installed. Traditional, small, farm-based digesters have been used in China, India,
and elsewhere for centuries. The number of digesters of this type and scale is esti-
mated to exceed six million. European (EU) companies are world leaders in devel-
opment of AD technology.
2
Adapted from California Energy Commission http://www.energy.ca.gov/research/renewable/
biomass/anaerobic_digestion/index.html.
2.8 Fuels 73
Carbon Based
Material
Heat
Carbon Based
Inert Gas or Nothing Pyrolysis Gas+Tar
(inert atmosphere) Char
Heat
Figure 2.38. Block diagram of the gasification and pyrolysis processes (adapted from Belgiorna
et al., 2003).
H2 8.6 61.9
CO 22.3 55.4
CO2 1.6 30.0
CH4 0.1 8.2%
N2 + Ar 0.2 49.3
H2O 0 39.8
that involve the use of high-hydrogen fuels. One representative composition was
selected for each of the following scenarios in addition to pure hydrogen:
Table 2.8 shows the representative syngas composition on a dry, volumetric basis
as produced by different gasification methods. Pure hydrogen represents the case
where carbon sequestration is utilized, though it is becoming apparent that 90 per-
cent carbon removal may be a reasonable upper limit (meaning the fuel would be 90
percent hydrogen and balance methane). The process and refinery gas blend reflects
the interest of smaller-scale combustion systems, while the gasified coal/petcoke is
indicative of current IGCC central power plants. Note that for these larger-scale
applications, expectations are that an air separation unit (ASU) will be used and, as
a result, the gasifier will be fed with oxygen as opposed to air. In a smaller unit, the
air separation unit may not be cost-effective, so gasification with air is more likely.
Finally, diluted fuel is representative of syngas diluted with nitrogen, which is the
current practice for combustion systems.
Hence, while the different ranges of concentrations of the minor species deter-
mines which specific fuel it represents, it may be reasonable from an experimental
efficiency viewpoint to treat all three of these as composition variations on the same
base fuel.
76 Ground-Based Gas Turbine Combustion
2.8.2.6 Summary
In summary, Table 2.9 presents a general perspective on the typical compositional
variation that might be found in gaseous fuels for use in gas turbines.
2.9 Summary
Gas turbines for power generation represent a large installed base and will play
a significant role in meeting future energy needs. The drivers for gas turbines for
power generation are much different from aero engines, and regulation of environ-
mental impact for these systems often drives technology advancement. Gas turbines
for power generation can be configured in a myriad of overall systems designed and
applications, often requiring trade-offs to be made. Advancements in gas turbine
technology for natural gas fuels has resulted in significant decreases in criteria pollu-
tant emissions and increasing concern for climate change has driven these systems to
impressive efficiencies to minimize mass emissions of greenhouse gases. The outlook
for power generation gas turbines includes a need to further fuel flexibility and also
to provide fast dynamic response. The coupling of these future needs with further
reductions in pollutant emissions while maintaining or improving operability will
require further understanding of the combustion system characteristics and of the
general combustion process. The desire for higher efficiencies will lead to higher
temperatures and/or pressures within the cycle, leading to further challenges for
References 77
operational and fuel flexibility for gas turbines. In short, while gas turbine combus-
tion technology has evolved substantially, numerous opportunities for improvement
in operational flexibility are apparent.
References
Araki, H., Koganezawa, T., Myouren, C., Higuchi, S., Takahashi, T., and Eta, T. (2012).
Experimental and Analytical Study on the Operation Characteristics of the AHAT
System. J. Engr Gas Turbine and Power 134(5): 0517011: 8.
Ballal, D. R., and Lefebvre, A. H. (1979). Weak Extinction Limits of Turbulent Flowing
Mixtures. ASME J. Engineering for Gas Turbines and Power 101: 343.
Balling, L. (2010). Fast-cycling and Starting Combined Cycle Power Plants to Backup
Fluctuating Renewable Power. Industrial Fuels and Power August 27.
Beerer, D. J., and McDonell, V. G. (2008). Autoignition of Hydrogen and Air inside a
Continuous Flow Reactor with Applications to Lean Premixed Combustion. Journal of
Engineering for Gas Turbines and Power 130: 0515071.
Beerer, D. J., McDonell, V. G., Samuelsen, G. S., and Angello, L. (2011). An Experimental
Ignition Delay Study of Alkane Mixtures in Turbulent Flows at Elevated Pressure and
Intermediate Temperatures. Journal of Engineering for Gas Turbines and Power 133:
101503111.
Belgiorno, V. et al. (2003). Energy from Gasification of Solid Wastes. Waste Management
23: 115.
Bolszo, C. D., and McDonell, V. G. (2009). Evaluation of Plain-Jet Air Blast Atomization
and Evaporation of Alternative Fuels in a Small Gas Turbine Engine Application.
Atomization and Sprays 19(8): 77185.
Bradley, D. (1992). How Fast Can We Burn? Symposium (International) on Combustion
24(1): 24762.
Bunz, W. J., Ziady, G. N., and Von E. Doring, H. (1984). Crude Oil Burning Experience in
MS5001P Gas Turbines. Journal of Engineering for Gas Turbines and Power 106(4):
81219.
Chaos, M., Burke, M. P., Ju, Y., and Dryer, F. L. (2010). Syngas Chemical Kinetics and Reaction
Mechanisms, chapter 2 in Synthesis Gas Combustion: Fundamentals and Applications,
Lieuwen, T., Yang, V., and Yetter, R. eds., Taylor & Francis, Boca Raton, FL.
Cheng, R. (2010). Turbulent Combustion Properties of Premixed Syngas, chapter 5 in
Synthesis Gas Combustion: Fundamentals and Applications, Lieuwen, T., Yang, V., and
Yetter, R. eds., Taylor & Francis, Boca Raton, FL.
Chaudhuri, S., Kosta, S., Renfro, M. W., and Cetegen, B. M. (2010). Blowoff Dynamics
of Bluff Body Stabilized Turbulent Premixed Flames. Combustion and Flame 157:
790802.
Collot, A. G. (2006). Matching Gasification Technologies to Coal Properties. International
Journal of Coal Geology 65: 191212.
Dmkholer, G. (1947). Influence of Turbulence on the Velocity of Flame in Gas Mixtures.
Z.Elektrochem 46: 60126 (1950). English translation, NACA-TM-I 112.
Demirbas, A. (2007). Progress and Recent Trends in Biofuels. Progress in Energy and
Combustion Science 33(1): 117.
DeZubay, E. A. (1950). Characteristics of Disk-controlled Flames. Aero Digest, 54(6):
1024.
Ducruix, S., Schuller, T., Durox, D., and Candel, S. (2005). Combustion Instability Mechanisms
in Premixed Combustors, in Combustion Instabilities in Gas Turbine Engines, Lieuwen,
T. and Yang, V. eds., AIAA Press., pp 179212.
Eichler, C., Baumgartner, G., and Sattelmayer, T. (2012). Experimental Investigation of
Turbulent Boundary Layer Flashback Limits for Premixed Hydrogen-air Confined in
Ducts. Journal of Engineering for Gas Turbines and Power 134: 01150218.
78 Ground-Based Gas Turbine Combustion
EPRI (1993). A Feasibility and Assessment Study for FT4000 Humid Air Turbine (HAT),
Report TR-102156.
Fritz, J., Krner, M., and Sattelmayer, T. (2004). Flashback in a Swirl Burner with Cylindrical
Premixing Zone. Journal of Engineering for Gas Turbines and Power 126: 26783.
George, D.L., and Bowles, E.B. (2011). Shale Gas Measurement and Associated Issues,
Pipeline and Gas Journal 238: 7.
Guin, C. (1998). Characterization of Autoignition and Flashback in Premixed Injection
Systems. AVT Symposium on Gas Turbine Engine Combustion, Emissions, and
Alternative Fuels, Lisbon.
Hauserman, W. B. (1994).High-Yield Hydrogen Production by Catalytic Gasification of Coal
or Biomass. International Journal of Hydrogen Energy 19: 41319.
Higuchi, S., Koganezawa, T., Horiuchi, Y., Araki, H., Shibata, T., and Marushima, S. (2008).
Test Results from the Advanced Humid Air Turbine System Pilot Plant: Part 1 Overall
Performance. TurboEXPO.
Jones, R. M., and Shilling, N. Z. (2003). IGCC Gas Turbines for Refinery Applications. GE
Power Systems Report GER-4219, 05/03.
Kavanagh, R. M., and Parks, G. T. (2009a). A Systematic Comparison and Multi-Objective
Optimization of Humid Power Cycles Part I: Thermodynamics. Journal of Engineering
for Gas Turbines and Power 131(4): 041701.
(2009b). A Systematic Comparison and Multi-Objective Optimization of Humid Power Cycles
Part II: Economics. Journal of Engineering for Gas Turbines and Power 131(4): 041702.
Kharchenko, N. V. (1998). Advanced Energy Systems, Taylor & Francis, Washington DC.
Kido, H., Nakahara, M., Hashimoto, J., and Barat, D. (2002). Turbulent Burning Velocities of
Two-Component Fuel Mixtures of Methane, Propane and Hydrogen. JSME International
Journal Series B 45(2): 35562.
Kiesewetter, F., Konle, M., and Sattelmayer, T. (2007). Analysis of Combustion Induced
Vortex Breakdown Driven Flame Flashback in a Premixed Burner with Cylindrical
Mixing Zone. Journal of Engineering for Gas Turbines and Power 129: 92936.
Knothe, G. (2010). Biodiesel and Renewable Diesel: A Comparison. Progress in Energy and
Combustion Science 36(3): 36473.
Konle, M., and Sattelmayer, T. (2010). Time Scale Model for the Prediction of the Onset
of Flame Flashback Driven by Combustion Induced Vortex Breakdown. Journal of
Engineering for Gas Turbines and Power 132: 041503.
Krner, M., Fritz, J., and Sattelmayer, T. (2003). Flashback Limits for Combustion Induced
Vortex Breakdown in a Swirl Burner. Journal of Engineering for Gas Turbines and
Power 125: 693700.
Lefebvre, A. H., and Ballal, D. (2010). Gas Turbine Combustion, 3rd Edition, CRC Press, Boca
Raton, FL.
Lefebvre, A. H., Freeman, W., and Cowell, L. (1986). Spontaneous Ignition Delay
Characteristics of Hydrocarbon Fuel/Air Mixtures. NASA Contractor Report 175064.
Leonard, P. A., and Mellor, A. M. (1983). Correlation of Lean Blowoff of Gas Turbine
Combustors using Alternative Fuels. Journal of. Energy 7(6): 72932.
Lewis, B., and Von Elbe, G. (1943). Stability and Structure of Burner Flames. Journal of
Chemical Physics 11: 7598.
(1987). Combustion, Flames and Explosions of Gases, 3rd Edition, Academic, New York.
Li, S. C., and Williams, F. A. (2002). Reaction Mechanism for Methane Ignition. Journal of
Engineering for Gas Turbines and Power 124: 47180.
Lieuwen, T. C., and Yang, V., eds. (2005). Combustion Instabilities in Gas Turbine Engines:
Operational Experience, Fundamental Mechanisms, and Modeling. Progress in
Astronautics and Aeronautics, 210, AIAA, Virginia.
Lieuwen,T. C., McDonell,V., Petersen, E., and Dantavicca, D. (2008).Fuel Flexibility Influences
on Premixed Combustor Blowout, Flashback, Autoignition, and Stability. Journal of
Engineering for Gas Turbines and Power 130(January): 0115061 01150610.
Lieuwen, T. C., Yang, V., and Yetter, R., eds. (2010). Synthesis Gas Combustion: Fundamentals
and Applications, CRC Press, FL.
References 79
Liss, W. E., and Rue, D. (2005). Natural Gas Composition and Fuel Quality. Information
Report from the Gas Technology Institute.
Liss, W. E., Thrasher, W. H., Steinmetz, G. F., Chowdiah, P., and Attari, A. (1992). Variability
of Natural Gas Composition in Select Major Metropolitan Areas of the United States.
GRI Report 92/0123.
Littlejohn, D., Cheng, R. K., Noble, D. R., and Lieuwen, T. C. (2010). Laboratory Investigations
of Low-Swirl Injectors Operating With Syngases. Journal of Engineering for Gas Turbines
and Power 132(1): 011502011508.
Liu, Y. and Lenze, B. (1992). Investigation of Flame Generated Turbulence in Premixed
Flames and Low and High Burning Velocities. Experimental Thermal and Fluid Science
5(3): 41015.
McDonell, V. (2008). Lean Combustion in Gas Turbines, in Lean Combustion Technology
and Control, Dunn-Rankin, D. ed., Academic Press, San Diego, pp. 121160.
Moore, J. J., Kurz, R., Garcia-Hernandez, A., and Brun, K. (2009). Experimental Evaluation
of the Transient Behavior of a Compressor Station During Emergency Shutdowns. Paper
GT200959064, TurboEXPO 2009, Orlando, FL, June.
NGC+ (2005). White Paper on Natural Gas Interchangeability and Non-Combustion End Use.
Ni, M. et al. (2006). An Overview of Hydrogen Production From Biomass. Fuel Processing
Technology 87: 46172.
Nigam, P. S., and Singh, A. (2011). Production of Liquid Biofuels from Renewable Sources.
Progress in Energy and Combustion Science, 37(1): 5268.
Notari, B. (1991). C-1 Chemistry, A Critical Review. Catalytic Science and Technology 1:
5576.
Peschke, W. T., and Spadaccini, L. J. (1985). Determination of Autoignition and Flame Speed
Characteristics of Coal Gases Having Medium Heating Values. Final Report for AP-4291
Research Project 23571, November.
Petersen, E. L., Kalitan, D. M., Barrett, A. B., Reehal, S. C., Mertens, J. D., Beerer, D. J., Hack,
R. L., and McDonell, V. G. (2007). New Syngas/Air Ignition Data at Lower Temperature
and Elevated Pressure and Comparison to Current Kinetics Models. Combustion and.
Flame 149(12): 2447.
Rayleigh, J. S. W. (1945). The Theory of Sound, Vol. 2, Dover: New York.
Reale, M. J. (2004). New High Efficiency Simple Cycle Gas Turbine GEs LMS100TM.
GER-4222A, June.
Rao, A. D., Francuz, D. J., and West, E. W. (1996). Refinery Gas Waste Heat Energy Conversion
Optimization in Gas Turbines. IJPGC Conference, pp. 47383.
Richards, G. A., McMillian, M. M., Gemmen, R. S., Rogers, W. A., and Cully, S. R. (2001). Issues
for Low-Emission, Fuel-Flexible Power System. Progress in Energy and Combustion
Science. 27: 14169.
Rizk, N. K., and Lefebvre, A. H. (1986). Relationship between Flame Stability and Drag of
Bluff-body Flameholders. Journal of Propulsion and Power 2(4): 361.
Rocca, E., Steinmetz, P. and Moliere, M. (2003). Revisiting the Inhibition of Vanadium-Induced
Hot Corrosion in Gas Turbines. Journal of Engineering for Gas Turbines and Power 125:
664.
Rojey, A., Jaffret, C., Cornot-Gandolphe, S., Durand, B, Jullian, S., and Valais, M. (1994). Natural
Gas Production, Processing, Transport, Imprimerie Nouvelle, Paris.
Rukes, B. and Taud, R. (2004). Status and Perspectives of Fossil Fuel Power Generation,
Energy 29: 185374.
Ryan, N. E., Larsen, K. M., and Black, P. C. (2002). Smaller, Closer, Dirtier, Diesel Backup
Generators in California. Environmental Defense Fund Report.
Schfer, O., Koch, R., and Wittig, S. (2001). Flashback in Lean Prevaporized Premixed
Combustion: Non-swirling Turbulent Pipe Flow Study. ASME Paper 2001-GT-0053.
Schelkin, K. I. (1943). On Combustion in Turbulent Flow. Zh.Tekh. Fiz 13: 52030.
Shanbhogue, S. J., Husain, S., and Lieuwen, T. C. (2010). Lean Blowoff of Bluff Body
Stabilized Flames: Scaling and Dynamics. Progress in Energy and Combustion Science
35: 98120.
80 Ground-Based Gas Turbine Combustion
Singer, B. C. (2007). Natural Gas Variability in California: Environmental Impacts and Device
Performance Literature Review and Evaluation for Residential Appliances. Report
CEC-5002006110, February.
Soares, C. (2008). Gas Turbines, A Handbook of Air, Land, and Sea Applications, Elsevier,
London.
Standby Systems, Inc. (20016). Propane Peak Shavingan Overview.
Spath, P. L., and Dayton, D. C. (2003). Preliminary Screening Technical and Economic
Assessment of Synthesis Gas to Fuels and Chemicals with Emphasis on the Potential for
Biomass-Derived Syngas. National Renewable Energy Laboratory.
Stephens-Romero, S., Carreras-Sospedra, M., Brouwer, J., Dabdub, D., and Samuelsen, S. (2009).
Determining Air Quality and Greenhouse Gas Impacts of Hydrogen Infrastructure and
Fuel Cell Vehicles, Environmental Science and Technology. online November 5.
Stiegel, G. J., and Ramezan, M. (2006). Hydrogen from Coal Gasification: An Economical
Pathway to a Sustainable Energy Future. International Journal of Coal Geology 65:
17390.
Tillman, D. A. and Harding, N. S. (2004). Fuel of Opportunity: Characteristics and Uses in
Combustion Systems, Elsevier, Oxford.
Trommer, D. et al. (2005). Hydrogen Production by Steam-Gasification of Petroleum Coke
Using Concentrated Solar Power I. Thermodynamic and Kinetic Analyses. International
Journal of Hydrogen Energy 30: 60518.
Walsh, P. P., and Fletcher, P. (2004). Gas Turbine Performance, 2nd Edition, Blackwell
Publishing, Oxford.
Wender, I. (1996). Reactions of Synthesis Gas. Fuel Processing Technology 48: 189297.
Wohl, K. (1952) Quenching, Flash-Back, Blow-Off Theory and Experiment. 4th Symposium
(Int.) on Combustion, pp. 6989.
Wright, F. H. (1959). Bluff-body Stabilization: Blockage Effects. Combustion and Flame 3:
31937.
Zinn, B. T., and Lieuwen T. C. (2005). Combustion Instabilities: Basic Concepts, in
Combustion Instabilities in Gas Turbine Engines. AIAA Press.
Zou, J. H. et al. (2007). Modeling Reaction Kinetics of Petroleum Coke Gasification with
CO2. Chemical Engineering and Processing 46: 6306.
Zukoski, E. E. and Marble, F. E. (1955). The Role of Wake Transition in the Process of Flame
Stabilization on Bluff Bodies. AGARD Combustion Research and Reviews, p. 167.
3 Overview of Worldwide Aircraft
Regulatory Framework
Willard Dodds
Whether they operate in the air, on the ground, or at sea, gas turbine engines must
deliver safe and reliable operation, high efficiency, and environmentally acceptable
emissions. However, engines designed specifically for each of these applications have
different environmental impacts and operating constraints that affect the range of
technologies that can reasonably be applied and the type of emissions regulatory
framework that is best suited to regulate their design, qualification, and operation.
Natural gas fuel reduces conventional combustor NOx emissions by nearly 50 per-
cent compared to jet fuel, primarily by reducing adiabatic flame temperature.
Water injection or use of dry low-emissions combustor technology reduces
remaining NOx by approximately 90 percent.
Catalytic exhaust gas cleanup reduces what emissions are left by approximately
90 percent.
Aircraft and industrial engines both emit significant amounts of carbon dioxide and
water formed during oxidation of the carbon and hydrogen in the fuel, as well as
a smaller quantity of sulfur oxides (SOx) from trace amounts of sulfur in the fuel.
81
82 Worldwide Aircraft Regulatory Framework
SOx emissions at the engine exhaust primarily occur in the form of gaseous SO2,
but a small portion of the fuel sulfur (estimated at 3.3 percent by Wayson et al.,
2009) is typically emitted as a range of heavier sulfur species classified as volatile
particulate matter.
NOx emissions are primarily comprised of NO and NO2. At idle conditions,
NOx emissions are about half NO and half NO2, while at high power, roughly 90
percent of NOx is in the form of NO.
Unburned hydrocarbons contain many hydrocarbons ranging from large fuel
molecules to very light hydrocarbons, and include several species classified as
hazardous air pollutants (HAPs) because they are toxic or carcinogenic (see, for
example, Spicer et al., 1994 and Herndon et al., 2009).
Particulate matter includes both nonvolatile soot (small carbonaceous parti-
cles that make up visible smoke) and volatile hydrocarbons and sulfur oxides
that condense as the engine exhaust plume cools through mixing with the
ambientair.
At low altitude (within the atmospheric boundary layer that typically extends up
to about three thousand feet, on the average), aircraft emissions affect local air
quality in much the same way as other ground-based emissions sources. Effects of
low-altitude aircraft engine emissions have been studied in detail (see, for example,
Ratliff et al., 2009). Some species in the exhaust are of direct concern, while other
species participate in chemical reactions that form pollutants:
Because of emissions potential health impacts, the United Nations World Health
Organization and many individual states define national ambient air quality standards
(NAAQS) for pollutants such as ozone, particulate matter, carbon monoxide, nitro-
gen dioxide, and sulfur dioxide. In the United States, the Environmental Protection
Agency (EPA) determines the potential health impacts and sets NAAQS for pollut-
ants that may be harmful to public health (http://www.epa.gov/air/criteria.html).
Aircraft engine emissions are unique in that approximately 90 percent of the total
NOx, SOx, CO2, and water formed in the combustion process are emitted at altitudes
3.1 Aero and Industrial Engines Contrasting Requirements 83
between three thousand and forty thousand feet. These high-altitude emissions are
not currently regulated, but they can contribute to potentially significant impacts on
air quality over a wide geographical area and can also affect climate change.
A global atmospheric modeling study described by Barrett and colleagues
(2010) indicates that the health impact of NOx emitted at high-altitude climb and
cruise conditions may be several times greater than the impact of low-altitude emis-
sions. The study indicates that much of the NOx emitted at high altitude is trans-
ported to ground level via subsiding air masses, where it adds to formation of ozone
and secondary PM.
With respect to climate change, the most significant greenhouse gas is CO2, and
the impact of CO2 on climate change is independent of the altitude at which it is
emitted. However, according to the Intergovernmental Panel on Climate Change
(1999), NOx emitted by aircraft during climb and cruise affect climate by increas-
ing ozone, which leads to warming, and reducing methane, which leads to cooling.
Relative impacts of ozone, methane, and CO2 are difficult to compare because the
three gases have much different lifetimes in the atmosphere. The warming due to
ozone only lasts a few months, while the reduction in methane lasts for about a
decade. Aviations impact also varies at different latitudes. Since most aviation
occurs in the northern hemisphere, most of the ozone is formed there, but because
of the short lifetime of ozone, it is destroyed before it can migrate to the southern
hemisphere, so a majority of the warming effect is in the north. On the other hand,
because of its long lifetime, the cooling effect of methane is evenly distributed over
the hemispheres.
The magnitude of climate impacts due to NOx-induced changes in ozone and
methane are of the same order of magnitude as that of CO2, but the impacts are
offsetting, so the combined impact is still uncertain.
The foremost requirement for aviation has always been safety. On the ground or at
sea, an engine failure can be inconvenient and expensive, but an engine failure in an
aircraft at thirty-five thousand feet can be a disaster. As described in Chapter 1, safety
considerations dictate that the aircraft engine combustor must be designed to:
National or regional regulatory agencies such as the U.S. Federal Aviation Agency
(FAA), the European Aviation Safety Agency (EASA), and civil aviation agencies
84 Worldwide Aircraft Regulatory Framework
(CAA) of other countries enforce safety regulations. These agencies issue airworthi-
ness certificates for engines and aircraft that have demonstrated that they meet all
airworthiness requirements.
From a combustor design standpoint, combustors designed to meet aircraft cer-
tification requirements for high-altitude air starts and fast transients have more strin-
gent ignition and stability requirements than ground-based engines. Furthermore,
flight operations require extended operation at low power. For a short-range flight,
nearly half of the operating time can be at approach (about 30 percent of rated
thrust) and taxi or descent points that require less than 10 percent of rated engine
thrust. The aircraft engine combustor must be designed for very low CO and HC
emissions at these low-power conditions.
3.1.4 Fuels
Aircraft engines typically only operate with high-quality liquid jet fuels having a
relatively narrow range of properties. Industrial engines may be required to operate
with fuels ranging from low Btu gas to heavy oils, but natural gas is the predominant
fuel for most industrial engine operation where low emissions are required. Use of
natural gas is a great benefit from an NOx emissions standpoint. For a conventional
aeroderivative combustor, simply changing from jet fuel to natural gas reduced NOx
emissions by almost half, primarily because of the lower adiabatic flame temperature
of natural gas. For advanced low-emissions lean premixing combustor designs, natu-
ral gas has advantages in that evaporation doesnt have to be considered; the fuel has
high thermal stability, so it is relatively easy to use a multiple point injection system
without having to worry about fuel coking; and ignition delay time of natural gas is
roughly an order of magnitude longer than jet fuel, so there is more time to premix.
industrial operating engine requirements with relatively small fuel injector modifica-
tions to enable use of the wide range of liquid and gaseous fuels required in industrial
applications. However, as described in Chapter 1, aircraft engines have severe weight
and volume constraints that limit the emissions reduction technologies that can be
applied. Any increase in weight requires an equal reduction in aircraft payload.
For example, water injection is a proven technology for reducing NOx emissions
from a conventional gas turbine combustor. Injecting equal parts water and fuel
through the fuel injector of an industrial engine reduces NOx by an order of magni-
tude, reduces the temperature of the combustor exhaust products that the turbine
has to accommodate, and increases power by increasing mass flow. Unfortunately,
this technology is not easily adapted to aircraft engines because of the weight of the
water that would have to be carried.
Catalytic exhaust gas cleanup is another proven technology that can signifi-
cantly reduce emissions. A selective catalytic reduction (SCR) system can reduce
NOx, CO, and unburned hydrocarbons, but available industrial systems are larger
and heavier than the aeroderivative engine itself, so it is not practical for use on an
aircraft.
Lean premixing combustors have been used in industrial engines for over fifteen
years (Leonard and Stegmaier, 1994), but are just beginning to be used on aircraft
engines. Several issues have had to be resolved to enable use of this technology on
aircraft engines (Foust et al., 2012):
emissions with conventional technology, and make it more difficult to apply lean
premising technology.
Climb-out
85% 2.2 min
Taxi in
Take-off
100% 0.7 min
7% 26 min
Final approach
30% 4 min
Taxi out
Figure 3.1. Illustration of ICAO landing-takeoff cycle.
Rated thrust (Foo) refers to maximum rated thrust at sea level, standard day condi-
tions, with no forward air speed.
The standards for gaseous emissions (CO, HC, and NOx) are based on the total
of each species emitted over the LTO Cycle (Dp in grams), divided by rated thrust
(Foo in kN) to account for engine size. Emissions are measured in engine tests con-
ducted as part of the airworthiness certification process, where a series of tests and
analyses are conducted under the supervision of the national airworthiness author-
ity (e.g., the U.S. Federal Aviation Administration, FAA, or the European Aviation
Safety Agency, EASA) to show that the engine meets all airworthiness, emissions,
and noise requirements for safe and environmentally acceptable operation.
During an emissions certification test, a representative engine is operated at the
four thrust levels representing the flight phases specified by the LTO Cycle. Fuel flow
and emissions of CO, HC, NOx, and smoke are measured at each thrust level, using
sampling, gas analysis, and smoke number measurement methods specified in ICAO
Annex 16, Volume II. The smoke is intended to be a measure of visibility of the
exhaust plume. Smoke is measured by passing a standard volume of engine exhaust
through white filter paper. The resulting smoke level, expressed as smoke number, is
based on the change in reflectance of the filter paper.
For gaseous emissions, the levels measured at each thrust level are expressed as
grams of CO, HC (as methane), or NOx (as NO2), per kilogram of fuel. For each spe-
cies, the mass of emissions generated during each phase of flight is calculated as
mass produced = (time in mode) (fuel flow) (emission index).
The emissions in each mode are then summed to calculate Dp for each species. Details
of emissions in each mode are reported in the ICAO data bank and on ICAO data
sheets (European Aviation Safety Agency, http://easa.europa.eu/environment/edb/
aircraft-engine-emissions.php, 2012). An example of an ICAO data sheet is shown
in Figure 3.2.
The ICAO Emissions Standards set limits on maximum Dp/Foo of CO, HC, and
NOx. For smoke, the standard sets the limit for maximum smoke number at any
operating condition. The first standards were set in 1981 and updated at the second,
88 Worldwide Aircraft Regulatory Framework
fourth, sixth, and eighth meetings of CAEP (designated CAEP/2, CAEP/4, CAEP/6,
and CAEP/8). The standards that apply to subsonic aircraft engines having rated
thrust above 26.7 kN (6,000 lb) are formally defined by several paragraphs in chap-
ter 2 of ICAO Annex 16, Volume II, but are popularly referred to by the CAEP
meeting where they were defined (e.g., CAEP/2 Standard). Standards have not
been set for smaller subsonic engines because small engines are a minor contributor
3.2 Regulatory Framework 89
Certification Production EPR Foo, kN Smoke, SN CO, g/kN HC, g/kN NOx, g/kN
160
140
ICAO LTO Cycle NOx limit, g/kN
120
100
80
60 CAEP/1
CAEP/2
40 CAEP/4
CAEP/6
20
CAEP/8
0
15 20 25 30 35 40 45 50 55
Engine pressure ratio
Figure 3.3. ICAO/CAEP NOx limits for engines above 89 kN thrust.
60
Maximum allowable smoke number
50
40
30
20
10
0
0 100 200 300 400 500 600
Engine rated thrust, kN
Figure 3.4. ICAO/CAEP smoke limit.
nonvolatile PM emissions standard. Both mass of PM emissions and particle size will
be measured.
CAEP has shown that for current aircraft combustor designs, modifications
to meet more stringent LTO NOx emissions standards provide equivalent reduc
tions at high altitude, so there has not been a strong need for separate standards
for higher-altitude climb and cruise operations. However, application of advanced
combustor concepts similar to industrial dry low emissions combustors will likely
change the relationship between low- and high-altitude emissions, and could enable
optimization of cruise emissions. In light of the potential importance of high-altitude
NOx emissions on climate and health, CAEP is revisiting means to reduce NOx
emissions at climb and cruise conditions.
When the original ICAO emissions standards were set, the EPA only adopted HC
and smoke requirements, so CO and NOx standards were not harmonized with
the requirements of other countries that adopted the entire set of standards.
In the case of the CAEP/4 standard, in November 1999, CAEP published the
standard meant to go into effect in January 2004. After the U.S. process, FAAs
Part 34 was finally amended in April 2009.
However, even though local requirements vary, engines operate around the world, so
the de facto requirement for manufacturers is to meet all current CAEP standards.
aircraft with new models, and since new engines typically operate at higher engine
pressure ratios and produce more thrust, there is incentive to use low-emission com-
bustor designs on new military engines reducing environmental impacts. Similarly,
an EIS must be prepared for a civil airport expansion project that adds capacity. In
such cases, there is an option to reduce total facility emissions by operating more
efficiently or reducing emissions from other sources such as heating or power gen-
eration facilities, road traffic, or ground support equipment.
Current aviation emissions rates are relatively high because many available
emissions reduction technologies cannot be applied within practical airplane
weight and volume limitations.
94 Worldwide Aircraft Regulatory Framework
ICAO is implementing emissions standards for CO2 and nonvolatile particulate mat-
ter and is actively investigating the impact of NOx emitted at cruise altitude. The
impact of black carbon particulate matter emissions on climate, particularly on
the melting of arctic ice is a new area of investigation within the FAA PARTNER
Environmental Center of Excellence.
In light of these current activities, efforts of understanding the impact of avia-
tion emissions, setting increasingly stringent standards, and evolving or adapting
improved emissions reduction technology are likely to continue through the next
decade.
References
Barrett, S. R. H., Britter, R. E., and Waitz, I. A. (2010). Global Mortality Attributable to
Aircraft Cruise Emissions. Environmental Science and Technology 44: 773642.
Foust, M. J., Thomsen, D., Stickles, R., Cooper, C., and Dodds, W. (2012). Development of the
GE Aviation Low Emissions TAPS Combustor for Next Generation Aircraft Engines.
AIAA Paper 20120936.
Herndon, S. C., Wood, E. C., Northway, M. J., Miake-Lye, R., Thornhill, L., Beyersdorf, A.,
Anderson, B. E. et al. (2009). Aircraft Hydrocarbon Emissions at Oakland International
Airport. Environmental Science and Technology 43: 17306.
Intergovernmental Panel on Climate Change. (1999). Aviation and the Global Atmosphere,
Cambridge University Press, Cambridge, UK.
Leonard, G., and Stegmaier, J. (1994). Development of an Aeroderivative Gas Turbine Dry
Low Emissions Combustion System. Journal of Engineering for Gas Turbines and Power
116: 5426.
Ratliff, G., Sequeira, C., Waitz, I. A., Ohsfeldt, M., Thrasher, T., Graham, M., Thompson,
T. (2009). Aircraft Impacts on Local and Regional Air Quality in the United States.
Report No. PARTNER-COE-2009002, PARTNER.
Spicer, C. W., Holdren, M. W., Riggin, R. M., and Lyon, T. F. (1994). Chemical-Composition
and Photochemical Reactivity of Exhaust from Aircraft Turbine-Engines. Annales
Geophysicae Atmospheres Hydrospheres and Space Sciences 12(1011): 94455.
Wayson, R. L., Fleming, G. G., and Iovinelli, R. (2009). Methodology to Estimate Particulate
Matter Emissions from Certified Commercial Aircraft Engines. Journal of the Air and
Waste Management Association. 59: 91100, January.
4 Overview of Worldwide Ground-Based
Regulatory Framework
Manfred Klein
Pollution prevention and energy conservation with system efficiency are key ele-
ments in arriving at cost-effective long-term solutions that address sustainability to
implement national clean energy and energy security initiatives. Low air pollution,
greenhouse gases, and water impacts are all important to local and regional areas
and can be dealt with by some degree of regulatory oversight, with trade-offs appro-
priately evaluated. International emission standards and regulatory policies for gas
turbines described here have developed over the past decade to address some of
these challenges.
Gas turbine cogeneration and district energy plants with efficient cycles and reli-
able dry low NOx combustion can provide important environmental improvements
to cleaner energy production. Until recently, GHG emissions and system energy
efficiency have not been closely studied in most permitting processes. Pollution pre-
vention planning and environmental assessments may require a more comprehen-
sive strategy, with balanced economic and environmental implementation to allow
consideration of a wide range of renewable and cleaner energy choices, including
various gas-turbine-based applications.
95
96 Worldwide Ground-Based Regulatory Framework
Air emissions
to create a combined impact on regional air quality, and they form the basis of many
important international initiatives (Figure 4.1).
In addition to air pollution, the climate change debate has led to concerns about
future increases in CO2, methane, and N2O emissions leading to a continued rise
in atmospheric concentrations. Early conferences to deal with this issue included
the 1988 event in Toronto, Canada, which led to the Intergovernmental Panel on
Climate Change (IPCC), the 1992 Rio Earth Summit on Global Climate Change and
Biodiversity conference in Brazil, and the 1997 Kyoto Protocol agreement, followed
by several other more recent international events (Copenhagen, Cancun, Durban,
and Dubai). Many of the worlds scientists now agree that there is a substantial
risk of not only higher global mean temperatures, but of consequent unpredictable
severe climatic and weather events. Recent evidence indicates that the frequency of
storms may be enhanced, and scientists are investigating certain linkages developing
between greenhouse gas interactions with ice melting, sea level rise, flooding and
droughts, and El Nino impacts (IPCC, 1992).
One of the key differences between the air pollution situation and the climate
problem is that time will allow for the cleanup of traditional forms of pollution.
However, the buildup of greenhouse gases (GHGs) and climatic change likely can-
not be reversed within a reasonable time frame. Energy-related activities are a key
source of these GHG emissions (and various air pollutants too), and thus a new
approach to energy and material conservation may be necessary, including vari-
ous forms of cleaner energy production (National Academy of Sciences, 2009; U.S.
Department of State, 2010).
Air toxics are another important type of emission produced from various indus-
trial processes and fuels. Mercury is one of the leading inorganic trace elements to be
regulated when in higher concentrations than naturally occur. Other heavy metals
such as vanadium, chromium, nickel, and cadmium are found in the emissions from
coal- and oil-fueled systems. These elements can bioaccumulate in air and water
supplies, with effects on humans and wildlife resulting in health and nervous system
4.2 Air Pollution and Greenhouse Gas Emissions from Gas Turbine Systems 97
damages. Organics produced by some types of gas turbine systems, such as ammonia
and formaldehyde, are also of concern to health, as are a wide range of persistent
organic pollutants (UNECE, 1979). Chlorofluorocarbons (CFCs) used in various
electric-driven cooling systems are linked to stratospheric ozone depletion, while
their replacements, hydrochlorofluorocarbons (HFCs), are also a strong greenhouse
gas. All of these types of trace elements occur in very small quantities, but often
from the same energy systems as those producing criteria air pollutants and CO2
emissions.
Clean and sustainable water supplies, as well as the general integrity of local
fresh and salt water, are very important to all communities for their health and their
economy. Energy industries often use large amounts of water for in-plant cooling
processes to move low-pressure steam back into warm water for boilers or HRSGs.
These condensers lose a lot of energy and cause thermal discharge into the river,
which can lead to increased water temperatures. This can harm growth and repro-
duction of sensitive organisms, stimulate the growth of algae, and decrease levels
of dissolved oxygen, also harmful to fish and other aquatic life forms. Facilities
can choose to use air-cooled condensing, which may have visible vapor plumes,
some residual particulate emissions, and some low-speed fan noise (IFC World
Bank,2007).
NOx oxides of nitrogen NO and NO2, precursors of ground-level ozone, smog, and acid
rain; formed from high temperature and pressure combustion and from N2 content in
liquid fuels;
CO carbon monoxide, a gas resulting from incomplete combustion, improper combus-
tor cooling air, and fuel mixing;
UHC unburned hydrocarbons from incomplete combustion, an indicator of trace
emissions;
PM particulates and smoke, mostly from incomplete combustion in liquid fuels, too rich
fuel-air mix, trace amounts possible from gas fuel with impurities and from SCR controls
(or swallowed with incoming inlet air);
SO2 sulphur dioxide, from S content in liquid fuel (~ 0.20.5%), or in natural gas (46
ppm), possibly with mercaptan odorants;
NH3 ammonia used in selective catalytic reaction systems for back-end NOx
reduction.
The trend toward high power and thermal efficiency, derived through high firing tem-
peratures and air compression ratios, has been translated into increased emissions of
nitrogen oxides (not nitrous oxide, which is the greenhouse gas N2O). As discussed
98 Worldwide Ground-Based Regulatory Framework
Carbon dioxide
1000
750
kg/MWhr
500
250
CCS
0
Coal Oil Gas GTCC GTCHP Bio IGCC
Because of its high hydrogen content, natural gas integrated systems can have a
CO2 rate of 220360 kg/MWhr. This represents a 6080 percent net GHG reduction
from current coal technology, with 9095 percent fewer NOx, SOx, and PM (oxides
of nitrogen, sulphur, particulate, and mercury) emissions. Onsite CHP can also have
important local co-benefits of energy process reliability, fewer power transmission
losses, and some CFC reductions with absorption chilling.
It can also be seen that both woodwaste energy and solid fuel gasification have
reasonably low levels of CO2 emissions, when carbon capture is employed in IGCC,
coupled with lower air pollutant emissions (from Figure 2.7). While carbon capture
may become a common CO2 management method, the increased use of cogenera-
tion and CHP systems may turn out to be an overall more cost-effective prevention
and efficiency concept. Fuel flexibility and gasification of solid materials into syngas
and hydrogen fuels, with CO2 capture, and liquid fuel from coal Fischer-Tropsch pro-
cesses, will play a vital role for energy security and provide economic and environ-
mental benefits for North America, Asia, and Europe.
Methane emissions are also an important GHG in the natural gas industry, with a
global warming potential of about twenty-one times CO2, depending on the timescale
assumptions. Several measures are employed to minimize CH4 leakage and venting
from gas turbine compressor stations, including dry seals on the compressors, methane
monitoring surveys, gas transfer units, and ensuring reliability for fewer blowdowns
of station piping. Waste heat recovery coupled with reliability of the gas turbine and
the pipeline system are significant elements of this strategy. Nitrous oxide, N2O, is a
more minor contributor in small amounts (GWP = 310), often produced as a result of
a low-temperature combustion or in a catalytic chemical reaction. It is important to
note that N2O as a GHG is a different type of emission than smog-forming NOx, and
this is sometimes confused in the literature and in some regulations.
100 Worldwide Ground-Based Regulatory Framework
NOx emissions criteria can be converted from one form to another, based on
the fuel and the oxygen content at the measured exhaust. Gas turbine emission con-
centrations at the stack are usually corrected to 15 percent O2, but measured at a
slightly different O2 level (1316 percent) depending on the engine efficiency. Some
source- and ambient-concentration-based rules give an incentive to dilute the emis-
sion using airflow or stack height, although excess air can be limited by defining the
standard oxygen content. Typically, a rate of x ppm can be multiplied by about two
to estimate a mg/m3 rate (i.e., 25 ppm equals about 50 mg/m3). For conversion to
fuel-based criteria, more information on fuel and airflow is needed.
While there is some advantage of simplicity for the concentration-based method,
it causes the emission to be viewed as only linked to the combustion source, the
engine, or exhaust stack, without recognizing the whole energy power plant seeks
the environmental permit. Under these criteria, there is little environmental stimu-
lus to conserve energy, as plant output and system efficiency is not directly consid-
ered. Back-end controls also tend to be encouraged because parasitic power losses
are ignored, and plant efficiency objectives are not usually considered in the analysis.
Energy conservation is clearly pollution prevention, and therefore waste heat recov-
ery and CHP could be specifically recognized as an emission prevention technology.
One can question why gas turbines are designed and regulated on the basis of
NOx concentration standards, when recip engines and cars use output-based stan-
dards such as g/kWhr or g/mile. The aviation sector employs criteria of kg of NOx
per 1000 kg of thrust for aircraft in a specified landing and takeoff cycle a practical
systems approach. Climate change risks and the clean energy situation may demand
a new way of thinking about energy systems and emissions prevention and that stack
and process emissions (and exhaust energy) be gathered and concentrated for utili-
zation and capture. Traditional emission standards and assessment methods do not
consider CO2 emissions, a necessary product of heat release. As discussed in Sections
2.4 and 4.6, many trade-offs and synergies exist that can lead either to comprehen-
sive solutions or to inconsistencies with other objectives, including water impacts,
energy efficiency, and noise issues.
102 Worldwide Ground-Based Regulatory Framework
An energy output-based standard, such as kg per MWhr, would allow the plant
designer and operator to take advantage of all available system efficiencies to reduce
fuel consumption and parasitic losses or to increase output to offset other emissions
(Bird et al., 2002). An output-based standard could allow for:
More stable and reliable operation at part load or changing ambient conditions,
if combustor mechanical design is not tied to ppm constraints, but rather to
power, fuel rate, and mass flow.
Recognition of high-specific power (kW per mass of airflow), the effectiveness
of how hot compressed air is used in the engine.
Avoidance of inefficient emission control practices, such as increased dilution,
compressor air bleed, and other fuel/air management practices. These techniques
may be unnecessary if the actual mass of NOx produced was already reduced at
lower power settings.
More flexibility and reliability in system design if applied to the complete
power plant instead of the unit, designers could optimize the whole plants effi-
ciency for power and heat, while minimizing total plant emissions. This opportu-
nity would be attractive for cogeneration systems.
In the practice of emissions trading, presenting measurements in tangible units,
such as kg of pollutant per MWhr of energy output, would clarify and compare
the real effects on the environment for a variety of nontechnical stakeholders.
Assist the promotion of industrial waste heat recovery, cogeneration, and dis-
trict energy as key energy and environmental solutions toward 8090 percent
energy system efficiency (EERE, 2009).
4.5 International Emission Rules for Air Pollution from Gas Turbines
A general survey of international policies for gas turbine emissions reveals some
similarities and differences in approaches to air pollution over the last fifty years,
and, more recently, to climate change and GHG reduction objectives from the 1992
Rio conference.
Early energy and environmental policy in the United States centered around the
Clean Air Act Amendments of 1977 and the Industrial Fuel Use Act of 1978. The
latter discouraged the use of high-value natural gas for power generation during
the two international oil crises of that decade. The first New Source Performance
Standards (NSPS) for gas turbines was issued in 1979. The primary pollutant of con-
cern had been NOx only, and the NSPS did not regulate the emissions of CO or
UHC because the levels were very low at base-load conditions. A notable provi-
sion in the NSPS was the heat rate correction to encourage efficient gas turbine
design and operation. At this time, the attainment area and non-attainment area
regional criteria for various pollutants was adopted, bringing about the Prevention
of Significant Deterioration (PSD) program.
4.5 International Emission Rules for Air Pollution from Gas Turbines 103
The best available technology (BAT) and best available control technology
(BACT) assessments and emission rules relied on state regulation, New Source
Review, and PSD determinations. In 1987, the EPAs top-down approach for deter-
mining the BACT became a requirement, reducing allowable gas turbine NOx levels
much lower than the existing NSPS (75 ppm on gas, with the heat rate correction
factor). As NOx levels decreased, with steam/water injection, CO emissions from the
large amounts of steam or water used became a concern. This was followed by dry
low NOx combustion technology advances in some large gas turbine units and new
add-on SCR emission controls to achieve very low levels of NOx without injection.
The 1990 Clean Air Act Amendments resulted in new emission control require-
ments, not only for NOx, but also for CO and VOCs in ozone non-attainment areas.
In some cases, the Lowest Achievable Emission Rate (LAER) rules for NOx and
CO have become dominant in state regulatory practice for non-attainment areas
such as California and the U.S. NorthEast Ozone Transport Region. NOx and SO2
emissions trading and offset systems have also come into play for the utilities sectors
with state implementation plans. Hazardous pollutants have also become a concern
during this time, as Maximum Available Control Technology (MACT) policies were
implemented (Schorr, 1999).
There has been much debate as to the appropriate emission standards for new
gas turbine plants, in part because of the BACT policies in the United States, which
have evolved toward ultralow NOx emission levels without significant regard to
GHG and overall system efficiency. The U.S. EPA and state regulators have usually
used concentration-based standards (ppmv at 15 percent oxygen), coupled with state
daily, monthly, or annual tonnage caps and emission offset rules. BACT practices
included steam/water injection, dry low NOx combustion technology, and add-on
SCR back-end emission controls. When LAER became dominant in certain regions,
these ultralow NOx control solutions with DLN plus SCR were often required, with
limitations on ammonia slip emissions. Over 250,000 MWe of gas turbine simple,
combined cycle, and cogeneration facilities were installed over the twenty-year
period, with varying degrees of annual load factors, system efficiency, and NOx con-
trol as low as 2 ppmv. Most gas pipeline turbine engines have been regulated in the
1542 ppm range, depending upon individual state determinations.
At the same time, for several decades the National Academy of Sciences, NASA,
the EPA, and other organizations have studied the climate change situation and
atmospheric CO2 levels. Although the specific climate change and GHG debates
are fairly recent, several regions of the United States have over the past few years
openly endorsed some types of GHG limits, cap-and-trade programs, or carbon pric-
ing mechanisms. National energy efficiency concerns are again on the rise, along
with the very important subjects of energy security and foreign policy (National
Academy of Sciences, 2009; U.S. Department of State, 2010).
In the year 2000, the U.S. EPA sent out a request for comments on BACT
choices for NOx control in combined cycle turbines. Subsequently, after some years
of consultation, the U.S. EPA released in July 2006 a new national NSPS regula-
tion for gas turbines used for pipeline compressors, utility combined cycles, and
104 Worldwide Ground-Based Regulatory Framework
<50 MMBTU/hr
> 30 MW 96 4.7
industrial cogeneration plants (Figure 4.3; U.S. EPA 2006, subpart KKKK). Key
elementsinclude:
The new rules would also give a choice for using ppm criteria, or new output-based
criteria in lb/MWhr, at somewhat higher NOx levels to encourage more energy
system efficiency.
Larger mechanical-drive GT engines over 10 MW on pipelines would only have
to meet 100 ppm NOx or 5.5 lb/MWhr, with small units and Arctic applications
at 150 ppm or 8.7 lb/MWhr, essentially uncontrolled.
Exemptions were made for small units and for gasification systems subject to
the EPA Sub-part Dd emissions legislation for coal systems (U.S. EPA 40 CFR
Part 60).
Rather than just using continuous measurement, the new rules allow for flexibil-
ity in monitoring important parameters for predictive emissions monitoring.
For gas-fired power plants, the use of toxic ammonia-based SCR control is now
being called into question because of health and safety issues.
The topic of GHG reductions is relatively new to the United States regulatory
system, although research has been developed in the United States since the 1980s.
Depending upon U.S. climate change and GHG policy developments, many of these
provisions have yet to be fully incorporated into state and regional assessment poli-
cies. However, they do point to a fairly remarkable change in philosophy of national
regulation of gas turbines in relatively clean energy applications. This new U.S. EPA
rule applies to units built after 2005, but state and local implementation may be held
up by legal challenges due to perceived pitfalls in the loosening up of NOx permit
levels, needed to achieve a more balanced sustainable energy system.
GHG emissions in the United States have grown from about 6,100 Mt in 1990 to
almost 7,000 Mt in 2010, with 80 percent being CO2, and a large 25 percent portion
4.5 International Emission Rules for Air Pollution from Gas Turbines 105
coming from the coal-based power industry. While many national programs have
developed around energy efficiency (i.e., EnergySTAR), natural gas repowering, sev-
eral IGCC demo plants, and lower-vehicle fuel consumption, the more comprehen-
sive clean energy and energy security bill proposals are just beginning to be offered
for consideration by the U.S. Congress (EIA, 2010).
4.5.2 Canada
The Canadian Council of Ministers of the Environment (part of the CCME Smog
Management Plan) published Canadian emission rules for stationary gas turbines in
1992 to promote system efficiency and reasonable pollution prevention technology
to achieve a sizeable reduction in NOx emissions. Energy efficiency to minimize CO2
emissions was deemed important, as were considerations of operational reliability
and cost-effectiveness.
A national consultation in 1991 recommended an energy output basis for the
guideline, with NOx levels directly tied to the overall demonstrated overall plant
efficiency. This is believed the first regulatory standard for the gas turbine sector
that helped to establish pollution prevention, combustion modifications, and overall
system CHP efficiency as best available technology. The guideline uses an energy
output basis for power and heat, in grams of NOx per gigajoule of energy output
(g/GJout). With a heat recovery allowance, this allows higher-efficiency engines and
systems to have a higher exhaust ppm NOx concentration and attempts to provide
incentive toward continuous improvement in system efficiency (Figure 4.4).
The guideline targets were established at a certain efficiency in each chosen size
category, for gaseous and liquid fuels. For large units > 20 MWe, the Power Output
Allowance at 140 g/GJout relates the mass of NOx emitted to the number of gigajoules
of power output (3.6 GJ = MWhrs output). This allowance results in large units,
fired on natural gas, having to meet a full-load NOx emission target of about 2733
ppmv in simple cycle applications, and 3742 ppmv in a combined cycle plant. A
higher emission level is available through the 40 g/GJ Heat Recovery Allowance to
encourage cogeneration applications. Units of 3 to 20 MW have targets set about 70
percent higher (240 g/GJ). Additional provisions for lower limits on very large gas
turbine units are now under consideration. These output-based criteria may allow
for simpler conversion of tonnage rates to economic and cost-effectiveness data, as
well as use for emission trading purposes (CCME, 1992).
The guideline was developed to promote high-efficiency applications of gas
turbines with a reasonably achievable low level of emissions of NOx and CO. It
was not generally felt that there was a need on a national basis to go to ultralow
NOx levels that would require SCR back-end cleanup and its attendant difficul-
ties. The CO limit was set at 50 ppm to ensure good combustion and reliable
operation. New plants will be required to measure their emissions of NOx and
other contaminants to document their performance relative to emissions targets,
either with continuous emission monitoring or with methods of comparable effec-
tiveness, such as steam/water injection flow rate measurement or some type of
106 Worldwide Ground-Based Regulatory Framework
Canada
Outputbased emission criteria power and heat
(grams NOx per GJoutput)
60 Heat 40 g/GJ
50
NOx (ppm)
40 Power
240 g/GJ
30 320 MW
140 g/GJ
> 20 MWe
20
10
0 20 40 60 80 100
Overall plant thermal efficiency (%)
Figure 4.4. National emission guideline for stationary gas turbines (CCME, 1992).
predictive emissions monitoring systems based on empirical data for the specific
engine (Klein, 1999).
On the climate change topic, Canada had seen a 25 percent increase in GHG
emissions since 1990, to a level of about 750 Mt/yr by 2008. To ensure GHG reduction
and comprehensive environmental measures align with policies in the United States,
federal commitments are for a 17 percent GHG reduction by 2020. Commitments
include fuel switching to renewable and low-carbon power sources, industrial and
commercial energy efficiency, carbon capture and storage, and higher-efficiency
transportation. Significant amounts of coal-based power units have been closed or
put on standby duty, and the recent 15,000 MW of installed gas turbine systems have
contributed to about 30 Mt/yr of GHG reductions to lower total annual GHG emis-
sions to the 700 Mt level by 2010 (plus about 200 kT/yr of less air pollution).
European countries have had several forms of air regulatory policy since the
mid-1970s. Various EU jurisdictions have since 1996 discussed various policies that
may integrate air pollution issues (only acid gases, air toxics, water and soil quality)
in the Directive on Integrated Pollution Prevention and Control (IPPC). The Large
Combustion Plant Directive (LCPD) first became official in 2001, and progress has
been made with revised LCPD directives in 2005, combining several directives
into one document in 2007, including strengthened BAT Reference documents
(BREFs) for atmospheric pollutants. However, industry and government still
engage in much debate on the various assessment strategies in European countries,
and to what extent the evaluations employ GHGs, system efficiency, carbon cap-
ture, and gasification solutions. Further changes to the new Directive on Industrial
Emissions are presently under review by the European Commission (IPPC, 2008).
Most EU countries still have their own individual emission policies for gas
turbines, as shown in Table 4.2. Europe has traditionally used concentration-based
4.5 International Emission Rules for Air Pollution from Gas Turbines 107
Table 4.2. Gas turbine air pollution emission standards in European countries (2005)
Source: Adapted from L. Witherspoon, Solar Turbines (personal communication, May 2005).
standards (milligrams per m3) or fuel-input-based levels (grams per GJ fuel input).
Note that size ranges are often in MW of thermal input. Previously, the emis-
sion requirements for pipeline units ranged from about 100200 mg/m3 in various
countries.
The LCPD tightened minimum emission limit values for large combustion plants
and introduced minimum provisions on environmental inspections of installations,
the review of permit granting and reporting of compliance for combustion plants
of between 20 and 50 MW thermal input. Should the 2007 Large Combustion Plant
Directive be required as regulation, limits for any fuel would be based on plants over
50 MW thermal input capacity (~1520 MW output). This would set emission limit
values for SO2, NOx, and PM for most industrial plants, and combines permits with
trading allowances for existing and new facilities with emission trading for plants
greater than 20 MWth.
For any gas-fired plants, the NOx emissions are set at a 50 mg/m3 level, or about
25 ppm. For both gas pipeline compressors and cogeneration facilities, the limit will
be 75 mg/m3 (37 ppm).
Natural gas
50 mg/m3 (simple) or 75 mg/m3 (cogeneration with 75 percent efficiency)
Combined Cycle: 50 / 35 efficiency
Mechanical drives: 75 mg/m3
Liquid and other gaseous fuels: 120 mg/m3
108 Worldwide Ground-Based Regulatory Framework
At this time there is an ongoing discussion on how EU member states will adopt
the BREF guidelines into their permitting, and how the EU GHG Emission Trading
and proposed NOx/SO2 emission trading systems will be incorporated into national
policy, with a few countries including emissions taxation on both air pollutants and
GHGs. The possible integration of these linked environmental issues may offer the
chance to provide more clarity on clean energy and energy security opportunities
while balancing as much simplicity as possible for implementation. Some flexibility
still exists for EU15 and EU27 countries that have unique economic or environmen-
tal circumstances.
The EU Emission Trading Scheme (ETS) for CO2 will involve about twelve
thousand carbon-intensive facilities in various power and industrial sectors to buy
and sell permit allowances for about 40 percent of the EU total GHG emissions.
It began in 2005 as part of the overall Kyoto commitments with national alloca-
tion plans, and will evolve into the 2020 period as Europe attempts to continue its
GHG mitigation path toward a 30 percent overall reduction. Additional tools will
include a monetary valuation of CO2, allowance auctions, and credits for carbon
capture and storage facilities. Research for CCS will come from the Zero Emissions
Fossil Fuel Technology Platform (2007) and the new CCS Directive (2009), sup-
ported by the European Turbine Network (ETN). Individual European countries
are also developing their own GHG emission programs in conjunction with the ETS
system, someof which are summarized in the following sections (UNFCC, Annex 1
Countries, 2010).
United Kingdom
The United Kingdom has significantly reduced its GHG emissions from 1990 levels
of about 780 Mt/yr of CO2e to levels near 570 Mt/yr in 2009, a 26 percent reduc-
tion caused mainly by a switch from coal-based power to natural gas and renewable
sources, and recently, to some degree, reduced economic activity. The country has
set additional mid-term targets of 34 percent by 2020 and 80 percent by 2050, and
was the first national government to enact specific low-carbon policies via its 2008
Climate Change Act.
Sweden
Sweden and other Scandinavian countries also have a long history of dealing with
climate change issues and GHG reductions stemming from the 1980s, including a
modest carbon tax enacted in 1991. It has a 40 percent reduction target by 2020,
4.5 International Emission Rules for Air Pollution from Gas Turbines 109
Germany
Although Germany has always relied heavily on coal-based power generation, it
has also been a leader in GHG and air pollution reduction policies. GHG emissions
have been reduced from about 1,200 Mt/yr to just under 1,000 Mt/yr. Presently the
2020 goals are a 20 percent GHG reduction in its Integrated Energy and Climate
Programme, but a 30 percent objective is now under discussion for a 270 Mt/yr
reduction target. Carbon capture, cogeneration, and other cleaner electricity direc-
tives are being introduced for a 100 Mt/yr reduction opportunity for thermal energy
by 2020, some of these involving gas turbine systems. Germany eliminated taxation
for natural gas systems used for power production in 2006.
France
Policy is based on the 2004 Climate Plan and the more recent Grenelle Environment
Forums of 2007 and 2010. Because of high reliance on nuclear power and some
hydro, Frances energy-related emissions are quite low, about 13 percent of the
national total of about 530 Mt/yr, a slight decrease from 1990. Among various poli-
cies in transportation, nuclear efficiency, and renewables, there are ongoing discus-
sions about a proposed carbon tax.
Italy
The Italian White Certificates system is a cross-sectoral initiative that promotes
energy efficiency in all the energy end-use sectors by 2020, including use of cogen-
eration and some combined cycle capacity. The Green Certificates system aims to
attract additional renewable energy capacity to reduce GHG emissions, which have
risen slightly from 1990 (516 Mt) to present levels of about 550 Mt/yr.
Japan
Japan has a varied mix of energy options with fossil fuel, nuclear, and some hydro,
although most fuels such as oil and natural gas are imported. The nuclear power
situation has now become very uncertain. While the country has always had very
strict air pollution regulations, its GHG emissions have risen from 1,200 Mt in 1990
to almost 1,400 Mt in 2008. Several mitigation policies are being discussed, and
notably the first coal gasification plant has been built at Nakoso in 2010. In terms of
gas turbine NOx emissions, the 1992 rules required about 30 ppm in populated areas
and 42 ppm elsewhere. Newer, stricter rules have often followed the U.S. BACT
policies, requiring very low DLN designs or back-end SCR systems.
110 Worldwide Ground-Based Regulatory Framework
Australia
The Australian Environmental Protection Authority Guidance Statement (May
2000) addresses emissions of oxides of nitrogen from gas turbines and recommends
DLN combustion systems based on its 1985 national guidelines. Gas-fired units over
10 MW required a 34 ppm limit, gas and oil units under 10 MW had a 44 ppm limit.
A more recent publication entitled the Protection of the Environment Operations
(Clean Air) Regulation of 2002, has NOx emission limits of 25 ppm for gas firing and
45 ppm for distillate oil fuel (Western Australia, 2000).
Selection of the best power generation technology for the fuel chosen to balance
the environmental and economic benefits, that is, the use of higher energy-efficient
systems, such as combined cycle gas turbine systems for natural gas and oil-fired
units, and supercritical, ultra-supercritical, or integrated coal gasification com-
bined cycle (IGCC) technology for coal fired units, respectively;
Considering use of combined heat and power (CHP, or cogeneration) facilities.
By making use of otherwise wasted heat, CHP facilities can achieve thermal
efficiencies of 7090 percent, compared with 3245 percent for conventional
thermal power plants.
In 2007, the International Finance Corporation of the World Bank updated this
policy with a new document, General Environmental, Health, and Safety (EHS)
Guidelines, for industrial sectors. (IFC World Bank, Table 1.1.2, 2007). New rules for
gas turbine NOx emissions are summarized in Table 4.3.
How clean are these plants, what are the fundamental issues for them, and how
important are low NOx emissions compared to other environmental concerns,
such as GHGs, toxics, and water impacts?
Should energy conservation be a more integral part of the environmental regu-
latory strategy, and is there a need to integrate plant efficiency and system reli-
ability considerations into the permitting rules can waste heat use be treated
the same as renewable energy?
Should ultralow NOx concentration reductions methods based on back-end
control take precedence over more comprehensive pollution prevention
methods?
Many of the environmental solutions for GHG reductions and energy security
make economic sense regardless of the degree of proof in anthropogenic climate
change what used to be termed no regrets measures. Best available technology
considerations for EAs will differ greatly depending on the objectives and environ-
mental issues to be mitigated and the extent to which prevention and conservation
is encouraged rather than controls and dilution. Discussion on these topics can lead
to improved clarity on what constitutes clean energy, some aspects of which are out-
lined in Figure 4.6.
A system-based approach allows the integration of our seemingly complex
industrial and community energy infrastructure, transportation, water, and waste
management systems. Industrial ecology principles as reflected in the natural step
and natural capital approaches recognize that all of these systems are intercon-
nected, and certain actions can lead to cost-effective and synergistic solutions. These
would optimize internal energy and materials flows so as to minimize imports of
112 Worldwide Ground-Based Regulatory Framework
Economic performance
Figure 4.5. Balancing several important objectives.
Low air pollution, GHG emissions, air toxics and water impacts
natural gas, conserve water, reduce wastes, and support CO2 recovery (Figure 2.9).
Processes can be intensified and coupled with recycled streams in polygeneration,
and energy quality can be evaluated as an integrated system. All losses could be
measured, and they can be recycled or used as energy input where possible using
temperature and pressure recovery.
source or fuel, as well as plant operation. This is often termed full fuel cycle analysis,
and is frequently associated with integrated resource planning, life cycle analysis of
products, and externality analysis as an economic impact tool.
Fuel Delivery
Fuel Production
Fuel Processing
End-Use Combustion
Because of the interaction and trade-offs of emissions throughout the cycle, there
are several reasons for EA practitioners to be aware of the full fuel cycle of air pol-
lution and GHG emissions. The following points relate to the natural gas industry,
for example:
The pollution prevention improvements in DLN combustion and waste heat recov-
ery for gas turbines over the past twenty years has been very successful, likely
unmatched by any other major industry. Given the existing mix of high-emission
energy sources, the ecosystem will likely not notice the difference between engines
emitting a moderate 2540 ppm level of NOx and those forced to further limit emis-
sions to single digit levels of 29 ppm. The latter come with high cost and low mar-
ginal benefits, especially if back-end control with ammonia and selective catalytic
114 Worldwide Ground-Based Regulatory Framework
reduction (SCR) is used after DLN combustion. Examples have varying degrees of
environmental profiles and benefits:
NOx and CO2 emissions often increase in opposite directions, and smaller
combustors can be allowed a higher NOx level for efficient CHP applications
with a high heat-power ratio with a lower GHG profile.
Back-end controls (SCR) have generally been shown to be less cost-effective
than pollution prevention measures, as they often give rise to other collateral air,
water, or safety impacts as well as efficiency losses and increased GHGs.
Pipeline system upsets can result from unreliable DLN combustion, causing
problematic shutdowns with stops and starts and station blowdowns.
For gasification, hydrogen-rich fuels have significant flame speed, autoignition,
and flashback characteristics in high-pressure combustion, and are often used
with nitrogen or steam dilution to minimize NOx emissions. Because this results
in a very clean and integrated coal-syngas energy solution, there may not be
a need for very low NOx requirements, which can lead to inoperability, safety
issues, or the need for additional back-end SCR systems.
the prevention of all GHG, toxic, and CAC air emissions and cooling water
impacts;
long power transmission lines and T&D losses;
the impending phaseout costs of CFC chillers;
the energy security provided by district energy loops;
process reliability of onsite generation with two or more units.
It is common to quantify the monetary impact on GHGs without also adding the
common reductions in regional acid rain, smog, and air toxic emissions, with less
cooling water usage. Clean-fueled CHP, renewables, and waste heat-to-energy can
provide all of those benefits, and portions of capital and operating costs could be
allocated to each emission reduction to show multi-pollutant $/ton cost-effectiveness.
Following are two typical examples of energy or environmental choices, comparing
cogeneration with separate production on a reduction value basis, and assessing the
relative merits of SCR and DLN air pollution control with a total damage costing
valuation.
Consider the simple case where the customer needs 1 MWe of electricity and 1
MW thermal energy (~ 4 GJ). Separate production will use 10 GJ of mixed fuel in a
boiler to make 3.6 GJ of purchased import electricity from a utility, and another 5 GJ
purchased natural gas fuel for 4 GJ of steam heat (loss = 7 GJ). Cogenerated power
and heat would require more capital cost and gas fuel, but would avoid the 1 MWe
of power purchase. The total reductions in air pollution and CO2 in the CHP case
would be 90 percent and 50 percent respectively, each day saving 0.16 tons of CAC
air pollution and 13 tons of CO2.
The total daily value of these reductions would be ($320 + 260) = $580 (if there is
an emissions trading or other economic driver for these savings). With an 80 percent
annual load factor, this would be worth $170,000 per year, a significant improvement
in the total payback value of the CHP system. Valuation of CFC cooling avoidance
and process and power grid reliability could also be worth several thousands, so that
a net economic comparison would show the CHP operation to have much higher
value than the separate purchase of electricity and gas. A key question is the nature
or profile of the avoided electricity import.
116 Worldwide Ground-Based Regulatory Framework
Separate production
1 1
10
10 GJ fuel 1 MWe
Power 5 Process
coal/oil/gas
boiler 5 GJ gas fuel heat 4
(36%) (80%)
3.6
1
10 GJ gas fuel
500 kg CO2 4
0.5 kg CAC Gas turbine WHR
1 MWe (36%)
3.6 1
2.NOx Reduction with Dry Low NOx and Selective Catalytic Reduction.
A common trade-off occurs with considering SCR systems for additional NOx control
after a DLN gas turbine system. The slightly lower NOx can be judged against the
collateral emissions and safety of ammonia-based SCR, when additional NH3, PM,
and GHGs would be created, as in the simple example of Figure 4.8 (a 300 MWe
combined cycle, 2 million MWhrs/yr). When all of the emission damages are valued
in a simple sense to gauge environmental and health impact, the total environmental
cost can be judged for design evaluation and future financial risk. As the various
contaminants and GHGs are totaled for each case, the DLN without SCR option
has a less costly environmental footprint.
Such comprehensive allocations can be used for planning purposes, but are sensi-
tive to assumptions around what emissions are avoided and the energy mix replaced.
They would allow stakeholders to clearly deal with and balance whatever important
trade-off issues may be included in the project objectives. Valuations added together
would assist in defining a common denominator for these choices, sometimes regard-
less of the actual monetary amount for each issue.
4.9 Summary
Todays design of a modern gas turbine plant to replace aging coal and heavy oil
power generation represents a dramatic 9099 percent decrease in total regional air
pollutants, as well as reduced carbon dioxide and cooling water discharges. Together
with conservation and all types of renewable energies, the gas turbine industry using
4.9 Summary 117
Ammonia 50 250 0 0
(5 ppm)
PM2.5 50 250 0 0
$15440 K $15200 K
clean hydrogen-based fuels, cogeneration, and solid fuel gasification with CO2 cap-
ture and storage represents a reasonably sustainable solution for at least the next
fifty years.
A definition of best available technology for some environmental issues (such
as only NOx) may not be consistent with best practices with respect to other envi-
ronmental issues. Best practices and BAT may vary by application and will differ
greatly depending on the objectives and environmental issues to be mitigated and
the extent to which prevention and conservation is encouraged rather than back-end
controls and dilution. System characteristics and technical choices for various sec-
tors will determine how the balancing act of low-criteria air pollutants, greenhouse
gases, and air toxics can be optimized for compression, combustion, turbine output,
and heat recovery in efficient gas turbine energy systems. A recent NAS excerpt:
Market forces have, for the most part, guided the development of the current United
States energy system. But to date, they have undervalued the changes necessary for
movement toward sustainable energy supply and use, such as the environmental costs
of burning fossil fuels and dependence on imported fuels. Decisions about future energy
options require technology choices that involve a complex mix of scientific, technical,
economic, social, and political considerations. A key message from Americas Energy
Future is that a suite of current and emerging technologies have the potential to move
the nation towards a more secure and sustainable energy system. (National Academy of
Sciences and National Academy of Engineering, 2009)
Abbreviations
BACT Best Available Control Technology
BREF BAT Reference documents
CO Carbon Monoxide
CFC Chlorofluorocarbons
CAC Criteria Air Contaminant
CCS Carbon Capture and Storage
CCME Canadian Council of Ministers of the Environment
CHP Combined Heat and Power
DLN Dry Low NOx
EIA Energy Information Administration (United States)
EPA Environmental Protection Agency (United States)
EU European Union
ETS Emission Trading Scheme
ETN European Turbine Network
EHS Environmental, Health, and Safety
GJ Gigajoule
GHG Greenhouse gases
GT Gas turbine
GWP Global Warming Potential
HRSG Heat Recovery Steam Generator
HFC Hydrochlorofluorocarbons
IPCC Intergovernmental Panel on Climate Change
LAER Lowest Achievable Emission Rate
MACT Maximum Available Control Technology
MWhr Megawatt hour
IGCC Integrated Gasification Combined Cycle
LCPD Large Combustion Plant Directive
IPPC Integrated Pollution Prevention and Control
NASA National Aeronautics and Space Administration (United States)
NOx Oxides of Nitrogen
NSPS New Source Performance Standards
NH3 Ammonia
ppm Parts Per Million by Volume
PM Particulates
SCR Selective Catalytic Reaction
T&D Transmission and Distribution
UHC Unburned Hydrocarbons
UNECE United Nations Economic Commission for Europe
UNFCC United Nations Framework Convention on Climate Change
References 119
References
Bird, J., DePooter, K., and Klein, M. (2002). Investigations into the Reporting of Gas Turbine
Emissions. National Research Council of Canada, March.
Canadian Council of Ministers of Environment (CCME). (1992).National Emission
Guidelines for Stationary Combustion Turbines. Canadian Council of Ministers of
Environment, CCME-EPC/AITG-49E, December (ccme.ca/assets/pdf/pn_1072).
Energy Information Association (EIA). (2010).EIA Annual Energy Outlook 2010. U.S.
Energy Information Administration, DOE/EIA-0383(2010) April.
European Commission. (2006).Integrated Pollution Prevention and Control Reference
Document on Best Available Techniques for Large Combustion Plants. July.
European Parliament. (2001).Directive 2001/80/EC. The European Parliament and the
Council of 23 October 2001, on the limitation of emissions of certain pollutants into the
air from large combustion plants. (GT units, pgs. 2123) OJ L 309, 27.11.2001.
European Turbine Network (ETN). (2010).EU Emissions Policies across the Member States.
September.
European Union (2004).Directive 2004/8/EC of the European Parliament and the Council of
11 February 21. Official Journal of the European Union.
IFC World Bank. (2007). Environmental, Health, and Safety Guidelines General EHS
Guidelines: Environmental Air Emissions and Ambient Air Quality Section
1.0 Environmental, April 30. http://www.ifc.org/ifcext/sustainability.nsf/Content/
EHSGuidelines.
Intergovernmental Panel on Climate Change (IPCC). (1992). Climate Change 1992,
Supplementary Report to the IPCC Scientific Assessment. Report prepared for IPCC
by Working Group I.
(2008). The IPPC Directive, Summary of Directive 2008/1/EC concerning integrated pol-
lution prevention and control www.ec.europa.eu/environment/air/pollutants/stationary/
ippc/summary.htm.
International Energy Agency (IEA). (2009). Power Generation, chapter 6 in World Energy
Outlook 2009.
Klein, M. Environmental Benefits of High Efficiency, Low Emission Gas Turbine Facilities.
Paper for CEA Conference, Toronto, April.
(1999). The Need for Standards to Promote High Efficiency, Low Emission Gas Turbine
Plants. ASME/IGTI Paper 99-GT-405, Indianapolis, June.
National Academy of Sciences and National Academy of Engineering. (2009). Americas
Energy Future (sites.nationalacademies.org).
Office of Energy Efficiency and Renewable Energy (EERE). (2009).Combined Heat and
Power A Vision for the Future. U.S. Department of Energy, August, pg. 18.
Schorr, M. et al. (1999). Gas Turbine NOx Emissions-Approaching Zero Is it Worth the
Price. GE Electric Power Systems, NY, Paper GER 4172.
Solar Turbines Tool Box Unit Converter. (http://mysolar.cat.com/cda/layout?m=43042&x=7).
UNECE. (1979). The 1979 Convention on Long-Range Transboundary Air Pollution on
Heavy Metals.
United Nations Framework Convention on Climate Change. (UNFCC). (2010). Annex I,
National Communications, Annex 1 Countries, January 1. http://unfccc.int/national_
reports/annex_I_natcom/submitted_natcom/items/4903.php.
U.S. Department of State. (2010). U.S. Climate Action Report 2010, U.S. Department of State,
Global Publishing Services, Washington, DC, June.
U.S. EPA. (2006). U.S. EPA Clean Energy Environment Guide to Action, Section 5.3:
Output-Based Environmental Regulations to Support Clean Energy Supply, April.
(2006). Standards of Performance for Stationary Gas Turbines. U.S. EPA Code of Federal
Regulations, 40 CFR Part 60, docket EPA OAR-20040490, pgs 38482506, July.
40 CFR Part 60 Subpart Da Standards of Performance for Electric Utility Steam
Generating Units.
120 Worldwide Ground-Based Regulatory Framework
5.1Introduction
Carbonaceous materials emitting from the exhaust of gas turbine engines are fre-
quently referred to as soot emissions, nonvolatile particulates, or smoke. Frequently,
these terms are used interchangeably. Such emissions typically consist of single par-
ticles ranging from 1080 nanometers that may agglomerate into a complex fractal
chain structure with much larger dimensions. A series of photomicrographs from
transmission electron microscope (TEM) analysis at different levels of magnifica-
tion for a combustor operating at 80 percent power is shown in Figure 5.1.
These carbonaceous soot particles should be contrasted with volatile particu-
lates (see Chapter 6), although volatiles may condense onto soot particles. In partic-
ular, soot particles can act as carriers of condensed polycyclic aromatic hydrocarbons
(PAHs), some of which are carcinogens. Additional discussion of these subjects can
be found in this chapter and in Chapter 6 of this book. Recent results indicate that
the morphology of these soot particles may change with power levels and even fuel
makeup (Anderson et al., 2011).
It can be estimated that, worldwide, aircraft emit approximately 25 million
pounds of carbonaceous particulate matter into the atmosphere each year, although
there is likely a factor of three uncertainty in this estimate. Higher levels exist for
emissions of total (volatile plus carbonaceous) particulate matter. Virtually all of the
particulate matter emissions in the United States and its fate in the ambient environ-
ment are subject to the new PM2.5 standards set by the Environmental Protection
Agency (EPA) in the United States for local ambient particulate levels. Local
resistance to expansion of airports is growing due to a possible increased impact
of aircraft emissions on ambient particulate levels. Historically, only the empirical
smoke number has been used for regulating emissions from commercial and mili-
tary aircraft. This empirical information is insufficient to determine local emissions
levels. Over the past five to ten years, the concerns regarding increased ambient
levels of PM 2.5particulates have led to NASA-funded studies (Wey et al., 2006,
123
124 Particulate Formation
100 nm 20 nm 5 nm
Figure 5.1. Photomicrographs of soot particles extracted from the exhaust of aircraft engines
at different magnification (Anderson et al., 2011).
Emission regulations for aircraft are set by the International Civil Aeronautics
Organization (ICAO) utilizing a standard landing takeoff cycle (LTOC), as described
in Chapter 3. Originally, emission limits were developed based on exhaust plume vis-
ibility and had no direct quantitative link to health issues. Hence, smoke emissions
from commercial aircraft are regulated by limiting the maximum Smoke Number
(SN) in the exhaust, based on an empirical scale from 0 to 100, linked to plume
5.1 Introduction 125
Figure 5.2. Visible smoke emissions looking along length of exhaust plume for legacy (circa
1960s), uncontrolled engines during takeoff. Image courtesy of William E. Harrison.
visibility. When soot/smoke emission limits were first established in the 70s, substan-
tially less information was available on health and other environmental effects. In
addition, there was generally poor understanding of the control of particulate emis-
sions and trade-offs against other engine requirements.
The maximum SN value allowable for a given engine depends on its exhaust
diameter (see Chapter 3). It is essentially a control on limiting visible emissions from
the aircraft such that smoke is barely discernable by eye when looking across the exit
plume at conditions of maximum sooting levels. Typically, this condition occurs dur-
ing takeoff. To limit the perceived visible smoke emissions, larger-diameter engines
must meet lower SN limits (<10), whereas small engines and auxiliary power units
(APUs) have more relaxed constraints (>25, for example). Hence, per pound of fuel,
large engines are substantially cleaner burning from the soot/smoke point of view.
Because of the manner in which smoke emissions are regulated, it is not unusual
to see visible soot emissions when looking along the length of the plume. Under
such conditions, the optical path length is much greater, and hence smoke emissions
can be observed. Extreme examples of such views are available from legacy (1950s
era) military engines (e.g., TF-33 on B52s), which have uncontrolled emissions (see
Figure 5.2).
Smoke Number (SN) is measured according to SAE/ARP 1179c, the standard
smoke evaluation test based on staining of a Whatman #4 filter by sample size of
0.0230 lbm/in2 of exhaust gas of filter area. The SN is reported on a scale of 0 to 100
(as shown in Figure 5.3). The measurement method typically can achieve precision
of +/3.
This diagnostic provides only empirical information on soot particulates. Hence,
scientists have conducted a variety of studies to relate SN data to quantitative emis-
sions of soot. This topic is discussed in subsequent pages.
Technical issues and varying international concerns must be balanced. None of the
measurement methods truly have the desired accuracies, and significant sample-line
issues exist because of rapid diffusive loss of small particles to the sample-line walls
that destroys particle number and mass (Liscinsky and Hollick, 2010). Meanwhile,
126 Particulate Formation
Reflectance
filter paper
(%)
0.023 lbm exhaust gas/
sq. in. filter at 0.5 cfm 0
0 Smoke number 100
Figure 5.3. Smoke measurement by SAE/ARP 1179c.
the United States prefers a mass-based measurement standard while Europe focuses
on number-based standards, as indicated by standards imposed for diesel emis-
sions (EURO 5 in 2009 and EURO 6 for 2014). Note that mass-based diagnostic
approaches (when average particles are greater than 30 nanometers) should have
lower fractional error, since smaller particles lost to walls have low mass. Still, we
must recognize that number-based approaches may well be a better metric related
to health impact.
Submicrometer particle size distribution (aerodynamic size) and particle con-
centration can be obtained using Differential (Electrical) Mobility Analysis (DMA).
Typically, a Scanning Mobility Particle Sizer (SMPS) electrostatic separator or clas-
sifier (see Figure 5.4) is coupled to a condensation nucleus counter (CNC). The elec-
trostatic separator is a standard aerosol characterization instrument that separates
particles according to their electrical mobility based on the equivalent aerodynamic
diameter. After the particles are separated by size, they are counted to obtain size
distributions. By assuming a soot mass density of 1.8 grams/cc and spherical particles,
total particle mass can be estimated from the particle size distribution data.
Alternatively, soot mass can be estimated using multi-angle absorption photom-
etry (MAAP). This technique monitors the optical properties of soot as it deposits
onto filter paper to deduce soot mass. Additional discussion of the methods and
sampling issues can be found in Marsh and colleagues (2010).
As described previously, smoke emissions from commercial gas turbine engines are
regulated via the Smoke Number. Historical data for engines are provided in the
engine exhaust data bank (ICAO, 1995). However, these data files are awkward to
use directly for estimating total engine or fleet emissions of soot, as the manufac-
turer needs to report only the maximum SN, as measured at any power point, yet the
operating point at which this maximum is reached is not reported. Maximum soot
levels are not always attained at takeoff conditions and emissions from all operating
5.1 Introduction 127
Sheath air in
(Charged)
Particle-laden
exhaust stream in
High voltage
rod
Excess air
Monodisperse stream to
condensation nucleii counter
Figure 5.4. Drawing of electrostatic classifier. Image courtesy of TSI Incorporation.
points are not provided. Furthermore, the SN is an empirical value related to smoke
emissions rather than a quantitative measure of smoke mass or number emissions.
Recent studies (Wayson et al., 2009) have needed additional information to make
estimates of total emissions in and around airports. Hence, engine manufacturers
have provided such data sets to analysts. Furthermore, there has been a need to
understand quantitatively the reported SN values.
Many studies have attempted to quantify a link between this empirical param-
eter (SN) and fundamental properties of soot. Prior studies (see, for example,
Champagne, 1971; Eckerle and Rosfjord, 1987; Hurley, 1993) have indicated a link
to total soot mass emissions. Despite a significant amount of prior work, uncertainty
remains regarding the relationship between soot mass and smoke number. Reported
correlations developed in the early 70s and 80s are reproduced as a set of curves
in Figure 5.5.
Substantial variability appears in this set of curves. If one had reported SN data
from an engine and needed to compute soot mass emissions, then the uncertain-
ties in Figure 5.5 lead to a factor of up to three in the uncertainty in predicted mass
emissions.
Researchers have speculated that part of the uncertainty in Figure 5.5 may be
due to variations in particle size. Using experimental data collected from a combustor
rig at UTRC (Colket et al., 2003), no relationship between SN and primary particle
size could be confirmed. Soot mass was not measured directly; however, the mass was
estimated from the algorithms within the software provided by the manufacturer for
the scanning mobility particle sizer (SMPS), assuming the particles detected by the
SMPS were each primary, spherical particles with a density of 1.8 grams/cc. While
the assumption of single primary particles is not consistent with prior understanding
of soot particle emissions, the data sets collected by Colket and colleagues support
arguments that at least some of the particle agglomeration occurs in the sampling and
128 Particulate Formation
10
Wood, A.D., 1975
Eckerle and Rosfjord, 1987
8
Norgren and Ingebo, 1975
Soot mass (mg/m3)
0
0 5 10 15 20 25 30
Smoke No.
Figure 5.5. Literature correlations relating soot mass and smoke number.
When sampling soot particles, investigators must consider a variety of loss mech-
anisms. For example, a large fraction (>90 percent) of particles less than 10 nm
in diameter can be lost due to diffusion to walls in typical gas sampling systems.
Fortunately, this loss does not largely impact the total soot mass, but will effect
total particle numbers. A general discussion of the characteristics of aerosols (solid
or liquid particles suspended in a gas) is beyond the scope of this chapter, but
excellent introductions to the general topic are available (Hinds, 1982; Willeke and
Baron, 1993). Applications of the related physics to particle line loss in probes and
sampling lines can be found in Liscinsky and Hollick (2010).
5.1 Introduction 129
60
Champagne, 1971 (filtered)
Champagne, 1971, unfiltered
50 Wood, A.D., 1975
Norgren and Ingebo, 1975
Soot mass (mg/m3)
10
0
0 10 20 30 40 50 60 70 80
Smoke No.
Figure 5.6. Correlations between soot mass and SN.
1.0
0.9
Transport efficiency
0.8
0.7
0.6
0.5
0 20 40 60 80 100 120 140 160 180 200
Particle diameter (nm)
Figure 5.7. Comparison of computed and experimental particle transport efficiency (Liscinsky
and Hollick, 2010).
Particle line loss issues can be viewed in Figure 5.7, in which computed losses due
to diffusion, thermophoresis, electrostatics, and so forth are delineated for particles of
different size and for typical sample line under typical operating conditions of aircraft
exhaust measurements.This figure shows (1) very significant losses in typical sample lines
due to diffusional losses of the small particles (< 30 nanometers) to sample-line walls;
and (2) if careful assessment of sampling conditions (flow rates, geometry, pressures,
etc.) are performed and documented, particle losses can be reasonably estimated.
130 Particulate Formation
Surface growth
Inception Oxidation
& coalescence
Alkylated-
CxHy
aromatics
PAH
Jet A H H2 CO
C2H2
C2H2 N aphthenes
Spray Gas-phase Ageing & Emissions
vaporization kinetics coagulation
Figure 5.8. Physical process of soot formation.
acetylene and PAH, and the main reactants in soot oxidation are normally OH and
O2. In most combustion systems, oxidation by OH radicals appears to dominate.
Reactive species produced by fuel-rich combustion lead to the formation of aro-
matic species (e.g., benzene, naphthalene, phenanthrene, pyrene, etc.) and acetylene.
The aromatic species lead to inception, while acetylene is recognized as a key growth
species. PAH also contributes to surface growth (Benish et al., 1996). Inception gen-
erally occurs primarily near the reaction front, while surface growth occurs in the
post-flame (fuel-rich) gases. Temperature is a key primary variable: at low tempera-
tures (e.g., <1500 K), kinetics are not fast enough to support rapid ring formation,
and at elevated temperatures (>1900 K), the ring structure is thermodynamically
unstable and ring growth is slowed. The H/C atomic ratio of the fuel is important, as
is the overall equivalence ratio, since these parameters affect the carbon available
for growth as well as the local temperature.
Appel and colleagues (2000) proposed that inception occurs when aromatic
structures dimerize to form co-planar structures. Strong evidence indicates that incip-
ient particles may be characterized as liquid droplets (see discussion in Bockhorn
etal., 2009). Liquid particle droplets soon undergo carbonization (Dobbins, 1996)
and release hydrogen, while particle density decreases. Particle mass and size
increase downstream in the post-flame zone due to a combination of surface growth
and coalescence. It is important to note that there are two types of particle-particle
interactions. In early stages, at least one of the particles is liquid-like (Bockhorn
et al., 2009), and the collision results in a coalescence of the two particles into a
single, nearly spherical particle with surface area significantly lower than the two
separate particles. The second type of collision occurs between two nearly solid par-
ticles, resulting in negligible loss to total surface area and producing a binary (or
aggregate) structure and, upon continued collisions with other particles, produces a
large agglomerate, such as that shown in Figure 5.1. Balthasar and Frenklach (2005)
have proposed an alternative hypothesis of sequences leading to nearly spherical
particles, based solely on solid-solid interactions and surface growth. Regardless, the
physical processes associated with aerosol dynamics are critical in defining total sur-
face area as well as particle growth and oxidation and thus the mass of soot present
in or emitting from a flame.
The characteristic time scale of a flame and related heat release processes is
much faster than the soot growth steps. The flame creates the environment such
132 Particulate Formation
that soot formation takes place in the post-flame region of a premixed flame or the
fuel-rich side of a non-premixed flame. In practical flames, turbulence and air addi-
tion continually mix and alter the local conditions. Hence, the scenario depicted pre-
viously can be considered highly idealized relative to that occurring in a combustor
wherein all processes may occur simultaneously.
Quantitative modeling of soot formation and oxidation in practical (turbu-
lent, diffusion) flames is difficult because of the complexities of the physical and
chemical processes involved; it remains one of the greatest challenges to computa-
tional modeling of combustion. First, formation of soot is an incompletely under-
stood process. Models do provide quantitative predictions of simple laboratory
flames burning pure fuels at atmospheric pressure, but generally they are accurate
over very limited conditions. Typically, existing models poorly predict trends over
a large range of experimental conditions. Second, surface growth and oxidation
rates of these particles depend on accurate knowledge of the active surface area of
the particles (Woods and Haynes, 1994; Appel et al., 2000), which in turn depends
on the formation, ageing, and collision processes. Third, the total soot emissions
from a practical burner are the difference between two large terms, the formation
and the oxidation, neither of which is known well. At full power, for example, exit
plane soot emissions can be two to three orders of magnitude less than the levels
in the primary zone (Hurley, 1993; Brundish et al., 2003), and within the fuel-rich
portion of the combustor 10 percent or more of the fuel carbon may be temporar-
ily converted to soot. Finally, turbulent mixing and reacting flow complicate simu-
lations with time scales of soot formation and oxidation substantially different
from the time scales of heat release. Tuning is often required to match the model
with data, and, even then, significant discrepancies arise in attempts to describe
soot emissions over a range of combustor conditions. Hence, it is not surprising
that authors are willing to present modeling results that agree only to within one
to two orders of magnitude from the experimental values (Brocklehurst et al.,
1997; Tolpadi et al., 1997). For laboratory flames, the models proposed by Appel,
Bockhorn, and Frenklach (2000) are typically employed, but even for such flames,
variations in empirical parameters or key rate parameters are required to enable
quantitative comparisons, even for such simplified flames (e.g., Marchal et al., 2009 ;
Zhang et al., 2009).
The following sections provide more details of gas-phase chemical kinetics, soot
nucleation, growth, and oxidation, as well as aerosol processes, with a focus on tradi-
tional modeling methods.
5.3Inception
Aromatic rings participating in the soot inception and soot growth steps can be
formed from cyclization of lower molecular weight hydrocarbons, produced from
dehydrogenation of cylcoalkanes, or provided by the parent fuel. It is now believed
that, for many flames fueled by low molecular weight hydrocarbons, the dominant
5.3 Inception 133
H + C 6 H 6 (benzene) C 6 H 5 (phenyl) + H
C 6 H 5 +C 2 H 2 C 6 H 5CHCH C 6 H 5CCH 2
C 6 H 5CCH 2 +C 2 H 2 =C 6 H 5C(CHCH)CH 2 => C 9 H 6 CH 2 (methyleneindene)+ H
C 9 H 6 CH 2 (methyleneindene) => C 10 H 8 (naphthalene)
Such steps are analogous to those steps involving species with radical sites on inte-
rior carbons in the formation of the first aromatic.
An alternate simulation of the inception process is utilized by Smooke and
colleagues (2005), who employ a series of steady-state assumptions on interme-
diate species to estimate the formation of a large polycyclic aromatic structure.
134 Particulate Formation
Fragmentation Fragmentation
+C2H2
CC-H
C=C
H H
Figure 5.9. Reaction pathway for formation of two-ringed aromatics (Frenklach et al. 1985).
For any computation method for modeling soot production in cases for which
a substantial amount of the fuel carbon is converted to soot, the source terms
for conversion of gaseous species to soot must be accounted for by including
appropriate terms in the gas-phase species equations. Likewise, an enthalpy term
should be added to the gas-phase energy equation because of the formation
of soot.
Perhaps the most realistic or practical inception model utilized today is
pyrene-pyrene dimerization as suggested by Frenklach and Wang (1990). The mass
production rate is described by:
C 14 H 10 + C14 H 10 = soot
dm
= 2 MWC14H10 k[C 14 H 10 ]2
dt
5.4Surface Growth
In premixed flames, surface growth has been shown to be first order in acetylene
concentration (Harris and Weiner, 1983). This observation resulted in the creation of
many models based on acetylene addition (e.g., Fairweather et al., 1985; Frenklach
and Wang, 1990; Colket and Hall, 1994). Colket and Hall utilized a surface growth
model in numerical simulations of non-premixed flames based on the premixed flame
data of Harris and Weiner (1983). The most widely used surface growth model is that
developed by Frenklach and Wang (1990) and demonstrated by Appel, Bockhorn,
and Frenklach (2000) and many others. It is referred to as the Hydrogen-Abstractio
n-Carbon-Addition (HACA) model (Table 5.1).
The Arrhenius expression is assumed to be k = A Tn exp(E/RT), where the
pre-exponential, A, is given in units of cc, mole, and seconds, and activation energies,
E, are given in kcal/mole. For this reaction set, only the first two were considered
reversible.
5.4 Surface Growth 135
Table 5.1. HACA mechanism for surface growth as utilized by Appel and Colleagues (2000)
Table 5.2. Modified version of the Frenklach and Wang soot growth mechanism as developed
by Colket and Hall (1994)
Appel and colleagues proposed that the overall surface growth rate would
depend on the fraction of surface area available for surface growth, which they
denoted as . They proposed that:
= tanh ( log1 + b)
where 1 is the first moment of the soot particle distribution and a and b are fit-
ted parameters determined separately for each experiment. Zhang et al. (2009)
instead utilized a value suggested by Xu et al. (1998), which depends on temperature:
= 0.004exp(10,800/T) but also pointed out that there are no universal values for
that work for all experiments.
An alternate reaction sequence, proposed by Colket and Hall, is reproduced in
Table 5.2. For this reaction model, oxidation steps are separately treated. The net
specific rate of soot mass growth (grams/second/cm2) via this sequence can be com-
puted by assuming steady-state levels of intermediate species:
where mc is the mass of a carbon atom and is the surface density of Csoot - H sites
(~ 2.310 15 cm-2, according to Frenklach et al. (1985). This corrected rate expression
(CH) is provided by Xu and colleagues (1997). Xu and colleagues (1997, 1998) have
shown that this expression describes surface growth rates in the post-flame regions
of laminar fuel-rich premixed flames as well as the Frenklach and Wang growth rate.
136 Particulate Formation
The soot growth rates from the CH expression do not employ the particle ageing
equation that reduces growth rates with increasing particle time. Instead, revers-
ibility in Reaction 3 is incorporated. Hence, in systems where elapsed time of the
soot particle is not available to compute temporal-dependent ageing effects, the CH
mechanism may be preferred.
Despite reasonable success in modeling of a variety of flames, it is now widely
recognized that PAH addition (or condensation) will also affect total soot mass,
especially in the early phases of growth (Benish et al., 1996; Bockhorn et al., 2009).
5.5Soot Oxidation
In principle, soot can be oxidized by any of many oxidizing species, including O2,
OH, O, CO2, and H2O. The analogous reactions associated with each of these oxidiz-
ing species may be viewed as:
O2 + C(s) => CO + O
OH + C(s) => CO + H
O + C(s) => CO
CO2 + C(s) => 2 CO
H 2O + C(s) => CO + H 2
Soot oxidation begins to occur as fuel-rich conditions (when O2 and O-atom con-
centration levels are negligible) begin to transfer toward fuel-lean conditions.
Specifically, this threshold occurs when equivalence ratios decrease below approxi-
mately 1.5. Under such conditions, oxidation by OH begins to occur. For the slightly
fuel-rich and perhaps stoichiometric conditions, molecular O2 concentrations are
vanishingly small and O-atom concentrations are negligible. Oxidation by thermo-
dynamically stable species such as CO2 and H2O is slow, as these processes are endo-
thermic by about 41 and 36 kcal/mole, respectively. Oxidation by OH also dominates
in leaner portions of laboratory flames; this result is somewhat surprising as equi-
librium values of OH suggest that this process is too low to dominate over oxida-
tion by O2. In such regions, however, super-equilibrium radical concentrations of
OH are about a factor of ten over equilibrium levels, resulting in soot oxidation
dominance by OH radicals; such conditions persist well into post-flame zones for
atmospheric pressure flames (Fenimore and Jones, 1967; Mulcahy and Young, 1975;
Neoh et al., 1980, 1984; Smooke et al., 1999; Xu et al., 2003). Under high-pressure gas
turbine engines, super-equilibrium levels relax much more quickly to equilibrium
conditions, yet inlet temperatures and hence flame temperatures are also higher,
resulting in still high OH concentrations, whereas oxidation rates by O2 at high tem-
peratures are diffusion limited and therefore constrained. Thus, oxidation by OH
also dominates in gas turbine combustors. Rates for these various processes have
been reexamined over the past fifteen years (e.g., Von Gersum and Roth, 1990; Roth
et al., 1991; Xu et al., 2003) and validate earlier values from over thirty years ago
(Nagle and Strickland-Constable, 1962; Neoh et al., 1980, 1984). Very recent results
by Lighty (Echavarria et al., 2011) confirm earlier suggestions by A. F. Sarofim that
5.5 Soot Oxidation 137
O2 can enhance particle breakup (versus simple erosion of the surface), but interpre-
tations of this latter work are incomplete. Particle breakup simply may be enhanced
in Echavarrias recent experiments by allowing O2 to diffuse into soot pores at low
temperatures and then, upon reheating, reaction and related heat release lead to
particle fracturing. This specific mechanism may be less important in conventional
flames or combustor situations, in which time at low temperatures is not available for
O2 diffusion into pores without reacting.
Exceptions to the rule that OH dominates soot oxidation do occur. One exam-
ple is in stirred reactor experiments. In this case, unreacted molecular oxygen may
be strongly mixed with the fuel-rich soot-forming gases. In such cases, OH levels are
low, but O2 may be unusually high because of incomplete combustion. Such condi-
tions may also exist in a strongly mixed fuel-rich region in the front end of a gas
turbine combustor, resulting in a possible role of O2 in soot oxidation in gas turbine
combustors (Colket et al., 2004).
Specific rate expressions for reactions of oxidizing species with soot are rea-
sonably well established, although researchers have done limited work in verifying
product identification. The rate for the reaction with O2 was developed by Nagle
and Strickland-Constable (1962) and is frequently referred to as the NSC rate. The
expression for the specific oxidation rate of soot mass (grams/sec/cm2) is reproduced
here:
1 dm kA po 2
= x + kB po 2 ( 1 x)
12 dt 1 + kC po 2
1
k
x = 1+ T
kB po 2
The model upon which this expression was developed assumes two types of sites, A
and B, for attack by O2. The latter, B, sites are less reactive. The fraction of sites
occupied by A sites is x and the remainder (1-x). The second equation accounts for
B sites reverting to A sites during the oxidation. The rate constants as defined by
Nagle and Strickland-Constable are reproduced in Table 5.3.
Curves showing the pressure, temperature, and oxygen dependencies as
described by Table 5.3 are shown in Figure 5.10. At lower temperatures (<1800 K),
the calculations depicted in Figure 5.10 suggest the rate is relatively independent of
the O2 concentration, but strongly dependent on temperature. However, at stoichio-
metric flame temperatures in a combustor (~2500 K) at which 1/T are ~ 4, then the
impact of the depleted O2 (due to combustion) is accentuated dramatically (> order
of magnitude change). Shock tube experiments (Brandt and Roth, 1989) in more
recent years have confirmed the rates depicted by the set of NSC expressions.
In computing oxidation rates, just as for surface growth processes, it is criti-
cal to know the (active) surface area of the soot particles, as the oxidation rate is
usually defined as a specific oxidation rate, or oxidation rate per surface area
(gm/cc/sec). Most models simply compute the surface area available for combus-
tion as the geometric equivalent of a set of spherical particles, of diameter, d. Some
138 Particulate Formation
models consider ageing effects in which the active surface area changes as a function
of time, not just through agglomeration/coalescence processes, but also via chemical
maturation effects (ageing) on the particle surface. Hence, oxidation rates are usu-
ally given by:
where kx is the applicable rate constant, [X] is the concentration of the oxidizing
species, SA is the available active surface area that can be oxidized, and MWc is the
molecular weight of carbon. Note that for oxidation by O2, the equation is modified
according to the NSC expression. Also note that the active surface area is usually
equal to or less than the geometric surface area, as active surface area is lost in the
ageing and agglomerization processes.
In models, oxidation of soot by OH radicals is assumed to proceed at a gas kinetic
collision frequency multiplied by a collision probability of 0.13. This probability has
been empirically determined by Neoh, Howard, and Sarofim (1984). Thus, with NOH
and NA representing the OH number density and Avogadros number, respectively,
the OH oxidation is:
RgT 12 P
ROH = 0.13NOH = 16.7 OH
2 WOH NA T
where POH is the OH partial pressure in the atmosphere, and the specific growth
rate is in gram/cm2/sec units. For reference, the collision probability for reaction
with O-atom (Von Gersum and Roth, 1992) is even higher (0.23), but oxidation by
OH dominates because of its higher concentrations where soot concentrations exist
(fuel-rich and stoichiometric regions).
1.E-02
Po2(atm)
0.5
Oxidation rate (gm/cm2/sec)
0.2
1.E-03
0.1
0.05
0.02
1.E-04
0.01
0.005
1.E-05
3 4 5 6 7
4/T(1/K)
10
Figure 5.10. Dependence of oxidation rates (NSC) by O2 on temperature, pressure, and O2.
other. The loss of surface area during agglomerization is small, but finite, because of
loss of the contact area among the touching spheres. The coalescence rates, , in
the previous equation is given by:
(m1 , m2 ) = PR (d1 + d2 )2 V
4
where m1 represents the soot particle masses of the colliding particles 1 and 2, d1 are
the particle diameters, and the relative particle velocity is given by:
8kT 1 1
Vm1m2 = +
m1 m2
A nominal of value of 1.5 is recommended for the Van der Vaals enhancement fac-
tor, PR, although values as high as 2.2 have been suggested (Frenklach and Wang,
1994).
A variety of methods have been developed to treat particle size distribution as it
evolves over time. The simplest approach is to assume a monodisperse distribution.
A variety of researchers, including Magnussen and Hjertager (1976), Fairweather
and colleagues (1992), and Lindstedt (1994), have utilized such methods. Colket and
colleagues (2003) have developed the simplified modification:
NM 2
SA = 4.29 3
2
to alter surface growth and oxidation rates to account for size distribution functions
when utilizing a monodisperse particle size model. In the previous equation, M is the
total soot mass (grams/cc gas), SA is the total surface area (cm2 soot/cc gas) based on the
particle number density, N (#/cc gas) is the number density, and (grams/cc soot) repre-
sents the primary particle mass density. Such monodisperse models offer a tremendous
140 Particulate Formation
As depicted in the schematic of Figure 5.8, soot formation occurs in the post-flame
region of fuel-rich (premixed) flames. The reactions forming soot require high tem-
peratures and substantial acetylene concentrations and hence do not initiate until
the main combustion process is nearly complete. For atmospheric pressure flames,
the flame zone is complete on the order of one to three milliseconds, while the soot
formation zone occurs on a time scale about an order of magnitude longer. For gas
turbines at elevated pressures and temperatures, these time scales are both reduced
by about an order of magnitude.
While the majority of soot studies in laboratory burners have been conducted
at atmospheric and subatmospheric conditions, some studies have been done at
5.7 Related Phenomena 141
elevated pressures. The results of these studies indicate that the pressure depen-
dence of soot production varies with flame type. In premixed flames, Bonig and col-
leagues (1991) studied flat ethylene flames up to pressures of 7 MPa and reported
a pressure dependence of the final soot volume fraction of approximately two.
In methane diffusion flames, Thomson and colleagues (2005) showed a pressure
exponent of maximum soot concentration of about two for pressures up to 2 MPa,
but a lower exponent of about 1.2 between 2 and 4 MPa. Thomson and colleagues
reported that their work is consistent with the earlier work of Flower and Bowman
(1986) and of Lee and Na (2000). McCrain and Roberts (2005) report different
results: they found that the peak soot volume fraction, for their study of laminar
methane and ethylene air flames, scales with pressure exponents of 1.2 and 1.7,
respectively, for pressures up to 2.5 MPa. So not only does this work disagree with
the findings of Thomson and colleagues, it also suggests that the pressure expo-
nent may depend on the fuel structure. McCrain and Roberts (2005) noted that
an exact comprehensive explanation of the high-pressure mechanisms and how
they differ from atmospheric mechanisms remains elusive, supporting the need
for more work in this area.
Soot Ageing: Soot primary particles reach a maximum size because of active sur-
face site deactivation (Woods and Haynes, 1994). This process is also referred to
as ageing of surface sites. Singh and colleagues (2005) tested functional depen-
dences of surface reactivity on age in their study of high-pressure coagulation using
Monte Carlo techniques. Using premixed flame data, Appel and colleagues (2000)
constructed an empirical expression for the fraction of active sites () that is a func-
tion of the average particle size and gas temperature, but not explicitly to individ-
ual particle age (see Chapter 5.4). Modeling this effect in a diffusion flame is more
difficult. Smooke and colleagues (2005) introduced a simple step function depen-
dence of surface reactivity on particle size at which growth is shut off above a cutoff
particle size (25 nm in their simulations). The 25 nm cutoff appears an appropriate
value for atmospheric pressure flames because the diameter of primary particles
in such flames are typically twenty to thirty nanometers for most fuels. Acetylene
flames, however, produce larger particles (4050 nm). Primary particle sizes may
increase with increasing pressure; although limited information is available on the
actual values, agglomerates with primary sizes of fifty to eighty nanometers have
been observed in emissions from gas turbines. This is not a rule, however, as primary
particle diameters of carbonaceous particulates in gas turbine emissions have been
observed to be as low as ten nanometers.
Dobbins (1996) has measured carbonization rates of young liquid-like soot. The
aerosol particles undergo a density change, increasing from about 1.2 grams/cc to
1.8 grams/cc and decreasing their H/C ratio. This is a very different physical pro-
cess than the ageing process described earlier, and it is logical to assume it will also
impact surface growth rates.
142 Particulate Formation
For atmospheric pressure lean flames with sub-ppm soot volume fractions, the power
radiated arises from narrow, spectral gas bands primarily from CO2, H2O, and CO.
When soot is present, the radiation becomes broadband, with spectral characteristics
similar to Plancks black body radiation leading to increased energy loss per unit
time. For such flames, residence times are often substantial (>20 milliseconds), and
it is common for 2030 percent of the chemical energy to be lost to surroundings
via radiation. Such losses can directly cause the reduction of the local flame tem-
perature by hundreds of degrees Kelvin. In fact, in many atmospheric combustion
devices, including home fireplaces, such radiative loss by soot particles is a key fea-
ture and critical to desired operation. Since this energy loss has a first order impact
on reducing local flame temperatures, gas-phase reaction rates and soot formation
rates are altered (reduced), which in turn alters the total soot volume fractions and
hence radiation losses. Hence, numerical solutions of flame structure and soot levels
require methodologies that couple the reaction chemistry, soot formation/oxidation
processes, and radiation (see, for example, Smooke et al., 2005).
In gas turbines, number densities (of gaseous species and soot) are much higher
(because of increased pressure) and temperatures are higher, hence radiation levels
are higher. The flames may well be optically thick, which limits the increase in energy
loss rates. More important, this increase is counterbalanced by the relatively short
residence times in gas turbine combustors, which for advanced aero engines may be
<5 milliseconds and even shorter for the fuel-rich front end of the combustor. Hence,
treatment of radiation losses in simulations of gas turbine flames may not be as criti-
cal as it is for atmospheric pressure flames.
For flames with low sooting levels, the optically thin approximation (i.e., ignor-
ing absorption of radiated energy) may be utilized using expressions such as those
developed by Hall (1994). For higher soot loading in flames, the radiation levels
increase and the optically thin model overestimates the radiation losses. The front
end of a combustor will have high soot loadings, and the optically thin assumption
will significantly overestimate radiation losses. In principle, some reabsorption of
thermal emissions can occur, particularly on or near the centerline of a coflow flame,
which receives emissions from surrounding regions of the flame. This optical thick-
ness effect reduces the net rate of thermal radiation energy loss and locally raises
the temperature. Smooke and colleagues (2005) provide a method to compute such
losses. While temperature changes associated with radiation reabsorption are not
large, the great sensitivity of soot growth (and NOx formation) to temperatures
makes incorporation of these effects important. At high soot loadings in gas turbine
combustors, reabsorption of radiated energy becomes dominant and nearly black
body radiation levels limit radiation to liners or to gases.
The fuel has a substantial impact on the production of soot in a gas turbine engine.
Gaseous and prevaporized fuels tend to produce low levels of soot, while polycyclic
5.7 Related Phenomena 143
2.00
J79
Smoke No. /(Smoke No.)14%H
1.75
TF30
1.50
F101
TF33 F100
1.25
TF41
1.00
11 12 13 14 15
Hydrogen content of fuel % by weight
Figure 5.11. Relationship between smoke number and weight percent hydrogen in various
engines, normalized by the smoke numbers obtained with 14.5 percent hydrogen in the fuel.
Alternative fuels have received a great deal of attention recently. Fuels based on
the Fischer-Tropsch (F-T) process and hydrotreated renewable jet (HRJ) fuels have
won approval (ASTM D7566, Annexes 1 and 2) for addition to petroleum jet fuels
up to a 50 percent level. Because of the high hydrogen content of these fuels, they
are composed primarily of normal alkanes, iso-alkanes, and some cycloalkanes but
with little or no aromatics. Blends with up to a maximum of 50 percent alternative
fuels are approved because of (1) conservatism, (2) gravimetric densities of F-T and
HRJ fuels below the minimum allowed for petroleum fuels, and (3) field tests that
demonstrate the importance of minimum aromatic levels in the fuel. Historically,
petroleum-based fuels have had 825 percent aromatic levels and the low aromatic
levels in alternative fuels create a problem for the seals in fuel lines. Any new blended
fuel must have a minimum level of 8 percent aromatics as that level is consistent with
the experience base. The impact on soot emissions through the use of alternative
fuels can be gathered from Figure 5.11 and the knowledge that the fuel components
of the F-T and HRJ fuels are predominantly n-alkanes and branched (iso) alkanes.
The H/C (molar) ratio of such fuels will be about 2.16 or a hydrogen mass fraction of
15.3 percent, which falls far to the right in the data presented earlier. In practice, pure
F-T fuels have been found to reduce soot from engines from about 50 to 90 percent
(Bulzan et al., 2010). The actual benefit by utilizing alternative fuels on reduced soot
emissions will be less, as no more than 50 percent can be utilized.
Recent work by Vander Wal (2011, Personal communication) in collecting soot
particles in engine exhaust and analyzing them using transmission electro micro-
graphs suggests differences in the character/structure of the soot particulates with
changes from petroleum jet fuels to alternative jet fuels. This is an area of active
research: the result may be related to the lower sooting propensities of the high H/C
ratio alternative fuels.
5.8 Particulate Formation in Combustion Systems 145
Direct dilution by adding inerts can noticeably reduce soot production (Gomez
etal.,1987). The inert may be added with the fuel or the air, but this effect is stron-
gest when the inert is added with the fuel, as it dilutes the fuel density and lowers
local temperatures, both of which reduce soot formation rates. Such additions are
challenging and costly to implement in gas turbine engines, but might be done to
control other phenomena (e.g., reduced NOx emissions, increased power output,
etc.). Examples include the Humid Air Turbine (HAT) cycle (Rao, 1989) and flame-
less combustion (Bruno and Vallini, 1999), because internal recirculation within
the burner of burned products into the flame front should also have some positive
impact in reducing soot formation. Unfortunately, large additions of termolecular
species (i.e., water or CO2) can have a negative impact on system (thermodynamic)
efficiency because of decreased ratio of the gaseous specific heats.
Early evidence of reduced soot production in flames with added inerts (Gomez
and Glassman, 1988) led to work on the cause for soot reduction. Primary causes
were identified as (1) reduction of the number densities of the reactive species to
slow down reaction rates, (2) reduced flame temperature, and (3) altered transport
rates (diffusivities and thermal condition). The relative fraction by which one or
another of these mechanisms plays a role depends on the flame conditions (e.g., the
fuel, flame structure, and diluent).
combustor in recirculation zones with long residence times. These conditions were
ideal for flame stabilization and also provided a good turndown ratio. Unfortunately,
the high local fuel-rich conditions with long residence times produced a large mass
of soot with large diameters that would not burn up in the subsequent leaner zones.
Hence large amounts of soot mass were emitted. Following the implementation of
the Smoke Number tests and regulations, manufacturers adopted either airblast or
air assist fuel nozzles that enabled creation of smaller fuel droplets but also added
significant amounts of high-velocity air into the fuel-rich regions. This reduced the
local residence time and the local fuel-air ratios, while increasing air-fuel mixing
rates, with a direct reduction of soot emissions.
Staged addition of secondary air can further control soot emissions. By adding
some air, following the primary zone, the local mixture ratio would approach equiva-
lence ratios of 1.3 < phi < 0.8, which are ideal for consumption of soot as well as
unburned hydrocarbons and CO. Additional secondary air was injected downstream
and allowed for tailoring of the temperature profile (pattern factor) prior to flow
into the turbine. Unfortunately, it became apparent that the near stoichiometric con-
ditions between the secondary air injection ports was also ideal for NOx formation.
To avoid the high NOx formation rates between the air dilution jets, the
rich-quench-lean (RQL) combustor was devised. This design resulted in a single row
of secondary air injection holes following the primary zone. The objective was to rap-
idly transition the flow from fuel-rich conditions at which NOx formation rates are
negligible or small to overall lean conditions at which NOx production rates are also
small. The challenge was to add the additional air rapidly and to minimize the time
at near stoichiometric conditions, but in a manner that allows for tailoring of the exit
temperature pattern. Unfortunately, the near-stoichiometric regions to be avoided
for NOx reductions are ideal for consuming soot particulates. Rapid air addition and
mixing may result in low NOx production rates, but also provides insufficient time for
particle burnout during the quench process. Since oxidation rates at fully mixed-out
conditions are small, particulate emissions could increase. Hence, careful control of
front-end fuel-to-air ratios and residence times to minimize initial soot formation
is desired for RQL combustors. Alternative designs of the RQL combustor include
those with air injection sleeves inserting the secondary air injection directly to the
center of the combustor.
More recently, lean-direct-injection (LDI) designs have been developed for
control of NOx emissions from engines. In such designs, the primary zone remains
overall lean to minimize NOx production. To enable turndown and flame stability,
such burners are typically staged and their primary zones may be larger than nec-
essary for RQL burners. Inherently, these burners have a different fingerprint for
particulate emissions. Soot is formed in thin fuel-rich zones that quickly mix out to
avoid formation of large particulates. The particulates thus have little time to grow
significant mass, but cool down rapidly and further oxidation is minimal. Hence par-
ticle emissions consist of noticeably smaller particles (d < 30 nm vs. d > 40 nm at full
power for other combustor designs), although the particle number densities may
be high. Despite significant benefits to NOx emissions, these burners generally have
increased complexity because of the requirements for fine-scale fuel-air mixing and
References 147
fuel staging; furthermore, these burner designs are more susceptible to combustion
dynamics as they have performance characteristics similar to those of lean premixed
combustors for industrial gas turbine engines.
Ground-based gas turbine engines typically have much lower soot emissions than
aero engines. Early designs consisted of aeroderivative engines in which combustors
are similar to those used in aero engines. However, such installations frequently utilize
gaseous fuels (natural gas or propane) and produce much less soot because of more
rapid fuel-air mixing and high H/C ratios of the fuel. Furthermore, most installations
are subject to local limits on NOx emissions; hence, in such aeroderivative engines,
systems were installed to inject water directly into the combustor to reduce local flame
temperatures and NOx productions. The water injection has two secondary advantages:
it allows for increased power output and it reduces production of soot due to dilution
effect. These benefits are achieved not only for gaseous fuels, but also for liquid-fueled
engines. Even without water injection and for liquid fuels, soot production from an
aeroderivative may be similar to or even greater than that from the equivalent aero
engine at the same operating condition. However, it is primarily at takeoff power when
soot emissions are the highest for the aero engine. Sustained operation at such condi-
tions for an aeroderivative would severely degrade its durability and hence soot emis-
sions from ground-based aeroderivatives are relatively low, even with liquid fuels.
Combustors in frame machines mostly employ various dry low NOx (DLN)
approaches. Typically, these are lean premixed systems. While some soot can be pro-
duced from continuous or intermittent pilots, the main flames produce virtually no
soot emissions and so such combustors may be considered non-sooting. Substantially
redesigned aeroderivative engines can offer similar capabilities.
Acknowledgments
The author is indebted to a host of persons who have provided financial support,
challenging questions, intellectual guidance, and mentoring over a period exceed-
ing twenty years and is grateful for their many helpful discussions and suggestions.
These persons include Michael Frenklach, Irv Glassman, Bob Hall, Dave Liscinsky,
Tom Litzinger, Mel Roquemore, Bob Santoro, Dan Seery, Mitch Smooke, Sadaat
Syed, Julian Tishkoff, and Hai Wang.
References
American Society for Testing and Materials. (1989). Standard Test Methods for Surface and
Interfacial Tension of Solutions of Surface-Active Agents, D133189.
Anderson, B. E., Beyersdorf, A. J., Hudgins, C. H., Plant, J. V., Thornhill, K. L., Winstead, E.L.,
Ziemba, L. D. et al. (2011). Alternative Aviation Fuel Experiment (AAFEX), NASA/
TM2011217059, February.
Appel, J., Bockhorn, H., and Frenklach, M. (2000). Combust. Flame 121: 122. www.me.berkeley.
edu-soot-codes.f..
Bachmann, J. D., Damburg, R. J., Caldwell, J. C., Edwards, C., and Koman, P. D. (1996). Review
of the National Ambient Air Quality Standards for Particulate Matter: Policy Assessment
of Scientific and Technical Information, EPA-452/R-96013 (NTIS PB97115406), Office
of Air Quality Planning and Standards, Research Triangle Park, NC.
148 Particulate Formation
Balthasar, M., and Frenklach, M. (2005). Proc. Combust. Inst. 30: 146775.
Bendetto, D., Pasini, S., Falcitelli, M., Marca, C. La., and Tognotti, L. (2000). Comb. Sci. & Tech.
153: 27994.
Bengtsson, K. U. M., Benz, P., Schren, R., and Frouzakis, C. E. (1998). Proc. Comb. Inst. 27:
13939.
Benish, T. G., Lafeur, A. L., Taghizadeh, K., and Howard, J. B. (1996). Proc. Comb. Inst. 26:
231926.
Bhargava, A., Colket, M., Sowa, W., Casleton, K., and Maloney, D. (2000). An Experimental
and Modeling Study of Humid Air Premixed Flames. ASME Journal of Engineering For
Gas Turbines and Power 122(3): 40511.
Blanquart, G., and Pitsch, H. (2007). Parameter Free Aggregation Model for Soot Formation.
Presented at the fifth U.S. Combustion Meeting, San Diego, CA, March 258.
Blazowski, W. S., and Jackson, T. A. (1978). Evaluation of Future Jet Fuel Combustion
Characteristics, AFAPL-TR-7793, July.
Bockhorn, H., DAnna, A., Sarofim, A.F., and Wang, H., eds. (2009). Combustion Generated
Fine Carbonaceous Particles, KIT Scientific Publishing, Karlsuher.
Bonig, M., Feldermann, C., Jander, H., Luers, B., Rudolph, G., and Wagner, H. Gg. (1991).
Soot Formation in Premixed C2H4 Flat Flames at Elevated Pressure, Proc. Combust.
Inst 23: 15817.
Brandt, O., and Roth, P. (1989). A Tunable IR-Diode Laser Technique for Measuring Reaction
Rates of High Temperature Aeroslos. Combustion and Flame 77: 6978.
Brocklehurst, H. T., Pridden, C. H., and Moss, J. B. (1997). Soot Predictions within an Aero
Gas Turbine Combustor Chamber. ASME Paper 97-GT-148.
Brundish, K. D., Miller, M. N.,Wilson, C.W., Hilton, M., and Johnson, M. P. (2003).Measurement
of Smoke Particle Size and Distribution Within a Gas Turbine Combustor. Proceedings
of ASME Turbo Expo, Atlanta GA, GT200338627, June.
Bruno, C., and Vallini, L. (1999). Flameless Combustion and its Application to Aero-engines.
Fourteenth International Symposium on Air Breathing Engines, September 510,
Florence, Italy.
Bulzan, D., Anderson, B., Wey, C., Howard, R., Winstead, E., Beyersdorf, A., Corporan, E.,
et al. (2010). Gaseous and Particulate Emissions Results of the NASA Alternative
Aviation Fuel Experiment (AAFEX). Proceedings of the ASME Turbo Expo 2010,
GT201023524, Glasgow, Scotland.
Champagne, D. L. (1971). Standard Measurement of Aircraft Gas Turbine Engine Exhaust
Smoke. ASME 71-GT-88, 1971.
Colket, M. B. (1988). The Pyrolysis of Acetylene and Vinylacetylene in a Single-Pulse Shock
Tube, Proc. Comb. Inst. 21: 85164.
Colket, M. B., and Hall, R. J. (1994). Successes and Uncertainties in Modeling Soot Formation
in Laminar, Premixed Flames, in Soot Formation in Combustion, Mechanisms and
Models, Bockhorn, H., ed., Springer Verlag, Berlin, p. 417.
(1997). Mechanisms Controlling Soot Formation in Diffusion Flames. AFOSR Contract
F4962094-C-0059, UTRC975.906.00011, August 1, 1994 to July 31, 1997.
Colket, M. B., Hall, R. J., and Stouffer, S. (2004). Modeling Soot Formation in a Stirred
Reactor. ASME Turbo-Expo Meeting, GT200454001, Vienna, Austria.
Colket, M. B., Liscinsky, D. L., Chiappetta, L., Leong, M., Madabhushi, R., Zeppieri, S., and
Hautman, D. (2003). Mitigation of Particulate Emissions in Engines via Fuel Additives.
UTRC Report No. R036.100.00071, AF Technology Investment Agreement F33615
0022001, April.
Colket, M. B., and Seery, D. J. (1984). Presentation at the Twentieth Symposium (International)
on Combustion, Ann Arbor, MI.
(1994). Proc. Comb. Inst. 25: 88391.
DAnna, A., DAlessio, A., and Minutolo, P. (1994). Spectroscopic and Chemical Characteriza-
tion of Soot Inception Processes in Premixed Laminar Flames at Atmospheric Pressure,
in Soot Formation in Combustion, Bockhorn, H., ed., Springer Verlag, Berlin, p. 83.
References 149
Liscinsky D. S., and Hollick H. H. (2010). Effect of Particle Sampling Technique and
Transport on Particle Penetration at the High Temperature and Pressure Conditions
Found in Gas Turbine Combustors and Engines. NASA contractor report, NASA/
CR-2010-NNC07CB03C.
Liu, A. B., Mather, D., and Reitz, R. D. (1993). Modeling the Effects of Drop Drag and
Breakup on Fuel Sprays. SAE Technical Paper 930072.
Magnussen, B. F., and Hjertager, B. H. (1976). On Mathematical Modeling of Turbulent
Combustion with Special Emphasis on Soot Formation and Combustion, Proceedings
of the Combustion Institute 16: 71929.
Malecki, R. E., Rhie, C. M., Mckinney, R. G., Ouyang, H., Syed, S. A., Colket, M. B., and
Madabhushi, R. K. (2001). Application of an Advanced CFD-Based Analysis System to
the PW6000 Combustor to Optimize Exit Temperature Distribution Part I: Description
and Validation of the Analysis Tool. Proceedings of ASME Turbo EXPO (2001-GT-
0062), New Orleans, LA, June.
Marchal, C., Delfau, J.-L., Vovelle, C., Morac, G., Mounam-Rousselle, C., and Mauss, F. (2009).
Modelling of Aromatics and Soot Formation from Large Fuel Molecules, Proceedings
of the Combustion Institute 32: 7539.
Marinov, N. M., Pitz, W. J., Westbrook, C. K., Lutz, A. E., Vincitore, A. M., and Senkan, S. M.
(1998). Chemical Kinetic Modeling of a Methane Opposed-Low Diffusion Flame and
Comparison to Experiments, Proc. Comb. Inst. 27: 60513.
Marsh, R., Crayford, A., Petzold, A., Johnson, M., Williams, P., Ibrahim, A., Kay, P., et al. (2010).
Studying, Sampling and Measurement of Aircraft Particulate Emissions II (SAMPLE
II) Final Report. Research Project EASA, 2009, European Aviation Safety Agency,
November.
Melius, C. F, Miller, J. A., and Evleth, E. M. (1993). Unimolecular Reaction Mechanisms
Involving C3H4, C4H4, and C6H6 Hydrocarbon Species, Proc. Comb. Inst. 24: 6218.
Mensch, A., Santoro, R. J., Litzinger, T. A., and Lee, S.-Y. (2010). Sooting Characteristics Of
Surrogates For Jet Fuels, Combust. Flame 157: 1097105.
McCrain, L. L. and Roberts, W. L. (2005). Measurements in the Soot Volume Field in Laminar
Diffusion Flames at Elevated Pressures. Combust. Flame, 140: 609.
McEnally, C., and Pfefferle, L. (2009). Sooting Tendencies of Nonvolatile Aromatic
Hydrocarbons, Proc. Comb. Inst. 32: 6739.
McEnally, C., Shaffer, A., Long, M. B., Pfefferle, L., Smooke, M. D., Colket, M. B., and Hall,
R. J. (1998). Computational and Experimental Study of Soot Formation in a Coflow,
Laminar Ethylene Diffusion Flame, Proc. Comb. Inst. 27: 1497505.
McVey, J. B., Russell, S., and Kennedy, J. B. (1987). High-Resolution Patternator for the
Characterization of Fuel Sprays. Journal of Propulsion and Power 3: 2029.
McKinnon, J. T., and Howard, J. B. (1992). Twenty-Fourth Symposium (International) on
Combustion, The Combustion Institute, Pittsburgh, PA, p. 965.
Miller, J. A., and Klippenstein, S. J. (2001). The Recombination of Propargyl Radicals: Solving
the Master Equation, J. Phys. Chem. A 105: 725466.
(2003). The Recombination of Propargyl Radicals and Other Reactions on a C6H6
Potential, J. Phys. Chem. A 107: 778399.
Miller, J. A., and Melius, C. F. (1992). Combust. Flame 91: 2139.
Moses, C. (1984). U.S. Army Alternative Gas-Turbine Fuels Research: MERADCOM.
Combustion Problems in Turbine Engines, AGARD-CP-353, January, pp. 71 to 710.
Mosier, S. A. (1984). Fuel Effects on Gas Turbine Combustion Systems. Combustion
Problems in Turbine Engines, AGARD-CP-353, January, pp. 51 to 515.
Mueller, M. E., Blanquart, G., and Pitsch, H. (1975). Hybrid Method of Moments for Modeling
Soot Formation and Growth, Combust. Flame 156: 114355.
Mulcahy, M. F. R., and Young, B. C. (1975). Reaction of Hydroxyl Radicals with Carbon at
298 K, Carbon 13: 115.
Nagle, J., and Strickland-Constable, R. F. (1962). Oxidation of Carbon between 1000-2000C,
in Proceedings of Fifth Carbon Conference 1:154, Pergamon Press, Inc., Oxford.
152 Particulate Formation
Neoh, K. G., Howard, J. B., and Sarofim, A. F. (1980). Effect of Oxidation on the Physical
Structure of Soot, in Particulate Carbon, Siegla, D. C. and Smith, B. W. eds., Plenum Press,
New York, p. 261.
(1984). Effect of Oxidation on the Physical Structure of Soot, Proc. Combust. Inst. 20: 951.
Norgren, C. T., and Ingebo, R. D. (1975). Particulate Exhaust Emissions from an Experimental
Combustor. Lewis Research Center, NASA Technical Memorandum, NASA TM X-3254,
June.
Odgers, J., and Kretschmer, D. (1984). The Effects of Fuel Composition upon Heat Transfer in
Gas Turbine Combustors, in Combustion Problems in Turbine Engines, AGARD-CP-353,
January, pp. 81 to 810.
ORourke, P. J. (1981). Collective Drop Effects on Vaporizing Liquid Sprays, PhD dissertation,
Princeton University.
ORourke, P. J., and Amsden, A. A. (1987). The TAB Method for Numerical Calculation of
Spray Droplet Breakup. SAE Technical Paper 872089.
Palmer, H. B., and Cullis, C. F. (1965). The Formation of Carbon from Gases, in Chemistry
and Physics of Carbon, Walker, P. L. ed., Marcel Dekker, NY, p. 265.
Park, S. H. and Rogak, S. N. (2004). A Novel Fixed-Sectional Model for the Formation and
Growth of Aerosol Agglomerates. Journal of Aerosol Sciences, pp. 1385.
Penner, J. E., Lister, D. H., Griggs, D. J., Dokken, D. J., and McFarland, M., eds. (1999). Aviation
and the Global Atmosphere, in Intergovernmental Panel on Climate Change, A Special
Report of Working Groups I and II, Cambridge University Press.
Priemus, H., and Schutte-Postma, E. (2009). Notes on the Particulate Matter Standards in the
European Union and the Netherlands. Int. J. Environ. Res. Public Health 6(3): 115573.
Rao, A. D. (1989). Process for Producing Power. U.S. Patent 4,829,763, May.
Richter, H., Granata, S., Green, W. H., and Howard, J. B. (2005). Detailed Modeling of PAH
and Soot Formation in a Laminar Premixed Benzene/Oxygen/Argon Low-Pressure
Flame, Proc Comb. Inst. 30: 1397405.
Roth, P., Brandt, O., and Von Gersum, S. (1991). High Temperature Oxidation of Suspended
Soot Particles Verified by CO and CO2 Measurements, Proc Comb. Inst. 23: 148591.
Sampath, P., and Gratton, M. (1984). Fuel Character Effects on Performance of Small
Gas Turbine Combustion Systems, in Combustion Problems in Turbine Engines,
AGARD-CP-353, January, pp. 61 to 612.
Singh, J., Balthasar, M., Kraft, M., and Wagner, W. (2005). Stochastic of Soot Particle Size and
Age Distributions in Laminar Premixed Flames, Proc. Comb. Inst. 30: 145766.
Smooke, M. D., Hall, R. J., Colket, M. B., Fielding, J., Long, M. B., McEnally, C. S., and Pfefferle,
L. D. (2004). Investigation of the Transition from Lightly Sooting Towards Heavily
Sooting Co-Flow Ethylene Diffusion Flames, Comb. Theory and Modelling 8: 593606.
Smooke, M. D., Long, M. B., Connelly, B. C., Colket, M. B., and Hall, R. J. (2005). Soot
Formation in Laminar Diffusion Flames, Comb. Flame 143: 61328.
Smooke, M. D., McEnally, C. S., Pfefferle, L. D, Hall, R. J., and Colket, M. B. (1999).
Computational and Experimental Study of Soot Formation in a Coflow, Laminar
Diffusion Flame, Comb. Flame 117: 11739.
Smooke, M. D., Yetter, R. A., Parr, T. P., Hanson-Parr, D. M., Tanoff, M. A., Colket, M. B., and
Hall, R. J. (2000). Computational and Experimental of Ammonia Perchlorate/Ethylene
Counter Flow Diffusion Flames, Proc. Comb. Inst. 28: 201320.
Smoluchowski, M. (1916). Drei Vortrge ber Diffusion, Brownsche Molekularbewegung
und Koagulation von Kolloidteilchen, Physik. Zeit. 17: 55771, 58599.
Snyder, T. S., Stewart, J. F., Stoner, M. D., and McKinney, R. G. (2001). Application of an
Advanced CFD-Based Analysis System to the PW6000 Combustor to Optimize Exit
Temperature Distribution. Part II: Comparison of Predictions to Full Annular Rig Test
Data. Proceedings of ASME Turbo EXPO (2001-GT-0064), New Orleans, LA, June.
Thomson, K. A., Gulder, O. L., Weckman, E. J., Fraser, R. A., Smallwood, G. J., and Snelling, D. R.
(2005). Soot Concentrations and Temperature Measurements in Annular, Non-premixed,
Laminar Flames at Pressures up to 4 MPa. Combustion and Flame, February, 140(3): 222.
References 153
Tolpadi, A. K., Danis, A. M., Mongia, H. C., and Lindstedt, R. P. (1997). Soot Modeling in Gas
Turbine Combustors. ASME Paper 97-GT-149.
Tsang, W. (Organizing Chair) (2003). Workshop on Combustion Simulation Databases for
Real Transportation Fuels. National Institute of Standards and Technology, Gaithersburg,
MD, September 45.
U.S. Environmental Protection Agency. (1997). National Ambient Air Quality Standards for
Particulate Matter: Final Decision, Federal Registrar, 62 FR 38652, July 18.
Violi, A., Kubota, A., Truong, T. N., Pitz, W. J., Westbrook, C. K., and Sarofim, A. F. (2002). A
Fully Integrated Kinetic Monte Carlo Molecular Dynamics for the Simulation of Soot
Precursor Growth, Proc. Combust. Inst. 29: 23439.
Violi, A., Sarofim, A. F., and Voth, G. A. (2004). Kinetic Monte Carlo-Molecular Dynamics
Approach To Model Soot Inception, Combust Sci. Technol. 176: 9911005.
Von Gersum, S., and Roth, P. (1990). High temperature oxidation of soot particles by. O
atoms and OH radicals, J. Aerosol Science 21: S31S34.
(1992). Soot oxidation in high temperature N2O/Ar and NO/Ar mixtures, Proc.
Combust.Inst. 24: 9991006.
Wagner, H. Gg. (1979). Soot Formation in Combustion, Proc. Combust. Inst. 17: 3.
Waldmann, L., and Schmitt, K. H. (1966). Chapter 6 in Aerosol Science, Davies, C. N. ed.,
Academic Press.
Wang, H. (2011). Formation of Nascent Soot and Other Condensed-Phase Materials in
Flames. Proc. Combust. Inst. 33: 4167.
Wayson, R. L., Fleming, G. G., and Lovinelli, R. (2009). Methodology to Estimate Particulate
Matter Emitting from Certified Commercial Aircraft Engines. J. Air & Water Management
Assoc. 59: 91100.
Wey, C. C., Anderson, B. E., Hudgins, C., Wey, C., Li-Jones, X., Winstead, E., Thornhill, L.K.
et al. (2006). Aircraft Particle Emissions eXperiment (APEX). NASA/TM-2006
214382, ARL-TR-3903, Cleveland, OH, September.
Wey, C. C., Anderson, B. E., Wey, C., Miake-Lye, R. C., Whitefield, P., and Howard, R. (2007).
Overview on the Aircraft Particle Emissions Experiment. Journal of Propulsion and
Power 23: 898905.
Willeke, K., and Baron, P. A., eds. (1993). Aerosol Measurement, Principles, Techniques, and
Applications, Van Nostrand Reinhold, New York.
Wolff, G. T. (1996). Closure by the Clean Air Scientific Advisory Committee (CASAC) on the
Staff Paper for Particulate Matter. EPA-SAB-CASAC-LTR-96008, U.S. Environmental
Protection Agency, Washington, DC.
Wood, A. D. (1975). Correlation between Smoke Measurements and the Optical Properties
of Jet Engine Smoke. Society of Automotive Engineering Paper 751119.
Woods, I. T., and Haynes, B. S. (1994). Active Sites in Soot Growth, in Soot Formation in
Combustion Mechanisms and Models, Bockhorn, H. ed., SpringerVerlag, Berlin,
pp.27589.
Wu, C. H., and Kern, R. D. (1987). Shock-Tube Study of Allene Pyrolysis, J. Phys. Chem 91:
62916.
Xu, F., El-Leathy, A. M., Kim, C. H., and Faeth, G. M. (2003). Soot Surface Growth in Laminar
Hydrocarbon/Air Diffusion Flames, Combust. Flame 132: 4357.
Xu, F., Lin, K-C., and Faeth, G. M. (1998). Soot Formation in Laminar Premixed Methane/
Oxygen Flames at Atmospheric Pressure, Comb. Flame 115: 195209.
Xu, F., Sunderland, P. B., and Faeth, G. M. (1997). Soot Formation in Laminar Premixed
Ethylene/Air Flames at Atmospheric Pressure, Combust. Flame 108: 47193.
Yang, Y., Boehman, A. L., and Santoro, R. J. (2007). A Study of Jet Fuel Sooting Tendency
Using the Threshold Sooting Index (TSI) Method, Combust. Flame 149: 191205.
Zhang, Q., Guo, H., Liu, F., Smallwood, G. J., and Thomson, M. J. (2009). Modeling of Soot
Aggregate Formation and Size Distribution in a Laminar Ethylene/Air Coflow Diffusion
Flame with Detailed PAH Chemistry and an Advanced Sectional Aerosol Dynamics
Model, Proc. Combust. Inst. 32: 7618.
6 Gaseous Aerosol Precursors
Richard C. Miake-Lye
6.1 Introduction
Concerns about particle emissions from aircraft were first raised in the 1960s because
of the visible smoke trails left behind by jet aircraft on takeoff. These concerns led
to an emission certification requirement for aircraft engines in the 1970s that man-
dated a smoke number (SN) measurement, which served to control the visible opac-
ity of the emitted exhaust. Understanding of the effects of particle emissions has
progressed dramatically since the 1970s, and more knowledge now exists both about
carbonaceous soot particles that contributed strongly to the black smoke trails of the
1960s and about how other emissions can condense and add to particle numbers and
mass. These latter contributions arise because of gaseous emissions that are prod-
ucts of combustion and also have low-vapor pressures. Having low-vapor pressures,
they are thermodynamically disposed to condense as the exhaust mixes and cools in
the atmosphere. These condensable species are gaseous aerosol precursors and their
contributions to particulate matter pollution are the subject of significant scientific
research and regulatory interest.
Like all consumers of hydrocarbon fuels, gas turbine engines emit products of
combustion dominated by carbon dioxide and water vapor. In addition to these major
products of combustion, the exhaust emissions also include products of incomplete
combustion, due to small combustion inefficiencies (very small for modern aircraft
engines at cruise), and pollutants formed in the combustion process, like NOx and
SOx. Beyond combustion-related emissions, recent work has identified emissions
from the lubrication system that also contribute to particle emissions in the exhaust.
These various emissions include gaseous species and particles.
Because aircraft gas turbine jet engines use their exhaust to propel the vehi-
cle, the temperature and velocity of the exhaust are uniquely high for such aircraft
engines relative to other types of exhaust. In turn, these exhaust conditions deter-
mine that the details regarding gaseous and particulate emissions are different for
a turbofan, turbo shaft, or turbo jet engine compared to a typical ground vehicle or
factory smokestack. The high temperatures at the engine exit plane mean that many
species that will eventually end up in the condensed phase as particulate emissions
are emitted from the engine as gaseous species.
154
6.1 Introduction 155
Figure 6.1. Research programs have measured engine exhaust emissions at the engine exit
plane and at various downstream locations. In Aviation Particle Emissions eXperiment
(APEX 1), the PM emissions from a NASA-owned CFM552C1 were measured at 1, 10, and
30 m. (Wey et al., 2007: JPP special section on APEX). Certification measurements are done
in dedicated engine test cells, not in on-wing measurements, with multipoint probes at the exit
plane, and so are much different than this image (NASA photo credit).
Thus the degree and nature of how gaseous aerosol precursors shift from the gas
phase to the particle phase depend on the local environment in which the airplane
is operating. And, while the primary processes of interest involve low-vapor pres-
sure species forming new particles and/or condensing on existing emitted particles,
under certain circumstances, particles or their coatings may reevaporate, especially
as the aerosolprecursor concentrations drop as the exhaust plume further dilutes
(Miracolo et al., 2011).
This chapter focuses on the various species that can be considered gaseous
aerosol precursors, by identifying important species and discussing their emission,
their measurement, and how they contribute to particle emissions. This is an area of
continuing active research, and researchers will likely make major advances in the
coming years. However, much has been learned in the past few decades, both about
particle microphysics generally and about the evolution of aircraft particle emissions
specifically, and this chapter provides a summary of some of the key understanding
presently available.
The exhaust flow contributions from primary pollutants NOx and CO have been
regulated for decades. Beyond the emissions of NOx and CO, the exhaust also con-
tains emissions of oxidized sulfur compounds, SOx, due to sulfur contained in the
fuel. Fuel sulfur levels are also regulated, although through fuel specification limits
rather than via exhaust measurements. Partially combusted fuel is also regulated as
unburned hydrocarbons, UHCs, and is represented in the exhaust emissions by
gaseous products of incomplete combustion (beyond CO) made up of a wide variety
of organic species.
The primary species that constitute gaseous aerosol precursors in aircraft
gas turbine engines are components of the families of SOx and organic emissions
(Table 6.1). Because the largest mass of SOx is emitted as SO2 and the majority
of the organic emissions are small (C1, C2, C3 ) molecules, the gaseous aero-
sol precursors are a smaller subset of the SOx and organic species that have suf-
ficiently low-vapor pressures and contribute to the condensed phase at ambient
conditions. In the case of SOx, the relevant specie is sulfuric acid, H2SO4. For the
organic species, the relevant species include larger hydrocarbons (including PAHs)
and partially oxidized HCs resulting from the incomplete combustion of the fuel
HCs. These organic emissions increase at low-power operation, when combustion
efficiency drops off, although in concert with much lower fuel usage as well at these
lower-power conditions, and decrease dramatically at powers higher than about
idle. The organic species range from (1) fuel hydrocarbons (HCs) and fragments
of fuel HCs through (2) partially oxidized HCs and (3) aromatic species such as
polycyclic aromatic HCs (PAHs) formed during the combustion process, to (4)
lubrication oil that enters the exhaust flow. These two classes of species, SOx and
organics, are the primary gaseous aerosol precursors that play a role in particles
present in the exhaust plumes of aircraft engines.
6.1 Introduction 157
Table 6.1. Sources of species that contribute to volatile PM mass in near-field aircraft plumes
Engine sources Engine combustor exit Engine turbine and Near-field plume species
species early plume species
Note: Both SOx and organic emissions in the near-field plume are mostly gaseous species, and only a small fraction
of SOx becomes condensable H2SO4 and only a small fraction of the emitted HCs, see Section 6.1.4, become con-
densable organic species.
The existing regulations require the measurement of CO, NOx, and unburned
organic emissions for the certification of aircraft gas turbine engines above a thrust
rating of 26.7 kN (6,000 lb) (Lister and Norman, 2003). Of these, CO plays no role in
particle processes. NOx has long been considered an important pollutant, and while
the role of aviation NOx in contributing to particles in the atmosphere is understood
(Woody et al., 2011), measurements of aircraft plumes indicate that NOx does not
play an important role in particle microphysics close to the aircraft. Since NOx and
CO are discussed extensively in Chapter 7, they will not be discussed further here in
any detail.
6.1.2 Existing Regulations on SOx (via Fuel Sulfur Level) and UHC
SOx: Most petroleum stock contains some level of sulfur in such species as sulfides,
disulfides, and benzothiophenes, which can vary greatly according to the petroleum
feedstock source. The level of sulfur in the fuel is constrained by fuel specification
for Jet A or Jet A1 to be less than 3,000 ppmm (parts per million by mass) or 0.3
weight percent. Most fuels fall well below that specification, and the future possibil-
ity exists of mandated reduction in fuel sulfur content in aviation fuel, as has been
in force for diesel fuels in many parts of the world in recent years. Typical fuel sulfur
levels for aviation jet fuel are in the range of 2001200, and emission levels of SOx
are limited by the fuel sulfur content, which is constrained more by feedstock and
refinery logistical issues than by the regulatory limit.
UHCs: Historically, the means of controlling organic emissions collectively has
been through a certification requirement using a Flame Ionization Detector (FID)
to quantify unburned hydrocarbons (UHCs). However, not all organic species are
measured equally well with the FID, and not all of the species have the same health
158 Gaseous Aerosol Precursors
or environmental impacts, and only a modest subset of these organics have a vapor
pressure low enough to contribute to particle processes immediately in plume. A
larger fraction of the organic emissions can contribute, later in the atmosphere after
further processing, to ambient atmospheric particle processes (Miracolo et al., 2011;
Presto et al., 2011), but remain in the gas phase through the initial exhaust plume.
For all of organic emissions, interest in addressing the emission levels of specific spe-
cies is growing (Wood et al., 2008) because of their individual toxic properties, their
carcinogenic tendencies, their role in pollution chemistry, and, per the focus of this
chapter, their role in forming particles.
The current regulatory FID technique measures the carbonaceous component
of the gaseous emissions by measuring the ionization of carbon atoms in a hydro-
gen flame. Thus, in that sense, the FID counts carbons and identifies the carbon
content of the carbon-containing gases emitted. Thus, there is no speciation of those
carbon-containing gases, only a quantification of the total amount of carbon collec-
tively in the gaseous species (one carbon for methane, two for ethene, etc.).
There are also limitations with the FID measurement. The first is that, as
already stated, there is no speciation information. A measurement of ethene does
not look any different than a measurement of pyrene once one accounts for ethene
having two carbons and pyrene having sixteen carbons. These two species would
look the same to an FID if the ethene had a eightfold higher molar concentra-
tion to account for the different carbon contents (or, equivalently, if their mass
concentrations were adjusted to account for their differing carbon contents). Yet,
in the context of gaseous aerosol precursors, pyrene is an important condensable
species, while ethene will not contribute to aerosol processes in the plume. Thus
the current regulation on UHC and the method used to measure these species
does not distinguish between UHCs that remain in the gas phase and those that
can contribute to aerosol mass (usually a small, but important fraction of the total
UHC emissions).
The second limitation of the FID is that, because of the chemistry associated
with the ionization process, any oxygenated HCs have a reduced sensitivity in the
FID measurement. Basically, any carbon bonded to an oxygen atom in the initial
species chemical structure is not counted by the FID measurement. Thus, form-
aldehyde, an important UHC emission for aircraft engines, is not accounted for by
an FID measurement. Larger aldehydes (e.g., the C4 aldehyde, butanal) contribute
to the UHC measurement with one less carbon than they actually contain (butanal
is counted as three carbons, even though it has four) because of having one already
oxygen-bonded carbon. This limitation has two ramifications in that the FID mea-
surement under-counts the HC emissions and that it measures with a reduced sensi-
tivity some fraction of oxygenated species that have lower vapor pressures and play
an enhanced role in aerosol processes.
Beyond these problems with the FID measurement for organic gaseous aerosol
precursors, its quantification is further confounded by the fact that the emission lev-
els of organics decrease in concentration as the carbon number increases and that
the larger molecules that participate in condensation processes have, per force, a
6.1 Introduction 159
5
Total HC (FID)
PTR-MS:
methyl naphthalene
Carbon emission ratio, ERC (ppb C/ppm CO2)
naphthalene
C5-benzene
4 C4-benzene
C3-benzene
m/z 107
styrene
hexanal
3 phenol
toluene
m/z 83
high ralkenes
benzene
2 m/z 73
m/z 59
m/z 57
acetaldehyde
m/z 43
1 m/z 41
TILDAS:
C2H4
HCHO
0
4 5.5 7 15 30 40 60 65 70 75 85 100
Engine power (%)
Figure 6.2. Speciated hydrocarbon emissions varying with engine power (Yelvington et al.,
2007, see also Knighton et al., 2007) comparing FID measurements with more detailed infra-
red (TILDAS) and mass spectrometric (PTR-MS) measurements. The most relevant con-
densable species are larger than even the naphthalenes presented here and include larger
alkanes, PAHs, and oxygenates such as organic acids (credit NASA/TM-2006214382 APEX
Report).
The processes by which particles form, grow, and interact have received increas-
ing attention in the last several decades as their impact on human health and envi-
ronmental pollution have been brought into focus, both for scientific understanding
and for regulatory control. These microphysical processes that control particle
properties such as number, size, composition, and morphology have motivated great
advances in understanding and in measurement capabilities in the same time period.
160 Gaseous Aerosol Precursors
Attention on aircraft PM has also become a focus, and significantly better under-
standing of aviation particle microphysical processing has come about.
The sulfur-containing compounds in jet fuel are oxidized with good efficiency in
an aircraft combustor, such that most of the fuel sulfur is emitted as SO2. This SO2
is important for particles in the atmosphere, since its primary removal process takes
place via further oxidation to SO3 and then interaction with H2O to form sulfuric
acid. Sulfuric acid has a very low-vapor pressure and is strongly favored to condense
in existing or new particles. The SO2 that leaves aircraft engines mixes with ambient
gases and later oxidizes and contributes to particles and to chemistry and particle
microphysics in concert with the ambient pollutant levels where the emissions occur,
whether in and around airports or in the upper atmosphere during cruise. This pro-
cessing occurs on time scales much longer than the plume of the aircraft.
Even though most of the fuel sulfur is emitted as SO2, a few percent or so is
emitted in the exhaust as SO3 (Krcher et al., 2000 and references therein). While
this seems like an almost negligible amount, the very low-vapor pressure and high
nucleation propensity of sulfuric acid makes the modest emission levels of this
more oxidized form of the sulfur emissions important in the immediate vicinity
of the emissions. The emitted SO3 is rapidly converted to H2SO4 (sulfuric acid)
(Kolb et al., 1994), with the abundance of combustion water available, and the sul-
furic acid can form new particles, as well as form coatings on emitted soot particles
(Miake-Lye etal., 1994, Krcher et al., 1995, Brown et al., 1996). So, even though a
small fraction of the fuel sulfur is available as sulfuric acid when first emitted, the
high propensity for sulfuric acid to form new particles and condense on existing
particles makes SOx an important emission for particles both in the plume and later
in the atmosphere.
The role of sulfuric acid in forming new particles in the near-field aircraft plume
was first studied (Miake-Lye et al., 1994; Zhao and Turco, 1995) in environmental
programs focused on stratospheric flight of supersonic aircraft and the potential for
stratospheric impacts such as ozone depletion. Detailed microphysical models of
sulfuric acid aerosol formation were developed (Krcher et al, 1995; Brown et al.,
1996), and flight measurements (Fahey et al., 1995; Anderson et al., 1998; Toon and
Miake-Lye, 1998; Schumann et al., 2002) corroborated that volatile particles were
present in aircraft plumes in flight.
Quantification of the sulfur levels associated with those particles presents a
challenge because of the difficulty in obtaining accurate measurements of the reac-
tive and sticky sulfur acid species, but estimates that up to a few percent of the fuel
sulfur is converted to sulfuric acid were broadly supported (Krcher et al., 2000). In
hindsight, one of the additional complicating factors in quantifying the sulfur levels
was the fact that the volatile particles have organic contributions as well as sulfate
contributions. This is important in determining the mass to use in calculating the
conversion of fuel sulfur to SO3 or H2SO4, because not all the volatile particle mass
is due to SOx when organics are also contributing. In ground measurements, the
organic contributions vary widely between engine designs, and need to be accounted
for properly to quantify the specific contribution from sulfuric acid.
6.1 Introduction 161
From modeling results (Krcher et al., 1995; Brown et al., 1996; Krcher and
Yu, 2009; Wong and Miake-Lye, 2010), researchers have learned that the very high
nucleation potential of sulfuric acid results in the rapid formation of many very
small particles composed of clusters of sulfuric acid and water. These numerous
small particles coagulate and grow to form a small nucleation mode of particles
determined by sulfuric acid concentrations at the engine exit plane and mediated
by the dilution history of the exhaust in the plume and wake of the airplane. The
dilution history determines the temperature and mixing of the exhaust with the
ambient air, and this determines how the available sulfuric acid partitions to new
particle formation, new particle growth, and deposition on existing soot particles.
Thus, the volatile contributions to particulate matter from sulfate add to particle
number, through new particle formation and to particle mass, both because of
these growing new particles and because of coatings on soot particles. The balance
between volatile contributions to the nucleation mode new particles and volatile
contributions as coatings on soot particles depends on levels of the emitted sulfuric
acid and the number of soot particles, in combination with the mixing history of the
exhaust plume.
While most modeling work has focused on soot, sulfuric acid, and water, recent
work includes the interactions of organic species with these other emissions (Jun,
2011). This new work explores the role of organics species in adding to the mass of
both nucleation mode particles and coatings on soot particles. As such, it focuses
attention on organic species that have the relevant physical properties to activate
soot surfaces and to participate in the condensation processes, represented as pyrene,
benzo[a]pyrene, coronene, acetone, propanoic acid, and butanoic acid in studies to
date. However, the dominant role of sulfuric acid in nucleating new particles is rein-
forced in these modeling studies, given the range of organic species concentrations
available and the expected sulfuric acid levels with current fuel sulfur levels.
Based on the early recognition of the role of sulfuric acid in nucleating and contrib-
uting to volatile particles, initial focus on volatile PM emissions revolved around
quantifying the sulfate levels and contributions (Fahey et al., 1995; Brown et al.,
1996; Schumann et al., 2002). Later, detailed size-resolved compositional measure-
ments of aircraft PM were made to quantify PM emissions from the existing com-
mercial fleet using an aerosol mass spectrometer (Anderson et al., 2005; Onasch
et al., 2009; Timko et al., 2010). These compositional measurements clearly demon-
strated that organic species not only make significant contributions to volatile PM,
but typically dominate the volatile PM mass at low-power, near-idle engine operat-
ing conditions (Onasch et al., 2009; Timko et al., 2010). At higher powers, where
the engines are operating closer to peak combustion efficiency, the balance between
sulfate and organic often becomes more nearly equal in the near-field plume, but
organics are still important for current engine technologies and representative fuel
sulfur levels for fuels today.
162 Gaseous Aerosol Precursors
Of the organic species emitted by aircraft, only a small fraction has a sufficiently
low-vapor pressure to participate in condensation processes in the near-field plume.
Comprehensive organic speciation measurements of aircraft exhaust emissions have
resulted in a detailed species profile of these emissions, especially at low-power
engine operation (Spicer et al., 1992, 1994; Knighton et al., 2007; Yelvington et al.,
2007; Wood et al., 2008). The EPA SPECIATE database has documented a robust
and relatively invariant emissions profile from operations using standard Jet A fuel
(Table 6.2; FAA-EPA, 2009). Commercial aircraft gas turbine engines burn Jet A or
Jet A1, and most commercially available jet fuel has a typical fuel composition some-
what narrower than the fuel specification, notably narrower for aromatic species and
for sulfur content. Interesting, burning commonly available jet fuel, the total amount
of organic emissions varies significantly with power and ambient temperature, but
the organic speciation profile is relatively insensitive to the varying level of organic
emissions, whether because of changes in the engine power near idle, in ambient
conditions, or in the specific engine technology tested. That is to say, as the organic
emissions rise or fall, their relative proportions do not change much, and the individ-
ual species concentrations mostly rise and fall together for near-idle operation and
are similar for all aircraft gas turbine engine types (Knighton et al., 2007). However,
most of these organics have too high a vapor pressure to contribute to particles.
It is worth noting that many of the organic emissions, both those that remain
gaseous and those that are aerosol precursors, are considered hazardous air pol-
lutants (HAPs) by the EPA, and thus their better quantification is motivated for a
number of reasons. In this context, it is also relevant that the nonvolatile soot emis-
sions are also considered HAPs by the EPA and that some organics, like PAHs, are
produced via the same pyrolytic pathways that result in soot formation. Thus, the
PAH organic aerosol precursors are produced in parallel with the nongaseous HAP
represented by soot.
The larger organic molecules that participate in particle processes have low con-
centrations at all power conditions (Knighton et al., 2007; Yelvington et al., 2007;
Wood et al., 2008), and do not decrease as sharply as the lighter HCs that dominate
the SPECIATE profile (Timko et al., 2010). A detailed profile of these condensed
species has not been determined yet, but a number of studies (Corporan et al., 2007;
Agrawal et al., 2008; Kinsey et al., 2011) suggest that PAHs and organic acids are
important, among others. Detailed analysis of the aerosol mass spectrometer data
using a method called positive matrix factorization identifies three classes of exhaust
species contributing to the organic PM contributions in the plume (Timko et al.,
2010), along with any ambient background organics that may participate. These
classes are aliphatics, aromatics, and lubrication oil. Presumably, a wide range of
related species contributes small concentrations, especially for the aliphatics and
aromatics, which are measured as sets of spectral peaks contributing to the total
organic measurement of the aerosol mass spectrometer.
Despite the relative invariance of the organic emissions profile to engine type
and power, recent work on alternative fuels has shown that dramatic changes in
fuel composition, especially changes in the aromatic content, can have dramatic
Table 6.2. Speciated gas-phase organic gas profile for aircraft equipped with turbofan, turbojet,
and turboprop enginesc
a
CAS = Chemical Abstracts Service b See discussion of unidentified species in Section 2.1 of this report. For commercial,
military, general aviation, and air taxi aircraft equipped with turbofan, turbojet, and turboprop engines. d Identified as
a HAP in Section 112 of the CAA (shaded above). * Identified in IRIS as having toxic characteristics (shaded above).
f
Values were adjusted from those shown in the Technical Support Document to account for rounding and to facilitate
inclusion of the data in the SPECIATE database (where the required sum of the values is 1.00000).
Note: Values in this table may be revised in the future as additional engine data are available.
Source: FAA-EPA, 2009
164 Gaseous Aerosol Precursors
At the high temperatures present in aircraft engine combustion, the sulfur contained
in the fuel hydrocarbon matrix is completely oxidized to SO2. At these high tem-
peratures, additional oxidation to higher oxidation states is not thermodynamically
favored (SO2 is S[IV]; SO3 and H2SO4 are hexavalent S[VI]), so the oxidized sulfur
leaves the combustor as SO2 (Tremmel and Schumann, 1999; Lukachko et al., 2008;
and references therein). As the exhaust passes through the turbine and exhaust noz-
zle, the exhaust gases cool and expand and the thermodynamic state favors complete
conversion to the higher oxidation state S(VI). However, the driving oxidative reac-
tions with OH (and to a lesser extent atomic oxygen) are not sufficiently fast with
166 Gaseous Aerosol Precursors
2000
0.0005%
0.0001% 3%
0.00
0.001%
5%
0.00 %
0.0 7
0
1500 0.01%
Temperature (K)
0.1%
1000 1%
5%
500
0 10 20 30 40
Pressure (atm)
Figure 6.3. Map of SOx (SO3) production potential and representative p-T trajectories
through the combustor and turbine as p and T drop before the engine exit. S(VI) produc-
tion occurs in the turbine where the higher potentials are realized, but the total conversion is
kinetically limited to the order of a percent or so (Lukachko et al., 2008; originally published
by ASME).
the available species concentrations and time spent during the cooling and expan-
sion to completely transition from S(IV) to S(VI). Thus the amount of S(VI) at the
engine exit plane is kinetically controlled.
Since the measurement of SO3 and H2SO4 is difficult and the resulting conversion
is kinetically limited to being only a few percent (Krcher et al., 2000 and references
therein), precise quantification of the sulfur conversion is still not straightforward. If
direct measurements of SO3 and H2SO4 were easier, their direct quantification would
be routine. If the overall conversion were more than a few percent, the quantification
of SO2 and the fuel sulfur level could be subtracted to determine the missing sulfur
represented by the difficult-to-measure S(VI) species. However, the conversion is
small enough that uncertainties in measurements of SO2 and fuel sulfur levels are
typically as large or larger than the expected S(VI) fractions.
However, modeling results (Tremmel and Schumann, 1999; Lukachko et al.,
2008) and the best measurements of sulfate in the exhaust constrain the fuel sulfur
conversion to S(VI) species to be several percent with significant uncertainties. Firm
lower bounds indicate that the conversion must be more than a few tenths of a per-
cent (Onasch et al., 2009; Timko et al., 2010), and upper bounds are generally below
or well below 10 percent, so roughly a percent to several percent appears a broad
and robust estimate. Modeling results suggest that engine power setting and engine
cycle variation (especially as higher pressure/temperature cycles are considered) have
some effect (Lukachko et al., 2008), but these variations are still modest compared to
remaining uncertainties in simply measuring the conversion itself (Figure 6.3).
6.3 Organic Precursors Formation 167
Clearly, the sulfur in the fuel directly determines the amount of total sulfur in the
exhaust. Again, the Jet A fuel specification limits the total fuel sulfur content to
below 3,000 ppmm, and this and the overall fuel-air ratio of the engine determine
the maximum concentrations possible at the engine exit plane. As mentioned previ-
ously, the conversion of SO2 to S(VI) species is kinetically controlled in the turbine,
so other details of the fuel and combustion process occurring upstream in the com-
bustor have little impact on either the total SOx levels or the SOx speciation among
SO2, SO3, and H2SO4. However, fuel sulfur level does control the total SOx emitted,
and this may be important as the industry considers reducing fuel sulfur for a variety
of environmental and maintenance reasons, as has already been accomplished for
diesel fuel for ground vehicles. This question of reductions in fuel sulfur content
may become even more important as alternative fuels, such as biofuel and other
nonpetroleum HC sources, are considered for powering aviation. Alternative fossil
sources for jet fuel, such as Fischer-Tropsch synthesis of jet fuel from coal or natural
gas, can also have very low sulfur levels, and these fuels are of interest for reduced
dependence on petroleum fuel supplies, even if they are not as attractive from a sus-
tainability or carbon footprint perspective.
the particle phase, the pathway and detailed organic speciation may depend on the
power setting and the chemical pathways dominant for those engine conditions.
While the effect of fuel sulfur on emitted SOx is a simple and direct relationship, the
impact of fuel properties on emitted organic species is more complicated. Emitted
organics depend partly on the fuel HC matrix, but also on combustion chemistry.
Work to date suggests that for the typical narrow range of Jet A/Jet A1 (often much
narrower than the specification) the emitted organic species are very similar, as
discussed previously and documented in the EPA SPECIATE database. However,
recent tests with alternative fuels (Anderson et al., 2011) suggest significant variation
6.4 Summary and Open Questions 169
in PM organic in the plume can result when different fuels are used. Only a limited
number of fuels have been explored, and systematic understanding of the connections
between fuel composition and organic emissions, including light volatile species and
low-vapor pressure condensable species, awaits further experimental and theoretical
exploration. And, with the many and significant constraints on aviation fuel proper-
ties for proper operation across typical engine operating profiles, wide variations of
fuel composition are not possible. However, the decrease in total organic emissions
and organic PM levels observed with low-aromatic alternative fuels suggests there is
room for optimization of fuel properties that could result in significant reduction in
PM, both for soot and for volatile contributions from organic aerosol precursors, via
a tailoring of the HC matrix in the fuel used.
For the global atmosphere, like SOx, the role of organic PM in affecting climate is
through how it mediates the aircraft PM properties. Since the biggest questions per-
tain to how aviation particles affect cloud cover, the question for organic PM is: How
are aircraft-emitted PM properties affected by the organic contributions? Is cloud
condensation nucleation affected by soot, and is that different depending on how
the organic coating on the soot interacts with water vapor in the upper atmosphere?
Research continues in this area in laboratory studies and limited in-flight field cam-
paigns, but many first order questions have still to be answered.
For local air quality, an understanding of the composition of aviation PM could
be very important for interpreting health effects due to small particles and the pop-
ulation exposed to them. While strong evidence suggests that aged atmospheric
particles have organic PM contributions that all converge to similar, highly oxi-
dized compositions (Ng et al., 2010), the initial composition can be very source
dependent. So, for populations close to the source, initial PM compositions may
still be quite important, especially if the composition includes particularly toxic
species. For assessing the impact on these nearby exposed populations, the relevant
species in the particles as well as the total amounts emitted need to be quantified.
For instance, in the case of flight line workers and communities close to airports,
these specific condensable organics present in the near-field emitted PM may be
important.
6.4.1 Summary
Gases emitted in the exhaust of aircraft gas turbine engines include a number of spe-
cies that contribute to the formation and growth of particles in the near-field aircraft
plume. The important gaseous precursors that contribute to PM mass are sulfuric
acid (H2SO4), a variety of larger organic species related to the combustion of jet fuel,
and engine-lubricating oil compounds. Other gaseous species like NOx and SO2, and
170 Gaseous Aerosol Precursors
many other organic emissions, remain in the gas phase through the plume, but can
contribute to atmospheric PM processes on time scales of hours or days.
These gaseous precursors add to PM mass in the plume by creating new volatile
particles, driven by the H2SO4 formed in small amounts from fuel sulfur compounds,
and by condensing and forming coatings on existing soot that left the engine as non-
volatile particles. The H2SO4 is typically only a few percent of the total sulfur that
leaves as SO2, yet is key in forming new particles for current petroleum-derived
aviation fuels.
Organics that contribute to PM in the plume arise from incomplete oxidation
of the jet fuel and from pyrolysis of the fuel that occurs in parallel with the produc-
tion of carbonaceous soot particles. The total amounts of the organics are affected
by engine power conditions, fuel composition, and ambient conditions, and, in turn,
affect the amounts of organic contribute to the new volatile particles and to coat-
ings on soot and, thus, the increases in PM mass. Lubrication oil can also add organic
mass to PM, and is an emissions source separate from the combustion processes in
the engine combustor. Lube oil has not traditionally been considered a contributor
to aircraft emissions, and thus the lube oil emissions loading varies widely across dif-
ferent engine types.
These volatile contributions to PM are of scientific and regulatory concern for
their potential effects on climate and on local air quality because of their associated
health effects. While current regulations focus on the mass of the total particles emit-
ted, increasing understanding of the size-dependent composition of newly formed
volatile particles and coatings on the larger soot particles may focus attention on
specific compounds as well as on the effects of number, size, morphology, and com-
position on the eventual impacts of the emitted PM.
Because of its relative infancy as a research topic, there is much to be learned about
the contributions of gaseous precursors to volatile PM in the near-field plume, as
well as to contributions to atmospheric aerosol on the regional and global scales.
Because of the dominant role of fuel sulfur in initiating new particle formation
and in activating soot surfaces for condensation, the level of sulfur in commercial
aviation fuel is a key parameter in determining the concentrations of the gaseous
precursor sulfuric acid. Thus, quantification and possible control of the fuel sul-
fur content in commercial aviation fuel supplies will remain key for volatile PM
processes.
While fuel sulfur plays a controlling role, the mass of organics in volatile PM is
as large or larger than that from sulfur with current fuels. Thus, their role and which
species are most important needs to be better understood. Significant advances have
occurred in recent years, but more accurate and more detailed measurements and
more advanced predictive tools are important to better understand how microphysi-
cal processes and engine and fuel technologies can affect the composition of par-
ticles formed in the aircraft plume and beyond.
6.4 Summary and Open Questions 171
in the plume of operating aircraft. As such, the nature of gaseous precursors and the
volatile PM that they form can be described, and the contributions that they make
to ambient aerosol can be bounded. Thus, the question of how they may impact the
environment can be properly posed, and will allow future research to provide more
detailed and specific assessments in coming years.
References
Agrawal, H., Sawant, A. A., Jansen, K., Miller, J. W., and Cocker, D. R. (2008). Characterization
of Chemical and Particulate Emissions from Aircraft Engines. Atmospheric Environment
42: 438092.
Anderson, B. E., Beyersdorf, A. J., Hudgins, C. H., Plant, J. V., Thornhill, K. L., Winstead, E.L.,
Ziemba, L.D., Howard, R., Corporan, E., Miake-Lye, R.C., Herndon, S. C., Timko, M.,
Woods, E., Dodds, W., Lee, B., Santori, G., Whitefield, P., Hagen, D., Lobo, P., Knighton,
W. B., Bulzan, D., Tacina, K., Wey, C., Vander Wal, R., Bhargava, A., Kinsey, J., and
Liscinsky, D.S. (2011). Alternative Aviation Fuel Experiment (AAFEX). NASA/
TM2011217059.
Anderson, B. E., Branham, H.-S., Hudgins, C. H., Plant, J. V., Ballenthin, J. O., Miller, T. M.,
Viggiano, A. A., Blake, D. R., Boudries, H., Canagaratna, M., Miake-Lye, R. C., Onasch,
T. B., Wormhoudt, J. C., Worsnop, D. R., Brunik, K. E., Culler, C., Penko, P., Sanders, T.,
Han, H.-S., Lee, P., Pui, D. Y. H., Thornhill, K. L., and Winstead, E. L. (2005). Experiment
to Characterize Aircraft Volatile Aerosol and Trace-Species Emissions (EXCAVATE).
NASA Report No. NASA/TM-2005213783.
Anderson, B. E., Cofer, W. R., Bagwell, D. R., Barrick, J. W., Hudgins, C. H., and Brunke,
K.E. (1998). Airborne Observations of Aircraft Aerosol Emissions 1: Total Nonvolatile
Particle Emission Indices. Geophysical Research Letters 25: 168992.
Brown, R. C., Miake-Lye, R. C., Anderson, M. R., Kolb, C. E., and Resch, T. J. (1996). Aerosol
Dynamics in Near-Field Aircraft Plumes. Journal of Geophysical Research (Atmospheres)
101(D17): 2293953.
Corporan, E., DeWitt, M. J., Belovich, V., Pawlik, R., Lynch, A. C., Gord, J. R., and Meyer,
T. R. (2007). Emissions Characteristics of a Turbine Engine and Research Combustor
Burning a Fischer-Tropsch Jet Fuel. Energy & Fuels 21: 261526.
Fahey, D. W., Keim, E. R., Boering, K. A., Brock, C. A., Wilson, J. C., Jonsson, H. H., Anthony,
S., Hanisco, T. F., Wennberg, P. O., Miake-Lye, R. C., Salawitch, R. J., Lousinard, N.,
Woodbridge, E. L., Gao, R. S., Donnelly, S. G., Wamsley, R. C., Del Negro, L. A., Solomon,
S., Daube, B. C., Wofsy, S. C., Webster, C. R., May, R. D., Kelly, K. K., Loewenstein, M.,
Podolske, J. R., and Chan, K. R. (1995). Emission Measurements of the Concorde
Supersonic Aircraft in the Lower Stratosphere. Science 270(5233): 704.
Federal Aviation Administration (FAA) (2009). Office of Environment and Energy, and
U.S. Environmental Protection Agency, Office of Transportation and Air Quality.
Recommended Best Practice for Quantifying Speciated Organic Gas Emissions from
Aircraft Equipped with Turbofan, Turbojet, and Turboprop Engines. Version 1.0, May 27.
Also Knighton, W. B., Herndon, S. C., and Miake-Lye. R. C. Aircraft Engine Speciated
Organic Gases: Speciation of Unburned Organic Gases in Aircraft Exhaust. Technical
Support Document for Recommended Best Practice Version 1.0.
Jun, M. (2011). Microphysical Modeling of Ultrafine Hydrocarbon-Containing Aerosols
in Aircraft Emissions. PhD Thesis, Department of Aeronautics and Astronautics,
Massachusetts Institute of Technology, May.
Krcher, B., Peter, T., and Ottmann, R. (1995). Contrail Formation Homogeneous Nucleation
of H2SO4/ H2O Droplets. Geophysical Research Letters 22(12): 15014.
Krcher, B., Turco, R. P., Yu, F., Danilin, M. Y., Weisenstein, D. K., Miake-Lye, R. C., and
Busen, R. (2000). A Unified Model for Ultrafine Aircraft Particle Emissions. Journal of
Geophysical Research (Atmospheres) 105(D24): 2937986.
References 173
Krcher, B., and Yu, F. (2009). Role of Aircraft Soot Emissions in Contrail Formation.
Geophysical Research Letters 36: L01804, doi:10.1029/2008GL036649.
Kinsey, J. S., Hays, M. D., Dong, Y., Williams, D. C., and Logan, R. (2011). Chemical
Characterization of the Fine Particle Emissions from Commercial Aircraft Engines dur-
ing the Aircraft Particle Emissions eXperiment (APEX) 1 to 3. Environmental Science
Technology 45: 341521.
Knighton, W. B., Rogers, T. M., Anderson, B. E., Herndon, S. C., Yelvington, P. E., and
Miake-Lye, R. C. (2007). Quantification of Aircraft Engine Hydrocarbon Emissions
using Proton Transfer Reaction Mass Spectrometry. Journal of Propulsion and Power
23(5): 94958.
Kolb, C. E., Jayne, J. T., Worsnop, D. R., Molina, M. J., Meads, R. F., and Viggiano, A. A. (1994).
Gas-Phase Reaction of Sulfur-Trioxide With Water-Vapor. Journal of the American
Chemical Society 116(22): 1031415.
Lee D. S., Pitari, G., Grewe, V., Gierens, K., Penner, J. E., Petzold, A., Prather, M. J., Schumann,
U., Bais, A., Berntsen, T., Iachetti, D., Lim, L. L., and Sausen, R. (2010). Transport Impacts
on Atmosphere and Climate: Aviation. Atmospheric Environment 44: 4678734.
Lister, D. H., and Norman, P. D. (2003). Aircraft Engine Emissions Certification: A Review of
the Development of ICAO Aneex 16, Volume II. EC-NEPAir: Work Package 1, QinetiQ/
FST, CR030440, GRD-CT-200000182.
Lukachko, S. P., Waitz, I. A., Miake-Lye, R. C., and Brown, R. C. (2008). Engine Design and
Operational Impacts on Particulate Matter Precursor Emissions. Journal of Engineering
for Gas Turbines and Power 130(2): 021505.
Miake-Lye, R. C., Brown, R. C., Anderson, M. R., and Kolb, C. E. (1994). Calculations of
Condensation and Chemistry in an Aircraft Contrail, in Impact of Emissions from Aircraft
and Spacecraft upon the Atmosphere, Schumann U. and Wurzel, D. eds., Proceedings of
an International Scientific Colloquium, Cologne, Germany, April 1820, DLR-Mitteilung
9406, Deutsches Zentrum fr Luft- und Raumfahrt, Oberpfaffenhofen and Cologne,
Germany.
Miracolo, M. A., Hennigan, C. J., Ranjan, M., Nguyen, N. T., Gordon, T. D., Lipsky, E. M., Presto,
A. A., Donahue, N. M., and Robinson, A. L. (2011). Secondary Aerosol Formation from
Photochemical Aging of Aircraft Exhaust in a Smog Chamber. Atmospheric Chemistry
and Physics 11: 413547.
Ng, N. L., Canagaratna, M. R., Zhang, Q., Jimenez, J. L., Tian, J., Ulbrich, I. M., Kroll, J. H.,
Docherty, K. S., Chhabra. P. S., Bahreini, R., Murphy, S. M., Seinfeld, J. H., Hildebrandt,
L., Donahue, N. M., DeCarlo, P. F., Lanz, V. A., Prevot, A. S. H., Dinar, E., Rudich, Y.,
and Worsnop, D. R. (2010). Organic Aerosol Components Observed in Northern
Hemispheric Datasets Measured with Aerosol Mass Spectrometry. Atmospheric
Chemistry and Physics 10: 462541.
Oberdrster, G., Oberdrster, E., and Oberdrster, J. (2005). Nanotoxicology: An Emerging
Discipline Evolving from Studies of Ultrafine Particles. Environmental Health
Perspectives 113(7): 82339.
Oluwole, O. O., Barton, P. I., and Green, W. H. (2007). Obtaining Accurate Solutions Using
Reduced Chemical Kinetic Models: A New Model Reduction Method for Models
Rigorously Validated over Ranges. Combustion Theory and Modelling 11: 12746.
Onasch, T. B., Jayne, J. T., Herndon, S., Worsnop, D. R., Miake-Lye, R. C., Mortimer, I. P., and
Anderson B. E. (2009). Chemical Properties of Aircraft Engine Particulate Exhaust
Emissions. Journal of Propulsion and Power 25(5): 112137.
Penner, J. E., Lister, D. H., Griggs, D. J., Dokken, D. J., and McFarland, M. (1999). Aviation and
the Global Atmosphere: A Special Report of the Intergovernmental Panel on Climate
Change. Cambridge University Press.
Presto, A. A., Nguyen, N. T., Ranjan, M., Reeder, A. J., Lipsky, E. M., Hennigan. C. J, Miracolo,
M. A., Riemer, D. D., and Robinson, A. L. (2011). Fine Particle and Organic Vapor
Emissions from Staged Tests of an In-use Aircraft Engine. Atmospheric Environment
45: 360312.
174 Gaseous Aerosol Precursors
Schumann, U., Arnold, F., Busen, R., Curtius, J., Krcher, B., Kiendler, A., Petzold, A., Schlager,
H., Schroder, F., and Wohlfrom, K. H. (2002). Influence of Fuel Sulfur on the Composition
of Aircraft Exhaust Plumes: The Experiments SULFUR 17. Journal of Geophysical
Research (Atmospheres) 107(D15): 4247.
Spicer, C. W., Holdren, M. W., Riggin, R. M., and Lyon, T. F. (1994). Chemical Composition
and Photochemical Reactivity of Exhaust from Aircraft Turbine Engines. Annales
Geophysicae 12(1011): 94455.
Spicer, C. W., Holdren, M. W., Smith, D. L., Hughes, D. P., and Smith, M. D. (1992). Chemical
Composition of Exhaust from Aircraft Turbine Engines. Journal of Engineering for Gas
Turbines and Power 114(1): 11117.
Timko, M. T., Onasch, T. B., Northway, M. J., Jayne, J. T., Canagaratna, M. R., Herndon, S.
C., Wood, E. C., Miake-Lye, R. C., and Knighton, W. B. (2010). Gas Turbine Engine
Emissions Part II: Chemical Properties of Particulate Matter. Journal of Engineering
for Gas Turbines and Power 132: 061505.
Toon, O. B., and Miake-Lye, R. C. (1998). Subsonic Aircraft: Contrail and Cloud Effects
Special Study. Geophysical Research Letters, 25(8): 110912.
Tremmel, H. G., and Schumann, U. (1999). Model Simulations of Fuel Sulfur Conversion
Efficiencies in an Aircraft Engine: Dependence on Reaction Rate Constants and Initial
Species Mixing Ratios. Aerospace Science and Technology 7: 41730.
Wey, C. C., Anderson, B. E., Hudgins, C., Wey, C. C., Li-Jones, X., Winstead, E., Thornhill,
L.K., Lobo, P., Hagen, D., Whitefield, P., Yelvington, P. E., Herndon, S. C., Onasch, T. B.,
Miake-Lye, R. C., Wormhoudt, J., Knighton, W. B., Howard, R., Bryant, D., Corporan,
E., Moses, C., Holve, D., and Dodds, W. (2006). Aircraft Particle Emissions Experiment
(APEX). Report No. NASA/ TM-2006214382.
Wey, C. C., Anderson, B. E., Wey, C. C., Miake-Lye, R. C., Whitefield, P., and Howard, R. (2007).
Overview on the Aircraft Particle Emissions Experiment. Journal of Propulsion and
Power 23: 897905.
Wong, H.-W., and Miake-Lye R. C. (2010). Parametric Studies of Contrail Ice Particle
Formation in Jet Regime Using Microphysical Parcel Modeling. Atmospheric Chemistry
and Physics 10: 326172.
Wood, E., Herndon, S., Miake-Lye, R. C., Nelson, D., and Seeley M. (2008). Aircraft and
Airport-Related Hazardous Air Pollutants: Research Needs and Analysis. ACRP
Project 0203, ACRP Report 7, ISBN 9780309117456, Transportation Research
Board, Washington, DC.
Woody, M., Baek, B. H., Adelman, Z., Omary, M., Lam, Y. F., West, J. J., and Arunachalam, S.
(2011). An Assessment of Aviations Contribution to Current and Future Fine Particulate
Matter in the United States. Atmospheric Environment 45: 342433.
Yelvington, P. E., Herndon, S. C., Wormhoudt, J. C., Jayne, J. T., Miake-Lye, R. C., Knighton,
W. B., and Wey, C. (2007). Chemical Speciation of Hydrocarbon Emissions from a
Commercial Aircraft Engine. Journal of Propulsion and Power 23: 91218.
Zhao, J., and Turco, R. P. (1995). Nucleation Simulations in the Wake of a Jet Aircraft in
Stratospheric Flight. Journal of Aerosol Science 26(5): 77995.
7 NOx and CO Formation and Control
Ponnuthurai Gokulakrishnan and Michael S. Klassen
7.1 Introduction
The majority of the worldwide demand for electricity and transportation is currently
met through the combustion of fossil fuels such as natural gas, petroleum-based
liquid fuels, coal, and biomass. As a result, combustion remains one of the major
anthropogenic sources of pollutant emissions. Key pollutants generated by combus-
tion of hydrocarbon fuels include nitrogen oxides (NyOx), carbon monoxide (CO),
sulfur oxides (SOx), unburned hydrocarbons (UHC), and particulate matter (PM).
The primary nitrogen oxides generated from combustion systems are nitric oxide
(NO), nitrogen dioxide (NO2), and nitrous oxide (N2O). The sum of NO and NO2
is generally referred to as NOx. Nitrogen oxides are a primary air pollutant linked
to photochemical smog, acid rain, tropospheric ozone, ozone layer depletion, and
global warming (Prather and Sausen, 1999; Skalska et al., 2010). When released in
the atmosphere, NOx can react photochemically with organic compounds to generate
O atoms, which combine with O2 to form ozone (Brasseur et al., 1998). Ground-level
ozone formed in this way is one of the major components, along with particulate
matter, of photochemical smog (Grewe et al., 2002). NOx can also eventually form
N2O5, which reacts with water to form HNO3 (nitric acid), one of the components of
acid rain (Brasseur et al., 1998).
Nitrogen oxides and carbon monoxide are primary pollutant emissions formed
during the combustion of hydrocarbon fuels in gas turbine engines. Emissions of UHC
and PM can also be an issue in gas turbines that operate in non-premixed combus-
tion mode, such as aircraft engines. In addition, the combustion of sulfur-containing
liquid fuels, coal, and biomass can generate sulfur oxides (SOx). SOx are generally
not a consideration for natural gas combustion as this fuel has a negligible amount
of fuel-bound sulfur. Interested readers are encouraged to review the chapter on gas
aerosol precursors for a detailed discussion on SOx emissions. Formation of H2O and
CO2 is a major fraction of the gas turbine exhaust during the combustion of hydro-
carbon fuels, and these substances play a role in global climate change as they act as
greenhouse gases (Prather and Sausen, 1999).
Reduction of pollutant emissions has been a driving force in the design of
efficient gas turbines for many years because of the increasing environmental
175
176 NOx and CO Formation
awareness among the public and the implementation of more rigorous environ-
mental regulations in many countries (Skalska et al., 2010). The Committee on
Aviation Environmental Protection (CAEP) of the International Civil Aviation
Organization (ICAO) regulates aircraft emissions and noise standards (Committee
on Aviation Environmental Protection, 2012). The latest emission standard,
CAEP/8 (2010), for example, introduced new NOx Stringency Options to reduce
NOx emission levels by up to 20 percent relative to the CAEP/6 (2004) standard
for future aircraft engines (CAEP, 2010). Therefore, it is critical to understand the
mechanics of pollutant formation and control during the combustion process to
economically achieve the required levels of pollutant emissions and to improve
energy efficiency and performance in gas turbine combustors. Several reviews
on the chemistry of the formation and destruction of combustion-generated pol-
lutants (Bowman, 1975; Hanson and Salimian, 1984; Miller and Bowman, 1989;
Hayhurst and Lawrence, 1992; Kramlich and Kinak, 1994; Dean and Bozzelli, 1999;
Glarborg et al., 2003) and its relevance to industrial NOx and CO control tech-
nologies (Correa, 1992, 1998; Smoot et al., 1998; Sturgess et al., 2005) have been
published over the years.
This chapter aims to provide a critical review of the formation of CO and NOx
emissions from gas turbine combustors and to discuss recent progress on the under-
lying chemical kinetics that control the production of these pollutants. The effect of
combustion conditions and the role of different formation pathways for the overall
NO production is discussed. A brief review of hydrocarbon fuel oxidation is first pre-
sented to describe the chemistry of CO formation. This discussion also includes the
generation of combustion radicals important for NOx formation. This is followed by
a discussion of various NOx formation pathways and the effect of NOx on hydrocar-
bon oxidation when it is present in a vitiated air stream. A discussion of NOx control
strategies, namely thermal-deNOx and reburning, is then presented. An analysis on
the role of pressure on CO and NOx formation is provided, followed by a description
of NO2 formation in combustion systems.
12
i+c-Paraffins
10 n-Paraffins
Aromatics
8
Volume (%)
0
C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17
Carbon number
Figure 7.1. Chemical class composition distribution for petroleum-derived commercial avia-
tion fuel Jet A-1. (Reproduced from Edwards, 2007; originally published by ASME.)
Fuel
RH
Products + OH
OH
H-atom Propagation
RH abstraction
R + H2O2 HO2 + Alkene R RO2 QOOH
O2
Branching O2
-scission
Alkene + small alkyl O2QOOH
radicals (C1, C2, C3)
OH + OH OH + ROOH
H + Alkene
Branching
Intermediate Branching O2
temperature oxidation RO + OH
OH + O
Low temperature
High temperature oxidation
oxidation
Figure 7.2. Main reaction pathways for chain-branching route during low- and high-temperature
oxidation. (key: RH paraffinic fuel molecule; R alkyl radical; RO2 alkylperoxy radical;
QOOH hydroperoxyalkyl radicals; ROOH alkyl hydroperoxide).
the different temperature regimes during the oxidation of a paraffinic fuel molecule,
which is the major hydrocarbon group in most liquid gas turbine fuels. Although the
temperature range of each pathway depends on the operating pressure, temperatures
above 1200 K can generally be considered the high-temperature oxidation regime
at typical gas turbine conditions. Therefore, most gas turbine combustor operations
fall under the high-temperature regime. In this regime, combustion radical genera-
tion is dominated by a chemical chain-branching reaction between the H atom and
O2 molecule.
In the low-temperature regime, a chain-branching reaction sequence initiated
by reactions between alkyl hydrocarbon radical (R) and O2 molecules is dominant.
Low-temperature oxidation kinetics, which occurs between 600 and 800 K, can play a
significant role in the premixing section where, for example, premature autoignition
of the fuel can occur (Correa, 1998). As the chemical processes of pollutant forma-
tion at typical gas turbine operating conditions are influenced by high-temperature
fuel oxidation pathways, combustion kinetics of hydrocarbons for this regime is pre-
sented in this chapter. The low-temperature kinetic pathways are relatively more
complex than the high-temperature oxidation route, since the formation of alkylper-
oxy (i.e., RO2) intermediates are favored at low temperatures (Walker and Morley,
1997). The low-temperature oxidation of long-chained hydrocarbon fuels has been
extensively studied in the context of spark and diesel engines, and a detailed review
on the low-temperature oxidation phenomenon of various hydrocarbon fuels can be
found elsewhere (Walker and Morley, 1997).
In the high-temperature regime, most of the long-chained hydrocarbon fuels
share very similar reaction pathways. Several comprehensive reviews have been
7.2 Hydrocarbon Oxidation and CO Formation 179
H + O2 <=> OH + O R1
H + O2 + M <=> HO2 + M R2
The onset of the dominance of the branching reaction R1 over the recombination
reaction R2 generally marks the transition from the low-temperature oxidation
region into the high-temperature regime. Detailed discussion of this topic can be
found elsewhere (Miller et al., 2005).
During the oxidation of higher-order hydrocarbons, the fuel molecule (RH) first
undergoes H-atom abstraction via reaction R3 to produce alkyl radicals (symboli-
cally shown as R),
RH + X <=> R + XH R3
CH 2 O + M<=>HCO + H + M R4
CH 2 O + M<=>CO + H 2 + M R5
CH 2 O + H<=>HCO + H 2 R6
CH 2 O + O<=> HCO + OH R7
CH 2 O + OH <=> HCO + H 2 O R8
CH 2 O + HO2 <=> HCO + H 2 O2 R9
180 NOx and CO Formation
CH2O is a precursor for the formation of HCO, which in turn becomes the main
source of CO via reactions R10 and R11. Reaction R11 acts as a branching sequence
as it produces two H atoms in concert with the HCO formed in reaction R4.
200
PB = 0%
PB = 92%
100 PB = 100%
1.0E+03 2500
Temperature
2000
Chemical time scale (sec)
1.0E+00
CO CH2O
H2O2 HO2
Temperature (K)
CH3 H 1500
O OH
1.0E-03
CH HCO
1000
1.0E-06
500
1.0E-09 0
2.0 2.5 3.0 3.5 4.0 4.5 5.0
Flow residence time from burner surface (msec)
Figure 7.4. Chemical time scale variation of selected species within one-dimensional, pre-
mixed laminar flame zone (highlighted region) for an atmospheric pressure CH4/air flame
(equivalence ratio = 0.9) using the GRI3.0 mechanism (Smith et al.).
8 2000
Temperature
Total NO
6 1500
Exp. Data-NO
Temperature (K)
NO (ppmv)
4 1000
N2O Route
Prompt-NO
2 500
0 0
0.0 0.2 0.4 0.6 0.8 1.0
Distance (cm)
7.3.1 Thermal NO
The relative importance of an individual chemical route to the overall NOx emis-
sions largely depends on the combustor operating conditions. When no fuel-bound
nitrogen is present, the thermal-NO route is one of the major sources of NOx in
7.3 Formation of Nitrogen Oxides 185
100 15
Prompt
N2O Route
Percent contribution to total NO
80 NNH Route 12
Thermal
Total NO (ppmv)
Total NO
60 9
40 6
20 3
0 0
0.7 0.8 0.9 1.0
Equivalence ratio
Figure 7.6. Computational results for NO and percent contribution of different chemical
pathways at a constant residence time of 10 msec as a function of equivalence ratio using the
conditions and the flame temperature profile described in Figure 7.5.
many practical combustion devices with flame temperatures greater than 1800 K. In
this pathway, NO is produced as molecular nitrogen reacts with oxygen atoms (via
reaction R16) at high temperatures. This initiates a chemical kinetic process followed
by reaction R17 where N atoms react with molecular oxygen. This set of reactions is
also known as the Zeldovich mechanism (Zeldovich, 1946). Also, the nitrogen atom
generated in reaction R16 reacts with an OH radical to form NO via reaction R18,
known as the extended Zeldovich mechanism. Reaction R16 is the rate-limiting step
for the formation of thermal NO because of its high activation energy.
N 2 + O<=> NO + N R16
N + O <=> NO + O R17
2
N + OH <=> NO + H R18
Since the concentration of atomic oxygen in the flame front is largely an exponential
function of temperature, NO formation via the Zeldovich mechanism has a similar
relationship with flame temperature. In addition, reaction R16 propagates the chain
reaction by producing a nitrogen atom, which then reacts with molecular oxygen to
produce NO and O atoms via reaction R17. As shown in Figure 7.3, as the flame tem-
perature increases, the NOx increases exponentially due to post-flame thermal-NO
production for both the premixed and diffusion systems. The rate of production of
NO via the Zeldovich mechanism can be estimated through the equilibrium concen-
tration of oxygen in the post-flame zone using Equation 7.2 (Bowman, 1975):
d[NO] 0.5
dt = 6 10 Teq exp
16
( 69,090
Teq )[O ]0.5
2 eq
moles
[N 2 ]eq 3
cm sec
(7.2)
However, since super-equilibrium concentrations of radical species such as O and
OH are typically present in the flame front, more detailed knowledge of the com-
bustion chemistry of given fuel is necessary to accurately predict the thermal NO
186 NOx and CO Formation
generated in the flame zone. Drake and colleagues (1990) experimentally investi-
gated the influence of super-equilibrium radical concentrations on flame-front NO
formation in premixed ethane flames at varying pressures. As shown in Figures 7.5
and 7.6, simulation results of these premixed experiments indicate that less than 10
percent of the total NO was produced via the thermal-NO route in the relatively
low-temperature laminar flame. If the flame temperature profile used in Figure 7.6
was altered to rise with equivalence ratio, the contribution of thermal NO would
increase dramatically with temperature. Flame temperatures higher than 2000 K will
produce a significant amount of thermal NO, which is the case in most non-premixed
combustion systems. Though a minor contributor in the example provided here, the
thermal-NO pathway must be taken into consideration for most practical lean, pre-
mixed combustion systems operating at high inlet pressures and temperatures.
Molecular nitrogen reacts with an O atom to form NO via reaction R16 in the
thermal-NO pathway. An alternative route to form N2O is a recombination reaction
R19 favored at low temperatures. N2O can further proceed to react with an O atom
to produce NO via reaction R20. Additionally, N2O can also react with an H atom to
produce NO while forming NH via reaction R21.
N 2 + O + M <=> N 2 O + M R19
N 2 O + O<=> NO + NO R20
N 2 O + H <=> NO + NH R21
However, N2O can also react with O and H atoms through alternative channels,
as shown in reactions R22 and R23. Therefore, the proper branching ratio of these
multiple channels must be taken into account to correctly model NO formation via
the N2O pathway.
N 2 O + O<=> N 2 + O2 R22
N 2 O + H <=> N 2 + OH R23
The relatively slow kinetic rates of these reactions within the N2O pathway reduce its
significance at most conditions, except under fuel-lean, low-temperature situations.
However, N2O formation via the recombination reaction R19 is increased at higher
pressures, while the destruction of N2O to form NO (reaction R20) is enhanced by
super-equilibrium O atoms in the flame-front region. This is evidenced in Figure 7.5,
where approximately 50 percent of the total NO was produced via the N2O route
for a low-temperature, premixed ethane flame at 6 atm. For practical gas turbines
that operate under lean-premixed conditions at higher pressures, one of the major
chemical pathways of NO formation is via the N2O route.
7.3.3 Prompt NO
Unlike the thermal NO and N2O reaction sets driven by the interaction between
molecular N2 and an O atom, prompt-NO formation is an attribute of hydrocarbon
7.3 Formation of Nitrogen Oxides 187
flames where smaller hydrocarbon radicals such as CH are available to react with
molecular N2. Fenimore (1971) initially proposed the prompt-NO formation pathway
to explain the nitric oxide found in the thin reaction zone close to the burner surface
in experimental data obtained from CH4, C2H4, and C3H8 flames. The thermal-NO
route does not explain this observation because of the lack of atomic oxygen or
nitrogen at this relatively cold location. Fenimore (1971) suggested that the likely
path must involve the reactions of hydrocarbon radicals formed in the flame zone
with molecular N2 to produce amines and cyano compounds that can further react
to form NO. Hayhurst and Vince (1980) proposed that the primary route for the
prompt-NO formation involves the reaction between N2 and hydrocarbon radicals,
predominantly CH, to form HCN and N via reaction R24:
Bozzelli and Dean (1995) have proposed that NNH, an intermediate species formed
by the reaction between N2 and an H atom (reaction R30), can react with an O
atom to produce NO (reaction R31) under certain conditions. This route is partic-
ularly viable at low flame temperatures. Furthermore, super-equilibrium O-atom
188 NOx and CO Formation
concentration at the flame front can also increase the rate of NO formation via the
NNH route.
N 2 + H <=> NNH R30
NNH + O<=> NO + NH R31
This effect can be noted in the burner stabilized ethane premixed flame example
described in Figure 7.5 where the contribution of the NNH pathway to the total
NO formation is small, but not insignificant. It can also be noted in Figure 7.6 that
NO formation via the NNH route increases with the equivalence ratio. However,
the importance of the NNH route for NO formation is diminished at higher flame
temperatures by the dominance of other reaction pathways that consume NNH mol-
ecules (reactions R32 and R33) and prevent reaction R31 from contributing to NO
formation (Klippenstein et al., 2011). In a recent publication, Klippenstein and col-
leagues (2011) discussed in detail the role of the NNH species in NO formation as
well as NO destruction in the thermal-deNOx process.
A ready source of NOx formation from combustion devices occurs with liquid and
solid fuels that contain chemically bound nitrogen (e.g., fuel oils, coal, and biomass).
Typical nitrogen concentrations in distillate fuels can range from 0 to over 0.65 wt per-
cent (Bowman, 1975). Table 7.1 lists the fuel-bound nitrogen content of the solid fuels
used in either incineration or power generation applications (Dagaut et al., 2008).
The reaction pathways involved in NOx formation from chemically bound nitro-
gen are rather complex because of the varying structure of the nitrogen bonding
to the parent molecule. The fuel-bound nitrogen in coal and biomass adds further
complexity because of the heterogeneous oxidation kinetics between volatiles and
char. During the oxidation of liquid and solid fuels, most of the fuel-bound nitro-
gen is converted to HCN and/or NH3 intermediates, which then react with combus-
tion radicals to form NOx (Glarborg et al., 2003). Dagaut and colleagues (2008) and
Skreiberg and colleagues (2004) have provided a detailed discussion on the role
of HCN and NH3 chemistry, respectively, in the production of nitrogen oxides. The
formation of NCO from HCN and NH from NH3 acts as the main precursor for the
formation of NO from fuel-bound nitrogen. Reactions R34 to R38 show the main
reaction pathways for the formation of NO from HCN, while reactions R39 to R41
show the major NH3 oxidation route for the formation of NO.
Biomass 0.13.5
Peat 0.52.7
Coal 0.52.5
Household waste 0.51.0
Sewage sludge 2.56.5
NH + O<=> NO + H R40
N + O2 <=> NO + O R41
Several secondary reaction pathways allow the formation of NO from HCN and
NH3 through various nitrogenous intermediate species such as HNCO and HNO.
A detailed chemical kinetic mechanism for HCN and NH3 chemistry can be found
elsewhere (Skreiberg et al., 2004; Dagaut et al., 2008).
Figure 7.7 summarizes the major reaction pathways for the formation of NOx
discussed previously and their relationship to overall NOx production. Although
the chemistry of NH3 and HCN oxidation plays a major role in the formation of
NO from fuel-bound nitrogen, they also share common reaction pathways with the
prompt-NO and NNH pathways. In addition, NH3 and HCN chemistry are responsi-
ble for the destruction of NO in the thermal-deNOx and reburning processes, respec-
tively. The chemical kinetic pathways of HCN and NH3 for the destruction of NO in
the context of reburning and thermal-deNOx processes, respectively, are presented
in the NOx abatement section.
Thermal NO Prompt NO
N2
N2
th
O
pa
pa
+H +O+M
th
NN
+O NNH N2O +CH
+O +O +H
+OH +OH
N NO NCN
+O2 +O
+H +OH +OH +O
NH +H NCO +O2 CN
(HNCO)
Figure 7.7. Major NOx formation reaction pathways and their interconnections.
(Fuller et al., 2012) properties of hydrocarbon fuels showed that NO has a signifi-
cant chemical kinetic influence in promoting ignition of the fuel, while CO2 has the
kinetic effect of reducing the laminar flame speed. The presence of CO2 in the oxi-
dizer stream diminishes the flame speed by reducing the concentration of H atoms
via reaction R12 (Fuller et al., 2012). The reduction in the radical concentration due
to the presence of CO2 in the oxidizer stream will indirectly contribute to the reduc-
tion in NO production, as shown by Fackler and colleagues (2011).
The NOx formation pathways discussed previously are largely controlled by
the nitrogen and hydrocarbon chemistry in the flame front and/or the post-flame
zone. However, researchers observed (Bromly et al., 1992; Bendtsen et al., 2000)
that the presence of NO in the oxidizer stream tends to promote the oxidation of
hydrocarbon fuels at low temperatures, especially at fuel-lean conditions. Several
experimental and computational studies (Bromly et al., 1992; Amano and Dryer,
1998; Bendtsen et al., 2000; Faravelli et al., 2003; Gokulakrishnan et al., 2005; Moreac
et al., 2006; Fuller et al., 2011) have shown that a small amount of either NO or NO2
promotes the oxidation of hydrocarbon fuels at low temperatures by accelerating
the formation of the radical pool. At low and intermediate temperatures (i.e., 600
to 1200 K), HO2 radical formation is favored via the recombination reaction R2, as
opposed to the chain-branching reaction R1. However, in the presence of vitiated
air, the relatively unreactive HO2 radical combines with NO to produce NO2 and
OH radical via reaction R42. NO2 then reacts with the CH3 radical (which itself is
relatively unreactive at low temperatures) to generate CH3O radicals via reaction
R43, converting the NO2 back to NO. Similarly, NO also reacts with CH3O2 to pro-
duce CH3O via reaction R44.
NO + HO2 <=> NO2 + OH R42
7.5 NOx Abatement Strategies 191
Although these reaction pathways have been mainly validated for methane oxida-
tion (Bromly et al., 1992; Bendtsen et al., 2000), they also play a significant role
during the oxidation of higher-order hydrocarbons in the presence of NOx. Radical
species such as HO2 and CH3 form the basic building blocks of hydrocarbon oxida-
tion. Reactions R42 to R44 are fuel-independent pathways to promote oxidation
by NOx irrespective of the type of hydrocarbon molecule. In addition, NOx interac-
tions with alkyl (R) and alkylperoxy (RO2) radicals (see Figure 7.2) generated from
the oxidation of paraffinic fuel molecules (see Figure 7.2), which are found in most
gas turbine liquid fuels, can also play a significant role in the promotion of hydro-
carbon oxidation by NOx (Chan et al., 2001). H-atom abstraction from paraffinic
molecules by NO2 can also be an important initiation reaction for the oxidation of
long-chained hydrocarbon fuels at low temperatures (Fuller et al., 2011). Further
research work is needed to fully understand the interaction between long-chained
hydrocarbon radicals and NOx, since most of the experimental work reported in the
literature (e.g., Amano and Dryer, 1998; Bendtsen et al., 2000) is related to natural
gas surrogate fuels.
Fuller and colleagues (2009) have investigated the chemical kinetic effect of
vitiated air composition (NO, CO, CO2, H2O, and O2 in N2) on the autoignition of jet
fuel in an atmospheric pressure flow reactor. They found NO has the largest effect
among the vitiated air components on altering the ignition delay time of the fuel.
Figure 7.8 shows the change in ignition delay time of JP-8 (U.S. military jet fuel),
with an inlet temperature of 900 K, as a function of initial NO concentration in the
inlet oxidizer stream (Fuller et al., 2011). It can be noted that a small amount of NO,
on the order of 500 pm, can reduce the ignition delay time by more than 50 percent.
A similar trend was observed in the high-pressure, jet-stirred reactor experiments
of Moreac and colleagues (2006), in which the addition of NO promotes the oxida-
tion of n-heptane, iso-octane, and toluene at initial temperatures of 700 to 1000 K.
However, at temperatures below 650 K, the addition of NO was found to inhibit
oxidation of n-heptane by scavenging the radical pool primary through the termina-
tion reaction R45 (Moreac et al., 2006). The OH depletion caused by reaction R45 is
offset by reaction R42 as the temperature is increased.
NO + OH + M <=> HONO + M
R45
1.0
20 mole% O2
Relative reduction in ignition delay time
12 mole% O2
0.8
0.6
0.4
0.2
0.0
0 200 400 600 800 1000
[NO] (ppmv)
Figure 7.8. Reduction in ignition delay time as a function of NO at varying O2 levels for stoi-
chiometric JP-8 mixtures at an initial temperature of 900 K (Fuller et al., 2011). Ignition delay
times have been normalized by the values measured in the absence of NO in the oxidizer
stream.
by the thermal-NOx (through lower flame temperatures) and the prompt-NOx (by
operating at fuel-lean conditions) pathways. However, when burning liquid fuels,
it is difficult to achieve fuel/air premixing without premature autoignition as these
fuels have much shorter ignition delay times than natural gas at typical gas turbine
inlet conditions. Alternative techniques such as lean direct injection (LDI) (Sturgess
etal., 2005) and lean prevaporized, premixed (LPP) combustion (Gokulakrishnan
etal., 2008) are being developed to extend the lean, premixed combustion capability
to liquid fuels, but these systems are not yet widely used. Hence, gas turbines that
operate on liquid fuels can produce significant amounts of NOx as most systems
operate in diffusion (or non-premixed) mode, which leads to higher flame temper-
atures (>2000 K). Therefore, for combustors that operate in non-premixed mode,
employing alternate NOx control strategies is critical to meeting the increasingly
stringent NOx emission levels required by environmental regulations.
In post-combustion control techniques, such as selective catalytic reduction
(SCR) and selective non-catalytic reduction (SNCR), the chemical reduction of NOx
is effective only within a very narrow set of operating conditions. In SNCR, also
known as the thermal-deNOx process, additives such as ammonia or urea are used
in the post-combustion zone to promote the conversion of NO into N2. This process
must take place within an appropriate temperature range to be effective. The main
advantage of a post-combustion control technique is that the combustor does not
require any significant design modifications. However, additional capital and oper-
ating costs are incurred to maintain the post-combustion NOx control zones within
specified bounds. This approach is most suitable for stationary gas turbines from
logistical and operational points of view.
7.5 NOx Abatement Strategies 193
For in situ combustion control techniques, such as air staging and fuel reburning,
NOx reduction is achieved by selectively changing the local fuel-air ratio during the
combustion process. Although in situ combustion control techniques are suitable for
aircraft gas turbines, this approach must be incorporated into the combustor design.
For example, the concept of a rich-quench-lean (RQL) system is currently employed
in various forms in aircraft combustor designs (Correa, 1998; Sturgess, 2005). In this
system, air staging is carried out in such a way that the fuel-rich primary combustion
zone is followed by a secondary burnout zone with excess air to reduce the NOx
formation via lower flame temperatures. However, fuel/air mixing, primary zone
equivalence ratio, and secondary zone residence time play crucial roles in achieving
desired NOx levels in RQL systems.
NH 3 + OH <=> NH 2 + H 2 O R46
NH 3 + O<=> NH 2 + OH R47
NH + H <=> NH + H R48
3 2 2
NH 2 + NO<=> N 2 + H 2 O R49
However, for this reaction scheme to sustain itself in the post-combustion zone, reac-
tive species such as OH, H, and O are needed for the conversion of NH3 into NH2.
NO consumption is achieved by the reaction of NO with NH2 via an alternate route
to produce NNH (R50), which then decomposes via reaction R51 to form N2 and H
(note: this is the reverse reaction of R30).
At 1100 K, the branching fraction must be at least 0.25 so that NO is reduced (via
reaction R49) while sustaining the branching reaction pathways that generate the
radical species pool necessary to maintain the NH3 initiation (via reactions R50 and
R51) (Miller et al., 2005). This balancing act has a limitation in terms of temperature,
and it was experimentally found (Kjaegaard et al., 1996) that the thermal-deNOx
process with NH3 is effective between 1100 and 1400 K. At temperatures above 1400
K, the subsequent increase in the reactive radical pool favors the formation of addi-
tional NO from NH3. However, this window can be shifted toward higher tempera-
tures when the system pressure is increased (Kjaegaard et al., 1996), mainly because
of the competition between branching reaction (R1) and the recombination reaction
(R2). It was also found that a higher O2 concentration enhances the deNOx process
efficiency.
7.5.2 Reburning
As discussed previously, the presence of NOx in the oxidizer stream promotes the
breakdown of hydrocarbon fuels at relatively low temperatures (600 K < T < 1200
K) by increasing the concentration of active combustion radicals at both lean and
rich conditions. However, at high temperatures (above 1200 K), the chain-branching
reaction R1 dominates the oxidation process by greatly increasing the combustion
radical pool, and hence diminishing the role of NO in promoting the oxidation of
hydrocarbon fuels. The interaction between NO and hydrocarbon radicals at high
temperatures was found to favor the destruction of NO, especially under fuel-rich
conditions (Myerson, 1974). This phenomenon is known as reburning. Reburning,
used in some form in many practical gas turbine systems, is an in situ NOx control
technique that can be described as a multi-zone process. Fuel is added to the exhaust
of the primary combustor to create a fuel-rich zone (or reburning zone), while air is
added after the reburning zone to complete the fuel oxidation (known as the burn-
out zone). In the reburning zone, the NOx produced in the primary combustion zone
can be converted to free N2 or to other nitrogenous intermediate species (such as
HCN or NH3) by the gas phase chemical kinetic interaction between NO and hydro-
carbon radicals generated from the added fuel. Smoot and colleagues (1998) provide
a comprehensive review on this subject.
7.6 Effect of Pressure on CO and NOx Formation 195
12 1750
Temperature (K)
NO at 14.6 atm
8 1500
NO at 6.1 atm
4 NO at 1 atm 1250
0 1000
0.0 0.2 0.4 0.6 0.8
Distance from burner surface (cm)
Figure 7.9. Model comparison of Gokulakrishnan and colleagues (2012) with the experimen-
tal data of Thomsen and colleagues (1999) for the axial profiles of NO concentration and
temperature in premixed CH4/O2/N2 flames (0.6 equivalence ratio) at pressures of 1.0 and 14.6
atm. Key: symbols experimental data (closed symbols NO; open symbols temperature);
lines model predictions (solid lines NO; broken lines temperature: dotted line 1 atm,
dot-dashed line 6.1 atm, dashed line 14.6 atm).
O-atom concentrations and the role of the N2O and prompt-NO pathways influence
the effect of pressure on NO formation.
Experimental and modeling studies of natural gas and methane combustion
(Correa, 1992; Leonard and Stegmaier, 1994) conducted at lean, premixed condi-
tions (with equivalence ratios less than 0.7) found that NOx emissions have very little
dependence on pressure. Bhargava and colleagues (2000) performed an experimen-
tal study with different fuel injection schemes to investigate the effect of pressure
on NOx between 7 and 27 atm. The experimental results showed that the pressure
power factor (i.e., n in Pn) for NOx emissions increased when the equivalence ratio is
changed from 0.43 to 0.65. The pressure power factor varied between 0.77 and 1.6,
depending on the type of nozzle used in these experiments. This indicates that the
effectiveness of the fuel/air mixing plays a significant role in determining the influ-
ence of pressure on NOx formation.
Figure 7.9 shows the effect of pressure on an NO profile obtained in burner stabi-
lized premixed flame experiments of Klassen and colleagues (1995), which were fur-
ther refined by Thomsen and colleagues (1999). The NO measurements shown were
obtained in a CH4/O2/N2 flame at pressures of 1 and 14.6 atm at an equivalence ratio
of 0.6 with a dilution ratio of 2.2. The figure also shows the model predictions obtained
from a burner-stabilized one-dimensional laminar flame simulation (solving for
flame temperature with the energy equation) in Cantera using the chemical kinetic
mechanism of Gokulakrishnan and colleagues (2012). The model predictions for the
temperature profile and NO agree fairly well with the experimental data. Figure 7.9
198 NOx and CO Formation
100
80
Thermal-NO
60
40
N2O Route
20
NNH Route
0
1 3 6 9 12 15
Pressure (atm)
Figure 7.10. Contribution of different NO pathways to the total NO as a function of pressure
at an equivalence ratio of 0.6 calculated at 0.3 cm from the burner surface for the experimen-
tal conditions described in Figure 7.9.
also shows the model prediction at a pressure of 6.1 atm for comparison. It can be
noted that the flame-front temperature increases as the pressure was raised from 1
to 14.6 atm. This results in increased flame-front NO formation at 14.6 atm because
of an increase in the super-equilibrium O-atom concentration. A steeper rise in NO
production is observed in the post-flame zone at 14.6 atm compared to the 1 atm
case as shown in Figure 7.9. This is due to greater thermal-NO formation enhanced
by the larger equilibrium O-atom concentration and longer residence time in the
post-flame zone at 14.6 atm.
Figure 7.10 shows the contribution of different NO pathways to the total NO
at 0.3 cm from the burner surface as a function of pressure at an equivalence ratio
of 0.6 for the experimental conditions described in Figure 7.9. The one-dimensional
laminar flame simulations were performed using the chemical kinetic mechanism
of Gokulakrishnan and colleagues (2012). As shown in Figure 7.10, the N2O route
is the largest contributor to the total NO. In general, the NO production from the
N2O route will increase with pressure at low temperatures. A decrease in N2O route
contribution to the total NO was observed above 6 atm in Figure 7.10 because of
increasing flame temperature with pressure. This results in an increase in thermal
NO with pressure as shown in Figure 7.10. The contribution of the prompt-NO path-
way is around 30 percent, while the NNH route has a negligible impact on the total
NO for the experimental conditions described in Figure 7.9.
Figure 7.11 shows the NO measurements reported by Thomsen and colleagues
(1999) at 0.3 cm from the burner surface as a function of pressure, along with the
model predictions at equivalence ratios of 0.6, 0.7, and 0.8. The one-dimensional lam-
inar flame simulation results shown in Figure 7.11 were performed using the chem-
ical kinetic mechanism of Gokulakrishnan and colleagues (2012). The figure also
7.6 Effect of Pressure on CO and NOx Formation 199
20
= 0.8
(P0.42)
NO (ppmv, wet at 15% O2)
15
= 0.7
(P0.35)
10
= 0.6
(P0.39)
5
0
0 5 10 15 20
Pressure (atm)
Figure 7.11. NO formation as a function of pressure at equivalence ratios of 0.6, 0.7, and 0.8
obtained at 0.3 cm from the burner surface for the experimental conditions described in
Figure 7.9. The experimental data (symbols) of Thomsen and colleagues (1999) compared with
the model predictions. The experimental error is indicated for the measurement at 14.6 atm
and 0.6 equivalence ratio.
20
Fuel oil #2: = 0.70 P0.17
16
NOx (ppmv, dry at 15% O2)
4
Methane: = 0.70 P0.44
0
1 4 7 10
Pressure (atm)
Figure 7.12. NOx emission as a function of pressure for methane and fuel oil #2 (Gokulakrishnan
et al., 2008; originally published by ASME).
oil are 0.44 and 0.17, respectively, at 0.7 equivalence ratio as shown in Figure 7.12. This
may be due to the dominance of fuel-bound nitrogen as the major source of NO in
fuel oil. Moreover, a number of factors, including the heat loss to the burner surface
and level of premixing in the swirl burner, make the determination of the role that
pressure plays difficult to ascertain for the measurements presented in Figure 7.12.
In practical combustion systems, the ability to deliver perfectly premixed air/
fuel mixtures can be difficult to obtain. In non-perfectly premixed systems, pockets
of relatively rich mixtures will contribute significantly to NOx production because of
the exponential temperature dependence of the thermal-NO route. Mongia and col-
leagues (1996) measured and modeled NOx production in a CH4/air flame as a func-
tion of fuel/air unmixedness, equivalence ratio, and pressure. Their results showed
that for well-mixed, lean flames, there is little pressure influence on NOx production.
However, as the unmixedness level increased, pressure dependence was exhibited.
This finding reflects back to the role of local equivalence ratio (and flame tempera-
ture) on the relative importance of different NO formation pathways on total NO
production.
However, NO2 can be formed during the mixing of hot exhaust gases with dilution
air in regions downstream of the combustor chamber via reaction R42 (Sano, 1984):
NO + HO2 <=> NO2 + OH R42
The levels of NO2 exiting the combustor is a small fraction (~5 percent) of the total
NOx in the exhaust stream (Feitelberg and Correa, 1999). Significant conversion of
NO to NO2 can occur under the proper conditions after the product gases have left
the combustion chamber. The conversion process predominately occurs at relatively
low temperatures (8001000 K), and is aided by reactions with unburned hydro-
carbons and CO. Temperatures in the range of 800 to 1000 K are present at some
locations after the exhaust gases have left the combustion chamber (e.g., in the heat
recovery steam generator (HRSG) of a combined cycle gas turbine plant) for the
period of time necessary for the conversion of NO to NO2. Hori and colleagues
(1992, 1998) investigated the conversion of NO to NO2 in flow-reactor experiments
at low temperatures. As discussed previously, NO can react with the HO2 radical (via
reaction R42) to produce the OH radical while converting NO into NO2. In addition,
trace amounts of hydrocarbons in the exhaust can also convert NO into NO2 via
reaction R59. Unburned hydrocarbons in the low-temperature exhaust can lead to
the presence of alkylperoxy (RO2) radicals (Faravelli et al., 2003), which react with
NO to form NO2.
NO + RO2 <=> NO2 + RO R59
In general, larger alkanes more effectively oxidize NO to NO2 over a wide tem-
perature range, while alkenes and methane are less efficient at this conversion. As is
seen in Figure 7.13, the composition of the exhaust stream is critical in the conver-
sion of NO to NO2 via this route. The presence of higher hydrocarbons, which can
be found in unprocessed natural gas and some LNG imports, increases the tendency
for NO2 conversion by up to several orders of magnitude as compared to methane.
Hence, when unburned hydrocarbons bypass the combustion zone, the presence of
any higher hydrocarbons in the exhaust will increase the propensity for the conver-
sion of NO to NO2. This will not change the total NOx production, but may lead to
other environmental issues such as a visible plume from the exhaust stack.
7.8 Summary
Reduction of pollutant emissions from combustion has been a driving force in the
design of gas turbines for decades. A thorough understanding of the chemistry of
combustion is required to produce cleaner, more efficient engines. This chapter has
focused on the complex interactions of hydrocarbon oxidation and the formation of
202 NOx and CO Formation
1.0
n-C4H10
C3H6
NO2 / NOx
CH4 CO
H2
0.5
C2H6
i-C4H10
C2H4
C3H8
0
1 10 102 103 104
Added fuel in air (ppm)
Figure 7.13. Proportion of NO2 to NOx as a function of amount of fuel added and the fuel type.
(Reproduced from Hori and colleagues 1992 with permission from Elsevier.)
pollutants such as NOx and CO. The current status of the combustion chemistry for
the production and control of NOx and CO emissions at gas turbine relevant con-
ditions has been reviewed. The combustion radicals and fuel fragments generated
during the consumption of the hydrocarbon fuel molecule play a critical role in the
generation of CO and NOx. A fraction of the nitrogen, introduced with the combus-
tion air or bound to the fuel itself, reacts with the radical species to form nitrogen
oxides. The amount of CO in the exhaust stream is largely determined by the bal-
ance between the fast reaction kinetics for CO formation and the relatively slow CO
consumption route to form CO2. Therefore, combustor design must allow an optimal
residence time for CO oxidation while minimizing the formation of NOx during the
CO burnout process.
The combustion mode (i.e., premixed or diffusion) plays a critical role in the
amount of pollutant formed. This is especially true for NOx and the chemical path-
ways critical for its formation. The NOx emissions from diffusion or non-premixed
combustion systems are largely dominated by thermal-NOx because of higher flame
temperatures. However, N2O and prompt-NO routes contribute significantly at the
lower flame temperatures relevant to lean, premixed combustion systems such as
DLE systems. The contribution of the N2O-route to the total NOx concentrations
increases with pressure at typical DLE conditions.
Areas of needed research have also been identified. In particular, additional
work is needed to better understand the chemical kinetic interactions between NOx
and long-chain hydrocarbon fuels at low temperatures (i.e., fuel oxidation chemistry
in the presence of NO) and at high temperatures (i.e., reburning chemistry where
NO is reduced by fuel addition). There is a consensus understanding of the effect
of pressure on NOx formation under fuel-lean, premixed combustion conditions.
However, additional work is needed to better understand the role of pressure on
NOx formation in non-premixed combustion systems.
References
Allen, M. T., Yetter, R. A., and Dryer, F. L. (1997). High Pressure Studies of Moist Carbon
Monoxide-Nitrous Oxide Kinetics. Combustion and Flame 109: 44970.
References 203
Amano, T., and Dryer F. L. (1998). Effect of Dimethyl Either, NOx, and Ethane on CH4
Oxidation: High Pressure, Intermediate-temperature Experiments and Modeling.
Proceedings of the Combustion Institute 27: 397404.
Bendtsen, A. B., Glarborg, P., and Dam - Johansen, K. (2000). Low Temperature Oxidation of
Methane: The Influence of Nitrogen Oxides. Combustion Science and Technology 151:
3171.
Bhargava, A. et al. (2000). Pressure Effect on NOx and CO Emissions in Industrial Gas
Turbines, in Proceedings of TurboExpo, ASME Paper No. 2000-GT-97.
Bowman, C. T. (1975). Kinetics of Pollutant Formation and Destruction in Combustion.
Progress in Energy Combustion and Science 1: 3345.
Bozzelli, J. W., and Dean, A. M. (1995). O + NNH: A Possible New Route for NOx Formation
in Flames. International Journal of Chemical Kinetics 27: 1097109.
Brasseur, G. P. et al. (1998). European Scientific Assessment of the Atmospheric Effects of
Aircraft Emissions. Atmospheric Environment 32: 2329418.
Bromly, J. H., Barnes, F. J., Mandyczewsky, R., Edwards, T. J., and Haynes, B. S. (1992). An
Experimental Investigation of the Mutually Sensitised Oxidation of Nitric Oxide and
n-Butane. Proceedings of the Combustion Institute 24: 899907.
Chan, W., Heck, S. M., and Pritchard, H. O. (2001). Reaction of Nitrogen Dioxide with
Hydrocarbons and its Influence on Spontaneous Ignition. Physical Chemistry and
Chemical Physical 3: 5662.
Chaos, M., Burke, M., Ju, Y., and Dryer, F. L. (2010). Syngas Chemical Kinetics and Reaction
Mechanisms, chapter 2 in Synthesis Gas Combustion: Fundamentals and Applications,
Lieuwen, T., Yang, V., and Yetter, R. eds., Taylor & Francis, Inc., pp. 2970.
Colket, M. B. et al. (2007). Development of an Experimental Database and Kinetic Models
for Surrogate Jet Fuels, in 45th AIAA Aerospace Sciences Meeting and Exhibit, Reno,
NV, pp. AIAA 2008972.
Colket, M. B., Guile, C. R., Zabielski, M. F., and Seery, D. J. (1982). Internal Aerodynamics of
Gas Sampling Probes. Combustion and Flame 44: 314.
Committee on Aviation Environmental Protection (CAEP). (2010). Report of the Committee
on Aviation Environmental Protection, Eighth Meeting. Montreal, Canada, Doc 9938,
CAEP/8.
(2012). [Online]. April. http://www.icao.int/environmental-protection/Pages/CAEP.aspx.
Connors, C. S., Barnes, J. C., and Mellor, A. M. (1995). Semi-empirical Predictions and
Correlations of CO Emissions from Utility Combustion Turbines. Journal of Propulsion
and Power 12(5): 92532.
Correa, S. M. (1992). A Review of NOx Formation under Gas-Turbine Combustion
Conditions. Combustion Science and Technology 87: 32962.
(1998). Power Generation and Aeropropulsion Gas Turbines: From Combustion
Science to Combustion Technology. Proceedings of the Combustion Institute 27:
1793807.
Dagaut, P. (2002). On the Kinetics of Hydrocarbons Oxidation from Natural Gas to Kerosene
and Diesel Fuel. Physical Chemistry and Chemical Physics, 4: 207994.
Dagaut, P., Glarborg, P., and Alzueta, M. U. (2008). The Oxidation of Hydrogen Cyanide and
Related Chemistry. Progress in Energy and Combustion Science 34: 146.
Dagaut, P., Lecomte, F., Chevailler, S., and Cathonnet, M. (1998). Experimental and Detailed
Kinetic Modeling of Nitric Oxide Reduction by a Natural Gas Blend in Simulated
Reburning Conditions. Combustion Science and Technology 34: 32963.
Dagaut, P., Luche, J., and Cathonnet, M. (2000). The Kinetics of C1 to C4 Hydrocarbons/
NO Interactions in Relation with Reburning. Proceedings of the Combustion Institute
28: 245965.
Dean, A. M., and Bozzelli, J. W. (1999). Combustion Chemistry of Nitrogen, chapter 2 in
Gas-Phase Combustion Chemistry, Gardiner, W. ed., Springer, New York, pp. 125343.
Dooley, S. et al. (2010). A Jet Fuel Surrogate Formulated by Real Fuel Properties. Combustion
and Flame 157: 23339.
204 NOx and CO Formation
Drake, M. C., Ratcliffe, J. W., Blint, R. J., Cater, C. D., and Laurendeau, N. M. (1990).
Measurements and Modeling of Flamefront NO Formation and Superequilibrium
Radical Concentrations in Laminar High-Pressure Premixed Flames. Proceedings of the
Combustion Institute 23: 38795.
Edwards, T. (2007). Advancements in Gas Turbine Fuels From 1943 to 2005. Journal of
Engineering for Gas Turbines and Power 129: 1320.
Edwards, T., and Maurice, L. Q. (2001). Surrogate Mixtures to Represent Complex Aviation
and Rocket Fuels. Journal of Propulsion and Power 17(2): 4616.
Fackler, B. K. et al. (2011). Experimental and Numerical Study of NOx Formation from the
Lean Premixed Combustion of CH4 Mixed with CO2 and N2, in Proceedings of ASME
Turbo Expo, pp. GT201145090.
Faravelli, T., Frassoldati, A., and Ranzi, E. (2003). Kinetic Modeling of the Interactions
between NO and Hydrocarbons in the Oxidation of Hydrocarbons at Low Temperatures.
Combustion and Flame 132: 188207.
Feitelberg, A. S., and Correa, S. M. (1999). The Role of Carbon Monoxide in NO2 Plume
Formation, in Proceedings of ASME TurboExpo, pp. 99-GT-53.
Fenimore, C. P. (1971). Formation of Nitric Oxide in Premixed Hydrocarbon Flames.
Proceedings of the Combustion Institute 13: 37380.
Fleck, B. A., Sobiesiak, A., and Becker, H. A. (2000). Experimental and Numerical
Investigation of the Novel Low NOx CGRI Burner. Combustion Science and Technology
161: 89112.
Frassoldati, A., Faravelli, T., and Ranzi, E. (2003). Kinetic Modeling of the Interactions
between NO and Hydrocarbons at High Temperature. Combustion and Flame 135:
97122.
Fuller, C. C. et al. (2012). Effects of Vitiation and Pressure on Laminar Flame Speeds
of n-Decane, in 50th AIAA Aerospace Sciences Meeting, Nashville, TN pp. AIAA
20120167.
Fuller, C. C., Gokulakrishnan, P., Klassen, M. S., Roby, R. J., and Kiel, B. V. (2009). Investigation
of the Effects of Vitiated Conditions on the Autoignition of JP-8. 45th AIAA/ASME/
SAE/ASEE Joint Propulsion Conference & Exhibit, vol. AIAA 20094295, August.
(2011). Investigation of the Effect of Nitric Oxide on the Autoignition of JP-8 at Low
Pressure Vitiated Conditions. 49th AIAA Aerospace Sciences Meeting, Orlando, FL, pp.
AIAA 201196.
Glarborg, P., Alzueta, M. U., Dam - Johansen, K., and Miller, J. A. (1998). Kinetic Modeling
of Hydrocarbon/Nitric Oxide Interactions in a Flow Reactor. Combustion and Flame
115: 127.
Glarborg, P., Jensen, A. D., and Johnsson, J. E. (2003). Fuel Nitrogen Conversion in Solid Fuel
Fired Systems. Progress in Energy and Combustion Science 29: 89113.
Gokulakrishnan, P. et al. (2008). A Novel Low NOx Lean, Premixed, and Prevaporized
Combustion System for Liquid Fuels. Journal of Engineering for Gas Turbines and
Power 130: 0515011.
Gokulakrishnan, P., and Lawrence, A. D. (1999). An Experimental Study of the Inhibiting
Effect of Chlorine in a Fluidized Bed Combustor. Combustion and Flame 116:
64052.
Gokulakrishnan, P., Fuller, C. C., Joklik, R. G., and Klassen, M. S. (2012). Chemical Kinetic
Modeling of Ignition and Emissions from Natural Gas and LNG Fueled Gas Turbines, in
Proceedings of ASME Turbo Expo 2012, Copenhagen, Denmark, pp. GT201269902.
Gokulakrishnan, P., Gaines, G., Currano, J., Klassen, M. S., and Roby, R. J. (2007). Experimental
and Kinetic Modeling of Kerosene-Type Fuels at Gas Turbine Operating Conditions.
Journal of Engineering for Gas Turbines and Power 129: 65563.
Gokulakrishnan, P., McLellan, P. J., Lawrence, A. D., and Grandmaison, E. W. (2005). Kinetic
Analysis of NO-Sensitized Methane Oxidation. Chemical Engineering Science 60:
368392.
References 205
Lu, T., and Law, C. K. (2009). Toward Accommodating Realistic Fuel Chemistry in Large-scale
Computations. Progress in Energy and Combustion Science 35(2): 192215.
Lyon, R. K. (1975). Method for the Reduction of the Concentration of NO in Combustion
Effluents Using Ammonia. U.S. Patent 3,900,554.
Lyon, R. K., and Benn, D. (1978). Kinetics of the NO-NH3-O2 Reaction. Proceedings of the
Combustion Institute 17: 60110.
Maghon, H., Kreutzer, A., and Termuehlen, H. (1988). The V84 Gas Turbine Designed for
Base-Load and Peaking Duty. Proceedings of the American Power Conference 60:
21828.
Mellor, A. M. (1976). Gas Turbine Engine Pollution. Progress in Energy and Combustion
Science 1: 11133.
Miller, J. A., and Bowman, C. T. (1989). Mechanism and Modeling of Nitrogen Chemistry in
Combustion. Progress in Energy and Combustion Science 15: 287338.
Miller, J. A., Durant, J. L., and Glarborg, P. (1998). Some Chemical Kinetics Issues in
Reburning: The Branching Fraction of the HCCO + NO Reaction. Proceedings of the
Combustion Institute 27: 23543.
Miller, J. A., and Glarborg, P. (1999). Modeling the Thermal De-NOx Process: Closing in on a
Final Solution. International Journal of Chemical Kinetics 31: 75765.
Miller, J. A., Klippenstein, J., and Glarborg, P. (2003). A Kinetic Issue in Reburning: The Fate
of HCNO. Combustion and Flame 135: 35762.
Miller, J. A., Pilling, M. J., and Troe, J. (2005). Unravelling Combustion Mechanisms through
a Quantitative Understanding of Elementary Reactions. Proceedings of the Combustion
Institute 30: 4388.
Mongia, H. C. (2010a). Correlations for Gaseous Emissions of Aero-propulsion Engines from
Sea-Level to Cruise Operation, in 48th AIAA Aerospace Sciences Meeting, Orlando, FL,
AIAA Paper 20101530.
(2010b). On Initiating 3rd Generation of Correlations for Gaseous Emissions of
Aero-Propulsion Engines, in 48th AIAA Aerospace Sciences Meeting, AIAA Paper
20101529.
Mongia, R. K., Tomita, E., Hsu, F. K., Talbot, L., and Dibble, R. W. (1996). Use of an Optical
Probe for Time-resolved in situ Measurement of Local Air-to-fuel Ratio and Extent
of Fuel Mixing with Applications to Low NOx Emissions in Premixed Gas Turbines.
Proceedings of the Combustion Institute 26: 274955.
Moreac, G., Dagaut. P., Roesler, J. F., and Cathonnet, M. (2006). Nitric Oxide Interactions
with Hydrocarbon Oxidation in Jet-stirred Reactor at 10 atm. Combustion and Flame
145: 51220.
Moskaleva, L. V., and Lin, M. C. (2000). The Spin-Conserved Reaction CH + N2 H +NCN:
A Major Pathway to Prompt NO Studied by Quantum/Statistical Theory Calculations
and Kinetic Modeling of Rate Constant. Proceedings of the Combustion Institute 28:
2393401.
Mueller, M. A., Yetter, R. A., and Dryer, F. L. (1999). Flow Reactor Studies and Kinetic
Modeling of the H2/O2/NOx and CO/H2O/O2/NOx Reactions. International Journal of
Chemical Kinetics 31: 70524.
Myerson, A. L. (1974). The Reduction of Nitric Oxide in Simulated Combustion Effluents by
Hydrocarbon-Oxygen Mixtures. Proceedings of the Combustion Institute 15: 108592.
Myerson, A. L., Taylor, F. R., and Faunce, B. G. (1957). Ignition Limits and Products of
the Multistage Flames of Propane-Nitrogen Dioxide Mixtures. Proceedings of the
Combustion Institute 6: 15463.
Naik, S., and Laurendeau, N. M. (2004). LIF Measurements and Chemical Kinetic Analysis
of Nitric Oxide Formation in High-Pressure Counterflow Partially Premixed and
Nonpremixed Flames. Combustion Science and Technology 176: 180953.
Newburry, D. M., and Mellor, A. M. (1996). Semiempirical Correlations of NO(x) Emissions
from Utility Combustion Turbines with Inert Injection. Journal of Propulsion and Power
12(3): 52733, MayJune.
References 207
Nguyen, Q. V., Edgar, B. L., Dibble, R. W., and Gulati, A. (1995). Experimental and Numerical
Comparison of Extractive and In Situ Laser Measurements of Non-Equilibrium Carbon
Monoxide in Lean-Premixed Natural Gas Combustion. Combustion and Flame 100:
395406.
Prather, M., and Sausen, R. (1999). Potential Climate Change from Aviation, chapter 6 in
Aviation and the Global Atmosphere, Penner, J. E. et al., eds., The Press Syndicate of the
University of Cambridge, New York, USA, pp. 187213.
Rink, K. K., and Lefebvre, A. H. (1989). Influence of Droplet Size and Combustor Operating
Conditions on Pollutant Emissions. International Journal of Turbo and Jet Engines 6:
11321.
Rizk, N. A., and Mongia, H. C. (1993). Semi-analytical Correlations for NOx, CO and UHC
Emissions. Journal of Engineering for Gas Turbines and Power 115: 61219.
(1995). A Semi-analytical Emissions Model for Diffusion Flame, Rich/lean and Premixed
Lean Combustors. Journal of Engineering for Gas Turbines and Power 117: 290301.
Rutar, T., Horning, D. C., Lee, J. C. Y., and Malte, P. C. (1998). NOx Dependency on Residence
Time and Inlet Temperature for Lean-Premixed Combustion in Jet-Stirred Reactors, in
ASME/ IGTI Turbo Expo, Paper No. 98-GT-433.
Sano, T. (1984). NO2 Formation in the Mixing Region of Hot Burned Gas with Cool Air.
Combustion Science and Technology 38: 12944.
Schmidt, C. C. (2001). Flow Reactor Study of the Effect of Pressure on the Thermal DE-NOx
Reaction. Stanford University, PhD Dissertation.
Schoenung, S. M., and Hanson, R. K. (1981). CO and Temperature Measurements in a Flat
Flame by Laser Absorption Spectroscopy and Probe Techniques. Combustion Science
and Technology 24: 22737.
Skalska, K., Miller, J. S., and Ledakowicz, S. (2010). Trends in NOx Abatement: A Review.
Science of the Total Environment 408: 397689.
Skreiberg, O., Kilpinen, P., and Glarborg, P. (2004). Ammonia Chemistry below 1400 K under
Fuel-rich Conditions in a Flow Reactor. Combustion and Flame 136: 50118.
Smith, G. P., Golden, D. M., Frenklach, M. et al. GRI-Mech 3.0. [Online]. Available at http//
www.berkeley.edu/gri-mech/.
Smoot, L. D., Hill, S. C., and Xu, H. (1998). NOx Control through Reburning. Progress in
Energy Combustion Science and Technology 24: 385408.
Steele, R. C., Malte, P. C., Nicol, D. G., and Kramlich, J. C. (1995). NOx and N2O in
Lean-Premixed Jet-Stirred Flames. Combustion and Flame 100: 4409.
Sturgess, G. J., Zelina, J., Shouse, D. T., and Roquemore, W. M. (2005). Emissions Reduction
Technologies for Military Gas Turbine Engines. Journal of Propulsion and Power 21(2):
193217.
Thomsen, D. D., Kuligowski, F. F., and Laurendeau, N. M. (1999). Modeling of NO Formation
in Premixed, High-Pressure Methane Flames. Combustion and Flame 119: 30718.
Vereecken, L., Sumathy, R., Carl, S. A., and Peeters, J. (2001). NOx Reduction by Reburning:
Theoretical Study of the Branching Ratio of the HCCO+NO Reaction. Chemical Physics
Letters 344: 4006.
Violi, A. et al. (2002). Experimental Formulation and Kinetic Model for JP-8 Surrogate
Mixtures. Combustion Science and Technology 174: 399417.
Walker, R. W., and Morley, C. (1997). Basic Chemistry of Combustion, in Low-Temperature
Combustion and Autoignition, Elsevier, New York, NY: Elsevier, vol. 35, ch. 1,
pp. 1120.
Warnatz, J. (1984). Rate Coefficients in the C/H/O System, chapter 5 in Combustion
Chemistry, Gardiner, W. ed., Springer-Verlag, pp. 197360.
Wendt, J. O., Sternling, C. V., and Matovich, M. (1973). Reduction of Sulfur Trioxide and
Nitrogen Oxides by Secondary Fuel Injection. Proceedings of the Combustion Institute
14: 897904.
Westbrook, C. K., and Dryer, F. L. (1984). Chemical Kinetic Modeling of Hydrocarbon
Combustion. Progress in Energy and Combustion Science 10: 157.
208 NOx and CO Formation
Yetter, R. A., Dryer, F. L., and Rabitz, H. (1991a). A Comprehensive Reaction Mechanism
for Carbon Monoxide/Hydrogen/Oxygen Kinetics. Combustion Science and Technology
79: 97128.
(1991b). Flow Reactor Studies of Carbon Monoxide/Hydrogen/Oxygen Kinetics.
Combustion Science and Technology 79: 12940.
Zeldovich, Y. B. (1946). The Oxidation of Nitrogen in Combustion and Explosions. Acta
Physicochimica U.R.S.S. 21: 577628.
8 Emissions from Oxyfueled or High-Exhaust
Gas Recirculation Turbines
Alberto Amato, Jerry M. Seitzman, and Timothy C. Lieuwen
8.1 Introduction
This chapter discusses emissions from systems with extensive levels of exhaust gas
recirculation (EGR)1 or that use oxygen rather than air as a reactant (referred to
here as oxyfuel combustion). Such systems have unique attributes that warrant a
dedicated chapter in this treatment. First, the systems in which EGR or oxyfuel
would be deployed have different degrees of freedom and requirements. For exam-
ple, both are prominent candidates for carbon capture and storage (CCS) (Griffin
et al., 2008; Budzianowski, 2010), where emissions requirements are driven by pipe-
line or geologic reservoir constraints rather than by atmospheric pollution consider-
ations. Second, while CO2 and H2O dilution have been discussed in Chapters 5 and 7,
their presence at very high levels in systems with EGR can provide a significant per-
turbation of the nominal reactant kinetics (such as in the radical pool) and requires
a focused treatment.
As noted earlier, EGR and oxyfuel combustion for gas turbine applications
are promising approaches to implement CCS in gas turbine power plants. EGR has
also been proposed as a means of promoting fuel flexibility (enabling use of fuels
with low heating value (Danon et al., 2010) and high hydrogen content (Lckerath
et al., 2008)), and for increasing static stability (resistance to flashback/blowout)
(Kalb and Sattelmayer, 2004) and dynamic stability (ElKady et al., 2009) relative to
lean premixed combustors, while enabling low levels of pollutant emissions.
Exhaust gas recirculation has been extensively utilized in industrial burners and
internal combustion engines as a strategy for controlling NOx emissions (Turns, 2000).
Recirculation of flue/exhaust gases increases the heat capacity of the burning mixture
and, for a given quantity of heat release, reduces peak temperatures in the combus-
tor, decreasing the formation of thermal NOx. For carbon capture applications, the
main advantage of EGR is the higher CO2 concentration in the flue gases compared
to common air-fired combustors. This allows a decrease in volume and costs of CO2
separation facilities with a significant reduction of overall CO2 capture costs.
1
Also denoted as flue gas recirculation (FGR).
209
210 Emissions from Oxyfueled/EGR Turbines
The use of oxyfuel combustion, on the other hand, has typically been limited to
processes requiring high temperatures like glass or steel production (Wall, 2007), or
for rockets or subsea vehicles (Yossefi et al., 1995). In oxyfuel combustion, the fuel
is burned utilizing almost pure oxygen, usually obtained from air separation units
(ASU) under stoichiometric conditions to avoid requiring excess supplies of fuel
and O2. The flame temperature is controlled by diluting the reactants with either
steam or CO2. This results in a flue gas almost entirely constituted of water vapor
and CO2. The water can be easily removed by condensation, allowing the remaining
CO2 to be sequestered.
The presence of high levels of combustion products in the reactants is a key dif-
ferentiater of oxyfuel and EGR combustion relative to hydrocarbon-air combustion.
Dilution by H2O and CO2 impacts combustion primarily through changes in: (1)
mixture-specific heat and flame temperature, (2) transport properties (thermal con-
ductivity, mass diffusivity, and viscosity), (3) radiative heat transfer, and (4) chemical
kinetic rates. These are each discussed next.
As triatomic molecules, CO2 and H2O have higher molar specific heat values
compared to N2. Consequently, to achieve comparable turbine inlet temperatures,
combustion with CO2 or H2O dilution requires either lower diluent levels or equiva-
lence ratios closer to stoichiometric compared to operation with air. These changes
in fuel and oxygen concentration, in turn, influence the mixtures kinetic character-
istics as discussed later.
Difference in gas transport properties have important influences upon quan-
tities such as laminar flame speed, sensitivity to aerodynamic stretch, entrain-
ment processes in mixing layers, and heat transfer in boundary layers. Figure 8.1
compares values of thermal conductivity, , binary mass diffusivity (with O2), D,
and dynamic viscosity, , for N2, CO2, and H2O obtained from the transport data-
base of CHEMKIN (Kee et al., 2007). While the thermal conductivity and viscos-
ity of CO2 are very close to those of N2, the mass diffusivity of oxygen in CO2 is
approximately 20 percent lower than in N2. The thermal conductivity and mass
diffusivity of H2O are considerably higher than the values for N2 and CO2, while
waters viscosity is much closer to that of the other two diluents, especially at high
temperatures.
CO2 and H2O dilution also affect radiation heat transfer, as these species are
far more effective absorbers and emitters than O2 or N2. For compact combustors
and low operating pressures (conditions that lead to an optically thin gas), dilu-
tion with significant amounts of H2O and CO2 is most likely to result in additional
heat loss from the combustion zone and hot products gases to the combustor walls
(Andersson and Johnsson, 2007). However, for high operating pressures and/or large
enough combustors, 1-D laminar flame simulations (Guo et al., 1998; Ju et al., 1998;
Ruan et al., 2001; Chen et al., 2007; Maruta et al., 2007) suggest that the presence
of significant levels of CO2 or H2O in the reactants leads to additional preheating
through radiative absorption of heat emitted from product gases, thus increasing
flame speeds, extending flammability limits, and influencing pollutant formation pro-
cesses (Naik et al., 2003).
8.1 Introduction 211
x 104
8
N2
CO2
6
H2O
D (m2/s)
0
500 1000 1500 2000
T (K)
x 105
8
N2
CO2
6
H2O
(Pas)
0
500 1000 1500 2000
T (K)
0.3
N2
CO2
0.2 H2O
(W/(mK))
0.1
0
500 1000 1500 2000
T (K)
Figure 8.1. Binary diffusion coefficient (with O2) at P = 1atm, D, dynamic viscosity, , and ther-
mal conductivity, , of N2, CO2 and H2O calculated from the transport database of CHEMKIN
(Kee et al., 2007).
Finally, CO2 and H2O are not passive diluents like N2, rather they have a direct
impact on chemical kinetics. The presence of CO2 and H2O in the reactants chemi-
cally affects hydrocarbon combustion and pollutant formation processes, mainly by
altering the size and partitioning of the H/OH/O radical pool. The principal reaction
212 Emissions from Oxyfueled/EGR Turbines
of CO2 with this radical pool occurs via the reaction (Liu et al., 2001, 2003; Glarborg
and Bentzen, 2007),
CO2 + H CO + OH R1
which is important in establishing a CO/CO2 partial equilibrium in high-temperature
conditions (see also Chapter 7). Water participates in the O/OH/H pool reactions
directly through the reaction (Le Cong and Dagaut, 2009)
H 2 O + O OH + OH R2
and indirectly through the recombination reactions (Hwang et al., 2004. Le Cong et
al., 2010)
H + O2 (+ M) HO2 (+ M) R3
H + OH(+ M) H 2 O(+ M) R4
because of the high third body efficiency of water (16 larger than N2 and 4.3
greater than CO2) (Anderlohr et al., 2010). In rich conditions, such as those found
on the fuel side of diffusion flames, its influence through the reaction
H 2 O + H OH + H 2 R5
can also become significant (Zhao et al., 2002; Renard et al., 2003; Hwang et al.,
2004). The competition between all these reactions and the main branching reaction
for high-temperature combustion (see Chapter 7)
H + O2 O + OH R6
generally controls the size and partition of the O/OH/H radical pool. In typical EGR
and oxyfuel combustion conditions, H2O and CO2 dilution tend to increase OH con-
centration while decreasing O and H concentrations relative to combustion in lean
air at the same temperature (Williams et al., 2008; Guethe et al., 2009). The actual
partitioning of the radical pool, however, is also strongly dependent on other com-
bustion parameters (e.g., stoichiometry and temperature).
In the remainder of this chapter, we first present a brief overview of the emission
requirements imposed by CCS processes. These requirements are markedly differ-
ent from those imposed by the interaction of pollutants in the terrestrial atmosphere,
and are necessary to understand constraints on the combustion process. Industrial
standards for the purity of CO2 for storage are still lacking, but the principal issues
caused by impurities present in sequestered CO2 stream are clear and, as such, are
briefly reviewed. This is followed by separate sections on EGR and oxyfuel com-
bustion. The final part of each section provides insight into CO and NOx formation
processes and emissions characteristics under EGR and oxyfuel conditions.
capture process is to produce a CO2-rich stream from the exhaust gases, with a purity
level imposed by the two successive phases. Transport (usually at high density) can
take place with pipelines or on ships. CO2 must be transformed into a form with high
density (liquid, solid, or in supercritical phase). The transported CO2 can be seques-
tered in geological formations, such as deep saline aquifiers/depleted gas reservoirs,
or used for enhanced oil recovery (EOR) or enhanced coal bed methane recovery
(ECBM).
With respect to emissions, the key issue in designing a gas turbine cycle accom-
modating CCS is to reduce the concentration of compounds other than CO2 in the
exhaust gases to acceptable levels for transport and storage while minimizing exces-
sive energy and economical penalties in the capture process. Control of pollutant
emissions in the turbine exhausts may have a significant impact on the costs asso-
ciated with flue gas cleanup. The following paragraphs describe the emission con-
straints imposed by CCS on EGR and oxyfuel combustion systems.
example, introducing into the MEA solvents chemical inhibitors (like hindered
ammine (Aaron and Tsouris, 2005)), or modifications of the absorption-desorption
process involving deoxidation of the CO2-rich ammine (Chakravarti et al., 2001).
Development of new solvents, membranes, and process integration is the subject of
ongoing research (IEA, 2010).
CO2 capture from oxyfuel exhaust gases generally consists of a water condensa-
tion step, usually performed by vapor-liquid separator drums at different pressures,
followed by a volatiles removal step performed by distillation or flash columns
(Aspelund and Jordal, 2007; Pipitone and Bolland, 2009; White et al., 2009). The
additional need for gas cleanup, in addition to water condensation, is potentially a
significant drawback for oxyfuel combustion technologies compared to capture with
ammine absorption for EGR cycles, which has only one purification step (Li, Yan,
Anhenden, 2009; Sass et al., 2009). These multi-pollutant control issues in oxyfuel
combustion have been studied for coal plants (Toftegaard et al., 2010) and will need
to be investigated for oxyfuel gas-turbine-based power plants (Aspelund and Jordal,
2007). This calls for careful consideration of the trade-offs between the CO2 purity
specifications imposed by the different phases of CCS.
Large-scale transport and storage of CO2 is not a new technology, and CO2 has
been captured, transported, and used for enhanced oil recovery (EOR) purposes
in the United States, Canada, and Norway for several decades. Vandenhengel and
Miyagishima (1993) present an overview of capture, conditioning and transport
technologies, costs, and specifications for some sources and sinks in the United
States and Canada. As shown in Table 8.1, the industry has already set specifications
for CO2 transport and storage. The first column shows flue gas composition for the
Canyon Reef Project (Metz, 2005), where CO2 from Shell Oil Company process-
ing plants is moved to the Val Verde basin. The second column shows composition
for the Weyburn Pipeline (Metz, 2005), where CO2 is transported from the Great
Plains Synfuels Plant in the United States to the Weyburn-Midale EOR project in
Saskatchewan, Canada.
In spite of the experience in EOR applications, the gas purity requirements for
CCS have not yet been established, mainly because they depend highly on actual
CCS scenarios. The issues related to impurities contents in the sequestered CO2
stream on transport and storage processes are numerous. Here we report only a brief
description of the main issues (drawn mainly from De Visser et al., 2008) and a list
of several others in Table 8.2. With respect to storage, the main risk due to impuri-
ties (especially from SO2 and O2) present in the sequestered CO2 is represented by
the formation of precipitates that can decrease the porosity of the reservoir rock.
The main issues for gas transportation via pipelines are related to corrosion. In the
presence of water vapor, CO2 can cause carbonic acid corrosion (so-called sweet cor-
rosion) and the formation of solid ice-like crystals known as hydrates, while SO2 or
H2S can cause sulfuric acid (or sour) corrosion, which depends on the dew point of
8.2 CCS Emissions Requirements 215
Table 8.1. Specifications for two CO2 transport pipelines for EOR
Table 8.2. Issues arising from the presence of impurities in captured CO2 stream
the mixture. The presence of noncondensable species can also significantly increase
the required compression work, which accounts for a significant part of the overall
CCS energy consumption. Finally, the concentrated CO2 stream might be deemed a
hazardous waste depending on the content of contaminants such as H2S, SOx, NOx,
and hydrocarbons, thus requiring special treatment.
Naturally, there is an economic optimum between the costs of CO2 capture and
of transport/storage facilities. To minimize the cost of the capture process, the prefer-
able option is to co-store as many of the impurities (SOx, NOx, noncondensable gas
species, and water) as possible. In turn, this will require more expensive materials
for the transport facilities to withstand potentially corrosive environments and an
increase in the compression work and size of the storage site. These technical con-
siderations, together with the limitations on the flue gas composition imposed by the
combustion process, will set the actual purity requirements, which can then be used
216 Emissions from Oxyfueled/EGR Turbines
Proposed configurations of EGR burners for gas turbines engines are based on
either internal or external recirculation. Internal approaches mix products with
fresh reactants by direct recirculation of flue gases within the combustion cham-
ber while external (or recuperated) approaches recirculate products only after
they exit the gas turbine. Interest in internal EGR for gas turbine combustors has
been frequently motivated by its success in industrial furnaces as a technique for
NOx abatement (Wnning and Wnning, 1997; Cavaliere and de Joannon, 2004); in
these applications, the acronyms flameless or mild combustion are sometimes
employed because combustion often occurs without a visible flame front. The com-
bustion chamber of a gas turbine, unlike an industrial furnace, is adiabatic (no heat
extraction), operates at high pressure, and typically operates with higher oxygen
levels (before combustion and in the combustion products) (Levy et al., 2004).
These differences tend to accelerate the kinetics of the combustion process, making
it difficult to achieve sufficient fuel/air mixing with exhaust gases before they react.
To overcome these difficulties, several combustor geometries have been proposed
(Kalb and Sattelmayer, 2004; Neumeier et al., 2005; Li et al., 2006; Gopalakrishnan
et al., 2007; Bobba et al., 2008; Duwig et al., 2008; Lckerath et al., 2008; Schtz
etal., 2008; Undapalli et al., 2009; Danon et al., 2010; Lammel et al., 2010; Lv et al.,
2010; Sadanandan et al., 2011).
8.3 Exhaust Gas Recirculation 217
(a) 100
Baseline, air
With EGR 25 atm
1 atm
NO (ppm, dry)
10
1
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
(b) 100
Baseline, air
With EGR 25 atm
NO (ppm, dry at 15%O2)
10
1 atm
1
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
Figure 8.2. Dependence of NO emissions uncorrected (a) and corrected to 15 percent O2 (b),
on adiabatic flame temperature for a fixed residence time (25 ms), initial temperature (650 K),
and for pressures of 1 and 25 atm.
from the beginning of the flame, defined as the point where the temperature rises by
five degrees to 655 K.
Figure 8.2 plots NO emissions as a function of the adiabatic flame temperature
(Tad) of the mixture for the range 17502000 K at 1 and 25 atm pressure. The equiv-
alence ratios needed to obtain these flame temperatures are ~ 0.50.64 when air is
used as the oxidizer, and ~ 0.670.86 for EGR; significantly higher levels of excess
oxygen are available in the undiluted air case. Figure 8.2a shows that EGR gener-
ally leads to lower NOx emissions for a given flame temperature and residence time,
except at low temperature and pressure. However, when corrected to 15 percent O2
(Figure 8.2b), NOx levels are always lower with EGR. In these cases, calculated NO2
values are less than 1 percent of the total NOx concentration.
Some insight into the source of NOx can be obtained by investigating the NO
levels through the 1-d combustor. Figure 8.3 shows the NO time history for the air
8.3 Exhaust Gas Recirculation 219
50
Baseline, air
With EGR
40
P = 25 atm
NO (ppm, dry)
30
20 P = 1 atm
10
0
0 5 10 15 20 25 30
Residence time (ms)
Figure 8.3. EGR effect on NO concentration profiles in the 1-d flame for a fixed adiabatic
flame temperature of 1900 K.
and EGR cases at 1 and 25 atm for a fixed adiabatic flame temperature Tad = 1900
K (corresponding to ~ 0.58 for the baseline-air case and ~ 0.78 for the case
with EGR). At atmospheric pressure, much of the NO is formed in the flame zone,
indicated by the sudden jump in NO concentration at low residence time. At high
pressures, most of the NO is formed in the post-flame zone, and this is where EGR
has the most noticeable inhibiting effect on NOx formation. The results can also be
interpreted as indicating that EGR can be used to maintain low NOx emissions even
if combustor residence times are increased.
The mechanistic pathways responsible for the total NO emissions are detailed
in Figures 8.4 and 8.5. They show the relative importance of the primary kinetic
pathways for the 1 and 25 atm cases, following Rutar and Malte (2002). Changes
in the relative importance of the different mechanisms because of EGR are most
evident at low pressure. These changes are primarily brought about by the decrease
of oxygen concentration in the oxidizer. For the Zeldovich (thermal) pathway, which
dominates at high temperatures, EGR causes both a decrease in the absolute value
of thermal NOx yield and also, especially at low pressure, in its contribution relative
to other pathways. The lower O2 levels in the EGR-diluted reactants lead to a reduc-
tion in atomic oxygen O levels (Guethe et al., 2009; Li et al., 2009). This, in turn, low-
ers the reaction rate of reaction R7, which is the rate-limiting step of thermal NOx
formation. Similar considerations apply for the N2O pathway, because its initiation
reaction (R9) also depends on the presence of atomic oxygen. The N2O pathway is
also responsible for a greater amount of the NOx emissions at high pressure, as R9
is a three-body reaction.
The relative importance of prompt NOx increases for EGR-diluted oxidizer
compared to air, at a given flame temperature. For the higher fuel-to-oxygen ratios
associated with EGR, hydrocarbon radical concentrations increase, which pro-
motes prompt NOx formation through reactions like R8. This is more evident at
220 Emissions from Oxyfueled/EGR Turbines
(a) 80
Zeldovich
70
N2O
60 NNH
NO contribution (%)
Prompt
50
40
30
20
10
0
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
(b) 80
Zeldovich
70
N2O
60 NNH
NO contribution (%)
Prompt
50
40
30
20
10
0
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
Figure 8.4. Contribution of primary NO production mechanisms to total NO yield for 1 atm,
res = 25ms (a) baseline, air (b) with EGR.
low pressure, because most of the high-pressure NOx emissions are produced in the
post-flame zone for the chosen residence time (see Figure 8.3). For premixed com-
bustors with sufficiently low residence times, the prompt pathway can be the most
important NOx formation mechanism controlling EGR effects (Fackler et al., 2007).
This can also be true for rich, partially premixed, or non-premixed flames where
hydrocarbon fragments are present in a wider zone than in a perfectly premixed
flame (Li and Williams, 1999).
The absolute level of NOx produced by the NNH pathway decreases in the pres-
ence of EGR, but its relative importance relative to other mechanisms remains nearly
unchanged. Like the thermal and NO2 pathways, the NNH mechanism depends on
the concentration of oxygen atoms (through the reaction NNH + O NO + NH),
which decreases with EGR. However, this is slightly compensated by the increase of
H concentration at higher equivalence ratios, which favors reaction R10.
8.3 Exhaust Gas Recirculation 221
(a) 80
70
60
NO contribution (%)
50
40
30 Zeldovich
20 N2O
NNH
10 Prompt
0
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
(b) 80
70
60
NO contribution (%)
Zeldovich
50 N2O
40 NNH
Prompt
30
20
10
0
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
Figure 8.5. Contribution of primary NO production mechanisms to total NO yield for 25 atm,
res = 25 ms (a) baseline, air (b) with EGR.
EGR changes NOx production not only through its influence on excess oxygen
levels, but also through the additional kinetic effects associated with the H2O and
CO2 reactions R1 through R5. This effect can be isolated computationally by replac-
ing the CO2 and H2O present in the oxidizer with non-reacting versions that have
the same thermal and transport properties. The adiabatic flame temperature changes
only a few degrees (<7 K) for all the cases considered here when the artificial species
are substituted. This is a useful technique for identifying purely chemical effects (Liu
et al., 2001, 2003; Park et al., 2004; Guo et al., 2008; Le Cong and Dagaut, 2009).
Figure 8.6a shows results for the EGR conditions of Figure 8.2. The kinetic effects
associated with CO2 and H2O tend to reduce NOx emissions. As seen in Figure 8.6b,
OH concentrations increase at the expense of H and O because of CO2 and H2O
chemical interactions. The decrease of O and H radicals reduces the rate of reac-
tions R7 (Zeldovich), R9 (N2O), and R10 (NNH). It also inhibits the oxidation of
CH4 and C2H6, decreasing the level of hydrocarbon radicals and, as a consequence,
222 Emissions from Oxyfueled/EGR Turbines
(a) 100
25 atm
NO (ppm, wet)
1 atm
10
x 103
(b) 5
OH Reacting CO2, H2O
Non-reacting CO2, H2O
4
Mole fraction
3
O
2
H
0
0 1 2 3 4 5
Residence time (ms)
Figure 8.6. Total NOx concentration at 1 and 25 atm with res = 25 ms, for normal EGR calcu
lation and with non-reacting CO2 and H2O added in the oxidizer (a) mole fraction profiles of
O, H and OH in the EGR flame at 1 atm and Tad = 1900 K (b).
the prompt NO pathway. This decrease of NO yield associated with all NO forma-
tion pathways is illustrated in Figure 8.7. While the kinetic effects associated with
CO2 and H2O are generally weak on a quantitative basis, they may be significant for
ultralow NOx systems.
Figure 8.6b also illustrates that changes in the H/O/OH radical partition are
mostly limited to the primary flame zone. O and H concentrations are minimally
influenced by CO2 and H2O kinetic effects in the post-flame zone, which is why these
kinetic effects are even weaker at high pressure, where most NO is produced in the
post-flame region (see Figure 8.3). For non-premixed flames, which have reactions
distributed on larger zones compared to premixed flames, CO2 and H2O chemical
effects can be more significant (Liu et al., 2001; Park et al., 2002).
In contrast to NOx, CO emissions tend to increase with EGR. This is mainly
due to the higher CO equilibrium levels associated with EGR. Considering the
equilibrium partition between CO and CO2 (e.g., CO + 1 2 O2 CO2 ), it follows
8.3 Exhaust Gas Recirculation 223
10
Reacting CO2, H2O
9
Non-reacting CO2, H2O
8
NO contribution (ppm, wet)
Zeldovich
7 Prompt
6
5
4
NNH
3
2
1
N2O
0
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
Figure 8.7. NO contribution of different formation mechanisms to the total NO yield for the
1 atm cases of Figure 8.6a.
that CO levels will increase because of the higher content of CO2 in the combustion
products associated with CO2 dilution and with the higher fuel-air ratio for EGR
compared to pure air combustion at the same flame temperature. In addition, the
increased level of CO2 caused by EGR reduces the net rate of CO oxidation accord-
ing to reaction R1. With increasing pressure, CO equilibrium levels are lower and
the CO oxidation rate increases (Guethe et al., 2009) (see also Chapter 7), although
CO emissions still tend to be higher than in air-fired combustors. For all the flames
simulated here, the oxidation rate of CO is fast enough to have reached the equilib-
rium CO concentration for a residence time of 25 ms, even at 1 atm. However, this
conclusion is a strong function of dilution levels, and CO levels are a major issue at
higher dilutions.
Finally, we include effects of fuel-air unmixedness in our discussion. We can
construct an approximate model to illustrate quasi-steady unmixedness effects by
assuming that the equivalence ratio is a random variable with a Gaussian distribu-
tion, characterized by its average, , and standard deviation, (Lyons, 1980; Li et al.,
2009), as given by
1 ( )2
f ( ) = exp
2 2 2
Perfectly premixed NO values are then weighted by the Gaussian distribution func-
tion to predict the total NOx emission according to the expression
NO = NO() f ()d
Figure 8.8 compares the effect of EGR on NO emissions for perfectly mixed
reactants with those for partially premixed reactants having a relative unmixedness
level / = 7.5 percent. The results are plotted against adiabatic flame temperature at
224 Emissions from Oxyfueled/EGR Turbines
100
/ =7.5%
NO (ppm, dry)
10
/ =0 Baseline, air
With EGR
1
1750 1800 1850 1900 1950 2000
Adiabatic flame temperature (K)
Figure 8.8. Effects of different degree of fuel-air unmixedness on NOx emissions at 25 atm,
res = 25 ms.
combustion temperature. Basic designs of these cycles require only a slight modifica-
tion of combined cycle systems (Bolland and Mathieu, 1998; Kvamsdal et al., 2007).
More sophisticated designs (Yantovski, 1996; Mathieu and Nihart, 1999; Staicovici,
2002) strive to integrate the gas turbine cycle and the CO2-compression step needed
for underground storage or transport so as to limit the associated energy penalties.
Alternatively, H2O can be used as the working fluid and diluent. The main charac-
teristic of this second class of cycles is the integration between the bottoming steam
cycle, from which the H2O is taken, and the topping gas turbine cycle (Sanz et al.,
2005; Anderson et al., 2008). Development and application of either approach, CO2
and H2O oxyfuel cycles, for CCS depend on similar techno-economical issues. Some
studies favor CO2 cycles despite slightly lower thermodynamic efficiencies than H2O
cycles (Sanz et al., 2008).
In addition to CCS, oxyfuel combustion cycles have also been proposed for inte-
grated gasification combined cycles (IGCC) (Casleton et al., 2008). For this appli-
cation, nearly zero NOx emissions provides a compelling motivation relative to air
combustion options. This is enhanced by the fact that oxygen-blown gasifier technol-
ogy for syngas production already requires an ASU. Moreover, the oxyfuel system
does not require the use of water-gas shift reactors to produce H2, with their atten-
dant capital and operating costs.
There are also a few studies of gas turbine oxyfuel combustors (see, for example,
Chorpening et al., 2003, 2005; Williams et al., 2008; Amato et al., 2011a, 2011b; Kutne,
2011). There does not appear to be a clear choice for the design of an oxyfuel com-
bustor (diffusion/premixed/partially premixed). From an operability point of view,
a key concern for oxyfuel combustors is blowoff (Amato et al., 2011b), since H2O/
CO2-diluted stoichiometric flames have lower flame speeds and extinction stretch
rates than lean air flames at the same temperature. For the same reason, flashback
and autoignition issues are much less problematic, and thermoacoustic instability
trends are mixed (Chorpening et al., 2005; Ditaranto and Hals, 2006).
This section describes emission trends for oxyfuel systems, with a focus on CO and
O2 emissions for the reasons discussed in section 8.2.2. A key consideration in the
analysis is the difference in the degrees of freedom oxyfuel systems have compared
to air-fired combustion. For air-fired systems, the flame temperature and stoichiom-
etry are strongly coupled. In contrast, flame temperature and fuel-oxidizer ratio are
decoupled in diluted oxyfuel combustors, with the reactant mixture near stoichio-
metric to avoid wasting fuel and oxygen. Flame temperature is controlled indepen-
dently by the diluent level. In the following calculations, we examine CO2-diluted,
near-stoichiometric oxyfuel combustion of pure CH4; similar considerations apply
for H2O-diluted systems.
Equilibrium calculations provide a useful perspective on emission trends, espe-
cially for long residence time combustors. Figure 8.9 demonstrates the dependence
of equilibrium CO and O2 emissions for oxyfuel combustion on flame temperature
226 Emissions from Oxyfueled/EGR Turbines
(a) 106
Rich CH4/air
105
=1.05
104
CO (ppm, dry)
103 =1.0
102
101
=0.95
100
Lean CH4/air P=1 atm
101
P=15 atm
102
1350 1450 1600 1750 1900 2050 2200 2350 2500
Adiabatic flame temperature (K)
(b) 106
Lean CH4/air
105
104 =0.95
103
O2 (ppm, dry)
=1.0
102
101
100 =1.05
Rich P=1 atm
101
CH4/air
P=15 atm
102
1350 1450 1600 1750 1900 2050 2200 2350 2500
Adiabatic flame temperature (K)
Figure 8.9. Dependence of CO (a) and O2 (b) equilibrium levels for CH4/O2/CO2 mix
tures with initial reactants temperature Tin = 533 K and at pressure of 1 and 15 atm. Results
for CH4/air mixtures at the same temperature and pressure conditions are also shown for
reference.
(and therefore diluent level) for several equivalence ratios and two pressures. Flame
temperatures between 1400 K and 2500 K are achieved by varying the CO2 mole
fraction in the reactants from approximately 84 percent to 50 percent. Also shown
for reference are results for a methane/air flame, where flame temperature is varied
by adjusting the equivalence ratio. As seen in Figure 8.9a, the CO levels are gen-
erally much higher for the CO2-diluted system than the fuel-lean air system at the
same flame temperature (note the logarithmic scaling of the axis). They are also
lowest at oxyfuel lean conditions and increase monotonically with equivalence ratio.
Increasing pressure or decreasing temperature reduces the oxyfuel CO emissions.
The pressure and temperature sensitivities are much lower, however, for rich mix-
tures. Under rich conditions, the trade-offs represented by the water-gas shift reaction
( H 2 + CO2 H 2 O + CO ) dominate. At lower temperatures, the reaction modestly
8.4 Oxyfuel Combustion 227
1000
P = 1 atm
P = 15 atm
= 0.90
Lean CH4/air
100
COmax/COeq
= 1.0
10
= 1.10
1
1600 1700 1800 1900 2000 2100 2200 2300 2400
Adiabatic flame temperature (K)
Figure 8.10. Calculated dependence of CO overshoot (i.e., ratio between maximum and
equilibrium CO molar fraction) upon adiabatic flame temperature (Tin = 533K, P = 1 atm and
15 atm).
favors CO2, but lower temperatures are achieved by adding more CO2, so this shifts
the mixture toward CO. The net result is the weak dependence on Tad observed in
Figure 8.9a. The weak pressure dependence follows from the equimolar nature of
the water-gas shift reaction. For lean mixtures, however, the equilibrium CO level is
dominated by the balance between CO, CO2, and O2, which is highly dependent on
temperature and pressure. Figure 8.9b shows the corresponding results for equilib-
rium O2 emissions. As should be expected, O2 increases as the oxyfuel equivalence
ratio is decreased because of the excess oxygen in the reactants. Similar to CO, the
mole fraction of O2 increases with flame temperature and decreases with pressure.
Unlike CO, the least sensitivity to temperature and pressure occurs for lean mix-
tures, reflecting the fact that O2 is a major product under lean conditions.
Finite rate chemistry effects can be examined with the same modeling approach
described in section 8.3.2. Beginning with CO emissions, there is an important differ
ence between premixed, CO2-diluted, oxyfuel combustion and lean (premixed) air
combustion that should be emphasized. Consider the ratio between the maximum
value of CO (which occurs in the primary flame zone) and its equilibrium value
(which would be obtained at sufficiently long residence time downstream of the pri-
mary flame zone), as shown in Figure 8.10. There is a large CO overshoot in the
oxyfuel flame relative to equilibrium for = 0.9. At 1800 K and 15 atm, for example,
peak CO levels in the flame are roughly seventy times the equilibrium level. Much
greater overshoots occur for lean-air combustion at the same flame temperature,
and higher pressure increases the CO overshoot. This overshoot is a major design
challenge for lean, premixed systems, which must achieve both acceptable CO levels
at low power and NOx at higher power. Significant attention is given to designing
systems to achieve sufficient CO burnout across the operating range. For oxyfuel
systems, likely to operate close to stoichiometric, and possibly slightly rich, this sen-
sitivity is decreased, as indicated by the lower amount of the CO overshoot for those
228 Emissions from Oxyfueled/EGR Turbines
(a) 105
CO (ppm, dry)
104 = 0.90
= 1.00
= 1.05
= 1.10
res = 40 ms
Equilibirum
103
1500 1700 1900 2100 2300 2500
Adiabatic flame temperature (K)
(b) 105
104
O2 (ppm, dry)
= 0.90
103
= 1.00
= 1.05
= 1.10
102
res = 40 ms
Equilibirum
101
1500 1700 1900 2100 2300 2500
Adiabatic flame temperature (K)
Figure 8.11. Dependence of CO (a) and O2 (b) levels for CH4/O2/CO2 combustion on flame
temperature ( Tin = 533 K, P = 1 atm).
cases. While CO levels are substantially higher than for lean systems, the low amount
of overshoot implies substantially less sensitivity of exhaust CO levels to combustor
design details. In other words, similar CO levels will be experienced regardless of
quenching by walls or dilution jets.
We next consider the post-flame relaxation (burnout) of CO. Figure 8.11a shows
the dependence of CO mole fraction on adiabatic flame temperature, under both equi-
librium and fixed residence time conditions at atmospheric pressure. Here, post-flame
residence time is based on the point in the flame where the heat release rate drops
to 10 percent of its maximum value. The results demonstrate the strong influence of
flame temperature on relaxation rates. At high temperatures (Tad>2000 K), the fixed
residence time and equilibrium values are nearly the same. At lower temperatures,
the 40 ms residence time and equilibrium CO concentrations quickly diverge as the
temperature decreases. Similar behavior is observed for the relaxation/burnout of
O2 (Figure 8.11b). The O2 and CO emission trends are correlated because CO is
8.5 Concluding Remarks 229
102
Mole fraction
103
104 O2
CO
res = 40 ms
Equilibrium
105
0.90 0.95 1.00 1.05 1.10
Equivalence ratio
Figure 8.12. Comparison of calculated and experimental data for CO and O2 concentration
in dry exhaust gases: emission levels are represented as function of equivalence ratio at fixed
flame temperatures (1600 K); lines with open symbols are numerical data; filled symbols are
experimental data (Amato et al., 2011a with permission from Elsevier).
in the requirements for systems in which they would be applied. For example, NOx
emissions from systems with EGR are generally lower than air-fueled systems. This
is primarily a result of the reduced amount of excess O2 in the reactants; EGR has
a secondary influence through the impact of CO2 and H2O reactions on the radical
pool. EGR can also have an impact on the level of premixedness needed to achieve
low NOx emissions. Similarly, oxyfuel and EGR combustion can be motivated by
carbon capture and storage (CCS) applications, where emissions requirements are
not driven by impact of atmospheric pollutants, but rather by pipeline or geologic
reservoir considerations. In this case, the level of O2, not normally considered a pol-
lutant, is an important concern. In addition, CO can be an issue for CO2-diluted
oxyfuel systems. Even for steam-diluted approaches, near stoichiometric operation
may lead to excessive CO emissions.
References
Aaron, D., and Tsouris, C. (2005). Separation of CO2 from Flue Gas: A Review. Separation
Science and Technology 40(1): 32148.
Amato, A., Hudak, B., DSouza, P., DCarlo, P., Noble, D., Scarborough, D., Seitzman, J., and
Lieuwen, T. (2011a). Measurements and Analysis of CO and O2 Emissions in CH4/CO2/
O2 Flames. Proceedings of the Combustion Institute 33(2): 3399405.
Amato, A., Hudak, B., DCarlo, P., Noble, D., Scarborough, D., Seitzman, J., and Lieuwen, T.
(2011b). Methane Oxycombustion for Low CO2 Cycles: Blowoff Measurements and
Analysis. Journal of Engineering for Gas Turbines and Power 133(6): 61503.
Anderlohr, J. M., da Cruz, A. P., Bounaceur, R., and Battin-Leclerc, F. (2010). Thermal and
Kinetic Impact of CO, CO2, and H2O on the Postoxidation of IC-Engine Exhaust Gases.
Combustion Science and Technology, 182(1): 3959.
Anderson, R. E., MacAdam, S., Vitieri, F., Davies, D. O., Downs, J. P., and Paliszewski, A. (2008).
Adapting Gas Turbines to Zero Emission Oxy-fuel Power Plants. ASME TurboExpo,
Berlin, Germany.
Andersson, K., and Johnsson, F. (2007). Flame and Radiation Characteristics of Gas-fired
O2/CO2 Combustion. Fuel 86(56): 65668.
Anhenden, M., Rydberg, S., and Yan, J. (2008). Consideration for Removal of Non-CO2
Components from CO2 Rich Flue Gas of Oxy-fuel Combustion. IEA Oxyfuel Workshop,
Yokohama.
Aspelund, A., and Jordal, K. (2007). Gas Conditioning The Interface between CO2
Capture and Transport. International Journal of Greenhouse Gas Control 1(3):
34354.
Bobba, M. K., Gopalakrishnan, P., Periagaram, K., and Seitzman, J. Flame Structure and
Stabilization Mechanisms in a Stagnation-point reverse-flow Combustor. Journal of
Engineering for Gas Turbines and Power 130: 031505.
Bolland, O., and Mathieu, P. (1998). Comparison of Two CO2 Removal Options in Combined
Cycle Power Plants. Energy Conversion and Management 39(16): 165363.
Budzianowski, W. M. (2010). Mass-recirculating Systems in CO2 Capture Technologies: A
Review. Recent Patents on Engineering 4(1): 1543.
Buhre, B. J. P., Elliott, L. K., Sheng, C. D., Gupta, R. P., and Wall, T. F. (2005). Oxy-fuel
Combustion Technology for Coal-fired Power Generation. Progress in Energy and
Combustion Science 31(4): 283307.
Burdet, A., Lachaux, T., de la Cruz Garcia, M., and Winkler, D. (2010). Combustion under
Flue Gas Recirculation Conditions in a Gas Turbine Lean Premix Burner. ASME Turbo
Expo, Glasgow, United Kingdom.
References 231
Casleton, K. H., Breault, R. W., and Richards, G. A. (2008). System Issues and Tradeoffs
Associated with Syngas Production and Combustion. Combustion Science and
Technology 180(6): 101352.
Cavaliere, A., and de Joannon, M. (2004). Mild Combustion. Progress in Energy and
Combustion Science 30(4): 32966.
Chakravarti, S., Gupta, A., and Hunek, B. (2001). Advanced Technology for the Capture of
Carbon Dioxide from Flue Gases. First National Conference on Carbon Sequestration,
Washington, DC.
Chen, Z., Qin, X., Xu, B., Ju, Y., and Liu, F. (2007). Studies of Radiation Absorption on Flame
Speed and Flammability Limit on CO2 Diluted Methane Flames at Elevated Pressures.
Proceedings of the Combustion Institute 31(2): 2693700.
Chorpening, B. T., Casleton, K. H., Richards, G. A., Woike, M., and Willis, B. (2003).
Stoichiometric Oxy-fuel Combustion for Power Cycles with CO2 Sequestration. Third
Joint Meeting of the U.S. Sections of The Combustion Institute, Chicago,IL.
Chorpening, B. T., Richards, G. A., Casleton, K. H., Woike, M., Willis, B., and Hoffman, L. (2005).
Demonstration of a Reheat Combustor for Power Production with CO2 Sequestration.
Journal of Engineering for Gas Turbines and Power 127(4): 7407.
Danon, B., De Jong, W., and Roekaerts, D. (2010). Experimental and Numerical Investigation
of a FLOX Combustor Firing Low Calorific Value Gases. Combustion Science and
Technology 182(9): 126178.
De Visser, E., Hendriks, E., Barrio, M., Mlnvik, M. J., De Koeijer, G., Liljemark, S., and Le
Gallo, Y. (2008). Dynamis CO2 Quality Recommendations. International Journal of
Greenhouse Gas Control 2(4): 47884.
Ditaranto, M., and Hals, J. (2006). Combustion Instabilities in Sudden Expansion Oxyfuel
Flames. Combustion and Flame 146(3): 493512.
Duwig, C., Stankovic, D., Fuchs, L., Li, G., and Gutmark, E. (2008). Experimental and
Numerical Study of Flameless Combustion in a Model Gas Turbine Combustor.
Combustion Science and Technology 180(2): 27995.
ElKady, A. M., Evulet, A. T., Brand, A., Ursin, T. P., and Lynghjem, A. (2009). Application
of Exhaust Gas Recirculation in a DLN F-class Combustion System for Postcombustion
Carbon Capture. Journal of Engineering for Gas Turbines and Power 131(3):
034505.
Evulet, A. T., ElKady, A. M., Branda, A. R., and Chinn, D. (2009). On the Performance and
Operability of GEs Dry Low NOx Combustors Utilizing Exhaust Gas Recirculation for
Post Combustion Carbon Capture. Energy Procedia 1(1): 380916.
Fackler, K. B., Karalus, M. F., Novosselov, I. V., Kramlich, J. C., and Malte, P. C. (2011).
Experimental and Numerical Study of NOx Formation from the Lean Premixed
Combustion of CH4 Mixed with CO2 and N2. Journal of Engineering for Gas Turbines
and Power 133(12): 121502.
Glarborg, P., and Bentzen, L. (2007). Chemical Effects of a High CO2 Concentration in
Oxy-fuel Combustion of Methane. Energy & Fuels 22(1): 2916.
Gopalakrishnan, P., Bobba, M., and Seitzman, J. (2007). Controlling Mechanisms for Low NOx
Emissions in a Non-premixed Stagnation Point Reverse Flow Combustor. Proceedings
of the Combustion Institute 31(2): 34018.
Griffin, T., Bcker, D., and Pfeffer, A. (2008). Technology Options for Gas Turbine Power
Generation with Reduced CO2 Emission. Journal of Engineering for Gas Turbines and
Power 130(4): 041801.
Guethe, F., Stankovic, D., Genin, F., Syed, K., and Winkler, D. (2011). Flue Gas Recirculation
of the ALSTOM Sequential Gas Turbine Combustor Tested at High Pressure. ASME
TurboExpo, Glasgow, United Kingdom.
Guethe, F., de la Cruz Garca, M., and Burdet, A. (2009). Flue Gas Recirculation in Gas
Turbine: Investigation of Combustion Reactivity and NOx Emission. ASME Turbo
Expo, Orlando, FL, USA.
232 Emissions from Oxyfueled/EGR Turbines
Guo, H., Ju, Y., Maruta, K., Niioka, T., and Liu, F. (1998). Numerical Investigation of CH4/
CO2/Air and CH4/CO2/O2 Counterflow Premixed Flames with Radiation Reabsorption.
Combustion Science and Technology 135(16): 4964.
Guo, H., Neill, W., and Smallwood, G. (2008). A Numerical Study on the Effect of Water
Addition on NO Formation in Counterflow CH4/Air Premixed Flames. Journal of
Engineering for Gas Turbines and Power 130(5): 054501.
Hwang, D. J., Choi, J. W. Park, J., Keel, S. I., Ch, C. B., and Noh, D. S. (2004). Numerical Study
on Flame Structure and NO Formation in CH4-O2-N2 Counterflow Diffusion Flame
Diluted with H2O. International Journal of Energy Research 28(14): 125567.
I.E.A. (2010). International Network for CO2 Capture. Available from http://www.ieaghg.org/.
Ju, Y., Masuya, G., and Ronney, P. (1998). Effects of Radiative Emission and Absorption on
the Propagation and Extinction of Premixed Gas Flames. Proceedings of the Combustion
Institute 2(27): 261926.
Kalb, J., and Sattelmayer, T. (2004). Lean Blowout Limit and NOx Production of a Premixed
Sub-ppm NOx Burner with Periodic Flue Gas Recirculation. Journal of Engineering for
Gas Turbines and Power 128(2): 24754.
Kee, R. J., Rupley, F. M., Miller, J. A., Coltrin, M. E., Grcar, J. F., Meeks, E., Moffat, H. K., Lutz,
A. E., Dixon-Lewis, G., and Smooke M. D. (2007). CHEMKIN Release 4.1. 1. Reaction
Design: San Diego, CA.
Kutne, P., Kapadia, B. K., Meier, W., and Aigner, M. (2011). Experimental Analysis of the
Combustion Behaviour of Oxyfuel Flames in a Gas Turbine Model Combustor.
Proceedings of the Combustion Institute 33(2): 338390.
Kvamsdal, H., Jordal, K., and Bolland, O. (2007). A Quantitative Comparison of Gas Turbine
Cycles with CO2 Capture. Energy 32(1): 1024.
Lammel, O., Schtz, H., Schmitz, G., Lckerath, R., Sthr, M., Noll, B., Aigner, M., Hase,
M., and Krebs, W. (2010). FLOX Combustion at High Power Density and High Flame
Temperatures. Journal of Engineering for Gas Turbines and Power 132(12): 121503.
Le Cong, T., Bedjanian, E., and Dagaut, P. (2010). Oxidation of Ethylene and Propene in
the Presence of CO2 and H2O: Experimental and Detailed Kinetic Modeling Study.
Combustion Science and Technology 182(4): 33349.
Le Cong, T., and Dagaut, P. (2009). Experimental and Detailed Modeling Study of the Effect
of Water Vapor on the Kinetics of Combustion of Hydrogen and Natural Gas, Impact on
NOx. Energy & Fuels 23(2): 72534.
Lefebvre, A. (2010). Gas Turbine Combustion, third edition, CRC Press, Ann Arbor, MI.
Levy, Y., Sherbaum, V., and Arfi, P. (2004). Basic Thermodynamics of FLOXCOM, the
Low-NOx Gas Turbines Adiabatic Combustor. Applied Thermal Engineering 24(11):
1593605.
Li, G., Gutmark, E. J., Stankovic, D., Overman, N., Cornwell, M., Fuchs, L., and Vladimir, M.
(2006). Experimental Study of Flameless Combustion in Gas Turbine Combustors.
AIAA Paper 546.
Li, H., ElKady, A. M., and Evulet, A. T. (2009). Effect of Exhaust Gas Recirculation on NOx
Formation in Premixed Combustion Systems. 47th AIAA Aerospace Sciences Meeting,
Orlando, FL.
Li, H., Yan, J., and Anheden, M. (2009). Impurity Impacts on the Purification Process in Oxy-fuel
Combustion Based CO2 Capture and Storage System. Applied Energy 86(2): 20213.
Li, S., and Williams, F. (1999). NOx Formation in Two-stage Methane-air Flames. Combustion
and Flame 118(3): 399414.
Liu, F., Guo, H., Smallwood, G. J., and Glder, . L. (2001). The Chemical Effects of Carbon
Dioxide as an Additive in an Ethylene Diffusion Flame: Implications for Soot and NOx
Formation. Combustion and Flame 125(12): 77887.
Liu, F., Guo, H., and Smallwood, G. J. (2003). The Chemical Effect of CO2 Replacement on
N2 in Air on the Burning Velocity of CH4 and H2 Premixed Flames. Combustion and
Flame 133(4): 4957.
References 233
Lckerath, R., Meier, W., and Aigner, M. (2008). FLOX Combustion at High Pressure with
Different Fuel Compositions. Journal of Engineering for Gas Turbines and Power 130:
011505.
Lv, X., Cui, Y., Fang, A., Xu, G., Yu, B., and Nie, C. (2010). Experimental Test on a Syngas
Model Combustor with Flameless Technology. ASME Turbo Expo, Glasgow, United
Kingdom.
Lyons, V. J. (1980). Fuel/air Nonuniformity Effect on Axisymmetrically Pulsed Turbulent Jet
Flames. AIAA Journal 20(5): 6605.
Maruta, K., Abe, K., Hasegawa, S., Maruyama, S., and Sato, J. (2007). Extinction Characteristics
of CH4/CO2 versus O2/CO2 Counterflow Non-premixed Flames at Elevated Pressures up
to 0.7 MPa. Proceedings of the Combustion Institute 31(1): 122330.
Mathieu, P., and Nihart, R. (1999). Zero-emission MATIANT Cycle. Journal of Engineering
for Gas Turbines and Power 121(1): 11620.
Metz, B. (2005). IPCC Special Report on Carbon Dioxide Capture and Storage, Cambridge
University Press, New York.
Naik, S. V., Laurendeau, N. M., Cooke, J. A., and Smooke, M. D. (2003). Effect of Radiation
on Nitric Oxide Concentration under Sooting Oxy-fuel Conditions. Combustion and
Flame 134(4): 42531.
Neumeier, Y., Weksler, Y., Zinn, B., Seitzman, J., Jagoda, J., and Kenny, J. (2003). Ultra Low
Emissions Combustor with Non-premixed Reactants Injection. 41st AIAA/ASME/SAE/
ASEE Joint Propulsion Conference & Exhibit, Tucson, AZ.
Park, J., Kim, S. G., Lee K. M., and Kim, T. K. (2002). Chemical Effect of Diluents on Flame
Structure and NO Emission Characteristic in Methane-air Counterflow Diffusion
Flame. International Journal of Energy Research 26(13): 114160.
Park, J., Hwang, D. J., Kim, K. T., Lee, S. B., and Kell, S. I. (2004). Evaluation of Chemical
Effects of Added CO2 according to Flame Location. International Journal of Energy
Research 28(6): 55165.
Pipitone, G., and Bolland, O. (2009). Power Generation with CO2 Capture: Technology for
CO2 Purification. International Journal of Greenhouse Gas Control 3(5): 52834.
Rao, A., and Rubin, E. (2002). A Technical, Economic, and Environmental Assessment of
Amine-based CO2 Capture Technology for Power Plant Greenhouse Gas Control.
Environmental Science and Technology 36(20): 446775.
Renard, C., Musick, M., Van Tiggelen, P. J., and Vandooren, J. (2003). Effect of CO2 or
H2O Addition on Hydrocarbon Intermediates in Rich C2H4/O2/Ar Flames. European
Combustion Meeting, Orleans, France.
Ruan, J., Kobayashi, H., Niioka, T., and Ju. Y. (2001). Combined Effects of Nongray Radiation
and Pressure on Premixed CH4/O2/CO2 Flames. Combustion and Flame 124(12):
22530.
Rutar, T., and Malte, P. C. (2002). NO Formation in High-Pressure Jet-Stirred Reactors with
Significance to Lean-Premixed Combustion Turbines. Journal of Engineering for Gas
Turbines and Power 124(4): 77683.
Sadanandan, R., Lckerath, R., Meier, W., and Wahl, C. (2011). Flame Characteristics and
Emissions in Flameless Combustion under Gas Turbine Relevant Conditions. Journal of
Propulsion and Power 27(5): 97080.
Sanz, W., Jericha, H., Moser, M., and Heitmeir, F. (2005). Thermodynamic and Economic
Investigation of an Improved Graz Cycle Power Plant for CO2. Journal of Engineering
for Gas Turbines and Power 127(4)): 76572.
Sanz, W., Jericha, H., Bauer, B., and Gttlich, E. (2008). Qualitative and Quantitative
Comparison of Two Promising Oxy-fuel Power Cycles for CO2 Capture. Journal of
Engineering for Gas Turbines and Power 130(3): 031702.
Sass, B. M., Farzan, H., Prabhakar, R., Gerst, J., Sminchak, J., Bhargava, M., Nestleroth, B., and
Figueroa, J. (2009). Considerations for Treating Impurities in Oxy-Combustion Flue Gas
Prior to Sequestration. Energy Procedia 1(1): 53542.
234 Emissions from Oxyfueled/EGR Turbines
Schtz, H., Lckerath, R., Kretschmer, T., Noll, B., and Aigner, M. (2008). Analysis of the
Pollutant Formation in the FLOX Combustion. Journal of Engineering for Gas
Turbines and Power 130(1): 011503.
Smith, G. P., Golden, D. M., Frenklach, M., Moriarty, N. W., Eiteneer, B., Goldenberg, M.,
Bowman, C., Hanson, R. K., Song, S., Gardiner, W. C., Lissianski, V. V., and Z., Qin
Available from http//www.berkeley.edu/gri-mech/.
Staicovici, M. (2002). Further Research Zero CO2 Emission Power Production: The
COOLENERG Process. Energy 27(9): 83144.
Toftegaard, M. B., Brix, J., Jensen, P. A., Glarborg, P., and Jensen, A. D. (2010). Oxy-fuel
Combustion of Solid Fuels. Progress in Energy and Combustion Science 36(5): 581625.
Turns, S. R. (2000). An Introduction to Combustion, second edition, McGraw Hill, New York.
Undapalli, S., Srinivasan, S., and Menon, S. (2009). LES of Premixed and Non-premixed
Combustion in a Stagnation Point Reverse Flow Combustor. Proceedings of the
Combustion Institute 32(1): 153744.
Vandenhengel, W., and Miyagishima, W. (1993). CO2 Capture and Use for EOR in Western
Canada 2. CO2 Extraction Facilities. Energy Conversion and Management 34(911):
11516.
Wall, T. (2007). Combustion Processes for Carbon Capture. Proceedings of the Combustion
Institute 31(1): 3147.
White, V., Torrente-Murciano, L., Sturgeon, D., and Chadwick, D. (2009). Purification of
Oxyfuel-derived CO2. Energy Procedia, 1(1): 399406.
Williams, T. C., Shaddix, C. R., and Schefer, R. W. (2008). Effect of Syngas Composition
and CO2-diluted Oxygen on Performance of a Premixed Swirl-stabilized Combustor.
Combustion Science and Technology 180(1): 6488.
Wnning, J., and Wnning, J. (1997). Flameless Oxidation to Reduce Thermal NO-formation.
Progress in Energy and Combustion Science 23(1): 8194.
Yantovski, E. (1996). Stack Downward: Zero Emission Fuel-fired Power Plants Concept.
Energy Conversion and Management 37(6): 86777.
Yossefi, D., Ashcroft, S. J., Hacohen, J., Belmont, M. R., and Thorpe, I. (1995). Combustion of
Methane and Ethane with CO2 Replacing N2 as a Diluent. Fuel 74(7): 106171.
Zhao, D., Yamashita, H., Kitagawa, K., Arai, N., and Furuhata, T. (2002). Behavior and Effect
on NOx Formation of OH Radical in Methane-air Diffusion Flame with Steam Addition.
Combustion and Flame 130(4): 35260.
Part 3
9.1Introduction
Premixed combustors for aero engines have been under development for nearly
forty years, yet, at the time of writing, the first airplane with premixed combus-
tion still awaits its entry into service. On the other hand, industrial gas turbines
have made the transition to premixed combustion within ten years and the level of
emissions of nitrogen oxides has decreased tenfold. The differences are due to the
peculiarities of gas turbines in flight and a large part of the chapter will be devoted
to the understanding of the consequences of those differences for premixed or par-
tially premixed combustion. An obvious difference between both applications lies
in the fuels, which are predominantly gaseous for industrial gas turbines and exclu-
sively liquid for aero engines and will continue to be for the foreseeable future.
Therefore, premixing in aero combustors always needs to be discussed together
with prevaporization, and the differences imposed on the liquid fuel preparation
by full or partial prevaporization and premixing are responsible for a large part of
the overall development effort. The other determining differences result from the
thermodynamic cycles specific to high bypass ratio engines and the impact of the
flight profile on the implementation of part load operation. The latter has already
been described in Chapter 1.5 and the concept of staging in lean aero engines has
been presented in Section 1.5.3 such that this chapter will concentrate more on the
implementation of staging and its consequences on the design of the combustor
components.
The chapter consists of three parts that partly also follow a historical order:
Some results of research are presented that are relevant for lean premixed, pre-
vaporized (LPP) combustion, which for a large part were concurrently achieved with
development efforts on LPP combustors. Understanding the limitations and diffi-
culties in the way of fully prevaporized premixed combustion, the concept with the
highest emissions reduction potential, will then supply the base for the discussion of
partially premixing combustors and their operability aspects.
237
238 Premixed Aero Engine Combustors
But first, the combustor emissions need to be put into the perspective of the
other design criteria and the path a combustor development takes to fulfill all of
them. The following list is taken from NASA (Rhode, 2002):
The presenter goes on to state: Emissions are normally the fourth or fifth combus-
tor design priority, a statement that will find the agreement of everybody board-
ing an airplane. However, for the development of premixed combustors the order
is different. After all, there are safe combustors to fly with, and without sufficient
advantage to justify the cost and risk of introduction there is no point in satisfying
the more important airworthiness criteria. Therefore, the gap in emissions needs to
be demonstrated at every step in development before the next is taken. Some of
the operability criteria can only be proven with successive integration of the com-
bustor into the engine, and safety ultimately needs to be established by flight test-
ing. Changes made to the combustor to achieve airworthiness may result in higher
emissions, which might be lowered again by an iterative redesign cycle in the earlier
stages; however, in the later stages, that will no longer be possible. Hence, it is most
important for the candidate designs to show a convincingly high reduction potential
at the beginning of development.
When the need for emissions reduction in aero engines became apparent in the
early 1970s (see Chapter 3.2.1.1), premixed combustion was quickly identified as
a promising candidate. Homogeneous reactor calculations showed that, in the-
ory, a three order of magnitude reduction of NOx should be possible for homo-
geneously lean premixed combustion compared with stoichiometric values.
Experimental proof was furnished in numerous flame tube studies cited, for
instance, in Tacina (1990). A result of one of them is shown in Figure 9.1 (Anderson,
1975). NOx reductions of more than two orders of magnitude were achieved for
= 0.4. CO equilibrium was reached at all conditions after 1.5 ms within exper-
imental uncertainty and together with that, the 99 percent combustion efficiency
benchmark was surpassed. Comparison of the NOx formation curves with time
9.2 Prevaporized Combustion in Aero Engines 239
10
Combustion Equivalence ratio
efficiency (%)
Nitrogen oxides (NOx) emission index (gNO2/KgC3H8)
0.6
99.70
99.50
0.5
99.25
0.4
97.00
shows that = 0.4 and 0.5 curves display an asymptotic behavior with time, which is
not apparent for = 0.6. These differing characteristics are due to the dominance of
different NOx formation mechanisms, as explained in Chapter 7.3. The quick rise at
the beginning of the curves is related to the prompt mechanism. If the flame temper-
ature stays below 1900 K, the thermal NOx formation is very slow. However, for =
0.6, the rate of thermal NOx formation in the experiment was so high that the NOx
concentration rose appreciably and residence time becomes important.
The influence of air pressure is also connected to the relative importance of the
two mechanisms. For perfect premixing below 1900 K (see chapter 10.3.2.1), Leonard
and Stegmaier (1994) were able to demonstrate the same NOx concentration for dif-
ferent pressures because the prompt mechanism does not depend on it. For higher
temperatures, the thermal mechanism takes over and displays the square root depen-
dence on pressure, well known from conventional combustors. Unfortunately, it is at
those temperatures that lean premixed combustion is most urgently required. The
most important applications are large turbofan engines having high-pressure ratios
and combustor exit temperatures for maximum efficiency, which are installed on
airplanes for long-range transport, which burn a large part of the overall fuel used in
aviation. The increase in the reaction rate of the thermal NOx mechanism at stoichio-
metric condition makes the control of NOx emission with RQL combustors very dif-
ficult because the process of dilution through the stoichiometry or quenching cannot
be accelerated accordingly (see Chapter 1.5.2). Figure 9.2 (Correa, 1992) illustrates
240 Premixed Aero Engine Combustors
104
Forward thermal NO formation rate (ppm/ms)
30 atm, 900 K
102
10 atm, 600 K
100
1 atm, 300 K
102
104
106
0.4 0.6 0.8 1.0 1.2 1.4
Equivalence ratio
Figure 9.2. Forward NO formation rate for a laboratory combustor at 1 bar, utility gas turbine
at 10 bar, and aero engine at 30 bar from Correa (1992).
the effect for methane as fuel. At stoichiometric condition, the reaction accelerates
by more than one order of magnitude between a moderate and a high-pressure gas
turbine cycle. On the other hand, burning at lean conditions can slow down the pro-
duction rate by more than three orders of magnitude before blowout.
However, the choice of a target equivalence ratio for lean combustion is not
free. Combustor exit temperatures for high bypass ratio engines are already quite
high. Table 9.1 shows some values for a large turbofan as anticipated for a pressure
ratio of forty-five (Wedlock et al., 1999). Real engines will have different values,
but a pressure ratio of forty-two has become reality. With 1792 K, the combus-
tor exit temperature is still below the critical temperature level for thermal NOx
formation; however, it is above the level where purely convective cooling with
regenerative use of the cooling air is feasible with proved materials. Therefore,
the air needed for film cooling has to be deducted from the overall budget and
the exit temperature of the remaining combustion air is considerably higher. With
20 or more percent used for cooling, the combustion temperature will surpass the
thermal NOx limit, and it is impossible to restrict the NOx formation solely by pre-
mixing; an additional limitation of the residence time is needed. Because of this
coupling of cooling and emission at high combustor outlet temperatures, a reduc-
tion or ultimately abolition of film cooling flow by high-temperature materials for
combustor walls could have a great impact on the emissions of lean combustors.
Ceramic matrix composites hold the promise to deliver these advantages, yet com-
bustors now in development still need to respect the limitations given by the avail-
able metallic materials.
Combustion of kerosene in the gas turbine ends with the oxidation of CO to
CO2. Compared to the heat release connected to the visible flame, it is a slow reac-
tion (see Chapter 7.2). It limits the possibility to control the NOx production by short
residence times. Film-cooled combustors face an additional complexity: if more
9.2 Prevaporized Combustion in Aero Engines 241
S=0
0.1
10
0.2
0.28
NOx emission index (gm/kg fuel)
1
S = Non-uniformity
parameter
S = 0 is a uniform
fuel/air profile
0.1
0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
Equivalence ratio
Figure 9.3. Calculated NOx emission for various nonuniformity parameters and equivalence
ratios for 600 K preheat, 3 bar air pressure, and 2 ms residence time (Lyons, 1982).
residence time is needed, the combustor volume and surface will grow, requiring
more cooling air thus forcing higher combustion temperatures. As NOx formation is
fastest at high power and CO oxidation slowest at low power (see Chapter 1.5.1), the
low-power situation will dictate the residence time.
9.2.2.1 Homogeneity
In a practical combustor, the mixture of fuel and air and, consequently, combustion
product temperatures will always be inhomogeneous in time and space. Because
of the highly nonlinear dependence of NOx formation rate with temperature, the
resulting NOx distribution will be skewed toward higher temperatures. Therefore,
the final emission does not only depend on the global equivalence ratio but also on
the width of the fuel-air ratio distribution of the combusting mixture. This effect
was quantified with homogeneous reactor calculations and flame tube experiments.
Figure 9.3 shows the result of calculations with curves of NOx over mean equiva-
lence ratio for several values of the nonuniformity parameter S (Lyons, 1982). In
this study, it was defined as the average standard deviation of equivalence ratio over
a cross section of the flame tube. The different curves cross around = 0.8 for the
selected flame tube condition of 600 K preheat at 3 bar and 2 ms residence time. For
242 Premixed Aero Engine Combustors
9.2.2.2 Prevaporization
For liquid fuels, the degree of premixing is coupled to prevaporization because
droplets entering the flame will never burn fully premixed. If they burn in a diffu-
sion flame mode, the huge difference between NOx formation at stoichiometric and
lean conditions can fully counteract the beneficial effect of premixing. Hence, the
influence of the degree of prevaporization on emissions is of great interest but hard
to quantify in general terms. If prevaporization is achieved in a premixing passage
within the burner, the liquid can be measured at the burner outlet. But further pre-
vaporization will happen on the way to the flame and by heat conducted from the
flame in the vicinity of the flame front. For lifted flames, this effect can lead to full
prevaporization even with direct injection of the liquid fuel from the burner into the
combustor. Varying the degree of prevaporization for realistic gas turbine condi-
tions and quantifying it right before the flame has not been achieved yet. However,
experimental investigations exist that quantify the degree of prevaporization enter-
ing the combustor, and detailed calculations of single droplet combustion enable
some conclusions.
One of the earliest experiments (Cooper, 1979) used a pressurized flame tube at
3 bar with preheat from 600 to 800 K and two atomizers at different positions before
the flameholder. The first was far enough to assure prevaporization and the second
was placed before the flameholder. Prevaporization varied between 70 and 100 per-
cent. At = 0.7 and 700 K preheat, no influence of prevaporization was measured,
whereas at = 0.6, the NOx emission doubled for the lower prevaporization. These
results were reproduced with a superposition of reactor calculations from pure diffu-
sion burning for the liquid part and pure premixed combustion for the prevaporized
fuel. The calculated NOx variation of 6 percent for = 0.7 was within the experimen-
tal accuracy. However, modern aero engine combustors almost exclusively use air
blast atomizers, where droplets are accelerated by the airflow from small velocities
to the air velocity according to their size. Hence, droplets entering the flame might
have an appreciable slip velocity, which then influences the mode of droplet burning.
For higher droplet Reynolds numbers, the spherical flame extinguishes and a flame
stabilizes in the wake of the droplet. The fuel entering the flame is partially premixed
by the turbulence of the wake. The ensuing development of the NOx production is
shown in Figure 9.4. The steep NOx reduction is due to the change of burning mode.
Here the probable situation of big droplets burning in the post-flame region at high
9.2 Prevaporized Combustion in Aero Engines 243
10
8
EINO (g/kg)
0
0 5 10 15 20
cslip (m/s)
Figure 9.4. Emission index of NOx for a 50 m heptane droplet versus relative droplet veloc-
ity at air temperature of 1567 K and atmospheric pressure (Beck et al., 2007).
temperatures leads to the high EI (Emission Index = gNOx/Kg fuel) at low Reynolds
number.
For practical lean combustors, the following conclusions can be drawn:
At full load and good premixing, remaining droplets will probably be smaller
than 50 m and their spherical flames will contribute disproportionately to NOx
emission.
If the level of premixing is not good or the equivalence ratio is very high, pre-
vaporization doesnt matter.
In part load, emissions from the droplet contribution will always be high because
the background emission from vapor phase combustion will be low. The amount
of turbulence in the post-flame zone determines relative droplet velocities and
hence emissions from the droplet phase.
These conclusions, however, provide no guidance for the designer who wants to know
how much effort should be spent in producing small droplets with the atomizer to
prevent single droplet burning and how much pressure loss should be invested to do
so. No clear answer is at hand. A hint can be derived from a recent investigation of
a generic air blast atomizer (Freitag et al., 2010). At 4 bar, 700 K inlet temperature
and a pressure loss of 3 percent, a flame with a lift-off height of 15 mm was produced.
Nearly complete prevaporization measured by a two order of magnitude reduction
of the Mie-scattering signal from the droplet phase occurred for the mean starting
SMD, around 15 m.
The available literature reports lean blowoff for premixed combustion around
1650 K. The operating range of aero engine combustors and its consequences on com-
bustor design have already been described in Chapter 1.5. Even with a pilot stage, it
is desirable to extend the blowoff limit of lean combustion. Accordingly, experiments
244 Premixed Aero Engine Combustors
were performed with different flameholder geometries, catalytic surfaces, and fuel
injection in the wake of the flameholders (McVey and Kennedy, 1979). A substantial
extension of the stability range was only possible with 5 percent pilot fuel injection,
which allowed stable combustion at = 0.25 for 600 K preheat. The study was con-
fined to the recirculation rates, possible with bluff body flame holders. As the recircu-
lation of these flame holders is coupled to the size of the wake, dependent to the first
order on the blockage of the flow path, the allowable pressure loss of the combustor
sets a limit to the recirculation rates, which are practically achievable. Swirl stabiliza-
tion or large-scale, jet-driven stabilization, as used in the so-called flameless combus-
tion, will enable larger recirculation rates and zones; the price to pay then will be a
higher residence time, which rises together with the size of the recirculation.
Despite this limitation of the investigation, it is possible to draw some general
conclusions. Of the ingredients for stable flame holding at the point of stabilization:
a sufficient amount of incoming fuel, heat, and radicals as well as good mixing, only
mixing and supply of heat can be influenced in fully premixed systems. The fuel
concentration is fixed by the prescribed fuel-to-air ratio, and the concentration of
radicals upstream of the stabilization point depends on the combustion tempera-
ture, which is again set by AFR. Here the OH radical plays a double role as pre-
cursor of ignition as well as decisive intermediate of the Zeldovitch mechanism for
the thermal NOx formation. As both the equilibrium OH concentration as well as
the NOx formation rate displays the same exponential characteristic, there is only
a small temperature interval where both requirements can be met. If, however, a
small amount of extra fuel can be brought to the stabilization point, vigorous mixing,
necessary for stabilization anyway, can dilute the local fuel concentration quickly
enough to beat the thermal NOx formation.
Besides the combustor pressure loss, combustion efficiency is the only factor that
directly influences the overall efficiency of the engine (see chapter 1.3). Current
RQL combustors have 99.9 percent efficiency at all high-power operating points.
The airlines expect a similar performance from lean combustors. For a staged com-
bustor, the lowest operating point for the lean main stage is the cruise condition, at
which this expectation is the hardest to fulfill.
Of the two constituents of the exhaust gas influencing efficiency, unburned
hydrocarbons and CO, CO is the more difficult component to reduce. Significant
amounts of unburned hydrocarbons usually are a sign of imperfect match of fuel
placement and stabilization zone and can be reduced by modifications to the burner
during development. However, control of NOx and CO oxidation place conflicting
demands on residence time as discussed in Section 9.2.1. If the combustor is convec-
tively cooled and the mixture is fully premixed before combustion, satisfying the res-
idence time requirement can be achieved by a suitable sizing of combustor volume
or length. For a direct fuel injection, combustion at lean mixtures must be achieved
with vigorous mixing; this will not stop immediately after the flame. On its way to
9.2 Prevaporized Combustion in Aero Engines 245
2.0
NOx/NOx1050 ratio
1.8
CO / CO 1050 ratio
1.6 Loads 94%, Pressure 4bar
1.4
Ratio
1.2
1.0
0.8
0.6
0.4
1000 1050 1100 1150
Max. combustion chamber wall temp. (C)
Figure 9.5. NOx and CO emission as a function of combustor wall temperature (Altemark
and Knauber, 1987).
the combustor exit, film-cooling air might be entrained, which will further dilute the
mixture and slow down the CO oxidation. In the extreme case, the reacting mixture
will be fully quenched near the wall. Hence, the mixing process must be controlled
up to the end of the combustor.
The influence of the wall temperature on premixed combustion was demon-
strated by Altemark and Knauber (1987). A fully premixed natural gas combustor
was operated in the temperature regime below 1900 K up to 10 bar. The wall was
convectively cooled and its temperature could be controlled by a separate air stream.
The influence of the wall temperature on NOx and CO is shown in Figure 9.5. The
wall temperature was varied between 1000 and 1100C. Its result was a threefold
increase in CO, whereas NOx only varied by 20 percent. A comparison of flame tube
studies carried out for propane (Anderson, 1975) with water-cooled walls and with
air cooling (Roffe and Venkatarami, 1978) for similar flame holders and preheat
finds an increase in the residence time to achieve equilibrium CO from 1.5 ms to 12
ms for the cooler wall (Roffe and Venkatarami, 1978).
9.2.5 Autoignition
some degree of scatter will be inevitable. This effect is especially important for the
high-pressure regime of interest where autoignition times are low and therefore the
allowable premixing passages short. An alternative route was taken by Guin (1999)
by the instrumentation of burners with fiber optic arrays with fast data acquisition
and continuous storage. The operating conditions were changed until stable combus-
tion resulted either in autoignition or in flashback with the additional advantage of
being able to discern between the two mechanisms. Pressures up to 30 bar and tem-
peratures up to 900 K were investigated. It turned out that the data of three of the
investigated burners having very different designs for different operating regimes
could be fitted to one correlation.
0.508 3377
= exp (9.1)
P 0.9 T
Here is the autoignition time in ms, P the pressure in bar, and T the air tempera-
turein K.
This result gives the reassuring message that for well-designed premixers with
good mixing and flashback prevention, the mixing effects are not so important as to
prevent a reasonable prediction of autoignition. The second message is that some
premixing within the burner with kerosene is also possible for higher operating
conditions. Extrapolating for the subsonic takeoff of Table 9.1 yields 0.65 ms. This
picture is completely different from the situation for premixing in industrial gas tur-
bines. Most of the engines burn natural gas in a composition dominated by methane.
Although there is also some scatter in the correlations, especially for low tempera-
tures, the data suggest autoignition times to be more than an order of magnitude
higher as for kerosene (Chen et al., 2004).
9.2.6 Flashback
Next to autoignition, flashback is the second risk connected with premixing pas-
sages. Flashback can occur when the flame velocity is higher than the gas velocity at
the point of flame stabilization. As both the local turbulent flame and gas velocity
depend on many factors, there are as many reasons for flashback. Three different
classes of flashback are described in the literature:
Table 9.1. Subsonic turbofan performance data, ISA SLS rated take-off thrust 459 kN
(Wedlock et al, 1999)
From the description of these mechanisms, it is clear that there are no simple rules
to prevent flashback. Therefore, the development of premixing burners needs to
embody tests to demonstrate a suitable safety margin against autoignition and flash-
back. During such tests, the burner flow function is continually reduced at the rel-
evant condition until autoignition or flashback occurs. A 50 percent reduction is
considered as a necessary value. Based on this assumption, the allowable time for
premixing and hence premixing passage length is cut in half.
even if combustors with only partial premixing are the most likely candidates for
market introduction today. On one hand, there are possible future applications in
supersonic airplanes and in smaller engines where full premixing remains possible.
On the other, the present premixing combustors incorporate the experiences with
premixing and the lessons learned that are most easily explained at the place of their
generation.
8.61 21.74
15 mounting struts
3.81 14.58
60 main stage
Fuel spraybars
0.103 dia. 78.08 dia.
83.70 dia.
61.31 dia.
60 flameholders
30 3
m
32c
75.03 dia. 8, 1
l.=3
72.29 dia. Vo
Vol.=14,979 cm3
63.55 dia.
30 Pilot stage
Linear dimensions in centimeters
Fuel injectors
56.22 dia.
2.16
30.35
Figure 9.6. Radial/axial staged GE combustor of ECCP Ph I, all dimensions in cm (Bahr and
Gleason, 1975).
50
II-6,7
Extrapolated to hot day takeoff
NOx emission index (g/kg fuel)
II-10
40 II-12,15
II-14
III-2
30 f = 0.024+ 0.001
20
10
0
0.0 0.0 0.2 0.3 0.4 0.5 0.6
Pilot fuel flow
Total fuel flow
Figure 9.7. Variation of NOx emissions with pilot fuel fraction of radial/axial staged GE com-
bustor of ECCP Ph. I extrapolated to take off. Symbols are for geometrical variants (Bahr and
Gleason, 1975).
that escape the flame. With a diffusion pilot, its NOx production will be so high that
the whole endeavor becomes questionable. However, reducing pilot fuel and air to
preserve peak temperatures is limited by the staging requirements.
At cruise condition, the stability of the main combustion zone was unsatisfac-
tory. A circumferential staging was introduced where every second main fuel bar was
switched off. To attain the same combustion efficiency as for the takeoff then also led
to the same level for NOx. Although the cruise condition is not regulated and there-
fore no cruise emissions can be found in the ICAO data bank, it can be assumed that
250 Premixed Aero Engine Combustors
Main combustion
zone
Dilution air
Pressure atomizing
nozzle for main zone
Ignitor 60 outer main swirlers
Figure 9.8. Swirl Vorbix combustor of Pratt & Whitney ECCP Phase I (Roberts et al., 1975).
the NOx of the lean combustor would have been higher than the emission of the
competing RQL technology. To avoid staging at cruise, an augmentation of the main
stage stability is necessary. As already discussed in Section 9.2.3, addition of pilot
fuel is one method to achieve that goal.
For the approach at 30 percent thrust, operation in the pilot-only mode was
planned. With 23 percent pilot air, the equivalence ratio at that point was 1.9, already
near the critical boundary for excessive soot formation (see Chapter 6). A partially
premixing pilot would be another solution. The goal would be a reduction of the
NOx-producing peak temperature zone, while temperatures in the flow interacting
with the main stabilization zone should stay the same as with a diffusion pilot. To
sustain idle conditions, a piloted partially premixing pilot stage would be necessary,
and Oda et al. (2003) realized such a combustor.
In the same program, Pratt & Whitney developed a concept under the name of
Vorbix combustor, presented in Figure 9.8 as the first explicitly partially premixing
low-emission concept (Roberts et al., 1975). It was axially staged. Behind the pilot
combustor, the main fuel is sprayed with an angle into the throat of the combustor
liner by a pressure-atomizing nozzle. The main air was introduced after the injection
by cylindrical swirl channels, which mixed spray and main air. The spray autoignites
in the hot stream. The aim was partial premixing of the main fuel before combus-
tion. No data exist to tell where stabilization actually happened. Since pilot exhaust
temperatures resulted in very short autoignition times, it can be suspected that igni-
tion started at stoichiometric or rich conditions, even if the cooling from the heat
transfer to the liquid phase is taken into account. Nevertheless, dilution by the main
air effected lean combustion of the bulk of the fuel. Tests were done in a 90 sector
up to 6.8 bar. As in the radially staged combustor, a strong influence of the pilot fuel
split was noted, however with 20 percent pilot fuel, a EI NOx of 12.4 at 99.7 percent
efficiency was measured. Using the scaling formula of the report for NOx, the con-
cept was worse than the radially staged premixer for 99 percent efficiency, but better
for 99.7 percent.
9.3 Development of Lean, Prevaporized Combustors with Premixing Ducts 251
passage with rising velocity. However, this now rather commonplace conclusion was
mainly reached within the development of premixing combustors of industrial gas
turbines.
The HSCT program differed from earlier programs as it comprised an inte-
grated development of airplane and propulsion systems from which an engine cycle
with more benign conditions for premixed combustion evolved. For the mixed flow
turbofan (MFTF), the start of supersonic cruise was the most critical condition for
the premixing system (Greenfield et al., 2004). Combustor inlet pressure was 9 bar,
inlet temperature 938 K, and exit temperature 1911 K. The higher efficiency require-
ment of 99.5 percent coupled to a higher pressure loss allowance of 5.3 percent for
the combustor liner. The goal for cruise NOx was relaxed to an EI of five. In view of
autoignition experience and higher inlet temperature, premixer residence time was
shortened to 1 ms (Guin, 1999). For flashback prevention ~ 100 m/s premixer veloc-
ity was deemed necessary. Combustor residence time was limited to 2 ms to achieve
the NOx goal, since the exit temperature was already within the thermal NOx regime.
With the prescribed 50 percent reduction of cooling air to 18 percent, a main zone
equivalence ratio of 0.5 could be realized. To that effect, a parallel program for the
development of CMC combustor liner materials was initiated. The uniformity goal
for the mixture was set to a standard deviation of 0.15 of the mean to lower the
maximum flame temperature, setting the prevaporization requirement of 90 percent.
At cruise, a staging split of 85 percent fuel through the main was used whereas idle
emissions required 30 percent airflow through the pilot.
With the requirements for main homogeneity and relative fuel flow at cruise, sta-
ging to achieve the operating range became a more difficult task as variable geome-
try was ruled out in the end because of safety reasons. Hence, more than one staging
point and more than two burner rings were considered. Figure 9.9 shows the staging
scheme employed for the jet mix LDI modules. Five different operating modes were
required to stay above the stability limit at 1654 K marked by the lowest horizontal
line and in between the EI NOx limit of five, marked by the highest horizontal line
at 2130 K and the 99.5 percent efficiency line in the middle at 1854 K. Such a staging
scheme might be seen as a necessary evil in industrial gas turbines, but will appear
dreadful in an aero engine and very hard to sell to any airline. GE investigated five
different premixer and lean direct injecting burners with pilot stages. For some con-
figurations, the residence time requirement could not be achieved simply because of
the disproportionate height of the liner, housing three rows of swirling burners. The
curvature radius of the liner in the combustion zone would have been excessively
small. So even if higher complexity is accepted, it produces unrealistic requirements
and the solution must come from premixing burners, which combine low emissions
with a high enough stability. To that effect, a quality indicator for burners was for-
mulated that divided the equivalence ratio for which the cruise NOx was reached
through the one where the efficiency goal was met. The number is a measure of the
operating bandwidth from which the necessary staging points can be deduced. It is
the quantitative equivalent to the qualitative requirement of a wide U-shaped form
of the combined NOx and CO curves over temperature used to judge the useful
254 Premixed Aero Engine Combustors
5000
4000
3000
Tflame (R)
Pilot only
2000
Fully fueled
Figure 9.9. Staging scheme for jet mix module of the HSCT program (Kastl et al., 2005).
operating range of any lean burner. Table 9.2 gives the ranking. IMFH stands for
integrated mixer flame holder, LDI for lean direct injection. The higher stability range
of flame stabilization by swirling recirculation is coupled to a larger diameter of the
premixer because of the swirlers, which lead to an unacceptable liner height for the
three-dome architecture. Interesting, the lowest bandwidth was noted for an LDI
injector. Observation of the flame in an optical sector showed a quick transition
from premixed to diffusion burning during the enrichment of the mixture. This prob-
lem will be discussed more thoroughly in the section on LDI burners. On the other
hand, the IMFH carried the risk inherent to a perforated plate flameholder.
The IMFH was chosen and sector tested up to 8 bar and 783 K. Tests were per-
formed for two cooling schemes, convectively with impingement cooling and with
Laminalloy, the most advanced effusion cooling technique at the time with minimum
air requirements. For the convective scheme, efficiency and NOx goals were reached.
However, the comparison shows that CO rose by a factor of 2.5 for Laminalloy.
Because the impingement cooling was not sufficient for the real operating condition,
the positive results achieved with the combustor architecture depended on the avail-
ability of the ceramic matrix composites for the combustor walls, which were not
ready at the end of the program and are not today. Although significant advances
have been made, the continuous strive for higher pressure ratios in aero engines con-
tinues to provide a moving target for these materials. During the optimization of the
combustor for the IMFH, attention also turned to the problem of uniform inflow to
the premixer tubes. As the ratio of flow to the combustor head to flow through the
secondary channel is inverted for the lean against the RQL combustor, providing
a uniform flow to the burner is a more difficult task. The architecture selected for
the sector test (Greenfield et al., 2004) was called multistage radial axial and put the
IMFH tubes in line with the split diffuser while the pilot was moved on the outboard
9.3 Development of Lean, Prevaporized Combustors with Premixing Ducts 255
Table 9.2. Comparison of useful operating ranges for premixing modules and LDI injectors
tested by GE in HSCT program (Kastl et al. 2005)
side with a 38 angle. A partially premixing burner, the cyclone swirler of Table 9.2,
was chosen.
To transfer what has been learned about premixed combustion in this program
to subsonic engines, a quick reminder of the specifics of the supersonic cycle is use-
ful. Lower pressure ratio, higher combustor inlet and exit temperatures, as well as
lower variations in stoichiometry throughout the cycle, all are in favor of LPP. With
the smaller pressure ratio comes a longer autoignition time, the higher inlet tem-
perature promotes vaporization, lower AFR variation reduces staging requirements,
and the high exit temperature makes a strong case for LPP. Yet without ceramic
combustor walls, the goals were not reached. The conclusion is that the NOx level of
an EI of five is unrealistic for large aero engines and even more so for high-pressure
ratio cycles of subsonic turbofans. Accordingly, even the most ambitious goal cur-
rently formulated, the 80 percent NOx compared to year 2000 technology goal of
ACARE or the 75 percent CAEP 6 of NASAs ERA program, would result in a
higher NOx emission at takeoff. Combined with the harsher conditions for premix-
ing, full premixing and prevaporization in premixing channels no longer appear an
option for subsonic turbofans. Accordingly, recent development has focused on par-
tial premixing and prevaporizing lean burning combustors, presented in the remain-
der of the chapter. But before that, some aspects of mixing in premixing passages,
common to LPP and LDI combustors, shall be discussed.
9.3.2.1 Prevaporization
Designing premixing channels for complete prevaporization is not possible, nor is
it necessary in view of the autoignition time limit and the results of Section 9.2.2.
But how much residence time is needed and what is the best way to atomize for
quick vaporization? The prediction of vaporization rate is directly coupled to the
prediction of atomization via the cumulative surface of the spray. So far, no reli-
able method to predict atomization is at hand. The air pressure has a beneficial
effect on atomization such that its influence is important for pressure and air blast
256 Premixed Aero Engine Combustors
50
P = 3 bar
P = 6 bar
40
Relative liquid mass flux (%)
P = 9 bar
P = 12 bar
30 P = 14.5 bar
20
10
0
0 20 40 60 80 100 120 140 160
x (mm)
atomization alike, which are the methods most commonly used in aero engines.
Hence, measurements should be carried out at elevated pressure, posing an addi-
tional hurdle.
Prefilming airblast atomizers, plain jets, and pressure-swirl injectors in co- and
cross flow were investigated by Brandt and colleagues (1994) at room temperature
and air pressure up to 5.2 bar for velocities between 60 and 100 m/s. Similarity of
airblast atomization is generally assumed for the same dynamic pressure of the air-
flow; hence, a typical cruise condition with similar air density and velocity was cov-
ered. Good atomization with Sauter diameters below 20 m resulted from almost
all configurations at high pressure and velocity. These velocities are achievable with
the range of liner pressure losses found in aero engine combustors. However, for
prevaporizer-premixer configurations with bluff body flame holders, that pressure
loss is not fully available because the porosity of the flameholder produces addi-
tional pressure loss. For the radially staged combustor of the advanced low emissions
combustor program, only two-thirds of the pressure loss was converted to velocity
at the atomization point; later combustors employed low-porosity flame holders for
better stability with up to 70 percent blockage (Greenfield et al., 2004). For that
case, the air velocity at the atomization point is so low that pressure injection in
cross-stream appears the only method to achieve good atomization.
Experimental investigations of vaporization rates at elevated pressure were car-
ried out by Brandt and colleagues (1997) on a flat prefilming injector and by Becker
and Hassa (2002) on a plain jet in cross flow, both placed into a uniform flow field.
Figures 9.10 and 9.11 show the liquid mass flux measured along the vaporizer chan-
nel for different pressures and temperatures. For the prefilming airblast atomizer,
vaporization is almost complete after 1 ms residence time in the prevaporizer for
a velocity of 120 m/s, pressures above 6 bar, and temperatures above 750 K. For
a cruise condition of the subsonic engine of Table 9.1, 90 percent vaporization is
9.3 Development of Lean, Prevaporized Combustors with Premixing Ducts 257
40 T= 650 K
30
20 T= 750 K
10 T= 850 K
0
0 20 40 60 80 100 120 140 160
x (mm)
Atomizer I, P = 9 bar, w =120 m/s, Ir =125 g/s/m
Figure 9.11. Relative mass flux of kerosene in prevaporizer channel for different air tempera-
tures (Brandt et al., 1997).
attained after 0.5 ms. Hence, it is possible to prevaporize within the autoignition time
calculated to 0.65 ms with Equation 9.1 for the most advanced enginecycles.
For the jet in cross flow configuration, atomization was not quite as good, in part
because the fuel was not heated in the fuel lines before atomization. For elevated
pressures, the fuel heating part of the total energy needed for full vaporization rises
with the rising boiling temperature, and fuel preheat can play an important role. To
investigate the effect, a parameter variation with a jet in cross flow was performed
with a fuel preheater, where fuel and air temperature were the same and air den-
sity and velocity were held constant (Hassa and Wiesmath, 2012). Because of the
lower surface tension, the Sauter diameter drops from 20 to 15 m. This may seem
not much, but according to the drop size correlation for the atomizer, the effect is
equivalent to doubling the pressure loss. The influence of the initial fuel tempera-
ture on vaporization rate was also demonstrated. For a residence time of 0.5 ms at
the investigated condition, the temperature difference of 110C was responsible for
a 25 percent higher vaporization. In the engine, it is difficult to control initial fuel
temperature. The first aim is to avoid coking, and therefore above all the thermal
management has to prevent overheating of the fuel. Hence, some uncertainty on
prevaporization will remain in the design process.
9.3.2.2 Premixing
Having stated that prevaporization is possible, the question remains what degree of
homogenization can be achieved in a premixing channel with the given liner pres-
sure loss without autoignition or flashback. During NASAs advanced low-emission
combustor program, studies for a number of injection and mixing devices were
performed (Dodds and Ekstedt, 1985). The conditions for the premixer foresaw
a pressure loss of 4 percent with a residence time of 2 ms before the bluff body
flameholder. Vortex generators (A), multitube arrangements (B), perforated plates
(C), and annular ducts with multi-tip injectors in cross flow (D) were investigated.
Figures 9.12 and 9.13 show the geometrical concepts. The configurations were tested
in a medium-pressure premixing duct rig at supersonic cruise condition. The ratio
of maximum to mean fuel air, its relative standard deviation, and pressure loss were
monitored. The mean equivalence ratio was 0.6. From Section 9.2.2, it is clear that
258 Premixed Aero Engine Combustors
Figure 9.12. Premixing duct design concepts of GE (Dodds and Ekstedt, 1985).
the mixture should be between 0.4 < < 0.8 to prevent blowout on the lean and too
high temperatures on the rich side, so a max to average well below 1.33 is desired. No
configuration satisfied all conditions. The configurations relying on turbulent mixing
particularly needed more than the available pressure loss to achieve the required
homogeneity. The best configuration was the smooth annular channel with six injec-
tion points per fuel bar. A density of 0.6 injector tips per cm2 was deemed necessary.
With this strategy, the problem of the fuel dispersion was transferred to the distribu-
tion of fuel in the fuel system before injection. With a multitude of injection points,
the fuel flow per injector tip gets very low with the danger of coking. Industrial sup-
pliers solved that problem for a class of matrix burners, described in a later section
(Tacina et al., 2002, 2003, 2005).
As a result from this investigation, it can be stated that sufficient premixing is
not possible with turbulent mixing alone. In addition, the following methods have
been used:
Core flow
Secondary flow
Air
Atomization
occurs here
2. Mixing by cross-stream momentum with jets as seen, for instance, in the radially
staged GE combustor of ECCP Ph. I or pressure-swirl injectors in cross flow, as
in the first Vorbix combustor;
3. Additional turbulence production by swirl-induced shear;
4. Swirl-induced radial droplet slip.
Mixing enhancement with swirlers has been applied in tubular and annular premix-
ing channels. An example of the first is the carburetor tube of the Vorbix combustor
(Zeisser et al., 1982). Figure 9.14 shows a cross section with the necessary elements.
The fuel is supplied in the center of the tube by a Simplex atomizer; the air enters
through a radial swirler. The fuel partially vaporizes and partially hits the wall. There
it is driven to the end of the carburetor tube and atomized from the edge with the
help of a secondary airflow from a concentric axial swirler. On this configuration,
some general problems of liquid premixing can be explained. With a central fuel
injection, the fuel distribution will display a central peak. With swirl or turbulence, the
fuel will move radially, with the outermost parts eventually hitting the wall. Therefore,
the liquid part of the fuel can never display a fully flat profile, because it is either cen-
trally peaked or has a large part of the fuel on the wall and mixing has to compromise
between the two extremes. To prevent flashback in the boundary layer, the wall film
has to be atomized as in the example, but preferentially with a recess of the atomizer
lip to avoid its wake reaching into the combustor, thereby ensuring that the airflow at
the entrance of the combustor is faster than the turbulent flame velocity.
Consequently, ideal premixing is only possible in a sequential way, first with fast
and almost full prevaporization, such that only a negligible part of the liquid reaches
the wall and vaporizes as film before reaching the end of the channel. After the
vaporization, further mixing of the gaseous fuel eventually produces a fully homog-
enized mixture in the second phase of premixing. However, this sequence of events
is normally made impossible by the short autoignition time of kerosene. Luckily,
ideal homogenization is hardly the goal for aero engines, since we have seen that
practical engines with a realistic number of staging points have LBO requirements
that cannot be fulfilled with ideal premixing.
Having identified the problem the wall interaction poses to premixing liquid
fuels, and having reached the conclusion that, even with premixing channels, only
260 Premixed Aero Engine Combustors
Air
Fuel Air
Center body
partial premixing can be achieved, evidently the rest of the homogenization has to
be effected in the combustor on the way to the flame. In this respect, bluff bodies are
not ideal, because the distance to the flame is coupled to the size of the bluff body,
which is naturally smaller than the channel height, and therefore ties the standoff
distance of the flame to the dimension of the channel.
One way to get around the problem is the transition from tubular to annular
premixing ducts with stabilization through swirl-induced recirculation, analogous to
the way aero engines moved from tubular to annular combustors. There are two
advantages, free dispersion of fuel in the circumferential direction and a flame lift-
off, which is geometrically linked to the diameter of the annulus and not to its height.
The latter enables additional homogenization of the near wall fluid of the premixing
channel. Hence, the mixing process in the channel can be stopped at the point where
the walls begin to wet significantly and the premixer can be shortened to gain some
additional security margin against autoignition.
This solution was embodied in a number of lean combustors developed with
premixing lengths varying with the application. Figure 9.15 shows one example
(Bittlinger and Brehm, 1999). The LPP module consists of an annular swirled chan-
nel with jet in cross flow injection and an inner swirled stream to cool the center
body. Flame stabilization is effected by the recirculation behind the center body sup-
ported by the swirl of the annular channel. The other task of the swirl is the circum-
ferential mixing of the fuel. The LPP module was integrated in an axially staged
combustor with a short RQL pilot. As seen from the schematic, the flow accelerates
from the point of injection to the outlet. The design was free of flashback down to 55
percent of the nominal pressure loss at 29 bar and 916 K.
Another method to accelerate mixing is the injection of fuel into a cross flow
with a gaseous carrier jet that can also be used to effect a jet in co-flow atomiza-
tion. Experiments performed at pressures up to 5 bar uncovered different operation
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 261
Separation bubble
modes depending on the pressure loss of the carrier jet, which are characterized by
the scheme of Figure 9.16 (Leong et al., 2001). A low pressure loss of 1 percent did
not result in full atomization, 2 percent gave the required atomization, and 3 percent
resulted in impingement of spray on the tube walls of the jet, with reatomization at
the trailing edge of the jet orifice. Hence, the system can be optimized if the pressure
loss can be tuned.
If the air velocity of the carrier jet is sufficiently higher than the fuel jet velocity,
the momentum-flux ratio of the compound jet in cross flow is nearly independent of
the fuel flow rate, which is synonymous with AFR in terms of the engine cycle. As it
has been shown that jet in cross flow penetration at higher pressures depends only
on momentum-flux ratio, the configuration with the carrier jet would have the con-
siderable advantage of constant fuel placement over the operating range. However,
it turns out that the penetration also depends to some degree on pressure. Higher
pressure influences atomization, and with smaller droplets the lower inertia will lead
to lower penetration. According to Brandt and colleagues (1994), the biggest drop-
lets of jet in co-flow atomization are bigger than for jet in cross flow, possibly because
of less effective secondary atomization. Because the effect of pressure on atomiza-
tion becomes smaller at higher pressure, differences in penetration at higher operat-
ing points will be also smaller, and if the system is only used for main injection, the
independence of operating might be nearly achieved.
Fuel in
Air
Air
Swirler
Fuel holes
Figure 9.17. Schematic of the GE cyclone swirler premixer of the HSCT program (Kastl et
al., 2005).
Andrews and coworkers presented various solutions (Andrews et al., 1990). One
of them, the radial swirler with passage injection, shall be mentioned here because
it is the simplest derivation from a lean direct injector for gaseous fuels and was
indeed derived from an industrial turbine. Some LDI and partially premixing con-
figurations, also from Andrews and colleagues (1990), were tested at elevated pres-
sures by GE in the HSCT program (see Table 9.2). Among those, the cyclone swirler
premixer was selected as pilot for the sector tests. Its design features are shown in
Figure 9.17 (Kastl et al., 2005). The fuel enters through a central rod from where it is
injected by eight fuel jets in the airflow of the inner body. The droplet-laden stream
is directed outward into the swirl chamber by radial holes where it mixes with the
main air entering through radial swirlers. The swirl and the wake of the inner body,
which is air cooled to protect its surface, stabilizes combustion. The burner residence
time in later versions was about 0.26 ms. In principle it is similar to the configuration
of Figure 9.15 in using the jet in a swirled cross flow mixing, but uses radial swirlers
and a shorter premixing annulus. The radial transport of the fuel by air jets simplifies
the design of the fuel path, because a central rod instead of an annulus can be used,
the comparatively wide rod diameter simplifying the thermal management. In its
principal features, it is a predecessor of the cyclone premixer of the GE-TAPS mod-
ule. Although performing well at medium conditions (see Table 9.2), NOx emissions
degraded rapidly for the same flame temperature when the preheat was raised from
727 to 900 K; the best version had an EI of nineteen at 10 bar, 922 K preheat, and
1944 K flame temperature (Greenfield et al., 2005). This behavior points to incom-
plete premixing at the exit of the premixer. Higher preheat causes earlier ignition. It
results in flame stabilization nearer to the exit, with less dilution by premixing and a
richer flame, giving higher local flame temperatures and, in the end, higher NOx. As
mentioned earlier, in the extreme this can lead to a continuous transition to a diffu-
sion flame. The dependence of the degree of premixing on the operating conditions
is a common problem for all partially premixing lean burners and will be the focus
of later case studies.
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 263
Fuel
Air
Figure 9.18. Fuel and airflow schematic of the multipoint integrated LDI module of Tacina
and colleagues (2002).
A way to diminish this problem is the minimization of the size of LDI burners. The
fuel is distributed to as many points as possible already in the fuel path. Figure 9.18
provides an example of this method. The remaining spatial dispersion of the fuel
necessary to have a sufficiently lean mixture is much smaller. If it is possible to build
a separate stabilization device for each injection point to yield a separate flame, the
variation of premixedness over the operating envelope will be smaller, because the
progress of mixing and, hence, also the distance to a possible stabilization point, also
scales inversely with the size of the stabilization device. NASA achieved such solu-
tions to yield matrix combustors composed of small LDI burners (Tacina et al., 2002,
2003, 2005).
The multipoint integrated module described by Tacina and colleagues (2002)
has thirty-six fuel injectors. The fuel-injector system consists of plates that contain
the fuel manifold to distribute the fuel to each injector, the simplex fuel injectors,
and a plate with an air gap that provides thermal protection to the fuel. The plates
that make up the fuel-injector assembly are diffusion bonded to form an integral
assembly. An air-swirler assembly of diffusion-bonded plates, which have the radial
air-swirler geometry chemically etched into them, follows the fuel-injector assem-
bly downstream. The schematic drawing in Figure 9.18 visualizes fuel and airflow
as well as the intended location of the flame. Note that again, there still is a volume
for premixing and prevaporization with nonzero residence time in the fuel-injector
assembly.
Low emissions were achieved in an uncooled ceramic flame tube at high pressure.
For 28 bar and 810 K preheat, an EI NOx of six was achieved at a flame temperature
of 1900 K. Varying the preheat between 590 K and 810 K with the same flame tem-
perature resulted in a NOx increase from EI four to eight for a twenty-five-injector
264 Premixed Aero Engine Combustors
100
EINOx (g-NOx/kg-fuel)
10
1
1400 1600 1800 2000 2200
Flame temperature (K)
Figure 9.19. NOx emission for LDI configurations of Tacina and colleagues (2002) for 28 bar,
810 K preheat, and 4 percent pressure loss.
configuration. That is still an appreciable difference, but much less than the fivefold
increase observed for the cyclone premixer with a smaller temperature variation.
Figure 9.19 shows the NOx emission over flame temperature for the different con-
figurations at the benchmark condition. All curves have a much lower slope than
the assembly of fully premixed flame tube data from earlier NASA-funded research
(Tacina, 1990), which, according to Figure 9.3, indicates a lower degree of premixing.
Although considerable differences arise in absolute values, the fully premixed data
agree quite well on the slope of the temperature dependence for a given residence
time, yielding about an order of magnitude NOx increase between 1800 and 2200 K
flame temperature. The differences in the slopes in Figure 9.19 suggest that the
higher number of injectors and the higher swirl had a better degree of premixedness.
However, the decreasing emissions while increasing pressure loss from 3 to 5 percent
indicates that, although the combustor cross section might be used in the best possi-
ble way for mixing, it is still far from a homogeneous reactor for the range of allow-
able pressure losses. Another task that cannot be fully completed with minimized
burners is prevaporization. Atomization quality does not depend on burner size. On
the contrary, it will be more difficult to arrange for high relative velocities in the
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 265
atomization region, such that vaporization time or length will be, at best, constant
with burner size. With the flame front getting nearer to the burner face with a shrink-
ing burner, the amount of nonvaporized fuel will rise with the connected potential of
partial diffusion burning as described in Section 9.2.2.
Operability aspects were also tested with a sector combustor at pressures
up to 48 bar (Tacina et al., 2004). Twenty-one percent of the air was used for the
effusion-cooled liners. With the outer burner rather near the walls, the highest tem-
perature was restricted to 1900 K. The injector orifice was enlarged from 0.25 to
0.51 mm to prevent fuel clogging. High-power EI NOx increased by a factor of 2.5 at
1800 K. Part load operation was tested with fuel grouped and partially switched off
(Tacina et al., 2004). Since the burners are densely packed, the mixing with unfueled
burners at the boundaries of fueled sectors dilutes the flow, such that suppression of
UHC or combustion efficiency becomes harder to reach.
The work performed up to this point provided a good picture of the poten-
tial of minimized LDI combustors. For large engines, for example the engine E of
Table 9.1, the takeoff condition with 21 percent cooling, as realized for the sector
tests, would result in 2024 K flame zone temperature, yielding an attractive NOx
reduction if operability issues could be resolved. Since miniaturized burners also
considerably shrink the time and, consequently, the space needed for the first part
of combustion up to the CO reduction step, the liner could be reduced in length
accordingly. In comparison to the sector tests, the higher temperature might then
be supported with the same amount of cooling air. The approach remains attractive
because it opens the potential of a very compact low-emission combustor, if an ele-
gant solution for part load operation can be found.
The preceding sections demonstrated that for high-pressure ratio engines, full pre-
mixing is not possible, but pure lean direct injection with minimal premixing has
some unresolved operability issues. So more partial premixing is needed. While this
prescription is vague, it is possible to name some necessary elements of the solution.
Since the additional partial premixing has to happen in the combustion chamber
itself, lifted flames are necessary. But the operation of a lifted flame in an aero engine
is an ambitious proposition, because a flame that is not anchored to a specific point
might blow off more easily, especially during transient operation or when subjected
to pressure oscillations. Therefore, piloting is necessary. In addition, the lift-off height
of lifted flames and, with it, the degree of partial premixing varies with operating
conditions such that means to influence it are desirable. Hence, generic information
on factors influencing lifted flames and piloting of lifted flames is given before the
most recent partial premixers are discussed.
10
6
Z /R 0
1 .4 4
1 .2
1 .0
0 .8
0 .6 2
0 .4
0 .2
4 2 0 2 4
X /R 0
Figure 9.20. Flame stabilization of lifted flame: equivalence ratio in false colors, vector plot of
UV velocity, and position of flame front with blue hatched line (Fokaides et al., 2008).
off, when the exit velocity exceeds the burning velocity, the attached flame stabi-
lized by a recirculation at a point near the zero streamline will often not exhibit a
marked lift-off regime before blowoff (see the discussion of the difference in blowoff
behavior between jet and bluff body stabilized flames in Shanbhogue et al. 2009). If
the mixture is within the flammability limits near the recirculation, burning velocity
does not play a big role because local turbulence alone governs the local velocity
near the recirculation boundary. To create a lifted flame, the fuel placement must
be shifted away from the recirculation such that the flammable mixture at the zero
streamline is reached more downstream. Hot gas with products is entrained from the
recirculation or the flame into the core flow and the equilibrium of flow and burning
velocity is reached at a later point. Figure 9.20 depicts an example of such a situation
(Fokaides et al., 2008). Here stabilization is through the outer recirculation, and the
color plot of the equivalence ratio gives clear evidence of lean burning.
Having established a lifted flame, the factors known to decrease stability will
in general also lead to higher lift-off. Obviously, this is the case for a leaner global
equivalence ratio. Higher preheat temperature has the intuitive effect of quicker
stabilization via its acceleration of the laminar burning velocity, Sl ~ T2 (Lefebvre,
1998). Because the mixture will be richer there, the burning velocity is additionally
increased. In the extreme case, the effect can lead to a stabilization where the mix-
ture has reached stoichiometry and burning in a diffusion mode will ensue. At con-
stant pressure drop, the flow velocity at the nozzle outlet will increase with a square
root dependence to preheat temperature and the former effect will be diminished.
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 267
1500 K 2500 K
Figure 9.21. OH-LIF-based single-shot temperature at 6 bars and 700 K at AFR = 0.92
AFRdesign (Hassa et al., 2005).
20% pilot
30% pilot
50% pilot
0 [a.u.] 1
Figure 9.22. Mean kerosene-LIF and deconvoluted OH* distributions at 6 bars and 700 K,
AFR = 0.92 AFRdesign at varying degrees of pilot fuel injection (Hassa et al., 2005).
burner. Adding just 10 percent more pilot fuel causes the pilot heat release to change
the flow field considerably by widening the fuel cone angle and the heat release zone
with it. The main fuel still burns lifted and the flame is shorter. This is coincident with
the rapid reduction of CO emissions to a level that does not significantly change for
higher pilot fuel. With an even split of pilot and main fuel, rapid heat release begins
right at the burner outlet, the pilot fuel burning in diffusion mode.
The comparison for different piloting shows how small the optimum band-
width for pilot fuel addition can be. A straightforward conclusion is that the
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 269
amount of pilot fuel has to be varied continuously with the operating condition,
which demands a flexible control system. However, since the position of this
band changes considerably with operating conditions and is further influenced
by other aspects, like cooling and the shape of the combustor liner, even knowl-
edge such as from Figure 9.22 will not be enough to predict optimum pilot fuel
addition. Designing burners for operation with a much smaller pilot fuel frac-
tion as in some industrial gas turbines, where a pilot burning in diffusion mode
does not entirely spoil the NOx emission, is not an option for aero engines, as the
pilot has to cover a considerable part of the load spectrum in pilot-only mode.
For these reasons, the additional complexity and cost of a set of swirlers for the
pilot is outweighed by the potential to achieve high stability for the whole pilot
operating range and more robust flame holding of the main. To some degree, the
aerodynamics then can be optimized separately for the different load regimes and
the sensitivity of the main flame lift-off toward the amount of pilot fuel can be
reduced. Thus some designs with internal pilots will be presented in the following
section.
Compared to the supersonic engines, for a long time the driver of premixed combus-
tion, the rationale for market introduction is different, as there is no stratospheric
cruise emission restriction. In subsonic engines, the advantage of lower emissions
in comparison to the competing RQL technology must be weighted against pos-
sible higher complexity. GE has developed double dome combustors that made it
into engines, but the experience of the CFM 56, where the airlines had a choice,
suggests that more than two domes will not be accepted and a low-emission archi-
tecture with a single dome will be a definite advantage. Also, it seems unlikely
that ceramics will enter the combustor with a sudden step change of the cooling
requirements. However, with higher pressure ratios, economizing on cooling air
is made difficult for two reasons: the temperature difference between exit and
allowable liner temperature rises, and the driving temperature difference for cool-
ing between liner and combustor inlet shrinks. GE acknowledged that the lean,
double dome GE 90 DAC 1 combustor had higher emissions than the rich DAC 2
(Mongia, 2003) because cooling enforced a primary zone equivalence ratio of 0.9
at takeoff. Here single annular architectures (SAC) offer the principal advantage
of lower surface. It is therefore not surprising that the current efforts to realize pre-
mixed combustion all try to realize internally piloted burners in a single annular
combustor.
As all the internally staged premixed combustors remain under development,
quantitative data are not available and published information is scarce. Therefore,
the following section will be more qualitative than before. The boundary conditions
of subsonic lean combustors lead to a certain convergence of designs. Three of them
are presented in an exemplary manner before the operability issues of SACs with
internally piloted burners are discussed.
270 Premixed Aero Engine Combustors
Main mixer
Main flame
68.5 mm
Figure 9.24. Isothermal flow of burner D in single-sector combustor (Yamamoto et al., 2009).
Rich-lean
combustor
7% MTO
30% MTO
85% MTO
100% MTO
Staging
fuel nozzle
0 10 20 30 40 50
Nox emission in each LTO condition against CAEP4 (%)
Figure 9.25. Comparison of TechCLEAN RQL and premixed combustor (Yamamoto et al.,
2009).
of all modern designs. Common to these designs is the need to protect the separating
geometry, in this example with a bleed airflow from the outer pilot (Figure 9.23).
Earlier in this chapter, it was said that the two domes of a double dome combustor
are too far away to enable efficient piloting; now the message is that efforts have to
be undertaken to separate the zones in internally piloted systems.
The large swirlers of the main are co-rotating, which results in the large and
stable recirculation exhibited in Figure 9.24, needed to ensure good combustion effi-
ciency at part load (Yamamoto et al., 2009). Because of the air split between pilot
and main, the pilot nozzle runs rich in the pilot-only mode of idle and approach,
such that the main recirculation needs to support the burnout of the pilot fuel. To
preserve the better mixing of counter-rotating swirl in the premixing passage, a third
swirler in the middle was added with opposite swirl direction to the large swirlers.
For the 17 bar cycle, because of the rather low combustor exit temperature of the
small engine at climb out, circumferential staging of main injectors had to be used to
ensure efficiency. For the high-pressure cycle, staging at climb out wasnt necessary, such
that the resulting ICAO NOx parameter was lower than for the low-pressure cycle.
Figure 9.25 (Yamamoto et al., 2009) compares the contribution of each condi-
tion in the ICAO landing and takeoff cycle for the RQL and premixed combustor.
272 Premixed Aero Engine Combustors
Cyclone
Pilot
Pilot/cyclone
interaction zone
Emission during takeoff and climb out are significantly reduced; however, for the
two pilot-only conditions, approach and idle, the NOx emissions increase. Idle is the
largest contributor to the aggregate emission parameter. In this condition with a
rich pilot, the combustor is a radially staged rich-lean combustor without a zone
explicitly designed for quick quench. One can see with this example that the change
of combustion system can also bring about a different weighting of the two environ-
mental concerns: local air quality around airports and atmospheric impact.
Pilot flame
Main fuel
PRZ
Pilot air
LRZ
Pilot fuel
Figure 9.27. Schematic drawing of TAPS and main flow features (Dhanuka et al., 2009).
Figure 9.28. CFD RANS temperature contours at high power (Foust et al., 2012).
main, as it is believed to store radicals from the pilot combustion. In comparison to the
TechCLEAN flow field in Figure 9.24, the LRZ recirculation is rather small. This has
some relevance for the larger GenX combustor with higher exit temperatures, as the
small recirculation rate reduces the residence time of the main flow. A CFD represen-
tation of the temperature field of the TAPS combustor, Figure 9.28 (Foust et al., 2012),
displays a lifted main heat release zone, which will have a long premixing time for the
related operating condition of the combustor. Apart from the burner design, GE notes
that standard design rules for the combustor dimensioning were used (Mongia, 2007),
that is, dome and reference velocities, cold and hot residence times, fuel nozzle spacing,
dome height and length, as well as combustor loading.
The combustor for the GenX seen on its Web site seems fairly convention-
ally cooled. A sketch from a patent (EP 1445 540 A1, 2004), Figure 9.29, dis-
plays the same general cooling arrangement with lipped rings. It also features an
impingement-cooled heat shield. Emission measurements for the GEnx-1B64 at the
design pressure ratio of 42.7 gave an ICAO NOx parameter 50 percent lower than
CAEP 6 (Maurice et al., 2009). The benchmark value for the best RQL combustor for
this pressure ratio is 35 percent lower than CAEP 6. So far the reduction is coupled
to higher CO emissions. However, the NOx CO trade-off seems to be the same as
for advanced RQL configurations, as exemplified in a comparison for CFM 56 SAC
and TAPS tests in Figure 9.30 (Mongia, 2007). TAPS was superior to DAC, which
274 Premixed Aero Engine Combustors
Figure 9.29. Scheme of the TAPS combustor (EP. 1445 540 A1, 2004).
60
55 CFM56-5B/P
CFM56-7B
50 CFM TAPS
LTO NOx (g/kN)
45
40
35
30
25
20
20 30 40 50 60 70 80 90
LTO CO (g/kN)
Figure 9.30. Comparison of ICAO LTO NOx and CO for two SAC-RQL and the CFM 56
TAPS combustor (Mongia, 2007).
carries the load of a larger combustor surface. The pilot AFR should govern the soot
emission of a lean combustor. A flexible control system allows reducing the pilot
fuel exactly at the higher operating conditions, where soot production in RQL com-
bustors is the highest. Comparing LTO smoke of CFM 56 SAC, DAC, and TAPS in
Figure 9.31 supports this assumption. A 90 percent margin to CAEP 6 is the goal for
TAPS II (Mongia, 2007).
Staging is qualitatively shown in Figure 9.32 for the GenX TAPS (Maurice et al.,
2009). Having noticed the low LTO NOx values, it comes as a surprise that the graph
actually suggests the takeoff NOx to be higher than in the RQL reference engine.
The choice or necessity of zonal AFRs obviously leads to a comparatively high main
zone equivalence ratio. The span of the cruise condition during an exemplary flight
mission is also displayed. Because of the burned fuel, the airplane is getting lighter
and less fuel burn is necessary, such that at some point, staging may become nec-
essary to maintain efficiency, which enforces a jump in the NOx emission. Here, in
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 275
15
CFM56-5B/P
CFM56-7B
Max smoke No.
CFM56-5B/PDAC
10 CFM56-7BDAC
CFM TAPS
0
20 25 30 35
Engine pressure ratio
Figure 9.31. Comparison of ICAO LTO smoke for CFM 56 SAC, DAC and TAPS combustor
(Mongia, 2007).
Takeoff
Current RQL Fueling mode Climb
GEnx TAPS switch-over point
EINOx(gNOx/lbFuel)
Pilot +
Main mode
Cruise
Pilot only mode
Approach
Taxi
return for the high takeoff temperatures, stability and efficiency at lower operating
points are improved, such that the largest part of cruise can be accomplished with
fueled mains. Furthermore, the diagram has only one staging point. Since the ICAO
cycle has no cruise, some of its relevance stems from the assumption that cruise
emissions scale with takeoff, and for RQL this might be true. As we see, this is not
necessarily true for premixing combustors. Depending on the AFR split at the sta-
ging point, it can be difficult to suppress excessive soot formation around the richest
pilot condition. The air distribution, which in RQL combustors is tailored to allow
quick soot oxidation in the intermediate zone, can make it difficult to do that once
the two air streams mix, because the mean temperature might be too low.
Figure 9.33. CFD-Research burner cross section (U.S. Patent No. 6,272,840, 2001).
swirlers with a simplex pilot atomizer. Air-blasted pilots with three swirlers were
also investigated. As with the previous specimen of piloted burners, it separates
the pilot and main flow field, here by a splitter, in Figure 9.33 (U.S. Patent No.
6,272,840, 2001). The downstream face (Figure 9.33) of the splitter sets up a wake,
which separates pilot and main. The result is called a bifurcated flow field leading
to the separated pilot and main flame. Their interaction is further influenced by
the converging and diverging passage of the outer main swirler, called flare. Its
outlet angle defines the direction of the main flow near the nozzle and prevents
attachment of the flow to the combustor head. Fuel injection to the main airflow
is accomplished with a prefilming airblast. The prefilmer lip, as well as the head
of the simplex, have a short recess from the combustor head, leading to some
premixing within the injector. However, the prefilmer lip is located almost at the
smallest diameter of the flare.
The development since then produced variants with different flow characteristics
(Lazik et al., 2008). Figure 9.34 shows photographs of three of them, C-A with pres-
sure atomizer, C-F with air-blasted pilot and V-shroud, and C-G with simplex and
V-shroud. The V-shroud is a derivative of the V gutter flame holder with different
9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors 277
Ptot comb.
Figure 9.35. Iso-thermal CFD prediction of velocity and pressure distribution (Lazik et al., 2008).
lengths and angles of the two surfaces optimized for the swirler combination and
allows for an even stronger separation of the two flows. Figure 9.35 displays the iso-
thermal flow and the pressure distribution in the combustor, illustrating the bifur-
cation of pilot and main. The pilot flow does not exhibit an inner recirculation and
the main flow follows a wide cone angle, leading to rapid dilution of the spray. The
temperature field of the reacting flow resulting from the calculation of a similar var-
iant is depicted in Figure 9.36. Here, the lifted flame seems to stabilize in the outer
recirculation whereas the pilot flame is restricted to the central area. It can also be
seen that the combustor liner uses effusion-cooled tiles. For the air split, a bandwidth
of 6070 percent is given for the burners and consequently 3040 percent for cooling
(Lazik et al., 2007).
The staging strategy consists of two steps with pilot only, pilot, and part of the
mains for approach and a second staging at mid-power before cruise with all mains.
The staging points are during transients, such that potential peaks of smoke and CO
are insignificant for the mission. A principal efficiency deficit at the staging points,
compared to conventional RQL combustion technology, remains because of the
nature of fuel staging and lean operation. The fuel repartition is accomplished with
individual fuel manifolds supplying each main injector line and the pilot injectors,
and a splitting device to the fuel-injector groups. Continuous variation of injector
fuel splits is assured by electronic engine control software (Lazik et al., 2008).
The Rolls Royce combustor has been reported to demonstrate a NOx reduction
of 60 percent against the CAEP 6 LTO at TRL 6 for an engine with a pressure ratio
278 Premixed Aero Engine Combustors
Figure 9.36. RRD E3E Core single annular combustor configuration with calculated temper-
ature field (Lazik et al., 2008).
of 39 (Maurice et al., 2009). As already stated by Rhode (2002), this reduction might
be smaller once TRL 9 is reached.
Although operability is more important than emissions, the latter drives premixed
combustor development, and operability issues are only worth solving if the emis-
sion advantage is preserved. Hence, operability is treated toward the back end of the
chapter. In addition, the discussion is restricted to internally staged single annular
combustors because the other specimen did not get far enough to warrant a solution
to all operability aspects. Having been discussed in more general terms in Chapter 1,
the description will restrict itself to the differences of premixing combustors to those
in active duty where all those problems are solved. It will start again with part load
emissions and then take a tour around the combustor.
Figure 9.37. Temperature prediction for pilot-only approach with decentral left and central
right pilot recirculation (Lazik et al., 2008).
to operate on the same part load AFRs. This can be done to some extent only as the
combustor is missing the quenching porting. The quenching has to be accomplished
with the excessive air of the nonoperative main stage of the injector. However, there
are conflicting requirements. Early on, it was clear that the pilot can be responsible
for most of the high power NOx so it should be small. Depending on the staging
scheme, approach is part of the pilot-only operating range and soot can become
an issue if this range is too small, that is, if the pilot stage operates too rich. Trading
NOx too heavily against efficiency can be a problem for the LTO NOx because of the
long idle interval in the cycle. As both global and local emissions impacts have to
be reduced, the emissions within the ICAO cycle as well as cruise emissions have to
taken into consideration for future low-emission combustion technology.
Lazik and colleagues (2008) give an example of such a trade-off in terms of the
aerodynamic solution for approach. Figure 9.37 shows a comparison of the temper-
ature fields for a pilot-only approach condition with outer and inner recirculation
zones. The outer recirculation has the bifurcated flow field, giving lower high-power
NOx, but does not mix the over stoichiometric part of the fuel well into the main air.
Therefore, the mixture remains rich, forming a rather hot core and high NOx.
Stability at idle is not the only requirement. Transient stability during slam decel-
eration must also be fulfilled. For premixing combustors, this obviously implies fuel
staging. In addition, hail and rain ingestion during flight require a good stability mar-
gin for weak extinction. For RQL systems, there will be a stoichiometric zone during
the high-power operation and some margin in descent and approach. For premixed
combustors with lean burning mains, this condition can set a boundary on the stoichi-
ometry of main and pilot as well. Generally, the main stability will always be worse
than the primary zone of a RQL combustor, hence during climb it will not be tolera-
ble to run the pilot leaner than the main to provide the required stability margin.
9.4.4.3 Ignition
It has already been noted that the internally piloted systems have a disadvantage
with conventional spark plug igniters because they have to achieve ignition across an
unfueled main zone airflow. Yet sufficient ignition loops for high altitude relight up
to thirty thousand feet have been achieved (Lazik et al., 2007). To better understand
the differences relative to RQL systems, it is useful to discuss them along the differ-
ent phases of the ignition process (see Lefebvre, 1998):
velocity is considerably smaller than the forward flow velocities. Having filled the
recirculation zone, in the case of the flow in Figure 9.27, enough energy needs to pen-
etrate the mixing layer to ignite the spray issuing from the pilot. Again, large-scale
mixing helps to do that. From here on, the process is the same as with unstaged burn-
ers; enough recirculation is necessary to establish self-supported combustion in the
pilot zone. As with LBO, it is obvious that with largely different recirculation zone
sizes and rates, ignition loops for the combustors will not be the same as for RQL
combustors.
The light around again must happen with the handicap; that energy has to pen-
etrate to the neighboring pilot burners through the cold main flow. Obviously, this
can lead to design constraints on the distance of the burners. According to Mongia
(2007), however, it is sufficient to adhere to the known design rules on burner spac-
ing to satisfy the issue.
For the development of the aerodynamics of internally staged premixing com-
bustors, many aspects have conflicting tendencies requiring very careful balancing
of the flow features. To satisfy those conflicting demands, modern premixers have a
much higher degree of sophistication and complexity than, for instance, the staged
combustors of the ECCP program (Roberts et al., 1975). It is, therefore, understand-
able that all earlier developments tried to separate the critical function of staging
into separate zones. The ability to mature internally staged premixers to airworthi-
ness is also evidence for the coming of age of design methods and a more rational
combustor design process.
1. Reduction of fuel flow passage area to decrease hot section residence time
2. Insulating air gaps
3. Use insulating materials like ceramics
4. Avoidance of flow separation in the fuel path
For staged burners, this is not sufficient for two reasons: the first problem burners
with fuel staging face is the fuel standing in the fuel line that has been shut down and
that subsequently has to be purged. However, purging with air at compressor exit
temperature above idle will cause deposits. Second, the prevention of fuel overheat-
ing during operation is exacerbated because the correlation of high thermal loads
with high fuel flow rates is lost, as we have learned that a flexible degree of pilot and
main flow is one of the prerequisites of a successful staging scheme. A number of
ingenious solutions have been proposed, therefore only an example shall be quickly
282 Premixed Aero Engine Combustors
Mains air
Mains fuel
Pilot fuel
Cooled cooling-air
Figure 9.38. Detail of fuel nozzle assembly with pilot and main fuel lines, purge airflow, and
fuel well to the swirl channel (EP 1445 540 A1, 2004).
provided to give an idea of the complexity required to fulfill the cooling requirement.
In EP 1445 540 A1 (2004), a method pointed to the TAPS combustor is described.
Pilot and main flow is transported in a strip consisting of two parts, with grooves
fused together by diffusion bonding (Figure 9.38). The grooves form channels for
pilot and main fuel or air used for purging. Because of the constant pilot fuel flow,
the other channels are cooled. The main fuel issues from the jet orifice and is guided
through the heat shields by a cylindrical insert, which is flushed by the cooled cooling
air. It is clear that the two-phase flow developing from the orifice will be influenced
by the surroundings largely imposed by the necessities of the thermal management.
emission. Active control has been investigated for premixed aero engine combustors
too (Umeh et al., 2007); however, no information is available on implementation.
Additional damping is usually provided in the form of Helmholtz resonators. Apart
from their narrow damping bandwidth, their weight and size requirement makes
their use in aero engine combustors almost prohibitive. Wideband damping can be
achieved by perforated liners (Maquisten et al., 2006), which will necessitate some
extra cooling air.
9.4.4.9 Outlook
Having satisfied all operability aspects, the future will show how well the premixing
combustors will fare on the market. Besides their advantage on emissions and of
course their future track record on dependability, this will be decided by cost, weight,
and efficiency, the two last items again entering into cost of ownership and by the
ease with which the combustors can be scaled for smaller engines.
Premixing combustors face a disadvantage on cost because of the more complex
fuel handling and control systems, bigger and more complex burners, and possibly
9.5 Summary 285
more complex cooling schemes. The additional components of the control system as
well as the bigger burners and possibly double-walled liners also result in additional
weight. Efficiency at cruise seems to remain a difficult subject. Some small difference
to the unstaged combustors seems tolerable, but there is an estimate (Maurice et al.,
2009) that additional inefficiencies above 0.5 percent would be a concern to airlines.
On the other hand, with a more homogeneous temperature distribution at the outlet,
an efficiency gain could be realized in the turbine.
The two combustors having progressed the farthest in TRL scale are in large
engines for long-range transport, where it makes sense for several reasons. Although
a smaller segment of the market by number, long-range transport has a dominant
share of the CO2 emitted in the atmosphere. Weight and size penalties of the fuel
and control system, which does not scale in the same way as the engine, are rela-
tively smaller, and the NOx advantage of premixed combustion is the highest with
high-pressure ratios (see Figure 9.2). The last point might be the most important
driver to bring the premixing technologies to smaller engines because the aim for
higher efficiency will also lead to higher pressure ratios for mid-sized engines of
the future. Scaling a premixing combustor to a smaller size and lower OPR is not
trivial; however, the CFM TAPS demonstrator showed the same NOx reduction at
OPR 29 in terms of CAEP 6 margin as the bigger GenX engine, although on lower
TRL. Scaling to smaller size, the physical aspects of the two-phase flow, fuel place-
ment, evaporation times, and eventually the heat release remain unchanged; hence
pure geometrical scaling of burners will fail. Scaling to smaller pressure ratios, the
staging changes as the temperature limits for stable premixed combustion stay the
same. Referring to Table 9.1, the combustor exit temperature of the generic engine
at cruise is just above the minimum temperature for stable fully premixed combus-
tion with kerosene, for lower temperatures it is not. Although the actual combustion
temperatures are higher, the efficiency-NOx trade-off will be different. In summary,
scaling a combustor is always a test for the design methods and all the more so for
premixing combustors.
9.5Summary
This chapter reviewed partially premixed and premixed aero engine combustors.
Apart from a description of the state of the art, it tried to explain how the design
constraints of airborne combustors enforced the breed of premixing combustors
that will soon enter service and why the NOx levels emitted from airplanes will con-
tinue to be so different from those of industrial gas turbines. The efficiency required
from aero engines of long-range aircraft forces exit temperatures that cannot be sus-
tained by pure convective cooling with the airworthy materials available today. The
airflow for combustion thus being reduced by the necessary film cooling air results
in combustion temperatures above the onset of thermal NOx production, which
must be limited by a short residence time. As premixing results in poor stability,
enforcing piloted combustion, the trade-off between NOx emission at full power and
combustion efficiency at cruise defines combustor size and premixed combustion
286 Premixed Aero Engine Combustors
References
Altemark, D., and Knauber, R. (1987). Ergebnisse von Untersuchungen an einem
Vormischbrenner unter Druck mit extrem niedriger NOx Emission. VDI-Berichte 645:
299311.
Anderson, D. (1975). Effects of Equivalence Ratio and Dwell Time on Exhaust Emissions
from an Experimental Premixing Prevaporizing Burner. ASME-75-GT-69.
Andrews, G. E., Abdul Aziz, M. M., Al Dabbagh, N. A., Ahmad, N. A., Al Shaikhly, A. F., Al
Kabie, H. S., and Kowkabi, M. (1990). Low NOx Combustor Designs without Premixing
for Aeroengine Applications. Proc. European Propulsion Forum: Future Civil Engines
and the protection of the Atmosphere, DGLR, AAAF, RAeS, Paper 90020, 16174.
Bahr, D. W., and Gleason, C. C. (1975). Experimental Clean Combustor Program, Phase I
Final Report GE. NASA-CR-134737.
Beck, C. H., Koch, R., and Bauer, H.-J. (2007).Investigation of the Effect of Incomplete Droplet
Prevaporization on NOx Emissions in LDI Combustion Systems. ASME-GT-27654.
Becker, J., and Hassa, C. (2002). Breakup and Atomization of a Kerosene Jet in a Crossflow
at Elevated Pressure. Atomization and Sprays 12: 4967.
Bittlinger, G., and Brehm, N. (1999). High Pressure Combustion Test of Lean Premixed
Prevaporized (LPP) Modules in an Axially Staged Combustor using a Multisector Rig.
ISABE 997008.
Brandt, M., Gugel, K. O., and Hassa, C. (1997). Experimental Investigation of the Liquid Fuel
Evaporation in a Premix Duct for Lean Premixed and Prevaporized Combustion. J. of
Eng. For Gas Turbines and Power 119: 81521.
Brandt, M., Hassa, C., Kallergis, K., and Eickhoff, H. (1994). An Experimental Study of Fuel
Injectors for Premixing Ducts. ICLASS-94, Paper VII-9, 7507.
Brundish, K. D., Miller, M. N., Morgan, L. C., and Wheatley, A. J. (2003). Variable Fuel
Placement Injector Development. ASME GT-200338417.
Bulzan, D., Anderson, B., Wey, Ch., Howard, R., Winstead, E., Beyersdorf, A., Corporan, E.
et al. (2010). Gaseous and Particulate Emissions Results of the NASA Alternative
Aviation Fuel Experiment (AFEX). ASME GT201023524.
Chen, J., McDonnell, V., and Samuelsen, S. (2004). Effects of Ethane and Propane Additives
on the Autoignition Behaviour of Natural Gas Fuels. Spring Meeting of the Western
States Section of the Combustion Institute, 04s-24.
Cooper, L. P. (1979). Effect of Degree of Fuel Vaporization upon Emissions for a Premixed
Prevaporized Combustion System. AIAA-791320.
References 287
Correa, S. M. (1992). A Review of NOx Formation under Gas Turbine Conditions. Combust.
Sci. and Tech. 87: 32962.
Croft, J. (2010). Engine Buzz: Blisks, Sabers and Safran. Flight International (October
26November 1): 22.
Dhanuka, S. K., Temme, J. E., Discroll, J. F., and Mongia, H. (2009). Vortex Shedding and
Mixing Layer Effects on Periodic Flashback in a Lean Premixed Prevaporized Gas
Turbine Combustor. Proceedings of the Combustion Institute 32: 29018.
Dodds, W. J. (2005). Twin Annular Premixing Swirler (TAPS) Combustor. The Roaring 20th
Aviation Noise & Air Quality Symposium.
Dodds, W. J., and Ekstedt, E. E. (1985). Evaluation of Fuel Preparation Systems for Lean
Premixing Prevaporizing Combustors. ASME 85-GT-137.
Eichler, C., and Sattelmayer, T. Experiments on Flame Flashback in a Quasi-2D Turbulent
Wall Boundary Layer for Premixed Methane-hydrogen-air Mixtures. Journal of
Engineering for Gas Turbines and Power 133: 011503.
Ekstedt, E. E., and Fear, J. S. (1987). Advanced Low Emissions Combustor Program.
AIAA-872035.
Environmental Protection Agency, Title 40, Protection of environment; Part 871-Aircraft and
aircraft engines Proposed standards for control of air pollution. Fed. Regist. 37, 26488
26503, 1972.
European Patent (2004). 1445 540 A1
Fiorentino, A. J., Greene, W., Kim, J. C., and Mularz, E. J. (1980). Variable Geometry, Lean,
Premixed, Prevaporized Fuel Combustor Conceptual Design Study. ASME 80-GT-16.
Fokaides, P. A., Kasabov, P., and Zarzalis, N. (2008). Experimental Investigation of the
Stability Mechanism and Emissions of a Lifted Swirl Nonpremixed Flame. J. Eng. for
Gas Turbines and Power 130.
Foust, M. J., Thomsen, D., Stickles, R., Cooper, C., and Dodds, W. (2012). Development of the
GE Aviation Low Emissions TAPS Combustor for Next Generation Aircraft Engines.
AIAA 20120936.
Freitag, S., Meier, U., Heinze, J., Behrendt, T., and Hassa, C. (2010). Measurement of Initial
Conditions of a Kerosene Spray from a Generic Aeroengine Injector at Elevated
Pressure. ILASS-Europe.
Fritz, J., Krner, M., and Sattelmayer, T. (2004). Flashback in a Swirl Burner with Cylindrical
Premixing Zone. Journal of Engineering for Gas Turbines and Power 126(2): 27683.
Futamura, H., Fukuyama, Y., Nozaki, O., and Hayashi, S. (2009). TechCLEAN Project: A
Technology Development Program for Environmental Aspects of Small Passenger
Aircraft. ISABE-20091196.
Greenfield, S. C., Herberling, P. V., Kastl, J., Matulaitis, J., and Huff, C. (2004). HSCT Sector
Combustor Evaluations for Demonstration Engine. NASA/CR-2004213132.
Greenfield, S. C., Heberling, P. V., and Moertle, G. E. (2005). HSCT Sector Combustor
Hardware Modifications for Improved Combustor Design. NASA/CR 2005213322.
Guin, C. (1999). Characterisation of Autoignition and Flashback in Premixed Injection
Systems in Gas Turbine Engine Combustion, Emissions and Alternative Fuels. RTO-MP
14, AVT Symp On Gas Turbine Engine Combustion, Emission and Alternative Fuels,
paper 30.
Hassa, C., Heinze, J., Rackwitz, L., and Drr, Th. (2005). Validation Methodology for the
Development of Low Emission Fuel Injectors for Aero-engines. ISABE-20051143.
Hassa, C., and Wiesmath, P.-F. (2012). The Effect of Initial Fuel Temperature on Vaporization
in Aeroengine Combustors with Prevaporization. Twelfth Triennial International
Conference on Liquid Atomization and Spray Systems, Heidelberg, Germany,
September26.
Heeger, C., Gordon, R. L., Tummers, M. J., Sattelmayer, T., and Dreizler, A. (2010).
Experimental Analysis of Flashback in Lean Premixed Swirling Flames: Upstream
Flame Propagation. Exp Fluids 49: 85363, Springer-Verlag.
Jones, R. E. (1978). Gas Turbine Engine Emissions Problems, Progress and Future. Prog.
Energy Combust. Sci. 4: 73113.
288 Premixed Aero Engine Combustors
Kastl, J. A., Herberling, P. V., and Matulaitis, J. M. (2005). Low NOx Combustor Development.
NASA/CR-2005213326.
Kokanovic, S., Guidati, G., Torchella, S., and Schuermans, B. (2006). Active Combustion
Control of NOx and Pulsation Levels in Gas Turbines. ASME GT200690895.
Lazik, W., Doerr, Th., and Bake, S. (2007). Low NOx Combustor Development for the Engine
3E Core Engine Demonstrator. ISABE-20071190.
Lazik, W., Doerr, Th., Bake, S., v.d. Bank, R., and Rackwitz, L. (2008). Development of
Lean-burn Low NOx Combustion Technology at Rolls Royce-Deutschland. ASME
GT200851115.
Lefebvre, A. H. (1998). Gas Turbine Combustion, second edition, Taylor and Francis,
Philadelphia.
Leonard, G., and Stegmaier, J. (1994). Development of an Aeroderivative Gas Turbine Dry
Low NOx Emissions Combustion System. J. Eng. Gas Turbines and Power 116: 5426.
Leong, M., McDonnell, V., and Samuelsen, S. (2001). Effect of Ambient Pressure on an
Airblast Spray Injected into a Crossflow. J. of Prop and Power 17 107684.
Lyons, V. J. (1982). Fuel/air Nonuniformity Effect on Nitric Oxide Emissions. J. AIAA 20:
6605.
Maquisten, M. A., Holt, A., Whiteman, M., Moran, A. J., and Rupp, J. (2006). Passive Damper
LP Tests for Controlling Combustion Instability. ASME GT200690874.
Marchione, E., Ahmed, S. F., and Mastorakos E. (2007). Ignition Behavior of Recirculating
Spray Flames using Multiple Sparks. Mediterranean Combustion Symposium, MCS 5,
Monastir, Tunisia.
Maurice, L., Ralph, M., Tilston, J., and Kuentzmann, P. (2009). Report of the Independent
Experts to CAEP/8 on the Second NOx Review & Long Term Technology Goals.
London, March.
McVey, J. B., and Kennedy, J. B. (1979). Lean Stability Augmentation Study.
NASA-CR-159536.
Midgley, K., Spencer, A., and McGuirk, J. (2007). Vortex Breakdown in Swirling Fuel Injector
Flows. ASME GT200727924.
Mondragon, U. M., Brown, C. T., and McDonnell, V. G. (2011). Evaluation of Atomization
Characteristics of Alternative Fuels for JP 8. ILASS-Americas, Ventura, CA, May.
Mongia, H. (2003). TAPS A 4th Generation Propulsion Combustor Technology for Low
Emissions. AIAA 20032657.
Mongia, H. (2007). GE Aviation Low Emission Combustion Technology Evolution. SAE
2007013924.
Nickolaus, D. A., Crocker, D. S., and Smith, C. E. (2002). Development of a Lean Direct Fuel
Injector for Low Emission Aero Gas Turbines. ASME GT-200230409.
Oda, T., Kinoshita, Y., Kobayashi, M., Nimomiya, H., Kimura, H., Hayashi, A., Yamada,
H., and Shimodaira, K. (2003). The Development of LPP Combustor for ESPR.
IGTC2003Tokyo TS-150.
Plee, S. L., and Mellor, A. M. (1978). Review of Flashback Reported in Prevaporizing/
Premixing Combustors. Comb. and Flame 32: 193203.
Rhode, J. (2002). Overview of the NASA AST and UEET Emissions Reduction Projects,
http://adg.stanford.edu/aa241/emissions/NASAUEETemissions.pdf.
Roberts, R., Peduzzi, A., and Vitti, G. E. (1975). Experimental Clean Combustor Program,
Phase I Final Report P&W. NASA-CR-134736.
Roberts, R., Peduzzi, A., and Vitti, G. E. (1976). Experimental Clean Combustor Program,
Phase II, Final Report P&W. NASA CR-134969.
Roffe, G., and Venkatarami, K. S. (1978). Experimental Study of the Effects of Flameholder
Geometry on Emissions Performance of Lean Premixed Combustors. NASA
CR-135424.
Shanbhogue, S. J., Husain, S., and Lieuwen, T. C. (2009). Lean Blowoff of Bluff Body Stabilized
Flames, Scaling and Dynamics. Progress in Energy and Combustion Science 35: 98120.
Sokolowski, D. E., and Rohde, J. E. (1981). The E3 Combustors: Status and Challenges.
AIAA-19811353.
References 289
Sturgess, G. J., Zelina, J., Shouse D., and Roquemore, W. M. (2005). Emissions Reduction
Technologies for Military Gas Turbine Engines. J. of Propulsion and Power 21: 193217.
Tacina, R. (1990). Low NOx Potential of Gas Turbine Engines. AIAA-900550.
Tacina, R., Lee, Ph., Wey, Ch. (2005). A Lean-direct-injection Combustor Using a 9 Point
Swirl-venturi Fuel Injector. ISABE-20051106.
Tacina, R., Mansour, A., Partelow, L., and Wey, Ch. (2004). Experimental Sector and Flame
Tube Evaluations of a Multipoint Integrated Module Concept for Low Emission
Combustors. ASME ITGI-200453263.
Tacina, R., Mao, Ch-P., and Wey, Ch. (2003). Experimental Investigation of a Multiplex Fuel
Injector Module for Low Emission Combustors. AIAA 20030827.
Tacina, R., Wey, Ch., Laing, P., and Mansour, A. (2002). A Low NOx Lean-Direct Injection,
Multipoint Integrated Module Combustor Concept for Advanced Aircraft Gas Turbines.
NASA/TM 2002211347.
Umeh, Ch., Kammer, L., and Barbu, C. (2007). Active Combustion Control by Fuel Forcing
at Non-coherent Frequencies. ASME GT200727637.
U.S. Patent No. 6,272,840, 2001.
Wedlock, M. I., Tilston, J. R., and Seoud, R. E. (1999). The Design of a Piloted, Lean Burn,
Premixed, Prevaporized Combustor. RTO-MP 14, AVT Symp On Gas Turbine Engine
Combustion, Emission and Alternative Fuels, Paper 23, 111.
Yamamoto, T., Shimodaira, K., Kurosawa, Y., Matsuura, K., Iino, J., and Yoshida, S. (2009).
Research and Development of a Staging Fuel Nozzle for Aeroengine. ASME
GT200959852.
Yamamoto, T., Shimodaira, K., Kuosawa, Y., Yoshida, S., and Matsuura, K. (2010).
Investigations of a Staged Fuel Nozzle for Aeroengines by Multi-sector Combustor
Test. ASME GT201023206.
Zeisser, M. H., Greene, W., and Dubiel, D. J. (1982). Energy Efficient Engine Combustor Test
Hardware Detailed Design Report. NASA CR-167945, PWA-5594197.
10 Industrial Combustors: Conventional,
Non-premixed, and Dry Low
Emissions (DLN)
Thomas Sattelmayer, Adnan Eroglu, Michael Koenig,
Werner Krebs, and Geoff Myers
10.1 Introduction
Modern industrial gas turbine combustors have to meet a wide range of technical
requirements governed by the thermodynamic cycle, the demand for a low environ-
mental impact, the compatibility with other machine components, and the safety
and reliability aspects. Although the evolution of combustion chamber technology
has resulted in substantially different hardware solutions, a high level of similarity
exists as to how the subprocesses take place in combustion chambers of different
design and how they interact with each other. What the basic designs of combustors
for industrial gas turbines relevant to emissions have in common will be discussed in
the first part of this chapter to create the basis for a better understanding of the sec-
ond part, where several specific design solutions are presented. These are grouped
into different classes of combustors according to fuel type and the pollutant emis-
sion abatement technique employed. The present state of the art is the result of the
simultaneous optimization of several subprocesses relevant for the performance of
combustors. Expanded discussion of many of the basic combustor technology driv-
ers and how they interact with the electrical grid as a whole, the power plant, and the
gas turbine are also included in Chapter 1.
290
10.1 Introduction 291
has led to the more complex and delicate multi-burner systems, which are the state
of the art today.
Fuel Flexibility
Fuel specifications for low-emission combustors are usually restrictive, and increased
fuel flexibility is the most important driver for current research and development.
Besides natural gas, fuels containing large amounts of higher hydrocarbons or fuels
with hydrogen and carbon monoxide are becoming of increasing interest.
In many cases, multi-fuel capability of industrial gas turbines is required, allow-
ing operation of the engines on at least two fuels for standard operation or one
standard plus one backup fuel, respectively. Fuel switchover capability between pri-
mary and backup fuels during operation is commonly required. Purging of the fuel
distribution system not used during continuous operation on the other fuel or after
fuel switchover should be technically simple and economic or even be avoided.
Environmental Compatibility
The general aim of modern combustion technology for gas turbines is to keep the
emissions of nitrogen oxides (NOx), carbon monoxide, and volatile organic com-
pounds (VOC) low over an increasing operational power range. These emissions
are detailed in Chapters 6 and 7. Compared with the demand for high combustion
efficiency, the emission limits for volatile organic compounds (unburned fuel and
hydrocarbon intermediates), particularly carbon monoxide, are much greater chal-
lenges to modern gas turbine combustion technology. This is especially true for
part-load operation of gas turbines. Above all, regulations limiting nitrogen oxide
292 Industrial Combustors
emissions to only a few parts per million have led in the past to fast innovation and
to frequent combustor design changes. The continuation of this trend is expected
to drive combustion technology further toward almost pollutant-free combustion
products with respect to the typical emissions generated by other thermal engines,
such as aero engines or reciprocating engines. Increasing the operation range in
which extremely low NOx and CO emissions can be achieved is an important cur-
rent driver of research.
Stability
During steady operation under load, fast load changes, sudden load rejection, and
when starting up the machine, combustors must provide sufficient margin against
flame extinction. Permitting sufficient so-called static stability is particularly chal-
lenging for premixed, low-emissions combustors and their control system because
they operate very close to the lean combustion limit.
Furthermore, combustors are required to provide enough dynamic stability
in the entire operation envelope, that is, sufficient margin against the onset of ther-
moacoustic combustion oscillations. Since the technology changes associated with
the introduction of low-emission combustors in the past were often accompanied by
severe instability problems, for many years, research on thermoacoustics has been
the most prominent topic of low-emission combustor development.
Combustor Integration
The overall engine design imposes numerous geometric constraints that limit the
combustion chamber topology and length and that influence the thermoacoustic sys-
tem properties.
One of the very few constants in the historical evolution of gas turbine combus-
tors is the design of the aerodynamics of combustion chambers for low-pressure
losses not exceeding 34 percent of the absolute pressure.
Sufficient azimuthal uniformity of the temperature profile at the combustor exit
must be guaranteed by proper tailoring of the combustor aerodynamics.
In industrial gas turbines with premixed combustion, the high radial uniformity
of the temperature profile is a challenge for turbine cooling as it does not allow
the reduction of the thermal load at the vane and blade roots and tips at the
expense of a hotter core flow.
10.2 Flame Types 293
All parts of the combustor exposed to heat input must support cycling capabili-
ties of the gas turbine with steep load gradients up to 50 MW/min during tran-
sient operation modes, such as frequency response or protective load shedding.
Although designers have always strived to preserve the simplicity of the fuel
distribution and control systems of older machine generations, the fuel feed sys-
tems of low-emission gas turbines have become gradually more complex and
costly with more demanding emission limits.
For gaseous fuels, the required fuel pressures have to be kept as low as possible.
This limits the fuel momentum available for injection into the air in the burners
and for fuel air premixing in low-emission burners.
Low flow rates for water injection, predominantly used during oil operation for
the reduction of the formation of NOx, are required.
Economic Aspects
Service friendliness of the combustion system is commonly of great customer con-
cern. Extension of component lifetimes, provision of long service intervals, and short
exchange times of the parts exposed to high temperatures are essential.
Finally, meeting the cost targets imposed on the gas turbine combustion sys-
tem stimulates continuous design efforts aiming at decreasing the complexity of
current solutions and at using components that can be manufactured with better
economy.
Air
Woi = LHVvol (10.1)
Fuel
The Wobbe index relates to the lower heating value (LHV) of the fuel and denotes
that fuels characterized by the same Wobbe number will cause the same pressure
loss in the fuel piping system when releasing the same heat in the combustor. The
lower the Wobbe Index, the larger the cross-section areas of the fuel delivery system
must be. Combustion systems, which can be fired with natural gas and syngases or
even blast furnace gas, need to cover an extremely wide Wobbe Index range. This
requires two separate fuel supply systems for the burners. Interesting, the Wobbe
Indices of natural gas and pure hydrogen are of the same order, although their heat-
ing values are considerably different.
The second important property of gaseous fuels is their reactivity, characterized
in Figure 10.1 by the reactivity time according to the characteristic time scale Chem:
a
Chem (10.2)
sl 2
The chemical time scale is given as the ratio of the thermal diffusivity a divided by
the square of the laminar burning velocity s1. A drop in the chemical time scale will
result in an increased reactivity of the gaseous fuel blend. Figure 10.1 illustrates that
extending the fuel spectrum from natural gas toward syngases or even hydrogen
requires reliable low-emission combustion technologies for highly reactive fuels.
10.2 Flame Types 295
0.1
Syngas (undiluted)
Carboncapture Hydrogen
process gas
0.01
Figure 10.1. Chemical time scale and Wobbe Indices of gas turbine fuels.
Because of its high chemical stability and its low laminar flame speed, sl, natural
gas is particularly suitable for the premixed combustion at low flame temperatures.
Even in machines with sequential reheat combustion (see Section 10.5.3.2), pre-
mixing at hot gas temperatures of approximately 1300 K at the inlet of the second
stage can be accomplished without premature self-ignition. For machines in peak
load operation, non-premixed combustion with water injection for simultaneous
NOx abatement and power augmentation may be an attractive alternative, assuming
operation without water injection leading to excessive NOx formation is permitted
or not required.
Premixed burners in gas turbines are designed primarily for natural gas as fuel.
Because of the lower self-ignition time and energy and higher laminar flame speed,
respectively, higher hydrocarbons or hydrogen can substantially increase the pro-
pensity to flashback and self-ignition. For this reason, non-premixed combustion is
generally applied for reactive HBtu fuels, if necessary with water injection.
Most of the heating value of typical fuels from oxygen-blown coal or oil resi-
dues gasification is bound in hydrogen and carbon monoxide. Their inert compo-
nents, predominantly nitrogen and carbon dioxide, reduce the heating value (MBtu
fuels) and Wobbe Indices (see Figure 10.1). Such fuels are often further diluted to
LBtu fuels prior to non-premixed combustion. With sufficient dilution, NO forma-
tion is almost fully suppressed because the peak temperature in the flame is nat-
urally limited. Only burners with optimized and intensified quick mixing permit
partially premixed combustion of MBtu fuels with low NOx emissions (see Section
10.5.1.2).
With the decreased heating value of LBtu fuels, the operational window for
burner operation becomes increasingly narrowed around stoichiometric fuel-air
ratios. The reason for this effect is the large concentration of inert components. As
the flame temperatures generally remain low despite low excess air, non-premixing
combustion is suitable, but special techniques for large-scale mixing of the large fuel
mass flows with the combustion air at the burner outlet must be provided to avoid
burnout problems stemming from incomplete mixing in the combustor. Because
of the fuel-air mixing in the combustion chamber, flashback or self-ignition are not
296 Industrial Combustors
mean concentration of combustion products with increasing distance from the outlet
of non-premixed burners to the point when all fuel has been consumed.
An important aspect of pollutant formation is the diffusion of the temperature
from the chemically active layer toward the fuel, leading to heating of areas of high
fuel concentration causing reaction in fuel-rich zones with the formation of soot
in the flame, as detailed in Chapter 5. Even globally lean gas flames at typical gas
turbine combustor chamber pressures exhibit a broadband radiation spectrum emit-
ted by the soot particles in the flame. In the soot oxidation zone downstream of the
primary zone, this spectrum is gradually transformed to a spectrum with the spectral
bands stemming from the radiation of the gaseous combustion products (carbon
dioxide and water vapor).
One particular advantage of non-premixed flames is that in gas turbines operat-
ing with different global fuel-air ratios, the equivalence ratio in the main reaction
zone is largely decoupled from the global fuel-air ratio of the burner. By suitably
designing the fuel injection system, the combustion process can be maintained up to
very low global equivalence ratios. The flame responds to a decrease of the fuel flow
rate with a shift of the reaction zone toward the fuel injector where the zones with
the most favorable equivalence ratio for reaction are located.
In the premixed case, diffusion of the temperature (heat conduction) and diffu-
sion of active combustion intermediates toward the approaching fuel-air mixture lead
to heating and radical attack of the fuel molecules, respectively, and finally to the onset
of reaction in the fuel-air mixture. The burning rate of a given mixture is the result of
the interplay of the volumetric heat release in the flame front, the diffusion processes,
and the heat required for the preheating of the mixed reactants up to the ignition
temperature. The laminar burning velocity sl increases with the temperature of the
fuel-air mixture approximately in the form of an Arrhenius equation. For typical gas
turbine fuels, an increase of pressure reduces the laminar burning velocity because the
required preheating per volume is proportional with pressure, in contrast to the local
diffusive transport of temperature and species that is almost independent of pressure.
Under gas turbine conditions, the much higher turbulent flame speed st is essen-
tial. It is mainly determined by the structure of turbulence in the primary zone of
the combustion chamber. In the area of interest for gas turbine combustion with air
overshoot (lean regime), the burning rate drops with increasing air ratio until abrupt
flame extinction occurs. When burning natural gas or other fuels with low reactivity,
the equivalence ratio of premixed gas turbine flames is selected as lean as possible
in the normal operating range, because low NOx formation rates are obtained only
in the narrow equivalence ratio range close to the lean extinction limit. This major
disadvantage makes it difficult for the adaptation of premixed combustion to the
wide range of combustor equivalence ratios required for the load range covered by
gas turbines. In any case, fuel and/or air staging are required, which increases the
complexity of premixed combustors significantly with respect to their non-premixed
counterparts exhibiting much higher NOx emissions. The design of premixed gas tur-
bine combustors has always been restricted by the trade-off between emissions and
complexity that cannot be circumvented.
298 Industrial Combustors
0 20 m/s
40 60
20 50
0 40
20 30
Premix 20
10
burner 0
10
20
30
40
Combustion
chamber
Figure 10.2. Flow field in a swirl-stabilized combustor: distribution of the axial component of
the velocity field in the mid-plane of a square combustion chamber.
Instan-
-
Premix taneous
at lean blowout
burner
Combustion
Average at full load chamber
Instan-
taneous
at full load
time scales and to keep flame stabilization intact. Within the permitted range for the
burner pressure drop (24 percent of the combustor pressure), the flow velocity also
has only a low influence on the lean blowout temperature.
As the comparison of the two inserted single shot images in Figure 10.3 shows,
the front of premixed gas turbine flames is wrinkled and corrugated. When the
equivalence ratio approaches the lean extinction limit, the structure changes toward
an ensemble of isolated reacting zones. This effect deteriorates the permanent igni-
tion mechanism because of recirculation of hot products until the cycle breaks down
and the flame extinguishes.
Swirl-induced flame stabilization dominates also in non-premixed gas tur-
bine burners. For standard gas turbine fuels, the ignition zones are generally
found in areas where the local equivalence ratio is not too far from stoichiomet-
ric. If a high static stability of the flame is required, the fuel is injected in a clas-
sical manner from a central lance into the shear layer between the incoming air
300 Industrial Combustors
K
1500 1700
1100
1500
Premix 1300
1700
burner
900
1100
1300
900
700
Combustion chamber
that sets the time for the turbulent mixing of fuel and air in relation to a character-
istic time required for reaction:
tMix
Da = (10.3)
tChem
In the hypothetical case of extremely small Damkhler numbers, the reaction would
proceed in partially premixed areas even though the fuel and air were fed into the
combustion chamber separately. However, the non-premixed combustion of typical
gas turbine fuels under pressure is characterized by Damkhler numbers usually
larger than 1. This indicates that mixing intensity governs the average reaction prog-
ress in combustors, not reaction kinetics. This finding is of utmost importance for
non-premixed combustion because it is possible to control the longitudinal reaction
progress by simultaneous design of the flow field of the air in the combustor and the
fuel injection strategy in a wide range.
Important design criteria for non-premixed gas turbine combustors are achiev-
ing complete chemical reaction within a short distance downstream of the burner
exit, a high ratio of volume to surface ratio of the combustion chamber, and low
soot formation. As these requirements are met only by intensely mixing burners that
produce a compact flame, the design of gas turbine burners usually seeks to attain
reduced Damkhler numbers.
Since the formation and oxidation of volatile organic compounds (VOC) and
hazardous air pollutants (HAP) are discussed in detail in Chapter 6, while Chapter
7 provides the essentials of CO kinetics, the subsequent explanations concerning
fuel oxidation and emission limits are focused on aspects particularly relevant for
industrial gas turbines.
For newly installed gas turbines, emission regulations usually limit carbon mon-
oxide to be fulfilled to a few ppm in the load range covered in regular operation of the
engines. This requires almost full conversion of the fuel-air mixture to the thermody-
namic equilibrium in the combustion chamber. Carbon monoxide concentrations in
the single digit ppm range can be reached in gas turbine combustors only in zones of
high excess air. Equilibrium calculations for CO at typical combustor pressures show
that an equilibrium concentration below 10 ppm (1 ppm) requires equivalence ratios
below 0.51 (0.43) and flame temperatures below about 1800 K (1650 K). For refer-
ence, the CO equilibrium is approximately three orders of magnitude higher for stoi-
chiometric adiabatic combustion. For typical hydrocarbon fuels, the adiabatic flame
temperature range between 1800 K and 1650 K lies within the limits of flammability
and near the lean extinction limit of premixed flames, and is also the range of lowest
nitrogen oxide formation in the primary zone of premix burners. This range provides
the most favorable conditions for premixed flames regarding low NOx as well as CO
emissions. In this range, premixed flames do not emit VOCs and soot.
From thermodynamic equilibrium calculations, it can be further deduced that
the carbon monoxide concentrations directly after completion of the reaction in the
hot, almost stoichiometric, reaction zones of non-premixed flames are always far
304 Industrial Combustors
above the limits permitted for exhaust gases. The required drop to acceptable levels
occurs by admitting air to the combustion products, leading to a gradual increase of
the air overshoot. Non-premixed flames in gas turbines require, therefore, a dilution
zone downstream of the main reaction zone with the lower temperature level to
where the CO equilibrium concentration decreases. The axial gradient of this drop is
determined by turbulent mixing.
Besides thermodynamic equilibrium, the kinetics of the CO burnout must also
be considered. The large concentration of OH radicals produced during the break-
down of the fuel is essential for the conversion of CO to CO2 (see Chapter 7). For
temperatures and pressures typical for gas turbines, it can be shown that the kineti-
cally induced burnout takes approximately only one millisecond. This is usually
about one order of magnitude faster than the mean residence time in well-designed
combustion chambers of industrial gas turbines.
At low power, this is no longer the case: with falling temperatures in the com-
bustion chamber, the equilibrium CO concentration decreases but, simultane-
ously, the reactions leading to the conversion of CO to CO2 slow down significantly
because of the effect of temperature on their reaction rate and the concentration
of the required radicals (e.g., OH). Finally, the kinetics of CO burnout becomes the
determining factor for combustion efficiency and emissions in the low-temperature
regime. Interesting, since similar effects also occur during the expansion in the tur-
bine, no effective carbon monoxide oxidation processes are observed in the exhaust
gases after expansion.
10.3 NO Formation
Chapter 4 lists the emission requirements for NOx for various countries. In power
production, the emissions are measured in parts per million (ppm) of dry exhaust
gas, which is then diluted with air so that it finally has a volumetric oxygen concen-
tration of 15 percent. The most severe emission requirements are found in parts of
the United States, where some places require less than 2 ppm of NOx. To date, such
low values can only be achieved by using selective catalytic reduction of NOx with
ammonia in catalysts mounted in the exhaust stack.
The equilibrium concentration of NO typical for gas turbine pressures lies sev-
eral orders of magnitude above the required limits detailed in Chapter 4. The reason
such low values can be met with appropriate combustor designs is the very low NO
formation rate in comparison with the much faster fuel oxidation reactions. In indus-
trial gas turbines, the required reduction of the NO formation to acceptable levels
is achieved by the control and the reduction of the temperature during combustion
and by the limitation of the residence time at high temperatures in the combustor to
the minimum required for complete burnout of the fuel.
The nitrogen chemistry leading to the emission of oxides of nitrogen has been
presented in detail in Chapter 7. Five mechanisms were identified and their character
was explained. Four of them, plus the formation of nitrogen dioxide, are of particular
relevance for gas turbine emissions. Their individual contribution to the overall NOx
10.3 NO Formation 305
emissions mainly depends on the fuel and the flame type as well as the pressures and
temperatures (Dean and Bozzelli, 2000). The contribution of these mechanisms to
the local generation rate of NOx is different in the reaction zone represented by an
ensemble of flamelets, and in the post-flame region of the combustor. From the view-
point of combustion in industrial gas turbines, the following aspects are of particular
design relevance:
Nitrous Oxide (N2O) Pathway. Since the N2O formation finally leading to NO is
based on the reaction of molecular nitrogen with oxygen radicals in the presence of
a third collision partner, the reaction is favored by high pressure. It is activated at
lower temperatures than the thermal route. This pathway contributes substantially
to NO generation at very lean conditions and can even dominate near the lean blow-
out limit of premixed flames.
Non-premixed flames allow the tailoring of the spatial heat release distribution by
control of fuel-air mixing, but the stoichiometry of the reaction zone can only be
influenced to a limited degree. The disadvantage of non-premixed flames is mainly
that the maximum temperature cannot be sufficiently limited and that, because of
the fairly long residence time in the hot zones of such flames, NO formation is one to
two orders of magnitude higher than commonly permitted. Even with the strongly
increasing air excess associated with load reduction of gas turbines, NOx emissions
only drop slowly and, at a certain lower limit, CO emissions start rising. This situa-
tion is qualitatively shown in Figure 10.5 for unstaged non-premixed combustors.
There, the CO and NOx emissions divided by their permitted limits are sketched as
functions of combustion temperature and load. The combustion temperature given
here is a computational bulk value obtained from calculating the adiabatic flame
temperature from the given air and fuel mass flows.
At higher temperatures, the NOx emissions increase exponentially with com-
bustion temperature and load. However, the emissions remain far beyond the com-
monly permitted values (see Chapter 4), even at very low load and near idle. The
comparison with the CO emission curve shows that the conversion to CO2 is suf-
ficient above a minimum flame temperature and load, but that both emission lim-
its cannot be reached simultaneously without applying additional NOx abatement
techniques, such as the injection of water in liquid or gaseous phase into the pri-
mary zone. This makes non-premixed combustion highly unattractive for industrial
gas turbines. The effect of water injection is also illustrated in Figure 10.5.
Inherent air staging can be achieved with proper control of the heat release
in the longitudinal direction of non-premixed flames by tailoring the fuel-air
mixing. For this purpose, fuel-rich and chemically reducing zones are created near
the fuel injector. The reburn mechanism in such fuel-rich zones limits the conver-
sion of chemically bound nitrogen to NO. However, this effect is too weak and its
NOx abatement potential is not compatible with the low emission limits listed in
Chapter4.
If the reaction in hot and unmixed areas is suppressed through the injection of
water, other inert species, or by fuel dilution, an improved mixing of the air with the
fuel before the onset of the reaction is achieved. Additionally the fuel conversion
10.3 NO Formation 307
NO/NOlimit
CO/COlimit
CO/COlimit
Limits
1
0% 100%
Increasing flame temperature Power output
Figure 10.5. NOx and CO emissions of non-premixed flames as function of the flame tempera-
ture and the gas turbine load.
near the injector is also affected by the reactions, for example, from steam reforming
of hydrocarbons. However, primarily the improvement of the mixing in the reac-
tion zone leads to a drop in the NOx emissions (Figure 10.5) toward the low levels
known from premixed combustion. At the same time, the routes for NO formation
become more similar to the formation routes observed for the premixed case. How
effective injected water reduces NOx formations depends on the specific configura-
tion. However, as a rule of thumb, water-to-fuel mass flow ratios of 1 lead to a drop
of the NO formation by approximately one order of magnitude. Since much higher
water-to-fuel mass flow ratios usually lead to dynamic and, finally, static flame insta-
bilities for standard gas turbine fuels with high heating values, the NOx abatement
potential of water injection is lower than the potential of premixed combustion and
the lowest NOx targets listed in Chapter 4 remain unattainable.
6 4 6 4
Figure 10.6. NO formation in laminar flames (TProducts = 1840 K). The chemical reaction pro-
ceeds from left to right; (TProducts-T) / (TProducts-TReactants) equals 1 and zero in the reactants and
equilibrium products, respectively.
concentration and the production rate, it can be deduced that diffusion from the
product side influences the NO concentrations in the colder part of the flame. At
high pressure, the production rate for NO is negative below 40 percent tempera-
ture rise, because in this temperature range a shift of NO to NO2 takes place that is
reversed at higher temperatures.
A particularly interesting finding stemming from the analysis of the two exam-
ples shown in Figure 10.6 is that, at the chosen adiabatic flame temperature, almost
no pressure dependency of the NO concentration is present when the comparison is
made on the basis of a normalized temperature, such as the one used as the abscissa
scale of the two graphs. However, Figure 10.6 also shows that the NO production
rate correlates with pressure and, consequently, that NO formation is much faster at
engine pressure.
Figure 10.7 shows the reaction progress in terms of the temperature rise with time.
At high pressure, the reaction time is on the order of one millisecond. Afterward, the
products approach their final temperature, but the NO formation continues in the
post-flame gases. This clearly shows that limiting the residence time of the combus-
tion products in the combustion chamber to the time required for chemical reaction
(burnout) would be beneficial. Unfortunately, these short times cannot be realized
because the time required for the turbulent flame propagation through the mixture
(see Section 10.2.4) is typically an order of magnitude longer than the chemical time
and the length of the combustion chambers has to be designed for considerably less
favorable reaction conditions at partial load.
The impact of the residence time in the post-flame zone of combustors on NO
emissions is summarized in Figure 10.8. The lower curve represents the NO forma-
tion in the flame zone within the time interval tminimum required for CO oxidation
(limit set to 150 percent of the equilibrium value). The set of curves above represents
the emissions for different post-flame residence times t- tminimum. It can clearly be seen
that at low flame temperatures, very low nitrogen oxide emissions can be achieved
310 Industrial Combustors
NO (ppmv)
0.1 MPa
T(K) 0.1 MPa CO (vol %)
T(K) 1.5 MPa NO (ppmv)
1.5 MPa
2000 CO (vol %)
5
1500 4
Mole fraction
3
T(K)
1000
2
500
1
0 0
0 1 2 3 4 5
t (ms)
Figure 10.7. Comparison of the time required for heat release in laminar flames (TProducts =
1840 K).
100 100
t-tminimum 0.1 MPa t-tminimum
0 ms 0 ms
5 ms 5 ms
10 ms 10 ms
NO (ppm, 15% O2)
20 ms 20 ms
40 ms 40 ms
80 ms 80 ms
10 10
1.5 MPa
1 1
1650 1700 1750 1800 1850 1900 1950 2000 1600 1650 17001750 1800 18501900 1950 2000
Tproducts (K) Tproducts (K)
Figure 10.8. Influence of adiabatic flame temperature and the residence time on the NO for-
mation of perfectly premixed flames.
products upstream of the second fuel injection. However, achieving complete mix-
ing of the secondary fuel with the combustion products from the first stage prior to
the onset of the reaction is a challenge if the temperature exceeds 12001300 K. This
limit is only valid for natural gas and wet combustion of fuel oil. For more reactive
fuels, the prevention of premature self-ignition requires even lower temperatures.
Consequently, the technical potential of the staged process for reducing NO forma-
tion below the values given for unstaged processes is limited.
In general, the high NO reduction potential of premixed combustion is restricted
to fuels without fuel-bound nitrogen because the breakup of fuel molecules contain-
ing nitrogen leads to almost full conversion to NO in lean, premixed flames.
air and fuel at the molecular level. The logarithmic plot in Figure 10.8 further shows
that the NO emissions generally increase exponentially with the flame temperature,
a nonlinearity that has far-reaching consequences for the combustion technology of
gas turbines and that makes the improvement of mixedness a major task in connec-
tion with the reduction of NOx emissions.
In this context, the following two subtasks have to be tackled. First, the fuel must
be evenly distributed over the cross-section of the incoming air, and second, the
temporal fluctuations of fuel concentration in the burner exit plane must be reduced
as much as possible. Providing temporal uniformity represents the main technical
challenge.
The attempt to achieve large-scale mixing of the combustion air with gaseous
fuels injected at a single or a low number of injection locations solely through the
turbulent motion of the largest eddies (stirring the mixture) leads to very long pre-
mixing ducts unsuitable for gas turbine combustors, as they would make the combus-
tion chamber too long and the long residence times associated with them would lead
to risk of self-ignition in the premixing section. Gas turbine burners must therefore
always provide a transport mechanism that leads to fast, large-scale mixing of the
fuel and the inflowing of the air over the entire cross-section of the inflow. Basically
two methods can be identified.
Mixing Energy Derived from Fuel-Jet Momentum. For standard gaseous gas tur-
bine fuels, the main problems encountered here are the small mass ratio between the
fuel and air and the limited pressure of the fuel. Both limit the kinetic energy input
into the airflow and lead to limited penetration of the fuel jets. As a consequence,
the inflow paths for the air in gas turbine burners are commonly segmented (mul-
tiple air inlet slots or swirl vane registers of burners; see Figure 10.39) and exhibit
very small aspect ratios. In order to provide good relative penetration of the fuel
jets injected along the length of the air inlet ports, the width of each passage should
not exceed twice the possible fuel jet penetration depth. For the premixing of liquid
fuels, a similar injection strategy can be applied. However, since much higher pres-
sures are generally allowed for liquid than gaseous fuel, liquid fuel can be injected
at a substantially lower number of injection points, especially if the fuel momentum
is used optimally to increase the penetration of the fuel into the air. Premixed sys-
tems employing a very low number of compact liquid fuel jets were introduced more
than ten years ago (see Figure 10.46). The challenge of designing such systems is to
achieve an adequate dispersion of the liquid fuel jets over the flow cross-section and
to optimize the liquid jet breakdown into droplets, the droplet motion, and their
evaporation, so that finally at the outlet of the burner a uniform gaseous fuel-air
mixture is obtained. A particular problem here is the strong sensitivity of the mix-
ture formation to the combustion chamber pressure, the air temperature, and the
mass flow rates of fuel and air.
Mixing Energy Derived from the Airflow. Since the mass flow of the air is by
more than one order of magnitude higher than that of standard gas turbine fuels,
10.3 NO Formation 313
1 3 2
4 5
using the capability of the air for the promotion of large-scale mixing is an alterna-
tive to using fuel momentum alone. This is usually accomplished by the generation of
vortical secondary flows in the inlet ports of the air (see Section 10.5.3.2). Particularly
effective mixing is achieved when the fuel is injected toward the centers of the vorti-
ces. It is then transported by the vortex motion across the entire cross-section of the
inflow and distributed in the air on the macroscopic scale.
The spatially homogeneous large-scale distribution of the fuel is a precondition
for low NOx emissions. But of equal importance is the mixing of the microscales of
the turbulent flow, thereby providing temporally constant fuel content in the mixture
approaching the flame front. For optimized gas turbine burners, the necessary length
required to obtain a uniform spatial fuel distribution is considerably smaller than
the distance required to reduce the local mixture fluctuations on the microscales.
The basic problem of turbulent fuel-air mixing in gas turbine burners is qualita-
tively illustrated in Figure 10.9, employing the simple example of the mixing of a row
of fuel jets injected into a co-flow of air, a configuration illustrative but also fairly
uncommon in premixed gas turbine burners. The local probability density function
of the mixture fraction of the fuel in the fuel-air mixture representing the tempo-
ral mixture fluctuations is analytically described by the -function, having two free
parameters, the local mean f and local variance g of the fuel mixture fraction f. The
standard deviation g characterizes the width of the distribution.
In a cross-section just downstream of the injection, either pure fuel (Position 1:
f = 1, g = 0 ) or air (Position 2: f = 0, g = 0 ) are found over the cross-section of the
314 Industrial Combustors
flow. In the thin shear layer between the air and fuel, a turbulent vortical structure
with strong intermittency of fuel and air (Position 3: 0 < f < 1 ) is present. However,
mixing on the molecular level is very low in this zone. After the shear layers have
merged further downstream, the average concentration gradients are successively
reduced to one average value that is almost constant perpendicular to the flow direc-
tion (Position 4). However, mixing on the microscale is not yet complete at that
point. Even for moderate variances, g, pockets of near stoichiometric or even fuel-
rich mixtures are present that significantly increase NOx emissions. Therefore, an
additional mixing length is needed to reduce the fluctuations to noncritical values
(Position 5).
The step from Position 4 to Position 5 is essential from the point of view of the
influence of finite mixing quality on the NOx emissions of gas turbine burners. A
straightforward measure for the degree of small-scale mixing achieved in premix
burners is the normalized standard deviation, s, of the mixture fraction:
g
s= (10.4)
f
The analysis of Figure 10.8 for 1.5 MPa reveals that all local zones with equivalence
ratios more than approximately 5 percent above the average contribute to a per-
ceptible increase of the integral NOx emissions of the flame because corresponding
leaner zones cannot fully compensate for the high NOx formation in the zones with
higher equivalence ratios. Assuming a normal distribution representing the mix-
ing quality in the fine-scale mixing zone, this implies that the standard deviation
of the equivalence ratio distribution upstream of the turbulent flame front must be
reduced to approximately 2 percent, of the mean mixture fraction ( s = g f = 0.02 )
to reduce the probability of the formation of pockets with more than 5 percent devi-
ation toward zero.
The degree of mixing is often characterized by the unmixedness parameter, U,
which relates the variance of the mixture g to the variance gunmixed = f (1 f ) obtained
in the hypothetical and entirely unmixed case:
g
U= (10.5)
f (1 f )
For perfect mixing, we obtain U=0, and for the entirely unmixed case, U = 1. With
g = s 2 f 2 = 4 10 4 f 2 (2 percent normalized standard deviation) and with f = 0.03,
for example, natural gas-air mixtures in the lean regime, we obtain U 105 from
Equation 10.5 for the upper limit of the mixing quality of burners that approach
minimum NOx formation. For a given normalized standard deviation s = g f , this
limit scales almost linearly with f for gas turbine fuels burned in premixed flames
( f 0.5 ).
The extremely small value required for U highlights the outstanding role of mix-
ing in the design of premix burners. The optimization of the mixing zone by the
intensification of turbulent mixing is an important objective for the development
10.3 NO Formation 315
0.1
0.06
0.09
0.08 0.05
0.07
0.04
0.06
0.05 0.03
0.04
0.02
0.03
0.02
0.01
0.01
0 0
Figure 10.10. Mixing quality in the exit plane of two premix burners (laser induced fluores-
cence, water channel experiments, left: instantaneous concentration, right: temporally aver-
aged fuel concentration).
of gas turbine burners. Providing additional mixing length between the fuel injec-
tors in the swirler and the flame is the most effective measure for the reduction of
the unmixedness of the fuel-air mixture of premix burners. Swirling flows producing
high turbulence have significant advantages over other types of flows without swirl.
In the turbulent free jet emerging from the burner exit, additional mixing takes
place that further reduces unmixedness upstream of the flame front (see Figure 10.4).
For natural gas combustion, this effect is essential for the moderate NO generation
of older burner designs that exhibit a high level of unmixedness in the burner exit
plane. However, it has been observed that this beneficial effect vanishes for fuels with
high reactivity when higher flame speeds lead to a shift of the heat release toward
the burner exit plane. At constant flame temperature, such burners show a substan-
tial increase in the NO emissions if reactive components are added to thefuel.
Figure 10.10 shows a comparison of the fuel concentrations of premix burn-
ers representing the status of 1990 and the actual technology, respectively. Different
scales have been selected to cover the high local maximum concentrations in the
instantaneous images without clipping. Only the mixing in the burner shown in the
upper part of Figure 10.10 approaches the well-mixed regime, whereas mixing in the
burner in the lower part of Figure 10.10 is much poorer.
A rather special method of improving fuel-air mixing is the seeding of the com-
bustor air with fuel far upstream of the burners. Because of the large mixing length,
excellent mixing is usually achieved, but great care must be taken that no zones with
combustible mixtures are formed that are exposed to large residence times, leading
to self-ignition of the fuel upstream of the combustor. In practice, safety consider-
ations limit the amount of fuel injected in that way to a few percent of the overall
316 Industrial Combustors
fuel flow. A further important aspect of applying this method is that it gives addi-
tional flexibility with regard to the suppression of thermoacoustic instabilities.
8
Tproducts
1500 K
8 6 1550 K
1600 K
Ignition delay time (ms)
1650 K
6 4
1700 K
4 2
0
2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Mole fraction products products
0
1100 1200 1300 1400 1500 1600 1700
Mixture temperature (products-reactants) (K)
Figure 10.11. Ignition delay times of fuel-air mixtures of 700 K mixed with their combustion
products at p = 2 MPa.
that the combustion chamber must already be operated in staged mode at higher
load, immediately after the potential of the variable compressor for the reduction of
the air mass flow rate has been fully exploited.
To highlight kinetic effects that contribute to burnout in fuel-staged operation
of premixed combustors, self-ignition of the fuel-air mixture in its own recircu-
lated combustion products is now considered for the adiabatic case. Self-ignition is
commonly perceived as a threat to the reliability of premixed combustors, but it has
also beneficial effects, promoting burnout in staged operation. In fact, self-ignition is
actually the basis of elegantly designed fuel-staging methods with low complexity of
the fuel supply system.
A direct link between the local concentration of the combustion products in the
mixture and the local mixture temperature exists because of the analogy between
the turbulent transport of energy and species. The local temperature increases with
the proportion of the combustion products in the mixture. In Figure 10.11, the
influences of the mixture temperature and the share of the combustion products
in the mixture on the ignition delay times are shown. These were calculated for
different flame temperatures using a plug flow reactor model and detailed meth-
ane chemistry. Interesting, when the results are plotted over the mixing tempera-
ture (Kalb and Sattelmayer, 2006), the results are almost wholly independent of
the flame temperature. The share of the combustion products required for a given
delay time decreases with increasing adiabatic flame temperature. As additional
318 Industrial Combustors
computations with hot air instead of hot combustion products delivered very sim-
ilar ignition delay times as a function of mixture temperature, it can be concluded
that the composition of the hot gas is of minor importance for self-ignition and that
the relationship between the ignition delay and the mixture temperature is very
similar for other temperatures of the air-fuel mixture in the typical range relevant
for gas turbines. For all mixtures with temperatures above about 12001300 K, the
ignition delay times are in the range of one millisecond, even for pure methane.
Since the mean residence time in combustion chambers is more than one order of
magnitude larger than this value, it can be assumed that such mixtures are fully
oxidized upstream of the turbine, independent of turbulent flame propagation. In
reality, the hot gas temperature must be at approximately 200 K above the limit
value for the mixing temperature derived from kinetics. The only regions poten-
tially leading to the emission of unburned hydrocarbons or combustion interme-
diates are streaks with temperature deficit in the combustor flow. At higher load,
such zones can be avoided by sufficient intensity of turbulent mixing in the com-
bustion chamber. With falling average temperatures in the combustion chamber, a
lower limit is reached where self-ignition of lean mixtures breaks down. Like other
phenomena governed by self-ignition, the lower limit for the breakdown of burn-
out depends on pressure.
For the part-load staging of premixed combustion chambers, a variety of tech-
nical solutions was found. These will be described later in the case studies in more
detail. However, all variants have common underlying principles, which are sche-
matically shown in Figure 10.12.
In principle, two essentially different approaches can be distinguished, depend-
ing on whether the air or the fuel distribution is controlled.
Using a bypass system for the air, the fuel-air ratio of burners can be kept in
the desired window. In the ideal case, the bypass air is injected downstream of the
CO burnout zone. If this can be accomplished, air staging provides ideal conditions
for low-emission operation over a wide load range. However, air staging leads to an
undesired decrease of the flow velocity and an increased tendency for flame flash-
back and self-ignition in the burners if the velocity drops too far below the full load
value. This imposes limits to the load range in which air staging in gas turbine com-
bustors can be employed and requires fuel staging in addition. Fuel staging alone
is often preferred, largely because of the high complexity of air staging with active
components exposed to high temperatures. The following three basic staging prin-
ciples are commonly applied:
Air staging
Fuel flow
Burner Combustor
Air flow & Hot gas flow
Fuel staging
Premix stage
Non-premixed / Pilot stage
Burner Combustor
Figure 10.12. Staging methods employed in combustion chambers with premix burners.
NO
0% 100%
100% 0% Power output
Increasing fuel flow to pilot
Figure 10.13. Emissions achieved with fuel staging in premix burners (dashed line: switching
between premixed and non-premixed operation, solid line: gradual fuel shift between pre-
mixed and non-premixed stage).
approached (vertical line in Figure 10.13). However, since this leads to an increase
of NOx emissions by about one order of magnitude, this staging strategy is applied
only if the switching point lies below the normal operating range of the gas turbine.
If, with decreasing global equivalence ratio, the fuel is not fully switched to the direct
injection but divided between the two injectors initially starting with a small fraction
fed to the second stage, the sudden sharp increase in NOx formation can be avoided
with careful tailoring of the fuel split versus load. This mixed operation is often
referred to as piloting of premixed flames. Interesting, full burnout can be achieved
in an intermediate range of the flame temperature, although the equivalence ratio
of the premixed flame becomes lower because of the partial direct injection of fuel
into the combustion chamber, and the fuel concentration in the mixture of the pre-
mix stage drops even further below the lean extinction limit. Since the lean mixture
320 Industrial Combustors
reacts in an afterburning mode stabilized by the hot gases of the non-premixed pilot
flame, an increase of the flame length is often observed. However, burnout does not
suffer if the combustor provides sufficient residence time. The reason for the onset
of the afterburning is the ignition kinetics of lean mixtures, which even ignite at
high pressures at temperatures much lower than the flame temperature at the lean
extinction limit (Figure 10.11). This effect is essential for the combustion technology
of gas turbines. The limit requiring to switch entirely to the direct injection is reached
when the afterburning process collapses and the burnout begins to become incom-
plete as the result of the further decrease of the equivalence ratio in the premixing
stage. This is the case when the temperature in the combustor downstream of the
main combustion zone drops below about 1400K in larger zones.
Piloting with very small fractions of the fuel fed to the pilot stage at high and
full load has also been applied because, in some cases, the dynamic stability of the
combustor benefited by this measure.
At part load, the inherent potential of the piloting method is not fully exploited
by non-premixed pilot stages because the non-premixed stage even at very lean
operation generates substantial NOx emissions. This problem has led to the devel-
opment of burners with premixed pilot stages. In burners of this type, the interaction
of two premixed flames with different equivalence ratios is used for flame stabiliza-
tion. The pilot stage is operated at its lean blowout limit to control the NOx emissions
stemming from the pilot fuel, and the rest of the fuel is fed to the main stage, where
a lean mixture beyond its lean blowout temperature reacts almost free of NOx gen-
eration because of the contact with the products of the pilot stage in an afterburning
mode.
Independent of the flame type generated by the pilot stage, a critical effect is
the undesired intense mixing of the fuel with the entire amount of air fed to the
burner. This does not allow the air overshoot to bypass the reaction zone. However,
bypassing the air overshoot would be very beneficial, because this would lead to
an increase of the effective temperature of the flame, leading to improved burnout
without adverse effects on the NOx emissions at part load. The observed limitations
stem from the concentration of fuel staging in the same burner that does not allow
for the exploitation of the full potential of the piloting method. Fortunately, this
intrinsic kind of air staging without moving parts can be better exploited if the inter-
action between burners is used for staging.
Combustion efficiency
100%
CO & UHC
0% 100%
NO Variable guide vane control
or
Compressor bleed
Figure 10.14. Emissions and burnout achieved by switching of groups of premix burners with
changing load.
equivalence ratio between a value close to the lean extinction limit of the flame and
a higher value necessary to complete the switching process without passing the lean
blowout limit of the burners in operation. Unfortunately, this leads to a periodic
increase and decrease of the nitrogen oxide emissions with the load accompanied
by an oscillation of the combustion efficiency with the NOx emissions at low load.
At low and mid-load, CO emissions are highest at the operating points with lowest
NOx emissions and vice versa (Figure 10.14). This corresponds with the classical NOx
and CO trade-off situation mentioned in Section 10.3.1. Since the emissions limits
have become very strict and usually do not have trade-off clauses for CO emission
versus NOx emission, these oscillations are clearly an inherent disadvantage of the
methods because the operating points with the highest NOx as well as the CO emis-
sions count.
Combustion efficiency
100%
CO & UHC
0% 100%
NO Variable guide vane control
or
Compressor bleed
Figure 10.15. Emissions and burnout achieved by piloting of groups of premix burners with
changing load.
emissions over the full load value (Figure 10.15). This lower temperature limit is usu-
ally reached in the medium load range. The deterioration of burnout below this limit is
strongly dependent on the mixing intensity in the combustion chamber. On one hand,
the intense mixing of the mixture of the piloted burners with the hot products from
the flames generated by the larger number of stably operated burners is beneficial for
burnout and CO emissions; on the other hand, intense mixing with the air stemming
from the burner with no fuel has to be avoided as far as possible. To achieve these con-
flicting goals at part load and idle, the heat release is usually concentrated in dedicated
zones of the combustion chamber to achieve a certain degree of intrinsic air staging by
keeping part of the excess air away from the combustion zone active at part load.
Longitudinal Fuel Staging. Commonly, in the latter two methods all burners are
arranged in one plane. In the case of the piloting of burners, therefore, the interac-
tion of the hot gases from the stably operated burners with the mixture of piloted
burners already begins directly in the primary zone downstream of the burner exit
zone and the undesired early mixing with the air of the burner not supplied with
fuel cannot be fully avoided. Longitudinal fuel staging of the combustion chamber
is beneficial with respect to the part-load emission characteristics. But longitudi-
nal staging requires air supply to both longitudinal stages, making combustors more
complex. Furthermore, each stage requires its own fuel injection systems and premix-
ing ducts. Since in low-load operation, only the upstream stage is in operation and
the second-stage air does not come into contact with the flame, improvements can
be achieved. However, the longitudinally staged process also requires intense mixing
of the very lean mixtures injected in the second stage at part load with the hot main
flow from the first stage to accomplish afterburning effectively. Combustors with
longitudinal staging are longer than designs with all burners arranged in one plane.
with split turbines (see Section 10.5.3.2) and with two combustors in series fall into
this class with the difference that the enthalpy of the exhaust gas of the first combus-
tor is reduced by partial expansion and, therefore, primarily fuel is injected in the
second combustor.
10.5Case Studies
Beginning with the first utility gas turbine in the 1940s, non-premixed combustion
was the prevailing method employed in gas turbine combustors until the mid-1980s.
After this point, NOx emissions reduction gradually became the dominating crite-
rion for the evolution of gas turbine combustors.
Combustion Retractable
cover spark Combustion Combustion Turbine
Flow
plug liner wrapper shell
sleeve
Combustion
air Film
cooling
air
Fuel nozzle
Combustion Compressor Transition
Support
casing discharge casing clamp piece
Combustion liner
Body
Cap
Water injection connection
Cowl
Nozzle body
Gas fuel jet
Atomizing
air passage
Inner oil tip
Fuel oil
connection Air swirler
Water spray
nozzle
(1 of 12)
Atomizing air connection
Gas connection
Figure 10.17. Cross-section of the dual-fuel nozzle of the General Electric standard
non-premixed combustor.
Figure 10.17 shows a typical dual-fuel standard nozzle assembly with fuel gas
jets in the axial swirler and a central single-circuit liquid fuel injector. Like virtually
all GE Energy liquid injectors, the design shown is an air-assisted pressure atomizer
with relatively large oil passages and atomizing air supplied by an independent atom-
izing air compressor at a pressure 40 percent above the compressor discharge level.
The atomizing air is primarily for creating an ignitable spray at firing speed in spite
of the large oil passages and minimal pressure drop in the nozzle tip. The liquid fuel
injectors rely on pressure atomization for operation at load. The fuel nozzle assem-
bly shown also has an array of smaller pressure atomizers just upstream of the axial
air swirler inlet to provide finely atomized water for emissions control and power
10.5 Case Studies 325
Dilution
Figure 10.18. General Electric standard diffusion flame combustor combustion cap and liner
assembly.
augmentation. The water is carried into the reaction zone along with the compressor
discharge air. Inlet flanges for the fuel gases, atomizing air supply, and connections for
the liquid fuel and water are on the forward side of the fuel nozzle assembly.
The combustion cap and liner assembly is shown in Figure 10.18. The combustors
are provided with a reaction zone length of about a meter, and diameters between
250 and 350 mm. Residence times in the liner at load are at least thirty milliseconds.
In addition to the recirculation zone created by the axial swirler, one or more rows
of primary air jets are provided to help stabilize the flame. Including the airflow from
the axial swirler, the conical cap cowling louvers, and the primary jets, the primary
zone fuel-air ratio is close to stoichiometric. Large dilution holes near the exit of
the liner mix additional air in to produce the desired combustor exit temperature
and radial profile. If the system is provided with additional wrapper steam for
power augmentation, the steam is distributed throughout the combustor with the
air flowing through all the metering features. Cross-fire tubes provide a source of
hot combustion products for ignition of adjacent combustion chambers after a spark
ignition system initiates the process in two of the chambers. Although the circum-
ference at the combustion chambers can be more than thirty meters, the process of
cross-firing takes less than 600 msec, as the pressure differential between fired and
unfired chambers completes the process quickly.
Liners are constructed from alloys that demonstrate good high-temperature
strength and resistance to creep and oxidation, such as Hast-X, formed from sheet
metal. The liners are protected from hot combustion products by film cooling, either
by an array of louvers along the entire length in the older combustors or by a series
of film-cooling skirts in the later designs. Ceramic thermal barrier coating provides
additional thermal protection. Usually the coating consists of yittria-stabilized zirco-
nia topcoat with a strain-compliant metallic bond coat.
In the 1970s, tighter air pollution standards established a need for reducing the
NOx levels from 42 ppm toward 25 ppm in combustion turbines employed for power
generation. The single-nozzle combustors could reach 25 ppm NOx with increased
326 Industrial Combustors
Figure 10.19. Liner and cap assembly of the General Electric MNQC combustor.
steam or water injection, but the resulting combustion instabilities shortened the
inspection intervals and reduced reliability. Carbon monoxide limits were also
exceeded when operating at reduced load and at cooler ambient conditions.
To improve burnout and stability, the Multi-Nozzle Quiet Combustor (MNQC)
system was developed. Better distribution of the fuel, air, and diluent over more of
the available combustor volume was used to improve the combustion efficiency and
produce a more compact reaction zone. The single fuel injector per combustor was
replaced by an array of six smaller nozzles arranged around the circumference of
the combustor, with fuel, air, and diluent supplied equally. As expected, this nozzle
arrangement provided more effective use of the injected steam or water diluent,
reducing the NOx levels toward 25 ppm without excessive CO or combustion noise.
The resulting cap and liner are shown in Figure 10.19 and one of the six gas tips
in Figure 10.20.
The gas tip design on the MNQC borrowed the axial swirler, gas injection geom-
etry, and center oil tip design from the earlier single-nozzle architecture, scaled down
to the multi-nozzle requirements. The smaller Frame 3, 5, and 6 (roughly 250 MM
diameter) combustors used an arrangement of five similar gas tips, reduced in diam-
eter but otherwise very similar to the original MNQC parts. The resulting combustor
was indeed quiet, reducing dynamic pressure amplitudes from several psi peak-to-
peak by a factor of ten, to tenths of a psi when operating in emissions compliance
with water-fuel ratios> 1 for 25 ppm NOx. Carbon monoxide and the available load
and ambient range for emissions compliance were also expanded for both gas and
liquid fuels.
Operating at higher firing temperatures and pressure ratios of the initial F-Class
gas turbine, the transition piece (TP) cooling had to be improved to prevent a reduc-
tion in inspection interval. The use of impinging jets to enhance convection was a
significant innovation in the design of the first F-Class combustion systems, and con-
tinues in the DLN 2.6+ and all other GE DLN systems today (see Section 10.5.4).
The compressor discharge air is directed through an impingement sleeve surround-
ing the transition piece, taking a pressure drop of 1 to 2 percent relative to the com-
pressor discharge pressure (Figure 10.21). Once directed against the cold side of the
TP body, significantly multiplying the heat transfer coefficients, the air flows forward
under the sleeve and into the flow sleeve surrounding the liner. The compressor
10.5 Case Studies 327
discharge air used for impingement cooling the TP then produces further convective
cooling of the liner before being introduced into the combustor reaction zone and
after gaining 25 to 80C in temperature. This cooling scheme allowed the use of film
cooling or transpiration cooling of the TP body to be minimized, eventually allowing
a majority of the combustor through-flow to be available for premixing in the later
lean-premixed, dry low NOx architecture.
We next consider Alstom designs. The first industrial gas turbine put into oper-
ation in 1939 by BBC Brown Boveri (now Alstom) has one single horizontally posi-
tioned silo combustor mounted on top of the turboset (Figure 10.22).
328 Industrial Combustors
Single combustor
Starting motor
Generator Compressor Gas turbine
Figure 10.22. First industrial gas turbine for electricity generation (BBC Brown Boveri).
Burner
Swirl body
Tiles
Mixing
nozzle
Liner
sheet
Figure 10.23. GT13E gas turbine from Alstom with a traditional silo combustor.
Later, the silo combustor design evolved into a vertical geometry that stands
on top of the gas turbine. One example of this architecture is shown in Figure 10.23
(Alstom GT13E gas turbine).
Non-premixed silo combustors consist of one large single burner with a large
swirler and a fuel lance on the centerline (Kenyon and Fluck, 2005). This burner-lance
assembly is mounted on the front panel of a hot gas casing (flame tube). Air from
a compressor flows first through an annulus between the outer cylindrical pressure
vessel and the inner hot gas casing, thereby cooling the hot gas casing before enter-
ing the swirler, the primary zone cooling, or the mixing nozzles.
In the primary zone of the combustor with the highest heat loads, the flame tube
consists of metal cooling elements (tiles) with finned back sides. First the cooling
air convectively cools the segments before it enters the combustor along the walls,
providing film cooling of the downstream wall segments. The ratio between swirler
air and mixing air is selected so that optimum soot burnout is achieved and that the
NO2 formation (see Section 10.3) at part load is minimized.
As can be seen from Figure 10.24, gaseous or liquid fuel is injected through a
central lance at the downstream end of the swirlers directly into the recirculation
zone formed by the swirling airflow. The second small swirler in the center provides
additional flexibility for tuning the flow field in the combustor. To achieve enough
fuel penetration, gaseous fuel is injected through multiple holes on the outer radius
of fuel lance, directed outward in radial direction. Liquid fuel is injected through a
10.5 Case Studies 329
Air channels
Figure 10.24. Typical non-premixed burner from an Alstom silo combustor (GT11N2).
central pressure atomizer as a hollow cone in axial direction. A large flame is estab-
lished immediately downstream of the swirler, stabilized through the recirculation
of hot gases. The flame is anchored and stabilized in the recirculation zone and the
swirler acts as a flame holder. While the fuel is mixed with the air and reaction prod-
ucts, the fuel burns as a typical non-premixed flame. Such diffusion flames usually
burn in the zones with near stoichiometric air-to-fuel ratio ( =1) with a relatively
high flame temperature, leading to high thermal NOx formation. To meet emission
requirements, steam or water injection through a number of nozzles between swirl
blades is used for NOx abatement. These nozzles can be used for power augmenta-
tion as well.
Such non-premixed burners can burn a wide range of fuels such as natural gas,
liquefied natural gas, heavy fuel oil, crude oil, distillate, and naphtha. If required,
steam or water injection is employed for NOx control. The burners have proved a
robust solution for the combustion of crude and residual oil, even with high ash
content. Because of the large sizes of all burner elements, clogging and coking of
injection holes is avoided.
An interesting early approach (1985) for reducing the NO formation of
non-premixed flames is the very compact combustor shown in Figure 10.25. The
basic principle is equivalent to the lean direct injection method, which is becoming
an important technique for low NOx aero engines.
Before emission limits were introduced, gas turbine operators were accustomed
to gas turbines with simple fuel systems and operational concepts. The combustor
shown in Figure 10.25 preserves this simplicity of early engines without NOx con-
trol. This is achieved by the unstaged operation throughout the entire load range.
The basic principle for NOx control employed is the reduction of the flame size and
the minimization of the residence times associated with the distribution of the heat
release to many very small flames (384 for a 60 MW engine). The air at the combus-
tor inlet flows either through one of the many swirlers or through the channels in
between them. The swirlers have fuel injectors in their centers and generate small
recirculation zones with fuel overshoot for flame stabilization. The hot combustion
products are then mixed with the air flowing through the axial channels to achieve
330 Industrial Combustors
10 1 2 3 9 4 7 5 8 6
full burnout. Because of the air overshoot, the temperature drops and NOx forma-
tion is reduced.
Although this type of combustor showed an NOx reduction potential of a fac-
tor of approximately five with respect to the NOx emissions of non-premixed silo
combustors, such as the example shown in Figure 10.23, the commercial success of
this elegant solution was limited because the NOx emission limits dropped quickly
below the NOx abatement potential of non-premixed systems. A drawback of this
unstaged solution was the strong quenching of the burnout reactions due to the large
air overshoot in the flame zones at part load with reduced turbine inlet tempera-
tures. However, full burnout was reached in the load range covered by inlet guide
vane control of the compressor (see Section 10.4).
between the burner outlet and the flame front because of the high turbulence in the
annular swirling jet produced by the conical swirler. This principle, similar to the
direct lean injection method currently pursued for low-emission aero engine com-
bustors, allows safe operation even with highly reactive fuels. With respect to the
standard EV Burner for premixed natural gas operation, the MBtu Burner has been
shortened to reduce the swirl number and to increase the air velocity in the annular
swirling jet at the exit of the burner. Both modifications increase the standoff dis-
tance of the flame from the burner exit and reduce the thermal load on the burner
exit section. The high air speed, along with thorough premixing, minimizes NOx emis-
sions and improves the long-term thermal integrity of the burner. This leads to a very
robust and reliable solution. For syngas with heating values of less than 78 MJ/kg,
almost fully premixed conditions are reached. Syngas with higher heating values
require dilution to this limit. Avoiding excessive degrees of dilution or even omitting
dilution omits air extraction from the compressor, which substantially complicates
gas turbine and the power plant topologies.
Since the aerodynamics of the standard EV Burner have been preserved in
designing the MBtu Burner, additional operation on backup fuels like oil or natural
gas is also permitted.
A combustion technology of Siemens for syngas is based on the Siemens Hybrid
Burner (see Section 10.5.2), which has been modified for the combustion of fuels
with low heating values and high reactivity; see Figure 10.27. Since syngas is not
premixed with the air, it must be diluted with nitrogen and/or steam for low NOx
operation.
As discussed in Section 10.2.1.1, the Wobbe Index of undiluted syngas from
oxygen blow processes is much smaller than the Wobbe Index of natural gas. The
low Wobbe Index of the fuel and the additional flow of diluent result in a massive
increase of the capacity of the fuel lines. To provide this extra capacity, an extra
syngas passage surrounds the pilot air passage. Syngas fed through this fuel line
is injected between the main air passage and the pilot air passage (Figure 10.28).
Moreover, syngas may also be fed through the diffusion gas line.
General Electric solutions for syngas and blast furnace gas are detailed by Hall
and colleagues (2011) and Payrhuber and colleagues (2008). The syngas combustors
332 Industrial Combustors
Fuel oil
(diffusion)
Fuel oil
Air Syngas 1
Natural gas
(premix) Fuel oil
(premix)
Figure 10.27. Comparison of the Siemens hybrid burner designs for natural gas and syngas
operation.
Air passage
Central swirler
passage
Syngas fuel
passge
Figure 10.28. Hybrid burner for syngas, natural gas, and fuel oil operation.
are a diffusion flame derivative of the MNQC, first fielded in the late 1970s and
described in Section 10.5.1.1. Like the turbine nozzle, the fuel injectors have to be
redesigned to handle the much higher volumetric flow of the synthetic fuel gas and
the nitrogen diluent used to suppress NOx. For the initial startup and shutdown, the
system must also operate reliably on natural gas. The syngas fuel injectors divide the
fuel and oxidizer into separate arrays of discrete jets that merge and converge at
theexit of the fuel nozzle to provide fast large-scale mixing for quick burnout. Direct
fuel injection is also provided at the center of each gas tip. The system can also be
modified to use diesel fuel as a backup source.
flow due to the significantly reduced heating value of LBtu gases, the combustor also
has to be significantly modified. A typical LBtu combustor (Alstom GT11N2 gas
turbine) is shown in Figure 10.29.
The main changes of the combustor geometry compared to natural gas engines
are the increased combustor volume required to compensate for the reduced reac-
tivity and a burner that can accommodate approximately ten times higher fuel flow
than the natural gas burner. Details of the LBtu burner are shown in Figure 10.30
and in Kenyon and Fluck (2005). This burner contains a swirler with alternating
passages for air and LBtu gas. Additionally, the LBtu gas passages are divided into
inner and outer channels, thereby providing staging opportunities. This geometry
provides a large recirculation zone for intense flame stabilization and leads to opti-
mum initial conditions for the large-scale mixing of the combustion air with the
high-volume flow of LBtu gas. The latter limits the axial extension of fuel-rich zones,
334 Industrial Combustors
Natural gas
Air
LBtu inner channel
despite the slow mixing of the air with the fuel. This slow mixing is a character-
istic of parallel flow configurations, as in the swirler shown in Figure 10.30. Since
the maximum flame temperature in non-premixed combustion of LBtu gases is
naturally restricted to low values, even single-digit NOx emissions can be achieved.
Usually, requirements can be fulfilled without any dilution or other NOx reduction
measures.
In addition, this design allows mixed operation with either natural gas or oil,
both of which are injected through a central injector. Usually, during mixed opera-
tion, some form of NOx abatement is needed because of higher NOx values stem-
ming from the higher flame temperatures.
The application of General Electrics syngas combustion systems to burn very
low Btu content blast furnace gas (BFG) derived from the steel-making process is
explained in Hall et al. (2011). Figure 10.31 shows predictions of the fuel and oxi-
dizer concentration in this unique combustor. The high fuel volume flow is injected
through the gas nozzles located in the combustor cap. The plot illustrates how the air
admitted though the liner participates in combustion and that flame length is gov-
erned by admixing of enough air to the fuel until locally lean conditions are reached
over the entire cross-section of the liner. The Frame E BFG gas turbine power train
developed by General Electric has been extended to operate with process gas fuels
down to approximately 4 MJ/kg.
Silo combustors are predominantly applied to gas turbines with moderate firing tem-
peratures that permit enough cooling air for the geometrically complex transition
section between combustor and turbine inlet. An advantage of such silo combustors
10.5 Case Studies 335
%O2
=1 21
15
10
Figure 10.32. Alstoms first-generation premix silo combustor (left: basic configuration, right:
burner section).
is their high robustness resulting in very high reliability. The hot reaction zones can
be kept away from the combustor walls. To achieve low NOx operation, either the
combustion temperatures are kept moderate when operating in premix mode or
water is injected when operating in diffusion mode.
The first generation of Alstoms premixed silo combustors consisted of a large
number of small premix burners, replacing the single swirler of non-premixed com-
bustors, as shown in Figure 10.32. These first-generation premix burners have long
mixing tubes, each with a central fuel lance and a swirler at the exit. In premix mode,
the fuel is injected several burner diameters upstream of the swirler from the fuel
lance radially outward into the combustion air, thus achieving a significant improve-
ment in fuel-air mixing quality. Liquid fuel is prevaporized in the tubes. Direct
non-premixed fuel injection downstream of the swirler is also permitted.
Only at high and full load are all burners supplied with fuel in premixed opera-
tion. At low load, the fuel flow to groups of burners is sequentially shut off in accor-
dance with the fuel staging principle shown in Figure 10.14.
336 Industrial Combustors
Gas
Gas
Figure 10.34. Alstom GT11 silo combustor with multiple premix burners.
A close-up view toward the exit plane of these premix burners from downstream
shows the hexagonal burner matrix with the swirler. This technology enabled the
reduction of the NOx emissions of non-premixed silo combustors from several hun-
dred to below 75 vppm. The same burners could be used for gaseous and liquid fuels,
including water or steam injection for emission reduction or power augmentation.
The initial idea to use the long tubes also for the prevaporization of liquid fuel at
moderate compressor exit temperatures of machines with low pressure ratios could
not be realized, primarily because deposits were observed in the mixing section and
on the swirler during the development of the system.
10.5 Case Studies 337
Figure 10.35. Operation of an Alstom premix silo combustor from idle to base load.
Fuel oil
(diffusion)
Natural gas
(diffusion)
Air
Air Air
premix groups, cover the complete load range from ignition to base load. Usually,
six burners are connected to a separate fuel line that serves as a second premix
group. Liquid fuel operation relies on the same burner switching and grouping as
gas operation.
The Siemens SGT52000E series engines utilize two silo combustors, and its
scaled version SGT62000E for the 60 Hz market. The SGT52000E and the SGT6
2000E are equipped with one large vertical silo combustors on each side of the tur-
boset. These silo combustors are equipped with Siemens hybrid burners, shown in
Figure 10.36.
The hybrid burner got its name from its ability to burn fuels in premixed and
non-premixed (diffusion) mode. When required, the diffusion mode can be extended
up to base load of the silo combustors. The hybrid burner consists of three main com-
ponents concentrically arranged with respect to each other. The central part con-
sists of the fuel oil lance in the center, which is surrounded by the pilot air passage
through which about 10 percent of the burner air is discharged into the combustor.
The fuel gas supplied by the diffusion gas fuel line is injected into this air passage
by jets perpendicular to the airflow. The pilot air passage is confined by the main
air passage with a diagonal swirler. Natural gas injected into this flow path flows
through the hollow swirler vanes. The aerodynamic design of the swirl vanes and the
injector provides flashback resistance in the premix mode in the subsequent air/fuel
premixing passage.
Although originally developed for silo combustors, this hybrid burner is also
widely used in annular combustors (see Section 10.5.3.1) and has continuously
evolved for almost over two decades.
Engines with hybrid burners are started in diffusion mode. In natural gas opera-
tion, the fuel to the diffusion line is closed at an intermediate load level while fuel
is fed to the premix fuel lines. Premixed operation is also possible for liquid fuels. If
10.5 Case Studies 339
Hybrid
burner
required, the burner can operate in diffusion mode up to base load, providing the
opportunity for adapting the combustion system to a wide range of fuels.
High fuel flexibility is one of the most attractive features of silo combustors. In
this framework, the large residence time is an additional important aspect, along
with the flexibility of the burner. Fuels ranging from coke oven gas with very low
Wobbe Indices (see Figure 10.1) to LNG fuels with a very high Wobbe Index of
almost 50 MJ/m, as well as all different types of liquid fuels including heavy distil-
lates, can be burned in silo combustors.
10.5.3 Premixed Annular Combustors for Natural Gas and Liquid Fuel
with or without Water Injection
The main driver behind the transition from silo to annular combustors was increased
turbine inlet temperature, as well the need for more primary air to further reduce
the NOx emissions of premixed combustors. Silo combustors have a long residence
time and neither the flow pattern nor the temperature field near the turbine inlet
are fully homogeneous because of the flow curvature upstream of the turbine. As
shown in Figure 10.8, the post-flame residence time increases NOx formation, espe-
cially at elevated flame temperatures. Typical annular combustors have average resi-
dence times of approximately 25 ms. Such residence times are also achieved with
can-annular systems as explained in Section 10.5.4.
twenty-four large recirculation zones are generated downstream of the burner exits
by the co-swirling burner flows, which provide flame stabilization around the circum-
ference over a wide operation range.
The hybrid burner used in the annular combustor is derived from the older ver-
sion used in silo combustors (see Section 10.5.2). The further evolution of the burner
design is described by Strebe and Prade (2001). To further reduce the NOx emis-
sions, fuel air mixing in the burner was considerably improved over the previous
silo design. In Figure 10.38, the main development step toward better premixing is
highlighted and is the introduction of the premixed pilot stage. On the left, a hybrid
burner with diffusion pilot is shown for comparison.
The modified burner consists of the same three main components (see Section
10.5.2). The locally concentrated injection of the pilot fuel in the non-premixed
pilot results in fuel-rich concentration peaks at the outlet of the burner passage. A
much better distribution of the fuel is achieved by introducing the fuel through the
vanes in the main stage and the pilot burner, too, as indicated on the right of Figure
10.38. Injection fuel over several holes at different radial positions distributes the
fuel much better over the pilot burner passage than in the previous diffusion-type
design.
In addition, further design efforts were made to optimize the fuel concentration
profile of the main burner passage. A result of a computer simulation of fuel air mix-
ing is shown in Figure 10.39. The trajectories of the fuel injected through the swirler
vanes of the main burner passage are illustrated, showing mixing in the main burner
passage. In addition, the resulting concentration profile at the exit of the burner pas-
sage is shown. The calculation reveals good mixing near the hub of the annulus and
higher degrees of unmixedness is visible near the shroud. This unmixedness is caused
by the limited mixing length and is further reduced further downstream toward the
flame front. The improved fuel-air mixing leads to a drop in the NOx emissions at a
given firing temperature, or alternatively, enables an increase of firing temperature
without deterioration of the emissions.
10.5 Case Studies 341
Figure 10.39. Fuel concentration profiles for the hybrid burner in premix operation.
Oil jets
premixed stage
Figure 10.40 illustrates the liquid fuel injection methods employed in the
Siemens hybrid burner. The hybrid burner has two separate stages for liquid fuel.
In the diffusion stage, fuel oil enters the lance through an annulus, enters the swirl
chamber of a pressure atomizer, and is then divided in the flow, leaving the nozzle
and the returning flow. The reverse flow control ensures a minimum flow through
the swirl chamber even at low engine load. A minimum flow rate through the swirl-
ing chamber is needed to provide a stable swirling flow in the atomizer, leading to
sufficient atomization quality under all operating conditions. To provide better fuel
distribution, additional injection nozzles are mounted at the hub of the premix pas-
sage. When this premix stage is switched on at higher loads, lower NOx emissions
342 Industrial Combustors
Carrier structure
Purge air
Metallic tile holder
Figure 10.41. Wall structure of the annular combustor.
Figure 10.42. Outer shell of the annular combustor with ceramic heat shields.
are achieved because of better fuel-air distribution. However, since complete evap-
oration of the fuel oil and its full dispersion in the airflow are not achieved, further
reduction in NOx emissions requires the addition of water to the liquid fuel and the
discharge of emulsion through the nozzles. Emulsions with a water content of 80
percent typically reduce NOx emissions by a factor of five compared with operating
the system without water.
Hybrid burners for annular combustors are available with a variety of lengths
of the cylindrical outer wall of the main swirl register. These extensions (cylindrical
burner outlet, CBO) allow the selection of the convective time lag of the premixed
gases from the injector to the flame zone. This parameter is the most essential mea-
sure for the optimization of the thermoacoustic stability of the combustor. Adding
CBOs to burners moves the heat release further downstream and the heat release
distribution becomes broader. Proper tuning of the mean time lag to the most crit-
ical instability and wider time lag distributions lead to an extension of the stable
operation range.
Lower NOx emissions over the entire operation range are achieved by an opti-
mized use of the air from the compressor, with the aim of maximizing the fraction
10.5 Case Studies 343
Control valve
Gas
of air that enters the combustor through the burners. Consequently, the adiabatic
flame temperature will drop and approach the turbine inlet temperature as closely as
possible. To this end, considerable efforts were made to reduce the air consumption
for wall cooling or purging gaps to a minimum. In modern combustors, the cooling
air consumption has been reduced to 3 to 8 percent of the compressor inlet mass
flow. The SGT54000F uses ceramic heat shields to protect the sustaining metal-
lic structure from hot combustion gases. Since the ceramic heat shields are made
of Al2O3 and can sustain temperatures of at least 1500C, they do not need to be
cooled. However, each tile is connected to the metallic support structure by metallic
clamps (tile holder), and these require some cooling with purge air as depicted in
Figure 10.41.
The design of the entire wall structure has been carefully optimized with regard
to minimum cooling air requirements for the clamps and also for the prevention of
the ingestion of combustion products in the gaps between the ceramic heat shields
and the carrier structure (Krebs et al., 1999). The outer shell of the annular combus-
tor is illustrated in Figure 10.42, showing the combustor front panel segments with
the orifices for the hybrid burners. The outer combustor wall is equipped with four
rows of ceramic tiles, followed by a row of curved metallic heat shields that guide the
flow toward the turbine. Where the ceramic heat shields have not yet been mounted,
the metallic support structure remains visible.
The Alstom EV burner introduced in Section 10.5.2 has also been used in annu-
lar combustors. The GT13E2 gas turbine was the first engine with this combina-
tion (Senior et al., 1993). Annular combustors provide the optimum conditions for
low-emission combustion with premix burners. The very compact designs associated
with annular combustors reduce the flame tube surface to be cooled and the cost of
the combustor compared to silo combustors. The GT13E2 annular combustor has
seventy-two EV premix burners arranged in two staggered rows.
The burners of the GT13E2 combustor are divided into two groups the main
and the lean burner group, as shown in Figure 10.43 (Dbbeling et al., 2007). The
344 Industrial Combustors
107 103
105
111 109
108 101
106 104 611
201 112 110
102
203 612
202 609
204
610 607
205
608
206
207 605
208 606
209 603
210 604
Figure 10.44. Distribution of premix gas groups in the GT13E2 gas turbine.
distribution of burners in these two groups is shown in Figure 10.44. At high and
base load, tuning of the fuel spit between both groups allows optimization of com-
bustion with respect to emissions and dynamic stability. At part load, the fuel flow
to the smaller group is gradually reduced to keep the flame temperature of the large
group above the lean extinction limit.
In addition to external staging via two premix groups, EV burners in GT13E2
utilize internal diffusion piloting, which is needed to cover the operation range from
ignition to about 50 percent load. Additionally, the fuel distribution system is divided
into six sectors, which also enables switching off some sectors after ignition to cope
better with transients during startup and acceleration.
Liquid fuel operation uses a similar operation concept to gas with the exception
that no internal piloting is required because of significantly lower lean blow out limit
with oil.
Further details of the GT13E2 combustor are illustrated in Figure 10.45. Air
from the compressor is divided into two separate flow paths, one flowing through a
combustor hood cooling the outer combustor wall, and a second cooling the inner
wall and the first vane row. Combustor walls are primarily convectively cooled. The
cooling air flows into a plenum upstream of the burner section, and then through the
EV burners into the combustor.
In the 1990s, dry low NOx operation with light oil became increasingly attractive
for smaller gas turbines (3050 MW class). This resulted in the development of the
AEV burner with a cylindrical mixing tube, based on the aerodynamic principle of
the Alstom EV burner (Figure 10.46).
10.5 Case Studies 345
GT casing
Cooling
shirt Zone 1 EV burners
Zone 2
Vane 1 Front segments
Combustor hood
Compressor airflow
Temperature 1600
Axial velocity
1700
(heat release) 1300
900
T (K)
1000 1700
0.0 1.0 2.0 3.0
1500
Oil
nozzle Gas
Air inlet Combustor
injector
The swirler with four slots generates a swirling flow with a strong peak on the
axis for flashback prevention in the critical center of the flow. The swirling flow
travels downstream and breaks down in the combustor, there forming the desired
recirculation zone required for flame stabilization. The most important achievement
was true low NOx operation on engines with almost twenty bar combustor pressure
346 Industrial Combustors
LP Turbine Compressor
HP Turbine
Figure 10.47. Sequential combustion system of GT24 and GT26 gas turbines.
EV lance
EV front segment
EV inner liner EV inner segment
Figure 10.48. Alstom GT24 and GT26 high-pressure combustors.
for oil without water injection, although liquid fuels are injected only through one
single nozzle at the apex of the conical swirler. By proper tailoring of the oil momen-
tum at the nozzle exit, full fuel evaporation and almost perfect radial dispersion of
the fuel vapor is achieved in the mixing tube upstream of the burner exit. As a bene-
ficial side effect, the long distance between the swirler with the gas injectors and the
flame also reduces mixture fluctuations to a very low level for gaseous fuels. AEV
burners are employed in Siemens gas turbine models SGT-700 and SGT-800. Both
engines are equipped with a single row of burners equally distributed around the
circumference of the combustor.
The EV combustors of GT24 and GT26 gas turbines utilize internal staging fea-
tures (staged premix) whereby each burner is connected to two separate fuel sup-
ply lines. This way, it is possible to operate the burners in a wide operation window
in full premix mode without relying on diffusion-type internal pilots. Both premix
stages are continuously in operation, from ignition to base load. Additionally, exter-
nal piloting can be carried out with separate burner groups.
The SEV combustor does not require any internal staging or piloting features.
However, external staging with fixed or variable groups is also possible.
Liquid fuel operation does not require an internal pilot feature, neither in EV
nor in SEV combustors. External staging can be employed with fixed or variable
groups.
The EV combustor utilizes a single row of retractable EV burners installed
through a front plate, as seen in Figure 10.48. Because of the higher combustion pres-
sure, the number of burners is substantially lower than in non-reheat engines. Both
inner and outer walls of the EV combustor consist of convectively cooled segments
installed on support structures. This segmented structure with cold carrier parts allows
complete convective cooling with minimum leakages. The air enters the burner ple-
num from inner and outer liner cooling passages and flows to the burners.
The SEV combustor has a similar structure to the EV combustor, as seen in
Figure 10.49. The SEV burners contain four tetrahedral vortex generators, similar to
delta wings familiar from aeronautical applications (Eroglu et al., 2001). In each of
these burners, the vortex generators introduce four pairs of streamwise vortices on
the hot gas flow from the high-pressure turbine. Gaseous or liquid fuel is injected
with carrier air through a retractable lance into the center of the four vortex pairs.
This permits effective large-scale distribution of the fuel over the entire cross-section
of the flow.
The cooling scheme of the SEV combustor is similar to that of the EV combus-
tor with convectively cooled segments connected to support structures that form
annular cooling passages. The air from inner and outer cooling passages enters a
plenum around the SEV burners. These are equipped with effusion cooling fed from
the cooling air from the plenum.
348 Industrial Combustors
Partially Fully
premixed premixed
Low NOx/CO
Fuel staging
Figure 10.50. General Electric DLN 1 fuel staging and combustion modes.
Primary FNs
Secondary FN Dilution
Venturi
Cap/Centerbody Liner
TP
Figure 10.51. General Electric DLN 1 combustor architecture.
of air from mixing holes on the liner. This interaction between the primary fuel air
mixture and the primary jets causes premixing in the primary zone.
Further downstream is the main combustion zone. In premixed mode, two dis-
tinct flames form the combustion zone. One flame is anchored behind the venturi
and burns the fuel that comes from the primary zone, whereas the other flame is
anchored on the secondary fuel nozzle in the middle of the centerbody. Typical base-
load operation has ~ 80 percent of the fuel supplied to the primary zone and the
remaining 20 percent through the secondary nozzle.
When the head-end primary zone is operating as a nonreactive gas premixer at
load, the bulk fuel-air mixture velocity across the narrow throat formed between
the inner edge of the venturi and the outer edge of the adjacent centerbody is main-
tained well above the turbulent flame speed, preventing the associated premixed
flame from propagating upstream.
The secondary nozzle in the middle of the centerbody also has a sub-pilot fuel
circuit at its tip that stabilizes the flame. Typically, in the DLN 1 configuration, the
fuel flow through the sub-pilot circuit is a fixed percentage of the fuel flow through
the secondary fuel nozzle circuit. The sub-pilot flame is a diffusion flame and, hence,
produces significant levels of NOx per unit fuel flow. However, the sub-pilot also pro-
vides a stabilizing feature for the secondary flame and suppresses dynamics ampli-
tudes. But not all operating conditions require the same amount of sub-pilot fuel for
combustion stability. Therefore, controlling the sub-pilot fuel to the amount required
for combustion stability can provide the benefit of low combustion dynamics for a
minimal increase in NOx. This balance in independently controlled sub-pilot gas cir-
cuit flow is implemented in the DLN 1+ configuration (Thomas et al., 2011).
Following the operating mode illustration in Figure 10.50, the annular DLN 1
primary stage operates as a diffusion flame combustor from firing speed (ignition) up
to 20 percent load, at which point fuel gas is introduced into the secondary premixer,
a second annular cavity within the centerbody at the forward end of the liner. The
second stage gas is injected from the secondary pegs. It varies over the load range up
to about 20 percent of the total, mixes with the compressor discharge air flowing in
the annulus formed by the outer surface of the secondary fuel nozzle and the inner
surface of the centerbody. The tip end of the secondary nozzle protrudes through
350 Industrial Combustors
Figure 10.52. Upstream view through the venturi into the primary zone of the DLN 1 com-
bustor with the second-stage injector in the center.
a large axial swirler mounted to the cap centerbody (Figure 10.52). The premixed
fuel-air charge is discharged through the axial swirler into a cavity at the down-
stream end of the centerbody. The cavity and the two swirlers, visible in Figure10.52,
provide a stable anchor point for the flame associated with the secondary nozzle.
The DLN1 combustor operates in a lean-lean mode, with a diffusion flame in the
primary zone, and a partially premixed flame anchored to the central premixer, up
to about 50 percent GT load.
The premixing approach of the DLN 1 primary zone has the advantage of being
able to tolerate flashback or primary stage reignition events. Since the primary sec-
tion has been designed with film cooling and thermal protection to tolerate flame
temperatures when operating in the initial primary-only mode and the lean-lean
modes, the result is simply a brief period of higher emissions. The combustion system
is then returned to premixed mode by momentarily turning off the fuel to the pri-
mary gas tips, operating the machine only on the partially premixed central second-
ary zone in each chamber. Fuel flow is then restored to the primary and premixed
operation in emissions compliance resumes.
Cap a
nd pr
tube em
assem ixer
bly
X-fire
Premix
chamber Compressor
discharge casing Impingement-cooled
transition piece
Figure 10.53. General Electric DLN 2.0 combustion system.
lean-premixed systems. However, the amount of cooled liner surface area and film
cooling required for adequate durability in the initial DLN 1 scheme limited the
amount of air available for premixing with the fuel, and the NOx entitlement. The
film cooling and secondary flows also became a carbon monoxide liability at part
load. Although DLN 1 combustion systems were developed and fielded for the 7F
machines as an interim measure, a second step in the DLN evolution was made
to balance the emissions and efficiency expectations with acceptable durability.
Reduced cooled surface area, reduced film cooling, revised materials and coatings,
and a new approach to fuel gas premixing and staging are employed in the DLN 2
system for the 7F and 9F turbines (Lovett and Mick, 1995; Davis, 1996; Vandervort,
2000; Bailey et al., 2002).
In the DLN 2 combustion system architecture, a single reaction zone is used in
a simplified liner with reduced cooled surface area per unit volume (Figure 10.53).
Highly efficient cooling, with a minimum of film cooling and seal leakage, was critical
to allow roughly 90 percent or more of the combustor through flow to enter through
the premixers as required for meeting the NOx target at the higher F-Class firing
temperatures. Film cooling is used only sparingly, at the interface between transi-
tion piece exit frame and the first-stage turbine nozzle, at the joint between the liner
discharge and transition piece inlet, at the interface between the cap and liner, and
in the array of effusion cooling holes through the hot surface of the cap. The liner
diameter at the cap end for the F-Class DLN liners is about 450 mm, substantially
larger in diameter than the DLN1 systems, but also shorter at 800 mm, providing
similar reaction zone volume and residence time at the higher F-Class cycle operat-
ing conditions.
352 Industrial Combustors
Figure 10.54. Premixers of the General Electric DLN 2.0 combustion system.
Unlike the DLN 1 primary stage in premixed mode, DLN 2.0 premixers are not
designed to tolerate flame in the premixing annulus. This requires careful tailoring
of the swirler and fuel injector aerodynamics. Extensive modeling and experimental
work resulted in the gas premixer geometry shown in Figure 10.54 with a large axial
air swirler followed by an array of fuel injection pegs immediately downstream. Ten
swirler airfoils and ten pegs are used in each premixer, with the premixing annulus
sized to produce a mixture velocity well above the worst-case turbulent flame speed.
Each premixer is equipped with two gas circuits, a diffusion circuit where fuel is
injected from the end of the diffusion tip through a series of angled jets directly
into the reaction zone, and a premixed circuit feeding ports in the pegs upstream.
The diffusion circuits were intended to be used from ignition to part load, when NOx
emissions compliance was not required, with the fuel switched off and purged con-
tinuously with cooled compressed air when in fully premixed mode.
The array of five identical gas premixers was used to supply the reactants to a
single combustor zone. Five premixers and the chamber count of fourteen for the
50 Hz machine and eighteen for the 50 Hz 9F matched the F-Class pressure ratio of
16:1 and flow per chamber of about 24 kg/s, at a turbine rotor inlet temperature of
about 1300C, with a combustor exit temperature of over 1400C. The fuel staging
required to maintain peak temperatures below the NOx threshold as firing tempera-
ture is increased from ignition through full load was accomplished by manipulat-
ing individual fuel circuits within each premixer, and by fueling different groups
of premixers. Primary, secondary, and tertiary gas fuel ring manifolds supply each
endcover.
During commissioning of the first engines with DLN 2 combustion systems,
quaternary fuel was added and other combustor modifications were made to con-
trol combustor pressure oscillations. The quaternary external fuel circuit was added
10.5 Case Studies 353
to inject a small percentage of the fuel, again via an array of short, hollow posts
arranged around the circumference of the casing pressure vessel. The quaternary
fuel peg array is fed via a manifold built into the forward casing, supplied by a ring
manifold. The pegs are just upstream of the head end of the combustor, at the axial
plane where the compressor discharge air exits the annulus formed between the
liner and flowsleeve, and the flow turns through 180 degrees and enters the inlet of
each gas premixer. Although typically much less than 10 percent of the total fuel,
the convective delay between the fuel injection point and the flame front associated
with the quaternary circuit proved very useful in managing dynamics across the load
and ambient range.
For dual-fuel versions of the DLN 2 system, each premixer is also provided with
an oil injector. Fuel oil and the water used for NOx suppression are directly injected
into the reaction zone from the center of each diffusion tip. The single oil circuit is
essentially a large pressure atomizer, provided with cooled, dried, higher pressure
atomizing air from an external compressor system for assisting atomization quality
at firing speed. Water is injected through a similar annular pressure atomizer, just
outboard of the oil circuit.
Regulatory pressures in the early 1990s led to the need to develop a 9 ppm
combustion system for the Frame 7FA. The result of this development is the DLN
2.6 (Davis 1996; Vandervort 2000). Reduction of NOx levels from the DLN 2 at 25
ppm to 9 ppm required that approximately 6 percent additional air was needed to
pass through the premixers in the combustor to achieve the maximum reaction zone
temperature that could be tolerated with a single-digit NOx target of approximately
1600C. This change in air splits was accomplished through reductions in cap and
liner cooling airflows, requiring increased cooling effectiveness. However, without
changes in the operation of the DLN 2.0 system, certain penalties would have been
incurred for achieving 9 ppm baseload performance. The turndown of a DLN-2 com-
bustor tuned to 9 ppm NOx and 9 ppm CO operation was estimated to be about
70 percent load, compared to 40 percent load for the 25 ppm NOx and 15 ppm CO
system.
For this reason, a new combustor configuration was conceived based on the
DLN 2 burner to overcome these difficulties. The key feature of the new configu-
ration is the addition of a sixth, smaller gas premixer located in the center of the
five existing DLN 2.0 premixers. The five outer nozzles are identical to those used
for the DLN 2.0, while the center nozzle is similar but with simplified geometry to
fit within the available space. The presence of the center nozzle enabled the DLN
2.6 to extend its turndown well beyond what the five-nozzle DLN 2.0 could attain.
By fueling the center nozzle separately from the outer nozzles, the fuel-air ratio
can be modulated relative to the outer nozzles, leading to approximately 100C of
turndown from baseload with 9 ppm NOx. Turning the fuel split down in the center
premixer does not result in any additional CO generation.
Absent any other changes in the DLN 2.0, other than the addition of the
center nozzle, the DLN 2.6 combustor would have required five fuel manifolds,
one more than the DLN 2.0. An alternative scheme was proposed to operate the
354 Industrial Combustors
Fuel
stage
sequence
Mode 1 Mode 2 Mode 3 Mode 4 Mode 6.3
Local fuel/air
Fuel/Air (ratio
Power output
PM1 PM2
distribution (%)
PM3
PM2 PM3
Fuel
Power output
Figure 10.55. General Electric DLN 2.6 combustor fuel staging sequence.
machine at startup and low load, which eliminated diffusion mode. The result was
a combustor with four manifolds: three premixed manifolds (PM) staging fuel to
the five outer and the single center premixers, and a fourth premixed manifold for
injecting the external quaternary fuel for dynamics abatement. The first manifold
fuels the center nozzle, the second manifold fuels the two outer nozzles located at
the cross-fire tubes, and the third manifold fuels the remaining three outer nozzles.
With the elimination of the diffusion circuits, the DLN 2.6 loads and unloads dif-
ferently than DLN 2.0. Additional mode changes are necessary to maintain the
premixed flames within their burnable zones and so prevent combustor blowout
(Figure 10.55).
The liner geometry of the DLN 2.6 combustor is completely conical, rather than
a cylindrical section followed by a short cone (Figure 10.53). The gas premixers have
also been updated with more aerodynamic vane-like fuel injection pegs replacing
the exclusively cylindrical fuel pegs used initially on the DLN 2.0. The DLN 2.6
system is capable of 9 ppm NOx and 9 ppm CO values at and above 50 percent load
with natural gas fuel.
In late 1996, a higher power output version of the Frame 9FA was introduced.
Called the 9FA+e, the cycle for this machine increased the air and fuel flow to
the combustion system by approximately 10 percent and the combustor exit tem-
perature to over 1470C. To meet the emission requirements for this engine, an
updated version of the DLN2.0, called the DLN 2+, was developed. The DLN 2+
retained the basic architecture of the DLN 2.0 with adaptations for both the new
requirements and to improve the operability and robustness of the existing system.
In comparison to the DLN2.0, the major changes are concentrated in the fuel noz-
zle and endcover arrangement. Both the endcover and fuel nozzle were supplied
with enlarged fuel passages for the increased volumetric flow of fuel, required to
accommodate weaker fuel gases. The DLN 2+ system introduced the swozzle gas
premixer, a combination of the words swirler and nozzle. The aerodynamics of this
DLN 2+ gas premixer, and subsequent swozzles applied on the derivative DLN
10.5 Case Studies 355
2.6+ and 2.5H systems, was redesigned for further improvements in flame-holding
margin, reduced pressure drop, and improved diffusion flame stability (Lewis et
al., 2011). The additional gains in flame-holding velocity margin resulted from
cleaner aerodynamics in the premixers. This was achieved via a new swirler design
that incorporates fuel injection directly from the swirler surface. Each swirler
vane comprises a turning vane and an upstream straight section. The straight sec-
tion is hollow and houses the fuel manifolds plus the discrete injection holes. This
contrasts with the earlier DLN 2.0 design with two elements in the flow path, a
swirler and pegs (later vanes) from which the gas fuel was injected. Upstream of
the swirler, an inlet flow conditioner improves the character of the flow entering
the premixer, while an integral outer shroud located downstream eliminates any
potential flow disturbances after the point of fuel injection. The nozzle-tip geom-
etry and the improvements in diffusion flame stability allow the use of a diffusion
flame on every nozzle.
This eliminates the lean-lean mode of the DLN 2.0 and resulted in the simpli-
fied staging methodology. The diffusion gas circuit eliminated in the DLN 2.6 was
reintroduced on the DLN 2+ and 2.5H to cope with weaker fuels.
Support
Pilot housing
C-stage
Transition
Basket
combination with the pilot nozzle. The typical loading sequence for the DLN com-
bustor is shown in Table 10.1.
The fuel is injected through four holes on the side of each of the main nozzles,
called rockets, just downstream of the main swirlers, through eight holes at the end
of the pilot nozzle, and through approximately thirty holes in the C stage ring. The
C stage fuel injection has a number of impacts on the combustion system. One key
feature of the C stage fuel injection is that it is well premixed with the incoming air-
flow. Therefore, the combustor can be tuned for combustor dynamics by shifting fuel
between the main nozzles and the C stage without any NOx penalty.
The liner and transition piece were initially cooled using air in plate-fin and
MT-fin geometries, respectively. These cooling systems consisted of two layers
of metal with the inner, thicker layer having a series of parallel cooling channels
machined into it. In the case of the liner, the cooling air exited directly at the end of
the channels and the inner and outer layers were only connected at the start of the
channel. There are a series of four of these panels with increasingly larger diameters
so that the cooling air exiting the first set of cooling channels forms a thin layer of
shield air at the start of the second panel and so on. Photographs of the DLN sup-
port housing and pilot nozzle are shown in Figure 10.57.
For the transition piece, the two layers are bonded together and each channel
is fed independently by an inlet hole and the cooling air from this channel exited
10.5 Case Studies 357
through an exit hole. The combination of the inlet hole and exit hole serves to regu-
late the cooling airflow to the channel.
As the system evolved, the liner was changed to have a rear side convectively
cooled panel, coated with a thick thermal barrier coating. In addition, the current
liner has an array of resonators mounted toward the aft end of the liner that serve to
damp out high-frequency combustor acoustic modes in the kHz range. Figure 10.58
shows pictures of the DLN baskets with a plate-fin cooled liner and with a rear side
cooled liner and resonators, respectively.
Improvements to the transition cooling arrangement have been more subtle and
involved optimization of the cooling channel arrangement and transition shape to
minimize heat transfer coefficients and maximize the mechanical robustness of the
design. Design modifications and optimization of the liner, transition, and seals of
these components have eliminated more than 50 percent of the cooling and leakage
air from the original design, enabling lower NOx and/or higher turbine inlet tem-
peratures at the same NOx levels.
Figure 10.59. Siemens ULN nozzle and support housing for can-annular combustors.
pilot nozzle, with both a premixed and a diffusion pilot stage. The premixed pilot also
injects the fuel through the swirler vanes.
The fuel staging strategy for the ULN combustion system is shown in Table 10.2.
The diffusion pilot stage is used predominantly for ignition and loading of the engine
and the premixed pilot is used predominantly at higher loads to control NOx emis-
sions to single-digit levels. Figure 10.59 shows pictures of the premixed pilot nozzle
and support housing.
The key advantage of the ULN system over the DLN system is that it allows the
fuel to be precisely placed within the premixing passage, allowing for an optimum
fuel profile in terms of circumferential mixing and radial profile. This radial pro-
file is tailored to ensure that the optimum pattern is achieved, not only in terms of
NOx levels, but also in relation to combustion dynamics, flashback, and lean stability.
Since the position of the injection holes, far more than the momentum of the fuel
jets, determines the mixing pattern in the ULN system, this system has the additional
advantage that it is less sensitive to changes in fuel temperature and Wobbe Index
than the DLN system.
In addition, the main stages are arranged in an AA-BB pattern, meaning that
two adjacent nozzles belong to the same stage. This evolution in the design results
in reduced CO emission level at loads when only the A-stage is active (below
30 percent load), reducing the startup emissions from the engine, as shown in
Figure10.60.
The ULN combustion system was first demonstrated in a trial operation run at
Renaissance power plant in Michigan in 2004 (Bland et al., 2004) and subsequently
was introduced to the market in 2008 at the Idaho Power, Evander Andrews project
(Johnson et al., 2008).
10.6 Nomenclature 359
normalized to peak
CO, corrected
15 ppm design
9 ppm design
10.6 Nomenclature
a m2/s thermal diffusivity
Da Damkhler number
f mixture fraction
f average mixture fraction
g variance of the fuel mixture fraction pdf
g standard deviation of the mixture fraction pdf
LHVvol MJ/m3 volumetric lower heating value
s normalized standard deviation
s1 m/s laminar flame speed
st m/s turbulent flame speed
T K temperature
U unmixedness parameter
Woi MJ/m3 Wobbe index
360 Industrial Combustors
Greek
equivalence ratio
mole fraction
kg/m3 Density
s characteristic time
Subscripts
Air air
Chem chemical
Fuel fuel
Mix turbulent mixing of fuel and air
Products combustion products
Reactants reactants, mixture of fuel and air
Unmixed entirely unmixed flow state of fuel and air
Abbreviations
HAP hazardous air pollutants
pdf probability density function
UHC unburned hydrocarbons
VOC volatile organic compounds
References
Aoyama, K. et al. (1991). Development of a Dry Low NOx Combustor for a 120MW Gas
Turbine. Proceedings of the ASME/IGTI TurboExpo, Paper ASME 91-GT-297.
Bailey, J. C., Intile, J., Fric, T. F., Tolpadi, A. K., Nirmalan, N. V., and Bunker, R. S. (2002).
Experimental and Numerical Study of Heat Transfer in a Gas Turbine Combustor
Liner. Proceedings of the ASME/IGTI TurboExpo, Paper 2002-GT-3018.
Bland, R., Ryan, W., Abou-Jaoude, K., Bandatu, R., Haris, A., and Rising, B. (2004). Siemens
W501F Gas Turbine: Ultra Low NOx Combustion System Development. Proceedings of
the Power-Gen International.
Bradley, D. (1992). How Fast Can We Burn? Twenty-fourth Symposium (International) on
Combustion, 24762.
Davis, B. (1996). Dry Low-NOx Combustion Systems for GE Heavy-Duty Gas Turbines.
Proceedings of the ASME/IGTI TurboExpo, Paper 96-GT-27.
Dean, A. M., and Bozzelli, J. W. (2000). Combustion Chemistry of Nitrogen, in Gas-Phase
Combustion Chemistry, Gardiner W. C. ed., Springer-Verlag.
Dbbeling, K., Pacholleck, J., and Hoffs, A. (2007). Combining Operational Flexibility with
Clean, Reliable Power Generation in the Alstom Gas Turbine GT13E2. Proceedings of
the Power-Gen Asia Conference.
Eroglu, A., Dbbeling, K., Joos, F., and Brunner, P. (2001). Vortex Generators in Lean-Premix
Combustion. Transactions of the ASME, Journal of Engineering for Gas Turbines and
Power 123: 419.
Eroglu, A., Flohr, P., Brunner, P., and Hellat, J. (2009). Combustor Design for Low Emissions
and Long Lifetime Requirements. Proceedings of the ASME/IGTI TurboExpo, Paper
GT200959540.
Gruschka, U. et al. (2008). ULN System for the New SGT58000H Gas Turbine: Design and
High Pressure Rig Test Results. Proceedings of the ASME/IGTI TurboExpo, Paper No.
GT200851208.
References 361
Hall, J. M., Thatcher, R. T., Koshevets, S., Thomas, L. L., and Jones R. M. (2011). Development
and Field Validation of a Large-Frame Gas Turbine Power Train for Steel Mill Gases.
Proceedings of the ASME/IGTI TurboExpo, Paper GT201145923.
Johnson, C. et al. (2008). Ultra Low NOx Combustion Technology. Proceedings of the
Power-Gen International.
Joos, F., Brunner, P., Schulte-Werning, B., Syed, K., and Eroglu, A. (1996). Development
of the Sequential Combustion System for the ABB GT24/GT26 Gas Turbine Family.
Proceedings of the ASME/IGTI TurboExpo, Paper 1996-GT-315.
Kalb, J. R., and Sattelmayer, T. (2006). Lean Blowout Limit and NOx Production of a Premixed
Sub-ppm NOx Burner with Periodic Recirculation of Combustion Products. Journal of
Engineering for Gas Turbines and Power 128: 24754.
Kenyon, M., and Fluck, M. (2005). Using Non Standard Fuels in the ALSTOM GT11N2 Gas
Turbine. Proceedings of the Power-Gen International.
Krebs, W., Walz, G., Judith, H., and Hoffmann, S. (1999). Detailed Analysis of the Thermal
Wall Heat Transfer in Annular Combustors. Proceedings of the ASME/IGTI TurboExpo,
Paper 99-GT-134.
Lewis, S., Thomas, S. R., Joseph Citeno, J., and Natarajan, J. (2011). F-Class DLN Technology
Advancements: DLN2.6+. Proceedings of the ASME/IGTI TurboExpo, Paper
GT201145373.
Lovett, J. A., and Abuaf, N. (1992). Emissions and Stability Characteristics of Flameholders
for Lean-Premixed Combustion. Proceedings of the ASME/IGTI TurboExpo, Paper
92-GT-120.
Lovett, J. A., and Mick, W. (1995). Development of a Swirl and Bluff-Body Stabilized Burner
for Low-NOx, Lean-Premixed Combustion. Proceedings of theASME/IGTI TurboExpo,
Paper 95-GT-168.
Matsuzaki, H. et al. (1984). Investigation of Combustion Structure Inside Low NOx
Combustors for a 1500C-class Gas Turbine. Transactions of the ASME, Journal of
Engineering for Gas Turbine and Power 106: 795800.
Payrhuber, K., Jones, R. M., Scholz, M. H. (2008). Gas Turbine Flexibility with Carbon
Constrained Fuels. Proceedings of the ASME/IGTI TurboExpo, Paper GT200850556.
Reiss, F., Griffin, T., and Reyser K. (2002). The ALSTOM GT13E2 Medium Btu Gas Turbine.
Proceedings of the ASME/IGTI TurboExpo, Paper GT 2002 30108.
Sattelmayer, T., Felchlin, M. P., Haumann, J., Hellat, J., Styner, D. (1992). Second Generation
Low-Emission Combustors for ABB Gas Turbines: Burner Development and Tests
at Atmospheric Pressure. Transactions of the ASME, Journal of Engineering for Gas
Turbine and Power 114(1): 11825.
Sattelmayer, T., Polifke, W., Winkler, D., and Dbbeling, K. (1998). NOx-Abatement Potential
of Lean-Premixed GT-Combustors. Transactions of the ASME, Journal of Engineering
for Gas Turbine and Power 120: 4859.
Senior, P., Lutum, E., Polifke, W., and Sattelmayer, T. (1993). Combustion Technology of the
ABB GT13E2 Annular Combustor. Proceedings of the Twentieth CIMAC Conference,
Paper G22.
Streb, H., and Prade, B. (2001). Advanced Burner Development for the Vx4.3A Gas Turbines.
Proceedings of ASME IGTI TurboExpo, Paper 2001-GT-0077.
Thomas, L. L., Simons, D. W., Popovic, P., Romoser, C. E., Vandale, D. D., and Citeno, J. V.
(2011). E-Class DLN Technology Advancements, DLN1+. Proceedings of the ASME/
IGTI TurboExpo, Paper GT201145944.
Vandervort, C. L. (2000). 9 ppm NOx / CO Combustion System for F Class Industrial Gas
Turbines. Proceedings of the ASME/IGTI TurboExpo, Paper 2000-GT-0086.
Index
363
364 Index
maximum available control technology policy, 103 Alstom LBtu combustor, 332
methane, 36, 99 Alstom MBtu combustor, 330
multi angle absorption photometry, 126 can annular, 323
multi nozzle quiet combustor, 326 design criteria, 303
municipal solid wastes, 71 General Electric systems, 323
Siemens designs, 331
national environmental policy act silo combustors, 328
military engines, 92 very compact combustor, 329
natural gas, 68 nozzle aerodynamics
new source performance standards, 102 design criteria, 54
July 2006 regulation, 103
nitrogen oxides organic Rankine cycles, 37
control Orimulsion, 65
landing fees, 93 oxyfuel combustion
post combustion applications, 210
selective catalytic reduction, 39, 85, 116, 192 carbon dioxide capture, 214
selective non-catalytic reduction, 192 combustor considerations, 224
water injection, 39, 43, 48, 76, 85 blowoff, 225
process modification flame temperature vs. residence time, 229
air staging, 193 emissions, 225
lean direct injection, 192 carbon monoxide, 226
lean premixing, 40, 85, 191 post-flame relaxation, 228
lean prevaporized premixed, 192 molecular oxygen, 227
reburning, 194 requirements, 214
emission criteria, 101 integrated gasification combined cycle, 225
concentration vs. output based, 101 oxyfuel combustion vs. exhaust gas
emissions recirculation, 214
exhaust gas recirculation, 21624
lean direct injection, 263 partial premixing, 261
lean premixed combustion, 238 particulate matter
prevaporization, 242 break-up vs. erosion, 137
formation, 5, 183, 200 carbonization rates, 141
effect of equivalence ratio, 184 coalescence, 138
effect of pressure, 195, 198 collision frequencies, 138
burner stabilized premixed flame, 197 Smulochowski equation, 138
counterflow diffusion flame, 196 particle size distribution, 139
fuel-air mixing, 197 method of moments, 140
lean premixed swirl burner, 199 monodisperse distribution, 139
other factors, 196 quadrature method, 140
fuel source, 188 sectional method, 140
N2O pathway, 186, 305 rates, 139
NNH pathway, 187 control, 1446
non-premixed flames, 306 lean premixing, 143
pathways, 183, 189 dilution effects, 145
premixed flames, 307 formation process, 130
effect of mixing quality, 311 modeling issues, 132
perfectly premixed, 308 fuel effects, 142
effect of pressure, 308 gas vs. PAH laden, 143
effect of residence time, 309 jet fuel content, 143
turbulent flames, 310 physical vs. chemical properties, 143
prompt mechanism, 186, 305 global warming, 124
thermal mechanism, 184, 305 ground vs. aero engines, 147
jet fuel ignition, 191 impact of combustor design, 145
jet methane reduction, 83 inception, 132
jet ozone formation, 82 aromatic ring formation, 133
nitrous oxide, 97 multi-ring aromatics, 133
world bank guidelines, 110 pyrene-pyrene dimerization, 134
non-premixed combustor mass production rate, 134
Alstom designs, 327 indirect control, 124
368 Index